Imagine you are a federal judge, and a company walks into your courtroom claiming it owns the color of its product. Not the name. Not the logo. The color. A specific shade of green-gold, used on the felt pads of a dry-cleaning press. The other side says that is absurd--nobody can own a color, any more than they can own the alphabet or the number seven. Run out the logic, opposing counsel argues, and we will be reduced to fighting over whose blue is bluer. Who wins?

In 1995, the Supreme Court said the color owner wins, and in doing so it taught the rest of us something profound about what trademark law is actually for. The case was Qualitex Co. v. Jacobson Products Co., 514 U.S. 159 (1995), and the answer turned not on metaphysics but on a single, practical question: does that green-gold tell consumers who made the pressing pads? If it does--if shoppers have come to see the color and think "ah, that's the Qualitex brand"--then the color is doing a trademark's job, and the law will protect it. If it does not, it is just a color.

That is the whole secret of this subject. Trademark law does not protect words, pictures, shapes, or colors because they are words, pictures, shapes, or colors. It protects them because, and only to the extent that, they identify a source. Get that idea firmly in mind and the rest of trademark's strange-looking rules--the spectrum that rates names from "generic" to "fanciful," the bar against owning useful features, the fights over scandalous logos and a former president's surname--all snap into focus. They are not a grab bag of arbitrary technicalities. They are all expressions of one question asked over and over: does this thing identify a source, and may the law let one company own it without choking off legitimate competition?

This article is Part 1 of a four-part overview of United States trademark law. Here we tackle the threshold question: what can be a trademark in the first place? Part 2 takes up the substantive standards for protection--distinctiveness in depth, secondary meaning, and the use requirement. Part 3 covers obtaining protection and licensing. Part 4 covers infringement and related rights. If you want the gentlest possible on-ramp before diving in here, the firm's trademark basics primer is a good warm-up, and the trademark FAQs answer the questions clients ask most. By the end of this piece you will understand the categories of things that can be marks, the famous distinctiveness spectrum that decides which of them earn protection, the special trouble with generic and functional designations, and the statutory bars that can sink even a distinctive mark.

What a Trademark Is, and What It Is For

A trademark is any word, name, symbol, device, or combination of them that a person uses, or intends to use, in commerce to identify and distinguish goods and to indicate their source. That definition is not ours; it is, almost verbatim, the definition Congress wrote into Section 45 of the Lanham Act, 15 U.S.C. § 1127. A service mark is the same thing for services rather than goods--the Geico gecko, the AAMCO jingle, the McDonald's golden arches selling the experience of fast food rather than a tangible widget--and by long convention lawyers use "trademark" or simply "mark" to cover both. We will do the same here unless the distinction matters.

Notice what that definition emphasizes: identify, distinguish, indicate source. A trademark is fundamentally a pointer. It points away from the symbol and toward the unseen producer behind the goods. When you grab a red can of soda with cursive white script, you are not buying the script--you are relying on the script to tell you that the familiar fizzy formula is inside, made by the company that makes it consistently. The mark is a shorthand, a compressed promise about who stands behind the product. The law sometimes calls this the mark's "source-identifying function," but you can think of it more plainly as the difference between a label that says what something is and a label that says whose it is. Only the second kind can be a trademark.

This is why trademark law is usually described as serving two masters at once, and the leading treatise, McCarthy on Trademarks and Unfair Competition, organizes the entire field around this dual purpose. First, marks protect consumers. They lower the cost of shopping. You do not have to chemically test every aspirin tablet, because the brand on the bottle stands in for a track record. The mark lets you find again the thing you liked last time--and avoid the thing you did not. Second, marks protect producers. They let a business reap the goodwill it has earned by making good products, and they stop a free rider from slapping the same brand on inferior knockoffs and siphoning off that reputation. The Supreme Court put both halves together in Qualitex: trademark law "reduce[s] the customer's costs of shopping and making purchasing decisions," and at the same time "helps assure a producer that it (and not an imitating competitor) will reap the financial, reputation-related rewards associated with a desirable product." 514 U.S. at 163-64.

Here is the consequence that trips up newcomers, including many smart entrepreneurs: trademark law does not reward creativity for its own sake. That is copyright's department. You do not get a trademark because you thought of a clever name; you get one because that name has come to function as a source identifier in the marketplace. A breathtakingly original product name that nobody has used yet is, on the day you coin it, only a potential mark. Whether it becomes a real one depends on use and on the rules we are about to explore. The flip side is just as important: an utterly unoriginal word--a real, ordinary, dictionary word like "Apple"--can be one of the strongest marks on earth, so long as it is being used to label something the word has nothing to do with. Strength in trademark is not about novelty. It is about the gap between what a designation means and what it labels. For the bigger picture of how this fits alongside patents and copyrights, see copyright vs. trademark--what is the difference and the firm's overview of the legal protection of software, which lays out how the regimes layer.

One more foundational point before we open the toolbox of mark types. Federal registration with the United States Patent and Trademark Office (USPTO) is not what creates trademark rights. In the U.S., rights spring from use in commerce--actually using the mark to sell goods or services--so an unregistered user can own a mark under common law. We explore that in trademark rights under common law. Registration is enormously valuable (nationwide constructive notice, a legal presumption of validity and of the owner's exclusive right to use the mark, the path to incontestability after five years, the ability to recover certain enhanced remedies, and more), but it is a way of strengthening and recording rights, not the origin of them. The mechanics of getting registered live in the trademark process, how to file a trademark application with the USPTO, and the trademark registration guide; Part 3 of this series ties it together. For now, our concern is subject matter: assuming you will use a designation as a brand, can it be a trademark at all?

