One of my first impulses in examining the Comédie-Française Registers Project (CFRP) and in perusing the registers’ digitized pages, was to imagine a performance. The impulse to imagine came from the peculiar way the registers’ printed categories work with the manuscript numbers that were added to each page on a daily basis. Like any form, the general categories—Billets à repeated at the top; Frais journaliers… Pensions, Gages, &c… following reliably below—visually anchor the numbers that change regularly: tickets sold, price of places, dates of performances. Balancing the impulse to attempt a close reading of them with a fear of the perils of calculation, I instinctively took something like a formalist-performance approach to understanding the pages and pages of digits which danced, lightly shifting and transforming themselves, with each flip of the page. The performance assigned each piece of information a role in the greater meaning of each sheet. I flip through R69: 1715-1716. [Fig. 1]

The visualizer tells me I’m on page 541 of this volume; the register notes that it’s that season’s 264th day: du jeudy 5e jour de mars 1716, première représentation d’Athalie. I imagine the crowd jockeying in front of the Théâtre des Fossés Saint-Germain, buying seats: 4 loges basses à 40 livres; 5 loges hautes à 20 livres; 56 billets à 7 livres 10 sous; 26 billets à 5 livres; 66 billets à 3 livres; 19 billets à 2 livres; 280 billets à 1 livre 5 sous. Is this a big crowd? Bigger, one might assume, than the day before, when a double billing of Molière brought a total of 97,10 to the company. The double of this register, R70, tells me the same thing, this time on p. 575: this performance of Racine’s 1691 tragédie sold tickets to a public that paid a total of 1396 livres, 10 sous that day. What else could that total tell us about the performance that evening? What does such a sum of seated, standing, crowded, spacious parts tell us? What do its constituent characters—whether relative price or sheer quantity—contribute to the dynamics of this whole experience? What kind of an audience was constituted by 1396 livres, 10 sous, and what can we know about it? What did its presence mean for the performance which its members not only attended but, by virtue of their very bodily presence, helped create and shape acoustically?

Ticket-holders make their way into the building. This is the first performance of Athalie at the Comédie-Française. (Was this an anticipated affair, and so the reason for which Athalie was the sole performance on offer that evening? Might audiences nevertheless have expected a double billing, a one-act comedy which accompanied so many tragedies? Or was a late seventeenth-century Christian tragedy with machinery worth as much as one regular tragedy and a one-act comedy?). They exchange their tickets for the appropriate contremarque (Did they know what was playing, did they care, or was the theater a destination for reasons entirely distinct from what the actors were offering?). Some make their way down toward the parterre or up the stairs to the loges (How much better were the tickets at 3 livres than those at 2? How many loges were there and were any empty?). Candles are lit, money is counted, perhaps a baton is pounded on the floor to announce the beginning of the spectacle.

What then happens? In my experience with the CFRP, mediated through my laptop’s monitor and trackpad, a click of the mouse makes time advance. The registers are an archive of daily records, so a turn of the page enforces a respect of Aristotelian unity of time; the sun rises and begins to set again. For each click, the page flips and the vision disappears: audiences for that performance of Athalie are gone. Vendredi, 6 mars 1716: two comedies are up: Les Ménechmes and Le Médecin malgré lui; no loges and just over 200 tickets sold for a total that day of 288 livres. A smaller crowd than the day before (Do comedies garner fewer sales than a tragedy, or is it Regnard and Molière who are not so attractive? Are Fridays slower than Wednesdays? Is there another event enticing a public away from the theater that same afternoon?) Click again and the page flips; Athalie is back and the recettes are higher than two days ago: 1995 livres, with 404 tickets at 1 livre 5 sous. Are we seeing the positive effect of word-of-mouth praise regarding the performance from the other night? With each click crowds move in and out, tickets are sold and sold again; there seem to be patterns to the waxing and waning crowds and the expenses associated with each day. They are drawn in by friends, rivals, habit, anticipating the declamation, the music, the pathos, the spectacularity of the stage or the public. They are propelled out by the staging of the play of course, but sometimes they are likely lured away by a different spectacle or drama elsewhere. It’s as if I can see them file in and out, trickles and waves.

All this clicking and imagining creates something like a flipbook or time-lapse video, where time and action move forward at times haltingly, and otherwise quickly, like a windy beach where the tide can appear to ebb or flow more quickly. The scene is more or less the same from day to day, even if the accoutrements are different, the crowds are different, the changes incremental with the season and years. In this picture the horizon is constant, any interruptions are rhythmic: each page follows the same form, exceptions barely visible. The anomalies can be jarring—when I happen to catch a glimpse of one: terrible revenues one evening (mercredi, le 18e jour de mars 1716, Molière’s L’École des femmes: 56,5) or a missing day (le mercredi 25 mars, whose absence is explained on the next page). Pages featuring Athalie always offer a slight disruption. Along with the usual additional expense of 30 livres for “lustres et decorations,” these stagings warrant another 500 livres “retiré[es] pour les frais de la piece d’athalie” (Why don’t the registers tell us more specifically what they were for, these 500 livres? We can guess: was it the machinery revealing the Temple’s inner sanctum, revealed so spectacularly at the dénouement of the play when the Christian Prince Joas emerges to be crowned and soldiers leap out to defeat Athalie? Did the crowd appreciate the effects of these extra costs?)

Is this theater I am visualizing? If so, it must be observed who are the major players in this flow: they are not the comédiens on stage. The scene is of the house, its seating, its doorways, its divisions, its heterogeneity, habits, approbation and boredom. Even when the action seems disjointed, the imagination remains continuously focused on the spectacle of the public who attend, not of the players themselves—the actors appear on the verso of the pages, if at all, and in the registers of the feux and assemblées. The public imagined becomes a material participant in what might called the paraperformance: all the elements of a play that frame the spectacle, including the chaotic crowd outside waiting to purchase tickets, the finding of places, the interactions with other members of the public, as well as reactions to the theater and to the heat, cold, stench of the room itself. Then, perhaps its exits: in the middle of a performance, whether by an officer of the police; before the second play begins; at the end of the séance. It is therefore most materially the audience members I envision, even though the only trace of them that we have are the sums of money that the CFRP shows they relinquished for admission.

Despite the cool rhythm that the registers’ patterns enforce, and once we think of the bodies whose movements and transactions they trace, the scenes we create of them are noisy: we fill them with the extradiagetic and ambient sound of eating, dancing, shoving, pickpocketing and brawls. Jeffrey Ravel has shown that these were constitutive of the early modern spectator’s experience—at least in the parterre. So different from that of the “quietly contemplative aesthetic event” of the modern period, they affirm that “the chaos of the pit only heightened the intellectual and emotional intensity of the theatergoing experience.”1 What was that sound like, in all its moderate diversity, day after day, performance following performance? How did it apprehend the acting before it, what did it return to the players and dramaturges, how did the sounds of the crowd affect what theater meant and how it changed?

All of this reminds us how much theater performances are informed and created by the collision of people with material objects: from actors with scripts to machinists and ropes, to the public with the building itself (and just as possible: actors with the building, the public with ropes, the machinists with scripts). But where are these collisions in a digital archive? Where are the sounds that create the theater?

A danger of the CFRP then, is that it appears to silence theater, making of it a mute, transactional space that is best understood through calculation. If, on the other hand, continuing the last 10-20 years of scholarship, we consider the practice of theater as a multi-sensorial affair, and theater historiography as an endeavor to elucidate these multiple senses in the archives of theater, how might we integrate the multiplicity of sounds that were fundamental to the cultural production of the Comédie-Française into our study of the CFRP? What can shifting numbers on the pages tell us besides income and debt? Might they not offer insight into the individuals and groups whose bodies are represented by each sale, and especially how these bodies collectively contribute to theater?

To what extent must we accept the huge holes in the data, or the ambiguity of how the data might be mapped on to the physical and temporal space of the theater, as serious impediments to listening to these data? Further, how can we unite all of these characteristics: the value of over a century of material with the granular specificity of day-by-day ticket sales, and the lacunae that make the work of interpretation a precarious affair?

