Notice Failure And Defenses In Trademark Law

Preparing to load PDF file. please wait...

0 of 0
100%
Notice Failure And Defenses In Trademark Law

Transcript Of Notice Failure And Defenses In Trademark Law

NOTICE FAILURE AND DEFENSES IN TRADEMARK LAW
ROBERT G. BONE*

INTRODUCTION ............................................................................................. I. THE NATURE AND EXTENT OF NOTICE FAILURE IN TRADEMARK LAW ................................................................................................... A. Notice in General ...................................................................... B. Notice in Trademark Law .......................................................... II. THE COSTS OF NOTICE FAILURE IN TRADEMARK LAW ..................... A. Costs in General ........................................................................ 1. Discourage Innovation......................................................... 2. Impede Licensing ................................................................ 3. Increase in Information Costs.............................................. 4. Increase in Litigation-Related Costs.................................... B. Evaluating the Sufficiency of Trademark Notice ....................... 1. Socially Valuable (Favored) Uses ....................................... 2. Probability of Chilling Socially Valuable Uses...................
III. SOLUTIONS TO NOTICE FAILURE IN TRADEMARK LAW..................... A. Targeted Defenses ..................................................................... 1. Rules Versus Standards ....................................................... 2. The Difficulty with Categorical Rules in Trademark Law......... ............................................................................. a. Defining Favored Uses ................................................. b. Justifying Categorical Exclusions ................................ B. Sanctions and Fee-Shifting ........................................................ C. Modifying Liability Standards ................................................... 1. Limiting Actionable Confusion ........................................... a. Inadequate Justification ............................................... b. Inadequate Efficacy ...................................................... 2. Adjusting the Likelihood-of-Confusion Test ......................
CONCLUSION .................................................................................................

1246
1249 1249 1252 1256 1257 1257 1258 1259 1259 1262 1263 1265 1267 1267 1268
1271 1272 1274 1280 1282 1282 1283 1286 1286 1292

Notice is important in intellectual property law, but notice works differently in trademark law than in copyright and patent. Trademark law is not based on

* G. Rollie White Professor of Law, The University of Texas at Austin School of Law. I am grateful to the participants in the Notice and Notice Failure in Intellectual Property Law Symposium at Boston University School of Law and to Rebecca Tushnet in particular for helpful comments and suggestions on an earlier draft of this article. I am also grateful to Alexandra Fulcher for her able research assistance.
1245

1246

BOSTON UNIVERSITY LAW REVIEW

[Vol. 96:1245

a property theory and the rights it creates are not property rights in the usual sense. As a result, the costs of notice failure are not necessarily as high in trademark as in patent and copyright, and solutions to notice problems take a somewhat different form. This Article examines notice problems in trademark law, identifies their source, evaluates their costs, and explores ways to remedy them. The principal notice deficiency has to do with uncertainty in determining the scope of trademark rights (rather than the existence or boundaries of protected subject matter), and the main source of the problem is the vague likelihood-of-confusion test for liability, which produces high litigation costs and risks. These costs and risks can chill the descriptive and expressive use of marks, the use of marks in resale markets, and the use of trade dress to compete in product markets. The seriousness of these chilling effects, however, depends on the user and the type of use.
After linking notice failure to chilling effects, the Article then explores possible ways to reduce the chilling effect problem. In this regard, an important distinction must be drawn between uses that implicate rights and uses that are valued only for efficiency reasons. Categorical exclusions work well for the former but not very well for the latter (except perhaps for productdesign trade dress). For cases involving non-expressive descriptive and nominative uses and resale uses, trademark owners should be required to plead (and prove) sufficient facts to support the existence of actual source or sponsorship confusion (not just a likelihood) and a significant risk of serious trademark-related harm, and judges should review preliminary injunction motions strictly. These reforms should go far toward reducing chilling effects while still allowing trademark relief in those exceptional cases involving serious confusion-related harm.

