Plaintext
INTERNET LAW
JO U R N A L O F
VOLUME 12
NUMBER 10
APRIL 2009
EDITED BY DLA PIPER
How to Fix the Google Book
Search Settlement
By James Grimmelmann To keep the Registry from overreaching, we’ll
T
need more checks than the settlement currently
he proposed settlement in the Google Book Search contains.
case should be approved with strings attached.1 The 2. Google poses an antitrust threat. By virtue of the
project will be immensely good for society, and the settlement, Google will have the book search market
proposed deal is a fair one for Google, for authors, and the download market for orphan works largely
and for publishers. The public interest demands, how- to itself. The structure of a class action settlement,
ever, that the settlement be modified first. It creates two in itself, is a highly effective barrier to entry. We’ll
new entities—the Books Rights Registry Leviathan and need similar checks on Google’s actions to keep it
the Google Book Search Behemoth—with dangerously from acting anticompetitively in book markets.
concentrated power over the publishing industry. Left 3. Consumers need protection. The settlement as it
unchecked, they could trample on consumers in any stands is often good at ensuring minimum standards
number of ways. We the public have a right to demand to benefit libraries, institutions, and consumers.
that those entities be subject to healthy, pro-competitive But in some areas, such as price discrimination and
oversight, and so we should. privacy, the settlement leaves the door open for
After laying out the basics of the proposed settle- Google to behave oppressively. We’ll need to close
ment agreement and explaining how it will benefit that door.
both the parties to the lawsuit and society more broadly, 4. Public goods should be widely available. Just by
I’ll work through the implications of five overriding providing the search and download services and
principles:
Continued on page 11
1. The Registry poses an antitrust threat. The Book
Rights Registry will be a new collecting society
representing the interests of authors and publish- IN THIS ISSUE
ers. This kind of collective action poses an obvious
HOW TO FIX THE GOOGLE BOOK
antitrust risk: What if the Registry becomes the SEARCH SETTLEMENT . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
instrument of a cartel to fix the price of books? By James Grimmelmann
LAW IN VIRTUAL WORLDS. . . . . . . . . . . . . . . . . . . . . . . . . 3
By Rónán Kennedy
James Grimmelmann is Associate Professor of Law, New York Law
School. An earlier version of this article was originally published on the INTERNET LAW IN THE COURTS . . . . . . . . . . . . . . . . . . 21
Laboratorium weblog at http://laboratorium.net/archive/2008/11/08/ By Evan Brown and Brian Beckham
principles_and_recommendations_for_the_google_book, on
November 8, 2008. Aislinn Black, Joseph Gratz, Joris van Hoboken,
Fred von Lohmann, Alexander MacGillivray, Dominic Mauro, and Derek
Slater gave helpful comments. This article is available for reuse under
the Creative Commons Attribution 3.0 United States license, http://
creativecommons.org/licenses/by/3.0/us/.
April 2009 J O U R N A L O F I N T E R N E T L AW
How to Fix the Google Book Search Settlement than many authors and publishers would like; they see
Continued from page 1 Google as a trespasser willing to drive through the yards
of people who aren’t at home.
by processing payments, Google and the Registry The proposed settlement is complex, but its central
will assemble some immensely useful databases about deal is simple. Looking backward, Google will be released
book copyright information. These databases are clas- from liability for its scanning, searching, and displaying2
sic public goods, and neither Google nor the Registry in exchange for a set of one-time payments totaling about
will need exclusive rights over them as an incentive. $125 million.3 Looking forward, it will be allowed to
We’ll need to make sure that those databases are continue scanning and displaying books4 in exchange for
made available to the public. 63 percent of its net revenues from the advertising that it
5. Transparency and accountability matter. If Google shows on search results and book display pages.5 This deal
becomes a chokepoint for getting books to the public, is most significant for out-of-print and orphan works: It
there’s a risk that it could secretly censor. Google, of effectively lets Google go ahead with making substantial
course, doesn’t want to distribute books that it strong- uses of these works, while holding a share of the revenues
ly disagrees with (or fears legal liability from). These in trust for the copyright owners.
two goals can be reconciled. We’ll need to make sure Of course, many of the copyright owners who’ll be
that when Google chooses not to make books avail- paid as part of the settlement don’t currently have deals
able, it leaves open suitable alternative channels for with Google. A new Book Rights Registry will be in
getting them to the public. charge of mediating all these relationships.6 It will take
payments from Google and pass them along to the appro-
These five principles give rise to a number of specific priate authors and publishers. In the other direction, it
recommendations. Each proposed change is incremental will communicate to Google copyright owners’ requests
and fully consistent with the settlement’s overall goals for the levels of access allowed to their books.
