IPSC: Fair Use

Matthew Sag, Loyola University Chicago School of Law

The Missing Theoretical Foundation of Transformative Use

Campbell v. Acuff Rose had lots of changes to the work and explicit critical stance v. original—the only reason the Sixth Circuit held not fair use was the prejudices of old white men. But then things get interesting. Cases w/no textual transformation at all—image search, plagiarism detection, HathiTrust. Some people start saying transformative use has gone too far; we need to jettison the whole thing. Doesn’t agree, but thinks that Campbell and Judge Leval’s foundational essay leave open important questions. There’s no theory of what transformative use should mean. There’s a good argument for why we need it but no argument for what its contours are. This is particularly problematic when we need to figure out how much change is enough and what kind of changes are enough—how to distinguish transformative works from derivative works.

What’s the theory? Leval: transformative use furthers copyright’s utilitarian objectives by balancing authorial incentives with freedom to generate new meanings from existing works.  Souter says basically the same.  But this is really a why, not a what.  Utilitarianism is a goal, not a decision rule in individual cases (except for perhaps the most obvious like Campbell). Bespoke cost benefit analysis in every case can’t be done: judges lack the info and the skills. Even if they could, that’s not a good idea b/c litigated cases don’t just resolve conflict b/t the parties; they’re supposed to tell us how to shape our future actions, and if it’s individualized CBA then there’s functionally no precedent because of changes in markets, individuals. Also, why use this lever to optimize © rather than the other possible levers?

Leval wasn’t himself suggesting individual CBA; that’s not the critique.  It’s useful to understand why fair use can’t work this way to see that the utilitarian imperative doesn’t explain much—Leval is interested in broad categorical rules too.

His friendly amendment: fair use should work in harmony with ©’s utilitarian objectives, but also with its means.  But then: is there a deeper structure to © law other than interest group negotiation?  You can’t posit a single theory of © that explains every detail of the Act, but there is a deeper, more fundamental theory, and part of our obligation as scholars is to try to figure that out.

What makes something copyrightable, what determines authorship attribution, what determines whether infringement occurred in substantial similarity cases: all of these show a concern with not the technical act of copying but the communication of original expression to the public. Screenplay cases: courts say that they don’t care about intermediate drafts if the final work wasn’t substantially similar.  In Tasini: court doesn’t care how the database works, only how it appears to the public. Primary purpose of fair use is to allow technical acts of copying in ways that don’t interfere with the author’s interest in communication of original expression to the public.

Nonexpressive uses like big data mining don’t communicate original expression to the public and obviously don’t interfere with the author’s interest.  Well-supported by HathiTrust and Google Books. Expressive uses: the same explanation, essentially: transformative if content/purpose change is sufficient that they pose no real threat of substituting for the author’s own communication of her original expression to the public.

Thus, we have to ask what was the new author’s purpose?  Was the amount reasonable in light of that purpose?  Is expressive substitution likely?  The factors are not independent: factor 4 (factor two is omitted b/c it’s meaningless) depends a lot on factors 1 and 3.  Not a question of complements and substitutes in product markets.  Look for uses that are beyond/outside the intrinsic purpose of the work. Commentary on the original clearly satisfies, subject to amount reasonable in context; debate will be over whether there is commentary. Otherwise new expressive uses are not likely to be transformative, unless radical transformation essentially negates substantial similarity.  He thinks many of these fair use cases should be “no substantial similarity” cases.

Q: Your proposal could get rid of classist stuff that appears in some of the cases about the original author v. the famous appropriator.

A: does think that courts may get star-struck, as in the Green Day case. [The likelihood that Article III judges thought that Green Day were worthy celebrities seems to me to be low, though I make no such statements about their attitude towards the lawyers for the parties.]

RT: This presentation confirms my sense that complaints about Google Books are best at destabilizing support for cases like Cariou, a completely different set of cases—it’s just easier to argue about whethercontent transformation is enough than to argue that the purpose is the same for big data than for individual works. But are you just moving the uncertainty around by stuffing it into “sufficient change to pose no real threat”? Dorling Kindersley/I think you will be forced to determine the original author’s purpose as a matter of law v. matter of fact, like contracts.

A: He thinks he’s reducing uncertainty but also that DK is the kind of case that should always have to go to a jury [which doesn’t sound uncertainty reducing to me].

