Introduction

[1]Copyright is a federal legal regime of exceptional influence on both American art and commerce. Literature, film, music, sculpture, architecture, and software all rely primarily upon federal copyright law for protection against unauthorized copying, which, if left unconstrained, would unquestionably squelch the incentives to create, or to invest in such creation.

[2]Yet the legal standard relied upon to determine copyright infringement is deeply fissured between the two major copyright courts--the Second and Ninth Circuits.⁽²⁾ The remaining regional circuits appear to be unaware of this fissure, as evidenced by the fact that several of them apply both tests interchangeably. Additionally, the infringement test is, by consensus, complicated, time-consuming, and multi-pronged, having vague and redundant nomenclature, and for these reasons alone, resists straightforward explanation. No serious attempt has been made in the scholarly literature to reconcile or to harmonize these two disparate standards, nor to urge one over the other, despite the urgent need to do so.⁽³⁾ This shall be the focus of this article.

[3]More particularly, the purpose of this article is to explain in detail the two legal standards with particular emphasis upon the points of disparity, to discuss the source of this disparity, and then to discuss its significance. Finally, a nominally revised--actually, not so much revised as clarified, or just relabeled--infringement standard is offered. This revised standard is shown in Figure -E-.

[4]In this article, I shall demonstrate that although the Second and Ninth Circuits apply different standards for copyright infringement today, many years ago the two circuits applied the same test. As we shall see, the current Ninth Circuit standard is a direct descendant from a twenty-year old Ninth Circuit decision, which created a de novo standard based on an inadvertent misreading of Second Circuit precedent. Moreover, although the Ninth Circuit test is a sincere imitation of the venerable Second Circuit test, the two are in fact sharply distinct.

[5]I shall also classify the infringement tests applied in the remaining regional circuits according to whether they apply either the Second or the Ninth Circuit standards--as well as demonstrate that although the Second/Ninth Circuit distinction is substantial, it is essentially ignored by the remaining circuits, who continue to apply the two tests more or less interchangeably.

The Second Circuit's Standard for Copyright Infringement

[6]The contemporary legal standard for copyright infringement applied in the Second Circuit is shown in Figure -A-. The acknowledged progenitor of the contemporary standard is Arnstein v. Porter⁽⁴⁾; indeed, the current standard applied in the Second Circuit is virtually identical to the test recited in that 1946 decision.⁽⁵⁾ According to the Supreme Court in Feist Publications v. Rural Telephone,⁽⁶⁾ a prima facie case of copyright infringement is established by proving ownership of a valid copyright and "copying of constituent elements of the work that are original."⁽⁷⁾ Or, in the Second Circuit's own words: "Copyright infringement is established when the owner of a valid copyright demonstrates unauthorized copying."⁽⁸⁾ The Feist standard is depicted in Figure -B-. At this level of abstraction, the standard is not useable; it requires further refinement before it can be applied. Indeed, the inter-circuit disparity that is the focus of this article resides in the definition of the latter requirement, "unauthorized copying."

[7]In the Second Circuit, "unlawful copying," the second prong of the Feist standard (the first being establishing ownership of a valid copyright), is deliberately separated into two requirements: "a plaintiff must first show that his work was actually copied...[and] then must show that the copying amounts to an improper or unlawful appropriation."⁽⁹⁾ This is depicted in Figure -A-. Hence, the "unlawful copying" requirement is traversed when the plaintiff proves that the accused text contains material that comprises the plaintiff's copyright. But the Second Circuit requires that the copyright plaintiff first prove that the accused text is derived from the plaintiff's, and then that the material borrowed qualifies for copyright protection.

[8]The Second Circuit's two prongs of the "unlawful copying" requirement, and their relationship to one another, require careful explanation. These two prongs are not distinct requirements, nor are they two components that together comprise a whole. Rather, the first of these prongs, "actual copying," is actually subsumed within the second, "improper appropriation." Again, "actual copying" means that the accused text is derived from the plaintiff's original work. "Improper appropriation" means that the material borrowed from the plaintiff's work must be copyrightable. Obviously, if the borrowed material is copyrightable, then it is, by definition, original to the plaintiff, and the accused text was by implication derived from the plaintiff's text. Put another way, the "actual copying" prong ("derivation") could be disposed of entirely, and in theory at least, one would arrive at the same result. I shall discuss in a moment the practical consequences of doing so.

[9]Therefore, in the Second Circuit, the plaintiff must show first that the text was derived from the plaintiff's text. Next, the plaintiff must show that at least some of the material that the defendant borrowed qualifies for copyright protection--if so, then copying becomes "unlawful copying" (Second Circuit's term) or "unauthorized copying" (Supreme Court's term). The first of these requirements (derivation) can be shown in one of two ways: either the plaintiff can provide direct evidence that the defendant copied from the plaintiff's text (this is very rare, almost requiring an express admission, or an implied admission, by the defendant) or the plaintiff can offer evidence of access--defined roughly as the "opportunity to view"--plus evidence that is probative of derivation ("probative similarity"). Evidence of access can be generally placed into one of three categories: (1) chain-of-events (the plaintiff gave the text to someone who gave it to defendant's personal secretary⁽¹⁰⁾ (2) widespread dissemination (bestseller status)⁽¹¹⁾; or (3) purely circumstantial (e.g., defendant prepared the accused text in less than half the time that it took the plaintiff, defendant has a history of plagiarizing, etc.).⁽¹²⁾

[10]Probative similarity--which along with access proves derivation--is a simple concept.⁽¹³⁾ It is easy to understand and easy to prove. The evidence--which consists of portions of the accused text compared with the plaintiff's--can be either protectable or unprotectable material. Additionally, as one might infer from this discussion, analytic dissection--performed either by experts or by counsel--is permitted to show probative similarity. By contrast, expert testimony is not permitted in the second prong of the unlawful copying requirement, "improper appropriation." Note also that to show derivation, the plaintiff need not show that the two texts are "substantially similar," in any sort of quantitative sense; indeed, highly probative evidence of derivation may consist of a single phrase, if there is no other plausible source of that phrase, other than the plaintiff's text. The point is this: probative similarity and substantial similarity are different concepts.

[11]If the plaintiff successfully proves that the accused text is derived from his or her text, the plaintiff must then show that the copying amounts to "improper appropriation." Mere copying or derivation might not rise to the level of improper appropriation for three reasons. First, the material copied might not qualify for copyright protection--either because the plaintiff is not the author of the work or because it is ineligible for copyright protection by statute or common law decision, e.g., idea/expression dichotomy, scenes-a-faire, etc. Second, the copying may be quantitatively insignificant--i.e., it is de minimis.⁽¹⁴⁾ Third, the material copied may be quantitatively significant, and therefore protectable expression, yet it may still not give rise to infringement because: (1) the "ordinary observer" overlooks the similarity; or (2) the copying is judged a "fair use" since it is unlikely to harm the market (or any probably unexploited market) for the plaintiff's text.

[12]These three safe harbors are actually subsumed within the ordinary observer standard--the test used to determine improper appropriation. The ordinary observer test is a deliberately un-analytical, almost superficial glance at the two texts side-by-side: "[w]hether the 'ordinary observer, unless he set out to detect the disparities, would be disposed to overlook them, and regard their aesthetic appeal as the same.'"⁽¹⁵⁾ The only restriction imposed upon the ordinary observer is that she limit her comparison to the copyrightable features of the plaintiff's text: "This test for illicit copying to prove infringement demands that the similarities relate to protectable material...."⁽¹⁶⁾

[13]This concludes my discussion of the legal standard for copyright infringement currently applied by courts in the Second Circuit. Again, as evidenced by Figure -A-, the test has four essential requirements: ownership of a valid copyright, access, similarity that is probative of derivation, and substantial similarity as judged by the casual observer. As we shall see, the Ninth Circuit's test also possesses these four requirements, though subsumed within a different structure.

The Ninth Circuit's Legal Standard for Copyright Infringement

[14]The current standard used in the Ninth Circuit, as recited in Apple Computer, Inc. v. Microsoft Corp.,⁽¹⁷⁾ a recent decision in that circuit, is shown in Figure -C-. As evidenced by Figure -C-, in comparison with Figure -A-, the two "primary" prongs, "ownership" and "unlawful copying," are identical in both the Second and Ninth Circuit tests. The disparity relates to requirements subsumed within the second of the two fundamental requirements, the "unlawful copying" requirement.

