Fair use is one of modern law’s most fascinating and troubling doctrines. It is amorphous and vague, as well as notoriously difficult to apply. It is, at the same time, vitally important in copyright and perhaps the most frequently raised and litigated issue in the law of intellectual property.

This Article synthesizes themes from the rich literature on fair use to fashion a novel theory of fair use that provides both a better understanding of the underlying principles and better tools for applying the doctrine.

In contrast to the dominant understanding of fair use in the literature—that fair use addresses market failures—the Article proposes that fair use is a tool that allocates a large bloc of uses directly to the public in order to limit the size of property rights that are granted to authors. The fair use doctrine, we argue, is an integral part of copyright’s sorting mechanism for, on the one hand, granting authors intellectual property rights based on their expected incentives for creation and, on the other hand, granting the public privileges based on the expected utility from direct allocation. The Article’s theory thus accords with recent Supreme Court cases by conceptualizing fair use not as an exception for costly transactions, but rather as a central feature of the copyright system that ensures productive and allocative efficiency.

This theory supports a reconceptualization of the basic structure of copyright law that both broadens fair use and makes the doctrine easier to apply. This Article favors a prima facie finding of fair use whenever the user’s category of use is one that produces widespread follow-on utility to nonusers (such as the categories of political speech or what we call “truth seeking”). This prima facie finding can be defeated only by showing that allowing such uses with respect to the particular copyrighted work would eliminate sufficient incentives for its creation.

TABLE OF CONTENTS

I.  Current Understandings of Fair Use

A.    The Market-Failure Understanding of Fair Use

The predominant theoretical justification of the fair use doctrine stems from an article by Professor Gordon over three decades ago.26 In an article that has become a classic among intellectual property scholars,27 Gordon tied fair use to the economic notion of market failure resulting from high transaction costs.28 To understand Gordon’s important insight, it is necessary to take a step back and begin with the animating philosophy that underlies the Copyright Clause of the Constitution. Article I, § 8 of the Constitution empowers Congress “[t]o promote the Progress of Science and useful Arts, by securing for limited Times to Authors and Inventors the exclusive Right to their respective Writings and Discoveries.”29 As many scholars have pointed out, the constitutional language suggests that the Framers viewed intellectual property protection as a means to an end, namely, to encourage the production of inventions and works of authorship.30

The grant of exclusive rights in intellectual goods in the form of copyright (and patent) protection has a double effect, both negative and positive. The negative effect of conferring protection is to take away certain rights from the public. If a work is protected by the law of copyright, the public may no longer copy, adapt, or otherwise use the work at will. At the same time, the grant has the positive effect of depositing all these rights in the control of the authors. Together, these two effects direct those users who wish to use copyrighted works to transact with authors and pay authors for the privilege of using their works of authorship. Importantly, it is the latter effect—authors’ ownership of rights—that induces users to pay authors through voluntary market transactions and that thereby incentivizes authors to keep creating.

It is imperative to understand, however, that the grant of exclusivity via copyright protection does not by itself guarantee pecuniary rewards to authors. Indeed, most authors discover that even though their works receive extensive copyright protection, no commercial success is forthcoming.31 On its own, legal exclusivity gives little to authors. It is merely a background condition necessary for the formation of markets. The actual remuneration of authors is decided by two parameters that are partly exogenous to the law. The first is the utility enjoyed by users. The second is transaction costs.

As Professor Ronald Coase stressed in his seminal article The Problem of Social Cost, it is only in the imaginary world without transaction costs that we can be certain that voluntary market transactions will inexorably lead to the efficient allocation of resources.32 In the real world, transaction costs are an ever-present influence on economic activity. These costs include expenditures on identifying the relevant counterparty, negotiating with it, formulating legal agreements, and, on some views, incurring enforcement costs.33 If transaction costs are too high relative to the gains of transacting, no voluntary transactions will take place—and as a consequence, legal entitlements will not gravitate to higher-value users through market transactions.34 As an illustration, consider the following example. Assume that Abbie, a PhD student, wishes to quote a short paragraph from a foreign book in her dissertation. Assume further that Abbie is willing to pay $12 for the right to quote from the book. Assume as well that the author’s asking price is $10. Although at first blush it may appear that the conditions will support a voluntary transaction, a closer look demonstrates the need for caution. Once we incorporate even modest transaction costs in the amount of $5 to the calculus, we can see that the transaction will not occur. Abbie will not agree to pay $15 ($10 in price plus $5 in transaction costs) to obtain a use permit that is worth only $12 to her. Doing so would represent a net loss of $3 to Abbie. In a world without fair use, therefore, Abbie will abstain from using the paragraph and forgo a potential gain of $2.

Building on this insight, Gordon suggested that fair use may be used to benefit users without (unduly) harming authors when transaction costs prevent voluntary exchanges from taking place.35 Gordon showed that allowing fair use in cases like Abbie’s would lead to a superior outcome. Considering Abbie a fair user permits her to use the paragraph for free, creating a net societal gain. If the author’s asking price actually reflects a cost to the author (in the form of dilution of the copyrighted work, for example), society is better off by $2 as a result of the fair use. In the more likely case, in which the author faces little or no marginal cost as a result of Abbie’s use, allowing Abbie the use without permission creates a net societal gain of $12 with all of the gain going to Abbie, yet not leaving the author worse off.

Gordon proposed three cumulative conditions that must be present for fair use to be recognized in her market-failure theory. First, transaction costs must bar voluntary exchanges between copyright owners and users. Second, the use for which fair use status is sought must promote the public interest. And third, allowing the contested use to continue under the rubric of fair use must not significantly diminish the author’s incentives to produce new works.36

Notwithstanding the elegance and analytical prowess of the market-failure theory of fair use, it suffers from several problems. First, as one of us has pointed out, the usefulness of the market-failure theory as a judicial tool is largely undermined by the fact that it relies on information that is nonobservable and nonverifiable by the courts.37 To use the proposed transaction cost framework to evaluate fair use claims, courts must know the value a user ascribes to a certain use—but this datum is private information that is known only to the user and may not be easily ascertained by the court. In litigation, any user who availed herself of a work without permission could argue that her use was of minimal value and that an agreement between the parties would therefore not have been feasible in light of the transaction costs.

Second, and no less significantly, the market-failure theory of fair use portrays fair use as an exception to the general rule that all use rights should reside with copyright owners. Indeed, in later writings, Gordon herself addressed this problem with the market-failure theory.38 Elaborating on this point, Professor Raymond Ku wrote: “Under [the market-failure] approach, fair use is an exception to the otherwise exclusive rights of copyright justified by the presence of market barriers such as high transaction costs, externalities, non-monetizable benefits, or anti-dissemination motives.”39 However, as we show in the next Section, fair use is not understood by courts as an anomaly or an exception. It is seen as central to the structure of copyright law. The market-failure theory is thus problematic as a descriptive theory of the law.

Furthermore, as many commentators have pointed out, the actual amount of transaction costs associated with many transfers of rights changes over time.40 An important factor that depresses transaction costs is technology. The ascent of the Internet, in particular, makes it possible for copyright holders to communicate to the public at large (by means of a website posting, for instance) elaborate menus of prices for various uses of their expressive goods and thereby to obviate the need for extensive negotiations concerning the use of intellectual goods.41 This phenomenon was noted, for example, by Professor Robert Merges, who questioned the need for limitations on intellectual property rights in an online world without “friction.”42 Taking this point to its logical extreme, Professor Tom Bell argued in a provocative article that, in the Internet age, “fair use” should be replaced with “fared use,” suggesting that no unauthorized uses of intellectual works should be free any more.43 Indeed, any theory of fair use that is based solely on market failure due to transaction costs necessarily suggests an ever-diminishing fair use doctrine in an age in which many kinds of transaction costs continue to decline. Gordon’s own discomfort with this conclusion in later writings44 demonstrates that she, like others, believes that there is more to fair use than simply a strategy for dealing with the market failures created by transaction costs.

It should be noted that technology does not eliminate all relevant transaction costs. In support of Gordon’s theory, Professors Ben Depoorter and Francesco Parisi have pointed out that even with declining transaction costs, the fair use doctrine still plays an important role under the market-failure theory in situations involving transactions with multiple rights holders.45 In such settings, each right holder may strategically hold out and deny an aspiring user the power to use her work in the hope of extracting the full value of the user’s project. Fair use mitigates this problem by setting a cap on the ability of rights holders to withhold consent strategically.46 While Depoorter and Parisi’s point is both elegant and well-taken, it potentially limits the usefulness of the market-failure theory to only cases in which a user must clear multiple rights and has no other alternatives.

Finally, as Professor Matthew Sag points out, fair use is highly important in the context of promotion of “copy-reliant” technological innovations. When analyzing the impact of transaction costs on such innovations, it is immaterial that the transaction costs associated with the making of individual copies are considerably lowered because of technological progress. In such instances, we need to factor in “the total volume of transaction costs faced by copy-reliant technologies,” not “the costs attending any one transaction.”47 Moreover, such technologies “typically rely on . . . default rules and . . . opt-out mechanisms to reduce transaction costs,”48 these default rules in turn depending on the fair use doctrine for their legal validity or acceptance as industry practice.

It is noteworthy that Gordon has often pointed out that her earlier work was misunderstood.49 In a later article written with Daniel Bahls, the authors explained that market failure resulting from relatively high transaction costs constitutes a prima facie reason to recognize fair use, but they by no means exhausted all the cases in which fair use ought to be considered. Rather, fair use should also be available in cases of

patterns of creative production that are not consistent with bureaucratic behaviors; anticommons, hold-out and bilateral monopoly problems; distributional inequities; positive externalities; use of another’s work not as expression but as a fact; use of another’s expression as a means to access the public domain; and critical, nonmonetizable and/or “priceless” uses of copyrighted works.50

Gordon’s expanded framework addresses many of the problems that inhere in the market-failure model, but it does so at the cost of a coherent account that is guided by a single principle. We view the clarification she offered in her article with Bahls as an invitation to supplement the market-failure theory of fair use with a broader theoretical conceptualization that would move the discussion in the direction Gordon and Bahls pointed out. It is precisely this call that we answer in the next Part.

B.    Recent Trends in Fair Use Case Law

One of the many strengths of Gordon’s article when it was published was the fit between the market-failure theory and the case law as it existed at the time. While judges did not expressly refer to market failure in fair use decisions, Gordon’s theory had strong explanatory power. The case that Gordon used as the primary illustration of her theory was Universal City Studios, Inc v Sony Corp of America.51 The legal issue was whether Sony, the manufacturer and distributor of Betamax, the first videocassette recorder (VCR), was secondarily liable for copyright infringement by VCR consumers. In particular, the question was whether Sony should be held to have contributed to the unauthorized recording of copyrighted television programs by home users of its technology.52 Gordon wrote in opposition to the decision of the Ninth Circuit to hold Sony liable for its contribution to consumer infringements.53 Gordon argued that Sony should not be found liable because the home uses of the VCR should be seen as fair uses due to market failure.54

Following the publication of Gordon’s article, the Supreme Court in Sony Corp of America v Universal City Studios, Inc55  reversed the Ninth Circuit’s holding, declaring that Sony did not secondarily infringe. In holding for Sony, the Supreme Court ruled, inter alia, that the practice of recording programs for later viewing—a practice it labeled “time shifting”—was fair use.56 The Sup­reme Court’s decision vindicated Gordon’s framework of analysis. After all, it is unrealistic to expect technology providers, such as Sony, to secure permission prior to putting their technology on the market from the thousands upon thousands of owners of rights in the numerous copyrighted television works. The large number of rights holders involved makes reliance on voluntary transactions impractical. Not only would the large number of rights holders require Sony to expend considerable resources on identifying the relevant counterparties and carrying out negotiations with them, but it would also mean that each of the counterparties could strategically hold Sony up by refusing to tender permission in the hope of extracting an especially high payment from Sony.

Sony, it turned out, was the high-water mark for the market-failure theory. Recent years have seen the case law moving away from the market-failure theory. First, in Campbell v Acuff-Rose Music, Inc,57 the Supreme Court had to decide whether an unauthorized rap version of the song “Oh, Pretty Woman” came under the auspices of fair use.58 In a pathbreaking decision, the Court ruled that it did, explaining that the rap version may be reasonably perceived as a parody of the original.59 As importantly, in a sharp deviation from prior decisions, the Court elevated transformativeness above all other fair use factors.60 Following Campbell, recent fair use decisions appear to focus on the transformativeness of the defendants’ works, not transaction costs, as the key factor in fair use cases.61 This trend is especially strong in the Second Circuit and can be seen in that court’s decisions in Blanch v Koons62 and Cariou v Prince.63 In Blanch, the Second Circuit was required to decide whether the defendant’s painting, which incorporated the plaintiff’s photograph without her permission, qualified as fair use.64 The court ruled that it was fair use, even though the defendant’s work was commercial and did not parody the plaintiff’s work.65 In this case, there were only two parties involved and no market failure seemed to have been present. The court, however, did not focus on these factors. Rather, it concluded that in its opinion “copyright law’s goal of ‘promoting the Progress of Science and useful Arts’ . . . would be better served by allowing Koons’s use of ‘Silk Sandals’ than by preventing it.”66

The Second Circuit’s decision in Cariou was even more extreme. In that case, the defendant, Richard Prince, was an “appropriation artist” who lifted many of the plaintiff’s photographs of the Rastafarian people of Jamaica and transformed them into his own artistic style by, “among other things, painting ‘lozenges’ over their subjects’ facial features and using only portions of some of the images.”67 Prince did not even attempt to negotiate a license from Patrick Cariou before using the images.68 Worse yet, Prince did not bother to offer a justification for his actions. When asked whether he intended his adaptations to be transformative, he answered that “he ‘[didn’t] really have a message’” and that “he was not ‘trying to create anything with a new meaning or a new message.’”69 Nonetheless, the court explained that

[w]hat is critical is how the work in question appears to the reasonable observer, not simply what an artist might say about a particular piece or body of work. Prince’s work could be transformative even without commenting on Cariou’s work or on culture, and even without Prince’s stated intention to do so.70

The court proceeded to find fair use with respect to twenty-five of Prince’s works that used Cariou’s images.71 From a pure market-failure perspective, there was little reason to find fair use. The defendant in this case simply bypassed the market, notwithstanding the fact that the transaction costs in this case seemed very low relative to the value of the uses. Indeed, with respect to the five images whose use was not ruled fair by the Second Circuit, the parties settled.72

No less importantly, in two recent decisions, Eldred v Ashcroft73 and Golan v Holder,74 the Supreme Court conceptualized fair use not as an anomaly or an exception, but as an intrinsic and indispensable part of the design of the constitutional copyright system. Eldred involved a facial challenge to the Sonny Bono Copyright Term Extension Act,75 in which Congress extended the term of protection by twenty additional years.76 The legislation offered the extension not only prospectively to works that were yet to be created, but also retroactively to works that had already come into existence under the previous, shorter protection term.77 A group of publishers whose business model was predicated on public domain materials argued that the extension was unconstitutional because it exceeded Congress’s power under both the Copyright Clause and the First Amendment.78 The Court ultimately rejected the challenge and approved the extension, explaining that the extension came under the aegis of the Copyright Clause and that it did not violate the First Amendment.79 In this vein, the Court stated that “copyright law contains built-in First Amendment accommodations.”80 Among those accommodations, the Court listed the idea-expression dichotomy, under which copyright protection extends to original expression but is withheld from the ideas underlying it, as well as the fair use doctrine, which “affords considerable ‘latitude for scholarship and comment.’”81

The Court’s language in Golan is even more instructive. In Golan, the lead petitioner, a conductor and music professor, challenged the United States’ implementation of § 514 of the Uruguay Round Agreement Act, now found in § 104A of the Copyright Act.82 Section 104A reinstates copyright protection for multiple foreign works that had fallen into the public domain for failing to comply with various formalities.83 Lawrence Golan argued that the legislation violated his free speech rights by impoverishing the public domain.84 Once again, the Supreme Court dismissed the challenge. Reiterating the reasoning in Eldred, the Court stated that the “traditional contours” of the copyright system are delineated by the idea-expression dichotomy and fair use.85 The Court proceeded to explain that legislative changes limiting the availability of copyrighted works to the public will pass judicial muster as long as they do not affect the idea-expression dichotomy or curtail fair use.86 As Professor Neil Netanel has pointed out, one clear implication of the Supreme Court’s ruling is that the First Amendment forbids legislation that would significantly restrict fair use.87

II.  A New Theory of Fair Use

In this Part, we present our alternative approach to the fair use doctrine. We argue that our approach better fits both the economic imperatives of copyright law and the recent case law on the subject.

