The First Amendment Problem of Stalking: Counterman, Stevens, and the Limits of History and Tradition

In Counterman v. Colorado, the Supreme Court decided an imaginary case. It held that Billy Ray Counterman’s conviction could not stand because it did not meet the First Amendment requirements for prosecutions based on threats. But, in fact, Counterman was not convicted for making threats. He was convicted of stalking, under a law that does not require that the defendant threaten anyone to be guilty of the crime. This Article argues that the Supreme Court’s confusion about the most basic facts of the Counterman case was not an aberration but instead reflects broader pathologies in First Amendment jurisprudence. These pathologies are a consequence of the impoverished view of the First Amendment’s boundaries depicted in the Court’s recent decisions, which suggest that the First Amendment’s doctrinal terrain can be described by a simple list of historically unprotected categories. These thin accounts of the First Amendment and the doctrinal distortions it creates are not inevitable, however. The Article argues for an alternative, more nuanced and dynamic approach to the question of the First Amendment’s boundaries—one that rests on a richer understanding of the traditions of speech regulation in the United States—and sketches out its implications for the law of stalking and, potentially, many other areas of free speech law. The takeaway is that courts do not need to distort the facts of the cases they adjudicate to craft a First Amendment jurisprudence that is doctrinally coherent, historically informed, and normatively desirable.

Table of Contents Show

Introduction

In 2023, the Supreme Court decided an imaginary case. In Counterman v. Colorado, a 7–2 majority of the Court overturned the conviction of Billy Ray Counterman because it found that the First Amendment required the prosecution to prove Counterman acted at least recklessly when he threatened a local musician, Coles Whalen.[1] Because the prosecution had only proven that Counterman acted negligently, the Supreme Court remanded to the district court for a do-over.[2]

The decision was hailed as a significant ruling that meaningfully reshaped—and imposed limits on—the First Amendment law of threats.[3] And so it did. The question of what mens rea standard the First Amendment requires to convict someone for threatening another has bedeviled courts for many years.[4] Counterman appeared to answer this question definitively, if not uncontroversially.[5] The only problem: this landmark ruling on the constitutional law of threats arose in a case that did not in fact involve a prosecution for making threats.

Counterman was instead prosecuted and convicted under a Colorado stalking law. Colorado’s stalking law does have a prong that requires proof that a defendant stalked another by means of “credible threat[s],”[6] but this was not the prong under which Counterman was convicted. Indeed, Counterman was initially charged under this provision,[7] but the prosecutor dropped the threats charge before trial. Instead, Counterman was tried and convicted under an entirely separate part of the stalking statute—a provision that made no mention of threats at all.[8] To convict him under this provision, the jury had to, and did in fact, find that Counterman repeatedly communicated with Whalen “in a manner that would cause a reasonable person to suffer serious emotional distress” and that these communications did cause Whalen “to suffer serious emotional distress.”[9] The jury was not required to find that Counterman ever made a credible threat, a true threat, or any other kind of threat. But you would not know this from reading the Supreme Court’s opinions. When the case was handed down, all but two members of the Court insisted, incorrectly, that Counterman was convicted, as Justice Barrett put it, “under a Colorado law that prohibits true threats.”[10]

The Court’s failure to understand the most basic fact about the case it decided was not just a simple quirk, mistake, or result of bad lawyering. As we argue in this Article, the confusion and misrepresentation that pervaded the Justices’ opinions in Counterman were a predictable consequence of how both courts and scholars have conceived of the First Amendment’s boundaries in the wake of the Court’s 2010 decision in United States v. Stevens.[11] They were also an illustration of the profound and detrimental effect that this conception of the First Amendment’s boundaries has had on the development of doctrine, including the doctrine that governs the regulation of stalking.

In Stevens, the Court declared that the First Amendment’s boundaries are historically determined. More specifically, it asserted that the strong presumption against the content-based regulation of speech, which courts had previously applied to laws that regulate political speech and other speech on matters of public concern, in fact applies to all laws that regulate speech on the basis of its content, except when the speech falls into one of the few “well-defined and narrowly limited” categories of low‑value expression that have, from 1791 to the present, been considered beyond the scope of constitutional concern.[12] The Court noted the possibility that there may be additional “categories of speech that have been historically unprotected, but have not yet been specifically identified or discussed as such in [its] case law,” but it declined to conclude as much definitively.[13]

The Court’s refusal in subsequent cases to recognize additional categories of low-value speech has led many federal and state courts to conclude that the boundaries of the First Amendment are, for all practical purposes, fixed.[14] Courts have assumed, in other words, that, except in the most extraordinary circumstances, the government can only constitutionally regulate speech based on its content if that speech falls into one of the half dozen or so historically established categories of low-value speech identified as such in Stevens and subsequent cases.

On its face, this approach to First Amendment analysis may seem to have the twin virtues of being both easy to apply and highly speech protective. In practice, however, implementing the Stevens test has not proven so simple or necessarily so speech protective. This is because, notwithstanding all its talk of history and tradition, the test does not map well onto either traditional judicial practice or judicial intuitions about how far the First Amendment extends.

Before 2010, courts relied upon a very different view of the First Amendment’s boundaries than the one Stevens suggested. Rather than assuming that content-based regulations are presumptively unconstitutional unless they regulate speech that falls within a historically established exception, courts instead interpreted the First Amendment to strongly protect speech that contributed to public debate, but to extend very different kinds and degrees of protection to speech that occurred in different contexts and served different purposes. That is to say, courts relied upon a positive vision of how far the First Amendment’s protections extend, rather than a negative vision of what it does not include. As a result, they routinely applied limited or no constitutional scrutiny to the content-based regulation of speech when it appeared to add little value to public discourse and its regulation did not otherwise threaten First Amendment values.[15] This was true even when that speech did not fit into one of the “historic and traditional categories [of low-value speech] long familiar to the bar.”[16] By defining the First Amendment’s boundaries in this way, courts were able to reconcile strong constitutional protections for democratically important expression (newsgathering, protest, consensual private conversation, etc.) with legislation that sought to protect the private rights of individuals against invasive, threatening, or harmful uses of language.

Stevens suggested that this approach to delimiting the First Amendment’s boundaries was not just wrong but “startling and dangerous.”[17] Nevertheless, in the wake of Stevens, courts have not always been willing to reverse these long-standing practices or deny their intuitions about what kinds of speech are entitled to constitutional protection and what kinds of speech are not. Instead, they have justified decisions that comport with these intuitions by expanding the scope of the recognized low‑value categories in sometimes implausible ways—or by distorting the facts of the cases before them.[18]

The stalking cases illustrate how complicated courts have found it to apply Stevens in practice. Stalking, as that term is used today, generally refers to the practice of repeatedly following, surveilling, contacting, or communicating with another against their wishes and in a manner that causes significant emotional distress, including fear.[19] Stalking can occur through non-verbal means—indeed, the term stalking was borrowed from the hunting context to emphasize the similarities between the way a hunter stalks their prey and the practices of surveillance and following that characterized the first well-publicized stalking cases.[20] But stalking frequently involves the use of language and often can be perpetrated entirely through speech.[21] Especially in the internet age, stalkers send direct messages, emails, texts, and letters to their victims, obsessively follow and interact with victims’ social media accounts, and intimidate them by means of symbolic gestures. Stalkers also may publicly post about their victims, on social media or in other speech forums.

When the government proscribes this kind of powerfully expressive behavior, its actions clearly implicate the First Amendment. Nevertheless, as we show in this Article, there are good reasons to think that, even in cases in which the act of stalking consists entirely of speech or symbolic conduct—even in cases where the stalker intimidates or scares his victim solely by emailing her, texting her, or leaving unwanted gifts at her door to signal his presence—this behavior can nevertheless be constitutionally regulable if proscribed by sufficiently narrowly drawn and clearly defined laws. Stalking is tightly correlated with violence and imposes emotional harms on its victims that are significant in their own right. In most cases, the speech that is part of stalking contributes little to public debate about matters of public concern. When considered in light of the purposes and principles that the First Amendment safeguards, there is consequently little reason to apply the withering standard of strict scrutiny to this kind of deeply harmful speech.[22]

Nevertheless, despite the strong normative argument that supports regulation of stalking-by-speech, it is very hard to argue that the speech that comprises the act of stalking falls into one of the historical and traditional categories of low-value expression explicitly recognized as such by the Supreme Court.[23] Although stalking can involve true threats, it often does not.[24] For this reason, many stalking laws were specifically drafted to apply to more than threatening speech alone.[25] Nor can stalking-by-speech easily be characterized as a kind of “speech integral to criminal conduct.”[26] Stalking is the crime; in most cases, there is no other criminal act to which it is integral. There is also little evidence of a longstanding tradition of regulating stalking as such. States only enacted stalking laws in the 1990s. And although the precursors to these laws, including verbal harassment laws, were enacted decades earlier, this was still centuries after the Stevens Court insisted that the boundaries of the First Amendment were set in stone.

The fact that stalking causes obvious and serious harms, in most cases possesses little social value, and yet cannot easily be slotted into one of the traditional categories of low-value speech has caused courts no end of problems in the post-Stevens era. Before Stevens, courts engaged in a relatively nuanced—albeit rather free-wheeling—First Amendment analysis of these laws. Since Stevens, however, analysis in stalking cases has grown increasingly tortured. Courts have either struck down these laws, construed them in ways that belie their original purposes, or made unpersuasive and often circular arguments to justify their constitutionality.

What this has produced is a messy, sometimes incoherent, and deeply unsatisfying body of law—and one that has, as in Counterman, sometimes led courts to distort the basic facts of the cases they adjudicate in order to make sense of them. The result is a distortion not only of reality but of the free speech interests at stake in these cases. As we show, the somewhat ironic consequence of a strict adherence to the Stevens categories is that it has led courts to engage in what we might call “categorical bloat.” Faced with what they view as a finite list of unprotected categories, courts have, in some cases, significantly expanded the scope of particular categories to reach results that accord with their substantive intuitions. In other cases, courts have interpreted Stevens to mean that the government may not prosecute many kinds of stalking, even when that stalking causes serious harm.

Counterman provided an opportunity for the Court to help clarify this doctrinal confusion. Instead, by failing to recognize the distinction between the expressive phenomena of stalking and threats, the decision only further muddied the waters.

In this Article, we use Counterman and the broader doctrinal morass it reveals and may exacerbate to illustrate both the power and problems of Stevens’s rigid historical approach to the First Amendment. We argue that the Court’s inability to understand the basic facts of Counterman demonstrates the need for a more nuanced and dynamic approach to the delimitation of the First Amendment’s boundaries than the approach Stevens articulates. Evidence of a historical tradition of regulating particular kinds of speech or regulating speech to prevent certain kinds of harms may be relevant to the question of what kind of scrutiny the First Amendment requires courts to apply, but it cannot be dispositive. This is because the existence of such a tradition may demonstrate that regulation of this kind of speech does not pose the kinds of dangers of government censorship and manipulation that the First Amendment guards against—but then again, it may not. Indeed, much of modern First Amendment law represents a break from (often quite repressive) historical traditions of speech regulation, rather than continuity with them.[27] And First Amendment law has in the past adapted, and should in the future continue to adapt, to changing social conditions and the novel problems they create.

Members of the Court have, in recent decisions, signaled their approval of a less rigid, more principled approach to delimiting the First Amendment’s boundaries than the one Stevens provides. This was most evident in last term’s Vidal v. Elster, in which the Court unanimously concluded that a provision in the federal Lanham Act prohibiting the registration of federal trademarks that include the name of a president should not be considered presumptively unconstitutional, even though it regulates speech based on its content.[28] Three members of the Court argued that this “names clause” did not trigger heightened scrutiny because of the long tradition of restricting the trademarking of names without constitutional concern.[29] But six members of the Court argued that this kind of historical evidence either might not be or affirmatively was not necessary to conclude that the names clause was compatible with the First Amendment.[30] And Justice Barrett argued forcefully in a concurring opinion (which three other Justices joined) that evidence of a historical tradition extending back to the Founding Era was not necessary to establish the constitutionality of the provision, because “[m]ore than a century’s worth of precedent reflects that trademark law has always been content based without functioning as a ready tool of Government censorship.”[31]

A similar analysis can and should be applied to the regulation of stalking. Stalking laws further interests that the First Amendment cases have long recognized as significant. For almost one hundred years now, courts have recognized that the First Amendment permits the legislature to protect speakers and listeners against unwanted communications and that the right to listen and to associate also includes a right not to associate that the government can protect. This right is not unlimited. The democratic values that the First Amendment protects sometimes require individuals (politicians, for example, or public figures) to listen to ideas and information they would rather not. But the First Amendment has never been interpreted to categorically prevent the government from vindicating the important right to be let alone or to require an intrusion upon autonomy to be accompanied by true threats or integral to some other crime. It should not be so interpreted now.

We argue instead that courts should develop a First Amendment doctrine of stalking that is responsive to the specific characteristics of this kind of speech, as well as the dangers its regulation poses to free speech values. In suggesting what the constitutional law of stalking should entail, we draw from Justice Sotomayor’s concurring opinion in Counterman. Justice Sotomayor’s concurrence was the only one of the three opinions written in the case that resisted the gravitational pull of the Stevens framework sufficiently to recognize the real rather than imaginary facts of the case, as well as the only opinion that explained why stalking laws raise different constitutional questions than laws that prohibit true threats.

In her concurring opinion, Justice Sotomayor argued that the rigid rule that Stevens insisted on need not be extended to laws that regulate stalking-by-speech for three reasons. First, laws of this kind protect an important right that it is within the legislature’s power to protect (namely, the right of the individual to be left alone). Second, they do not reflect a governmental preference for speech of a particular kind and therefore pose a much lesser risk of governmental manipulation of the marketplace of ideas than the kinds of laws that are typically subject to strict scrutiny. And third, it is less likely they will be applied to inadvertent or otherwise misunderstood communications than other kinds of speech prohibitions (for example, laws punishing the making of threats) because what they regulate are repeated communications, not one-off utterances.

We agree with Justice Sotomayor that the Stevens framework is both inappropriate and unnecessary to apply to stalking laws. However, we argue that these basic features of stalking laws are not sufficient to entirely inoculate them from constitutional concern. Instead, we suggest that to protect against any interference with the First Amendment’s broad democratic purposes, courts should require proof of some kind of wrongful mens rea for those charged with stalking (but not necessarily the recklessness mens rea that Counterman requires for prosecutions that are founded on threats). We also think that courts should limit the reach of stalking laws to exclude speech that touches on important public matters or concerns.

These are only suggestions. Courts may devise the constitutional rules that apply to stalking cases differently than we suggest. What ultimately matters, however, is that they do so because they view particular kinds of restrictions on the legislature’s power to protect individuals from stalking to be necessary to vindicate the values and interests that the First Amendment protects, not because a rigid historical view of the First Amendment’s boundaries compels them to do so. Counterman illuminates the grave problems that viewing what is in many respects a novel social problem through this rigid historical lens can create.

The Article proceeds in three parts. In Part I, we describe the very different way in which courts conceived of the First Amendment’s boundaries prior to Stevens, and the impact that Stevens had on both the discourse and the practice of constitutional boundary-drawing in the First Amendment context. In Part II, we explore the impact of these changes on the law of stalking, the doctrinal problems they have created, and the Court’s failure to grapple with, or even perceive, those problems in Counterman. And finally, in Part III, we lay out an alternative view of how the First Amendment should apply to stalking-by-speech and to other kinds of invasive communication and argue for a more principled, multi-dimensional, and less rigidly historical approach to the profoundly important question of what kinds of speech and expressive activity the First Amendment protects.

I. Stevens & the Solidification of the First Amendment’s Boundaries

The Speech Clause of the First Amendment guarantees that the government shall make “no law . . . abridging the freedom of speech.”[32] Despite the seeming absolutism of its language, however, the First Amendment has never been understood to require the government to treat all speech the same. With respect to certain kinds of speech—speech considered to possess significant, or at least more than de minimis, constitutional value—courts have interpreted the First Amendment to impose very serious constraints on the government’s ability to intentionally restrict its dissemination or sanction its expression.[33] With respect to other kinds of speech, however, the First Amendment has been interpreted to provide little or no protection at all. In other words, the First Amendment is not and has never been absolute. To the contrary, there is, as Fred Schauer notes, a “vast expanse of human communication that lies beyond the boundaries of the First Amendment.”[34]

Where and how courts have drawn the First Amendment’s boundaries have shifted over time, however. And for much of the twentieth century, the question of where these boundaries lie proved highly contentious—as one might expect of a question with such significant constitutional stakes.[35]

The Court’s 2010 decision in United States v. Stevens purported to end that contention once and for all by declaring the boundaries of the First Amendment to be historically determined and fixed, and therefore beyond the power of courts to change. Stevens’s depiction of the First Amendment’s boundaries has turned out to be incredibly influential, significantly shaping how both judges and academics talk about the scope and limits of the Speech Clause.[36]

In many respects, however, the decision’s success in reshaping how courts talk about the Speech Clause has only made the actual demarcation of its boundaries more difficult. This is because Stevens’s apparently simple map of the First Amendment does not accord with judicial practice or intuitions about what speech should or should not be constitutionally protected. Courts do not treat and have never treated as fully protected all speech or expressive conduct that falls outside the half-dozen or so “traditional and historical” categories of low-value speech that the Court identified in Stevens or some other longstanding, but unrecognized, tradition of speech regulation.

The result has been to force courts to fit square pegs into round holes to formally comply with the picture of the First Amendment that Stevens draws while also reaching a result that a given court believes is intuitively correct. This approach has produced a body of law riven with boundary struggles, in which courts have struggled to apply the Stevens categories to the facts at hand with little additional guidance from the Court itself. In this Part, we describe how the Court conceptualized the First Amendment’s boundaries prior to Stevens and the significant shift that Stevens enabled, before exploring its problematic effects on the law of stalking in Part II.

A.      The Emergence of the Boundary Question

The question of how broadly the First Amendment’s coverage extends first emerged as a topic of serious judicial concern in the 1930s and 1940s, after the Supreme Court embraced a much more speech-protective interpretation of the First Amendment.[37] Before that, as one of us has documented in detail, the question of what kinds of speech were or were not covered by the First Amendment received little judicial attention because its answer simply did not matter much.[38] Under the dominant interpretation of the First Amendment, and the parallel state constitutional free speech and press guarantees, the government was permitted to restrict even core political or religious speech, so long as a jury concluded that the speech was made with bad intent and threatened the public welfare.[39] Consequently, courts did not need to reflect on the different protection afforded to different kinds of speech. Except when it came to a few narrow areas of free speech law, such as the law of libel, which emerged as a political flashpoint during the Sedition Act debates, the First Amendment’s boundaries remained largely unmarked and unremarked upon in judicial decisions.[40]

This state of affairs changed in the 1930s and 1940s, when the Court embraced the more assertive interpretation of freedom of speech first advanced by Justices Holmes and Brandeis in the aftermath of World War I. The Court’s insistence that the First Amendment permitted the government to restrict speech only when it posed a “clear and present danger” of harm made the question of what kinds of speech triggered constitutional protection an urgent and unavoidable one.[41] After all, almost everything that humans do collectively they do by means of language or symbolic communication. [42] If the government were required to show a clear and present danger of harm anytime it punished members of the political community for actions that involved speech, the result would be to gum up the work of the regulatory state. Few members of the New Deal Court were interested in an interpretation of the First Amendment that led to this result.[43] What followed was the emergence, for the first time, of extended judicial discussion of what kinds of speech or expressive conduct were protected by the First Amendment.

At first, the tone of this discussion was casual. Members of the Court appeared to believe that the question of how far the First Amendment’s boundaries extended was an easy one to answer. For example, in Cantwell v. Connecticut, the Court rather breezily asserted, without much explanation, the existence of important limitations on the First Amendment’s coverage.[44] In that case, the Court struck down the conviction of a Jehovah’s Witness for disorderly conduct after he played a record to two Catholic passersby that described organized religion as an “instrument[] of Satan” and insulted the Catholic Church specifically.[45] The conviction could not stand, the Court held, because the First Amendment protected even sharp and intemperate speech about religion and politics to ensure that all members of America’s diverse society were able to participate in democratic public debate.[46] The Court nevertheless noted that its analysis might have been different had there been evidence not only of “assault or threatening of bodily harm,” but also “intentional discourtesy” or “personal abuse.”[47]

The Court did not explain why the First Amendment protected clearly offensive anti-Catholic dogma but not “intentional discourtesy,” other than to note that “[r]esort to epithets or personal abuse is not in any proper sense communication of information or opinion safeguarded by the Constitution.”[48] But the distinction clearly reflects a strong intuition that although certain kinds of speech—including certain kinds of uncivil speech, like the speech that Cantwell played—needed to be protected in order to safeguard the diversity of a pluralist democracy like the United States, other kinds of speech—epithets, threats, speech that constituted “personal abuse”—did not deserve constitutional protection because speech of this sort did not communicate anything useful to democratic citizens.[49]

In later cases, the Court developed these intuitions into a more worked-out theory of the First Amendment’s boundaries. The Court first sketched out this theory in the important 1940 decision Thornhill v. Alabama.[50] The case involved the question, vexed at the time, of whether participating in a peaceful labor picket was the kind of expressive conduct the First Amendment protected. The Court held that it had to be protected, given the purpose of the First Amendment’s free speech guarantee was to ensure that the public had an “adequate . . . supply [of] information and education with respect to the significant issues of the times.”[51] In “the circumstances of our times”—a time characterized by significant labor strife and intense political debate about the regulation of the workplace—it was obvious to the Court that labor pickets communicated to members of the public information about a significant issue (namely, the causes and nature of labor unrest) and therefore, if peaceful, were entitled to constitutional protection.[52] Thornhill thus strongly suggested that to determine whether speech or expressive conduct had constitutional value, courts should look at the role it played in public discourse and, more specifically, should make a realistic appraisal of whether, considering the circumstances, the kind of speech involved contributed “publicly and truthfully” to the discussion of “matters of public concern” that the First Amendment protected in the name of democracy.[53]

Although the holding of Thornhill proved quite controversial—some at the time rejected the idea that labor pickets contributed anything meaningful to public debate[54]—its theory of the First Amendment’s boundaries emerged over the next few decades as the primary framework the Court used to determine whether different kinds of speech or expressive conduct were covered. In case after case, the Court found that speech that contributed to the vigorous public discussion of public matters required for a healthy democratic system of government—or, at least, that could not be restricted without undermining that vigorous public discussion—had constitutional value and that its regulation therefore triggered heightened judicial scrutiny.[55] Speech that did not contribute much to public debate, or that could be restricted without harming it, was in contrast deemed, if not entirely beyond the scope of constitutional protection, then to possess “lesser” constitutional value and, for that reason, to be more easily subject to regulatory control.[56]

This theory of the First Amendment’s boundaries did not explain all the cases that the Court decided over the course of the twentieth century. The terrain of free speech law turned out to be too broad and the decisions too “eclectic.”[57] But the theory did explain great swathes of the case law. It explained, for example, why constitutional protection extended even to crassly commercial movies and pulp magazines (because the Court recognized their capacity to shape public attitudes and beliefs about important matters of public concern)[58] and why the same was true of certain kinds of government employee speech,[59] sexually explicit speech,[60] libelous speech,[61] professional speech,[62] tortious speech,[63] and (eventually) commercial advertising.[64]

The Thornhill framework also explained why the First Amendment was interpreted not to protect speech imbricated with violence—like the violent labor pickets that the Court, a year after Thornhill, found to be outside the scope of constitutional protection.[65] This kind of speech, Justice Frankfurter explained in his majority opinion in Milk Wagon Drivers v. Meadowmoor Dairies, was not entitled to First Amendment protection because it functioned primarily to coerce rather than to persuade or inform and thereby violated the norms that were supposed to undergird public discussion and render it meaningfully democratic.[66] In other words, because the Court believed that violence-laden picketing did not contribute to the reasoned public debate about public matters, it concluded that the speech “was not meant to be sheltered by the Constitution,” no matter how expressive it might be.[67]

The theory of the First Amendment outlined in Thornhill justified and helps explain the Court’s conclusion in Chaplinsky v. New Hampshire in 1942. The Chaplinsky Court held that “insulting or ‘fighting’ words”—words directed at an individual that are sufficiently outrageous they can reasonably be expected to lead to violent retaliation or “by their very utterance inflict injury [on their target]”—constitute categorically unprotected speech. The Court found that speech of this kind failed to contribute anything significant to the public debate on matters of public concern—that it possessed what Justice Murphy described in his majority opinion as “such slight social value as a step to truth” and formed “no essential part of any exposition of ideas.”[68] The opinion suggested the same to be true of other kinds of speech with “such slight social value,” such as “lewd and obscene” and “profane” expression (as Cantwell had suggested).[69]

The theory of the First Amendment’s boundaries, laid out in somewhat embryonic form in Thornhill and fleshed out in dozens of later decisions, provided a powerful and flexible framework for deciding First Amendment boundary questions. But it was also a framework for deciding border disputes that depended on contested, or at least eminently contestable, value judgments about the norms that structure public debate in a democratic society, and about what kinds of speech have to be protected to ensure that members of the public have an “adequate . . . supply [of] information and education with respect to the significant issues of the times.”[70] After all, what counts as “adequate,” what counts as “information and education,” and what counts as a “significant issue” are all contestable matters of opinion, not simply identifiable matters of fact. The result was intense disagreement about precisely how to apply the Thornhill test in specific cases.

Even though, at least prior to the 1990s, the Justices tended to agree that whether speech was covered by the First Amendment primarily depended on whether it touched on matters of public concern, they disagreed, often quite strenuously, about what this meant in practice. Members of the Court disagreed, for example, about whether commercial advertising had anything to say about important public matters or instead implicated merely private concerns.[71] They disagreed about whether “scurrilous” or hateful speech contributed anything of value to democratic public debate,[72] and they disagreed about whether the Court was correct to deny protection to obscene speech,[73] to libelous speech,[74] or to profane speech,[75] given the democratic interests that the First Amendment safeguarded.

