Opera is a genre that prizes interior expression. How then does it cope with characters who are defined by reticence and inscrutability? Donizetti's 1837 opera Roberto Devereux thematises these issues – a reflection, in part, of the nineteenth century's growing anxieties over public self-expression. Engaging with contemporary 1830s criticism, this article focuses on two numbers in the work that were identified as problematic by early audiences, arguing that their perceived failure is a consequence of the work's inventive depiction of its characters' vacant subjectivities. Donizetti's opera can be seen to stretch primo Ottocento operatic conventions to their limits, in the process undoing the traditional correspondence between interior expression and self-definition.

It has often been suggested that a renewed fascination with Verdi’s Don Carlos coincided with the advent of Regieoper (or radically revisionist staging) in Germany over the past few decades. However, Don Carlos already counted among the most frequently revived operas in German-language theatres during the first half of the twentieth century. This article argues that neglect of this rich performance tradition is linked both to a German-centred narrative of the history of operatic production, which constructs the 1930s and 1940s as a gap in the development of ‘avant-garde’ direction, and to an over-emphasis on the visual side in recent academic discourse on operatic staging. These attitudes are challenged by a close look at selected German productions of Don Carlos from the 1920s to 1940s. Treatment of the opera's most difficult scenes – the mystical elements of the auto-da-fé finale and the dénouement – reveals striking continuities between the Weimar and Nazi eras, as well as conceptual affinities to some of the most acclaimed recent stagings. These findings call for a more historically grounded approach to operatic production.

The subject of this article is the failure of the Stalinist Soviet opera project. Although similar proposals had appeared years before, the project was inaugurated in 1936, and its realisation was placed in the hands of the State Committee for Artistic Affairs. The archival materials discussed in the article (including transcripts of the Committee's meetings) demonstrate that even publicly acclaimed productions were seen as failures by these senior bureaucrats. On the one hand, there were demands for realism and contemporary topics, and on the other, for monumentality and elevated musical language; these demands proved to be in deep conflict with each other. In addition to this crippling problem, it soon became apparent that any treatment of a contemporary topic was bound to become unacceptable before long, given the ever-shifting political landscape. While novels and films were certainly under close scrutiny, many operas were subjected to so many demands for revision that they never saw production at all. The article's central claim is that the 1939 Soviet reworking of Glinka's A Life for the Tsar as Ivan Susanin fulfilled the state's needs much better than any newly created Soviet opera could have, resulting in the effective curtailment of the project by 1946.