Artistic Practice as Music Research: Towards a Generative Theory of Musical Meaning and Expression

By Dillon Parmer (Tenor), University of Ottawa Canada

I begin with the quotation at the top of the accompanying handout: “We who spend our paid lives thinking about music…need to spend a lot of time talking with our colleagues in the studios and practice rooms, in order to develop answers to [our] questions.” My paper today embodies what one such colleague might say were he asked “how do meaning and expression work in classical music?” My answer proposes a new model for musical meaning, one which, with no intention of invoking Noam Chomsky, I will call “generative.” I will illustrate its workings in the domain of only operatic performance practice in the Western Classical Tradition. And I will conclude with a brief polemic.

Current theories of musical meaning tend to locate meaningful content in musical structure, form, and design; in how musicians give music shape through parametric and sonic variance; and in the vocal, facial, and other physical gestures musicians add in performance. Whether formulated in in isolation from, or in relation to, contexts of composition and reception, such theories tend to reduce music into objects for listener-spectatorship, and musicians into the research subjects. The generative model comes at the matter from another vantage, from critically reflecting on the performed experience of music itself. Such reflection might seem to resonate with what such scholars as Christopher Small, Nicholas Cook, and Carolyn Abbate among others have been advocating for since the new millennium, to shift the musicological gaze from abstract musical works to live and recorded performance events. But reflection from having to perform or having performed, precisely because it is grounded in real-world artistic practice, also reveals what a gaze that focuses on performance cannot readily see, that a more fundamental creative act underpins all the structural variants, vocal inflections, and physical gestures left behind when musical works, as well as the structures defining them, are decentered.

In the performance of vocal music, this creative act interpolates a supplemental text non-identical but coextensive with music, word, and gesture. Theatrical artists call it a “subtext” or a “meta-text,” and training manuals for singers identify it as an essential component for vocal performance, one without which what’s on the page fails to come to life on the stage. For present purposes, I will use a term of my own making, Infrastructural Urtext. I say “infrastructural” because the content of this text resides beneath the threshold of what is perceivable in score, sound, and gesture. And I say “Ur” because, like the Schenkerian Ursatz, it is a basic idea underlying and driving the performance of musical works even though performers create it after the fact of what composers and, in the case of vocal music, librettists and poets have written down.[1] For ease of reference, I will call this infrastructural Urtext an Infra-text. Three gradated examples will illustrate incrementally where the infra-text text resides, how it works in performance, and, finally, what its implications for understanding meaning and expression might be.

The first example comes from an improvisatory exercise I learned in an opera-training program in France. The exercise requires two partners. While facing each other, one partner makes a gesture which the second partner imitates and follows with a contrasting gesture. The first partner then imitates the contrasting gesture and, in response, enacts a third which the other partner imitates replying with a fourth. The alternation continues until the participants reach a certain level of fluency. At this point, they add nonsensical vocal utterances to their gesturing. The gibberish can be sung or spoken or both. Once fluency is achieved, the partners add a further layer, improvised intentions from which each iteration of gibberish and gesture takes its impetus. Once fluency is reached, the partners gradually morph this combination of gibberish and gesture into improvised singing and action all the while striving for fluency. The exercise will feel silly to the unpracticed: indeed, novices tend to laugh. But as fluency and speed develop, laughter is replaced with an awareness of how conscious control of vocal utterance and physical gesture is relinquished to another process. In that process, the intentions arising in reaction to what one partner does motivates the other partner’s vocal utterance and physical gesture. It is precisely at this moment that an immaterial force takes form, one that comes to control the form utterance and gesture take in material reality.

The exercise can involve more participants, to form trios, quartets, quintets, etc. But we don’t need to go that far to get to the point. Even though it might be difficult for the unpracticed to reach fluency, it is not difficult to see how the exercise applies to music making. For even when improvised utterance and gesture are turned into fully structured singing and acting, the process—of allowing what is sung and done to arise in reaction to some anterior event—should not have changed at all. In this framework, performance becomes less a vehicle for presenting meanings found in musical works, and more an exercise in creating the illusion that music, text, and gesture arise spontaneously out of something else, an anterior force generating how performers move through the framework specified by score and libretto.

