Reverse Confusion and the Justification of Trademark Protection

Category: Print Issues

Reverse Confusion and the Justification of Trademark Protection

Jeremy N. Sheff

Volume 30

, Issue 1

30 GEO. MASON L. REV. 123 (2022)

Introduction

In late 2021, social media giant Facebook was facing an increasing array of legal and regulatory pressures over its business practices.¹ In the midst of this scrutiny (and perhaps in hopes of changing the subject), Facebook CEO Mark Zuckerberg announced with alacrity that he was rebranding, changing his company’s name to “Meta” (a reference to the “metaverse” of augmented and virtual reality that the company sees as its most promising area of future growth).² But there was a problem: there was already a company in the computer technology field calling itself Meta (or, more specifically, MetaCompany).³ In fact, as of the date of Zuckerberg’s announcement, there were twenty live federal trademark registrations for the mark META (and another for MÈTA, and yet another for METAS), covering goods and services such as software and computer services (Registration No. 6,292,538), video game controllers and headsets (Registration No. 6,253,957), and even augmented and virtual reality eyewear (Registration Nos. 5,201,102 & 4,870,701).⁴ Moreover, before Zuckerberg’s announcement Facebook’s lawyers had allegedly approached MetaCompany—without revealing their client’s identity—to try to negotiate the purchase of MetaCompany’s trademark rights, and had been rebuffed, but went ahead with the Facebook rebranding anyway.⁵ Following Zuckerberg’s announcement, MetaCompany immediately threatened suit.⁶

Billions of people know Facebook and use the company’s services, and at least an appreciable fraction of them will hereafter associate the term “Meta” with that company (particularly given the substantial press coverage the name change received).⁷ But very few people will have heard of MetaCompany prior to its dispute with Facebook, and many who do learn about MetaCompany may never know anything more about it other than its involvement in this dispute. Indeed, to the extent they do not know about this dispute, many who learn about MetaCompany after Facebook’s name change may mistakenly believe that MetaCompany is a licensee or affiliate of Facebook. Yet if MetaCompany’s trademark registration is not invalidated and establishes priority over any registrations acquired by Facebook, trademark law could well afford the small, little-known company the power to enjoin any use of the disputed trademark by the global behemoth (at least in the United States), causing all of Facebook’s investments in rebranding to go to waste. The source of that power is the doctrine of reverse confusion.

Trademark law imposes infringement liability when a defendant uses a mark on or in connection with the defendant’s goods or services in a manner that is “likely to cause confusion, or to cause mistake, or to deceive” as to either the “affiliation, connection, or association” between the plaintiff and the defendant or “as to the origin, sponsorship, or approval” of the defendant’s goods or services.⁸ The most typical form of trademark infringement involves “passing off” (or “point-of-sale confusion”), which arises when the defendant puts the plaintiff’s previously-established trademark on the defendant’s goods in order to deceive consumers into thinking the goods actually originated with the plaintiff.⁹ A reverse confusion claim, in contrast, arises when a relatively little-known senior (i.e., earlier) user of a trademark faces a much more well-known junior (i.e., later) user of the same mark. In such cases, consumers encountering the little-known senior user’s mark on the senior user’s goods are likely to be confused into thinking that the goods are somehow connected to the better-known junior user. As the Seventh Circuit Court of Appeals explains:

Reverse confusion occurs when a large junior user saturates the market with a trademark similar or identical to that of a smaller, senior user. In such a case, the junior user does not seek to profit from the good will associated with the senior user’s mark. Nonetheless, the senior user is injured because “[t]he public comes to assume that the senior user’s products are really the junior user’s or that the former has become somehow connected to the latter. The result is that the senior user loses the value of the trademark—its product identity, corporate identity, control over its goodwill and reputation, and ability to move into new markets.”¹⁰

Disputes like that between Facebook—one of the planet’s largest and best-known companies—and MetaCompany—a company relatively few people have ever heard of—raise interesting questions about the theoretical justifications for trademark doctrines. In particular, they challenge conventional “search costs” justification that depends on the effect of trademark uses on the costs of obtaining and transferring information in the marketplace—justifications particularly associated with the Seventh Circuit, and its most prominent member, Judge Richard Posner. Judge Posner’s theory runs as follows:

The fundamental purpose of a trademark is to reduce consumer search costs by providing a concise and unequivocal identifier of the particular source of particular goods. The consumer who knows at a glance whose brand he is being asked to buy knows whom to hold responsible if the brand disappoints and whose product to buy in the future if the brand pleases. This in turn gives producers an incentive to maintain high and uniform quality, since otherwise the investment in their trademark may be lost as customers turn away in disappointment from the brand. A successful brand, however, creates an incentive in unsuccessful competitors to pass off their inferior brand as the successful brand by adopting a confusingly similar trademark, in effect appropriating the goodwill created by the producer of the successful brand. The traditional and still central concern of trademark law is to provide remedies against this practice.¹¹

Proponents of this search-costs theory generally claim that trademarks provide consumers with information not only about the source of a marked product, but about the qualities of the product itself.¹²

In previous works, I have explained how certain trademark doctrines do not seem to be consistent with the search-costs account of trademark law’s purposes.¹³ From traditional “passing off” or point-of-sale confusion,¹⁴ to post-sale confusion,¹⁵ to dilution,¹⁶ numerous forms of trademark liability appear to be pursuing aims having little to do with lowering consumer search costs or providing incentives to produce goods of high and consistent quality—and indeed sometimes in opposition to such economic concerns. Others have pointed out how initial-interest confusion and sponsorship, affiliation, and approval confusion similarly lack strong justification from conventional economic theory—though they typically argue this is a failure of doctrine, rather than of theory.¹⁷ But reverse confusion—being a relatively rare form of infringement—has not attracted similar theoretical scrutiny. In my own prior works, I have examined two possible alternative theoretical accounts for various forms of trademark liability that—I argue—are descriptively more successful than the standard account. One—misappropriation—is an accepted theoretical basis for trademark liability outside the United States, but seldom invoked (at least openly) within the United States, and it generates substantial criticism on grounds of both incoherence and normative undesirability.¹⁸ The other—my own theory of trademark as promise—offers a more flexible and more coherent account for the doctrine, though it is an account that remains in development.¹⁹

The key claim of trademark as promise is that trademark doctrine appears (as a matter of positive law) to have developed in such a way as to help producers make credible assurances (promises) to consumers, and to impose liability against conduct that interferes with the making, performance, or reliability of such promises.²⁰ Trademark as promise justifies these observed features of doctrine under a contractualist theoretical framework founded on the equal autonomy of all moral agents, and by implication reserves to consumers the right to choose whether to accept a substitute for performance of such a promise. I believe that this account of trademark law is at least as normatively attractive as the currently dominant but descriptively inadequate search-costs account, and perhaps more so.

Reverse confusion offers an opportunity to test and refine the theory of trademark as promise, and this Article represents my first sustained effort to apply that theory to relations among producers (as opposed to relations between producers and consumers). Reverse confusion is a particularly congenial subject for such theoretical development precisely because, as I will argue, it is not well explained by the criteria of economic analysis on which search-costs theory depends (insofar as finding against senior users in reverse confusion cases would likely increase aggregate welfare, at least as a matter of first-order effects), nor by a misappropriation-based account (because the junior user in reverse confusion cases is concededly not extracting any value from the senior user’s goodwill). But, as I will further argue, the harm in reverse confusion cases can be understood as an interference with the ability of a senior user to credibly make promises to consumers that the senior user would otherwise be in a position to make and perform. This is an understanding that is consistent with our moral intuitions, but difficult to express in economic terms. Trademark as promise gives us a vocabulary and framework to express those intuitions in a systematic way, and in so doing helps us make sense of the case law in this area (and potentially others). It is, in short, a better theory than either the search-costs theory or the misappropriation theory, certainly as a descriptive matter and at least plausibly as a normative matter. It also offers an example of the extent to which decades of theorizing about private law in economic terms can lead well-meaning analysts to misunderstand and mischaracterize relevant doctrines and to blind themselves to private law’s broader normative commitments and potential.

The remainder of this Article proceeds as follows. Part I reviews the law of reverse confusion and identifies key themes in the cases: the relative importance of the junior user’s state of mind and of the relative economic power of the parties, and the relative unimportance of the value to consumers of the information transferred by the mark (in either the junior or the senior users’ hands). Part II attempts to model reverse confusion in economic terms, both in terms of consumer search costs and as a form of non-market transfer—as theft—but finds that these standard law-and-economics accounts are a poor fit for reverse confusion cases. Part III then turns to my developing contractualist theory of trademark as promise, and argues that reverse confusion doctrine fits well within the theory. The Conclusion sketches implications for theories of private law and analysis of market institutions more generally.

I. Reverse Confusion: The Case Law

A. Early Developments

The earliest reported federal decision to explicitly consider a reverse confusion theory of trademark infringement rejected the theory as irrational. In Westward Coach Mfg. Co. v. Ford Motor Co.,²¹ the Seventh Circuit reviewed an award of summary judgment to Ford in an action that accused Ford of infringing Westward Coach’s mark MUSTANG for campers and trailers—a mark for which Westward Coach obtained an Indiana state registration in 1962 and a federal registration in 1967, having used the mark since 1960.²² Ford’s famous MUSTANG sports car hit the market in 1964, supported by an eight-figure advertising campaign. By the time of the suit Ford had sold over half a million MUSTANG cars, while Westward Coach had sold just over 1,200 campers and trailers—and only about two-thirds of those under the MUSTANG mark.²³

Affirming the district court’s award of summary judgment to Ford, the Seventh Circuit concluded that campers and trailers were too different from sports cars to generate confusion simply through the use of an identical mark, and furthermore deemed the plaintiff’s theory of reverse confusion unsupported by law or reason:

Ford contends that under Indiana law and under federal law, if it is applicable, the second use of a trademark is not actionable if it merely creates a likelihood of confusion concerning the source of the first user’s product. . . . The Indiana appellate courts have not considered whether a second use creating the likelihood of confusion about the source of the first user’s products is actionable. The tests of unfair competition and trademark infringement formulated by the Indiana courts . . . do not embrace the latter type of confusion. . . . Appellants apparently put forth a suggested doctrine of “reverse confusion” concerning which we find no rational basis for support.²⁴

Thus, in the view of the court that first used the term “reverse confusion” in a reported opinion, the very concept was irrelevant and nonsensical: actionable confusion required confusion as to the source of the accused junior user’sproduct.²⁵

