Federal Circuit Reversed Fair Use in Java-Android Case But Strengthened Fair Use Overall

A Riddle We Didn’t Know We’d Been Asked Is Answered.

It was legitimately shocking when the Federal Circuit reversed a jury verdict that Google’s use of Java to make the Android operating system was a fair use. Not because the Federal Circuit disagreed with the trial court about fair use—reasonable minds can almost always disagree about that—but because it was a jury verdict.

An android celebrates its jury-trial victory with a steaming cup of Java, moments before tragedy ensued.

Juries Are Scary, Part 2

Jury verdicts are almost never reversed. To reverse one, an appellate court must rule that there was no way for the jury to reach its verdict, given the evidence available to it. If there was some way to view the evidence to reach the verdict, appellate courts will not disturb the jury verdict, no matter how unlikely it seems, or how much the appellate court might disagree with the jury’s conclusions. If the verdict, for example, required the jury to disbelieve five witnesses but believe one, well, then, it stands.

Juries are also black boxes. Evidence and a description (often confusing of the relevant law goes in one end, and a verdict pops out the other. Most verdicts just reach the ultimate conclusions, sometimes with a dollar figure attached. Q. Did Robin Thicke infringe the copyright in “Got To Give it Up?” A. Yes. Q. If your answer was yes, what are the damages? A. $4,000,000.00.¹The actual verdict form in the “Blurred Lines” case is here. It’s a bit more detailed, but not by much.

Now think about fair use. Think about how there are four factors, some of which have sub-factors, and how you’re supposed to balance them in some undisclosed way. It should be impossible to reverse a jury verdict on fair use (either way). Even if there were just one factor favoring (or disfavoring) fair use, you can’t say there was no evidence for the verdict.

But you almost never see jury verdicts on fair use in the wild. Instead, the question is almost always handled by a judge on a pre-trial motion. Normally, judges are reluctant to rule on these pre-trial motions in a way that takes the case away from the jury, but when fair use is the issue, they lack such reluctance. Which is the opposite of what you’d expect.

Paradox: Classic Jury Question That Never Goes to the Jury

For years, we’ve hand-waved this paradox: fair use looks like a classic jury question, yet it hardly ever reaches a jury. We’ve overlooked the paradox because generally we were pleased judges were keeping the issue away from juries. Fair use, the thinking goes, is fundamental to copyright, and it’s a right we all have. It often involves unpopular speech. It often involves people without the resources to assert fair use. Best to keep fair use far away from juries, and best to get a ruling on fair use as early in the lawsuit as possible.

Consider: Campbell v. Acuff-Rose, the case involving 2 Live Crew’s impolite parody of Roy Orbison’s “Oh Pretty Woman,” was first brought in Nashville in 1990. Do you think a Nashville jury was going to be terribly sympathetic to a bunch of outsiders mocking their beloved Roy? Or, if you were representing an artist whose works baked, rotisseried and threatened to blend nude Barbie dolls, how confident would you feel about a jury on fair use? If cases like these always went to juries, we’d be too afraid of the expense of litigation to exercise our right to fair use.

How do you think a jury full of folks who grew up with Barbie dolls would feel about this art?

In this piece of her performance art, Barbie demonstrates the perils of fair use.

And, yet: questions of fair use can and do go to juries. We always assumed that was technically correct, but judges were bending the rules because of the importance of the defense.

Strangely, until now, no appellate court has definitively weighed in on the issue: is fair use a question for the jury or for the judge?

Now, appellate courts, and the Supreme Court, have answered a closely related question, albeit in an unhelpful way. Is the question of fair use one of fact (and thus for the jury) or one of law (and thus for the judge)? Guess what the answer is. It’s a “mixed question of fact of law.” Big help.

