Download a PDF version of this article here.

By Nicole Lieberman*


By refusing to acknowledge[1] the aesthetic judgments inherent in determining copyright disputes,[2] American courts have plagued our copyright law with subjective bias[3] and doctrinal confusion. To avoid the appearance of impropriety, since at least 1903[4] courts have side-stepped clearly defining foundational concepts such as “originality,”[5] “authorship,” and “infringement.” As such, they have failed to provide a meaningful methodology for determining when a work infringes the copyright of another.[6] By instead relying on the impossibly vague “substantial similarity” test,[7] courts have crafted an impressionistic doctrine that has drifted far from copyright’s original economic purpose of incentivizing creation.

While copyright infringement requires proof of copying, mere copying is not the end of the inquiry, as “[t]rivial copying is a significant part of modern life.”[8] Thus, proof of copying, or “copying-in-fact,” is only a threshold issue for proving infringement.

Copying-in-fact can be shown through direct evidence, such as testimony, but with witnesses and honest thieves often lacking, copying is most often shown by circumstantial evidence. Indirect proof of copying is provided by evidence creating an inference that the defendant copied – typically a combination of evidence of access to the plaintiff’s work and similarities probative of copying. While courts allow expert analysis and dissection to aid them in inferring copying, the largely unguided impression of lay observers determines the more exacting question of misappropriation.[9]

Yet determining misappropriation requires an understanding of the “axiom of copyright law that the protection granted to a copyrightable work extends only to the particular expression of an idea and never to the idea itself.”[10] Application of the “idea/expression” distinction requires delicate line-drawing to decide the appropriate “‘level of abstraction’ at which one defines the ‘idea’ that merges with the subject’s expression.”[11] But fact finders are unlikely to understand on their own which “ideas” are excluded or what elements fall into the category of “ideas.” While jury instructions can theoretically work to inform jurors to exclude such elements, “in practice jurors aren’t going to know what things are, for example, scène à faire[12] in the music industry, without some testimony on standard chord progressions.”[13] Thus, jurors are not likely to understand such an ephemeral distinction between ideas and expression, especially when applied to areas in which they lack expertise, as is often the case with copyright.[14] Because the issue of misappropriation is so dependent on the interpretation of these underlying principles of copyright law, classifying the issue as purely a question of fact for the jury requires reconsideration.[15]

Courts recognize the need for expert analysis and dissection in determining infringement in cases involving computer software. Distinguishing computers as “complex” and having elements dictated by limited options, courts apply a special test to ensure only protected elements are considered for infringement purposes. Yet they proscribe such guidance when the “aesthetic arts” are at issue, failing to recognize traditions unique to genres, that all art is capable of being broken down into constituent elements, and that such elements are dictated by genre and functional constraints. Courts have assumed that art is intuitive, simply reflecting emotions, and capable of being understood by anyone. Their narrow understanding of art comes from our law’s founding.

Copyright arose in an era where the courts viewed creativity as coming from a place of pure autonomous genius,[16] but this romantic view of aesthetics is a relic of the past: a counteraction to the age of enlightenment and rationalization. The reality is that creative borrowing is almost unavoidable and results in widespread use of unprotected elements from preexisting works. Without expert guidance and the ability to dissect protectable and unprotectable elements, judges and jurors are “more likely to find infringement in dubious circumstances, because they aren’t properly educated on the difference between protectable and unprotectable elements.”[17]

Due to the prevalence of music copyright infringement suits, and the fact that music is more perceptively derivative than other media,[18] it seems disproportionately plagued by the courts’ bias for traditional aesthetics. But music, like all arts, is inherently complex and technical,[19] and few “ordinary observers” know the elements and factors that go into its creation,”[20] especially with works of less familiar genres. Thus, fact finders are easily misled into finding substantial similarity based on unprotectable elements.[21]

While music may be uniquely crippled by our current copyright regime, the problems plaguing music copyright stem directly from a lack of guidance where it is arguably most needed: the technical issue of misappropriation.[22] With fact finders less likely to detect similarities attributable to common sources in unfamiliar aesthetic, the current system results in a prejudice against lesser-known aesthetics, and a bias for the traditional. [23] The result is far from encouraging aesthetic progress.

This paper will argue that to create a more encompassing[24] and objective copyright law, that fosters progress in all arts, it is vital to expand the role of analytic dissection and expert testimony to the misappropriation prong of the infringement test.

Part I of this paper provides background on the history of court treatment of music copyright and lays out the two major approaches to copyright infringement. In addition, this part outlines the foundational principle that only the expression of an idea is protectable. Part II illustrates how the tests have veered away from the original purposes and values underlying the inquiry. It argues that by relying on the ordinary observer test for misappropriation, the tests fail to accurately account for the idea-expression distinction. In outlining the problems facing music under our current copyright regime, this section shows how the problems with the audience test are particularly problematic for music, a medium in which the line between idea and expression is often not “spontaneous and immediate” to the ordinary observer.[25] The recent “Blurred Lines”[26] lawsuit serves to illustrate how the lack of objectivity in our current law results in inconsistent application, thereby diminishing incentives to create new works. More broadly, this section considers that while the problems for music are often more noticeable than for other media, they merely expose the larger inaccuracies of the audience test. Finally, Part III considers proposals for creating a more guided and objective infringement analysis. Ultimately, this paper concludes that the best solution is adopting the test for computer software, the abstraction-filtration-comparison method (AFC),[27] as a uniform test for infringement.

Requiring careful dissection of unprotected elements by the court would ensure educated decisions, and reserving the intuitive question of whether the defendant copied those elements for the trier would preserve the economic rationale of the lay listener test.[28] Effectively reversing the analysis of proof “will likely result in greater attention to the limiting doctrines of copyright law”[29] and the evolution of reasoned rule of law.[30] By basing aesthetic nondiscrimination in objective and reasoned criteria, as opposed to the “anti-intellectual and book burning” philosophy[31] of visceral impressions, the courts can determine actual illicit copying while being receptive to unconventional aesthetics.[32]

I. Historical Background of Music in Copyright

While copyright law struggles to deal with the fine arts as a whole,[33] particular problems arise in the context of musical works. These issues are rooted in the history of copyright law. Many of the problems facing music copyright lie in the fact that creators are seeking protection under a scheme created for the distinct purpose of protecting works of literature.[34] However, these problems are not unique to music. American copyright law is based on a concept of authorship ill-suited to progress in general. This section will outline the evolution of our copyright infringement doctrine. In considering the historical application of the doctrine to musical works, this section analyzes the aesthetic norms embedded within judges’ and jurors’ findings of infringement.

A. Music’s Initial Encounters in Early Legislation and Case Law

Article 1, Section 8 of the Constitution authorizes federal legislation “[t]o promote the Progress of Science and useful Arts,”[35] but gives little guidance in defining the scope of the copyright system. The original Copyright Act of 1790 extended protection only to maps, charts, and books.[36] Though musical compositions were routinely registered under the 1790 Act as “books,”[37] it was not until the Copyright Act of 1831 that Congress expressly extended protection to musical compositions. Congress’s early failure to provide well-crafted protection for musical compositions is hardly surprising given the 1790 Act’s roots in Great Britain’s Statute of Anne, which covered only the distinct category of “books.”[38]

With no other protection available against infringers, composers naturally came to seek protection of their works through copyright.[39] Yet utilizing a scheme created for “written” works meant obtaining copyright protection solely with respect to the underlying composition, “the notated, written score, including the music and any lyrics.”[40] While seemingly analogous, music as a performing art is “often related in some way to performance” and must be understood by reference to its context, that is, elements outside the composition.[41] Though federal law since 1976 has applied copyright protection to musical recordings, including some performance elements such as percussion, recordings are treated as distinct expressions with separate copyright protection.[42] Consequently, musical compositions are protected only within the restrictive framework of the “musical work,” which is defined as a combination of melody and harmony.[43]

More problematically, courts analyzing music copyright cases tend to place undue weight on melody, rather than harmony and rhythm,[44] failing to consider the complexity of music and a realm of possible distinguishing features. Focusing on elements of music that “lend themselves to notation”[45] may seem adequate in analyzing works from European musical traditions, which typically have predominant harmonic and melodic structures,[46] but doing so fails to consider music in its totality. Because music is inherently relational,[47] our perception of musical works, and their meaning, is dependent on the context in which notes and pitches in the melody are played.[48] Elements such as timbre and spatial organization are also relevant to the way we hear music and to the similarities we perceive. Consequently, “originality is better viewed as a function of the interaction and conjunction of these elements than of any element alone.”[49]

Neglecting to consider the totality of elements in musical works, while ill-suited even to classical traditions, most drastically affects works outside of Western traditions. The main aesthetic features of non-Western music often fall outside the confines of the court’s emphasis on melody and notation. Most notably, hip-hop, which finds its roots in certain African musical traditions, features a dominant “oral tradition,” evidenced in the practice of rapping, and complex rhythmic structures with less emphasis on melodic and harmonic structures. Moreover, such traditions predominately feature the element of musical borrowing through the practice of sampling, looping, and interpolation. These features are also found in other African American musical genres, including blues, jazz, rhythm and blues, gospel, Soul, rock, reggae, funk, disco, and rap, and are found mixed with all types of music today.[50] Electronic music producers are now producing hip-hop tracks,[51] and even pop-country artists are making rhythm-centric tracks that reference hip-hop culture. Not surprisingly, entering the arena puts artists at risk of facing a copyright suit. Taylor Swift recently faced a $42 million infringement claim for using the lyric “haters gone hate,” a staple in hip-hop culture and music, [52] in her recent dance-pop track “Shake it Off.”[53] Despite the prevalence of non-notational elements, copyright’s bias for written work places works that do not fit the mold “at the bottom of the hierarchies of taste,”[54] making findings of original elements in allegedly infringing works more difficult to obtain.