and design. The settlement also authorizes some exciting new
programs. An Institutional Subscription system will
THE SETTLEMENT allow Google to sell all-you-can-eat digital access to
the entire catalog of scanned books to companies,
First, a little history. Google Book Search is actually colleges, and potentially even individuals.7 A Public
two related programs. The Partner Program works with Access Service will provide a restricted version of the
books that it receives from publishers (either physically Institutional Subscription to colleges and public librar-
or as PDFs). Meanwhile, the Library Project has been ies for free.8 Meanwhile, a Consumer Purchase system
scanning books en masse from the collections of uni- will allow Google to sell electronic access to complete
versity and civic libraries. Either way, Google folds the individual books.9 Copyright owners can set their sale
scans into its gigantic Book Search index. Enter a search price or delegate that decision to Google’s algorithmic
term, and Google will show you a list of books contain- whizzes.10 The revenues from these new programs, along
ing the term. with other new revenue models worked out by Google
What happens when you click on a result depends and the Registry, will also flow to copyright owners in
on where the book came from. For public-domain books the same 63/37 split.11 A section on “Non-Consumptive
from the Library Project, you go directly to a page where Research” also allows researchers to run gigantic auto-
you can browse the book online or download a PDF copy. mated statistical studies on the entire corpus of scanned
For books from the Partner Program, Google lets the pub- books.12 Copyright owners won’t be paid for these uses
lisher decide how much users can see. Generous publishers on the theory that they don’t involve people “read-
enable extensive previewing of sample chapters; stingy ing portions of a Book to understand [its] intellectual
publishers allow you to see only a snippet of a few lines content.”13
containing the search term; paranoid publishers prohibit Meanwhile, the libraries that have been supplying
even that. books for scanning aren’t actually parties to the lawsuit.
This leaves open a large middle ground: in-copyright The settlement, however, contains provisions that enable
books from the Library Project. Most of these books are them to sign agreements with the Registry that effectively
out of print; many of them are “orphan works” for which release them from liability.14 Libraries that get back digi-
no copyright owner can be found. Google’s policy has tal copies of their books from Google must agree to keep
been to make these books searchable and available in those copies secure and limit access;15 libraries that don’t
snippets. That may not seem like much, but it’s still more have few other obligations.16
11
J O U R N A L O F I N T E R N E T L AW April 2009
THE SETTLEMENT SHOULD comprehensive, however, that it gives us perhaps 80 or 90
B E A P P R OV E D percent of the actual uses of books that a positive fair use
finding would have enabled. Those who favor a judicial
My starting point is that the settlement is a good showdown on the fair use issue would do better to join
thing. Everyone is better off than they would be in a world that battle in some other case, one in which the equities
without Google Book Search: tip more clearly in favor of fair use.
Second, even if scholars and observers would like
• Google makes piles of money from selling ads, sub- a fair use fight, it’s not our call to make. It’s Google’s.
scriptions, and e-books. Google was the defendant; it earned that dubious privilege
• Authors and publishers get the lion’s share of that by actually scanning and searching books. Having stepped
money. up to the plate to risk a lawsuit, and having been beaned
• The publishing industry benefits from the incidental with one, Google now has the right to choose whether to
cleanup of copyright records. settle that suit. Google’s choice to settle takes away no
• Universities, schools, companies, and other institu- legal rights from anyone else; one else loses the fair use
tions can subscribe to Google’s convenient fire hose argument because Google didn’t chance it.
of books. I’ve seen people raise the argument that Google’s
• Public and college libraries get free access to the fire capitulation means that there’s now a functioning licens-
hose. ing market for searching books. If true, this fact would
• The libraries participating in scanning books get digi- undercut fair use claims by Google’s competitors, since
tal copies of the scanned books from their collections. it would imply that scanning and searching without pay-
• Individual readers get an increasingly comprehensive ment take away revenue that copyright holders could have
book search engine. realized. The critical hole in this argument is that this isn’t
• Readers also get free PDF access to millions upon a market that one can effectively negotiate in without the
millions of public-domain books and two paid-but- device of the class-action lawsuit. Even after the settle-
convenient digital sources for in-copyright books. ment, there will still be far too many potential plaintiffs
• The public as a whole gets a substantial leg up on for any competitor to be able to cleanly license all the
solving the orphan-works problem. rights that it needs to start large-scale scanning. It’s also
• Researchers running automated studies can advance significant that Google won’t pay royalties for scanning or
human knowledge on algorithms, natural language, searching under the settlement; the paid uses only start
the history of publishing, and other topics. once a user clicks through from a search results page to a
page with content from a specific book.17 If the settlement
These are serious benefits, and the settlement is a creates a distinct market for scanning and searching, it’s
universal win compared with the status quo. Still, some a market in which these uses are worthless to copyright
commentators dislike the settlement because they that owners.
think we’d be even better off if the case proceeded to trial Third, given that Google and the copyright owners
and judgment. They’d been hoping that Google would are the parties here, there’s no way to force the court to
have established that its scanning and searching features consider the fair use issue. The adversarial legal process
were fair uses. If Google had prevailed on the fair use issue, allows parties to present their cases and controversies to a
it would have opened the book search business to anyone, judge for resolution. It doesn’t generally allow outsiders to
free from legal taint. What’s more, it would have given us compel the parties to litigate issues not of their choosing.
a powerful, portable fair use principle that could do a lot of This fact means that, to the extent that the rest of us want
other good in this digital age. While a definitive finding of the court to modify the settlement or even to think about
fair use would have been better than the settlement, that’s particular issues, we need to find a hook to put them prop-
not the choice on the table. There are three good reasons erly before the court. There are some such hooks, but the
that we should settle for the settlement. fair use question isn’t one of them. There is no convincing
First, there’s no guarantee that Google would have way to categorize that specific copyright issue as part of the
won on the fair use issue. A finding against fair use, general public interest that the court is directed to take
especially given the possibility that the Supreme Court into account in approving a settlement.