The Toolbox: Things That Can Serve as Marks

The Lanham Act's "word, name, symbol, or device" language is famously roomy. The Supreme Court has read "symbol" and "device" to "include almost anything at all that is capable of carrying meaning." Qualitex, 514 U.S. at 162. Practitioners sometimes call this the "universe of registrable subject matter," and the point of the phrase is that there is no closed list. There is only a test, applied to whatever shows up: can this thing identify and distinguish source, and is it non-functional? If the answer is yes, the only remaining questions are about distinctiveness and the statutory bars. Let us tour the toolbox.

Words and Names

The most common marks are words: APPLE for computers, NIKE for shoes, GOOGLE for search. Word marks are the workhorses of the system, and they are usually registered in "standard characters"--meaning the owner claims the word itself, in any font, size, or color, rather than a particular stylized rendering. That is the broadest possible protection for a word, because it covers every way the word might be displayed.

Personal names can be marks too, though they carry special baggage. A surname is generally treated as not inherently distinctive--more on that below--because the law is wary of stopping people from honestly using their own names, and because a surname tells the public little about source until the public learns to associate it with one. "Ford," "Disney," and "McDonald's" are all surnames that became towering brands only after years of use built the association. If you are weighing whether to brand a venture under your own name, see trademarking your own name, which walks through the surname rule and its consequences.

Logos, Designs, and Symbols

A logo--the Nike swoosh, the Apple silhouette, the Target bullseye--is a classic "symbol or device." Design marks can be purely graphic, purely textual in a stylized form (the cursive Coca-Cola script), or a composite of both (a word inside a distinctive shield, say). They often coexist with a word mark covering the same brand, which is why sophisticated owners frequently register both the standard-character word and the stylized logo. The two registrations protect different things: the word registration follows the word wherever it appears; the design registration protects the specific visual treatment even when the words are stripped away (think of the Nike swoosh with no "Nike" anywhere near it).

Slogans and Taglines

A slogan can be a mark when it does more than describe--when it functions as a source identifier. "Just Do It," "I'm Lovin' It," and "The Ultimate Driving Machine" are registered marks that consumers instantly trace to a single company. The USPTO refuses slogans that are merely informational, ornamental, or commonplace expressions consumers would not perceive as brands. An examiner is unlikely to register "Have a Nice Day" for greeting cards or "Drive Safely" for bumper stickers, because the public reads those as messages, not as badges of origin. This "failure to function" objection has become one of the USPTO's most-cited refusals, and it is worth understanding that it is conceptually distinct from a distinctiveness refusal: the problem is not that the phrase is weak, but that consumers do not perceive it as a trademark at all. The line between an enforceable tagline and an unprotectable bit of puffery is, in the end, the distinctiveness and failure-to-function question we reach in the next section.

Trade Dress: The Look and Feel

Trade dress is the overall appearance and image of a product or its packaging--shape, color combinations, textures, graphics, even the décor of a restaurant or the layout of a retail store. The Lanham Act protects it as a form of mark when it identifies source. The Supreme Court confirmed in Two Pesos, Inc. v. Taco Cabana, Inc., 505 U.S. 763 (1992), that the festive, distinctive interior of a Mexican restaurant chain could be protectable trade dress, and that inherently distinctive trade dress is protectable without separate proof that consumers associate it with a source. Trade dress comes in two flavors with different rules, and the difference is worth internalizing because it determines whether you need years of marketplace evidence before you have anything to enforce:

  • Product packaging (the box, the bottle, the wrapper, the restaurant's interior) can be inherently distinctive, like an arbitrary word. An unusual, distinctive package design is protectable on use.
  • Product configuration--the design of the product itself, like the shape of a chair or a phone--can never be inherently distinctive. The Court held in Wal-Mart Stores, Inc. v. Samara Brothers, Inc., 529 U.S. 205 (2000), that product-design trade dress always requires proof of secondary meaning (acquired distinctiveness) before it can be protected, because consumers do not ordinarily assume that the shape of a product signals its maker. As Justice Scalia put it, consumers are "predisposed" to see a product's design as serving a purpose other than source identification.

When packaging and configuration blur together--and they often do, as with a uniquely shaped bottle that is both container and product silhouette--Wal-Mart told courts to err on the side of treating the trade dress as product design and demand secondary meaning. Trade dress is one of the richest and most litigated corners of this field. For a deep treatment, see the intricate world of trade dress protection--balancing brand identity and functional design; for the long arc of the case law, the evolution of trade dress protection for product design and configuration; and for the boundary line with industrial-design patents, design patents vs. trade dress protection for product configurations. We will return to trade dress below when we discuss its great limiting principle, functionality.

Color

We opened with color, and Qualitex settled the basic question: a single color can be a trademark, provided it has acquired secondary meaning and is not functional. There is no special rule that color is too "basic" to own. Tiffany's robin's-egg blue, UPS brown, T-Mobile magenta, Owens-Corning pink insulation, and Christian Louboutin's red shoe soles are all protected color marks. But the Qualitex logic carries two important caveats baked in. Because a single color, like a product shape, is never inherently distinctive, a color-mark owner must always prove that consumers see the color as a brand. And the color cannot be functional--more on the famous example of why John Deere may struggle to own green for farm equipment in a moment.

There is a wrinkle worth knowing, because it shows how the law keeps refining even settled rules. In In re Forney Industries, Inc., 955 F.3d 940 (Fed. Cir. 2020), the Federal Circuit distinguished between single colors and multi-color combinations used on product packaging. The court held that a particular combination and arrangement of colors on packaging can be inherently distinctive--protectable without secondary meaning--if consumers would perceive that specific arrangement as indicating source. So Wal-Mart's "never inherently distinctive" rule still governs a single color on a product, but a distinctive multi-color design on a package may get the more generous packaging treatment from Two Pesos. The reconciling principle, as always, is consumer perception: the more a feature looks like deliberate branding rather than an inherent attribute of the goods, the closer it edges toward inherent distinctiveness.

