1 CONCEPTUAL SEPARABILITY AS CONCEIVABILITY: A

CONCEPTUAL SEPARABILITY AS CONCEIVABILITY:

A PHILOSOPHICAL ANALYSIS OF THE USEFUL ARTICLES DOCTRINE

Mala Chatterjee*

In copyright law, the useful articles doctrine plays an invaluable role in defining the limits of

“conceptual separability” has proved to be difficult to define, and the Supreme Court’s effort to

define it in the recent case Star Athletica, L.L.C. v. Varsity Brands, Inc. is unsatisfying. In an effort

to resolve this challenge, the present paper puts forth a novel test for conceptual separability, one

that draws inspiration from the philosopher’s idea of conceivability. The test is the following

question: “When you conceive of the relevant useful article as lacking the design element in

question, is the article you imagine functionally identical to the actual article?” If the answer to

this question is yes, then the design element is conceptually separable from the article’s utilitarian

aspects; if not, then the element has failed the test, and it is not entitled to copyright protection.

The present paper explores why this novel proposal avoids many of the pitfalls of existing tests

(including the Court’s own in Star Athletica), why it best achieves the aims of the useful articles

doctrine, and what questions remain once the challenge of conceptual separability has been

resolved.

INTRODUCTION…2

I. THE PATENT/COPYRIGHT DIVISION OF LABOR, THE USEFUL

ARTICLES DOCTRINE, AND THE EXISTING SEPARABILITY TESTS…5

A. Copyright, Patent, and the Useful Articles in Between…5

B. Conceptual Separability: The Old Tests…7

II. THE CONCEIVABILITY TEST…11

A. The Proposal…11

B. Applying the Test…12

C. The Camouflage Example…13

D. The Virtues of Conceivability…15

III. PHILOSOPHICAL ISSUES IN CONCEIVING…20

A. Conceivability & Necessity…20

B. Conceiving How: The Closest, Uncreative, Alternative Design…22

IV. ESSENTIALITY & FUNCTIONALITY…26

A. Essentiality…27

B. Functionality…28

V. STAR ATHLETICA & OTHER CONCLUDING REMARKS…35

A. Conceivability & Cheerleading Uniforms…35

B. Conclusion…38

INTRODUCTION

Philosophers differentiate concepts and objects—and delineate what is necessary to a

particular concept or object’s identity—on the basis of conceivability.

They ask questions of the

form: Can you conceive of X as lacking property Y? If you cannot, then property Y is necessary

to the identity of X. By way of example, a philosopher would say that because one can look at a

red, two-by-two square and conceive of it as instead being blue or four-by-four, the color and

size of the square are not necessary parts of the square’s identity. On the other hand, because one

cannot conceive of a square that is also round, being non-round is a necessary part of what it is to

be a square. This idea is both simple and intuitive, but the lens of conceivability can nonetheless

bring something novel and helpful to a problem in copyright law, one that has yet to find a

satisfying solution. This is the problem of conceptual separability.

*Joint-JD/PhD in Philosophy candidate at New York University; B.A. in Philosophy from

Stanford University, 2014. I warmly thank Professors Jeanne Fromer, David Chalmers, C. Scott

Hemphill, Barton Beebe, and Barry Friedman for their extremely helpful comments and

questions, as well as Erick Sam and my fellow members of the Furman Academic Scholars

Program in the 2016–2017 year.

See David J. Chalmers, Does Conceivability Entail Possibility?, in CONCEIVABILITY AND

POSSIBILITY 145, 145–200 (John Hawthorne & Tamar Gendler eds., 2002).

First, some background: In copyright law, the protection granted to certain pictorial,

graphic, and sculptural works is limited by the useful articles doctrine.

“Useful Articles” are

defined as objects having an intrinsic utilitarian function that is not merely to portray the

appearance of the article or convey information; for example, dresses, belt buckles, containers,

and industrial designs broadly.

The design of a useful article is considered a pictorial, graphic,

or sculptural work—protectable under copyright law—“only if and only to the extent that such

design incorporates pictorial, graphic, or sculptural features that can be identified separately

from, and are capable of existing independently of, the utilitarian aspects of the article.”

This is

to say that, if a useful article’s expressive features are inseparable from its utilitarian aspects,

those features are not entitled to copyright protection. This “separability” requirement includes

both physical and conceptual separability.

But although physical separability is relatively

straightforward, conceptual separability has proven much more difficult to pin down.

The issue of conceptual separability emerged into focus in the Supreme Court’s recent

case Star Athletica, L.L.C. v. Varsity Brands. In this case, the Court addressed the question of

whether copyright protection extends to the design elements on cheerleading uniforms; namely,

the arrangement of colors, stripes, and chevrons on the fabric.

Uniforms—and articles of

17 U.S.C. § 101.

Id.

H.R. REP. NO. 94-1476, at 55 (1976), reprinted in U.S.C.C.A.N. at 5668 (“Unless the shape of

an automobile, airplane, ladies’ dress, food processor, television set, or any other industrial

product contains some element that, physically or conceptually, can be identified as separable

from the utilitarian aspects of that article, the design would not be copyrighted”).

Star Athletica, L.L.C v. Varsity Brands, Inc., 580 U.S. _ (2017).

clothing more generally—fall squarely within the purview of the useful articles doctrine, as they

possess the intrinsic utilitarian function of covering their wearers’ bodies.

Because the design

elements in question in Star Athletica were printed directly onto the fabric of said uniforms, they

were certainly not physically separable. The Supreme Court thus articulated its own test for

conceptual separability, and, in applying it, concluded that these design elements were entitled to

protection.

However, the Court’s holding did not successfully solve the issue of conceptual

separability. We therefore still need a satisfying test in order to delineate the limits of copyright’s

reach.

In the present paper, I propose a novel test for conceptual separability that draws

inspiration from the way philosophers think about necessity and conceivability. In particular, I

argue that the correct test is the following question: “When you conceive of the article as lacking

the design element in question, is the article you imagine functionally identical to the actual

article?” If the answer to this question is yes, then the design element is conceptually separable

from the article’s utilitarian aspects; if not, then the element has failed the test, and it is not

entitled to copyright protection. I will argue that this test follows from the correct way to

interpret the text of the useful articles doctrine, that it promotes the purpose of the doctrine, and

that it possesses several important virtues that many of the existing separability tests lack.

In Section I, I give an account of the division of labor between patent and copyright law,

and the role that the useful articles doctrine plays in maintaining this division. I then proceed to

17 U.S.C. § 101 (defining “useful articles” as ones with intrinsic utilitarian functions that are

not merely to portray the appearance of the article or to convey information).

Id.

canvass the history of conceptual separability

—starting with the pre-Star Athletica tests, then

turning to the Supreme Court’s Star Athletica proposal—and highlight the problems faced by

each account. In Section II, I propose the conceivability test and illustrate how it works when

applied to example articles, and also explain why I take it to be the correct way of thinking about

conceptual separability. Section III tackles philosophical concerns regarding what precisely is

meant by conceivability in this context—such as what kind of conceivability I have in mind, and

what property I take it to be useful for guiding us to. Section IV addresses challenges faced by

the test, such as whether it is impermissibly essentialist and what “having identical functionality”

really means. In Section V, I apply the conceivability test to the articles of Star Athletica and

explain how the Court should have ruled in this case, and conclude by noting what further

questions remain.

I. THE PATENT/COPYRIGHT DIVISION OF LABOR, THE USEFUL

ARTICLES DOCTRINE, AND THE EXISTING SEPARABILITY TESTS

Before turning our attention to conceivability, we must understand the purpose, history,

and current status of the useful articles doctrine. Section I.A thus provides an overview of

division of labor. Section I.B. then summarizes the major, pre-Star Athletica tests for conceptual

separability, a number of their many problems, and the Supreme Court’s own attempt to resolve

the matter in Star Athletica.

A. Copyright, Patent, and the Useful Articles in Between

See cases cited infra notes 22-26 for the pre-Star Athletica tests of conceptual separability.

In the United States, patent and copyright law are both grounded by the Constitution’s

Progress clause, which gives Congress the power to “promote the Progress of Science and useful

Arts, by securing for limited Times to Authors and Inventors the exclusive Right to their

respective Writings and Discoveries.”

But the domains of copyright and patent are nonetheless

distinct: Patent is the province of useful inventions, whereas copyright is home to expressive

creations.

This division of labor is important to maintain, as patent and copyright are quite

different with respect to their standards for granting protection, as well as the scope of protection

that they grant. Inventions seeking patent protection are required to be useful, novel, and

nonobvious.

