Copyright Theory II
Brad Greenberg Columbia Law School
Does having lots of different antennas in one place v. all over the place make a difference to whether there is public performance? Central question of Aereo: looks like cable, must be cable. Aereo follows a long line of cases about new tech asking whether judges should look inside the machine or just look at the inputs and outputs. The case is thus fairly ordinary. Why it matters: constanct conflicts over behavioral perspective (majority) or technological/structural perspective (dissent). Who should decide how © touches new tech: the answer is determined by behavioral v. structural perspective. Statutory signals theory applied in Aereo and Google Book Search.
Napster and Grokster: different tech, same result. Goes back at least to White-Smith Music; also occurred in video game cases. Majority position in White-Smith is structural: absent machine, piano roll is unintelligible. Dissent takes behavioral perspective: no reason to decouple roll from machine; makes a sound that sounds like the song.
Structural perspective: narrower rights and narrower limits. Behavioral: broader rights and broader limitations, like fair use.
Who should decide? Congress has increasingly punted on tailoring © to new tech. Presumptively, courts and CO articulate limitations. Two key features of 1976 signal congressional deference: tech neutrality and standards over rules. Tech neutral: §101, 102, 106 exclusive rights. Tech specific: §108, 111, 115-119, 122. The choice of provisions is important signal for how © should be applied by future courts—behavioral or structural.
Google Books: outcome would be different if it were a full digital library, not just snippets, and so it should be. Fair use is tech neutral and flexible. Behavioral is the right perspective because §106 rights are tech-neutral. Even if the cable analogy is wrong, it’s likely that it doesn’t matter b/c the court had to find some way to shoehorn its decision into §106.
Counterargument: Congress has tinkered plenty w/satellite tech, compulsory license, etc. That’s true, and when you ask about §119 etc. you probably do need a structural inquiry/ask about Congress’ particular mindset, but for now everything else falls under §106 by default.
Christina Mulligan: If Aereo is rightly decided, is the RAM copy doctrine wrong?
A: It should be wrong, but it’s probably not b/c Congress did tinker w/§117 in response. First sale is a similar issue.
Jake Linford: One way to read Aereo is just Breyer as Cong. intent and Scalia as text—isn’t this the same dumb point that isn’t really about ©?
A: would make the same argument regardless. Second Circuit was much more about behavior v. structure.
Q: Ambiguity in question of who should decide. What has Congress done? Who ought to be deciding various parts of the © system?
A: taking an imperfect world—judges regularly look to statute for answers. Judges often choose a perspective w/o explaining why or even saying that they are. Maybe Congress isn’t the best source of innovation policy, but it’s what we’ve got.
Peter DiCola: White-Smith isn’t as clean as that either. The president had weighed in publicly; there were already negotiations in Congress. I’ve interpreted that case as being about the Court pushing the parties to bargain; composers weren’t coming to the table and were being punished for bad behavior in licensing only one company. 1976: §106(4) and 106(6) are tech-specific and Congress did decide to do it that way. They’re broad in rights and narrow in most exceptions—which signal is Congress giving? It will always be mixed.
A: see those as tethered to tech mindset of Congress—intent to be tech neutral to the extent that something can be publicly performed (?) regardless of whether tech was known in 1976.
Sag: number of aerials didn’t matter in Aereo. What mattered was interposition of lots of copies (even though that wasn’t before the Court).
A: Aereo is a tough example b/c so many people don’t like it, but that still shows his point.
RT: Behavior: is that right? B/c I think you mean audience reception. Connects Aereo to Tasini, a connection made by the Court itself. Connected to what you mean by tech neutrality. Both parties claimed the mantle of tech neutrality. Neutrality means different things to different people, as in equal protection.
Kevin Hickey, University of California, Berkeley, School of Law, Center for Law & Technology
Paternalism in Copyright
Aims: explore tension between incentive model of © and its paternalistic provisions; envision legal structure of © if we took paternalistic impulse seriously—a thought experiment, not a normative endorsement.
What do we mean by paternalism? State limits choices of individuals to protect them from consequences of their own decisions. Nothing inherently wrong with intervention.
Soft paternalism tends to fail for the very reasons that motivated the regulation in the first instance. (RT: And the existence of entities motivated to make it fail, pace Lauren Willis!)
Potential behavioral market failures: creation stage (intrinsic motivation; judgment under uncertainty—© as distant, uncertain, indirect incentive, but contrary to that are lottery effects). Also assignment stage: bounded willpower/short-sightedness; social preferences for “fair” allocations v. the endowment effect.
Paternalism in ©: termination rights; elimination of formalities; limitations on alienability (writing requirement); moral rights. Due to lack of bargaining power, short-sightedness, poverty, uncertain valuation.
