Concept
Why Prior Restraints Are Disfavored
Prior restraint
Why Prior Restraints Are Disfavored
A prior restraint is an administrative system or a judicial order that stops speech before it occurs,1 and Near v. Minnesota made that timing the center of the doctrine. Under Near, prior restraints are generally disfavored, the government carries a heavy burden to justify one, and only narrow exceptions, such as a ban on publishing troop movements in wartime, survive that burden. The Court has repeated the point in harsher terms. It calls prior restraints the most serious and least tolerable infringement on First Amendment rights, and it says that any system of them comes to the Court bearing “a heavy presumption against its constitutional validity.”2 The earlier sections taught how to spot a prior restraint by the moment the government acts. This section asks why a rule about timing carries so much weight.
The question is sharper than it first appears. A large enough penalty after the fact can silence a speaker as completely as an order, and a prior restraint can look careful by comparison, because it targets particular speech and usually follows some judicial or administrative hearing.3 Some scholars press the point and doubt that the Court’s sharp distinction between the two kinds of remedy can be fully justified. What follows is the case for the distinction as its supporters make it, and each argument must explain why a hearing and a narrow target do not cure the problem.
The case rests on four arguments, and each names a different mechanism. History supplies the first: the First Amendment arose against a licensing regime. The second compares what an order does to speech with what a threat of punishment does. The third asks who decides, and on what basis, when the speech has not yet been heard. The fourth concerns how a restraint is enforced once it is entered. Together they answer Justice Butler’s suggestion in Near that an order issued by a court differs from a censor’s veto, and they give an exam answer the vocabulary to explain why a particular restraint deserves heightened suspicion. Behind all four stands the Court’s theory that a free society prefers to punish the few who abuse speech after they break the law than to throttle them and everyone else in advance.4
Historical Roots
The doctrine begins with English licensing. Until 1694, England allowed no publication without a government-granted license, so an author needed official approval before a single copy could be printed.5 The defect lay in where the system placed the decision. A licenser acted before the public could read the work, and acted alone. Blackstone described the evil in those terms: to subject the press to a licenser is to subject all freedom of sentiment to the prejudices of one man and to make him “the arbitrary and infallible judge of all controverted points in learning, religion, and government.”6 It is widely accepted that the First Amendment was meant, at the very least, to abolish such prior restraints on publication.7
Blackstone defined the liberty of the press by the same target: it “consists in laying no previous restraints upon publications, and not in freedom from censure for criminal matter when published.”8 The American Court drew the same lesson when it struck down a city ordinance that required the City Manager’s written permission before anyone distributed literature. In Lovell v. Griffin, a Jehovah’s Witness handed out religious tracts without applying for a permit, and the Court held the ordinance invalid on its face because it “strikes at the very foundation of the freedom of the press by subjecting it to license and censorship.”9 The struggle for the freedom of the press, the Court observed, “was primarily directed against the power of the licensor,” and the liberty of the press “became initially a right to publish ‘without a license what formerly could be published only with one.’”10 Because the ordinance was void on its face, she did not have to seek a permit before contesting it, and she was entitled to contest its validity in answer to the charge against her.11 The section on licensing schemes returns to that consequence.
