“Cute Little Tiny Baby Pig” by Michael Kappel CC-BY-NC
Last week, a district court judge upheld a jury’s decision that Google’s use of Java declaring code in the Android operating system was protected by fair use. If terms like “declaring code” and “application programming interface” aren’t common parlance for you, don’t worry, you’re in very good company (and may be eligible for a seat on the Federal Circuit). No matter how familiar you are with the technical details of the case, however, it offers a fascinating example of how fair use has evolved in recent years. It also suggests we take a closer look at one generally-overlooked aspect of the law: the second fair use factor.
The Google case goes back to a 2014 decision in the Federal Circuit finding that APIs, – described in detail in this article, but basically specifications that allow programs to communicate with each other – were protected by copyright despite the fact that most experts believe that they are functional and thus should not be eligible for copyright protection. Confronted with the specter of a $9.3 billion verdict, Google raised a fair use defense for use of the APIs and, at trial, went to great lengths to better-explain the technical issues, comparing APIs to a QWERTY keyboard, a library, a restaurant menu, and, amazingly, to a “literal actual physical filing cabinet labeled ‘java.lang’ that Google dragged into the courtroom for the benefit of the jury.”
In May, the jury sided with Google and found that the use of Oracle (which owns Java)’s code was fair. The jury’s decision supporting Google’s fair use of the APIs is an important win for open source, but also shines a light on the way that fair use has evolved over time. As chronicled from the courtroom by livetweeters like Sarah Jeong and Parker Higgins, much of the non-file-cabinet discussion centered on the four statutory fair use factors. Grounded in common law and enumerated in the 1976 Copyright Act, these are:
- the purpose and character of your use
- the nature of the copyrighted work
- the amount and substantiality of the portion taken, and
- the effect of the use upon the potential market.
Oracle, arguing that the use was not fair, focused primarily on the first factor (the purpose and character of the use) and asserted that Google’s copying was commercial, not transformative, and was done in bad faith – all strikes against first factor support of fair use. This emphasis on the first factor was no surprise, since it is often cited as the “key to the fair use inquiry,” but disagreement about the purpose of Google’s use may not have been what ultimately persuaded the jury.
As noted above, the Copyright Act lists four factors “to be considered” with no indication that any factor be given special weight. In the years immediately following the passage of the Act, however, courts looked primarily to the fourth factor (market harm), as in Harper v. Row, where Justice O’Connor wrote that this factor was “undoubtedly the single most important element of fair use.”
Beginning with the seminal 1994 Campbell case, however, the Supreme Court adopted a new test that placed the first factor front and center. In an influential early study on the use of the four factors, Barton Beebe (whose chart is included here) notes that, before Campbell, 59% of opinions gave primacy to the fourth factor. A more recent article by Neil Weinstock Netanel notes that since 2000 better than 70% of cases looked to the first factor’s “transformation” test. Since then, scholars have argued back and forth about the importance of those factors, with a general consensus today that the first factor’s transformation test is the guiding star.
As these approaches have risen and fallen – often with the third factor (amount) implicitly tied to the analysis of the first or fourth – one factor has been neglected: the second. The second factor (nature of the original work) has been mostly overlooked by scholars, who have asked “is that all there is?” and by courts, as in the Georgia State case where Judge Tojflat found that the second factor was “of relatively little importance.”
In this case however, something unusual happened: the nature of the copyrighted work became really important.
Considering the use of APIs, which probably shouldn’t be protected in the first place, the functional character of the declaring code brought the second factor to the fore. Google’s use is clearly commercial and Oracle would argue that market harm exists, but a jury nevertheless found that the use was fair.
Because the case was decided by a jury, there was no written opinion, just a simple “yes, the use was fair” so we can only speculate about the jury’s motivations. But the court’s subsequent order upholding the decision as reasonable does explicitly point to the “functional character” at issue. Whether the jury was ultimately swayed by the functional character of APIs, by the norms of the open source community, or just by the giant file cabinet Google dragged into the courtroom, fair use saved the day by patching up the damage done by a misguided decision.
It was able to do this by relying on the full scope of the statute, referencing all four factors. Even as the first factor’s transformation analysis empowers us in so many ways, we would be wise not to overlook the rest of the text. Copyright experts often remind us that loving fair use sometimes means giving it the day off once in a while. The same is true for transformation. In modern copyright, the first factor may be first among equals, but the law of fair use is more robust and more versatile than we often give it credit for.