Borrowing similarly conflicts with Western ideals of creativity and originality, with the result that music has historically been disvalued. Records from the time of the Statute of Anne’s enactment are telling of the hostile attitudes facing music. While literature was held in high esteem for its educative role, music was seen as an unnecessary luxury that served merely as entertainment.[55] In Pyle v. Falkener,[56] an early case brought under the Statute of Anne, defendant publishers argued that, in contrast to works of literature, authorship of music required “a high standard of originality to qualify for protection under any legal theory.”[57] Underlying their challenge was the commonly held notion that composers “merely borrowed” from “[o]ld [t]unes which had been [u]sed in [c]ommon by all persons for many years before…” and as such have no proprietary rights.[58]

Such disparaging views of music are less surprising when one considers the rise of the Romantic view of authorship in the nineteenth and twenties centuries.[59] Unlike the classical conception of authorship, which “conceives of art as imitating universal truths and ideas,” and thus contemplates the evolutionary nature of art, the Romantic view conceptualizes the creation of art “as a process that reflect[s] the emotions and personality of the individual artist.”[60] With the Romantic view informing cultural assumptions, originality often came to be defined as requiring independent creation, “which essentially appears to rule out or significantly limit borrowing.”[61]

With the functional and genre constraints inherent to music,[62] tensions existed early on in applying copyright to musical works. Yet “use of existing works has historically been a core feature of the musical composition process”[63] and the artistic process in general. The courts’ neglect to appreciate the reality of borrowing has often resulted in overbroad copyrights, extending protection to more than just the particular arrangement of the literal elements of a work.[64] Additionally, by failing to recognize distinguishing features of songs that lie outside the melody and notation, courts often find infringement based on unprotected elements. Genres that explicitly sample existing works, such as hip-hop, have been hit hardest. As a result, the courts are perpetuating a bias for traditional aesthetics at the expense of progressive and unfamiliar artistic movements.[65]

B. The Idea-Expression Distinction

Early on, courts using the idea-expression dichotomy to distinguish between unprotectable and protectable aspects of works did so on the basis of tangibility. Though the courts still use these terms in filtering out unprotectable elements, changing views on the nature of the artistic process have distorted the original tangibility basis, leading to ad hoc judicial determinations. With the rise of the Romantic view, artistic works in their entirety came to be regarded as reflecting the artist’s contributions.[66] As a result, perceptions regarding the moral (and thus, intellectual property) rights of an artist expanded to include more than just the particular arrangement of the literal elements of a work.

Originally, American copyright law viewed “ideas” as “intangible, unexpressed concept[s] that existed only in the author’s mind.”[68] Therefore, copyright protected only the tangible[69] “expression,” or the “arrangement of words which the author has selected to express the idea.”[70] The rationale served the purposes of the intellectual property clause well, since free access to ideas is critical to the development of creative works.[71] Moreover, the right granted did not include a right over certain words used, because “they are the common property of the human race.”[72] This early approach was consistent under the classical conception of the creative process, which views the artist as portraying an intangible idea or truth which “cannot and should not be captured or controlled by one artist.”[73]

However, with the rise of the Romantic view in the nineteenth and twenties centuries,[74] Congress no longer limited “expression” to the arrangement of the literal elements of the copyrighted work, but expanded it to include underlying “original” conceptual elements as well.[75] In 1909, Congress both enlarged the category of works eligible for protection and expanded the rights provided to copyright owners, including use of the work in a different medium.[76] Protection under the Act was no longer limited to the literal form or features of the expressed idea, but extended to elements of a work that are intangible and conceptual.[77] In applying the new act, the Supreme Court in Kalem Co. v. Harper Bros.[78] found the defendant’s film to infringe upon the plaintiff’s copyright in the book Ben Hur because the film expressed the same underlying idea, or plot, albeit in an entirely different medium.[79]

It became clear that balanced against the idea-expression distinctions is the countervailing consideration that copyright infringement cannot be limited to exact copying, “else a plagiarist would escape by immaterial variations.”[80] The problem is one of line-drawing: at what point is a variation distinguishable enough to “sufficiently alter a work’s substantial similarity to another so as to negate infringement,” without extending protection to the underlying idea of the plaintiff’s work?[81]

Views on the nature of art and the creative process have only continued to evolve and become more inconsistent with the idea-expression dichotomy.[82] The conceptual art movement advanced the rejection of any distinction between an artist’s idea and the ultimate expression.[83] As conceptual artist Sol LeWitt stated, “the idea or concept is the most important aspect of the work. When an artist uses a conceptual form of art, it means that all of the planning and decisions are made beforehand and the execution is a perfunctory affair. The idea becomes a machine that makes the art.”[84] In rejecting the Formalist tradition, which defined art by its form and structure, conceptual art judges art by what it contributes to the conception and definition of “art.”[85] Even an unchanged item from the grocery store, like a box of Brillo soap pads,[86] can be art if framed in a new way.

With Romantic and neo-romantic views challenging classical aesthetic theory, no universally accepted philosophical or objective basis remains for distinguishing ideas from expression in works of art. Continuing to use the terms leaves courts to make infringement decisions on the basis of their own subjective assessments of a work’s artistic value.[87] Judicial determinations of what constitutes the “idea” versus the “expression” have come to reflect personal assumptions and experiences. Courts tend to find elements of a work to be an “idea” when they are familiar with the work’s aesthetic tradition and can recognize the elements as commonplace.[88] Conversely, courts are more likely to find elements of works in less familiar traditions to be original “expression,” making them more inclined to find later uses infringing.[89] As the Ninth Circuit admitted, “‘At least in close cases, one may suspect, the classification [of idea and expression] the court selects may simply state the result reached rather than the reason for it.’”[90] Thus, with changing views on the creative process, “it is no longer necessary or valuable or even possible to dissect a work of art to uncover the universal truths or ideas which must remain freely available to all future authors.”[91]

Distinguishing between ideas and expression is perhaps most illusory in the context of music, due to the relatively limited number of compositional choices when compared with literary works.[92] Western music, at issue in most copyright suits, is primarily written in the tonal system, an organized and relational system of tones (e.g., the notes of a major or minor scale) in which one tone becomes “the central point to which the remaining tones are related.”[93] Because there are a limited number of possible pitch and harmonic relationships, options within tonal music are somewhat dictated by the system.[94] Moreover, because the tonal system is built on a hierarchy of predominate chords and pitches,[95] certain “patterns and tendencies are . . . common to virtually all musical works composed in the tonal system.”[96] The distinction between these unprotectable ideas and the original expression thereof is difficult to see, and thus the room for bias is most apparent.[97]

C. Evolution of the Copyright Infringement Tests

Courts since the nineteenth century have attempted to separate the issue of copyright infringement into two issues. First, “Copying-in-Fact”: did the defendant see and copy from the copyrighted work or did he create his work independently?; second, “Misappropriation”: did the defendant appropriate too much of the protected work?[98] The first question is used as an evidentiary tool to infer copying from access or “striking similarity,”[99] while the second focuses on the liability issue.[100] The degree of similarity between the two works is relevant to both the inquiries;[101] the phrase “probative similarity” is often used in reference to the first inquiry, while “illicit similarity” is used for the second.

