The Evolving Common Law Doctrine of Copyright Misuse: A Unified Theory and Its Application to Software
By Brett Frischmann and Dan Moylan
Abstract
This Article explores the common law defense of copyright misuse from a variety of angles in an effort to refine and unify existing views. The unified model that emerges is then applied to software copyright, addressing the tension that software creates within copyright law as well as between copyright, patent, and antitrust law. Part II develops a jurisprudential model for understanding the substantive relationship between the copyright misuse doctrine and copyright, patent, and antitrust laws, and the procedural approaches taken by courts when formulating and applying misuse principles—per se rules and the rule of reason. Part III examines four Supreme Court cases that provide guidance for the application of misuse principles in the copyright context. It then turns to an analysis of the application of copyright misuse in the federal courts of appeals. These discussions enable a distillation of guiding principles from the case law in an attempt to clarify the “current state of the copyright misuse doctrine.” Part IV applies the principles derived in Parts II and III to software copyrights and proposes a per se rule against licensing restrictions upon reverse engineering that complements antitrust-based misuse and the fair use doctrine. Rather than attempt to provide a comprehensive set of public policy-based misuse rules, this Article instead presents a single rule as an illustration of how further doctrinal development might proceed.
Intellectual property misuse is a common law defense to infringement that derives from the equitable doctrine of “unclean hands.”1 The defendant raising the defense need not be affected by the plaintiff’s inequitable conduct.2 However, in some jurisdictions, a defendant may be barred from raising the defense if the defendant’s hands are unclean.3 When defendants successfully use misuse defenses, the courts bar immediate relief from the “guilty” plaintiffs. However, the misuse doctrine does not bar future reliance on the courts. The intellectual property owners may return to court once they have “purged” the misuse, for example, by striking anticompetitive provisions in their licensing agreements.4
Judicial creation of intellectual property misuse doctrines has been piecemeal, beginning with patent misuse and only recently moving into copyright misuse. Both trademark and trade secret misuse remain subjects for academic discussion without practical force in the courts.5 Although intellectual property misuse has been mentioned or alluded to in its various forms by courts for over a century, the Supreme Court did not establish the patent misuse doctrine until 1954 in Morton Salt Co. v. G.S. Suppiger.6 Since then, patent misuse has developed significantly. Yet copyright misuse, which was mentioned in dictum in the Morton Salt opinion,7 remained in limbo until the 1990 Lasercomb America Inc. v. Reynolds decision of the Fourth Circuit, which expressly upheld the doctrine’s existence.8 Since 1990, both the Fifth and Ninth Circuits have established copyright misuse as a viable defense in their jurisdictions.9 The Supreme Court and the remaining circuit courts have not established this defense, leaving the doctrine’s fate uncertain.
Today, patent misuse is a well-established doctrine where courts generally apply antitrust principles to determine whether a patentee’s use is misuse.10 In fact, Congress amended the patent law to require a showing of “market power in the relevant market for the patent or patented product” for a misuse defense to be successful in “tying” cases—when a patent owner conditions a license or the sale of a patented product on the “acquisition of a license to rights in another patent or purchase of a separate product.”11 It is not surprising that patent misuse depends on patentees’ attempts to affect market dynamics, since the superceding public policy behind the patent system is to promote the creation and dissemination of utilitarian or functional innovation. This innovation is a primary upstream force behind downstream market dynamics.12
Copyright misuse, on the other hand, is far from well-established. Only recently have federal courts of appeals begun to apply affirmatively this doctrine. In three court of appeals cases, the courts have explicitly relied on public policy in lieu of antitrust principles in evaluating the misuse defense.13 Other courts of appeals, particularly the Seventh Circuit, have been less receptive to copyright misuse grounded in public policy. In Saturday Evening Post Co. v. Rumbleseat Press, Inc.,14 Chief Judge Posner extended his rationale regarding antitrust analysis in patent misuse cases to copyright misuse:
‘If misuse claims are not tested by conventional antitrust principles, by what principles shall they be tested? Our law is not rich in alternative concepts of monopolistic abuse; and it is rather late in the day to try to develop one without in the process subjecting the rights of patent holders to debilitating uncertainty.’ This point applies with even greater force to copyright misuse, where the danger of monopoly is less.15
Are claims of copyright misuse and violation of antitrust law really two sides of the same coin, both to be determined with “conventional antitrust principles”? Or, as many suggest, does copyright misuse embody more than market-based concerns? If so, by what principles should public policy-based misuse claims be tested? Should courts fashion common law rules to limit misuse? Can they do so without generating friction with the intellectual property and antitrust laws? Finally, even where antitrust principles govern, is the danger of monopoly really less for copyrighted innovations than for patented innovations?
This Article examines the copyright misuse doctrine and the principles with which courts evaluate misuse defenses. It is important to remember that copyright misuse is an evolving common law doctrine and that this Article provides only a static “snapshot” of its current state. However, when considering what copyright misuse ought to be in the future, one should take into account what the law is today. Accordingly, we devote substantial attention to the influential case law before turning to general principles.
Part II provides a brief introduction to the jurisprudential functions of the copyright misuse doctrine. It first develops a schematic model for understanding the jurisprudential relationship between the copyright misuse doctrine and copyright, patent, and antitrust laws. Next, it considers two approaches to formulating and applying misuse principles: per se rules and the rule of reason.
Part III analyzes the case law in the Supreme Court and the federal courts of appeals. Part III.A examines four foundational Supreme Court cases: Morton Salt, United States v. Paramount Pictures, Inc.,16 United States v. Loew’s, Inc.,17 and Broadcast Music, Inc. v. Columbia Broadcasting System, Inc.18 These cases provide lower courts guidance on what misuse principles should be applied in the copyright context. Parts III.B and III.C consider the case law in the federal courts of appeals, looking more carefully at the express application of copyright misuse as a defense to infringement to determine on what principles courts refuse to enforce the copyright, or apply the doctrine and fail to find misuse. Part III.B covers precedential case law, and Part III.C covers persuasive case law. Although the case law and legal commentary seem in disarray, Part III.D distills a set of guiding principles for evaluating copyright misuse and provides our “unified theory” of copyright misuse. It concludes that courts should ask first whether a challenged action amounts to per se misuse by looking to the facts for evidence of blatantly egregious conduct. Two sets of per se rules may be fashioned by the courts. The first type identifies misuse violating the public policy behind the intellectual property grant while the second type identifies misuse violating the antitrust laws. If a challenged action does not fit within either set of per se rules, we suggest that the courts then engage in a rule of reason analysis similar to that used by courts in the patent misuse and antitrust contexts.
Part IV applies the jurisprudential model and procedural principles developed in Part II to the unique context of software. Part IV.A explores the nature of software both as an innovation and in its wider landscape, emphasizing the ways that software challenges standard assumptions about copyright law. Part IV.B focuses on how the nature of software creates a substantive need for common law using the jurisprudential model from Part II. Part IV.B next turns to existing doctrines, namely fair use and antitrust-based copyright misuse, identifying their limitations. We argue that these doctrines form part of an answer to the challenges posed by software, but that a need for doctrinal development remains. Finally, Part IV.C proposes refinements to the current approach to software, recommending that narrow public policy-based per se rules supplement a core antitrust-based defense. In particular, we advocate one per se rule against licensing restrictions that bar reverse engineering, although we envision that other per se rules might eventually be added along the guidelines we provide.
This Article provides a unified approach for development of the copyright misuse doctrine. The approach both recognizes a substantive need for common law using the jurisprudential model and recommends carefully crafted rules to meet that need.
Before exploring the copyright misuse case law, it is helpful to consider the justifications for and limitations of the common law doctrine.19 When courts formulate and apply misuse principles, whether based in equity or antitrust, they affect the statutory scheme created by Congress. At first glance, only the copyright statute seems affected. However, as the case law analysis in Part III demonstrates, other areas of law, especially patent and antitrust, are implicated. This Part briefly develops a schematic model for understanding the jurisprudential relationship between the copyright misuse doctrine and copyright, patent, and antitrust laws. Next, this Part considers two approaches to formulating and applying misuse principles: per se rules and the rule of reason.
The misuse doctrine is a mechanism that operates on at least three distinct levels. First, it gives courts the flexibility to “fill in gaps” left in statutory law; we label this the corrective function of the misuse doctrine.20 Second, the misuse doctrine allows courts to coordinate related and interdependent bodies of law; we label this the coordination function. Third, it allows courts to safeguard the public interest generally; we label this the safeguarding function. This subsection briefly explains the substantive nature of these three functions.
