Part 2 of 2
3. Remaining IndeterminacyAt bottom, the transformative use doctrine sets out a test for whether the defendant’s use is transformative that may be precisely enunciated, but which is susceptible of incoherence and judicial manipulation as applied in practice. The test quite clearly requires the court to identify the expressive purpose for which the author of the copyrighted work created that work and the expressive purpose for which the defendant copied from the work, and then to compare the two to determine if the defendant’s expressive purpose materially differs from that of the author. But neither the test nor precedent provides dispositive rules for how broad the relevant categories of expressive purpose should be, and just how different the defendant’s expressive purpose must be from that of the author to qualify as a transformative use.
Nor does the test definitively constrain judicial discretion on how to characterize the author’s and defendant’s expressive purpose. For example, in Calkins v. Playboy, [150] one of the cases alluded to above, a photographer sued Playboy for reproducing his high school yearbook portrait of a teenage girl who later became a Playboy model in the magazine’s photographic centerfold feature of the model. The court characterized the high school portrait photographer’s expressive purpose as creating a gift for family and friends, and Playboy’s use as one designed to “inform and entertain” its readers through personalizing the model. [151] However, a court that wished to find that Playboy’s use was not transformative could fairly and more broadly characterize the high school yearbook portrait as informational or biographical, while narrowly characterizing Playboy’s choice to reproduce its model’s high school portrait as serving an informational and biographical purpose within a glossy photographic spread otherwise designed for entertainment.
As courts have in other cases, the Calkins court relied on evidence of the defendant’s motive for using the copyrighted work, as presented in the defendant’s own statements and testimony. [152] The defendant’s statement of motive provides an evidentiary basis for determining the expressive purposes for the defendant’s use. But reliance on such statements, obviously, opens the door for users’ self-serving, strategic statements to build a fair use defense in anticipation of a possible copyright infringement lawsuit.
D. Case OutcomesThe transformative use paradigm merely means that the key question for fair use analysis is whether the court characterizes the use as transformative. Despite the paradigm’s favorable view of fair use as integral to the overall purposes of copyright law, its ascendancy does not necessarily portend that more uses will be found fair than under pre-2006 case law. So what impact has the transformative use paradigm had on case outcomes? The data, of course, can show correlation, but not causation. Nonetheless, there seems to be a strong correlation between increased win rates for fair use defendants and judicial adoption of the transformative use doctrine at the district court level. The appellate court level shows at best a weak correlation.
Beebe reported that during the period he studied, only 30.4% of the unreversed district court preliminary injunction decisions, 24.1% of the unreversed bench trial opinions, and 37.5% of the unreversed district court rulings on cross motions for summary judgment found fair use. [153] The combined total in such cases in Beebe’s study is a fair use win rate at the district court level of only 32.1%. The unreversed circuit court rulings compiled by Professor Beebe were somewhat more favorable to fair use. Their holdings in favor of fair use were, respectively, 40% for preliminary injunction decisions, 38.5% for bench trial opinions, and 55% for cross motions for summary judgment, representing a combined total of 43.75% fair use wins for the defendant in such cases. [154]
In reporting fair use case outcomes, Beebe presents convincing reasons for excluding uncrossed motions for summary judgment. [155] As he notes, summary judgment motions brought only by one side tend to present a clearer case in that party’s favor than when both parties bring conflicting summary judgment motions. Moreover, Beebe cites data showing that courts are more likely to publish an opinion granting summary judgment than denying it. [156] My data supports Beebe’s arguments. During the period 1995–2010, the plaintiff prevailed in 84.21% of the reported opinions in which only the plaintiff brought a motion for summary judgment, and the defendant prevailed in 70% of the reported opinions in which only the defendant brought a motion for summary judgment.
Further, my data provides an additional reason for excluding uncrossed motions for summary judgment in an empirical analysis of historical trends in fair use outcomes: there is a statistically significant variation in the proportion of reported rulings on uncrossed plaintiff versus defendant motions for summary judgment over time. In 2001–2005 only 19% of uncrossed motions for summary judgment were brought by plaintiffs, while in 1995–2000, the previous period, and 2006–2010, the later period, that percentage was, respectively, 33% and 42%. Thus, a longitudinal study of case outcomes that includes uncrossed motions for summary judgment will present results that are skewed by the significant variation in the mix of uncrossed plaintiff versus defendant summary judgment motions over time. Accordingly, like Beebe, in my analysis of fair use case outcomes I exclude motions for summary judgment brought only by one party or the other, and include data only from unreversed preliminary injunction decisions, bench trial opinions, and rulings on cross motions for summary judgment.