[15]According to Apple Computer, unlawful copying requires that the plaintiff offer "circumstantial evidence of access and substantial similarity of both the general ideas and expression between the copyrighted work and the allegedly infringing work."⁽¹⁸⁾ "Substantial similarity" in turn is proved by a two-part test: an "objective" test and a "subjective" test.⁽¹⁹⁾ Again, the copyright plaintiff in the Ninth Circuit must first prove access, though unlike the plaintiff in the Second Circuit, he does not have to prove probative similarity. Indeed, the sharp separation between derivation and improper appropriation--the cornerstone of the Arnstein test--is conspicuously absent in the Ninth Circuit's test. Thus, the Ninth Circuit plaintiff proves access, then proceeds directly to the substantial similarity requirement. Under the objective test (the first of the two tests comprising the substantial similarity requirement), the plaintiff's text is dissected into both protectable and unprotectable elements, and compared element-by-element with the accused text.⁽²⁰⁾ According to various Ninth Circuit decisions, the purpose of this objective test is to determine the scope of the plaintiff's copyright:
 

Analytic dissection [i.e., the objective test] is relevant not only to the copying element of a copyright infringement claim, but also to the claim's ownership element. One aspect of the ownership element is the copyrightability of the subject matter and, more particularly, the scope of whatever copyright lies therein.⁽²¹⁾

So, once the scope of the plaintiff's copyright is determined, then the only question left is to determine whether the accused text--or some material part of it--lies within that scope. Apparently that determination is made in the objective test as well, since it involves a detailed, analytical "comparison" of the two texts.⁽²²⁾

[16]Next, the court proceeds under the Ninth Circuit standard to the subjective test. Under this test, the ordinary observer--deliberately eschewing an analytical perspective--assesses the overall similarity of the two texts. In light of the thorough analysis that comprises the intrinsic test, the extrinsic test is difficult to justify. For instance, in the intrinsic test, the features common to both texts are identified then assessed for copyrightability. Thus, by the completion of this test, the fact finder knows whether the accused text contains protectable expression owned by the plaintiff (and since protectable expression is by definition original to the plaintiff, then that implicitly proves that the accused text is derived from the plaintiff's). Provided the amount of protectable expression borrowed traverses some de minimis threshold, what more is needed to establish copyright infringement?

[17]Yet the Ninth Circuit's test--just like the Second Circuit's--permits the accused infringer who borrowed substantial protectable expression to exculpate himself if the casual observer would not recognize the accused text as having been derived from the plaintiff's text. Put another way, merely borrowing some nontrivial, indeed even significant, amount of protectable expression from another's text does not necessarily make one an infringer. If the two texts appear different enough to conceal the similarities from the deliberately unreflective impression of the ordinary observer, then the accused infringer borrows with impunity.

A Comparison of the Second and Ninth Circuits' Standards for Copyright Infringement

[18]If the Second and Ninth Circuit standards are in fact identical, then their disparate vernacular and structure needs to be reconciled. And if they are indeed different, then the differences need to be accounted for--and perhaps one standard urged over the other.

[19]Two legal standards can differ from one another in two essential ways: content/substance or procedure. As we shall see, the Ninth Circuit's test has the same essential requirements as the Second Circuit standard, though the requirements have been complexly reshuffled, almost beyond recognition. The two legal standards discussed in this article have the same content; their difference relates to structure, which in turn confers a procedural distinction with significant consequences.

[20]As evidenced by this discussion (Figure -A- versus -C-), the two tests are at least facially different (e.g., the Second Circuit has no "objective" nor "subjective" tests). The two primary differences between the circuit tests (under which all other lesser differences are subsumed) are as follows: While the Second Circuit extracts from the "unlawful copying" prong a separate derivation requirement, the Ninth Circuit test embeds no such distinction. Derivation and copying copyrightable material are blended into a single inquiry ("substantial similarity"). Second, while the Ninth Circuit divides the substantial similarity inquiry into two components, an objective and a subjective test, the Second Circuit's substantial similarity inquiry only has a single component.

[21]And yet upon closer inspection by displaying and comparing the two standards diagramatically (Figures -A- and -C-), they may not be so different. Most significantly, they have identical requirements, viewed piecemeal, though they appear at different points in the test, and are sometimes performed with different endpoints in mind. More specifically, the Arnstein test (Figure -A-) has four specific requirements (the elements appearing at the ends of the flow chart): ownership, access, an analytical test (probative similarity), and a subjective test (substantial similarity based on an audience impression). Likewise, the Apple test (Figure -C-) also has four requirements: ownership, access, an analytical test (the objective test), and a subjective test. Therefore, the Second and Ninth Circuit's tests have the same four requirements.

[22]The ownership and access requirements are not discernibly different between the two circuits. That leaves the latter two requirements: the analytical test and the audience test. As I shall explain below, these two tests vary between the circuits not only with respect to their content but more significantly with respect to their position in the overall infringement standard.

[23]The Second Circuit's analytical test--"probative similarity"--is directed solely at identifying discrete similarities that are probative of derivation. Here, there is no need that such similarities be "substantial," nor need they relate to protectable material. Probative similarity is used, along with the access requirement, to answer the question, "Is the accused text derived from the plaintiff's original work (Figure -A-)?" By contrast, the Ninth Circuit's analytical test--the "objective test"--is one half of that circuit's substantial similarity determination (Figure -C-). There, the combined objective of these two tests is to identify unlawful copying (which equals derivation and copying protectable material). Thus, "derivation" is not at issue--indeed, it is never explicitly determined in the Ninth Circuit's test but is embedded within the overall unlawful copying determination (i.e., proved by implication during execution of the objective plus subjective tests).

[24]Furthermore, the Ninth Circuit's objective test differs from the Second Circuit's probative similarity requirement in that the former is apparently directed in part at determining copyrightability of each element that appears in both texts. The Second Circuit's probative similarity determination is focused solely upon identifying whether the accused text is derived from the plaintiff's--copyrightability is virtually irrelevant.

[25]The Second and Ninth Circuit tests both contain an audience-impression test. As evidenced by Figure -A-, the Second Circuit's audience test is the sole vanguard that distinguishes mere derivation from copyright infringement. Therefore, this prong must perform two tasks: first, it must separate the protectable and unprotectable elements (though only nominally); next, considering only those protectable elements common to both texts, it must determine whether the two are--according to the casual observer--" substantially similar." By contrast, the Ninth Circuit's audience test (or subjective test) has become almost vestigial--since the objective test has been expanded after Krofft to include an analysis of both protectable as well as unprotectable material. Thus, the Ninth Circuit's audience test does not separate protectable from unprotectable material (that was done in the objective test) like the Second Circuit's audience test does. Moreover, just like the analytical test, perhaps the most significant difference between the audience tests in the two circuits is their position in their respective standards. In each circuit's test, the audience test is performed last. But by the time Ninth Circuit courts perform the audience test, they have already determined which of the common elements qualify for copyright protection. By contrast this analysis is left for the Second Circuit courts to perform during, or within the audience test.

[26]Why bother to separate the "unlawful copying" requirement into two parts (mere derivation and improper appropriation), as the Second Circuit has done? This separation is significant because, as we shall see, the Ninth Circuit's standard embeds no such separation. Since the separation adds nothing--again, both tests have the same four requirements--it can at most confer only a procedural advantage.

[27]Applying any legal test is more or less a process of classifying a particular set of actions. Complex legal tests involve a series of classifications to identify finer and finer distinctions. Aside from the particular classifications chosen--generally a legislative decision, not a judicial one--how one chooses to apply the classification is just as important. Put another way, the taxonomy chosen is really only as reliable as the process used to execute it. Since it is law we are applying, we insist that the legal standard--i.e., the classification process--be fair, which means, in part, that it be precise, or repeatable. Beginning with the same facts, one should insist that the standard produce the identical result each time it is applied. The best way to ensure this result is to systematize the standard, which will eliminate the subjective element as much as possible. This is typically done by requiring that the judge explain/justify/rationalize his or her result by careful step-wise titration, rather than just a single determination, whose methodology is much more difficult to assess ex post. For instance, we do not permit a judge to determine whether a particular element in a literary text qualifies for copyright protection merely by nodding her head; rather, we insist that she systematically apply, step-wise, a series of cumulative partial tests. Thus, she applies, for instance, the scenes-a-faire doctrine, which alone does not tell her whether the particular element is eligible for copyright protection, though if the element flunks the test, then it most certainly is not. Hence also, the procedural advantage to fracturing an abstract legal requirement like "unlawful copying" into discrete manageable elements. In the example just presented, if the particular element fails the scenes-a-faire test, then the accused text is released without the need for further inquiry.

