Free Speech and the Confluence of National Security and Internet Exceptionalism
Free Speech and the Confluence of National Security and Internet Exceptionalism
Alan K . Chen 0 1
0 Thi s Symposium is brought to you for free and open access by FLASH: The F ordham Law Archive of Scholarship and History. It has been accepted for inclusion in Fordham Law Review by an authorized editor of FLASH: The F ordham Law Archive of Scholarship and History. For more information , please contact
1 University of Denver Sturm College of Law
Recommended Citation Alan K. Chen, Free Speech and the Confluence of National Security and Internet Exceptionalism, 86 Fordham L. Rev. 379 (2017). Available at: http://ir.lawnet.fordham.edu/flr/vol86/iss2/2
After years of wrangling with the issue, the Court in Brandenburg v. Ohio4
established a strongly speech-protective orientation that prohibits the
government from regulating “advocacy of the use of force or of law violation
except where such advocacy is directed to inciting or producing imminent
lawless action and is likely to incite or produce such action.”5 In recent years,
commentators have questioned whether, in light of contemporary events, the
Court should reconsider Brandenburg in favor of a more lenient standard that
would permit the government to regulate more speech.6 As it happens, the
* Professor of Law, University of Denver Sturm College of Law. This Article was prepared
for the Fordham Law Review symposium entitled Terrorist Incitement on the Internet held at
Fordham University School of Law. For an overview of the symposium, see Alexander Tsesis,
Foreword: Terrorist Incitement on the Internet, 86 FORDHAM L. REV. 367 (2017). Thanks to
Alexander Tsesis for organizing this symposium and inviting me to participate and to all
symposium participants, whose questions and comments helped me to think about these very
difficult issues. I am grateful to my research assistants, Justin Martin and Sarah Spears, for
their help with this Article. Any errors are mine.
two primary justifications for a proposed modification arose at roughly the
same time, around the beginning of the twenty-first century. First, there have
been calls to relax the Brandenburg standard to accommodate the rise of
global and domestic terrorism, particularly since the terrorist attacks of
September 11, 2001.7 Here, the idea is that the increasing (and, it is argued,
increasingly successful) efforts of terrorist organizations to recruit and incite
people to commit unlawful acts suggest the need for greater leeway for
government intervention to prevent death, injuries, and property destruction.8
Second, scholars have called for a relaxed Brandenburg standard in light of
the rapid expansion of digital communication technologies made possible
through social media platforms.9 This claim suggests that there has been
substantial growth in opportunities for terrorist incitement because of the
wide availability of internet communication.
In this Article, I argue that, notwithstanding these contemporary
developments, the Court got it mostly right in Brandenburg. Or, I want to at
least suggest that it is premature to reconstruct the Brandenburg test to
address perceived changes in our global environment. For the most part,
Brandenburg has succeeded in mediating the balance between protecting
political or ideological advocacy and enabling the government to regulate
actual incitement, even in the contemporary era. Moreover, I argue that
society should be especially wary of calls to narrow Brandenburg’s
speechprotective standard because such changes might be significantly influenced
by the confluence of two forms of exceptionalism—national security
exceptionalism and internet exceptionalism—both of which are continuing
to evolve in real time.
In development of this argument, this Article contains three parts. Part I
discusses how the law of incitement is situated in the evolution of modern
free speech doctrine. Next, Part II identifies and explains how national
security exceptionalism and internet exceptionalism may work together to
influence the relaxation of the Brandenburg test. Finally, Part III argues that
there is insufficient evidence at this point to suggest a strong need to
recalibrate the Brandenburg test.
_present_an_unprecedented_danger.html [https://perma.cc/YPC4-3AW3]. For arguments in
support of retaining the Brandenburg test notwithstanding contemporary developments, see
generally GEOFFREY R. STONE, PERILOUS TIMES: FREE SPEECH IN WARTIME: FROM THE
SEDITION ACT OF 1798 TO THE WAR ON TERRORISM (2004); Judge Lynn Adelman & Jon
Deitrich, Extremist Speech and the Internet: The Continuing Importance of Brandenburg, 4
HARV. L. & POL’Y REV. 361 (2010); Thomas Healy, Brandenburg in a Time of Terror, 84
NOTRE DAME L. REV. 6
7. See infra notes 52–80 and accompanying text.
8. Anthony Faiola and Souad Mekhennet, From Hip-Hop to Jihad, How the Islamic State
Became a Magnet for Converts, WASH. POST (May
9. See infra notes 82–108 and accompanying text.
The modern understanding of the free speech doctrine is only about 100
years old.10 During the past century, several understandings about the limits
of government power to ban or regulate speech have become canonical. The
Supreme Court has typically followed the so-called two-level theory of
speech.11 Under this model, speech that has high value in terms of advancing
the goals of the Free Speech Clause (i.e., advancing democratic
selfgovernance, facilitating the search for “truth,” and promoting individual
autonomy)12 is considered to have the highest level of constitutional
protection from government regulation. It is presumed that in adopting and
enforcing legal regulations of high-value speech, the government may not
discriminate against speakers based on their viewpoints or against speech
because of its content.13 Such regulations are subject to the strictest form of
But, in Chaplinsky v. New Hampshire,15 the Court began carving out a
distinct category of no- or low-value speech. There, the Court defined
categories of expression that are not “covered” by the First Amendment—
that is, their regulation is not even subject to scrutiny under the Free Speech
Clause.16 This means that the government may not only regulate but may
potentially even prohibit speech, such as obscenity17 and fighting words,18
because it is outside the scope and concerns of the First Amendment. The
rationale typically provided for the Court’s exclusion of these forms of
communication from the First Amendment’s coverage is that they have little
or no social value. But just as important is that these types of speech are also
presumptively considered to be harmful.19 Recently, and somewhat
controversially, the Court has suggested that this list of no- or low-value
speech is based on historical recognition rather than categorical balancing.20
Under this two-level theory, First Amendment doctrine has never quite
known what to do about government regulation of speech that advocates for
others to violate the law. The evolution of the relevant doctrine has been
recounted numerous times, so a brief summary should suffice here. Unlawful
advocacy falls presumptively into neither the high- nor no-value categories.
In part, this is because pure advocacy, even of such extreme ideas as the
violent overthrow of government, can be a form of core political expression.
