IN the current issue of this journal, Jeremy Orosz evaluates the reliability of entries on the whosampled.com web site, which hosts a user-generated database of musical borrowing between works of popular music (especially rap/hip-hop and R&B) as well as classical music. The online database distinguishes between two types of appropriation: 1) a direct sample, which involves sound from one recording being electronically copied (via analog or digital mechanisms) into a new recording; and 2) an interpolated sample, which involves a re-performance of lyrics or musical material from another work. In this regard, the web site takes the term "sample" in a much broader way than its traditional definition, which normally includes only the direct method (Fulford-Jones, 2001). In either case, the appropriation of the sample into the new work may involve some sort of manipulation, whether through signal processing (as with a direct sample) or recomposition (as with an interpolated sample).

Orosz's primary finding is that the reports of direct samples in the database are generally more accurate than the reports of interpolated samples. This is perhaps not surprising, since a source recording will have a unique sonic signature that is retained through direct sampling, which often makes the source recognizable even under significant manipulation. In contrast, a report of an interpolated sample may simply be due to two songs using similar musical material, without any genuine borrowing from one work to the other. Orosz concludes, though, that entries for interpolated samples are more likely to be correct if the source material is distinctive in some way. He also notes that the reliability of the database loosely correlates to the musical style, with reports of sampling in rap/hip-hop, for example, as generally being more accurate than those in classical music.

Orosz does not support any of his conclusions with statistical measures. Instead, he describes four case studies (on entries for Bruno Mars, Janelle Monáe, Dua Lipa, and classical music), assessing the accuracy of each individual report on a case-by-case basis. Orosz cannot be blamed for not including any quantitative data. His own discussion often centers on just how difficult it is to assess whether a sample—particularly an interpolated sample—is truly borrowed from another work or whether that similarity is merely coincidental. To be clear, Orosz brings both common sense and musical intuition to his analyses, and he is obviously knowledgeable about the history of many popular styles. The issue is simply that music (perhaps especially popular music) is so often composed using a shared grammar and syntax—such as stock chord progressions, limited scalar collections, and characteristic rhythms—that trying to make a binary decision about whether a particular passage of music is "borrowed" or "not borrowed" from an older work can be challenging even for an expert, at least in the absence of any evidence that documents the compositional process or thinking behind the new work.

But while Orosz does as good a job as can be reasonably expected at assessing the reliability of the whosampled.com database, a central question remains unanswered: What is the value or utility of a database like this? To put it another way: What are we supposed to do with the information on this web site? In his closing paragraph, Orosz asserts that the database—despite its flaws—is "nevertheless an invaluable resource for DJs, producers, and scholars alike" (p. 17). But how so? It is doubtful that this database has the same value to a DJ, for example, as it does a scholar (not that those roles are mutually exclusive). What, then, will different users take away from this database, depending on their perspective?

From Orosz's perspective, the reliability of the database may not actually matter all that much in terms of its value. Orosz writes in his closing sentence, for example, that "even the entries that do not convincingly identify a single source are still informative" (p. 17), since every entry—whether truly an example of borrowing or not—has recognized some interesting musical similarity between different musical works. This is the perspective we might expect from a music theorist. After all, music theorists are typically concerned with classifying common structures and patterns within a music style. Cases of similarity are thus valued as scholarly knowledge, whether they involve true musical borrowing or not. In other words, even if an entry incorrectly identifies borrowing, there will nonetheless be presumably at least some remarkable relationship between the two works. This observed relationship may then catalyze a new idea about common musical structures in popular music.

From other perspectives, however, the reliability of the database—which is in part a byproduct of the how the database is structured—is problematic. In the following section, we consider these issues with the whosampled.com database from a perspective other than that of the music theorist.