To its credit, the Federal Circuit decided that it needed to figure this out once and for all. Perhaps it had no choice. The easy way out would have been to declare, “Welp, a jury did it!” and affirm the verdict, even while tut-tutting it, a lá the “Blurred Lines” decision.²Thus, when I predicted that the Federal Circuit would uphold the jury verdict, I was rather assuming that there would be a jury verdict to uphold. “It’s not how we would have ruled, but we have to respect the jury.”³And then not bother to really analyze why it was a jury verdict, so it’s all kind of dicta. But if the Federal Circuit felt Google’s use of Java SE to make Android shouldn’t be fair use, it needed to do away with the “jury” part of the “jury verdict.”

Facts and Law Are Totally, Clearly Different. Except for All the Times They’re Not.

Generally speaking, we rely on juries to find facts and judges to determine the law. Thus, we say that questions of fact go to the jury and questions of law go to the judge. Thus, the questions, “What color was the traffic light?” and “Did the driver exercise reasonable care in making the left-hand turn?” are jury questions, but the test for whether a car accident was the result of actionable negligence is for the judge.

On appeal, the difference between questions of fact and of law is crucial. Determinations of fact are accorded a great deal of deference by the appellate court, while questions of law are accorded none at all.

But it’s not that easy. For example, procedural questions, which can be crucial to a case’s outcome, are considered questions of law even though they depend on facts. Thus, whether a police search of your home violated your Constitutional rights is a question of law, even though its answer depends on a lot of facts and circumstances: was there a warrant? If not, why were the police there? Was there a risk that evidence would be destroyed if the police didn’t act right away?

To take another example: what a written contract means is usually a question of law, unless the contract is ambiguous. If it’s ambiguous, then the court must look at a lot of outside evidence—about the negotiations leading up to the contract, for example—and the question of interpretation becomes one of fact.

So, a question of fact isn’t one that happens to involve the determination of a fact. But it’s one that where a jury would be helpful and practical to help determine. Whether the search of your home was Constitutional turns on facts, but they’re facts that are pretty easy to determine and aren’t all that contentious. Further, we need to know whether the search was Constitutional before we go to trial, because otherwise the jury might be exposed to things that should have been excluded. Similarly, if the contract can be interpreted easily, then we’d just as soon leave the jury out of that process. But once different sides are allowed to dispute about “what happened,” we need a jury to make that determination.

To make matters more complex, sometimes a question is a “mixed question of fact and law.” And the Supreme Court has told us that fair use is a “mixed question.” But that doesn’t tell us whether it should go to the jury or to the judge, and, thus, whether the Federal Circuit must accord the jury verdict a lot of deference or none at all.

A Stroke of Luck in Bankruptcy

As luck would have it, the Supreme Court had just addressed the issue of how to treat mixed questions of fact and law—on March 5, less than three weeks before the Federal Circuit handed down its decision.⁴It was Google that made the Federal Circuit aware of this decision. I’ve read this case, U.S. Bank v. Village at Lakeridge, so you don’t have to. It’s a bankruptcy case, and the question was whether one of the creditors was an “insider,” because insiders get treated differently from (I guess) outsiders. Funny thing about bankruptcy court is that there is never a jury. But it still has that distinction between questions of fact and questions of law. The difference is that the judge determines both, but the judge’s determination of facts is accorded deference—but not as much as a jury’s—while determinations of law continue to be accorded no deference at all.

Luckily for us, whether someone is an “insider” is one of those mixed questions. The big question in U.S. Bank was whether the putative insider was too cozy with a real insider and thus was effectively also an insider. The accusation was a bit salacious: that the putative insider and actual insider were romantically involved. The Supreme Court noted that the test for whether someone is an insider is purely a question of law, and the test is whether the putative insider’s transactions with the debtor were at arms’ length. The Court also noted that whether those transactions really were at arms’ length depended on “historical facts,” by which it means not facts you find in history books, but facts that are defined by a particular time and place: a particular somebody did something at a particular time at a particular place.⁵We also ask juries to determine whether historical actions were reasonable or not. I’m not sure whether that’s a “historical fact” or not. In U.S. Bank, the three most important historical facts were that the real insider did not cohabit with the putative insider, that she did not pay the putative insider’s bills, and that the putative insider acquired the debt in question as a speculative investment, not to help out his friend. Are those historical facts enough to make the putative insider’s acquisition of the debt something less than arms’ length? What counts as arms’ length sounds like a question of law, but it’s bound up with those historical facts. That’s why it’s a “mixed question of fact and law.”

