Some people concerned about the proposed Google Book Search settlement, including me, have complained that the settlement is a ticket good for Google only. Google gets a set of copyright licenses that would be difficult to obtain through traditional negotiations–and effectively impossible in the case of orphan works. No one can compete on equal terms; the public at large can’t take full advantage of the orphan works Google has access to.
Google has consistently replied that anyone is free to do what it has done: scan books and seek a class-action settlement. I’ve expressed public skepticism about this claim, noting that there’s no legal right to such a settlement, only an uncertain prospect of seeking one. I’ve also worried in private that the settlement is so precisely tailored to Google’s circumstances that even the same exact deal would be of little use to anyone else.
But what if I’m wrong? What if others could obtain their own Google-style settlements without fuss? If the Google settlement were on the books, what other settlements would be on the table? What rights could Google’s competitors obtain to books; what could open access advocates do for other works?
This brief essay is an attempt to answer those questions by turning the settlement inside-out. Rather than focusing on what makes the Google settlement unique, I’ll examine what it would take to replicate it. David Balto has argued that the Google settlement would be a “precedent” for future, similar settlements. If so, then we should treat it the way that lawyers are trained to evaluate precedents from the first day of law school: by asking which of its features of it are essential to its precedential value, and which could be changed the next time around.
Before you can obtain your own Google-stye settlement, you need a lawsuit to settle. Thus, your first order of business is to be sued in a way compatible with the settlement you seek. You need to find a group of copyright owners, create a genuine copyright case with them, and make sure they’re neither too friendly nor too hostile towards you.
Your goals for your collection will drive the choice of who ought to be in the plaintiff class of copyright owners. Commercial vendors will probably focus on high-value works; open-access fans may be more concerned with comprehensiveness. Both will want to minimize litigation risk. If some subset of your ideal class would be particularly hard to work with, consider excluding it. Google made this strategic choice when it came to illustrations in books, for example.
It’s useful to put orphan works in your class, since they’re impossible to clear through other means. On the other hand, the class probably can’t consist entirely of orphan works. You need someone to negotiate with, and orphan owners, by definition, can’t fill that role.
Copyright’s registration threshold complicates matters. It’s unclear whether the owners of unregistered copyrights would be proper members of the class (the Supreme Court will weigh in on this question in the coming term). Depending on your risk tolerance, you might not want to bother with a lawsuit at all for unregistered United States works; the worst you face is actual damages. This option may not be realistic, however, if it’s hard for you to tell whether a work is domestic or foreign (books are probably at the easier end of the spectrum here, since their bibliographic metadata is reasonably good).
Since your goal is a settlement that gives you a copyright license, your lawsuit will almost certainly be a copyright infringement suit in federal court. The question is what you do to trigger the copyright suit. You could be aggressive, actually making many copies of many different works. You could be conservative, merely suggesting that you might make a few copies of a few works. Or you could pick any point in between.
There’s a tough tradeoff here. The more aggressive you are, the stronger the copyright case against you. That’s good, because it creates a real “case or controversy” that will support a lawsuit. It’s bad because it opens you up to greater liability. Google threaded this needle by actually digitizing millions of books but only using them in ways for which it had a strong fair use defense. The settlement involves further uses of the same works, but not uses that Google was willing to make pre-settlement.
Your relationship to the plaintiff class representatives will require another balancing act. They should be hostile enough to sue you, and friendly enough to be willing to settle, but not so friendly that the settlement becomes collusive. Google managed this feat, in part, through initial clumsiness. By scanning books without even notifying copyright owners, it aroused enormous hostility in the publishing community. The actual lawsuits, however, were overseen by people imaginative enough to recognize the settlement possibilities and seize them. Ironically, it may take more planning for you to replicate this result deliberately.
One way of ducking this issue would be to announce your plans and then immediately file a declaratory judgment action against copyright owners as a class. A declaratory defendant class is an exotic piece of civil procedure, but not out of the question. The two hard pieces of this strategy are finding suitable class representatives (since the representatives won’t have volunteered themselves) and making sure that copyright owners have issued sufficiently specific threats to make the lawsuit ripe.