Sound

Sounds can be marks. The NBC three-note chime is one of the oldest registered sound marks; so are the THX "Deep Note," the Intel five-tone bong, the MGM lion's roar, and the Netflix "ta-dum." A sound mark must, like the rest, identify source--and here the law draws a line that mirrors the distinctiveness spectrum. Under the test the TTAB articulated in In re General Electric Broadcasting Co., 199 U.S.P.Q. 560 (T.T.A.B. 1978), a sound that is "unique, different, or distinctive" can be inherently distinctive and registrable on use, while a "commonplace" sound (one consumers are already used to hearing in everyday life) requires proof of secondary meaning. A short, arbitrary jingle that a streaming app plays on launch is the easy case. The actual noise a product makes in normal operation is the hard case--and if the sound is essential to the product's function, it crosses from "merely commonplace" into outright barred. In In re Vertex Group LLC, 89 U.S.P.Q.2d 1694 (T.T.A.B. 2009), the Board refused to register the alarm sound of a children's personal-security alarm: the sound was not distinctive, the public did not recognize it as a source indicator, and--decisively--it was functional, because an audible alarm is essential to the product's purpose and the applicant's particular pattern of pulses and silence actually made the alarm work better. A sound that helps the product do its job cannot be owned.

Scent

Yes, scent. The USPTO has registered scent marks--famously, "a high impact, fresh, floral fragrance reminiscent of plumeria blossoms" applied to sewing thread and embroidery yarn. That registration came out of In re Clarke, 17 U.S.P.Q.2d 1238 (T.T.A.B. 1990), where the Board reversed a refusal and accepted that a scent with no functional relationship to the goods could serve as a mark. Others have followed: a cherry scent for synthetic vehicle lubricants, a bubble-gum scent for flip-flops, and--registered to Hasbro in 2018--the unmistakable smell of Play-Doh, described in the registration as "a sweet, slightly musky, vanilla fragrance, with slight overtones of cherry, combined with the smell of a salted, wheat-based dough." The bar is high. Because a scent is presumed to be a characteristic of the goods rather than a brand, it can never be inherently distinctive, and the applicant must show substantial acquired distinctiveness. See In re N.V. Organon, 79 U.S.P.Q.2d 1639 (T.T.A.B. 2006) (taste; same logic). And a scent is barred outright if it is functional. A perfume cannot trademark its own smell--that is the product, the whole reason anyone buys it. But a scent applied to goods whose ordinary purpose has nothing to do with smell can, in principle, mature into a source identifier.

Taste, Texture, Motion, Holograms, and Other "Non-Traditional" Marks

Taste is the frontier nobody has crossed in the U.S.: in Organon, the TTAB doubted that the flavor of an orally administered medicine could ever function as a mark, partly because consumers do not encounter the flavor until after purchase (so it cannot influence the buying decision) and partly because flavor is so bound up with the product itself. Texture has fared a little better as a component of trade dress. And motion is well established: the Lamborghini scissor-door movement, the Columbia Pictures torch-lady sequence, the Twentieth Century Fox searchlights, and the way certain software animates have all been registered as motion marks. Holograms and color-changing features have been registered in appropriate cases.

Practitioners group color, sound, scent, taste, texture, motion, and holograms under the umbrella non-traditional trademarks, and the practical lesson is the same across all of them, and worth stating as a rule of thumb: the more unusual the mark, the heavier two burdens become. First, distinctiveness--because most non-traditional features start life looking like ordinary attributes of the goods, the examiner and the courts will demand convincing proof (surveys, "look-for" advertising, long exclusive use) that consumers actually treat the feature as a brand. Second, functionality--because non-traditional features so often do something for the product, the functionality doctrine is aimed at them with special force. The closer the feature sits to the working guts of the product, the less likely trademark will ever protect it.

Service Marks, Trade Names, and the Things That Look Like Marks but Are Not

A quick but important clarification, because clients confuse these constantly. A service mark identifies the source of services rather than goods--brokerage, fast food, package delivery, search--and is governed by the same rules as a trademark. A trade name, by contrast, is the name a business operates under (the corporate or "doing business as" name). A trade name is not, by itself, a trademark. "Acme Coffee Roasters, LLC" on your incorporation papers and your storefront sign is a trade name; "Acme" stamped on the bag of beans as the brand is a trademark. The same word can be both, but only the source-identifying use on the goods or services creates trademark rights. Likewise, registering a domain name with a registrar, or forming an LLC with a state, is an administrative act that creates no trademark rights at all. We return to this in the FAQ, because the misunderstanding is so common and so costly.

Two Specialized Cousins: Collective Marks and Certification Marks

The Lanham Act also recognizes two specialized categories that are easy to confuse with each other and with ordinary marks. Both are protectable "in the same manner and with the same effect" as trademarks, 15 U.S.C. § 1054, but each does a different job.

A collective mark comes in two sub-types. A collective membership mark simply indicates that its user belongs to an organization--a union, a trade association, a fraternal order, a cooperative (the "ILGWU" union label, the "REALTOR" designation reserved for members of the National Association of Realtors). A collective trademark or service mark is used by the members of a collective on their goods or services to indicate that the member produced them. The defining feature is membership: the mark signals "produced by, or affiliated with, a member of this group."

A certification mark is the genuinely strange one, and the strangeness is the point. It is used by someone other than its owner to certify a characteristic of goods or services--their regional origin, the material or mode of manufacture, the quality, or that the work was performed by members of a particular organization. 15 U.S.C. § 1127. Think "UL" for safety-tested electronics (owned by UL LLC, used by any manufacturer whose products pass UL's tests), "ENERGY STAR" for efficient appliances, the "OU" or "K" symbols for kosher goods, "ROQUEFORT" for cheese cured in the caves of a particular French region, or the "WOOLMARK" for pure new wool. The certification can even attest to a person's qualifications, like a board certification in a medical specialty--but only if the public genuinely perceives it as a certification, with real, ongoing review behind it, and not as a mere title or training-course credential. See Medical-Legal Consulting Inst., Inc. v. Evans & Assocs., Ltd., 2007 WL 9752840 (S.D. Tex. 2007) (designation that was just a label for people who took the owner's course was not a valid certification mark).