Courts in the 1900s maintained the distinction between the copying-in-fact and misappropriation inquiries. A “substantial similarity” test was used for the copying-in-fact inquiry to determine whether the degree of similarity between the defendant’s and the plaintiff’s work was substantial to the point of being probative of actual copying. [102] The focus was solely on whether the defendant had copied “the labors of the original author.”[103] As such, before comparing the two works for similarities, the court filtered out unprotected elements from the plaintiff’s work, including those that were “well known and in common use.”[104]

In determining misappropriation—that is, whether the defendant copied enough of the plaintiff’s work to be held liable—courts looked to the economic or aesthetic value of what the defendant copied: if the portions extracted “embody the spirit and the force of the work…they take from it that in which its chief value consists.”[105] In this context, courts often used the adjective “substantial” to refer to the qualitative value of what was copied.[106]

As precedent evolved, courts began to combine the structure of these two prongs. As a result, courts have often confused the economic purpose of the misappropriation prong: finding infringement based solely on quantitative similarity without taking account of the unprotected elements in the original work. The analysis of the two major copyright tests below outlines how this confusion arose and focuses on the problems the misappropriation prong are causing for copyright.

1. Second Circuit Copying/Unlawful Appropriation Test

In Arnstein v. Porter,[107] the litigious Ira B. Arnstein sued the American songwriter and composer Cole Porter, alleging that many of Porter’s songs infringed the copyrights of songs written by Arnstein.[108] The Second Circuit conducted an influential bifurcated test,[109] which requires a plaintiff to prove (1) copying-in-fact and (2) illicit copying (unlawful appropriation) to establish infringement.[110]

i. Copying-in-Fact

The first prong of the Arnstein test is satisfied with the showing of (a) access; and (b) sufficient similarity, which is “probative” of copying: “The stronger the proof of similarity, the less the proof of access that is required.”[111] Thus, if similarities are so “striking” so as to “preclude the possibility that plaintiff and defendant independently arrived at the same result,” evidence of access may not be necessary.[112] Of course, the converse is not true because “access without similarity cannot create an inference of copying.” [113]

To evaluate the likelihood of copying versus independent creation, expert testimony and “analytic dissection”[114] are admissible.[115] However, the two works are to be compared in their entirety, including both protectable and non-protectable material.[116]

ii. Unlawful Appropriation

Only if the threshold issue of copying-in-fact is shown does the court move to the question of misappropriation.[117] Having established copying-in-fact, the issue of unlawful appropriation is a question of fact. Therefore, the fact finder must determine whether the taking went so far as to constitute infringement under the “substantial similarity” test.[118] That is, would the ordinary observer, unless he set out to detect the disparities, be disposed to overlook them and regard the aesthetic appeal of the two works as the same?[119] The second part of the Arnstein test is “related to the nineteenth century concern with the value of what the defendant had copied” as it asks whether the similarity “relates to material of substance and value in plaintiff’s work.”[120] However, the Arnstein test departs in some ways from earlier definitions of infringement by looking to the reaction of the “ordinary observer.”[121]

In determining misappropriation, the Arnstein test looks to “the response of the ordinary lay hearer.” That is, rather than making a purely subjective determination, the trier of fact is meant to determine the issue “in light of the impressions reasonably expected to be made upon the hypothetical ordinary observer.”[122] Because the court reasoned that the value of the work lay solely in the opinion of its intended audience, it held that expert testimony on the “impression made on the refined ears of musical experts” was “utterly immaterial.”[123] While seeming to realize the difficulty in discovering the views of the imaginary “ordinary observer,”[124] the court stated that expert testimony was permitted for the limited purpose of “assist[ing] in determining the reactions of lay auditors.”[125]

Moreover, because the court determined that the value of the works lay in their final form as impressed upon the ordinary observer, it instructed that detailed analysis and careful dissection were inappropriate.[126] Therefore, according to Arnstein, works were to be considered in their entirety, again including both protectable and non-protectable material.[127] The trier was left to depend on “some visceral reaction” as the basis for determining misappropriation.[128]

If the case involves “comprehensive non-literal similarity”[129] – that is, similarity in the overall structure of the works – the trier must make a value judgment of “whether defendant took from plaintiff’s works so much of what is pleasing to the ears of lay listeners, who comprise the audience for whom such popular music is composed, that defendant wrongfully appropriated something which belongs to the plaintiff.”[130] In a case of “fragmented literal similarity,” or verbatim copying of constituent elements, an analogous value judgment must be made, but here only with respect to the protectable portions of plaintiff’s work that have been taken.[131] Dissimilarities between materials alleged to be infringing are “significant because they mitigate any impression of similarity.”[132] Dissimilarities in other aspects of the defendant’s work, except to the extent they create an overall different impression, “typically are not significant.”[133] As Judge Learned Hand said, “no plagiarist can excuse his wrong by showing how much work he did not pirate.”[134] Thus, if the defendant copies from the plaintiff’s work, it does not matter if he adds significant material of his own,[135] resulting in what might be a transformative new work.

Consequently, under Arnstein, “[i]nstead of using some objective standards or criteria based on economic impact or quantity, courts [are] to determine infringement on an unpredictable, impressionistic basis.”[136]

iii. Further Developments and Confusions of the Arnstein Test

Although the Second Circuit in Arnstein conducted two separate inquiries into the level of similarity between the two works,[137] namely to establish copying-in-fact and then to determine misappropriation, confusion ensued from the dual use of the term “substantial similarity.” As a result of this confusion, in Ideal Toy Corp. v. Fab-Lu Ltd,[138] the Second Circuit essentially combined the issues into one subjective test. By misinterpreting the element of misappropriation identified in Arnstein as “merely an alternative way of formulating the issue of substantial similarity” rather than as an independent step, the Second Circuit stated that copyright infringement is shown solely by “substantial similarity” between the two works based on the reaction of the ordinary observer.[139] The court effectively reduced the test for prima facie copyright infringement to (1) access, and (2) misappropriation, thereby failing to consider copying-in-fact.[140] Therefore, the court rejected dissection, analysis, and expert testimony entirely and did not think it necessary to analyze the similarities between the works to determine the likelihood of independent creation.[141] By basing the entire copyright infringement inquiry on the subjective impression of those untrained in the arts, the court neglected to protect against a finding of infringement based on purely unprotectable and unoriginal elements. The Ideal Toy test fails to deal with the fundamental principle of copyright law that seeks to protect merely the expression of ideas rather than ideas themselves.[142] Unfortunately, Ideal Toy’s interpretation of the Arnstein test largely influenced the way modern courts use “substantial similarity” in determining infringement.[143]

Luckily, some courts have maintained the Arnstein two-part inquiry. In Universal Athletic Sales Co. v. Salkeld,[144] the Third Circuit restored Arnstein’s bifurcated approach in holding that a plaintiff must prove copying and “that copying went so far as to constitute improper appropriation.”[145] Moreover, the court recognized that “substantial similarity to show that the original work has been copied is not the same as substantial similarity to prove infringement.”[146]

More inherently problematic is the fact that the Salkeld court maintained Arnstein’s limits on expert analysis and dissection in determining misappropriation.[147] As a result, the Salkeldtest likewise fails to provide any objective standards or criteria for determining how much similarity is necessary to constitute misappropriation.[148]

Though courts instruct the fact finder to find misappropriation only if the defendant’s work copies not merely the idea, but ‘the expression of the idea,’ this “vague formula” is a reformulation, not a solution, to the problem of determining “what sort of similarity short of the verbatim will constitute substantial similarity.”[149] Thus, the ordinary observer continues to be left with “the impossible task of comparing only protected expression in determining substantial similarity without engaging in any thoughtful dissection or analysis of the works.”[150]

2. Ninth Circuit: “Total Concept and Feel”

The Ninth Circuit’s framework, laid out in Sid & Marty Krofft Television Prods., Inc. v. McDonald’s Corp,[151] represents the second main approach to determining copyright infringement. Although not a music infringement case, the “extrinsic/intrinsic” test developed by the Krofft court has been influential in many copyright disputes, including those involving music.[152] In recognizing that the ordinary observer is unlikely to be able to separate idea from expression in comparing two works without dissection or analysis, the Ninth Circuit proposed its own two-step test that attempts to ensure that there is “substantial similarity not only of the general ideas but of the expressions of those ideas as well.”[153] Though the test was later reformulated to include specific expressive elements during the extrinsic inquiry, as discussed in in Part I.ii.c., understanding the original formulation is key to examining its foundational flaws.

i. Extrinsic Test

The first step, or the “extrinsic” analysis, as originally cast by the Krofft court, was an objective comparison by the court for similarity in ideas.[154] Only if a substantial similarity of objective criteria under the “extrinsic” test is found do courts consider misappropriation under the “intrinsic” analysis.[155] Thus, the extrinsic test aims to limit protection to protectable elements by first filtering out unprotectable elements, including ideas, facts, and scènes à faire, and then determining whether the allegedly infringing work is “substantially similar to the protectable elements of the artist’s work.”[156]

According to the Ninth Circuit, in filtering out unprotected elements the extrinsic test incorporates the idea-expression dichotomy by limiting the scope of copyright protection to expression. As the court stated:

By creating a discrete set of standards for determining the objective similarity of literary works, the law of this circuit has implicitly recognized the distinction between situations in which idea and expression merge in representational objects and those in which the idea is distinct from the written expression of a concept….[157]

Courts conducting the extrinsic test “must take care to inquire only whether the protectable elements, standing alone, are substantially similar.”[158] Therefore, analytic dissection and expert testimony presented by the plaintiff on the similarities between the plaintiff’s work and defendant’s work are recommended.[159]

In performing the analytic dissection, courts are instructed to list and analyze the “measurable, objective elements” of the works,[160] including “the type of artwork involved, materials used, and the subject matter.”[161] For literary works, courts have listed such elements as “plot, theme, dialogue, mood, setting, pace, characters, and sequence of events.”[162] Because these factors do not readily apply to visual works of art, the court looks to the “objective details of the appearance.”[163] Without attempting to provide an exhaustive list of relevant factors, the court in Ninth Circuit listed such elements as “the subject matter, shapes, colors, materials, and arrangement of the representations.”[164]

Though described as a factual question, because it bases the question on objective criteria rather than the response of the trier, the extrinsic test may often be decided as a matter of law.[165]

ii. Intrinsic Test

Much like the Second Circuit’s ordinary observer test, the intrinsic test is entirely subjective and based on the “response of the ordinary reasonable person” to the “total concept and feel” of a work,[166] excluding expert testimony and dissection. Similar to the Arnstein court’s language of “lay [persons], who comprise the audience,”[167] the Krofft court suggested that the fact finder’s reaction be geared towards that of the intended or likely audience.[168] In a suit involving the characters in a children’s television show, the court stated, “The present case demands an even more intrinsic determination because both plaintiffs’ and defendants’ works are directed to an audience of children.”[169] Therefore, the court limited the inquiry to the understanding of a child and found substantial similarity despite noted differences. Without expert testimony to aid the trier in determining whether children might detect distinctions, the court relied on the triers’ subjective belief that children would be unlikely to notice minor distinctions.[170]

iii. Further Developments and Confusions of the Krofft Test

In Shaw v. Lindheim,[171] the Ninth Circuit modified the extrinsic/intrinsic test in recognition of the fact that district courts were not limiting the extrinsic test to a comparison of ideas.[172] Recognizing that the similarity of ideas prong is often shown by “focusing on the similarities in the objective details of the works,”[173] the Shaw court explained that the extrinsic/intrinsic test is no longer divided by an analysis of ideas and expression.[174] Rather, the extrinsic test is an objective analysis of specific “manifestations of expression,” while the intrinsic test is a subjective analysis of expression by the fact finder, which is no more than the lay observer’s visceral reaction which is “virtually devoid of analysis.”[175] Though the Shaw court recognized that the test was “more sensibly described as objective and subjective”[176] courts have confusingly continued to use the extrinsic/intrinsic language. Moreover, subsequent cases have left the analysis of improper appropriation to the jury analyzing the works as a whole.

In Swirsky v. Carey,[177] the Ninth Circuit applied the extrinsic/intrinsic test to a case involving musical works. Swirsky and his co-writer filed a copyright infringement suit claiming that Mariah Carey’s song “Thank God I Found You” plagiarized their song “One of Those Love Songs.”[178] The court rejected the district court’s approach to the extrinsic test, which involved a “measure-by-measure comparison of melodic note sequences.”[179] The Ninth Circuit felt comparing notes would fail to consider other relevant elements such as “harmonic chord progression, tempo, and key” as “it is these elements that determine what notes and pitches are heard in a song and at what point in the song they are found.”[180] precisely which elements the court should consider, explaining in dicta that the copyright framework is difficult to apply to aesthetic works such as music which are “not capable of ready classification into . . . constituent elements” the way literary work can be classified into “plot, themes, mood, setting, pace, characters, and sequence of events.”[181]

The Ninth Circuit’s opinion in Swirsky is also relevant for its proposition that substantial similarity can be found based on a combination of elements, “even if those elements are individually unprotected.”[182] For example, in Three Boys Music Corp. v. Bolton,[183] the Ninth Circuit upheld the jury’s finding that two songs were substantially similar due to the presence of the same five individually unprotectable elements: “(1) the title hook phrase (including the lyric, rhythm, and pitch); (2) the shifted cadence; (3) the instrumental figures; (4) the verse/chorus relationship; and (5) the fade ending.”[184] Even though courts filter out unprotected elements such as expression that are commonplace within a genre, they reconsider these elements in examining whether there is a unique combination of elements.[185] However, the protection granted to a unique combination of elements is “thin,” applying only to the combination itself, not the individual elements, and protecting only against “virtually identical” copying.[186]

II. The Roots, Flaws, and Legacy of Arnstein

Though Arnstein remains the majority approach to analyzing copyright infringement,[187] the opinion has attracted much criticism. This part will begin by detailing the flaws of Arnstein, and its progeny, including Krofft. Arguing that Arnstein lacks objectivity by relying on the “impression” of the lay observer and limiting the use of expert testimony,[188] this section will consider courts’ classification of misappropriation as a subjective factual question for the jury, rather than as a legal question with its own standard. This part will argue that the treatment of the “aesthetic arts” as incapable of technical analysis is the root of its subjective treatment of misappropriation. Finally, this part will argue that in relying on the general public, untrained in artistic assessment, the courts risk finding infringement based on similarities that are attributable to common sources.[189]

A. Problems with Arnstein and the Ordinary Observer

Without detailed analysis, filtering out unprotectable ideas, or guidance from experts on the artistic merits of the works at issue, little assurance remains that jurors will decide the issue of misappropriation in keeping with the law. This section will explore why relying on the reaction of laymen is problematic in the context of copyright law. While jury instructions attempt to solve the problem, this part will examine why such abstract guidance is often more confusing than helpful, as judges themselves seem baffled by the blurry line of where an idea ends and its expression begins.

While the ordinary observer test attempts to utilize the “reasonable person” standard found in other areas of the law, its application to copyright is not analogous. In areas such as tort law, the trier is capable of placing himself in the defendant’s shoes to assess the defendant’s actions.[190] In copyright, because the trier often lacks the defendant’s artistic background, the trier cannot reasonably put himself in the defendant’s shoes to consider whether he would have been constrained to copy from the plaintiff in order to achieve the given result.[191] Consequently, the trier is asked to assess the defendant’s actions as if he is an “average lay observer” reacting to whether the defendant’s work appears to have been copied from the plaintiff’s work.[192] Yet copyright is meant to protect artists from the theft of the fruits of their labor, not from the “impression” of theft.[193] While the “spontaneous and immediate” impression of theft is “important evidence,” it cannot be the end-all-be-all test.[194] Given the complexities of copyright law,[195] the ordinary observer simply is not capable of accurately detecting very real appropriation.[196]

While the ordinary observer test has logical value in protecting the artist’s interest in the potential fruits of his labor by looking to the response of the specific market from which those fruits would derive,[197] its methodology for making that determination is lacking. One problem is that it is not clear the trier has knowledge of what constitutes the “lay listener’s” response,[198] especially considering the multitude of possible reactions even among a “target” audience.[199] While expert testimony is permitted to inform the fact finder on the views of the target audience, it is questionable whether qualifying as a music expert establishes “an expertise in the aural perceptions of a lay hearer.”[200] Put another way, “whether an expert, highly educated in the field of music theory, analysis, and history, can in fact hear again as a lay listener is speculative at best.”[201]

Prohibiting expert dissection and analysis on the ultimate issue of misappropriation deprives the trier of information that may be helpful in hearing the music through the ears of the “audience for whom such popular music is composed.”[202] As Professor Nimmer pondered:

If what is to be protected is literary theft, and not the impression of literary theft per se, why, we may wonder, must the view be “uncritical,” and why must there be no suggestion and pointing to similarity, if that suggestion would prove helpful to the trier in seeing that all or a part of plaintiff’s work formed the basis for all or a part of defendant’s work?[203]

Experts could note similarities dictated by the particular type of work at issue that are “most likely insignificant to the ears of the targeted audience familiar with that form or type of work,”[204] thereby helping to ensure that infringement is not found based on common sources or coincidence,[205] and conversely, ensuring that very real appropriation does not go undetected.[206] As Justice Clark, dissenting in Arnstein, noted, the jury is not “‘pre-eminently fitted to decide questions of musical values,’ which are different from an ordinary fact-finding exercise.”[207]

Perhaps the outcome in Arnstein, and thus the bases of our modern law, can be explained by the views of Justice Frank, who wrote for the majority. Frank believed that some judicial decisional processes “like the artistic process, involve[d] feelings that words cannot ensnare” since they contain “overtones inexpressible in words.”[208] For Frank, music was the prime example “of a hunch that was intrinsically incapable of disaggregation.”[209] He wrote, “a melody does not result from the summation of its parts; thus to analyze a melody is to destroy it. It is a basic, primary unit.”[210] Similar language is found in opinions today, including the Ninth Circuit case Swirsky cited earlier.

