What did it mean to sing a song with a refrain in the Middle Ages? The refrain was an explicit marker of form and, to a degree, function for poets, composers, scribes, and singers. Refrain forms were consciously adopted for certain subjects, gathered together, and ordered in specific ways in manuscript sources. The refrain song was created, performed, and transmitted within and among communities for whom forms of musical and poetic expression rooted in the calendar year and religious narratives were meaningful – communities for whom retelling saints’ miracles in song or celebrating the Feast of the Circumcision as the New Year, for example, contributed to a vibrant and communal devotional and musical life adjacent to the liturgy. The refrain was also a formal axis linking Latin and vernacular song, a flexible and mobile unit of text and music across songs, and a possible signal for movement.

All of these aspects related to the performance, textualization, and function of the refrain testify to its currency within medieval song culture, but do not necessarily address the question of why the repetition of text and music in performance – sometimes of content as brief as a single word – carried so much cultural, musical, and poetic weight.Footnote ¹ Why the refrain? How did the performance of refrains in Latin song foster and maintain certain sets of cultural and ritual associations within and across clerical, monastic, and pedagogically oriented religious communities?

The evidence of the songs themselves, their poetry, musical settings, and contexts of manuscript transmission, reveals one way in which the refrain operates in medieval Latin song that embraces and even fosters its cultural impact in performance. The refrain is powerful not solely due to the rhetorical potential of repetition – although that is significant – but also by means of the refrain’s relationship to poetic, musical, and liturgical performance practices. Evinced in poetic language and grammar, form, musical settings, and scribal cues, the refrain in devotional Latin song brings individuals together in the moment of performance through the act of remembering together, responding collectively, and worshipping communally. With roots in the responsorial, litaneutical, and ejaculatory refrains and repeated prayers of the liturgy, the refrain evokes both real and imagined performance practices around Latin song that contribute to its agency within the devotional and musical lives of medieval communities. The refrain carries cultural, musical, and poetic weight precisely because it rarely reflects the utterance of an individual, but instead expresses the voice of the community, whether defined as narrowly as a choir in a local church, or as broadly as the clergy, or even the entire church.

This chapter explores the meaning of the refrain in and for communal performance. For the communities who created, sang, and transmitted devotional Latin song, the refrain afforded a moment for the collective, shared experience of remembering and responding. It has long been assumed by scholars and performers that refrain forms in medieval song initiated, on a practical level, a call-and-response format, with strophic material performed by soloists and the refrain by a choir.Footnote ² While there is an attractive simplicity in the equivalence of performance practice to form, the reality is more complex. Scholarship by Ardis Butterfield, Jennifer Saltzstein, and others has shown that this is not the case for the French refrain, for example; instead, the refrain evokes variable ideas of performance through its dissemination and interpolation into other genres, including narratives and the polyphonic motet. The relationship between form and performance has yet to be similarly questioned for the Latin refrain, in part because its structural, rather than intertextual and citational, identity seems to discourage deeper examination. Moreover, responsorial and refrain forms in the liturgy, most significantly processional hymns with refrains, have offered appealing and convincing touchstones for the performance of devotional Latin songs.Footnote ³

Rather than taking responsorial – solo and choral – performance for granted, in this chapter I reexamine the ways in which the refrain functions as a moment of literal or figurative coming together in song. Drawing on the Latin refrain song’s self-theorization of refrains and refrain forms through rhetoric, grammar, form, and musical texture, I illustrate how poets and composers embedded ideas about performance, community, and communal participation in their compositions. Crucially, the songs themselves define communities rooted in the structures of the church; singers and songs belong to nested and hierarchical communities of clergy, choir, rank, and in some cases age. Working within these communities, the efforts of scribes and compilers contribute additional layers of information to the meaning and performance of refrains, with rubrication and textual cues offering rare insight into specific and relatively demarcated contexts for the refrain song’s responsorial performance.

A reevaluation of assumptions around the vocal and choreographic performance of the Latin refrain song reveals the nuance of performance practices around refrain forms. Most significantly, refrains and refrain forms did not have to be sung responsorially in order to invoke the voice of the community, nor was dance inherent in their meaning and realization, by contrast to the frequently applied label of “clerical dance songs.” Whether sung by soloists or a choir, standing still or moving, the Latin refrain cultivated a sense of community and belonging that buttressed its multivalent cultural, theological, musical, and poetic functions and made it possible for the structural repetition of a unit of text and music to accrue a range of meanings in performance and interpretation.

Song, Refrain, and Singing with One Voice

The fifteenth-century refrain song Presens festum laudat clerus, sung at Matins in an Austrian abbey and copied uniquely into the St. Pölten Processional (fol. 9^r–v), provides internal evidence of its vocal performance. It begins in its first strophe with a call to the clergy to sing sweet melodies “with one voice” (“una voce”). This cue for a unified voice leads into the refrain, which repetitively and rhetorically plays with the stem “mir-,” or “wonderful”:Footnote ⁴

Repetition saturates this song beyond the refrain; the five strophes share portions of repeated text (such as “bis intonent eia et eia” in strophes 2 and 4) as well as similar exhortations to sing and rejoice on the feast day, named in the third strophe as Christ’s Nativity. While the rubrication of the song, “conductus infra noctem,” suggests a functional role in the Matins liturgy adjacent to a reading (lectio), the final strophe of the song formulaically cues the lector to begin following a final iteration of the “wonderful” refrain.Footnote ⁵

A responsorial performance for Presens festum laudat clerus is overtly cued in this final strophe as the whole choir is directed to respond (“totus chorus sic respondeat”) with the refrain. In previous strophes, implicit performance cues frame the refrain as a vocalic choral response, a semantically simplistic manifestation of the joyous song repeatedly elicited in the strophes by commands to let “everyone,” “all mankind,” and the “church” express their joy. The framing of the first strophe, however, is key: the clergy is asked to sing their sweet, sincere melodies “with one voice” (“una voce”). While in literal terms this could refer to the monophonic (literally unison) performance of Presens festum laudat clerus, the expression originates in Daniel 3:51 with the three boys who were thrown into the furnace singing the canticle Benedictus es “as with one mouth” (“quasi ex uno ore”).Footnote ⁶ The doctrine of singing quasi una voce became integral to how liturgical chant and, indeed, the aural performance of sacred texts more generally, was understood by the Church Fathers and liturgical commentators throughout the Middle Ages.Footnote ⁷ As Clement of Rome states, “we too assembled with one accord should earnestly cry out without ceasing to him as with one voice”;Footnote ⁸ in this he is followed by Clement of Alexandria, who asserts that “the union of many, which the divine harmony has called forth out of a medley of sounds and divisions, becomes one symphony.”Footnote ⁹

The doctrine of quasi una voce pertains both to performance practices – specifically monophonic chant – and to the metaphor of performance. As Judith Peraino writes regarding Augustine’s views on devotional song, the doctrine leads to “many bodies coming together as one voice, producing one sound, and becoming one body (the Church).”Footnote ¹⁰ The music that results – namely, chant – is seen as a “solution to the dangers of song, for the words are completely controlled … by preventing individuality.”Footnote ¹¹ Univocal song and praise are a way for the church to control and divert the individual voices of a community (i.e. the church) into a shared expressive goal. And while chant as a whole can be conceived as a form of singing una voce, liturgical responses and refrains (e.g. “amen”) are especially rich moments of explicit participation by the totality of the community, bringing a greater number of voices into alignment literally and figuratively.

