In recent years, the nature of analysis and performance relations has become a concern among music theorists. While many wish to see the practical applications of analysis to performance, they refuse to concede to any simplistic position that sees a direct correlation between structural insights and performance. However, it is rare that the investigations get beyond attempts to make technical analysis more relevant to performers, which virtually turns the analysis-and-performance problem into one for analysts alone. This thesis attempts to reconstrue the issue of analysis and performance in order that performers can play a genuine part in the discussion of it. Its premise is that the nature of musical performance is such that neither analysts nor performers can ever describe it adequately, and its prime concern is to identify the real issue behind that we ought to be dealing with. This I see as the lack of authentic discourse in academic musical writings: writers do not normally attend to the fictional status of words about music. Consequently, most studies of performance by analysts distort performances by moulding it according to analytical terms. With this in mind, I endeavour to expose the limitations of structuralist approaches to performance in a broad sense, those which see performance primarily as interpretations. This is achieved through critiquing contemporary theoretical literature on analysis and performance, and through commentaries based on a comparative listening of recordings. The discursive nature of musical writings about performance and analysis is then enhanced through a reading of performers' own writings and a case study of different discourses on a performance attribute.