Prof. Randal C. Picker, University of Chicago Law School
August 6, 2014
As the New York Times 2004 headline put it, “New Google Service May Strain Old Ties in Bookselling.” Strain indeed. Google announced its mass book digitization project, which it then called Google Print, at that year’s Frankfurt Book Fair. Now almost 10 years later, at least some of the legal pieces of the puzzle are coming into place. The Second Circuit’s June 10, 2014, decision in the HathiTrust case is a key piece and that decision in turn helps to frame an important outstanding challenge to the original Google project.
A quick review. Google’s strategy was audacious: enter into contracts with the world’s great libraries to copy millions of the books that they held. Bring those books online as a fully searchable text database made available to the world. If I copy a single book in its entirety, there is a very good chance that I would violate the copyright in that book, but Google was betting that when it did that with millions of books, it somehow would not.
And Google would not just digitize the physical books and make copies, it would multiply those copies. Less cryptically, although the contracts between Google and individual libraries vary, Google typically returned not only the original physical book to the library but at least one digital copy of that book as well. Google kept a new digital copy for itself and also gave the library a copy that it could put to use. To do what? The HathiTrust was the answer. The HathiTrust was formed in 2008 as a digital repository for the copies received from Google as well as those from other digitizers.
Both Google and the HathiTrust had to understand that they were in uncharted legal waters and knew to expect the litigation that would come. The Authors Guild describes itself on its website as “the published writer’s advocate for effective copyright, fair contracts, and free expression since 1912.” That sounds a lot like apple pie and motherhood, but the actual posture of the Authors Guild is more contested and competing organizations, such as the Authors Alliance, have emerged.
The Authors Guild separately sued both Google and the HathiTrust. The Authors Guild suit against the HathiTrust raised many issues, some of which I will ignore here (such as whether the Authors Guild could actually assert claims on behalf of authors (the associational standing issue – they could not) or whether possible orphan works’ uses had been appropriately teed up (ripeness – they had not). Instead, I will focus on the core copyright issues presented in the case.
The case against the HathiTrust focused on three different uses that the repository was making of the digital copies that it held. First, the repository offers a fully searchable text database of both public domain works and the digitized copyrighted works it holds. For public domain works, it shows the results in full view – meaning the entire book – but for works in copyright for which it does not have permission to display more, the search results only indicate the number of times the relevant search item appears in the book and the pages on which those terms appear. The HathiTrust does not return so-called snippet views in which the search terms are displayed surrounded by the other words near those terms. Second, the repository also makes books available more broadly to those with disabilities. Third, Section 108(c) of the Copyright Act offers specific use rights for libraries and archives to reproduce books that are deteriorating if they can’t purchase a new copy of the book at a fair price.
Three different sections of the Copyright Act might apply to the repository’s activities. These are Section 107 on unfair use, Section 108 on the special use rights of libraries and archives, and Section 121 that creates protected uses for individuals with disabilities. Each of these operates as a limitation on the rights of copyright holders under Section 106. The Authors Guild contended that the existence of the institution-specific rights under Section 108 foreclosed the ability of libraries or archives to take advantage of fair use rights under Section 107, but the Second Circuit rejected that, given the clear language of Section 108(f)(4) (“[n]othing in this section … in any way affects the right of fair use as provided by section 107… ).
Section 107 sets out four factors to consider: “the purpose and character of the use”; “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.” The Second Circuit focused most of its attention on the first and fourth factors.
Modern first-factor analysis focuses on the question of transformative use, meaning exactly how transformative is the use in question? A use that substitutes for the original use is much more likely to be found to be infringing, but a use that genuinely changes the original work – transforms it into something new – is much more likely to be found to be fair use. That is the framework that the U.S. Supreme Court applied in its 1994 decision in Campbell v. Acuff-Rose, Inc. As the Supreme Court put it there, “although such transformative use is not absolutely necessary for a finding of fair use, … the goal of copyright, to promote science and the arts, is generally furthered by the creation of transformative works.” The Second Circuit itself has built upon that analysis in its 2006 decision in Bill Graham Archives, which involved miniature versions of Grateful Dead concert posters republished in an illustrated Grateful Dead biography, and much more recently in 2014 in Cariou v. Prince, which involved the reuse of Rastafarian photographs in large-scale (and quite expensive) artworks. In HathiTrust, the Second Circuit regarded the full-text search capability as being clearly transformative.
As to the fourth factor, the Second Circuit focused on whether the full-text search use would impair any specific existing or potential traditional market that the copyright holders operated in and concluded that it would not. In that regard, the Second Circuit rejected a frequent line of analysis seen in fourth-factor discussions. Copyright holders will contend that they should be able to sell to subsequent users whatever use is at stake, and if they can’t do that, a potential market for their works is being impaired. There is clearly a fundamental circularity there and resolving that argument in favor of copyright holders would mean that any new subsequent use would be fair for only a brief period of time. Once the subsequent users had made clear that that use had economic value, the original copyright holders would jump in and seek to license the use.
That would be very much a heads-I-win, tails-you-lose situation. Subsequent users who invested resources in new failed uses would suffer losses, but in cases where they identified new valuable uses, initial copyright holders could scoop up that value by offering to license those uses. Sensible fourth-factor analysis puts the burden on the copyright holders to identify the markets for their works and protects subsequent users who discover new markets for works. The Second Circuit’s analysis in HathiTrust is consistent with that framework. The Authors Guild also argued that the creation of multiple copies created security risks and potential harms down the road, but the Second Circuit declined to try to forecast the future.
