Scholarly article on topic 'The First Amendment and Political Risk'

The First Amendment and Political Risk Academic research paper on "Law"

Share paper
Academic journal
Journal of Legal Analysis
OECD Field of science

Academic research paper on topic "The First Amendment and Political Risk"


Mark Tushnet1


First Amendment doctrine is at its core about the correct response to the fact that speech O

can increase the risk of social harm. First Amendment risk varies along several dimen- |

sions, including distribution of risk, its magnitude, and the magnitude of social benefit. g

After describing several cases in which the Supreme Court's assessment risk or harm d

seems mistaken, I describe the tendency over time for courts to replace doctrine articu- O

lated as standards with doctrine articulated as rules with exceptions. I explain why that h

tendency occurs and can be normatively justified, but that it can produce pathologies :

when the courts resist, for a variety of reasons, the proliferation of exceptions to the rules. 3


Speech can directly inflict harm, and can increase the risk that harm will occur. O

False statements about a person's life can injure that person's reputation, to the /

point where the person might lose a job. Speech vigorously criticizing a gov- ^

ernment policy can increase the risk that some people will take unlawful actions e that interfere with the government's ability to implement that policy. Such

speech can increase the likelihood that the government will abandon the 8,

policy, and adopt one that promotes social welfare to a lesser extent. First h

Amendment doctrine is at its core about the correct response to the fact that g

speech can increase the risk of social harm. §

Like all risks, First Amendment risk varies along several dimensions, of which §

I will focus on three.2 Risk can be distributed broadly or in a more focused way: ,

criticisms of government policy increase the risk that large numbers of people 1

1 William Nelson Cromwell Professor of Law, Harvard Law School, Cambridge, USA. E-mail: I thank John Goldberg, Robert Kagan, and participants in the Workshop on Public Law at Harvard Law School for helpful comments on an earlier version of this Essay.

2 I do not focus on the mechanism by which risk can be realized, for example by persuasion, or through psychological mechanisms that are thought to bypass cognitive professes, or through the reaction by those who disagree with the speech. Variation in these mechanisms explains some aspects of judicial doctrine, in ways that I explore in connection with variation along other dimensions.

© The Author 2012. Published by Oxford University Press on behalf of The John M. Olin Center for Law, Economics and Business at Harvard Law School.

This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (, which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.

doi:10.1093/jla/las005 Advance Access published on June 19, 2012

will bear the social losses attendant on law-breaking, whereas real threats to identified individuals primarily increase the risk that those individuals will be attacked.3 Risks can vary in magnitude as well: an anonymous pamphlet by a puny anonymity, to adopt one of Justice Holmes's phrases (Abrams v. United States, 250 U.S. 616, 629 (1919) (Holmes, J., dissenting)), increases the risk of law-breaking by a small amount; a speech by a person with great oratorical skills might increase the risk more.4 The circumstances under which words are uttered may affect the magnitude of the risk they impose.5 And, finally, speech varies in the magnitude of the social benefit it occasions: as the Court put it, some utterances are of low social value in their contribution to the discovery of truth or the development of public policy (Chaplinsky v. New Hampshire, 315 U.S. 568, 572 (1942)). In Justice Stevens's words, "though the First Amendment protects communication in th[e] area [of 'Specified Sexual Activities'] from total suppression", "few of us would march our sons and daughters off to war to preserve the citizen's right to see" such activities (Young v. American Mini Theatres, 427 U.S. 50, 70 (1976)).

To this point I have put my points as assertions about what speech does. But, a more precise formulation is required to understand free speech doctrine as a response to risk. Each of the assertion I have made should be prefaced with the phrase, "Democratically elected legislatures reasonably believe ...." So, for example, First Amendment questions arise when democratically elected legislatures reasonably believe that speech of a particular sort causes a specific kind of harm that they believe to be distributed in a particular way.6 Consideration of g

the institutional relationship between courts and legislatures is more central to O

3 Or increase the likelihood that they will take precautions against attack. Here too I forgo discussion era' of the argument that real threats do more than increase risk but upon their communication directly O inflict harm by instilling fear in their target. This latter effect can perhaps be understood as an J

assertion that the communication of a real threat in itself increases the risk of harm to 100%. §

4 The implicit reference is to Debs v. United States, 249 U.S. 211 (1919). "

5 Again, Holmes: "When a nation is at war, many things that might be said in time of peace are such a 5 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." Schenck v. United States, 249 U.S. 47, 52 (1919).

6 Two observations about this formulation: (i) Legislation can be challenged on the ground that the

belief that it addresses some identifiable harm is unreasonable. But, such a challenge should be understood as resting on general libertarian principles (substantive due process, or a reasonableness test applicable to all legislation) rather than on the First Amendment. (ii) The formulation is applicable to legal rules imposing liability pursuant to the common law, on two theories: (a) In most states the judges who develop the common law are elected or at least subject to retention elections, and (b) common law rules not displaced by state legislation, as they always can be, can be taken to have the implicit endorsement of the democratically elected legislature. I believe that the latter of these theories is the stronger one, but nothing much in what follows turns on it. (A more complete account would address the claim that state legislatures might have incentives to address some

analyzing judicially developed free speech doctrine than is a direct assessment of risk, its magnitude and distribution, and the social benefits of speech. Courts cannot completely avoid such direct assessments, but after they make a rough judgment about these matters, they must consider the institutional question of whether, and more important how, they should respond when their assessments differ from the legislature's. So, for example, how should a court respond when a legislature reasonably believes that speech made in connection with a commercial transaction increases the risk that social insurance schemes will pay too much to compensate consumers for their purchases (and mechanisms other than regulating the speech are less effective in reducing that risk)? The answer does not lie primarily in improving the courts' estimates of risk and the like, though that might help, but in judicial doctrine predicated on judgments of differential institutional competence.

Vince Blasi's classic article, The Pathological Perspective and the First Amendment (1985), argued that First Amendment doctrine rested on the accurate view that courts could reliably identify certain pathologies in the legislative process that predictably generated systematically excessive legislative assessments of the degree to which speech increased the risk of harm. For Blasi, First Amendment doctrine should be structured by those pathologies: For each pathology there should be a doctrine that corrected the excessive estimate of the risk of harm.7 Courts could then determine whether the regulation the legislature adopted was an appropriate response to the risk of harm accurately assessed.8

This article addresses a different pathology, located in the judicial branch rather than the legislative one. I motivate the argument by describing several cases in which the courts' assessment of the risk that speech causes harm seems

common law rules—for example, on products liability—but not others, such as the tort of intentional infliction of emotional distress.)

7 Some pathologies might be quite general. So, for example, John Hart Ely's perception that legislatures will overestimate the risk that criticism of their policies will undermine the policies' effective implementation seems to be a pathology applicable to all legislative policies, and so to all legislative regulation of speech critical of government policy. Other pathologies might generate specific ways in which legislatures overestimated risk, so doctrine responsive to one pathology might be inappropriate with respect to another.

8 My view is that Blasi's analytic structure is that the First Amendment identifies and rectifies legislative pathologies, and that due process rationality requirements apply to the evaluation of legislative policies with that rectification in hand. I need not argue here, though, that other understandings of Blasi's analytic structure are inferior to mine. Specifically, for present purposes it does not matter if one wants to say that First Amendment doctrine has two components, the identification and rectification of legislative pathologies, and in addition the constitutional evaluation of legislative policies with an accurate assessment of the risk that speech will cause harm in hand.

mistaken,9 either because the courts seem to be mistaken in thinking that the legislature's estimates of the risk of harm are excessive (a phenomenon that I will sometimes describe as a judicial underestimation of the risk of harm, to parallel Blasi's concern that legislatures overestimate harm), or because the courts are insensitive to questions about the distribution of harm.10 In conjunction with that description I offer a diagnosis of the judicial pathology, which, following Duncan Kennedy, I call the rule-ification of doctrine, that is, the tendency over time for courts to replace doctrine articulated in the form of standards with doctrine articulated in the form of rules with exceptions.11 I explain why that tendency occurs and can be normatively justified, but that it can produce pathologies when the courts resist, for a variety of reasons, the proliferation of exceptions to the rules. I conclude with a discussion of the obvious treatment, given that diagnosis—the injection of standards into the rule-ified system. I observe, though, that such an injection might not occur for the reasons that lead courts to rule-ify, and that in any event the tendency to rule-ification will assert itself even after an injection of standards.12

10 At some points I discuss the courts' constitutional evaluation of rules developed by other courts acting in their common-law capacity. Given his central concerns, Blasi sensibly did not treat common-law rules separately from legislatively enacted ones. The pathologies he identified are not directly applicable to common-law courts. Perhaps we might elide the differences by observing that legislatures' ability to modify common-law rules might justify attributing common-law rules to legislatures that failed to modify or repeal the common-law rules. See supra note 6.1 note, though, that the real-world processes for adopting legislative policy and repealing or modifying judicial policy are different enough that we should not assume without detailed argument that the pathologies that affect the former processes affect the latter in the same way or to the same degree. Learned Hand suggested that judges were subject to the same pathologies in his comment on the inadequacy of Holmes's "clear and present danger" test, which he regarded as leaving too much discretion in the hands of "Tomdickandharry, D.J.", observing that "the Nine Elder Statesmen have not shown themselves wholly immune from the 'herd instinct'" common to legislators and jurors. Cited in Gunther (1975, 749).

