The cycle Ave domine Iesu Christe (henceforth: ADIC), consisting of eight motets for four and five voices, belongs to the ‘core’ group of the so-called motetti missales.^[1] The unique source, Librone 1 (the first of the Gaffurius Codices), does not provide any attribution, neither in the heading of the first motet nor in the index.^[2] In the latter, compiled by Franchinus Gaffurius, ADIC is listed in the column of the ‘motetti missales consequentes’ and placed between two cycles explicitly attributed to ‘Loyset’ [Compère] (Ave virgo gloriosa caeli iubar and Hodie nobis). As pointed out first by Ludwig Finscher and then by Paul and Lora Merkley, the absence of an attribution is, in this constellation, rather an argument against Compère’s authorship – ‘precisely because of the explicit ascriptions on either side of it’, as Fabrice Fitch has recently observed.^[3] It might be worth noting a peculiarity of Gaffurius’s attributions in the index: whereas the other two cycles attributed to ‘Loyset’ [Compère] are also attributed by the scribe (Scribe A)^[4] at the start of the first motet,^[5] the attribution to ‘Gaspar’ [van Weerbeke] of the cycles Ave mundi domina and Quam pulchra es finds no indication in the body of the manuscript, where they were copied by the same scribe. In other words, Gaffurius’s attributions are at least in part independent from the data entered by (or available to) Scribe A. Therefore, we might infer, if Gaffurius did not enter any attribution for ADIC into the index, it was not because he was mechanically copying from the corresponding pages of the manuscript, but rather because he was unaware or unsure about the cycle’s authorship. This is, however, hardly a strong argument against the possible attribution to Compère. Gaspar – whose music, probably indeed the motetti missales, Gaffurius knew even before his tenure at the Duomo^[6] – began his second Milanese stay in 1489, and might have been personally in touch with Gaffurius while Librone 1 was being prepared^[7] (in his Angelicum ac divinum opus musice of 1508 the Duomo chapel master famously recalled his discussions about the notation of sesquialtera with ‘Iusquin despriet & Gaspar dignissimi compositori’ of many years before).^[8] Compère, in contrast, left Milan in 1477, and there was no possibility for Gaffurius to be directly in contact with him. Gaffurius might simply have found our cycle unattributed among other Duomo music materials from the 1470s; his own unawareness or unsureness does not ultimately rule out Compère’s authorship.

Turning to stylistic evaluations, it is the consistency of the cycle’s style with that of the other cycles by Compère that induced Finscher to ascribe it to the French master. In particular, Finscher judged the presence of the idiosyncratic ‘split tenors’ in the fourth and in the last motet as an unmistakable link with the cycle Ave virgo gloriosa caeli iubar, entirely permeated by this feature (both in the three-motet version present in Librone 1 and in the eight-motet version copied in Librone 3 with the rubric Galeazescha).^[9] As is well known, the split tenors create what Finscher described as a ‘strangely “unreal” five-part notation’.^[10] Subsequent scholars mostly accepted Finscher’s attribution (consecrated by the inclusion of the cycle in his edition of Compère’s works in the series Corpus Mensurabilis Musicae):^[11] the Grove article about Compère, for instance, lists the cycle among the surely attributed compositions and discusses it as such in the section ‘Works’.^[12] Though remaining open to the possibility of an attribution to Compère, Joshua Rifkin, in his influential article about Josquin’s Ave Maria … virgo serena of 2003, preferred to label the cycle as anonymous.^[13] More recently, Clare Bokulich has highlighted the differences in mensural practice between ADIC and the other motetti missales cycles, suggesting that ‘we may gain more by keeping an open mind about the work’s authorship’.^[14] Plainly contrary to the attribution to Compère is, instead, Fabrice Fitch, who in a monographic article argued that the cycle’s composer was ‘another member of Galeazzo [Maria Sforza]’s choir intent on emulating his famous colleague’:^[15] someone with ‘a lesser degree of facility and invention’, who gave himself away by a few instances of maladroit management of rhythm and counterpoint.^[16] Fitch sees the adoption and treatment of the split tenors as a further sign of this imperfect imitation of Compère: in his view, the obbligato passage involving both tenors in ADIC vita dulcis (mm. 39–43) shows that the anonymous composer opportunistically adopted that most special feature of Compère’s Ave virgo gloriosa caeli iubar without fully grasping its working, and ended up ‘abandon[ing] it for the sake of expediency’.^[17] (We should note, by the way, that Finscher, Rifkin, and Fitch all seem to agree on one point: that ADIC was composed after both Compère’s Hodie nobis and Ave virgo gloriosa caeli iubar).