Courts insist that arms-length negotiation between skilled attorneys is essential for class-action settlements not to be thrown out as collusive. For this reason, it’s almost certainly not possible to go into the lawsuit with the end result prearranged. Helpful indicia of a genuinely adversarial lawsuit include compromises by both sides, a willingness to proceed to trial if negotiation fails, and real bargaining room.
Once everyone is in the same room and talking, it’s time to make a deal. The three central questions are what uses you’ll be allowed to make, what you’ll pay for the privilege, and how copyright owners can opt out.
Of course, the nature of the works and your goals will will drive the substantive copyright terms of the settlement. What makes sense for books may not work for abandonware. An important limit on what you can get is what copyright owners are willing to give you. You should almost certainly be able to get by with copyright licenses, rather than needing transfers of ownership.
The Google settlement shows that the allowed uses can include some of the core revenue-generating uses protected by copyright: making and distributing copies to the public for money. The two core programs in the settlement–individual book sales and all-you-can-eat subscriptions–both fit that mold. You shouldn’t feel constrained to mimic those two precise programs; they’re just reasonable ways of making use of copyrighted works.
Anything within copyright is potentially on the table. You will probably need to be specific about the exact contemplated uses, however. You won’t be able to public-domain everything, nor will you be able to take broad, open-ended rights without specific reference to the programs in which they’ll be used. If you need to leave open future options, consider creating a registry representing copyright owners, as Google did, which can negotiate the details with you later. (If you’re working with books, joining the Book Rights Registry as part of your settlement may take some finesse, since it’s not a copyright-owning plaintiff.)
Now for the hard part: the money. The Google settlement “works” because some active copyright owners like the deal’s financials. Their willingness to accept Google’s deal for themselves provides a reason to think it’s fair to the absent class members. Your deal, therefore, needs to be sweet enough to draw in some revenue-maximizing self-interested copyright owners.
Setting prices can be difficult, especially for new and unprecedented uses. The Google settlement dodges most of these issues, in three ways. First, it specifies most payments in terms of percentage shares, rather than fixed amounts. Second, it gives copyright owners control over individual retail prices. Third, it requires Google and the Registry to collaboratively set prices for the all-you-can-east subscription at a later date. These techniques avoid locking the settlement into one, possibly incorrect, pricing structure.
Note that now the Google settlement is on the books, its 63/37 revenue split may be a lodestar. Its mere existence creates a colorable argument that any split less favorable to copyright owners is automatically unfair to them. If you want to negotiate better terms for yourself, you may want to focus on non-price terms.
Open access advocates face a particularly difficult challenge here; many of the uses they’d like to encourage aren’t revenue-generating. Since the profit motive may not be motivating you as strongly, copyright-owner plaintiffs may want a greater say in setting prices. If your open-access principles make you unwilling to sell access to works, consider whether an advertising-based business model would be appropriate.
The Google settlement also provides for an “inclusion fee” of $60 for every book scanned before the opt-out date. You’ll want to minimize these up-front costs. One option is to negotiate no (or a lower) inclusion fee. Another is to scan only enough works to launch the lawsuit, and save the bulk of your efforts for after the settlement is approved.
The most expensive part of your settlement, ironically, could be paying the lawyers. The Google settlement took years to negotiate, and its legal fees are in the tens of millions. Your fees could be lower, since the Google settlement provides some of the groundwork–but the need to avoid collusion puts a floor on how quickly the negotiations could go. Even if you can get your own lawyers to volunteer their time, the plaintiffs’ lawyers are unlikely to be cheap, and the money to pay them will almost certainly need to come out of the settlement.
The Google settlement gives copyright owners a one-time chance to opt out from the settlement entirely. It also gives them a perpetual right to pull their books from any or all of Google’s uses. These opt-out rights defuse many objections to the settlement, since they allow copyright owners to protect their individual interests. If they don’t think the deal is good for them, they don’t need to take it.
It’s probably impossible for your settlement not to contain a similar opt-out provision. That means that you, like Google, need either to be able to remove works from your programs if their owners opt out or you need to be willing to bear any legal risk from using them (e.g. for indexing uses you consider fair). The granularity of the opt-out is a fair subject for negotiation. Copyright owners may demand to set prices and control the economic details of their participation; you may be able to require that they not participate in one of your programs unless they also participate in another.
