Professor Jennifer Rothman has written a long and thoughtful article whose central thesis is that we should be cautious about using customary practices to decide intellectual property cases, especially in the copyright area.1 At the theoretical level, her skepticism about custom is at odds with the defense of custom that I have offered in previous writings and still defend in a wide range of tort and contractual contexts.2 I am grateful that the editors of the Virginia Law Review have invited this brief response, which accepts some of Professor Rothman’s main points but dissents on others. It is convenient to divide this commentary into dubious and useful customs.DUBIOUS CUSTOMS
Rothman’s basic critique is that the overextension of custom in copyright cases has unduly narrowed the scope of the fair use doctrine. As she notes, judges and lawyers have struggled gallantly to make sense of the maddeningly vague statutory definition of the defense.3 In my view, she stands on very strong ground in insisting that a narrow reading of the fair use defense should not be established through the aggressive behavior of individual copyright holders who frighten off potential users of their copyrighted material. The mechanism that she outlines is as deadly as it is effective. The copyright holder—she refers to Steven Joyce—adopts an aggressive litigation strategy that leads potential users either to back off their original project or to pay the demanded license fee.4 The former cases are lost to history, but the latter are then (mis)interpreted by courts as evidence of the narrow scope of the fair use privilege. She is right to reject any alleged custom that is formed in the shadow of improper legal threats, as courts are better advised to make some independent judgment as to whether the copying falls within the scope of the fair use defense—assuming someone is brave enough to risk the very heavy statutory damages available in copyright cases. Rothman presents no systematic evidence about the frequency of these practices, but her cautionary remarks to judges strike me as fully justified.
Her point must be placed in context, however, for it is critical to remember that defenders of custom such as myself fully embrace this conclusion. Any custom worthy of the name has to result from repeated voluntary interactions among parties from the relevant groups; otherwise, it offers no evidence that a purported custom maximizes the joint welfare of the parties whom it governs. Without doubt, no custom should bind strangers to its formation who lose systematically from its application. Yet that is exactly what happens when the threat of litigation shapes the behavior of the submissive parties. We do not have to worry about what kinds of cooperative interactions generate customs. It is quite enough to rule out the purported legitimacy of any threat-induced custom.USEFUL CUSTOMS
The differences between myself and Rothman arise because she extends this critique to customs that I think work well in their broader institutional context. In this brief comment I shall address three such cases. The first deals with the brief use of copyrighted material in set dressings. The second concerns the major settlement between the publishers and universities over the scope of the fair use privilege, and the last concerns the ability of individual scientists and researchers to make personal copies of copyrighted materials.
A central theme in Rothman’s work is to decry the use of a “clearance culture” that has arisen in the movie industry.5 I understand Rothman’s frustration, for example, with Judge Newman’s failure in the U.S. Court of Appeals for the Second Circuit to accept as a matter of law the fair use defense in Ringgold v. Black Entertainment Television.6 That case involved a defendant who used a poster of plaintiff’s “Church Picnic Story Quilt” as a background prop for a combined period of less than thirty seconds, albeit it in the center of the screen, but to which no dialogue or plot line was directed. This work was a distinctive form of art, pioneered by Faith Ringgold, which involved the combined use of “a painting, a handwritten text, and a quilting fabric,” and conveyed messages about African-American life in the early part of the twentieth century.7 The defendant’s show was a sitcom called ROC about the experiences of a middle-class black family living in Baltimore.
Looking at the overall context, however, it is far from clear that any industry custom in favor of licensing is necessarily misguided. Nor does Judge Newman seem incorrect when he found that the defendant’s limited use should not be treated as a fair use as a matter of law, which was the only point he decided. Initially, there is no evidence of the kind of bullying behavior that Rothman found in other cases. If, moreover, the fair use defense is denied, what is the big deal? By assumption, there is no real connection between the displayed poster and any thematic element of the sitcom episode. If we read the facts that way, there is no particular reason to use it at all, for any other poster should do as well. Accordingly, if the custom is respected, one of three things ought to happen. First, the license fee in this case will be trivial because of the credible threats to use other decorative art in the background. Second, ironically, an astute filmmaker may well ask for product placement fees, which the owners of some copyrighted works might well be prepared to pay. Or third, the defendant (by assumption) loses little by putting up some other art work that lies in the public domain. In fact, in this case, Ringgold had turned down offers for use of the type that the defendant made. All in all, I see little to criticize in Judge Newman’s nuanced account of the fair use doctrine. In particular I think that he is clearly right on the fourth fair use factor when he writes:
Ringgold is not required to show a decline in the number of licensing requests for the “Church Picnic” poster since the ROC episode was aired. The fourth factor will favor her if she can show a “traditional, reasonable, or likely to be developed” market for licensing her work as set decoration.8
In addition, Rothman has missed some of the advantages of following a customary practice to license. The first is that the industry norm avoids drawing the hard lines that arise when any of these variables change. When the poster is displayed for a longer period of time, or when it becomes a focal point for dialogue or the thematic development of the movie, does the result change? This custom therefore has a beneficial sorting effect that Rothman does not discuss. The cases where the art has no distinctive role are cases where the set designer can choose artwork in the public domain. The advantage of denying the fair use defense is that it stimulates artistic production by opening up a second revenue stream. If in fact there is widespread agreement on both sides of the industry, then the custom may well advance global efficiency by clarifying property rights and avoiding the endless litigation under the unavoidably vague statutory fair use standard. After all, it is not clear that the defendant should prevail under a straight statutory reading. The use made of the poster is, but only after a fashion, commercial; the creative nature of the copyrighted work—art—cuts in the plaintiff’s favor; the use of the work is entire, albeit for a short period of time; and the use of the work might stimulate the demand for posters, or alternatively, undercut that for set-dressings. This is a close case under the statute. Yet no matter which way it comes out, Rothman puts far too great a weight on the individual case relative to the need for stable institutional arrangements to deal with mass transactions. Before challenging this custom, I would like to see some systematic industry dissatisfaction with the arrangement, either by newcomers or established firms. On balance, this industry practice, if established, looks quite defensible.