As this survey of the literature makes clear, the question of the attribution is a complex one, which touches upon many different aspects. In the absence of truly conclusive proofs, I consider it preferable and heuristically more fertile to maintain the doubtful attribution to Compère in the heading of the present edition, instead of burying the cycle under the nondescript label ‘anonymous’. I feel encouraged to do so not only by the scrutiny of the partially contradictory findings by previous scholars, but also by several fresh elements and considerations that I will discuss in what follows.

A first observation regards the repeatedly mentioned split tenors present in both the fourth and the last motet. If Fitch blamed the composer of this cycle for not having fully understood the principle adopted by Compère in Ave virgo gloriosa caeli iubar, I contend (as I explain more fully in the introduction to MCE 3) that the real nature of this compositional device has been only imperfectly recognized so far. Its employment in Ave virgo gloriosa caeli iubar is inherently connected with a certain way of selecting and treating pre-existing materials: the two ‘antiphonal’ tenors of that cycle patently consist in their entirety of cantus prius facti (whether identified of not) deriving from sequences and other metrical songs. The pre-existing melodies are not treated as cantus firmi in long notes, but retain their tuneful character and phraseological geometry. Furthermore, they are given simple rhythmic and metric shapes probably redolent of the ways in which the corresponding songs were performed.^[18] The split tenor should not be seen as a whimsical choice of performance practice, but rather as part, and indicator, of a special compositional project: a tour de force of pre-existing melodies multiplied and transformed in polyphony. Having said that for Ave virgo gloriosa caeli iubar, it becomes clear why in ADIC only two motets have the split tenors: it is because only those two incorporate pre-existing materials treated with an approach similar to that of Ave virgo gloriosa caeli iubar, though not as regular and pervasive.

Let us consider first the last motet, Da pacem domine, in which the presence and treatment of pre-existing material is more consistent. In the first section (‘Da pacem, domine … deus noster’, mm. 1–27), there is a clear reference to the incipit of the corresponding chant antiphon Da pacem domine;^[19] only the first six notes of the antiphon, however, are sung by the tenor(s), with a modal rearrangement (from d-c-d-f-g-f to f-e-f-g-a-g).^[20] In the following section (‘Ave, Maria, gratia plena, Dominus tecum virgo serena’, mm. 28–45), the tenors alternate in singing a melody closely derived from that of the corresponding sequence.^[21] Subsequently (‘Ecce ancilla … verbum tuum’, mm. 45–52), the tenors answer each other intoning a common two-pitch reciting formula. To these derivations already noted by Finscher,^[22] we can add a new one, covering the entire following (and final) section of the motet (‘In honore … plena gratia’, mm. 53–74).^[23] The Marian sequence In honore matris dei was apparently unique to Cambrai: Analecta Hymnica lists as the only sources two missals printed for that diocese in 1495 and 1507,^[24] adding that the melody probably corresponded to that of the more widespread sequence for St. Nicholas Sospitati dedit aegros.^[25] Furthermore, Barbara Haggh counted In honore matris dei among the ‘sequences of later date found first or only in … manuscripts from Cambrai and probably composed in that city or its surrounding region’.^[26] A confirmation that the melody of In honore was indeed borrowed from Sospitati derives from the only polyphonic setting of the former sequence I could track down: Eustachius Barbion’s In honore matris dei, published in the Liber sextus ecclesiasticarum cantionum quinque uocum vulgo moteta uocant (Antwerp: T. Susato, 1553; RISM B/I 1553/13), closely paraphrases the melody of Sospitati.^[27] We are therefore in a position to realize that the composer of our motet derived from the sequence not only the words, but also the melody, as becomes clear by comparing the melody of Sospitati (for instance as attested in the notated breviary Paris, Bibliothèque nationale de France, Département des manuscrits, MS latin 15181, from Notre Dame, ca. 1300, see Fig. 1) with T1-T2, mm. 53–63 and 64–74. As in the case of the opening quote from the antiphon Da pacem domine, here too the original melody undergoes a modal change in order to fit with the F-mode of the whole motet (and cycle).

The split tenors of the last motet in the cycle, then, derive largely, if not entirely, from pre-existing materials and retain the tuneful character of the corresponding cantus prius facti. The situation for the fourth motet is more mixed. The text of its first half derives from the Ave domine Iesu Christe prayer, for which we are not even sure that one or more traditional melodies existed.^[28] The next section derives its text from the Christmas Benedicamus trope Verbum patris hodie, but the melody in the tenors does not resemble the one most frequently associated with the trope (which was also frequently set in cantus planus binatim^[29]). Interestingly, the notation of such settings as that of the Florentine laudario Banco Rari 19,^[30] ff. 72v–73r, clearly suggests a rhythmicized, ternary performance for the trope, and this resonates with the rhythmic treatment of the text in our motet; but it is also true that the first phrase of T2, mm. 48–52, precisely on the words ‘Verbum patris hodie’, exactly replicates a melody already sung in the previous section by A and T1, mm. 35–43, clashing with the idea of a pre-existing melody proper to Verbum patris hodie. The only passage in which pre-existing material comes to the fore is, in fact, the last one, with the litany for St. Augustine (‘Sancte Augustine, ora pro nobis’, mm. 77–92).