Perhaps the easiest way to understand our approach is by contrasting it to the market-failure approach. In some ways, ours is the polar opposite of the market-failure approach. The market-failure theory assumes that all rights associated with copyrighted works should be presumed to lie with the owner of the copyright. In this view, fair use is an exception that allows transfers of those rights to users on a transaction-by-transaction basis, and only when a particular transaction is likely to be blocked by market failure.88

Our theory, by contrast, assumes that copyright law cons­ciously allocates to users many privileged uses, irrespective of the desires of the author. The law divides the rights associated with copyrighted works between author and public. We assume that certain privileges associated with copyrighted works should be presumed to lie with the general public and that those privileges should be denied only in extreme cases. As we show, our proposed version of the doctrine would have courts recognize a contested use as “fair” whenever two conditions are met: First, the use is within the scope of several presumptively public uses, such as political speech or research. These presumptively public uses can be identified by category and are characterized by the creation of significant value for follow-on users. Second, granting the use the status of “fair use”—or, more precisely, granting the status of fair use to the class of uses that would be encompassed both by the user’s proposed fair use and by similar uses by similarly situated users—would not entirely eliminate the incentive to create this particular work. The second condition largely overlaps with one of the elements of Professor Gordon’s proposal,89 with some differences that we highlight below.90 We acknowledge that it also shares one of the imperfections of the similar element in Gordon’s account, namely, that the determination will not be mathematical—though, as we show,91 it should produce clearer results than Gordon’s proposal. In particular, we show that our approach has the advantage of focusing on the character of the use rather than the character of the user. Thus, we are able to avoid individualized assessments of users in every fair use case.

At the outset, however, we must state that while our analysis differs in significant respects from Gordon’s original market-failure theory, doctrinally our proposal should be seen as a cumulative addition to market-failure analysis, rather than a repudiation of it. While our theory of fair use leads us to embrace a new doctrine for protecting uses that would not be protected under the market-failure theory, our theory is not incompatible with the idea that uses that would be subject to market failure due to relatively high transaction costs should be treated as fair uses in addition to the other uses that we identify as presumptively fair uses. Likewise, we emphasize that our goal in using the market-failure theory as a backdrop against which we portray our theory is not to discredit Gordon’s many invaluable contributions to the theory of fair use. On the contrary, it is due to the significance of her illuminating and multifaceted scholarship in this area that her theories provide the starting point for our discussion.92

A.    The Economics of Copyright

The point of departure for our theory of fair use is an incentives-based or economic understanding of copyright. There is little doubt that the foundations of American copyright law, as found in the Constitution, are economically based and aimed at incentivizing authorship.93 The Copyright Clause tells us that Congress should grant exclusive rights in “Writings” to “Authors” in order to “promote the Progress of Science.”94 Copyright law aims to incentivize the creation of expressive works by granting authors exclusive rights in their creations—enough rights to make creation pecuniarily worthwhile for authors while not unduly removing expressions from the public domain.95

Professor William Landes and Judge Richard Posner’s classic article on the economics of copyright law explains the incentive structure created by copyright law.96 There are two outstanding features of expressions that must be taken into account by lawmakers in establishing the law of copyright. First, expressions, like other items that are encompassed within the world of intellectual property, have features of what the economic literature calls “public goods.”97 This means that the market alone will not produce the goods and that some legal incentives are necessary to prompt the market to produce them. Second, the legal incentives created by copyright law are themselves potentially market distorting.98 Like all intellectual property law, copyright law can create monopolies or other anticompetitive features.99 Copyright law thus eternally seeks to balance the beneficial and harmful aspects of the rights that it grants to authors.100 Unfortunately, the very same tools used by lawmakers to create and encourage copyright markets are the ones that distort those markets.

There is a vast literature on public goods and the reasons why ordinary markets will not produce them. Public goods are items whose consumption is nonrivalrous and nonexcludable—that is, all users can use the items without diminishing the ability of others to consume them, and there is no way to block potential users from consuming the items.101 Pure public goods cannot be sold in the market effectively and therefore cannot bring returns to their producers. Consequently, there is no reason a private producer will ever decide to manufacture public goods.102 Expressions, or more generally information, are in many ways the quintessential public goods.103 Information can be consumed by any number of people, and the knowledge of one does not in any way diminish the knowledge of another. In the ordinary course of events, once information is revealed, producers have no practical ability to block further transmission of the information to others. While there are various nonpecuniary benefits that a producer may get from creating expressions or information, the absence of markets would mean that in a world without intellectual property rights, fewer expressions would be created because there would be fewer incentives to create such expressions.

Copyright law supplies the missing incentives by giving creators a package of exclusive legal rights in expressions for an extended period of time. The legal protection makes the expressions excludable. That is, once copyright law exists, the producer can police the use of the expressions and prevent nonpaying users from consuming the good.104 Legal protection thus significantly expands the possibilities for markets in expressions and thereby incentivizes creativity.

However, the market created by the legal protection is itself a distorted market. During the period of exclusive protection, the copyright owner enjoys monopoly rights over the expression.105 Because expressions often do not have perfect substitutes, the market for expressions may often display the characteristic flaws of monopolistic markets—excessively high prices and subcompetitive supply of the good.106 Legal protections for copyright, in other words, impose a cost on society at the same time as they produce the societal benefit of incentivizing creation. As a result, noted Landes and Posner, while some copyright protection is beneficial for society, “beyond some level copyright protection may actually be counterproductive.”107 Indeed, since all copyrighted expressions invariably rely to some degree on earlier works, excessive copyright protection can directly restrict the cost and availability of future copyrighted works.108 Copyright law must therefore always seek to find the balance that maximizes the societal benefit of incentivizing creation net the cost of restricting the availability of protected expressions. Here, Landes and Posner, like the scholars that have followed them, frankly acknowledge that they have no formula for determining the precise amount and type of protection that would be optimal.109

B.    The Allocation of Rights

The exclusive rights granted to an author are significant and designed to give the author ample opportunity for pecuniary benefit. In general, copyright law specifies that, for a fixed period of time,110 only the author is permitted to copy, adapt, display, perform, or distribute a copyrighted work.111 This means that during this period, most potential users (other than those enjoying the benefit of a legal exemption or license)112 must receive permission from the author. The author is expected to sell such permission and thereby realize the pecuniary rewards of creation.

As we have written elsewhere, it is highly doubtful that the entire package of exclusive rights is necessary to incentivize the amount of creation of expressive works we see today.113 For most copyrighted works, the authors are inframarginal; they would have created the works even if they had been granted only a much smaller selection of rights. More broadly, the vast expansion of authors’ rights in recent decades has led to widespread criticism that compensation of all authors is excessive.114

How can lawmakers trim the package of authors’ rights to give only enough rights to incentivize creation? There are many ways to adjust the optimal balance of copyright rights so as to attempt to maximize social benefits. The most straightforward—but far from the best—means of adjusting is to change the temporal margins of protective terms or rights. For instance, lawmakers can shorten or lengthen the term of protection for copyrighted works from its current seventy years plus the life of the author115 to fifty or ninety years plus the life of the author. Term adjustments straightforwardly subtract or add a certain quantum of incentives (the expected pecuniary benefits to be realized by the author during that future period of time) at the same time as they expand or restrict the public’s ability to use the work by changing the point in time when the work will enter the public domain. Yet, while it is relatively easy for the state to administer adjustments of the term of protection, such term adjustments are blunt instruments. During the legal term of protection, whatever it might be, authors enjoy the usual monopoly protection for their expressions, meaning that even if the term adjustments are perfectly tailored to provide socially desired incentives for creation, those incentives might be provided at a higher social cost than necessary. The law could have achieved the same level of incentives by granting fewer rights for a longer period of time. Or it could have given the same rights for a longer period of time, but expanded the number of privileged uses.

Our approach to the fair use doctrine reads the Copyright Act as aiming to adjust the amount of incentives granted to authors by setting aside for the public a number of privileged uses. We argue that the fair use doctrine seeks to balance copyright through a conscious division of legal entitlements between authors and users. In our theory, instead of dividing the rights along the axis of time or by the type of exclusive activity granted to the author (reproduction, adaptation, distribution, etc.), the fair use doctrine is intended to divide legal protections between author and public according to the nature of the intended use. The purpose in allocating entitlements this way is to allocate to the public certain privileged uses that are likely to be of significant public value, while reserving for authors sufficient valuable rights to maintain incentives to create.

In allocating certain privileges to the nonowner users, fair use doctrine consciously departs from the usual practice in property law, and it is important to explain why. Specifically, we must show why it would not be better to just give the copyright owner all the rights and allow her to sell to consumers the uses that they want.

To understand the best way to divide these legal protections, it is worth taking a second look at the problem of allocating property rights. Consider a world in which lawmakers know the optimal users of all uses of assets, the appetite for such uses is fixed and stable over time, and allocation among users is costless. In this idealized world, the optimal approach would be to allocate directly to each potential user the use right in the assets, even if this meant dividing up the use rights for a single asset among hundreds of users. The allocation would be free, final, and optimal. High-value users would have all the use rights they needed in assets. Society would avoid the costs of providing mechanisms for efficient allocation. While subsequent transactions in the asset might well be difficult and costly, they would also be unnecessary because the high-value user of each use would already own what she needs.

Of course, the imaginary world of perfect initial allocations does not exist. Appetites for uses are not fixed and stable, and even if they were, information about appetites for uses is not freely available. Initial allocations are expensive. It is generally impossible for a central planner to identify and measure tastes anywhere near as efficiently as a marketplace can. In the real world, it is difficult to find the circumstances in which a central planner dividing uses among many claimants produces net societal gain.116

However, the point made by the example is an important one: a good initial allocation is of great social benefit. If the allocation is done well, it can increase societal welfare by reducing the need to transfer rights, thereby saving transaction costs. Indeed, a good initial allocation can be of such high social benefit in some cases that it can compensate for subsequent societal losses created by legal regimes that make transactions more difficult (and by the attendant higher cost of those transactions that must take place).

This observation is doubly true when it comes to copyright. Because copyright uses are generally nonrivalrous—that is, the use of the work by one consumer does not in any way diminish the ability of a different consumer to use it—the cost of misallocation is low. Granting a use right to the wrong consumer does not require a new (and costly) transaction; it simply eats away from the incentives of the author. Misallocation is not costless. But it is not nearly as costly as with ordinary property.

Our understanding of the fair use doctrine in copyright law is informed in large part by the allocative question in light of the need for limited rights. The optimal grant of copyright rights would not give the owner all rights to the copyrighted work in perpetuity. Rather, the optimal grant would give the owner exclusive authority over the work as necessary to incentivize creation. At the same time, the fair use doctrine necessarily bestows on consumers of the expression exclusively uses of positive value. Compare the fair use doctrine, for example, to a rule that curbs owners’ terms of protection by ten years. The term limitation grants additional privileges to the public, but there may be no audience for such privileges. It may be that there is no market for using a given work in its 115th year of protection or, for that matter, even in its 15th year. Indeed, the value of most expressive works disappears long before the protection term runs out.117 The fair use doctrine, under any interpretation, applies only to uses that at least one consumer desires. If there is no actual use, there can be no fair use. A fair use doctrine allocates directly to users privileges that they want and value.

In this sense, there is no better way to achieve division of legal protections between the public and the copyright owner than by granting use privileges via such doctrines as fair use. Such doctrines necessarily grant to the public something valuable and only something valuable. This point deserves emphasis. In general, the main societal loss stemming from misallocation of an asset is underuse of the asset. If the state allocates an asset to a low-value user and transaction costs are excessive, the asset will likely not be used by some or all high-value users. In other words, some high-value uses will never take place. But this cost of misallocation does not appear when the state grants use privileges and uses are nonrivalrous. In such cases, users will, by necessity, exercise their use privileges only when the value of the use is positive for them and for society as a whole. Even misallocated use privileges will not lead to underuse. By contrast, allocating legal protection directly to the user is valuable to society, because it saves the need to transact and the cost that comes with it.

Unfortunately, use allocations are not flawless, even in the world of copyright. There are two chief pitfalls that use allocations must avoid to be societally beneficial. First, the use allocation must leave authors with sufficiently valuable rights. The authors must retain the right to veto at least some transactions to profit from the expression. This profit, indeed, is the sine qua non of copyright law, for it must generate the incentives for authors to create. Thus, the public use (or fair use) privilege cannot be so large as to eliminate all profit and thereby destroy the incentives to create.

Second, adjudicating use divisions is not costless. Use divisions are not easily administered, and judicial resources in resol­ving disputes over the use privilege may be significant. For a use allocation to be efficient, the use privilege must not be so difficult to administer as to dissipate all the ex ante transaction cost savings in ex post judicial administration costs.

In the next Part, we examine how a doctrine of fair use can achieve a superior allocation of rights without eliminating the incentives to create and without creating excessive judicial administrative costs.

III.  A New Doctrine of Fair Use

In this Part, we argue that the best way to implement the division of legal protections created by the dual-grant theory of fair use is to allocate uses to ultimate users in discrete categories that are relatively easy to identify, that do not require excessive amounts of private information, that provide some identifiable widespread value across the public (that is not exclusive to the actual user of the expression), and that are not so great as to cross the margin at which the author loses incentives to create. Fair use doctrine is the mechanism by which the law allocates directly to users those uses with the greatest net societal benefit from direct allocation and removes them from the grant of exclusive rights to the author.

Drawing on this approach, we present the actual doctrine of fair use we propose, drawing from the theory we expounded in the previous Part. We present both an idealized doctrine and a more realistic one, compare them to existing doctrine, and illustrate our proposed doctrine with several examples. Focusing on the mechanics of our proposal, we pay particular attention to the way uses might be categorized as presumptively fair. We begin by contrasting our proposal with extant fair use doctrine.

A.    The Extant Doctrine of Fair Use

Fair use is one of the most widely litigated and amorphous doctrines in the law118 and has been part of US copyright law at least since 1841119 (and much longer than that, according to some scholars120 ). Its modern form is codified in § 107 of the Copyright Act. Section 107 does not give clear guidance on when uses of a copyrighted work should be considered “fair” and therefore legally permitted without license from the owner. Rather, § 107 gives courts a list of illustrative purposes of fair use—“criticism, comment, news reporting, teaching (including multiple copies for classroom use), scholarship, or research”—and four factors to consider in determining whether the use is fair: (1) “the purpose and character of the [allegedly fair] use”; (2) “the nature of the copyrighted work;” (3) “the amount and substantiality of the portion” of the copyrighted work used by the allegedly fair use; and (4) “the effect of the use upon the potential market for or value of the co­pyrighted work.”121

In a classic article, Judge Pierre Leval highlighted the shortcomings of both the statutory formulation and the cases that implement it:

What is most curious about this doctrine is that neither the decisions that have applied it for nearly 300 years, nor its eventual statutory formulation, undertook to define or explain its contours or objectives. . . . [They] furnish little guidance on how to recognize fair use. The statute, for example, directs us to examine the “purpose and character” of the secondary use as well as “the nature of the copyrighted work.” Beyond stating a preference for the critical, educational, and nonprofit over the commercial, the statute tells little about what to look for in the “purpose and character” of the secondary use. It gives no clues at all regarding the significance of “the nature of” the copyrighted work. Although it instructs us to be concerned with the quantity and importance of the materials taken and with the effect of the use on the potential for copyright profits, it provides no guidance for distinguishing between acceptable and excessive levels. Finally, alth­ough leaving open the possibility that other factors may bear on the question, the statute identifies none.

Curiously, judges generally have neither complained of the absence of guidance, nor made substantial efforts to fill the void. Uttering confident conclusions as to whether the particular taking was or was not a fair use, courts have treated the definition of the doctrine as assumed common ground.

