Review: International Music Score Library Project (IMSLP)

Six pièces faciles, op. 13 (published in 1925), is a set of solo guitar pieces by Antonio Alba (see Figure 1).1 I discovered this score thanks to the “random page” navigation feature on the International Music Score Library Project (IMSLP) home page. Alba was new to me. Through a quick Internet search, I learned that he was a Catalan guitarist, composer, and music teacher who lived in Chile from 1873 to 1940. He composed more than 400 pieces, 250 of which were printed by commercial publishers in Valparaíso, Barcelona, Madrid, and Paris.2 Works for guitar make up the bulk of Alba's output. I clicked on his IMSLP artist profile and saw twenty-two of his compositions available to download. For each work on IMSLP, I could download a scan of the first printed edition, a nonprofessionally engraved modern edition, and tablature.IMSLP is a digital archive of public domain music scores available for free download. The site was founded in 2006 by Edward Guo, then a composition student at the New England Conservatory. Thirteen years later, IMSLP has grown from an obscure experiment by a nineteen-year-old undergraduate to a major resource accessed by more than 3.4 million unique monthly visitors from around the globe.3 Comprising more than 480,000 scores, parts, and arrangements by roughly 17,700 composers, IMSLP has become an essential Internet resource for classically trained musicians (see Figure 2).4 The website is a nonprofit entity managed by Project Petrucci, a limited liability company with twenty-five full-time employees. It supports its operations by selling advertising space and soliciting donations from individuals and foundations. Like Spotify, it has a premium (paid) subscription option with fewer advertisements and faster file downloads. Most of the scores on the site are uploaded by volunteer users, and these volunteers also scour the files in the archive to identify copyright issues.5 In addition to music scores, IMSLP has a partnership with Naxos Music Library that provides streaming access to thousands of recordings for premium (paid) subscribers.6Given its operating model, it is not surprising that IMSLP has raised serious intellectual property concerns. The site faced high-profile lawsuits in its early days and even shut down for several months in 2007 on account of copyright infringement issues.7 When I interviewed him in 2018, Guo stated that the site is governed by Canadian copyright law because that is where its servers are located. IMSLP also maintains smaller servers in the United States and Europe to host scores that are still under copyright in Canada but in the public domain in other countries. Notwithstanding IMSLP's efforts to comply with Canadian copyright law, the site places the onus on individual uploaders and downloaders to ensure that they are not violating copyright in their own countries.8As Napster and Spotify have done for audio, the archive has undoubtedly transformed the distribution of musical scores. Visitors to IMSLP can access an array of printed music from their own computers at no cost. In this sense, the site's labor and business model reflects broader trends in the twenty-first-century cultural landscape toward a data- and advertising-driven framework in which consumers no longer pay directly for goods and services. But these transformational aspects should not obscure deeper continuities. IMSLP's representation of Western classical music in digital format reinforces many traditional aspects of this repertoire. In this respect, the site's full disruptive potential remains unrealized, despite its democratizing aspirations.Like other cost-cutting disruptors, IMSLP depends on cheap labor. The website is kept running by a legion of volunteers. The level of commitment varies widely, from providing a rating for a downloaded file to more active involvement such as uploading scans, posting on discussion boards, or even contributing original arrangements and compositions to the archive.9 A contributor known as “Marieh Marieh,” for example, uploaded all the tablature editions of Antonio Alba's works on the site. Marieh's IMSLP profile lists 7 original compositions, 66 arrangements, and 263 scanned scores (see Figure 3).10 Marieh's compositions encompass a variety of genres, including “Classical,” “Jazz,” “Modern Jazz,” and “Traditional Folk.” The instrumentation is also diverse: solo and combination guitar, flute, saxophone, and drum set. Marieh does not have a biography or a photograph on IMSLP, just a link to a Musescore profile with even more compositions.11 With so many volunteer contributors like Marieh, IMSLP has amassed hundreds of thousands of scores at no cost. The site's twenty-five paid employees perform many of the same tasks as volunteers—patrolling the files for copyright infringement and inaccuracy, uploading new scores, and translating song lyrics and descriptive text into other languages. Advertising revenue and donations pay their salaries, but there is not enough funding to compensate every user's work on the site.12Still, IMSLP's business model resembles the traditional concept of music-as-commodity upon which commercial publishing was built. Volunteers upload scores because they choose to and know that others will benefit. They measure the value of their own labor by the popularity of the uploaded files, not by monetary compensation. A visitor to the site has many reasons for downloading a score, but for the uploader the fulfillment of the exchange comes from high ratings and growing download counts. Meanwhile, visitors also help IMSLP fulfill a more traditionally recognizable exchange: advertising revenue for user impressions. Thus, IMSLP depends on the all too common ecology of the culture industries, whereby music, art, and information are collectively produced but selectively compensated.13In addition to the work by volunteers, IMSLP depends on the work of thousands of long-dead collaborators. Though most (if not all) of the original composers, arrangers, and editors of the public domain works available on the site are