A Radical Restructuring? A Response to Bell and Parchomovsky, Restructuring Copyright Infringement, 98 Texas. L. Rev. 679 (2020)

Response - Online Edition - Volume 99


In Restructuring Copyright Infringement, Abraham Bell and Gideon Parchomovsky call for a “radical reform in the way copyright law assigns liability.”[1] The authors propose to split copyright infringement into three distinct categories—inadvertent, standard, and willful—and to tailor the remedies offered to the copyright holder in each case. Under this system, the inadvertent infringer would be subject to less severe penalties than the standard and willful infringer. Overall, this proposal is interesting, timely, and contains a good deal of common sense.

However, is it truly a “radical” proposal? In my opinion, the article’s recommendations do not go far enough. Rather than reforming copyright remedies, the underlying liability rule needs a more significant overhaul to meet the challenges of the twenty-first century. Elsewhere, I have argued that IP law (copyrights and patents) must be reformed around a negligence principle.[2] In the Information Age, it has become alarmingly easy to accidentally infringe IP rights. For reasons of efficiency, equity, and fairness, our law ought to hold accidental infringers liable only when they have failed to adopt reasonable care to prevent the infringement. Accidental infringers who have acted with due care ought to be subject to no liability at all.

In Part I of this response, I summarize the Bell and Parchomovsky remedial proposal in greater detail. In Part II, I highlight the major strengths of the proposal. In Part III, I turn to the weaknesses of the proposal. I conclude by arguing that copyright law could more efficiently solve the accident crisis by adopting a negligence liability rule.

I. A Summary of the Proposal

Copyright law employs a “one-size-fits-all strict liability regime.”[3] Defendants will be held liable regardless of whether their infringements were intentional, negligent, or entirely innocent. Take, for example, the challenges faced by publishing companies. Palgrave Macmillan is a highly respected international publisher of academic texts. But strangely, many of its books start with an apology. Open one of Palgrave’s recently published books and within the first pages you may find the following statement: “While every care has been taken to trace and acknowledge copyright, the publishers tender their apologies for any accidental infringement where copyright has proved untraceable.”[4] Palgrave offers this boilerplate apology upfront because it understands that it will be held strictly liable for any infringement contained in its books, even when it has taken “every care” to prevent that infringement. When this happens, Palgrave will be subject to the same set of remedies as a wilful infringer, i.e., actual damages, statutory damages, disgorgement of profits, seizure of infringing copies, and injunctions.[5]

There is one exception to the uniformity of remedies. Under U.S. copyright law, courts have the discretion to reduce statutory damages to $200 per infringed work in cases of “innocent” infringement.[6] However, it is questionable how often this provision is used in practice. Pamela Samuelson and Tara Wheatland argue that the innocent infringer provision has “virtually no significance” in litigation and is “essentially never used” by courts.[7] Furthermore, this provision does not apply if the copyright owner elects to receive actual damages or disgorgement of profits. And courts continue to grant injunctions more or less routinely in these cases.

Bell and Parchomovsky propose to introduce some more nuance into copyright’s remedial practices. The reform proposal would involve formally separating copyright infringement cases into three types. The first type is “inadvertent infringement.” These are infringements which take place either subconsciously (i.e., where the defendant copies without awareness of their copying) or when the copying is “so small or questionable that a court could conclude that the infringer reasonably believed in good faith that nothing protectable was copied.”[8] In these cases, the infringement should result in compensatory damages only. On the other end of the spectrum is the second type of infringement, i.e., “wilful infringement.” This category includes “blatant and inexcusable” infringements like piracy.[9] In this category, courts should be able to award the “full panoply of copyright remedies.”[10] Lastly, in the middle is “standard infringement.” The standard infringer is “not blameless” but equally “not as culpable” as the wilful infringer.[11] This would include cases where a defendant has, for example, an arguable claim to fair use but has mistakenly gone too far and copied too much. In such cases, courts should award the full range of monetary remedies but typically withhold injunctive relief. All infringements would be “presumptively standard” unless either “the defendant could prove her innocence or the plaintiff could prove willfulness or malice.”[12]