The example of liturgical responses, however, brings up an important point. Singing quasi una voce, as if with one voice, does not necessarily mean everyone (choir, school, clergy, congregation) participates equally or at the same time; as Richard Crocker observes, “more than one Christian cannot sing literally with one voice; the doctrine is some kind of spiritual direction for ‘performance practice.’”Footnote ¹² To sing with one voice is to imagine a collective voice being expressed by singers regardless of their number. Steven Connor employs the term “chorality” to think about these sorts of collective speech and musical acts, which include prayers, statements of fealty, protest songs, and learning songs, and how what he terms the “choric voice” most often builds solidarity – a sense of singing as if with one voice – within communities, however delineated.Footnote ¹³ Importantly, chorality does not simply mean a “collective voice,” or many voices coming together, but rather a voice that is understood as representing the whole.Footnote ¹⁴

The Latin refrain and refrain song resonate with Connor’s “choric voice.” Mirroring the ideal of liturgical chant to produce a single, unified voice for the church, the refrain song is constructed around the voice of the community and its rhetorically anticipated, if not actual, participation cued by the refrain. Whether or not an entire choir responds with the refrain of Presens festum laudat clerus as cued in the final strophe, the entire Augustinian community in St. Pölten – and, indeed, the entire church – is ventriloquized through performance in the collective voice of the song and its responsorial form. Although a collective yet unified voice resonates throughout Latin song, chant, and liturgical tropes, the refrain is a specific signal. Repeating over the course of a song and, in some cases, across songs, refrains belong to a register of voice capable of eliciting both real and imagined community participation.Footnote ¹⁵

A Grammar of Performance and Plurality

The poetry of the refrain song – grammar, language, rhetoric – illustrates through the use of plural and collective nouns and verbs the ways in which communities of singers are linked through a unified voice in song. Like many medieval songs that range across language and register, the Latin refrain song is self-reflexive; references to song and singing, as exemplified in Presens festum laudat clerus, are commonplace.Footnote ¹⁶ Interest in song itself most often arises around the issue of a song’s performance, whether imagined or actual. Initial strophes of Latin song often offer an assertion of the implied performing forces; as I discuss in Chapter 1, incipits can also position the performance of song within the calendar and seasonal year. Accounting for elements of performance in the language and grammar of refrain songs reveals a repertoire of songs focused on communal song, repeatedly evoking, if implicitly, the doctrine of singing una voce. This manifests in two ways in refrain songs, often, albeit not exclusively, within the confines of the refrain itself: first, through repeated returns to the idea of a community, and second, through a consistent emphasis on plurality by means of the widespread use of the plural first-person and collective nouns. The poetry of Latin refrain songs creates a link between performance and the expression of a unified voice of the church through consistent terminological and grammatical choices.

The contribution of language to the creation and maintenance of feelings of community is not a new concept but has been well explored by cultural and social historians, as well as musicologists. Peter Burke in particular has examined the role of language in the process of community formation for the early modern period.Footnote ¹⁷ His insights regarding the role of Latin as a shared, albeit second, language within communities across Europe and the significance of plural grammar as signifier of identity have relevance for the high Middle Ages and the lyrics under consideration here. The resolute Latinity of these songs situates them within literate, clerical, and monastic milieus, and gives them the potential, in Burke’s words, of creating “a sense of distance from everyday life and a sense of universality” as well as “a sense of tradition.”Footnote ¹⁸ Close associations with the Latin liturgy (including direct and indirect borrowings of prayers, chants, biblical passages, etc.) further enhance Latin song’s ties with the linguistic communities created and supported by the church. The largest community circumscribed by the Latinity of Latin song is, consequently, that of the Catholic church writ large. This is not surprising; the spheres in which Latin song was principally cultivated were also ones in which Latin served as the key mode of communication and education within and without the liturgy.

While language delineates the church at the broadest level, the refrain song repeatedly evokes performance by groups of singers through the consistent use of collective nouns and first-person plural verb forms attached to self-referential descriptions of performance. The systematic use of the first-person plural sets refrain-form songs apart from the grammatical practices of Latin song more broadly in ways that have yet to be examined.Footnote ¹⁹ The final two fascicles in F serve as a case study for the way in which grammatical choices differ between songs with and without refrains. The collection of conducti in Fascicle X in F begins with a highlighting of singular voices in Homo natus ad laborem (fol. 415^r–v), attributed to Philip the Chancellor.Footnote ²⁰ In a dialogue that alternates between the voice of the body and soul by strophe, Homo natus ad laborem dramatically privileges the subjectivity of individual voices. By contrast, an Easter song, De patre principio, begins Fascicle XI (fol. 463^r). Its five strophes and two-part refrain are markedly different in tone and grammar than Homo natus ad laborem, emphasizing the unified voice of the church, as the first strophe illustrates:

The Easter rondellus repeats “gaudeamus eya” twice in each strophe as part of the two-part refrain, whose subject is identified as the church (“ecclesia”); the song is not the expression of an individual, subjective voice, but instead reflects the voice of the collective church celebrating Christ’s Resurrection. De patre principio sets the grammatical tone for the entire final fascicle of F. Only two songs out of sixty consistently employ a singular voice; the remaining fifty-eight are grammatically unified around either the first-person plural or collective nouns (e.g. “contio”) that take the third-person singular.Footnote ²¹ By contrast, the conducti of Fascicle X, beginning with Homo natus ad laborem, showcase a wider variety when it comes to verbal person and number, with a notable emphasis on the first-, second-, and third-person singular rather than plural forms. Fascicles X and XI in F offer a striking opportunity to compare the poetic grammar of songs that take different forms and subject matter. In addition to ordering according to form and the liturgical calendar, the refrain songs of Fascicle XI also show an undeniable poetic and grammatical tendency toward voicing the collective rather than the individual.Footnote ²²

Throughout devotional Latin refrain songs, certain verbs occur more often than others in the first-person plural imperative, most significantly “gaudeamus,” “exultemus,” “iubilemus,” “celebremus,” and “psallamus” (let us rejoice, exult, sing out joyously, celebrate, and sing, respectively). Over half the time, these calls to the community can be found in the refrain itself, as in De patre principio. In three songs, “gaudeamus” comprises the entirety of the refrain, marking it as a refrain word, or response, comparable to the “alleluia” or “amen.” Like these liturgical refrains, “gaudeamus” also implicitly supports group participation. The three songs with the “gaudeamus” refrain were closely linked to, or performed within, the liturgy.Footnote ²³ Voce resonantes is a neumed Benedicamus Domino song-trope transmitted in the thirteenth-century troper and songbook from a Swiss cloister, Stuttg, also known as the Stuttgart Cantionale (fols. 52^v–53^r and 74^v–75^r); Salve festa dies veneranda is rubricated “conductum sancti Iacobi” in the twelfth-century troped liturgy for St. James in the Codex Calixtinus (fol. 132^r–v); and Exultantes in partu virginis survives unnotated in the twelfth-century versarium St-M A and also serves as a Matins conductus for the Feast of the Circumcision in sixteenth-century manuscripts from Le Puy Cathedral.Footnote ²⁴ In each case, the invariable, single-word refrain is treated differently, showcasing a range of approaches to the composition and performance of refrain forms and, importantly, the potential of the refrain to be performed by a soloist or larger ensemble.

In Voce resonantes, “gaudeamus” repeats between each line of the Benedicamus Domino song-trope, the brevity of the refrain at odds with the way it textually dominates the brief work:Footnote ²⁵

The poem includes repeated references to song and singing, the incipit immediately signaling its performance by means of a “resonant voice.” At times, the “gaudeamus” refrain appears to fit grammatically into the poem; at other times, it sits apart from the simplistic text of praise and thanksgiving that tropes the Office versicle. The structure of the poem is uncommon, but not unique among tropes; within Stuttg alone, Marian refrains are inserted across liturgical chants in a pattern similar to the “Gaudeamus” refrain in Voce resonantes.Footnote ²⁶ Significantly, in Voce resonantes the “gaudeamus” refrain is textually identical throughout, but musically variable, alternating between two melodic profiles, one brief and neumatic and the other melismatic (see Figure 3.1). The refrain mirrors the musical style of the nonrefrain material, matching it in complexity and style; the neumes make it challenging, however, to discern whether melodic repetition occurs among the textual elements.

