Owning mark(et)s. - Free Online Library (2024)

Trademark owners regularly rely on claims that the defendant is"free riding" on their mark by making money using that mark,money the trademark owners say should belong to them. We analyze thosefree-riding claims and find them wanting. The empirical data shows thatdefendants in unrelated markets can benefit from using a well-knownmark, but that neither mark owners nor consumers suffer any injury fromthat use. A legal claim that a defendant is unjustly benefiting by usinga plaintiff's mark is hollow unless it is accompanied by a theoryof why that benefit should rightly belong to the plaintiff. And unlikereal property, or even other types of intellectual property, trademarklaw has no such theory. The result is that free-riding claims fall backon empty circularity. Yet these arguments are--explicitly orimplicitly--behind the most problematic expansions of trademark law inrecent years. We suggest that trademark law needs a theory of trademarkinjury that distinguishes harm to legitimate interests the law shouldprotect from a mere desire to capture a benefit enjoyed by another.

TABLE OF CONTENTSINTRODUCTION I. THE TRADITIONAL--AND THE IMPLICIT-CASE FOR TRADEMARK PROTECTION II. THE GROWING ROLE OF MARKET PREEMPTION AND FREE-RIDING ARGUMENTSIII. EVALUATING CLAIMS TO OWN MARK(ET)S A. Traditional Quality Feedback Arguments Lack Empirical Support B. Empirical Evidence Supports the Existence of a Benefit to Junior Users C. Evidence of Market Preemption Through Restricted Expansion D. Summary IV. EVALUATING CLAIMS TO OWN MARK(ET)S A. Claims of Producer Harm B. Free Riding 1. Preventing Consumer Confusion 2. Incentive Theory 3. Accession and Default Ownership 4. Natural Rights and the Property Instinct C. Spillovers V. TOWARD A "TRADEMARK INJURY" DOCTRINECONCLUSION: INSTINCTS AND EVIDENCE

INTRODUCTION

Black & Decker recently focused its marketing efforts for itsDEWALT line of power tools on Hispanic soccer fans, seeking to grow theDEWALT brand by "building and executing relevant cultural eventsand contests.'" In its El Tricolor Contest by DEWALT, forexample, Black & Decker gave soccer fans a chance to win "asigned Adidas Mexico jersey, $1,000 dollars in Eurosport/SOCCER.COM giftcards, DEWALT Compact Lithium-Ion combo tool kits, DEWALT Worksite RadioChargers and gear signed by Mexican Primera Division players." (2)Black & Decker also ran a variety of promotions linked to soccermatches, like the "Futbol con DeWalt" promotion in which Black& Decker gave two VIP tickets to the September 27, 2007, matchbetween DC United and Chivas to anyone who purchased $1,500 worth ofDEWALT products. The flyer Black & Decker distributed for thispromotion referred to the match as the 2007 DEWALT CUP, though this wasnot an official name] For Black & Decker, these promotions weresimply targeted marketing, using Hispanic soccer fans' interest inthe sport to draw attention to the new DEWALT products.

Soccer United Marketing ("SUM"), owner of commercial andmatch promotion rights for Major League Soccer, the US SoccerFederation, the Mexican Men's national team, and otherinternational teams such as Chivas and FCBarcelona, saw thingsdifferently. To SUM, Black & Decker was engaged in "ambushmarketing," wrongly profiting from SUM's trademarks. SUMtherefore filed suit. Though it nominally framed the issue as trademarkinfringement, SUM's complaint makes abundantly clear that its mainobjection to Black & Decker's promotions was that Black &Decker was interfering with SUM's ability to license its marksexclusively, in this case particularly because SUM had sold Makita anexclusive license to use the MLS and DC United marks on power tools. (4)Black & Decker, according to SUM, was confusing the public intobelieving that Black & Decker had a fight to utilize the names ofthe teams SUM represents, to depict the jerseys it was giving away, topromote and give away tickets to games in which SUM's teamsparticipated, or to set up product displays outside the stadiums inwhich those teams were playing. (5) On SUM's account, only thosewho pay money to SUM should be able to do these things.

Given the expansiveness of modem trademark law, SUM'sarguments are not obvious losers. SUM is arguing that consumers whoencounter Black & Decker's promotions are likely to be confusedabout whether there is some type of sponsorship relationship betweenBlack & Decker and SUM's soccer teams. If consumers areconfused in this way, of course, it's only because courts haveaccepted that sports teams should have such exclusive licensing rightsand have thereby created the very expectations they presume consumershave. (6) But even if consumers do wonder if Black & Decker hadpermission, what harm could possibly befall DC United if Black &Decker sells DEWALT tools in the parking lot outside one of its games?Consumers are extremely unlikely to attribute to the soccer team anydisappointment they have with DEWALT tools--indeed, they would be veryunlikely to do so even if they were explicitly told there was asponsorship relationship between DC United and DEWALT. (7) It'spossible consumers acquire some information about DEWALT tools fromtheir misperception of a relationship with one of the soccer teams, butwe find it hard to believe that information has anything to do with thequality of DEWALT tools. And if it doesn't, any confusion aboutsponsorship or affiliation is unlikely to affect consumers'decisions to purchase DEWALT tools and ought to be irrelevant totrademark law. (8)

Given the difficulty of articulating harm in conventional trademarkterms, it is not surprising that SUM's complaint makes very littleattempt to demonstrate that consumers will hold DC United responsiblefor the quality of DEWALT tools, or that Black & Decker'spromotions will harm consumers. Instead, the complaint is filled withaccusations that Black & Decker is free riding and wrongfullybenefitting from the soccer teams' goodwill, and that Black &Decker's actions impair SUM's ability to reap the benefits ofthe soccer teams' investments in their brands because "MLS andSUM can no longer guarantee that, by entering into a licensingagreement, [licensees] would have the exclusive use of the SUMMarks." (9)

SUM's free-riding argument is based on the assertion that theowner of a mark in one context should have the right to controlancillary uses of that mark in other contexts because the later usersare mere free riders, reaping what they have not sown. These claims aresometimes justified on the theory that failure to enforce rights againstjunior, noncompeting users would impede the senior user's abilityto enter ancillary markets itself. In other words, trademark ownersargue they should be able to preempt uses of a mark even in markets inwhich they do not currently participate, against the possibility thatthey will want to use the mark in that market in the future. (10) Thisclaim is contrary to trademark law's long-held maxim that trademarkrights are not rights in gross. (11) But ultimately, the free-ridingclaims are even more sweeping than ownership of marks: trademark ownerssometimes are effectively asserting the right to own markets themselvesbecause, as in SUM's case, the relevant market owes its origin totheir brands.

Free-riding and market preemption arguments are not new intrademark law. But those arguments generally have been tacked on toprimary arguments regarding the potential that a later, noncompetitiveuse might negatively impact the senior mark owner's reputation forquality. And because those primary arguments have been so widelyaccepted, neither courts nor scholars have expended any serious energyevaluating the market preemption or free-riding arguments. That has beena mistake. First, we believe the sentiments behind the market preemptionand free-riding arguments have actually motivated courts to imposeliability in a number of questionable cases. Second, as we demonstratedin our prior work, the empirical case for the reputational dilutionclaim is much weaker than one might expect, (12) and the claim thatconsumers are injured by the defendant's use of a mark in anunrelated market is implausible except under specialized circ*mstances,circ*mstances that trademark plaintiffs should have to prove. (13) Bycontrast, the empirical evidence confirms both that third parties canbenefit from uses of known marks in markets ancillary to the senior markowner's and that those third-party uses can impair the senioruser's ability to expand its own product lines. (14) Put anotherway, the evidence suggests that third parties like Black & Deckermight benefit from use of, or proximity to, SUM's trademarks, butnot that SUM is harmed by such use. The result is a puzzle for the law:what, if anything, should trademark law do to prevent practices thatbenefit the defendant but do not harm the plaintiff?

In this Article, we confront the market preemption and free-tidingarguments directly. We believe these arguments actually departfundamentally from the traditional bases of trademark law and theory,and in ways that could prove quite troubling in a competitive economy.The claim that trademark owners are injured by not being able to controluse in a remote market is ultimately a circular claim--mark owners areinjured if, but only if, we define their trademark rights ex ante toinclude control over that remote market. The arguments in favor of doingso, however, turn out to be remarkably weak, and we argue that they donot justify expanding trademark law. If that means that others free rideon the effort a trademark owner puts into building a brand or making amarket, so be it. Free riding itself is bad only if we think there issome reason to prevent it. We have such a reason, generally, in thepatent and copyright contexts: we affirmatively want to encourage thecreation of new works or inventions, and we're concerned that freeriding might undermine incentive to create them. But we don't haveany reason to affirmatively encourage the creation of new brands. Infact, absent confusion regarding responsibility for the quality of thegoods at issue, free riding on brands is likely a good thing, not a badone, and in any event isn't the sort of harm trademark law isparticularly good at addressing. The relevant issue should in fact beharm to consumers or to the market as a whole, not simply the benefit amark confers on others.

As a result, we propose a "trademark injury" requirementakin to the "antitrust injury" requirement currently used toweed out undeserving antitrust plaintiffs: to sustain an infringementclaim, a plaintiff should have to demonstrate that the defendant'sconduct is likely to cause material confusion in the minds of consumers,and allegations of other types of "harm" should beinsufficient.

If we're right, our argument has implications for a number ofcurrent trademark doctrines in which the market preemption andfree-riding arguments have come to the fore: broad sponsorship oraffiliation claims, including those in the merchandising context and inthe context of expressive works; initial interest and post-saleconfusion; and dilution. But it also has broader implications for theway trademark cases are argued and reasoned. The anti-free-ridingimpulse can corrupt even cases ostensibly decided on more traditionaltrademark grounds. As a result, courts must be particularly vigilant toavoid finding confusion in unlikely circ*mstances because of the pull offree-tiding concerns.

In Part I, we explain the market preemption and free-ridingarguments and how they differ from the traditional justifications fortrademark law. In Part II, we explain why those arguments are important,and why they are increasingly driving results in trademark cases. InPart III, we evaluate those arguments on the merits, finding themwanting. In Part IV, we draw some broader implications from thoseevaluations, and in Part V, we suggest that a trademark plaintiff shouldhave to show "trademark injury" in order to prevail.

I. THE TRADITIONAL--AND THE IMPLICIT--CASE FOR TRADEMARK PROTECTION

The arguments for trademark protection in the core cases--thoseinvolving directly competing goods--are fairly straightforward and notterribly controversial. Trademark law prevents parties from using a markthat is likely to confuse consumers about the source of their goods. Inthe context of competing goods, this protects mark owners from divertedtrade and consumers from making mistaken purchases. When a secondcompany comes along and sells Apple computers, that company cancapitalize on consumers' confusion about the source of the newproducts to divert to itself customers who are familiar with Appleproducts and who are trying to patronize the same Apple company fromwhich they've bought computers before. Apple, Inc. is harmeddirectly by this confusion because it loses sales it otherwise wouldhave made, and indirectly because consumers are likely to loseconfidence that they can rely on the Apple mark to distinguish desirablefrom undesirable products. Consumers are harmed here too--directlybecause they bought something different than they expected, andindirectly because they won't get to rely on the Apple mark in thefuture to tell them which computer products to buy. Hence, mark ownerand consumer interests converge to support trademark protection in thesecases of competitive goods. (15) Trademark law protects the integrity ofthe marketplace by preventing deceptions that change the way consumersbuy goods.

But beyond cases of directly competing goods, the arguments get alittle more complicated. The predominant producer-side arguments infavor of protection against noncompeting uses have focused on thepossibility of negative feedback to the producer's original market.Specifically, the claim has been that consumers who believe their Applewatches come from the same company that produced their iPod are likelyto direct any dissatisfaction with their watches at the iPod-producingApple, leading them to question whether iPods are really that good afterall. The worry is that consumers' quality experience in the newmarket "feeds back" to the original market. We addressed thatargument in our prior paper, surveying the marketing literature andfinding surprisingly little evidence that consumers punished the corebrand for what they saw as poor quality extensions into new markets.(16)

Two other arguments have frequently been offered in addition to thequality feedback arguments, and those additional arguments are ourprimary focus here. First is the argument that, regardless of whether amark owner is held responsible for negative experiences with a junioruser's products, the mark owner is harmed by a noncompeting usebecause that use might impede the mark owner's ability toexpand---either into the same market as the junior use or other relatedmarkets. This argument is not about the present harm to the producerfrom third-party uses, but rather the potential future benefitsforeclosed by those uses. Thus, for example, even if Apple, Inc. is notdirectly affected by another party's use of the Apple mark forwatches because consumers do not hold Apple, Inc. responsible for thequality of Apple watches, Apple, Inc. could still claim to be harmedbecause it would be unable to expand into the watch market using theApple mark.

The notion that a mark owner is preempted by a junior user fromentering another market operates on a background assumption that thesenior user ought to have a superior right to enter other markets under"its" mark. Specifically in the context of the hypotheticalabove, this argument assumes that any value the Apple mark has in thewatch market rightfully belongs to Apple and no one else. (17) It is theloss of this market opportunity that qualifies as "harm" toApple. This is a market preemption argument analogous to the derivativeworks right in copyright. (18) The claim is that the trademark holderowns not just the right to use a particular mark in connection withparticular goods, but the mark itself, wherever it has value.

A related theory of harm from preemption might focus on the costs aproducer would be required to incur in re-educating consumers about thequality of its goods if and when it entered the same market as thejunior user. If, for example, Borden--the maker of condensed milk--wasunable to prevent another company from using the Borden name for icecream (19) and was forced to enter the ice cream market under adifferent name (since the other company would have established priorityin the ice cream market), it would lose the benefits of name recognitionamong consumers who bought both ice cream and condensed milk. Here toothe claimed injury is that the defendant's use interferes with theplaintiff's expansion, as this "harm" could come tofruition only if the senior user in fact expands. (20)

The market preemption claim is based on a theory of harm to thetrademark owner. Because the harm is bound up with the trademarkowner's anticipated entry into another market, the plausibility ofthe claim is a function of how likely that entry is. That is afact-specific question; it seems reasonable to conclude that Bordenwould move from making condensed milk to making ice cream, but perhapsless reasonable that Apple will go into the watch-making business.Notably, however, when trademark owners argue market preemption theyrarely tie it to concrete evidence of planned entry and likely injury.(21) Instead, what is nominally a factual question about injury to thetrademark owner is often simply presumed as injury: I must have beenhurt because I no longer have an opportunity I once had.

In fact, however, the preemption "injury" also presumesthat the trademark owner had a legal fight to exclusive use of the markin that secondary market. Even if the mark owner can show actualpreparations to expand, there is a deeper question about whether--andwhy--they should have a superior right to use the mark in the ancillarymarket. (22) True, in many of the cases in which the mark owner is infact planning to expand, the goods are sufficiently related thatconsumers are likely to believe the mark owner is responsible for theirquality. And if consumers do believe that, the mark owner willeffectively be protected in their ability to expand under a theory ofconsumer harm. (23) But if the goods are not so closely related thatconsumers will draw inferences about product quality from the use of amark, it is far less clear that the mere fact that a mark owner wasplanning to enter a distant market should give it the right to do so.Indeed, in the antitrust context, courts have rejected a very similarargument that firms suffer cognizable injury when excluded from a marketwhich they are not currently prepared to enter, but hope to enter in thefuture. (24)

The second argument advanced in support of a trademark owner'sright to control the use of the mark in unrelated markets is afree-riding argument that is at least nominally distinct from marketpreemption. On this theory, a third party that uses Apple for watchessimply free tides on the goodwill developed by Apple, Inc., and thirdpatties should not be allowed to claim for themselves any value theApple mark has in the watch market. This argument appears to be distinctfrom the market preemption argument in that the objection to free ridingdoes not depend on a claim that the Apple watch company harms Apple,Inc., but simply reflects the belief that Apple ought to own whatevervalue the Apple mark has in the watch market. While the marketpreemption argument says that a mark owner will eventually be harmed bythe defendant's conduct, the free-riding argument claims insteadthat the defendant will benefit from the plaintiff's conduct, andthat any such benefit is itself unjust and ought to be paid as awindfall to the trademark owner.

But these claims of harm and benefit ultimately run together,particularly in intellectual property ("IP"), where the entireconcept of harm is in some sense an artificial construct based on thegovernment's decision to create a right. Trademark owners mightclaim a harm from free riding by using the circular argument that ifsomeone benefits from the use of a mark they own, that benefit belongsto them, and therefore they have been injured by not being paid alicense fee for the right to authorize that use. A similarly circularargument seems to have carried the day in copyright, undermining thefair use doctrine. (25) Despite the ability to turn any benefit to athird party into a claim of harm to oneself, we think it is important tounderstand that claims that the trademark owner is injured by marketpreemption are ultimately claims that are based on the asserted abilityto capture benefits from others. The idea that a mark owner is harmedbecause a defendant interferes with its ability to expand operates on apresumption that the mark owner ought to have the fight to expandwithout interference. It's not a harm to the mark owner tointerfere with expansion if we don't define the right ex ante toinclude the right to expand.

Thus, market preemption arguments that are not based on actualconsumer harm are in essence claims of free riding. Both are premised onthe idea that the law should give a particular trademark owner a rightto control someone else's use of the mark even if consumerdecisions aren't affected, on the theory a mark owner has a claimto any value derived from use of that mark. The next Part explains whythese arguments are important, and why they need to be addresseddirectly.

II. THE GROWING ROLE OF MARKET PREEMPTION AND FREE-RIDING ARGUMENTS

The most obvious reason to take market preemption and free-ridingarguments seriously is that those arguments are, in fact, actuallymotivating courts' decisions in a variety of cases, particularlythose at the edges of trademark protection.