11 I believe I take the term from Kennedy, but neither he nor I can locate an original source. In a private communication Robert Kagan suggested an alternative term, such as "hard" rule-ification, to distinguish the concept presented here from another prevalent in a different literature, on the proliferation of rules regulating some defined subject matter. I have chosen "rule-ification" for ease of exposition.

12 Rule-ification is a general form of acontextual decision making of the sort Paul Horwitz explores in his forthcoming book on First Amendment institutions.

For present purposes I set aside the possibility that the decisions rest on simple value disagreements 8

between other law-makers (legislatures, state court judges, juries) and the Court's majority, because n

(i) the Court's opinions are not cast in terms of such disagreements, and (ii) such disagreements are g

generally regarded as inappropriate grounds for constitutional doctrine (however often they might S2.

be the actual grounds). o


I begin by describing several recent constitutional controversies in the United States, which I treat as the symptoms of the pathology I am interested in: (i) Snyder v. Phelps, the funeral protest case; (ii) United States v. Stevens, the "animal crush video" case. (iii) Sorrell v. IMS Health Inc., involving state limitations on the ability of druggists to sell information about prescriptions to drug companies, which use that information to target specific doctors for their sales efforts.13

2.1. Distribution of Harms and Snyder v. Phelps d

2.1.1 Description r

Snyder v. Phelps is the well-known "funeral protest" case (131 S.Ct. 1207 I

(2011)). Members of the Westboro Baptist Church believe that the deaths of :

U.S. soldiers in combat express God's judgment that the United States has a

become sinful because of its toleration of homosexuality. They express that f

belief by engaging in demonstrations near soldiers' funerals, at which they O

display signs such as "Thank God for Dead Soldiers", "Fags Doom Nations", n

and "You're Going to Hell". They held one such demonstration just before the O

burial of Matthew Snyder, who had been killed in Iraq. They displayed their a

signs at a place where they were entitled to be, about 1000 feet from the church Un

where the funeral was held.14 Snyder's father sued the Westboro church mem- r

bers for the tort of intentional infliction of emotional distress. He had seen only I

the tops of their signs but not the signs' content as he drove to the funeral, but M

learned of the content when he saw a television news broadcast about the h

demonstration. Instructed that liability required that it find that the g

Westboro church members "intentionally or recklessly engaged in extreme J

and outrageous conduct that caused [Snyder] to suffer severe emotional dis- I

tress", the jury returned a multi-million dollar judgment against the church I

members. 1

In 1988, the Supreme Court held that the First Amendment barred recovery for intentional infliction of emotional distress on a public figure (Hustler

13 There is some overlap between this article and Schauer (forthcoming). Schauer focuses on the three cases dealt with here, emphasizes that, unlike some classic cases, they are ones in which legislators (and sometimes jurors) could reasonably believe that the proscribed actions caused significant social harm, and discusses some aspects of the distributional issues I discuss as well.

14 A subsequently enacted state law prohibits demonstrations within 100 feet of funerals. The

Westboro church members' demonstration would not have been unlawful under that statute. See 131 S.Ct. at 1218 n. 5.

Magazine, Inc. v. Falwell, 485 U.S. 46 (1988)).15 In Snyder v. Phelps the Court extended this rule to cover actions filed by private figures, where the speech was "of public ... concern, as determined by all the circumstances of the case" (131 S. Ct. at 1215). The Court did assert that its holding was "limited by the particular facts before us" (id. 1220), but under its analysis those facts were relevant only to the determination of whether the church members' speech was on a matter of public concern,16 which it clearly was.17

The Court chose the "matters of public concern" rule because, in its view, the constraints on jury discretion embodied in the instructions—"severe harm", "outrageous and extreme" behavior, and (in my view most important) that the behavior be "intentional"—failed adequately to take both public benefit and private harm into account. As I understand the Court's opinion, the difficulty with the instructions is that they are too complicated.18 A rule that precludes recovery for statements on matters of public concern is certainly simpler than one allowing recovery under some rather limited circumstances. Yet, I find it unclear why simplicity should control, at least where the statements were made intentionally, which I understand to mean, in the present context, made for the very purpose of inflicting harm (in addition to contributing to discussion of a matter of public concern).19 I do not find it obvious that a rule allowing the

15 There a pornographic magazine had published a "parody" advertisement, labeled as such, saying that the "first time" the prominent political preacher Jerry Falwell had had intercourse was in a drunken incestuous encounter with his mother. Falwell, not his mother, filed the action.

16 See also 131 S.Ct. at 1218 footnote 4 ("The fact that Westboro conducted its picketing adjacent to a public street does not insulate the speech from liability, but instead heightens concerns that what is era at issue is an effort to communicate to the public the church's views on matters of public concern."). This shows that the location is relevant not as part of a general assessment of all the circumstances, but only to the determination of whether the speech was on a matter of public concern. I think it is an interesting question: Why is it permissible to take the circumstances into account in determining whether the speech is about a matter of public concern but not with respect to the broader question of whether regulating the speech was justified by the balance of harms and benefits?

17 The Court conceded the possibility that a few of the signs, which used the word "You", might have been taken to be references to Matthew Snyder as an individual, but stated, correctly in my view, that "the overall thrust and dominant theme of Westboro's demonstration spoke to broader public issues." Id. at 1217.

18 For my discussion of the possibility that the "outrageousness" standard raises questions of de facto viewpoint-discrimination, see text accompanying note 26 infra.

19 Were there some ambiguity about the interpretation of the terms "intentionally or recklessly" in the jury instructions, the Court could have stated that recovery was possible only where the jury concludes that the statements were made for the very purpose of inflicting harm, and vacated the judgment on the ground that the instructions did not conform to that requirement. See also 131 S. Ct. at 1223 (Alito, J., dissenting) ("When grave injury is intentionally inflicted by means of an attack like the one at issue here, the First Amendment should not interfere with recovery.").

imposition of liability for statements on matters of public concern made for the very purpose of inflicting emotional harm (and succeeding in that purpose) would "stifle public debate" in a manner inconsistent with the choices "we" have made in the First Amendment.

2.1.2 Critique

Chief Justice Roberts closed his opinion in the funeral protest case with this paragraph:

Speech is powerful. It can stir people to action, move them to tears of 0

both joy and sorrow, and—as it did here—inflict great pain. On the I

facts before us, we cannot react to that pain by punishing the speaker. d

As a Nation we have chosen a different course—to protect even hurtful f

speech on public issues to ensure that we do not stifle public debate. |

That choice requires that we shield Westboro from tort liability for its t

picketing in this case (131 S.Ct. at 1220). l

The distributional question is apparent here: "We have chosen... to pro- ^

tect... hurtful speech on public issues to ensure that we do not stifle o

public debate." The public benefits from the debate, but the victim bears the |

harms. o

How should we think about the Court's argument? I divide my discussion a

into two parts: (i) Whether the Court should adopt a rule, and (ii) What rule the I

Court should adopt. This section addresses the first question, with some dis- |

cussion of the second; later sections address the second. 4

The Court adopts a rule that a victim cannot recover for a speaker's g

intentional infliction of emotional distress if the vehicle for inflicting that dis- i

tress is a comment on a matter of public concern. Consider an alternative, g

suggested indeed by the Court's expressions of sympathy for the damage the J

victim suffered in Snyder.20 The Court could make a case-specific all-things- g

considered judgment about whether, taking everything into account, the pub- 4

lic benefit of the protester's speech outweighed the harm inflicted on the 2 victim.21 Assume that the Court's expressions of sympathy mean that, had it

20 "Westboro's choice to convey its views in conjunction with Matthew Snyder's funeral made the expression of those views particularly hurtful to many, especially to Matthew's father. The record makes clear that the applicable legal term—'emotional distress'—fails to capture fully the anguish Westboro's choice added to Mr Snyder's already incalculable grief." Id. at 1217-1218.

