The Semiotic Analysis of Trademark Law

Part IV offers a historical account of the fall of source distinctiveness and the rise of differential distinctiveness as the cynosure of the law. This is the story of the emergence of the modern trademark as a "floating signifier." While the economic account continues to profess that trademarks do little more than minimize consumer search costs, much of modern trademark law is now directed towards the commodification of semiotic "sign value." Part V analyzes the concept of trademark dilution and distinguishes between two modes of antidilution protection: antiblurring protection, which seeks to protect relations of source distinctiveness and which is essentially an inverted form of anti-infringement protection, and uniqueness protection, which seeks to protect relations of differential distinctiveness and which is absolute in scope. Courts and commentators have embraced the former, "compromise" mode of antidilution protection as an alternative to the latter "radical" mode. Nevertheless, as if by its own semiotic logic, antidilution protection of whatever mode inevitably degenerates into the provision of absolute, in gross property rights. Part V also considers other issues in dilution doctrine and argues on semiotic grounds, most notably, that trademark tarnishment is not a form of trademark dilution and that the requirement that a trademark be inherently distinctive to merit antidilution protection is incoherent.

The Chicago School of law and economics has long offered a totalizing and, for many, quite definitive theory of American trademark law. This theory conforms to a programmatic thesis: "[T]rademark law, like tort law in general . . . can best be explained on the hypothesis that the law is trying to promote economic efficiency." 1 The Chicago School asserts that trademarks serve two efficiency-enhancing functions: First, trademarks lessen consumer search costs by making products and producers easier to identify in the marketplace, and second, trademarks encourage producers to invest in quality by ensuring that they, and not their competitors, reap the reputation-related rewards of that investment. 2 As a historiographical proposition, the school further asserts that American trademark law has evolved over time an efficient system of rules designed to facilitate these two functions. 3 The influence of this analysis is now nearly total. It has been adopted at the highest levels of American law. 4 No alternative account of trademark doctrine currently exists.

Though powerful, the economic analysis of trademark law remains incomplete. It cannot explain, predict, or justify certain outcomes in the law, nor can it articulate the need for necessary reforms. This is nowhere more apparent than in its treatment of the concepts of trademark "distinctiveness" and trademark "dilution." The economic analysis falls short for two reasons. First, trademark doctrine is a hybrid doctrine. It is not simply an economic doctrine elaborating the principles of the market. It is also, and at the same time, a semiotic doctrine elaborating the principles of sign systems, of language. If there is a "language of commodities," 5 then trademark doctrine is its grammar, and this grammar must be understood not simply in economic, but also in linguistic terms. The second reason follows in part from the first. In asserting that trademarks do no more than facilitate search and encourage quality, the Chicago School has long declined to acknowledge what is obvious: that firms produce trademarks as status goods, that consumers consume trademarks to signal status, and that courts routinely invest trademarks with legal protection in an effort to preserve this status-signaling function. The culture industries-and what industries aren't?-have long sold trademarks as commodities in their own right. Entire areas of trademark doctrine cannot be understood except as systems of rules designed to facilitate the commodification-indeed, the "industrial production" 6 -of social distinction.

This Article undertakes a semiotic analysis of trademark law in an effort to explain and reform what the economic analysis cannot. The Article L. REV. 199, 214-17 (1991) proceeds in five parts. Part I introduces semiotic thought and sets forth, in simplified form, certain concepts that are indispensable to a semiotic account of trademark law and the trademark system. Part II analyzes the internal structure of the trademark. Traditionally, trademark commentators have conceived of the trademark as a three-legged stool, a relational system consisting of a "signifier" (the tangible form of the mark), a "signified" (the semantic content of the mark, its meaning), and a "referent" (the product to which the mark is affixed). Much of twentieth-century trademark doctrine was developed in light of this framework. In recent decades, however, the structure of the trademark has changed radically. Consistent with what Henri Lefebvre has called the "decline of referentials" 7 in modern culture, the trademark need no longer identify any particular commodity (other than itself) in order to receive protection. While the realities of trademark merchandising have forced a merger of signifier and referent, firms' efforts to assert exclusive rights over functional product features have forced a merger of signified and referent. In many cases, the trademark referent is now little more than a legal fiction. The modern trademark is dyadic in structure.

Part III proposes a revised theory of trademark distinctiveness. The concept of distinctiveness is the hinge on which trademark law turns. Yet for all of its importance-or perhaps precisely because of its importance-distinctiveness has never been adequately theorized. Traditional notions of "inherent" and "acquired" distinctiveness tend to confuse more than they clarify. This part argues that trademark law should reconceptualize trademark distinctiveness as consisting of source distinctiveness and differential distinctiveness. Corresponding to the semiotic relation of signification, source distinctiveness describes the extent to which a trademark's signifier is distinctive of its signified. Corresponding to the semiotic relation of value, differential distinctiveness describes the extent to which a trademark's signifier is distinctive from other signifiers in the trademark system. This reconceptualization recommends, among other things, a new approach to trademark infringement analysis. To determine whether a trademark falls within the subject matter of anti-infringement protection, a court should consider whether or not it possesses source distinctiveness. To determine the scope of anti-infringement protection an eligible trademark should receive, a court should consider the extent of its differential distinctiveness.

Part IV proposes a doctrinal history of the relation between source distinctiveness and differential distinctiveness. Specifically, it tells the story of the fall of the former and the rise of the latter as the cynosure of the law. The 7.

See result of this inversion has been the emergence in recent times of trademarks as "floating signifiers." As Rosemary Coombe has noted, the brand name is the "quintessential self-referential sign." 8 Part V analyzes the concept of trademark dilution. When Frank Schechter, the founder of dilution doctrine in America, first argued in the 1920s that "the preservation of the uniqueness of a trademark should constitute the only rational basis for its protection," 9 he failed to specify which form of uniqueness antidilution protection should protect: the uniqueness of the trademark signifier in itself as against all other signifiers in the trademark system, or merely the uniqueness of the relation between that signifier and the source it signifies. Since Schechter's time, trademark commentary has embraced the latter, "blurring" form of antidilution protection on the unstated assumption that it entails a limited grant of property rights in the intramark relation of source distinctiveness, while the former entails an unlimited grant of property rights in intermark relations of differential distinctiveness. Yet despite the best efforts of courts and commentators, the blurring form of antidilution protection has proven to be no less absolute than its alternative. This part explains the semiotic logic that drives this result. It also seeks to bring semiotic learning to bear on a number of ambiguities and unresolved questions in dilution doctrine relating to the subject matter and scope of antidilution protection, and argues, most notably, that trademark tarnishment is not a form of trademark dilution and that the requirement that a trademark be inherently distinctive to qualify for antidilution protection is incoherent.

A representative introductory guide defines semiotics as "a domain of investigation that explores the nature and function of signs as well as the systems and processes underlying signification, expression, representation, and communication."

10 A more advanced guide describes "semiosis," the subject matter of semiotics, as "the processes and effects of the production and reproduction, reception and circulation of meaning in all forms, used by all kinds of agent[s] of communication." 11 Defined in these terms-and they are typical 12 -semiotics would appear to be everything and nothing. Indeed, semiotics has traced its origins at least as far back as Heraclitus and Hippocrates, 13 and extended its reach at least as far out as proxemics, 14 the semiotics of space, and chronemics, 15 the semiotics of time. Even so, semioticians know their field when they see it. What defines semiotics is not so much its subject matter as its intellectual tradition and intellectual sensibility.

1. The Modern Semiotic Tradition: Saussure and Peirce Modern semiotics rose from inauspicious beginnings. Its two principal founders, the Swiss linguist Ferdinand de Saussure (1857-1913) and the American philosopher Charles Sanders Peirce (1839-1914), were entirely unaware of each other's existence. 16 Both would become famous only posthumously. Saussure's masterwork, the Course in General Linguistics (the Course), was published three years after his death, and was not actually written by Saussure. Two of his colleagues compiled the work from lecture notes-Saussure's, their own, and those of various students-and the few remaining manuscripts he failed to destroy before his death. 17 In the Course, Saussure projected a bold extension of his research in structural linguistics: "A science that studies the life of signs within society is conceivable . . . I shall call it semiology (from Greek semeîon 'sign'). Semiology would show what constitutes signs, what laws govern them." 18 involving the study of such phenomenon as "symbolic rites, polite formulas, military signals, etc.," 19 but did not live to conduct this work. The Course itself inspired others to do so, however, and much more. The Course is the foundational text of the twentieth century's structuralist and poststructuralist thought.

That modern semiotics should arise in part from the work of Charles Sanders Peirce is even more surprising. 20 Largely ignored by philosophers of his time, Peirce is now remembered as an eccentric and prickly character, and revered as a brilliant polymath, whom Bertrand Russell called "the greatest American thinker ever" 21 and Karl Popper called "one of the greatest philosophers of all time," 22 and who died isolated and in poverty. Peirce's work in logic, linear algebra, and the philosophy of pragmatism 23 is at least as important as and thoroughly informed by his contribution to semiotics. Peirce's thoroughgoing, even radical pansemiotism-"all this universe is perfused with signs, if it is not composed exclusively of signs" 24 -formed the basis of his "semeiotic" 25 and of his more general philosophy of categories. 19. SAUSSURE (Baskin), supra note 18, at 16. 20.

"Peirce" is pronounced "purse." For an account of Peirce's life, see JOSEPH BRENT, CHARLES SANDERS PEIRCE: A LIFE (1993). See also LOUIS MENAND, THE METAPHYSICAL CLUB (2001). Some biographical detail may also be found in Roberta Kevelson's extensive research on the commonalities between Peirce's thought and legal theory. See Peirce preferred the term "pragmaticism," which he considered to be "ugly enough to be safe from kidnappers." 5 CHARLES SANDERS PEIRCE, COLLECTED PAPERS OF CHARLES SANDERS PEIRCE ¶ 414 (Charles Hartshorne & Paul Weiss eds., 1934). On the connection between Peirce and American pragmatism, see generally A.J. AYER, THE ORIGINS OF PRAGMATISM: STUDIES IN THE PHILOSOPHY OF CHARLES SANDERS PEIRCE AND WILLIAM JAMES (1968).

24. 5 PEIRCE, supra note 23, ¶ 448 n.1. On Peirce's pansemiotism, consider his statement to Victoria Lady Welby that "it has never been in my power to study anything,-mathematics, ethics, metaphysics, gravitation, thermodynamics, optics, chemistry, comparative anatomy, astronomy, psychology, phonetics, economic [sic], the history of science, whist, men and women, wine, metrology, except as a study of semeiotic From its origins in Saussure and Peirce, semiotics has progressed along two sometimes converging, sometimes diverging lines of development. 27 The Peircean "philosophical" or "logical" tradition has continued in the works of Charles Morris, 28 I.A. Richards, 29 and Charles Ogden. 30 The Saussurean "linguistic" tradition has continued in the works of Roman Jakobson, 31 Claude Lévi-Strauss, 32 Roland Barthes, 33 Julia Kristeva, 34 and Jean Baudrillard,35 among others. If the latter names sound both foreign and more familiar to the American reader, it is because, for better or worse, the Saussurean tradition has dominated the American scene in the postwar period, sometimes under the very broad rubric "theory."

2. The Modern Semiotic Sensibility: Sign, Structure, and Difference Despite their apparent schism, the philosophical and linguistic traditions of semiotics have much more in common than what separates them. There is, as the leading American semiotician Thomas Sebeok asserts, a consistently shared point of view. 36 I will have much more to say about this point of view through the course of this Article, but a few generalities concerning it should be listed here.

See SEBEOK, supra note 16, at 63 (discussing the "two traditions" of semiotic thought); THOMAS SEBEOK, CONTRIBUTIONS TO THE DOCTRINE OF SIGNS 55 (1976) (discussing the difference between semiotics and semiology).

28. First, just as atomic physics' primary object and instrument of analysis is the atom, and psychoanalysis' is the psyche, so semiotics' is the sign. The semiotic tradition defines the sign not by what it is, but by what it does. That is, the sign is defined functionally rather than ontologically. 37 As Peirce wrote, the sign is "something which stands to somebody for something in some respect or capacity." 38 The elaboration of the sign-form, of intrasign structure and intersign relations, is probably the central achievement of the semiotic tradition. 39 Second, semiotics is an emphatically structuralist enterprise. It is, in essence, an extension of structural linguistics to all aspects of experience.

As such, it holds that no element of experience is meaningful in itself. Meaning arises only in the structural relations or, more precisely, in the structural oppositions among elements. 41 The axial idea in semiotic thought is difference, which it holds to be prior to identity and to meaning. To view the world through the semiotic lens is to view things for what they are not rather than for what they are.

Third and closely related, semiotics is generally systems-theoretical in orientation. The semiotic method typically attributes to semiotic structures or 37.

On the distinction between functional and ontological definitions, consider Terry Eagleton's discussion of the meaning of "literature":

There is no "essence" of literature whatsoever. Any bit of writing may be read "nonpragmatically," if that is what reading a text as literature means, just as any writing may be read "poetically." If I pore over the railway timetable not to discover a train connection but to stimulate in myself general reflections on the speed and complexity of modern existence, then I might be said to be reading it as literature. John M. Ellis has argued that the term "literature" operates rather like the word "weed": weeds are not particular kinds of plant, but just any kind of plant which for some reason or another a gardener does not what [sic] around. Perhaps "literature" means some thing like the opposite: any kind of writing which for some reason or another somebody values highly. As the philosophers might say, "literature" and "weed" are functional rather than ontological terms: they tell us about what we do, not about the fixed being of things. They tell us about the role of a text or a thistle in a social context, its relations with and differences from its surroundings, the ways it behaves, the purposes it may be put to and the human practices clustered around it. "Literature" is in this sense a purely formal, empty sort of definition. systems-the two concepts are often used interchangeably 42 -such characteristics as self-regulation and self-reference or autopoiesis. 43 Semiotics investigates the relation between "structure" (a language's underlying system of rules and regularities) and "event" (specific language uses). This relation is typically understood to be dialectical 44 and, in chicken-and-egg fashion, paradoxical: 45 Event assumes and informs structure and structure assumes and informs event. The semiotic method also tends carefully to observe the distinction between diachronic inquiry (the study of a system's evolution over 42.

Consider Lévi-Strauss's definition of "structure": [A] structure consists of a model meeting with several requirements. First, the structure exhibits the characteristics of a system. It is made up of several elements, none of which can undergo a change without effecting changes in all the other elements.

Second, for any given model there should be a possibility of ordering a series of transformations resulting in a group of models of the same type.

Third, the above properties make it possible to predict how the model will react if one or more of its elements are submitted to certain modifications.

Finally, the model should be constituted so as to make immediately intelligible all the observed facts. LÉVI-STRAUSS, STRUCTURAL ANTHROPOLOGY, supra note 32, at 279-80.

See, e.g., JEAN PIAGET, STRUCTURALISMS (Chaninah Maschler ed. & trans., 1970) (1968) (describing structure as having the characteristics of "wholeness," "transformation," and "self-regulation"). We tend to think that what we call the meaning of a word depends on the fact that it has been used by speakers on various occasions with the intention of communicating or expressing this meaning, and we thus might want to argue that what can in general be called the structure of a language-the general system of its rules and regularities-is derived from and determined by events: by acts of communication. But if we took this argument seriously and began to look at the events which are said to determine structures, we would find that every event is itself already determined and made possible by prior structures. The possibility of meaning something by an utterance is already inscribed in the structure of the language. The structures, of course, are themselves always products, but however far back we try to push, even when we think of the birth of language itself and try to describe an originating event that might have produced the first structure, we discover that we must assume prior organization, prior differentiation. For a caveman successfully to originate language by making a special grunt signify something like "food" is possible only if we assume that the grunt is already distinguished or distinguishable from other grunts and that the world has already been divided into categories of food and non-food. Signification always depends on differences: contrasts, for example, between food and non-food that allow "food" to be signified. JONATHAN CULLER, ON DECONSTRUCTION: THEORY AND CRITICISM AFTER STRUCTURALISM 95-96 (1982); see also time) and synchronic inquiry (the study of a system, of its intrarelations, at a given point in time), and ultimately to privilege the latter. 46 Fourth and finally, semiotics conceives of human language as the archetype for all cultural sign systems, and thus of linguistics as the "patron général" or "master-pattern for all branches of semiology." 47 "What semiotics has discovered," writes Julia Kristeva, "is that the law governing or, if one prefers, the major constraint affecting any social practice lies in the fact that it signifies; that is, that it is articulated like a language." 48 Enormously expansive in reach and ambition, 49 semiotics has studied as forms of language such practices as painting, 50 architecture, 51 urban space, 52 advertising, 53 fashion, 54 and most relevant for our purposes, consumerism and consumption. 55 As Jean

Saussure distinguished between synchronic and diachronic linguistics as follows: Synchronic linguistics will be concerned with the logical and psychological relations that bind together coexisting terms and form a system in the collective mind of speakers.

Diachronic linguistics, on the contrary, will study relations that bind together successive terms not perceived by the collective mind but substituted for each other without forming a system. SAUSSURE (Baskin), supra note 18, at 99-100. On the privileging of the synchronic over the diachronic, see JEAN-BAPTISTE FAGES, COMPRENDRE LE STRUCTURALISME 45-53 (1968).

Consider Roland Barthes' projected science of "arthrology": [T]he future task of semiology is far less to establish lexicons of objects than to rediscover the articulations which men impose on reality; looking in to the distant and perhaps ideal future, we might say that semiology and taxonomy, although they are not yet born, are perhaps meant to be merged into a new science, arthrology, namely, the science of apportionment. BARTHES, ELEMENTS, supra note 33, at 57. 50. 342 (1966) (describing a commodity as a sign consisting of a signifier "susceptible of exchange" and a signified "potential satisfaction"); WINFRIED NÖTH, HANDBOOK OF SEMIOTICS 445 (Indiana Univ. Press, 1990) (1985) (describing the "language of commodities" as a "semiotic system par excellence"); DON SLATER, CONSUMER CULTURE & MODERNITY 137-47 (1997); Umberto Eco, Social Life as a Sign System, in STRUCTURALISM: AN INTRODUCTION 57, 70 (David Robey ed., 1973) (describing the commodity system as "a system, structured by means of oppositions, as semiotic systems are"). Baudrillard, among many others, has suggested, consumption may be understood and analyzed "as a system of communication and exchange, as a code of signs continually being sent, received and reinvented-as language." 56

Since its first appearance in the classical period, semiotic thought has generated a great profusion of sign structurations. Most have fallen into one of two categories: those that are dyadic in structure, consisting of two subsign elements, and those that are triadic in structure, consisting of three subsign elements.

