On October 7, the Supreme Court finally heard oral arguments in the decade-long copyright software slugfest Google v. Oracle. Thomas Goldstein represented Google, Joshua Rosenkranz represented Oracle, and Deputy Solicitor General Malcom Stewart represented the United States as amicus curiae in support of Oracle. The major arguments discussed were the following:
- whether the Java declaring code Google copied into the Android OS is unprotectable under the limitation in copyright called the merger doctrine;
- whether the Federal Circuit applied the correct standard when it reviewed the case de novo and overturned a fair use finding by a jury; and
- whether Google’s use was a fair use, particularly whether it was transformative; and
- whether the Court’s decision for either party would risk upending the American software industry.
While it is folly—let me underscore, folly—to make predictions based on the questions justices ask during oral arguments, I will presume to make a few comments and some soft predictions. First, I believe the Court will have a hard time accepting Google’s merger argument and may even find unanimously for Oracle on that question.
Second, the Court may remand to the Court of Appeals on the standard of review question, but if so, it will have to address the question raised about what effect a decision could have on the practice to resolve fair use issues on summary judgment. Alternatively, the Court may hold that the Federal Circuit did not err with regard to civil procedure and will then comment solely on the appeals court’s holding that, as a matter of law, Google’s use was not a fair use.
Third, with regard to the fair use defense itself, the Court did not devote much attention to the question of transformativeness, upon which Google’s claim rests. So, good luck reading those tea leaves. And, fourth, by contrast, the justices did direct more questions toward Google’s claim that a finding for Oracle would cause “the sky to fall” upon the software industry. And while that claim may—may—concern the Court, readers should not confuse this argument with the fourth fair use factor, which asks whether market harm is done, or will be done, to a plaintiff in a case, not whether a finding against fair use conflicts with widespread to market practices overall.
Part of what makes this case historic is that Google, the darling of the contemporary tech world, has effectively asked the Court to upend copyright law in order to (allegedly) avoid upending the software industry. If that is a fair summary, then the Court’s path should be clear. If the legal arguments presented by Google are untenable and might, therefore, weaken copyright for every author under in its protection, the Court cannot accept Google’s claims on the merits. Further, Google’s broad appeal to avoid industry disruption (unless of course Google is doing the disrupting) is at odds with its own claim that the works it copied are not copyrightable.
Copyrightability & the Merger Argument
On the matter of copyrightability of the Java declaring code, Google conceded to Justice Gorsuch that it would focus its argument on the merger doctrine, pulling back from its earlier arguments that the declaring code constitutes an unprotectable “method of operation” under the statute. The merger doctrine denies copyright protection to works when there is only one way, or very few ways, to express an idea. In such instances, the expression and the idea (or in computer code, the function) are said to be merged.
As addressed in this post, one disconcerting aspect of Google’s appeal to merger is that it is based on its claim that it “had to” copy the Java code at issue. That need, however, appears to be a business decision to attract Java programmers to develop apps for Android, rather than a genuine limit of coding options available at the time the copying occurred.
As such there is no distinction between Google’s merger argument and the claim that someone “had to” copy any other protected work. A different creator might argue that he “has to” copy a lot of Star Wars material in order to make a new work that appeals to Star Wars fans, but this business goal (need) does not void the protection of the works he wants to use. This is as basic as copyright gets.
The justices homed in on this aspect of Google’s appeal to merger, using the word circular more than once to probe Mr. Goldstein on the question; and several justices voiced their concern that if merger is misapplied upon the condition of a use, rather than as a limitation at the moment of authorship, the result is tantamount to penalizing a work for its own success. This is another potentially hazardous aspect of Google’s merger argument for all creators: it implies that once a work is highly popular, it is practically, if not entirely, in the public domain. Though many copyright critics would love to go there, the Court knows this is not the law.
Having said that, several of the justices, most notably Justice Breyer, did challenge Mr. Rosenkranz with what Terrica Carrington at Copyright Alliance calls some “less-than-stellar” analogies. Justice Breyer referred a few times to the QWERTY keyboard as a comparison to the Java declaring code, Justice Roberts referred to the standard arrangement of a restaurant menu, and Justice Kagan described a hypothetically ingenious way to arrange products in a grocery store.
Of course, none of these examples is properly a subject of copyright because none entails a modicum of creative expression fixed in a tangible medium. Even if we imagine viewing each of these arrangements of non-protectable elements—the alphabet, restaurant dishes, or grocery items—for the very first time, it would be very hard to identify creative expression that transcends mere utility. By contrast, Google has already conceded that the code at issue is expressive, and this is supported by expert amici, who tell us that there can be tremendous creativity in the kind of code Google copied. Perhaps the justices here reflect what I mentioned in an earlier post—that analogies can be very hard to draw to computer code because code is illegible to most humans and inherently functional, neither of which is true of other works protected by copyright.
