What is a derivative work is somewhat mysterious, especially given the expansion of the reproduction right.
Courts and litigants have debated whether a photograph of a sculpture is a derivative work of the sculpture or merely a reproduction, with its copyrightable creativity present somewhere other than its representation of the sculpture. The creativity in a photograph of a sculpture, in the dominant understanding, comes from the photographer’s choices with respect to lens, angle, timing, and so on, rather than from transforming, recasting, or adapting the sculpture itself. So the new work has a standalone copyright (and can be a reproduction of the three-dimensional work)—a new work fundamentally dependent on the existence of the prior work, but not a derivative of that prior work. But applied to termination rights with respect to preexisting works, such as songs incorporated into a motion picture, this same reasoning gives a counterintuitive result.
When it comprehensively rewrote the Copyright Act and created the current definition of derivative works, Congress tried to solve an economic problem—what happens to works that incorporate other works as a necessary component—with a conceptual definition. That didn’t work.
Cases about photos of sculpture seemed to make sense when I read them—the copyrightable elements of the photo generally come from the choices made by the photographer, not from the subject matter, and that suggests that the photo isn’t a derivative work of the subject matter. If the subject matter were an uncopyrightable object, we’d still recognize the photo itself as copyrightable, so it seems that the photo’s copyrightability must be independent of the subject matter.
But I think on further reflection that’s wrong. A photo is a translation of a sculpture into another medium, and translations are generally recognized as derivative works. Indeed, I think the comparison to a photo of an uncopyrightable object actually points in that direction. All representational creativity takes uncopyrightable subject matter as its starting point. It’s the resulting creative work that is protected, and the fact that the subject matter was uncopyrightable—whether fact, public domain, or something else—affects the scope of the copyright as applied to other works that might allegedly infringe the representational work. But this is also true to just the same extent when the subject matter is copyrightable. Creative elements already present in the underlying work can’t be part of the protectable increment of the derivative work. This is the same as saying that uncopyrightable elements like facts or public domain elements can’t be part of the protectable increment of the freestanding work. That is, both derivative works and nonderivative works contain a golden nugget of separately copyrightable material, along with the noncopyrightable elements. It’s for that reason that the separability of photo and subject matter doesn’t inherently take the photo out of the derivative works category.
The cases reaching the conclusion that photographs aren’t derivative works were distracted by an analytically separate and easily answerable question, which was whether a photographer who had permission to create a photo of a copyrightable 3-D object needed separate permission to claim and enforce the copyright in the photo. The answer is no, obviously so, but the courts then went on to say that anyway the photo wasn’t a derivative work so hypothetical rules about exploitation of derivative works didn’t apply.
And that holding is more troublesome. Consider the insight, largely incorporated into copyright doctrine only via fair use, that change of context changes meaning? The photo is, by general agreement, a new copyrightable work, and it incorporates the original sculpture in a way that copyright recognizes: isn’t incorporation of the copyrightable elements of the original plus addition of new copyrightable expression another standard way of defining a derivative work, as in the classic Batlin case?
There is a conceptual claim made in the photography cases: the creativity of the photo does not interact with the creativity of the sculpture. But what can that really mean? There is a physical change in the new work: the sculpture has been flattened into two dimensions. So it can’t be that there was no physical transformation (though I suspect that our tendency to conflate the image with that which it represents helps courts ignore the physical transformation). There is also a change in meaning. Entire books have been written about photographs of sculpture and how they change what the viewer experiences: the viewer does not see the sculpture, she sees a photo of the sculpture, and that matters to what she sees, including what she sees of the creativity of the sculpture.
Consider this photo of a mosaic:
(sculpture by Jack Mackie, photo by Mike Hipple)
Why doesn’t that change the meaning of the mosaic through the addition of the human following the steps and the framing choices that create a particular experience of the mosaic?
Likewise, putting a song into a movie can often have profound effects on the song and how it’s experienced. The final paper will go into the literature on the relationship between sight and sound in TV and movies; the presence of music affects both what people see and what they understand the music to mean, which sounds very much like what we’ve come to understand in law as transformativeness. Juxtaposition regularly changes meaning: Kuleshov effect/commutation test: “trying out a sound change and observing whether a meaning change is produced or not.” If you’ve ever seen fake movie trailers, such as The Shining as lighthearted comedy or Diff’rent Strokes as a story about a pedophile who adopts young children, you’ll notice that the music has done almost all the work of effecting the genre change.
From that perspective, it often would be fair to say that the movie transforms the recording that appears in it, into a bearer of meaning of the narrative. When Jon Cryer lipsynchs to Otis Redding in Pretty in Pink, for example—or choose your own favorite example—the song becomes part of the larger work, and not one that could be extricated without making profound changes to the movie.
