This Article introduces the concept of amplification into First Amendment law. Amplification, or the size of the audience reached by speech, lies at the heart of many contemporary free speech struggles. Yet the concept is surprisingly absent as a category of analysis from constitutional doctrine and virtually undiscussed in legal scholarship. Amplification deserves its own set of legal rules and doctrines, because the right to amplify one’s speech serves the two core types of First Amendment interests—those of audiences and those of speakers—differently than the right to choose the content of one’s speech. The higher the degree of amplification, the greater the disparity. When it comes to audience interests, amplification via mass media platforms has unique potential to distort the marketplace of ideas that informs voting audiences. When it comes to speaker interests, greater amplification has only diminishing marginal returns for the speaker’s primary interest in autonomy, understood as the capacity for living one’s own life, because speakers need very large audiences neither to (a) form their own life plans nor (b) have the motivation to act on them. Thus, the right to amplify speech to very large audiences is justified by its benefits for audience interests rather than speaker interests, and so may be constitutionally regulated to preserve the integrity of democratic discourse for audiences. A central practical upshot is that certain carefully drafted legal rules on amplification, including campaign finance laws and social media regulations, should survive constitutional scrutiny.
Introduction
A man walks into a courthouse wearing a jacket embroidered with the words “F*** Big Oil,” in order to protest a federal permit recently granted to a multinational corporation for building an oil pipeline. Later that day, the corporation’s CEO pays her media consultant $100,000 to place local advertisements on billboards, television, radio, and YouTube about the jobs that the pipeline will generate.
Under current First Amendment law, these two speech acts have identical constitutional status. Both express views in public on matters of public concern. Both are robustly protected against government meddling. It is commonly thought, even, that a central strength of the First Amendment is its blindness to other details about the acts. This Article, however, argues that the First Amendment should not remain blind to one difference between these acts: their amplification, or the size of their audience. The ad payor, unlike the jacket wearer, is using a media platform to amplify her speech to be heard by a potentially massive audience.
The history of free speech is, in many ways, a history of the struggle for control over the means of amplification, from printing presses1 to megaphones2 to electromagnetic waves.3 The struggle continues even in this era of YouTube and Facebook, because, while over four billion humans have instant access to tools of self-publication,4 the new means of amplification are algorithms that draw attention toward some users’ content but not others’.5 As the system of amplification and its inequalities become more transparent, inquiring minds outside of law have begun to wonder whether the freedom of speech includes a “freedom of reach.”6
Yet First Amendment law remains oddly silent about amplification. The doctrine contains categories, and separate lines of analysis, for other speech properties, such as time, place, and manner, or content. But amplification is only mentioned in passing and is not treated consistently across cases. Constitutional scholars have not written about the theory of amplification in any detail, either.7
Surely there must be a right to speak to other people, or else free speech rights would have little value. The law could not restrict the man in our example to wear his jacket only at home. But to how many people does one have a right to speak? This Article starts the process of thinking through this question.
Some might argue that, in principle, speakers are entitled to as large a willing audience as they can buy or otherwise acquire through their private resources. Arguably the Supreme Court adopts this view of amplification rights in some cases. The view—call it the libertarian view—seems facially plausible because of the venerable tradition of the First Amendment as guardian of individual liberty. Speech rights, like all rights, preserve for individuals control over some zone of our lives, despite the necessary compromises of communal living. The libertarian view of amplification implies that having control over not only what we say, but exactly how we say it and to whom we say it, is part of the bargain.
This Article argues, to the contrary, that the right to amplify speech is—even in principle—restricted: at the highest levels of amplification, the right obtains only insofar as it is consistent with the basic foundations of democratic discourse.
Understanding why takes a closer philosophical look at the relationship between amplification and the core values that undergird the First Amendment. The Free Speech Clause has a split purpose: serving both the (a) individual speaker and (b) overarching system of democratic discourse.8 In most cases, this split goes unnoticed because, ordinarily, both individual and structural values are served simultaneously: while the speaker expresses herself, democratic discourse gains a new idea or perspective. But the decision to amplify speech works differently. As speech reaches larger and larger audiences, it has a smaller impact on the speaker’s own interests, properly understood, and has a greater impact on democratic discourse. Past some threshold audience size, the right to amplify speech is justified almost entirely because of its benefits to democratic discourse rather than speakers.
I do not mean that mass-amplified speech promotes none of a speaker’s interests. The argument rests on a distinction between speakers’ interests that lie at the core of the constitutional guarantee and those that do not. The core speakers’ interests are autonomy and political participation. It is these core interests—which ultimately account for the protection of a very broad range of speech—that matter for the purposes of constitutional interpretation. But large-scale amplification is necessary for neither.
Begin with autonomy. Autonomy is the capacity for living one’s life according to one’s own reasons. Constitutional doctrine emphasizes two components of autonomy: (a) the ability to freely form one’s own mind, including one’s beliefs and identity; and (b) the opportunity for one’s beliefs and identity to be recognized and affirmed by others. Generally, speaking freely to an audience is crucial for both aspects of autonomy. However, this Article makes the case that the best audiences for these purposes have at least some of the characteristics of diversity, responsiveness, familiarity, and supportiveness. But past some threshold audience size, just adding listeners does little to enhance these characteristics, and can actually undermine them. So amplification past that threshold has only diminishing marginal benefits for autonomy.
Political participation is the interest in taking part in democratic self-governance. For this interest to be meaningfully fulfilled, we must be able not only to talk to our family and coworkers, but to protest, post on social media, and otherwise join en masse with other speakers. But it is not so clear why a right justified by this interest would be unlimited. If the right is to take part in a democratic process, then it is unclear how the justification could support anything more than a claim to equal amplification for each citizen—and we cannot all be heard by millions. Political participation justifies my shouting on the street corner but not my shouting everyone else down, too.
This leaves only structural values to justify large-scale amplification rights. Democracy is the chief such value found in the doctrine.9 Speech rights can foster a vibrant marketplace of ideas that serves two functions in the democratic process: (a) informing voters and (b) legitimizing political decisions. But a long tradition of scholarship suggests that not just any marketplace of ideas will effectively produce these benefits.10 Even the Supreme Court seems to recognize that the marketplace, to function adequately, needs three crucial features: diversity, meaning that a wide range of views are heard by most citizens; mobility, meaning that views have some realistic chance of being heard and—depending on their inherent appeal—moving upward in public discourse; and at least occasional antagonism, meaning that views are presented in passionate and responsive exchanges. I call these the preconditions of epistemic competition, because they make it likelier that truth rises to the top of public discourse.
Mass-amplified speech is uniquely positioned to disrupt these preconditions. To be amplified to a large audience, speech needs the right platform. Yet mass media platforms are scarce, and access to them is enormously (and not randomly) unequal. Those who speak on them can crowd out other speakers and have little need to respond to lesser-amplified opponents. So, if every speaker is entitled to the largest audience money can buy or connections can broker, then democracy may significantly suffer. If amplification rights are grounded primarily in democratic values, then it would be a “topsy-turvy” world in which those rights are interpreted to destabilize the very conditions that enable effective democratic discourse.11
In other words, the right to amplify speech to very large audiences should be subject to a democratic qualification. The First Amendment should protect the oil CEO’s ability to amplify her speech to, say, hundreds of thousands of listeners only if her and others doing so facilitates rather than undermines epistemic competition—e.g., diversity, mobility, and antagonism—in the marketplace of ideas. If this analysis is correct, then regulations on mass-amplified speech that, narrowly and carefully, seek to ensure a functioning marketplace of ideas may well survive constitutional scrutiny.
This conclusion ripples across First Amendment law. Many free speech cases turn out to be mass-amplification cases upon closer examination. For instance, campaign finance laws regulate campaign spending, which largely goes toward amplifying political advertisements over mass media. The Supreme Court has struck down an array of such laws—from caps on campaign donations12 to bans on corporate campaign advertising,13 most notably in Citizens United v. FEC, to specialized public financing schemes14—under the strictest form of constitutional scrutiny. But if these regulations are understood as narrowly seeking to promote epistemic competition, then these cases may have been wrongly decided. In addition, how we think about mass amplification matters for a range of cases involving regulations on mass media platforms, including net neutrality rules recently before federal courts and proposed social media regulations.15
A defense of such regulations that focuses on the degree of amplification regulated offers theoretical and doctrinal advantages over the two other central democratic lines of defense in recent literature.16 The first such line is egalitarian-democratic, interpreting the First Amendment to encapsulate principles of equality among speakers as democratic citizens.17 Doctrinally, these defenses face an uphill battle, because the Court has already firmly rebuffed such egalitarian principles as antithetical to the First Amendment.18 I join a second, and older, scholarly line that is epistemic-democratic, interpreting the First Amendment to require a marketplace of ideas that actually (rather than theoretically) informs voters. My focus is on diversity among ideas and views, rather than equality among speakers. However, proponents of epistemic-democratic views have traditionally sidelined the First Amendment speaker interests that the Court holds dear.19 An amplification-based view takes these claims seriously. It offers a justification for embracing epistemic-democratic principles, with an explanation for why those principles triumph over individual liberty in some contexts (those of mass amplification) but not others (those of minimal amplification).
The Article proceeds in four Parts. In the first, I introduce the concept of amplification and explain constitutional doctrine’s implicit but inconsistent treatment of it. In the second, I explain how massively amplified speech can serve the democratic process but can—though the Court has so far missed it—also undermine the epistemic reliability and legitimacy of that process. In the third Part, I canvass the speaker interests with constitutional status—autonomy and political participation—and conclude that amplification offers only marginal benefits for the former and only slightly greater benefits for the latter. In the final Part, I explore how my analysis of amplified speech offers a justification for the constitutionality of certain campaign finance and media regulations.
I. An Introduction to Amplification
When you think of a speech act,20 you might think first of its content, or the message conveyed in its words or symbols. Content is the core of speech’s power, because it is what most reliably changes listeners’ minds and even conduct. Perhaps because of content’s power, it is also the feature of speech most often suppressed by governments trying to preserve the political or cultural status quo.
First Amendment law now fiercely protects content, more than any other feature of speech, against government suppression.21 Laws singling out specific content for restriction are—with few exceptions22—subjected to the highest bar of constitutional review, strict scrutiny.23 This demanding test requires that a law must “further[] a compelling interest” and be “narrowly tailored to achieve that interest.”24 The test applies even when the content is indisputably false.25 And it applies not just to one’s general message but to one’s specific choice of words.26
But a speech act is also more than its content. It is an event, rooted in a concrete context. It is uttered by a speaker, in a certain time, place, and manner. The Gettysburg Address, to take a well-known example, was delivered by Abraham Lincoln in the midst of the U.S. Civil War, on August 28, 1863, at the dedication of a new military cemetery in Gettysburg, Pennsylvania, the site of one of the bloodiest battles of the war. This context matters for understanding the act.27 The 271 words of the Gettysburg Address would take on different meanings if spoken by Lincoln on the same battlefield in 1865, by a protestor in Hong Kong in 2019, or by a Black Lives Matter activist in 2021.
Another aspect of a speech act’s context is its audience. While technically we can speak to ourselves, in a diary or voice memo, ordinarily we speak in order to communicate to others. Indeed, we often intend to speak to specific others. Just as the speaker or the time, place, or manner of a speech act can change its perceived message, so, too, can the audience to whom it is uttered. If Lincoln had been speaking to the members of Congress, gathered before him, rather than 15,000 citizens of the Union, his words would have been understood differently.
Some of these non-content aspects of speech are protected by the First Amendment.28 Speech may not be regulated based on the identity of its speaker (or even based on the fact that the speaker is a corporation or union).29 Nor may it, without adequate justification, be restricted in its time, place, or manner.30
The First Amendment also, arguably, guarantees a right to an audience. However, that does not mean a right to be heard by exactly the audience one hopes for; audiences have their own rights not to listen. They can typically do so by averting their eyes or covering their ears.31 And when they cannot do so with relative ease, such as when speech is mailed or broadcast directly into their homes, they are protected as “captive” audiences.32 Nonetheless, constitutional doctrine implies that speech rights are meaningless without the opportunity to reach an audience.33 As the Supreme Court expresses it, “[t]he right to free speech, of course, includes the right to attempt to persuade others to change their views.”34
This Article zeroes in on one characteristic of speech’s audience: its size. Amplification is the term I use for the size of the audience of a speech act. To increase the audience size of a speech act is to “amplify” it.