Within those parameters, my study adds a temporal dimension to Beebe’s findings and shows a remarkable shift in favor of finding fair use in such cases at the district court level since 1995. During the period 1995–2000, only 22.73% of the district court opinions in such cases found that the defendants had made a fair use of the plaintiffs’ work. During the period 2001–2005, the percentage of fair use wins for the defendant in such cases increased to 40.91%. During the years 2006–2010, the period in which courts have overwhelmingly embraced the transformative use doctrine, 58.33% of the opinions found in favor of the defendant on the issue of fair use. [157]
The parallel data for appellate rulings on motions for preliminary injunction, bench trial appeals, and cross motions for summary judgment does not match the shift in favor of fair use that we see for such opinions at the district court level. In fact, while the appellate rulings in such cases show an increase in fair use wins for defendants from 1995–2000 to 2001–2005, moving from 46.67 to 57.14%, they exhibit a slight decrease in findings of fair use from 2001–2005 to 2006–2010, when the fair use win rate for defendants in appellate opinions declined to 50%. The defendant win rate at the appellate level during the years 2006–2010, and indeed during the previous two periods following the Campbell decision as well, is slightly higher than the defendant win rate at the appellate level during the entire period of Beebe’s study. But given the small sample size of appellate opinions and relatively small increase in fair use wins for the defendant, the difference may well be just a reflection of statistical randomness.
As Beebe discusses, the high plaintiff win rates on the issue of fair use for the pre-2006 district court cases he studied run contrary to the much cited “Priest–Klein selection hypothesis,” which predicts that civil litigation plaintiff win rates at trial should typically approach 50%, largely because it is only the close cases that survive settlement or summary judgment. [158] Indeed, the plaintiff win rate of 67.9% during Beebe’s period of study is all the more striking given that empirical studies of the Priest–Klein selection hypothesis have demonstrated that, for various possible reasons, an overwhelming majority of types of lawsuits in fact exhibit plaintiff win rates of considerably less than 50%. For example, a study of all district court outcomes in civil cases reported in the Seventh Circuit between 1982 and 1987 found that plaintiffs initially won at trial in only 26.9% of the cases and won only 31.4% of time after appeals had been decided. [159] Notably, that study also provides ancillary, contextual support for Beebe’s finding that fair use outcomes were highly skewed in favor of plaintiffs in comparison to other types of litigation generally. In the Seventh Circuit study, copyright infringement plaintiffs initially won at trial 71.4% of the time, and after appeal in 57.1% of the cases, a far greater win rate for copyright infringement plaintiffs than for plaintiffs generally. [160]
Beebe offers two possible explanations for the discrepancy between win rates for plaintiffs on fair use and the Priest–Klein hypothesis. First, as Priest and Klein recognize, their 50% hypothesis does not apply when the stakes of the parties differ; repeat players and other parties with greater stakes in, or risk aversion to, the litigation are likely to have a higher degree of success in adjudicated outcomes because they will be more likely to settle near-close, as well as close, cases. [161] In that vein, Beebe conveys William Landes’s argument that intellectual property plaintiffs tend to have higher stakes in a favorable litigated outcome than civil litigation plaintiffs generally since IP plaintiffs face the risk that an adverse judgment will limit or extinguish their rights vis-à-vis third parties. [162] As a result, we would expect IP plaintiffs to settle near-close, as well as close, cases rather than take that risk. Second, Beebe adds that some of those who have decided to defend a copyright infringement lawsuit might raise a frivolous fair use defense in addition to more credible defenses. [163] To the extent that either explanation has purchase—and Beebe raises them merely as possibilities—they would apply equally to post-2005 cases, which makes the significant shift in favor of fair use even more notable.
We now turn to the correlation between increased defendant win rates and judicial embrace of the transformative use doctrine. As noted above, there has been a sharp increase in the embrace of the transformative use doctrine for all judicial opinions from the first post-Campbell period of my study, 1995–2000, to the most recent period, 2006–2010. To isolate the possible correlation between defendant win rates and judicial adoption of the transformative use doctrine, I examined judicial adoption of the doctrine and findings on transformativeness for the mix of reported district court opinions that form the basis for Beebe’s and my reporting of case outcomes—that is, preliminary injunction decisions, rulings on crossed motions for summary judgment, and bench trial judgments. As indicated in Table A, I found, in those opinions, a sharp increase over time in (1) judicial assessment of whether the use is of a type favored under the transformative use doctrine (regardless of whether the court expressly uses the term “transformative”); (2) the percentage of cases in which the court found that the defendant’s use was favored under the transformative use doctrine; and (3) the percentage of cases in which the defendant won on fair use when the court considered the transformativeness of the use. [164] I also found (4) a consistently high rate of defendant wins, reaching 100% in 2001–2005 and 2006–2010, in those cases in which the court found that the defendant’s use was, in fact, unequivocally transformative. At bottom, the respective increases in district courts’ embrace of the transformative use doctrine and in their findings that the use in question is, in fact, transformative during 2006–2010 each correlate significantly with (5) the increase in defendant win rates in fair use case outcomes for our mix of district court cases since 1995–2000.