INTRODUCTION
While notice of legal rights and obligations is important generally, it is especially important in property law. Reasonably clear notice of property ownership, boundaries, and rights facilitates efficient investment and exchange. When ownership and boundaries are clear, those developing property can be confident that their investments will benefit their own property and not their neighbor’s, and those acquiring property can be confident of what they are buying. In the case of land, the physical nature of the property helps considerably with identifying boundaries; title registries simplify the search for conflicting claims, and the standardization of ownership interests makes valuation easier and reduces transaction costs.1 Notice, however, is much more difficult for intellectual property. The nonrivalrous nature of IP greatly

1 Thomas W. Merrill & Henry E. Smith, Optimal Standardization in the Law of Property: The Numerus Clausus Principle, 110 YALE L.J. 1, 33, 43 (2000) (“Because land is permanent and immobile, recording can give notice to prospective purchasers not just of the physical dimensions of land but also of the legal dimensions . . . .”).

2016]

NOTICE FAILURE AND DEFENSES

1247

multiplies the number of potentially conflicting claimants; the abstract nature of the subject matter complicates boundary determinations, and the complexity and vagueness of IP rights make valuation difficult and transactions costly.2 Moreover, IP registries, when they exist, can be incomplete, or so crowded that searching is extremely costly.3 The resulting notice problems impede upstream innovation, chill downstream creativity, encourage strategic holdup, and compound litigation costs.
These problems are particularly acute for patent and copyright because both types of IP grant broad rights against copying and enforce those rights through injunctive relief.4 Without clear notice of competing claims, prospective innovators face a number of risks that can discourage investment in innovation. For one thing, they are likely to have difficulty estimating the value of the IP they create when that value depends on the uncertain existence of conflicting claims.5 Also, they face the risk of holdup. Owners of conflicting rights might wait until innovation costs are sunk and then demand a high licensing fee. This is the problem with many so-called non-practicing entities (“NPEs”), or more derisively, patent trolls. NPEs collect patents only to assert them against unsuspecting inventors and obtain disproportionately high settlements.6
Trademark law is different than patent and copyright in ways that require a different analysis of notice problems. In contrast to patent and copyright, trademark is not based on a property theory in the usual sense.7 At its core, trademark law protects marks not as things of value in themselves, but as

2 See, e.g., JAMES BESSEN & MICHAEL J. MEURER, PATENT FAILURE: HOW JUDGES, BUREAUCRATS, AND LAWYERS PUT INNOVATORS AT RISK 53-72 (2008) (comparing IP with real property and noting that “[t]here are sharp differences between patents and tangible property regarding the law and institutions that promote clear notice”); Peter S. Menell & Michael J. Meurer, Notice Failure and Notice Externalities, 5 J. LEGAL ANALYSIS 1, 2-7 (2013) (“[Intangible resources] can be difficult to navigate because of the amorphous nature of intangible boundaries . . . .”).
3 BESSEN & MEURER, supra note 2, at 68-71.
4 Patent law goes even further and grants rights against independent replication, not just copying. Kewanee Oil Co. v. Bicron Corp., 416 U.S. 470, 478 (1974).
5 BESSEN & MEURER, supra note 2, at 9 (asserting that unclear patent boundaries chill incentives to innovate).
6 See Mark A. Lemley & A. Douglas Melamed, Missing the Forest for the Trolls, 113 COLUM. L. REV. 2117, 2125-27 (2013) (evaluating the NPE problem); Robert P. Merges, The Trouble With Trolls: Innovation, Rent-Seeking, and Patent Law Reform, 24 BERKELEY TECH. L.J. 1583, 1614 (2009) (identifying rent-seeking and its adverse effects on innovation as the core problem with patent trolls).
7 See Robert G. Bone, Hunting Goodwill: A History of the Concept of Goodwill in Trademark Law, 86 B.U. L. Rev. 547, 561-65 (2006).

