Googling the Book Settlement
Wednesday, May 26, 2010
Clashes between competing visions of the good are more compelling than humdrum stories of good versus evil. This is no different.
Federal District Judge Denny Chin recently spent a long day hearing arguments on the Amended Settlement Agreement (ASA) proposed to resolve the litigation brought by copyright holders protesting Google’s digitization of the world’s books.
The matter justifies the attention it is receiving, because by any standard Authors Guild Inc. v. Google Inc. qualifies as “a great case.” Sheer scope is one factor; the outcome will affect tens of millions of books and hundreds of millions of people, including generations yet unborn. Another is the richness of its context. Thinking about the ASA quickly leads to considering the nature of property rights, the implications of the digital revolution, the industrial structure of the information world, and all points in between.
The dispute also meets the criterion that great cases should provide an arena for the dominant conflicts of their time: the shift from physical to electronic media qualifies. Even more dramatic power derives from the reality that both sides are right. Google is trying to do a fine thing, and makes compelling arguments that it deserves a reward for acting to promote the public interest (less a handling fee, of course). From this perspective, the failure of the ASA would be most unfortunate. On the other hand, many objections are cogent, both the known and unknown unknowns are considerable, and should Google really be rewarded for using its unique status to game the legal system (for driving 90 miles an hour in a speed zone, as its opponents say)? Perhaps not, and one can make an equally passionate case that approving the ASA would be a mistake. Clashes between competing visions of the good are more compelling than humdrum stories of good versus evil.
On balance, the ASA should and probably will be rejected, but this is by no means an entirely happy conclusion.
The case stems from Google’s decision in December 2004 to start digitizing the world’s books. The program has been modified over time, but at inception it was broad indeed. Google would partner with major libraries to digitize their collections, including works covered by copyright. Bibliographical information would then be available to a searcher. Contents would be made available differentially. Works in the public domain would be fully online; for other works, snippets would be shown, with safeguards against excessive cut-and-paste gaming.
Safeguards were established, but they are porous. Digital versions would be shared with library partners, so, given the broad views of their rights taken by some libraries, Google’s promises to protect against leakage were regarded by rights holders as a bit of vaporware. A copyright holder’s major protection was that it could notify Google and opt out of the program.
The outcome will affect tens of millions of books and hundreds of millions of people, including generations yet unborn.
Publisher and author organizations sued in 2005, and the litigation has dragged on. In the meantime, Google stopped digitizing works under copyright, partly in deference to the concerns of the rights holders and partly because Google has a voluntary “partners” program with rights holders under which it digitizes their offerings. This program gathered immense momentum, so, despite the cessation of the separate effort, many books are still being digitized.
Google’s claim that it can legally digitize bibliographic information is ironclad, under several of copyright’s fundamental principles. Its claim that it can copy content is more attenuated, because it is difficult to contend that copying a work does not transgress the right to copy, which is after all the derivation of the term “copyright.” Google’s justification rests on “fair use,” the rule that some uses of copyrighted material are deemed “fair,” and thus not violating the right. This is a complex doctrine, with a four-pronged test based on the nature of the use and the impact on the rights holder. Some uses are favored, such as news reporting, political debate, teaching, and scholarship, and a fair-use escape hatch from copyright is probably constitutionally required by the First Amendment.
Google’s fair-use contentions are aggressive, but not outlandish. They rest heavily on a transaction cost argument, that a use is fair when the transaction costs of obtaining permission become too large to be practical. Many find this argument persuasive, but it is not usually at the heart of fair-use analysis. In the context of real property, considerations of transaction costs and practicality do not justify wiping out claims. Furthermore, although copying a whole work rings an alarm bell, even if the copy is held confidentially, Google has reasonable arguments that this is reasonable in the context of the digital age.
But Google’s argument has several problems. A big one is that its practices do not translate well into a general rule. Rights holders might trust Google’s promise of protection, but the law has no special “Google clause.” If Google can copy freely for purposes of indexing and snippeting, can others also copy freely if they make the same promise of careful custody? A sensible rights holder might rely on Google, or on Microsoft and Amazon, but what about Fly-by-Night Digital, Inc.? What about the libraries that receive a digital file on a vague promise not to do anything wrong? Why should a rights holder be forced to rely on this promise when deciding whom to trust is an important part of ownership, and what is the remedy if the library does an “oops”?
Even more dramatic power derives from the reality that both sides are right.