First, market preemption and free-riding arguments have playedprominent roles in courts' expansion of trademark law to encompassnoncompeting goods. In the late nineteenth and early twentieth century,courts allowed mark owners to assert claims only against directcompetitors, and they routinely denied claims when the parties'goods were even modestly different. (26) But courts gradually began totake a more liberal view and allowed claims against uses of a mark forgoods that were only related to those of the senior user. In Aunt JemimaMills Co. v. Rigney & Co., (27) one of the earliest cases to enforcetrademark rights against a noncompetitor, the court allowed the owner ofthe AUNT JEMIMA mark for syrup to assert its rights against a later userof AUNT JEMIMA for flour. (28) Though flour and syrup are clearly notthe same goods, so "no one wanting syrup [can] be made to takeflour," the court believed the products were "so related as tofall within the mischief which equity should prevent." (29)Confusion about the source of the syrup would put Aunt Jemima'sreputation in the defendant's hands and "will enable [thedefendant] to get the benefit of the complainant's reputation andadvertisem*nt." (30) Likewise, in Stork Restaurant, Inc. v. Sahati,(31) the court allowed the owner of a New York club called The StorkClub to assert its rights against a San Francisco club of the same name,even though such geographically remote clubs clearly do not compete witheach other. (32) The court was impressed with the value of the StorkRestaurant mark which it thought "wholly adventitious, [and]brought about by continued, expensive, and spectacular advertising--suchas the giving away of one thousand dollar bills." (33) In light ofthis value, the court thought "[t]he conclusion [was] inescapablethat the appellees [were] seeking to capitalize on the publicity thatthe appellant ha[d] built around the name." (34)

In Precision Tune, Inc. v. Tune-A-Car, Inc., (35) the courtcharacterized the defendant's use of a confusingly similar mark ina different geographic market as "depriv[ing] [the plaintiff] of anopportunity to expand its market." (36) As long as the defendant"continue[d] to employ the deceptively similar marks and tradedress, [the mark owner could not] attempt to open a franchise because it[could not] guarantee its franchisee's exclusive use of themark." (37) And in Scarves by Vera, Inc. v. Todo Imports Ltd., (38)the court found that the defendant's use of Vera for cosmetics andtoiletries infringed the plaintiff's rights in the same mark, whichit had used for women's scarves, sportswear, and linens. (39) In sodoing, the court emphasized the plaintiff's interest in being ableto enter a related field at some future time. (40)

We think the influence of the free-riding and market preemptionarguments also explains the merchandising cases. These cases involveuses of the names or logos of professional sports franchises oruniversities to adorn clothing or other merchandise. They straintraditional trademark principles because, at least at the time the casesfirst arose in the 1970s and 1980s, (41) there was little reason tothink consumers believed the franchises or universities were the sourcesof merchandise simply because it depicted their marks. In spite of theweakness of the confusion arguments, courts in some cases foundinfringement because they were moved by their belief that the defendantswere mere free riders. In Boston Professional Hockey Association v.Dallas Cap & Emblem Manufacturing, Inc., (42) for example, the firstcase to expand trademark rights to include merchandising, the courtfound the conclusion "inescapable that, without plaintiffs'marks, defendant would not have a market for his particular productamong ice hockey fans desiring to purchase emblems embroidered with thesymbols of their favorite teams." (43) And the case that gave theBoston Athletic Association exclusive control over Boston Marathont-shirts was even clearer about its rationale: "Defendants'shirts are clearly designed to take advantage of the Boston Marathon andto benefit from the good will associated with its promotion byplaintiffs. Defendants thus obtain a 'free ride' atplaintiffs' expense."

The "right" to control merchandise that includes a schoolor team name--a right unheard of before 1975--has since expanded apace,to the extent that you can now violate a university's trademarkrights without using its name or logo at all, merely by selling t-shirtsin the colors of the school. (45) The culprit again is theanti-free-riding impulse: "Smack's alleged competitivedisadvantage in the ability to sell game day apparel relates solely toan inability to take advantage of the Universities' reputation....This is not an advantage to which it is entitled under the rubric oflegitimate competition." (46)

Something similar seems to be driving the results in a third set oftrademark cases, those involving products based on copyrighted works.Authors of books, comics, and movies create characters and worlds. Theyare entitled to copyright in the works they create, including charactersthat are sufficiently well delineated. (47) But they have also managedto persuade courts that they should have trademark rights not just intitles or brands, but in those same characters and other objects thatappear within their creative works. The argument is again a free-ridingargument--that no one should be able to sell products or write booksthat evoke Wolverine or James Bond, not because doing so hurts thecreator, but because it makes money for the user, and that money shouldbelong to the creator. In Warner Bros., Inc. v. Gay Toys, Inc., (48) thecourt justified a copyright owner's trademark control over toyreplicas of the General Lee, an automobile featured in Warner'stelevision series The Dukes of Hazzard, by claiming that "deny[ing]Warner Bros. injunctive relief would.., enable Gay Toys 'to reapwhere [i]t bald] not sown.'" (49) And DC Comics has prevailedin a number of cases against writers who featured similar superheroes.(50) Notably, the courts in those cases found it unnecessary to engagein classical likelihood-of-confusion analysis, instead resting theirliability findings on evidence of similarity between the characters.(51)

Courts have reached similar results in replica cases outside thecopyright context. Vehicle manufacturers have sued companies that maketoy replicas of the vehicles in which their manufacturers claimtrade-dress rights. In General Motors Corp. v. Lanard Toys, Inc., (52)for example, GM sued a company that sold toy military vehicles thatresembled the Humvee military vehicle, claiming the toy vehicleinfringed its trade-dress rights in the design of the real Humvee. (53)Though the markets for toy military vehicles and Humvees obviously arequite distinct, the court found the "relatedness of the goods andservices" factor to favor General Motors, suggesting "the toycar is quite closely related to the actual car on which the registeredtrademark of the grille is found." (54) Influenced by thatdetermination and its conclusion that Lanard had bad intent because"there [was] undisputed evidence that the design of the frontgrille of Lanard's toys was copied directly from the Hummervehicle," (55) the court concluded that Lanard's toy waslikely to cause confusion and enjoined the rise. (56) GM itself nevermade replica Humvees, nor had it licensed anyone else to do so. (57) Andit's hard to imagine how toy vehicles in any way harm GM'ssales of real military vehicles. GM's concern was thatLanard's toy Humvees would interfere with any future licensing ofthe Humvee design to toy makers. Perhaps inspired by GM's success,Paccar, Inc. recently filed a similar suit against the maker of replica,scaled model trucks that resembled Paccar's KENWORTH and PETERBILTtrucks. (58) Paccar claimed that the defendant's replica trucksharmed it because the replicas competed with scaled model trucksproduced under license from Paccar. (59)

Courts have understood in these cases, as in the merchandisingcases, that substantial value was at stake, and their sense that thevalue belonged to the mark owner clearly influenced their decisions. Andthe same is true of many commentators. Indeed, in arguing against theso-called "trademark use doctrine," Dinwoodie and Janisexpress concern that reading the "'use as a mark'requirement strictly as incorporating the notion of the mark as a'source-identifier[]'... might undermine the multi-billiondollar industry of brand merchandising and product design." (60)

Fourth, we see evidence of free-tiding concerns in keywordadvertising cases. Competitors commonly buy from search engines theright to have their advertisem*nts triggered by search queries thatinclude others' trademarks (a process known as "keywordadvertising"). While trademark owners sometimes articulate anargument that advertisers are confusing consumers by placing ads thatappear opposite search results, a growing number of cases allege thatthe direct infringer is not the advertiser but the search engine itself.(61) No one would think that an internet search engine necessarilysponsors or endorses the whatever products are found in the search,which is probably why the plaintiffs in these cases do not even allegeconfusion about the source of the search engines' services. (62)Nonetheless, courts have permitted claims against search engines toproceed on the theory that the search engine is making money in a waythat involves some internal use of a trademark. (63)

Fifth, we see the influence of the free-riding impulse ininitial-interest and post-sale confusion cases. Mark owners claiminitial-interest confusion in cases where the defendant does somethingthat draws attention to itself in a way that might momentarily confuseconsumers, but where any such confusion is likely to be dispelled beforeconsumers make any purchasing decisions. Precisely becauseinitial-interest confusion is temporally disconnected from purchasingdecisions, these cases are difficult to justify with ordinaryconfusion-based arguments. Hence, not surprisingly, courts tend to fallback on epithets here, too. In Brookfield Communications Inc. v. WestCoast Entertainment Corp., (64) for example, the court concluded thedefendant's use of "moviebuff' in the metatags for itswebsite violated the plaintiff's rights in that mark because searchengines used the metatags to generate search results in which thedefendant's site appeared higher in the rankings than theplaintiff's. (65) While the court conceded that confusion wasunlikely, (66) it believed consumers, now presented with both websitesin response to a search employing "moviebuff' as a searchterm, might choose the defendant's website rather than theplaintiff's:

 Although there is no source confusion in the sense that consumers know they are patronizing West Coast rather than Brookfield, there is nevertheless initial interest confusion in the sense that, by using "moviebuff.com" or "MovieBuff" to divert people looking for "MovieBuff" to its web site, West Coast improperly benefits from the goodwill that Brookfield developed in its mark. (67)

Similarly in Mobil Oil Corp. v. Pegasus Petroleum Corp., (68) thecourt found that Pegasus Petroleum infringed Mobil Oil's flyinghorse (a Pegasus) by adopting the name Pegasus Petroleum for its oiltrading company. (69) The Second Circuit was moved by its belief that anoil trader "might listen to a cold phone call from PegasusPetroleum ... when otherwise he might not, because of the possibilitythat Pegasus Petroleum is related to Mobil." (70) And in ElvisPresley Enterprises v. Capece, (71) the court noted that"initial-interest confusion is beneficial to the Defendants becauseit brings patrons in the door" which "is even more significantbecause the Defendants' bar sometimes charges a cover charge forentry, which allows the Defendants to benefit from initial-interestconfusion before it can be dissipated by entry into the bar." (72)

In the post-sale confusion cases, courts have found use of a markinfringing even when there is no evidence consumers of thedefendant's goods are likely to be confused at the time ofpurchase. Mark owners in these cases claim, and courts have accepted,that it is enough if others who come into contact with the goods afterpurchase might be confused. Courts in these cases do not demand proofthe allegedly confused nonpurchasers are ever likely to be consumers ofthe mark owner's goods, nor do they demand evidence the allegedconfusion away from the point of sale would actually affect purchasingdecisions. They don't demand this evidence, in our view, becausethese are really cases about free riding, not confusion.

In the earliest post-sale confusion case, Mastercrafters Clock& Radio Co. v. Vacheron & Constantin-Le Coultre Watches, Inc.,(73) Judge Frank was concerned that:

 [A]t least ... some customers would buy [the copier's] cheaper clock for the purpose of acquiring the prestige gained by displaying what many visitors at the customers' homes would regard as a prestigious article. [The copier's] wrong thus consisted of the fact that such a visitor would be likely to assume that the clock was an Atmos clock. (74)

As Professor McCarthy has noted, the real concern of these cases isthat "consumers could acquire the prestige value of the senioruser's product by buying the copier's cheap imitation."(75)

More recently, the Ninth Circuit made abundantly clear what it sawas the basis for the post-sale confusion doctrine: "Post-purchaseconfusion creates a free-rider problem." (76) And in the NinthCircuit's view, free riding is such a problem that it trumps thefirst-sale doctrine, collapsing yet another trademark"defense" into the prima facie case. (77) Hence, inAu-Tomotive Gold the court found the defendant's conduct infringingwhen it purchased genuine Volkswagen badges from an authorizedVolkswagen dealer and used the badges to adorn license plate frames.(78) It found infringement even though no purchasers of the licenseplate frames were confused about the source of the frames and Auto Goldpackaged the frames with labels making clear that the frames were notproduced or sponsored by Volkswagen. (79) The concern was simply that"a person on the street who sees an Auto Gold marquee license platewith a VW badge will associate the plate with Volkswagen," because"customers buy marquee license plates principally to demonstrate tothe general public an association with Volkswagen." (80) Thispost-sale "association" was enough for the court to findinfringement: "If the producer purchases such a trademarked productand uses that product to create post-purchase confusion as to the sourceof a new product, the producer is free riding even though it has paidfor the trademarked product." (81)

This result is particularly problematic because it takes the ideaof owning markets to another level. (82) In the noncompeting-goodscontext, mark owners are arguing that they should have the exclusiveright to use a mark in the ancillary markets. They claim, in otherwords, to own the right to use the mark in any market in which that markhas value. But in Au-Tomotive Gold, as in the merchandising cases, thecourt allowed a mark owner to use trademark law to capture an actualmarket, as Volkswagen owners were not likely to be satisfied withaccessories bearing a Toyota logo or without any logo at all. (83)Volkswagen owners want Volkswagen accessories for the same reasonfarmers with green farm equipment want other green farm equipment (84)and owners of baroque-style silverware want other baroque silverware.(85) Yet the court was unconcerned about giving Volkswagen control overthe market for Volkswagen accessories: "It may be true that AutoGold's activities serve to reduce the price paid by consumers formarquee plates. But trademark law protects trademark holders from thecompetition that results from trademark infringement, irrespective ofits effect on prices." (86) And in the Ninth Circuit, at least, itapparently prevents competition even by those who are reselling goodsthey have already bought from the trademark owner.

Finally, theories of market preemption also seem to underlie thedoctrine of trademark dilution. (87) One of us has argued that, properlyconstrued, dilution is based on a theory of consumer harm. (88) Butthere is no question that much of the rhetoric used by both courts andcommentators to justify dilution has focused on the idea of a trademarkas a property right that confers control over noncompeting uses whetheror not consumers are hurt. (89) Indeed, the Supreme Court--in the courseof restricting dilution law--said that "[u]nlike traditionalinfringement law, the prohibitions against trademark dilution ... arenot motivated by an interest in protecting consumers." (90) AndJudge Posner has even suggested that free riding itself can beconsidered a form of dilution. (91)

Several cases that might be regarded as dilution-by-tarnishmentcases depend in significant part on the courts' perceptions thatthe defendants were making illegitimate use of the plaintiff'smarks as "triggering mechanisms" in order to sell their owngoods. In Chemical Corp. of America v. Anheuser-Busch, Inc., (92) forexample, the court found the defendant's use of the slogan"Where there's life, there's bugs" to infringe theplaintiff's slogan "Where there's life, there'sBud" because of the defendant's "brazen ... effort ... tocapitalize on the good will created by ... the plaintiff" and the"peculiarly unwholesome association" the slogan created withBudweiser beer. (93) Anheuser-Busch also won a tarnishment claim againstthe Buttwiper brand of dog toys. (94) Not surprisingly, the success ofthese "association" claims has encouraged others, like TheNorth Face's recent lawsuit against a company called "TheSouth Butt," which makes clothing featuring, well, a butt. (95)

And it is not just in the United States that courts strive toprevent free riding even absent a theory of harm to the plaintiff. ACanadian court interpreted a provision of the Canadian Trade-Marks Act(96) to prevent truthful comparative advertising on the theory that thatadvertising enabled the defendant to "free ride" on theplaintiff's brand investment. (97) And more recently, the EuropeanCourt of Justice declared that European Union law prevents companiesfrom taking advantage of the prestige of well-known marks in order toboost sales, even if the copycat products do not directly harm the ownerof the mark or create a likelihood of confusion. (98) In that case,L'Oreal objected to Bellure NV, a Belgian perfume manufacturer,advertising and selling perfume that imitated L'Oreal products butobviously did not come from L'Oreal. In particular, L'Orealwas upset with the packaging of Bellure's perfume, which wassimilar to that of some of L'Oreal's famous fragrance brands,including Tresor and Miracle. L'Oreal also found Bellure'sadvertising unfair, as Bellure advertised its scents on a list alongsideL'Oreal's scent names. The court found that, despite the lackof any consumer confusion, Bellure took advantage of the distinctivecharacter of L'Oreal's marks in order to benefit from theprestige and power of attraction of those marks and to exploit themarketing efforts expended by L'Oreal. (99) The ruling byEurope's highest court opens the door for brand owners to takerenewed action against the producers of look-alike products andrepresents a major shift in European IP law, which has typicallysupported the legitimacy of comparative advertising. (100)

We should be clear that our objection is not to any of thesevarious doctrines per se. Trademark owners may--or may not--be able toshow that consumers are injured by sports merchandise sold by amanufacturer unaffiliated with a team, by Google selling ad spaceopposite search results, by web sites momentarily divertingsurfers' attention with a URL, or by homeless people wearingknock-off Rolexes. (101) Similarly, trademark owners may or may not beable to show that consumers are actually made worse off by the blurringor tarnishment of marks. What is notable is that each of these doctrinesare extensions of basic trademark law into areas in which the harm toconsumers is questionable at best. (102) And yet courts don't seemto be asking the hard questions about that putative consumer harm. Themarket preemption and free-riding arguments have served to distractattention from the question of whether consumers are in fact confused totheir detriment-or, in the case of dilution, whether they are otherwiseharmed--by such uses. In so doing, those arguments have contributed tothe expansion of trademark law beyond its traditional conceptualmoorings.

III. EVALUATING CLAIMS TO OWN MARK(ET)S

Notwithstanding their strong emotional appeal, free-riding claimsactually run counter to what has long been a fundamental principle oftrademark law: its grant, not of rights in a mark "in gross,"but of rights tailored to the actual use made of a mark by its owner.Trademark owners do not own words, at least not traditionally. (103)Rather, they own the right to prevent the use of certain similar wordsin certain places in connection with the sale of certain goods in a waythat is likely to confuse consumers. (104) Market preemption claimspresuppose a fight to control use outside those boundaries. While somesuch claims--like the Borden case--can be justified on consumer harmgrounds, most cannot. They therefore stretch the traditional account oftrademark theory. And if market preemption claims stretch trademarktheory, flee-riding claims discard that theory altogether, claimingtrademarks are precisely what trademark law and theory have longmaintained they aren't: rights in gross over a mark itself.

The fact that these arguments run counter to tradition is not, ofcourse, sufficient reason to reject them. But it does mean that thearguments are worthy of close scrutiny--scrutiny that has so far beenlacking. For if we are persuaded that market preemption and free ridingare appropriate considerations in shaping the scope of trademark law, wemay well find ourselves with a radically different trademarklaw--indeed, one that would have to change even more than it alreadyhas. Specifically, if we are to take seriously the logic of marketpreemption and free-riding arguments, trademark law need not retain thelikelihood-of-confusion determination as its central feature. (105)Market preemption arguments are essentially about allocation of markets.That allocation need not have anything to do with whether consumers areconfused. It is probably no wonder, then, that we see these argumentscarrying the most weight precisely in those cases in which the consumerconfusion arguments are weakest.

We believe the time has come for these arguments to be givenserious attention in the trademark context. In this Part, we examine theempirical evidence behind both the market preemption and the free-ridingclaims.

A. Traditional Quality Feedback Arguments Lack Empirical Support(106)

The empirical case for the argument most frequently offered infavor of broad trademark protection--the quality feedback argument--ismuch weaker than most suppose. Research regarding brand extensionssuggests that consumers generally do not alter their global evaluationsof brands (i.e., their assessments of the brand's quality) whenthey encounter negative information about related products offered underthe same mark. (107) Even negative personal experiences with a productare likely to have an effect only for closely related products, and eventhen only for family branded rather than sub-branded extensions. (108)Moreover, even in the context of successive extensions the only apparentrisk to a core brand from a failed extension is that consumers willevaluate future extensions more negatively than they otherwise mighthave. (109) And even when negative information does affect a parentbrand, it does so only in an abstract sense; it does not impact brandimage for the parent brand in the context of the goods the parentpreviously offered. (110)

Extension information is also unlikely to negatively impactspecific brand beliefs, even when the extension is incongruent withthose brand beliefs. (111) And just as with global brand beliefs, anyimpact an extension has on specific brand beliefs is likely limited tothe parent brand generally. Stated differently, an extension has littleor no impact on the brand in the context of particular products. Thus,even if an extension affects consumers' general view of theNeutrogena brand as "mild," it is unlikely to affect theirbelief that Neutrogena hand lotion is mild.

This is true, generally speaking, regardless of how familiar thebrand is to consumers. Consumers are less motivated to processinformation about less familiar brands, and unmotivated consumers tendto use a "sub-typing" strategy, storing information about theatypical extension in a separate cognitive category. (112) As a result,relatively unfamiliar brands are likely to be insulated from feedbackfrom incongruent extensions.

