187 F.3d 363 (4th Cir. 1999)

ASHLEY FURNITURE INDUSTRIES, INCORPORATED, A Wisconsin corporation, Plaintiff-Appellant,
v.
SANGIACOMO N. A. LIMITED, a New Jersey Corporation; CARLO BARGAGLI-STOFFI, Defendants-Appellees. DREXEL HERITAGE FURNITURE COMPANY, Amicus Curiae.

No. 98-2228 (CA-97-309-2).

UNITED STATES COURT OF APPEALS, FOR THE FOURTH CIRCUIT.

Argued: May 18, 1999.
Decided: August 10, 1999.

Appeal from the United States District Court for the Middle District of North Carolina, at Greensboro.

Frank W. Bullock, Jr. Chief District Judge.[Copyrighted Material Omitted][Copyrighted Material Omitted]

COUNSEL ARGUED: D. David Hill, MCANDREWS, HELD & MALLOY, LTD., Chicago, Illinois, for Appellant. Gilbert Julian Andia, Jr., RHODES, COATS & BENNETT, L.L.P., Greensboro, North Carolina, for Appellees. ON BRIEF: John J. Held, Thomas J. Wimbiscus, MCANDREWS, HELD & MALLOY, LTD., Chicago, Illinois, for Appellant. C. Robert Rhodes, James L. Lester, RHODES, COATS & BENNETT, L.L.P., Greensboro, North Carolina, for Appellees. Stephen M. Trattner, John J. Dabney, TRATTNER & ASSOCIATES, Washington, D.C., for Amicus Curiae.

Before MICHAEL, MOTZ, and KING, Circuit Judges.

Reversed and remanded by published opinion. Judge Motz wrote the opinion, in which Judge Michael and Judge King joined.

OPINION

DIANA GRIBBON MOTZ, Circuit Judge:

1

This case requires us to determine under what circumstances the configuration of a product can constitute inherently distinctive trade dress that is protectable under federal law. Because we conclude that a product's configuration qualifies as inherently distinctive trade dress if it is capable of functioning as a designator of an individual source of the product, we reverse the contrary ruling of the district court and remand for further proceedings.

I.

2

Ashley Furniture Industries, Inc., and SanGiacomo N.A. Ltd. manufacture home furniture. In this action, Ashley alleges that SanGiacomo copied the design of one of Ashley's bedroom suites, in violation of federal trade dress law and an oral contract between the parties.

3

Both companies market furniture in the mid-level price range and sell numerous lines of furniture nationwide. Between them, for example, the two companies produce over thirty different bedroom suites. Ashley and SanGiacomo sell their furniture to retailers, who display the furniture of various manufacturers and fill consumers' orders either from inventory or by ordering from the wholesaler.

4

Generally, those seeking to buy this sort of furniture cannot determine the manufacturer of a piece or set of furniture simply by looking at it. Manufacturers in this market do not carve or emboss their names or trademarks on the exterior of their furniture; instead they rely on hang tags and emblems inside drawers to identify their goods. Often, retailers do not identify the manufacturer or the name of the design to the consumer, refusing even to use the hang tags. Retailers provide pictures of the furniture they offer in flyers, newspapers, and on television, but again these advertisements may not include any reference to the manufacturer.

5

Ashley and SanGiacomo aggressively compete for customers, and their rivalry has previously led to litigation. In that action, filed in 1993, the parties' roles were reversed: SanGiacomo sued Ashley charging that it had infringed the trade dress of SanGiacomo's bedroom furniture and seeking a declaratory judgment that a design patent obtained by Ashley was invalid as a matter of law. The parties settled the claims during a one-on-one, closed-door negotiation between Carlo Bargagli-Stoffi, the chief executive of SanGiacomo, and Ronald Wanek, the chairman and chief executive of Ashley. Wanek and Bargagli-Stoffi agreed that both sides would walk away from the suit, and that neither company would, in the future, copy the other's furniture designs. Bargagli-Stoffi characterizes this deal as a "gentlemen's agreement" that they would not make any "Polaroid copies" of the other's designs. Although the parties then entered into a written settlement agreement drawn up by counsel, that formal agreement contains no reference to any mutual covenant not to copy.

6

Approximately one year after that settlement, in the fall of 1995, Ashley introduced a neoclassical bedroom suite under the trade name "Sommerset." The suite consists of a headboard, two night stands, an armoire, and a dresser with a mirror. The design of the Sommerset suite combines a "modern" high-gloss polyester look and feel with "classical" elements including a finish that suggests marble or travertine, fluted columns, arches, and entablatures.

7

According to Ashley's witnesses, the design and overall appearance of the Sommerset suite were unique and unlike any other bedroom furniture ever sold. For example, one Ashley industry expert, a retired home furniture retailer with 25 years of experience, explained:

8

Based on my experience, the common high gloss polyester finish in all the pieces of the Sommerset suite, in either the Carmelstone or lighter Goatskin finish, combined with the off white moldings and classic columns and flutings, provide a unique and unusual appearance for the Ashley Sommerset bedroom suite. The combination of features provides both a traditional and contemporary appearance. Although these individual features have been used in other bedroom suites, I do not recall seeing the combination of such features in a single bedroom suite. Therefore, the Ashley Sommerset has a unique appearance in the furniture industry . . . [that] distinguishes it from other bedroom suites in either the contemporary or the traditional furniture markets.

9

Another expert opined that "the overall appearance of Ashley's Sommerset bedroom suite is not a common design in the bedroom furniture market" but rather is "different and distinct from every other bedroom suite on the market or that had previously been on the market" in his 17-year retail home furniture experience. SanGiacomo offered no contradictory evidence.