Figure 3.1 Neumed “gaudeamus” refrain in Stuttg, fol. 75^r. Reproduced by kind permission of the Württembergische Landesbibliothek, Stuttgart

Considering the weight of the refrain in terms of repetition and its frequent melismatic setting, it is unclear what performance practices the “Gaudeamus” refrain entailed. The refrain text itself, “let us rejoice,” might provide some slight grammatical support for the implied choral versus soloistic performance of the refrain. The liturgical context is also noteworthy here. The Benedicamus versicle is comprised of both a verse, typically performed by two or more soloists, and a choral response; the structure of the verse and response is outlined in the text of Voce resonantes above by the citation of Benedicamus Domino midway through and “Deo dicamus gratias” at the end. While the refrain runs throughout the song-trope, the implied responsorial format may reflect the versicle’s associated liturgical practices – certainly, the choir would have been ready to participate.

A conductus for St. James in the Codex Calixtinus, Salve festa dies veneranda, shares with Voce resonantes the textually identical yet musically variable repetition of “gaudeamus” following each verse line, yet with details regarding performance added by the rubricator (see Figure 3.2):

Figure 3.2 Codex Calixtinus, fol. 132^r–v, Salve festa dies veneranda with “gaudeamus” refrain.

Copyright © Cabildo de la Catedral de Santiago de Compostela – all rights reserved. Partial or whole reproduction is prohibited

The musical settings alternate with the verses of variable lengths, lengthening or shortening as needed. Akin to Voce resonantes, the refrain shifts between two melodies, characterized by descending versus ascending figures on “gau-” of “gaudeamus” (the scribe gradually abbreviates the music of the refrain moving down the folio).

By contrast to Voce resonantes, however, the scribe in Codex Calixtinus includes a performance indication directly before the first iteration of the refrain: “the boy repeats this [refrain], going between two singers” (“puer hoc repetat pergens inter duos cantores”). The choral performance of the refrain suggested by its plural form and function as a response is rejected entirely – instead a single boy is seemingly directed to repeat it between the verses sung by the two singers. In Salve festa dies veneranda, “gaudeamus” is a boy soloist’s response, not the expected communal interjection implied by the first-person plural form. In the Codex Calixtinus, Salve festa dies veneranda concludes a series of five conducti with liturgical functions as Benedicamus Domino tropes and lectionary introductions, functions suggested by rubrics and their poetry.Footnote ²⁷ All five employ refrains, and three include similar rubrication to Salve festa dies veneranda, prescribing the refrain’s solo performance by a boy. The performance practices of refrain songs in the Codex Calixtinus thus invert the expectations of the choral refrain, although not the responsorial framework – the boy soloist still responds to other singers. Moreover, by repeatedly proclaiming “let us rejoice,” a single boy takes on the voice of those around him; in the case of the Codex Calixtinus, we might imagine this would involve not only clergy, but possibly also the imagined voices of pilgrims.

A conductus sung during the first Nocturne of Matins during the troped Circumcision liturgy transmitted in Le Puy A, Exultantes in partu virginis, presents an entirely different structure and performance situation than Voce resonantes and Salve festa veneranda:Footnote ²⁸ The one-word refrain is sung between longer strophes and, in Le Puy A alone, the texture changes between strophe and refrain, with the refrain functioning as a monophonic interjection in a polyphonic setting:Footnote ²⁹

The single-word refrain, combined with a monophonic setting, underscores the responsorial structure of the song, and may have suggested the deliberate juxtaposition of polyphony and monophony for a composer who added a polyphonic line in an appendix at the end of the manuscript; as it occurs initially in Matins, the entire work is monophonic.Footnote ³⁰ The invariable musical setting of the refrain is short and simple, outlining a descent of a third, including a lower neighbor extending the gesture to a fourth, by means of small clusters of pitches (see Figure 3.3). The voices of the choir come together both literally in monophony and figuratively here on “let us rejoice,” the musical setting affirming the communal message of the refrain. Alternating textures work at the poetic as well as musical level in Exultantes in partu virginis to convey meaning in the ritual act of celebrating Christ’s birth. Whether the soloists come together in unison for the refrain or they are joined by the choir (and other conducti sung at Matins do specify the choir, chorus, in rubrics), the “gaudeamus” refrain in Exultantes in partu virginis functions as a moment of singing quasi una voce.

Figure 3.3 Le Puy A, fol. 28^r, monophonic “gaudeamus” refrains.

Reproduced by kind permission of the Bibliothèque municipale, Grenoble

A variety of performance practices are implied by text, form, and rubrication among songs employing “gaudeamus” as a refrain, ranging from the unison response of Exultantes in partu virginis to the explicitly singular performance of the refrain in the Codex Calixtinus. Less clearly attuned to performance, Latin refrain songs invoke three particular communities as the “we” in their plural grammar. The communities who rejoice and sing together are nearly always identified as cleri and pueri (clergy and choirboys); more generally as a contio/concio (assembly, company, or congregation); and, most broadly, as ecclesia (the church). The first category pertains both to the clerical milieu of the repertoire’s creation and dissemination and to the performative aspect of collectivity, while the latter two are far more general in scope, although always bounded by the limits of the church. While this terminology is not surprising given the religious milieus and institutions in which Latin song was created, the usage of such language in refrain songs specifically frames ideas of performance. Cleri, contio, and ecclesia are not passing references or invoked as subjects alone, but are positioned in the poetry (and in some cases in rubrics) as the agents responsible for producing song on behalf of their community.

When clergy are positioned as singers and listeners in Latin song, they are typically identified by the plural cleri or the diminutive clericuli (choirboys).Footnote ³¹ The clergy is seldom far removed from the larger assembly of which they are a part, leading to the linking of cleri and concio, as in phrases from refrain songs including “letatur cleri concio” or “clericalis concio.”Footnote ³² Returning regularly throughout individual songs and across the repertoire, the refrain frequently serves as the locus for the clergy’s self-identification. Communities of faithful, moreover, are rarely silent; instead, their verbal actions emphasize the sounding of devotion: the clergy resonet, gaudeat, psallat, congaudeat, and exultet.Footnote ³³ Notably, verbs are nearly always in the hortatory subjunctive mood, expressing the poet’s desire for communities to convey their devotional joy through sound and song.

Since examples of this formulaic combination of collective noun and subjunctive verb are abundant, a small selection of songs readily illustrates how this linguistic practice works poetically. The first is an Aquitanian versus, Promat chorus hodie, often described as a formal precursor to the French rondeau and Latin rondellus due to its internal refrain:Footnote ³⁴

The expected subjunctive mood appears twice in the first strophe of four in the incipit and in the refrain. The imperative mood also appears in the refrain, varying the repetition of the verb psallere between “psallite” and “psallat.” From the initial line, the versus frames itself as a communal outpouring of praise (for Christ, although this is not explicit until the second strophe). The “chorus” is immediately made central to its imagined performance as “bringing forth” joyous song, underscored by the repetition of “contio” in each strophe thanks to the internal refrain. The final three refrain lines bring the sound back to the contio once again, with the psalmic verb providing the action for the assembly.