dead, their work lives on and interacts productively with current users, who upload, download, reengrave, and even compose new orchestrations for or arrangements of the works hosted by the site. This integration of “dead” and “living” labor is an example of what Jason Stanyek and Benjamin Piekut call “intermundane collaboration,” which they define as effective colaboring between the living and the dead.14 IMSLP's living users work closely with partners who cannot respond, and since these works are no longer under copyright, even their estates and descendants have no involvement.To be sure, some contemporary composers (or their estates) whose works are not in the public domain, such as Frederic Rzewski and Leo Ornstein, have granted explicit permission to IMSLP to include their works on the site.15 But many contemporary composers resist the inclusion of their works on IMSLP for fear of losing the prestige of a commercial publishing contract or royalty income.16 While Guo acknowledges that these concerns are valid, he justifies the archive's approach by noting that most contemporary composers earn little from published scores and rely instead on commissioning fees and performances.17Moreover, even traditional music publishers are experimenting with new, rent-based models that may one day replace the traditional buying and selling of scores. Apps developed by Bärenreiter, Henle, and other publishers offer access to scores for a fixed amount of time in an attempt to replicate the way Spotify, Apple Music, and other streaming platforms have transformed traditional models of musical exchange while still paying lip service to the moral and economic rights of artists.18 By contrast, IMSLP neither sells nor rents access to musical scores, and in this respect its philosophy is similar to that of the MP3-file-sharing sites that flourished in the early 2000s, such as Napster. Napster users argued that recorded music in MP3 format should not be treated as private property because it is nonrival and immaterial.19 The advent of IMSLP demonstrates how this manner of thinking has converted published music from a rival to a nonrival good, at least if it has entered the public domain.IMSLP's practice of giving music away for free suggests a radical break with the music industry by cutting out traditional middlemen. The website's guiding philosophy is that “music should be something that is easily accessible for everyone,” and the site proclaims that “[t]he ultimate goal of IMSLP is to gather all public domain music scores, in addition to music scores of all contemporary composers (or their estates) who wish to release them to the public free of charge.”20 The sheer number of scores available for download on the site (a number increasing by the hour) suggests that IMSLP could perhaps one day fulfill its goal of making all known notated music available to everyone. But the scope of repertoire available in the archive is limited to music that is in the public domain, notated, and selected by its volunteer contributors. And despite the breadth of musical offerings, most visitors to IMSLP do not seek lesser-known and obscure repertoire. The site's most downloaded score is Beethoven's String Quartet no. 13, op. 130, and the most popular composers, according to download statistics, are Beethoven, Brahms, and Mozart (see Figure 4).21The dominance of certain types of music in the archive and the relative absence of others reflect both internal and external constraints that prevent IMSLP from amassing a truly comprehensive collection of the world's music. Besides user preference, Western classical music is one of the only musics in the world to rely largely on written scores for dissemination, and the robust history of commercial music printing means that public domain copies of Western classical scores are relatively easy for users operating in the United States and Europe to acquire and scan. Traditional African music, for example, which is not usually notated and often composed collectively, is more challenging to include on IMSLP and therefore almost entirely absent from the archive's offerings.22 While IMSLP's philosophy implies universality, in practice “making music available to everyone” actually means “making Western, notated music available to everyone with a high-speed Internet connection.”Yet the predominance of “great” Western classical composers in the archive is complemented, to some degree, by the presence of works by more obscure composers from all over the world, such as Antonio Alba and Marieh Marieh. What is more, the large volume of arrangements and original compositions by anonymous (often amateur) individual uploaders like Marieh demonstrates that many of the site's most active contributors view IMSLP as an outlet for sharing their work with a wider audience, rather than merely a repository for the greatest hits of the classical canon. Even for the well-known works, there are some surprising adaptations. These include Beethoven's Piano Sonata no. 14, op. 27, the “Moonlight” Sonata, arranged for five recorders; Bach's Prelude and Fugue in C Minor BWV 847 adapted for clarinet, trombone, bass guitar, drum set, and piano, and retitled “Bach Was a Cool Dude” (see Figure 5); and Mozart's Eine kleine Nachtmusik, K. 525, arranged for four saxophones (see Figure 6). While the majority of public domain scores currently date from the nineteenth century or earlier, copyrights on twentieth-century compositions are quickly expiring and these works will likely appear on IMSLP at some point.23 To be sure, anonymous Bach and Beethoven arrangements seem a far cry from twentieth-century avant-gardist compositions. Nonetheless, the tongue-in-cheek nature of these unconventional adaptations recalls the irreverent attitudes that generated The Rite of Spring and “total serialism,” opening the way to further innovation.24Despite the presence of these delightfully unconventional arrangements and obscure compositions, most individuals who use the site may never know that they exist. The majority of the site's web traffic involves downloading a specific score or