For ease of reblogging, etc., here is the full text of the comment that I, Peter Jaszi, and Michael Carroll submitted to the Copyright Office on Wednesday.
In re: Docket No. 2012–12, Orphan Works and Mass DigitizationComments of Brandon Butler, Michael Carroll, and Peter Jaszi
American University Washington College Of Law
May 21, 2014
Several myths and misstatements about fair use proliferated at the series of roundtables held at the Library of Congress in early March to discuss orphan works and mass digitization. As experts who have studied fair use and counseled practitioners on its proper application, we write to briefly highlight and respond to some of the most alarming misunderstandings about fair use and its application.
Myth: Fair use is unpredictable, and people who are not highly risk tolerant need more certainty than fair use currently provides. Representatives of the Copyright Office asked repeatedly whether the flexibility of the four-factor framework is a hindrance to all but the most courageous, risk-tolerant actors,[1] and rights holders claimed there was widespread confusion about what constitutes fair use.[2]
Fact: Fair use has become a stable, predictable, coherent doctrine. The courts are applying a unified view of fair use grounded in the concept of transformativeness,[3] first suggested by Judge Leval and endorsed by the Supreme Court in 1994[4], and for many common categories of use it is possible to make powerful predictions about how a court will assess specific examples.[5] There are more and more tools available to users to help them make these determinations, including best practices statements developed by user communities.[6] In reality, more and more people and institutions are relying on fair use on a daily basis, and only the myth of an arbitrary and capricious fair use doctrine is preventing others from joining them.
Myth: Courts’ solicitude of mass digitization is warping the fair use doctrine. June Besek of Columbia’s Kernochan Center has argued in several fora that the fair use doctrine is ‘incredibly expanding’ due in part to judges’ unwillingness to bar socially useful mass digitization projects such as the HathiTrust, and she repeated that claim at the roundtable.[7] She advocated in the Roundtable for a specific legislative exception, narrowly tailored to account for a list of specific policy concerns, that would channel mass digitization away from fair use and allow the doctrine to shrink to a more “reasonable” size.
Fact: Mass digitization is not really new to the courts, and it benefits from well-established fair use jurisprudence. The decisions applying fair use to mass digitization of books, Authors Guild v. Google and Authors Guild v. HathiTrust, are of extremely recent vintage, too recent to have exerted any real influence on doctrine in other cases. However, the fair use jurisprudence that informed the outcomes of both cases has been growing and entrenching itself in the courts in cases involving a wide variety of uses, some of which can be fairly described as “massive” in scale, others not.[8] Channeling “mass digitization” into a special exception will thus have little effect on the fair use doctrine’s overall development, though it would very likely slow the progress of socially beneficial digitization projects themselves.
Myth: Fair use case law has developed in a disturbing new direction in certain courts, or in recent years. A more generalized version of the previous myth is the claim, repeated several times during the roundtables, that mass digitization is taking advantage of an already-distorted, aberrant strain of fair use jurisprudence.[9] Critics insist on distinctions between judicial circuits, or between transformative “purpose” and transformative “content,” or between reasonable cases decided in an unspecified bygone age and unreasonable ones decided since.
Fact: Fair use as applied by courts has evolved into a clear, coherent, unified doctrine. Recent scholarship has demonstrated that over nearly 20 years courts have moved decisively away from a series of confusing and contradictory rules of thumb focused on market harm and toward an emphasis on transformative purpose under the first factor.[10] This trend embraces both the technology cases from the Ninth Circuit[11] and the publishing and fine art cases from the Second Circuit.[12] It is impossible to isolate an ‘aberrant strain’ of fair use thinking without rewinding fair use case law into the 1980s, before the Supreme Court’s foundational opinion in Campbell v. Acuff-Rose. Today, as Peter Jaszi testified before the House Judiciary Committee earlier this year, “fair use is working.”[13]
Myth: The Google and HathiTrust decisions are unusual in giving substantial weight to the public interest as part of the fair use calculus. Several representatives at the roundtables expressed surprise and outrage that Judge Chin and Judge Baer had both been moved by the public’s interest in the services provided by Google and HathiTrust.[14] One participant even suggested that she had read a great deal of fair use case law in the wake of the Google Books case and she had not seen any mention of the public interest.[15]
Fact: Fair use jurisprudence has always included consideration of the public interest. Professor Alan Latman’s 1958 Fair Use Study[16] explains that among the theoretical bases of the fair use doctrine is the principle that “as a condition of obtaining the statutory grant, the author is deemed to consent to certain reasonable uses of his copyrighted work to promote the ends of public welfare for which he was granted copyright.”[17] The courts have consistently invoked the public interest in applying Section 107.[18] In the Second Circuit, where both HathiTrust and Google were decided, the Court of Appeals has declared that “The ultimate test of fair use … is whether the copyright law’s goal of promoting the Progress of Science and useful Arts would be better served by allowing the use than by preventing it.“[19]