The assumption of common ground is mistaken. Judges do not share a consensus on the meaning of fair use. . . . The opinions reflect widely differing notions of the meaning of fair use. Decisions are not governed by consistent principles, but seem rather to result from intuitive reactions to individual fact patterns. Justification is sought in notions of fairness, often more responsive to the concerns of private property than to the objectives of copyright.122

Case law tends to focus on the four factors in determining whether a use is fair, but decisions are far from uniform. Recent years have seen courts paying growing attention to whether the allegedly fair use is transformative; transformativeness has thus become a key element in all four factors.123 However, other features of the cases, such as market analysis, remain important, making outcomes difficult to predict.124

One outstanding feature of current case law is the degree to which the four factors are pliable. Consider, for instance, the treatment of the third factor (the amount and substantiality of the portion of the copyrighted work used by the allegedly fair use). Some courts have found that using the entirety of the copyrighted work is no great hindrance to a finding of fair use,125 while others have found the use of much smaller amounts of the original work to be critical to the denial of a finding of fair use.126 In most cases, courts analyze the proportion of the original work that has been used127 —but there have been cases in which courts have looked instead at the proportion of the new, allegedly fair work that was copied from the original.128 While hard guidelines on proportions have been suggested129 and even incorporated in nonbinding guidelines,130 many cases suggest that the very idea of hard guidelines is an affront to the nature of fair use.131

Recent years have seen some refinement of the thesis that fair use is in doctrinal disarray. In a study of fair use opinions from the years 1978–2005, Professor Barton Beebe found that judicial applications of the doctrine were surprisingly stable over time and even displayed some resistance to changes in doctrine announced by the Supreme Court.132 However, Beebe also found some manipulation of the four statutory factors, as well as some variance in the way the factors were interpreted.133 Using these findings, Professor Pamela Samuelson emphasized the degree to which extant fair use law is clustered around what she called “policy-relevant” patterns.134 Essentially, Samuelson argued that much of the apparent disarray of existing fair use doctrine disappears if we note that courts’ approaches to fair use cases depend almost entirely on the category to which the allegedly “fair” use belongs. Unfortunately, the picture of policy-relevant categories is not entirely clean. Samuelson identified a large number of policy categories—“promoting freedom of speech and of expression, the ongoing progress of authorship, learning, access to information, truth telling or truth seeking, competition, technological innovation, and privacy and autonomy interests of users”—some originating in the language of § 107 and others originating in case law.135

B.    An Idealized New Doctrine of Fair Use

The pliability surrounding the working of the fair use doct­rine necessitates clear earmarking of permitted uses, in line with our proposal of dual grants. Legislators can implement the dual-grant theory in two ideal ways: author-centered and user-centered. Unfortunately, neither of these ideal approaches is workable in reality, leaving legislators to choose an imperfect compromise.

An author-centered approach would explicitly key fair use to the quantity and quality of the grants of rights to authors. Any expansion of author rights would automatically lead to an expansion of fair use. If, for instance, Congress were to grant an additional twenty years of copyright protection to owners—as it did in the Sonny Bono Copyright Term Extension Act136 —courts would automatically expand fair use to recapture uses for the public and to restore the balance of the dual grant. The scope of fair use would move up and down in lockstep with changes in the grant of rights to authors.

A user-centered approach, by contrast, would grant large chunks of presumably fair uses directly to the public. These grants of uses and changes in the scope of fair use would, in turn, affect the rights of authors. An expansion of fair use would expand authors’ rights, and a contraction of fair use would shrink authors’ rights.

Unfortunately, both ideal approaches suffer from a huge flaw—they are very expensive to administer. Both ideals demand constant vigilance. The world of copyright is already in continuous flux due to technological innovations, market developments, and legal changes. To match user privileges to owner rights (or vice versa), courts would have to monitor all these changes and update standards on a regular basis.

The law of copyright thus adopts a compromise. Both user privileges and author rights are defined with respect to uses rather than by comparison with one another. The rights granted to authors are defined in terms of specific acts (such as copying or performing) for specified times. The privileges granted to users are defined more vaguely, though, we argue, no less broadly.

Although user privileges and author rights are defined separately by statute,137 their interrelationship ought to be borne in mind by courts. Specifically, courts must be sensitive to the breadth of author rights when they define uses as fair or not. An expansion in author rights should be accompanied by a corresponding judicial openness to new claims of fair use.

C.    A Proposed New Doctrine of Fair Use

Since it is impracticable to create a fair use standard that is precisely matched to author rights, we propose a simplified set of fair use rules that make fair use easier to apply and more oriented toward achieving allocative efficiency, even if the doctrine is incapable of precisely measuring the set of rights necessary for author incentives. These fair use rules could be applied by judges within the existing fair use framework, given its pliability.

1.   Two steps and four factors.

Our proposed doctrine cuts through the confusion by refocusing courts on two basic questions. Under our proposal, courts would evaluate fair use claims through a two-step process. In the first step, the fair use claimant (the alleged infringer) would est­ablish a prima facie right to use the fair use defense by showing that the use falls within certain privileged categories—political speech, truth seeking, or any other category of uses that produces widespread nonpecuniary follow-on benefits for nonusers. In the second step, the owner of the copyright would have the opportunity to rebut the fair use defense by showing that granting the fair use claim in this and similar cases would have the effect of denying the minimal incentives necessary for creation of this work.

Importantly, our proposal would easily accommodate the four statutory factors that § 107 of the Copyright Act indicates should guide courts in their decisions. In our proposed doctrine, two of the factors would be examined to determine whether the use was prima facie privileged, while the other two would help the court figure out whether recognizing the use as fair would eliminate the incentives for creation.

Part one of a court’s fair use determination, in our proposed doctrine, would require the court to decide whether the use is within a privileged category by evaluating the purpose and character of the allegedly fair use (factor one). The court would also have to pay attention to “the amount and substantiality of the portion”138 of the copyrighted work used by the allegedly fair use (factor three) to determine whether the claimed use actually matched reality. Consider, for instance, a decision by the political campaign of Harry to distribute for free the autobiography of his opponent Bernice, in order to criticize Bernice’s onetime support for socialism. It is clear that this is a political use, which is a use that is prima facie privileged in our proposed doctrine. However, Bernice might convincingly argue that Harry could have achieved his proposed political use by distributing portions of the autobio­graphy rather than the whole work. A court convinced of the merits of Bernice’s claim would determine that Harry had not established the prima facie validity of use, since the distribution of the entire work could not be reconciled with the privileged purpose.

Part two of a court’s fair use determination, in our proposed doctrine, would involve an assessment of whether recognizing the proposed use as fair would, in the context of the work, eliminate sufficient incentives for creation. Naturally, this examination would have to consider the effect of the proposed fair use “upon the potential market for or value of the copyrighted work”139 (factor four), and the extent to which nonmarket factors might sufficiently motivate the creation of this kind of work (factor two). However, the court would examine these elements to determine only whether the proposed fair use is so extreme as to eliminate the incentives to create the work. In making this determination, the court would consider not only the particular use engaged in by the defendant, but also the effect on incentives to create the work if similar users with similar uses could likewise enjoy the protection of the fair use doctrine.140

Consider again Harry’s use of Bernice’s autobiography. A court would have to examine how Harry’s distribution of parts of Bernice’s autobiography, and like distribution by other candidates or political opponents of Bernice, would affect the market for Bernice’s book. However, finding a significant adverse effect on the market would not constitute sufficient reason to defeat a claim of fair use. Rather, the court would have to determine that if all Bernice’s opponents were to distribute portions of the autobiography without license, Bernice would lack sufficient incentive to publish the autobiography in the first place. Here, the nature of the work would be particularly important, because a court might readily determine that the incentives for production of a political autobiography are not primarily pecuniary.

2.   Privileged uses.

The key to our proposed fair use doctrine is a list of privileged uses that are presumptively fair. A court’s determination that a use is presumptively fair greatly increases the likelihood that the defendant will succeed in a fair use defense; it will not be easy for a copyright owner to show that recognizing the fair use claim would suffice to eliminate incentives to create the work.

What types of uses should be recognized as presumptively fair in our proposed doctrine? As we noted in Part II.B, fair use doctrine should aim to privilege uses in discrete categories that are relatively easy to identify and that provide some identifiable, widespread value across the public that is not exclusive to the actual user of the expression. Our aim is to identify those uses that create utility not only for the direct user of the expression but also for many others. Stated differently, we seek uses of expressive works in which the uses themselves create identifiable and widely spread positive externalities. The list of privileged uses must take account of the needs of administrable doctrine as well. The nature of the use that is privileged must not rely on private information, for example.

In this Section, we propose several obvious candidates for the list of privileged uses and also explore several categories of uses that are already important in current fair use litigation. Our theory does not always direct courts to find fair use. In fact, it calls into question many aspects of fair use jurisprudence, as it currently exists. At the same time, our theory directs courts to find fair use in many instances that would not be recognized as such under extant case law. Together, these categories demonstrate how our approach requires us to rethink some foundational notions that have shaped fair use law.

a) Political speech.  The most obvious example of a privileged category for fair use doctrine is political speech. Political speech is already conferred an exalted status in First Amendment doctrine141 and for very important reasons. Political speech is an integral ingredient in self-governance, and therefore it is critical to the functioning of democracy.142 From an economic point of view, the social value of political speech is somewhat difficult to describe. Professor Robert Cooter suggests that political speech should be seen as a kind of “merit good”;143 Professor Richard Musgrave controversially developed the term “merit goods” to describe goods that individuals in society “need” notwithstanding their unwillingness to pay.144 Specifically, Cooter argues that free speech enhances societal liberty (and therefore utility) in a manner that is greater than the public’s preference for consuming the good.145 But it is not necessary for us to endorse this controversial theory in order to establish that political speech is a valuable category for prima facie protection under fair use doctrine. It is enough for us to note that political uses of speech are thought to promote self-governance in a democratic society, and thus create widespread nonpecuniary follow-on benefits throughout society. These benefits accrue not only to the political speaker but to many others that do not directly use the expression. Political speech thus falls directly within our definition of categories for prima facie protection.

Controversies regarding fair use and the political use of speech have arisen in the past. One journalist has gone so far as to say, “If you’re a candidate for president, there’s a good chance you’ve been sued—or at least threatened—by a musician . . . [for] using popular songs at [your] campaign rallies.”146 The American Society of Composers, Authors, and Publishers (ASCAP) has published guidelines urging politicians to purchase licenses to use music in their campaign rallies; the guidelines do not reference the possibility of a fair use claim.147

Our theory suggests that all political uses of expressions should be presumptive fair uses. This does not mean that political speech should be unprotected by copyright law; the fair use doctrine protects uses, rather than entire expressions. Thus, a commercial use of a political tract by Newt Gingrich would infringe Gingrich’s copyright, while a political use of a Fleetwood Mac song would presumptively be a fair use even if the song had nothing whatsoever to do with political affairs.

Our approach would constitute a departure from extant jur­isprudence. This is best seen by reexamining one of the Supreme Court’s most important pronouncements on the scope of fair use—its ruling in Harper & Row, Publishers, Inc v Nation Enterprises.148 Harper & Row involved the publication of a news story by The Nation that included between three and four hundred words of unlicensed quotations from former President Gerald Ford’s A Time to Heal: The Autobiography of Gerald R. Ford. Drawing directly from the autobiography, the story in The Nation described Ford’s perspective on his decision to pardon former President Richard Nixon of all crimes related to the Watergate affair.149 The Court acknowledged that The Nation was engaging in news reporting and that First Amendment concerns came into play, particularly since Ford was a public figure and the speech concerned politics.150 Nevertheless, the Supreme Court dismissed The Nation’s claim of fair use and found that The Nation had unlawfully infringed on Harper & Row’s copyright in the Ford autobiography by copying text without a license.151

Two factors appeared to weigh heavily on the Court in its considerations. First, Harper & Row had sold to Time for $25,000 the right to publish an excerpt of the autobiography prior to publication of the book as a whole. The 7,500-word excerpt in Time would have concerned Ford’s pardon of Nixon. However, after The Nation’s “scoop,” Time cancelled the contract and refused to make payment on the outstanding $12,500, because there was little news value to the excerpt after The Nation story appeared.152 The Court viewed this as clear evidence of the fact that the claimed fair use had imposed a significant cost on the market for the work.153

Second, Harper & Row had not yet published the full autobiography. Rather, the autobiography’s first appearance was in the quotations within The Nation’s news story, violating what the Court labeled the “author’s control of first public distribution.”154 While the Court acknowledged that “first publication” by an unlicensed user should not be an absolute bar to a claim of fair use,155 the Court also noted that fair use was traditionally understood to be based on the author’s implied consent to a “reasonable and customary use,” and it was unlikely that authors would consent to giving a first-publication license to an uncoordinated news “scoop.”156 The Court thus suggested that the use was presumptively unfair.

The Court’s approach in Harper & Row has been criticized for its miserliness toward fair use claims.157 In particular, critics have asked whether three to four hundred words should truly be considered a significant portion of a full-length autobiography.158 By refusing The Nation’s claim to fair use, the Court found the market effects of unlicensed copying of an extremely small amount of the copyrighted work to be significant enough to warrant upholding the rights of the author.

Under our approach, the Court would have to uphold the claim of fair use. Given The Nation’s use of the quotations for political speech, it would be unnecessary for the Court to engage in any extended analysis of the length of the quotations or the effect on prepublication contracts. The political use of the quotations would make The Nation’s use prima facie fair. As indicated by the very small portion of the original work used, it would not be possible to argue that The Nation had an ulterior motive in quoting the original. Indeed, The Nation might have directly excerpted far larger portions of the original and could still have credibly argued that its aim was to engage in political speech. Thus, the Court would be left to ask only whether the loss of prepublication contracts (and the attendant diminution of the advance that would be offered by Harper & Row) would be sufficient to defeat the incentives for Ford to write his autobiography. The near-certain answer would be “no.” Ford would certainly want to write his autobiography, even if Harper & Row had to reduce his advance fee by $25,000 or more.159 This is due not only to the continued expected market for the full autobiography but also to the presumed interest of a retiring president in presenting his own version of his biography to the public.

b) Truth seeking.  A second example of a privileged category for fair use doctrine is truth seeking. What we mean by “truth seeking” is uses of copyrighted expressions that aim at establishing or identifying the truth of a matter. This category would encompass not only news reporting but also, and perhaps more importantly, research and teaching.

Research and news reporting are not only explicitly mentioned in § 107 as two of the types of uses at which fair use is aimed,160 but they are also fairly well implied by the constitutional copyright scheme, which explicitly aims to promote “[s]cience,” or knowledge.161 This suggests that uses of expression aimed at establishing the truth of some matter should be treated as presumptively fair. Showing clips of a film in order to show that digital effects have troublesome color variations would therefore be presumptively within the scope of fair use, while using the same clips of film to promote the sale of DVDs would apparently be outside of fair use.

Truth seeking produces clear follow-on benefits beyond the immediate users of a copyrighted expression. Science advances by cumulative understandings of examined phenomena. As advances are reported, others examining the same phenomena can incorporate the findings (or criticize them) in their own research. The same is true of history or news reporting. The aggregate growth in understanding benefits not only the particular person who uses any given copyrighted expression but also all others who benefit from the growth in knowledge.

Deciding whether uses of an expression constitute truth seeking may sound daunting and appear difficult to administer judicially. Fortunately, judicial determinations of the use of the expressions in truth seeking are familiar from an entirely different context: the rules of evidence. More specifically, the admissibility of hearsay evidence depends crucially on the use to which it is put. Hearsay evidence is inadmissible when used to prove some propositions, but admissible to prove others. To rule on the admissibility of hearsay, a judge must determine not only whether the evidence is used to prove the truth of a matter but also what particular matter it is being used to prove.162 The precise determination of the purpose of the use of truth-seeking speech in the context of fair use should therefore be a familiar one to courts, and one that they can carry out.

As with political speech, extant case law fails to show sufficient solicitousness to truth-seeking uses, which in our theory are presumptively treated as fair. Consider the case of American Geophysical Union v Texaco Inc.163 The case concerned the practices of hundreds of scientists working for the oil company Texaco in research aimed at producing new products and technology in the petroleum industry. The scientists, working at several of Texaco’s research centers, photocopied articles they found of interest from journals to which the Texaco library subscribed. They used the articles for archival research purposes—that is, they kept the articles on the shelf for consultation as they researched relevant topics.164 Acknowledging that the articles were copyrighted and that the photocopying was unlicensed, Texaco claimed fair use.165 In analyzing the viability of the fair use defense, the Second Circuit—both the majority and the dissent—focused on the issue of transaction costs. The main inquiry was whether there existed a market mechanism that enabled research institutes to efficiently clear reproduction rights. The majority found that such a mechanism existed in the form of the Copyright Clearance Center—a clearinghouse that dealt with copyrights in scientific publications on behalf of the publishers. Based on this finding, the majority rejected the defendant’s fair use defense.166 The dissent, by contrast, believed that the Copyright Clearance Center did not function effectively and that consequently it could not have been relied on to streamline transactions between copyright owners and researchers. Hence, the dissent was willing to recognize a fair use defense.167 The case is a textbook example of how courts should work under the transaction cost view of fair use. It also provides a vivid illustration of the limitations of the transaction cost theory; after all, the majority and the dissent could not agree on the magnitude of the cost of transacting in this case or whether it justified a finding of fair use.

