Thursday, October 20, 2022

Warhol v. Goldsmith, purpose, and character

A small thought on Warhol v. Goldsmith: Justice Jackson seemed interested in whether there was anything to be gained by parsing “purpose” and “character” separately in factor one. Although I don’t think there is given the intentional breadth and flexibility of Congress’s adoption of the common law concept into a still-broad and flexible test, it occurred to me that Professor Tony Reese presciently identified a division in the cases that corresponds to those two words. Professor Reese’s Transformativeness and the Derivative Work Right, 31 Colum. J.L. & Arts 467 (2008), whose analysis I followed in Content, Purpose, or Both?, pointed out that many of the big data/evidentiary use-type fair use cases are well-described by the idea of a transformative purpose—a purpose orthogonal or unrelated to the expressive content of the original work or works used. Cases involving criticism and parody, by contrast, tend to involve changes in content, which could be mapped on to a parodic/critical/etc. “character.” 

The reason I’m not convinced this gets us very far is that one still needs to identify favored “purposes” and “characters,” since that’s not self-evident.  For example, although the justification for appropriation art is different than the justification for parody, appropriation art could easily be said to have a distinctive favored “character” in the context of how it is understood by audiences. And though I classified evidentiary use (Bond v. Blum, Dorling Kindersley, etc.) as a different “purpose,” the way the audience reacts to an evidentiary use as compared to how it reacts to an ordinary consumptive use could easily justify saying that it has a different “character” as well.

However, this approach might help make sense of the otherwise fairly incomprehensible approach advocated by the SG, with respect to some favored “purposes.” The SG wanted to split the baby by allowing the Foundation to continue to display the Warhol lithographs, but not allow it to license them for use in articles about Prince, but if the creation of the lithographs was fair (or authorized) then it’s not clear why further uses would be unfair, and if the uses are unfair then why would there be a valid copyright in the lithographs under §103?

Moreover, evaluating every single use of a newly created work for fairness seems like a really bad idea. That is, one should never have to relitigate whether a parody or review is a fair use even if one, say, licenses it for film adaptation, or distributes it in a different market. (Shades of the old obscenity standard where something that was ok for a gentleman would not be ok for a chambermaid.) Such a standard is likely to substantially suppress publishers’ circulation of fair uses in case “this time is different.” Especially when the accused use is argued to be a derivative work instead of a pure reproduction—as here—this test is very dangerous to fair users who would never be confident that they had their own valid copyrights.

On the other hand, there are purpose-based fair uses, usually involving reproductions and not derivative works, where it does make sense to say that the fairness is inherently tied to the specific use: We believe Google made its big database of books only to show snippets and not to distribute full copies, so its backup copies are fine. 

So one reason the SG’s position seemed so odd (along with its terrible endorsement of some sort of “necessity” standard) was that it probably shouldn’t make a difference in this case, where there is a new work that either has a valid copyright, or, under §103, doesn’t. 

Relatedly: it would be deeply ironic if the result of this round of fair use cases was that big data uses are definitely fair use, so Google wins its cases, while each individual artist has to establish the importance of their specific fair use, even with very explicit parody or criticism.

No comments: