CHAPTER 7

COPYING: THE SUBSTANTIALITY OF SIMILARITIES

Forensic analysis includes multiple reductions and comparisons designed to accomplish two purposes: (1) to ensure that the analyst has made all relevant reductions and is firmly committed to those reductions,[1] and (2) to ensure that the analyst makes exhaustive comparisons of surface to surface, surface to middleground and background, and function to function.[2] He needs to achieve a three-dimensional view of the works at issue in order to place similarities in perspective.

The crucial question remains unanswered when all the steps of forensic analysis are complete: Did the defendant copy the plaintiff's work? This is the ultimate question historically reserved as the sole province of the trier of fact but now open to the opinions of expert witnesses.[3] A music analyst may base his opinions on his expertise regarding similarities, derivations, compositional norms, stylistic influences, and the creative process itself. This expertise enables him to characterize similarities and judge their significance. But similarities alone do not constitute infringement; they provide only circumstantial evidence of copying.[4] Like the trier of fact, the analyst can only draw inferences from this circumstantial evidence. The question for the expert then is what analytical criteria to apply to the evidence in order to draw conclusions about the existence of copying.

Legal proof of infringement requires a showing of access and substantial similarities.[5] Access may be proved through direct testimony, in which case the expert plays no role, or through a showing of "striking similarities," which generally requires expert testimony.[6] Plaintiff must show "substantial similarities" in all cases. Court opinions have described "striking similarity," but the term lacks a coherent definition. "Substantial similarity" remains largely a subjective question to be examined on a case-by-case basis.

The expert's testimony on musical similarities, substantial or striking, has become an essential aspect of infringement litigation.[7] But the guidelines provided to the expert by the court on these questions are very general and vague. The generality of the legal guidelines, unfortunately, does not render them generally applicable.

Because the tests for striking and substantial similarities have not been defined at law, musical criteria may fill the void. The law understandably hesitates to draw up rigid standards to be applied to the highly abstract endeavor of artistic creation. Even musical criteria probably cannot produce absolute standards for assessing musical similarities, because musical relationships are too numerous and complex to be codified. But music does follow recognized principles that may suggest certain guidelines appropriate for general application. The musician must remember that the distinction between striking and substantial similarities is a legal one: substantial similarities plus access prove copying; striking similarities allow an inference of access.

This chapter explores whether musical similarities will submit to general guidelines, what the nature of those guidelines might be, and whether such guidelines will prove useful and enlightening. This chapter proposes tests, cognizable to the court as well as the expert, to assist the trier of fact in reaching intelligent conclusions.

Striking Similarity

The law provides unclear and potentially conflicting guidelines on striking similarity. The Seventh Circuit's opinion in Selle alone contains the following various formulations of the same guideline (distinguished by the emphasis added). The law allows an inference of access where "such similarities are of a kind that can only be explained by copying, rather than by coincidence, independent creation, or prior common source."[8] But the criteria is more stringent when the similarities must be "so striking and of such nature as to preclude the possibility of coincidence, accident or independent creation."[9] Somewhat less rigidly, similarities must be "the sort that cannot satisfactorily be accounted for by a theory of coincidence, independent creation, prior common source, or any theory other than that of copying."[10] In addition, Selle notes: "The similarities must appear in a sufficiently unique or complex context as to make it unlikely that both pieces were copied from a prior common source,"[11] or "that the defendant was able to compose the accused work as a matter of independent creation."[12] Selle finally notes the very loose standard that requires only a "highly unique pattern" which "makes copying more likely."[13]

The Selle opinion never reconciles the difference between a standard that precludes the possibility of coincidence and one that merely finds copying a more likely explanation. Rather, the Selle court contributes without comment two additional formulations that fall somewhere between the extremes: "similarity which reasonably precludes the possibility of any explanation other than that of copying,"[14] and similarities "of a type which would eliminate any explanation of coincidence, independent creation or common source. . . ."[15]

Absolute preclusion of the possibility of coincidence, the most rigid standard quoted above, is an unrealistic goal. If that were truly the standard, then the law could eliminate the doctrine of striking similarity altogether. A plaintiff can never refute coincidence entirely; coincidence can only become an implausible explanation. On the other hand, if the trier of fact needs to find only that copying is more likely than coincidence, then the standard is merely one of preponderance. Such a weak standard would reduce the doctrine of striking similarity to no standard.

The Selle court also fails to explain the difference between coincidence and independent creation. Plaintiff sought to prove striking similarity in Baxter v. MCA, Inc.[16] The defense argued at trial that plaintiff's proof should include a showing that the defendant was not "competent" to compose the accused work.[17] For authority, the defense pointed to the statement in Selle quoted above, that similarities must make it unlikely that defendant was "able" to compose the work as a matter of independent creation.[18] The proper interpretation of this statement is far from clear.[19] In all cases, the defendant copied the plaintiff's work, copied some other work, or did not copy. If the defendant created the work independently, then similarities must be coincidental.[20]

The illogic and inconsistency surrounding the doctrine of striking similarity suggest that the court has resorted more to slogans than to reason. Certainly the expert can find little guidance in these various formulations. Plaintiff's expert can only follow the path laid out by the courts: refute the alternative explanations of coincidence, independent creation, and common prior source, and buttress this testimony with reference to the work's uniqueness and complexity. But the plaintiff's expert bears the burden here of proving a negative, and he must do so without a clear standard. An astute cross-examination seems almost certain to raise an appealable issue.

Assuming that the doctrine of striking similarity survives in spite of its flaws, the expert must deal with it. It must mean something less than the preclusion of the possibility of coincidence and something more than mere preponderance in favor of copying. The relevant question seems to be: At what point do the possibilities of coincidence and common source lose credibility?

Musical Guidelines for Striking Quantity

Some courts and legal commentators view striking similarity as something greater than substantial similarity, suggesting a difference of degree instead of kind.[21] Under this approach, proof of striking similarity would eliminate the need to make a separate showing of substantial similarity. The theory that striking similarity implies and subsumes substantial similarity suggests a quantitative approach: striking similarity results from many substantial similarities. The seemingly qualitative interpretation, that striking similarity results from very substantial similarity, does not work well. Determining which similarities are substantial is difficult enough without dividing substantiality into subcategories.

Refuting the Possibility of Coincidence

An extensive quantity of similarities might render coincidence unlikely. But because coincidence can never be refuted entirely, any quantitative test should be a stringent one. The dangers of quantifying similarities have been explained above. Quantity must at the very least be supported by functional similarities. If quantity makes the similarities striking, then those similarities must also qualify as substantial. Quantity of similarities, then, should tend to disprove coincidence only where there is virtual identity of substantially similar features occurring without interruption over a sustained period of time.

Further, the court should not find striking similarity where the accused work's internal structure suggests a credible explanation for the similarities. In other words, if the similarities occur in a passage that develops the accused work's internal and dissimilar thematic materials, then independent creation remains likely. A striking feature that adheres to the organic unity of the defendant's work does not, without more, suggest coping. The defendant cannot be presumed to have ignored his own thematic materials.

Disproving Prior Common Source

The extent or quantity of similarities tends to disprove only coincidence. Even the most extreme quantity of similarities can never disprove prior common source.[22] To prove striking similarity, the expert is asked essentially to testify that similarities between the accused and complaining works outstrip those between the accused work and all other works. But an assessment based on quantity cannot look beyond specific works actually compared. Because the expert cannot compare all works, he can testify only that the complaining work's unique qualities make a prior common source unlikely. Thus, on the issue of common source, the expert must look to quality rather than quantity.

The difficulty of disproving prior common source also raises questions concerning burden of proof. The Selle court established that the burden of showing striking similarities rests with the plaintiff:

At oral argument, plaintiff's attorney stated that the burden of proving common source should be on the defendant; however, the burden of proving "striking similarity," which, by definition, includes taking steps to minimize the possibility of common source, is on the plaintiff.[23]

Plaintiff in Selle attempted to prove striking similarity with reference solely to the two works at issue. He presented an argument based essential on quantity. But plaintiff's expert had not inspected any prior works, even those of the defendants, and the court justifiably found his testimony insufficient to disprove common source.

A search for musical precedents within the defendant's own compositions seems an obvious and essential step in analyzing possible derivations. But the Selle court indicated that the plaintiff must take steps to minimize the possibility of common source in compositions created by the defendants or by others.[24] How extensively must plaintiff's expert search the vast repertoire that includes works "by others?" An effort made in utmost good faith can only scratch the surface. Even an intelligent and accurately focused search accomplishes little more. Therefore, the basis for refuting common source cannot lie in a search of musical precedents, because a search can never cover even a representative sample of the literature.[25] An expert must use his general knowledge of music, rather than his knowledge of specific works, to refute common source.

