Harvard Law Review Harvard Law Review Harvard Law Review

First Amendment: Speech

In re Tam

Federal Circuit Holds the Lanham Act's Antidisparagement Provision Unconstitutional.

Passed in 1946, section 2(a) of the Lanham Act1×1. Ch. 540, 60 Stat. 427 (1946) (codified as amended in scattered sections of 15 U.S.C.). (§ 2(a)) authorizes the Patent and Trademark Office to deny registration of marks that “disparage . . . persons, . . . institutions, beliefs, or national symbols.”2×2. 15 U.S.C. § 1052(a) (2012). The Lanham Act is the federal law of trademarks, and § 2 specifies conditions under which marks cannot be registered. The list includes, but isn’t limited to, descriptive, functional, deceptively misdescriptive, deceptive, confusing, scandalous, immoral, and disparaging marks. See id. § 1052(a), (d), (e). Though mark holders have challenged the provision on free speech grounds in the past, courts have categorically rejected such claims, deeming registration uncovered by the First Amendment.3×3. See, e.g., In re Mavety Media Grp. Ltd., 33 F.3d 1367, 1374 (Fed. Cir. 1994); see also Test Masters Educ. Servs., Inc. v. Singh, 428 F.3d 559, 578 n.9 (5th Cir. 2005). Recently, in In re Tam,4×4. 808 F.3d 1321 (Fed. Cir. 2015) (en banc). the Federal Circuit reversed course.5×5. See id. at 1328. Because the provision was a content- and viewpoint-based regulation of trademark’s “expressive” aspect, the en banc court held that it was subject to and failed to satisfy strict scrutiny. As a normative matter, Tam effected a partial shift in regulatory control from the legislature to the judiciary in an area of law particularly unfit for judicial management. Doctrinally, it constitutes a problematic compromise with the recent Supreme Court case Reed v. Town of Gilbert,6×6. 135 S. Ct. 2218 (2015). which significantly expanded the definition of content-based laws deemed presumptively unconstitutional.7×7. See id. at 2226–27. In an effort to insulate trademark law’s central provisions from Reed, Tam claimed the unprecedented power to distinguish between trademark’s expressive and commercial regulations and subject each to varying levels of scrutiny. But this move is too tenuous to cabin Reed’s deregulatory implications for key areas of trademark law.

In 2011, Simon Shiao Tam, “‘front man’ for the Asian-American dance-rock band The Slants,” sought federal registration for his band’s mark.8×8. Tam, 808 F.3d at 1331. The examining attorney refused under § 2(a), finding the band name likely disparaged people of Asian descent.9×9. Id. Tam appealed, and the Trademark Trial and Appeal Board affirmed, holding that his mark satisfied both prongs of the governing test: (1) its “likely meaning . . . refer[red] to identifiable persons,” and (2) it disparaged “a substantial composite of the referenced group.”10×10. In re Tam, 108 U.S.P.Q.2d (BNA) 1305, 1309, 1311–13 (T.T.A.B. 2013). Almost as an afterthought, the Board cited to In re McGinley11×11. 660 F.2d 481 (C.C.P.A. 1981). for the proposition that because Tam could still call his band The Slants, registration refusal didn’t implicate his speech rights.12×12. Tam, 108 U.S.P.Q.2d (BNA) at 1313 (citing McGinley, 660 F.2d at 484). Again Tam appealed, this time raising a First Amendment challenge, and the Federal Circuit affirmed.13×13. In re Tam, 785 F.3d 567, 568–69, 571 (Fed. Cir. 2015). Writing for the panel, Judge Moore confirmed the Board’s analysis, noting McGinley foreclosed Tam’s constitutional claim.14×14. Id. at 572. Judge Moore was joined by Judges Lourie and O’Malley. Discontent with the result, Judge Moore penned “additional views,” urging the court to sit en banc and overturn its own precedent.15×15. Id. at 573 (Moore, J., additional views).