10

Retail sales of the Sommerset suite began in March 1996, and it quickly became one of Ashley's best-selling bedroom designs.

11

In December 1996 or January 1997, SanGiacomo began selling a line of bedroom furniture under the name "La Dolce Vita." This bedroom set, like the Sommerset suite, consists of a headboard, two night stands, an armoire, and a dresser with a mirror. Also like Ashley's Sommerset, SanGiacomo's La Dolce Vita includes a high-gloss marble-like finish, fluted columns, arches, and entablatures. The entire Sommerset suite retails for approximately $2,500; the La Dolce Vita suite retails for approximately $1,800. Ashley asserts that the materials and craftsmanship used in making SanGiacomo's La Dolce Vita furniture are far inferior to those used in making its own Sommerset furniture. In the months that followed introduction of the La Dolce Vita collection, it displaced sales of Ashley's Sommerset suite at many retailers.

12

Ashley maintains that SanGiacomo's marketing of the La Dolce Vita design infringes upon its rights in the Sommerset design. Accordingly, it filed this action asserting a violation of federal trade dress law, breach of the parties' alleged oral agreement not to copy, and other state law claims. After discovery, Ashley moved for summary judgment on its breach of contract claim and on the issue of whether SanGiacomo copied the Sommerset design. SanGiacomo, in turn, moved for summary judgment on Ashley's trade dress claim. The district court denied Ashley's motion and granted SanGiacomo's. Ashley appeals, asserting that summary judgment should not have been granted on either the trade dress or contract claim.

II.

13

The trade dress of a product consists of its "total image and overall appearance," including its "size, shape, color or color combinations, texture, graphics, or even particular sales techniques." Two Pesos, Inc. v. Taco Cabana, Inc., 505 U.S. 763, 764 n.1 (1992) (quoting district court instruction); Tools USA & Equip. Co. v. Champ Frame Straightening Equip., Inc., 87 F.3d 654, 657 (4th Cir. 1996) (quoting Two Pesos).

14

In addition to protecting registered trademarks, the Lanham Act protects unregistered marks and trade dress. Specifically, § 43 of the Lanham Act provides that

15

[a]ny person who, on or in connection with any goods or services . . . uses in commerce any word, term, name, symbol, or device, or any combination thereof . . . which . . . is likely to cause confusion, or to cause mistake, or to deceive as to the affiliation, connection, or association of such person with another person, or as to the origin . . . of his or her goods . . . shall be liable in a civil action by any person who believes that he or she is or is likely to be damaged by such act.

16

15 U.S.C.A. § 1125(a)(West 1998).

17

The Supreme Court has explained that "[p]rotection of trade dress, no less than of trademarks, serves the Act's purpose." Two Pesos, 505 U.S. at 774. That purpose is two-fold: § 43(a) seeks (1) "to secure to the owner of the mark the goodwill of his business" and (2) "to protect the ability of consumers to distinguish among competing producers." Id. (quoting Park 'N Fly, Inc. v. Dollar Park & Fly, Inc., 469 U.S. 189, 198 (1985)).

18

Thus, § 43 does not prohibit copying per se. Rather, the statute only prohibits a copy that can be passed off as the product of the originator, thereby confusing the consumer and interfering with the originator's rights in the goodwill of its business. See Bonito Boats, Inc. v. Thunder Craft Boats, Inc., 489 U.S. 141, 157 (1989) (citing Crescent Tool Co. v. Kilborn & Bishop Co., 247 F. 299, 301 (2d Cir. 1917) (Hand, J.)). By prohibiting copying just in this narrow class of circumstances, the statute protects the substantial public interest in free imitation. "[I]mitation is the life blood of competition" and only the ready "availability of substantially equivalent units . . . yield[s] the fair price society" seeks. American Safety Table Co. v. Schrieber, 269 F.2d 255, 272 (2d Cir. 1959).

19

To prove a case of trade dress infringement, a party must demonstrate that (1) its trade dress is primarily non-functional; (2) the alleged infringement creates a likelihood of confusion; and, (3) the trade dress either (a) is inherently distinctive, or (b) has acquired a secondary meaning. See Two Pesos, 505 U.S. at 769.

20

Without reaching the questions of functionality or consumer confusion, the district court granted summary judgment to SanGiacomo because it concluded that Ashley had not forecast sufficient evidence to demonstrate that its Sommerset design was inherently distinctive or had acquired a secondary meaning. On appeal, Ashley does not contest the secondary meaning ruling but does contend that the district court erred in holding as a matter of law that the Sommerset design's trade dress is not inherently distinctive.

21

In its careful opinion, the district court properly recognized that courts have differed as to how to determine inherent distinctiveness where, as here, the trade dress at issue is the product's configuration or design rather than its packaging. This question has bedeviled courts because determining inherent distinctiveness typically requires verbalizing a judgment about the visual character of the product. This is always difficult, but verbally describing the considerations that should inform that judgment in a manner that will assist the analysis in cases where the asserted trade dress is the actual design of the product presents a particularly formidable challenge. Justice Holmes's words in the trademark context seem especially apt here:"[b]eyond stating the principles to be applied there is little to be said except to compare the impression made by the two" product configurations. Joseph Schlitz Brewing Co. v. Houston Ice & Brewing Co., 250 U.S. 28, 30 (1919) (Holmes, J.). Below, we state and apply the relevant principles. Before doing so, however, we note that although we ultimately disagree with the district court, the thoroughness of its discussion has greatly assisted us.