How does and should the Constitution protect amplification? Those are the overarching questions of this Article. This Section tackles the description question. To begin, it introduces readers to the tools speakers use to amplify their speech to the largest audiences—often with great difficulty and cost—in the contemporary media environment. Then it explains how those tools of amplification are currently protected under constitutional doctrine. In short, the Constitution contains an implicit right to speak to the largest willing audience that one can acquire through private means.
A. Amplification Mechanics
Most speakers will attempt to amplify their speech at some point. They will post messages on Twitter, rather than just talking to family and close friends. They will seek to publish their book or article, rather than just circulating it within their network. They will advertise their couch for sale on Craigslist, instead of just by word of mouth. Increased amplification brings the increased potential to change the beliefs and conduct of others, to find persons with a mutual interest, and to raise one’s own profile.
So how do we do it? Increasing amplification involves drawing attention to speech.35 I will name five methods here; these can be employed alone or in combination. First, attention can be drawn by changing the content of speech to be more intriguing or shocking, like a newspaper changing its headlines into clickbait. This method is easiest for speakers to control, but, on its own, has minimal amplification potential; it works best only in certain contexts or to augment an already large audience. A comment posted to a Twitter account with three followers will almost certainly not achieve much amplification, no matter how incendiary it is. Second, the speaker can change the manner of speech to be eye- or ear-catching. For instance, speakers on a street attract attention using a bullhorn or flashing lights. Third, mass attention can in rare cases be drawn by the speaker’s identity—e.g., if they are a celebrity, or occupy a temporary position of importance. Fourth, attention can be drawn by popularity cascades. An orator surrounded by a small crowd may draw a still larger one; or a YouTube video initially liked by thousands may eventually “go viral” and be seen by millions. But while popularity cascades can potentially draw heaps of attention, they are unpredictable. Fifth, the speaker can place the speech in a venue at which attention already pools, such as a newspaper or television channel. Call these venues amplifying platforms. The degree of amplification that results depends on the size of the platform’s existing audience.
Notice that amplification is almost always a function of other properties of speech, such as its content, speaker identity, or “time, place, or manner.” All of these properties are inextricably linked. Why should anyone bother treating amplification as a distinct property? One reason is that different degrees of amplification (whatever the associated content and context) share distinctive qualities. For instance, the larger amplification is, the more it tends to be motivated by a desire to influence—whether to sell products, achieve fame, or change hearts and minds—and to in fact bring about that influence. In these respects, a speech in a printed copy of the New York Times has more in common with a speech uttered on CNN than with a speech printed in the physical pages of the United Teachers of Wichita monthly newsletter.
Occasionally speakers seek, sometimes in combination with others, to amplify speech on a very large scale, in order to influence public opinion as a whole and thereby bring about social or political change. To do so, they need enough amplification to at least affect the subjects and arguments heard by a significant portion of the targeted public. I will call this degree of amplification mass amplification. I do not want to tie mass amplification to an exact number of listeners, because that number is more of a range and will vary by population size and other circumstances. Just to provide a concrete anchor for thinking, I will very roughly estimate that, if the target population is the United States, mass amplification means at least a few million people.36 Mass amplifying platforms include well-known media sources like, inter alia, the New York Times, the Wall Street Journal, Fox News, CNN, and certain YouTube channels.37
Mass amplification does not come easily. Of the five methods mentioned above, the only one that can reliably draw mass attention for the noncelebrity speaker is the use of amplifying platforms—and only then the highest-traffic platforms. Yet these platforms are scarce, relative to the number of potential speakers. Few media sources can attain truly mass reach;38 those that do often achieve quasi-monopoly status.39 Attention tends to pool on a subset of platforms due to two primary phenomena: network effects and cognitive scarcity. Network effects occur when platforms acquire high status in a community, drawing in new users and retaining others who do not want to miss out.40 Cognitive scarcity—limited time and attention span—ensures that few audience members seek out additional media sources.41
Mass-amplifying platforms are not only scarce; access to them is unequal. Most mass media organizations are privately held corporations. While these corporations possess a variety of motivations, they primarily seek to raise advertising revenue. They do this by expanding their audience and increasing existing audiences’ engagement with and time consuming their content—all of which present additional advertising opportunities. To borrow Tim Wu’s phrase, media corporations are “attention merchants.”42 This often means choosing to amplify speakers who will either attract audience attention or pay directly. Those who attract attention tend to be public figures or celebrities, or say incendiary things. Fox News, in particular, is famous for its media model that maximizes emotionally inflammatory content.43
For the non-famous citizen, payment is the most reliable route through these platforms’ gates. But prices for advertising over mass platforms are well beyond the means of the average citizen; and the larger the audience reached, the more exorbitant the price. Just to offer a quick survey, it costs about a dollar per click to advertise on Facebook, over $400,000 for a thirty-second ad during a popular television show, and about $1,000 per column inch to advertise in the New York Times.44 These expenses start to add up once one realizes that effective amplification requires repetition.45 To reach most of the viewers of one platform, one must pay for multiple spots. Moreover, to maximize one’s amplification, one needs to pay for multiple spots on multiple platforms.
One might think that in the Internet era these scarcity problems would wane. Free and equal access to social media makes everyone, in a sense, their own self-publisher or self-producer. But technology cannot remove the cognitive and sociological drivers of mass media scarcity, and so, while more speech is formally published today than in the past, proportionately less of what is published is actually heard. Mass-amplification scarcity is thus not a passing contingency but a deep and enduring feature of human life as we know it.
What has changed with the Internet are the loci of mass attention. On the Internet, one does not gain mass attention simply by posting speech on Facebook or Twitter, as one does by having one’s speech simply appear on a popular television channel. One’s speech also needs to be selected by the algorithms that “promote,” or draw attention to, some over other speech on the same platform.46 Algorithms promote content in two main ways. One is by placing content in high-visibility locations, such as in the margins or banner of a webpage. (Money, of course, can also often be used to buy these spots.) The second is by raising the “rank” of content so that it appears at the front or top of the primary information flow seen by each user, whether a Facebook news feed, a Google search result list, or a YouTube recommended videos list. While such rankings cannot always be directly bought, sophisticated speakers, such as political campaigns, are pouring money into “beating” algorithms.47
What this means is that a speaker can rarely achieve a significant amount of amplification on her own. To do so, she needs access to one of a few selective, privately controlled platforms or algorithms. In later Sections, I will use this fact about mass amplification to argue that constitutional law should protect the use of mass amplifiers differently from the use of other amplifiers. But first I explain how amplification is currently protected.
B. Amplification in First Amendment Doctrine
Constitutional doctrine’s official approach to amplification might be described as “amplification blindness.” The Supreme Court, at least, almost never directly discusses audience size. As Justice Douglas put it in a rare opinion that explicitly mentions the issue, “the size of the audience has heretofore been deemed wholly irrelevant to First Amendment issues. One has a right to freedom of speech whether he talks to one person or to one thousand.”48 The explanation for this blindness is fairly simple: the speaker’s claim in litigation is never to directly reach an audience of size n. Instead, the claim is to use a particular means to assist in amplification, which only indirectly and imprecisely corresponds to an estimated audience size. If the First Amendment protects access to an amplifier, then it will in effect protect access to whatever size of audience the amplifier happens to reach. (Of course, this is only protection against government intervention in amplification, not a guarantee of amplification; the speaker must acquire her “soapbox” or “television frequency” on her own, through private means.) In other words, amplification blindness—like so-called color blindness—usually most benefits the most privileged.
But amplification blindness is only the Court’s approach in principle. In practice, not all degrees of amplification are treated the same for First Amendment purposes. But the wrinkle is certainly not to the advantage of the lesser-privileged: it is the higher levels of amplification that receive the greater solicitude. The Court never expressly states this. Instead, how much protection use of an amplifier receives depends on whether it is employed in a publicly or privately owned space—which in turn often correlates with the level of amplification it provides. The speaker’s choice to use the widest-reaching amplifying platforms, like newspapers or television, is treated as a core part of the freedom of speech. Regulations on use of these platforms are subject to the highest constitutional standard, strict scrutiny, and rarely survive. By contrast, the speaker’s choice to use low-level amplifiers in public forums, like sound trucks and signage, is viewed as peripheral to the speaker’s free speech interests. Regulations on them are subject only to the lower intermediate scrutiny and usually stand.
The result, painting with broad strokes, is that the right to amplify is most robust when one is trying and able to reach the largest numbers. This Part illustrates the Court’s formal amplification blindness approach and implicit preferential treatment of higher levels of amplification. In Parts II and III, I will explain why the Court should take greater explicit notice of the degrees of amplification offered by different amplifiers. The Court gets its implicit normative analysis exactly backwards: the biggest amplifiers are the ones that should be most easily regulable.
1. Amplification in Public Forums
The one constitutionally guaranteed opportunity for amplification—the sort that anyone, irrespective of their resources, must be granted—resides within public forum doctrine. The State must hold open to all access to public forums, or spaces historically reserved for “assembly, communicating thoughts between citizens, and discussing public questions.”49 Public forums are usually (though not necessarily) publicly owned spaces, such as roads, sidewalks, parks, and squares in which pedestrian or vehicular traffic is common.50 No speech may—with rare exceptions—be outright banned in them.
Public forum access is not granted just so that friends can talk in the fresh air. The Court describes public forums as places to engage in persuasion and influence within the broader community—i.e., to take at least some small part in the formation of public opinion. For that, one needs to speak to those with whom one disagrees.51 In addition, while audience size is not often mentioned, it is implied: to take part in the formation of public opinion, you want not just to successfully change minds but to change many minds. And public forums often provide access to crowds.
Yet the right to access public forums usually provides only limited amplification. Audiences there consist primarily of passersby. On an average day, the number that can be expected to pass by is not large—especially compared to the number of minds one must change to bring about major social change.52 It is therefore no surprise that public forums are primarily used for amplification by relatively small and poor grassroots movements—which may not be able to afford other amplification strategies.
But even the small potential audience within a public forum can rarely be reached in its entirety by one speaker. The portion of that audience reached depends on how much attention the speaker can draw to her speech, and for how long. In other words, it depends on her use of additional amplifiers. A person shouting from a soapbox, as nineteenth-century orators did, may catch a few extra ears during the hours her voice lasts. Someone with a megaphone or someone who can post durable signs will probably do better.