AN ALTERNATIVE PERSPECTIVE

As mentioned above, Orosz contends that the whosampled.com database is an invaluable resource not just for scholars but also for DJs and producers, i.e., those people who create or aspire to create music that may involve borrowing musical material from a preexisting work. From the perspective of an aspiring content creator, the database seemingly reveals extensive borrowing across many musical styles, especially hip-hop and rap music. Indeed, Justin Williams has argued that "the fundamental element of hip-hop culture and aesthetics is the overt use of preexisting material to new ends" (2014, p. 1). As a result, the database may give the sense (at least to a naive user) that there is nothing problematic about borrowing material from preexisting works; rather, it is par for the course. We are reminded here of the old aphorism that "good artists borrow; great artists steal," 2 which implies that more successful creators take materials without permission.

Unfortunately, this attitude runs afoul of the law, at least in the United States. To better understand this conflict, it is worth briefly summarizing U.S. copyright law as it currently stands. Generally speaking, the United States Copyright Act of 1976 (Title 17, U.S.C.) lays out two relevant provisions. The first provision pertains to whether or not a work is copyrightable at all. The second provision pertains to whether using material from a copyrighted work requires consent from the owner of the copyrighted work or whether the use of this material can be considered "fair use" and thus escape liability.

As for the first provision, the issue of whether something is sufficiently original to be copyrighted was addressed at length in Atari Games Corp. v. Oman (1992), the opinion written by then U.S. Court of Appeals D.C. Circuit Judge Ruth Bader Ginsberg. Ginsberg relied heavily upon the Supreme Court's then recently settled "test" for originality in copyright as a de minimus standard (in deciding whether geometric shapes were copyrightable within the design of a videogame). Generally speaking, the barrier to copyright with regards to the "originality" of the work is low, as one might confirm considering the vast but shallow lake of copyright-protected pop songs that are extremely similar in many ways yet have made their way into the marketplace as individual copyrights. In contrast is the test for a patent, for instance, which is instead a deep but narrow well requiring the design be novel and non-obvious. A patent requires a high bar of proof, with the inventor needing to show that their creation is the first of its kind (novel) and that its design cannot be deduced without its inner workings revealed (non-obvious). A copyright is thus easy to claim at the outset, comparatively speaking. The copyright vests as soon as it is fixed in a tangible medium, with registration only required if in federal copyright litigation (albeit highly recommended as notice of ownership and as a deterrent to infringers).

The financial implications of copyright law arise when one party claims that another party's work is infringing on theirs. In this situation, the U. S. Copyright Act's provisions (supported by case law) on derivative works and fair use come into play. "Derivative works" are defined in USCA Section 101 as "a work based upon one or more preexisting works, such as a translation, musical arrangement, dramatization, fictionalization, motion picture version, sound recording, art reproduction, abridgement, condensation, or any other form in which a work may be recast, transformed or adapted." While derivative works can themselves by copyrighted—and the creators of these derivative works can profit from their creations—USCA Sect 103(a) states that no copyright can be claimed for any part of a derivative or collective work that has used the pre-existing material unlawfully. In other words, the creator of a new work cannot profit from the new work if that new work uses material from a copyrighted work without the copyright owner's consent, since the new work is barred from copyright protection due to the unlawful use.

The "test" for whether something is a derivative work is laid out well in Pickett v. Prince (U.S. Court of Appeals, 7th Circuit (2000)). In this case, a guitar maker (Pickett) made a guitar in the shape of the "Artist formerly known as Prince" hieroglyphic. The guitar maker then sued Prince for copyright infringement when Prince started performing with his own guitar in the same shape. Prince then countersued, claiming that Pickett's use of the hieroglyph infringed on Prince's copyright. Ultimately, the court decided that since Pickett did not secure permission from Prince to use the hieroglyph in the guitar design, Pickett could not claim that his guitar was a derivative work and thus could not copyright his design. In summary, if a preexisting copyright is pervasive throughout the entirety of a new work (as it was with the guitar), it is not a protectable "derivative work" under the law and thus cannot be copyrighted by the new creator. If, however, there is consent by the underlying owner and the use of the pre-existing work is distinct and distinguishable, then it may indeed be a derivative work which can be claimed as a new copyright by the new creator, to hold for the duration of a copyright, i.e., the life of the author plus 70 years.