It is precisely this striking passage that may provide a new clue about the attribution to Compère, and leads me to my second observation on this matter. Fitch has pointed out that at least three other pieces of the period combine a litany tone with a technique very similar to the split tenors:^[31] Isaac’s Fortuna disperata/Sancte Petre/Ora pro nobis,^[32] Japart’s Vray dieu d’amours/Sancte Johannes baptista, ora pro nobis,^[33] and the anonymous Ic zei den claren dach/Sancte Johannes baptista, ora pro nobis.^[34] We should, however, observe that the three pieces differ from our motet in many ways: not only for the genre and the polytextual layering, but, most importantly, for the compositional concept. On the one hand, in all three pieces the two voices that alternate singing the litany are motivically and rhythmically independent from, and unrelated to, the other parts: this is a substantial difference from our split tenors. On the other, the litany device is at work for the entire duration of the piece, whereas in ADIC vita dulcis the litany characterizes the last segment only. Leaving aside now the split tenors and concentrating on the second aspect, we may note that there is a motet surely attributed to Compère in which a litany breaks into the texture at the very end: the five-voice tenor motet Gaude prole regia dedicated to St. Catherine of Alexandria.^[35] The litany appears there in the top voices, starting from m. 109: the formula used in the altus is similar to the one of ADIC vita dulcis, whereas that of the cantus is different (and more common). It is not the melodic similarity, however, that should interest us here, but rather the twofold compositional idea of (1) closing a motet with a sudden shift to a litany for a single saint and (2) giving the litany sonic prominence over the texture of the last segment. This idea, though accomplished by different means and in the context of a differently conceived motet, recalls that of ADIC vita dulcis, and finds, to my knowledge, no parallel in any other motet cum litany of the time.^[36] The temporal distance is conspicuous,^[37] and alternative explanations are surely possible, but it seems a remarkable coincidence that Compère, of all composers, resorted to such a similar idea in the same genre.

A final observation on matters of attribution can be added thanks to a ‘feature’-based computational analysis of the cycle conducted by Cory McKay in Spring 2020, which gave a largely positive result about Compère’s authorship.^[38] As McKay himself admits, his pioneering model cannot be trusted without question, especially since he had a limited number of digitally processed works by Compère with which to properly train the system. Since, however, the model tended to err rather on the side of ‘not Compère’, the fact that seven out of eight motets were classified as ‘Compère’ is all the more significant – though of course insufficient, per se, to draw any firm conclusions.

* * *

That this cycle belongs to the special category of motetti missales is revealed not only by its inclusion in the list of the ‘motetti missales consequentes’ in the index of Librone 1 and by the specific phrasing used there by Gaffurius (‘Ave domine Iesu Christe | cum reliquis totius misse’), but also by the presence of the so-called loco rubrics and of a motet clearly designed for the elevation of the Host (no. 6, Adoramus te Christe). In the Gaffurius Codices, only two other cycles have the loco rubrics (interestingly, the two by Compère); outside the Libroni, the rubrics appear in the two anonymous cycles of Munich 3154,^[39] but for some motets only.^[40] In ADIC the rubrics are arranged as follows:

It is the same exact pattern as in the Galeazescha of Librone 3,^[41] whereas Hodie nobis has a different organization of the Sanctus–Elevation complex,^[42] with a hybrid Sanctus followed by a ‘post Elevationem’ motet (see the introduction to MCE 1).