Myth: Fair use cannot apply to "massive” uses because users must evaluate use of each work on a “case by case” basis. Citing the courts’ frequent admonition that the flexible, four factor framework should be applied on a ‘case by case’ basis, some participants suggested that it is impossible for a user, and by extension, for a court to find that the use of many thousands or millions of works is fair without embarking upon many thousands or millions of individual fair use evaluations, rendering mass digitization de facto beyond the reach of a fair use analysis.[20]
Fact: Courts and ordinary people can, and do, generalize about classes of uses where the four factor analysis will apply in the same way. Judges, for example, have done so in the Sony v. Universal case (consumer time-shifting of broadcast programming), in iParadigms (copying of thousands of student works in many genres to detect plagiarism), in Perfect 10 v. Amazon (copying and display of web content for a search engine), and in Arribasoft (same). From academic libraries to The Daily Show to ordinary DVR users, a wide variety of people and institutions exercise their fair use rights regularly across hundreds of thousands of cases without resort to an elaborate use-by-use four factor determination. Instead, these practitioners reasonably rely on the belief that certain kinds of uses—time-shifting, parody, commentary, preservation—are fair across broad swaths of essentially similar use scenarios.
Myth: Specific exceptions like Section 108 occupy the field for covered communities, preempting fair use. Not surprisingly, the Authors Guild was the source of this claim, which it has pressed in its lawsuit against the HathiTrust and its member libraries. Guild General Counsel Jan Constantine said that libraries who relied on fair use to conduct their digitization projects were "violating Section 108.”[21]
Fact: Fair use can and often does supplement the specific exceptions in order to provide favored users with flexibility to engage in protected activities. Section 108 itself contains an express savings clause at 108(f)(4) that reserves the general right of fair use to libraries notwithstanding the specific rights granted in Section 108. There could not be a clearer expression of congressional intent on this issue. In addition, the House Report on the 1976 Copyright Act describes preservation of movies stored on fragile nitrate film stock as an example of a use ‘beyond 108’ that would “certainly be fair use.”[22] Jonathan Band has explained at length why Section 107 can act as a supplement to Section 108,[23] and the argument that Section 108 occupies the field has been decisively rejected in litigation.[24]
Myth: When fair use is applied in the same areas as existing narrow exemptions, those exemptions become useless and meaningless. Section 108, for example, was said to lose its reason for being if fair use can be invoked as a justification for library preservation.[25]
Fact: Specific exceptions describe safe harbors where Congress favors use without regard to the four factors. While fair use’s “uncertainty” is often exaggerated, there is still real value in establishing that certain uses are categorically favored and authorized without regard to the fair use balancing test.[26] These safe harbors can also help to guide courts in applying the fair use doctrine, because they give courts information about which uses are seen as favoring the public interest.[27]
Myth: Best practices in fair use developed by practice communities are less legitimate and useful than guidelines negotiated with rights holder representatives. Throughout the meeting representatives from the Copyright Office raised the question whether the fair use best practices documents developed by a growing list of user communities[28] were somehow deficient for failure to include rights holder representatives in their development.[29] Several rights holder groups answered in the affirmative.[30] Both question and answer misunderstand the nature of the best practices project.[31]
Fact: Fair use is a user’s right, and user communities need not apologize for deliberating together about their rights. On the contrary, it is absolutely essential for communities with shared missions and shared values to deliberate together about how the law affects their work and the ways they feel comfortable deploying their rights to achieve legitimate and socially beneficial shared goals. This is true for journalists who have shared views about the First Amendment, it is true for labor unions who have shared views about wage and hour laws, and it is true for user communities who are developing shared views about fair use. No one accuses journalists of being one-sided when they establish their own views about the legality of publishing allegedly secret government information, without first consulting the government. By definition this intra-community deliberative project differs fundamentally from the process of negotiating among groups with adverse interests. At the same time, because the best practices and the rationales that undergird them are declared publicly, they serve as invitations to dialogue with effected rights holders, who have not been hesitant to declare (and act upon) their own views about fair use.