Saussure may be taken as the representative theorist of the dyadic sign model. 57 In the Course, he proposed that the sign is a "double-entity" 58 formed by the relation between a "signifier" and "signified," 59 as illustrated in Figure 1. Generally speaking, the signifier is the perceptible form of the sign (for example, the aural sound of a word, the visual appearance of a street sign, the olfactory smell of the ocean). 60 The signified consists of the particular mental idea for which the perceptible form stands. 61 In the sign BOOK, for example, the sound or appearance of the word "book" (the signifier) typically stands for the concept of a book (the signified). 62 The sound "book" is not itself a "sign" of the concept of a book. Rather, the sound and See id. at 65-67. Strictly speaking, the Saussurean signifier is not the material sound, a purely physical thing, but the psychological imprint of the sound, the impression that it makes on our senses. The sound-image is sensory, and if I happen to call it 'material,' it is only in that sense, and by way of opposing it to the other term of the association, the concept, which is generally more abstract. Id. at 66. 61.

Saussure's sign model is far subtler than this schematic presentation suggests. In his view, the sign is a purely psychic form, a "two-sided psychological entity," rather than a physical substance. Id. at 66. This form consists of the relation between (1) a concept and (2) a "sound-image" in the mind of the sound that represents that concept. From "concept" and "sound-image," Saussure proceeds to the more familiar terminology of signifier and signified. Id. at 66-67. The most sophisticated treatment of Saussure's sign model is found in PAUL J. the concept together form the sign. The sign is "the whole that results from the associating of the signifier with the signified." 63 FIGURE 1 THE SAUSSUREAN SIGN Saussure held that, at least in spoken and written language, the relation between the sign's signifier and its signified is "arbitrary." By this he meant that there is no natural connection between the concept of a book and the sound or appearance of the word "book"-or "Buch" or "livre" for that matter. Except in rare instances of onomatopoeia, signifiers are "unmotivated" by their signifieds. Their relation is established and sustained by convention alone: "It is because the linguistic sign is arbitrary that it knows no other law than that of tradition, and because it is founded upon tradition that it can be arbitrary." 64 As Claude Lévi-Strauss explained, the sign is arbitrary a priori, but non-arbitrary a posteriori. 65 The Saussurean distinction between signifier and signified may appear to be a simple matter of form and content (of word and idea), and semiotics is not embarrassed to restate the obvious. There is, however, one element missing from the Saussurean dyad: the actual physical object, the material thing referred to, the book in one's hands. 66 Saussure's "bracketing" of the referential object is an expression of his "linguistic turn." 67 67.

The "linguistic turn" is a [c]ollective designation for a range of otherwise quite disparate trends in twentieth-century thought. What they all have in common is an appeal to language, to discourse, or forms of linguistic representation as the furthest point that philosophy can reach in its quest for knowledge and truth. There are no "facts" outside language, and no "reality" other than that which presents itself under some linguistic description. Christopher Norris, Linguistic Turn, in THE OXFORD COMPANION TO PHILOSOPHY, supra note 12, at 492. analytic divisibility, the Saussurean signifier and signified form a structural relation of mutual interdependence; one cannot exist without the other. This Saussure asserts in a famous passage analogizing language to a sheet of paper:

Language might be called the domain of articulations . . . . Each linguistic term is a member, an articulus in which an idea is fixed in a sound and a sound becomes the sign of an idea.

Language can also be compared to a sheet of paper: thought is the front and the sound the back; one cannot cut the front without cutting the back at the same time; likewise in language, one can neither divide sound from thought nor thought from sound; the division could be accomplished only abstractedly, and the result would be either pure psychology or pure phonology.

Linguistics then works in the borderland where the elements of sound and thought combine; their combination produces a form, not a substance.

In emphasizing form rather than substance, Saussure rejects the "nomenclaturist" fallacy that language involves "a list of terms corresponding to a list of things" 69 or that "ready-made ideas exist before words." 70 Rather, language and experience are mutually constitutive and fully coextensive: "There are no pre-existing ideas, and nothing is distinct before the appearance of language." 71 To the extent that language forms every aspect of perceptual existence, it is not possible for language to refer to any objectified reality that is independent of linguistic structure. In Saussurean orthodoxy, the relation between sign and referent is identical to the relation between signifier and signified. Strictly speaking, the relation is not one of "standing for," but rather one of mutual linguistic "construal." 72 All the world is articulated as and through language: "There is nothing outside of the text." 73 (1979) See THIBAULT, supra note 62, at 214-15. Thibault explains: [T]he extra-linguistic domain is not an objectified reality which is independent of language form. Instead, the categories which are internal to a given language system are themselves incorporated into extralinguistic [sic] phenomena in the act of construing and constructing them. . . . . . . [L]anguage users use signs not to reflect an already given reality, but to intervene in and to structure reality along the parameters afforded by the internal design features of the sign itself. The sign imposes digital order on the analogue flux of perceptual experience.

Though many before him conceived of the sign as consisting of three elements, 74 Peirce may be taken, at least on the American scene, as the representative figure of the triadic structuration. Early in his semiotics, Peirce spoke of the sign in nontechnical terms as "a triple connection of sign, thing signified, cognition produced in the mind." 75 This ordinary language approach was not typical. Peirce's terminology is difficult, oft-changing, and sometimes inconsistent, as his most frequently cited formulation of his sign model evinces:

A sign, or representamen, is something which stands to somebody for something in some respect or capacity. It addresses somebody, that is, creates in the mind of that person an equivalent sign, or perhaps a more developed sign. That sign which it creates I call the interpretant of the first sign. The sign stands for something, its object. It stands for that object, not in all respects, but in reference to a sort of idea . . . . 76 Peirce's sign consists, then, of three elements, each of which corresponds to one of Peirce's three categories of Firstness, Secondness, and Thirdness. The first element, comparable to Saussure's signifier, is the representamen, the perceptible object, the "vehicle conveying into the mind something from without." 77 The second element is the object, or "referent" as this Article will call it, which can be a physical "object of the world" or a mental entity "of the nature of thought or of a sign." 78 The third element, comparable to the Saussurean signified, is the interpretant, which Peirce defined as "[creating] something in the Mind of the Interpreter," 79 the "proper significate effect," the "proper effect of the sign." 80 Peirce did not do so, but others have rendered this formulation graphically, in the form of a so-called "semiotic triangle" 81 shown in Figure 2.

See Peirce was an enthusiastic, even flamboyant taxonomist of signs. He proposed at one point no less than 3 10 or 59,049 different types of signs, 82 ranging from "[t]he feeling of the skylark's song in the morning" 83 to the "[s]ign of law." 84 In his most fundamental and still influential typology, he reduced his categories to three: a given sign is (i) iconic, and/or (ii) indexical, and/or (iii) symbolic. 85 Peirce classifies a given sign according to the categorical nature of its representamen-object (signifier-referent) relation. As its name suggests, an iconic sign is one in which the representamen resembles or imitates to some degree the object to which it refers, for example, representational paintings, imitative gestures, and scale models. An indexical sign is one in which the representamen is related in some causal, physical way to the object, for example, smoke, medical symptoms, sundials. Finally, a symbolic sign is one in which the connection between the representamen and its object is purely a matter of convention. As we saw above, Saussure called such signs "arbitrary."

The "master-pattern" for Peirce's semiotics is not human language, but human cognition, the structure and processes of thought. The philosophical See id. ¶ ¶ 247-49. Though they may seem to array themselves along a spectrum, Peirce's types are not mutually exclusive. An individual sign can be a symbol, an icon, an index, or any combination, depending on its use within particular contexts. For example, a map may fit within each of the categories simultaneously. As the result of a photochemical process, a photograph can be understood as both iconic and indexical of its subject, if not also symbolic. Due to its expense, a VOLVO brand automobile may be indexical of wealth, while also symbolic and perhaps iconic of wealth as well.

Interpretant (Signified) Object (Referent) Representamen (Signifier) Before beginning in earnest a semiotic analysis of trademark doctrine, we must establish the meaning of certain concepts regarding the process of semiosis itself, specifically, signification, value, and sign value. These will have a direct bearing on our analysis of trademark distinctiveness and of the difference between trademark infringement and trademark dilution.

Saussure's structural linguistics proceeds from a fundamental opposition: the opposition between langue and parole. Langue is the abstract system of rules and conventions that exists prior to and independently of individual instances of language use, or parole. 91 Langue is structure; parole is event. Langue is sometimes translated as "language," "code," or "competence," parole 86.

Eco as "speech" or "performance." When an individual learns a language, she assimilates langue; when she uses that language, she participates in instances of parole. Early on in the acquisition of a language, her parole may be significantly out of joint with langue. Over time, she corrects herself to conform to langue, or perhaps, if there are enough like her, langue will correct itself to conform to her parole. The relation between langue and parole, as between structure and event, is dialectical. The distinction between langue and parole implies a further distinction, that between signification and value. Signification describes the vertical, intrasign relations between signifier and signified within an instance of parole (Figure 3). 92 Value describes the horizontal, intersign relations among signifiers, signifieds, and signs generally across the system of values, of "reciprocal delimitation[s]," 93 of "articulations," 94 that constitutes langue (Figure 4). 95 Thus, whereas signification refers to the positive meaning of the sign, value refers to the negative difference or distinctiveness of the sign as against all other signs. Signification is one-dimensional equivalence; value is n-dimensional difference. To the extent that "differences carry signification," 96 value is that by virtue of which signification occurs. Signification, in other words, cannot obtain without value; identity cannot obtain without difference. 97

See BARTHES, ELEMENTS, supra note 33, at 48 ("[S]ignification can be conceived as a process; it is the act which binds the signifier and the signified, an act whose product is the sign.").

93. SAUSSURE (Harris), supra note 18, at 110. 94. SAUSSURE (Baskin), supra note 18, at 112. 95.

J.M. Balkin, The Hohfeldian Approach to Law and Semiotics, 44 U. MIAMI L. REV. 1119, 1121 (1990) ("[T]he relation between signifier and signified is mediated by the relationship of signifiers to each other in a general system of signification. Meaning in language, then, comes from the play of differences.").

Id. at 108 ("The linguistic mechanism is geared to differences and identities, the former being only the counterpart of the latter.").

Signification would appear to be an easily understood concept. The meaning of signification is ultimately based on the meaning of value, however, and value is probably the most obscure and unstable concept in all of Saussurean semiotics. It is also the most important (and of utmost importance to an understanding of the concept of distinctiveness in modern trademark doctrine). In general terms, value is a consummately structuralist notion. It conceives of identity not as something intrinsic, but rather as something dependent entirely on extrinsic, oppositional relations, that is, on relations of differential value to other identities in a system. "A language is a system of interdependent elements in which the value of any one element depends on the simultaneous coexistence of all the others." 98 Value describes place-value, context, and situation: "Signs function . . . not through their intrinsic value but through their relative position." 99 Value is the name given to the set comprising a language element's differential, oppositional relations with all other elements in a language system. 100 Consider Saussure's explanation of the workings of value on the level of the signified, of mental concepts:

In a given language, all the words which express neighbouring ideas help define one another's meaning. Each of a set of synonyms like redouter ("to dread"), craindre ("to fear"), avoir peur ("to be afraid") has its particular value only because they stand in contrast with one another. If redouter did not exist, its content would be shared out among its competitors . . . . So Saussure draws the same conclusions with respect to the workings of value on the level of the signifier: "Phonemes are characterized not, as one might think, by their own positive quality but simply by the fact that they are distinct. Phonemes are above all else opposing, relative, and negative entities." 102 In Saussure's description of language generally and of the workings of value specifically, one senses the operation of two overriding themes. The first is that language is empty, that it is form rather than substance. Language is the articulation of the world, the cutting-out of shapes, but not the filling-in of those shapes: "Language, in a manner of speaking, is a type of algebra consisting solely of complex terms." 103 A form's value is determined not by substance, but by the values of all other forms in the language. "[T]he value of a term may be modified without either its meaning or its sound being affected, solely because a neighboring term has been modified." 104 The second theme is that language involves competition, a form even of agon, and that this competition is characterized by zero-sum tradeoffs and disparate values in which "[s]ome of the oppositions language includes are more important than others." 105 In the operation of both themes, Saussure, of Geneva, shows the influence on his thinking of marginalist economic theory, specifically that of the School of Lausanne. 106 The influence is all the more apparent in Saussure's most cited explanation of the relation between value and signification, which takes the form of an extended market analogy:

To determine what a five-franc piece is worth one must therefore know: (1) that it can be exchanged for a fixed quantity of a different thing, e.g. bread; and (2) that it can be compared with a similar value of the same system, e.g. a one-franc piece, or with coins of another system (a dollar, etc.). In the same way a word can be exchanged for something dissimilar, an idea; besides, it can be compared with something of the same nature, another word. Its value is therefore not fixed so long as one 162-82 (1977). In Saussure's view, linguistics and political economy had much in common. Most fundamentally, both were "value sciences": "Here as in political economy we are confronted with the notion of value; both sciences are concerned with a system of equating things of different orders-labor and wages in one and a signified and signifier in the other." SAUSSURE (Baskin), supra note 18, at 79. As a result, both fields also required a strict separation between the diachronic, or historical, study of their respective phenomena and the synchronic, or static, study of those phenomena. Just as there is a division in the "economic sciences" between "political economy" and "economic history," explains Saussure, so in linguistics there is a "radical duality" between what he called "synchronic linguistics," on the one hand, and "diachronic linguistics," on the other. SAUSSURE (Harris), supra note 18, at 213.

Sign value is an often invoked but rarely explained term in social theory. It is typically used simply to refer to the capacity of status goods to signal high status, their "expression and mark of style, prestige, luxury, power, and so on." 111 Thus, it is said that the BMW has sign value and the DODGE does not. This definition of sign value descends from our notions of use value and exchange value, both of which draw upon essentially utilitarian conceptions of the term "value" ("worth in usefulness or importance to the possessor" 112 "[m]onetary or material worth" 113 ), rather than upon, say, a painterly conception of "value" ("the relative darkness or lightness of a color," 114 "the relation of one part or detail in a picture to another with respect to lightness and darkness" 115 ). In this utilitarian sense, sign value is understood as a special form of use value-a commodity may have various use values, one of which may be that it signals high status. This is not the definition of sign value that I want to apply to trademark doctrine, where it will yield few nonobvious insights. Rather, I want to establish here the concept's more technical meaning, particularly as it is set forth in the early work of Jean Baudrillard. To develop the concept of sign value, Baudrillard works not from the classical economic notions of use value and exchange value, but from the linguistic, Saussurean notion of value, that is, value as relational difference. In Baudrillard's social-semiotic theory, sign value describes a commodity's differential value as against all other commodities, and thus the commodity's capacity to differentiate its consumer. Sign value does not necessarily involve the conspicuous display of prestige or wealth or of scarce "positional goods." 116 Rather, it involves something more essential: the conspicuous display of distinctions, of "marginal differences," 117 of which there would appear to be an inexhaustible and easily renewable supply. Sign value is Saussurean structural value made explicit, signaled, displayed. It is formal, differential value performed. It is the abstract essence of what Justice Frankfurter called a trademark's "commercial magnetism." 118 To be sure, connotations of prestige may and often do issue from difference, but such connotations are merely an effect of sign 113.

See FRED HIRSCH, SOCIAL LIMITS TO GROWTH 27 (1976) ("[P]ositional goods" are valued in part for their scarcity and social exclusiveness.).

117. BAUDRILLARD, CONSUMER, supra note 6, at 90; see also BAUDRILLARD, CRITIQUE, supra note 35, at 75. Baudrillard states:

The origin of meaning is never found in the relation between a subject (given a priori as autonomous and conscious) and an object produced for rational ends-that is, properly the economic relation, rationalized in terms of choice and calculation. It is to be found, rather, in difference, systematizable in terms of a code (as opposed to private calculation)-a differential structure that establishes the social relation, and not the subject as such.

118 value; they are not sign value itself. 119 Such connotations merely give content to the differential form. Placing this in more concrete terms, it is well understood that firms seek to maximize the sign value of their brands and products. But, this does not mean that firms seek to enhance the prestige of their products. Rather, it means simply that firms seek to enhance their products' distinctiveness, to raise their products' profile. Consider a notorious print advertisement campaign undertaken by the apparel company United Colors of Benetton. The campaign consisted of graphic images of war, famine, disease, and sacrilege accompanied, inexplicably, by nothing but the company's logo. 120 The goal was certainly not to develop the meaning of the brand. Rather, the goal was to enhance the brand's sign value, its pure, abstract difference, as against other brands, and in this it succeeded. 121 The trademark IBM may stand for technological achievement, COKE and PEPSI for youth, WAL-MART for affordability-while none of these marks is a symbol of high status, each possesses a relatively high degree of sign value. That is to say, each is highly distinctive and each conveys its message, its signification, by means of its distinctiveness as against other marks. It is marketing orthodoxy that a trademark's most important quality is not the "esteem" in which it is held by consumers or its "relevance" to the lives of consumers, nor is it the "knowledge" consumers have of what the mark stands for. Rather, strong brands are characterized above all by "differentiation." 122 Distinction is their lifeblood and arguably the primary characteristic they offer for 119.

Cf. COOMBE, supra note 8, at 56 ("Baudrillard's discussion of the 'brand' suggests that the postmodern commodity/sign operates not primarily to signal the product, but to mobilize connotations of affect."). 120.

