The Registration Is Not Bulletproof

Picture the moment a registration certificate arrives from the U.S. Patent and Trademark Office, embossed and official, and a brand owner files it away like a deed to a house. The instinct is understandable. A federal registration is a formidable asset: it carries a presumption that the mark is valid, a presumption that the registrant owns it, a presumption of the exclusive right to use it nationwide, and—after five years—the near-armor of incontestability. We catalog those advantages in our companion piece on the benefits of federal trademark registration. But the certificate is not a deed, and the analogy to real property misleads. A registration is more like a driver's license: a document the government issues based on representations you made, that confirms a status the law recognizes, and that the government can take back if the representations were false or the status no longer holds.

That is what cancellation does. The same statute that creates a registration—the Lanham Act—also supplies the tools to undo it. When a registration should never have issued, or should no longer stand, a court or the USPTO's own tribunal can cancel it and order the federal register corrected to reflect the truth.

Cancellation is the trademark system's mechanism for cleaning the register. A competitor blocked by a registration resting on a mark that is generic, abandoned, fraudulently obtained, or confusingly similar to the competitor's own senior mark is not stuck; it can petition to strike that registration. A defendant sued for infringement can counterclaim to cancel the very registration the suit is built on—and if the counterclaim succeeds, the plaintiff's case can collapse at its foundation. And a federal court presiding over a trademark dispute holds an extraordinary power: it can order the register itself rectified, with its order certified to the Director of the USPTO for action. For the registrant, cancellation is the nightmare scenario—the loss of the asset itself. For the challenger, it is among the most powerful weapons in trademark litigation, because erasing a registration does not merely win a point; it can remove the other side's right to register, weaken its mark in any confusion analysis, and reshape the entire dispute.

This guide explains how cancellation works, with particular attention to cancellation inside federal litigation—as a claim or counterclaim in a district court case under Section 37 of the Lanham Act—rather than the standalone administrative proceeding before the Trademark Trial and Appeal Board. It is written for both lawyers and non-lawyers and takes a deliberately neutral view, useful whether you are seeking to cancel a registration or defending one. Because cancellation so often turns on priority, use, distinctiveness, and the registration's statutory presumptions, it connects to topics we treat in depth elsewhere: the constructive-use priority that registration confers, the common-law priority that can defeat a junior registration, and the bona fide intent required for intent-to-use applications.

Three Venues: The TTAB, the District Court, and the USPTO's New Ex Parte Proceedings

Most discussions of cancellation describe two forums. As of late 2021, there are effectively three, and the choice among them is the first strategic decision a challenger makes.

The first is an administrative cancellation proceeding before the Trademark Trial and Appeal Board (TTAB), authorized by Section 14 of the Lanham Act, 15 U.S.C. § 1064. A party with a sufficient stake files a petition to cancel; the matter is litigated before the Board under the Trademark Rules of Practice (37 C.F.R. §§ 2.1–2.209) and the guidance collected in the Trademark Trial and Appeal Board Manual of Procedure (TBMP), largely on a paper record. There is no in-person trial—the Board takes testimony by deposition and declaration and decides on the briefs. Crucially, the TTAB's jurisdiction is limited to the register. It determines only the right to register a mark; it cannot decide the right to use a mark, cannot adjudicate infringement or unfair competition, and cannot award an injunction, damages, attorneys' fees, or costs. A party that wants money or an injunction has to be in court. We treat the Board's machinery—including its distinctive discovery regime—in our guide to discovery practice in TTAB trademark proceedings.

The second—and the focus here—is cancellation in federal district court under Section 37 of the Lanham Act, 15 U.S.C. § 1119. Section 1119 provides that, in any action involving a registered mark, the court may determine the right to registration, order the cancellation of registrations in whole or in part, restore cancelled registrations, and otherwise rectify the register, with the court's order certified to the Director. This is the power that lets a district court, in the middle of an infringement case, strike the very registration in suit—and, unlike the TTAB, the same court can simultaneously enjoin the loser, award profits, and shift fees in an exceptional case under Section 35(a).

The third venue is newer and worth flagging because it has quietly changed the landscape. The Trademark Modernization Act of 2020 created two ex parte proceedings at the USPTO—expungement and reexamination—that took effect on December 27, 2021. Expungement lets a third party (or the Director) ask the Office to cancel a registration, in whole or in part, on the ground that the mark was never used in commerce on some or all of the registered goods or services; reexamination targets registrations where the mark was not in use as of a relevant date for use-based filings. Both are streamlined, examiner-driven proceedings—no adverse party to litigate against, no discovery, lower cost—and they exist alongside a new inter partes ground: as of the same date, a petition to cancel may be filed against a registration at least three years old on the basis of nonuse of the mark on some or all of the registered goods, under 15 U.S.C. § 1064(6). For a challenger whose only complaint is that the registrant never actually used the mark, expungement may be a far cheaper path to clearing the register than a full court fight—though it cannot resolve an infringement dispute or yield damages.