21 A more common term, though less transparent, is "ad hoc balancing". I note that case-specific all-things-considered judgments have no precedential effect: the next similar case will be evaluated according to its own facts, without reference to those in the first one.

made such a judgment, it would have upheld the liability award.22 Why would the Court abjure making that judgment in favor of the "matters of public concern" rule?

The usual argument is that rules are better than case-specific all-things-considered judgments when invoking the rule in all the cases to which it applies produces a better set of results than the results produced by case-specific all-things-considered judgments. But, this argument needs unpacking. Were the Court in a position to make case-specific all-things-considered judgments in every relevant case (that is, in every case in which a jury imposed liability for intentional infliction of emotional distress), such judgments would necessarily be better than those produced by applying the "matters of public concern" rule.23 My sense is that there are relatively few such cases,24 which suggests that there might be no need for a rule of any sort.25

Suppose, though, that the Court believes that it is not in a position to review all cases in which liability is imposed. It could direct those who will effectively make the final decision—those who I will call the Court's "targets"—either to make case-specific all-things-considered judgments, or it could direct them to follow some rule, in which event it will have to determine what rule the targets should follow. To simplify exposition, I will identify the targets in IIED cases as the judges who develop the instructions given to juries in such cases.26 A rule

22 I note the possibility that these expressions of sympathy were designed to take the sting out of the Court's ruling, explaining that a result that might seem harsh to the uninformed is actually justified when seen in some larger frame.

23 Because the only effect of such a rule is, as in Snyder v. Phelps, to deny recovery in cases whether the case-specific all-things-considered judgment is that liability is appropriate.

24 Zipursky (2011) provides an overview of the common law IIED tort, generally supportive of the observation in the text. Zipusky also notes that cases involving IIED in connection with funerals form a reasonably well-defined subset in the IIED universe, which suggests to me that recognizing liability in Snyder would not interfere with speech on matters of public concern outside that specific setting. (Zipursky's article was largely completed before the Curt decided Snyder, but it contains a "Postscript" on the Court's decision.)

25 Note that the analysis would be different if a victim could claim that his or her constitutional rights were violated by a failure to impose liability for intentional infliction of emotional distress, because the relevant set of cases would then include all IIED cases. European human rights law, which generally recognizes a constitutional right to the protection of human dignity, might be a model, and the case for rules would accordingly be stronger in such a system. But, in the United States victims cannot generally claim that failure to find liability violates their constitutional rights, and so the relevant set of cases is as described in the text.

26 I believe that the analysis would be the same were the targets identified as the juries who impose liability (because juries do so according to instructions given them by judges). I note, though, that

juries might be susceptible to their own pathology—roughly, a systematic tendency to find liability

when the victims are socially valued people (and the speakers socially disvalued ones), and not when the victims are socially disvalued and the speakers valued. Cf. Snyder, 131 S.Ct. at 1219 (" 'Outrageousness'... is a highly malleable standard", and its application raises an "unacceptable"

would be appropriate if the Court could not trust those judges to make the correct case-specific all-things-considered judgments, where "correctness" is defined as matching the judgments the Court would make were it to review all the lower court decisions.27 And, perhaps the Court might properly lack such trust, because those judges are less able to make "good" case-specific all-things-considered judgments.28 A rule would then be appropriate.29

What rule, though? For present purposes, there are two candidates. One would allow liability where juries found that "the defendant intentionally or recklessly engaged in extreme and outrageous conduct that caused the plaintiff

risk of suppressing "' "vehement, caustic, and sometimes unpleasant[]"' " statements."). The ques- e

tion of viewpoint discrimination is, I think, a bit more complex in the context of common-law r

actions for intentional infliction of emotional distress than in the usual context of statutory regu- |

lations. The problem is one of underinclusiveness: people who intentionally inflict emotional dis- t

tress in an outrageous manner on socially favored groups will be held liable while those who /

intentionally inflict emotional distress on socially disfavored ones will not. But, those held liable a

have no real claim that they are being treated wrongly, having themselves inflicted harm. The usual X

justification for finding underinclusiveness troublesome is that doing so encourages law-makers to o be more careful in defining liability standards, and that greater care—meaning, in the present

context, willingness to define such standards in ways that will lead to imposing liability on those g

who harm members of socially disfavored groups—might lead law-makers to eliminate the possi- S

bility of liability altogether. Cf. R.A.V. v. City of St. Paul, 505 U.S. 377 (1992) (holding unconstitu- g

tional as viewpoint discriminatory, an ordinance punishing as hate speech, a subcategory of fighting T

words defined with reference to the group characteristics of the speech's target). Yet, finding a rule ^

unconstitutional for underinclusiveness is strong medicine, and one would want to be reasonably V

sure that liability standards could not be refined to reduce the risk of viewpoint discrimination. That S

assurance would come from a careful examination, forgone in Snyder, of the jury instructions taken S

as a whole. S

27 A note on another possible target—the Supreme Court itself in the future—is probably appropriate here. For reasons I develop in more detail elsewhere (Tushnet 1997), I doubt that future Courts are era sensibly treated as potential targets for rules. Briefly: today's Court cannot ensure that its successor will follow the rule it sets out, rather than distinguish or in the extreme overrule it, unless it (x) accurately believes that successor Courts will be less talented in distinguishing and other techniques of legal analysis than it itself is and (y) accurately believes that successor Courts will recognize their lower level of talent and act on that recognition. My view is that condition (y) will almost never be satisfied.

28 The selection processes for lower courts (elections in most states, a patronage-influenced process in many states and the federal system) compared to the high-attention process for Supreme Court justices suggests that the average quality of lower courts will be lower than the quality of judges on the Supreme Court (although the small number of Supreme Court justices suggests that variance might be higher there). Concretely, lower courts are better targets for rules than successor Courts because the former are less likely than the latter to develop rationales for distinguishing prior decisions that "feel" satisfying to well-socialized lawyers.

29 Notably, the strongest justification for the Court's stringent rules for allowing criminal punishment for speech lies precisely in the judgment the Court reached over time that even juries given instructions that appeared to confine their decisions quite a bit would too often find liability, and that lower court judges were disinclined to refuse to submit cases to juries unless the judges' discretion were quite tightly confined.

to suffer severe emotional distress".30 The jury is not asked to make a case-specific all-things-considered-judgment; rather, it is asked to find that the defendant's actions were intentional or reckless, extreme and outrageous, and caused severe harm. The other allows liability where the defendant satisfies all those requirements, but only with respect to statements that were not comments on matters of public concern. The relevant question then is: which of these rules produces a set of results that better matches the results the Court itself would reach were it to make case-specific all-things-considered judgments in the cases it is unable to review?

The distributional question does not disappear when the question is properly formulated, because the right answer might be that the Court's "matters of public concern" rule is better than the rules embedded in the IIED tort ("better", again, defined with reference to the case-specific all-things-considered judgments the Court would make). I can only express my sense that the rules embedded in the tort probably confine liability tightly enough that the harm inflicted in cases where properly instructed juries find liability probably does outweigh the public benefit of the speaker's comments.31 Whatever one's views on the alternatives, it remains true that the Court—consistent with its course of rule-ification—chose a simpler over a more complex rule.

2.2. Magnitude of Harms and United States v. Stevens

2.2.1 Description

United States v. Stevens held unconstitutional a federal statute making the production of "animal snuff films" illegal (130 S.Ct. 1577 (20 1 0)).32 Defending the statute, the government argued that such films were not covered by the First Amendment based on the description of uncovered categories of speech in Chaplinsky v. New Hampshire (315 U.S. 568 (1942)). According to the unanimous Court in Chaplinsky, "There are certain well defined and narrowly limited classes of speech, the prevention and punishment of which have never been thought to raise any Constitutional problem. These include the lewd and obscene, the profane, the libelous, and the insulting or 'fighting' words ...." (id. 571-572). The Chaplinsky Court continued with an explanation of why speech

30 These are the requirements for finding liability under Maryland law as applied in Snyder v. Phelps. 131 S.Ct. at 1215.

31 I note, though, that the "outrageousness" component of the IIED tort might be particularly susceptible to the kind of as-applied disparate impact noted above, note 26 supra. My sense is that the component could be tweaked to reduce that possibility, so that the results under a refined definition of the tort would better match the case-specific all-things-considered judgments the Court would reach. But, in any event, I believe that it is that question on which the better analysis should focus.

