Vol. 97, No. 4

Infringement Episodes

Shani Shisha*
Copyright Law, Damages
Articles

For decades, copyright scholars have waged a spirited campaign against statutory damages. Our remedial system, critics say, is an incoherent mess. The core problem is that copyright holders can recover a separate award of statutory damages for every infringed work. As a result, damages can rapidly add up in any case involving multiple works. Because the number of statutory awards is tethered to the number of works, even trivial claims can lead to crippling damages. Commentators, policymakers, and judges have criticized this system as arbitrary and overbroad. And yet it endures.

This Article argues that copyright’s per-work scheme has obscured, and at times eclipsed, a more compelling paradigm of copyright damages—one that attends more closely to the defendant’s course of conduct. This new approach would allow courts to examine whether the defendant’s actions arose out of, and were rooted in, a single infringement episode. By infringement episode, I mean a chain of infringing acts that together constitute a larger factual event. When the defendant’s conduct is traceable to a single episode, courts should issue only a single statutory award—no matter how many works are at stake. This framework, in short, would substitute rigidity for flexibility. It would displace copyright’s one-award-per-work scheme and instead introduce a contextual inquiry into the defendant’s course of conduct. Doing so can mitigate the risk of outlandish awards, encourage courts to properly calibrate damages, and infuse a degree of much-needed pragmatism into our system.

INTRODUCTION

Copyright scholars have long taken a decidedly dim view of statutory damages. The federal copyright statute entitles plaintiffs to recover statutory damages without proof of harm.¹ But our remedial system, critics say, is an incoherent jumble. Courts wield “virtually unfettered discretion” in assessing damages.² Statutory awards often seem inexplicable or arbitrary.³ Courts are also deeply divided over the justificatory basis for statutory damages. Some believe that statutory damages derive from a compensatory rationale, while others lean more readily on punitive instincts.⁴ And statutory damages sometimes lead to inflated awards that far outstrip any reasonable assessment of harm.⁵ The law of statutory damages is a disaster.

Underlying these pathologies is a peculiar feature of our copyright system: its per-work structure. Copyright holders can recover a separate statutory award for every infringed work.⁶ As a result, statutory damages can add up rapidly in any case involving more than a single work.⁷ In one recent case, copyright holders leveraged this per-work scheme to sue for a breathtaking award of $75 trillion in statutory damages—more than “the combined gross domestic product and national debt of the United States.”⁸ In another notable case, a record label brought action against Jammie Thomas-Rasset, a single mother from Minnesota, for illegally downloading and distributing twenty-four songs.⁹ A jury found the defendant liable and returned an award of $80,000 per infringed song, resulting in a total award of $1.92 million.¹⁰ Stunningly, by some measures, this award was 35,000 times larger than the plaintiff’s actual loss.¹¹

As these examples illustrate, the prospect of outlandish damages is rather startling. Because the number of statutory awards is tethered directly to the number of infringed works, statutory damages can quickly balloon in relatively minor cases. The risk of financial ruin looms large.

Copyright scholars, in turn, have proposed a number of potential reforms.¹² Matthew Sag, for instance, suggests that we reduce or cap the range of available damages in certain file-sharing cases.¹³ Oren Bracha and Talha Syed contend that, at least under a compensatory framework, statutory damages should approximate actual harm.¹⁴ Ben Depoorter contemplates a variety of reforms that would make excessive damages both less likely and less costly.¹⁵ James DeBriyn calls for eliminating statutory damages.¹⁶ Michael Carrier would proscribe recovery of statutory damages in cases involving secondary liability.¹⁷ And Alan Garfield endorses a host of different reforms—interpretive, legislative, and constitutional—to restrict judicial discretion in assessing statutory damages.¹⁸

None of these proposals, however, confront the core issue: the per-work structure of copyright remedies. To meet the urgency of the moment, this Article suggests a legislative reform that would do away with per-work remedies altogether. Instead, I propose that courts police the defendant’s conduct based on a more holistic approach.¹⁹ Our current system encourages courts and juries to engage in a rote exercise of counting infringed works.²⁰ A better framework, I argue, would afford courts greater discretion to assess whether the defendant’s conduct gave rise to, and was squarely rooted in, a single infringement episode. By infringement episode, I mean a chain of related infringing acts that together constitute a larger factual event. When the defendant’s conduct is traceable to a single larger episode, courts should be able to issue only a single statutory award—no matter how many works are at stake. By focusing on infringement episodes rather than the number of infringed works, this approach breaks with copyright’s one-award-per-work system. It substitutes rigidity for flexibility, relying instead on a context-sensitive inquiry into the defendant’s course of conduct. And it offers an alternative analytical paradigm for thinking about copyright remedies.

To operationalize this approach, the Article sketches a richly nuanced account of when and why courts might treat a series of infringing acts as sufficiently intertwined to constitute a single episode.²¹ In particular, I argue that courts should attend to an array of interrelated factors, including the nature of the infringing acts; the time and place of each act; whether the evidence supporting one infringement is necessary or sufficient to sustain liability for another; whether the defendant’s actions were executed in pursuit of a common plan or a larger creative enterprise; and the potential consequences of a statutory award. If adopted, this multifactor framework would allow courts to avoid per-work damages when the defendant’s actions derive from a single infringement episode.

This approach may seem radical. Immersed as we are in a legal culture that prizes per-work awards, it may be difficult to conceive of a better system for assessing damages—one that is both administrable and normatively attractive. But if we want to put an end to excessive awards, we need to get judges and juries out of the business of counting infringed works. The way to do that is through structural reform that would enable courts to scrutinize the defendant’s overarching course of conduct.

The Article also brings this approach into conversation with other legal disciplines.²² Across a variety of seemingly siloed areas of law—criminal law, civil procedure, and immigration law—courts have already fashioned doctrinal tools to evaluate whether a cluster of wrongful acts might be collectively reducible to a single transaction, episode, scheme, or series of occurrences. Typically, courts do so by undertaking a multifactor inquiry into the defendant’s entire course of conduct. Rather than examining every single action in isolation, courts approach the defendant’s conduct from a more holistic perspective. And this framework, subject to a few notable modifications, could be brought to bear on modern copyright law.

Ultimately, this Article seeks to chart a new horizon for copyright remedies. What distinguishes the proposed approach from previous proposals is that it takes seriously the core issue: per-work remedies that allow courts and juries to aggregate damages. After decades of faint-hearted debates, policymakers ought to try something different. If left unabated, copyright’s remedial system will continue to provoke all manner of mischief. It is time to rethink per-work damages in toto.

The argument proceeds in four parts. Part I discusses two of the most contestable aspects of our current system. The first problem is that infringement is principally understood to be a strict liability tort.²³ In other words, liability for copyright infringement does not depend on the infringer’s intent, knowledge, or negligence. Accidental infringement, then, is a very real possibility. The second problem is that copyright owners can elect to recover a separate award of statutory damages for every infringed work.²⁴ Because these awards can range anywhere from $200 to $150,000 per work, there is much at stake.

And when one considers the caselaw, the picture grows darker still. Part II delivers a diagnosis of the central issues surrounding the law of statutory damages. First, courts often disagree over the appropriate justification for statutory damages.²⁵ They vacillate between compensatory, punitive, and deterrence-based rationales, and rarely offer anything more than a threadbare explanation for why one justification is preferable to another. Second, copyright’s remedial scheme tends to produce unpredictable or otherwise arbitrary outcomes.²⁶ Third, statutory damages can lead to grossly excessive awards.²⁷ Due to per-work aggregation, statutory awards can swell up in any case involving more than a single work. Fourth, this per-work structure can discourage courts from properly developing copyright law.²⁸ Because statutory damages attach to infringed works rather than infringed rights, courts have little incentive to systematically address or unpack certain rights. The result is an impoverished body of law.

Part III outlines an alternative framework for assessing damages. It begins by surveying previous reform proposals.²⁹ It next catalogs a range of doctrinal analogues borrowed from other legal disciplines, including criminal law, civil procedure, and immigration law.³⁰ Based on this cross-disciplinary analysis, I then chalk out a more granular account of how courts might go about identifying infringement episodes.³¹ In so doing, I show that the proposed approach can offer a practical blueprint for evaluating the defendant’s conduct in different kinds of cases: file-sharing cases, artistic appropriation cases, and commercial infringement cases.

Part IV considers and rejects a number of potential objections. I defend my use of doctrinal analogues; clarify the scope of this framework by explaining what it can—and cannot—do; and examine how my proposal might ensure that plaintiffs are compensated for the harms they suffered.³² A brief conclusion follows.

I. A TALE OF TWO WOES

Copyright law suffuses our world. An incredible constellation of seemingly mundane activities—emailing, posting photos and videos on social media, texting, and even sharing memes—can give rise to claims of copyright infringement.³³ The result is that copyright “pervades our cultural universe.”³⁴ It has come to “dominate vast swaths of everyday life.”³⁵ As one commentator put it, copyright law “touches everyone and everything.”³⁶

This Section discusses two of the main features that render modern copyright law so permissive. The first is the fundamental nature of our system as a strict liability regime. The second is a unique remedial framework that entitles copyright owners to recover statutory damages without proof of actual harm. These two features strike at the heart of a formidable regime that has grown precariously overbroad. The risk of accidental infringement, coupled with the attendant threat of exorbitant damages, is all but inescapable. Copyright law is a slumbering giant.

A. Strict Liability

In the traditional telling, ours is a strict liability regime. Courts routinely treat copyright infringement as a strict liability tort.³⁷ Scholars take a similar view.³⁸ It is easy enough to understand why. Copyright liability, after all, does not depend on the infringer’s knowledge, negligence, or intent.³⁹ It is simply irrelevant whether the infringer knew that their actions constituted copyright infringement. Liability is not conditioned upon any kind of fault. To establish a prima facie case, the copyright owner need not make a showing of knowledge or negligence on the part of the defendant. Copyright infringement, in short, is a strict liability tort.

But why is that so? One standard answer is that copyright is a form of property entitlement, and violations of property rights ought to be punishable regardless of the transgressor’s fault.⁴⁰ Another explanation is that the defendant should face absolute liability because they are better positioned to avoid loss, at least as compared to the copyright owner.⁴¹ The idea is that “as between owners and infringers, it is more efficient for infringers to bear the costs of infringement.”⁴² Yet another familiar theme is that a strict liability standard serves an evidentiary function: it relieves plaintiffs of the burden of proving the infringer’s state of mind.⁴³

To understand how our strict liability system came to be, it is critical to consider its historical genesis. As others have noted, Judge Learned Hand first laid the foundations for a strict liability regime in a spate of early-twentieth-century decisions dealing with copyright infringement.⁴⁴ In these cases, Judge Hand dismissed the defendants’ claims of ignorance on the theory that the copyrights at issue were registered with the Copyright Office: the defendants, Judge Hand suggested, were on notice as to the existence of the copyright.⁴⁵ At the time these decisions were issued, copyright registration was mandatory.⁴⁶ Failure to register the rights would lead to copyright forfeiture.⁴⁷ Because the infringed works were registered, Judge Hand could reasonably surmise that the defendants had proper notice of the status of the works they had copied.