In exchange for meeting these requirements, the patent regime grants inventors a

total monopoly in the technology such that even independent inventors are defenseless against an

infringement claim.

This allows the inventors oto charge monopoly-high prices and maximize

returns for the twenty-year patent term.

U.S. CONST. art. I, § 8, cl. 8.

E.g., Christopher J. Buccafusco, A Theory of Copyright Authorship, 102 VA. L. REV. 1229

(2016); Jeanne C. Fromer, A Psychology of Intellectual Property, 104 NW. U. L. REV. 1441

(2010).

35 U.S.C. §§ 101–03 (requiring that patent-eligible inventions be a “new and useful process,

machine, manufacture, or composition of matter, or any new and useful improvement thereof,”

that is also not “obvious…to a person having ordinary skill in the art”).

35 U.S.C. §271 (2000); Kewanee Oil Co. v. Bicron Corp.,416 U.S. 470,490 (1974) (“While

trade secret law does not forbid the discovery of the trade secret by fair and honest means, e.g.,

independent creation or reverse engineering, patent law operates ‘against the world,’ forbidding

any use of the invention for whatever purpose for a significant length of time.”).

E.g. Mark A. Lemley, The Economics of Improvement in Intellectual Property Law, 75 TEX. L.

REV. 989, 996 (1997).

being “fixed in any tangible medium of expression.”

creative works be novel or non-obvious; they need only be original, where “originality” is

interpreted as requiring “independent creation plus a modicum of creativity.”

recognize an independent creation defense, which means that exclusive rights over works of

authorship are only enforceable against those who have “actually copied” the work in question.

But copyright protection also lasts significantly longer than a patent, as the term is typically the

whole lifetime of the author plus an additional seventy years.

Given the varying standards and

scopes of protection— carefully crafted by Congress to optimize incentives for production—it is

important that the line between patent and copyright does not get blurred. In particular, scholars

emphasize that copyright must not be abused by inventors hoping for a “backdoor patent” on

their useful invention, as such abuse would result in a significantly longer exclusive right with a

much lower standard for receiving protection.

17 U.S.C. § 102(a).

Feist Publ’ns, Inc. v. Rural Tel. Serv. Co., 499 U.S. 340, 346 (1991) (holding that information

alone without a modicum of original creativity cannot be protected by copyright, and therefore

that Rural’s telephone directory was not copyrightable).

Castle Rock, Entertainment, Inc. v. Carol Pub. Group, Inc., 150 F.3d 132, 137 (2d Cir. 1998)

(noting that establishing copyright infringement requires demonstrating actual copying and that

the copying amounts to an improper appropriation).

17 U.S.C. § 302(a).

E.g., Viva R. Moffat, Mutant Copyrights and Backdoor Patents: The Problem of Overlapping

Intellectual Property Protection, 19 BERKELEY TECH. L.J. 1473 (2004) (arguing that the

availability of overlapping protection threatens the intellectual property system by undermining

the goals of intellectual property law and disrupting the balance struck by Congress in fashioning

the copyright and patent systems).

The useful articles doctrine is a mechanism whereby copyright law screens out functional

elements of an object—including elements that are both functional and expressive— so that

functionality remains the domain of patent alone. The doctrine states that copyright law only

protects the pictorial, graphic, or sculptural features of a useful article to the extent that these

features “can be identified separately from, and are capable of existing independently of, the

utilitarian aspects of the article.”

If a design element is not physically or conceptually separable

from the article’s utility, then it will fall outside of the purview of copyright protection. The

element is then channeled to the utility patent domain, where—if it meets the higher bars of

utility, novelty, and non-obviousness—it will be eligible for patent protection.

The useful

articles doctrine is thus an important safeguard in copyright, blockading against utilitarian

elements and quelling attempts at backdoor patents.

But in order to correctly apply the doctrine,

we must determine the correct test for conceptual separability.

B. Conceptual Separability: The Old Tests

The Supreme Court’s holding in Star Athletica, LCC is where the law of separability

presently stands. But we cannot appreciate the challenges raised by this doctrine without first

examining its history. Prior to Star Athletica, a number of mysterious, problematic tests for

separability were proposed and utilized. To name only the most well-known of these tests:

Professor Paul Goldstein has argued that a design is conceptually separable “if it can stand on its

17 U.S.C. § 101.

35 U.S.C. §§ 101–03, 131.

E.g., Moffat, supra note 19, at 1500 (noting that the useful articles doctrine is a “channeling

doctrine” that directs functional works to the patent realm in order to “maintain the distinction

between the two regimes”).

own as work of art traditionally conceived, and if the useful article in which it is embodied

would be equally useful without it”;

Nimmer On Copyright claimed that conceptual separability

exists when there is “substantial likelihood that, even if the article had no utilitarian use, it would

still be marketable simply because of its aesthetic qualities”;

Keiselstein-Cord v. Accessories by

Pearl, Inc. holds that the expressive aspect of the article is conceptually separable whenever it

can be said to be “primary,” and the utilitarian function can be said to be “subsidiary”;

Brandir

International, Inc. v. Cascade Pacific Lumber Co. applied the test articulated by Robert Denicola

in his paper Applied Art and Industrial Design: A Suggested Approach to Copyright in Useful

Articles, holding that copyrightability is dependent on “the extent to which the work reflects

artistic expression uninhibited by functional considerations”;

and Judge Newman’s dissent in

Carol Barnhart v. Economy Cover Corp. proposed a finding of copyrightability where a work

“stimulate[s] in the mind of the beholder a concept that is separate from the concept evoked by

its utilitarian function.”

PAUL GOLDSTEIN, COPYRIGHT: PRINCIPLES, LAW & PRACTICE § 2.5.3, at 109 (1989).

1 MELVILLE B. NIMMER & DAVID NIMMER, NIMMER ON COPYRIGHT (1963).

Kieselstein-Cord v. Accessories by Pearl, Inc., 632 F.2d 989 (2d Cir. 1980) (holding that belt

buckles had conceptually separable sculptural elements and were therefore entitled to copyright

protection).

Brandir Int’l, Inc. v. Cascade Pacific Lumber Co., 834 F.2d 1142 (2d Cir. 1987) (holding that

functional considerations of the Ribbon bike rack overtook the aesthetic considerations, such that

the form and function of the rack became “inextricably intertwined,” and therefore precluding

Approach to Copyright in Useful Articles, 67 MINN. L. REV. 707 (1983).

Carol Barnhart Inc. v. Economy Cover Corp., 773 F.2d 411 (2d Cir. 1985) (Newman, J.,

dissenting).

However, each of these tests face problems. For instance, the Goldstein and Nimmer tests

required that judges evaluate the artistic merit and marketability of the expressive elements,

which they are arguably ill-equipped to do, and which might also introduce unnecessary

subjectivity into the analysis.

The Kieselstein-Cord test presupposed an untenable “primary”

versus “subsidiary” distinction between the aesthetic versus the utilitarian aspects of the useful

article, and without providing criteria for pinning down which is which, thereby lacking the

predictability and stability needed from a legal rule.

The Brandir “process” test rested on an

unrealistic picture of how Useful Articles are made—namely, that their design process is wholly

determined either on the basis of aesthetic considerations or on the basis of functional ones—

when, in reality, the process probably involves both.

And Judge Newman’s test faced both an

audience and a vagueness problem: Who exactly is the relevant beholder, and what does it really

John B. Fowles, The Utility of a Bright-Line Rule in Copyright Law: Freeing Judges from

Aesthetic Controversy and Conceptual Separability in Leicester v. Warner Bros., 12 UCLA ENT.

L. REV. 301 (arguing that relying on judges’ aesthetic understanding—or lack thereof—in

deciding copyright cases could lead judges to sink deep into a “quagmire of conceptual

separability”). Nimmer’s test has also been criticized by Nimmer himself on the grounds that a

“likelihood of marketability” standard is foreign to copyright, could disproportionately favor

more conventional or popular forms of art, and is very difficult to prove. NIMMER & NIMMER,

supra note 24, § 2.08[B][3], at 2-101.

See, e.g., Barnhart, 773 F.2d at 421 (Newman, J. dissenting) (arguing that Kieselstein-Cord

“offers little guidance to the trier of fact, or the judge endeavoring to determine whether a triable

issue of fact exists, as to what is being measured by the classifications ‘primary’ and

‘subsidiary’”).

Judge Winter’s dissent argued against the Brandir test on the grounds that it makes protection

depend on “largely fortuitous circumstances concerning the creation of the design in issue.”