Table A. The Transformative Use Doctrine in Unreversed District Court Preliminary Injunctions, Bench Trials, and Crossed Motions for Summary Judgment
As Sag emphasizes, the increasing rate at which judges invoke the transformative use concept does not necessarily mean that the transformative use paradigm is actually driving the trend towards significantly greater win rates for fair use defendants at the district court level. [165] However, particularly in light of the empirical studies discussed above, concluding that legal doctrine does generally impose some constraint on judicial discretion and thus impact case outcomes, [166] I hypothesize that the judicial turn towards the transformative use paradigm is, at the very least, partly responsible for greater defendant win rates. Moreover, Sag’s principal results would appear to comport with that possibility.
In his study of district court cases, Sag concludes that the case attributes that have the most significant correlation with a fair use defendant’s successful outcome are the plaintiff’s legal personality and whether the defendant’s use constitutes a “Creativity Shift,” i.e., a use of a creative work for informational purposes or visa-versa. [167] With regard to legal personality, Sag concludes that fair use defendants are more likely to win if the plaintiff is a natural person than if a corporation. [168] Interestingly, however, there is no statistically significant difference between the proportions of plaintiffs who are natural persons and corporations in unreversed district court opinions during 2006–2010 than the relative proportion of plaintiffs with those respective legal personalities during 1978–2006, the period of Sag’s study. [169] This suggests that even though the plaintiff’s legal personality significantly correlated with defendant win rates overall during 1978–2006, it was not a factor driving increased win rates for fair use defendants over time into the period 2006–2010.
With regard to Creativity Shift, Sag concludes that when a defendant’s use falls within that category, the defendant’s predicted probability of a favorable finding on fair use almost doubles, from 33 to 62%. [170] I have not attempted to determine the proportion of cases that might be characterized by a Creativity Shift, as Sag defines it, during 2006–2010. Yet the use of a creative work for informational purpose or an informational work for purposes of creative expression falls solidly within the widely accepted definition of transformativeness in that they are uses in which the defendant’s expressive purpose differs from that of the copyrighted work’s author. Hence, as Sag notes, Creativity Shifts are a “kind of transformative use.” [171] Accordingly, even if the proportion of Creativity Shifts sharply increased and remained a significant predictor of fair use outcomes for 2006–2010, the data would not show whether courts were applying the transformative use concept to find that Creativity Shifts are fair use, as the transformative use paradigm suggests they should, or merely labeling Creativity Shifts as “transformative” to justify case outcomes post hoc. [172]
More speculatively (and without quantitative, empirical support), I offer a further alternative hypothesis for the increase in defendant win rates: the tilt towards fair use defendants and the judicial embrace of the transformative use doctrine are, together, part of a broader, growing judicial skepticism towards copyright holder rights, for which the Supreme Court’s decision in Eldred v. Ashcroft, [173] issued in 2003, might have been the watershed. In Eldred, the Supreme Court rejected a constitutional challenge to the Copyright Term Extension Act of 1998, in which Congress added an additional 20 years to the copyright term for new and existing works. [174] But, as the next day’s New York Times headline proclaimed, while Eldred was a “corporate victory” for the copyright industries who lobbied for the term extension, it was one that “raise[ed] public consciousness” about copyright excess, [175] and tarnished the copyright industries as greedy and overreaching. Since Eldred, courts have interpreted other 1998 amendments to the Copyright Act to provide far-reaching immunity to Internet service providers, like YouTube, which host infringing material posted by users; [176] held that a cable operator that stores television programs for subscribers on a company server has not itself reproduced or publicly distributed the copied programs; [177] limited the availability of statutory damages in lawsuits against peer-to-peer file trading sites and other secondary infringers; [178] and held that copyright infringement plaintiffs are not entitled to a presumption of irreparable harm in motions for preliminary injunction. [179] In addition, in Eldred’s wake, the Supreme Court sharply curbed copyright holders’ ability to use federal trademark rights to control uses of formerly copyrighted material that is in the public domain, [180] and in a possible step towards further limiting Eldred, the Supreme Court has recently granted cert to hear a constitutional challenge to a 1994 law that restored copyright protection to certain works that had gone into the public domain. [181] To the extent these rulings are representative of a backlash against a perceived expansion of copyright holder rights and demands, the tilt towards defendants in fair use cases might be a part of that general move.
In any event, the Priest–Klein selection hypothesis teaches that the judicial shift in favor of defendants in fair use outcomes is unlikely to persist even if the transformative use paradigm continues to dominate fair use analysis and is the driving force in that shift in outcomes. At some point, we would expect that copyright holders will settle or cease to litigate copyright infringement claims in which the user has a colorable argument of transformative use, leaving only materially uncertain areas of fair use doctrine for adjudicated outcomes. Indeed, the data presented in Figure 4 on the proportion of fair use defendant wins over time suggests that defendant win rates might have already begun to return to a mean of somewhat below 50% during the last half of 2006–2010, although the number of opinions is too small to discount the possibility of statistical randomness.
