by Noelle Higginson*
Trademark law allows producers of commercial goods to identify and differentiate their goods in the market by a specific mark. In this way, trademark law protects producers and consumers alike: A recognizable mark, exclusively used by a single producer, allows consumers to make informed choices about what goods to purchase and allows producers to benefit from the fruits of their labor—when they make good products, consumers associate their mark with that good reputation. Trademark infringers—those who use a confusingly similar mark to the trademark holder’s mark—interfere with that goal. However, certain products like works of art or parody that intentionally refer to another entity often must use the other’s trademark (or something similar) to achieve their purpose. These alleged infringers therefore risk trademark liability for the sake of artistic expression. The difficulty for courts is determining when the use of a mark in art or parody might defeat an infringement claim or vice versa. That determination is the province of the Rogers framework as recently considered by the Supreme Court in Jack Daniel’s Properties, Inc. v. VIP Products, LLC.1 This Contribution (i) assesses the Rogers v. Grimaldi2 framework’s application in two recent cases in lower federal courts and (ii) considers how courts can factor in First Amendment principles even when, following the VIP Products holding, Rogers does not apply.
Trademark law, a subset of unfair competition generally, has two principal goals: protect consumers from confusion and protect trademark owners’ reputation and goodwill, identified by a mark and earned by their investment of time and energy.3 Trademark achieves this by conferring a property right on the trademark holder who may use their mark4 to identify their product(s) and enforce their right against other would-be users, thereby preventing them from identifying their products with that mark in such a way as might create a likelihood of confusion among consumers as to who is the source of the product.5
But these unfair competition policies are not without their own competition. Trademark protection may conflict with First Amendment protections for expressive speech when, for example, trademark infringement claims are brought against parodies or artistic works.6 The public has an interest in not being misled, to be sure, but the public has a constitutionally protected interest in free expression as well.7 This Contribution is concerned with how that tension—between trademark property rights and First Amendment expression rights—currently stands with the understanding that the narrowness or capaciousness of a First Amendment defense to trademark infringement claims will have carry-on effects in the marketplace, in particular when products cannot easily be separated along a commercial-expressive binary.
To set the stage, a run-of-the-mill trademark infringement claim brought under the federal Lanham Act requires the plaintiff (the trademark owner) to establish that they own a valid and protectable mark and that the defendant’s use of that mark—or some imitation of it—creates a likelihood of confusion.8 Circuits have adopted similar but not identical tests for assessing what uses may create a likelihood of confusion.9 The Second Circuit for example, applies a multifactor balancing test to determine “likelihood of confusion” that considers
(1) the strength of the trademark; (2) the degree of similarity between the two marks; (3) the proximity of the products and their competitiveness with one another; (4) the likelihood the prior owner may ‘bridge the gap’ in the markets for the products; (5) evidence of actual consumer confusion; (6) the defendant’s good faith in adopting its imitative mark; (7) the quality of defendant’s product compared with the plaintiff’s product; and (8) the sophistication of the buyers.10
However, when a defendant’s imitative use is “a parodic or artistic expression” of the plaintiff’s mark, many circuits will consider a First Amendment protection claim to counter the infringement claim.11 This asserted protection is assessed under the Rogers v. Grimaldi12 test: When parodic or artistic expressive uses are at issue, a court may find that the First Amendment protects the imitative use (i.e., that the Lanham Act’s protection of the trademark owner’s right does not apply) when the use is “artistically relevant” and is not “‘explicitly misleading’ as to the source or content of the work.”13 This test is seen as more defendant-friendly, given that it protects imitative uses so long as they are minimally artistically relevant and do not explicitly mislead, even if they may confuse.14
With those basics in mind, this Contribution first considers the problem the Rogers framework rose to address. Second, it summarizes how the Supreme Court’s recent decision in VIP Products affects application of the Rogers framework. Finally, it will assess the outcomes of two post-VIP Products cases. How lower federal courts resolve those and similar cases will determine the actual impact of the VIP Products holding in light of what the Court left open in VIP Products: what is source-identifying use and what standard to apply to parodies and expressive works in the absence of Rogers. Not all courts default back to standard likelihood of confusion inquiries, and the expressive content at stake—even if not enough to justify application of Rogers—might justify a modified standard.