The advantages of this proposal, according to Bell and Parchomovsky, are twofold. First, the proposal has “intuitive moral appeal.”[13] It seems unfair to subject the innocent infringer to the same remedies as a deliberate infringer. But even more importantly, the reform would help fulfil the utilitarian objectives of federal copyright law. The authors highlight the problem of “risk of legal liability” in copyright.[14] The cause of such risk is the “substantial information costs” produced by the copyright system.[15] In the twenty-first century, a vast amount of creative material is protected by copyright; the copyright in that material lasts for a very long time; the scope of the copyright is often uncertain (due to vague doctrines such as substantial similarity and fair use); and without mandatory formalities such as registration and notice requirements, it is frequently impossible for actors to determine the legal status of works ex ante. Against this backdrop, limiting the remedies that the inadvertent infringer faces would have two effects. First, it would reduce “unnecessary barriers to entry by new authors.”[16] And second, the proposal would incentivize copyright holders “to make their claims known as widely as possible, so as to foreclose the possibility of innocent infringements.”[17] In a nutshell, if copyright owners are unable to claim the full panoply of remedies from innocent infringers, then they will have a stronger incentive to engage in activities (like registration and notice) that reduce the probabilities of innocent infringement and increase the chances of ex ante licensing deals being reached.

II. Strengths

Bell and Parchomovsky’s proposal is clearly a step in the right direction. Indeed, a number of us have been working on the same problem for some time. In 2010, Avihay Dorfman and Assaf Jacob released a paper called Copyright as Tort.[18] In the article, the authors highlighted how easy it is for individuals to infringe copyright accidentally due to underlying features of the copyright system such as long copyright terms and the vagueness of various copyright doctrines. They argued that copyright was therefore closer to the paradigm of “accident law” (i.e., the law of negligence in tort law) than the paradigm of “trespass.”[19] In 2012, Stewart Sterk also highlighted that negligence liability rules ought to be used more frequently in IP (and property generally) as a way to deal with the information costs leading to inadvertent infringement.[20] In patent law, Tun-Jen Chiang made a similar argument.[21]

Subsequently, Oren Bracha and I published Copyright Accidents in 2016.[22] This article built on the prior literature in two directions.

First, we refined the concept of “copyright accidents” or “accidental infringement” of copyright. We argued that the key conceptual feature of accidental infringement (which is shared by other types of accidents found in tort law) is the idea of “ex ante risk”.[23] Users like Palgrave, mentioned above, cannot easily ascertain the legal status of the material they wish to copy. In such cases, users cannot bargain for permission ex ante because they do not know whether the work is protected or who the owner is. They are therefore faced with a situation where copying entails a risk of infringement (but not a certainty).

Second, we used the economic literature from accident law to determine what is the most efficient liability rule to govern such accident cases. In a nutshell, we argued that accidental infringement ought to be governed by a negligence liability rule.[24] The key advantage to a negligence liability rule is that it sets incentives for optimal bilateral care. Under a strict liability rule, the user has an incentive to adopt an efficient level of care to avoid accidental infringement. Unfortunately, the copyright owner does not have a similar incentive. Because copyright owners can receive damages in cases of accidental infringement, they have insufficient incentives to engage in activities, such as registering their work with the Copyright Office or attaching notice, that would help reduce accidental infringement to an appropriate level. This problem is solved under a negligence regime. Under a negligence regime, users still have an incentive to take reasonable levels of care because doing so helps them avoid liability. But, unlike strict liability, owners now have an incentive to also take care. Since they are no longer compensated automatically in cases of accidental infringement, their best financial strategy is to engage in activities that reduce the probability of accidents happening in the first place and increase the chances of ex ante licensing deals being reached.