32 The following five paragraphs are adapted from Tushnet 2011.

in these classes was not covered: because the words "are no essential part of any exposition of ideas, and are of such slight social value as a step to truth that any benefit that may be derived from them is clearly outweighed by the social interest in order and morality" (id. 572).

In Stevens the government argued that this explanation was as applicable to animal snuff films as it was to the previously identified "classes of speech". The reason for treating those classes of speech as outside the First Amendment's coverage being equally applicable to animal snuff films, the prohibition of animal snuff films should not be "thought to raise any Constitutional problem".

Chief Justice Roberts, writing for the rest of the Court save Justice Alito, rejected the government's argument. According to the Chief Justice, "Our Constitution forecloses any attempt to revise that judgment [about the social benefits and costs of First Amendment protection] simply on the basis that some speech is not worth it" (Stevens, 130 S.Ct. at 1585). Yet, Chaplinsky shows that there are some categories of speech that are not covered by the First Amendment "on the basis that some speech"—fighting words, obscenity, and the other enumerated categories—"is not worth it". And, the government contended that animal snuff films fit into the generalized description of uncovered categories Chaplinsky offered.

The Chief Justice responded to the government's argument by denying that the reasons Chaplinsky offered for the categories' exclusion from coverage were relevant to the fact of their exclusion. Those reasons, according to the Chief Justice, were merely "descriptive", not justificatory, and so were irrelevant (id. 1586). What mattered was the seemingly prefatory comment in Chaplinsky, that these categories "had never been thought to raise any Constitutional problem". The Chief Justice said that those categories were to be determined by a purely historical test: uncovered categories were those "that have been historically unprotected" (id.). The reasons for the exclusion that Chaplinsky gave were irrelevant. The only uncovered categories of speech, according to this argument, are those that have been held to be uncovered from the time of the Framing.33

To the extent that the Court offered a defense of choosing the historical rather than the functional interpretation of Chaplinsky, it relied on a common but misleading trope in recent decisions. As the Chief Justice

33 The Chief Justice introduced a purely ad hoc exception to the historical test for the recently recognized category of child pornography, a category that includes material that would not fit the Court's definition of obscenity. According to the Chief Justice, child pornography could be prohibited because "[t]he market for child pornography was 'intrinsically related' to the underlying abuse, and was therefore 'an integral part of the production of such materials, an activity illegal throughout the Nation'" Id. (quoting New York v. Ferber, 458 U.S. 747 759, 761 (1982).

interpreted the government's argument, it offered a "free-floating test for First Amendment coverage", in which "[w]hether a given category of speech enjoys First Amendment protection depends upon a categorical balancing of the value of the speech against its societal costs" (130 S. Ct. at 1585). That "startling and dangerous" proposition had to be rejected because "[t]he First Amendment itself reflects a judgment by the American people that the benefits of its restrictions on the Government outweigh its costs " (id.)34 That, though, is unresponsive to the government's position. The issue in the case is: what are the restrictions the First Amendment places on the government? Once we know that the First Amendment covers a category of speech, of course we do not do the balancing the Chief Justice describes, and for the reason he gives. But, observing that covered speech is not subject to a balancing test tells us nothing about what speech is covered.35

The Court's trope reflects what I have described elsewhere as a fear of judgment (Tushnet 2009).36 The Court adopts originalism in part to avoid the judgments it thinks required by balancing and similar analyses. Yet, even ori-ginalism cannot avoid judgments. The historical materials almost never deal with the precise problem presented today. An originalist must therefore either confine himself or herself to what scholars have called the actually expected applications of constitutional terms37 or devise some method for analogizing

34 Id. For the trope's use elsewhere, see Sorrell v. IMS Health Inc., 131 S.Ct. 2653, 2665 (2011) ("The g Constitution 'does not enact Mr. Herbert Spencer's Social Statics.' It does enact the First S2 Amendment."); District of Columbia v. Heller, 554 U.S. 570, 636 (2008) ("the enshrinement of ° constitutional rights necessarily takes certain policy choices off the table."). I note the possibility g that the trope has the effect of shifting responsibility for decision from the Court to the First Amendment's adopters. For a peculiar juxtaposition using the trope, see Texas v. Johnson, 491 era U.S. 397, 420-421 (1989) (Kennedy, J., concurring) ("we are presented with a clear and simple statute to be judged against a pure command of the Constitution. The outcome can be laid at no door but ours. The hard fact is that sometimes we must make decisions we do not like. We make them because they are right, right in the sense that the law and the Constitution, as we see them, compel the result.").

35 I think that the Chief Justice is able to write what he does because he engages in a transparent—and therefore ineffective—sleight-of-hand. The government had argued for using a categorical balancing approach to determine whether a category of speech is covered. The Chief Justice describes the government's position as asking for some sort of "ad hoc balancing of relative social costs and benefits" to determine whether a category of speech is covered. See also id. at 1586 (implicitly rejecting the argument that Chaplinsky justifies lack of coverage when "an ad hoc calculus of costs and benefits tilts in a statute's favor"). Frankly, I simply cannot understand what the Chief Justice means by this odd blend of categorical and ad hoc balancing.

36 I argue below that the fear of judgment reflected in the trope operates only a bit below the surface of the Court's decisions, and that seeing the Court as affirmatively attracted to rule-ification provides a deeper understanding of these decisions.

37 See Greene (2009, 662) (arguing that the difference between original meaning and original expected

application is "a question of the level of generality").

such expected applications to the problem at hand. So, for example, in District of Columbia v. Heller, the Court rejected as "bordering on the frivolous" the argument that the "arms" that individuals had a right to bear were muskets and the like (Heller, 554 U.S. at 582). Rather, the term referred ("prima facie") to all "bearable" weapons, and then to weapons in common use for purposes of self-defense in cases of confrontation (id.). Drawing that conclusion required the Court to identify the characteristics of the relevant weapons—common use, self-defense, and the like. Yet, those weapons had other characteristics, such as their accuracy, the degree to which they were lethal in the hands of those who possessed them, which might in turn depend on the amount of training in weapons use people generally had, and the like. Choosing the characteristics treated as relevant requires the Court to exercise judgment, and, I think, a judgment not different in the respects bearing on the Court's fear of judgment from that required by a functionalist interpretation of Chaplinskys language.38

Fear of judgment cannot support the Court's choice of the historical over the functionalist readings of Chaplinskys language.39 I suggest below that the better explanation lies in a different jurisprudential direction, one that points to the relation between the Court as rule-deviser and other law-makers as

38 For another example, consider the exchange between Justices Alito and Scalia in the oral argument of Brown v. Entertainment Merchants Ass'n, 131 S. Ct. 2729 (2011):

JUSTICE ALITO: Well, I think what Justice Scalia wants to know is what James Madison thought about video games.

JUSTICE SCALIA: No, I want to know what James Madison thought about violence.

Transcript of Oral Argument, Brown v. Entertainment Merchants Ass'n, No. 98-1448 (Oct. Term 2010), p. 17. In the event Justice Scalia, writing for a plurality, chose violence as the proper characterization. Justice Thomas, though, argued forcefully that the proper characterization was something like "material to which the founding generation would have allowed children access without their parents' permission". The disagreement between Justice Thomas and Justice Scalia shows that originalists cannot avoid making some kinds of judgments, though of course different ones from those that functionalists make.

39 A related possibility is that rule-ification reflects something like what Karl Llewellyn called a "period style", a characteristic mode of argument associated with a historical era and explicable, if at all, by judicial preferences rooted in psychological and sociological grounds, the nature of which Llewellyn left largely unexplored. For an exposition of Llewellyn's views on period-style, see Llewellyn 1960, 36. If there is a period-style of rule-ification, one would expect to find it expressed in other doctrinal areas as well. It maybe worth noting that Justice Alito resisted the pull of rule-ification in Stevens and Phelps, and that Justice Breyer does so generally (though not uniformly, as his agreement with the outcome in Stevens indicates). These observations in turn suggest that we should be wary of associating rule-ification with a specific political tendency. (That point has been made in connection with the related distinction between rules and standards by, among others, Sullivan 1992.)

rule-devisers, and, because of that relation, a preference on the Court for what it believes to be simple rather than complex analyses.40

2.2.2 Critique

The IIED problem might be addressed through case-specific all-things-considered judgments because the number of cases may well be small enough to allow the Court to review them all. Other problems will clearly generate too many cases to be dealt with directly. One alternative is category-based balancing.41 Here the Court identifies a category of expression—fighting words, hate 0 speech, obscenity, political discourse—and aggregates the social harms (the | increased risk of social harm) associated with all instances in that category, j aggregates the social benefits similarly associated, and permits or prohibits f regulation depending on whether in its judgment the aggregated harms g exceed the aggregated benefits. p The analytic structure for evaluating category-based balancing is similar to a that described above: were the Court in a position to make the category-specific f

judgments with respect to every possible category, it should do so. But, legal jj

categories are not natural objects, and their proliferation depends primarily on |

