Tag: standard of review

Google v. Oracle XII : SCOTUS Should Not Remand on Fair Use Review

As mentioned in my last Google v. Oracle post, the Supreme Court devoted considerable attention during oral argument to addressing the standard applied by the Federal Circuit when it determined as a matter of law that Google’s copying of Oracle’s code was not fair use. Google maintains that the Federal Circuit failed to show proper deference to a jury decision, and I discussed the constitutional and historical aspects of this argument in an earlier post. Oracle’s response argues that the Federal Circuit performed essentially the same function on appeal that courts perform all the time at summary judgment, where the vast majority of fair use cases are decided.

The reason why so much fair use jurisprudence occurs at summary judgment is simply a practical reality of litigation. When a party moves for summary judgment—for instance, a defendant who presents a fair use defense to an infringement claim—the court must assume that the facts favor the non-moving party and then consider whether there is any genuine dispute as to a material fact and whether the moving party is entitled to judgment as a matter of law. Once a court rules as to whether (or not) a fair use defense will survive, that decision is so important to the legal positions of the two sides it is often the end of the case.

Google asserts that the Federal Circuit erred when it decided the fair use question. Indeed, several amici in support of Google argued that overturning the jury in this instance was a misstep under the rules of civil procedure and a violation of the Seventh Amendment. Counsel for Oracle, Joshua Rosenkranz, noted at oral argument that Google went so far as to assert that “only a jury can balance the [fair use] factors.” Assuming Oracle’s characterization of Google’s argument is accurate, Google is simply incorrect. Obviously, fair use is a matter that may be considered exclusively by the courts, since most of the time, it is. As the Federal Circuit opinion in this case states:

“The fair use question entails … a primarily legal exercise. It requires a court to assess the inferences to be drawn from the historical facts found in light of the legal standards outlined in the statute and relevant case law and to determine what conclusion those inferences dictate….the exercise of assessing whether a use is fair in one case will help guide resolution in all future cases.”

The Supreme Court may consider whether remanding to the Federal Circuit for review “under a more deferential standard [of review],” as Justice Gorsuch put it, would have a deleterious effect on future summary judgment proceedings. And I may be over-speculating here, but remand seems unlikely. If the Court finds that the Federal Circuit correctly applied its fair use analysis and solely weighed questions of law, then any lingering matters regarding deference—if they are found to exist at all—do not seem compelling enough to delay the Court from ruling on both the review standard and the fair use finding. After all, this case is already more than ten years old.

No Reasonable Jury…

According to Oracle, the Federal Circuit correctly applied the “no reasonable jury” standard in deciding fair use as a matter of law. Does this mean the jury was unreasonable? No. It means fair use, to put it in technical terms, is a pain in the butt. Professional creators struggle with it; attorneys struggle with it; and courts struggle with it. And the trickiest aspect of fair use is acquiring a sensitivity to the interdependence of the four factors, which, unsurprisingly, a jury is unlikely to possess.  

What the Federal Circuit determined was that the jury resolved disputes of historical fact in favor of the verdict (and this was reviewed with deference), but that the jury arrived at incorrect legal inferences drawn from those facts (and this was reviewed de novo). Thus, the appellate court found that, even deferring to the fact finding, Google’s copying was not fair use.

I suspect the Supreme Court will not find that the Federal Circuit applied the wrong standard, unless there is some arcane bit of civ pro flaw in the mix that eludes me and the litigants. The appellate court followed Ninth Circuit precedent (because that is where the parties would have filed had there not originally been a patent claim in the suit), and it seems to have very clearly articulated how it separated questions of fact from questions of law before proceeding with an in-depth fair use analysis. And the review considered the same questions of law judges consider on summary judgment, where fair use is so often decided. For these reasons, I imagine the Supreme Court will not remand but will instead rule on the fair use analysis itself and either agree or disagree with the Federal Circuit’s findings.  

Separating the Factual from the Legal in Fair Use

Under factor one (the purpose and character of the use), the Federal Circuit correctly separates the factual question of a use with a commercial purpose from the legal implication of a use with a commercial purpose. The opinion notes that Google’s commercial intent here is undisputed (how could it be when Android is worth billions?) and that the jury was instructed that commerciality generally weighs against fair use as a matter of law. The court then correctly identifies transformativeness as a legal question and spends nearly six pages explaining why Google’s use of the Java code superseded the purpose of Oracle’s original work and was, therefore, “not transformative as a matter of law.”