Since 2016, I have expanded upon this accidents problem in IP. In a forthcoming monograph, I argue that IP, particularly copyright and patents, requires serious reform to meet the challenges of the twenty-first century.[25] In the nineteenth century, the probability of accidental infringement was relatively low. As explored previously by R. Anthony Reese, and expanded upon by me, the nineteenth-century patent and copyright regimes contained numerous “safeguards” that limited the possibility of accidental infringement, such as mandatory notice and registration.[26] With such safeguards in place, strict liability in IP was the appropriate rule. Illustratively, Judge Learned Hand went as far as to say that the legitimacy of strict liability in copyright was conditioned upon the presence of such safeguards. Without the registration and notice provisions of the copyright system, “it could not be a tort innocently to copy a copyrighted work,” he said.[27] However, strict liability is clearly not the appropriate liability rule in the twenty-first century. In the past 150 years, the “safeguards” that operated to reduce the probability of accidents have been eroded. The consequence is that creative risk and levels of accidental infringement have skyrocketed. Much as tort law developed the tort of negligence to deal with the skyrocketing of accident litigation in the nineteenth century,[28] today IP law must adopt a negligence liability rule to efficiently, equitably, and fairly apportion responsibility in cases of accidental IP infringement. Accidental infringers of copyright and patents should be liable for their infringements only if they have failed to take a reasonable level of care to prevent the infringement.

As a result, I agree with Bell and Parchomovsky’s proposal in a number of ways. They are clearly right to highlight how high levels of risk lead to cases of inadvertent infringement (although I would prefer we use the term “accidents” to highlight the relationship between these accidents and their cousins in tort law). They are also right that leniency, in some cases, needs to be shown to the infringer in order to encourage bilateral care and thus minimize the social cost associated with IP accidents. And one way to achieve that outcome is indeed to change the remedial structure of copyright. All of this strikes me as sound and perhaps even uncontroversial.

III. Weaknesses

So how ought we treat the accidental infringer? What solution will best encourage copyright owners and copyright users to take efficient levels of care and thus minimize the social costs associated with accidents? Putting aside the moral claims for the moment (due to space constraints),[29] will the proposal adequately achieve the utilitarian policy of copyright? The answer is “probably not.”

Despite some language to the contrary, Bell and Parchomovsky are not recommending any changes to the liability rule in IP. The proposal as it relates to accidental infringers is to impose liability upon the infringer strictly, regardless of whether the user has taken reasonable care or not, and to require the infringer pay compensatory damages. This is the very definition of a strict liability rule![30] The proposal is therefore not really to change the liability rule but merely to impose a less draconian set of remedies on the infringer within the confines of a strict liability regime.

Because Bell and Parchomovsky propose that the accidental infringer ought to be held strictly liable, the proposal is unlikely to achieve its stated aims. Under this proposal, the copyright owner will be compensated for the infringement every time that someone accidentally infringes their copyright. Why then would the copyright owner go to the expense and effort of trying to avoid the infringement? Taking care, such as registering their work, or ensuring that appropriate notice appears on all copies, etc., requires a certain amount of diligence, expenditure, and work on the part of the copyright owner. Why would the copyright owner bother engaging in these ex ante activities when instead they can simply wait for the infringement to happen and claim damages ex post?

To illustrate, consider the case of the University of Michigan’s Orphan Works Project (OWP). In May 2011, the University of Michigan announced it would begin to digitise out-of-print books from its library.[31] The project would increase worldwide access to rare books and save some works from obscurity. But the project ran into troubles. It was difficult to determine whether the works were protected by copyright and, if so, who owned the rights. To avoid this problem, the university performed a search for any potential copyright owners, published a list of the suspected ownerless works online, and called for rightsholders to come forward.[32] When no copyright owners came forward, the university began to digitise the books. Later, the Authors Guild, a collective of American authors, alleged the project infringed its members’ copyrights and began legal action to halt digitisation. Mired in a legal quagmire, the project was suspended indefinitely.[33]

The standard economic theory of accidents tells us that allowing the owner to claim compensatory damages in cases of accidents, like that involved in the OWP case, will result in the owner taking less than optimal care.[34] Consider the situation that the university found itself in. It was considering whether to digitise the out-of-print books. At that point in time, there was a certain probability (p) that digitisation would result in the accidental infringement of someone’s copyright causing financial loss (A) to the copyright holder (measured in terms of lost license fees). Both the university and the copyright owner could, however, take steps to reduce the probability of an accident (or pA). The university could search for the copyright owner and make its plans public. Doing so would cost a certain amount of time and money (Bu).[35] Meanwhile, the copyright owner could register her rights with the Copyright Office and respond to the university’s calls for information about the works. This would likewise cost a certain amount of time and money (Bo).