[28]Consider the case of an accused text that is similar to the plaintiff's text (for which the plaintiff owns a valid copyright that subsumes that text within its scope). Despite the similarity, there are only two ways that the accused text may not infringe--i.e., that it may not have "copied unlawfully." First, the infringer may not have derived his text from the plaintiff's--that is, the similarity is due to coincidence, or more likely, to the fact that both authors derived their texts from a common third-party text. Second, even if the accused infringer did prepare his text from the plaintiff's--i.e., he copied from the text--he is only an infringer if what he took was eligible for copyright protection. If the accused infringer borrowed only uncopyrightable elements (e.g., not original or not sufficiently concrete to qualify as protectable expression) he is not an infringer no matter how much material he copied. Derivation alone does not determine infringement. Moreover, if the accused infringer borrowed copyrightable elements from the plaintiff's text, then by definition, he derived his text from the plaintiff's; hence, there is no need, in theory at least, to separately determine derivation. Indeed, dividing the "unlawful copying" inquiry into derivation and improper appropriation--as the Second Circuit has done--is really just a procedural short-cut. Hence, the piecemeal analysis is not just a heuristic device but a time-saving one as well; moreover, truncating the analysis, on average, reduces the error rate.

[29]And so it is with the Second Circuit's separation of the unlawful copying requirement into two prongs: derivation (my term) and unlawful appropriation. The procedural advantage is the simplest way to justify the division: if the accused text is not derived from the plaintiff's text, then he is not an infringer, regardless of the degree of similarity between the two texts, and more importantly, regardless of whether the material in common qualifies for copyright protection. Consider how much simpler this inquiry is than the subsequent inquiry (or an inquiry blending the two, which the Ninth Circuit requires). In the Second Circuit, to prove derivation, the plaintiff proffers a list of a few discrete similarities (either protectable or not, it doesn't matter here) between the two texts, from which, along with proof of access the fact finder can infer derivation. These proffered similarities may be slight indeed--e.g., that the accused city map contains a misspelled street name also found on the plaintiff's map, but not found on the actual street sign, nor anywhere else. But like this example, it can nevertheless be highly probative of derivation, which can make this inquiry unusually facile--facile in the sense that it is unusually objective and quickly performed.⁽²³⁾ Second, and most importantly, the emphasis in the derivation inquiry inevitably converges to the source of the proffered similarity. Thus, if the accused infringer could show that an old map in the public domain also contains the misspelled street name, then the probative value of that evidence vanishes. Identifying that pubic domain map may not be easy--which is certainly not the fact finder's problem--but if it is proffered by the accused infringer, then the fact finder's job is straightforward. Notice also that the derivation inquiry will never drift towards protectability, which is a far more difficult inquiry--whether the particular element proffered by the plaintiff to prove derivation qualifies for copyright is irrelevant. Often the most probative evidence of derivation is unprotectable subject matter, as evidenced by the map example. Indeed, one way to view the Second Circuit's infringement test is that it is structured so that it avoids, as much as possible, the virtually intractable determination of protectability. Thus, before the two texts are handed to the ordinary observer for a final verdict on infringement, we know that at least one element in the plaintiff's text is copyrightable (a presumption triggered by proffering the copyright registration), and that the accused text was derived from the plaintiff's text. Granted, these facts, coupled with the ordinary observer's impression, do not guarantee that the accused infringer copied protectable expression, they just make it very likely. Thus, these three elements combined are a proxy, though a coarse one, for the generally tedious and ad hoc task of determining whether the elements common to the two texts qualify for copyright protection. In my opinion, the copyright infringement test has evolved towards eliminating this task as much as possible.⁽²⁴⁾

[30]Now contrast the derivation prong with the "improper appropriation" prong. (Again, the former is subsumed within the latter--explicit in the Second Circuit's test, implicit/invisible in the Ninth Circuit's test). In the "improper appropriation" prong, the emphasis is not on the source of the proffered similarity, but on whether it qualifies for copyright protection. Granted one way that something can be uncopyrightable is if it is not original (hence the derivation and improper appropriation inquiries overlap), but more often than not, the focus of the protectability inquiry is whether the particular element qualifies as protectable expression according to the "idea/expression dichotomy," which is beyond question a very time-consuming and highly subjective (and therefore error-prone) inquiry.⁽²⁵⁾ And so the Second Circuit's test places this determination temporarily to one side--to the extent that it performs it all--avoiding the test until it is absolutely necessary. By contrast, the protectability determination is the undisputed center of gravity of the Ninth Circuit's test.

[31]It is clear that the Second Circuit and the Ninth Circuit use very different legal standards to determine copyright infringement. What is interesting is how they got that way.

The Origin of The Difference Between the Second and Ninth Circuits' Standards

The Krofft Standard

[32]At one time, the Second Circuit and the Ninth Circuit applied the same legal standard--both were derived from Arnstein. The two standards diverged as a result of the 1977 decision of Sid & Marty Krofft Television v. McDonalds Corp.⁽²⁶⁾--an a priori overhaul of the then-current standard.⁽²⁷⁾ The Krofft test is shown schematically in Figure -D-, so that it can be compared with the Arnstein test (depicted in Figure -B-) and the current Ninth Circuit standard (shown in Figure -C-).

[33]As evidenced by Figure -C-, the Krofft Court divided substantial similarity into two components: an "extrinsic" test and an "intrinsic" test.⁽²⁸⁾ Prior to Krofft, the substantialsimilarity test had consisted of a single element (like the Second Circuit's, Figure -B-). To summarize, the Krofft Court revised the substantial similarity test because of a deficiency it perceived in the then-current test. More precisely, the Krofft Court believed that the then-current substantial similarity test would condemn an accused text that copied only ideas from the plaintiff's text (which are per se unprotectable). Thus a refinement was needed, according to the Krofft Court, to assess separately whether the accused text copied protectable expression or just ideas.

[34]Thus, the purpose of the extrinsic test is to determine whether there is substantial similarity with respect to ideas. The Krofft Court labeled it the extrinsic test because:
 

"[I]t depends not on the responses of the trier of fact, but on specific criteria which can be listed and analyzed. Such criteria include the type of artwork involved, the materials used, the subject matter, and the setting for the subject. Since it is an extrinsic test, analytic dissection and expert testimony are appropriate. Moreover, this question may often be decided as a matter of law."⁽²⁹⁾

[35]By contrast, the purpose of the intrinsic test is to determine whether the two texts are substantially similar with respect to expression. Unlike the extrinsic test, which is analytic, this test is administered by the ordinary observer, who assesses similarity after a casual glance at the two texts.⁽³⁰⁾

[36]This is a peculiar reformulation of the venerable Arnstein standard, to say the least. For one thing, it is not clear why one needs an entire test (the extrinsic test) just to determine similarity with respect to (unprotectable) ideas; by definition, ideas can be freely copied. Therefore, that an accused infringer has copied ideas from the plaintiff's work tells you nothing relevant to infringement.⁽³¹⁾

[37]In fact, the Krofft standard has been harshly and consistently criticized; there is little point in reiterating its shortcomings here.⁽³²⁾ Instead, what is of interest is why the Krofft Court departed from established law, and whether the departure was necessary or even justified. What is also of interest is the path from the Krofft standard to the current Ninth Circuit standard. So why did the Krofft Court depart from the established standard?