But this point appears generalizable: after reviewing the relevantliterature to distill "main tendencies," Sjodin and Tornconclude that negative evaluation of incongruent extension informationwill not affect evaluation of the parent brand. (113) The authorsexplain this somewhat counterintuitive result by suggesting consumersgenerally use a sub-typing strategy to resolve incongruous information.(114) But whether or not sub-typing explains the lack of impact onfamiliar brands in addition to unfamiliar ones, Sjodin and Tom'sconclusion that incongruous information is unlikely to impact parentbrands fits comfortably with other research demonstrating that thatconsumer perceptions of well-known brands are quite resistant to change.(115) It therefore seems unlikely that incongruence negatively impactseven well-known brands. This likely explains why numerous cultural iconsthat lack any trademark protection and are therefore free for use--UncleSam, the Eiffel Tower, the Statue of Liberty, the Mona Lisa, MountRushmore, the works of Shakespeare, and the characters Frankenstein (andhis monster), Dracula, Scrooge, and King Arthur--nonetheless seemimpervious to either blurring or tarnishment from incongruous uses.(116)

Moreover, even if incongruous uses did impact a few moderatelywell-known brands--those for which consumers would likely be highlymotivated to process information but which are not well known enough tobe impervious to change--consumers with high levels of involvement areless likely to be confused by uses of a similar mark in the first place,or at least are likely to be confused by different types of similaritythan are consumers in low-involvement situations. According to Howard,Kerin, and Gengler, consumers in high-involvement situations areunlikely to be confused by uses of a mark that is similar to the coremark in sight or sound, though relatively more likely to be confused byuses of marks that are similar in meaning. (117) Together these lines ofresearch suggest both that consumers are unlikely to be confused bysimilarity in circ*mstances of high involvement, (118) in which harm ismost likely to follow from confusion, and that confusion is unlikely tobe harmful in those cases where confusion is more likely.

Finally, any effect on consumers' specific brand beliefs isunlikely to matter to purchasing decisions because evaluations ofspecific brand beliefs generally don't impact consumers'decisions: consumers evaluating a new product tend to rely on globalattitudes towards a brand rather than attempting to recall and processspecific brand attributes.

The argument that mark owners are harmed by reputational feedbackin the context of pure sponsorship or affiliation relationships is evenweaker. (119) The most relevant information here comes from studies ofnegative spillover from information about brand alliance partners. (120)In one such study, Votolato and Unnava focused on the consequences to apartner of negative information about a supplier or a celebrity endorserof the partner's products. (121) Specifically, the authors soughtto measure the effect of information that a fictitious clothingcompany's partners had behaved immorally or had been incompetent onconsumers' attitudes toward the company. (122) There wasn'tany. The lesson is simple, if remarkable: negative information does nothave any feedback effect on the partner absent some additionalinformation about the partner's culpability for the failing. Thisis true regardless of whether the information relates to competence or amoral failing and regardless of whether the information is about anothercompany or a person with which the partner is associated. The authorsnote:

 [A] host brand may generally be quite impervious to negative publicity surrounding its partner brand; the host brand [in the study] was only affected when participants were led to believe that the host knew of and condoned the partner's behavior. Spillover from the partner brand to the host brand did not occur unless this condition was present. (123)

We think these findings cast serious doubt on the assumption thatconsumers will hold a mark owner responsible for the quality ofunrelated goods beating the same mark even if they are confused aboutthe mark owner's relationship with those goods. As a result, theycall into serious question the first-line argument that trademark ownersare harmed by a defendant's use of the same mark in another market.

B. Empirical Evidence Supports the Existence of a Benefit to JuniorUsers

On the other hand, the available empirical evidence suggests thatjunior users can benefit significantly from use of a known mark, andthat junior uses can sometimes impede a mark owner's own expansion.In other words, known marks do sometimes add value to new products orservices, and third-party uses do sometimes affect expansion.

One branch of marketing literature attempts to study the conditionsunder which information about the existing ("parent") brandtransfers to a new product offered under that brand, thereby reducingthe costs of entry and increasing the likelihood that consumers willaccept the new product or service. The effects these studies attempt tomeasure are spillover effects--that is, they focus on the potentialbenefits accruing to a new product offered under a known brand namerather than the effects on the brand name in its original context. Thesestudies therefore measure the expansion potential of a brand and speakmost directly to the arguments regarding free riding, as they seek toidentify when the junior use will benefit from goodwill associated withthe senior use.

According to this literature, the success of any particularextension is primarily a function of three factors: (1) the perceivedquality of the core brand; (2) the similarity or "fit" of theproposed extension with the family (or core) brand; and (3) theperceived credibility of the family brand. (124) Consumers favorablyevaluate extension products when the core brand is high quality and theyperceive the extension product as a good fit with existing products.(125) Conversely, when an extension is not perceived as a good fit, thegoodwill associated with the core brand will not lead to favorableevaluations of the extension products. (126)

Fit in this context can be measured in terms of complementarity,substitutability, and transferability. Complementarity refers to theextent to which consumers view the extension product as a complement tothe goods offered by the parent brand. Products are complementary ifboth are consumed jointly to satisfy some particular need. (127) Forexample, ski clothing is a complementary extension for the ROSSIGNOLbrand, which originally was known for downhill skis. (128)Substitutability refers to the extent to which consumers view theextension product as a substitute for the goods offered by the parentbrand. Substitute products tend to have common applications and usecontexts such that one product could replace the other in usage andsatisfy the same needs. (129) Cross-country skis or ice skates thereforewould be substitute extensions for ROSSIGNOL. (130) Finally,transferability relates to the relevance of a brand owner'sexpertise in the extension product category. Specifically,transferability depends on the extent to which consumers believe aparent brand owner can use its people, facilities, and skills to makethe new product or offer the new service. (131) Transferability isrelated to credibility, which is a function of perceived expertise andtrustworthiness. (132)

Extension products also benefit to the extent particular concreteor abstract brand attribute associations transfer to the new product.(133) As Keller and Aaker explain, "[e]xtension evaluations willdepend primarily on whether the specific attribute or benefitassociations for the core brand are viewed as relevant in the extensionproduct category and, if so, how favorable those inferred associationsare [in the context of the extension product]." (134) Consumers mayvalue a particular brand association highly in one context but not inanother, even when there is fit between the products or services.Despite the similarity between products, for example, consumers mayvalue thickness in tomato-based juices but not in children'sfruit-flavored drinks. (135) Likewise, pulp is related to high qualityin orange juice but to low quality in apple juice. (136) The extent towhich particular brand attribute associations transfer to the extensionproduct therefore depends "not only on the strength ofassociation" with the parent brand, but also on the"appropriateness of the association" in the new context and"whether cues are present to activate an association." (137)

Importantly, all of these findings depend on the consumerconnecting the new extension to the original brand. (138) If theconsumer does not believe that the original trademark owner manufactured(or at the very least is partly responsible for) the new extension, shewill not attribute the characteristics of the original brand to theextension. As a result, the evidence for benefits to third parties fromadopting an established brand from another market frequently depends onevidence that consumers are actually confused about the relationshipbetween those brands. Indeed, in the studies we cite, there is noambiguity about source: subjects are told explicitly that the extensionscome from the brand owner or that the alliance actually involvesparticular brand owners.

Moreover, the extent to which new products benefit from use of theknown mark depends substantially on the nature of the respectiveproducts and how they are positioned in the marketplace. Whether globalassessments of a parent brand will benefit a new product depends on thenature of the new product and its "fit" with the parent brand.And product characteristics also determine whether many relevant brandattribute associations--like thickness--will transfer to the newproduct. Thus, while it is undeniably true that modem marketing focuseson brand identity and producers' ability to create associationswith a brand, it is also clear that the product or product class towhich the mark has been applied is a central component in the meaning ofa mark to consumers.

Cumulatively, this research suggests to us that a third partystands to benefit from use of an established brand to offer a newproduct or service when (1) consumers believe that the original brandowner makes or is somehow responsible for the new product, (2) that newproduct or service is perceived as a good fit with the brandowner's other products, and (3) particular brand attributeassociations transfer to the new product from the old. These benefitsparticularly redound to uses of brands believed to be high quality.Hence, a third party that makes use of a known high-quality brand tooffer its own products may, in some cases, reap benefits from theassociation with the established brand that it would not if it createdits own new brand.

A final way in which third parties might benefit from the use of aknown brand is by drawing on the emotional valence of that brand.Specifically, consumers' familiarity with a brand may translate toan initial preference for new goods bearing that mark, a preference thatis independent of specific brand attributes. (139) This effect might beparticularly pronounced in situations of low involvement, whereconsumers lack motivation to process information about the products withwhich they are confronted. (140) In those situations, consumers maysimply rely on their familiarity with the known brand, which willredound to the benefit of the junior user.

Brand alliance studies also confirm that certain alliances havesynergistic effects, positively affecting consumers' evaluations ofthe alliance product or service in ways neither alliance partner coulditself. Similar to the lessons of the brand extension literature, thesestudies suggest that consumers' evaluations of brand alliancesdepend on pre-existing attitudes towards the alliance partners asindividual entities (and particularly their levels of perceived quality)and on the fit of the brands in the alliance. (141) Specifically, brandsprove beneficial to a brand alliance if they can "signal highquality cues that transfer to the other alliance brand, or provideinformation on product attributes that benefits the alliance."(142) Indeed, "transfer of perceived quality is enhanced whenbrands fit together." (143)

Fit in the context of brand alliances, like in the brand extensioncontext, can relate to the types of products or services primarilyassociated with the alliance partners or to brand personalitycharacteristics. (144) And fit in both senses can enhance consumerevaluations of the alliance product or service. Thus, according to onestudy, an alliance between Filofax and Sony to offer an electronicpersonal organizer is better received by consumers than an alliancebetween Filofax and Calvin Klein to offer the same product. (145) At thesame time, either of those alliances (Filofax and Sony or Filofax andCalvin Klein) is likely to be better received than an electronicorganizer offered jointly by Calvin Klein and Vidal Sassoon, neither ofwhich offers any expertise in electronics.

C. Evidence of Market Preemption Through Restricted Expansion

A third-party use can directly preempt use of a mark in the latermarket because it establishes priority in the junior user. If the brandowner was in fact going to enter the market using the same brand name,it may now be unable to do so. And as some of the research we discussedabove indicates, some uses can indirectly impact a brand owner'sability to expand by affecting consumer perceptions of the parent brandsuch that, even if the brand owner gets to a new market first, consumersare less likely to accept the extension product or will be lessattracted to it. These uses need not be negative in the sense that theproducts with which they are used are of poor quality or carry offensiveconnotations. A junior use can hamper the brand owner's ability toexpand when the junior product fails, (146) leading consumers toevaluate later extensions more negatively than they otherwise mighthave. (147) Or the junior use can interfere by affecting global beliefsabout the brand (e.g., the belief that Neutrogena products are"mild") and therefore making certain other extensions lesslikely to be accepted by consumers.

Such effects on global brand beliefs, however, operate at anabstract level, and not in the context of the goods the parentpreviously offered. (148) The effect of an extension on specific brandbeliefs also tends to be limited to the parent brand in the abstract,but specific brand beliefs are less important for extension purposes.Research suggests that consumers evaluating a new product tend to relyon global attitudes toward a brand rather than attempting to recall andprocess specific brand attributes. (149) Thus, any interference withexpansion would result primarily from those few cases in which athird-party use negatively impacted global brand attitudes. And ofcourse any such impact wouldn't prevent product expansion entirely;it would merely make it more difficult for the owner of a mark in onemarket to use the same mark in other markets. Many companies in fact useentirely different marks when they expand into new markets; nothing inthe literature we discuss suggests that a brand owner that sought toexpand under a different mark would be harmed at all by anothercompany's use of the brand owner's core mark in a differentmarket.

D. Summary

In short, the available marketing evidence suggests several things.First, the traditional claim of harm to trademark owners fromnoncompeting uses--that consumers will attribute the poor quality of thedefendant's goods to the plaintiff, undermining theplaintiff's reputation--is not persuasive. Even when consumers doattribute the defendant's goods to the plaintiff, they do notpunish the plaintiff for perceived lapses in a different market. Second,some uses can in fact interfere with a brand owner's own potentialuse in a new market, either by getting to that market before the brandowner or by making consumers less willing to accept certain extensions,perhaps by giving those consumers a more negative initial impression ofthe brand extensions. This might be seen as a disadvantage to brandowners, but only if they were in fact going to enter that market usingthe same name. And third, junior users can benefit from using the marksof an established brand in a different market. In the next Part, wediscuss the implications these findings have for trademark law.

IV. EVALUATING CLAIMS TO OWN MARK(ET)S

We begin with the market preemption arguments that purport to fitwithin the framework of producer harm. We then turn to claims thattrademark owners should be able to own mark(et)s themselves,appropriating any value that can trace its origin to the mark or themark owner. In the absence of consumer injury, neither a claim that theproducer has been injured in a market in which it does not compete nor aclaim that someone else has benefited from using the mark provides apersuasive reason to hold a defendant liable for trademark infringement.

A. Claims of Producer Harm

Once we have set aside cases in which consumers are confused aboutthe actual source or quality of the goods, the "injury" atrademark owner suffers from the adoption of the same mark by anothercompany for unrelated goods seems to be that it is more difficult forthe trademark owner to expand into new markets using the same mark. Thatdifficulty can be a result of market preemption--the mark ownercan't enter the same market as the defendant with the same markwithout causing confusion. It can also be a more diffuseproblem--consumers may react differently to brand extensions in a thirdmarket if they have seen what they mistakenly believed were extensionsby the mark owner in other markets.

But merely because the trademark owner wants the ability to enteran unrelated market using its "home" trademark doesn'tmean it has a legal right to do so. After all, many, perhaps most,trademarks are not unique; mark owners frequently coexist with thirdparties that own the same mark in other markets. (151) The question,then, is whether trademark law ought to prefer a particular mark ownerover a junior user in those cases in which the third party uses a markin a way that does not suggest the mark owner is responsible for thequality of the junior user's goods. More particularly, the questionis whether a senior user ought to have superior rights to use a mark inancillary markets because third-party use of the mark would interfere,directly or indirectly, with the senior user's ability to enterother markets under that mark.

It is an important question because extending trademark rightsbroadly into unrelated markets has real costs. We detailed those costsin a prior paper. (152) Briefly, they include the generation of numerouslegal conflicts between marks that would otherwise coexist; theunsuitability of existing legal frameworks to resolving those conflictswhen not directed at traditional forms of confusion; the risk ofactually increasing consumer confusion by changing consumer beliefsabout the role of brands; and the chilling effect on speech abouttrademarks and trademark owners. (153) Because of these costs, we cannotbe indifferent about the initial allocation of trademark rights. Markowners allocated broad trademark rights are likely to exploit thoserights to their specific benefit even if consumers and other producersare hurt as a result. Nor can we blithely assume that parties willsimply bargain to reallocate rights when third parties value the usemore than the mark owner. Many socially valuable uses--such as parodiesand criticisms--are precisely those that trademark owners would neverlicense. And cognitive biases are likely to interfere even for thoseuses mark owners ordinarily would be inclined to license. For one thing,mark owners tend to be strongly attached to "their" marks, andas a result, the "endowment effect" is likely to beparticularly pronounced here. (154) Mark owners may also be overlyoptimistic about their own ability to expand into ancillary markets.(155) The claim of injury to trademark owners from the foreclosure ofmarket entry is circular. Because consumers have not been injured andtrademark owners have not suffered injury in their core markets, thedifficulty mark owners have in expanding their brands to new markets isa cognizable legal harm if, but only if, trademark owners have a legalfight to control the use of their marks in unrelated markets. But thatis precisely the question we are trying to decide. To decide it, we mustdo more than simply assume the conclusion. We must explore the reasonswe might want to grant or withhold control over brand extensions. Thatdebate--and the entire claim of producer harm through marketpreemption--is bound up with the free-riding arguments, to which we nowturn.

B. Free Riding

The market preemption argument is a specific instance of a moregeneral claim: that trademark owners should control words or marketsaltogether, on the theory that the defendant who uses a mark onunrelated goods, or who makes goods that derive from theplaintiff's original market, is free riding on the goodwill builtup by the trademark owner. The claim is not that either consumers ortrademark owners are harmed by this free riding, except in the circularsense that if the law allowed trademark owners to get paid for it theywould have made more money. Rather, the claim is that defendants benefitfrom the use of the mark, the benefit is unjust, and so it should bepaid to the plaintiffs.

As we saw in Part III, there is some truth to the claim that adefendant starting up can benefit from using an established brand nameon unrelated goods, although the likelihood and degree of the benefitdepends on the context of the use and how consonant it is with the imageof the underlying brand. And of course if a defendant hopes to sellgoods based directly on the plaintiff's mark--t-shirts featuring acartoon character or a famous line from a movie, for instance--it haslittle choice but to use the established character or movie line. Anddespite the fact that characters and lines often will not constitutetrademarks, (156) both types of uses are regularly swept within therubric of free riding on a plaintiff's mark.

"Unjust enrichment" is not redundant, however. The factthat defendants are enriched by using the plaintiff's mark shouldbegin the inquiry, not end it. A defendant's enrichment is onlyunjust if there is some reason to believe the value of a trademark inancillary markets should belong to the plaintiff in the first instance.The unjust enrichment rationale simply assumes that conclusion, (157)and it therefore fails to provide a standalone explanation for expandingtrademark law.

The following sections look more closely at four theories of why atrademark owner should own the value of a mark in ancillary markets. Aswe demonstrate, none is terribly persuasive.

1. Preventing Consumer Confusion

The most straightforward reason to prevent defendants from using aplaintiff's mark is that consumers are likely to be confused by theuse in a way that harms them. That last point is critical, however: itis not enough that consumers misunderstand the relationship between theplaintiff's and the defendant's goods. If thatmisunderstanding has no consequence for consumers--if they are not hurtas a result--it is not something trademark law should care about. (158)

The evidence we discussed in the previous section suggests thatconsumers are actually quite sophisticated about brand extensions. Theyare able to think of the extension separately from the underlying brand,and to evaluate the quality of each independently. As a result, in theordinary run of cases, consumers are unlikely to be harmed by adefendant's use of a plaintiff's brand in an unrelated market.There may be circ*mstances, however, in which consumers are confused totheir detriment by an apparent brand extension. We think thosecirc*mstances are likely to be ones where the junior use is sufficientlyrelated to the mark owner's use that consumers are likely tobelieve the mark owner is responsible for the quality of the junioruser's goods. If a defendant's confusing use of aplaintiff's mark actually injures consumers, there is good reasonfor the law to prevent it.

It is worth emphasizing, however, that this is not in fact a theoryof producer harm or of free riding at all. In this set of cases, weenjoin the defendant in order to protect consumers. We may also requiredefendants to disgorge profits in some limited circ*mstances in order todeter intentional deception. (159) But any such disgorgement is awindfall to the trademark owner, who is unlikely to have suffered harm,not compensation for something rightfully theirs.