would have granted certiorari in such a socially signifi- Thus, I start from a baseline of believing that the
cant case, could have done immense damage to other settlement should be approved. It makes all of us better
large-scale projects making beneficial but incidental uses off, and it is, in an intuitive and meaningful sense, quite
of a great many copyrighted works. The settlement is so fair to all involved. I have concerns and critiques, but I see
12
April 2009 J O U R N A L O F I N T E R N E T L AW
them as patches to make the settlement better, not as do- books.) Similarly, its board is equally divided between
or-die clauses that the settlement must absolutely contain. authors and publishers, with any action requiring a
My suggestions flow from the principles identified earlier. majority.23
Further, the settlement ensures that the licenses that
P R I N C I P L E 1 : T H E R E G I S T RY P O S E S copyright owners grant to participate are non-exclusive;
A N A N T I T R U S T T H R E AT they can strike side deals with anyone they like.24 They’re
guaranteed more than two years to decide whether to
The basic deal embodied in the settlement—authors remove their books from the program entirely.25 Even after
and publishers who don’t opt out will receive payments out that, they can opt out of almost everything beyond basic
of Google’s revenues—will create a huge administrative searchability, in which Google will tell users what page of
workload. Someone needs to maintain a database of who a book contains a search term, but won’t show them any
owns which copyrights, mediate in ownership disputes, of the actual text around the term’s appearance on that
process payments, audit Google, and so on. Although page.26 Even within the Consumer Purchase system, indi-
collecting societies currently do many of these jobs for vidual copyright owners are always free to set the price for
other kinds of works, such as ASCAP for musical works e-books sold through Google.
and SoundExchange for sound recordings, the United These facts keep the Registry from acting like a classic
States doesn’t currently have a suitable similar organiza- price-fixing cartel; individual publishers can easily defect
tion for books. Nor can these tasks be trusted to Google, and charge less (or more). It’s true that the Registry could
which would have an obvious conflict of interest if it were facilitate coordination, but that’s not a big issue in the
charged with tracking down AWOL authors or being its publishing industry. Publishers can already see what each
own auditor. Thus, the settlement agreement establishes a others’ suggested prices are by looking on Amazon.
new collecting society, the Book Rights Registry.18 The Registry, however, will be an anticompetitive
The Registry doesn’t just have ministerial tasks, threat for the same reason that this lawsuit will transform
though. It also has substantial authority to negotiate on publishing: the settlement’s class-binding effect. If Google
behalf of authors and publishers. It has approval power would like to negotiate, say, an encryption standard and
over the security standards that Google and the various DRM terms for book downloads, without the Registry, it
libraries must create and live by.19 It has broad discretion needs to negotiate one-on-one with authors and publish-
to work out an equitable formula for dividing revenues ers. But the Registry is authorized to negotiate on their
among publishers and authors.20 It can even negotiate the behalf, all of their behalf. It could agree with Google on a
terms of new revenue models (e.g., print-on-demand, PDF privacy-intrusive DRM standard that fed back usage infor-
downloads, and coursepacks).21 mation into a database used to do industry-wide price-
If your antitrust sensors aren’t pinging wildly at this fixing in the guise of price discrimination. Other examples
point, please make sure that they’re properly calibrated. abound; the Registry’s centralized negotiating role permits
The Registry is a centralized entity with the authority to various anti-competitive practices to be coordinated and
negotiate on behalf of all registered book copyright own- laundered. What should we do about these risks?
ers. As such, it walks and quacks like a cartel. There’s a
reason that ASCAP and BMI, which play similar roles for • Antitrust consent decree. The Registry should be
musical compositions, live under antitrust consent decrees; under ongoing antitrust supervision from the day of
otherwise, they could make or break radio stations and its birth. The Department of Justice (DoJ) should
musicians and drive up the effective price of music. Were require that it negotiate and sign an antitrust con-
all authors and publishers (i.e., the plaintiff class, more sent decree. That decree would enumerate various
or less) to sign a piece of paper giving the Registry these forbidden anticompetitive practices, including those
powers, it could barely lift a finger without violating the called out in the settlement, ones pertaining to hid-
Sherman Act. That fact doesn’t change just because the den price-fixing, and whatever else the experts in the
agreement is the result of a class action settlement rather Antitrust Division think necessary to add. In addi-
than a meeting in the world’s largest smoke-filled room. tion, the DoJ should be given the authority to review
The settlement recognizes this danger, and accord- all contracts entered into by the Registry and reject
ingly puts some very important limits on the Registry. It’s any with anticompetitive effect.
specifically prohibited from representing any subgroup • Nondiscrimination among copyright owners. For
of copyright owners; it has to act in all their interest.22 understandable but regrettable jurisdictional rea-
(This restriction keeps it from being used by one group sons, the plaintiff class doesn’t include owners of
of authors, say, to suppress the market for another group’s unregistered copyrights, nor does it include future
13
J O U R N A L O F I N T E R N E T L AW April 2009
authors. Only currently registered copyright owners version of the fire hose. The settlement just sets up a series
are part of the lawsuit, which means that only they of deals with Google; it leaves everything else open for all
can be part of the settlement.27 The most immediate comers.
danger is that the Registry might adopt policies that The problem with this argument is that you can’t
operate to the benefit of past authors and against actually just go out and do what Google has done. One of
future authors (policies with narrow views of fair use Google’s fair use arguments would have been the insane
and broad views of a derivative works right come to transaction costs of trying to negotiate with every pos-
mind). There’s an easy way out, which is that the sible copyright claimant, particularly for out-of-print and
Registry must be explicitly required to represent any orphaned books. The settlement gives Google a clean
copyright owners who agree to its standard deal and release from the transaction-cost madness. All those pesky
explicitly forbidden from offering non-plaintiff and claims from authors who can’t be found or won’t play ball
future copyright owners any materially different deal. just go away.