For decades, this disagreement was simply part of practicing First Amendment law at the Court. Indeed, one might credibly say that a huge part of the Court’s First Amendment jurisprudence in the mid- and late-twentieth century was consumed by these kinds of boundary questions and debates.

This fact did not prevent First Amendment scholars, and dissenting members of the Court, from registering strong complaints about the extent to which contestable value judgments were driving First Amendment law. In 1960, Harry Kalven, for example, called the idea that courts had the power to “weigh[] the social utility of speech” before granting it First Amendment protection “difficult to accept as a doctrine.”[76] Fifteen years later, Kenneth Karst argued similarly that the idea that speech might have high or low constitutional value was “radically inconsistent with the principle of equal liberty of expression” that was supposed to undergird the First Amendment.[77] And Justice Douglas declared himself a First Amendment absolutist who believed that the First Amendment “was designed to preclude courts as well as legislatures from weighing the values of speech against silence.”[78]

At one level, these critiques of the doctrine were nonsensical. Given how much human conduct occurs through language, how could a First Amendment jurisprudence compatible with virtually any form of modern government refrain from judging in one way or another the “social utility of speech”? Surely, knowing whether speech has social utility is necessary to determine whether and to what extent it can be regulated as part of the ordinary practice of contract law, or criminal law, or any of the many other branches of law that restrict what individuals can say.

But what Kalven and Karst and Douglas meant to criticize by criticizing the value-laden nature of the Court’s First Amendment jurisprudence was not so much its focus on the protection of public and truthful speech on matters of public concern, or its concomitant exclusion of areas such as contract law and antitrust law from constitutional coverage.[79] What they challenged instead was one piece of the Court’s First Amendment boundaries jurisprudence: namely, the denial of protection in low-value speech cases to speech that addressed a public audience and was disseminated through the ordinary distribution channels of the public sphere. Their argument, in essence, was that judicial judgments of the social value of speech should not by themselves be sufficient to deny speech constitutional protection.[80]

Over time, the Court responded to this powerful but nuanced critique of its jurisprudence by embracing an increasingly absolutist view of the First Amendment’s protection of public speech. A solid majority of the Justices proved increasingly unwilling to deny constitutional protection to speech that addressed a public audience merely because it could be said to lack significant social value.[81] The result was the almost continual narrowing of many of the categories of low-value speech identified in the Cantwell and Chaplinsky line of cases, as well as the extension of constitutional protection to other kinds of heretofore unprotected speech.[82]

The Court proved unwilling to get rid of the low-value categories altogether, however, or to more broadly embrace an absolutist vision of freedom of speech. It continued to exempt from full constitutional protection many kinds of speech that, although not generally regarded as low value, nevertheless had not traditionally been granted First Amendment protection. The Court made clear, for example, that government employees’ speech could be subject to discipline and other sanctions, both when it did not and in some cases when it did touch on matters of public concern.[83] The Court reached the same conclusion about classified and secret speech that was restricted as part of a contractual relationship between a speaker and the government.[84] And, as we discuss in the next Part, lower courts routinely found the same to be true of harassing speech and other kinds of speech that invaded personal privacy, at least when such speech did not implicate public matters.

Throughout the end of the twentieth century and into the twenty-first, the First Amendment’s boundaries remained, in other words, shaped by the theory laid out in Thornhill, and by courts’ view of the social and democratic value that different kinds of speech possessed. Beginning in the 1990s, however, members of the Court began to articulate a different theory of the First Amendment’s scope. At least on its face, this theory did not give courts as much room to determine the First Amendment’s boundaries by reference to their contestable judgments of the “social utility” of speech.

Most notable in this respect was Justice Scalia’s opinion in R.A.V. v. St. Paul in 1992.[85] The case involved the prosecution of a juvenile defendant under a municipal hate crimes law for burning a cross in the yard of the only Black family on the block. The Court reversed the conviction because it found that, even though the law targeted only unprotected fighting words, by criminalizing only certain kinds of fighting words—those that targeted another because of their race or color or gender or religion—the law violated the First Amendment’s prohibition on discriminating on the basis of content.[86] In insisting that the principle against content discrimination applied even to low-value speech, the opinion in R.A.V. reflected the majority’s increasing discomfort with the potentially idiosyncratic value judgments that underpinned the low-value categories. After all, as Justice Stevens pointed out in his concurrence, the majority’s assertion that the First Amendment principle of viewpoint neutrality had to apply even to the regulation of low-value speech in order to prevent governmental manipulation of the marketplace of ideas reflected an implicit recognition that the Chaplinsky Court had been wrong when it concluded that fighting words and other kinds of low-value speech possessed such “slight social value as a step to truth” that their regulation posed no threat to the democratic interests that the First Amendment safeguarded.[87]

Nevertheless, rather than actively rejecting the existence of the low-value categories entirely, the opinion instead reimagined them as historically determined fixtures of the constitutional landscape. Specifically, Justice Scalia asserted in his majority opinion in R.A.V. that the low-value categories had been a part of the First Amendment landscape “since 1791” and that, although their boundaries had changed over time—they had been narrowed by an increasingly speech-protective judiciary—“‘the freedom of speech’ referred to by the First Amendment d[id] not include a freedom to disregard these traditional limitations.”[88] On Scalia’s account, although the scope of the low-value categories might be judicially malleable, their existence did not depend on the circumstances of the time. Instead, it was a historically fixed feature of the “free but civilized” society which was the United States.[89]

Justice Scalia’s emphasis in R.A.V. on the historical fixity of the low-value categories represented a notable discursive shift in the Court’s description of the First Amendment’s boundaries. Although Justice Murphy’s opinion in Chaplinsky had made weak historical claims about the provenance of the low-value categories when it identified them as “well-defined and narrowly limited classes of speech, the prevention and punishment of which have never been thought to raise any Constitutional problem,”[90] the Court had never previously made historical continuity a requirement for delimiting the First Amendment’s boundaries. To the contrary: one of the primary critiques of how the Court conceptualized the First Amendment’s meaning and scope during the mid and late-twentieth centuries was that it broke so much with the past.[91] One cannot help but be struck, reading many of the Court’s most important decisions creating the modern First Amendment, how little history mattered at all.

And history certainly did not seem to constrain the Court when it came to the question of whether the First Amendment required continued recognition of existing low-value categories of speech (Justice Scalia’s claims in R.A.V. notwithstanding). Quite the opposite: in Joseph Burstyn v. Wilson, the Court unequivocally denied that blasphemy was low-value speech for First Amendment purposes, even though courts had presumed the opposite for over one hundred years.[92] Similarly, in Virginia State Board of Pharmacy v. Virginia Citizens Consumer Council, the Court rejected the idea that commercial speech was categorically unprotected, notwithstanding the fact that this was a well-established, traditional category of exclusion.[93] More recently than that, in New York v. Ferber, the Court recognized the existence of a wholly new category of low-value speech—what it called at the time “child pornography” and now tends to be referred to as child sexual abuse materials (CSAM)—on the basis of what appeared to be an entirely presentist analysis of the costs and the benefits of its regulation.[94]

The opinion in R.A.V. thus reflected a significant change in how some members of the Court were beginning to think about the First Amendment’s boundaries, one that clearly reflected anxiety about how claims of social utility could be manipulated to grant novel powers to the government to restrict speech.[95]

At the time, however, this shift in rhetoric failed to impact how First Amendment law actually developed. The Court for its part did not cite the historical language from R.A.V. for almost two decades. Instead, to describe the non-absolute nature of the First Amendment’s protections, it continued to quote the language from Chaplinsky that suggested that speech could be denied constitutional protection whenever it possessed “such slight social value as a step to truth that any benefit that may be derived from [it] is clearly outweighed by the social interest in order and morality.”[96]

The Court only resuscitated the historical claims from Scalia’s opinion in 2010, in what would turn out to be one of the early Roberts Court’s most important First Amendment decisions: United States v. Stevens.[97] This time around, however, the Court’s assertion that the First Amendment’s boundaries were historically fixed and determined would have much more of an impact on First Amendment law—to a degree unanticipated even by the author of the Stevens opinion himself.

B.      Stevens and the Embrace of a Historically Determined First Amendment

Stevens involved a First Amendment challenge to a federal law that made it a crime to knowingly “create[], sell[], or possess[] a depiction of animal cruelty . . . for commercial gain,” when the act of animal cruelty depicted was “illegal under Federal law or the law of the State in which the creation, sale, or possession [of the depiction] takes place.”[98] Congress had enacted the law primarily to stamp out the trade of what is known as animal “crush videos”—videos in which small animals were filmed being tortured to death by stiletto-wearing women for the sexual pleasure of the viewer.[99] For whatever reason, however, the first case the Department of Justice chose to pursue all the way to trial under the new law involved dog-fighting videos instead.[100]

The owner of a business called “Dogs of Velvet and Steel” was charged and convicted of three counts of violating the new law after he produced and sold documentaries that he claimed were intended to encourage a love of the pit bull breed but which included, among other things, clips of illegal dogfights.[101] The defendant argued that his conviction violated the First Amendment because the law under which he was convicted restricted high-value speech without adequate justification.[102] In response, the government argued that the law was not unconstitutional because the speech it regulated—“depictions of animal cruelty”—contributed so little to the “‘free dissemination of ideas of social and political significance’ that lie at the core of the First Amendment” that its benefits, if any, were clearly outweighed by its costs.[103] More specifically, the government argued that, like the speech recognized as low value by the Ferber Court, the depictions of animal cruelty criminalized by the new law contributed virtually nothing of value to the democratic public sphere but instead merely showed, and thereby encouraged, criminal and extremely harmful conduct.[104]

Although in many respects the analogy the government drew to Ferber was strong—in both cases almost all the images depicted illegal and violent activity—the claim that the speech possessed as little social utility as CSAM was a hard one to make given the facts of the case. The defendant’s videos contained, after all, plenty of ideas and information about the history and merits of the pit bull as a breed apart from the (in some cases) brief clips of dog fighting.[105] And even those videos, as the defendant’s brief noted, “lack[ed] . . . images of blood or serious injury to the dogs.”[106] They were, in this respect, far tamer than the “gruesome” videos of dogfighting distributed by animal rights organizations as part of their political efforts to halt the practice.[107]

The Court therefore could have disposed of the case somewhat narrowly by rejecting the specific analogy that the government drew to Ferber and insisting that the speech the federal law regulated possessed enough social utility to warrant the full protection of the First Amendment. This was indeed how the Third Circuit, sitting en banc, justified its decision to strike down the law.[108]

Instead, the Court decided to use the case to articulate a general theory of the First Amendment and its boundaries. Rather than holding that the law was unconstitutional because it got the balance of social utility and harm wrong, the Court instead rejected the idea that First Amendment protections depended on this kind of balancing at all. In his majority opinion, Chief Justice Roberts expressed shock and outrage at the idea that speech might be categorically excluded from the First Amendment merely because its social value appeared outweighed by the harms it caused. This idea, he wrote, “is startling and dangerous,” even if it reflected what the Court had asserted for decades.[109] Chief Justice Roberts explained:

The First Amendment’s guarantee of free speech does not extend only to categories of speech that survive an ad hoc balancing of relative social costs and benefits. The First Amendment itself reflects a judgment by the American people that the benefits of its restrictions on the Government outweigh the costs. Our Constitution forecloses any attempt to revise that judgment simply on the basis that some speech is not worth it.[110]

Thus, in order to maintain the original judgment made by the American people about the costs and benefits of allowing the government to regulate different kinds of speech, the opinion insisted that speech could be categorically excluded from constitutional protection only when it fell into one of the low-value categories that “[f]rom 1791 to the present” had “never been thought to raise any Constitutional problem” or, alternatively, comprised a “historically unprotected” category of speech that had “not yet been specifically identified or discussed as such in our case law.”[111] It was on this basis that the majority concluded that the speech targeted by the federal law could not be categorically denied First Amendment protection. After all, “depictions of cruelty to animals” was not a heretofore recognized category of low-value speech.[112] And even if, as Chief Justice Roberts noted, laws protecting animals against acts of cruelty extended back to the Founding, laws prohibiting depictions of these acts had no such history.[113]

In other words, and in contrast to R.A.V., which made historical claims about the First Amendment merely in dicta, the Stevens opinion incorporated a historical boundary test into its holding and into the core of First Amendment law. It thus announced a significant revision not just in how the Court discussed the First Amendment’s boundaries but in how it applied them. Just a few months earlier, the lower court in Stevens had declared that the idea that “the [traditional] categories [of unprotected speech] may be supplemented” by courts, when circumstances required it, to be an “unassailable proposition” under the existing precedents.[114] After Stevens, however, this idea was not only not unassailable; the Supreme Court seemed to have declared it flat-out wrong.

The effect of this decision on the development of First Amendment doctrine was almost immediately apparent. The following year, in Brown v. Entertainment Merchants’ Ass’n, the Court relied upon Stevens’s new boundary test to determine the constitutionality of a California law that made it a crime to sell extremely violent video games to minors.[115] The majority held that because there was no historical tradition of restricting minors’ access to depictions of violence, the California law triggered strict scrutiny—and, on that basis, struck it down.[116] Justice Alito, in an opinion joined by Chief Justice Roberts, disagreed with the majority’s conclusion that the violent video games were high-value speech (although he concurred in the judgment because he thought the California law was unconstitutionally vague).[117] Contra the majority, Justice Alito argued that depictions of violence for minors should be interpreted to fall within the well-established low-value category of obscenity for minors because to do anything else would create a bizarre funhouse world in which the government could “prohibit the sale to minors of . . . ‘girlie magazines,’” but would face great difficulty “prevent[ing] children from purchasing the most violent and depraved video games imaginable.”[118]

That the author of the Stevens opinion, just a year after it was handed down, joined an opinion that argued that the majority was misapplying the Stevens test illustrates how malleable this apparently simple test would turn out to be. Notwithstanding the Court’s claim to the contrary, this was largely due to the fact that the existing categories of low-value speech were anything but “well[ ]defined” and “narrowly limited.”[119] Instead, they were often pliable, evolving, and indeterminate categories of exception, many of which (like the fighting words category) were introduced into First Amendment law with apparently little thought or judicial discussion.[120] This meant that courts retained a great deal of wiggle room under the Stevens test to set the First Amendment’s boundaries where they wished by defining the recognized categories of low-value speech more or less narrowly.

The decision in Brown also makes clear, however, how quickly and thoroughly the map of the First Amendment from Stevens captured the judicial imagination and became the terrain on which disagreement about the limits of constitutional protection for speech occurred. After all, notwithstanding the vigorous disagreement between the majority and the concurrence about the status of extremely violent speech under the First Amendment, both parties to that debate took it for granted that determining the constitutionality of California’s law depended on figuring out whether it regulated a kind of speech that fell into a historically established low-value category. The only member of the Court who questioned that assumption was its most originalist member, Justice Thomas, and he did so (ironically enough) because he believed that, however you applied the Stevens test, it led to a conclusion that was inconsistent with all the historical evidence of what the First Amendment was understood to mean in 1791.[121]

Brown was no aberration. Today, courts routinely invoke the Stevens test to determine whether a given content-based regulation of speech triggers strict scrutiny or some other, less-stringent standard. And the decision’s framing of the First Amendment’s boundaries has come to be recited as the natural starting point in scholarly articles as well.[122] Both in theory and in practice, the opinion has profoundly reshaped how both courts and scholars think about the First Amendment.

The judicial embrace of the Stevens test had some advantages: it provided (ostensibly at least) a much clearer rule of decision that courts could apply in boundary cases than the rather complex and murky “matter of public concern” test that preceded it. And it made it harder, albeit not impossible, for judges to deny protection to speech simply because they believed it to be unworthy of such protection.

But the test has real disadvantages too. One of these disadvantages is that, notwithstanding all its talk of tradition, the Stevens test is inconsistent with how courts traditionally resolved First Amendment boundary questions. And although the decision very successfully reshaped how courts talked about the First Amendment’s boundaries, it did not reshape judicial intuitions about what speech is entitled to constitutional protection and—equally importantly—about what kinds of interests the government can take account of when regulating speech. Predictable but significant doctrinal problems followed, as the stalking cases discussed in the next Part make clear.

II. The Confused Constitutional Law of Stalking

The case law addressing First Amendment challenges to harassment and stalking laws provides a rather dramatic example of the problems the Stevens test created due to the fact that, notwithstanding all its talk of history and tradition, the decision did not accurately represent the evolving tradition of First Amendment law. This is because, as we show in this Part, in the decades prior to the Stevens decision, courts routinely upheld laws that prohibited verbal harassment and stalking-by-speech without subjecting them to strict scrutiny. They did so for the same reason that the Cantwell Court rejected the idea that constitutional speech protection extends to “epithets or personal abuse”: namely, because they recognized that speech of this kind is not generally seeking to communicate “information or opinion” but is being used instead to further other, less constitutionally important aims (for example, to annoy, harass, intimidate, or dominate).[123] The constitutional analysis in these cases was not always simple. Courts did not always agree about where the precise boundary lay between persuasive and abusive or harassing speech, or what kinds of laws were sufficiently well specified to avoid vagueness concerns.[124] But on the whole, what the cases reveal is a widely shared judicial intuition that the government can protect individuals against unwanted, persistent, and fear-producing speech without significant constitutional constraint, when doing so does not interfere with the freedom of individuals to engage in more legitimate and democratically important forms of communication.

As we also show, things changed in the post-Stevens period. Although the judicial intuition that stalking was not the kind of speech that the First Amendment protected did not disappear in the aftermath of the decision, it became considerably harder to justify doctrinally. The result was, in some cases, judicial acrobatics, as courts pushed and pulled at the doctrine to get the results they believed to be intuitively correct. In other cases, courts submitted to a rigid interpretation of the First Amendment boundaries that left many stalking victims unprotected. This produced an unsatisfying and inconsistent body of law.

In theory, the Court’s decision in Counterman could have helped matters by clarifying the rules that should apply in stalking cases. But, instead, the power of the Stevens framing prevented all but two members of the Court from even recognizing that Counterman was a case about stalking at all.

A.      The History and Nature of Stalking Laws

Concern about abusive, threatening, or coercive speech in a highly gendered and unequal society has motivated lawmaking in the United States for a long time. In the nineteenth century, civil laws against the tort of seduction provided fathers, and ultimately the victims themselves, a limited but nevertheless meaningful cause of action when “daughters were subjected to sexual coercion while they worked outside the confines of their father’s home.”[125] These early efforts to protect against sexual coercion in the workplace, including coercive speech, were precursors to the modern regime of workplace sexual harassment laws.[126]

Beginning in the mid-twentieth century, state and federal lawmakers took steps to protect (usually) women against coercive and abusive speech in other contexts too. In the 1950s and 1960s, in response to an apparent epidemic of telephone abuse,[127] states and the federal government passed laws that made it a crime to repeatedly call someone, call someone anonymously, or make harassing phone calls with the intent to “abuse, threaten, or harass” the recipient of the call or in a manner that could reasonably be expected to have that result.[128] Over the next few decades, many states added provisions to these laws that made it unlawful to communicate with another in an intentionally harassing or abusive manner, even when not on the telephone.[129]

When legislatures decided to take action against what began to be recognized in the early 1990s as the separate but related problem of interpersonal stalking, they built on this existing legal regime but also changed it in important ways. California enacted the country’s first stalking law in 1990 after the actress Rebecca Schaeffer was murdered by an obsessive fan who had stalked her for three years and five other California residents (all women) were murdered by their stalkers.[130] Other states were quick to follow suit after similar crimes occurred around the country.[131] By 1993, all states and the District of Columbia had created criminal sanctions for stalking.[132] In 1996, Congress enacted a federal stalking law.[133] Although the federal law initially prohibited only physical stalking that crossed state lines, Congress amended it in 2000 and 2006 to address the problem of internet or cyberstalking.[134] So too did many states.[135] The result was that, by the early 2000s, all fifty states and the federal government prohibited stalking-by-speech as well as physical stalking.[136]

These new stalking laws shared a number of features that were intended to fill in the gaps left by existing harassment laws. In many of the cases in which women were murdered by their stalkers before the passage of these laws, the stalking had occurred for months—or, in Schaeffer’s case, years—without law enforcement being able to stop it.[137] This was because existing criminal prohibitions on harassment did not usually make it a crime to follow or surveil another. Moreover, when stalkers did engage in verbal harassment, the existing laws treated each act of communication as a separate offense and typically imposed only very light penalties. The result was that, even when prosecutors could use harassment laws to penalize stalkers, the legal slaps on the wrist that they imposed “often served to aggravate the circumstances for the victim” rather than provide them meaningful protection.[138]

The new stalking laws were therefore crafted in an effort to “make it possible for the police to intervene [in an escalating pattern of behavior] before the violence began.”[139] They accordingly punished not only harassing behavior but also the acts of surveilling and/or following another.[140] They also defined the crime of stalking not as a single act but as a “course of conduct” that typically required proof of a series of two or more prohibited acts in pursuit of the same aim.[141] They thus allowed the criminal law to take account of the specific harm that could accrue from the repetition and persistency of harassing behavior. And in some cases, although far from all, they imposed heavier penalties on the act of stalking than harassment.[142]

In other respects, the laws were quite diverse. Different states made different choices about how to define the relevant criminal acts. This reflected the fact that, like the harassment laws that preceded them, what these laws prohibited was a kind of harm-creating social interaction that took different forms in different contexts.[143]

The mens rea the new stalking laws required, for example, varied considerably. Some laws required proof that a defendant who engaged in stalking behavior did so “knowingly and maliciously.”[144] Other statutes, borrowing from the existing harassment laws, instead required proof that the stalker intended to harass or annoy another person or acted in reckless disregard of this result.[145] And some statutes required only that the stalker intentionally commit a series of acts that had the result of placing another in reasonable fear of their safety or the safety of others.[146]

The stalking laws also varied greatly in the kinds of harm they required the victim to suffer in order for stalking to be criminal. Some statutes required that the stalking place its victim in fear of physical harm to themselves or others.[147] Other statutes required only that the stalking cause the victim serious or substantial emotional distress.[148]

The new laws varied in the kinds of behavior that could give rise to stalking. Initially, most state stalking statutes required a “credible threat” of violence, but this proved problematic because many stalkers do not threaten their victims.[149] Many stalkers are motivated by delusion or obsession and seek affection or attention from those they stalk, rather than intending to threaten or harm them.[150] A National Institute of Justice survey found that “[l]ess than half of all stalking victims are directly threatened by their stalkers.”[151] Nevertheless, the study reported that all of the victims surveyed reported feeling great distress, including fear, for understandable reasons.[152] Even in cases where no threats are made, violence can result.[153] And the emotional harms caused by being a target of another’s invasive and obsessive attention can be very significant in their own right.[154] For this reason, when the National Institute of Justice drafted a model stalking law in 1993, it recommended against including a credible threat requirement in stalking laws.[155] Lawmakers listened. Some states criminalized the act of stalking even when it did not involve the making of threats, or they treated stalking-with-threats as an aggravated version of the basic stalking crime.[156]

Some states also created new stalking statutes to deal specifically with the increasing tendency of stalkers to use the internet to stalk their victims.[157] In 1999, Vice President Al Gore asked the Attorney General to write a report on the “emerging problem” of cyberstalking and the adequacy of federal and state laws for dealing with this increasing phenomenon.[158] As the Attorney General’s report concluded, the internet gave stalkers new ways to prey on their victims, as they took “advantage of the ease of communications as well as increased access to personal information.”[159] Despite widespread recognition of these new forms of technologically enabled harms, diverse legislative responses meant that, as with other stalking laws, “no uniform definition of cyberstalking” emerged.[160]

The diversity that characterized the stalking legislation resulted in a complex and voluminous body of First Amendment case law when, as was also true of the earlier generation of anti-harassment laws, defendants challenged their convictions on constitutional grounds. In some cases, courts found that the stalking laws violated the First Amendment—just as courts had in some cases found to be true of some of the earlier anti-harassment laws—either because they were vague or overbroad.[161]

For the most part, however, courts upheld these laws against First Amendment challenges. Just as courts had upheld the vast majority of anti-harassment laws, courts similarly did not interpret stalking laws to imperil (at least not substantially) people’s ability to engage in political protest or other kinds of legitimate expression. As the Third Circuit put it in United States v. Lampley, in 1978, when upholding the federal telephone harassment law against an overbreadth and vagueness challenge: “Not all speech enjoys the protection of the first amendment, and in enacting [the federal telephone harassment law] the Congress had a compelling interest in the protection of innocent individuals from fear, abuse or annoyance at the hands of persons who employ the telephone, not to communicate, but for other unjustifiable motives.”[162] Or, as the Florida Supreme Court put it in Bouters v. State in 1995, to justify its conclusion that the state’s newly enacted stalking law was constitutional: “While the First Amendment confers on each citizen a powerful right to express oneself, it gives the [citizen] no boon to jeopardize the health, safety, and rights of others.”[163] Hence, because the law applied only to speech that “serves no legitimate purpose,” the court concluded that it did not violate the First Amendment free speech guarantee.[164]

To avoid approving statutes that could be used to punish speech that in fact was intended to communicate ideas and information—even if also achieving other aims—courts sometimes interpreted the harassment and stalking laws restrictively, to require proof that the speaker was solely motivated by an intent to harass or frighten, rather than by other more legitimate desires.[165] In other cases, they interpreted the harm requirement narrowly, to avoid an interpretation of the laws that could potentially implicate the sometimes annoying but often important communicative activity that life in a vigorous democratic society entails. In State v. Lee, for example, the Washington Supreme Court interpreted a state stalking law that required the stalking victim to be “intimidated, harassed, or placed in fear” to require proof in all cases that the target of the stalking felt reasonable fear.[166] Even though this construction created statutory redundancy, the court feared that any other construction could implicate a “great deal of constitutionally protected activity” insofar as “[r]eporters or protesters commonly pursue their activities with knowledge that they ‘vex, trouble, or annoy’ others.”[167] The court held, however, that the statute was not facially overbroad because, even though it could “conceivabl[y] [apply to] a reporter or protester [who] [], in the course of following another with a legitimate First Amendment purpose, provoke[d] a reasonable but mistaken sense of fear in the person being followed,” “the risk of such misunderstandings [was] small” and could be addressed on a case-by-case basis.[168]

As the decision in Lee suggests, courts in the pre-Stevens era assumed that the First Amendment constrained the reach of stalking and harassment laws and, more specifically, prevented these laws from being used to criminalize protest, newsgathering, and other “legitimate” communicative acts. Nevertheless, they vigorously endorsed the idea that the First Amendment posed little bar to the government’s power to regulate, at least in cases where the speech was not made with the intent of communicating ideas or opinions or persuading anyone of anything, and the law did not permit the government too much power to criminalize newsgathering, protest, or other obviously important speech.[169] And they concluded as much notwithstanding the fact that neither harassment nor stalking-by-speech fell into any of the categories of speech traditionally identified as low value.[170]

Indeed, in some cases, courts expressly rejected the idea that the speech targeted by harassment or stalking laws had to be considered “low-value speech” in order to be constitutionally regulable. In Yates v. Commonwealth for example, the Kentucky Supreme Court rejected the idea that the Kentucky harassment law, which made it a crime to “[c]ommunicate[] with a person,” with “intent to harass, annoy, or alarm” them, “by telephone, telegraph, mail or any other form of written communication in a manner which causes annoyance or alarm and serves no purpose of legitimate communication,” had to be construed to target only “fighting words” to be constitutional.[171] The court explained:

The First Amendment protects the right of free speech. However, there is a difference between communication in a public forum and the type which this statute regulates. This form of communication intrudes upon a justifiable privacy interest of the recipient and therefore, this right to communicate must be considered in light of a person’s right “to be left alone.”. . . Freedom of speech does not include freedom to convey messages when, where, and how one chooses. That right must be adjusted to the rights of others.[172]

As support for this claim, the Kentucky Supreme Court relied not upon Chaplinsky but instead upon a 1970 Supreme Court decision, Rowan v. United States Post Office Department, which upheld a federal law that permitted householders to notify the postmaster general that they did not wish to receive mailings they considered to be “sexually provocative.”[173] The Court found the law to be constitutional even though it acknowledged that the law could be used to restrict the dissemination of speech that was not considered, for constitutional purposes, low value, but that the homeowner for whatever reason did not want to receive. “To make the householder the exclusive and final judge of what will cross his threshold,” the Court wrote, “undoubtedly has the effect of impeding the flow of ideas, information, and arguments that, ideally, he should receive and consider.”[174] Nevertheless, “[w]eighing the highly important right to communicate . . . against the very basic right to be free from sights, sounds, and tangible matter we do not want” led the Court to conclude that “a mailer’s right to communicate must stop at the mailbox of an unreceptive addressee.”[175]

The decision in Rowan reasserted what was by 1970 a well-established First Amendment principle: namely, that the constitutional guarantee of freedom of speech does not prevent the government from using its coercive and regulatory powers to reinforce, in certain contexts, the right of the individual “to be let alone.”[176] In so doing, the opinion also made clear that what matters to the delimitation of the First Amendment’s boundaries is not merely the content of the speech targeted by state action but also the context in which it occurs, and the nature of the expressive relationship (voluntary or unwilling) between the speaker and audience. Rowan recognized, in other words, that the task of delimiting the First Amendment’s boundaries is a multi-dimensional exercise.

