Gaspar van Weerbeke’s motet cycle Ave mundi domina is transmitted in full only in the oldest of the Gaffurius manuscripts, Librone 1 of the Archivio della Veneranda Fabbrica del Duomo in Milan, completed in June 1490.^[1] Ascribed to ‘Gaspar’ in the index of the manuscript, this motet cycle was copied together with Weerbeke’s cycle Quam pulchra es by Scribe A in the section of the manuscript gathering labeled ‘motetti missales consequentes’ in the index, namely motet cycles meant to be sung during the celebration of Mass.^[2] Unlike Loyset Compère’s motet cycles Hodie nobis de virgine and Ave virgo gloriosa^[3] as well as the anonymous cycle Ave domine Iesu Christe,^[4] which include specific ‘loco’ rubrics referring to parts of the mass,^[5] both motet cycles by Weerbeke do not contain any rubrics. The layout of the cycles, however, similarly structured in eight (or seven) motets, including a section for the Elevation,^[6] clearly points to the same idea, connected with their use within liturgical services, as shown here:

Librone 2 transmits just two of the eight motets of Ave mundi domina and notably the ones corresponding to ‘loco Sanctus-ad/post Elevationem’, covering the Canon of the Mass and the mystery of the transubstantiation (nos. 5 and 6). In the same manuscript, immediately before, the motets with the same function from the cycle Quam pulchra es have also been copied, detached from the larger cyclic layout as well.^[7] In the context of a manuscript like Librone 2, almost entirely dedicated to Mass Ordinary cycles, these motets seem to represent additional music for the Eucharistic liturgy. This would in fact help explain the inclusion of these motets in Librone 2 as well (1492), that is, in a choirbook for the same institution chronologically close to Librone 1.^[8]

Compared with the other motet cycles mentioned above, Ave mundi domina is the only one for which partial concordances for single motets are found in sources from outside Milan. The Silesian manuscript Pl-Wu 5892 contains two motets from the cycle, namely Anima mea liquefacta est and Quem terra pontus aethera; the choirbook of the Siena cathedral I-Sc K.I.2 includes O virginum praeclara, whereas in the Flemish manuscript known as the Occo Codex (B-Br MS IV.922) the elevation section O salutaris hostia from Ave regina caelorum is copied as the opening piece.^[9] As the Table of Sources illustrates, the fifth motet of the cycle particularly, Anima mea liquefacta est, had a wide circulation. In fact, it is transmitted in seven sources of different provenance (besides Librone 1) not directly stemming from the Milanese copy, among them the richly decorated Florentine manuscript Banco Rari 229 (I-Fn B.R. 229) and the famous choirbook Brussels 228 (B-Br MS 228), belonging to Margaret of Austria, regent of the Netherlands.^[10] Anima mea liquefacta est thus represents the most widely transmitted work not only within the motet cycles but also among all of Weerbeke’s motets.^[11]

Thus, it becomes clear that motets originally conceived as parts of a cycle – or to be sung in a longer sequence of thematically interconnected motets – could find their way as ‘autonomous’ works in different places and contexts, being copied in diverse and unrelated sources.^[12] The independent circulation of individual motets, however, leads to the question whether this cycle was composed from the beginning as a unified project, namely as a cyclic composition formally defined and specifically arranged in the order given by Librone 1. Even though the other sources were copied later than Librone 1, the fact that their readings do not directly derive from it, but are instead related to lost sources, suggests that some of them may rely upon a different and possibly earlier channel of transmission; therefore, the motets might have circulated as single pieces at an early time.

