and a lower B, pitches that were excluded from the
Guidonian gamut. But it is intriguing that the Old Roman sources, representing
a milieu supposedly innocent of such theoretical considerations, also notate
these protus chants on A.Chant following the Epistle in the Roman Mass.
3. Repertory, texts and annual cycle.
JAMES W. McKINNON
The gradual, a chant of great melodic elaboration, is so named because it was sung on one of the higher steps – gradus – of the ambo (the same step on which the subdeacon read the Epistle, one below that on which the deacon read the Gospel). It was sung at every Mass throughout the year except during Paschal Time (the Sunday after Easter to the Saturday after Pentecost), when it was replaced by an Alleluia (two alleluias were sung then rather than the more typical gradual and alleluia, the arrangement for the rest of the year; on penitential occasions the alleluia was omitted entirely).
The gradual is a responsorial chant, that is, it bears some resemblance in its manner of performance to a responsorial psalm, where the psalm verses are chanted by a soloist and answered by a choral response. The medieval gradual consisted of a response and single verse. In the later Middle Ages the response was intoned by a cantor until its final phrase, which was sung by the chorus. This is apparently not the original arrangement. It is widely believed that at first a fourfold pattern was observed: the singing of the response by a cantor; its repetition by the chorus; the singing of the verse by a cantor; and a final choral repetition of the reponse. Such a format is plausible in view of the exigencies of oral transmission and the workings of responsorial psalmody, but it is not given explicit support (neither is it denied) in the sources. The following passage from Amalarius of Metz (d c850) is typical of the earlier descriptions: ‘Responsorium ideo dicitur, eo quod uno cantante ceteri respondeant ... ipse idem qui inchoavit solus, solus versum cantat’ (‘The responsory is so called because to that which is sung by one singer others should respond … he who has begun alone also sings the verse alone’, Amalarii episcopi opera liturgica omnia, ed. J.M. Hanssens, Vatican City, 1948–50, ii, 259). There is no reference here to a final repetition of the response, and it is not altogether clear whether the soloist is to begin with the entire response or only the intonation. However, the full notation of the response in the early 10th-century cantatorium of St Gallen (CH-SGs 359), a book reserved for the cantor, would seem to indicate that at the time the entire response was sung by the soloist. The later, abbreviated manner of performance is implicit in the alternation of polyphony and chant in early 13th-century Notre Dame organa, where only the solo portions of the chant were set polyphonically. However, for at least one gradual, Priusquam te formarem, for the feast of John the Baptist, the final words of the verse, ‘et dixit mihi’, would appear to call for the repetition of the response. That this is the case, but nevertheless an exception, is indicated by the 14th-century gradual of the Thomaskirche in Leipzig (D-LEu 391), in which the incipit of the response is given after the verse.
The origins of the gradual can be traced to the responsorial psalmody that accompanied the readings of the first portion of the Mass. The earliest unambiguous references to this psalmody appear in the patristic literature of the second half of the 4th century, particularly in the writings of John Chrysostom and Augustine of Hippo. Contrary to what is widely reported in both liturgical and music histories, psalms that were sung at this point in the service did not function as responses to a reading (in the manner of the medieval Matins Responsory) but were looked upon as independent readings. Augustine, for example, is particularly clear on this point when he refers to the Epistle, psalm and Gospel as ‘divine readings’: ‘We heard the Apostle, we heard the psalm, we heard the gospel; all the divine readings sound together’ (Sermon 165).
A related misconception about the psalmody of the early Christian Fore-Mass is that two psalms were always sung, one in response to the Old Testament reading (the gradual psalm) and another in response to the Epistle (the alleluia psalm). Martimort (1984, 1992) has shown that there was no single pattern of readings before the Gospel, and that a single reading was at least as common as multiple readings. As for psalmody, the sources call for a similar conclusion: multiple psalms are suggested by the occasional patristic reference, but more typically, especially in the West (as in the above passage from Augustine), a single psalm is indicated.
