Scott Gant is an author, a partner at Boies, Schiller, and a self-appointed crusader against the settlement. This Wednesday, he filed a 47-page objection to the settlement. It is the most comprehensive and aggressive attack on the settlement to date; the issues it raises are the most fundamental that anyone has brought up in a submission to the court. His view is that the settlement—or at least the Revenue Models portions of it—is fundamentally unsalvageable.
Gant starts by framing the settlement as “fundamentally a commercial transaction” (5). Compare the framing in our amicus brief, where we call the settlement’s approach “prototypically legislative.” We both agree that the settlement’s complexity takes it out of the realm of ordinary class actions. But he focuses on the commerical rights involved, whereas we focus on the public interests. This is a classic duck-bunny problem; both perspectives are valid. Indeed, if you don’t recognize both components, not only will you miss important dangers of the settlement, but you’ll also miss the reasoning for it in the first place.
Others have called attention to notice problems, most notably the literary estates represented by Andrew DeVore. Gant, however, uses them to go for the jugular.
He starts from a familiar slogan in class action law, “Individual notice must be sent to all class members whose names and addresses may be ascertained through reasonable effort,” quoting from the famous Eisen case. Eisen involved a class action over stock-trading commissions; the six-million-member class contained about two million members whose names and addresses could be extracted from the defendants’ teletype records. The Supreme Court rejected the plaintiff’s attempt to avoid sending them individual notices, even though he made a convincing argument that it would be so prohibitvely expensive as to end the lawsuit. The notice rules have something of a fiat justitia ruat caelum flavor about them.
Gant himself never received individual notice, and to all accounts, neither have many other authors. He points out that the Copyright Office has records of copyright registrations, and thus “all of the authors of the works at issue in this case are known or knowable.” (16) Actually, this isn’t quite true; foreign works need not be registered to be in the settlement. Gant’s point, however, stands, because it still applies to the subclass of United States works, which are registered.
In a bit of extremely clever jiu-jitsu, Gant also quotes the studies that Google and the Authors Guild pointed to to argue that there are relatively few orphan works. If it’s possible to find authors with a moderate search in the orphan context, he asks, how can the parties justify not having found authors in the class notice context? He does a similar trick with the claims made about how effective the Registry will be in tracking down rightsholders.
I think this is a powerful argument, but it doesn’t strike me as conclusive. The unexamined term in Gant’s brief is what level of effort is “reasonable.” Note that in Eisen, it was only the two million class members whose teletype records were right at hand who mattered, and that all that was at issue was whether to send an envelope to the last-known address. Gant wants to hold the parties to the higher burdens of tracking down all authors, to carrying out searches that involve more individualized investigation, and to ensuring that the letters actually reach everyone.
My guess is that the notice program was designed to comply with Eisen’s specific requirements. Gant starts off by speculating that “hundreds of thousands, if not millions” (15) of authors didn’t receive notices. By the end of the section, this speculation has become a dead certainty: only a “small fraction of the millions of class members were directly contacted through individual notice.” (22) While the individual mailings weren’t great, I know of enough authors who received them that I doubt things are so unambguously inadequate as Gant claims.
That said, however, this is an area in which even if Gant loses on a bright-line rule, he can fall back to the underlying standard of fairness. He raises arguments about the complexity and size of the settlement and its massive implications for class members. Arguments of this sort should be familiar from our amicus brief. Gant uses them to call for a searching inquiry into the quality of notice. Thus, he turns to the Summary Notice mailed out to authors.
Gant quotes its opening paragraph, correctly noting that it doesn’t say anything abut future claims, only past actions. Thus, he argues that authors whose books have not yet been canned will stop reading right there. His claim that the Summary Notice is misleading depends on this argument, because the very next paragraph opens by saying that the settlement, if approved “will authorize Google to scan in-copyright books.” (emphasis added). I think he’s right that the Summary Notice is unclear and doesn’t make it obvious that the settlement deals with future book sales, but I’ve never in my life met a class-action notice that was clear. The courts have set that bar unfortunately low.
You Call This Fair?
Gant’s strongest arguments, I think, concern his attack on the settlement’s economic terms, though he opens on a weak note. He contrasts the $60 cash payments with jaw-dropping $750 to $150,000 range of copyright statutory damages. Had the case gone to trial, though, the plaintiffs wouldn’t have gotten even $750 a work. They’d have gotten nothing. Google’s scans were fair use for the indexing purposes for which they were used, and $60 a work is a fair and reasonable compromise of that long-shot lawsuit.
He’s on firmer ground raising difficult questions about the Registry. He raises the quite realistic possibility that the Registry will not be economically viable—in which case, participants in the settlement will receive nothing from the forward-looking revenue models. Moreover, he asks what the Registry’s overhead will be and how authors can trust that they will actually see most of their 63%. At the very least, the parties ought to be required to give a clearer accounting of why they expect the Registry’s operating expenses to be acceptably small.