That said, U.S. copyright law does allow for a narrow, special exemption from the requirement to gain consent. This exemption is known as "fair use," and it entails four criteria for using copyrighted material without permission: (1) the purpose of the use, i.e., whether it was commercial or educational, (2) the nature of the copyrighted work, i.e., whether the original work was more creative (such as a song) or factual (such as scientific data), (3) the portion or amount of material used, and (4) the effect of the new work on the market for the original work. Each of these four criteria is assessed one-by-one, and the standard legal threshold is that three out of four of these criteria must be met to be considered "fair use." As a result, fair use is often a delicate and subjective claim, with the outcome of legal challenges never predictable and often contingent on a particular judge in a particular court. In other words, fair use is generally not a dependable argument for failing to get the consent otherwise required to use someone else's copyrighted work. And even when a fair use argument is ultimately successful, it may require years of (usually expensive) legal challenges in court (as, for example, in the case of Bill Graham Archives v. Dorling Kindersley, Ltd.).

In light of U. S. copyright law, both statutory and case precedent, using material (a "sample") from a copyrighted song in a new song should ideally be done with the consent of the copyright owner. Normally, a license is procured that is highly limited in scope, with its licensors being the owner of the underlying composition and the master recording, the licensee being the owner of the new sound recording. The use of the sample will be restricted to the precise use and placement in the new work, defined in the license agreement, for a limited duration (running time), for a license fee to be negotiated. The license is usually a grant for the life of the copyright of the new work, restricted to this onetime embodiment of a particular number of seconds of the pre-existing work in the derivative work. To be clear, there are no shortcuts with the law on the issue of infringement; if someone samples or copies without permission and incorporates this material into a new work, that person is infringing and thus cannot copyright or profit from the new work. The only exemption is fair use, which is rarely a successful argument, limited to special and narrow cases such as a clear parody or obviously critical commentary.

With these legal issues in mind, it is worth noting that the whosampled.com database contains no information on which samples were properly licensed and which were not. For those samples that were not properly licensed, the database also does not indicate what financial or legal consequences did or did not occur as a result of appropriating the unlicensed musical material. The Verve song "Bitter Sweet Symphony" (1997), for example, is listed on whosampled.com as containing an "Interpolated (Replayed Sample) of Multiple Elements" from the Rolling Stones song "The Last Time" as performed by the Andrew Oldham Orchestra (1965). This entry gives no indication, however, that the Verve did initially secure a license to sample the song but were then sued—and lost—because they used more material than was originally licensed. As a result, the Verve had to forfeit all royalties from the song to the Rolling Stones until 2019, when Mick Jagger and Keith Richards of the Rolling Stones volunteered to sign publishing rights for the song back to Richard Ashcroft of the Verve. 3

Scrolling down in the "Comments" section of the entry for "Bitter Sweet Symphony," we find users discussing the lawsuit and some of these issues, both musical and legal. But this is a relatively famous case. Many if not most entries on the whosampled.com web site have no comments. We thus wonder: Which samples on this web site were properly licensed, and which were not? And of those samples that were not properly licensed, which resulted in a legal settlement between the copyright holder and the borrower? In some cases, an album will acknowledge a sample in its liner notes, by which we presume the sample has been properly cleared (as Orosz notes on p. 2). But this type of information is not currently incorporated into the web site in any official or organized way.

These questions about licensing would admittedly be difficult for an online database to answer in a comprehensive way (especially given that some settlement agreements may have been privately negotiated). But without any attempt to address these issues, the web site gives no sense of the legal ramifications of sampling. Yet these legal ramifications are (or at least should be) a central consideration when sampling another work, specifically when sampling a work under copyright, as most entries on the site involve.

