…but not the Court's own bad writing. Two civil procedure decisions masquerading as copyright matters were decided by the Supreme Court this morning. One of them is unexceptional (literally) and relatively quickly explained. The other… not so much, exposing a century of Congressional greed and incompetence, with inadequate consideration of the practical consequences and sources of information on them (especially in a time in which more and more individuals and small businesses are copyright claimants and defendants).
First up, the easy one. In Rimini St., Inc. v. Oracle USA, Inc., No. [20]17–01625 (04 Mar 2019) (PDF), the Court held that despite goofy language in the Copyright Act (obviously inserted by a combination of nonlitigation-aware staff attorneys and lobbyists), "full costs" (17 U.S.C. § 505) means "costs as provided in the general litigation statute at 28 U.S.C. § 1821 and § 1920." This case concerned primarily expert witness fees, which are not explicitly included in § 505 (unlike, say, 42 U.S.C. § 6972(e), also enacted in 1976, showing that at least some Congressional committees knew how to include expert witnesses… but in conjunction with attorneys' fees). In a unanimous opinion, Justice Kavanaugh declared that "expert witnesses" are not "costs" unless explicitly so stated… and because § 505 does not explicitly include expert witness fees (as consistent with prior decisions in other contexts), the word "full" does not suffice. Oracle's purported historical arguments were rejected as unsupported by the actual historical record. What I found most frustrating about Oracle's various briefs in this case, both at certiorari stage and on the merits, is a refusal to directly engage with the opposing argument. There is a good argument to be made that expert witness fees should be recoverable routinely when a matter must be tried to an inexpert jury… but it's at least as much a policy matter for Congress as it is a single-case-that-sets-precedent matter for the Court.
The Court might as well (and perhaps should) have pointed out that § 505 already makes copyright claimants special snowflakes by allowing them — under appropriate circumstances — to recover their reasonable attorneys' fees, not just the "costs" recoverable by any winning party under § 1920. But this just means that the next round of copyright statute revisions will include probably-stealth-mode lobbying to explicitly add expert witness fees to § 505 — a change that ultimately benefits almost exclusively transferees who are well insured, whichever side of the dispute they're on. A freelance author/artist/musician/composer much below the wealth of a Stephen King simply cannot afford expert witness fees as either a winner or a loser. I therefore call for rejecting such a change, unless accompanied by removal of copyright matters from a jury's provenance, preferably to a special master with specific expertise (which would require constitutional gyrations).
Now the harder decision. Justice Ginsburg for a unanimous Court came to a defensible conclusion in Fourth Estate Pub. Benefit Corp. v. Wall-Street.com, No. [20]17–00571 (04 Mar 2019) (PDF) that a litigant may not file a copyright action until the Registrar actually acts on the application. This "must register" requirement (17 U.S.C. § 411) is a holdover from prior law, particularly the 1909 Act, and it has two fundamental problems with its premises before getting to the factual and doctrinal issues.
- It's probably inconsistent with our obligations under the Berne Convention; Article 5(2) disfavors formalities but unfortunately distinguishes between "enjoyment" and "protection" (which are otherwise undefined). Sadly, Fourth Estate does not engage with the Berne Convention outside of a single sentence, slip op. at 9, although a treaty is of higher order than a "mere" statute (if slightly below constitutional provisions). The most-natural reading of the English-language official text (remembering that this Article was originally written in French… by nonlitigating lawyers) is that a failure of a formal registration requirement may restrict remedies, but not the very act of suit. Which leads to:
- It's Congress's cheapskate way of adding to the Library of Congress without actually appropriating money to do so — functionally, it's a tax on original and hence copyrightable expression. Not only must registrants provide two copies of the best (presumably "library-ready") edition of their works, but they have to pay a processing fee on top of it! And I'm not just being cynical; this rationale was explicitly provided and/or acknowledged in both reports to Congress and in Congressional proceedings prior to the 1909 Act, and indeed for a number of years before that. Those who look only at the legislative materials for the 1976 Act would have to delve deeper to find it… but it's there, in the early 1960s reports and hearings. <SARCASM> Apparently, Congress had gotten smart enough by the time there was actual legislative language on the table to keep its conflict of interest less visible. </SARCASM> The irony that the actual appropriated cost of properly acquiring materials for the Library of Congress in 1976 would have been just about the operating cost of an Enterprise-class aircraft carrier is a bit much.
