The Laboratorium
January 2009

What’s Going On with Google Book Search’s Terms of Service?


Sometime in the last few weeks, Google quietly changed its guidelines on what you can do with the PDF versions of public-domain books it makes available. The new rules now warn, “Don’t sell digital or physical copies, or help other people buy and sell them.” Given that these are public-domain books, why is this any of Google’s business? I heard about the change from Yakov Shafranovich, who runs PublicDomainReprints.org and is rightly concerned that he may be forced to stop supplying the public with copies of books they have every right to read.


Kevin Drum: “Our long regional nightmare is finally over.”


This rant against Billy Joel at Slate is officially credited to Ron Rosenbaum, but reads more like the transcript of a drunken barroom tirade. It’s so unhinged, you’d think that Billy Joel sadistically murdered the entire Rosenbaum clan while young Ron looked on helplessly. It also shifts tone abruptly and repeatedly, as though the only editing it received was a note reading, “We paid for 2,000 words but this is only 500. Maybe you could add a digression about Jeff Jarvis or an interlude in the form of an imagined dialogue with the reader?”

Planet Google Book Club


My apologies for the late notice, but I was part of a book club at TPM Café: a roundtable discussion of Randall Stross’s new book, Planet Google.

Famous for Puzzling


A crew from G4’s Attack of the Show hung out with my team at the Mystery Hunt this weekend. Although they shot several hours worth of footage, it was boiled down into a three-minute segment that aired earlier this evening. I make a brief speaking appearance.

Obama Keeps His Blackberry


Third, messages from the president will be designed so they cannot be forwarded.

Does anyone know how they implemented this feature? Restricting email distribution is a hard problem. Will Obama only be emailing with a small set of locked-down White House blackberries? Did the reporter get the facts wrong? Is someone blowing smoke? Enquiring minds want to know.

Virtual World Feudalism


I have a short (1500-word) essay online at the Yale Law Journal Pocket Part as part of a new symposium issue on virtual worlds. My contribution, “Virtual World Feudalism,” looks inside the common metaphor of virtual worlds as “feudal societies” and concludes that feudalism might not be such a bad thing for them. Feudal property—based on a personal relationship between a lord and a vassal—turns out to be a pretty good fit for the current state of Second Life. It gives players some stability and a powerful ally while giving Linden Labs the flexibility it needs to keep he world running.

I’m extremely happy with how this one turned out. There’s something about short-form academic writing that’s especially satisfying, and my editors genuinely cared about making every word count. t also provided me a good excuse to renew my acquaintance with S.F.C. Milsom and some of the other fascinating scholarship on feudal property law. The resulting piece is a bit of a provocation, but it’s also a serious argument about the purpose and nature of “property.”


Res publica refecta est.

Annals of Poorly-Targeted Marketing


“Revive & Rejuvenate at the Women’s Wellness Weekend”

And We’re Back


My computer is working again, no thanks to Apple. The email backlog is long, and I’m away this weekend at ZyzzlCon 3009, so please be patient if you’re expecting something from me.

Ancient Order of Hibernation


It’s still winter break at school, so this is as good a chance as I’m likely to get for a little electronic hibernation. I’m taking my computer in for service and it’s likely to be out of my hands for a few days at least. Expect Lab updates, email correspondence, and other signs of my digital life to be sporadic at best.

The Best Things in Life Are DRM-Free


So the iTunes Store will be DRM-free. That’s not big news. No, the big news is that the iTunes Store is also going to 256kbps, and that, with variable pricing, many songs are going to 69 cents. With these changes, iTunes will leapfrog past Amazon’s (also DRM-free) MP3 Store in quality and price.

The message is: don’t buy music. The price is falling, and it’s falling fast. It’s not just the sticker price, either: with higher bitrates and no DRM, the annoyance surcharge is also falling. Sure you can buy now at a bargain compared with last year, but next year’s bargains are likely to be even better. Wait.

Think of it like the housing market. The record companies are sitting on an inventory of dilapidated mansions. Sure, they could try to wait for the end of the the bad times, but you can’t eat mansions. The record companies need to do whatever it takes to keep from going out of business now, and if that means dropping the price to keep the inventory moving, so be it. Trouble is, all the other record companies are also trying to survive, and they have their own mansions to move. The rational buyer, like Cadmus, should sit this one out.

