The Ninth Circuit’s amended opinion in Garcia v. Google is now out, and the majority is sticking to its guns in granting Garcia preliminary injunctive relief against Google for its alleged copyright infringement. The facts are somewhat strange: Garcia had a minor part in a film, and she claims that she, and not the filmmaker, holds a protectable copyright right in her performance. Once the film appeared on YouTube, Garcia sent Google eight different DMCA takedown notices to have it removed. After Google failed to act on her requests, Garcia filed suit in the district court seeking injunctive relief against Google.
The district court denied that relief, finding that Garcia was unlikely to succeed on the merits of her copyright claim against Google. Garcia then turned to the Ninth Circuit, and in a divided opinion, that court has reversed. In the majority’s opinion, Garcia is likely to succeed on the merits of her copyright claim. In fact, the majority finds that all four preliminary injunction factors favor Garcia, and as such, it leaves in place its order from this past February: “Google, Inc. shall take down all copies of [the film] from YouTube.com and from any other platforms under Google’s control, and take all reasonable steps to prevent further uploads of [the film] to those platforms.”
The case has drawn so much attention that the Ninth Circuit created a webpage devoted to it. The appellate court even made a public request for amicus briefs, which it has added to that page. One of the arguments being made by Google and its amici is that the Ninth Circuit’s injunctive order against Google constitutes a prior restraint in violation of the First Amendment. The Ninth Circuit, in its amended opinion, brushes off this claim with little explanation. In this post, I’ll explain why I think the Ninth Circuit gets it right on the prior restraint issue.
Prior restraint is one of the more confusing—and confused—First Amendment doctrines. The origin of the doctrine is simple enough: Under English licensing schemes, publication of printed materials required the prior approval of a censor. Such schemes eventually fell out of favor, and the press became free to publish without first obtaining permission. As Blackstone noted in his Commentaries:
The liberty of the press is indeed essential to the nature of a free state; but this consists in laying no previous restraints upon publication, and not in freedom from censure for criminal matter when published. Every freeman has an undoubted right to lay what sentiments he pleases before the public; to forbid this is to destroy the freedom of the press: but if he publishes what is improper, mischievous or illegal, he must take the consequence of his own temerity.
Thus, even though there was no prior restraint on what the press could publish, the press could nevertheless face subsequent punishment if it published something illegal. This represents the classic dichotomy between prior restraint and subsequent punishment, and it rests on the principle that it’s better to punish people after they’ve spoken than it is to prevent them from speaking in the first place. The notion that freedom of the press equates to freedom from prior restraint was baked right into the First Amendment. That’s what freedom of the press means.
The prior restraint doctrine has evolved quite a bit since Blackstone’s time, and the distinction between prior restraint and subsequent punishment has been rendered all but unrecognizable. Nevertheless, the doctrine still plays an important role in First Amendment analysis. True to its roots, the prior restraint doctrine is concerned with censorship of protected speech. It is not a substantive doctrine for determining whether certain speech is protected by the First Amendment; it instead focuses on the adequacy of any procedures used to restrain speech—even speech that may ultimately be unprotected. The doctrine is invoked in several different contexts, but the common denominator is that prior restraint ferrets out and condemns the unbridled discretion of a censor.
For example, the classic prior restraint is a licensing scheme, where the censor decides whether certain materials can be published in the first place. The Supreme Court has endorsed such a scheme, despite acknowledging that “the censor’s business is to censor,” so long as “it takes place under procedural safeguards designed to obviate the dangers of a censorship system.” The constitutionality of the scheme turns on whether the safeguards employed adequately constrain the censor’s discretion to restrain speech, and the Court finds it important that a restrained speaker has access to swift recourse in the courts so that a neutral party can double-check the censor’s work for error.
The prior restraint doctrine is also invoked in the context of a pretrial seizure of expressive materials. As with a licensing scheme, the issue here is whether the censor’s discretion is adequately constrained by the procedural safeguards employed. When the line between protected and unprotected speech is difficult to discern, such as with obscenity, the Supreme Court permits a pretrial seizure so long as “rigorous procedural safeguards,” such as a prior adversarial hearing, are employed before the seizure takes place. When circumstances dictate, the unbridled discretion of the agent-in-the-field is replaced by the neutral decisionmaking authority of the court, and the censor’s work is again double-checked by a neutral party.
Prior restraint also comes up quite often in the context of an injunction, as we have here in Garcia v. Google. With an injunction, the court itself is the censor, enforcing its speech restraint via its power of contempt. As with the prior restraint doctrine generally, the issue is whether the procedures employed adequately protect against the discretion of the censor. But the fact that the censor is the court changes things a bit. With a licensing scheme or pretrial seizure, the concern is that an agent of the executive branch will have too much discretion in separating protected from unprotected speech. We turn to a court to double-check the censor’s work because we trust its discretion, circumscribed, as it is, both substantively and procedurally.