This pair of rules guarantees to anyone who comes Consider the unappetizing options facing a would-be
along in the future effectively the same opt-out right competitor like Yahoo! If it goes ahead and starts doing
that the settlement class enjoys, while offering them large-scale scanning, it’ll get sued just the way Google did,
exactly the same terms if they choose not to opt out. but there’s no guarantee that the plaintiffs there would
This way, the Registry will truly be a fair, impartial feel any interest in settling on terms comparable to the
representative of all authors and publishers. ones that Google got. Indeed, as long as there were poten-
• Library and reader representation at the registry. tial plaintiffs out there, Yahoo! couldn’t feel safe, even if
Beyond that, the Registry’s structural protections it had struck agreements with 99 percent of them. The
should be supplemented to include voices from others could still pull enough of copyright’s harsh remedial
outside the publishing industry. It’s acceptable and levers to scotch the whole enterprise.
understandable for the Registry’s charter to require No, Yahoo! would need the same magic device of
that at least one author representative and at least the class action that Google is now taking advantage of.
one publisher representative consent to any action Would the plaintiffs bother to organize themselves as a
that it takes, but that veto principle doesn’t require class for its benefit? There’s no guarantee they would.
that they be the only members of its board. It should Yahoo! would be in the especially tricky situation of filing
also contain members representing libraries and the a declaratory judgment action against a class of copyright
reading public. In addition to objecting if the Registry owner defendants. It’d be hard even just to pick proper
takes anti-competitive, anti-reader actions, these class representatives and appoint appropriate class counsel
additional members would be able to monitor the without some kind of collusion. And once there were class
Registry’s actions, bringing important transparency to representatives, would they settle on comparable terms?
this new quarter-ton gorilla of the book industry. There’s no guarantee of it, especially given the guaranteed
good deal that they’re getting from Google.
PRINCIPLE 2: GOOGLE POSES I’ve also heard floated the idea that competitors are
A N A N T I T R U S T T H R E AT perfectly free to lobby Congress on orphan works legisla-
tion. So they are, but the argument that lobbying is an
The Registry isn’t the only entity that’ll have market acceptable substitute for free competition in the book
power as a result of the settlement. Google will too. It’ll market as it currently exists is laughable. If I manufacture
become the only game in town for scanning and searching widgets and my competitor is monopolizing the widget
books on anything resembling this scale. Yes, it was the market, it’s no answer to my pleas to say that I can ask
only game in town on this scale before, but that was when Congress for widget subsidies. Orphan works legislation,
there was a legal threat hanging over it. Now Google will done right, would be a great thing. But no one should
have the legal okay to go full-steam ahead, with some have to count on it happening as a condition of entry to a
exceedingly tasty markets all to itself. market that Google is already in.
The immediate rejoinder from Google, of course, is Thus, Google’s first-past-the-post status here could
that these markets aren’t closed to entry. Microsoft used easily turn into a durable monopoly. That might be the
to have a book-scanning program; it could have one inevitable result anyway; this is a market with substantial
again. There’s nothing in the settlement to prevent any- economies of scale and positive network effects. One may
one involved from doing side deals with others. Authors or may not think that a book-search and distribution
could license Yahoo! to scan, index, and sell. The Registry monopoly built on such structural bases is legitimate; one
could split the take. Libraries could subscribe to Yahoo!’s may or may not favor government intervention if it just so
14
April 2009 J O U R N A L O F I N T E R N E T L AW
happens that Google is the only player in this game. I take Google or by a competitor—the previous rule would
no position on these questions. But the court reviewing kick in and require that the new project license be
this settlement should not set up its own power—in the available to anyone on nondiscriminatory terms.
form of its ability to bind absent class members—as a bar- • Beware a scanning monopoly. If Google remains the
rier to entry in the online-books and book-search markets. dominant player in actually scanning books, we ought
What, then, should the court do? to be concerned about steps that it takes to preserve
or extend that monopoly.30 This concern is especially
• No most-favored-nation clause. The most pressing acute for public-domain books, which ought not to
problem is that the settlement explicitly guarantees be under anyone’s exclusive control. Google’s Web
Google a privileged position, via a most-favored- search engine has the right model: It returns results
nation clause in Google’s favor. For 10 years, the from Google-run sites alongside results from sites run
Registry can’t give anyone else better “economic and by others. Google’s book search engine should be
other terms” than Google gets.28 Notice, for example, similarly ecumenical and treat book scans put online
that this term would preclude the Registry from offer- by others on an evenhanded basis. Google should also
ing a better revenue-sharing deal to Yahoo! even if be prohibited from using its terms of service to put
the Registry thinks that this better deal is necessary copyright-like terms-of-service restrictions on what
to turn Yahoo! into a serious competitor to Google. others do with its public-domain scans.31 In the oppo-
This clause alone might be enough to deter any other site direction, Google’s collection of public-domain
serious entry. Google’s concern about being undercut books should be open to other search engines.
is real, but provided that the Registry itself is under
proper antitrust scrutiny (see above), it has nothing PRINCIPLE 3: CONSUMERS
legitimate to fear. The most-favored-nations clause N E E D P R OT E C T I O N
should be struck.