Yet the use of these intra-public forum amplifiers is not itself constitutionally guaranteed. Doctrine ordinarily classifies these amplifiers as the time, place, or manner (TPM) of speech. This classification makes analytical sense. As described in the last Section, nearly all amplifiers operate by affecting time, place, or manner; in a public forum, this usually means raising volume or visibility. But with this classification comes a demotion in constitutional status: limitations on TPMs, unlike content, are constitutionally permissible. So a city’s ordinance forbidding the posting of signs in certain public areas may well be upheld. The speaker’s choice of her speech’s TPM is thought to be much less important than her choice of its content.53 TPM rules are still subject to constitutional constraints: they must pass the intermediate level of constitutional scrutiny and, as such, must “restrict[] no more speech than necessary” to achieve a legitimate government interest, and leave open “ample alternative channels for communication.”54 But they ordinarily do pass, so long as the government offers an adequate justification.55 The prevention of social disruption—in the form of noise or aesthetic blight—is a commonly cited justification, and it is usually deemed adequate.56
Amplification might still be safeguarded even from TPM restrictions if, in the last prong of the constitutional test, “ample” alternatives had to offer similar levels of amplification to the time, place, or manner restricted. But they do not. Courts seldom deem an alternative channel inadequate just because it reduces potential audience size. For instance, in Ward v. Rock Against Racism, a New York City rule required performers using the Central Park bandshell to use sound-amplification equipment and a sound technician provided by the city—partly in order to keep sound levels down.57 Rock Against Racism was hosting their annual, multihour concert, said to reach 50,000 residents, and wanted to use their own technician.58 The Court upheld the city rule, on the ground that the city was “protecting its citizens from unwelcome noise.”59 According to the Court’s opinion, “[t]hat the city’s limitations on volume may reduce to some degree the potential audience for respondent’s speech is of no consequence, for there has been no showing that the remaining avenues of communication are inadequate.”60 All that mattered was that “the guideline continue[d] to permit expressive activity” and had “no effect on the quantity or content of that expression”—the more highly valued properties of speech.61
The Court also almost uniformly dismisses a stark reality of amplification in public forums: that the most effective methods of amplification within those forums also take more resources. For instance, in Kovacs v. Cooper, the City of Trenton, New Jersey, had passed an ordinance prohibiting the operation of sound amplifiers affixed to vehicles that emitted “loud and raucous noises.”62 The defendant in the case was charged under the statute for using a so-called “sound truck” to comment on an ongoing labor dispute.63 The Court upheld the regulation as protecting against “distractions . . . dangerous to traffic at all hours” in business streets and promoting “quiet and tranquility” in residential areas.64 In addressing the alternative channels of communication, the Court found them adequate because “[t]here is no restriction upon the communication of ideas or discussion of issues by the human voice, by newspapers, by pamphlets, by dodgers.”65 In doing so, it observed “[t]hat more people may be more easily and cheaply reached by sound trucks, perhaps borrowed without cost from some zealous supporter, is not enough to call forth constitutional protection for what those charged with public welfare reasonably think is a nuisance when easy means of publicity are open.”66
In sum, the right to amplification within public forums is only weakly guarded in First Amendment doctrine. Unlike the content of speech, it is seen as a fairly peripheral property of a speech act that can be easily sacrificed to the right social objectives. Alternatives that reduce amplification significantly or are prohibitively expensive for the speakers in question are often accepted as adequate. This will prove a major contrast with the doctrine’s treatment of amplification examined in the next Section.
2. Mass Amplification: Campaign Finance
When mass media is involved, amplification is protected more like content, as an indispensable aspect of a speaker’s speech. Courts show more sympathy here than they do in the public forum context with the speaker’s right to choose her degree of amplification—at least to the extent that she can afford it. The clearest example of this appears in cases challenging regulations of spending on campaign speech.
In 1972, Congress passed the Federal Election Campaign Act (FECA), the most comprehensive attempt in American history to regulate money spent on political campaigns.67 Among other things, FECA placed strict dollar limits on expenditures for campaigning speech.68 In the seminal case Buckley v. Valeo, the Court interpreted the First Amendment to forbid any such caps, unless the funded speech was coordinated with a political candidate.69 Indeed, any limits on campaign expenditures would be required to clear the highest constitutional bar, strict scrutiny, which demands that a law must “further[] a compelling interest” and be “narrowly tailored to achieve that interest.”70 Since Buckley, regulations of expenditures have been struck down under this standard as unconstitutional, including caps on expenditures by a self-financing (wealthy) candidate,71 corporate and union expenditures in favor of a candidate,72 and payments to petition gatherers for ballot initiatives.73 The Court has on similar grounds invalidated some schemes for publicly financing candidates, as explained below.74
The Court’s primary concern about expenditure limits is that they “necessarily reduce[] the quantity of expression.”75 Quantity can mean the number of unique campaign ads produced.76 But it can also mean—and the Court seems, the majority of the time, to take it to mean—the number of listeners—i.e., amplification—for any given ad. When the Court discusses expenditures, it ordinarily contemplates their application to amplification. As the Court said in Buckley:
[V]irtually every means of communicating ideas in today’s mass society requires the expenditure of money. The distribution of the humblest handbill or leaflet entails printing, paper, and circulation costs. Speeches and rallies generally necessitate hiring a hall and publicizing the event. The electorate’s increasing dependence on television, radio, and other mass media for news and information has made these expensive modes of communication indispensable instruments of effective political speech.77
Similarly, the Court states in a later case that “allowing the presentation of views [in a national presidential campaign] while forbidding the expenditure of more than $1,000 to present them is much like allowing a speaker in a public hall to express his views while denying him the use of an amplifying system.”78 This focus is consistent with a great deal of research confirming that most money spent on campaigning goes to fund airtime or the equivalent.79
The description of amplification as quantity—a classification that the Court expressly denies amplification in public forums80—suggests that more amplification is more speech. That is, the total number of speech acts funded by an expenditure depends on the number of listeners. One might, alternatively, think that a speech amplified broadly is just one speech act with many listeners. This distinction matters, because it sets the stage for arguing that one speech amplified to a thousand has the same constitutional weight as a thousand separate speeches. Indeed, the Court seems not to see much difference when it comes to campaign finance. As the Court says in a famous footnote in Buckley, “[b]eing free to engage in unlimited political expression subject to a ceiling on expenditures is like being free to drive an automobile as far and as often as one desires on a single tank of gasoline.”81 In cases in which amplification is conceived of as quantity rather than TPM, the normative value of amplification is also rated highly.
But we see a very different model of amplification in the TPM cases. There, the Court takes the opposite view described above: it sees the amplification of a speech as just one speech act. As a result, the Court treats the amplification as less integral to the speech act, less on a par with content. What matters is being able to get your message out, not getting it out to the maximal number of persons.
We can see further just how central the idea of amplification-as-quantity is to the Court’s analysis of campaign expenditures by contrast with the Court’s treatment of another sort of campaign spending that it deems not to involve amplification: contributions. The original FECA legislation capped not just “expenditures,” or money spent directly on campaigning speech of one’s own, but also “contributions,” or individual and corporate donations to candidate campaigns or campaign committees. For instance, if I spend $10,000 to make and air my own thirty-second ad in favor of a candidate, then I have made an “expenditure” under federal law; but if I send $10,000 directly to my candidate, then I have made a “contribution.” In Buckley, the Court drew a firm constitutional line between the two categories: while Congress could not cap independent expenditures, it could cap contributions.82 Contribution regulations are, unlike expenditure caps but like TPM rules, subject to the low bar of intermediate constitutional scrutiny.83 Since Buckley, caps on contribution limits are expected to survive unless they are set very low.84
Contributions, in the Court’s view, do not involve amplification. What contributions do is best explained by the Court’s oft-cited description of a contribution in Buckley:
A contribution serves as a general expression of support for the candidate and his views, but does not communicate the underlying basis for the support. The quantity of communication by the contributor does not increase perceptibly with the size of his contribution, since the expression rests solely on the undifferentiated, symbolic act of contributing. At most, the size of the contribution provides a very rough index of the intensity of the contributor’s support for the candidate.85
In other words, a contribution is, whatever its amount, just one speech act. It expresses association and does not communicate much more with any precision, so the amount of money contributes only minorly to the content of the act. As a result, a ceiling on contributions “does not in any way infringe the contributor’s freedom to discuss candidates and issues”86 and “entails only a marginal restriction upon the contributor’s ability to engage in free communication.”87 Contributions, in other words, are far from the core of the free speech right.
Here we see the parallel with the TPM cases. There, the Court sees the use of a given amplifier as just one manner, among many, of speaking; and nearly any alternative amplifier is adequate. Similarly, the Court also sees almost any alternative to contributing money for speaking about election campaigns as adequate. Buckley describes contribution limits as “leaving persons free to engage in independent political expression, to associate actively through volunteering their services, and to assist to a limited but nonetheless substantial extent in supporting candidates and committees with financial resources.”88 As in the TPM cases, the Court is relatively unfazed by the fact that these methods of speaking vary drastically in their expected audience size.
Yet the Court’s distinction between contributions and expenditures is largely a fiction. Both forms of spending are primarily used to increase amplification. One can either pay a team employed by the candidate’s campaign to create an ad and then air it many times, or one can pay one’s own team to create an ad and then air it many times. Indeed, Justice Thomas, who has for decades opposed the description of contributions as mere associative acts, clearly sees how they amplify speech.89 In his words, “[c]ontributions do increase the quantity of communication by ‘amplifying the voice of the candidate’ and ‘help[ing] to ensure the dissemination of the messages that the contributor wishes to convey.’”90 As a result, it is hard to see why the distinction between contributions and expenditures has the normative significance that it does.
The Court’s commitment to the individual’s freedom to buy amplification over mass media has only deepened over the last decade. Now the Court finds that freedom is threatened not just by expenditure caps but also when expenditures come with burdens, like state funding for a poorer opponent. As a result, public financing schemes that in any way tether the public funding levels or contribution/expenditure limits of one candidate to the expenditure of money by a privately and better-financed opponent are constitutionally vulnerable. For example, in Davis v. FEC, the Supreme Court struck down the Millionaire’s Amendment to FECA, which tripled federal contribution limits for candidates whose opponent declared an intent to spend over $350,000 of their personal wealth on their campaign.91 According to the Court, the resulting burdens on the speech of privately financed candidates were—while not as severe as direct caps—still too great.
In sum, the most powerful speakers’ amplification remains the freest from regulation.
3. Mass Amplification: Media Regulation
There is one more puzzle piece to add with respect to the doctrine of mass amplification: how the rights of media gatekeepers themselves are analyzed. The Supreme Court has primarily maintained that gatekeepers to amplifying platforms have their own rights as speakers to determine which speech is welcomed onto their platforms. For instance, in the seminal Miami Herald Publishing Co. v. Tornillo, the Court overturned a Florida statute that required newspapers to print opposing views on any public issue covered in their pages, in light of the newspaper owners’ editorial discretion.92
While the Supreme Court has not yet extended such rights to tech companies, such as search engines and social media platforms, other courts have. In 2014 in Zhang v. Baidu.com, Inc., prodemocracy advocates sued the search engine Baidu, a popular alternative to Google in China, under various New York antidiscrimination laws for blocking their articles from its search results.93 The district court ruled that Baidu has free speech rights to block whatever content it would like. Three years later, a Florida district court found Google to have similar rights.94 Given the recent de-platforming decisions by social media networks, the Court is likely to be asked to weigh in on these questions.
The answer seems predictable, given that, in the recent litigation over net neutrality, lower courts have just barely denied full-blown editorial discretion rights even to Internet service providers (ISPs), like AT&T and Comcast—and even then, with a current member of the Supreme Court dissenting. Net neutrality rules seek to ensure that ISPs do not discriminate against different content creators (usually companies) in granting access to the Internet, by blocking them, slowing connections to them, or charging them higher rates. Federal-level net neutrality rules have ebbed in and out with alternating Democratic and Republican administrations over the past couple of decades, though many states have passed their own.95 Most recently, the Obama Administration passed net neutrality rules in 2010, and these were functionally repealed by the Trump Administration in 2018 by the reclassification of broadband Internet as a service not subject to them.96 It was before the reclassification, in 2016, that the D.C. Circuit upheld the rules against a First Amendment challenge in United States Telecom Ass’n v. FCC.97
The court concluded that ISPs lack First Amendment rights because they do not make substantive judgments about content and therefore are mere “conduits” for the speech of their users.98 While it is surely correct that ISPs currently operate as mere conduits,99 they could choose to be otherwise. As then-Judge Brett Kavanaugh pointed out, the court’s version of editorial rights seems to rest on a “use it or lose it” theory of free speech rights, whereby, counterintuitively, a decision not to speak or interfere is considered no speech at all.100 If editorial rights are analogous to individual speech rights, then surely Kavanaugh is right. One cannot lose one’s future speech rights by remaining silent in the present. As a result, if net neutrality rules are enacted again and come before the Court on which Justice Kavanaugh now sits, they could well be overturned.
In other words, the owners of the most powerful amplifying platforms appear to also have unlimited rights to choose whom to amplify.
II. Amplification and Democracy
The speech rights guaranteed by the First Amendment are meant to serve not just the interests of individual speakers, but also structural interests.101 The First Amendment’s structural interests are those served by the entire system of freedom of expression, rather than by individual acts of speech. The primary structural interest identified in the doctrine interpreting the Free Speech Clause is democracy. This Part explores the conditions under which mass amplification serves democracy.