As a practical matter, then, the burden of proving striking similarity does not rest entirely with the plaintiff. Plaintiff has no incentive to find real similarities; his tendency will be to seek out straw men. A search benefits the defendant alone: similarities found prove something; similarities not found are irrelevant. The court asks plaintiff's expert only to assert a credible basis for his belief that no common source exists. Defendant bears a much heavier burden: if some prior work is similar, the defendant must produce it. Defendant's expert must back his assertion that the similarities derive from the literature with examples.[26]

Musical Guidelines for Striking Quality

Selle states that something more than quantity must support a finding of striking similarities.[27] The Selle court criticized plaintiff for the absence of any testimony on the relative complexity or uniqueness of the similarities in the two works at issue.[28] Thus, the court introduces two qualitative measures that might support a showing of striking similarity.

Taken at face value, the statement applies complexity and uniqueness only to the works as a whole. The terms thus seem to raise the question of competence once again. Mozart's operas are unique and complex.[29] The Bee Gees, for example, who never learned to read music, would be unlikely to create independently a work of such innovation and complexity.

However, uniqueness and complexity deserve some scrutiny as applied to specific features within the work. Complexity in this context suggests a multitude of relationships, something not merely episodic or isolated from other features. Thus, complexity might be found in a feature's functional aspects. Forensic analysis will uncover these functions, some more complex than others. But this adds no weight, because any similarity worth noting will include similarities of function. Classifying some as striking and others as merely substantial either returns the inquiry to quantification or creates subcategories of substantiality. Complexity, therefore, provides minimal assistance to the analyst trying to draw inferences from similarities.

Disproving Prior Common Source

Uniqueness seems merely to state the requirement of no prior common source in different terms. The court's invocation of this measure does not inform the expert regarding appropriate evidence of uniqueness. Potentially, he might show uniqueness by conducting a comprehensive but nevertheless fruitless search of musical precedent; but this approach, requiring a comparison of specific work to specific work, has been shown to be too ambitious. Rather, the expert must resort to showing that the complaining work is unconventional‑-a more general indicia of originality than uniqueness and a term that seems preferable to "relative uniqueness." Even under this less rigid standard where the analyst compares a specific work to musical conventions, proof remains difficult. The expert still faces the question of which musical conventions provide relevant points of comparison. The expert might compare the potentially unconventional features to the external factors of notational rules, theoretical principles, or stylistic vocabulary. On the other hand, he might compare the potentially unconventional features to the internal principles of the individual work.

Here the analyst begins to encounter the notion of common errors. Some examples are easy: notational errors such as an incorrect key signature or the wrong number of beats in a measure. Yet these do not properly constitute musical errors because they are not sounded.

True musical error is not so easy to define. Shafter's treatise suggested that voice-leading errors indicate copying. Parallel octaves and fifths, Shafter asserted, are banned by the rules of harmony.[30] This is not so. Rules of harmony (more accurately, rules of counterpoint) no longer ban anything in contemporary music.[31] Only theory students learning contrapuntal practices suffer such dire constraints. Composers have managed to break every rule of the common practice period; the literature of the past hundred years abounds with such examples. Thus, theoretical rules might guide the discovery of unusual features, but they cannot form a test for real error.

Musical errors in the more metaphorical sense suggest things that are out of place. One reference point seems to that of stylistic vocabulary. Complex harmonic movement does not belong in country music nor yodelling in opera. But in either case, independent creation cannot be ruled out entirely. Something more is needed to prove copying; the oddity must somehow be compounded. Suspicions are warranted only regarding the second opera to employ yodelling. "Common error" exists only where the strikingly similar feature is foreign to both works.

The internal makeup of a work provides another reference point. Some element that resists all norms of theme, imitation, variation, and contrast might earn special scrutiny. But music combines elements of unity and disunity all the time; any work that lacks elements of disunity will sound trite. By what standard can contrast be measured, and when does an essential aspect of music, disunity, suggest error?

Louis Nizer provided what he considered an example of common error in music.[32] He described the word painting in plaintiff's "L'Année Passée," a chromatic chord denoting pathos where the text speaks of a girl turned "streetwalker."[33] The same chord was out of place in defendant's work where it accompanied "They make you feel so very glad."[34] Several factors are at work here. First, the chromatic chord was unconventional in the plaintiff's work; the calypso style makes little use of such chromaticism.[35] Second, the chord appears only once in plaintiff's song. It can thus be attributed to the specific purpose Nizer mentions; it is not readily explained (although it might be explainable) in purely musical terms. Third, defendant's work employed the chord at a corresponding point.[36] The chord bore all of the peculiarities within defendant's work that it possessed in plaintiff's.[37] However, in defendant's work the explanation of the text was gone.

The expert faces considerable difficulty in demonstrating this kind of error because the criteria vary. Although Nizer makes a partially credible showing in this regard, he did not have to show striking similarities. Nizer's common error, if it is such, went only to the question of substantial similarity.

The theorist Donald Tovey offered Tchaikovsky's Symphony No. 5 (Figure 48) as an example of stylistic peculiarity:

[G]reat harmonic distinction is given to this theme by its first note. Those who misremember it as B will learn a useful lesson in style when they come to notice that this note is C and not B.[38]

The C is striking in the musical sense. Convention suggests an opening B so strongly that many casual listeners probably remember it that way. An expert could not testify that the feature is unique without knowing every precedent. But because Tchaikovsky contradicted such a strong convention, an expert could assert that the literature will probably yield no prior common source.

Example 46: Tchaikovsky Symphony No. 5.

Refuting the Possibility of Coincidence

The duplication of an unconventional feature still will not prove striking similarity. Tchaikovsky created the unconventional theme of Figure 43 independently, and another composer might coincidentally do the same. Similarly, in Nizer's example above, the one chord seems an insufficient basis from which to infer access. There seems to be nothing attributable to quality alone that cannot be explained by coincidence.

Striking similarity thus evades any single test. The quantity or extent of similarities will not refute prior common source, and qualitative measures such as uniqueness do not refute coincidence. Some combination of the two seems essential. The guideline for a quantity tending to disprove coincidence was stated above: Quantity of similarities should tend to disprove coincidence only where there is virtual identity of substantially similar features occurring without interruption over a sustained period of time. The qualitative considerations mentioned above can be summarized: (1) the complaining work should employ an unconventional or stylistically foreign feature; (2) the defendant's work should duplicate that feature in a way that retains the same unconventional or stylistically foreign nature; and (3) the feature as it appears in defendant's work should not be readily explained by the internal characteristics of defendant's work.

These two sets of guidelines are imperfect. Music presents so many parameters and functional relationships that it is difficult to say what kind of feature might be striking in the qualitative sense. In addition, infringement cases infrequently assert striking similarity, providing little opportunity for the court to fine-tune the guidelines. Courts have too little experience to evaluate effectively this or any other set of guidelines.

Because copying requires access, the burden of showing striking similarity without a showing of access should be a heavy one. Some evidentiary standard close to that of "beyond a reasonable doubt" should be preferred over a mere "preponderance of the evidence" standard.

Finally, although the commentator Jeffrey Sherman presents a good case for considering striking similarity to be separate and apart from substantial similarity, this analysis suggests that he is only partly correct. Similarities may be musically striking without being substantially similar, but access should not inferred in such cases. The qualitative side of the guideline put forward here implies a substantive requirement. The quantity of similarities required to refute the reasonable possibility of coincidence requires an unbroken chain of otherwise substantial similarities. As a practical matter, this burden may be impossible to meet in any case short of literal duplication.

Substantial Similarity

If one were to examine two chess games in progress and discover that all of the pieces in one game were in the same position as the other, he might infer that this striking similarity could result only from the literal duplication of every move. Because of the low probability that this literal duplication would occur by coincidence, he could justifiably infer that one pair of players had copied the moves of the other. However, if the pieces were in the same position with one or two exceptions, the observer might conclude only that the two games were substantially, even remarkably, similar. The players must have followed both a similar strategy and similar tactics to arrive at this correspondence. It would be logical to inquire whether one pair of players knew what the other was doing. But if all pieces were in the same positions except for the black king, the observer could not draw these conclusions. The strategy and tactics of defending and capturing the black king would have been entirely different, and the similarities would suggest only a profound coincidence.

The question of substantial similarities in music should involve essentially this same procedure. Too often, however, the process focuses on finding similarities of any kind without regard to their significance or relatedness. If two composers have different purposes, wish to express different ideas, and happen to manifest some similarities in a different context, the probability that one copied the other is low. Analysts cannot merely focus on the position of pawns without asking why they were arranged in that manner. Similarities are significant only to the extent various elements support each other and reveal the same overall strategy.

