The Google Book Search Settlement

Topic Owners: Gwen, Lee, Jon

Topic Date: March 30

back to syllabus

Concrete Questions of the Week

How will the Google book digitization project affect various interests, including those who were parties to the settlement between Google and the Authors Guild/American Association of Publishers, and those who were not? In particular, what changes will libraries (public, private, and university) and readers face going forward, and how should they respond?

Brief Overview of the Google Book Search Settlement

History of the Settlement

In 2004, Google began to digitally scan books from the university library collections of Harvard, Stanford, Oxford, and the University of Michigan, as well as from the New York Public Library system. Google provided a service, Google Book Search (GBS), whereby users could search within these digitized books. The service allowed users to search and view the entire content of books in the public domain, but it also permitted searching copyrighted books, providing short "snippets" containing the search term and some surrounding text.

Crying foul, the Authors Guild, a writers' advocacy group, initiated a class action law suit on September 20, 2005 in the U.S. District Court for the Southern District of New York, claiming copyright infringement. The Association of American Publishers (AAP) followed with a suit of its own on October 19, 2005. Google responded that its actions were lawful under the doctrine of fair use. Rather than fight that battle, however, Google decided to settle with the Authors Guild and the AAP.

The parties announced a settlement on October 28, 2008. Judge Denny Chin is supervising the settlement process. Under the current schedule, authors have until September 4, 2009 to opt out of the settlement, and the court will hold a final hearing on October 7, 2009, after which it will accept or reject the settlement. On April 14, 2009, Professor Charles Nesson filed a motion to intervene in the lawsuit as counsel on behalf of two professors and the Open Access Trust, and on April 17, 2009, the Internet Archive, an online library, also sought to intervene, but Judge Chin denied both of these requests. On May 4, 2009, a number of library associations filed a brief requesting that the court "vigorously" supervise implementation of the settlement, should it be approved. Meanwhile, the Justice Department has reportedly commenced a review of the settlement's legality under antitrust law.

Content of the Settlement

The settlement creates an entirely new legal regime for book digitization. Under the settlement, Google will continue to offer GBS, but it will now have to pay authors and publishers for books still under copyright from the revenue it makes from advertising and selling access to the books. To facilitate payment, the settlement creates a new entity called the Book Rights Registry (BRR), an independent body chaired by an equal number of author and publisher representatives and charged with maintaining a database of book copyrights and implementing the terms of the settlement. Google will be able to continue to digitize books, free from fear of litigation (at least from those copyright owners who do not opt out of the settlement and who published a book prior to January 5, 2009). In return, 63% of revenue from advertising and book sales will be channeled to the BRR for distribution to rights owners. The BRR also possesses other powers. For instance, the BRR has some discretion over dividing revenues between publishers and authors, has some approval power over GBS’s security standards, will help set book prices, and can even license copyrights to third parties besides Google. The settlement is thus non-exclusive to Google, though it contains a “most favored nation clause” that requires the BRR to give Google at least as good terms as any other third-party for the 10 years after the settlement’s approval.

Users of GBS will continue to be able to search the contents of books, but instead of returning "snippets" for copyrighted books, search results will depend on the type of book. The settlement places each book in one of three categories. First are public domain books, which users will continue to be able to view in their entirety. It is estimated that about 20% of GBS books are in the public domain. Second are books that are in-copyright and commercially available, which means that they are available for sale new through a “customary channel of trade” (e.g. it is available new at Amazon). For these books, GBS will display only bibliographic information and "front material" (copyright page, table of contents, index, etc.). About 10% of GBS books are estimated to be in-copyright and commercially available. Third are books that are in-copyright but not commercially available, which account for an estimated 70% of GBS books, and thus make up the bulk of what is covered by the settlement. For these books, users will be able to view up to 20% of the book (with some restrictions). Rightsholders may choose to deviate from these default settings and individually set the amount of each of their books available for users to view. Institutions and individual users will also have the option of paying for permanent online access to the full content of digitized books. The initial price of the institutional subscription will be set with reference to the prices of products and services “comparable” to GBS and will vary based on the type of institution (e.g. whether it is a corporation, a library, or a government office) as well as how many people are members of the institution.

Libraries are also specifically addressed by the settlement. Each public library will be provided a single GBS terminal that will display the entire content of the Institutional Subscription Database (ISD), essentially a database comprised of all books that are in-copyright but not commercially available. Academic libraries will be allowed to have multiple terminals with such access, based on the number of full-time equivalent students enrolled at their respective schools. Institutions may also purchase subscriptions to the ISD. "Fully participating libraries" are given digital copies of any book scanned from their collections, as well as digital copies of books that are in their collection but were scanned from another library, provided that a sufficient proportion of their own collection has been digitized.