It was this view of the First Amendment that the Yates court and many other state and federal courts relied upon to justify the government’s power to punish verbal harassment and stalking-by-speech when the resulting laws did not intrude too much on what they viewed as “legitimate” communication, particularly legitimate communication in the public sphere. The result was to vest the government with broad, although by no means unlimited, power to regulate in this area.

B.      The Post-Stevens First Amendment Law of Stalking

The constitutional analysis in stalking and harassment cases became considerably trickier after Stevens was handed down. This is because, although by 2010 there was a well-established tradition of regulating the speech that comprises the acts of harassment and stalking, this tradition did not extend all the way back to the Founding period, or even to the nineteenth century. Nor was this tradition easily cognizable within the Stevens framework because the cases in which courts affirmed the constitutionality of personally invasive harassing and stalking speech tended not to rely upon the terminology of low-value speech or cite Chaplinsky. As a result, this body of law became something of a doctrinal orphan once Stevens was handed down.

Perhaps for this reason, it did not appear to occur to litigants—or to courts—that stalking or harassment, or personally invasive speech more broadly, could be treated as a category of low-value speech in its own right under the Stevens test. Instead, courts turned to the familiar low-value categories to justify the constitutionality of stalking legislation, and to two categories in particular: “speech integral to crime” and “true threats.” This approach often provided courts a doctrinal hook that enabled them to do what they clearly wanted to do in most cases: affirm the constitutionality of stalking laws. However, because neither of these categories was designed to address the kind of privacy and autonomy harms that harassment and stalking laws sought to prevent, the result was a doctrinally incoherent or very restrictive body of law. The many appeals courts cases analyzing the constitutionality of the federal cyberstalking law, as well as other stalking statutes, bear this out.

1.       The Federal Cyberstalking Law

In 2000, Congress amended the federal stalking law, 18 U.S.C. § 2261A, to prohibit not only the act of physically travelling across state lines to stalk another person but the use of the “mail or any facility of interstate or foreign commerce to engage in a course of conduct that places [another] in reasonable fear of . . . death . . . or serious bodily injury” when the purpose of that use was “to place a person . . . in reasonable fear of the death of” themself, their family member, or intimate partner.[177] In 2006, Congress broadened the law further to prohibit “use[] [of] the mail, any interactive computer service or any facility of interstate [] commerce to engage in a course of conduct that causes substantial emotional distress” to another, when the purpose of that conduct was “to place a person . . . in reasonable fear of the death of” themself, their family member, or intimate partner.[178] These amendments significantly expanded the law to reach stalking that consisted of speech alone—and that did not necessarily have to involve either a threat or any other kind of low-value speech. Unsurprisingly, then, the federal law has been the subject of many First Amendment challenges, particularly since 2010.

Prior to 2010, few First Amendment challenges to § 2261A made their way to federal court. This may have been a consequence of the narrower reach of the law, although even its pre-2006 version was not obviously limited to any of the “familiar” categories of low-value speech.[179] But it may also have reflected the less-attractive doctrinal environment in which these challenges occurred. Certainly, in those few cases in which federal courts of appeal were asked to analyze the constitutionality of the law prior to 2010, they affirmed the constitutionality of the law quite easily.

Consider for example United States v. Bowker.[180] In that case, the defendant was indicted for violating the federal stalking law as well as the federal telephone harassment law after he sent aggressive and persistent sexual messages via email, phone, and mail to a television reporter he had never met.[181] He challenged the constitutionality of both statutes on First Amendment overbreadth grounds, but the Sixth Circuit Court of Appeals dismissed both arguments rather summarily.[182] With respect to the overbreadth challenge to § 2261A, the Court wrote only that it “fail[ed] to see how a law that prohibits . . . intentionally using the internet in a course of conduct that places a person in reasonable fear of death or seriously bodily injury” could be found to sweep in “a substantial [amount] of constitutionally protected conduct.” “It is difficult to imagine,” the Court added, “what constitutionally-protected political or religious speech would fall under these statutory prohibitions.”[183] The Sixth Circuit engaged in a more extended discussion of the overbreadth challenge to the telephone harassment law, but here too it ultimately concluded that the challenge had no merit. The court reasoned that “the thrust of the statute is to prohibit communications intended to instill fear in the victim, not to provoke a discussion about political issues of the day,” and therefore, even if in some cases the statute could be interpreted to prohibit protected speech, it was not sufficiently overbroad to warrant invalidation.[184]

In the wake of Stevens, constitutional analysis of the federal cyberstalking law became considerably more tortured, because courts could no longer—as the Sixth Circuit had in Bowker—simply conclude that the law was constitutional because it was unlikely to have a significant impact on “legitimate” communication. Courts instead construed Stevens to require them to show that the law regulated only speech that fell into one of the historically established low-value categories of exception.

At first, courts turned to the category of speech integral to criminal conduct to make this showing. This category of speech had been recognized as unprotected at least since the Court declared in Giboney v. Empire Storage & Ice Co. in 1949 that “it has never been deemed an abridgement of freedom of speech or press to make a course of conduct illegal merely because the conduct was in part initiated, evidenced, or carried out by means of language, either spoken, written, or printed.”[185] In the decades after that decision, however, the category was rarely invoked. But, as Eugene Volokh has chronicled, beginning in the 2000s, the category enjoyed a doctrinal resurgence.[186]

This resurgence can be explained, at least in part, as a response to the Court’s embrace of a historically fixed test of the First Amendment’s boundaries in Stevens. Because the Supreme Court never clearly demarcated the boundaries of the category of speech integral to crime, courts were able to invoke the category to uphold many apparently novel kinds of speech regulation notwithstanding the Stevens test.[187] This is certainly how the Stevens Court employed the category. By redescribing child pornography as an example of the broader category of speech integral to crime, the majority was able to argue that its recent decision in Ferber, which appeared on its face to recognize an entirely novel category of low-value speech, was consistent with the historically fixed account of the First Amendment it now provided.[188] The speech integral to criminal conduct category allowed the Court to transform something new into something old and traditional.[189]

In the wake of Stevens, courts attempted to employ this same magic trick to justify the constitutionality of the federal cyberstalking law. In some cases, they were able to do so quite successfully, at least if one did not ask too many questions. For example, in United States v. Petrovic, the Eighth Circuit argued persuasively that the speech that led to the defendant’s stalking conviction under § 2261A (a series of texts and emails in which he threatened to disseminate naked pictures of his ex-lover if she did not pay him $100,000) fell into the category of speech integral to crime because it also led to the defendant’s conviction under the federal extortion law, 18 U.S.C. § 875(d).[190] Because the speech alleged to constitute stalking also constituted a separate federal crime, it was easy enough for the court to conclude that the speech was “integral” to criminal conduct. Of course, the fact that the defendant was already subject to criminal penalties for his speech also raised the question of why conviction under the cyberstalking statute was necessary. But the other convictions certainly made it a simple task to slot the speech into a familiar low-value category.

In other cases, however, it was not so easy to characterize the speech targeted by the cyberstalking law as speech integral to criminal conduct. Consider the Ninth Circuit’s decision in United States v. Osinger.[191] In that case, the defendant was convicted of cyberstalking under § 2261A after he sent multiple harassing texts and emails to his ex-lover and also created a Facebook page in which he posted nude photos of her (photographs he also sent to her coworkers) to punish her for leaving him.[192] The defendant raised a facial and as-applied First Amendment challenge to his conviction, but the Ninth Circuit rejected both.

The court held that the law was not overbroad because it primarily targeted conduct—specifically, the act of stalking—not speech.[193] The Ninth Circuit also held that the law was not unconstitutional as applied because, although many of the acts for which the defendant was convicted involved speech, that speech was integral to criminal conduct.[194] In this case, however, the defendant was not charged with violating a separate criminal statute. Hence, to substantiate the claim that the speech was integral to criminal conduct, the Ninth Circuit argued that Osinger’s speech was integral to the conduct made criminal by the federal stalking law.[195]

This is a fundamentally circular argument and thus deeply problematic. Under the logic adopted by the Osinger court, the government could make an end-run around the First Amendment simply by declaring speech to be an element of a crime that the law described as “criminal conduct.” But speech cannot be transformed into conduct simply by government declaration.

The loose reasoning in Osinger and cases like it generated significant criticism by scholars as well as courts.[196] As a result, in later opinions, federal courts tended to make a somewhat narrower argument to justify the conclusion that the speech regulated by § 2261 was integral to crime. They argued that the speech targeted by the law was integral to criminal conduct because it was speech motivated by criminal intent. As the First Circuit put it in United States v. Ackell, on this view, the federal stalking law was not overbroad because it prohibited “conduct performed with serious criminal intent”—specifically, speech that was motivated by an intent “to kill, injure, harass, intimidate, or place under surveillance”—rather than speech protected by the First Amendment.[197]

This argument has obvious problems too, however. We do not ordinarily assume that mere bad intent is sufficient to make conduct—and certainly not speech—criminal. The criminal law generally requires, after all, both a bad act and a bad intent.[198] And indeed, decades of First Amendment precedent had made clear that the bad act of harassing someone could not always be criminalized, even when the harasser possessed a bad intent, because to do so would be to sweep too far into the (not always pleasant or even well-intentioned) practices of the democratic public sphere.[199]

The First Circuit clearly recognized as much and so defined the terms “intent to harass” and “intent to intimidate” narrowly. In order to avoid constitutional difficulties, the court interpreted the statutory term “intent to intimidate” to mean “[i]ntimidation in the constitutionally proscribable sense of the word[, which] is a type of true threat.”[200] Intent to harass, the court meanwhile argued, must be read to mean “intent to [criminally] harass.”[201] The court drew on a similar reading from an earlier First Circuit case, United States v. Sayer, in which the court found the defendant’s decision to post fake online sex ads about his ex-girlfriend that “deceptively enticed men to [her] home, put[ting her] in danger and at risk of physical harm,” constituted the kind of harassment that the First Amendment does not protect.[202] The Sayer court suggested, in other words, that criminal harassment meant physical harassment that puts someone in fear of physical harm, and it was only speech motivated by an intent to engage in this kind of harassment that could be considered “integral to criminal conduct.” Ultimately, then, to avoid the circularity that plagued the earlier cases while remaining true to the Stevens framework, the First Circuit interpreted the federal cyberstalking law to prohibit only speech that constituted the familiar low-value category of true threats or that encouraged others to engage in behavior that posed a similar threat of physical harm.

In subsequent decisions, other courts of appeal adopted equally narrow interpretations of the federal law. In United States v. Yung, for example, the Third Circuit strongly suggested that § 2261A could be interpreted to apply only to violent, threatening speech even though this was not what the most natural reading of the text of the statute suggested.[203] The same year, the Eighth Circuit similarly concluded that the statute could only be constitutionally applied to speech that constitutes a true threat, or that is integral to another crime that “does not involve protected speech, such as antitrust conspiracy, extortion, or in-person harassment,” or to speech that falls into some other familiar low-value category, such as obscenity or defamation.[204] The Eighth Circuit construed Stevens, in other words, to preclude courts from recognizing stalking as a distinctive act that imposed distinctive privacy and autonomy harms.[205]

The result of these decisions was a significantly straitened reading of § 2261, one that made it inapplicable in almost all cases in which the defendant acted with an intent not to physically frighten but to psychologically dominate or control—that is, in the very cases stalking laws had been enacted to reach.[206] If a stalker happened to defame his victim in the course of stalking her, the Eighth, First, and Third Circuit decisions made clear that the federal law could do something to punish or prevent it. But the harm of the stalking itself was now defined for the most part as beyond Congress’s power to punish, except and unless the speech involved in the stalking included explicit or implicit threats of physical harm. Courts adopted this reading of the law, moreover, despite the fact that in 2006 Congress amended § 2261A specifically to expand the reach of the statute beyond speech or conduct that placed another “in reasonable fear of . . . death . . . or serious bodily injury,” and did so because its members recognized that it was not only stalking of this kind that caused the kinds of serious harm that § 2261 was intended to protect against.[207] And yet neither the First, Third, Eighth, nor Ninth Circuits engaged at all with this problem of statutory construction and the deeper normative problems that it raised. Their focus was on executing the elaborate doctrinal gymnastics needed to justify the cyberstalking law within the Stevens frame.

The result was a body of law that was either worryingly expansive in its (ultimately circular) interpretation of the low-value category of speech integral to crime or very restrictive in the kinds of speech acts that could be constitutionally criminalized. Courts’ efforts to make sense of the federal cyberstalking law within the Stevens framework produced, in other words, a chaotic body of law that engaged very little with the deep normative questions raised by the First Amendment problem of stalking.

2.       Other Stalking Statutes

Similar doctrinal gymnastics characterize the case law addressing constitutional challenges to other stalking laws in the wake of Stevens. When faced with challenges to state and local stalking laws, courts also pushed and pulled to make the regulation of stalking fit within the Stevens framework.

In some cases, courts appeared to find those doctrinal gymnastics quite easy to perform. They rather straightforwardly relied upon the helpfully amorphous category of speech integral to criminal conduct to replicate more or less the same kind of analysis that courts employed in the pre-Stevens era and affirm the constitutionality of state and local stalking laws, albeit in a manner that was formally consistent with the Stevens test.

In State v. Hemmingway, for example, the Wisconsin Court of Appeals upheld the state’s stalking statute against an overbreadth challenge because it found that the law, which made it a crime to “intentionally engage[] in a course of conduct directed at [another] that [one] knows (or should know) will instill fear in [the target], does instill fear in [that person], and would instill such fear in a reasonable person under similar circumstances,” regulated only conduct, not speech, or, to the extent it regulated speech, regulated only speech that was integral to crime.[208] To justify the conclusion that the speech the statute regulated was integral to crime, the court did not identify a separate crime that the speech might violate. It was enough that the statute restricted only speech that was either intended or reasonably likely to cause, and did in fact cause, fear. “Such intimidating conduct,” the court noted, “serves no legitimate purpose and merits no First Amendment protection.”[209]

A few years later, the Wyoming Supreme Court made a similar argument to uphold that state’s stalking statute against an overbreadth challenge.[210] The court found that the statute did not implicate the First Amendment because it criminalized speech because the speech was communicated with an improper purpose, not because it communicated any particular idea.[211] In reaching this conclusion, the court expressly relied upon the decades of prior stalking and harassment cases that reached similar conclusions, albeit by somewhat different doctrinal routes.[212]

In other jurisdictions, meanwhile, Stevens had a more drastic impact on courts’ interpretations of state or local stalking laws. Like the Eighth, First, and Third Circuits, these courts interpreted Stevens to permit the government to criminalize stalking-as-speech only when that speech fell into one of the listed low-value categories. As a result, in order to uphold the stalking laws, these courts significantly narrowed those laws’ reach—and did so even when that required disregarding both legislative history and the most natural reading of the statutory text. For example, in its 2023 decision in Mashaud v. Boone, the D.C. Circuit Court of Appeals interpreted the local stalking law to apply only to communications that fell into the low-value categories of “threats, obscenity, defamation, fraud, incitement, and speech integral to criminal conduct,” even though—as the government argued—this was not at all how the D.C. Council intended the statute to apply. The court dismissed that argument as little more than a claim that the “the Council intended to pass an unconstitutional statute” and instead read the statutory savings clause for all that it was worth.[213]

Other courts also interpreted local and state stalking laws restrictively in order to save them from invalidation. Like the D.C. Circuit, the Connecticut Supreme Court insisted that the state stalking law could be constitutionally applied only to recognized low-value categories of speech.[214] The Supreme Judicial Court of Massachusetts meanwhile interpreted the state’s criminal harassment law to apply only to speech that constituted a “true threat”—and did so notwithstanding strong evidence that the state legislature enacted the law for the very purpose of providing “a remedy to [stalking] victims before ‘nonthreatening’ harassment escalates into life-threatening assault.”[215] Ironically, in order to uphold the constitutionality of a conviction under the law, the court also interpreted the category of “true threat” very expansively, and in a manner that was inconsistent with the Supreme Court’s true threats precedents. Like Osinger and the other decisions that interpreted the category of speech integral to crime very generously in order to reconcile stalking prosecutions with Stevens, here, too, the court engaged in categorical bloat in order to reach what it believed to be the intuitively desirable result—as one of the justices on the court noted forcefully in dissent.[216]

In contrast, in other states, courts struck down stalking laws when they found the statutory text unable to bear the restrictive interpretation that Stevens seemed to require. For example, in 2017, the Illinois Supreme Court struck down a newly added provision to the state stalking law that made it a crime to communicate with another person on two or more occasions in a way that the speaker knew or should have known would cause a reasonable person significant emotional distress because the court found that the new law could be used to criminalize speech that was neither a true threat nor proximately integral to the commission of a separate crime.[217]

For similar reasons, the Minnesota Supreme Court struck down the state stalking-by-mail statute as unconstitutionally overbroad and narrowed the state mail harassment statute considerably to comply with the Stevens framework.[218] Perplexingly, although the court concluded that the first law was constitutionally overbroad because it criminalized speech that was neither integral to crime nor a true threat, it found that the second law could be rendered consistent with the “judgment [of] the American people” embodied in the First Amendment so long as its reach was limited to those who repeatedly sent mail with the intent to abuse another—even though the definition of “abuse” the Court read into the statute extended far beyond the definition of a true threat provided in the First Amendment case law.[219]

And the North Carolina Court of Appeals invalidated a defendant’s conviction under the state stalking law because it found that none of the communications for which he was convicted fell into one of the recognized low-value categories of speech.[220] The court reached this conclusion notwithstanding the fact that, amazingly enough, just a few months earlier a different panel of the court had upheld a conviction under the law without engaging in a careful analysis of the low- or high-value nature of his speech because it found that “the stalking statute[] does not restrict any speech based on its content,” but “[r]ather . . . criminalizes conduct directed toward another person when done for an illegitimate purpose.”[221]

These (what we might charitably describe as) deeply uncertain and inconsistent approaches taken by the different panels of the North Carolina Court of Appeals vividly illustrate the confusion and disagreement that has characterized the First Amendment stalking case law in the wake of Stevens. The confusion and disagreement reflect in part the contested status of some of the low-value categories themselves, particularly the uncertain status of the category of speech integral to criminal conduct. More fundamentally, however, they reflect the difficulties courts have faced reconciling the Stevens boundary test with the approach they had taken for decades in harassment and stalking cases—and the doctrinal gymnastics they have adopted in order to solve these difficulties.

The Supreme Court’s decision to grant certiorari in Counterman v. Colorado—a case that involved a prosecution under a very typical state stalking law—represented an opportunity for the Court to clean up some of this mess by providing guidance to lower and state courts on the clearly vexed question of where stalking laws fit into the First Amendment after Stevens. Instead, if anything, the Court made things worse by failing to recognize the existence of the question in the first place.

C.      The Square Peg of Counterman and the Round Holes of Stevens

No case illustrates the gravitational force of Stevens and the way it has influenced how courts think about the First Amendment’s boundaries as starkly as Counterman v. Colorado. By the time Counterman reached the Supreme Court, it was almost universally understood as a case about threats. The case was seen as neatly teeing up the question that the Court had failed to resolve seven years earlier in Elonis v. United States[222]—namely, what mens rea the First Amendment requires the government to prove when prosecuting someone for making threats. This was a question that had divided the Court in part because there were strong free speech arguments that could be made both in favor and against a high mens rea standard for threats cases.[223] Perhaps for this reason, the parties’ arguments, and all but a few of the amicus briefs, focused exclusively on this interesting, difficult question.[224]

The public conversation about the case also tended to frame it exclusively as a case about threats. The New York Times described Counterman as a case that raised the question of “whether intent counts in criminal cases based on online threats.”[225] The New Republic declared that “[t]he Supreme Court [w]ill [d]ecide the [l]ine [b]etween [f]ree [s]peech and [v]iolent [t]hreats.”[226] Another headline announced, with no apparent awareness of the elision, “Supreme Court takes up convicted stalker’s plea to decide what constitutes a ‘true threat.’”[227] The oral argument at the Court likewise focused on the constitutional doctrine of true threats. The word “threat” or some variation thereof was said 143 times, while the word “stalking” or some variation thereof was uttered a mere thirty-seven times.[228]

And yet Counterman was not a threats case.[229] Billy Counterman was never prosecuted for making threats. Indeed, as noted above, at trial the prosecution dropped a charge that would have required proof of a “credible threat.”[230] Instead, Counterman was convicted under a Colorado statute that provides for the prosecution of someone who “[r]epeatedly follows, approaches, contacts, places under surveillance, or makes any form of communication with another person . . . in a manner that would cause a reasonable person to suffer serious emotional distress and does cause that person . . . to suffer serious emotional distress” after he sent hundreds, and perhaps thousands, of messages over a period of three years to a local musician he had never met, even after she attempted to stop him from doing so.[231] That is, Counterman was prosecuted for, and convicted of, stalking.

The fact that Counterman was convicted of stalking, not threats, did not appear to matter to how courts analyzed the constitutionality of his conviction, however. The First Amendment arguments in his case focused from the very beginning to the end on the question of whether Counterman’s messages to Whalen could be considered a “true threat.” At trial, for example, Counterman’s lawyer argued that Counterman should be acquitted based on the First Amendment and the free speech provision of the Colorado state constitution, but the trial judge rejected this argument because he found that, when considered in light of “the totality of the circumstances, a reasonable jury could find that defendant’s statements rise to the level of a violation of law and that of a true threat.”[232] On appeal, the Colorado Court of Appeals also exclusively focused on the question of whether Counterman’s messages to Whalen constituted true threats.[233]

Similarly, when the Supreme Court issued its opinions, both the majority and the dissents entirely ignored the fact that Counterman was convicted for stalking. Justice Kagan’s opinion for the Court began with the matter-of-fact declaration that “[t]rue threats of violence are outside the bounds of First Amendment protection and punishable as crimes” but then proceeded to make the much less obvious (and in fact untrue) assertion that the case before the Court involved “a criminal conviction for communications falling within that historically unprotected category.”[234] The rest of Justice Kagan’s opinion mostly ignored the facts of the case at hand and focused entirely on the question of how to define the historically unprotected category of threats.[235]

In a footnote, Justice Kagan did acknowledge that “Counterman was charged with violating . . . a ‘stalking’ statute” but suggested that this did not matter to the First Amendment analysis because “the prosecution based its case solely on Counterman’s ‘[r]epeated[ ] . . . communication[s]’ with [Whalen].”[236] That is, Justice Kagan accurately identified that Counterman was charged with the crime of stalking but treated this fact as entirely irrelevant because Counterman’s stalking happened to take the form of speech.[237]

Justice Barrett’s dissent also mischaracterized the nature of Counterman’s conviction. The first line of Justice Barrett’s opinion asserted baldly that “Billy Counterman was convicted under a Colorado law that prohibits true threats,” but this is simply not true. [238]

The fact that Counterman had been thoroughly transformed from a stalking prosecution into a case about threats by the time it made its way to the Supreme Court is in large part due to the tick-the-box analysis of the First Amendment’s exceptions that Stevens encouraged. This approach not only led the Justices to get the most basic facts about the case wrong, but also prevented the Court from fully engaging with the harm that Counterman’s speech produced.

