Lemley & McKenna: Trademark Spaces

Where a trademark is located matters. Some physical spaces on products are more likely to contain trademarks. Some obvious examples are the left breast of a t-shirt, the side or tongue of a sneaker, or the label on the front of a food container.  Consumers are more likely to find trademarks located in those spaces, and they are probably more likely to perceive features as trademarks if they are in those locations.

In their fascinating new article, Trademark Spaces and Trademark Law’s Secret Step Zero, Mark Lemley and Mark McKenna draw out the relevance of physical space for trademark protection, and in particular for non-verbal marks like logos and trade dress.  The authors argue that where a trademark is located determines how the law currently treats it, but that this is currently done by judges and trademark examiners as a "secret step zero." They think this is not ideal. Consideration of trademark spaces should instead be both more explicit and better thought-out. 

The article is now published in the Stanford Law Review (and posted here on SSRN).

Prior scholarship on "trademark failure to function"

Trademark Spaces builds directly on Alexandra Roberts' recent article on whether a mark functions as a trademark. (Laura Heymann previously blogged on Roberts' work on JOTWELL). In Trademark Failure to Function, Roberts compellingly showed that trademark law contains the legal requirement that a mark be used as, and function as, a trademark, and that this requirement is grounded in the statute. See 15 U.S.C.§ 1127.  Use-as-a-mark generally means the putative mark must be perceived by consumers as a source-identifier, instead of as something else, like merely ornamental or informational content.  

This use-as-a-trademark requirement is separate from the requirement of inherent or acquired "distinctiveness." Distinctiveness asks: How unique or unusual is the word, logo, or device in its own right? Does it stand out enough conceptually to consumers? Use-as-a-mark asks: Does it look like a source-identifier or not?   In practice, this means the putative trademark "must appear where consumers expect a trademark to appear, and it must be sufficiently set off from the surrounding text and images to attract notice." (Roberts, 105). Otherwise, it does not function as a trademark to consumers and is invalid, no matter how conceptually distinctive it might be on the Abercrombie spectrum. 

Use-as-a-mark has been like (in Roberts' words) the "stepchild" of trademark law, receiving very little attention compared to distinctiveness. Many trademark case books don't cover it. Students are unlikely to learn about it in IP survey.  However, Roberts showed that in practice the trademark office regularly rejects trademark applications as merely "ornamental" if the specimen of use submitted by the applicant does not show the mark is being used in a way that consumers will recognize as a trademark. (Roberts, 5). The office states in the Trademark Manual of Examining Procedure (TMEP) § 1202.03, that 

"[s]ubject matter that is merely a decorative feature does not identify and distinguish the applicant’s goods and, thus, does not function as a trademark. A decorative feature may include words, designs, slogans, or trade dress. This matter should be refused registration because it is merely ornamentation and, therefore, does not function as a trademark, as required by §§1, 2, and 45 of the Trademark Act, 15 U.S.C. §§1051, 1052, and 1127."

The office provides some examples of proper use (e.g. a tag on clothing) and improper use, such as words or phrases "used on items such as t-shirts and sweatshirts, jewelry, and ceramic plates ... that purchasers will perceive as conveying a message rather than indicating the source of the goods." TMEP. § 1202.03(f)(i).  

Roberts observed that, despite how often examiners issue failure-to-function-style these refusals, the office has not given examiners concrete guidance on what constitutes ornamental use, as compared to other refusals like distinctiveness.  (122-123).  Instead, a lot of the analysis is apparently performed without much transparency and based on decision-makers' gut instincts.

The broader relevance of trademark spaces 

Building on Roberts' failure-to-function scholarship, Lemley and McKenna's article argues that whether a trademark is legally protectable depends in part on where it is located in physical space, and on whether it is located in what the authors call a "trademark space." 

Whether a mark is in a trademark space can obviously help illuminate whether a mark fails to function, in the way Roberts described. Physical location is not the only factor the examiners assess in deciding whether to reject as merely "ornamental," but it is a big one. So trademark spaces is highly relevant to the question of whether a putative trademark fails to function as a trademark.

However, in contrast to Roberts' article, which focused on word marks, Lemley and McKenna's article focuses on how trademark spaces inform trademark protection for non-verbal marks, such as logos and trade dress.  Their article does deal with the question of whether a non-verbal mark functions as a trademark. However, one of the main take-homes I got from their article is that, especially for non-verbal marks, neither examiners nor courts have properly separated out the questions of use-as-a-mark and distinctiveness. The issues are often jumbled together, and a lot of the real analytical work is happening (as the authors say) in "secret."  