It would be an oversimplification, however, to see in these passages the ancestor of the gradual and nothing more. It is true that many passages refer explicitly to a responsorial psalm and hence suggest a proto-gradual, as when Augustine reminds his congregation of the brief psalm that they had just heard sung and to which they themselves had sung in response: ‘brevis psalmus ... quem modo nobis cantatum audivimus, et cantando respondimus’ (In psalmo cxix). But when one of the 20 psalms with ‘alleluia’ superscribed in the Book of Psalms was sung responsorially, that Hebrew exclamation appears to have functioned as the response. This might prompt speculation about a proto-alleluia, although since a single psalm only was involved, it could be said that the gradual psalm was occasionally an alleluia psalm. There is a further complication in that many passages, while mentioning that a psalm had been sung, are not explicit about the involvement of a response. Some passages may indeed have referred to direct psalmody, and hence a proto-tract, but on balance there are many indications to suggest that in the late 4th-century Fore-Mass a single responsorial psalm was the norm: first, the sheer number of references in the patristic writings to responsorial psalmody; secondly, the scattered references to it in 5th- to 7th-century Western sources, with nothing comparable for direct psalmody (McKinnon, 1996); and finally, the central position of the gradual in the emergent Roman liturgy of the 8th century.
To see the origins of the gradual in the 4th-century responsorial psalm of the Fore-Mass, however, is not the same as to claim that there was a continuous evolution between the psalmodic repertory of the patristic period and the early medieval gradual, the implication being that psalms were permanently assigned to specific dates in the 4th-century calendar and at a later date abridged when the manner of chanting them had become so elaborate as to prolong the singing unduly. Indeed there is much evidence against such a model. It is clear from the patristic sources that the psalms (and the Epistle and Gospel too) were not yet permanently assigned to liturgical occasions but were chosen each time by the celebrant. Only a small number of responses show continuity between patristic usage and appearance in the medieval repertory, for example, Haec dies (Psalm cxvii.24) for Easter and In omnem terram (Psalm xviii.5) for Apostolic feasts.
An examination of medieval gradual texts fails to support the notion of a psalm from which all verses except the first had been cut. More often than not the gradual verse was chosen from some later verse of the psalm, and a significant number of graduals have either non-psalmic texts or texts compiled from more than one psalm. Moreover, many gradual texts are not literal reproductions of psalmic verses but free compositions based on the biblical text; they give the impression of having been fashioned as artful ‘librettos’ for original chants. It is true that the more extreme examples of this, such as Ecce sacerdos magnus (Sirach xliv.16–20), tend to be graduals with non-psalmic texts, but there are many examples with psalmic texts as well, such as Qui sedes (Psalm lxxix.2–3) and Constitues eos (Psalm xliv.17–18).
The gradual, then, seems to derive in form, function and placement from the early Christian responsorial psalm, but to have experienced a period of radical transformation and re-creation at some point before its appearance in the medieval sources.
There is a ‘core repertory’ of 105 graduals, that is, those chants that appeared in both the Old Roman and the early Frankish manuscripts and hence made up the repertory transmitted from Rome to the Carolingian realm in the second half of the 8th century. (De necessitatibus, Domine audivi and Domine exaudi are not included in this number; see Tract.) Two other graduals appear only in the Old Roman books and a further nine only in the Frankish manuscripts (see §5 below).
With regard to the texts, 11 of these core repertory chants are derived from sources other than the Psalter (ten from various books of the Bible and one, Locus iste, that is non-biblical); four are compiled from more than one psalm (Hesbert, 1981) and the remaining 90 use single psalms. Thus the gradual repertory contains a significantly higher proportion of psalmic chants than most other items of the Mass Proper.