A certification mark carries three rules that flow from its odd structure, and they are heavily tested in practice:

  • The owner does not sell the certified goods. The owner sets a standard and authorizes anyone who meets it to use the mark. If the owner starts producing the goods it certifies, it can lose the registration--because it can no longer be a neutral certifier.
  • The owner must control the certification and cannot discriminate. The owner must police use of the mark so it continues to mean what it says, but it also may not refuse certification to anyone who actually meets the published standard. A certifier that arbitrarily withholds its mark from a qualifying producer risks cancellation.
  • The mark certifies a characteristic, not a single source. Where a trademark says "made by this one company," a certification mark says "meets this standard, whoever made it."

For brand owners exploring these categories, the unifying rule never changes: source (or, for certification marks, standard) identification, plus no functionality, plus enough distinctiveness. Geographic certification marks like ROQUEFORT are the one place where the usual hostility to geographic terms flips--there, the whole point is to certify regional origin, so geographic meaning is a feature rather than a bug.

The Distinctiveness Spectrum: The Heart of the Matter

Now to the single most important idea in all of trademark subject matter. Even among words--the most ordinary marks--the law treats some as strong and instantly protectable and others as weak or unprotectable. The framework that sorts them is the distinctiveness spectrum, and it comes from a 1976 opinion by the great Judge Henry Friendly: Abercrombie & Fitch Co. v. Hunting World, Inc., 537 F.2d 4 (2d Cir. 1976). Almost every American trademark lawyer thinks in Abercrombie categories, examiners apply them daily, and courts cite them reflexively, so it is worth learning them cold.

The spectrum arranges designations into five categories, from weakest to strongest:

  1. Generic. The common name of the thing itself. "Bicycle" for bicycles, "Computer" for computers, "Convenience Store" for convenience stores. A generic term can never be a trademark--not even with the most extensive advertising and the most devoted customers in the world. The reason is competitive necessity: if one seller owned the word "Apple" for apples, no other apple seller could describe its own fruit. Genericness is the death sentence of trademark law, and we devote a full section to it below.

  2. Descriptive. A term that directly describes a feature, quality, function, characteristic, or purpose of the goods. "Creamy" for yogurt, "Speedy" for delivery, "Cold and Creamy" for ice cream, "Holiday Inn" (a place to stay on holiday), "Vision Center" for an eyewear shop. A descriptive term is not protectable as a trademark on its own--but it can climb into protection if it acquires secondary meaning, meaning consumers have come to associate it primarily with a single source rather than with the described quality. "American Airlines" and "Sharp" televisions are descriptive marks that earned protection through long, heavy use. (Surnames and geographic terms are treated as a special case in the descriptive zone--ordinarily unprotectable absent secondary meaning, which is why "McDonald's" needed years of use to lock in.) Part 2 of this series treats secondary meaning at length, because it is the mechanism that lets a whole second category of marks survive.

  3. Suggestive. A term that suggests a quality of the goods but requires the consumer to exercise imagination, thought, or perception to make the connection. "Coppertone" for suntan lotion, "Netflix" for streaming, "Greyhound" for fast bus travel, "Microsoft" for software for microcomputers, "Jaguar" for a sleek car. Suggestive marks are inherently distinctive--protectable immediately, with no proof of secondary meaning required. The line between descriptive and suggestive is famously slippery; courts use the "imagination test" (does the consumer need a mental leap to get from the word to the product?) and the "competitors' need" test (do rivals need this term to describe their own goods?). A huge share of trademark prosecution fights are fought right on that border, because suggestive is the cheapest protectable category--inherently distinctive, but using a word with marketing resonance.

  4. Arbitrary. A real word used in a way unrelated to its ordinary meaning. "Apple" for computers, "Camel" for cigarettes, "Amazon" for an online store, "Shell" for gasoline. Inherently distinctive and strong.

  5. Fanciful. A coined or invented word that exists only as a mark. "Kodak," "Xerox," "Exxon," "Pepsi," "Verizon," "Häagen-Dazs." These are the strongest marks of all, because they carry no prior meaning to dilute their source-identifying power, and no competitor has any legitimate need to use them. The trade-off is marketing: a fanciful mark tells the consumer nothing about the product, so the owner has to spend to build meaning from zero.

Categories 3, 4, and 5--suggestive, arbitrary, and fanciful--are inherently distinctive and protectable the moment they are used as marks. Category 2, descriptive, is protectable only with secondary meaning. Category 1, generic, is never protectable. Memorize that hierarchy and you can predict the outcome of a large fraction of trademark disputes.

A worked example makes the spectrum concrete. The following is a hypothetical. Suppose Acme Corp. is launching a line of artisanal coffee. Consider the candidate names:

  • "Coffee" — generic, dead on arrival.
  • "Fresh Roast" — descriptive; Acme cannot stop competitors from using ordinary roasting language, and would need years of secondary meaning to claim it.
  • "Morning Mountain" — suggestive; it hints at a wake-up product and a coffee-growing highland without describing the goods, and would be inherently distinctive.
  • "Penguin" — arbitrary; penguins have nothing to do with coffee, so it is strong and protectable instantly.
  • "Zafloo" — fanciful; an invented word, the strongest of all.

A good trademark lawyer counseling Acme would steer it toward the right end of the spectrum, because a strong mark is not just easier to register--it is easier and cheaper to enforce later, and far less likely to collapse under a validity challenge. The whole architecture of clearance and selection is built on this spectrum; see how to conduct a comprehensive trademark clearance search and the firm's guidance on when to trademark your brand.