Under factor two (the nature of the work), the analysis is generally dominated by separating factual works, which are more often subject to fair use, from expressive works, which are less often subject to fair use. Acknowledging that computer code can be complicated because it is an expressive work of a highly functional nature, the Federal Circuit sensibly held that the jury could have decided that the functionality favored fair use under factor two, but that the factor itself did not weigh heavily overall. This opinion should be little surprise to Google and several of its amici, as they have made precisely the same argument in cases involving the use of expressive books and photographs.

The more that a single factor teeters on the centerline, the more determinative the other factors become. But to reiterate the broader point, there appears to be nothing novel about the Federal Circuit’s approach to separating the factual from the legal and opining solely on the legal under factor two.

Factor three (the amount and substantiality of the portion used) tricks a lot of people. A user might copy a fraction of a work and be justly found guilty of infringement because the third factor is a qualitative analysis that asks whether a user has copied the “heart” of the work. Alternatively, a user might copy a whole work and have the use fall within the fair use exception.  This is not to say that the fair use analysis is capricious—rather, this is one case where we really see the interdependence of the four factors.

Google argued that it only copied the bare minimum of what it needed to achieve its purpose, and it further adds that it used only a fraction of all Java code. But this ignores the Federal Circuit’s finding that the parties did not dispute that only 170 lines of code were necessary to “write in the Java language.” Thus, the appeals court found it significant that Google copied 11,330 more lines of code than it needed. Further, the Federal Circuit opined that “no reasonable jury could conclude what was copied was qualitatively insignificant, particularly when the material copied was important to the creation of the Android platform.”

Related to this consideration, note that although the word “purpose” is not stated in the third factor, Google’s defense and the Federal Circuit’s reference to how much code Google “needed to” copy demonstrates how the third prong consideration is influenced by the first factor finding. Purpose is highly determinative of fair use.  For example, in Campbell v. Acuff-Rose, the Supreme Court made clear that 2 Live Crew’s intent to parody “Oh, Pretty Woman” allowed a very liberal taking of the “heart” of the original material.

Every user has a purpose, and juries can be tripped up when that purpose results in something new and significant like a mobile operating system that attains Android’s market share. But it is important not to be distracted by the fact that Android is a big deal because this market reality does not tell us whether Google made a transformative use of Oracle’s code as a matter of law. And if transformativeness does not exist under factor one, then copying what may seem like a small portion of a work is more likely to tilt away from a finding of fair use under factor three. Ultimately, the Federal Circuit found that the third factor was either a tie or weighed against fair use.

Under the fourth factor (effect on the actual or potential market for the original work), if factor one weighs against fair use—especially with a commercial use—then odds are that the effect upon the potential market for the original author will be considered harmful. Further, under the fourth prong, the courts assume that upon a finding of fair use that the use in question will be repeated by other users. Thus, courts take an expansive view of “potential” market harm in the factor four analysis.

Here, the Federal Circuit held that the jury drew incorrect inferences about potential market harm and held that no reasonable jury could have imagined that no actual or potential market harm would fall upon Oracle’s protected work. It rejected Google’s arguments that Oracle 1) is not a device maker; and 2) had not yet built its own smartphone platform. To the untrained juror, these two points seek to emphasize Google as innovator using a work to exploit a market the original author never intended to enter, but the appellate court rightly observes that these facts are irrelevant as a matter of law under factor four.  

The problem with Google’s arguments here is that they both clearly ignore that a copyright owner has a right to the potential market for lawful derivative works. One does not say to an illustrator that because she has never been a comic book producer, she is denied the exclusive right to license her imagery for a graphic novel. One also does not say to the sculptor that if another party makes tea cozies out of her designs, she’s out of luck because she did not think of doing so first. These would be errors of law, both with regard to fair use factor four and Section 106(2) of the Copyright Act, which provides the exclusive right to prepare derivative works.

So, based on my rereading of the Federal Circuit fair use opinion, it appears to have properly and extensively separated the factual and legal considerations in Google v. Oracle, considered the legal questions de novo, and correctly found as a matter of law that under the historical facts, no reasonable jury could have found fair use. I will be very surprised if the Supreme Court does not agree with the appellate court’s analysis and, therefore, even more surprised if they find any complaint about the standard of review compelling enough to remand.

Google v. Oracle VIII: On Juries Deciding Fair Use

Soon after the pandemic forced the Supreme Court to delay proceedings in Google v. Oracle, it directed the parties to “file supplemental letter briefs addressing the appropriate standard of review” with respect to the Federal Circuit’s decision in 2018, concluding that no reasonable jury could find that Google’s copying to create Android was fair use. On August 7, both parties filed their letters in response to the Court’s inquiry.