The concern with an injunction, the Supreme Court has explained, is this:
The special vice of a prior restraint is that communication will be suppressed, either directly or by inducing excessive caution in the speaker, before an adequate determination that it is unprotected by the First Amendment.
Thus, with an injunction, the question becomes whether the court has made an “adequate determination” that the speech at issue is unprotected. Once this has happened, the speech can be enjoined without violating the First Amendment as an unconstitutional prior restraint. But how does this standard apply in a copyright case? As Professors Mark Lemley and Eugene Volokh have noted, when it comes to copyright claims, “preliminary injunctions are granted pretty much as a matter of course.” I think this makes sense, because, at least with wholesale copying, there is very little doubt that the defendant’s speech is unprotected by the First Amendment. To see why this is so requires a bit of unpacking.
When speech infringes a copyright right, that speech is unprotected by the First Amendment. Despite forbidding laws “abridging the freedom of speech,” the First Amendment is not abridged by the copyright laws that grant authors exclusive rights to their works. The Supreme Court has held that copyright rights and the First Amendment are balanced at the definitional level through two doctrines internal to copyright law, namely, the idea/expression dichotomy and fair use. I wrote about this substantive First Amendment analysis of copyright rights in two prior posts (here and here). The idea is simply that, as a matter of law, these two doctrines provide sufficient breathing space for the competing First Amendment values.
When the speech at issue is wholesale copying—by which I mean copying that would be “consumptive infringement,” as Terry Hart has called it, of entire copyrighted works—the court can very easily rule out the two doctrines that internally balance copyright rights with the First Amendment. This is so because a defendant that has engaged in such wholesale copying will usually have no colorable defense based on the idea/expression dichotomy or fair use doctrines. The defendant might have another defense that excuses the copying, such as a license, but then it’s the license that protects the copying—not the First Amendment. If a court enjoins speech that ultimately proves protected only by a license, then it cannot be a prior restraint since the doctrine applies only to constitutionally-protected speech. The licensed speech is simply not protected by the First Amendment—that’s why the licensee needs the license in the first place.
I think the hardest thing for people to keep separated is the fact that the First Amendment interests at issue in a copyright case depend on who the speaker is. For example, if I write a book about copyright law, my speech is fully protected by the First Amendment. If you come along and distribute copies of my book over the internet without my permission, your wholesale-copying-as-speech doesn’t get my full First Amendment protections. The problem for you is that it’s my speech that’s at issue—speech in which I hold a property right to exclude you from speaking. To be sure, your wholesale copying of my speech is to some extent also your speech, but your speech has to stand on its own two feet. And unless you can demonstrate that your wholesale copying of my speech is excused under the idea/expression dichotomy or fair use doctrines, your speech is not protected by the First Amendment.
Thus, I think the prior restraint doctrine is rather simple to apply when it comes to wholesale copying: Unless the defendant has some colorable defense under the idea/expression dichotomy or fair use doctrines, the court can enjoin the allegedly-infringing speech without the risk of restraining constitutionally-protected speech. That speech may turn out to be noninfringing (such as when the defendant can prove a license), but without either of these two First Amendment-grounded defenses, the speech isn’t constitutionally protected. The prior restraint doctrine is concerned with restraining speech that may ultimately turn out to be protected by the First Amendment, not with speech that is constitutionally unprotected yet potentially noninfringing.
But what about when it’s not wholesale copying? What if, to borrow Terry Hart’s terms again, the alleged infringement is “creative,” not “consumptive”? Or, put another way, what if the defendant does have a colorable defense based on the idea/expression dichotomy or fair use doctrines? Here, the situation is obviously different, and it’s possible that constitutionally-protected speech will be restrained. Yet, unlike with a licensing scheme or pretrial seizure, we’re not concerned about agents of the executive branch having too much discretion in separating protected from unprotected speech. The censor is the court, the neutral decisionmaker we entrust to make these difficult decisions.
When the issue is wholesale copying, the court can rule out prior restraint by simply looking to whether the defendant has raised either of the two First Amendment-grounded defenses. The court would then apply that initial merits determination to the rest of the four-factor preliminary injunction test, perhaps granting the injunction if the balance of equities tip in that direction. The same is true when it’s creative copying, only now there’s a chance the court will find that the defendant is likely to succeed on the merits of a First Amendment-grounded defense. Enjoining speech that is likely to be protected by the First Amendment is a prior restraint, and unless some exception based on the balance of equities applies, the injunction is unconstitutional. On the other hand, if the court finds that the defendant isn’t likely to succeed on a First Amendment-grounded defense, then it’s not a prior restraint since no constitutionally-protected speech is likely to be enjoined.
Take, for example, the Suntrust case, where the Eleventh Circuit vacated a preliminary injunction issued by the district court against an alleged infringer who argued a colorable fair use defense. The appellate court found that the alleged infringer was likely to succeed on the merits of her fair use defense, and as such, “the issuance of the injunction was at odds with the shared principles of the First Amendment and the copyright law, acting as a prior restraint on speech because the public had not had access to [the alleged infringer’s] ideas or viewpoint in the form of expression that she chose.” Having found that the use was likely fair, and thus protected by the First Amendment, the Eleventh Circuit struck down the preliminary injunction as a prior restraint.
