May 20, 2005
Mr. Alexander Macgillivray
Senior Intellectual Property and Product Counsel
1600 Amphitheater Parkway
Mountain View, CA 94043
Dear Mr. Macgillivray:
I am the Executive Director of the Association of American University Presses, whose 125 members are all non-profit scholarly publishers. Most of them are affiliated with research universities in the United States and Canada, but our membership also includes scholarly societies, museums, and non-degree-granting research institutions.
The common mission that unites all our members is to help the advancement of knowledge by making the results of scholarly research known through their publications, and it is generally recognized that those peer-reviewed publications set the gold standard for excellence of information. Quantitatively the members of AAUP publish around 750 academic journals and 11,000 books each year; qualitatively each year their publications win hundreds of prizes for their superiority in their fields. Major research libraries in the English-speaking world routinely buy all the books and journals published by AAUP members.
Last year when Google representatives were soliciting publishers' agreement to participate in Google Print for Publishers, many AAUP members signed on with enthusiasm. They continue to believe the program has great potential. However, our members were understandably confused about Google Print for Libraries when it was first publicly announced last December. Even though, according to the publicity, the program was being developed during the same period as Google Print for Publishers, news of Google Print for Libraries came as a complete surprise. It had not been mentioned by Google representatives during any of the discussions they were having with our members, and Google's subsequent explanations of Google Print for Libraries have only increased that confusion and transformed it into mounting alarm and concern at a plan that appears to involve systematic infringement of copyright on a massive scale.
I am referring, of course, to the portion of Google Print for Libraries that calls for the eventual digitization of copyrighted works in the collections of the university libraries at Harvard, Michigan and Stanford. Google asserts that it can make these copies without seeking permission as a fair use under Section 107 of the Copyright Act, and Google plans to give copies of those digitized works to the participating libraries.
The idea that once this gigantic digitization project has been completed anyone with a computer and internet access will be able to use Google to search the collections of these libraries -- including the public domain material from the New York Public Library and the Bodleian Library at Oxford -- is enormously seductive. However, it also appears to be built on a fundamental violation of the copyright act, and this large-scale infringement has the potential for serious financial damage to the members of AAUP.
Although they are nonprofits and many of them receive an operating subsidy from their parent institutions, our members still have payrolls to meet and bills to pay, and in 2003, the most recent year for which we have such data, total university support only averaged about 13% of their operating revenue. Virtually all the rest of the money required to recover costs and stay in business must come from the sale and licensing of their publications, and as in any publishing business, copyright plays an utterly fundamental role in establishing the legal basis on which their business rests. Respect for copyright is essential to scholarly publishing. (For your information, I'm appending a brief AAUP statement on the relationship between copyright and cost recovery. You may also find it on our website at http://www.aaupnet.org/aboutup/issues/index.html#statement.) For the members of AAUP, most of whom struggle to break even in the best of times, that risk is serious indeed.
Below is a series of questions to which I would appreciate your answers so that I can better inform our members about Google's practices and intentions. Some of the questions are fairly specific and technical, for which I apologize, but my purpose in asking them is to fill in some of the gaps in information that have been left by Google's public statements about the program, and to clarify others. I hope you will agree that it is worthwhile to make this effort. Without better communication and greater understanding, I'm afraid publishers' alarm and concern about Google Print for Libraries can only continue to grow.
Here are the questions.
1. Google recognized that permissions agreements with publishers were necessary for Google Print for Publishers; why do you believe agreements are unnecessary after the books have been sold to libraries?
2. Google's claim that it is fair use to make copies of every copyrighted work in even one major library, let alone three of them, is completely unprecedented in scale; it is tantamount to saying that Google can make copies of every copyrighted work ever published, period. Courts have never recognized a fair use claim of that magnitude. What is your argument that they should do so now?
3. This claim is also completely unprecedented in sweep. Under U.S. law, a fair use determination requires an analysis of the four factors specified in Section 107, and is highly fact- and circumstance-specific. What is your argument to justify treating books of haiku, dictionaries, novels, collections of letters, engineering handbooks, biographies, musical scores, works of literary criticism -- in short, every copyrighted work in a library's collection -- as identical so that they can all fall under the same four factors analysis?
4. If Google's position that copying copyrighted works in order to index and display snippets of them in search results is valid, what is to prevent, not just Google's immediate competitors, like Microsoft, Yahoo! and Alta Vista, but all other companies with a search engine from making the same claim? In a fair use analysis the courts have made plain that the analysis of market harm under the fourth factor can't be limited to the question of immediate market harm, but must also consider what would happen if the practice at issue were to become widespread.
5. At the recent meeting of scientific, technical and medical publishers in Washington, your colleague J.R. Needham, if I heard him correctly, told us that it was unnecessary for Google to clear permissions for Google Print for Libraries with those publishers who had agreed to participate in Google Print for Publishers, because they had already given their consent. These facts are simply wrong. Publishers' contracts for Google Print are title-specific and can't be interpreted as a blanket license. Furthermore, no publisher knew about Google Print for Libraries until it was announced in mid-December, and at the time of those initial announcements the librarian at the University of Michigan was quoted as saying how surprised he was that Google and the libraries had been so successful at keeping the program a secret for the two years or more it was in development.
On what basis did you conclude that publishers' agreement to participate in Google Print for Publishers automatically extended to approval of Google Print for Libraries, when in fact they knew nothing about this program?
6. Hugh Jones has reported that Tom Turvey recently claimed, while speaking to the General Meeting of the Publishers Association in the U.K., that all rights in the digital files you are creating, including copyright, would vest in Google. If Google is, or will be, asserting a copyright interest in its digital files of publishers' works, would you please explain the extent and basis for such an assertion?