Laws that punish threats—true, credible, or otherwise—punish speech because it communicates a frightening message: namely, that the speaker plans to do something terrible to someone else. The harm that results from this message, as courts have recognized for decades, is the fear and uncertainty that it engenders and all the collateral costs that fear and uncertainty can impose.[239] Laws that punish stalking also strive to protect listeners against fear-causing speech. But the speech targeted by these laws engenders fear not so much because it communicates a frightening message but because it violates basic social norms of privacy and consent by persisting in the face of a lack of consent.[240] It is this violation of social norms and its disregard for the autonomy of the listener—her right to withdraw from the relationship and not have to hear what the speaker has to tell her—that makes stalking speech so frightening. It also threatens the target’s sense of agency and freedom because of its (seemingly unstoppable) intrusion into the victim’s private life.[241] After all, if the speaker will violate these norms, what other norms will he violate?

Scholars of stalking describe this as the harm of “life invasion,”[242] and it is a harm that the stalking and harassment cases prior to Stevens recognized as well.[243] For good reason. Depriving another of autonomy and privacy is a serious injury in a liberal society, and one that cases like Rowan recognize the government should be able to do something about.[244] Indeed, protecting the right of the individual to be left alone follows logically from the strong protection the First Amendment gives individuals to gather collectively.[245] Just as the freedom to speak only means anything if one is also free not to speak, freedom to associate only has meaning if one has a concomitant right not to associate.[246]

Nevertheless, as the previous Part illustrates, in a post-Stevens world, the right of the individual to be left alone has become a difficult thing for courts to make sense of, at least when speech is involved.[247] The facts of Counterman make this rather brutally evident.

Counterman had never met Whalen, but he insisted and persisted in sending her a barrage of messages over years.[248] Whalen never replied—indeed, she repeatedly blocked him. But Counterman would not be dissuaded: he created new accounts, and the messages continued. Some of these messages, particularly those he sent after Whalen failed to respond, were threatening in nature. Counterman told Whalen at one point to “[f]uck off permanently.” [249] Later, he told her that “[s]taying in cyber life is going to kill you.”[250] These were the messages on which the Court focused. But the vast majority of his messages to Whalen were not threatening. Many of these messages—“Good morning sweetheart”; “I am going to the store would you like anything?”—would have been, as Justice Kagan put it, “utterly prosaic,” sweet even, except for the (important!) fact that they came from a “total stranger” and presumed a relationship that did not exist.[251] Collectively, “[t]he messages put [Whalen] in fear and upended her daily existence.”[252] In an amicus brief she filed with the Court, Whalen described the “life altering” impact of Counterman’s behavior and how it caused her to “step back from her dream” of a music career because of her fear that this clearly deluded and obsessive man would turn up at one of her performances.[253]

Importantly, however, this intense fear was not caused by any specific thing Counterman said. The content of the messages was not their most salient feature. In fact, Whalen deleted many of Counterman’s messages without reading them because, as she testified at trial, if she had read all the messages, she “would not have been able to function.”[254] It was instead the “endless barrage” of messages that caused Whalen severe anxiety.[255] Their mere presence in her inbox, the fact of their constant arrival, that they just did not stop, made them distressing regardless of what they said.

But because the case was framed around the problem of threats, both the majority’s and the dissent’s analysis of the constitutionality of Counterman’s conviction focused entirely on the content of the messages he sent Whalen—and more specifically, the half-dozen or so messages that could be interpreted as threatening. The thousands of other persistent, disturbing, prosaic, or nonsensical messages that Whalen received were entirely ignored. Devoid of this context, during oral argument, some Justices even expressed amusement at the idea that Counterman’s only elliptically threatening messages to Whalen could have caused her alarm and suggested instead that Counterman was being “solicitous” of Whalen’s welfare when he messaged her to tell her, for example, that “staying in cyber life is going to kill [her].”[256]

Commentators have blamed the laughter in the courtroom on the failure of the conservative male Justices to take women’s experiences seriously.[257] That may certainly have been a factor. But the Justices’ inability to understand how frightening Counterman’s messages were to Whalen also clearly reflects the framing of the case, and specifically the exclusive focus of the arguments on the question of whether one or more of Counterman’s texts, considered in isolation, conveyed a “serious expression of an intent to commit an act of unlawful violence.”[258]

That the case was framed in this way rather dramatically illustrates the profound and detrimental impact that Stevens has had on how courts conceive of the possibilities and limits of government regulation of harmful or abusive speech. It was because Counterman’s messages frightened Whalen that the trial court assumed that his prosecution for those messages had to be based on true threats. This was, after all, the only one of the recognized low-value categories that involves fear-causing speech. Hence the court assumed that true threats had to be the relevant category of constitutional analysis. And so too did all the other courts that heard the case subsequently.

As the facts of Counterman make clear, this is an incredibly impoverished view of the constitutional landscape. And it is a view that may have significant and unanticipated doctrinal consequences for the First Amendment law of stalking. This is because the Court’s conclusion that the First Amendment prevented the government from punishing Counterman for stalking unless it could prove that he uttered a true threat could be interpreted to mean that in all future prosecutions involving stalking-by-speech, the government would similarly have to show not only that the defendant communicated a “serious expression of an intent to commit an act of unlawful violence,” but that he did so recklessly. And, indeed, this is precisely how some courts have interpreted the decision.[259] They have essentially equated the crime of stalking-by-speech to the crime of making threats.

Other courts have not interpreted the decision so strictly, however. For example, the Maine Supreme Judicial Court recently interpreted Counterman only to hold that, in stalking cases in which the defendant is prosecuted because of what he says, the state must prove that the defendant “recklessly disregarded the effect that his words [would] have upon the victim.”[260] The court interpreted Counterman, in other words, to require proof of subjective mens rea in content-based stalking prosecutions, but a different kind of subjective mens rea than is required to establish a true threat.[261] Moreover, it held that Counterman only applied in cases in which the defendant was prosecuted because of what he said when he communicated to the victim of his alleged stalking, not in cases where the defendant was prosecuted because he continued to communicate at all, despite being told to stop.[262] In these cases, the court insisted, the First Amendment does not require proof of any kind of subjective bad intent because what is being regulated is the defendant’s conduct, not his speech.

By construing Counterman in this manner, the Maine Supreme Judicial Court was able to both clarify the decision’s otherwise deeply unclear relationship to stalking prosecutions and affirm the constitutionality of the defendant’s conviction and the Maine stalking statute generally (a statute which the court acknowledged looked a great deal like the Colorado stalking statute under which Counterman was convicted).[263] It minimized the decision’s impact on the regulation of stalking in Maine, while not denying its applicability entirely. But to do so, it had to depart quite dramatically from what the majority opinion in Counterman actually said—which was nothing about the law of stalking at all.

In fact, courts need not, and we think should not, jump through hoops to justify the prosecution of defendants who violate the autonomy of another by sending them unwanted and frightening or harassing communications but who do not necessarily intend or understand that their speech will be interpreted to communicate a serious threat of physical harm. The act of sending messages to another may be conduct, but it is also certainly speech. Laws punishing this kind of behavior can chill all kinds of potentially valuable expressive activity, particularly if they do not require evidence that the defendant knew his or her communications were unwanted. In this context, as in others, judicial efforts to minimize First Amendment concerns by treating speech as conduct, however well intentioned, may end up overshooting the mark by permitting the legislature to prosecute not only harmful and unimportant speech but many other kinds besides.

Instead, courts should interpret the Counterman decision as if it were the imaginary case the Court assumed it to be—that is, as if it were a case that had nothing to do, really, with stalking prosecutions that do not require proof of true threats. They should recognize that whatever powerful effects Stevens may have had and may continue to have on other areas of First Amendment law (including, in complex ways, the law of threats[264]), there are good reasons why the rigid categorical boundary test that Stevens announced, and that the Counterman majority assumed as a fixed principle of First Amendment law, does not apply to the regulation of stalking.

These reasons were eloquently, if rather succinctly, articulated by Justice Sotomayor in the concurring opinion she wrote in Counterman itself. In her opinion, Justice Sotomayor not only recognized that it might be constitutionally significant that Counterman was convicted for stalking, not threats—as neither the majority nor the dissent had—but also provided a way of reconciling the law of stalking with Stevens without requiring courts to stretch the low-value categories in unjustifiable ways, or transform speech into conduct, or perform any other magic trick. In the next Part, we explore Justice Sotomayor’s alternative vision of Stevens’s place in the First Amendment firmament and what it means for the constitutional analysis in stalking, harassment, and other cases.

III. Clarifying the Constitutional Law of Stalking

The Court’s decisive embrace in recent years of what is often described as a “history and tradition” approach to constitutional law has generated significant critique in the First Amendment context, as elsewhere.[265] Some scholars (including one of us) have expressed alarm at the consequences of such an approach, particularly when it comes to the judiciary’s ability to adequately analyze the Constitution’s relationship to novel technological and social problems.[266] The worry is that a history and tradition approach to constitutional interpretation will bind courts, and consequently all of us, too firmly to the past.

But this need only be true if the history and tradition that constitutional jurisprudence relies upon is a static history and a static tradition. And yet, what Parts I and II make clear is that, as a matter of actual practice, the First Amendment tradition has been quite dynamic. It has been, to use another metaphor, a living tradition. In response to novel social problems—the problem of telephone harassment, for example, or the problem of stalking and, later, cyberstalking—courts have not simply denied protection to speech because they think the speech is “not worth it.” Instead, they have translated older principles— ideas of the democratic public sphere, as well as liberal commitments to individual autonomy and privacy—into new social contexts and regulatory terrains.

The fixed picture of the First Amendment’s boundaries that Stevens provides threatens this tradition. But as Part II shows, Stevens need not be interpreted so expansively as to exclude from constitutional analysis any consideration of the distinctive features of stalking as a social practice. As this Part suggests, there are reasons to think that the rigid rule it articulates should not apply in cases involving stalking speech. In her concurring opinion in Counterman, Justice Sotomayor suggested what these reasons might be.

In this Part, we sketch out the alternative view of the First Amendment’s boundaries that Justice Sotomayor lays out in that concurrence and argue that it provides both a normatively persuasive and doctrinally elegant solution to the problem of how stalking fits into a post-Stevens First Amendment. This is true even if Justice Sotomayor’s view of the constitutional landscape would require courts to reconceive the place of Stevens in the First Amendment landscape and recognize what has in practice always been true: that it does not apply in all cases in which the government engages in the content- or quasi-content-based regulation of speech.

As Fred Schauer notes, it is simply “not . . . plausible to think that the Court in [Stevens and Brown] genuinely intended to create a presumption in favor of First Amendment coverage for the entire universe of linguistic (or pictorial or informational or communicative) activity,” even if that is what the Court in these cases “might be understood as saying.”[267] And indeed, when it comes to the regulation of contractual speech, securities, government employee speech, and much else besides, courts have not interpreted Stevens to require the application of strict scrutiny unless the government can show a “heretofore unrecognized” tradition of regulation dating back to 1791.[268] But courts have interpreted Stevens to apply in stalking cases, as Part II makes clear. This is presumably because stalking cases tend to involve speech like that in Stevens, but unlike that in most student speech or professional speech cases, that occurs in the sites and via the technological media that transmit otherwise high-value “public discourse.” These cases have proven as a result less easily insulated from Stevens because, from afar, they look just like it.

Despite this resemblance, however, Justice Sotomayor’s concurrence suggests that there are principled reasons why the Stevens rule should not be applied to stalking cases, even if it can and should be applied to other kinds of speech. In what follows, we describe Justice Sotomayor’s arguments and suggest that lower courts seeking guidance in navigating the messy and confused First Amendment stalking doctrine should take Sotomayor’s opinion as a beacon guiding their way out of the morass. This is not to say that the opinion answers all the First Amendment questions that arise in stalking cases. But it provides a theoretical framework for courts to answer those questions directly and on their own terms, without having to distort the basic facts of the cases they analyze, or the relevant doctrinal categories, in order to reach normatively desirable conclusions.

A.      What Stevens Obscures

Justice Sotomayor’s concurrence in Counterman was the only one of the four opinions written in the case that did not assume that the constitutionality of Counterman’s conviction depended entirely on whether his messages to Whalen constituted true threats.[269] In other words, it was the only opinion that could see the case on its own terms and not merely in terms of Stevens’s categories.

This is not to say that the opinion ignored Stevens. Like Justice Kagan’s majority opinion and Justice Barrett’s dissent, Justice Sotomayor’s concurrence quoted heavily from the Stevens majority, including its assertions about the historically fixed nature of the First Amendment’s boundaries.[270]

Nevertheless, despite its foregrounding of the language from Stevens, the concurrence did not assume—as the other opinions did—that the Stevens list was the only determinant of whether the content-based regulation of Counterman’s speech triggered strict or more lenient judicial scrutiny. Justice Sotomayor did acknowledge that one of the questions courts should ask when determining whether a content-based law is constitutionally permissible is whether it regulates speech that constitutes a “true threat” or is “integral to criminal conduct.”[271] But she recognized that “[t]rue-threats doctrine came up below only because of the lower courts’ doubtful assumption that petitioner could be prosecuted only if his actions fell under the true-threats exception.”[272] And she noted that speech could “warrant less First Amendment protection for other reasons” beyond constituting a true threat.[273] Stalking laws regulate such speech, Sotomayor argued, for a few reasons: they are not content-neutral as such, but neither are they content-based in the usual sense of the term; they vindicate the important individual right to be “left alone” that has long-standing provenance in First Amendment case law; and they only proscribe repeated communications, rather than one-off interactions. We explain the significance of each of these features of stalking laws in turn.

First, in cases in which a defendant is prosecuted for stalking-by-speech, the “content of the repeated communications can sometimes be irrelevant [to their effect on the listener].”[274] This is because, as the Maine Supreme Judicial Court recognized, the speech that comprises stalking can cause harm not only, or not even primarily, because of the messages it communicates. It can instead cause harm by the mere fact of its intrusion into the life of another. To illustrate the point, Sotomayor invoked the example of a stalker who causes his victim emotional distress by “persistently call[ing] [her] and hang[ing] up” or by sending her “a stream of ‘utterly prosaic’ communications.”[275]

The fact that the harm that stalking laws seek to prevent does not arise solely from the content of the speech—and, equally importantly, that the emotional distress the speech causes, and not speech itself, is the primary trigger of regulation—matters because it means that when the government punishes someone for stalking or harassment, it is not punishing them for their ideas.[276] Billy Ray Counterman was not prosecuted because he believed if Whalen stayed in cyberlife it would kill her, or because he thought he and Whalen were soulmates. He was prosecuted because he persistently invaded Whalen’s privacy with hundreds, perhaps thousands, of messages even after she communicated her lack of consent by blocking him.

This feature of stalking cases has led some courts to insist that stalking-by-speech laws are content-neutral time, place, and manner regulations that should trigger lesser scrutiny for that reason.[277] That is certainly what the text of many stalking laws suggests. For example, the Colorado law under which Counterman was prosecuted defines stalking as the act of communicating with one another on multiple occasions “in a manner that would cause a reasonable person to suffer serious emotional distress and does cause that person . . . to suffer serious emotional distress.”[278] For this reason, the Criminal Justice Legal Foundation argued in an amicus brief to the Court that Counterman’s conviction was constitutional because he was convicted “for the manner of his communications,” not for the content of his speech.[279]

This is not quite right, however. It is certainly true that the Colorado stalking law, and many others like it, do not make any facial content distinctions but instead sweep into their ambit any speech that is made with the requisite bad intent and has the requisite bad effect. But it is too simple to conclude from this fact that laws of this kind target the manner of speech, rather than its content. This is because it will frequently be the content of the speech that a stalker communicates to his or victim that will make it the kind of speech that can and does cause emotional distress.[280] This is particularly true in cases in which stalkers do not (unlike Counterman) send hundreds of texts but only a few painful ones. Even in Counterman, although the sheer profusion of the messages clearly upset Whalen greatly, their content also impacted her. She testified at trial, for example, that she became especially frightened when “the tone of [Counterman’s] messages . . . escalated” and his messages changed from “just invasively intimate [to] more aggressive.”[281] In this stalking case, as in many others, the upsetting manner of Counterman’s messages was closely linked to their content—albeit not their substantive content so much as the emotions they expressed and forced Whalen to have to deal with. In addition, many stalking laws (including the Colorado law) require not only that the speech be emotionally distressing to its recipient but that this distress be reasonable.[282] As a result, to convict someone of these crimes, the jury has to find that the content of the defendant’s messages was the kind of content that could cause a reasonable person to suffer emotional distress. This inevitably involves examination and judgment about the nature of different speech acts.[283]

For all these reasons, it is not accurate to say that stalking laws like the one in Colorado are content-neutral time, place, and manner regulations that trigger only intermediate scrutiny.[284] Instead, these kinds of laws can best be characterized as falling on the boundary line between manner- and content-based regulations of speech. We might describe them, therefore, as content-triggered or content-related regulations of speech, rather than as content-based regulations. But the bottom-line is that they are certainly not always content neutral in the regular sense of the term.

Nevertheless, because laws of this kind do not generally demarcate any particular topic or viewpoint for disfavored treatment—stalkers can say anything and be guilty of stalking, or say nothing at all—they pose far less risk that the government will use them “to give one side of a debatable public question an advantage in expressing its views to the people” than do laws that reflect a governmental content preference.[285] And it is this risk that the application of strict scrutiny is intended to guard against.[286] This provides on its own a very good reason to think that a presumption of unconstitutionality should not be unthinkingly extended to these laws, as Stevens suggests. Stalking laws simply do not pose the same risk of governmental manipulation of the marketplace of ideas as do laws that exclude certain kinds of topics or subjects from public discourse—like the law the Court struck down in Stevens itself.

In her concurrence, Justice Sotomayor identified a second reason why stalking laws might enjoy a different status under the First Amendment: they safeguard a right that the Court has recognized as crucially important to protect since the beginning of the modern First Amendment tradition. This is the right of the individual to be “left alone.” Justice Sotomayor cited Rowan as support for the proposition that “[r]epeatedly forcing intrusive communications directly into the personal life of ‘an unwilling recipient’ . . . enjoys less protection” than other kinds of speech.[287] Indeed, in Rowan the Court was sweeping in its conclusion that “[n]othing in the Constitution compels us to listen to or view any unwanted communication, whatever its merit.”[288] But, as Part I makes clear, the Court’s solicitude for the right of the individual to refuse to receive communication has much earlier roots and is evident in many of its earliest First Amendment boundary decisions too. In Cantwell, for example, the Court insisted that Cantwell’s speech was protected because he was engaged in an “effort to persuade a willing listener.”[289] A few years later, in Martin, the Court insisted that, although the government could not decide for homeowners whether they wished to receive door-to-door canvassers, the legislature could protect the right of the individual to decide for themself “whether distributers of literature may lawfully call at a home” by making “it an offense for any person to ring the bell of a householder who has appropriately indicated that he [or she] is unwilling to be disturbed.”[290] And a few decades later, in Lamont v. Postmaster General, Justice Brennan similarly acknowledged the importance of “fully safeguard[ing] . . . the sensibilities of the unwilling recipient” by protecting her from receiving speech she would rather not.[291]

The cumulative force of these decisions is that, as the Court stated in 2000, “[t]he unwilling listener’s interest in avoiding unwanted communication has been repeatedly identified” and affirmed in First Amendment jurisprudence.[292] This is for good reason. Regulation of such unwanted direct communications, as Sotomayor recognized, poses little threat to the values the First Amendment is supposed to vindicate. Such speech, not spoken to a public audience but to an uninterested recipient alone, impinges very little on the “uninhibited, robust, and wide-open” public discussion that the First Amendment protects.[293] Such communications are “highly unlikely to persuade or inform anyone” when the sole recipient does not want to receive them.[294] Indeed, in Counterman, Whalen stopped opening the messages at all, so there was no chance she could be informed, let alone persuaded, by them. And laws that prohibit communications with unwilling recipients do not constrain the speaker from communicating their ideas, to the extent their statements involve them, to anyone else.

For these reasons, the right to be left alone prohibits only speech that does little to advance the search for truth or promote democratic self-governance.[295] And yet it is a right that none of the low-value categories of exception work to protect.[296] The Stevens framework therefore risks leaving it under-enforced. This provides another very significant reason why Stevens is simply not the correct framework to apply to these kinds of privacy- and autonomy-protecting laws—at least not if we think that the right to be free of unwanted communication deserves constitutional solicitude.

Finally, Justice Sotomayor noted that stalking laws might trigger less constitutional concern than other kinds of content-related laws because they penalize speech only when it is repeated. This matters to the constitutional analysis, she explained, because “while there is considerable risk with a single intemperate utterance that a speaker will ‘accidentally or erroneously incur liability,’ that risk is far reduced with a course of repeated unwanted contact.”[297] Stalking laws are therefore likely to impose far less of a chill on what we might consider the ordinary practices of public discourse than are laws that punish single speech acts.

One can see this rather vividly if one compares the Colorado stalking law under which Counterman was convicted with the law at issue in the last important true threats case the Court decided, Elonis v. United States.[298] The defendant in that case was convicted under the federal threats law, 18 U.S.C. § 875(c), which makes it a crime to transmit in interstate or foreign commerce “any communication containing . . . any threat to injure the person of another.”[299] Because the law required only a single instance of a threat—and, on its face, appeared to require only general intent—it could be used to punish someone who made an offhand comment, on a social media site, about their desire to kill another, or blow things up, or commit some other act of violence.[300] And indeed, similar laws have been used in the past to punish similarly offhand comments.[301] The chilling effect of this feature of the law is what led the Court, in Elonis, to read a recklessness mens rea into the law and, in Counterman, to find such a mens rea to be constitutionally compelled for threats prosecutions.[302]

In contrast, like most stalking laws, the Colorado law under which Counterman was convicted not only requires that the emotionally distressing communications be made on multiple occasions but also requires proof that those speech acts have a significant emotionally distressing effect on their recipient, and reasonably so.[303] All of these requirements cabin the reach of the law and ensure that what it punishes is sustained and abusive communication, not an isolated, and therefore potentially misconstrued, speech act.

These features of stalking laws make them much less likely to chill ordinary communication. This is not to say that these kinds of laws pose no risk of chilling valuable speech. As we explore below—and as the stalking cases have long recognized—like any laws that sanction speech, stalking laws can be abused and require careful crafting to ensure that they further, rather than undermine, free speech values. But the reduced risks of chilling that stalking laws pose means that it isn’t necessary to read into these laws all of the stringent mens rea and actus reus requirements that courts read into true threats laws to safeguard the vitality of public discourse and guard against political abuse. The two kinds of laws function very differently. To assume that safeguarding First Amendment values requires courts to treat stalking speech as a kind of true threat is to make a fundamental category error.

This was certainly what decades of pre-Stevens case law had recognized. Indeed, all of the points that Justice Sotomayor made in her concurrence about stalking laws had been made previously, in the many pre-Stevens decisions in which state and federal courts grappled with the constitutional questions raised by these kinds of laws.[304] But Sotomayor’s concurrence suggests that these older cases might still be relevant in a post-Stevens world: she suggests, implicitly, that what James Weinstein describes as the Stevens “All-Inclusive Approach” should apply only to laws that regulate decontextualized speech acts that address a broad, public audience, rather than those that occur in relationship to, or are directed towards, a particularized other.[305] In other words, Justice Sotomayor’s opinion suggests a less totalizing view of Stevens and a more multi-dimensional understanding of the First Amendment and its boundaries. This is a view that courts should embrace not only because doing so will help them avoid resorting to the kind of doctrinal gymnastics outlined above, but also because doing so will make it possible to assess the difficult First Amendment questions raised by stalking laws on their own terms, rather than by attempting to squash them into categories of speech regulation with different histories and different purposes.

B.      The Resuscitated First Amendment Law of Stalking

To say that courts should assess the constitutionality of stalking laws on their own terms is, of course, only to beg the question: What are the constitutional principles that should govern the prosecution of stalking-by-speech? What kinds of speech may the legislature criminally sanction—or, alternatively, civilly restrain—in order to protect the privacy, safety, and autonomy of others?[306]

These are difficult questions to answer because, even if stalking laws do not pose the same threat of government manipulation of the marketplace of ideas that other, more squarely content-based laws do, they nevertheless require courts to reconcile two competing but important sets of interests: on the one hand, the right of listeners not to be targeted by unwanted communications; on the other hand, the right of speakers to communicate unpleasant or inconvenient or simply unwanted ideas and opinions. Ensuring that the government has sufficient power to protect the first set of interests without doing too much harm to the second set of interests—the task that any constitutional law of stalking must accomplish—is no easy feat. In this Article, we cannot do justice to all the dimensions and aspects of that question. For our purposes, it is enough to sketch out a few basic principles that have in the past guided, and we think should in the future continue to guide, the constitutional law of stalking.

