By Graham Ballou*

A pdf version of this article may be downloaded here.

In Our Bizarre System of Copyright Infringement, Mark Lemley argues that copyright law is exactly backwards: we should allow the jury to decide whether a defendant has copied the plaintiff’s work as a factual matter, and leave the question of improper appropriation to experts.[FN1] Expert testimony on substantial similarity would, at the least, clear judicial fog at this stage of a copyright infringement analysis: courts could abandon the fiction of an objective, “ordinary observer” perspective – the controlling test for substantial similarity – and allow specialists to conduct what is in fact a highly technical analysis.  But Lemley overlooks two realities of our copyright system.  First, the court – not the jury – most often has the final word on substantial similarity.  If a defendant loses at trial, she can seek de novo review of this issue.[FN2] Second, courts’ decisions on substantial similarity are much less technical than their elaborate rhetoric would suggest.  Despite the myriad judicial tests that courts employ – from “ordinary observer” to “more discerning ordinary observer,” from “filtration” to “total concept and feel” – their decisions ultimately stampede a formal analytical framework in favor of a basic “we know it when we see it” approach.

Part I of this paper gives a brief overview of the Second Circuit’s substantial similarity doctrine.  Part II examines application of this doctrine at the district court level, where summary judgment opinions from 2008 to 2010 display wide judicial latitude in the framing of the test.  In Part III, I propose an alternative.  Rather than allow expert testimony at this stage of an infringement analysis, copyright law should discourage summary judgment on and de novo review of substantial similarity.  By empowering the jury on this inherently subjective question, courts would free themselves of what has become an unworkable doctrine.

I.  Substantial Similarity in the Second Circuit

A. The Ordinary Observer

In Arnstein v. Porter, a 1946 case that remains good law today, Judge Frank held that the test for substantial similarity “is the response of the ordinary lay hearer.”[FN3] Plaintiff Arnstein, a serial litigant, had accused Cole Porter of stealing his songs.  The Court heard expert testimony on whether Porter, as a factual matter, copied Arnstein’s works.[FN4] But as to whether such copying touched the protectable elements of Arnstein’s expression, and therefore rose to improper appropriation, it deemed expert testimony inapplicable.[FN5] “The question…is whether defendant took from plaintiff’s work 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 that belongs to plaintiff.”[FN6] The perspective of this lay observer remains the primary lens through which courts examine substantial similarity.  The contemporary framing of this test asks “whether ‘the ordinary observer,’ unless he set out to detect the disparities [between the parties’ works], would be disposed to overlook them, and regard their aesthetic appeal as the same.”[FN7]

When the allegedly infringed work contains both protectable and unprotectable expression, the Second Circuit uses a “more discerning ordinary observer” test.[FN8] Under this heightened level of scrutiny, a court “must attempt to extract the unprotectable elements from…consideration and ask whether the protectable elementsstanding alone, are substantially similar.”[FN9] Unfortunately, the central assumption of this test – that certain works contain unprotectable expression, and others do not – creates a problem.  In a literal sense, every work incorporates unprotectable expression, if a court only examines the work at a sufficiently high level of detail.  Even though the substance of Starry Night is “original” to van Gogh, for example, he could not claim protection in the blue-white swirl of brushstrokes used to depict the Milky Way.  This element is an “idea,” not “expression,” and therefore unprotectable.   But should its presence in the painting, all other factors aside, require a court to view the work with a “more discerning” eye?  Whether the “more discerning” test should apply begs the very question that the substantial similarity test purports to address.  It forces a game of line drawing between idea and expression, but rehashes the analysis as a threshold inquiry.

Two examples illustrate the problem.  In Knitwaves, Inc. v. Lollytogs Ltd. (Inc.),[FN10] the defendant urged the Court to apply the “more discerning” test to plaintiff’s designs for children’s sweaters, which “incorporated design elements intended to express a ‘fall’ motif, such as leaves, acorns, squirrels, and the like.”[FN11] Though these elements, standing alone, are unprotectable, the Court applied only the “ordinary observer” test because the plaintiff had claimed protection in the combination of these elements, which in fact was protectable.[FN12] The Court distinguished Folio Impressions, Inc. v. Byer California, a previous Second Circuit decision and the origin of the “more discerning” language.[FN13] Folio Impressions concerned fabric designs of stylized roses on a complex background that the designer had photocopied from a document in the public domain.[FN14] The Folio Court “extracted” this photocopied background and compared only the remaining elements of the plaintiff’s work to the defendant’s, concluding that the two works were not substantially similar.[FN15] The plaintiff’s sweater designs in Knitwaves, on the other hand, contained no such wholesale appropriation of unprotected expression.[FN16] Knitwaves, then, suggests that the “more discerning” test only applies where (i) plaintiff has appropriated expression verbatim from the public domain and (ii) plaintiff claims protection in the elements themselves, not in their selection and arrangement.