The corrective function of the misuse doctrine involves both judicial interpretation of express statutory language and congressional intent, and judicial lawmaking where gaps in the substantive law exist.21 For the most part, judges are expected to exercise these functions within reasonable discretion under the copyright, patent, and antitrust statutes. For example, much of the federal antitrust law that exists today derives from decades of dynamic common law-making by the federal courts.22 The broad precepts of the Sherman and Clayton Acts have produced a complex set of rules to effectuate the Acts’ procompetitive agenda in light of changing social, technological, and economic factors under a large variety of factual settings.23 Courts have also historically formulated common law in the patent and copyright areas.24 Consider, for example, the fair use doctrine. Fair use was originally a common law defense before Congress stepped in to codify the doctrine in 17 U.S.C. § 107.25 The fair use example highlights the fact that Congress can always, and sometimes does, supercede common law by passing a statute.26 Thus, corrective common law-making can also be viewed as a signal to Congress that a gap exists.
As the label we attach suggests, the gap-filling and interpretive functions exercised by courts applying corrective common law are internal to the statutory body of law at issue. Courts correct legal ambiguities (of varying scope) within a statutory scheme.27 For example, the set of judicially-crafted rules governing vertical price fixing derives from the express and implied policies underlying the Sherman Act, and the judicially-crafted fair use doctrine embodies equitable principles concerning the societal trade-off made via the copyright grant. In either case, the judicial power is cabined by the express and implied scope of the statute being interpreted or “filled in.”28 As will be explored in significant detail in Parts III and IV, the copyright misuse doctrine may be a vehicle for correcting various ambiguities or gaps in the copyright law, particularly as it is applied to software. For example, the inclusion of software within copyrightable subject matter exposes the absence of a disclosure requirement in the copyright law. While the traditional expression gaining statutory protection is naturally disclosed when encountered by the public—consider, for example, books, songs, and paintings, among others—the expression in the source and object code of software is not, jeopardizing the societal trade-off established by the copyright statute. The copyright misuse doctrine may fill the gap in the statute and protect public access to copyrighted expression.
The coordination function of the misuse doctrine involves the reconciliation of external (or interstatutory) relationships between the related and interdependent bodies of antitrust, copyright, and patent law. While similar to the corrective function in that coordination involves statutory interpretation and gap-filling, it operates externally to any single body of law. The express or implied statutory objectives derived internally—from myopic consideration of a single body of law—may not lead to principled rules at the interfaces.29 Common law misuse allows courts to develop rules that evolve dynamically. For example, in the patent misuse context, courts have coordinated patent and antitrust law. Over the course of the twentieth century, courts, as well as legislators, enforcement officials, and commentators, have struggled to resolve “conflicts” between the two bodies of law in a variety of ways, ranging from near preemption in favor of patentees to strict antitrust-based limits on patentees’ behavior in the marketplace to a moderated contemporary approach.30 Today, the patent misuse doctrine’s reliance on antitrust principles reinforces the importance of the market mechanism in achieving the public policies embodied in both statutory schemes.31
In the copyright misuse context, courts coordinate both copyright and antitrust law as well as copyright and patent law. As will be seen in Parts III and IV, coordinating copyright and antitrust law leads to a copyright misuse doctrine that is identical to the patent misuse doctrine—both rely on antitrust principles for a finding of misuse. However, the coordination of copyright and patent law is a more complicated task because it involves important policy decisions as to the appropriate social cost-benefit trade-off for promoting development of different types of innovation.32 For the most part, this task should be left to Congress because of its constitutional authority under the Intellectual Property Clause and its presumptive institutional competence in developing policy. However, judges may be forced to coordinate, or at least to signal to Congress that coordination is necessary, where friction between the two intellectual property regimes arises.33 As will be explored in more detail in Part IV,34 copyright protection of computer software, which derives its economic value from the functional ends it facilitates rather than its expressive content, gives rise to such friction.35
Finally, when exercising the safeguarding function, courts play their traditional role as “balancer of equities” in specific cases. Misuse doctrines grew from the equitable notion that a plaintiff with “unclean hands” could not use the courts to obtain redress. In addition to balancing the equities between the parties to a case, courts safeguard the public policy underlying the governing statutory scheme where internal statutory protections are lacking and an external conflict is not implicated.36 For example, copyright holders who use their copyrights to gain leverage through licensing provisions that broaden the scope of their copyright may be misusing their copyright even if the leveraging is insufficient to raise antitrust concerns.37
The final Part of this Article demonstrates that the copyright misuse doctrine may apply in the computer software context under each jurisprudential function. First, it may correct for the absence of an explicit disclosure requirement in copyright law.38 Second, it may be used to coordinate copyright, patent, and antitrust law. And finally, it may safeguard several interrelated public interests underlying the copyright laws: (1) promoting creative transfers; (2) preserving the unlimited reuse of ideas and unprotected expression; and (3) stimulating downstream innovation and competition in software development. In addition to the “substantive” demand for correction, coordination, and equitable safeguards,39 the case-by-case nature of judicial rulemaking has a “procedural” component that must also be considered.
This section provides a brief explanation of the types of rules that judges create when formulating the substantive common law discussed above. The creation of common law depends upon case by case adjudication and the evolution of judicial precedent. The legally binding nature or precedential value of a substantive rule depends in part upon the manner in which it is applied in subsequent cases.
“Per se” and “rule of reason” describe procedural approaches to judicial decisionmaking. Numerous formulations and explanations of what these labels mean can be found in the antitrust case law and literature.40 Generally, they differentiate the degree to which judges entertain evidence and certain defenses when applying the law in a given case.41 In theory, an absolute per se rule would preclude any defense and restrict the presentation of evidence to what is needed to characterize the alleged conduct as fitting within the per se classification. In practice, defenses creep into the characterization.42 At the opposite extreme from per se rules is the rule of reason approach. The rule of reason involves a comprehensive, fact-intensive inquiry where, in the end, courts assess the reasonableness of contested conduct.43 The “quick look” approach is something between strict per se and rule of reason, either a slightly more involved consideration of evidence and defenses in the per se context or a slightly less involved balancing approach.44
In Broadcast Music, an important case discussed fully in Part III, the Supreme Court mandated that courts apply a per se rule when “the practice facially appears to be one that would always or almost always tend to restrict competition and decrease output.”45 Judge Posner reframed the Broadcast Music test in General Leaseways: “if the elimination of competition is apparent on a quick look, without undertaking the kind of searching inquiry that would make the case a [r]ule of [r]eason case in fact if not in name, the practice is illegal per se.”46 In General Leaseways, the Seventh Circuit concluded “on the basis of the record compiled in the preliminary injunction hearing” that the market division at issue was a per se violation of the Sherman Act.47
In this Article, per se rules are similarly narrow legal constructions that fulfill the “facially apparent” and “always or almost always” requirements.48 On the other hand, the rule of reason encompasses a full consideration of the evidence and relevant defenses in a balancing or reasonableness inquiry. This Article rejects a rule of reason approach outside the antitrust context because the requisite balancing of public interests—market and non-market-based—involves policy decisions best left to the legislative branch. By the same token, courts should also be wary of creating per se rules where the corrective, coordination, and safeguarding functions involve a complex balancing of public interests.
The unified approach to assessing misuse, presented below in Part III.D.3, argues that (1) courts should apply traditional antitrust principles, consisting of per se rules and the rule of reason, when coordinating copyright and antitrust or gap-filling along the interface of these bodies of law; and (2) courts should avoid any form of reasonableness or balancing test and should constrain themselves to formulating per se rules when filling gaps in the copyright laws, coordinating copyright and patent law, or safeguarding the policies underlying the copyright laws. This approach limits judicial lawmaking to the formulation of per se rules based on recognizable and egregious factual settings that implicate the concerns mentioned in Part II.A. A per se rule is a clearer signal to Congress and the public of a gap or conflict in the law; additionally, per se rules more clearly delineate the intellectual property rights themselves.49 A reasonableness or balancing test inquiry would be problematic because judges are left to value and balance public interests in the absence of statutory guidance against other public interests recognized in the statute. Restricting common law-making, particularly of common law defenses to statutorily-recognized infringement, to per se rules limits the frequency and effects of judicial activism to egregious cases.50
Although intellectual property misuse can be traced back to early notions of equity, the doctrine has experienced different evolutionary tracts in the patent, trademark, and copyright areas. Patent misuse is a technical defense to infringement that relies largely on antitrust analysis for resolution.51 Trademark misuse, on the other hand, remains an inchoate collection of principles largely based in equity and is not a recognized defense in most jurisdictions.52 Copyright misuse resembled trademark misuse until 1990, when the Fourth Circuit recognized the former doctrine.53 Since 1990, the common law development of copyright misuse has yielded two schools—(1) courts that apply a “public policy approach” to assessing misuse, and (2) courts that apply an “antitrust approach.” After reviewing the significant case law, we analyze these two schools of thought and explain how they can be reconciled.