2. Incentive Theory

The argument that a senior user of a trademark ought to havesuperior rights to ancillary markets is troubling in a market economybecause it presupposes the right to control use of a mark in a market inwhich that mark owner does not compete. Such a fight to controlancillary markets bears striking resemblance to the derivative workfight in copyright law. (160)

a. Incentives in Copyright Law

Precisely because such rights are departures from ordinaryconceptions of competition, they are usually justified in copyright lawby the same sort of incentive theory as copyright law as a whole. (161)According to Paul Goldstein, for example, the derivative work right"enables prospective copyright owners to proportion theirinvestment in a work's expression to the returns expected not onlyfrom the market in which the copyrighted work is first published, butfrom other, derivative markets as well." (162) Landes and Posnersimilarly suggest that derivative rights increase the incentive toengage in creative activities, encourage earlier publication of anoriginal work by making it unnecessary to withhold the publication inorder to gain a lead time in derivative markets, and reducetransactional costs by concentrating the control over derivative workson the copyright owner. (163) To make it more concrete, Landes andPosner suggest that an author will have greater incentive to create (andthe publisher greater incentive to distribute) the next great Americannovel if the author or her publisher can prevent third parties fromcreating a movie or a Broadway play based on the novel. Alternatively,consolidating the fight to control follow-on works to the novel issometimes asserted to be more efficient. (164)

These arguments are far from universally accepted even incopyright. Scholars have argued, for example, that the derivative rightis not necessary to induce production of creative works because creatorsrely primarily on returns from the original work itself to recover theircosts. (165) And some have argued that, whatever marginal incentive thederivative right provides--particularly in light of courts'expansive construction of the reproduction right--is outweighed by thecosts of impeding vigorous competition for works that build on anoriginal work. (166) But the terms of the debate are more or lessaccepted: derivative rights are evaluated under the same sort ofutilitarian calculus as copyright law generally. (167) Copyright lawgenerally--and the derivative works right in particular--intentionallyreduce market competition in order to overcome a public goods problemthat leads to the underproduction of hard-to-create, easy-to-copy works.(168)

b. Incentives to Create New Marks

It is hard to conceive of a similar public goods problem intrademark law that would justify market allocation of noncompeting goodsin incentive-based terms. Whether or not we need to allocate additionalmarkets to authors and playwrights in order to motivate them to createbooks and plays, it seems implausible to think that we need to awardrights in words themselves in order to motivate the creation and use ofnew trademarks. (169) The Fourth Circuit once famously said that areasonably smart person could come up with plenty of new marks in anafternoon's work. (170) While that may overstate the case--there iswork that goes into proper branding--the effort expended is hardly ofthe same order of magnitude as the creation of a new movie, much less anew drug. Moreover, if we did think the incentive-to-create theorymotivated trademark law, we would probably design that law ratherdifferently, giving much more weight to creation and relatively less touse as a central feature. And we would surely not allow firms to claimownership rights in terms they didn't invent, but borrowed fromothers, like Federated or American or United or National. (171)

c. Incentives to Invest in Quality or Market Entry

To be sure, there is an incentive-based argument that is typicallyoffered in support of trademark law generally. Courts and commentatorsboth frequently suggest that trademark rights allow a producer to investin the quality of its goods or services, as "a firm with a valuabletrademark would be reluctant to lower the quality of its brand becauseit would suffer a capital loss on its investment in the trademark."(172) Hence, "legal protection of trademarks encourages theproduction of higher-quality products." (173) On this account,trademark rights don't themselves create the incentives. Marketcompetition provides plenty of incentive to brand and distinguishone's goods. Trademark law merely preserves business incentives bymaking sure the party in control of the quality of the goods getsappropriate credit or blame for that quality. This is a criticaldistinction between trademarks and the rest of IP law. Patents andcopyrights are government efforts to interfere with the operation of themarket, skewing production away from what would happen in a competitiveeconomy in hopes of enhancing social welfare. Trademark law, bycontrast, is not designed to create above-market incentives, but toensure that the competitive market functions well.

Further, this traditional incentive-based argument for trademarklaw is not an argument for granting rights for the purpose of preventingmarket preemption, nor is it an argument particularly focused onancillary markets. Indeed, the argument makes much more sense in thecontext of competing, or at least closely related, goods. If a companyis not producing goods in a market or actively considering entering thatmarket, it has no reputation to protect in that market, and there are nogoods whose quality might be influenced. And the derivative worksrationale surely doesn't justify an anti-free-riding impulse. Thefact that an unrelated company in an unrelated market benefits fromusing the same mark as me doesn't reduce my incentive to makehigh-quality products in any way.

Incentive-based arguments for trademark protection must thereforefocus on any possible feedback that the sale of unrelated productsbearing the same mark might have on consumers' perceptions of thetrademark owner's core product, and hence on its incentives tomaintain the quality of that product. We agree it would be a problem iftrademark law were structured in a way that had negative effects onthese incentives. But we think it is important to think carefully aboutthe scope of trademark protection that is necessary to preserve theseincentives. First, the available empirical evidence suggests thatconsumers are very unlikely to alter their perceptions of the quality ofa core brand based on negative information about noncompetitive goodsoffered under the same mark. (174) And if the image of the core productis intact, the trademark has served its quality-protection function.

Even in the few cases in which the poor quality of an unrelatedgood might affect consumers' views of a brand, mark owners shouldhave sufficient reason to invest in quality under a trademark systemthat affords a claim against uses that imply control over quality. Thequality of any product that is sufficiently remote, such that use of thesame mark causes no confusion about who is responsible for the qualityof the product, won't be attributed to the mark owner. Hence, themark owner has no reputation for quality at stake.

We can imagine a few cases in which the prospect of being able tocontrol ancillary markets could affect the quality of a markowner's goods in its core market even though consumers don'tpenalize the mark owner for the quality of the goods in those ancillarymarkets. But for this story to hold, we would have to believe that atrademark owner planned to expand into separate markets hoping foradditional returns from doing so, and that the mark owner wouldn'tinvest enough in product quality in its core market unless we give itthose returns. We are skeptical that this often happens in practiceoutside of specialized contexts like Disney, which produces movies withone eye on the marketing tie-ins they will create. But even if it doeshappen, it doesn't follow that trademark law should encourage it.Trademark law is designed to ensure that consumers can draw the fightconnection between the product and its maker, so that they understandthe quality of the products they are buying. Allowing a trademark ownerto capture spillovers in other markets distorts that relationship.Companies who try to build up a core brand in hopes of parlaying it intosales elsewhere are actually over-investing in product quality in thecore brand, hoping to use that investment to capture rents in adifferent market. It is as though we took one and only one area of theeconomy--say, carpentry--and gave out lottery tickets to carpenters.More people would become carpenters (and fewer people would go intoother professions of equal worth) because they were offered anadditional, government-provided bonus found nowhere else.

Nor do we think broader trademark rights are necessary to preserveincentives to enter new markets, even assuming we want to encourage suchentry. The current trademark infringement paradigm grants mark ownersclaims against any uses likely to confuse consumers about who isresponsible for the quality of the goods at issue, and mark ownerstherefore have sufficiently robust protection to prevent against wastedinvestments preparing to enter closely related markets. We can, however,imagine circ*mstances in which a mark owner should not have to take thechance. The risk here would materialize only in cases in which a markowner is actively preparing to enter a new market under its existingbrand, but that new market is sufficiently remote from the brandowner's core market that consumers are not likely to believe thebrand owner is currently responsible for the quality of any thirdparty's goods.

The most likely set of cases are those in which the preparation ispublicly known and is of sufficient importance to attract attention fromthose who seek to divert consumer attention. For example, sports teamsthat change cities or names, or build new stadiums, may invite such rentseeking. (175) And large companies that plan to expand abroad often findtheir marks in the hands of so-called "trademark pirates." Inthose cases, the brand owner stands to risk its investments preparing toenter the new market if another party can enter the market and establishpriority there.

This is, in our view, the best justification for an intent-basedregistration option. (176) And we think the intent-to-use system is thebest way to deal with these circ*mstances. In fact, we think a markowner should not be able to simply assert a trademark infringement claimagainst a junior user in these contexts and try to prove itspreparations to enter the market. Instead, a mark owner that wantsprotection of its investment should have to file the intent-basedapplication to signal a real investment in entering the new market.(177) But if a party does file an ITU application, the law should treatthat party as senior to any other party that subsequently makes actualuse, even though the applicant was actually the later entrant into themarket. (178) And indeed courts generally find a way to achieve thatresult, though sometimes through the problematic rhetoric of freeriding. (179)

d. Incentives to Invest in the Brand

An alternative incentive-based argument might focus on a markowner's incentive to invest in the brand itself--particularly inthe brand's personality or "atmospherics." This is a morecontroversial argument because it isn't clear we want to encouragegreater investment in brands themselves. Several scholars have suggestedsuch investments can be wasteful, as they may encourage"irrational" brand loyalty and create barriers to entry. (180)But even assuming for the sake of argument that we do want mark ownersto invest at least to some degree in their brands above and beyond theirinvestment in the goods they represent, we think the value the markowners can realize in their core markets gives them ample incentive todo so. Marketing research suggests that most consumers categorize marksin terms of their product categories and that brand personalitiesdistinguish brands within their markets. And while it's true thatsome aspects of a brand image play a role in determining the likelihooda particular extension will be accepted by consumers, the payoff frominvestments in the brand should generate sufficient returns in a markowner's primary market to make such investments worth while.Moreover, a trademark infringement standard that focuses onresponsibility for quality will generate rights broader than justdirectly competitive uses.

Having said this, we concede it's possible that refusing toprotect a mark against some uses outside the control-over-quality rangewe have advocated will somewhat reduce the incentive to invest in thisbrand "personality." But we think that whatever incentive iscreated by this incremental difference in scope is small, particularlysince third-party uses in ancillary markets can also benefit mark ownersby making their marks more familiar and therefore more likeable. Such asmall incentive effect clearly doesn't warrant the costs ofexpanding trademark protection. As we explained in our prior work, thosecosts are substantial: claims to own rights in noncompeting goods areharder to evaluate, interfere with free speech and social dialogue, andmay actually make consumers more rather than less confused. (181) And itis no accident that those costs are at their greatest precisely in thecases in which the incentive story is at its weakest.

3. Accession and Default Ownership

Some argue that defendants should be prevented from using a markbecause, even if the producer is not harmed by that use, someone mustown the right to use the mark in ancillary markets, and the trademarkowner has a better claim than anyone else. This argument finds itsorigin in property theory and is primarily concerned with thedisposition and use of real property. Tom Merrill has argued that thereal-property principle of accession justifies assigning certainintellectual assets to someone, and that the logical owner is the one"most prominently connected" to the asset--the person who canmake the highest and best use of the property. (182)

One of us has elsewhere criticized the effort to fit IP into therubric of property, pointing out that doing so often ends up warping IPrules because the nature of intangible property is so different than thenature of real and chattel property around which property doctrines werebuilt. (183) The idea of accession seems to us a perfect case in point.The principle of accession is a way of allocating rights in newproperty--land that has been deposited on a river bank, for instance, orfixtures and other improvements that have been attached to land bysomeone other than the owner of that land. (184) The underlyingpresumption--as it seems to be with real property in general--is thatsomeone must own the new asset, and the owner of some existing, nearbyasset--the owner of existing land to which the new land is attached, forexample--seems the most natural choice, particularly since dividedownership of attached things creates its own set of problems. (185)

Whether or not the presumption that all property must be ownedmakes sense for land or other tangible property--and there are someimportant reasons to question it (186)--it certainly doesn't makesense for IP. (187) Patent and copyright law not only permit butaffirmatively demand that the residuum of ideas and creations remainsunowned, free for all to use. This public domain does not lead to a"tragedy of the commons," as some have argued an analogouscommons would in the physical world, (188) for the simple reason thatconsumption of ideas is non-rivalrous. (189) And while it's truethat the value of that new asset may be partly rivalrous, (190) therivalry of value generally does not warrant assigning exclusive control.There is substantial, and rivalrous, value in an exclusive right tooperate in any market, but we ordinarily prefer that value be shared bycompetitors, who create consumer welfare through their competition. In amarket economy it is not reasonable to simply assume that someone mustown the right to compete in particular ways. (191) As one of us hasnoted:

 It is true that if we gave only one person control over a particular [trademark], that person would restrict the [use of that trademark], raise its price, and make more money than providers do in a competitive market. But society as a whole would be worse off, since buyers who could afford to pay more than what it costs to provide the information still would not receive it. (192)

In other words, exclusive rightsholders prevent dissipation of thevalue of exclusivity by acting like monopolists, complete with therequisite supracompetitive prices. "But that supracompetitivereturn is not found money; it comes directly out of consumer surplus.And basic economics teaches us that what the owner gains from exclusivecontrol is less than what consumers lose." (193) Absent someoffsetting consideration--like material consumer confusion--there is noreason to accept this outcome.

If we really believed the principle of accession ought to beapplied to trademarks--that someone must own them, and that ownershipshould entail control of use in ancillary markets--there is no reason tostop with existing marks that have already acquired goodwill. We couldparcel out ownership rights to every word in the English language todifferent claimants, by lottery or auction, and record those words in aregistry so everyone would know with whom they needed to deal to use thewords. The fact that we don't--that the whole idea seemsludicrous--is pretty good evidence that words are not like land when itcomes to ownership.

Moreover, even if we were to accept that someone must own the rightto use a trademark in a particular ancillary market, it's not clearwhy we would give ownership of the term ("Exxon" as applied togloves, for example) to the first user of that mark in a differentmarket (Exxon Oil company) rather than the first user in the new market(the company that had first used the term for gloves). Indeed, trademarklaw has always assumed the opposite: trademark rights are awarded on thebasis of first use of a mark in a particular geographic and productmarket. (194)

Accession, as Merrill acknowledges, is an alternative to firstpossession, and the case for preferring it over first possession dependson a determination that the party to whom the asset is assigned is themost prominently connected to the asset. (195) Prominent connection, ifit can be reliably deter. mined, (196) is important to Merrill bothbecause it serves as a proxy for being a competent manager of the assetand because it appears to lower the cost of determining who owns theasset so that coordination can take place. (197) But information costsare not really at stake here, because as trademark law has longrecognized, first use of a mark in a particular market is generally asufficiently clear indication of ownership. And the only cases in whichwe may have reason to believe the senior user is a more competentmanager of a mark are those in which the junior use is for relativelyclosely related products where the quality signal conveyed by a markmight be meaningful. Yet those are precisely the cases in which ourapproach to infringement--focusing on responsibility for quality--wouldalready extend protection. Outside of those cases there is no reason tosuspect that the senior user is a better, more competent manager of theasset. Indeed, we might well assume the opposite: by getting to a newmarket first, a junior user might be thought to have demonstratedsuperior foresight or greater ability to exploit that new market.

Nor is there any reason to fear costly races in the trademarkcontext. Development of trademark rights in a new market entailsrelatively little investment because the amount of use necessary totrigger trademark rights is small. And while there are costs indeveloping new branding, those costs by definition are reduced if themark has already been created.

In the end, the persuasiveness of accession depends on whether wethink there is any harm to the "overuse" of a mark in anunrelated market in a way that does not confuse consumers about thesource or quality of the goods being sold. As we have seen, the evidenceof such harm simply isn't there. The only "injury" thetrademark owner can be said to suffer is a possibly reduced ability touse the same mark in an entirely different market. But that is an injuryonly if we define the scope of the trademark owner's legalentitlement to include control over those markets. The idea of accessiondoesn't help here, because it assumes the conclusion--that someonemust own the right to use a term across all markets.

4. Natural Rights and the Property Instinct

Another possible argument for such broad rights is that trademarkowners have some natural fight not just to use their mark to brand theirgoods, but to prevent others from similarly branding unrelated goods, oreven from selling goods that draw their market in part from the allureof something the plaintiff has created. The instinct here is oneRochelle Dreyfuss has derided as "if value, then right" (198):there is value to the use of a mark on unrelated goods, and so thetrademark owner deserves that value. This type of argument increasinglyshows up in scholarly commentary about trademarks. (199) The naturalrights argument is striking because trademark law in the nineteenth andearly twentieth century was frequently justified in natural rightsterms--and courts that relied on natural rights principles createdsignificantly narrower rights than modern trademark law recognizes.(200) Specifically, courts relying on a natural rights theory oftrademark rights strictly limited trademark rights to markets in whichthe mark owner actively competed. They did so because, while they aimedto protect the fruits of a producer's honest labor by preventingcompetitors from stealing its trade, (201) they also felt compelled toavoid interfering with the rights of others to develop their own trade.(202) Courts distinguished between legitimate and illegitimatediversions of the mark owner's trade by focusing on deception.Indeed, this distinction is critical to understanding why trademarkinfringement was grouped with other forms of unfair competition:competition was an essential element of the claim, and courts gavecontent to the "unfair" component by distinguishing betweenhonest and dishonest actions. Use in different markets wasn'tdishonest because consumers could not be tricked by those uses intobuying one product believing it to be another. (203)

Modern natural rights arguments seek something muchdifferent--broad ownership of a mark independent of any particularmarket on the theory the senior user owns any value attributable to themark. And here the argument is not simply disconnected from history; itloses its conceptual coherence. Natural rights theories--at least thosethat don't depend directly on the will of a god--grant propertyrights on the basis of productive use of an asset, and courts operatingin the natural rights tradition viewed customer patronage as therelevant asset in which a producer had a property interest. Trademarkrights therefore existed solely for the purpose of preventingcompetitors from luring away a producer's customers bymisrepresenting themselves as the mark owner. Whatever one's viewof the natural rights theory generally, customer patronage is much moreobviously the result of a mark owner's productive labor than is atrademark itself, and uses of the mark by direct competitors to divertcustomers who otherwise would have gone to the mark owner are much moreclearly interfering with that productive labor. The value modem naturalrights arguments seek to protect is a much broader brand value, andsubstantially less of that value is clearly attributable to the markowner. Indeed, in many of the cases in which this argument would havesignificance, the uses at issue draw on the mark's value as asocial referent--value created largely by the public. (204)

If we did think that the use of a word--or control over amarket--in fact belonged as a matter of fight to the person who createdits value, it is not obvious to us that the rightful owner would be thecompany that first used the mark in an unrelated context rather than thefirst user of that mark in the new market. (Remember that natural rightsare here being used to justify ownership of the mark for unrelatedgoods, not simply for goods on which the first user has actually usedthe mark.) In other IP contexts, advocates of a natural rights theoryoffer it in order to justify conferring rights on authors or inventors,not on record companies or semiconductor manufacturers. (205) The factthat trademark law expressly protects companies, not individuals, alsomakes a natural rights theory seem out of place, at least to the extentthat theory arises from claims of desert based on individual dignity.

Finally, if we did accept a natural rights theory of trademark law,trademark doctrine would look radically different than it does today.Trademark law has long countenanced simultaneous use of the same orsimilar marks in different product and geographic markets. But if theadoption of a mark were to confer on its adopter a natural right tocontrol the word in other markets, that coexistence would becomeproblematic. Every mark would have to be unique, or at least soinsulated from any other that the users wouldn't share consumers. Anatural rights approach would similarly sweep away the likelihood ofconsumer confusion test that lies at the heart of trademark law, since adefendant would run afoul of the first user's rights whether or notconsumers were confused. After all, natural rights are being used hereprecisely to justify ownership of a mark in circ*mstances in whichconsumers are not materially confused by a use. The hierarchy of markprotection would probably have to go as well: if I have a natural rightto my mark once I create it, it is not clear why it should matter whatmarket I use it in, and whether it is descriptive, suggestive, orarbitrary in that market. On the other hand, if creation is the act thatimbues a mark with its natural rights, perhaps trademark law shouldprotect only fanciful marks, as those are the only marks that are reallycreated. Either way, the protectability spectrum doesn't seem tomesh with a natural rights approach. Nor do defenses such asfunctionality, abandonment, nominative use, and certainly notgenericide. (206) If I own the mark, politicians, newspapers, andparodists should have just as much obligation to refrain from using itas companies who sell products in different markets under that mark.After all, they are profiting from the use of the mark, generally farmore directly than is a company in a different market that happened toadopt the same name. So, too, are gas stations and stores thatdeliberately locate across the street from their branded competitors,(207) and for that matter ordinary consumers who use a brand asshorthand in conversation. (208)

The problem is that an a priori decision to define the scope of alegal entitlement can easily be an arbitrary one unless it is tied tosome sort of social welfare calculus. We could have given the inventorof the steamboat (whomever that turned out to be) (209) a potentiallyperpetual legal right to prevent the making of steamboats that workedthe same way, or a legal right to prevent the making of any steamboats,or a legal right to prevent the making of any transportation devicesthat relied on steam power, or a legal right to prevent the making ofany transportation devices that relied on any machine for power. (210)Similarly, we could decide a priori that trademarks confer only rightsto prevent competing uses, or rights to prevent competing and analogoususes, or rights to prevent all commercial uses, or rights to prevent alluses of any type, including use of the word in conversation. Saying"someone must (or deserves to) own this," even if true,doesn't help answer the question of who should own it and what thescope of their ownership right should be. Those questions can beanswered only by resort to social welfare. Absent a reason to believethe world would be a better place if we created a new property right,the fact that it is a property fight is no reason to do so. And the onlyplausible reason that has been offered to protect trademarks is toprotect consumer perceptions and prevent mistaken purchasing decisions.