In contrast, the Court in Laureyssens v. Idea Group, Inc.[FN17] interpreted Folio Impressions as requiring the “more discerning” test where functional considerations lead designers to certain inevitable choices.  These choices result in design elements that the “more discerning” observer must overlook – here, the shapes of plaintiff’s three-dimensional puzzles.  The Court wrote, “[I]n order to express the idea of a perfect hollow cube puzzle that can also be assembled in flat form, a designer must use pieces that interlock through fingers and notches cut at right angles.”[FN18] Applying the “more discerning” test, the Court found no substantial similarity.  Yet the rationale at work in this decision – that the “more discerning” test applies when the plaintiff claims protection in functional elements of a chosen subject – sits uneasily alongside that of Knitwaves.  Starry Night once again illustrates the tension.  The Knitwaves Court would deem the “more discerning” test inappropriate; though van Gogh’s use of a blue-white swirl to depict the Milky Way is unprotectable, he did not “extract” this expression wholesale from the public domain.  The Laureyssens Court, however, would interpret this same element as necessary to van Gogh’s choice of subject matter – an impressionistic rendition of a starry sky.  The “more discerning” test would therefore apply.

B. Filtration

Regardless of whether a court adopts the “ordinary observer” or “more discerning” perspective, the mandate of copyright law – to protect original expression, but not ideas[FN19] – requires the factfinder to compare only the protectable elements of plaintiff’s expression to the defendant’s alleged copy.  “Filtration,” or the process of separating protectable from unprotectable expression, and has its roots in a 1930 opinion from Judge Hand:

Upon any work…a great number of patterns of increasing generality will fit equally well, as more and more of the incident is left out.  The last may perhaps be no more than the more general statement of what the [work] is about, and at times might consist only of its title; but there is a point in this series of abstractions where they are no longer protected, since otherwise the [author] could prevent the use of his “ideas,” to which, apart from their expression, his property is never extended.[FN20]

The modern incarnation of Judge Hand’s abstractions formula is more technically refined but no less arbitrary.  Williams v. Crichton, a literary infringement case, typifies the analysis.  After the release of the film Jurassic Park, Williams sued a host of defendants for infringing his copyright in several fictional children’s books.[FN21] With titles like Dinosaur World and Lost in Dinosaur World, the plaintiff’s books described the adventures of Young Tim in “an imaginary present day man-made animal park for dinosaurs.”[FN22] The district court awarded summary judgment for defendants, and Williams appealed.

After referencing Judge Hand’s “abstractions” formula, the Court identified the proper level of “generality” from which to analyze the works.  Though the works seemed analogous at first blush, the Court overlooked broad similarities and focused on discrete elements – theme, characters, plot, sequence, pace, and setting – and filtered unprotected expression from the plaintiff’s works at each turn.  It dismissed any similarity between the works’ themes, for example, because the similarities related only to the “unprotectable idea of a dinosaur zoo.”[FN23] It dismissed similarities in settings with equal brevity, based on the somewhat arbitrary level of abstraction at which it chose to examine to examine this element: “While both the Dinosaur World books and the Jurassic Park works share a setting of a dinosaur zoo or adventure park, with electrified fences, automated tours, dinosaur nurseries, and uniformed workers, these settings are classic scenes à faire that flow from the unprotectable concept of a dinosaur zoo.”[FN24] The Court arrived at similar conclusions as regards characters and plots.[FN25] At each stage in its analysis, then, the Court first determined which level of “abstraction” should govern, and this decision itself controlled its finding on substantial similarity.

Returning once more to Starry Night, we can draw two conclusions.  First, as a practical matter, the distinction between the “ordinary observer” and “more discerning” perspective matters little to the ultimate conclusion of fact, provided a court incorporates some variation of the filtration process.  If a court were to follow Knitwaves and consider Starry Night from the less exacting “ordinary observer” perspective, it would still, as part of the filtration process, exclude van Gogh’s blue-white swirl of brush strokes before comparing his work to the defendant’s.  As such, the “more discerning” test is largely a judicial fiction, in that a court should arrive at the same result regardless of which perspective it adopts.  Second, the amount of protection available to van Gogh depends entirely on an “arbitrary” exercise of judicial discretion.  A court might define the level of generality appropriate to Starry Night broadly – as a painting of Saint-Rémy, at night, in heavy brush strokes – and thereby award van Gogh a virtual monopoly over the subject matter.  More likely, of course, the court would define the level of generality in narrower terms, and compare Starry Night to an alleged infringement in such details as van Gogh’s precise framing of the town, his inclusion of a cypress tree in the foreground, the number of homes that appear in the work, etc.  The Second Circuit’s filtration doctrine provides no guidance as to where, between these extremes of generality, a court should situate the analysis.