See GOLDMAN & PAPSON, supra note 53, at 48-53 (discussing the Benetton advertising campaigns of the 1980s and 1990s).

See id. at 51 ("Through the process of advertising alchemy, Benetton turns these advertising taboos into a sign that represents difference. Their value is constituted by the mere fact of being different: images on the margins of what is socially acceptable have not previously been seen in advertising.").

See DAVID A. AAKER, BUILDING STRONG BRANDS 304 (1996). Borrowing from Young & Rubicam's proprietary Brand Asset Valuator model, Aaker identifies four factors contributing to the strength of a brand: (1) differentiation, which "[m]easures how distinctive the brand is in the marketplace"; (2) relevance, which "[m]easures whether a brand has personal relevance for the respondent"; (3) esteem, which "[m]easures whether a brand is held in high regard and considered the best in its class"; and (4) knowledge, which is "[a] measure of understanding as to what a brand stands for." Id. Aaker and Erich Joachimsthaler write elsewhere that "[o]ne conclusion that emerges from the Young & Rubicam Brand Asset Valuator, based on a structured inventory of more than 13,000 brands in nearly three dozen countries, is that differentiation is the key to a strong brand, more so than esteem, relevance, and knowledge. consumption. This is the lesson of BusinessWeek 123 as much as it is of Baudrillard. 124

Working from the two leading semiotic structurations of the sign, this part elicits from trademark cases and commentary two corresponding legal structurations of the mark. This structural analysis will lay the groundwork for the semiotic analysis of trademark distinctiveness and trademark dilution to follow in later parts. It will also clarify the nature of certain conceptual problems common to both semiotic thought and trademark doctrine.

While semiotic theory has sought primarily to describe the structure of the sign as it is, trademark commentary has sought primarily to prescribe the structure of the trademark as it must be. Traditionally, the law has prescribed that the trademark be triadic in structure. To be sure, this requirement does not spring unambiguously from the Lanham Act's section 45 definition of "trademark":

The term "trademark" includes any word, name, symbol, or device, or any combination thereof . . . used by a person . . . to identify and distinguish his or her goods, including a unique product, from those manufactured or sold by others and to indicate the source of the goods, even if that source is unknown.

In an analysis endorsed by the U.S. Supreme Court in Qualitex Co. v. Jacobsen Products Co., 126 J. Thomas McCarthy reformulates the Act's definition of the mark:

The requirements for qualification of a word or symbol as a trademark can be broken down into three elements: (1) the tangible symbol: a word, name, symbol or device or any combination of these; (2) type of use: actual adoption and use of the symbol as a mark by a manufacturer 123.

See Gerry Khermouch, The Best Global Brands, BUS. WK., Aug. 6, 2001, at 50. 124.

See BAUDRILLARD, CONSUMER, supra note 6 (discussing "the industrial production of differences").

Qualitex Co. v. Jacobsen Prods. Co., 514 U.S. 159, 166 (1995) (citing 1 MCCARTHY, infra, § 3.1 for the proposition that the "'requirements for qualification of a word or symbol as a trademark' are that it be (1) a 'symbol,' (2) 'use [d] . . . as a mark,' (3) 'to identify and distinguish the seller's goods from goods made or sold by others,' but that it not be 'functional'"); see 1 J. or seller of goods or services;

(3) the function: to identify and distinguish the seller's goods from goods made or sold by others.

Here, the triadic structure is clearly revealed. First, the trademark must take the form of a "tangible symbol." This "word, name, symbol or device or any combination thereof" constitutes the trademark's signifier. When a court speaks of the trademark as, in the Third Circuit's phrase, a "signifier of origin" 128 or, in the Seventh Circuit's, a "signifier[ ] of source," 129 they mean to refer specifically to the perceptible form of the mark. Some courts use the more general term "symbol." 130 Second, the trademark must be used in commerce to refer to goods or services. These goods or services constitute the trademark's referent, as when Judge Rya Zobel explained that "a descriptive mark describes a property or ingredient of its referent." 131 Third and finally, the trademark must "identify and distinguish" its referent. Typically, it does so by identifying the referent with a specific source and that source's goodwill. This source and its goodwill constitute the trademark's signified. Thus, in the case of a trademark such as NIKE, the signifier is the word "nike," the signified is the goodwill of Nike, Inc., and the referent is the shoes or other athletic gear to which the "nike" signifier is attached. 132 127.

A & H Sportswear, Inc. v. Victoria's Secret Stores, Inc., 237 F.3d 198, 222 (3d Cir. 2000) ("This is because the classification system's primary purpose is to determine whether the mark is protectable as a trademark in the first place-that is, to determine whether consumers are likely to perceive the mark as a signifier of origin, rather than as a mere identification of the type of product.").

Publ'ns Int'l, Ltd. v. Landoll, Inc., 164 F.3d 337, 343 (7th Cir. 1998) ("[T]rademark and trade dress law do not protect originality; they protect signifiers of source.").

See, e.g., Boston Prof'l Hockey Ass'n, Inc. v. Dallas Cap & Emblem Mfg., 510 F.2d 1004, 1011 (5th Cir. 1975) ("The fact that the symbol covers the entire face of defendant's product does not alter the fact that the trademark symbol is used in connection with the sale of the product."); Warner Bros., Inc. v. Gay Toys, Inc. 724 F.2d 327, 332 (2d Cir. 1983) ("Gay Toys claims that to show secondary meaning a trademark holder must show that consumers are motivated to buy the allegedly infringing goods in question because they believe the goods are sponsored or manufactured by a particular source identified by the trademark symbol in question."). The problem with this usage is that the term "symbol" is more often used to refer to the "the use of the ® registered trademark symbol," Calvin Klein Cosmetics Corp. v. Parfums de Coeur, Ltd., 824 F.2d 665, 667 (8th Cir. 1987), or the "™ symbol," Gift of Learning Found., Inc. v. ("Trademark law has traditionally imposed restrictions on the right to exclude others from using certain 'descriptive' symbols to ensure that the opportunity for all to associate such symbols with their common referents remains unencumbered.").

132. At least two trademark commentators have previously conceived of the trademark in structuralist terms. In his work on trade dress, Tom Bell has outlined a structural model of the trademark that takes into account the distinction between the signifier and the referent within the trademark structure. Specifically, Bell invokes Gottlob Frege's division of sense and reference to criticize the legal protection of the new "virtual trade dress," which "merg [es] Trademark law has developed sophisticated rules to ensure-or at least, to attempt to ensure-that its triadic prescription is followed. For example, the "affixation" and "use in commerce" requirements enforce the linkage between the trademark signifier and its referent. 133 The signifier will not qualify for protection if it its owner does not affix it to a good or service offered in commerce. Assignment doctrine enforces the linkage between the signifier and the goodwill to which it refers. Trademark law will deny protection to trademarks that have been assigned "in gross," separate from the going business that is the source of their underlying goodwill. It will do so because a "trade name or mark is merely a symbol of goodwill; it has no independent significance apart from the goodwill it symbolizes." 134 Finally, licensing doctrine enforces the linkage between the trademark's signified and its referent. If a mark owner engages in "naked licensing," if she fails to exert some control over the quality of the goods with which a licensee is associating her goodwill, then her mark will be deemed abandoned. 135 To maintain the structural integrity of the mark, the law does not merely enforce linkages among the mark's three elements. It also enforces separations Per Mollerup has also developed a highly sophisticated triadic model of the mark. See generally PER MOLLERUP, MARKS OF EXCELLENCE: THE HISTORY AND TAXONOMY OF TRADEMARKS (1999).

The Lanham Act's section 45 definition of "use in commerce" sets out both the affixation and use in commerce requirements that a mark must meet to qualify for federal registration:

The term "use in commerce" means the bona fide use of a mark in the ordinary course of trade, and not made merely to reserve a right in a mark. For purposes of this Chapter, a mark shall be deemed to be used in commerce-(1) on goods when-(A) it is placed in any manner on the goods or their containers or their container or the displays associated therewith or on the tags or labels affixed thereto, or if the nature of the goods makes such placement impracticable, then on documents associated with the goods or their sale, and (B) the goods are sold or transported in commerce, and (2) on services when it is used or displayed in the sale or advertising of services and the services are rendered in commerce . . . . 134. Marshak v. Green, 746 F.2d 927, 929 (2d Cir. 1984). See also 15 U.S.C. § 1060(a) ("A registered mark or a mark for which an application to register has been filed shall be assignable with the good will of the business in which the mark is used, or with that part of the good will of the business connected with the use of and symbolized by the mark.").

135. Gorenstein Enters., Inc. v. Quality Care-USA, Inc., 874 F.2d 431, 435 (7th Cir. 1989). The court stated:

The owner of a trademark has a duty to ensure the consistency of the trademarked good or service. If he does not fulfill this duty, he forfeits the trademark . . . . The purpose of a trademark, after all, is to identify a good or service to the consumer, and identity implies consistency and a correlative duty to make sure that the good or service really is of consistent quality, i.e., really is the same good or service. among them. The mark's elements must be related, but they may not be identical. The law maintains these separations primarily under the rubric of "functionality." It will deny protection to certain features of the referent, specifically those which are "essential to the [referent's] use or purpose" 136 or are otherwise a "competitive necessity," 137 even when those features are perceived by consumers as signifiers of source. 138 In holding that a popular song could not function as a "trademark for itself," 139 the Second Circuit held more fundamentally that a trademark signifier cannot be identical to its referent. It did so for the same reason that the Third Circuit denied protection to the design of a Grecian-style plastic planter. 140 The design was "constitutive of the product itself."

141 Unlike a trademark, it had no "dialectical relationship to the product." 142 The signifier at issue did not refer to something external to itself. The trademark's signifier may consist of certain features of the trademark's referent, but it may not consist of the totality of those features or of those features that make up the essence of the referent, that are "essential to [its] use or purpose," for in both cases, identity would preclude reference.

The trademark, then, is not an irreducible, indivisible thing, but rather a set of relations, specifically, of semiotic relations of reference. It is more wave than particle. Its relational nature creates a host of conceptual problems, to two of which I now turn.

The first problem involves the meaning of the term "trademark." By referring to "New York," one may be referring to either the city or the state, to either the part or the whole. There is a similar ambiguity in both semiotic and trademark talk. In the Saussurean scheme, the term "sign" refers to the relational structure consisting of the signifier and the signified. It does not refer to the signifier alone. On this, the Saussurean tradition is emphatic. To use the term sign to refer only to the signifier, to the perceptible form of the sign, is to give the name of the whole to what is in fact only the part. 143 much a category mistake as to use the term "atom" to refer to the nucleus. 144 More to the point, referring to the signifier (the part) as the sign (the whole) obscures the fact that the signifier is as much constituted and defined by the signified as the signified is by the signifier. Neither precedes the other and neither can exist independently as a sign. The two are mutually interdependent. Because of its commonalities with trademark commentary, Saussure's discussion of this point is worth quoting at length:

Our definition of the linguistic sign poses an important question of terminology. I can call the combination of a concept and a soundimage a sign, but in current usage the term generally designates only a sound-image, a word, for example (arbor, etc.). One tends to forget that arbor is called a sign only because it carries the concept "tree," with the result that the idea of the sensory part implies the idea of the whole.

Ambiguity would disappear if the three notions involved here were designated by three names, each suggesting and opposing the others. I propose to retain the word sign [signe] to designate the whole and to replace concept and sound-image respectively by signified [signifié] and signifier [significant]; the last two terms have the advantage of indicating the opposition that separates them from each other and from the whole of which they are parts. As regards sign, if I am satisfied with it, this is simply because I do not know of any word to replace it, the ordinary language suggesting no other. 145 This ambiguity, synecdochic in nature, appears in trademark talk as well. The difference is that, in technical trademark talk, we persist in using the same term to refer to the whole or to the part. When it is used to describe the whole, the term "trademark" refers to the relational system consisting of the tangible signifier (or set of signifiers), the source or goodwill signified, and their connection to a referent (or more commonly, a set of referents). This is the meaning of the term that the treatise writer William Browne meant to invoke in 1873: "A trade-mark is nothing more nor less than one's commercial signature to his goods; and the mark and the goods the broader sense of a semiotic entity which unites a sign vehicle with a meaning, or it has the narrower sense of a sign vehicle only . . . . In semiotics, the distinction between sign vehicle and sign was introduced in various terminological versions . . . . But because of the ordinary language usage, this distinction has never been strictly observed.

Cf. SAUSSURE (Baskin), supra note 18, at 103. Saussure writes: The two-sided linguistic unit has often been compared with the human person, made up of the body and the soul. The comparison is hardly satisfactory. A better choice would be a chemical compound like water, a combination of hydrogen and oxygen; taken separately, neither element has any of the properties of water. Id bear the same relation to one another as do the positive and negative forces of electricity to each other; and in their opposition they mutually uphold and sustain." 146 In a metaphor that Saussure might have been content to call his own, McCarthy writes more recently that "[g]ood will and its trademark symbol are as inseparable as Siamese Twins who cannot be separated without death to both." 147 Examples of this broader meaning of the term are found in the definition of "trademark" given in section 45 of the Lanham Act and McCarthy's reformulation of that definition. We mean to invoke this broader meaning as well when we speak of "trademark rights" and "trademark infringement" and when we analyze the distinctiveness of a "trademark" by considering the nature of the relation between the trademark's signifier and its referent.

When it is used to describe the part, the term "trademark" refers to the trademark signifier alone, separate from any goodwill or product. We rely on this narrower meaning of the term when we speak, for example, of "trademark y for product x," of the various forms of word, image, or shape that a trademark may take, of the similarities of sound, sight, and meaning among trademarks, or of the literal meaning or functionality of a trademark.

Why is this ambiguity a problem in trademark talk? As an initial matter, the failure to recognize that the trademark is not merely the signifier, but is rather a full-blown sign is often the cause of judicial error. For example, in the recent case of Pro-Football, Inc. v. Harjo, 148 Harjo, a Native American, sought to cancel the registration of the WASHINGTON REDSKINS trademark on the grounds that it was disparaging of Native Americans at the time of its registration. The Trademark Trial and Appeal Board (TTAB) agreed and ordered the cancellation of the mark. 149 On appeal, however, the District Court for the District of Columbia reversed. One ground for reversal was that the TTAB failed to find that the word "redskins" (the signifier) was disparaging when used specifically in connection with sports entertainment services (the referent). 150 The TTAB had committed, in other words, a classic error in trademark thinking (if not also in semiotic thought): It mistook the signifier for the sign. The failure to recognize the interdependence of the signifier with the signified and referent has broader implications relating to the scope of trademark rights. Strictly speaking, the trademark owner has no rights in the signifier alone. Indeed, on its own, in isolation from its signified and referent, the trademark's signifier is not a "trademark" deserving of "trademark rights." Nevertheless, to expand the scope of their property rights, trademark owners have sought to define their property right as an exclusive right to the signifier in itself. The ambiguity of the term "trademark" invites a slippage in the doctrine towards "in gross" rights. This ambiguity may largely account for the increasing liberalization of licensing and assignment doctrine in recent decades to the point now where the trademark owner may license or assign the "trademark" (the signifier), with only token acknowledgment of that signifier's relations to a signified or referent.

The second conceptual problem relates to intrasign reference within the trademark structure. We saw above that in the view of Browne in the nineteenth century and of McCarthy in the twentieth, the trademark's signifier must be understood as merely one element, one relatum in an interdependent relation. Both employed colorful metaphors to make their point. Browne spoke of "positive and negative forces of electricity," McCarthy of inseparable "Siamese Twins." The problem is that while Browne was speaking of the relation between the signifier and its referent (its product), McCarthy is speaking of the relation between the signifier and its signified (its goodwill). What Browne and McCarthy are disagreeing over, perhaps inadvertently, is a question, in semiotic terms, of "mediation." This question is simply stated: To what does the trademark's signifier refer? Does it refer in the first instance (1) to the source of the product, as Judge Richard Posner suggests when he speaks of trademarks as "signifiers of source," 151 or (2) to the product itself, as Judge Edward Becker suggests when he speaks of the trademark as "a symbol according to which one can relate the signifier (the trademark, or perhaps the packaging) to the signified (the product)," 152 in speaking of the trademark's "referent" as "the product or source for which the trademark stands." 153 This question is of more than merely academic concern. The doctrine's inability to settle on an answer has prevented it from clearly defining the nature of the tort of trademark dilution-does "blurring" involve a blurring of the link between the signifier and the source it signifies or between the signifier and the product to which it refers? Of this I will have more to say in Part V. Here, it is appropriate more generally to note that the unanswered question of intrasign reference has seriously complicated the task of defining the concept of trademark "distinctiveness." In the controversial Anti-Monopoly, Inc. v. General Mills Fun Group, Inc. cases, 154 the Ninth Circuit held that the trademark MONOPOLY for the board game was generic because Parker Brothers failed to show that "the primary significance of the term in the minds of the consuming public is not the product but the producer." 155 Parker Brothers failed to show, in other words, that purchasers are motivated to buy the Monopoly game because they like Parker Brothers products rather than simply because they like the game itself. 156 In what the court called a "reductio ad absurdum" 157 of one of the plaintiff's surveys, Parker Brothers introduced survey evidence relating to the TIDE trademark for detergent. The survey evidence showed that "when asked 'Would you buy Tide primarily because you like Proctor and Gamble's products, or primarily because you like Tide detergent?' about 68% [of those surveyed] indicated the latter reason." 158 These results led the court to suspect that "the general public regards 'Tide' as the name of a particular detergent, having particular qualities, rather than as one producer's brand name for the same detergent which is available from a variety of sources," and that "Proctor and Gamble[, the owner of the TIDE mark,] might have cause for alarm." 159 Indeed, trademark owners everywhere had cause for alarm after this decision was handed down. Courts had long neglected to distinguish between marks that are "distinctive of plaintiff's product" 160 product's source," 161 and had, in fact, used such phrases interchangeably. Now, however, the former type of mark risked being found generic.