• Competitors offered the same deal as Google. Any Despite being the product of a lawsuit in which the
other entity willing to assume the same payment and reading public isn’t directly represented, the proposed
security obligations that Google assumes in the settle- settlement often takes the public interest seriously. Many
ment should be allowed to offer the same services clauses in it aren’t strictly necessary to resolve the dispute
that Google will, or any subset of them. This kind of between the parties to the lawsuit but nonetheless go a
competition is fair to authors and publishers because good way toward making sure that the results will provide
the various payment and security terms are already books to the public on fair terms:
presumably acceptable to them. It’s fair to Google,
which gets the same deal it currently does. It’s fair • Google specifically promises that it won’t use pop-up
to competitors, who could enter on a level playing or pop-under ads, and the Registry is also authorized
ground, without needing to roll the dice on invoking to take swift action to opt authors out of having
the legal system’s power to intervene. And it saves their books shown with “animated, audio or video
the legal system the work of having to deal with the advertisements.” 32
Microsoft Book Search lawsuit, the Yahoo! Book • In providing the Institutional Subscription, Google
Search lawsuit, the Facebook Book Search lawsuit, guarantees that its terms and conditions will “not
and so on. prohibit any uses . . . that would otherwise be permit-
• Registry authority to negotiate with Google ted under the Copyright Act.”33 This is significant;
competitors. The previous rule suffices for the services it’s a commitment that the subscriptions won’t require
the settlement describes, but there are also the new readers to surrender their fair use rights, for example.
business models, which can’t be specified in detail • Similarly, the Institutional Subscription will never
precisely because they don’t exist yet. Under the pro- offer an “experience and rights” worse than those
posed settlement, the Registry can give its blessing to enjoyed by readers who purchase e-books through the
Google on plenty of projects; it should be allowed to program,34 and copyright owners won’t be allowed to
give the same blessing to anyone else. Crucially, that sell out-of-print books while excluding them from the
blessing would have the same effect of binding all Institutional Subscription.35
authors and publishers to the deal it strikes. Again, • Colleges and public libraries will receive free comput-
authorizing the Registry to do such things is fair er terminals with complete access to the Institutional
because copyright owners could still opt out of any Subscription.36 The service is stingy by default: Four-
new uses.29 Once the Registry did bless a project—by year colleges will receive one terminal per 10,000
15
J O U R N A L O F I N T E R N E T L AW April 2009
students, and public libraries one per building. Google about copyright owners.42 That’s insufficient. The
isn’t even obligated to provide the service at all. settlement should contain explicit privacy guaran-
(There’s authority for Google and the Registry to tees that user information and reading habits should
expand this program, but it’s unclear that the Registry be monitored only to the minimal extent necessary
would ever approve more generous terms.) for billing, auditing, and security; that no such data
• Various provisions of the Copyright Act operate in be used for any other purpose, that all such data be
favor of libraries and other public-service entities, promptly destroyed when no longer needed; that
and the settlement makes sure that those promises Google not reveal any information about any user
remain intact. Libraries that allowed their books to or users’ reading habits to any other entity, including
be scanned get back digital copies, which they can the Registry; and that Google be legally responsible
use for accessibility purposes and to replace damaged for any security breaches resulting in third-party
or lost copies of physical books.37 access to reader information.
• Reasonable terms and conditions. The settlement
These provisions are all to the good. There are, how- should protect reader rights under the Copyright
ever, other consumer-protection matters on which the Act across the board. There’s a good first cut at such
settlement is silent or ambiguous. Google insists on its language in the settlement already. Google promises
good intentions in many of these areas. (”Don’t be evil,” reasonable terms and conditions for the Institutional
and all that.) I would prefer to see stronger protections Subscription, specifically that it will “not prohibit
than Google’s continued promises of non-evilness. It’s in any uses . . . that would otherwise be permitted under
the nature of such assurances that they’re still offered long the Copyright Act.”43 Similarly, the Library-Registry
after they’ve ceased to be true. To the extent that Google agreements should contain explicit statements that
really means to abide by them, it should have no objection library terms of service will not require readers to give
to putting these terms explicitly in the settlement agree- up any of their other rights under the Copyright Act.
ment or in an FTC consent decree. Any new business models should come with the same
protections, as well.
• No price discrimination. The pricing structure for the
Institutional Subscription will be largely fair. (Google PRINCIPLE 4: PUBLIC GOODS
will set pricing on an FTE basis with different pric- S H O U L D B E W I D E LY AVA I L A B L E
ing buckets for different categories of institutions, for
example, higher education, corporate, government, The Google Book Search project would be impossible
etc.38) For individual buyers, either booksellers can without some crucial bibliographic databases. A pile of
pick a price, or they can let Google’s algorithms set a scans is useless unless it’s linked to a database of publi-
price for them based on buying patterns.39 The settle- cation metadata. Since only out-of-print books will be
ment doesn’t explicitly say that Google won’t charge previewable by default, Google needs a database telling
different readers different prices for the same book, it which books are in and out of print.44 To convey copy-
something that its immense computational power right-owner requests to Google and to convey payments
and huge pricing corpus might make dangerously from Google to them, the Registry will need a database of
attractive. Google has no current plans to do so, and book copyright ownership. Some of these databases exist
it’s true that the market punished Amazon harshly already; some will be built or supplemented as part of the
when it tried the same stunt a few years back, but still. project.