Thus, a strict liability regime was defensible on more functional grounds. It arose at a time when copyright registration was mandatory. Registration provided notice of a work’s legal status. And copyright protection was also conditioned on an additional notice affixed to copies of the work—every published copy had to include a notice specifying who the copyright owner was and when the copyright was registered.⁴⁸

But these formalities were eliminated or rendered optional in the late twentieth century.⁴⁹ Consequently, under our current system, it has become increasingly difficult to track down and identify the owners of certain works.⁵⁰ Copyright protection today vests automatically—authors need not take any affirmative steps to register their rights or apply for copyright protection.⁵¹ This complicates matters. Because copyrights vest automatically but are not registered, the risk of accidental infringement has grown measurably. It is indeed harder to justify a strict liability regime in a world where the risk of accidental infringement is so high.

Meanwhile, copyright’s strict liability regime has been the subject of sustained scholarly attention. A strict liability regime, scholars contend, is no longer workable. By increasing the likelihood that inadvertent infringers face liability, a strict liability scheme threatens to disproportionately expand the scope of copyright entitlements. Dane Ciolino and Erin Donelon, for instance, argue that strict liability frustrates the very objectives underlying our copyright system—it overprotects preexisting works, thus decreasing accessibility and preventing future authors from engaging with and building on existing works.⁵² Patrick Goold rails against copyright’s strict liability regime on moral grounds. Quite apart from any of the practical consequences that might attend a strict liability system, Goold explains that liability for accidental infringement is morally unjust—it is simply unfair to impose liability on a defendant who did everything they reasonably could to avoid infringement.⁵³ Avihay Dorfman and Asaf Jacob claim that strict liability makes little sense in a system focused on intangible goods—among other problems, intangible goods are difficult to define and their legal status is often uncertain.⁵⁴ Finally, Jacqueline Lipton asserts that strict liability aligns poorly with the realities of the digital era.⁵⁵ In a world where technology allows for the mechanical, accidental, and often involuntary copying of works, strict liability is inappropriate.⁵⁶

In 2012, Shyamkrishna Balganesh offered a more systematic account of copyright infringement.⁵⁷ As Balganesh points out, the basic structure of the infringement action does not require proof of harm. To make a prima facie case of copyright infringement, the plaintiff must prove (1) ownership of a valid copyright in the work and (2) infringement of one of the plaintiff’s rights, requiring both copying-in-fact and improper appropriation.⁵⁸

The question of harm is curiously absent from this basic scheme. As Balganesh notes, it is only later in the process, when confronting the question of fair use, that the issue of harm comes into play. Fair use is copyright’s most important defense to claims of infringement.⁵⁹ In determining whether the defendant’s copying qualifies as fair use, courts consider a range of statutory factors, including the question of “market harm”—the “effect of the use upon the potential market” for the copyrighted work.⁶⁰ So although harm is not an element of the infringement tort, it surfaces as part of the fair use inquiry. The issue of harm, then, is baked into the defense stage. Nevertheless, although the question of harm is central to the fair use analysis, the larger point remains: our copyright system is essentially a no-fault regime, one in which it “makes little difference for liability whether the copying was intentional, negligent, or a genuine mistake.”⁶¹

Others, though, question whether our system is truly a strict liability regime. In a series of articles published over the past few years, Goold articulated an intricate account calling into question the conventional view that copyright is best conceptualized as a strict liability regime.⁶² Goold posits that the infringement tort does, in fact, accommodate a fault element. This element, however, has gone largely unnoticed thanks to copyright’s messy structure.⁶³ The fair use principle is again central to the story: Goold suggests that the fairness standard reflected in the fair use inquiry is a kind of fault standard, somewhat analogous to the reasonableness standard undergirding negligence law.⁶⁴

But whatever one makes of Goold’s analysis, it is not at all clear that the fair use defense can meaningfully relax the no-harm regime that lies at the heart of copyright’s prima facie case. For many would-be infringers, the fair use defense is no silver bullet. The fundamental problem with fair use is that it relies on an ex-post analysis of four factors.⁶⁵ By design, the fair use inquiry is case specific and can only be tested by a court after the fact.⁶⁶ To prevail on fair use, the defendant must withstand lengthy and uncertain litigation.⁶⁷ For that reason, risk-averse defendants would often choose to settle out of court.⁶⁸ With the specter of statutory damages looming in the background, the risks of going to court are too grave. In such circumstances, the safe choice—and perhaps the only sensible one—is to settle.

The idea here is clear: whether the fair use doctrine introduces a fault standard, as Goold claims, is beside the point. For many would-be infringers, the fair use defense never really comes into play. And while it may be true that, as an analytical matter, copyright infringement is best characterized as a fault liability tort, this is largely tangential to the broader argument I advance here—that our current regime is overly burdensome.

B. Statutory Damages

A successful plaintiff in an infringement action may pursue one of two remedial avenues.⁶⁹ The first entitles a victorious plaintiff to collect actual damages plus infringement-related profits.⁷⁰ The second relieves a prevailing plaintiff of the burden of establishing actual harm, allowing instead for the recovery of statutory damages.⁷¹ Ordinarily, a plaintiff may recover statutory damages ranging from $750 to $30,000 for any infringed work.⁷² The number of statutory awards depends on the number of infringed works.⁷³ Each statutory award may be reduced to a sum of no less than $200 per work in cases of innocent infringement,⁷⁴ or increased—up to a sum of $150,000 per work—in cases of willful infringement.⁷⁵

As a practical matter, the Copyright Act offers little guidance on how courts might assess statutory damages.⁷⁶ As § 504 (“section 504”) of the copyright statute makes clear, the court is at liberty to adjust the award as it “considers just.”⁷⁷ In practice, in the absence of a jury trial, courts often assess damages by reference to four principal factors: (1) the harm suffered by the plaintiff; (2) the profits collected by the defendant; (3) the infringer’s state of mind, namely, whether the infringement was willful, knowing, or innocent; and, finally, (4) whether either of the parties had violated its contractual obligations.⁷⁸ More broadly, some courts also attend to the objectives animating our statutory regime. The Second Circuit, for example, often undertakes a more elaborate inquiry into “the deterrent effect on the infringer and third parties,” as well as “the infringer’s cooperation in providing evidence concerning the value of the infringing material” and “the conduct and attitude of the parties.”⁷⁹

Moreover, as noted above, the statutory award may be decreased to a sum of no less than $200 per work in cases of innocent infringement.⁸⁰ In such cases, the defendant bears the burden of proving that they were not aware, and had no reason to believe, that their actions were infringing.⁸¹ The burden here is twofold: the defendant must establish both that they did not believe their acts constituted infringement and that their belief was reasonable. Whether an award is ultimately reduced, though, remains a discretionary matter, and a court may refuse to do so even in cases involving innocent infringers.⁸² Finally, it is also important to note that the Copyright Act precludes reliance on the “innocent infringement” defense when a defendant had access to published copies bearing a copyright notice.⁸³

Similarly, in cases of willful infringement, the court may enhance the amount of statutory damages to a sum of “not more than $150,000” per work.⁸⁴ Whether the infringement was “willful” depends on the infringer’s state of mind.⁸⁵ And although a determination of willfulness can prove somewhat elusive,⁸⁶ the prototypical case of willful infringement involves a defendant who knowingly infringed the plaintiff’s rights⁸⁷ or otherwise displayed “reckless disregard for, or willful blindness to, the copyright holder’s rights.”⁸⁸ The point is that, even absent actual knowledge of infringement, the defendant may be held liable for willful infringement so long as they displayed reckless disregard of the plaintiff’s rights.

One strong indication of reckless disregard is a history of past infringements.⁸⁹ Another piece of evidence probative of willfulness is the infringer’s past record of seeking authorization prior to using other copyrighted works.⁹⁰ Indeed, a defendant who was previously vigilant in obtaining licenses but failed to do so at a later point may be deemed to have engaged in willful infringement. In addition, the Ninth Circuit has suggested that a jury could find willfulness where a company failed to promulgate any procedures for “establishing or reporting on who holds the rights to the [works] whose use is proposed.”⁹¹ And a court may likewise draw an inference of willful infringement when the defendant ignores a letter informing them, in concrete terms, of infringement.⁹²

At the same time, a host of countervailing factors might cut against a finding of willfulness. For one thing, when fair use is a close call, some courts insist that it would be inappropriate to infer willfulness even if the defendant is ultimately found liable.⁹³ Take the case of Princeton University Press v. Michigan Document Services, Inc., in which a university press sued a copy shop for selling unauthorized “coursepacks.”⁹⁴ The court concluded that the defendant’s copying did not constitute fair use⁹⁵ but took pains to note that the question was a close one—indeed, there were “forcefully argued dissents” on the issue.⁹⁶ As a consequence, the court declined to determine whether “the defendants’ belief that their copying constituted fair use was so unreasonable as to bespeak willfulness.”⁹⁷

Courts have also refused to countenance claims of willful infringement when the defendant had a reasonable belief that their actions were legally permissible, even if they had been notified that their use of the work was potentially infringing.⁹⁸ Furthermore, some plaintiffs may struggle to establish claims of willful infringement when the defendant did not consult a lawyer⁹⁹ or when the infringer performed a search of Copyright Office records but could not locate any records for the plaintiff’s work.¹⁰⁰

In short, a successful plaintiff may choose to recover statutory damages at any point before a final judgment is rendered. Typically, a statutory award can range from $750 to $30,000 for every infringed work.¹⁰¹ The award may be reduced to a sum of no less than $200 per work in cases of innocent infringement or increased to a sum of no more than $150,000 per work in cases of willful infringement.¹⁰² The number of statutory works is tied to the number of infringed works and, in some cases, the number of infringers.¹⁰³

This Section draws out two of the central problems ailing our copyright system. The first is that copyright infringement is principally understood to be a strict liability regime. And while this regime is tempered by the flexible fair use standard, it is also true that fair use is mostly irrelevant for a great many defendants—namely, those who might choose to settle out of court. The second major issue confronting our system is that copyright owners can choose to recover a separate award of statutory damages for every implicated work. Since these awards can range anywhere from $200 to $150,000 per work,¹⁰⁴ vast fortunes hang in the balance.