I’d say there’s a better than even chance that within the next five years, almost all recorded pop music becomes free. This enormous overhang of recordings will have very interesting effects on contemporary creativity. Existing music is both a complement to and a substitute for new music. We’ll see more Girl Talks: musicians who take inspiration from an insanely rich knowledge of what’s come before. But on the other hand, why should you pay for this month’s hit single when you can get last month’s, which sounds almost the same, for zilch?

Research Center on Fictitious Economy and Data Science


Before clicking the link, guess whether this is about virtual worlds or about finance and credit.

Privacy and the New Virtualism


Law review article making the argument that online privacy ought to be conceptualized as privacy of the embodied digital self. I’m not sure that this move actually simplifies the analysis, but it had me saying “hmmmm.” The point that we can’t theorize how the Internet changes privacy without also theorizing how the Internet changes us is elegantly argued and bears repeating.

Why Is Google Restricting Access to Government Books?


Per 17 U.S.C. § 105, any “work of the United States Government” is in the public domain from the moment it’s created. And yet, Google Book Search restricts access to snippet view: short one-inch sections around your search term. Compare the full download available for other clearly public-domain books.

My best guess is that Google hasn’t yet figured out that these government reports actually are public-domain. Here’s a hint for a future update to their copyright database: if it’s published by the U.S. Government Printing Office, it’s probably not in copyright. Open-source America’s operating system!


A simple rule for making sense of the financial crisis and the recession: ignore anyone who speaks with complete certainty.

Lavinia I give it 2 stars


It seems like a good idea: a gentle tribute to Virgil that gives voice to one of the Aeneid’s flatter characters. The problem is that the last six book of the Aeneid are the boring half, no matter whose eyes they’re seen through. It’s like having ringside seats to the All-England Summarize Proust Competition: interesting only to the extent it compresses the original.

And the Best Student Films Are Made with, Um, Borrowed Equipment


The DigiPen Institute of Technology has a policy that it owns the copyright in any games its students submit for coursework. At first blush, it sounds outrageous—a digital sweatshop disguised as a school!—but there are in fact non-exploitative reasons for the rule:

“We are not here to compete with the games industry,” [DigiPen president Claude Comair] says. “We are not here for people to come and make a game in a less-expensive manner utilizing equipment and software that has student licenses.”

“Just as importantly, we are not equipped to properly firewall our projects in the sense that we really don’t know legally speaking how many or which students created which games. We don’t know whether they received input from other students who have not been credited.”

“These are just a few of the reasons why we have this policy,” he adds, “but the bottom line is that DigiPen has never sold any of its students’ games nor do we intend to.

Whether you agree or not (and I don’t), it’s an interesting perspective on how IP rights create both gigantic opportunities and gigantic messes.


There is a woman down on the street calling out for “Jimmy!” every few seconds.

I am tempted to open the window and shout back, “Run, Jimmy! It’s a trap!”

But it’s winter out, so temptation will have to wait.

Parsing the Seventeenth Amendment


Norman Williams raises an interesting issue in interpreting the Seventeenth Amendment. Here’s the relevant passage:

When vacancies happen in the representation of any state in the Senate, the executive authority of such state shall issue writs of election to fill such vacancies: Provided, that the legislature of any state may empower the executive thereof to make temporary appointments until the people fill the vacancies by election as the legislature may direct.

Note that:

  • “Vacancies” is always plural.
  • “Representation” is singular .
  • “Any state” is always singular, and so is “such state.”
  • “Executive authority, “legislature,” and “executive thereof” are singular.
  • “Writs of election” and “appointments” are plural.

Williams argues that the plural use of “appointments” suggests that a governor, empowered by the legislature to make an appointment to fill a Senate vacancy, is in fact empowered to make multiple appointments. Thus—here comes the punch line—once Blagojevich is impeached and removed from office, his successor will be able to make another appointment to fill Barack Obamas Senate seat. This second appointee would therefore simply replace Burris, making the whole problem go away.

Howard Wasserman takes this argument apart by pointing out that “vacancies” is also plural. Thus, the “appointments” simply mirror the “vacancies,” and no such inference from the use of the plural is warranted.