Our analysis points courts in a different direction. To us, the issue is not the level of transaction costs or the feasibility of market transactions. Rather, it is the basic question whether archival research uses of works produce nonpecuniary benefits to nonusers. Accordingly, the focus of the inquiry in Texaco should have been on the nature of the use. Specifically, the court should have inquired into whether the reproduction of the scientific articles significantly advances research projects (in which case the use should have been deemed fair) or is strictly for archival purposes unrelated to the advancement of research (in which case fair use should have been denied). The question whether there is a functioning market for article reprints is largely beside the point. The only question that courts should address is whether archival storage of the kind practiced by Texaco would defeat the incentives to produce the journal articles that were subject to litigation in the Texaco case. We would hazard a guess: not likely.168

Digital libraries provide another interesting context in which fair use has been debated recently and on which our theory casts light. The issue has come up in a trio of cases: Authors Guild, Inc v Google Inc,169 Authors Guild, Inc v HathiTrust,170 and Cambridge University Press v Becker.171 All three cases raised a common legal issue: What is the status of digital libraries under the fair use doctrine? Google, the most famous case of the three, concerned the Google Books Library Project, Google’s scanning of twenty million books in order to include them in a fully searchable database.172 The second case, HathiTrust, involved a much smaller version of the same type of project. HathiTrust, a research community that is comprised, inter alia, of several universities, had set up a searchable digital library.173 Finally, Becker addressed the legality of Georgia State University’s plan to put digital copies of books and other academic materials on electronic course reserves.174

The first two cases—Google and HathiTrust—resulted in findings of fair use.175 In the third case, the district court also found fair use.176 However, on appeal, the Eleventh Circuit disapproved of the district court’s use of rigid quantitative benchmarks, such as allowing the use of one chapter or up to 10 percent of a work, and remanded the case back to the district court177 with clear instructions to adhere to the highly contextual and flexible approach adopted by other courts.178 In our analysis, all three cases reached (roughly) the correct result.

The aggregation of copyrighted works to form a searchable database creates an unparalleled research tool that cannot be otherwise duplicated. This means that the use of works in this fashion should be considered a prima facie fair use until the point at which incentives to create are eliminated. In both Google and HathiTrust, the organizations putting together the databases employed sufficient restrictions to prevent the databases from serving as complete substitutes for the original works. Google Books, for example, limits the ability of database readers to see the full works; readers are able to see only “snippets” of the full text.179 These restrictions mean that those who want to read a book will still have to purchase a copy; the main use of the database is to find relevant books or quotations regarding a point of interest. The reasoning of the Second Circuit in finding fair use in both Google and HathiTrust, particularly the special emphasis the judges put on public benefit, is highly consistent with our theory. The use of searchable digital depositories of copyrightable works creates immeasurable benefit for the public at large. In the words of Judge Denny Chin:

The benefits of the Library Project are many. First, Google Books provides a new and efficient way for readers and researchers to find books. . . . Second, in addition to being an important reference tool, Google Books greatly promotes a type of research referred to as “data mining” or “text mining.” . . . Third, Google Books expands access to books. In particular, traditionally underserved populations will benefit as they gain knowledge of and access to far more books. . . . Fourth, Google Books helps to preserve books and give them new life. . . . Finally, by helping readers and researchers identify books, Google Books benefits authors and publishers.180

Importantly, the benefits listed by Chin lie outside any of the individual books included in the project, and no work on its own could produce them. Furthermore, the fact that Google allowed readers to view only small portions of the works without being able to reproduce them led Chin to conclude that the project “advances the progress of the arts and sciences, while maintaining respectful consideration for the rights of authors and other creative individuals, and without adversely impacting the rights of copyright holders.”181 The same is true of the contested use in HathiTrust.

The Georgia State case, by contrast, calls for a more cautious approach. The inclusion of digital copies in courses’ electronic reserves clearly provides a significant benefit to students. It is conceivable that some of this benefit comes in the form of advancing research, but a better way of characterizing the use is educational (or teaching), which is a category we address below. As we note, it is not certain that the category ought to be treated as entitled to the status of a prima facie fair use. Additionally, if universities and other educational institutions were permitted under the scope of fair use to include complete copies of mandatory, copyrighted course materials in course portals, it is likely that authors’ incentives to create at least some of these works would be in jeopardy. Thus, a more contextual approach is warranted.

As the Georgia State case shows, the categories of claimed fair uses will not always be crystal clear. For instance, there will naturally be some cases in which the categories of truth seeking and political speech overlap. Harper & Row can be considered such a case. It involved the use of copyrighted expression to report the news; the news it involved was essentially political, and it thus concerned political speech. In our view, this simply makes the Harper & Row case easier to identify as one involving a presumptively fair use of an expression. Once the use is presumptively fair, it is not relevant whether it is presumptively fair due to one or several categories. Likewise, there is no need to characterize the intensity of the use in order to establish the presumption of fair use. The only question remaining for the court—regardless of whether Harper & Row is treated as a case of presumptive fair use due to news reporting or due to political speech—is whether recognizing the specific kind of fair use would eliminate the incentives for creation.

c) Criticism and comment.  Although we have noted that there is no reason to restrict the categories of presumptive fair use specifically to those mentioned in the first paragraph of § 107, there is likewise no good reason to automatically discount § 107. Section 107 offers an illustrative list of fair use purposes: “criticism, comment, news reporting, teaching (including multiple copies for classroom use), scholarship, [and] research.”182 We have already assimilated scholarship, research, and news reporting into our analysis within the broader framework of truth seeking. We now turn to the remaining trio of criticism, comment, and teaching.

Criticism and comment, in our minds, can be lumped together as uses aimed at evaluating existing works for a wider audience. In essence, criticism and comment allow an audience to experience a work indirectly, through the medium of a different user’s reaction to it. Audiences can also combine direct and indirect experiences, consuming the work together with the user’s opinion of it. For instance, a listener may read a critical review of a musical performance while listening to a recording of it to help herself form her own opinion of the piece or to compare notes with the critic.

In one sense, all criticism produces follow-on benefits for nonusers. The critic records her impressions of the copyrighted work for others to consume, rather than simply for herself. At the same time, criticism does not neatly fall into the category of a use that provides widespread nonpecuniary follow-on benefits among nonusers. In many cases, the consumption of criticism is localized among users of the copyrighted work itself. In the most extreme cases, the criticism and comment are enjoyed solely by those who are consumers of the original work. Some kinds of works by fans of cult television series, for instance, can be enjoyed only by those who have already experienced every second of every episode of the series. That group will almost always be very small and the benefits will not be widespread. Criticism and comment are therefore controversial candidates for categories of prima facie fair uses in our theory.

Altogether, we come out on the side of treating criticism and comment as prima facie fair uses—but it is worth adding two notes. First, recall the classic case of Folsom v Marsh,183 which introduced the English rule of fair use to American jurisprudence. The case involved a two-volume biography of President George Washington, comprised almost entirely of copied letters and interstitial text. The interstitial text was new, but the copied letters were used without license, and they comprised 353 (!) pages of the text.184 While the case has been cited for a number of propositions, the most important for our purposes is the court’s reliance on a line between, on the one hand, good faith use of the original “for the purposes of fair and reasonable criticism,” and on the other hand, using the original “with a view, not to criticise, but to supersede the use of the original work, and substitute the review for it.”185 Even if we recognize criticism and comment as prima facie fair uses, there will still be cases in which courts should reject the claim that the infringing use is actually criticism and comment. In some cases, as in Folsom, the court should recognize that the claim of criticism and comment is merely a subterfuge. Interestingly, from the point of view of our theory, Folsom involved a different category of use—political speech, rather than criticism and comment. Nonetheless, the result should be the same. It is not credible that the wholesale copying of 353 pages involved a good faith writing of a biography of Washington for political purposes.

Second, even if one were to reject our classification of criticism and comment as prima facie fair uses, it would not follow that criticism and comment should fall outside the scope of fair use altogether. As we noted earlier, our theory is intended to be cumulative. It still makes sense under our theory to allow the defendant to prove that a finding of fair use is warranted when excessive transaction costs are expected to block a socially optimal transaction.

d) Education.  Educational uses, too, make a mixed case for recognition as prima facie fair uses. “Teaching” is explicitly mentioned in § 107 as a fair use purpose. However, it is not entirely clear that education neatly fits within the description of a use that produces widespread nonpecuniary benefits for nonusers. Education is generally a bilateral transaction between teacher and student (or multiple students); the users of educational copyrighted materials, whether teachers or students, are the beneficiaries. In this sense, there is little reason to include education as a prima facie category of fair use. On the other hand, education is widely thought to create societal benefits beyond those conferred on the direct participants in and beneficiaries of education, for example, a reduction in crime.186 Additionally, education has a clear and obvious connection with uses that are prima facie fair uses in our theory, such as truth seeking.

We suggest that the stronger argument would omit education from a list of prima facie fair uses. Fair use in educational settings could be claimed, therefore, only when the use could fairly be categorized as political speech, research, truth seeking, criticism, or another prima facie fair use, or when the use was small enough that it could be justified by a transaction cost argument. This would still allow for a wide scope of fair uses in educational settings, but not nearly as many as for political speech, for instance.

Our approach would accord with a number of recent decisions that have been conservative with findings of fair use in education settings. A series of cases in the 1980s and 1990s dealt with photocopying materials for university classes.187 Courts were generally unsympathetic to the claim that universities and copy shops could, within the framework of fair use, photocopy all the course materials needed by a student for a given class.188 Negotiations between some of the interested parties led to two sets of informal guidelines for determining whether course-related copying is fair: the Agreement on Guidelines for Classroom Copying in Not-for-Profit Educational Institutions with Respect to Books and Periodicals189 and the Guidelines for Educational Uses of Music.190 The guidelines include research within the category of educational uses; we would separate the two categories. Otherwise, however, the guidelines are largely positive from our perspective. They attempt to demark uses that are sufficiently small to warrant a finding of fair use. In many cases, the guidelines offer quantitative benchmarks. This is a means of offering greater certainty within a transaction cost analysis of fair use.191 While the two sets of guidelines have not been adopted into law in any formal proceeding, they have received largely sympathetic treatment in case law.192 In some ways, the Georgia State case mentioned above also reflects this approach.

D.    Other Illustrations of the Proposed Fair Use Doctrine

To further illustrate how our proposed doctrine would work, and to contrast it with existing doctrine, we also consider several issues that have arisen in prior fair use litigation that do not necessarily fall into categories of prima facie fair uses. Applying the proposed new test to some old cases provides some interesting results, confirming the results in some cases and rejecting them in others.

1.   Parodies.

Courts applying fair use have long struggled with the question of how to treat parodies. Perhaps the best illustration of the courts’ confusion can be found in two Ninth Circuit cases decided by the same judge in 1955. Both Jack Benny and Sid Caesar had parodied movies on their television shows: Benny’s sketch parodied the film Gaslight,193 and Caesar’s parodied From Here to Eternity.194 While there was little to distinguish the two parodies, the court approved Caesar’s as fair use while rejecting Benny’s as an unacceptable unlicensed appropriation of a copyrighted work.195 Along the way, the court in each case emphasized that its central concern was with the amount of material used by the parody, rather than the style or target of the humorous work. In Benny’s case, for example, the court noted that “the words ‘burlesque,’ ‘parody,’ ‘satire’ and ‘travesty’ are often used interchangeably.”196

A new course was charted by the Supreme Court in Campbell. Campbell involved an infringement suit by the holder of the copyright in the song “Oh, Pretty Woman” (composed by Roy Orbison and Bill Dees) against the rap group 2 Live Crew. 2 Live Crew released to the public a song entitled “Pretty Woman” that borrowed extensively from Orbison’s hit without permission.197 While 2 Live Crew acknowledged seeking and failing to obtain permission from the copyright owner, the Supreme Court found the rap group’s use of the original could be a fair use and therefore noninfringing. Writing for a unanimous court, Justice David Souter explained that 2 Live Crew’s song “reasonably could be perceived as commenting on the original or criticizing it.”198 He argued that parodists have a special claim to fair use for two reasons. First, original authors are unlikely to license parodies of their works.199 Second, parodies are especially valuable to society on account of their transformative nature.200 Importantly, the Court denied that all parodic works were of equal value. The Court took pains to distinguish the favored parodies from the less protected satires. The “nub” of parody, according to the Court, is “the use of some elements of a prior author’s composition to create a new one that, at least in part, comments on that author’s works,”201 while satire “has no critical bearing on the substance or style of the original composition [but rather] merely uses [it] to get attention or to avoid the drudgery in working up something fresh.”202

The Campbell decision changed preexisting understandings of fair use in two important respects. First, it emphasized the importance of transformativeness in fair use determinations—an aspect that led to a revolution in the way fair use cases were decided.203 Second, the Court effectively created a “safe harbor” for parodies within fair use; although the Court never said that parodies are per se fair, commentators and courts alike understood the instruction they received from the Court as suggestive of this outcome.204 The reaction to the ruling was swift. After Campbell, defendants have sought to emphasize that their use, whatever it might be, is transformative and therefore fair.205 Likewise, defendants in fair use cases have gone to great distances to persuade courts that their putatively infringing works were parodies, and whenever they were successful, a fair use ruling ensued.206 Practically, parodies are almost guaranteed a fair use finding. Strikingly, satires are not entitled to the same status and are treated like all other putatively infringing works—that is, they are subject to the balancing test of § 107 and they are often denied fair use.207

Indeed, in many cases, the entirety of the court’s decision turns on whether the allegedly infringing work is best viewed as a parody or as a satire. Sometimes, the alleged infringer attempts to shape the likely outcome with labeling. Consider the case of Dr. Seuss Enterprises, LP v Penguin Books USA, Inc.208 Writing under the pen name “Dr. Juice,” Alan Katz authored a book called The Cat Not in the Hat! A Parody by Dr. Juice, which, as the district court noted, mimicked the distinctive style of the Dr. Seuss209 family of works in rhyme, illustration, and packaging.210 The district court rejected Katz’s claim that his book was a parody of Dr. Seuss’s works that could copy from the Dr. Seuss books under the rubric of fair use. Instead, the court found Katz’s book to be an unprotected satire that made “no [ ] attempt to comment upon the text or themes of The Cat in the Hat” and other Dr. Seuss books.211 Indeed, said the Court of Appeals, the use of Dr. Seuss’s work was “pure shtick” intended merely to retell the story of O.J. Simpson’s murder trial.212

Viewed from our vantage point, the Supreme Court got it backward in Campbell. Under our test, it is political satires that should most obviously be considered prima facie fair, whereas it is less obvious that parodies and other types of works that function primarily as criticisms of the original should receive solicitous treatment. Political parodies, such as Saturday Night Live’s famous spoof of Sarah Palin’s interview with Katie Couric,213 are clearly political speech and play a central role in democratic self-governance. Parodies that make light of the original work, like book reviews and other critical works, may also fall within the scope of fair use,214 but their demand for the protection of fair use is less pressing. Many parodies, however, serve nonpolitical purposes and thus do not provide as substantial a follow-on benefit to society.

2.   Personal use.

Personal uses of works have enjoyed a mixed reception in fair use jurisprudence. The high mark of personal uses, without doubt, was the Supreme Court’s ruling in Sony.

As previously noted, in Sony the Supreme Court ruled that the practice of home consumers recording programs for their own later viewing—“time shifting”—was a protected fair use.215 This would appear to anchor the idea that even large-scale infringements—including copying of entire works—would fall within the bounds of fair use so long as the copying were for personal, noncommercial uses.