Ignoring Wal-Mart

One of the main doctrinal issues where Lemley and McKenna see trademark spaces playing a role is in how courts apply, or don't apply, the Supreme Court's analytical framework for trade dress.

Under the Supreme Court's holding in Wal-Mart, product packaging trade dress, as well as other non-verbal marks like logos, can be inherently distinctive. In contrast, product design trade dress, such as the distinctive "look" of children's clothing, is harder to protect because, like a single color, product design trade dress always requires secondary meaning.  

Identifying product designs (and single colors for that matter) has proven very difficult. The closest thing to a definition that the Court in Wal-Mart gave is that "almost invariably, even the most unusual of product designs—such as a cocktail shaker shaped like a penguin—is intended not to identify the source, but to render the product itself more useful or more appealing."  Even Justice Scalia conceded in Wal-Mart  that this is not a very helpful standard. (24-25). 

As a descriptive matter, Lemley and McKenna argue that examiners and courts are not really applying Wal-Mart at all, or at least they aren't just doing that. (31, 42).  Instead, they are (implicitly) asking whether a non-verbal feature looks enough like a trademark to be protected. In effect, they are asking whether it functions as a trademark in the way Roberts discussed in her work. Moreover, in asking whether the feature functions as a mark, examiners and courts are looking at where the feature is located. If the feature is in a trademark space (e.g. left breast of shirt), it is more likely to be classified as a logo or product packaging. If it is not located in a discrete trademark space, it is more likely to be treated as product design trade dress.  

One of the clearest examples Lemley and McKenna give of this phenomenon is the fact that decision makers have treated the three-stripes feature on the side of Adidas' shoes as a logo that's capable of inherent distinctiveness because the three-stripes are in classic trademark space, "without ever acknowledging" the possibility that Adidas’s three-stripes might be product design. (5, note 7).  In other words, they look at the three-stripes and they say "that's a trademark, because it's in a trademark space," and then they give it more favorable treatment, classifying it as a logo.  This in turn allows Adidas to rely on inherent distinctiveness rather than having to show secondary meaning, as would be the case if this were product design trade dress.

Obviously, the Adidas three-stripes mark would survive anyway, because it has secondary meaning now. But the point is that this sort-of secret "step zero" analysis is occurring, and is affecting other cases going forward.  For example, other places on shoes, like the toe, aren't seen as a normal place for a trademark to be, and so features in those places are assessed as product design trade dress and denied protection without a showing of secondary meaning. Compare LVL XIII Brands, Inc. v. Louis Vuitton Malletier S.A. (2d Cir. 2017) (holding "rectangular metal plate affixed to the toe of LVL's sneakers" is a product design feature only protectable with secondary meaning).

Courts and examiners are also likely assessing trademark spaces to some degree when they evaluate the inherent distinctiveness of logos and product packaging trade dress. For these types of marks, courts typically use the Seabrook Foods test in lieu of the Abercrombie spectrum. Seabrook Foods doesn't compare the meaning of a term to the product or service at issue; instead, it looks at factors like how unusual the logo, imagery, or packaging is in the industry, and whether it "is capable of creating a commercial impression distinct from the accompanying words."  (12-13, 68-69).  One of the easiest ways to create a distinct commercial impression is to place a feature in a discrete physical space.

    

The case for considering trademark spaces more explicitly

The authors' normative claim is that this "secret step zero" should be made explicit and done with more precision. They recommend the following framework.  If a mark is located in a trademark space, it should benefit from a rebuttable presumption that it functions as a trademark. Distinctiveness, either inherent or through secondary meaning, still must be met separately. But generally speaking, this would mean that marks placed in trademark spaces will be easier to protect than marks that are not placed in trademark spaces. (7-8, 65-67).

There are several potential payoffs from adopting this approach. One benefit is simply helping trademark examiners and courts conduct difficult inquires for which they currently lack guidance -- like deciding whether a feature is perceived by consumers as a trademark, or deciding whether to classify a feature as product design trade dress. Explicitly considering whether the feature is located in a trademark space could in theory give decision-makers a helpful tool for classifying non-verbal features and for "determining whether a claimed design functions as a mark[.]" (7-8, 64-74).

Another payoff is that limiting trademarks to certain spaces can help constrain what can serve as a trademark in the first place and, to quote the authors, help reign in "the problem of trademark creep." (72-73).  If trademark spaces becomes an accepted part of establishing a protectable trademark, then some spaces can be trademark spaces; but other spaces can remain free from trademark protection. "Recognizing certain trademark spaces may also open the door to acknowledging that ... just as there are trademark spaces on products and packaging, there are also non-trademark spaces." (73).  

