This article examines implication and expectation in music, taking as its starting point music-theoretical and music-psychological work ranging from the seminal thinking of Meyer (1956, 1967, 1973) to its development in the theories of Narmour (1990, 1992) and subsequent empirical and theoretical investigation by, for example, Schellenberg (1996, 1997), Von Hippel and Huron (2000) and Aarden (2003). Other psychological approaches, such as those adopted by Jones (1981, 1982, 1992) and Bharucha (1987, 1999), are considered too. The most important contemporary reference point, however, is Huron’s latest extended thinking on expectation (forthcoming), which summarizes, consolidates and develops a wide range of theoretical and empirical work in the field. These diverse perspectives on musical implication and expectation are analysed using the ‘zygonic’ theory of musical understanding recently developed by Ockelford (for example, 1999, 2002, 2004, 2005a, 2005b). This holds that the cognition of structure stems from a sense of derivation arising from the presence of repetition in certain contexts. Using this framework, a new, composite theory of expectation in music is developed, which acknowledges the potential implications of three sources of regularity in music: patterns within groups of notes, and between them - as encoded in short-term memory and long-term, both veridically and schematically. Finally, the phenomenological relevance of the new model to ‘typical’ listening experiences is discussed, and the need for future empirical work is set out.