Figure 4. Twenty-Opinion Moving Average of the Proportion of Unreversed District Court Opinions (Preliminary Injunctions, Bench Trials, and Crossed Motions for Summary Judgment) in Which Defendant Wins on Fair Use, 1978–2010.
_______________
Notes:85. My data set is available as a link alongside this Article at
http://www.law.ucla.edu/faculty/bibliog ... tanel.aspx.
86. The dissent was in Bouchat v. Baltimore Ravens Ltd. P’ship (Bouchat II), 619 F.3d 301, 317 (4th Cir. 2010) (Niemeyer, J., dissenting).
87. Id. (holding that defendant’s use of plaintiff’s design in team logo in a football team’s highlights film was not fair use but that its use in photos displayed in the team’s corporate headquarters was fair use); Peter Letterese & Assocs., Inc. v. World Inst. of Scientology Enters., Int’l, 533 F.3d 1287 (11th Cir. 2008) (defendant church copied from plaintiff’s sales techniques book in two separate publications, one for internal church purposes and another marketed for secular use); Henley v. DeVore, 733 F. Supp. 2d 1144 (C.D. Cal. 2010) (defendant used two of plaintiff’s songs).
88. There are significant shifts in the mix of case posture over time, but with the exception of uncrossed motions for summary judgment, none impact the transformation of fair use doctrine that I describe. As I discuss below, both Beebe and I eliminate uncrossed motions for summary judgment from our analysis of case outcomes in order to avoid skewing the results as a result of fluctuations in the mix of plaintiff versus defendant uncrossed motions for summary judgment. See infra notes 155–157 and accompanying text.
89. For a chilling account of pervasive flaws and random impacts in major statistical studies, as well as scientific publications’ selection biases, see Jonah Lehrer, The Truth Wears Off; Is There Something Wrong with the Scientific Method?, NEW YORKER, Dec. 13, 2010, at 52.
90. Sag, supra note 13 (manuscript at 6). The pioneer article on the reported case selection bias is George L. Priest & Benjamin Klein, The Selection of Disputes for Litigation, 13 J. LEGAL STUD. 1 (1984). See also Samuel Issacharoff, The Content of Our Casebooks: Why Do Cases Get Litigated?, 29 FLA. ST. U. L. REV. 1265 (2002).
91. See Wendy Parker, Lessons in Losing: Race Discrimination in Employment, 81 NOTRE DAME L. REV. 889, 910–11 (2006). Some studies also find differences in result between reported cases that are certified for publication and those that are not. See Lee Epstein & Gary King, The Rules of Inference, 60 U. CHI. L. REV. 1, 106–08 (2002) (surveying the literature). I did not attempt to compare reported with unreported fair use decisions. However, in examining all reported fair use opinions issued since 1994, I found no statistically significant difference between published opinions and those opinions reported in Westlaw or LexisNexis but not certified for publication for either of the two critical factors in my analysis: fair use outcomes and judicial adoption of the transformative use doctrine.
92. See Beebe, supra note 11, at 565 (noting the paucity of reported fair use opinions as compared to the number of copyright infringement complaints that are filed and surmising that “many fair use disputes may never reach the courts”).
93. See generally PATRICIA AUFDERHEIDE & PETER JASZI, CTR. FOR SOC. MEDIA, UNTOLD STORIES: CREATIVE CONSEQUENCES OF THE RIGHTS CLEARANCE CULTURE FOR DOCUMENTARY FILMMAKERS (2004); LESSIG, supra note 3; James Gibson, Risk Aversion and Rights Accretion in Intellectual Property Law, 116 YALE L.J. 882 (2007).
94. See Issacharoff, supra note 90, at 1270–71 (noting the impact of existing case law on prospects for settlement); Sag, supra note 13 (manuscript at 6 n.22) (“[T]he selection effect may not even be constant as the results of prior cases necessarily inform the expectations of future litigants.”).
95. See infra Part IV.B.
96. Wendy J. Gordon, Fair Use as Market Failure: A Structural Analysis of the Betamax Case and its Predecessors, 82 COLUM. L. REV. 1600 (1982) [hereinafter Gordon, Structural Analysis]. Gordon subsequently distanced herself from that market paradigm, emphasizing the importance of non-monetary values in fair use doctrine and in copyright law generally. See Wendy J. Gordon, A Property Right in Self-Expression: Equality and Individualism in the Natural Law of Intellectual Property, 102 YALE L.J. 1533 (1993); Wendy J. Gordon, Market Failure and Intellectual Property: A Response to Professor Lunney, 82 B.U. L. REV. 1031 (2002).
97. Sony Corp. of Am. v. Universal City Studios, Inc., 464 U.S. 417, 478 (Blackmun, J., dissenting).
98. Harper & Row, Publishers, Inc. v. Nation Enters., 471 U.S. 539, 566 n.9 (1985). See also id. at 559 (“[T]o propose that fair use be imposed whenever the ‘social value [of dissemination] . . . outweighs any detriment to the artist,’ would be to propose depriving copyright owners of their right in the property precisely when they encounter those users who could afford to pay for it.” (alteration in original) (citing Gordon, Structural Analysis, supra note 96, at 1615)).