[38]According to the Krofft Court's own admission, it did so to restrict the scope of a plaintiff's copyright from reaching to (unprotectable) ideas. Thus, it sought to construct a mechanism to separate ideas from protectable expression:
 

[I]nfringement would be established upon proof of ownership, access, and substantial similarity. Application of this rule, however, would produce some untenable results. For example, a copyright could be obtained over a cheap manufactured plaster statue of a nude. Since ownership of a copyright is established, subsequent manufacturers of statues of nudes would face the grave risk of being found to be infringers if their statues were substantially similar and access were shown. The burden of proof on the plaintiff would be minimal, since most statues of nudes would in all probability be substantially similar to the cheaply manufactured plaster one. ⁽³³⁾

[39]Having identified the problem, the Krofft Court continued "Clearly the scope of copyright protection does not go this far. A limiting principle is needed."⁽³⁴⁾ The Court found this limiting principle in the "idea/expression" dichotomy:
 

The test for infringement therefore has to be given a new dimension. There must be ownership of the copyright and access to the copyrighted work. But there also must be substantial similarity not only of the general ideas but of the expression of those ideas. Thus two steps are implied by the requirement of substantial similarity.⁽³⁵⁾

[40]Thus, the Krofft Court departed from the Arnstein test in two crucial respects: (1) it disregarded the extraction of the copying (or derivation) prong from the more general requirement of unlawful copying (Arnstein subsumed a mechanism to release the accused infringer who can prove he did not copy from the plaintiff's text, even though the two are alike); and (2) it incorporates a two-part substantial similarity inquiry. But was a new legal standard really needed?

[41]The Krofft Court begins by saying: "[I]nfringement would be established upon proof of ownership, access, and substantial similarity." Everything in the Krofft opinion that flows from that statement depends upon its veracity--and unfortunately the statement is materially incorrect. Indeed, this statement, by itself, proves that the Krofft Court misunderstood the Arnstein standard. Under the Arnstein test, one does not establish infringement by showing ownership, access, and substantial similarity; Arnstein said no such thing. Indeed, as evidenced by Figure -A-, one proves copyright infringement under Arnstein by showing that the plaintiff owns a valid copyright, by showing that the accused text is derived from the plaintiff's (which can be proved by access plus probative similarity), and by proving "unlawful appropriation," which in turn is shown by "substantial similarity." The source of the Krofft Court's concern (and confusion) was the term "substantial similarity." This term has never included similarity with respect to protectable portions of the work-in-suit. Courts have always instructed the ordinary observer to consider unprotectable subject matter, as we shall see.

[42]Indeed, the Arnstein Court never used the term "substantial similarity." The Arnstein Court did, however, use the term "improper appropriation," which it referred to as the second prong in the infringement analysis, the first being copying (more precisely referred to as "derivation"). The Arnstein Court decreed a legal standard for copyright infringement in which the plaintiff was first required to prove that the accused text was derived from a text in which the plaintiff had a valid copyright. If it traversed that requirement, then it still had to prove that the copying was "illicit," or "unlawful." Therefore, not all copying was actionable--a point that Arnstein emphasized not once but three separate times:
 

In applying that standard here, it is important to avoid confusing the two separate elements essential to a plaintiff's case in such a suit: (a) that defendant copied from the plaintiff's copyrighted work and (b) that the copying (assuming it to be proved) went so far as to constitute improper appropriation.⁽³⁶⁾ ... If copying is established, then only does there arise the second issue, that of illicit copying (unlawful appropriation).⁽³⁷⁾ ... Assuming that adequate proof is made of copying, that is not enough; for there can be 'permissible copying,' which is not illicit.⁽³⁸⁾  

[44]The Krofft Court was also concerned about the scenario where two texts could be "substantially similar" yet it would be unjust to condemn the copying because the material copied consisted of (unprotectable) ideas rather than (protectable) expression. The Arnstein Court did not expressly mention this scenario; nevertheless, its standard subsumes this concern by distinguishing copying from illicit copying (though one can copy without the copying rising to the level of illicit copying for reasons other than that the material consists of ideas. For instance, the borrowed material could have been in turn borrowed by the plaintiff from a text whose copyright has long since lapsed, or the material could be scenes-a-faire, and so forth.⁽³⁹⁾

[45]Indeed, subsequent Second Circuit decisions that closely follow Arnstein have further clarified the role of the substantial similarity prong, which further evidences that the Krofft Court was fixated on a concern already accounted for in the Second Circuit's test:
 

[S]ubstantial similarity" requires that the copying [be] quantitatively and qualitatively sufficient to support the legal conclusion that infringement (actionable copying) has occurred. The qualitative component concerns the copying of expression, rather than ideas.... ⁽⁴⁰⁾

It is only after copying is established that one claiming infringement then proceeds to demonstrate that the copying was improper or unlawful by showing that the second work bears "substantial similarity" to protected expression in the earlier work.⁽⁴¹⁾

[C]opying...is usually proved by...substantial similarities as to protectable material in the two works.⁽⁴²⁾

[46]Beyond a shadow of a doubt, the Krofft standard is based on a flawed premise: that the extant standard would condemn texts that incorporated from the plaintiff's text only unprotectable ideas. Indeed, the Arnstein test already subsumed a reliable mechanism to ensure that would not happen.

The Role of Second Circuit Law in the Krofft Standard

[47]One must keep in mind that Krofft purports to follow Arnstein: "We believe that the court in Arnstein was alluding to the idea-expression [by distinguishing copying from illicit copying] which we make explicit today."⁽⁴³⁾ And later: "We believe Arnstein is still good law."⁽⁴⁴⁾ This is surprising--that is, one would never suspect Krofft to be the faithful prodigy of Arnstein once one has examined the former's de novo infringement standard.

[48]Therefore, Krofft is based upon a misreading of Arnstein; this is not difficult to prove.⁽⁴⁵⁾ The Krofft Court apparently believed that the extant standard overprotected original material because it subsumed no mechanism to distinguish mere (unprotectable) ideas from (protectable) expression.

[49]The interesting question therefore is precisely what led to the obvious error by the Krofft Court. While Arnstein is the undisputed ancestor of the infringement standard applied in every regional circuit, it nevertheless introduced two slight but insidious flaws into the standard which much later metastasized into chronic confusion over how to apply the standard. And yet one must be careful to state that while Arnstein introduced these flaws, it did not itself contain them; indeed, the flaws relate solely to nomenclature--the particular labels affixed to the prongs comprising the test--yet the Arnstein Court knew what it meant. Other courts have been less sure, as we shall see.

[50]Refer again to Figure -B-. In its most general form, copyright infringement is proved by copying and ownership of a valid copyright. And as evidenced by Figure -A- (the Arnstein test) this second requirement in turn has two prongs, copying and improper appropriation. For this latter requirement, Arnstein used the term "illicit copying"--which has three occurrences of the term "copying," in one legal standard, and they each refer to different things.⁽⁴⁶⁾ This casual usage of the already vague term "copying" has directly caused chronic, material confusion over the proper infringement standard--confusion that has persisted today, and has led to a different standard used in the Ninth Circuit, a standard which originally deviated from Arnstein because of a misunderstanding over Arnstein's multiple uses of the term "copying," as I have shown.

[51]Thus, Arnstein said that, "[i]f there is evidence of access and similarities exist, then the trier of fact must determine whether the similarities are sufficient to prove copying."⁽⁴⁷⁾ In that instance it was referring to what I have relabeled the derivation prong in Figure -E- (the upper most appearance of the term "copying" in Figure -A-. Yet--and this is important--many courts after Arnstein confused the term "copying" as Arnstein used it above, with "unauthorized copying" (the leftmost appearance of the term "copying"in Figure -A-, or as it appears in Figure -B-). In other words, while Arnstein was speaking about mere derivation, other courts believed it was talking about the entire "unauthorized copying" inquiry, of which derivation is only a part.

[52]Most relevant here, the Reyher Court, an influential 1976 Second Circuit decision, said--with the intention of applying the Arnstein standard--that "[c]opying...is...proved by circumstantial access to the copyrighted work and substantial similarities as to the protected material in the two works."⁽⁴⁸⁾ One can compare this excerpt with the one from Arnstein above, and see that they are roughly the same: i.e., copying is shown by access plus similarity--though again, in the passage reproduced above, Arnstein was referring to the derivation prong. Yet the Reyher Court misread Arnstein, and believed that it was referring to the "unlawful copying" prong (in Figure -A-).⁽⁴⁹⁾ In other words, while the Arnstein Court was reciting just a portion of the test (the other part being "unlawful appropriation," in Figure -A-) the Reyher Court thought that the Arnstein Court was reciting the entire test--due solely to the multiple occurrences of the term "copying" in the standard--and therefore it ignored, effectively severing from the standard, Arnstein's "unlawful appropriation" (or "illicit copying") prong.