The law treats substantial similarity on a case-by-case basis. Court opinions offer only the proviso that substantial similarity is to be determined by the reactions of a lay audience.[39] The expert can offer his opinion on how a lay person hears the works based on his expertise in musical perception. But reactions of lay listeners have not been studied sufficiently to provide objective data, and few music theorists study public reactions to music in a systematic way.

Before suggesting guidelines for determining the substantiality of similarities, it seems appropriate to examine the various contexts in which similarities appear. Some factors surrounding music composition point toward copying, intentional or subconscious, but not necessarily copying that is legally actionable. Other factors suggest merely coincidence, an over-developed proprietary interest in commonplace expressions, or, less charitably, the intentional assertion of inconsequential similarities for personal gain.

Motivations of the Plaintiff

Some attorneys divide plaintiff in infringement actions into two categories. The first category involves working, published writers who tend to file credible claims. The second category includes the numerous writers with little output and no publications. Typically, a plaintiff in this second category has sent his song unsolicited to a publisher or entered it in a songwriting contest. When something remotely similar shows up on the popular music charts, he presumes that it must be derived from his work.[40]

This latter scenario seems to play itself out often. Many such cases are settled well before trial. Defending one's work as an independent creation poses real difficulties, because most composers generate no dispositive evidence of their creative processes. Although the Bee Gees introduced a tape of their composition session that they claimed proved independent creation, that evidence did not warrant the serious consideration it received.[41] Tapes do not record thought processes and cannot, therefore, record the compositional process. The Bee Gees could have as easily recorded themselves composing "Happy Birthday."

Once a plaintiff demonstrates access, however, all similarities become suspect. Virtually all plaintiffs assert some real similarities, even though those similarities sometimes have no musical significance. Therefore, the defendant is at risk concerning any material he has seen or heard. For this reason, many composers and publishers, the Bee Gees included, establish a policy of never opening unsolicited materials.[42] The defendant is more likely to prevail through plaintiff's failure to show the concrete element of access than on the abstract and nebulous issue of substantial similarities.

A successful work of music is a highly valuable asset. This fact, coupled with the difficulty of defending independent creation, makes a musical work an attractive target for litigation.[43] The time, expense, and risk involved make defending an infringement suit an unattractive alternative. A well-known defendant may be subject to unfavorable media scrutiny and may find his career in jeopardy if in the end a jury brands him a plagiarist.

Naturally, a writer senses a strong proprietary interest in his creative work. That interest tends to be heightened when an unknown composer believes his labors are supporting the flaunted success of a famous personality. But too often litigation focuses on that success and the prejudices it engenders. The jury may be asked to champion the underdog or, conversely, to believe that dissimilar success proves the works dissimilar. Such approaches do little to advance a just resolution.

The Impetus to Copy

Some factors motivate a composer to write a work of music that is similar to another either intentionally or subconsciously. Composers employ many established practices and tend to build on what has been done before. Composers sometimes quote another work, although there are historical examples of wholesale borrowing with no hint of attribution. Composers almost inevitably conform to various stylistic forces, which, by definition, result in similarities and shrink the number of elements that might convey originality. Finally, composers almost always include a certain amount of non-original materials, because listeners demand a familiar frame of reference.

Historical Views of Borrowing

Modern conceptions of plagiarism are diametrically opposed to the historical traditions of western music, which posit that a new work should be based on one already in existence.[44] The earliest efforts in polyphony consisted of adding a new voice to an existing plainsong melody.[45] The existing melody was employed as a cantus firmus (fixed song) or cantus prius factus (song previously made).[46] By the 15th century, composers were writing Masses based on a cantus firmus taken either from plainsong melody or from some secular source. The Mass would bear the name of that secular source, for example Dufay's Mass Se la face ay pale (If my face is pale). Nearly every composer of the Renaissance wrote a Mass on the song L'Homme armé (The armed man).[47]

In his Missa L'Homme armé, Obrecht follows the lead of still another forerunner‑-Busnois. The general plan of the Mass is, indeed, so close to that of Busnois's that the work has been described as a "parody" of the earlier composition . . . and also "a tribute, on Obrecht's part, to the `authority' of his model. . . ." Despite all [similarities], however, Obrecht's Mass has its own striking individuality.[48]

The term "Missa parodia" is the only one, now employed for the various Renaissance Mass types, that was actually used in the 16th century. It was included by the German, Jakob Paix, in the title of his Mass based on Crequillon's Domine, da nobis, 1587.[49]

The practice of basing Masses on these secular tunes was widespread:

We have introduced an artificial and theatrical music into the church, a bawling and agitation of various voices, such as I believe had never been heard in the theatres of the Greeks and Romans. Horns, trumpets, pipes vie and sound along constantly with the voices. Amorous and lascivious melodies are heard such as elsewhere accompany only the dances of courtesans and clowns. The people run into the churches as if they were theatres, for the sake of the sensuous charm of the ear. . . .[50]

The church eventually took action, instituting sweeping reforms at the Council of Trent (1545-1563):

In the case of those Masses which are celebrated with singing and with organ, let nothing profane be intermingled, but only hymns and divine praises. . . . They shall also banish from church all music that contains, whether in the singing or in the organ playing, things that are lascivious or impure.[51]

The issue was not whether to borrow an existing work, only the appropriate choice of material.

Composers have interwoven popular tunes into more serious works throughout much of the history of western music. The term "quodlibet" (what you please) was applied to works of the 15th to 18th centuries in which disparate tunes were made to fit together. J.S. Bach's Goldberg Variations include a quodlibet that combines two contemporary popular songs, "Long Have I Been Away from Thee" and "Cabbage and Turnips."[52]

Composers borrow wholesale when they write variations on a theme of another composer. For example, Brahms' Variations on a Theme of Handel, a work used to illustrate some analytical principles in Chapter 5, begins with Handel's theme in its original form‑-an act of literal duplication. The twenty-five variations and fugue that follow are all derivatives of Handel's theme. The music world does not excuse the taking because Handel's theme had entered the public domain or because Brahms cited the theme's origin; Brahms complemented Handel and transplanted the theme into new harmonic and formal contexts.[53] Thus, Brahms made a significant original contribution to the repertoire. He took Handel's middleground and elaborated it into his own foreground. The original foreground sufficiently distinguishes Brahms' work from Handel's.

A composer who takes a short portion of another's work and employs it in a way that does not seem to pass it off as his own should not be considered a plagiarist. The music often sets off the borrowed material, as though putting it in quotes. For example, the Shostakovich Symphony No. 15 quotes Rossini's "William Tell Overture," and Bartok's Concerto for Orchestra (1944) quotes the Shostakovich Symphony No. 7 (1942). Both composers admitted the taking. It would seem that some aspect of the fair use provision would permit this kind of borrowing, some examples of which greatly enrich the repertoire, provide testimonials to the original composer, and disserve no one.[54]

Only as musical works lost utilitarian value and began to be perceived primarily as the personal expression of an individual did plagiarism earn serious condemnation. The term "plagiarism" seems to have entered the English language around 1600.[55] Discussions of plagiarism concerning music, however, do not occur in the literature until the 18th century. In 1774, Petit de Bachaumont noted unattributed borrowings from Christoph Gluck:

The music of Orphée et Euridice [of Gluck] was printed in Paris eight or ten years ago, but it has excited such little interest that at the end of that time the printer has not sold twelve copies. Nonetheless it seems that this mine of harmony is not unknown to everyone. We have been very much surprised to find in performance that Messrs. Philidor, Gossec, Floquet, etc., have borrowed from it at their convenience and that entire pieces from their works may be found there, which has embarrassed them slightly.[56]

Philidor's taking was not condemned, and Gluck himself reportedly brushed aside the event.[57]

More recently, discovery of plagiarism has seemed to obsess some, particularly in the field of music criticism. Sigmund Spaeth advanced himself as a "tune detective." Shafter's treatise is replete with references to cunning plagiarists who disguise their thefts with ingenious little twists:

Clever infringers attempt to deceive composers by alterations and changes in their musical ideas, these disguises taking form, as the English Act says, of "colourable imitations." Colorable is another word for "camouflaged"; and a musical idea so treated is just as much an infringement as one taken openly. The difficulty of proving the theft does not lessen the liability of the thief. . . .

Much has been made of the fact that Brahms took the melody of the Westminster Chimes as the theme for his famous horn motif in his First Symphony. . . .

