Google’s Fair Use Shell Game

“[Google is] pretending that moving the same copyrightable expression from one place to another doesn’t affect the market for that work. That approach would excuse verbatim copying applied to virtually any other device or context. Like the shell game, the only thing that’s certain is the copyright owner will lose.”

Google shell game - has admitted it copied over 11,000 lines of Oracle’s creative Java code to build its Android smartphone platform, that it makes commercial use of the code it copied, and that it uses the code for the same purpose as Oracle and those who license its products.

Now, Google wants the Supreme Court to hold that this is fair use under American copyright law. Stripping away the diversions that Google and its amici offer the Court, these are the core facts which show how extreme Google’s invocation of fair use in this case truly is.

I’ve previously written two articles for IPWatchdog on the high-profile Google v. Oracle case before the Supreme Court, addressing the history of Google’s Java code use in Android and its copyrightability arguments.  As the October 7 oral argument date nears, I’d like to build on the fair use discussion by Washington, D.C. attorney Terry Campo published in August to further analyze Google’s claims of fair use to excuse its copying. The company’s claims aren’t just insufficient, they’re undermined by Google’s own arguments.

The four factors codified in Section 107 of the U.S. Copyright Act by which  fair use is assessed are familiar to any copyright lawyer:

  • The purpose and character of the use, including whether such use is of a commercial nature or is for nonprofit educational purposes;
  • The nature of the copyrighted work;
  • The amount and substantiality of the portion used in relation to the copyrighted work as a whole; and
  • The effect of the use upon the potential market for or value of the copyrighted work.

Purpose and Character

There can be no credible claim that the multi-billion dollar revenue stream from Google’s Android smartphone platform is anything but a commercial venture; Google conceded as much in a brief to the Supreme Court. Rather, it hopes to overcome that strong weight against fair use by claiming its copying was “transformative.”

The Supreme Court ruled in the 1990s fair use case over musical compositions, Campbell v. Acuff-Rose, that a use is transformative and may overcome its own commercial nature if it “adds something new, with a further purpose or different character, altering the first with new expression, meaning, or message.”

Google has a problem in meeting this standard – its own statements prove it doesn’t. In the context of its arguments about the copyrightability of Oracle’s code, Google spent many pages in its briefs telling the Court only one “precisely written set of declarations will perform the function,” and that if “the declaration is changed, the call will fail.” The company concluded each “declaration that Google reused [copied] is thus the only instruction that performs the function….”

I’ve already discussed Google’s merger argument and why it fails. But here, in the context of it’s fair use claims, Google cannot escape that once it made the decision to copy Oracle’s code, by its own admission, it locked itself into using exactly the same code for exactly the same purpose. Google’s argument is built on the foundation that in order to copy Oracle’s code, it had to copy it exactly. This is the opposite of a transformative use – it’s verbatim copying.

The Federal Circuit agreed with that assessment, ruling the copied declaring code served the “same purpose” in both Java SE (a version of Java for PCs and servers) and Android. Google calls that an error and complains “[t]hat legal rule would dramatically limit any fair use of computer code.” It seems to have forgotten that the Federal Circuit didn’t invent the “legal rule” that the purpose of the use matters in fair use analysis. It’s encoded in the law. If that constrains Google’s definition of fair use, it’s more indicative of the company’s overly expansive views than any flaw in the reasoning of the Federal Circuit or the statute.

Google argues that even though the same Java code was used for the same purpose, its use was transformative because Google used the code in smartphones instead of PCs. It cited mobile devices’ constraints, including differences in the power, memory, and processing chips. This fundamentally misapplies the transformative use doctrine.

By its own admission, Google made no changes to the code it copied from Oracle. Instead, it’s trying to claim transformation based on other code Google wrote for use with what it copied. Adding creative expression alongside the unaltered original simply is not “altering the first with new expression, meaning, or message.” Google can’t escape the centerpiece of its own brief: it copied Oracle’s Java code without alteration.

Even if the Court were open to considering such extraneous matters, none of the differences Google highlights as unique to smartphones indicate creative expression; they’re utilitarian. And utilitarian considerations are anathema to creativity, copyrightability, and transformative use, which is longstanding settled law reaffirmed by the Supreme Court in the 2017 Star Athletica v. Varsity Brands case.

Google also seeks credit for “transformative use” based not on what it did, but for the apps written by third-party programmers after it released its Java-powered platform. Case law is clear that Google can’t stand in the shoes of third parties, any more than a commercial photocopy shop can claim to be engaged in nonprofit, educational uses because it sells its services to students. In both cases, verbatim copying is neither transformative nor fair use.

Nature of the Copyrighted Work

Google takes the opportunity of the second fair use factor to repeat its claim that Oracle’s code deserves little or no copyright protection because it performs a function. Of course, all computer code performs a function, so this is another example of Google’s wholesale attacks against the statute and all software copyright protection.

Once again, Google’s own statements undermine its arguments. As the brief of the United States recognized, Google conceded that the Copyright Act’s originality requirement “is satisfied for both the 11,330 lines of declaring code that it copied and the SSO [structure, sequence, and organization] of the Java Standard Library.” Because Oracle seeks protection for its particular creative expression and not the underlying function, the second factor of the test weighs against fair use.

Amount and Substantiality

Google emphasizes that the amount of computer code copied in relation to the entirety of Oracle’s Java code base is quite small. However, this factor is not merely a quantitative analysis, it’s also qualitative. The Supreme Court memorably ruled in Harper & Row v. Nation Enterprises that when a copier takes “the heart” of the work, the third factor weighs against fair use.