One feature of § 1119 is essential and frequently misunderstood: it is not a freestanding grant of federal jurisdiction. Its text requires "an action involving a registered mark," which means there must already be an independent basis for the court's jurisdiction—most commonly a pending infringement claim, a declaratory-judgment action over the parties' rights, or a similar live controversy. A party generally cannot file a lawsuit whose only purpose is to cancel a registration and expect § 1119 to carry it into federal court; cancellation under that section rides along with a genuine, jurisdiction-supporting dispute. In practice, then, cancellation in litigation almost always arrives as a counterclaim by a defendant accused of infringement, or as a claim folded into a broader fight about the parties' rights.

The strategic calculus is straightforward. A party that wants only to clear the register, and has a clean nonuse story, may reach for expungement or a TTAB petition. A defendant already staring down an infringement complaint will almost always counterclaim for cancellation under § 1119—both to defend itself and to obtain a court order erasing the registration and rectifying the register in one proceeding. And because parallel tracks can collide, courts coordinate: a district court may stay an infringement suit pending a TTAB decision, or the Board may suspend its proceeding in deference to a court, under primary-jurisdiction and efficiency principles. The choice of forum is not a formality; it determines the record, the cost, the timeline, and—as we will see when we reach B&B Hardware—sometimes who wins the issue everywhere.

Who Can Bring the Claim: Entitlement After Lexmark

Before reaching the grounds, the challenger has to clear a threshold question that trips up the unwary: does it have the right to bring a cancellation claim at all? Not everyone who dislikes a registration may attack it. For decades, courts and the Board framed this as "standing." The vocabulary has changed, but the gate remains.

The Lanham Act provides that "any person who believes that he is or will be damaged" by a registration may petition to cancel it. 15 U.S.C. § 1064. The Supreme Court reframed the analysis in Lexmark International, Inc. v. Static Control Components, Inc., 572 U.S. 118 (2014)—a false-advertising case—by holding that a statutory cause of action under the Lanham Act extends only to plaintiffs whose interests fall within the zone of interests the statute protects and whose injuries are proximately caused by the challenged conduct. The Federal Circuit then imported that framework into trademark registration disputes, holding in Australian Therapeutic Supplies Pty. Ltd. v. Naked TM, LLC, 965 F.3d 1370 (Fed. Cir. 2020), and Corcamore, LLC v. SFM, LLC, 978 F.3d 1298 (Fed. Cir. 2020), that the Lexmark test now governs entitlement to bring opposition and cancellation proceedings. The court confirmed that this entitlement test is substantively the same as the older "real interest" formulation the Board had long applied—a real interest in the proceeding plus a reasonable basis for the belief of damage. See Corcamore, 978 F.3d at 1305–06; Ritchie v. Simpson, 170 F.3d 1092, 1095 (Fed. Cir. 1999).

In practice, the bar is low and the courts construe it liberally. A challenger need not plead or prove actual damage; a direct and personal stake in the outcome suffices. Entitlement is readily found where, for example, the challenger owns a confusingly similar mark or has prior use, where its own application is being blocked (or it reasonably believes will be blocked) by the registration, or where it wants to use a descriptive or generic term that the registration locks up. See, e.g., Cunningham v. Laser Golf Corp., 222 F.3d 943 (Fed. Cir. 2000); Empresa Cubana del Tabaco v. General Cigar Co., 753 F.3d 1270 (Fed. Cir. 2014). And there is a convenient shortcut in litigation: a defendant sued for infringement on a registration has an inherent entitlement to counterclaim to cancel it, because the suit itself supplies the stake. The lesson is not that entitlement is hard—it usually is not—but that it is a real element the challenger must allege at the outset and ultimately prove, and a careless pleading that fails to articulate the challenger's interest invites an early dismissal of the cancellation claim before the merits are ever reached.

What You Are Attacking: The Presumptions and Incontestability

Cancellation is an uphill fight by design, because the registrant does not start at zero—it starts ahead. A registration on the Principal Register is prima facie evidence of the validity of the registered mark, of the registrant's ownership, and of the registrant's exclusive right to use the mark in commerce on the goods or services listed. 15 U.S.C. §§ 1057(b), 1115(a). The practical consequence is that the burden sits on the challenger: the registration does not have to prove itself, at least at the outset, and the challenger must come forward with evidence sufficient to overcome the presumption. The quantum of proof varies by ground—most grounds by a preponderance of the evidence, but fraud by the demanding clear-and-convincing standard discussed below.

The registrant's armor thickens with time. After five years of continuous use and the filing of a Section 15 affidavit, a registration can become incontestable under 15 U.S.C. § 1065, and incontestability forecloses many avenues of attack. (Maintaining a registration to that milestone—the Section 8 declaration of continued use, the combined Section 8 and 15 filing—is its own discipline, and a registrant that lets a maintenance deadline lapse can lose the registration without anyone ever filing a cancellation petition.) But—and this is the point everyone half-remembers wrong—incontestability is not invulnerability. A registration, incontestable or not, remains subject to a defined set of grounds preserved by §§ 1064(3), (5), (6) and 1115(b). An incontestable mark can still be cancelled because it has become generic, is functional, has been abandoned, was obtained or maintained by fraud, is being used to misrepresent the source of the goods, or runs afoul of certain provisions of § 2(a)–(c), among others.

So the real effect of incontestability is to take the "weaker," registrability-based grounds off the table—principally a challenge that the mark is merely descriptive and lacks secondary meaning—while leaving the "fundamental" grounds available at any time. That makes the first question in any cancellation analysis a clerical-sounding one with enormous consequences: how old is the registration, and is it incontestable? The answer dictates which grounds even exist before anyone argues a single fact.