But it also underscores the very real concern that when advocacy is strongly
connected to unlawful conduct, its regulation might be necessary to avoid the
Constitution becoming, as Justice Robert Jackson once wrote, “a suicide
In light of this tension, the Supreme Court first struggled with defining the
boundary between protected advocacy and dangerous incitement in a series
of early twentieth-century cases concerning critics of U.S. involvement in
World War I and sympathizers with the Soviet revolution. These cases—
involving prosecutions under the Espionage Act of 1917 and the Sedition Act
of 1918—ultimately led the Court to adopt a First Amendment test that
permitted government regulation of advocacy of unlawful conduct when
“words used are used in such circumstances and are of such a nature as to
create a clear and present danger that they will bring about the substantive
evils that Congress has a right to prevent.”22 Although the “clear and present
danger” test sounds like a fairly robust standard, the Court upheld numerous
prosecutions under this analysis. Notably, in Schenck v. United States,23 the
Court found it irrelevant that the defendant’s advocacy for others to avoid the
military draft was unlikely to influence any audience member because his
intent for such an effect was sufficient to justify his conviction.24
During the period of peace between World War I and World War II, the
Court employed an even more deferential test that permitted government
regulation of advocacy of law violation where the regulations were
“reasonable.”25 Some form of the clear and present danger test reemerged
during the McCarthy era, as federal prosecutors pursued convictions under
the Smith Act, which criminalized advocacy or even teaching about the
“duty, necessity, desirability, or propriety of overthrowing or destroying any
21. Terminiello v. City of Chicago, 337 U.S. 1, 37 (1949) (Jackson, J., dissenting).
Interestingly, unlawful advocacy cases have never drawn an express distinction among
different types of law violations. Thus, at least in theory, the Brandenburg test would be
applied both in a case prosecuting an individual for advocating that another person commit a
minor infraction, such as littering, and a case when a person advocated that another pursue the
violent overthrow of the government. See Margot E. Kaminski, Incitement to Riot in the Age
of Flash Mobs, 81 U. CIN. L. REV. 1, 44 (2013) (describing this as Brandenburg’s “most
significant ambiguity”); see also Goedert v. City of Ferndale, 596 F. Supp. 2d 1027, 1032
(E.D. Mich. 2008) (noting that it is unclear whether Brandenburg includes incitement to
nonviolent lawbreaking and holding that the city could not invoke the incitement doctrine to
justify an ordinance banning signs encouraging motorists to honk their car horns in support of
22. Schenck v. United States, 249 U.S. 47, 52 (1919).
23. 249 U.S. 47 (1919).
24. Id. at 52.
25. Gitlow v. New York, 268 U.S. 652, 668–69 (1925).
government in the United States by force or violence.”26 In Dennis v. United
States,27 the Court rejected facial and as-applied challenges to the Smith Act
because such speech created a clear and present danger.28 In response to the
claim that the danger of violent revolution resulting from the defendants’
speech was remote, the Court noted that because the seriousness of the harm
sought to be prevented was so high, the low probability of success did not
render the speech protected.29 Dennis substantially diluted First Amendment
protection for unlawful advocacy.30 As Erwin Chemerinsky has observed,
“the approach taken by the plurality in Dennis makes probability and
imminence—two seeming requirements of a clear and present danger test—
Both the World War I-era and McCarthy-era cases were decided during
wartime or during a time when many public officials engaged in alarmist
posturing about substantial dangers to national security. It is therefore not
surprising that the Court’s next foray into the incitement doctrine did not
come until 1969. Since, at the time, the United States was heavily involved
in military action in Southeast Asia, and the country’s deep divisions over
that policy were starting to emerge, the perceived concern about undermining
national security on American soil was not as clear.32 Brandenburg
concerned a prosecution under a criminal syndicalism statute dating back to
the post-World War I era, but the relevant expression involved racial epithets
rather than workers’ rights or antiwar rhetoric. The case came to the attention
of prosecutors when a local television station broadcast a report from a rally
by a local Ku Klux Klan faction.33 Clarence Brandenburg, a leader of that
group, was shown on film giving a speech calling for “revengeance [sic]”
against the government for suppressing the white race and making derogatory
remarks about African Americans and Jews.34 Although the film also
captured images of some weapons, Brandenburg’s speech did not overtly call
for any type of violence, and he was never shown carrying a weapon.35
Importantly, there were only about twelve Klan members present, along with
the television reporter and camera operator for the story, who were the only
The Court overturned Brandenburg’s conviction, overruled Whitney v.
California,37 and held that the First Amendment prohibits the government
from punishing “advocacy of the use of force or of law violation except where
such advocacy is directed to inciting or producing imminent lawless action
26. 18 U.S.C. § 2385 (2012).
27. 341 U.S. 494 (1951).
28. Id. at 516–17 (plurality opinion).
29. Id. at 510.
30. CHEMERINSKY, supra note 1, at 1046.
32. STONE, supra note 6, at 543.
33. Brandenberg v. United States, 395 U.S. 444, 445 (1969) (per curiam).
34. Id. at 445–47.
37. 274 U.S. 357 (1927).
and is likely to incite or produce such action.”38 This Brandenburg test is
most commonly understood to impose three prerequisites on the state.39
First, it must show that the speaker intended to (“directed to”) incite
imminent lawless action.40 Second, the speech must be likely to be successful
in its provocation of unlawful behavior in the specific context.41 Finally, the
law violation must be likely, under the circumstances, to be imminent.42
Subsequent cases have wholeheartedly embraced the Brandenburg
Unlike the categories of no-value speech, such as obscenity and fighting
words, at no point has the Court ever categorically excluded unlawful
advocacy from the First Amendment’s coverage. Nor, for what I think most
would agree are obvious reasons, has the Court treated such advocacy as pure
speech, both covered and fully protected by the First Amendment’s
guarantees. Rather, the Brandenburg test reflects a preference for a regime
of ad hoc, case-by-case (rather than categorical) balancing regarding laws
targeting terrorist expression.
As in all areas of law, in an ideal world, the relevant First Amendment
doctrine would strike a perfect balance to achieve optimal deterrence. That
is, the Constitution would permit laws to regulate speech that is truly
dangerous (that either has caused, or is very likely to cause, tangible harm),
but would forbid any law prohibiting or even deterring pure political
expression that has utility in promoting democracy, the search for truth, or
individual autonomy. A First Amendment standard that is too capacious
could permit speech that causes devastating harm. A free speech doctrine
that is too narrow will prohibit or chill much expression, and the marketplace
of ideas will be less robust.
As numerous scholars have observed, Brandenburg adopted a highly
speech-protective test. Or at least, maybe it did. Uncertainty over the
definition of “imminence” and questions about how to measure likelihood
raise concerns for some.44 Others question whether the Brandenburg test
implicitly incorporates any sort of Learned Hand-like balance, whereby
speech might be restricted even if there is little likelihood of actual incitement
because the potential product of such incitement presents a substantial danger
to human lives.45
38. Brandenburg, 395 U.S. at 447.
39. See Kaminski, supra note 21, at 42.
43. NAACP v. Claiborne Hardware Co., 458 U.S. 886, 928 (1982); Hess v. Indiana, 414
U.S. 105, 108 (1973) (per curiam). But see Holder v. Humanitarian Law Project, 561 U.S. 1,
8 (2010) (rejecting a facial challenge to the constitutionality of a federal law prohibiting the
provision of material support to organizations identified as “foreign terrorist organizations”).
As discussed below, the Holder Court did not even cite to or invoke Brandenburg in its
analysis. See infra note 111.
44. See, e.g., Kaminski, supra note 21, at 43–46.
45. See, e.g., RICHARD A. POSNER, NOT A SUICIDE PACT: THE CONSTITUTION IN A TIME OF
NATIONAL EMERGENCY 122 (2006) (“A huge harm unlikely to materialize for several more
The Supreme Court’s treatment of such unlawful advocacy reveals
important things. It is not inevitable that the law should treat advocacy of
unlawful conduct as speech. From a categorical-balancing perspective, one
could make a normative claim that such advocacy is much more likely to
cause broad, tangible social harm than, say, defamation of a private person.