This is a nuanced matter, and I will leave the civil procedure questions to the practicing attorneys and those who have a detailed record on the day-to-day proceedings in this decade-long case. But I was intrigued by the broader copyright law consideration underlying the Court’s review standard question because it is one I have often thought about myself:  is fair use best weighed by a jury, by a court, or by some appropriate combination of the two? 

As a general rule, we look to juries to consider disputed issues of fact and courts to weigh or apply principles of law, and any copyright expert will tell you that a fair use analysis is a mix of fact and law. But as we see in Google v. Oracle, there can be ample disagreement about the line between fact and law that exists in a given fair use consideration.

As a simple example, consider factor three of the four-prong test—the amount and substantiality of the portion of the work used. A jury can reasonably arrive at a factual finding as to how much of a copyrighted work has been used. But the same jury may struggle somewhat with how much weight to give these facts with respect to the overall, four-factor, fair use analysis. Particularly because fair use is judge-made law, case law can be the most instructive guide, yet one which juries are not expected to know.

With that in mind, it is worth thinking about an argument presented by several legal scholars to the Supreme Court in defense of Google on this issue. They argued that the Federal Circuit’s failure to defer to the jury verdict on its ultimate conclusion of fair use was in error for three reasons. First, they argued it was unprecedented for a fair use finding by a jury to be overturned; second, they asserted that the court of appeals’ de novo standard for reviewing the matter was inconsistent with the Supreme Court’s standard; and third, they argued that the decision was an unconstitutional abridgement of the Seventh Amendment right to jury trial in civil litigation.

Is the “Exceptional Overruling” Truly Exceptional?

The first argument—that this was an unprecedented overturning of a fair use finding by a jury—may not carry much weight with the Court for the simple fact that jury decisions on fair use are not very common in the first place. Copyright cases involving fair use are typically decided on summary judgment, and fair use case law decisions encompasses a large anthology of discussion among judges that guides most fair use considerations before those questions get anywhere near a jury. [1] So, the supposed anomaly in Google v. Oracle, while it may be good for blog headlines, may not be as compelling a legal argument as it appears.

The Standard of Review: Are the Facts Really That “Complicated”?

The amici’s second argument—that the Federal Circuit erred when it engaged in de novo review—turns on that mix of law and fact bugaboo inherent to fair use analyses. When dealing with mixed questions of fact and law, the Supreme Court has stated that a court must decide whether the question tilts more toward fact or more toward law.  One reason for this is that when juries commit errors of factfinding, justice may not be served to the parties involved, but the law itself is not usually altered. Conversely, when errors of law are left intact without review, the law itself may be changed substantially, causing problems far beyond the specific case. Or as the Federal Circuit stated, in seeking to harmonize its standard with the Supreme Court’s opinion in U.S. Bank Nat’l Ass’n v. Vill. at Lakeridge, LLC (2018): 

Where applying the law to the historical facts “involves developing auxiliary legal principles of use in other cases—appellate courts should typically review a decision de novo. But where the mixed question requires immersion in case-specific factual issues that are so narrow as to “utterly resist generalization,” the mixed question review is to be deferential. [Citations omitted]

The Federal Circuit opinion devotes considerable time in its opinion explaining its approach before finding that de novo review was justified.  Nevertheless, amici for Google assert that the fair use analysis in this case is more factual than it is legal, tilting toward deference to the jury. Their basis for this is that, “the facts [in this case] are undeniably complicated.”

But are the facts in Google v. Oracle terribly complicated?  For instance, the Federal Circuit only highlighted eight facts it found to be relevant to the fair use analysis, including the finding on which the parties agreed that Google “copied material for the same purpose as in the original work” by using Java’s declaring code in Android. Even though that undisputed fact resolves the vast majority of the first prong of the fair use test, Google’s primary appeal to fair use is that its use of the Java packages to build Android —was “transformative” on the grounds that the work was used “in a new context.”

Here, the Federal Circuit correctly found this defense to be based on errors in both fact and law. Though I am admittedly jumping past the standard of review question, the record does not support Google’s claim of fair use. It deployed Oracle’s code for the same purpose in a small computer (a mobile device) instead of a large computer (a PC), which answers a question of fact; and even if this had been a new use, courts have largely held that “use in a new context” does not necessarily support a finding of “transformativeness,” which answers a question of law.