The first judicial acceptance of the theory of reverse confusion²⁶ arose nearly a decade later. In Big O Tire Dealers, Inc. v. Goodyear Tire & Rubber Co.,²⁷ the Tenth Circuit became the first federal court of appeals to endorse the theory, albeit as a matter of state law. Big O, a relatively small regional tire distributor, had begun using the mark BIG FOOT on some of its tire models beginning in February 1974, but did not seek to register the mark with the United States Patent and Trademark Office (“USPTO”). Goodyear—the largest tire manufacturer in the world, and the owner of a federal trademark registration for the mark BIGFOOT for snowmobile replacement tracks—decided in July of 1974 to extend its BIGFOOT mark to its primary market: radial automobile tires.²⁸ Despite conducting a clearance search that does not appear to have been negligent, Goodyear did not discover Big O’s products before making this decision.²⁹ Two months later, on the eve of launching a multimillion-dollar advertising campaign in September of 1974 (for which the money had already largely been spent), Goodyear first became aware of the existence of Big O’s BIG FOOT tires, and entered into negotiations with Big O over use of the mark.³⁰ Goodyear sought a covenant not to sue and was willing to pay Big O for it, but Big O sought a commitment from Goodyear to discontinue use of the BIGFOOT mark as soon as possible, and the parties failed to reach agreement.³¹ Goodyear went ahead with its advertising campaign, and Big O sued for false designation of origin under section 43(a) of the Lanham Act, for trademark infringement under Colorado common law, and for “trademark disparagement.”³² The jury found for Goodyear on the Lanham Act claim but for Big O on the other claims,³³ awarding compensatory damages of $2.8 million and punitive or exemplary damages of $16.8 million.³⁴ The district court went on to award an injunction against any use of the contested mark by Goodyear.³⁵

In affirming the district court judgment as to all but the size of the damages award, the Tenth Circuit noted the Colorado Supreme Court’s generally expansive view of trademark rights, and concluded that “the Colorado courts, if given the opportunity, would extend its [sic] common law trademark infringement actions to include reverse confusion situations.”³⁶ Moreover, it approved an argument for this result, originally made by the district court, grounded in distrust of economic power:

The logical consequence of accepting Goodyear’s position would be the immunization from unfair competition liability of a company with a well established trade name and with the economic power to advertise extensively for a product name taken from a competitor. If the law is to limit recovery to passing off, anyone with adequate size and resources can adopt any trademark and develop a new meaning for that trademark as identification of the second user’s products. The activities of Goodyear in this case are unquestionably unfair competition through an improper use of a trademark and that must be actionable.³⁷

Indeed, the district court had gone even further in condemning Goodyear as a “wanton and reckless” bully:

What Goodyear did was to determine that the plaintiff was an insignificant competitor and that Goodyear’s investment of time, money and effort was too great to give up the use of the BIGFOOT material even though it would clearly damage and perhaps destroy the plaintiff’s use of its trademark. Goodyear then proceeded with an intentional and deliberate infringement of the plaintiff’s trademark. In short, Goodyear elected to ignore completely the property rights of Big O. It is difficult to characterize such conduct in any manner more favorable than as a wanton and reckless disregard of the plaintiff’s rights.³⁸

From this seed grew the doctrine of reverse confusion. The Fifth Circuit was next to adopt the theory, in an unfair competition case alleging plagiarism of an alternative history film imagining the trial of Lee Harvey Oswald—including copying of the film’s title, claimed as an unregistered trademark.³⁹ Again the federal court embraced reverse confusion by inferring what state courts (this time those of Texas) would hold under broad theories of unfair competition.⁴⁰ In 1981, leading trademark law commentator J. Thomas McCarthy highlighted the “New Concept” of reverse confusion in an article on important trends of the previous decade,⁴¹ reviewing the Big O case in detail and connecting it to a dictum in Justice Holmes’s concurrence in the (in)famous misappropriation case of International News Service v. Associated Press.⁴² Over the next decade, the Court of Customs and Patent Appeals⁴³ and the Second,⁴⁴ Sixth,⁴⁵ Seventh,⁴⁶ and Ninth⁴⁷ Circuits would expressly or impliedly recognize the theory’s validity (if not its direct applicability to the cases before them).⁴⁸ But cases in which a plaintiff prevailed on a reverse confusion theory remained rare, and some district courts expressed doubt that federal law—and particularly section 43(a) of the Lanham Act—encompassed reverse confusion.⁴⁹

Judge Sweet of the Southern District of New York was the first to hold that the Lanham Act does in fact encompass reverse confusion, in Elizabeth Taylor Cosmetics Co. v. Annick Goutal, S.A.R.L.⁵⁰673 F. Supp. 1238, 1248–49 (S.D.N.Y. 1987). Like Big O, this case involved a junior user with deep pockets (a joint venture of Elizabeth Taylor and Cheesebrough-Ponds) who discovered on the eve of launching a multimillion-dollar brand campaign that there was a small but established senior user of the mark they had selected (PASSION for perfumes), and went ahead with the launch anyway over the senior user’s protest.⁵¹ Judge Sweet’s opinion enjoining the junior user echoed the distrust of economic power from the district court’s opinion in Big O: “The Lanham Act was designed to protect all legitimate businesses from unfair competition. It was not merely intended to preserve the kind of good will that is purchased for millions of dollars.”⁵²

Soon thereafter, the Second Circuit became the first federal court of appeals to affirmatively hold that reverse confusion is actionable under the Lanham Act in Banff, Ltd. v. Federated Department Stores, Inc.⁵³ That case, however, did not involve a much larger company bullying a senior trademark user out of existence; rather, it involved a dispute between a sizeable garment manufacturer and the Bloomingdale’s department store over the latter’s introduction of a house mark identical to the manufacturer’s established trademark.⁵⁴ The court noted that the dispute could be argued under both a forward-confusion and a reverse-confusion theory, but took the opportunity “to hold explicitly that a likelihood of reverse confusion warrants an injunction under the Lanham Act,” because otherwise “a larger company could with impunity infringe the senior mark of a smaller one.”⁵⁵ Around the same time the Trademark Trial and Appeal Board began citing reverse confusion as a basis for refusal to register (as where luxury brand Tiffany sought to enter the cosmetics market and was successfully opposed by the owner of a registration for the mark TIFFAN “E” for skin cream),⁵⁶ and for cancellation of existing registrations.⁵⁷

B. Representative Examples of Reverse Confusion

Despite its broad acceptance, reverse confusion is rare. It is a theory premised on an unsuccessful or only moderately successful business blaming its failures (or threatened failures) on a more successful company. Perhaps because business failures are often overdetermined, because vigorous competition is presumed to be a virtue, and because wealthy companies may attract vexatious litigants in search of a payday, the theory can often draw skepticism from federal judges. As the Third Circuit has explained:

The chief danger inherent in recognizing reverse confusion claims is that innovative junior users, who have invested heavily in promoting a particular mark, will suddenly find their use of the mark blocked by plaintiffs who have not invested in, or promoted, their own marks. Further, an overly-vigorous use of the doctrine of reverse confusion could potentially inhibit larger companies with established marks from expanding their product lines . . . .⁵⁸

Or, as Judge Walker of the Northern District of Florida more colorfully explained, reverse confusion liability “should not serve as a bazooka in the hands of a squirrel, used to extract from a more fearsome animal a bounty which the squirrel would never be able to gather by his own labors—at least not when the larger animal is mostly without sin.”⁵⁹

Courts thus often reject reverse confusion claims before trial on the ground that they are not supported by sufficient evidence of likely confusion. The particular failures of evidence vary, and often overlap. Even minor dissimilarities between the parties’ marks can defeat a reverse confusion claim.⁶⁰ Such dissimilarities may include the juxtaposition of the junior user’s house mark alongside the contested mark, though some courts have opined that such a juxtaposition actually increases the likelihood of reverse confusion (by cementing the association between the contested mark and the better-known junior user).⁶¹ Moderate divergences between the parties’ products or target markets have also been held sufficient to reject reverse confusion claims.⁶² And courts are unsympathetic to reverse confusion plaintiffs whose commercial struggles can be attributed to causes other than the better-known defendant’s adoption of a similar mark.⁶³

Finally, the evidence that the defendant was unaware of the plaintiff’s mark when entering the market can help defeat a reverse confusion claim even where the parties’ marks are identical—particularly if any of the other deficiencies identified above are present. For example, in Pump, Inc. v. Collins Management, Inc.,⁶⁴ a little-known rock band held a federal registration for the stylized word mark “pump,” which they had used as their band name for over two years before the world-famous arena rock group Aerosmith adopted “PUMP” as the title of what would become their multi-platinum album of 1989.⁶⁵ Notwithstanding the nationwide constructive notice established by the plaintiff’s federal registration,⁶⁶ the court found that there was no evidence that the defendants were in fact aware of the plaintiff’s existence when they adopted the disputed mark (explicitly distinguishing Big O on this ground).⁶⁷ Combined with dissimilarities in the parties’ stylization of the word “pump” in their marks and the presence of Aerosmith’s band logo on its album cover, this was deemed a sufficient basis to award summary judgment to the defendants.⁶⁸ Similarly, in Walter v. Mattel, Inc.,⁶⁹ the parties both used the mark PEARL BEACH in conjunction with a pearl-and-seashell logo: plaintiff in connection with her commercial illustration business, and Mattel in connection with a new variety of Barbie doll.⁷⁰ The district court found (and the Ninth Circuit affirmed) that Mattel’s trademark clearance search, which failed to discover plaintiff’s unregistered business name, established its good faith.⁷¹ This, when combined with different stylization of the parties’ marks and dissimilarity between their goods and services, was sufficient to entitle Mattel to judgment as a matter of law.⁷² Finally, in J.T. Colby & Co. v. Apple Inc.,⁷³ both parties used the word mark IBOOKS: plaintiff as a publishing imprint, and defendant as the name of an e-book reader app and marketplace service.⁷⁴ Apple had performed a clearance search that did not discover the plaintiff’s business or its unregistered mark, though it did discover that plaintiff’s bankrupt predecessor had abandoned an application to register the mark.⁷⁵ The district court held (and the Second Circuit affirmed) that this was evidence of Apple’s good faith, and (again citing differences in the parties’ stylization of the word mark and the context of their uses) granted summary judgment in Apple’s favor.⁷⁶

