Inside Judge Chin's Opinion

By James Grimmelmann

Posted on March 24, 2011

I have now had the chance to go through Judge Chin’s opinion rejecting the proposed Google Books settlement for the third time. I am struck by how much — and how little — it says. Its holding is clear and direct: the settlement “would release Google (and others for certain future acts … an arrangement that exceeds what the court may permit under Rule 23.” (21) The legal analysis supporting this conclusion takes perhaps five pages out of forty-eight. The rest of the opinion is … well, it’s complicated.

The opinion discusses seven kinds of objections to the settlement:

1. Notice
2. Representation
3. Future releases
4. Copyright
5. Antitrust
6. Privacy
7. International
According to my reading, “Future releases” is the only one of the seven that Chin holds to be a sufficient ground for rejecting the settlement on its own. He dismisses notice and privacy relatively quickly, writing, “I am satisfied that the class received adequate notice” (19), and while the “privacy concerns are real,” they are not “a basis in themselves to reject the proposed settlement.” (39) The other four all receive various expressions of concern, but even as Chin repeatedly pronounces himself “troubled,” he avoids making direct statements that they require rejection. Instead, there is only a “substantial question” about the adequacy of representation (20), he “need not decide” the most specific copyright question (32), he “need not decide whether the ASA would violate international law,” (43) and I can’t find anything in the antitrust section that constitutes a conclusion of law. (The only cited case is a “Cf”!)

What is going on here? The future releases issues are sufficient by themselves to reject the settlement; indeed, having concluded that the settlement “exceeds what the court may permit under Rule 23” (21, emphasis added), Judge Chin left himself no choice but to reject it. The rest of the opinion provides reasons to support that result — but the opinion is cagey as to which of these are additional legal rationales, and which are just policy arguments.

I can imagine two reasons for this caginess. The first is that Judge Chin deliberately wrote a minimalist opinion. If hard cases make bad law, then perhaps big cases make strange law. It may be Judge Chin resolved this one on a relatively narrow basis so he could avoid the need to make detailed rulings on dozens of different issues at once in the context of a highly unusual case. The other possibility is that the ultimate legal question Judge Chin faces — is the settlement “fair, adequate, and reasonable?” (14) — is a flexible all-things-considered standard. His point, then, is not that there are a dozen reasons why the settlement must be rejected, but rather that there are a dozen important considerations all tipping against the settlement, which, taken together add up to an easy call: reject.

The other thing that struck me immediately about the opinion was the remarkable diversity of objectors whose views it cites. It quotes from dozens of different filings, including one notable passage on pages 33 and 34 that pull together concerns from four authors and a pair of literary agents about the settlement’s opt-out structure. This is a quietly effective piece of judicial rhetoric: it emphasizes the range of objectors as well as the range of objections. This goes along with its emphasis on the “great in number” objections and the “extremely high number of” opt-outs: persuading the reader that class members disapprove of the settlement.

While I’m talking about citations, let me also congratulate Emily Anne Proskine and Andrea Glorioso, both of whose student notes were cited in the opinion. Even on such a big and complicated case, good student work can make a valuable contribution, and theirs did. Three of the pieces in our D is for Digitize symposium issue were cited, as was — permit me a moment’s pride — our supplemental letter the court.

That’s the high-level view. Let’s look into the details.

Future Releases

I’ve long believed that the Rule 23 aspects of the settlement are absolutely central to what make it both powerful and dangerous. Judge Chin agrees. He starts by making a crucial distinction:

The ASA can be divided into two distinct parts. The first is a settlement of past conduct and would release Google from liability for past copyright infringement. The second would transfer to Google certain rights in exchange for future and ongoing arrangements, including the sharing of future proceeds, and it would release Google (and others) from liability for certain future acts. (21)
The emphasis is Judge Chin’s. This passage is the key to the settlement, and the key to the opinion:

I conclude that this second part of the ASA contemplates an arrangement that exceeds what the Court may permit under Rule 23. (21)
By way of contrast, he cites — seemingly with approval — the United States’s position that the past-conduct half of the settlement could appropriately be settled. Throughout the opinion, almost every time Judge Chin raises an objection, that objection has bite only as applied to the future-conduct portions of the settlement. Thus, the question this part of the opinion resolves is how far a court can go in approving a settlement in which a class gives up claims against a defendant for things it hasn’t done yet.