To what extent does seeing the multiple senses implicit in the archive make us change not just our vision of the theatrical experience lurking behind it, but the modes and practices of interrogating history for our own purposes today? That is, to what extent can a “resensorization” of theater’s past demand a different kind of archival practice? Perhaps we need to rethink both the politics of these performances, and our own methods of interpretation—methods which almost necessarily touch questions of understanding, discovering, even recreating the past. What then, might it mean to imagine hearing the sounds of which the registers are an archive, and how might such imaginings change how we use the CFRP? 

In what follows, I offer a thought experiment in order to explore the methodological and historiographical issues at stake in working with the registers, given new research in theater historiography. It proposes an unlikely tack, one allowing for a consideration not only of the meaning of theater performance in the early modern period and the material constitution of performance, but also the benefits and challenges of data-driven humanistic inquiry. I propose to consider exploring the CFRP as an archive of sounds, and what it might mean to create a variable acoustic model of a play in the repertoire of the Comédie-Française. In so doing, I hope to underscore some of the less acknowledged potential of digital humanities projects.

I. Why Sound?

We have long established as a significant basis for the art of theater its engagement with sound—whether through strategies of rhetoric and declamation, the vital role of music, or the crucial questions of projection and theater design.2 From the perspective of theater historiography, attention to sound not only challenges the primacy of the written text as archive, it also alters the history of aesthetics and performance associated with these texts. The production of sound—from actors’ voices through machinists’ pulleys, musical instruments, special-effects thunder to audience intervention—is at the heart of theater.

Perhaps uniquely, however, sound also adds to one aspect of the twin histories of performance and of theater, and to a recent call to recognize more fully what Christian Biet and Christophe Triau have called “l’hétérogénéité, le champ de forces des instances” of a theater where the presence and dynamic ephemerality of the public contributes to who—and what—creates a performance.3 That is, admitting the primacy of sound in the creation of the theatrical event compels the inclusion of a far greater range of actors into the event, from those who declare, those who pull or play, those who jeer or laugh or applaud to even those sound-making objects without which nothing would be heard. Sound then becomes not only a significant element in the creation of a performance and in our understanding of it, it also demands a very different model for such a performance. Sound therefore is both a material aspect of theater to be privileged and arguably the grounds for developing a theoretical schema for a holistic understanding of the social complexity of theater performance. Theater is created by actors and technicians, audience members and ticket takers, machines that simulate flight and seats that creak.

If dominant visual models privileged the politics of performance by determining, for example, who sat where and saw what, sound fundamentally challenges the isolation of the senses. To hear is also to feel: as media archeologist Wolfgang Ernst has asserted, sound has been recognized as an embodied phenomenon ever since Archytas of Taras considered the phenomenon of acoustics as physiological and not cognitive.4 Further, considering the full palette of sounds that create a performance, such physical apprehension is a collective endeavor: consider how sound waves emanate from people, but also are absorbed by or rebound off of their bodies, and connecting people as they inhabit a physical space. We might even consider these interactions as crucial components of theater’s “soundscape,” the now-classic term coined by composer Murray Schafer to denote the ensemble of sounds that define the acoustic ecology of an environment in a certain time and place, including the work and ordering we do when we, “earwitnesses,” perceive, integrate, or reject sounds.5 The soundscape of theater is not just the sound of actors, of music or machinery, nor even just that of the spectators. It is a combination of these noise makers, and therefore the network they create with the space which they occupy. The soundscape of theater is this network, of literally productive, reverberating, absorbing relationships among people, physical materials, time, and space that can be perceived and interpreted.

By thinking theater through sound, we expand the agents—or actors—that create theater. Both comédiens and spectators make noises that influence the theater experience. Oboes and stage machinery do so as well, as do the columns against which the sound waves bounce and refract, or by which they are absorbed and muffled. An acoustic understanding of theater implicates individuals and groups of people acting in concert or in conflict, within the physical plant—its materials, its spatiality—in the making of a performance. There is no sound, then, of the theater, if not a networked one. Privileging the sound archive of the theater then, is one way to further develop an understanding of those networks that constitute the theater experience and that undergird its archive.

I am here repeating and attempting to answer positively a question asked by Ernst, “Is there a sound of the archive?”6 In tracing sound we are forever following the reverberations shaped by various players amid the networks of participants in the making of sound. Sound is like performance insofar as each is defined in part by its ephemerality. In Peggy Phelan’s classic analysis, performance “becomes itself through disappearance.”7 Existing only in the present moment of its unfolding, performance can only be determined as such insofar as it ends, disappears, has been and is no longer. If sound, like the theatrical event, has meaning most fully in its evanescence, sound then is fundamentally a historical category. Thus conceived, sound—not unlike Phelan’s notion of performance—is to be considered as it vanishes. Our analysis of sound is always based in trace, whether of memory, archive, or physical evidence.

Paradoxically then, sound is crucial to theater but there is no sound of the theatrical event, which is always in the past. But there is a history of its sounds. Ernst suggests, “Written records or printed texts necessarily miss sound matter. But in the deeper sense there is implicit sonicity even in images, diagrams and graphs which are derived from sound sources; any sonagram keeps an indexical relation to the sonic event.”8 For Ernst, all sound representation is an archive of a sound source, and so bears some trace of those sounds, no matter its form.

Ernst here is referring to the maps made through ultrasound technology, in addition to scores or other diagrammed records; can the claim that there is “implicit sonicity” apply even to a written account of one aspect of the event? In other words, is it possible to see in the CFRP an archive of a sonic event? Clearly the registers represent performances, and record, however incompletely or enigmatically, certain contours of a major sonic contribution to such performances, its audience. An archive of a theatrical performance quite literally records—or registers—in some form or another, its sounds. This might be particularly true in the early modern period, where theater was not just delivered as a performance, but printed and circulated in part so that it could be orated in smaller venues. Print itself then is a form of archived sounds, not least because at the foundation of so much of early modern theater is the complex tradition of poetry as an oral art. Print conventions of early modern plays also seem to involve complex relations to orality, for example in the ways printers confound censors and the regime of privilèges by relying on actors’ scripts or audience memory.9

Arguably, then, the registers are nothing if not a sound archive, insofar as they offer a schematic representation of a noisy event. As they necessarily exclude the sounds of those performances, the data for each day dress the contours of what catalyzed them: the actors, the spectators, the porters and carriage drivers, and the musicians. Returning to my impulse toward imagining the audience of a performance which initiated the thinking behind this essay, it seems now that it is precisely the absence of sense data that inspired me to consider these as archives of events at all. What might then the CFRP, as an archive of theater’s evanescence, tell us about theater’s soundscape? How could the digital data of the registers participate in the turn to see theater of the premodern period as an embodied experience? Can an archive be an actor? What would it mean to make the registers perform?

II. The Imperfect Archive

At this point, a sensible reader might pause at these theoretical, even speculative musings about an archive as actor in performance, observing that neither this specific archive nor this period is an ideal fit for an investigation of sound in theater.

Such a reader might also wonder at the utility of such a projecting, pointing out that historians of theater have long considered sound in relation to eighteenth-century theater. Despite a paucity of records, scholars have already mined what we have, establishing a rich and varied understanding of the sounds of the early-modern stage and playhouse. And, despite such an incomplete archive, what we do know illuminates our understanding of theater studies, eighteenth-century French culture, and Enlightenment philosophy and science. Scholarship treating the question of sound in relation to the Comédie-Française has shed light on the noises of the theater in all their complexity, including the diversity of the public, acoustics of the buildings, and the ways in which theater was influenced by the multisensorial attentions of materialist philosophy.

Jeffrey Ravel has filled those spaces with many of the people who occupied them, showing just how diverse in socioeconomic status, temperament, and tendency to abide the law members of the parterre were, and argues for a dynamic evolution of the experience of the parterre throughout the century that crucially “enabled the French and French speakers in the eighteenth century to suggest a positive sense of community, political and otherwise, which differed from the models favored by the absolutist monarchy or the social elites who sat on the stage and in the loges.”10 As Ravel’s influential work has suggested, by the first decades of the eighteenth century, the emergence of interventionist spectators in the parterre meant that theater-goers became “irrevocable participants in the making of meaning in the playhouse.”11 The placement of benches in 1782 was one of many failed attempts to control the parterre.12 Its failure to exert control notwithstanding, it had an impact on the acoustics of the theater itself.

