In Hein v. Harris (1923) the defendant attempted to persuade the court that prior art songs were similar enough to plaintiff's "Arab Love Song" for the court to conclude that in writing his song the plaintiff had simply borrowed existing generic musical conventions without adding sufficient original expression on which to stake an independent claim of copyright.
Learned Hand, who wrote the Hein opinion, appears to have taken defendant's prior art defense seriously. One finds in the docket on the sheet music of the so-called prior art, Hand's meticulous annotations marking the common pitches among these numbers. Holding for the plaintiff, however, Hand found that the similarities among the prior art works and the plaintiff's extended only to generic stylistic features, and that defendant's melody was much more similar to plaintiff's than were those of any of the alleged prior art songs.
Here is a sample of Learned Hand’s handwriting used to confirm that it was Hand who marked up the sheet music offered by the defendant, and found in the case docket.
This handwriting sample is from a letter Learned Hand wrote on Christmas Eve, 1939 to Charles E. Wyzanski, who was the first law clerk of Learned's cousin Augustus Hand, and who was about to become a judge himself. It is taken from a reproduction of the letter in The Remarkable Hands: An Affectionate Portrait, edited by Marcia Nelson, and published in New York in 1983 by the Foundation of the Federal Bar Council.
Charles Brown's "Mobile Prance" - published by defendant Harris himself - was one of the numbers offered to the court as prior art evidence. This is a page of the copy that Hand marked up (taken from the case docket); the tidy hatch marks over certain notes are in keeping with the emphasis Hand placed in this, and similar cases, on the precise number of pitch concurrences between disputed works as highly probative on the questions of similarity and copying.
If Learned Hand were trying this case today how likely would it be that he would still hold for the plaintiff? Below we consider how the parties in Hein v. Harris might have used music notation software to support their respective positions. The reader might consider the question whether, on balance, use of music digital technology makes infringement more or less discernible, and more specifically the likelihood that music evidence in digital form might have affected Hand's decision in this case.
Plaintiffs could point up similarities between musical works using notation software to normalize melodies that at first blush might appear unrelated to non-musicians. Using programs like Finale one can, for instance, with the proverbial "click of the mouse," transpose melodies in disparate keys to a single key; give a uniform tempo and instrumentation to the MIDI playbacks of these melodies; assign a single meter to works that have different meters; and transpose lines from less commonly read clefs into G and F clefs. Because digitized music evidence is unfettered by the coils of physicality and place judges could thoughtfully review and manipulate this evidence in the privacy of their offices, as could jurors in their jury room.
If I were to testify on behalf of the plaintiff in Hein v. Harris I must try to convince the court that the defendant took an economically valuable part of plaintiff's work. I would try to minimize the significance - economic or musical - of the differences between these parts. Because disputed musical works in these cases are never identical I will need to "normalize" their melodies in order to compare them meaningfully. I might transpose both melodies to a common key, program the MIDI files to perform both works at the same tempo, at the same volume and with similar attack and with the same instrumentation. I might drop or add upbeats, change enharmonic spellings and flatten syncopated rhythms. These ministrations accomplished, superimposing the defending melody over the plaintiff's might help indicate how the essential musical expression of the latter work is derived from the former.
Here are the disputed themes from Hein v. Harris in their original form.
Arab Love Song / Woodpecker Original Score
To normalize the defending theme let us transpose it to E-flat and adjust its syncopated rhythm a bit.
After having so tinkered with the melody, when one superimposes it over plaintiff's one sees and hears relatively few dissonances, or even divergences. Moreover, if one sets this slightly doctored melody over the unaltered accompaniment of the plaintiff's theme, one finds a felicitous correspondence between them.
Arab Love Song / Woodpecker Correspondence
Finally, one finds a telling coherence to a new theme derived from every other measure of the disputed themes, set over the plaintiff's accompaniment (and, one notes that none of the prior art melodies that defendant cites are tractable to such contortion).
Even without these bits of arguably contrived evidence supporting the plaintiff's position one must concede a family resemblance between the themes based on their sharing a duple meter, light syncopation, and a strutting opening motive rising by step. To counter inferences of substantial similarity that one might draw from the similar melodic and harmonic scansions of the songs, the defendant will need to emphasize not only the number of differences between the melodies, but also the significance of these differences to the overall sound of performances of the two works.
The defendant could, for instance, simply superimpose his melody over plaintiff's, without making any adjustments to either, to underscore the differences between them.