Olin Downes says that what Brahms changed is not the notes, but their rhythm. In the case of a copyrighted composition this change might be the colorable variation mentioned; but in the instance of music in the common domain, the copy is permissible. What is important in this respect is the fact that one may copy a melody by changing the rhythm‑-and still be infringing.[58]

Shafter believed that the value of experts lay in their ability to ferret out the devious ways in which composers disguised their thefts.[59] This school of commentators seems to find composers at once too lazy to be original and yet masters of creative deception.[60] Schenker railed against this naive approach to music:

It can only be regarded as a ridiculous attempt at debasement and disparagement of the diminutions of the masters when a certain literature busies itself with finding "wandering melodies" in the foreground, maintaining that similarities exist where they do not, with drawing lines of historical connection in every direction, where none in fact exist, or with pointing out plagiarisms where none are to be found. The employment of comparable superficial methods in language and literature would call forth general laughter and head-shaking over the deplorable intellectual state of any such writers and teachers.[61]

Schenker's reference to wandering melodies in the foreground suggests that the plagiarism enthusiasts tend to uncover similarities located in surface diminutions alone. These would-be detectives usually fail to take into account the context in which these surface diminutions appear. A more recent commentator has taken aim at those who delight in discovering rampant plagiarism:

These self-styled music critics seem to envision a world full of composers endeavouring mightily to conceal acts of plagiarism, even going so far as to change all the notes.[62]

Dictates of Style

Musical similarities should be judged with reference to stylistic context. Some substantial similarities may be explained by the particular vocabulary and idioms of the style to which they belong. When a work is classified as belonging to a particular style, the relevant questions for the trier of fact are: What musical elements define that style, and what musical characteristics do those elements share.

Style is determined by certain aspects held in common by a body of works. Therefore, the elements that define a particular musical style cannot be protected by copyright. Although, in one sense, a composer may develop his own individual style, the elements that properly define that composer's style should not include protectable elements. As a general proposition, "style" and "protectability" should each be defined in a way that excludes the other.

Virtually all musical elements may be subject to stylistic treatment:

In a musical composition, "style" refers to the methods of treating all the elements‑-form, melody, rhythm, etc. In practice, the term may be applied to single works (e.g., the style of Tristan compared to that of Die Meistersinger); to composers (the style of Wagner compared to that of Beethoven); to types of composition (operatic style, symphonic style, motet style, church style); to media (instrumental style, vocal style, keyboard style); to methods of composition (contrapuntal style, homophonic style, monodic style); to nations (French style, German style); to periods (baroque style, romantic style); etc.[63]

Obviously, no style can embrace all aspects of music simultaneously. Room must be left for individual creation.[64]

Elements that comprise a particular style may occur in the background (such as tonal vocabulary), middleground (such as formal structure), foreground (such as common rhythmic figurations), or even super-foreground (such as embellishments and performance techniques). The typical treatment of those elements should yield a general theoretical statement concerning their characteristics. Extra-musical descriptions, often used in common parlance to distinguish styles, do not advance the analysis and offer nothing relevant to the issue of copying.

Music analysts work with all styles and describe them in purely musical terms. Style analysis forms an important aspect of musicology.[65] It looks at normative aspects of music as opposed to the characteristics of individual works.[66] Jan LaRue describes the scope of style analysis as one of:

. . . exceptional completeness to ensure our observation of every nook and cranny of a piece, studying each of its musical elements in turn at various magnifications to fit all dimensions. Next we must try to understand the functions and interrelationships of these elements, so that we can make meaningful interpretations, identifying the significant aspects of each piece in relation to its composer and the stylistic relationship of each composer to his milieu.[67]

LaRue's approach varies little from that set forth in Chapters 5 and 6 of this work. Style analysis has the object of discovering and evaluating similarities, the same as the more critical approach to analysis suggested for the discovery of plagiarism. The distinction lies in some of the comparisons. Whereas critical analysis compares one work to another, style analysis compares one work to many.

Nothing concerning the question of style argues for reduced reliance on expert witnesses trained in all aspects of music analysis. The assertion that style dictates certain similarities does not limit the inquiry; it merely adds a specific question of protectability. It creates a need to explain how a certain style limits compositional choices and where the potential for original expression is greatest. If some elements slavishly follow formula, then originality or infringement must lie elsewhere.

Style does not change musical principles. The public often misperceives the distinctions that determine style; the public divides music into "kinds" rather than "styles." No "kinds" of music exist that would require a different kind of expert; music theorists are not taught to analyze a particular kind of music. Many theorists focus on more serious and complex genres simply because those genres present the greatest challenges and rewards.

The court in Selle v. Gibb criticized plaintiff's expert for not being well-versed in popular music:

Defendants are perhaps to some degree correct in asserting that Parsons, although eminently qualified in the field of classical music theory, was not equally qualified to analyze popular music tunes.[68]

There is only one kind of music theory; the modifier "classical" has no meaning. Further, anyone qualified to analyze classical music needs no additional qualifications to analyze popular tunes. One might as well question the ability of an expert in calculus to add and subtract or the ability of a lawyer to read a simple sales receipt.[69] Parsons' lack of familiarity with popular music was relevant in only one regard: he was unable to compare meaningfully the similarities he found between the two works with similarities that would normally appear in popular songs. That inability did not derive from his training in classical music, or any lack of expertise, but merely from his failure to investigate the stylistic traits of the works he analyzed.[70]

Classical music is made up of the parameters, functions, and relationships examined in Chapter 5. Popular music employs those same elements in the same ways. However, popular music presents less material, less complex designs, and less innovation. Popular idioms tend to employ limited musical vocabularies in rather predictable ways. Popular music elaborates middleground aspects less extensively than more serious forms of music. In general, most popular forms present simplified versions of musical materials common in the 19th century. Although some popular styles have developed complex and sophisticated approaches to certain musical parameters, popular forms tend to be characterized by timidity and adherence to familiar formulae.

It has been argued that popular music, because it presents relatively simple musical materials, should be subjected to less extensive analysis.[71] This approach prejudges the music; it requires laymen to characterize the complexity of the works involved and dictate how far the analyst may go. The argument subscribes in many ways to the separation of music into "kinds." Infringement litigation concerns the issue of copying, not the value of particular forms of music.[72] Thus, the necessity of a complete forensic analysis should not hinge on style, form, or complexity.

Economic Demands

The music typically litigated has economic value that derives from its mass appeal. Music that attempts to challenge the public and change public tastes stands considerably less chance of "making the charts." The public likes what it is accustomed to hearing.[73] Similarities, imitations, and repetitions‑-both of internal and extraneous materials‑-form an important aspect of the art:

The most primitive impulse toward artistic creativity finds its expression in imitation, so the first response on the part of the recipient is the joy of recognition.[74]

In large-scale serious works, repetition of original, internal materials provides the listener with the joy of recognition. In smaller-scale popular works, composers have a greater need to imitate extraneous materials.

Many highly successful writers of popular music have learned to fit slightly new surface diminutions onto tried and true middleground formulae. A comparison of such works will yield many similarities, but the trier of fact must understand those similarities for what they are. The importance of designating similarities as belonging to foreground, middleground, or background cannot be over-emphasized.

Formula writing is not disingenuous or even significantly different from the practices of some composers of more serious music. Virtually all composers adhere to certain stylistic formulae. Like music's pervasive dichotomy of unity and disunity, music also contrasts the familiar with the new. Classical music uses familiar conventions but injects more originality. Its complexity necessitates more intense listening and a greater number of repetitions before understanding is achieved. Complexity, subtlety, variety, and unity of expression account for the longevity of classical music's appeal.

Popular music generally has a higher ratio of familiar to new. It needs to strike a familiar chord immediately. Unlike serious forms, popular music relies on first impressions. The listener must find instant gratification if commercial success is to be achieved. A popular song has perhaps three minutes to capture the listener's interest and comprehension. The composer achieves this immediate appeal by evoking a sense of familiarity, which can only come from imitation of external materials and formulae.

Composers respond to many historical practices, stylistic formulae, and economic demands. Although these forces in and of themselves do not provide a defendant with an ironclad defense, the trier of fact should recognize that not all similarities reflect a composer's desire to profit from the labors of others. Guidelines for the substantiality of similarities should separate similarities that music considers traditional and necessary from those that misappropriate protectable elements.