After the Supreme Court denied certiorari review of the second of the Anti-Monopoly cases, 162 Congress passed the Trademark Clarification Act of 1984 163 (TCA) to override the decision. The TCA amended the Lanham Act to provide that a trademark may identify a "unique product" and to require that "[t]he primary significance of the registered mark to the relevant public rather than purchaser motivation shall be the test for determining" genericness. 164 But the TCA's embrace of the "primary significance" standard has done little to solve the problem at the root of the genericness inquiry and more generally of the concept of distinctiveness. On the contrary, under the language of the TCA strictly construed, a trademark that is more distinctive of its product than its producer, as most trademarks are, would appear still to risk being found generic. 165 Semiotics has long struggled to make sense of a similar problem of intrasign reference in its own study of the triadic sign structuration, and has proposed as a solution to the problem the concept of "mediation." Mediation describes the process by which the signified mediates between the signifier and the referent. 166 In this process, the signified is understood to function as the hinge within the sign. This formulation rings true in trademark terms: The signifier identifies goodwill in order to distinguish the product; it does so by identifying goodwill with the product. To what, then, does the signifier refer within the trademark structure? Notwithstanding the ambiguous language of the Lanham Act's definition of the trademark, the trademark does not identify and distinguish goods; it identifies and distinguishes the goods' source, and the identification of the goods' source identifies and distinguishes in turn the goods themselves.

At the same time that the triadic structuration creates conceptual problems it also creates conceptual opportunities, the most significant of which is considered here. The triadic structuration is generative of the basic syntax of trademark talk, in which we speak of trademark y for product x, as in the trademark "FORD for cars" or the trademark "ACE for hardware, but not for bandages," and, in doing so, assume a third, unmentioned variable, z, the source of the product or the goodwill associated with that source. The triadic structuration underlies the law's basic two-dimensional model of trademark infringement, which conceives of any given trademark as forming a point in a two-dimensional features space consisting of a signifier dimension and a referent dimension. 167 The signifier dimension approximates in a collapsed, one-dimensional continuum what semiotics calls, among other things, a "sign field." 168 For purposes of trademark law, this field comprises the constellation of actual or hypothetical signifiers that because of similarities of "sound, sight, or meaning" surround any given signifier. 169 For example, a series such as NIKE, NIKON, NEIKAY, NIKKEI, NOKAY, NOKIA might form a portion of such a continuum of marks. Similarly, the referent dimension approximates in a collapsed, one-dimensional continuum what trademark law calls a product's "related goods," goods which, because of shared features "may be reasonably thought to originate from the same source." 170 A series such as unleaded gasoline, leaded gasoline, diesel fuel, kerosene might form a portion of the referent dimension. Thus, as in Figure 5, "the trademark BASS for loafers" or "the trademark BASS for ale" each form a point in two-dimensional product features space, the former at (loafers, BASS), the latter at (ale, BASS).

McCarthy's commentary suggests such a framework: In a case where the goods are non-competitive and the marks are not identical, two issues must be resolved: (1) Are the marks themselves confusingly similar? and (2) Are the goods of the respective parties so "related" that an ordinarily prudent purchaser will be likely to be confused as to source, connection or sponsorship? 4 J. THOMAS MCCARTHY, MCCARTHY ON TRADEMARKS AND UNFAIR COMPETITION § 24.27 (4th ed. 2002).

See ROSSI-LANDI, supra note 55, at 162-63. Rossi-Landi explains: When normally developed languages are concerned, it is not usually possible to keep in mind at any given moment, or even in the course of a single piece of research, the position of every word (its "value") with regard to all the other words of that language. This position is therefore to be determined by studying the limited sign field of the word under examination. The word one has decided to examine will therefore exercise the function of a provisional nucleus around which the sign field is arranged.

Every sign field belongs in turn to broader sign fields; and so on until we arrive at the sign totality of the language. Id. (citations omitted); STEPHEN ULLMANN, SEMANTICS 250-61 (1977) (discussing "semantic fields").

Cf. SAUSSURE (Baskin), supra note 18, at 126 ("A particular word is like the center of a constellation; it is the point of convergence of an indefinite number of co-ordinated terms . . . Distance in this features space is a measure of two concepts. First, distance is a measure of difference. The distance between any two points in product features space represents the degree of difference between them. Second, and related, distance is a measure of the likelihood of consumer confusion. The closer two points are in features space, the greater the proportion of consumers in the relevant consumer population who will likely confuse them.

In order to prevent consumer confusion as to the source, z, of a trademark-product combination, trademark law typically grants to the producer of that combination exclusive rights over more than simply a point in features space. Otherwise, a competitor could come very near to that point, as in (stout, BASS), (ale, BOSS) or (stout, BESS), and by confusing some proportion of consumers as to source, unfairly appropriate as to those consumers the goodwill of the BASS ale brand. Trademark law prevents this form of unfair competition by investing a trademark with scope, or in Judge Hand's phrase, a "penumbra." 171 The closer a junior user's signifier-referent combination comes to the signifier-referent combination of a senior user, the greater the proportion of consumers who will confuse the junior's with the senior's use. At some proximity to the senior's use, trademark law declares that too high a proportion of consumers are or will be confused, and establishes a border inside of which no competitor may come. This border, enveloping any given trademark, describes the scope of that trademark's protection and the extent of the producer's property right.

Orthodox trademark doctrine abides even now in the belief that the trademark consists of three interdependent but separate elements: the trademark signifier, which refers to a signified source, which refers in turn to a product referent. Underlying this structuration is a pair of naïve assumptions. The first is that consumers consume things, not signs, actual goods or services, not the meanings of those goods or services. In terms of the triadic structuration, the assumption is that consumers consume material referents, not ideational signifieds. The second assumption is that trademark law merely protects means, not ends. Trademark law only protects signs, the economic value of which is exhausted once the thing is found, while the protection of things themselves is left to patent or perhaps to copyright law.

In the past decades, if not the past century, trademark producers have made a mockery of both these assumptions. They have discovered in the trademark a remarkably productive tool for the commodification of distinction, for the packaging and selling of sign value. The emergence of a sign economy of "semiurg[y]" rather than "metallurg[y]," 172 of "pure image production and marketing," 173 has worked enormous stresses on the triadic edifice of the mark and on the law designed to preserve it. It has ultimately led to the merger of the signified and referent, and the emergence of the dyadic trademark structuration.

At the same time, as the definition of the trademark signifier has become more comprehensive, so as to include "almost anything at all that is capable of carrying meaning," 174 and as producers have more aggressively sought to 172.

See BAUDRILLARD, CRITIQUE, supra note 35, at 185. 173.

DAVID HARVEY, THE CONDITION OF POSTMODERNITY 89 (1989) ("Given the ability to produce images as commodities more or less at will, it becomes feasible for accumulation to proceed at least in part on the basis of pure image production and marketing."); see also FREDERIC JAMESON, POSTMODERNISM, OR THE CULTURAL LOGIC OF LATE CAPITALISM 4-1 (1991) ("[A]esthetic production today has become integrated into commodity production generally: the frantic economic urgency of producing fresh waves of ever more novel-seeming goods . . . at ever greater rates of turnover, now assigns an increasingly essential structural function and position to aesthetic innovation . . . invest their products' characteristics with not only utilitarian but semantic distinction, trademark protection has been extended to the very product features that consumers seek to consume. In such instances, the ends of consumption are protected in the name of its means. Four times in the past decade, the Supreme Court has judged it necessary to intervene in this area, at first boldly, perhaps too boldly, 175 then with ever greater circumspection. 176 In each case, the Court has sought to establish principled grounds on which to distinguish between use value and sign value, between referent and signifier.

In sum, the triadic structuration is being attacked on two fronts, on one front by the granting of protection to trademarks as products themselves, and on the other, by the granting of protection to products (or product features) as trademarks. I turn first to the emergence of the referentless trademark.

The social theorist Mark Poster has observed that "today increasingly meaning is sustained through mechanisms of self-referentiality, and the non-linguistic thing, the referent, fades into obscurity, playing less and less of a role in the delicate process of sustaining cultural meanings." 177 A host of trademark commentators have made a similar observation with respect to trademarks. Judge Alex Kozinski, for one, has noted that trademarks have "begun to leap out of their role as source-identifiers and, in certain instances, have effectively become goods in their own right." 178 In such a situation, the 175.

See id. at 165 (holding 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,' that is, if exclusive use of the feature would put competitors at a significant non-reputationrelated disadvantage." (quoting Inwood Labs., Inc. v. Ives Labs., Inc., 456 U.S. 844, 850, n.10 (1982))); Two Pesos, Inc. v. Taco Cabana, Inc., 505 U.S. 763 (1992) (holding that trade dress may be inherently distinctive).

176. referent, of whose source the consumer is ostensibly being informed by the trademark, is reduced to a nullity. Its absence collapses the trademark's conventional triadic structure by forcing a merger of signified and referent. The trademark's goodwill is commodified and sold as its own product.

In a constellation of early promotional use cases known as the "patch cases," 179 the federal courts first squarely confronted the phenomenon of the dyadic structuration. The 1975 Fifth Circuit case Boston Professional Hockey Ass'n v. Dallas Cap & Emblem 180 is the best known. Its facts are straightforward: Dallas Cap & Emblem was engaged in the unauthorized manufacture and sale of National Hockey League team logos in the form of patches. The Boston Bruins and other teams sued for trademark infringement and won. The case is notable in that it is not a common trinket-sale case, involving the affixation of a team's logo to a shot-glass or bobble head or the like. Rather, the defendant was replicating and selling the trademark in itself, unaffixed, raw. An unwitting Saussurean, the defendant was selling only the signifier-signified combination. The court explains: "The difficulty with this case stems from the fact that a reproduction of the trademark itself is being sold, unattached to any other goods or services. The statutory and case law of trademarks is oriented toward the use of such marks to sell something other than the mark itself." 181 The court was thus confronted with a very simple question: Where is the referent, the product, as to whose source consumers are confused? In answering source identifiers, but also as part of the language and as profitable commodities in their own right"). Rochelle Cooper Dreyfuss was among the first to observe that trademarks had become independent commodities:

[I]deograms that once functioned solely as signals denoting the source, origin, and quality of goods, have become products in their own right, valued as indicators of the status, preferences, and aspirations of those who use them. Some trademarks have worked their way into the English language; others provide bases for vibrant, evocative metaphors. In a sense, trademarks are the emerging lingua franca: with a sufficient command of these terms, one can make oneself understood the world over, and in the process, enjoy the comforts of home. the question, the court resorts to sleight of hand. The thinnest of referents briefly materializes:

Defendant is in the business of manufacturing and marketing emblems for wearing apparel. These emblems are the products, or goods, which defendant sells. When defendant causes plaintiffs' marks to be embroidered upon emblems which it later markets, defendant uses those marks in connection with the sale of goods as surely as if defendant had embroidered the marks upon knit caps. The fact that the symbol covers the entire face of defendant's product does not alter the fact that the trademark symbol is used in connection with the sale of the product. The sports fan in his local sporting goods store purchases defendant's fabric and thread emblems because they are embroidered with the symbols of ice hockey teams. Were defendant to embroider the same fabric with the same thread in other designs, the resulting products would still be emblems for wearing apparel but they would not give trademark identification to the customer.

The actual product that defendant sells, we are told, is not the "trademark symbol," but "fabric and thread emblems." The defendant improperly attempts to sell these emblems, these referents, by imprinting upon them signifiers-"plaintiff's marks"-that associate the emblems with the plaintiff's signified. This ingenious version of the facts is unmistakably in the grand tradition of the cause of action arising in Minorca, to wit, in London. 183 It is a legal fiction that enables the court to maintain the familiar triadic structure of the mark. Yet this structure is unsustainable on Boston Hockey's facts. The defendant's "fabric and thread" is not the product that the consumer seeks, nor is any other material object. The product that is being sold is the plaintiff's signified, its goodwill, its positive affect, commodified in the form of the plaintiff's signifier. If there is a material referent in this structure, then that referent is the consumer himself. He is the commodity-form about which the trademark is designed to convey information. Though the reasoning in Boston Hockey has gained few followers, the outcome in the case is no longer controversial. If trademark law has begun to resemble more and more a copyright-like regime of enumerated exclusive rights, then the plaintiff's exclusive right to make promotional uses of its mark 182.

See is one such right which is well accepted. Either under the broad language of section 43(a) of the Lanham Act or on the common law theory of unjust enrichment, courts have proven quite willing to protect the promotional extension of the trademark far beyond its traditional product category, indeed, to the zero degree of no product category at all. Even so, the problem of the referent in the promotional goods context has by no means been resolved. It remains an essentially structural rather than doctrinal problem. It threatens to reappear any time a careful judge looks too closely into the doctrine before her. This very scenario played out when, in the remarkable case of United States v. Giles, 184 Boston Hockey's legal fiction could not survive the scrutiny of a criminal appeal. Giles, the owner of store called "Fabulous Fakes," was charged with trafficking in counterfeit goods in violation of 18 U.S.C. § 2320, the federal criminal trademark infringement statute. In a sting operation, Giles sold an FBI informant one thousand counterfeit "patch sets" bearing the DOONEY & BOURKE trademark:

A "patch set" consists of a leather patch and a gold medallion, which both bear the Dooney & Bourke logo, and a leather strap which is used to attach the medallion to a purse or piece of luggage . . . . Once the patch set is applied to a generic purse or piece of luggage, the bag will appear to have been made by Dooney & Bourke.

In reviewing the promotional goods cases, Judge Stephanie Seymour candidly acknowledged the Boston Hockey problem before her:

The issue is made complex because a trademark is easily conceived of as an abstract symbol or design which is carried by the tangible goods at issue. In reality, this mark itself often consists of a tangible item such as a label, a patch, or a medallion (collectively referred to in the remainder of this opinion as "labels"). The question comes down to whether this label can be considered a good if it is disconnected from its host good and sold for a profit.

In the terminology of semiotics, the judge is troubled by the "materiality of the signifier" and the danger in particular of mistaking the tangible signifier for its equally tangible referent. Unpersuaded by the government's logic that anything sold for money is a good, Judge Seymour dismissed the indictment: "The 'goods' at issue in this case are the purses and handbags to which the patch sets could be applied. The patch sets are not goods, but labels." 187 Though the result in the case may appear absurd, the judge's reading of the statute is altogether sound-and semiotically quite sophisticated. At least in criminal cases in the Tenth Circuit, the trademark remains a three-legged stool.

The likes of Boston Hockey and Giles may appear to be special cases. Unlike the patch cases, most promotional uses of trademarks involve a tangible and ostensibly useful referent of some sort, if only in the form of a paperweight or a sweatshirt. Such referents would appear to sustain the trademark's traditional triadic structure. Yet the differences among the "fabric and thread emblem" on which is imprinted the BRUINS logo, the t-shirt on which is imprinted ABERCROMBIE, and the watch on which BATMAN tells the time are differences of degree, not kind. In each case, the utilitarian component of the referent is merely "an alibi for the function of invidious distinction." 188 More or less useful, the referent is a more or less persuasive legal fiction.

If, by allowing the merger of signified and referent within the trademark structure, trademark law has begun to claim for itself subject matter traditionally reserved to copyright protection, then, by allowing the merger of signifier and referent, trademark law risks claiming for itself subject matter traditionally reserved to patent protection. As above, so here, when the triadic structuration collapses, trademark law begins to assume the responsibilities of other forms of intellectual property protection. a.

Functionality and the "Semiotic Threshold" "Functionality" doctrine has long existed to prevent the merger of signifier and referent. In the typical functionality case, the plaintiff asserts that a utilitarian or aesthetic feature of its product is sufficient in itself to denote source and should therefore be protected as a trademark signifier. The defendant responds that even if consumers perceive the feature at issue to be source-denotative, protection should be denied because the feature at issue is "functional." "Functional" is a term of art in trademark law: A product feature may be "de facto functional" (functional under the standard dictionary meaning), but not "de jure functional" (functional as a matter of trademark law and thus unprotectable). 189 For example, certain distinctive design features of a ZIPPO lighter serve a utilitarian purpose, but the TTAB determined that the exclusive use of those design characteristics would not restrict the ability of other firms to "compete effectively," and thus that those characteristics were nonfunctional as a matter of trademark law. 190 The TTAB has ruled, in contrast, that the fluted design of a beverage can's sidewalls is "one, if not the best, way" of strengthening a can's sidewalls and is thus functional. 191 Aesthetic features, such as the black color of outboard motors 192 and the baroque design of dinnerware, 193 have also been found to be functional on the ground that granting exclusive rights in such features would put competitors at a non-reputation-related competitive disadvantage.

The problem of functionality in trademark law is closely related to the problem of the "semiotic threshold" in semiotic thought. At this threshold between the nonsemiotic and semiotic universes, objects cease to be purely functional (or otherwise semantically inert) and begin to signify, to represent. Roland Barthes has observed that, by an "inevitable" process of "semantization," even the most mundanely functional objects may cross this threshold and become "pervaded with meaning. . . . [A]s soon as there is a society, every usage is converted into a sign of itself." 194 The semiotic threshold, in other words, is very low. This is especially true, Barthes argues, of mass-produced commodities, which are "unavoidably realizations of a model, the speech [parole] of a language [langue], the substances of a significant form. To rediscover a non-signifying object, one would have to imagine a utensil absolutely improvised and with no similarity to an existing model." 195 Taken seriously, Barthes' observation is of great relevance to trademark thinking, particularly as to the merger of signifier and referent. As Barthes explains, an individual instance of a mass-produced commodity is akin to an instance of parole. In Peircean terms, it is a token of a type (the word "the" appears many times on this page, but each of these tokens of "the" invoke the term's one type 196 ). As such, the individual instance signifies the model; the Peirce explains the distinction by reference to the word "the": A common mode of estimating the amount of matter in a MS. or printed book is to count the number of words. There will ordinarily be about twenty thes on a page, and of course they count as twenty words. In another sense of the word "word," however, there is but one word "the" in the English language; and it is impossible that this word should lie visibly on a page or be heard in any voice, for the reason that it is not a Single thing token becomes a signifier of a general type. Trademark law has drawn upon this aspect of mass-produced commodities to develop the doctrine of "anonymous source." Even if consumers cannot identify precisely to what factory or company a trademark refers, still, consumers will assume that a specific instance of a trademark is referring to the same source as are other instances of the trademark. This "anonymous source" will receive trademark protection. The tendency of consumers to construct an anonymous source, a model, from which is produced specific instances of a commodity, is not limited to conventional word or image marks, however. In the case of massproduced commodities, consumers will also construct such a source for product features. 197 Just as language speakers will, upon exposure to a handful of tokens, quickly construct their type, so consumers will, upon exposure to a handful of instances of a commodity, quickly construct a single source for that commodity, and associate that commodity's features with that source. Barthes' observation that the semiotic threshold is exceedingly low may be transposed to the commodity system: The threshold for product features to develop "secondary meaning," the "source-distinctiveness threshold," is also very low. The law is left continually to struggle to limit protection to features that serve some aesthetic or utilitarian purpose and that have at the same time developed, as tokens, a source-denotative meaning, a type.