Google shrewdly adopted a couple of additional provisions designed to buttress its fair-use argument. By avoiding direct damage to rights holders by limiting the public’s ability to view the works, Google could argue, certainly correctly, that making books findable enhanced their value rather than appropriated it. This made the rights holders’ objection to the indexing and snippets appear, probably correctly, to be a strategic effort rather than a genuine concern. Also, Google did not attach advertising to the books program, so it could not be accused of trying to profit by free-riding on the property of others—a non-commercial purpose is a key factor in fair-use analysis.
The case is evenly balanced. Law Professor Jonathan Band has written an excellent dialectic on the arguments back and forth, and the one certainty is his observation that “the case could easily have marched up and down the federal courts for years without resolution.”
Given these uncertainties, settlement makes a lot of sense. Furthermore, let there be no doubt that digitizing the world’s literature is a great idea, and one with immense public benefit.
Judged by the values of transparency, consistency, and clarity, the copyright system is a mess. Length, registration requirements, and renewal have been changed so often over the years that a ten-page table is required to set out all the relevant expiration dates, which vary with the date of publication, date of creation, nature of the author, and the author’s longevity.
Nor is the relevant data recorded anywhere. A fundamental function of any government is to establish systems that make property rights transparent—land registries and repositories for liens under the Uniform Commercial Code are vital to the smooth functioning of markets and finance. For copyrights, some information is on file with the Copyright Office, but the records are neither comprehensive nor digital. Private parties have now undertaken digitizing the Copyright Office records, but most relevant documents reside in private filing cabinets somewhere. Many works are orphans in that the author cannot be located or the publisher has gone out of business or the rights holder is for some other reason unknown, often even by the holder itself.
The situation presents serious problems to a company that wants to digitize the world’s information. It is hard to find out if works are covered by copyright, and, for many that are, it is hard to find the owner. Even if efforts are focused on currently active works, the number of rights holders is immense and the transaction costs of sorting out permissions insurmountable.
The situation will steadily grow worse. Until 1976, some formality was required; a work received protection only if it was registered with the Copyright Office and a © notice attached. Since then, no formality is needed. If sorting out rights seems difficult now, just you wait. Copyright attaches automatically to everything, including every email and Twitter. Going forward, the Internet is enabling micropublishing, and copyright status will grow even more difficult to track as firms go in and out of business. A huge “gray literature” exists—defined as “information produced on all levels of government, academics, business and industry in electronic and print formats not controlled by commercial publishing i.e. where publishing is not the primary activity of the producing body”—and only rarely is it part of the discussion. The gray literature is apparently not within the ambit of Google’s current digitization effort, but it would be a logical extension, and would compound the problems of finding rights holders and obtaining permission.
The ASA is a well-crafted and ingenious document that creates the equivalent of a joint venture between Google and the rights holders, under which out-of-print books are made available, the rights holders retain control if they want, and the revenues from book sales, advertising, and other sources are split, with two-thirds going to the rights holders. Books that are in print and available are not covered; those are left to the partner program, or to whatever other arrangements the rights holder cares to make. (For details on the settlement, consult Band again.)
For Google, it is an excellent deal, since it will have a de facto monopoly on the commercial uses of the digitized body of books, which it can mine for data and to which it can attach advertising, since the self-imposed stricture against commercial uses, made in the interests of strengthening the fair-use argument, will no longer be necessary. As far as the rights holders go, it looks like a good deal.
However, there are interests beyond those represented by Google and the rights holders, and these push against accepting the settlement. Normally, a settlement good for the parties should be no one else’s business, but the ASA is more than this. It is, as a practical matter, a legal rulemaking proceeding in the guise of a settlement.
There are interests beyond those represented by Google and the rights holders, and these push against accepting the settlement.
The first of these problems involves the legal system. The ASA stretches the use of the class-action mechanism, and creates not just a possibility but a certainty that the judge would become the ongoing Digital Books Administrator, as parties (and non-parties) kept running back to court with new thoughts and complaints. Such judicial excursions into long-term oversight rarely end happily, and the need is for continuing administration is a warning that something is being tried that exceeds the institutional competence of a court.
The other set of interests involve competition and industrial structure. The Department of Justice opposes the ASA because it fails to solve the “core issue” that the settlement would allow “Google, and no other entity, to compete in a marketplace that the parties seek to create,” and that “this outcome has not been achieved by a technological advance in search or by operation of normal market forces [but as] the direct product of scanning millions of books without the copyright holders’ consent and then using Rule 23 to achieve results not otherwise obtainable in the market.”
This objection is powerful. Even strong skeptics of antitrust law agree that government-created market power is a bad thing, but that is indeed what the settlement would do, even though the problem is well camouflaged.