*****
The tension of weighing trademark protection against First Amendment rights became the core concern of the Rogers framework.15 There, the Second Circuit considered a trademark infringement claim arising from a film, entitled Ginger and Fred, brought by Ginger Rogers (who frequently and famously performed alongside co-star Fred Astaire, hence Ginger and Fred).16 Rogers objected to the title of the film, arguing it implied that “she sponsored, endorsed, or was otherwise involved in the film.”17 The filmmakers and producers argued, and the District Court concluded, that the title of the film was expression “entitled to the full scope of protection under the First Amendment.”18 The Second Circuit reasoned that, in the name of constitutional avoidance, a two-part test would apply to determine the application of the Lanham Act—and its traditional “likelihood of confusion” inquiry for trademark infringement—when a claim like Rogers’s might conflict with “the public’s interest in free expression.”19 Thus, the Rogers framework was born and required dismissal of infringement claims brought against artistic works unless the alleged infringing use either has “no artistic relevance to the underlying work” or “explicitly misleads as to the source or the content of the work.”20
Over the course of the last thirty-five years, most of the federal courts of appeals have adopted Rogers.21 When applied, the test generally precedes the traditional likelihood of confusion analysis and requires dismissal of a trademark infringement claim when the plaintiff cannot show either of the two factors above (artistic relevance or explicitly misleading). But the Supreme Court’s recent holding in Jack Daniel’s Properties v. VIP Products ensures that the Rogers framework will remain “a cabined doctrine.”22
In VIP Products, VIP Products made a dog toy named “Bad Spaniel’s” that mimicked many of the elements of Jack Daniel’s trademarked whiskey bottles, but the company asserted that its object was to parody the famous Jack Daniel’s bottle, therefore Rogers applied given the expressive nature of its parodic products.23 The Court disagreed. It reasoned that historically Rogers only applied—if ever—when a mark was used to “perform some . . . expressive function,”24 not “to designate source.”25 Confirming that limitation, the Court in VIP Products ensured that Rogers would not apply—if it will or should apply at all26—in circumstances in which use of another’s mark is in a “source-identifying way.”27 Because VIP Products conceded in the course of litigation that it used the marks in a source-identifying manner it could not avail itself of protection under Rogers.28
But after VIP Products the Court left open two questions: (1) what qualifies as a “source-identifying” use and (2) when a use is source-identifying, how is a court to apply the likelihood of confusion test under the Lanham Act—with or without an eye to protecting expression?
*****
Identifying use of a mark as “source-identifying” for purposes of applying Rogers, is not so easy. Indeed, the Second Circuit in Rogers itself acknowledged that titles for artistic works can take on double meaning: “Titles, like the artistic works they identify, are of a hybrid nature, combining artistic expression and commercial promotion.”29 The Supreme Court did not dispute that trademarks, while source-identifying, “are often expressive, in any number of ways.”30 So a lingering question for practitioners and courts alike is how to categorize the use of a mark. And that uncertainty leads to a second lingering question: What inquiry, then, is a federal court to engage in when a title or use of a mark is both source-identifying and expressive?
Consider two recent examples—the Second Circuit’s decision in Vans, Inc. v. MSCHF Product Studio, Inc.31 and the District Court for the Southern District of New York’s decision in Hermès International v. Rothschild32—both of which show the need to balance trademark protections against protections for expression when trademark infringement claims are brought against use of a mark that falls into a grey area between commercial and expressive speech, or, more accurately, where those two categories overlap. MSCHF and Hermès provide some post-VIP Products guidance. They appear to confirm (1) use of a mark will often have both expressive and source-identifying elements; (2) when that is true, what matters for determining the unavailability of a First Amendment defense is the presence of source-identifying use; and (3) in light of source-identifying use, the relevant test is the likelihood of confusion inquiry.