Q: 2d Cir treated Sony as a transformative use case: was that right?

A: He isn’t saying that transformative use + nonexpressive use are the sum total; other things can also be fair use. Everything the Second Circuit said about transformativeness in TVEyes was garbage, though.

Mehtab Khan, UC Berkeley School of Law

Fair Use and Educational Uses: Commercial Endeavors for Public Interest

Interested in intermediary uses.  Intermediaries have been important in pushing boundaries of fair use. Open access/Creative Commons; controlled digital lending by the Internet Archive; mass digitizaiton under Google Books and HathiTrust.  The threshold set by those cases isn’t achievable by individual users, though: it’s not possible for an individual to create a database and then ask a court to find that’s fair use. Thus, we need to rethink relationship b/t intermediary and end user. Tiered approach: look at what the end user is doing with the copy, not just at what the commercial intermediary is doing.

TVEyes criticism: licensing as a barrier to public interest (note that Fox would delete certain clips that ended up not furthering the Fox narrative, so there was no access to them even through licensing).  Court should have considered users’ (lack of) other options due to Fox licensing practices, not just that TVEyes was a commercially successful endeavor.  Great Minds v. FedEx: there might be some commercial activity involved in an ultimately noncommercial use, so a copying company could make money if the end user was noncommercial (for purposes of a CC noncommercial license).  Implications beyond the US: India Delhi High Court case finding that textbooks are necessary for public education.

Betsy Rosenblatt: fits w/in larger issue of how we treat commerciality generally. Is the initial owner making commercial use?  Is the intermediary?  Is the end user?  If you push it all the way to the end, do you gobble up all of ©? 

A: not black and white, especially w/digital technologies.

Rub: if the end user has money, why shouldn’t the © owner have a claim on it? At least that’s the reasoning.

A: if they make money the should have to show transformative fair use. But sometimes the commercial activity arises along the line.

RT: Agree w/ Rosenblatt--when I watch a movie for pleasure, I’m making a noncommercial use. Probably not enough!  Another place to look is Canada’s UGC exemption—even though it’s called the YouTube exception in public discussions, equivocation about what noncommerciality means when an intermediary is involved makes the provision itself very hard to parse.

Cathay Y. N. Smith, University of Montana Blewett School of Law

Political Fair Use

Peterman v. RNC: photo of candidate appeared in attack ad on candidate: lost on sj on fair use grounds. There haven’t been many fully litigated cases, but once the court determines a secondary use was political, the court’s balancing of the factors changes. Similar to what happens when a court determines that a work was a parody. Makes it easier to find fair use, in political as well as parody cases.  Factor two (Sag’s nonfactor) seems to be key: if the original work at issue was politically related, e.g. a headshot taken for a campaign ad, the court is more likely to find fair use. Affects other factors, e.g., market harm: if you made a photo for one campaign, you’re unlikely to authorize use by the other side. If the original work was unrelated to politics, such as an engagement photo or a song, the political use leans away from fair use.  This seems to trump even new meaning added by the new use. Courts seem interested in moral rights arguments in here, but find market harm.

Questions: how broadly should this project define “political”? How far into the con law abyss should it go?  How should we think about moral rights/dignity rights concerns? Does it deserve treatment alongside false endorsement/Lanham Act type concerns?

Rosenblatt: European students in particular often are surprised that the original author’s objection makes it more likely to be a fair use (both b/c of factor one, reuse in unapproved ways/unapproved changes, and because of factor four—there’s no market b/c author would never allow such a market).

Jennifer Rothman: is this really about moral rights? These authors are objecting not b/c they want to be paid but b/c they object to the underlying politics. That fits into a story told in recent papers by Shyam Balganesh about “censorial” copyright/privacy interests, or Andrew Gilden’s work on nonmarket harms. You could fit this story into those stories and think about nonmarket functions of © law. Something more empirical might ask: is something funky going on with these political © cases in that it makes factor two matter? You could do case counting.   Not sure you can fully extricate Lanham Act/RoP claims b/c they indicate what’s really at stake when brought alongside ©.

Q: lawsuits as performative: showing the world that you don’t endorse this—though that strategy is more effective for musicians than for photographers.

Roberta Kwall: there are other categories that are similar, like religious uses. Religion and politics go together; related dignity based harms.  USCO’s recent report on moral rights gave no indication that it’s concerned about objectionable contextual uses—a gap in our law. [Or, you know, a distinctive First Amendment tradition.]