The Question of Coincidence

Other factors beyond the composer's control contribute to the perception of similarities. Shafter's treatise cites a paucity of possible combinations in music that creates inevitable similarities, and he even makes the preposterous statement that "it is generally agreed that the original fund of melodic ideas has been exhausted."[75] Sherman, a more astute commentator, falls into the same trap:

It must be remembered that there is not the same number of pleasing combination of notes as there is of words. "Popular songs, particularly, lie within a very small radius. In a confined space, similarity of tone construction is inevitable. Practically every original idea the composer can think of has appeared somewhere before; it is a matter of probabilities, and every day the number of new possibilities grows less."[76]

The statement contains two different propositions: that musical combinations have been exhausted, and that popular idioms have been exhausted. Further, one cannot tell whether Shafter believes the fund of ideas to be truly exhausted or merely dwindling at a startling rate.[77] The court has reiterated the notion that musical expression is severely limited, and defense counsel almost invariably resort to this argument to support the coincidence of similarities.[78]

Exhaustion of Musical Combinations

A kernel of truth can be found in Shafter's statement regarding musical combinations: the number of pitches in traditional western music is limited to twelve, not counting octave equivalents, and the combinations that convey tonality are fewer than the total mathematical possibilities.[79] However, Shafter's statement distorts this truth by attaching significance to the number of pitches. The number is irrelevant to both music and law. One might note also that there are only twenty-six letters in the alphabet, and of the many possible combinations, only a few are intelligible. Indeed, there are only ten numerical symbols in our mathematical system, yet these symbols can express an infinite array of values.

A court would not indulge a literary comparison based on letters or a numerical one based on numerical symbols in corresponding columns. There is no substantial similarity between "the rapist" and "therapist" nor between the numbers 109,623 and 189,623.[80] A court cannot be expected to recognize the same fallacy applied to music; basic education no longer extends to music as it does to math and language. But this kind of false analysis, sometimes perpetrated by the experts themselves, characterizes many plagiarism cases. The court has cast a wary eye on music experts, suspecting that their analyses do not always comport with aural perception. Similarities between two works can almost always be found and, barring literal duplication, divergences can be found as well. In either case, the analyst should help the trier of fact to distinguish the relevant from the irrelevant. Legal relevance and musical relevance should converge. Thus, the analyst, despite his ignorance of law, cannot present what he knows to be musically insignificant.

Exhaustion of Musical Vocabularies

Serious composers have historically sought out new methods of expression, often venturing beyond the public's tastes. This progression is not simply a result of the exhaustion of musical ideas within established styles. Elements of style ebb and flow; changes often represent a retreat to past practices as much as the adoption of new ones. Yet Shafter claims not only that past stylistic vocabularies have been exhausted but that no possible combinations of musical elements remain that can be deemed original. Composers, Shafter seems to believe, can only repackage old ideas.

The attitude voiced by Shafter denies the power and expressive capacity of music. Although the layman may perceive some logic in Shafter's false proposition, the layman does not act in conformity with this notion. Even in the more limited arena of popular music, where the court believes few permutations will please the public's infantile demands,[81] the industry continues to generate musical materials at an astonishing rate. The public apparently finds these materials sufficiently new and satisfying to continue purchasing them. To some extent then, the lay listener has already judged a vast number of works with many surface similarities and found them to be substantially original. If music retains these almost inexhaustible expressive qualities within the rigid confines of popular idioms, then the art as a whole need not fear a shortage of new possibilities.

The fact that music retains many possibilities for original expression does not mean that similarities should always be difficult to find. But defendants in infringement actions need a better argument than Shafter's to explain the existence of numerous apparent similarities. For example:

(1) A quantity of similarities that seems impressive is often in fact quite meager when compared to the myriad relationships that identify each individual work.

(2) A cumulation of similarities appearing in different parameters is musically irrelevant except to the extent those parameters combine in a common purpose.

(3) Note-by-note comparisons made without reference to context and function fail to analyze the music and account only for a meaningless sharing of isolated elements. Such correspondence can be found in the comparisons of many pairs of indisputably original works.

(4) Note-by-note comparisons purporting to show that prior works do not correspond to the same degree as the works at issue demonstrate only empty quantifications.

The defendant needs to convince the trier of fact not so much that similarities are inevitable but that inevitable similarities are not substantial. Shafter's proposition relies on demonstrably false data, but showing the insubstantiality of inevitable similarities is inherent in a careful and discerning approach to forensic analysis.

Musical Guidelines for Substantial Similarity

The more comprehensive analytical approach suggested in this work should provide the means of separating meaningless surface correspondence from similarities with musical significance. Surface features will always contain coincidental similarities. If, in addition to sharing these surface features, two works assign the same function to those features, then the similarities begin to be perceptible. If the functional similarities also appear in the same context, episodic similarities may be apparent. Finally, if all of these similarities are strung together in a sustained and concerted fashion, then the lay listener may recognize the similarities as a derivation of another work.

Even musically significant similarities, however, may lack legal significance. Given the complex nature of music, similarities that lack complexity and depth should not earn the label "substantial."[82] The law must look to the nature of the art it examines in these cases. Music does not exist in a set of pitches or static values, and an analysis of similar notes does not reveal the important relationships that control perception. Musical perception is a complex process; it deserves to be considered an intelligent process rather than a purely reflexive one. The lay ear processes the same data and connects the same relationships as that of the expert, but the lay listener generally does not understand the nature of the process or possess the means to articulate the experience. The expert witness should illuminate the process of listening for the lay trier of fact, and a trier of fact should combine rather than separate the power of his ear and mind.

The principal attribute of similarities that deserve to be considered substantial is that they should be perceptible simultaneously by the ear and mind.[83] Bifurcation of this process has served a useful, analytical purpose; but like all analytical segmentations, the sum of the segments does not provide a complete answer.[84] The trier of fact must assimilate the data before making a final evaluation. The courts' seriatim treatment of their bifurcated tests has hindered assimilation. Courts have not successfully combined parts into a meaningful whole.[85]

In addition to this general proviso, three separate and specific guidelines are suggested that must be made with direct reference to the analysis accomplished through the procedures outlined in Chapters 5 and 6. In order to reach a conclusion that substantial similarities exist, the trier of fact should find all three of the following conditions to be met:

(1) Plaintiff should be required to show similarities in the middleground. Function and context can best be determined at this level. The middleground accounts for larger-scale and more distant relationships that control how musical details are heard. Middleground aspects also guide the discovery of greater abstractions than foreground materials. Music cannot be substantially similar in foreground features alone, because similarities confined to foreground features are likely to be the result of idiomatic figurations or pure coincidence.[86] If only minor changes have been made to disguise an act of plagiarism, similarity of middleground will remain intact and support the plaintiff's argument.

(2) Plaintiff should be required to show similarities in the foreground. Because the middleground tends to reflect more abstract elements, similarities confined to the middleground may be unprotectable[87]. Stylistic formulae, key relationships, rules of counterpoint and harmony, as well as other middleground aspects all possess a greater likelihood of similarity. The critical question for the trier of fact is whether the defendant realized, elaborated, or prolonged middleground similarities using the same foreground materials as the plaintiff. Only the foreground contains elements specific enough to corroborate middleground similarities and support a claim of infringement.

(3) Plaintiff should be required to show a nexus between specific foreground and middleground similarities. The random cumulation of disparate elements, some taken from foreground and some from middleground, does not suggest copying. Isolated similarities will each fail under one of the first two guidelines. Substantiality is not measured by quantity. Only the relationship between foreground and middleground similarities tends to show copying as opposed to coincidence. The nexus ties similarity of surface to similarity of function‑-the same means of expression employed to express the same thing.

Application of the Guidelines

Finally, these guidelines can be applied to the analytical comparisons made in Chapter 6 of the works litigated in Gaste v. Kaiserman.[88] Subject matter comparisons of the foreground revealed similarities in Paradigm A, the most prevalent motive, but not in Paradigms B through F.[89]

The functional analysis of harmony revealed some similarities and some differences. In Phrase 2, "Feelings" employed a more elaborate foreground progression, relegating the similarities to a slightly more abstract level. Phrase 1 showed substantial divergence in the progression of measures 1 through 5, but correspondence in the modulation to the relative major. Phrase 1 therefore lacks significant foreground similarities in the progression.[90]

Rhythmic reductions yielded significant background similarities, particularly of Phrase 1. But the foreground remained quite dissimilar throughout, and these dissimilar foreground figurations were deemed an important internal aspect of each piece.[91]

Similarities revealed through the hierarchical analysis were most pronounced in the background. Middleground similarities appeared but were interspersed with significant distinctions as well. The foreground lacked correspondence except for the figuration already noted as Paradigm A.[92]

Formal analysis revealed notable correspondences located exclusively in the middleground. Some distinctions, albeit less significant than that revealed in other comparisons, also appear in the formal analysis.[93]

Temporal comparisons tended to confirm that the majority of similarities discovered could not be located in the foreground. Besides Paradigm A, only the outlines of a similar sequence in Phrase 2 could be noted in the melodic comparison.[94]

"Pour Toi" and "Feelings" exhibit relatively significant similarities in the middleground, but foreground similarities are quite limited. Although a comparison of the two works meets the first guideline, similarities in the foreground, and the second guideline, similarities in the middleground, no nexus can be established between the two. The foreground similarity is limited to a short repeated figure that conveys different harmonies and rhythms. Thus, the plaintiff would fail to meet the third guideline. These guidelines, if employed, would lead the trier of fact to find that "Feelings" did not copy "Pour Toi"‑-a conclusion contrary to the one reached at trial.