Brandir Int’l, Inc., 834 F.2d at 1151 (Winter, J., dissenting). Others have argued that the

approach has no basis in statute or legislative history, and is “out of step” with the rest of

Conceptual Separability and Copyright in the Designs of Useful Articles, 37 J. COPYRIGHT

SOC’Y U.S.A. 339, 372 (1990).

mean to “stimulate” two separate concepts in his mind, rather than, say, one united concept?

sum, given all of these (as well as other) concerns, it is unsurprising that courts have been unable

to uniformly embrace one proposal.

Thus, the state of the useful articles doctrine has historically been problematic.

Some

circuits have shared the same test for conceptual separability, but others, like the Second Circuit,

have had so many different and potentially incompatible tests that it is not even clear which one

predominated within that circuit.

This has led to inconsistency and unpredictability, as well as a

litigation-generating effect with more extreme sorts of cases.

Moreover, the lack of clarity with

respect to where the law stands has distorted its ability to produce the incentives it aims to

produce. Copyright law has thus long needed a test that can be universally embraced by courts,

as well as depended on by plaintiffs and defendants alike.

The Supreme Court set out to resolve these challenges in Star Athletica LLC.

Rather

than choosing to apply any one of the aforementioned tests, the Court instead opted for a much

more minimal analysis. In particular, it held the following:

The majority in Barnhart argued against Judge Newman’s test on the grounds that it is “so

ethereal as to amount to a ‘non-test’ that would be extremely difficult, if not impossible, to

administer,” and that almost any utilitarian article may be viewed by some as art depending on

how it is displayed, citing examples of a Campbell’s Soup can or a pair of scissors. Barnhart,

773 F.2d at 419 n.5.

See Varsity Brands, Inc. v. Star Athletica, LLC, 799 F.3d 468, 484 (6th Cir. 2015)

(summarizing nine existing tests for conceptual separability and describing the resultant circuit

split).

Id.

E.g., Brief for Intellectual Property Professors as Amicus Curiae in Support of Petitioner at 21,

Varsity Brands v. Star Athletica (2016) (No. 15-866).

Star Athletica, L.L.C v. Varsity Brands, Inc., 580 U.S. _ (2017).

[A] feature incorporated into the design of a useful article is eligible for copyright

protection only if the feature (1) can be perceived as a two- or three-dimensional work of

art separate from the useful article and (2) would qualify as a protectable pictorial,

graphic, or sculptural work—either on its own or fixed in some other tangible medium of

expression—if it were imagined separately from the useful article into which it is

incorporated.

This account of conceptual separability has become the law of the land, thereby purporting to

dissolve the aforementioned inconsistency and unpredictability. But, nonetheless, it is also

crucially mistaken. I will elaborate on the Star Athletica, LCC test and its problems – ones which

the conceivability test itself successfully avoids – in the section below.

II. THE CONCEIVABILITY TEST

A. The Proposal

Conceivability—broadly understood as the imaginative ability “to represent a scenario

involving actual or non-actual things in actual or non-actual configurations”

—is a

philosophically significant notion, traditionally taken to be a guide to possibility and necessity.

The thought is that, if a certain configuration of an object is conceivable, then this is a “possible”

configuration; conversely, if an object cannot be conceived of as lacking a certain property, then

this property is necessary to the object in question.

We can see conceivability in action when

Id. at 1, 10.

Tamar S. Gendler & John Hawthorne, Introduction: Conceivability and Possibility, in

CONCEIVABILITY AND POSSIBILITY 1,1 (Tamar S. Gendler & John Hawthorne eds., 2002).

Id.

Id. There is a philosophical issue here regarding what kind of necessity I have in mind for the

purposes of the conceivability test. This issue is addressed in Section IV.

we look at examples. For instance, we can conceive a certain blond-haired human as having red

hair. We can imagine humans with any number of different hair colors, in fact, because hair

color is a contingent property. On the other hand, we cannot conceive of a human that is not an

organism. We might conceive of things that look and act like humans—perhaps that are even

qualitative duplicates of humans—which we might call robots, or something similar. But these

nonetheless cannot be actually human, because being human necessarily involves being an

organism.

The conceivability framework is thus well-suited for conceptual separability, for it can

tease apart what design elements are necessary to the functionality of the article. In other words,

once we have determined the function of the article, we can utilize the lens of conceivability to

determine if a certain expressive element is necessary to the article’s performance of that

function. The conceivability test is thus as follows: When you conceive of the article as lacking

the design element in question, is the article you imagine functionally identical to the actual

article? If the answer is yes, then the design element is not necessary to the article’s function. It

is conceptually separable from the article’s utilitarian aspects, and it is entitled to copyright

protection. On the other hand, if the answer is no, then the design element has failed the test and

falls out of the purview of copyright law.

To further elucidate the distinction between contingent and necessary properties as illustrated

by conceivability, consider the following examples: 1) I can conceive of the pain I am

experiencing in my left arm being instead located in my right arm, but I cannot conceive of the

pain I am experiencing being such that I have it without feeling it, because feeling a pain is

necessary to having a pain. 2) I can conceive of my pet black poodle Bertrand being a white

poodle, but I cannot conceive of Bertrand being a plush toy poodle while still being himself,

because being an actual dog is necessary to being Bertrand. 3) I can conceive of the number 2

being denoted by the symbol “&” instead of the symbol “2,” but I cannot conceive of the number

2 not being a prime number, because being a prime number is necessary to being the number 2.

B. Applying the Test

Before canvassing the virtues of this proposal, and to help us get a better sense of how the

proposal works, let us apply it to particular examples of useful articles from actual separability

cases to see what outcomes it generates. First, pulling from Kieselstein-Cord v. Accessories by

Pearl, imagine a belt buckle, a utilitarian object, that is also a sculptured design “cast in precious

metals – decorative in nature – and used as jewelry.”

The question presented is whether this

sculptural design is entitled to copyright protection. Applying the conceivability test, the right

question to ask is following: In conceiving of the same belt buckle but without the decorative,

sculptural design, do we conceive of something that has the same functionality as the existing

belt buckle? The answer in this case is yes. The conceived-of buckle is, though less beautiful,

still able to be worn, to be fastened, to perform all of the actual buckle’s functions. The design

thus passes the conceivability test and is entitled to protection, which is what the Kieselstein-

Cord court held

(but, unfortunately, using the mysterious primary/subsidiary test).

On the other hand, consider the articles of Carol Barnhart, Inc. v. Economy Cover Corp.

This case regarded the design of human torso mannequins, used for the purpose of displaying

clothes as they would look when worn.

The mannequins were designed to imitate the muscular

chest of a man, and the breasts and waist of a woman; the question at hand was whether these

elements were conceptually separable from the mannequin’s functionality. Applying the

Kieselstein-Cord v. Accessories by Pearl, Inc., 632 F.2d 989, 990 (2d Cir. 1980).

Id.

Carol Barnhart Inc. v. Economy Cover Corp., 773 F.2d 411, 419 (2d Cir. 1985) (holding that

the mannequins were not copyrightable because their design as “inextricably intertwined with the

utilitarian feature” of displaying clothes).

conceivability test, we start by conceiving of the mannequins without the design elements, and

we then ask whether the articles we conceive of are functionally identical to the actual articles. In

this case, the answer is no. Given that the mannequin’s function is to show how clothing will

look when worn by actual human beings, the body-like design elements are necessary for

performing this function. If the mannequin did not have this design, it would instead be shaped

as a flat rectangle, incapable of emulating the look of the clothing on an actual person’s figure.

Therefore, these design elements of the mannequins are not entitled to copyright protection.

C. The Camouflage Example

It is worth noting that an implication of the conceivability test is that the same design

element might be conceptually separable on one object but not on another. For instance, consider

the example of a camouflage design, which was discussed in the oral arguments heard by the

Supreme Court for Star Athletica.

This camouflage design might be printed on a military

uniform, or it might instead appear on a fashion accessory such as a purse. In the context of the

military uniform, the design would not pass the conceivability test. If we conceive of the uniform

without the camouflage print, then what we imagine is not functionally identical to the actual

uniform because it is not able to successfully hide the soldier in the woods.

Given that

concealing the soldier is one of the primary functions of the uniform (along with protecting the

Transcript of Oral Argument at 4, Star Athletica, L.L.C v. Varsity Brands, Inc. (2016) (No. 15-

866).

Id. at 23.

soldier’s body from the elements), the camouflage print is clearly not conceptually separable in

this context.