III.

A.

22

In 1976, Judge Friendly articulated a methodology in Abercrombie & Fitch Co. v. Hunting World, Inc., 537 F.2d 4 (2d Cir. 1976), that has been cited, quoted, and applied in numerous subsequent trademark cases to determine inherent distinctiveness. See, e.g., Time, Inc. v. Petersen Publ'g Co., 173 F.3d 113, 118 (2d Cir. 1999); Perini Corp. v. Perini Constr., Inc., 915 F.2d 121, 124 (4th Cir. 1990); Union Nat'l Bank of Texas, Laredo v. Union Nat'l Bank of Texas, Austin, 909 F.2d 839, 844-45 (5th Cir. 1990); Walt-West Enters., Inc. v. Gannett Co., 695 F.2d 1050, 1055-56 (7th Cir. 1982). A court applying the Abercrombie analysis asks whether the trademark in question is (1) generic, (2) descriptive, (3) suggestive, (4) arbitrary, or (5) fanciful. See Perini, 915 F.2d at 124.

23

A generic mark refers to the genus or class of which a particular product is a member and can never be protected. A descriptive mark describes a characteristic of a product and can only be protected if it has acquired secondary meaning, i.e., if it has become distinctive of the maker's goods over time. A suggestive mark connotes a characteristic of a product, thereby enabling a consumer to infer something about the product from the mark. Both arbitrary and fanciful marks are totally unrelated to the product; the two differ, however, in that a mark qualifies as arbitrary if it is well-known in a different context, and fanciful if it is newly invented. With suggestive, arbitrary, and fanciful marks, "the association between the mark and its source is presumed and the mark is eligible for trademark protection." See Perini, 915 F.2d at 124-25. Because their "intrinsic nature" thus "serves to identify a particular source of a product," suggestive, arbitrary, and fanciful marks "are deemed inherently distinctive and are entitled to protection." Two Pesos, 505 U.S. at 768.

24

The rub, of course, is in trying to distinguish generic marks (ineligible for trademark protection) or descriptive marks (eligible only if accompanied by secondary meaning), from suggestive, arbitrary, or fanciful marks (all automatically eligible for trademark protection). This almost always depends on the particular facts of the case. For example, the trade name "Pet Store" for a shop that sold pets would be generic. The word "penguin," by contrast, would be descriptive when used in the name of a shop that specialized in items relating to a certain polar species; suggestive when denominating an air conditioning company; and arbitrary when the name of a book publishing company. (If the book publisher wanted a fanciful mark, it could use a made-up word, like "Penquell.")For two decades, the Abercrombie analysis has guided litigants and courts in determining "inherent distinctiveness" in trademark cases. In 1992, the Supreme Court in Two Pesos described Abercrombie as the "classic formulation" of inherent distinctiveness. 505 U.S. at 768. Moreover, the Two Pesos Court held that the court of appeals had been "quite right" in "follow[ing] the Abercrombie classifications consistently" to determine "whether trade dress for which protection is claimed under § 43(a) is inherently distinctive." Id. at 773 (emphasis added).

25

Despite the Supreme Court's unqualified and express approval of the application of the Abercrombie analysis in the trade dress context, some courts have held that the Abercrombie analysis does not apply where, as here, the alleged trade dress is the product's design or configuration. See, e.g., Knitwaves, Inc. v. Lollytogs Ltd., 71 F.3d 996, 1007-08 (2d Cir. 1995); Duraco Prods., Inc. v. Joy Plastic Enters., Ltd., 40 F.3d 1431, 1440-41 (3d Cir. 1994). These courts maintain that although Abercrombie's system of categories works well when a plaintiff seeks trade dress protection for things like logos or packaging, which can be conceptually separated from the product with ease, it is "unsuited for application to the product itself." Duraco, 40 F.3d at 1440-41. This criticism rests on the faulty premise that the trade dress being asserted in a product configuration case is "the product itself." Id. In fact, the trade dress of a product, even in a product configuration case, consists not of the entire product but only of those nonfunctional features of the product that, taken together, make up its total image. See Stuart Hall Co. v. Ampad Corp. , 51 F.3d 780, 788 (8th Cir. 1995) (rejecting criticism of Abercrombie because "trade dress is not the entire product itself, but specific features of the product").

26

The refusal to apply the Abercrombie analysis to trade dress cases involving product configuration also ignores the facts of Two Pesos itself. The plaintiff in Two Pesos asserted that a rival had infringed the trade dress of its Mexican restaurant; a jury agreed and awarded damages to the plaintiff, which were affirmed on appeal. The Supreme Court stated that the courts below had been correct in using Abercrombie to determine whether the decor and design of the restaurant constituted an inherently distinctive trade dress. Two Pesos, 505 U.S. at 773. Decor and design of a restaurant are not separable packaging. Rather they are a part of the product being sold at a restaurant, which consists not just of the food but of the entire dining experience. As such, a restaurant's decor and design are neither more nor less separable from it than furniture design is from furniture. Thus the Supreme Court's approval of the application of Abercrombie to restaurant design and decor surely indicates, if it does not require, that lower courts should use that analysis in product configuration cases generally.

27

Furthermore, contrary to suggestions made by courts that would limit the application of Abercrombie, see, e.g., Duraco, 40 F.3d at 1440-41, it is not inherently impossible, illogical, or anomalous to apply the Abercrombie categories to product configuration. The configuration of a banana-flavored candy, for example, would be generic if the candy were round, descriptive if it were shaped like a banana, suggestive if it were shaped like a monkey, arbitrary if it were shaped like a trombone, and fanciful if it were formed into some hitherto unknown shape.