Attempting to break down what was underlying Frank’s opinion, one commentator argued that the Arnstein court’s exclusion of expert testimony and dissection seemed rooted in the court’s fact-skepticism and unwillingness to entrust the explanation of similarities to experts, preferring instead a “subjective determination,” perhaps in the hopes that a lay-jury trial would confirm the judge’s own subjective hunches.[211]

While courts typically rely on lay jurors to apply the law to the facts of a case at hand,[212] it seems naïve to believe jurors are capable of understanding the complexities of copyright law, particularly the ever elusive idea-expression distinction.[213] Federal judges themselves have found the doctrine difficult to apply,[214] and for good reason as “precision in marking the boundary between the unprotected idea and the protected expression . . . is rarely possible.”[215] Determining what is an “idea” versus an “expression” requires more than mere application of the law; it requires interpretation of the law and consideration of policy. Even with a basic understanding of copyright law, applying the doctrine to music is still more difficult. “What of a song’s music is ‘idea’ and what is ‘expression’?”[216]

While jury instructions attempt to inform jurors of the law,[217] in practice instructions are often an inconsistent, “confusing welter of legal jargon” that may wrongly suggest that any copying, including copying of an idea, counts as infringement.[218] Moreover, clarity in keeping the two different “substantial similarity” inquiries distinct, both in purpose[219] and procedure, is often lacking. Because both the Arnstein and Krofft tests forbid expert testimony on the ultimate issue of misappropriation, fact finders will often have to consider expert analysis and dissection on alleged similarities during the copying-in-fact or extrinsic inquiries, but somehow disregard such testimony on the misappropriation or intrinsic inquiry.[220] “Especially in complex cases, [it is] doubt[ful] that the ‘forgetting’ can be effective when the expert testimony is essential to even the most fundamental understanding of the objects in question.”[221] Moreover, courts often fail to clearly explain the two-step inquiry and the need to disregard testimony presented on the first issue in determining the second.

For example, consider a well-known music plagiarism case, Selle v. Gibb,[222] where the plaintiff, a part-time musician and composer, sued the Bee Gees, alleging that the Bee Gees’ “How Deep is Your Love” infringed the copyright of his song, “Let It End.”[223] Without evidence of access, Selle sought to establish copying-in-fact by showing substantial similarity between the two songs, relying heavily on expert testimony.[224] Yet in making the misappropriation determination, the jury was neither instructed to disregard the expert’s testimony on substantial similarity, nor informed of the two distinct steps of the Arnstein test.[225] Instead, the jury instructions combined the copying-in-fact and misappropriation steps, explaining that “[t]o prove substantial similarity plaintiff must establish…that the average person would recognize [defendant’s song] as having been appropriated from parts of [plaintiff’s song].”[226]

Consider also the jury instructions in the recent “Blurred Lines” trial. Judge Kronstadt’s first instruction to the jury on the copyright infringement standard said that any copying of original elements is unlawful:

Anyone who copies original elements of a copyrighted work during the term of the copyright without the owner’s permission infringes the copyright….[If] 1. the Gaye Parties are the owner of a valid copyright; and 2. the Thicke Parties copied original elements from the copyright work….your verdict should be for the Gaye Parties.[227]

According to this instruction alone,[228] whether the copied element is an unprotectable idea is irrelevant. In attempting to explain the extrinsic/intrinsic analysis, Kronstadt further suggested that copying an idea can count as infringement. Kronstadt instructed the jury to consider similarities in ideas, as well as expression, on the extrinsic test, but he failed to tell the jurors to disregard any similarity in ideas when considering the “concept and feel” during the misappropriation inquiry.[229] Following this questionable guidance seems to be the exact mistake the jury made in finding that Robin Thicke and Pharrell Williams infringed the copyright in Marvin Gaye’s song “Got to Give it Up” by creating their stylistically similar song, “Blurred Lines.”[230]

Fundamentally, the problem with the audience test is that by wholeheartedly relying on the lay juror, the test erroneously treats the question of misappropriation as a pure question of fact. Infringement is far less intuitive, and more complex, than ordinary negligence. Just as the testimony of medical experts is necessary in negligence cases in the context of medical malpractice, copyright is dependent on a technical analysis of works which jurors know little about. Determinations of infringement necessarily involve ad hoc line-drawing that affects artistic incentives and the public’s access to art works. While it is possible to limit and tailor the test, as it stands, it has been disappointingly inaccurate and has been often used as a “verbal formula to explain results otherwise reached.”[231]

B. Krofft: Reconciling Arnstein and the Idea-Expression Dichotomy?

While the court in Krofft seemed to recognize the problems with the Arnstein test in its failure to ensure copyright infringement was only found on a similarity of expression, the Ninth Circuit neglected to explain how its own test would resolve any of the problems presented by the Second Circuit approach.[232] In laying out vague criteria for defining ideas versus expression, the Ninth Circuit left lower courts and artists to their own devices in figuring out where to draw the line.[233] The court itself noted that the extrinsic test is “turbid waters.” “Nevertheless,” it continued, “the test is our law and we must apply it.”[234] Applying the test to artistic works is even more problematic. As the court said in Swirsky, “the extrinsic test provides an awkward framework to apply to copyrighted works like music or art objects, which lack distinct elements of idea and expression.”[235]

By failing to state whether a lack of substantial similarity in ideas is relevant for copying or misappropriation, or both,[236] the extrinsic/intrinsic test fails to isolate the issue of copying from the issue of misappropriation. Focusing on the idea-expression distinction improperly frames the question. Similarity of ideas found during the extrinsic inquiry may be probative of copying, but such similarity does not prove copying of protected expression. Furthermore, the Ninth Circuit never made clear how the ordinary observer could determine if there is a similarity of expression while refraining from dissection and analysis. If anything, fact finders in the Ninth Circuit are left with less expert guidance than in the Second Circuit, as the Second Circuit at least allowed testimony to inform the trier of the intended audience’s likely views rather than leaving them guessing.

As in other jurisdictions using the term, courts applying Krofft struggle to define “substantial.” The focus ranges from audience confusion,[237] to the substantiality relative to the overall work,[238] to the aesthetic or financial value of the portion of the work copied.[239] In the Krofft decision itself, because the works at issue were directed at children, the court focused on the impact of the “total concept and feel” of the works “upon the minds and imaginations of young people.”[240] Thus, some courts adopting the extrinsic/intrinsic framework utilize an “intended audience” test for works that appeal to an audience with “specialized expertise,”[241] therefore allowing the use of expert testimony. Yet such works have been narrowly construed and usually only include computer software.[242] Most courts applying the Ninth Circuit test determine the likelihood of audience confusion by simply comparing the “total concept and feel” of the works to the ordinary observer as determined by jurors with no specialized training or expertise.[243]

Courts that look to the value of the elements at issue criticize the “total concept and feel” approach for failing to maintain the idea-expression distinction during the misappropriation analysis. As the Ninth Circuit stated in Cooling Systems & Flexibles, Inc. v. Stuart Radiator,[244] “What is important is not whether there is substantial similarity in the total concept and feel of the works . . . but whether the . . . amount of protectable expression in Cooling Systems’ catalog is substantially similar to the equivalent portions of [the defendant’s] catalog.”[245] The court in Cooling Systems stated that the intrinsic test must take into consideration the nature of the elements that were allegedly infringed,[246] reasoning that “the fewer the methods of expressing an idea, the more the allegedly infringing work must resemble the copyrighted work in order to establish substantial similarity. . . .”[247] Yet the court continued to prohibit the aid of dissection, analysis, and expert testimony, failing to explain how an ordinary observer lacking expertise could make such a determination.[248]

C. The Modern Music Dilemma

While creators of musical works have struggled to extend copyright beyond its natural borders since the Statute of Anne was enacted, the inexact fit of music in our copyright doctrine has only become more obvious with the invention and popularity of sampling technology. Since the advent of digital sampling technology in the 1960s, the art of taking a portion, or sample, of a sound recording and repurposing it to make a different song has become engrained in nearly every popular musical genre.[249] Yet, after the industry-shattering decisions in Grand Upright Music, Ltd. v. Warner Brothers Records Inc.,[250] and Bridgeport Music, Inc. v. Dimension Films,[251] which resulted in a bright-line rule against de minimis copying of even three notes of a sound recording,[252] artists need to seek licensing rights to sample a sound recording. However, the costs to do so are often prohibitive, both in terms of negotiating fees and contacting often-elusive copyright owners.[253] Consequently, many artists have turned to re-creating segments of prior musical compositions for use in their own works, a practice known, at least in the hip-hop industry, as “interpolation.”[254] Though creative borrowing of this type is deeply embedded in art history, after the Blurred Lines verdict it seems songs that interpolate prior works are also at risk of extinction.