One might inquire, however, as to the interaction between these two aspects of our copyright system. While the defendant’s state of mind is wholly absent from the liability analysis, it does inform the damages calculus. As this Section makes plain, whether the infringer had real or constructive knowledge that their acts were infringing is key to determining whether they face reduced or increased statutory damages. Therefore, in a sense, these two features of our system—strict liability and statutory damages—work in unison. Theoretically, when liability is imposed on an innocent infringer, courts may choose to reduce the award under the category of innocent infringement (in which case, the award could be as low as $200 per work).¹⁰⁵ When one mechanism (strict liability) fails, judges can tap into the other (statutory damages) to prevent an unjust outcome—that is, to prevent an inadvertent infringer from facing an excessive award.

There is just one problem: in reality, the law of statutory damages is arbitrary, haphazard, and unpredictable. Part II will explore the law in action. It will show that different courts harbor very different ideas about what might qualify as an appropriate statutory award. Indeed, courts issue wildly divergent awards in factually similar cases. And, worse, courts cannot even agree on what statutory damages are meant to achieve: some believe statutory damages are rooted in a compensatory rationale, while others insist that statutory damages are meant to punish or deter future infringers. The result is a body of caselaw that often seems excessive or unprincipled.

II. STATUTORY DAMAGES IN ACTION

This Part discusses the principal issues plaguing the law of statutory damages. First, courts often disagree over the appropriate justification for statutory damages, oscillating between compensatory, punitive, and deterrence-based concerns. Second, the caselaw turns out to be unpredictable or otherwise arbitrary—in cases sharing a similar fact pattern, courts mete out divergent awards. In part, this is because various courts employ different multipliers to calculate damages. Third, statutory damages can at times lead to shockingly excessive awards. Because statutory damages are computed on a per-work basis, awards can dramatically rack up in cases that implicate multiple works. Fourth, and perhaps most bafflingly, the per-work structure of statutory damages can produce some unintended consequences: because damages attach to infringed works rather than infringed rights, courts tend to systematically gloss over certain exclusive rights. The net result is an underdeveloped body of law. I take up these issues in turn.

A. Justifications

In many cases, courts award statutory damages “with little to no attention . . . to their underlying purpose.”¹⁰⁶ Debates about statutory damages tend to home in on three standard justifications. First, some courts believe that statutory damages are grounded in a compensatory-evidentiary justification. The general idea is that copyright owners may struggle to prove actual harm.¹⁰⁷ Indeed, proponents of statutory damages believe that there is something distinctive about the evidentiary challenges facing copyright plaintiffs.¹⁰⁸ Imagine, for example, a copyright owner whose work was illegally distributed online and then downloaded by third parties. In such circumstances, a court might struggle to “assess how many of these third-party downloads dislodged actual sales that would have otherwise taken place if the work had been distributed by the plaintiff.”¹⁰⁹ In other instances, the information needed to prove actual damages might lie “uniquely within the infringers’ control.”¹¹⁰

Given these challenges, the argument goes, statutory damages serve a crucial role in ensuring that rightsholders get compensated for their losses. Without statutory damages, copyright holders would face an all-or-nothing regime; they would not be able to obtain compensation at all if they cannot prove actual loss. On this view, statutory damages are directed at a particular problem—the difficulty of proving actual damages—and should thus seek to account for actual harm. They should approximate, if only imperfectly, “the amount that would be recovered as actual compensation.”¹¹¹

This compensatory justification, simple and intuitive as it may seem, has found favor with many courts.¹¹² The legislative history, too, seems to indicate that compensatory instincts partly underlie our statutory scheme.¹¹³

Yet, as Bracha and Syed point out, the evidentiary justification suggests two related principles. First, if our goal is to address the difficulties of proving harm, statutory damages should only be available when evidence of actual harm is unavailable or difficult to obtain. Second, statutory damages should approximate actual harm as best as possible.¹¹⁴

The problem, of course, is that neither principle finds support in the caselaw or in the language of the copyright statute. Nowhere does the Copyright Act limit the availability of statutory damages only to circumstances where evidence of actual harm is difficult to recover.¹¹⁵ Courts, in turn, treat statutory damages “as a matter of unqualified right.”¹¹⁶ And statutory awards often seem wholly unmoored from compensatory instincts—courts have developed no rules to cap statutory damages based on an assessment of actual damages. In fact, courts sometimes calculate the proper compensatory sum and then multiply it to arrive at a statutory award that is, by definition, supracompensatory.¹¹⁷ The compensatory rationale, then, does not quite map onto the caselaw.

Another justification courts occasionally invoke is that statutory damages serve a punitive or retributive role—punishing the infringer for their wrongful conduct.¹¹⁸ Many courts endorse this idea of punishment as a justification for statutory damages.¹¹⁹ And while judges seldom clarify exactly what they mean in describing statutory damages as “punitive,” the basic idea is simple: it is fair for the wrongdoer to suffer in direct proportion to the wrong they inflicted. The wrongdoer deserves to be punished.

This punitive rationale meshes rather neatly with the language of the Copyright Act. The statute, after all, establishes a tripartite framework to distinguish among different categories of infringement—regular, innocent, and willful—based on the infringer’s mental state.¹²⁰ That is, the infringer’s level of culpability is measured by reference to their state of mind. Correspondingly, the appropriate remedy is adjusted in proportion to the infringer’s degree of culpability—namely, whether they had actual or constructive knowledge that their actions were infringing.

But the larger picture, again, is quite messy. As Bracha and Syed observe, it would be peculiar for our system to enforce a punitive rationale through copyright’s statutory damages scheme.¹²¹ That is because our copyright system already classifies certain acts as criminal offenses.¹²² Retribution is widely understood to be the classic justification for criminal sanctions.¹²³ Criminal enforcement is also subject to a host of procedural and substantive safeguards, as well as a heightened burden of persuasion.¹²⁴ Accordingly, one might question whether copyright’s scheme of civil remedies should serve any punitive purposes. Why should we impose punitive penalties for conduct that does not meet the heightened standard for criminal culpability under the Copyright Act?

Moreover, the caselaw remains inconsistent and unpredictable. Some courts appear to recognize both compensatory and punitive instincts as proper justifications for statutory damages.¹²⁵ Others embrace a more constricted view, insisting that punitive damages should be available only in cases of willful infringement.¹²⁶ And sometimes courts dismiss the punitive rationale altogether, implying that statutory damages that extend beyond the purely compensatory level reflect an impermissible “windfall” for the plaintiff.¹²⁷ The bottom line is that the courts’ treatment of the punitive rationale is unpredictable or otherwise inconsistent.

A third justification that courts occasionally marshal in support of statutory damages is a more intuitive one: deterrence. Typically paired with the punitive rationale, the need to deter future infringements figures prominently in many cases discussing statutory damages,¹²⁸ as well as jury instructions.¹²⁹ The corollary is that statutory damages, justified by the need for deterrence, should reflect a supracompensatory penalty.¹³⁰

Yet doubts abound here as well. First, the idea of deterrence, if taken to reflect a broad justification for statutory damages, tends to fit poorly with our statutory scheme. It is not clear why we would think deterrence appropriate in cases involving innocent infringers. When a case implicates accidental infringement—for example, a defendant who had no reason to believe that their acts were infringing—the concept of deterrence seems irrelevant. As Bracha and Syed ask, “Why would we want to deter people from engaging in reasonable behavior?”¹³¹

And even in cases of willful infringement, it is doubtful whether the deterrence rationale is necessarily appropriate. If statutory damages are truly about deterring individuals from engaging in wrongful conduct, it should not matter what their mental state is, be it willfulness, reckless disregard, or even malice. If we aim to discourage infringements, it should not matter whether the infringer held a particular mental state.

Second, courts and commentators rarely offer anything more than a bare-bones explanation for why one might think deterrence is even warranted.¹³² Is there any evidence to support the proposition that purely compensatory penalties cannot deter would-be infringers? What is the appropriate level of deterrence? And why should we want to achieve a level of deterrence that goes beyond what is already produced by compensatory damages? As Depoorter has explained, the most colorable justification for supracompensatory damages—and one that courts regularly skirt—is that there is a strong need for deterrence given the substantial costs of detecting and enforcing infringement.¹³³ The optimal level of deterrence, then, would depend on enforcement costs.

Third, as with the other justifications sketched above, the jurisprudence governing statutory damages has been unpredictable. Courts tend to vacillate between the standard justifications for statutory damages. While some adopt both punitive and deterrence-based rationales,¹³⁴ others are perfectly content to focus exclusively on compensatory impulses.¹³⁵

In sum, all three justifications for statutory damages are under siege from different directions. The caselaw is muddled and the statute is underspecified. Courts have struggled to articulate a proper justification for statutory damages and have largely failed to clarify why one justification is preferable to another. The jurisprudence thus boils down to a hodgepodge of conflicting and alternating views on what might qualify as a valid rationale for copyright’s robust scheme of statutory damages.

B. Predictability and Consistency

This much is clear: the caselaw is not a model of clarity. Critics have long criticized the law of statutory damages as indeterminate or otherwise arbitrary.¹³⁶ The problem is that courts appear to have only a rudimentary understanding of copyright’s remedial scheme, and they rarely attempt to apply the law in a manner that is consistent and predictable.

Begin with one of the central challenges in assessing damages: the use of multipliers. In calibrating statutory awards, courts and juries often use a simple, two-step formula. First, they identify the plaintiff’s actual loss—say, the licensing fee the plaintiff would have charged if their rights had not been infringed. Second, courts apply a multiplier to the amount identified in the first step. For example, if the lost fee was $5,000 and the court elected to use a multiplier of three, the final award would be $15,000 ($5,000 x 3).

Yet courts have done little to generate any sense of consistency in the application of multipliers. Cases among and within circuits reveal large discrepancies. For example, in R.A. Guthrie Co., Inc. v. Boparai, the court rejected the plaintiff’s argument that a multiplier of five would be appropriate.¹³⁷ Instead, the court identified a standard multiplier of two to three,¹³⁸ and then decided to apply a larger multiplier of four to account for the defendant’s willful infringement.¹³⁹ The court’s determination was based on its “sense of justice.”¹⁴⁰ The final award totaled $240,000.¹⁴¹

In another case, Strober v. Harris, a district court in Florida applied a multiplier of six.¹⁴² Notably, the multiplier here consisted of two elements: (1) a multiplier of three, designed to account for the defendant’s willful infringement; and (2) a “scarcity multiplier” of two, meant to account for the “unique” and “rare” attributes of the plaintiff’s work.¹⁴³ After assessing the lost licensing fee at $5,000, the court employed a multiplier of six to arrive at a total award of $30,000 in statutory damages.¹⁴⁴

Similarly, in Sadowski v. Primera Plana NY, the court invoked a more modest multiplier of three, explaining that the plaintiff did not assert that he ever notified the defendant of infringement.¹⁴⁵ Other courts have applied multipliers of two, five, and seven.¹⁴⁶ Of course, the practical difference between a multiplier of two and a multiplier of eight could be dramatic, depending on the plaintiff’s actual loss and the number of works at play.