In the 1995 case of Qualitex v. Jacobson Products, 198 which involved the protectability of the green-gold color of Qualitex's dry-cleaning press pads, the Supreme Court plumbed the depths of the semiotic and sourcedistinctiveness threshold and refused to find a bottom. The Lanham Act's section 45 definition of the trademark inventories various formats that the mark may, but need not, take: "[T]he term 'trademark' includes any word, name, symbol, or device, or any combination thereof." 199 In Qualitex, truly or Single event. It does not exist; it only determines things that do exist. Such a definitely significant Form, I propose to term a Type. . . . In order that a Type may be used, it has to be embodied in a Token which shall be a sign of the Type, and thereby of the object the Type signifies. I propose to call such a Token of a Type an Instance of the Type. Thus, there may be twenty Instances of the Type "the" on a page. the Chakrabarty 200 of American trademark law, the Court interpreted this language to define "the universe of things that can qualify as a trademark" in "the broadest of terms," so broad, in fact, as to include "almost anything at all that is capable of carrying meaning." 201 The Court noted that the Patent and Trademark Office had allowed the registration of "a particular shape (of a Coca-Cola bottle), a particular sound (of NBC's three chimes), and even a particular scent (of plumeria blossoms on sewing thread)," 202 because each of these formats proved "capable of satisfying the more important part of the statutory definition of a trademark," 203 that the signifier "identify and distinguish." 204 As Justice Breyer explained, "[i]t is the source-distinguishing ability of a mark-not its ontological status as color, shape, fragrance, word, or sign" 205 that qualifies it for protection.

As the Court made clear, the Qualitex definition of the signifier is a functional rather than an ontological definition. 206 To determine if a given perceptible form qualifies as a signifier for purposes of trademark law, the Qualitex approach does not ask what that form is, but rather what it does: Does it signify? That is, does it convey a signified? In this sense, the Qualitex definition of the signifier is a relational definition: A trademark signifier is any perceptible phenomenon to which the process of signification connects a trademark signified. If there is an ontological requirement, it is imposed on the signified. The signified cannot involve just any meaning; it must involve a source-denotative meaning.

At the same time that it established an exceedingly inclusive standard for what product features may qualify as a trademark signifier, the Qualitex Court formulated, perhaps unknowingly, an exceedingly narrow exception to that standard directed at product features that are de jure functional. It did so, arguably in dicta, in its two-part definition of de jure functionality: The first part of this definition of functionality, from Inwood, is an ontological definition. A product feature is intrinsically functional if it is "essential to the use or purpose of the article" or otherwise "affects the cost or quality of the article." The second part of the definition is a functional, relational definition. To determine if "exclusive use of the feature would put competitors at a significant non-reputation-related disadvantage," courts were logically required to consider the "availability of equally satisfactory alternatives." 208 Here, the term "functional" was little more than a shorthand label for the outcome of a balancing test, one in which the "anticompetitive consequences" of exclusive rights outweigh "the public and private interest" in protecting distinctive designs. 209 In the wake of Qualitex, the circuits tended to use the second, relational definition, 210 which constituted a far narrower exception to protectability than the first, intrinsic definition. The functional trumped the ontological.

To the extent that it represented an enormous expansion in the subject matter of trademark protection, Qualitex opened a Pandora's box, one which the Court attempted to close six years later in TrafFix Devices v. Marketing Displays, Inc. 211 The TrafFix Devices Court found that a dualspring configuration at the base of traffic signs, designed to keep the signs upright in windy conditions, was de jure functional and thus unprotectable under trademark law. In making this determination, the Court restated its definition of functionality, calling Qualitex's second, relational definition "incorrect as a comprehensive definition." 212 The TrafFix Devices Court explained that "a feature is also functional when it is essential to the use or purpose of the device or when it affects the cost or quality of the device. The Qualitex decision did not purport to displace this traditional rule." 213 207.

Id Here, the Court signaled that that the ontological definition of functionality should henceforth trump the functional definition. In doing so, the Court betrayed its unease not simply with the economic definition of functionality, but also by implication with the functional definition of the signifier it had established six years earlier. In this sense, TrafFix Devices is a subtle, but unmistakable repudiation of Qualitex. As a general matter, producers seek to exploit the semiotic potential inherent in even the most coldly utilitarian objects. They do so by pushing the source-distinctiveness threshold as close to the semiotic threshold as possible, in the hope that any similar feature across a product line will be interpreted as a source-denotative feature, as a mass-produced token of a more general type. TrafFix Devices has established an ontological definition of functionality to push back.

The retrenchment begun in TrafFix Devices remains incomplete, however. At the same time that the TrafFix Devices Court endorsed an ontological definition of utilitarian functionality, it endorsed a functional definition of aesthetic functionality:

It is proper to inquire into a "significant non-reputation-related disadvantage" in cases of esthetic functionality, the question involved in Qualitex. Where the design is functional under the Inwood formulation there is no need to proceed further to consider if there is a competitive necessity for the feature. In Qualitex, by contrast, esthetic functionality was the central question, there having been no indication that the green-gold color of the laundry press pad had any bearing on the use or purpose of the product or its cost or quality. 214 There appear to be two reasons why the TrafFix Devices Court retained a functional definition of aesthetic functionality, neither of which goes to the merits of the issue. The first reason is that the precise terms of the Inwood definition of functionality are rarely appropriate to an aesthetic functionality inquiry. It seldom makes sense to ask whether a color, decoration, or other aesthetic feature is "essential to the use or purpose of the article" or otherwise minimizes the article's "cost or quality." The baroque design of dinnerware, for example, is not essential to its use or purpose and, if anything, increases its cost of manufacture. 215 The second reason is that in Qualitex, the Court made a mistake. When it formulated Qualitex's two-part definition of functionality, it failed to anticipate the lowering of the functionality bar that would result.

See TrafFix Devices proceeded to raise the bar anew, but perhaps in an effort to save appearances, rather than overrule Qualitex outright, the TrafFix Devices Court distinguished Qualitex on the ground that that case dealt with an issue of aesthetic functionality. At least one commentator has noted, however, that Qualitex had little, if anything, to do with aesthetic functionality. 216 Ultimately, there is no logical reason why an ontological definition of functionality should not also be established for purposes of determining aesthetic functionality. The "ontological" standard is simply stated: Is the feature at issue dictated primarily by aesthetic considerations? Is it intrinsically aesthetically functional? To be sure, this standard would be far more restrictive than TrafFix Devices's current "competitive necessity" standard for aesthetic functionality, in that it would ask not whether the feature at issue is competitively necessary, but rather simply whether it is competitively important. But so, arguably, was the TrafFix Devices standard of utilitarian functionality when it was introduced. It is unclear when, if ever, the Court will find an opportunity to finish what it began in TrafFix Devices, but until it does, the subject matter of trademark protection will continue to be exceedingly overbroad. Aesthetically important features of the referent will continue to be protected in the name of the signifier.

C. The Floating Signifier and the Hypermark Saussure himself would not likely have accepted the proposition, but Saussurean semiotics has since his time explored the possibility that while signification cannot obtain without value, value can nevertheless obtain without signification. From this follows the radical hypothesis that a signifier can be articulated, can achieve form, without being connected to any particular signified-you can cut out one side of the sheet of paper without cutting out the other. Such an "empty" or "floating signifier" may refer to, or at least imply, a signified, but that signified is so indefinite or contested as to constitute an "empty category." 217 In such a situation, a sign "only means that it means." 218 216.

1 MCCARTHY, supra note 126, § 7:80 ("[A]esthetic functionality was not the question in any way in Qualitex, let alone the 'central question.'").

217. JOHN LECHTE, FIFTY KEY CONTEMPORARY THINKERS: FROM STRUCTURALISM TO POSTMODERNITY 64 (1994). The concept of the floating signifier has its origins in the postwar structuralist anthropology of Claude Lévi-Strauss, see id. at 73, the criticism of Roland Barthes, see, e.g., ROLAND BARTHES, IMAGE-MUSIC-TEXT 39 (Stephen Heath trans., Noonday Press ed., 1988) (referring to "a floating chain of signifieds"), and the psychoanalysis of Jacques Lacan, see JACQUES LACAN, ÉCRITS 154 (Alan Sheridan trans. 1977) (discussing the "incessant sliding of the signified under the signifier"). Iterated by so many in the decades since, the term itself has become a leading example of the phenomenon it purports to label.

218 The phenomenon of the floating signifier is most readily identifiable in the context of nonrepresentational art and modernist literary texts (or indeed in ideological categories such as race, 219 democracy, 220 "1968," 221 or "postmodernity"). It is also increasingly identifiable in modern visual culture, particularly in the rapid succession of provocative, obscurely meaningful images that characterizes music video, 222 and in modern celebrity culture. 223 As Jean Baudrillard has written, the increasingly common condition of Disney-like or Las Vegas-like hyperreality, in which reality itself is constructed of simulations of imagined realities that themselves never existed, is especially conducive to the "floatation" of the signifier and the liberation of value from signification, of difference from identity:

Referential value is annihilated, giving the structural play of value the upper hand. The structural dimension becomes autonomous by excluding the referential dimension, and is instituted upon the death of reference . . . . The emancipation of the sign: remove this archaic obligation to designate something and it finally becomes free, indifferent and totally indeterminate, in the structural or combinatory play which succeeds the previous rule of determinate equivalence. . . . The floatation of money and signs, the floatation of needs and ends of proand signified to semes: these units of meaning become meaningful only insofar as they are understood to mean at all"). 219. duction, the floatation of labor itself. . . . the real has died of the shock of value acquiring this fantastic autonomy.

In a condition of hyperreality, in other words, differences are not built upon designation, upon the equivalence of signifier and signified. There is only distinctiveness from, not of. This involves more than simply the "bracketing of the referent." It involves the bracketing of reference altogether, of any intrasign relation among subsign elements.

The dyadic structuration stops short of the condition Baudrillard describes. It still insists on reference, on source, however "anonymous." But in recent times, there has emerged a truly radical structuration of the mark, what might be termed the monadic structuration. In such a structuration, the trademark signifier has broken free from its moorings in a signified or referent. It signifies still, but signifies nothing. Such hypermarks-and Times Square is filled with them-are not designations of source, but commodified simulations of such designations. Mass-produced, nonrepresentational canvases of great value, they invest the products to which they are affixed with pure unarticulated distinctiveness-distinctiveness of nothing, distinctiveness from everything. Their sign value is the source of their economic value. They represent the total collapse-the "implosion," Baudrillard might say 225 -of the triadic structuration. Of this, I will have more to say below.

This part first reviews the law's traditional theory of distinctiveness and explains its flaws. It then considers the synchronic relation between source distinctiveness and differential distinctiveness.

Under current doctrine, to fall within the subject matter of trademark protection, a trademark must be found to be inherently distinctive or to possess acquired distinctiveness. An inherently distinctive trademark is one whose signifier cannot reasonably be understood to be descriptive or decorative of the product to which it is affixed-for example, APPLE computers, CLOROX bleach, MARS candy bars. As a matter of "consumer predisposition" 226 or simple marketplace literacy, consumers will "automatically" 227 recognize such signifiers as designations of source. Not all marks are inherently distinctive of source, however. Many marks consist of signifiers that may be understood to be descriptive or decorative of the products to which they are affixed-for example, SUPER BLEND on a can of motor oil, BEST BUY on an electronics catalogue, three stripes on tennis shoes. Such marks will qualify for trademark protection only if their producers can show that they have "acquired" through use in the marketplace and advertising a "secondary meaning" as a designation of source.

Through the course of the last century, the law developed a taxonomy of trademarks more refined than this simple inherent/noninherent dichotomy. Judge Friendly authoritatively set forth this taxonomy in the 1976 case of Abercrombie & Fitch v. Hunting World: 228 "Arrayed in an ascending order which roughly reflects their eligibility to trademark status and the degree of protection accorded, these [four different categories of terms with respect to trademark protection] are (1) generic, (2) descriptive, (3) suggestive, and (4) arbitrary or fanciful." 229 This is essentially a hierarchy of figurativeness. It ranks marks according to the degree to which their signifiers are, in semiotic terms, "motivated" 230 by their referents. The more motivated the signifier, the less intrinsically distinctive of source it is. Thus, in generic marks such as THE COMPUTER STORE, 231 LIGHT BEER, 232 or LITE BEER, 233 the signifiers are so highly motivated by their referents that there would be little if any "consumer predisposition" to recognize the signifier as a designation of source. Descriptive marks, such as AMERICAN AIRLINES or PERRIER are slightly less motivated by their referents. If their producers can show that they have acquired a secondary meaning, that is, a nonliteral, figurative meaning, as a designation of source, then they will qualify for protection. 234 51:3 Beebe Beebe Handcrafted.doc (2/17/2004 9:48 AM)

Suggestive, arbitrary, and fanciful marks are held to be inherently distinctive of source. Suggestive marks, such as ATLAS for moving services or ROACH MOTEL for insect traps, are textbook metaphors and are described as such by the doctrine. They require "imagination, thought and perception to reach a conclusion as to the nature of the goods." 235 In arbitrary marks such as APPLE for computers or SHELL for gasoline, the figurative relation is not so much metaphorical as it is catachrestic. The relation involves a signifier whose ordinary meaning could not reasonably be understood to describe or suggest any characteristic of its referent. This is akin to Saussure's usage of the term "arbitrary." 236 Finally, in a fanciful trademark, the signifier takes the form of a neologism that conveys no definite semantic content. Strictly, speaking, FLEXITIZED, 237 BASSER, 238 and STEEM 239 are neologisms, as is BREADSPRED 240 for jam, but these terms are not fanciful. They are rather suggestive. In contrast, EXXON, XEROX, CLOROX, and ACELA are both neologistic and fanciful.

But for all of the nuances of the Abercrombie hierarchy, we must not lose sight of the simple dichotomy underlying the hierarchy when it is used to determine the subject matter, rather than the scope, of trademark protection. For purposes of determining whether or not a trademark is eligible for protection, we should not be concerned with the extent to which a trademark is distinctive of source. In the eligibility inquiry, there are only two important categories of marks: (1) marks which are distinctive of source, either inherently or by acquisition of secondary meaning, and which are thus eligible for protection, and (2) marks which are not distinctive of source and which are thus not eligible for protection. It is not useful to speak, as the Abercrombie opinion does, of one mark as being "more eligible" for protection than another, of, say, the fanciful mark IORMYX as being more eligible for protection that the descriptive mark NFL. This is akin to saying that one mark is "more of a trademark" than another, that ACURA is more of a trademark than FORD. The distinction of eligibility is a binary distinction.

The Abercrombie hierarchy properly recognizes that the scope inquiry involves distinctions of degree. But it instructs courts to grant a broader scope of protection to marks which are inherently distinctive of source than to those whose distinctiveness of source is merely acquired. In other words, "[if a descriptive] mark has achieved consumer recognition, it is not denied the protection of the trademark law, but it commands a lesser degree of protection than arbitrary, fanciful marks." 241 On this reasoning, arbitrary marks such as KING or SWAN should in all cases enjoy a greater scope of protection than descriptive marks with secondary meaning such as, say, COCA-COLA, IBM, GENERAL ELECTRIC, or MCDONALD'S.

There are a number of problems with this formulation. Most fundamentally, the degree to which a mark is distinctive of source, either inherently or by acquisition, is simply not relevant to the infringement inquiry. Trademark law traditionally assumes that the more "distinctive" a mark is, the more likely it is that consumers will be aware of it and thus the more likely they are to mistake other marks for it. This is an argument from exposure: the more exposed a mark is in the marketplace, the more likely it is to be confused with other marks. 242 Especially "distinctive" marks shine especially brightly in features space, with the result that the light of nearby marks will be mistaken for their brighter neighbor. But this form of distinctiveness is not distinctiveness of source. It is rather a mark's distinctiveness from other marks, its salience, that makes consumers aware of it and that will affect whether consumers are likely to confuse it with a junior, similar mark. We grant a broader scope of protection to more "distinctive" marks in an effort to preserve a greater expanse of difference (in features space) between them and other marks. Our assumption in doing so is that any junior mark which comes within that expanse of difference may be confused with the senior mark. This explains the perhaps counterintuitive principle in trademark law that the stronger mark succeeds to greater protection. In the case of a senior mark that is highly distinctive as against all other marks in the marketplace, consumer confusion may result from the appearance of a junior mark that is not comparatively as distinctive from the senior mark. Furthermore, Abercrombie ignores the importance of a mark's fame, its notoriety in the marketplace. In its formulation of a hierarchy that "reflects [a mark's] eligibility to trademark status and the degree of protection accorded," 243 Abercrombie converts what is properly a conclusive presumption as to the existence of distinctiveness of source into a conclusive presumption as to the extent of distinctiveness from other marks. To be sure, a mark may be inherently distinctive from other marks, perhaps because of its especially clever suggestiveness or especially appealing sound, but the scope of its protection will be quite small indeed unless it is used in commerce and advertised, unless it acquires further distinctiveness. In the scope inquiry as opposed to the eligibility inquiry, there is no sound without listeners-mere "use in commerce," while sufficient for purposes of registering an inherently distinctive mark, will do little good for the plaintiff seeking to defend that mark against infringement. The binary distinction of kind, between marks which are and are not inherently distinctive, is properly understood as merely the first in a continuum of distinctions of degree, along which are arrayed marks of more or less fame.