First, in Hermès, the fashion brand Hermès—designer and producer of Birkin bags—sued a digital content creator, Rothschild, for creating a non-fungible token (NFT) collection wherein the underlying images “depict[ed] fur-covered Birkin handbags—so-called ‘MetaBirkins.’”33 Rothschild’s MetaBirkins project sold NFTs to consumers, under which contracts buyers then owned the MetaBirkin image associated with the NFT.34 Thus, on one hand Rothschild asserted his use of the Birkin marks and trade dress were artistic expression, protected by the First Amendment; on the other hand, Hermès argued that Rothschild’s use was infringement, likely to confuse consumers as to the creator or source of the MetaBirkins, especially as Hermès submitted evidence that it was contemplating entering the NFT space itself.35
Second, in Vans, the sneaker company Vans sued an art collective, MSCHF, for alleged trademark infringement. MSCHF created a “Wavy Baby” shoe (“Wavy Babys”) that resembled Vans’ “Old Skool” shoe.36 MSCHF claimed its Wavy Babys were works of art, commentary on consumerism and “sneakerhead culture,” and meant for collection, not use.37 Yet, the Second Circuit upheld the district court’s refusal to apply Rogers. It pointed to the visual similarities between MSCHF’s design and Vans’ trademarks and trade dress, and it considered MSCHF’s concession that the Wavy Baby design process “‘start[ed]’ with Vans’ marks” as evidence of intent to “benefit from the ‘good will’ that Vans . . . had generated over a decades-long period.”38 In the eyes of the courts, MSCHF’s use was “source-identifying” use.
In both cases, the “source-identifying” nature of the use of the mark determination was relatively one dimensional.39 As indicated above, the Second Circuit relied heavily on a perceived intent to “benefit from [Vans’] good will . . . generated over a decades-long period” to determine MSCHF’s use of the mark was source-identifying.40 Meanwhile, in Hermès the court refused to disturb the jury’s conclusion that First Amendment protection did not apply to Rotschild’s use of “MetaBirkins” in part because the use of “Birkins” in the title alone suggested that the use was source-identifying and therefore, post-VIP Products, not entitled to a more “stringent threshold test.”41 That is, even when the true source might be readily apparent from other aspects of the product and marketing, since the uses of the mark suggested an intent to appropriate goodwill and did not sufficiently differentiate the real producer from the trademark owner (in MSCHF’s case) or since the use was in the title (in Rothschild’s case), the courts found the use of the marks to be source-identifying.42
Ultimately, in both cases, the alleged infringer ultimately lost: in Hermès because both the jury and the judge were unconvinced by Rothschild’s First Amendment argument (and went on to find likelihood of confusion)43 and in MSCHF because the court did not deem the Rogers applicable at all, “notwithstanding the Wavy Baby’s expressive content.”44
*****
What standard for determining infringement should be applied in circumstances like MSCHF and Hermès—i.e., when expressive and source-identifying use is at issue—going forward? VIP Products dictates that at step one, once a court decides that use of a mark is source-identifying, it cannot apply Rogers. Both MSCHF and Hermès conform to that principle. But at step two, must a court then default back to its standard “likelihood of confusion” analysis? The Court in VIP Products arguably left open the possibility that the answer to this second question—what “likelihood of confusion” analysis to undertake—may well be adjusted when an expressive work is at issue: “[A]lthough VIP’s efforts to ridicule Jack Daniel’s does not justify use of the Rogers test, it may make a difference in the standard trademark analysis . . . . [W]e remand that issue to the courts below.”45 Turning to that issue now, lower federal courts may take different approaches to tackling “the standard trademark analysis” when expressive works are at issue.
Tellingly, the Second Circuit in MSCHF affirmed the district court’s likelihood of confusion analysis without modification. Indeed, the lower court there only considered the parody argument after having determined a likelihood of confusion.46 The Hermès case had no occasion to consider it, given that the defense failed under Rogers itself. Some courts have, indeed, concluded that the likelihood of confusion factors alone provide ample protection for expression. However, the Supreme Court’s decision in VIP Products seems to leave open whether a situation like MSCHF’s Wavy Babys should be subjected to the Polaroid factors independent of some consideration for MSCHF’s parodic use argument. What’s more, the Court hinted at how parodies should be analyzed, giving a Fourth Circuit decision, Louis Vuitton Malletier v. Haute Diggity Dog, as an example.47 Indeed, Louis Vuitton and cases in circuits that declined to adopt Rogers from the get-go may prove useful for artists (and their lawyers) seeking to protect their works from trademark infringement claims in a world in which Rogers is a far more distant source of protection, given the proclivity lower courts may have to find source-identifying use, as Hermès and MSCHF may demonstrate.