Successful reverse confusion cases, in contrast, follow a similar pattern to Big O and Elizabeth Taylor Cosmetics. That pattern consists of a larger company entering the market under a trademark with knowledge or good reason to know of a smaller company’s established rights in the same or a closely similar trademark in the same or a closely similar area of commerce, thereby foreseeably threatening to damage or destroy the smaller company’s ability to continue operating under its mark. For example, in Dreamwerks Production Group, Inc. v. SKG Studio,⁷⁷ “a small Florida company” that had been engaged in staging Star Trek conventions throughout the Northeast and Midwest since 1984 obtained a federal registration for its trademark in 1992 but “ha[d] never really taken off” by the time the giant Hollywood production company DreamWorks SKG was founded in 1994.⁷⁸ The evidence showed that DreamWorks SKG had “conducted a diligent search and discovered the Dreamwerks mark, yet failed to make accommodations or select a different mark.”⁷⁹ The Ninth Circuit held that given the relatedness of the parties’ goods and services—”movies and sci-fi merchandise are now as complementary as baseball and hot dogs’’⁸⁰—the danger of reverse confusion entitled the smaller company to a jury trial. The court thus reversed an award of summary judgment in favor of the larger defendant.⁸¹

Likewise in A & H Sportswear, Inc. v. Victoria’s Secret Stores, Inc.,⁸² a successful swimsuit manufacturer operating under the registered mark MIRACLESUIT sued “lingerie leviathan” Victoria’s Secret when the latter company entered the swimwear market under its own registered THE MIRACLE BRA trademark.⁸³ Though the defendant was aware of the plaintiff’s mark at the time it entered the swimwear market and was found to have willingly run the risk of an infringement finding, the district court ultimately concluded its actions, while “risky,” were not “wilfully tortious” insofar as they “did not deliberately undertake to cause consumer confusion or destroy Plaintiff’s business . . . or push Plaintiffs out of the swimsuit market.”⁸⁴ Nevertheless, in a tortuous seven-year litigation that traveled back and forth between the district court and the court of appeals three times, the plaintiff ultimately obtained an injunction against use of defendant’s marks on swimwear, but no monetary relief.⁸⁵

Not one but two unrelated small, established companies successfully sued the ubiquitous ridesharing company Uber Technologies when that company began encroaching on their areas of business. In one case, a moderately successful marketing and design services company operating under the unregistered mark UBER—which had declined the defendant’s offer of $120,000 to change its name—survived a motion to dismiss.⁸⁶ In the other, a limousine and event planning company operating under the unregistered mark UBER PROMOTIONS won a preliminary injunction against the advertisement or use of the ridesharing company’s mark in connection with the latter’s new event transportation service in the single county where the plaintiff was already operating.⁸⁷ The second of these opinions included colorful and telling characterizations of the parties’ relative size. Noting that the junior user was not aware of the senior user at the time it adopted its mark, but must have been aware when it entered the plaintiff’s market, the court downplayed any suggestion of culpable intent with the quip: “Elephants don’t look out for gerbils when they plow through the bush.”⁸⁸ (This is the same opinion that analogized reverse confusion doctrine to a “bazooka in the hands of a squirrel.”)⁸⁹

In the recent case of Ironhawk Technologies, Inc. v. Dropbox, Inc.,⁹⁰ a software company serving mainly military clients sued the well known consumer- and business-oriented cloud storage service provider over its adoption of the mark SMARTSYNC, which plaintiff had registered ten years earlier.⁹¹ Sealed evidence apparently suggested that Dropbox was aware of Ironhawk’s mark before it launched its own SmartSync brand, and some of Ironhawk’s potential customers had allegedly inquired as to the existence of a connection between the two companies.⁹² The Ninth Circuit, over a dissent, reversed the district court’s dismissal of the claim at the summary judgment stage and remanded for trial.⁹³

Even relatively large senior users can take advantage of reverse confusion doctrine when facing off against a true multinational behemoth. In Fleet Feet, Inc. v. Nike Inc.,⁹⁴ the owner-franchisor of a nationwide chain of running and fitness stores had held a registration for the mark CHANGE EVERYTHING for several years, and was in the process of registering the mark RUNNING CHANGES EVERYTHING.⁹⁵ Defendant, a world-renowned athletics gear manufacturer, was apparently aware of the plaintiff’s marks but nevertheless launched an eight-figure advertising campaign around the slogan SPORT CHANGES EVERYTHING, targeting several sports including running.⁹⁶ On a motion for a preliminary injunction, the district court noted that the plaintiff’s advertising expenditures over the past five years were only a fraction of the defendant’s planned expenses for the single challenged campaign, that the defendant was almost certainly aware of the plaintiff’s marks, and that the defendant had continued its campaign with only minor modifications after receiving plaintiff’s cease-and-desist letter.⁹⁷ The court concluded that the plaintiff was likely to succeed on its reverse confusion claim and issued a preliminary injunction against any use of the challenged slogan.⁹⁸

Finally, in Attrezzi, LLC v. Maytag Corp.,⁹⁹ plaintiff opened a single-location kitchen products store and had obtained a state registration (but not a federal registration) for the mark ATTREZZI.¹⁰⁰ Between the opening of plaintiff’s store and the issuance of its state registration, internationally known appliance maker Maytag began developing plans to adopt the same mark for a new line of electric kitchen appliances.¹⁰¹ A trademark clearance search discovered plaintiff’s mark and in-house counsel advised Maytag that it presented “a problem,” but Maytag requested “another look” and counsel thereupon reversed their position.¹⁰² Maytag applied for a federal registration; plaintiff opposed the application and sued for infringement.¹⁰³ A jury found willful infringement and the district court granted a permanent injunction against any use of the disputed mark (subject to a twelve-month window to sell off existing inventory), enhanced damages, and attorney’s fees; the First Circuit affirmed.¹⁰⁴

As the reverse confusion case law has matured, several themes have developed in these and other cases that distinguish reverse confusion from other trademark infringement theories and characterize successful reverse confusion claims. The rest of this Part reviews those distinctions.

C. Distinctive Features of Reverse Confusion Cases

1. The Nature of the Injury

The mine run of trademark infringement cases—what we might call “forward confusion” cases—involve an intuitive theory of injury. In the classic point-of-sale source-confusion case the defendant, in passing itself off as the plaintiff, deceives the plaintiff’s would-be customers into purchasing from the defendant instead. The defendant thereby defrauds those customers and deprives the plaintiff of sales it would otherwise have enjoyed. Other, more attenuated theories of infringement such as sponsorship and affiliation confusion, post-sale confusion, and initial-interest confusion involve theories of injury grounded in antipathy toward misappropriation: the idea that it is wrong to gain a benefit from another’s labor and investment—to reap where one has not sown.¹⁰⁵ These misappropriation-based forms of infringement all involve the defendant gaining an advantage from consumers’ associations with the plaintiff’s mark, and thus generate liability even where the plaintiff has not suffered any direct or tangible injury as a result.¹⁰⁶

Reverse confusion is quite different. The stereotypical reverse confusion case has been described as a “David and Goliath” scenario,¹⁰⁷ but in truth it aligns more closely with the story of David and Uriah.¹⁰⁸ Specifically, it resembles the parable of the ewe lamb with which the prophet Nathan condemns David for the rape of Uriah’s wife Bathsheba (the punchline of the Old Testament story turns on a lamentably retrograde analogy between women and livestock).¹⁰⁹ In that parable, a rich and powerful man takes from a poor man his only beloved ewe lamb and slaughters it to host a feast for a guest, rather than selecting from among his own vast flocks to satisfy his wants.¹¹⁰ David “burned with anger” against the rich man in Nathan’s parable, unwittingly condemning himself.¹¹¹

Successful reverse confusion claims follow an analogous pattern to the parable of the ewe lamb. A large, wealthy company wishes to adopt a trademark, finds out that a small but established company already has rights in that mark, and goes ahead with its plans to adopt the mark, knowing that it will be disastrous for the smaller company but unwilling to accept the cost of changing its plans. This pattern, it seems, causes an injury that demands compensation and even punishment—as David “burned with anger” at the prophet Nathan’s parable of the ewe lamb,¹¹² we may find ourselves similarly inflamed against the junior user in reverse confusion cases. But the injury that generates our opprobrium is quite specific, and involves not only the harm to the senior user but the relative position of the junior user.

As the Second Circuit explains, in reverse confusion cases “the commercial injury . . . is an erosion of goodwill and a loss of control over [the senior user’s] reputation, and as such is a more subtle injury than the customary diversion of trade engendered by direct confusion.”¹¹³ The Sixth Circuit characterizes the injury in similar terms: “[T]he senior user loses the value of the trademark—its product identity, corporate identity, control over its goodwill and reputation, and ability to move into new markets.”¹¹⁴ In “‘saturat[ing] the market’ with a trademark that is confusingly similar to that of a smaller, senior registrant for related goods or services,” a reverse confusion defendant “does not seek to benefit from the goodwill of the senior user”¹¹⁵—the loss to the plaintiff is not a transfer to the defendant. Thus, unlike in ordinary trademark infringement cases, the defendant in reverse confusion cases gains nothing that the plaintiff would otherwise have enjoyed (such as diverted sales or reflected goodwill), it merely avoids the inconvenience (and perhaps expense) of having to select a different mark—of having to work around the plaintiff’s rights. Destroying the senior user’s goodwill—the injury that reverse confusion doctrine seeks to redress—is in this sense a kind of collateral damage, a byproductof the junior user’s indifference to the senior user’s existence.¹¹⁶ Proving this distinctive injury requires somewhat different evidence than is required in other trademark infringement cases.

2. Evidentiary Distinctions

While the typical infringement case attempts to assess the likelihood of confusion of actual or potential purchasers of the defendant’s products (on the theory that consumers might purchase such products in the mistaken belief they originate with the plaintiff), in reverse confusion cases, “[t]he relevant confusion is confusion that affects the purchasing decisions of actual or prospective purchasers of the plaintiff’s product.”¹¹⁷ Either confusion as to source or, more likely, confusion as to sponsorship or affiliation can be relevant.

This difference affects a number of the evidentiary factors considered in the multifactor tests each circuit deploys to assess likelihood of confusion. For example, survey evidence of actual consumer confusion in forward confusion cases typically requires surveying potential purchasers of the defendant’s products—they are, after all, the most likely to encounter the defendant’s allegedly misleading trademark.¹¹⁸ By contrast, in reverse confusion cases, “[w]here . . . the relevant issue is whether consumers mistakenly believe that the senior user’s products actually originate with the junior user, it is appropriate to survey the [plaintiff] senior user’s customers.”¹¹⁹ Unfortunately, such a survey is unlikely to generate any useful results where the defendant has not yet fully rolled out their new brand because consumers will not yet have been exposed to the ubiquitous marketing a large junior user can deploy. And of course, once that rollout has happened, the damage to the plaintiff is done and perhaps irreversible. Thus, Professor McCarthy has gone so far as to say that “a survey cannot be run in a reverse confusion case.”¹²⁰

Similarly, while in the ordinary infringement case the renown of the plaintiff’s mark correlates with a higher likelihood of confusion (on the theory that consumers encountering the defendant’s mark will be reminded of the plaintiff’s already-familiar trademark), in reverse confusion cases it is the renown of the junior user’s mark (i.e., the defendant’s mark) that makes confusion more likely.¹²¹ Indeed, the less well-known the senior plaintiff’s mark is compared to the defendant’s mark, the more likelyreverse confusion becomes.¹²² This analysis of mark strength highlights a key characteristic of successful reverse confusion claims: a large disparity in the size and economic power of the parties (hence the gerbil-trampling elephant and squirrel-with-a-bazooka metaphors of Uber Promotions).¹²³

Finally, as noted above, it seems that the plaintiff must show that the defendant knew of and intentionally disregarded the plaintiff’s rights in order to make out a reverse confusion claim.¹²⁴ This intent element is an especially distinctive feature of the doctrine that bears further investigation.