In his dissertation on Weerbeke’s motets, Gerhard Croll first analysed his motetti missales, highlighting all the shared elements that contribute to create a cyclical layout.^[13] Composed in a common mode for the same range of voices (i.e. with unity of key and clefs), the motets, largely characterized by syllabic melodic lines stressing clear text declamation (with frequent repetitions of notes), have a series of compositional features in common that confer unity and enforce the cyclical design. Besides narrow vocal ranges, contrapuntal writing articulated by frequent cadences, and homorhythmic passages largely in trochaic rhythm, the motets also share melodic material, which Croll tried to characterize in melodic turns (Melodieverläufe). He recognized thematic relationships between the individual motets, underscoring particularly the occurrence of related motifs in their openings.^[14] The last three motets in particular show more pronounced affinities: not only do they share motivic material based on the melody of Fit porta Christi pervia in the incipits (paraphrased in the last motet of the cycle), but they are also interrelated with regard to their texts, in which stanzas from the hymns Quem terra, pontus, aethera and A solis ortus cardine are differently combined.^[15]

Some years later, Thomas Noblitt relativized the meaning of the melodic unifying means distinguished by Croll. In fact, he pointed out that Croll’s ‘Melodieverläufe’ belong to a ‘common melodic ductus’ related to the selected mode, to ‘be found in the works of numerous contemporary composers’.^[16] The discussion of the unifying means continued some decades later in Nolan Gasser’s study on the Marian motet cycles of the Libroni, the only contribution after Croll’s dissertation in which Weerbeke’s Marian cycles have been comprehensively examined as a whole with a certain degree of detail.^[17] Gasser recognized in Ave mundi domina a cycle ‘of considerable stylistic unity, despite the diverse nature of the texts’. Agreeing with Noblitt’s observation on the generic nature of the recurring melodic outlines identified by Croll, Gasser rightly called attention to the fact that in Ave mundi domina the cyclical unity ‘is engendered through a near constant variety of texture, counterpoint, and rhythmic character’, forming ‘a kind of model for one of the reigning paradigms of Renaissance aesthetics, unity in diversity’.^[18]

Despite the different approaches to the analysis of the compositional means and the interpretation of the composer’s work, the cyclical design of Ave mundi domina as transmitted in Librone 1 was never seriously questioned by these or other scholars.^[19] Indeed the compositional idea of a sequence of interconnected motets can hardly be argued in Ave mundi domina, although some distinctions can be made with regard to the single motets.^[20] Yet a comparison with Weerbeke’s second Marian cycle, Quam pulchra es, and its variable transmission within the Libroni (with a different number and order of the motets),^[21] in addition to the individual circulation of the motets mentioned above, does justify the question about the formal layout of Ave mundi domina. In a certain measure, in fact, from a textual and analytical point of view it would be plausible to think that the cycle was the result of a non-linear process, one that possibly was not clearly defined as such from the beginning. This may sound like mere conjecture, especially in consideration of the stylistic homogeneity of the cycle, but some textual choices, as discussed below, do support this hypothesis – which in turns frames the separate transmission of single motets in a different perspective.

In this light, the absence of rubrics in both Weerbeke’s cycles as well might acquire a new shade of meaning, and could be read in a double perspective. In fact, if from a philological point of view it can be assumed that the source from which Scribe A copied did not contain any ‘loco’ rubrics or instructions on the performance during Mass, the missing instructions could also point to an order (or function) of the motets as preserved in Librone 1 that was originally not completely prescriptive. In other words, behind the  easily recognizable cyclic façade, the performance of the motets might have been handled with a certain freedom, in terms of the number and order of the pieces – perhaps the motets were even conceived with this kind of flexibility in mind – at least partially, excluding the Elevation section. This, in turn, would suggest a more flexible approach to the idea of a cycle than the one characterizing the modern understanding of cyclic works.^[22] On the other hand, the separate transmission of the motets corresponding to ‘loco Sanctus-ad/post Elevationem’ in Librone 2 endorses the specific function and position of the ‘Elevation complex’, the only portion within the cycle properly characterized in relation to the service.^[23]

By assessing the transmission of Ave mundi domina within the Libroni and outside of Milan, it is clear that questions concerning the compositional genesis of the cycle should be considered separately, in principle, from issues related to its copying in Librone 1 – namely to choices, needs, or circumstances connected to the production of the manuscript itself as a collection of polyphony for the Milanese cathedral. In other words, one should take into account all the external elements that might have affected the manner of presenting the music in the manuscript. Nevertheless, the absence of other sources for the motet cycle as a whole necessarily makes the transmission of Librone 1 the only reference for any discussion on matters of composition and of cyclic design; in consideration of the parallel transmission of similarly structured motet cycles, it is in any case an authoritative one. Although the transmission raises many questions, there can be no doubt that Ave mundi domina reveals the composer’s intention to create a sequence of motets suitable (even if not exclusively destined) for a longer, consecutive performance, similar to the other similar settings in Librone 1, with which Weerbeke’s cycles share several stylistic features.