Looked at another way, if a trademark becomes more distinctive over time from other marks, as many do, this cannot be because the mark has become more inherently distinctive. This is not possible-though in the case of genericide, it can certainly become less inherently distinctive. Rather, acquired distinctiveness from other marks, not inherent distinctiveness of source or from other marks, is what expands scope; a mark will gain in scope in features space what it has gained in fame in the marketplace. For purposes of determining the scope of protection, inherent distinctiveness is inevitably contingent on acquired distinctiveness. The essentially empirical evaluation of a mark's actual acquired distinctiveness in the marketplace logically incorporates the essentially formal evaluation of its inherent distinctiveness. 244 In sum, Abercrombie is entirely extraneous to the infringement inquiry.

One reason why trademark law has traditionally granted broader protection to inherently distinctive marks is to provide an incentive to trademark producers to generate new words rather than appropriate from the language preexisting words for commercial purposes. On this reasoning, we give broader protection to ALTRIA than ALTAR in an effort to encourage a producer to choose the former. Yet, if the producer is producing anything other than religious paraphernalia, it is likely that both words would qualify as inherently distinctive-the former would be fanciful, the latter arbitrary-and the incentive to choose the former would be lost. This scheme would make more sense if its grant of broader protection were limited only to fanciful marks.

In any case, even in situations in which the plaintiff's mark is descriptive with secondary meaning, the empirical inquiry into the trademark's actual distinctiveness from other marks in the marketplace tends now to trump the old formalisms, however well intentioned, laid down by Abercrombie. 245

The failures of the Abercrombie scheme recommend the adoption of a more logical framework. To be sure, a trademark may possess inherent distinctiveness and/or acquired distinctiveness. But this tells only part of the story, and the less important part. It ultimately only tells us how the trademark has come to be distinctive. It does not tell us what kind of distinctiveness the trademark has come to possess. Indeed, a trademark may possess source distinctiveness either inherently or through acquisition. It may also possess differential distinctiveness either inherently or through acquisition. But in any case, a trademark will not be eligible for anti-infringement or antidilution protection if it does not possess source distinctiveness, be it inherent or acquired, 246 and the scope of its protection will be determined by the extent of its differential distinctiveness, again, be it inherent or acquired. While the formal relation between inherent distinctiveness and acquired distinctiveness is notoriously obscure, 247 the formal relation between source distinctiveness and differential distinctiveness is straightforward: A mark's differential distinctiveness from other trademarks facilitates its distinctiveness 245.

See RESTATEMENT (THIRD) OF UNFAIR COMPETITION § 21 cmt. i (1995) (noting that the Abercrombie classification is not "conclusive of 'strength,' however, since the issue ultimately depends on the degree to which the designation is associated by prospective purchasers with a particular source"); see, e.g., Oxford Indus., Inc. v. JBJ Fabrics, Inc., 6 U.S.P.Q.2d (BNA) 1756, 1760 (S.D.N.Y. 1988) (explaining that while conceptually strong, JBJ is commercially weak because it has "little actual consumer recognition value"); Frank Brunckhorst Co. v. G. Heileman Brewing Co., 875 F. Supp. 966, 976 (E.D.N.Y. 1994) ("A finding that a mark is inherently distinctive does not guarantee a determination that the mark is a strong one, since inherent distinctiveness does not guarantee distinctiveness in the marketplace (else, where would the public relations business be?).").

There is one exception to this statement. In the Second Circuit, antidilution protection is denied to non-inherently distinctive marks. See infra notes 332-338 and accompanying text.

Consider the paradox of Two Pesos, Inc. v. Taco Cabana, Inc., 505 U.S. 763 (1992). In the trial court, the jury found the trade dress at issue to be inherently distinctive but to possess no secondary meaning as a designation of source. Id. at 766. The jury would have done better to determine, first, whether the trade dress was distinctive of source, and second, to what extent it was distinctive from other trade dresses used by competing restaurants. There is no identity without difference, no source distinctiveness without differential distinctiveness. The semiotic distinction between the relation of signification and the relation of value, and the corresponding legal distinction between source distinctiveness and differential distinctiveness, is recapitulated in the difference between the prohibition against trademark infringement and the prohibition against trademark dilution. The former prohibition is a prohibition against interference in the intramark relation of signification between a trademark's signifier and its signified. Notwithstanding its name, trademark infringement is not infringement of a "trademark," if by trademark we mean simply the trademark's perceptible form, its signifier. Rather, trademark infringement is a trespass on goodwill, one which is accomplished by means of a "trademark," that is, by means of a confusingly similar signifier (and referent). The engrafting onto trademark doctrine of the syntax of copyright law has long confused this matter. The mere "reproduction, counterfeit, copy, or colorable imitation" of a trademark's signifier is not in itself trademark infringement, 248 nor, in fact, is the mere creation of confusion, mistake, or even deception. These are highly probative tests of whether trademark infringement has occurred, but they are not bases for relief. Copyright law prohibits the infringement of the signifier. Trademark law, in contrast, prohibits the infringement of the signified. Anti-infringement protection ultimately seeks to protect exclusive rights in the idea, not the expression.

To determine whether a defendant has infringed a plaintiff's trademark, a court should thus proceed in two steps. First, the court should determine whether the plaintiff's trademark is distinctive of source and thus eligible for trademark protection. Underlying this inquiry is the assumption that if the plaintiff's signifier-referent combination is not itself distinctive of the plaintiff's signified, then no similar signifier-referent combination will also be distinctive of, and thus trespass upon, that signified. The eligibility determination is not difficult to make, and traditional tools are sufficient. An eligible mark is either inherently source distinctive or acquires its source distinctiveness. Either form of distinctiveness will do; neither is privileged. Having determined that the plaintiff's signifier-referent combination is itself distinctive of the plaintiff's signified, the court should then determine whether the defendant's signifier-referent combination is sufficiently similar to the plaintiff's as also to be distinctive of the plaintiff's signified. Here, the court should consider not the source distinctiveness of plaintiff's mark, but its differential distinctiveness, the distance between it and the nearest, most similar marks, other than the defendant's, in features space. Consumer confusion surveys, as opposed to secondary meaning surveys, test for precisely this form of distinctiveness. Such surveys are essentially tests of comparative similarity. In the absence of reliable survey evidence, the most important factor in estimating the differential distinctiveness of the plaintiff's signifier is its acquired distinctiveness, its fame, renown, notoriety, in short, its salience. As explained above, a mark's actual differential distinctiveness in the marketplace logically incorporates its inherent differential distinctiveness.

While trademark infringement involves the infringement of source distinctiveness, trademark dilution involves the dilution of differential distinctiveness. As I will explain in more detail in Part V, dilution occurs when, because two signifiers are similar, they lessen each other's differential distinctiveness. (We conventionally say that the junior signifier dilutes the distinctiveness of the senior signifier, though, strictly speaking, dilution occurs as to both signifiers; they are engaged in a zero-sum struggle.) In the typical dilution situation, the plaintiff's and the defendant's signifiers are very near, if not identical, to each other on the signifier dimension, but because their referents are sufficiently different, consumers are not confused as to source, with the result that no infringement action will lie. In this sense, trademark dilution constitutes not a trespass on the plaintiff's signified, but rather a kind of nontrespassory nuisance as to the plaintiff's signifier. The action for trademark dilution is designed to prevent such nuisances and, in doing so, to preserve the differential distinctiveness of the plaintiff's signifier, regardless of to what referent it is affixed. The prohibition against dilution is thus a prohibition against interference in intermark relations of value between the plaintiff's signifier and all other signifiers in the trademark system. If protection from trademark infringement prohibits synonyms (two different signifiers pointing to the same signified), protection from trademark dilution prohibits homonyms (two closely similar signifiers pointing each to its own signified). Antidilution protection ultimately seeks to protect exclusive rights in the expression, not the idea.

The test for trademark dilution remains highly unstable as does the definition of dilution set forth in the case law. The proposition that dilution involves the relation of value, whereas infringement involves the relation of signification, will only take us so far. Before engaging in a more detailed semiotic analysis of the concept of dilution, we must first review the concept's historical origins, which account for many of its idiosyncrasies. This is the subject of the next part.

The evolution of American trademark law in the twentieth century is characterized by a shift in emphasis. At the beginning of the century, the overriding purpose of the law was to protect distinctiveness of source. The stated rationale for this protection was three-fold: (1) the protection of property, (2) the promotion of competition, and (3) the prevention of fraud upon the individual merchant as well as upon the public. 249 By the end of the century, source distinctiveness lost its status as the cynosure of the law. Now, the overriding purpose of the law is to protect distinctiveness from other marks. The rationale for anti-infringement protection is currently articulated in the terminology of economics and is said to involve the minimization of search costs and the promotion of consistent levels of product quality. 250 Yet these are largely alibis, not so much rationales as rationalizations. The law's true desideratum is now the promotion of semiotic sign value.

The fall of source distinctiveness is an oft-told story. 251 Indeed, it was first told by Frank Schechter in the 1920s, 252 and most modern accounts, including this one, draw from his. 253 The story begins in the second half of the nineteenth century. During this time, in order to qualify for protection, the trademark was required to indicate the physical, "personal" 254 source of the goods to which it was affixed. This rigid "strict source theory" reflected the earlymodern origins of the trademark. In pre-industrial Europe, the guild systems commonly required the use of individual "production marks" not simply to control competition, but also to fix responsibility in the case of poor craftsmanship. 255 The trademark was not so much an asset as a liability. It was a regulatory device imposed on, rather than adopted by, the individual producer. That the mark could somehow be licensed or franchised to others, or affixed to goods made under contract, was inconceivable. The mark was no more alienable than the craftsman's own individual skill.

Through the course of the first quarter of the twentieth century, the rise of mass consumption and modern advertising 256 and the consolidation of a national market 257 rendered the strict source theory untenable. Trademark doctrine could no longer assume that consumers of mass-produced brand name goods knew or cared to know the actual source of the brand they were buying. As the Seventh Circuit put it at the turn of the century, "[w]e may safely take it for granted that not one in a thousand knowing of or desiring to purchase 'Baker's Cocoa' . . . know of Walter Baker & Co., Limited." 258 In the 1910s and '20s, the "anonymous source theory" was developed to meet a felt need. According to this theory, in order to qualify for protection, a trademark need only designate a single, though possibly anonymous, source. 259 This was essentially a theory of goodwill. The trademark was understood to function and was protected as a guarantee of quality.

This shift from the strict source theory to the anonymous source theory proved to be of epochal significance. First, and most importantly, it inverted the traditional hierarchy of trademark doctrine, which viewed goodwill as prior to the trademark. Nineteenth-and early-twentieth-century trademark thinking assumed that the trademark "merely" gave a name to a preexisting 256.

See 259. See Schechter, supra note 9, at 816. In Schechter's view, for example, the trademark indicates, not that the article in question comes from a definite or particular source, the characteristics of which or the personalities connected with which are specifically known to the consumer, but merely that the goods in connection with which it is used emanate from the same-possibly anonymous-source or have reached the consumer through the same channels as certain other goods that have already given the consumer satisfaction, and that bore the same trademark. Id.; see also Manhattan Shirt Co. v. Sarnoff-Irving Hat Stores, 164 A. 246 (Del. Ch. 1933). The Manhattan Shirt court speaks of the trademark in similar terms:

When the courts speak of the public's identifying the source of origin, they do not mean thereby that the purchasing public can identify the maker by his specific name or the place of manufacture by precise location. What they mean by such expression is that the purchaser of goods bearing a given label believes that what he buys emanated from the source, whatever its name or place, from which goods bearing that label have always been derived. and autonomous goodwill. As Edward Rogers, a leader of the trademark bar, argued:

[J]udges have begun to appreciate that the trade mark in and by itself is of little importance, that it is but the visible manifestation of a much more important thing, a business good will, that the good will is the substance, the trade mark merely the shadow, and that this business good will is the property to be protected against invasion.

This was a kind of Platonic theory of the mark, and trademark writing of the time is filled with it. 261 This is because such talk met an important need in the doctrine: the need to counter the argument from circularity. Felix Cohen argued that much of trademark law, if not of property law generally, was based upon circular reasoning: The law "purports to base legal protection upon economic value, when, as a matter of actual fact, the economic value of a sales device depends upon the extent to which it will be legally protected." 262 If only as a makeweight argument, this view could present goodwill-"reputation," "name," "business style"-as prior to rights, as something that had nothing to do with rights, as something that preceded sociolegal construction. Goodwill was, and remains for some, the "transcendental signified" 263 of trademark theory. The anonymous source theory-or more specifically, Schechter's analysis of it-made untenable the view that goodwill was prior to the trademark. I mentioned above that Saussurean linguistics initiated a radical break with the "nomenclaturist" assumption of some previous linguistics that words are simply names for preexisting ideas. 264 signifiers and signifieds, are mutually constitutive. Schechter brought the same analysis to trademark law (and did so, interestingly, about a decade after the publication of the Course). 265 In an oft-quoted passage, he argued that "today the trademark is not merely the symbol of good will but often the most effective agent for the creation of good will, imprinting upon the public mind an anonymous and impersonal guarantee of satisfaction, creating a desire for further satisfactions. The mark actually sells the goods." 266 Schechter recognized that trademark law's traditional nomenclaturist conception of goodwill required a strict source theory of the mark, and that any departure from the strict source theory-from what he called the "antiquated neighborhood theory of trade" 267 -necessarily involved a departure from nomenclaturist orthodoxy:

Superficially it may appear to be very fine hair-splitting to say that while the consumer does not know the specific source of a trademarked article, he nevertheless knows that two articles, bearing the same mark, emanate from a single source. However, the precise distinction is vital in the present connection, for it brings out clearly the creative and not merely symbolic nature of the modern trademark or tradename.

In the trademark structure, the signifier creates the signified as much as the signified does the signifier. The two mutually articulate each other. If one should be privileged before the other, it is the signifier that is the semiotic engine of the mark: "The mark actually sells the goods." 269 In Schechter's time, such a declaration of the priority of the signifier was heresy. It openly invited the charge of circularity. For better or worse, Schechter was the first to untether the doctrine from the metaphysics of source. 270 Paradoxically, as the trademark was understood to mean less and less, it was understood to be more and more creative of whatever meaning it did convey. The second implication of the abandonment of the strict source theory is more social-theoretical in nature than doctrinal, but is nevertheless worthy of mention. In trademark law, the term "source" is now a legal fiction. The vacuum left by the abandonment of the strict source theory has been filled by the products themselves. Here, again, we confront the problem of mediation, the problem of whether the signifier refers to the signified or to the referent. In Schechter's view, the trademark pointed not to a specific source, but rather to a specific product: "The true functions of the trademark are . . . to identify 265.

There is no evidence to suggest that Schechter was at any time in his career aware of developments in linguistic theory. 266. a product as satisfactory and thereby to stimulate further purchases by the consuming public." 271 In this sense, the metaphysics of source was replaced by the fetishism of commodities-or, indeed, of trademarks themselves. 272 Trademarked products no longer came from any particular source. It was not clear, or at least not important, whether they were the products of workers or machines or social relations. Their source was "anonymous." To the extent that they had an identity, that identity was determined not by the source of their manufacture, but by their own relations to each other, relations mediated not by producers or consumers, but by trademarks.

Schechter's indictment in the "Rational Basis" of the strict source theory and his exposure of the signifier's priority are both significant, but these are not what made the article revolutionary. Rather, it is what is missing in the article that it makes it worth reading today. Throughout "Rational Basis," Schechter scrupulously avoids any consideration of the semantic content of the trademark, its positive or negative affect, 273 nor does he have anything to say about the "tarnishment" of a mark. Instead, his concern is far more abstract. He is concerned with the "preservation of the uniqueness or individuality of the trademark," 274 its "arresting uniqueness," 275 its "singularity" 276 and "identity," 277 for these, and not its positive affect, are what ultimately constitute the mark's "selling power." 278 An unmistakable emptiness pervades this continual emphasis on the importance of a mark's being "unique and different." 279 This is because Schechter is talking about form, not content, about relations of value rather than relations of signification. He is arguing for the protection of uniqueness per se. Consider the four principles of trademark protection he sets forth near the end of "Rational Basis":

(1) that the value of the modern trademark lies in its selling power; (2) that this selling power depends for its psychological hold upon the public, not merely upon the merit of the goods upon which it is used, but equally upon its uniqueness and singularity; (3) that such uniqueness or singularity is vitiated or impaired by its use upon either related or non-related goods; and (4) that the degree of its protection depends in turn upon the extent to which, through the efforts or ingenuity of its owner, it is actually unique and different from other marks. 280 Notably, in his fourth principle, Schechter argues that the scope of a trademark's protection should be determined by its degree of differential distinctiveness, by the degree to which the mark is "actually unique and different from other marks." Here, his concern is not so much with a mark's inherent differential distinctiveness, as with its "actual[ ]" differential distinctiveness acquired "through the efforts or ingenuity" of the mark's owner. But more importantly, Schechter argues that what is valuable in the trademark is that it conveys above all an abstract characteristic of itself, namely, that it is unique, that it is neither good nor bad but simply different. This difference is not the trademark's semantic content, but rather a formal description of that content. The trademark signifies this description, this difference. It stands for its own difference as against all other marks. In short, the mark possesses "selling power" to the extent that it is distinctive of its distinctiveness from other marks. This is, of course, a formulation, a half-century before its time, of the semiotic concept of "sign value," of the appeal, the "commercial magnetism," 281 of difference. As Schechter appears to have understood, the experience of sign value is ultimately a purely aesthetic, rather than an ethical or moral, experience. It is akin to the "shock of the logo" described by Leslie Savan: "The presence of a logo when unexpected (or its absence when expected) acts as a flash of consciousness-not good or bad consciousness-but as commenter, an interlocutor, a readjuster of the usual." 282 The sheer force of the charge overwhelms considerations of whether it is a positive or a negative charge. Indeed, the sheer force of the charge is what makes it in the end positive, even sublime, and pervaded with "selling power."