Yet notwithstanding this ruling, subsequent attempts to anchor personal use copying as fair uses were far less successful. The clearest example of the mixed fate of personal uses in fair use litigation can be found in a number of attempts to win fair use protection for “space shifting”—the copying of works from one medium to a different medium for personal, noncommercial uses. This occurs, for instance, when someone records music from an older format (such as a CD) to a newer format (such as an MP3 file).216

Perhaps the most resounding failure of the personal fair use arguments was the Napster litigation.217 Napster was a company that created and marketed software and network services that permitted users to transfer and share files through the internet.218 Napster featured two central innovations. First, it employed a peer-to-peer transfer technology that permitted users to transfer files directly to one another without having to go through the central server. This eliminated bottlenecks and sped up file transfers. Second, Napster offered directory software that permitted users to locate files of other users and initiate downloads.219 Napster was aimed, in particular, at consumers of digital music files.220 Napster permitted millions of users to obtain digital copies of copyrighted music from one another without ever obtaining a license for the distribution or copying.221 Music companies sued Napster for its facilitation of and direct involvement in the infringing acts.222

Napster’s legal strategy was based on the Supreme Court’s ruling in Sony. Napster claimed that one of the main uses of Napster software was “space shifting”—the practice of copying from obsolete format to usable digital format music to which the consumer had already purchased a license.223 Of course, it was not Napster that directly engaged in copying. Rather, it was Napster users who did so. Napster users who wanted to space shift their music from computer to computer would make MP3 copies of their older CD formats and then use Napster to transfer the files. Napster claimed this was a near-perfect analogue to the Sony case. As in Sony, Napster claimed that it was doing nothing more than distributing an article of commerce that was sometimes used by consumers for forbidden copying, but was at other times used for permissible fair use copying.224 The key for Napster’s success was to convince the court that “space shifting” was, like “time shifting,” a fair use. On this point, unfortunately, Napster failed.225

The district court dismissed the space-shifting argument on multiple grounds. It found that the amount of pure space shifting was small, such that even if space shifting were a personal use recognized as fair, Napster could not enjoy the benefit of the Sony ruling to protect itself from other infringing activities by Napster and its users.226 But the district court went beyond simply rejecting the applicability of Sony. The court cast doubt on whether space shifting should be recognized as a fair use at all. While not entirely ruling out the possibility that space shifting might be a fair use, the court rejected several of the arguments that were proffered in favor of considering space shifting a fair use, and it carefully refrained from ever endorsing the claim that space shifting is fair.227 On appeal, the Ninth Circuit agreed that space shifting should not be viewed as a fair use, at least in the context of Napster’s business. The court emphasized that, given the way that Napster’s technology worked, any user who utilized Napster to space shift would necessarily expose the “shifted” files to other Napster users who would wish to make their own copies.228

The Napster ruling has engendered a great deal of discussion, as well as some controversy.229 However, space shifting has not won general endorsement as fair use in case law since then, notwithstanding the plausible argument that space shifting is a quintessential personal, noncommercial use. From the perspective of our theory, this makes sense. It may well be that Congress has no interest in barring unauthorized, noncommercial, personal uses. However, such personal uses do not fit into the broader purposes of fair use as we understand them. By definition, personal uses are not likely to produce widespread follow-on utility to nonusers. While the doctrine is underinclusive of particularly high-value personal uses, these uses are so rare that they need not factor into the doctrine.

This is not to say that there can be no argument for limiting enforcement of copyright against purely noncommercial personal uses. In fact, there are three possible arguments. First, given what appear to be strong social norms in favor of viewing such uses as permitted,230 it may be wise to limit enforcement in order to keep the law in line with widespread social expectations. Second, many noncommercial personal uses will be sufficiently small that transaction costs may prevent bargaining between the user of the work and the copyright owner, even though it would be socially optimal for the work to be used. And third, the promise of nonenforcement may be necessary to encourage the development and adoption of new technologies. For instance, purchasers of new music-storage technology (such as dedicated digital music players like the iPod) might want some assurance that their CD collect­ions will not become completely obsolete before they purchase the new technology.

The strength of these arguments will vary by technology. The social norms surrounding the copying of music for personal enjoyment may not apply to artwork, for example. Transaction costs will not only differ by technology; they will also change over time. It is now evident—though it was not so evident twenty years ago—that it is cost-effective to sell listeners digital recordings of music by the song. And, at least in some cases, no assurances to consumers were necessary to encourage them to adopt new technologies. DVDs supplanted videocassettes without giving consumers the ability or the right to transfer their recorded material from videotape to DVD. DVDs, in turn, are being supplanted by streaming video, again without the legal ability to transfer rights. The optimal legal treatment for such personal uses may, therefore, vary by technology. And, indeed, the strongest judicial pronouncement in favor of legal protection of space shifting as within the rights of users as a personal, noncommercial use came in Recording Industry Association of America v Diamond Multimedia Systems Inc,231 a case that construed a provision of the Audio Home Recording Act of 1992232 (AHRA) that specifically permitted the making of personal copies of digital music in certain circumstances.233 While that case found such space shifting to be a “paradigmatic noncommercial personal use entirely consistent with the purposes of the Act,” the court did not rule that all noncommercial, personal uses are fair uses.234

3.   Fan fiction.

Fan fiction provides an interesting application of our theory. For our purposes, the category of fan fiction encompasses noncommercial production—and often dissemination—of a broad range of derivative works that rely on prior successful works. The Harry Potter and The Hunger Games book series, the Star Wars movie series, and the Star Trek television series are just a few well-known examples of works that engendered a significant body of derivative works by consumers.235 The unlicensed derivative works, in many cases, are the vehicles for extending enjoyment of the fictional works to contexts that the original creators would never have thought to explore.

Some scholars have called for the creation of a new legal privilege for fan fiction on grounds of self-expression,236 harnessing cultural theory to support this initiative. Others have sought to assimilate fan fiction within existing fair use doctrine. Professors Anupam Chander and Madhavi Sunder, for example, have contended that fan fiction should be entitled to a status of a presumptively fair use on the grounds that it is parodic and highly transformative.237 Using this rationale, they have suggested that even commercial fan fiction should be considered fair.238

Our model offers a more nuanced and, we believe, clearer position. On our view, which once again focuses on the production of dispersed nonpecuniary benefits to nonusers, fan fiction should enjoy a presumptively fair status only if it is noncommercial and only if it is disseminated to the public. Hence, noncommercial literary, musical, and audiovisual adaptations of works should be considered fair, as long as they are made available to the public. The same is true of noncommercial parties and costume parades. Commercial merchandise, video games, and other derivative works that are produced for profit would not be sheltered under our theory.

IV.  Exploring the Interface with Other Theories

Having presented our view of fair use in theory and in doctrine, we turn in this Part to an exploration of the way our proposed view of fair use interacts with related theoretical scholarship in the fields of property and copyright.

A.    Free Speech, the First Amendment, and Copyright

We begin by looking at the obvious connections between our theory and broader theories of the importance of free expression in society, particularly in the context of the First Amendment and its guarantee of freedom of speech. A number of scholars have noted that there is some obvious tension between the strategies chosen by free speech law and copyright law to encourage expression.239 Copyright law seeks to support expression by restricting its use; free speech law seeks to support it by sweeping away restrictions. As we have noted, copyright law aims to give authors the opportunity to realize pecuniary gain from expressions they have created by giving them exclusive property rights over various uses of the expressions.240 Free speech law, by contrast, aims to expand the scope and reach of various kinds of expression by preventing government interference with the expressive activity.241

Many scholars have tried to explain where the boundary between these two strategies lies. In particular, they have explored when concerns of free speech should trump copyright law and permit unfettered expression that utilizes copyrighted speech.242 It is undeniable that the two bodies of law adopt fundamentally opposite approaches to promoting speech in society. Professor Rebecca Tushnet has provocatively labeled copyright “book burning mandated by law.”243 Professor Julie Cohen has labeled copyright a “form of censorship.”244 Several scholars have suggested broader exemptions to copyright protections in order to accommodate free speech concerns.245

The natural starting point for exploring this scholarship remains Professor Melville Nimmer’s classic 1970 article, Does Co­pyright Abridge the First Amendment Guarantees of Free Speech and Press?, which first identified the “largely ignored paradox” of the Copyright Act forbidding speech.246 Several decades later, Professor Netanel picked up these themes in a number of works, including a book247 and several articles.248 Netanel has argued that much of modern copyright law exceeds constitutional free speech limits.249 Netanel’s critique is not based on the fair use doctrine, per se; rather, it looks holistically at all copyright law, shining light on numerous doctrines such as the tests for determining nonliteral infringing copying.250 After the Supreme Court’s ruling in Golan,251 Netanel argued that the Supreme Court had adopted Nimmer’s proposed “definitional balancing” by constitutionalizing the fair use doctrine and what is called the idea-expression dichotomy.252 However, Netanel was less interested in sketching out the scope of the fair use doctrine as required by the constitutional guarantee of free speech.253

Tushnet’s justly celebrated Copy This Essay: How Fair Use Doctrine Harms Free Speech and How Copying Serves It is the most direct exploration of how fair use doctrine must be viewed in light of free speech concerns. Tushnet has celebrated the focus of fair use doctrine in recent years on transformativeness; she has agreed that transformative uses should generally be viewed as fair. At the same time, she has argued that transformativeness cannot be the sine qua non of fair use doctrine, as it is excessively narrow. Even with her recommendation that personal, noncommercial uses also be considered fair uses, she has acknowledged that she “[has] no [ ] solution to offer,”254 adding that “[a] judicial attitude of tolerance toward copying that is not quantitatively large would help when fair use cases were actually litigated, but any standard would be difficult to codify and would still subject many uses to a chilling uncertainty.”255

While our Article touches on many of the same concerns as the fascinating free speech–copyright literature, our aim is different. We do not attempt to draw the boundary lines between the concerns of free speech and copyright. Rather, we seek to explicate the proper scope of the fair use doctrine. That said, we acknowledge that our approach may contribute to understanding how to harmonize the concerns of copyright law with those of free speech law. We argue that our doctrine offers an easier way to accommodate the tension—one that has been alluded to by the Supreme Court in the years since Tushnet wrote her essay. Essentially, our version of the fair use doctrine would involve a broadening and a reinterpretation of the doctrine that would cover enough speech to take care of the free speech concerns.

B.    Spillovers

Another body of scholarly work whose concerns clearly overlap with ours is one addressing the issue of positive externalities. In an impressive body of work, Professor Brett Frischmann, alone256 and with Professor Mark Lemley,257 has discussed the important phenomenon of positive externalities (“spillovers,” in his terminology) and its importance for policymaking in the realm of intellectual property. The gist of Frischmann’s theory is that since intellectual goods produce many important spillovers for society, it would be unwise to grant strong protection to copyright owners lest the all-important spillovers be lost.258 Our theory touches on many of the concerns addressed by the work of Frischmann and others—such as Lemley,259 Professor Katherine Strandburg, Professor Michael Meurer,260 and Professor Michael Carroll261 —but as far as the bottom line is concerned, our theory diverges from these in critical respects.

Not only does our theory not depend on the concept of spillovers, but also we believe that the presence of spillovers actually mitigates in favor of stronger author rights, all things being equal. The accepted lore in economics is that all externalities—negative and positive alike—present a similar challenge.262 When actors fail to take account of negative externalities, they are likely to carry out too much of the activity. When it is positive externalities that they overlook, actors are likely to carry out too little of the activity.263 A central, though by no means exclusive, strategy for countering over- or underproduction of an activity is internalization.264 When the actor has to fully bear the costs of negative externalities and may fully realize the gains of positive externalities, the actor’s cost-benefit analysis will mirror society’s, and the actor will undertake the socially optimal amount of the activity. As Professor Harold Demsetz famously observed, property rights implement a strategy of internalization by giving an owner a set of exclusive rights regarding an asset.265 This is true even if the asset in question is intangible. It is unsurprising, therefore, that Professor Polk Wagner has pointed out that in the intellectual property domain, stronger protection may actually create more information for the public than weak protection.266

To be sure, as Professor Coase, Demsetz, and others have observed, absolute property rights are not a panacea.267 Setting up and enforcing property rights is not costless.268 Regulation or other nonproperty methods may prove more cost-effective in incentivizing the optimal level of activity.269 Likewise, the initial allocation of property rights may prove “sticky” as a result of high transaction costs, making property rights inferior at times to regulated access. As we noted, the nonrivalrous nature of consuming intellectual property goods may make it worthwhile to allocate certain uses directly to users.270 Indeed, as Demsetz pointed out in his response to the literature on spillovers, the desirability of property rights is highly context dependent.271 From an economic standpoint, there is no universally correct approach. The appropriate response to externalities depends on the relative costs and expected benefits of different methods of incentivizing behavior. But certainly one cannot conclude from the mere presence of positive externalities that it is societally beneficial to deny the owner the right to capture the value of those externalities.

The theory we develop here is independent of the concept of spillovers. Ours is an allocative theory that strives to divide uses of expressive works such that certain types of uses will initially be allocated to consumers in order to save on transaction costs and encourage uses otherwise favored by the state. On our vision, certain uses of works are of such great significance to the public that they must be allocated to it, at least in principle. We allow courts the power to change the initial allocation in appropriate cases, but not because of the existence of spillovers. Rather, courts can take away uses from the public and shift them to authors if doing so is absolutely necessary to preserve incentives to create. In this sense, our theory converges with the market-failure theo­ry: like Professor Gordon,272 we believe that uses that threaten to undermine incentives to create should be judicially reapportioned.

C.    Other Fair Use Theories

Fair use is not only one of the most litigated topics in the law, but it is also one of the most frequently commented on.273 It is simply not possible within the scope of a law review article to give proper credit and acknowledgement to the many invaluable scho­larly contributions to thinking about fair use. That said, there are only a small number of comprehensive attempts at rethinking the theory behind the fair use doctrine and recrafting the doctrine accordingly. In this Section, we distinguish our comprehensive attempt from two others—one by Professor William Fisher, and another one by one of us. We also address doctrinal analyses by Professor Glynn Lunney and Professor Samuelson that share some interesting features with our approach in this Article.

1.   Fair use and allocative efficiency.

In his ambitious 1988 article, published after the controversial Harper & Row decision by the Supreme Court, Fisher suggests two possible lines along which fair use doctrine might be reformulated. One possible approach is using the fair use doctrine to promote certain views of the “good society.”274 The other, which is much closer to our concerns in this Article, is allocative efficiency.275

In suggesting an allocative-efficiency approach to the fair use doctrine, Fisher shares our basic assumption: that the fair use doctrine aims to allocate certain uses directly to the public in order to reduce the inefficiency of the copyright owner’s monopoly power, while still retaining the minimum incentives necessary for the creation of copyrighted works.276 However, Fisher chooses a more direct strategy for allocating uses to the public. Rather than identify categories of presumptive fair uses, as we have done, Fisher calls for measuring all potential uses of all similar works whenever a fair use is claimed. For each of these uses, Fisher calls on judges to estimate two items: the amount of incentives that would be created for the author if the use were found unfair (that is, the profit the author could realize by selling the use) and the societal-efficiency loss that would be engendered by finding the use unfair (that is, the loss to society resulting from the owner not licensing a use).277 A high ratio between the two would indicate that the use is not fair; a low ratio would show that the use is fair. The cutoff point between high and low should be set at the use with the highest marginal aggregate social gain (gain to society from creation minus loss to society from monopoly control over uses).278

Fisher’s approach is obviously appealing in that it seeks to reach a more precise allocation of rights and privileges between author and public. If Fisher’s method could be followed by judges, fair use could achieve precise allocative efficiency, ensuring that authors receive as much incentive as necessary to create while leaving for the public whatever is unnecessary.

Unfortunately, Fisher’s method also appears to be impossible to implement in the real world. It requires of judges extensive knowledge of the entire field of creative works, potential uses, and values to society. These factors change over time, so judges could not consistently rely on prior rulings; essentially, each judge facing a fair use decision would have to recalculate and set the bar anew. It is not surprising, therefore, that even Fisher acknowledges that the framework he constructs is impractical.279 Thus, while Fisher’s approach is important for understanding the trade-offs involved in fair use decisions, his proposal offers little purchase to judges and actors.

Interestingly, Fisher’s suggestions for a fair use doctrine oriented around promoting the aims of a “good society” come closer to our doctrinal approach because they privilege certain kinds of uses, such as education.280

2.   Fair use and custom.

A different approach at regularizing fair use doctrine aims to anchor it in custom. A 2004 article by Professor Michael Madison,281 as well as a 1997 article by one of us,282 suggests that fair use doctrine can be reorganized around existing social practices. The 1997 article proposes a focused analysis on a particular social concern, namely, the reciprocity of risk.283 Fair use, on this understanding, should shield infringements by users who are themselves authors whose own works can be similarly infringed.284 This is not as narrow a reading of fair use as it might seem on first blush, given the low bar for the creation of protected works under copyright. Indeed, it would be unusual if users of copyrighted works could not be considered authors. The article therefore proposes limiting fair use findings to users or authors whose practices fall within extant societal norms.285 In those cases, the risks imposed by the purported fair use can be considered to be the kinds of reciprocal risks that escape legal liability, since penalizing such risks would do little more than incur administrative costs.286

Madison’s piece undertakes a far broader review of the state of fair use law, though its ambitions for rewriting fair use law are more limited. Madison’s analysis is grounded in the case law, and it abstracts from the cases to broader principles. In general, Madison argues that fair use cases fall into patterns clustered around certain social contexts and associated customs, and that fair use law should aim to enforce societal expectations of legitimate use.287 Reciprocity is important to Madison, but “the pattern-oriented approach does not require evidence of reciprocal behavior as such in all cases.”288 Rather, “[i]t requires commonality of behavior and expectation.”289

In this Article, our approach is different. We do not seek to identify extant societal behavior that ought to be protected through fair use doctrine, nor do we seek to enforce it through reinterpretation of fair use. Our organizing principle for the law of fair use is not reciprocity of risk or the assumption of efficient societal norms. Rather, it is the centrality of the fair use doctrine for allocating copyright-related rights and privileges efficiently.