Sitting en banc, the court held the disparagement provision unconstitutional.16×16. Tam, 808 F.3d at 1328. Writing again for the majority, Judge Moore17×17. Judge Moore was joined by Chief Judge Prost and Judges Newman, O’Malley, Wallach, Taranto, Chen, Hughes, and Stoll. argued the provision was content and viewpoint discriminatory, and, therefore, subject to strict scrutiny.18×18. Tam, 808 F.3d at 1334–37. A law is content based if “its reach is defined . . . by the topic (subject matter) of the covered speech,”19×19. Id. at 1334 (citing Reed v. Town of Gilbert, 135 S. Ct. 2218, 2230 (2015)). and viewpoint based if it aims to exclude specific “viewpoints from the marketplace.”20×20. Id. at 1335 (quoting Simon & Schuster, Inc. v. Members of the N.Y. State Crime Victims Bd., 502 U.S. 105, 116 (1991)). The provision not only singled out disparaging speech as a subject matter for regulation,21×21. Id. (“It is beyond dispute that § 2(a) discriminates on the basis of content . . . .”). but also discriminated among viewpoints in attempting to eliminate disparaging, and not dignifying, marks from the market.22×22. Id. at 1337. Hence, it faced and failed strict scrutiny.23×23. Id. at 1339 (citing Sorrell v. IMS Health Inc., 131 S. Ct. 2653, 2667 (2011)).

The majority then rejected the government’s characterization of the provision as a regulation of “commercial speech,” subject to a lower standard of scrutiny.24×24. Id. at 1337–38. Under Central Hudson Gas & Electric Corp. v. Public Service Commission, 447 U.S. 557 (1980), commercial speech is subject to a brand of intermediate scrutiny. See id. at 566. Trademarks, in the majority’s view, embody two “inextricably intertwined” aspects25×25. Tam, 808 F.3d at 1339 (quoting Riley v. Nat’l Fed’n of the Blind of N.C., Inc., 487 U.S. 781, 796 (1988)). : they are both commercial “source identifier[s]” and expressive mediums of social discourse.26×26. Id. at 1338. Because the disparagement provision specifically targeted “the mark[’s] expressive content,” it was subject to strict scrutiny.27×27. Id. In contrast, § 2’s most central provisions — § 2(d) (bar on marks causing consumer confusion) and § 2(e) (bar on misdescriptive marks) — target misleading commercial speech, thereby evading strict scrutiny.28×28. See id. at 1329 (noting that § 2(d) and § 2(e) “do not run afoul of the First Amendment”).

Turning next to McGinley, the majority rejected the view that because Tam could still call his band The Slants, the disparagement provision left his speech rights undisturbed.29×29. See id. at 1333–34 (discussing In re McGinley, 660 F.2d 481, 484 (C.C.P.A. 1981)). Perry v. Sindermann30×30. 408 U.S. 593 (1972). repudiated this reasoning, barring the government from conditioning receipt of benefits on the relinquishment of a constitutional entitlement.31×31. See id. at 597; see also Sorrell v. IMS Health Inc., 131 S. Ct. 2653, 2664 (2011). Per Perry, the provision conditions the benefits of registration on the use of a nondisparaging mark and therefore “creates a serious disincentive to adopt [offensive] mark[s]” like The Slants.32×32. Tam, 808 F.3d at 1341. The majority referenced the numerous procedural and substantive benefits that attend to federal registration. Id. at 1340–41. The court deemed this “chilling effect” sufficient to violate the First Amendment.33×33. Id. at 1342. The court defended the proposition with citations to amici briefs. Id. at 1341.

The majority then considered and rejected the government’s classification of registration as either a “subsidy” or “government speech,” both of which lie largely outside the bounds of the First Amendment.34×34. Id. at 1345, 1348. As to the first, the majority argued that subsidies come in only two forms — monetary gifts or use of government property — and registration assumes neither.35×35. Id. at 1351. Even conditions on subsidies cannot leverage conduct outside the contours of the program. Because the disparagement provision “bears no relation to the [commercial] . . . purpose of . . . registration,” it would be unconstitutional even as a subsidy. Id. at 1354. As to the second, the court rejected the government’s position that it could “convert the underlying speech to government speech” simply by affixing a symbol (®) or listing the mark in a register.36×36. Id. at 1346. If this were so, copyright — which uses these same “indicia of registration” — would also constitute government speech, a result clearly “[in]consistent with the First Amendment.”37×37. Id. Judge Moore concluded that even if strict scrutiny didn’t apply, the provision would still fail the commercial-speech test in Central Hudson Gas & Electric Corp. v. Public Service Commission38×38. 447 U.S. 557 (1980). for lack of a legitimating state interest.39×39. Tam, 808 F.3d at 1355–57. Judge O’Malley, joined by Judge Wallach, concurred, arguing that the disparagement provision is void for vagueness. Id. at 1358–63 (O’Malley, J., concurring).