3. Registration, Geography, and Notice

The emphasis in reverse confusion cases on the junior user’s awareness of and indifference toward the senior user’s rights requires a brief digression into the mechanics of notice and priority in federal trademark law. Federal trademark registration provides nationwide constructive notice of the registrant’s claim of rights in the registered trademark,¹²⁵ as well as nationwide priority of right to that trademark (with some limited exceptions).¹²⁶

“if the use of the marks by the registrant and the unauthorized user are confined to two sufficiently distinct and geographically separate markets, with no likelihood that the registrant will expand his use into defendant’s market, so that no public confusion is possible, then the registrant is not entitled to enjoin the junior user’s use of the mark.”

Dawn Donut Co. v. Hart’s Food Stores, Inc., 267 F.2d 358, 364 (2d Cir. 1959); accord Minn. Pet Breeders, Inc. v. Schell & Kampeter, Inc., 41 F.3d 1242, 1246 (8th Cir. 1994). With respect to unregistered marks, however, the rights of a senior user have long been circumscribed by the so-called “Tea Rose/Rectanus” doctrine. In Hanover Star Milling Co. v. Metcalf,¹²⁷ the Supreme Court held that two different users of the “Tea Rose” mark for flour, in two different parts of the country, both unknown to one another at the time each adopted the mark, might in good faith establish concurrent trademark rights, each in their own areas of use.¹²⁸ And in United Drug Co. v. Theodore Rectanus Co.,¹²⁹ the Court held that the senior of two geographically remote concurrent users could not rely on its national-level priority to dislodge the established rights of the junior user in those parts of the country where the junior user had used the mark first—that indeed the senior national user could be enjoined from expanding into that area.¹³⁰ Professor McCarthy encapsulates the doctrine emerging from these cases in two complementary rules:

(1) The territorial scope of an unregistered mark is limited to the territory in which the mark is known and recognized by those in the defined group of potential customers.

(2) The national senior user of an unregistered mark cannot stop the use of a territorially “remote” good faith national junior user who was first to use the mark in that territory.¹³¹

The Tea Rose/Rectanus doctrine calls into question the imposition of reverse confusion liability in cases involving a senior user’s unregistered mark. At most, those plaintiffs should be entitled to injunctive relief only with respect to the geographic area of their actual use (and perhaps damages corresponding to compensable injuries in that area).¹³² To be fair, where a senior user has not registered their mark they are often hard put to establish that they have any rights at all, let alone rights sufficient to displace a much larger junior user whom consumers have already come to associate with the mark.¹³³ But reverse confusion plaintiffs can discharge this burden—and have done so—obtaining far broader relief than the Tea Rose/Rectanus doctrine would suggest they have any right to. The district court in Big O enjoined Goodyear from advertising or using the contested mark “at any locality in the United States,”¹³⁴ and the Tenth Circuit did not review this aspect of the judgment (perhaps because Goodyear did not specifically address it in its appeal).¹³⁵ Likewise, the court in Attrezzi awarded a permanent injunction against any use of the disputed trademark, despite the fact that the plaintiff had not even applied for a federal registration for the mark by the time the defendant adopted and sought to register the mark itself.¹³⁶ In contrast, the court in Uber Promotions did in fact limit its injunction to the area of the plaintiff’s actual use—but this seems to be a result of the court’s finding that the defendant was “mostly without sin” insofar as it adopted its mark in ignorance of the plaintiff’s use, only becoming aware of that use when it later expanded into the plaintiff’s market.¹³⁷

In short, something about the factual predicates of reverse confusion cases seems to be taking precedence over background principles of trademark law. That something, I will argue, is the combination of economic power and a particular form of intent.

4. Notice, Power, and Bad Faith

The cause of action for trademark infringement evolved out of the law of unfair competition, which in turn evolved out of the common-law action on the case in the nature of deceit (and related equitable claims).¹³⁸ As such, early forms of the cause of action required some element of bad intent on the part of the defendant—typically, the intent to pass off the defendant’s goods as those of the plaintiff.¹³⁹ But as rights to certain “technical” trademarks (i.e., inherently distinctive marks applied to particular goods) came to be separately recognized as propertyrights, infringement with respect to such “technical” trademarks could be found even in the absence of proof of the defendant’s intent to deceive (on the theory that the defendant could have no good-faith reason to use the plaintiff’s mark).¹⁴⁰ Over the course of the early twentieth century, as trademark law shed the historical distinction between “technical” trademarks (which could be infringed without a showing of culpable intent) and mere “trade names” (which could not),¹⁴¹ it came to focus on the state of mind of consumers rather than that of accused infringers.¹⁴²

Thus, by the time Big O was decided (and continuing to today), in most areas of trademark law a defendant could not avoid liability by claiming they were ignorant of the plaintiff’s trademark—i.e., by arguing good faith. The only question is whether consumers are likely to be confused by the presence of both parties’ marks in the marketplace, and the defendant’s knowledge or state of mind is—at least according to the case law—not a prerequisite to a finding of infringement. Indeed, some courts have gone so far as to discount even the relevance of the defendant’s intent to a finding of likelihood of confusion—the key element of trademark infringement.¹⁴³ But while good intent may be no defense to trademark infringement, badintent still seems to matter in ordinary infringement cases. As Barton Beebe has shown, a finding of bad faith almost always leads to liability: while not a necessary basis for a finding of infringement, it seems to be a sufficient one.¹⁴⁴ In light of this history, the role of intent in reverse confusion cases is unique and curious: good faith does seem to operate as something close to a defense, and bad faith as a necessary element of the cause of action.

As the Third Circuit explains, the relevant state of mind differs in reverse confusion cases compared to traditional trademark infringement cases:

[I]ntent to confuse is relevant to both reverse confusion and direct confusion. The difference is that the tenor of the intent to confuse evidence changes from the deliberate intent to palm off or exploit the goodwill of the senior user’s mark (deliberate [forward] confusion), to the deliberate intent to push the senior user out of the market (reverse confusion).¹⁴⁵

In this framework, something beyond simple negligence is required to establish relevant bad faith: “mere carelessness” is not enough.¹⁴⁶ Courts have thus been at pains to assert that bare knowledge of the existence of the senior user’s mark, by itself, is not sufficient. They are especially likely to emphasize this point where plausible reasons are offered for the junior user to believe that the senior user held rights too narrow in scope to limit the junior user’s freedom to use the contested mark, particularly when the defendant relied on advice of counsel.¹⁴⁷ This gloss is perhaps designed to guard against the problem, described above, of small, unsuccessful companies unfairly blaming larger companies for their own failures.

The upshot is that infringement by reverse confusion is treated almost as if it were an intentional tort. As noted above, a defendant’s ignorance of the senior user’s mark is a substantial obstacle to a reverse confusion claim.¹⁴⁸ Conversely, all of the cases reviewed above in which reverse confusion was found involved—at a minimum—the defendant’s awareness of the plaintiff’s mark at the time it adopted the contested mark or expanded into the plaintiff’s area of use, and a conscious decision to proceed regardless.¹⁴⁹ As Professor McCarthy puts it: “The paradigm case is that of a knowing junior user with much greater economic power . . . .”¹⁵⁰

McCarthy’s point about economic power is the other distinguishing feature of reverse confusion, and it leads to the theoretical questions that will occupy the rest of this paper. The power imbalance between the parties in reverse confusion cases wouldn’t seem to matter if all we care about is consumer confusion. Moreover, if efficient allocation of trademarks were our goal we might favor larger firms on the theory that they enjoy greater economies of scale in their marketing. But that same power imbalance seems to trigger powerful moral intuitions that generate the outcomes in reverse confusion cases. Those intuitions, and those outcomes, are hard to square with prevailing theories of trademark law. This suggests two alternative possibilities: either the cases are wrong, or the theories are.

II. Misappropriation and Search Costs: Inadequate to the Task

A. Reverse Confusion and Misappropriation

Reverse confusion is essentially an inversion of misappropriation. As I have put it elsewhere, “[t]he anti-misappropriation principle, at its core, is that it is wrongful and therefore actionable for a competitor to gain a commercial advantage from the efforts of another, even if that advantage does not directly harm the person whose efforts have been misappropriated.”¹⁵¹ Thus, misappropriation arises when the defendant extracts value from some aspect of the plaintiff’s investment or effort—in trademark cases, we call this value goodwill. But the key feature of misappropriation is that it does not necessarily result in any direct injury to the plaintiff. The plaintiff need not lose a sale, or suffer reduced goodwill going forward, for misappropriation to be deemed wrongful.¹⁵²

Reverse confusion presents exactly the opposite scenario. A reverse confusion defendant injures the plaintiff by wiping out their goodwill, but the injury is entirely gratuitous. The defendant gains nothing from the plaintiff’s goodwill; it is building goodwill of its own, through its own investment and effort. As discussed above, the injury to the plaintiff is a byproduct of the defendant’s goodwill-creating efforts—an afterthought. To be sure, there is a harm to the plaintiff in reverse confusion cases—even a grievous harm. But that harm has nothing to do with misappropriation.

Rather, what the defendant takes from the plaintiff in reverse confusion cases is the ability to effectively use a particular symbol to invoke its own goodwill. That may well leave the plaintiff worse off than they would have been otherwise, but it is not a misappropriation injury. This is because the symbol appropriated by the defendants in reverse confusion cases (as opposed to the goodwill the plaintiff’s customers associate with that symbol) was not created by any labor or investment of the plaintiff. None of the disputed marks in the reverse confusion cases reviewed above are the type of coined or invented terms referred to in trademark doctrine as “fanciful” marks.¹⁵³ And in each of those cases, the defendant identified the disputed mark for adoption independently, only later discovering the plaintiff’s use of the mark. In short, reverse confusion defendants adopt a mark already in use by the plaintiff not because of but despite that prior use.

