10 May 2009

Google Library Project Settlement (tangent)

How You Lose Matters

The Settlement (in essay form)
The Lawsuit (in essay form)

One of the major problems with the Google Library Project settlement is buried in the field definition: It is limited to books for which there is a valid, and renewed, copyright registration. This limitation comes from a long, misguided series of opinions — largely from the Second Circuit, and adopted without all that much independent analysis elsewhere — holding that § 411(a) of the Copyright Act of 1976 is a jurisdictional gatekeeper; that is, the certificate of registration is the sine qua non of federal jurisdiction over copyright claims.

This is clearly wrong, but understanding why requires both reading the whole statute and having some understanding of what subject-matter jurisdiction really is. It is, instead, a historical-continuity argument, because it continues the jurisprudence of what was properly before a court under the 1909 Act. I have always been under the apparently mistaken impression that one looks at the statute that applies now first.13 Instead, this line of cases from the Second Circuit — stretching back to the 1940s and an incoherent set of both district and appellate decisions regarding photographs, during the period when ownership of the master negative was considered ownership of the copyright itself — actually relies upon the appellate courts' collective merging of the certificate of registration and the copyright itself under the 1909 Act.

As I've mentioned before, this matters to the GLP settlement: A court cannot enforce a settlement over which it has no jurisdiction, but can enforce a settlement for incomplete claims over which it does/would have jurisdiction. Justice Thomas's unanimous decision for the Court earlier this week in Carlsbad Technology, Inc. v. HIF Bio, Inc., 556 U.S. ___ (04 May 2009) comes at this particular question rather sideways, but nonetheless instructively... because it implies that the Court is well along the path of not calling a statutory requirement "jurisdictional" unless the statute explicitly says it's jurisdictional.14 Although this does represent a slight retreat from the purportedly federalism-respecting trend of the early and middle Rehnquist Court, keep in mind that most of those decisions revolved around a subset of federal jurisdiction: State immunity.15

Although there is little doubt that the jurisdictional interpretation of § 411(a) represents the accepted wisdom, it does not represent a clear paradigm distinct from "that's the way it's always been". (It is well beyond the scope of this forum, but that is a valid criticism of much of the Second Circuit's copyright jurisprudence since adoption of the 1976 Act: That the past is to be followed, even in the face of both legislative and treaty changes since.) That jurisdictional interpretation is a fatal defect in the GLP settlement, which fails on procedure, on substance, and on simple fairness. Carlsbad Technology makes clear that "accepted wisdom" is not good enough for this Supreme Court on defining the jurisdiction of the courts... because the accepted wisdom (at least in the Second and Ninth Circuits, where a disproportionate amount of copyright litigation occurs) was that a remand of nondiverse state-law claims once all federal-question and diverse state-law claims had been dismissed (or otherwise resolved) was a "jurisdictional" decision insulated from appellate review (for reasons that we don't need to go into here).16 Carlsbad Technology, however, holds that it is an exercise of judicial discretion and not a question of jurisdiction per se.

So, then, what do all of those technicalities mean for the GLP settlement, given the grant of certiorari in Muchnick (the Post-Tasini parallel case on periodicals) to decide whether § 411(a) is jurisdictional? First, there's the timing issue. Since it is known to the judge — or, at minimum, known to all counsel who represent parties to the matter, who in turn have an ethical obligation to inform the judge — that the Supreme Court intends to rule by a date (relatively) certain on a critical limitation in the definition of the field for the GLP settlement, it would border on an abuse of discretion to continue proceedings that depend upon the Supreme Court ruling a certain way.

Second, there's the substantive question of how this matters to the settlement itself... which, as I've indicated previously, is not nearly clear enough to justify certifying any class, let alone an opt-out class with the inadequate notice being provided. It also brings in the spectre of cross-claims between members of the respective classes when a publisher did not register a work that was scanned; despite the three hundred pages of obfuscation in the settlement documents, there is no consideration of this problem whatsoever, even though it is clearly implied by Morris (the decision that "justifies" the restriction to registered works), and for that matter Tasini itself.