His complaints about the Author-Publisher Procedures are also likely to resonate with many class members. The settlement doesn’t so much look to ascertain who owns the appropriate electronic rights as it does allocate them by fiat. This is, in essence, a side deal between authors and publishers, one that replaces the messy diversity of contacts with a few standardized ones. I like that; it cleans up an anticommons and makes it easier for books to reach paying publics. But I can also understand why some authors don’t like it, particularly those who gave more thought to negotiating electronic rights. I would love to see a detailed walk through the Author-Publisher Procedures that tries to find the worst-case and average-case scenarios for both sides.
Gant’s analysis of the role of the Authors Guild is very sharp and is the most original part of the brief. He observes that while the Authors Guild is named as an “Associational Plaintiff,” the motion for approval of the settlement does not “appear to advance the Authors Guild as a class representative” (35). The authors’ memorandum supporting the settlement focuses on the five individually named authors as “adequate representatives.” This fact is striking, because the Authors Guild has always been the public face of the lawsuit, and it played a central role in the negotiations. This gives Grant a good rhetorical basis for asking what role the Guild—as opposed to the five individual plaintiffs—played in negotiating the settlement. It also lets him raise some more familiar questions about how representative the Authors Guild is of authors in general and how good a deal it struck on their behalf.
Gant also emphasizes the increase in the class from the initial lawsuit—which defined the class in terms of the contents of the University of Michigan’s library—to the settlement—which defined the class in terms of all books. This expansion is one of the reasons some authors, particularly international ones, have been freaked out by the settlement. They had no particular reason to expect they’d be implicated by the settlement until they realized that the settlement class was so much bigger. People don’t like surprises.
I’d also point out that the switch moved the class from being enumerable to being open-ended. You could, in theory, write out the complete University of Michigan catalog, produce a list of authors and publishers, and try to track them all down. But when your class consists of all copyright holders of all books, there’s no way to create such a list, because there is no authoritative master list of books.
Within this immense class, Gant thinks that the Authors Guild and its pets don’t make for appropriate representatives. In our amicus brief, we point to the line between orphan and non-orphan works and suggest separate representation for the orphans. Gant adds another line of division: that between works scanned as of the opt-out date and those not yet scanned. The settlement does take account of the division—only the former are eligible for the $60 cash payments—but Gant argues that they have a conflict of interest, as the former would want to push for the pot of statutory damages at the end of the litigation rainbow. Thus, he seeks not two subclasses, but four, making up the full two-by-two matrix.
I continue to see the orphan works issue as fundamental, for all sorts of public-interest reasons. I’m less convinced by the alleged split between those whose works have been scanned and those who haven’t. Given the weakness of the copyright argument in the scanning itself, the same reasons that make $60 fair also make combined representation reasonable.
Who Can Be Against Trust?
The antitrust portions of Gant’s argument, although excellently written, largely rehearse arguments familiar from elsewhere. (Indeed, Gant attaches a number of the antitrust papers mentioned in this space as exhibits to his filing.) He contributes to the conversation by dropping a footnote suggesting that the right term of art in antitrust law for the Registry may be “joint venture.” He also picks up on Randy Picker’s “no-Noerr” argument with an analysis of the text of the settlement’s Bar Order’s antitrust implications.
Proving Too Much?
Gant makes a few arguments that strike me as implausible. He claims, for example, the settlement is impermissible because it would “transfer class’ members intellectual property rights to Google and its partners.” (10) This, he claims, would “alter the substantive rights of class members,” but the Federal Rules of Civil Procedure (including class actions) are not to “abridge, enlarge, or modify any substantive right.” (10)
This is wrong in two ways. First, it misuses copyright law’s idioms. Copyright distinguishes non-exclusive “licenses” from exclusive “transfers.” The settlement is quite explicit that it gives Google only a license, not any copyright ownership rights. Gant has things exactly backwards when he says that the settlement is a “transfer” of copyrights to Google, “in effect a license permitting specific uses of copyrighted works.” (4). (I and others have worried that the settlement is effectively exclusive, but that’s only the case for orphan works, and it takes more work to reach that conclusion. The switcheroo has the effect of making the settlement sound more dire for copyright owners’ rights than it is.
More seriously, Gant’s argument, if accepted, would strike down the Federal Rules themselves. Of course lawsuits affect substantive rights; that’s what lawsuits are for. Of course settlements affect substantive rights; otherwise, they’d be useless. Of course the Federal Rules provide a framework for lawsuits and settlements; that’s their job. The Rules aren’t supposed to affect substantive rights except by providing the procedural framework to adjudicate them. And that’s exactly what this lawsuit does.
Similarly, Gant overreaches when he argues that the court lacks jurisdiction to approve the settlement. He claims that the main elements of the settlement involve the compromise of “potential future claims” and therefore can’t be justified either as damages or injunctive relief. Thus, he concludes, the settlement violates the case or controversy requirement of Article III.
I looked into this idea when working on our amicus brief and rejected it. Google announced specific plans to scan every book it could. The act of scanning is a prima facie violation of the copyright of each class member (though it’s likely not ultimately an infringement). That’s sufficient to create a live case or controversy between Google and each class member — the test being whether the class member would face a concrete enough threat of scanning to be able to sue Google for declaratory or injunctive relief to stop it. (The only exception I could see would be for class members whose books are obscure enough — not in any partner libraries, perhaps — that the scanning wasn’t imminent.) Once in court, though, parties can voluntarily enter whatever settlement they want. Patent cases, for example, regularly end with the purchase by the defendant of a forward-looking license, which of course would release future claims of infringement. There cannot be a categorical jurisdictional bar on releasing claims for future conduct, as Gant’s argument would imply.