Some tricky issues

I found the descriptive aspect of this article extremely insightful. I was convinced by the claim that decision makers often take physical space into account when determining what can be protected as a trademark. I also love the normative proposal to use trademark spaces more explicitly to keep trademarks in their proper place. I think this could make trademarks a more valuable informational tool for consumers, while also helping control inappropriate uses of trademarks, and especially trade dress, protection. 

I will point out below a few difficult issues that decision makers will likely face in applying this framework, and how the authors seek to address them. I'll also consider possible interplay between trademark spaces and the functionality bar.

1. How do we know what counts as a trademark space?

The first and most obvious difficulty in applying the trademark spaces framework is how to decide what counts as a trademark space. Putting aside the obvious examples like the left breast of a t-shirt this is not an easy task. Lemley and McKenna don't give a mathematical definition or exhaustive list. 

In fact, one of their main points is that what count as a trademark spaces for consumers today is not inevitable or dictated only by efficiency. Trademark spaces evolve over time depending on where companies choose to place their trademarks on products, and how successful those products are. For example, the success of companies' like Nike and Adidas in selling shoes with a Swoosh and three-stripes, respectively, on the side or tongue of their shoes turned this into a trademark space. And now consumers perceive as such. (32-35).

While recognizing the innate difficulty in predicting or predetermining what happens in the marketplace, the authors suggest that what counts as a trademark space should not be based on the gut instinct of a single examiner or judge, and definitely should not be solely up to companies to make those decisions through well-resourced branding campaigns that can allow companies to effectively "buy up" new and unprecedented trademark spaces. (66).

Going forward, one very practical recommendation the authors make is that the trademark office give more guidance "about what constitutes a trademark space on various forms of consumer goods." (71). To some degree, the office has already done so in the course of addressing so-called ornamental use in TMEP. § 1202.03, quoted above. On its website, and in a mildly entertaining broadcast, the office helpfully gives some specific examples of non-ornamental uses, which I think we can call trademark spaces:

• "[d]iscrete wording or design on the pocket or breast portion of a shirt", 
• a "tag on the inside of a hat or garment",
•  a "[l]ogo on a tag above the back pocket of a pair of jeans", or
•  "[a] small logo stamped on the back of a dinner plate or bottom of a coffee mug."  

The office even provides a few examples of improper uses of non-verbal marks and trade dress. See, e.g., TMEP § 1202.03(f)-(g), showing uses that failed inherent distinctiveness under the Seabrook Foods, or appeared to be "ornamental," "decorative," or otherwise lacking in "trademark significance."  

But the office could do a lot more work to identify locations that examiners and more importantly consumers see (or do not see) as trademark spaces.  The office could keep up to date on where marks are being used and on what examiners and courts have said about the trademark-appropriateness of certain physical spaces. Surveys could be performed or collected from cases and academic work. All sorts of evidence could be considered. (71-72). Again, the payoff from doing this could be two-fold: help examiners and courts make difficult decisions more transparently; and constrain the places that are subject to trademark protection, keeping trademarks in their proper place.

 

2. Is the feature separable from, or part of, a product's design?

A second extremely tricky issue in applying trademark spaces is that it can be hard to decide whether a feature is in fact in a trademark space, as opposed to part of a product's overall design. It's easy to say a logo on the left breast of a shirt is in a trademark space and call it a logo. But what if a feature is sort-of in a trademark space, and also sort-of integrated into the design as a whole? 

For example, on many sneaker models, Adidas' three-stripes mark is very big. It's not confined to the usual trademark space on the side of the shoe. Is this a logo that is being used in a trademark space and that clearly functions as a trademark? Or is it in fact product design trade dress and also not clearly functioning as a trademark?

To solve this issue, Lemley and McKenna develop a concept of "separability."  They suggest that it is possible to separate out a feature located in a trademark space from the rest of the product's design, but that this is not always the case.  For example in In Re Slokevage, the putative trademark was a cut-away flap configuration on the back pocket of jeans, which is a place where companies like Levis have long placed their trademarks. But the examiners and the Federal Circuit held this was product design under Wal-Mart. The authors' separability concept could explain this case by saying that the cut-away flap feature was product design because it was not sufficiently separable from the design as a whole.

It strikes me that this separability concept is going to be hard to apply. Copyright law's useful article doctrine has a statutory concept of separability, 17 U.S.C. § 101 (protecting pictorial, graphic, or sculptural features "that can be identified separately from, and are capable of existing independently of, the utilitarian aspects of [a useful] article."), and that has been notoriously hard to apply in practice. The authors notably do not draw on copyright's separability concept, presumably because it's such a mess. Instead, they find support in the Seabrook Foods test, which many courts use for assessing the inherent distinctiveness of non-verbal mark such as logos and product packaging in lieu of Abercrombie.  This test expressly considers, among other factors, whether the feature is "capable of creating a commercial impression distinct from the accompanying words." This analysis can be re-jiggered to take into account trademark spaces.  We can ask, is the feature "capable of creating a commercial impression" apart from the rest of the product's design, or is it in fact part of the design? 