The second example is not a single song, but rather a refrain that appears, similar to “gaudeamus,” in different songs and sources: gaudeat ecclesia. As a generic expression of the church’s joy, “gaudeat ecclesia” occurs dozens of times in Latin poems, not always functioning as a structural refrain, but instead a formulaic invocation of the church. Altogether, nearly one dozen Latin refrain songs utilize “gaudeat ecclesia” as all or part of their refrain (including the variation “congaudeat ecclesia”).Footnote ³⁵ The refrain of a Nativity song in the Moosburger Graduale, Gaudeat ecclesia inventa, fol. 232^v–233^r, intertwines the phrase with vocalic repetition and an expanded description of the church in question: “Eya et eya, | regia egregia, | gaudeat ecclesia!” (“eia and eia, let the splendid royal church rejoice”). A well-known song employing the vocabulary of music theory in honor of Christ’s Resurrection, Diastematica vocis armonia (Later Cambridge Songbook, fol. 2 v [1v], twelfth century), features a longer refrain that concludes with “gaudeat ecclesia”:Footnote ³⁶

In a similar fashion to Gaudeat ecclesia inventa in the Moosburger Graduale, the vocalic repetition of “eya” precedes the church’s rejoicing; unlike the later cantilena, this earlier twelfth-century song includes a reference to the people who comprise the ecclesia: the “plebs fidelis” (faithful people).

Collective nouns and verbs of identity and collectivity form a grammatical and poetic trope throughout Latin refrain songs, one heightened and emphasized through the repetition of the refrain itself. The deliberate choice of the first-person plural (as well as collective nouns) is part of the delineation of communities, solidifying a sense of togetherness that might already exist, while also serving to redefine and fortify boundaries. In the context of song, this community formation and/or maintenance occurs within the space of the lyric and its performance, with the potential to interact with performance practice.

In other words, pluralizing language in song may reflect a song’s performance by a plurality of voices within a community – who better to sing “let us rejoice” than the community as a whole? Yet the language of song does not always neatly connect to its performance forces or practices, as alluring as the language of refrains might be in this regard. The rubric of Salve festa dies veneranda makes this point by asking a single boy to represent the community in his plural “we” refrain. Grammar and language construct the actual as well as imagined performance of this repertoire in ways that always refer to the plural and collective. When one boy sings “gaudeamus,” listeners understand that he performs a kind of ventriloquism of the group to which he belongs. Form, manuscript context, and poetics work together to convey information about song’s performance in time and space, underscoring how song and refrain functioned as an expression of a community’s unified voice, regardless of whose voices, and how many, were actually heard.

All as One and Many Together: Musical Settings of Refrain Songs

Musical texture plays a significant role in shaping the performance of refrains. Above all, the combinations of voices in either monophony or polyphony provides a commentary on the performance of refrain forms that speaks to the question of responsorial performance and the communal identity of the refrain. For Latin song, the greater part of the repertoire is monophonic; only a small percentage survives for two and three voice parts. Scholarly assumption has been that the performance of the repertoire generally involved a choir or group of soloists, by contrast to the soloists required for the frequently more complex polyphonic conductus.Footnote ³⁷ However, not all Latin refrain songs lend themselves to a choral performance in which the ideal of singing quasi una voce can be achieved. The survival of polyphonic settings of Latin refrain songs, alongside several exceptional works that alternate monophonic and polyphonic textures between strophes and refrains, offers different perspectives on the performance of the Latin refrain as an expression of the choric voice and as a soloist’s art.

Several songs alternate polyphonic and monophonic textures between strophe and refrain, creating a sonic contrast between the two formal elements.Footnote ³⁸ Although examples are scarce and predominantly date from the fourteenth century and beyond, four extant songs set strophes polyphonically and interpolate monophonic refrains. Three are preserved in Swiss sources chiefly transmitting liturgical contents, tropes, and some polyphony, while one is found in a French source for the Feast of the Circumcision. Two are uniquely preserved in the Engelberg Codex, Unicornis captivatur (fol. 150^v–152^r) and Ovans chorus scholarium (fol. 153^r–153^v); a third is the more widely transmitted Nove lucis hodie in SG 392, pp. 88–89 (see Chapter 4); and the fourth is Exultantes in partu virginis (fols. 27^v–28^v and 162^r) in the Le Puy Feast of the Circumcision sources, discussed in the previous section.Footnote ³⁹ Only Nove lucis hodie is set for three voices in SG 392 (notated successively); the others feature strophes for two voice parts. The two songs in the Engelberg Codex are notated in score format, and the lower of the two voices in Exultantes in partu virginis is transmitted in an appendix to Le Puy A. In all cases, refrains are set monophonically. The alternation of texture between strophes and refrain is unambiguous across all the sources, further enhanced by scribal cueing of the refrain in the Engelberg Codex and SG 392.

The transition from polyphonic strophes to monophonic refrain is exemplified by Unicornis captivatur in the Engelberg Codex, in which the refrain is visually marked by a textural change and textual cue in red ink. The layout is noteworthy, since the upper voice drops out before the end of a system and the lower line alone continues, leaving the upper stave blank (see Figure 3.4). A textual abbreviation of et repetitur cues the monophonic refrain, and capital letters emphasized with red figuration further distinguish refrain from strophe.

Figure 3.4 Engelberg Codex, fol. 151^r, Unicornis captivator, from “[me]dicatur sauciatus” in strophe 1 to “misera” in strophe 2.

Reproduced by kind permission of Stiftsbibliothek Engelberg

The differing performing forces for strophes and refrain are not indicated; the monophonic refrain could have been performed by the same singers responsible for the polyphonic strophes. The coming together as one voice in the refrain is one that could occur not only by means of a choral interjection, but also by the coming together in unison of previously independent polyphonic voices.

The musical setting and poetry of the refrain in Unicornis captivatur work together to create a moment of song framed within a well-known collective response, “Alleluia.” The poetry of Unicornis captivatur as a whole is striking, employing animal allegories familiar from medieval bestiaries to construct a poem about Christ’s Resurrection. Unambiguous animal signifiers of Christ, real or mystical, are introduced in quick succession throughout the poem: The strophes name the unicorn, pelican, phoenix, hydrus, lion, and basilisk, while the refrain names the lamb and the lion.Footnote ⁴⁰ The poem rehearses the traditional allegorical meanings of the animal symbols in turn, each strophe narrating the period between the Passion and Resurrection. The strophic narrative is bounded by the refrain, which provides a two-part summary of Easter – first Christ as the dying paschal lamb and then second as the risen, victorious lion. The summative effect of the refrain within the Easter narrative is enhanced by its imperative to sing alleluia. Refrains in narrative song, as I show in Chapter 2, often invoke song; Unicornis captivatur further invokes the singing of a familiar liturgical refrain, “Alleluia,” which was removed from the liturgy during Lent and reintroduced on Easter Sunday. The textural shift from polyphony to monophony, in other words, mirrors the textual shift from narrative to invocations of song and refrain.

A New Year’s song transmitted in St-M A and the Le Puy Feast of the Circumcision sources, Annus novus in gaudio, upsets the textural affiliation of the refrain with monophony exemplified by Unicornis captivatur by setting the refrain polyphonically alternating with monophonic strophes.Footnote ⁴¹ Much discussed due to the successive notation of its polyphonic refrain in St-M A, Annus novus in gaudio stands as the sole example in either manuscript context, and more broadly among medieval Latin songs, to invert the expected texture of refrain versus strophes (see Example 3.1).

