James Madison Carpenter (1888-1984) was a Harvard-trained scholar who recorded traditional singing in Britain and, to a lesser extent, in his native United States, in the period c.1928-40. His extensive collection includes 179 Dictaphone cylinders, totalling some 35 hours of examples (https://hdl.loc.gov/loc.afc/eadafc.af010002). Although he worked on the collection throughout his life, Carpenter’s hopes for publication were never realized and he eventually sold it to the Library of Congress. It has since been digitized and is now available online via the Vaughan Williams Memorial Library (https://www.vwml.org/archives-catalogue/JMC), facilitated though the work of an ongoing project by a team of UK- and US-based scholars towards surfacing the collection in a critical edition (https://www.abdn.ac.uk/elphinstone/carpenter/).
Carpenter was the first to consistently use a recording device in British folk song collecting. The cylinders thus have the potential to provide insights into traditional performance style and the comparative study of folk song melodies, as well as providing a relatively untapped source of repertoire for contemporary performers.
The quality of the sound recordings is, however, disappointingly poor, making Carpenter’s own transcriptions an essential complement. A self-trained musician, he found the process laborious, taking inspiration from the dictum that ‘a wrestler with sounds is a wrestler with shadows’.
This paper focuses on Carpenter’s approach to recording and evaluates his transcriptions, as well as outlining the issues involved in producing new ones. As we continue to wrestle with these historic sounds today, what is their potential for scholarship and performance, and what is the role of transcription in this context?