The first and most important of these is the principle that stalking laws should not be written so broadly that they can apply to speech that contributes to what the Court has described as the public “exposition of ideas.”[307] This limitation undergirds virtually all of the stalking and telephone harassment cases, as Part II makes clear.[308] It also reflects a principle that pervades the American free speech tradition more broadly: namely, that safeguarding pluralist democracy requires tolerating a great deal of hurtful and offensive and even malicious speech.[309] In other words, when it comes to democratically important speech, it is only when “substantial privacy interests are being invaded in an essentially intolerable manner” that the government can step in to protect the right of a listener to be free of unwanted communications.[310]

This principle means that stalking laws need to be drafted, enforced, and interpreted in a manner that does not criminalize, or chill the expression of, speech on matters of public concern—speech about politics, or religion, or any other “issues about which information is needed . . . to enable the members of society to cope with the exigencies of their period.” [311] Nor should stalking laws sweep so broadly they could be used to criminalize the expressive practices (interviewing sources, talking to politicians, canvassing voters) that create this kind of publicly important, or at least publicly relevant, speech. Most of the time, stalking prosecutions will not violate these principles. This is because the speech that typically constitutes the act of stalking is usually either very personal or very mundane, or both. Demands for another chance at love, requests to meet, personal images of the victim’s naked body, and other similarly intimate expressive acts may cause the person they target great anxiety, pain, and fear, but they do not provide information that helps “members of society . . . cope with the exigencies of their period.”[312] To see this, one need only consider the speech for which Counterman was convicted. “Good morning sweetheart,” “I am going to the store would you like anything,” and “Staying in cyber life is going to kill you” may have been relevant (terrifyingly so) to Coles Whalen, but they do not provide any useful or interesting information to anyone else.[313] This is not speech that contributes to the public “exposition of ideas.”[314]

This is not always the case when it comes to stalking speech, however. Consider, for example, the reporter who persistently calls a politician to talk to him about a politically damaging scandal the politician is implicated in, or the outraged consumer who publishes a series of damning blog posts about the character failings of a prominent corporate executive. These acts of speech could plausibly meet the definition of stalking under some state laws, including the one Counterman was convicted under. After all, both examples involve repeated communications that could reasonably cause another person “to suffer serious emotional distress” and might, in both those hypothetical situations, actually cause that other person serious emotional distress.[315] Yet, as the decision in Thornhill makes quite plain, this is not the kind of speech that the government is ordinarily permitted to sanction, and it is not the kind of speech we would want it to.

The potentially problematic scope of stalking laws may not always be accidental. Legislatures may intentionally structure stalking or harassment laws to permit prosecutors to target this kind of democratically important speech. Consider, for example, the “hunter harassment law” that the Wisconsin legislature enacted in 1990 and amended in 2015, which makes it a crime to “[e]ngag[e] in a series of 2 or more acts carried out over time, however short or long, that show a continuity of purpose and that are intended to impede or obstruct a person who is engaged in lawful hunting.”[316] The language of the statute clearly tracked and was modeled after the Wisconsin criminal stalking law (as its nickname suggests).[317] But, as the Seventh Circuit found when it struck the law down, the law was clearly intended to “target the expressive activities of . . . anti-hunting advocates” by making it harder for them to film people hunting and distribute those images publicly.[318] The legislature evidently hoped that dressing up what was, as one supporter put it, an attempt to ensure that hunters are not “scrutinized by those that do not share [their] views” in the garb of a stalking law, it would escape serious judicial (and perhaps public) scrutiny.[319] This is obviously a move that should be resisted. As the majority in Stevens explained at great length, hunting videos can add a great deal to the public debate about public matters that the First Amendment protects, if only by illustrating how gruesome the act of hunting can be.[320]

Courts should therefore insist, as the Seventh Circuit did when it struck down the hunter harassment law, that even when the elements of a stalking prosecution are successfully established, sanctions cannot be imposed on speech that “contribute[s] to public discourse.”[321] This won’t always be easy to do. As Part I made clear, and as the Supreme Court itself has noted, judges have often disagreed about what kinds of speech contribute to the public debate about public matters that the First Amendment protects against government regulation.[322] But while there will be hard cases, for the reasons provided above, in most stalking-by-speech cases it will be relatively straightforward to distinguish these two kinds of speech. This is because, often, one of the primary factors that distinguishes speech on matters of private concern from speech on matters of public concern is a pre-existing (or desired or presumed) private relationship, out of which the speech emerges.[323] This should hardly be surprising, given that what distinguishes the public sphere as a sociological and communicative space is the fact that it involves communication between and among strangers who have no ongoing bond.[324] The private sphere is instead comprised largely of relationships between those we know, or want to know, or, at least, those with whom we have relationships that extend beyond the mere fact of the communication.

The First Amendment cases have long recognized as much and, for that reason, have tended to treat speech that occurs within a private relationship as speech that touches on matters of private concern—at least when it relates solely to the details of that relationship, particularly the intimate details. Courts have tended to conclude that, absent special circumstances—allegations of criminal activity, or a close connection to a social movement like the Me Too movement—this is not the kind of information that members of the public have any need to know about.[325] For the most part, then, reconciling stalking laws with the Thornhill principle will not be terribly difficult because it is this kind of intimate speech that most stalking involves. But it is important that courts make sure that the category of stalking is interpreted in a manner that is consistent with the Thornhill rule, given the democratic interests the First Amendment protects.

Second, courts should permit legislatures to punish as stalking only speech that serves no legitimate purpose. This principle, which may at first seem simply a repetition of the prior principle, in fact expands upon it. This is because speech may serve legitimate, even important aims, even when it does not touch on matters of public concern—as courts have, in the past, recognized. As Judge Mansfield put the point in his concurring opinion in the influential Second Circuit decision in Gormley v. Director, Connecticut State Department of Probation:

Telephone calls by irate citizens to their Congressmen, by collectors seeking payment of legitimate bills overdue, by customers voicing to a seller dissatisfaction with goods or services purchased, and calls of like tenor, are likely to be annoying, even harassing, to the recipients. Such calls are usually made to influence the person called. [But n]o one could seriously question the caller’s free speech right under the First Amendment to engage in such telephone conversations.[326]

The first example that Judge Mansfield provides clearly involves speech on matters of public concern. But the other examples do not. And yet, it would be a very impoverished understanding of the free speech principle that denied speakers the right to engage in this important, but private, speech. After all, as Jack Balkin has argued, the First Amendment protects not only the communicative processes that are necessary to the operation of democratic government, but something broader that he calls a “democratic culture.”[327] Surely one of the hallmarks of a democratic culture is a system of government in which the government intrudes as little as possible on the private expressive choices of members of the democratic political community, even when these do not touch on matters of broad public concern.

Certainly, many stalking statutes recognize as much by making it an element of the crime that the communication serve “no legitimate purpose.”[328] But even in cases where they do not, courts should safeguard from the reach of stalking laws the unwanted but legitimate acts of communication, like those Mansfield listed above, that emerge out of relationships (between government officials and those they represent, or sellers and buyers, or creditors and debtors) in which one party has a right to be heard by the other. They should instead reserve stalking laws for the very important classes of cases in which the speaker has no right to be heard but insists on speaking anyway. To do otherwise would undermine the vitality of the rich private sphere of expressive activity on which the more familiar public sphere depends (public opinions are, after all, often formed by means of private communications, such as those around the kitchen table, on the telephone, in a bar, etc.).

Finally, and relatedly, prosecutors in stalking cases should have to prove that the defendant possessed some kind of wrongful mens rea—even if not the mens rea required to convict someone for making true threats. Justice Sotomayor suggested as much in her concurrence. She asserted that, in cases like Counterman’s, which involved stalking (supposedly) by means of threats, a showing of “recklessness [would be] amply sufficient,” but that a lower standard of mens rea would be required in cases where the prosecution did not depend on the content of the speech.[329]

The requirement that prosecutors prove some kind of culpable mens rea flows directly from the requirement that they establish that the speech that constitutes the stalking served no legitimate purpose. It ensures that stalking prosecutions are not in fact used to chill misguided but not malicious or illegitimate speech. For this reason, what mens rea is required may differ in different contexts and circumstances. In some circumstances, it may be sufficient to establish that the defendant knew or should have known that his speech was unwelcome. Consider, for example, the facts of the Counterman case. There is no reason to believe that Counterman texted Whalen so many times because he wanted to harass or scare her. Like many stalkers, he instead hoped to win her love and to sustain what he believed to be a pre-existing relationship.[330] Nevertheless, after she blocked his messages multiple times, Counterman surely should have recognized that his understanding of their relationship and its prospects was erroneous. At this point, he was not engaged in anything we might call legitimate activity when he continued to demand that Whalen communicate with him.

In other contexts, meanwhile, a higher burden may be required to avoid chilling important or legitimate speech. Consider the facts of United States v. Sryniawski.[331] In that case, the ex-husband of the wife of a candidate for the state legislature sent a number of emails to the candidate’s campaign e-mail address that accused the candidate’s stepdaughter of being a pedophile and included links to two websites that established that the stepdaughter had been charged with child solicitation and sexual assault charges in another state.[332] The emails also accused his ex-wife of making up the fact that she had been molested as a child at the hands of her father (a story she had told him in confidence while they were married) in order to obtain prescription drugs and threatened to continue to expose the family’s secrets unless the candidate “Quit the Race” and “Never [ran] for any Political Office again.”[333] None of these emails clearly involved speech on matters of public concern. The fact that the stepdaughter of a political candidate was at one point in time charged with a crime is not the kind of speech that is or should be relevant to voter’s decisions about the candidate’s merits. None addressed a public audience, or provided any information that could help voters cope with the exigencies of their time. The emails might have served a legitimate purpose, had the defendant’s intent in sending these emails been to discover more information about the candidate, or to encourage him to disclose this information himself to the public. But the same would not be true were the purpose (as it likely was) to harass and intimidate the candidate in an attempt to punish the ex-wife for leaving him. In cases like these, a higher mens rea than negligence should therefore be required to validly prosecute the defendant for stalking.

This kind of calibration of the mens rea that is constitutionally required is not new to First Amendment law. For example, the First Amendment law of defamation requires proof of differing mens rea, depending on the status of the defendant and the kind of speech involved, in order to account for the different kinds of chilling effects that liability can impose in different contexts and the differing abilities of variously situated defendants to engage in effective counterspeech.[334] The considerations relevant to the inquiry in stalking cases will obviously be different than those relevant in defamation cases in some respects. But the bottom line is that, however the analysis plays out, courts should insist on proof of whatever mens rea is necessary to prevent the specter of stalking prosecutions from imposing too great a chill on the willingness of private persons to speak to one another in intemperate but potentially valuable ways.

Crafting such a rule will, of course, require courts to take positions on what are or are not legitimate or valuable modes of communication. But this should be based on contextual and sociological factors, rather than a judgment of the value of the content of the speech in any particular case. As the First Amendment has recognized in other contexts, there is no other means by which to effectively reconcile the competing interests implicated by the prosecution of stalking-by-speech.

Professor Eugene Volokh has suggested that, instead of this highly value-laden inquiry, courts should distinguish permissible from impermissible stalking prosecutions on the basis of the kinds of speech they involve. Prosecutions based on what Professor Volokh calls “one-to-one” speech—“speech said to a particular person in a context where the recipient appears not to want to hear it, whether because the recipient has expressly demanded that the speech stop or because the speaker intends to annoy or offend the recipient”[335]—should ordinarily be considered consistent with the First Amendment. In contrast, prosecutions based on “one-to-many” speech—speech addressed to listeners who are not themselves the targets of the stalking and not necessarily unwilling to hear it—should be considered unconstitutional. This is because, Professor Volokh argues, one-to-many speech looks a lot more like the public discourse that the First Amendment protects. Because “it involves the expression of facts and opinions aimed at informing and persuading potentially willing listeners,” and in that respect resembles conventional forms of public discourse (political speeches, advertisements, etc.), it cannot serve as the basis of a stalking prosecution without violating the First Amendment.[336]

Such a rule would certainly provide a simple means of distinguishing between permissible and impermissible stalking prosecutions. But that simplicity would come at serious cost. This is because there is good reason to think that, in a social media age, a great deal of destructive and malicious communication can take the form of “one-to-many” speech. Indeed, stalkers may use one-to-many speech as much as one-to-one speech to exert the kind of domination over their victims that stalking laws are aimed at protecting against. In order to coerce, control, and punish their victims, stalkers publicly humiliate them, spread rumors about them to their friends and family members, and blog about them, or at least threaten to do so—as the facts of Petrovic, Osinger, and Ackell all make clear.[337] And the internet makes this kind of speech more possible and pervasive, as victims and their social networks are much more visible and accessible from a distance than ever before.

And yet, although one-to-many speech looks more like what we typically think of as public discourse, it is not always public discourse. A picture of a private individual, published without her consent, with the tag line “I am just a whore 4 sale” contributes nothing—or at least, nothing significant—to the public debate about matters of public concern.[338] Even if it is not directed to its victim, this speech nevertheless emerges out of and relates to a private relationship. It is an attempt to use the norms and forms of public discourse to carry out what are essentially private affairs. And although maliciously intended speech of this kind may communicate useful information by revealing, as Volokh puts it, “who in the [audience’s] social circle is potentially untrustworthy or even dangerous,” these informational benefits are outweighed by the serious harms this kind of reputationally damaging speech can cause both to free speech values and to other interests when done not to contribute to public discourse but to harass, intimidate, or control.[339] As Danielle Citron and others have pointed out, stalking-by-speech can have a profoundly silencing effect on its victim.[340] And it can and does cause great distress, fear, and insecurity.[341]

Allowing the government to protect against these harms, even when it comes to speech directed at a public audience, is also not at all new when it comes to First Amendment law. Indeed, this is the entire justification for the doctrine of true threats. As Justice Scalia explained in R.A.V., “the reasons why threats of violence are outside the First Amendment” is because of the need to “protect[] individuals from the fear of violence, from the disruption that fear engenders, and from the possibility that the threatened violence will occur.”[342] The true threats cases grant the government the power to protect individuals against these kinds of emotional effects because they recognize the serious impact the speech can have on its targets’ sense of autonomy and agency. As Kenneth Karst notes of death threats, “when fear is severe enough, it takes away your normal life and seriously diminishes your sense of self.”[343] This loss of autonomy, or what Karst describes as a “deformation of self,” has obvious implications for an individual’s ability to participate in public discourse. Therefore, courts permit the government to punish threats not only to protect the welfare of the individual, but to protect the collective interest of us all in vibrant, inclusive, and uninhibited public debate.[344]

All of these justifications apply equally to cases involving stalking-by-speech. As we discussed in Part II, the core purpose of stalking laws is to protect individuals against fear and insecurity. Indeed, “[f]ear is the fundamental justification for stalking laws. . . . [W]ithout fear, there is no crime of stalking.”[345] And, as with threats, the fear stalking victims feel can be very real, very damaging, and a very common precursor to harm. Indeed, a stalking victim’s sense of fear “is a fairly accurate and robust predictor of actual violence,” regardless of whether it arises from public or private speech.[346] Thus, the governmental interest in protecting against the disruptive effects of fear and the possibility of actual violence is as compelling in the context of stalking as it is in the context of threats. No doubt, this familial similarity of threats laws and stalking laws explains why courts keep reaching for the true threats category in trying to categorize stalking laws. But the particular pathologies of stalkers mean that explicit threats of harm—the kind generally required to punish speech as a true threat—are much less common. Hence, there is a good reason to think that the government should also be constitutionally permitted to punish stalking speech when that speech causes significant fear and is done intentionally, knowingly, or, as Justice Sotomayor’s concurrence suggests, with reckless disregard for the possibility that the speech will cause great fear or emotional distress in the person it targets, even when it takes the form of emails to the victim’s friends and family or blog posts about her.

Whether one agrees with this conclusion or not, the point is that courts should be engaging in this kind of careful consideration of the specific interests, concerns, and dangers that laws regulating stalking-by-speech implicate when deciding their constitutionality. They should not engage in the entirely distinct and often highly conclusory exercise in box-checking that a rigid reading of Stevens encourages.

C.      Seeing Past Stevens

Recognizing the limited reach of the Stevens test has implications beyond just the law of stalking (although, given the volume and seriousness of the stalking cases, that is no small “just”). This is because stalking laws are not the only kinds of laws that operate in ways that are quite distinct from the animal cruelty or violent video game laws that were the focus of Stevens and Brown. Indeed, stalking laws are not the only kinds of laws that regulate something other than isolated expressive utterances, made to a broad public audience, in the public sphere.

The same is obviously true of harassment laws, which differ from stalking laws only in that they punish one-time utterances and do not always require that the victim suffer emotional distress.[347] These distinctions may mean that harassment laws need to include more stringent mens rea requirements than stalking laws require or may have to be limited in other ways to vindicate First Amendment values. But otherwise, the same general considerations that guide the analysis of stalking laws should guide the analysis of harassment laws, as the overlapping history discussed above suggests.

Laws that prohibit the non-consensual disclosure of intimate imagery (“NCII,” or colloquially known as “revenge porn”) also resemble stalking laws in many ways.[348] Like stalking laws, they punish speech that emerges out of and relates to a social relationship—typically speech that is motivated by the destruction of that relationship—rather than speech that expresses ideological commitments or aesthetic desires and that is addressed to a general public audience.[349] Like stalking laws, they regulate speech with respect to, but not solely on the basis of, its harmful content.[350] And like stalking laws, they do not regulate a type of speech that fits easily into any of the recognized low-value categories, but they nevertheless protect what both courts and scholars have recognized to be important autonomy and privacy interests.[351]

For this reason, the constitutional analysis in NCII cases has been almost as tortured as the analysis in the stalking cases. Five state appellate courts have upheld their state NCII laws against First Amendment challenges, but they have often had to engage in complicated and problematic doctrinal gymnastics to do so.[352] In People v. Austin, for example, the Illinois Supreme Court held that the state NCII law was content neutral rather than content based because the law penalized only the dissemination of sexual images that were “obtained . . . under circumstances in which a reasonable person would know that the image was to remain private and knows or should have known that the person in the image has not consented to the dissemination.”[353] Because the NCII law did not prohibit all sexual images, the Court concluded that the law regulated speech on the basis of its manner, not its content, and therefore was constitutional.[354] However, this conclusion, which echoes some courts’ approach to stalking laws, is impossible to square with the Supreme Court’s holding in Reed v. Town of Gilbert that laws are content based whenever they make distinctions between speakers based on the subject matter of their speech.[355] Similarly creative doctrinal gymnastics characterize many of the other state supreme court decisions that affirm the constitutionality of NCII laws.[356] These maneuvers reflect, just as they do in the stalking context, how difficult courts have found reconciling their intuitions about what kinds of laws actually threaten free speech values with the map of the First Amendment that Stevens provides.

The anti-doxing laws enacted in a number of states in recent years also bear a striking similarity to stalking laws. Like stalking laws, these laws impose liability on those who disseminate speech about a particular person and do so with the purpose of harming them or with knowledge that their actions will have that effect. In some cases, they also require proof that the dissemination of private information did in fact have that effect. Here too, we see the law attempting to grapple with the ways in which new technology can permit strangers and non-strangers to harass, intimidate, and punish one another.

As these examples make quite clear, it would be a mistake to think of stalking laws as an isolated exception to the general Stevens framework. What they reveal is an entire category of speech regulations that do not fit easily within the map of the First Amendment that Stevens provides, but also do not fit into any other well-recognized institutional exception, such as, for example, the exceptions for student speech or government speech.

As the examples of doxing and NCII also suggest, developments in technology are likely to push lawmakers to continue to expand on this category of laws and therefore put only further pressure on the Stevens framework. Of course, “the basic principles of freedom of speech and the press . . . do not vary when a new and different medium for communication appears.”[357] But new technologies are likely to enable modes of communication that do not neatly resemble the historically identified categories of exception, even if they cause the same kinds of harms and threaten First Amendment values in the same way. As Andrew Koppelman observes, a rigid reading of Stevens means that “harms will accumulate and the law will be helpless to remedy them.”[358]

We take no position here on how specifically First Amendment doctrine should address the problem of harassment, or NCII, or doxing—or, for that matter, other kinds of expression that are likely to pose similar problems. Instead, our argument is only that these problems should be dealt with head-on and not through the distorted kaleidoscope produced by a rigid interpretation of Stevens’s categories. This is not to engage in a “free-floating test for First Amendment coverage” that Stevens declares “startling and dangerous.”[359] It is to accept only the obvious reality that a complete map of the vast legal terrain that is speech regulation was not, and should not be taken to have been, comprehensively described by the Court in a single paragraph in 2010.

To the extent that claims about the First Amendment’s history and tradition justify the Stevens test, these claims rely on an inaccurate account of the past, as one of us has written elsewhere and this Article has further shown.[360] And to the extent that Stevens is justified as a methodological tool to provide a more determinate and restrained mode of judicial reasoning, Counterman and the stalking cases have shown that this is an inaccurate forecast of the future. Courts, including the Supreme Court, will struggle when they encounter novel problems that cannot neatly be slotted into the boxes Stevens describes, but they will not simply abandon their judicial intuitions and engage in dry, rote reasoning based on taxonomical labels.

What the stalking cases show, in short, is that Stevens cannot bear the doctrinal weight it is being asked to carry and cannot and should not be regarded as providing a comprehensive test of the First Amendment’s boundaries. Instead, the First Amendment tradition is far richer than Stevens reflects.

Conclusion

Stalking is a very serious social problem. Justice Sotomayor was right to warn that “[o]verly constraining our society’s ability to respond to stalking would come at a real cost.”[361] Indeed, in 2019, 3.4 million people were victims of stalking in the United States, a number that represents profound physical, mental, and social harms.[362] As this Article has documented, stalking also poses a serious doctrinal problem for courts trying to make sense of the government’s power to regulate this kind of harmful speech in a post-Stevens terrain. This Article also suggests that the doctrinal problem of stalking, at least, need not be quite as difficult as courts have found it in recent years, because the First Amendment provides more latitude to legislatures to protect individuals from this kind of fear-inducing speech than a superficial reading of the Court’s recent precedents implies. Recognizing as much is only going to be increasingly important in a rapidly changing world where technological advances create new forms of communication and new types of communicative harms. This Article has accordingly challenged the conventional story of the First Amendment’s past in order to open up new avenues for courts to imagine its future.

Copyright © 2025 Lakier & Douek

                *     Professor of Law & Herbert and Marjorie Fried Teaching Scholar, University of Chicago Law School.

             **     Assistant Professor, Stanford Law School.

This article builds on an amicus brief we filed in Counterman v. Colorado with Eugene Volokh, and we thank and acknowledge Eugene for his contributions to that brief, as well as Charlotte Taylor and Alyxandra Vernon for their incredible work on it. We would like to thank Helen Norton, Eugene Volokh, participants at the University of Chicago Work in Progress Workshop, the Stanford Law School Faculty Workshop and the Free Expression Scholars Conference at Yale Law School for very helpful comments, and the editors of the California Law Review for their excellent editorial help.

          [1].     Counterman v. Colorado, 600 U.S. 66, 82–83 (2023).

          [2].     Id.

          [3].     See Adam Liptak, Supreme Court Puts First Amendment Limits on Laws Banning Online Threats, N.Y. Times (June 27, 2023), https://www.nytimes.com/2023/06/27/us/politics/supreme-court-first-amendment-colorado.html [https://perma.cc/693E-FCH4]; Melissa Quinn, Supreme Court Sets Higher Bar for Prosecuting Threats Under First Amendment, CBS News (June 27, 2023), https://www.cbsnews.com/news/supreme-court-threats-counterman-colorado-first-amendment/ [https://perma.cc/777B-5E7Q].

          [4].     See, e.g., Jennifer E. Rothman, Freedom of Speech and True Threats, 25 Harv. J.L. & Pub. Pol’y 283 (2001) (analyzing courts’ various approaches when considering First Amendment threat cases).

          [5].     Justice Barrett wrote a spirited dissenting opinion, which Justice Thomas joined. For scholarly criticism of the opinion, see Danielle Keats Citron, From Bad to Worse: Stalking, Threats, and Chilling Effects, 2023 Sup. Ct. Rev. 175 (2024).

          [6].     Colo. Rev. Stat. Ann. § 18-3-602(1)(b) (West).

          [7].     People v. Counterman, 497 P.3d 1039, 1043 (Colo. App. 2021).

          [8].     Id. at 1044.

          [9].     Colo. Rev. Stat. Ann. § 18-3-602(1)(c) (West).

        [10].     Counterman v. Colorado, 600 U.S. 66, 106 (2023) (Barrett J., dissenting); see also id. at 69 (“True threats of violence are outside the bounds of First Amendment protection and punishable as crimes. Today we consider a criminal conviction for communications falling within that historically unprotected category.”).

        [11].     559 U.S. 460 (2010).

        [12].     Id. at 468–69.

        [13].     Id. at 472.

        [14].     State v. Katz, 179 N.E.3d 431, 453 (Ind. 2022) (concluding that, given the Court’s reluctance to recognize new categories, “[t]his Court ‘leave[s] it to the Supreme Court in the first instance to designate nonconsensual pornography as a new category of speech that falls outside the First Amendment’s full protections.’”); People v. Austin, 155 N.E.3d 439, 455 (Ill. 2019) (same); State v. VanBuren, 214 A.3d 791, 801–02, 806–07 (Vt. 2019) (same); In re Welfare of A. J. B., 929 N.W.2d 840, 846 (Minn. 2019) (recognizing that “[t]he Supreme Court has been reluctant to expand these categories of unprotected speech” and declining to identify a new category as a result).

        [15].     See infra notes 162–175.

        [16].     Stevens, 559 U.S. at 468 (quoting Simon & Schuster, Inc. v. Members of N.Y. State Crime Victims Bd., 502 U.S. 105, 127 (1991) (Kennedy, J., concurring in judgment)).

        [17].     Id. at 470.