Example 3.1 Le Puy A, fols. 1^v–2^r and 171^r, strophes 1–2 and refrain of Annus novus in gaudio

Annus novus in gaudio is unquestionably a song about singing, framed as the simultaneous celebration of the New Year (understood as the Feast of the Circumcision on January 1) and its octave and the cantor, leader of the choir (see Chapter 1). The strophes follow the model of the first two underlaid in Example 3.1, beginning with a statement on the New Year and then prescribing sung praise in honor of the cantor. The message of the strophes is emphasized in the refrain, which asks the assembled worshippers to raise their voices in honor of the cantor. Performative language characterizes the entirety of the poem as does the collective voice, making the playful move from monophony to polyphony a musical complement to the poetry. The singers make their voices heard (“sonantia”) in a special way when the refrain enters, joining together (“concurrent”) in harmony, that is, polyphony. While the monophonic “gaudeamus” of Exultantes in partu virginis utilizes monophony to signify communal rejoicing and singing quasi una voce, in Annus novus in gaudio the conjoining of voices in polyphony is deployed to express the collective sentiments of the refrain.Footnote ⁴²

Apart from the unusual witness of Annus novus in gaudio, a preference emerges for monophonic over polyphonic settings of refrain forms, brought into relief through the alternation of musical textures. In some cases, focusing on musical texture reveals the seeming elimination of repeating structures in polyphonic versions of monophonic refrain songs. While monophonic textures do not universally correlate to communal performance, simpler, monophonic settings of refrains increase the potential for group, rather than solo, performance. Variable texture works likewise to highlight an affiliation between monophony and refrains, showcased through alternatim performance based on the division between strophe and refrain. The interjection of the refrain offered composers, in many cases, a poetic “break” that could be exploited texturally by shifting from, more commonly, polyphony to monophony, or more rarely, monophony to polyphony.Footnote ⁴³

Only a small number of Latin refrain songs were set polyphonically for two or three voices across strophes and refrain.Footnote ⁴⁴ Perhaps most surprisingly, only one Latin rondellus survives in a polyphonic setting, Luto carens et latere (W1, fol. 80^r); all other polyphonic textures occur in settings of strophic songs with refrains. Luto carens et latere, moreover, is a unique case study.Footnote ⁴⁵ Extant in five sources, its text appears alone in Tours 927 (fol. 18^v), with one voice in F (fol. 463^v), LoB (fol. 48^r), and Bord 283 (fol. 134^v), and for three voices in W1 (fol. 80^r); all the polyphonic versions share their bottom-most voices with the monophonic settings. Notably, in the one-voice and unnotated sources, a minimum of three strophes are included, with eight surviving in Bord 283. As evidenced by the monophonic sources, Luto carens et latere is a rondellus, with one line of the refrain inserted between the first two lines of the strophe. Yet in its polyphonic form in W1, only the first strophe is copied.

The continuation of the text from a previous conductus in the upper right-hand corner of the folio and the space below makes it clear, however, that the scribe had space to copy additional strophes if desired to the right of the systems (see Figure 3.5).Footnote ⁴⁶ Throughout W1, multiple strophes of polyphonic songs are typically included, making the single strophe of Luto carens et latere an outlier. Either the scribe of W1 did not have access to a version with multiple strophes or they deliberately chose to exclude further strophes.

Figure 3.5 W1, fol. 80^r (73^r), Luto carens et latere.

Reproduced by kind permission of the Herzog August Bibliothek Wolfenbüttel

The effect of W1’s poetically foreshortened polyphonic transmission of Luto carens et latere is an effacement of the rondellus form. As transmitted in W1, the reoccurrence of lines 2 and 4–6 in subsequent strophes is elided, leaving only the repetition of line 1 as line 5. Moreover, while the lower voice repeats the melody of the A-line of the refrain, the upper voices present different music at the repetition of the refrain text (lines 1 and 5), further obscuring the repetition of the text.Footnote ⁴⁷ In its polyphonic form, Luto carens et latere resembles any other short, rhymed, rhythmical poem set for multiple voices in W1 and elsewhere. In performance, it is possible that the shared tenor line with its repetitive patterning and the repetition of text may have clued the listener into the rondellus form of the song that becomes obvious when further strophes are present; it is also possible that singers could have supplied additional strophes as they performed.Footnote ⁴⁸ As it stands in W1, however, Luto carens et latere resembles less a strophic refrain song than a through-composed setting of a short poem.

Assuming the scribe of Luto carens et latere in W1 deliberately left out additional strophes, which seems likely considering the wider – and in the case of Tours 927, earlier – transmission of the song, did the textural shift to polyphony motivate the obscuring of the underlying refrain form? In other words, does the polyphonic rendering of Luto carens et latere in W1 align it better with other polyphonic conducti since refrains were more commonly the purview of chorally performed monophony rather than soloist-performed polyphony? Of the over 400 songs listed in the Appendix, fewer than 40 (depending on source and reading) survive in polyphonic settings for part or all of the poem. Sources for the versus transmit the greatest number of polyphonic refrain forms, and in these the refrain is typically cued consistently. Moreover, the transmission of polyphonic refrain forms is often complicated by absent or inconsistent cueing of refrains, as well as inclusion of only one strophe or a single iteration of the refrain, as in Luto carens et latere.

The refrain-form lament Eclypsim patitur exemplifies the ambiguities around polyphonic refrain forms in two particular sources, W1 and F.Footnote ⁴⁹ By contrast to the transmission of Luto carens et latere, which survives monophonically in F with seven strophes and polyphonically in W1 with one, all four strophes of the two-voice Eclypsim patitur are included in W1 (fol. 110^r) (the first notated, and the following ones copied below the final system). At no point, however, is the refrain (beginning “Mors sortis”) indicated as such, nor is its return indicated textually after strophes 2–4. The only evidence of a refrain is the poetic and musical form itself in W1, which suggests by means of the contrasting scansion and musical setting of the refrain that it is indeed a refrain, and not the continuation of the strophe or an entirely new one.Footnote ⁵⁰ The situation is more obscure in F, where only one strophe and one statement of the refrain are included (fol. 322^v–323^r), making the refrain structure impossible to derive from this manuscript alone.

A similarly ambiguous situation obtains for many of the polyphonic conducti cum refrains transmitted in W1 and F, calling into question the very identification of polyphonic refrains in these polyphonic settings. Only a few polyphonic conducti include refrains and multiple strophes (eight between W1 and F), and not all cue the return of the refrain with any regularity, similar to Eclypsim patitur.Footnote ⁵¹ Works like Luto carens et latere are indicative of this ambiguity, the addition of voices and subtraction of strophes essentially erasing the refrain form. In the absence of extant monophonic versions that clearly indicate a repeating refrain, single strophes and single iterations of refrains in polyphonic settings can only be tentatively labeled refrain forms. The differing treatment of a single refrain form in monophonic and polyphonic settings points toward a correlation of monophony with specifically repeating refrain forms, as opposed to single strophes and refrains set polyphonically. The performance implications of such an association relate to choral versus solo performance – monophonic songs and refrains are more likely, albeit not exclusively, sung by groups whereas polyphonic settings, at times eliminating refrains, are more likely to be sung by soloists. Indeed, this resonates with Anonymous IV’s association of conducti without caudae, or musically less complex conducti, with performance by “minores cantores.”Footnote ⁵²

Shaping Performance: Strophic Refrain Songs and Rondelli

The noticeable lack of polyphonic settings of rondelli and the ambiguous example of Luto carens et latere in W1 points toward a larger divide in terms of performance and transcription between rondelli and strophic+refrain songs. Textual cues for refrains in manuscripts, including “chorus,” “repetitio,” and others, are exclusively associated with strophic+refrain songs, as opposed to rondelli. As I detail in the Introduction, all Latin refrain songs fall into one of these two camps, with strophic+refrain songs more common. Given the marked preference to indicate form and performance for strophic+refrain songs alone, what does this suggest concerning the performance of the Latin refrain song more generally?