On the other hand, a fashion purse with a camouflage print does not do anything to

conceal the wearer of the purse, nor does it aim to itself be concealed. Such a print is placed on a

purse precisely to be seen, to “portray the appearance” of the print itself, for purely aesthetic and

expressive purposes.

In this context, the print is analogous to a polka dot or a paisley, or any

other stylish pattern printed on a fabric. And if we imagine the purse in question without the

print, then what we imagine is functionally identical to the original purse: It still holds the

wearer’s belongings and can be worn by the wearer over her arm. The camouflage in this context

is thus a purely expressive design choice, conceptually separable from the functionality of the

purse.

D. The Virtues of Conceivability

As simple and minimal as the conceivability test might seem, it has several important

virtues lacked by many of the previous tests. First, the conceivability test is in accordance with

the statutory text. Recall that the design of a useful article is protectable “only if and only to the

extent that such design incorporates pictorial, graphic, or sculptural features that can be identified

separately from, and are capable of existing independently of, the utilitarian aspects of the

article.”

Under the conceivability test, if the “utilitarian aspects of the article” are so

See, e.g., Fulmer v. United States, 103 F. Supp. 1021 (Cl. Ct. 1952) (finding no copyright

infringement when the government produced parachutes printed with the plaintiff’s camouflage

design because of the useful articles doctrine).

17 U.S.C. § 101. (“A ‘useful article’…[has] an intrinsic utilitarian function that is not merely

to portray the appearance of the article or to convey information.”)

Id.

intertwined with the design element that they are modified when we conceive of the two pulled

apart, then it follows that those design elements cannot be identified separately from the article’s

useful aspects. To put it another way, conceivability allows us to distinguish which elements of

the article are actually part of the article’s functionality. By focusing on the persistence of the

article’s utility when stripped of design elements, rather than on the persistence of the design

element’s aesthetic value when stripped of utility, the test draws our attention to what the statute

is trying to safeguard against.

The statute does not ask us whether the pictorial, graphic, or

sculptural (PGS) features have artistic worth once separated, or whether they could stand on their

own as a copyrightable work.

It instead specifically asks whether the PGS features can exist

independently from the utilitarian aspects. Therefore, the correct test for conceptual separability

must inquire about the PGS features’ relationship to these utilitarian aspects.

This reading of the statute is undoubtedly aligned with the purpose of the useful articles

doctrine; namely, assuring that no functional elements—including elements that are both

This approach has been adopted by a number of courts. See, e.g., Whimsicality, Inc. v. Rubie’s

Costume Co., Inc., 891 F.2d 452 (2d. Cir. 1989) (holding that the decorative elements of clothing

are particularly unlikely to meet the test of conceptual separability because they are intrinsic to

the decorative function of the clothing, rather than asking the question of whether the elements

could stand on their own as a work of art), and Eliya, Inc. v. Kohl’s Dept. Stores, No. 06 Civ

195(GEL), 2006 WL 2645196 (S.D.N.Y. Sept. 13, 2006) (holding that the design elements of the

shoe are unprotected because they all serve a functional purpose on the shoe without discussing

whether or not they would be aesthetic works in their own right when stripped of functionality).

Other courts have taken the opposite approach, however. See, e.g., Universal Furniture Int’l, Inc.

v. Collezione Europa USA, Inc., 618 F.3d 417 (4th Cir. 2010) (holding that the design elements

of the furniture pieces are aesthetic objects in their own right and therefore protected, without

discussing whether the furniture itself would be functionally different when stripped of the

ornamentation), and Galiano v. Harrah’s Operating Co., Inc., 416 F.3d 411 (5th Cir. 2005)

(holding that the uniform designs are unprotected because they are unable to “moonlight as a

piece of marketable artwork,” without considering whether they are necessary to the

functionality of the uniforms).

17 U.S.C. § 101.

expressive and functional—receive “backdoor patent” copyright protection.

If we read the

statute such that it only asks whether the PGS features can stand as works of art on their own,

then PGS features which do count as works of art, but which are also necessary for the

functionality of the article (such that their removal diminishes functionality), will receive

statute presupposed by the conceivability test successfully maintains the division of labor

between copyright and patent. If a design element has some functional content and therefore fails

the conceivability test, then it will not be granted protection, regardless of whether it could be a

work of art on its own. This forecloses copyrights in elements wielding utilitarian power, leaving

the business of utility monopolies squarely in the province of patent law.

In fact, failure to accord with this purpose of the useful articles doctrine is precisely

what’s wrong with the Supreme Court’s own proposal, as put forth in Star Athletica. Recall that

the Court articulated the following test for separability:

[A] feature incorporated into the design of a useful article is eligible for copyright

protection only if the feature (1) can be perceived as a two- or three-dimensional work of

art separate from the useful article and (2) would qualify as a protectable pictorial,

graphic, or sculptural work—either on its own or fixed in some other tangible medium of

expression—if it were imagined separately from the useful article into which it is

incorporated.

E.g., Moffat, supra note 19.

Star Athletica, slip op. at 1, 10.

In a sense, this test makes use of an insight underlying the conceivability test; that, in

simply imagining the feature in question separately from the useful article, we can arrive at

genuine insight into its copyrightability or lack thereof. However, what the Supreme Court’s test

gets wrong is what it asks us about what we imagine. By focusing on whether the feature would

count a PGS work when separated from the article, the Court’s test pays no attention to whether

the feature is useful at all, or whether the functionality of the article would be the same without

it. Again, this is problematic because a feature could both qualify as a PGS work as well as have

functionality. Thus, under the Court’s test, it is entirely possible for creators to receive backdoor

patents in functional features, as long as they can also meet the notoriously thin requirements of

originality and a modicum of creativity.

Star Athletica’s test thus does not promote the useful

articles doctrine’s aim of keeping the domains of copyright and patent separated. It instead

undermines this purpose, opening the floodgates for backdoor patents. The conceivability test, on

the other hand, successfully preserves this desired division of labor.

Next, the conceivability test minimizes many of the challenges faced by the pre-Star

Athletica test proposals. For instance, this test does not require that judges make aesthetic

judgments about the artistic merit of the design elements as works of art, nor does it require them

to make empirical judgments about their marketability without utility, the process that was

behind their creation, or whether the aesthetic elements are perceived by viewers as primary or

secondary to the utilitarian ones. To the contrary, the conceivability test only asks judges to make

a metaphysical judgment about whether the design element is necessary to the utility. Luckily,

See, e.g., Bleistein v. Donaldson Lithographing Co., 188 U.S. 239, 251 (1903) (noting that

originality creativity in copyright law need only pass “the narrowest and most obvious limits,” as

persons trained only in the law are not equipped to evaluate creative works beyond this low bar).

judges make such metaphysical judgments all the time: To name only a few, they are tasked with

determining whether a certain action caused a certain effect, whether something was actually

copied from another work, and whether one expressive feature of a work is identical to a feature

of another. Tasking a judge with the conceivability test is thus asking her to make a kind of

judgment she is already equipped and responsible for making.

Further, compared to the tests of Nimmer, Goldstein, Kieselstein-Cord, and Judge

Newman, the conceivability test does not raise as challenging of an “audience” problem. If we

ask the questions of whether some expressive element is a work of art, is marketable, is a

“primary” rather than “subsidiary” feature, or triggers two separate concepts rather than one,

then we will likely get different answers depending on who we ask. An artist or designer will

have different judgments than a mere appreciator of art or design, who will see things differently

still from someone uninitiated or uninterested in aesthetic works. Of course, one might insist that

all separability tests will necessarily raise an audience problem: After all, who gets to decide

whether the element is conceptually separable, and why is their point of view the favored one?

But the audience problem faced by the conceivability test seems to be comparatively less severe.

When asked the question of whether something is “conceivable,” people are much less likely to

disagree. Conceivability is a property so closely tied to logic, to the imaginative limitations

shared by us all, that only in controversial cases will opinions seriously diverge.

Sidestepping

One might worry that this argument overlooks the fact that there are many different kinds of

conceivability. In order for there to really be uniformity in different applications of the

conceivability test, it is important that all those applying the test have precisely the same kind of

conceivability in mind. This issue is addressed in Section IV.

the more daunting challenge of designating some particular point of view thus minimizes a

problem—both theoretical and practical in nature—that many other tests cannot avoid.

Another virtue of the conceivability test is that it treats conceptual separability as

analogous to physical separability. This virtue is contrasted against the many existing,

cumbersome tests asking us something that looks and operates totally differently from the

physical separability inquiry.

In the case of physical separability, the question to ask is whether

we can actually take away the design element from the article without affecting the article’s

utility.