Better safe than sorry. These databases are all public goods. They’ll be use-
• Reader privacy. Your choice of reading matter ought ful to readers and researchers. They’re also going to be
to be highly private. There’s a real concern that immensely useful to players in the book business. The
Google could identify and track readers, page by page, in-print database will help libraries understand their rights
minute by minute. Indeed, the security standard that under copyright law; the rights-owner database will help
Google must comply with requires it to keep exten- publishers gather the rights that they need to publish new
sive logs of user activity.40 Similarly, libraries that and exciting editions.
open their digital copies for scholarly and classroom Moreover, these databases are byproducts of the
uses must “keep track of and report[] all such uses Google Book Search project, not its goals. Google isn’t
of Books to the Registry.”41 The only explicit pri- compiling them because it can make money selling access
vacy protections in the settlement, though, are about to the databases. Instead, it’s compiling them because it
keeping private the information that the Registry has can’t offer Book Search without them. Whether or not
16
April 2009 J O U R N A L O F I N T E R N E T L AW
Google (or the Registry) can monetize these databases problems. Pseudonyms and proxies are reasonable,
directly won’t substantially affect the incentive to compile provided that the database in general is made avail-
them. able so that others can use it as a point of contact in
Taken together, these propositions imply that these finding copyright owners or in verifying that no one
databases should be opened to broad public access. That’s knows who the owner is or where she can be found.
exactly the policy that the settlement takes with the • Best practices for open access. In making these
Books Database—Google’s list of books it has or plans to databases available and in providing some of the
scan—which is required to be online and searchable.45 The other core services (the exact set to be determined),
same policy should be adopted wherever else possible. Google should be required to use standard APIs and
That’s not everywhere. I’m informed that Google has open data formats, as well as to allow programmatic
assembled its database of bibliographic metadata about access and bulk download where appropriate. Google
publication largely by licensing it from other sources. currently does this as a matter of policy in many of
Scholars may contest whether such databases should be its other lines of business, and it’s already providing
capable of exclusive licensing, but even those who want to PDF downloads of public-domain books. These good
pick that fight shouldn’t pick it here. Google didn’t gener- policies should be enshrined as actual requirements.
ate this data; it shouldn’t be forced to reveal it. Among other things, they’ll ensure that Google’s
The correct principle, instead, is that to the extent that competitors behave reasonably, too.
Google and the Registry create new and useful metadata
databases as part of the Book Search project, those data- P R I N C I P L E 5 : AC C O U N TA B I L I T Y
bases should be offered to the public, gratis, and without A N D T R A N S PA R E N C Y M AT T E R
legal or technical restrictions. The settlement agreement
contemplates at least two such databases, both important Google has been repeatedly criticized by scholars
(though the principle might also apply to others). and activists upset at its lack of institutional transpar-
ency. It’s also learned from that criticism. The settlement
• Publicly available in-print information. Consider agreement contains some reassuring provisions to provide
first the database of in-print information that Google accountability. As good as they are, they should be supple-
needs to decide whether works are “commercially mented with a few more.
available” and thus restricted by default. Google cur- Institutionally, Google and the Registry are given
rently synthesizes this information from a variety of mutual rights to audit each others’ relevant books.49 While
sources (such as looking at used book sales online).46 these audits are themselves confidential, the arrange-
Google is required by the settlement to make this ment creates a healthy system of mutual accountability.
information available to the Registry on behalf of Similarly, research users and libraries are subject to secu-
copyright owners;47 it should be required to make rity audits under suitable procedures.50 When there are
the database public, as well. This isn’t likely to be a disputes about public-domain or in-print status, they’re
practical problem, since Google will all but inevitably subjected to a low-stakes initial process that lets the par-
expose this information when it lets users either see ties sort out the facts.51 Larger disputes go first through
preview pages or not. But still, Google shouldn’t be executive-level mediation and then arbitration, on rea-
given the option to restrict its availability; in case sonably balanced terms.52
Google does wind up having competitors in this Google has also accepted a fairly stringent set of rules
space, there will be an inevitable temptation to cut that prohibit it from altering the texts of the books that
off access as a way of slowing down the other guy. it scans. There are the usual, sensible carve outs: Google
• Publicly available copyright ownership information. can hyperlink indices, it can link from books to sources
There’s also the database of information about copy- they cite, it can highlight user search queries, and it can
right claims that the Registry will need to use to even add a limited social-networking annotation-sharing
distribute payments among copyright owners. The feature.53 These are specific exceptions, however, from the
Registry is required to share much of this information general principle that it won’t change one word of the
with Google;48 it should be required to share almost author’s writings without permission.54 Good.
as much of the copyright-owner database with the Google has even agreed to procedures that limit its
public. There are privacy concerns here, since it will editorial discretion to exclude books from being displayed.