Williams then claims that the text would have been unambiguous in requiring only one appointment if it had said “a temporary appointment until the people fill the vacancies by election,” so the choice not to say that must be meaningful. The counter to that argument, however, is that “a temporary appointment” would (by the same kind of reasoning) be telling us that the governor is only allowed to make one appointment, even when there are two vacancies.

In fact, there’s an even more striking reading of the Amendment, one that Wasserman passes over as “unlikely” and Williams doesn’t even mention. If “vacancies” is plural but “state” is singular, why are we even bothering with an appointment to fill Obama’s seat? Dick Durbin is still in office; Illinois only has one “vacancy.”

I can even spin out a story about why this is the most structurally plausible reading. The point of this passage of the Seventeenth Amendment is to make sure that no state is wholly unrepresented in the Senate, I’d say. Being down a Senator is acceptable, because the other can speak for the state. Only when the state is down both Senators does the election-or-appointment clause kick in.

This argument has got to be wrong, but its wrongness is no longer a matter of the kinds of textual exegesis or democratic-republican first-principles you see in so many constitutional arguments. This is why I hate most modern constitutional law scholarship that seeks to interpret particular provisions. Having relied on such profoundly junky forms of argument, it no longer has the intellectual purchase to reject obviously fallacious counter-arguments.

I know just enough linguistics to know that lawyers can quickly get out of their depth when they try to make linguistic arguments about legal texts. When a scholar with actual linguistic training looks at a disputed phrase, she doesn’t see two possible meanings; she sees nine. I would be interested in hearing from any linguists in the readership who could give an analysis of the number of acceptable readings of this passage. It’s been too long since my own computational linguistics course for me to do this properly myself, but I’d note the following:

  • Does the scope of “when” run to the entire sentence, or only through the colon?
  • I’m inclined to say that any reading in which “the vacancies” in the second half doesn’t refer back to “vacancies” in the first should be starred.
  • In the second half, “as the legislature may direct” is susceptible of both high and low attachment.
  • We also need to decide where to attach the entire second half; perhaps the executive authority should only issue writs of election if the legislature of the state isn’t allowed to empower the executive to make appointments.
  • Should we imply an “authority” to follow the second “executive,” or treat the second occurrence as a noun? If the latter, perhaps they need not refer to the same thing.
  • Presumably, “writs of election” don’t in and of themselves “fill such vacancies,” or we could have the Senior Writ from the Great State of Illinois.
  • “Any state” appears twice; are you certain that both instances refer to the same state?

An Informal Rant About Formalities


One of the most infuriating details of the Google Book Search lawsuit is the way the plaintiff class is defined: everyone who owns a “Copyright Interest” in a “Book” under United States law. That sounds fine, and most of it is, but lurking in the definition of “Book” are two surprises. First, for stuff published in the United States (technically, a “United States work,” which is a slightly larger category), it’s not a “Book” unless it’s also been registered with the Copyright Office. Second, if it was first published somewhere else, the copyright owner is part of the lawsuit even without registration.

Both of these rules are unintuitive. The registration rule means that a fair number of authors and publishers won’t be part of the settlement at all. The new Registry isn’t authorized to collect money and pay them; Google isn’t released from copyright liability for scanning and distributing their books. If the settlement is good for books that have been registered (and it is), it ought to be almost equally good for books that haven’t been. It costs $35 and up to register with the Copyright Office; that fee alone will cut off a part of the Long Tail of books the settlement tries to set up.

As for foreign works, the consequences are even more surprising. If your book has been published in another country but you’ve never even thought about the United States, well, congratulations. You’re part of the plaintiff class, which means that when the settlement goes through, your book will be online, searchable, part of library subscriptions, and for sale. If you object—or if you just want your share of the money—it’s on you to get in touch with this new American institution. A registration threshold might have restricted the lawsuit’s impact to foreign copyright owners who had some inkling that they had something big at stake over in the U.S. But no, everyone’s in the hotchpot now.

This is why the settlement notice will require extensive international advertising to tell copyright owners around the world about the consequences for them. It’s also why security provisions about restricting access to United States users will be important in the new digital search-and-download world. Your U.K. publisher, which thought it had a territorially exclusive publication right may not be very happy to have competition from the Google download service. It’s not clear to me that authors and publishers around the world have grasped how quickly copyright’s territorialism may be about to crumble.