Q: is an underlying work ever not political?  Artists are political.  [In some ways this is the standard question of transformativeness/criticism—every work has its politics, but the interaction of that politics with the challenged use may not be that obvious or direct.]

Gilden: distinction in what motivated the initial work is really interesting—© misuse may come into this. But should © care about what motivated the initial work?

A: also shows the flexibility of fair use—changes the normal operation as soon as you find a political work.  [I’m not sure this is the right characterization. The presentation itself said that it’s exactly what happens when you find a parody—or, these days, a transformative use in general.]

Christopher Yoo, University of Pennsylvania Law School

The Transformation of Transformative Use: The Infiltration of Functionality into Copyright

Original formulation of Leval/Campbell assumed that new material would be added as well as a different purpose.  The Google Books type cases take away the requirement of adding new content. Reformulate Campbell to make that possible; not a big deal b/c that’s how interpretation works. The 9^th Circuit introduced utility/functionality as a defining feature of fair use, but that’s more a patent idea. Purpose is now overdrawn.  General sense that fair use protects valuable uses. Is that factually true?  Is boiling transformative use down to a single purpose a good idea?

Kelly v. Arriba Soft/Perfect 10 ask whether the use benefits the public. That’s an incomplete question: fair use isn’t required just b/c there’s a public benefit.  Hathitrust rejects value and utility as a basis for transformativeness—not enough to provide more access to the work for, e.g., disability access.  Focus is on whether use has a different purpose, and snippet view has different purpose and provides access to different information about the works and not the works themselves.  Leval says don’t take transformative literally—a translation is a derivative work.  The fact that a collateral tech will make works more accessible is not properly a © consideration.  And we shouldn’t assume that you won’t get the use if it’s not fair use.  In the original Google Books settlement, Google was going to pay—so you could have gotten the Google Books settlement [if class action doctrine were completely different than it is].

Requiring new content would be a better idea.  Authors have incentives to license for other purposes. Fair use requires an explanation for why authors won’t license; fair use when authors have incentives to license is just a wealth transfer, not a legitimate use for fair use.  Private ordering should be preferred. 

Precedent: contributory infringement has obscured things like the influence of Sony, which is about who made the copy [though also and pretty famously it is about fair use].  We have precedents for this idea in trademark—permitting references to names, titles.  You have to be able to refer to works by their titles in order to have meaningful discourse.  That could count as a nonexpressive use.

RT: Your argument is that you could have gotten the Google Books settlement but that’s only true if class action doctrine were completely different than it is, but it’s not different. Also [and not unrelatedly] class action settlement is not private ordering!  It’s extended collective licensing which is just a different kind of wealth transfer, and notably not to authors in many circumstanes.

[Boiling down fair use to one purpose is not how I read the big data cases; the Perfect 10 circuit is the Green Day/Jersey Boys circuit; they don’t seem super committed to a single purpose for fair use.]

TVEyes: there was evidence that Fox discriminated politically in what clips it made available for licensing and in its licensing terms—you couldn’t criticize.  That is characteristic of supposedly blanket licensing schemes.  Is that ok?

A: The revealed preference of Google was willingness to pay.  [But there was no one they could legally pay in the absence of extended collective licensing, which is not private ordering.]

Boiling down fair use: yeah, but that language has been influential.

Blanket licensing: we are seeing one-off licensing, showing the existence of a market.  [A market for what, though? For licensing that supports Fox political positions?]

Sag: the settlement was not on fair use; the settlement also allowed Google to sell books. You can’t use the settlement the way you are.  Also doesn’t deal w/concerns of people w/o Google’s cash, like HathiTrust or other text miners.  (Plagiarism detection.)

A: one actor can monetize something, another can’t—the question is whether Google should transfer some of the value it’s creating to others [something about property rights].

Holdout is a normal part of property rights.  But there has to be a framing about why the system isn’t working; if you gave people full entitlements, they still have high powered incentives to contract.

Buccafusco: the claim is that © doesn’t extend legitimate rights to control nonexpressive uses like big data uses. What is your answer to that?

A: copyright owners desire monetization.  [Well, I too would like a million dollars.]  Perfect 10 is outrageous in folding factor four into factor one analysis.