Myth: Fair use may give substantial comfort to educational and non-profit users, but it is of little use to commercial actors. Representatives of the Copyright Office raised concerns that, although the library community seems increasingly disinterested in orphan works legislation for its purposes, commercial actors might be differently situated, perhaps because they could not rely on fair use.[32] Some rights holder representatives suggested that the presence of a commercial entity tainted fair use arguments in cases like the Google Books case.[33]
Fact: The vast majority of cases where judges have found fair use have involved commercial users or uses. Since Campbell, in which the Supreme Court held that commercial uses should not be presumed to harm the market for an underlying work, courts have repeatedly blessed commercial uses as fair when the purpose is appropriately transformative and the relevant market, therefore, is not a market the plaintiff has the right to control.[34] To be sure, the law rightly provides special protections for non-commercial educational users, and especially for state institutions, that cannot be claimed by private actors. But fair use is certainly available to commercial users, at least for uses where there is a plausible claim of transformativeness.
Myth: It is impossible to “export fair use” to countries who do not currently have a fair use regime. Citing the need for relative harmony with our treaty and trading partners, several commentators said a narrow special exception should be preferred over fair use as an answer to the orphan works problem. It would be better to “import” the approaches being tried in Europe (extended collective licensing, for example) than to “export” fair use.[35]
Fact: Foreign jurisdictions are perfectly capable of adopting and applying flexible exceptions like fair use. Some countries already have expanded their flexible exceptions,[36] and others are considering joining them.[37] While they are of course not bound by US case law, foreign judges could use the rich material in US fair use case law as a basis for developing their own law. Scholars have developed a model flexible exception for adoption in non-US jurisdictions.[38] In Australia, a copyright review process has resulted in the recommendation that an open-ended fair use provision be adopted, in part to encourage the development of a technology sector that could rival America’s thriving fair use-based economy.[39]
Myth: If there were a new narrow exception for orphan works or mass digitization, a fair use savings clause would be necessary and sufficient to protect fair use from any implied derogation of its scope. Some of the very same commentators who worried that fair use had “gone too far” were quick to reassure fair use allies that a savings clause could be added to any new exception to ensure the doctrine is not adversely affected.[40]
Fact: Fair use is a First Amendment doctrine. The Supreme Court has twice held that fair use is a built-in First Amendment safeguard in the copyright law, without which government grants of copyright privilege would be inconsistent with the public’s right of free expression.[41] Courts should read statutes as preserving rather than constricting fair use rights, even in the absence of an express savings clause.[42] At the same time, fair use can be shrunk in practice by offering apparent certainty in exchange for more conservative practice. This is why one commentator suggested that a specific exception for mass digitization would halt the “incredibly expanding” trend in fair use, by channeling mass digitization into a more conservative safe harbor. A savings clause would not necessarily forestall this channeling effect.
Myth: Fair use doesn’t adequately provide for “security” of digital files created in digitization projects. Some panelists argued that concerns about security, primarily online ‘hacking’ and downstream sharing of digital files, provide an important reason to carefully limit digitization projects with a narrow exception rather than flexible fair use.[43] This view treats a work’s analog nature as a kind of technical protection measure, and digitization as a form of circumvention that should be per se infringement, regardless of the purpose or legitimacy of the use.
Fact: Courts can and do consider reasonable security as part of the fair use determination. Judge Chin and Judge Baer both weighed the Authors Guild’s arguments about security as part of their four factor analysis, but there turned out to be no evidence of security risk in those cases.[44] A court could easily find differently were a plaintiff to show that lax security measures belied a defendant’s allegedly transformative purpose, for example. Security could also be considered under the third factor, the amount of the underlying work used, as sufficiently lax security arguably “exposes” (and hence “uses”) the work in ways not appropriate to the described purpose. The fourth factor may be the best place to consider this issue, as lax security may increase the risk of a harmful effect on the market for the work. And, as Jonathan Band noted at the roundtables, courts are not limited to the four statutory factors; they can consider security on its own as a factor.[45]
Conclusion
While the precise metes and bounds of the fair use doctrine as applied to orphan works and mass digitization can and should be the subject of reasoned debate (and judicial deliberation over time), that debate must proceed on the basis of a shared understanding of the basic facts about the doctrine. We hope that by countering myth with fact we have helped to lay a reasonable foundation for discussion.
Signed,
Brandon Butler
Michael Carroll
Peter Jaszi
[4] Campbell v. Acuff-Rose, 510 U.S. 569 (1994).
[31] See Patricia Aufderheide and Peter Jaszi, Reclaiming Fair Use: How To Put Balance Back in Copyright (2011).