3.   Presumptive fair use.

Another interesting perspective on the fair use doctrine can be found in Lunney’s 2002 article on the Sony decision and its proper place in fair use analysis.290 Nominally, Lunney’s article aims only to correct a long-standing misinterpretation of the Sony decision. In fact, Lunney’s article strives for something far more ambitious. Presenting fair use as the means for balancing bet­ween copyright’s conflicting goals of incentivizing creation and encouraging use of creative works, Lunney argues that essentially all uses should be presumed fair.291 In Lunney’s account, the reason that the time shifting in Sony was a fair use was not because it was small enough that it was likely to be foiled by transaction costs.292 Rather, it was fair because, like all uses, it was to be considered fair unless the copyright owner could show that the use caused great loss to the owner—which in that case was a burden the copyright owner could not meet.293

In general, Lunney suggests that the proper test for fair use ought to presume that all uses are fair unless proved otherwise, and also suggests that owners should be able to prove that uses are unfair only by demonstrating both that the use has an adverse effect on the market for the creation, and that the adverse effect would affect marginal incentives for production.294 Lunney’s version of fair use, in other words, is even broader than ours, as it makes all uses of copyrighted works presumptively fair. Nonetheless, Lunney’s approach shares a great deal with ours. Like us, Lunney views fair use as an allocative tool for balancing the incentive effects of granting rights to authors with the benefits of widespread use of works created by public privileges to unlicensed use of creative works. Like us, Lunney suggests that fair use doctrine can best be shaped to this allocative function by viewing at least some uses as presumptively fair.295 However, we believe that a fair use doctrine that views all uses as presumptively fair strikes the wrong allocative balance.

An interesting contrast with Lunney’s approach is provided by Samuelson’s Unbundling Fair Uses. As noted above, Samuelson emphasizes the degree to which extant fair use law is clustered around what she calls “policy-relevant” patterns.296 In privileging some uses by giving them the status of presumptive fair uses, while denying that status to other uses, our proposal echoes a pattern of extant fair use law that is both identified and lauded by Samuelson. Samuelson cautiously suggests that existing fair use law is normatively sound, as it befits treatment of copyright as “limited monopoly.”297 She too proposes limited burden shifting, writing that

[g]iven the important role that fair use plays in mediating tensions between copyright law and the First Amendment and other constitutional values, it would be appropriate for the burden of showing unfairness to be on the copyright owner. . . . At the very least, copyright owners should bear the burden of proving unfairness in free speech/expression, personal use, and litigation use cases.298

While neither Lunney nor Samuelson replicates our theoretical approach or doctrinal prescriptions, the doctrinal implications of their suggestions belong to the same family of approaches as ours.

Conclusion

Fair use is the principal mechanism employed by copyright law to balance the interests of authors and copyright-reliant industries against those of the public. It is therefore not surprising that it has attracted considerable attention from copyright theor­ists, who offer myriad theories to justify and explicate this remarkable doctrine.

In this Article, we have offered a novel approach to the fair use doctrine based on two central insights. First, we show that the fair use doctrine is best understood as an allocative tool that grants use categories directly to users to ensure efficient utilization of the rights associated with copyright. Second, we show that the categories allocated to users by the fair use doctrine should be those that provide follow-on users with significant nonpecuniary benefits.

Our novel approach to the fair use doctrine can provide some needed clarity and coherence to a doctrine that has puzzled scholars and confounded courts. Naturally, in a scholarly space rife with many outstanding preexisting contributions, it is impossible to be completely innovative. Indeed, as we point out throughout the Article, on various dimensions our theory overlaps with preexisting ones. Like many others before us, we also clearly benefit from standing on the shoulders of giants. Yet the core theoretical thesis, as well as the enumeration of fair and unfair use categor­ies we advance, breaks new ground and can introduce a great deal of coherence and clarity into judicial determinations and academic discussions of the most remarkable doctrine in the entire world of copyright.