While we may be lacking strong evidence that copyright actually encourages creativity,[255] copyright can suppress the creation of works of a specific nature.[256] After Grand Upright put “the fear of God” in recording companies, artists releasing a record on a major label were forced to clear every sample used.[257] As a result, songs composed of various samples from multiple sources were no longer possible. Records like Paul’s Boutique by the Beastie Boys, which is almost entirely composed of 104 samples, would be “financially and bureaucratically impossible” today.[258] Each sample would have to be cleared by obtaining two different licenses from two different rights holders: the owner of the sound recording and the owner of the underlying composition. To make matters more complicated, additional licenses are needed if the sampled song contains samples itself, as is increasingly the case today.

While artists can obtain a compulsory license to create a cover of a prior song, using only a portion of a song while substantially altering “the melody or fundamental character of a work” falls outside of the range of the statutory license,[259] even though doing so entails more originality and less substantial similarity.[260] Newcomers are back in the position of having to seek permission and pay whatever rights holders demand,[261] or face the risk of hefty legal judgments and court costs.

For well-known songs, the costs to license are often as high as 100% of the royalties generated by the new song, and sometimes higher. Mark Ronson, a music producer and sampling guru, analogizes the process of creating his debut album, Here Comes the Fuzz, to the process of producing the fictional musical in the play, The Producers. In creating the track, “Ooh Wee,” Ronson wanted to sample two different songs by two different artists: a drum sample from Dennis Coffey’s song “Scorpio,” and a string sample from a cover of Boney M’s song “Sunny.” Collectively, the owners of the rights wanted 125% of the song’s royalties. Believing the samples to be necessary to the “emotional effect” and “toughness of the beat,” Ronson said he had to do it: “I had to pro-rate my entire album down so I could rob this song to pay that guy….”[262] In hindsight, the decision may have made his career, as the song ended up being the lead single and a chart-topper.[263]

Rights holders can always refuse clearance entirely, as they often do. When musician and producer Danger Mouse put out The Grey Album, which splices together samples from The White Album[264] by the Beatles with an a cappella version of The Black Album by rapper Jay-Z, the recording company that holds the Beatles’ copyright was able to prevent the album’s distribution, despite approval from both Jay-Z and the surviving members of the Beatles.[265]

Since Grand Upright, it has become largely impossible to create songs using more than one or two samples. As a result, industry practice and the sound of hip-hop music abruptly changed. Artists turned to heavily interpolating a few samples per song, particularly from artists who are amenable to having their music sampled. Dr. Dre’s production style, which was highly influential in spawning an entire era in hip-hop, changed from outright sampling to heavily interpolating the 1970s funk band, Parliament.[266] Dre’s style changed hip-hop forever, as G-funk, defined by Parliament’s influence, became the most popular genre in hip-hop during the 1990s.[267]

While some artists are still producing heavily-sampled albums, it seems only those with the most obvious fair use defense are confident enough to do so without a license, perhaps recognizing the strength of their defense and realizing record companies would rather not risk setting bad precedent.[268] The D.J. Gregg Gillis, better known as Girl Talk, is perhaps the most notorious sampler; he uses hundreds of small samples on a single album, never licenses anything, generates tons of publicity, and is never sued. As Gillis put it, with so many samples, “[i]t would take you hundreds of hours of work and hundreds of thousands of dollars to clear the rights to this album even if you wanted to.”[269] Yet, Girl Talk’s business is not without harm. Both iTunes and a CD distributer refused to carry his most recent album, Night Ripper, because of legal concerns.[270]

While Girl Talk might feel confident in his defense, anyone using a sample in a more qualitatively and quantitatively significant way is likely to fear settlement fees, or worse, high statutory damages. Even with a seemingly good defense, those with less fame than Girl Talk, or Pharrell Williams, are less likely to take the risk. When part of the beauty of digital sampling technology lies in its removal of bars to entry, allowing a twenty-year-old kid to create a critically acclaimed album with cheap technology,[271] cases like Grand Upright and “Blurred Lines” leave the next Dr. Dre on the margins.

Although Dr. Dre’s debut album immortalized Parliament’s style as an entire hip-hop genre in and of itself, and the release of “Blurred Lines” landed “Got to Give it Up” back on the Billboard 200 after a decade’s long hiatus,[272] the courts have ignored the economic reality of homage and have placed it on par with theft. A justified reason for deeming musical borrowing as theft is lacking when artists of all mediums, from classic literature to appropriations art, have borrowed from their predecessors without anyone taking notice, or under the defense of “fair use.”[273] One rationale is that our copyright law is based in the romantic conception of authorship, that is, the notion that the author has the right to control the context of his work and should be able to object to “her work [being] sampled and added into a work she finds repugnant.”[274] However, applying moral rights arguments to the issue does not serve the Intellectual Property Clause’s purpose of striking a balance between economically incentivizing the production of creative works and maintaining public access to those works.[275]

The current tests do not adequately balance the impact of copyright on creativity. Rather, courts focus more on the potential harm to the plaintiff’s market[276] than on assuring “access to the raw materials that artists need to create in the first place.”[277] Progress and culture depend on the accumulation of aesthetic works. By granting property rights in creative building blocks and requiring new works to be wholly different from prior works, the courts are treating the arts like science and technology, where progress means improving. However, progress in the arts is valued for reasons beyond efficiency. Society benefits from the accumulation of artistic works and the mere experience of making them.[278] “Culture . . . is a social phenomenon. It is not the creation of one or another artist, but of many doing somewhat similar things.”[279] The scènes à faire doctrine is one way of protecting free access to artistic materials. But with jurors unable to appreciate the balance of these competing interests, we risk skewing it in a way that stifles creativity.

1. Scènes à Faire in Music: “Blurred Lines” as a Case in Point

Credited as a Billboard’s Song of the Summer for 2013,[280] and the best-selling single of 2013,[281] it is no wonder the controversies over Robin Thicke and Pharrell Williams’s song “Blurred Lines” made so many headlines. The song was controversial from its release date, featuring arguably misogynistic, “rapey” lyrics,[282] and a music video that was removed from, and then censored on, YouTube for violating the site’s policies regarding nudity.[283] Then came the copyright controversy. Marvin Gaye’s family accused Thicke and Williams of copying the “feel” and “sound” of Gaye’s “Got to Give It Up,” [284] the song Thicke even publicly noted as the influence for “Blurred Lines.”[285] Therefore, the issue in the “Blurred Lines” case revolves around whether Thicke and Williams illegally interpolated original elements of Marvin Gaye’s musical composition.

In response to the allegations, Thicke and Williams sought a declaratory judgment that “Blurred Lines” did not infringe upon “Got to Give It Up.”[286] When that failed, and inevitably backfired with the Gaye family filling a cross-complaint, the plaintiffs moved for summary judgment, claiming that the eight alleged similarities between the songs are based on elements of Gaye’s sound recording that are not part of the copyright that Gaye’s family claims to own.[287] According to the Gaye family’s expert report, these eight“substantially similar features” include: (1) the signature phrase; (2) hooks; (3) hooks with backup vocals; (4) the core theme in “Blurred Lines” and the backup hook in “Got to Give It Up”; (5) backup hooks; (6) bass melodies; (7) keyboard parts; and (8) unusual percussion choices.[288] The family’s expert also alleged that the songs share similar “departures from convention such as the unusual cowbell instrumentation, omission of guitar and use of male falsetto.”[289]

Though most of these elements are not part of the underlying composition, and the melodies of the two songs are arguably completely different (“one’s minor and one’s major. And not even in the same key”[290]), the court found that the Gaye family made a sufficient showing of substantial similarity to satisfy the extrinsic test and the issue went to trial.[291] The jury found against Thicke and Williams with a judgment topping $7.3 million, one of the largest ever in music copyright history.[292] While Thicke’s and Williams’s fortune and fame seem unlikely to be tarnished, with each of them earning over $5 million for the song itself, [293] the verdict shows how misguided results can be under our infringement tests. In misapplying the already confusing Ninth Circuit test for substantial similarity, the court protected musical style and genre under the guise of protecting an original combination of elements. As one composer and producer put it, the court “made it illegal to reference previous material. . . . I’m never going to come up with something so radically different that it doesn’t contain references to something else.”[294] While the creators of “Blurred Lines” have already decided to appeal the verdict,[295] if it is upheld, artists on the margin, who are less willing to take on legal fees and whose work is less in tune with romantic authorship, will be discouraged from “creating any new songs that evoke the feel of the music that inspired them in their youth.” Further, “with the length of copyright we have these days, artists who want to feel confident that their musical influences are in the public domain are going to have to go all the way back to ragtime.”[296]

The verdict reached is procedurally problematic for a few reasons. Under the current Ninth Circuit test, the court erred in allowing the jury to hear excerpts of the sound recordings for “Got to Give It Up” and “Blurred Lines,”[297] which contained elements that are not part of Gaye’s written composition and, therefore, not part of his copyright.[298] Though the background chatter, party noise, and percussion, are common to both songs and contribute to the instinctive feeling that “Blurred Lines” “reminds” us of “Got to Give It Up,” they are totally irrelevant to the issue of substantial similarity.[299] Yet, these elements seemed to sway the jury in reaching its verdict.

