This article is a guest post about ways to answer the challenges presented by COVID-19 to the making and use of editions in scholarship, peformance, and teaching. To submit an idea for this blog category, contact

By Loren Ludwig

Donald Rumsfeld once infamously opined that it was the “unknown unknowns”—the stuff you don’t know that you don’t know—that make foreign policy so difficult. A similar problem confronts those fascinated by lacunae in Renaissance and baroque polyphonic music—compositions for which one or more polyphonic parts have been lost to the ravages of history. To those seeking to reconstruct missing parts, and thereby render incomplete pieces playable, a primary challenge is figuring out what, exactly, is unknown. Much can be inferred from surviving voices, particularly if the primary structural voices (often including the cantus and tenor) survive.[1]

A particularly famous case of missing parts in Renaissance music is Carlo Gesualdo’s first book of Sacrae cantiones, a sei voci (1603), which has come down to us missing the sextus and bassus partbooks. Here, many aspects of the missing voices can be recovered using the contrapuntal language of the period as it appears in the surviving voices, marked as they are by Gesualdo’s distinctive chromatic style. A less certain—and perhaps more challenging—lacuna appears in British Library Additional MSS 18936–9, four surviving partbooks out of an original set of five. Add. 18936–9 holds the only known copy of Robert Stevenson’s (fl. 1570–1600) seven instrumental Misereres, which are labeled “6 voc:” in each of the four surviving partbooks. Not only is the cantus—an important structural voice—missing, but those interested in attempting a reconstruction are also left with uncertainty about how many voices are lost. Are Stevenson’s Misereres really six-part pieces that were copied into five partbooks (a familiar practice)? Or did some scribe make a mistake at some point and title the pieces incorrectly? Unknown unknowns . . .

Both the Gesualdo and Stevenson examples appear in online initiatives inviting musicians to reconstruct missing polyphonic voices—an activity that seems perfect for the legions of performers and music scholars now sheltering in place indoors with no access to physical library collections. The Gesualdo Project, run by the Centre d’études supérieures de la Renaissance of the University of Tours, invites scholars around the world to participate in a crowdsourced editing project that includes all of the composer’s known works. While the Gesualdo Project went live in 2017, Stevenson’s Misereres can be found as part of a COVID-era project (“MiserereFest”) intended to “help creatively pass the time as we shelter in place from the virus.”

Prominent recent projects aimed at reconstructing (and often subsequently performing) missing voices include the Boston vocal ensemble Blue Heron’s The Lost Music of Canterbury, a five-CD set completed between 2010 and 2017 of much of the incomplete music in the Peterhouse Partbooks. Blue Heron partnered with musicologist Nick Sandon, who originally reconstructed the missing voices as part of his doctoral dissertation at the University of Exeter, but recent projects are more apt to rely on a crowdsourced approach to the laborious process of reconstructing idiomatic polyphonic parts. These current initiatives include the Lost Voices Project, dedicated to the sprawling oeuvre of more than 400 chansons published in Paris by Nicolas Du Chemin during the middle years of the sixteenth century (the contratenor and bassus voices of more than 80 chansons have not survived), as well as the many incomplete editions discussed by scholar and singer Marty Morell on his website, the Italian Madrigal Resource Center. Both of these online initiatives, like the Gesualdo Project and MiserereFest, provide resources and guidance and invite visitors to take a crack at reconstructing “lost voices.”

Idiomatic reconstruction of lost polyphonic parts requires access to manuscripts or scans of manuscripts, as well as knowledge of counterpoint and familiarity with a given composer’s specific style. Recent digitization efforts by libraries around the world have greatly (though not totally) lowered both these barriers by making many manuscripts of Renaissance and baroque music available, as well as primary and secondary sources that teach idiomatic counterpoint. With the COVID-related cancelation of public musical performances and in-person college and university music classes, professional and amateur musicians and scholars have found themselves at home with time on their hands. Musicologist and viola da gamba player Sarah Mead, for example, says that MiserereFest, “has been a great occupation during these last few weeks of ‘captivity’ and I’m grateful for the diversion.”

