Whether we’re aware of it or not, the doctrine of “fair use” built into copyright law is one of the most important protections available to scholars, librarians, and students. Every time you quote from someone else’s work, every time you photocopy an article for a student, every time you read a passage aloud to your class, you are technically in violation of copyright.
The reason that an army of publishers and FBI agents aren’t smashing down your office door is that U.S. jurisprudence has long understood that a totalizing approach to copyright would be disastrous. Fair use is the only way we as individuals can together do what is fundamentally a collective endeavor, scholarship, in an information ecology that otherwise lives and dies by the intensely individualizing force of the marketplace.
But fair use has been carrying a heavy load lately, and it’s starting to show its limitations. Over the last few decades and especially amid the recent “copyright wars,” a powerful new philosophy has emerged: Rather than seeing copyright as a careful balance between the interests of private owners and the public, powerful content industries have argued that robustly protecting private interests is always the best way to serve the public. It’s the trickle-down theory of knowledge: Give the power to the producers and get out of the way, and it will eventually get to everyone who needs it. And digital technologies have handed copyright owners further power to regulate the use of their work, to further commodify information in ways never before imagined.
While most of us in higher education are little content industries ourselves, we should not be seduced into forgetting our role first and foremost as the keepers, distributors, and developers of our society’s body of public knowledge. We must fight for the promise copyright made to the public: All these economic rights are only in the service of intellectual progress. However, our rhetorical arsenal in this battle seems to be only to trot out fair use, i.e. the right to violate copyright for progressive reasons. Technical copy protection? Don’t forget about fair use. Restricting peer-to-peer networks? Don’t forget about fair use. Suing our students for downloading? Don’t forget about fair use. Automatic permission systems in educational courseware? Don’t forget about fair use. It’s a wonder the poor statute can barely stand, considering how often it is invoked as defense, criticized as folly.
This dependence on fair use, to somehow safeguard all of the myriad “public interest” elements of copyright’s balance, risks crushing it altogether -- no more so than in the pending battle around Google Book Search.
For those who don’t know, the search engine giant recently announced its aspiration to digitize every book ever printed. To do this, it partnered with the university libraries of Stanford, Harvard, Michigan, and Oxford, and with the New York Public Library. Together they have already begun the process of digitizing works whose copyright protection has run out -- right now, those published before 1924. These books would be full-text searchable and could be read in their entirety online, for free. For more recent books still protected under copyright, Google intends to digitize and make them searchable as well; however, the text returned in response to the search query would only be a short excerpt around the located word or phrase. Publishers who don’t want their work to appear at all can opt out of the system. Links will lead users to vendors where the book in question can be purchased.
To be clear, Google’s project does require making copies of numerous copyrighted books, and an unauthorized copy at that. Google says this copy is a fair use. And in lawsuits brought in September and October of 2005, the Author’s Guild ( complaint) and the Association of American Publishers ( complaint) argue it is a violation of their rights, and an attempt to unfairly capitalize on their work.
Unlike battles around digital music that have occupied the courts’ attention of late, this case will be of vital importance for the academic community. What is at stake is the possibility of a digital database of all written knowledge, and the question of who gets to produce it and under what conditions. Some think this is the Library at Alexandria finally realized; others think it's risky to have just one company running the stacks. But the case will live or die not on the question of the value of such a database to users, but on the narrower legal question of whether Google has the right to scan the books to begin with.
Perhaps this case will settle -- Google certainly has the funds to do so if it chooses. If it does get heard by the courts, what is of greatest importance, I believe, is how well the doctrine of fair use can carry the weight of this particular dispute. Lawrence Lessig has argued that fair use is being stretched thin because copying is so fundamental to the digital environment; uses that never even rang copyright’s bell, because they now require a copy to be made in the process, find themselves under legal scrutiny. I believe this is true. But fair use has already been pulled in too many directions, well before the Internet stretched it to its breaking point.
Fair use has a century-long history in U.S. courts, as a handy way for judges to stave off copyright claims when the use in question is socially valuable. At first, it was a way to protect small amounts of copying for the sake of criticism; as Justice Story noted in Folsom v. Marsh (1841), “no one can doubt that a reviewer may fairly cite largely from the original work, if his design be really and truly to use the passages for the purposes of fair and reasonable criticism. On the other hand, it is as clear, that if he thus cites the most important parts of the work, with a view, not to criticize, but to supersede the use of the original work, and substitute the review for it, such a use will be deemed in law a piracy.”
As such, one of the important criteria used by the courts to judge a use fair has been whether the new work is “transformative,” rather than merely replacing the old. The most famous of these is Acuff-Rose v. Campbell (1994), in which a surprisingly culturally savvy Supreme Court found that 2 Live Crew’s sampling of the Roy Orbison classic “Pretty Woman” was a kind of parody, however crude, and should be protected as fair -- it “adds something new, with a further purpose or different character, altering the first with new expression, meaning, or message.”
However, when fair use was finally codified in 1976, the primary motivation was not to protect criticism or parody but to accommodate the increasing use of the Xerox machine, particularly in education. University libraries did not want to risk liability when they made copies of journals and book chapters for faculty and students, and aggressively lobbied Congress for some legal protection to do so. When fair use became law, it included the four factors that had developed through court precedent, but also specified “multiple copies for classroom use” alongside parody, criticism, journalism, and scholarship as the likely contexts for the use to be considered fair.