In an effort to establish what the copyright misuse doctrine is today, this Part takes a careful look at the precedential case law.54 Accordingly, Part III.A analyzes four foundational Supreme Court cases that impact the evolution of the copyright misuse: Morton Salt, Paramount Pictures, Loew’s, and Broadcast Music. These cases provide lower courts guidance on misuse principles that can be applied in the copyright context.55 Parts III.B and III.C consider the case law in the federal courts of appeals, looking more carefully at the express application of copyright misuse as a defense to infringement to determine on what principles courts refuse to enforce the copyright or apply the doctrine and fail to find misuse. Part III.B focuses on the precedential case law where a copyright misuse defense is expressly established, and Part III.C focuses on persuasive case law where a copyright misuse defense is considered but not found to exist. Each section ends with interim conclusions to provide a comprehensive outlook on how the cases relate to each other. Finally, Part III.D extracts guiding principles from the case law in an attempt to clarify where the copyright misuse doctrine stands today.
Morton Salt56 lays the foundation for copyright misuse. The reasoning employed by the Court addresses patent misuse, while the dicta and commentary address copyright misuse. Moreover, the Court cited two cases “for application of the like doctrine in the case of copyright,” indicating the Court’s recognition that misuse principles can be applied broadly.57
In 1942, the Supreme Court firmly established the patent misuse doctrine.58 Although courts have applied patent misuse since 1917,59 pre-1942 patent misuse resembles today’s copyright misuse—evolving doctrinally with an uncertain future. Morton licensed its patented salt-depositing machine with a condition that licensees exclusively use Morton’s salt tablets. The Court found that using the patent to restrain competition in a market for unpatented goods (i.e., the salt tablets) was patent misuse. The Court stressed that the public policy behind the patent system involved a delicate balance between the social benefits of improved progress in “Science and the useful Arts” and the social costs of granting a “limited monopoly.”60 Tipping the balance by using the “granted monopoly” to “secure an exclusive right or limited monopoly not granted” is forbidden and contrary to the public policy behind the system.61 Courts of equity “may rightly withhold assistance from such use of the patent by declining to entertain a suit for infringement, and should do so at least until it is made to appear that the improper practice has been abandoned and that the consequences of the [patent misuse] have been dissipated.”62
The Morton Salt misuse defense was based in equity and did not mirror the statutory antitrust law. Notably, the Court distinguished the patent misuse defense from traditional antitrust analysis under the Clayton Act and reversed the Seventh Circuit opinion which relied on a finding that the patent did not substantially lessen competition for salt tablets.63 Thus, although patent misuse in tying cases was later modified by legislation to resemble an antitrust-based defense,64 the original basis in equity may also apply to copyright misuse today. Moreover, the two copyright cases cited by the Court are both copyright infringement actions wherein a defense based in equity prevailed.65 Neither case involved antitrust law nor anticompetitive behavior that would rise to the level of an antitrust violation today.
Although Paramount Pictures66 did not involve a copyright misuse defense, it sets the stage for more intensive judicial scrutiny at the interface of copyright and antitrust. The Court relied on the Morton Salt rationale and emphasized the inherent dependence of the success of the copyright laws in achieving their ends on efficient operation of the market. The opinion demonstrates interstatutory coordination at the interface of antitrust and copyright law.
In a multi-faceted antitrust action brought by the United States against various companies in the motion picture industry,67 the Court addressed the issue of block-booking of copyrighted material for the first time. Block-booking is a particular tying arrangement where the sale or license of one or more copyrighted works is conditioned on the sale or license of other copyrighted works. The Court condemned the practice as a means used to enlarge the scope of individual copyrights and directly cited Morton Salt for support.
Paramount Pictures focused extensively on the public policy behind the issuance of an intellectual property right and the carefully prescribed balance achieved by the statutory scheme.68 Moreover, the Court went further and recognized that the public policy behind granting intellectual property is dependent upon successful operation of the market mechanism, which allows consumers to differentiate among innovations based upon quality.69 This concept acknowledges a role for the market in shaping the scope and value of an intellectual property right and reaffirms the role of antitrust law in protecting market dynamics. The citation to Morton Salt acknowledges (or suggests) that the misuse doctrine also plays a role in cabining the scope of an intellectual property right.
Along with Paramount Pictures, the Loew’s70 opinion is often cited as support for copyright misuse claims.71 While Loew’s clearly supports extension of antitrust-related presumptions to the copyright misuse context, it does not appear to support extension of misuse principles from patent to copyright.72
Loew’s presented the Supreme Court with the issue of whether block-booking violated section 1 of the Sherman Act. The Court upheld the lower court’s finding that the practice was an antitrust violation.73 Relying on International Salt74 and Paramount Pictures, the Court recognized a presumption that a copyright confers sufficient economic leverage to induce purchasers to go along with a tying arrangement.75 The Court also reasoned that the existence of substitutes does not destroy the presumption of market power.76
Statutory developments since the Loew’s decision make it unlikely that the Court’s presumptions regarding copyrights could operate under an antitrust-based misuse analysis today.77 For example, in the patent misuse context, where contemporaneous antitrust-based presumptions stood on firmer ground, the law now requires a showing of market power.78 Courts, enforcement officials, and commentators alike have generally leaned away from presuming that intellectual property confers market power.79 Nonetheless, this issue has not yet been visited in the copyright context, and Loew’s has yet to be overruled.
In Broadcast Music,80 the Court recognized the economics of copyrights, including the inherent transaction costs associated with monitoring and enforcement and the implications for efficient licensing. As a result, the court required a rule of reason approach to evaluating the lawfulness of the blanket licenses used by the American Society of Composers, Authors and Publishers (“ASCAP”) and Broadcast Music, Inc. (“BMI”). Although copyright misuse is not expressly considered in its opinion, the Court’s antitrust analysis illuminates a set of guiding principles to be applied when an asserted misuse defense is based upon anticompetitive behavior.81
Columbia Broadcasting System, Inc. (“CBS”) brought suit against ASCAP, BMI, and their members and affiliates seeking an injunction under the antitrust laws82 and a declaratory judgment of copyright misuse.83 After dismissal by the district court, the Second Circuit reversed and remanded.
Similar to the block-book licensing of Paramount and Loew’s, ASCAP and BMI each managed portfolios of copyrighted musical works and issued blanket licenses to perform each and every composition contained therein. The Supreme Court left intact the Second Circuit’s holding that the blanket licenses were distinguishable from block-booking due to the available opportunity to negotiate individual licenses and, therefore, were not unlawful tying arrangements. The Supreme Court focused instead on the blanket license as an unlawful means of price fixing.
The Supreme Court reversed both the Second Circuit’s holding that ASCAP’s and BMI’s blanket licensing policies amounted to per se violations of the Sherman Act and “the copyright misuse judgment dependent upon it.”84 The Second Circuit did not hold that blanket licenses were per se violations; it held “that the ASCAP blanket license in its present form is price-fixing and with respect to the television networks cannot be saved by a ‘market necessity’ defense. It therefore constitutes a violation of Section 1 of the Sherman Act.”85 Thus, because price fixing is per se unlawful, this particular form of blanket license violated antitrust law.86
The Supreme Court refused to “wholly equate[]” the collective arrangement among copyright holders to the “simple horizontal arrangement among competitors.”87 Commercial copyright transactions require negotiation, monitoring, and enforcement that can be prohibitively costly for individuals but minimized through collective arrangements. Although not quite a “market necessity,” blanket licenses may increase the private and social benefits of copyrighted works and improve market functionality by overcoming prohibitive transaction costs through economies of scale. Therefore, the Court concluded that blanket licenses do not “facially appear[] to . . . always or almost always tend to restrict competition and decrease output”88—analysis of actual competitive effects was needed.
Justice Stevens, in his dissent, applied a rule of reason analysis and found the particular blanket licenses to be unlawful.89 He argued that “[t]he rules which prohibit a patentee from enlarging his statutory monopoly by conditioning a license on the purchase of unpatented goods, or by refusing to grant a license under one patent unless the licensee also takes a license under another, are equally applicable to copyrights.”90 In doing so, he cited, inter alia, two leading patent misuse cases, Paramount and Loew’s.
Although the Supreme Court has not expressly applied copyright misuse in an infringement action, it has acknowledged the doctrine’s existence, suggested that it is derived from principles of equity applicable to intellectual property in general, and provided a rule of reason framework for evaluating conduct at the antitrust-intellectual property interface.91 There is little doubt that copyright misuse will eventually resurface in the Supreme Court and be reconciled. On this point, lower courts and commentators alike agree; however, divergent opinions abound as to what should be the proper scope and guiding principles for copyright misuse.