* * *

Over the years studies dealing with the Milanese corpus of the motetti missales have pointed out the peculiarities of these cycles in terms of style and function, seconding the idea that this music was composed following a Milanese tradition. The ‘local’ component has been identified both in some stylistic characteristics (such as writing for  paired voices, the declamatory and homophonic style, the insertion of sesquialtera sections – all elements gathered under the umbrella of a ‘Milanese style’)^[24] and in the practice of performing motets at Mass, and notably motets based on different texts from those recited or sung during the liturgical action. The idea that the motet cycles with their specific loco rubrics may point to Ambrosian services, proposed by Knud Jeppesen, was taken up by Gerhard Croll and discussed by Ludwig Finscher and Thomas Noblitt on the basis of the motets’ transmission in the Libroni of the Duomo.^[25] After the publication of Patrick Macey’s fundamental study on the musical patronage of Duke Galeazzo Maria Sforza, however, the connection of the missales with the Ambrosian liturgy of the Duomo has been substantially questioned and essentially discarded, while the motetti missales repertory by Compère and Weerbeke has been convincingly related to the court environment, in which these composers operated, and to the Roman rite followed by the Sforza dukes.^[26] Weerbeke was employed as singer and ‘vice-abate’ at the Sforza court from winter 1471–72 until 1481 and then later in the early 1490s. The composition of his two Marian motet cycles has been allocated to his first and longer Milanese period, analogous to Compère’s cyclic works, chronologically framed in the period of the composer’s stay in Milan from 1474 to February 1477.

In his study of Compère’s works Ludwig Finscher addressed the question of the mutual relationship between Compère’s and Weerbeke’s cycles, pointing out that the latter ones show more advanced stylistic features (with regard to the contrapuntal writing and the use of imitation).^[27] Yet Finscher’s astute observations do not help to establish a concrete chronology for the cycles transmitted in Librone 1, nor to define the specific role played by each of the composers in developing the design of the motet cycles. Regardless of personal contributions or priorities around an idea of cyclicity, which cannot be established so far, the main issue regarding the motetti missales remains the question of how this repertory has to be historically assessed in a proper way. If the belief that the motet cycles were rooted in the Milanese tradition and that the composers created them in reaction to that environment has found general acceptance, substantial objections to this view have been raised recently, especially in the context of the Motet Cycles research project, during which new investigations of this repertory have been carried out.^[28]

The reconstruction of the tradition of the texts set in the cycles, of their circulation and use, has in fact brought new elements to the discussion of this repertory, leading to new possibilities of reading the Milanese transmission. Some relevant findings concerning in particular the cycle Ave mundi domina are especially telling in this regard. A large part of the Latin prayers set in polyphony by Weerbeke has been recognized as circulating specifically in the Franco-Flemish region: the fact that the motet texts were drawn from sources outside of Milan suggests an association of Weerbeke’s motets with Franco-Flemish singing traditions. The poem of the first motet, Ave mundi domina, focused on the joys of Mary, for instance, has been tracked so far only in northern sources, and notably in a book of hours for Blanche of Burgundy.^[29] Circulating with the incipit ‘Audi’, instead of ‘Ave’, at least from the fourteenth century the text of the second motet, Ave mater gloriosa, is documented in French and Franco-Flemish sources, which point again to the area of provenance of the composer.^[30] Thomas Aquinas’s stanza O salutaris hostia, sung in the Elevation section of the fifth motet, enjoyed particular favour as an autonomous prayer in connection with the Bleeding Host of Dijon that was especially venerated by the dukes of Burgundy; it was indeed used as an Elevation prayer in the Franco-Flemish area.^[31] Furthermore, Fit porta Christi pervia, formerly believed to be a composite text from different hymns, is attested in the same version used by Weerbeke in books of hours from Flanders.^[32]