To understand what the French parterre might sounded like, or what its sonic contributions were to the overall acoustic environment, is, however, a different matter, and assessing what the spectatorship of the Comédie-Française might have heard generally, let alone any one performance, is particularly complicated, because the troupe occupied four different buildings. Additionally, the science, philosophy, and importance of acoustics evolved over the eighteenth century. These shifts afford us an opportunity to consider how ideas about acoustics shifted in time, as Jan Clarke and Pannill Camp have examined.13

In the context of her exhaustive study of the material history of the Comédie-Francaise’s early years, Jan Clarke has attended to the first two permanent theaters, the Guénégaud and the Salle des Machines. Recently she studied the early Salle des Machines, used sporadically during its first years, before becoming first a pantomine theater and then a storage facility for scenery. She explored archival plans, drawings, and reports to understand why the Salle des Machines had the reputation for terrible acoustics, and why the jeu de paume continued to be a viable model for theaters. Such factors as concern for the aesthetic of the exterior façade, the fear of combustible materials and the designer’s interest in constructing an impressive space, contributed to the creation of a building whose materials, enormous volume, unfortunate proportions, and decorative embellishments worked against acoustic success. By understanding theater as a product and effective of various aesthetic, social, and political forces, Clarke shows how the success—and periodic failure—of the Comédie-Française was based significantly in the materials of its construction and the space of its play. The salle itself becomes party to the performance, a veritable actor. Clarke also makes a convincing argument, based on contemporary understandings of acoustics, as to why the jeu de paume’s long narrow architecture was quite a successful form for the acoustics of the theater. Here, the awkward sight-lines and lack of potential for spectatorial performance—it is quite difficult to privilege being seen at a jeu de paume—were over-ridden, Clarke speculates, by the physics of the space which allowed for excellent transmission of sound. We might wonder, then, to what extent the multiple changes to the seating arrangements attested to by changes in loges were the result of an attention to acoustic problems. 

Questions of acoustics were not just important to the structure of theaters, they entered into debates about the nature of experience itself. Following on the work of Pierre Frantz14 who suggests that the fashion for the theatrical tableau emerged alongside a heightened consideration for the visual, including gesture, pose, as well as costumes and scenery, Pannill Camp has convincingly and thoroughly argued for a revision of how we see the relation between theater architecture and changing philosophies of experience: they were mutually constitutive, especially in the last sixty years of the eighteenth century. Following his argument, we see how questions of optics and visuality remain primary for the conceptualization of both the theater experience and its influence on natural philosophy—its spaces, methodologies, and spectators. Over the course of the century, suggests Camp, sensationist philosophy helped craft an idea of the theater and its public as a multi-sensorial affair. Whether in reformist projects for theater architecture, arguments about theatrical illusion, or debates about acting style, Enlightenment writers came to emphasize acoustics alongside sightlines. By 1765, some designers and theorists were imagining new theaters that cultivated “a stage-directed spectatorial act composed of sound and sight,” advocating theater design that might address the problems of actors’ voices being either “lost in the wings” or overly powerful and exaggerated to those sitting close to the stage.15 As Camp also points out, however, theater reformers’ treatises and drawings did not have an immediate effect on the real spaces in which Parisian performances were held, and in any event, experimental physics recommended a limited set of precautions to facilitate the distribution of stage sounds while mitigating those of the spectators.16

This kind of research begs for further work on the sound of theater. The CFRP clearly offers outstanding resources for a longitudinal analysis of these spaces and their performances. Given the length of the registers’ collection of sales, pursuing an understanding of how theater changed sonically over the hundred-plus years is tempting. We know it did change, if only because we know that acting practices, theatrical genres, and the acoustic space of the theaters changed. All of these changes, however, create seemingly insurmountable challenges. Indeed, the Comédie-Française occupied four different locations over the course of the century—not including performances at court or other venues—a fact hindering or at least complicating any long-term understanding of shifts in spectatorship. And as Clark asserts, the documentation for any one of the four houses is sparse and what we have is idiosyncratic enough that comparisons are not easy to establish. Further, as Ravel argues, the first-hand accounts of what theater sounded like are necessarily limited in their witness.

We should perhaps turn to a shorter horizon of time for greater consistency. Taken season by season, or run by run, the matter seems simpler, but the devil is in the details, of course. Performance by performance, the registers offer us ticket sales in a variety of categories. These categories tell us how much people spent to go to the theater. But they do not tell us how many people were in the theater. While tickets for the parterre were sold on a per-person basis, loges were purchased for a single price while the number of people in each loge varied. Additionally, some loges were rented by the season, accounts for which were kept separately, and are currently in the archives of the Comédie-Française. Furthermore, the capacity of loges could change with renovations to the theaters. Thus the distinction between quantifiable and “less accurate” information must be considered.17 Extrapolation from any one set of data is as yet untested: it is tempting to speculate that the ticket sales for the parterre might also offer a measure of how full the loges were; but this speculation is unproven. To complicate matters, it is difficult to determine to what extent record-keeping habits changed over the course of the century, given that categories of seating multiplied: how did new understandings of seating work with exclusively increased precision in recording-keeping? 

How can we further measure, chart, and understand the vagaries of what we know about the troupe’s record-keeping habits? How can we use ticket sales to better understand the impact of the public, those who purchased those tickets, entered the theater, and were witnesses to the performance? What contribution did they make to the event? How easy was it to hear? What did attendance mean acoustically, to spectators and to comédiens? How might it be possible to use such a diverse and changing dataset to make assertions about change over time? In merely envisioning the evidentiary grounds upon which we might start thinking about the sound of the Registres, the questions proliferate. With each question, a new investigation might be proposed. And with each new proposal, new lacunae in the data can be identified, delegitimizing the proposal. So why attempt to model a play as an acoustic event based on a dataset that is at once too much and too fragmented, too homogeneous and yet too scattered?

III. Sound Archive as Quasi-Object

Perhaps it is through the very questions that we ask of this seemingly vast, but maddeningly uncooperative archive, that we might ground an investigate of the theater’s sound. Such questions help us trace a different portrait of what the registers’ might be as an object of investigation. Given the complexity of the sounds the registers’s traces indicate, given the diversity of players—human and otherwise—contributing to the creation of the sonic event, and given the ways in which subjects and objects both invite participation from each other and yet obscure each others’ presence in the archive, perhaps we have at hand something more than a simple network of relations. Treating them as a possible sound archive pushes us to position the registers also as a kind of subject: one which proliferates questions, demands more data, requires more investigative models. It becomes something like a classic Latourian “quasi-object.” Bruno Latour’s original mobilization of the quasi-object worked primarily to define what he called “hybrids,” those forces that defy the divisions between “nature” and “culture,” between “human” and “non-human,” or “subject” and “object;” divisions which underpin scientific modes of interpreting knowledge and society and secure our mastery of the world around us. Latour borrows from philosopher Michel Serres the concept of “quasi-object” to express how these non-human actors, straddling the subject/object divide, not only have themselves a certain form of agency, but also bring out, in the subjects engaging with it, their own work as objects.18

Latour’s development of Actor-Network Theory (ANT), along with his abjuration and later recall and revision of it as “Actor-Network-Theory” in Reassembling the Social, has urged a method of inquiry that refuses just as much the fallacy that Nature is an object to be discovered as it does the idea that Society is one too.19 Instead, for Latour, the social is a kind of temporary assemblage; our goal should be the “tracing of associations.”20 While Latour and his colleagues began working at the crossroads of Sociology and STS (Science and Technology Studies), variations on their work have generated multidisciplinary practices examining the complexity of agency and meaning-, even “world-making,” practices that redistribute heretofore stabilized structures and stave off any need to recognize new power entities.