In short, it is hard to credit some sort of pre-existing naturalentitlement to a trademark. And even if we did credit it, itwouldn't justify trademark law in anything like its current form.

How, then, to explain the seemingly universal instinct that freeriding is bad and must be stopped? We begin by questioning theuniversality of that instinct. People react to the sale of t-shirtsbearing university logos as free riding, but seem to have no similarobjection to Little League teams adopting the names of professionalbaseball clubs. People who think companies should be entitled to controluses of a common English word like "apple" in markets entirelyremoved from computers and phones see nothing wrong with a gas stationdeciding to locate across the street from a competitor. The instinct, inother words, seems to be contingent--dependent not on the economic factsof a use but on how accustomed we are to seeing it and how it can becharacterized.

One might perhaps turn to sociobiology: it may be that we arehardwired with some version of the Golden Rule, and that freeriding--when painted as such--offends our sense of justice. But if so,our genes are serving us ill. For as we discuss in the next section,what might make sense in the world of rivalrous goods makes little sensewhen applied to non-rivalrous goods like ideas.

C. Spillovers

Our point is not merely that there is no evidence to justifyextending trademark law beyond cases of consumer harm. An unjustenrichment approach--one that attempts to identify and weed out freeriding--may actually do affirmative economic harm. While law andeconomics theorists focused on real property have often spoken of adesire to internalize externalities, thereby allowing a producer tocapture the full benefits of its product, students of the economics ofinnovation have increasingly come to recognize that in factuncompensated positive externalities, or spillovers, serve a valuablesocial function. Far from interfering with incentives, empiricalevidence suggests that these spillovers actually drive furtherinnovation. Industries with significant spillovers generally experiencemore and faster innovation than industries with fewer spillovers. (211)Dietmar Harhoff finds empirical evidence that firms in high-technologyindustries (the most innovation-intensive ones) are likely to increaserather than decrease their investment in research and development in theface of significant intra-industry spillovers. (212)

The computer industry shows this dynamic at work. Both AnnaleeSaxenian and Ron Gilson have shown that spillovers drove innovation inthat industry: Silicon Valley thrived while Boston's Route 128withered in the 1980s and 1990s in significant part because employeesand knowledge moved freely to new companies in Silicon Valley, but notin Boston. (213) And as Alan Hyde puts it, no shortage of innovationresulted:

 In California, employees are normally free to change jobs without a lawsuit alleging ... breach of a covenant not to compete.... There is no evidence of any social harm from this. In particular, there is no evidence that firms lack incentives to invest in the production of information. (214)

More generally, Brett Frischmann and Mark Lemley have argued:

 [T]here is no reason to think that complete internalization of externalities is necessary to optimize investment incentives; at some point, there are decreasing returns (in terms of improved incentives) to allowing property owners to capture more of the value from their inventions. Spillovers do not always interfere with incentives to invest; in some cases, spillovers actually drive further innovation.... [E]ven where internalizing externalities increases incentives to invest, the social costs of relying on property rights to do so still may exceed the benefits. (215)

If consumers aren't hurt by the use of a mark on noncompetinggoods or on products that relate to and draw strength from the markowner's goods, there is little reason to worry even if defendantsin these cases are free riding on the plaintiff's mark. And theremay be good reason to celebrate that costless use. If the mark has valuein the other market, and if another firm can get to that market first,there is consumer benefit to allowing that firm to get there, since theconsumers aren't confused to their detriment by the new mark, and(by hypothesis) affirmatively value it. For instance, in our world theUniversity of Southern California and the University of South Carolina,which have both been in existence for over 125 years, could continue tocoexist even though they both go by the monikers "SC" and"USC." There is no room for that coexistence in theanti-free-riding world: if consumers overlap, someone's use of amark must surely be a detriment to someone else. (216) But this issimply a positive externality generated by the first mark owner, andthere is no reason it should be stopped absent good evidence it somehowharms consumers or the mark owner.

A more concrete case of spillover benefits comes in cases in whichdefendants sell products built on a literary or movie character or amark that owes its origin to the plaintiff. If consumers aren'tconfused about the source or quality of the goods and if the trademarkowner doesn't need to control the market for these derivativeproducts in order to induce it to invest in quality, (217) there issimply no harm to allowing "ambush marketing" by companiesthat manufacture goods designed to appeal to sports fans, or who writefan fiction featuring literary characters from the trademarkowner's work, or who sell plush dolls featuring those characters.And there is substantial social benefit, not only to the seller but toconsumers, who get a wider variety of mark-related goods, generally at alower price and of higher quality than the mark owner alone wouldlicense, (218) and who get takes on the trademarked goods and charactersthat the trademark owner would never permit if given control. Trading onthe goodwill of an established brand without confusing consumers inmaterial ways can also serve as a way for a new product to enter amarketplace and therefore expand competition by giving consumersinformation about the new product. (219) In fact, there is good evidencethat consumers benefit when they can use familiar packaging features togeneralize product qualities. (220) More generally, there are consumerbenefits to autonomy--to being able to "make whatever associationsshe wants with the marks she encounters, even if those associations arenot the ones the mark holder would prefer." (221) Some think ofthese uses as harmful because they involve free riding. But we shouldthink of them as positive spillovers--benefits conferred on defendantswithout corresponding harm to plaintiffs or consumers.

Wendy Gordon explained two decades ago that "[a] culture couldnot exist if all free riding were prohibited within it." (222)Trademark law seems perfectly designed to prove her point. It isdesigned to facilitate a competitive marketplace by allowing consumersto know what they are buying, or at least from whom. But a trademark lawthat is distorted into a right to own markets--one that seeks out andtries to forbid all free riding on a mark--ends up interfering withrather than enabling competition.

V. TOWARD A "TRADEMARK INJURY" DOCTRINE

Arguments about market preemption that assume a trademarkowner's right to control use of a mark in remote markets, andfree-riding arguments that take for granted that defendants should notmake money using marks plaintiffs first adopted, carry substantialcurrency in court. Upon examination, though, those arguments arecircular and lack empirical support. The simple solution, then, would beto reject the arguments outright. In fact, however, the world may bemore complicated. While the market preemption and free-riding instinctsseem to be driving the results in a disproportionate number of the mosttroubling trademark cases, those cases are often nominally decided underdoctrines that at least claim to find consumer harm, such asinitial-interest confusion, post-sale confusion, and dilution. Whenclaims of consumer injury are mixed with unexamined but attractiveclaims about producer injury or unjust enrichment, it has proven all tooeasy for courts to find a violation without seriously considering theclaims of consumer injury.

We propose to solve this problem by creating a "trademarkinjury" doctrine. Just as courts in antitrust cases created anantitrust injury doctrine in the 1970s to try to weed out faciallyappealing but ultimately anticompetitive antitrust claims, (223) courtsshould require trademark plaintiffs to show trademark injury as acondition of standing. The situations are quite parallel. (224)Antitrust plaintiffs are supposed to be vindicating not just their owninjuries, but the injuries of consumers more broadly, just as trademarkplaintiffs are. And, as with antitrust plaintiffs, trademark plaintiffssometimes complain of conduct that injures them personally but actuallybenefits competition overall.

The antitrust injury doctrine requires plaintiffs to show that theyhave suffered "injury of the type the antitrust laws were intendedto prevent and that flows from that which makes defendants' actsunlawful." (225) Something similar would be appropriate intrademark cases. Trademark owners should be permitted to bring trademarksuits only if they can prove that their "injury" is in factrepresentative of injury to consumers rather than merely a benefit tosomeone else. In our view, this would mean that trademark plaintiffsshould have to demonstrate (I) that their injury flows from confusionabout the actual source of the defendant's goods or about who isresponsible for the quality of those goods, or (2) that thedefendant's use causes confusion about some other relationship thatis material to consumer purchasing decisions. When consumers areconfused about actual source or about responsibility for quality, the"injury" to mark owners, if there is one, is caused by conductthat harms the marketplace more generally. But other forms of confusionhave, at best, ambiguous effects on competition. In order to assure thatmark owners are allowed claims only when consumers are negativelyimpacted in their ability to make decisions, in these cases plaintiffsshould bear the burden of demonstrating materiality. If mark owners canneither show confusion about source or responsibility for quality northat the alleged confusion is material, then any "injury" themark owner suffers is not a trademark injury.

In some of the cases lacking a trademark injury there will be otherlegal mechanisms through which perceived harms can be addressed. In thecases involving characters, for example, authors can rely on copyrightlaw to remedy certain forms of free riding. But where copyright does notoffer a remedy, trademark law should not step in. Copyright servesdifferent values, and it is not trademark law's role to backstopcopyright protection to ensure that the author captures all of the valueof a work. Copyright leaves certain uses open for a reason.

A trademark injury doctrine is not a panacea. Courts determined tosneak market preemption and free-riding concerns into trademark casescould probably still do so. But requiring courts to assess the ways inwhich a defendant's conduct does or does not affect competitionshould give them a greater awareness of the need for plaintiffs to proveinjury. It will require the parties and the courts to focus on exactlywhat is asserted to be wrong with the defendant's use. And ifnothing else, it will make clearer the role of free-riding claims indeciding trademark cases.

CONCLUSION: INSTINCTS AND EVIDENCE

The anti-free-riding impulse is a deep-seated one, despite itsquite recent importation into trademark law. Perhaps this results fromapplying our intuitions about land, or perhaps we have internalized theincentive stories of other, quite different IP regimes. Whatever thereason, the impulse has seeped sufficiently far into the public moralconsciousness that the reader's instinct may well rebel at the ideathat someone else should be able to use my mark to make money.

If we persuade you of nothing else, we hope at least to haveconvinced you that this instinctive reaction is worth examining. Andonce we do examine it, the free-tiding instinct proves remarkably hardto support. Not history, nor economics, nor logic support giving theowner of a mark in one market the power to control all uses of that markeverywhere. And the consequences of trying to squelch all free tidingwill be substantial. Owning marks should not mean owning markets.

(1.) Press Release, DEWALT Partners with Eurosport to Launch NewCompact Lithium Cordless Products (Oct. 6, 2009),http://www.hispanicprwire.com/News/in/15642/18/dewalt_partners-with-eurosport-to-launch-new-compact%20_lithium_cordless_products (internal quotation marks omitted). Black & Decker focused thesemarketing efforts primarily on "key Hispanic markets":Chicago, Houston/Austin, and Miami. hi.

(2.) Id. The Mexican national team is often referred to by itsnickname, El Tricolor. See, e.g., La Plaza,http://latimesblogs.latimes.com/laplaza2009/08/mexico_shows_us_is_no_match_in_ soccer.html (Aug. 13, 2009, 10:25 PST).

(3.) Complaint at [paragraph] 38, Soccer United Mktg., L.L.C. v.Black & Decker Corp., No. 09-CIV-10378 (S.D.N.Y. Dec. 22, 2009).

(4.) Id. at [paragraph] [paragraph]15, 31.

(5.) Id. at [paragraph][paragraph] 31, 36.

(6.) See infra notes 35-39 and accompanying text (discussing thesecases).

(7.) See Mark P. McKenna, Testing Modern Trademark Law'sTheory of Harm, 95 IOWA L. REV. 63, 114-15 (2009) (describing marketingstudies focusing on brand alliances and concluding that consumers do notroutinely blame a host brand for its partner's mistakes).

(8.) Mark A. Lemley & Mark McKenna, Irrelevant Confusion, 62STAN. L. REV. 413, 436-46 (2010).

(9.) Complaint, supra note 3, at [paragraph] 37.

(10.) See infra notes 16-24 and accompanying text.

(11.) United Drug Co. v. Theodore Rectanus Co., 248 U.S. 90, 97(1918) ("The asserted doctrine is based upon the fundamental errorof supposing that a trade-mark right is a right in gross or at large,like a statutory copyright or a patent for an invention, to either ofwhich, in truth, it has little or no analogy.").

(12.) Lemley & McKenna, supra note 8, at 429-32; McKenna, supranote 7 at 1010-14; cf. Rebecca Tushnet, Gone in Sixty Milliseconds:Trademark Law and Cognitive Science, 86 TEX. L. REV. 507 (2008)(demonstrating the weakness of empirical evidence of dilution).

(13.) Lemley & McKenna, supra note 8, at 432-35,449-56.

(14.) See infra Section III.B and sources cited therein.

(15.) See WILLIAM M. LANDES & RICHARD A. POSNER, THE ECONOMICSTRUCTURE OF INTELLECTUAL PROPERTY LAW 167-68 (2003); NicholasEconomides, Trademarks, in 3 THE NEW PALGRAVE DICTIONARY OF ECONOMICSAND THE LAW 601, 602 (Peter Newman ed., 1998) (describing the savingsfor consumers in product searches as one of "it]he primary reasonsfor the existence and protection of trademarks"); Nicholas S.Economides, The Economics of Trademarks, 78 TRADEMARK REP. 523, 525--270988) (discussing the economic benefits of marks that apprise consumersof products' unobservable features); William M. Landes &Richard A. Posner, Trademark Law: An Economic Perspective,

30 J.L. & ECON. 265, 268-70 (1987) (identifying the lowering ofbrand recognition costs to consumers as the justification for trademarklaw); Mark A. Lemley, The Modern Lanham Act and the Death of CommonSense, 108 YALE L.J. 1687, 1690-94 (1999) (describing economicjustifications for trademarks and advertising); I. P. L. Png & DavidReitman, Why Are Some Products Brandedand Others Not?, 38 J.L. &ECON. 207, 208-1l, 218 (1995) (analyzing empirical search cost data andsuggesting that "consumers of products subject to performanceuncertainty will pay for brand-name assurance"); John E Coverdale,Comment, Trademarks and Generic Words: An Effect-on-Competition Test, 51U. CHI. L. REV. 868, 869-70 (1984) (noting that trademark law encouragescompetition, which potentially decreases the cost to consumers); BrianA. Jacobs, Note, Trademark Dilution on the Constitutional Edge, 104COLUM. L. REV. 161, 164 (2004) (noting search costs rationale); see alsoQualitex Co. v. Jacobson Products Co., 514 U.S. 159, 163-64 (1995)(explaining that trademark law "reduce[s] the customer's costsof shopping and making purchasing decisions," and "helpsassure a producer that it (and not an imitating competitor) will reapthe financial, reputation-related rewards associated with a desirableproduct") (internal quotation marks omitted); Union Nat'l Bankof Tex., Laredo v. Union Nat'l Bank of Tex., Austin, 909 F.2d 839,844 (5th Cir. 1990) ("The idea is that trademarks are'distinguishing' features which lower consumer search costsand encourage higher quality production by discouragingfree-riders."); cf. Mishawaka Rubber & Woolen Mfg. Co. v. S.S.Kresge Co., 316 U.S. 203, 205 (1942) ("A trademark is amerchandising short-cut which induces a purchaser to select what hewants, or what he has been led to believe he wants.").

(16.) Lemley & McKenna, supra note 8. For a detailed analysisof the relevant marketing literature, see McKenna, supra note 7.

(17.) Schechter clearly accepted this proposition:

 Quite apart from the destruction of the uniqueness of a mark by its use on other goods.., once a mark has come to indicate to the public a constant and uniform source of satisfaction, its owner should be allowed the broadest scope possible for 'the natural expansion of his trade' to other lines or fields of enterprise.

Frank I. Schechter, The Rational Basis of Trademark Protection, 40HARV. L. REV. 813,823 (1927).

(18.) See infra notes 160-168 and accompanying text.

(19.) This was indeed the result in an early case, Borden Ice CreamCo. v. Borden's Condensed Milk Co., 201 F. 510, 514 (7th Cir.1912), though we are quite confident courts would not reach the sameresult today. See Lemley & McKenna, supra note 8, at 428 & n.59.

(20.) We thank Stacey Dogan for suggesting this alternativepreemption argument.

(21.) Indeed, because the traditional likelihood of confusion testincludes the likelihood that the plaintiff will expand into thedefendant's market as a factor, AMF Inc. v. Sleekcraft Boats, 599F.2d 341,354 (9th Cir. 1979); Polaroid Corp. v. Polarad Elecs. Corp.,287 F.2d 492, 495 (2d Cir. 1961 ), there is arguably no reason to assertmarket preemption as a separate theory of injury except in the verycases in which that expansion is unlikely.

(22.) In the geographic context, the Lanham Act has long permittedsuch concurrent uses. See 15 U.S.C. [section] 1057(c) (2006).

(23.) We explained this in Irrelevant Confusion. See Lemley &McKenna, supra note 8, at 433-36.

(24.) IIA PHILLIP AREEDA ET AL., ANTITRUST [paragraph] 349a (3ded.).

(25.) See Am. Geophysical Union v. Texaco, Inc., 60 F.3d 913 (2dCir. 1994); Mark A. Lemley, Should a Licensing Market Require Licensing?, 70 LAW & CONTEMP. PROBS. 185 (2007).

(26.) See Borden Ice Cream Co. v. Borden's Condensed Milk Co.,201 F. 510 (7th Cir. 1912) (rejecting the plaintiff's claim thatuse of the BORDEN mark for ice cream infringed its rights in BORDEN formilk and related products); see also JAMES LOVE HOPKINS, THE LAW OFTRADEMARKS, TRADENAMES AND UNFAIR COMPETITION [section] 3, at (2 (2d ed.1905) ("The qualified fight in the tradename [or a trademark],--aright to prevent a defendant from passing off his goods as those of theplaintiff by the use of it--exists only with regard to goods of the kindfor which the plaintiff uses it, and to which the connection with hisbusiness suggested by the use of the name extends." (quoting DUNCANM. KERLY, THE LAW OF TRADEMARKS, TRADE-NAME, AND MERCHANDISE MARKS 475(2d ed. 1901))).

(27.) 247 F. 407 (2d Cir. 1917).

(28.) Id. at 410.

(29.) Id. at 409-10.

(30.) Id.

(31.) 166 F.2d 348 (9th Cir. 1948).

(32.) Id. Because geographically remote uses cannot divert tradethat otherwise would have gone to the mark owner, trademark lawtraditionally limited a party's rights to the geographic areas inwhich it actually used a mark. See, e.g., Thomas J. Carroll & SonCo. v. McIlvaine & Baldwin, 183 F. 22, 26-28 (2d Cir. 1910)(distinguishing between the Baltimore and New York markets for whiskeyand denying the plaintiff, which had prior rights in Baltimore, thefight to prevent use of the same mark in New York).

(33.) Stork Rest., Inc., 166 F.2d at 356.

(34.) Id.