Generally speaking, the mass amplification of speech is indispensable to the democratic process. Mass amplification makes easier the widespread dissemination of information and ideas. It allows the approximation of responsive nationwide “conversations”: public figures and pundits are able to respond to one another precisely because they all hear one another; and citizens are able to converse with one another more coherently given their shared common knowledge of the latest news and opinions expressed in widely circulated newspapers, television shows, and tweets. Without mass amplification, we would likely all be stuck in our own silos, uninformed and unchallenged.
But just because mass amplification is generally speaking indispensable to democracy does not mean that every instance of it is. Consider a single speaker who commandeers an entire mass-amplifying platform and crowds out all other speakers. Her amplification presumably serves her own interests, but at the same time undermines the democratic interest in educating voters about competing views.
This Part explains in detail how speakers’ amplification and democracy come to collide in certain cases. It also offers a diagnosis for why the Supreme Court has denied any such collision. In short, the Court has been captivated by an erroneous, long-discredited laissez-faire vision of the marketplace of ideas that sees all speaker interests and democratic interests as harmonious.
The tension I am introducing between speakers and democracy at first appears in a troublesome constitutional guise. Speakers’ interests and democratic interests both appear to be protected by the First Amendment. Unless one interest is less weighty than the other, how are courts to adjudicate conflicts between them? However, I will ultimately argue in Part III that the constitutional nature of the conflict is merely apparent in cases involving mass amplification. The specific speaker interests at stake in these cases are not the sort protected by the First Amendment. As a result, sometimes constraints on speakers’ amplification will on balance clearly promote First Amendment values. For this Part, however, I set aside the constitutional protection of speakers, and focus exclusively on the democratic threats they can pose.
A. Democracy in First Amendment Doctrine
The first point to establish is that, in spite of the doctrine’s primary emphasis on speakers, structural values are in fact relevant in interpreting free speech rights.
Generally speaking, rights are justified because they advance fundamental interests of persons.102 Ordinarily, the interests advanced are those of the right-holder, or the individual who exercises the right.103 This is the most familiar understanding of rights, and in law it is furthered by our legal process: it is individual right-holders who bring lawsuits. However, it is well established in political philosophy, and accepted in law, that the moral or legal recognition of a right can also be justified because of the value that third parties, or even society as a whole, derive from that recognition. For instance, parents’ custodial rights benefit not just parents but children, too; education rights benefit individual students, but also create a more educated population of citizens. And these “structural” interests are considered in interpreting the scope and application of the right.
Free speech rights are likewise grounded in both individual and structural interests.104 As First Amendment scholars well know, when interpreting the Free Speech Clause, courts consider the benefits that speech rights can have not just for speakers but for audiences and society at large, too. These nonspeaker benefits inhere not in the value of the individual’s freedom of choice but in the consequences for the entire system of freedom of expression. That system provides for the general dissemination of information and ideas, including about the economic marketplace,105 and opportunities for exchange and debate.
The primary structural value of the First Amendment, and the one on which this Article focuses, is democratic discourse.106 Free speech scholarship also now overwhelmingly emphasizes the democratic dimension of the First Amendment, and for good reason.107 The Court variously describes free speech and discussion as “at the foundation of free government by free men,” “vital to the maintenance of democratic institutions,”108 “essential to free government,”109 the only means through which “government remains responsive to the will of the people and peaceful change is effected,”110 and even “the guardian of our democracy.”111 One of the Court’s most common refrains in First Amendment cases is that “speech concerning public affairs is more than self-expression; it is the essence of self-government.”112 The structural democratic value of free speech was an important—if not leading—reason why the free speech, press, and assembly clauses of the First Amendment were adopted.113
I will sometimes call this value “democracy” for short, but I always mean to refer to democratic discourse in particular. On nearly all democratic theories, political decisions—by voters and officials alike—must be made against an ongoing background of public discourse. By “public discourse,” I mean the aggregation of all public discussions, which may be more or less centralized and more or less overlapping. While the First Amendment could serve democracy more broadly, it is generally interpreted to safeguard this discursive aspect of democracy in particular.
Public discourse serves two core functions in the democratic process: (1) disseminating the information and ideas that shape public opinion and public character; and (2) legitimizing democratic decisions.114 Call these the epistemic function and the legitimacy function. The epistemic function dominates the doctrine.115 The interest and even right of the people to “receive information and ideas” has been cited in free speech cases since the early twentieth century, and consistently set on a par with the quintessential speaker interest in self-expression.116 As Justice Brandeis says in his seminal concurrence in Whitney v. California, the freedom of speech is “indispensable to the discovery and spread of political truth.”117 The Court often tellingly cites the free speech scholar Alexander Meiklejohn with approval.118 Meiklejohn argued that the sole purpose of the First Amendment is to protect the informed judgment process of the democratic citizenry, and as such that it should be used to ensure that public discourse aided in that process.119 Most of what I say in the following relates to the epistemic function of democratic discourse, because of its doctrinal primacy, but I will explain as I go how the legitimacy function can usually be served in conjunction with the epistemic function.
While many of these decisions recognizing the democratic value of the First Amendment are old classics, the Roberts Court continues the tradition. Take Citizens United v. FEC, in which the Court struck down federal restrictions on corporate campaign spending. Many have assumed that the Court focused on the value of the speech to the corporations. Yet much of the argument was instead about democratic discourse. The Court reiterated that the value of speech lies at least partly in its democratic value: “Speech is an essential mechanism of democracy, for it is the means to hold officials accountable to the people. . . . The right of citizens to inquire, to hear, to speak, and to use information to reach consensus is a precondition to enlightened self-government and a necessary means to protect it.”120
The Court went on to devote many words to establishing that the corporate speech in the case was valuable to public discourse. The Court’s primary concern was that, through the law regulating corporate expenditures, “the electorate [had] been deprived of information, knowledge and opinion vital to its function.”121 In particular, the government had “muffle[d] the voices that best represent the most significant segments of the economy.”122
B. Competing Models of Epistemic Competition
The bedrock American free speech principle is that government should stay mostly out of the picture when it comes to speech. This principle has obvious advantages for speakers, but it is understood to advance the democratic goals of the First Amendment, too. The Court has offered two different explanations for how democratic discourse thrives best on a strict government-hands-off approach; one applies to its primary epistemic function of discourse and the other applies to its secondary legitimacy function. This Section discusses the epistemic side of the matter, while the next Section treats the legitimacy side.
No notion has a firmer grip on First Amendment doctrine than the marketplace of ideas.123 Sometimes conceived of as a theory, sometimes as a metaphor, it (very roughly) analogizes public discourse to an economic marketplace. Yet in this marketplace, ideas, rather than products and services, compete against one another for acceptance. And truth is thought to be more likely to emerge through this competition.124 The origin of the theory in Supreme Court doctrine is usually traced to Oliver Wendell Holmes’s dissenting opinion in Abrams v. United States, in which he remarked that “the best test of truth is the power of the thought to get itself accepted in the competition of the market.”125
Constitutional doctrine contains two different strands of thinking about what “the competition of the market”—i.e., epistemic competition—entails. The first is the more familiar laissez-faire model, which ties the health of epistemic competition to the absence of government intervention. The second is what I call the process model, which ties the health of epistemic competition to the existence of certain background conditions of discourse. This Section examines each in turn. The Court often regards these models as reconcilable, but the final passages of this Section explain how they are actually in tension (and why the latter should prevail).
1. The Laissez-Faire Model
On the laissez-faire model, the government is the great nemesis of epistemic competition: the marketplace of ideas functions best to produce truth when government refrains from intervening in private speech. The model takes the sensible fear that the government will distort the marketplace toward views it favors and elevates it to the primary First Amendment concern.
Sometimes, the Court’s laissez-faire model rests only on a distrust of government and assumes nothing at all about the epistemic merits of an unregulated marketplace of ideas. On this view, the truth may or may not emerge from public discourse, but it has at least a fighting chance only if government does not rig the process. This view of laissez-faire as the best we can hope for is reflected in the Court’s description of the First Amendment as removing “governmental restraints from the arena of public discussion, putting the decision as to what views shall be voiced largely into the hands of each of us, in the hope that use of such freedom will ultimately produce a more capable citizenry and more perfect polity.”126
Other times, the Court’s laissez-faire model draws on the further assumption that the marketplace of ideas itself has epistemic merit; the competition of the market is such that truthful speech possesses inherent advantages over false speech. In this vein, the Court sometimes implies that all speech has some value in the pursuit of truth, or at least cannot hurt it. John Stuart Mill, often thought to be the originator of the marketplace of ideas theory and an inspiration of Holmes’s, laid out the case for this view with particular elegance in his nineteenth-century writings.127 First, speech believed to be false or harmful may in fact be true or helpful. Humans, even when acting in good faith, are notoriously unreliable in distinguishing between truth and falsity, good and evil.128 Second, even if speech truly is false, its existence out there somewhere can actually contribute to the elevation of truth, either because it often contains at least a grain of truth itself or because responding to it sharpens our understanding of the truth.129
When drawing on this thicker version of the laissez-faire model, the Court maintains that government intervention in private speech is harmful not just because it might distort discourse but because it reduces the total volume of speech in competition.130 For instance, the doctrine’s well-known remedy for false speech is “more speech.”131 This also explains why the Court, when striking down regulations on speech, tends to stress the contribution that the targeted speech would make to the search for truth—as it did in Citizens United when it emphasized that corporate speakers offer a unique perspective on the economy.132
Both the thinner and thicker versions of the laissez-faire model hold up fairly well in at least one context: outright censorship. The government has the unique power to effectively censor views in all or most forums and ulterior motives to exercise it. If certain views are removed from discourse altogether, it takes little imagination to see the damage to epistemic competition—if we take seriously Mill’s contention that human beings are fallible arbiters of truth. But later in this Section, I will offer arguments that show why, in other contexts, even the thinner version of the laissez-faire model falters.
2. The Process Model
A second model of how democratic discourse achieves epistemic competition can be distilled, with some care, from the doctrine. This model—the process model—and the laissez-faire model are not logically contradictory, though I explain below why in practice they conflict.
The process model construes epistemic competition not as an absence of government intervention but as a positive process with certain defining features that are conducive to the emergence of truth. These features are ones that allow the average citizen, consuming and potentially participating in public discourse, to make an informed, rational judgment about public issues (if she is so inclined). They are also—consistent with the “process” name—purely procedural, in the sense of being neutral with respect to what the truth is.
I cannot here offer a comprehensive accounting of all of the truth-conducive, purely procedural features of democratic discourse. More empirical research than we currently have would help in doing so. However, at least three such features already have firm footholds in constitutional doctrine: diversity, mobility, and antagonism.
Diversity obtains when a wide range of views from differing perspectives are included in the mainstream of public discourse.133 Discourse lacks diversity when certain widely held or important voices are generally not heard. Mobility obtains when a pathway—even if a difficult one—exists for views to move from the fringes to the mainstream of public discourse. Antagonism obtains when diverse views are articulated at least sometimes in an adversarial fashion. That is, the same media consumer will hear these views in close proximity, or even dialogue, and these views will be articulated by actual proponents who are best positioned to passionately advocate for them. Antagonism is especially missing in so-called echo chambers.134
Diversity and antagonism are both straightforwardly endorsed in great classics of First Amendment law. In AP v. United States, the Court expressly states, and has repeated many times since, that the First Amendment “rests on the assumption that the widest possible dissemination of information from diverse and antagonistic sources is essential to the welfare of the public, that a free press is a condition of a free society.”135 Parallel language can be found in one of the most oft-quoted lines from New York Times Co. v. Sullivan, which warns that government intervention in public discourse can “dampen[] the vigor and limit[] the variety of public debate.”136
Similar themes echo up and down the cases in later years, with the Court declaring that government cannot limit “the range of information and ideas to which the public is exposed”137 and that “assuring that the public has access to a multiplicity of information sources is a governmental purpose of the highest order, for it promotes values central to the First Amendment.”138 Even in Citizens United, the Court states that “it is inherent in the nature of the political process that voters must be free to obtain information from diverse sources in order to determine how to cast their votes.”139 In addition, in many cases the Court praises the vigor and vibrance of public discourse and the need for citizens to hear views they do not like.140
The Court’s endorsement of mobility is implied in its embrace of the marketplace of ideas theory. Roughly, the idea is that no views may be suppressed, so that they may make their way into larger discourse where they can compete for acceptance. As the Court says, “[b]y protecting those who wish to enter the marketplace of ideas from government attack, the First Amendment protects the public’s interest in receiving information.”141 This would not work unless a path to receipt existed.