Following in line, the conceivability test asks if we can imagine the article’s utility

unaffected with the design element taken away. The similarity between these two tests is both

conceptually satisfying and practically convenient; and, by linking the conceptual separability

reasoning processes to the processes of the straightforward physical case, the conceivability test

promotes uniformity and predictability.

The preceding discussion helps clarify the way the conceivability test works and why it is

desirable as an alternative. But it does not clarify everything. There are important remaining

This is not to say that there isn’t still some audience problem in the conceivability test: It is

true that we must still always ask whose point of view is the one from which we should ask the

separability question. It also might turn out that all of the difficult separability cases are

controversial enough for opinions to divulge. Nonetheless, it is reasonable to posit that the range

of disagreement produced by the conceivability question will be comparatively more constrained

than it is in the question of aesthetic value or marketability.

See Brief for Intellectual Property Professors as Amicus Curiae in Support of Petitioner, supra

note 34, at 9 (arguing that the correct test for conceptual separability considers it to be “a kind of

a coda to physical separability”).

COMPENDIUM III § 924.2(A) (“Physical separability means that the useful article contains

pictorial, graphic, or sculptural features that can be physically separated from the article by

ordinary means while leaving the utilitarian aspects of the article completely intact.”).

questions requiring what precisely is meant by conceivability, and what sorts of conceptions I

have in mind. I answer these questions in the following section.

III. PHILOSOPHICAL ISSUES IN CONCEIVING

Conceivability is a complex and philosophically loaded property, and we must therefore

pin down how exactly it is to be understood in this test. I have argued that conceivability is an

appropriate lens for separability because it is a guide to necessity, and can thus illuminate

whether an element is necessary for an article’s functionality. But, as philosophers have

recognized, there are multiple types of necessity—for instance, “logical,” “metaphysical,” or

“nomological” necessity—and so we must clarify which type we are interested in illuminating.

in Section III.A briefly addresses this issue. A further, more pressing issue is that there are a

number of ways in which one can conceive of some article without some design element. For

instance, one might imagine the article with the element ripped away, or with a replacement

element, or instead with an additional, unrelated modification. The problem with this range of

possibilities is that different conceptions might yield different answers to the test. Thus, we must

specify which is the correct way of conceiving. This question is answered in III.B.

A. Conceivability & Necessity

The philosopher’s first question is one that might not plague lawyers and judges, but we

are nonetheless able to quickly set it aside. Recall the idea that, if a certain configuration of an

object is conceivable, then this is taken to be a “possible” configuration.

There are a number of

See, e.g., Kit Fine, Varieties of Necessity, in CONCEIVABILITY AND POSSIBILITY (Tamar S.

Gendler & John Hawthorne eds., 2002).

E.g., Gendler & Hawthorne, supra note 37. Many philosophers argue that conceivability is not

a perfect guide to metaphysical possibility, as there are some propositions that are conceivable in

some sense but are not metaphysically possible. For example, it is prima facie conceivable that

both Goldbach’s conjecture, as well as its negation, are true, even though one of these is

ways in which we can understand possibility. For instance, metaphysical possibility is a

understood as a particularly broad notion: It is arguably broader than nomological or natural

possibility, which refers only to what is possible in worlds in which the laws of nature are the

same as those in the actual world.

The idea here is that certain things are both conceivable and

metaphysically possible which nonetheless violate the laws of nature. For instance, it is

conceivable and metaphysically possible for a ball dropped from a rooftop in some possible

world to not fall at all, even though the laws of nature of the actual world are such that the ball

would fall (and accelerate at a rate of 9.8 meters per second squared).

This technical distinction pushes us to clarify the correct understanding of the

conceivability test, because it shows that a certain notion of conceivability would be too broad

for our purposes. To see this, consider a useful article that has some functional element X which

is necessary given the laws of nature, but which would not be necessary in certain worlds with

different laws. Suppose that element X enables the article to function whilst handling

gravitational effects, and it therefore would not be necessary if there were no gravity. In this

case, one might worry that, if we allow ourselves to conceive of the article without element X in

worlds with no gravity, then element X would pass the conceivability test. However, intuitively,

this element is still functional, as it does serve an important functional role in the actual object

(namely, that of handling gravitational effects).

The preceding consideration highlights that what we are concerned with in the useful

articles doctrine is not which elements are necessary to the article’s functionality in just any

metaphysically impossible. Nonetheless, philosophers still tend to think that for the most part

conceivability is a reliable guide to possibility. See, e.g., Chalmers, supra note 1.

Fine, supra note 58.

possible world, but crucially, only in worlds where the laws of nature are the same as the actual

world. Thus, the kind of conceivability deployed in this test is narrowly tailored in a certain

sense, as it should serve as a guide to nomological necessity specifically, and not the broader

notion of metaphysical necessity.

This is to say that the conceivability test should be

understood as employing a conditional conceivability, whereby we restrict ourselves to only

imagining worlds close enough to ours such that they have the same laws of nature as the actual

world. The more precise formulation of the test can thus be stated as follows: “Holding the laws

of nature fixed, when you conceive of the article as lacking the design element in question, is the

article you imagine functionally identical to the actual article?”

B. Conceiving How: The Closest, Uncreative, Alternative Design

It is worth pointing out that some philosophers do hold that the laws of nature are

metaphysically necessary. See, e.g., Fine, supra note 58 (arguing for the “weak” metaphysical

necessity of laws of nature); Sydney Shoemaker, Causality and its Properties, in TIME AND

CAUSE 109–35 (Peter Van Inwagen ed. 1980) (arguing for the “strong” metaphysical necessity of

laws of nature); Christopher Swoyer, The Nature of Natural Laws, in AUSTRALIAN JOURNAL OF

PHILOSOPHY, 203–23 (same). Weaker versions of this view hold that the necessity arises for

Kripkean reasons. The thought here is that the reference of terms such as "mass" are fixed a

posteriori to a certain specific property in our world, such that worlds with different laws do not

contain mass, and therefore, that the laws actually governing mass are metaphysically necessary.

See generally SAUL KRIPKE, NAMING AND NECESSITY (1980). Stronger versions of this view

argue that the laws of nature must be metaphysically necessary in order for there to be a robust

difference between the laws of nature and merely true accidental generalizations. For these

philosophers, metaphysical and nomological necessity are, in some sense, one and the same,

which might initially seem to entail that these philosophers do not face with the preceding

paragraph’s challenge requiring a narrower conception of conceivability. However, this only

follows if these philosophers also accept that conceivability is a reliable guide to possibility. It is

implausible to think the laws of nature are necessary as well as that conceivability is a reliable

guide to possibility, because we can easily conceive of possible worlds in which the laws of

nature are different from our own (such as the gravity-free world imagined above). Thus, for the

purposes of our test, we still need a narrower conception of conceivability that serves as an

epistemic guide only to the worlds that have the same laws of nature as our own, and not to the

possible—or impossible—worlds in which the laws are different.

The second philosophical question is much more practically pertinent, and it demands

that we specify the manner of conceiving. Given the range of possible conceptions, it might be

that certain conceptions will tell us that the element we are considering is functionally neutral,

whereas other conceptions will hold that it is not. To see this, consider the example of a

creatively shaped pocket on a shirt. We could conceive of this shirt as though we have simply

ripped the pocket from it (in which case what we imagine would not be functionally identical to

the actual shirt, as it would lack the “carrying around small objects” functionality); or, we could

conceive of the shirt with a pocket of the same size and location, but with a wholly uncreative

square shape (in which case what we imagine would be functionally identical, as it would still be

able to carry around small objects). Instead, we could imagine a very different pocket replacing

the original one (which may or may not pass the conceivability test, depending on the size or

usefulness of this new pocket); or, finally, we could imagine the pocket to be designed in very

different way, in a different size and with a different, unique shape (which also may or may not

pass the conceivability test depending on the nature of the new design). In light of these various

outcomes, it is clear that we need a constraint on what sort of conception is called for by this test,

so that it yields consistent applications and results.

At this point, it is important to remember that the conceivability test only asks us to

imagine away the design element itself. It does not ask us to imagine away the part of the article

that the design element is on, nor does it ask us to imagine the article with some additional

component or design. With this in mind, we can see that the first conception noted above—

namely, ripping the whole pocket from the shirt—gets things wrong because it removes more

than just the design of the pocket. The third conception—which adds a new and different pocket

to the article—also goes beyond the test by doing more than merely removing the design.