        [18].     See infra Part III.A.

        [19].     See Katrina Baum, Shannon Catalano & Michael Rand, Bureau of Just. Stat. Special Rep., National Crime Victimization Survey: Stalking Victimization in the United States 1 (Jan. 2009), https://permanent.fdlp.gov/gpo31367/stalking-victimization.pdf [https://perma.cc/NQ9J-3637] (defining stalking as “a course of conduct directed at a specific person that would cause a reasonable person to feel fear”); Julie Miles Walker, Anti-Stalking Legislation: Does It Protect the Victim Without Violating the Rights of the Accused?, 71 Denv. U. L. Rev. 273, 276 (1993) (“The activity of stalking involves repeated following, harassing, or threatening another or acting in such a way as to create a credible threat of harm in the mind of the victim.”).

        [20].     See Patricia G. Tjaden, Stalking Policies and Research in the United States: A Twenty Year Retrospective, 15 Eur. J. Crim. Pol’y & Rsch. 261, 262 (2009); see also Patricia Tjaden & Nancy Thoennes, Stalking in America: Findings From the National Violence Against Women Survey, Nat’l Inst. Just. Ctrs. for Disease Control & Prevention 8 (Apr. 1998), https://www.ojp.gov/pdffiles/169592.pdf [https://perma.cc/5G8A-RJ2U] (highlighting the report’s conclusions that “stalkers do not always threaten their victim verbally or in writing”).

        [21].     For example, in its report on stalking victimization in the United States, the Department of Justice listed as typical stalking behaviors “making unwanted phone calls . . . sending unsolicited or unwanted letters [and] e-mails . . . leaving unwanted items, presents, or flowers, [and] posting information or spreading rumors about the victim on the internet, in a public place, or by word of mouth,” as well as “following or spying on the victim, showing up at places without a legitimate reason, [and] waiting at places for the victim.” Baum, Catalano & Rand., supra note 19; see also U.S. Dep’t of Just., 1999 Report on Cyberstalking: A New Challenge for Law Enforcement and Industry (Aug. 1999), https://webharvest.gov/peth04/20041022072652/http://www.usdoj.gov/criminal/cybercrime/cyberstalking.htm [https://perma.cc/XXR5-FY75] (listing examples of serious stalking that occurred entirely through language).

        [22].     See infra Part III.

        [23].     Notwithstanding their importance to the Stevens schema, there is no universally agreed upon list of historically recognized categories of low-value speech. But there is broad consensus that the list includes, at minimum, the categories of “obscenity, defamation, fraud, incitement, and speech integral to criminal conduct,” Stevens, 559 U.S. 460, 460 (2010), “fighting words,” Brown v. Ent. Merchs. Ass’n, 564 U.S. 786, 791 (2011), and “true threats,” Virginia v. Black, 538 U.S. 343, 359 (2003).

        [24].     See infra Part II.A.

        [25].     Id.

        [26].     United States v. Alvarez, 567 U.S. 709, 717 (2012) (identifying this as a historically recognized category of low-value speech).

        [27].     See notes 90–94 and accompanying text.

        [28].     602 U.S. 286 (2024).

        [29].     Id. at 301 (plurality opinion) (“The Lanham Act’s names clause has deep roots in our legal tradition. Our courts have long recognized that trademarks containing names may be restricted. And, these name restrictions served established principles. This history and tradition is sufficient to conclude that the names clause—a content-based, but viewpoint-neutral, trademark restriction—is compatible with the First Amendment.”). Justices Gorsuch and Alito joined with Justice Thomas on this portion of the opinion.

        [30].     Id. at 311 (Kavanaugh, J., concurring in part) (concurring in the conclusion that “the names clause is constitutional, particularly in light of the long history of restricting the use of another’s name in a trademark” but noting that “a viewpoint-neutral, content-based trademark restriction might well be constitutional even absent such a historical pedigree”); id. at 311–12 (Barrett, J., concurring in part) (concluding that content-based but viewpoint-neutral trademark “restrictions, whether new or old, are permissible so long as they are reasonable in light of the trademark system’s purpose of facilitating source identification”). Chief Justice Roberts joined Justice Kavanaugh’s opinion. Justices Sotomayor, Kagan, and Jackson joined Justice Barrett’s opinion.

        [31].     Id. at 312 (Barrett, J., concurring in part).

        [32].     U.S. Const. amend. I.

        [33].     As we explain below, the Court’s increasing unwillingness to recognize distinctions between different kinds of public speech has led it to extend constitutional protection to speech it believes to possess only marginal social value or, as the Court put it with respect to the expressive practice of nude dancing, speech “within the outer perimeters of the First Amendment.” Barnes v. Glen Theatre, Inc., 501 U.S. 560, 566 (1991). For most intents and purposes, however, the same First Amendment rules that apply to much more valuable speech are applied to this kind of marginal expression. Hence, the term “high-value speech” generally includes all speech with more than de minimis value.

        [34].     Frederick Schauer, The Boundaries of the First Amendment: A Preliminary Exploration of Constitutional Salience, 117 Harv. L. Rev. 1765, 1784 (2004).

        [35].     Id. at 1767 (noting that “questions about the involvement of the First Amendment in the first instance are often far more consequential [to case outcomes] than are the issues surrounding the strength of protection that the First Amendment affords the speech to which it applies”).

        [36].     See infra Part II.B; see also Frederick Schauer, The Politics and Incentives of First Amendment Coverage, 56 Wm. & Mary L. Rev. 1613, 1624 (2015) (describing the language in Stevens as “especially important” to understanding the “politics, psychology, and sociology of [First Amendment] coverage”).

        [37].     For clarity’s sake, in what follows, we use the distinction between First Amendment coverage and protection that Fred Schauer first developed to distinguish between speech that is understood to implicate First Amendment values, and thus whose regulation triggers heightened judicial scrutiny (covered speech), and speech that is affirmatively protected against government regulation after heightened judicial scrutiny has been applied (protected speech). See Schauer, supra note 34, at 1769 (“The acts, behaviors, and restrictions not encompassed by the First Amendment at all . . . are the ones that are simply not covered by the First Amendment. . . . When the First Amendment does show up, the full arsenal of First Amendment rules, principles, standards, distinctions, presumptions, tools, factors, and three-part tests becomes available to determine whether the particular speech will actually wind up being protected.”).

        [38].     Genevieve Lakier, The Invention of Low-Value Speech, 128 Harv. L. Rev. 2166, 2179–80 (2015) (“Courts did little to delimit the boundaries of the constitutional categories of speech and press because they did not need to. . . . The freedom that the First Amendment and state provisions guaranteed . . . was freedom of expression—but not freedom from responsibility for the ill effects of what one expressed.”).

        [39].     As Jud Campbell notes, “[n]early everyone who spoke on the issue [of freedom of speech and press] agreed that well-intentioned statements of one’s thoughts were constitutionally protected,” but courts were quick to read bad intentions into statements that violated social norms. Jud Campbell, Natural Rights and the First Amendment, 127 Yale L.J. 246, 318 (2017). The result was that “even though the Founders broadly acknowledged that speaking, writing, and publishing were among their natural rights, governmental limitations of expressive freedom were commonplace.” Id. at 276. “Speaking, writing, and publishing were . . . ordinarily subject to restrictions under laws that promoted the public good.” Id.

        [40].     For discussion of the Sedition Act debates about libel, see Geoffrey R. Stone, The Story of the Sedition Act of 1798: “The Reign of Witches,” in First Amendment Stories 13, 23 (Richard W. Garnett & Andrew Koppelman eds., 2012).

        [41].     Lakier, supra note 38, at 2199–2200.

        [42].     As the Court noted in Dallas v. Stanglin, “[i]t is possible to find some kernel of expression in almost every activity a person undertakes—for example, walking down the street or meeting one’s friends at a shopping mall.” 490 U.S. 19, 25 (1989). But the expressive nature of this kind of activity has never been considered by itself “sufficient to bring the activity within the protection of the First Amendment.” Id.

        [43].     Indeed, the New Deal Court understood itself, quite explicitly, as having the responsibility of developing a body of First Amendment law that vindicated free speech values without undermining the operations of the regulatory state. See W. Va. State Bd. of Educ. v. Barnette, 319 U.S. 624, 639–40 (1943) (“[T]he task of translating the majestic generalities of the Bill of Rights, conceived as part of the pattern of liberal government in the eighteenth century, into concrete restraints on officials dealing with the problems of the twentieth century, is one to disturb self-confidence. These principles grew in soil which also produced a philosophy that the individual was the center of society, that his liberty was attainable through mere absence of governmental restraints, and that government should be entrusted with few controls and only the mildest supervision over men’s affairs. We must transplant these rights to a soil in which the laissez-faire concept or principle of non-interference has withered at least as to economic affairs, and social advancements are increasingly sought through closer integration of society and through expanded and strengthened governmental controls. . . . But we act in these matters not by authority of our competence but by force of our commissions.”).

        [44].     310 U.S. 296 (1940).

        [45].     See id. at 309–11.

        [46].     Id. at 310 (asserting that the “essential characteristic of [the] liberties [guaranteed by the First Amendment] is, that under their shield many types of life, character, opinion and belief can develop unmolested and unobstructed,” and that although “[t]here are limits to the exercise of these liberties,” they protect “resorts to exaggeration, [the] vilification of men who have been, or are, prominent in church or state, and even . . . false statement”).

        [47].     Id.

        [48].     Id. at 309–10.

        [49].     Robert Post has argued that the Cantwell Court rejected the idea that the government could constitutionally “enforce the civility rules of one community . . . against those of another” because it believed that if it did so “the state would in effect be using its power and authority to support some communities and repress others.” Robert C. Post, The Constitutional Concept of Public Discourse: Outrageous Opinion, Democratic Deliberation, and Hustler Magazine v. Falwell, 103 Harv. L. Rev. 601, 630 (1990). This is certainly true with respect to some civility rules—those that govern general public address. But the Court also clearly recognized that other civility rules (for example, the rules against calling someone by an epithet or using profanity when speaking to them) could be upheld to safeguard what it described as “good order.” Cantwell, 310 U.S. at 304. We might distinguish between these two kinds of civility rules as rules that are understood to limit the scope and diversity of public debate in a pluralist society like the United States and rules that are believed to enable it. However one understands the distinction, the important point is that Cantwell did not categorically reject the possibility of state-imposed civility rules; it simply rejected the idea that these could extend to intemperate speech that reflected a genuine “effort to persuade a willing listener to buy a book or to contribute money [to a religious cause].” Id. at 310.

        [50].     310 U.S. 88 (1940).

        [51].     Id. at 102.

        [52].     See id. at 102–03.

        [53].     See id. at 101–02.

        [54].     See, e.g., Ludwig Teller, Picketing and Free Speech, 56 Harv. L. Rev. 180, 200–01 (1942); Charles O. Gregory, Peaceful Picketing and Freedom of Speech, 26 A.B.A. J. 709 (1940).

        [55].     See infra notes 48–53 and accompanying text.

        [56].     Dun & Bradstreet, Inc. v. Greenmoss Builders, 472 U.S. 749, 759–60 (1985) (noting that while “speech on matters of purely private concern . . . is not totally unprotected by the First Amendment[,] . . . its protections are less stringent”); Connick v. Myers, 461 U.S. 138, 146 (1983) (“When employee expression cannot be fairly considered as relating to any matter of political, social, or other concern to the community, government officials should enjoy wide latitude in managing their offices.”).

        [57].     Steven Shiffrin, The First Amendment and Economic Regulation: Away from A General Theory of the First Amendment, 78 Nw. U. L. Rev. 1212, 1251–52 (1983) (noting that the Court’s pursuit of a “general balancing methodology” in its First Amendment cases produced an “eclectic approach” that reflected “the complex accommodations needed in each context”).

        [58].     Joseph Burstyn, Inc. v. Wilson, 343 U.S. 495 (1952).

        [59].     See, e.g., Pickering v. Bd. of Educ., 391 U.S. 563, 571–72 (1968).

        [60].     Roth v. United States, 354 U.S. 476, 484 (1957).

        [61].     N.Y. Times Co. v. Sullivan, 376 U.S. 254, 270 (1964).

        [62].     NAACP v. Button, 371 U.S. 415, 429 (1963).

        [63].     Snyder v. Phelps, 562 U.S. 443, 458 (2011); Hustler Magazine, Inc. v. Falwell, 485 U.S. 46, 50–52 (1988).

        [64].     Va. State Bd. of Pharmacy v. Va. Citizens Consumer Council, Inc., 425 U.S. 748, 764 (1976).

        [65].     Milk Wagon Drivers Union of Chicago, Loc. 753 v. Meadowmoor Dairies, Inc., 312 U.S. 287, 293–94 (1941).

        [66].     Id. at 293 (“[U]tterance in a context of violence can lose its significance as an appeal to reason and become part of an instrument of force. Such utterance was not meant to be sheltered by the Constitution.”).

        [67].     Id.

        [68].     315 U.S. 568, 572 (1942).

        [69].     Id.

        [70].     Thornhill v. Alabama, 310 U.S. 88, 102 (1940).

        [71].     See, e.g., Va. State Bd. of Pharmacy v. Va. Citizens Consumer Council, Inc., 425 U.S. 748, 787 (1976) (Rehnquist, J., dissenting).

        [72].     Terminiello v. City of Chicago, 337 U.S. 1, 14 (1949) (Jackson, J., dissenting).

        [73].     Roth v. United States, 354 U.S. 476, 514 (1957) (Douglas, J., dissenting).

        [74].     N.Y. Times Co. v. Sullivan, 376 U.S. 254, 299 (1964) (Goldberg, J., concurring).

        [75].     FCC v. Pacifica Found., 438 U.S. 726, 773 (1978) (Brennan, J., dissenting).

        [76].     Harry Kalven, Jr., The Metaphysics of the Law of Obscenity, 1960 Sup. Ct. Rev. 1, 10, 19 (1960).

        [77].     Kenneth L. Karst, Equality as a Central Principle in the First Amendment, 43 U. Chi. L. Rev. 20, 31 (1975).

        [78].     Roth, 354 U.S. at 514 (Douglas, J., dissenting).

        [79].     Kalven, for example, focused exclusively on the problem of obscenity. See Kalvin, supra note 76. Karst’s critique was broader, but similarly focused on public speech addressed to a public audience. See Karst, supra note 77, at 30–35.

        [80].     Karst, supra note 77, at 32 (arguing that the existence of the low-value categories “is radically inconsistent with the principle of equal liberty of expression” because although “the equality principle . . . does not prohibit all content regulation,” it permits regulation upon a showing of harm, not lack of value); Kalven, supra note 76, at 4 (“It is hard to see why the advocacy of improper sexual values should fare differently, as a constitutional matter, from any other exposition in the realm of ideas.”).

        [81].     The epitome of this tendency was the Court’s grudging extension of constitutional protection to nude dancing—a species of symbolic expression that the Justices clearly believed failed to communicate much of substance to its audience but, because it addressed a public audience, was nevertheless, the Justices recognized, “within the outer perimeters of the First Amendment, though . . . only marginally so.” Barnes v. Glen Theatre, Inc., 501 U.S. 560, 566 (1991).

        [82].     The most obvious example of this was with respect to the commercial speech cases, which over time expanded protection to a wider variety of commercial and corporate speech. See, e.g., Robert Post & Amanda Shanor, Adam Smith’s First Amendment, 128 Harv. L. Rev. F. 165, 167 (2015); Genevieve Lakier, The First Amendment’s Real Lochner Problem, 87 U. Chi. L. Rev. 1241, 1242–43 (2020).

        [83].     See Garcetti v. Ceballos, 547 U.S. 410, 418 (2006).

        [84].     See Snepp v. United States, 444 U.S. 507, 509–10 (1980).

        [85].     505 U.S. 377, 379 (1992).

        [86].     See id. at 391.

        [87].     Id. at 418 (Stevens, J., concurring) (describing the majority’s argument as “an adventure in a doctrinal wonderland” because it relied on the “fantastical” idea that speech could be both obscene and have political content and concluding that “‘[o]bscene antigovernment speech,’ then, is a contradiction in terms”).

        [88].     Id. at 383.

        [89].     Id. at 382.

        [90].     Chaplinsky v. New Hampshire, 315 U.S. 568, 571–72 (1942).

        [91].     For example, in Bridges v. California, 314 U.S. 252 (1941), the Court held that speech about an ongoing litigation could be punished for contempt of court only if it posed a clear and present danger of grave harm to the administration of justice, notwithstanding the fact that this conclusion “[broke] with the uninterrupted course of constitutional history,” as Justice Frankfurter noted in his dissent. Id. at 279 (Frankfurter, J., dissenting). For similar criticisms of the significant breaks the modern cases made with prior practice, see Near v. Minnesota, 283 U.S. 697, 723 (1931) (Butler, J., dissenting) (“The decision of the Court in this case . . . gives to freedom of the press a meaning and a scope not heretofore recognized, and . . . is without precedent”); Gertz v. Robert Welch, Inc., 418 U.S. 323, 380 (1974) (White, J., dissenting) (“[T]he Court discards history and precedent in its rush to refashion defamation law in accordance with the inclinations of a perhaps evanescent majority of the Justices”); United States v. Sineneng-Smith, 590 U.S. 371, 390 (2020) (Thomas, J., concurring) (“The overbreadth doctrine appears to be the handiwork of judges. . . . It seemingly lacks any basis in the text or history of the First Amendment”).

        [92].     See 343 U.S. 495, 505 (1952).

        [93].     425 U.S. 748, 764 (1976).

        [94].     458 U.S. 747, 763–64 (1982).

        [95].     The shift may also have been informed by the growing appeal of originalism as a method of constitutional interpretation. It is notable that the Justices who signed on to Stevens were all the primary proponents of originalism on the Court, although Justice Thomas’s dissent in Stevens makes clear the tension between “the First Amendment traditionalism” espoused by Justice Scalia in R.A.V. and Chief Justice Roberts in Stevens and original-public-understanding originalism of the kind Justice Thomas embraced. See infra Part I.B.

        [96].     See, e.g., Virginia v. Black, 538 U.S. 343, 358–59 (2003) (quoting R.A.V., 505 U.S. 377, 382–83 (1992)).

        [97].     559 U.S. 460 (2010).

        [98].     18 U.S.C. § 48 (2001).

        [99].     Stevens, 559 U.S. at 465–66.

      [100].     See Brief of the Center on the Administration of Criminal Law as Amicus Curiae in Support of Petitioner at 4, Stevens, 559 U.S. 460 (No. 08-769).

      [101].     Brief for the Respondent at 2–5, Stevens, 559 U.S. 460 (2010) (No. 08-769).

      [102].     See Stevens, 559 U.S. at 473.

      [103].     Brief for the United States at 21–22, Stevens, 559 U.S. 460 (No. 08-769) (quoting Young v. American Mini Theatres, Inc., 427 U.S. 60, 61 (1976)); see also id. at 21 (“The graphic depictions of the torture and maiming of animals reached by Section 48 are ‘no essential part of any exposition of ideas.’”) (quoting Chaplinsky v. New Hampshire, 315 U.S. 568, 572 (1942)).

      [104].     See id. at 36–37.

      [105].     For example, one film instructed viewers how to train pit bulls to hunt prey and “show[ed] footage both of well-trained dogs performing properly . . . and poorly trained dogs and hunters making errors.” Brief for the Respondent at 3, Stevens, 559 U.S. 460 (2010) (No. 08-769).

      [106].     Id. at 21.

      [107].     Id. at 12.

      [108].     See United States v. Stevens, 533 F.3d 218, 224–32 (3d Cir. 2008) (en banc).

      [109].     United States v. Stevens, 559 U.S. 460, 470 (2010).

      [110].     Id.

      [111].     Id. at 468–69, 472 (internal quotes omitted).

      [112].     Id. at 469.

      [113].     Id.

      [114].     Stevens, 533 F.3d at 224.

      [115].     Brown v. Ent. Merchs. Ass’n, 564 U.S. 786, 791–93 (2011).

      [116].     See id. at 794–95, 799.

      [117].     Id. at 814 (Alito, J., concurring).

      [118].     Id.

      [119].     Lakier, supra note 38, at 2177 (noting the lack of “historical evidence” to support the claim that these categories were well defined or narrowly limited in the eighteenth or nineteenth centuries); id. at 2204 (noting the intense disagreement on the Court in the twentieth century about how the categories should be defined).

      [120].     The notes Justice Murphy made when writing Chaplinsky were very short and contained very little theoretical or even doctrinal discussion of the problem of fighting words. Murphy noted merely that, although the language in Cantwell about the exclusion of epithets and personal abuse appeared to “control the present case” since Cantwell involved a different kind of statute, “it perhaps is better to note probable jurisdiction and write this one out.” Notes by Justice Murphy, Walter Chaplinsky vs. State of New Hampshire (#255), Folder 5, Box 65, microformed on Roll 124, Frank Murphy Papers (on file with the Bentley Historical Library, University of Michigan). The opinion itself includes almost no reasoned argument for the exclusion other than a vague invocation of historical tradition that is also unsupported by any authority. Lakier, supra note 38, at 2206.

      [121].     Justice Thomas argued in dissent that it didn’t matter, ultimately, whether violent speech was within the scope of the First Amendment because, at the Founding, children did not have First Amendment rights. See Brown, 564 U.S. at 821 (Thomas, J., dissenting). In effect, he challenged the unidimensionality of the Stevens Court’s approach to delimiting First Amendment boundaries.

      [122].     See, e.g., Clay Calvert & Mary-Rose Papandrea, The End of Balancing? Text, History & Tradition in First Amendment Speech Cases after Bruen, 18 Duke J. Const. L. & Pub. Pol’y 59, 69 (2023) (“In short, as Chief Justice Roberts wrote for the eight-Justice Stevens majority, the precluded category of speech must have been ‘historically unprotected’”); Francesca L. Procaccini, Equal Speech Protection, 108 Va. L. Rev. 353, 415–16 (2022) (arguing that after Stevens, “the Court . . . began to realign its methodological approach away from myth and towards a more accurate description of the doctrine”); Claudia E. Haupt & Wendy E. Parmet, Public Health Originalism and the First Amendment, 78 Wash. & Lee L. Rev. 231, 264 (2021) (accepting the framing of Stevens but arguing that the Court’s opinions suggest “the categories are to be determined based on longstanding understandings, not the understanding at the moment of ratification”). This is not to say that the case’s hold on the First Amendment has not been challenged. A vigorous body of commentary has emerged in the wake of the decision that challenges its core assumptions. See infra note 265.

      [123].     Cantwell v. Connecticut, 310 U.S. 296, 309–10 (1940).

      [124].     Gormley v. Dir., Conn. St. Dep’t of Adult Prob., 449 U.S. 1023, 1025 (1980) (White, J., dissenting from denial of cert.) (“[S]tate courts are not in agreement concerning application of First Amendment principles in this area of the law.”).

      [125].     Lea VanderVelde, The Legal Ways of Seduction, 48 Stan. L. Rev. 817, 817, 894–96 (1996) (noting that by the end of the nineteenth century, courts found in a number of cases that “the coercive force of words of economic threat were sufficient to render the sexual predation redressible”).

      [126].     See Reva B. Siegel, Introduction: A Short History of Sexual Harassment, in Directions in Sexual Harassment Law 1–39 (Catharine A. MacKinnon & Reva B. Siegel eds., 2004).

      [127].     Indeed, the problem appears to have been significant. According to a 1966 House Report, the Bell Telephone System (which provided telephone service to some 80 percent of American households) estimated that it received approximately 376,000 complaints a year about abusive or harassing telephone calls and 46,000 complaints just in the month of March. 1966. H.R. Rep. No. 89-1334, at 2 (1966).

      [128].     See, e.g., 47 U.S.C.A. § 223(a)(1)(C)–(E) (West 2013) (making it a crime, punishable for up to two years in jail, to make a telephone call “without disclosing [their] identity and with [the] intent to abuse, threaten, or harass any specific person” or to “make or cause[] the telephone of another repeatedly or continuously to ring, with intent to harass any person at the called number” or to make repeated phone calls “solely to harass any specific person.”); Ark. Code Ann. § 5-71-209 (West 2019); La. Stat. Ann. § 14:285 (2018); Ohio Rev. Code Ann. § 4931.31 (West 2010) (repealed); S.D. Codified Laws § 49-31-31(4) (1983); Andrea J. Robinson, A Remedial Approach to Harassment, 70 Va. L. Rev. 507, 523–24 (1984) (describing these laws).

      [129].     Robinson, supra note 119, at 525.

      [130].     Robert A. Guy, Jr., The Nature and Constitutionality of Stalking Laws, 46 Vand. L. Rev. 991, 991–92 (1993).

      [131].     Id. at 992; Walker, supra note 19, at 280.

      [132].     Tjaden, supra note 20, at 262.

      [133].     18 U.S.C. § 2261A (1996).

      [134].     See infra notes 177 and 178 and accompanying text.

      [135].     See Legislative Attorney, Cong. Rsch. Serv., RL34651, Protection of Children Online: Federal and State Laws Addressing Cyberstalking, Cyberharassment, and Cyberbullying 21–31 (2009), https://www.everycrsreport.com/files/20091019_RL34651_b7f143a4e8a4c744a3b60f5c66268fa796b6711b.pdf [https://perma.cc/QRT7-DNE6].

      [136].     Nat’l Ctr. for Victims of Crime, The Model Stalking Code Revisited: Responding to the New Realities of Stalking 11 (2007), http://www.markwynn.com/stalking/model-talking-code-revisited-2007.pdf [https://perma.cc/2L65-SR3L].

      [137].     Rebecca Schaeffer was stalked for about three years. In other states, reports surfaced around this time of women being stalked for even longer—in one case for eight years—and of others going into hiding to escape their stalkers. Guy, supra note 130, at 996–97.