Remarkably, the selection of texts for Ave mundi domina does not reveal any specific connection with the Milanese environment. No specific Milanese prayers were chosen, nor particular prayers well attested in the Milanese area like Mater digna dei or Christi mater ave, on which Weerbeke based later motets.^[33] Instead, it shows that, in composing his motets, Weerbeke looked at prayers with which he was acquainted in his place of origin or of former activity.^[34] The identification of all the Latin texts set in polyphony in Ave mundi domina also shows that Weerbeke preferably used ‘ready made’ texts already in circulation instead of creating centos himself. The different strategy emerging from the textual design of Ave mundi domina compared to Compère’s cycles^[35] endorses the idea that the composer himself selected the prayers to set and that therefore his work was probably not substantially affected by ‘external’ requests of his Milanese patrons.

Yet, what the textual findings concretely mean in terms of cyclic design and composition, especially with regard to the precise circumstances that led to the origin of Ave mundi domina as a cyclical work, cannot easily be guessed. On basis of the texts selected, however, it would be quite plausible that the motets of this cycle (or of a part of them) originated in Flanders instead of in Milan.^[36] It would even be conceivable that one or more motets might have been composed in the North and later incorporated into a cyclic design such as the one transmitted in Librone 1 (or added to an already existing ‘block’ of motets).^[37] Although the idea of an origin of the cycle or part of it in Flanders has never been concretely considered in previous scholarship, such a hypothesis indeed deserves proper attention, since it would logically explain the particular textual choices revealed by recent research.

However, one could conceive of other plausible reasons for Weerbeke’s compositional choices without questioning the Milanese origin of the cycle. The composer might have followed a specific idea in designing his suites of motets and in choosing texts mostly belonging to the rich tradition of rhythmical poetry in his country. The hypothesis that the texts might have been selected with regard to specific Marian feasts in connection to a particular religious order was therefore explored during the aforementioned project, without, however, reaching firm conclusions.^[38] Pending further research, it is nevertheless worth remarking that the consequences of establishing a Flemish origin for at least some of the motets in Ave mundi domina would be far-reaching:^[39] the whole historical construct of the motetti missales as a Milanese phenomenon would require at least to be re-thought and its defining elements substantially reassessed.

The investigation of the textual tradition of Ave mundi domina has been particularly fruitful in relation to the stylistic features associated with the motetti missales as well, since it has raised awareness of the close relationship between compositional style and the texts set in polyphony. In the motets of Ave mundi domina we can recognize compositional means redolent of the tradition of singing hymns and sequences, and rhythmic verses in general: narrow vocal ranges, homorhythmic declamation, paired voices, and sesquialtera passages represent stylistic features associable with such singing traditions.^[40] In particular, the sesquialtera sections, considered as one of the most remarkable characteristics of the motetti missales and thought to be specifically linked to Milanese singing habits (or more in general to the Italian lauda), seem rather to be moulded on practices of singing rhythmic verses widespread in the European context, in Italy as well as in Flanders.^[41] Likewise in relation to the most striking feature of the missales, the chordal style of the Elevation section, visually marked by fermatas in the Libroni, the path opened by tracking the history of texts reveals a direct connection with the Franco-Flemish area. A musical setting of O salutaris hostia such as that in the manuscript Amiens 162 D, copied around 1500, in fact provides evidence of chordal singing at the Elevation in northern France similar to the examples in the Libroni.^[42] Furthermore, the French setting makes extremely evident the association of homophonic choral singing with practices of improvised counterpoint on a cantus prius factus – in this case the tune of Aquinas’s hymn Verbum supernum prodiens – that must have been much more widespread across Europe than it is recognizable today.^[43]

From the point of view of the liturgical use of a motet cycle like Ave mundi domina, as yet unexplored is the relationship with usages attested at the French court, where the performance of motets during the Mass liturgy is documented at least from the early sixteenth century. Even if scholars have related it to Milan,^[44] as a kind of emulation of a Milanese local custom, the origin of this practice is still unknown.^[45] Considering that Bona of Savoy, wife of Galeazzo Maria Sforza, grew up at the French court and that Galeazzo himself served the French King in his youth, one may wonder instead whether it was Galeazzo’s court that took over that practice from France in the 1470s: by a happy chance, the results remained ‘captured’ within the output of the Franco-Flemish Sforza composers later included in the Duomo Libroni. Without going too far, it is worth underlining that recent research has significantly contributed to enlarge the perspective on this repertory, placing it in the broader context of polyphonic singing practices diffused in various liturgical and devotional contexts, and fed by a rich tradition of improvised counterpoint on chant melodies.