The use of theatrical terminology in ANT also offers a critique of a traditional notion of performance as a matrix for social critique, one in which the authority to critique is never in question. To use the word “‘actor,’” for example, “means that it’s never clear who and what is acting. . . . [A]n actor on stage is never alone in acting. . . . [T]he very word actor directs our attention to a complete dislocation of the action, warning us that it is not a coherent, controlled, well-rounded, and clean-edged affair.”21 Theater terminology here is not part of an extended analogy or simile; rather there is a complex epistemological charge to this call to revise where knowledge comes from, in particular how the very means by which we trace such knowledge-making affects what it is and what it can do. Thus performance-studies scholar and theater historiographer Margaret Werry, pursuing an “Oceanic history of performance” which integrates indigenous theories of historical consciousness and the radical alterity of Polynesian epistemologies, argues that ANT and the new materialism inform a framework in which we can understand “performance itself as not only a way of knowing but also a way of knowing how one might come to know.”22

Such an approach to the epistemological reordering which ANT allows suggests that it can also change, in a radically unpredictable and unpredictive way, the paths and modes of understanding we can take. As theater mobilizes objects, building materials, and people amidst each other, in a network of engagements, compelling them all to work together, it cannot act independently but forms its own network of relation among its actors of people and pillars, pulleys and brawls. The model of sound, and the related dynamics of acoustics, brings this quasi-objectivity of theater to the fore. As we consider theater as such a quasi-object, shaping relationships among people, objects and space, and shaped by them, we can see how sound might even be an apt analog to or figure for theater performance thus imagined. Of course this model in many respects assumes a dynamic, active site: the theater is quasi-object when it is in process, when it is producing sounds and when sound is producing it. So, can an archive be a quasi-object itself?

In part the question as to whether an archive can be a quasi-object might also appear to be asking whether archives can be permitted to have not just agency but also the agential ability to change relations with other objects and subjects. Can remnants of the past perform such work? I don’t mean this in a spooky manner at all—and thereby refuse many tantalizing ideas of haunting, revival, or other metaphors of temporal endurance to consider rather what a sound archive as a quasi-object might do, specifically.23 An archive as quasi-object does not represent the past as much as it appears as a monstrous figuration: with parts that are unexplained, extraneous, irrelevant, functions that are unintelligible, information that might be mistakes or perhaps crucial to an understanding to come. An archive as quasi-object does not represent the past as much as it offers a set of associations among actors. The quasi-object of the archive then mobilizes so-called subjects and objects into action, shows their associations, their influences, their impacts. Much like how sound travels and how acoustics are studied, an archive as quasi-object reveals relations, but it also creates connections. Far removed from the archive as a body of documentation ready for assessment and interpretation, the quasi-object-archive is more like a reflexive performance itself, whose significance might be more available to us if we use it less as grounds for affirming our own understanding of the past than as an interlocutor who participates in setting the terms for our inquiry. In this way, the quasi-object-archive is only intelligible as a repository for historical evidence when we let it also ask the questions. The quasi-object-archive must somehow participate in its own self-investigation, in its own experimentation.

This optic might be something like what Latour is hoping for when he urges us to “feed off uncertainties” in our approach, and enumerates five potential sources for this uncertainty, emphasizing the refusal of a fixed constitution or division of power.24 While Latour’s major aim—the trumped-up objectivity of sociology and its related disciplines—is particularly presentist, his injunction holds for historical approaches. Consider the impulses of so many archive-based scholars, whose fixation on an ideal of accurate historical portrayal is paradoxically affirmed by the commonplace that “we can never know what really happened,” which gives license to the original impetus toward “reconstruction.” And it’s particularly true for for Latour’s last area of uncertainty, the scholarly account itself. Whether article, performance, oral exam, file or website, the scholar’s intervention must be available for the same level of destabilization of entrenched habits and the reopening of its constituent pieces as the other elements of inquiry. It must be seen not as a conclusion but a “laboratory,” as an experiment itself.25

What kind of a “scholarly account” would do justice for a quasi-object-archive? My hunch is that a particular kind of model for this archive might allow us not just to see more of the complexities of early modern theater. It might also help us ask better questions of these archives, and of ourselves as we shape them through our research. It would do so by accepting this archive’s incompleteness, by mobilizing and not effacing its gaps and unknowns, by resisting a prioritization of our epistemological standards over its subjects’ understanding of what theater, acoustical physics, performance, success, sensation, might mean. It goes so far as to seek out its subjects’ understanding of these matters by searching for and even highlighting the fissures and ruptures within the archive. An acoustic model would eschew a positivist understanding of what it might represent and instead seek to create a framework for asking more questions.

IV. The Uncertain Laboratory

What I am suggesting works against most acoustic historical models today, including and especially the most elaborate ones. They are worth considering in relation to the CFRP, however, for two reasons. First, they show the successes and gaps in a traditional acoustic model. Second, their comparison elucidates what is particular about the CFRP’s potential as a site of experiment.

Perhaps one of the most elaborate models was that developed by English-literature scholar John Wall, which considered the phenomenon of public preaching in Renaissance England. The model he developed can tell us much about what it can mean to make a variable acoustic model, both for the archive and for historiographical questions more broadly. It is an inspiration—and yet also something of a cautionary tale—for what recreation of the past can mean, and what it can offer. 

In contrast to the vastness of the CFRP data, Wall’s object was singular in nearly every way: The Gunpowder Day Sermon by preacher and poet John Donne. The sermon, supporting a new edict by King James against charismatic preaching, was a planned performance, written for and delivered on Tuesday, 5 November 1622. The sermon was written for delivery at Paul’s Cross, the outdoor preaching space of St Paul’s Cathedral, a yard that could contain numbers as large as 6000. Wall wanted to understand the sermon as he says, “as a social phenomenon, as an interactive occasion, and as a public event.”26 Wall’s project began with a small question not about Renaissance culture writ large, or about Donne or his sermon, but one that tied the performance to the public’s role in its realization: “How many people could hear an unamplified human voice?”27 With such a precise question, the team began with the question of acoustics and then developed a visual model.

Through an NEH grant, Wall’s team created the Virtual Paul’s Cross Project (VPCP). They used architectural modeling, acoustic simulation software, and actors to offer experiential access to this event, despite a lot of variables to the historical record. Paul’s Church, which burnt down during London’s Great Fire, is known only through some prints made from archeological excavation and artistic depictions. They began with an actor recording Donne’s 2-hour sermon in seventeenth-century pronunciation. The recording was done in an anechoic chamber for later manipulation. The team then accounted for phenomena affecting the experience of a speech: they took into account the typical weather for a London November day (overcast, light winds), the smog from coal, the animals (dogs and birds, mostly), and the materials composing the church yard (the cathedral’s stone, the neighboring houses’ timber).28

The project has been housed in North Carolina State’s James B. Hunt Library, which opened in 2013. Designed as a technology laboratory and incubator for the people of the state of North Carolina, its Teaching and Visualization Lab allowed an installation that designed 21 speakers and 10 projectors “to provide a 270-degree seamless wraparound image of Paul’s Churchyard, as well as immersive surround sound for the audio portion of the installation.”29 The diverse technological tools and myriad forms of labor used in the VPCP are amply documented on its site; the visual modeling allowed for the construction of a wireframe image for the acoustic model, which could then reconstruct the conditions of the sound’s behavior from the appropriate corner of the churchyard. So, while the impetus behind the process was a question of acoustics, the visual recreation of the lost Paul’s Churchyard was a significant step that appeared foundationally and again in the final display.

The visual modeling approach taken in the Paul’s Cross Project has been used recently for the Pudding Lane Project and the Rome Reborn project, two historical-recreation projects that create immersive virtual experiences. In early phases of the Rome Reborn project, whose first version was completed in 2007, its creators resisted the idea of historic recreation, calling this work a “model” or “a representation of the state of our knowledge (and, implicitly, of our ignorance).”30 Now owned by Flyover Productions, Rome Reborn is described as a “a series of products for personal computers and VR headsets that make it possible to visit the now-vanished ancient city.”31 There is no epistemological meditation, and questions of historical speculation are absent. Aimed at the general public, including tourists to Rome, the site sells a series of apps and videos serving as individual guided tours of the city, while also providing quizzes and social-media rating capacity.