99. Id. at 549–50.
100. Id. at 562.
101. Id.
102. Id. at 568 (quoting Sony, 464 U.S. at 451).
103. Id.
104. 510 U.S. 569, 576 (1994).
105. Id. at 578 n.10.
106. The percentage of opinions in which judges adopted the transformative use doctrine during 2006–2010 represents a statistically significant increase since 1995–2000 but is not statistically significant when one takes into account the increase from 1995–2000 that already occurred during 2001–2005. However, as I discuss below with respect to case outcomes, the results for 2001–2005 are skewed in favor of defendants by a relatively low percentage of uncrossed motions for summary judgment brought by plaintiffs during that period. The increase in judicial adoption of the transformative use doctrine during 2006–2010 is statistically significant for district court opinions when uncrossed summary judgment motions are eliminated from the analysis. See infra notes 155–164 and accompanying text.
107. That outlier decision was Thomas M. Gilbert Architects, P.C. v. Accent Builders & Developers, LLC, 377 Fed. App’x 303 (4th Cir. 2010).
108. The data for appellate opinions only does not show the same upward curve. During the first 6 year period after Campbell, 19 out of 20, or 95%, of appellate opinions invoked the transformative use doctrine, which is slightly higher than the appellate court’s embrace of the doctrine during the most recent period.
109. See, e.g., Lorimar Music A. Corp. v. Black Iron Grill Co., No. 09-6067-CV, 2010 WL 3022962, at *4 (W.D. Mo. July 29, 2010) (quoting Campbell, 510 U.S. at 579).
110. As Judge Leval noted, “[t]ransformative uses may include criticizing the quoted work, exposing the character of the original author, proving a fact, or summarizing an idea argued in the original in order to defend or rebut it. They also may include parody, symbolism, aesthetic declarations, and innumerable other uses.” Leval, supra note 54, at 1111.
111. See, e.g., Bouchat v. Baltimore Ravens Ltd. P’ship (Bouchat I), 587 F. Supp. 2d 686, 695 (D. Md. 2008) (citing Bill Graham Archives v. Dorling Kindersley Ltd., 448 F.3d 605, 609 (2d Cir. 2006), a leading Second Circuit decision on the importance of the transformativeness of the use for the fair use analysis).
112. See Beebe, supra note 11, at 604.
113. The data regarding express reference to the term “transformative” as broken down by appellate and district court opinions is more erratic, showing a marked increase in appellate opinions that did not make such a reference in 2001–2005, followed by a sharp decrease in 2006–2010, and a slight increase in district court opinions that failed to make that express reference in 2006–2010 as compared to the previous five-year period. It is the combined total of all opinions, appellate and district court, that suggests that there has been a steady increase in influence of the transformative use doctrine using Beebe’s measure.
114. See supra note 72 and accompanying text. Although copying for purposes of litigation does not appear among the favored uses enumerated in section 107’s introductory clause, the House Report accompanying the Copyright Act of 1976 listed the “reproduction of a work in . . . judicial proceedings” as an example of “the sort of activities that the courts might regard as fair use under the circumstances.” House Committee on the Judiciary, H.R. Rep. No. 94-1476, at 65, (1976), reprinted in 1976 U.S.C.C.A.N. 5659, 5678.
115. Kelly v. Arriba Soft Corp., 77 F. Supp. 2d 1116 (C.D. Cal. 1999); Kelly v. Arriba Soft Corp., 336 F.3d 811 (9th Cir. 2003).
116. The three opinions involving the Google search engine were: Perfect 10 v. Google, Inc., 416 F. Supp. 2d 828 (C.D. Cal. 2006); the same case on appeal, Perfect 10, Inc. v. Amazon.com, Inc., 508 F.3d 1146 (9th Cir. 2007); and Field v. Google Inc., 412 F. Supp. 2d 1106 (D. Nev. 2006). The two opinions involving the digital plagiarism detection service were: A.V. v. iParadigms, LLC, 544 F. Supp. 2d 473 (E.D. Va. 2008); and the same case on appeal, A.V. ex rel. Vanderhye v. iParadigms, LLC, 562 F.3d 630 (4th Cir. 2009).
117. The court recently rejected the proposed settlement in that case. See Authors Guild v. Google, Inc., No. 05 Civ. 8136 (S.D.N.Y. March 22, 2011).
118. Bridgeport Music, Inc. v. UMG Recordings, Inc., 585 F.3d 267, 278 (6th Cir. 2009). The decision might also be explained by the court’s use of “transformative” to mean altered aesthetic character rather than different expressive purpose, which cuts against the prevailing understanding of “transformative.” It found the defendant’s song to be transformative given that it had “a different theme, mood, and tone” from the plaintiff’s composition. Id.