Implications of the Settlement

What does this settlement mean for libraries and the reading public? Many have hailed it both for improving access to knowledge by creating "the long dreamed of universal library" (or at least a "universal bookstore") and for providing more access than may have been permitted under fair use if the case had gone to trial (though some question that latter assessment). But is this settlement optimal for all interested groups? The ramifications of the settlement will affect not only those parties who participated in the negotiations (Google, authors, and publishers), but also libraries and the reading public, neither of whom had a direct voice in the settlement-drafting process. Each of these groups will likely face a different set of benefits and problems. With a focus on libraries (public, private, and academic), we aim to identify the main challenges that the Google digitization project will entail and develop ways of working within the bounds of the settlement to mitigate these concerns.

Substantive concerns generally fall within one of the following categories:

• Access: Are the permitted free uses of copyrighted books sufficient? Is the number of public terminals provided to libraries adequate? Are subscription costs for individuals, and especially institutions, going to be too high? Could more be done to increase access to the wealth of knowledge in GBS?
• Antitrust and Monopoly: Does Google's de facto monopoly of digitized book search violate antitrust law? Even if not, are there protections that could be implement to ensure that it does not misuse its monopoly? And what about the BRR's monopoly control over licensing digitized works? Will this harm authors or would-be competitors of GBS?
• Copyright: What are the substantive legal implications of the settlement for copyright law? Was Google right to settle rather than take the case to trial on a fair use defense? Does Google's settlement make it more difficult for competitors to rely on a fair use defense? And what consequences will the settlement have for orphan works? Are Google and the BRR reaping profits from orphan works illegally or unfairly, as some have argued? And how will the BRR's copyright database help future licensees track down copyright owners? Will the database be openly available and readily updated?
• International Implications: What impact will the settlement have on foreign works in the U.S.? Will foreign authors be exploited, as some have suggested? And what access, if any, will located outside of the U.S. have to GBS?
• Privacy and Security: Will GBS monitor users' reading habits? If not, what security will it provide so that others are not able to determine this information? If so, what protections will Google provide readers so that the information is not used or disseminated in harmful ways?
• User Interface: Could the settlement provide better or more imaginative options for readers, such as better annotation or modification features? Does GBS's scanning fragment works in harmful ways? And how will Google choose which books and which editions to show at the top of GBS searches? Could the settlement have done more to resolve these issues?

Outline of Class Plan

We will be working from the assumption that the current proposed settlement will be approved by the court. Within the framework that this settlement establishes, then, how can we mitigate the concerns that have been raised by libraries and the reading public? We hope to explore the contours of these problems and produce an array of possible solutions.

Class time will be segmented as follows:

• The class will begin with a short overview of the settlement itself, provided by the student presenters responsible for this class session.
• Then, each of our three guests will be given ten minutes to present his views on the settlement and begin to answer the question posed above.
• After our guests have introduced their perspectives on the settlement, we will follow up with questions we have prepared for them. These questions are designed to ensure that all important aspects of the settlement and its expected effects are brought up.
• Finally, we will open the session to class discussion, guided by the Berkman Question Tool (see Class Participation). The discussion will be focused on normative suggestions for how institutions and readers should respond to the settlement in the wake of its (presumed) approval by the court.

Guests

Three guests will be participating in our class session. Two of them will be joining us live, and one will join us over videoconference. All three guests will be present for the duration of the class session.

• Jeffrey P. Cunard (via videoconference), managing partner of the Washington, D.C. office of the law firm Debevoise & Plimpton LLP.
• John G. Palfrey, Henry N. Ess III Professor of Law at Harvard Law School, Vice Dean of Library and Information Resources at Harvard Law School, and Faculty Co-Director of the Berkman Center for Internet and Society. (Blog located here.)
• Allan A. Ryan, Director of Intellectual Property at Harvard Business Publishing.

Class Participation

We will be using our old friend, the Berkman Question Tool! (The instance name for our class session is IIFGBS, and it is located here.) Before the day of our class session, we will seed the question tool with some questions that we feel are particularly relevant and topical. We encourage everyone to take a look at these questions prior to the class and to make comments, add more questions, and vote up the questions you find most interesting. We will be using this list of questions to guide the class discussion, especially in the final portion of the class.

