The growing number of performances and recordings of Feldman’s music in recent years attests to increased interest among performers and listeners; yet his music remains an uncommon subject for detailed music analysis. Proceeding on the premise that this disparity is no accident, I argue that certain distinctive qualities of the music render it difficult to analyse with tools, methods, and practices developed in response to other repertories. This paper investigates the analytical challenges posed by Feldman’s music. A survey of such challenges as they relate to his output in general is followed by an account of two particular issues associated with his late work: scale and repetition. Two case studies address these issues in turn, advancing relevant conceptual and methodological approaches. In the first study, on Coptic Light for orchestra (1985), I suggest that analysts might reconsider part–whole relationships in music analysis, and use the idea of ‘populations’ (with their attendant features of range of variation and distribution) to develop a non-reductive (and non-constructive) approach to scale. In the second study, on Piano, Violin, Viola, Cello (1987), I encourage analysts to rethink the role of repetition in music analysis, such that repetition is no longer (only) a goal, but becomes a point of departure. Throughout the essay I take the view that analysis is an investigation of experience; that a particular difficulty of analysing Feldman’s music is the self-knowledge it requires; and that the concerted inquiry that is music analysis can well be used to expand – not only condense – the realm of musical experience.