  • 26See generally Gordon, 82 Colum L Rev 1600 (cited in note 5).
  • 27See Raymond Shih Ray Ku, Consumers and Creative Destruction: Fair Use beyond Market Failure, 18 Berkeley Tech L J 539, 542 (2003); Christina Bohannan, Copyright Harm, Foreseeability, and Fair Use, 85 Wash U L Rev 969, 972 n 15 (2007).
  • 28See generally Gordon, 82 Colum L Rev 1600 (cited in note 5).
  • 29US Const Art I, § 8, cl 8.
  • 30Balganesh, 122 Harv L Rev at 1576–77 & n 16 (cited in note 14). See also Landes and Posner, 18 J Legal Stud at 326 (cited in note 16) (emphasizing that one of copyright law’s central goals is to maintain incentives for individuals to produce creative works).
  • 31See Nigel Parker, Music Business: Infrastructure, Practice and Law 272 (Sweet & Maxwell 2004); Jane C. Ginsburg, Creation and Commercial Value: Copyright Protection of Works of Information, 90 Colum L Rev 1865, 1869 (1990) (“Moreover, to the extent that the worth of the work lies in the information, rather than in the form[,] . . . modern copyright’s emphasis on personality-manifesting characteristics fails to secure the commercial value of these kinds of endeavors, even though the demand for productions such as directories . . . is ever increasing.”).
  • 32R.H. Coase, The Problem of Social Cost, 3 J L & Econ 1, 15 (1960).
  • 33Id.
  • 34Id at 15–16.
  • 35Gordon, 82 Colum L Rev at 1626 (cited in note 5) (“When severe market failure is present, injury to the copyright owner may not follow from infringement.”).
  • 36Id at 1626–27.
  • 37Gideon Parchomovsky, Fair Use, Efficiency, and Corrective Justice, 3 Legal Theory 347, 357–61 (1997).
  • 38See, for example, Wendy J. Gordon and Daniel Bahls, The Public’s Right to Fair Use: Amending Section 107 to Avoid the “Fared Use” Fallacy, 2007 Utah L Rev 619, 651–54 (addressing the “need [for] a comprehensive definition of the public’s rights in the realm of expression” in the doctrine of fair use); Gordon, 50 J Copyright Society USA at 151–54 (cited in note 7) (describing the claim in earlier work that “fair use should generally be denied if recognizing the defense would cause substantial injury to the copyright owner” as “overbroad”); Gordon, 82 BU L Rev at 1035 (cited in note 7) (asserting that her original use of the term “market failure” was overly narrow and should be more expansive).
  • 39Ku, 18 Berkeley Tech L J at 542 (cited in note 27), citing Gordon, 82 Colum L Rev at 1627–35 (cited in note 5).
  • 40Randal C. Picker, From Edison to the Broadcast Flag: Mechanisms of Consent and Refusal and the Propertization of Copyright, 70 U Chi L Rev 281, 295 (2003) (“As transaction costs drop through a combination of institutional arrangements such as the Copyright Clearance Center, and as the internet creates a ubiquitous structure for micro-transactions—microconsents with micropayments—fair use might cease to play a meaningful role.”); Gideon Parchomovsky and Kevin A. Goldman, Fair Use Harbors, 93 Va L Rev 1483, 1499–1500 (2007) (“[T]ransaction costs are not static.”).
  • 41Parchomovsky and Goldman, 93 Va L Rev at 1499–1500 (cited in note 40) (“The rise of the Internet and the advent of digital platforms, together with the development of copyright clearance agencies, have reduced many of the transaction costs that previously stood as barriers to cost-effective bargaining.”) (citation omitted).
  • 42Robert P. Merges, The End of Friction? Property Rights and Contract in the “Newtonian” World of On-Line Commerce, 12 Berkeley Tech L J 115, 130, 136 (1997).
  • 43Tom W. Bell, Fair Use vs. Fared Use: The Impact of Automated Rights Management on Copyright’s Fair Use Doctrine, 76 NC L Rev 557, 579–600 (1998).
  • 44Gordon and Bahls, 2007 Utah L Rev at 619–24 (cited in note 38).
  • 45Ben Depoorter and Francesco Parisi, Fair Use and Copyright Protection: A Price Theory Explanation, 21 Intl Rev L & Econ 453, 458–62 (2002).
  • 46Id at 463–64 (“The emergence of a ‘click and pay’ economy . . . does not necessarily eliminate the strategic pricing of copyright licenses. Whenever anticommons costs are serious enough to undermine the viability of the transaction, fair-use doctrines become a valuable tool for mitigating the resulting deadweight losses.”).
  • 47Matthew Sag, Copyright and Copy-Reliant Technology, 103 Nw U L Rev 1607, 1664 (2009).
  • 48Id at 1666.
  • 49See Gordon, 50 J Copyright Society USA at 150 (cited in note 7) (asserting that her “1982 article has been often misapplied, by both courts and commentators”); Gordon and Bahls, 2007 Utah L Rev at 623–24 (cited in note 38) (clarifying that her original market-failure theory “never purported to displace the other justifications for fair use,” but has been used by commentators to suggest that transaction costs are “the sole basis for fair use”).
  • 50Gordon and Bahls, 2007 Utah L Rev at 635 (cited in note 38).
  • 51659 F2d 963 (9th Cir 1981). Gordon’s article was written in the interim between this decision and Sony Corp of America v Universal City Studios, Inc, 464 US 417 (1984). The article focused, naturally, on the Ninth Circuit’s decision.
  • 52Sony, 464 US at 420.
  • 53See Gordon, 82 Colum L Rev at 1652–57 (cited in note 5).
  • 54See id at 1655–57.
  • 55 464 US 417 (1984).
  • 56Id at 455–56.
  • 57510 US 569 (1994).
  • 58Id at 571–72.
  • 59Id at 581–83.
  • 60Id at 579 (“[T]he more transformative the new work, the less will be the significance of other factors, like commercialism, that may weigh against a finding of fair use.”).
  • 61See Kathleen K. Olson, The Future of Fair Use, 19 Communication L & Pol 417, 420–21 (2014) (highlighting the importance of transformativeness in court decisions post-Campbell); Sag, 73 Ohio St L J at 76–77 (cited in note 8) (concluding, based on empirical studies of court decisions, that although transformativeness is “not the ‘alpha and omega’” of fair use, it is nevertheless a significant core component of fair use); Neil Weinstock Netanel, Making Sense of Fair Use, 15 Lewis & Clark L Rev 715, 736 (2011) (finding that, after 2005, courts’ use of the fair use doctrine is “overwhelmingly dominated” by transformativeness); Rebecca Tushnet, Copy This Essay: How Fair Use Doctrine Harms Free Speech and How Copying Serves It, 114 Yale L J 535, 550–51 (2004) (tracing the rising importance of transformativeness in fair use back to 1990); Matthew D. Bunker, Eroding Fair Use: The “Transformative” Use Doctrine after Campbell, 7 Communication L & Pol 1, 2 (2002) (analyzing the history of transformativeness and arguing that courts’ increasing reliance on it overprotects copyright holders). Transformativeness may sometimes overlap with high transaction costs, but as a practical and analytical matter, there is no necessary correlation between the two benchmarks.
  • 62467 F3d 244 (2d Cir 2006).
  • 63714 F3d 694 (2d Cir 2013).
  • 64Blanch, 467 F3d at 246.
  • 65Id at 254–55.
  • 66Id at 259, quoting US Const Art I, § 8, cl 8.
  • 67Cariou, 714 F3d at 699.
  • 68See id.
  • 69Id at 707.
  • 70Id.
  • 71Cariou, 714 F3d at 712.
  • 72Brian Boucher, Landmark Copyright Lawsuit Cariou v. Prince Is Settled (Art in America, Mar 18, 2014), archived at http://perma.cc/B8LX-5K8Q.  
  • 73537 US 186 (2003).
  • 74132 S Ct 873 (2012).
  • 75Pub L No 105-298, 112 Stat 2827 (1998), codified as amended at 17 USC §§ 108, 203, 301–04.
  • 76Eldred, 537 US at 193.
  • 77Id at 195–96.
  • 78Id at 193.
  • 79Id at 221–22.
  • 80Eldred, 537 US at 219, citing Harper & Row, Publishers, Inc v Nation Enterprises, 471 US 539, 560 (1985).
  • 81Eldred, 537 US at 219–20, quoting Harper & Row, 471 US at 560.
  • 82Golan, 132 S Ct at 878.
  • 8317 USC § 104A(a)(1), (h)(6).
  • 84Golan, 132 S Ct at 891.
  • 85Id at 890.
  • 86See id at 890–91.
  • 87Neil Weinstock Netanel, First Amendment Constraints on Copyright after Golan v. Holder, 60 UCLA L Rev 1082, 1128 (2013).
  • 88See Part I.A.
  • 89Gordon, 82 Colum L Rev at 1618–19 (cited in note 5). Gordon terms this the “substantial injury hurdle” required to be overcome in cases of “intermediate market failure.” Id.
  • 90See Part III.C.1.
  • 91See Part III.C.2.
  • 92Indeed, in one important sense, our proposal is very close to Gordon’s. Specifically, as we show, our proposal, like Gordon’s, is—at least in some cases—focused on transaction costs. However, our treatment of the transaction cost issue is very different from Gordon’s. In our view, one of the central concerns of the fair use doctrine is granting users the rights to carry out uses that would otherwise be attended by high transaction costs. However, the reason for this is not that we are concerned that the high transaction costs might block otherwise-valuable transactions. Rather, we worry that the transaction costs themselves are costs to society that can sometimes be avoided. Simply put, when the fair use doctrine allocates a use to a user who can derive benefit from the expression, the doctrine saves the user and the author the cost of negotiating a license and thereby benefits society. Viewed in this way, transaction costs are potentially enough to justify fair use any time they are above zero (or at least above the cost of administering the doctrine of fair use), even if they are not so high as to bar voluntary transactions.
  • 93See Yochai Benkler, Siren Songs and Amish Children: Autonomy, Information, and Law, 76 NYU L Rev 23, 59–60 (2001) (“The Supreme Court has held consistently and unanimously that American law explicitly treats intellectual property rights in utilitarian terms.”); Yochai Benkler, Through the Looking Glass: Alice and the Constitutional Foundations of the Public Domain, 66 L & Contemp Probs 173, 175–77 (Winter/Spring 2003).
  • 94US Const Art I, § 8, cl 8.
  • 95Julie E. Cohen, Copyright and the Perfect Curve, 53 Vand L Rev 1799, 1801 (2000) (“Copyright law conceives and promotes progress in two distinct but related ways: First, it seeks to increase both the quantity and quality of creative output. Second, it seeks to broaden public access to creative works.”); S.J. Liebowitz, Copyright Law, Photocopying, and Price Discrimination, 8 Rsrch L & Econ 181, 184–86 (1986); Ian E. Novos and Michael Waldman, The Effects of Increased Copyright Protection: An Analytic Approach, 92 J Polit Econ 236, 244 (1984).
  • 96See generally Landes and Posner, 18 J Legal Stud 325 (cited in note 16).
  • 97Depoorter and Parisi, 21 Intl Rev L & Econ at 454 n 4 (cited in note 45); Mark A. Lemley, The Economics of Improvement in Intellectual Property Law, 75 Tex L Rev 989, 994–96 (1997); Landes and Posner, 18 J Legal Stud at 326 (cited in note 16); William W. Fisher III, Reconstructing the Fair Use Doctrine, 101 Harv L Rev 1659, 1700–05 (1988); Gordon, 82 Colum L Rev at 1610–11 (cited in note 5); Stephen Breyer, The Uneasy Case for Copyright: A Study of Copyright in Books, Photocopies, and Computer Programs, 84 Harv L Rev 281, 281 n 4 (1970).
  • 98See Niva Elkin-Koren, Copyrights in Cyberspace—Rights without Laws?, 73 Chi Kent L Rev 1155, 1171 (1998); Sterk, 94 Mich L Rev at 1204–05 (cited in note 12). See also Christopher S. Yoo, Copyright and Product Differentiation, 79 NYU L Rev 212, 227–28 (2004); Dan L. Burk, Muddy Rules for Cyberspace, 21 Cardozo L Rev 121, 133–34 (1999); Neil Weinstock Netanel, Asserting Copyright’s Democratic Principles in the Global Arena, 51 Vand L Rev 217, 248–49 (1998); Landes and Posner, 18 J Legal Stud at 350–51 (cited in note 16).
  • 99See Mark A. Lemley, Property, Intellectual Property, and Free Riding, 83 Tex L Rev 1031, 1058–59 (2005); Burk, 21 Cardozo L Rev at 133 (cited in note 98); Elkin-Koren, 73 Chi Kent L Rev at 1171 (cited in note 98); Sterk, 94 Mich L Rev at 1205 (cited in note 12).
  • 100Wendy J. Gordon, An Inquiry into the Merits of Copyright: The Challenges of Consistency, Consent, and Encouragement Theory, 41 Stan L Rev 1343, 1344 (1989). See also Netanel, 51 Vand L Rev at 248–49 (cited in note 98) (“The incentive rationale posits, however, that such a ‘tax on readers’ is a necessary evil because, without it, audiences would have little creative expression to choose from in the first place.”).
  • 101David W. Barnes, Congestible Intellectual Property and Impure Public Goods, 9 Nw J Tech & Intel Prop 533, 537–38 (2011); R.A. Musgrave, Provision for Social Goods, in J. Margolis and H. Guitton, eds, Public Economics: An Analysis of Public Production and Consumption and Their Relations to the Private Sectors 124, 126–29 (Macmillan 1969). See also Christopher S. Yoo, Copyright and Public Good Economics: A Misunderstood Relation, 155 U Pa L Rev 635, 644–46 (2007).
  • 102See Daphna Lewinsohn-Zamir, Consumer Preferences, Citizen Preferences, and the Provision of Public Goods, 108 Yale L J 377, 380–90 (1998) (discussing multiple theories explaining why private markets underproduce public goods).
  • 103See Michael J. Burstein, Exchanging Information without Intellectual Property, 91 Tex L Rev 227, 235–36 (2012); Peter S. Menell and Suzanne Scotchmer, Intellectual Property Law, in A. Mitchell Polinsky and Steven Shavell, eds, 2 Handbook of Law and Econ­omics 1473, 1476–79 (Elsevier 2007); Mark A. Lemley, Ex Ante versus Ex Post Justifications for Intellectual Property, 71 U Chi L Rev 129, 129 (2004); Gordon, 82 Colum L Rev at 1610–11 (cited in note 5).
  • 10417 USC § 106.
  • 105See 17 USC § 106.
  • 106See Lemley, 83 Tex L Rev at 1058–60 (cited in note 99).
  • 107Landes and Posner, 18 J Legal Stud at 332 (cited in note 16).
  • 108Id.
  • 109See id at 361 (“Whether the law has struck the right balance is not readily determinable.”).
  • 11017 USC § 302.
  • 11117 USC § 106.
  • 112See, for example, 17 USC § 110.
  • 113See Abraham Bell and Gideon Parchomovsky, Reinventing Copyright and Patent, 113 Mich L Rev 231, 276–77 (2014) (asserting that a voluntarily tailored copyright regime would “preserv[e] the incentives to produce intangible assets while reducing the cost to society”). See also Bracha and Syed, 92 Tex L Rev at 1858–59 (cited in note 14) (explaining how increased copyright protection can block access to some works that would have been created even without the additional incentive).
  • 114See Bell and Parchomovsky, 113 Mich L Rev at 241–42 (cited in note 113).
  • 11517 USC § 302(a).
  • 116See Israel M. Kirzner, Economic Planning and the Knowledge Problem, 4 Cato J 407, 412–14 (1984); Robert D. Cooter, Decentralized Law for a Complex Economy: The Structural Approach to Adjudicating the New Law Merchant, 144 U Pa L Rev 1643, 1644–46 (1996).
  • 117Eldred, 537 US at 248 (Breyer dissenting).
  • 118See Nimmer, 66 L & Contemp Probs at 266–67 (cited in note 2) (highlighting the number of Supreme Court decisions interpreting fair use).
  • 119See generally, for example, Folsom v Marsh, 9 F Cases 342 (CCD Mass 1841).
  • 120See William Patry, How to Fix Copyright 215 (Oxford 2011) (tracing the origins of the fair use doctrine back to the 1710 Statute of Anne in England); Matthew Sag, The Prehistory of Fair Use, 76 Brooklyn L Rev 1371, 1387–93 (2011) (arguing that the development of fair use “predates Folsom v. Marsh by at least 100 years”); Kylie Pappalardo and Brian Fitzgerald, Copyright, Fair Use and the Australian Constitution, in Brian Fitzgerald and John Gilchrist, eds, Copyright Perspectives: Past, Present and Prospect 125, 129–30 (Springer 2015) (tracing the English history of copyright law to 1740 and noting how the English cases informed the American doctrine of fair use).
  • 12117 USC § 107.
  • 122Pierre N. Leval, Toward a Fair Use Standard, 103 Harv L Rev 1105, 1105–07 (1990) (citations omitted).
  • 123Sag, 73 Ohio St L J at 76–77 (cited in note 8); Beebe, 156 U Pa L Rev at 605–06 (cited in note 1).
  • 124See Michael J. Madison, A Pattern-Oriented Approach to Fair Use, 45 Wm & Mary L Rev 1525, 1666 (2004) (describing the unpredictable nature of fair use and calling it “a lottery argument”). See also Thomas F. Cotter, Fair Use and Copyright Overenforcement, 93 Iowa L Rev 1271, 1284 (2008) (modeling the similar uncertainty surrounding an indiv­idual’s ex ante determination of whether a particular use of a copyrighted work will qualify as a fair use); James Gibson, Once and Future Copyright, 81 Notre Dame L Rev 167, 192 (2005).
  • 125See, for example, Swatch Group Management Services Ltd v Bloomberg LP, 756 F3d 73, 90 (2d Cir 2014); Authors Guild, Inc v HathiTrust, 755 F3d 87, 98 (2d Cir 2014) (holding that copying entire books did not prevent a finding of fair use because it was necessary to enable full-text searching); A.V. v iParadigms, LLC, 562 F3d 630, 642 (4th Cir 2009) (noting that the use of a work in its entirety was not dispositive of fair use); Bill Graham Archives v Dorling Kindersley Ltd, 448 F3d 605, 613 (2d Cir 2006); Chicago Board of Education v Substance, Inc, 354 F3d 624, 629 (7th Cir 2003).
  • 126See, for example, Ringgold v Black Entertainment Television, Inc, 126 F3d 70, 75–77 (2d Cir 1997) (finding that a 26.75-second exposure of the plaintiff’s posters on a television show weighed against a finding of fair use); Allen-Myland, Inc v International Business Machines Corp, 746 F Supp 520, 534–35 (ED Pa 1990); Iowa State University Research Foundation, Inc v American Broadcasting Companies, Inc, 621 F2d 57, 61–62 (2d Cir 1980) (holding that a defendant’s use of 8 percent of a film in its broadcasts was significant and concluding that a fair use defense was unavailable).
  • 127See, for example, Arrow Productions, Ltd v Weinstein Co, 44 F Supp 3d 359, 371 (SDNY 2014) (comparing the length of three scenes from a copyrighted movie with the overall length of the movie); NXIVM Corp v The Ross Institute, 364 F3d 471, 480–81 (2d Cir 2004) (comparing the amount copied from a copyrighted work with the work’s total length); Wright v Warner Books, Inc, 953 F2d 731, 738–39 (2d Cir 1991) (comparing the quotations taken from a copyrighted book with the length of the book itself); New Era Publications International, ApS v Carol Publishing Group, 904 F2d 152, 158–59 (2d Cir 1990).
  • 128See, for example, Salinger v Random House, Inc, 811 F2d 90, 98–99 (2d Cir 1987) (noting that copyrighted material appeared on 40 percent of an infringing book’s pages); Harper & Row, Publishers, Inc v Nation Enterprises, 471 US 539, 564–66 (1985); Meeropol v Nizer, 560 F2d 1061, 1070–71 (2d Cir 1977) (finding it significant that copyrighted letters comprised a substantial portion of promotional materials for an infringing book although they made up less than 1 percent of the book itself).
  • 129See Gregory K. Klingsporn, The Conference on Fair Use (CONFU) and the Future of Fair Use Guidelines, 23 Colum-VLA J L & Arts 101, 113 (1999); Kenneth D. Crews, The Law of Fair Use and the Illusion of Fair-Use Guidelines, 62 Ohio St L J 599, 630–31 (2001) (highlighting the minutely detailed CONFU multimedia guidelines for educational use).
  • 130Parchomovsky and Goldman, 93 Va L Rev at 1505–06 (cited in note 40) (detailing the Classroom Guidelines that found a place in the House Report accompanying the Copyright Act); Crews, 62 Ohio St L J at 637 (cited in note 129) (contending that the National Commission on New Technological Uses of Copyrighted Works (CONTU) guidelines of 1979 had strong legislative authority backing them, more so because they were built on the foundation of the judicial precedent of Williams & Wilkins Co v United States, 487 F2d 1345 (Ct Cl 1973)).
  • 131See Cambridge University Press v Patton, 769 F3d 1232, 1273 (11th Cir 2014), quoting Campbell, 510 US at 584 (“[T]o treat the Classroom Guidelines as indicative of what is allowable would be to create the type of ‘hard evidentiary presumption’ that the Supreme Court has cautioned against, because fair use must operate as a ‘sensitive balancing of interests.’”); Perfect 10, Inc v Amazon.com, Inc, 508 F3d 1146, 1163 (9th Cir 2007), quoting Campbell, 510 US at 577 (“We must be flexible in applying a fair use analysis; it ‘is not to be simplified with bright-line rules, for the statute, like the doctrine it recognizes, calls for case-by-case analysis.’”); Basic Books, Inc v Kinko’s Graphics Corp, 758 F Supp 1522, 1537 (SDNY 1991) (“We . . . refuse to hold that all unconsented anthologies are prohibited without a fair use analysis.”).