More generally, it is worth noting that “transformativeness” is one of the most vexing doctrines in the fair user’s handbook; it has confounded judges in district courts, sparked heated debate among legal scholars, and split circuits across the country. Very recently, in Brammer v. Violent Hues (2019), the Fourth Circuit held that a “…difference in purpose is not quite the same thing as transformation.” So, assessing “transformativeness” sounds an awful lot like a question of law amenable to review. As the Federal Circuit states in its opinion on this question:

… Google’s use of the API packages is not transformative as a matter of law because: (1) it does not fit within the uses listed in the preamble to § 107 ; (2) the purpose of the API packages in Android is the same as the purpose of the packages in the Java platform; (3) Google made no alteration to the expressive content or message of the copyrighted material; and (4) smartphones were not a new context.

Google characterizes the Federal Circuit’s holding as merely a challenge to “the sufficiency of the evidence,” based on its conclusion that “no reasonable jury could find that Google’s verbatim and entirely commercial use of the declaring code and SSO to compete against the Java platform was a fair use.’” (Citation omitted). Again, I will let the civil procedure question go, but note that to this layman, who probably knows fair use better than the average juror, that all reads as a mix of at least equal parts fact and law. “Verbatim,” “commercial,” and “potential” threat to the market are all bright yellow flags tilting against a finding of fair use as a matter of law. 

In fact, this is a pretty good example of what I referred to above as the distinction between juries and courts in fair use cases. Neither “verbatim” copying nor “commercial” use are disputed facts in this case. Yet a jury, even understanding these facts with little difficulty, can still err in apportioning weight to these factors in the overall fair use analysis. Hence review seems more than justified.

It is also notable that at least two of those subjects—“potential market” and “commercial use”– are frequently confused by the general public, and occasionally misapplied by district courts, as it was in Brammer. So, can a jury get these legal considerations wrong? I believe they can. Or as the letter for Oracle succinctly states, “A ‘primarily legal’ mixed question that is reviewed on appeal de novo does not somehow become a ‘factual’ one reviewed deferentially just because the factfinder was a jury rather than a judge.

What Does History and the Seventh Amendment Say?

Keep in mind that the procedural question at issue here is not whether the Federal Circuit erred in its holding that Google’s use of Oracle’s code was not a fair use (SCOTUS should rule on that question in the end), but only whether the appellate court applied the correct standard of review to the jury’s fair use finding , “including but not limited to any implications of the Seventh Amendment.”

To this, amici for Google present a final historical argument in order to draw two intertwined conclusions: 1) fair use in general is highly amenable to consideration by juries; and 2) evidence of fair use jury verdicts in eighteenth century English case law, establishes a jury right in the common law at the time of ratification of the Seventh Amendment.  

Citing two cases, Sayer v. Moore (1785) and Cary v. Kearsley (1802) amici for Google assert, “They represent common-law courts expressly recognizing that the jury should decide whether copyrighted material was used fairly by a defendant….History thus satisfies the constitutional test for whether to apply the Seventh Amendment to the issue of fair use.” These two cases are instructive, say the amici, because they entailed fair use questions presented to a jury. But according to a recent paper by Justin Hughes of Loyola Law School on this topic, those cases were not fair use cases—at least not cases consistent with the modern, American fair use doctrine. Hughes explains:[2]

As much as there are antecedents in these cases to our own thinking about copyright, there was also much going on that reflects concerns incongruent with today’s range of copyrighted works and ideas incompatible with our current themes.

In Sayer, for instance, Hughes observes that the jury charge contained prototypical elements of the idea/expression dichotomy more than facts relevant to a contemporary fair use consideration. And in Cary, Hughes agrees with New Zealand legal scholar Alexandra Sims who writes, “[Cary] represents the beginning of a judicial recognition of fairness in relation to the use of factual materials in the creation of new works, but not fairness in the sense of using material for the purpose of review and criticism or even quotation.” 

On that note, I will presume to interject here that it is a shaky proposal at best to imply that our modern concept of fair use was a living principle at the founding period, let alone one that was so well ingrained in England’s common law that one might hope to demonstrate that it was always viewed as a question of fact to go before a jury. I have not personally traced the fair use pedigree, but I have traced other copyright principles back a few centuries, and I would propose that the ways in which the Americans split with England are often more instructive to jurisprudence than the ways in which some historians identify precedent in the raw ingredients of a bygone world.