So what? Most of the time the reproduction and derivative works rights overlap and so the distinction doesn’t make a difference, but it is extremely common for films and particularly television shows to license the inclusion of existing masters into their works, as fairly significant elements of the narrative. For that matter, it’s relatively common to license things like the appearance of within-copyright sculptures in the background of scenes, after the BET case. Section 203 was designed to protect film and TV producers, among others, by making terminations inapplicable to prevent the continued exploitation of derivative works: “A derivative work prepared under authority of the grant before its termination may continue to be utilized under the terms of the grant after its termination.”
Now return to those photos of sculptures: if they’re not derivative works, then the right to keep disseminating the photo can be terminated. But the whole point of giving copyright in photos based on the creative choices made by the photographer was to recognize the creativity of the photographic enterprise, which is precisely the reason that termination was thought to be problematic when a derivative work was involved. The theory was that the original copyright owner shouldn’t be able to completely suppress the extra creative elements added by the creator of the second work. If “derivative work” is a subset of “creative works with separate copyrights based on/requiring license from a prior copyright owner,” then that scheme fails.
This is even more true as applied to movies incorporating previously recorded music; true, very few movies will have significant value 35 years after release, but those that do will face substantial termination risk if they used existing masters unless we have a very expansive definition of derivative work. Depending on lawyers’ cleverness, there could be other time bombs hidden in movies. For example, there are at least lawyers out there willing to argue that actors’ performances are separately copyrightable independent of the audiovisual works in which they appear. If the overall movie or show isn’t a derivative work of such a performance, it would follow that any transfer that wasn’t a work for hire agreement would be terminable.
Possible theory to rationalize the result that the owner of the copyright in the master of the song used at a crucial moment in the movie could terminate a transfer even though the owner of the copyright in the screenplay couldn’t: assumption of investment: movie is unusable/unexploitable without the underlying book; could regularly edit out the song/sculpture in the background and continue to extract economic value from the edited work? I don’t find that very persuasive, though.
One piece of evidence from the 1976 House Report suggests that, though Congress didn’t think this through, it assumed that movies would count as derivative works with respect to songs without addressing the question of “derivative of what?”:
The definition of “joint works” has prompted some concern lest it be construed as converting the authors of previously written works, such as plays, novels, and music, into coauthors of a motion picture in which their work is incorporated. It is true that a motion picture would normally be a joint rather than a collective work with respect to those authors who actually work on the film, although their usual status as employees for hire would keep the question of co-ownership from coming up. On the other hand, although a novelist, playwright, or songwriter may write a work with the hope or expectation that it will be used in a motion picture, this is clearly a case of separate or independent authorship rather than one where the basic intention behind the writing of the work was for motion picture use. In this case, the motion picture is a derivative work within the definition of that term, and section 103 makes plain that copyright in a derivative work is independent of, and does not enlarge the scope of rights in, any preexisting material incorporated in it.
Idea that I’m playing with: Addition of copyrightable material plus recontextualization equals derivative work. Not a standard of whether the use opened a new market, as the 9th Circuit held in the terrible Mirage case, but would find a derivative work where the new copyrightable material and the existing material affect each other’s meaning. Recontextualization alone might be a fair use, but it wouldn’t count as creating a derivative work. One way to think about that would be to say that the pure recontextualizer just adds an idea—wouldn’t this be productive in a search engine, or a database of potentially plagiarized papers, or in a new frame on the wall—but I’m still trying to think about what that means. Mike Madison will probably say that what’s going on here is an implicit theory of things and which things are unitary (the painting and its frame and the gallery in which the frame hangs are all separate things, despite the fact that the meanings of each influence the meanings of the others). So, a collage that reflects selection, coordination, and arrangement by the artist has sufficient creativity to be copyrightable, and is its own fixed thing, meaning that at least in my model it’s a derivative work.
Mike Madison: are you using perspective of the creators or perspective of the audience? Perspective of some objective third party?
A: I’m most interested in the audience, but creators are clearly thinking of the ways use of existing works affects their own output—lots of literature on music selection for movies.
Glynn Lunney: are you just trying to solve termination?
Lemley: bizarre to treat derivative works separately; also questions of fixation where there’s no reproduction: Midway versus Galoob—my approach could make many more things into derivative works, and that might be a reason to reject it. (I agree that’d be a reason to reject it; I’m not sure Galoob would be affected by my theory, since it essentially imposes a fixation requirement, which I don’t think I have to disagree with. Tony Reese has done the most persuasive work on this, convincing me that there should never be a violation of the derivative works right without a violation of either the reproduction right or the performance right.)
Jeremy de Beer and Mira Burri, Transatlantic Copyright Comparisons: Hyperlinking and Making Available as Communication (presented by de Beer)
Treaties contain ambiguity about how signatories shall provide protection about making available rights, in order to get agreement by member states, mainly because US and European approaches dramatically differ. US: dealt with in context of distribution right; Europe: communication right (closer to US public performance). Canadian: hybrid approach: depends on purpose for which you’re making work available. If for streaming, under brand new Canadian law, it will likely implicate the communication right. But if for downloading, other rights such as reproduction/authorization/distribution come into play. It can’t impact all of those rights because of a decision, ESA v. SOCAN, saying you can’t stack royalties for different rights in the particular context of transmitting music as part of video games online. One trial case: Publication for defamation case: don’t break the internet—linking to a defamatory publication can’t be itself defamation. Safe to say that defamation cases and SCt’s approach in copyright, hyperlinking to copyright infringing content is unlikely to constitute a communication itself.