What this suggests is that even the early twentieth-century Supreme Court
recognized that advocacy of unlawfulness has social value, even if its
decisions did not always reflect that. Without some type of meaningful
constitutional scrutiny, government regulation of such expression could
realistically suppress or chill what we might recognize as pure expressions of
Legal commentary often invokes the idea of exceptionalism across several
areas. Exceptionalism conveys the notion that, in specified areas of law
where the courts consistently apply a particular legal doctrine or analytical
framework, there exist certain subcategories of cases in which the courts
depart from that framework because they view those subcategories as
exceptional, or requiring a different set of rules.46 Often, this occurs without
explicit recognition about what the courts are doing. This suggests that the
courts themselves do not recognize that they are departing from legal norms
or that, if they do recognize it, they do not wish to publicly acknowledge it.
Not surprisingly, there are debates about how to identify exceptionalism
when the courts do not expressly articulate what they are doing and disputes
over whether it is actually occurring in any given area. In addition, even
assuming that there is a form of exceptionalism occurring in an area of the
law, there are frequently serious normative questions about whether it is a
This Part argues that two distinct forms of exceptionalism may skew the
development of First Amendment doctrine in an era of concerns about
terrorist incitement on the internet. First, there is the danger of national
security exceptionalism, a concept that has frequently been invoked by First
Amendment scholars to suggest that courts may relax free speech protection
in cases (or in specific eras) involving acute concerns about national
security.47 Second, there is the real possibility of internet exceptionalism, the
idea that courts may create new First Amendment rules to reflect the
“newness” of digital communication platforms because of concerns that the
years is not a lesser threat to the nation than a much smaller harm likely to materialize
46. See, e.g., Erik Luna, The Bin Laden Exception, 106 NW. U. L. REV. 1489, 1491 (2012)
(defining exceptionalism as “a comparative concept involving a contrast among sufficiently
analogous sets of values and practices, where an apparent anomaly or special case is subject
to descriptive and normative assessments”).
47. See generally STONE, supra note 6.
internet has fundamentally transformed human communication in ways that
previous generations of doctrine do not adequately accommodate.48
The contemporary “war” on terror roughly coincides with the proliferation
of easy and fast electronic means of communication. This implies that the
modern era is one in which free speech doctrine may be particularly
vulnerable to the dual pressures of these different forms of exceptionalism.
Email communication became widely available in the mid-1990s, shortly
before the 2001 terrorist attacks.49 Facebook and Twitter, two of the most
popular social media platforms, went online in 2004 and 2006, respectively.50
Thus, it is unsurprising that Brandenburg, itself a doctrinal correction to
exaggerated national security concerns from a previous generation, is under
great scrutiny as domestic terrorist incidents appear to be on the rise and
organized terrorist groups seem to use social media and other internet
communication platforms to expand their networks.51
First Amendment doctrine should be sensitive to the possibility that calls
to adjust the Brandenburg test and narrow the scope of speech protected by
its admittedly capacious standard may be skewed by the way these two forms
of exceptionalism come together (or as I say, get “squared”). Their
confluence may have the substantial capacity to distort free speech law in
ways that each form of exceptionalism alone might not (or might not to the
same degree). Courts should therefore tread cautiously before departing from
a legal regime that has, in retrospect, been largely successful in distinguishing
between actual incitement and other forms of speech that should remain
A. National Security Exceptionalism
The national security exceptionalism narrative runs strongly through a
broad band of legal scholarship. The general theory of this type of
exceptionalism is that courts show more deference to the state, and are
correspondingly less protective of civil liberties, during times of war or other
national security crises.52 From a descriptive standpoint, many legal scholars
and historians have argued that such exceptionalism has occurred during
various times in our nation’s history, most obviously during World War I,
World War II, and the Cold War.53 Although this Article focuses on speech,
others argue that it also extends to constitutional guarantees of equality and
privacy. Perhaps the case most often associated with this phenomenon is
Korematsu v. United States,54 in which the Court upheld the mass internment
of persons of Japanese ancestry living in the western United States during
World War II.
When it comes to judicial review of laws restricting expression during
wartime, the pattern, and historical evidence, is pretty clear. Although some
scholars have questioned the descriptive accuracy of this phenomenon,55 it is
not uncommon to see the Supreme Court openly embrace exceptionalism in
its First Amendment decisions. As Justice Oliver Wendell Holmes wrote in
Schenck, “[w]hen a nation is at war[,] many things that might be said in time
of peace are such a hindrance to its effort that their utterance will not be
endured so long as men fight and that no Court could regard them as protected
by any constitutional right.”56 The Espionage Act of 1917 itself, moreover,
applied by its own terms only “in time of war.”57 The subversive advocacy
cases that emerged during the Cold War were unequivocally influenced by
national security concerns, often expressly stated in the Court’s decisions.58
Some scholars have argued that national security exceptionalism is
normatively desirable. One common argument is that during emergencies
courts may be uncomfortable shaping long-standing constitutional rules
because they may not have full access to the intelligence information on
which government decisions are based.59 Additionally, some commentators
argue that there are credible concerns about institutional competence and that
courts are justifiably deferential when national security is at stake.60
Some of these same theorists have suggested that Brandenburg in
particular needs to be modified in response to the most recent era of threats
from terrorist actions, roughly post-9/11. Eric Posner, for example, notes that
“[t]he pattern in American history—and, in the other democracies as well,
even today—is that during times of national emergency, certain limits on
speech will be tolerated.”61 Posner views this not only as tolerable but
strongly desirable.62 He suggests that never before have individuals and
organizations from outside of the country been so easily able to use social
media to recruit and “radicalize” supporters and encourage them to engage in
violent or other criminal acts against the United States and its people.63 And,
glossing over Brandenburg, Posner notes that prior to the 1960s, people
could be punished for engaging in “dangerous” speech.64
Posner has proposed an unprecedented law that would make it a “crime to
access websites that glorify, express support for, or provide encouragement
for ISIS or support recruitment by ISIS; to distribute links to those websites
or videos, images, or text taken from those websites; or to encourage people
to access such websites by supplying them with links or instructions.”65 To
stave off concerns about legitimate uses of such sites, he would allow an
exemption for those who could show that they have research, journalistic, or
professional security justifications for accessing them.66
Critics of Posner’s approach have already identified several serious
concerns with his proposal. First, the difficulty in administering such a law
without censoring legitimate political advocacy is a significant danger.67
Second, it has been argued that Posner may substantially overestimate the
level of influence that ISIS propaganda has had in the United States and the
extent to which it might have instigated tangible harms.68 As Paul Gowder
points out, Posner’s column extrapolated his projection of harm from only a
single anecdote.69 Reliable estimates suggest that there have been
seventyone deaths from domestic terrorism events over a ten year period (2005
through 2015), which amounts to about seven per year.70 Moreover, that
figure includes deaths caused by terrorist attacks that are not related to
There are certainly other reasons to quarrel with Posner’s reasoning. First,
the speech doctrine relating to unlawful advocacy is, for obvious reasons,
focused on the speaker, rather than the persons incited to act. The dilemma
turns on the concern about punishing someone for pure speech; it was never
in doubt that the incited person could be punished for his or her actual
Second, in the case of social-media-inspired violence, most speakers
conveying ISIS propaganda are likely to be located outside of the United
States and therefore beyond the territorial jurisdiction of American criminal
laws.72 To be sure, this is why Posner’s proposal targets not the speakers but
those who domestically access, consume, and redistribute their speech on
social media or through more traditional means.73 But, as he points out, such
a law clashes with existing First Amendment precedent protecting the right
to receive information.74 Insofar as I can determine, the concept of punishing
the audience for a speaker’s expression is completely foreign to American
Third, there is an interesting and potentially troubling assumption about
human agency related to both Posner’s proposal and the Brandenburg test
itself. Part of the justification for punishing the inciting speaker is that
speakers in some circumstances will engage in such powerful rhetoric that it
will virtually overcome the will of the listener, compelling him to engage in
criminal conduct he would not otherwise have carried out. The central idea
here is not that the speaker was successful by virtue of her rhetorical
persuasiveness but that she has somehow been so mesmerizing or
provocative that the listener cannot help but act. To some degree, this
liberates the listener of any agency or moral responsibility. It suggests that
he cannot engage in autonomous cognitive function and could not possibly
have listened to the speaker and made his own decision to act based on
rational deliberation about the speech. Posner’s proposed law assumes that
cutting the audience off from such speech will protect them from what is
essentially mind control through internet postings. It is never questioned
whether this understanding of communication is consistent with what we
know about social psychology.