The legal analysis here turns on the relationship of those future-conduct claims to the underlying lawsuit. Here, the parties had been arguing that the scope of the underlying lawsuit should be determined by looking at their complaints against Google, which they construed broadly to involve Google’s “digital copies.” (25) But Judge Chin rejects that move, writing instead:

This case was brought to challenge Google’s use of “snippets,” as plaintiffs alleged that Google’s scanning of books and display of snippets for online searching constituted copyright infringement. … There was no allegation that Google was making full books available online, and the case was not about full access to copyrighted works. The case was about the use of an indexing and searching tool, not the sale of complete copyrighted works. (24-25)
A page later, he adds:

Google did not scan the books to make them available for purchase, and, indeed, Google would have no colorable defense to a claim of infringement based on the unauthorized copying and selling or other exploitation of entire copyrighted books. Yet, the ASA would grant Google the right to sell full access to copyrighted works that it otherwise would have no right to exploit. (26)
This move is what sinks the settlement. Chin has set up a dichotomy: Google’s past conduct in scanning and searching was the subject of the lawsuit, but it is Google’s future conduct in selling whole books that would authorized by the settlement. The case “was not about” the same things the settlement is. Thus, under the two governing precedents — the “identical factual predicate” test of Wal-Mart Stores and the “general scope of the case made by the pleadings” test of Firefighters — the settlement goes beyond the underlying lawsuit, and is therefore impermissible.

This strikes me as broadly correct, but the legal analysis is surprisingly thin. If it had been me, I might have said that Firefighters is simply not on point because it concerned promises by the defendant, not releases by the class. I would have added that the reason the Wal-Mart “identical factual predicate” test makes sense here is that it is grounded in preclusion law: win or lose, a lawsuit resolves all of the plaintiffs’ claims against the defendant that are based in the same facts.

The “same facts” line helpful distinguishes Google’s actual past scanning from its hypothetical future book sales. Since Google hasn’t sold complete books — and wouldn’t without the settlement — the class members couldn’t sue Google on claims connected with whole-book sales. And thus, since those claims couldn’t properly be before the court in litigation, they’re not properly before it in settlement. Q.E.D.

(I’ve been working out this theory for a while now. It also makes sense of ripeness, personal jurisdiction, defendant class actions. If this is how the Google Books settlement ends — not with a bookstore, but a whimper — the theory may need to end up back on the shelf.)

Also in the “future releases” section of the opinion, Judge Chin discusses Congress’s institutional role and representation of class members. (This last one is puzzling, as there is also a separate “adequate representation” section. I’ll discuss it there.) He emphasizes, in several places throughout the settlement, that this is a job for Congress. Interestingly he does so not on Rule 23 civil-procedure institutional-choice arguments, but rather on substantive copyright grounds:

The questions of who should be entrusted with guardianship over orphan books, under what terms, and with what safeguards are matters more appropriately decided by Congress than through an agreement among private, self-interested parties. Indeed, the Supreme Court has held that “it is generally for Congress, not the courts, to decide how best to pursue the Copyright Clause’s objectives.” Eldred v. Ashcroft, 537 U.S. 186, 212 (2003); accord Sony Corp. of Am. v. Universal City Studios, Inc., 464 U.S. 417, 429 (1984) (“[I]t is Congress that has been assigned the task of defining the scope of the limited monopoly that should be granted to authors or to inventors in order to give the public appropriate access to their work product.”). (23)
This stops short of being a straight legal conclusion that the court cannot pass on such matters. It reads more like a prudential argument: the court shouldn’t.


Now that we’ve seen the five pages of future-conduct arguments that doomed the settlement, let’s turn back to the other forty-plus pages. I’ll take up the issues in the order Judge Chin does. He is strikingly unconcerned with notice — indeed, this is the one part of the opinion that I count as a clear win for Google and the plaintiffs:

The objections to the adequacy of the class notice are rejected. I am satisfied that the class received adequate notice. More than 1.26 million individual notices in thirty-six languages were sent directly to copyright owners, potential class members, and publisher and author associations worldwide. Plaintiffs also established a website to provide information about the case, the original proposed settlement, and the ASA. Of course, the case has received enormous publicity, and it is hard to imagine that many class members were unaware of the lawsuit. (19)
This last phrase reads as though it were directed to Scott Gant, who pushed the notice issue vigorously at the fairness hearing, only to be asked, “You’re here, though?


Judge Chin has some nice words for class counsel, whom he calls “qualified, experienced, and able to conduct the litigation.” (20) That’s just the setup, though, because he then states that there is a “substantial question” (20) as to whether the named plaintiffs adequately represented their peers, and that the differences of opinion among class members are “troubling.” (21) (Note that the “substantial question” is asked but not answered.) He identifies three groups who may not have been adequately represented:

First, academic authors care about access to knowledge; commercial authors and publishers (like the named plaintiffs) care about revenue. This point is a coup for Pam Samuelson’s academic-author coalition, for whom it was a central argument.

Second, foreign rightsholders have interests “that may be at odds” (29) with the class representatives for reasons that are alluded to but never quite made explicit. I’ll discuss this below, but while the opinion tallies their objections to the settlement, it never quite identifies how their interests qua interests diverge.