) • 42 1

advocates' skills. Legislatures and lower courts will find it relatively easy to O

identify some new category of speech different from those the Court has pre- c/

viously recognized, and the Court will be unable to police those decisions u

through direct review. Instead, it must provide criteria for determining when g

a new category can be created. gg

That was the problem addressed in Stevens, the crush video case. To return to M

the argument's structure: the government contended that Chaplinsky provided h

the criteria, in its statement that "There are certain well defined and narrowly |

limited classes of speech, [a] the prevention and punishment of which have §

never been thought to raise any Constitutional problem.... [S]uch utterances g

40 For a brief discussion bearing on whether historical inquiries are indeed simpler than functional ones at the level of trial and intermediate appellate courts, see note 49 infra. I am inclined to think that they are not simpler even at the Supreme Court level, except in exceptional cases such as Heller, where a large body of historically oriented scholarship was available to the Court. Of course, the Court's preference for historical inquiries may elicit relevant scholarship, thereby increasing the accessibility of the material the Court treats as relevant, to both the Court and lower courts.

41 I substitute the term "balancing" for "all-things-considered judgment" here because making the category the relevant unit of evaluation necessarily screens out some case-specific features of the cases, and "all-things-considered with respect to the category" is too cumbersome.

42 For a good example, see Brown v. Entertainment Merchants Ass'n, 131 S.Ct. 2729 (2011), where the Court, through Justice Scalia, described the relevant category as "speech depicting violence," and Justice Thomas, in dissent, described it as "speech directed at minors without their parents' consent".

are [b] no essential part of any exposition of ideas, and are of such slight social value as a step to truth that any benefit that may be derived from them is clearly outweighed by the social interest in order and morality" (315 U.S. at 571-572). According to the government, description [b] identified functional criteria for determining when utterances were not within the First Amendment's coverage.43 The Court disagreed, holding that description [a] set out historical criteria for making that determination.44

Here we face the two questions described above: is a rule appropriate, and if so, what should it be? The analytic structure is as before: identify the targets and ask (i) whether they are likely to do worse than the Court in identifying relevant categories, and (ii) whether, if so, they will do better using historical criteria or functional criteria to identify categories. The ultimate target in Stevens is the legislature, with lower courts as a more proximate target.

Blasi's analysis suggests that we should address the first question by asking whether there is some pathology associated with the legislature's identification of categories.45 The inquiry into pathologies formally has two steps. First we ask whether there is some pathology associated with the particular category the legislature has created: is there some reason to think that the legislature has overestimated the risks associated with the category of expression or underestimated its benefits? If so, we proceed to the question of devising a rule to guide the legislature's creation of categories. If not, we then ask whether there is some pathology associated with legislatures' assessment of the existence of 3 categories of expression in toto: is there some reason to think that the legislature is prone to overestimate risks or underestimate benefits of a system of free expression in which some categories of expression are excluded from the First Amendment's coverage (or, as commercial speech cases suggest, receive less First Amendment protection than other categories do)? If so, we ask whether the functional or the historical test is more likely to lead legislatures

43 For the distinction between coverage and protection, see Schauer 1981, 270-271.

44 Description [a], according to the Court, explained why the historical criteria made functional sense, but did not displace a historical inquiry with a functional one.

45 For completeness, I note that a complete formulation would deal with two forms, one in which the legislatures (pathologically) proscribe a category of speech and the other in which they identify a category of speech and delegate to juries the decision about penalizing speech within that category. As to the latter a pathology might affect the legislative decision to delegate or a different pathology might predictably occur when juries decide. The distinction corresponds roughly to the well-established distinction in First Amendment doctrine between Gitlow v. New York, 268 U.S. 652 (1925), exemplifying the first type, and Schenck v. United States, 249 U.S. 47 (1919), exemplifying the second. Yet a further complication is that Snyder v. Phelps shows that some analysis of judicial pathologies would be required where state courts develop the common law. I hope that my discussion in the text helps identify considerations that would be deployed in dealing with these complexities.

to develop categories that match those the Court would choose were it in a position to develop all the relevant categories.

According to Blasi, legislatures are prone to overestimate the risks associated with speech critical of government policy, particularly in heated times. What of the risks and benefits associated with crush videos? For myself, I find it difficult to identify a pathology that would lead to the overestimation of risks or underestimation of benefits associated with the dissemination of such videos.46

If that is right, we must then ask whether there is some pathology associated with creating a system of expression in which some utterances are covered and others are not.47 The best candidate I have been able to devise is this: understanding themselves to have the power to create categories of uncovered expression, legislators will be tempted to respond to constituents' concerns about forms of expression the constituents find particularly distasteful by seeking to regulate the utterances without seriously considering whether the harms caused by the utterances truly outweigh the social benefits of their dissemination. This could be understood as a generalization of the "interest-group capture" pathology: a system that allows legislatures to generate new categories of un- or less-protected expression will prompt interest groups to pressure legislatures to do so, and sometimes the legislatures will succumb to those pressures.48 For myself, I am again skeptical about this claim's merit. But, the Court's choice between the "[a] and [b]" options in Stevens is defensible if the Court accepted the claim, and for present purposes I do so as well.

As a first approximation, I suggest that legislatures authorized to use a functional approach will generate "better" categories—that is, those in which the aggregated harms outweigh the aggregated benefits, at least as the legislatures evaluate those harms and benefits—than will legislatures required to use a historical one. They are likely to have access to relevant functional information, which will come to them as they consider a wide range of policy questions,

46 A common claim is that interest-group capture produces a pathology relevant to First Amendment analysis. Interest-group capture, it is thought, produces legislation that overestimates the harm averted and the overall social benefit provided by the statute, in the service of distributing gains to the interest group. In Stevens an interest-group capture account of this sort does not seem especially strong. Certainly animal-welfare groups favored the statute, not to gain competitive advantage against someone, but rather to advance their vision of what the public interest demands. (The statute is Stevens was almost certainly overbroad under standard doctrine, but the interesting question is whether the legislature could define a category that included the crush videos that were its primary concern. I am quite confident that it could.)

47 Or, again, in which some categories receive more protection than others.

48 I put it in terms of "serious consideration" rather than "actually outweigh" because the argument's predicate is that the Court believes that the legislature will not do a good job (with "good" defined as above) in determining whether aggregated risks outweigh aggregated benefits.

whereas they are unlikely to deal with historical questions in any setting other than the one at issue.49 But, for the same reason they are likely to generate more categories using a functional than a historical approach. And, the more categories the more likely a mismatch between the legislatures' evaluation of the aggregated harms and benefits and the Court's. The Court's choice in Stevens would then be justified by a judgment that targets—legislatures, in the current example—are more likely to get it right from the Court's point of view if they are instructed that they can develop a rule-system with a few broad rules and only a few exceptions. The historical criterion [b] is on the surface an expression of the Court's commitment to originalism, but it operates more importantly to o induce the creation of a simple rule-set by reducing the number of legislatively I

created exceptions. To that extent, some preference for the specific form that e rule-ification has taken might be justified. r

2.3 Magnitude of Benefits and Sorrell v. IMS Health, Inc. /

2.3.1 Description 0

Sorrell v. IMS Health, Inc. held unconstitutional a Vermont statute barring the r sale of certain pharmacy records to data-mining companies that intended to U resell them to pharmaceutical manufacturers (131 S. Ct. 2653 (2011)). Vermont | requires pharmacists to keep records of the prescriptions written by physicians, ° classified by the doctors' names. It barred the pharmacists from selling lists 3 showing the number of prescriptions for each drug written by each physician.50 i That restriction was limited to sales of the lists for marketing uses, and did not preclude the pharmacists from selling the lists to those who intended to use them for research.51 Vermont believed that the ban would reduce medical costs because it would make it more difficult for manufacturers of expensive drugs, typically those protected by patents, to identify doctors who were prescribing similar but less expensive drugs. Drug manufacturers send "detailers" to visit doctors to describe the benefits of using the manufacturers' drugs.52 They explain why the substitutes the doctors are using might be less effective than the drugs their employer makes, for example. Vermont's legislators believed that this sales technique induced doctors to switch from less to more expensive but

49 Though the Court's adoption of the historical approach might induce interest groups and others to supply legislatures with the required historical information.