One key question facing anyone attempting a reconstruction is whether they seek to create an idiomatic solution or the solution—in other words, to reconstruct the exact part that has been lost (it’s worth remembering that, in nearly all cases, the rules of counterpoint allow for many correct solutions to any particular reconstructed part). The perpetual interest over the last century in J. S. Bach’s incomplete Die Kunst der Fuge presents a case of the former—musicians and scholars of all stripes have published dozens of completions of the incomplete contrapunctus 14, many purporting to be motivated by an attempt to recover the solution. Generations of musicians have used numerology (gematria), paleography, codicology, and a dizzying variety of musico-theoretical approaches to attempt to reconstruct Bach’s own exact solution to the final section of the final fugue.

Inspired in part by such attempts, recent research using statistics and information theory uses various methods of data analysis to determine how closely a given reconstruction imitates the historical composer’s favored compositional gambits. This new suite of approaches, of course, offers the potential of evaluating past and future reconstructions in such a way as to elide the difference between the search for an idiomatic reconstruction and the exact polyphonic voice lost to history (see, for example, the forthcoming Approaches to the Computational Analysis of Classical Music: Methodology and Applications, edited by Pablo Padilla and Francis Knights).[2]

While the COVID era may present a perfect opportunity for many musicians and scholars to while away their confinement reconstructing lost voices, the absence of a comprehensive list of incomplete compositions from the Renaissance and baroque eras complicates this work. While reference sources listing known printed and manuscript sources of historical repertory often note missing partbooks and/or voices, these resources are increasingly hard to access in hard copy and rarely feature an efficient way to gather and assess references to incomplete works. What is needed is a widely accessible resource that compiles known incomplete historical compositions. Such a resource, whether in published form or presented as an online wiki to which users can contribute known lacunae, would encourage the growing trend of crowdsourced reconstruction and contribute an important “known” to Rumsfeld’s list of unknowns.

Loren LudwigLoren Ludwig is a scholar/performer based in Baltimore, Maryland. He researches what he describes as “polyphonic intimacy,” the idea that music in the Western tradition is constructed to foster social relationships among its performers and listeners. Current research/performance projects include the use of the viola da gamba in eighteenth-century British colonial America and the reconstruction of a lost tradition of Early Republic New England string ensemble playing. Loren is a co-founder of LeStrange Viols and Science Ficta and performs with ACRONYM, Ruckus, and numerous ensembles in the US and abroad. He also serves as program coordinator for the program in the Arts, Humanities, & Health at the Johns Hopkins School of Medicine.

[1] Additional comment from A-R Editions: Musicologists have been reconstructing missing parts for as long as they have been editing the music of the past. Such reconstructions feature in several of A-R’s Recent Researches in Music volumes. Many of Luzzasco Luzzaschi’s printed madrigal collections have come down to us missing one or more parts—in some cases almost all, as in his Sesto libro de’ madrigali a cinque voci (1596, for which we have only a quinto partbook). Anthony Newcomb’s editions of Luzzaschi’s complete madrigals (R136, R139, R150, and R156) attempt editorial reconstructions of some pieces with missing parts—see, especially, “Solea lontana in sonno consolarme,” no. 3 in his Primo libro di madrigali, a cinque voci (1571; R156), which shares a text with a madrigal by Cipriano de Rore—while others can be completed from later anthology prints and manuscript copies, and others from a combination of both (like “Non sono, oimè, questi miei luci degne,” no. 4 in the Sesto libro; R136). Other volumes have featured reconstructions of entire missing pieces or sections in addition to missing parts—a notable example being Ross Duffin’s reconstruction of the four-voice passion setting of early sixteenth-century English composer Richard Davy preserved in the Eton Choirbook (Y2-017). In this volume, Duffin not only reconstructs the missing polyphony of Davy’s passion but also supplies the interspersed plainchant passages for the Evangelist and Jesus from contemporary liturgical sources.

[2] An article by Francis Knights, Mateo Tonatiuh Rodríguez, and Pablo Padilla, using computational methods to assess newly composed voices and titled “Reconstructing Renaissance Polyphony,” will appear in the autumn 2020 issue of the National Early Music Association Newsletter. See for details.