Third, “rightsholders who do not come forward to register” (29) (e.g., orphan copyright owners) don’t have interests that align with the parties to the lawsuit. Instead, the “parties have little incentive to identify and locate the owners of unclaimed works, as fewer opt-outs will mean more unclaimed works for Google to exploit.” (29) Although the direct financial incentives that encouraged claiming rightsholders not to look for unclaiming ones are gone, I think this point is correct. A rightsholder who doesn’t step forward can’t opt out, which means more power for the Registry, more revenue for Google, and a venture more likely to achieve critical mass for the other rightsholders.

On this last point, the parties had argued that orphan owners here shouldn’t be thought of any differently than any members of any other class who don’t step forward — every class action, almost by definition, will have absent class members. I’ve been arguing against this view for two years now. I think that the fact that we know with substantial confidence that many copyright owners cannot and will not be found makes this class action particularly troubling, and that the fact that their future claims are being compromised compounds the trouble. In a subtle and important passage, Judge Chin comes down on the “orphans are special” side of the debate:

While it is true that in virtually every class action many class members are never heard from, the difference is that in other class actions class members are merely releasing “claims” for damages for purported past aggrievements. In contrast, here class members would be giving up certain property rights in their creative works, and they would be deemed — by their silence — to have granted to Google a license to future use of their copyrighted works. (30)
If you ask me for proof that Judge Chin gets it, I’ll cite this passage. Future-conduct releases by individuals are okay, because well-informed individuals can make choices for themselves about the future. Past-conduct releases by a class are okay, because the most they have to lose is possible compensation for harms that have already happened to them. But the combination of future-conduct releases and a class with known absent members is explosively dangerous, because so much more is on the table for them: they could lose important property rights, or have fresh harms inflicted on them.


This section is an interesting mix of policy and statutory arguments. It starts with a reprise of the “copyright is a job for Congress” song from Act I. Then it develops an argument that the settlement is an “expropriation” of class members’ copyrights in violation of Section 201(e) of the Copyright Act. I tend towards the parties’ position here: the settlement is a license, not a transfer of ownership, so this section doesn’t apply at all. Judge Chin, though, finds it “at a minimum a fair question” (31), although he ultimately concludes that he “need not decide the precise question.” (32)

I found this passage notable also because it illustrates David Nimmer’s advocacy at work; you can see why Amazon is paying him what I assume are the big bucks. I scoffed at the § 201(e) argument when he made it in his briefing for Amazon, but here it is in the opinion. What authority does the opinion cite to bolster its analysis? “3 Melville B. Nimmer & David Nimmer, Nimmer on Copyright, § 10.04 (Rev. Ed. 2010).” (32) Self-link FTW!

This leads into perhaps the most impassioned part of what is generally a fairly staid and restrained opinion: a discussion of whether a settlement could ever impose an opt-out system of assumed consent on copyright owners. It quotes liberally from Nimmer’s oral argument from individual authors who opposed the settlement, and from an “eloquent[]” (35) letter from two literary agents. All of them object to the idea that the burden is on them to step forward if they wish to object, instead of the burden being on Google to secure their permission. Judge Chin’s summary paragraph is worth reading closely:

While the named plaintiffs and Google would argue that these authors can simply opt out, the comments underscore certain points. First, many authors of unclaimed works undoubtedly share similar concerns. Second, it is incongruous with the purpose of the copyright laws to place the onus on copyright owners to come forward to protect their rights when Google copied their works without first seeking their permission. Third, there are likely to be many authors — including those whose works will not be scanned by Google until some years in the future — who will simply not know to come forward. (35-36)
There’s a lot going on in here. In particular, the point about “authors of unclaimed works” is so subtle I almost missed it. It’s not just a statement that the silent majority of authors who haven’t yet claimed their works would object to the opt-out system, although it is that. It’s also a point that those authors who haven’t claimed their works are not as effectively served by the opt-out option as those who have written to the court to object. Orphans by their nature can’t opt out. That said, I do wonder whether there is a bit of tension between the claim, above, that the notice program was effective and the claim that many authors won’t know to opt out.


This section is thin and a bit odd. It contains only a single citation to caselaw, and that one’s only a “Cf.” The rest is factual analysis and discussion of objections to the settlement. My take, therefore, is that this is an outline of reasons to be concerned about the settlement’s competitive effects, but not a holding that it actually violates any of the antitrust laws. What are those effects? Google would obtain “a de facto monopoly over unclaimed works” (36), the settlement “would arguably give Google control over the search market” (37), and “Google’s ability to deny competitors the ability to search orphan books would further entrench Google’s market power in the online search market.” (37)


Tim Lee tweeted something interesting about this section this afternoon. He said, “I was really confused by the privacy section. Objections were about user privacy, response about rightholder privacy.” Perhaps that lack of clarity explains this section’s waffling tone:

The privacy concerns are real. Yet, I do not believe that they are a basis in themselves to reject the proposed settlement. … I would think that certain additional privacy protections could be incorporated, while still accommodating Google’s marketing efforts. (40)
This is vague and unhelpful. I’m sure the parties would like to know what “certain additional privacy protections” Judge Chin would like to see. The CDT and EFF proposals seem like they’d be a good starting point for discussion, but there’s not even a gesture in the direction of the specific suggestions in their filings.