Posner is by no means the only legal scholar to call for at least some
modification to the Brandenburg test. Alexander Tsesis, one of the most
thoughtful of these critics, has criticized Brandenburg’s conclusion that only
laws that prohibit imminently dangerous speech are constitutional.
According to Tsesis, the Court’s decision “ignored a plethora of empirical
evidence about the long-term effects of racist and ethnocentric propaganda”
and showings that “[of]ten, ideologues prepare their followers for
broadbased, organized destruction through systematic, long-term
indoctrination.”75 This observation is important because it pinpoints part of
the danger to national security posed by the particular ideologies of the
radical groups associated with contemporary terrorism. But it is unclear how
the law could, consistent with generations of First Amendment doctrine,
address long-term invocation of rhetoric to influence even racially or
religiously hateful opinions, teachings, and beliefs.
In contrast, many theorists such as Geoffrey Stone have been critical of
national security exceptionalism, arguing that
[t]hroughout our history, judges have erred on the side of deference in times
of crisis. Like other citizens, judges do not want the nation to lose a war,
and they certainly do not want to be responsible for a mass tragedy. . . .
Moreover, . . . judges, like other citizens, are not immune to the fears and
anxieties of the moment. This makes them even more prone—indeed,
perhaps too prone—to err on the side of deference.76
Similarly, David Cole has explained that
there is reason to think that as a general matter in times of crisis, we will
overestimate our security needs and discount the value of liberty. Liberty
is almost by definition abstract; it is measured by the absence of control or
restraint. Fear, by contrast, is immediate and palpable; it takes physical
form as stress, anxiety, depression, a pit in the stomach, a bad taste in the
mouth. It is easy to take liberty for granted, and to presume that
government powers to intrude on liberty are not likely to be directed at
one’s own liberty. Fear affects us all, especially after an attack like that of
In my view, the current threats to national security are, like those in our
nation’s past, serious but also likely to be overestimated. That is not to
minimize the devastating loss of American lives that has already occurred in
the first quarter of the twenty-first century. But historical lessons should
nonetheless breed skepticism, at least for now. It is easy for the government
to assert abstract national security concerns, and for the public to believe
those assertions because citizens generally have limited access to information
about the degree and likelihood of current threats. It is, and always will be,
difficult to objectively assess such threats in real time.78 This is why
understanding our past record of addressing such threats is so crucial.
Moreover, those who view national security exceptionalism as a good
thing maintain that the dangers are not as great as we might think, since courts
typically return to more robust protection of civil liberties during
peacetime.79 But that argument assumes finite periods of national security
concerns, such as traditionally declared wars against foreign nations. To the
extent that current terrorism threats justify an “emergency” exception to First
Amendment speech protection, what do we do about the fact that the War on
76. STONE, supra note 6, at 544.
77. David Cole, Enemy Aliens, 54 STAN. L. REV. 953, 955–56 (2002).
78. A stark reminder of this is the granting of a writ of coram nobis invalidating Fred
Korematsu’s original conviction after the information on which the detention order was
premised became declassified. Korematsu v. United States, 584 F. Supp. 1406, 1420 (N.D.
79. POSNER & VERMEULE, supra note 53, at 42.
Terror is not really a war, but an ongoing global problem with no apparent
end? As Stone observes, “if the [Bush] administration is correct that the war
on terrorism will grind on indefinitely, that is all the more reason to be
scrupulous in scrutinizing proposed restrictions of civil liberties.”80
Yet another reason for skepticism about relaxing Brandenburg in the
current era is that, in the current global climate, the targets of restrictions on
civil liberties are more likely to be Muslims, a religious minority group (even
if those who are engaged in violent acts are at the outer fringes of Islamic
thought). As Korematsu should remind us, it is certainly not unknown to our
history for Americans to be particularly fearful of perceived national security
threats from those belonging to minority religious, cultural, and ethnic
Finally, skepticism brought on by exaggerated fears of national security
threats must, in this case, be multiplied by the corresponding fears of new
communication technologies. This claim is elaborated below.
B. Internet Exceptionalism
In considering whether to modify the Brandenburg test, we must also
account for concerns about internet exceptionalism. Mark Tushnet has
defined internet exceptionalism as “whether the technological characteristics
of the Internet (and, more generally, twenty-first century information
technologies) justify treating regulation of information dissemination
through the Internet differently from regulation of such dissemination
through nineteenth- and twentieth-century media, such as print, radio, and
television.”82 Translated to the current topic, the question is whether these
characteristics should push us toward different types of regulations of
Commentators have generally identified three distinctive features of the
internet and social media that might justify different treatment under First
Amendment doctrine. First, speech on the internet can be communicated
broadly and instantaneously (amplification).83 Second, communication on
the internet is much less expensive than other modes of communication
(cost).84 Finally, speakers using the internet can often mask their identities
to listeners, at least in the absence of technological tools unavailable to the
general public (anonymity).85 There is little dispute that these features make
communication on the internet distinctive from traditional media. The
80. STONE, supra note 6, at 554.
81. See, e.g., Aziz Z. Huq, The Political Psychology of Counterterrorism, 9 ANN. REV. L.
& SOC. SCI. 71, 83 (2013) (“[S]tudies . . . furnish support for the proposition that perceptions
of terrorism threat will tend to correlate with more stereotypical thinking, greater disapproval
of minorities, and increasingly authoritarian attitudes, although they disagree about the precise
mechanism at work.” (citation omitted)).
82. Tushnet, supra note 48, at 1638.
83. Id. at 1651–54.
84. Id. at 1654–58.
85. Id. at 1658–62.
question is whether those distinctions are sufficiently meaningful to require
a new or modified legal regime for the regulation of internet speech.