*****
Again, the question is how are courts to make infringement determinations for expressive works when Rogers does not apply? The Second Circuit condoned a return to the standard Polaroid test for likelihood of confusion in Hermès and MSCHF. But practitioners need not accept such a full pendulum swing back; indeed, other federal courts adjust their likelihood of confusion inquiries to include additional considerations when parodies and expressive works are at issue. One method would be to assess the success of a parody in the first instance, whereupon “[t]he finding of a successful parody only influences the way in which the [likelihood of confusion] factors are applied.”48 This is the approach adopted by the Fourth Circuit in Louis Vuitton, and condoned by the Court in VIP Products.49 A second method would be to consider additional factors to aid in the assessment of likely confusion that recognize the artistic expression at stake, e.g., whether “the junior user ha[d] a genuine artistic motive for using the senior user’s mark” based on timing, the mark’s relation to the underlying work, what expressive content the artist added beyond the use of the mark, among other factors.50
Ultimately, pushing for these “heightened” standards will require an appeal to the reason Rogers was adopted in the first place: That expressive works deserve special consideration in light of First Amendment protections. Take, for instance, MSCHF whose product was intended to be a social commentary—one that, perhaps, Vans did not like—not a commercial product. The Second Circuit’s reasoning suggests trepidation, but ultimately deference, to the district court’s finding that MSCHF’s shoe was not a work of art, but a commercial product, acknowledging that “[a]lthough it is hard to see why some people would wear the Wavy Baby as a functional shoe, we owe that finding deference.”51 In that circumstance, although not going so far as to apply Rogers, a lower court might well adjust its likelihood of confusion analysis—indeed, Polaroid and other likelihood of confusion inquiries are concededly non-exhaustive lists of potentially relevant factors52—to include additional factors protecting expressive uses.53
In sum, VIP Products does not eviscerate special consideration for expressive uses even if it “cabined” the application of Rogers when source-identifying use is established. Although MSCHF did not prevail in the Second Circuit, which affirmed application of the traditional Polaroid factors, artists in similar circumstances might well push for adjusted standards, e.g., the Louis Vuitton approach to parody or the Stouffer approach to artistic expression. Hermès suggests that at extremes, tinkering with standards may not make a difference. Rothschild lost under Rogers and Polaroid. There is no reason to believe instructing a jury to consider some intermediate or different standard would be needed or useful. But MSCHF is a far closer call. Assessing likelihood of confusion where successful parodying is baked into the analysis—or where special consideration is given to expressive content—may well have altered the outcome there.
Taken together, Hermès and MSCHF indicate that defendants will face a tougher time seeking First Amendment protection for works that use another’s trademark in a way that seems to identify that other as a source even when they have, as MSCHF had, a perfectly legitimate claim to expressive content. Source-identifying use will preclude application of Rogers and may be found quite readily as Hermès and MSCHF suggest. But defendants can avail themselves of the open questions in VIP Products to argue for heightened protection nonetheless: Lower courts are already flexible in applying likelihood of confusion analyses—considering (1) the success of a parody in the first instance before weighing the remaining factors or (2) more granularly the parodist’s intent—and may be open to adding more factors relevant when expressive content is at issue. These approaches might protect artistic expression like MSCHF’s social commentary, even if they will not protect mere “swindl[ing],” like Rothschild’s opportunism in Hermès.
* Noelle Higginson is a J.D. Candidate (2024) at New York University School of Law. This Contribution originated in the problem at the 2023 Cardozo BMI Entertainment Moot Court Competition, hosted by Cardozo Law School. The question presented was whether use of the word “NUESTRO” in an NFT digital art collection infringed upon a design company’s registered “NUESTRA” trademark. This Contribution considers the continued viability and application of a First Amendment defense in light of the Supreme Court’s recent decision in Jack Daniel’s Props., Inc. v. VIP Products, LLC, 599 U.S. 140 (2023). The views expressed herein do not necessarily represent the author’s views.
1. 599 U.S. 140 (2023).
2. 875 F.2d 994, 999 (2d Cir. 1989).
3. See J. McCarthy, Trademarks & Unfair Competition § 2:2 (5th ed. 2019).
4. “‘[A] trademark [is] any word, name, symbol, or device, or any combination thereof’ that a manufacturer uses to distinguish the manufacturer’s goods from those manufactured or sold by others and to ‘indicate the source of the goods.’” Vans, Inc. v. MSCHF Prod. Studio, Inc., 88 F.4th 125, 135 (2d Cir. 2023) (quoting the Lanham Act, 15 U.S.C. § 1127) [hereinafter “MSCHF II”].
5. For example, one of the more recognizable trademarks today is the adidas “Three-Stripe” trademark (among other adidas trademarks), rights to which adidas seeks to enforce often. See, e.g., adidas America, Inc. v. Skechers USA, Inc., 890 F.3d 747, 752 (9th Cir. 2018).