The Grounds for Cancellation: A Map by Timing

The grounds for cancellation, drawn principally from § 1064 and the registrability provisions of § 1052 (Section 2 of the Act), sort into two timing buckets. Understanding the buckets is half the battle.

Within the first five years after registration, a Principal Register registration is exposed to essentially the full menu of grounds that could have blocked registration in the first place. 15 U.S.C. § 1064(1). These include a likelihood of confusion with the challenger's senior mark—a priority-plus-confusion theory, and the single most common ground in competitor disputes; mere descriptiveness without acquired secondary meaning; the mark being primarily merely a surname without acquired distinctiveness; primarily geographically descriptive or deceptively misdescriptive marks; deceptive marks; marks that falsely suggest a connection with a person or institution; and likelihood of dilution of a famous mark. Once the five-year window closes and the registration becomes incontestable, these "registrability" grounds generally drop away—which is precisely why a challenger who spots a vulnerable young registration should move promptly rather than let the clock run.

At any time, even against an incontestable registration, a registration is exposed to the grounds that go to the fundamental propriety of the registration. Under §§ 1064(3), (5), and (6), a registration may be cancelled because the mark has become the generic name for the goods or services; because the mark (or its claimed feature) is functional; because the mark has been abandoned; because the registration was obtained or maintained by fraud; because the mark is being used to misrepresent the source of the goods; because the mark was never used in commerce; and, under the 2021 addition for registrations at least three years old, on the basis of nonuse of the mark on particular goods. These "anytime" grounds are the ones that matter in long-running disputes over older registrations, because they survive incontestability.

A few constitutional footnotes round out the map. The Lanham Act still recites, on its face, bars on registering "immoral," "scandalous," and "disparaging" marks under § 2(a). But the Supreme Court struck the disparagement bar in Matal v. Tam, 582 U.S. 218 (2017)—the case that grew out of the band The Slants and resonated with the long-running Washington Redskins trademark dispute—and struck the immoral/scandalous bar in Iancu v. Brunetti, 588 U.S. 388 (2019). A challenger should not build a cancellation on a ground the Constitution has already gutted.

A challenger's first analytical task, then, is to lay the available grounds against the registration's age and status. If the registration is young and contestable, the full menu is open, and the challenger should pick the strongest available theory and move before the window closes. If it is incontestable, the challenger must find a fundamental ground—genericness, functionality, abandonment, fraud, misrepresentation, or nonuse. The rest of this guide examines the grounds that actually decide cases.

Fraud on the USPTO After In re Bose: The Claim Everyone Overpleads

Of all the cancellation grounds, fraud is the most dramatic, the most overpleaded, and the most misunderstood. Its modern contours were set by the Federal Circuit's landmark decision in In re Bose Corp., 580 F.3d 1240 (Fed. Cir. 2009), and understanding Bose is essential because it transformed fraud from a routine claim into a rarely successful one.

To appreciate what Bose did, you have to know what it undid. The Board's earlier decision in Medinol Ltd. v. Neuro Vasx, Inc., 67 U.S.P.Q.2d 1205 (T.T.A.B. 2003), had effectively permitted a fraud finding where an applicant "knew or should have known" that a statement in its application or maintenance filing was false—an objective, negligence-flavored standard. Under Medinol, an applicant who overstated the goods on which it used the mark, or who carelessly signed an inaccurate declaration, risked losing the entire registration for fraud. Predictably, fraud claims proliferated; reciting "fraud" became a near-reflex in trademark answers and cancellation petitions.

In re Bose overruled that approach and restored a stringent, intent-based standard. To prove fraud on the USPTO, the challenger must show that the applicant or registrant knowingly made a false, material representation with the intent to deceive the Office. Every element bites. The misrepresentation must concern a material fact—one that would have affected the USPTO's decision to issue or maintain the registration. The registrant must have acted with subjective intent to deceive; the Federal Circuit was emphatic that "mere negligence is not sufficient to infer fraud or dishonesty," and that even gross negligence does not establish the requisite intent. A false statement made with a reasonable, honest belief that it was true is not fraud. And all of it must be proven by clear and convincing evidence—the court's memorable formulation, borrowing from earlier cases, is that fraud must be proven "to the hilt," with any doubt resolved against the party charging fraud.

The practical consequences are large and run in both directions. Post-Bose, fraud claims rarely succeed, because subjective intent to deceive is hard to prove and is seldom present in the ordinary case of sloppy or overbroad filings. A registrant who listed goods on which it was not actually using the mark has not committed fraud unless the challenger can show it lied with intent to deceive. The more likely remedy for such overclaiming is not cancellation of the whole registration but restriction—deleting the unused goods or limiting the registration to what was actually used. Bose itself made the point: where the only proof is that the registrant overstated its goods, the appropriate response is to narrow the registration, not to wipe it out.

For the challenger, the discipline is to bring a fraud claim only when there is genuine evidence of a knowing, deceptive, material lie—and to plead it with the particularity that Federal Rule of Civil Procedure 9(b) demands, identifying the who, what, when, and why of the deception—rather than dressing up an ordinary inaccuracy as fraud and watching it get dismissed. (For the discovery work that either unearths real fraud or exposes its absence, see our guide to discovery practice in TTAB trademark proceedings.) For the registrant, Bose is reassuring but not a license for carelessness: honest, well-documented filings have little to fear, but overclaiming still invites a restriction challenge even where it falls short of fraud, so keeping the registered goods and services accurate remains plain good hygiene.