C. The Emergence of the Trademark as a "Floating Signifier" As Sara Stadler Nelson has recently pointed out, 283 notwithstanding the sophistication of his semiotics, Schechter's revolution remained incomplete.

He was clearly of two minds about the nature of differential distinctiveness and remained something of a traditionalist. He held on dearly to the notion of reference. As we saw above, for all of his apparent admiration of the "arresting uniqueness" of strong marks, he was not prepared to cut the cord entirely between this uniqueness and some thing to which it uniquely referred. That thing was not source, however, but the product. Schechter was especially concerned with protecting marks that "have, from the very beginning, been associated in the public mind with a particular product, not with a variety of products, and have created in the public consciousness an impression or symbol of the excellence of the particular product in question." 284 But a funny thing happened on the way to Schechter's new vision of the trademark. In the same century that Schechter's influence was working to render the trademark's signified a nullity, trademark producers were themselves working to render the trademark's referent a nullity as well. This is the story told in Part II of the rise of the dyadic mark, unaffixed to any particular product, referring to no product other than its own distinctiveness. The signified and referent having been thrown over, all that was left was the signifier itself (the part which we now, appropriately, mistake for the whole of the trademark structure). I also explained in Part II that, synchronically speaking, relations of signification require relations of value, but relations of value do not require relations of signification. Stated more simply, there cannot be meaning without difference, yet there can be difference without meaning, signifiers untethered to signifieds. Baudrillard describes the nature of this new "form-sign":

The form-sign describes an entirely different organization: the signified and the referent are now abolished to the sole profit of the play of signifiers, of a generalized formalization in which the code no longer refers back to any subjective or objective "reality," but to its own logic . . . . The sign no longer designates anything at all. It approaches its true structural limit which is to refer back only to other signs. All reality then becomes the place of a semiurgical manipulation, of a structured simulation.

Perhaps as an unintended consequence, Schechter is primarily responsible for the development of the doctrine dedicated to the preservation of such "signforms" as commercial symbols, without regard to the meaning or product to which they refer, and of the culture of "semiurgic manipulation" that has resulted. This is the doctrine of trademark "dilution," to which I now turn. V. TWO MODES OF TRADEMARK DILUTION Dilution is as difficult a concept to understand in trademark law as value is in semiotic thought. This is not surprising. The law's recognition of the phenomenon of dilution is essentially the law's recognition of the semiotic relation of value. Antidilution protection entails the protection of a trademark signifier's set of relations of difference with all other signifiers in the trademark system. Yet courts have clearly been uncomfortable with granting such a broad expanse of property rights and so they have sought repeatedly to reconfigure antidilution protection as the protection not of intermark relations of value, but of intramark relations of signification. Under the rubric of "blurring," they have sought to reduce antidilution protection to a form, albeit a topsy-turvy form, of anti-infringement protection. Two modes of antidilution protection have thus uneasily coexisted since the passage into law of the Federal Trademark Dilution Act (FTDA): 286 a "radical" mode that would grant exclusive rights over system-wide relations of value and a "compromise" mode that would only grant exclusive rights over sign-specific relations of signification. The problem is that the compromise mode invariably evolves into the radical mode. Antidilution protection is by its nature absolute and unlimitable. This part seeks to explain why.

Because trademark commentary has traditionally been at such a loss to explain the nature of dilution, courts and commentators have tended to return to Schechter's writing on the issue on the assumption that the original founder of the concept must have understood it and that we need only try to figure out what he was trying to say. This elevation of Schechter to the status of prophetic and all-knowing framer has not served us particularly well. As we saw in Part IV, Schechter had much to say about the importance of a mark's uniqueness. On the nature of dilution and its prevention, however, he was practically Delphic. "Rational Basis" is a ten-page article, most of which is given over to criticism of then current doctrine. It defines dilution as the "gradual whittling away or dispersion of the identity and hold upon the public mind of the mark or name by its use upon non-competing goods." 287 As he did in "Rational Basis," "Rational Basis" argues that "the preservation of the uniqueness of a trademark should constitute the only rational basis for its protection." 290 In defending this thesis, the article finesses the concern that antidilution protection will result in in gross rights and "an undesirable monopolization of language." 291 Schechter agrees that trademark producers should not be allowed to "farm out" their mark to others without also controlling the quality of the goods to which the mark is attached, but "this rule that a trademark must be appurtenant to a going concern should not in any way set limits to the extent of protection of such a mark when so appurtenant." 292 Moreover, since, in Schechter's view, only inherently distinctive marks should receive antidilution protection, there would be little risk that a firm's antidilution rights in a mark would limit the commercial speech of other firms: Aside from plaintiff's mark, "[a]ll the rest of infinity is open to defendant." 293 Schechter believed antidilution protection was necessary for two reasons. First, he believed, with some justification, that the anti-infringement law of his time did not adequately protect the senior user against junior uses of its mark on noncompeting goods. The law more or less subscribed to a simple maxim: "[I]f there is no competition, there can be no unfair competition." 294 In other words, if there is no diversion of a plaintiff's trade, there can be no trademark infringement. 295 Thus, the use of the mark BORDEN on ice cream was held not to infringe on the use of the mark BORDEN on milk. 296 The scope of trademark protection horizontally along the product axis in Schechter's time was, by modern standards, extraordinarily narrow. This provided an opening for an unscrupulous firm to sell its goods under a famous mark when its goods were unrelated to those traditionally sold under that famous mark, in the hope that the persuasive appeal of the famous mark would attract buyers even if they well knew that the goods did not come from the firm with which they normally associated the mark. Because there was no confusion as to source, the injunction against trademark infringement would not reach such uses. Schechter's solution was to grant to owners of qualifying marks an extraordinarily broad, essentially absolute scope of rights along the product axis, so that Kodak, for example, could enjoin the use of its mark on any other product, regardless of how far afield that product's characteristics lay from photographic film-in other words, regardless of actual consumer confusion as to source. In Schechter's view, the provision of such a broad scope of protection horizontally along the product axis was the only means of protecting the differential distinctiveness of a qualifying mark and preventing the mark's "disassociation from the particular product with which it has been used." 297 Second, and not generally recognized, Schechter believed that antidilution protection would bring greater predictability to the law. Though he had the critical instincts of a realist, he sought to rationalize trademark law by returning it to formalism. At least in his trademark jurisprudence, Schechter was emphatically not an empiricist. As any structural semiotician would, he preferred the predictability of langue to the vagaries of parole. He preferred system to event. Near the end of his authoritative history of trademark law, he digresses to criticize the consumer-confusion basis for trademark protection: Any theory of trade-mark protection which . . . does not focus the protective function of the court upon the good-will of the owner of the trade-mark, inevitably renders such owner dependent for protection, not so much upon the normal agencies for the creation of goodwill, such as the excellence of his product and the appeal of his advertising, as upon the judicial estimate of the state of the public mind. This psychological element is in any event at best an uncertain factor, and "the so-called ordinary purchaser changes his mental qualities with every judge Schechter believed that the bright line rules of antidilution doctrine, at least as he envisioned it, would relieve the courts of having to make an "estimate of the state of the public mind." 299 The subject matter of antidilution protection would be limited to arbitrary and fanciful marks, 300 categories whose boundaries were relatively easy to determine. As for the dilution inquiry itself, this would consist of a purely formal question: are the marks at issue closely similar? If so, then there is dilution. Because antidilution protection would enjoin even the slightest "whittling away," it did not matter how extensively defendant was using its closely similar mark. The vagaries of the consumer confusion test could thus be avoided.

There is a two-fold irony to Schechter's justifications for antidilution protection. The first is that, for the kind of marks Schechter hoped to protect, the scope of anti-infringement protection has by now expanded so broadly, particularly horizontally along the product axis, that antidilution protection is no longer necessary to protect their uniqueness or product identification. 301 Under current law, the junior use of marks such as COKE, KODAK, MICROSOFT, and NIKE would likely be held to confuse consumers as to source regardless of the good or service to which such marks might be affixed. Of this, Schechter might have approved. But the second irony is bitter in nature. Modern antidilution protection is characterized by precisely the kind of unpredictability and discretion that Schechter hoped to avoid in first setting forth the concept. This is the subject of the next subpart.

B. The Ambiguities of Antidilution Doctrine Then and Now 1. The Meaning of "Uniqueness" As mentioned in Part IV, Schechter was of two minds about the nature of differential distinctiveness and, consequently, of dilution. In the end, he never specified what exactly he meant by "uniqueness" when he urged that "the preservation of the uniqueness of a trademark should constitute the only rational basis for its protection." Two interpretations have been offered: first, that he meant to refer to the uniqueness of the link between the mark and the product to which it is affixed, and second, that he meant to refer to the The first interpretation suggests that Schechter intended antidilution protection to protect vertical relations of signification, that is, the mark as an identifier uniquely of a source or product. The second suggests that he intended it to protect horizontal relations of value, that is, the mark as a unique identifier, regardless of its source or product. The first would appear to be a kinder, gentler mode of antidilution protection than the second. As does anti-infringement protection, the first would appear only to enhance the trademark owner's exclusive rights over relations internal to the trademark, relations between the trademark's signifier and signified or referent, rather than over relations external to the trademark between the trademark's signifier and all other signifiers in the trademark system. Remarkably, this ambiguity in "Rational Basis" is recapitulated, apparently inadvertently, 302 in the language of the FTDA. Section 45 of the Lanham Act defines "dilution" in the following terms:

The term "dilution" means the lessening of the capacity of a famous mark to identify and distinguish goods or services, regardless of the presence or absence of-

(1) competition between the owner of the famous mark and other parties, or (2) likelihood of confusion, mistake or deception.

The section 45 definition conceives of dilution as involving the "blurring" of the vertical relation of signification between the trademark's signifier and its signified or referent. Section 43(c) of the Act, however, speaks of commercial activities which "cause[ ] dilution of the distinctive quality of the mark." This usage has been read to suggest that dilution involves some diminution in the mark's "distinctiveness per se," 304 could, theoretically, seek to address. The first involves a loss of what the court terms a mark's "'distinctiveness' as such, in the narrow sense of its singularity as a word symbol." 306 In the court's view, this conception of harm stems from Schechter's "radical dilution proposal, whose practical effect if fully adopted would be to create as the whole of trademark protection law property rights in gross in suitably 'unique' marks." 307 The court notes with relief that this conception of dilution "never has been legislatively adopted by any [state] jurisdiction in anything approaching that extreme form," 308 and further argues, that with respect to the FTDA, "we simply cannot believe that Congress could have intended, without making its intention to do so perfectly clear, to create property rights in gross, unlimited in time . . . even in 'famous' trademarks." 309 The Ringling Bros. court instead determines that "the end harm at which [the FTDA] is aimed is [the loss of] a mark's selling power, not its 'distinctiveness' as such." 310 The court offers no definition of "selling power," other than to refer to a mark's "selling power as an advertising agent for its goods or services." 311 The court thus takes advantage of the ambiguity of a term which Schechter himself left undefined and which is mentioned nowhere in the FTDA to equate dilution with the blurring of the link, of the vertical relation of signification, between the mark and the product to which it is affixed. 312 The result is that antidilution protection is reduced to little more than the protection of a mark's "typicality," that is, "its ability to conjure up a particular product category." 313 It is less clear that other courts appreciate the difference between the two modes of dilution identified in Ringling Bros. In the 1999 Second Circuit case of Nabisco, Inc. v PF Brands, Inc., 314 for example, the court states that the "antidilution statutes rest on a judgment that the 'stimulant effect' of a distinctive and well-known mark is a 'powerful selling tool' that deserves legal protection." 315 This seems to refer to the distinctiveness as such of the mark, as does the court's reference to the possibility that an unauthorized use of a mark may reduce "the public's perception that the mark signifies something unique, singular, or particular." 316 Yet the Nabisco court then suggests that antidilution protection seeks to protect the "selling power" that a mark "has engendered for a product in the mind of the consuming public." 317 Similarly, in the 2000 case Federal Express Corp. v. Federal Espresso, Inc., 318 the Second Circuit refers to "blurring" as involving a mark's loss of "its ability to serve as a unique identifier," 319 which suggests a loss of differential distinctiveness, but then the court adds, "of the plaintiff's product," 320 which suggests instead a loss of distinctiveness of source or product. Other circuits have similarly conflated the two modes of harm. 321

In the 2003 case Moseley v. V Secret Catalogue, Inc., 322 the Supreme Court appears to have resolved the question of whether the FTDA seeks to protect the trademark's distinctiveness as such (its intermark relations of value) or rather the trademark's distinctiveness of a source or product (its intramark relations of signification). Though the Court declined explicitly to define dilution, it stated, arguably in dicta, that the "statutory requirement for dilution under the FTDA" is a showing that the defendant's mark will "reduce the capacity of the famous mark to identify the goods of its owner." 323 This would appear to be a decisive repudiation of what the Third Circuit called Schechter's "radical dilution proposal" and its entailment of "in gross property rights."

The Moseley Court's endorsement of the kinder, gentler productidentification mode of dilution did not constitute its primary holding, however. The Court resolved a far more urgent problem in the law when it held that the text of section 43(c)(1) "unambiguously requires a showing of actual dilution, rather than a likelihood of dilution." 324 This holding has been interpreted as a major setback for trademark producers. After all, proving actual blurring is tantamount to proving a negative. Conceived of as an impairment of the immediacy of the relation of signification between signifier and signified or referent, dilution is essentially an inverted theory of trademark infringement. Where anti-infringement protection is a shield that prevents consumer confusion as to source, antiblurring protection is a sword that promotes consumer identification as to source (or product). 325 The blurring theory of dilution seeks to give the consumer better than 20/20 vision. Proving that a defendant's use has degraded the typicality of a mark is especially difficult when the mark is famous, as it must be to receive protection under the FTDA. But while Moseley did indeed represent one step back for trademark producers, it also represented many more steps forward. In Ringling Bros., the Fourth Circuit held that the plaintiff must show actual dilution, and so, in this sense, the Fourth Circuit was on the winning side of Moseley. But the Fourth Circuit also held in Ringling Bros. that actual dilution may not be presumed simply from the close similarity or identity of the plaintiff's and the defendant's marks. 328 It did so for a very good reason. If the identity of the plaintiff's and the defendant's marks was sufficient on its own to show actual dilution, then, at least in such cases, the plaintiff would by force of this presumption enjoy exactly the same regime of in gross rights that 324.

Consider in this connection what might be termed "reverse blurring." In a conventional reverse passing-off situation, a defendant passes off the plaintiff's goods as its own. If, for example, the Coca-Cola Company bought Pepsi and resold it under the Coca-Cola trademark, the Coca-Cola Company would be liable for reverse passing off. See Dastar Corp. v. Twentieth Century Fox Film Corp., 123 S. Ct. 2041, 2047 (2003). Under our current theory of blurring, the Coca-Cola Company might also be liable for blurring the link between the product (Pepsi) and the trademark normally affixed to it. This result seems a bizarre extension of antidilution protection, yet it follows logically from blurring doctrine as it is currently conceived.

326 Schechter's "radical" "'distinctiveness' as such" conception of dilution would provide. 329 In Moseley, the Supreme Court explicitly repudiated the Ringling Bros. holding on this issue. 330 The Court instead held open the possibility that "direct evidence of dilution such as consumer surveys will not be necessary if actual dilution can be reliably proved through circumstantial evidence-the obvious case is one where the junior and senior marks are identical." 331 Thus what Moseley took away in holding that actual dilution must be shown, it gave back in allowing, if not encouraging, the possibility that when the plaintiff's and the defendant's marks are identical, actual dilution may be presumed. Reports of the death of Schechter's "radical" antidilution proposal are apparently exaggerated.

The Supreme Court left at least two issues unresolved in Moseley. The first is the question of whether marks must be inherently distinctive of source, rather than merely descriptive with secondary meaning, to qualify for federal antidilution protection. 332 The Court acknowledged the existence of this question in two cryptic footnotes to the opinion, but answered it in neither. 333 The Second Circuit has held in a series of cases that, under the language of section 43(c)(1), only inherently distinctive marks so qualify. 334 The Second Circuit's reasoning boils down to essentially one argument: that the term "distinctive" in the FTDA means "inherently distinctive." The circuit's reading of the statutory language, though ingenious, is wholly unpersuasive. For at least a century-and-a-half of American trademark jurisprudence, "distinctiveness" has been used as a generic term to refer collectively to two species of distinctiveness: inherent distinctiveness and acquired distinctiveness. The FTDA itself posits as one factor in determining whether a mark is "famous" the "degree of inherent or acquired distinctiveness of the mark." The legislative history also supports the proposition that Congress meant to use the term generically. 335 Even so, the Second Circuit's heart is in the right place. Underlying its strained textual interpretation is a policy argument, the force of which is undeniable. Like the Fourth Circuit in Ringling Bros., the Second Circuit is concerned about the grant of in gross property rights, particularly in terms that are commonplace source-identifiers. In the 2001 case of TCPIP Holding Co. v. Haar Communications, Inc., 336 Judge Pierre Leval sets forth what might be termed the "American Airlines argument": Descriptive marks, often asserting geographical identity or nation-wide prominence, or claiming merit or strength, abound in the U.S. marketplace. A few well-known examples are American, National, Continental, Metropolitan, Pacific, Southern, Texas, Chicago, Federated, United, Consolidated, Allied, First National, Acme, Merit, and so forth. Some of the holders of these inherently weak marks are huge companies; as a function of their commercial dominance, their marks have become famous. It seems unlikely that Congress could have intended that the holders of such non-distinctive marks would be entitled to claim exclusivity for them throughout all areas of commerce. Innumerable good-faith junior users of the same weak marks, who have developed goodwill in these marks, would be denied further use of their marks to their detriment and that of their customers. And nation-wide, throughout all areas of commerce, the use of ordinary, descriptive marks like American would be restricted to one famous user (and others whose use pre-dated the plaintiff user's achievement of fame). 337 Judge Leval is certainly correct that if AMERICAN AIRLINES falls within the subject matter of federal antidilution protection, then its holder could conceivably enjoin any other company from using the signifier AMERICAN in the American marketplace if such company began its use after AMERICAN AIRLINES became a "famous" mark. The same would hold true of the other examples Judge Leval gives.