No matter how close they hew or far they stray, these recreations depend on a connection to an archive—a body of historical artifacts that authorize the experience as indexed to what was it really like. They also depend on a contemporary audience’s participation in and connection to the world they create. This same technology is a part of what makes Ubisoft Assassin’s Creed games so popular. Their historical accuracy is so impressive that fans document and share their brotherhood tours using game footage and on-site video. Never mind that in Assassin’s Creed players inhabit virtually the body of a time-travelling assassin who scales walls and jumps with the strength and powers of Marvel Comic book heroes. Some of these pilgrimages offer well-researched comparisons between existing monuments and geographies and the game’s depiction of their historical versions, often forsaking the open-ended and variable nature of game play for the overarching narrative of what should happen to the central character, as if the game were itself an account of historical events. These games are invested in representing worlds of the past in such a way that we can believe we can inhabit them, and their players explore them as if they are portals to a version of history. Resisting an archeological or historically factual strategy of representing the past, such large-scale multi-sensory digital simulations are “polychronia:” sites in which multiple timelines and temporalities cohabitate and thus are a “model of the contemporary public’s collective idea” of the past.32 Neither strictly historical nor entirely imagined, “they exist in a curious landscape situated somewhere between the two, emphasizing the epistemological impact of particular visual modes when communicating historical times, events and places.”33 Such attentive examinations have allowed historians to reconsider the work of dynamic digital immersive worlds. For Adam Chapman, video-game developers are engaging in a form of historiography similar to the work of “empirical-analytical written representationalist” historians committed to the “verticality of the book;” video-game play then becomes its own postmodern historiographic practice.34 

While these simulations concentrate on visual worlds, the same principle of recreation and the same struggle for a patina of authenticity inform how video game software is used by Mylène Pardoen for the Projet Bretez, a reconstruction of the Grand Châtelet quarter under Louis XV which also uses 3D visual modeling while privileging sound.35 Visual modeling in particular tends to satisfy users to the extent that it feels like a reliable “recreation,” giving experiential context, which is perhaps why Pardoen, who is a musicologist, contextualized the 70 soundscapes she created within the visual experience of a walk between the pont au Change and le pont Notre-Dame. Pardoen chose this area because of its soundscape: 

[I]l concentre 80 % des ambiances sonores du Paris de l’époque. Que ce soit à travers les activités qu’on y trouve – marchands, artisans, bateliers, lavandières des bords de Seine… – ou par la diversité des acoustiques possibles, comme l’écho qui se fait entendre sous un pont ou un passage couvert…36

The labor that went into the creation of these soundscapes must have been immense—Pardoen suggests that they are all “naturels,” with even the sound of machinery having been recorded on antique equipment. To call a sound natural in this context is to underscore its authenticity; to index its quality to the original, lost sound which it imitates. Calling her process a form of “archéologie du passage sonore,” Pardoen asserts a need to balance the work of the archive with the more aesthetic or creative work of composition.37 The matter of authenticity here is a fascinating one, since in a certain way as a standard, it excludes particularity. That is, Pardoen was not recreating a particular day or a specific sound, but rather an accurate sonic ambience. The natural, or the authentic, is also then the general.

This philosophy is quite different from, for example, Braxton Boren and Malcom Longair’s modelling of Venetian Church music of the Renaissance in situ, or that of Bissera V. Pentcheva on the Hagia Sophia, Icons of Sound. These soundscapes are so specific that they achieve the value of authenticity by the precise recreation of a moment in time. Similarly, Vincent Dumestre, Benjamin Lazar, and Cécile Roussat’s recreation of a period-faithful Bourgeois Gentilhomme (2004) was based on what was deemed “truly archeological work:” its researchers strove to determine period-appropriate lighting, diction, gestural repertory, and affect in its staging. Their efforts drew praise and some criticism: praise for the way in which the performance seemed so distant from ours (and therefore so much closer to a mid-seventeenth century experience), and criticism for the impossibility of the task.38

Paradoxically, then, the models that attempt to offer an ambient recreation intersect with those that target a very particular resuscitation. They both appeal to an audience’s understanding of what it might be like to have attended, to have participated, to have lived at a certain moment in time and in a certain soundscape. In some ways they all depend on and even cultivate the value of a certain kind of speculative history. While traditional speculative history outlines stages or trends over long periods of time to explain where we are today and perhaps where we are going tomorrow, the experiential model of historical recreation instead suggests that we can know what it was like to sense any given moment.39 That is to say, the broad sweep of the speculative history and the complex sensory modeling of the virtual historical world-maker both achieve a kind of empire over the past intending to make our relationship to it seamless and comfortable. There is a danger to historical recreation: the belief in the perfect recreation of the past, whether sweeping or precise, general or particular, necessitates an attention to detail that often escapes the archive.

These endeavors seem very different from the VPCP, which seems to be terribly precise, even singular in its eventual scope and corresponding research methodology: it avails itself of an existing manuscript text to represent what seems to be a precise sermon, given at an exact time in a definable location. Perhaps this is because Donne’s open-air sermon was never uttered in the space for which it was written, the space Wall’s team reconstructed. The sermon was moved “because of the weather” into the cathedral. This historical fact is generally not featured prominently among the project’s findings, in part because I think it goes against the sense many scholars (and grantors) have, that historians should be discovering and documenting what happened, not what didn’t happen.

In this regard, it might be argued that Wall was engaging in a specific form of speculative history, distinctive in its gesture toward pasts that never really existed. For this reason it is often called hypothetical, virtual, counterfactual, or “what-if” history. What if France had become Protestant? What if the 1830 Revolution had a different outcome? What if the Industrial Revolution occurred primarily in France, and not in England?40 While historiographers engage in deep debates as to the value of a model that tests an alternative outcome to the past, what interests me is the way in which these examples unrelentingly participate in the valorization of a certain teleology, even as they propose methodologically to undo it. We are in the regime of La Grande Histoire, for sure: of battles, of tides changing, of cultures built and destroyed. Counterfactual historians would never offer a counterfactual history of a minor event (what if the beggar Rose Keller had not gone home with the Marquis de Sade on Easter 1768? What if Prévost had not published Book 7 of his Mémoires et Aventures d’un homme de qualité qui s’est retiré du monde separately in 1733 as L’Histoire du chevalier Des Grieux et de Manon Lescaut? What if Beaumarchais had never made watches?) Whether speculative or virtual, written or engineered, historical models—visual, acoustic, schematic, or other—tend toward contributing to a progressivist sense of history marked by large-scale events whose importance we have already anchored. 

What if the Gunpowder Day Sermon had been given in Paul’s Yard? If for historians of English religious life, and scholars of John Donne, the question is not insignificant, it cannot be said by many scholars to be a question of singular historical importance; it cannot promise to yield a change in the order of history as we know it. One of the limitations of the VPCP was also one of its strengths—its specific attention to a moment, albeit an unrealized moment—in order to closely examine what constituted this sermon, from its structure, to its cadence, to its acoustic reception, and all the variables in between. At the same time, of course, it must be observed that this large project—over 25 people participated in the undertaking, funded by no fewer than 5 entities including the National Endowment for the Humanities, resulted in the recreation of one sermon. But the limitation of this project was the specificity of its endeavor. Once the recreation was over, the project is complete. What is the payoff of this project? Will more instructors integrate the Gunpowder Day Sermon into their classrooms? Will students gain a better appreciation for religious discourse? For whom might it matter how many people could have attended the sermon in the Yard?

If it is nevertheless tempting to consider the VPCP as a kind of counterfactual or hypothetical virtual model; Wall calls it “restorative;” the VPCP was not endeavoring to represent a historical event, “but an event that should have happened, but didn’t. It now finally happens, in virtual space though within the real two hours of time originally allocated for it, almost 400 years later. That captures for me both the restorative and the speculative character of digital modelling, what it can do and what kind of reality it constitutes.”41 Digital modeling does not recreate the past at all. It rectifies a lost opportunity.