One subtlety worth flagging, because it is the source of endless confusion: distinctiveness is not assessed in the abstract but in relation to the goods or services. "Apple" is arbitrary for computers but generic for fruit; "Ivory" is arbitrary for soap but descriptive for products actually made of ivory; "Sharp" is descriptive for knives but merely suggestive (and strong) for televisions. The same word can sit at different points on the spectrum depending on what it labels. This is also why two businesses in unrelated fields can sometimes coexist with the same name (Delta the airline and Delta the faucet; United the airline and United the moving company), a point we develop in Part 4 on confusion. The mark is never the word alone; it is the word plus the goods.

The Generic Trap--and the Booking.com Twist

Genericness deserves its own spotlight, both because it is so unforgiving and because the Supreme Court recently complicated it in an instructive way.

A term is generic if the relevant public understands it primarily as the name of a class of goods or services rather than as a brand. The test is sometimes framed as "the primary significance of the term to the relevant public"--do consumers hear the word as a what or a whose? Genericness can be a designation's original condition (you simply chose the common name) or a fate it suffers later. Genericide is the grim process by which a once-strong mark becomes so successful that the public adopts it as the everyday word for the product category, and the owner loses all rights. The graveyard is well populated: "aspirin," "escalator," "thermos," "cellophane," "dry ice," "yo-yo," and "trampoline" all began as proprietary brands and ended as generic nouns. There is a bitter irony here--the better your marketing works, the more your brand risks becoming the public's word for the whole category, which is the one outcome that kills the mark.

That threat is exactly why you see legal teams insisting that "Xerox" is a company, not a verb ("photocopy, please, not Xerox"), why Adobe polices "Photoshop" as a noun rather than a thing you do to an image, and why Google has spent real money discouraging "to google" as a synonym for "to search." Google's vigilance paid off: it survived a genericide challenge in Elliott v. Google, Inc., 860 F.3d 1151 (9th Cir. 2017), where the court held that verb use alone does not make a mark generic, so long as the public still understands the word as a source identifier for the goods. The lesson for brand owners is concrete: use your mark as an adjective modifying a generic noun ("KLEENEX tissues," not "a kleenex"), never as the noun or verb itself, and correct misuse in your own materials and, where feasible, in the press. For the doctrine in depth, see Part 2's treatment of distinctiveness and the firm's clearance materials.

The fascinating modern wrinkle is what happens when you bolt a generic term onto a generic top-level domain. The travel company Booking.com sought to register "Booking.com." The USPTO refused, reasoning that "booking" is generic for hotel-reservation services and ".com" is generic for a commercial website, so generic + generic must equal generic--just as "Generic Company, Inc." or "The Wine Company" would be unregisterable. The logic had a certain algebraic neatness. The Supreme Court disagreed in United States Patent and Trademark Office v. Booking.com B.V., 591 U.S. 549 (2020). The Court rejected a per se rule that a "generic.com" term is automatically generic. The dispositive question, the Court held, is whether the term as a whole, in the eyes of consumers, signifies the class of goods or a particular source. Because consumers did not understand "Booking.com" to mean online hotel-reservation services in general--they understood it to mean one specific company--the composite was not generic and could be registered (on a showing of secondary meaning, since it remained, at most, descriptive). 591 U.S. at 554-58.

The Court was careful to cabin the win. It did not bless every "generic.com" mark; it warned that such marks may be weak, that the secondary-meaning bar is real, and that classic trademark limits--like descriptive fair use, which lets competitors honestly use the descriptive words--will keep "Booking.com" from monopolizing the word "booking." A rival can still run advertisements about "booking a hotel" all day long. Justice Breyer dissented, worried that the ruling hands near-monopolies to first movers who grabbed the obvious domain name. For our purposes, Booking.com teaches a clean lesson that runs through this entire article: genericness, like distinctiveness, is ultimately a question of consumer perception, decided on evidence (surveys, dictionaries, media usage, the way the public actually talks), not on a tidy formula. There is no shortcut around what real people actually understand the words to mean.

The Functionality Bar: Why You Cannot Trademark a Better Mousetrap

Return now to that color, that bottle shape, that scent. A second great limiting principle stands guard over all of them, and especially over trade dress: the functionality doctrine. The rule is simple to state and crucial in practice: a product feature is not protectable as a trademark if it is functional, no matter how distinctive it is or how strongly consumers associate it with a source. The Lanham Act now codifies the bar directly, listing functionality as a ground for refusal and for cancellation. 15 U.S.C. § 1052(e)(5).

Why does the law refuse protection even to a famous, source-identifying feature? Because trademark protection is potentially perpetual--you can renew a mark every ten years forever, so long as you keep using it. Patent protection, by contrast, lasts only twenty years (utility) or fifteen (design) and then the invention belongs to the public. If a company could use trademark to lock up a useful product feature, it could win an eternal monopoly on functional innovation, evading the patent bargain entirely. Functionality is trademark law's way of refusing to let brand protection swallow the public's right to copy useful, unpatented (or expired-patent) features. The doctrine, in other words, is a border guard between two regimes that would otherwise overrun each other.

The Supreme Court has given us the working test. Under Inwood Laboratories, Inc. v. Ives Laboratories, Inc., 456 U.S. 844, 850 n.10 (1982), a feature is functional if it is "essential to the use or purpose of the article or if it affects the cost or quality of the article." That is the test for utilitarian functionality, and every circuit applies it. In TrafFix Devices, Inc. v. Marketing Displays, Inc., 532 U.S. 23 (2001), the Court sharpened the analysis, holding that when a feature was the subject of an expired utility patent, that patent is strong evidence the feature is functional, and the would-be trademark owner bears a heavy burden to show otherwise. (The road-sign stand in TrafFix had a dual-spring mechanism that the very patent claims described as the useful innovation; that was close to fatal.) TrafFix also recognized a second branch, aesthetic functionality: even a feature that is not mechanically essential can be functional if granting one company exclusive rights to it would put competitors at a significant non-reputation-related disadvantage. 532 U.S. at 32-33.