50 As compiled by the pharmacists the lists did not identify the patients for whom the prescriptions were written.

51 In dissent, Justice Breyer observed that the state's ethics rules for pharmacies define release of practitioner information "to unauthorized persons" as unprofessional conduct. 131 S. Ct. at 2680 (Breyer, J., dissenting).

52 For a journalistic account of this sales method, see Elliott (2010: ch. 3).

not more effective drugs, thereby increasing the cost of medical care. By making it more difficult for detailers to identify possible targets for their sales efforts, the ban on selling information for use by drug manufacturers would help control medical costs.

The Supreme Court held that Vermont's ban violated the First Amendment because it regulated speech—the dissemination of information lawfully acquired by pharmacists—in a way that discriminated on the basis of both content and speaker: the regulation "disfavors marketing, that is, speech with a particular content" and "disfavors specific speakers, namely pharmaceutical manufacturers" (131 S.Ct. at 2663). Even though the regulation dealt with O

commercial speech, and even on the assumption that the information was 1

simply a commodity, the regulation was subject to heightened scrutiny because e

it was not justified on some basis independent of the speech's content or the g

recipient's identity. The Court noted that "the outcome is the same whether a h

special commercial speech inquiry or a stricter form of judicial scrutiny is /

applied" (id. 2667). The state had "to show at least that the statute directly .

advances a substantial governmental interest and that the measure is drawn to O

achieve that interest" (id.). That standard "ensure[s] not only that the State's jo

interests are proportional to the resulting burdens placed on speech but also 3

that the law does not seek to suppress a disfavored message" (id. 2668). Because g

the regulation allowed researchers and journalists—an "almost limitless audi- 3

ence" (id.), in the Court's words—access to physician-related information, the |

regulation did little to protect physicians' privacy. Doctors who did not like to g

discuss their prescription practices with detailers could simply refuse to meet gg

with them. Nor was the connection between cost-control and the regulation g

close enough. Under the regulation costs are controlled because detailers cannot i

present doctors with accurate information that might lead them to choose to 3|

prescribe more expensive drugs, perhaps mistakenly. But, "[t]hose who seek to §

censor or burden free expression often assert that disfavored speech has adverse g

effects" (id. 2670). g

Justice Breyer in dissent offered a different characterization of the regulation. 5

Rather than a regulation of information lawfully acquired, Justice Breyer saw it as "inextricably related to a lawful governmental effort to regulate a commercial enterprise" (id. 2673 (Breyer, J., dissenting)). Much of his opinion was devoted to an explication of what he saw as the Court's traditional "test-related distinction^' between regulations of commercial and noncommercial speech (id. 2674 (Breyer, J., dissenting)). According to Justice Breyer, regulations of "expression inextricably related to" business regulations were almost inevitably content and speaker based. As to content, regulated businesses engage in many forms of communication—commercial advertisements, advertisements on public policy proposals of interest to the businesses, securities

disclosures—but regulators will select only some of those, on the basis of their content, for regulation.53 And as to speaker-based regulations, regulators might identify a specific subset of an industry where problems are especially severe and impose requirements on manufacturers in that subset without imposing similar regulations on others in the same industry.54

As Justice Breyer's opinion emphasized, Sorrell's most problematic aspect is its continuation of the Court's trend to assimilate commercial speech to political speech, subjecting both to demanding requirements for justification, or, in Justice Breyer's words, an "unforgiving brand of 'intermediate' scrutiny..." (id. 2679 (Breyer, J., dissenting)). As the Court observed, SorrelFs analysis follows the path marked by a series of cases in which the Court refused to distinguish sharply between commercial and other forms of protected speech, asserting that whether the stringent tests for regulating non-commercial speech or the supposedly less stringent tests for regulating commercial speech were applied, the result would be the same.55

Yet, at least on the surface the tests could produce different results. In Sorrell, for example, the Court made much of the fact that everyone but pharmaceutical manufacturers could use physician-identified prescription information. So, the Court said, the information was available to "an almost limitless audience". l Formally speaking, perhaps so. Listing all the possible users of such informa- gg tion—pharmaceutical manufacturers, university researchers, journalists, any ►

interested citizen—the statute does indeed exclude only one group from g.

using the information. One might think, though, that the relevant question 3 was not the length of the list of possible users—which depends on the criteria 0 for determining how to enumerate users—but the likelihood that any potential 3 user would be interested in identifying specific physicians. The pharmaceutical manufacturers of course were interested in doing so, and an occasional journalist might think it important to use a doctor's name to bring to life some story about prescription practices, but it is hardly clear that university researchers, or anyone else, would think it important to do so. So, in practical terms, the

53 See 131 S.Ct. at 2677 (Breyer, J., dissenting) ("Electricity regulators, for example, oversee company statements, pronouncements, and proposals, but only about electricity.").

54 See id. at 2677 (Breyer, J., dissenting) ("An energy regulator, for example, might require the manufacturers of home appliances to publicize ways to reduce energy consumption, while exempting producers of industrial equipment."); 2680 (quoting Burson v. Freeman, 504 U.S. 191, 207 (1992) (plurality opinion) ("The First Amendment does not require States to regulate for problems that do not exist.").

55 For earlier discussions, see Lorillard Tobacco Co. v. Reilly, 533 U.S. 525 (2001) (refraining from deciding whether the tests for commercial and non-speech were different because the regulations at issue did not satisfy the nominally lower standard for permissible regulations of commercial speech); Thompson v. Western States Medical Center, 535 U.S. 357 (2002) (providing a similar analysis).

regulation covered the vast bulk of those who might actually use the information.56 And, in doing so, it would do a reasonably good job—the kind of job the nominal test for commercial speech demands—of promoting the government's permissible regulatory goal of protecting physician privacy.

Justice Breyer's doctrinal alternative would preserve commercial speech as a separate category for constitutional analysis, by allowing regulation of speech inextricably bound up with lawfully regulated activities. This would make the doctrinal structure a bit more complicated, but hardly beyond the capacity of judges to apply.57 A constitutional rule dealing with speech in connection with lawfully regulated activities does not seem to me difficult for legislatures to o

apply, nor is it obviously subject to any more manipulation in the service of l

narrow interest groups than is substantive regulation. The only justification for e

the Court's assimilation of commercial to noncommercial speech seems to be a 0

preference for simplicity as such. h

2.3.2 Critique .

For the present I take the historical approach of Stevens as given. The conclusion 0

that categories of un- or less-protected speech should be identified through °

historical rather than functional inquiry leaves open the question of how to 0

define the boundaries of those categories the Court finds have adequate histor- o

ical support. And here, I believe, the preference for simplicity as such is more t

difficult to justify. The reason emerges when we consider the two possible i

targets for the Court's rules, legislatures and courts. g

As to legislatures, the historical inquiry, as we have seen, inevitably generates O

questions about the level of generality at which the historical inquiry should be g

conducted. There will inevitably be insufficient evidence to support a judgment i that a category of speech described at a sufficiently low level of generality— described in relatively specific terms such as "violent video games"—has historically been given no protection or less protection than other categories receive. The Court's forays into the "level of generality" problem have not yielded much in the way of clear doctrine.58

56 See 131 S. Ct. at 2680 (Breyer, J., dissenting) ("The record contains no evidence that prescriber--identifying information is widely disseminated."); 2681 ("noting in the record indicates that any 'counterdetailing'... has ever taken place in fact in Vermont.").

57 The business of the "media" consists of selling information, and some of that information might well be "in connection with lawfully regulated activities". The courts would then have to develop an exception to Justice Breyer's rule for the media, which would introduce another degree of complexity into the doctrinal structure.

58 One thing is clear: a purely historical inquiry cannot determine the level of generality at which the historically determined categories should be described. Rather, some analysis, typically predicated on a broader view of the proper scope of judicial review, is required. Consider here Justice Scalia's

59 Described earlier as the libertarian position.

60 The definition offered in early commercial speech cases.

61 Justice Breyer's definition.

The historical inquiry might generate relatively few categories of un- or less-protected speech, categories described in relatively general terms. Generality means that defining the precise boundaries of the categories will be the Court's—and the lower courts'—task. Once again, we have to ask whether the Court will be in a position to police directly what the lower courts do (probably not) and what rule or rules it could develop to ensure that the lower courts' actions conform to the Court's judgment about the best set of outcomes overall.