Judge Dyk concurred in part and dissented in part.40×40. Id. at 1363–74 (Dyk, J., concurring in part and dissenting in part). Judges Lourie and Reyna joined all but the as-applied portions of Judge Dyk’s opinion. Though he agreed the provision was unconstitutional as applied to Tam, he rejected the majority’s facial invalidation. First, he noted that many marks, devoid of expressive content, warrant only intermediate scrutiny. When applied to these purely commercial marks, the provision survives intermediate scrutiny because it “protect[s] underrepresented groups . . . from being bombarded with demeaning messages.”41×41. Id. at 1364. He then rejected the majority’s narrow view of subsidies, deeming registration a nonmonetary subsidy subject to mere reasonableness review.42×42. See id. at 1372. Writing for himself, Judge Dyk concluded the provision was still unconstitutional as applied to Tam’s “core political” speech.43×43. Id. at 1374.

Judge Reyna dissented as well.44×44. Id. at 1376 (Reyna, J., dissenting). Judge Lourie also dissented, invoking stare decisis and separation of powers concerns and expressing skepticism as to the burdens of nonregistration. Id. at 1374–76 (Lourie, J., dissenting). He argued that registration, “considered as a whole, is indisputably commercial, not political.”45×45. Id. at 1378 (Reyna, J., dissenting). Because the provision could have survived Central Hudson scrutiny, he would have upheld it both facially and as applied to Tam’s mark.46×46. Id. at 1378–82. In his view, the disparagement provision served the substantial government interest of “promot[ing] the orderly flow of commerce.” Id. at 1378.

Tam’s novelty — that denial of registration burdens a constitutional speech interest — effected an undesirable shift in regulatory control from the legislature to the judiciary, one that could be amplified by Reed’s broadened definition of “content-based” laws. Tam attempted to insulate trademark’s key provisions from Reed, but its doctrinal method proves wanting. According to the majority, courts can distinguish between trademark’s “expressive” and “commercial” regulations and subject each to varying levels of scrutiny. But this apparently unprecedented move is tenuous at best, leaving open the possibility that trademark’s most central provisions could face strict scrutiny.

As a normative matter, trademark presents a clash of speech-versus-speech, where vindicating one speech interest entails suppressing another.47×47. See id. at 1378 (“[I]f the expressive content of the mark precludes regulation, on what authority may the government grant Mr. Tam the exclusive right to use this mark . . . ?”). Consider the case of two bands both calling themselves The Slants: Tam’s group and a racist counterpart. As the majority observes, the disparagement provision burdens Tam’s speech interest in reappropriating the slur. But if Tam were to receive registration, he could then enjoin the racist group from exercising its racist speech interest in the mark. In this zero-sum game, § 2 reflects legislative choices as to whose speech should prevail.48×48. Cf. id. at 1374 (Lourie, J., dissenting) (“[The majority] interfere[s] with the long-standing Congressional policy of . . . filter[ing] out . . . undesirable marks . . . .”). Before Tam was decided, when registration was uncovered by the First Amendment, courts generally let these legislative choices lie.49×49. See, e.g., In re McGinley, 660 F.2d 481, 484 (C.C.P.A. 1981). Tam’s novelty was to constitutionalize a speech interest in registration and thereby wrest power from the legislature to partially redraw trademark’s boundaries.50×50. To be clear, courts have long recognized that registration can threaten the constitutional speech interests of enjoined parties. See, e.g., Lamparello v. Falwell, 420 F.3d 309, 313 (4th Cir. 2005); Rogers v. Grimaldi, 875 F.2d 994, 998 (2d Cir. 1989). The inverse claim — that registration refusal threatens the mark holder’s speech interests — is Tam’s novelty.