My second example illustrates how such a force seems to function in the performance of actual music. See your Figure 1. It’s an excerpt from the fugal Terzetto for Alto, Tenor, and Bass from Giacomo Rossini’s Petite Messe Solennelle (1863). For one production I was involved in, we (the three soloists) turned this movement into a competition, each statement of the subject serving as an opportunity for one soloist to outdo the others. Here’s how it worked. The first statement of the subject is allotted to the Bass soloist. The longest solo line in the Mass thus far, it can be sung as if it were an aria proper, an impression reinforced were the other soloists to remain seated until it is their turn to sing. Indeed, when the Alto suddenly enters with the fugal subject, the Bass is taken aback by her stealing center stage from him. For a moment, he sings more softly but quickly regains composure to reassert his vocal prominence at mm. 35-36. The competition for center stage builds through the next phrase, with the Alto having to get louder as she descends into her lower voice in order to compete with the Bass, who gets louder because he ascends into his upper voice. See Figure 2. At precisely the moment they converge on middle C, the Tenor suddenly stands up and bursts forth with the third full statement of the subject on the same pitch. In the ensuing trio proper, the Tenor takes center stage, while the other two leave off their vocal battle as they contend with a new competitor. Attempts to upstage the other singers becomes what motivates the rest of the number. See Figure 3. After a series of Trios, the Alto seems to win the game at mm. 92-94 where she gets to sing all by herself in her low register. The Tenor, however, gets a last word in when, at mm. 94-96, he answers the Alto’s line in exactly the same register, which allows him to deliver a sparkling high note outshining her low voice singing. And the Bass, who had the first word, now resigns himself to providing support for the Tenor’s last.

What I am describing here is neither the musical work as found in the score, nor any meanings supposedly inscribed in its tones, its words, or in the relationship between the two. What I am describing is a scenario governing how we performed the number. In such a performance, the musical work ceases to be that which performance presents and becomes instead the framework within which a predetermined sequence of actions and reactions is enacted. That sequence constitutes the infra-text for that specific performance. In our case, it arose not from score analysis but out of a social dynamic that existed quite apart from the score: we (the soloists) were all friends who enjoyed teasing one another, and we continued doing so during the first rehearsal. Having spilled over into the act of music making, all this teasing gave rise to the idea of how we might perform the Trio. Our interpretive decisions about volume and projection, phrasing and shaping, color and characterization derived not from the usual routines of contrapuntal singing—of bringing the subject out when you have it, of backing away when you don’t, of maintaining textural clarity. They derived rather as outcomes of playing the game of one-upmanship. And in that game, what we call the music itself becomes the acoustic by-product of enacting a set of anterior intentions that, although un-notated, nevertheless generate what the notation specifies.

To get one’s level of music making to a point where anterior intentions function as a prime mover of musical performance is admittedly difficult and time-consuming. It takes hours of repetition for technique, music, word, and, in the case of opera, stage action to be “fully embodied.”[2] And it takes even more time to bring what is fully embodied to levels of fluency necessary for performance—professional or otherwise. With the Terzetto, the idea of enacting competition came to us wholly by chance at the first rehearsal. Even if we were able to render it in performance, with only one more rehearsal scheduled we did not have the time to work out all the details not to mention making them embodied and fluent. This is certainly not the case in operatic production where, as Paul Atkinson has explained, performers have much more time: staging and musical rehearsals spread out over 14-21 days, a production run that can last anywhere from 3-4 shows spread out over a week to 20 shows across a month. But Atkinson’s ethnographic account doesn’t take into account the months of preparation that precede production, in the memorizing of notes and words, in the acquisition of technique to sing them, in the building of stamina for endurance, in the acquaintance with performing traditions and styles. Nor does it take into account what happens on the other side of production, when you are engaged to sing the same opera elsewhere and must, therefore, not only revise all that preparation, but also adapt yourself to different production concepts, performers, and conductors. Because operatic performance requires so much prep time, repetition, and revision, what you come to see regarding musical meaning and expression is that the matter is hardly a function of locating content on the page and then presenting it on the stage, what might be termed “structural performance.” It’s a function of a generative act. In the case of opera, that act requires you to create the illusion that everything you do on stage—all the notes you sing, all the dynamics, articulation, phrasing, coloring, all the words, gestures, and stage action—emanates from living through the life of the character you are portraying, as reactions to anterior events and actions. At this level of fluency, you realize that, in order to make that life whole and continuous, you must create thoughts, feelings, motivations, intentions, actions, histories, in excess of what the score and libretto specify. The infra-text resides in the aggregate of all these bits of experience, the total set of intentions performers must enact for any given performing situation. It is this set of intentions that drives performance and gives the music that is its byproduct any ensuing meaning.

To illustrate how this generative process works in operatic performance, I’ll take a passage from Tamino’s Portrait aria from Mozart’s Die Zauberflöte, and set it in relief against a hypothetical close reading of those excerpts. See Figure 4. The passage encapsulates what David Freedberg calls “Love by Sight of Image”: the portrait is a “Götterbild” that induces in Tamino both veneration of and terrible desire for Pamina. A close, hermeneutic reading might take note of two salient features of Mozart’s setting—a gap-fill melodic structure for the first two lines of text, and the evaded cadence at the end of the fourth line of text—and then interpret those features as embodying Tamino’s reaction to this love at first sight experience: even though he wants to fill the void in his heart with the love of this woman, he hesitates to embrace the new feeling fully. Thus, the musical setting can be said to “express” or “signify” his predicament. Even when conducted with more analytic rigor or critical sophistication, the approach entails that musical meaning is a function of grafting suggestive metaphors onto, or drawing them out of, the technical descriptions of musical structure. Such a critical practice might give the impression of opening hermeneutic windows onto musical works, but it does so by reducing musical meaning to score reading, it restricting the flow of meaning between composers and listener-spectators, and by limiting performance to a channel for such flow.