C. Total Concept and Feel

In 1970, Roth Greeting Cards accused the United Card Company of copying seven of its greeting card designs.[FN1] The district court found that Roth’s designs lacked copyrightable expression, and dismissed.[FN26] The Ninth Circuit reversed, and in its opinion devised a tool for substantial similarity analysis that runs contrary to that most basic tenet of copyright law – that copyright does not protect ideas.  Ever since, the “total concept and feel” test (“TC&F”), as it is known, has allowed courts in certain circumstances to avoid the difficult line drawing between expression and idea.

The Roth Court first admitted that the actual text of plaintiff’s greeting cards – “common and ordinary English words and phrases” – was in the public domain prior to plaintiff’s use and thus not original expression.[FN27] Roth, however, did not seek protection in the phrases themselves, but rather in the “arrangement of the words, their combination and plan, together with the appropriate art work” contained therein.[FN28] After agreeing with the district court that the specific elements of Roth’s cards, taken alone, were unprotectable,[FN29] the Court nonetheless held for the plaintiff on account of the cards’ overall similarities: “It appears to us that in total concept and feel the cards of United are the same as the copyrighted cards of Roth…[T]he characters depicted in the art work, the mood they portrayed, the combination of artwork conveying a particular mood with a particular message, and the arrangement of the words on the greeting card are substantially the same as in Roth’s cards.”[FN30] The filtration test would have required the Court to overlook each of these discrete elements in comparing the parties’ works.  Indeed, the “total concept and feel” of Roth’s greeting cards would have struck Judge Hand as an untenably “high” level of generality at which to compare the designs.  TC&F, by contrast, allowed the Ninth Circuit to sidestep the strictures of filtration and reach what it viewed as an equitable decision – but a decision precluded by copyright doctrine to date.  In essence, its TC&F analysis took to an extreme the Arnstein proposition that substantial similarity should be judged from the perspective of a lay observer.

Professor Nimmer traces the Second Circuit’s first use of TC&F to a 1976 case involving a children’s book.[FN31] Plaintiff, author of My Mother Is the Most Beautiful Woman in the World, claimed that the producers of Sesame Street had copied her book in a televised skit.[FN32] Both parties’ works, it turns out, were based on an underlying story in the public domain.[FN33] But the Court’s substantial similarity analysis mirrored that in Roth: “We must first note that both stories, intended for children, are necessarily less complex than some other works submitted to the pattern analysis [from Nichols]…Therefore, in addition to the essential sequence of events, we might properly consider the ‘total concept and feel’ of the works in question.”[FN34] The Court then examined the TC&F of each party’s work, despite its previous determination that the underlying public domain story was unprotectable.  Unlike in Roth, where filtration and TC&F led to contradictory results, both tests in Reyher pointed toward non-infringement.[FN35] Nonetheless, the Court managed to reincorporate unprotectable expression into its substantial similarity analysis by way of TC&F – this time, under the rationale that the works at issue appealed to children, who might be less likely to pick out subtle variations.

Whereas Roth and Reyher demonstrate a “pure” application of TC&F, under which a court compares the entirety of the plaintiff’s work to defendant’s, the test also appears, somewhat illogically, in combination with filtration.   Most often in the context of literary works, a court will include TC&F as an “element” alongside other variables (e.g. plot, character) in its substantial similarity analysis.  In Williams, as we have seen, the Court filtered unprotectable elements from plaintiff’s work before comparing it to defendant’s.[FN36] But it included TC&F within the context of this filtration framework: “[W]e examine the similarities in such aspects as the total concept and feel, theme, characters, plot, sequence, pace and setting of the Dinosaur World books and the Jurassic Park works.”[FN37] Its conclusion on the issue of TC&F is striking:

[T]he total concept and feel of the two works differ substantially.  The Jurassic Park works are high-tech horror stories with villainous characters and gruesome bloodshed.  Books II and III of Dinosaur World series, by contrast, are adventure stories and, although suspenseful in places, have happy endings…The total concept and feel of the Jurassic Park works is of a world out of control, while Williams’s Dinosaur World is well under control.[FN38]

Such elements as “adventure,” “suspense,” a “happy ending” and a “world under control” are unprotectable elements of plaintiff’s Dinosaur World books.  In Judge Hand’s abstractions rhetoric, these levels are too “general” to constitute original expression.  TC&F, however, allowed the Court to adopt a birds-eye view of the plaintiff’s works; it nudged the Court away from the technical rhetoric of filtration, and toward the truly lay perspective that Judge Frank envisioned in Arnstein.