1248

BOSTON UNIVERSITY LAW REVIEW

[Vol. 96:1245

symbols of goodwill and devices to communicate information to the market.8 The primary goal is not to incentivize new creations, as it is for patent and copyright, but rather to protect the quality of market information, reduce consumer search costs, and safeguard the reputation of sellers.9 These distinctions have important implications for notice and the costs of notice failure, as elaborated more fully in Parts I and II below.
This Article examines notice failure in trademark law, identifies its sources, evaluates its costs, and explores ways to address it. The discussion is divided into three parts. Part I reviews some aspects of notice in general and then identifies the most serious notice problems in trademark law. Because trademark rights are based on public use, notice problems differ from those in patent and copyright. The principal concern has to do with ascertaining the scope of rights, and the main source of the problem is the vague likelihood-ofconfusion test for liability.
Part II analyzes the costs of vague notice. The most serious cost is the chilling effect on expressive, competitive, and other socially valuable uses. The need to conduct a likelihood-of-confusion analysis in each case creates litigation risks and costs that can discourage perfectly lawful and socially beneficial uses of marks. Still, as Part II argues, it is not clear how serious this problem actually is. Its magnitude depends on the type of use and the type of user.
Part III discusses several possible solutions to notice failure on the assumption that the problems are serious enough to warrant a regulatory response. One possibility is to rely on defenses in the form of categorical rules or balancing tests.10 Categorical rules that exclude liability without regard to confusion risks work reasonably well for core expressive use and perhaps for some kinds of trade dress use. But they work poorly for other types of socially valuable use. In these latter cases, categorical exclusions can be difficult to implement and not easy to square with trademark policy. Moreover, balancing tests, while pervasive in trademark law, do little to solve notice problems. Some commentators have proposed using sanctions and fee-shifting to

8 See id. at 549. Some recent trends in trademark law come close to protecting marks as property, but these are controversial and they expand liability far beyond the relatively settled core. See id. at 621.
9 See, e.g., Dastar Corp. v. Twentieth Century Fox Film Corp., 539 U.S. 23, 34 (2003); Qualitex Co. v. Jacobson Prods. Co., 514 U.S. 159, 163-64 (1995). For a more extensive overview of trademark policies, see Robert G. Bone, Enforcement Costs and Trademark Puzzles, 90 VA. L. REV. 2099, 2105-14 (2004).
10 See Graeme B. Dinwoodie, Developing Defenses in Trademark Law, 13 LEWIS & CLARK L. REV. 99, 99 (2009) (“Some defenses will operate as mechanisms by which to balance competing policy concerns on a case-by-case basis . . . .”); William McGeveran & Mark P. McKenna, Confusion Isn’t Everything, 89 NOTRE DAME L. REV. 253, 301-06 (2013) (proposing categorical exclusions for some favored uses).

2016]

NOTICE FAILURE AND DEFENSES

1249

discourage illegitimate claims and reduce chilling effects.11 But sanctions are not likely to work well and fee-shifting can produce strategic costs of its own.
A much more promising approach targets the liability standard directly. Part III outlines a proposal along these lines. While core expressive use and possibly some trade dress use would be subject to categorical rules, other socially valuable uses would be handled differently. Trademark owners in those cases would be required to plead (and prove) sufficient facts to support the existence of actual source or sponsorship confusion (not just a likelihood) and a significant risk of serious trademark-related harm, and judges would review preliminary injunction motions strictly in favored use cases. This proposal should go far toward reducing chilling effects while still allowing trademark relief in those exceptional cases involving serious confusion-related harm.