There is nothing in the normative framework of misappropriation—the prohibition against reaping where one has not sown—to justify reverse confusion liability. This distinguishes reverse confusion from most other innovations in the scope of trademark liability over the past century.¹⁵⁴ But as the next Section argues, the other standard theory for trademark liability—the economic theory of search costs—fares no better.

B. Economic Theories of Trademark

1. Search Costs

Economists generally argue that “[t]he fundamental purpose of a trademark is to reduce consumer search costs by providing a concise and unequivocal identifier of the particular source of particular goods,” and that “[t]his in turn gives producers an incentive to maintain high and uniform quality, since otherwise the investment in their trademarkmay be lost as customers turn away in disappointment from the brand.”¹⁵⁵ An infringer under this theory is a kind of parasite, feeding on the information value stored in another’s trademark in a way that prevents the owner from recouping their investment in the mark and undermines both the incentive to make that investment and the value of the mark to consumers as a reliable signal of source or quality. Even the misappropriation theory of trademark protection can be squared with this account to the extent that trademark owners might be disinclined to invest in an information resource unless they can capture 100% of its value as a marketing device.¹⁵⁶

But any attempt to extend this logic to reverse confusion must grapple with two complications. The first is the point raised above in the context of misappropriation: the reverse confusion defendant is not trying to extract any of the informational value—the goodwill—associated with the plaintiff’s use of the mark, and does not in fact do so. But this point only goes so far. After all, the defendant need not extract value for itself from the plaintiff’s investment in order to destroy the value of that investment, and both the incentive effects toward the plaintiff and the effect on the value of the trademark to consumers as an information resource that would result from such destruction remain negative. In these respects, reverse confusion could still be said to decrease aggregate welfare.

The real problem for search-costs theory is the second complication. Even though reverse confusion may destroy substantial informational value for the plaintiff and its customers, the relative size of the parties in such cases means that the defendant is creating far more informational value than they are destroying. This follows of necessity from the factual predicate of a reverse confusion claim: a far larger and better-known junior user “saturates the market”¹⁵⁷ with advertising and products—lowering the search costs of a far larger population than the senior user’s customer base. This is a key distinction between reverse confusion cases and any other type of trademark claim: in other trademark cases the value of the defendant’s use is typically lower than (or at best comparable to) the value of the plaintiff’s use. The fact that the value of the defendant’s use of a contested trademark is so clearly higher than that of the plaintiff in reverse confusion cases poses a fundamental challenge to economic theories of reverse confusion liability, because (under the Kaldor-Hicks criterion at least), the first-order (or “static”) effects of the defendant’s behavior on allocative efficiency would seem to be positive.¹⁵⁸ Indeed, even if we view the choice in reverse confusion cases as deciding which party will be forced to suffer the sunk costs and new burdens of developing and marketing a second-best trademark, as the size difference between the parties increases it becomes less plausible that imposing those costs on the senior user is justified.¹⁵⁹ It would be a strange welfarist theory of trademarks that prohibited the increases in aggregate welfare that result from the destruction of a plaintiff’s goodwill as a byproduct of the simultaneous creation of a defendant’s far more valuable goodwill.¹⁶⁰

There is, however, another possible argument grounded in economic analysis that might be offered as a justification of reverse confusion doctrine. This argument depends not on reverse confusion’s first-order (static) effects, but on its second-order (dynamic) effects. According to this argument, the problem with reverse confusion is not that aggregate welfare will be lower (due to higher aggregate search costs) if the plaintiff’s few customers lose a poor information resource while the defendant’s many customers gain a new, more valuable one. Rather, it is that the very process by which that substitution takes place is a pure transfer from the plaintiff and its customers to the defendant and its customers: a non-market exchange, or a theft. Because of the disconnect between a reverse confusion defendant’s use of a mark and the plaintiff’s goodwill, this argument already strains against a fair amount of black-letter law about the absence of “rights in gross” in trademarks,¹⁶¹ because it requires us to treat the plaintiff’s interest as in the mark itself rather than in the associations consumers have with the mark as it is used on a particular producer’s goods. But it merits consideration, because theft is an intuitively accessible characterization of the harm of reverse confusion, and because non-market exchanges pose particular problems in standard economic theory.

2. Reverse Confusion as Theft

Nonconsensual transfers—thefts—may indeed increase aggregate welfare as a matter of first-order effects if the thief values the stolen property more than the owner, but such thefts are also believed to give rise to indirect or “dynamic” costs that may outweigh any such gains.¹⁶² These costs include “the investments of capital and labor in the activity of theft and in protection against theft,” often represented by the opportunity costs of those investments.¹⁶³ If I am worried you will rob my house, I’ll buy a lock, but if you still really want to rob my house, you’ll buy a lock pick, and then I’ll have to buy an alarm system, and you’ll have to spend time practicing catburglary, and on and on. Each of us spends time and resources that might be put to actually producingvalue into the effort to effectuate or prevent mere transfers of value, leaving everyone worse off.

But there are obstacles to any effort to convert this economic argument against theft into a justification of reverse confusion doctrine. The first is that there are indirect costs of market exchange as well, and they may, in turn, exceed the indirect costs of theft. As Richard Hasen and Richard McAdams explain:

Notwithstanding the existence of indirect costs, theft is efficient if incurring those costs avoids incurring larger transaction costs from a voluntary sale and the thief values the good more than the owner does. More generally, an efficiency analysis requires comparing the indirect costs of legalized theft with the transaction costs of market sales. . . . Thus . . . the real problem with theft is its indirect costs.¹⁶⁴

Thus, to formulate an economic defense of reverse confusion doctrine grounded in the theory of theft, we would need an account of the indirect costs of reverse confusion itself (the theft case), as well as an account of what voluntary transfer would look like in reverse confusion scenarios, and what transaction costs such transfers would generate. We would conclude that reverse confusion ought to be prohibited only if reverse confusion scenarios fail one of the efficiency criteria for theft offered by Hasen and McAdams: “(i) the thief values the good at some amount greater than the owner does, (ii) the indirect costs of the theft are less than the transaction costs of a sale, and (iii) the theft avoids the transaction costs.”¹⁶⁵

As discussed in the previous Section, in reverse confusion cases, by definition, the junior user (thief) and its much greater number of customers will value the trademark at a greater amount than the senior user (owner) and its much smaller number of customers do simply by virtue of their relative size. So the first criterion for permitting theft seems satisfied. The remaining criteria require us to identify and compare the indirect costs of reverse confusion—the types of measures senior users are likely to undertake to prevent it, and the types of measures junior users are likely to undertake to circumvent the senior users’ measures, and the social cost of both—with the transaction costs of a negotiated sale of the senior user’s rights to the junior user.

What might the indirect costs of reverse confusion be? Trademarks are intangible assets; unlike physical property they cannot be locked in a safe or guarded behind a fence. They are also assets whose only value is in their public use in the marketplace; attempting to keep one’s trademark secret or obscured is not a plausible protective measure. Thus, if the senior user’s goal to try to deter potential junior users from adopting the senior user’s mark, the most effective way to do so is to put those potential junior users on notice of the existence of the senior user’s claim of rights in the mark, and to threaten costly retaliation against any junior user who encroaches on that claim.

The most obvious way to achieve these goals in the trademark system is to obtain a federal registration—a tool designed precisely to lower the costs of providing notice of trademark claims.¹⁶⁶ As noted above, registration provides nationwide constructive notice and priority,¹⁶⁷ and its cost is quite low—the application and maintenance fees that fund the registration system are a few hundred dollars per registration, and the full cost of obtaining a registration is typically no more than two to three thousand dollars when effective clearance searches and professional legal assistance are factored in (assuming the application is not contested).¹⁶⁸ Once notice and priority are established, the senior user would also have to credibly threaten to litigate any encroachment on their registered rights, thereby deterring any potential junior users by threatening them with losses stemming from litigation expenses and an uncertain prospect of being able to retain any value from their adoption of the trademark. The senior user would also have to be prepared to follow through on that threat, incurring legal expenses of their own. Conversely, we might expect would-be junior users to attempt to circumvent these protective measures with legal machinations of their own, such as obtaining legal opinions supporting their right to adopt the senior user’s trademark and spending large amounts on litigation themselves.

As it happens, we observe all these measures and countermeasures in the reverse confusion cases described in Part I. Many (though, importantly, not all) reverse confusion plaintiffs own trademark registrations,¹⁶⁹ all of them obviously litigated their claims, many defendants cite advice of counsel in mitigation of their allegedly wrongful behavior,¹⁷⁰ and reverse confusion litigation can drag on for years.¹⁷¹ By far the largest component of these indirect costs of reverse confusion are legal costs—mainly litigation expenses, which range from a few hundred thousand dollars in low-value cases to over three million dollars in high-value cases.¹⁷² We must thus determine whether these costs are likely to be greater than or less than the transaction costs of a voluntary sale—an assignment or license of the senior user’s rights, a concurrent use arrangement, or a covenant not to sue. Again, we should only expect an economic theory to demand prohibition of reverse confusion if these indirect costs of reverse confusion are greater than the transaction costs of a voluntary sale.

There is good reason to think they are not. Such transaction costs arise from multiple sources. In the first instance, the junior user must be able to identify and locate the senior user to be able to bargain with them.¹⁷³ In the trademark context, these are the costs of a clearance search, which are quite low for federally registered marks but increase substantially for unregistered marks—and as noted above, not all successful reverse confusion plaintiffs own trademark registrations.¹⁷⁴ This achieved, however, there still remains the substantial problem of strategic behavior. Knowing that the junior user has already sunk at least some costs in pursuing the trademark, and has much deeper pockets, the senior user has strategic incentives to hold out for a much higher price than their own subjective valuation of the trademark and its goodwill. It is precisely this danger of holdouts—and the potential rent-seeking behavior that might lead to holdout situations—that motivates the law’s strict insistence on a trademark owner’s actual use of a mark in commerce as a condition of acquiring and maintaining trademark rights.¹⁷⁵ Conversely, the junior user, knowing they are better able to bear the costs of litigation, has strategic incentives to lowball the senior user in negotiations over the transfer of rights in the hopes that the senior user will capitulate rather than face expenses it cannot bear. Indeed, in Big O and Uber, Inc., bargaining over an allocation or transfer of the disputed trademark rights was attempted but failed.¹⁷⁶

These types of reciprocal holdout problems—or bilateral monopoly bargaining situations—generate substantial transaction costs, though economists generally expect them not to prevent most positive-sum transactions assuming bargainers are acting rationally (though this is by no means a safe assumption, and there are reasons to think holdout problems can often cause bargaining to fail).¹⁷⁷ This does not tell us whether the costs of bilateral monopoly are less than the indirect costs of reverse confusion (i.e., of reverse confusion litigation). This is the question we are ultimately trying to determine, which if answered in the affirmative would provide an economic justification for prohibiting reverse confusion. And there is one more, perhaps most important, category of transaction costs we have yet to discuss.

