This dissertation begins with a source-critical problem with consequences for performance: the many differences between autographs and disseminated sources of Mozart's Viennese instrumental music. This problem serves as a foil for two lines of inquiry: the notion of performance in the late-eighteenth century and the place of performance in historical scholarship. The first chapter traces the recent history of Mozart scholarship in the context of wider debates about writing history; here, the central methodological question of the dissertation -- is there a musical way to write music history? -- is raised for the first time. The second chapter compares traditional musicological techniques of textual editing with more recent developments in the fields of "performance studies" and considers their application to the "multi-textuality" of some of Mozart's Viennese instrumental music. In the third chapter, the growing interest in the expressive qualities of human language in German aesthetic thought beginning around 1770 serves as background to an examination of shifts in thinking about the place of expression in musical performance. String quartets by Mozart and his contemporaries are explored for evidence of this change, and for the clues they might offer to their composers' opinions about fixing musical expression in text. The fourth chapter appraises two performances of Mozart's Fantasy for Clavier in c Minor K. 475, one by the composer himself and one in Wilhelm Heinse's 1795 novel Hildegard von Hohenthal. If performance is a criteria for interpretation, it is argued, then these performances yield two different meanings for K. 475. In the fifth chapter an instance of 'multi-textuality' in the String Quintet K. 593 is studied in detail. K. 593's material sources (autograph manuscript, manuscript copies, and early engraved editions) are considered in relation to documents of an early performance involving Mozart. Read together, these "echoes" of the expressive acts of performance paint a complex picture. Finally, a short epilogue draws major themes of the dissertation together and suggests that the arrival of both new theories of linguistic expression and the birth of "historicism" around the time Mozart was active as a composer is no coincidence.