II.  District Court Application of Second Circuit Law

Not surprisingly, district court application of the Second Circuit substantial similarity “test” is a mess.  A court may use the “ordinary observer” or “more discerning” test, and cherry pick among Second Circuit doctrine to support either choice in almost any circumstance.  Under either rubric, it may further choose between filtration, “pure” TC&F, or filtration that incorporates TC&F as an “element” of the broader test.  Summary judgments opinions at the district court level from 2008 to 2010 display at least seven permutations of these factors.[FN39] In sum, despite the Second Circuit’s attempt to objectify the question of substantial similarity, district courts ultimately manipulate the test to reach the conclusion they feel is most equitable.  Their analyses stampede any formal analytical framework, and their conclusions reflect the true perspective of a lay observer.  They are, in a word, using common sense.

The following sections highlight the more common approaches to substantial similarity among the opinions studied.  I argue that each example demonstrates the practical failure a cohesive “substantial similarity” doctrine, and the common-sense approach behind courts’ analyses on this issue.

A. Category 1:  Ordinary Observer – Filtration

Four cases[FN40] combined the Arnstein ordinary observer test with filtration.  Each case merely cites the Arnstein standard in passing without first explaining why the “more discerning” test does not apply,[FN41] even though works at issue – architectural plans, song lyrics, and a work of fiction – admittedly contained “unprotectable expression.”  Because each court incorporates filtration, however, the cases establish a formal framework identical to the “more discerning ” test.  In each instance, the court found no substantial similarity and awarded summary judgment to defendant.

This category of analysis – ordinary observer combined with filtration – takes its cue from the Nichols-Arnstein line of Second Circuit doctrine.  Though the analyses appears straightforward, we must ask why these courts chose not to apply the “more discerning” test, “pure” TC&F, or TC&F as an “element” of the broader filtration analysis.  The answer to the first question seems clear: a “more discerning” perspective would only have favored the defendant by heightening the level of scrutiny applied to the plaintiffs’ works, thus complicating the analysis without altering the result.  An application of TC&F, however, might very well have materially affected the outcomes.  Though one case, Blakeman, cited Williams for the proposition that a court may consider TC&F alongside other variables in the substantial similarity analysis of a literary work,[FN42] its actual application of the test was silent on this issue.

One possible explanation is that the inclusion of TC&F would have run contrary to the merits, which, according to the Blakeman Court, favored the defendant.  It noted at the outset that the “ideas” of the works are “entirely different.”[FN43] Of course, a difference of underlying ideas will not save a defendant from liability if it otherwise appropriates the plaintiff’s protected expression.  But the Court’s instinctual response to the parties’ works – “They’re just too different!” – is reflected in its construction of the substantial similarity test, a construction that allows it to easily reach the conclusion it feels is most logical.

B. Category 2:  Ordinary Observer – Total Concept and Feel (“Pure”)

Two cases[FN44] combined the Arnstein ordinary observer test with “pure” TC&F.  The first, R.F.M.A.S., Inc. v. Mimo So, concerned jewelry designs, and the Court explicitly found the “more discerning” perspective inapplicable:

The combinations of common elements have resulted in designs that are original and protectable in their entirety.  It makes no sense…to look at the designs by excluding elements such as cable and gemstones and consider what is left.  Consequently, plaintiffs would only have to satisfy the ordinary observer test and show that the ordinary observer would tend to overlook the similarities between the plaintiff’s and defendant’s jewelry designs.[FN45]

The second, Axelrod & Cherveny, Architects v. T.& S. Builders, Inc., involved architectural plans, and the Court did not explicate its reason for avoiding the “more discerning” perspective.  Neither court granted summary judgment.

In both opinions, application of “pure” TC&F resulted in relatively brief substantial similarity analyses.  In R.F.M.A.S., for example, the Court compared the works in a single paragraph.  It concluded, “While the…Defendants point to discrepancies between the two works, the jewelry designs need not be absolutely identical [to be substantially similar].”[FN46] Since neither court felt able or compelled to filter protectable expression, their decisions included straightforward and sweeping comparisons of the parties’ works.