There is a substantial collective action problem in trademark transactions (and, for that matter, litigation).¹⁷⁸ This is because the value of a trademark is not enjoyed exclusively by its owner; the trademark owner’s customers capture a great deal of that value. This, indeed, is the very premise of the search costs theory of trademarks: that they serve as an information resource that consumers can access in making purchasing decisions. But the value of this resource is widely dispersed over the relevant consuming public—which in the case of a reverse confusion defendant is, by definition, a huge number of people. The problem of bargaining for the transfer of trademark rights from a smaller senior user to a larger junior user is thus a kind of free rider problem: bargaining is frustrated by the impossibility of coordinating contributions from all of the millions of consumers who might stand to gain from the large junior user’s acquisition of the smaller senior user’s underutilized trademark rights. And unless these consumers’ valuations of the mark in the hands of the larger junior user are taken into account, the junior user’s willingness to pay for a transfer of the senior user’s rights will be too low,¹⁷⁹ and even moderately aggressive holding out by the senior user could cause bargaining to fail. Thus, even if the holdout problems discussed above did not generate transaction costs greater than the cost of litigating reverse confusion claims, combining those transaction costs and the costs of bargaining with all the consumers with a stake in the allocation of the disputed trademark is almost surely greater than the costs of such litigation. It seems the second criterion for permitting theft is likely to be satisfied in the reverse confusion context.

This leaves only the third criterion: whether allowing theft avoids the transaction costs discussed above. And for obvious reasons, it does: if the parties are not required to bargain over allocation of a trademark right, they do not face the transaction costs of such bargaining. Of course, allowing large junior users to simply appropriate already-appropriated trademarks would generate other dynamic costs. Opportunistic and deep-pocketed junior users might be marginally more likely to try to expropriate a smaller junior user’s trademark, perhaps without even trying to negotiate, hopeful that they would be able to convince a court that the result is allocatively efficient. Courts would face marginally higher costs of discriminating between welfare-enhancing and welfare-harming transfers. And small businesses, understanding these risks, might expect marginally higher marketing and litigation costs over the long run, which could lead to marginally lower small business formation and marginally higher small business failure. But it seems unlikely that this alternative set of indirect costs of theft would be substantially greater than the costs of the reverse confusion litigation we already observe. It is hard to believe they would add enough to the indirect costs of reverse confusion to outweigh the transaction costs of bargaining reviewed above combined with the welfare gains resulting from allocation to the larger junior user.

In short, if reverse confusion is trademark theft, it is probably efficient theft—an economist would conclude that the law ought to permit it. But just saying that a much larger junior user ought to be permitted to appropriate the mark of a much smaller senior user does not tell us all we need to know about the conditions under which such transfers ought to take place. The main question is whether the junior user’s commandeering of the senior user’s trademark ought simply to be tolerated, with no relief available to the dispossessed plaintiff, or whether the senior user ought to be compensated for the loss of value they once derived from the repository of goodwill represented by the transferred trademark. This is a question that the copyright regime—with its mix of uncompensated fair use and compensated statutory licenses—deals with in detail.¹⁸⁰

An economist would approach this question by trying to formulate a legal framework that minimizes the sum of transaction, administration, and error costs attending the transfer of rights—whether voluntary or otherwise—where such transfers would increase aggregate welfare. But obviously, the existence of reverse confusion liability is inconsistent with the view that a small junior user ought to be left without any remedy whenever a larger senior user adopts their trademark. So the toleration or fair-use strategy—even if it were called for by economic analysis of reverse confusion scenarios—does not appear to be able to explain reverse confusion doctrine. This brings us to the final potential economic justification for reverse confusion doctrine: the familiar property rule versus liability rule framework of Guido Calabresi & Douglas Melamed.¹⁸¹

3. Liability Rules

If we do not want to disturb our general rule allocating rights in trademarks according to priority in time,¹⁸² the economist’s task is to establish a legal framework for reverse confusion scenarios that minimize the transaction costs attending the transfer of trademark rights from the lower-valued senior user to the higher-valued junior user. And if, as was argued in the previous Section, the transaction costs of voluntary transfers under what Calabresi and Melamed call a “property rule”¹⁸³ are excessive, we need an alternative rule. As Calabresi and Melamed famously argued, where the transaction costs of market transfers are too high, we ought instead to fashion a “liability rule” to govern such transfers.¹⁸⁴ “Whenever someone may destroy the initial entitlement if he is willing to pay an objectively determined value for it, an entitlement is protected by a liability rule.”¹⁸⁵ And “a very common reason, perhaps the most common one, for employing a liability rule rather than a property rule to protect an entitlement is that market valuation of the entitlement is deemed inefficient, that is, it is either unavailable or too expensive compared to a collective [i.e., government-determined] valuation.”¹⁸⁶ Such liability rules create costs of their own—principally the costs of the state apparatus set up to enforce them, and the error costs attending government estimation of the value of the entitlement.¹⁸⁷ But the types of bargaining costs typical of reverse confusion cases—particularly holdout problems created by vastly unequal stakes and collective action problems among parties not represented before the court—are analogous to the types of bargaining costs that lead courts to adopt liability rules in other contexts.¹⁸⁸

Treating reverse confusion as subject to a liability rule would be consistent with an economic approach to reverse confusion that depends less on allocative efficiency in individual disputes and more on the overall systemic efficiencies generated by a rule that uses priority in time as a basis for allocating trademarks—a rule-utilitarian rather than an act-utilitarian approach.¹⁸⁹ Under such a view, the fact that in some cases a junior user will be the higher-value user of a disputed trademark is not sufficient to justify the departure from a rule allocating trademarks to their first adopter—a rule that, if followed, allows trademark users to rely on settled rights and thereby provides the incentive to invest in their trademarks in the first place. An alternative rule that required—or indeed, even allowed—case-by-case judicial evaluation of which party is making the highest value use of any disputed trademark might fail to engender sufficient confidence in trademark rights to incentivize the investment in quality that search-costs theory promises. On this view, the allocative question is resolved in favor of the senior user as a rule, with the understanding that this rule, while best for the system as a whole, might result in allocative inefficiency in particular cases. To mitigate those cases, complementary rules might be deployed. For example, trademarks allocated according to priority in time might be made transferrable in voluntary transactions, to allow them to move from lower-value users to higher-value users. Moreover, the transaction costs associated with such transfers could be mitigated by providing a centralized registry providing notice of priority-based claims (allowing would-be junior users to identify bargaining partners at low cost). Finally, consistent with Calabresi’s and Melamed’s analysis, infringement remedies available to senior users might be limited to compensation in categories of cases (such as reverse confusion cases) where other transaction costs would clearly and systematically frustrate such welfare-increasing transfers.

This is a plausible theory, but reverse confusion doctrine does not conform to it. First and most obviously, reverse confusion cases do not adopt liability rules. All but one of the successful reverse confusion claims reviewed in Sections I.A. and I.B. above that proceeded to judgment awarded an injunction—the hallmark of a property rule—and as discussed in Section I.C.3 above the injunctive relief awarded was often far broader in scope than background principles of trademark law would provide.¹⁹⁰ Second, the cases do not appear to adopt any of the other subsidiary rule-based strategies we would expect from a system whose goal is efficiency. For example, in tort law (generally governed by liability rules), “economic efficiency asks for that combination of entitlements to engage in risky activities and to be free from harm from risky activities which will most likely lead to the lowest sum of accident costs and of costs of avoiding accidents,”¹⁹¹ which often entails allowing the costs to fall on the party who could have prevented a loss from occurring at the lowest cost.¹⁹² But reverse confusion law conspicuously fails to do this. The most obvious example is the imposition of reverse confusion liability in cases where the plaintiff neglected to obtain a federal registration—a singularly low-cost way of putting potential junior users on notice and allowing them to avoid reverse confusion conflicts in the first place—even where the defendant ran up substantial expenses preparing to use the mark before it became aware of the plaintiff’s claim.¹⁹³ For example, in Big O, the record reflected Goodyear’s position that “Big O had failed to protect itself by spending the money to register the [disputed] name,” that “Goodyear could not let go of its massive investment in the BIGFOOT campaign,” but that Goodyear was willing to compensate Big O for the loss of goodwill or investment resulting from the public’s association of the disputed mark with Goodyear.¹⁹⁴ Big O nevertheless sought and obtained a permanent injunction.¹⁹⁵ Similarly, in Elizabeth Taylor Cosmetics, the defendant had undertaken roughly a year of work to select and prepare to market a perfume under the PASSION mark by the time the plaintiff applied for a federal registration; the plaintiff nevertheless obtained an injunction against any use of the mark by the defendant in marketing channels that overlapped with the plaintiff’s.¹⁹⁶ Conversely, the mere fact that the plaintiff’s mark is registered does not guarantee liability: defendants often prevail in reverse confusion cases even where they are aware of the plaintiff’s senior registered use.¹⁹⁷ In short, to the extent that economic theory might dictate a liability rule regime to compensate senior users in reverse confusion cases, the doctrine does not seem to adopt economic theory.

4. Intentional Torts and Deterrence

One final economic argument might begin from the observation that the apparent notice and intent elements of reverse confusion claims¹⁹⁸ change the calculus of which party is the cheapest cost avoider of a dispute. According to this argument, only defendants on notice of the plaintiff’s claim to a mark at the time they entered the market are held liable for infringement by reverse confusion, and this is because such defendants could (and should) simply have given up and chosen a different mark: a low-cost option to avoid the dispute entirely. This argument reflects the view that reverse confusion operates in practice like an intentional tort, and thus comes under the economic logic of deterrence.¹⁹⁹ But there are multiple problems with this argument. The first is that it does not align with the cases: as just noted, not all defendants on notice of the plaintiff’s use are held liable for reverse confusion.²⁰⁰ The second is theoretical: the logic of deterrence has already been implicitly addressed in rejecting the theory of reverse confusion as theft, by considering both the higher value of the larger junior user’s use of the disputed mark and the indirect costs of nonvoluntary transfers of trademark rights from much smaller senior users to much larger junior users. We do not wantto deter defendants from adopting a disputed mark where the elements of reverse confusion that generate obstacles to private bargaining are present. The fact that an economic account would seem not to call for deterrence in reverse confusion cases also seems to render the distinctive knowledge and intent elements of those cases superfluous—never a good sign for a descriptive theory.