But there is far more nuance to the caselaw. Some courts have suggested that double-digit multipliers—that is, multipliers of ten and above—are grossly excessive and potentially impermissible. Consider Affordable Aerial Photography, Inc. v. Palm Beach Real Estate, in which the court rejected the plaintiff’s request that a multiplier of thirty-six be used for each of its three infringed works.¹⁴⁷ The plaintiff argued for an award of $108,000 in statutory damages, based on an assessment of $1,000 in lost licensing fees for each of the plaintiff’s three works, multiplied by thirty-six ($1,000 x 3 x 36 = $108,000).¹⁴⁸ After rejecting the proposed multiplier, the court suggested that even a more modest multiplier of 14.4 would “create a windfall for the [p]laintiff.”¹⁴⁹

Or take the case of Markos v. Yacht Charters of Miami.com.¹⁵⁰ The plaintiff asked for three separate multipliers to be applied: a scarcity multiplier of three, a “quality . . . factor” multiplier of four, and a “willful [infringer]” multiplier of five, thereby settling on an aggregate multiplier of sixty.¹⁵¹ The court was quick to reject the requested multiplier as excessive, opting instead for a multiplier of three.¹⁵² Likewise, the court in Stockfood America, Inc. v. Fernando Arcay Special Events Corp. reasoned that a multiplier of twenty was excessive and thus inappropriate.¹⁵³ And a court in another case, Schwabel v. HPT Service, LLC, dismissed a multiplier of fifteen and instead applied a multiplier of three.¹⁵⁴

However, in a long line of cases, courts and juries have applied far larger multipliers—sometimes thousands of times the amount of actual damages. Think, for instance, of a case referenced earlier, Capitol Records, Inc. v. Thomas, in which a record label sued a defendant for illegally downloading twenty-four songs.¹⁵⁵ The jury ultimately awarded the plaintiff $9,250 per infringed song for a total award of $220,000.¹⁵⁶ This award was particularly jarring given the fact that, as the court emphasized, the plaintiff’s actual loss amounted to $50.¹⁵⁷ The resulting award, therefore, was roughly 4,000 times larger than the sum of the plaintiff’s actual damages.

In another notable case, Zomba Enterprises, Inc. v. Panorama Records, Inc., the court awarded the plaintiff $31,000 per infringed work for an aggregate award of $806,000.¹⁵⁸ The defendant argued that the award was grossly excessive—it was thirty-seven times the plaintiff’s actual damages.¹⁵⁹ On appeal, the Sixth Circuit upheld the award.¹⁶⁰ Other courts have variously used multipliers of 10, 20, and 300.¹⁶¹

Moreover, as Samuelson and Wheatland explain, inconsistent awards in factually similar cases are “easy to find.”¹⁶² Often, a series of cases brought by the same plaintiff (such as a record label) in connection with a set of nearly identical infringing acts can lead to wildly divergent awards.¹⁶³ And in many cases—especially ones involving excessive awards rooted in double-digit multipliers—the size of the final award seems inexplicable.¹⁶⁴ Equally troubling is the fact that cases dealing with similar categories of infringing acts sometimes yield markedly inconsistent awards.¹⁶⁵

And broader issues persist. Looking at the caselaw, one gets the distinct sense that courts do not have a firm grasp of copyright’s remedial scheme. A few different problems intersect. First, and perhaps most perplexingly, judges sometimes jump straight to the statutory maximum, even when the defendant was not especially culpable and the plaintiff suffered little to no harm.¹⁶⁶ Second, and relatedly, Samuelson and Wheatland find that courts often treat the statutory maximum as a starting point and then work backwards from there.¹⁶⁷ And in one of the cases discussed before, Zomba Enterprises, the court appears to have erroneously assumed that, if the defendant is found to be a willful infringer, the court must award above $30,000 in damages—which would explain why the court issued an oddly specific and otherwise atypical award of precisely $31,000 per work.¹⁶⁸

The picture, in other words, is a profoundly bleak one. Courts dole out inconsistent awards in factually similar cases. They cannot quite agree on the appropriate multiplier range. They do not understand the law’s remedial structure. And, as the previous Section showed, they disagree over the proper justification for statutory damages.

C. Excessive Awards

The law is not just inconsistent but can also produce outlandish statutory awards that seem entirely divorced from any conception of actual harm or culpability. Partly at fault for this state of affairs is the per-work structure of our remedial system. Even a reasonable award can become excessive thanks to the aggregation requirement: courts and juries are simply compelled to multiply the award by the number of infringed works.

A few examples might prove instructive. Consider again a case discussed throughout this Article, Capitol Records, Inc. v. Thomas.¹⁶⁹ The final award in that case was 4,000 times the sum of the plaintiff’s actual damages.¹⁷⁰ The jury assessed statutory damages of $9,250 per work, although the trial judge observed that the plaintiff’s actual damages—for the 24 works combined—amounted to about $50.¹⁷¹

It bears emphasizing, though, that even at the statutory minimum of $750 per work, the total award in Capitol Records, Inc. v. Thomas would have been excessive. Because the case involved 24 infringed works, and because statutory damages are awarded on a per-work basis, the jury would have to issue a minimum aggregate award of no less than $18,000 ($750 x 24).¹⁷² But recall again that the actual harm suffered by the plaintiff in that case was roughly $50. So even under the statutory minimum, the aggregate award would have been 360 times the plaintiff’s actual damages—an appallingly excessive award by any measure.

Now consider the procedural history of Capitol Records, Inc. v. Thomas. After the jury returned an award of $220,000 against the defendant, the trial judge vacated the judgment and ordered a retrial on the basis of a jury instruction error.¹⁷³ In doing so, the judge also cautioned that Congress did not intend for large statutory awards to be “applied to a party who did not infringe in search of commercial gain.”¹⁷⁴ Remarkably, however, the second trial turned out to be even more controversial: the jury increased the statutory award to $80,000 per work, for a final award of $1.92 million.¹⁷⁵ Following the second trial, the defendant filed a post-trial motion requesting, among other things, that the court reduce the jury award either by way of remittitur or under the Due Process Clause.¹⁷⁶ The court agreed. It lambasted the second award as “simply shocking” and remitted it to a sum of $2,250 per work, for a total award of $54,000.¹⁷⁷ As the court stressed,

despite the . . . justifications [for increased damages] and the Court’s deference to the jury’s verdict, $2 million for stealing 24 songs for personal use is simply shocking. No matter how unremorseful Thomas-Rasset may be, assessing a $2 million award against an individual consumer for use of Kazaa is unjust. Even Plaintiffs admit that Thomas-Rasset is unlikely to ever be able to pay such an award.¹⁷⁸

The record companies then exercised their right to seek a new trial on the question of damages.¹⁷⁹ The district court held a third trial, and the jury returned an award of $62,500 per work, yielding a total award of $1.5 million.¹⁸⁰ Once again, the defendant moved to amend the judgment.¹⁸¹ The district court again granted the defendant’s motion and reduced the award to $2,250 per work, leading to a final award of $54,000.¹⁸² The record companies appealed, and the Eighth Circuit reinstated the first award of $222,000.¹⁸³

What makes Capitol Records, Inc. v. Thomas particularly egregious is that the case appears to pit a poorly resourced defendant against a group of deep-pocketed plaintiffs that suffered little material harm. But the issues here were further compounded by the aggregation requirement. Since the defendant technically infringed the plaintiffs’ rights in twenty-four individual works, even a minimum statutory award would have been grossly excessive. Given copyright’s per-work structure, the prospect of crippling damages was virtually inescapable.

Another example is Columbia Pictures Television v. Krypton Broadcasting of Birmingham, Inc.¹⁸⁴ Columbia Pictures sued the owner of several television stations for broadcasting four copyrighted television programs without a license.¹⁸⁵ The court found for Columbia and entered an award of $20,000 per episode.¹⁸⁶ Because the infringed shows consisted of 440 episodes, the court issued a total award of $8.8 million.¹⁸⁷ After the Ninth Circuit affirmed the judgment, the Supreme Court reversed and remanded for failure to grant the defendant a jury trial on the issue of damages.¹⁸⁸ After a jury trial was convened, the jury returned a staggering award of $72,000 per infringed episode, for a total award of $31.68 million.¹⁸⁹ Here, too, aggregation proved fatal. Because each episode constituted a separate work, the per-work award had to be multiplied by 440, leading to a final award in excess of $30 million.¹⁹⁰

The number of infringed works is often a matter of bitter contestation for that reason precisely. Take, for example, Yellow Pages Photos, Inc. v. Ziplocal, LP.¹⁹¹ Yellow Pages Photos licensed stock photos for the yellow pages industry.¹⁹² It owned a library of thousands of photos that were sorted into themed collections.¹⁹³ Yellow Pages Photos brought action against two companies for infringing its rights in 178 collections that contained 10,411 photos.¹⁹⁴ The jury returned an award of $123,000 in statutory damages against one of the defendants.¹⁹⁵ On appeal, the plaintiff argued that the court erred in treating each of its 178 collections, rather than each of its 10,411 photos, as a separate work.¹⁹⁶ The relevant benchmark, the plaintiff asserted, was the number of individual photos, not the number of collections.¹⁹⁷ As the Eighth Circuit explained, “[b]y taking the position that its 10,411 individual photos are each separate works, [the plaintiff] presumably [sought] to raise the statutory damages award in this case from $123,000 to a minimum of $1.5 million and a potential maximum of $300 million.”¹⁹⁸ Ultimately, the court rejected the plaintiff’s claims, holding that each collection was an individual work under the Copyright Act.¹⁹⁹

Still, this case offers a neat illustration of the broader issues. Although the plaintiff’s claims were eventually rejected, the critical point is that the aggregation requirement presented a serious threat. The scope of the defendants’ potential exposure was startling: depending on how the court might have chosen to count the number of implicated works, the defendants could have faced an award ranging anywhere from $123,000 up to $300 million. Indeed, as this brief discussion demonstrates, our per-work system is broken. Statutory damages can add up quickly and dramatically in cases involving multiple infringed works. And that should give us pause.

