Just as politics can be analysed as a cultural and symbolic enterprise (that is, as theatre in the broadest sense), so too can theatre or opera (in a narrower sense) be analysed as political. Jonathan Dollimore identifies various conflicting processes at work in Renaissance English theatre: the ‘consolidation’ of power by a dominant order; the ‘subversion of that order’ and the ‘containment of ostensibly subversive pressures’. We need not accept Dollimore's essentially Marxist analysis of these processes in order to recognise the validity of his assertion that ‘the theatre [is] a prime location for the representation and legitimation of power’. But the way such power is consolidated, subverted or contained depends on the political and social systems in which the theatre operates. The issues are complex enough when one focuses on plays produced in Elizabethan or Jacobean London. They become even more difficult to sort out when single works or groups of related works are performed over a period of time in various locations, each with its own societal configuration, as in the different political entities that comprised the Italian peninsula during the first half of the nineteenth century (to which might be added the other European and even American audiences to which they were played). Under such circumstances, how can we measure the political implications of these works? Where does their meaning reside? How does that meaning change as a function of time or geography?

Three thematic clusters can be inferred from the libretto of Billy Budd: the sexual, the spiritual and the social. Each exists dialectically with the other two, creating a characteristic apparent synthesis of their respective internal and relational contra-dictions. Indeed, Billy Budd's status as a masterpiece traditionally depends on an assumed successful and harmonious resolution of its thematic material within the overall musical context. However, this article will argue that no such resolution takes place and that this ‘failure’ is the opera's most striking and interesting characteristic. I do not wish to argue that exposing these contradictions, along with the failed strategies designed to effect customary artistic reconciliations, lessen the opera's status: they are better regarded as evidence of its most challenging and worthwhile aspects.

A natural starting point for a critical reading of Verdi's Rigoletto might be the protagonist's ‘Cortigiani, vil razza dannata’ in Act II: an utterance around which much of the emotional intensity of the opera is centred. Rigoletto's outburst can be discussed to great advantage in terms of current musicological fashion, as it alters conventional forms in fascinating and provocative ways, and to great dramatic effect. Yet such an approach presupposes that the key to understanding operas lies in their Great Moments – those passages of intense musical expression that tend to be quoted in movies and television commercials. Of course these moments are a crucial aspect of our delight, and can be a rich source for interpretative ventures. But there is more to opera: various levels of meaning invite our exploration and enjoyment; hermeneutic ‘secrets’ lurk behind seemingly ‘trivial or irrelevant’ passages, and can lead to new perspectives on familiar works.

It is sometimes remarked that Wagner's prose works would not be of much interest if their author had not also turned his creative energies in a very different direction. What is not often observed, however, is the influence this truism has exerted on critical approaches to the theoretical writings. A tacit consensus dictates that these texts be seen as marginalia, ancillary documents whose value lies in the light they shed on immeasurably more important achievements. We have inherited the prose works in the form of useful gnomic tags: Gesamtkunstwerk, ‘endless melody’, ‘deeds of music made visible’ – fragments of text extracted from their context and pressed into service as authoritative pronouncements about the operas. More extensive analysis has usually been concerned with the relation between the written theories and the works they are understood to describe. There is a powerful hierarchy implicidy at work: text (theory) is subservient to music (practice); the former is at best an interpretative tool.

David Rosen has written a detailed critique of my essay, ‘Opera Talk: A Philosophical “Phantasie”’, which appeared in a recent issue of this journal. I am grateful to him for his interest in what I had to say there. But at the same time I fear that he has misconstrued it. The result is that he has set up a parade of straw men which (not surprisingly) he has knocked – indeed bludgeoned – to the turf. There are a lot of straw men standing in for a lot of real ones, some of the real ones more important than others. It would be a bore for me as well as for my readers were I to run through them all (and if I tried, the vigilant editors of the journal would, quite rightly, put a stop to my profligacy). So I am going to concentrate on a very few of what I take to be the most important ones, and ones that cluster around the same basic issue. In the process, I hope not only to perform the negative task of refuting Rosen's ‘refutation’, but the positive one of pushing my project forward, at least to the extent of clarifying it.

‘If music is a language, who is speaking?’ asked Edward T. Cone on the first page of The Composer's Voice (1974), later turning to such questions as ‘does a vocal persona [whether protagonist of a song or character in an opera] know he is singing?’ And does that persona hear the accompaniment? Cone's answer then was that ‘Consciously, he neither knows that he is singing nor hears the accompaniment; but his subconscious both knows and hears’.

The policies of centralisation pursued by Louis XIV and his ministers affected most aspects of French life and culture. From 1645 opera had been imported from Italy by Louis' minister Cardinal Mazarin, originally out of political motives. When it had become ‘naturalised’, assuming its characteristic French guise under the despotic direction of Lully's Académie Royale de Musique, it continued to serve political purposes. In return for a monopoly on theatre music, Lully saw to it that opera served not only as entertainment for the nobility and bourgeoisie, but also as propaganda for the state and for the divine right of the King. An incidental effect of these policies was that the number of French operas produced was small compared to the number in Italy. This was due to the monopoly; to the centralisation, which meant that with few exceptions ‘French’ opera really meant ‘Parisian’ opera; and to the lavishness of the productions, which made frequent changes of repertory impractical even with subsidies. Each première was an event of note, chronicled in official and unofficial sources – the archival documents, mémoires, correspondence, periodicals, pamphlets and books of the day. This profusion of documentation frequently makes possible a degree of precision about the history of early French opera that can rarely be attained for other national schools.

One of the tasks of opera scholarship, broadly defined as an interdisciplinary enterprise, should be to include within its purview areas of investigation that have received little attention in purely musicological research. While it is obvious that libretto studies might form one such focus, there is a broader area that should not be ignored. This is the historical and cultural situation of opera and its reception, a subject usually excluded in formalist analyses and studies of musical sources or performance practice, often with the implicit assumption that the ‘aesthetic world’ of opera is self-contained or has nothing to do with the ‘real world’. Even if one does not subscribe to contemporary theory's penchant for subsuming art and history into textuality, the interactions between opera and its cultural context are many-sided and complex, and deserve full scholarly attention.

Propped up in his bed, for all the world the quintessential fin-de-siècle invalid, Marcel Proust listened to the perplexing sound of music far away. He heard it from beyond the walls of his room, through a connecting tube: the famous théâtrophone, a permanent subscription telephone line that could connect Proust's apartment in the boulevard Haussmann to a number of Parisian theatres, opera houses and concert halls. The operatic scenes that succeeded in penetrating those walls were not scenes at all: they were disembodied voices, issuing instructions for the visual imagination. Those moments that progressed further — onto the pages of A la recherche du temps perdu — were of course even less corporeal: both invisible and soundless. In the passage from opera house to author to novel, who can say how much was lost? All that remains are words, hundreds of thousands of them, pouring noiselessly into a space where the music has sunk without trace; the fevered patient added reams more supplementary material (inflations, substitutions, emendations) as fast as the opera came in through the wall, papering — soundproofing — the room with words.