Arab Love Song / Woodpecker Original Score
One wonders, however, whether the resulting notational gibberish and aural cacophony of the MIDI rendering would have probative value on the question of melodic similarity regardless of the extent to which musically uneducated jurors and judges might be impressed by such testimony.
The question of the defendant's testimony raises again the issue of normalization of the allegedly misappropriated thematic material. One considers certain pitches of a melody to be more important than others based on a variety of factors, e.g. place in measure, key, duration, meter and genre of work. As discussed, in order to make a dramatic showing of musical similarities between "Arab Love Song" and "Woodpecker" (in the form of highly similar notation and audio files) the plaintiff must adjust several features of the defendant's theme. The defendant might argue, therefore, that plaintiff's normalization of defendant’s theme not only stripped away the very elements that made the defendant’s work original and viable as an independently copyrightable work, but also that if one performed similar ministrations on the plaintiff's theme one might reduce it to a mere collection of generic conventions associated with ragtime.
Prior Art Defense
In Hein v. Harris the defendant used an "even if" argument along these lines: even if his song were substantially similar to the plaintiff's, the elements shared between the works were commonplaces that are found in earlier works of this genre. Like the plaintiff who would normalize defendant's melody to underscore its similarities to his, the defendant would similarly need to normalize plaintiff's theme to underscore its similarity to works preceding it. The more information that the defendant strips from the plaintiff's theme to make a plausible case for a prior art claim, however, the greater the inference that the plaintiff's number is in fact original.
The defendant offered four popular songs to support his prior art argument.
This evidence did not impress the court; of the four numbers that the defendant put forth Learned Hand found only "Bon Bon Buddy" to be similar to "Arab Love Song," and thought the similarity too brief to support the defendant's claim of derivation.
To identify the four prior art works, the defendant's music expert relied upon his familiarity with the popular song repertory of his era. One wonders whether his prior art testimony would have been more persuasive if he had had access to a searchable database of the themes or scores of all popular numbers in America from 1880 to 1910. This database does not yet exist, but publishers' and archivists' adoption of digital notation technology points to a time when it will.
Like digital text, digital music notation is tractable to indexing and searching. Using queries of note names, interval sequences, and rough and refined melodic contours, one can search data collections of complete scores or indexes of themes. Digital notation searching technology is young and developers ponder questions of how to massage queries and searchable data so as to improve their elasticity for octave placement and enharmonic spellings, while accommodating more information about rhythm. (Digital music audio file searching offers different, but equally nettlesome problems.) One of the most extensive digital theme indexes to date is Themefinder , a project of Stanford's Center for Computer Assisted Research in the Humanities. Themefinder is a searchable collection of over 35,000 themes of Folksongs, Classical, and Renaissance works.
Melodic profiles are influenced by various factors, including the genre of a work, and whether the melody is modal or tonal, instrumental or vocal. Themefinder’s repertories are inapposite to those encompassing the popular works of music copyright cases. One might then, best explore the potential usefulness of resources like Themefinder in plagiarism disputes using a melody with Classic characteristics.
In 1794, Beethoven's tuition with Franz Josef Haydn ("Papa" Haydn - a curious sobriquet given that he had no children) ended with misgivings on the part of both composers. Let's assume that Beethoven wrote a piano trio in 1796, the third movement of which, opens with the following theme:View Video Clip
Does this look and sound familiar? Haydn might have thought so given the opening of the elegant and playful Menuet from his "London" Symphony, first performed in the city in 1795:
Accused of plagiarism (a not uncommon accusation lobbed among 18th and 19th century composers, although usually without the gruesome legal and economic baggage that attends these charges today) Beethoven might have claimed that both his and Haydn's themes were derived from prior art, as indicated by information obtained from Themefinder . First, as indicated in the example above, both melodies are reduced to strings of searchable information. Searching the Folksong and Classical modules of Themefinder using a query of the first five pitches of Haydn's theme retrieves Haydn's work, a Schubert symphony (not, of course, prior art) and four folksongs; a rough and refined contour query of the first eight directions retrieves thirteen works, but of these only the Folksongs show any kinship with Haydn's melody.
Themefinder Search Results
Something to ponder…
What should one make of these results? Leaving aside the question how closely Haydn's theme maps any one of the folk tunes, is the fact that our search results contain a preponderance of music with no individual attribution (i.e. echt public domain material in the form of folk melodies) useful to either party in this dispute? If you were deciding this case would your decision be affected by the Themefinder results? In your decision, which party should prevail, and to what extent would your ruling be influenced by musical analysis of the works versus the undisputed close personal contact between the litigants?