In view of all the issues discussed above, Weerbeke’s compositional output deserves equal attention both in its dimension as a cycle and with regard to the individual motets. The particular fortune of Anima mea liquefacta est, unique within Weerbeke’s surviving motets, places in the foreground the only motet of the Ave mundi domina cycle setting a prose text and devoid of final passages in triple metre – which occur in the motets composed on rhythmic poetry, excluding the motet with the Elevation passage. Why specifically this motet enjoyed such considerable circulation cannot be properly determined, but the contents of the text, deriving from the Song of Songs, metaphorically understood as the quest of the human soul for divine love, the wide liturgical use of Anima mea as an antiphon for various feasts, as well as the powerful expressiveness achieved by the composer in his polyphonic setting might have been significant factors for its outstanding reception. Unrelated to a sacred context, however, is the transmission of the motet in sources like the Florentine manuscripts Banco Rari 229 and Magl. 178, from the early 1490s, in which the motet, not supplied with the text, apparently was not copied to be performed vocally. These sources attest to an autonomous circulation of the piece already at a relatively early stage and parallel to its inclusion in collections of mostly secular pieces.

Particularly interesting in this regard is to find the motet also in a choirbook such as Brussels 228, prepared for Margaret of Austria between 1516 and 1523, mainly containing secular works as well (and hence known as Chansonnier of Margaret of Austria).^[46] Provided with the entire text, this copy of the motet witnesses to an interest completely unconnected to its original function or destination decades after its composition and thus probably due to its semantic contents and the effective polyphonic setting. Notably, some major variants occur in Brussels 228 and in particular changes made to the motet reveal the wish to clarify some points of articulation of the setting in accordance with the textual syntax.^[47] It would therefore be conceivable to ascribe them to the composer himself as a result of a process of revision undertaken years after the composition of the motet, possibly in the period between 1495 and 1498, when Weerbeke worked in Flanders. Still, an external intervention is also possible and would confirm that the piece was an object of special attention in a Flemish environment, and certainly under different circumstances than those suggested by its earliest source.

The same motet is transmitted with a major variant reading also in the manuscript Barcelona 454 (Cancionero musical de Barcelona), dating from the early sixteenth century and including a large amount of sacred repertory.^[48] In this source the motet ends with an additional ‘Amen’ set in polyphony with a few additional notes. Given the composite nature of this source, copied in Catalonia by several Dutch and Spanish scribes, no clues to a direct relationship with Weerbeke can be found to suggest an authorial ascription of the ‘Amen’ addition.

Despite the availability of chant melodies for Anima mea liquefacta est, Weerbeke apparently did not compose his motet on an easily recognizable cantus prius factus. However, at least in two cases in the cycle he used pre-existing melodies. The chordal section on O salutaris hostia is based on a melody circulating with Dutch and German texts.^[49] Although it might be a contrafactum of a still unidentified Latin text, the melody again points to music from beyond the Alps. Sung by the Tenor voice, the tune is enriched by the homophonic setting of the other voices following models of note-against-note counterpoint.^[50] As mentioned above, Weerbeke’s O salutaris hostia found a place as an autonomous piece in the Occo Codex, prepared for Pompeius Occo around 1515–17 and including further settings of the same hymn stanza.^[51] In this case too we have evidence of the circulation of Weerbeke’s motets many years after their composition and, remarkably, also a corroboration of the direct association of this prayer with Corpus Christi in the Franco-Flemish region.^[52]