This section starts by explaining how the settlement is no longer as worldwide as it previously was, but is still broad enough to encompass many foreign rightsholders. It gives a list of countries with objecting copyright owners. The most specific point about how their interests differ from those of domestic copyright owners is a point made by various international publishers, that they will face a more difficult problem searching United States copyright records, due to historical twists in our law on when registration was required.

There then follows a survey of arguments that the settlement violates international treaties … but no legal analysis of whether it does or doesn’t violate them. Instead:

In any event, I need not decide whether the ASA would violate international law. In light of all the circumstances, it is significant that foreign authors, publishers, and, indeed, nations would raise the issue.
I’m not quite sure how to read this. Is it a statement that the fact that foreign rightsholders objected, just by itself, cuts against the settlement? Is it a special sense of deference to foreign nations? Perhaps the best reading is that this is another argument for deferring to Congress:

The fact that other nations object to the ASA, contending that it would violate international principles and treaties, is yet another reason why the matter is best left to Congress.
What Now?

The opinion concludes with a passage recommending that the settlement drop back from an opt-out system to an opt-in system:

In the end, I conclude that the ASA is not fair, adequate, and reasonable. As the United States and other objectors have noted, many of the concerns raised in the objections would be ameliorated if the ASA were converted from an “opt-out” settlement to an “opt-in” settlement. I urge the parties to consider revising the ASA accordingly.
The next judicial step will be a status conference on April 25 at 4:30 PM. That leaves a month for all concerned to figure out their plans for what happens as the case goes forward. I can see three possible scenarios:

  • The parties could try to appeal Judge Chin’s ruling. (This one has procedural complications, which I thought I understood, but now am less sure about.)

  • They could redraft a new settlement and offer it for approval.

  • They could give up on settlement and go back to litigation.

Google is giving noncommittal answers; the authors and publishers are sounding more like they’re inclining towards revise-and-resubmit. Hilary Ware of Google stated via press release:

This is clearly disappointing, but we’ll review the court’s decision and consider our options.
Scott Turow of the Authors Guild stated via press release:

[W]e’ll be studying Judge Chin’s decision and plan on talking to the publishers and Google with the hope that we can arrive at a settlement within the court’s parameters that makes sense for all parties.
And John Sargent of Macmillan stated via press release:

For that reason, publishers are prepared to modify the Settlement Agreement to gain approval. We plan to work together with Google, the Authors Guild and others to overcome the objections raised by the Court and promote the fundamental principle behind our lawsuit, that copyrighted content cannot be used without the permission of the owner, or outside the law.
If I had to bet, I would guess that we’ll end up with a revised settlement drafted to meet Judge Chin’s specification, which will be approved relatively quickly (at least compared to this last go-round). His opinion is short on specific guidance, but it’s relatively easy to extract the essentials. Here’s what I predict:

  • Google is allowed to continue scanning and searching in exchange for cash payments on the order of (but perhaps not exactly) the $60 in the present settlement, and it’s required to provide an opt-out. Very few people have argued that this form of settlement would be beyond the court’s power. The precise explanation of how this would be distinguishable from the present settlement, although quite feasible, will require some nuance and subtlety.

  • The Display Uses — Consumer Purchase, Institutional Subscription, etc. — are either gone entirely ore are offered on an opt-in basis. The difference between these two possibilities is not large, since, in effect, Google already offers an opt-in through the Partner Program.

  • The libraries receiving digital copies are released from liability but are even more tightly restricted in the uses they can make than under the present settlement.

  • The fates of other facets of the settlement such as the Research Corpus, will be hammered out in the negotiations.

My read is that the parties are not enthusiastic about litigation. This has been a long road, they are tired, and the publishing world has moved very quickly from underneath the settlement. They will be happy to have a settlement that lets everyone claim a kind of minor victory, and to be done with the ordeal. A few of the author objectors, who would like to see Google razed to the ground and Mountain View sowed with salt, will continue to object, but most of the others will quietly shuffle away.

And then, the action will shift to Congress. Will Google start putting together a coalition to push for a legislative solution? Who will sign up? What will the proposed compromises look like? Who will oppose it, and with what arguments? And is this the route by which we will get a national digital library?

The Google Books settlement is dead. Long live the digitized book.

The preceding is re-published on TAP with permission by its author, Professor James Grimmelmann. Inside Judge Chin's Opinion was originally published 3/22/11 on The Laboratorium. The Laboratorium publishes under the Creative Commons License.