Concerning distribution, 73 graduals are assigned to the Temporale (the three festivals of 26 to 28 December included) and 32 to the Sanctorale. Chants for the Advent–Christmas season show particular care in their selection and arrangement; virtually all are uniquely assigned and contain explicit reference to the theme of their liturgical occasion. There is, moreover, an almost unbroken series of chants in the A-mode melody type (see §4), followed by a similar series of F-mode chants. A large proportion of the Lenten graduals are also uniquely assigned, the most notable exceptions being those for the Saturday of the Lenten Ember Days, and those for the first five Thursdays, which are shared with the Sundays after Pentecost. The gradual for Thursday in Passion week, Tollite hostias, is uniquely assigned to that date, a remarkable circumstance since the chant would thus appear to have been composed when Pope Gregory II (715–31) established the Thursdays of Lent as liturgical. The Easter gradual Haec dies (Psalm cxvii.24) is unique in that its response is repeated on each day of Easter week with a different verse from the same psalm. It is widely believed that the entire group of verses was originally sung on Easter day – the Mont Blandin gradual (Hesbert, 1935, p.100) does in fact assign them in this way – and they were subsequently apportioned throughout the week.
The graduals of the post-Pentecostal Sundays differ sharply from those of the Advent–Christmas season as regards the care with which they have been assigned and arranged. Of the 22 chants in the Old Roman series, only six are unique; the majority are shared with Lent, and two of these, Benedicam Dominum and Timebunt gentes, are repeated internally. The Old Roman series, with minor adjustments, is duplicated in the first of the two sets appearing in the Mont Blandin gradual (op. cit., p.lxxvi). It would appear, then, that the second set, which is numerically ordered according to psalmic text derivation, represents a Frankish arrangement. The liturgical assignment of graduals in the Old Roman and Frankish sources displays nearly absolute continuity for the Temporale (with the exception of the post-Pentecostal series); Sanctorale assignments are continuous in an approximate ratio of 2:1, a considerably lesser proportion than that observed in introits and communions.
The
105 graduals of the core repertory maintain substantial continuity between Old
Roman finals and Gregorian modal assignments (see Table 1);
only 16 of the chants contrast in this respect, a proportion roughly comparable
with other items of the Mass Proper. The modal distribution is thus
substantially the same in the two dialects, with a slightly more pronounced
preference for the protus in the Old Roman and the tetrardus in the Gregorian.
The Gregorian assignments (with the exception of the mode 2 melody type) show a
pronounced preference for authentic designations, perhaps because of the
generally high tessitura of the verses, a trait that appears to reflect their
virtuoso, soloistic character. What stands out especially in both Old Roman and
Gregorian graduals is the large number of chants (more than two-fifths of the
total) in the so-called F-5 mode (i.e. Old Roman F and Gregorian 5) and in the
A-2 mode (roughly a fifth). It is no surprise that in the Gregorian sources the
latter group have their final on A; this is to avoid an E and a lower B, pitches that were excluded from the
Guidonian gamut. But it is intriguing that the Old Roman sources, representing
a milieu supposedly innocent of such theoretical considerations, also notate
these protus chants on A.
|
TABLE 1: Old Roman finals and Gregorian modal assignments |
|
|||||||
|
|
|
|||||||
|
|
|
|||||||
|
old roman |
|
|
|
|
|
|
|
|
|
Final: |
D |
A |
E |
|
F |
|
G |
|
|
No. of graduals: |
18 |
21 |
12 |
|
44 |
|
10 |
|
|
|
|
|
|
|
|
|
|
|
|
gregorian |
|
|
|
|
|
|
|
|
|
Mode: |
1 |
2 |
3 |
4 |
5 |
6 |
7 |
8 |
|
No. of graduals: |
13 |
19 |
13 |
2 |
43 |
0 |
12 |
3 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|||||||
While both the A-2 and F-5 groups make liberal use of standard melodic formulae, as do most graduals to varying degrees, this trait is most pronounced in the A-2 type (that generally referred to in the musicological literature as the Justus ut palma type – an infelicitous designation since the latter is a Frankish addition and does not appear in the original Old Roman repertory). In most examples of this type the standard formulae are used in such regular succession that they can be said to approximate a genuine melody type, such as that named for the Alleluia, Dies sanctificatus. A representative example of the A-2 type, A summo coelo, is given in ex.1. Remarkably, there is a nearly exact correspondence between the use of the Old Roman and Gregorian formulae. It is possible to take the A-2 formulae singled out by Apel in his indispensable analysis of Gregorian graduals (Gregorian Chant, 344–63), find their Old Roman equivalents and observe this precisely analogous usage. The correspondence of the formulae (marked in ex.1 with Apel’s sigla) is exact except for a slight deviation from the standard Old Roman formula near the beginning of the phrase ‘opera manuum eius’. To be observed also is the exceptionally close melodic relationship between the two dialects, although this is a characteristic not just of this melody type but of melody types in general. In keeping with this melodic similarity the Old Roman makes less use of the trill-like oscillation that characterizes much of the Old Roman repertory. It is not known, of course, precisely how much the two versions changed during their respective histories of oral transmission, but at the least the relationship evident here suggests that both versions are close to the original.