Historiographically, the rondellus has been understood as inherently responsorial. Musical evidence for the responsorial performance of rondelli rests on the assumption that the soloist’s introduction of all the melodic material before the refrain enters familiarizes the choir with the melody – if the form is that of the early rondeau or rondet, aAabAB – thereby enabling a choral response. As far as this argument goes, the initial statements by the soloists would serve as a reminder to the chorus, who then sing their refrain (the text of which is not introduced by a soloist) to the same melody.Footnote ⁵³ Evidence for such a performance has been shaped by the reception of the repertoire over time and the gradual solidification of this perspective by means of contemporary performance and recording practices; no primary evidence of which I am aware, theoretical, manuscript, or otherwise, suggests that rondelli were necessarily performed alternatim between soloist and chorus.

Conversely, manuscript evidence does exist for the alternatim performance of strophic+refrain songs. Although the forms of refrains interjected between strophes vary more widely than those in rondelli since they lack the same structural constraints of syllable count and rhyme, rubrication in several manuscript sources points both to their identity as refrains (by means of cues such as “repetitio”) and their choral performance.Footnote ⁵⁴ In the cantionale of the Moosburger Graduale, which also furnishes a preface attributed to the dean of the song school, Johannes de Perchausen, in which he discusses the songs, textual cues for refrains are widely employed (see Chapter 1).Footnote ⁵⁵ In two Benedicamus Domino song-tropes, Christus vicit resurgendo (fol. 248^r–v) and Florizet vox dulcisonans (fol. 249^v–250^r), refrains are marked off according to both their refrain forms and performance. In Christus vicit resurgendo the cue for the refrain changes from “R.” to “chorus” midway through the song, before the third repetition of the refrain, as a page turn occurs. For the remainder of the song, the fourfold repetition of the liturgical refrain “alleluia” is cued with first “chorus” and then the abbreviation “chor” (see Figure 3.6).Footnote ⁵⁶ Nowhere does the scribe specify performing forces for the strophes. Instead, the “chorus” cue implies the potential for the nonchoral performance for the strophic material. The use of “chorus” as a refrain cue also speaks to the community addressed by Johannes in his preface – the songs were an offering to the clericuli of the song school, the very choir we could imagine singing the refrain in Christus vicit resurgendo.

Figure 3.6 Moosburger Graduale, fol. 248^v, Christus vicit resurgendo, excerpt.

Reproduced by kind permission of the Universitätsbibliothek, Munich

An identical refrain cue is attested to in the Seckauer Cantionarium, a mid-fourteenth-century liturgical book with a musical cantionale featuring neumed songs and tropes from the collegiate church in Seckau.Footnote ⁵⁷ The scribe of the songs and tropes in the Seckauer Cantionarium, unlike in the Moosburger Graduale, only periodically included rubrication and cues for refrains. Only one song, Stella nova radiat, includes thoroughgoing cueing of the refrain.Footnote ⁵⁸ Beginning “ergo novis laudibus,” the refrain’s choral performance is indicated by the red abbreviated “Chor.” for “Chorus” appearing prominently before its first iteration (see Figure 3.7).

Figure 3.7 Seckauer Cantionarium, fol. 186^r, Stella nova radiat.

Reproduced by kind permission of the Universitätsbibliothek, Graz

The case of Stella nova radiat, moreover, becomes more interesting after the page turn. Only the first occurrence of the refrain on fol. 186^r is preceded by a prompt for the choir; while it is not uncommon for cues to disappear in the course of a song’s inscription (see Chapter 4), the continuation of Stella nova radiat onto fol. 186^v also marks a change in the refrain’s prescribed performance. From “chorus” the scribe switches to “pueri” for the remaining iterations of the refrain (see Figure 3.8). This is an unusual moment in which a scribe not only changes a cueing strategy, but also clarifies the performance forces in question. “Chorus” could mean boys, but not necessarily; by using “pueri” the precise subgroup of performers is made explicit. Given the possible connections of the Seckauer Cantionarium to the bishop’s school in Seckau, it may be that the “pueri” in question were, like Johannes’s clericuli in Moosburg, local students and choirboys.

Figure 3.8 Seckauer Cantionarium, fol. 186^v, Stella nova radiat continued.

Reproduced by kind permission of the Universitätsbibliothek, Graz

A final example is an early sixteenth-century antiphonary produced for the Monastery of St. Nicolas in Freiburg and whose contents reflect the use of the Lausanne diocese.Footnote ⁵⁹ A liturgical source, the antiphonary includes accretions in the form of songs rubricated as prosae that function as Benedicamus Domino substitutes. On Vespers for the Feast of St. Stephen (December 26), the somewhat widely transmitted refrain song and Benedicamus Domino song-trope Dulcis laudis tympano is included along with performance instructions that place boys (pueri) in musical conversation with the choir (see Figure 3.9).Footnote ⁶⁰ A rubric precedes the song, reading “in place of the Benedicamus Domino, the boys sing the following prosa” (“Loco de benedicamus domino pueri canunt prosam sequentem”). Another rubric appears before the refrain, beginning “Oy mira virtus stephani,” directing the choir to sing it as a response (“Chorus respondet”). After the refrain, a third rubric appears, this time directing the singers (boys and choir) to continue as follows for the remaining verses, until after the final verse when the choir sings the conclusion of the Benedicamus Domino versicle, “Deo dicamus gratias alleluia” (“et sic de versibus consimilibus; post ultimum versum Chorus respondet”).

Figure 3.9 Antiphonarium Lausannense III.3.1, p. 103, Dulcis laudis [Dulces laudes] tympano.

Reproduced by kind permission of the Archives de l’État de Fribourg

The responsorial framework of the Benedicamus versicle is reflected in a detailed way in the performance structure outlined for Stella nova radiat. Compared to most items in the antiphoner for the Feast of St. Stephen, and indeed for most other feasts, greater attention is given to the performance of Dulcis laudis tympano. A striking aspect of this example, moreover, is that boys were tasked with singing the strophes, while the whole choir sang the refrain – a contrast to the boy soloist’s performance of the refrain in Salve festa dies veneranda in the Codex Calixtinus and Stella nova radiat in the Seckauer Cantionarium.

The question remains whether the performance practices indicated by scribes in these examples, all dating from the fourteenth century and later, can be applied to other sources and repertoire. Notably, the manuscripts just discussed transmit more than a single Latin refrain song or song-trope, but scribes did not systematically indicate the choral or group performance of refrains in these. Is the absence of similar performance indications for other Latin refrain songs merely an accident or oversight? What was special about certain songs that demanded clearer performance directions? Significantly, refrain songs that include scribal performance directions each carry liturgical designations. In these examples, Christus vicit resurgendo and Dulcis laudis tympani are Benedicamus Domino song-tropes, and Stella nova radiat is rubricated for performance “in die nativitatis Domini super magnificat,” suggesting its performance at Vespers before the Magnificat.Footnote ⁶¹ The liturgical orientation of these songs is noteworthy since similar performance indications, whether referring to the choir, or even boys, were not uncommon in liturgical manuscripts, especially for special feast days.

Scribes seemed to have adapted practices used in copying chant and applied them to devotional song, with the intention of clarifying the performance of songs added, in many cases, to the Office. Conversely, rubrics or annotations specifying performing forces are absent for songs and sources that lack liturgical ties. While this does not mean comparable performance practices could not have been shared with Latin refrain songs absent liturgical ties, it does mean that, as far as scribal intentions are concerned, details concerning the performance of refrain forms were only deemed important to include when songs were integrated into the liturgy. Rondelli, significantly, are all devotional but never explicitly incorporated into the liturgy and, consequently, never received scribal intervention in terms of their prescribed performance. In manuscripts such as F, scribes were not concerned with outlining the performance of rondelli, or other Latin songs, leaving little material evidence concerning possible performance practices.