This section analyzes the controlling case law in three circuits that have recognized and applied the copyright misuse doctrine as a defense to bar recovery in a copyright infringement suit. Part III.C next discusses the status of the doctrine in other circuits that have yet to affirmatively apply it.
In 1990, the Fourth Circuit firmly established the copyright misuse doctrine as a viable defense to copyright infringement.92 Remarkably, the decision was the first affirmative application of the copyright misuse defense in a federal court of appeals.
Lasercomb licensed its copyrighted CAD/CAM software to the Holiday Steel Rule Die Corporation (Holiday Steel).93 The software enabled steel rule die manufacturers to utilize a computer to create a design and then direct manufacturing systems.94 Holiday Steel first copied the software for unauthorized private use, circumventing protective devices and avoiding additional fees, and then marketed its own software that was nearly identical to Lasercomb’s program. The district court found the defendants liable for copyright infringement and dismissed their affirmative copyright misuse defense. The Fourth Circuit reversed both the injunction and damages award granted by the district court based upon its conclusion “that Lasercomb’s anticompetitive clauses in its standard licensing agreement constitute misuse of copyright.”95 The defense was available “even though [Reynolds was] not part[y] to the standard licensing agreement.”96
The damning provisions of Lasercomb’s standard licensing agreement prohibited licensees from writing, developing, producing or selling “computer assisted die making software.”97 The district court found that even if copyright misuse were an available defense, the anticompetitive provisions were “reasonable in light of the delicate and sensitive area of computer software.”98 The Fourth Circuit rejected the lower court’s reliance on the “‘rule of reason’ concept” and stressed the independence of copyright misuse from antitrust law:99
[W]hile it is true that the attempted use of a copyright to violate the antitrust law probably would give rise to a misuse of copyright defense, the converse is not necessarily true—a misuse need not be a violation of antitrust law in order to comprise an equitable defense to an infringement action.100
The circuit court found that Lasercomb’s attempt “to use its copyright in a manner adverse to the public policy embodied in the copyright law” is sufficient to support a misuse defense.101 The license provisions were not only anticompetitive but they also “suppress[ed] any attempt by the licensee to independently implement the [unprotected] idea which [Lasercomb’s software] expresses,”102 thus reaching beyond the scope of the granted property right.
The Fourth Circuit dedicated a significant portion of its opinion to explaining the historical basis and public policy behind the intellectual property systems and the related development of the patent misuse defense as a limitation on the scope of the grant. Adhering to the reasoning of Morton Salt, the circuit court extended the scope limitation rationale from patents to copyrights.103 In doing so, the court also cited Paramount Pictures, quoting the Supreme Court’s discussion of block-booking as a means for enlarging the granted monopoly.104
In applying the copyright misuse doctrine to the facts, the court paid particular attention to the language of the licensing agreement and did not focus on the actual effects on competition or market power of the plaintiff, as it would in an antitrust analysis. The attempt to use the copyright “to control competition in an area outside the copyright, i.e., the idea of computer-assisted die manufacture [constitutes copyright misuse] regardless of whether such conduct amounts to an antitrust violation.”105
In Practice Management,106 the Ninth Circuit appeared to follow the Fourth Circuit’s analysis in Lasercomb to establish copyright misuse as a defense to infringement. Expressly agreeing with the Fourth Circuit, the Ninth Circuit reiterated that a finding of copyright misuse does not require that an antitrust violation be shown.107 In establishing the copyright misuse defense within its jurisdiction, the circuit court did not discuss the historical development or rationale behind the doctrine.108
As in Lasercomb, anticompetitive licensing provisions were the basis for the misuse holding. The American Medical Association (“AMA”) granted the Health Care Financing Administration (“HCFA”) a “nonexclusive, royalty free and irrevocable license to use, copy, publish and distribute” a coding system for medical procedures (compiled into the “Physician’s Current Procedural Terminology” or “CPT”) on the condition that the HCFA exclusively use the CPT and require its use in “programs administered by the HCFA, by its agents, and by other agencies whenever possible.”109
The circuit court dismissed various arguments made by the AMA that likely would be considered in a “rule of reason” antitrust analysis and focused exclusively on the plain text of the licensing agreement.110 Prohibiting the HCFA (and those within its regulatory reach) from using competing products gave the AMA “a substantial and unfair advantage” in the marketplace.111 The improper use of the CPT copyright as leverage to gain this advantage was sufficient for the court to hold that “Practice Management established its misuse defense as a matter of law . . . .”112
An important issue on which the Ninth Circuit may diverge from Lasercomb is whether a defendant with unclean hands may invoke the doctrine. In Atari Games Corp. v. Nintendo Co., Ltd.,113 the Federal Circuit applied Ninth Circuit law on copyright infringement claims and affirmed the district court’s holding that “Atari’s unclean hands prevent it from invoking equity” as a defense.114 The Ninth Circuit did not address this issue in Practice Management.
Alcatel115 was a significantly more complex case than either Lasercomb or Practice Management. In those cases, a rather simple reading of the licensing provisions provided a sufficient basis for finding copyright misuse. Although Alcatel ultimately depended similarly on the text of the license, the Fifth Circuit looked more carefully at the context and the effects of the plaintiff’s anticompetitive behavior.
Alcatel USA, Inc. (formerly known as DSC Communications Corporation, “DSC”) produced telephone switching equipment used in routing long distance telephone calls. Its copyrighted operating system software controlled the switching equipment, much like Lasercomb’s software was ultimately used to control manufacturing equipment when producing tool dies. DSC licensed its software through an agreement providing:
1) the operating system software remains the property of DSC;
2) the customer has the right to use the software only to operate its switch;
3) the customer is prohibited from copying the software or disclosing it to third parties;
4) the customers are authorized to use the software only in conjunction with DSC-manufactured equipment.116
In order to keep up with growing demand, DSC customers often expanded their capacity by adding “cards” to the switch.117 DGI Technologies, Inc. (“DGI”) was formed in 1989 precisely to produce expansion cards for use with DSC switches. Yet, because the cards had to be compatible with the controlling operating system, DGI was forced to download and copy DSC’s operating system software for testing and development.
DSC brought suit against DGI seeking relief for, inter alia, copyright infringement. The Fifth Circuit affirmed (1) the jury’s determination that damages for trade secret misappropriation were due, (2) those parts of the district court’s injunction against DGI based upon the trade secret claim, and (3) the district court’s dismissal of DGI’s antitrust counterclaim. However, the circuit court reversed those parts of the injunction that were based upon DGI’s copyright infringement because Alcatel had misused its copyright.118
As opposed to Lasercomb and Practice Management, in which Lasercomb and the AMA each attempted to suppress competition with its copyrighted product, copyright misuse in Alcatel stemmed from DSC’s attempt to use its “copyrights to indirectly gain commercial control over products DSC does not have copyrighted [or patented].”119 DSC’s licensing agreement prevented competing “card” manufacturers from developing products that were compatible with DSC’s copyrighted operating system. The Ninth Circuit found that “DGI was effectively prevented from developing its product, thereby securing for DSC a limited monopoly over its uncopyrighted [and unpatented] microprocessor cards.”120 In effect, this licensing arrangement gave DSC a de facto monopoly on the microprocessor cards.121
The Fifth Circuit acknowledged that copyright misuse is an equitable or “clean hands” defense, and concluded the licensing provisions alone dirtied DSC’s hands sufficiently for relief to be withheld on the infringement claim. As in Lasercomb and Practice Management, the court did not explicitly consider the extent of DSC’s market power in the “card” market or the actual effects on competition, as it would in an antitrust action. Instead, the text of the licensing agreement was controlling.
The three circuit court opinions discussed in this section illustrate the tip of the iceberg with regard to potential applications of the copyright misuse defense. In all three cases, the copyright licenses contained anticompetitive provisions that were relatively easy to discern as “overreaching” and, accordingly, were treated essentially as per se misuse.122
In Lasercomb, the Fourth Circuit established the defense as parallel to patent misuse, derived from the equitable doctrine of unclean hands, and distinguished from antitrust law.123 The Fifth and Ninth Circuits, in Alcatel and Practice Management respectively, followed suit without adding much substantive guidance. Although the circuits’ adoption of the doctrine was a necessary first step in jurisprudential development, it was primarily aimed at safeguarding the public interest in limiting the scope of the copyright and the corresponding statutory trade-off. It remains to be seen whether the defense will be limited to these seemingly per se misuse situations or extended beyond the text of licensing agreements.