Cancellation for Nonuse and Abandonment: Against a Famously Low Bar

Because trademark rights and registrations ultimately rest on use, nonuse and abandonment are among the most important cancellation grounds. They are also the grounds most often misjudged, because challengers underestimate how little use the law actually requires.

Start with the doctrine. Under § 45 of the Lanham Act, a mark is abandoned when its use has been discontinued with intent not to resume, and—critically—nonuse for three consecutive years constitutes prima facie evidence of abandonment, creating a rebuttable presumption that shifts to the registrant the burden of producing evidence of an intent to resume use. 15 U.S.C. § 1127. (The three-year window replaced a two-year window in a 1996 amendment.) Abandonment is a powerful "anytime" ground that survives incontestability, 15 U.S.C. § 1115(b)(2), and it is frequently litigated over older registrations whose marks have quietly fallen into disuse. Once a mark is genuinely abandoned, it is gone for good: it cannot be revived, a later user cannot claim priority dating back to the abandoned use, and an attempt to assign the dead mark is an invalid assignment in gross. See L. & J.G. Stickley, Inc. v. Canal Dover Furniture Co., 79 F.3d 258, 264–65 (2d Cir. 1996).

The element that does the work is intent not to resume. The registrant facing the three-year presumption rebuts it not by wishing to use the mark someday but by producing concrete evidence of an intent to resume use within the reasonably foreseeable future—documented business plans, steps toward relaunch, licensing efforts, or a credible explanation for excusable nonuse. See Crash Dummy Movie, LLC v. Mattel, Inc., 601 F.3d 1387, 1391 (Fed. Cir. 2010) (registrant rebutted the presumption with evidence of research, development, and licensing activity). A bare assertion of "we always meant to bring it back" will not do; vague hopes are not plans.

Now the counterpoint that defeats so many nonuse theories: the bar for "use in commerce" is famously low, which means a cancellation premised on insufficient use can fail even when the registrant's commercial activity looks modest. The Federal Circuit made the point vividly in Christian Faith Fellowship Church v. adidas AG, 841 F.3d 986 (Fed. Cir. 2016), reversing a Board decision that had cancelled a registration for lack of use in commerce. The court held that the sale of just two hats bearing the mark to an out-of-state purchaser constituted "use in commerce" sufficient to support the registration—because even a small interstate sale falls within Congress's commerce power, and the Act does not impose a minimum-volume threshold. The lesson for challengers is blunt: a "they barely used it" theory is weak. What matters is whether there was bona fide use in the ordinary course of trade—genuine commercial use, not "token" sales made merely to reserve rights, see Wells Fargo & Co. v. ABD Insurance & Financial Services, Inc., 758 F.3d 1069, 1072 (9th Cir. 2014)—not whether the volume was impressive. The stronger nonuse theory is genuine abandonment: a real cessation of use with no intent to resume, not a quibble about small numbers.

It is worth keeping three related-but-separate attacks distinct, because conflating them is a classic error:

  • Nonuse asks whether the mark was ever used as the registration required (never used at all, or not in use as of a relevant date). This is the theory behind the new § 1064(6) ground and the TMA's expungement and reexamination proceedings.
  • Abandonment asks whether use, once begun, was discontinued with intent not to resume—the three-year-presumption ground.
  • Fraud asks whether the registrant lied about use with intent to deceive.

They have different elements and different burdens, and a challenger should plead the theory the facts actually support rather than blur them together. The point matters most where fraud is concerned: fraud's clear-and-convincing "to the hilt" standard is far harder to meet than abandonment's preponderance standard riding on a three-year presumption. A challenger who can prove genuine cessation should not bet the case on a fraud theory it cannot prove.

Genericness: When the Mark Becomes the Name of the Thing

A mark that started life as a distinctive brand can collapse into the common name for the product itself—and when it does, it can be cancelled at any time, incontestability notwithstanding. 15 U.S.C. § 1064(3). The graveyard of "genericide" is familiar: aspirin, escalator, thermos, cellophane—all once brands, all now the names of the things they once distinguished. The legal test asks for the term's primary significance to the relevant public: does the public understand the word to identify a particular source, or merely to name a category of goods? The classic framing comes from Bayer Co. v. United Drug Co., 272 F. 505 (S.D.N.Y. 1921) (Judge Learned Hand asking, in essence, "What do the buyers understand by the word?"), and the modern statutory test is codified in § 14(3).

Genericness is proved with the evidence of usage: dictionary definitions, media and trade usage, competitor usage, the registrant's own materials, and—often decisively—consumer survey evidence designed to capture the public's understanding. (The methodology and the fights over it are their own field; we cover them in our guide to consumer survey expert methodology in trademark cases and the Daubert challenges those experts face.) The Supreme Court refined the doctrine for the internet age in USPTO v. Booking.com B.V., 591 U.S. 549 (2020), holding that a "generic.com" term is not automatically generic; whether BOOKING.COM is generic depends on whether consumers in fact perceive the composite as a brand or as the name of a class of services. The decision is a reminder that genericness is a fact question about perception, not a mechanical rule—and that survey evidence frequently carries the day.