Tension between incentive model with its rational actors and paternalistic provisions; rational actors respond to 20 additional years of protection after life +50, but paternalistic provision supposes an author incapable of protecting her own interest, making bad deals—if those bad deals are made, then the prospect of future rewards wasn’t driving creation in the first place.
Normative and policy implications: current hybrid model risks costs of moral rights like system without the benefits. Social costs of expansive, default © through eliminating formalities. Fails in curing perceived unfairness and actually securing a piece of the profits for the author. Existing author protections are weak and ineffective for the vast majority of authors (whose rights aren’t that valuable 35 years later). Writing requirement might help a bit, but it just means they sign away rights for a pittance instead of orally granting them for a pittance.
Taken seriously, we’d either eliminate paternalistic provisions and reintroduce formalities, or do more in the way of author protection such as EU-style voidable terms, mandated fair compensation; incentives would focus on present-time, direct, certain outcomes.
Betsy Rosenblatt: how much is paternalism v. market uncertainty/accounting for an unknown future? Formalities are value-neutral; others are harder to chalk up to that.
A: often given as justification for termination, but publishers are also subject to uncertainty; transferring risk to publishers may make sense given publishers’ portfolio. Regardless whether it’s short sighted or difficulty in valuing, it’s paternalistic in operation.
RT: What if it doesn’t matter what the law is? Abraham Drassinower & Jessica Silbey: if all you’re about is incentives, then there’s really just one entity, TM/©/patent/IP. Also Laura J. Murray, S. Tina Piper, & Kirsty Robertson, Putting Intellectual Property in Its Place: Rights Discourses, Creative Labor, and the Everyday: Rather than listening to experts, “people actually choose to understand the law through information and opinion gathered from friends, strangers, coworkers, and the media.” If you gave up on people understanding the law, at least at the point of creation, then you could essentially ignore the creation stage in terms of making authors better off. You’d be led to focus on the assignment or commercialization stage.
A: thinks the law matters. (But I think you can’t have that as the assumption for this project.) Would also be led to think about access to lawyers as well. (Again, that’d be after creation, which has implications for which provisions would work to help authors.)
Sag: Add in compulsory remuneration. How is this different to distributional concerns in ©?
Q: difference b/t irrational and uninformed. (Hmm, not sure that’s true. Chicago school people would disagree b/c the choice to become informed is itself a cost that people choose to bear or not bear.) In formalities, formalities limit my freedom. Creators wouldn’t understand that they needed to give notice. There’s a nonpaternalistic explanation in terms of market structure. (? But notice requirements are useful in forcing information so that markets can more easily form.)
A: There is evidence that termination rights exist to protect the impecunious/irresponsible author. These are fairly weak default rules; if we are serious we’d want something stronger. Push back on formalities elimination as nonpaternalistic. Depends on whether your baseline is © as natural right or statutory right. Elimination does depend on conception of authors as careless.
Q: Keep an eye on the context of the time: eliminating formalities was also about joining Berne. (Which pushes the issue back to a different entity: why were formalities eliminated from Berne? Answer may well be that it wasn’t paternalistic but based on authors’ rights view, a kind of romanticism—but consider that elevating authors as more important than other kinds of people and thus giving them special rights is just the flip side of paternalism, as the sexism inherent in the term indicates—putting on a pedestal = putting in a cage in many ways.)
VARA allows for waiver, but not transfer. That may hurt authors.
Margot Kaminski & Guy Rub, Ohio State University College of Law
Zoom-in Zoom-out of Copyright
Outcome of cases/scope of copyright protection is determined by court’s point of reference. If you evaluate a small unit (God’s finger touching Adam’s in the Sistine Chapel), you give strong copyright protection/shrink the public domain. You get lots of different IP rights/transaction costs. If you zoom way out (the whole ceiling) you frustrate other interests by weakening ©, and maybe unfairly punish people who create complex works.
Natural tendency of © in US is that © owners argue for zooming in as much as possible for statutory damages/fair use, but only up to the point of invalidity. Statutory framework: the word “work” is not defined. After formalities were limited, authors can do whatever they want at the point of claiming infringement, and registration is a huge mess for these purposes.
Our work is about zooming as a framing decision that judges or regulators or legislators make. It’s not about defining a work. Judicial framing move takes place across doctrines. Initial findings: courts often set level of zooming without noticing the issue; if noticed, do it without principled reasoning; if do reason, adopt different tests across and even within doctrines.
Example: statutory damages. Factors: independent economic value; can each TV episode be consumed separately; was each episode produced separately; registration as an additional factor. W/two albums, 2d Cir. rejected independent economic value test because (1) statutory text says one compilation is one statutory damages award; (2) © owner decided to bundle the songs together. Under Arista Records, SDNY 2010, court says that if, before infringement, the copyright owner offers an unbundled product, zooming in is appropriate. (Of course even before iTunes there was sample licensing for individual songs, so there was “unbundling” of a sort even w/in songs since the 1990s at least.)