Perhaps this is why what is most intriguing about the VPCP is not its final outcome, its virtual performance, which is at once ephemeral in that it is no longer occupying the space which saw its inauguration at NCSU, and enduring in that its visual and sound components are archived. Rather it is the process of building this experience that seems to really matter—to such an extent that a good deal of the website is devoted to an account of this process. This process also documents the unanticipated outcomes which emerged along the way. Because Wall and his team were not trying to recreate an original moment but to create a moment that never happened, the project could “embrace the ambiguity and indeterminacy of an archive” while highlighting what scholars know about open-air preaching performances at Paul’s Cross. The reality constituted in their digital modelling was not just speculative but generative.

Wall’s team constructed a model that foregrounds both of these qualities. The resulting model performs the sermon amid ambient noises—water, dogs barking, even the timing of the church bells—with certain variables: a crowd of sizes varying from about 250 to 5000, and positions from various locations in the courtyard. [Fig. 2] What that means is that the model refuses to let us experience the sermon as a singular event, suggesting instead that we must consider how where one is (and therefore who one is, whether elite or non-elite) makes all the difference to one’s apprehension of and participation in the performance. 

As I asserted at the beginning of this piece, to imagine a performance of the registers is to imagine a series of fleeting performances, like a time-lapse film edited for discontinuity, and with the camera turned toward the house. It offers a vision of theater as an event created by actors, spectators, and material objects. Might these data serve to create a model which honors this angle, while acknowledging the necessarily halting editing, the missing information and the visible lacunae?

The registers are tantalizing precisely for the ways in which they seem to make available a completely different path into history from the speculative, hypothetical, or generative. For example, the registers present 1036 plays, a challenge to the current repertory of fewer than 200 to which literary historians routinely make reference. While these famous eighteenth-century plays make a stunningly abundant appearance, they do so along with unknowns and forgottens, one-hit wonders, holdovers from the seventeenth century and other interlopers onto the main stage of the Enlightenment. The architecture of the CFRP digital interface further breaks down the notion of greats. Its search screen addresses the Comédie-Française’s custom of playing two plays in one séance with a layout that privileges visually this pairing, prompting literary historians to think of two different scripts together. The question of ticket sales and prices itself, one of the motivating features of the registers, has proven difficult to interpret and chart in a manner yielding immediately solid interpretations: the four different locations, and with each house the changing configuration of seating arrangements are merely the beginning of the complexity of a history of the theater through the registers.

The registers demand a methodology befitting of at once the small, varied stories they tell, and of the impulse they seem to grant: that there is something greater, if not more unified, to experience through this dataset. What kind of a model both marshals the size of a dataset that accounts for over 34,000 performances, and the scale of a phenomenon which, lasting over a century and in 4 major edifices, assuredly did happen? What model would push us not to recreate but to test, explore, and expand what we think we can know about eighteenth-century theater?

V. Toward an Ambiguity-driven Model for the Comédie-Française: Pitfalls to Avoid

A virtual digital approach to historical archives such as the registers must therefore adopt a model not of recreation but of variability, even ambiguity. This would help us avoid such pitfalls of theater history as the celebration of anomalies and outliers, the revelatory impulse, and the hidden drive of positivism. The danger of these three fallacies—the quantitative fallacy, the teleological fallacy, and the “discovery” fallacy—are exacerbated by the push toward quantitative organization of interpretation that is natural to data-driven inquiry.

The quantitative fallacy is the tendency to look for the quantitatively outstanding as the best (most successful, most lucrative, most frequent, most important), as if we know what “best” meant for the period—and therefore to take the data to be offered a whole and intelligible representation of success or failure. Doubtless, singular cases whether of play or of author, whether Le Mariage de Figaro or Shakespeare offer a glimpse into something unique and therefore interesting, but they also tend to reinforce deliberately or otherwise, our own categories of value.42 Rather, attending to that which does not indicate a quantitative success might allow the other variables of theater’s sounds to come to the fore and be tested. Thus while it might go without saying that a model should go beyond economic measures, it should also be constructed to adapt to other conditions. If we were to focus on one play, then, it should be not be an outlier, but rather one that shares features with other plays.

The second pitfall, that of an attempt to uncover the past, takes the form of an attempt to recreate the past or discover the truth in performance. The tendency in any kind of model is to hope that it might offer a representation of the past which adequately and totally explains it and makes it accessible, as Jeffrey Ravel’s exploration of earlier print adaptations of the registers suggests. A model not only does nothing of the sort, but also it offers—as Jeffrey Peters argues—something entirely new.43 Just as it should avoid the quantitative fallacy by exploring unexceptional or middle-of-the-road objects, a model that informs our understanding of the past should shift the burden of significance away from a singular account of any moment or condition, and toward a dynamic apprehension of historiographical questions. A variable acoustic model offers a new space for us to ask questions about performance and the power of social networks. The idea is not to recreate an experience but to understand the range of possibilities the archive offers us.

The last pitfall is the tendency to confirm (or disprove) teleological models. As the first two pitfalls suggest, we have a tendency to read what we already know into data. How do they support or confirm the evolution of Enlightenment theater? Or, in a more caricatured vein, how do the data predict the Revolution? For this reason, we need a model which is not seeking to show us the past, but to help us ask questions of our archive. If again the idea is not to recreate an experience we cannot recreate, then it should also be to expose these data to new forms, to offer them new expression that allows us to ask different questions. A variable acoustic model might be one example not of an attempt to represent the past, but a new form of engaging with the questions that the past can compel us to ask. 

Even the question of acoustics as an ideal way of understanding performance as an embodied and networked event must be modeled at a smaller scale, and this is where the issue of variability comes in. Can we use modelling to test various scenarios? For example, one vexed issue among scholars of the Comédie-Française is the relationship between ticket sales and numbers of people in attendance at any given performance. Christophe Schuwey and Christopher Morse tackled this issue during the May 2016 CFRP conference (Harvard-MIT), when they created a visualization of the data, “to reveal how crowded the theater was on a given night in the form of a heat map—the hotter the color, the busier the performance.” Schuwey and Morse charted the 1784-85 season. This experiment not only makes good on its goals to create a visualization of a crowd’s density over the season, and to show (more accurately) how we see revenue changing over the same period. More significant, it reveals when the data need another look. Schuwey and Morse isolated an anomaly: 27 April, 1784. That night, the heat map suggests that the third ring of loges is blue—empty. This is a historic improbability since it was the much-anticipated opening night of Le Mariage of Figaro. “Could this be an error in the data, or perhaps attributable to special seating arrangements, or season passes purchased in advanced? With a visual guide to each performance it becomes far easier to discover and query these inconsistencies.”44 Schuwey and Morse are now extending their project to include all 113 seasons, including graphics of the four different spaces in which the Comédie-Française performed. 

Could we also use the Registres to understand what poor attendance might have sounded like, in contrast to robust attendance of the same play, and how each condition might have affected the event? Could we not add to that model how the sounds of spectators might reverberate also, changing the spectacle? To what extent were those expensive loges to which Jan Clarke refers places not to see or hear the play, but to be seen and therefore to be heard? Additionally, this might be extremely useful for considering some of the important mutations that performance incurred over the eighteenth century; for example what was the acoustic effect of the removal of spectators from the stage after 1759, both for actors and for the soundscape altered by the change in ticket categories and seating. To what extent might Voltaire’s concern for theatrical illusion have been informed as much by what he saw of spectators on the stage as by what he heard from them?

