Google This Georgia Harper

Google began an ambitious project last winter to digitize books from five large library collections and make the works' contents searchable online. It's called the Google Print Library Project ("Library Project" or "Project") and the participating libraries include Harvard, Stanford, University of Michigan, New York Public Library, and Oxford. As reported in various news stories over the last 10 months, Google will digitize both public domain and protected works. It will use the full text versions of protected works "behind the scenes" in a database capable of being queried by searchers using Google's popular web-based search engine interface. If a user's search term appears in a protected work, the search results page will show "snippets" of text surrounding the search term (a sentence or two). That search results item would also show links facilitating purchase of the work, if it is available, and links to the library's collection where a physical copy may be found. If the search term appears in a public domain work, the entire work will be available when one clicks on the search result. Everyone agrees that digitizing public domain works poses no copyright problems, but for those works still under protection, copyright issues came to a head in September when the Author's Guild filed suit against Google in New York, home to the publishing industry. Five major publishers, all members of the Association of American Publishers (AAP) quickly followed suit filing their own complain on October 19. Courts in New York have considerable experience with complex copyright issues. They'll need it.

The Library Project promises a fascinating legal dispute. Ostensibly, the battle revolves around the issue of whether the Project makes fair use of the literary property of others, since digitizing a protected work in its entirety would otherwise be an infringement. But that's only part of the story. It's unlikely that either side will win this case on the strength of a fair use analysis regardless of how subtle, clever or even brilliant it may be. Rather, this case is fundamentally about what we as a society should do with our older books sitting on library shelves gathering dust. It will present hard choices among good competing alternatives where the good can be the enemy of the best. Nevertheless, the Library Project undeniably presents opportunities to examine copyright principles, to explore the meaning of recent technological advances and the implications of digital technologies for decades to come, and of course, to really understand the effect our law has on the achievement of its own ends in light of these changes. This is phenomenal, and so exciting. The dispute around the Library Project touches broadly upon the problems those who would use the works of others encounter where it is not clear who the "others" are or how to reach them -- the orphan works issue. It will raise issues regarding our obligations under international treaties, the rights of libraries to hire others to digitize the works on their shelves and their rights to use the digital copies provided, and the role of libraries and of publishers more generally. It will usher in changes that we won't fully appreciate for a long time to come.

And it may help more people understand how copyright law and fair use really work today, for better or worse. A lot of people will notice and care about this case.

The Four Factor Fair Use Test: Putting the Cart Before the Horse?

More than two decades ago, Wendy Gordon published, "Fair Use as Market Failure: A Structural and Economic Analysis of the Betamax Case and its Predecessors" (82 Colum. L. Rev. 1600), observing generally that courts seemed to decide fair use cases based on underlying principles involving market failures. In essence, she said that if market failures were present and couldn't be easily remedied by a finding of liability, courts would tend to find a use fair. If they were not present, the use would most likely be judged not fair. It wasn't that simple, of course, but the details are not critical for understanding this point. To reiterate: she noticed that the four factor fair use test was not what judges seemed to use to decide whether a use was fair, though it's what they must use to explain the conclusion they may already have reached based on other criteria. To the extent that she was right, or at least to the extent what she described is operating in this case or any case at issue, litigants waste a lot of time arguing about the four factors. It may be quite reasonable to conclude that they would be irrelevant if Section 107 didn't require judges to explain their decisions in those terms. The four factors serve quite well in this limited capacity though, because they are so malleable that practically any use gray enough to reach a courtroom can be characterized as fair or not fair by simply choosing among the many facts about the use those to emphasize and those to gloss over. A review of any important fair use decision will show battling four factor analyses in the majority and dissenting opinions, and usually at the different levels of review as well. No offense to Justice Story, who believed that "[p]atents and copyrights approach nearer than any other class of cases belonging to forensic discussions, to what may be called the metaphysics of the law, where the distinctions are, or at least may be very subtle and refined, and sometimes, almost evanescent[,]" and who devised within that understanding the considerations that have become our modern-day four factors (Folsom v. Marsh, 9 F.Cas. 342, 344 (C.C.D. Mass. 1841) (No. 4901)), but using the four factor framework in this way is child's play.