7. The single case you have cited to support Google's fair use claim, Kelly v Arriba Soft, has a pattern of facts substantially different from those in Google Print for Libraries. Among many other important differences, Arriba Soft was making copies of images that had already been digitized and posted on the web by their copyright owners. Google is presuming the authority to digitize many works whose copyright owners have not taken that step, and given the ease with which digital files can be duplicated and further transmitted, may have good reason for deciding not to do so.
Additionally, the full resolution copies Arriba Soft made in order to create the low-resolution thumbnails were deleted from Arriba Soft's server after the thumbnails were made. Google claims the right to retain the digital copies it makes -- the full resolution copies, if you will -- even in those cases when a publisher asks them not to display any text from particular works.
Given these significant differences, how does Kelly v Arriba Soft support your claim? There are also recent cases in which the courts found against a claim of fair use, like Buena Vista v Pipeline Video (which, interestingly, cites the Arriba Soft decision). How does your fair use argument deal with these countervailing decisions?
8. Snippet is used so consistently in describing Google Print for Libraries that it's taking on the status of a technical term, and thus requires a specific definition. How long is a "snippet?"
9. How many digital copies do you intend to make in order for your indexing system to work? How many copies do you intend to make for archiving, and how often do the archival copies need to be refreshed?
10. How do you intend to protect these copies against misuse?
11. What protection do copyright owners have against a future owner of Google deciding to exploit them directly?
12. What protection do copyright owners have against Google itself deciding to adopt a new business model that involves the direct exploitation of these copies by, for example, offering Google users access through the pay-per-view system for which Google has a patent application pending?
13. Google's response to publishers' objections to Google Print for Libraries that they may "opt out" of the program seems both legally irrelevant and factually disingenuous. Among other reasons, it is irrelevant because all a publisher can do under this option is assert its control over the right of display by Google after the infringing copies have been made. It ignores the fundamental exclusive right of copyright owners to make copies in the first place, and it ignores the exclusive right of distribution, since a copy or copies will have already been given to the participating libraries.
And disingenuous because Google's practice so far has been to honor the "opt out" provision only on a very narrow and limited basis, if at all. At least two publishers have asked that the works to which they hold copyright not be included in Google Print for Libraries, and to date, Google has not complied. Several publishers associations have written to Google expressing concern on behalf of their members about the copyright infringement that appears to be built into Google Print for Libraries, and Google's replies have treated their concerns simply as a public relations issue. Finally, as far as I have been able to determine, there's no public information on Google's website about how a publisher may, in fact, opt out.
Does the "opt out" option actually exist? What is the process and where is it described? Does it permit a publisher to opt all of its copyrighted works out of Google Print for Libraries? If not, why not?
14. AAUP members, working individually, in collaboration, and with other vendors are already investing in and selling digital copies of works they publish to libraries through, for example, Rotunda, the ACLS History E-book project, NetLibrary, and so on. Section 108 of the Copyright Act permits neither the indiscriminate and wholesale digitization of copyrighted works in a library's collection, nor does it permit digitization for any of the purposes specified in Section 108 done by an entity other than the library or its employees. Moreover, Google is giving copies of the files for the works it has digitized to the libraries, not as a public service, but as a consideration for the libraries allowing Google to scan and digitize their collections.
In view of these facts, what is the legal justification for Google providing digital copies of copyrighted works to the participating libraries? Why aren't these copies infringing? How can the libraries claim that these copies have been lawfully acquired?
15. AAUP is very concerned about many libraries' extremely permissive use of digitized materials in their e-reserves systems. How many digital copies of copyrighted works is Google giving to the participating libraries? Under your agreements with them, what terms or restrictions have you placed on the libraries' use of these copies? Would you please provide us with a copy of your agreements with the participating libraries?
16. What is the status of the digitization program for Google Print for Libraries? Has Google started giving copies of the copyrighted works it has digitized to the libraries? When do you anticipate that Google users will be able to start using this program?
Google Print for Libraries has wonderful potential, but that potential can only be realized if the program itself respects the rights of copyright owners and the underlying purpose of copyright law. It cannot legitimately claim to advance the public interest by increasing access to published information if, in the process of doing so, it jeopardizes the just rewards of authors and the economic health of those nonprofit publishers, like the members of AAUP, who publish the most thoroughly vetted and highest quality information in the first place.
Thank you for your consideration, and I will look forward to your answers to these questions. In view of their pressing nature and the ongoing character of the program, please respond by June 20. If you think it would be useful I would be happy to meet with you or other qualified Google representatives to discuss them further, and to see if we can find a way they might be resolved.
Cc: Allan Adler, Association of American Publishers
Jens Bammel, International Publishers Association
Pieter Bolman, International Association of Scientific, Technical and Medical Publishers
Hugh Jones, The Publishers Association
Sally Morris, Association of Learned and Professional Society Publishers
Copyright and the Costs of Scholarly Publishing
On average, university presses recover 87% of the cost of publishing scholarly books from sales. An important component of this revenue comes from payments received for permission to reproduce works in, for example, anthologies, paperback editions, coursepacks, electronic reserves, and document delivery services.
Federal copyright law is the legal foundation on which this method of cost recovery rests. Copyright protects the right of authors to be recognized for their work and be appropriately compensated for it, and by limiting distribution to authorized rights holders it provides the basis for market-based recovery of publishing costs.
Respect for copyright is essential to making this system work. Copyright infringement violates authors' rights and, like any other form of theft, increases the burden on those who abide by the law. It puts pressure on prices, reduces publishing capacity, increases deficits, and shrinks resources needed for change, experimentation, and growth.
AAUP calls on all members of the university community -- students, faculty, and administrators -- to respect the obligation of university presses to strike a balance between the need for access to the information they publish, and the twin imperatives of protecting the legal rights of their authors and recovering publishing costs.