I return here to John Wall’s VPCP, for what most intrigued me about it is what the process of modeling this sermon afforded him and his team. Two aspects of the process seem particularly relevant. The first concerns what Wall learned about the oral performance of such a piece. Donne’s speech was written to be given outside, near stone walls and the church facade that would have created echoes, whereas the timber-frame houses would have absorbed sounds. Under these conditions, the size of the crowd would have affected greatly the sound range and quality—how many people how far away could hear Donne, and how well they heard it. Wall and his group discovered that due to echo, Donne must have planned to utter his speech slowly to avoid reverberations masking his words. To what extent would modeling the oration of a performance, including an accounting for the comparative intensities of French phonemes, as Bruce Smith has considered for English theater, and the ways theatrical discourse borrowed from oratory (both civic and religious) to make itself heard, tell us something about what theater performance sounded like—I think in particular of Pierre-Alain Clerc’s work on the débit, the speed of alexandrines. An acoustic model of a particular play that was performed throughout the century might allow us then to model the kinds of changes over time that Clerc suggests happens to oration, and that others—Sabine Chaouche for example—have argued occurred in acting style.45

The second conclusion I want to borrow from Wall’s project is how it put the work of the religious faithful—the crowd—front and center, in its understanding of what preaching, religious persuasion, might have meant. In Wall’s model the crowd emerges as a lynchpin for understanding the sermon as an event, affirming what sociologist Bruce Reed, in The Dynamics of Religion has argued: that the practice of religious faith is, in Wall’s words, “a disposition, a physical as well as intellectual orientation, formed through participation in a set of practices that establishes and supports an understanding of the self in terms of social networks and shared participation in the formation of relationships.”46 Might modeling the potential crowds in the registers allow us to see further the making of theater as precisely that of a set of practices establishing the self and one’s beliefs within social networks? Might we identify, perhaps, a power of the crowd to influence performance through its presence, a power is that is not financial, but nevertheless economic in its ability to transform energy and resources?

VI. The example of Athalie: a thoroughly adequate play

What would it mean to create a variable acoustic model starting with the data of the registers? Such a model would include input from theater practices from treatises on acting, song and music, and what we know about performances of plays. It would also include information from the architecture of the buildings and their materials; the history of set design, the testimony of players, attendees, and dramaturges, the insight from police records and memoirs. It is of course mainly to this larger acoustic complex that the digital data of the CFRP might contribute directly. 

Of course the thousands of plays, the four performance spaces that were modified repeatedly, the shifting performance schedules, and the outside historical elements that would have influenced performances, and other variables make any idea of a model as a miniature of a physical space almost untenable. Which space? At which moment? Given the scope of the data and the divergences among them, the first step then would be to identify some parameters, some elements whose relative stability will shape the model. It is tempting, therefore, to pick one theater, one season, or one play. In such a narrowing, I think the key would also be to develop a model that embraces variability and even underscores ambiguity, one that lets the nature of theater as “quasi-object” remain and even grow. Here I think we can also learn from the VPCP, which started with one sermon, the nature of which was oral but immediately recorded, improvised but rehearsed, made for the yard but ultimately preached in the Cathedral. That is to say, the goal should not be to recreate an event or to capture history. Rather we should strive for a model that allows for a variety of questions to be posed—questions both about theater performance as well as about the nature of the dataset upon which the model depends.

To create a variable acoustic model of a play as an event created by players, space, and public would allow us a laboratory for exploring how we consider theater of the eighteenth century. In this model, a recording of the text by actors would be manipulated according to the acoustics of the theater house, which would include what we know of its materials and space, but also what we know changed about its layout and the crowds who were present from performance to performance. The text in question should be chosen, however, not because it represents the quantitatively outstanding. Indeed, it would be an error to take what we would consider the most successful (whether the most frequent or the most aesthetically or politically important), as if we know what “best” meant for the period. To do so would be to set the data up as a complete and intelligible representation of success or failure. Rather, I would think that if one might record a text, it should be one that was in fact not a quantitative success—not an outlier, as it were, not an exception. Rather, a play that has endured without fanfare, one that appears regularly but perhaps without great consequences, might clear the way for the other variables of theater’s sounds to come to the fore, be tested, heard, and followed.

It is therefore not entirely perverse to suggest modeling not the most popular play of the century, Beaumarchais’s Mariage de Figaro (1784), or one of the most revolutionary, Chénier’s Charles IX (1789), or one of the “firsts,” like the first “French” tragedy, Voltaire’s Zaïre (1732), but rather something quite the opposite: a holdover from the seventeenth century, Jean Racine’s 1691 Athalie. Performed at first only 3 times, not even in a proper theater but in front of members of Louis XIV’s court, and then prohibited publicly for decades, Racine’s last play appears at first quite atypical: a product of an aging court and its relationship to the noble school girls of Saint-Cyr, its genesis and first life is distinct from the repertory lives of many Comédie-Française plays of the eighteenth century. Any exceptionality, or lack thereof, might be less significant for my purposes, however, than the ways in which Athalie was drawn into the repertoire and so integrated continually into the theater life of the eighteenth century.

Athalie was first performed in 1691 on 5 January and again on 8 February, by Mme de Maintenon’s students at “La maison royale de Saint-Louis,” Saint-Cyr, for a select group of nobility, and for the last time in front of the Queen and King of England in exile at Saint-Germain on 22 February. A prohibition against performing it in public ensured that no troupe would take it up, though performances behind closed doors did occur. It was not until 1702 that Athalie emerged from its cloistered state, during Carnival of that year, but not to be played by professionals. Members of the court performed for the king aided by Michel Baron, “le vieux Baron,” of Molière’s troupe. According to Georges Forestier, “on garde la trace d’une représentation privée qui eut lieu en 1714 chez la duchesse du Maine. . . . Pour la première fois des acteurs professionnels tinrent les rôles principaux. Et l’on est en droit de penser que ce ne fut pas un cas isolé.”47 The prohibition was formally lifted in 1716 by the duc d’Orléans, who had himself played in the 1702 performances, without the chorus; Forestier asserts that it was performed 14 times while the CFRP record 13.

Over the course of the century, the play was performed 209 times according to Forestier—206 according to the CFRP. What’s so remarkable about this record is that it is not among the top in any category of Racine’s plays. Phèdre was performed 424 times, Andromaque 296 times, Britannicus 289 times. But it was presented more frequently than Bajazet (184), Bérénice (78), Esther or La Thébaïde (each 8 times), not to mention Alexandre le Grand (3). Its performance history remains remarkable well into the nineteenth century, where according to Forestier, it is performed even more frequently (255 times). This is particularly interesting not because of the slight quantitative increase, nor as a signal of increased popularity—according to Forestier, it passes from 6^th most popular to 5^th most popular, well ahead of Mithridate. Rather, it speaks to a continuity of interest for the play: Athalie persists. A relic of the paradoxically decadent austerity of Louis XIV survived the eighteenth century, along the way becoming what Renaud Bret-Vitoz suggests was “davantage une tragédie du dix-huitième siècle que du dix-septième siècle.”48

Forestier is measuring Athalie against Racine’s plays only. Within the context of the Comédie-Française, the question of impact must be contextualized differently. Derek Miller has visualized the “afterlife” of the Comédie-Française’s 1680 repertory, the longevity of this season’s plays through the eighteenth century.49 Miller astutely observes how a use of brute numbers of productions necessarily privileges older plays. Instead his list is organized by the length of annual runs of plays running more than 5 years, and thus reveals the enduring popularity of Le Mariage de Figaro through the century. This “afterlife” approach also underscores another phenomenon: that of the staple, the play that remains in a repertory consistently—if not strikingly or remarkably. Athalie is one such enduringly adequate play. Starting with the 1715 season, Athalie was performed as part of 55 seasons through 1791, in every decade, with the largest period of dormancy a five-year stretch starting in 1733.50

This means that Athalie was performed in all three of the eighteenth-century buildings to house the Comédie-Française. It is an intriguing idea to imagine an acoustic model that, unlike immersive virtual realities, could be adapted to each of the four buildings which housed the Comédie-Française. Privileging the sound, perhaps at the expense of a sophisticated recreation of the physical plant of each theater, might offer insight into the changing fortunes of theater from building to building, by allowing the transformation of sound to guide the way we understand these changes, instead of what we think we could have seen.

Given its performance record, we might consider Athalie to be an example of a play whose success waned over the years: it was performed never again as many times as it was in the first years of its reintroduction, and even as few as one performance in a season (1725-26, 1741-42, 1747-48, 1778-79, 1780-81, 1782-83). Yet with the exception of the 1730s, when Athalie was absent from the scene for the greater part of the decade, each full decade saw between 20 and 30 performances. This isn’t a lot, but it’s consistent. It suggests that Athalie might be a satisfactory test case for a long-range interpretation of the life of one unexceptional play over the course of the Comédie-Française.