D. The Copyright Wasteland

I have argued elsewhere that per-work statutory damages have another profound yet overlooked consequence: they discourage courts from properly developing copyright law.²⁰⁰ Modern copyright law, to be sure, is underdeveloped. Typically, courts focus on one exclusive right—the right to reproduce the copyrighted work—while glossing over all other copyright entitlements.²⁰¹ As a result, we do not have a clear sense of what some exclusive rights mean or how they might be applied. Courts have not really grappled with many questions central to the scope of these rights.²⁰²

One example is the exclusive right to distribute the copyrighted work.²⁰³ It is an open question whether the distribution right extends only to “actual dissemination,” namely, to circumstances where copies of the work were offered to and received by members of the public.²⁰⁴ Thus, in cases centering on claims of online distribution, it is unclear whether the plaintiff must show that copies of the work were not only uploaded online but also downloaded by third parties. District courts have been somewhat divided on the issue, although most courts now seem increasingly more hospitable to the view that distribution does, in fact, require proof of actual dissemination.²⁰⁵ But, stunningly, circuit courts have not weighed in on the question.²⁰⁶ Nor has the Supreme Court.²⁰⁷

Why is that the case? The answer is that courts never really have to address the issue. Virtually every copyright infringement case involves claims of unauthorized reproduction.²⁰⁸ And once courts establish that the right of reproduction has been infringed, all other rights turn out to be inconsequential. It does not matter how many separate exclusive rights were infringed; a single infringement suffices. That is because statutory damages attach to infringed works rather than infringed rights. So, it is irrelevant whether there is one right at issue or a few separate ones—the result, in terms of the final statutory award, will be the same.²⁰⁹ The only thing that matters is the number of infringed works, not infringed rights.²¹⁰

In that sense, there is little incentive for courts to properly address copyright’s full suite of exclusive rights. Instead, courts tend to fixate on one exclusive entitlement—the right of reproduction—to the exclusion of all others. And this pathology, again, is an artifact of copyright’s per-work scheme of statutory damages.

III. TOWARD STRUCTURAL REFORM

The discussion thus far offers a grim perspective. It paints a picture of a system run amok. It demonstrates that copyright’s scheme of statutory damages is troublesome for any number of reasons: it rests on ever-shifting and poorly developed justifications; it is rooted in a body of caselaw that is both arbitrary and inconsistent; it can lead to outlandish awards; and it allows courts to systematically shun certain exclusive rights.

Where does this leave us? I offer here a high-level sketch of one possible solution. In simple terms, I suggest that we disentangle the number of statutory awards from the number of infringed works. Instead, I develop a framework for assessing damages by reference to the concept of infringement episodes. An infringement episode, in a nutshell, is a series of related infringing acts that together make up a single, larger event. When courts identify a single infringement episode, they may choose to issue only a single statutory award—no matter how many individual works are in play. The proposed framework would thus introduce a degree of much-needed flexibility into our system. Rather than engage in a mechanical exercise of counting works, courts and juries would enjoy a significant measure of discretion to assess whether the defendant’s actions are sufficiently interconnected to constitute a single, larger episode.

This Part begins by offering a brief overview of the proposed framework. I explain that previous proposals have not gone far enough in confronting the main problem, and then describe in general terms how an approach rooted in the concept of infringement episodes might work. I next turn to demonstrate that a similar approach, based on a holistic assessment of the defendant’s actions, already exists and is widely employed in the fields of criminal law, civil procedure, and immigration law. I close by cashing out my proposed framework and developing a more fine-grained account of how courts might identify infringement episodes.

A. Overview

To reiterate, statutory damages are awarded on a per-work basis. And this is a problem for two related reasons. First, copyright’s per-work scheme forces courts to aggregate damages in ways that lead to unpredictable—and often excessive—awards. In cases like Yellow Pages Photos,²¹¹ for example, conflicting views on how to measure the number of works could potentially produce earth-shattering consequences. Under the defendants’ count, the statutory minimum was $120,000 to $135,000, and under the plaintiff’s, the

statutory award could have been as high as $1.5 billion.²¹² A divergence of this scope and intensity is indefensible. Second, as discussed above, per-work damages generate a legal equilibrium in which the number of infringed rights, as opposed to infringed works, is of little practical significance.²¹³ And that is why courts tend to glaze over certain exclusive rights. The result is an impoverished body of caselaw.²¹⁴

Per-work damages, then, have worked considerable harm to the copyright system. And copyright scholars, for their part, have suggested a number of doctrinal or legislative reforms to address the issue. Let us first consider what is arguably the most serious attempt to confront the problem of excessive damages in recent years: Depoorter’s multifaceted proposal for reducing the incidence and availability of excessive awards.²¹⁵ Based on a close analysis of docket records and court decisions, Depoorter finds that “a vast majority of plaintiffs accuse defendants of willful copyright infringement.”²¹⁶ Indeed, “[t]the sheer number of willful infringement claims is remarkable,” especially when one considers that enhanced damages were originally designed to address only exceptional cases.²¹⁷ Nevertheless, Depoorter also finds that, of all cases where the plaintiff won on the merits, courts awarded increased damages for willful infringement in only 2.8% of cases.²¹⁸ This glaring mismatch between what plaintiffs allege and what courts find suggests that plaintiffs engage in remedy overclaiming.²¹⁹

Based on these findings, Depoorter puts forward a number of prescriptive proposals. First, he advances a series of proposals meant to make remedy-overclaiming more costly to opportunistic plaintiffs. One way to do so is to impose sanctions on plaintiffs who make inflated or exaggerated damage claims.²²⁰ Under the current statute, a plaintiff seeking statutory damages need not justify their damage claims on the basis of actual harm.²²¹

And attorney cost fee-shifting is only available to the prevailing party,²²² so a defendant who lost but faced an outlandish award claim (which was ultimately rejected by the court) has no effective recourse. Depoorter thus recommends that Congress amend the Copyright Act to allow for fee-shifting in cases in which the plaintiff is ultimately found to have made inflated damage claims, even if the plaintiff prevailed in litigation.²²³

Second, Depoorter also presses a few proposals designed to make remedy-overclaiming less risky to defendants. One way to achieve this is by developing a set of guidelines and standards to govern the application and assessment of statutory damages.²²⁴ A second useful step would be to simply reduce the statutory range—which currently extends from $750 to $150,000 in most cases.²²⁵ Yet another helpful measure would seek to increase transparency by forcing courts to explain, in detail, the “motivation and calculation involved with every statutory award,” as well as by establishing a central database to collect statutory award judgments.²²⁶ Depoorter also suggests that Congress eliminate enhanced damages for willful infringement or otherwise limit their availability only to cases of commercial infringement.²²⁷ Another possibility is to impose a cap on the size of the total award.²²⁸ In addition, Congress should amend section 504(c) “so that statutory damages are not available when the defendant has offered credible evidence of its profits and/or the plaintiff’s damages.”²²⁹

Similarly insightful is recent work by Bracha and Syed.²³⁰ Focusing more keenly on the justification for statutory damages, they propose a few ways to think about the role that statutory damages ought to serve in the copyright system. One possible approach is to conceive of statutory damages as serving a compensatory role.²³¹ On this view, courts should strive to award statutory damages that approximate actual harm.²³² Alternatively, courts can frame statutory damages as a means of achieving optimal deterrence.²³³ But optimal deterrence, at least on its face, would seem to justify imposing very high awards on a small number of defendants in an effort to deter future infringement.²³⁴ Thus, an “optimal deterrence” approach would only be defensible to the extent that it is subject to “judicial safeguards against the incurrence of serious inequitable harms.”²³⁵

Two other reform proposals merit attention. One is a study by Pamela Samuelson and Tara Wheatland.²³⁶ Drawing on a comprehensive survey of the history and doctrine of statutory damages, Samuelson and Wheatland make two central claims: first, that statutory damages should ordinarily approximate actual damages; and, second, that enhanced damages should be available only in cases involving the sort of willful infringement Congress had originally contemplated—that is, patently outrageous cases implicating mass-infringement or repeat infringers.²³⁷

Another proposal, introduced by Garfield, focuses on three different approaches to calibrating statutory damages: interpretive, statutory, and constitutional.²³⁸ On the interpretive front, Garfield embraces the set of proposals articulated by Samuelson and Wheatland—proposals that would compel courts to approximate actual damages in most cases and reserve enhanced damages for nakedly egregious cases.²³⁹ But Garfield also notes that such an interpretive fix would not necessarily “translate[] into clear signals that potential users . . . can identify and confidently trust.”²⁴⁰ Garfield thus reasons that additional measures are necessary. On the legislative front, Garfield proposes an amendment of section 504 to (1) clarify that statutory damages should be primarily compensatory, (2) limit enhanced damages to exceptional cases, (3) empower courts to adjust the total award below the statutory minimum even in cases involving multiple works, and (4) give courts discretion to reduce the statutory minimum when a user has a strong innocent infringement claim, even if that user had access to works that bore a valid copyright notice.²⁴¹ Finally, on the constitutional front, Garfield recommends that courts reinvigorate and harness both Due Process and free-speech limits on statutory damages.²⁴²

There is much to admire in these proposals. To be sure, copyright scholars have devoted a great deal of energy to crafting proposals that would limit the scope of statutory damages. But none of these proposals go nearly far enough: they all stop short of severing the link between statutory damages and the number of infringed works. To meet the challenge of the moment, I suggest a legislative reform that would do away with copyright’s per-work remedies altogether. A better framework, I argue, would afford courts a significant degree of discretion to look beyond the number of implicated works and instead evaluate whether the defendant’s conduct gave rise to, and was grounded in, a single infringement episode. By infringement episode, I mean a chain of related infringing acts that together constitute a larger factual event. When the defendant’s conduct is attributable to a single infringement episode, courts should issue only a single statutory award, even if multiple infringed works are at stake.

To illustrate, consider Sony BMG Music Entertainment v. Tenenbaum.²⁴³ A group of record companies brought action against the defendant, Joel Tenenbaum, for downloading and distributing thirty songs.²⁴⁴ The jury found that Tenenbaum had willfully infringed the plaintiffs’ rights and returned an award of $22,500 per infringed song for a total award of $675,000.²⁴⁵ The defendant here was accused of having downloaded works en masse. But what if the court were able to treat the defendant’s conduct as reflecting a single infringement episode? And what if, because the defendant’s actions were sufficiently intertwined, the court were able to issue only a single statutory award?

There is reason to think a single award would have been appropriate in cases like Tenenbaum or Thomas.²⁴⁶ The value of the copyrighted works in both cases was negligible, and the harm to the plaintiff was marginal.²⁴⁷ In both cases, the defendants were private individuals who pursued a noncommercial objective by downloading songs for private consumption. And in both cases, the aggregation of per-work awards proved fatal. After all, a person of average means may never be able to withstand (or fully recover from) a damages award approaching $700,000.

Now, to be clear, courts today already consider some of these factors in calibrating the size of the per-work award.²⁴⁸ Yet the larger problem remains: courts are forced to multiply the award by the number of infringed works.²⁴⁹ So even if a court determines that the per-work award should be lower because the infringer was not especially culpable, the total award may nonetheless prove disproportionately large after the per-work award has been multiplied by the number of infringed works. Recall that in Tenenbaum, the per-work award itself ($22,500) was arguably quite modest, amounting to less than a fifth of the statutory maximum.²⁵⁰ But after it was multiplied by thirty, the award grew dramatically.²⁵¹ The rub is that even a modest per-work award can grow significantly due to aggregation.