As useful as it has proven for opening proverbial hermeneutic windows onto music, the method ultimately reduces music making into a delivery system for presenting sonic structures and verbal patterns for listener-spectator enjoyment. By contrast, the generative theory turns away from such a hermeneutic practice to one grounded in action, specifically the act of performing itself, and thereby opens the door for making a more radical claim for how musical meaning works. It says that musical structure takes on the attributes of the performing intentions that give rise to it. And that those intentions are found not in tonal structure, textual meaning, or text-music relationships; nor are they found in the vocal inflections and physical gestures performers which on their own are only superficial markers of expression. They are found in how an infra-text created by performers gives rise to the delivery of music, text, and gesture. In the case of the Portrait aria, for instance, it is entirely possible to play Tamino as a wimp, as someone who moves through the passage with fear and hesitation. Indeed, this is how Lawrence Kramer sees him in his feminist reading of the opera. Both of these qualities might be marked in a vocal/gestural approach involving tentativeness and instability, even tension (of course, these would be deliberate artistic choices on the singer’s part and not the result of vocal faults). In this wimpy portrayal, the first evaded cadence becomes a moment when the discomfort Tamino experiences causes him to recoil from the new feeling of love blossoming in his heart. But it is also entirely possible to play Tamino in other ways, as a hero for example. Heroic Tamino goes through the same passage with quite another attitude. He would be just as astonished at seeing the portrait as his wimpy counterpart, but heroic Tamino launches himself through the aria with vocal assuredness, facial excitement, and overall anticipation, as part of his journey into manhood. In this portrayal, the evaded cadence marks that moment when a mild sense of disorientation envelops him as he crosses the threshold into new experience. I could concoct other infra-texts, but I think my point is clear enough. The process of making music making meaningfully expressive becomes not one of executing meanings found in scores as it is one of creating and then coordinating the character’s experience, which has to be constructed apart from the score, with what the score specifies, and then executing both in such a way as to make the performance believable, convincing, compelling. That a correspondence arises between an enactment of wimpy Tamino and the terminology used to describe musical structure is nothing but coincidence. Such coincidences hardly justify the appropriateness of playing Tamino as a wimp, and the close hermeneutic readings built upon such coincidences are not just irrelevant in performance, but are off the mark when it comes to giving an account not only of how real musicians communicate content in real-world performance situations, but also of how musical meaning and expression subsequently work in general.

Formulating theories of musical meaning and expression that correspond to what transpires in those situations entails more than entering into dialogue with musicians, as Cusick suggested twenty years ago. It entails rethinking the very subject who speaks to the art. As I have argued elsewhere, shifting from understanding music as text to understanding it as performance does little to alter how much it is still the listener-spectator who dominates music discourse, as if musical understanding boils down to how you hear a piece, listen to a recording, watch a video, attend to a performance. Rather than limit understanding to only these ways of experiencing music, the generative theory brings another subjectivity to the discussion table, what Linda Kaastra calls the performer-scholar. Moving from a scholarship based in listener-spectatorship to one grounded in having to perform music may not be enough to offset the hegemony of listening. But it does yield quite a different answer to the question I posed for this paper. Since music is not just something to listen to but also, and more importantly, something you do, meaning and expression become less things seen in scores, less things heard in structural and sonic variance, less things observed in vocal and physical gesturing. Rather, meaning and expression become emergent properties of enacting a deeper, underlying process that generates music, text, and gesture. From the vantage of the performer-scholar, the outcome of that process, the performance event, doesn’t express anything at all. It is that which is expressed. And that finding leads in turn to a whole array of other questions, of which one stands out as being worth a million dollars: how is it that the same musical structure can be adapted to so many different artistic intentions, to so many different infra-texts? But that is a question for another time. My time today is up. Thank you.


[1] Obviously, I am appropriating “Urtext” from the domain of editorial practice, where it is used to refer to an authoritative score in which an original source, like a manuscript in the composer’s hand for instance, is transcribed into modern notation with no editorial additions or alterations, the ideal being that this transcription allows unadulterated access to the composer’s conception.

[2] By “embodied,” I do not intend to mean the word in the superficial sense of embodied cognition, in the notion that how “we make sense of music through metaphors [derives] from our general bodily experience of the world as well as through our specific bodily experiences of engaging with music.” (Martin Clayton and Laura Leante, “Embodiment in Music Performance,” 191.) I mean it in the sense of having reached that moment of mastery when the configuring of your body into all the various positions it needs to be in to sing and act within the confines dictated by the score, the text, and the production concept can be carried out without at all having to think about it.