Columbia fair use conference: beyond fair use

Panel 3: Beyond Fair Use

Moderator: Clarisa Long

Jessica Litman: (See her article Lawful Personal Use.) Changes in fair use aren’t necessarily stretching or shrinking it, but just moving it. Copyright owners want expansions in rights, but are unwilling to countenance an expansion of fair use. In the time of Williams & Wilkins, the consensus was generally that personal copying wasn’t infringing, but also wasn’t fair; Sony yanked fair use to cover personal copying, but that led to trouble with commercial parodies, so Campbell yanked it back.

We’re asking fair use to do too much. If we think that unauthorized use of works is always infringing, we end up relying on fair use to get us out of any problems that arise. It’s hard for fair use to do what we ask it without expanding it, but if we want to confine its boundaries, we should redefine its tasks. The problem is not §107 but §106 – an essentially nonstatutory expansion, since Congress hasn’t revisited the key issues in 30 years, but judicial and business understandings of the rights have expanded.

Current fair use doctrine is not optimized for Google Book Search, no matter how you want it to turn out, or for individual personal consumptive uses, though it is optimized for The Wind Done Gone. Fair use dates from an era in which copyright didn’t cover individual personal uses (and in which there was no Google, not incidentally). It’s bad at distinguishing lawful and unlawful personal uses. Making that distinction should be understood as a §106 problem. Congress didn’t seem to think it was adopting laws that reached the ordinary behavior of listeners, readers, and viewers.

We need to reexamine our complicity in the sloppy and unbounded expansion of §106 into an unbounded use right.

The Copyright Act should be understand to grant rights to readers etc., not just copyright owners and their assignees. This is a currently unpopular claim. Copyright owners are afraid that even raising reader rights is an attack on the rights of authors and distributors. But authors and distributors are only some of the people that the copyright law is supposed to benefit. The fact that this claim raises hackles is a symptom of our intense polarization, and it’s keeping us from fixing things that are clearly broken.

Copyright has special protection from First Amendment scrutiny, because the freedom to read, listen and view are essential elements of human freedom to think. These are foundational liberties on which all copyright systems are built. So long as copyright leaves individual readers, listeners, and viewers alone, we don’t notice this. Recently, though, copyright owners have asserted more control over how readers read, listeners listen, and viewers view. The traditional contours of copyright law include its white spaces, copyright liberties that are implicit in the scheme and essential to the system’s design. These liberties shouldn’t be invaded without thought and sometimes there is good reason to expand them.

Copyright law was designed in a time when intermediating was expensive; we needed to direct many of the rewards to intermediaries so they’d buy the paper and build the towers. The rationale for reserving so much of the proceeds to distributors is, if not obsolete, at least optional. We could reallocate that surplus to authors or to readers, listeners, and viewers, or both.

This is an argument to make to Congress, but also a reason to think carefully before acquiescing in a construction of §106 rights that allows them to seep everywhere.

Tim Wu: Tolerated use: slightly different from fair use. One of the signal features of modern copyright law is the presence of what we would call technical infringements of copyright law that are not subject to enforcement action, but are tolerated. Maybe this is a consequence of the expansion of §106 as Litman says; you look at these uses and say “maybe fair, maybe infringing.”

Example: Lostpedia, a fan-created encyclopedia for Lost. The site is notable for sheer volume: 3 million changes. The details accumulated for one character, Kate Austen, involve a huge amount of work – and her entry is longer than the entries of most presidential candidates on Wikipedia. There are pages and pages of theories about Lost as well. And it would be hard to analyze Lostpedia for fair use – pictures seem possibly infringing, but there’s commentary that seems like traditional fair use. There are also transcripts for every show. Almost everyone would agree the transcript is a violation of the reproduction right of the underlying script. Yet there’s been no legal action against this site. Every once in a while, the show actually advertises on this site. Is this an implied license? Probably not, but definitely a tolerated use.

Another example: Guyz Nite’s video celebrating Die Hard. The lawyers for the studio sent a takedown notice, but then the marketing department wanted to pay Guyz Nite to put it back up because it was good promotion; the band says it has no formal license for the footage.

Three ways in which the law is or will be adapting to these mass quantities. (1) The rise of systems of opt-in copyright. To activate statutory rights, copyright owners have to give some type of notice to the putative infringer before that person actually becomes liable in some way. (Comment: What I’ve called informal formalities.) It’s like the exceptions for certain public performances that exist in the statute and allow nonprofit performances so long as the copyright owner doesn’t object. Section 512 is like that, pending the outcome of the Viacom case. Orphan works proposal: flip the duty to object to the rights owner. Google Book Search also has that pattern, where Google wants an opt-out system. (Wu suggests teaching a course called “Google,” which I’ve thought about too.)

His purpose here is not normative, but descriptive: the law seems to be dealing with these developments by establishing opt-in. The traditional answer to mass violations of copyrights, or claimed violations, was compulsory licensing. The alternative emerging solution to transactions costs is opt-in.

(2) In reaction to something like Lostpedia, there may be more pressure on interpretation of the adaptation right and perhaps a narrowing. There are two lines of cases, mostly unreconciled, about what counts as an adaptation. The Seventh Circuit has left more white space, in Litman’s terms – not everything you do with a work makes it a derivative work. A Beanie Baby guide is not a derivative work of a Beanie Baby.

(3) Perhaps companies like ABC want to encourage things like Lostpedia. They might consider clear No Action Policies about what they’re not going to act on. This is different from licensing because such policies could change. If they think it’s good for business – which Wu thinks it probably is – they could make the rules clearer. Since people are doing it anyway, that might not be all that important, but perhaps signalling makes a difference.