The F-5 graduals are not nearly as regular as the A-5 group in their use of formulae, particularly as regards the responses, although there is more similarity among the verses, most of which conclude with the same formula. However, three of the F-5 graduals, Christus factus est, Exiit sermo and Sacerdotes eius, are virtually identical throughout. The verse of Locus iste also approximates this group, but the response does not. These exceptional chants aside, the F-5 graduals manifest about the same degree of formulaic regularity as does the rest of the repertory. The E-3 and E-4 graduals are among the more interesting. They tend to be particularly long and melismatic and to manifest a comparable degree of formulaic usage in response and verse, with a tendency towards melodic rhyme between the final cadences of the two sections. A number of the E-3 chants, including Eripe me, Benedicite and Juravit Deus, are quite similar, sharing a distinctive opening gesture, while the E-3 Speciosus and E-4 Domine praevenisti are entirely free. The D-1 graduals share with the F-5 chants the tendency to be considerably less formulaic in their responses than in their verses. A noteworthy detail of the G-7 category is the extraordinarily long melisma appearing on the word ‘pacem’ of Benedictus Dominus (see ex.2) and on ‘corde’ of Clamaverunt justi. That the Gregorian version with its skips and higher tessitura appears somewhat more virtuoso than the Old Roman is generally characteristic of gradual verses.
The chronological implications of formulaic usage, particularly with respect to the A-2 type, has been a matter of dispute among chant scholars. Reasons have been advanced for both its antiquity and its later origins, but what is perhaps more likely than either extreme is that a venerable melody would have been drawn upon at a later date to supply a repertorial need. In all probability, such hastily provided chants would tend to manifest greater melodic similarity. Of the A-2 graduals, the Easter exemplar Haec dies and the Christmas exemplar Tecum principium are virtually the two most irregular representatives, while the group of four assigned to the Saturday of the Advent Ember Days are among the most regular. (Several French musicologists have followed Jean Claire in seeing the Advent A-2 graduals as adopted by Rome from Gaul in the mid-6th century, but this view has not found wide acceptance among other chant scholars.) As for the F-5 chants, Viderunt omnes for Christmas is quite irregular, while the extremely homogeneous group mentioned above – Christus factus est (Philippians ii.8–9) for Holy Thursday, Exiit sermo (John xxi.22–3) for the feast of St John the Evangelist, and Ecce sacerdos magnus (Sirach xliv.16–20) for St Sylvester – share the presumably late traits of non-psalmic texts and assignment to dates in the calendar that are subject to liturgical adjustment. Similarly, F-5 chants are clustered in Passion week of Lent, and along with the A-2 type they comprise the six (three each) uniquely assigned post-Pentecost chants. It is also noteworthy that no less than nine of the 11 non-psalmic graduals are of the F-5 type.