---

[1] As explained in Chapter 5, the analyst should reduce both works at issue to show the entire spectrum of internal temporal segmentations and a hierarchical analysis from foreground, through middleground, to background. In addition, the analyst may choose to isolate parameters in a way that illuminates the various internal functional relationships.

[2] As explained in Chapter 6, the analyst should compare all surface segments of both works in a systematic way (organizational aspects of semiotics comprising the suggested approach), compare the structures of both works through traditional formal analysis, compare the harmonic and rhythmic functions of the works, and compare middleground and background layers through hierarchical reductions. (The author suggests Schenkerian analysis for these comparisons.)

[3] "Testimony in the form of an opinion or inference otherwise admissible is not objectionable because it embraces an ultimate issue to be decided by the trier of fact." Fed. R. Evid. 704.

[4] "Proof of copying is crucial to any claim of copyright infringement because no matter how similar the two works may be (even to the point of identity), if the defendant did not copy the accused work, there is no infringement." Selle v. Gibb, 741 F.2d 896, 901 (7th Cir. 1984) (citing Arnstein v. Edward B. Marks Music Corp., 82 F.2d 275 (2d Cir. 1936).

[5] Sherman, Musical Copyright Infringement: The Requirement of Substantial Similarity, 22 Copyright L. Symp. (ASCAP) 81, 83 (1977).

[6] "Striking similarity is an extremely technical issue‑-one with which, understandably, experts are best equipped to deal." Id. at 96 (cited with approval in Selle, 741 F.2d at 904). "Internal proof of access may rest in an identity of words or in the parallel character of incidents or in a striking similarity which passes the bounds of mere accident." Wilkie v. Santly Bros., Inc., 91 F.2d 978, 979 (2d Cir.), cert. denied, 302 U.S. 735 (1937).

[7] Sherman, supra note 5, at 93 (citing Marks v. Leo Feist, Inc., 290 F. 959 (2d Cir. 1923); Life Music, Inc. v. Wonderland Music Co., 241 F. Supp. 653 (S.D.N.Y. 1965); Gingg v. Twentieth Century-Fox Film Corp., 56 F. Supp. 701 (S.D. Cal. 1944); Carew v. R.K.O. Radio Pictures, 43 F. Supp. 199 (S.D. Cal. 1942); Hirsch v. Paramount Pictures, Inc., 17 F. Supp. 816 (S.D. Cal. 1937)).

[8] Selle, 741 F.2d at 904 (citing Testa v. Janssen, 492 F. Supp. 198, 203 (W.D. Pa. 1980) (quoting Stratchborneo v. Arc Music Corp., 357 F. Supp. 1393, 1403 (S.D.N.Y. 1973)).

[9] 741 F.2d at 904 (citing Scott v. WKJG, Inc., 376 F.2d 467, 469 (7th Cir.), cert. denied, 389 U.S. 832 (1967); Arnstein v. Porter, 154 F.2d 464, 468 (2d Cir. 1946); Scott v. Paramount Pictures Corp., 449 F. Supp. 518 (D.D.C. 1978)).

[10] 741 F.2d at 904 (quoting Sherman, supra note 5, at 96).

[11] 741 F.2d at 904 (citing Sheldon v. Metro-Goldwyn Pictures Corp., 81 F.2d 49, 54 (2d Cir.), cert. denied, 298 U.S. 669 (1936)).

[12] 741 F.2d at 904 (citing Nichols v. Universal Pictures Corp., 45 F.2d 119, 122 (2d Cir. 1930), cert. denied, 282 U.S. 902 (1931)).

[13] 741 F.2d at 904 (citing ABKCO Music, Inc. v. Harrisongs Music, Ltd., 722 F.2d 988, 998 (2d Cir. 1983).

[14] 741 F.2d at 905.

[15] Id.

[16] Case No. 88-6660 (C.D. Cal. 1988), aff'd, 907 F.2d 154 (9th Cir. 1990).

[17] See Baxter, Record at 205-11. Because the jury found the portion allegedly copied to be unprotectable, the question of striking similarity was not decided and, thus, never argued on appeal.

[18] 741 F.2d at 904. See supra pages 420 and 203.

[19] The competence argument also fails to distinguish coincidence from independent creation. If the defendant produces a work that exceeds his creative abilities, then copying is the only explanation. The defendant's lack of competence rules out independent creation, but it rules out coincidence at the same time. Lack of competence merely becomes the reason that the similarities are not coincidental.

[20] Selle's citations to Nichols and Darrell further blur any distinction between coincidence and independent creation by allowing the issue of protectability to seep in. Selle, 741 F.2d at 904 quotes Darrell: "simple, trite themes . . . are likely to recur spontaneously . . . and [only few] . . . suit the infantile demands of the popular ear." Darrell v. Joe Morris Music Co., 113 F.2d 80 (2d Cir. 1940). The cite to page 122 of Nichols seems to refer to the following passage: "It is indeed scarcely credible that [the plaintiff] should not have been aware of those stock figures, the low comedy Jew and Irishman. The defendant has not taken from her more than their prototypes have contained for many decades. . . . Even though we take it that [the plaintiff] devised her figures out of her brain de novo, still the defendant was within its rights." Nichols v. Universal Pictures Corp., 45 F.2d 119, 122 (2d Cir. 1930), cert. denied, 282 U.S. 902 (1931).

[21] "Even without proof of access, plaintiff could still make out her case if she showed that the two works were not just substantially similar, but were so striking similar as to preclude the possibility of independent creation." Ferguson v. National Broadcasting Co., Inc., 584 F.2d 111, 113 (5th Cir. 1978) (citing 3 M. Nimmer, Nimmer on Copyright § 13.01[A]). See also, Sherman, supra note 5, at 90.

[22] Two works copied in every detail from a common source will match each other precisely.

[23] Selle, 741 F.2d at 905.

[24] Id. (emphasis added).

[25] A representative sample probably does not exist. Computer searches, although perhaps capable of making a sufficient number of comparisons, fail in two respects: the literature has not been computerized, and no criteria exist that would allow the computer to determine what is similar.

[26] An example of prior art introduced by the plaintiff to prove the negative will necessarily be viewed as a straw man.

[27] "`Striking similarity' is not merely a function of the number of identical notes that appear in both compositions. Cf. Wilkie v. Santly Brothers, Inc., 13 F. Supp. 136 (S.D.N.Y. 1935), aff'd, 91 F.2d 978 (2d Cir.), cert. denied, 302 U.S. 735 (1937), aff'd on reargument, 94 F.2d 1023 (2d Cir. 1938) (comparison of note structure demonstrates striking similarity), and Jewel Music Publishing Co. v. Leo Feist, Inc., 62 F. Supp. 596 (S.D.N.Y. 1945) (in light of plaintiff's inability to establish access, degree of similarity despite identity or near identity of several bars was not striking). An important factor in analyzing the degree of similarity of two compositions is the uniqueness of the sections which are asserted to be similar." Selle, 741 F.2d at 903-04.

[28] 741 F.2d at 905.

[29] "Mozart's ability to define character by purely musical means, to write for each of the three sopranos . . . in an individual and characteristic way, provides the variety necessary for his structure. The essential innovation that was the keystone of his success was his extraordinary development and expansion of the ensemble. . . . The tour de force of this new conception of musical continuity in drama as an increasing complexity of independent units is the famous second act finale, which moves from duet, through trio, quarter, and quintet to septet in a magnificently symmetrical tonal scheme.

"This synthesis of accelerating complexity and symmetrical resolution which was at the heart of Mozart's style enabled him to find a musical equivalent for the great stage works which were his dramatic modes. . . . For the first time in the history of opera, the musical version could accept, and even welcome, comparison with the greatest dramatic achievements." Charles Rosen, The Classical Style (New York: Norton, 1972), 182-83.

[30] Alfred M. Shafter, Musical Copyright, 2d ed. (Chicago: Callaghan, 1939), 216.