28

This is not to say that the Abercrombie analysis may not be difficult to apply in some product configuration cases. But the contention that we should therefore refuse to apply it at all in such cases is, as the Supreme Court remarked with respect to another thorny Lanham Act issue, "unpersuasive . . . because it relies on an occasional problem to justify a blanket prohibition." Qualitex Co. v. Jacobson Prods. Co., 514 U.S. 159, 168 (1995).

29

Moreover, difficulties in applying Abercrombie to product configuration can often be mitigated by considering the principles stated in Seabrook Foods, Inc. v. Bar-Well Foods Ltd., 568 F.2d 1342 (C.C.P.A. 1977). The Seabrook court explained that in determining inherent distinctiveness a court looks to whether the alleged trade dress is "a ``common' basic shape or design," "unique or unusual in a particular field," or "a mere refinement of a commonly-adopted and well-known form of ornamentation for a particular class of goods viewed by the public as a dress or ornamentation for the goods." Seabrook, 568 F.2d at 1344. Seabrook makes plain that a product's overall design cannot be found inherently distinctive if it constitutes a "well-known" or "common" design, even if that design had not before been "refine[d]" in precisely the same way. Rather, to qualify as inherently distinctive a design must be "unique or unusual" in the "particular field" at issue.

30

In essence, Seabrook clarifies when a trade dress should be considered arbitrary or fanciful, on the one hand, or generic, on the other. This elaboration of Abercrombie is particularly useful in product configuration cases because the nonverbal character of the product features at issue in such cases makes them less likely to turn -as word mark cases frequently do -on whether the alleged trade dress is descriptive. At least compared with word marks, few product configurations will have substantial descriptive capabilities, and therefore descriptiveness is likely to be a real issue only rarely. Cases in which a product configuration has enough communicative value to raise a question under the suggestive category may arise more often. But we think it likely that, by and large, the crucial question in product configuration cases will be the question that Seabrook helps answer: whether an alleged trade dress can be considered arbitrary or fanciful or whether it must be ruled generic.

31

In sum, we hold that the Abercrombie categories, as Two Pesos suggests, provide the appropriate basic framework for deciding inherent distinctiveness in product configuration cases. Furthermore, we believe that the rules that SanGiacomo urges as alternatives to Abercrombie, see I.P. Lund Trading v. Kohler Co., 163 F.3d 27, 41 (1st Cir. 1998); Landscape Forms, Inc. v. Columbia Cascade Co., 113 F.3d 373, 378, 378 n.3 (2d Cir. 1997); Duraco , 40 F.3d at 1450, are at odds with the analysis set out in Two Pesos .

32

In ruling that an alleged trade dress, like a trademark, could be protectable absent a showing of secondary meaning, the Two Pesos Court exhibited a desire for a uniform approach to claims under § 43(a). See 505 U.S. at 773 (noting that inherent distinctiveness rule applies to trademark, that "protection of trademarks and trade dress under § 43 serves the same statutory purpose," and that"[t]here is no persuasive reason to apply different analysis to the two"). Moreover, in reaching its conclusion the Court emphasized that nothing in the language of § 43(a) would support a differentiation between trade dress and trademark of the kind for which the defendant had argued. Id. at 774. Similarly, the creation of a new rule to be applied only to product configuration would comport neither with the language of the statute nor with the Court's preference for uniformity. See Stuart Hall, 51 F.3d at 787 (reading Two Pesos as "resting on a presumption that ``trade dress' is a single concept that encompasses both product configuration and packaging").

33

The alternative tests that have been proposed also conflict with Two Pesos on a more specific level. These tests require that in order to qualify for trade dress protection a product configuration must "act as a[] . . . signifier of origin," Duraco , 40 F.3d at 1450, be "likely to serve primarily as a designator of origin," I.P. Lund, 163 F.3d at 41, or be "likely to be understood as an indicator" of the source, Landscape Forms, 113 F.3d at 378, 378 n.3, of the product. Yet the Two Pesos Court described inherently distinctive trade dress as trade dress that is "capable of identifying products or services as coming from a specific source." 524 U.S. at 773 (emphasis added). This clearly suggests that actual or likely consumer recognition of a product design as a source-designator is not a prerequisite for inherent distinctiveness.

34

Moreover, any "source-identification" requirement impermissibly undermines the distinction that the Supreme Court established in Two Pesos between "inherently distinctive" analysis and "secondary meaning" analysis. Prior to Two Pesos, some circuits had held that inherent distinctiveness in trade dress cases could only be established by proof that a trade dress had acquired a secondary meaning in consumers' minds. See Merchant & Evans, Inc. v. Roosevelt Bldg. Prods. Co., 963 F.2d 628, 633 (3d Cir. 1992); Wallace Int'l Silversmiths, Inc. v. Godinger Silver Art Co., 916 F.2d 76, 79 (2d Cir. 1990); Keebler Co. v. Rovira Biscuit Corp., 624 F.2d 366, 378 (1st Cir. 1980). In Two Pesos, the Supreme Court categorically rejected that view. 505 U.S. at 776 ("proof of secondary meaning is not required to prevail on a claim under § 43(a) of the Lanham Act where the trade dress at issue is inherently distinctive" (emphasis added)). Requiring a plaintiff to produce evidence that a design "acts" as a designator of origin directly subverts the Court's holding in Two Pesos by requiring a plaintiff who seeks to show inherent distinctiveness to produce the same kinds of evidence that are required for a showing of secondary meaning. See Stuart Hall, 51 F.3d at 788. If a product is inherently distinctive then its function as a designator -not only of origin, but of an individual origin -is presumed; that is, no evidence of source-identification by consumers need be presented. See Two Pesos, 505 U.S. at 770-71 (describing as "sound" the Fifth Circuit's statement that inherently distinctive trade dress is protectable "regardless of whether substantial consumer association yet bestows the additional empirical protection of secondary meaning"); cf. Perini, 915 F.2d at 124-25 (noting that association between suggestive, arbitrary, or fanciful mark and its source is presumed).