Orosz does not consider these issues, since his perspective is not that of a content creator. Yet we would estimate that many of the students Orosz teaches aspire to be content creators (if they are not content creators already). The School of Music at the University of Memphis (where Orosz currently teaches), for example, has a Bachelor's degree program in Commercial Music Performance. The current list of required courses for these students includes many semesters of coursework on music theory, ear training, music history, arranging, and performance, as we would expect for any music program. 4 Additional coursework specific to commercial music relates to the music industry (MUID 2201), music technology (MUID 3905), and entrepreneurship (MUID 4840). But despite the department offering courses on the legal aspects of the music and entertainment industries (MUID 4301 and MUID 4604), the Commercial Music Performance program does not currently require students to take any dedicated courses that pertain to copyright law. Since sampling appears to be such a central component of hip-hop and R&B, though, and since R&B/hip-hop now enjoys the largest proportion of listenership in the United States of any musical genre, 5 it would seem that coursework aimed at helping students navigate issues of intellectual property would be a critical (and thus required) component of a modern musical curriculum.

Moreover, many if not most music departments in the United States do not offer commercial music programs or offer any courses dedicated to copyright law. We thus estimate that the typical music student does not have even a basic education on issues of intellectual property by the time they graduate. This is problematic, given how much intertextuality and borrowing has been and continues to be part of music making. As web sites like whosampled.com proliferate and expand their databases, more and more connections between songs will presumably be identified, whether real or imagined. How can a modern music creator ensure that what they thought was simply a stock pattern would not be construed as copyright infringement? Orosz, for example, refers (on p. 5) to the pending case against Ed Sheeran for his song "Thinking Out Loud" (2014), which allegedly copies (among other elements) the chord progression (I–iii–IV–IV or I–I⁶–IV–V) from Marvin Gaye's "Let's Get It On" (1973). Yet this same chord progression (or bass line) can be found in many songs, including "Make Your Own Kind of Music" by the Mamas and the Papas (1969), "Man in the Mirror" by Michael Jackson (1988), "Right Now" by Van Halen (1991), and "Dreamin'" by Weezer (2008).

Some chord progressions are so common that they are typically considered to not be copyrightable, such as the 12-bar blues or the so-called "Axis" progression, I–V–vi–IV (Richards, 2017). Perhaps the chord progression shared by "Thinking Out Loud" and "Let's Get It On" falls into this category. But these two songs share more than just a chord progression (or bass line). They also share a similar "feel," including the tempo and the eighth-note anticipations on every other chord. The issue of one song appropriating the "feel" of another cannot be dismissed, as it is essentially the winning argument made in 2015 by the family of Marvin Gaye against the songwriters of "Blurred Lines" (specifically, Robin Thicke and Pharrell Williams), a multi-million dollar verdict that has withstood appeal at the district court level.

What does it mean, therefore, that one artist can sue another for millions of dollars of damages for copyright infringement based on something so blurred as the "feel" of a song? It means that songwriters and music creators need to take particular care when they are producing music and be especially knowledgeable about the laws surrounding intellectual property. Historically, music departments have been largely populated by students who are interested in careers in classical performance or music education (i.e., secondary school teaching). But as more departments of music seek to attract a wider body of students, these departments will likely find that their new students are more interested in fields that intersect with music creation and production. 6 If so, music departments owe it to these students to train them in how to navigate issues of intellectual property.

So, while an emerging field in music theory has been the study of intertextuality in popular music, 7 these studies mostly ignore the types of questions that would be central to a content creator and producer. We do not mean to imply that it would necessarily be the role of the musicologist to answer these questions, per se, although musicologists are increasingly becoming the central figures in legal disputes, employed as expert witnesses on both sides. Yet the musicological work on intertextuality in popular music—specifically with how this intertextuality intersects with U.S. copyright law—is fairly thin. (The work by Burkholder (2018) offers a good starting framework, albeit in the context of classical music.) While we thus agree with Orosz that a web site like whosampled.com will be an invaluable resource for those interested in discovering connections between musical works, the implications of these connections—and how to navigate these potential implications as an artist on either end of an allegation of copyright infringement—is a huge area that unfortunately remains mostly unaddressed within the academic music community. Hopefully, music theorists and other musicologists will work to help bring these blurred lines into better focus soon.