Google’s briefs repeatedly emphasize the importance of the particular declaring code. “Google created its own computer code whenever possible,” but without copying Oracle’s declaring code, the third-party app developers Google wanted to attract to make Android a more attractive platform would have “to learn thousands of new calls,” Google wrote in its January 6 brief to the Court. Google therefore chose to copy what it considered necessary “to meet [the] expectations of Java programmers’ creating smartphone applications,” it confessed in its March 11 reply brief.

I’ve written previously that Google’s myopia – ignoring alternative code to achieve the desired functionality – was proven by both Apple’s iOS operating system and Microsoft’s Surface devices. Only after Google made the decision to copy Oracle’s code, the Java programming language purportedly “permitted no other options,” as Google wrote in its reply brief. It was the declaring code, and third-party app developers’ familiarity with it, that drove Java’s popularity and Google’s corresponding desire to free-ride on the value of that specific expression.

When Google decided to copy, it took the heart of the work – the declaring code. The third factor clearly weighs against Google.

Potential Market or Value of the Work

Despite the fact that Google is using Oracle’s code for the very same purpose for which it was written, it nonetheless tries to argue that deploying Oracle’s code in smartphones is so different from use in other computing devices that it doesn’t harm the value of the work or effect the potential market.

This is Google’s version of the street hustler’s shell game, pretending that moving the same copyrightable expression from one place to another doesn’t affect the market for that work. That approach would excuse verbatim copying applied to virtually any other device or context. Like the shell game, the only thing that’s certain is the copyright owner will lose.

Actual Supreme Court precedent is much more straightforward and logical. The Court held in Harper & Row, a case involving copying of a work (in this case, memoirs of former U.S. President Gerald Ford) that was actively being licensed, that only rarely does an infringement case “present such clear-cut evidence of actual damage.”

Given the undisputed fact of Oracle’s vibrant licensing market for its Java-based code, and the concrete example of Amazon leveraging the availability of Oracle’s code through Android to secure a 97.5% discount for a license from Oracle, this case is another such example.

Courts vs. Engineers

Google and its amici go to great lengths to browbeat the Court with policy claims that copying others’ code is “long-standing, ubiquitous practice,” in the tech industry, which it emphasizes by referencing the software engineers who’ve filed friend of the court briefs.

Claims by Google and its amici offer a false dichotomy: The Court’s choice isn’t between copyright and copying; it’s between licensed and unlicensed use. If the Court were to rule as Oracle asks, it would not prevent future “reimplementation.” Future innovators who don’t want to write their own code would simply need to abide by the licensing terms Oracle has long offered (including the very open source licenses touted by Google’s amici).

The manner in which these policy arguments are put forward by Google and its amici also carry the implication that the Court isn’t competent to rule on technology issues. While it may be true that few of the Supreme Court Justices or their clerks have the specialized knowledge of software engineers, few software engineers have specialized legal knowledge. Google’s framing invokes a fundamental question of the rule of law: Do courts interpret and apply the laws? Or do software engineers get to decide what’s best for society?

Google spent the first part its briefs to the Court arguing that Oracle’s declaring code isn’t creative, and the latter part arguing its verbatim copying is transformative. They managed to be wrong both times: the code is creative and copyrightable, but nothing Google did in transposing it into Android was transformative – it was just plain old copying. Google’s arguments in this case all depend on dramatic and unprecedented narrowing of copyright protection or expansion of fair use. The law is the domain of the Supreme Court and it shouldn’t hesitate to apply that law — as it did in the peer-to-peer Grokster and TV retransmission Aereo cases — to rule that neither engineering gimmicks nor arguments that twist the intent and meaning of the Copyright Act can avoid accountability for building a business by free-riding on others’ copyrightable works.

*Special note: Justice Ruth Bader Ginsburg had an historic positive influence on copyright law, penning several major copyright decisions and leading the Court with some of the most clear-headed and logical analysis of those issues. In this regard, her loss seems irreplaceable. Nonetheless, I trust that the Court will continue to be guided by her legacy in this and other copyright cases in the future.

Image Source: Deposit Photos
Author: BrianAJackson
Image ID: 24525713


Warning & Disclaimer: The pages, articles and comments on do not constitute legal advice, nor do they create any attorney-client relationship. The articles published express the personal opinion and views of the author as of the time of publication and should not be attributed to the author’s employer, clients or the sponsors of

Join the Discussion

4 comments so far.

  • [Avatar for CharlieSeattle]
    October 4, 2020 10:53 pm

    Good Article.

  • [Avatar for Pro Say]
    Pro Say
    September 26, 2020 01:23 pm

    Big +1 Tim @ 2.

    Vringo was one of the worst cases — and darkest days — of IP jurisIMprudence in the history of our Nation.

    The law and juries be d.a.m.n.e.d.

  • [Avatar for Tim]
    September 25, 2020 10:51 am

    All of those at Google in high positions should be in Leavenworth with judges Mayer and Wallach after these two flipped “Vringo vs IP Internet” findings where 12 jurors found them guilty on every single charge 12-0 on that Infringement case in Judge Jackson’s Norfolk Court. Mayer and Wallach tossed the case for Google, AOL, Target and Gannett at the Appeals Court, as the only qualified judge Chen, former head of the USPTO highly dissented that they were wrong. And our “wonderful” Supreme Court wouldn’t even see the case! All part of the swamp! Vringo and s now XSPA symbol on Nasdaq, and has been reverse split 3 times and at $1.83 is now a wellness company and is worthless. Many of us lost bigtime because of Google crooks and corrupt judges.

  • [Avatar for Pro Say]
    Pro Say
    September 24, 2020 05:58 pm

    Google, to:

    Copyright holders: “Hey! That’s patentable, not copyrightable!”

    Patent holders: “Hey! That’s copyrightable, not patentable!”

    All other IP holders: “Hey! . . . er, um, gulp . . . well . . . just hey!”