35

Nor does requiring that a design simply be "likely to serve" or "likely to be understood" as source-designating eliminate the tension between Two Pesos and source-identification tests. True, these more modest requirements would not necessarily mandate that a plaintiff produce evidence of the kind required to show secondary meaning; rather a plaintiff could, for example, rely on evidence that product design functions as an important means of source-designation in the relevant market, together with evidence that the product in question is unusual. But even these requirements misdirect the analysis because they improperly suggest that consumers would have to consciously recognize a product design as indicative of source in order for the design to qualify as inherently distinctive. Not only does this approach continue the improper emphasis on source identification, which is the province of secondary meaning, but it also ignores § 43(a)'s intent to prevent customer confusion. Customers could certainly end up being confused about the source of a product by a deceptive trade dress without consciously recognizing the trade dress of the original product as a source-designator upon first seeing it.

36

Thus, in determining inherent distinctiveness the question is not whether a product design "identifies" a product, is "likely to identify" a product, or is "likely to be understood" as identifying a product. We believe, rather, that the question is whether the design is "capable of identifying a product," Two Pesos, 524 U.S. at 773, that is, whether, consciously or not, consumers would be impressed enough by a product design that if they were to see something sufficiently like it, it would be reasonable for them to assume either that it is exactly the same product (and therefore implicitly the product of the same producer) or that it was a different product made by the same producer.

37

We follow the Supreme Court's lead in Two Pesos and hold that the venerable Abercrombie analysis provides the appropriate method for answering this question. See also Pebble Beach Co. v. Tour 18 I Ltd., 155 F.3d 526, 540 (5th Cir. 1998) (holding Abercrombie analysis applicable to product configuration trade dress); Stuart Hall Co., 51 F.3d at 788 (same). We turn to application of that analysis.

B.

38

When the district court applied the Abercrombie analysis it concluded as a matter of law that "[t]he Sommerset design is merely descriptive of itself and, as such, is not inherently distinctive." This ruling relies upon the same false premise that animated the Duraco and Knitwaves holdings, i.e., that the asserted trade dress in a product design case is "the product itself." Duraco, 40 F.3d at 1440-41; Knitwaves, 71 F.3d at 1007-08 (quoting Duraco).

39

As noted above, the trade dress in such cases is not"the product itself," but rather the nonfunctional aspects of the product that make up its total image. If this were not so, then a product design could never be protectable as an inherently distinctive trade dress because it would always be part of the product. Two Pesos simply does not permit that result -rather it clearly contemplates trade dress protection for inherently distinctive product configurations. Indeed, given that both the district court and the court of appeals in Two Pesos approved the jury's finding that the design and decor at issue there did not "describe" the restaurant to which it was attached, 505 U.S. at 770, a jury in this case could just as properly conclude that the Sommerset design does not "describe" the furniture to which it is attached.

40

The district court also erred in holding that the Sommerset design could not fall within the "arbitrary" or "fanciful" Abercrombie categories. The court reasoned that "[t]he use of a high-gloss finish, columns, arches and entablatures on bedroom furniture cannot be characterized as arbitrary or fanciful because, as Ashley admits, such features are common to the furniture industry."

41

At the heart of this holding is the erroneous belief that if the elements of a design are common in the industry, the design cannot be inherently distinctive. In fact, trade dress is the"total image" of a product, and thus the relevant inquiry is not whether the individual components of a design are common or not, but rather whether the alleged trade dress as a whole is inherently distinctive. See Two Pesos, 505 U.S. at 764 n.1; Tools, 87 F.3d at 657; Roulo v. Russ Berrie & Co., Inc., 886 F.2d 931, 935 (7th Cir. 1989); Hartford House, Ltd. v. Hallmark Cards, Inc., 846 F.2d 1268, 1271 (10th Cir. 1988). To be sure, a basic overall design that is"common" or simply a "refinement of a commonly-adopted and well-known" design cannot be inherently distinctive. See Seabrook, 568 F.2d at 1344. But the fact that design elements have been used separately before does not foreclose the possibility that their combination in a new, unique way will create an inherently distinctive trade dress.

42

For instance, in Ambrit, Inc. v. Kraft, Inc., 812 F.2d 1531 (11th Cir. 1986), the Eleventh Circuit affirmed the district court's finding of inherent distinctiveness in the packaging of the Klondike ice cream bar despite the fact that various elements of the alleged trade dress had been previously used by others. Id. at 1537. Likewise, in Roulo, the court upheld a finding that a line of greeting cards was inherently distinctive even though the cards "incorporated several common features such as stripes, dots, [and] hand-writing . . . which are indigenous to all greeting cards" because of evidence that "the same combination of elements" had not before been used in other cards. 886 F.2d at 936; see also Computer Care v. Service Sys. Enters., Inc., 982 F.2d 1063, 1069 (7th Cir. 1992) ("Where the plaintiff's overall trade dress is distinctive, the fact that it uses descriptive (or generic) elements does not render it nonprotectable.").