In Sorrell the candidate definitions are (i) no categorical exemption for commercial speech,59 (ii) "speech proposing a commercial transaction",60 and (iii) o "speech inextricably related to a lawful governmental effort to regulate a com- l

61 ... °

mercial enterprise". I put aside the first position simply because it is incon- e

sistent with existing doctrine, including the doctrine at least nominally followed 0

in Sorrell. On the face of things, I find it hard to distinguish between the h

remaining possibilities, except perhaps on the ground that the traditional def- /

inition is a bit simpler for lower courts to administer, though the difference in .

simplicity seems to me small. The question then becomes whether the move 0

from a simple boundary-definition to a slightly more complex one is likely to o

introduce more errors (from the Court's point of view) than improvements in ° the set of overall outcomes. That question, in turn, requires that we ask whether r the latter, slightly more complex definition would open the way for some path- ° ology to operate at the legislative level. The thought would be that the "inex- n tricably incident" test would license a bad interest group dynamic.

Precisely what that dynamic would be is unclear. The legislature defended the statute in Sorrell as a cost-containment technique aimed at practices by a perhaps unpopular group, "Big Pharma", and perhaps to the advantage of manufacturers of generic versions of out-of-patent prescription drugs. But, the

analysis elsewhere. In Michael H. v. Gerald D., Justice Scalia, joined only by Chief Justice Rehnquist, argued that in determining the level of generality at which traditional protections of family autonomy should be described, wrote, "We refer to the most specific level at which a relevant tradition protecting, or denying protection to, the asserted right can be identified." 492 U.S. 110, 127 n. 6 (1989). The reason for selecting that level of generality is that it appropriately accommodates the constitutional interest in protecting liberty to the states' general regulatory power. Any higher level of generality would, on this view, authorize the courts to invalidate "too many" legislative innovations, thereby intruding on valuable experimentation, a potential engine of progress or at least learning through self-government. I leave for discussion elsewhere whether or how this approach could be used to deal with the problem ofrule-ification, noting only that it would counsel in favor of defining categories in ways that licensed legislative experimentation and the like.

interest-group story about Sorrell seems to me the kind of "just-so" story that people make up when forced to fit a statute enacted in pursuit of a vision, contested to be sure, of the public interest, into a theory that calls for interest-group explanations of all statutes.62 Perhaps in other regulatory contexts, the "inextricably incident" test would lead to an obvious overestimation of benefits (or underestimation of costs) associated with interest-group capture. But, my view is that the burden should have been on the Court to explain what that dynamic would be before it rejected Justice Breyer's test. Without such an explanation, we are left with a preference for simplicity as such.

2.3.3 Conclusion

In each of these cases the Court reached counter-intuitive results from the perspective of almost any foundational theory of the First Amendment. I have treated these cases as "symptoms" of a judicial pathology. The symptoms result, I suggest, from the form that rule-ification has taken in First Amendment doctrine. The judicial and scholarly debate about the relative merits of rules and standards is well-known and of long-standing. Rules provide clear guidance to people charting their course of action, but they necessarily fail to take into account normatively relevant features of actual cases. They are, in the jargon, inevitably over and underinclusive with respect to their purposes. Standards are better tailored to the circumstances, but for that reason are worse at providing ex ante guidance, because actors cannot know what "the circumstances" will be and, more important, what courts assessing their actions ex post will find to the "the circumstances".

There is a long-standing scholarly tradition arguing that, as cases accumulate, courts are driven to move from standards to rules. Experience accumulates, and judges get familiar with some generic features of situations they repeatedly confront. They become increasingly confident that they can devise rules that capture so many of the normatively relevant features of those situations that the benefits of ex ante guidance from rules come to outweigh the benefits of applying standards after the event.

Free speech doctrine in the United States has reached the point where rules have generally replaced standards.63 The pathology is not rule-ification in itself. Rather, it is the decision to have an extremely simple rule-system. So, for

62 An alternative interest-group story pervades discusses of health care. In that story "Bog Pharma" is the concentrated interest group that is able to "distort" legislative deliberations through campaign contributions and the like. Those who offer the counter-story about Sorrell might be asked for state-specific reasons explaining Big Pharma was unable to capture the Vermont legislature.

63 I note that Fredrick Schauer has argued that the European preference for standards over rules,

through the proportionality doctrine, arises from the relatively brief experience those courts have

example, in the funeral protest case, the Court held that the stringent standard for allowing states to impose liability for intentionally inflicting emotional distress on public figures should be used in cases involving the intentional infliction of emotional distress on ordinary, private people where the means of inflicting the injury is through the "discussion" of public matters such as the proper social response to homosexuality. And, in the crush video case, the Court expressly limited the categories of speech outside the scope of the First Amendment to a short list of categories historically recognized, not functionally defined.

Yet, the rules/exceptions/exceptions approach is actually quite standard.64 In another recently decided case, involving a regulation of legislative ethics, the Court had no difficulty in finding that the act of casting a vote was not an exercise of free speech—that is, was categorically excluded from the First Amendment's coverage (Nevada Comm'n on Ethics v. Carrigan, 131 S.Ct. 2343 (2011)). This is the rule/exception model in action. Adding an exception to the exception is hardly beyond the capacity of judges either, as the leading U.S. decision on hate speech regulation shows.65

So, the pathology is the use of a simple rule-set rather than a rule-set that includes more elements. What accounts for that choice? The arguments roughly parallel those in the rules-standards debate, though on a higher level. The question is, Should a court take a rule-like approach to developing the rules that regulate primary behavior, or should it take a standard-like approach? The i U.S. Supreme Court has pretty clearly chosen a rule-like approach to developing rules. In the seminal gun-rights/gun-control case the Court relied on ori-ginalism—a rule-like method, in my terms—and derided Justice Breyer's use of a standard-like method. It similarly criticized the government's attempt to defend the crush video statute through what it variously described as a categorical balancing or a simple balancing approach.

had with free speech problems, and has predicted that European jurisprudence will move in a more rule-oriented direction over the next decades. Schauer 2003.

64 The Court's opinion in Stevens struggled with the rules/exceptions/exceptions problem in its unsuccessful effort to distinguish uncovered crush videos from covered child pornography by arguing that, though child pornography (not fitting the Court's definition of obscenity) did not fit within a historically defined category excepted from First Amendment coverage, coverage was justified because the production of child pornography was inextricable from child abuse, unlawful everywhere in the country. The same could be said about crush videos and animal abuse.

65 In R.A.V. v. City of St. Paul, 505 U.S. 377 (1992), an ordinance was held to violate the First Amendment because it made people liable for fighting words on the basis of the particular viewpoint expressed through those words. The analytic structure is—Rule ("no viewpoint discrimination")/ Exception ("fighting words are outside the First Amendment's coverage")/Exception to exception ("but not when the regulation is viewpoint-based").


The Court has not given adequate reasons for preferring a simple rule-structure

to a slightly more complex one. In other settings the Court sometimes knowledges that it is dealing with a relatively new phenomenon, as in a recent decision involving how the prohibition of unreasonable searches applied to a public employer's examination of an employee's text messages (City of Ontario v. Quon, 130 S.Ct. 2619 (2010)). When it does so, it generally says that it should be cautious and apply either a narrow rule, a rule with many qualifications and therefore possible exceptions, or a standard.67

My sense is that the pathological cases I have described are "new" in the relevant sense, although I confess that I have not figured out exactly why. But, if that sense is right, the prescription is simple. The Court should de-rule-ify the area, or, more precisely, accept an analytic structure with a reasonably thick set of rules. Full de-rule-ification is straightforward. It would be the direct application of a proportionality analysis, either at the level of the case itself (ad hoc balancing, in the U.S. terminology) or at the level of devising the appropriate rule (categorical balancing).

Thickening the rule-set is a bit more complicated, and probably cannot be described in general terms but only with reference to specific problems. So, in the crush video case, the Court should not have resisted creating a new category of uncovered speech, but should have adopted a functional analysis when presented with arguments in favor of doing so. And, in the funeral protest case, the Court should have acknowledged that the interests at stake when someone intentionally inflicts emotional distress on a private person, even through discussion of a public matter, are different from those at stake when the target is a public figure.68

The obvious prescription for pathological rule-ification is the reintroduction of some degree of complexity into the applicable doctrine. Here I sketch three possibilities, all inspired by constitutional concepts more common in Europe

66 In the U.S. literature, there is something of a confusion between ad hoc balancing, which ordinarily is the simple invocation of a standard to evaluate the constitutionality of a regulation of primary behavior, and categorical balancing, which operates on the next higher level where rules to regulate primary behavior are defined.

67 For recent examples, see Caperton v. A.T. Massey Coal Co., 556 U.S. 868 (2009); Missouri v. Frye, 132 S. Ct. 1399 (2012).