If you wanted to, you could make an argument that these rules—U.S. books require registration, while foreign ones don’t—are exactly backwards. After all, isn’t it American authors who might be more expected to be aware of Google’s English-language scanning efforts and better able to collect from a U.S.-based collective rights organization? Whereas, given the difficulty of tracking down authors worldwide, wouldn’t requiring registration assist in that task? Instead, it’s the local authors who have to put up with the registration hurdle, whereas foreign authors are deprived of its protections.

Let me be clear. These complications aren’t the fault of Google or of the plaintiffs. They’re the fault of United States copyright law. We’ve blundered our way through a series of inconsistent rules about “formalities” like registration (and also the notice of copyright you see on the publication information page of most books). The bill for this folly is coming due.

The context here is that the U.S. has long been an exception to the international rule that copyright doesn’t depend on formalities. According to the 1886 Berne Convention, copyright sprang into being as soon as the work was created, and those rights “shall not be subject to any formality.” The U.S. rule that you had to register your work (and, later on, renew that registration), put a notice of copyright on each copy, and supply deposit copies with the Library of Congress, made it an outlier. Particularly under 1909 Copyright Act, many authors were bitten badly by the U.S.’s rule; they’d fail to comply with one of the formalities (mistakes in the notice were especially common) and thereby forfeit copyright. Everyone else laughed at us. We laughed back; the rules gave some certainty to potential defendants; if you confirmed that a work had never been registered, you could feel comparatively safe in running off copies of your own.

After a century of mutual snickering, the U.S. joined the Berne Convention and, in 1988, really dropped the formalities. (We’d taken some steps in that direction before, but the 1988 changes gave us the system we’re living under today.) Registration became optional.

Well, almost optional. You have a copyright as soon as you put pen to paper, but you can’t sue on that copyright until you register. Congress wanted to hang on to the evidentiary value of the registration records and to keep the Copyright Office involved, but it also wanted to be part of Berne. Requiring registration prior to suit seemed like an acceptable compromise; you can get a registration request processed quickly if you’re trying to file suit, so it doesn’t really harm the copyright owner.

Even still, registration was still a “formality,” which led to a further compromise. Registration before suing would only be a requirement for United States works; owners of copyrights in foreign works could just go ahead and sue directly. It was a clever dodge: Berne required only that foreign authors be treated as well as local ones, and here, the United States was treating foreign authors even better than it treated its own. And thus, the bifurcation we see in the Google Book Search suit—registration for U.S. works, but not for foreign ones—was born.

But here the mischief begins. The courts have interpreted the registration requirement as “jurisdictional”: that is, they can’t hear copyright disputes about unregistered works. That sounds all well and good, but it has some ugly side effects. If the court can’t pass judgment on unregistered works, it means it can’t fully resolve a case that involves both registered and unregistered works.

Imagine a plaintiff who takes cat photos and registers some of them, and a defendant who starts a lolcat site with the plaintiff’s photos. The court can order the defendant to take down the registered photos, but can it order the defendant to take them all down? Not if you’re a jurisdictional purist, it can’t. The court would have no power over the unregistered ones! The plaintiff needs to file registrations for them, and only then turn around and sue again. (Imagine the chaos as to the photos the plaintiff hasn’t even taken yet.)

That one might sound like the plaintiff’s fault, depending on how you spin it, but what about this one? Take a class of plaintiffs, some of whom have registered and some of whom haven’t. Is this a proper class? In a continuation of the famous Tasini case, the Second Circuit said no. Freelance writers had sued the New York Times and other publications for putting their articles into electronic databases such as Nexis. After the Supreme Court held that this might be copyright infringement, the parties hammered out a settlement that divided up a pool of money among the various writers.

You can probably see where this is going. Some of the authors involved had registered their copyrights. Others hadn’t. The Second Circuit decided that the courts had no authority to hear a class action in which a substantial part of the class had no individual right to be in court in the first place. The settlement was bounced. Keep in mind that, since we’re talking about individual articles, the registration fees might well have exceeded the payout these authors could expect. The result is that the register-before-suing doctrine eliminated the one mechanism that could practically have resolved their perfectly valid claims of infringement of their perfectly valid copyrights. Formalities = ouch.