Similarly, the fourth conception replaces the relevant design with a different creative design, and

it thus also adds something that the conceivability test does not require. The second approach,

however, simply strips the pocket of its unique design. The pocket in this conception is the same

pocket, only without the design in question, and also without the addition of anything further.

This is thus the approach that accords with the demands of our test.

At this point, we might wonder if we can always imagine away only and all of the design

element in question. For one thing, doing so will require a precise account of what exactly

constitutes the design element. This alone should not be problematic, because we already need

such an account in order to understand the copyright claim itself. But there is a further, more

challenging worry here: Once we’ve stripped the article of the creative design, we might then

need to replace it with some alternative design, in order for the article to continue to exist at all.

The issue of alternative designs arose in the preceding pocket example, in which I

deployed the notion of an uncreative square-shaped pocket. As noted above, the conceivability

test does not ask us to replace the design under consideration with an alternative creative design.

This is an important limitation on our conceptions. Otherwise, there might be cases where the

design element in question—call it element A—really is functional and therefore should not pass

the test, and yet it does pass because we are improperly conceiving of the article with some

alternative, creative element B that is itself functional (in fact, functionally equivalent to element

A) and that therefore substitutes the functionality of element A. Thus, to avoid such cases, our

conceptions must be limited so that we may only replace the creative design element with an

uncreative one, such as the basic square shape of the pocket described above.

In addition to uncreativeness, we need one more constraint on our conceptions in order to

avoid the “functional supplement” worry above; namely, the constraint of closeness.

To see

why, consider the pocket example once again. In conceiving of the pocket with only an

uncreative square shape, we might imagine it as twice the size of the original pocket (A), half the

size of the original pocket (B), or the exact same size as the original pocket (C). All three of

these conceptions involve uncreative alternative designs, but only C is functionally identical to

the original pocket; A has more functionality, whereas B has less. Under conceptions A or B, the

design element would fail the conceivability test. But this is problematic, because the change in

functionality is clearly not attributable to the loss of the design element itself. Conceptions A and

B both deviate from the original pocket on an unnecessary dimension, thereby modifying the

article in an unnecessary and functional way. This deviance eliminates the control in our

conception experiment; it sabotages our efforts to test the specific element at hand. Luckily, such

sabotage can be easily avoided. If we restrict our conception to the alternative design closest to

the actual one—which, in this case, would be C—then we prevent unnecessary deviance from

interfering with our inquiry.

The complete constraint on our conceptions is thus as follows: In conceiving of the article

without the relevant design element, we must replace the creative design element with the closest

uncreative alternative design. This constraint is both effective and intuitive. If we modify the

article in our conception in unnecessary ways, then this creates the possibility that we have added

or changed something that compensates for the function of the element lost; but if we have

compensated for lost functionality, then the element lost must itself be functional. On the other

For a discussion elaborating the idea of “closeness” of possible worlds, see DAVID K. LEWIS,

COUNTERFACTUALS (Oxford 2001).

hand, straying too far from the original article could result in functional modifications for which

the element itself is not to blame, thereby failing elements which should pass the conceivability

test. The range of permissible conceptions can leave no room for such functional slight of hand,

and the constraints of uncreativeness and closeness achieve this goal.

One might wonder if the notion of “uncreative designs” needs further explication. But

creativity already plays a central role in copyright law, and it is therefore not something requiring

elaboration uniquely for this test. As explained in Section I, in order for an expression to be a

candidate for copyright protection to begin with (whether or not it is a part of a useful article), it

must first fulfill the requirement of possessing a modicum of creativity.

In Feist, for example, it

was held that a phonebook, organized by alphabetization, was not sufficiently creative to meet

the modicum of creativity requirement.

The question of whether a certain design element will

qualify as creative will depend on the context and article in question, and so we will not say

more about it here. For our purposes, it is enough to note that I use the notion of “uncreative

designs” to refer specifically to ones not entitled to copyright protection, because they fall short

of a modicum of creativity.

I have argued that restricting ourselves to the closest uncreative alternative design will

prevent us from imagining articles with functional alterations, so that we can clearly identify the

functionality (or lack thereof) of the design element at hand. But we are still left with a

substantial challenge: functionality itself.

IV. ESSENTIALITY & FUNCTIONALITY

E.g., Feist Publ’ns, Inc. v. Rural Tel. Serv. Co., 499 U.S. 340, 340 (1991).

Id.

The preceding section has clarified how to understand conceivability. But in order to

apply the conceivability test, we must understand what it means for articles to be functionally

identical. This challenge is taken up in the present section. First, Section IV.A responds to the

objection that the conceivability test understands functions in an impermissibly essentialist or

idealistic way. Section IV.B then turns to the challenge of defining articles’ functions, ultimately

concluding that all tests for separability—and not just the conceivability test—are faced with this

challenge. It is therefore the challenge we should turn our attention to next.

A. Essentiality

One might object that the conceivability test is problematically essentialist. Is it really the

case that design elements either will or will not be essential to the functionality of a certain type

of article? To illustrate this worry, consider one commonly encountered example: a table.

Initially, we might think that a four-legged design is essential to the functionality of a table, as a

table would not be able to perform its function of supporting objects without such legs. But this

thought is quickly undermined when we conceive of a triangle-shaped, three-legged table,

serving the function of supporting objects just as well as its four-legged counterparts. We can

also conceive of a long, thin, two-legged table, or even a square-shaped table supported by one

large leg at its very center. Given this variety in possible, functional table forms, does it make

sense to think of any of the particular legs of a table as essential to its functionality? If we did

think in this way, then which leg(s) would be the essential ones?

The problem with this worry is that it hinges on a misreading of what the test really

demands, and the level of generality at which it operates. The conceivability test is to be applied

at the level of particular useful articles: It does not ask us to consider the kind of article at the

platonic level, or about what properties are essential to the functionality of “tables” abstractly

construed. Instead, the test asks us what is necessary to the functionality of the particular article

in question. When examining particulars, the task of determining what is necessary to their

function is much easier than in the context of abstract object types. We do not need some

background conception of what is essential to tables or backpacks or watches generally; we must

only reflect on the table or backpack or watch at hand. This entails that, although a certain design

element might not be necessary for the functionality of all backpacks, it could still be necessary

for the function of the particular backpack in question. But this particular backpack is the true

object of our inquiry. Thus, we can and should apply the test on the design elements of said

backpack, without any commitments about what all backpacks essentially require.

B. Functionality

Nonetheless, even at the level of particulars, we are still faced with the challenge of

defining an article’s function. How fine-grained should our notion of functionality really be?

Consider the example of one hookah water container, a useful article shaped in a particularly

distinctive and creative way, as discussed in Inhale, Inc. v. Starbuzz Tobacco, Inc.

In applying

the conceivability test, we ask: When we conceive of this hookah water container without this

distinctive and creative shape—and instead with the uncreative shape of a straightforward tube—

do we imagine something functionally identical to the actual hookah water container? Clearly,

the answer here depends on the fineness of grain of our specification of the function. If the

See, e.g., Margot E. Kaminski & Guy A. Rub, Copyright’s Framing Problem, 64 UCLA L.

REV. 1102, 1129 (2017) (discussing the widespread “framing” problems in copyright law and

noting that, in useful articles doctrine cases, the court is faced with a framing problem in

deciding whether to zoom in by considering each possible utilitarian function of parts of the

work, or zoom out by focusing on the main functionality of the work as a whole).

Inhale, Inc. v. Starbuzz Tobacco, Inc., 755 F.3d 1038 (9th Cir. 2014) (holding that the shape of

a hookah water container is not entitled to copyright protection “because the shape accomplishes

the function” of holding the water).

function of the container is merely “containing water,” then the imagined container—stripped of

the distinctive and creative design—is functionally identical to the actual container. But if the

function is defined more specifically, say, as “containing a certain amount of water in a certain

shape,” then the imagined container has different functional properties than the actual one. Given

this difference in outcome of the two inquiries, it is clear how much hinges on how we specify

the function. So, how do we determine the right specification?

A different strand of the worry with pinning down an article’s function is that we limit

ourselves to legitimate functions, such that not every possible use of the article is deemed a part

of its functionality. After all, useful articles likely could be used in a variety of ways, including

ways in which they are not designed or intended to be used. A pair of sneakers could be used as a

doorstopper, or as a paperweight, or as a hammer to pound in a nail. A pair of socks could be

used as a handkerchief. Should these possible but peripheral functions factor into our separability

analysis? Certainly not. After all, such an overly permissive definition of “utilitarian aspects”

would take us through a downward spiral, seemingly implying that basically every article—

useful or not—has countless utilitarian aspects, and that basically every property of said article is

relevant to one of these aspects or another. But the mere existence of a distinction between

utilitarian and non-utilitarian aspects suggests that this overly permissive definition of “utilitarian

aspects” is not the correct way to interpret the statutory text. Thus, in our application of the

conceivability test and our determination of the article’s functional identity, our inquiry should

be limited to its central, legitimate utilitarian aspects, and not all of the many possible functions

that the article could be used to perform.