Below the Supreme Court, the lower courts often run features through the four Morton-Norwich factors (from In re Morton-Norwich Products, Inc., 671 F.2d 1332 (C.C.P.A. 1982)) to organize the evidence: (1) the existence of a utility patent disclosing the utilitarian advantages of the design; (2) the owner's own advertising touting the design's utilitarian advantages; (3) the availability to competitors of functionally equivalent alternative designs; and (4) whether the design results from a comparatively simple or cheap method of manufacture. Note the trap built into factor (2): a company that has spent years advertising how well its design works may have authored the evidence that destroys its own trademark claim. Marketing and legal should compare notes before the brochure goes to print.

Three illustrations make the doctrine click. First, recall Qualitex: the green-gold color was protectable because it was not functional--the color did nothing for the pressing pad's performance and served only to identify Qualitex. Had the color helped the pad work better (say, by hiding stains, a benefit any cleaner would want), it would have been functional and unprotectable. Second, recall In re Vertex Group: the security-alarm sound was barred under Inwood precisely because the noise was essential to--and made the product better at--alarming. Third, consider John Deere's green. The shade is iconic, and consumers absolutely associate it with Deere. But courts have wrestled with whether farmers want their equipment to match (so the implements pulled behind a green tractor look coordinated), which would make the color aesthetically functional for farm machinery. That tension is exactly the aesthetic-functionality problem TrafFix flagged, and it shows why even a wildly famous feature can fail the subject-matter test. Aesthetic functionality is also why a court might let one company own the Louboutin red sole as a contrasting accent but not as the color of an entire monochrome red shoe--where red is just the color you wanted the shoe to be.

The practical upshot for product designers is blunt: if a feature is there because it works better, costs less, or is the only good way to do the job, do not count on trademark to protect it. The right tool is a utility patent (for how it works) or a design patent (for how it ornamentally looks), each with its own bargain and expiration date. We compare these instruments in design patents vs. trade dress protection for product configurations and the intersection of design and utility patents. Functionality is the doctrine that keeps these regimes in their lanes.

The Section 2 Bars: When Even a Distinctive Mark Is Refused

Suppose a designation clears every hurdle so far: it is the right kind of thing, it is distinctive, it is not functional. It can still be refused federal registration if it runs into one of the statutory bars in Section 2 of the Lanham Act, 15 U.S.C. § 1052. These are the grounds an examining attorney cites most often, and a few of them have been at the center of dramatic recent Supreme Court litigation. Let us take the most important ones.

Section 2(d): Likelihood of Confusion

Section 2(d) bars registration of a mark that so resembles an existing registered or prior-used mark "as to be likely . . . to cause confusion, or to cause mistake, or to deceive." This is the single most common ground of refusal, and it is the registration-side cousin of the infringement test. It is why clearance searches matter so much: a great mark on the merits is worthless if a confusingly similar mark already sits on the register for related goods. The multifactor confusion analysis is the subject of Part 4 and of navigating the maze of trademark confusion--key considerations for brand owners; for the screening that catches 2(d) problems early, see how to conduct a comprehensive trademark clearance search. A 2(d) refusal can sometimes be overcome with a consent agreement or coexistence agreement with the prior owner, but never count on one.

Section 2(e): Descriptive, Deceptively Misdescriptive, Geographic, and Surname Marks

Section 2(e) gathers several related refusals. It bars marks that are merely descriptive or deceptively misdescriptive of the goods; primarily geographically descriptive; primarily geographically deceptively misdescriptive; primarily merely a surname; or functional. Most of these overlap with the distinctiveness spectrum we have already walked: a 2(e)(1) descriptiveness refusal is simply the registration-side expression of the rule that descriptive terms need secondary meaning. The cure for many 2(e) refusals is to prove acquired distinctiveness under Section 2(f)--essentially, secondary meaning--or to register on the Supplemental Register and build toward the Principal Register over time as the public learns the mark. (The two registers are explained in Part 3.) But note the two exceptions that no amount of secondary meaning can fix: a deceptive mark under 2(a) (discussed next) and a functional feature under 2(e)(5). Functional matter is barred, full stop, because the policy behind functionality is not about consumer recognition at all--it is about keeping useful features free for competition.

Section 2(a): The Bar That Keeps Reshaping the First Amendment

Section 2(a) historically barred a grab-bag of marks: those that are deceptive, that falsely suggest a connection with persons or institutions, that are scandalous or immoral, that disparage persons or beliefs, and certain geographic indications for wines and spirits. Two of those grounds--disparagement and scandalousness/immorality--collided with the First Amendment in a remarkable pair of recent decisions, and they have left the provision substantially rewritten.

In Matal v. Tam, 582 U.S. 218 (2017), an Asian-American dance-rock band called "The Slants" sought to register its name, deliberately reclaiming an ethnic slur as an act of pride. The USPTO refused under the disparagement clause. A unanimous Supreme Court struck that clause down as unconstitutional viewpoint discrimination, holding that trademark registration is not government speech, and that the government may not refuse to register a mark because it disapproves of the message. The decision had immediate ripple effects--it is what cleared the legal cloud over the long-running "Washington Redskins" registration dispute, which we discuss in Washington Redskins trademark troubles--disparagement and the aftermath.

Two years later, Iancu v. Brunetti, 588 U.S. 388 (2019), finished the job on the neighboring "immoral or scandalous" clause. The applicant sought to register the streetwear mark "FUCT." The Court struck down the immoral/scandalous bar as well, again as impermissible viewpoint discrimination, because it disfavored marks the government found offensive while favoring "wholesome" ones. (Some Justices suggested a narrower, viewpoint-neutral ban on vulgar or profane matter might survive; Congress has not yet enacted one.) After Tam and Brunetti, the disparagement and immoral/scandalous prongs of 2(a) are effectively dead letters. What survives in 2(a) are the viewpoint-neutral grounds: deceptiveness and false suggestion of a connection. Those bars do not care whether the government likes the message; they care whether the mark lies to consumers or fraudulently claims a tie to someone famous.