As Thomas Noblitt observed, ‘in most [motetti missales] cycles, the texts are related, not to a specific feast in the liturgy, but to a festival group, making possible the performance of the cycle on a number of occasions throughout the year’.^[43] This makes all the more sense if, as I argued elsewhere, the cycles were meant to be performed during low masses, both votive and festive.^[44] Finscher and Noblitt assigned our cycle to feasts ‘de domino nostro Jesu Christo’ throughout the year.^[45] The disparate character of the texts has probably been exaggerated in the past,^[46] as content-wise there is a distinct focus on the themes of the Eucharist and the Incarnation, theologically interconnected and most appropriate for accompanying a Mass.^[47] The texts of the first four motets mainly derive from one and the same prayer to Christ, often suggested in prayer books and hours of the time as accompaniment to the Elevation,^[48] and motets 5–6 consist entirely of two nested Elevation prayers. Several other texts converge on praising the Incarnation: from ‘Verbum patris hodie…’ in motet 4 to the Marian portions of motet 8, via a passage of the Elevation prayer in motet 5. The dense network of textual passages connecting the two themes and ‘cross-glossing’ each other is not to be overlooked:

Even the apparently incongruous penitential text of motet 7 can be tied to the Eucharistic thread through the mention of the redeeming Blood of Christ (‘quem redemisti, Christe, pretiosissimo sanguine tuo’). Undoubtedly, however, the composite structure of the cycle text is at times disconcerting, most notably in the interpolation of the litanic invocation to St. Augustine at the end of motet 4, which still awaits a plausible explanation. By comparison, the texts of Compère’s Hodie nobis cycle present a similar mix of prose and verse, and the same penchant for textual patchwork, but are more coherent content-wise, all thematizing Christmas (see the introduction to MCE 1). The texts of Ave virgo gloriosa caeli iubar again derive from many different sources, mostly sequences and rhymed prayers with few interpolations, but are all in verse and share a Marian focus (see the introduction to MCE 3).

It is debated whether the composers of the motetti missales were directly responsible for the choice of texts and for their assembling;^[49] in turn, this problem affects the question about the ‘Milanese’ and/or ‘northern’ roots of the cycles. As I have demonstrated elsewhere, Milan surely shared a widespread stock of devotional and liturgical practices, texts as well as prayers, that made the motetti missales possible (knowledge of which makes the missales appear less isolated and peculiar than they are usually considered).^[50] Some specific ‘ingredients’, however, seem indeed to come from the north.^[51] Within the texts of ADIC, the specific quote from In honore matris dei discussed above suggests the agency of a northern composer. Not only is this rare sequence attested exclusively in liturgical books for Cambrai, but its one known polyphonic reworking is, as discussed above, by Barbion (ca. 1515–56), a native of Antoing (County of Hainaut), who was later active in Bruges and, during the ten years before his In honore matris dei was published, in Kortrijk (Courtrai).^[52] Compère, by the way, was born either in the diocese of Arras (35 km north-west of Cambrai) or in the County of Hainaut, adjacent to the Cambrésis.

Another issue often discussed in regard to the motetti missales is their unitary design: were they conceived as cycles right from the start, or were they assembled from previously independent motets, perhaps with the intervention of various hands? In the case of ADIC the evidence is at least in part contradictory. As Finscher has accurately demonstrated, motets 1–3, 5, and 7 are ostensibly interconnected through recurring subjects, notably in the tenor:^[53] this strong effort to unify borders, at times, on literal repetition (compare motet 2, mm. 1–15, with 7, mm. 1–14, and motet 1, mm. 30–65, again with 7, mm. 15–55). Motets 4, 6, and 8, on the contrary, are motivically unrelated with the rest and among each other. Motets 4 and 8, in turn, share the split tenors feature, combined with the use of cantus prius facti. The choice of mensuration signs is consistent throughout the cycle; in particular, the highly unusual metrical progression from q to r3 binds together motets 2, 3, and 7.^[54] From the point of view of the text, as we have seen, motets 1–4 patently constitute a close-knit kernel (all drawing from the same prayer, even though with interpolations), while the other motets have looser, and thematic more than formal connections (motets 5 and 6 compose, however, a diptych). In sum, there are different and partially conflicting signs of unification at the various levels: pending further research, it cannot be ruled out that the cycle took shape after discrete phases of compositional work.

* * *

The cycle and all its component motets are unica in Librone 1. Even though a few emendations seem necessary (see the Critical apparatus), the version, copied by a single hand (Scribe A), appears to be substantially correct and reliable, also thanks to at least one intervention by a different, unidentified hand.

The most apparent differences with the previous editions by Finscher and by Dino Faggion^[55] are that the present one provides a critical apparatus and a separate annotated edition of the texts, and that it retains the original mensuration signs and rhythmic values. Rifkin has already pointed out a small series of faulty passages in Finscher’s pioneering edition (which are instead correct in Faggion’s one).^[56] To his list I can add the following:

As stated in the MCE Editorial methods, the text underlay adopted in the score ‘should be regarded more as a proposal than as a prescriptive solution’. Unsurprisingly, the distribution of the text under the notes in Librone 1 is often untenable, and has required the insertion of repetitions or added words (marked in italics in the score), and the shifting of words and syllables (only major interventions of this kind are recorded in the Critical apparatus).

Milan, May 2020