But just because the Court raised imperfect analogies attempting to divine the nature of code, this does not mean that Google’s merger claim gained much purchase on the bench. Once we separate all the rough analogies, the Court is still left with one question: whether Google could have chosen to author its own declaring code and sequence, structure, and organization to achieve its purpose of developing a popular mobile OS. And the answer to that is yes.
Indeed, the availability of alternative expression is the focus of the merger inquiry in the 1979 report to Congress written by the National Commission on New Technological Uses of Copyrighted Works (CONTU), which established the conceptual foundations for copyright protection over computer code. As a result, if Google could have written its own expression to achieve the same result, the Copyright Act provides that the Java declaring code is not merged. In the brief submitted by Professor Arthur Miller, former member of CONTU, he writes:
“CONTU described how it envisioned the ‘idea-expression identity’ exception’ would operate in the software context: when specific instructions, even though previously copyrighted, are the only and essential means of accomplishing a given task, their later use by another will not amount to an infringement.’ But ‘[w]hen other language is available, programmers are free to read copyrighted programs and use the ideas embodied in them in preparing their own works.’ In so saying, CONTU found that ‘[t]he availability of alternative noninfringing language is the rule rather than the exception.”
Google appeals to a different authority: the purportedly established practice of the tech industry to freely copy APIs. Google’s broad claim that “everyone in the software industry knows APIs are not protected” attempts to serve both as evidence to support its merger argument and as an implied warning to the Court that a finding in favor of Oracle will overturn decades of tech industry practice and, therefore, stifle innovation. But the “everyone knows” argument contains a few significant flaws.
First, as Mr. Rosenkranz stated, many other commercial users have paid to license Java, including licensing only the declaring code. This practice belies the allegation that “everyone knows” these works are not protectable. In fact, prior to the time Google copied the Java declaring code, it had been negotiating a license with Oracle and only chose to reject that agreement because of Java’s interoperability requirements—not because the works at issue were so obviously not subject to protection.
Next, and perhaps most importantly, Google’s “everybody knows” argument is circular reasoning with respect to the merger argument. Justice Thomas correctly asked whether merger applies at the moment of authorship or at the moment of an alleged infringement, and Mr. Goldstein answered that it is the latter. But aside from the fact that this is not the correct standard under §102(b), Google is deceptively asking the Court to hold that merger applies in both instances. Or perhaps more pointedly, Google wants merger to apply at the time that is most convenient to the alleged infringer.
Because if indeed “everybody knows” that API packages have never been copyrightable, then Google is asserting that merger applied the day Sun authored the code in 1995. Either Google is saying that the universal practice of “reimplementation” does not date back to the 1990s, or it is playing shell game with the merger doctrine. Hence, the reason the merger claim sounds circular is because it is. It is an argument of convenience, not law.
It Will Probably Come Down to Fair Use, But…
The justices did not focus a lot of attention on Google’s fair use defense itself, instead emphasizing whether the Federal Circuit applied the correct standard of review to the 2016 jury verdict. As discussed at length in this post, the simple explanation is that the Federal Circuit applied the correct standard of review if it assumed the facts weighed in favor of Google but that the jury below made errors of law in the fair use analysis.
Whether the Court will remand for another review is difficult to predict, especially given that it involves esoteric questions of civil procedure. However, the discussion about the standard of review did raise an interesting topic as to the potential effect on summary judgment practice. And because summary judgment is where most fair use defenses live or die, this seems like a topic worthy of its own post.
If the Court does remand, this will likely imply a finding that Oracle’s code is copyrightable and not merged. Otherwise, a fair use analysis is an absurdity because one cannot make a fair use of a work unprotected by copyright. But if I had to guess, I think this monster of a case probably will come down to the fair use question because the Court appeared to be unable to find a hook under statute or caselaw that the code at issue is subject to merger. If that’s the way it breaks, many will say “the Court just doesn’t understand the coding world.” Or, as Mr. Goldstein put it…
“Why would Congress want a rule that says: ‘okay, these developers are extremely familiar with these commands. They’re used to write creative computer programs. Let’s just make it as inefficient as possible for them? The only upshot of Oracle’s rule that it wants you to adopt is to make computer programming incredibly inefficient so that we have fewer creative computer programs.”
But stepping away from the nuanced legal questions for a moment, a reasonable prediction on the industry effect of this case militates against Google’s allegations and its all-too familiar attempt to stand in the shoes of small and independent producers. Google is the 900lb gorilla in just about any dispute, and there is at least as much risk (if not more) that a finding on its behalf will lead to the simplest result that when the biggest kid in the sandbox wants to take something, he will. Imagine the software developer who does not have Oracle’s resources trying to survive just to the appellate phase of this epic battle. More likely, they would tap out early and settle, but how would this outcome serve innovation better than the opposite scenario in which even the biggest gorillas are required to license protected works?
As such, the broadest market-based considerations return us to the same bugaboo: the fact that hundreds of other commercial users licensed Java code and, presumably, added their contributions to the world of computing. So, if Google ultimately gets a pass just because it is big enough to infringe as-needed and litigate indefinitely, that hardly seems like justice or the best possible outcome for future creators.