Making multiple copies of an article for use in the classroom does not claim to produce a new work, in the way that sampling Orbison’s tune does. The value of the use is not that it is “transformative,” but that it is “distributive.” Now fair use is saddled with two aspirations. If the first understands that new work often needs to lean on and borrow from existing work, the second understands that the market mechanisms and distribution technologies that circulate work do not always achieve the scope and access we would like, or that other socially valuable activities require.
The courts have since used fair use in this ‘distributive’ sense, allowing cable TV to retransmit copyrighted broadcasts to audiences who could not otherwise receive them, prohibiting Kinko’s from producing course packets without paying a fee but leaving open the possibility that universities could do so as long as they do not enjoy direct commercial gain, and most notably in Sony v. Universal (1984), granting VCR manufacturers immunity to copyright penalties because some VCR users do make unauthorized copies of protected movies. The court argued that users have the right to record shows in order to watch them at other times, that this in fact “enlarges the television viewing audience” -- even the beloved Mr. Rogers testified that he wanted public school teachers to be able to tape his show and show it in class the next day. Again, these fair uses are not transformative, but distributive.
Is Google’s book search project fair use? This was the question vigorously debated, but by no means settled, at the recent “Battle over Books” debate at the New York Public Library and the blog-off that followed. Most copyright watchers largely agree that, if it makes it to court, the legal answer will come down to a battle of precedents. (See, for example, Jonathan Band’s “The Google Print Library: A Copyright Analysis.”) Google will come out on top if the court sees the case as akin to Kelly v. Arriba-Soft (2003), which allowed an image search engine to copy images from the Web so as to make thumbnail versions available to user queries.
The publishers and authors will likely triumph if the court turns to UMG Recordings et. al. v. MP3.com (2000), where MP3.com was found to be infringing when it made single copies of 400,000 CDs in order to stock a digital locker from which users could stream music they could prove they already owned. Google needs fair use to accommodate an activity that is neither “transformative” in the classic sense, or “distributive” in the Sony sense. Neither precedent did either, and the solutions were work-arounds to force the square pegs of searching and streaming in the oddly-shaped hole fair use offers them.
Let’s give fair use a break by sending in a legislative relief pitcher, one that can better allow for the role search engines play in facilitating the circulation of digital information. If fair use has been protecting both ‘transformative’ and ‘distributive’ uses, today we need a statute that can cover the kind of “indexing” uses that Google is after.
If we recognize that the Internet offers us the chance to make much more of our society’s culture and knowledge available to more people, and we recognize that to make this massive resource most useful requires ways to navigate and search it, and we further recognize that search engines like Google need to make copies of that work in order to make it searchable, then we have a genuine and reasonable public interest in ensuring that that they and others can do so. At the same time, we should also ensure that doing so doesn’t undercut the possibility of selling these works, and ideally should help their sales.
The publishers’ concern is not that Google shouldn’t make books searchable, but that they should have to pay a fee to do so. Such a fee represents the compensation not for lost sales, but to match what they might have earned had they provided this search function themselves. So let’s imagine that they do just that; Harper & Collins has already announced that it will develop a digital database of their books, following the lead of academic journal publishers like Sage. We could decide that this is a reasonable exploitation of one’s copyright, and forbid Google from building a library.
What this is likely to produce is a bunch of different, publisher-specific archives, all searchable under different criteria in different ways, all with different rules for how much text you can view and under what conditions -- and price. Smaller publishers will be less able to afford to do any of this, so once again we will be incidentally privileging those represented by the larger publishers when what we want is all work to be as available as possible.
And all publishers will be in a position to exclude some of their works from public view, for whatever idiosyncratic (or, more likely, financial) reasons they fancy. Perhaps someone would develop a meta-search that could query many of these archives simultaneously and return the results together -- in all likelihood, it would be Google. But this does not solve the systemic problem posed by letting publishers also govern access to their content.
What I think we’re after is something more straightforward, but nearly impossible to achieve. In this dream scenario, every author would make his or her work available in a digital form that is searchable but cannot be redistributed, in a widely compatible format, marked with the same kinds of metadata. We wouldn’t need Google Book Search, because these book “footprints” would all be online and available for searches just as Web sites are. But this is certainly an unreasonable and prohibitive request to make of authors, at least right now. For all intents and purposes, this is what Google seems willing to provide for us, with the promise of some ad revenue in return. As a less than perfect version of that ideal, it’s quite good.
Waiting for fair use to shield this expanding range of uses is slowing the innovation in information, knowledge, and culture the Internet seems ready to facilitate. And every time it does, we risk a court setting a retrograde precedent that cements digital culture into place for good. We need a new statute that acknowledges and accommodates the common sense recognition that search is good, that it requires incidental copying, and that it should not be left to individual, competing publishers to make their work part of the public trust.
In a moment when we are handing content owners much more control not only over the use of their work but over access to it, we need to make a parallel commitment to ensuring and expanding access of a different kind, as an aggregate collection of all things thought and written that can be easily explored. And, we need to let fair use protect the activities it’s designed to protect, instead of letting it fray as it stands in as the only protection against a locked and licensed digital world.