This is not to suggest that there's anything particularly untoward in the way fair use disputes are resolved. Rather, examining fair use in this way emphasizes what's really at issue, what's really important, and, I believe, what really wins or loses fair use cases like those involving Google. To ease ourselves away from reliance on the four factors for understanding this dispute, let's just ignore them for now. Thus, I hypothesize that if a court had before it a clear copyright violation involving a new use of old works by a defendant who had raised fair use as a defense, it could, or more strongly, would, decide the case based on other criteria besides the four factors but report the decision, either way it goes, within the four factor framework.

The essential question facing the court is simply, who *should* get the benefit of a new use of old works, the copyright owner or the person or entity who wants to use the copyright owner's work? Obviously, without the four factors, the court must come up with other criteria for deciding who should initially get the benefit of this new use. When they truly lack statutory or precedential guidance for resolving a dispute, courts usually examine the underlying principles, goals or objectives of the law. Any resolution must comport with these principles and further the legal objectives. The goal of copyright law is to promote the progress of knowledge. The law accomplishes this goal by giving copyright owners exclusive rights to the works they create for a limited time. At issue in this case is the scope of those rights: Should the copyright owner's rights include the right to digitize and authorize others to digitize a work in its entirety for the purpose of including the work in a searchable Internet index? Will giving this right exclusively to copyright owners or reserving the right to the public best promote the progress of knowledge?

The court will put aside initially the fear that these digital files could "get loose" along with other similarly tangential fears such as what the Library Project might do to libraries in the future, or to booksellers, or to publishers, and dreams about attenuated benefits that might result from a decision either way. Instead, initially it will focus only on the immediate implications of a decision for the use that's at issue.

Each side wants to persuade the court to give it the benefit of this new "digitizing analog content for web-based indexing" right. Clearly, whomever the court awards the benefit to has an initial advantage in the negotiations that should take place thereafter if the other side desires to make some use of the new right. Is there a logical starting point for this decision-making process? Many, though not all, copyright scholars believe that as a starting point, courts should award the benefit of each novel use of his works to the copyright owner if the use deals with works of a type that copyright typically covers or with the kinds of rights the law typically protects. This case deals with books and making copies, both well within the ambit of the law's current protections. Thus, for those who believe in this default position for awarding new uses, the court should initially award the right to this new use to the content owners. Other scholars believe that new uses should be reserved for the public unless the copyright owner can demonstrate a certain level of need for the revenue from exploitation of the right and perhaps a present capability to exploit it. There are many other points of view about where to initially place the right and what the other side should have to show to overcome a default presumption. But for this case, let's assume that our judge is not persuaded by any automatic rule. He just wants to think through what will happen if he places the right with one side or the other. An interesting issue in this case comes up immediately: for many works, and by some estimates for most works, awarding the right to the copyright owner will not lead to fruitful negotiations or any negotiations at all. In many cases there won't be anyone at the negotiating table to represent the copyright owner. This is the "orphan works issue." For a majority of protected works found on America's library bookshelves today, placing the right with the content owner is simply throwing the right away, literally, giving it to the dead. This is the market failure at the heart of this case: no market can form around this right if it belongs to someone who is either dead and buried or who does not know or care that he has the right.

From the Least to the Most Speculative Evidence

Of course, the only cases that make it to court are the close ones, so even though this market failure is apparent early on, our judge is going to have to look carefully at both sides' arguments. He will ignore the competing fair use analyses, in accordance with our hypothesis. He will focus on the harms that both sides claim will result from a decision that goes against them and the benefits they'll tout of a decision in their favor. Nonetheless, the judge will likely toss out all the gloom and doom predictions that both sides will have provided him because they are not hard evidence. They don't prove anything. He's going to want to rely on his least speculative, currently credible, verifiable evidence. So, limiting himself to the most credible facts, at the outset he will likely note that:

If he gives the right to the content owners:

• Indexers will have to negotiate with content owners
  • Many content owners will license their rights to indexers
  • Many will not license their rights to indexers
• In many cases, indexers will not be able to identify or locate rights holders due to the market failure caused by orphan work status
  • If the indexer uses an orphan work and its rights holder later comes forward, the indexer could be liable for willful infringement, so these types of works will probably not be indexed by anyone

If he gives the right to the indexers:

• Indexers need not negotiate with content owners to include a work, but content owners will have to negotiate with indexers if they do not want their works digitized
  • This would validate Google's "opt-out" approach -- Google agrees not to include in its index any work a content owner indicates it wants Google to exclude
  • Orphan works by definition will not be excluded because there is no identifiable owner to negotiate with Google to exclude the work
• Indexers will digitize large numbers of protected works making some significant proportion of all written works searchable after a period of time

Now it gets tricky. If the statements above represent facts stripped of attenuated predictions, over which there is minimal disagreement, what may the judge reasonably conclude will be the immediate ramifications of a decision either way?

If he gives the right to the content owners:

• Conclusion: The project will cost more and take more time
• Conclusion: The project may fail to reach a critical mass of coverage and the result may not justify the time, money and effort that the threat of infringement liability will necessitate, principally because of the orphan works problem (the market failure that will necessitate high expenditures to determine that ultimately there is no one to negotiate with) and to some extent because of holdouts (those who want more than the indexer is willing to pay)
  • On the other hand, no index is perfect and it may still be worthwhile for Google to pursue only those works whose publishers affirmatively give permission
    • The Yahoo! Open Content Alliance suggests that an index that excludes orphan works is still valuable and viable though this project's goals are somewhat different than Google's
• Conclusion: Some content owners will likely digitize their own works and perhaps work cooperatively with other content owners to create indexes of their collective works
  • but no one will be able to include orphan works in these indexes;
  • coverage dependent upon clearing all rights will always be spotty reducing the effectiveness of any one index and of all indexes

If he gives the right to the indexers:

• Conclusion: Indexers will have an unambiguous right to digitize content and display snippets, under the circumstances Google presents in its fact pattern
• Conclusion: The public will benefit generally by having improved access to information contained within older analog works
• Conclusion: Copyright owners will benefit if there is an increased awareness of and demand for their works

Now, even trickier: what are the less likely, but still considerable ramifications about which there's probably going to be significant disagreement? The judge may discount these in his actual decision-making because they are speculative and disputed, but he may cite to them as further support for his decision in the end.

If he gives the right to the content owners:

• As a practical matter, the public will lose knowledge about and access to information contained in older works that remain available only in analog form as more and more content becomes available and searchable electronically
  • This will apply to orphan works and works whose copyright owners do not permit their works to be digitized
  • It will be ameliorated as those works begin to enter the public domain (assuming no further copyright term extensions) or to the extent Congress addresses this issue as described below
• There may be ramifications for the copies that search engines make of content already online
  • Perhaps implied license (and "opting out" of implied licenses) adequately covers these copies obviating the need to rely on fair use

If he gives the right to the indexers:

• Some content owners will lose revenue or other consideration they might have obtained through negotiations with indexers
• If the digitized files "get loose," there may be some uncontrolled uses of them by the public
  • The market for permissions and for books themselves may be undercut by unauthorized uses of the digital files
    • Print on demand technologies may take off once most books are digitized, increasing low-priced alternatives to attract customers away from otherwise illegitimate outlets
    • Content owners may prevent illegitimate alternatives from outperforming legitimate ones with regard to the breadth of available content by acting promptly to make the widest breadth of materials available though authorized outlets
• Google will give a copy to libraries of digitized books they contributed to an index, facilitating their use of those works, if they were so inclined, beyond the scope that content owners' believe to be fair use, or beyond the uses authorized by Section 108
  • Libraries and content owners are collaborating with the Copyright Office to revise Section 108 to allow libraries to make better use of digital copies of the works in their collections; many anticipate that these negotiations could include clarifications regarding fair uses such as electronic reserves
• There will be evolutionary effects on content owners, distributors and libraries, but these are so attenuated and speculative as to be very difficult or even impossible to integrate into the reasoning process

Overall Societal Benefit: How Many Ways Can We Solve A Problem?

The judge now has a rough idea of what will probably happen, and some thoughts about what may happen, as a result of a decision either way. It may simply be a matter now of comparing the two outcomes to see which result overall provides the most social benefit with the least harm; or, in terms of our underlying copyright principles, which best promotes the growth of knowledge. But there is more information that he must integrate into his decision-making process. Although the facts indicate that the market failure caused by orphan works will be a major problem for indexers if he gives the right to the copyright owners in that a majority of protected published works will remain incapable of being indexed until they enter the public domain, the Copyright Office has begun to address this issue and hopes to recommend to Congress a solution that will give all those who want to use orphan works (even other content owners themselves) a way forward. The Copyright Office could effectively address what is otherwise a significant market failure by recommending to Congress a workable solution to the problems posed by works whose copyright owners can not be identified or located, or who do not respond to inquiries. But any recommendation that proposes public uses of protected works will face perhaps insurmountable resistance by some copyright owners. Congress will not pass a bill opposed by powerful content community interests. But, can Congress broker a compromise that preserves the rights of owners of orphan works who ultimately come forward without so hobbling the public's ability to use these works that the use right is use-less? Unfortunately, it is impossible to know the answer to that question right now.

Nevertheless, the mere possibility that Congress might solve this problem will affect the judge's decision. Congressional action is a wild card.

As an aside, one might assume that the problem orphan works pose to the eventual digitization of most published works gives the content community a choice: it could opt for a balanced legislative orphan works solution that provides sufficient protection for the rights of owners of orphan works who come to light, but that truly does otherwise release these works to all public uses, or it could opt to treat index-related digitizing as fair, leaving other uses of orphan works to be evaluated on a case-by-case basis as all fair uses must be. Logically though, so long as an orphan work is an orphan work, it presents by definition a significant and continuing market failure, so one might hypothesize that many, perhaps all socially beneficial uses of orphan works would likely be fair. But, as many have noted, there are those too risk averse to rely on fair use, even when obvious market failure militates against a finding of liability. Thus, characterizing this use as fair still leaves some ambiguity with which some will be uncomfortable. On the other hand, there has never been a legislative solution to any copyright problem that didn't create nearly as many questions as it answered...

Could the content community have it both ways: perpetuate the orphan works problem by resisting a meaningful legislative solution and demand that indexers must get content owners' permission? It is difficult to imagine how one might responsibly take that position, but the content community has taken surprising positions. Similarly, could indexers have it both ways? May they demand a workable legislative solution to the orphan works market failure and still argue that index-related digitizing is fair use? It would seem out of the question, but that is, of course, assuming a truly functional, economically efficient orphan works solution, which may elude Congress.

The Finishing Touch: Framing the Decision within the Four Fair Use Factors

So, regardless of how our judge decides to allocate the benefit of the indexing right, he now must report his decision in Section 107's four fair use factor terms. This should not pose a challenge. He will have been provided competing fair use analyses by the two litigants (and from countless commentators if he has time to read them). He may adopt one that reasonably explains his conclusion or fashion his own. The dissent will have come to a different conclusion about overall social utility and will adopt a different four factor analysis emphasizing a different subset of the facts. It's that simple. If the case goes up on appeal, the majority and dissenting opinions may flip, and they may flip again if the case goes to the Supreme Court, but the process for deciding would be the same at every level.

Wendy Gordon was right. The side whose fundamental case resonates best with those who adhere to market failure principles will win. Litigants must address the four factors, of course, but their most important arguments must hammer hard on the raw economics of the positions they take. It may appear that Google has the better of this dispute because of the near-term, empirical ramifications of an award of the right to indexers, but content owners can point to legislative potential to fix this market failure. It would be better were a legislative solution on the table before the court must decide, but that is not indispensable. After all, if Congress' solution does not really work, the truth will out. The market will not hide that fact. And it's all about the market. And those of us who watch fair use cases should get used to that.

Top | Search
Crash Course in Copyright | Intellectual Property Section | Office of General Counsel