England paired the ban on previous restraints with another tool. Punishment after publication fell outside the ban, and seditious libel made criticism of the government a crime. Chief Justice Holt explained the need in 1704: “If people should not be called to account for possessing the people with an ill opinion of the government, no government can subsist.” Truth was no defense, because the goal was to punish all criticism of the government, and true speech seemed worse because it might do more to damage the government’s image.12
Americans abandoned seditious libel in the eighteenth century. In Zenger’s Case (1735), Andrew Hamilton argued that the jury should decide whether the defendant had spoken the truth about the government. The judge told the jury that truth was no defense, yet the jury acquitted, and the verdict is generally seen as the beginning of the end of seditious libel prosecutions in the United States.13 Even the Sedition Act of 1798 declared that true criticism of the government was constitutionally protected.14 By the time of ratification, commentators agreed that freedom of the press at a minimum meant that speech could be punished only after publication.15
The words “at a minimum” mark a floor, and the guarantee later rose above it. Lovell said that freedom from previous restraint “cannot be regarded as exhausting the guaranty of liberty,” although preventing that restraint was a leading purpose of the First Amendment’s adoption.16 It is no longer true that the First Amendment means only freedom from prior restraint, but prior restraint remains a more serious restriction on speech than subsequent punishment, and freedom of speech and of the press “has meant, principally although not exclusively, immunity from previous restraints or censorship.”17
The result is a two-part inheritance that an exam answer should keep apart. At its core the doctrine is the historical objection to licensing. Any modern order that works like a licensing board, whether it comes from an agency or a court, revives the old fear that one authority will decide in advance what the public may read. A city that demands a permit before leaflets may circulate sits at that core, and so does a judge who enjoins an unpublished article. A statute that punishes a publication after it appears is not a prior restraint, but that fact does not end the inquiry, because the guarantee reaches subsequent punishment as well. Concluding that a measure is valid because it operates afterward repeats the old view that the amendment guards only against previous restraints. The modern doctrine applies the guarantee to both kinds of measure and applies it with particular vigor to prior restraints.18
Freezing Versus Chilling
The functional case against prior restraints starts from a comparison the Court drew in Nebraska Press Association v. Stuart. A state court barred members of the press from publishing confessions by an accused in a murder trial, along with other facts “strongly implicative” of the accused. The order was a judicial command forbidding communication in advance, and the press itself challenged it, so the case put the First Amendment’s hostility toward prior restraints directly in issue.20 The Court unanimously held the order invalid.21
The Court explained why such orders draw its sharpest scrutiny with a single contrast. A prior restraint “has an immediate and irreversible sanction,” and if a threat of criminal or civil sanctions after publication “chills” speech, a prior restraint “freezes” it at least for the time.22 That contrast is the working explanation for the heavy presumption against validity that attaches to any system of prior restraints.23
Chilling and freezing place the sanction at different points in time. A speaker who faces only the threat of later punishment still decides whether to speak, and a speaker who goes ahead puts the message before its audience at once. Cautious speakers hold back under that threat, which is the chill, yet the speech that is made reaches the public. An advance order reverses the sequence. Prior restraints prevent speech from ever occurring,24 so the speaker’s only road to the audience runs through getting the order lifted, and the message stays out of circulation while that effort proceeds. The harm is irreversible in a practical sense: a speaker who wins on appeal regains the right to speak but cannot regain the moment at which the speech would have been heard.
The damage is greatest when the restraint falls on news and commentary about current events.25 News is perishable. A report on a council vote, a pending trial, or an official’s conduct is worth most while readers can still act on it, and a delay of weeks may leave little to report. A restraint reversed on appeal therefore returns the speaker’s right without returning the audience, since protected speech may be delayed until it has lost its relevance.26 Truthful reports of public judicial proceedings already receive special protection even against subsequent punishment, so an order that bars them in advance reaches speech the Court guards most closely.27
Both measures burden speech, and the Court ranks them by when the burden arrives. After-the-fact liability pressures the decision to speak, while an advance order removes the decision. That ranking explains why the doctrine concentrates its suspicion on timing, and an essay applying it should name the effect on the facts by asking whether the order keeps the message out of circulation now or whether the government responds only once the message is out.
Because the freeze begins at once, the Court has shaped the procedures that may accompany a restraint so that they keep the freeze short and put a judge in charge. Where prior restraints are permissible at all, the Court has insisted on careful procedural provisions, and it finds no place for an ex parte order against speech unless it is impossible to notify the opposing parties.28 Any restraint imposed before a final judicial determination on the merits must be limited to preserving the status quo for the shortest fixed period compatible with sound judicial resolution of the issue.29 The burden of proving that expression is unprotected rests on the censor, and the censor’s determination may not carry an effect of finality.30 The section on procedural safeguards develops these requirements. Each is a device for shortening the freeze.