I. THE NATURE AND EXTENT OF NOTICE FAILURE IN TRADEMARK LAW
A. Notice in General
Notice is important in the law even when no specific type of notice is legally mandated.12 Notice is about predictability—the ability of persons to predict the legal consequences of their actions in advance. When we refer to B as having notice of X, we mean that B is aware of X or can acquire knowledge of X with reasonable effort under the circumstances. This is the sense of notice that underlies discussions of notice failure in IP law. For example, when commentators complain that patent law does not furnish clear enough notice of the metes and bounds of a patented invention, they mean that the public cannot acquire the necessary information with reasonable effort despite the Patent Act’s requirements.13
Four points follow from this observation, all of which are important to bear in mind when considering the analysis in Parts II and III below. First, notice is probabilistic and predictive. For example, a patent claim provides notice by helping a prospective inventor estimate the probability that her contemplated invention will infringe the claim. So too for trademark law. Someone wishing to parody a mark but concerned about possible infringement will look to the statute, case law, and other sources of notice in order to estimate the

11 See, e.g., William McGeveran, Rethinking Trademark Fair Use, 94 IOWA L. REV. 49, 122 (2008) (suggesting fee shifting); cf. Irina D. Manta, Bearing Down on Trademark Bullies, 22 FORDHAM INTELL. PROP. MEDIA & ENT. L.J. 853, 858-64 (2012) (evaluating the use of sanctions to deter trademark bullying).
12 Sometimes the law requires notice and also specifies its form and content. For example, before the United States joined the Berne Convention in 1989, the federal copyright statute required authors to give formal notice of copyright ownership. 2 MELVILLE B. NIMMER & DAVID NIMMER, NIMMER ON COPYRIGHT § 7.02[C] (2015).
13 See, e.g., Menell & Meurer, supra note 2, at 6.

1250

BOSTON UNIVERSITY LAW REVIEW

[Vol. 96:1245

probability her parody will infringe and the consequences if it does. In both cases, notice provides the basis for predicting expected liability from undertaking the use (i.e., the amount the user expects to lose if a court imposes liability discounted by the probability that liability will be imposed).14
Second, it follows from this first point that notice is a matter of degree. Its quality depends on the accuracy and the certainty of the expected liability estimate. To see this point more clearly, suppose that B, concerned about possible liability for using a mark, collects information about past infringement suits in similar cases. Suppose that some of those previous suits ended in large judgments for the plaintiff, some ended in small judgments, and some ended in judgments for the defendant.15 B, if rational, will take an average over all these cases to estimate her expected liability from using the mark. In other words, the mean of the distribution (with a judgment for the defendant counting as zero) is the same as a rational B’s estimate of expected liability.16
Now suppose that B’s research uncovers only cases ending in a trial judgment and misses weak cases that were dismissed early. Under these circumstances, the mean of the sample distribution, and thus B’s estimate of expected liability, will be too high relative to the true mean over all relevant cases, including those dismissed early. As a result, B might decide not to use the mark even though she would have used it had she had a more accurate estimate. Thus, the quality of notice is better when the estimate is closer to the true mean.
Next suppose that for some reason B finds only cases lying at the extremes—those with large judgments for the plaintiff and those with judgments for the defendant—and misses all the cases in between. The mean of this sample might be very close to the true mean—and thus a reasonably accurate estimate of expected liability—despite the missing cases. But the variance would be high because the sample consists of only extreme cases.17

14 For an explanation of expected liability and the more general concept of expected value, see ROBERT G. BONE, THE ECONOMICS OF CIVIL PROCEDURE 18-29 (2003).
15 A large or small judgment can be measured in terms of the size of monetary awards or the burdens imposed by injunctions of varying scope.
16 To illustrate, consider the following simple numerical example. Suppose that all the cases fall into one of three categories: 20% have large judgments in the amount of $1 million per case; 60% have small judgments in the amount of $50,000 per case; and 20% have judgments for the defendant. The mean of this distribution is: 0.2 × 1,000,000 + 0.6 × 50,000 + 0.2 × 0 = $230,000. Thus, a rational B will assume an expected liability of $230,000.
17 To illustrate with a simple example, suppose that the true distribution of cases is as follows: 20% have judgments of $1,000,000 per case; 30% have judgments of $600,000 per case; 30% have judgments of $400,000 per case; and 20% have judgments for the defendant. The true mean of this distribution is: 0.2 × 1,000,000 + 0.3 × 600,000 + 0.3 × 400,000 + 0.2 × 0 = $500,000. Now suppose that for some reason B finds only the $1,000,000 cases and the cases with judgments for the defendant. B will calculate the mean