The third problem is that treating reverse confusion as an intentional tort as a matter of economic theory requires us to take into account the costs of deterrence. To conclude that deterrence is efficient, we would have to address cases where a defendant has already innocently incurred substantial expenses preparing to launch a disputed trademark prior to becoming aware of the plaintiff’s senior use. Finding against such a defendant where the plaintiff has failed to take advantage of easily available tools like federal registration (as in Big O and Elizabeth Taylor Cosmetics) is not only flatly inconsistent with the cheapest cost avoider principle, it requires the pure waste of the defendant’s sunk costs, which can often exceed the value of the mark to the plaintiff.²⁰¹

Moreover, requiring a junior user to adopt an alternative mark is not costless—indeed, it can be quite costly, for a number of reasons. Notwithstanding the theoretical musings of early economic theorists of the trademark system,²⁰² there appears to be substantial congestion in the market for useful source identifiers.²⁰³ Acceptable substitutes for the senior user’s mark will not always be available, and even where available they may be quite costly to identify, requiring multiple clearance searches and reliance on extensive legal and other professional advice. Finally, many of the marks in dispute in the reverse confusion cases—marks like PASSION for perfumes, MIRACLE[BRA/SUIT] for figure-shaping garments, and UBER for, well, anything—even if they are not descriptive in the technical sense under the law of trademark distinctiveness,²⁰⁴ nevertheless carry connotations that many competitors might benefit from incorporating into their marketing.²⁰⁵ As the hornbook law of inherent distinctiveness explains, priority in time alone offers a meager excuse to allow a single competitor to monopolize such terms.²⁰⁶ That principle is only stronger where, as in reverse confusion scenarios, a junior user (and its customers) can derive far more value from the term than the senior user (and its customers).

In short, while the reverse confusion theory of infringement seems to have some legal characteristics of an intentional tort, an attempt to model it with the economic theory of intentional torts comes up short.

5. Review of the Case Against Economic Theories of Reverse Confusion

To summarize, economic theories of trademark law do not seem to be able to satisfactorily account for reverse confusion cases. To the extent trademark protection is designed to lower consumer search costs and provide incentives to produce goods of high and consistent quality, assigning trademark rights to a producer with a far smaller number of customers over one with a far larger number of customers would seem to be contrary to theory as a matter of first-order static effects. To the extent that an economic theory of reverse confusion doctrine might depend on preventing non-consensual transfers of trademark rights once those rights have been initially allocated, the theory is challenged by strong reasons to believe that non-consensual transfers will generate far fewer indirect costs than the transaction costs of bargaining over transfers of rights in reverse confusion scenarios. To the extent that this indirect-costs analysis counsels in favor of a compensatory liability rule regime in reverse confusion scenarios, such a regime is not to be found in the case law. And to the extent that this inconsistency leads us to consider a deterrence rationale, we run back into the problem that deterring defendants in reverse confusion cases probably costs far more than it is worth.

In fairness, it is possible that this estimation of the indirect costs of reverse confusion and of the transaction costs of bargaining between a small senior user and a large junior user is mistaken. If so, the reverse-confusion-as-theft theory might be a plausible one, as might the deterrence theory. But if that is the basis of critiquing the analysis of this Part, the critic must admit that the effects of reverse confusion seem analogous to the effects of other situations in which economic analysis concludes that transaction costs counsel against assigning an entitlement to the plaintiff to be protected by a property rule, and moreover that the relative magnitude of transaction costs attending any rule is an empirical question that economic theorists are not currently in a position to settle. And if the estimation is not mistaken—or if at the very least it is plausible—there are two alternative possibilities. Either the case law of reverse confusion is mistaken on a wholesale basis, or the economic theory of trademark law—at least in the area of reverse confusion—is wrong. Assuming the premise that the best theory is the one that agrees most closely with observation, the latter possibility is the one we should accept. The question that remains is whether a better theory exists.

III. Trademark as Promise: Reverse Confusion as Interference with Promising

A. A Review of Trademark as Promise

In prior work I have begun developing a non-economic theory of trademark law grounded in contractualist moral philosophy. By “contractualist,” I am referring to the philosophical tradition, stretching from Immanuel Kant²⁰⁷ to Thomas Scanlon,²⁰⁸ that “purport[s] to derive moral principles from hypothetical reason giving and consensus, the common thread being the ideal of the social contract and the underlying assumptions of equality and mutual respect among moral agents.”²⁰⁹ My key claim is that trademark doctrine appears (as a matter of positive law) to have developed in such a way as to help producers make credible assurances (promises) to consumers, and to impose liability against conduct that interferes with the making, performance, or reliability of such promises. These features of trademark doctrine can be justified under a contractualist theoretical framework that privileges the autonomy of moral agents, and by implication reserves to individual consumers the right to choose whether to accept a substitute for performance of such a promise.

The normative appeal of this theory derives from its grounding in the interpersonal morality of promising. Contractualists hold that making, receiving, and relying on promises is an important part of how each of us secures the cooperation of others necessary to pursue our own ends, and is thus an important aspect of our autonomy.²¹⁰ The equal respect each moral agent owes to the autonomy of others requires us to take due care not to mislead others into relying on expectations of our future conduct that we either cannot or will not fulfill.²¹¹ Thus, when we knowingly create such expectations in others by making promises, we have an obligation to perform according to such expectations, out of respect to our promisee’s autonomy, unless the promisee releases us. A rule requiring promisors to keep their promises, and allowing promisees to enforce performance of promises, is therefore one that a contractualist holds could not be reasonably rejected.²¹²

Trademark as promise extends these moral obligations concerning promising into the domain of legal obligations.²¹³ The basic theory of trademark as promise in the context of classic infringement by passing off is straightforward.²¹⁴ Imagine a consumer C who wants and expects a product made by the plaintiff P. Where a defendant D uses P’s trademark on D’s goods, and C purchases those goods in the belief that he is acquiring a good made by P, contractualist analysis could find D’s conduct wrongful on multiple grounds.

In the easier case in which D’s product is inferior to P’s products, we would say that D has caused C to take an action (purchasing the product) that he would not otherwise have taken, in the expectation that he would receive something from D (a product consistent with the quality of P’s products) that D failed to provide, thereby causing C an injury (acceptance of a lower-value product in place of a higher-value one) that C would not have suffered but for D’s deception. D’s act is wrong under a contractualist framework because D has made C worse off than he would otherwise have been by failing to respect C’s autonomy: the right to “direct one’s efforts and resources toward aims one has chosen and not to have one’s planning co-opted . . . whenever this suits someone else’s purposes.²¹⁵ In Kantian terms, D is, without C’s consent, treating C merely as a means to D’s ends, and not as an end in himself.²¹⁶

But even if D’s product is not of inferior quality to P’s, we would still find D’s actions wrongful, even as against C, under a contractualist theory, according to the following reasoning: Assume C doesn’t simply want a product consistent with the qualities he has come to expect of P’s products. Assume C also wants to be assured that the product was in fact made by P—whether as a guarantee of the product’s unobservable qualities or for any other subjective reason—and in the absence of such assurance, C would not have purchased the product. In such a circumstance, we would still say that D has wronged C, in this case by inducing C to rely on an assurance that only P was in a position to provide. Importantly, this conclusion does not in any way depend on the actual qualities of D’s product—it depends on our respect for C’s autonomy and for the moral value of C’s power of choice. Marketing research has taught us that consumers often make purchasing decisions for reasons we might find dubious, but which are neither irrational nor founded on mistake of fact,²¹⁷ and a contractualist may be bound to respect those reasons as part of their respect for consumer autonomy.²¹⁸ D (knowing of C’s desire for assurance) therefore may not defend the delivery of a product made by D rather than P on the grounds that D’s product is just as good. Crediting such a defense would violate C’s autonomy-based right to make purchasing decisions on the grounds that seem best to C, and would instead substitute a producer’s (or a court’s) judgment that C ought to be satisfied with what C received. In this view, the use of a trademark is tantamount to a promise—an assurance concerning the nature or provenance of the good to which it is affixed—that the seller must perform unless the buyer (C) releases him. Indeed, because the harm of a false promise implicates C’s autonomy rather than the welfare consequences for C of accepting a substitute for what they were promised, the appropriate remedy is specific performance—compensation is inadequate unless C accepts it or unless specific performance is impossible.²¹⁹ For reasons I explain in earlier work, economic theory cannot justify the imposition of liability in this latter case, at least not by reference to the effects on C.²²⁰

B. A Promise-Based Account of Producer Obligations

Because trademark law regulates not only the relationships between producers and consumers, but also the relationships among producers (as a matter of unfair competition), the theory of trademark as promise must also provide an account of the latter category of relationships. This account remains largely in development. In earlier work, I have suggested principles at two different levels of generality that may form the basis for such an account.²²¹ At the higher level of generality, and deriving from broad contractualist principles such as Kant’s formula of universal law or Scanlon’s principle of right and wrong,²²² we might ask whether a rule proscribing certain competitive behavior is one that all rational, reasonable producers would accept as a limitation on their own conduct.²²³ But given the broad potential disagreement as to the answer to this question, it may not be a productive tool to generate legal rules from a blank slate.

Moving to a lower level of generality, we might attempt to formulate principles grounded in the relationship between promising and autonomy. When considering a defendant’s challenged behavior with respect to a trademark, and understanding the use of a trademark on a product to constitute a kind of assurance or promise, we could ask ourselves two questions. First, we might ask: In engaging in the challenged conduct, did the defendant make a promise he was not in a position to perform? Answering this question in the affirmative in the case of passing off (because D was not in a position to provide a good made by P) leads us to conclude that passing off is wrongful, even if the defendant’s product is one we would otherwise expect the plaintiff to be satisfied with. An affirmative answer to this question even outside the context of passing off would generally lead us to conclude the challenged conduct is wrongful if it induces the promisee to act in reliance on a promise under circumstances where their reliance will almost certainly be frustrated, thereby interfering with their autonomy in precisely the same way as in the passing off case.

The second question is in essence an inversion of the first. Precisely because making, receiving, and relying on promises is an important part of how each of us secures the cooperation of others necessary to pursue our own ends, interference with the making, performance, or enjoyment of such promises could be wrongful because of the importance of promising to the pursuit of one’s own ends. We would be reasonable in rejecting a rule permitting all such interference, and it would be unreasonable to reject a rule prohibiting at least some interferences, particularly interfering acts that are in themselves unreasonable or are not privileged on other grounds that cannot be reasonably rejected.²²⁴ We might therefore ask, with respect to a defendant’s challenged conduct, whether that conduct prevents someone else from either (i) making or performing a promise they would otherwise have been in a position to make or perform, or (ii) receiving the benefit of a promise they would otherwise have been in a position to receive.²²⁵ In the rest of this Part, I will consider whether this question helps us make sense of reverse confusion doctrine.