43

Similarly, the evidence here supports a possible finding that the Sommerset's total overall image was arbitrary or fanciful and therefore inherently distinctive. A factfinder could reasonably conclude that the total image created by the Sommerset's neo-Roman design has no more to do with bedroom furniture than a penguin does with a publishing company. Ashley produced uncontroverted evidence that the "the combination of features in the Sommerset-features which include the high gloss polyester Carmelstone and Goatskin finishes, off white moldings on the night stands, dresser, and armoire, classic columns and fluting on the headboard and mirror, and arched tops on the mirror and headboard frames -have never before been utilized together in a bedroom suite." Also undisputed was the evidence that "[t]he overall appearance" of the Sommerset "is not a common design in the bedroom furniture market," and that it has a "unique and unusual appearance" that "distinguishes it from other bedroom suites." From this evidence, a reasonable fact finder could conclude that even if the individual elements of the Sommerset design were common in the furniture industry, i.e., even if some of these elements fit within Abercrombie's "generic" category, the Sommerset's combination of these elements was sufficiently unique to be arbitrary or fanciful and therefore eligible for trade dress protection. See Roulo, 886 F.2d at 936.

44

Our determination that the Sommerset design could qualify for trade dress protection does not mean that every unique or novel product design will be protectable as an inherently distinctive trade dress. For an item to be considered inherently distinctive, it must differ from its predecessors in a more than trivial way; an alleged trade dress is not protectable if it is "a mere refinement of a commonly-adopted and well-known form of ornamentation for a particular class of goods." Seabrook, 568 F.2d at 1344. Thus if an overall product design differs only slightly from the designs of preexisting competing products, it does not qualify for protection as inherently distinctive trade dress.

45

For example, in EFS Mktg., Inc. v. Russ Berrie & Co., 836 F. Supp. 128 (S.D.N.Y. 1993), aff'd in part, rev'd in part, vacated in part on other grounds, 76 F.3d 487 (2d Cir. 1996), the plaintiffs contended that a line of troll dolls similar in appearance to dolls that had originally appeared in the 1960s were nonetheless inherently distinctive because of slight design changes, e.g.,"chin coming after smile line; rounded ears; brown [e]ye color; eyes closer together." Id. at 131. The court found, however, that "most of the product features" of the plaintiff's troll doll had figured in the design of troll dolls "for 25 years"; the plaintiff's changes to these features -rounding ears, changing eye color, etc. -constituted "mere refinements." Id. at 135-36. The court thus properly rejected the contention that some novelty or uniqueness of design is sufficient, in and of itself, to qualify a product as inherently distinctive. Id.

46

In contrast to the plaintiff in EFS, Ashley presented evidence that its Sommerset design is not merely unique, but also significantly and remarkably different from other designs. One expert testified that the design creates a "unique and unusual appearance" and that "the unique appearance of the Sommerset distinguishes it from other bedroom suites." Another opined that "[t]he overall appearance of Ashley's Sommerset bedroom suite is not a common design in the bedroom furniture market." If SanGiacomo had offered undisputed evidence that a design like its La Dolce Vita had preceded the Sommerset, then the district court would have been entirely justified in finding, as a matter of law, that the Sommerset constitutes a "mere refinement." But no evidence in the record here tends to show that such designs were on the market prior to the Sommerset. Rather, this record provides evidence (at the present, undisputed) from which a reasonable factfinder could conclude that the Sommerset was not merely different from predecessor designs, but that it was noticeably and significantly different and therefore inherently distinctive.

C.

47

Before leaving the trade dress question, we pause to examine the considerations that seem to have motivated courts to raise the bar for inherent distinctiveness in product configuration cases. On examination, we conclude that none of these considerations gives cause for alarm; indeed, most of them are adequately accommodated by traditional trademark and trade dress law.

48

One motive behind the efforts to create more rigorous inherent distinctiveness standards for product configuration cases is the fear that trade dress protection in such cases could have anti-competitive effects. The rule that a product feature is unprotectable if it is functional, however, fully safeguards competitors from the danger that a producer will, through trade dress law, obtain a monopoly over a useful product feature when the alternatives are limited. See Two Pesos, 505 U.S. at 775 (noting that functionality rules"serve[] to assure that competition will not be stifled by the exhaustion of a limited number of trade dresses"). Thus, as far as trade dress law is concerned, a company can compete by copying every useful feature of a competitor's product. The only thing a company cannot do under trade dress law is compete by producing products that imitate the nonfunctional, inherently distinctive features of existing products so closely that they are likely to cause consumer confusion. And that, of course, is precisely the purpose of § 43 of the Lanham Act.

49

Another concern about applying traditional (rather than more rigorous) trade dress rules in product configuration cases is that they will improperly allow the protection of product features that are aesthetically pleasing, or, as some authorities have put it, features that contribute to the inherent appeal of a product or serve a merely ornamental purpose. See Duraco, 40 F.3d at 1441, 1449; Restatement (Third) of Unfair Competition § 16, comment b (1995). In order to prevent such protection from being accorded, courts have required that a product configuration must be likely to serve"primarily" as a designator of origin in order to qualify as inherently distinctive; under such an approach, configurations deemed "primarily" aesthetic would be unprotectable. See Knitwaves, 71 F.3d at 1009; Duraco, 40 F.3d at 1450. The assumption that a feature must be either "primarily" source-designating or "primarily" aesthetic is, however, misconceived. The aesthetically pleasing features of a product generally serve the dual purpose of attracting customers to the product and making them distinctive in a manner that is capable of identifying them as having derived from an individual source. See Knitwaves, 71 F.3d at 378 n.3 (citing Krueger Int'l, Inc. v. Nightingale Inc., 915 F. Supp. 595, 606 (S.D.N.Y. 1996)).