Issue also alive in EU; pending before ECJ. Svensen: a fairly typical news aggregator case. Does there have to be an actual transmission/download? European approach in general, and copyright scholars, think that transmission is necessary, and that the work has to be made available to a new public—a public without previous access—for there to be a violation of the making available right. 3 national courts: German news aggregator; Norwegian case against Napster; Dutch case about pornography (unpublished Playboy images). German news aggregator was held liable, but Norwegian court held that linking alone wasn’t enough for liability. Dutch case: unpublished nature made the difference.
Paper: comparative analysis is useful. Hopes the ECJ will look at Canadian courts streamlining royalty payments and limiting defamation liability. This could create a validation for the Canadian SCt.
Note that no one is saying that linking could never lead to liability, but that it’s not a transmission.
Matthew Sag, The Perils of Compulsory Licenses in Copyright Law
Another timing failure! I caught enough to know that part of his argument was that compulsory licensing for digitization is offered as a solution to orphan works problems, and really only makes sense in that context, but such licenses have structural problems making statutory damages/remedies reform more appealing if what we really care about is orphan works. I look forward to the full paper.
Jack Lerner, Regulation and Oversight of Collective Management of Copyright
Well known problems of corruption, self-dealing, etc. Transparency, organizational governance, and institutional design literature offer insights. Also has looked comparatively at collecting societies around the world. There’s little regulation in the US beyond antitrust law, Copyright Royalty Board. Other regimes require permission to operate; membership may be regulated, or leadership. Transparency requirements that aren’t present here. Different people can have standing to enforce rights, and different forms of oversight.
What are our goals? Are we trying to empower small creators? In developing world, is there a way to empower local creators, since CMOs mostly just take revenues out of those countries? Is it just about decreasing transaction costs?
What are our tools? Need to assess whether CMO is doing what it’s supposed to, and that’s difficult to measure.
Various scandals of abuse, including in Brazil. The window for reform is now here: the question is how to generate legitimacy. Discloser and recipient must both benefit from transparency—need carrots and/or sticks. Maybe users need an organization separate from CMOs.
Oren Bracha & Talha Syed, Beyond the Incentive-Access Paradigm? Copyright & Product Differentiation Revisited
Product differentiation is monopolistic competition: many offerings in the market are variations of each other (e.g. action films), mapping onto consumer tastes heterogenously. Christopher Yoo: Can better explain competitive/monopolistic dynamics in IP industries. Seems to allow us to finesse incentives/access: strong copyright but also narrow copyright. Michael Abramowicz: foregrounds an aspect of monopolistic competition, rent dissipation. More than one entrant incurs costs of R&D in creating work that is nonrival and can satisfy consumer demand; that’s waste. He suggests weak and broader copyright—reduced protection may reduce incentives but that’s ok to deter wasteful competition; strong derivative works rights allows control of subsequent development and reduces wasteful activity.
Bracha & Syed: you can’t get rid of the deadweight loss by appealing to monopolistic competition. Property v. monopoly rhetoric is inconsequential. You can’t provide the incentive of copyright or patent without the deadweight loss unless there is costless perfect rights discrimination, which is a fantasy. The point is to allow a markup over a work that should be provided at marginal cost; every markup over marginal cost incurs deadweight loss unless there’s perfect price discrimination, which there’s not. It’s a red herring to talk about monopolistic competition. Product differentiation suggests that the drawback doesn’t have to be access.
As you expand protection, you capture more supramarginal works, but you lose inframarginal works. Product differentiation shows that incentives/access isn’t the only tradeoff. It is possible that if you increase protection, you induce more competition in inframarginal markets over existing innovation, and therefore reduce or fail to increase price competition. The cost is increasing duplication/rent dissipation. Each increase in level of protection provides or enables generation of supramarginal innovation or increased variety in inframarginal markets. The drawbacks may be increased inframarginal barriers to access, or increased inframarginal price competition, which comes at the cost of rent dissipation. Need a general equilibrium analysis, as Glynn Lunney has written. May distort investment away from non-IP activities. Tradeoffs involve a series of factors, but there’s no free lunch.
Yoo: We never claimed we weren’t in a second-best world. The rent dissipation problem has always been clearest to Yoo with patent races because of patent’s first past the post. Even if you come in second in copyright, you can still split the rents; this isn’t perfect, but it’s a healthier equilibrium—people can compete to divide the market.
A: there’s a race to be first and inventing around. Both involve social waste—there’s more variety than is socially optimal because of cannibalizing rents. That happens at either stage of the race. The second comer in patents can be the big winner, too.
Yoo: but you can have insufficient entry, with nothing to dissipate. The differentiation literature recognizes dissipation as an issue.