Many, if not all, of the elements in “Got to Give It Up” are unoriginal staples in funk music, from the walking down funky bass line to the falsetto and melisma hook,[300] but Judge Kronstadt, applying the extrinsic test on summary judgment, held that a genuine issue of material fact existed either on the similarity to protectable elements, or on the similarity to an original combination of elements.[301] To show that the allegedly infringing elements are unprotectable scènes à faire, the defendants’ expert cited multiple prior songs that used many of the same allegedly infringing elements.[302] Judge Kronstadt still ruled that the testimony was not sufficient, citing Swirsky for the idea that the defendants failed to show that the plaintiff’s work is “more similar” to prior works than it is to the defendants’ work.[303] Yet the issue in Swirsky, and here, was whether the individual elements are scènes à faire, not whether the works as a whole are unoriginal.[304] Thus, Kronstadt’s reliance on Swirsky’s “more similar to prior works” language is misplaced, and bypasses actual consideration of the protectability of the elements themselves.[305] The court thereby failed to dispose of the issue of whether the defendants copied original elements on summary judgment, leaving the question to the jury.

In erroneously applying the “more similar” to prior works test to the works as a whole, Kronstadt further failed to consider the actual test for copying a combination of unprotectable elements: whether the works are “virtually identical.” As Pharrell and Thicke argued, the requirement for originality in a combination of unoriginal elements is much more stringent than the usual “some minimal level of creativity.”[306] The elements need to be “numerous enough and their selection and arrangement original enough that their combination constitutes an original work of authorship.”[307] Furthermore, infringement can only be found if the defendant’s work is “virtually identical” to the copyrighted work.[308] While many might say the two songs sound similar, with entirely different melodies and lyrics, one cannot maintain that the songs are “virtually identical,” especially when the misappropriation inquiry is supposed to be based solely on the written melodies, chords, and lyrics.[309] Even still, copying that is virtually identical may fall within the merger doctrine – that is, it may be an unavoidable product of using the same idea, which as discussed above is not a protectable interest.[310]

Because Judge Kronstadt determined that Pharrell and Thicke failed to meet the stifling burden of “more similar” to prior works on summary judgment, the final decision went to the jury to be based on their spontaneous reaction unguided by musicologists and uninstructed by the court on the “virtually identical” standard. Thus, the simple instinct that “Blurred Lines” felt like “Got To Give It Up” determined the verdict, regardless of the law. Yet, if the “feel of a work is sacrosanct,” many songs would be illegal for simply being a part of a “derivative and rigid genre,” as is often the case in music.[311]

When genres and subgenres have been inspired by one song, privatizing common elements can only do harm. Musicians like Marvin Gaye still have the incentive to create, especially with the notoriety and revitalized publicity that comes with a new artist referencing an old song. A reference from today’s stars can make a music legend. Snoop Dogg’s cover of Slick Rick’s 1985 hit “La Di Da Di,” and Notorious B.I.G.’s sample of it in “Hypnotize Me,” made Slick Rick’s song the most sampled rap song of all time. Dr. Dre’s use of the drum-break from The Wintons’ “Amen, Brother” started a narrative that led this little known band to have arguably the most sampled song of all time. The Amen Break has been used in over 1,700 songs and has become the basis for drum-and-bass and jungle music.[312]

Despite the notoriety and financial benefits of being sampled, by looking to our skewed notions of authorship, both judges and juries alike are trigger-happy to find copyright infringement when there is so much as a waft of homage. Their intentions are well wrought in a desire to protect the “genius” of artists loved by generations. However, that desire is based on notions of fairness, and copyright was never meant to protect artists’ moral rights in their works. Incentives to create have always been the only concern, balanced against maintaining public access to creative works. Yet “discussions of copyright and its goals frequently conflate the compensatory and control aspects of copyright on the incentive side.”[313]

If we expand the copyright in an original selection and arrangement to find infringement when any song includes just some of its elements, the blurry standard determining how many elements count as copying will result in risk-averse content creators. When nothing is truly original or avant-garde in music, the safe route in creating content is not clearly laid out either. Even if artists can be completely progressive, it is not clear they will be rewarded for doing so. Popular content is popular for a reason, and the benefits of creating it will be limited to the lucky first comers who hold a monopoly on elements they did not even create simply because they strung them together first. The success of “Blurred Lines” might be due, at least in part, to its ability to incorporate so many nostalgic funk clichés at once, all in the context of a modern pop song. Moreover, the enjoyment derived by the public from having an expanse of options to choose from will be lost. Of course, the effects of prohibiting “inspired works” will touch on industries beyond music, covering the fine arts, dance, and anything covered by copyright.

III. Considering Dissection and Reversed Questions of Law and Fact

When copyright infringement is meant to prohibit the copying of protectable elements of a work, it seems that the audience test, unguided as it is, can “play no useful role” in fulfilling the goals of copyright law.[314] Expert testimony and analytic dissection are necessary to maintain the distinction, both at the copying and misappropriation stages of the inquiry. More specifically, courts should be informed fully of the broader contexts within which specific artistic works are created.

While some courts recognize Arnstein’s limits and allow expert testimony when works are of a “highly technical nature,” thus far only computer software has met this characterization.[315]

The courts’ unwillingness to see “aesthetic arts, such as music, visual works or literature” as sufficiently complex to warrant the aid of expert guidance is merely an artifact of indoctrinated opinions on the nature of aesthetics. Yet with culture becoming ever more aware of its recombinant nature,[316] and the arts becoming increasingly technical in the age of digitalization, it is necessary to alter the current infringement tests in order to encompass varying artistic theories.

Courts need to act as gatekeepers, preventing an onslaught of needless and threatening litigation by deciding whether there is a cognizable claim of misappropriation and identifying the unprotectable elements in a work before sending the intuitive issue of copying to the trier. There have been many scholarly suggestions for reforming the two-part test. This part considers adoption of a proposal first suggested by Nimmer and expanded by Professor Lemley: extending the “abstraction-filtration-comparison” (AFC) test for computer software to all cases deciding copyright infringement.[317]

First adopted by the Second Circuit in Computer Associates v. Altai,[318] the AFC test requires the court to identify which aspects of the program constitute its expression versus ideas (“abstraction”), remove from consideration unprotectable ideas (“filtration”), and only then compare whether the defendant copied the protectable elements (“comparison”).[319]

Following Altai, the AFC test was widely adopted in determining substantial similarity in the non-literal aspects of computer programs. However, as Nimmer notes, “there is no reason to limit it to that realm.”[320] In fact, the AFC method is more consistent with the Supreme Court’s definition of infringement (along with ownership of a valid copyright) as the “copying of constituent elements of the work that are original.”[321] Whether expert testimony is permitted, and to what extent, is left to the discretion of the district court.[322] This approach allows the court flexibility in determining the necessity of expert testimony depending on the complexity and nature of the works at issue. A court might therefore determine that the issue of substantial similarity in a case involving generic pop songs would best be determined by lay jurors, but in most cases where songs mix genres and traditions, the court would need the testimony of experts well versed in the nature of the art at issue.

Given that nearly every circuit already permits expert testimony on both prongs of the copyright infringement test in software cases,[323] adopting the AFC test wholeheartedly seems like a more feasible path to altering the current framework than other suggested reforms.[324] In fact, the Tenth Circuit already applies the AFC test for all copyright infringement cases, and the Sixth Circuit uses a variation of the Tenth Circuit test.[325] As discussed above, expanding the role of analytic dissection and expert testimony will result in better maintenance of the idea-expression distinction and respect for the limits to copyright protection. Moreover, allowing expert testimony in guiding the jury will solve the problem of juries having trouble disregarding the testimony they hear on the copying-in-fact prong in deciding misappropriation.