So where does that leave us with Garcia v. Google? The prior restraint argument has been briefed at length by both of the parties and several amici. Google vigorously argues that enjoining it from providing the film on its platforms violates its own First Amendment rights. Think about this for a minute. Normally, Google would not be arguing that the user-generated content available on its platforms is its own speech. It would instead point to Section 230 or the DMCA to disavow its role as the speaker. But I think it’s true that Google is the speaker when it publicly performs works uploaded to its platforms by its users, so I’m glad to see it acknowledge its role as the speaker of the content here.
In its first motion for a stay, Google argues that the Ninth Circuit’s injunction “infringes on Google’s First Amendment rights because it prevents Google from exhibiting the Film on its websites.” This, Google argues, “is a clear infringement on Google’s First Amendment expression, given that the Supreme Court ‘has long held’ that movies ‘are protected expression.’” Google goes ever farther in its second motion for a stay, arguing that the risk of restraining constitutionally-protected speech here means that “only a particularly strong showing of likely success” can justify the injunction. While the Ninth Circuit did find Garcia likely to succeed on the merits, it nevertheless admitted that the point was “fairly debatable,” and Google claims this measurement of likely success is insufficient under Ninth Circuit prior restraint precedent.
Google makes the same arguments in its supplemental brief and in its petition for rehearing en banc. But all of Google’s hand-waving about its First Amendment rights misses the point. There is no doubt that the film is protected speech, and there is no doubt that enjoining Google will restrain speech that is protected. But that doesn’t mean it’s a prior restraint. Since this is a copyright claim, we have to keep track of whose speech it is—and it isn’t Google’s. While the film is undoubtedly protected speech, those full First Amendment protections belong to the copyright owner, not Google. Whether the owner is Garcia or the filmmaker doesn’t matter, because either way the question to be decided is whether the First Amendment protects Google’s copying of someone else’s copyrighted speech.
Google is arguing the wrong analytical track when it claims that the Ninth Circuit should have found a “strong showing of likely success” instead of the likely-yet-fairly-debatable amount of success it ultimately settled upon. That distinction would make a difference if Garcia were seeking an injunction against the filmmaker, since there the issue of the identity of the true author—i.e., the party with the full First Amendment protections—would be germane. If Garcia weren’t likely to succeed on the merits against the filmmaker, then enjoining the filmmaker’s speech, which would then be likely protected by the First Amendment, would be a prior restraint. But that’s not the case here, and there is no doubt that Google is not the author. Google’s prior restraint argument instead is properly analyzed as with any other instance of wholesale copying: The speech can only be protected by the First Amendment if Google has a colorable defense based on the idea/expression dichotomy or fair use doctrines.
Google, quite naturally, offers no such defense for its wholesale copying, and this has not gone unnoticed by the Ninth Circuit. The appellate court’s entire rebuttal to Google’s prior restraint argument is this:
The problem with Google’s position is that it rests entirely on the assertion that Garcia’s proposed injunction is an unconstitutional prior restraint of speech. But the First Amendment doesn’t protect copyright infringement. Cf. Eldred v. Ashcroft, 537 U.S. 186, 219–220 (2003). “First Amendment protections are ‘embodied in the Copyright Act’s distinction between copyrightable expression and uncopyrightable facts and ideas,’ and in the ‘latitude for scholarship and comment’ safeguarded by the fair use defense.” Golan v. Holder, 132 S. Ct. 873, 890 (2012) (quoting Harper & Row Publishers, Inc. v. Nation Enters., 471 U.S. 539, 560 (1985)). Google hasn’t raised fair use as a defense in this appeal, see page 11 supra, so we do not consider it in determining its likelihood of success. This does not, of course, preclude Google from raising the point in the district court, provided it properly preserved the defense in its pleadings.
Thus, the Ninth Circuit comes to the same conclusion—albeit in a much more direct fashion—as I do: Since the issue is Google’s use of someone else’s copyrighted speech, Google’s own speech is not protected by the First Amendment unless it can claim a defense based on the idea/expression dichotomy or fair use doctrines. And since Google has never argued that either of these doctrines excuses its copying, the court can enjoin Google’s speech without the risk of restraining constitutionally-protected speech. By failing to establish that it has even the possibility of First Amendment protection for its wholesale copying of someone else’s copyrighted speech, Google has guaranteed that the injunction against it cannot possibly be a prior restraint. Of course, nothing is stopping Google from spreading whatever messages it pleases, and the marketplace of ideas likewise continues to thrive none the lesser.
Follow me on Twitter: @devlinhartline
© 2014 Devlin Hartline. Licensed under the Law Theories Public License 1.0.