2016]

NOTICE FAILURE AND DEFENSES

1251

This variance makes B’s estimate of expected liability uncertain, and if she is risk-averse, she might not make the parody at all even if the expected liability estimate, by itself, would not deter her.18
This observation leads to a third point. Evaluating the adequacy of notice is a normative judgment and depends on the choice of normative metric. In particular, a utilitarian metric, such as economic efficiency, can produce a different result than a rights metric. From an efficiency standpoint, one is justified in concluding that there is a failure of notice only when the marginal social benefits of providing better notice exceed the marginal costs of improving the notice scheme.19 By contrast, a rights metric deems notice deficient if the failure to give better notice impairs rights even if improving the notice is not cost-justified on efficiency grounds.20 This point will be particularly important to the discussion of categorical defenses in Part III.21
Fourth, it is important to bear in mind that notice can work quite well without being certain (i.e., without having a small variance). To illustrate, consider a case in which a rational and risk-neutral innovator is uncertain about the scope of a potentially conflicting patent claim. She knows that there is some chance the claim will be construed very narrowly, some chance that it will be construed very broadly, and some chance that it will be construed in between. Under these circumstances, she will estimate the expected cost of future infringement based on the mean of this distribution. But—and this is the important point—the mean of the distribution should closely reflect the proper scope of the patent claim if courts on average construe patent claims correctly. As a result, our innovator will invest optimally from a social point of view despite her uncertainty. Hence the main point: notice can be very good even if it only communicates information about the average case.
Of course, this assumes risk-neutrality and perfect rationality. As noted above, if our innovator is risk-averse, she will be affected by the variance of the distribution as well as its mean, and should invest less as the variance increases, all other things equal. Moreover, innovators with bounded

of this distorted sample as follows: 0.5 × 1,000,000 + 0.5 × 0 = $500,000. The sample mean is exactly the same as the true mean of the underlying population and thus yields an accurate expected liability estimate despite the distorted sampling. However, the variance is large— larger than the variance of the actual population distribution.
18 For the effects of variance on risk-averse actors, see A. MITCHELL POLINSKY, AN INTRODUCTION TO LAW AND ECONOMICS 53-58 (2d ed. 1989).
19 See BESSEN & MEURER, supra note 2, at 219-23 (using an efficiency metric to analyze notice failure).
20 For more on the difference between utilitarianism (of which efficiency analysis is an example) and rights analysis, see RONALD DWORKIN, TAKING RIGHTS SERIOUSLY 193 (1977), and JOHN RAWLS, A THEORY OF JUSTICE 3 (1971).
21 See infra notes 139-56 and accompanying text.

1252

BOSTON UNIVERSITY LAW REVIEW

[Vol. 96:1245

rationality are likely to arrive at different probability estimates.22 Costs matter, too. The lower the cost of acquiring information necessary to make a prediction, the more information will be obtained and the better the prediction should be.
In sum, notice must be evaluated along three dimensions: accuracy, certainty, and cost. Moreover, any evaluation depends on the normative metric. Efficiency balances the social benefits of better notice in terms of accuracy, certainty, and cost against the social costs of improving the notice scheme. By contrast, a rights metric gives priority to certain interests with special social value and focuses on how notice affects those interests.
B. Notice in Trademark Law
In analyzing notice failure, it is useful to distinguish three different aspects of IP in which notice can be a problem: ownership, subject matter boundaries, and the scope of rights and remedies.23 While notice problems in copyright and patent manifest in all three areas, notice problems in trademark law focus mostly on the scope of rights.
First, consider ownership. The fact that copyright does not require publication or registration complicates the task of identifying owners of conflicting copyrights.24 As a result, a copyright owner can keep her copyrighted work private and still assert rights against another creator. Registration and publication are both required for patents, but the costs of searching the registry and identifying conflicting claims can be prohibitive.25 The results are well known: orphan works chill creativity in copyright, and NPEs create holdup problems in patent that can discourage invention.26