Registration to the rescue? Author expressing intent through registration system; cost of registration at least helps. But it doesn’t work because it has its own zoom problem. 9th Cir. found that registration of database of stock photos registers every photo; SDNY has found to the contrary. Having the DB registration reach individual works w/o having to list indiv. works eliminates the utility of registration and don’t have transaction costs identifying each work as a work.
New Copyright Office compendium: (1) registration policy sometimes conflicts w/statutory damages approach—allows author to register one CD as unit of registration, which reaches each individual song. That’s the opposite of the result in the courts. (2) registration policy is calibrated against statutory damages! CO sets policy as a result of what CO thinks result would be in statutory damages case. CO assumes that unpublished works should be registered as individual works, not compilation, b/c it wants to preserve statutory damages.
There are multiple inconsistent tests where they exist. Four factor fair use test also has zooming problems: factor three amount (factor four also has a zooming problem). Photocopying cases show this: Williams & Wilkins looked at all the magazines; Texaco looked at each article and was the only one that engaged in reasoning; Princeton Univ. Press looked at each article; Cambridge Univ. Press v. Patton zoomed out & thought the issue was raised too late.
Zooming and substantial similarity: total concept and feel of rugs in 2d Cir.; Calabresi wants zoom in to make sure that public domain works are discounted.
Zooming and authorship: Garcia v. Google. Majority says: this is one work. If so, she’s not a joint author. Kozinski fights the zooming and looks at 5 seconds as relevant reference point. Maybe she can be the author of that. (I think it’s worse than that—it’s not the 5 seconds but the performance in that 5 seconds.)
Also finds zooming in other areas: Notice; revision of collective work; separability with PGS works; etc.
Questions: other areas of IP law where this happens? Scope of project—descriptive v. presecriptive? Judges versus regulators v. legislatures?
Q: some of these questions are more about sets than zooming in and out.
Andrew Gilden: is this an evidentiary q—how to courts prove their conclusions? But if it’s more a conceptual issue, then looking at briefs in the cases might help.
Ramsey: songs in an album are easier to count than characters in a book—there might be a bunch of ways to count in a book. [I’d also note, as above, that chunks of a song can be counted in particular contexts, such as ringtones and sampling.]
Sag: some approaches have no limiting principle—turtles all the way down. Those approaches are probably wrong.
A: we can probably agree that Garcia’s position won’t work.
Q: market effect in fair use, definition of nascent markets—it also occurs. Choosing which it is pervades © generally.
Q: if you take the 2d Cir. approach, I have serious issues with newspapers/photo collections. You just couldn’t register at an acceptable cost. But newspapers will license individual articles.
Niva Elkin-Koren, Haifa Center for Law & Technology, & Orit Fischman-Afori, The Haim Striks School of Law, College of Management Academic Studies
Rulifying Fair Use
Presented by Fischman-Afori. The trigger was the Cambridge Univ. Press case, a rule against fair use rulification. Fair use is a standard; can courts develop the standard into more concrete guidelines?
Their position: rules/standards is a spectrum, not a dichotomy. Fair use was not meant to foreclose evolution into more concrete guidelines. Rulification can serve copyright goals. Courts should do more rulifying and the 11th Cir. is wrong. Ancillary rules to assist the court—does fair use prohibit any such ancillary rules or is it a mandatory rule against rulification? Designed as a permissive standard. Lower courts have allowed some rulification, with Cambridge Univ. Press being the big (and bad) exception.
Advantages: avoid uncertainty creating a chilling effect; allows both flexibility and certainty. You can have your cake and eat it too because rules/standards are a spectrum. Transparency: avoid manipulation of the fair use four-factor analysis. Rulification may force judges to fully disclose underlying analysis.
Q: Appellate court in Cambridge Univ. Press was concerned w/ 10% being pulled of thin air; we have more rulification than you suggest. Timeshifting = fair use is also a rule. Transformativeness is also quickly becoming a rule of its own—Neil Netanel’s work.
Q: SCt doesn’t say every parody is fair use. DCt went much further than any previous court. Safe harbor = not the same as a rule.
A: our paper does address difference b/t common law style rulification and safe harbors. We should check what the rules are. There are 75% safe harbors, etc.
Sag: doesn’t think categories are the same as rules—reverse engineering, genuine parodies, digitizing library to make a search index—but maybe we’re just debating terminology. Clear categories do provide guidance. 10% rule also, but that’s subject to gaming, so there’s a huge difference. If we could rulify fair use we wouldn’t need fair use.