[53]The Reyher Court restated the Arnstein test with two effects. First, the Reyher Court believed it was reciting the entire test, but it was not, thus completely omitting the "improper appropriation prong" from the "revised" test that it decreed. The significance of this inadvertent corruption will be discussed later. Second, precisely because the Reyher Court believed it was reciting the entire test, it reworked slightly the first portion of the Arnstein test. Thus (to paraphrase the two passages reproduced verbatim above) the Arnstein Court said that copying (derivation) is proved by access plus similarity probative of derivation." Yet the Reyher Court said: copying (unlawful copying) is proved by access plus "substantial similarities as to protectable material...."⁽⁵⁰⁾ Thus the term "substantial similarity" was introduced into the infringement test. The Reyher Court then requires access plus "substantial similarity." But the Arnstein standard for derivation requires only similarity that is probative of copying; this similarity need not be "substantial" (i.e., it need not be pervasive) nor must it relate to protectable material. Subsequent courts, attempting to reconcile the Arnstein and Reyher standards, introduced a dual usage of the term "substantial similarity"--it referred to both similarities probative of derivation, and similarities probative of unlawful appropriation.

[54]Reyher's misreading of Arnstein is significant because Krofft followed Reyher, hence Reyher's obvious misreading of Arnstein was instantly infused into Ninth Circuit law.

From Krofft to the Present Ninth Circuit Standard

[55]As evidenced by a comparison of the Krofft standard (Figure -D-) with the current Ninth Circuit standard (Figure -C-) one can see that it has changed very little. Indeed, its essential structure has not changed at all; aside from re-labeling the two prongs of the substantial similarity test (swapping "extrinsic" and "intrinsic" for "objective" and "subjective") the only observable change has been to the extrinsic test. The post-Krofft modifications shall be discussed below.

[56]From its inception, the extrinsic test--to determine similarity of (unprotectable) ideas--had little justification. And yet since the test permitted analytical dissection, courts gradually began to use it to separate ideas from protectable expression, in effect, to determine the scope of the copyright claim, since once the court reached the intrinsic test, then it could no longer scrutinize individual elements but would be forced to rely upon the subjective impressions of the casual observer, who would inevitably ignore the crucial distinction between protectable and unprotectable expression. Once the ideas were separated from protectable expression, then the court could proceed to the intrinsic test, examining the two texts only in light of the similarities that related protectable expression. Thus, the inevitable metamorphosis of the extrinsic test took place.

[57]Again, the major difference in the Krofft standard and the one used today is not a structural one, but a slight change in nomenclature (from extrinsic/intrinsic to objective/subjective). Though the terms "objective" and "subjective" had appeared in Second Circuit opinions earlier,⁽⁵¹⁾ Shaw v. Lindheim, a 1990 Second Circuit decision, is generally cited as the genesis of the reformulation:
 

(Shaw) Now that it includes virtually every element that may be considered concrete in a literary work, the extrinsic test...can no longer be seen as a test for mere similarity of ideas. Because the criteria incorporated into the extrinsic test encompass all objective manifestations of creativity, the two tests are more sensibly described as objective and subjective analyses of expression, having strayed from Krofft's division between expression [intrinsic test] and ideas [extrinsic test].⁽⁵²⁾

(Apple) As [the Krofft test] has evolved, however, the extrinsic test now objectively considers whether there are substantial similarities in both ideas and expression, whereas the intrinsic test continues to measure expression subjectively.⁽⁵³⁾

[58]Moreover, probably by Brown Bag, and at least by Apple, the extrinsic test has evolved into a full-blown analysis of the boundaries of the plaintiff's property right--i.e., which elements borrowed by the accused infringer are protectable and which are not: "[T]he purpose of analytic dissection is to define scope of copyright protection."⁽⁵⁴⁾

The Remaining Regional Circuits

[59]In the first portion of this article, I sought to show that a genuine difference exists between two standards for copyright infringement. Having shown that it is significant, I shall now argue that the distinction is under appreciated--indeed virtually ignored--among the regional circuit courts of appeal.

[60]According to my research, the regional circuit courts of appeal are roughly split between the two polar standards decreed by the Second and Ninth Circuit respectively. Of the nine remaining regional circuits, four follow the Second Circuit; five have adopted the Ninth Circuit approach. This division roughly follows geography: The First⁽⁵⁵⁾, Fifth⁽⁵⁶⁾, Tenth⁽⁵⁷⁾, and Eleventh⁽⁵⁸⁾ Circuits follow the Second Circuit approach, while the Third⁽⁵⁹⁾, Fourth⁽⁶⁰⁾, Sixth⁽⁶¹⁾, Seventh⁽⁶²⁾, and Eighth⁽⁶³⁾ circuits follow the Ninth Circuit approach. This list is a little less than precise--a few of the circuits have (occasionally) blended the two approaches (in a single opinion), or switched back and forth between the two from one opinion to the next. This suggests that the circuits generally regard any disparity between the circuits as mere difference in nomenclature, rather than as a genuine difference in outcome.

[61]For instance, the term "probative similarity" appears infrequently in Ninth Circuit opinions and in opinions from courts that ostensibly follow the Ninth Circuit approach. Again, the term "probative similarity" is properly used to distinguish evidence that tends to prove derivation versus evidence that tends to prove legally actionable copying (infringement); the latter is referred in Second Circuit vernacular as "substantial similarity." Beyond question--and as discussed in this article--the Ninth Circuit eschews this distinction. A recent Ninth Circuit decision, Fodor v. Time Warner, used the term "probative similarity," though interchangeably with the term "substantial similarity."⁽⁶⁴⁾ A recent trial court decision in the Eighth Circuit (again, a circuit that purports to follow the Ninth Circuit's approach) also refers to "probative similarity," though like Fodor, uses it interchangeably with "substantial similarity."⁽⁶⁵⁾ And therefore, these courts do not extract from the infringement analysis a distinct derivation requirement, as a faithful reading of Arnstein requires.

[62]Consider this remark from a frequently cited Fourth Circuit decision:

 

Some courts use a different set of labels for the two-prong inquiry, referring to the prong as establishment of copying and the second prong as establishment of illicit appropriation. The difference in labels need not concern us because the apparent consensus as to the nature of the tests applicable to each prong of the substantial similarity inquiry smoothes over, as a practical matter, underlying differences in the inquiry's two characterizations.⁽⁶⁶⁾
 

An Interim Summary and a Proposed Relabeled Test for Infringement

[64]To summarize: these are the indisputable facts: (1) the current tests for copyright infringement applied in the Ninth Circuit (Apple Computer) and in the Second Circuit (Castle Rock) are not the same; (2) both tests claim to be faithful progeny of Arnstein; (3) the current Ninth Circuit test originated with Krofft, which claims to follow Arnstein; (4) Krofft is based on a substantial, though inadvertent, misreading of Arnstein; and therefore (5) the current Ninth Circuit test is ultimately based on a misreading of Arnstein.

[65]The entire preceding discussion should reveal that the confusion over the copyright infringement standard is more likely than not due to the multiple usage of the same terms to describe different legal requirements. Indeed, not only are two key terms used multiple times in a single legal standard, but the two terms are themselves vague--"substantial similarity," and "copying."

[66]One source of the confusion has been largely cured--at least in the Second Circuit. The "substantial similarity" requirement that once appeared in the derivation portion of the test has been replaced with "probative similarity" so that it cannot be confused with the "substantial similarity" requirement in the second prong of the test (illicit copying or improper appropriation).⁽⁶⁷⁾ Still, the term "copying" appears too many times in the infringement test. The most abstract formulation of the test is: ownership of a valid copyright and copying. The latter requirement subsumes two requirements: mere copying and illicit copying. That amounts to three occurrences of the term "copying." To avoid confusion, and to better clarify the objectives of the test, the term derivation should be used to refer to the requirement that the accused text was prepared from the plaintiff's text. This requirement--which does not appear separately in the Ninth Circuit test--is not distinct from illicit copying, instead, derivation is subsumed within illicit copying. Thus, a text can be derived from another, yet still not be an "illicit" copy of the first text, because, for instance, the borrowed material was not protectable by copyright. The next prong--whether the derivation amounts to unlawful copying--should be referred to as unlawful appropriation. This is not a new term, for one, it was used in Arnstein, though interchangeably with the term "illicit copying," and it is used in Second Circuit decisions today.⁽⁶⁸⁾ And the first time the term is used--i.e., to refer to the most abstract form of the test--it should be replaced with "actionable copying," which captures the two requirements that it subsumes: derivation and unlawful appropriation. Figure -E- illustrates this suggested nomenclature.