C. Reverse Confusion and Interference with Promising

When applying our question regarding interference with promising to reverse confusion doctrine, we see good reasons to find reverse confusion to be wrongful. We can even see explanations for some of the more distinctive features of the case law.

First, unlike in passing-off cases (where the harm is in the defendant’s making of a promise it is in no position to perform), the requirement of likely confusion provides the basis for concluding that a reverse confusion defendant’s behavior will interfere with the plaintiff’s ability to make and perform promises in the future through the use of its mark, as well as the ability of consumers to rely on such promises.²²⁶ If the plaintiff’s potential customers see the plaintiff’s mark and believe it constitutes some assurance about the relationship between the marked goods and the defendant, the plaintiff will be both unable to deliver on that assurance and unable to make a truthful assurance about the relationship between the marked goods and the plaintiff—an assurance upon which the plaintiff is in a position to deliver. Of course, this would be true even if the plaintiff were not a prior user of the mark: any one of us could just as easily complain that we are unable to use the mark BIGFOOT to give consumers assurances about the source or quality of our own radial tires now that Goodyear has adopted the mark for its goods. The plaintiff’s distinctive grievance in such a circumstance depends on some special claim of the plaintiff to use the mark as an instrument of promising, and in the context of the trademark system this claim is grounded in the principle of priority in time.²²⁷ But as I have noted elsewhere, “[w]hen considering whether a moral principle can be justified, a contractualist looks to principles of fairness: ‘We have reason to object to principles simply because they arbitrarily favor the claims of some over the identical claims of others: that is to say, because they are unfair,’” and priority in time is an apparently arbitrary basis to favor one trademark user over another.²²⁸ It is for that reason that any plausible contractualist argument about wrongful interference with the use of a trademark to give assurances about the promisor’s future performance

appears to depend on the autonomy of third parties—consumers—and the ways such consumers might come to rely on trademarks in making their way through the world. This justification, in turn, seems to depend on a particular kind of consumer reliance—reliance not on stable meaning, which nobody in a free and dynamic society is in a position to provide, but on performance of promises to deliver goods and services.²²⁹

This observation not only rescues our conclusion that reverse confusion wrongfully interferes with promising, it explains another observed feature of reverse confusion case law: the readiness of courts to summarily find in favor of defendants where the plaintiff’s own past business failures suggest it lacks a reliable reputation among consumers.²³⁰

The involvement of the plaintiff’s customers in our evaluation of the wrongfulness of the defendant’s conduct toward the plaintiff takes us back to the origins of trademark law in the common law action on the case in the nature of deceit.²³¹ As in that ancient form—and indeed in the modern action for tortious interference with prospective contractual relations²³²—the law’s function is to provide a mechanism to recover for an injury done indirectly to the plaintiff by means of an independently wrongful act to a third party. The reverse confusion defendant’s independent wrong is the deception of the senior user’s potential customers—leading them to believe and perhaps rely on some assurance about the plaintiff’s goods or services that is not true. Though the deception works against the customers, the injury is also felt by the senior user, who has lost a previously enjoyed ability to communicate through their trademark to their actual and potential customers.

Of course, we could tell much the same story about forward confusion doctrine as we have been about reverse confusion. Both create similar interference with promising by the plaintiff and reliance by consumers (albeit a different group of consumers). While this consistency and parsimony is a good result for the theory of trademark as promise in terms of its descriptive power, it remains to be seen whether trademark as promise can account for the other distinctive features of reverse confusion doctrine, particularly those that we could not account for with economic theory. There is good reason to think it can. The remainder of this Part will focus on two: the size disparity between the parties that defines reverse confusion cases, and the apparent requirement that a defendant must have had knowledge of the plaintiff’s senior use and intentionally disregarded any rights arising from that use in order to be held liable.²³³

Notably, unlike the search-costs theory, the theory of trademark as promise does not necessarily need to separately account for the focus on the relative size of the parties in reverse confusion cases, because of its assumption of the equality of moral agents.²³⁴ While the welfare-maximization principle of the economic account means such an account must explain why a higher-value user of a mark should give way to a lower-value user of the same mark, the contractualist account of trademark as promise bears no such burden. The normative prescriptions of trademark as promise turn on neither the number nor the wealth of consumers (nor, for that matter, of producers), and therefore the theory faces no theoretical challenge in applying liability against a larger user of a contested trademark.

The theory of trademark as promise can, however, provide some insight into the strong moral intuitions pumped by the “David and Uriah” framework of reverse confusion cases,²³⁵ precisely because of its assumption of the equality of moral agents. By treating a smaller senior user as if its trademark interests (or the interests of its customers) must give way because of the larger junior user’s putatively more important trademark interests, the junior user in reverse confusion cases signifies by its actions that it rejects the foundational assumption of the equal autonomy of moral agents. Rejection of the equality of others is not a position that can coherently be adopted by moral agents seeking principles that none could reasonably reject: it is logically impossible for each producer to be more important than every other producer.²³⁶ Thus, behavior motivated by such arbitrary self-preference fails the various contractualist standards for right conduct.

Moreover, the theory of trademark as promise can account for both the apparent presence of an intent element in reverse confusion cases and the absence of such an element in forward confusion cases. On the latter point, consider hornbook contract law regarding intent of a promisor to mislead or deceive a promisee. If I intentionally mislead you into entering into an agreement with me that I know I will not or cannot perform, and you are injured as a result, you can be excused from performance and I can be held liable for fraud.²³⁷ But my intent to mislead you is not required for you to hold me to my promise if I knowingly promised you something it turns out I am not in a position to deliver.²³⁸ A contractualist would infer such principles from my obligation to fulfill promises I made knowing others would rely on them, out of respect for the autonomy of those others, whether or not I realized I was promising beyond my ability to perform.²³⁹ So where the theory of trademark as promise designates adoption of another’s trademark as wrongful because it constitutes the making of a promise the promisor is not in a position to perform, the promisor’s prior knowledge of the senior user would not seem to be a necessary condition of that designation, though it would be a sufficient condition. Indeed, this observation provides a theoretical explanation for Barton Beebe’s empirical finding that a showing of bad faith by the defendant nearly always leads to a finding of infringement in forward confusion cases.²⁴⁰

Where, however, the designation of conduct as wrongful is grounded in its interference with others making, performing, or relying on promises, things are more complicated. There are any number of acts a person could take that might make it more difficult for others to pursue their goals effectively. It is not a requirement of contractualism that we avoid all such actions, only that we avoid those actions that would be forbidden by a rule that others could not reasonably reject. It would be reasonable to reject a rule that forbids any unintentional interference with the pursuit of goals by others, because avoiding all such interference is in practice impossible. But if I am aware of the effect my actions will have on others’ ability to pursue their goals, I have an obligation to take due care not to lead those others to rely on false information, either about what I will do or about other relevant facts in the world within my knowledge.²⁴¹

This obligation helps explain the intent element of reverse confusion doctrine. Where the defendant knows when placing its goods into the stream of commerce that consumers will rely on the mark as a promise regarding the source or quality of both the plaintiff’s goods and the defendant’s goods, the defendant acts wrongfully because it knows that some consumers will rely on such a promise and that it cannot be performed.²⁴² However, there is no reason that a defendant ignorant of the plaintiff’s mark should know that its use of its trademark on its own goods would cause consumers to rely on others’ trademarks on those others’ goods in ways that those others will be unable to perform. Such reliance is a necessary basis for reverse confusion liability because the defendant in reverse confusion cases is in a position to perform the promise represented by the use of the trademark on the defendant’s own goods. Absent notice of the plaintiff’s trademark, then, the reverse confusion defendant has not made any conscious choice to cause anyone to rely on another producer’s promise that could not be performed, nor to make it more difficult for such producers to make or perform their own promises. A contractualist theory would therefore find no fault with the defendant’s conduct. Thus, unlike economic theory, trademark as promise can explain both why intent is not a necessary (though it might be a sufficient) condition for liability in forward confusion cases, and why it is a necessary condition for liability in reverse confusion cases.

Finally, the obligation to avoid intentionally creating false expectations through promising also helps to explain the puzzling remedial features of reverse confusion doctrine. As noted above, contractualism holds that intentionally making a false promise is a wrong that cannot be adequately remedied with compensation, because it implicates the promisee’s autonomy rather than their well-being.²⁴³ This aspect of trademark as promise helps explain why broad injunctions are the typical remedy in reverse confusion cases, despite the fact that economic theory and even background principles of trademark law would suggest either a damages remedy or a geographically limited injunction.²⁴⁴ Because it is the reverse confusion plaintiff’s saturation of the nationwide market that threatens to create mistaken reliance in the senior user’s customers, the geographically restricted injunctions provided for by the Tea Rose/Rectanus doctrine would fail to prevent the harm of that mistaken reliance—they are, from the contractualist point of view, an ineffective remedy. The broader injunctions we see in the reverse confusion cases are the only plausible way to prevent (as opposed to compensate for) the knowing creation of such mistaken reliance in reverse confusion cases. Conversely, the fact that contractualism accepts compensation as a second-best remedy only where specific performance of a knowingly false promise is impossible helps to explain the exceptions to the preference for broad injunctive relief in reverse confusion cases. For example, in Uber Promotions, Inc. v. Uber Technologies, Inc. and Stuart v. Collins,²⁴⁵ such broad injunctive relief was denied precisely because any goodwill the plaintiff might once have enjoyed had already been supplanted by the defendant’s independently created goodwill.²⁴⁶ In such circumstances, a nationwide injunction would not help restore the senior user to their former ability to use their trademark as a vehicle for promising, and second-best remedies such as compensation and cost-shifting are all that remains. And indeed, those are precisely the circumstances in which we see broad injunctive relief being denied, and such second-best remedies being awarded.

Conclusion

Reverse confusion doctrine poses a puzzle and a test for justificatory theories of trademark law. As I have shown, reverse confusion cases have a number of distinctive features that set them apart from ordinary trademark infringement cases. These distinctions challenge theorists to account for reverse confusion doctrine in ways that do not undermine their theories’ explanatory power in other areas. The standard economic theory of trademark law struggles to meet this challenge. In contrast, my own theory of trademark as promise, grounded in contractualist moral theory, faces far fewer difficulties meeting the challenge set by reverse confusion doctrine. In doing so, I have expanded the theory’s reach from an account of relationships between producers and consumers to include an account of relationships among producers. This result should persuade readers that trademark as promise is a theory worth taking seriously. More broadly, this demonstration provides further support to legal theories of private law that treat markets as normative institutions warranting analysis from diverse normative-theoretical perspectives.

Category: Print Issues

Reverse Confusion and the Justification of Trademark Protection

Share this article