(35.) 611 F. Supp. 360 (W.D. La. 1984).

(36.) Id. at 368.

(37.) Id.

(38.) 544 F.3d 1167 (2d Cir. 1976).

(39.) Id. at 1172.

(40.) Id. The court identified two other relevant interests: themark owner's interest in "protecting the good reputationassociated with his mark from the possibility of lit] being tarnished byinferior merchandise of the junior user," and the"public's interest in not being misled by confusingly similarmarks." Id.

(41.) See generally Stacey L. Dogan & Mark A. Lemley, TheMerchandising Right: Fragile Theory, or Fait Accompli?, 54 EMORY L.J.461 (2005) (discussing the recent vintage of merchandising claims).

(42.) 510 F.2d 1004 (5th Cir. 1975).

(43.) Id. at 1011.

(44.) Boston Athletic Ass'n v. Sullivan, 867 F.2d 22, 33 (1stCir. 1989); see also Univ. of Ga. Athletic Ass'n v. Laite, 756 F.2d1535, 1547 (11th Cir. 1985) (enjoining use of Battlin' Bulldog beerwhen "the confusion stems not from the defendant's unfaircompetition with the plaintiff's products, but from thedefendant's misuse of the plaintiff's reputation and good willas embodied in the plaintiff's mark"); Sigma Chi Fraternity v.Sethscot Collection, No. 98-CV-2102, 2000 WL 34414961, at *9 (S.D. Fla.Aug. 7, 2000) ("[T]he confusion factor is met where, as here, theregistered mark.., is the triggering mechanism for the sale of theproduct."); cf. Univ. Book Store v. Bd. of Regents, 33 U.S.P.Q.2d(BNA) 1385 (T.T.A.B. June 22, 1984) ("[The] antiquated view oftrademarks as harmful monopolies which must be rigorously confinedwithin traditional bounds [is] outmoded and not in accordance with morerecent cases.").

(45.) Bd. of Supervisors for La. State Univ. Agric. & Mech.Coll. v. Smack Apparel Co., 550 F.3d 465,477-78 (5th Cir. 2008) (holdingthat universities' color schemes are protectable and thatothers' use of those colors on t-shirts evoking the universitiesinfringed their rights).

(46.) Id. at 488.

(47.) Judge Learned Hand famously described the difference betweenprotectable and unprotectable characters in Nichols v. UniversalPictures Corp.:

 If Twelfth Night were copyrighted, it is quite possible that a second comer might so closely imitate Sir Toby Belch or Malvolio as to infringe, but it would not be enough that for one of his characters he cast a riotous knight who kept wassail to the discomfort of the household, or a vain and foppish steward who became amorous of his mistress. These would be no more than Shakespeare's "ideas" in the play, as little capable of monopoly as Einstein's Doctrine of Relativity, or Darwin's theory of the Origin of the Species. It follows that the less developed the characters, the less they can be copyrighted; that is the penalty an author must bear for marking them too indistinctly.

45 F.2d 119, 121 (7th Cir. 1930).

(48.) 724 F.2d 327 (2d Cir. 1983).

(49.) Id. at 334 ("It is because of that association, theidentification of the toy car with its source, Warner's televisionseries, that the toy car is bought by the public. That is enough [for aninfringement claim against an imitator].").

(50.) See, e.g., DC Comics Inc. v. Unlimited Monkey Bus., 598 F.Supp. 110 (N.D. Ga. 1984); DC Comics Inc. v. Filmation Assocs., 486 F.Supp. 1273 (S.D.N.Y. 1980).

(51.) Filmation, 486 F. Supp. at 1276-77. For criticism of thesedecisions, see Valerie McConnell, The Expansion of Trademark Rights inGraphic Characters and the Need for a Trademark Misuse Defense (workingpaper 2010) (on file with author).

(52.) 468 F.3d 405 (6th Cir. 2006).

(53.) Id. at 405. Lanard originally produced and sold a toy vehiclecalled the "Mudslinger," which was modeled after the Humvee,including a similar grille design. In fact, the box for the Mudslingerreferred to the toy vehicle as a "Hyper Humvee." Aftercorrespondence with AM General, however, Lanard agreed to stop using the"Humvee" name on its toys. Lanard continued to manufacture theMudslinger toy, but later ceased selling the Mudslinger and beganproducing another vehicle with a similar design called "THE CORPS!ATK." That toy was the subject of the lawsuit. Id. at 411.

(54.) Id. at 414.

(55.) Id. at 413.

(56.) Id. at 405. The court also rejected Lanard'sfunctionality defense, finding that "the plain appearance of thevehicle shows that the elements which comprise its trade dress areinherently nonfunctional." Id. at 417. We don't know what"inherently non-functional" means either.

(57.) This explains the court's findings that:

 The factors of "marketing channels used" and "likelihood of expansion of product lines" do not strongly favor either party .... There appears to be no evidence of how the marketing for either product might overlap, and while General Motors states it has considered making Hummer toys, there is no real proof that the company is seriously considering this possibility.

Id. at 414. It makes the finding on relatedness of goods ratherodd, however.

(58.) Complaint, Paccar Inc. v. Malibu Int'l Ltd. (W.D. Wash.Apr. 8, 2009), available athttp://seattletrademarklawyer.com/storage/Paccar%20Inc.%20v.%20Malibu%20International%20Ltd.%20-%20Complaint.pdf.

(59.) Id. at [paragraph] 16.

(60.) Graeme B. Dinwoodie & Mark D. Janis, Confusion Over Use:Contextualism in Trademark Law, 92 Iowa L. REV. 1597, 1654 (2007).

(61.) See, e.g., Rescuecom Corp. v. Google Inc., 562 F.3d 123 (2dCir. 2009); Am. Airlines, Inc. v. Yahoo!, Inc., No. 4:08-CV-626-A, 2009WL 381995 (N.D. Tex. Jan. 16, 2009); Google, Inc. v. Am. Blind &Wallpaper Factory, Inc., No. C 03-5340 JF (RS), 2007 WL 1159950 (N.D.Cal. April 18, 2007); Gov't Emps. Ins. Co. v. Google, Inc., 330 F.Supp 2d 700 (E.D. Va. 2004).

(62.) As one of us has explained, aside from the "trademarkuse" issue, these claims ought to fail because they do not allegeactionable infringement under the Lanham Act. See Mark P. McKenna,Trademark Use and the Problem of Source, 2009 U. ILL. L. REV. 773,819-21 (noting that these cases fail to allege confusion about thesource of the search engines' services).

(63.) See, e.g., Rescuecom, 562 F.3d at 129.

(64.) 174 F.3d 1036 (9th Cir. 1999).

(65.) Id. at 1066.

(66.) Brookfield, 174 F.3d at 1062 ("[I]t is difficult to saythat a consumer is likely to be confused about whose site he has reachedor to think that Brookfield somehow sponsors West Coast's website.").

(67.) Id. (emphasis added).

(68.) 818 F.2d 254 (2d Cir. 1987).

(69.) Id. at 260.

(70.) Id. at 259. Importantly in that case, Mobil had used the namePegasus in addition to its flying horse logo in its own oil business,Id. at 256.

(71.) 141 F.3d 188 (5th Cir. 1998).

(72.) Id. at 204.

(73.) 221 F.2d 464 (2d Cir. 1955).

(74.) Id. at 466.

(75.) 4 J. THOMAS MCCARTHY, MCCARTHY ON TRADEMARKS AND UNFAIRCOMPETITION [section] 23:7, at 23-54 (4th ed. 2010).

(76.) Au-Tomotive Gold Inc. v. Volkswagen of America, Inc., No.08-16005, 2010 WL 1794018, at *5 (9th Cir. May 6, 2010).

(77.) See William McGeveran, Rethinking Trademark Fair Use, 94 IOWAL. REV. 49, 66-71 (2008) (describing the ways courts tend to subjecttrademark defenses to the condition that the use not cause confusion).

(78.) 2010 WL 1794018, at *4-5.

(79.) Id. at *1.

(80.) Id. at *4. This is consistent with Professor McCarthy'sacknowledgement that the real concern in post-sale confusion cases isthat "consumers could acquire the prestige value of the senioruser's product by buying the copier's cheap imitation."MCCARTHY, supra note 75, [section] 23:7, at 23-54.

(81.) Au-Tomotive Gold, 2010 WL 1794018, at*5.

(82.) The result is problematic as a pure doctrinal matter as well.The first-sale doctrine is a defense to a trademark cause of action. Anda defense that does not apply whenever infringement is shown--which iswhat the court holds--isn't much of a defense. See KP PermanentMake-up, Inc. v. Lasting Impression I, 543 U.S. 111, 121 (2004)(rejecting a parallel conclusion by the Ninth Circuit about the fair usedefense).

(83.) This is the intuition expressed by Justice Breyer in the oralargument for American Needle, Inc. v. National Football League:

 And I just heard you say that.., you want the Red Snx to compete in selling T-shirts with the Yankees; is that right? [MR. NAGER: The ability to compete. Yes.] Yes. Okay. I don't know a Red Sox fan who would take a Yankees sweatshirt if you gave it away. I mean, I don't know where you're going to get your expert from that is going to say there is competition ... between those two products. I think they would rather--they would rather wear a baseball, a football, a hockey shirt.

Transcript of Oral Argument at 17, 130 S. Ct. 2201 (No. 08-661),available at http://www.supremecourt.gov/oral_argument_transcripts/08-661.pdf.

(84.) See Deere & Co. v. Farmhand, Inc., 560 F. Supp. 85, 96-98(S.D. Iowa 1982), aff'd, 721 F.2d 253 (8th Cir. 1983) (finding thecolor green functional for front-end loaders because farmers prefer thecoloring of their loaders to that of their tractors).

(85.) See Wallace Int'l Silversmiths. Inc. v. Godinger SilverArt Co., 916 F.2d 76, 80-81 (2d Cir. 1990) (finding the design ofWallace's GRANDE BAROQUE line functional because baroque designelements are necessary to compete in the market for baroque silverware);see also Keene Corp. v. Paraflex Indus., 653 F.2d 822, 826 (3d Cir.1981) (finding the design of a lighting fixture functional because itwas designed to match the architecture of the building in which it wasmounted).

(86.) Au-Tomotive Gold, 2010 WL 1794018, at *6. This statementwould be accurate, and unproblematic, if it were confined to cases inwhich the defendant's use caused confusion about the actual sourceof a product. Our ability to sell "M&M" candies for lessthan Mars, Inc., for example, should not lead courts to conclude thatour use of the M&M mark does not infringe when applied to candies.But it's much different to say that trademark law protects Marsfrom competition from an ice cream shop that buys genuine M&M'sand mixes them into its ice cream.

(87.) See 15 U.S.C. [section] 1125(c).

(88.) See Dogan & Lemley, supra note 41. The other of us doubtsthe empirical premises on which this argument rests; cf. Tushnet, supranote 12 (exploring the foundations of blurring claims in cognitivescience). But that's an argument for another day.

(89.) See Jerre B. Swann, An Intuitive Approach to Dilution, 89TRADEMARK REP. 907 (1999); Jerre B. Swann& Theodore H. Davis, Jr.,Dilution, An Idea Whose Time Has Gone; Brand Equity as ProtectableProperty, The New/Old Paradigm, 84 TRADEMARK REP. 267 (1994).

(90.) Moseley v. V Secret Catalogue, Inc., 537 U.S. 418, 429(2003). Courts are not alone in this regard. Though he is careful tonote that an investment in goodwill is not enough by itself to createtrademark rights, Tom McCarthy, the author of the leading trademarktreatise, argues that "[t]he creation of value in a trademarkrequires 'the expenditure of great effort, skill and ability'and a competitor should not be permitted to take a 'free ride"on the trademark owner's good will and reputation." 1MCCARTHY, supra note 75, [section] 2:30, at 2-54.

(91.) Ty Inc. v. Perryman, 306 F.3d 509, 512 (7th Cir. 2002).

(92.) 306 F.2d 433 (5th Cir. 1962).

(93.) Id. at 437-38. But see Ringling Bros.-Barnum & BaileyCombined Shows, Inc. v. Utah Div. of Travel Dev., 170 F.3d 449 (4th Cir.1999) (rejecting a similar claim: that the Utah slogan "GreatestSnow on Earth" diluted the plaintiff's slogan "GreatestShow on Earth").

(94.) Anheuser-Busch, Inc. v. VIP Prods., LCC, 666 F. Supp. 2d 974(E.D. Mo. 2008). Amazingly, the court also found for Anheuser-Busch onlikelihood-of-confusion grounds after crediting Anheuser-Busch'ssurvey, which purported to show "that 30.3% of those surveyed hadthe mistaken belief that 'Buttwiper" is made or put out by orwith the approval or sponsorship of the maker of'Budweiser'--Plaintiff--or that there is a businessrelationship between the maker of 'Budweiser' and the maker of'Buttwiper.'" Id. at 983.

(95.) See Sarah Netter, The North Face vs. The South Butt:Entrepreneurial Teen Undaunted by Lawsuit Threat, ABCNEWS, Oct. 1, 2009,http://abcnews.go.com/Business/teens-south-buttapparel-irks-north-face/story?id=8712101.

(96.) Trade-marks Act, R.S.C., ch. T-13, [section] 22(1) (2010)("No person shall use a trade-mark registered by another person ina manner that is likely to have the effect of depreciating the value ofthe goodwill attaching thereto.").

(97.) Clairol Int'l Corp. v. Thomas Supply & Equip. Co.,[1968] 2 Ex. C.R. 552.

(98.) Case C-487/07, L'Oreal SA v. Bellure NV, 2009 ECJEUR-Lex LEXIS 532.

(99.) Id.

(100.) As Dev Gangjee and Robert Burrell state:

 It must be remembered that the [Misleading and Comparative Advertising Directive ("MEAD")] exists to prevent a blanket prohibition on comparative advertising and the default position under the MeAD is that such advertising is allowed. By adopting an expansive reading of Art. 3a(l)(h), the "imitations and replicas" exclusion, the ECJ may have insulated reputed marks from a range of competitive comparisons.

Dev Gangjee & Robert Burrell, A Brief Note on L'Oreal andthe Prohibition on Free Riding 14 (Dec. 9, 2009) (unpublishedmanuscript), available at http://ssrn.com/abstract=1491402.

(101.) Though we rather doubt they can.

(102.) On the merits of these various claims, see, for example,Lemley & McKenna, supra note 8 (sponsorship or affiliation claims);Dogan & Lemley, supra note 41 (merchandising); Stacey L. Dogan &Mark A. Lemley, Trademarks and Consumer Search Costs on the Internet, 41Hous. L. REV. 777 (2004) (keyword advertising); Eric Goldman,Deregulating Relevancy in Internet Trademark Law, 54 EMORY L.J. 507(2005); Michael Grynberg, The Road Not Taken: Initial InterestConfusion, Consumer Search Costs, and the Challenge of the Internet, 28SEATTLE U. L. REV. 97 (2004) (initial interest confusion); AlexKozinski, Trademarks Unplugged, 68 N.Y.U.L. REV. 960 (1993) (luxurygoods and post-sale confusion).

(103.) Ringling Bros.-Barnum & Bailey Combined Shows, Inc. v.Utah Div. of Travel Dev., 170 F.3d 449, 459 (4th Cir. 1999) ("[W]esimply cannot believe that, as a general proposition, Congress couldhave intended, without making its intention to do so perfectly clear, tocreate property rights in gross, unlimited in time (via injunction),even in 'famous' trademarks.").

(104.) To that end, the Supreme Court has held:

 There is no such thing as property in a trade-mark except as a right appurtenant to an established business or trade in connection with which the mark is employed. The law of trademarks is but a part of the broader law of unfair competition; the right to a particular mark grows out of its use, not its mere adoption; its function is simply to designate the goods as the product of a particular trader and to protect his good will against the sale of another's product as his; and it is not the subject of property except in connection with an existing business.

United Drug Co. v. Theodore Rectanus Co., 248 U.S. 90, 97 (1918).

(105.) See Swann & Davis, supra note 89, at 269 (recognizingthat courts fudge the likelihoodof-confusion analysis in cases in whichthe defendant is perceived to have taken a free ride). Swann and Davisargue that "[g]iven the current economic function and acceptance oftrademarks, and recognition of their value and performance in the marketplace," evidence of the defendant using the mark as a"triggering mechanism" to capitalize on the good will createdby the plaintiff "alone should mandate relief whether or notconfusion is present." Id.

(106.) Sections III.A and B are adapted from McKenna, supra note 7,and Lemley & McKenna, supra note 8.

(107.) The few studies that have found some effect on global brandassessments have involved umbrella branding of extremely closely relatedproducts. In one study, for example, Tulin Erdem estimated a model usingscanner data regarding purchases of umbrella-branded toothbrushes andtoothpaste. Tulin Erdem, An Empirical Analysis of Umbrella Branding, 35J. MARKETING RES. 339, 347 (1998). Erdem concluded that variance in thequality of toothbrushes given away as free sampies from the owner of aknown toothpaste brand had some cross-category effects (i.e., consumersupdated their quality expectations of the toothpaste and bought less ofit). But even here the effects were "small in magnitude," andthat was in a study in which the brand owner explicitly tied theextension product to the core product. Any uses that would fall in thiscategory of sufficiently closely related uses almost certainly fallwithin the zone we describe in Irrelevant Confusion. Lemley &McKenna, supra note 8.

(108.) See Sanjay Sood & Kevin Lane Keller, The Effects ofBrand Name Structure and Product Experience on Brand ExtensionEvaluations and Parent Brand Dilution (unpublished manuscript) (draft onfile with authors). Sood and Keller tested extensions of the Pepsi andTropicana brands to vitamin fortified cola and sodium free orange juice.Subjects in their pretest regarded vitamin fortified cola as similar toPepsi but dissimilar to Tropicana, and sodium free orange juice asdissimilar to Pepsi and similar to Tropicana. Sood and Keller testedthese similar and dissimilar extensions under family branding (Tropicanacola) and sub-branding (Quencher by Tropicana cola) conditions. Theyfound that "dilution effects in parent brand evaluations areevident only when consumers I) have a negative experience with a brandextension that 2) has a high degree of similarity (e.g., in the case ofa family-branded line extension)." Id. at 21. Sood and Keller alsofind that "sub-branding strategies can offer the 'best of bothworlds" by both enhancing extension evaluations and protecting theparent brand from any unwanted negative feedback." Id. at 22.

(109.) See infra note 146.

(110.) Thus, for example, Joseph Chang found that the general brandimage of the parent brand of Sprite products was diluted by both of twounfavorable extension products: Sprite orangeades and Sprite washing-upliquids. Joseph W. Chang, Will a Family Brand Image be Diluted by anUnfavorable Brand Extension? A Brand Trial-Based Approach, 29 ADVANCESCONSUMER RES., 299, 302 (2002). At the same time, however, neitherunfavorable extension diluted the image of Sprite lemonades, theoriginal product offered under the parent brand. Id.

(111.) Congruence here is defined in terms of keeping with thedominant concept of the brand, which is conceived of as primarilyfunctional, symbolic, or experiential. Congruence differs from productcategory-related effects in that an extension could be seen asincongruent even when it is in a similar product class. Consumers, forexample, might regard a new Rolls-Royce economy car as incongruent withthe Rolls-Royce brand image, even though the new car is in the samebroad category as Rolls-Royce luxury vehicles. Helge Thorbjornsen, Brandextensions: brand concept congruency and feedback effects revisited, 14J. PRODUCT & BRAND MGMT. 250, 250-51 (2005); see also Henrik Sjodin& Fredrik Torn, When communication challenges brand associations: aframe work for understanding consumer responses to brand imageincongruity, 5 J. CONSUMER BEHAV. 32, 38 (2006).