By “unexceptional” I meant to underscore the contrast between the reception of the public and that of critics. Especially in retrospect, with our sense of genre, of Racine’s career and the fortunes of his œuvre in the eighteenth century and beyond, Athalie is exceptional. But that is not the case, for the reception that the registers record. For example, recettes vary from 276 livres (27 March 1722, with Du Fresny’s Esprit de contradiction) to a high of 5708,8 (1 August 1788, with Joseph Pilhes’ Le Bienfait anonyme). All of this to say that Athalie represents a range of success and mediocrity that might be considered to overlap of so much of what is “typical” of the registers as they record eighteenth-century theater. This is a striking claim, at odds with literary-historical assessments of Racine and his last play especially. Criticism generally positions Athalie as an anomaly in many respects—whether among Racine’s works, early modern tragedies from the court of Louis XIV, those of machine or religious plays. This position often establishes itself on such literary-historical grounds as the play is Racine’s only full-length tragedy drawn from scripture, or it is exceptional for its end-of-the-century on-stage expression of violence. 

In the writings of eighteenth-century critics, Athalie’s status was rather multivalent. Generally, Racine’s tragedies were held up in the eighteenth century for their poetic genius—but as non-theatrical poetry. While Athalie was lauded for the beauty of its verse by, for example, Voltaire, D’Alembert, and Marmontel, it was also derided for the absurdity of its action and characters.51 Nevertheless, some saw within its final specular scenes, so exceptional for Racine’s plays, a model for the “pathétique” that would be so celebrated by eighteenth-century critics generally. Athalie becomes significant for the history of the Comédie-Française for the ways its spectacle was variously rejected and imitated, and so echoed, by diverse aspects of eighteenth-century theater practices.

VII. Echoes of Athalie in Enlightenment Theater

Despite those eighteenth-century critics who decried older tragic forms overtly or implicitly, Athalie appears, sometimes paradoxically, woven into both the writings and performances of theater over the whole eighteenth century. In this regard, its sonority is decisive. Multi-faceted, integrated into the play’s composition and structure, sound is a major element, both lauded and decried, of its critical reception and its diverse performance history. 

First, as a foil for the establishment of the spectacular, vision- and sense-driven theater that eighteenth-century critics developed, Athalie occupied a double role. The appreciation of its poetic verse made it a prime source for imagining the legacy of theater under Louis XIV, as in Julie Philipaut’s early nineteenth-century depiction of Racine reciting the play. [Fig. 4] In this image, Racine is not acting but declaiming; Madame de Maintenon and the King’s hands are also captured in movement echoing Racine’s gestures. It encapsulates the double vision of seventeenth-century tragedy that evolved over the course of the eighteenth century and into the nineteenth: at once a static, even stilted relic of an elitist and obsolete political regime, and an exemplar of compelling poetry that affectively moved auditors.

While it should go without saying that an acoustic rendering of a play would allow researchers to consider evolving ideas about declamation and acting, Athalie is particularly interesting for the ways in which it became over the course of the eighteenth century both an example of poetic excellence as well as a source of spectacle. Philipaut’s rendition of the play’s first years celebrates one of these legacies, but the other is just as important.

Over the course of the eighteenth century Athalie had a series of well-known performers in the titular role whose performances might be tracked: we know from the registers who performed them for each run, and such additional sources as Le Mercure galant confirm not only the distribution of roles, but also the reception of their performances.52 Might a variable recording of Athalie allow us to consider what we know of the different acting styles—and reception—of Desmares, Dumesnil, Duclos, Lecouvreur, Clairon, not to mention Baron, who played Joad to great acclaim in 1721 after his return to the stage, or the children who played Joas? [Fig. 5]

Critics at the time compared their interpretations; a variable acoustic model of the play might allow for a better understanding of the aesthetics of dramatic interpretation and the controversies around acting styles in which critics engaged.

 Athalie was first written to have music accompanying the chorus. The score was composed by Jean-Baptiste Moreau, which a scholar of the text might nevertheless consider an essential part of the play. Remarkably, however, Athalie was performed only 8 times with music. And the first performance with music was not 1691—not with Moreau’s music but a score by François-Joseph Gossec, and solos by Haydn. In fact the only time Athalie was performed with Moreau’s music was in 1770 at the Opéra de Versailles, yet with certain pieces inserted from other scores.53 An acoustic modelling of Athalie might include these variables: an ideal recording of a performance that never was, with Moreau’s score for the chorus, and then variations with the other pieces of music included. One might also offer a counter-factual scenario using Moreau’s original score, to ask whether the original music was itself a problem for spectators of the play. 

Indeed, we know that Athalie was—perhaps like other performances with music—subject to emendation: the entire choral portion could be shortened or excised entirely. The Mercure Galant of June-July 1721 reports that, for that season’s performance, “Les Comediens en ont retranché la plus grande partie des Chœurs & tout le Chant.” [Fig. 6] Athalie’s sonic landscape is therefore quite rich: chorus, music, and distinct actresses and actors all shape its potentional sound archive, as do their absence.

Perhaps most interesting for the ways in which the stage side of performance might have add an impact on the experience of the theater is that Athalie is perhaps most celebrated as a machine play. Its critical scenes in Act 5, especially scenes 4 and 5, literally perform its revelation of Joas as the rightful king, as machinery opens up the rear decor to reveal a sanctuary within the sanctuary where the young king emerges, crowned on his throne and surrounded by soldiers who will drag Athalie off stage to her demise. As Renaud Bret-Vitoz has argued, it is precisely for its machinery that Athalie became “le parangon de la tragédie sublime et spectaculaire qui rompt avec l’esthétique antécédente.”54 The spectacularity of this mise en scène of revelation is unique in Racine’s œuvre, and was likely achieved through stage technology when played by the Comédie-Française. The machines offer a particular category of sound. What did the grinding of pulleys or the operation of wheels in tracks add to the cacophony of this scene? How did it affect the declamation of the verse? Is there an interplay between the machinery’s sound and actors’ voices that might have been accounted for as the performance built toward the play’s important tableau?

Perhaps just as importantly, a variable acoustic model would allow us to experiment with the sound of disruption: what was it like when a performance was booed? What does it sound like when the parterre throws oranges at an arrogant gentleman whose on-stage seat blocks their view? How many people might drown out a performance in order to hasten the transition to the second play on the bill? A variable acoustic model would challenge us to take up the charge of listening for the anomalies that gave the theater its potential to not just display, not just disrupt, but also to change society.

Conclusion: May the Uncertainties of the Past Drown Us Out

With Wolfgang Ernst, I do believe that the archive can have a sound. My suggestion of creating a variable acoustic model from the CFRP, coupled with other qualitative archival sources, and a variable acoustic recording of a particular play, would allow us to join quantitative and qualitative sources. In doing so, we can go beyond the idea that we can create a past performance or series of performance. We can attempt to use the multi-faceted and multi-actor model of acoustics to challenge ourselves to think with the historical archive and to let it ask us questions, instead of, as Bruno Latour, suggested, “deciding in advance what the furniture of the world should look like.”55 What kind of uncertainties can we hear once we allow the matter and people of the past to articulate themselves in concert with its spaces? 

Most importantly, I have a hope that a variable acoustic model, in all its potential adaptability and dynamism, might challenge us to resist the kind of master narratives to which we still cling. In short, by pointing us to the kinds of questions we should be asking, a variable acoustic might productively drown us out. For, I think we can agree that the loudest sound of the archive is often that of the historian her- or himself (coughing, gasping, snoring, or clicking, swearing, and rebooting) who ventriloquizes through interpretation, mute data. In moving-image and sound studies this might be what Michel Chion called the acousmêtre, that invisible voice seemingly just off the screen whose power comes from not being seen within the frame, but from threatening to break through at any moment—an acoustic version of a Foucaultian panopticon, as it were.56 Above all, I think that we need to be very careful about pretending that we, as acousmêtres of the archive and the stories we tell about the past, can ever let the data speak for themselves. We are always present; the first sound of the archive is our own noise.57