Trademark’s speech-versus-speech quality compounds the countermajoritarian difficulty here on two accounts.51×51. See generally Alexander M. Bickel, The Least Dangerous Branch 16–23 (Yale Univ. Press 2d ed. 1986) (1962) (identifying legitimacy concerns that arise when unelected judges strike down laws enacted by democratically accountable bodies). First, the moral quandary: between racist and reappropriative Slants, whose speech right ought to prevail? Should both groups be granted nonexclusive rights to the mark, or should the first to register win the power to exclude?52×52. Intuitions diverge on this point. See, e.g., Sonia K. Katyal, Trademark Intersectionality, 57 UCLA L. Rev. 1601, 1686 (2010) (“[C]ensorship does not stem from section 2(a), but from the trademark owner’s power to control alternate commentaries on a specific mark . . . .”); Alex Kozinski & Stuart Banner, Who’s Afraid of Commercial Speech?, 76 Va. L. Rev. 627, 644 (1990) (“[O]ur usual method of correcting for bias in public debate is not to suppress speech but to encourage counterspeech.”). In settling this rights contest, trademark regulators must decide on the moral status of intellectual property rights,53×53. Cf. Intellectual Property and First Principles, Cato Inst. (Feb. 10, 2016), http://www.cato.org/events/intellectual-property-first-principles [http://perma.cc/9AT2-LYZQ] (discussing intralibertarian dispute as to the moral status of intellectual property rights). “essentially contested concepts”54×54. See generally W.B. Gallie, Essentially Contested Concepts, 56 Proc. Aristotelian Soc’y 167 (1956) (describing concepts whose contents invoke disputed normative judgments). best left to the legislature.55×55. See Obergefell v. Hodges, 135 S. Ct. 2584, 2622 (2015) (Roberts, C.J., dissenting) (“[A] Justice’s commission does not confer any special . . . insight sufficient to justify imposing [his or her own moral] perceptions on fellow citizens . . . .”); Planned Parenthood of Se. Pa. v. Casey, 505 U.S. 833, 1001 (1992) (Scalia, J., dissenting) (“Value judgments . . . should be voted on, not dictated . . . .”). Second, the epistemic difficulty: to draw a defensible boundary, regulators require contextual, empirical data as to the incentive effects of registration refusal.56×56. See generally Am. Booksellers Ass’n v. Hudnut, 771 F.2d 323, 329 n.2 (7th Cir. 1985) (“[A]s judges we must accept the legislative resolution of . . . disputed empirical questions.”); William McGeveran, Rethinking Trademark Fair Use, 94 Iowa L. Rev. 49, 65 (2008) (considering the complex incentive effects of trademark’s doctrinal details). Consider Tam: did the provision really “chill” Tam’s speech?57×57. Compare Tam, 808 F.3d at 1341 (describing a chilling effect), with id. at 1374–75 (Lourie, J., dissenting) (disputing the existence of a chilling effect in Tam’s case). Even so, might it have had the justifying effect of maximizing aggregate speech?58×58. Compare Katyal, supra note 52, at 1686 (“[S]ection 2(a) might indirectly . . . add[] more speech, rather than less, to the marketplace . . . .”), with Erwin Chemerinsky, Balancing Copyright Protections and Freedom of Speech: Why the Copyright Extension Act Is Unconstitutional, 36 Loy. L.A. L. Rev. 83, 83 (2002) (“Without copyright protections, less speech would occur.”). In short, Tam’s novelty invites courts to implicitly adjudicate trademark’s murkiest moral and empirical questions.

A recent shift in the Supreme Court’s First Amendment jurisprudence risks exacerbating this normative problem. Last Term in Reed, the Court vastly expanded the category of content-based regulations deemed presumptively unconstitutional.59×59. See Norton v. City of Springfield, 806 F.3d 411, 413 (7th Cir. 2015) (Manion, J., concurring). Before Reed, laws were content based when animated by disagreement with the regulated speech.60×60. See id. at 412 (majority opinion) (“[Before Reed,] the absence of an effort to burden unpopular ideas implie[d] the absence of content discrimination.”). Reed’s definition swept more broadly, including any law “that depend[s for its application] on an evaluation of the content of the speech.”61×61. Rebecca Tushnet, Essay, The First Amendment Walks into a Bar: Trademark Registration and Free Speech, 91 Notre Dame L. Rev. (forthcoming 2016) (manuscript at 2), http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2745016 [http://perma.cc/PBG2-CRES]; see also id. (manuscript at 22–23) (discussing Reed v. Town of Gilbert, 135 S. Ct. 2218 (2015)). Decided in our post-Reed world, Tam rightly branded the disparagement provision content based.62×62. See Tam, 808 F.3d at 1334 (citing Reed, 135 S. Ct. at 2226). But as Professor Rebecca Tushnet has argued, the same could be said of § 2(d) and § 2(e): whether a mark is confusing or misdescriptive will similarly depend upon its content.63×63. See Tushnet, supra note 61 (manuscript at 2) (“Section 2 is almost nothing but content-based.”); see also Reed, 135 S. Ct. at 2234 (Breyer, J., concurring in the judgment) (“Regulatory programs almost always require content discrimination.”). Taken together, Tam’s novelty and Reed’s broad definition of “content based” might be used to justify an expansion of strict scrutiny to § 2’s key content-based provisions.