[31] What were rules for the common practice period may be better understood as good advice for contemporary tonal works. Parallel fifths were found in "Pour Toi," a tonal work. See supra page 388. "Pour Toi" illustrates both the advisability of avoiding parallel fifths and the fact that composers sometimes ignore that advice.

[32] Louis Nizer, My Life in Court (Garden City, N.Y.: Doubleday, 1961), 252.

[33] The chord is a diminished seventh (vii⁰⁷/ii) in C Major accompanying an A (a non-chord tone) in the melody.

[34] Nizer, supra note 32, at 253.

[35] Id. at 252.

[36] Id. at 253.

[37] Nizer also asserts that the resolution of the chord provided further evidence of copying. "Once a composer has inserted a chromatic chord among a series of diatonic chords, he cannot easily get back to his simple chords. He must build a musical bridge back, or, as the musicians put it, he must resolve it. There are eight choices for this resolution. Once more [plaintiff] Maurice Baron chose an unusual resolution. Remarkably enough, the same resolution was found in [the accused work] `Rum and Coca-Cola.'" Id. See Appendix H, at 1278, measure 9. Nizer's analysis is at best misleading. Chromatic chords cannot be contrasted to "simple" chords, and diatonic chords must be resolved also. Because the diminished seventh chord is symmetrical (its pitches are all separated by the same interval), it has a highly ambiguous nature that allows numerous resolutions‑-more than eight. Its symmetry allows any of its four pitches to function as the root, it follows three conventional resolutions, and it can progress to either major or minor. 4 x 3 x 2 = 24 conventional resolutions. Plaintiff's work followed one of these conventions, albeit not the most obvious one. (vii^o7/ii resolved to V⁷.)

[38] Nicholas Cook, A Guide to Music Analysis (New York: Braziller, 1987), 12; quoting Donald Tovey, Essays in Musical Analysis VI: Miscellaneous Notes, 61.

[39] Trademark law has employed consumer surveys to solve the similar question of likelihood of confusion. The actual confusion of a statistical sample provides evidence of the likelihood that potential buyers will be confused by similar trademarks. Consumer surveys, however, are conducted routinely. Scientific standards have been imposed not for litigation purposes but by businesses that demand accurate marketing data. Many elements of music industry tend to exploit markets rather than establish them; consequently, the industry has not developed these marketing tools in a way that is readily translatable to litigation purposes.

[40] The cases examined in this work, however, do not entirely confirm these classifications. Ronald Selle, an unpublished antique dealer, presented a case not without merit. Selle v. Gibb, 567 F. Supp. 1173 (N.D. Ill. 1983), aff'd, 741 F.2d 896 (7th Cir. 1984). Leslie Baxter and Louis Gasté, both published but relatively minor composers, did not present cases substantially better than Selle's. Baxter v. MCA, Inc., 812 F.2d 421 (9th Cir.), cert. denied, 484 U.S. 954 (1987), on remand, Case No. 88-6660 (C.D. Cal. 1988), aff'd, 907 F.2d 154 (9th Cir. 1990); Gaste v. Kaiserman, Case No. 86 Civ. 5671 (S.D.N.Y. 1986), aff'd, 863 F.2d 1061 (2d Cir. 1988). See Chapter 4. Steve Karmen, a highly successful jingle writer, brought a suit without musical merit, although non-musical facts may have supported his claim. Anheuser-Busch, Inc. v. Elsmere Music, Inc., 633 F. Supp. 487 (S.D.N.Y. 1986). The most spurious claims, however, do seem to come from unknown and infrequent composers.

[41] "Barry Gibb's testimony included a detailed explanation of a work tape which was introduced into evidence and played in court. This tape preserves the actual process of creation during which the brothers, and particularly Barry, created the tune of the accused song while Weaver, a keyboard player, played the tune which was hummed or sung by the brothers. Although the tape does not seem to preserve the very beginning of the process of creation, it does depict the process by which ideas, notes, lyrics and bits of the tune were gradually put together." Selle, 741 F.2d at 899.

[42] Keyt, An Improved Framework for Music Plagiarism Litigation, 76 Calif. L. Rev. 421, 424 (1988).

[43] "The existing legal framework for copyright actions tends to create perverse incentives through misguided analysis and unclear liability standards. Plagiarism suits have enormous nuisance value: a defense can cost upwards of $100,000 in attorneys fees, in addition to the time lost by musicians in attending depositions and trial instead of in composing, performing, or recording. . . ." Id. at 423-24.

[44] Charles Michael Carroll, "Musical Borrowing‑-Grand Larceny or Great Art?" College Music Symposium 18 (1978): 11.

[45] See Gustave Reese, Music in the Middle Ages (New York: Norton, 1940), 249 ff. The earliest description of liturgical polyphony, "Organum" (part of the body), is found in the 9th-century Musica enchiriadis (Handbook on music), but scholars generally believe polyphony originated in earlier nonliturgical music never documented.

[46] Donald Jay Grout, A History of Western Music, 3d ed. (New York: Norton, 1980), 98.

[47] Id. at 165.

[48] Gustave Reese, Music in the Renaissance (New York: Norton, 1959), 197; quoting Oliver Strunk, Origins of the "L'Homme armé Mass, abstract in Bulletin of the American Musicological Society, No. 2 (1937), 25.

[49] Id. at 202.

[50] Id. at 448 (quoting Erasmus).

[51] Id. at 449 (quoting the canon dealing with music to be used in the Mass adopted by the Council of Trent, 10 September 1562).

[52] Percy Scholes, Oxford Companion to Music (London: Oxford University Press, 1970), 853. Bach also wrote a series of imitative works. He made keyboard arrangements of Vivaldi's violin concertos and borrowed themes of Corelli, Legrenzi, and Albinoni. Bach's purpose in making these adaptations may have been simply to learn the Italian style. Manfred F. Bukofzer, Music in the Baroque Era (New York: Norton, 1947), 276. Whereas Bach's unattributed takings were intended primarily for performances in church or school, Handel passed off more extensive borrowings as his own. Carroll, supra note 44, at 15.

[53] Variations on a theme as a genre constitute an implied honor, often rescuing themes and even composers from obscurity. Beethoven resurrected nearly forgotten works of Diabelli, Dressler, Dittersdorf, Haibl, Righini, and Winter. Carroll, supra note 44, at 13.

[54] Carroll cites as prime examples of the best kind of borrowing: Beethoven's Missa Solemnis quoting Handel's Messiah; Schubert's Symphony No. 7 (written the year after Beethoven's death) quoting Beethoven's Symphony No. 9; Schumann's Fantasy in C Major quoting Beethoven's An die ferne Geliebte; and Brahms' Symphony No. 3 (written immediately after Wagner's death) quoting Wagner's Tannhaüser. Id. at 17-18.

[55] Id. at 12. The date precedes music's emergence as self-expression but follows the printing press by 150 years.

[56] Id. at 15-16 (quoting Petit de Bachaumont, Mémoires secrets.)

[57] Id. at 16-17.

[58] Shafter, supra note 30, at 193-99 (emphasis in original). See also, Orth, The Use of Expert Witnesses in Musical Infringement Cases, 16 U. Pitt. L. Rev. 232 (1955). "The propensity has existed for a long time to steal the tunes of other composers. Lizst made free use of gypsy melodies and Handel was a notorious plagiarizer. Upon listening to Brahms' First Symphony, for instance, one might well recognize that the broad chorale of the fourth movement has been transplanted from the "Ode to Joy" motif in Beethoven's Ninth Symphony, which, in turn, had its origin in the minuetto of Mozart's Haffner Symphony, which came from Sandrina's aria "Una voce vento al core." Id. (citing Shafter, supra note 30, and Fox, Evidence of Plagiarism in the Law of Copyright, 6 U. of Toronto L.J. 414, 415 (1946)).

[59] Shafter, supra note 30, at 189.

[60] "If it were not for the fact that juries are almost never called in [popular music cases], the average ear test would be somewhat ineffective in holding copiers for infringement, because infringing tunes are usually written‑-probably with success‑-to deceive this very class of listener. Thus the infringer, who had cleverly disguised the tune by rhythmic and/or harmonic tricks, would be given great leeway, amounting in the case of a skillful and careful job of plagiarism to virtual immunity." Orth, supra note 58, at 236.

[61] Heinrich Schenker, Free Composition, trans. and ed. Ernst Oster (New York: Longman, 1979), 105.