The freeze argument is the Court’s most quoted justification, and it meets serious objections. An injunction gives the speaker something a prosecution withholds. A person who faces only a criminal statute cannot know beforehand whether a judge will find the planned speech illegal, while a person facing a request for an injunction is heard before the judge acts, so the injunction is arguably the more speaker-friendly choice.31 The Court has itself said that the label “prior restraint” does not settle a case. In Kingsley Books, Inc. v. Brown, a 5-4 decision, it upheld a New York procedure authorizing an injunction against the sale of allegedly obscene printed matter pending an expedited trial, and Justice Frankfurter wrote that “prior restraint” is “not a self-wielding sword.” A further objection holds that the doctrine fixes on timing when the real danger lies in who restrains and with how much discretion. Freedman v. Maryland drew that line: a censor, whose business is to censor, may well be less responsive than a court to protected interests, and only a judicial determination in an adversary proceeding ensures the necessary sensitivity to freedom of expression.32 On that view, a judge who applies clear standards after a full hearing raises few of the concerns that make an administrative censor dangerous.33
Those objections leave the freeze standing for as long as an order is in force, but they show why defenders of the doctrine do not rest on it alone. The advance ruling that an injunction offers is a thin advantage, because a speaker who wants a determination of legality before speaking can bring an action for a declaratory judgment.34 A legislature that wants more deterrence can obtain it more directly by increasing the penalty for criminal violations.35 Otherwise the main difference between a prior restraint and subsequent punishment is that an injunction deters more effectively, so there is little to be said affirmatively for injunctions against speech and at least some risk to free speech if they are allowed.36 The next two subsections develop the forecasting problem and the collateral bar rule, which carry the argument past what the freeze contrast can bear alone.
Speculation and Institutional Bias
Every request to restrain speech asks a decision-maker to predict what a message that nobody has yet heard will do once it is heard. The previous subsection showed what an advance order costs the speaker, a freeze that lasts as long as the order does. This subsection explains why the person who decides whether to impose that cost is likely to err in one direction, and the explanation has three parts: the forecasting problem together with the lopsided cost of forecasting wrongly, the bias of censors, and the severity with which courts treat disobedience of their own orders. These points also help explain why the government carries a heavy burden and must generally show a special societal harm, such as a serious threat to national security, before a prior restraint can stand.
The forecasting problem comes first. A court that punishes speech after publication can look at what the speech actually did, such as whether a reputation was damaged, a secret exposed, or a person hurt. A court asked to restrain the same speech in advance has only a prediction to examine. Before speech takes place, any forecast regarding its effects contains an element of speculation, and that can lead a judge to misjudge the constitutional issue and to overuse injunctions “just to be on the safe side.”38 The same tendency appears elsewhere in the law. It is easier to get a judge to authorize seizure of potentially dangerous medications than to bring a criminal prosecution afterward against the manufacturer, because judges are reluctant to make mistakes that could cause a public health crisis.39
Nebraska Press shows the difficulty in a setting with no national security stakes. The trial judge found only “a clear and present danger that pre-trial publicity could impinge upon the defendant’s right to a fair trial.”40 The Supreme Court said that his conclusion about the effect of publicity on prospective jurors was “of necessity speculative, dealing as he was with factors unknown and unknowable.”41
A forecast would be tolerable if mistakes in either direction cost the decision-maker the same, but they do not, and that imbalance is the best explanation for the presumption against restraint. A judge who lets the story run shares the blame if damage follows, because the refusal then looks like the cause.42 A judge who enjoins the story faces no comparable exposure. The story is never printed, or the copies are seized, so the public never learns whether the predicted harm would have come, and the court loses no esteem when it over-predicts the danger.43 Over-prediction is safe and invisible while under-prediction is risky and public, so forecasts drift toward restraint, and the benefits of the speech are under-predicted along with the danger. The drift operates on a conscientious judge, because it works through which mistake can be seen. The public bears the cost, since it never gets to judge for itself the importance of the speech that was suppressed.44
National security cases show the asymmetry at full strength, because the government’s showing is itself a forecast. In New York Times Co. v. United States, the Pentagon Papers case taught in full later in this chapter, the government asked a court to bar two newspapers from publishing a classified study of Vietnam policy on the ground that publication could harm the national interest. A request of that kind gives the judge nothing to examine except an executive’s prediction, and the asymmetry pushes the judge toward accepting it. Once the paper has published, the court’s position changes, because it can examine what actually happened.