6. See, e.g., Novalogic, Inc. v. Activision Blizzard, 41 F. Supp. 3d 885, 897–902 (C.D. Cal. 2013) (recognizing expressive, artistic values in video games and applying Rogers and First Amendment analysis to find for the alleged infringers). Note that caselaw preceding Jack Daniel’s Props. v. VIP Prods., Inc, 599 U.S. 140 (2023) might start to erode after the Supreme Court’s ruling, as discussed supra.
7. U.S. Const. amend. I.
8. MSCHF II, 88 F.4th at 136 (summarizing the Lanham Act as requiring the plaintiff to show, in relevant part, that “defendant’s use of a similar mark is likely to cause consumer confusion as to the origin or association of the goods or services”).
9. Compare Playboy Enterprises, Inc. v. Netscape Commc’ns Corp., 354 F.3d 1020, 1026 (9th Cir. 2004) (setting out the eight-factor “Sleekfcraft” test used to assess likelihood of confusion in the Ninth Circuit), with A&H Sportswear, Inc. v. Victoria’s Secret Stores, Inc., 237 F.3d 198, 211 (3d Cir. 2000) (setting out the ten-factor “Lapp” test used to assess likelihood of confusion in the Third Circuit).
10. Vans, Inc. v. MSCHF Prod. Studio, Inc., 602 F. Supp. 3d 358, 367 (E.D.N.Y. 2022) (citing Polaroid Corp. v. Polarad Elecs. Corp., 287 F.2d 492, 495 (2d Cir. 1961)) [hereinafter “MSCHF I”].
11. See, e.g., id. at 370 (noting that the product in question was “a parodic or artistic expression of Plaintiff’s [mark] and thus protected by the First Amendment”).
12. 875 F.2d 994, 999 (2d Cir. 1989).
13. MSCHF II, 88 F.4th at 136 (citing Rogers, 875 F.2d at 999).
14. See VIP Prods., 599 U.S. at 157 (“[T]he Rogers test . . . offers an escape from the likelihood-of-confusion inquiry and a shortcut to dismissal.”).
15. See Rogers v. Grimaldi, 875 F.2d 994, 999 (2d Cir. 1989) (reasoning that “the public’s interest in free expression” did not support application of the Lanham Act’s usual trademark infringement provisions “unless the [allegedly infringing use of the mark] has no artistic relevance to the underlying work whatsoever, or . . . explicitly misleads as to the source or content of the work”).
16. Id. at 996.
17. Id. at 997.
18. Id.
19. Rogers, 875 F.2d at 999.
20. Id.
21. Despite this, the Supreme Court declined to adopt or overturn the Rogers doctrine. See VIP Prods., 599 U.S. at 163 (“We do not decide whether the Rogers test is ever appropriate.”).
22. VIP Prods., 599 U.S. at 155–56 (“Rogers . . . has always been a cabined doctrine. . . . . The test has not insulated from ordinary trademark scrutiny the use of trademarks as trademarks.”).
23. Id. at 144–45.
24. Id. at 154.
25. See MSCHF II, 88 F.4th at 137 (citing VIP Prods., 599 U.S. at 153). The Supreme Court did not provide much to guide lower courts in assessing what “source-identifying” use is.
26. VIP Prods., 599 U.S. at 155 (“[W]hatever you make of Rogers—and again, we take no position on that issue—it has always been a cabined doctrine.”) (emphasis added).
27. Id. at 163 (“[O]n infringement, we hold only that Rogers does not apply when the challenged use of a mark is as a mark,” or, in other words, is “source-identifying.”).
28. Id. at 159–60 (“VIP conceded [that it used its Bad Spaniel’s trademark as a source identifier] below.”).
29. Rogers v. Grimaldi, 875 F.2d 994, 998 (2d Cir. 1989).
30. See Jack Daniel’s Props., Inc. v. VIP Prods., LLC, 599 U.S. 140, 158 (2023).
31. 88 F.4th 125 (2d Cir. 2023).