Outside the Fourth, Fifth and Ninth Circuits, the viability of the copyright misuse defense remains uncertain. Although district courts have entertained the defense, the appellate courts have generally denied the existence of misuse based on the facts presented. Thus, their holdings may be limited to the facts, and their discussion and analysis of the doctrine most likely should be considered persuasive dicta, without binding precedential value. Accordingly, not all of the circuit court opinions mentioning copyright misuse are analyzed; instead, attention is briefly given to the more influential, substantive decisions in order of importance.
The Seventh Circuit has been adamant in its antitrust approach to misuse analysis. In Saturday Evening Post Co. v. Rumbleseat Press, Inc., a pre-Lasercomb decision, Judge Posner extended his view of patent misuse to copyright misuse, characterizing the doctrine as a defense that should be evaluated under traditional antitrust principles.124 The case involved a no-contest clause in a licensing agreement that prohibited Rumbleseat from challenging the validity of the Saturday Evening Post Company’s copyright. Judge Posner illuminated the potential procompetitive and anticompetitive effects of a no-contest clause and expressed doubts that copyrights generally confer “economically significant monopoly” power.125
Refusing to create “a federal common law rule that would jostle uncomfortably with the Sherman Act” and “[n]oting the convergence of patent-misuse principles with antitrust principles,” the Seventh Circuit held that “a no-contest clause in a copyright licensing agreement is valid unless shown to violate antitrust law.”126 While subsequent Seventh Circuit cases have not evaluated copyright misuse claims on antitrust grounds,127 Posner’s supposition that copyrights are less likely to confer market power than patents has been reiterated by the Seventh Circuit.128
In United Telephone Co. of Missouri v. Johnson Publishing Co., Inc.,129 the Eighth Circuit rejected a copyright misuse defense based on the facts presented. However, the court considered the relevant case law supporting the extension of the misuse doctrine from patent to copyright, including Morton Salt, Paramount Pictures, and Loew’s. Then, assuming “that judicial authority teaches that the patent misuse doctrine may be applied as a defense to copyright infringement,” the court evaluated whether United Telephone actually restrained competition by its pricing scheme or its “effort to require Johnson to purchase a license in its entire white pages listing” and concluded that it did not.130 Similar to the Seventh Circuit, the Eighth Circuit approach extends the misuse doctrine from patents to copyrights and applies a rule of reason analysis along the lines of traditional antitrust principles.131
In Data General Corp. v. Grumman Systems Support Corp.,132 the First Circuit entertained a complex copyright infringement action in which antitrust counterclaims paralleled a copyright misuse defense.133 Although the First Circuit acknowledged the copyright misuse doctrine’s development and discussed Lasercomb,134 it expressly chose not to “decide whether the federal copyright law permits a misuse defense.”135
Interestingly, the circuit court acknowledged that “the Lasercomb court [did] not require proof of an antitrust violation” for a successful misuse defense,136 but relied on the lack of an antitrust violation to reject Grumman Systems’ misuse defense. Grumman Systems based its misuse defense solely on the alleged anticompetitive behavior underlying the antitrust counterclaims, which the court held to be without merit, and the court found the misuse defense to be “equally devoid of merit.”137 In addition, the court noted that Grumman Systems might be precluded from raising the misuse defense because it had engaged in inequitable conduct itself.138
In 1996, the Eleventh Circuit seemed to endorse the copyright misuse doctrine where copying is required to achieve compatibility in the computer software context.139 It expressly joined “other circuits in finding that external considerations such as compatibility may negate a finding of infringement.”140 The court refused to enforce the contested copyright, and emphasized the fact-specific nature of its holding.141 Accordingly, the holding most likely is limited to the compatibility issue for computer software.142
Although many circuit courts have not affirmatively established the copyright misuse defense in their jurisdictions, some have considered the doctrine at length, notably the First, Seventh, and Eighth Circuits. The Seventh and Eighth Circuits have tended towards the antitrust-based approach to evaluating alleged misuse.143
This section extracts guiding principles from the case law in an attempt to clarify where the copyright misuse doctrine is today. Despite the existence of two seemingly divergent approaches to assessing misuse (antitrust-based copyright misuse and public policy-based copyright misuse), the final subsection presents a unified view.
Most of the opinions rejecting the copyright misuse defense on the facts have applied antitrust principles. The Seventh and Eighth Circuits have employed a rule of reason analysis to assess the anticompetitive effects of alleged misuse. Although criticized by some commentators for misunderstanding the equitable nature of the misuse doctrine,144 Judge Posner advocated a principled antitrust approach to evaluating misuse claims, citing a lack of equally tested or applicable alternatives. Antitrust-based copyright misuse resembles modern day patent misuse, which applies per se misuse to certain activities (which are established as such by the Supreme Court).145 The following two-part rule of reason test is applied to all other activities:
Is the restraint within the scope of the patent claims (and thus the patent grant)?
If the answer is yes, then the activity is per se legal. If the answer is no, does the activity on the whole promote or restrict competition?
The latter part of the test requires courts to balance the anti- and procompetitive effects of a particular activity, e.g., a licensing provision, and determine the net competitive effect.146
The Seventh Circuit in Saturday Evening Post and the Eighth Circuit in United Telephone engaged in a similar analysis. The presumptions upon which the rule of reason analysis of misuse is based are 1) that intellectual property generally is procompetitive and does not confer market power in the antitrust context; 2) that any market power given via an intellectual property right is valid; and 3) that intellectual property owners are entitled to exploit their intellectual property through the most efficient means available.147 These presumptions have emerged in the antitrust-intellectual property dialogue during the 1980s and 1990s and represent the current institutional opinion of the Department of Justice Antitrust Division.148 This view is supported by the Supreme Court holding in BMI that courts generally must conduct a more probing analysis into the market conditions and the procompetitive need for seemingly anticompetitive practices at the antitrust-intellectual property interface.149
Extending the rule of reason approach to copyright misuse requires courts first to evaluate the scope of the rights and the breadth of protection granted to copyright holders, and, second, to evaluate the net competitive effects of any questionable conduct. Both inquiries are commonly undertaken by courts and are readily within their adjudicative capacity. Removing the latter part of the test transforms the rule of reason approach into a per se rule.
Courts and commentators have attempted to distinguish the copyright misuse defense from antitrust law by focusing on the equitable nature of the doctrine as a clean hands defense and on the scope limitation function that it provides.150 Inequitable conduct on the part of the copyright holder need only offend the public policy behind the copyright system to trigger the defense.151 As the Fourth Circuit noted in Lasercomb:
[A] misuse need not be a violation of antitrust law in order to comprise an equitable defense to an infringement action. The question is not whether the copyright is being used in a manner violative of antitrust law (such as whether the licensing agreement is “reasonable”), but whether the copyright is being used in a manner violative of the public policy embodied in the grant of a copyright.152
Courts applying this rationale have looked specifically at copyright licensing provisions and decided whether the scope of the “limited monopoly” granted by the copyright is being expanded.
The three circuit court decisions establishing copyright misuse, discussed supra, are illustrative.153 In Lasercomb, the licensor restricted the licensee’s ability to develop similar software, effectively expanding the scope of the copyright beyond the protected expression to the unprotected idea.154 Similarly in Practice Management, license provisions restricting the licensee’s ability to use competing products were found to be expansive and thus misuse.155 Finally, in Alcatel, license provisions that restricted a licensee’s ability to develop products to compete with the licensor’s uncopyrighted products similarly were found to be misuse.156
In all three cases, the copyright was used as leverage to gain competitive advantage over licensees in areas beyond the scope of the limited privileges conferred by the copyright. While the scope limitation function of the misuse doctrine seems particularly fitting in these cases, the per se approach will likely be less straightforward in more complicated situations in the absence of blatantly objectionable license text. A more nuanced approach is inevitable, as seen in the patent misuse context,157 and will likely involve a set of per se rules applicable to easily recognizable factual settings and a rule of reason analysis in the remaining cases.158 It remains to be seen whether a rule of reason analysis not rooted in competitive effects but instead focused on the more amorphous concept of public policy will arise in the courts. As the next section indicates, we caution against such a development.
Most courts evaluating a copyright misuse defense acknowledge the equitable nature of the doctrine and the scope limitation function it plays, yet none have enunciated a form of public policy misuse that does not involve some degree of anticompetitive effects. This is not surprising because of the nature of intellectual property. Any use of a copyright has some anticompetitive effect. The public policy behind the copyright system is premised upon an exchange between short-term “monopoly” costs and long-term efficiency gains in investment, production, and dissemination of innovation.159 Thus, one might conclude that the proper inquiry is not what the behavior in question violates—public policy or antitrust law—but rather whether the social costs arising from copyright use exceed the expected short-term social costs inherent in the intellectual property grant.160 However, this formulation proves too much because assessing the social costs necessarily involves a careful look at the underlying public policies at stake. Yet the social cost inquiry returns us to a modified rule of reason analysis: Is the behavior in question a valid exercise of statutorily granted rights over the particular innovation? If not, what are the net effects from the behavior?