[67]Besides nomenclature, another reason for the persistent confusion over the infringement standard is that the two prongs subsumed under the actionable copying requirement are not actually separate requirements, but rather one is actually a part of the other (i.e., derivation is part of the unlawful appropriation prong). In other words, if the accused text includes copyrightable expression from the plaintiff's text, then derivation is implicitly proved, since if the material was copyrightable, then it is by definition original to the plaintiff (he is the author) and therefore the only source of it is the plaintiff's text. Thus, one inherently proves derivation by proving unlawful appropriation. The significance of this is that one could avoid altogether the derivation prong, and instead proceed to the second prong (unlawful appropriation). Arnstein insists upon extracting from the infringement test the predicate sub-requirement of derivation--no doubt as a way of avoiding both the arbitrary and labor-intensive unlawful appropriation analysis (to assess substantial similarity under the ordinary observer) in those (inevitably frequent) instances in which it is not needed. By contrast, the Ninth Circuit test blends the two inquiries of derivation and unlawful appropriation. At least in theory, the Ninth Circuit's test should converge on the same result as the Second Circuit's test, even without an explicit predicate derivation requirement. This is because under the substantial similarity test, an accused text that does not embody any protectable expression found in the plaintiff's text will be released. Under the Second Circuit's test, it might very well have been determined that the accused text was derived from the plaintiff's (e.g., it was proved by unprotectable material common to the two texts) though it would later be determined during the unlawful appropriation prong that what was taken did not qualify for copyright protection, hence no infringement. The Ninth Circuit would have reached the same result, though without having bothered to first determine that the accused text was derived from the plaintiff's.

[68]But what about the opposite scenario? Suppose the accused text was not derived from the plaintiff's, though they are highly similar. The Second Circuit test would detect that right away, and analysis would conclude, avoiding the unlawful appropriation requirement. The Ninth Circuit would, in theory, reach the same result, but it would take longer (which always increases the potential for error). Without a mechanism to detect derivation, or the absence of it, a Ninth Circuit tribunal must plod through the substantial similarity test--both the objective and subjective prongs. And after doing that, the court would (again, in theory) conclude that the accused text does not infringe, not because it was not derived from the plaintiff's text (the court has no way to discern that), but because it does not embody protectable expression from the plaintiff's text. As I said earlier, it is much easier to argue over the source of the similarities, than it is to argue over whether elements in the plaintiff's text (for which there exists a similar element in the accused text) qualifies for copyright protection. The practical significance of this distinction cannot be overstated.

[69]So perhaps these differences alone are reason enough to favor the Second Circuit test over the Ninth Circuit test--aside from the fact that Ninth Circuit test is derived from a 50-year-old inadvertent misreading of a Second Circuit decision. The Second Circuit standard is far more summary judgment-friendly, since it allows the accused infringer--particularly in those instances involving two highly similar texts, in which the plaintiff's is characterized by low originality--to walk away from the dispute once they can show that they did not derive their text from the plaintiff's. Yet in the Ninth Circuit, which again should produce the same result, the accused infringer must suffer through the substantial similarity analysis (or at least the objective portion of it anyway) since the plaintiff's prima facie case does not include proof of derivation apart from the requirement to prove similarity with respect to protectable elements. Therefore, splitting the actionable copying prong into two separate requirements is not only sensible but procedurally (and heuristically) useful--i.e., it allows the fact finder to stop and assess derivation separate from the far more tedious and complicated task of determining whether what was borrowed qualifies for protection by copyright--and it is also procedurally superior--since it allows the court to truncate the dispute if the plaintiff is unable to prove derivation, and so avoid becoming mired in the next part of the test.
 
 

1. Schlumberger Technology Corporation, Houston, Texas. J.D., University of Chicago; M.S. Massachusetts Institute of Technology. I am grateful for the excellent editorial assistance of Amy Werner.

2. The two courts most significant (past and present) to the development of copyright law are the Second and Ninth Circuits. Indeed, the major dichotomy in the infringement standard occurs between these two circuits. Granted, the legal standards for copyright infringement may differ among the other regional circuits; nevertheless, in copyright law, the operative legal standards are generally borrowed from the Second and Ninth Circuits.

3. The only critical comparison in the scholarly literature that I have been able to locate is: Aaron M. Broaddus, Eliminating the Confusion: A Restatement of the Test for Copyright Infringement, 5 J. Art & Ent. Law 43 (1994). Yet this article does not discuss the positions of circuits other than the Second and Ninth, nor does it discuss what I believe to the most significant difference between the two standards, i.e., the procedural disparity that arises from the Second Circuit's "derivation" requirement, which is not part of the Ninth Circuit test.

4. 154 F.2d 464, 468 (2d Cir. 1946).

5. Though Arnstein is viable authority in the Second Circuit with respect to the legal standard for copyright infringement it recites, it has been overruled with respect to its holding on the standard for summary judgement in copyright infringement suits. Specifically, the Second Circuit in Arnstein reversed the trial court's ruling in the defendant's favor on summary judgment, reciting its infamous "slightest doubt" standard: "The principal question on this appeal is whether the lower court, under Rule 56, properly deprived plaintiff of a trial of his copyright infringement action. The answer depends upon whether 'there is the slightest doubt as to the facts.'" Arnstein, 154 F.2d at 468. In fact, this standard has been repudiated everywhere, including the Second Circuit. See, e.g., Denker v. Uhry, 820 F. Supp. 722, 729 (S.D. N.Y. 1992) ("The Arnstein court's conclusions regarding summary judgment, however, have been undermined by recent case law both on summary judgment in general and summary judgment in particular. The mere existence of disputed facts is no longer sufficient to defeat a motion for summary judgment." (citing, inter alia, Knight v. United States Fire Ins. Co., 804 F.2d 9, 11-12 (2d Cir. 1986)); see also, e.g., Beal v. Lindsay, 468 F.2d 287, 291 (2d Cir. 1972). The Ninth Circuit has relied heavily upon the liberalized summary judgment standards in film-adaptation disputes. See Douglas Y'Barbo, Aesthetic Ambition Versus Commercial Appeal: On Adapting Novels to Film and the Copyright Law, 10 St. Thomas L. Rev. 299 (1998)). Again, one should not for a moment confuse Arnstein's demise in this limited aspect with a more general repudiation of its core holding: the standard for infringement.

6. Feist Publications, Inc. v. Rural Tel. Serv. Co., 499 U.S. 340, 361 (1991).

7. Feist., 499 U.S. at 361.

8. Castle Rock Entertainment, Inc. v. Carol Publishing Group, Inc., 150 F.3d 132, 137 (2d Cir. 1998) (quoting Repp v. Webber, 132 F.3d 882, 889 (2d Cir. 1997)).

9. Castle Rock, 150 F.3d 132, 137 (quoting Laureyssens v. Idea Group, Inc., 964 F.2d 131, 139-40 (2d Cir. 1992)).

10.  See, e.g., Moore v. Columbia Pictures Indust., Inc., 972 F.2d 939, 942 (8^th Cir. 1992); Meta-Film Assocs., Inc. v. MCA, Inc., 586 F. Supp. 1346, 1358 (C.D. Cal. 1984

11.  See, e.g., ABKCO Music, Inc. v. Harrisongs Music, Ltd., 944 F.2d 971 (2d Cir. 1991); Cholvin v. B&F Music Co., 253 F.2d 102 (7^th Cir. 1958).

12. See, e.g., Toksvig v. Bruce, 181 F.2d 664, 666 (7^th Cir. 1950 ); Lauratex Textile Corp. v. Allton Knitting Mills, 517 F. Supp. 900, 903 (S.D. N.Y. 1981).