C. Category 3:  Ordinary Observer – Filtration – Total Concept and Feel as “Element”

Four cases[FN47] combined the ordinary observer test with filtration, but included TC&F as an “element” of the analysis.  Not surprisingly, three of these cases concerned novels or screenplays.  Each drew heavily on Williams for the proposition that courts should consider TC&F alongside other elements in the filtration analysis of a literary work.[FN48] All three courts dismissed plaintiff’s claims.

In Cabell, for example, the plaintiff claimed that the defendant, producer of You Don’t Mess with the Zohan, had copied its comic books about the escapades of Jayms Blonde, a “gay former U.S. Navy SEAL who divides his time between work as a hairdresser and service as a secret agent.”[FN49] The Court conducted a traditional filtration analysis, then turned to TC&F: “[T]he total concept and feel of the Blonde Works are distinct from the Zohan films.  The Jayms Blonde stories are parodies of the James Bond stories, and much of the humor is double entendre and innuendo.  In contrast, You Don’t Mess With The Zohan derives much of its humor by exaggerating Arab and Israeli stereotypes.”[FN50] These broad generalities – the use of double entendre and innuendo, the status of a work as a parody – are certainly not copyrightable elements of Cabell’s work, nor did Cabell claim that the defendant had unlawfully copied them.  The Court’s use of TC&F as an “element” in its filtration analysis, therefore, focused its attention on the works’ broad ideas rather than on plaintiff’s protectable expression, even though it had already determined that the defendant had not improperly appropriated the plaintiff’s protected expression.

D. Category 4: More Discerning Ordinary Observer

Two cases[FN51] used the “more discerning ordinary observer” test without any reference to TC&F.  Recognizing at the outset that the works at issue contained both protectable and unprotectable expression, the Court in these cases “ignor[ed] those aspects of the work that are unprotectible” in making the substantial similarity comparison.[FN52] Since the “more discerning” perspective implies filtration, the analyses in this category closely mirror those in Category 1 (ordinary observer combined with filtration).

E. Category 5:  More Discerning Ordinary Observer – Total Concept and Feel as “Element”

Five cases[FN53] used “more discerning” perspective, but included TC&F as an “element” of the filtration test.  Since the “more discerning” perspective mirrors the “ordinary observer” test combined with filtration, the analyses in this category closely resemble those in Category 3.

Porto v. Guirgis is typical of this category.  Plaintiff, author of Judas on Appeal, accused the defendant of improperly appropriating his novel into a play entitled The Last Days of Judas Iscariot. Both works envisioned Judas’ trial, at which a wide variety of commentators debated his forgiveness.[FN54] The Court initiated its substantial similarity analysis by filtering unprotectable elements from plaintiff’s novel.  It found the theme of “Judas standing trial” to be “a highly general description of both works, but…[not to] cover the protectable expression in each work.”[FN55] It next filtered characters common to both works – Satan, Caiaphas, Pontius Pilate, and Peter[FN56] – as unprotectable scenes à faire that “necessarily result from the choices of a setting or situation.”[FN57] The Court further filtered themes such as “predestination” and “forgiveness” as “stock themes in Christian theology.”[FN58] Ditto for setting, plot and language.[FN59] After filtration, no alleged similarities remained: “Stripped of unprotectible elements – such as the biblical characters and the biblical story – the works are not substantially similar.”[FN60]

But before reaching this conclusion, the Court reincorporated certain unprotectable elements into its discussion.  Under the heading of “Overall Thrust, Mood and Feel,” it noted that the “substantial similarity test is ‘guided by comparing the total concept and feel of the contested works,’”[FN61] and for a moment, left filtration by the wayside:

The overall feel of both works is so distinct that an ordinary observer would not regard the defendants’ play as substantially similar to the plaintiff’s novel.  One particularly striking point that underscores the substantially different feel between the two works is the differences between the endings of the works.  The plaintiff’s novel ends in a practicable, straight-forward manner…The defendant’s play, on the other hand, ends in an entirely different and very powerful manner.[FN62]

In a strict application of the filtration test, such broad references to “overall feel,” or the fact that an ending is “predictable and straight-forward,” would be immaterial.