Administrative censors bring a different bias to the same task. Freedman observed that because the censor’s business is to censor, there inheres the danger that the censor will be less responsive than a court to the constitutionally protected interests in free expression.45 If seeking judicial review is made unduly onerous, by delay or otherwise, the censor’s determination may in practice be final.46 These institutional points weigh most heavily against administrative licensing. A judge who decides after an adversary hearing raises fewer of them, since only a judicial determination in an adversary proceeding ensures the necessary sensitivity to freedom of expression, so the case against judicial injunctions rests mainly on the forecasting asymmetry and on the contempt power.47
Contempt supplies the last piece. Judges tend to be severe with criminal contempt because a violation breaks the law like any other crime and also undermines the stature of the judiciary.48 A prosecutor who has obtained an injunction has also pre-committed to pursuing violations, while one facing an ordinary complaint might or might not bring a charge.49 The injunction can also short-circuit what would otherwise be a right to a jury determination of whether a work is obscene, defamatory, or otherwise unprotected. The judge decides that issue in granting the injunction, and any jury at a criminal contempt trial decides only whether the defendant violated the order.50 The collateral bar rule, taken up next, adds to this enforcement power.
An essay asking why a judge’s order against speech deserves suspicion should lead with the over-prediction asymmetry, because it identifies a mechanism that works on a fair-minded judge, while a charge of general bias would need proof. The censor’s professional stake and the severity of contempt then serve as supporting points for licensing systems and injunctions.
The Collateral Bar Rule
Under the collateral bar rule, a person who violates an injunction or other prior restraint on speech is ordinarily subject to punishment even if the speech was constitutionally protected, because it is usually no defense that the restraint was unconstitutional.52 The rule governs the judicial form of prior restraint, an order that directs a person not to engage in expression on pain of punishment,53 and it is perhaps the most persuasive reason to distrust such orders, because it changes what happens to a speaker who defies a restraint that turns out to be invalid. A statute that merely criminalizes certain speech may be violated by a speaker who is confident that the statute is unconstitutional, and in the prosecution the First Amendment may be raised as a complete defense.54 The injunction offers no such route. The only way to challenge it is ordinarily to obey it while contesting its validity in court, which can postpone the speech for a long time.55 In jurisdictions that follow the rule, this is a clear-cut reason for disfavoring injunctions against speech.56 A judicial restraint can be more dangerous to the speaker than a criminal ban on the very same words.
The leading illustration is Walker v. City of Birmingham, in which civil rights marchers in Birmingham defied a state court injunction. By a vote of 5-4, the Court held that the marchers could not defend against contempt charges by asserting that the ordinance or the injunction was unconstitutional.57 Justice Stewart, writing for the Court, declined to hold that the petitioners were “constitutionally free to ignore all the procedures of the law and carry their battle to the streets.”58 The Alabama courts had justifiably relied on the general rule that court orders must be obeyed until “reversed for error by orderly review.”59 Justice Brennan, joined by Chief Justice Warren and Justices Douglas and Fortas, dissented on the ground that the Court had elevated a “rule of judicial administration above the right of free expression.”60
Justice Stewart also wrote Shuttlesworth v. City of Birmingham, which grew out of the same demonstration but involved the permit ordinance itself rather than an injunction.61 The ordinance was the administrative form of prior restraint, a permit requirement for expression, and it forbade a permit for a protest march if “the public welfare, peace, safety, health, decency, good order, morals or convenience” required that it be refused.62 The Court reversed the convictions of marchers who violated it, because it was unconstitutionally vague, and Justice Stewart said that a person faced with such an unconstitutional licensing law “may ignore it and engage with impunity in the exercise of the right of free expression for which the law purports to require a license.”63
The pair is the centerpiece of the doctrine, because one permit scheme produced opposite results depending on its form. Marchers who violated the scheme as an ordinance went free, and marchers who violated it after a court enforced it by injunction were punished.64 Walker did not call the injunction free of constitutional problems. The Court said that the breadth and vagueness of the injunction would unquestionably raise substantial constitutional questions, but that the way to raise them was to apply to the Alabama courts to have the injunction modified or dissolved.65 What separated the cases was the form of the command and the route for contesting it. An ordinance addressed to the public is tested in the prosecution that enforces it, and violating a statute is often the best or only way to attack its constitutionality.66 An injunction addressed to particular litigants has a forum built in, the court that issued it, and those litigants are expected to use that forum before they speak.67