It’s not just plaintiffs who suffer from the rule. In a recent case, the Eleventh Circuit, taking the “jurisdictional” theory seriously, held that a potential copyright defendant can’t go to court for a declaratory judgment of noninfringement of an unregistered copyright. That means that a clever copyright owner could hold off on registering while it sent threats and demand letters to a victim. The victim can’t go to court for a declaratory judgment that it’s safe to proceed with whatever it wants to do. But the copyright owner, any time it wants, can fire off the registration paperwork, show up in court, and get the party started.

Put these two rules together and you see the issue with the definition of the plaintiff class in the Google Book Search settlement. There are who knows how many owners of unregistered copyrights in books. But there’s no good way to bring them into court as part of what’s supposed to be a comprehensive settlement. The court doesn’t have authority, on the jurisdictional theory—binding on it, since it’s in the Second Circuit—over all the owners who haven’t registered. And, on the Eleventh Circuit’s reasoning, there’s no way that Google could bring them in, either, by filing for a declaratory judgment. They’re outside in the cold, looking in. And—to recall the absurdity of the situation—keep in mind that foreign authors and publishers are inside, tied up right next to the fire, whether they’d like to be there or not.

That’s why I’m convinced that the U.S. formality system, such as it is, is irreparably broken. The “compromises” of our accession to the Berne Convention fit with neither the policies of formalities nor the policies of rejecting them. Worse, they create gigantic, unnecessary divisions between classes of copyright owners. I’m attuned to the consequences for the Google Book Search project, because that’s where I roll, but that’s far from the only damage. Someone oughta do something about it.

As for the Google Book Search settlement itself, I’m still pondering ways to patch things up. The antitrust and discrimination concerns that arise when we slice up copyright owners in this way are real. Perhaps that reality is the ground on which intervention may be justified. We might not be able to require that all copyright owners, past and future, registered and unregistered, be brought into the class. We might, however, be able to require that the same terms authors in the settlement class are offered be extended to all other authors who wish to join in. Or Congress could bless the new order of things and legislatively bring everyone in. (It should be obvious to students of copyright law and politics why I regard that last option as a last resort.)

Who’d have thought that it was possible to get so worked up about copyright formalities? Until I put these pieces together, I wouldn’t have. Perhaps I haven’t convinced you to get equally outraged, but have I at least convinced you that one can?

The Weeds of Con Law


I’ve never been much of a con law junkie. The field sits at an unpleasant intersection of intense partisan political pressure and excessive academic attention. Too many lay observers assume that any decision they disagree with is a cynical and lawless ploy; too many scholars split hairlets arguing over which theory best justifies decisions everyone agrees with. Spend too long debating the same few words, and a version of the madness that James Thurber described sets in:

I fell to repeating the word “Jersey” over and over again, until it became idiotic and meaningless. If you have ever lain awake at night and repeated one word over and over, thousands and millions and hundreds of thousands of millions of times, you know the disturbing mental state you can get into. I got to thinking that there was nobody else in the world but me, and various other wild imaginings of that nature. Eventually, lying there thinking these outlandish thoughts, I grew slightly alarmed. I began to suspect that one might lose one’s mind over some such trivial mental tic as a futile search for terra firma Piggly Wiggly Gorgonzola Prester John Arc de Triomphe Holy Moses Lares and Penates. I began to feel the imperative necessity of human contact.

These forms of madness, however, are primarily confined to the big and enduring controversies of constitutional law. The Constitution also contains plenty of other, more endearing provisions. (For pure phrasal catchiness, I’m fond of the prohibitions on titles of nobility and corruption of blood.) Precisely because they’re less-litigated and less all-consuming, it becomes possible to have polite and rational conversations about what they actually mean and require.

The last two months have been especially good at kicking up genuine, interesting issues from the forgotten Constitution. Are Barack Obama and John McCain “natural born Citizen[s]” of the United States and therefore eligible to be President? Can Hillary Clinton be appointed Secretary of State, even though “the Emoluments whereof shall have been increased during” her time as Senator? Does George W. Bush’s “Power to Grant Reprieves and Pardons for Offenses against the United States” allow him to retract his attempted pardon of Isaac Toussie? And can the Senate, as “Judge of the Elections, Returns and Qualifications of its own Members,” refuse to seat Roland Burris?

I’d say yes, yes, it depends, and no. But I can’t tell you what a relief it is that at least some of the arguments about these issues are being joined on the basis of Constitutional text and structure, judicial precedents, and historical practice, rather than obvious ideology or grand high idiosyncratic theories.