The public policy approach focuses on the first step of the rule of reason analysis and ignores the actual effects. In fact, this appears to be the approach taken in the Lasercomb and Practice Management decisions. Both the Fourth and Ninth Circuits stopped short of the second prong of the rule of reason test. The Fifth Circuit in Alcatel, on the other hand, emphasized the restrictive effect that DSC’s behavior had on DGI’s ability to compete. Still, none of the three circuit courts considered the potential positive effects of the licensing provisions in question.161 Thus, it seems that the public policy approach may reduce the rule of reason approach to a single step, which is akin to a per se rule. Although no court explicitly declared the particular licensing practice to be per se misuse, the lack of additional guidance leaves little room for distinction. Therefore, restricting a licensee’s ability to innovate or tying an uncopyrighted or unpatented good to a copyrighted good is likely to be per se misuse.
In establishing these per se rules, the Fourth, Fifth, and Ninth Circuits were not creating per se antitrust rules based on the conclusion that in the general run of cases the questioned conduct is presumably anticompetitive on the whole. Instead, the courts were creating per se rules that keep intact the socially acceptable trade-off manifest in the copyright law. In a sense, the circuit courts were safeguarding the public interest in maintaining a copyright of limited scope while at the same time coordinating the various intellectual property systems.162 Choosing to protect one’s software through copyright should not and, at least in these jurisdictions, does not facilitate patent or trade secret protection on related functional innovations.
Limiting public policy-based misuse to per se misuse under a small set of factual settings readily identifiable both by courts and private actors allows a principled or unified theory of copyright misuse to emerge. Generally, copyright misuse should closely resemble patent misuse, having a small set of “known” per se rules and otherwise being determined under a rule of reason analysis. Copyright misuse may diverge, in practice, from patent misuse by its incorporation of per se rules based on the public policy behind the copyright system that complement per se rules based on antitrust principles.
This Part considers the copyright misuse defense as it relates to computer software, emphasizing the jurisprudential and procedural principles developed in Parts III.A and III.B. In building our approach to software copyright misuse, Part IV.A discusses the nature of software as an innovation and next turns to the landscape of software. This sets the stage for Part IV.B, which applies the jurisprudential model to identify the substantive need for common law. Part IV.B also considers the limits of the fair use doctrine and antitrust-based copyright misuse that make them incomplete protections for those who would reverse-engineer software. Finally, Part IV.C proposes a copyright misuse defense with an antitrust core, yet one complemented by a per se misuse rule against reverse engineering licensing restrictions.163
Importantly, the proposed refinement of existing doctrine does not endeavor to be a complete solution; instead, this approach stresses the need for precise and clear development of the law. Thus, we imagine that courts may craft additional public policy-based per se rules along the guidelines we provide,164 with antitrust doctrine applying in remaining cases.
In early cases involving software copyrights, a handful of district courts expressed doubt that unreadable programs could be protected on the grounds that they are not expressive.165 Congress, however, answered this question by carving out a limitation in copyright protection of software in its 1980 amendment to § 117 of the Copyright Act,166 thereby implying the existence of a general protection of software under § 102(a).167 Moreover, the Commission on New Technological Uses (“CONTU”), appointed by Congress to study the copyrightability of software, concluded in its final report that the Copyright Act should “make it explicit that computer programs, to the extent that they embody an author’s original creation, are proper subject matter of copyright.”168 In light of these developments, a consensus quickly emerged among federal courts that copyright does indeed protect software as a literary work,169 a view unlikely to change, particularly in light of international consensus on the issue.170
Despite software’s inclusion in the ambit of copyright law, software’s unique attributes continue to present courts with special pedagogical challenges. An understanding of these challenges helps to define the need for doctrinal development. We concentrate on particular attributes of software as an innovation before turning to the landscape of software.
Software challenges a basic assumption that copyrights should afford the public greater access to expressive works. This section first explains why software is hidden expression as a matter of technological coincidence. Further, it argues that copyright law is designed to promote creative expression not only by stimulating its production, but also by envisioning wider public access to such expression.171 Finally, this section briefly explores the tension created by software where the public acquires no expression.
Computer programs are instructions that manipulate computer hardware to useful ends. These instructions are typically written by humans in languages that, while resembling natural languages, are lexically scaled down in order to afford greater precision.172 These human-readable instructions are referred to as source code, and this code is written in one or more files called source files.
Programs usually must undergo additional processing in order to be executed by a computer. Sometimes this intermediate processing is done after the program has already been initiated: a process known as interpretation, where each line of a source file is translated into a format that can be executed by the processor. However, running such programs—commonly referred to as scripts—through an interpreter is often too slow to be practical for large applications. For this reason, most programs are prepared for execution in two stages: first, source code is “compiled” into object code, which is a simplified format that can be understood only by a digital processor; and, second, object files are linked together to resolve external references. This processing creates a single executable program.173 The executable program is not dependent on any of the files that went into creating it—indeed, the purpose of compilation and linking is to resolve dependencies to afford faster processing. As a consequence, a compiled and linked program can function entirely without the presence of its underlying source code.
If source code, apart from its executable program, is published in guides or on the Internet for instructive or other purposes, one may readily conclude that it is a literary work within the meaning of the Copyright Act.174 Source code can be an extraordinarily complex, structured expression of ideas. Yet source code as a literary work does not present courts with a novel problem, so existing principles of law can be adapted to define the scope of copyrights on source code.175 Courts have generally approved some form of dissection analysis, filtering ideas and functional elements from expression.176
Yet, protection of the object code embedded in executable files presents special challenges. The functional independence of executable programs effectively hides the program’s underlying expression, which allows software manufacturers to maintain greater secrecy and control over their designs. Because executable files do not depend on their source files in order to operate, source files are typically not made available to those who are given access to the executable files.177 In this way, users of an executable program cannot evaluate the program’s source files and thereby understand its algorithm. Although this fact does not afford complete control or secrecy, someone with only executable files must undertake relatively costly efforts in order to translate executable files back into source code—a process known commonly as decompilation, disassembly, or, more generically, reverse engineering.178 Increasingly, producers supplement their copyright protection of hidden expression through alternate frameworks like licensing restrictions or technology in order to raise the costs of reverse engineering.179
While the hidden nature of object code does not preclude copyright protection,180 it creates friction with a core premise of copyright law, namely that producers will distribute expressive works to the public.181 Congress chose to confer property rights to authors,182 rather than simply passing a criminal law against misappropriation,183 or relying on other mechanisms.184 Thus, the law lures authors to produce expressive works with exclusive, alienable rights,185 increasing the prospects that authors will profit from their labors by transacting in the market.186 The transactional framework allows for the efficient distribution of expressive works, thereby maximizing the public’s welfare in accessing expression.187 Therefore, Congress’s central reliance on property rights in copyright law evinces an intent to distribute ideas and expression to the public, as well as to stimulate their production.188
In addition to the incentive structure, copyrights are limited in scope and duration so that expression and ideas eventually pass into the public domain. This allows consumers priced out of the market for an expressive work during the term of the copyright to gain access when the copyright expires. Likewise, the consuming public may reuse unprotected ideas and expression even during the term of copyright. As with the incentive structure, Congress struck a balance between public and private interests, giving the public greater access to ideas and expression than it might otherwise have without copyright law.189
Although Congress designed the Copyright Act to promote creative expression, object code’s principal value lies not in its “expression,” but instead in its ability to perform tasks effectively.190 Indeed, object code is unique in copyright law in that it has no perceptible expression. By contrast, traditional subject matter, such as paintings or novels, bears ideas and protected expression on its face.191 By transacting, buyers of traditional works can perceive the expression as a matter of course. However, transactions involving software have little if anything to do with expression.
The hidden nature of software challenges the assumption that copyright naturally entails broader distributions of expression. As Part IV.B explores, the resulting imbalance helps to explain the emergence of copyright misuse and particularly the extension of the fair use defense to intermediate copying,192 both of which apply predominately in the software context.
In addition to being hidden expression, executable programs are stand-alone units that comprise the protected and unprotected elements of the underlying source code.193 This, too, challenges copyright doctrine because a producer can enforce its copyright over programs as a single black box, thereby acquiring a de facto monopoly over the unprotected material within.