The rule rests on a judgment about where disputes over an order’s validity belong. A person bound by an order can ask the issuing court to modify or dissolve it, so a neutral determination is available before anyone speaks.68 If each bound party could decide alone whether an order was valid and defy it on that judgment, the order would restrain only those who agreed with it, and every dispute would be fought in a contempt trial instead of in the review process built for it. A second reason concerns the judge. A defendant who violates an order has defied the very judge who issued it, and the judge may read the violation as an attack on the courts, so contempt draws more adamant enforcement than violation of a criminal statute, where the judge has no personal stake.69
The price of the rule falls on the speaker whose position is correct. A bound speaker who believes the order invalid must wait for relief, so the order freezes speech for at least the duration of the appellate proceedings, and even constitutionally protected speech may be delayed until it has lost its relevance.70 The rule therefore has qualifications, and each marks a place where one of the reasons above is missing: the order is so plainly defective that obedience serves no purpose, no prompt route to review existed, or no court has spoken at all.
An order that is transparently invalid may fall outside the rule, because Walker suggested that the case might have come out differently on that showing.71 The exception is narrow. The marchers in Walker lost even though the permit ordinance behind the injunction was later held unconstitutionally vague, so a defect that becomes clear only after a later decision does not make an order transparently invalid. The qualification is a safety valve for orders that no reasonable judge could defend, and it does not reopen every contempt case.
The rule also presupposes that the speaker could obtain prompt review of the order. If the state fails to provide strict procedural safeguards, including “immediate appellate review,” the injunction does not bind in the same way.72 The logic is that obedience is demanded in exchange for a fair chance at a prompt ruling. A different conclusion follows where no reasonable review procedure exists, but so long as a judicial order is outstanding, the legal process cannot tolerate a choice to disregard it and then attack it collaterally.73 The Court also limited Walker by regulating how speech-restrictive injunctions issue. In Carroll v. President & Commissioners of Princess Anne, the petitioners challenged an injunction against a rally in court rather than ignoring it as the Walker marchers did, and Justice Fortas found no adequate justification for the ex parte nature of the proceedings.74 Where prior restraints are permissible at all, the Court said, it insists on careful procedural provisions, and ex parte orders against speech have no place unless it is impossible to notify the opposing parties.75 Walker requires speakers to use the courts, and Carroll requires that the process they must use give them a hearing worth having. The section on procedural limits on injunctions develops Carroll in full.
A licensing law invalid on its face can be ignored, which is the lesson of Shuttlesworth, and Lovell had said as much: a speaker facing a void licensing ordinance did not need to seek a permit and could contest its validity in answer to the charge.76 The collateral bar rule has little to say here, because it concerns disobedience of court orders and no judge has issued a command, so the court’s own stake in its authority, one reason for the rule, is absent when an administrative scheme is the only restraint.77 The section on unbridled discretion explains how a court recognizes such a law.
A licensing law valid on its face must be followed, which is the lesson of Poulos v. New Hampshire. A protestor who disobeys a facially valid licensing system, an administrative requirement of permission before expression,78 must seek judicial review of the administrative action before taking to the streets.79 The difference between Poulos and Shuttlesworth is the facial validity of the law: a speaker could ignore the facially invalid scheme in Shuttlesworth, but had to challenge the facially valid one in Poulos through the channels available. A speaker who skips the procedures of a sound licensing system, or who fails to challenge a denial through them, cannot defend a prosecution by attacking the denial.80
Step 3 draws the most errors, because students tend to apply the bar to every licensing law. The bar applies only to a licensing law that is valid on its face, and a facially invalid one is treated like an unconstitutional statute. An exam question that gives the same permit language once as an ordinance and once as an injunction is testing steps 2 and 4 against each other.