Whether a Mixed Question More Law-y or More Fact-y Depends on What You Think Juries Are For.

Appellate courts need to know how much deference to give the trial court. Appellate courts give findings of fact a lot of deference because they weren’t there when the evidence was being adduced. It didn’t observe the witness’ demeanor; it wasn’t putting the story together in real time. But if the judge is just announcing the law, then the appellate court gives the judge zero deference. The law is the law, and it should be the same everywhere.

That last point informs the Supreme Court’s analysis of how much deference appellate courts give trial courts on mixed questions: a lot⁶Bearing in mind that juries get more deference than judges. or none at all. The law implies generalities; facts imply specificity. Is the question something that is likely to keep coming up over and over? Or is it something more unique? In U.S. Bank, the Supreme Court held that the mixed question was “about as factual sounding as any mixed question gets.” Change a small factual detail, and you might get a very different result. What if the real insider was cohabiting with the putative insider and they were sharing expenses? If you ran this film over and over but with different bankruptcies, you’d get different fact patterns that lead to different results.

Returning to Oracle v. Google, the Federal Circuit held that fair use is more like a question of law than a question of fact. It didn’t present any analysis on this point, unfortunately, relying instead on legal authority purporting to hold that fair use is a question of fact. But we can come up with a pretty good analysis on our own. Fact patterns for fair use tend to repeat. And those who are thinking about using someone else’s copyrighted work without permission would be well-served to know whether their planned use is like uses that have been judicially determined to be fair use, or not. Concepts like “transformativeness,” “expressiveness” and “substantiality” lean toward “law.” And, also, since fair use is somehow bound up with the First Amendment right of free expression, including the right to express unpopular opinions, maybe we don’t want juries as guardians of the right to fair use.

You Thought Jury Duty Was Bad? Try Being on an Advisory Jury

Before you start to feel sad for juries, they still have a role to play. The jury can still find those historical facts—if the judge wants. This is known as an advisory jury.⁷These are still pretty common in those rare jurisdictions, such as—ahem—Tennessee, that still have courts of equity. Advisory juries can’t make the sort of ultimate legal conclusions that regular juries can. All they can do is make specific factual findings. This means you can’t just give them a general verdict form. Instead, you have to present the jury with specific questions about the facts, which the jury answers after hearing the evidence. Ideally, the questions will be specific and admit yes-no or short answers (e.g., “Who was driving the vehicle at the time of the accident?”, “What color was Mr. Rogers’ car?”).

Even when an advisory jury is used, the judge is entitled to ignore some or all of the advisory jury’s findings. In the end, the findings of fact are the judge’s. The jury is only there to “advise.”

Alas, that’s not what happened at trial! All the court got was a general verdict: Google’s use of Java SE to develop and use Android was a fair use. They didn’t get into any of the specific factors, not that would have been good enough. Asking, for example, “Was Google’s use of Java SE transformative?” wouldn’t be good enough because that’s a conclusion, not a historical fact. You’d have to break that question down in to its constituent facts.

Was the Federal Circuit a Little Over-eager? And Wasn’t the Case Tried by Jury by Consent?

Instead of sending the case back to the trial court to do fair use all over again, the Federal Circuit assumed that the jury resolved all historical facts in favor of fair use. At least, that’s what the Federal Circuit said. It’s fairly obvious that it didn’t, or else how could it have reversed the jury? Anyway, the Federal Circuit wasn’t required to accept any of the advisory jury’s findings—no more than the trial judge had to. Technically, though, the Federal Circuit should have sent the case back to the trial judge so the trial judge could make his own findings, in light of the jury’s implied findings. Again, with an advisory jury, the findings are still ultimately the judge’s. That way, the Federal Circuit would have some actual findings of fact with which to work and wouldn’t need to guess.

