Abstract This presentation centres around the performance history of the choral work of Nigerian composer and leprosy patient activist Ikoli Harcourt Whyte (1905-1977). Harcourt Whyte wrote a huge number of hymns in the Igbo language, fusing Nigerian and UK Methodist musical modes, for performance by patients and fellow former patients at the Uzuakoli Leprosy Centre in southeastern Nigeria. Here, I develop a conceptual armature for the deployment of a research painting practice, in revisiting documentary and artist collaborations around Harcourt Whyte’s work.

The painting practice in question uses a mix of physical and digital processes to interrogate the role of machine learning algorithms in knowledge production, preservation and storage in relation to artistic performance practices, and in mediating access to, and custody and legibility of archives of practice. I investigate how we can use such practices, deriving from relations around painting, to prospect for ways to insinuate minoritised histories/heritage, community practices, and archives-at-risk into the evolving algorithmic landscape, without reinforcing existing harms or rendering new axes of vulnerability. To evidence this potential, I revisit my own collaborative work with composers, academics, medical workers, and patient activists to recover, reinterpret, perform, and share the work of Harcourt Whyte.