contain information about authors, but those con- If Google removes a book for “editorial reasons,” it will tell
cerns can be accommodated without much limiting the Registry about it and give the Registry a digital copy of
the usefulness of the database in solving orphan works the book.55 The Registry may then go out and commission
17
J O U R N A L O F I N T E R N E T L AW April 2009
a competitor to provide the display services that Google institute a review program that would take reports of
has refused to.56 Google believes it has a First Amendment distorted or mutilated digital versions, respond with
right not to be required to “speak” by passing along a book reasonable speed, and rescan the books if necessary
that it strongly objects to, and it’s chosen an honorable (and allowed by the owner of the physical copy).
and speech-friendly way of exercising that right. Google’s • Other scanning institutions. One last point of
waiver does not censor the book itself, which can still be accountability concerns an issue raised by Jean-Noël
made available through other means. Jeanneney: What books are scanned and in the col-
There are, however, some potential accountability lection at all? 59 Jeanneney’s specific concern—a lack
holes in this system. One is that the Registry need not, of Francophone sources—has an easy and obvious
or might not be able to, find a replacement for Google. response: The Bibliothèque nationale de France, of
I’m reluctant to intervene too strongly here, particularly which he is the president, could join with Google to
when no other potential partner is willing to step forward. scan its collections. Indeed, Google has indicated its
Others, bolder than I, might propose a positive duty on broad willingness to partner with libraries interested
the Registry’s part, but I take no position on the issue, in scanning large corpuses of books to get them into
noting only that it raises difficult issues of free speech law the digital collection more quickly. Once again,
and free speech policy. Fortunately, other potential holes Google’s sensible policy is one thing in the context
are easier and less controversial to close. of a private Google project and another in the con-
text of a massive remaking of the US system of book
• No secret censorship. If Google de-lists a book and copyrights that requires the blessing of a court of law.
the Registry doesn’t or can’t engage a replacement, So long as Google is the only serious player in book
the book will genuinely vanish from this new Library scanning and search, any institution that wishes to
of Alexandria. Perhaps that should happen for some provide books for scanning, or to perform scanning
books, but decisions like that shouldn’t be made in itself, should be allowed to take part in the scanning
secret. When Google chooses to exclude a book for effort and ensure that particular works are digitized.
editorial reasons, it should be required to inform the There will need to be appropriate provisions about
copyright owner and the general public, not just the capacity, financing, quality control, and so on, but
Registry. This path leaves intact Google’s option to a well-drafted consent-not-to-be-unreasonably-with-
be silent, but requires that it be exercised with trans- held clause can take care of many of them. Once
parency. If and when Google chooses not to speak, again, it’s worth emphasizing that this provision, like
it should own the ethical consequences, rather than all of the others, would apply both to Google and to
being able to hide from its decision to hide a book. any of its competitors who come in under the modi-
• Clear definition of non-editorial exclusion. The fied settlement.
settlement contains no clear distinction between
“non-editorial” and “editorial” reasons for Google to CONCLUSION
exclude a book from being displayed. This ambigu-
ity raises the possibility that Google might exclude The starting point for my analysis has been that
a book for editorial reasons but tell no one, not Google and the copyright owners are asking a federal
even the Registry, about it, and thereby completely court to put the US judicial power behind a document
suppress the book. There’s a danger of line-crossing that they have presented to it. The court’s consent should
wherever a line is drawn, but assuming that Google not be given lightly; the settlement should be approved
will act in good faith, a sharper definition of “non- only when the court is satisfied that it really will serve the
editorial reasons” should suffice. The current draft of interests of all parties, including the public. I have tried to
the settlement says “quality, user experience, legal, or offer general principles to think through what the public
other non-editorial reasons,” an unclear and impre- interest requires, along with specific, realistic recommen-
cise list that could easily be converted into a clear and dations to implement those principles.
precise one. At the same time, this is not a sentencing hearing
• Accurate scanning. Under the settlement, Google or a legislative chamber. The court is not in a position to
can’t “intentionally alter” the text of the books that rewire Google and the book industry to right all wrongs
it scans,57 but it doesn’t promise anything about the therein, nor should it try. Google’s other ventures are not
quality of the digitization process itself. Unfortunately, on the table; nor are the many other problems bedevil-
scanning mistakes have left some of its digital books ing copyright law. My recommendations respond to the
all but unreadable.58 Google should be required to specific question that the court faces: Should it use its
18
April 2009 J O U R N A L O F I N T E R N E T L AW
power to bind absent class members and approve this and those who think that it does only evil. Perhaps they’ll
settlement? prove equally frustrating. I offer them not as criticisms of
Thus, I hope that my recommendations all have the settlement, but refinements of it. As the chess adage
two things in common. Each takes off from some issue goes, “When you see a good move—wait—look for a
specifically raised by the proposed settlement, some way better one!”
in which approving the settlement could cause trouble
down the line. Each then offers a change to head off that N OT E S
trouble, a change more or less narrowly tailored to the
1. See Proposed Settlement, Authors Guild v. Google Inc., No. 05 CV 8136
issue it confronts. (S.D.N.Y. filed Oct. 28, 2008) [hereinafter Settlement]. The case has been
How do we get there from here? The parties can’t coordinated with and would be settled along with the related publisher
lawsuit, McGraw-Hill Cos. v. Google, Inc., No. 05 CV 8881 (S.D.N.Y.