Yet what these scribal cues offer is a glimpse into the variety of performance practices around strophic+refrain songs as well as into the specific communities tasked with singing Latin song in liturgical contexts. Boys, whether junior clergy, choirboys, or students, were clearly the intended performers for a portion of the refrain song repertoire, their vocal engagement overtly dictated by scribes. Moreover, for strophic+refrain songs, the balance of poetic and manuscript evidence points toward the choral performance of refrains. When exceptions occur, as with the single boy in the Codex Calixtinus or the polyphonic refrain of Annus novus in gaudio, it is nevertheless possible to point to the ways in which the poet, composer, and scribe through language, grammar, and form intended the refrain to be heard as a reflection of the choric voice, the collective voice singing una voce.

Reimagining the Clerical Dance Song

The ability of the refrain to participate through performance, grammar, poetry, and music in implicit and explicit discourses of community does not account for a central strand of scholarship that sees the Latin refrain as generating chorographic meaning in performance. Scholarship on the Latin refrain song as early as the eighteenth century has steadfastly connected it to the devout circling, stepping, and twirling choreographies of religious men and women across Europe, repeatedly privileging its lyric-choreographic meaning. As the general absence of dance in this chapter and elsewhere in the book suggests, however, movement was not a central performance referent for the Latin refrain song on the part of poets, composers, scribes, and compilers. Yet the association of the Latin refrain song with dance has extensive roots and, in some cases, seemingly strong evidence that supports the “clerical dance song” label. The strongest evidence has been historically located in the refrain songs of the final fascicle of F, initially argued by Yvonne Rokseth in 1947 and reiterated in a half-century of subsequent scholarship.Footnote ⁶² I turn to this fascicle now in order to reevaluate the lyric-choreographic identity of the refrain songs of F and align their perceived choreographic potential with their vocal realization and invocation of community.

The collective identity of the songs in the eleventh fascicle of F is strikingly signaled on the initial folio by a historiated initial depicting a group of tonsured clerics. The five clerics appear to hold hands, or, more accurately, link fingers, with feet turned out and positioned as if in motion, in parallel to depictions of social dancing in sources of vernacular song and romance (see Figure 3.10).Footnote ⁶³ The image is evocative, but not the sole evidence called upon to label the works that follow “clerical dance songs.” In addition to iconography, the abundance of rondelli and a thematic focus on Easter and festive and liturgical occasions associated with dance have been interpreted as markers of a choreographic function. Given this confluence of details, few dance scholars or musicologists have critically evaluated the probability that the eleventh fascicle transmits such a large number of dance songs in relation to the relatively few surviving instrumental dance works and vocal dances in other sources.Footnote ⁶⁴ If the refrain songs in F were intended to be danced to, the manuscript represents an atypical source within the history of medieval dance music, and the only source for Latin refrain songs that explicitly signals a choreographic function.Footnote ⁶⁵

Figure 3.10 F, fol. 463^r, illustrated “D-” of De patre principio. Reproduced by permission of Florence, Biblioteca Medicea Laurenziana, MiBACT. Further reproduction by any means is prohibited

A reevaluation, however, of the evidence deriving from the iconography, form, and poetry of the refrain songs in F weights the relationship of song to dance more on the side of the symbolic than the functional. These songs may have invoked the idea of dance, but a primary function as the accompaniment of clerical dance in thirteenth-century Paris cannot be assumed; poets, composers, scribes, and compilers in F, and in sources for the refrain song more broadly, show a greater interest in the voice, and the poems and melodies it realized, than in the body and its movements.Footnote ⁶⁶ Apart from infrequent, highly localized witnesses, manuscript sources for the Latin refrain song are silent when it comes to rubricating or otherwise gesturing textually or iconographically toward dance – with the exception of the image in F.Footnote ⁶⁷

In addition to the opening historiated initial, the internal evidence of the refrain songs themselves in F has most often been called upon to support an association with movement.Footnote ⁶⁸ Although the poetic register and vocabulary of the songs have been called upon as evidence of function, a vocabulary of sobriety infuses the refrain songs of F, illustrating the conservative and vocal performance-focused poetics of the fascicle, as I argue in Chapter 1.Footnote ⁶⁹ When it appears in the fascicle, dance-related vocabulary metaphorically frames the voice and signals a festive register. Similarly, the presence of a refrain has been interpreted as a signal of register, and especially a so-called “Lower Style,” characterized by links to dance and popular song.Footnote ⁷⁰ Formal parallels of the Latin refrain with vernacular song forms, most especially French rondeaux, have consequently underpinned the choreographic identification of the Latin rondellus in particular, even though the identity of the French rondeau as a “dance song” is itself precarious.Footnote ⁷¹ In Latin or French, the rondeau form does not supply incontrovertible evidence of a choreographic function.

The existence of Latin and vernacular contrafact pairs has also supplied evidence in support of the Latin refrain song as clerical dance music – if a Latin song is modeled after a vernacular dance song, the assumption has been that the two share a function. Yet, even if Latin contrafacts hypothetically shared the function of vernacular dance models – which is debatable – the paucity of vernacular contrafacts in Fascicle XI of F is noteworthy. Of the nearly sixty refrain songs in F alone, only two have vernacular contrafacts: Decet vox letitie and Fidelium sonet vox sobria.Footnote ⁷² Of these, neither can be unproblematically associated with dance via its vernacular counterpart. Decet vox letitie, for instance, has a Latin contrafact, Ave mater salvatoris, in St-M D, and together the songs share music with a French refrain in the motetus of the motet Tout leis enmi les prez/DOMINUS, a refrain found also in an unnotated rondet from Jean Renart’s Le Roman de la Rose, C’est la gieus, en mi les prez.Footnote ⁷³ The paths and chronologies of transmission and contrafacture among these four contexts are convoluted, to say the least, and there is little evidence to suggest that the rondet was danced to, or that either Latin song serves a choreographic function.Footnote ⁷⁴

Fidelium sonet vox sobria, by contrast, shares its music, but not rondellus form, with a contemporaneous strophic+refrain Galician-Portuguese cantiga, Maldito seja quen non loará from the Cantigas de Santa María (fol. 260^r): It is unlikely that either song is a contrafact of the other; instead, a third, unknown song probably served as an intermediary.Footnote ⁷⁵ Although the cantiga may have been a dance song, or may have borrowed a dance song melody, it would be a stretch to see dance as a shared function between Maldito seja quen non loará and Fidelium sonet vox sobria, especially in the absence of an additional work connecting the two songs.Footnote ⁷⁶ Latin-vernacular contrafacture, in other words, offers little support to the identification of the refrain songs of F as dance songs.