Readings

Assigned Readings

• This summary of the settlement, prepared by Jonathan Band on behalf of the Association of Research Libraries and the American Library Association.

• These selected provisions from the settlement agreement.

• This article by James Grimmelmann, outlining potential problems with and revisions to the settlement.

• This New York Review of Books article by Robert Darnton criticizing the settlement. (Or, for those who prefer to listen rather than read, hear Darnton's criticisms on National Public Radio.)

• This letter in response to Darnton by Paul Courant (head of the University of Michigan's library, a participant in Google Book Search), challenging Darnton's criticism of the settlement.

• This New York Review of Books piece by Darnton in response to Courant's challenges. (Skip the published version of Courant's letter here, which is less detailed than the one posted on his blog at the link above, and just read Darnton's response at the bottom of the page.)

Optional Readings

• Experiment with the current incarnation of Google Book Search.

• Skim the settlement agreement, focusing on Article 4 (Economic Terms for Google Book Search) and Article 7 (Obligations and Rights of Participating Libraries).

• For the highly ambitious, Walt Crawford of Cites and Insights has compiled this 30-page newsletter (PDF), which includes excerpts from an enormous number of blog posts, accompanied by his own commentary. It covers the landscape up until late February.

• This Economist piece introduces Brewster Kahle and his Internet Archive as an alternative or supplement to Google Book Search.

Further Research

• The full settlement, including all of its attachments, is available here.

• Timothy Vollmer of pureinformation.org has compiled a comprehensive and regularly updated list of links to articles and blog posts about the settlement.

Class Session Recap

The class session was dominated largely by discussion among our three guests, punctuated by questions from the class. What follows below is a thematic summary of the views that were expressed during this discussion. For a more detailed account of the class session (including diachronic notes on the guests' presentations, the live discussion, and the online discussion, as well as lists of our prepared questions), see our edited class notes.

A Private Settlement With Benefits for Everyone

Despite the many concerns (e.g. Darnton's) and settlement amendment proposals (e.g. Grimmelmann's) raised in newspapers and the blogosphere, it is important to remember that this settlement is the (presumptive) result of actual litigation between specific parties. The parties have taken care to incorporate other interests into the negotiation process, but at its core the settlement is designed to satisfy the concerns of publishers and authors that led them to sue Google in the first place. Nonetheless, major libraries had a presence during the negotiations, participating both directly and through Google. The public interest was also taken into account in a variety of ways, culminating in the provision of public access terminals to all public and academic libraries and substantial access provisions for the visually impaired. It should also be noted that copyright holders, as members of the reading public themselves, had a stake in ensuring that the public interest was not ignored during the negotiations. All three guests agreed that the settlement represents a substantial improvement over the status quo for everyone affected. Professor Fisher, however, suggested that "an ideal dream scenario" would be a more useful basis of comparison than the status quo, and once such an ideal was described, the settlement was found to fall short of the ideal.

Publishers and Authors: Upholding the Copyright Regime

From the publishers’ perspective, the litigation was essentially concerned with ensuring that copyright remains a permissions-based regime. When Google decided to start digitizing books on a massive scale, it did so without attempting to secure permission from rights holders; this move was a bold challenge to the existing regime. Publishers were also concerned about potential repercussions from Google’s practice of providing libraries with digital copies of the books scanned from their collections. Library users have special exemptions under the Copyright Act to copy books in a library’s collection, and these exemptions were carved out during an era where photocopying a book, page by page, was the only means of producing such a copy. In today’s digital era, this exemption for library users, coupled with the ease of online library "use" resulting from mass digitization, could lead to much more widespread copying of books in library collections without any revenue being generated for the rights holders. It was these concerns that motivated publishers and authors to bring suit against Google.

Libraries: Concerned About Resource Strain and the Proprietary Database

Although the settlement would result in a net increase in public access to information, there are causes for concern from the perspective of libraries. The provision of a free public access terminal to every library represents a boon to small-town local libraries, but it likely would act as a strain on a larger libraries with more user demand, such as the New York Public Library system. The digital book database that Google is amassing would be an invaluable corpus on which to perform various types of research, and research libraries are rightfully concerned about restrictions on the use of this information. Potential problems also arise from the concentration of this corpus in the hands of a single (or small handful of) powerful private actor(s). The issue of concentrated private power over vast information is particularly alarming because the settlement effectively precludes any potential competitors from offering a comparable service in the near future.