That's bad enough. The real problem, though, arises in the civil-procedure implications that are soft-pedalled in Justice Ginsburg's opinion.
Fourth Estate raises the specter that a copyright owner may lose the ability to enforce her rights if the Copyright Act’s three-year statute of limitations runs out before the Copyright Office acts on her application for registration. Fourth Estate’s fear is overstated, as the average processing time for registration applications is currently seven months, leaving ample time to sue after the Register’s decision, even for infringement that began before submission of an application.
True, the statutory scheme has not worked as Congress likely envisioned. Registration processing times have increased from one or two weeks in 1956 to many months today. Delays in Copyright Office processing of applications, it appears, are attributable, in large measure, to staffing and budgetary shortages that Congress can alleviate, but courts cannot cure. Unfortunate as the current administrative lag may be, that factor does not allow us to revise § 411(a)’s congressionally composed text.
Slip op. at 11–12 (internal citations omitted, typography corrected).
The tl;dr version: Bad and noncomparable data sources lead to poor conclusions. A longer version:
- In a practical sense, that "seven months" figure is not accurate, not even for just group registration of periodicals as it issue here — and even if it were, it's only an "average." Reading the GAO report from which it is drawn more closely (and with more knowledge of the actual context) discloses that this "average" is calculated by accreting noncomparable applications from noncomparable sources; more formally, it is a cross-population mean of uncertain variance. Seven months is probably about average for a single text-printed-on-paper work submitted as part of a frequent-but-not-otherwise-related batch by trained staff at a commercial publisher. It egregiously underestimates the delay for group registrations (such as material appearing in periodicals as at issue in this case), and for photographs and other fine art, and for first-time-or-occasional registrations, and for musical recordings (as distinct from compositions).
- The claimant's proposed remedy — a mandamus action against the Registrar (see slip op. at 8–9) — would not and does not extend the statute of limitations against the infringer absent a later application for "equitable tolling"… and recent Supreme Court decisions (as in "last week" (PDF)!) indicate increasing skepticism of tolling.
- I must disagree with this unanimous decision on the facts relating to "preregistration as a safety measure": Preregistration under § 408, particularly as implemented by the Copyright Office, simply is not available to many types of works likely to be infringed prior to the ability to actually obtain registration, and have all of their value destroyed (both retrospectively and prospectively) by that infringement. In particular, preregistration functionally is not available for periodicals, the equivalent of "hot news" (which may include photographs of breaking events)… and musical compositions, an area of copyright law even more dependent on the "expert testimony" issues raised by the other opinion issued today.
- This last point becomes more important in the internet age when determining who is actually infringing may not be as easy — in fact, almost never is as easy, cf., e.g., 17 U.S.C. § 512 — as it was in the case in front of the Court. "Sure, go ahead and file a Doe complaint." <SARCASM> Yeah, that always (ever, in copyright matters) turns out well. Even when you really aren't Prenda. </SARCASM> And, of course, this also assumes that infringements are detected immediately.
- Just blaming increased processing times on inadequate budget and staff — implying that there's a quick-and-easy solution if Congress has the will, and pretending that these were not that significant an issue until the "increased security" imposed after 11 Sep 2001 — ignores the explosion in copyrightable material eligible for registration since 1956; this blawg entry is just the tip of the tip of the tip of the iceberg, just considering that in 1956 it would have required at best a manual typewriter, and more probably my illegible scrawl. And that's before considering the explosion in sources of copyrightable material since 1956, when copyright attached not on fixation but on publication.
In the end, this is a bad decision that is a predictable consequence of bad Congressional drafting and bad Congressional intent. So-called "registration" should be eliminated unless and until Congress makes it like a car registration or real-property registration — including mandates for current contact information, current ownership, and everything else. Which cannot be done as long as the Copyright Office is a subunit of Congress rather than the Executive Branch. Guys, it wouldn't be that hard — the Patent and Trademark Office has had such systems in place for decades — but it's an executive function… meaning that to keep building the Library of Congress, materials would have to be transferred between branches, which is not as easy as it sounds. And then the registration issues will have to interface with bankruptcy matters, and state-law estate matters, and… naah, just do away with it. Use a carrot (if you want an injunction for seizure or destruction — US law remedies that are in excess of what is required by the Berne Convention — you have to provide the copies to the Library of Congress and prove up with the proof of receipt only) instead of a stick.