Then, in Vidal v. Elster, 602 U.S. 286 (2024), the Court addressed a different bar and reached the opposite result--an important corrective for anyone tempted to read Tam and Brunetti as having opened the floodgates. Steve Elster sought to register "TRUMP TOO SMALL" for shirts, a political quip. The USPTO refused under Section 2(c), which bars registering a mark that uses the name of a particular living individual without that person's written consent. Elster argued the refusal violated his free speech. The Supreme Court unanimously upheld the names clause. The key distinction: unlike the disparagement and scandalousness bars, the names clause is viewpoint-neutral--it does not turn on whether the government likes the message, only on whether you are appropriating a living person's name for commercial branding. Drawing on a long historical tradition of restricting the trademarking of others' names, the Court held the names clause does not offend the First Amendment. Elster thus marks the boundary cleanly: viewpoint discrimination in trademark registration is forbidden (Tam/Brunetti), but viewpoint-neutral registration conditions rooted in trademark's history (like the names bar) are permissible. Critically--and this point is often missed--none of these cases stopped anyone from using the words on shirts; they were only about the privileges of federal registration. Mr. Brunetti could sell his clothing; Mr. Elster could print his slogan. The fight was over the government register, not the right to speak.

Other Bars Worth Knowing

A handful of additional Section 2 and related bars round out the picture. Section 2(b) bars registration of the flag, coat of arms, or other insignia of the United States, a state, a municipality, or a foreign nation. Section 2(c), as we just saw, bars use of a living individual's name (or a deceased U.S. president's name during the life of the surviving spouse) without consent. Separately--and this is a subject-matter point rather than a Section 2 policy point--marks that are purely ornamental (a slogan splashed across the front of a t-shirt as decoration rather than as a brand on the tag or label) or merely informational are refused because they do not function as marks at all. This "failure to function" refusal, which the USPTO has emphasized heavily in recent years, is conceptually upstream of everything else: the designation never crossed the threshold of being a source identifier, so we never even reach distinctiveness. Specimens that show the designation used only as a trade name, a domain address, or background decoration--rather than as a source identifier on the goods--draw the same objection.

The throughline across all of Section 2 is worth stating plainly: some bars (2(d), much of 2(e), failure-to-function) enforce trademark's internal logic--no confusion, enough distinctiveness, genuine source use. Others (the surviving parts of 2(a), 2(b), 2(c)) reflect external policy judgments about deception, government symbols, and the appropriation of identity. After the 2017-2024 trilogy, the Supreme Court has drawn a clear constitutional line through that second set: bars keyed to viewpoint fall, bars that are viewpoint-neutral stand.

Putting It Together: A Subject-Matter Decision Walkthrough

Let us run one more hypothetical end to end, because seeing the doctrines work in sequence is the best way to lock them in. The following is a hypothetical.

Globex Industries invents a sleek new electric kettle and wants the broadest possible brand protection. It comes to counsel with five things it would like to "trademark": (1) the kettle's name, "Lumina"; (2) the kettle's distinctive teardrop silhouette; (3) the warm amber glow of its water-level window; (4) the chime it plays when the water is ready; and (5) the slogan "Boil Brilliantly."

Counsel walks the toolbox and the limits:

  • "Lumina" is a coined word for kettles--arbitrary at worst, fanciful at best (it gestures faintly at "luminous" but is not a real word). Inherently distinctive, strong, and registrable once used in commerce. The only real homework is to clear 2(d) by searching the register and the marketplace first, before a dollar goes into packaging.
  • The teardrop silhouette is product-configuration trade dress. Under Wal-Mart v. Samara, it can never be inherently distinctive, so Globex must build and prove secondary meaning before it can claim the shape. And counsel must scrutinize functionality under TrafFix and the Morton-Norwich factors: if the teardrop shape makes the kettle pour better, hold more, or cost less to manufacture, it is functional and unprotectable--and a design patent, filed early before any public disclosure, would be the right tool instead. Counsel should also check whether Globex has been advertising the shape as a pouring improvement, which would be self-inflicted functionality evidence.
  • The amber glow is a color/feature claim. Per Qualitex, a single color can be a mark, but only with secondary meaning and only if non-functional. The hard question: does amber light help the user judge the water level or temperature (a functional benefit)? If so, it is barred. If it is purely decorative branding, it can mature into a mark over time with "look-for" advertising teaching consumers to treat the glow as Globex's signature.
  • The ready chime is a sound mark--registrable if distinctive and non-functional. Under General Electric Broadcasting, a short, arbitrary, "unique" melody can even be inherently distinctive; a commonplace beep would need secondary meaning. But counsel must heed Vertex Group: if the chime is the functional signal that the water is ready (as opposed to a stylized flourish layered on top of a basic ready-indicator), functionality may bar it.
  • "Boil Brilliantly" is a slogan. It is at least suggestive and arguably laudatory/descriptive ("brilliantly" praises the product). Counsel must make sure it functions as a brand and is not merely an informational or ornamental phrase, and may need secondary meaning if an examiner reads it as descriptive or laudatory.

Notice how every answer routes back to one question--does this identify Globex as the source?--filtered through three gates: the right kind of subject matter, enough distinctiveness, and no functionality, all subject to the Section 2 bars. That is the entire architecture of trademark subject matter in a single case. Globex would also be smart to coordinate this trademark strategy with design-patent and (for any creative artwork, manuals, or marketing copy) copyright filings, layering protections the way we describe in copyright vs. trademark--what is the difference and, for app and product brands specifically, brand protection online--a strategic guide for businesses. And as virtual goods and digital storefronts proliferate, the same subject-matter questions are now being asked about marks in the metaverse, a frontier we cover in trademark challenges in the metaverse and virtual goods.