The Settlement Mechanism as a Problem-Solving Tool

Additionally, considerations of institutional competence were raised. The effects of the settlement on copyright doctrine -- and in particular, the fate of orphan works -- are uncertain but have the potential to be far-reaching. Are broad legal issues that so fundamentally affect the public interest best decided through private litigation, or should Congress step in? Is it wise to establish such a rigid, comprehensive system for the digital book realm before we, as a society, have a better idea of how we want to structure our digital world?

Concrete Normative Suggestions

• Professor Nesson rejected our premise that the settlement will be accepted in its current form. He views the threats posed by this settlement to digital freedom as urgently severe and alarming; thus, he advocates active intervention in the settlement process itself. Indeed, following our class, Professor Nesson moved to intervene as counsel on behalf of professors Lewis Hyde and Harry Lewis and the Open Access Trust, although the request was denied.
• Professor Zittrain sees the "golden copies" of digitized books that Google provides to participating libraries as holding great potential as instruments of change. He suggests that libraries cooperate in a large-scale creative use of these copies.
• Professor Fisher emphasized the importance of understanding the problem on the big-picture level of legal theory and social goals before electing a particular course of action by a particular institution.
• Professor Palfrey suggests three generalized improvements to the settlement that would begin to address many of the concerns that have been raised:
  • Ensure the possibility of a meaningful competitive landscape, such that second-comers are not barred from success.
  • Establish a means by which the public can have a meaningful level of control over the workings of the Book Rights Registry.
  • Create a system of periodic review for the settlement terms. This system would not need to involve periodic wholesale review of the entire settlement by the courts; it could instead merely involve libraries negotiating sunset provisions for individual works with publishers, authors, and other rights holders.

Teacher's Guide

This section is intended to provide guidance to anyone interested in teaching about the Google Book Search settlement in the future. Anyone who has taught on this topic is encouraged to use this wiki page to share additional wisdom gained from the experience. The development of our plan for this class session (including our selection of the Google Books Settlement as a focus within the broader topic of the internet and publication, as well as our decisions about guests, readings, and technology use) is documented on the discussion page.

Evaluation of the Class

Topic Breadth and Time Constraints

The Google Book Settlement is a large, diverse, and contentious topic. Even with our efforts to limit the scope of our discussion by (a) assuming the settlement would be approved and (b) focusing primarily on concerns raised by libraries and the reading public, two hours was enough time only to scratch the surface. One way to help alleviate this difficulty would be to focus the topic even more narrowly, albeit at the cost of precluding discussion on subjects that students might find more interesting. The most significant factor contributing to the palpable time constraint, however, is the sheer complexity of the settlement itself. At over 140 pages plus appendices, the settlement contains a lot of material that must be digested before potential concerns may be formulated and discussed. Band's summary does a good job of explaining the most relevant portions of the settlement, and Grimmelmann's article explains major concerns in the context of the settlement, but neither piece addresses all the relevant details. As a result, significant time was spent with Mr. Cunard answering specific questions on the terms of the settlement before the class could proceed to a more high-level discussion about concerns and potential solutions. Additionally, while each of our guests brought his own unique and insightful perspective to the settlement, we found two hours to be insufficient to appropriately accommodate three guests.

Reading Assignments

The amount of reading we assigned seemed proportional to a two-hour class session, and for the most part, the content of the readings was relevant to our discussion. We sought to provide a descriptive introduction of what the settlement encompasses and what issues it raises, as well as an introduction to the normative implications of the settlement. For descriptive pieces, we chose not to assign the entire settlement given its length and technical legal language (which would have been particularly difficult for the non-law students in our class). Instead, we assigned Band's piece to provide a readable summary of the settlement, some selected examples of provisions in the settlement to give a taste of the actual settlement, and Grimmelmann's primer on the chief issues raised by the settlement. As to normative pieces, we assigned what we felt were the most provocative (and the most discussed) pieces on the settlement: the Darnton and Courant articles, plus Darnton's response to Courant. We felt the readings all served their purpose well during the class (the students seemed relatively well informed of the descriptive aspects of the settlement and its normative implications), with the sole exception being the actual settlement provisions that we assigned but whose language we did not examine closely during class. We felt that it would be useful to have students--especially law students--glimpse what the settlement's actual language looked like, but given our time constraints, we were unable to delve into the particular language and whether it would implement the settlement in the manner that the various commentators expect it to.