F. Categories 6 and 7:  No Observer

Three cases[FN63] dove into the substantial similarity analysis without establishing either an “ordinary observer” or a “more discerning ordinary observer” framework.  The lack of a defined perspective in these cases is likely a factor of the highly technical nature of the works at issue: an historical novel based on the death of Pope John Paul I,[FN64] software code,[FN65] and classified advertisements.[FN66] Plaintiffs’ copyrights in all three instances were extremely “thin”; the perspective of a lay observer – more discerning or otherwise – seems logically less applicable.  As regards the classified advertisements, for example, the Court had to decide whether the font size, bolding of certain text, and use of certain categories in plaintiff’s publication gave rise to copyright protection at all.[FN67] The analysis in each of these cases, notwithstanding the lack of an “observer,” follows the traditional filtration approach.  Citing Williams, the Court faced with the historical novel further integrated TC&F as an “element” of its broader filtration analysis.

G. Analysis

The above seven categories demonstrate wide judicial discretion in the framing of the substantial similarity inquiry.  The flowchart below simplifies the distinctions between each category.  Most striking is the lack of any relationship between subject matter and analytical framework.  Literary works appear in Categories 2, 3, 4, and 5 – and Categories 6 and 7 if one considers computer code a “literary work.”  Jewelry designs appear in Categories 1 and 4.  Architectural plans appear in Categories 1, 3 and 5.  Even the relatively few summary judgment opinions studied display a diversity of approaches to a given subject matter.

In addition, the formal rhetoric of each category bears little relationship to the actual application of the substantial similarity test therein.  The analysis of a court under Category 2 (“ordinary observer” plus filtration, with TC&F as an “element” of the test) is, in practice, indistinguishable from that of a court in Category 4 (“more discerning” observer, with TC&F as an element of the test).  Category 3 is likewise indistinguishable from Category 5.  Category 1 – “ordinary observer” plus “pure” TC&F – stands out as on the only truly unique substantial similarity test.  But this test is unique less for what it includes – an examination of the plaintiff’s work writ large – than for what it omits: filtration.  In this category, the Arnstein “ordinary observer” approaches the works as a truly lay member of the public, and just as Judge Frank envisioned in Arnstein, bases its finding of improper appropriation on its “response.”  Though judicial discretion in the framing of the analysis means that courts in every category are, in reality, conducting this very same “lay” investigation, only Category 1 courts match word to deed.

Substantial Similarity Analysis Flowchart

III. Conclusion

Lemley responds to judicial confusion on the issue of substantial similarity by proposing that courts allow expert testimony on improper appropriation.[FN68] This would solve the problem of juries misunderstanding the distinction between protectable and unprotectable expression.[FN69] Alternatively, he proposes that the dissection of the unprotectable elements of a copyrighted work, and therefore the assessment of whether the defendant copied protectable expression, become a question of law rather than fact.[FN70] This, Lemley concludes, would “result in greater attention to the limiting doctrines of copyright law.”[FN71]

Lemley’s proposal overlooks the reality of analytic chaos on the issue of substantial similarity described in Section II.  If the substantial similarity “test” were less subject to judicial discretion, the inclusion of experts would indeed help both courts and juries parse protectable from unprotectable expression.  But as the law stands, application of the test is less responsible for judicial disarray than the number of analytic variations available.  Even assuming that a court allowed expert testimony on this issue, the court – not the experts – would continue to frame the inquiry, and in so doing, direct the outcome.  Further, because the procedural aspects of summary judgment and de novo appellate review most often place the court in the position of final arbiter on substantial similarity, our prime concern should not be jury confusion.  Rather, we should ask why our copyright system has placed the lay judgment of improper appropriation in the hands of the court at all.

As an alternative to expert testimony, we should discourage summary judgment and de novo review of substantial similarity.  This proposal is not as radical as it might first appear.  Because a plaintiff must demonstrate that its work is (i) original and (ii) contains some minimal degree of creativity before the court even reaches the issue of copying,[FN72] the system would continue to prevent plaintiffs from claiming protection in uncopyrightable material.  District courts could dismiss frivolous suits under this preliminary analysis.  Further, expert evidence on the issue of factual copying would provide the fact-finder with an objective perspective on the amount of material that the defendant appropriated from the plaintiff.  Though this perspective would not distinguish protectable from unprotectable expression – at least under the “probative” rubric that currently holds sway in the Second Circuit – it would nonetheless give the fact-finder a relative measure of the works’ similarities for use in its improper appropriation analysis.