A fundamental axiom of copyright law is that copyrights may not protect ideas, unoriginal expressions, or unfixed expressions.194 Ideas and such expressions reside in the public domain, and an author may not appropriate them through their inclusion in a work that qualifies for copyright protection.195
Yet protecting object code as a black box effectively extends the copyright to unprotected, internal elements.196 Reverse engineering entails copying object code into a computer’s memory, which violates the copyright unless such copying is authorized.197 If the defendant’s conduct does not qualify as a fair use,198 the law effectively protects all the elements in the black box irrespective of whether the elements could otherwise be protected. Thus, a mere act of discovery, without some form of common law involvement, could subject a user of software to liability.199
By enforcing black-box software copyrights, courts effectively seal the box from public viewing. Thus, protection of object code carries the potential to disrupt the public-private balance200 by limiting downstream innovation rather than promoting it.201
The third attribute setting software apart from other copyrightable media is its functional nature.202 Software creates tension at the patent-copyright interface because patent law presumptively covers functional innovations whereas copyright law covers expressive innovations. Moreover, fitting software into copyright law is inconsistent with established copyright doctrine that generally rejects protection for functional works.
In general terms, software has blurred the functional-nonfunctional dichotomy that traditionally delineated classes of innovation. Although functional innovations are the “proper subject matter” of the patent system, over the last few decades, innovators have increasingly looked for intellectual property protection through other institutional mechanisms. In some cases, innovations with functional attributes fall short of the patentability criteria, forcing innovators to look to other forms of protection. In other cases, innovators look to other forms of protection with the hope to gain exclusive rights for a longer duration than available under the patent system; some innovators simply prefer trade secrecy to fully disclosing their innovation.203
The emergence of the computer software industry provides the archetypal example. Computer software is currently protected through copyright because both source code and object code are considered “written expressions.” Yet the economic value of copyrighted object code is completely derived from the functional ends facilitated by the software.204 Courts and commentators alike have struggled to determine what form of intellectual property—patent, copyright, or some sui generis creation—is appropriate for computer software.205
Correspondingly, copyright misuse is an appropriate judicial mechanism for restricting the social costs of granting copyrights on functional innovations. In fact, as discussed in Part II, both the Fourth and Fifth Circuit decisions establishing copyright misuse arose in the context of copyrighted software with functional applications. Although some courts have intimated that patents are more prone to convey market power than copyrights,206 others have acknowledged otherwise in the software context:
Certainly, a monopolist’s refusal to license others to a commercially successful patented idea is likely to have more profound anti-competitive consequences than a refusal to allow others to duplicate the copyrighted expression of an unpatented idea (although such differences may become less pronounced if copyright law becomes increasingly protective of intellectual property such as computer software).207
In addition to the tensions between patent and copyright, software also creates internal tension within copyright law. Courts generally deny functional innovations protection if they fall outside the patent system. For example, in the context of pictorial, graphical, and sculptural (“PGS”) works, courts have developed the useful article doctrine to deal with these works’ functional nature.208 Under this doctrine, copyright protection is denied “if design elements reflect a merger of aesthetic and functional considerations . . . .”209 The reason for this strict refusal is because “the artistic aspects of [the] work cannot be said to be conceptually separable from the utilitarian elements.”210 The case of computer programs presents an even starker picture because there is no apparent expression: only the functional effects caused by the program’s operation. In trademark law, protection is denied for trade dress that is functional, i.e., “if it is essential to the use or purpose of the article or if it affects the cost or quality of the article.”211 Courts recognize that, as in the PGS context, trademark holders could otherwise acquire patent-like protection (in this case by exhausting competitors’ potential options for a comparable trade dress).212 Accordingly, courts outside of patent law and software have consistently denied protection for functional elements.
The functional value of software challenges courts to coordinate patent and copyright law. Patent law imposes more rigorous requirements than copyright, such as a disclosure requirement, a shorter duration, and more robust protection,213 arguably because functional innovations are more prone to confer market power.214 Allowing a functional innovation like software into copyright carries the potential to disrupt the public-private balance that each system has struck.215
The previous discussion considered attributes of software as a class of innovation, whereas this section looks at software in a broader context in order to evaluate its impact on innovation. While a careful analysis of the landscape demands far more attention than we can provide here, this subsection considers a sample of features that affect innovation.
As information-related technologies have pervaded the international economy, software producers have sought a fuller range of protection for their products. What has emerged is a rich tapestry of legal safeguards for innovation. This Article deals mainly with proprietary sources of protection, i.e., patent and copyright law.216 Trade secrecy, the traditional fallback for producers without intellectual property rights, confers weak legal protection to producers, exposing them to risks of lawful appropriation.217 Licenses remain a central form of legal protection, complementing or possibly even substituting for intellectual property rights.218 In recent years, Congress has increasingly looked to criminal law as a means of regulating access to information technologies including software.219
Legal protections are not, however, the complete picture. The wildcard in protecting software is technology itself, such as lock-out programs and encryption methods.220 Technology may allow clever producers to craft their own extra-legal protection, impacting existing legal doctrine in undetermined ways. Three possible scenarios emerge. First, producers may supplement their legal protection with technological protections. For example, a video game manufacturer might rely on lock-out technology and/or copyright law to protect its object code.221 Second, producers may circumvent legal rules through technology. For example, in the Microsoft case, the district court found that Microsoft reengineered its operating system, bypassing an earlier consent decree.222 Third, producers might rely exclusively on technology as a form of self-help. For example, AOL configured its servers to return large volumes of unsolicited e-mail to their senders in an effort to discourage “spamming.”223
Whether one considers legal or technological measures, today’s landscape affords software producers unprecedented protection over their products.224 While reduced risk may attract entrants to the market, excessive protection can create higher barriers to downstream innovation,225 limiting the synergistic benefits of innovation.226
Aside from the breadth of protections, the software landscape also impacts innovation through a set of structural factors. The first of these factors is the extraordinary pace of technology: leaps in technology may quickly allow actors to achieve market dominance. Technology can thereby promote competition in cases where it breaks down existing monopolies,227 yet it also may simply swap one monopolist with another.228 Software copyright protection favoring producers magnifies these sudden shifts in market power,229 particularly in networking markets like those relying on the Internet.230
Furthermore, the software market exhibits lead-time effects, meaning that producers have a window during which they can gain an advantage on competitors.231 The advantage acts as a barrier to entry, and it is particularly significant because innovation often acts as an input to further innovation.232 Lead-time effects may arise naturally or legally. In the first case, a producer has an advantage in developing derivative software to the extent that it understands its own technology and has processes and facilities in place to allow more rapid development. A competitor, on the other hand, would need to reverse-engineer and spend time learning the technology before developing it. Copyright law extends natural lead-time effects during the statutory term of protection by giving authors exclusive rights to produce derivative works.233 The consequence of lead-time effects is to allow dominant actors greater breathing room to maintain or extend their dominant status.
Software further exhibits application effects, or the increased value attributable exclusively to users’ familiarity with a product.234 Application effects could roughly be thought of as brand loyalty, though they are particularly significant in the software context because of its relative complexity compared to other products. Producers benefit because users are more likely to buy upgrades of familiar products than they are to buy and learn a substitute.
Closely related to application effects are network effects, or the increased value attributable to more users of a product.235 The paradigmatic example of network effects is the telephone: owning the only phone in town is less valuable than owning one if everyone else has one.236 Network effects in software can lead to standardization of technologies, which can confer considerable benefits to a producer whose product becomes the standard.237 In software, an example of network effects is Microsoft Windows, which is more valuable because so much of the market uses it, thereby reinforcing Microsoft’s dominant share of the PC operating system market.238
Application effects and network effects are similar in that they both create barriers to entry for competitors and pressure for standardization. They differ, however, in that application effects relate exclusively to a user’s desire for consistency, whereas network effects follow from others’ use of a technology. To illustrate the difference, traditionalists might retain familiar technologies (e.g., Atari consoles c. 1980) despite the network effects of emergent technologies (e.g., Sony Playstation II).
This discussion reveals the software landscape as one tending towards monopoly in important respects, and one affording producers an extensive range of protections. These both have important implications for copyright law, making excesses in the level of legal protection particularly severe.239
Software’s unique nature effectively magnifies the value of copyright protection by allowing producers to fence in ideas and expression.240 Producers can build fences through a variety of means: acquiring copyrights on hidden expression, exercising black-box copyrights, seeking protection through alternate legal frameworks, seeking protection through extra-legal technologies, and (if the producer has market power) withholding production thereby raising price.241
Information fencing benefits society by encouraging production of software, though it imposes an important social cost, namely the reduced public access to an author’s ideas and expression. Access is an integral part of “Progress,” in that it facilitates review, debate, and refinement of existing innovations.242 It also raises the prospects for derivative innovations that one or few producers might not have conceived of or been able to afford. By restricting access to innovation, producers fundamentally alter the nature of copyright protection for computer programs.243
Part IV.B expands on this analysis in two ways, first articulating a demand for substantive doctrinal development, and next explaining why fair use and antitrust-based copyright misuse by themselves do not fully meet the demand.