To address the issue, I suggest that Congress amend the Copyright Act to permit courts to break with this per-work scheme in cases where they find, based on a multifactor inquiry, that the defendant’s acts were attributable to a single infringement episode. Looking for sources of inspiration, the next Section will explore how courts have approached analogous questions in different fields. And the third Section will then offer a more detailed explication of the proposed framework, discussing the different factors that courts might consider in evaluating the defendant’s course of conduct.

B. Doctrinal Analogues

An assessment of the defendant’s conduct, as explained previously, would have to depend upon an open-textured, case-by-case analysis. But would this system be administrable? Are courts well-positioned to engage in this type of multifactor inquiry? Would the proposed framework spawn lengthy and uncertain litigation?

The answer is no. Across a variety of different legal fields—including criminal law, civil procedure, and immigration law—courts have already developed a rash of somewhat equivalent doctrines aimed at evaluating the defendant’s conduct. These doctrines all share a core commitment: they leverage a flexible, multifactor inquiry to assess whether the defendant’s actions can be lumped together as part of a larger transaction, episode, scheme, or series of occurrences. The following discussion will offer a perfunctory and necessarily incomplete overview of such doctrines. These doctrines, though at times controversial, demonstrate that the proposed framework is both plausible and administrable.

1. Criminal Law

The defendant’s course of conduct features most notably in the doctrine of double jeopardy. The Double Jeopardy Clause of the Fifth Amendment shields a defendant from being “twice put in jeopardy of life or limb” for the same offense.²⁵² The goal of the prohibition against double jeopardy is to protect an individual from “being subjected to the hazards of trial and possible conviction more than once for an alleged offense.”²⁵³ The Supreme Court has interpreted the Double Jeopardy Clause as protecting against a second prosecution for the same offense after either acquittal or conviction.²⁵⁴ More broadly, the Double Jeopardy Clause has also been understood to preclude multiple punishments for the same offense.²⁵⁵ In other words, the central question under the Double Jeopardy Clause is whether the defendant is being punished or tried twice for the same offense.

How do courts define the term “same offense”? Federal courts have long relied on a decades-old test first developed in Blockburger v. United States²⁵⁶ to determine whether the defendant is at risk of being tried twice for the same offense. Under the Blockburger test, when the same act violates two separate statutory provisions, the court is charged with examining “whether each provision requires proof of a fact which the other does not.”²⁵⁷ In essence, courts seek to ascertain whether one or two offenses are at issue. This same-elements test examines the statutory elements of each offense to determine whether one of the implicated offenses incorporates an element that the other does not. Accordingly, if a single act or transaction leads to multiple offenses—namely, if one offense requires additional proof beyond what is required by the other—the Double Jeopardy Clause does not apply.

Yet, ultimately, whether one is placed twice in jeopardy for the same offense is not just a matter of federal law. Many states also encode a robust protection against double jeopardy, either through constitutional provisions or through state statutes requiring joinder of certain offenses.²⁵⁸ And the picture here is more intricate. Some states, like Michigan, apply the Blockburger test.²⁵⁹ Other states, like Louisiana, employ a double-pronged test for defining “same offense”: a variation of the Blockburger same-elements test, coupled with an additional same-evidence test.²⁶⁰ Under the same-evidence test, courts examine whether the evidence necessary to support the second indictment was also sufficient to support a conviction for the first.²⁶¹

A number of states, however, reject the Blockburger same-elements test in favor of a more flexible test: the same transaction or course of conduct test. Consider, for instance, New Jersey. As the New Jersey Supreme Court has explained, courts should consider several different factors in assessing whether the defendant’s misconduct is part of a single transaction, including whether the offenses are “based on the same conduct or arose out of the same episode.”²⁶² And New Jersey courts, in turn, have applied this “same episode” test by considering a number of additional factors, including:

[T]he nature of the offenses, the time and place of each offense, whether the evidence supporting one charge is necessary and/or sufficient to sustain conviction under another charge, whether one offense is an integral part of the larger scheme, the intent of the accused, and the consequences of the criminal standards transgressed.²⁶³

Other states similarly consider whether the defendant’s actions are causally related and whether they were temporally close.²⁶⁴ Additionally, a 2017 study found that “[t]hirteen states apply no individual test, but rather multiple factors, to prohibit successive prosecutions.”²⁶⁵ These states use some combination of several existing tests—temporal and spatial tests, “same transaction” tests, evidentiary/same-elements tests, and others—to carry out a more open-ended, multifactor inquiry.²⁶⁶

To conclude, different states diverge radically in their approach to protecting against double jeopardy. Nevertheless, a few relevant factors stand out as particularly relevant. Some factors hinge on the defendant’s conduct: the spatial and temporal proximity of the defendant’s acts; whether they fit into a larger episode or transaction; and the nature of the causal relationship between each of them. A second set of factors look at the legal directives that were violated—do these provisions share the same elements and require the same evidence to establish guilt? A third category of factors lean more broadly on the defendant’s intent and the consequences of the defendant’s conduct.

2. Civil Procedure

The rule of permissive joinder, codified in Rule 20 of the Federal Rules of Civil Procedure, provides that multiple plaintiffs may join in one action if (1) they assert a right to relief arising out of the “same transaction, occurrence, or series of transactions or occurrences,” and (2) “any question of law or fact common to all plaintiffs will arise in the action.”²⁶⁷ A similar rule of permissive joinder applies when multiple defendants face actions arising out of a single series of occurrences and involving questions of law and fact common to all defendants.²⁶⁸ Rule 20 is permissive in that it does not compel joinder even in circumstances that satisfy its requirements.²⁶⁹

As a general matter, the rule of permissive joinder is applied liberally and flexibly. The Ninth Circuit, for instance, has adopted a liberal approach in an effort “to promote trial convenience and to expedite the final determination of disputes, thereby preventing multiple lawsuits.”²⁷⁰ And the Supreme Court has clarified that “[u]nder the Rules . . . joinder of claims, parties and remedies is strongly encouraged.”²⁷¹ It is also important to note that joinder is a procedural device and does not affect the substantive rights of the parties.²⁷² Hence, a judgment for or against one party need not lead to a similar outcome for another party in a joined action.²⁷³

More concretely, the first requirement under Rule 20—that the disputes arise out of the same series of occurrences—is most relevant for our current purposes. This requirement appeals to an open-ended assessment of the parties’ entire course of conduct. But, consistent with the flexible spirit animating Rule 20, courts have largely balked at developing “one generalized test” for identifying a single transaction or occurrence.²⁷⁴ Instead, courts leverage a case-by-case approach to examine whether the events at stake are all logically related and thus constitute a single series of occurrences.²⁷⁵ This logical-relationship test defies any rigid formulation and is flexible by design. Moreover, some courts complement the logical-relationship test with a variation of the evidentiary standard that criminal courts employ under the double jeopardy rule. They do so by asking whether different legal actions would lead to “overlapping proof and duplication in testimony.”²⁷⁶ When such overlap is likely, courts are more inclined to find that the claims arise out of the same transaction or occurrence.

Given the flexible nature of the permissive-joinder inquiry, it is hardly surprising that courts sometimes find that events stretching over many years could nonetheless be sufficiently connected to warrant joinder. One example is Burton v. American Cyanamid.²⁷⁷ In Burton, the plaintiffs alleged harm as a result of ingesting lead as children, arguing that their injuries over the years constituted a single series of occurrences that resulted from the defendants’ negligence in manufacturing, promoting, and selling lead paint.²⁷⁸ The court agreed and ordered joinder.²⁷⁹ In another case, a federal court ordered joinder of more than 400 defendants in an action involving some 400 separate insurance and retirement plans, finding that the claims, while addressing many different plans and defendants, all arose from the same series of transactions or occurrences.²⁸⁰ The rules of permissive joinder, in short, rest on a flexible standard and are thus applied somewhat liberally.

3. Immigration

Questions about the defendant’s course of conduct also crop up under the Immigration and Nationality Act.²⁸¹ The Act renders an alien deportable if, at any point after admission, they are convicted of two or more crimes of moral turpitude not arising out of a “single scheme of criminal misconduct.”²⁸² One of the central questions in this context is whether the alien’s convictions could be characterized as arising out of a single scheme of misconduct. While the statute does not define the phrase “single scheme of . . . misconduct,”²⁸³ some early authorities have suggested that a variety of factors might play into an analysis of the alien’s conduct, including the time and purpose of the crimes, the methods and procedures used, the identity of the participants, and the identity of the victims.²⁸⁴

Nonetheless, clouds of uncertainty enshroud the single scheme exception. Some circuit court cases, as well as decisions issued by the Board of Immigration Appeals (“B.I.A.”), take a restrictive approach. They do so by suggesting that, to qualify for the “single scheme” exception, the alien’s crimes must be temporally close.²⁸⁵ Likewise, the First Circuit has stated that a single scheme “must take place at one time; there must be no substantial interruption that would allow the participant to disassociate himself from his enterprise and reflect on what he has done.”²⁸⁶ These authorities come close to equating a “single scheme” with a “single act”—a uniform, “temporally integrated episode of continuous activity.”²⁸⁷

Other courts, by contrast, find that separate crimes committed under a preconceived plan—even if the two crimes were days apart—could qualify as a single scheme of misconduct.²⁸⁸ These courts suggest that two crimes executed in pursuit of a common plan could constitute a single scheme.²⁸⁹ And while this “common plan” approach has been rejected by the B.I.A. and the First, Fourth, Fifth, Seventh, and Tenth Circuits,²⁹⁰ it has been adopted to varying degrees by the Second, Third, and Ninth Circuits.²⁹¹ As a consequence, there remains a great deal of disagreement and confusion as to the scope of the “single scheme” exception.

* * *

In summary, each of the doctrines canvassed above seems to lend some support to the suggested framework. Each seeks to identify the circumstances under which a series of wrongful actions might give rise to a single transaction, episode, or scheme. And each appears to mobilize a case-by-case approach. Courts applying these doctrines often consider both the specific attributes of the defendant’s actions and the consequences associated with the larger scheme.

At the same time, it would be a mistake to overstate the significance of this cross-disciplinary survey. The preceding discussion does not aim to provide a fully exhaustive synthesis of three discrete bodies of law. Rather, it seeks only to highlight a few rough, high-level sources of inspiration that could bear in some loose sense upon the proposed framework.

C. Application

Copyright’s per-work scheme has been the source of much anguish. It tends to produce grossly excessive awards, and it discourages courts from properly developing modern copyright law. A better framework, I argue, would empower courts to look beyond the number of infringed works and focus instead on the defendant’s entire course of conduct. When a series of separate infringing acts can be traced to a single, larger infringement episode, courts should be able to issue only a single statutory award—regardless of how many works are at issue.

