Recently, the most exciting productions of operas have attracted attention by rebelling against established ideas of the opera's text and stagings - Peter Sellars' New York settings of Mozart operas are just one example among many. Likewise, the most stimulating developments in opera criticism have been in the area of performance, where a much-needed sharpening of opera-production theory has formed around such extraordinary re-stagings. This focus on performance is a welcome one, particularly for opera, where the visual component is of no less importance than the aural. Yet the nearly exclusive attention on extraordinary productions and a concomitant valorization of the provocative is troubling. Such selectivity, particularly when founded on loaded criteria such as "strong" and "innovative" runs the risk of creating more canons of "great works" or "great men."

This dissertation will seek to redress some of the problems with current methodologies for studying opera productions, illustrated with case studies of four of Leos Janacek's operas: Kat'a Kabanova, The Cunning Little Vixen, The Makropulos Case, and From the House of the Dead. My thinking on this subject has been filtered through work on Janacek's operas which, I have found, fit uneasily into existing models of opera studies. Unlike the Italian, German, and French operas that form the canon of opera criticism, Janacek's were notoriously slow starters. Only the premieres of his last few operas could be considered important musical events and even then only within the Czech Republic. Works such as Makropulos and The House of the Dead have acquired significance in international opera houses only relatively recently. The unusual relationship these pieces have with the operatic performance canon required developing new approaches to their study. First, I propose supplementing any examination of opera production with the very different information reception history provides. Alone, neither production nor reception can completely represent the impact of performance: on the one hand, the visual traces of productions, particularly those pre-dating video recording, are frustratingly ephemeral; on the other, the written texts that usually comprise reception history tell only part of the story. Bringing the two together, can fill in some of the pieces missing in either alone. Second, the myopic effect caused by focusing on single productions should be countered: as Gundula Kreuzer has recently argued, studies of newer productions often lack historical perspective. Thus I suggest along with Kreuzer, that the chronological purview of any such study be radically expanded to include stagings from the premiere up to recent years. Lastly, I suggest a shift in focus from difference to sameness. Reception histories in music have typically concentrated on changes in a work's meaning as indicators of shifts in broader historical, social, or political contexts. The problem with looking exclusively for difference is, as Jim Samson has argued, that a work's meanings may become so unstable as to render them meaningless. Tracing sameness or, to borrow from Jan Broeckx, "residual layers of receptional insight" through the history of an opera's production and reception not only reintroduces stability through continuously regenerated meanings of the work, but also provides us with new insights.