The substantive basis for common law derives from the jurisprudential model developed in Part II. The model supports common law to fulfill each function of common law rule-making, namely to correct an internal deficiency in the Copyright Act, to coordinate copyright law with patent and antitrust law, and to safeguard policies of copyright law.
Our model sets forth the three jurisprudential functions of correcting, coordinating, and safeguarding.244 Where a demand arises for courts to fulfill one or more of the functions, a case can be made that courts should intervene. The challenges posed by protecting software in copyright implicate all three jurisprudential functions, establishing a strong substantive basis for doctrinal development.
The hidden nature of software exploits a gap in the Copyright Act, namely the absence of an explicit requirement that expressive works be perceptible. Any such gap, of course, implicates the corrective function as described above. As discussed above, a core yet unarticulated premise of copyright law is that authors will distribute expression and ideas to the public, primarily through the market.245 Yet software copyrights challenge this premise by rendering expression imperceptible. The public loses an essential benefit of copyright law—greater knowledge—and instead gains a functional innovation.246 Correspondingly, producers gain the opportunity to fence information. Practically, courts cannot be expected to read a perceptibility requirement into the Copyright Act where Congress did not express one,247 yet they may scrutinize conduct by a plaintiff that frustrates a defendant’s access to a program’s underlying expression. By forbidding overreaching restrictions, the courts can thereby help to fill the gap.
The Copyright Act contains another gap reflected in its protection of black boxes, or works affording de facto protection of ideas and expression in the public domain.248 As discussed, software works like a black box, because standalone executable files cannot be decompiled without a predicate act of infringement. Enforcement of software copyrights would thus protect everything in the box whether or not certain elements belong to the producer. Common law could allow the box to be opened, so the public could reuse unprotected elements, thereby stimulating downstream innovation. If a producer maintained restrictions beyond the term of copyright, courts could likewise permit the opening of the box on the grounds that the full contents then belong to the public. As to the form of common law, a court could not plausibly read a decompilation rule into the statute, but it could scrutinize efforts to keep the black box sealed.
The functional nature of software presents the need for courts to coordinate copyright and patent law, particularly given differences in scope, duration, and criteria of protection.249 The patent-copyright interface is based on a distinction between functional and expressive innovations. Even though software is protectable under copyright law, its value lies in its functionality, which muddies the distinction between patents and copyrights. Allowing software to reside in copyright alters the public-private balance that each regime has struck.
While there is no easy answer to the tensions that software copyrights create with patent law, courts may scrutinize efforts that effectively expand rights as necessary intrusions into patent law. Producers expand their rights by frustrating decompilation of their source code. If producers can maintain a secret base of functional knowledge that is further grounded in a legal monopoly, copyright law can undercut the integrity of patents. Software patents appear less valuable by contrast (e.g., due to the longer duration of copyrights).
Software copyrights further present a coordination issue with antitrust law, because the magnified protection carries a greater potential for competitive harm than with traditional subject matter. As with many industries, software can be fiercely competitive, yet the scope and nature of protection can and does create monopolies—Microsoft being the most obvious example.250 Antitrust law serves as the obvious limit on overreaching just as it does in the patent regime. Thus, courts already recognize a substantive need for common law, and need only look to familiar law to ensure that producers have not unreasonably burdened competition.
Thus far, we have focused on particular attributes of software as an innovation; yet looking at software in its broader context, a striking imbalance in public and private interests emerges. In the absence of common law, software producers can effectively acquire a monopoly of underlying ideas and expression, all the while maintaining secrecy in their designs, for a term of life plus seventy years. The producers can potentially supplement their copyrights with patent protection,251 licensing restrictions,252 federal criminal law,253 and technological restrictions.254 And the industry itself may magnify emergent inefficiencies because of the high pace of change, thereby compounding lead-time effects and application effects.
The nature of software copyright protection marks the emergence of information fencing in copyright law; this, in turn, justifies common law to safeguard a broader public policy favoring openness. Doctrinal development might involve new rules that seek to balance public and private interests, or it might turn to familiar doctrine. We advocate a mix of both, namely narrowly-crafted public policy-based rules complementing antitrust law and the fair use doctrine.
This subsection has developed a basis for common law without rigorously defining a role for copyright misuse. We undertake this process in the following subsections by evaluating the limits of the fair use doctrine and antitrust-based misuse. What emerges is a remaining need for court action, calling for the extension of existing doctrines.
Fair use is a complete, statutory defense to copyright infringement that protects conduct such as reporting, scholarship, and parody.255 Several courts of appeals have held that fair use also protects certain forms of reverse engineering.256 Yet, the defense does not fully meet the substantive demand for common law rule-making because of limitations on its scope. Moreover, applying the defense involves an ad hoc balancing of statutory factors,257 leaving actors with lingering uncertainty.258
While the Supreme Court has not taken a reverse engineering case, the Court has stressed that fair use is generally reserved for uses that do not adversely affect the plaintiff’s market.259 For example, the Court rejected the defense where a defendant had copied factual news accounts in writing an article, based partly on the fact that the defendant’s article “adversely affect[ed] the potential market for the copyrighted work.”260 The Court has further recognized that the market effects analysis extends to uses that preempt a copyright holder’s entry into a foreseeable market.
The Ninth Circuit made the definitive statement on reverse engineering as a fair use,261 holding that “where disassembly is the only way to gain access to the ideas and functional elements embodied in a copyrighted computer program and where there is a legitimate reason for seeking such access, disassembly is a fair use of the copyrighted work.”262
Sega established the fair use defense for intermediate copying, yet the court established limits on its scope that made it an incomplete answer to the substantive need for further common law development. The Sega court suggested that where a defendant’s purpose in disassembling object code is not “legitimate,” the fair use defense would be unavailable.263 In qualifying legitimate conduct, the court emphasized that the defendant, Accolade, had not avoided an independent creative process: indeed, most of Accolade’s software had already been developed for other hardware systems.264 This reasoning implies that a defendant who decompiles simply to avoid “reinventing the wheel” might fare differently than Accolade.265 The court’s reasoning also suggests that the availability of noninfringing uses (such as licensing) would alter the analysis.266
Further, the Sega court’s analysis of market effects leaves uncertainty, and appears to be partially at odds with the Supreme Court’s guidance. The Sega court stressed that disassembly, in order to create competitive software, “affected the market for Genesis-compatible games in an indirect fashion.”267 Its reasoning is illuminating: “no consumer except the most avid devotee of President Ford’s regime might be expected to buy more than one version of the President’s memoirs, [yet] video game users typically purchase more than one game.”268 The court suggested that because one more competitor in an active market would not seriously impact a plaintiff, its economic loss was de minimus.269
Yet, if one considers a different market, it is unclear that Sega’s reasoning still applies. A competitor that decompiles Netscape or Internet Explorer in order to enter the browser market might significantly alter the existing balance in market share.270 Factors that may support a distinction between video games and browsers include the availability of substitutes, the barriers to market entry, and (potentially) the demand for the products in general. As consumers have demanded more sophisticated features in the browser market, the barriers to entry have steadily increased. As a result, the browser market, like President Ford’s memoirs, presents consumers with relatively few choices. Therefore, a software manufacturer that reverse-engineers and substantially borrows Netscape’s body of functionality could plausibly supplant Netscape in the browser market.
The Sega court’s market effects analysis also exhibits some internal tension. The court apparently embraced competition as a justification for intermediate copying by emphasizing the public benefit of market entry and competition. However, the court sought to distance the challenged conduct from the foreseeable competitive harm by casting the downstream market effects of intermediate copying as “indirect.” Particularly given the Supreme Court’s guidance, it remains unclear that the Court would recognize increased competition as consistent with Congress’ intent in § 107(4).
Further, as commentators have observed, a fair use defense does not preclude software developers from imposing higher technical and legal barriers against reverse engineering.271 Rather, it protects only prospective competitors who are clever or persistent enough to surmount these ever-increasing hurdles. Therefore, even a broader fair use defense, which could completely and unambiguously protect intermediate copying, would not suffice in promoting creative transfers if barriers to decompilation become unreasonably high.
Finally, even if fair use were sufficient to meet the substantive demand for furth