Discussion Framework and Presence of Guests

A substantial portion of the class discussion focused on the question of whether the pre-settlement status quo was the appropriate comparison against which the settlement should be judged. One specific alternative “dream scenario” was suggested as a potential foil to the settlement, but it seems that more could have been done with this approach. A more lively discussion of alternatives to the settlement might have emerged had the guests only been present for the first hour of the class. Planning for a portion of class discussion to take place outside the presence of the guests would have prevented monopolization of the conversation by the guests, thereby allowing for more student input and a greater diversity of ideas and proposals. In particular, by removing the “settlement expert” after a specified period of time, the amount of time spent asking detailed descriptive questions about the settlement could have been reduced, allowing more time for an open-ended normative discussion of concerns with the settlement and potential solutions. Of course, if fundamental questions regarding the mechanics of the settlement remained or arose after the guest's departure, an informed discussion of concerns and solutions would become more difficult. In order for a closed class discussion to be productive, those leading the discussion would need to have a firm grasp of the major points of the settlement and be able to act as a fall-back experts on descriptive matters once the guests have left.

Use of Technology

This class was served well by a minimal use of technology. The Berkman Question Tool was useful in providing a means for backchannel discussion while not being too distracting. It also provided fall-back questions to use if there was a lull in the discussion, with student voting (presumably) signaling which questions and topics the class found most interesting. Videoconferencing was an effective way of communicating with a remote speaker. However, the camera placement was not ideal, as it gave the remote guest a view of students' backs.

Suggestions for Future Iterations

• Topic Management - Even though our topic was narrowed to addressing the concerns of libraries and the reading public under the assumption that the proposed settlement agreement will take effect, we were able only to scratch the surface of this topic in the two hours allotted. Structuring the class around a more focused topic (perhaps looking at only one or two particular concerns) would likely allow for a deeper and more substantial discussion. As the topic becomes more focused, however, it may be more difficult to ensure that all students are interested and feel competent to contribute to the discussion. A class specifically addressing antitrust concerns, for example, would likely have alienated many of the non-law students in the class.

• Readings Management - Future iterations of this class would do well to use the readings we assigned, as they are comprehensive, yet readable. Furthermore, they are currently the pieces most mentioned by those discussing the settlement, so knowledge of them is crucial for being able to participate in the discussion. The one exception is the actual settlement provisions that we assigned. Careful thought should be put into assigning portions of the settlement agreement itself as reading. Any such settlement excerpts should be selected deliberately with an eye to how they will be used in class. Parsing the language of settlement terms might alienate any non-law students in a class, so the composition of student backgrounds should be taken into account in deciding whether to assign reading from the settlement agreement. Although a glimpse of what the actual settlement looks like can be educational, it could probably be done without. One further caution is that the settlement is part of a constantly changing landscape (see our summary of the settlement's current status), so future classes on the subject will need to make sure they are up to date on developments in status of the settlement -- for instance, whether it has been modified, accepted, or rejected. If a major change does occur, then substitute readings may be needed.

• Guest Management - Potential ideas of ways to help foster a lively class discussion that is neither dominated by the guests nor stuck on purely descriptive questions include the following:
  • Fewer guests. While each guest had a unique perspective on the settlement, more guests means less time for class discussion and student input, and this trade-off is especially problematic given the complexity of the settlement.
  • Pre-recorded material. In order to better control timing and ensure that only the most relevant and interesting guest statements are presented to the class, interviews with the guests and/or prepared statements by them could be recorded and edited. This edited final version would then be shown to the class instead of, or in conjunction with, live guests. Recorded interviews would also enable the presentation of guests who are not available on the day of the class session; for example, our class might have benefited from a recorded interview with Robert Darnton, who was enthusiastic about participating in the discussion but was unable to join us in real time because of a schedule conflict.
  • Time limits on guests. Both to address timing issues and also to lead students away from descriptive questions and toward more normative discussion, the amount of time for which the guests are available could be limited. One possible arrangement would be to have the guests available for introductions and questions during the first hour, and then to have student discussion proceed during the second hour without the guests present.

• Technology Management - Technology might be used more effectively in the following ways:
  • Interest polling for planning purposes. The Berkman Question tool, or even an online poll, might have been used earlier in the planning process in order to determine which of the major issues the class found most interesting. This technique could have alleviated some of the problems mentioned under "topic management" above, by allowing for a more narrowly focused class while ensuring that at least a plurality of students would be interested in the discussion.
  • Videoconferencing details. Appropriate camera placement should be considered in advance, in order to avoid awkward arrangements such as we experienced. Videoconferencing also offers a convenient way of limiting the period of time during which the guests are available, whereas ushering out live guests would be more difficult.