Now consider the effect that a compensatory damages rule would have on the parties’ incentives for care. Under such a rule, the university would be required to pay damages which are equal to the financial loss suffered by the copyright owner. Therefore, the university internalizes the costs of its precaution and the probable cost of the accident (via damages). Assuming it is trying to minimize its own private costs, the university will select a level of care that minimizes the sum of both Bu + pA. Thus, we can expect a rule of compensatory damages to encourage the university to take an efficient level of care to prevent such accidents. However, what about the copyright owner? Under such a rule, the copyright owner does not internalize the probable costs of the accident (pA) because these costs are fully internalized by the university. The copyright owner only internalizes the costs of her own precaution (Bo). Again, assuming she is trying to minimize her own private costs, the owner will select a level of care that minimizes Bo. In other words, the owner will take no care at all as this minimizes her private cost. Thus, under a compensatory damages rule, the copyright owners in the OWP case still have an incentive to sit back and wait to claim damages ex post rather than doing the hard work ex ante to prevent the infringement.

Consider how the situation would change under a negligence liability rule. Under such a rule, the university would only be held liable if it failed to take all reasonable care (where reasonable care is defined using the Hand formula, i.e. care is reasonable when B < PL). Under such a rule, the university will still take reasonable care to avoid the infringement. If it fails to take reasonable care, then it internalizes both the expected costs of the accident (pA) and the cost of whatever precaution it does adopt (Bu). By contrast, if it adopts reasonable care, it only internalizes the cost of its care (Bu). This sets an incentive for it to adopt a reasonable level of care and thus avoid liability. In practical terms, we can predict that even under a negligence liability rule, the university would continue to act carefully (by searching the register and publicizing their intentions etc.) in order to avoid paying damages.

By contrast, in cases where the user takes care, the copyright owner internalizes the remaining expected accident costs (pA) and whatever care she adopts (Bo). Therefore, she has an incentive to select a level of care that minimizes the sum of Bo + pA. Or more simply, in a case like the OWP one, the copyright owner could predict that under a negligence liability rule, the university would behave reasonably, and thus she would not be compensated for any lost license fees she suffers. In this case, her best strategy is to take reasonable car, such as registering and responding to calls for information, to avoid the accident and increase the chances of reaching licensing deals ex ante.

Thus, unlike the Bell and Parchomovsky solution of imposing a compensatory damages rule on the user, a negligence liability rule will lead to optimal incentives for bilateral care. Yet, Bell and Parchomovsky only cursorily consider the role that negligence ought to play within copyright at the end of the article and dismiss this option without a particularly clear reason.[36]


What is animating the Bell and Parchomovsky proposal is just how draconian copyright remedies currently are. Today, U.S. copyright law is in the unenviable position that it likely overcompensates copyright owners for the harm they suffer. What Bell and Parchomovsky have suggested is changing the rules so that copyright owners in cases of accident are merely compensated for their loss, rather than enjoying the supra-compensatory rents that statutory damages provide. This is clearly a step in the right direction, but it does not go far enough. Holding accidental infringers strictly liable and expecting them to compensate the copyright owner in all cases is better than the current status quo, but it still will result in copyright owners taking insufficient care to prevent accidents ex ante and merely waiting to claim damages ex post. A better solution would be to adopt a negligence liability rule. Under this rule, negligent users should pay compensatory damages, but careful and reasonable users should pay nothing. This would give both the copyright owner and copyright user an incentive to take reasonable care to avoid accidents and thus minimize the costs of copyright accidents. And, as explored in IP Accidents, this is not the only benefit of reforming IP law around a negligence liability principle.