[62] Keyt, supra note 42, at 424 n.17. "We are dealing, of course, with a romantic concept still widely shared by laymen: the composer creates in an intoxicated daze, at `white heat,' and if he uses material of earlier creation, he is a plagiarist, a swindler, and a thief. One wonders why the other romantic conception of the artist as an irresponsible individual who must not be measured by the standards of bourgeois ethics is not applied in this instance. . . . Only those who do not understand the process of musical composition, who cannot see and feel the subtlety of transfiguration that can be created by a changed melody, even a single note, rhythm, or accent, have made a moral issue of something that is a purely esthetic matter." Id. at 425 (quoting Paul Henry Lang, George Frederic Handel (1966), at 559-69).

[63] Harvard Dictionary of Music, 2d ed. (1969), s.v. "Style."

[64] In MCA, Inc. v. Wilson, defendant argued essentially that pervasive similarities were the result of stylistic demands, seemingly suggesting that the style left no room for originality. This expansive definition style swallows protectable expression. 425 F. Supp. 443 (S.D.N.Y. 1976), aff'd and modified, 677 F.2d 180 (2d Cir. 1981).

[65] See, e.g., Guido Adler, Der Stil in der Musik (Leipzig: Breitkopf & Härtel, 1911); see also, Chapter 3, notes 72-74.

[66] An analyst must first examine the musical environment of the work‑-the normative practices of the individual composer and historical period. John D. White, The Analysis of Music, 2d ed. (Metuchen, N.J.: Scarecrow Press, 1984), 11. See also, Michael R. Rogers, Teaching Approaches in Music Theory (Carbondale: Southern Illinois University Press, 1984), 76-77.

[67] Jan LaRue, Guidelines for Style Analysis (New York: Norton, 1970), 2.

[68] Selle v. Gibb, 741 F.2d 896, 904 (7th Cir. 1984). The Selle court was not the only one to misunderstand an analyst's training. For example, the court in Herald Square Music Co. v. Living Music, Inc., 205 U.S.P.Q. 1241 (S.D.N.Y. 1978), asked plaintiff's expert Chasins if his career had concerned only serious music. Record, at 338.

[69] One anecdote relates an occasion on which Albert Einstein was playing a violin sonata with Artur Schnabel. When Einstein made a mistake, Schnabel turned to him and said, "What's the matter, Albert? Can't you count?"

[70] The court qualified Parsons as an expert, and his testimony was not disregarded. The court found his testimony not sufficient rather than not credible or relevant. Had Parsons testified that the similarities could result only from copying, as he needed to do in order to prove striking similarity, a reasonable jury could have believed him. Judgment n.o.v. resulted from Parsons' failure to state that the similarities could only result from copying. Parsons could not rule out prior common sources because he had not researched the literature. See supra pages 149 and 157.

[71] "A high level test need not be given to low caliber material." Orth, supra note 58, at 253.

[72] "In testing popular music, L. Hand's ingenuous or "audience" approach is more realistic than the use of intelligence and astute scrutiny by musical experts advocated by [Judge] Clark [dissenting in Arnstein v. Porter]. Clark's footnote on the employment of intellect in "music appreciation" betrays that the author had a higher order of music under consideration." Orth, supra note 58, at 254 (citing Arnstein v. Porter, 154 F.2d 464, 476 (2d Cir. 1946)). Orth argues that defendants suffer most from extensive analysis, because the inquiry points out similarities perceptible only to the trained ear. Id. at 253-54.

[73] Schenker held public tastes in low esteem. "Concerning the law of large numbers: two, four, or eight people can easily be brought together by games, if necessary also by intellectual entertainment. Art can bring together as many as two or three thousand people. But to assemble and entertain 50,000 people‑-this can be accomplished only by bullfights, cock fights, massacres, pogroms: in short, a brutal ranting and raving, a demented and chaotic outcry." Schenker, supra note 61, at 159.

[74] Oswald Jonas, Introduction to the Theory of Heinrich Schenker, ed. and trans. by John Rothgeb (New York: Longman, 1982), 1.

[75] Shafter, supra note 30, at 197. Sigmund Spaeth said, "Music has gone as far as it can in originality." Orth, supra note 58, at 847. Shafter's statement is refuted by Keyt, supra note 42, at 432. It is difficult to reconcile Shafter's pro-plaintiff stance that composers regularly misappropriate previous works with his extreme pro-defendant belief that originality is virtually impossible. One can only conclude that Shafter discerns little subtlety, variety, and expressive capacity in music.

[76] Sherman, supra note 5, at 124 (citing Shafter, supra note 30, with approval).

[77] If musical combinations in general are truly exhausted, Shafter has no need to address the specific exhaustion of popular idioms. In fact, he leaves no room for originality of melody but admits a few possibilities remain in popular forms. Shafter's statements thus seem to be simple incantations, devoid of substance.

[78] See supra page 275.

[79] Possible combinations are extensive. For example, the possible sequences of the twelve tones without duplication of any pitch (the twelve-tone rows common to serialism) number over 479 million. Allowing duplications, an array of possibilities based on only ten choices adequately supports the nationwide telephone system with 10 billion numbers. The possible successions of twelve pitches, allowing duplications but still without exceeding the bounds of the octave, number almost 9 trillion.

[80] This kind of numerical similarity might have significance to telephone numbers where the succession of values is everything, but musical impressions rely on much more complex relationships than simple succession.

[81] Darrell v. Joe Morris Music Co., 113 F.2d 80 (2d Cir. 1940).

[82] This runs counter to Orth's proposition: "While the rule is that numerous similarities will create an inference of copying sufficient to show infringement if this circumstantial evidence is convincing, it has been held that this rule is not applicable where the claimed similarities in the two works are so technical or complex that the average person is unable to hear any substantial similarity. . . ." Orth, supra note 58, at 238 (citing Gingg v. Twentieth Century Fox Film Corp., 56 F. Supp. 701 (S.D. Cal. 1944). In fact Gingg refers only to similarities that are "highly technical in nature." Orth apparently added the term "complexity." Complexity of relationships tends to make similarities more apparent rather than calling for greater expertise.

[83] Although the author disagrees with many of Orth's premises, Orth reaches a similar conclusion regarding the combination of ear and intellect. "Expert testimony would be used to guide the jury, to point out significant similarities and possibly to explain and clarify psychological reactions to the two compositions. . . . Such testimony would be heard along with the impression made on the jury's ears by several careful playings, and, taking the whole testimony and complex of impressions together, the decision would be left to the jurymen, under appropriate summary and instructions by the judge. Included should be an instruction to the effect that the jury must, in whatever state of enlightenment the experts have developed in it, now hear substantial similarities between the works before holding defendant liable.

"This is not the approach Judge Frank [in Arnstein v. Porter] advocated for use with a jury. It contains no compartmentalizing of expert testimony and then putting it aside when the final determination of illegal copying is made. It is instead that relevant expert analysis and opinion is part of all the evidence receivable, and that the issue of infringement of popular music is then decided by a possibly enlightened consensus of lay ears." Orth, supra note 58, at 255-56 (emphasis in original).

[84] Judge Clark's dissent in Arnstein v. Porter points out the potential problem. "I find nowhere any suggestion of two steps in the adjudication of this issue, one of finding copying which may be approached with musical intelligence and the assistance of experts, and another that of illicit copying which must be approached with complete ignorance; nor do I see how rationally there can be any such difference, even if a jury‑-the now chosen instrument of musical detection‑-could be expected to separate those issues and the evidence accordingly. 154 F.2d 464, 476 (2d Cir. 1946).

[85] The bifurcation was intended to limit expert testimony, but Fed. R. Evid. 704, abolishing the ultimate issue rule, provides an adequate basis for allowing expert testimony on all infringement tests. Der Manuelian, The Role of the Expert Witness in Music Copyright Infringement Cases, 57 Fordham L. Rev. 127, 147-48 (1988).

[86] "Preservation of context must be a crucial element of copying. It is not enough to compare only strings of acoustical events. The comparison must include the structures that the sounds articulate." Keyt, supra note 42, at 437 (emphasis in original).

[87] Indeed, Chapter 8 will argue that the line between protectable idea and unprotectable expression in music should be drawn between the foreground and middleground. Thus, the guidelines for substantial similarity advanced here require that similarities extend to unprotectable aspects of the music. The guidelines follow one aspect of the Ninth Circuit’s intrinsic-extrinsic test: requiring first similarity of idea, and then similarity in the expression of that idea. See infra page 509.

[88] Case No. 86 Civ. 5671 (S.D.N.Y. 1986), aff'd, 863 F.2d 1061 (2d Cir. 1988).

[89] See supra page 362.

[90] See supra page 368.

[91] See supra page 380.

[92] See supra page 398.

[93] See supra page 404.

[94] See supra page 411.