Opt-out rights depend on proper notice. Pay close attention to the issues the Google notice program has faced. Large, worldwide classes require extraordinarily large and expensive notice programs.
Once the basic terms of the settlement are clear, you can think about details. They may or may not be essential to settlement approval. They give you flexibility, and may be useful in achieving some of your goals.
In the Google settlement, the Registry wears many hats. It helps channel payments to copyright owners. It provides tools for them to set prices or remove their books. It helps communicate important information to authors. It provides a clean database of information about copyright claims. And so on. Consider carefully whether your desired settlement has administrative functions that should be given to a dedicated entity. Think also about whether your settlement has a role for equivalents to the libraries: institutions that supply materials, hold copies, and/or have an ongoing role to play in providing access.
The Google settlement requires Google and the libraries that have digital copies to comply with industrial-strength security standards. Major copyright owners–fearful of their works escaping in unconstrained digital form–are likely to insist that you comply with similar security conditions.
Attachment A to the Google settlement–the “Author-Publisher Procedure”–is fascinating. In determining where the Registry should send payments from the Google programs, the settlement actually reallocates some copyright interests between authors and publishers. The procedures mimic standard industry patterns for assigning electronic publishing rights, but only up to a point. Uniform rules substitute for the precise details of every last publishing contract.
Consider whether you too might be able to solve some problems of rights fragmentation simply by changing the relevant rights. When you have all of the relevant rightsowners in the class-action, you can get them to strike side deals with each other to clean up the rights. You can also mimic the Google settlement by using the payment mechanism to help locate missing authors, or by empowering your equivalent to the Registry to collect rights and negotiate on behalf of copyright owners as a collective in the future.
The Google settlement includes extensive mediation and arbitration procedures. These function to reduce Google’s risk. They also lower the costs of the inevitable disputes that will arise about the meaning and implementation of the settlement.
Arbitration can also have a substantive face. Consider whether you might be able to use arbitration as a way of cleaning up rights. Similarly, think about using arbitration to resolve contentious questions. Google’s amended agreement with the University of Michigan uses a special-purpose arbitral body to review prices. Scholars have proposed special-purpose fair use administrative hearings; the Copyright Office already uses three-judge panels to set particular rates.
The Google settlement contains plenty of interesting uses that are ancillary to the main deal. The research corpus, for example, is both easy to create once Google has scans of enough books, and also not a major threat to copyright owners’ interests. It makes sense to offer it in addition to the main money-making programs. Google also negotiated many provisions–such as guarantees to libraries of certain specific rights under the institutional subscription, and the public access service that offers libraries one free full-access terminal–that were hardly required but offer public benefit.
Be creative in asking whether your deal could contain similar terms. A cost-conscious Google competitor might want to economize by omitting non-core programs; a munificent one could win a lot of public approbation by including new ones. Open-access advocates should consider whether some of their goals could be realized through programs that are a natural offshoot of a money-making deal. For example, perhaps building a universal, world-class clearinghouse for photography rights would create the conditions under which a version of the clearinghouse’s archive could be opened up to personal uses.
I’m not sure whether this exercise has been extraordinarily cynical or extraordinarily naive–or both at once. The Google settlement is still before the court, and it may be a long time before we know whether, and on what terms, it’s available as a precedent. Nonetheless, I learned three things from working through its terms as though in hindsight:
I’d like to close with one last, more radical thought. Is it now possible that the Google settlement creates a precedent for the governmental taking of old copyrights? Google has put in an offer to buy almost all of the economically significant interests in a whole class of copyrights–and thus put a price tag on them. We might well ask whether Google’s proposal really differs from a governmental seizure of unused copyrights, payment of just compensation, and transfer of them to Google. If others can obtain similar settlements, there’s a real sense in which the orphan works, at least, may be about to reenter the public domain.
August 4, 2009
Revision history:
This essay is licensed under a Creative Commons Attribution 3.0 United States License. It is canonically available at http://james.grimmelmann.net/essays/UnprecedentedPrecedent.
I welcome your comments, critiques, and corrections.