But how might courts evaluate the defendant’s course of conduct? In what follows, I map out a number of relevant factors that courts should consider, including the spatial and temporal nature of the defendant’s actions; whether the evidence supporting one infringement is necessary or sufficient to sustain liability for another; whether the defendant’s actions were executed in pursuit of a common plan or a larger creative project; and the potential consequences of a statutory award.

These factors operate along three interrelated dimensions: conduct/evidence, intent, and consequences. Some factors focus on the evidence required to distinguish between—or bring together—the defendant’s actions. A second subset of factors look to the defendant’s intent in pursuing these actions. A third subset of factors dwell on the larger consequences of the defendant’s actions.

Let us consider a few illustrative examples. Think, for instance, of Yellow Pages Photos.²⁹² A local phone book publisher, Ziplocal, licensed thousands of photos owned by the plaintiff.²⁹³ The licensing agreement authorized Ziplocal to distribute the photos to its users but prohibited the company from transferring the photos to outside companies or nonemployees.²⁹⁴ A few years later, Ziplocal subcontracted with an outside company to assist with editing the photos and producing a phone book.²⁹⁵ The plaintiff brought action for infringement, and the jury found Ziplocal liable for infringing the plaintiff’s rights in 123 individual works.²⁹⁶

The court, though, could have reasonably concluded that the 123 infringed works were all implicated in a single scheme—a commercial effort by the defendant to produce a phone book. Perhaps more crucially, it seems as though both parties understood the defendant’s actions to constitute a single episode. The plaintiff, after all, never identified with any degree of specificity the circumstances under which particular works were infringed.²⁹⁷ The plaintiff did not bother to break down the broader episode into smaller, concrete actions that the defendant carried out.

This becomes all the more apparent when the case is assessed against the same-elements/same-evidence question: Is the evidence supporting one infringement necessary or sufficient to sustain liability for another? Again, the answer here is yes. The plaintiff never adduced any evidence to separate one infringing act from another. Instead, the defendant’s liability was ultimately based on evidence relating to the broader episode rather than any particular infringing acts. Under the same-evidence test, then, the defendant’s actions could be characterized as arising out of a single infringement episode.

The same is true for some file-sharing cases, such as Tenenbaum.²⁹⁸ In Tenenbaum, the defendant was accused of illegally downloading and distributing thirty sound recordings.²⁹⁹ But, again, the plaintiffs here failed to present any evidence to distinguish among discrete actions. Consider the plaintiffs’ pretrial motion, which discusses at some length the infringed works and the plaintiffs’ process for detecting them. As the pretrial motion explains, the plaintiffs engaged an external company to assist in detecting infringement.³⁰⁰ The company was able to identify the defendant by pinpointing the IP address of his computer, and the company was also able to ascertain that the defendant had used a file-sharing software called Kazaa to illegally download sound recordings.³⁰¹ But the plaintiffs’ evidence did not identify when the infringed works were downloaded—it could only establish that, at some point, the defendant’s publicly viewable Kazaa folder contained and listed the infringed works.³⁰² So the plaintiffs could show that the defendant engaged in a long-running scheme of downloading songs via file-sharing software, but they could not identify any particular actions the defendant took at any particular point in time. The upshot, once again, is that a court may reasonably conclude that the defendant’s actions here—as reflected in the evidence presented and in the plaintiffs’ failure to identify particular acts—were reducible to a single infringement episode.

Now, to take another example, consider the case of Cariou v. Prince.³⁰³ Patrick Cariou published a book of portraits depicting Rastafarians in Jamaica.³⁰⁴ A few years later, a well-known appropriation artist by the name of Richard Prince tore out several of Cariou’s photos, altered them, and incorporated them into a series of paintings and collages that he exhibited at art galleries.³⁰⁵ Cariou sued Prince for copyright infringement, alleging that Prince had infringed his rights in dozens of separate photos.³⁰⁶ The court held that twenty-five of Prince’s artworks made fair use of the infringed photos.³⁰⁷ But let us assume, for the sake of argument, that Prince was held liable for his actions involving dozens of infringed works.

Could a court conclude that Prince’s actions arose out of, and were grounded in, a single infringement episode? The common plan test suggests a reason to think so. As the Second Circuit ultimately acknowledged, Prince’s actions were carried out in pursuit of a larger creative plan. Prince incorporated Cariou’s photos into a series of artworks displayed as part of an exhibition, called “Canal Zone,” and these works were later bound up and published in an accompanying exhibition catalog.³⁰⁸ In executing his creative vision, Prince altered the infringed photos and used them to create an exhibition that was far removed from Cariou’s original work. Prince’s exhibition “manifest[ed] an entirely different aesthetic,” turning Cariou’s serene portraits into “crude and jarring works.”³⁰⁹ So while Prince used a large number of Cariou’s works, these works were all incorporated into a larger project conveying a different brand of creative and social commentary. In other words, Prince’s actions arose out of a larger creative project—and were pursued in compliance with a preconceived plan. One could thus conclude that Prince’s infringing acts constituted a single infringement episode and may warrant only a single statutory award.

This suggests that the defendant’s intent, at least to the extent that it involves a larger creative enterprise, should inform the analysis. Why inquire into the defendant’s intent? The general idea is that the defendant’s creative motivation should matter because copyright law is fundamentally about encouraging creativity.³¹⁰ Nowhere is this more apparent than in the context of fair use. The most important factor in the fair use analysis is whether the purportedly infringing use is transformative.³¹¹ The defendant’s use is transformative when it introduces a new purpose, meaning, or message.³¹² This means that, generally speaking, injecting new creative expression into an original work is likely to be privileged as transformative.³¹³ The central point, then, is that courts should follow a similar path when adjusting damages. They should consider whether the defendant engaged with an existing work for the purpose of pursuing a new creative project—this is, after all, precisely the kind of creativity that our law seeks to foster.³¹⁴

In addition, there is another set of considerations that should bear on an assessment of the defendant’s conduct: the potential consequences of a statutory award. Here, courts need to consider two separate questions. The first is whether the defendant’s conduct prevented cost recoupment by the plaintiff. Did the plaintiff already recoup the costs of creating and distributing their works? If so, courts should be more inclined to find that the defendant’s actions derive from a single episode.

To see why, consider the dominant justification for copyright law in the United States: the incentive-access tradeoff. On this account, the copyright system provides authors with incentives to create intellectual works.³¹⁵ Because intellectual goods are non-rivalrous and non-excludable,³¹⁶ authors need some form of legal protection to prevent others from free riding on their creative efforts.³¹⁷ Without copyright protection, copyists will be able to distribute cheap copies and undercut the author’s prices. Consequently, authors will not be able to recoup the costs of producing their works, and so they may choose not to create at all.³¹⁸ Cost recoupment is central to this story. By providing authors with a legal entitlement to exclude copyists, copyright aims to ensure that authors are able to recover their costs.³¹⁹ Copyright law, then, provides authors with an economic incentive to invest in the production of intellectual works.

But copyright protection comes at a price. When a copyright owner enjoys legal exclusivity, they can increase the price they charge for access to their work.³²⁰ This means that more consumers will be priced out of the market.³²¹ And the problem runs deeper: by limiting access to existing works, copyright law runs the risk of frustrating, rather than encouraging, creativity.³²² Indeed, authors often create new works by engaging with existing ones.³²³ Therefore, if copyright law seeks to make good on its promise of encouraging creativity, it must ensure that copyrighted works are ultimately made accessible to current and future users.³²⁴ That is why copyright protection must be limited in scope and duration—it must “protect authors only to the extent necessary,”³²⁵ while allowing for “the creation of new works that build upon earlier ones.”³²⁶ To do so, our law should strike a balance between two conflicting interests: the need to incentivize authors on the one hand, and the need to ensure that works are accessible to society at large on the other.³²⁷

Consistent with this understanding of copyright law, it is crucial to figure out whether the copyright owner was able to recoup their costs. And that question should inform not only the duration and scope of copyright entitlements;³²⁸ it should also interact with the question of remedies. My core claim, in other words, is that in assessing whether to multiply the number of awards, courts should be able to confront the question of cost recoupment. If the plaintiff suffered little harm and reaped enormous profits—and if there is no risk the plaintiff might fail to recoup their costs—courts should find that the defendant’s actions constitute a single episode.³²⁹

Admittedly, the question of cost recoupment extends beyond the defendant’s immediate conduct. But assessing cost recoupment makes sense as a matter of copyright policy: it allows courts to account for the underlying objective of our system.

Finally, courts should also give pride of place to various distributional concerns. To assess whether it would be reasonable to treat the defendant’s conduct as arising out of a single scheme, courts must consider the distributional effects of an aggregate statutory award. As Bracha and Syed explain, distributional concerns should typically provoke a distinction between corporate and individual defendants.³³⁰ By and large, when the defendant is a corporation or a firm with dispersed liability, “neither distributive nor aggregation concerns are likely to prove worrying enough to require any [judicial] safeguard[s].”³³¹ However, when the defendant is a private individual, courts should ask whether a statutory award—multiplied by the number of infringed works—would “so eat into an average individual’s income or wealth”³³² as to trigger significant distributive concerns. A hefty award directed at an individual of average means may well “have serious effects in areas such as housing, health, education, and more generally the ability to make and pursue basic life choices.”³³³

It follows, then, that courts should account for the identity (and relative wealth) of the defendant in considering whether their conduct is best characterized as reflecting a single, larger event. When the defendant is a private individual of moderate means—like the defendants in Thomas or Tenenbaum—it would be more appropriate to conclude that the defendant’s actions were the product of a single episode.

In conclusion, the proposed framework turns on a number of related factors. It attends not only to the defendant’s immediate conduct, but also to some of the broader distributional and utilitarian concerns that underpin our system. In particular, it directs courts to take into account the spatial and temporal nature of the defendant’s actions; whether the evidence supporting one infringement is necessary or sufficient to sustain liability for another; whether they were executed in accordance with a common plan or a larger creative project; and the potential consequences of a statutory award.

IV. OBJECTIONS

This Part briefly considers three potential objections: that the proposed system draws inspiration from the wrong sources; that it fails to alleviate the indeterminacy that ails our law; and that it is poorly suited to the task of compensating aggrieved plaintiffs. For various reasons, none of these objections is particularly compelling. Below I explain why.

A. Doctrinal Transplants

One preliminary objection is that the doctrinal analogues I have identified above are irrelevant or otherwise inapplicable. The basic idea is that these doctrines emerged in very different settings and were designed with different goals in mind: streamlining civil proceedings, protecting criminal defendants, and demarcating grounds for deportation. These doctrines, put simply, serve different objectives, are subject to different limitations and judicial safeguards, and should not be applied to copyright lawsuits. Accordingly, it is not clear why these discrete bodies of law should bear in any meaningful sense on copyright remedies.