Because the burden of proving genericness rests on the challenger and the threshold is high, genericide is hard to prove and registrants fight it ferociously (the perennial "photocopy, don't Xerox" advertising campaigns are genericide-prevention in action). But where the evidence is one-sided, genericness is a strong cancellation ground precisely because it survives incontestability and disposes of the registration entirely.

Functionality: Why You Cannot Trademark a Useful Feature

When the "mark" is not a word but a product feature, a color, a shape, or trade dress, a distinct anytime ground comes into play: functionality. The Lanham Act bars registration—and supplies a ground for cancellation at any time—of matter that is functional, 15 U.S.C. §§ 1052(e)(5), 1064(3), because trademark law protects source identifiers, not useful features. Allowing a producer to monopolize a functional feature through a perpetual trademark would do an end-run around the patent bargain, in which exclusivity is the reward for disclosure and is strictly time-limited.

The Supreme Court's touchstone is TrafFix Devices, Inc. v. Marketing Displays, Inc., 532 U.S. 23 (2001): a feature is functional, and cannot serve as a trademark, if it is essential to the use or purpose of the article or affects its cost or quality. Where that "utilitarian" functionality is shown, the inquiry ends—no amount of secondary meaning can rescue a functional feature, and the existence of an expired utility patent claiming the feature is strong evidence that it is functional. Courts also recognize aesthetic functionality, which asks whether protecting the feature would put competitors at a significant non-reputation-related disadvantage. Functionality is the central battleground in trade-dress and product-configuration disputes, a subject we explore in our analysis of the evolution of trade dress protection for product design and configuration. For a challenger facing an incontestable registration covering a product shape or feature, functionality is often the most promising line of attack, because it survives incontestability and, where the feature is plainly utilitarian, can be resolved without the swearing contest that fraud requires.

Cancellation for Intent-to-Use Defects

Registrations that issued from intent-to-use applications carry a distinct cluster of vulnerabilities. The short answer to whether such a registration can be cancelled for intent-to-use defects is yes, and there are two principal defects to investigate.

The first is the absence of a bona fide intent to use the mark at the time the application was filed. An intent-to-use application under § 1(b) must rest on a genuine, good-faith intention to use the mark on the specified goods or services, evaluated on the objective evidence as of the filing date. The Federal Circuit's leading decision, M.Z. Berger & Co. v. Swatch AG, 787 F.3d 1368 (Fed. Cir. 2015), holds that bona fide intent is an objective inquiry: the absence of any documentary evidence of intent to use—no business plan, no product development, no marketing steps—can be enough to find that the applicant lacked the required intent, voiding the application and any registration that issued from it. A registrant who filed a placeholder application to "bank" a mark, with nothing behind it, exposes the resulting registration to attack. This is why intent-to-use applicants should document their launch plans contemporaneously, as we explain in our guide to intent-to-use trademark applications.

The second defect is an improper assignment of an intent-to-use application. The Lanham Act bars assigning an intent-to-use application before the applicant has begun using the mark, except to a successor to the portion of the applicant's ongoing and existing business to which the mark pertains. 15 U.S.C. § 1060(a)(1). An assignment that violates this anti-trafficking rule can void the application and any registration that grew out of it—a genuine trap that surfaces in mergers, acquisitions, and reorganizations, where a pending application changes hands in a bulk transfer and no one notices the restriction. A challenger examining an intent-to-use-based registration should therefore investigate both the bona fide intent at filing and the chain of any pre-use assignments, because either can supply a clean cancellation ground that sidesteps the harder merits questions entirely.

"Primarily Merely a Surname" and Its Ripple Effects

A mark that is primarily merely a surname occupies a special place in cancellation analysis, and a finding that a mark is primarily merely a surname carries consequences well beyond cancellation itself. Under § 2(e)(4) of the Lanham Act, 15 U.S.C. § 1052(e)(4), a mark that is primarily merely a surname is not registrable on the Principal Register absent acquired distinctiveness—it is, in the law's terms, not inherently distinctive. Such a mark may reach the Principal Register only on a showing of secondary meaning under § 2(f), or it may sit on the Supplemental Register while it works toward acquired distinctiveness.

Whether a term is "primarily merely a surname" is assessed through a set of recognized factors: whether the surname is rare; whether anyone connected with the applicant bears the surname; whether the term has a recognized meaning other than as a surname; whether it has the structure and pronunciation that the public would perceive as a surname; and, in some formulations, whether the stylization of the mark creates a separate, non-surname commercial impression. See In re Etablissements Darty et Fils, 759 F.2d 15 (Fed. Cir. 1985); In re Benthin Management GmbH, 37 U.S.P.Q.2d 1332 (T.T.A.B. 1995). The ultimate question is the term's primary significance to the relevant public.

A surname finding ripples in three directions, and a litigant on either side should trace all three:

First, registrability and cancellation. If a primarily-surname mark reached the Principal Register, it did so either under § 2(f)—an express concession that the mark is not inherently distinctive and depends on acquired distinctiveness—or it slipped through without that showing. In the latter case, within the five-year window, it is vulnerable to cancellation under § 2(e)(4) as a surname lacking secondary meaning. That is a contestable, registrability-based ground, which is why timing again matters: once the registration becomes incontestable, the surname challenge generally evaporates.