There has, of course, been much thoughtful scholarship addressing how
and to what degree the technological transformation that has occurred over
the past twenty years might require substantial alteration in the way existing
legal regimes operate.86 Moreover, calls for modifying First Amendment
doctrine in reaction to digital communication technology have ranged from
concerns about the ease with which the internet can act as a shield for people
engaging in hate speech and other forms of harassment,87 to the proliferation
of “fake news,”88 to the infringement on intellectual property rights.89
Here, the question is whether, in a manner similar to that reflected by
national security exceptionalism, courts may feel the need to defer to
government regulation of terrorist incitement made more possible because of
rapid advances in digital technologies. Such reactions may stem from fear
generated by legal decision-makers’ lack of familiarity with such
technologies just as they may stem from fear of threats to national security.
Much of the literature on internet exceptionalism is normative and argues
for or against idiosyncratic legal models for regulation of speech through
digital technologies.90 There are, however, doubts about the degree to which
the amplification, cost, and anonymity features change the qualitative nature
of speech, whether that relates to the speech’s value or the harms it might
cause.91 The amplification and cost factors no doubt make it possible for
speakers to reach a wider audience than traditional communication modes,
but do not necessarily change the content of the speech. For example,
socalled “hate speech” can be conveyed through old-school techniques, such as
direct verbal epithets, racist graffiti, or cross burning, but may potentially be
directed at a wider range of targets through the internet. Commentators argue
whether that ability creates substantially more harm to a larger group of
listeners than traditional communication.92 Claims regarding the anonymity
feature generally relate to the fact that speakers may be emboldened to make
more extreme statements when hiding behind artificial identities and may
also evade detection for internet speech that otherwise might be subject to
prosecution or other legal sanction.93
The argument for relaxing Brandenburg’s standard is that, to the extent
terrorist incitement is dangerous, that danger is substantially exacerbated
because of the availability of the internet. Of course, terrorist incitement can
take place through many different conventional communication methods,
including in-person conversation, written letters, leaflets, memos, telephonic
communication, and broadcasting over traditional television and radio
networks. But when conducted over the internet, that incitement is amplified
at low or no cost and more anonymous than any of the other potential means
Arguments for internet exceptionalism applied to incitement stem from
these features but also rely on other claims. First, some have argued that it is
easier for terrorists to radicalize their audience through social media and
other digital communication platforms.94 Eric Posner, again, writes:
Today, the Internet makes possible the constant circulation of captivating
videos, vivid images, and extremist text, creating a “radicalization echo
chamber.” It is the change in technology, more than the change in the
nature of foreign threats, that has given rise to a historic and unprecedented
danger from foreign radicalization and recruitment.95
If there is a meaningful causal link between internet communication and the
development of extremist views, and more importantly, actions based on
those views, then there might be a strong claim to allow greater regulation of
It has also been argued that the amplification feature of internet speech
provides another reason to permit greater regulation of inciting speech.96 Not
only can terrorists form networks around the globe, but they can also incite
violent actions by others located in distant lands with great speed. This calls
into question whether Brandenburg’s requirement that the speech in question
be likely to cause imminent harm is too narrow to permit regulation of
These claims in favor of internet exceptionalism for terrorist incitement
have materialized through various proposals for reform. As already
discussed, Posner calls for a law criminalizing the act of accessing terrorist
web sites or social media platforms.97 Lyrissa Lidsky argues for a more
anonymous internet users); Alexander Tsesis, Terrorist Speech on Social Media, 70 VAND. L.
REV. 651, 658 (2017) (“[T]he internet emboldens terrorists by providing them with the tools
to post anonymously and inexpensively”). But see Julie Seaman & David Sloan Wilson,
#freespeech, 48 ARIZ. ST. L.J. 1013, 1038 n.147 (2016) (citing a study finding little empirical
evidence of a causal relationship between internet anonymity and aggressive, antisocial
94. CASS R. SUNSTEIN, REPUBLIC.COM 2.0, at 69 (2007).
95. Posner, supra note 6. Underscoring that his main claim relates to technological
developments, Posner responds to some of his critics by pointing out that “[t]he question,
which no one seems to want to address, is whether social media and other forms of
Internetbased communication create opportunities for radicalization that pose a threat to public
safety.” Eric Posner, More on Anti-Propaganda Laws, ERIC POSNER (Dec. 19, 2015)
96. See, e.g., Tsesis, supra note 93, at 657–59.
97. See supra notes 61–66 and accompanying text.
limited adjustment to Brandenburg’s imminence requirement to address
some types of terrorist incitement. As she argues,
The imminence requirement serves to prevent suppression of speech based
on the government’s exaggerated fears of the danger posed by radical
speech. A satisfactory replacement for cyber incitement would focus on
ensuring a direct causal linkage between the speech and the harm, focusing
on factors such as the likely make-up of the target audience, whether there
was a prior history of violence by members of that audience, whether the
speaker supplied detailed instructions on carrying out the violent acts
advocated, and whether the violence took place with little delay upon
receiving the inciting speech.98
Thus, her claim is not so much for Brandenburg’s imminence requirement to
be relaxed but for a more nuanced consideration of the context in which the
speech takes place.
Tsesis does not overtly call for Brandenburg to be modified, but he argues
that its imminence requirement substantially limits the government’s
authority to regulate terrorist incitement on the internet. His argument is
more geared toward the idea that other legal doctrines might have to be
invoked to address such incitement. As he observes,
The incitement doctrine applies only to imminently dangerous statements
and is hence of limited value to combat internet terrorist incitement. A
statute containing such a component could be effective against immediate
calls for violence through applications such as Instagram or Snapchat. But
the bulk of internet terrorist speech seeks long-term indoctrination,
mentoring, recruitment, and so on; hence, policymakers need additional
Several things should give us pause before transforming or modifying
existing First Amendment doctrine in reaction to these changes. The same
things that make speech potentially more problematic also make it possible
to more effectively address those problems. Hateful speech is cheap and
easy, but so is counterspeech. Perhaps terrorist speech can be used to incite
a wider range of potential sympathizers, but so can propaganda that responds
to such provocation. One response to this argument, however, is that terrorist
social media networks may incite violence or other criminal conduct over
relatively secured networks so that counterspeech may not be a realistic
possibility in this scenario. But counterspeech need not necessarily take the
form of persuading a particular individual not to carry out a specific, unlawful
act. Widely broadcast propaganda refuting radical rhetoric might operate to
98. Lyrissa Barnett Lidsky, Incendiary Speech and Social Media, 44 TEX. TECH L. REV.
147, 164 (2011).
99. Tsesis, supra note 93, at 667. Unlike others, I do not address the so-called “true
threats” doctrine as applied to terrorist incitement on the internet. My hesitancy to do so stems
from the fact that this Article focuses on speech directed at a listener with the goal of inciting
that listener to harm a third person, not on speech that is directly threatening the listener.
Others suggest that the true-threats doctrine may be a better fit for terrorist speech on the
internet. See Martin H. Redish & Matthew Fisher, Terrorizing Advocacy and the First
Amendment: Free Expression and the Fallacy of Mutual Exclusivity, 86 FORDHAM L. REV.
565, 583–90 (2017).
dissuade potential terrorists. Admittedly, however, there are serious
limitations to the effectiveness of counterspeech if those who are predisposed
to accept radical terrorist views are less likely to consult or believe alternate
This brings us to the related issue of imminence. The expression of radical
terrorist groups through social media and other internet-communication
methods is often not the type of speech that fits neatly under the incitement
doctrine. For instance, the type of long-term radicalization or indoctrination
that these groups sometimes view as their goal relates to advocacy and
teaching about their ideology—albeit sometimes an ideology infused with
violence as a tactic—rather than immediate calls for a specific violent act.
