I begin with the premise that there can be no performance without some interpretation. In other words, except for wholly recorded compositions (e.g. computer-generated music) or improvisations unconstrained by tradition, every performance is an interpretation. Indeed we often use the words interchangeably. For, however much a performer seeks to understand and convey what a composer intended, music – including the eighteenth- and nineteenth-century music to be the focus of this study – cannot be said entirely to ‘speak for itself’. Obvious behavioural evidence of this is that we might choose to listen to ‘Norrington's Ninth’ (read: Roger Norrington's interpretation of Beethoven's Ninth Symphony) rather than Leonard Bernstein's, Arturo Toscanini's and so on.

Just as every performance is an interpretation, every interpretation is either a performance or, when written as analysis and criticism, construable as a set of ‘instructions’ for a performance, though seldom explicitly so. As Leonard Meyer (1973: 29) has written: ‘The performance of a piece of music is … the actualization of an analytic act – even though such analysis may have been intuitive and unsystematic. For what a performer does is to make the relationships and patterns potential in the composer's score clear to the mind and ear of the experienced listener.’ But which of the myriad ‘relationships and patterns’ (surely never all in a single performance), on what basis and to what degree? Most importantly, with what consequences for the way in which experienced listeners apprehend a musical work or part of a work?

My general concern in this essay is with two kinds of consequentiality which exist symbiotically: consequences of interpretation for performance, and consequences of performers' choices – whether made intuitively or deliberately – for experienced listeners’` understanding.