119. Shepard v. Miler, No. Civ. 2:10-1863, 2010 WL 5205108, at *6 (E.D. Cal. Dec 15, 2010).
120. Gulfstream Aerospace Corp. v. Camp Sys. Int’l, Inc., 428 F. Supp. 2d 1369 (S.D. Ga. 2006) (plaintiff sought to prevent reproduction of aircraft manual in order to stifle competition in providing maintenance service).
121. Super Future Equities, Inc. v. Wells Fargo Bank Minn., N.A., 553 F. Supp. 2d 680 (N.D. Tex. 2008) (posting of loan servicer’s image on webpage criticizing loan servicer’s business practices).
122. Peter Letterese & Assocs., Inc. v. World Inst. Of Scientology Enters., 533 F.3d 1287, 1318–19 (11th Cir. 2008). There are also three cases that involve two distinct uses, with the courts assessing the transformative character of each use separately: Peter Letterese & Assocs., Inc., 533 F.3d 1287 (defendant church copied from plaintiff’s sales techniques book in two separate publications, one for internal church purposes and another marketed for secular use); Henley v. DeVore, 733 F. Supp. 2d 1144 (C.D. Cal. 2010) (two songs); and Bouchat v. Baltimore Ravens Ltd. P’ship (Bouchat I), 587 F. Supp. 2d 686 (D. Md. 2008) (use of logo in a football team’s highlights film and in photos displayed in the team’s corporate headquarters); and another case in which the same use was held to be transformative vis-à-vis one of the plaintiff’s works but not vis-à-vis another of the plaintiff’s works. Warner Bros. Entm’t, Inc. v. RDR Books, 575 F. Supp. 2d 513 (S.D.N.Y. 2008).
123. Sag, supra note 13 (manuscript at 11) (citing Nimmer’s discussion of Second Circuit cases).
124. Id.
125. Whatever the weakness of the commerciality presumptions, my data shows that, for each of the three periods, a judicial finding that the use is commercial correlates significantly with a fair use win for the plaintiff, but that a judicial finding that the use is also unequivocally transformative trumps the significance of commerciality. For the entire period 1995–2010, the plaintiff won on fair use in 90.3% of the cases in which the court held that the use was commercial but not (unequivocally) transformative, versus only 12.5% of the cases in which the court held that the use was both commercial and unequivocally transformative. For the period 2006–2010, those plaintiff win rates were 88.2% versus 20%, respectively.
126. See Beebe, supra note 11, at 584–85.
127. See, e.g., Bill Graham Archives v. Dorling Kindersley Ltd., 448 F.3d 605, 615 (2d Cir. 2006).
128. In 2006–2010, a surprisingly high 30.43% of opinions that found that the use was not transformative, nevertheless found that the use did not harm the plaintiff’s market. For 1995–2000 and 2001–2005, that figure was 0.00% and 7.14%, respectively.
129. Indeed, in the prior two periods as well, a substantial percentage of opinions that acknowledged the Harper & Row dictum found no market harm and that the fourth factor favored the defendant. In 1995–2000, 50% of those opinions found no market harm and 44.44% found that the fourth factor favored the defendant. In 2001–2005, 42.86% of those opinions found no market harm and 37.5% found that the fourth factor favored the defendant. The variations from one period to the next, and as compared with 2006–2010, are not statistically significant.
130. More than 75% of the 2006–2010 opinions that found that the use was transformative expressly considered whether the defendant copied more than necessary for the defendant’s expressive purpose in the court’s analysis of factor three. In the 1995–2000 and 2001–2005, that figure was 44.44% and 57.14 %, respectively.
131. Campbell v. Acuff-Rose Music, Inc., 510 U.S. 569, 579 (1994).
132. For discussion of this confusion, as well as other ambiguities in the definition of transformative use, see Diane Leenheer Zimmerman, The More Things Change, the Less They Seem “Transformed”: Some Reflections on Fair Use, 46 J. COPYRIGHT SOC’Y 251, 256–268 (1998).
133. Campbell, 510 U.S. at 579.
134. Moreover, in his article, Judge Leval states that adding “new aesthetics” and creativing a “derivative work” can qualify as transformative, even if not necessarily fair use when copying is excessive. Leval, supra note 54, at 1111–12.
135. See, e.g., 2 PAUL GOLDSTEIN, GOLDSTEIN ON COPYRIGHT § 12:34–35 (3d ed. Supp. 2011) (“On principle, the rule [weighing transformativeness in favor of fair use] threatens to undermine the balance that Congress struck in section 106(2)’s derivative rights provision to give copyright owners exclusive control over transformative works to the extent that these works borrow copyrightable expression from the copyrighted work.”).
136. R. Anthony Reese, Transformativeness and the Derivative Work Right, 31 COLUM. J.L. & ARTS 467, 485 (2008).
137. See supra text accompanying notes 67–70.
138. In the single outlier decision between 2006 and 2010, Bridgeport Music v. UMG Recordings, the Sixth Circuit found the defendant’s song to be “certainly transformative” because it had “a different theme, mood, and tone” from the plaintiff’s composition, but then declined to overturn a jury’s verdict rejecting the defendant’s fair use defense. 585 F.3d 267, 278 (6th Cir. 2009).