This is not to say that determining an article’s central, legitimate functions will always be

straightforward. The question of what counts as a function poses great challenges in copyright as

well as design patent and trademark law, all of which channel functional elements (albeit, each

with slightly different understandings of “functionality”) away from their spheres of protection

and towards the utility patent regime.

The challenge was notably discussed in the oral

arguments heard by the Supreme Court in Star Athletica.

Exploring the ideas of Chris

Buccafusco and Jeanne Fromer in Fashion’s Function in Intellectual Property Law,

the Court

inquired about whether the function of a uniform is not merely to clothe the cheerleaders and

convey the fact that they are cheerleaders, but also, to make the cheerleaders look good by

altering the perception of their bodies in a flattering way. If “looking good on” counts as one of

these article’s central and legitimate utilitarian aspects, then the design elements of the uniforms

that make the article look good on—namely, the stripes and chevrons positioned in a flattering

and slimming way—would not pass the conceivability test. The outcome thus turns on our

understanding of the article’s legitimate function. But is it implausible to think that legitimate

functions can be so defined?

In one effort to dissolve this concern, we might argue that the legitimate functions should

be delineated on the basis of the purpose of the article. For example, the purpose of a shoe is not

to hold doors open, but rather, to enclose, protect, and support the wearer’s foot; therefore, being

a doorstop is not a part of the shoe’s legitimate function. This purposive understanding has

See supra Section I.

See Transcript of Oral Argument at 4, Star Athletica, LCC v. Varsity Brands, Inc. (2016) (No.

15-866).

Christopher Buccafusco & Jeanne C. Fromer, Fashion’s Function in Intellectual Property

Law, 93 NOTRE DAME L. REV. 51, 53 (2017) (arguing that clothing design elements which

perform the function of making the wearer look good should not be entitled to copyright

protection).

intuitive appeal, but it faces its own challenges. For instance, whose understanding of the

purpose is relevant for our analysis? If the relevant perspective is the article’s designer, then

“purpose” might turn out to be under-inclusive of functional aspects. To see this, imagine that an

artist creates a certain sculpture. It later turns out that the shape of the sculpture is particularly

well-suited for some legitimate utilitarian purpose, but it is one that the artist had not herself

contemplated. In this case, even though the artist did not create the design with this useful

purpose in mind, we would still not want to grant her a copyright in said utility. Such a grant

would undermine the statutory language of the useful articles doctrine,

because the sculpture’s

shape really is utilitarian, even if the artist herself did not know this whilst designing it. It would

also undercut the policy behind the useful articles doctrine, as this copyright would prevent the

rest of society from benefiting from a genuinely useful design.

On the other hand, if the relevant point of view for determining the purpose is that of the

article’s consumer, then “purpose” might become unreasonably expansive. Imagine that an artist

creates a particular abstract design for t-shirts, and it turns out that this design is extraordinarily

good at scaring away kittens. Although kitten-scaring is a function that, we can imagine, very

few people would see as legitimate, we can also imagine that some customers with particularly

strong aversions to kittens would find this to be extremely useful. These individuals might

therefore insist that the design should not be entitled to copyright protection, even though the

vast majority of us would not deem this a legitimate function of the design, instead seeing it as

purely aesthetic. If we do not want such an idiosyncratic function to count as legitimate, then we

cannot delineate legitimate functions on the basis of just any customer’s purpose.

See 17 U.S.C. § 101 (withholding copyright protection from the “utilitarian aspects” of an

article).

So we are clearly faced with a challenge in delineating legitimate functions. But this

challenge is universal. The conceivability test is an account of separability, and any such account

will also demand a conception of functionality. This becomes especially obvious once we recall

that, in order to determine if the article is a useful article at all, we must determine if it has an

intrinsic utilitarian function. In other words, the question of the article’s intrinsic functionality

must be answered before even asking the question of separability. To see this, recall that the

merely to “portray the appearance of the article or . . . convey information.”

Thus, an article

does not count as a useful article if it merely has some utilitarian function, or a possible utilitarian

function; it must have an intrinsic one in particular. The following passage from Christopher

Buccafusco and Jeanne Fromer’s paper Fashion’s Function in Intellectual Property Law further

illustrates this point:

Most paintings do not have an intrinsic utilitarian function. Instead, they exist to portray

their own appearance. When we think about why most paintings are painted and why

they are valued by consumers, we focus on their representational characteristics. Indeed,

virtually all of the decisions that an artist makes about the content of a painting are

concerned with representational or depictive ends…This is not to say that a painting can

have no function. It may successfully cover a hole in the wall, in much the same way that

a book can serve as a doorstep. Yet this sort of utilitarian function is one that is shared by

all physical objects and is thus not intrinsic, or essential, to a painting or a book in any

meaningful way. This sort of non-intrinsic functionality does not magically make

Id.

paintings…intrinsically utilitarian. Most paintings, then, are by application of the

Given that we must already engage in this sort of analysis in order to determine if the

object qualifies as a useful article, there is no reason to think we cannot engage in a similar

analysis in order to determine what the legitimate functions are. To be sure, it might still require

a further step to define the function, beyond merely confirming that it has some such function;

but again, every test of separability will require this.

We have good reason to believe we can rise to this challenge. To see an example of how

such functionality-defining analysis should go, consider again the Inhale, Inc. hookah water

container.

One might argue that the actual container is not functionally identical to the one we

conceive of without the distinctive shape, as it does not contain the same amount of water in the

same way. But this argument is unconvincing, since it is unlikely that the amount and manner of

containing water is part of the hookah water container’s legitimate function, which is simply to

contain water. The amount and manner of containing serve no further functional ends, nor are

they relevant to the utility of the hookah or its ability to operate properly. We simply don’t care

how much water the container holds, or what particular shape it holds in. We can thus conclude

that the conceived-of container stripped of its design elements is functionally identical to the

actual one, and therefore—contrary to what the Inhale, Inc. court held—the shape is

conceptually separable from its functionality.

Buccafusco & Fromer, supra note 69, at 66.

Inhale, Inc. v. Starbuzz Tobacco, Inc., 755 F.3d 1038 (9th Cir. 2014).

Contra id.

To summarize the dialectic of the preceding paragraphs, I began in Section IV.A by

considering the worry that it is unrealistic to think that certain design elements are essential to

certain types of useful articles, and I dissolved this worry by re-emphasizing that the

conceivability test is to be applied at the level of particular useful articles, rather than such

articles abstractly construed. I then turned in Section IV.B to the challenges in defining a

particular article’s function, such as the fact that a function can be defined with more or less

fineness of grain, or can include all of the possible functions of the article (even incidental ones).

I argued that we can and should restrict ourselves to considering the object’s legitimate

functions, and that this is a feasible restriction, because we must already consider whether the

object has legitimate functions in our “intrinsic utilitarian function” analysis. Finally, I gave an

example of how this “legitimate function” analysis should go.

As mentioned above, while it is true that the conceivability test must be paired with an

understanding of functionality in order for it to actually be applied, this requirement is by no

means uniquely faced by the conceivability test alone. All tests for separability require an

account of functionality, and it is therefore not a weakness of the present proposal. In fact, it is a

virtue of the conceivability test that it allows us to cleanly separate functionality and separability

analysis. We have fully teased apart the questions of what is the legitimate function of the article

and whether the design element is conceptually separable from that function (in other words,

whether it passes the conceivability test). And such a systematic approach to the messy and

tricky useful articles doctrine will no doubt better ensure that courts get things right with their

analysis, with respect to both separability and functionality.

V. STAR ATHLETICA & OTHER CONCLUDING REMARKS

A. Conceivability & Cheerleading Uniforms

What does the conceivability test tell us about the facts of Star Athletica? For one thing,

in applying the test, we can see even more clearly that the difficult remaining question is that of

functionality (and it is the question we should turn our attention to next, now that separability has

been resolved). Recall again that the dispute in Star Athletica regarded the design elements of

cheerleader uniforms: stripes and chevrons, placed in particular colors and locations on

otherwise white clothing articles.

Applying the conceivability test, we pose the following

question: When we conceive of the uniform as lacking these design elements, is the uniform we

conceive of functionally identical to the actual uniform?