Where Fidelium sonet vox sobria, uniquely among the refrain songs of Fascicle XI, gains potential meaning as a clerical dance song is in its complicated reception history outside of F. Since the publication of an anonymous French article published in 1742 identifying Fidelium sonet vox sobria as an example of a song danced to at Easter at the collegiate church of St. Mary Magdalene in Besançon, France, in the fifteenth century, scholars have repeatedly linked this song with clerical dance.Footnote ⁷⁷ This identification, however, is rooted in layers of historiographical reinterpretation and misreading and the witness of several lost manuscripts; moreover, the possible – if unlikely – role of Fidelium sonet vox sobria as a dance song in fifteenth-century Besançon has little bearing on its choreographic identity in thirteenth-century Paris.Footnote ⁷⁸

Although the internal evidence – poetry, form, and contrafacture – of the refrain songs as dance songs is ambiguous at best, the image of the clerical dancers at the head of Fascicle XI might provide clearer proof of function (see Figure 3.10). Not only is the image strikingly similar to contemporary depictions of caroles, the visual program in F also often showcases thematic links between miniatures and the music of the fascicles (or works) they introduce.Footnote ⁷⁹ Importantly, however, none of the other historiated initials allude to the function or performance of the songs that follow. The question, then, is whether the refrain songs dominating the final fascicle motivated the illustrator to switch from thematically related illustrations for the previous openings to this sole instance of an image suggestive of performance practice. Considering that the artist probably drew on stock images, and in at least one case misjudged the thematic match between illustration and the work that followed,Footnote ⁸⁰ one such factor may be the frontispiece of the manuscript depicting the Boethian division of music (musica universalis, humana, and instrumentalis) (see Figure 3.11).Footnote ⁸¹ Enthroned in panels on the left side, Lady Music gestures to the panels on the right representing each division in turn: The heavenly firmament represents musica universalis; a group of four dancers represents musica humana; and a vielle player in an instrumentarium represents musica instrumentalis.

Figure 3.11 F, frontispiece, Boethian division of music.

Reproduced by permission of Florence, Biblioteca Medicea Laurenziana, MiBACT. Further reproduction by any means is prohibited

A unique visual elaboration of Boethius’s tripartite division, the frontispiece philosophically frames the music contained within F, only indirectly relating to the music folios that follow. For instance, the artist chose instruments to depict the only audible music in Boethius’s division, musica instrumentalis, yet the manuscript transmits solely vocal music. The male figures reflect musica humana through synchronized motions symbolizing the harmonization of body and soul and, as Anne-Zoé Rillon-Marne has suggested, social harmony, by combining clerical and lay figures.Footnote ⁸² Moreover, by employing dancers in this context, the artist accessed a long-standing tradition in which dance mirrored celestial harmonies, exemplified by the heavenly ring dance described by Honorius of Autun and others.Footnote ⁸³

The images in the frontispiece and at the head of Fascicle XI, the final unit of music in the manuscript, while not identical, visually bookend the source. How might we interpret the illuminator’s reuse of the type of image associated with musica humana in the frontispiece for introducing devotional Latin refrain songs later in the manuscript? Is the image of dancers transformed over the course of hundreds of folios from a metaphorical to a literal representation of movement?Footnote ⁸⁴ Rather than assume the artist recognized the refrain songs of Fascicle XI as dance music and chose to represent them functionally – rejecting the working procedure of the visual program throughout the manuscript – a simpler answer might be that the scribe was thematically illustrating the poetry on fol. 463^r. Considering the collective spirit of the songs and refrains that follow, perhaps the artist was inspired to depict clerics absorbed in an activity deeply linked with the act of harmoniously rejoicing together, thereby thematically, rather than functionally, linking image with content. Clerics caught in a moment of bodily and communal rejoicing may have seemed an apt choice, considering that the songs copied on the folio celebrate Easter in highly communal and festive terms.

What is more, the artist had only to look to the frontispiece for a visual prototype of musica humana, the bodies of individuals moving together as symbolic of a higher order of harmony. In this interpretation, an image of group dance becomes another way of signaling the community-building enacted by the songs’ vocal performance, using a familiar image of musica humana to cue the “social harmony” of the clerical community represented visually by the dancers and sonically by the refrain songs.

While evidence in support of the “clerical dance songs” of F is less convincing than scholars have hoped, the songs and iconography in Fascicle XI could still evoke the idea, if not the reality, of dance for medieval singers and audiences. Just as the polyphonic rondeaux of Adam de la Halle and Guillaume de Machaut, or the later English carol evoke the idea of dance without its accompanying physical movements, so too could the Latin refrain songs of F resonate musically and formally with the gestures of dance.Footnote ⁸⁵ Their vocal, rather than bodily, performance manifests in the choreography implied by the poetry’s “joyful leaps” of the voice, with the voices of singers comprising a choreography of togetherness. Formulated in this way, the Latin refrain song might be understood as shaped by medieval perceptions of collective, social dance and its cultural associations with refrain forms, without necessarily serving as the accompaniment to dance.

If the refrain song, and those in F in particular, are demoted in importance within the narrative of the “clerical dance song,” what remains? Rather than relying on broad stylistic and formal markers such as the refrain, detailed archival research of local contexts and communities will best afford a richly textured, and historically accurate, understanding of devotional dance and its music in medieval Europe. This necessitates returning to the manuscripts and archives that have been repeatedly featured in histories of devotional dance, yet are mediated by layers of prior scholarship, hypotheses, and assumptions. For the Latin refrain song in particular, the label of “dance song” should not be applied without first reevaluating the reasons for so doing. Letting go of predetermined associations of refrains with dance allows the Latin refrain song to be resituated within a broader range of performance contexts in the Middle Ages.

Singing the Refrain: Form, Performance, and Community

Interrogating the poetry, form, inscription, and performance of the Latin refrain song reveals the invocation of communal, unified performance through different means and at different levels. The partiality of material evidence is an important caution: Rubrics or iconography in individual sources, or the relatively limited number of songs with variable musical textures between refrain and strophe, cannot speak to the performance of the repertoire writ large. Yet the identity of the refrain as either a real or imagined choral response – literally sung by a group or ventriloquized by a single singer or soloists – plays a role in fostering a sense of togetherness and community. Singing quasi una voce is powerful, as the chanted liturgy of the medieval church demonstrates. The refrain fosters a similar, even heightened, sensation of togetherness in devotional Latin song, underscoring through form and repetition the church’s doctrine of singing as if with one voice.

An implicit theory of collective song and performance develops in the poetry of the refrain song, rooted in a linguistic emphasis on communities of singers and grammatical preference for the first-person plural. Poetic, terminological, and grammatical choices made throughout the repertoire alert us to the implications of the refrain imagined by their creators for meaning and performance. However, the theory of collective performance expressed in the poetry of the refrain song does not uniformly correspond with performance practices described by scribes and compilers. Reappraising the evidence for the performance practices around refrain forms demonstrates that, in practical terms, refrains were performed responsorially, but the performing forces responsible for the refrain varied. Responsorial practices, moreover, are only rarely detailed in manuscript sources, and scribal evidence for performance along with musical settings emphasizes further the distance between rondelli and strophic+refrain songs already witnessed in the organization and compilation of refrain songs.

Above all, the refrain was framed and experienced as an expression of the collective voice – a moment of singing as if with one voice. The refrain enacted a sense of togetherness by symbolically and literally emblematizing the choric voice in form and performance, emphasized through repetition and the ensuing familiarity. This is the power of the repeated refrain in Latin song, and what affords the regular repetition of text and music a hefty cultural weight and ability to be meaningful in so many ways. Perhaps most importantly in light of Chapters 1 and 2, the ability of the choric voice to unify a group of individuals is part of the regulative function of singing and moving together.Footnote ⁸⁶ Implicitly and explicitly, the refrain coordinates, aligns, and even disciplines the voices of singers, while the words they sing persistently emphasize belonging in communities whose boundaries are defined by the church. The inclusiveness and plurality of the songs thus also speaks to those it excludes – for instance, those outside of the communities of clerics and choirboys in the collegiate churches and cathedrals of Moosburg, Seckau, Le Puy-en-Velay, and many other institutions and locales in medieval Europe where the Latin refrain song found willing singers and proponents. Finally, in this chapter I move away from the previous association of the refrain song with devotional dance practices, although, as I suggest, dance emerges in F as a metaphor for the social harmony of the clerical performers of the refrain song, once again reinforcing the communal spirit of this repertoire.