      [138].     Walker, supra note 19, at 278 (reporting testimony that “as a general rule, [an] arrest [under one of these laws was] perceived by the mentally ill stalker as a confirmation of the relationship, and by the less seriously ill stalker as an angering challenge”).

      [139].     Tamar Lewin, New Laws Address Old Problem: The Terror of a Stalker’s Threats, N.Y. Times (Feb. 8, 1993), https://www.nytimes.com/1993/02/08/us/new-laws-address-old-problem-the-terror-of-a-stalker-s-threats.html [https://perma.cc/3HAM-GADC].

      [140].     See, e.g., Cal. Penal Code § 646.9 (West 1990).

      [141].     See, e.g., id. at § 646.9(f).

      [142].     In many states, stalking remains a misdemeanor. But in some states, it is defined as a low-grade felony, particularly if repeated and/or if it includes the making of a threat. See, e.g., Mass. Gen. Laws Ann. ch. 265, § 43 (West 2014); N.M. Stat. Ann. § 30-3A-3 (West 2009). Penalties range widely but can be significant. Counterman, for example, was sentenced to four-and-a-half years in prison after he was convicted of violating the Colorado stalking law. People v. Counterman, 497 P.3d 1039, 1044 (Colo. App. 2021).

      [143].     Guy, supra note 130, at 994 (“Stalking behavior varies widely in both its intensity and form. Some cases involve threatening and unceasing harassment; other cases involve voluminous mailings of unsolicited love letters coupled with constant following; still other instances consist of merely incessant following and surveillance.”).

      [144].     See e.g., Idaho Code Ann. § 18-7906 (West 2004); see also Mass. Gen. Laws Ann. ch. 265, § 43(a) (West 2014) (“willfully and maliciously”); N.M. Stat. Ann. § 30-3A-3 (West 2009) (“knowingly”).

      [145].     See, e.g., 18 U.S.C.A. § 2261A; Haw. Rev. Stat. § 711-1106.5 (2009); 11 R.I. Gen. Laws § 11-59-2 (West 1956).

      [146].     See, e.g., Colo. Rev. Stat. Ann. § 18-3-602(1)(b) (West); N.H. Rev. Stat. § 633:3-a (2023); N.J. Rev. Stat. 2C:12-10 (2009).

      [147].     The federal stalking law, for example, requires proof that the stalking caused “reasonable fear of the death of, or serious bodily injury to . . . [the target of the stalking]; an immediate family member (as defined in section 115) of that person; a spouse or intimate partner of that person; or the pet, service animal, emotional support animal, or horse of that person.” 18 U.S.C.A. § 2261A(1)(A). In contrast, the New Mexico stalking law requires proof that the stalking was intended to “place the individual in reasonable apprehension of death, bodily harm, sexual assault, confinement or restraint of the individual or another individual.” N.M. Stat. Ann. § 30-3A-3 (West 2009). The federal law also prohibits stalking that causes emotional distress, but under a separate prong of the statute. 18 U.S.C.A. § 2261A(1)(B).

      [148].     See, e.g., N. J. Rev. Stat. § 2C:12-10; Wis. Stat. § 940.32; Wyo. Stat. Ann. § 6-2-506.

      [149].     Tjaden, supra note 20, at 266.

      [150].     See Ashley N. B. Beagle, Modern Stalking Laws: A Survey of State Anti-Stalking Statutes Considering Modern Mediums and Constitutional Challenges Comments, 14 Chap. L. Rev. 457, 460–62 (2011) (describing various types of stalkers).

      [151].     Tjaden & Thoennes, supra note 20, at 2; see also Baum, Catalano & Rand, supra note 19, at 8 (making similar, earlier findings: “Stalkers made one or more threats to 43% of victims.”).

      [152].     Tjaden & Thoennes, supra note 20, at 2.

      [153].     As the National Institute of Justices notes, stalkers can cause their victims to feel great fear even when they do not explicitly or implicitly threaten them. Nat’l Inst. of Just., Project to Develop a Model Anti-Stalking Code for States 45 (1993), https://www.ojp.gov/pdffiles1/Digitization/144477NCJRS.pdf [https://perma.cc/SAN8-ZFDG]; see also Barry Rosenfeld, Violence Risk Factors in Stalking and Obsessional Harassment: A Review and Preliminary Meta-Analysis, 31 Crim. Just. & Behavior 9, 14 (2004) (“An intuitively obvious predictor of violence is the existence of a previous threat. . . . Yet, the correspondence between threats and violence is not as simple as stalking victims might assume. Studies of stalking-related violence have demonstrated high rates of false positives, where threats were made but never carried out. Likewise, smaller, but nevertheless substantial, false-negative rates have also been observed, with many violent stalkers giving no warning of their violent intentions.”).

      [154].     For example, one study found that more than one-third of stalking victims moved residences in an attempt to prevent further stalking, and 11 percent purchased a gun. Tjaden, supra note 20, at 273.

      [155].     Nat’l Inst. of Just., supra note 153, at 45.

      [156].     See, e.g., Nev. Rev. Stat. § 200.575(1), (3) (West 2019); N.H. Rev. Stat. § 633:3-a (2018); Or. Rev. Stat. § 163.732 (1993); W. Va. Code § 53-8-4(a)(3) (2017); W. Va. Code § 61-2-9a(a).

      [157].     See Danielle Keats Citron, Hate Crimes in Cyberspace 5, 12 (2014) (describing the rise of the use of the Internet as a vehicle for stalking).

      [158].     Dep’t of Just., supra note 21, at 1.

      [159].     Id. at 3.

      [160].     Rachel E. Morgan & Jennifer L. Truman, U.S. Dep’t of Just., Bureau of Just. Stats., Stalking Victimization, 2019 6 (Feb. 2022), https://bjs.ojp.gov/content/pub/pdf/sv19.pdf [https://perma.cc/9EMJ-E4XE].

      [161].     For example, in Commonwealth v. Kwiatkowski, 637 N.E.2d 854, 857 (Mass. 1994), the Massachusetts Supreme Court held that the state’s newly enacted stalking law was unconstitutionally vague because it did not make clear either to law enforcement officers or regulated parties whether it required proof that a defendant engaged in multiple courses of stalking conduct, or only one. However, the Court cured the law’s vagueness problem by construing it to require proof of only one course of stalking conduct, and permitted it to, thus construed, go into effect. Id. And in Long v. State, in 1996, the Texas Court of Criminal Appeals found that the state’s stalking law was unconstitutionally vague because it could be used to criminalize the actions of a speaker who made a threat and then decided to engage in the kind of public, political expression that was ordinarily entitled to full First Amendment protection. 931 S.W.2d 285, 293–94 (Tex. Crim. App. 1996). The court expressed concern, for example, that “a political protester who crosses the line and makes a threat might find himself forever barred from engaging in peaceful, legitimate expression for fear of subjecting himself to punishment for stalking.” Id. at 293. Because it feared that the stalking law would intrude upon the realm of protected expression, the court struck it down—just as the Thornhill framework suggested it should.

      [162].     573 F.2d 783, 787 (3d Cir. 1978) (citations omitted).

      [163].     659 So. 2d 235, 237 (Fla. 1995) (quoting Operation Rescue v. Women’s Health Ctr., Inc., 626 So. 2d 664, 675 (Fla. 1993)).

      [164].     Id. at 236.

      [165].     See, e.g., United States v. Darsey, 342 F. Supp. 311, 313–14 (E.D. Pa. 1972); State v. Brown, 266 S.E.2d 64, 65 (S.C. 1980).

      [166].     State v. Lee, 917 P.2d 159, 165–166 (Wash. Ct. App. 1996), aff’d, State v. Lee, 957 P.2d 741 (Wash. 1998).

      [167].     Id. at 166 (citations omitted).

      [168].     Id.

      [169].     For cases in which courts reached this conclusion about anti-harassment laws, see, e.g., Thorne v. Bailey, 846 F.2d 241, 243–44 (4th Cir. 1988); State v. Elder, 382 So.2d 687, 692–93 (Fla. 1980); Gormley v. Dir., Conn. State Dep’t of Prob., 632 F.2d 938, 940 (2d Cir. 1980); People v. Shack, 658 N.E.2d 706, 710–11 (N.Y. 1995).

      [170].     As then-Judge Alito noted in Saxe v. State College Area School District, “[t]here is no categorical ‘harassment exception’ to the First Amendment’s free speech clause.” 240 F.3d 200, 204 (3d Cir. 2001).

      [171].     753 S.W.2d 874, 875 (Ky. Ct. App. 1988).

      [172].     Id. at 875–76 (citations omitted); see also People v. Weeks, 591 P.2d 91, 96 (Colo. 1979) (“The First Amendment does not extend to any person the right to use his power of speech as a battering ram to destroy the tranquility and repose of another person’s home. . . . The gravamen of the offense is the thrusting of an offensive and unwanted communication on one who is unable to ignore it. Thus, defendant’s assertion that his communication to the victim was not ‘obscene’ does not dispose of this case.”) (citations omitted).

      [173].     397 U.S. 728, 730 (1970) (upholding a federal law that permitted a householder to “insulate himself from advertisements that offer for sale ‘matter which the addressee in his sole discretion believes to be erotically arousing or sexually provocative’”) (quoting 39 U.S.C. § 4009(a)).

      [174].     Id. at 736.

      [175].     Id. at 736–37.

      [176].     Id. at 736; see also, e.g., Martin v. City of Struthers, 319 U.S. 141, 148 (1943) (affirming the right of homeowners to refuse people wishing to distribute literature to their home). This is discussed further infra Part III.C.1.

      [177].     Victims of Trafficking and Violence Protection Act of 2000, Pub. L. No. 106–386, 114 Stat. 1464 (2000).

      [178].     Violence Against Women Reauthorization Act, Pub. L. No. 109–162, 119 Stat. 2988 (2006). In 2013 Congress amended the Act again to include acts that not only cause but “attempt[] to cause, or would be reasonably expected to cause substantial emotional distress.” Violence Against Women Reauthorization Act of 2013, Pub. L. No. 113-4, 127 Stat. 54 (2013).

      [179].     The law did not require the stalker to communicate a serious intent to do physical harm. It could in fact be satisfied even by purely innocuous communications that nevertheless reflected the requisite intent. Hence, even this narrow version of the federal stalking law could not be properly characterized as a “true threats” regulation.

      180.     372 F.3d 365 (6th Cir. 2004), overruled on other grounds, 543 U.S. 1182 (2005).

      [181].     Id. at 371–75.

      [182].     Id. at 379–81.

      [183].     Id. at 379.

      [184].     Id. at 378–80.

      [185].     336 U.S. 490, 502 (1949).

      [186].     Eugene Volokh, The “Speech Integral to Criminal Conduct” Exception, 101 Cornell L. Rev. 981, 983 (2016) (noting that “[s]ince 2006, the Supreme Court has been reviving a long-dormant and little-defined First Amendment exception: the exception for ‘speech integral to criminal [or tortious] conduct’” and that “‘[s]peech integral to criminal conduct’ is now a standard item on lists of First Amendment exceptions”).

      [187].     Id. at 986, 988 (noting that the opinions identifying the exception “were not clear in their scope” and that the exception has become “at times, a tool for avoiding serious First Amendment analysis—a way to uphold speech restrictions as supposedly fitting within an established exception, without a real explanation of how the upheld restrictions differ from other restrictions that would be struck down”).

      [188].     United States v. Stevens, 559 U.S. 460, 471 (2010). In fact, as Volokh notes, the idea that the speech was integral to crime was “only one of the rationales the Court gave in Ferber”—and arguably not the most important one. Volokh, supra note 186, at 999. After Stevens, however, it has become “the Court’s principal explanation for the constitutionality of bans on distributing child pornography.” Id.

      [189].     See Amy Adler, The Perverse Law of Child Pornography, 101 Colum. L. Rev. 209, 210 (2001) (“Child pornography law is a remarkably recent invention. Not until 1982 did the Supreme Court consider the distinct problem of child pornography, create it as a special category of constitutional inquiry, and expel it from the protection of the First Amendment.”).

      [190].     701 F.3d 849, 855 (8th Cir. 2012).

      [191].     753 F.3d 939 (9th Cir. 2014).

      [192].     Id. at 941–43.

      [193].     Id. at 944.

      [194].     Id. at 947.

      [195].     Id. (“As described in 18 U.S.C. § 2261A(2)(A), Osinger engaged in a course of conduct ‘with the intent . . . to . . . harass, or intimidate, or cause substantial emotional distress to’ V.B. . . . Any expressive aspects of Osinger’s speech were not protected under the First Amendment because they were ‘integral to criminal conduct’ in intentionally harassing, intimidating or causing substantial emotional distress to V.B.”).

      [196].     See, e.g., Volokh, supra note 186, at 1040 (arguing that under the Osinger logic “any repeated online speech—including public political ridicule of politicians, journalists, businesspeople, religious figures, and others—that intentionally causes substantial emotional distress would be constitutionally unprotected”); James Weinstein, The Federal Cyberstalking Statute, Content Discrimination and the First Amendment, 54 U.C. Davis L. Rev. 2553, 2558 (2021); see also United States v. Sryniawski, 48 F.4th 583, 588 (8th Cir. 2022) (rejecting a similar argument as “circular and unpersuasive” and noting that “Congress may not define speech as a crime, and then render the speech unprotected by the First Amendment merely because it is integral to speech that Congress has criminalized.”); Matter of Welfare of A. J. B., 929 N.W.2d 840, 852 (Minn. 2019) (making a similar argument); Burroughs v. Corey, 92 F. Supp. 3d 1201, 1209 n.16 (M.D. Fla. 2015), aff’d, 647 F. App’x 967 (11th Cir. 2016) (same).

      [197].     907 F.3d 67, 72, 74, 77 (1st Cir. 2018) (“The statute does not, on its face, regulate protected speech, or conduct that is necessarily intertwined with speech or expression.”).

      [198].     George P. Fletcher, The Nature and Function of Criminal Theory, 88 Calif. L. Rev. 687, 701 (2000) (To be criminal “[e]very punishable act requires a union of actus reus and mens rea.”).

      [199].     See, e.g., State v. Brobst, 857 A.2d 1253, 1254, 1256–57 (N.H. 2004) (striking down a state harassment law that made it a crime to make a telephone call “with [the] purpose to annoy or alarm another” because it could be used to criminalize “a call from a neighbor warning of an approaching tornado or a dangerous animal that escaped from the zoo, or a call from a bill collector demanding that payments be made,” notwithstanding the fact that these “are, undoubtedly, legitimate communications”); People v. Klick, 362 N.E.2d 329, 331 (Ill. 1977) (“There are . . . many instances when, without breaching the peace, one may communicate with another with the possible intention of causing a slight annoyance in order to emphasize an idea or opinion, or to prompt a desired course of action that one is legitimately entitled to seek.”).

      [200].     Ackell, 907 F.3d at 76 (quoting Virginia v. Black, 538 U.S. 343, 344 (2003)).

      [201].     Id.

      [202].     748 F.3d 425, 434 (1st Cir. 2014).

      [203].     37 F.4th 70, 80 (3d Cir. 2022) (“To ‘intimidate,’ we hold, a defendant must put the victim in fear of death or bodily injury. And to ‘harass,’ he must distress the victim by threatening, intimidating, or the like. That reading limits intent to harass to ‘criminal harassment, which is unprotected because it constitutes true threats or speech that is integral to proscribable criminal conduct.’ It also limits ‘intent to intimidate’ to what it ‘especially’ means, a form of true threats or speech integral to a crime.”) (citing Ackell, 907 F.3d at 76).

      [204].     United States v. Sryniawski, 48 F.4th 583, 588 (8th Cir. 2022) (internal citations omitted).

      [205].     Id. at 587–88.

      [206].     See supra notes 137–142 and accompanying text.

      [207].     See State v. Lee, 917 P.2d 159, 165–66 (Wash. Ct. App. 1996), aff’d, 957 P.2d 741 (Wash. 1998).

      [208].     825 N.W.2d 303, 305–07 (Wis. Ct. App. 2012) (summarizing Wis. Stat. Ann. § 940.32k).

      [209].     Id. at 309.

      [210].     Dugan v. State, 451 P.3d 731, 738 (Wyo. 2019).

      [211].     Id. at 739.

      [212].     Id. at 737–38.

      [213].     Mashaud v. Boone, 295 A.3d 1139, 1160 (D.C. 2023).

      [214].     State v. Billings, 287 A.3d 146, 168 (Conn. App. Ct. 2022), cert. denied, 288 A.3d 217 (Conn. 2023) (reversing a stalking conviction on the grounds that the conviction relied on constitutionally protected speech conduct as evidence); see also State v. Nowacki, 111 A.3d 911, 928 (Conn. App. Ct. 2015) (holding that if a defendant “was prosecuted on the basis of the content of his communication,” a court must “examine that communication to determine if it falls within the wide aegis of protected speech.”).

      [215].     Com. v. Bigelow, 59 N.E.3d 1105, 1126 (Mass. 2016) (citing Com. v. Welch, 825 N.E.2d 1005, 1005 (Mass. 2005)).

      [216].     Id. at 1122, 1127 (Duffly, J., dissenting) (noting that “[u]nder the court’s analysis today . . . speech will be considered unprotected if the statements, ‘when viewed in context,’ could be found to increase the ‘potential to instill [in an intended target] a fear of future harm,’ because the recipient is unable to determine the speaker’s intent” and that this definition “stretch[es] the meaning of ‘true threat’ far beyond common understanding” and consequently “results essentially in the creation of a broad and amorphous category of unprotected speech”).

      [217].     People v. Relerford, 104 N.E.3d 341, 356 (Ill. 2017).

      [218].     Welfare of A. J. B., 929 N.W.2d 840, 863, 865 (Minn. 2019) (interpreting the state mail harassment statute to only criminalize the “mailing or delivery of a letter with an ‘intent to abuse’” and affirming its constitutionality against an overbreadth challenge on the grounds that “the primary aim” of the defendant’s conduct must be “causing mental, physical or emotional injury or harm” and that while this statute “may still chill speech, it chills much less protected speech”).

      [219].     Id. at 846 (quoting United States v. Stevens, 559 U.S. 460, 470 (2010)).

      [220].     State v. Shackelford, 825 S.E.2d 689, 698–99 (N.C. Ct. App. 2019).

      [221].     State v. Mazur, No. COA17-736, 2018 WL 4440576, at *7 (N.C. Ct. App. 2018); see also Nathan W. Wilson, Frog Eyes and Pig Butts: The North Carolina Stalking Statute’s Constitutional Dilemma and How to Remedy It, 99 N.C. L. Rev. 479, 481–83 (2021) (observing that in these two cases “two panels of the same court looked at the same issue . . . and reached opposite outcomes” based on a different application of First Amendment jurisprudence).

      [222].     575 U.S. 723 (2015).

      [223].     See generally Rothman, supra note 4 (canvassing these arguments).

      [224].     Indeed, several of the amicus briefs submitted to the Court did not even mention the word “stalking.” See, e.g., Brief of Amicus Curiae Foundation for Individual Rights and Expression in Support of Petitioner and Reversal, Counterman v. Colorado, 600 U.S. 66 (2023) (No. 22-138); Brief of Amicus Curiae Electronic Frontier Foundation and Student Press Law Center in Support of Petitioner, Counterman, 600 U.S. 66 (No. 22-138).

      [225].     Adam Liptak, When Do Creepy Facebook Messages Cross a Constitutional Line?, N.Y. Times (Feb. 20, 2023), https://www.nytimes.com/2023/02/20/us/politics/supreme-court-facebook-stalking-colorado.html [https://perma.cc/GJG4-GBED].

      [226].     Matt Ford, The Supreme Court Will Decide the Line Between Free Speech and Violent Threats, New Republic (Feb. 15, 2023), https://newrepublic.com/article/170586/counterman-colorado-stalking-first-amendment [https://perma.cc/FKV4-QJPW].

      [227].     Lawrence Hurley, Supreme Court Takes Up Convicted Stalker’s Plea to Decide What Constitutes a ‘True Threat,’ NBC News (Jan. 13, 2023), https://www.nbcnews.com/politics/supreme-court/supreme-court-takes-convicted-stalkers-plea-decide-constitutes-true-th-rcna59689 [https://perma.cc/67WG-QN6A].

      [228].     Transcript of Oral Argument, Counterman, 660 U.S. 66 (2023) (No. 22-138).

      [229].     This section draws on Brief of First Amendment Scholars as Amici Curiae in Support of Respondent, Counterman, 600 U.S. 66 (No. 22-138).

      [230].     Colo. Rev. Stat. § 18-3-602(1)(b) (West).

      [231].     Colo. Rev. Stat. Ann. § 18-3-602(1)(c) (West).

      [232].     J.A. Volume II at 346, Counterman, 600 U.S. 66 (No. 22-138) (emphasis added).

      [233].     See People v. Counterman, 497 P.3d 1039, 1045–50 (Colo. App. 2021).

      [234].     Counterman, 600 U.S. at 69.

      [235].     See id. at 72–82.

      [236].     Id. at 71 n.1.

      [237].     Id.

      [238].     Id. at 106 (Barrett, J., dissenting).

      [239].     See Virginia v. Black, 538 U.S. 343, 344 (2003); see also Rogers v. United States, 422 U.S. 35, 46–47 (1975) (Marshall, J., concurring).

      [240].     Robert M. Emerson, Kerry O. Ferris & Carol Brooks Gardner, On Being Stalked, 45 Soc. Probs. 289, 300 (1998) (noting that the experience of being followed by a stalker “generate[s] a distinctive quality of stalking—the sense of menace,” and that the “victim [frequently] experience[s] the pursuer’s persistent yet distanced attentions as ‘sinister’ or ‘creepy,’ and begins to anticipate and dread the other’s appearance jogging by their house, purchasing lunch in the restaurant where they regularly eat, or walking down the aisle of the supermarket they patronize”); id. at 303 (noting that communications from stalkers “often create or heighten the pursuer’s sense of being a victim of stalking, generating feelings of being threatened, menaced, or even shamed” by the fact that “another—an ex, a stranger—now knows more about one’s life than many of one’s intimates, and more than the victim wants known, that this information is ill gotten and illegitimate, and that the victim has lost control over what is known by whom”).

      [241].     See TK Logan & Robert Walker, Stalking: A Multidimensional Framework for Assessment and Safety Planning, 18 Trauma, Violence, & Abuse 200, 201 (2017).

      [242].     Id. at 204 (“Life invasion is a particularly common form of stalking and harassment [and] is referred to as invasion because it means entry directly or through electronic means into the private sphere of victim lives—even though it can occur in public settings . . . The level of stalking invasion has been associated with increased victim distress.”).

      [243].     In 1979, for example, the Colorado Supreme Court noted that “[t]he First Amendment does not extend to any person the right to use his power of speech as a battering ram to destroy the tranquility and repose of another person’s home” and accordingly permitted the government to impose sanction on one who “thrust[s] . . . an offensive and unwanted communication on one who is unable to ignore it.” People v. Weeks, 591 P.2d 91, 96 (Colo. 1979).

      [244].     Andrew Koppelman, Revenge Pornography and First Amendment Exceptions, 65 Emory L.J. 661, 684 (2016) (“The central ideal-regarding goal of free speech is a distinctive type of human flourishing, one that cannot occur except in what Mill called ‘an atmosphere of freedom.’”).

      [245].     De Jonge v. Oregon, 299 U.S. 353, 365 (1937) (holding that the “security of the Republic” is enshrined in the rights of individuals to gather collectively to maintain free political discussion).

      [246].     See generally Abner S. Greene, “Not in My Name” Claims of Constitutional Right, 98 B.U. L. Rev. 1475 (2018) (describing this parallel).

      [247].     Courts have had no trouble affirming the constitutionality of stalking laws when applied to non-verbal conduct, in cases where these laws have involved challenges based on the right to travel or some other constitutional provision. See, e.g., Falls v. State, 131 N.E.3d 1288, 1291 (Ind. 2019); Snowden v. State, 677 A.2d 33, 37 (Del. 1996). They have recognized, in general, the legitimacy and constitutionality of these kinds of laws.

      [248].     Counterman v. Colorado, 600 U.S. 66, 70 (2023).

      [249].     Id.

      [250].     Id.

      [251].     Id.

      [252].     Id.

      [253].     Brief of Coles Whalen as Amicus Curiae in Support of Respondent at 1, Counterman, 600 U.S. 66 (No. 22-138).

      [254].     J.A. Volume I at 127, Counterman, 600 U.S. 66 (No. 22-138).

      [255].     Brief of Coles Whalen as Amicus Curiae in Support of Respondent, supra note 253, at 12.

      [256].     Transcript of Oral Argument at 53–54, Counterman, 600 U.S. 66 (No. 22-138) (“I think that might sound solicitous of the person’s development. I mean, if—if we’re talking just about what the statements are, how is that—what tone would you use in saying that that would make it threatening?” . . . I can’t promise I haven’t said that [laughter].”) (statement of C.J. Roberts).

      [257].     See, e.g., Mary Anne Franks, How Stalking Became Free Speech: Counterman v. Colorado and the Supreme Court’s Continuing War on Women, Geo. Wash. L. Rev. On the Docket (July 28, 2023), https://www.gwlr.org/how-stalking-became-free-speech-counterman-v-colorado-and-the-supreme-courts-continuing-war-on-women/ [https://perma.cc/E8K2-SU2Y].

      [258].     Virginia v. Black, 538 U.S. 343, 359 (2003).