50

More fundamentally, though, the problem of when aesthetic features can be protected is already dealt with in trade dress law by the functionality requirement. As the Supreme Court has explained, a feature qualifies as functional not only if it "is essential to the use or purpose of the article" but also if it "affects the cost or quality of the article, that is, if exclusive use of it would put competitors at a significant non-reputation-related disadvantage." Qualitex, 514 U.S. at 165 (emphasis added) (internal quotation marks omitted). If trade dress protection for an aesthetic design "would deprive competitors of alternative designs, and, thus, foreclose competition from the relevant market," the design would be considered functional and could not be protected. Landscape Forms, 113 F.3d at 377. Because the functionality requirement adequately safeguards the competitive use of aesthetic features, we see no reason for altering the inherent distinctiveness doctrine to limit protection of such features.

51

Yet another justification for heightening trade dress requirements in product configuration cases is the perceived need to protect the boundaries between trade dress, copyright, and patent law. The argument is that a producer who could prevent others from copying the distinctive configuration of its product through trade dress law would have little need for a patent or a copyright. We recognize that an improperly expansive application of trade dress principles could conceivably frustrate patent and copyright protection in some circumstances. Such problems can and should be avoided, however, by "careful application of traditional bases for determining the propriety" of trade dress protection, rather than by rewriting trade dress law. See Kohler Co. v. Moen, Inc., 12 F.3d 632, 642 (7th Cir. 1993) (Two Pesos indicates traditional trade dress principles should be used to prevent trademark law from encroaching upon patent law).

52

Allowing traditional trade dress protection for product configuration will not undermine either utility or design patent law. Indeed, the trade dress rule that functional aspects of a product, although patentable, are not protectable as trade dress directly eliminates any conflict between patents for useful items and trade dress protection for product configuration. See Kohler, 12 F.3d at 638 (quoting W.T. Rogers Co. v. Keene, 778 F.2d 334, 337 (7th Cir. 1985) (Posner, J.)). Furthermore, permitting such trade dress protection would not impermissibly give the trade dress plaintiff a perpetual design patent, as SanGiacomo suggests. Trade dress rights do, of course, persist indefinitely, while design patents provide only 14 years of protection. See 35 U.S.C.A. § 173 (West Supp. 1999). But the trade dress plaintiff must establish likelihood of customer confusion in order to be protected at all "which the design patent owner need not do; there is therefore no necessary inconsistency between the two modes of protection." W.T. Rogers, 778 F.2d at 337. Moreover, "[c]ompared to patent protection, trademark protection is relatively weak because it precludes competitors only from using marks that are likely to confuse or deceive the public." Kohler, 12 F.3d at 637. It should also be noted that trade dress rights, although of indefinite duration, are not necessarily perpetual. Such rights terminate if the trade dress is abandoned, Roulo, 886 F.2d at 938-39, or if the trade dress becomes generic through public usage, cf. King-Seeley Thermos Co. v. Aladdin Indus., Inc., 321 F.2d 577 (2d Cir. 1963) (addressing trademark law). Thus, if Ashley pulled the Sommerset design off the market for good, or if high-gloss polyester, neo-Roman design became all the rage, any trade dress rights established in the design would lapse. A current design patent, by contrast, would continue to provide protection regardless of such eventualities.

53

For similar reasons, trade dress protection for product configuration poses no general threat to the copyright regime. Again, trade dress protection requires a showing of likelihood of consumer confusion, but copyright law, like patent law, does not. Thus, in contrast to copyright infringement, "mere reproduction" of a trade dress "is not an infringement." Quality Inns Int'l, Inc. v. McDonald's Corp., 695 F. Supp. 198, 218 (D. Md. 1988) (Niemeyer, J.) (addressing trademark law). Of course, trade dress rights should not be improperly extended so as to vitiate the copyright rule that protection is not available for an idea, concept, or generalized type of appearance. See Jeffrey Milstein, Inc. v. Geger, Lawlor, Roth, Inc. , 58 F.3d 27, 32-33 (2d Cir. 1995); see also Galerie Furstenberg v. Coffaro, 697 F. Supp. 1282, 1290 (S.D.N.Y. 1988). For example, if Ashley is found to have protectable trade dress rights in its Sommerset design, that protection should not be held to prevent SanGiacomo or anyone else from producing neo-Roman furniture, or even high-gloss polyester neo-Roman bedroom furniture; it should only prevent them from producing furniture so similar to the Sommerset design that it would be likely to cause customer confusion. For these reasons, we believe that if courts addressing claims of trade dress protection for product configurations bear all the elements of trade dress law in mind and pay careful attention to the boundaries of patent and copyright law, application of traditional trade dress rules will not in any way weaken copyright or patent protection.

54

A final concern that may have led courts to adopt restrictive inherent distinctiveness standards is the fear that, absent such standards, a plaintiff who merely presents an expert's affidavit asserting that its product is significantly unique would survive summary judgment on the issue of inherent distinctiveness in every case. Actually, a court would in some cases be justified in rejecting an expert's opinion on inherent distinctiveness as conclusory or lacking factual foundation. Expert opinion asserting that black and white newsprint is inherently distinctive, for example, would not enable a plaintiff to survive summary judgment.

