This reading note introduces Stephen Davies’s central claim in 'On defining music' (2012): although humans can immediately and almost infallibly recognize music, defining it proves remarkably difficult. Davies examines why major definitional strategies fail by analyzing two influential proposals - Jared Levinson’s intention‑based definition and Andrew Kania’s feature‑based account. Levinson’s view is considered by Davies as being too narrow, excluding everyday musical practices, while Kania’s is conceived as too broad, allowing 'Morse code, sirens, computer start‑up tones' to count as music. The test case of Cage’s 4'33" shows how divergent these approaches are, as each philosopher classifies the work differently, revealing the instability of definitional boundaries. Davies argues that four families of definitions (e.g., functional, operational, socio‑historical, and structural) collapse for internal reasons. For Davies, music is too heterogeneous in function, form, and cultural role to fit a single criterion. A hybrid view is proposed: music conceived as patterned sound that needs to be understood as a situated practice within socio-historical traditions, not a closed formal system (similar to mathematics).

--