Professor Lemley suggests a further change: reserving the issue of misappropriation to the court as a question of law, to be determined on summary judgment. The role of the jury would be preserved in leaving the intuitive issue of whether the defendant copied as a question of fact.[326] According to Lemley, this seems to be the “practical import” of the AFC test.[327] Reversing the judge-jury role would better serve the interests of copyright, as the issue of which elements are protectable and unprotectable in any given work implicates substantial policy considerations and interpretation of the law better suited for a judge well-versed in the law.

Moreover, by providing written judicial opinions, this altered framework would allow a reasoned jurisprudence to develop on the issue of which elements are protected. The transparency and predictability of the law would allow artists a safe harbor and guaranteed protections in creating new works, thereby preventing the chilling effects of a potentially devastating lawsuit.

Taking reform a step further, Professor Balganesh suggests reversing the order of the test.[328] That is, have the court determine misappropriation as a question of law on summary judgment, filter out what is sent to the jury, and then have the jury determine the question of copying-in-fact of “protected expression.” The jury would still hear evidence on the issue of “protected expression,” including expert testimony on whether elements are scènes à faire or original in the area or genre, and even on the ultimate issue of similarity between the works and probative similarity. The jury would also hear the judge’s own dissection from the first step.[329] Balganesh argues that reversing the framework would encourage disposition on summary judgment, providing judges with a real gate-keeping role.[330] Moreover, it removes the subjectivity of the audience test, which he argues was based in mere skepticism towards rules, judges, and the law.[331]

While reversing the ordering might offer the benefit of preventing the judge’s determination of misappropriation from improperly influencing the jury’s probative similarity analysis, almost every court has adopted some version of the Arnstein-Krofft test,[332] which might make it difficult to persuade courts to do exactly the opposite.

At least preliminarily, extending the AFC approach seems more feasible and the need for dissection and expert testimony most ripe. Allowing the judge to determine the question of misappropriation as a matter of law, while preserving the issue of copying-in-fact for the jury, removes the inherent subjectivity and messiness of the audience test while complying with the constitutional mandate of jury trials in copyright lawsuits.[333] Furthermore, it uses juries for the role they are best suited for: the intuitive and fact-dependent question of whether the defendant copied.

On both questions, expert testimony can aid the court in understanding the nature of the work at issue. The controversy over sampling, for example, focuses on the taking rather than the contribution. Yet, experts in musicology know that sampling can be transformative. As D.J. Shadow put it, sampling is a way to reintroduce a person’s music to people “in a completely different context than the way they originally intended.”[334] Mark Ronson similarly explained that through sampling artists insert themselves in a narrative and push that narrative forward.[335]

An expert might testify that sampling extends an African American oral tradition known as “signifying,” a type of word play that draws attention to the cultural significance of the words.[336] For example, an expert could explain that the Beastie Boys’ line “I shot a man in Brooklyn just to watch him die” places Johnny Cash’s line “I shot a man in Reno just to watch him die” in a new setting – a hip-hop anthem called “Hello Brooklyn” – thereby removing the ideas of consequences and regret found in the context of Cash’s murder ballad.[337] An expert could note that while sampling might intuitively seem like theft, it is no more so theft than quoting a passage from another book in a novel.[338] Rather, sampling is a form of textual revision and a literary device through which text speaks to text.[339] Through sampling, 2 Live Crew relied on repetition of the key elements of “Oh, Pretty Woman” to achieve a parody of the song.[340]

Our current framework for copyright infringement fails to inform courts of the artistic merits behind our most illustrative and progressive artistic movements. In perpetuating the courts’ biases, copyright is its own worst enemy, stifling innovation where it is most likely to happen: on the margins.


Though the problems with copyright pose unique problems for music, these problems reflect the larger difficulties in our current copyright law. Our infringement doctrine inhibits progress by making “substantial similarity” the end-all-be-all test. The generalized focus on substantial similarity leaves homages and similarities inherent in the genre subject to findings of infringement because the law, like our culture, sees derivative works as unworthy. But what is originality? As music producer and sampling guru Mark Ronson said in an interview with NPR: “Well, what’s the T.S. Eliot quote, which apparently he even stole from Picasso, about ‘Genius steals…?’ ‘Good artists borrow, great artists steal.’”[341] How we define “Progress of Science and useful Arts” is rooted in the dominant cultural beliefs of aesthetic value, but innovation comes with embracing progressive ideas and more of them.[342] The law needs to protect a new kind of originality, one that might not fit the Romantic mold, but that re-conceptualizes and re-frames preexisting works and provides listeners with a different experience than the original. The most important question should be what the newcomer added: how they took influences to make something new, because progress never involves creating something from nothing.

* J.D. Class of 2016, New York University School of Law; B.A. Philosophy, Politics, and Law, magna cum laude, Binghamton University. The author would like to thank Professor Christopher Sprigman for his inspiration and guidance and the 2016-2017 Editorial Board of the NYU Journal of Intellectual Property and Entertainment Law for their encouragement.

[1] See Bleistein v. Donaldson Lithographing Co., 188 U.S. 239, 251 (1903) (“It would be a dangerous undertaking for persons trained only to the law to constitute themselves final judges of the worth of pictorial illustrations, outside of the narrowest and most obvious limits.”).

[2] See generally Alfred C. Yen, Copyright Opinions and Aesthetic Theory, 71 S. Cal. L. Rev. 247, 251, 285–86 (1998) (arguing that “judges necessarily make decisions of aesthetic significance in copyright” as implicit in determining whether elements are “public domain,” or if they are “protected material.” Moreover, substantiality rests on “how sensitive” courts are to the degree of quantitative and qualitative similarity between two works).

[3] Id. at 251 (“[S]ince no aesthetic perspective can be neutral and all-encompassing, aesthetic bias becomes inherent in copyright decisionmaking . . . .”).

[4] See, e.g., Nichols v. Universal Pictures Corp., 45 F.2d 119, 121 (2d Cir. 1930); Brandir Int’l, Inc. v. Cascade Pac. Lumber Co., 834 F.2d 1142 (2d Cir. 1987); Carol Barnhart Inc. v. Economy Cover Corp., 773 F.2d 411, 415–18, 420 (2d Cir. 1985).

[5] See Olufunmilayo B. Arewa, From J.C. Bach to Hip Hop: Musical Borrowing, Copyright and Cultural Context, 84 N.C. L. Rev. 547, 565(2006) (“Although originality is not explicitly included in the Intellectual Property Clause of the U.S. Constitution, it is a fundamental assumption of current copyright law that originality is implicitly mandated by the Constitution’s references to ‘authors’ and their ‘writings.’”) (citations omitted); see also U.S. Const. art. I, § 8, cl. 8 (“The Congress shall have 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.”).

[6] See generally Amy B. Cohen, Masking Copyright Decisionmaking: The Meaninglessness of Substantial Similarity, 20 U.C. Davis L. Rev. 719, 720 (1987).

[7] H.R. Rep. no. 94-1476, at 61-62 (1976); S. Rep. No. 94-473, at 71 (1975) (“[A] copyrighted work would be infringed by reproducing it in whole or in any substantial part, and by duplicating it exactly or by imitation or simulation.”) (emphasis added).

[8] Davis v. Gap, Inc., 246 F.3d 152, 173 (2d Cir. 2001).

[9] Shyamkrishna Balganesh, The Questionable Origins of the Copyright Infringement Analysis, 68 Stan. L. Rev. 791, 805 (Forthcoming 2016).

[10] Atari, Inc. v. N. Am. Phillips Consumer Electric Corp., 672 F.2d 607, 614-15 (7th Cir. 1982).

[11] Melville B. Nimmer & David Nimmer, Nimmer on Copyright § 13.03[a][i] (2002).

[12] Scène à faire is the notion that certain similarities in the basic idea will require similarities in the expressions used to develop that idea. For example, “if two scenarios wish to treat the unprotected idea of police life in the South Bronx, one court has determined it would only be natural to depict ‘drunks, prostitutes, vermin and derelict cars,’ juxtaposed against hard drinking Irish cops chasing fleeing criminals.” Id. (citations omitted).

[13] Mark A. Lemley, Our Bizarre System for Proving Copyright Infringement, 57 J. Copyright Soc’y U.S.A. 719, 738 (2009) (alteration in original).

[14]See infra Part II.A (discussing how this need for, and lack of, special expertise in copyright makes the “ordinary observer” test a poor extension of the rationale underlying the negligence standard used in areas such as tort law).

[15] Balganesh, supra note 9, at 805 (explaining that the misappropriation inquiry requires procedural, substantive, and theoretical considerations).

[16] See generally Arewa, supra note 5 (noting that American copyright law is founded on the unrealistic conception that creativity necessarily comes from a place of pure autonomous genius).

[17] Lemley, supra note 13, at 739 (citing Ann Bartow, Copyrights and Creative Copying, 1 U. Ottawa L. & Tech. J. 77 (2003) (suggesting that this has been the result)).