In Fit porta Christi pervia, the last motet of the cycle, Weerbeke used the hymn melody in the first section based on stanzas from the hymn. Instead of placing the melody in just one register, the composer distributed it among the voices, especially Superius and Tenor, but also involving the Bass, thus making use of techniques of paraphrase still little studied for motets of that time. It is interesting to realize that imitations are introduced on purpose to highlight segments of the pre-existing melody; they accordingly assume a precise structural role.^[53]

The melodic cell with the descending (and then ascending) interval of fourth (g′–d′) that features the incipit of the chant on Fit pervia Christi also characterizes the openings of motets nos. 5–8, giving a strong sense of unity to the second half of the cycle, although each motet is individually built and preserves its own character. These interrelated opening motifs cannot be fortuitous: they suggest instead a compositional project aimed at connecting the individual pieces and to create a cyclic sequence of music.

The similar opening motifs of Ave mundi domina and Ave mater gloriosa (particularly the melodic outline in the Superius), also echoed in the incipit of Anima liquefacta est, seem to suggest that also in the first motets of the cycle Weerbeke might have utilized pre-existing, though still unidentified, melodies.^[54] Melodic relationships can be perceived as more or less strong, but the first motets appear to build a ‘subunit’ of sorts, somehow distinct from the group of the last four motets. In relation to the issue of the genesis of the cycle as discussed above, this kind of differentiation could be meaningful in terms of compositional process and chronology. However, what matters here is to underscore that the composer probably combined more pre-existing melodic material in the cycle than we can recognize today. Further elements support this idea, such as the quotation in motet 5, Ave regina caelorum, of the incipit of the Salve regina melody at mm. 30–32 (corresponding to the word ‘Salve’ of the verse ‘Salve gloriosa’), and the presence of a melodic segment in Quem terra pontus aethera (T, mm. 15–19 and 37–41) that exactly matches a passage in the Tenor of the first motet in Compère’s Hodie nobis de virgine (mm. 11–14).^[55] This common material probably derives from a melody associated with hymn singing and known to both composers.

These examples give evidence of compositional procedures that freely handle pre-existing melodic material – procedures that even in apparently ‘freely’ composed motets were probably more frequent than one can grasp today. It is to be hoped that further research will favour new identifications and discoveries, which may lead us to a deeper understanding of Weerbeke’s work in all its multifarious stylistic components – and on the background of a rich tradition of sacred polyphonic singing. New findings will possibly reveal more on the choices made by the composer in designing his motets and in creating mutual relationships to substantiate a cyclic dimension.

* * *

Compared to previous editions,^[56] the present one retains the original mensuration signs and rhythmic values and it provides a critical apparatus and a separate annotated edition of the texts. These aspects in particular take advantage of the output of new research, giving information not contained in previous editions and studies.

The versions of Librone 1 have been taken as basis of the edition, also for the motets transmitted in other sources. Although the scribe responsible for the copying, Scribe A, worked quite carefully, some inaccuracies and a few errors required editorial interventions. Particularly in relation to the text setting the readings of the concordances have been taken into account, especially in cases of missing text passages or textual repetitions (such as in O virginum praeclara, lacking two verses in Librone 1, instead given in the Siena manuscript, in which, remarkably, even some repetitions are recorded).

As stated in the MCE Editorial methods, the text underlay given in the score has to be considered ‘more as a proposal than as a prescriptive solution’. Notably in melismatic passages different choices in the accentuation of the syllables can be made, also depending on whether one prefers to expand on middle or final syllables or is oriented towards a French accentuation of the Latin poetry. In the passages in sesquialtera the text underlay mostly follows the trochaic rhythms of the text, but in the case of the final section of Salve virgo virginum iambic patterns contrasting with the textual accentuation seem to be required. As is common in this repertory, characterized by tfrequent exchanges between voice pairs, the motet texts are mostly not sung complete by all voices: the lack of a textual portion in one or more voices is, however, not indicated in the Critical Apparatus, and only major interventions in the matter of shifting words and syllables are recorded.

Due to the current technical limits of the MEI rendering, editorial ties added to connect notes of the same pitch which have to be sung on a single syllable are indicated as normal (instead of dashed) ties, but registered as such in the Critical Apparatus.

Salzburg, March 2021