      [259].     See, e.g., State v. Calloway, 550 P.3d 77, 86 (Wash. Ct. App. 2024) (interpreting a state harassment law to require proof that the “defendant [had] subjective awareness of the ‘substantial risk that [the] communications would be viewed as threatening violence’”) (quoting Counterman, 660 U.S. at 69) (italics in the original); People v. Merklin, No. 22CA0980, 2024 WL 3812297, at *3 (Colo. App. June 6, 2024) (interpreting the emotional distress stalking law under which Counterman was convicted to require, in a subsequent case, that the defendant “consciously disregarded a substantial risk that his communications would be viewed as threatening violence”) (italics in original); State v. Reeves, 316 A.3d 408, 432 (Del. Super. Ct. 2024) (interpreting Counterman to require proof of subjective intent in all cases in which a defendant is charged under a state stalking law based on the content of his communications but rejecting the idea that Counterman requires proof of recklessness in cases in which the communications cause distress for other reasons).

      [260].     State v. Labbe, 314 A.3d 162, 181 (Me. 2023).

      [261].     As applied to the Maine stalking statute, the mens rea test that the court articulated only required proof that the defendant recklessly disregarded the risk that his words would cause their target serious inconvenience or distress, not that they would be perceived to threaten violence. Id. at 181. This will obviously be a much easier hurdle to clear in many stalking cases.

      [262].     Id. at 179 (“Some stalking prosecutions . . . may rely in whole or in part on words used by a defendant to establish the ‘course of conduct’ and consequent effect upon the victim. It does not follow, however, that the Counterman standard applies to every stalking prosecution in which words are spoken or electronic communication devices are used. Rather, Counterman’s holding is clear: where the State relies on the content of a defendant’s expression as the basis for a stalking charge and to establish harm to the victim, the additional requirement to prove subjective mens rea of recklessness applies.”).

      [263].     Id. at 176, 179 (noting that clarification on “the question of Counterman’s effect on the instant case [is warranted] . . . because the specific provision in Colorado’s stalking statute at issue in Counterman is similar to Maine’s stalking statute” but concluding that defendant’s conviction was constitutional because “the stalking charge [against him] was not predicated on the content of those communications but rather on the act of communicating itself—the repeated, unwelcome contact carried out through electronic devices—even after he was asked to stop”).

      [264].     We describe the impact of Stevens as “complex” because, although the decision announces a historical test for determining the categories of speech that are unprotected, in its aftermath, the Court has continued to engage in relatively free-wheeling balancing analysis to determine the scope and requirements of those categories. Justice Kagan’s opinion for the Court in Counterman—which spends very little time examining how courts in the nineteenth century conceptualized the category of threats, and instead focuses almost entirely on the costs and benefits of different definitions of the category—illuminates this well. See Counterman v. Colorado, 600 U.S. 66, 79–82 (2023).

      [265].     Robert C. Post & Jennifer E. Rothman, The First Amendment and the Right(s) of Publicity, 130 Yale L. J. 86, 133 (2020) (“[I]t has been plain for a long time that the theory [of First Amendment coverage outlined in Stevens] is a poor and inaccurate description of actual First Amendment jurisprudence”); Amanda Shanor, First Amendment Coverage, 93 N.Y.U. L. Rev. 318, 340 (2018) (“[T]he Court’s articulated list of categories [in short or long form] does not account for the far greater range of regulations of what is colloquially understood as speech or expression that have long not been subject to First Amendment challenge, let alone strict review.”); Koppelman, supra note 244, at 661 (“If [Stevens] is the law, then harms will accumulate and the law will be helpless to remedy them.”); Weinstein, supra note 196, at 2567 (“Not only is the All-Inclusive Approach [in Stevens] descriptively inaccurate, it is normatively problematic.”); Schauer, supra note 36, at 1622–64 (2015) (suggesting Stevens does not accurately describe First Amendment doctrine); Lakier, supra note 38, at 2197.

      [266].     Lakier, supra note 38, at 2220–21; Koppelman, supra note 244, at 661.

      [267].     Schauer, supra note 36, at 1627–28.

      [268].     Id. at 1624.

      [269].     Justice Thomas wrote a separate dissent to register his objection to the majority’s reliance on New York Times Co. v. Sullivan but otherwise fully joined in Barrett’s opinion discussed above. Counterman v. Colorado, 600 U.S. 66, 105 (2023) (Thomas, J., dissenting).

      [270].     Counterman, 600 U.S. at 83 (Sotomayor J., concurring) (“From 1791 to the present. . . .”) (quoting United States v. Stevens, 559 U.S. 460, 468 (2010)); see also id. at 73–74 (Kagan, J.) (quoting Stevens); id. at 107 (Barrett, J., dissenting).

      [271].     Id. at 84.

      [272].     Id. at 86 (emphasis added).

      [273].     Id. at 84.

      [274].     Id. at 86.

      [275].     Id. In its opinion in the Labbe case, the Maine Supreme Judicial Court relied upon this passage from Justice Sotomayor’s opinion to support its claim that the Counterman standard should not “appl[y] to every stalking prosecution in which words are spoken or electronic communication devices are used.” State v. Labbe, 314 A.3d 162, 179 (Me. 2024).

      [276].     See supra notes 151–156 and accompanying text.

      [277].     See, e.g., State v. Brown, 85 P.3d 109, 113 (Az. Ct. App. 2004) (finding that, although the Arizona stalking law “criminalizes certain types of ‘communication,’ . . . criminal liability under the statute is based on the ‘manner’ in which certain communication is conveyed and the underlying purpose for the communication,” not the content per se).

      [278].     Colo. Rev. Stat. Ann. § 18-3-602 (West) (emphasis added).

      [279].     Brief Amicus Curiae of the Criminal Justice Legal Foundation in Support of Respondent at 24, Counterman v. Colorado, 600 U.S. 66 (No. 22-138).

      [280].     See Eugene Volokh, One-to-One Speech vs. One-to-Many Speech, Criminal Harassment Laws, and “Cyberstalking,” 107 Nw. U. L. Rev. 731, 769 (2013) (making this point).

      [281].     J.A. Volume I at 131, Counterman, 600 U.S. 66 (No. 22-138).

      [282].     Colo. Rev. Stat. Ann. § 18-3-602 (West).

      [283].     Even in cases where the defendant is prosecuted for (for example) breathing into the phone when the victim picks it up—as was the case in State v. Labbe, 314 A.3d 162, 168 (Me. 2024)—it is not strictly true that the prosecution of those acts has nothing to do with the content of the defendant’s communications. Given the sequence of phone calls in which it occurred, the heavy breathing phone call almost certainly communicated a powerful message—that the defendant wanted to be present in the victim’s life—that would not have been communicated by another phone call (a wrong number, for example) that similarly contained no words. The Maine Supreme Judicial Court was, for this reason, perhaps too quick to conclude that stalking prosecutions involving communications of this sort involved only the regulation of conduct, not speech, even if it obviously was correct that the prosecution did not reflect governmental hostility or favor towards any particular viewpoint or message. Id. at 80.

      [284].     See Brief Amicus Curiae of the Criminal Justice Legal Foundation in Support of Respondent at 24, Counterman, 600 U.S. 66 (No. 22-138) (making this argument).

      [285].     First Nat. Bank of Bos. v. Bellotti, 435 U.S. 765, 785–86 (1978).

      [286].     See, e.g., Madsen v. Women’s Health Ctr., Inc., 512 U.S. 753, 794 (1994) (Scalia J., concurring in part) (“The vice of content-based legislation—what renders it deserving of the high standard of strict scrutiny—is not that it is always used for invidious, thought-control purposes, but that it lends itself to use for those purposes.”) (emphasis in original).

      [287].     Counterman, 660 U.S. at 86 (Sotomayor J., concurring in part) (citing Rowan v. Post Off. Dept., 397 U.S. 728 (1970)).

      [288].     397 U.S. at 737 (emphasis added).

      [289].     Cantwell v. Connecticut, 310 U.S. 296, 310 (1940) (emphasis added).

      [290].     Martin v. City of Struthers, 319 U.S. 141, 148 (1943).

      [291].     381 U.S. 301, 310 (1965) (Brennan, J., concurring).

      [292].     Hill v. Colorado, 530 U.S. 703, 716 (2000).

      [293].     N.Y. Times Co. v. Sullivan, 376 U.S. 254, 270 (1964).

      [294].     Volokh, supra note 280, at 743.

      [295].     Id. at 744.

      [296].     Some members of the Court have argued that greater attention should be paid to the importance of this right when it comes to defining the familiar categories of low-value speech. Justice Brennan, for example, argued that the best reason for denying constitutional protection to obscene speech was to protect “juveniles and unconsenting adults from exposure to sexually oriented materials” and that the constitutional definition of unprotected obscenity should be formulated accordingly. Paris Adult Theatre I v. Slaton, 413 U.S. 49, 106–07 (1973) (Brennan J., dissenting) (quoting a commentator who argued that “exposure to (erotic material) is for some persons an intense emotional experience [and therefore a] communication of this nature, imposed upon a person contrary to his wishes, has all the characteristics of a physical assault”). This view, however, never managed to convince a majority of the Court, and the regulation of obscenity, like the regulation of the other low-value categories, remains today entirely insensitive to the question of whether the audience to the obscenity consents. Id. at 57. Notably, Justice Brennan’s view of the assaultive nature of unwanted obscenity is, however, reflected in the law of telephone harassment. Many such laws make it a crime not only to call someone repeatedly to harass them, but also to use the telephone to communicate obscene or salacious messages to an un-consenting recipient. Robinson, supra note 128, at 523–24.

      [297].     Counterman v. Colorado, 600 U.S. 66, 86 (2023) (Sotomayor, J., dissenting) (citations omitted).

      [298].     575 U.S. 723 (2015).

      [299].     18 U.S.C.A. § 875(c) (West).

      [300].     See Perez v. Florida., 580 U.S. 1187, 1187 (2017) (Sotomayor J., concurring) (noting the possibility that a similar state law could be used to put someone in jail for making threats “irrespective of whether his words represented a joke, the ramblings of an intoxicated individual, or a credible threat”); see also Lyrissa Barnett Lidsky & Linda Riedemann Norbut, #I🔫U: Considering the Context of Online Threats, 106 Calif. L. Rev. 1885, 1886–87 (2018).

      [301].     See Watts v. United States, 394 U.S. 705, 709–10 (1969) (Douglas, J., concurring) (describing historical examples).

      [302].     See Elonis, 575 U.S. at 736 (reading a recklessness mens rea requirement into § 875(c) in order to “separate wrongful conduct from ‘otherwise innocent conduct’” (quoting Carter v. United States, 530 U.S. 255, 269 (2000))); Counterman, 600 U.S. at 77–78 (holding that a “subjective element” is required in true threats prosecutions because without it “[t]he speaker’s fear of mistaking whether a statement is a threat; his fear of the legal system getting that judgment wrong; his fear, in any event, of incurring legal costs—all those may lead him to swallow words that are in fact not true threats”).

      [303].     See Colo. Rev. Stat. Ann. § 18-3-602 (West).

      [304].     See infra Part II.A.

      [305].     Weinstein, supra note 196, at 2558 (defining this as the view that “except for a few, traditional exceptions, government is constitutionally forbidden to punish speech based on its content”).

      [306].     While this article focuses primarily on criminal stalking laws, the constitutional law of stalking inevitably will impact the rules that govern civil protective orders and other civil restraints on speech. Indeed, Justice Barrett, in her Counterman dissent, assumed that the First Amendment rules must be the same for both contexts. 600 U.S. at 119 (Barrett, J., dissenting). This assumption appears to be correct: the First Amendment rules that apply to different kinds of speech do not ordinarily depend upon the civil or criminal nature of their regulation. See, e.g., Garrison v. Louisiana, 379 U.S. 64, 67 (1964) (“Where criticism of public officials is concerned, we see no merit in the argument that criminal libel statutes serve interests distinct from those secured by civil libel laws, and therefore should not be subject to the same limitations.”). It is beyond the scope of this Article to comment on whether this is normatively correct, but it does appear to be descriptively accurate.

      [307].     Chaplinsky v. New Hampshire, 315 U.S. 568, 572 (1942).

      [308].     See supra notes 161–69 and accompanying text.

      [309].     See, e.g., Terminiello v. City of Chicago, 337 U.S. 1, 4 (1949) (recognizing that because “a function of free speech under our system of government is to invite dispute . . . [i]t may . . . best serve its high purpose when it induces a condition of unrest, creates dissatisfaction with conditions as they are, or even stirs people to anger” and “[t]hat is why freedom of speech, though not absolute, is nevertheless protected against censorship or punishment, unless shown likely to produce a clear and present danger of a serious substantive evil that rises far above public inconvenience, annoyance, or unrest” (citation omitted)).

      [310].     Cohen v. California, 403 U.S. 15, 21 (1971) (“Any broader view of this authority would effectively empower a majority to silence dissidents simply as a matter of personal predilections.”).

      [311].     Thornhill v. Alabama, 310 U.S. 88, 102 (1940).

      [312].     Id.

      [313].     Counterman v. Colorado, 600 U.S. 66, 70 (2023).

      [314].     Chaplinsky v. New Hampshire, 315 U.S. 568, 572 (1942).

       [315].       Colo. Rev. Stat. Ann. § 18-3-602(1)(c) (West).

      [316].     Wis. Stat. Ann. § 29.083(2)(a)(7) (West 2016).

      [317].     Brown v. Kemp, 86 F.4th 745, 788 (7th Cir. 2023) (Kirsch, J., dissenting) (noting that the statutory language was “lifted directly from Wisconsin’s stalking statute”) (discussed in Part II).

      [318].     Id. at 780.

      [319].     Brief and Short Appendix of Plaintiffs-Appellants Joseph Brown, Louis Weisberg and Stephanie Losse at 7, Brown, 86 F.4th 745 (No. 21-1042).

      [320].     See United States v. Stevens, 559 U.S. 460, 481–82 (2010).

      [321].     Brown, 86 F.4th at 777.

      [322].     See City of San Diego v. Roe, 543 U.S. 77, 83 (2004) (per curiam) (acknowledging that “the boundaries of the public concern test are not well defined”); see also Volokh, supra note 280, at 785 (noting “the vagueness of the public concern/private concern line, which the Court has never defined clearly, and which has often been applied in surprising ways” and providing examples).

      [323].     See Snyder v. Phelps, 562 U.S. 443, 455 (2011) (concluding that speech touched on matters of public concern because “[t]here was no pre-existing relationship or conflict . . . that might suggest Westboro’s speech on public matters was intended to mask an attack on [the individual referenced by the speech] over a private matter”); Connick v. Myers, 461 U.S. 138, 153 (1983) (concluding that speech touched on a matter of private concern because it emerged in the context of a personal disagreement between the speaker and his supervisor); see also Cantwell v. Connecticut, 310 U.S. 296, 309 (1940) (distinguishing fighting words from protected but uncivil discourse by the fact that fighting words are addressed to the specific “person of the hearer”).

      [324].     See Michael Warner, Publics and Counterpublics, 14 Pub. Culture 49, 55–56 (2002) (defining the public as quintessentially a space organized around addresses to strangers).

      [325].     This explains why, in a number of the stalking cases cited in Part II, courts concluded that the prosecutions were constitutional not only because the stalking fell into a familiar low-value category like threats, but also because it involved only speech that touched on matters of private concern. In United States v. Petrovic, for example, the Eighth Circuit noted that because “the intimately private facts and photographs” the defendant revealed as part of his stalking conduct “were never in the public domain” and were not something the public had any “legitimate interest” in seeing, his conviction was for that reason also consistent with the First Amendment. 701 F.3d 849, 856 (8th Cir. 2012); see also United States v. Osinger, 753 F.3d 939, 948 (9th Cir. 2014) (“Osinger’s speech is not afforded First Amendment protection for the additional reason that it involved sexually explicit publications concerning a private individual.”); People v. Austin, 155 N.E.3d 439, 459 (Ill. 2019) (“Deciding whether speech is of public or private concern requires an examination of the content, form, and context of that speech, as revealed by the entire record. . . . Applying these principles to the instant case, we have no difficulty in concluding that the nonconsensual dissemination of the victim’s private sexual images was not an issue of public concern.”).

      [326].     632 F.2d 938, 944 (2d Cir. 1980) (Mansfield, J., concurring); see also United States v. Darsey, 342 F. Supp. 311, 313–14 (E.D. Pa. 1972); Mashaud v. Boone, 295 A.3d 1139, 1144 (D.C. 2023).

      [327].     Jack M. Balkin, Commentary, Digital Speech and Democratic Culture: A Theory of Freedom of Expression for the Information Society, 79 N.Y.U. L. Rev. 1, 3–4 (2004) (defining democratic culture as “a culture in which individuals have a fair opportunity to participate in the forms of meaning making that constitute them as individuals”).

      [328].     See, e.g., Ala. Code § 13A-6-90.1(a) (2012) (defining stalking in the second degree to mean “acting with an improper purpose” and “intentionally and repeatedly follow[ing], harass[ing] telephones, or initiat[ing] communication . . . with another person, any member of the other person’s immediate family, or any third party with whom the other person is acquainted” when doing so “causes material harm to the mental or emotional health of the other person” and “the perpetrator was previously informed to cease”); Cal. Penal Code § 646.9(a), (e) (West 2008) (defining stalking to include “maliciously harass[ing] another person” and defining harass to mean “knowing and willful course of conduct directed at a specific person that seriously alarms, annoys, torments, or terrorizes the person, and that serves no legitimate purpose”); Fla. Stat. Ann. § 784.048(1)(d)(2) (West 2021) (defining the crime of cyberstalking to mean “access[ing], or attempt[ing] to access, the online accounts . . . of another person without that person’s permission, causing substantial emotional distress to that person and serving no legitimate purpose”); Haw. Rev. Stat. Ann. § 711-1106.5 (West 2009) (defining the crime of harassment by stalking to mean “with intent to harass, annoy, or alarm another person, or in reckless disregard of the risk thereof . . . engag[ing] in a course of conduct involving pursuit, surveillance, or nonconsensual contact upon the other person on more than one occasion without legitimate purpose”).

      [329].     Counterman v. Colorado, 600 U.S. 66, 84, 86 n.2 (2023) (Sotomayor, J., concurring in part) (“[P]rosecuting threatening statements made as part of a course of stalking does not squarely present the hardest questions about the mens rea required to prosecute isolated utterances based solely on their content. . . . [And] stalking prosecutions that do not rely on the content of communications would raise even fewer First Amendment concerns.”).

      [330].     Brief for the Petitioner at 6, Counterman, 600 U.S. 66 (No. 22-138) (noting as much).

      [331].     United States v. Sryniawski, 48 F.4th 583 (8th Cir. 2022).

      [332].     Id. at 585.

      [333].     Id. at 585–86.

      [334].     See, e.g., Gertz v. Robert Welch, Inc., 418 U.S. 323, 348 (1974) (stating that the “accommodation of the competing values at stake in defamation suits” requires a different mens rea standard for defamation of private individuals than the standard of actual malice announced in New York Times Co. v. Sullivan and its progeny for public figures); Dun & Bradstreet, Inc. v. Greenmoss Builders, Inc., 472 U.S. 749, 761 (1985) (holding that a lower mens rea standard could be required in suits involving speech on matters of private concern because of the “reduced constitutional value of speech involving no matters of public concern”).

      [335].     Volokh, supra note 280, at 742.

      [336].     Id. at 751 (arguing that “one-to-many speech has full First Amendment value because it involves the expression of facts and opinions aimed at informing and persuading potentially willing listeners” and therefore “should . . . generally be constitutionally protected, notwithstanding the offense and distress it causes to its subjects”).

      [337].     See, e.g., Tjaden & Thoennes, supra note 20, at 8 (noting that life invasion can and does occur in “public setting”); United States v. Ackell, 907 F.3d 67, 71 (1st Cir. 2018) (describing how a defendant threatened to send sexually explicit photos of a young woman he described as his “caged butterfly” to her friends, classmates, and family if she did not continue to do what he told her, including by taking additional sexually explicit photographs of herself); United States v. Osinger, 753 F.3d 939, 942 (9th Cir. 2014) (describing how a defendant sent emails to his victim’s ex-boyfriend and coworker showing sexually explicit photographs of the victim and created a Facebook page that contained nude pictures of the victim, which defendant invited her friends and coworkers to view); United States v. Petrovic, 701 F.3d 849, 853 (8th Cir. 2012) (describing how a defendant “mailed dozens of homemade postcards to addresses throughout [the victim’s] community, including [her] workplace, [her] family members, [her children’s] home, and local businesses like the neighborhood drugstore . . . [that] portrayed a picture of [the victim] scantily clad . . . along with abusive language (for example, ‘I am just a whore 4 sale’) and directions to a website” on which additional intimate images and information could be found).

      [338].     Petrovic, 701 F.3d at 853.

      [339].     Volokh, supra note 280, at 761.

      [340].     Citron, supra note 5, at 21 (“Studies show that cyber stalking victims close their social media accounts; some quit the internet totally. They stop using their phones. Victims change how they express themselves; they are less controversial, more muted, and connect with fewer people.” (cleaned up)). This was certainly true in the Counterman case. As Citron points out, Counterman’s “threatening texts made Whalen so afraid that the defendant would confront and hurt her that she canceled concerts and stopped meeting fans.” Id. As this example makes clear, in stalking cases, courts have to weigh both the free speech interests of the stalker and the victim to fully understand the impact of these laws on First Amendment values.

      [341].     For example, the victim in Ackell told her stalker, at one point, that the coercive control he exercised over her—by, among other things, threatening to disseminate naked pictures of her to others—had left her feeling not only suicidal but “really scared, beyond scared.” Brief of Appellee, United States v. Ackell, 907 F.3d 67 (1st Cir. 2018) (No. 17-1784). “I’m terrified,” she told her victimizer. In response he texted “I do know that.” Id. Similar descriptions of terror, as well as profound emotional distress, pervade the stalking cases.

      [342].     R.A.V. v. City of St. Paul, 505 U.S. 377, 388 (1992).

      [343].     Kenneth L. Karst, Threats and Meanings: How the Facts Govern First Amendment Doctrine, 58 Stan. L. Rev. 1337, 1344–45 (2006) (concluding that “[i]n the language of today’s constitutional law, prevention of so serious a deformation of one’s sense of self must be an interest compelling enough to justify the state’s punishing of death threats”).

      [344].     See Koppelman, supra note 244, at 685 (“A person in fear is a poor participant in democratic deliberation.”).

      [345].     Tjaden, supra note 20, at 266.

      [346].     Logan & Walker, supra note 241, at 210.

      [347].     See infra note 138 and accompanying text.

      [348].     Nonconsensual Distribution of Intimate Images, Cyber C.R. Initiative, https://cybercivilrights.org/nonconsensual-distribution-of-intimate-images/ [https://perma.cc/65MV-P3P3] (noting that forty-eight states, Washington, D.C. and two territories have enacted laws of this sort).

      [349].     Danielle Keats Citron & Mary Anne Franks, Criminalizing Revenge Porn, 49 Wake Forest L. Rev. 345, 346 (2014) (noting that these laws punish images “obtained without consent . . . usually within the context of a private or confidential relationship”).

      [350].     As the Illinois Supreme Court has explained, these laws penalize only speech that depicts certain kinds of images (typically nude or sexual images of the human body) and the gravamen of the offense is not the image on its own but the fact that it is disseminated without the consent of its subject. People v. Austin, 155 N.E.3d 439, 452 (Ill. 2019) (noting that “it is the absence of consent to the image’s distribution that renders the perpetrator in violation of the law” (quoting Ava Schein, Note, When Sharing Is Not Caring: Creating an Effective Criminal Framework Free From Specific Intent Provisions to Better Achieve Justice for Victims of Revenge Pornography, 40 Cardozo L. Rev. 1953, 1955–56 (2019))). This is obviously not a content-based wrong.

      [351].     See Austin, 155 N.E.3d at 462 (finding the government interest in these laws to be “substantial”); State v. VanBuren, 214 A.3d 791, 808 (Vt. 2019) (recognizing the government interest to be “compelling”); Koppelman, supra note 244, at 662, 686 (arguing that “none of [the Stevens] exceptions” are applicable to such laws because they are a response to the social problems created by “entirely new technologies: digital photography and the Internet,” but that these laws nevertheless further important social interests by prevent the silencing of women and preventing the “permanently life-altering effect” that the dissemination of NCII can have).

      [352].     See Austin, 155 N.E.3d at 455; State v. Katz, 179 N.E.3d 431, 453 (Ind. 2022); State v. Casillas, 952 N.W.2d 629, 646 (Minn. 2020); Ex parte Jones, No. PD-0552-18, 2021 WL 2126172, at *6 (Tex. Crim. App. May 26, 2021); VanBuren, 214 A.3d at 814.

      [353].     Austin, 155 N.E.3d at 457.

      [354].     Id. (“The manner of the image’s acquisition and publication, and not its content, is thus crucial to the illegality of its dissemination.”)

      [355].     576 U.S. 155, 163–64 (2015).

      [356].     For example, the Indiana and Vermont Supreme Courts have held that, although NCII laws are content based and therefore trigger strict scrutiny, the strictness of this scrutiny can be relaxed because the speech that they regulate touches only on matters of private concern, not public concern. See Katz, 179 N.E.3d at 456 (“Because the speech here is of less First Amendment importance, it is easier to find the reasons for regulating it compelling.”); VanBuren, 214 A.3d at 808 (“We conclude that the State interest underlying [the NCII law] is compelling. We base this conclusion on the U.S. Supreme Court’s recognition of the relatively low constitutional significance of speech relating to purely private matters. . . .”). This approach to the strict scrutiny analysis is obviously in considerable tension with Stevens.

      [357].     Brown v. Ent. Merchs. Ass’n, 564 U.S. 786, 790 (2011) (cleaned up).

      [358].     Koppelman, supra note 244, at 661.

      [359].     United States v. Stevens, 559 U.S. 460, 470 (2010).

      [360].     See generally Lakier, supra note 38 (critiquing the Court’s “history and tradition” analysis of the First Amendment).

      [361].     Counterman v. Colorado, 600 U.S. 66, 89 (2023) (Sotomayor, J., concurring in part).

      [362].     Morgan & Truman, supra note 160, at 1.