Third, we need to preemptively ignore anything that the Register of Copyrights has to say on the matter. This is a matter of institutional conflict of interest: The Copyright Office's main income source is registration and other service fees it charges, and this financial interest in encouraging a formality that is at minimum in tension with US treaty obligations under the Berne Convention should at least give one pause when considering the Register's position. (That the Register has, since the mid-1950s "studies" supporting changes to copyright law that would eventually become the basis for the 1976 Act, virtually always acted consistent with that financial interest should not surprise anyone.) I have no doubt at all that the Register's position in front of the Supreme Court will/would be that § 411(a) is jurisdictional; Ms Peters and her staff have made that all too clear in past presentations, including amicus briefs to other courts. Ms Peters and her staff may well honestly believe that interpretation correct, but rejecting the views of a captured agency does not require imputing knowledge or bad faith.

Fourth, and last for now, we need to start by interpreting what § 411(a) says without relying on tradition, instead of the other way around. Reading the actual language at issue — and comparing it to § 301 — calls the intellectual basis of the GLP field definition into severe doubt, and therefore the viability of the entire settlement... if only because none of the proposed representatives has interests typical of or in common with authors whose books were not registered (for whatever reason).

  1. This is a common explanation:

    Federal courts treat registration as a rigid jurisdictional prerequisite — even when the defendant's fault is otherwise incontestable. Defendants' conduct in reproducing the Articles, verbatim or nearly verbatim, is both abject and reprehensible. Nevertheless, plaintiff has failed to protect her own rights because she failed to obtain valid copyright registrations in the Articles. Although Morris had the foresight to contractually protect her ability to exploit her work for profit, she neglected to protect her ability to prevent others from doing the same. Because plaintiff has failed to show a genuine issue of material fact for trial, defendants' motion for summary judgment is granted with respect to her first claim for copyright infringement.

    Morris v. Business Concepts, Inc., 54 U.S.P.Q.2d 1561, 1563 (S.D.N.Y. 2000), aff'd, 259 F.3d 65 (2d. Cir.), later op., 283 F.3d 502 (2d Cir. 2001). But cf. Arbaugh v. Y&H Corp., 546 U.S. 500 (2006) (number of employees is a substantive element of the claim, not a jurisdictional prerequisite); Carlsbad Technology, Inc. v. HIF Bio, Inc., 556 U.S. ___ (04 May 2009) (a district court’s order remanding a case to state court after declining to exercise supplemental jurisdiction over state-law claims under 28 U.S.C. § 1367(c) is not a remand for lack of subject-matter jurisdiction).

  2. This is far from the only similar problem in copyright law. For example, consider the problem of the parody/satire/fair use morass created by the Ninth Circuit's continued reliance on Air Pirates as good law... even though Air Pirates itself says that it's based in the 1909 Act, when fair use was a judge-made doctrine and not codified as it is now in § 107 of the 1976 Act, and that the Supreme Court had not yet explicitly held parody (for some meaning of "parody") as potential fair use in 2Live Crew. Cf. C.E. Petit, Cost Allocation and Copyright Orphans at notes 38–42 and accompanying text.
  3. I can almost hear the whingeing now that Congress can't be expected to predict that the word "jurisdiction" is a magic one that must be included in any provision regarding jurisdiction. That argument ignores that Congress knew how to do precisely that in taking jurisdiction over copyright actions away from the state courts when it drafted the 1976 Act. See 17 U.S.C. § 301, 28 U.S.C. § 1338(a). These are not only proof of institutional competence... they were drafted at the same time, and in one instance in the same part of the same bill, as 17 U.S.C. § 411.
  4. Which, in the GLP settlement, leads civil procedure geeks — like me — into the interesting, if hypertechnical, side issue of whether scanning at the University of Michigan libraries was somehow exempt from challenge... if, that is, the University of Michigan is an "arm of the state" of Michigan, a question left unanswered by the desegregation cases because they fall under the Fourteenth Amendment, not the Eleventh Amendment, for determining jurisdiction.
  5. I'd like to think I know a little bit about this, as I've been grappling with it since a few days after I took the bar exam. Channell v. Citicorp National Services, Inc., 89 F.3d 379 (7th Cir.1996) (the one time I've won anything in front of Judge Easterbrook...). The Second Circuit is still sidling up to this question. Itar-Tass Russian News Agency v. Russian Kurier, Inc., 140 F.3d 442, 445–47 (2d Cir.1998) (explicitly distinguishing its jurisprudence from the Seventh Circuit's). Ironically, at least three judges on the Second Circuit can sometimes recognize when a statutory change "has displaced, rather than codified, whatever validity inhered in the earlier view." Jones v. Ford Motor Credit Co., 358 F.3d 205, 208 (2004).