browsing works by certain composers.25 IMSLP even suppresses noncanonic versions of well-known works: site guidelines instruct volunteers to patrol the files and flag unusual arrangements for deletion.26 Though IMSLP offers tools—such as the “random page” feature—that expose visitors to some of the obscure offerings that fall outside the narrow constraints of the traditional canon, the number of scores in the archive is overwhelming. Ultimately, most downloaders use IMSLP to find music they already know, rather than to discover something new or unusual.A wiki-based website, IMSLP uses open-source code. Its home page and interface project have a scrappy, do-it-yourself air (see Figure 7). So much music seems to be only a few clicks away, but orienting oneself to the intricacies of the site can be challenging for novices. The easiest way to browse scores in the archive is to search by composer, keyword, or work title using a Google-powered search function on the home page. Because the information for the scores is user-generated, however, there are sometimes gaps. Depending on what interests the uploader, there could be a plethora of detail about the composer's biography yet little to no information about instrumentation, difficulty level, or genre. These discrepancies can render the site's offerings somewhat opaque to users without prior knowledge. Admittedly, the learning curve to fully benefitting from IMSLP's corpus is perhaps not that different from that required to navigate a physical collection of music in a college or university library. Despite its aspirations to universal access, the online archive assumes that its users already have the requisite knowledge, involuntarily replicating common barriers to entry found in other, more conventional classical music resources.And even when the site does promote noncanonic composers, its limitations quickly become apparent. At first glance, the archive appears to contain works by composers from a diverse array of countries and ethnic backgrounds, with ninety distinct nationalities represented (see Figure 8).27 But categorizing composers by national identity is hardly a straightforward affair. For example, though there are twenty names listed under the subcategory of “Indian people,” a cross-check with Grove Music Online and Wikipedia confirms that many of the “Indian” composers are actually people of British descent who inhabited colonial India in the eighteenth and nineteenth centuries.28 It is not part of IMSLP's mission to historicize the composers or works contained in its archive, and it would be unfair to expect a site run primarily by volunteers to spend the resources required to educate users on its platform to this kind of nuance. This example merely underscores how unstated assumptions can involuntarily guide and shape users' understanding of music history itself, even on a website that offers a vision of democratic transformation.The digital format of the scores on IMSLP marks a paradigm shift in the value and role of notated music for performers, composers, and audiences everywhere. Whereas previously consumers purchased authoritative, commercially produced versions of scores or even exhaustive “critical editions”—often edited by musicologists—from music publishers, many of IMSLP's users choose to bypass these expert publications. Instead, they turn to public domain editions, many of which were published more than a hundred years ago. Much like Dover Publications' “Thrift Editions” (perhaps a predigital analogue to IMSLP's digital scores and parts), they make up for their lack of accuracy with low cost, ease of access, and convenience.But some users turn to IMSLP for more than just free, near-instant access to scores and parts. One user maintains a list of popular and newly added scans of first editions and manuscripts.29 This list is prominently featured on IMSLP's home page and updated regularly (see Figure 7). From Beethoven's late string quartets to sixteenth-century madrigals, the scans of these “original” scores are in high demand among certain musicians. The Borromeo Quartet's members often play from files downloaded from IMSLP and extol the virtues of practicing and performing from full score and manuscript—albeit a digital copy displayed on an iPad.30 Borromeo's first violinist, Nicholas Kitchen, has said that he loves playing from manuscripts because “I can tell you that Beethoven's personal interaction with the page results in a completely different system than what we see in printed music.”31 In their attempts to understand the composers' intentions and uncover, as Borromeo violist Mai Motobuchi says, “exactly the direct picture [the composer] had in their mind,” these musicians turn to full scores and scanned copies of handwritten pages from the nineteenth century.32 Yet while the iPad makes the process easier, these modern musicians' belief that the autograph offers access to the Werktreue is not new.33 Deirdre Loughridge's work on the creation and circulation of lithograph facsimiles before photography demonstrates that amateur musicians and collectors have long believed that the original autograph can offer insight into the composer's inner being.34In sum, IMSLP provides more widespread access to printed music by making files available for free download to anyone with an Internet connection. Yet if it augurs a quantitative revolution by cutting out publishers and thus increasing the supply of scores, it has done less to qualitatively change the music industry than we would suspect. The ways musicians use the site are of a type with practices that have been deeply ingrained in the Western classical canon for centuries. Still, the site is only in its second decade. Music by composers such as Florence Price and Clara Wieck Schumann, not to mention Antonio Alba and Marieh Marieh, exists in the site's archive alongside more familiar works by Bach, Beethoven, and Brahms. It is not hard to envision a world in which IMSLP becomes more than just a means for users to download the old favorites over and over again.

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.