settle the case without the court’s approval, but the court complaint filed Oct. 19, 2005).
can’t just rewrite the settlement and impose it on them, 2. Settlement art. X. [
either. The best alternative would be for Google, the 3. Settlement art. V.
authors, and the publishers to modify the proposed settle- 4. Settlement §§ 3.1(a), 3.3(a)–(c), 3.4.
ment along the lines that I’ve suggested and then bring 5. Settlement §§ 1.86, 4.5(a)(ii).
the modified version to the court for approval. None of my 6. Settlement art. VI.
recommendations touches the basic deal at the heart of the 7. Settlement § 4.1.
settlement or seeks to impose terms that the parties should 8. Settlement § 4.8.
find onerous. 9. Settlement § 4.2.
10. Settlement § 4.2(b)–(c).
We don’t need to depend on the parties’ good graces,
11. Settlement § 4.5(a)(i).
however. Some of the concerns that I’ve discussed can
12. Settlement § 7.2(b)(vi), (d).
easily be raised by members of the plaintiff class as part
13. Settlement § 7.2(d)(vi)(1).
of an objection to the fairness of the settlement to them.
14. Settlement § 7.1.
Authors have a particular interest in not having their 15. Settlement § 7.2, attachment B1.
work unaccountably excluded from Google Book Search, 16. Settlement attachments B2, B3.
for example. They should enter objections to the settle- 17. Settlement § 3.14.
ment unless it’s modified. Amicus briefing on some of 18. Settlement art. VI.
these issues could help the court put the settlement in 19. Settlement § 8.2(a)(iv).
context, as would intervention by some of the many inter- 20. Settlement attachment C § 1.1(a).
ested non-parties. 21. Settlement § 4.7.
Even if the court is unwilling to fit all of these issues 22. Settlement § 6.2(b).
23. Settlement § 6.2(b).
into the copyright dispute before it, there are other legal
24. Settlement § 2.4.
avenues open. The DoJ could open an investigation into
25. Settlement § 3.5(a)(iii).
the antitrust issues and the Federal Trade Commission
26. Settlement § 3.5(b)(i).
into the consumer-protection ones. It matters less that 27. Settlement §§ 1.16, 1.38, 1.142. This statement leaves out copyright
these recommendations be embodied in the settlement interests in foreign works; under US law, one need not register such a work
than that they be enshrined somewhere enforceable, and before suit. See 17 U.S.C. § 411(a). Owners of unregistered copyrights in
foreign works are part of the plaintiff class and thus part of the settlement.
consent decrees are a perfectly reasonable alternative. 28. Settlement § 3.8(a).
Potential competitors in book-scanning can also raise 29. Settlement § 4.7.
many of these issues, for example, by way of a private 30. See Tim O’Reilly, “Book Search Should Work Like Web Search,” O’Reilly
antitrust suit; Microsoft, which previously had a book- Radar (Dec. 11, 2006), http://radar.oreilly.com/archives/2006/12/book-
search-sho.html.
scanning project, would be a natural plaintiff.
31. See Yakov Shafranovich, “Change in Google Book Search Guidelines
My goals here are pragmatic. I’m not proposing to take for Public Domain Books” (Jan. 30, 2009), http://www.shaftek.org/
public control of the Book Search project. In comparison blog/2009/01/30/change-in-google-book-search-guidelines-for-public-domain-
books/.
with the institutional reconfiguration of book copyright
32. Settlement § 3.10(c)(iii).
law that the settlement would enact, these tweaks are all
33. Settlement § 4.1(e).
quite minor. Nor am I proposing to leave Book Search 34. Settlement § 4.1(f).
entirely alone; the parties gave up on that possibility when 35. Settlement § 3.5(b)(iii).
they asked the court to approve this sweeping class-action 36. Settlement § 4.8
settlement. 37. Settlement § 7.2(b).
I hope that these recommendations will prove equally 38. Settlement § 4.1(a).
appealing to those who think that Google can do no evil 39. Settlement § 4.2(b)–(c)
19
J O U R N A L O F I N T E R N E T L AW April 2009
40. Settlement attachment D § 3. 51. Settlement § 3.2(d)(iv)–(v).
41. Settlement § 7.2 (b)(vii) 52. Settlement art. IX.
42. Settlement § 6.6(a)(vi), (d). 53. Settlement § 3.10(c)(ii).
43. Settlement § 4.1(e). 54. Settlement § 3.10(c)(i).
44. Settlement § 3.2(b). 55. Settlement § 3.7(e).
45. Settlement § 3.1(b)(ii). 56. Settlement § 3.7(e)(i).
46. Settlement § 3.2(d). 57. Settlement § 3.10(c)(i).
47. Settlement § 3.2(d)(ii). 58. See Paul Duguid, “Inheritance and Loss? A Brief Survey of Google Books,”
First Monday (Aug. 2007), http://firstmonday.org/htbin/cgiwrap/bin/ojs/index.
48. Settlement § 6.6(c).
php/fm/article/view/1972/1847.
49. Settlement § 4.6(e), 6.3(d). 59. See Jean-Noël Jeanneney, Google and the Myth of Universal Knowledge
50. Settlement § 8.2(c). (2007).
20