Diversity, mobility, and antagonism most directly produce epistemic competition, but they also bolster political legitimacy as a byproduct.142 Those who feel shut out of democratic discourse may well not respect democratic decisions.143 But mobility ensures that they have a fair shot, and diversity ensures that at least those whose views are held by substantial numbers are part of the debate of the polity.
In conclusion, note that the process model is not a model of what, overall, makes public discourse function best for democracy. Epistemic competition is just one crucial component—and one that happens to have solid roots in the Court’s jurisprudence. However, scholars have posited other, thicker accounts of the foundations of democratic discourse. For instance, many new democratic-egalitarian readings of the First Amendment emphasize the importance of equality and even civility among speakers in public discourse.144 These interpretations are unfortunately far from the traditional doctrine, but some of their goals can be realized just by safeguarding epistemic competition. For instance, disadvantaged voices will often benefit the most from the promotion of mobility and diversity, because it is these voices that struggle to gain traction. At the same time, antagonism can help to break down the echo chambers that fuel the spread of hatred and intolerance.
3. The Limitations of the Laissez-Faire Model
Ordinarily, the Court operates as though the laissez-faire and process models are in harmony: the absence of government intervention is what allows for the preconditions of epistemic competition. If true, this would reconcile the First Amendment’s democratic values and speaker values, too, because (at least plausibly) speakers are better off when the government does not intervene in their choices. The overarching First Amendment principle would be to resist government intervention. The trouble, highlighted in this Section, is that the two models of epistemic competition are actually in tension—and it is the laissez-faire model that is untenable. (At least this is true outside the one, simple context mentioned above: outright censorship.) Once the laissez-faire model is out, we can see how, in some cases, speakers threaten (the process model of) epistemic competition and therefore democratic discourse.
The laissez-faire picture of the marketplace of ideas goes awry in large part because it is flat.145 It evokes speakers sitting around a circular table, exchanging ideas. Everyone gets to speak and everyone listens. The only threat at the table is the censor. Yet real public discourse has tiers, defined by levels of amplification: only very few speakers rise to the topmost level to be heard by larger numbers. And they make it to the top because the gatekeepers of the amplifying platforms let them in.
These top gatekeepers have power over democratic discourse rivaling that of government. They can immediately transform a national conversation by changing their criteria of entry. We have seen this illustrated most recently by the de-platforming of Donald Trump by Twitter and Facebook.146 Largely, this power is due to the increasing concentration of the ownership of mass-amplifying platforms in a small number of hands.147 The big tech giants such as Facebook and Google have also bought up rivals (such as Instagram and YouTube, respectively) to further consolidate their power.148 At the same time, these powerful gatekeepers lack stable incentives to prioritize diversity, antagonism, and mobility when deciding who gets the most amplification. Rather, they face market pressure to favor paid and sensationalist speech and to entrench echo chambers.
The Court has occasionally recognized private threats to diversity, when under the sway of the process model rather than the laissez-faire model. For instance, in Red Lion Broadcasting Co. v. FCC in 1969, the Court upheld the now-defunct “fairness doctrine” because it promoted epistemic competition. That Federal Communications Commission directive required that—to maintain a license to broadcast—a radio or news channel had to (a) cover controversial issues of public interest and (b) devote some of its coverage to opposing views on those issues.149 The Court recognized that the broadcast companies had an outsized influence on public discourse, given the scarcity of the airwaves, and thus that the policy served the diversity of discourse. As the Court said, “[i]t is the purpose of the First Amendment to preserve an uninhibited marketplace of ideas in which truth will ultimately prevail, rather than to countenance monopolization of that market, whether it be by the Government itself or a private licensee.”150 The FCC’s intervention, according to the Justices, was supported by the people’s “collective right to have the medium [of radio] function consistently with the ends and purposes of the First Amendment.”151 While Red Lion remains good law, the Court has rarely followed its reasoning outside of the broadcast context.152
But it is antagonism that most visibly suffers in today’s media climate, dominated as it is by social media platforms. Facebook, Twitter, Instagram, and YouTube all use algorithms that promote content to users at least in part based on what they have previously liked.153 This algorithmic design can encourage the development of echo chambers, in which users hear only speech expressing views similar to their own.154 This can lead to an intolerance toward opposing views and, in extreme cases, radicalization.155
What this discussion suggests is that government is not only a potential threat to epistemic competition: it may be the only means of countering other potential threats. Only government is powerful enough to set basic ground rules for public discourse—rules that apply to either amplifying platforms or their users. This is a key objective behind campaign finance laws, net neutrality rules, and many other forms of regulation.
Perhaps government can be trusted here to do what it does in the antitrust context: police the background rules of competition—in this case, epistemic competition. Antitrust laws are designed to protect the process of competition, not individual competitors. To do so, they restrict free private transactions—such as collusion, horizontal mergers that lead to monopolies, and price discrimination—that distort that process. These laws also hold bigger players (those with greater market share) to different standards because of their scale.156
None of this is to deny that the risk of government intervention is heightened in the free speech arena. If the state privileged the amplification of certain views over others, it could substantially distort the marketplace. But in the absence of government intervention, discourse is already skewed by the choices of those who control the amplifiers. Indeed, the status quo skew could itself serve to entrench government power because public officials often have influential allies in industry and media. This is all the more likely when the voices missing from the conversation are those of dissenters or disadvantaged speakers.
It is thus not clear if state action or inaction poses a greater risk to epistemic competition. Arguably the answer is inaction if the state’s power is circumscribed to purely viewpoint-neutral regulations designed to promote epistemic competition. In many of the campaign finance laws, for instance, government is not skewing the market toward any particular view, but simply encouraging diversity. These sorts of regulations are certainly much less risky than the censorship or content discrimination of which First Amendment nightmares are made.
C. Competing Theories of Legitimacy
The doctrine also contains glimpses of the idea that wholly uninhibited public discourse is required not just for the proper working of the marketplace of ideas but also for self-government to be legitimate. In the vein of Jürgen Habermas’s deliberative democracy theory, democratic decisions are justified exercises of political authority only when they emerge from a reasonable discursive process engaged in by citizens—either preceding any public vote, or always ongoing in the background.157 For the Court, the reasonable discursive process has one crucial feature: the absence of government intervention in citizens’ speech.
One of the most prominent theories of how participation in public discourse legitimizes democratic decisions—and which maps fairly closely onto the doctrine—is that of Robert Post.158 Post argues, drawing on Habermas, that the process of opinion formation is legitimizing for democracy because it is by participating in this process, and crucially choosing how she will do so, that the individual citizen “experience[s]” democratic authority for herself and makes democratic decisions her own.159 Public opinion formation, then, must be permitted to proceed in unregulated fashion, at least in the public domain, because this is what allows each individual participant to choose her own manner of participation.160
Post’s theory of legitimacy, and the Court’s, runs into problems given scarcity, too. Post’s central argument for government restraint in involvement with public discourse is that the individual must, in order to acquiesce in democratic decisions she disagrees with, feel that she has been able to participate in the formation of public opinion on those decisions in the manner she pleases.161 In other words, legitimacy, on his view, hinges on individuals’ subjective experiences. But if the manner in which one individual wants to participate makes it harder for others to participate or to be heard in the manner that they want, then these other citizens may experience democratic decisions as illegitimate. Indeed, Post goes some way toward acknowledging this in later works, after Citizens United, in which he writes that campaign finance regulations might be justified if their absence creates the perception of electoral illegitimacy.162 But if one goes so far, then the ostensibly tight link between speaker choice and democratic discourse is broken.
The concern cuts deeper. If one ties democratic legitimacy to the occurrence of any particular procedure, then one must defend the significance of that procedure. Post’s legitimizing procedure is effectively the absence of government involvement in public discourse. But why is government intervention the only illegitimate influence on discourse? Why would not the influence of major concentrations of wealth, for instance, be similarly illegitimate? For instance, it seems plausible that Habermas, who was not so exclusively opposed to government interventions,163 might see campaign finance regulations as necessary to justify democratic elections.
***
If the conclusion of this Part is correct, and government intervention may sometimes further the democratic purposes of the First Amendment, then it leaves us with the other tension mentioned at the beginning of this Part: that between the First Amendment’s democratic values and its speaker values. Some laws regulating the amplification of speech that are necessary to maintain epistemic competition in public discourse also appear to encroach on speakers’ interests, as in Bennett, Citizens United, and Baidu.com. Given such an internal First Amendment conflict, should courts favor speakers or democracy?
III. Amplification and Speakers
The last Part explored how speech is indispensable to the democratic process—but can also sometimes undermine it. But speakers are protected under the First Amendment, too. And none of the scholars who advocate interpreting the First Amendment to protect democracy have seriously confronted the potential cost of that protection to speakers. Doing so is all the more pressing if law is needed to safeguard democratic discourse, because the Supreme Court has—albeit without acknowledging any conflict with democracy—usually sided strongly with speakers.164
Take Bennett, in which the Court struck down a public funding scheme for electoral candidates. Under the scheme, Arizona made initial funding grants to candidates who agreed to abide by certain conditions, and then gave additional public funds if and when these candidates’ privately financed opponents spent more than a threshold “trigger” amount against them.165 The law would have created more balanced electoral discourse. But the Court lamented that the privately financed candidates’ expenditures would be “less effective” once publicly financed opponents gained matching funds, because “an advertisement supporting the election of a candidate that goes without a response is often more effective than an advertisement that is directly controverted.”166 This remark provoked Justice Kagan, dissenting, to quip, “[e]xcept in a world gone topsy-turvy, additional campaign speech and electoral competition is not a First Amendment injury.”167 More importantly, it revealed the depth of the Court’s preference for speakers.
This Part explains why, in cases involving mass amplification, First Amendment claims based on speakers’ interests are weaker than the Court usually assumes. If one combs the doctrine for the speaker interests on which the First Amendment rests that could, in theory, ground a speaker claim to unlimited amplification, two emerge: autonomy and political participation.168 Addressing each of these interests in turn, this Part demonstrates that both interests gain decreasing marginal benefits from additional degrees of amplification—at least past a point. The fall is especially steep in the case of autonomy. The result is that any right to mass amplification is supported mostly by democratic structural interests rather than speaker interests, and thus cannot preclude carefully tailored efforts to secure epistemic competition.
A. Autonomy
Autonomy, on anyone’s accounting, involves living our own lives.169 In the words of John Stuart Mill, “[i]f a person possesses any tolerable amount of common sense and experience, his own mode of laying out his existence is the best, not because it is the best in itself, but because it is his own mode.”170 But what, exactly, is required for laying out one’s own existence is far from obvious.171
First Amendment doctrine is not much help. My use of the term “autonomy” comes primarily from the First Amendment scholarship,172 because the Court rarely uses the word “autonomy” in free speech cases. However, the Court does appear to employ the concept of autonomy, or something like it, using a range of apparently interchangeable words such as “self-determination,” “self-realization,” or “self-fulfillment.”173 The Court’s emphasis on the self is consistent with the general understanding of autonomy as living one’s own life. But the Court seems to justify the freedom of speech using several distinct conceptions of autonomy.
One conception equates autonomy with freedom simpliciter, i.e., with the absence of constraints on the will. Call this negative autonomy, insofar as it is chiefly characterized by an absence. In the first part of this Section, I argue that negative autonomy is too thin to justify speech rights.