22 For a discussion of bounded rationality, see Russell Korobkin & Thomas S. Ulen, Law and Behavioral Science: Removing the Rationality Assumption from Law and Economics, 88 CALIF. L. REV. 1051 (2000).
23 For a similar classification of notice problems, see Menell & Meurer, supra note 2, at 7-13 (dividing notice problems into two “buckets”: (i) notice of boundaries and ownership, and (ii) notice of the scope of the property rights).
24 The 1976 Copyright Act protects unpublished as well as published works and requires registration only as a condition to suit but not as a requirement for rights. See 17 U.S.C. §§ 102, 411 (2012).
25 See BESSEN & MEURER, supra note 2, at 55 (explaining that patent searches cost “about a hundred times more” than land title searches). This problem is a function of both the huge number of patents and the difficulty of being certain that any particular patent covers the contemplated invention. See id. at 54-71.
26 See U.S. COPYRIGHT OFFICE, REPORT ON ORPHAN WORKS 1 (2006) (addressing “the issue of ‘orphan works,’ a term used to describe the situation where the owner of a copyrighted work cannot be identified and located by someone who wishes to make use of the work in a manner that requires permission of the copyright owner”); see also Lemley & Melamed, supra note 6, at 2125-26 (describing the different types of NPEs).

2016]

NOTICE FAILURE AND DEFENSES

1253

Trademark is different. Trademark rights in the United States depend critically on public use.27 To be protected, a mark must be used in trade, and the use must be sufficiently public to furnish notice to others that the senior user claims rights in the mark.28 Moreover, if for some reason a junior user does not learn about a prior use, its lack of notice does not prevent it from acquiring trademark rights in its own geographic area.29 It is true that this result no longer holds when the senior user’s mark is registered, but then registration provides its own notice.30 Finally, because of the abandonment doctrine, a mark owner must continue to use the mark publicly in order to retain rights; otherwise, the mark enters the public domain free for anyone to use.31 This makes it hard for a trademark owner to conceal its mark and still retain trademark rights in it.32
Next consider subject matter boundaries. In patent law, the patent owner defines the subject matter of the patent through written patent claims. This means that subject matter boundaries depend on how the claim is construed, and claim construction is a notoriously messy business plagued by pervasive uncertainty.33 In copyright law, the boundaries of a protected work of

27 2 J. THOMAS MCCARTHY, MCCARTHY ON TRADEMARKS AND UNFAIR COMPETITION §§ 16:1, 16:6-:7 (4th ed. 2009). The secondary meaning requirement can produce some uncertainty about the existence of rights in descriptive marks. Id. § 11:15. This is not likely to be a problem, however, for the favored uses that I focus on in this Article because those uses tend to target well-known marks that clearly have secondary meaning.
28 See Couture v. Playdom, Inc., 778 F.3d 1379, 1381 (Fed. Cir. 2015) (holding that to register a service mark under the Lanham Act, the mark must not only be used publicly in an “open and notorious” way but must also be used in connection with the actual provision, not just the offering, of the service); Zazú Designs v. L’Oréal, S.A., 979 F.2d 499, 503 (7th Cir. 1992) (“Only active use . . . notifies other firms that the mark is [associated with particular goods].”); see also MCCARTHY, supra note 27, §§ 16:1-:14 (discussing the types of use necessary to acquire trademark rights and establish priority).
29 This is the result of the common law doctrine that a good faith junior user in a remote area does not infringe the senior user’s mark and can acquire rights of its own in that area. See 5 MCCARTHY, supra note 27, § 26:2 (explaining the Tea Rose-Rectanus doctrine). For this doctrine to apply, the junior user must be located outside the geographic area in which the senior user’s mark already enjoys a substantial reputation. Id. But this requirement does not undercut the notice function when the senior user’s reputation itself provides notice.
30 Registration notice, however, is only as good as the Patent and Trademark Office’s database.
31 See 3 MCCARTHY, supra note 27, § 17:9.
32 There is, however, one possible scenario where holdup might still occur. Suppose a senior user limits its use to a small area, shuns the Internet, and does not register the mark. A junior user could adopt the same mark, unaware of the senior user’s prior use, and then face a conflict when it expands into the senior user’s market. I have no idea whether or how often this scenario occurs. The Internet should make it unlikely, however.
33 See BESSEN & MEURER, supra note 2, at 54-62 (“[P]atent applicants sometimes game