68 If I am right in describing these cases as presenting new problems, we can expect that rule-ification will set in as experience accumulates. But, I think, we have benefited from the experience of working out what the right rules are, rather than simply assuming that the rules developed to deal—probably quite adequately—for familiar problems are necessarily suitable for these new ones.

than in the United States: setting some other constitutional right against the First Amendment, using proportionality on a case-by-case basis, and using proportionality at the level of case types or classifications of utterances.

(A) The first possibility builds on established doctrine where vindicating the First Amendment threatens to infringe another constitutional right. The standard example is the "free press/fair trial" conflict, in which the dissemination of information otherwise protected against regulation raises a high risk that a defendant's right to a fair trial will be infringed, for example by adverse publicity that cannot be cured through careful screening of jurors and the like.69 The doctrinal solution is balancing: after extended consideration of alternatives o to preventing the dissemination of information, the trial judge must balance the o risk to the defendant's right to a fair trial against the public interest in acquiring e information.70 r

European constitutionalists would readily recognize a competing constitu- h

tional right in the right to human dignity.71 They would see it obviously impli- /

cated in Snyder, and the doctrine of human dignity would treat it as implicated 0

in the self-degradation of crush-video purchasers in Stevens.72 But, the U.S. r

Constitution does not directly recognize human dignity as a right,73 although 0 one can find suggestions that the concept of human dignity ought to inform the

69 See, e.g., Nebraska Press Ass'n v. Stuart, 427 U.S. 539 (1976).

70 It is probably worth emphasizing that the conflict here is between two constitutional rights, not between a constitutional right and some interests protected to some extent by another constitutional right. The standard U.S. example is hate speech regulation, conceptualized as a conflict between a First Amendment right and interests in equality that are not themselves constitutionally guaranteed era in the context of, for example, private harassing speech. Where rights conflict with (mere) interests, the rights necessarily prevail. Were the government to be under a constitutional duty to promote equality, a question that depends on one's understanding of the state action doctrine, we would have a right-versus-right conflict to be resolved by balancing. For a good explication, see Michelman 1989.

71 European constitutionalists have a more robust concept of informational privacy than that used in the United States, though I am uncertain whether even a robust concept of informational privacy would protect the information in Sorrell. The candidates for privacy protection are patients and their prescribing physicians. The patients would clearly have a decent informational-privacy claim, but the information in Sorrell was anonymized as to them. I am not sure whether the physicians would have a robust informational-privacy claim as to the number of prescriptions they wrote for specific drugs, given their ability to decline to meet with the drug companies' detailers.

72 For some relevant cases, see Laskey v. United Kingdom, 24 Eur. Ct. H.R. 39 (1997) (finding no human rights violation in a prosecution for consensual sado-masochism); Views of the Human Rights Committee, Communication No. 854/1999 (submitted by Manuel Wackenheim) (finding no human rights violation in the prohibition of "dwarf tossing", where the person being tossed consents to and earns money from participating in the activity).

For discussions, see Rao 2011; Jackson 2004.

specification of the content of other constitutional rights.74 Were the Court willing to move even a slight distance from strong rule-ification toward a marginally more complex rule-set, the idea of human dignity might support different results in Stevens and Snyder.

(B) A second possibility is the application of a proportionality test on a case-by-case basis. Passages in Snyder suggest that the fact that the outrageous comments dealt with a matter of public concern were not entirely dispositive.75 Here the only observation worth making is that the spectrum of doctrinal possibilities stretches from rule-ification to case-specific balancing, and that even weak instrumental considerations might counsel against case-specific balancing.

(C) Finally, there is the possibility flagged by the functional approach offered by the United States in Stevens—proportionality at the level of types of cases. The Court's response to that possibility was plainly inadequate, criticizing case-specific balancing and applying those criticisms to case-type balancing without recognizing the difference between the techniques.76 On a generous reading, the Court's concern was not with case-type balancing itself, but with the identification of case-types as to which proportionality analysis would apply. That is, it was concerned that a functional test focusing on the contribution a class of utterances makes to discovering truth (appropriately defined) would lend itself to misapplication by other public officials. Perhaps so, but I have argued that the Court failed to make the case for its approach. i


Rule-ification does not mean that the Court's analyses in the cases discussed era

here were doctrinally innovative or extreme applications of existing doctrine. Nor does it mean that the outcomes the Court reached are indefensible.77

74 For an example, see Graham v. Florida, 130 S.Ct. 2011 (2010) (holding unconstitutional the imposition of a sentence of life imprisonment without the possibility of parole on an offender who committed a non-homicide offense when a juvenile).

75 See Snyder, 131 S. Ct. at 1220 ("the reach of our opinion here is limited by the particular facts before us."). See also id. at 1222 (Breyer, J., concurring) ("To uphold the application of state law in these circumstances would punish Westboro for seeking to communicate its views on matters of public concern without proportionately advancing the State's interest in protecting its citizens against severe emotional harm.").

76 Although the government's position could hardly have been more clear.

77 For example, the statute in Stevens was overbroad, though whether it was overbroad enough to violate the Court's overbreadth standard is a more difficult question. If it is overbroad, Stevens's conviction should indeed have been vacated. And, perhaps courts should refine the "outrageous-ness" instruction given the jury in Snyder v. Phelps to reduce the risk of de facto viewpoint

A rule-ified constitutional doctrine, by its nature, is unresponsive to the empirical fact that the risks speech creates vary along several dimensions. Such unresponsiveness might be justified on the kinds of institutional grounds I have explored in this article.

It may be worth ending on a skeptical note. Justice Hugo Black was committed to the most simple rule system possible: "No law means no law." The current Court is committed to a somewhat thicker rule system. Between the Court's system and case-specific all-things-considered judgments lie many other rule systems, some slightly thicker than the Court's, some a great deal thicker. I have suggested that the Court could have achieved more socially desirable results in Snyder, Stevens, and Sorrell with a rule system that was only modestly thicker than the one the Court adopted. I believe that the Court's position on First Amendment issues raises the question: why a quite thin rule system rather than an only modestly thicker one? Answering that question will, I think, require connecting First Amendment theory to general constitutional theory in what I believe would be productive ways.78


Blasi, Vincent. 1985. The Pathological Perspective and the First Amendment.

Columbia L. Rev. 85, 449-514. Elliott, Carl 2010. White Coat, Black Hat: Adventures on the Dark Side of

Medicine. Boston: Beacon Press. Greene, Jamal. 2009. Selling Originalism. Georgetown L. J. 97, 657-722. Gunther, Gerald. 1975. Learned Hand and the Origins of Modern First Amendment Doctrine: Some Fragments of History. Stanford L. Rev. 27, 719-774.

Jackson, Vicki C. 2004. Constitutional Dialogue and Human Dignity: States and

Transnational Constitutional Discourse. Montana L. Rev. 65, 15-40. Llewellyn, Karl. 1960. The Common Law Tradition: Deciding Appeals. Boston: Little, Brown.

Michelman, Frank. 1989. Conceptions of Democracy in American Constitutional Argument: The Case of Pornography Regulation. Tennessee L. Rev. 56, 291-320.

discrimination. If so, the judgment against the Westboro church members should indeed have been vacated.

78 For example, it would place Blasi's questions about identifying legislative (and judicial) pathologies at the core of First Amendment theory generally, and not only in connection with speech critical of the government.

Rao, Neomi. 2011. Three Concepts of Human Dignity in Constitutional Law.

Notre Dame L. Rev. 86, 183-272. Schauer, Frederick. 1981. Categories and the First Amendment: A Play in Three Acts. Vanderbilt L. Rev. 34, 265-308.

- 2003. The Convergence of Rules and Standards. New Zealand L. Rev.

2003, 303-328.

- 2012. Harms and the First Amendment. 2011 Supreme Court Rev.,


Sullivan, Kathleen. 1992. Foreword: The Justices of Rules and Standards.

Harvard Law Rev. 106, 22-123. Tushnet, Mark. 2011. 'Telling Me Lies': The Constitutionality of Regulating False Statements of Fact. Harvard Public Law Working Paper No. 11-02. Available at SSRN: or http://dx.doi. org/10.2139/ssrn.1737930.

-2009. Heller and the Critique of Judgment. 2008 Supreme Court Rev. 61

- 1997. Self-Formalism, Precedent, and the Rule of Law. Notre Dame L.

Rev. 72, 1583-1596. Zipursky, Benjamin C. 2011. Snyder v. Phelps, Outrageousness, and the Open Texture of Tort Law. DePaul L. Rev. 60, 473-520.