I then trace two other, more promising conceptions of autonomy that also appear in the doctrine as justifications for free speech. Unlike the negative conception of autonomy as freedom from constraint, these other conceptions are positive because they identify autonomy as the presence of certain basic preconditions for meaningful human agency. First is the capacity to think freely and independently. Second is a sense of dignity or self-respect.
In the final part of this Section, I explain how autonomy, understood in either of these latter two ways, gains only diminishing marginal returns from amplification to large audiences.
1. Negative Autonomy
Many scholars have worried that, when upholding free speech claims, the Supreme Court sometimes appeals to autonomy understood as no more than freedom from constraint by others—and, in particular, the government.174 This is what I have referred to as “negative” autonomy.175 The value of negative autonomy is thought to derive from the value of the human faculty for making choices and acting on them. Philosophers since Kant have argued that this faculty, often known as rational agency, is a uniquely valuable characteristic—if not the distinguishing characteristic—of humans.176 Nearly all sane, adult humans possess it. It is the faculty that allows us to “author” our own lives and to take credit and responsibility for our actions.177 If agency is so valuable, then—on this view—the way to respect it is by leaving agents alone to make decisions for themselves rather than allowing others to make decisions for them.178
If free speech rights are recognized in order to protect negative autonomy, then the law must—at least absent an exceedingly strong justification—refrain from invading speakers’ free choice about whether and how to speak.179 Such a negative autonomy-based protection for speech should in principle extend to all features of an individual’s speech—including her use of any amplifying platforms. If it did not, then an explanation would need to be offered for why some choices about speech are more valuable than others—just the sort of value judgment eschewed by a negative account of autonomy, which would rather leave that judgment up to the speaker herself.
Negative autonomy surely has normative force. It is arguably the consideration that weighs against paternalistic legislation: each individual should be able to decide for herself what is best for her, so long as her choice does not cause harm to others. It also works well as an explanation for a general presumption of individual liberty, or the idea that government should not be involved unless it has a justification to be.180 But it is of limited use in defining the scope and strength of First Amendment speech rights, which offer additional protection for individual choice. Insofar as the Court has invoked negative autonomy in the past, it should refrain from doing so further.
First, negative autonomy cannot explain what is so special about speech. We recognize any right because we have a justification for government to guarantee freedom within special zones of individual choice. The justification usually appeals to one or more fundamental interests of persons that freedom in the zone will advance. It is these justifications that courts draw on to interpret the contours of a right—i.e., whether it is implicated or can be overridden in a given case. For instance, the Supreme Court interprets Second Amendment rights so as to advance the interest in self-defense.181 But the justification for a right cannot simply be “freedom” in general, because this would offer no explanation for guaranteeing freedom in this specific zone—here, speech.182
Negative autonomy also cannot explain why some speech should be protected more than other speech. Yet the freedom of speech intuitively has a core, subject to maximal protection. Some acts of speech seem to go to the heart of the freedom of speech more than others—among them, artistic speech, speech about political affairs, speech on matters of conscience, speech in the workplace, speech in the home, and speech about one’s identity.183 Being hushed, whatever you say, in a library does not carry the same blow as being told that you cannot mention your political views at work.
None of this is to say that government should be in the business of finely gradating the value of different speech acts. Rights generally extend beyond their cores precisely to avoid such difficult line drawing. But a view of free speech rooted in negative autonomy is located at the opposite extreme: it has no core, and is therefore both counterintuitive and no guide to interpretation in hard cases.184 The next Section offers two more specific and plausible conceptions of autonomy that the Court has used to justify speech rights.
2. Positive Autonomy
Fortunately, negative autonomy is not the only conception of autonomy appealed to in First Amendment doctrine. The Court also often describes free speech as serving at least two other conceptions of autonomy: one associated with independence of thought, and the other with self-respect. These might be thought of as “positive” conceptions of autonomy, insofar as they locate autonomy in a positive process or characteristic that enables a certain healthy sort of agency. I explain each conception before examining the evidence that the Court embraces them.
A human being is continuously engaged in the process of forming and reforming his self, or the central features of his mind—including his values, principles, commitments, identity, opinions, and dispositions—that determine who he is and plans to be.185 For this process to be autonomous, however, requires some degree of free thinking, or independence from others. Only then is it true self-determination. An individual who is indoctrinated to become a specific sort of person or to carry out a very specific life plan, and then—without any further intervention—in fact becomes that person and carries out that plan, is not autonomous. Of course, no person can be detached from their community or culture, as we are fundamentally shaped by our culture.186 As Mill says, a person’s “desires and impulses are his own” when they “are the expression of his own nature, as it has been developed and modified by his own culture.”187 But, lest he lose his own nature in this process, the individual must not be subject to overly invasive pressures toward conformity with prevailing social norms and expectations, which must instead be either adopted or rejected on his own terms.188
Even having formed herself with the requisite independence, an individual is not autonomous unless she can act consistently with herself. To form intentions and act, she must be motivated to do so. Such motivation requires, above all, self-respect, or—to borrow the philosopher John Rawls’s definition—enough self-confidence in the worth of her plans and in her ability to carry them out.189 Self-respect can partly be found within ourselves.190 But, as Rawls recognized, we acquire most of our self-respect, at some point or another, from the recognition and affirmation of other human beings.191 The attitude is transferable; when others respect us, we tend to respect ourselves.192 For us to respect ourselves, then, we must have our self—our specific traits and talents and goals and commitments—assessed and praised or, ideally, admired. Child psychologists have documented the necessity of affirmation in the ordinary development of children, and research suggests that affirmation continues to play an important role in adult success.193
The Court often runs these two notions of freedom of thought and self-respect together. For example, in an oft-cited passage, the Court describes free speech as “an integral part of the development of ideas, of mental exploration and of the affirmation of self.”194 Justice Thurgood Marshall used similar but even loftier terms:
The First Amendment serves not only the needs of the polity but also those of the human spirit—a spirit that demands self-expression. Such expression is an integral part of the development of ideas and a sense of identity. To suppress expression is to reject the basic human desire for recognition and affront the individual’s worth and dignity.195
But the Justices have most often emphasized the freedom-of-thought conception, by tightly linking the freedom of speech with the freedom of thought.196 The Court has gone so far as to describe the freedom of speech as a “complementary component[] of the broader concept of ‘individual freedom of mind.’”197 Above all, when the Court mentions freedom of the mind it shows concern for the individuality of mind. For the Court, the speaker’s right is not just to speak but to “speak his own mind.”198 In Cantwell v. Connecticut, the Court declared that “[t]he essential characteristic of [speech] liberties is, that under their shield many types of life, character, opinion and belief can develop unmolested and unobstructed.”199 Similarly, in West Virginia State Board of Education v. Barnette, the Court extolled “intellectual individualism,”200 and announced its famous maxim that “[i]f there is any fixed star in our constitutional constellation, it is that no official, high or petty, can prescribe what shall be orthodox in politics, nationalism, religion, or other matters of opinion.”201
The Roberts Court also affirms the importance of freedom of thought, but, perhaps more than prior courts, emphasizes the self-respect conception of autonomy. For instance, three years ago, in Janus v. AFSCME, Council 31, the Court linked the speech right to dignity:
Whenever the Federal Government or a State prevents individuals from saying what they think on important matters or compels them to voice ideas with which they disagree, it undermines these ends. When speech is compelled, however, additional damage is done. In that situation, individuals are coerced into betraying their convictions. Forcing free and independent individuals to endorse ideas they find objectionable is always demeaning . . . .202
Similarly, in Citizens United, the Court praised the “right to use speech to strive to establish worth, standing, and respect for the speaker’s voice.”203 In these quotes, the Court emphasizes the human need for recognition and affirmation—for respect from others.
3. Amplification and Positive Autonomy
Here we see why the speaker’s speaking to ever larger audiences delivers only diminishing marginal benefits to both the freedom-of-thought and self-respect conceptions of autonomy.
a. Freedom of Thought
Speech is indispensable in the free development of one’s own mind because of how speech assists thinking. Verbalizing thoughts, especially aloud or in writing, tends to clarify and refine them.204 In particular, an audience helps to sharpen our words. And dialogue is better for this purpose than monologue. Others’ responses to our speech allow us to see more fully and vividly the space of possible views and how our own fit among them. Because these responses issue from other minds with different perspectives, experiences, and biases, they can also create forceful challenges to our views that pressure us to rethink and defend (or modify) our views. Or, when the responses reflect agreement with us, they can help us to deepen and refine our views.
For our speech to most helpfully aid our free thinking, it is essential that we have control over the exact verbalization—or content—of that speech.205 For one, we must not be cowed from honestly voicing any of our opinions, hypotheses, or doubts. We must be able to follow our thoughts exactly where they go. We must even try on for size ideas we do not believe, so that we can better understand if and why they are wrong. If we do not undergo this free-thinking and free-speaking process, we are more likely to simply absorb the beliefs that happen to be articulated around us—and not to develop our own, distinctive self.206
What is not needed for autonomy is an audience that is very large. It is true that larger audiences expand possibilities for diversity and responsiveness: the larger the audience, the more likely that clear and more capable interlocutors of diverse views will be reached. But is this a linear relationship? Does an audience of 500,000 offer 5,000 times the advantages of an audience of 100? Past a point, which I shall call n, the advantages that additional audience members provide for freedom of thought begin to decline. What we need to be autonomous agents is not every nuance of every view held on earth, but enough of a space of possibilities and enough challenge to our own views to gain a critical distance from them. In a random sampling of 100 people on the streets of New York City, one should quickly find a range of views—atheists, libertarians, doctors, lawyers, artists, religious adherents, etc.—and plenty of people generally willing to engage in conversation. If one lives in a community where even large crowds lack diversity, or willingness to engage, then one might achieve the same benefits to autonomy by just talking to 100 people outside one’s community. The idea is not to defend 100 as n, but simply to suggest that this number is not enormously different from, say, half a million for freedom-of-thought purposes.
Even if the largest crowds maximize diversity, they make the sort of responsive dialogue for which we wanted diversity harder to achieve. The greatest benefits of freedom of thought come not just from monologuing to a crowd and then taking questions, or from posting a video on YouTube and then scrolling through comments. They come from sustained, back-and-forth dialogue. While in theory a speech to 10,000 could generate 10,000 responses, we lack the time to hear all of them and the cognitive capacity to process all of them—much less to reply to all.
Yet this last point suggests an important counterargument. As mentioned above, speaking to larger and therefore more diverse crowds does offer the possibility of finding certain small subsets within them: perhaps a randomized, highly diverse set or, conversely, a set of like-minded people scattered across a population. But while it might be a sign of cosmopolitanism to seek engagement with ever more diverse groups, it is not clear that doing so is required in order to gain the critical distance from one’s own views and surroundings that is necessary for freedom of thought. Any further resulting refinements to the self would be, at most, subtle. Moreover, diverse and like-minded people may be sought in alternative ways, such as by traveling or searching online or making oneself available by search online. It does seem plausible, however, that speaking to very large crowds makes the task of finding these crowds considerably easier and likelier to succeed. To that extent, I acknowledge that this might contribute marginally to the benefits amplification offers autonomy.
b. Self-respect
Speech is critical, too, in the quest for respect and, therefore, self-respect. For others to respect us as individuals, we must express to them who we are. Sometimes such self-expression involves nonverbal representations of one’s emotions, preferences, style, artistic sensibility, etc.—say, through tattoos, clothing, interior design, or music. However, speech is especially important in that endeavor because of the nuance that it enables in conveying our particularity. As Milan Kundera says, “[E]veryone is pained by the thought of disappearing, unheard and unseen, into an indifferent universe, and because of that everyone wants, while there is still time, to turn himself into a universe of words.”207
Here, too, autonomy suffers if the broadest freedom to decide the content of speech is not protected.208 For self-respect to be instilled, we need our actual self, or at least our actual understanding of our self, to be recognized. Respect for some other self—perhaps one tailored to fit the prevailing social climate—will not give us the confidence needed to carry out our own life plans. It is possible that under certain circumstances, being recognized as and treated as someone we are not might be confidence boosting, such as when the false self carries high status. But the effects only run so deep. Psychological evidence suggests that hiding, or remaining closeted, may be harmful to one’s psychological health.209 It is therefore essential that speakers be allowed to be authentic whenever possible, if they choose. This is why the Court shows such empathy for those who are compelled to speak contrarily to their beliefs.210
To engender respect, our self-expression needs a certain type of audience—but not necessarily a large one. Respect is most easily conferred by like-minded persons who are predisposed to favor the existing characteristics, virtues, traits, values, and commitments of an individual. Thus, philosophical and psychological research suggests that recognition by a small community of supporters is for most people sufficient to establish self-respect.211 Rawls, for instance, suggests that all we need for self-respect is the praise of a community of like-minded persons.212 We just need enough recognition to come to believe that our life plans are worthwhile, not to come to judge ourselves to be Übermenschen. Indeed, psychologists find that the greatest boost to self-confidence comes from the development of close, enduring relationships.213 These relationships have the capacity to affirm a larger range of our qualities than superficial relationships between celebrities and fans can. Studies of social media users, for instance, suggest that social media use does not positively impact one’s self-confidence or mental health.214
That said, the desire to express oneself to large, adoring crowds is common. A teenager does not simply aspire to be a musician or an actress of medium renown but a rock star or a movie star. In the age of social media, many go to extreme lengths in the pursuit of likes and followers.215 Even academics often aspire to sell books to tens of thousands, not dozens. Nor do these pursuits seem irrational: high levels of recognition often confer social status; the pursuit of recognition for a particular talent or ability can motivate one to develop that talent or ability; and maybe there is even some rare thrill that can only be experienced by those who step onstage to see tens of thousands of fans below.