(112.) Thorbjornsen, supra note 111.

(113.) Sjodin & Torn, supra note 111.

(114.) Id.

(115.) See Stephen J. Hoch, Product Experience Is Seductive, 29 J.CONSUMER RES. 448, 451 (2002) ("Using a simply associative learningprocedure, [researchers] showed that, in a few trials, people learnbrand associations that later block the learning of new predictiveattribute associations."). Even Jacob Jacoby, perhapsdilution's biggest proponent, admits that truly well-known marksare essentially unshakable. Maureen Morrin & Jacob Jacoby, TrademarkDilution: Empirical Measures for an Elusive Concept, 19 J. PUB.POL'Y & MARKETING 265, 274 (2000) ("It appears that verystrong brands are immune to dilution because their memory connectionsare so strong that it is difficult for consumers to alter them or createnew ones with the same brand name"). There is abundant evidenceoutside the branding context of the robustness of initial judgments.See, e.g., Jon D. Hanson & Douglas A. Kysar, Taking BehavioralismSeriously: The Problem of Market Manipulation, 74 N.Y.U.L. REV. 630,646-54 (1999) (discussing a number of empirical demonstrations of thepersistence of initial judgments, even in the face of contradictory orambiguous hard data). Even conscious consumers who try to reason throughadditional information are unlikely to change their perceptions;attempts at rationalization may actually serve to increase confidence ina faulty intuitive judgment, a phenomenon known as confirmation bias.See id. at 647-50, 660-62; Nicholas Epley & Thomas Gilovich, TheAnchoring-and-Adjustment Heuristic: Why the Adjustments areInsufficient, 17 PSYCHOL. SCI. 311, 312 (2006) ("[P]eople evaluatehypotheses by trying to confirm them.").

(116.) See Mark A. Lemley, Ex Ante Versus Ex Post Justificationsfor Intellectual Property, 71 U. CHI. L. REV. 129, 146 (2004); JustinHughes, "Recoding" Intellectual Property and OverlookedAudience Interests, 77 TEX. L. REV. 923, 961 (1999). For a discussion ofthe marketing literature on why these well-known marks are essentiallyimpervious to harm, see McKenna, supra note 7, at 105. A significantupshot of this conclusion is that the preemption and free-ridingarguments are particularly important in the context of well-known brandssince the feedback arguments are even weaker for those brands.

(117.) Daniel J. Howard et al., The Effects of Brand NameSimilarity on Brand Source Confusion: Implications for TrademarksInfringement, 19 J. PUB. POL'Y & MARKETING 250, 257--58 (2000).Even these findings are probably overstated. The authors of the studytested confusion by showing participants a variety of goods, includingtwo different bottles of drain cleaner and two different containers ofcar wax. They varied the difference between the two brands such that thesecond sometimes was similar in meaning (Hurricane vs. Cyclone) andsometimes in sound (Hurricane vs. Hurri-Drain). The authors thenmeasured confusion about common source by telling study participants:

 The first set of products you reviewed yesterday included a brand of drain cleaner (car wax). The second set of products you just reviewed also included a brand of drain cleaner (car wax). How likely or unlikely is it that those two brands of drain cleaner (car wax) were made by the same company?"

Id. They then scored the responses on a seven-point scale rangingfrom "very likely" (1) to "very unlikely" (7). Thisformulation likely led to overstated levels of confusion generally,since the question seems to have primed respondents to consider commonsource. Confusion also was likely overstated because the authors usedthe same trade dress for all the test objects (both drain cleaners, forexample). See id.

(118.) If Thorbjornsen is right that high familiarity correlateswith higher involvement, see Thorbjornsen, supra note 111, then it seemslikely that consumers are less likely to be confused about third-partyuses that are similar to familiar brands only in sight or sound.

(119.) See Lemley & McKenna, supra note 8, at 416.

(120.) Brand alliances are "partnership[s] between twoentities in which efforts are combined for a common interest or toachieve a particular aim." Nicole L. Votolato & H. Rao Unnava,Spillover of Negative Information on Brand Alliances, 16 J. CONSUMERPSYCHOL. 196, 196 (2006). Such partnerships can take many forms, but thetwo most common forms are joint promotions (e.g., McDonald's usingKung Fu Panda toys in its Happy Meals) and co-branding arrangements(e.g., Edy's Loaded Cookie Dough Ice Cream with Nestle Toll Housecookie dough).

(121.) Id. at 198.

(122.) Previous research suggested that consumers might reactdifferently to different types of negative information--informationabout competence, on the one hand, and moral misdeeds on the other. Id.at 197 (citing Tom J. Brown & Peter A. Dacin, The Company and theProduct: Corporate Associations and Consumer Product Responses, 61 J.MARKETING 68 (1997); Bogdan Wojciszke et al., Effects of InformationContent and Evaluative Extremity on Positivity and Negativity Biases, 64J. PERSONALITY & Soc. PSYCHOL. 327 (1993)). Specifically, thisearlier research suggested that consumers react more negatively tocompetence-based information than moral failures when the target of theinformation is a company; just the reverse is true when the target ofthe information is a person. Id. at 197.

(123.) Id. at 201. These findings, as the authors also note, mayhelp explain why spillover effects are not frequently reported inpractice. Id. It is also worth emphasizing that respondents in thisstudy were told explicitly that the partner had a relationship with thethird party about which the negative information was provided. Thus,there was no ambiguity about affiliation--respondents understood thatthe partner was affiliated with the third party. Hence, spillover isunlikely to occur absent some information--additional information,beyond the mere fact of association-demonstrating the host brand'sspecific culpability. In other words, as Votolato and Unnava note, theevidence suggests that "'consumers do not routinely blame ahost brand for its partner's mistakes.'" Id. at 198.

(124.) See Kevin Lane Keller & David A. Aaker, The Effects ofSequential Introduction of Brand Extensions, 29 J. MARKETING RES. 35, 47(1992) (discussing the factors that contribute to a successful brandextension).

(125.) While in one study Keller and Aaker found that extensionsfrom high-quality brands may be evaluated favorably even when they aresomewhat more remote--that is, high-quality core brands "stretchfarther"--the relatively dissimilar products in that study werestill quite close to those offered under the core brand. See id. at40-44 (testing extensions deemed close, medium, and far from the corebrand product, where ice cream was the "far" extension of abrand known for potato chips); Sood & Keller, supra note 108.

(126.) Keller & Aaker, supra note 124, at 45.

(127.) David A. Aaker & Kevin Lane Keller, Consumer Evaluationsof Brand Extensions, 54 J. MARKETING 27, 30 (1990).

(128.) Id.

(129.) Id. (discussing the importance of product classes whendetermining complements).

(130.) Id.

(131.) Id.

(132.) See id. Perceived expertise and trustworthiness are highlycorrelated and may depend on the perception of previous extensions. Theeffect of previous extensions on new extension evaluation appears todepend more on the success of the previous extensions than the relativesimilarity of the intervening extensions. Aaker and Keller found nodifferences in perceived credibility (and presumably in evaluations ofproposed extensions) based on fit between an intervening extension andthe core brand. See id.

(133.) Keller & Aaker, supra note 124, at 36-37. Concrete brandattribute associations relate to tangible product characteristics, andabstract brand attribute associations relate to intangible imagecharacteristics. Id.

(134.) Id.

(135.) Valarie A. Zeithaml, Consumer Perceptions of Price,Quantity, and Value: A Means-End Model and Synthesis of Evidence, 52 J.MARKETING 2, 7 (1988).

(136.) Id.

(137.) Aaker & Keller, supra note 127, at 29.

(138.) In the studies we report, the brand extensions are in factmade by the trademark owner, not by unauthorized third parties.

(139.) See Laura R. Bradford, Emotion, Dilution, and the TrademarkConsumer, 23 BERKELEY TECH. L.J. 1227, 1264-66 (2008).

(140.) Id. It is worth emphasizing that while third parties benefitfrom the familiarity of the known mark in these circ*mstances, thatthird party's very use may increase the brand's familiarityand benefit the mark owner in its market as well. See McKenna, supranote 7, at 110-12.

(141.) Akshay R. Rao & Robert W. Ruekert, Brand Alliances asSignals of Product Quality, 36 SLOAN MGMT. REV. 87 (1994); Bernard L.Simonin & Julie A. Ruth, Is a Company Known By the Company It Keeps?Assessing the Spillover Effects of Brand Alliances on Consumer BrandAttitudes, 35 J. MARKETING RES. 30 (1998).

(142.) David O. James et al., Does the Tail Wag the Dog? BrandPersonality in Brand Alliance Evaluation, 15 J. PRODUCT & BRANDMGMT. 173, 174 (2006).

(143.) Id. at 175.

(144.) Id.

(145.) Id. at 176.

(146.) As we noted before, in Aaker and Keller's study,successful brand extensions increased evaluations of later extensionsand of the core brand itself, at least when the core brand was ofaverage quality. Aaker & Keller, supra note 127, at 43 (also notingthat successful extensions had no impact on high-quality core brands).

(147.) Id. at 46; Erdem, supra note 107, at 347. Notably, whileunsuccessful intervening extensions affected consumer evaluations offuture extensions, they did not affect evaluations of the core brand.Keller & Aaker, supra note 124, at 46. And even this risk regardingfuture evaluations appeared significant only for moderate-quality corebrands. Unsuccessful extensions had no impact on evaluation ofextensions by high-quality brands. An interesting parallel finding wasthat an unsuccessful extension, even when it affects credibility andprevents the core brand from expanding to less similar products, doesnot appear to prevent the core brand from "backtracking" andlater introducing a more similar extension. Id. at 48. This may bebecause subjects tended to find the core brand owner equally credibleeven after receiving information about a brand extension they regardedas a bad fit. Id. at 45.

(148.) See supra note 110.

(149.) See Girish N. Punj & Clayton L. Hillyer, A CognitiveModel of Consumer-Based Brand Equity for Frequently Purchased Products:Conceptual Framework and Empirical Results, 14 J. CONSUMER PSYCHOL. 124,125 (2004) (stating that consumers tend to rely predominately onattitudes toward a brand when evaluating new products): see alsoBradford, supra note 139.

(150.) Whether it is a legal injury is a question to which we turnin the next Part.

(151.) Among countless examples, consider Dell, United, American,Delta, Apple, Amazon, and Visa (notwithstanding Judge Kozinski'sclaim in Visa Int'l that there is only one Visa). See VisaInt'l Serv. Ass'n v. JSL Corp., 610 F.3d 1088 (9th Cir. 2010).

(152.) Lemley & McKenna, supra note 8, at 438-43.

(153.) Id.

(154.) The "endowment effect" refers to the difference invalue people attach to goods when they own them as compared to when theyare considering purchasing them. People are reluctant to part with theirproperty, and as a result, the price at which they are willing to sellit generally far exceeds the amount that others are willing to pay toacquire it. See Russell Korobkin, The Endowment Effect and LegalAnalysis, 97 NW. U. L. REV. 1227, 1231-35 (2003). For a demonstration ofthe endowment effect in the context of creative works, see ChristopherJ. Buccafusco & Christopher Jon Sprigman, Valuing IntellectualProperty: An Experiment, 91 CORNELL L. REV. (forthcoming 2010).

(155.) See generally David A. Armor & Shelley E. Taylor, WhenPredictions Fail: The Dilemma of Unrealistic Optimism, in HEURISTICS ANDBIASES: THE PSYCHOLOGY OF INTUITIVE JUDGMENT 334, 334 (Thomas Gilovichet al. eds., 2002) ("By a number of metrics and across a variety ofdomains, people have been found to assign higher probabilities to theirattainment of desirable outcomes than either objective criteria orlogical analysis warrants."). See also Christopher J. Buccafusco& Christopher Jon Sprigman, The Creativity Effect, 78 U. CHI. L.REV. (forthcoming 2010), available at http://ssrn.com/abstract=1647009(finding that creators substantially overvalue their works as comparedto mere owners and attributing the difference in substantial part tooptimism bias of the creators).

(156.) Trademarks are devices used to brand goods; characters in abook usually don't serve that purpose. Indeed, even titles ofmovies and books are generally not treated as trademarks unless thetitle is common to multiple works in a series. See Herbko Int'l,Inc. v. Kappa Books, Inc., 308 F.3d 1156, 1162 n.2 (Fed. Cir. 2002)("While titles of single works are not registrable, they may beprotected under section 43(a) of the Lanham Act upon a showing ofsecondary meaning."); Sugar Busters LLC v. Brennan, 177 F.3d 258,267-69 (5th Cir. 1999) (viewing single book titles as descriptive of thecontents, and requiring proof of secondary meaning); Estate of Jenkinsv. Paramount Pictures Corp., 90 F. Supp. 2d 706, 710 (E.D. Va. 2000),aff'd, 7 F. App'x 270 (4th Cir. 2001) ("[T]itles ofexpressive works are treated differently from other trademarks, in thattitles, even if suggestive, arbitrary, or fanciful, nonetheless requiresecondary meaning to receive trademark protection, while othersuggestive, arbitrary, and fanciful marks do not.").

(157.) See William P. Kratzke, Normative Economic Analysis ofTrademark Law, 21 MEM. ST. U. L. REV. 199, 223 (1991) (arguing thatdeclaring competitive behavior to be "free riding" is aconclusory epithet, not a workable economic principle).

(158.) We explain this in detail in our prior work. Lemley &McKenna, supra note 8, at 448-49.

(159.) See 15 U.S.C. [section] 1114(1) (2006) (permitting court toaward profits if infringement is intentional).

(160.) See 17 U.S.C. [section] 106(2) (2006) ("Subject tosection 107 through 122, the owner of copyright under this title has theexclusive right[] ... (2) to prepare derivative works based upon thecopyrighted work."). The Copyright Act defines a derivative work as"a work based upon one or more preexisting works, such as atranslation, musical arrangement, dramatization, fictionalization,motion picture version, sound recording, art reproduction, abridgment,condensation, or any other form in which a work may be recast,transformed, or adapted." Id. [section] 101.

(161.) See Sony Corp. v. Universal City Studios, Inc., 464 U.S.417, 429 (1984) (noting that a copyright is "intended to motivatethe creative activity of authors and inventors by the provision of aspecial reward").

(162.) Paul Goldstein, Derivative Rights and Derivative Works inCopyright, 30 J. COPYRIGHT SOC'Y 209, 216 (1983); see also PAULGOLDSTEIN, COPYRIGHT [section] 5.3 (2d ed. Supp. 2004) (repeating theanalysis).

(163.) William M. Landes & Richard A. Posner, An EconomicAnalysis of Copyright Law, 18 J. LEGAL STUD. 325, 353-57 (1989).

(164.) Edmund W. Kitch, The Nature and Function of the PatentSystem, 20 J.L. & ECON. 265, 275-76 (1977); see also E Scott Kieff,Property Rights and Property Rules for Commercializing Inventions, 85MINN. L. REV. 697 (2001). For an explanation of why thecommercialization argument falls short in intellectual property moregenerally, see Lemley, supra note 116, at 132-41.

(165.) Stewart E. Sterk, Rhetoric and Reality in Copyright Law, 94MICH. L. REV. 1197, 1215-16 (1996). Even Landes and Posner ultimatelyaccept that the derivative right is not necessary to give adequateincentives to create new works of authorship. Landes & Posner, supranote 163, at 354.

(166.) This criticism applies with almost equal force to manyreproduction cases in which the defendant is deemed to have copieddespite having created a new work that differs significantly from theoriginal work.

(167.) There are some alternative accounts of the derivative right.See, e.g., Michael Abramowicz, A Theory of Copyright's DerivativeRight and Related Doctrines, 90 MINN. L. REV. 317, 322 (2005) (arguingthe derivative right "is best understood not solely as a means offurthering the incentive to create works, but more significantly as ameans of providing an author control over the release of adaptations andlimiting the production of adaptations that would be close substitutesfor one another" and thereby reducing redundancy). But the onlyalternative account with any significant traction is one based on moralrights. On that argument, the right to control derivative uses flowsfrom the author's interests in the integrity of her work. This is,as one of us has recognized, a problematic justification even for thederivative right, not least because moral rights arguments focus onauthors, while the derivative right is alienable and therefore may wellbe held by someone other than the author. Mark A. Lemley, The Economicsof Improvement in Intellectual Property Law, 75 TEX. L. REV. 989,1031-34 (1997). And the moral rights argument would be particularlyunpersuasive in the trademark context, since moral rights are thought toderive from the intimate connection an author has with her work. The"authors" of trademarks, which generally are corporateentities, have no human dignity at stake when others use their marks. Wediscuss this moral claim in more detail infra notes 180-192 andaccompanying text.

(168.) Lemley, supra note 167, at 993-97.

(169.) See Ralph S. Brown, Jr., Advertising and the PublicInterest: Legal Protection of Trade Symbols, 57 YALE L.J. 1165, 1200-01(1948) (disapproving of cases in which "admiration for innovationobscum[s] the soundness of rules designed to foster free and easycompetition" and arguing that "the only interests in tradesymbols worth protecting are those against loss of sales or loss ofreputation"); Rochelle Cooper Dreyfuss, Expressive Genericity:Trademarks as Language In the Pepsi Generation, 65 NOTRE DAME L. REV.397, 399 (1990) ("[T]here is little need to create economicincentives to encourage businesses to develop a vocabulary with which toconduct commerce."). This isn't to say that there aren'tany reasons to want new trademarks--at the very least, trademarks mightcreate a placebo effect that enhances the effectiveness of certainproducts. But the question is not whether we want new trademarks:it's whether we need protection to incentivize their creation.

(170.) Ambrosia Chocolate Co. v. Ambrosia Cake Bakery, 165 F.2d693, 697 (4th Cir. 1947) ("[A] man of ordinary intelligence couldeasily devise a score of valid trade-marks in a short period oftime.").

(171.) This might mean granting trademark rights to the marketingfirms that thought up the brand names, rather than to the companies thatused them. Even more certainly, we would not endow companies withownership of marks created by the public, as many courts have. SeeNat'l Cable Television Ass'n v. Am. Cinema Editors, Inc., 937F.2d 1572, 1577-78 (Fed. Cir. 1991) ("Moreover, even without usedirectly by the claimant of the rights, the courts and the Boardgenerally have recognized that abbreviations and nicknames of trademarksor names used only by the public give rise to protectable rights in theowners of the trade name or mark which the public modified. Such publicuse by others inures to the claimant's benefit and, where thisoccurs, public use can reasonably be deemed use 'by' thatparty in the sense of a use on its behalf.") (footnote omitted);see also Johnny Blastoff, Inc. v. L.A. Rams Football Co., 188 F.3d 427,434 (7th Cir. 1999) (quoting Nat'l Cable Television, 937 F.2d at1577); Volkswagenwerk AG v. Hoffman, 489 F. Supp. 678, 681 (D.S.C. 1980)(recognizing VW's rights in "Bug" based on public usageof the nickname without VW's protest); Coca-Cola Co. v. Busch, 44F. Supp. 405, 407 (E.D. Pa. 1942) (finding the public use of"co*ke" to refer to Coca-Cola's soft drink sufficient tocreate rights for Coca-Cola in that term); Am. Stock Exch., Inc. v. Am.Express Co., 207 U.S.P.Q. 356, 362-64 (T.T.A.B. 1980) (attributing toAmerican Express rights in "AMEX" based on public use of thatdesignation to denote American Express); Norac Co. v. OccidentalPetroleum Corp., 197 U.S.P.Q. 306, 315 (T.T.A.B. 1977) (earlier use of"OXY" by public determined priority); Pieper v. PlayboyEnters., 179 U.S.P.Q 318, 320 (T.T.A.B. 1973) (recognizingPlayboy's rights in "Bunny Club").