139. See, e.g., Peter Letterese & Assocs., Inc. v. World Inst. of Scientology Enters., Int’l, 533 F.3d 1287, 1311 (11th Cir. 2008) (“Although the [defendant’s] course materials adopt a different format, incorporate pedagogical tools such as sales drills, and condense the material in the [plaintiff’s] book, these changes do not alter the educational character of the material taken from the book; they merely emphasize, rather than transform, the overall purpose and function of the book.”).
140. See Soc’y of the Holy Transfiguration Monastery, Inc. v. Archbishop Gregory, 685 F. Supp. 2d 217, 227 (D. Mass. 2010) (holding that making a religious text available for religious instruction on the Internet constitutes non-transformative repackaging). Courts have long been reluctant to find fair use when they find that the defendant has merely retransmitted the copyrighted work in a different medium. See, e.g., Infinity Broad. Corp. v. Kirkwood, 150 F.3d 104, 108–09 (2d Cir. 1998) (concluding that retransmission of radio broadcast over telephone lines is not transformative).
141. Sag, supra note 13 (manuscript at 16, 27).
142. A.V. ex rel. Vanderhye v. iParadigms, LLC, 562 F.3d 630, 644 (4th Cir. 2009); Perfect 10, Inc. v. Amazon.com, Inc., 508 F.3d 1146, 1165 (9th Cir. 2007); Kelly v. Arriba Soft Corp., 336 F.3d 811, 818–19 (9th Cir. 2003); Warner Bros. Entm’t, Inc. v. RDR Books, 575 F. Supp. 2d 513, 541 (S.D.N.Y. 2008).
143. Bill Graham Archives v. Dorling Kindersley Ltd., 448 F.3d 605, 615 (2d Cir. 2006); Warren Publ’g Co. v. Spurlock, 645 F. Supp. 2d 402, 423 (E.D. Pa. 2009).
144. Blanch v. Koons, 467 F.3d 244, 253 (2d Cir. 2006).
145.Calkins v. Playboy Enters. Int’l, Inc., 561 F. Supp. 2d 1136, 1141 (E.D. Cal. 2009).
146. Bouchat v. Baltimore Ravens Ltd. P’ship (Bouchat II), 619 F.3d 301, 314 (4th Cir. 2010).
147. While 86% of opinions considering author criticism between 1995 and 2010 found the use to be unequivocally transformative, courts are split on whether copying to criticize the author, as opposed to the work itself, constitutes a parody, which is especially favored for fair use purposes. See Henley v. DeVore, 733 F. Supp. 2d 1144, 1152–54 (C.D. Cal. 2010) (discussing split case law on the issue of whether copying to criticize the author as distinct from the work itself is a parody).
148. Between 2006 and 2010, 11 opinions considered uses in advertising and commercial promotion. All but two found the use to be non-transformative and the remaining two found the respective uses in question to be “somewhat transformative” and “minimally transformative at best.” Reyes v. Wyeth Pharmaceuticals, Inc., 603 F. Supp. 2d 289, 296–97 (D.P.R. 2009) (use of photograph of sculpture in an advertisement as part of public service education campaign); Designer Skin, LLC v. S & L Vitamins, Inc., 560 F. Supp. 2d 811, 823 (D. Ariz. 2008) (internet reseller’s use of depictions of plaintiff’s products to market those products to consumers was “minimally transformative at best”). In disfavoring uses for advertising and promotion, lower courts have followed Supreme Court dicta in Campbell, suggesting that “use . . . of a copyrighted work to advertise a product, even in a parody, will be entitled to less indulgence under the first factor of the fair use enquiry than the sale of a parody for its own sake.” Campbell v. Acuff-Rose Music, Inc., 510 U.S. 569, 585 (1994).
149. See, e.g., Video Pipeline, Inc. v. Buena Vista Home Entm’t, Inc., 342 F.3d 191, 198–200 (3d Cir. 2003) (display of movie trailers to sell videos was not a transformative use); Teter v. Glass Onion, Inc., 723 F. Supp. 2d 1138, 1153–54 (W.D. Mo. 2010) (art gallery’s display of digital images of artist’s work was not transformative since the images served only the promotional function of showing customers that the artist’s works were available in the gallery); United States v. Am. Soc’y of Composers, Authors & Publishers, 599 F. Supp. 2d 415, 432 (S.D.N.Y. 2009) (ringtone preview is advertising, not informational).
150. 561 F. Supp. 2d 1136.