Judge Frank intended his “ordinary observer” to be just that – ordinary.  The Circuit’s continued attempt to objectify the substantial similarity analysis has led to a variety of intellectually incompatible tests that remain as “arbitrary” as Judge Hand’s abstraction formula from 1930.  In the interest of calling a spade a spade, copyright law would be well served by re-empowering the jury on this inherently subjective inquiry, and freeing litigants of what has become a confused, unworkable and dishonest judicial doctrine.


* J.D. candidate, NYU School of Law, 2011; B.A., University of Chicago, 2006.  Graham will join the Corporate Department at Proskauer Rose, LLP in the fall of 2011.

[FN1] Mark Lemley, Our Bizarre System of Copyright Infringement, Stanford Public Law Working Paper No. 1661434, 3 Intellectual Prop. L. eJournal 105 (2010).

[FN2] Seee.g., Hamil America, Inc. v. GFI, 193 F.3d 92, 97 (2d Cir. 1999) and Folio Impressions Inc. v. Byer California, 937 F.2d 759, 766 (2d Cir. 1991) (“In considering substantial similarity between two items, we review the district court’s findings de novo – not on the clearly erroneous standard – because what is required is only a visual comparison of the works, rather than credibility, which we are in as good a position to decide as was the district court”).  Though the Second Circuit has more recently showed greater deference to a jury’s substantial similarity findings, see Yurman Design, Inc. v. PAJ, Inc., 262 F.3d 101, 111 (2d Cir. 2001), it continues to review bench decisions on substantial similarity de novo.  Seee.g., Boisson v. Banian, Ltd, 273 F.3d 262, 272 (2d Cir. 2001).

[FN3] Arnstein v. Porter, 154 F.2d 464, 468 (2d Cir. 1946).

[FN4] Id. at 468.

[FN5] Id.

[FN6] Id. at 473 (emphasis added).

[FN7] Peter Pan Fabrics, Inc. v. Martin Weiner Corp., 274 F.2d 487, 489 (2d Cir. 1960) (L. Hand, J.).

[FN8] Seee.g., Knitwaves Inc. v. Lollytogs Ltd., 71 F.3d 996, 1002 (2d Cir. 1995) (“[W]here we compare products that contain both protectible and unprotectible elements, our inspection must be ‘more discerning’”).

[FN9] Id.

[FN10] Id.

[FN11] Knitwaves at 1000.

[FN12] Knitwaves at 1004.

[FN13] See Folio Impressions, Inc. v. Byer California, 937 F.2d 759, 765-66 (2d Cir. 1991) (“Of course, the ordinary observer would compare the finished product that the fabric designs were intended to grace…and would be inclined to view the entire [work] – consisting of protectible and unprotectible elements – as one whole. Here, since only some of the design enjoys copyright protection, the observer’s inspection must be more discerning”).

[FN14] 766.

[FN15] Id.

[FN16] Knitwaves at 1004n3.

[FN17] Laureyssens v. Idea Group, Inc., 964 F.2d 131 (2d Cir. 1992).

[FN18] Id. at 141 (emphasis added).

[FN19] Seee.g., Feist Publ’ns, Inc. v. Rural Tel. Serv. Co., 499 U.S. 340, 350 (1991) (“No author may copyright facts or ideas”) (citations omitted).

[FN20] Nichols v. Universal Pictures Corp., 45 F.2d 119, 121 (2d Cir. 1930).

[FN21] See generally Williams v. Crichton, 84 F.3d 581 (2d Cir. 1996).

[FN22] Id. at 583.

[FN23] Id. at 589.

[FN24] Id.; copyright protection does not extend to scene à faire – that is, to scenes or other elements of a work that necessarily result from the choice of a setting or situation. See, e.g., Walker v. Time Life Films, Inc., 784 F.2d 44, 50 (2d Cir. 1986) and Reyher v. Children’s Television Workshop, 533 F.2d 87, 92 (2d Cir. 1976).

[FN25] Id. at 589-90.

[FN26] See Roth Greeting Cards v. United Card Co., 429 F.2d 1106 (9th Cir. 1970).

[FN27] Id. at 1108.

[FN28] Id. at 1109.

[FN29] Id.

[FN30] Id. at 1109.

[FN31] Id. at 1110.

[FN32] Reyhersupra, at 87.

[FN33] Id. at 88.

[FN34] Id. at 92-93.

[FN35] Id. at 91.

[FN36] Id. at 92.

[FN37] See Section 1(b)(ii), supra.

[FN38] Williams at 588.

[FN39] Id. at 589.