Yet this objection misfires for a number of reasons. One is that, as clarified above, I do not mean to suggest here that these doctrinal analogues are directly applicable to the law of copyright remedies. The point, rather, is that these doctrines could be treated as rough sources of inspiration. A doctrinal analogue is just that—an analogue. And while the proposed framework does incorporate some notable doctrines borrowed from other fields—the same-evidence test, the temporal question, and the common plan test—I nonetheless chose to pass over certain other doctrines.

I should also clarify that, in order to account for the underlying objectives of our system, I propose that we reorient some of these doctrinal analogues. Take the “common plan” test used by courts to determine whether an alien is deportable for committing two crimes of moral turpitude.³³⁴ Under this test, some courts ask whether the two crimes were executed in pursuit of a common, preconceived plan. When that is the case, the two crimes are thought to have arisen out of a “single scheme of criminal misconduct.”³³⁵ But under the framework developed here, the “common plan” test would be expanded to cover a larger creative enterprise.³³⁶ This is important because, as discussed above, modern copyright law seeks to stimulate precisely this type of creativity—namely, to allow users and authors to pursue creative projects.³³⁷ It would thus make sense for courts to evaluate the defendant’s creative motivation when calculating remedies. Even if the defendant is found to have infringed the plaintiff’s rights—such that the defendant’s conduct cannot be shielded by the fair use doctrine—we may still think it desirable to treat such infringements as less culpable so long as the defendant pursued a creative vision.

Moreover, in a broad sense, the framework envisioned here is a rather familiar one. Though the suggested system relies on a smattering of different factors, many of these factors are ones that courts already consider. It is just that courts typically do so when considering the size of the award, rather than the number of available awards. Applying these factors, then, should not require much adaptation—courts and juries have long attended to these factors anyway. Consider, for instance, the distributional issues I discussed above, which courts today sometimes assess by asking whether a large award would “put [the defendant] out of business.”³³⁸ In so doing, courts impose a proportionality requirement—a requirement that the final award not be so large or disproportionate as to put the defendant out of business. My suggestion, then, is that courts simply consider this factor in one additional context: not just when calibrating the size of the award but also when evaluating whether to issue one award or multiple ones.

So, to sum it up, my choice of looking outward for inspiration is defensible. I suggest that we do so by devising a careful and nuanced framework, one that accounts for the ultimate objectives of the copyright system. And, essentially, this system is structured around an assortment of factors that courts already consider in different contexts, both within and outside of copyright law.

B. Indeterminacy

Another objection is that the proposed system would not resolve the problem of indeterminacy. As everyone seems to agree, the caselaw is a mess. Courts issue divergent awards in similar cases and rarely offer any explanation for doing so. The caselaw, then, turns out to be inconsistent and arbitrary. And the proposed framework, in turn, does little to confront this issue. Instead, it affords courts and juries even more discretion, asking them to carry out a flexible, holistic inquiry on the basis of several unweighted factors. Won’t that make matters worse?

One possible response is that the proposed system will not make things any worse than they are now—the law is already an unmitigated disaster. As others have observed, courts tend to issue awards that are simply inexplicable, often dispensing radically different awards in similar circumstances.³³⁹ So even if my proposed system will not yield a measurable improvement in consistency, it is also unlikely my framework will make matters any worse. We’ve already hit rock bottom.

In any event, it is important to remember that a host of other issues bedevil our system. In a system driven by per-work aggregation, statutory awards can swell up dramatically.³⁴⁰ The result is that some defendants face a significant risk of excessive awards. Copyright’s per-work structure also discourages courts from developing substantive law.³⁴¹ And it is here that one can clearly see the benefits of a more flexible system: if we decouple the number of awards from the number of infringed works, at least in cases where the defendant’s conduct resulted from a single infringement episode, the twin problems of aggregation and underdevelopment are likely to diminish. Given that these problems derive from per-work remedies, we could mitigate their effects by allowing courts to break from our per-work scheme. This means that, while my approach cannot tackle all of the relevant issues, it can surely contribute to eliminating or addressing some of them.

This suggests a related point: I am not claiming here that the proposed framework is a catchall solution. Rather, my proposal must be accompanied by a suite of additional reforms designed to overhaul our remedial scheme. Many of the proposals other scholars have advanced—such as reducing the statutory range,³⁴² imposing a statutory cap,³⁴³ tying statutory damages to actual losses,³⁴⁴ and imposing sanctions on plaintiffs who make inflated claims³⁴⁵—could work in tandem with my proposed system. While these proposals seek to reorient or limit the courts’ discretion in adjusting the size of the award, my proposal would allow courts to do away with per-work damages altogether.

C. Compensation

One final objection is that the proposed scheme would make it harder for plaintiffs to obtain compensation for the actual harm they suffered. As noted above, most copyright scholars agree that courts are insufficiently attentive to the rationales underlying statutory damages.³⁴⁶ And while it is clear that Congress intended for courts to award enhanced damages only in a narrow subset of “exceptional cases”—those involving large-scale infringement or repeat transgressors—it is also clear that, in most other cases, statutory damages were meant to serve a largely compensatory role.³⁴⁷ Thus, with one limited exception, statutory damages were meant to ensure that copyright owners would be compensated for their losses.

Yet there is a problem lurking beneath the surface: if statutory damages were engineered to serve a compensatory role, courts would have to account for the number of infringed works. After all, each work represents a distinct source of harm.³⁴⁸ A copyright owner will often charge a licensing fee for access to each of their works. To account for the plaintiff’s lost fees, then, the court will have to take stock of each and every work at issue. My proposal, however, could frustrate this process. By severing the connection between the number of awards and the number of works, the suggested system threatens to deny plaintiffs the opportunity to obtain proper compensation for their actual damages.

But is that really the case? As I have emphasized before, there is more than one way to compensate a copyright holder for the harm they sustained.³⁴⁹ One way is to adjust the number of awards based on the number of infringed works. But there is another possibility, and perhaps an equally effective one: “control[ing] for the number of infringed works by increasing or reducing the size of the aggregate award” within the statutory range.³⁵⁰ Courts and juries have at their disposal an incredibly robust statutory range stretching from $200 to $150,000, depending on the type of case at hand. And they can increase or reduce the total award within that range based on the number of works infringed and the defendant’s level of culpability.

For example, suppose the defendant copied three of the plaintiff’s photos. Now suppose that the plaintiff typically charges a licensing fee of $1,000 for each of their photos, meaning that the plaintiff’s total loss is $3,000. In adjusting the award, the court can pursue one of two possibilities: it can issue three separate awards of $1,000 each; or, alternatively, the court can dispense a single, aggregate award in the amount of $3,000. In other words, the court should be able to ensure full compensation by calibrating a single, aggregate award within the statutory range.

Nevertheless, there is an additional wrinkle here: if courts are to rely on a single, aggregate award, the statutory range has to be sufficiently broad to accommodate cases involving a large number of works or a small number of particularly valuable ones. To be sure, there may be cases where so many works are at stake—hundreds or even thousands—that a single award, even at the statutory maximum of $150,000, simply won’t do. For example, if the defendant copied 5,000 songs, each valued at $50, the plaintiff’s total loss would amount to $250,000 (5,000 x $50). This means that a single award of $150,000 simply wouldn’t suffice. Or imagine a case implicating 25 works, each valued at $10,000—again, a total loss of $250,000 would exceed the statutory maximum. What this suggests is that the statutory range must be robust enough to accommodate an array of possible scenarios.

These concerns, however, are overblown. First, even under the proposed approach, courts may nonetheless find that a mass-infringement event does not qualify as a single infringement episode. Perhaps such an event would be better described as a series of separate episodes, each warranting an additional statutory award. Exceptional cases call for exceptional remedies. Second, there is good reason to believe that such cases would be vanishingly rare. The vast majority of copyrighted works command little market value, if any, and the current statutory range would likely be sufficiently broad to ensure proper compensation in a large majority of cases.³⁵¹ In a sense, any concrete cap on the amount of damages would be arbitrary—we simply have to draw the line somewhere. But given that most works have little value, the current statutory limit seems more than sufficiently robust to accommodate the typical infringement case.

CONCLUSION

The law of statutory damages is in a state of disrepair. It often produces “arbitrary, inconsistent, unprincipled, and grossly excessive awards.”³⁵² Courts command “a contested, somewhat obscured, and even outright confused” understanding of copyright’s remedial system.³⁵³ And statutory damages are susceptible to pervasive “overclaiming” across “virtually all areas of copyright law.”³⁵⁴ But most reform proposals to date have papered over the real issue. So entwined are statutory damages in our copyright system that any attempts at serious reform may seem hopeless. And although commentators uniformly agree that statutory damages present a problem of colossal proportions, they have not yet been able to tackle the core problem: per-work damages.

This Article attempts to do just that. It seeks to jumpstart a conversation about structural reform and offer a roadmap for legislative and judicial action to confront the risk of inflated damages. It proposes, in short, that we do away with per-work damages altogether. It develops an alternative system that would accord judges a significant degree of discretion to look beyond the number of infringed works. Courts could do so by examining whether the defendant’s actions are collectively attributable to a larger infringement episode. When the defendant’s conduct arises out of a single episode, courts should issue only a single statutory award—no matter how many individual works are at issue. If adopted, this system would reduce the risk of excessive awards, prompt courts to properly adjust damages, and introduce a degree of pragmatism into our system.

This framework may seem radical. That is so in part because it appears to mandate a sharp break from a centuries-old system of per-work damages. But, in a sense, the proposed system is, in fact, a familiar one—after all, courts have forged similar doctrines to address analogous problems across a variety of otherwise distinct areas of law, including criminal law, civil procedure, and immigration law.

In the end, this Article envisions a world without per-work damages. What distinguishes the proposed system from other proposals is that it takes seriously the core issue: per-work remedies make it too easy for courts and juries to aggregate damages in ways that lead to outlandish awards. After thirty years of faint-hearted debates, it is well past time that policymakers try something different.

97 S. Cal. L. Rev. 1029

Download

Shani Shisha*

* Assistant Professor, SMU Dedman School of Law; Fellow, Berkman Klein Center for Internet & Society, Harvard Law School, 2021–22. For valuable feedback and generative conversations, I thank William Fisher, William McCoy, Guy Rubinstein, Brian Soucek, Oren Tamir, and Rebecca Tushnet. I received helpful comments from participants at the Harvard Law School Art of the Response Paper Workshop. I am also grateful to Vice Dean Gabriella Blum and Assistant Dean Catherine Peshkin for providing institutional support for this project. I am indebted to the editors of the Southern California Law Review, particularly Madeline Goossen, Ariana Croll, Mariem Masmoudi, Carus Newman, and Chloe Williams, for superb editorial work. Finally, I would like to thank the Harvard Law School Graduate Program for generous financial support.