The Second Circuit seeks essentially to impose a functionality limitation on the subject matter of antidilution protection. The argument is that certain marks, specifically, those which are descriptive, are sufficiently functional that their owners should not be given the in gross, marketwide property rights entailed by antidilution protection. While the benefits of limiting consumer confusion are judged to outweigh the costs in functionality of anti-infringement protection for such marks, antidilution protection is a whole other order of magnitude, extending across the product axis into every conceivable noncompeting good or service. Where anti-infringement rights in such marks are limited by the requirement that consumer confusion must be shown, antidilution rights are absolute.

The problem is that while some imposition of a functionality limitation makes sense, drawing the line at inherently distinctive marks will not satisfy the goals of the limitation. On the one hand, it would deny antidilution protection to marks which, in Congress's judgment, should receive protection. In TCPIP, Judge Leval cites approvingly the examples Congress gives in the FTDA's legislative history of marks that should receive protection: "The three marks cited as possible beneficiaries of the [FTDA] were Dupont, Buick, and Kodak-all highly distinctive, arbitrary or fanciful marks." 338 Of course, DUPONT and BUICK are not, in fact, inherently distinctive. Both were used as surnames and are thus per se descriptive under basic trademark doctrine. On the other hand, the granting of absolute, antidilution rights in suggestive and arbitrary marks, both of which fall within the category of inherent distinctiveness, could conceivably result in the same functionality loss as the granting of such rights in descriptive marks. If I affix the signifier ATLAS to the computers I manufacture, my trademark is arbitrary and thus deserving of antidilution protection under the Second Circuit's standard. But that protection would allow me to enjoin the use of the term by any company in the marketplace, including companies that might use the term descriptively. The same would hold true for most words in the dictionary, even those whose use would be deemed "highly suggestive." Ultimately, to meet the goals of the Second Circuit's functionality limitation, only fanciful marks should receive antidilution protection, for they are, by definition, invented signifiers and exclusive rights to their use will not deplete the language's stock of potential source-identifiers.

How then can we limit the subject matter of federal antidilution protection to prevent the outright ownership of commonplace words? First, we should abandon the requirement that marks be inherently distinctive of source. Indeed, we should abandon any restriction based on distinctiveness of, that is, based on intramark relations of signification between the signifier and signified. As the Second Circuit rightly points out, this is far too generous a framework for determining the subject matter of antidilution protection. By showing that its mark is distinctive of source, the plaintiff secures rights 338.

Id. that involve distinctiveness from other marks, rights which are absolute. Instead, we should fit the semiotic relation underlying the subject matter of antidilution protection to the semiotic relation underlying the scope of that protection. In both cases, the value-relations of the mark, rather than its signification-relations, should be considered. The plaintiff should be required to show that its mark is absolutely distinctive from other marks in order to receive exclusive rights in such absolute intermark distinctiveness. In other words, we should impose the requirement logically called for by the Second Circuit's reasoning: Only fanciful marks-that are also "famous"-should receive federal antidilution protection.

A second issue that the Moseley Court left undecided is the question of whether "tarnishment" is a form of trademark dilution prohibited under the FTDA. 339 It is well understood that a trademark's signifier does more that denote the trademark's source or product. It also tends to connote various characteristics of the source or product, such as quality, prestige, exclusiveness, or "wholesome[ness]." 340 These characteristics form part of the trademark's signified. The question, then, is how can the law protect these connotations? How can the law prevent a defendant from using or invoking plaintiff's mark in such a way that consumers, otherwise unconfused as to source, may begin to associate the mark with shoddy quality or low prestige, perhaps even with scandalous conduct? In the 2002 case New York Stock Exchange v. New York, New York Hotel 341 the Second Circuit faced precisely such a situation. The New York Stock Exchange (NYSE) sued the operator of a Las Vegas casino, claiming that, among other things, the casino's use of a façade evocative of the NYSE's famous façade and its use of the phrase "New York, New York $lot Exchange" constituted trademark infringement and trademark dilution. On appeal from summary judgment, 342 the Second Circuit held that the NYSE could not show trademark infringement because the casino's uses were so clearly parodic in nature. Consumers were simply not confused as to source. 343 Nor, for similar reasons, could the NYSE show dilution by blurring. Because "[t]he humor or parody in the Casino's use of the modified NYSE marks depends upon the fact that the Casino is not claiming to be associated with NYSE," there is "no diminution of the capacity of NYSE's marks to serve as a unique identifier of its products and services." 344 The Second Circuit held, however, that the NYSE could show "tarnishment" of its business reputation. The court explained that "[t]arnishment occurs where a trademark is 'linked to products of shoddy quality, or is portrayed in an unwholesome or unsavory context,' with the result that 'the public will associate the lack of quality or lack of prestige in the defendant's goods with the plaintiff's unrelated goods.'" 345 To the extent that the casino's evocation of the NYSE might associate securities trading on the NYSE with mere gambling and with stacked odds, the Second Circuit held that the casino's uses could tarnish the NYSE brand's apparent connotations of rational action and fair dealing. 346 The tarnishment claim at issue in New York Stock Exchange proceeded not from the FTDA or any other section of the Lanham Act, but from New York law, specifically from section 360-1 of the New York General Business Law. This section, typical of state law, provides: Likelihood of injury to business reputation or of dilution of the distinctive quality of a mark or trade name shall be a ground for injunctive relief in cases of infringement of a mark . . . or in cases of unfair competition, notwithstanding the absence of competition between the parties or the absence of confusion as to the source of goods or services. 347 Importantly, the tarnishment claim that the Second Circuit considered was not a dilution claim, but rather a claim of "injury to business reputation." Under most state laws and in the view of many trademark commentators, these two forms of injury are, in the words of the Restatement Third of Unfair Competition, "conceptually distinct." 348 While "dilution of the distinctive quality of a mark" may in some sense constitute "injury to business reputation," "injury to business reputation" does not necessarily entail, as it did not in New York Stock Exchange, "dilution of the distinctive quality of the mark."

Unlike most state law, the Lanham Act makes no explicit mention of injury to business reputation. When the United States Trademark Association first proposed a federal antidilution provision in 1987, it proposed adding a new subsection to section 43(a) to address trademark tarnishment. 349 When the FTDA was enacted in 1995, however, no such subsection was created. Though the drafters of the FTDA stated in the House Report accompanying the FTDA bill that the language of the FTDA was intended to prohibit tarnishment, 350 the plain meaning of the actual statutory language passed into law does not. 351 Remarkably, this has not prevented a wide array of federal courts from reading into the FTDA a federal cause of action against tarnishment. They have done so on the assumption, made in light of the FTDA's legislative history, that the FTDA's definition of dilution "encompasses traditional state law doctrines of blurring and tarnishment." 352 Of course, the FTDA's definition of dilution does no such thing, nor does our traditional theory of dilution, be it dilution of source distinctiveness or of differential distinctiveness. There is ultimately no ambiguity here. The term "dilution" need not be itself a "floating signifier." It has a specific meaning: It involves the "dilution," the attenuation, the lessening, of the immediacy of relations of reference. In the case of the dilution of source distinctiveness, it involves a lessening of the immediacy of the relation of reference between the trademark's signifier and its signified or referent. In the case of the dilution of differential distinctiveness, it involves a lessening of the immediacy of relations of difference (which are a form of negative reference) between the trademark's signifier, on the one hand, and all other signifiers in the trademark system, on the other. The prohibition against dilution is a prohibition against interference in relations of meaning, not in what is ultimately meant. The concept of dilution goes to form, not substance.

The federal courts' reading of antitarnishment protection into the FTDA is probably the clearest sign of just how undisciplined our analysis and application of antidilution protection has been. Consider the subject matter limitations of the FTDA. By its terms, section 43(c) protects only those trademarks which qualify as "famous." This limitation makes sense when applied in connection with antiblurring protection. It is at least arguable that only truly renowned marks require, in order to protect their renown, the wide grant of property rights called for by antiblurring protection. But there is no good reason, and certainly none stated in the FTDA's legislative history, to limit antitarnishment protection to the category of famous marks. 353 It is likely that section 43 will be amended at some point explicitly to include a cause of action for damage to business reputation. Until then, federal courts that find a prohibition against tarnishment in the FTDA's prohibition against dilution, on the misguided assumption that the former is a form of the latter, will do so in error.

The final ambiguity in dilution doctrine worth addressing here goes to the problem of mediation. As explained above, the problem of mediation involves the problem of whether the signifier points to the signified or to the referent within the triadic sign. This appears in the context of antidilution protection as the problem of whether "blurring" involves a blurring of the link between trademark and product or between trademark and source. Conventional blurring theory would appear to be fairly clear on this issue. The formulations and examples of blurring mentioned so far in this part all involve a blurring of product identification, a lessening of typicality. Thus, in the 1998 case Luigino's, Inc. v. Stouffer Corp., 354 the Eighth Circuit distinguished between infringement and dilution: "Infringement depends on a likelihood of consumer confusion over the source of a product, while dilution by blurring 'concerns the lessening of the capacity of a famous mark to identify and distinguish goods or services.'" 355 353.

Differently stated, while, under antiblurring protection, we give protections of form only to marks of a certain form (specifically, marks which are famous), there is no analogous reason under antitarnishment protection to give protections of substance or meaning only to marks of a certain form (specifically, marks which are famous). To be sure, a mark whose meaning is tasteless or somehow objectionable could reasonably be denied antitarnishment protection on that ground, but this limitation of the subject matter of antitarnishment protection makes sense only with respect to the meaning of the mark, not its degree of fame.

354 or VIRGIN or SONY now typical, the typicality of which must be protected? I argued above that "source" is now little more than a legal fiction in trademark law. The trademark identifies "source" only to the extent that it identifies some anonymous entity that produced or was at least in some way affiliated with the production of previous products bearing that trademark. In the context of antidilution product identification protection, "product" has similarly become a legal fiction. In essence, it stands for a line of products, perhaps related, perhaps unrelated, whose one commonality is that each bears the signifier the other bears. Blurring would appear, then, to involve the blurring of the link between trademark and source, even if that source is anonymous. This formulation resolves the problem of self-blurring. Marketing literature suggests that brand extensions do not blur the source identification link but actually enhance it. 366 Nevertheless, this formulation raises a larger conceptual problem. If the trademark is no longer pointing to a product, and if the source it points to is anonymous, that is, based only on the identity of the trademark signifier across uses on an array of referents, then what semiotic relation is really being blurred? The answer is that it is the identity of the trademark signifier itself that is being diluted-"vitiation of identity" Schechter would have said 367 -irrespective of to what source or product the signifier refers. What is at issue, in other words, is not relations of reference, but relations of difference. What is being protected is not the trademark's distinctiveness of source or product, but its distinctiveness from all other trademarks, including, most importantly, identical marks. Because "product" and "source" are both little more than legal fictions in current trademark doctrine, the notion of blurring has become very abstract indeed. I mentioned in Part II the Peircean distinction between token and type: Each appearance of a word or other sign is a token that invokes that word or sign's one type. Ultimately, in protecting the identity of the trademark signifier itself, antiblurring protection preserves the uniqueness of the relation not between the trademark's signifier and its signified or referent, but between the trademark signifier's one type and the many tokens of that type in the marketplace. A defendant "blurs" this relation when it establishes a new similar or identical type to which plaintiff's or defendant's tokens might also refer. Antidilution protection thus necessarily provides exclusive, essentially in gross rights to types.

See C. The Absolutism of Antidilution Rights Hostile courts have long engaged in a tradition of "drawing the teeth" 368 of antidilution statutes, most brazenly through restrictions on the subject matter of antidilution protection, like those imposed by the Second Circuit in TCPIP, or through the imposition of nearly insurmountable evidentiary requirements, like those imposed by the Fourth Circuit in Ringling Bros. The most effective, because the most elusive, threat to Schechter's original "radical" proposal, however, has been the development of the blurring theory of dilution, a compromise theory that seeks to protect intrasign relations of reference just as anti-infringement protection does. Schechter's own indecisiveness about the nature of differential distinctiveness openly invited the development of this theory and the courts have made the best of it. Yet Schechter's "radical" proposal has survived, and if Moseley is any indication, it has only gotten stronger. There is a simple semiotic explanation for the radical proposal's perdurability. The Ringling Bros. court vaguely hinted at this explanation when it noted that Schechter's proposal is only coherent in its radical form:

[T]he cases demonstrate that once the dilution concept is sought to be given any form other than that of Schechter's simple original proposal it begins to lose its coherence as a legally enforceable norm. Specifically, it becomes difficult to identify the legal interest sought to be protected from 'dilution,' hence the legal harm sought to be prevented.

369 By its very nature, antidilution protection is either absolute or it is meaningless. As if by its own semiotic logic, the concept of dilution will resist any attempts to limit its reach. This is because "dilution," either in its radical or compromise forms, necessarily involves global intermark relations of value in addition to local intramark relations of signification. Consider Saussure's discussion of semiotic value in the Course:

[T]he idea of value . . . shows that to consider a term as simply the union of a certain sound with a certain concept is grossly misleading. To define it in this way would isolate the term from its system; it would mean assuming that one can start from the terms and construct the system by adding them together when, on the contrary, it is from the interdependent whole that one must start and through analysis obtain its elements. The blurring theory of dilution has sought to consider the trademark as simply the union of a certain trademark with a certain source or product. In doing so, it has sought to isolate the trademark from the trademark system. This has proven to be impossible, however, because relations of signification, which blurring seeks to protect, require relations of value. Distinctiveness of requires distinctiveness from, and while distinctiveness of is by its nature limited to relations within the mark, distinctiveness from is not. To protect against any mitigation, any "whittling away," of distinctiveness of, the law must prevent any mitigation of distinctiveness from, and this entails a commitment to global, absolute, systemic protection. Stated differently, the blurring of a trademark signifier's distinctiveness of its signified or referent is powerful evidence of dilution, but it is not dilution itself. Blurring is merely one effect of the lessening of the trademark signifier's distinctiveness from other signifiers. To prevent this blurring, this effect, the law must address its cause. The law does so by preserving the differential distinctiveness of the trademark's signifier, for this differential distinctiveness makes possible the signifier's distinctiveness of its signified or referent.

Try as we might to limit it, Schechter's original concept, even in its product or source identification form, resists our efforts. "Radical" outcomes spring even from compromise premises. This, I think, accounts for why the concept has for so long proven to be "bewilderingly intangible." 371 Strong judges simply cannot accept that exclusive rights should be given in types, and so they have repeatedly, and no doubt in some cases consciously, misinterpreted the concept in an effort to misinterpret it out of existence.

In Jonathan Swift's Gulliver's Travels, the academicians of Lagago developed a new object language:

[A] Scheme for entirely abolishing all words whatsoever . . . was urged as a great advantage in point of health as well as brevity. For it is plain, that every word we speak is in some degree a diminution of our lungs by corrosion, and consequently contributes to the shortening of our lives. An expedient was therefore offered, that since words are only names for things, it would be more convenient for all men to carry about them, such things as were necessary to express the particular business they are to discourse on. Few adopted this language, however. It was simply too burdensome to carry one's lexicon on one's back. 373 The rise of consumer culture may be characterized by, among other things, the adoption by consumers of the language of the Lagagoan scholars. It has long been a cliché, of social theory as much as of advertising practice, that consumers communicate with each other by the objects they consume. Of late, however, commodity culture has begun to unburden itself of the object language of material commodities. The trademark system has developed as an alternative language of consumption, and its development has been rapid indeed. No other language in history, and certainly no other language of distinction, has experienced such explosive growth, both extensively and intensively, in so short a time. The trademark system's classificatory scheme now orders culture as much as the market. To be sure, it is not the only such system to do so, but none exercises its classifying function so exoterically, in terms so easily and widely understood. Where other systems of distinction are opaque, even unknowable but to those whom they privilege, the power of the trademark scheme resides in its emphatic transparency.

The emergence of the trademark system as a language of social differentiation offers a deeper and perhaps more satisfying explanation for many of the phenomena that this Article has identified: the collapse of the triadic structure, the rise of differential distinctiveness, the advent of the hypermark, the development of and futile attempt to control antidilution doctrine. While the semiotic analysis-or the economic analysis, for that matter-may go far towards explaining how the trademark system has evolved as it has, it will not ultimately explain why. Underlying the semiotic logic of the trademark system and of trademark law is a more profound logic, what Baudrillard has termed the "social logic of differentiation." 374 This logic meets "an objective social demand for signs and differences," 375 for "the distinguishing processes of class or caste which are fundamental to the social structure and are given free rein in 'democratic' society." 376 The trademark system is ultimately both an agent and an object of "classification struggle": 377 an agent in its own struggle to establish itself as the preeminent system of classification, and an object in the struggle by producers and consumers within the trademark system to bend its classificatory scheme to their own economic or cultural ends. For all of our efforts to reform trademark law, it is unlikely that any reform of the law itself will alter this underlying social logic. Perhaps we can fortify the defense of fair use or expand the functionality bar to protection. Perhaps we can persuade courts that consumers do not so easily confuse two similar, but not identical marks and thus bring about a limiting of the scope of anti-infringement protection. Perhaps we can somehow even cabin the notion of trademark dilution. But none of these reforms will stem the cultural, "semiurgic" tide. More likely, the further rationalization of the law will only quicken it. Consumers will continue to demand signs, distinctions, differences. As presently conceived, the goal of trademark law is to meet that demand as efficiently as possible. Whether we can conceive of a different goal for trademark law remains an open question. This is not a question of law or economics, however, but ultimately one of aesthetics and politics.