151. Id. at 1141.
152. Id.
153. Beebe, supra note 11, at 577.
154. Id. at 578.
155. Id. at 576.
156. Id.
157. By point of comparison, the defendant win rates in all unreversed district court opinions, including rulings on uncrossed summary judgment motions, during the three periods was 31.34% in 1995–2000, 55.00% in 2001–2005, and 41.32% in 2006–2010. Those percentages show a modest increase in defendant wins comparing 1995–2000 and 2006–2010. But the highest rate of defendant wins for this open data set of all unreversed district court opinions was in 2001–2005, when there was a considerably greater proportion of uncrossed motions for summary judgment brought by defendants.
158. George L. Priest & Benjamin Klein, The Selection of Disputes for Litigation, 13 J. LEGAL STUD. 1, 4–6, 17–18 (1984).
159. Daniel Kessler, Thomas Meites, & Geoffrey Miller, Explaining Deviations from the Fifty-Percent Rule: A Mulitmodal Approach to the Selection of Cases for Litigation, 25 J. LEGAL STUD. 233, 249–51 (1996).
160. Id. at 251.
161. See id. at 237–48 (describing and building upon the Priest–Klein hypothesis).
162. Beebe, supra note 11, at 579. See William M. Landes, An Empirical Analysis of Intellectual Property Litigation: Some Preliminary Results, 41 HOUS. L. REV. 749, 772 (2004).
163. Beebe, supra note 11, at 580.
164. The statistical significance of the increase over time for each variable is set out in the Table footnotes.
165. See supra notes 123–124 and accompanying text.
166. See supra notes 82–83 and accompanying text.
167. See Sag, supra note 13 (manuscript at 27, 32).
168. Id. (manuscript at 29).
169. For the period of his study (1978–2006), Sag found that in 41.29% (64 out of 155) of the cases a natural person was the plaintiff and in 53.55% (83 out of 155) a corporation was the plaintiff. Matthew Sag, Predicting Fair Use (2011) (manuscript at 39), available at
http://works.bepress.com/matthew_sag/10. I found, for 2006–2010 (unreversed district courts), that in 46.7% (28 out of 60) of the cases a natural person was the plaintiff and in 48.3% (29 out of 60) a corporation was the plaintiff. The differences are not statistically significant—and, in any event, are far smaller than the percentage increase in defendant win rates from the period 1978–2006, which Sag measured at 32.28%, to the period 2006–2010, which for district court opinions, excluding uncrossed summary judgment motions, was 58.33% and for district court opinions including uncrossed summary judgment motions was 41.32%.
170. Sag, supra note 13 (manuscript at 28).
171. Id.
172. As noted in the text accompanying notes 78 and 79, supra, Sag also finds a significant correlation (although less so than that for plaintiff’s legal personality and Creativity Shifts) between fair use outcomes and whether the use is a Direct Commercial Use or Partial Copy. If the former, the defendant has a lesser chance of winning; if the latter, a greater chance. The definition of transformative use does not encompass all uses that are other than Direct Commercial Uses or all uses that are Partial Copies. Some transformative uses involve uses of the plaintiff’s work as part of a commercial product or service without any modification of the original copyrighted work, even if courts have declined to hold that uses for commercial advertising or promotion are transformative. Further, a use may be transformative even if the defendant has copied the plaintiff’s work in its entirety. Thus, if the proportion of uses that are Direct Commercial Uses have markedly declined during 2006–2010 or if those that are Partial Copies have markedly increased, this might suggest an alternative explanation for the significant increase in fair use defendant win rates during that period—or at least an additional explanation to supplement the judicial abandonment of the market paradigm and embrace of the transformative use paradigm, which is evident in how courts actually characterize their analysis of the four fair use factors.
173. 537 U.S. 186 (2003).
174. Id. at 218.
175. Amy Harmon, A Corporate Victory, But One That Raises Public Consciousness, N.Y. TIMES, Jan. 16, 2003, at A24.
176. See, e.g., Perfect 10, Inc. v. CCBill LLC, 488 F.3d 1102 (9th Cir. 2007) (interpreting the safe harbor provisions of the Digital Millennium Copyright Act); Viacom Int’l Inc. v. YouTube, Inc., 718 F. Supp. 2d 514 (S.D.N.Y. 2010) (same); IO Group, Inc. v. Veoh Networks, Inc., 586 F. Supp. 2d 1132 (N.D. Cal. 2008) (same).
177. Cartoon Network LP v. CSC Holdings, Inc., 536 F.3d 121 (2d Cir. 2008).
178. Arista Records LLC v. Lime Group LLC, No. 06 CV 5936, 2011 WL 1226277 (S.D.N.Y. March 29, 2011) (copyright plaintiff may not elect to receive the “extraordinary remedy” of statutory damages against a secondarily liable defendant when any individual direct infringer infringed the work in question prior to the work’s copyright registration).
179. Salinger v. Colting (Salinger II), 607 F.3d 68 (2d Cir. 2010).
180. Dastar Corp. v. Twentieth Century Fox Film Corp., 539 U.S. 23 (2003).
181. Golan v. Holder, 131 S. Ct. 1600 (2011), granting cert. to Golan v. Holder, 609 P.3d 676 (10th Cir. 2010).