this definition to music in Unsung Voices (1991), Carolyn Abbate observes the difficulty of making a similar sort of distinction, even in an opera. Most operas convey stories, but in few cases are we invited to imagine that there is a fictional entity responsible for presenting the entire story to us. Abbate concludes that operas are not narratives, but they may contain “moments of narration” where the story-discourse distinction can be perceived through discontinuities within the score or between the score and the libretto.1 The narrator of musical narratives remains unclear in Abbate’s account, however. As she repeatedly informs her readers, the voices of which she speaks are not those of the historical persons who created the work, nor the implied author, nor even the singers who make the work perceptually accessible.2 Through a process of elimination, these voices must refer to features of the work’s structure.
This suspicion is borne out in the evidence Abbate presents for or against a work being a narrative. Her argument that the epilogue to Paul Dukas’s symphonic poem L’apprenti sorcier (1897) constitutes a moment of narration rests on the appearance of the main theme in rhythmic augmentation. Abbate argues that the epilogue serves an analogous function to the quotation marks encasing the sorcerer’s words at the end of the poem on which Dukas’s work is based, Goethe’s Der Zauberlehrling (1797). Both imply the presence of a “third person narrator” who recounts the events to us.3 This storyteller is internal to the work’s structure. Absent from Abbate’s discussion is any consideration of the work’s context of performance: what its storyteller (Dukas) was attempting to accomplish with L’apprenti sorcier, whether he was successful, and how his audiences interpreted his work. Did they imagine an apprentice sorcerer’s futile attempts to put a stop to his spell, or did they regard the work as a purely abstract composition?
In the wake of Unsung Voices and other high-profile rejections of the possibility that musical works could constitute narratives, music scholars treaded more cautiously with regard to the narrative-definitional question, typically avoiding it altogether.4 An exception is Byron Almén, a music theorist who is not merely content to say that musical works are like narratives or that we may gain insights about them by regarding them as such. In A Theory of Musical Narrative (2008), he argues that musical works are narratives and presents a new “medium-independent” definition to support this claim. For Almén, a narrative consists of a hierarchy, established within a system of signs, that is subject to change over time—change that a listener interprets as a change in a cultural hierarchy of some sort.5
Almén outlines a method for interpreting virtually any musical work as a narrative. The first step is to identify the salient features of the music (pitches, keys, themes, instruments) that are brought into conflict. One observes the hierarchy in which they are found at the beginning and tracks changes to that hierarchy throughout the composition. Next, one classifies one’s findings according to the narrative archetypes that the literary theorist Northrop Frye proposed in his Anatomy of Criticism (1957): romance, tragedy, comedy, and irony (resulting from the permutations of order/transgression and victory/defeat). Finally, the analyst interprets these musical conflicts as representing conflicts taking place within a single agent, between agents or groups thereof, or between an individual and a group.
To highlight how his theory builds on existing practices in music theory and musicology, Almén illustrates it with discussions of preexisting analyses representing a variety of interpretive approaches. One such example is Susan McClary’s interpretation of the first movement of Bach’s Brandenburg Concerto no. 5 (1721). McClary focuses on the relationship between the harpsichord and the rest of the players (the ripieno, or large ensemble, as well as the other soloists). Bach’s concerto initially appears to be for flute and violin, with the harpsichord performing its customary “service role” as part of the continuo. Before long, the harpsichord begins to assert itself beyond its station, eventually “hijacking” the piece by inserting an inordinately long solo capriccio in which it “unleashes elements of chaos, irrationality, and noise until finally it blurs almost entirely the sense of key, meter, and form upon which eighteenth-century style depends.”6 Only then does it deign to allow the ripieno to reenter and restore order with its performance of the final ritornello.
McClary interprets the conflict between the harpsichord and the rest of the instrumentalists as representing the conflict between the growing individualism of the bourgeoisie in Bach’s time and European society, which was still largely under absolute rule. As Almén observes, a more typical concerto from this period would represent individualism and social stability as co-realizable through either “the appropriate submission of individual aspiration for the good of society” or “the reconciliation of the apparently contradictory aims of the individual and society.”7 Bach’s concerto, McClary argues, represents individualism that exceeds social acceptability. That the harpsichord eventually yields to the ripieno may appear to represent the individual submitting to the greater good of society, as in Almén’s first scenario. Nevertheless, McClary observes that “the subversive elements of the piece seem far too powerful to be contained in so conventional a manner.”8
Since the narrative resulting from Almén’s method is largely the listener’s confection, many different narratives may result from the same conflict. “Another analyst,” he speculates, “might have viewed the intrusive harpsichord music as a threat that is ultimately excised by the final ritornello—a romance narrative of the successful quest, if you will, rather than a comic narrative of a blocked society renewed or an ironic narrative of a fractured society.”9
Almén’s stipulation that the work must establish a hierarchy that undergoes change specifies some structural features the music must possess in order to be considered a narrative. But unlike Abbate’s definition, Almén’s may not be solely dependent on the work’s structure. Although he rejects Abbate’s requirement of a storyteller, he affirms the importance of a listener who interprets the work as a narrative.10 Precisely what role listeners play in determining a work’s narrative status remains unclear, however. It may be that McClary, by interpreting Brandenburg Concerto no. 5 in the way she did, makes it a narrative. If that is correct, the work is a narrative for McClary, but it would not have been a narrative for Peter Kivy, who rejected the validity of such interpretations.11 Alternatively, Almén may be arguing that the Fifth Brandenburg Concerto is a narrative, even if certain listeners refuse to regard it as such, because it is a work to which it is appropriate to adopt a method like the one he outlines. Given Almén’s lack of interest in authorial intentions or historical practices of music listening, what makes this approach appropriate appears to be structural features capable of supporting such interpretations.
Texts and Works
To understand what separates the foregoing definitions from the one I will put forth in the following section, it will be necessary to expose some of the assumptions about the nature of musical works underlying these definitions. Inspired by French literary theory, particularly the work of Roland Barthes, New Musicologists such as Abbate and McClary moved away from regarding their subjects of study as works and began to think of them as texts. This is not to say that opera scholars abandoned the study of scores and focused instead on libretti. What Barthes seems to have meant by the work-text opposition was the difference between interpreting what one is reading or listening to in light of the historical circumstances of its production and approaching it as a mere sequence of words or sounds, which could be interpreted in any way one pleased.12
Another way of understanding the opposition between texts and works is through the contrast between products and processes. As a text or product, a work of instrumental music is merely a sound structure.13 As a process, by contrast, it also includes all factors that contributed to its production, such as the performers, instruments, and performing circumstances for which it was written, and influences both artistic and nonartistic (e.g., religious or philosophical beliefs or events in the composer’s private life).
Barthes preferred texts to works because of his interest in maximizing interpretive freedom. Reducing works to mere texts licensed musicologists to put forth interpretations that were implausible accounts of composers’