Second, strength. A surname mark is not inherently distinctive, which weighs against its strength in the likelihood-of-confusion analysis—a factor we examine in our discussion of the Polaroid factors on summary judgment in the Second Circuit. A weaker mark claims a narrower scope of protection, so even where the registration survives, the surname finding shrinks the registrant's enforcement reach. The challenge to validity and the defense to infringement run on the same fact.

Third, register and priority. If the mark sits on the Supplemental Register rather than the Principal Register, it never carried the Principal Register's presumptions or its constructive-use priority in the first place—a distinction we unpack in our guide to common-law rights, the Supplemental Register, and the Principal Register.

In short, a surname finding is a flag waving over the whole dispute. It marks a non-inherently-distinctive mark whose registration may be cancellable within the window absent secondary meaning, and whose strength—and therefore whose reach—is diminished regardless. A registrant facing the finding will usually need to prove acquired distinctiveness to shore up both the registration and the mark; a challenger will press the cancellation ground (within the window) and the weakness in any confusion analysis at the same time.

Can Cancellation Be Won on Summary Judgment?

A practical question that often determines the shape of litigation is whether cancellation can be obtained on summary judgment rather than only after trial. The answer is "yes, for some grounds; rarely, for others," and the dividing line tracks a single distinction: whether the ground turns on objective facts or on contested intent.

Grounds resting on objective facts are well suited to summary judgment where the material facts are undisputed. Genericness can be resolved on summary judgment where the dictionary, media, competitor-usage, and survey evidence leaves no genuine dispute that the term is the common name for the goods. Functionality can be decided as a matter of law where the feature is plainly essential to the article's use—an expired utility patent and the registrant's own touting of the feature's advantages can foreclose any genuine dispute under TrafFix. Abandonment can be resolved on summary judgment where undisputed evidence establishes three or more consecutive years of nonuse—triggering the presumption—and the registrant fails to produce evidence of intent to resume. Mere descriptiveness, surname, and priority/likelihood-of-confusion grounds can likewise be decided on summary judgment where the underlying facts are undisputed and only the legal conclusion remains (with the standard caution that confusion is fact-intensive and frequently survives to trial, as we discuss in the Polaroid summary-judgment article).

Fraud, by contrast, is rarely a candidate for summary judgment, and for structural reasons that follow directly from Bose. Fraud requires proof of subjective intent to deceive, by clear and convincing evidence, "to the hilt." Intent is the quintessential fact question—it usually depends on credibility and on the inferences a factfinder draws from the registrant's conduct and explanations, exactly the sort of thing a court will not resolve on a paper record. A court will only rarely find, on summary judgment, that no reasonable factfinder could fail to find deceptive intent by clear and convincing evidence with all doubts resolved against the challenger. The strategic consequence is clean: a challenger hoping to win cancellation before trial should lead with an objective ground—genericness, functionality, abandonment, descriptiveness, surname, or priority—and treat fraud as a trial issue, if it is genuinely supported at all. The choice of trier of fact then matters too; which validity issues go to a jury and which to the bench is a recurring strategic question we take up in our guide to bench trial versus jury trial issues in trademark litigation.

Procedure and Remedy in Federal Court

In litigation, cancellation is ordinarily pleaded as a counterclaim (or as an affirmative claim) invoking § 1119, attached to a case that independently supports federal jurisdiction—most often an infringement suit. The challenger must plead the specific ground or grounds with the precision each requires—fraud with Rule 9(b) particularity, the others with enough factual content to be plausible—and then carry the applicable burden against the registration's presumptions. A defendant should plead the cancellation counterclaim early, both to anchor the § 1119 power and to make clear that the registration's validity is squarely in play.

The remedy is what makes the district court distinctive. The court does not merely declare the registration invalid between the parties; under § 1119 it may order the register itself rectified, directing the cancellation—or restriction—of the registration, with the order certified to the Director of the USPTO for action. The relief can be partial: a court (like the Board under § 18, 15 U.S.C. § 1068) may restrict a registration to the goods or services actually used, cancel it as to particular classes, or narrow an overbroad identification. That partial-relief power dovetails neatly with the Bose restriction remedy for overclaiming, and it gives a court a scalpel where a challenger has proven that some—but not all—of a registration is defective. A court may also coordinate with the Board, staying or deferring to a parallel administrative proceeding where efficiency or primary-jurisdiction concerns counsel.

Because cancellation grounds are statutory and shared across the two forums, the substantive standards do not change with the venue: the same § 1064 grounds, the same presumptions, the same Bose fraud standard, and the same three-year abandonment presumption apply whether the cancellation is litigated before the Board or in court. The TTAB's own practice and standards, collected in the TBMP and analyzed in McCarthy on Trademarks and Unfair Competition (the standard treatise), therefore inform the analysis even when cancellation is pursued in district court. And when cancellation rides alongside an infringement claim, the court's other powers come into play—injunctive relief, an accounting of profits, and the apportionment of damages among the infringing and non-infringing portions of the defendant's sales—so that a single proceeding can both correct the register and resolve the parties' commercial dispute.