[FN40] A Lexis search for “substantial similarity” among district court cases in the Second Circuit between 2008 and 2010 returned ninety-seven results, roughly thirty-five of which turned on the issue of improper appropriation.  A third of this subset does not fit cleanly into any of the seven categories described below.  As such, the data from this period is insufficient to support statistically significant conclusions as to whether certain variations of the test lead to findings of non-infringement more frequently than others.  A further study might attempt to code a larger sample set and plot test variations against outcomes.

[FN41] Blakeman v. Walt Disney Co., 613 F. Supp. 2d 288 (E.D.N.Y. 2009); Currin v. Arista Records, Inc., 2010 U.S. Dist. LEXIS 37575 (D. Conn. Feb. 25, 2010); Oldham v. Universal Music Group, 2010 U.S. Dist LEXIS 126697 (S.D.N.Y. Nov. 29, 2010); Peter F. Gaito Architecture, LLC v. Simone Dev. Corp., 2009 U.S. Dist. LEXIS 53253 (S.D.N.Y. May 22, 2009).

[FN42] Blakeman at 305; Peter F. Gaito Architecture at 14; Currin at 6; Oldhamat 6.

[FN43] Blakeman at 305.

[FN44] Blakeman at 306.

[FN45] R.F.M.A.S., Inc. v. Mimo So, 619 F. Supp. 2d 39 (S.D.N.Y. 2009); Axelrod & Cherveny, Architects, P.C. v. T. & S. Builders Inc., 2008 U.S. Dist. LEXIS 60212 (E.D.N.Y. Aug. 6, 2008).

[FN46] R.F.M.A.S. at 65 (quoting Yurman Design, Inc. v. Golden Treasure Imps., Inc., 275 F. Supp. 2d 506, 517 (S.D.N.Y. 2007)).

[FN47] Id. at 66.

[FN48] Flaherty v. Filardi, 2009 U.S. Dist. LEXIS 22641 (S.D.N.Y. 2009);Lapinesupra; Cabell v. Sony Pictures Entm’t, 714 F. Supp. 2d 452 (S.D.N.Y. 2010); Telebrands Corp. v. Del Laboratories, Inc., 719 F. Supp. 2d 283 (S.D.N.Y. 2010).

[FN49] See Section 2(b)(iii), supra.

[FN50] Cabell at 454.

[FN51] Id. at 461.

[FN52] North Forest Development, LLC v. Walden Ave Realty Assocs., LLC, 2009 U.S. Dist. LEXIS 125331 (W.D.N.Y. July 22, 2009); Bus. Mgmt. Int’l, Inc. v. Labyrinth Bus. Solutions, LLC, 2009 U.S Dist. LEXIS 24900 (S.D.N.Y. March 24, 2009).

[FN53] North Forest Dev., LLC at *12 (quoting Laureyssens).

[FN54] Lewison v. Henry Holt & Co., 659 F. Supp. 2d 547 (S.D.N.Y. 2009); Porto v. Guirgis, 659 F. Supp. 2d 597 (S.D.N.Y. 2009); Sheldon Abend Revocable Trust v. Spielberg, 2010 U.S. Dist. LEXIS 99080 (S.D.N.Y. Sept. 21, 2010); City Merchandise, Inc. v. Broadway Gifts, Inc., 2009 U.S. Dist. LEXIS 5629 (S.D.N.Y. 2009); Lombardi v. Whitehall, 2010 U.S. Dist. LEXIS 19288 (S.D.N.Y. March 3, 2010).

[FN55] Porto at 604.

[FN56] Id. at 597.

[FN57] Id. at 611-12.

[FN58] Id. at 612.

[FN59] Id. at 611 (quoting Williams at 587).

[FN60] Id. at 614.

[FN61] Id.

[FN62] Id. at 597.

[FN63] Id. at 615 (quoting Lapine).

[FN64] Id. at 615-16.

[FN65] Crane v. Poetic Prods. Ltd., 593 F. Supp. 2d 585 (S.D.N.Y. 2009); Want Ad Digest, Inc. v. Display Adver., Inc., 653 F. Supp. 2d 171 (N.D.N.Y. 2009); Ebsin & Alter, LLP v. Zappier, 2010 U.S. Dist. LEXIS 9487 (S.D.N.Y. Feb. 3, 2010).

[FN66] See generally Crane.

[FN67] See generally Esbin & Alter.

[FN68] See generally Want Ad Digest.

[FN69] Id. at 179.

[FN70] Lemley, supra note 1, at 29.

[FN71] Id.

[FN72] Id. at 31.

[FN73] Id. at 30.

[FN74] Seee.g.Feist, 499 U.S. at 345.