To move quickly to the other two uses by HathiTrust of the digitized works, the Second Circuit relied on the legislative history of the 1976 Act to indicate that use by blind persons was a particularly good example of a use that was likely to be fair. That result also turned on the fact the publishers did not offer many of their works in a format accessible to the disabled. The Second Circuit did not rely on the separate targeted-use provision for the disabled (Section 121). Finally, on the Section 108 reproduction right, the court concluded that the Authors Guild had not presented that question appropriately, as it wasn’t clear that any books of actual plaintiffs in the case had been reproduced by the HathiTrust.
The result in HathiTrust tees up the pending Second Circuit appeal between the Authors Guild and Google. Google won on fair use below, and even though the Authors Guild indicated in its press release on the HathiTrust decision that it had “mixed feelings” about the result in HathiTrust, it simply could not have been pleased to lose again on fair use.
The Authors Guild filed its initial brief in its Google appeal on April 7, slightly more than two months before the release of HathiTrust on June 10. Google filed its responsive brief on July 3 and unsurprisingly sought to invoke HathiTrust as often as possible. In its reply brief filed on July 24, the Authors Guild needed to figure out how to separate Google from HathiTrust and unsurprisingly focused on two issues. First, HathiTrust is a nonprofit while Google is, as the brief puts it, “a highly commercial search engine” and second, Google search results return snippets and HathiTrust does not.
Do those two differences matter? Many users are in it for the money; 2 Live Crew, the band in Campbell, was not a nonprofit and neither were the fair users in Bill Graham Archives or Cariou. Those cases tell us that for-profit use and fair use are perfectly consistent. Google is likely to carry off the same fair use analysis on the full-text searchable database, as that seems to follow almost immediately from HathiTrust. The key issue will be the incremental snippet use and how transformative that is. As a user I will say that the extra use is meaningfully more valuable and a very different use than occurs in just reading the text sequentially. My guess is that will suffice on first-factor Section 107 analysis and that there will be no shown harm on traditional fourth-factor uses or potential uses.
As to a bottom line, I expect Google to win. If it does, Google will be rewarded for the bold vision offered at the 2004 Frankfurt Book Fair. Fair use analysis has evolved over the last decade in ways that very much favor Google and that evolution flows from the 1994 decision in Campbell, but there were no guarantees here. This was both a big technical bet and a big legal bet and Google looks like it is sitting pretty right now.
Comments From Our Readers
John E. Miller:
The following quote from the Authors Guild v HathiTrust 2nd Circuit ruling was reccently submitted to the US Copyright Office as part of a Comment to the renewal of Section 1201 exemptions:
http://copyright.gov/1201/2015/comments-020615/InitialComments_LongForm_AFBetal_Class09.pdf
From that submission at their Page 11:
“This year, the Second Circuit affirmed that conversion of inaccessible copyrighted works into accessible digital formats for use by people who are blind, visually impaired, or print disabled is a fair use in Authors Guild, Inc. v. HathiTrust. As the HathiTrust court held, ‘the doctrine of fair use allows [the] provi[sion of] full digital access to copyrighted works to [the] print disabled.’╡
That is not what the 2nd Circuit ruling said. The exact words of the ruling as quoted and modified above are:
“Weighing the factors together, we conclude that the doctrine of fair use allows the Libraries to provide full digital access to copyrighted works to their print-disabled patrons.”
There is nothing in the 2nd Circuit’s ruling to indicate that their actual assessment of ‘fair use’ applies to anyone other than the eligible patrons of the HathiTrust Consortium Member Libraries and the Libraries themselves presuming that the standards and requirements of the HathiTrust Members as currently drafted are maintained and enforced.
So while AFB and others participating in their S. 1201 submission seem to have inferred that the 2nd Circuit decision applies in a wholesale manner to any person with a print disability, anywhere, and to anyone who might reproduce.
Doug Lichtman:
Randy,
Your analysis on the fourth factor doesn’t strike me as right. The point of the fourth factor is to figure out the degree to which there is money at stake for the copyright holders. The more money there is, the more we need to think about whether it makes sense for that money to go to the copyright holder and in that way encourage all the good things copyright law encourages. The mere fact that we acknowledge that money, however, does not tell us that the use is or is not fair. Instead, this consideration – how much money would go toward copyright holders if we deny fair use here – is then weighed against other factors, like the first factor’s concern about whether the absence of a fair use defense might meaningfully chill creativity of the sort at issue in the case. So factor four analysis is not circular in any bad sense when done correctly, and factor four definitely should not ignore the money at issue in the case at hand. Factor four instead should be used to highlight the best argument against fair use (we are denying some money that would otherwise go toward the copyright holder) and other factors should then highlight the best arguments in favor (like the possibility of losing this new creativity if the fair use defense is not accepted).
Doug Lichtman
Professor of Law
UCLA
John E. Miller:
From the above: “… the Second Circuit relied on the legislative history of the 1976 Act to indicate that use by blind persons was a particularly good example of a use that was likely to be fair.”
The exact wording of the 1976 94-1476 House Report says:
“For the most part, such copies and phonorecords are made by the Library of Congress’s Division for the Blind and Physically Handicapped with permission obtained from the copyright owners, and are circulated to blind persons through regional libraries covering the nation…. While the making of multiple copies or phonorecords of a work for general circulation requires the permission of the copyright owner … the making of a single copy or phonorecord by an individual as a free service for a blind persons would properly be considered a fair use under section 107.”
So is the 80-some member HathiTrust consortium the same as an “individual” or would it have been required – as did the Library of Congress / NLS from 1976 until the Chafee Amendment was passed in 1996 – to seek the permission of the copyright owner?