13. The late Professor Alan Latman was the first to suggest this change in nomenclature in, Alan Latman, "Probative Similarity" As Proof of Copying: Towards Dispelling Some Myths in Copyright Infringement, 90 Colum. L. Rev. 1187 (1990). Prior to Latman's article, the term "substantial similarity" appeared twice in the Second Circuit's infringement test: once in the copying prong, and once in the illicit copying prong (Figure -A-). See, e.g., Reyher v. Childrens' Television Workshop, 533 F.2d 87, 90. ("[c]opying...is usually proved by circumstantial evidence of access to the copyrighted work and substantial similarities as to protectable material in the two works.") But a few years after Latman's article, the (not trivial) relabeling he suggested was firmly reticulated in Second Circuit law. See, e.g., Castle Rock, 150 F.3d 132, 137 ("Actual copying may be established 'either by direct evidence of copying or by indirect evidence, including access to the copyrighted work, similarities that are probative of copying between the works, and expert testimony," (citing, Laureyssens, 964 F.2d at 140)).

14. Castle Rock, 150 F.3d at 138.

15.  Folio Impressions, Inc. v. Byer Cal., 937 F.2d 759, 765 (2^nd Cir. 1991) (quoting Peter Pan Fabrics, Inc. v. Martin Weiner Corp., 274 F.2d 487, 489 (2^nd Cir. 1960) (Hand, J)).

16. Folio Impressions, 937 F.2d at 765.

17. 35 F.3d 1435, 1442 (9^th Cir. 1994), I shall also rely upon Shaw v. Lindheim, 919 F.2d 1353, 1357 (9^th Cir. 1990), since it recites certain portions of the infringement standard with slightly greater precision than Apple Computer.

18.  Apple Computer, 35 F.3d at 1442 (citing Brown Bag Software v. Symantec Corp., 960 F.2d 1465, 1472 (9^th Cir. 1992)).

19. Shaw, 919 F.2d at 1357.

20.  Thus, for instance, in the case of literary texts, the objective test examines a more-or-less predetermined set of components: plot, theme, characters, dialogue. See, e.g., Brown Bag, 960 F.2d at 1477 citing Shaw, 919 F.2d at 1356-57.

21.  Brown Bag, 960 F.2d at 1476.

22.  A critical examination of the Ninth Circuit's test produces this question: what is the purpose of the subjective test? Its ostensible, or stated, purpose is to assess the overall similarity of the two texts. However, it is not clear is why this is necessary once the fact finder has performed the objective test. In the objective test--analogous, though not identical to the Second Circuit's "probative similarity" prong--the text-in-suit and the accused text are placed side-by-side, and the protectability of the elements common to both texts are broken down, element-by-element, as the protectability of each element is assessed. Again, the purpose of this test appears to be two-fold: (1) to determine the scope of the plaintiff's copyright; and (2) to determine whether the accused text contains any protectable expression from the plaintiff's text. (See, e.g., Id. ("Analytic dissection [a.k.a. "the objective test"] is relevant not only in the copying element of a copyright infringement claim, but also to the claim's ownership element. One aspect of the ownership element is the copyrightability of the subject matter and, more particularly, the scope of whatever copyright lies therein....Data East teaches that the source of the similarity must be identified and a determination made as to whether this source is covered by plaintiff's copyright.") (citing Data East USA, Inc. v. Epyx, Inc., 862 F.2d 204, 209 (9^th Cir. 1988)). Once the objective test is applied, it is not clear what is left to do to determine infringement. For instance, suppose that execution of the objective test reveals that the accused text contains several copyrightable elements from the plaintiff's text. Aside from fair use--which does not destroy a prima facie case of infringement, but rather is an affirmative defense to an infringement charge, and, in any event, is almost never invoked in the case of fiction works--is that not the end of the matter? Put another way, the presence of the subjective test--a.k.a., the "audience" or "ordinary observer" test--suggests that an accused text may contain substantial protectable subject matter from the plaintiff's text, yet still not infringe. This must be true, otherwise there would be no need for the subjective test. Indeed, one panel in the Ninth Circuit remarked that the intrinsic test is "virtually devoid of analysis...." Shaw, 919 F.2d at 1357. (The Court went on to say, "for the intrinsic test has become a mere subjective judgment as to whether two literary works are or are not similar," citing Berkic v. Crichton, 761 F.2d 1289, 1294 (9^th Cir. 1985) as a case "reaching a result under the intrinsic test in one paragraph."). Perhaps the subjective test embeds within the infringement standard the concept that a text, even though it copied protectable expression from another text, still does not infringe that text if the average consumer would not recognize meaningful similarity (e.g., derivation). This is a plausible explanation, it fits the data, it is not contradicted by any of it, and is not unnecessarily complicated--yet for this particular datum, a simpler explanation, whose source is the historical development of the current Ninth Circuit infringement standard, exists.

23.  See, e.g., Cooling Systems and Flexibles, Inc. v. Stuart Radiator, Inc., 777 F.2d 485, 492 (9^th Cir. 1985) ("[C]ourts have regarded the existence of common errors as the strongest evidence of piracy, but proof of common errors does not obviate the need for proving substantial similarity." (citation omitted) (citing Nimmer § 13.03 [C] at 13-44)).

24.  See, e.g., Douglas Y'Barbo, The Heart of the Matter: The Property Right Conferred by Copyright, 49 Mercer L. Rev. 643 (1998).

25.  See, e.g., Rehyer v. Children's Television Workshop, 533 F.2d 87, 91 (2d Cir. 1976) ("The difficult task in an infringement action is to distill the nonprotected idea from protected expression.")

26.  562 F.2d 1157 (9^th Cir. 1977).

27.  See, e.g., Dr. Seuss Enterprises, L.P. v. Penguin Books USA, Inc., 109 F.3d 1394, 1398 (9^th Cir. 1997) ("This court's two-part test for substantial similarity finds its roots in Sid & Marty Krofft....").

28.  Krofft, 562 F.2d at 1164.

29.  Id.

30.  Id.

31.  Of course, copying of any material, is probative of derivation, though that's certainly not what the Krofft Court had in mind.

32.  See, e.g., Melville and David Nimmer, 2 Nimmer on Copyright, § 13.03[D],13-77 n.198.6 (1997).

33.  Krofft, 562 F.2d at 1162-63. (citation omitted).

34.  Id. at 1163.

35.  Id. at 1164.

36.  Arnstein v. Porter, 154 F.2d 464, 468.

37.  Id.

38. Id. at 472.

39.  For a more complete discussion of the categories of material that comprise the public domain, see, e.g., Jessica Litman, The Public Domain, 39 Emory L.J. 965 (1990).

40.  Castle Rock, 150 F.3d 132, 137 (italics supplied).

41.  Castle Rock, 150 F.3d 132, 138, quoting, Repp v. Webber, 132 F.3d 882, 889 (2d Cir. 1997); and citing, Laureyssens, 964 F.2d at 140 (italics supplied).

42.  Reyher v. Children's Television Workshop, 533 F.2d 87, 90 (2d Cir. 1976); citing, Arnstein, 154 F.2d at 464, and Bevan v. Columbia Broadcasting System, Inc., 329 F. Supp. 601 (S.D. N.Y. 1971) (italics supplied).

43.  Krofft, 562 F.2d at 1165.

44.   Id.

45.  Actually, Krofft cites precedent from its own circuit, Reyher v. Children's Television Workshop, 533 F.2d 87, 90 (2^nd Cir. 1976).

46.  Actually, Arnstein never states the infringement test in its most abstract form (ownership plus unlawful copying), instead, since it is focused upon the second prong only, it never actually refers to it by name, but only to the two requirements (copying and illicit copying/improper appropriation) subsumed within it. Arnstein used the terms "illicit copying" and "improper appropriation" interchangeably.

47.  Arnstein, 154 F.2d at 468.

48.  Reyher, 533 F.2d at 90.

49.  How can we be sure that the Reyher Court misread Arnstein? First, we can simply compare the two legal standards, noting that in fact the Reyher Court faithfully cites Arnstein at critical points in the opinion. Second, Reyher begins by stating the infringement test in its most fundamental form: infringement requires ownership and "copying." Next Reyher states that "copying" is proved by access and "substantial similarities as to the protectable material in the two works." Id. Therefore, beyond question, Reyher, had in mind the broad meaning of the term "copying" ("unauthorized copying" in Figure -A-) and not the narrower usage of the term copying in which the Arnstein Court used it ("derivation" in Figure -B-). Again, the Arnstein Court said that copying ("derivation") was proved by access plus probative similarity of copying. Once plaintiff established this "factual type of copying," then he must prove that the copying is legally actionable, which is done by proving that some of the purloined material qualifies for copyright.