Key Takeaways

  • A trademark is a source identifier, not a reward for creativity. Anything "capable of carrying meaning"--word, logo, slogan, shape, color, sound, scent, motion--can be a mark if and only if it tells consumers who made the goods. (Qualitex, 514 U.S. 159.)
  • The distinctiveness spectrum decides protectability. Fanciful, arbitrary, and suggestive marks are inherently distinctive and protected on use; descriptive marks need secondary meaning; generic terms are never protectable, ever. (Abercrombie, 537 F.2d 4.)
  • Distinctiveness is relative to the goods. "Apple" is fanciful for phones and generic for fruit. The mark is always the designation plus what it labels.
  • Generic terms are forbidden, but composite "generic.com" marks may survive if consumers perceive the whole as a source rather than a class. Genericness turns on consumer perception, decided on evidence. (Booking.com, 591 U.S. 549.) And success can be fatal: genericide has killed "aspirin," "escalator," and "thermos." (Elliott v. Google, 860 F.3d 1151.)
  • Functional features cannot be trademarks, no matter how distinctive, because trademark's perpetual term would otherwise swallow the patent bargain. Watch utilitarian functionality (Inwood), expired-patent and Morton-Norwich evidence (TrafFix), and aesthetic functionality (the John Deere green problem). (15 U.S.C. § 1052(e)(5).)
  • Non-traditional marks face heavier burdens. Color, scent, taste, and many sounds can never be inherently distinctive and must prove secondary meaning--except that a distinctive multi-color packaging design can be inherently distinctive. (In re Forney, 955 F.3d 940; In re Clarke; In re General Electric Broadcasting.)
  • Certification and collective marks follow special rules. A certification mark is used by people other than its owner to certify a standard; the owner cannot sell the certified goods and cannot discriminate among qualifying applicants.
  • Section 2 of the Lanham Act adds statutory bars--confusion (2(d)), descriptiveness/surname/geographic/functionality (2(e)), and the policy bars in 2(a)/(b)/(c). After Tam (582 U.S. 218) and Brunetti (588 U.S. 388), viewpoint-based bars (disparagement, scandalousness) are unconstitutional; after Elster (602 U.S. 286), viewpoint-neutral bars like the names clause stand.
  • Registration strengthens rights but does not create them--U.S. trademark rights flow from use in commerce.

Frequently Asked Questions

Can I really trademark a color, a smell, or a sound? Yes, in the right circumstances. The Supreme Court confirmed color can be a mark in Qualitex, and the USPTO has registered sound marks (the NBC chimes), scent marks (plumeria-scented thread; the smell of Play-Doh), and motion marks. The catch is twofold. First, these "non-traditional" marks are almost never inherently distinctive--a single color, a scent, a taste, and many sounds can never be inherently distinctive, so you must prove consumers see them as a brand (secondary meaning). (A distinctive multi-color design on packaging is the rare exception, per In re Forney.) Second, they are barred if they are functional. The more unusual the mark, the heavier your evidentiary burden on both fronts.

What is the difference between a generic and a descriptive term? A generic term is the common name of the product class ("Bicycle" for bicycles) and can never be a trademark. A descriptive term names a feature or quality ("Fast" for delivery) and is not protectable by itself, but it can become a trademark if it acquires secondary meaning--that is, if the public comes to associate it primarily with one source. The practical line between descriptive and the next category up, suggestive (which is protectable immediately), is one of the most litigated questions in trademark law, decided with the "imagination test."

What is a certification mark, and how is it different from a regular trademark? A certification mark certifies a characteristic of goods or services--regional origin, quality, safety, ingredients, or union labor--and is used by people other than its owner. UL, ENERGY STAR, and the kosher symbols are classic examples. Two oddities define it: the owner sets a standard rather than selling the goods (and loses the mark if it starts producing them), and the owner must certify anyone who meets the standard rather than playing favorites. A regular trademark, by contrast, says "made by this one company."

Did the Supreme Court make it harder or easier to register edgy or political marks? Both, depending on why the mark is refused. Tam (2017) and Brunetti (2019) struck down bars that refused marks because the government found the message disparaging, immoral, or scandalous--that was unconstitutional viewpoint discrimination, so offensive or provocative marks can now be registered. But Elster (2024) upheld the bar on registering a living person's name without consent, because that rule is viewpoint-neutral. The dividing line is whether the refusal targets a viewpoint. And remember: all of these cases were about registration, not about the right to use the words.

If a feature of my product is both useful and distinctive, can I protect it? Not with trademark, if it is functional. Under TrafFix and Inwood, a feature that is essential to the product's use or purpose, or that affects cost or quality, is functional and cannot be a trademark even if consumers strongly associate it with you. The same goes for aesthetically functional features whose exclusive use would disadvantage competitors. Useful features belong to the patent system--a utility patent for how it works, a design patent for how it ornamentally looks--each lasting a limited term before passing into the public domain.

Is the name of my LLC or my domain name automatically a trademark? No. Forming a business under a name, or registering a domain with a registrar, does not by itself create trademark rights. Those are administrative registrations, not source-identifying use on goods or services. (A trade name and a trademark can be the same word, but only the brand use on the goods creates rights.) You acquire trademark rights only by using the designation as a brand in commerce--and even then, you should clear it against existing marks and consider federal registration. See the trademark process and trademark rights under common law.

Where does this fit in the four-part series? This is Part 1, on what can be a mark. Part 2, substantive standards for protection, digs into distinctiveness, secondary meaning, and the use requirement. Part 3, obtaining protection and licensing, covers registration and licensing. Part 4, infringement and related rights, covers enforcement, dilution, and defenses.

Related Articles

This article provides general information and is not legal advice. Trademark law is fact-specific and changes over time; consult qualified counsel about your particular situation.