I argue that the right answer to this question is no. The functions of a cheerleading

uniform—beyond those of any other, ordinary article of clothing—are to create the impression

that the wearer is a cheerleader, that all the wearers are members of the same cheerleading squad,

and that the whole cheerleading squad is uniquely attached to the team for which it cheers. Since

the uniform we conceive of must involve the closest uncreative alternative design, we cannot

replace the stripes and chevrons with some newly imagined design elements. But such blank

uniforms are not identical to the existing ones with respect to the aforementioned three functions.

A cheerleading uniform without stripes, chevrons, or any other such embellishment does not

look like a cheerleading uniform at all: It looks only like a white skirt and blouse. This is

unhelpful in identifying a squad or delineating it from another squad (for instance, in the context

of a game or large-scale cheerleading competition), nor does it do anything to suggest which

team the cheerleading squad is attached to. Therefore, the design elements in this case do not

Star Athletica, L.L.C v. Varsity Brands, Inc., 580 U.S. _ (2017).

pass the conceivability test, and thus the Supreme Court erred in extending them copyright

protection.

My analysis might be contested on the basis of how I characterize the functions of these

articles. For instance, it might be objected that the three functions listed constitute “conveying

information,” and that they are therefore irrelevant under the useful articles doctrine.

I have two

responses to this objection. First, under the statute, even if the article must have an intrinsic

utilitarian function beyond conveying information in order to count as a useful article, this alone

does not entail that, once the article does count (by virtue of having some other function), the

conveyance of information cannot be regarded as one of its legitimate functions. In the present

case, this is to say that although the cheerleading uniforms would not have qualified as a “useful

article” if they had no function beyond conveying information, this is irrelevant because (a) they

do have some other intrinsic utilitarian function (namely, covering the wearer’s body) and (b)

conveying information can still count among their legitimate functions in our separability

analysis.

My other response to this objection is that it would be a mistake to characterize the

aforementioned functions of cheerleading uniforms as merely conveying information. To the

contrary, creating an impression involves causing a certain kind of phenomenology, or creating

an experience in the viewers. All uniforms, by nature, convey some information about their

wearers, but cheerleaders are also performers of a certain kind. The impression that these

uniforms create of the wearers—their relationships with each other as fellow squad members, as

well as their relationship to the team for which they cheer—are functionally important to that

17 U.S.C. § 101 (“A ‘useful article’ is an object having an intrinsic utilitarian function that is

not merely to portray the appearance of the article or to convey information”).

performance. After all, it would not be the same if the cheerleaders wore different, generic

outfits, but with large nametags identifying themselves as cheerleaders of squad X cheering for

team Y.

The impression created by these uniforms is therefore more than mere information

conveyance, and it is no doubt a crucial part of the uniform’s function.

It might also be argued that cheerleading uniforms have other important functions beyond

the three I have listed. For example, they might perform the function of flattering the wearer by

causing her figure to be perceived as an hourglass shape.

However, this point is orthogonal to

our present purposes. Even if the legitimate functions of these uniforms are not exhausted by the

three that I have listed, these functions are nonetheless among them. And it is sufficient for our

purposes that I have identified only these three, because the design elements already fail the

conceivability test with respect to these functions alone. This is to say that adding more to our

list of the article’s legitimate functions would not change our answer to the conceivability

question, and therefore does not bear on what our test tells us about the right holding for this

case.

We have considered and addressed some objections to my analysis of Star Athletica;

however, I must emphasize again that these concerns were with my functionality analysis, and

By “not the same,” I do not mean merely for aesthetic reasons, but for functional ones. For

instance, rather than immediately getting the impression that cheerleader A is on squad X and

cheering for team Y, I would have to read her nametag. This is a functional difference!

Is “creating an impression” too broad to count as a legitimate function? Would it entail, for

instance, that paintings are functional, or is there room for a distinct impression-creating function

that is important to the cheerleaders’ performance? This is another compelling line of criticism,

and I’m not sure that it would be wrong; but given that it takes us too far down the rabbit hole, I

will cut off my exploration of functionality at this juncture, leaving it aside for further, important

work.

Buccafusco & Fromer, supra note 69, at 53.

not regarding the conceivability test itself. And although I have attempted here to provide a

satisfactory defense of my account of the uniform’s functionality, these issues are otherwise set

aside for later work. Our application of the separability test to Star Athletica has shown us

another instance of this test in action, as well as how it differs from the Court’s own, both in its

method as well as the outcomes it might generate. But there is no doubt that, armed with our

newly minted test for conceptual separability, we must next turn our attention to the challenges

of functionality.

B. Conclusion

I criticized the Supreme Court’s holding in Star Athletica on the grounds that it

undermines the very purpose of the useful articles doctrine. Far from allowing us to maintain the

division of labor between copyright and patent law, the test will permit creators to utilize the thin

requirements of copyright to obtain backdoor patents on functional features of articles. This is to

say that, as a general matter, the Star Athletica test presupposes a wider domain of copyright, one

which includes elements that are both expressive and functional as candidates for protection. In

comparison, the conceivability test’s conception of the scope of copyright is more restricted:

Only those features which are purely expressive—but not functional—are regarded as within

the conceivability test, such as whether it expands or contracts the domain of copyright in other

ways, or whether it forces us to take a stand on any other difficult copyright questions. After all,

some scholars have advanced policy arguments emphasizing the importance that the useful

articles doctrine either prevent or permit copyright protection for certain kinds of articles. To

consider one example, Kal Raustiala and Christopher J. Sprigman have argued that the lack of

cyclical industry forward.

On the other hand, Christopher S. Hemphill and Jeannie Suk have

argued that the limitations on copyright protection offered to fashion design results in under-

protecting designers from “close copying,” thereby undermining innovation and incentives to

create, as well as distorting the direction of innovation.

If the conceivability proposal had either a broadening or narrowing effect on the scope of

protection afforded to fashion designs, then it would also have implications for the policy

concerns raised by the aforementioned discussions. However, and happily, the test is rather

neutral. This is precisely because the conceivability test, unlike many proposals of the past,

allows for separability and functionality analysis to be independent of one another.

Many of the

policy concerns regarding the domain of copyright law will ultimately rise or fall depending on

our account of functionality, but the conceivability test does not presuppose any commitments

with respect to that question. For instance, accepting the conceivability test is compatible with

both accepting or denying Buccafusco and Fromer’s earlier-mentioned account of fashion and

“looking good on” functionality, according to which design elements that positively affect the

Kal Raustiala & Christopher J. Sprigman, The Piracy Paradox: Innovation and Intellectual

Property in Fashion Design, 92 VA. L. REV. 1687 (2006).

Christopher S. Hemphill & Jeannie Suk, The Law, Culture, and Economics of Fashion, 61

STAN. L. REV. 1149 (2009).

See, e.g., Brandir Int’l, Inc. v. Cascade Pacific Lumber Co., 834 F.2d 1142, 1142 (2d Cir.

1987) (holding that copyrightability is dependent on “the extent to which the work reflects

artistic expression uninhibited by functional considerations”); Kieselstein-Cord v. Accessories by

Pearl, Inc., 632 F.2d 989, 989 (2d Cir. 1980) (holding that the expressive aspect is conceptually

separable whenever it can be said to be “primary,” and the function can be said to be

“subsidiary”).

See supra Section IV.

way the garment’s wearer is perceived qualify as functional.

Undoubtedly, whether or not one

accepts the Buccafusco and Fromer view on the function of fashion—or any other proposed

account of functionality for that matter—will have broadening or narrowing effects on the

domain of copyright; but this can all be freely debated and determined in its own right,

independent from the question of conceptual separability.

We have seen that the philosophers’ tool for delineating necessities is well-suited to the

task of delineating whether or not an expressive element is independent from its article’s utility. I

have argued that this test follows from the appropriate reading of the statutory text of the useful

articles doctrine, and that it also promotes the purpose of the doctrine by maintaining the division

of labor between copyright and patent law. I have outlined a number of virtues that this test

possesses which many others lack, including that it treats conceptual separability as analogous to

physical separability, and that it does not require judges to make aesthetic judgments about the

value of expressive elements. I have responded to philosophical challenges raised by the notion

of conceiving and have remarked on how the notion of functionality fits with this test. And

although it is true that the conceivability test must be paired with the right account of

functionality in order for it to be successfully applied, the need for such an account is equally

shared by all other candidate tests. Fittingly, questions of functionality should remain

conceptually separate.

Buccafusco & Fromer, supra note 69, at 53.