The TTAB–Court Interplay and Issue Preclusion After B&B Hardware

Because cancellation can be pursued in more than one forum, a recurring strategic question is how a TTAB proceeding and a district court case interact—and the answer changed meaningfully with B&B Hardware, Inc. v. Hargis Industries, Inc., 575 U.S. 138 (2015). There, the Supreme Court held that a TTAB decision can have issue-preclusive effect in later federal-court litigation when the ordinary elements of issue preclusion are satisfied and the issue the Board actually decided is materially the same as the one before the court. A Board determination on likelihood of confusion, for instance, can bind the parties in a later infringement or cancellation action—provided the usages and issues actually adjudicated align.

That holding carries two practical consequences. First, a TTAB proceeding is not a low-stakes warm-up. Its findings may follow the parties into court, so a litigant cannot treat the Board as a throwaway forum, fielding a B-team and a thin record on the theory that the "real" fight is the lawsuit. The Board's record may become the real fight. Second, the alignment of issues is everything. B&B Hardware turned on whether the Board adjudicated the same likelihood-of-confusion question the court would face, and preclusion does not attach where the Board considered the marks in a materially different context than the marketplace usages at issue in the litigation—because the Board often compares marks and goods "as set forth in the application and registration," while a court examines real-world commercial use.

A litigant choosing between forums, or facing parallel proceedings, must therefore plan ahead: will a Board win (or loss) carry over, and is it wiser to seek a stay of one proceeding pending the other? Courts retain discretion to stay an infringement action in deference to a pending TTAB proceeding, or to press forward and let the § 1119 power resolve the registration directly; the Board, for its part, routinely suspends a proceeding when a civil action that may dispose of the registration is pending. The interplay rewards forethought. The forum in which an issue is first and fully litigated may control its resolution everywhere, and a party that drifts into a TTAB cancellation without appreciating its preclusive reach can find a court door quietly closing before it ever reaches it. The prudent course is to map, at the outset, which forum will decide which issue, what record each will build, and whether a stay serves the client better than fighting on two fronts at once.

A Worked Example (Hypothetical)

Consider an invented dispute that draws the threads together. The following is a hypothetical for illustration only.

"Hargrove," a maker of leather goods, holds a four-year-old Principal Register registration for HARGROVE for handbags. It sues "Hargrave Leather," a newer competitor, for infringement in federal district court. Hargrave Leather answers and counterclaims under § 1119 to cancel the HARGROVE registration, asserting several grounds.

Run the analysis the way a court would. Because the registration is four years old, it is not yet incontestable, so the full menu of grounds—including descriptiveness, surname, and priority—remains open. That timing is a gift to Hargrave; had it slept on the dispute for two more years, the registrability grounds would have closed.

Hargrave's strongest ground may be primarily merely a surname. If HARGROVE is primarily a surname and the registration issued without a § 2(f) showing of acquired distinctiveness, Hargrave can seek cancellation under § 2(e)(4) within the window—and can simultaneously argue that a surname mark is conceptually weak, narrowing its protection in the confusion analysis. The same evidence does double duty: the surname factors (rarity, alternative meaning, structure) attack validity, and the resulting weakness erodes the infringement claim. A single finding ripples through both.

Suppose Hargrave also pleads fraud, alleging that Hargrove listed "wallets" and "belts" in its registration but never sold those items. After Bose, this is a weak theory. Overclaiming goods is not fraud absent clear and convincing proof that Hargrove knowingly lied with intent to deceive the USPTO, and the realistic outcome is restriction—deletion of the unused goods—rather than cancellation of the whole registration. The fraud claim is also unlikely to survive summary judgment, because deceptive intent is a contested fact question that cannot be established "to the hilt" on a paper record. Hargrave would be better served pressing the surname ground and, if the facts support it, an abandonment or nonuse theory. But here Christian Faith Fellowship sounds a warning: if Hargrove can show even modest bona fide sales of handbags, a "they barely used it" argument fails; only genuine cessation with no intent to resume—say, three-plus years of no sales and no documented plan to relaunch—supports abandonment.

On summary judgment, Hargrave's objective grounds (surname; abandonment if the nonuse is undisputed and unrebutted) are viable candidates for resolution as a matter of law; its fraud claim is not. If Hargrave prevails on the surname ground for lack of secondary meaning, the court can order the register rectified—cancelling the registration, or, if only some goods are affected, restricting it—with the order certified to the USPTO. And even short of cancellation, the surname finding weakens HARGROVE in the confusion analysis, improving Hargrave's defense to the underlying infringement claim and, potentially, its position on damages and fees. The example shows the three lessons in miniature: grounds are chosen for the facts and the timing, Bose defangs casual fraud theories, and one determination—surname status—can ripple through validity, infringement, and remedy alike.