Perhaps most significantly, it is a unique feature of American copyright that fair use was, in part, codified so that copyright’s limitations would conform to our speech and press rights. And this is one area in which English and American copyright diverge significantly, especially when we travel back through the nineteenth and eighteenth centuries. For instance, English copyright law did not wholly shed its ties to anachronistic licensing acts (i.e. state authority to publish) until the mid-twentieth century. So, it is quite a stretch to assert that when English judges in 1785 and 1802 gave their jury instructions in Sayer and Carey respectively, they were asking those juries to weigh anything that truly resembles American fair use doctrine in the twenty-first century.

“The only analog Google’s amici have cited is the ancient doctrine of ‘fair abridgment,’ states Oracle. “Fair abridgment categorically excused a historical practice of shortening a longer work enough to ‘be called a new [work].” And this is not even always true under our fair use doctrine. Fair use, as we know it, simply did not exist in 1791 at the ratification of the Bill of Rights. Still, the controlling part of the Seventh Amendment is the clause that says, “…and no fact tried by a jury, shall be otherwise reexamined in any court of the United States, than according to the rules of the common law.”

This prohibition on reexamination is precisely why Google and its amici hope to emphasize the factual nature of fair use considerations and endeavor to straighten the historical line between England’s Lord Mansfield in Sayer, and America’s Justice Story, who became the judicial father of fair use when he presided over the case Folsom v. Marsh (1841). Citing Story in a subsequent case, Emerson v. Davies, amici for Google emphasize that Justice Story “described fair use as a ‘question of fact to come to a jury’ in 1845.”

But Professor Hughes asserts, quite reasonably, that Justice Story said nothing of the kind—at least not as a bright line rule that implicates a standard of review. Instead, Hughes elucidates:

Story is just quoting Mansfield in Sayer—and what Mansfield said in Sayer wasn’t specifically about fair use to begin with. Nowhere in Emerson v. Davies does Story say he himself believes that the test he set out in Folsom was a ‘question of fact to come to a jury.’

In fact, with respect to the standard of review, even if there were not a meaningful ideological divergence between England and the U.S., Hughes declares unequivocally:

Legal historians have yet to find any ruling or statement that a copyright infringement defendant in a court of equity had the right to remove the case to a court of law or to have any elements of the infringement action decided by a jury sitting in one of those courts of law.

Are Copyright Critics Begging for Uncertainty in Fair Use?

Finally, it should not go unnoticed that many of the same advocates for deference to the jury in Google v. Oracle share ideological kinship with some of copyright’s most ardent critics, who often write in papers, blogs, and social media comments that copyright is unjust because of its uncertainties. More specifically, asking for legal certainty in the software market is an explicitly stated goal by several of the amici who have filed on behalf of Google in this case. But it is hard to see this deference to the jury argument as anything other than a departure from that view, or as anything other than an argument of convenience in the moment. After all, if fair use considerations, already a complicated matter, are best left to juries without judicial review, it can only add uncertainty with implications for copyright law that go well beyond the short-term exigencies of software companies.  Or as Hughes puts it:

… if one wants a fair use eco-system in which there are some areas of more objective ex ante rule-like norms exempting certain reasonably defined categories of behavior from copyright liability – if one believes (as Justice Kennedy clearly did) that much of the conduct exempted by § 107 fair use is “amenable to regulation by rule” — one is likely to prefer de novo review as a means to clarify what is and is fair use.

On a very broad level, Google’s letter argues that a jury “properly decides fair use because it is well suited, from both a practical and a policy perspective, to determine how a ‘reasonable’ party would assess the defendant’s conduct.” And right there is where I return to my personal musings on this question and find that this “reasonableness” standard (i.e. a jury’s gut feel) is inherently disconcerting. I have consistently seen users of works and owners of works reveal a general misunderstanding of fair use. So, I believe Oracle’s letter is more on the money when it states that “fair use does not ask what feels fair to an ordinary person.” Instead, it “applies judge-made factors codified in the Copyright Act and elucidated by additional legal rules developed in a vast body of fair use precedent.”

[1] An analysis of district court opinions between 1978-2005 found that “more than half of the opinions addressed a motion or cross-motion for summary judgment, and of these 121 opinions, 86% granted the motion or one of the cross-motions. This supports the conventional wisdom that courts regularly resolve fair use issues at the summary judgment stage.”  Barton Beebe, An Empirical Study of U.S. Copyright Fair Use Opinions, 1978-2005, University of Pennsylvania Law Review, Vol. 156, No. 3.

[2] Justin Hughes. “The Respective Role of Judges and Juries in Fair Use Determinations.” Loyola Law School. Legal Studies Paper No. 2020-09