The Tam majority seemed to have understood that its holding could be extended beyond the disparagement provision.64×64. See Tam, 808 F.3d at 1330 n.1 (“We limit our holding in this case to the constitutionality of the § 2(a) disparagement provision. . . . [O]ther portions of § 2 may likewise constitute government regulation of expression based on message . . . .”). And so to insulate § 2’s key provisions from strict scrutiny, the majority offered a three-step doctrinal argument: (1) Reed covers only expressive speech;65×65. Tam made this move implicitly when it identified Central Hudson’s intermediate scrutiny as the governing standard for commercial speech. See Tam, 808 F.3d at 1355; see also id. at 1329 (citing to Central Hudson for the claim that § 2(d) and § 2(e) don’t “run afoul of the First Amendment”). Query whether this is the best reading of Reed. Compare Mass. Ass’n of Private Career Sch. v. Healy, No. 14-13706-FDS, 2016 WL 308776, at *10–11 (D. Mass. Jan. 25, 2016) (limiting Reed to noncommercial speech), and Note, Free Speech Doctrine After Reed v. Town of Gilbert, 129 Harv. L. Rev. 1981, 1990–92 (2016), with Sorrell v. IMS Health Inc., 131 S. Ct. 2653, 2664 (2011) (applying “heightened scrutiny” to some commercial speech restrictions). (2) commercial and expressive aspects are “inextricably intertwined” in registration;66×66. Tam, 808 F.3d at 1339 (citing Riley v. Nat’l Fed’n of the Blind of N.C., Inc., 487 U.S. 781, 796 (1988)). and (3) the standard of scrutiny for a particular trademark provision should depend upon which aspect it regulates. In defense of this last move, Tam cited to R.A.V. v. City of St. Paul67×67. 505 U.S. 377 (1992). for the proposition that “speech can be proscribable on the basis of one feature [e.g., its commercial function] . . . but not [necessarily] on the basis of another [e.g., its expressive content].”68×68. Tam, 808 F.3d at 1339 (quoting R.A.V., 505 U.S. at 385). On this view, Tam could limit Reed to the disparagement provision (a regulation of expressive content), while leaving § 2(d) and § 2(e) (regulations of trademark’s commercial function) completely undisturbed.69×69. See id. at 1338. What’s more, because these latter provisions bar distinctly deceptive commercial speech, they need only satisfy reasonableness review.70×70. See id. at 1329 (citing Cent. Hudson Gas & Elec. Corp. v. Pub. Serv. Comm’n, 447 U.S. 557, 563 (1980)).