  1. .Abraham Bell & Gideon Parchomovsky, Restructuring Copyright Infringement, 98 Texas L. Rev. 679, 683 (2020).
  2. .See generally Patrick R. Goold, IP Accidents: Negligence Liability in Intellectual Property (Cambridge University Press 2021) (forthcoming); Oren Bracha & Patrick R. Goold, Copyright Accidents, 96 B.U. L. Rev. 1025 (2016); Patrick R. Goold, Patent Accidents: Questioning Strict Liability in Patent Law, 95 Ind. L.J. 1075 (2020); Patrick R. Goold, Moral Reflections on Strict Liability in Copyright, 44 Colum. J. L. & Arts 123 (2021); see also Patrick R. Goold, Is Copyright Infringement a Strict Liability Tort?, 30 Berkeley Tech. L. J. 305 (2015).
  3. .Bell & Parchomovsky, supra note 1, at 679.
  4. .See e.g., Ralph Hall, Applied Social Research: Planning, Designing and Conducting Real-World Research iv (Kellie Hughes, ed., 2008) (published by Palgrave Macmillan); see also Sotirios Sarantakos, Social Research xxiii (4th ed. 2013) (published by Red Globe Press, a Palgrave Macmillan subsidiary).
  5. .See 6 William F. Patry, Patry on Copyright § 21:38 (describing copyright infringement as “a strict liability tort . . . for which there is joint and several liability”); 17 U.S.C. §§ 502–04 (providing such remedies for all copyright infringement and differentiating willful infringement only as to the possibility of increased statutory damages). See also Chavez v. Arte Publico Press, 204 F.3d 601, 607 (5th Cir. 2000) (stating that “[c]opyright infringement actions . . . ordinarily require no showing of intent to infringe” and noting that such knowledge or intent is “relevant in regard to damages”).
  6. .17 U.S.C. § 504(c)(2) (2018).
  7. .Pamela Samuelson & Tara Wheatland, Statutory Damages in Copyright Law: A Remedy in Need of Reform, 51 Wm. & Mary L. Rev. 439, 453–54, 460 (2009).
  8. .Bell & Parchomovsky, supra note 1, at 684–85.
  9. .Id. at 685.
  10. .Id. at 686.
  11. .Id. at 685.
  12. .Id. at 686.
  13. .Id.
  14. .Id.
  15. .Id.
  16. .Id. at 686.
  17. .Id. at 686–87.
  18. .Avihay Dorfman & Assaf Jacob, Copyright as Tort, 12 Theoretical Inquiries L. 59 (2011).
  19. .Id. at 93–94.
  20. .Stewart E. Sterk, Strict Liability and Negligence in Property Theory, 160 U. Pa. L. Rev. 2129, 2136 (2012).
  21. .Tun-Jen Chiang, The Reciprocity of Search, 66 Vand. L. Rev. 1, 5 (2013).
  22. .Bracha & Goold, supra note 2.
  23. .Id.
  24. .Id. at 1029.
  25. .Goold, IP Accidents, supra note 2.
  26. .R. Anthony Reese, Innocent Infringement in U.S. Copyright Law: A History, 30 Colum. J. L. & Arts 133, 135 (2007).
  27. .Haas v. Leo Feist, Inc., 234 F. 105, 107 (S.D.N.Y. 1916); see also Hein v. Harris 175 F. 875 (S.D.N.Y. 1910) (noting that whether or not defendant heard complainant’s song is irrelevant in determining if an infringement of the copyrighted song occurred); Stern v. Jerome H. Remick & Co., 175 F. 282, 282–83 (S.D.N.Y. 1910) (finding that a demonstration of defendant’s intent to violate a copyright was unnecessary where he had means to discover if a copyright existed).
  28. .Goold, IP Accidents, supra note 2, at chapter 3.
  29. .But considered further in Goold, IP Accidents supra note 2, at chapter 5.
  30. .See, e.g., Steven Shavell, Economic Analysis of Accidents Law 8 (2009) (“Under strict liability injurers must pay for all accident losses that they cause.”).
  31. .Authors Guild, Inc. v. Hathitrust, 755 F. 3d 87, 92 (2d Cir. 2014).
  32. .Id.
  33. .Id.
  34. .See Shavell, supra note 30, at 5–46 (Harvard University Press 2009); Robert Cooter & Thomas Ulen, Law and Economics 201–04 (6th ed. 2016).
  35. .More specifically, taking more care would require an additional marginal expenditure. The greater care would correspond to a marginal reduction in the expected accident costs. Whether the marginal expenditure on care is cost-justified depends on whether it is higher or lower than the marginal reduction in expected accident costs.
  36. Bell & Parchomovsky, supra note 1, at 731–32; cf Bracha & Goold, supra note 2, at 1040–62.