Randy Picker: There is no fair use right! It’s important to distinguish use and access rights. We understand the rules for owning a piece of paper. But composing a poem on a piece of paper doesn’t change the ownership regime for the piece of paper. Creating the poem doesn’t create an access right to it, and it’s hard to get to a use right without an access right. (Comment: I’m pretty sure I disagree, at least as to works that have been voluntarily disclosed. Wendy Gordon wrote about this extensively in her Yale piece.) Fair use doesn’t trump otherwise applicable law. The statute merely says that a fair use isn’t an infringement; other laws may apply, like obscenity. Also true with respect to the First Amendment: Cohen v. Cowles makes contract law applicable even when you seem to be in a speech situation.

It used to be true that disclosing a work necessarily involved surrendering control. But it’s now possible to distribute works digitally and still exercise control at a distance. So the question is what copyrighted work “locks” we are going to respect.

The Castaway: famous because of the Bobbs-Merrill case, an attempt to engage in retail price maintenance. The Supreme Court rejected this. The record companies tried to do the same thing: selling records labeled “not licensed for radio broadcast.” RCA v. Whiteman rejects this.

If legends fail, contracts don’t. The ProCD case allows the contract to be enforced. DRM is another option. Adding it on to an existing medium, like CDs, generally failed, but integrating it into a new product, as with DVDs and iTunes, has generally been successful.

Two types of use restrictions. The link between fair use and derivative works is exactly the place to be in our thinking. To think about fair use, we have to think about use generally.

The importance of incremental investment in copyrighted works: publisher won’t do the Mandarin translation until it’s seen whether the English version is a success. You get market feedback on the value of the work before you make more investments. Authors are overoptimistic about the commercial appeal of their works. If we force authors to compete with other translators, then the author will take advantage of initial monopoly access to release all versions at once. But that’s inefficient. Likewise, performance rights and printing rights for a play are separate and the author should be able to set different prices for them. He isn’t interested in creation incentives, but in inefficient over and underinvestment. Fair use is a kind of bundling: whenever you sell a CD, fair use rights are part of the bundle, and if you have to sell in a bundle you will increase the price and make consumers worse off. So unbundling is often a good idea for consumers.

Wu: Picker’s theory of derivative works is completely wrong. It’s not about incremental investment. That is part of the effect of the law. But why is the law involved in trying to discourage overinvestment in anything? It’s like the 1960s where we denied market entry because we thought it would encourage waste. If the adaptation right were fully enforced, there’d be no Lostpedia. But the original rightsholder may routinely be wrong about what the markets are. This is the basic problem with prospect theory.

Picker: This is a standard problem – tragedy of the commons, overfishing; you can’t answer the derivative works question without having a discussion about investment. Isn’t your assumption that if we assign the rights to ABC they’ll either agree or tolerate Lostpedia?

Wu: No, we’re lucky that we have lax enforcement. You believe we should eliminate the evils of competition. But competition is good.

Picker: It’s useful to have property rights to figure out what the entitlements are.

Wu: Yes. But I don’t want to assign the secondary market to the original rightsholder. Fair use and derivative works are both about who gets to control secondary markets. In most areas of the economy we don’t give secondary markets to the original rightsholder. Especially in telecom and patent, we’ve retreated from that.

Q: There is a whole lot of original material in Lostpedia from fans, including ideas and conjectures. Maybe what we’re talking about is the big unstated bargain: the law gives us a starting point, and we work from there, whether by license, laxity, or a wink and a nod. It’s hard for your writers to avoid access to all this stuff on the Web, whereas in the past controlling access to unsolicited works was usually a useful barrier to liability in copyright suits against successful Hollywood productions. But the other side of the bargain is: if you put your stuff out there on the web, don’t complain when your ideas show up in the series in the fourth season.

Wu: The question is whether the law should be so far out of step with what even most owners seem comfortable with. Even though Lostpedia is there, copyright is a sword of Damocles. That’s not an ideal outcome.

Q: ABC may have changed its mind about the site, but the law is fairly clear. The sword is there and should be there.

Wu: People who create these sites don’t have any idea whether what they’re doing is legal. And when Wu gave Lostpedia as an exam question, he got 90 different answers to the question of how much of that was infringing, fair use, impliedly licensed. The law is not clear. It’s also a funny situation to make illegal what people do on a regular basis, especially in the area of secondary creativity.

Q: How does JK Rowling’s reaction to the Harry Potter Lexicon fit into this?

Litman: JK Rowling said keep it noncommercial, online, and sexless and it’s okay; they’re suing over print publication of the Harry Potter Lexicon. This is different from what George Lucas does, which is to say “you can only do this on my website, and then I own it.”

Picker: Wu would say that the people creating the “dictionary of spells” are making a fair use and that the Seinfeld Aptitude Test case was wrongly decided.

Q: Lostpedia bears a Creative Commons license. What’s the impact of such licenses on how users/readers think of fair use?

Picker: It’s a powerful viral strategy for spreading a vision of how works should be used.

Q: What about giving fair users lesser rights?

Litman: Reminds her of how other countries treat “neighboring rights.” That doesn’t speak to the hardest fair use problem, which is whether you’re entitled to make the use in the first place, and lesser rights may not be any consolation to people who are unhappy with the idea that their works are being made use of without consent.

Picker: Anderson v. Stallone involved a Rocky sequel – someone who rights up a treatment without permission was denied copyright in that unpermitted derivative. It’s a twilight zone of the law, with uncertainty for the owner of the Rocky rights – what do they have if they use my treatment in their production?

Picker on “traditional contours”: Cohen is important to his analysis because control at a distance is an access question, and Eldred is a case about traditional uses.