Unlike speech spurring on an angry mob, there may be a substantial lag
between when speech is posted on a web page or Facebook and when an
audience member reads and acts on that speech. The ultimate goal of the
speaker might be to inspire the listener to engage in a violent act at some
indefinite time in the future, but that type of more abstract advocacy is what
Brandenburg was designed to protect. As Margot Kaminski observes,
If groups can form so quickly that police cannot react, there might be an
argument for ignoring the imminence requirement and allowing regulation
before the call to arms happens. This, however, is exactly why
Brandenburg has an imminence standard: the further back from actual
harm regulation gets, the more it impinges on free expression.101
Moreover, during the lag time, there could be other factors that cause a
reader to become radicalized as well as substantial opportunities for
counterspeech to dispute the call for violent action. Thus, Brandenburg takes
into account that there may be many reasons why the length of time between
the speech and the act may diminish the danger—counterspeech, doubt, and
regret developed internally by the listener, or concern about capture and
punishment, to name a few. These things are not likely to be true when an
angry mob is incited.
In addition, it is not clear just how much the breadth and speed with which
radical messages can be conveyed on the internet necessarily justifies internet
exceptionalism given our past experiences with other types of
communication. Other technologies prior to the internet also had the
advantages of relatively low cost and wide dissemination. The radical
rightwing “radio priest” Father Charles Coughlin, for example, is reported to have
reached an audience of as many as 30 million listeners through his weekly
radio broadcasts.102 The Egyptian Sheikh Omar Abdel Rahman—who was
later convicted of conspiring to bomb several New York City landmarks—
communicated and inspired his followers through conference calls and the
wide circulation of cassette tapes containing recordings of his messages.103
With regard to the anonymity that is enhanced by communicating over the
internet, there are two general responses. First, while anonymity could
certainly be a valid concern for law enforcement agencies trying to capture
and deter terrorists, as stated earlier, most inciters in this realm are likely to
be outside the United States and therefore beyond its criminal jurisdiction—
a much larger impediment to law enforcement than anonymity.104 Second,
as others have pointed out, the First Amendment was enacted against the
backdrop of a culture in which anonymous political speech was recognized
as a valuable feature of free expression.105
In general, I am sympathetic to concerns that the technological changes
wrought by the internet might not be as qualitatively substantial as initial
concerns might indicate. Or that, at the very least, we still do not know
enough about how digital communication might transform things such as
terrorist incitement to justify a major doctrinal change. Jamie Boyle’s book,
The Public Domain, describes a phenomenon that has useful parallels to the
arguments presented in this Article.106 Boyle argues that in the regime of
intellectual property, the content industry (i.e., movie studios and music
production companies) typically has adopted what he calls a “20/20
downside” vision of changes produced by digital communication and the
internet.107 This influences them to see only the downside of piracy threats
from the internet and to ignore or undervalue the corresponding potential
upsides.108 There is a degree to which that same danger might occur in the
context of terrorist incitement in the digital age. Some policy-makers, judges,
and scholars may have a 20/20 downside vision of what will happen if the
current First Amendment doctrine is not adjusted to deal with the dangers of
One might object that in the contemporary era, the national security and
internet issues are one and the same and that I am bootstrapping my
arguments by contending that their confluence is more reason for caution
than either factor alone. Moreover, one could argue that the multiplier effect
of both concerns—national security and internet communication—is
precisely the danger to society that justifies at least a temporary exceptional
approach to free speech doctrine.
In terms of the first claim, it is clear from our long history of contracting
civil liberties during crises that national security exceptionalism is by no
means limited to the advent of the internet. It is equally clear that internet
exceptionalism concerns are not confined to issues concerning national
security. This is exhibited by the discourse on how digital communication
technologies suggest a need to modify other areas of law, such as intellectual
property109 and privacy.110 With regard to the second argument, I maintain
that we should, at the very least, be skeptical about the potential to propose
exaggerated responses to multiple justifications for exceptionalism.
Moreover, as I will explain in Part III, if there are grave concerns about
national security and the internet leading to tangible social harms that cannot
be addressed in any way other than relaxing the Brandenburg standard, the
evidence of such a problem has yet to emerge in any concrete way.
III. BRANDENBURG TODAY
Although Brandenburg has been the subject of much academic debate,
particularly in recent years, it appears in practice to strike an appropriate
balance. In my view, even with some level of ambiguity, the test has largely
produced results that comport with the Court’s objectives. It seems in most
cases to protect speech that is merely teaching or advocating for ideas that
might be dangerous if implemented and to guard against the chilling effect
on such speech. But it also allows the government some latitude to prosecute
those who clearly intend to cause real harm to the United States and are
targeting an audience of people likely to cause that harm.
While there have been few calls to broaden the Brandenburg test to protect
more speech,111 there have been, as discussed, several suggestions about
narrowing its holding to permit broader regulation of terrorist incitement.112
Notwithstanding arguments for skepticism, one could argue that such reforms
would be justified if many individuals who engaged in dangerous speech are
successfully mounting Brandenburg defenses to their prosecutions. But an
examination of all reported federal cases since 9/11 that have cited to
Brandenburg does not really provide any such evidence.113 Most of those
cases have nothing to do with terrorism (again, perhaps because much
terrorist speech that might be considered as inciting may come from outside
109. Lessig, supra note 90, at 502.
110. See generally NEIL RICHARDS, INTELLECTUAL PRIVACY: RETHINKING CIVIL LIBERTIES
IN THE DIGITAL AGE (2015).
111. Unlike during earlier periods of unrest over national security concerns, there do not
appear to have been a plethora of abuses of criminal laws or enforcement of
subversiveadvocacy laws against academics or those who teach about different ideologies that might be
associated with contemporary terrorist movements. One might argue that the Court’s decision
in Holder v. Humanitarian Law Project, 561 U.S. 1 (2010), is an example of the unfortunate
relaxation of the Brandenburg test. Holder rejected a facial challenge—brought by lawyers,
journalists, and activists—to federal laws prohibiting the provision of material support to
groups identified as a “foreign terrorist organization.” The majority in Holder, however, did
not even cite Brandenburg, although Justice Breyer’s dissent did. Id. at 43–45 (Breyer, J.,
dissenting) (arguing that the Brandenburg standard would protect much of the plaintiffs’
speech). If Justices on the Court completely ignore Brandenburg’s applicability to a particular
type of criminal statute, it is hard to fault the rule of Brandenburg itself. Rather, the problem
isn’t Brandenburg per se, but the refusal to invoke it when it is quite arguably controlling.