Indeed, try to imagine settling this very case in a way that didn’t involve releasing at least some future claims. Google has scanned copies of books on its servers. Every time it backs up those servers, it makes fresh copies of the books (in a complex database, to be sure, but everything is there). Those copies are prima facie infringements of the reproduction right. Google can’t implement useful search without making at least some copies on an ongoing basis. Any release that let Google run Book Search at all would need to waive some future claims. Gant’s line in the sand against future-claim compromises is untenable.
This is not to say that releasing claims for future infringements would not be problematic on other grounds. Because releasing future claims can leave class members worse off than when they started the lawsuit, any settlement that deals with future conduct requires especially close scrutiny. Here’s what we currently say in the amicus brief:
Second, the Proposed Settlement waives future claims by members of the Settlement Class. “The settlement of future claims has been viewed as an area where there is a need for increased judicial scrutiny … .” WRIGHT, MILLER, AND KANE, FEDERAL PRACTICE AND PROCEDURE § 1797.3. As the Supreme Court stated when rejecting certification of a class of plaintiffs exposed to asbestos, “Many persons in the exposure-only category [i.e. those with future claims] … may not even know of their exposure, or realize the extent of the harm they may incur. Even if they fully appreciate the significance of class notice, those without current afflictions may not have the information or foresight needed to decide, intelligently, whether to stay in or opt out.” Amchem Products, Inc. v. Windsor, 521 U.S. 591, 628 (1997). Here, members of the subclass of orphan work book copyright owners are, by their very nature as orphan work owners, unable to “decide, intelligently, whether to stay in or opt out.” Indeed, the orphan work book copyright owners are even less able to recognize and defend their interests than the future claimants in Amchem. The Proposed Settlement does not merely compromise future claims for past conduct, as in Amchem. Instead, it would release Google from liability for its future conduct: scanning books it has not yet scanned, and selling copies it has not yet sold. It is thus particularly inappropriate for this Court to waive the future claims of the members of the orphan work book copyright owner subclass without a searching examination to ensure that their rights are adequately protected.
There is a genuine concern here, but note the difference. The court should be concerne with the practical effects on the interests of class members, a fact-intensive and nuanced inquiry. It does not need to worry about the categorical jurisdictional bar that Gant advances.
A Few More Fine Points
A few other scattered details I found interesting:
Gant’s sleuthing confirmed that the case was never heavily litigated. The $140,000 in expenses incurred by the authors’ attorneys are, according to Gant, “a tiny amount for a complex case or a large class action — further suggesting no significant discovery or expert work occurred.” (3) He also found some gems in Google’s 10-Q filing, which splits the settlement’s $125 million into a $95.1 million “settlement portion” and a $25.9 million “commercial portion.” (8)
In footnote 25 on page 13, Gant expresses a takings argument not unlike the one in the conclusion of my Unprecedented Precedent.
Gant and lead author counsel Michael Boni are not good buddies. In footnote 27, Gant describes a telephone call he placed to Boni three weeks ago; Boni promised to call back, but never did. In footnote 85 (the final text of the brief), Gant takes his revenge, arguing that even if the settlement is approved, the $30 million set aside for the plaintiffs’ legal fees is excessive and should be reduced.
This is a very good brief. It is the best-written and sharpest filing I have seen in the case. Only the proposed settlement itself tops it in imagination. That said, are places where it goes wrong. I think, for example, that Gant shot himself in the foot by joining other, more persuasive arguments, to his unpersuasive treatment of the Rules Enabling Act. But even filtering out those concerns, this is still a smart and searching attack on some fundamental aspects of the settlement. How fundamental? Let me quote from the second-last footnote:
I take no position at this time about specific provisions of an alternative agreement which I might support. The issue before the Court is whether or not to approve this particular agreement. It is not the responsibility of the Court, or objectors, to rewrite the agreement.
In this, he’s actually allied with Google and the plaintiffs. In the past, they’ve responded to calls for modification of the settlement by insisting that this is an up-or-down call for the judge to make. Gant continues:
However, the Court might instruct the settling parties to decouple the two components of the Proposed Settlement — the resolution of claims for past infringement, and the commercial transaction involving the transfer of rights from copyright owners to Google and others. If the settling parties believe the terms of the commercial component of the Proposed Settlement are attractive, they are free to implement them outside the context of a Rule 23 proceeding, and invite copyright owners to participate.
Let us be clear. If the settlement goes down this path, the orphan works component of it is dead. That part works only because of the class action. In an ordinary negotiation, you can build an e-book service, but you can’t bring orphan works into it. That would be a great loss—not just for readers, but for society. And that is why I continue to believe that the goal must be to salvage this settlement, to take the flaws in it that Gant and others have identified, and fix them.