To try to understand how this works, I asked the authors how it would apply to the famous Burberry “check mark” pattern, which appears on scarves, sunglasses, and clothing. Their answer, as I understand it, was that sometimes this can get trademark space treatment, but only if the pattern is used in a trademark space and in a way that creates a separate commercial impression from the product as a whole. As you can see below (from a complaint Burberry brought against Target) the pattern often takes up the entire article like a scarf. There's nothing in a trademark space that is separable from the scarf as a whole. So this is product design, even if the check mark touches on some trademark spaces. (One could even make the argument that it's akin to a single color).

 

In contrast, Burberry sometimes uses that pattern in discrete physical spaces, like the lapel of a jacket. Assuming the lapel of a jacket is a trademark space, then this use of the pattern would be, I think, conceptually separable from the design as a whole and count as use in a trademark space. Thus, there would be a rebuttable presumption that it functions as a trademark. (Distinctiveness under Seabrook would still have to be shown or a showing of secondary meaning.)  

3. Relevance for functionality?

One issue I'm still working through is how, or whether, trademark spaces can help courts and examiners apply the functionality bar.  This is important because functionality is the main backstop for features that achieve secondary meaning and can otherwise be protected. Functionality cabins protection for distinctive features that become beloved by consumers, and yet also serve non-trademark functions, such as making something useful or making something pretty. 

For example, the Burberry check mark pattern, shown above, has a lot of secondary meaning among fashionistas. Functionality (of the aesthetic variety) is an important limitation on how far Burberry can protect its pattern.  

Lemley and McKenna have each written at length about functionality in prior articles, so while they allude to a potential interaction between trademark spaces and functionality, they don’t address it in depth. (69, 77). It seems to me there is interplay between trademark spaces and functionality, because trademark spaces should, by definition, be non-functional. Trademark spaces should be sufficiently out of the way and not interfere with the product's non-trademark utility, aesthetically or mechanically-speaking.

Utilitarian functionality

One could argue that trademark spaces are necessarily defined in contrast to what has a mechanical function. In TrafFix, Supreme Court held that a product feature is "functional and cannot serve as a trademark, if it is essential to the use or purpose of the article or if it affects the cost or quality of the article." The Court suggested, though, the possibility of a carve-out for "arbitrary, incidental, or ornamental aspects of features of a product," "such as arbitrary curves in the legs or an ornamental pattern painted on the springs." (The Court said these features don't get a presumption of functionality, even if they are claimed in a utility patent). 

This suggests a division between functional features, such as a dual spring design that holds a road sign up in the wind, and non-functional "arbitrary, incidental, or ornamental features," which are or can become trademark spaces. In other words, utility can dictate what can be a trademark space.  Like a peacock's tail, trademark spaces have to be non-essential to how the product functions and cannot affect cost or quality.  They need to stay out of the way.  One example, suggested by a few courts, are the specific features making up the "look" of a Hummer vehicle that are not shared by other military vehicles.  

Aesthetic functionality

The aesthetic functionality bar is trickier. This holds that features cannot be protected if doing so would put competitors at a "significant non-reputation related disadvantage." This doctrine allows courts to narrow the scope of trademark protection, giving the owner only a sliver of protection for source-identifying features, while allowing others to use even-potentially-confusingly-similar features for non-source identifying reasons -- namely, because they are pretty.   

One possible interpretation of aesthetic functionality could be that it constrains trademark protection to certain physical trademark spaces. For example, a famous aesthetic functionality case involves a floral pattern on dish ware.  Trademark spaces would say, don't try to make the floral pattern on the front of the plate your trademark. That isn't a trademark space. Put a logo on the back of the plate instead. If you must, use the floral pattern in a discrete physical space. Likewise, for the Burberry check mark example, trademark spaces would say, you can't claim the check mark pattern as your trademark for all possible uses. It's only protected to the extent it is used in specific spaces, like on the lapel of a jacket or on the top of the fingers of gloves. Likewise, if you want the red outsole of a high-heel to be your trademark, do so in way that carves out a discrete physical space.  To quote the Second Circuit, "limit the trademark to uses in which the red outsole contrasts with the remainder of the shoe ..."   I don't claim to know whether that is what courts are doing in aesthetic functionality cases. But it does seem like one approach to aesthetic functionality could be to constrain trademark protection to certain spaces.