112. See supra Part II.A.
113. The author examined these cases in September 2017 by searching for “Brandenburg
v. Ohio” in the Westlaw Federal Cases database with the date restriction “All Dates After
of the country and is therefore beyond the territorial jurisdiction of American
criminal law).114 A handful of cases dismissed charges because the
government failed to prove one of the three prongs of Brandenburg (intent,
imminence, or likelihood of unlawful action).115
Decisions adjudicating the claims of accused terrorists show that
Brandenburg does not protect speech by those who advocate and provoke
serious and imminent harm to national security interests. In a handful of
these cases, the courts have entertained, but ultimately rejected, Brandenburg
defenses.116 In each of these cases, the courts have rejected the First
Amendment claim and concluded that the United States has easily satisfied
the intent, likelihood, and imminence elements of the test.117
Also, in reviewing the available information about both successful and
failed terrorist plots inspired by internet communications, it appears that
many of the most serious threats do not come from generalized calls for
action or radicalization. Rather, these threats come from what can only be
viewed as direct, step-by-step incitement and hand-holding that, were the
speakers subject to criminal jurisdiction in the United States, would surely
meet the Brandenburg standard and permit their prosecution.118 As one
terrorism analyst has reported, “‘If you look at the communications between
the attackers and the virtual plotters, you will see that there is a direct line of
communication to the point where they are egging them on minutes, even
seconds, before the individual carries out an attack.’”119 Thus, the actions of
these groups far exceed the general type of advocacy that Brandenburg
Finally, again, there may be a degree to which the concerns about national
security and the distinctiveness of the internet are at least somewhat inflated.
With regard to national security, as cited above, the number of deaths related
to terrorism incidents on U.S. soil are relatively few, particularly when
compared to other causes of death.120 Furthermore, from the available data,
the role of the internet in such incidents is not even clear. According to an
extensive report by the Cato Institute, the internet and social media played
either little or no role in the bulk of domestic terrorism prosecutions brought
This Article has argued that the legal system ought to be skeptical about
calls to relax the free speech protections surrounding unlawful advocacy
because we live in a time when such calls may be unduly influenced by
inflated concerns about both national security and rapid technological
change. In claiming that the Brandenburg standard is sufficiently protective
of general advocacy while still allowing room for prosecution of truly
inciting and dangerous expression, I do not mean to minimize the human
damage that terrorist-inspired violence has so tragically caused in our recent
history. At the same time, as Geoffrey Stone has thoroughly catalogued,122
our longer history shows a tendency to overreact to what are perceived as
new types of national security threats. That history should give us pause
when reconsidering First Amendment doctrine that has served the mutual
values of liberty and security quite well for over a generation. Societal
reactions to the newness of the internet and social media platforms can
similarly skew our thinking about the need for government regulation. The
confluence of these two types of exceptionalism should cause us to be
exponentially wary of tinkering with the law of unlawful advocacy in the
absence of strong evidence that the law has impeded the ability of the
government to protect us from imminent danger.
1. ERWIN CHEMERINSKY , CONSTITUTIONAL LAW 1038 (5th ed. 2015 ).
2. Martin H. Redish , Advocacy of Unlawful Conduct and the First Amendment: In Defense of Clear and Present Danger , 70 CALIF. L. REV. 1159 , 1159 ( 1982 ).
3. For a thoughtful and comprehensive treatment of these issues, see KENT GREENAWALT, SPEECH, CRIME, AND THE USES OF LANGUAGE ( 1989 ).
4. 395 U.S. 444 ( 1969 ) (per curiam).
5. Id. at 447.
6. By now, there is already a rich post-9/11 literature on this important topic. For an argument that Brandenburg should be modified to some degree, see Eric Posner, ISIS Gives Us No Choice but to Consider Limits on Speech , SLATE (Dec. 15 , 2015 ), http://www.slate.com/ 10. See Dennis v. United States , 341 U.S. 494 , 503 ( 1951 ) (plurality opinion) (“No important case involving free speech was decided by this Court prior to Schenck v . United States [in 1919 ].”). But see David M. Rabban, The First Amendment in Its Forgotten Years , 90 YALE L.J. 514 , 520 ( 1981 ) (surveying judicial decisions and legal scholarship concerning free speech prior to World War I).
11. See Geoffrey R. Stone , Kenneth Karst's Equality as a Central Principle in the First Amendment , 75 U. CHI. L. REV . 37 , 43 ( 2008 ) (discussing the application of the “two-level” theory).
12. These are the most commonly invoked reasons for protecting speech. For a more extensive discussion and critique of these theories, see Alexander Tsesis , Free Speech Constitutionalism , 2015 U. ILL. L. REV . 1015 , 1016 .
13. Turner Broad . Sys., Inc. v. FCC , 512 U.S. 622 , 641 - 42 ( 1994 ).
14. Id .
15. 315 U.S. 568 ( 1942 ).
16. For a comprehensive discussion of the coverage-protection distinction, see generally Frederick Schauer, The Boundaries of the First Amendment: A Preliminary Exploration of Constitutional Salience, 117 HARV . L. REV. 1765 ( 2004 ).
17. Miller v. California , 413 U.S. 15 , 35 - 37 ( 1973 ).
18. Chaplinsky , 315 U.S. at 572.
19. Id .
20. United States v. Stevens , 559 U.S. 460 , 468 - 69 ( 2010 ). But see Genevieve Lakier, The Invention of Low-Value Speech , 128 HARV. L. REV. 2166 , 2177 - 79 ( 2015 ) (disputing the Court's historical account of low-value speech). II. EXCEPTIONALISM (SQUARED)
48. See , e.g., Mark Tushnet, Internet Exceptionalism: An Overview from General Constitutional Law , 56 WM. & MARY L. REV . 1637 , 1639 ( 2015 ).
49. World Wide Web Timeline, PEW RES. CTR. (March 11 , 2014 ), http://www.pewinternet.org/ 2014 /03/11/world-wide-web-timeline/ [https://perma.cc/W9BTLRPQ].
50. Company, TWITTER, https://about.twitter.com/en_us/company.html [http://perma.cc/Q6EG-CGPL] (last visited Oct . 16 , 2017 ) (scroll down to “Our leadership” and select the “Bio” link under “Jack Dorsey”); Our History , FACEBOOK, https://newsroom.fb.com/company-info/ [https://perma.cc/RQN3-XB5G] (last visited Oct . 16 , 2017 ).
51. See Rukmini Callimachi, Not “Lone Wolves” After All: How ISIS Guides World's Terror Plots from Afar , N.Y. TIMES (Feb. 4 , 2017 ), https://www.nytimes.com/ 2017 /02/04/world/asia/isis-messaging -app-terror-plot .html [https://perma.cc/GC77-RNCD].
52. Of course, exceptionalism is not limited to courts and can influence decision-making by other government institutions as well . See Sudha Setty, Obama's National Security Exceptionalism , 91 CHI. -KENT L. REV . 91 , 91 - 92 ( 2016 ).
53. See , e.g., ERIC A. POSNER & ADRIAN VERMEULE, TERROR IN THE BALANCE: SECURITY, LIBERTY, AND THE COURTS 3 ( 2007 ) ; STONE, supra note 6 , at 12-14; David Cole, The New McCarthyism: Repeating History in the War on Terrorism , 38 HARV. C.R.-C.L.L. REV . 1 , 1 ( 2003 ).
54. 323 U.S. 214 ( 1944 ). National security exceptionalism can also occur with regard to government efforts to compel nationalistic feelings and loyalty in its citizens . See generally Alan K. Chen, Forced Patriot Acts , 81 DENV. L. REV. 703 ( 2004 ).