Such fame could translate into higher levels of self-esteem, or one’s appraisal of one’s own merit or accomplishments. But self-esteem is different from self-respect, as I understand the latter term.216 Self-respect is one’s assessment of the worth of one’s self, including one’s life plans, and is closely tied to dignity, motivation, and agency. Those who are self-respecting can hold their head high and execute their life plans without pathological self-doubt. Self-respect also can, unlike self-esteem, be seen as a binary: we are either self-respecting or we are not. At the same time, it is far from clear that self-esteem brings self-respect. Fame is generally not linked to mental health or motivation to carry out one’s plans.217 Some psychological studies suggest that celebrities are more depressed and have higher suicide rates.218 Similarly, new research on social media users suggests that the number of followers one has is not obviously correlated with ratings of self-confidence.219 Perhaps the sort of affirmation that comes with fame—of people showing up to hear you speak or “subscribing” to you on YouTube—is too vague or thin to establish self-respect.
But why should we think that the Free Speech Clause promotes self-respect rather than self-esteem? On most philosophical accounts of rights, rights are rooted in fundamental human interests. These tend to be interests that nearly all humans, across a wide variety of life paths, find desirable—precisely because they enable the pursuit of so many life paths.220 By contrast, rights are not granted in order to guarantee that any particular life path goes successfully, or even to guarantee most of the means to make that happen. One is not given a right to become wealthy, win elected office, or be awarded the Fields Medal. Fame is a life path; it is not a life path enabler. To put this intuitively, imagine that one state prevents its citizens from fostering close companions, while a second state prevents its citizens from becoming famous. Both states might well be unjust, but the first state seems significantly more so; friends are necessary for almost any life to go well, but fame is necessary for only certain lives—those of, for example, influencers and public figures—to go well.
The foregoing arguments are not meant to establish that amplifying speech provides no benefits to autonomy. Rather, they are meant to show that those benefits begin dropping off steadily past some threshold audience size, and eventually cease to be appreciable. Mass amplification, therefore, is not essential to safeguarding the autonomy at the core of free speech rights.
B. Political Participation
While free speech jurisprudence mainly stresses the speaker’s interest in autonomy, it sometimes indicates that the Free Speech Clause serves a second speaker interest: participating in democratic discourse. This is the speaker-interest equivalent of the democratic interest discussed in the last Part: our interest not in consuming democratic discourse but in actively participating in it ourselves, as speakers. The Court staunchly defends the First Amendment “right of citizens to participate in political affairs” and in political debate in particular.221 It describes the “core political speech” shielded by the First Amendment as speech intended to “persuade” that “involves both the expression of a desire for political change and a discussion of the merits of the proposed change.”222 It has shown special solicitude toward classical forms of persuasive political communication, such as leafletting, picketing, petitioning, and canvassing.223
Political participation is also the darling of contemporary theoretical First Amendment scholarship.224 On almost all theories of democracy, the legitimacy of democratic decisions requires not just that the people vote but that the vote be preceded by public deliberation.225 Because that deliberation affects voting outcomes, scholars argue that it is part of the process of “self-government” and thus that the people must be able to participate in it just as they participate in voting.226 Theoretical explanations for this entitlement abound. Some, such as Robert Post and Frank Michelman, argue that this contribution to the process of self-government allows us to consider ourselves authors of the laws.227 Some have argued that having the opportunity to contribute secures our status as equal citizens.228
These theories suggest that every individual in a democracy is entitled to at least a minimum of opportunities to influence public opinion through speech, and therefore to some threshold level of amplification. As the Supreme Court says, “The right of free speech is guaranteed every citizen that he may reach the minds of willing listeners and to do so there must be opportunity to win their attention.”229 To feel like a meaningful participant in the collective task of self-governance, or an equal citizen, one may need more than a miniscule such opportunity.230 Perhaps one is even entitled to a significant chance—at least when one’s voice is combined with other voices. Some have gone so far as to argue that this chance must be equal.231
Wherever the threshold lies, however, it is not extremely high. The interest in political participation cannot support a right to unlimited—or even mass—amplification by private means. An interest in participating in a process means participating in a process according to its own rules. An interest in winning chess cannot support a right to sweep all the pieces off the board when your king is in check. So, an interest in participating in a democratic process like the formation of public opinion on political issues comes with an internal caveat: it cannot be served by speech that disrupts the basic integrity of democratic discourse. And if we accept my arguments from Part II, this means the interest excludes speech that disrupts the basic conditions of epistemic competition, such as by crowding out other voices.232 Such internally qualified rights are not unfamiliar. The right to vote, for instance, represents an interest in participating in political decisions according to a legitimate electoral process. We might disagree about what constitutes a legitimate process, but we should be able to agree that no one has a claim to participate to whatever degree that their private resources will allow—say, by buying others’ votes.
IV. Toward a New Amplification Doctrine
As explained in Part I, the Court tends to be more concerned about the ability of advantaged speakers and media actors to amplify speech to an unlimited extent than about the ability of disadvantaged speakers to even have a chance of being heard.233 In other words, mass amplification seems to be more stringently protected by the Constitution than smaller-scale amplification. This is exactly backward.
The normative analysis of Parts II and III showed how the justification for protecting amplification under the Free Speech Clause shifts as amplification increases: while small degrees of amplification are protected because of their benefits for individual speakers, larger degrees of amplification are protected instead because of their benefits for the overarching system of free expression in a democracy.234 The implication is that some laws constraining individuals’ choices about how and when to use mass amplifying platforms may—depending on their purpose and execution—actually serve the First Amendment’s democratic values without undermining its individualistic values.
To be clear, not just any regulation of mass amplification will be constitutional. Such regulations would need to pass intermediate scrutiny: they would need to refrain from discriminating among viewpoints, and to be closely drawn to actually promote epistemic competition.
This Part starts to explore types of mass-amplification regulation that this Article’s analysis might make constitutional room for, and exactly how, doctrinally speaking, that room would be made. I explore in detail only the familiar areas of campaign finance and social media, though my argument may also have implications for defamation, incitement, and other areas of speech law.235 I also offer a few thoughts on how, as a result of my analysis, certain regulations of small-scale amplifiers, like those used in public forums, might be more constitutionally troublesome than previously recognized. I conclude by addressing whether we can trust legislatures and courts to implement my analysis.
A. Mass Amplification
Courts should treat mass amplification and the resources necessary for it in a manner similar to the time, place, or manner of speech—as just part of the context of speech, and not vital to any of a speaker’s First Amendment interests. When it comes to mass amplification, “[i]t is the right of the viewers and listeners . . . which is paramount.”236 Accordingly, courts should review interventions designed to assist listeners under the lower bar of intermediate scrutiny, like they do TPM rules.
1. Campaign Finance Doctrine
Reviewing campaign expenditure limits under intermediate scrutiny would bring the doctrinal treatment of expenditures in line with that of contributions. As mentioned above, campaign contribution limits are reviewed under intermediate scrutiny because they only marginally encroach on speakers’ expressive interests. The Court explains the marginal nature of the encroachment by reference not to amplification but to the “symbolic” nature of contributions.237 Yet the simpler and more persuasive explanation is that any large campaign spending—whether in the form of a donation or an expenditure—goes primarily toward mass amplification designed to change the political landscape and not toward the speaker’s expressive interests. Both forms of spending should therefore be subject to the same intermediate standard.
The result would transform campaign finance law. Legislatures would have more leeway to ensure that voters are exposed to a range of views on electoral questions, whether via direct caps on campaign spending or public funding schemes tailored to balance out electoral discourse. Again, even if justified by the interest in preserving epistemic competition in democratic discourse, any such interventions would still need to be closely tailored to achieve that interest. But many could be, if they strictly target mass-amplified speech and do not overcorrect. For instance, the careful expenditure limits that the Court has overturned in cases like Buckley v. Valeo, Citizens United v. FEC, and McCutcheon v. FEC238McCutcheon v. FEC, 572 U.S. 185, 193 (2014) (plurality opinion) (invalidating federal limits on individuals’ total donations to all candidates each election season). may turn out to pass constitutional muster. After all, the Court has approved other types of restrictions on campaign expenditures under the intermediate scrutiny standard, such as requirements that donors be disclosed.239 Public funding schemes like the one struck down in Arizona Free Enterprise Club’s Freedom Club PAC v. Bennett could be especially valuable for ensuring diversity in electoral discourse in a minimally restrictive manner, because they primarily level up.240
To be clear, I am not trying to revive a government interest that the Court has already unequivocally repudiated: the “antidistortion” rationale. In Austin v. Michigan Chamber of Commerce, the Court upheld a state law that required any corporation wishing to engage in election expenditures to do so from a separate, segregated fund rather than its general treasury. The Court endorsed the State’s aim to combat “the corrosive and distorting effects of immense aggregations of wealth that are accumulated with the help of the corporate form and that have little or no correlation to the public’s support for the corporation’s political ideas.”241 Exactly two decades later, Citizens United explicitly overruled Austin on this point.242
Many pages have been written defending Austin since Citizens United.243 But this Article is not Austin redux. Austin’s principle of antidistortion opposed what we might think of as democratic distortion: the representation of ideas in discourse out of proportion to their support from the population. According to Austin, making corporations spend on elections from separate funds made sense because this ensured that any such spending reflected the views of the corporation’s members. Indeed, Citizens United’s refutation of Austin focuses almost entirely on the latter opinion’s differential treatment of the corporate form.244 By contrast, the concern of this Article is better described as one about epistemic distortion, or interventions that prevent a reasonably diverse range of views from being heard. An epistemic antidistortion rationale may lead to many of the same outcomes as Austin’s would, because sometimes diversity will demand moving the microphone to different hands. But its epistemic spirit, as examined in Part II, has the Court’s own blessing.
2. Potential Internet Regulations
Today the greatest threats to the health of American public discourse can be found online.245 Especially since the 2016 presidential election, concerns have focused on the rapid spread of misinformation, the creation of echo chambers, the promotion of emotionally exploitative and sensationalist content, the encouragement of radicalism and harassment, and the demise of quality journalism.246 While not all of these problems are directly a matter of epistemic competition, most could be partially alleviated if the speech users saw online was more diverse and antagonistic.247 Even if these concerns are partly endemic to Internet communication, commentators insist that the dominant online amplifying platforms, including ISPs, search engines, and social media platforms, have at least exacerbated them.248 Further, tech companies possess broad and unsettling discretion to shape democratic discourse and democratic decisions into the future.249
Legislatures, law schools, political science departments, and the world of punditry are swimming with potential regulations to cabin that discretion in ways that will promote epistemic competition—several of which I will discuss below.250 These regulations would likely only pass constitutional review, however, if courts review them under intermediate scrutiny.