As the Justice Department notes, any company that tried to enter the market would face the same legal challenges as those faced by Google, only the new context would be that the opposing parties (plus Google) would have an incentive not to settle on equivalent terms because the judicial approval of the settlement with Google would create a possibility of monopoly profits. The result would be a major barrier to entry, with significant domino effects on markets for both content and search.
Google can argue that it deserves reward for its entrepreneurial spirit in pushing the issue, and that without its initiative the book world would remain in the swamp of murky rights and insurmountable transaction costs. It has a point, but the argument is double-edged. It is difficult to think of any other company that has the combination of resources and public prestige necessary to pull off this power play, which makes concerns about its dominance even greater. Napster was innovative, too, and pushed the edge of the legal envelope—and look what happened to it.
Even strong skeptics of antitrust law agree that government-created market power is a bad thing, but that is indeed what the settlement would do, even though the problem is well camouflaged.
Another powerful objection based on industrial structure goes unmentioned by the Justice Department. Much of the ASA is devoted to provisions concerning not rights holders but educational institutions and libraries, designed to privilege these institutions against disintermediation by new business models based on the Internet. This is objectionable—if any industry cries out for reform, it is contemporary higher education. Besides, why should a settlement dealing with a dispute between Google and copyright holders contain so many provisions so solicitous of the current educational structure? It is a strange detour, and needs more attention.
To repeat, the conclusion that the ASA should fail is not entirely welcome, because the choices are not between this settlement and Utopia. Many argue that the current mess should indeed be solved by legislation, but Congress has not acted, is largely responsible for the problems anyway, and cannot even pass a straightforward bill dealing with the worst problems of orphan works. Besides, since Congress has decided it is responsible for all of American life, it may be too busy running the auto industry, the financial system, and healthcare to turn its talents to the gritty issues of copyright. In this calculus, the ASA beats inaction.
On the other hand, the choice need not be between the ASA and nothing, because some alternative steps to improve things are within the realm of feasibility.
Letting the Tea Steep
A major possibility is that Google might win the lawsuit. Fair use is a common law, judge-made doctrine, and its statutory embodiment reflects this; the courts are allowed and expected to evolve the concept. They could conclude that digitizing and displaying snippets from an out-of-print book that is commercially unavailable is fair use, period, and if the rights holder shows up, it can demand takedown or a share of any revenues, but not damages. Or the use could be fair as long as the original work is acknowledged, but not if it is simply appropriated. Or this fair-use approach could be extended still further; if the snippets are made available, and the rights holder does not come forward within reasonable time, then it could be fair use for anyone (not just Google) to make the whole work available on the theory that property has been abandoned.
(One reason for Google to settle is the fear of such a victory, and it is lucky for the company that I am not the judge here, because this is what I would decide. It is reasonable to apply such ancient legal doctrines as adverse possession, laches, and the law of salvage to books; if you sleep on your rights for an inordinate time and do not cultivate your acres or find your wreck, the world rolls over you, and what’s wrong with that?)
If the courts do not act under the doctrine of fair use, it might be possible to get congressional attention long enough for it to pass a one-sentence statute telling the Copyright Office: “Solve this, and here is the rule-making authority.” This would allow a finer tuning of the best parts of the settlement, perhaps run by a private voluntary organization akin to that set up by the ASA. It is an interesting thought experiment to consider whether a Copyright Office rule-making would produce such a Google-centric result as the ASA. The answer is, obviously, no.
(An additional cautionary note is also required here, though: different media present different issues, and the prescription for books may not be right for illustrations, music, photographs, or other types of creations.)
Congress could also direct the Copyright Office to set up a system whereby rights holders in books published before some appropriate year in the 1900s must reassert their interest or be subject to having them enter the public domain. This alone would solve much of the problem, since few unavailable works have any significant economic value, and best estimates are that approximately 70 percent of published books are out of print, though still under copyright, 20 percent are in the public domain, and 10 percent are under copyright and in print. Back in the day when copyright was granted for a 28-year term with a renewal available, the renewal rate was only 15 percent.
So rejecting the ASA might delay things a bit, but probably not too long, given the obvious advantages of getting the world’s literature into digital form, and the fact that alternative channels do exist. Considering the stakes, it would be wise to let the tea steep a bit longer.
James V. DeLong is vice president and senior analyst of the Convergence Law Institute, and a visiting fellow at the Digital Society.
FURTHER READING: DeLong recently discussed “Making Finance Easy to Fix, not Hard to Break,” “Supreme Climate Folly,” and "Opening a Can of Worms: Government and Climate Change Data." He also outlined the failure of Congress and the administration to encourage biotech investment in "Bleeding Biotech," and predicted "The Coming of the Fourth American Republic" after America's special interest state decays.
Image by Rob Green/Bergman Group.