55. Aziz Huq has observed that most of the theoretical work assumes that national security exceptionalism exists and debates only its normative desirability . See Aziz Z. Huq, Against National Security Exceptionalism , 2009 SUP. CT. REV . 225 , 229 . He argues that, at least in the area of legal challenges to emergency detention policies, there is no empirical support to suggest that this descriptive account is accurate . Id. at 226.
56. Schenck v. United States , 249 U.S. 47 , 52 ( 1919 ).
57. Act of June 15, 1917 , ch. 30, tit . 1 , 40 Stat . 217 , 217 (codified as amended at 18 U.S.C. §§ 791 - 799 ).
58. See Dennis v. United States , 341 U.S. 494 , 509 ( 1951 ) (plurality opinion) (“[T]he words [clear and present danger] cannot mean that before the Government may act, it must wait until the putsch is about to be executed, the plans have been laid and the signal is awaited .”).
59. POSNER & VERMEULE, supra note 53, at 5-6.
60. Id . at 17. A somewhat middle-ground position is that national security exceptionalism occurs, but that over time the nation learns from the previous generation's overexaggeration of such concerns, thus pushing government toward more modest responses . See Mark V. Tushnet , Defending Korematsu?: Reflections on Civil Liberties in Wartime, 2003 WIS. L. REV. 273 , 307 .
61. Posner , supra note 6.
62. Id .
63. Id .
64. Id .
65. Id . There is some disagreement about the various terms used to describe the so-called Islamic State, such as ISIS, ISIL, and Daesh. See Callimachi , supra note 51; Max Fisher, When a Phrase Takes on New Meaning: Radical Islam , Explained, N.Y. TIMES (June 16, 2016 ), https://www.nytimes.com/ 2016 /06/17/world/when-a -phrase-takes-on-new-meaning-radicalislam-explained.html [https://perma.cc/Z6TK-C2TU]. For purposes of this Article, I refer to this coalition of groups as ISIS because that currently seems to be the most common usage .
66. Id .
67. Paul Gowder, Let's Just Censor Eric Posner , MEDIUM (Dec. 15 , 2015 ), https://medium.com/@PaulGowder/ let-s-just-censor-eric-posner-instead- 9975537eb548 [https://perma.cc/AWC8-XFZ4].
68. Id .
69. Id .
70. Id .
71. Id .
72. See Alexander Tsesis, Prohibiting Incitement on the Internet , 7 VA . J.L. & TECH. 1 , 4 ( 2002 ) (suggesting that domestic law alone cannot handle problems of internet incitement).
73. Posner , supra note 6.
74. Id .; see also Stanley v. Georgia , 394 U.S. 557 , 564 ( 1969 ).
75. Tsesis , supra note 72, at 13.
86. See , e.g., Jack M. Balkin , Old-School/New-School Speech Regulation, 127 HARV. L. REV. 2296 , 2297 - 99 ( 2014 ); Frank H. Easterbrook, Cyberspace and the Law of the Horse, 1996 U. CHI . LEGAL F. 207 , 210 - 14 .
87. See Tsesis, supra note 72.
88. See Anthony J. Gaughan , Illiberal Democracy: The Toxic Mix of Fake News , Hyperpolarization, and Partisan Election Administration, 12 DUKE J. CONST. L. & PUB. POL 'Y 57 , 66 ( 2017 ).
89. See JAMES BOYLE , THE PUBLIC DOMAIN: ENCLOSING THE COMMONS OF THE MIND 54- 82 ( 2008 ).
90. Compare Easterbrook, supra note 86 , at 208, with Lawrence Lessig, The Law of the Horse: What Cyberlaw Might Teach , 113 HARV. L. REV. 501 , 502 ( 1999 ).
91. For an excellent discussion disputing whether those features make incitement of mobs over the internet substantially more problematic , see Kaminski, supra note 21 , at 78-82.
92. Tsesis , supra note 72, at 30.
93. See Reno v . ACLU , 521 U.S. 844 , 889 - 90 , 894 ( 1996 ) (O'Connor , J., concurring in in part and dissenting in part) (discussing challenges in identifying and holding responsible
100. SUNSTEIN, supra note 94, at 69.
101. Kaminski , supra note 21, at 80.
102. Rick Perlstein, I Thought I Understood the American Right . Trump Proved Me Wrong., N.Y. TIMES MAG . (Apr. 11 , 2017 ), https://www.nytimes.com/ 2017 /04/11/magazine/ithought -i-understood-the-american-right-trump-proved-me-wrong .html [https://perma.cc/W3J3-DKQF].
103. Jim Dwyer , The Voice That Called to Terrorists, Unstilled by Time or Even Death , N.Y. TIMES (Feb. 20, 2017 ), https://www.nytimes.com/ 2017 /02/20/nyregion/omar-abdelrahman.html [https://perma.cc/XU7Q-RKJX].
104. See Tsesis, supra note 72 , at 4.
105. See generally Andrew Koppelman, Entertaining Satan: Why We Tolerate Terrorist Incitement , 86 FORDHAM L. REV. 535 ( 2017 ).
106. BOYLE, supra note 89, at 63.
107. Id .
108. Id .
114. See Tsesis, supra note 72 , at 4.
115. See , e.g., Bible Believers v . Wayne County , 805 F.3d 228 , 244 ( 6th Cir . 2015 ) (en banc) (holding that members of a Christian group who attended an Arab festival and displayed anti-Islamic messages did not “advocate for, encourage, condone, or even embrace imminent violence or lawlessness”), cert . denied, 136 S. Ct . 2013 ( 2016 ) ; United States v. McCrudden, No . CR-11 -061 DRH , 2015 WL 1198544, at *26 ( E.D.N.Y. Mar . 16 , 2015 ) (holding that an email calling for revolution was unlikely to incite such conduct ), aff'd, 655 F. App'x 23 (2d Cir . 2016 ).
116. See , e.g., United States v . Al Bahlul, 820 F. Supp . 2d 1141 , 1245 - 50 (Ct. of Mil. Comm'n Rev . 2011 ) (considering but rejecting a Brandenburg defense), aff'd sub nom . Bahlul v. United States , 840 F. 3d 757 (D.C. Cir . 2016 ), petition for cert . filed, No. 16 - 1307 (Mar. 28, 2017 ); United States v . Rahman , 189 F.3d 88 , 115 ( 2d Cir . 1999 ) (“[W]hile the state may not criminalize the expression of views, even including the view that violent overthrow of the government is desirable, it may nonetheless outlaw encouragement, inducement, or conspiracy to take violent action”); United States v . Sattar , 272 F. Supp . 2d 348 , 374 (S.D.N .Y. 2003 ) (stating that “such acts and statements that 'instruct, solicit, or persuade others to commit crimes of violence' are not protected by the First Amendment and may be prosecuted” without explicitly invoking Brandenburg (quoting Rahman, 189 F .3d at 117)).
117. See supra note 116.
118. Callimachi , supra note 51 (quoting one terrorism expert as stating that foreign terrorist groups frequently engage in very explicit direction and control of individuals carrying out attacks in the United States) .
119. Id .
120. Gowder , supra note 67.
121. See TERRORISM SINCE 9/11: THE AMERICAN CASES (John Mueller ed., 2017 ed.), https://politicalscience.osu.edu/faculty/jmueller/SINCE.pdf [https://perma.cc/EMF4-3YVK].
122. See STONE , supra note 6.