1254

BOSTON UNIVERSITY LAW REVIEW

[Vol. 96:1245

authorship are defined primarily in the course of infringement litigation and by the application of relatively open-ended rules, such as the substantial similarity test, the idea-expression dichotomy, and the merger doctrine.34
Trademark is similar to copyright in this regard. Trademark law does not protect marks as such; it protects the information content of marks, that is, the mark’s goodwill. Moreover, a trademark owner suing for infringement does not first define its goodwill and then show that the defendant injured or appropriated that goodwill.35 Instead, the owner proves likelihood of consumer confusion (“LOC”), and this determination defines the boundaries of protectable goodwill.36 Thus, subject matter boundaries in trademark law are largely defined by the scope of trademark rights, just as they are in copyright.37

the system by drafting ambiguous claims that can be read narrowly during examination . . . and broadly during litigation . . . .”). The ex ante approach used in patent law is known as “peripheral claiming,” and the ex post approach used in copyright is known as “central claiming.” Dan L. Burk & Mark A. Lemley, Fence Posts or Sign Posts? Rethinking Patent Claim Construction, 157 U. PA. L. REV. 1743, 1744-46 (2009) (explaining the concepts of peripheral and central claiming); Jeanne C. Fromer, Claiming Intellectual Property, 76 U. CHI. L. REV. 719, 721 (2009) (discussing peripheral claiming’s role in patent law).
34 A copyright-protected work must also be fixed in a tangible medium of expression, which helps reduce uncertainty to some extent. See 17 U.S.C. § 102 (2012). But the fixation requirement has limited utility in determining the boundaries of protection for a work because a work’s boundaries extend beyond the literal content of its fixed form.
35 Goodwill, of course, is not something definite with boundaries capable of clear definition and demarcation. In fact, the concept of goodwill has always eluded precise definition. See Bone, supra note 7, at 569-72 (reviewing the different attempts to define the term “goodwill”).
36 Id. at 558. I focus on confusion-based theories, but it is worth mentioning that trademark law also imposes liability for dilution, which does not require confusion. See, e.g., Moseley v. V Secret Catalogue, Inc., 537 U.S. 418, 432-33 (2003) (clarifying the basis of dilution claims in trademark law). Dilution occurs when the defendant’s use impairs the mark’s selling power either by tarnishing it with unsavory associations or by blurring its distinctiveness with multiple uses on different products. See id. at 433. I concentrate on confusion, however, because it is the core liability theory in trademark law. Dilution has limited scope and is highly controversial. Moreover, dilution is also subject to multi-factor tests that create similar notice problems. See, e.g., 15 U.S.C. §§ 1125(c)(2)(A)-(B) (2012) (listing factors for determining whether a mark is famous and whether the defendant’s use dilutes by blurring). For the history of the dilution theory and its controversial status in trademark law, see Bone, supra note 7, at 604-06, and Robert G. Bone, Schechter’s Ideas in Historical Context and Dilution’s Rocky Road, 24 SANTA CLARA COMPUTER & HIGH TECH. L.J. 469 (2008).
37 This is particularly true for product market boundaries. A trademark owner can protect its mark against uses on noncompeting products only insofar as those uses are likely to cause confusion (or dilution). Still, likelihood of confusion is also relevant to the geographic reach of trademark rights, especially for owners of unregistered marks and for the scope of
NoticeTrademark LawRightsLiabilityMark