Music is ancient, pancultural, and, given the spontaneous emergence of song in children, virtually universal. Moreover, we can immediately and recognize it, even where it comes from a culture that is foreign to us. Though I may be unable to predict how such will continue or to recognize errors, and though it may sound strange to me, I can be in no doubt that it is I am hearing.I say almost infallibly because there are marginal cases and a few possibilities for error. Some musical-sounding things are not music: infantdirected speech,1 tone languages,2 and sing-song linguistic accents, as in Welsh. In addition, some patterned might be mistaken for music: these include sound art,3 that are not humanly made, such as the nightly rice field frog chorus, and that are not primarily intended to have the aural character they have, such as incidental auditory effects in the factory making crystal glasses.In this paper I reflect on the enterprise of defining music. Although I critically review some proposed definitions in order to show the difficulties inherent in the task, I do not offer a counterdefinition. Instead, I consider four strategies that might be adopted by the would-be definer. I argue that none of these alone can succeed, though, in combination, they might.1. Some Definitions of MusicCompared to art, philosophers have not shown much interest in defining music. In fact, definitions of are rarely attempted. I am not sure why this is the case. Perhaps it is because, as was observed earlier, we are usually highly successful in identifying as such, and don't feel the need of a definition. And the task of producing a watertight definition of such a wide-ranging phenomenon as is not only thankless but also challenging. To indicate the obstacles that stand in the way and to consider strategies for negotiating them, I now briefly discuss proposals advanced by Jerrold Levinson and Andrew Kania.Levinson, whose goal is to define as an art, proposes that is temporally organized by a person for the purpose of enriching or intensifying experience through active engagement (e.g., listening, dancing, performing) with the regarded primarily, or in significant measure, as sounds (1990, 273). He is happy to exclude Muzak on the grounds that it is not intended for attention and birdsong because it is not humanly created or artistically motivated, though he allows birdsong to qualify under a secondary, legitimate use of music where it refers simply to a type of sonic phenomenon, namely one that musical. He concedes John Cage's silent piece 4 '33 as a limiting case of music, since Cage has organized for listening the anticipated but unpredictable that will occur during a (1990, 270, fh. 3). He denies, however, that this shows that any sounds, when not so framed, are music. At most it shows that any can be treated as i/they are music.Perhaps unsurprisingly, philosophers disagree about the status of Cage's 4 '33 , so before evaluating Levinson 's definition I sketch my view of the piece. Though I have not provided a definition, I have argued (Davies 2003) that if is organized sound, however liberally we construe the notion of organization, it must be the case that some are excluded such that, were they to occur during a performance, they would be ambient to it. Cage's work is usually understood as absorbing all the that take place in the performance period as its contents, so that none count as ambient. In that case, Cage's piece fails the necessary condition for that I have proposed. I conclude that, as an artwork, 4 '33 is a performance piece, a theatrical work if you will, about the performance of music, not a musical work as such.The main difficulty with Levinson's definition, I think, is that does not and is not always intended to call attention to itself. …