There’s a fly in this clouded ointment. The parties are allowed to have a real jury trial on issues that aren’t supposed to be tried by juries, so long as all the parties and the judge consent.⁸Don’t believe me? Read Fed. R. Civ. P. 39(c)(2). The consent doesn’t have to be express. If you don’t object, you’ll be deemed to have consented. This isn’t just some advisory jury, either. It’s a real jury.⁹So says 8 Moore’s Federal Practice § 39.15. You can give it a general verdict form. And the Federal Circuit would have had to give the jury verdict the deference juries normally get.

The Federal Circuit’s opinion makes it fairly clear that the parties and the judge consented to a jury trial on the issue of fair use: “Despite this case law, all aspects of Google’s fair use defense went to the jury with neither party arguing that it should not.” Oracle’s opening brief indicated that the parties agreed to a jury trial on fair use.¹⁰It was supposed to be fair use and damages, but the judge decided to separate out damages. If Oracle objected to the bifurcation, was that enough to withdraw its consent to the jury trial? The judge, in addressing the jury’s verdict, certainly treated it as though it were a regular jury!

Obviously, I’m missing something. Neither the Federal Circuit nor the parties have suggested the jury trial was by consent, which to my mind would utterly change the result. Jury trials by consent are unusual but not secret. No one commenting on this decision—and there are lots—has brought it up. Google didn’t bring it up in its appellate briefing; it just assumed there had been a regular jury trial.

The Takeaway: Fair Use Is Not a Jury Question, and it’s Stronger For It.

But, let’s not lose sight of the important thing here. For the first time, an appellate court has ruled that fair use is a question for the judge, not the jury. That means it can be ruled on early in the case, sometimes right at the beginning of the case. It means that fair use is probably strengthened overall, even though fair use was not found in this instance. Judges are probably more likely to find fair use than juries, especially for unpopular uses, which are most vulnerable.

Why do I think this, when statistically we have no evidence one way or the other (because there hardly have been any fair use jury determinations)? For one thing, I think we can all agree that, for better or for worse, fair use has grown broader over time—and judges, not juries, have been driving that. Maybe juries would have done the same, but I doubt it. Less certain is the sense that juries tend to favor rights holders. There have been studies showing this to be the case with patent holders¹¹But not as much as they like plaintiffs, as I recall. In declaratory judgment cases, juries still tended to side with the plaintiff, but not nearly as much as when the plaintiff was also the patent holder., at least, and I think juries would treat copyright holders similarly.

Perhaps as important, knowing that the question of fair use will be resolved early in litigation—that there is no threat that a copyright defendant will have to survive all the way to trial to have this crucial question decided—will encourage those who intend to rely on fair use. And the bolder—or better resourced—of them will actively seek to push the boundaries.

At the same time, this isn’t a seismic shift, either. For practical purposes, fair use had been a question for judges. It was really rare (but not unheard of) for fair use to be presented to a jury. And even those lawyers, scholars and industry professionals who thought fair use had gotten too big for his breeches haven’t been clamoring for juries to pay a bigger role in fair use.¹²Indeed, the trend in scholarship has been rather the opposite, i.e., that juries strengthen fair use and judges threaten it. That doesn’t make intuitive sense to me.

Also, other appellate courts, even the Ninth Circuit, are free to ignore the Federal Circuit’s holding. Although the Federal Circuit was trying to apply Ninth Circuit law, its treatment of Ninth Circuit law isn’t binding on the Ninth Circuit, to say nothing of the other circuits. Still, this decision is highly persuasive, and, right now, it’s the only one we’ve got, so I suspect it’ll carry a great deal of weight, on this one question.

Next time, I’ll get into the weeds about the Federal Circuit’s analysis of the fair use issue.

Thanks for reading!

Rick Sanders

Rick is currently General Counsel for Software Freedom Conservancy. Previously, he has been practicing law as an intellectual-property litigator since 2000.