  • 132See Beebe, 156 U Pa L Rev at 621–22 (cited in note 1).
  • 133See id at 610, 614–15, 622.
  • 134See Pamela Samuelson, Unbundling Fair Uses, 77 Fordham L Rev 2537, 2541–42 (2009). Samuelson acknowledges her intellectual debt to earlier attempts to divide fair use into clusters. Id at 2541 n 27, citing Madison, 45 Wm & Mary L Rev at 1645–65 (cited in note 124), Alan Latman, Fair Use of Copyrighted Works, Study No 14, Subcommittee on Patents, Trademarks, and Copyrights of the Senate Committee on the Judiciary, 86th Cong, 2d Sess 8–14 (1960), Paul Goldstein, Fair Use in Context, 31 Colum J L & Arts 433, 439–41 (2008), and William F. Patry, The Fair Use Privilege in Copyright Law 64–259 (BNA 2d ed 1995).
  • 135Samuelson, 77 Fordham L Rev at 2541–42 (cited in note 134).
  • 136Sonny Bono Copyright Term Extension Act § 102, 112 Stat at 2827–28.
  • 137See 17 USC § 107 (codifying users’ fair use rights); 17 USC § 106 (identifying copyrights owners’ exclusive rights).
  • 13817 USC § 107(3).
  • 13917 USC § 107(4).
  • 140If the court failed to consider the effects of similar uses, our doctrine would create a perverse situation in which the same type of use would be fair for an early user, but not fair for a later user.
  • 141See BeVier, 30 Stan L Rev at 309 (cited in note 18).
  • 142Brown v Hartlage, 456 US 45, 60 (1982) (noting that the First Amendment functions “as the guardian of our democracy”); Alexander Meiklejohn, Political Freedom: The Constitutional Powers of the People 24–25 (Harper & Brothers 1960).
  • 143Robert D. Cooter, The Strategic Constitution 313 (Princeton 2000).
  • 144Richard A. Musgrave, The Theory of Public Finance: A Study in Public Economy 13–15 (McGraw-Hill 1959).
  • 145Cooter, The Strategic Constitution at 310–12 (cited in note 143).
  • 146Joel Rose, Music in Political Campaigns 101 (NPR Music, Feb 29, 2012), online at http://www.npr.org/sections/therecord/2012/02/29/147592568/music-in-political-campaigns
    -101 (visited Jan 22, 2016) (Perma archive unavailable).
  • 147Using Music in Political Campaigns: What You Should Know (ASCAP), archived at http://perma.cc/9MQ7-2E2S. Interestingly, Professor Rebecca Tushnet argues that pol­itical uses of songs are almost inherently transformative and thus should be judged fair uses in a fair use doctrine that protects transformative uses. See Tushnet, 114 Yale L J at 572–73 (cited at note 61). From our perspective, the transformativeness of the use is neither here nor there. A political use is prima facie fair, whether transformative or not.
  • 148471 US 539 (1985).
  • 149Id at 542–45.
  • 150Id at 555–60.
  • 151Id at 569.
  • 152Harper & Row, 471 US at 542–43.
  • 153Id at 567.
  • 154Id at 555.
  • 155Id at 552–55.
  • 156Harper & Row, 471 US at 542, 550–51 (quotation marks omitted).
  • 157See, for example, Joseph P. Bauer, Copyright and the First Amendment: Comrades, Combatants, or Uneasy Allies?, 67 Wash & Lee L Rev 831, 860–61 (2010) (describing how the Court “got it wrong” in Harper & Row); Pamela Samuelson, Copyright and Freedom of Expression in Historical Perspective, 10 J Intel Prop L 319, 320–21 (2003) (highlighting the critical response to the Court’s decision and the congressional response); Neil Weinstock Netanel, Copyright’s Paradox 63–65 (Oxford 2008); L. Ray Patterson, Free Speech, Copyright, and Fair Use, 40 Vand L Rev 1, 64–65 (1987); Samuelson, 77 Fordham L Rev at 2565 (cited in note 134).
  • 158See, for example, Bauer, 67 Wash & Lee L Rev at 860 (cited in note 157); Samuelson, 77 Fordham L Rev at 2564 (cited in note 134).
  • 159The figure a court would have to use in calculating the effect of a fair use finding on incentives is $25,000 because this represents the full value of the prepublication excerpt contract. The finding of fair use ex post would result in only the $12,500 cost of the ex post cancellation of the contract. However, if Harper & Row knew that all prepublication stories of the kind The Nation published would be found as fair use, and it knew that the ability to keep the manuscript under wraps until publication was limited, it would likely expect that it could not realize any gain from the sale of prepublication excerpts from political biographies.
  • 16017 USC § 107.
  • 161US Const Art I, § 8, cl 8.
  • 162Clifford S. Fishman, 4 Jones on Evidence: Civil and Criminal §§ 24:1, 24:6, 26:1, 27:1, 28:2 (West 7th ed 2000).
  • 16360 F3d 913 (2d Cir 1994).
  • 164Id at 915.
  • 165Id at 914–15.
  • 166Id at 929–32.
  • 167Texaco, 60 F3d at 940–41 (Jacobs dissenting).
  • 168The dissenting opinion by Judge Dennis Jacobs in Texaco considered this issue and concluded that the incentives provided by copyright law were sufficient to secure an exponential increase in scientific publications. Id at 940 (Jacobs dissenting). The majority opinion, on the other hand, did not examine in detail whether the publisher’s incentives would be considerably reduced by the infringing use in question. It observed only that “evidence concerning the effect of Texaco’s photocopying of individual articles . . . is of somewhat limited significance in determining and evaluating the effect . . . ‘upon the potential market for or value of’ the individual articles.” Id at 928.
  • 169954 F Supp 2d 282 (SDNY 2013).
  • 170755 F3d 87 (2d Cir 2014).
  • 171863 F Supp 2d 1190 (ND Ga 2012).
  • 172Google, 954 F Supp 2d at 285–86.
  • 173It is noteworthy that HathiTrust also wished to grant access to the works to print-disabled users and to ensure the preservation of the works in its database. See HathiTrust, 755 F3d at 91.
  • 174Becker, 863 F Supp 2d at 1201.
  • 175Google, 954 F Supp 2d at 293–94; HathiTrust, 755 F3d at 101, 103.
  • 176Becker, 863 F Supp 2d at 1242–43, 1363–64.
  • 177Patton, 769 F3d at 1275, revg Becker, 863 F Supp 2d 1190.
  • 178Patton, 769 F3d at 1271–72, 1283.
  • 179Google, 954 F Supp 2d at 286–87.
  • 180Id at 287–88.
  • 181Id at 293.
  • 18217 USC § 107.
  • 1839 F Cases 342 (CCD Mass 1841).
  • 184See id at 345.
  • 185Id at 344–45.
  • 186See generally Lance Lochner and Enrico Moretti, The Effect of Education on Crime: Evidence from Prison Inmates, Arrests, and Self-Reports, 94 Am Econ Rev 155 (2004); Stephen Machin, Olivier Marie, and Sunčica Vujić, The Crime Reducing Effect of Education, 121 Econ J 463 (2011).
  • 187See Princeton University Press v Michigan Document Services, Inc, 99 F3d 1381, 1383 (6th Cir 1996) (en banc); Blackwell Publishing, Inc v Excel Research Group, LLC, 661 F Supp 2d 786, 788 (ED Mich 2009); Basic Books, 758 F Supp at 1526.
  • 188See, for example, Princeton University Press, 99 F3d at 1391; Blackwell Publishing, 661 F Supp 2d at 794; Basic Books, 758 F Supp at 1534–37.
  • 189See generally Copyright Law Revision, HR Rep No 94-1476, 94th Cong, 2d Sess 68–70 (1976).
  • 190See id at 70–71. See also Crews, 62 Ohio St L J at 615–20 (cited in note 129).
  • 191See Parchomovsky and Goldman, 93 Va L Rev at 1506–08 (cited in note 40).
  • 192See, for example, Princeton University Press, 99 F3d at 1390–91; Texaco, 60 F3d at 919 n 5; Marcus v Rowley, 695 F2d 1171, 1178 (9th Cir 1983).
  • 193Loew’s Inc v Columbia Broadcasting System, 131 F Supp 165, 167 (SD Cal 1955).
  • 194Columbia Pictures Corp v National Broadcasting Co, 137 F Supp 348, 351–52 (SD Cal 1955).
  • 195Contrast Loew’s Inc, 131 F Supp at 183, with Columbia Pictures Corp, 137 F Supp at 353–54.
  • 196Loew’s, 131 F Supp at 176 n 31.
  • 197Campbell, 510 US at 572–73.
  • 198Id at 583.
  • 199Id at 592.
  • 200Id at 579.
  • 201Campbell, 510 US at 580.
  • 202Id.
  • 203Michelle B. Lee and Erin M. Hickey, The Changing Landscape of Fair Use: ‘Transformative’ Is Key (Law360, Apr 22, 2015), archived at http://perma.cc/3VY6-8ZJ3 (noting that “courts began to increasingly emphasize whether a use was ‘transformative’ after the Campbell decision”). See also Sag, 73 Ohio St L J at 76–77 (cited in note 8).
  • 204See, for example, Burnett v Twentieth Century Fox Film Corp, 491 F Supp 2d 962, 968–69 (CD Cal 2007) (interpreting Campbell to hold that if a work is likely to be reasonably perceived as a parody, then it is a fair use); Mattel Inc v Walking Mountain Productions, 353 F3d 792, 806 (9th Cir 2003); Suntrust Bank v Houghton Mifflin Co, 268 F3d 1257, 1268 (11th Cir 2001); Lyons Partnership, LP v Giannoulas, 14 F Supp 2d 947, 955 (ND Tex 1998). See also Samuelson, 77 Fordham L Rev at 2550–51 (cited in note 134); Pamela Samuelson, Possible Futures of Fair Use, 90 Wash L Rev 815, 821 (2015) (“Although the Court in Campbell expressly declined to adopt a presumption that parodies of copyrighted works were fair uses, the parody case law after Campbell has resulted in many fair use rulings. . . . [This trend suggests] that parodies are de facto presumptively fair.”) (citations omitted).
  • 205See, for example, Google, 954 F Supp 2d at 290–91; HathiTrust, 755 F3d at 97–98; Cariou, 714 F3d at 698–99; Perfect 10, 508 F3d at 1164–67; Blanch, 467 F3d at 251–53; Kelly v Arriba Soft Corp, 336 F3d 811, 818–20 (9th Cir 2003); On Davis v Gap, Inc, 246 F3d 152, 174–76 (2d Cir 2001); Castle Rock Entertainment, Inc v Carol Publishing Group, Inc, 150 F3d 132, 142–43 (2d Cir 1998).
  • 206See, for example, Brownmark Films, LLC v Comedy Partners, 800 F Supp 2d 991, 1001–02 (ED Wis 2011) (holding that South Park’s partial use of a music video was parody and fair use); Burnett, 491 F Supp 2d at 971–72, 974–75; Suntrust Bank, 268 F3d at 1276; Leibovitz v Paramount Pictures Corp, 137 F3d 109, 117 (2d Cir 1998); Lyons Partnership, 14 F Supp 2d at 955 (holding that a mascot’s use of the “Barney dance” at sporting events was a parody and fair use).
  • 207See, for example, Blanch, 467 F3d at 254–55, 259 (considering the work more of a satire, yet favoring the defendant after considering the four factors); Williams v Columbia Broadcasting Systems, Inc, 57 F Supp 2d 961, 969 (CD Cal 1999); Columbia Pictures Industries, Inc v Miramax Films Corp, 11 F Supp 2d 1179, 1187 (CD Cal 1998); Dr. Seuss Enterprises, LP v Penguin Books USA, Inc, 924 F Supp 1559, 1569–70 (SD Cal 1996) (holding that a work’s nature as satire was insufficient for a fair use defense, but that parody was sufficient).
  • 208924 F Supp 1559 (SD Cal 1996).
  • 209Dr. Seuss was the pen name of Theodor S. Geisel. Id at 1561.
  • 210Id.
  • 211Id at 1569.
  • 212Dr. Seuss Enterprises, LP v Penguin Books USA, Inc, 109 F3d 1394, 1401–03 (9th Cir 1997).
  • 213Andrew O’Hehir, Tina Fey’s Sarah Palin: When “Saturday Night Live” Finally Got Political Satire Right (Salon, Feb 14, 2015), archived at http://perma.cc/SYQ8-ENYS.
  • 214See Suntrust Bank, 268 F3d at 1267–76; Leibovitz, 137 F3d at 114–17.
  • 215Sony, 464 US at 456.
  • 216A&M Records, Inc v Napster, Inc, 114 F Supp 2d 896, 904 (ND Cal 2000).
  • 217Id at 896, affd in part and revd in part, 239 F3d 1004 (9th Cir 2001), on remand, 2001 WL 777005 (ND Cal), affd, 284 F3d 1091 (9th Cir 2002).
  • 218A&M Records, 114 F Supp 2d at 901.
  • 219See A&M Records, 239 F3d at 1011–12.
  • 220Id at 1017.
  • 221Id at 1015.
  • 222Id at 1011, 1013.
  • 223A&M Records, 239 F3d at 1019.
  • 224Id at 1021.
  • 225Id at 1019.
  • 226A&M Records, 114 F Supp 2d at 904–05.
  • 227Id at 915–16.
  • 228A&M Records, 239 F3d at 1019.
  • 229See, for example, Raymond Shih Ray Ku, The Creative Destruction of Copyright: Napster and the New Economics of Digital Technology, 69 U Chi L Rev 263, 298–311 (2002); Lior Jacob Strahilevitz, Charismatic Code, Social Norms, and the Emergence of Cooperation on the File-Swapping Networks, 89 Va L Rev 505, 576–82 (2003).
  • 230See Gaia Bernstein, When New Technologies Are Still New: Windows of Opportunity for Privacy Protection, 51 Vill L Rev 921, 944 (2006); Lawrence Lessig, Free Culture: How Big Media Uses Technology and the Law to Lock Down Culture and Control Creativity 66–68, 124–26 (Penguin 2004); Steven A. Hetcher, The Music Industry’s Failed Attempt to Influence File Sharing Norms, 7 Vand J Enter L & Prac 10, 11 (2004); Ann Bartow, Electrifying Copyright Norms and Making Cyberspace More Like a Book, 48 Vill L Rev 13, 95 (2003); Dan L. Burk and Julie E. Cohen, Fair Use Infrastructure for Rights Management Systems, 15 Harv J L & Tech 41, 65 (2001).
  • 231180 F3d 1072 (9th Cir 1999).
  • 232Pub L No 102-563, 106 Stat 4237, codified at 17 USC § 1001 et seq.
  • 233Recording Industry Association of America, 180 F3d at 1079.
  • 234Id.
  • 235See generally Karen Hellekson and Kristina Busse, eds, Fan Fiction and Fan Communities in the Age of the Internet (McFarland 2006) (describing the plethora of fan fiction stories derived from different original works).
  • 236See, for example, Rebecca Tushnet, Legal Fictions: Copyright, Fan Fiction, and a New Common Law, 17 Loyola LA Enter L Rev 651, 655 (1997).
  • 237Anupam Chander and Madhavi Sunder, Everyone’s a Superhero: A Cultural Theory of “Mary Sue” Fan Fiction as Fair Use, 95 Cal L Rev 597, 615–17 (2007).
  • 238Id at 616.
  • 239See, for example, Bauer, 67 Wash & Lee L Rev at 873–78 (cited in note 157); David McGowan, Why the First Amendment Cannot Dictate Copyright Policy, 65 U Pitt L Rev 281, 287–91 (2004); Erwin Chemerinsky, Balancing Copyright Protections and Freedom of Speech: Why the Copyright Extension Act Is Unconstitutional, 36 Loyola LA L Rev 83, 95–97 (2002); Neil Weinstock Netanel, Locating Copyright within the First Amendment Skein, 54 Stan L Rev 1, 12–20 (2001); Rebecca Tushnet, Copyright as a Model for Free Speech Law: What Copyright Has in Common with Anti-Pornography Laws, Campaign Finance Reform, and Telecommunications Regulation, 42 BC L Rev 1, 4–27 (2000); Patterson, 40 Vand L Rev at 33–36 (cited in note 157); Eugene Volokh, Freedom of Speech and Intellectual Property: Some Thoughts after Eldred, 44 Liquormart, and Bartnicki, 40 Houston L Rev 697, 702–10 (2003).
  • 240See Sterk, 94 Mich L Rev at 1198–1204, 1207–08 (cited in note 12); William M. Landes and Richard A. Posner, The Economic Structure of Intellectual Property Law 13 (Belknap 2003).
  • 241See Oren Bracha, The Folklore of Informationalism: The Case of Search Engine Speech, 82 Fordham L Rev 1629, 1666 (2014); Harry H. Wellington, On Freedom of Expression, 88 Yale L J 1105, 1109 (1979).
  • 242See generally Christina Bohannan, Copyright Infringement and Harmless Speech, 61 Hastings L J 1083 (2010); Alan E. Garfield, The Case for First Amendment Limits on Copyright Law, 35 Hofstra L Rev 1169 (2007); Netanel, Copyright’s Paradox (cited in note 157); Patterson, 40 Vand L Rev 1 (cited in note 157); Bauer, 67 Wash & Lee L Rev 831 (cited in note 157).
  • 243Tushnet, 114 Yale L J at 540 (cited in note 61).
  • 244Julie E. Cohen, Intellectual Privacy and Censorship of the Internet, 8 Seton Hall Const L J 693, 693 (1998).
  • 245See generally Chemerinsky, 36 Loyola LA L Rev 83 (cited in note 239); C. Edwin Baker, First Amendment Limits on Copyright, 55 Vand L Rev 891 (2002); Jed Rubenfeld, The Freedom of Imagination: Copyright’s Constitutionality, 112 Yale L J 1 (2012); Netanel, 54 Stan L Rev 1 (cited in note 239); Yochai Benkler, Free as the Air to Common Use: First Amendment Constraints on Enclosure of the Public Domain, 74 NYU L Rev 354 (1999).
  • 246Melville B. Nimmer, Does Copyright Abridge the First Amendment Guarantees of Free Speech and Press?, 17 UCLA L Rev 1180, 1181 (1970). Another notable article published that same year was Paul Goldstein, Copyright and the First Amendment, 70 Colum L Rev 983 (1970), which touched on many of the same issues.
  • 247See generally Netanel, Copyright’s Paradox (cited in note 157).
  • 248See generally Netanel, 60 UCLA L Rev 1082 (cited in note 87); Netanel, 54 Stan L Rev 1 (cited in note 239); Neil Weinstock Netanel, Market Hierarchy and Copyright in Our System of Free Expression, 53 Vand L Rev 1879 (2000); Netanel, 51 Vand L Rev 217 (cited in note 98); Neil Weinstock Netanel, Copyright and a Democratic Civil Society, 106 Yale L J 283 (1996).
  • 249See Netanel, Copyright’s Paradox at 109–53 (cited in note 157).
  • 250Id at 59–60.
  • 251See text accompanying notes 82–86.
  • 252Netanel, 60 UCLA L Rev at 1085–86 (cited in note 87).
  • 253Id at 1112–13. Netanel did note several troubling aspects of the Supreme Court’s ruling in Harper & Row. Id at 1105.
  • 254Tushnet, 114 Yale L J at 587 (cited in note 61).
  • 255Id at 588.
  • 256See generally Brett M. Frischmann, Evaluating the Demsetzian Trend in Copyright Law, 3 Rev L & Econ 649 (2007); Brett Frischmann, Spillovers Theory and Its Conceptual Boundaries, 51 Wm & Mary L Rev 801 (2009).
  • 257See generally Brett M. Frischmann and Mark A. Lemley, Spillovers, 107 Colum L Rev 257 (2007).
  • 258See Frischmann, 3 Rev L & Econ at 669–72 (cited in note 256).
  • 259See Lemley, 83 Tex L Rev at 1046–69 (cited in note 99).
  • 260See Michael J. Meurer and Katherine J. Strandburg, Patent Carrots and Sticks: A Model of Nonobviousness, 12 Lewis & Clark L Rev 547, 558–65 (2008).
  • 261See Michael W. Carroll, One Size Does Not Fit All: A Framework for Tailoring Intellectual Property Rights, 70 Ohio St L J 1361, 1419–21 (2009).
  • 262See Giuseppe Dari-Mattiacci, Negative Liability, 38 J Legal Stud 21, 22 (2009); Harold Demsetz, Toward a Theory of Property Rights, 57 Am Econ Rev 347, 347–57 (1967) (arguing that property rights arise to effect internalization of externalities, both positive and negative); Ariel Porat, Private Production of Public Goods: Liability for Unrequested Benefits, 108 Mich L Rev 189, 190 (2009).
  • 263Dari-Mattiacci, 38 J Legal Stud at 22 (cited in note 262).
  • 264Id.
  • 265Demsetz, 57 Am Econ Rev at 349–50 (cited in note 262).
  • 266See R. Polk Wagner, Information Wants to Be Free: Intellectual Property and the Mythologies of Control, 103 Colum L Rev 995, 1024 (2003).
  • 267See Coase, 3 J L & Econ at 15 (cited in note 32); Demsetz, 57 Am Econ Rev at 354–59 (cited in note 262); Louis Kaplow and Steven Shavell, Property Rules versus Liability Rules: An Economic Analysis, 109 Harv L Rev 713, 741–44 (1996).
  • 268Demsetz, 57 Am Econ Rev at 353–54 (cited in note 262).
  • 269See Joseph P. Liu, Regulatory Copyright, 83 NC L Rev 87, 129–39 (2004) (proposing a regulatory approach to copyright, instead of a property rights approach); Kaplow and Shavell, 109 Harv L Rev at 741–44 (cited in note 267) (arguing that, in some instances, liability rules may be less costly than property rules for copyright enforcement).
  • 270See Part II.B. See also Lunney, 82 BU L Rev at 979, 993–94 (cited in note 11).
  • 271Harold Demsetz, Frischmann’s View of “Toward a Theory of Property Rights”, 4 Rev L & Econ 127, 128–29 (2008).
  • 272See generally Gordon, 50 J Copyright Society USA 149 (cited in note 7).
  • 273Beebe, 156 U Pa L Rev at 552 (cited in note 1) (noting that “the concept of fair use . . . ha[s] attracted an enormous amount of scholarly attention”); Goldstein, 31 Colum J L & Arts at 433 (cited in note 134) (describing fair use as “the great white whale of American copyright law”).
  • 274Fisher, 101 Harv L Rev at 1695–96 (cited in note 97).
  • 275Id.
  • 276Id at 1715–17.
  • 277Id.
  • 278Fisher, 101 Harv L Rev at 1715–17 (cited in note 97).
  • 279Id at 1739.
  • 280Id at 1769–72.
  • 281Madison, 45 Wm & Mary L Rev at 1622–77 (cited in note 124).
  • 282See Parchomovsky, 3 Legal Theory at 349 (cited in note 37).
  • 283Id.
  • 284Id at 370–71.
  • 285Id.
  • 286Parchomovsky, 3 Legal Theory at 370–71 (cited in note 37).
  • 287Madison, 45 Wm & Mary L Rev at 1687 (cited in note 124).
  • 288Id at 1635.
  • 289Id.
  • 290See generally Lunney, 82 BU L Rev 975 (cited in note 11).
  • 291See id at 996, 1022–23.
  • 292See id at 976–77, 992–96.
  • 293Id at 982, 991–92.
  • 294Lunney, 82 BU L Rev at 981–85, 999 (cited in note 11).
  • 295Others have suggested more limited use of the presumption of fair use. See generally, for example, Greenberg, 85 U Colo L Rev 53 (cited in note 8); Ned Snow, Proving Fair Use: Burden of Proof as Burden of Speech, 31 Cardozo L Rev 1781 (2010); Samuelson, 77 Fordham L Rev 2537 (cited in note 134).
  • 296Samuelson, 77 Fordham L Rev at 2541–42 (cited in note 134).
  • 297Id at 2617–18.
  • 298Id (citation omitted).