55

It is nonetheless true that an expert's affidavit will frequently enable the plaintiff to defeat summary judgment on the issue of inherent distinctiveness. This is not an unusual circumstance in the law; in many areas, an expert's affidavit will enable the plaintiff to survive summary judgment unless the defendant shows that the plaintiff's proof fails with regard to at least one element of the claim asserted. And indeed, results of this kind comport with the basic summary judgment principle that the moving party bears the burden of indicating why the nonmoving party's case fails to establish the existence of a genuine issue of material fact. See Celotex Corp. v. Catrett, 477 U.S. 317, 323 (1986).

56

Moreover, the question of inherent distinctiveness almost always involves heavily disputed facts and so is not ordinarily amenable to summary judgment. See, e.g., Mana Products, Inc. v. Columbia Cosmetics Mfg., 65 F.3d 1063, 1069 (2d Cir. 1995). In fact, our review of countless cases has only turned up a handful in which a court has dismissed a case or granted summary judgment on the basis of lack of inherent distinctiveness as a matter of law, and in two of those cases that ruling was reversed on appeal. See Paddington Corp. v. Attiki Importers & Distribs., Inc., 996 F.2d 577 (2d Cir. 1993); Abercrombie & Fitch Co. v. Hunting World, Inc., 461 F.2d 1040 (2d Cir. 1972). Conversely, because the concept of distinctiveness is an "intuitive" one, Publications Int'l, Ltd. v. Landoll, Inc., 164 F.3d 337, 340 (7th Cir. 1998), and heavily fact-dependent, when a fact finder does make findings on this question (either in determining whether to grant a preliminary injunction or after trial), an appellate court will naturally be exceedingly reluctant to rule such findings clearly erroneous.

57

In sum, we find no reason to depart from the traditional Abercrombie test for inherent distinctiveness, and we conclude that under that test Ashley's alleged trade dress in the Sommerset design could reasonably be found inherently distinctive on the present record. The district court therefore erred in granting summary judgment to SanGiacomo on Ashley's trade dress claim.

IV.

58

Ashley also contends that SanGiacomo, by copying the Sommerset design, breached an oral agreement between the parties not to copy each other's designs. The district court held that under North Carolina law, such an agreement had to be in writing to be enforceable.

59

North Carolina General Statute § 75-4 provides in part that "[n]o contract . . . limiting the rights of any person to do business anywhere in the State of North Carolina shall be enforceable unless such agreement is in writing and duly signed by the party who agrees not to enter into any such business within such territory." N.C. Gen. Stat. § 75-4 (1998). North Carolina courts have applied this provision to agreements prohibiting former employees or business associates from entering into businesses that compete with their former employers or associates, see, e.g., Manpower of Guilford County, Inc. v. Hedgecock, 257 S.E.2d 109, 113 (N.C. Ct. App. 1979) (employees); Keith v. Day, 343 S.E.2d 562, 567-68 (N.C. Ct. App. 1986) (business associates); agreements prohibiting competition by one of the parties to the sale of a business against the other, see Maola Ice Cream Co. v. Maola Milk & Ice Cream Co., 77 S.E.2d 910, 915 (N.C. 1953); and agreements between suppliers and distributors prohibiting the suppliers from using other distributors, see, e.g., Radio Elecs. Co. v. No. 386, 92 S.E.2d 664, 666-67 (N.C. 1956). The provision has never been applied to an agreement like the one at issue here.

60

More generally, North Carolina courts have repeatedly held that § 75-4 applies to an agreement that "limits substantially" a party's right to do business. See, e.g., Radio Elecs., 92 S.E.2d at 666; Norlin Indus., Inc. v. Music Arts, Inc., 313 S.E.2d 166, 169 (N.C. Ct. App. 1984). The courts have never defined what constitutes a "substantial" limitation, but no matter how leniently that standard is construed, the oral agreement at issue here cannot meet it. Without violating the oral agreement, SanGiacomo can (and indeed does) sell bedroom furniture, sell bedroom furniture with a high-gloss polyester finish, and sell bedroom furniture using columns and arches as design elements; the only thing SanGiacomo cannot do under the oral agreement is copy Ashley's bedroom furniture designs. The exact meaning of "copy" under the agreement was a matter of factual dispute between the parties. But although it is conceivable that Ashley could attempt to prove a meaning of "copy" that would be so broad as to turn the agreement into a substantial impairment of SanGiacomo's right to do business, therefore triggering the requirements of § 75-4, Ashley does not appear to have made this argument thus far and the normal meaning of the word does not suggest such a severe restriction. Thus we conclude that the district court erred in granting summary judgment to SanGiacomo on the contract claim based on § 75-4.

61

At oral argument, we raised the question of whether enforcement of this agreement would be inconsistent with the district court's local rules. The parties came to their agreement in the context of a court ordered mediation. The local rules governing such mediation efforts provide that "[u]pon reaching a settlement agreement at a mediated settlement conference, the parties shall forthwith reduce the agreement to writing," M.D.N.C.R. 83.10e(d) (1999), and that "the court will not permit parties in other litigations to conduct discovery regarding the mediation," id. at 83.10e(f)(8). In memoranda submitted at our request after oral argument, the parties disagreed about the applicability and significance of these rules in the instant case. The district court is certainly in a better position than we to interpret its rules in the first instance, and can, if the point is pressed, make this determination on remand.

V.

62

For the foregoing reasons, the judgment of the district court is

63

REVERSED AND REMANDED.