32. Hermès Int’l v. Rothschild, 2023 WL 4145518 (S.D.N.Y. June 23, 2023).
33. Hermès Int’l v. Rothschild, 654 F. Supp. 3d 268, 272 (S.D.N.Y. 2023).
34. Id. at 273–74.
35. See Hermès, 2023 WL 4145518, at *5. Notably, pre-trial argument on the Rogers issue preceded the VIP Products case. Thus, the jury was instructed to consider the Rogers defense: “[E]ven the modest elements of artistic expression contained in Rothschild’s works entitled him to total First Amendment protection against Hermès’ claims unless Hermès proved that Rothschild intentionally misled consumers.” Id., at *1. Even still, the jury decided against Rothschild at trial, and Judge Rakoff of the United States District Court for the Southern District of New York denied Rothschild’s post-trial motions. Id.
36. MSCHF, 88 F.4th at 129.
37. Id. at 129–30.
38. Id. at 139.
39. Perhaps understandably, given that VIP Products came down after trial in the Hermès case, the court dismissed Rothschild’s reassertion of a First Amendment defense in post-trial motions by simply opining that using “a trademark as a trademark . . . . is precisely what [Rothschild], with his ‘MetaBirkins’ NFTs and ‘MetaBirkins’ website, did here.” 2023 WL 4145518, at *3.
40. MSCHF, 88 F.4th at 139.
41. Hermès, 2023 WL 4145518, at *3 (stating that because Rothschild used “MetaBirkins” in the title of his NFTs and his website the jury was correct—and indeed “accurately predicted” VIP Products’ holding—in concluding that First Amendment protections did not attach, necessarily concluding that that sort of use was source-identifying).
42. In other words, it did not matter that in Rothschild’s case the NFTs themselves (the product consumers were buying) were divorced from the actual MetaBirkin image (the use that might confuse consumers) that would populate the object image after purchase. See Hermès, 654 F. Supp. 3d at 274. Nor did it matter that in MSCHF’s case, Wavy Babys were exclusively sold through MSCHF’s proprietary app and shipped with a manifesto describing who created them—an inclusion that may suggest a lack of intent to appropriate goodwill. MSCHF, 88 F.4th at 141.
43. Hermès, 654 F. Supp. 3d at 272–73.
44. MSCHF, 88 F.4th at 139.
45. Jack Daniel’s Props., Inc. v. VIP Products LLC, 599 U.S. 140, 161 (2023).
46. Vans, Inc. v. MSCHF Prod. Studio, Inc., 602 F. Supp. 3d 358, 367–68, 370 (E.D.N.Y. 2022) (addressing the First Amendment argument only after finding “the Wavy Baby shoes will likely cause consumer confusion”).
47. VIP Prods., 599 U.S. at 161 (citing Louis Vuitton Malletier v. Haute Diggity Dog, 507 F.3d 252, 265 (4th Cir. 2007)). In Louis Vuitton, the Fourth Circuit considered a claim not dissimilar to the one at issue in VIP Products: Louis Vuitton sought to prevent a dog toy company from mimicking its trademarks and trade dress on the dog toys. Louis Vuitton, 507 F.3d at 256. There, unlike in MSCHF, the Fourth Circuit first considered the parodic use and used its determination that the parody was successful to inform its likelihood of confusion analysis on the basis that a parodic use—although facially similar—under scrutiny served to highlight differences for the purpose of the likelihood of confusion analysis. Id. at 261.
48. See Louis Vuitton Malletier v. Haute Diggity Dog, 507 F.3d 252, 261 (4th Cir. 2007).
49. Jack Daniel’s Props., Inc. v. VIP Prods. LLC, 599 U.S. 140, 161 (U.S. 2023) (“[A] trademark’s expressive message—particularly a parodic one, as VIP asserts—may properly figure in assessing likelihood of confusion.”) (citing Louis Vuitton, 507 F.3d at 265, as exemplifying how parody “influences the way in which the [likelihood-of-confusion] factors are applied”)).
50. See Stouffer v. Nat’l Geographic Partners, LLC, 400 F. Supp. 3d 1161, 1179 (D. Colo. 2019).
51. Vans, Inc. v. MSCHF Prod. Studio, Inc., 88 F.4th 125, 140 (2d Cir. 2023).
52. See, e.g., Plus Prods. v. Plus Disc. Foods, Inc., 722 F.2d 999, 1004 (2d Cir. 1983) (“[The Polaroid factors] do[] not exhaust the possibilities—the court may have to take still other variables into account [to assess likelihood of confusion].”)
53. Cf. VIP Prods., 599 U.S. at 161 (“There is no threshold test working to kick out all cases involving ‘expressive works.’ . . . [A] trademark’s expressive message . . . may properly figure in assessing the likelihood of confusion.”).