50.  Id.

51.  At least as early as 1984 (seven years after Krofft) Second Circuit panels began to expand the extrinsic test to include the entire text, and to separate protectable from unprotectable material. Litchfield v. Spielberg, 736 F.2d 1352, 1356 (9^th Cir. 1984). See also, e.g., Data East USA, Inc. v. Epyx, Inc., 862 F.2d 204, 208 (9^th Cir. 1988) ("[the extrinsic test is] an objective test which rests upon specific criteria that can be listed and analyzed.")

52.  Shaw, 919 F.2d at 1357.

53.  Apple, 35 F.3d at 1442; citing, inter alia, Brown Bag, 960 F.3d at 1475.

54.  Apple, 35 F.3d at 1443; citing, inter alia, Brown Bag, 960 F.3d at 1475.

55.  CMM Cable Rep. Inc. v. Ocean Coast Properties, Inc., 97 F.3d 1504, 1513 [41 USPQ 2d 1065] (1^st Cir. 1996) ("To show actionable copying and therefore satisfy Feist's second prong, a plaintiff must first prove that the alleged infringer copied plaintiff's copyrighted work as a factual matter; to do this he or she may either present direct evidence of factual copying, or if that is unavailable, evidence that the alleged infringer had access to the copyrighted work and that the offending and copyrighted works are so similar that the court may infer that there was factual copying (i.e., probative similarity). The plaintiff must then prove that the copying of copyrighted material was so extensive that it rendered the offending and copyrighted works substantially similar.") (citations omitted)).

56.  Engineering Dynamics, Inc. v. Structural Software, Inc., 26 F.3d 1335, 1340 (5^th Cir. 1994) ("Two separate components underlie proof of actionable copying. First is the factual question whether the alleged infringer actually used the copyrighted material to create his own work. Copying as a factual matter typically may be inferred from proof of access to the copyrighted work and "probative similarity." Not all copying, however, is copyright infringement. The second and usually more difficult question is whether the copying is legally actionable. This requires a court to determine whether there is substantial similarity between the two works.") citing, inter alia, Plains Cotton Coop. Assn' v. Goodpasture Computer Serv., Inc., 807 F.2d 1256, 1260 [1 U.S.P.Q. 2d (BNA)1635] (5^th Cir. 1635).

57.  Country Kids 'N City Slicks, Inc. v. Sheen, 77 F.3d 1280, 1284 [38 U.S.P.Q. 2d 1017] (10^th Cir. 1996) ("Plaintiff can indirectly prove copying by establishing that Defendants had access to the copyrighted work and that there are probative similarities.... [Copying does] not end the court's inquiry, as liability for copyright infringement will attach only when protected elements...are copied. To impose liability, the court must find substantial similarity between those aspects of Plaintiff's dolls which are legally protectable and the Defendant's dolls.") citing, inter alia, Gates Rubber Co. v. Bando Chem. Indus., Ltd., 9 F. 3d 823, 832 n. 6 [28 USPQ 2d 1503] (10^th Cir. 1993).

58.  MiTek Holdings, Inc. v. Arce Engineering Co., Inc., 89 F.3d 1548, 1554 [39 U.S.P.Q. 2d (BNA)1609] (11^th Cir. 1996) (Birch, J.) ("Proof of copying may be shown either by direct evidence, or in the absence of direct evidence, it may be inferred from indirect evidence demonstrating that the defendant had access to the copyrighted work and that there are probative similarities between the alleged infringing work and the copyrighted work." (citations omitted)).

59.  Ford Motor Co. v. Summit Motor Products, Inc., 930 F.2d 277, 291 [18 USPQ 2d 1417] (3d Cir. 1991) ("[c]opying [is proved by] showing that the defendant had access to the allegedly infringed work...that the alleged infringing work is substantially similar to the copyrighted work....") citing, Whelan Associates v. Jaslow Dental Laboratories, Inc., 797 F.2d 1222, 1231 [230 USPQ 481 (3d Cir. 1986).

60.  Towler v. Sayles, 76 F.3d 579, 582 [37 U.S.P.Q. 2d (BNA)1785] (4^th Cir. 1996) ("Towler [the plaintiff] can raise a presumption of copying by showing both that Sayles had access to "Crossed Wires" [the work in suit] and that the two screenplays in question are substantially similar." citing, Dawson v. Hinshaw Music Inc., 905 F.2d 731, 732 (4^th Cir. 1990). Additionally: "[T]here are two prongs to the substantial similarity inquiry. [A]n 'extrinsic' or 'objective' inquiry [and] the 'intrinsic' or 'subjective' test.") citing, Litchfield v. Spielberg, 736 F.2d 1352, 1355 (9^th Cir. 1984).

61.  Marigold Foods, Inc. v. Purity Dairies, Inc., 966 F.2d 1453, 1457 (6^th Cir. 1992) ("When direct evidence of access is unavailable, evidence of the alleged infringer's access to the underlying work and a substantial similarity between the alleged copy...'[e]stablished a prima facie case of copying.'") citing, Blumeraft of Pittsburg v. Newman Bros, Inc., 373 F.2d 905, 906 (6^th Cir. 1967). See also, Baxter v. MCA, Inc., 812 F.2d 421, 424 (9^th Cir. 1987) ("Applying the two-part extrinsic/intrinsic test.... The extrinsic analysis requires a 'determination of substantial similarity between the ideas...[T]he intrinsic test involves 'a determination of substantial similarity between the expressions of the idea.") citing, Spectravest, Inc. v. Mervyns, Inc., 673 F. Supp. 1468, 1491 (N.D. Cal. 1987), in turn citing Krofft, 562 F.2d at 1164.

62.  Wildlife Exp. Corp. v. Carol Wright Sales, Inc., 18 F.3d 502, 508 [30 U.S.P.Q. 2d (BNA)1140] (7^th Cir. 1994) ("[C]opyright infringement may be inferred when it is shown that 'the defendants had access to the copyrighted work and that the accused work is substantially similar to the copyrighted work.' [S]ubstantial similarity is made by the 'ordinary observer test...." (citations omitted)).

63.  Hartman v. Hallmark Cards, Inc., 833 F.2d 117, 120 [17 USPQ 2d 1889] (Fed. Cir. 1991) ("To establish her claim for copyright infringement in the absence of direct evidence of copying, Hartman had to prove: (1) her ownership of the copyright...; access...; and (3) substantial similarity...." citing, inter alia, McCulloch v. Albert E. Prince, Inc., 823 F.2d 316, 318 (9^th Cir. 1987)). The Hartman Court goes on to say: "Determination of substantial similarity involves a two-step analysis. [citations omitted] There must be substantial similarity 'not only of general ideas but of the expressions of those ideas as well. [F]irst similarity of ideas is analyzed extrinsically, focusing on objective similarities...[s]imilarity of expression is evaluated using an intrinsic test....") Hartman, F.2d at 120, citing, Krofft, 562 F.2d at 1164.

64.  Fodor v. Time Warner, 1994 U.S. App. LEXIS 4250, *4 (9^th Cir. 1994). This is the only Ninth Circuit decision that refers to "probative similarity." In Fodor, the Court cites Nimmer's treatise for the proposition that proof of copying (which along with proof of ownership equals infringement) requires access plus probative similarity. Yet the Court states in the very next sentence that: "Even if two works are substantially similar, however, there is no infringement liability if the challenged work was independently created."

65.  Control Data Systems, Inc. v. Infoware, Inc., 903 F. Supp. 1316, 1320 (D. Minn. 1995). Consider this passage from this opinion: "The Court next considers whether there are "probative similarities" between [the two works]. Control Data asserts that the AlphaCyber source code is substantially similar to the NOS source code." Id. at 1321. Just a few sentences later the Court goes on to say: "[T]he Court finds that the similarities in these areas are indeed substantial and constitute probative evidence of copying." Id. These passages indicate that the Court has confused the two terms--beyond question, it believes the two terms are synonymous.

66.   Dawson, 905 F.2d at 733, n. 1. (citations omitted).

67.  Latman, "Probative Similarity" As Proof of Copying, 90 Colum. L. Rev. at 1190.

68.  Arnstein, 154 F.2d at 468; Laureyssens, 964 F.2d at 139.

Date of BLT Publication: May 11, 1999
© Douglas Y'Barbo 1999