I think that there is even stronger reason to defend the various commercial arrangements that have been developed over time for the reproduction of class packets and scholarly articles, both of which Rothman criticizes.9 These packets do not raise issues of artistic creativity such as those found in Ringgold. Rather, they are straight commercial disputes, involving the mass copying of large chunks of protected works, so that a heavy burden lies on the copier to show fair use even in the absence of custom. For example, the Classroom Guidelines set pretty strict limits for fair use that cover 250 words or less of poetry and other prose excerpts that range from 500 to 2500 words.10 Rothman notes that these industry minimums have quickly become standardized norms “in stark contrast to the open-ended nature of the fair use criteria set forth in Section 107.”11
This transformation is all to the good. Rothman notes that these guidelines have met with widespread compliance with most universities, but she never pauses to explain why that result does not make sense. Vague fair use tests do not always give an infringer a fighting chance to win. The usual course packet contains excerpts from writings of different length and type, and no prudent administrator wants to ask anyone to make a case by case determination with respect to each and every one of them. The fixed guideline provides an intelligible safe harbor and the particular word count for which permission is necessary. The clean line reduces uncertainty and saves administrative costs. Individual universities and libraries that are not signatories to the agreement have sufficient resources to litigate against the norms. That they do not suggests that this master settlement provides useful guidelines to all those other schools that did not participate in their negotiation. Nor is there a real risk of exploitation. There is every reason to believe that the participants to these negotiations had a large enough stake to be careful about the deal that they negotiated. Third parties are not excluded strangers who are hurt by negotiations that pay no weight to their concerns. Rather, they can confidently free ride on the standards that skilled universities use for their own affairs, without making any payment at all. This is hardly a case in which the nonparticipation of these parties flunks Rothman’s representativeness test.12 And it is worth noting that the license agreement supplies revenues to encourage the production of new works, which of course the fair use defense forecloses.Research Copies
I take the same view toward American Geophysical Union v. Texaco, where the noncompliance with industry custom figured in the Second Circuit’s rejection of the industry custom.13 Once again Rothman deplores a result that is eminently defensible when fully analyzed. At stake in the case was whether each of Texaco’s many research scientists could make a single copy of articles from a journal for their own personal use. I have no doubt that any individual subscriber is protected by the fair use doctrine when he makes a copy for personal use. But the cumulative impact changes when the same journal is circulated seriatim to 500 or more readers. In principle, the journal seller might increase the price of the subscriptions to research institutions and libraries to offset their more intensive use, as is commonly done with the higher subscription prices to libraries. But this approach would quickly lead to evasive tactics (i.e. institutional subscriptions that masquerade as individual ones). Or the journal seller could altogether cut out sales to institutional subscribers whose low internal use levels do not justify paying the higher price. The sensible way to handle this difficulty is to develop a reliable source of reprints that allows the publisher to pick up on the differential demands, just as a phone company monitors use by charging in accordance with intensity of use. The publishers sensed this opportunity and created a Copyright Clearance Center (“CCC”) (which Rothman did not discuss) which facilitated the massive reprint purchases at low rates. The upshot is a stronger incentive for the publication of journals in the first place. This institutional response removed any transaction-cost barriers to voluntary sales and allowed for the precise metering that cannot take place if the fair use defense is accepted. The metering also permits the collection of revenue to spur production that the fair use defense denies. To be sure, it was not crystal clear that the fair use defense should have been allowed before the CCC was developed. But a good case could be made in that direction because any such potential market would surely have been forestalled if the fair use defense was recognized before the formation of the CCC. The systemic risk here, contrary to Rothman’s contention, is that once the fair use exception gets embedded in common practice, it will act as a deterrent to some future property-rights solution. There are good and sufficient reasons then to think that customary practice points us in the right direction.
In sum, there is good reason to think that Rothman’s more ambitious attacks on industry custom should be rejected. In this context, custom is really a proxy for the creation of voluntary markets in a low transaction-costs environment. If these were inefficient, then the fair use defense, which calls for no compensation, might be justified. But the fair use defense means that the supplier of the copyrighted good gets no revenues from the use of its copyrighted material, which neglects the incentive function of copyright law. Rather than denigrate custom, the wiser strategy is to develop clear rules that aid in the emergence and enforcement of voluntary markets.
Preferred citation: Richard A. Epstein, Some Reflections on Custom in the IP Universe, 93 Va. L. Rev. In Brief 223 (2008), http://www.virginialawreview.org/inbrief/2008/01/21/epstein.pdf.
[Back to this issue]