Epstein on the Google Print Library Project

On Wednesday I introduced and commented on a presentation by Professor Richard Epstein of the University of Chicago on Google Print and Copyright, cosponsored by the Georgetown Federalist Society and Student Intellectual Property Law Association. Professor Epstein is an impressive speaker who speaks apparently extemporaneously. He also speaks fast with few wasted words, so my notes are necessarily incomplete.

Epstein identified a bicoastal split on academic assessments of Google. By this he meant that Jane Ginsburg at Columbia is pro-copyright owner, whereas the Left Coast inclines towards Google. Where you come from determines how you think on IP; he comes from real and personal property. A large number of academics start with a preference for the commons, with exceptions as islands of property therein. Epstein isn’t squarely on the other side, since his starting point is Roman law, which provides a framework for asking why some things are common and others privatized.

Land is relatively compact and permanent. To get development, the developers need a return on investment; thus, property. Reaping what you sow is the metaphor for a reason. Property allows internalization of costs/benefits. Water is more complicated; privatization destroys the going concern value of a river, so you get a mix of commons and private rights.

IP is more like water. With respect to land, use requires exclusion of others; not so with respect to ideas. So the explanation for any exclusion has to be instrumental, and it is: incentives to create. IP is a constant duet between the need for incentives and the waste generated by exclusions. Both are inescapable.

Law attempts approximation of a social ideal. With land, problems come up with exclusion: necessity (shelter from a storm) and eminent domain, and the law responds by allowing exceptions, with or without compensation. It’s the same with IP – private rights should yield in some circumstances. Epstein sees fair use as analogous to defamation’s fair comment privilege, necessary for open and robust debate. Fair use had the same narrow origins, protecting criticism by allowing quotation. We do, however, want to avoid surreptitious reproduction that undermines the core right – so we don’t allow reproduction of A Tale of Two Cities with a negative preface (odd example, given that we allow reproduction of that work in every possible way, but substitute in the latest Harry Potter if you like).

The question of Google Print is whether you have to bypass a market to create a good thing. Shall transfers be accomplished by consent (licensing) or necessity (fair use)? Google’s ambition is large, to make anything digital and easily searched, which allows extraction of value not replicable by traditional techniques. The really hot question is what Google will do with its copy. The publishers don’t fear the snippets; what they fear is unknown uses of the database to extract value – to show what works consumers prefer, for example, so as to advise young authors what’s hot. In Epstein’s discussions with publishers, this was their main concern; they feared being preempted out of their market (emphasis mine, not his).

What about opt-out as a solution? Google is using “opt-out” as a term of art, since it dictates the terms and conditions. Harper & Row won’t just be able to send a letter saying “take all our books out,” since Google is requiring specific ISBN and other information. There’s a deeper problem: If Google Print is fair use, Google’s entitled to do this independent of the publishers’ consent (I wrote about this before).

Google isn’t the only one who wants to organize the world’s information: Yahoo!, Amazon, etc. are looking for blanket licenses, suggesting that market transactions are possible. If they can work out ways to share advertising revenue with publishers, now that microtransactions are feasible, than no fair use will be necessary.

Epstein noted that his analysis hadn’t mentioned the statutory fair use factors. Indeed, those factors are not key; the critical issue is whether a viable market on a voluntary basis is possible. In earlier discussions, Robert Merges had said that Google’s project was no problem because there was no viable cash market, but later he qualified that claim – there may be a market in two years. It’s fair use now, but maybe not in 2008. (My question: Does the same logic apply to Sony v. Universal? My guess is that the answer’s yes, though we needn’t wait for 2008 for the market.) Then the issue is how to organize the transition.

Epstein believes that a voluntary market is feasible. If MSN and Yahoo! can deal, so can Google.

What about implied consent? People argue that, if the Google Print for Libraries model is defective, then Google itself is defective with respect to the web as a whole. Opt-out works the same way online. How to distinguish the situations? Epstein’s answer is both practical and theoretical. Practically, the publishers brought a class action rather than sending Google their ISBNs, showing that there is not really implied consent. We should imply consent when a formal consent requirement would debilitate the market too much, destroying the usefulness of the resource (here, the search engine); when there’s a high probability of people wanting in, as there is online, we can presume consent in the absence of objection, but that’s not so in print.

Ultimately, we have to balance the potential gain from the database versus transaction costs. All copyrighted works are not created equal (I’d say, they may be created equal, but they don’t make it to the finish line equal, but that’s either a quibble or a disagreement far too profound for this space). With a 1950 work, it’s so great to have online access that a copyright owner would probably be happy to have it online. But with the latest Stephen King (short review: creepy premise plus competent execution produces decent midlevel King), precise management of the property is critical and thus a voluntary market should prevail. In the latter case, the value of price discrimination and other work-specific techniques dwarfs the transaction costs of consent. We need to separate low- from high-value works and make different rules for each, as we do with respect to allowing solicitors onto property.

The only wrong argument, in Epstein’s opinion: We shouldn’t think of Google Print like the overflight problem (see the exquisite William Empson poem reproduced in comments: fair use?). The ancient ad coelum doctrine said that property owners could exclude others all the way down and all the way up into the sky, but once airplanes appeared, that created huge problems: everyone could exclude and no one could use the skies (the anticommons, though Epstein didn't use that term), destroying the value of this new mode of transport. To avoid this, we forced the top slice of property into the commons, redesignating the upper air as a highway. There were no losers from this change (my comment: really?). For technical reasons, we can never allow any opt-out from the air rights cutoff; only 1-2% optout would destroy the system.

This isn't the case with Google. No library requires all the books in the world to operate. Thus there's no overflight problem.

Back to the beginning: if the evidence is unclear (as it always is in copyright), people stick with their initial presumptions. Shall it be property or commons?

Long post already; comments period next. Possibly with more articles!