(172.) Landes & Posner, Trademark Law: An Economic Perspective,supra note 15, at 270.

(173.) LANDES & POSNER, THE ECONOMIC STRUCTURE OF INTELLECTUALPROPERTY LAW, supra note 15, at 179; Ariel Katz, Beyond Search Costs:The Linguistic and Trust Functions of Trademarks, 2010 B.Y.U. L. REV.(forthcoming). It would be more accurate to state that protection oftrademarks encourages production of products with consistent quality,rather than high quality. While it is true that trademark protectionallows a mark owner the opportunity to reap the benefits of investmentsin quality--since consumers will know who to credit for thatquality--how much any particular mark owner actually invests in qualitydepends on the position of the relevant goods or services in the market.There are plenty of well-known trademarks that are used withconsistently low-quality products.

(174.) McKenna, supra note 7, at 115.

(175.) See Md. Stadium Auth. v. Becker, 806 F. Supp. 1236 (D. Md.1992) (deciding a case in which the owner of the new stadium of theBaltimore Orioles, Camden Yards, brought a trademark infringement actionagainst a vendor that used the "Camden Yards" mark on t-shirtsand clothing items).

(176.) See 15 U.S.C. [section] 1051(b) (2006) (permittingapplications to register on the basis of a "bona fideintention" to use a trademark in commerce).

(177.) See Ayco*ck Eng'g, Inc. v. Airflite, Inc., 560 F.3d1350, 1360 (Fed. Cir. 2009) (declining trademark protection based onmere potential to create a competing product); RealNetworks, Inc v. QSAToolWorks, LLC, No. C07-1959MJP, 2009 WL 2512407, at *4 (W.D. Wash. Aug.14, 2009) ("A party may not allege trademark protection orconfusion based on its potential to develop a competing product.").

(178.) We think a similar efficiency-based argument can be maderegarding the registration system more generally, particularly thenationwide priority feature. To begin with, many geographic expansionsdon't need a market preemption or free-riding justification becauseconsumers are likely to believe that the Crate and Barrel store openingin their town is run or controlled by the national Crate and Barrelchain. But even when consumers in the new geographic market might not befamiliar with the mark owner, registration enables that mark owner tomake the necessary investments to expand without worrying thoseinvestments will be lost. But here, too, we think parties that want toclaim priority beyond their actual use should have to actually file fora registration as a signal of real investment. And we think the DawnDonut rule plays an important role in limiting the extent to which suchregistrations actually preempt remote geographic uses. See Dawn DonutCo. v. Hart's Food Stores, Inc., 267 F.2d 358 (2d Cir. 1959)(refusing to enjoin geographically remote use absent likely consumerconfusion even when the plaintiff had a federal registration).

(179.) Cf. Md. Stadium, 806 F.Supp. at 1241 (finding liabilitydespite the fact that the plaintiff wasn't in the market because ofworries about free riding).

(180.) See Brown, supra note 169, at 1181-83; Glynn S. Lunney, Jr.,Trademark Monopolies, 48 EMORY L.J. 367 (1999); cf. Barton Beebe, Searchand Persuasion in Trademark Law, 103 MICH. L. REV. 2020 (2005) (notingthe controversy over whether advertising expenditures create irrationalbrand loyalty).

(181.) See Lemley & McKenna, supra note 8, at 439-45.

(182.) Thomas W. Merrill, Accession and Original Ownership, 1 J.LEGAL ANALYSIS 459 (2009). To be clear, Merrill argues from theprinciple of accession, which he regards as a general principle ofproperty allocation. This is not to be confused with the doctrine ofaccession, which is only one doctrine among many Merrill identifies asflowing from the principle of accession. See id. at 465-66. In additionto the doctrine of accession (which deals with mistaken improvers ofpersonal property), Merrill identifies the doctrine of increase (whichawards offspring to the owner of the mother), the doctrine of accretion(which provides a riparian landowner whose land is gradually augmentedby alluvial formations ownership of the new land), and even the adcoelum doctrine (which declares that the owner of surface land owns fromthe depths to the sky) as examples of the principle of accession. Seeid. at 7-12.

(183.) Mark A. Lemley, Property, Intellectual Property, and FreeRiding, 83 TEX. L. REV. 1031 (2005); see also Richard A. Posner,Misappropriation: A Dirge, 40 Hous. L. REV. 621, 623 (2003) ("Theonly point I want to emphasize is that the free riding that creates aneed for IP law is economically distinct from that involved in the theftof tangibles.") (emphasis added).

(184.) See, e.g., 2 WILLIAM BLACKSTONE, COMMENTARIES *404-05; EarlC. Arnold, The Law of Accession of Personal Property,, 22 COLUM. L. REV.103, 118 (1922).

(185.) See, e.g., MICHAEL HEELER, THE GRIDLOCK ECONOMY (2008);Michael A. Heller, The Tragedy of the Anticommons: Property in theTransition from Marx to Markets, 111 HARV. L. REV. 621, 623-24 (1998).

(186.) For critiques of the claim that real property must be owned,see, for example, ELINOR OSTROM, GOVERNING THE COMMONS: THE EVOLUTION OFINSTITUTIONS FOR COLLECTIVE ACTION (1990); Brett M. Frischmann, AnEconomic Theory of Infrastructure and Commons Management, 89 MINN. L.REV. 917 (2005); and Carol Rose, The Comedy of the Commons: Custom,Commerce, and Inherently Public Property, 53 U. CHI. L. REV. 711, 720-21(1986).

(187.) For critiques of this assumption, see Dogan & Lemley,supra note 41, at 478-81 (critiquing unjust enrichment justification formerchandising rights in the trademark context); Lemley, supra note 116,at 144-47; and Diane Leenheer Zimmerman, Fitting Publicity Rights intoIntellectual Property and Free Speech Theory: Sam, You Made the PantsToo Long!, 10 DEPAUL-LCA J. ART & ENT. L. 283, 307-08 (2000).

(188.) E.g., Garrett Hardin, The Tragedy of the Commons, 162 SCI.1243, 1244-45 (1968).

(189.) This may well be why, notwithstanding the ancient pedigreeof many of the doctrines on which Merrill focuses, and grounding of theprinciple of accession in the writings of early English writers likeHume and Blackstone, early IP law, and particularly early trademark law,recognized nothing like a principle of accession. Derivative rights werenot recognized by copyright law, and the right of reproduction wasinterpreted much more narrowly than it is now. And this was not becauseauthors didn't recognize the potential value of broaderrights--copyright did not extend to translations, for example, eventhough translations were widely used and clearly contemplated by authorsand the creators of the copyright system. And trademark rights also weretraditionally much narrower, explicitly limited to directly competinguses. See Mark P. McKenna, The Normative Foundations of Trademark Law,,82 NOTRE DAME L. REV. 1839 (2007).

(190.) May, but not must. We can think of plenty of examples inwhich simultaneous use of an intangible asset by multiple partiesincreases, rather than decreases, its value. Facebook, for instance.

(191.) As Ralph Brown put it, "Competition is copying."Ralph S. Brown, The Joys of Copyright, 30 J. COPYRIGHT SOC'Y 477,481 (1983); see also Robert C. Denicola, Freedom to Copy, 108 YALE L.J.1661, 1661 (1999) ("[L]aws that restrain copying ... restraincompetition."). See generally Peter Jaffery, Merchandising and theLaw of Trade Marks, 3 INTELL. PROP. Q. 240 (1998) (noting that trademarklaw does not support a general merchandising right).

(192.) Lemley, supra note 116, at 144.

(193.) Id. at 145.

(194.) See Lanham Act of 1946 [section] 32, 15 U.S.C. [section]1114 (2006) (prohibiting "reproduction, counterfeit, copy, orcolorable imitation of a registered mark"); MCCARTHY, supra note76, [section] 2:9, at 2-16 to -18. The first-use rule has historicalpedigree. See Am. Washboard Co. v. Saginaw Mfg. Co., 103 F. 281, 287(6th Cir. 1900) ("It is the party who uses [a designation] first asa brand for his goods, and builds up a business under it, who isentitled to protection, and not the one who first thought of using it onsimilar goods, but did not use it. The law deals with acts, notintentions.") (quoting George v. Smith, 52 F. 830, 832 (C.C.N.Y.1892)).

(195.) Merrill, supra note 182, at 482-88.

(196.) This is a big "if." Merrill does not identify aprinciple by which this decision is to be made; rather, he more or lesssimply asserts that "[a]ccession ... awards ownership based onstatus--the status of owning something prominently connected to thedisputed object" id. at 481. But what is connected to what? Merrillargues that certain solutions have a "natural" prominence andthat ownership runs from ownership of "big" things to"small" things. Merrill, supra note 182, at 25-26. But in thetrademark context, it isn't obvious that use in one market is"bigger" or more valuable than the use in another market.

(197.) Id. at 482-93.

(198.) Dreyfuss, supra note 169, at 405; see also Felix Cohen,Transcendental Nonsense and the Functional Approach, 35 COLUM. L. REV.809, 815 (1935) (noting the circularity inherent in this argument).

(199.) See, e.g., David J. Franklyn, Debunking Dilution Doctrine:Toward a Coherent Theory of the Anti-Free-Rider Principle in AmericanTrademark Law, 56 HASTINGS L.J. 117, 140-42 (2004); Swann & Davis,supra note 89, at 276-77; Corina I. Cacovean, Note, Is Free Riding Aidedby Parody to Sneak Between the Cracks of the Trademark Dilution RevisionAct?, 31 HASTINGS COMM. & ENT. L.J. 441, 458-59 (2009).

(200.) For a detailed description of that natural rights theory andits application to traditional trademark law, see McKenna, supra note189, at 1873-95.

(201.) See, e.g., Lawrence Mfg. Co. v. Tenn. Mfg. Co., 138 U.S.537, 546 (1891) (describing a mark owner's interests as "thecustom and advantages to which the enterprise and skill of the firstappropriator had given him a just right," which is infringed whenit is "abstracted for another's use ... by deceiving thepublic, by inducing the public to purchase the goods and manufactures ofone person supposing them to be those of another"); Wolfe v.Barnett, 24 La. Ann. 97, 99 (1872) (noting that the leading principle ofthe law is to secure to the "honest, skillful and industriousmanufacturer or enterprising merchant ... the first reward of hishonesty, skill, industry or enterprise" and protect him fromdeprivation at the hands of another who "appropriates and appliesto his productions the same or a colorable imitation of the same name,mark, device or symbol, so that the public are, or may be, deceived ormisled into purchase of the productions of the one, supposing them to bethose of the other" (quoting FRANCIS H. UPTON, A TREATISE ON THELAW OF TRADE MARKS, WITH A DIGEST AND REVIEW OF THE ENGLISH AND AMERICANAUTHORITIES 97 (1860))).

(202.) See, e.g., Avery & Sons v. Meikle & Co., 81 Ky. 73,102 (1883) (referring to "that great generic rule which lies at thefoundation of all law, that a man must so use his own property as not toinjure the property of another").

(203.) And it's not as if courts never considered the questionof whether trademark rights should extend to noncompeting goods. Even inthose different commercial times, mark owners sought protection of theirmarks outside their own markets, only to be consistently turned away bycourts. See, e.g., Borden Ice Cream Co. v. Borden's Condensed MilkCo., 201 F. 510, 514 (7th Cir. 1912).

(204.) Cf. Deborah R. Gerhardt, Consumer Investment in Trademarks,88 N.C.L. REV. 427, 449-67 (2010) (noting the important role consumersplay in determining which brands succeed in a world in which eight outoften brands fail).

(205.) See, e.g., Roberta Rosenthal Kwall, Copyright and the MoralRight: Is an American Marriage Possible?, 38 VAND. L. REV. 1 (1985).

(206.) See 15 U.S.C. [section] 1064(3) (requiring cancellation ofmarks for abandonment, functionality, and genericide); Century 21 RealEstate Corp. v. Lendingtree, Inc., 425 F.3d 211, 218-21 (3d Cir. 2005)(establishing the nominative use defense in that circuit).

(207.) This seems to be a universal practice, not just one engagedin by discounters. See Deven Desai, Why Do Competitors" Set Up ShopNear Each Other?, MADISONIAN.NET, Dec. 21, 2009,http://madisonian.net/2009/l2/2l/why-do-competitors-set-up-shop-near-each-other/; see also Eric Goldman, Brand Spillovers, 22 HARV.J.L. & TECH. 381, 384-97 (2009) (describing retailers'ubiquitous practice of capitalizing on brand spillovers yet avoidingtrademark liability).

(208.) While one of us has argued that trademark law should applyonly to defendants who use the mark as a brand, Stacey L. Dogan &Mark A. Lemley, Grounding Trademark Law Through Trademark Use, 92 IOWAL. REV. 1669 (2007), a category that would exclude ordinary consumers,that argument has not carried the day, see, e.g., Rescuecom v. GoogleInc., 562 F.3d 123 (2d Cir. 2009) (finding that a search engine engagedin "trademark use" by permitting companies to run ads oppositesearch results).

(209.) For a brief history of the competing claims to steamboatpatents, see, for example, Dotan Oliar, The (Constitutional) Conventionon IP: A New Reading, 57 UCLA L. REV. 421, 449-50 (2009); and Frank D.Prager, The Steamboat Pioneers Before the Founding Fathers, 37 J. PAT.OFF. SOC'Y 486 (1955).

(210.) Cf. Yochai Benkler, Some Economics of WirelessCommunications, 26 HARV. J.L. & TECH. 25, 25-27 (2002) (drawing ananalogy to England in the 17th century, which could have defined aproperty right in "trading with India" and assigned thatproperty right to a single company, a result that would certainly havebeen inefficient).

(211.) See, e.g., Dietmar Harhoff, R&D Spillovers,Technological Proximity, and Productivity Growth--Evidence from GermanPanel Data, 52 SCHMALENBACH BUS. REV. 238, 258 (2000)("High-technology firms react more sensitively to spillovers interms of their R&D spending, and their direct marginal productivitygain from spillovers (in excess to the effect from enhanced R&Dspending) is considerably larger than the respective gain for lesstechnology-oriented firms."). Indeed, the positive relationship isso strong that some economists use spillovers as a measure ofinnovation! See Tobias Schmidt, An Empirical Analysis of the Effects ofPatent and Secrecy on Knowledge Spillovers 1 (Ctr. for Eur. Econ. Res.,Discussion Paper No. 06-048, 2006), available atftp://ftp.zew.de/publzew-docs/dp/dp06048.pdf.

(212.) Harhoff, supra note 211, at 258.

(213.) See Ronald J. Gilson, The Legal Infrastructure of HighTechnology Industrial Districts: Silicon Valley, Route 128, andCovenants Not to Compete, 74 N.Y.U.L. REV. 575, 577-78 (1999); AnnaLeeSaxenian, Inside-Out: Regional Networks and Industrial Adaptation inSilicon Valley and Route 128, CITYSCAPE, May 1996, at 41, 44-45.

(214.) ALAN HYDE, WORKING IN SILICON VALLEY: ECONOMIC AND LEGALANALYSIS OF A HIGH-VELOCITY MARKET 43 (2003).

(215.) Brett M. Frischmann & Mark A. Lemley, Spillovers, 107COLUM. L. REV. 257, 258 (2007); see also Frischmann, supra note 186. at1018; Lemley, supra note 183, at 1032.

(216.) See Univ. of S. C. v. Univ. of S. Cal., No. 2009-1064, slipop. (Fed. Cir. Jan. 19, 2010) (rejecting South Carolina's attemptto register the SC mark on the grounds that it was too similar to theSouthern California mark, despite absence of evidence of consumerconfusion).

(217.) See supra notes 144-163 and accompanying text (making theargument that this is extremely unlikely).

(218.) Indeed, as sports leagues have tightened up on merchandiselicensing, restricting the number of authorized producers ofmerchandise, see Am. Needle, Inc. v. Nat'l Football League, 538F.3d 736 (7th Cir. 2008), rev'd, 130 S. Ct. 2201 (2010), they havealso acted to raise prices across the board by terminating discounters,see. e.g., Ken Belson, The N.F.L. is Squeezing Discounters Over Apparel,N.Y. TIMES, Jan. 20, 2010, at B11.

(219.) Bradford, supra note 139, at 1286-96.

(220.) As we noted above, in the context of competitive products,research suggests that consumers are more likely to generalize productattributes from one brand to another when their packaging is moresimilar. "In other words, physical appearance of the brands had ameasurable influence on perceptions of brand performance andquality" Ellen R. Foxman et al., An Investigation of FactorsContributing to Consumer Brand Confusion, 24 J. CONSUMER AFFAIRS 170,173-74 (1990); see also George Miaoulis & Nancy D'Amato,Consumer Confusion and Trademark Infringement, 42 J. MARKETING 48, 54(1978) (finding that subjects in a field study often purchased thecompetitive brand because of product expectations "stimulated bythe visual impact of the product"). Interestingly, according toMiaoulis and D'Amato, the primary cue for association between thetwo brands was not the name but the visual appearance, ld.

(221.) Laura A. Heymann, The Public's Domain in Trademark Law:A First Amendment Theory of the Consumer, 43 GA. L. REV. 651, 656-57(2009).

(222.) Wendy J. Gordon, On Owning Information: IntellectualProperty and the Restitutionary Impulse, 78 VA. L. REV. 149, 167 (1992);see also Marina Lao, Free Riding: An Overstated, and Unconvincing,Explanation for Resale Price Maintenance, in How THE CHICAGO SCHOOLOVERSHOT THE MARK: THE EFFECT OF CONSERVATIVE ECONOMIC ANALYSIS ON U.S.ANTITRUST 196, 197 (Robert Pitofsky ed., 2009) ("[F]ree ridingexists throughout the economy and the law generally does not banit.").

(223.) Cargill, Inc. v. Monfort of Colo., Inc., 479 U.S. 104(1986); Assoc. Gen. Contractors v. Cal. State Council of Carpenters, 459U.S. 519, 540-45 (1983); Brunswick Corp. v. Pueblo Bowl-O-Mat, Inc., 429U.S. 477, 489 (1977).

(224.) Cf. Christopher Sprigman, Copyright and the Rule of Reason,7 J. TELECOM. & HIGH TECH. L. 317 (2009) (proposing to limitcopyright by drawing analogous doctrines from antitrust law).

(225.) Brunswick, 429 U.S. at 489.

[c] 2010 Mark A. Lemley & Mark R McKenna.

Mark A. Lemley, William H. Neukom Professor, Stanford Law School;partner, Durie Tangri LLP.

Mark P. McKenna, Associate Professor, Notre Dame Law School.

For comments on the Article and helpful discussions of the idea, wethank Robert Burrell, Tom Cotter, Deven Desai, Stacey Dogan, DeborahGerhardt, Rose Hagan, Ariel Katz, Rebecca Tushnet, and participants atthe Ninth Annual Intellectual Property Scholars' Conference andfaculty workshops at Notre Dame and the University of North Carolina.