Practical Takeaways

For the challenger seeking cancellation, the discipline is to match the ground to the facts and the timing—and to resist the temptation to throw everything at the wall. Determine first whether the registration is incontestable. If it is not, the full menu (confusion, descriptiveness, surname, dilution) is available within five years, so move promptly before the window closes. If it is incontestable, hunt for a fundamental ground—genericness, functionality, abandonment, fraud, misrepresentation, or nonuse. Confirm your entitlement under Lexmark/Corcamore and plead it; in an infringement suit, the counterclaim's entitlement is built in. Lead with objective grounds that can be won on summary judgment, and treat fraud with caution—plead it only with genuine evidence of a knowing, deceptive, material lie, with Rule 9(b) particularity, and expect it to be a trial issue under the clear-and-convincing "to the hilt" standard. Investigate intent-to-use registrations for bona fide-intent and improper-assignment defects, which can void a registration without a merits fight. And pick the right forum: for a pure nonuse complaint, the TMA's expungement or reexamination route may be faster and cheaper than litigation; for a live infringement dispute, plead cancellation as a § 1119 counterclaim and ask the court to rectify the register, including partial restriction where only some goods are vulnerable.

For the registrant defending a registration, the message is that the presumptions and incontestability are real protections—but they are not absolute, and good hygiene preserves them. Keep the registration's goods and services accurate to avoid overclaiming and the restriction it invites. Maintain genuine, documented use to defeat abandonment, and keep the records—invoices, advertising, dated specimens—that prove it. Preserve evidence of bona fide intent and a clean assignment history for intent-to-use registrations. Where the mark is a surname, descriptive, or trade dress, build and document acquired distinctiveness (and, for trade dress, be ready to meet a functionality attack) so the mark can withstand a § 2(e) challenge and carry real strength. File the Section 8 and 15 affidavits on time and reach incontestability as soon as the law allows, because incontestability removes the descriptiveness and surname grounds from the table. And when a fraud claim arrives, hold the challenger to the demanding Bose standard rather than conceding that an honest inaccuracy is deceptive.

For both sides, the unifying principle is that the register is supposed to reflect the truth about a mark, and cancellation is the mechanism for correcting it when it does not. The grounds are statutory and shared across the TTAB and the courts; the presumptions and incontestability shape which grounds are even available and how hard they are to prove; Bose has made fraud a rarely successful claim; the low use-in-commerce threshold means nonuse theories must rest on genuine abandonment rather than modest volume; and B&B Hardware means the forum where an issue is first fully litigated may decide it everywhere. Understand the grounds, respect the burdens, choose the forum with care, and the most powerful move in trademark litigation—erasing the other side's registration—becomes a disciplined, predictable exercise rather than a long shot.

Frequently Asked Questions

Can a federal court cancel a trademark registration, or only the USPTO? A federal district court can. Section 37 of the Lanham Act, 15 U.S.C. § 1119, empowers a court—in any action involving a registered mark—to order cancellation (or restriction) of a registration and to otherwise rectify the register, with the order certified to the Director of the USPTO. The catch is that § 1119 is not a standalone basis for federal jurisdiction; cancellation must ride along with an independent controversy, usually an infringement claim, most often arriving as a counterclaim.

Can an incontestable registration still be cancelled? Yes. Incontestability removes the weaker, registrability-based grounds (principally mere descriptiveness without secondary meaning, and surname challenges) but it does not protect against the fundamental grounds. An incontestable registration can still be cancelled because the mark has become generic, is functional, has been abandoned, was obtained or maintained by fraud, is being used to misrepresent source, or was never used. 15 U.S.C. §§ 1064(3), (5), (6), 1115(b).

How long do I have to bring a cancellation claim? It depends on the ground. The registrability grounds (likelihood of confusion, descriptiveness, surname, dilution, and most § 2 bases) must generally be raised within five years of registration. 15 U.S.C. § 1064(1). The fundamental grounds—genericness, functionality, abandonment, fraud, misrepresentation, and nonuse—may be raised at any time. Because the five-year window can close on a strong registrability theory, challengers should not delay.

Is fraud on the USPTO a realistic cancellation ground? Rarely, after In re Bose Corp., 580 F.3d 1240 (Fed. Cir. 2009). Fraud requires a knowing, material misrepresentation made with subjective intent to deceive, proven by clear and convincing evidence "to the hilt." Negligence—even gross negligence—is not enough. Overclaiming goods usually leads to restriction of the registration, not cancellation for fraud. Plead fraud only with real evidence of a deliberate lie, and expect it to be a trial issue, not a summary-judgment winner.

How much use does a registrant need to defeat a nonuse attack? Surprisingly little. In Christian Faith Fellowship Church v. adidas AG, 841 F.3d 986 (Fed. Cir. 2016), the sale of two hats across state lines was enough "use in commerce." The Act requires bona fide use in the ordinary course of trade, not a minimum sales volume, so a "they barely used it" theory is weak. The stronger attack is genuine abandonment—three or more consecutive years of nonuse (which creates a rebuttable presumption under 15 U.S.C. § 1127) with no documented intent to resume.

What is the difference between the TTAB and a district court for cancellation? The TTAB can decide only the right to register; it cannot decide the right to use, find infringement, or award an injunction or damages. A district court can do all of that and rectify the register in the same case. The TTAB is typically cheaper and runs on a paper record, but its decisions can have issue-preclusive effect in later litigation under B&B Hardware, Inc. v. Hargis Industries, Inc., 575 U.S. 138 (2015). For a pure nonuse complaint, the USPTO's newer ex parte expungement and reexamination proceedings may be the cheapest route of all.

Related Articles

This article is provided for general informational purposes and does not constitute legal advice. Cancellation is fact-specific and depends on the registration's age, status, and the available grounds, as well as on the rules of the chosen forum; consult qualified trademark counsel about any particular matter.