But this argument rests § 2’s constitutionality on tenuous doctrinal grounds. When the Court first introduced “inextricably intertwined” speech in Riley v. National Federation of the Blind of North Carolina, Inc.,71×71. 487 U.S. 781. it expressly rejected Tam’s limiting principle, holding that speech does not “retain[] its commercial character when it is inextricably intertwined with . . . fully protected speech.”72×72. Id. at 796. In Riley, the Court struck down a state law requiring fundraisers to disclose the percentage of charity they retained as commission in the prior year. See id. at 798. As recognized by other circuits, Riley requires a ratcheting up of protection whenever speech’s commercial and expressive aspects are sufficiently intertwined.73×73. See, e.g., Jordan v. Jewel Food Stores, Inc., 743 F.3d 509, 520–21 (7th Cir. 2014); Dex Media W., Inc. v. City of Seattle, 696 F.3d 952, 962 (9th Cir. 2012) (“We treat the entirety of the intertwined speech as fully protected expression.” (citing Riley, 487 U.S. at 796)). Assuming, as Tam did, that the commercial and expressive really are intertwined in registration,74×74. See Tam, 808 F.3d at 1339. any effort to regulate the former should trigger the full protections accorded to the latter.75×75. Whereas Riley refused to “parcel out the [intertwined] speech, applying one test to one phrase and another test to another phrase,” Riley, 487 U.S. at 796, Tam — though it recognized trademark as an instance of inextricably intertwined speech — sought to untangle the expressive and commercial dimensions of a single word, a move ruled out by Riley a fortiori. And though it’s true that § 2(d) and § 2(e) regulate purportedly deceptive speech, the Supreme Court has held in United States v. Alvarez76×76. 132 S. Ct. 2537 (2012). that expressive deception warrants full constitutional protection, absent a clear showing of harm.77×77. See id. at 2549; Tushnet, supra note 61 (manuscript at 23 n.75). Section 2(e) would likely fail this test because it “doesn’t require . . . that consumers would change their behavior because of the confusion or that the senior trademark owner would suffer . . . harm.” Id. (manuscript at 32). Riley, in requiring that intertwined speech be treated like expressive speech, would seem to trigger Alvarez, subjecting even § 2(d) and § 2(e) to strict scrutiny.78×78. See Tushnet, supra note 61 (manuscript at 23 n.75). Under the tetrarchy of Riley, Reed, Alvarez, and now Tam, just about every provision in § 2 could face the fatal test of narrow tailoring.79×79. Both Tam’s argument and the above criticism assume without argument that trademark constitutes inextricably intertwined speech. For a defense of this assumption, see Kristian D. Stout, Terrifying Trademarks and a Scandalous Disregard for the First Amendment: Section 2(A)’s Unconstitutional Prohibition on Scandalous, Immoral, and Disparaging Trademarks, 25 Alb. L.J. Sci. & Tech. 213, 242–49 (2015).

In reply, Tam would likely reemphasize R.A.V., where the Court struck down a statute — similar to the one in Tam — that singled out and criminalized disparaging “fighting words.”80×80. R.A.V. v. City of St. Paul, 505 U.S. 377, 380 (1992). Though fighting words are generally excepted from First Amendment protection, R.A.V. held they still cannot be regulated on the basis of “nonproscribable,” racist content.81×81. Id. at 385–86. Analogously, Tam held registration regulable on the basis of a mark’s commercial misleadingness, just not on the basis of its expressive, offensive message.82×82. See Tam, 808 F.3d at 1355. But this analogy engenders doctrinal tension: for R.A.V. has never been applied to intertwined speech,83×83. Tam appears to be the first and only case to use R.A.V. to untangle inextricably intertwined speech. Cf. Gaudiya Vaishnava Soc’y v. City and County of San Francisco, 952 F.2d 1059, 1065 (9th Cir. 1990) (explicitly rejecting an R.A.V.-style argument for regulating intertwined speech). and by extending its logic to the commercial/expressive distinction, Tam placed it on a direct collision course with Riley’s ratcheting-up requirement. Perhaps this tension reflects the unsettled nature of intertwined speech law.84×84. See, e.g., Katyal, supra note 52, at 1647; Stout, supra note 79, at 242. But with Reed lying in wait, courts should seek more than a contestable argument from an unresolved area of doctrine to shield § 2 from heightened scrutiny.

Within a few weeks of Tam, the government conceded that § 2(a)’s bar on immoral and scandalous marks was also unconstitutional.85×85. In re Brunetti, No. 2015-1109 at 1–2, In re Brunetti, No. 15-1109 (Fed. Cir. Jan. 21, 2016). This may have been Tam’s easiest extension; but given Reed, Riley, and Alvarez, it needn’t be the last. For the sake of both a functioning trademark system and a coherent free speech jurisprudence, courts should “hesitate, and then hesitate some more”86×86. Kansas v. Nebraska, 135 S. Ct. 1042, 1061 (2015). before constitutionalizing the mark holder’s right to silence dissent.87×87. Cf. Katyal, supra note 52, at 1686 (“[C]ensorship . . . stem[s] from . . . the trademark owner’s power to control alternate commentaries on a specific mark . . . .”). Even free speech absolutists can agree that a subject so fraught with rights conflicts and epistemic complexity should rest primarily with the legislature.