Courts will apply strict scrutiny if, as explained in Part I, a regulation infringes on the “editorial discretion rights” of a mass media corporation.251 These rights are apparently construed to serve the autonomy interest.252 So the argument about autonomy from Part III applies here, too. Because editorial discretion is exercised to decide who has access to amplifying platforms, it itself becomes a matter of mass amplification. Editors could choose to voice their preference for certain content in places other than at their platform gate—and could even choose, in their own voices, to disclaim certain content. No danger is posed to freedom of thought or self-respect. Editorial discretion therefore has marginal value for autonomy purposes.253
Instead, editorial discretion rights exist in service of democratic discourse. Editorial control, when exercised, serves a signaling function, pulling ideas upward in public attention. While this signaling will always involve a degree of arbitrariness, it can sort some better ideas from worse ones and plays an indispensable function in the operation of the marketplace. Then again, it can also be used to unacceptably reduce the diversity of ideas seen by the largest swaths of the population. The question for regulating media platforms is thus whether the exercise of complete editorial discretion in a certain class of cases is consistent with democratic discourse. Usually it will be, given the importance of uninhibited media to democratic discourse; but tensions will start to emerge to the extent that media undermines epistemic competition itself. For instance, in Red Lion Broadcasting Co. v. FCC, the radio company challenging the fairness doctrine invoked their editorial right “to use their allotted frequencies continuously to broadcast whatever they choose, and to exclude whomever they choose from ever using that frequency.”254 But the Court concluded that the scarcity of broadcast radio waves gave each channel too much control over what voters heard.255
Of course, many proposals for Internet regulation that seek to achieve goals beyond epistemic competition are arguably at odds with the First Amendment. For instance, Germany and France have in the past several years passed laws, which some would like emulated in the United States, that have forced social media companies to remove nonviolent, nonpornographic, and non-copyright-infringing content256 and would fail any interpretation of the First Amendment content-neutrality test.257
However, some regulatory proposals do seek to further epistemic competition in ways that are at least broadly in keeping with the recommendations of this Article—though the devil may be in the details. I describe two below: net neutrality and social media diversification.
Net neutrality rules, described above, ensure that Internet service providers do not discriminate in their service.258 While no such rules are currently in place, the Biden Administration will likely restore those from the Obama Administration.259 However, those may rest on shaky ground. While they were upheld in United States Telecom Ass’n, the D.C. Circuit opinion rested on the conclusion that ISPs lack editorial discretion rights because they do not make content moderation decisions.260 In theory, robust economic competition among a large number of firms might keep ISPs out of content moderation, but in fact ISPs are scarce because the cables and other infrastructure necessary for their operation are physically limited. Nothing in practice stops them from engaging in limited content moderation—as then-Judge Kavanaugh pointed out in his dissent from a denial of rehearing en banc in the case.261
Firmer ground for the affirmation of net neutrality comes from recognizing that ISPs are amplifying platforms with a unique role in maintaining epistemic competition: holding open the gates to the very lowest levels of public discourse for the widest possible array of views. By providing a gateway to the Internet, and to the mass-amplifying platforms found there, they facilitate mobility and epistemic competition.
Unlike ISPs, social media companies have not yet been the serious target of American regulation. But many proposals aim to preserve epistemic competition on them. One common strategy, endorsed by Senator Elizabeth Warren, proposes to break up social media companies like Facebook and Twitter, perhaps stimulating the creation of a wider range of competing forums.262 Another common strategy would require social media platforms to adopt independent review boards for content-moderation decisions.263 Other strategies actually tinker in nondiscriminatory ways with the content of social media, by imposing “must-carry” rules in the style of the old fairness doctrine,264 restricting sorting algorithms from solely using the metric of popular engagement,265 and requiring stricter disclosure of speakers’ identities and backers.266
Whether or not these particular proposals are advisable, nearly all of them—except for antitrust solutions—would face serious constitutional stumbling blocks if social media companies were deemed to have editorial rights analogous to autonomy- or political participation–based speech rights of individual speakers. The platforms have thus far held themselves out to be “neutral conduits” like ISPs.267 But all of them engage in exactly the sort of content moderation that courts have deemed ISPs not to.268 To name just a few, Facebook promotes “reliable” news stories,269 YouTube demotes conspiracy videos,270 and Twitter flags tweets that have been repeatedly identified by users as factually false and cuts off access to some.271 Yet even assuming these companies exercise editorial judgment in filtering their content, any editorial rights they have serve no autonomy purposes and must be interpreted instead to serve—and to extend only so far as they serve—the voters’ interest in epistemic competition online. As the Court itself has implied, social media platforms are the new soapboxes “in the modern public square.”272
To reiterate once more, nothing said so far provides a constitutional carte blanche for Internet regulation. Just as for campaign finance regulations, any Internet regulations would need to be closely drawn to achieve epistemic competition in democratic discourse. The Internet is still a relatively new beast, and any rules governing it must be enacted with caution. Indeed, the German regulations on social media appear to have produced serious chilling effects.273 Regulations must minimize chill by involving regulators only in broad structural interventions, not case-by-case line-drawing; and they must produce empirically verifiable results.
Perhaps, to avoid these pitfalls, Internet companies should for now be left to self-regulate, as some scholars have advocated.274 The major trifecta of platforms Facebook, Twitter, and YouTube have all expressed concern about their impact on democratic discourse, either out of a genuine sense of social responsibility or concern for their public image, and all have enacted and called for internal reforms that would increase epistemic competition. For instance, Twitter’s CEO Jack Dorsey has floated the idea of allowing users to choose their own content-filtering algorithms.275 YouTube has altered its “recommended videos” list to decrease radicalization.276 Facebook has established a “supreme court” to review content decisions.277
But the arguments of this Article are actually relevant to social media self-regulation, too. Major social media platforms have also intimated that they are at least informally bound by the First Amendment, as though they are public actors. Sometimes they express concern that any changes in their algorithms that demote the content of users will violate (informal) free speech rights.278 If the companies are taking their cues from the Constitution, then they should be less concerned about depriving users of mass amplification and more concerned about epistemic competition.
At the same time, other scholars have predicted that self-regulation will fail.279 What this Article shows is that, if failure happens, any proposals for state intervention to correct it should not be dismissed merely on the grounds that they violate the speech rights of the social media companies themselves.
B. Small-Scale Amplification
In the public forum cases discussed earlier, the Court correctly grasps that the speaker’s freedom to choose her amplification is not central to the freedom of speech.280 Ever-higher levels of amplification are unnecessary for autonomy. The Court’s chief error is in neglecting how the low-level amplification provided by some platforms can be critical for (a) the mobility of ideas in democratic discourse and, sometimes, (b) political participation. For many citizens lacking wealth and resources, speaking at certain public places or times, or in certain manners, are roughly their only gateways into broader public discourse and the formation of public opinion. Letters to the editor of newspapers are often not accepted; internet blogs and webpages will not be visited unless they are picked up by amplifying algorithms.
What the doctrine needs is a more robust inquiry into the amplification effects of any TPM regulation that blocks disadvantaged voices. In particular, courts must ask if alternative channels of communication with similar amplification potential are accessible to the speakers raising the claim; if not, then the regulation may stifle viewpoints that might otherwise not be expressed in significant numbers.
The Court does look carefully at degrees of amplification in an aberrant recent case, McCullen v. Coakley, which struck down a Massachusetts law that restricted activists’ advocacy in areas surrounding abortion clinics. The law required that protestors stand at least six feet from anyone entering a clinic. The Court criticized the restrictions in part because, since their enactment, antiabortion activists had “reache[d] ‘far fewer people’” and convinced fewer women to terminate their pregnancies.281 While scholars have actually challenged this description of the law’s effects, it is rare to see these sorts of before-and-after amplification comparisons made by the Court.282 While one may or may not agree with the outcome of the case, the inquiry about amplification is the right one.
Reasoning like this may well have changed the outcomes in a number of TPM cases over the decades. For example, in Kovacs, the case discussed in Part I that upheld a conviction for the use of a sound truck in the streets, the dissenters were right to question whether the communist activists had equally inexpensive alternative methods of amplifying their method. To take another example, Members of the City Council v. Taxpayers for Vincent involved a ban on affixing posters to utility poles. The Court sustained the regulation, suggesting that speakers could instead post signs on private property or else print handbills to distribute to passersby.283 Dissenting, Justice Brennan provided a better model for analyzing the ampleness of alternative channels of communication. Brennan objected that the law foreclosed what was perhaps the most effective speech platform available to poorer speakers.284 The Court’s proposed substitutes, he argued, were inadequate for these speakers, because handbills were more costly to produce and distribute, while reaching a smaller audience; and private property owners might not consent to the use of their property to disseminate these speakers’ messages.
C. Institutional Limitations
I have already mentioned that the amplified speech doctrines advocated here would not remove all limits on the regulation of mass amplification. In this Section, I explore the firmest boundaries at greater length and explain how courts can police them.
The first is content neutrality: government must not attempt to remove any views from the conversation, or seriously disadvantage them. Of course, legislative motives can always be mixed, with some legislators aiming to improve discourse and others aiming to gain political advantages for themselves. But this is true for every law. The routine judicial inquiry is whether the law is well supported by legitimate reasons. Indeed, courts have no trouble upholding contribution limits as anticorruption measures, even though these laws may offer benefits for incumbents, who typically have fundraising advantages over challengers.285 A unique advantage of diversity, mobility, and antagonism as government justifications is that they would provide exceptionally poor cover for any legislature or agency deliberately attempting to skew public discourse: a law that creates rather than dismantles extreme advantages or disadvantages for any viewpoint is not narrowly tailored to achieve these goals.
Second, government cannot attempt to engage in fine line drawing when it comes to democratic discourse. Diversity does not require that every view achieves exactly equal airtime; antagonism does not require that opposing views be juxtaposed fifty-fifty on every channel. When such hairsplitting is attempted, unanswerable questions proliferate. What is a view? Which views need to be heard by everyone? How much time should be given to any particular view? While public officials may be better placed to make political judgment calls than courts, they still cannot be relied upon to make such exact determinations.286 The history of the ill-fated fairness doctrine bears out this point. The doctrine was difficult to enforce, because the Agency lacked a clear answer to questions like those above, and media companies could not accurately predict which conduct would result in sanctions.287
More reasonable rules would opt for structural adjustments aimed at bringing out disadvantaged views.288 Such structural adjustments might concentrate on the resources for accessing media platforms, the options for pooling resources for speech, and amplification platforms themselves—that is, the sources of advantage and disadvantage—rather than specific views.
Conclusion
One might have thought, from reading recent constitutional doctrine and scholarship, that the First Amendment is divided against itself, torn between its value of individual liberty and its value of democracy. Some scholars and jurists have taken up the torch of the individual; others that of collective self-governance. This Article offers the beginning of a détente between the two sides, explaining that in one crucial field of speech, the two values are reconcilable.
If the arguments I have presented are correct, then the right to amplify speech to very large audiences is qualified: it is not protected when its exercise interferes with the background conditions that enable democratic discourse to function. The reason is that, as I argued in Parts II and III, amplification rights are justified based on their contribution to the First Amendment value of democratic discourse, rather than to the other, more individualistic First Amendment values. A right cannot be recognized that contradicts its own foundation.
This insight explains why certain speech-restrictive measures can be taken to preserve the integrity of democratic discourse without infringing on individuals’ free speech rights. These measures, from campaign finance laws to media regulations, target primarily speech amplified over mass-media platforms. These platforms are scarce, and access to them is granted to no small degree due to socioeconomic advantages. Thus, laws designed to ensure, in service of democratic discourse, that a diversity of voices are heard over those platforms, should, if narrowly tailored, survive constitutional scrutiny. At least in this field, the twin First Amendment goals—the free individual and the free society—are in harmony.