After the transmission of the core repertory to the north the Romans added two graduals and the Franks nine. The Roman additions, Uxor tua for the Nuptial Mass and Qui Lazarum resuscitasti for the Mass for the Dead, are easily explained by the need to provide chants for these newly established liturgical occasions. Both graduals are examples of the A-2 type; Qui Lazarum is regular throughout, but Uxor tua is free at the beginning of the response in order to accommodate its exceptionally short text. Frankish additions are, by and large, similarly explained as regards other items of the Mass Proper such as introits and communions, that is, they were composed to provide chants for new entries into the liturgical calendar. But while this may be true in the case of a gradual such as Benedictus es, created for the Carolingian festival of Trinity Sunday, it is not for a number of other graduals. Unam petii, for example, is assigned to the Friday after Ash Wednesday and to the post-Pentecostal cycle, while Justus ut palma and Justus non conturbabitur are each given several sanctoral assignments. Of the nine Frankish additions two are D-1 chants, two A-2 chants and five A-5 chants.
In subsequent centuries the central European sources display the same conservatism with respect to the provision of new graduals as they do with other items of the Mass Proper. Among the small number of added chants Benedicta et venerabilis for the votive Mass of the Virgin Mary and Requiem aeternam for the Mass for the Dead are the most frequently encountered. Aquitanian and Italian sources provide considerably more, but not as many as they do for the introit, offertory and communion (and, of course, the alleluia). A-2 and F-5 chants figure prominently among the later graduals.
P. Wagner: Einführung in die gregorianischen Melodien: ein Handbuch der Choralwissenschaft, i: Ursprung und Entwicklung der liturgischen Gesangsformen bis zum Ausgang des Mittelalters (Leipzig, 2/1901, 3/1911/R; Eng. trans., 1901/R), 72–80
P. Ferretti: Estetica gregoriana ossia Trattato delle forme musicali del canto gregoriano, i (Rome, 1934/R; Fr. trans., 1938), 159–75
R.-J. Hesbert: Antiphonale missarum sextuplex (Brussels, 1935/R)
J.A. Jungmann: Missarum sollemnia: eine genetische Erklärung der römischen Messe (Vienna, 1948, 5/1962; Eng. trans., 1951–5/R as The Mass of the Roman Rite), i, 421–42
H. Hucke: ‘Gregorianischer Gesang in altrömischer und fränkischer Überlieferung’, AMw, xii (1955), 74–87
H. Hucke: ‘Die gregorianischen Gradualweisen des 2. Tones und ihre ambrosianischen Parallelen’, AMw, xiii (1956), 284–314
W. Apel: Gregorian Chant (Bloomington, IN, 1958, 2/1990), 344–63
R. Le Roux: ‘Les graduels des dimanches après la Pentecôte’, EG, v (1962), 119–30
F. Haberl: ‘Die Gradualien des III. Modus und ihre musikalische Struktur’, Musicus – Magister: Festgabe für Theobald Schrems zur Vollendung des 70. Lebensjahres, ed. G. Köllner (Regensburg, 1963), 76–99
J. Claire: ‘Les psaumes graduels au coeur de la liturgie quadragésimale’, EG, xxi (1980), 5–12
H. Hucke: ‘Towards a New Historical View of Gregorian Chant’, JAMS, xxx (1980), 437–67
R.-J. Hesbert: ‘Le graduel, chant responsorial’, Ephemerides liturgicae, xcv (1981), 316–50
M. Huglo: ‘Le répons-graduel de la messe: évolution de la forme: permanence de la fonction’, IMSCR XIII: Strasbourg 1982a [Schweizer Jb für Musikwissenschaft, new ser., ii (1982)], 53–77
P. Jeffery: ‘The Introduction of Psalmody into the Roman Mass by Pope Celestine I (422–432): Reinterpreting a Passage in the Liber pontificalis’, Archiv für Liturgiewissenschaft, xxvi (1984), 147–55
A.G. Martimort: ‘A propos du nombre des lectures à la messe’, Revue des sciences religieuses, lviii (1984), 280–331
J. Claire: ‘La musique de l’office de l’avent’, Grégoire le Grand: Chantilly 1982, ed. J. Fontaine, R. Gillet and S. Pellistrandi (Paris, 1986), 649–59
J. McKinnon: ‘The Fourth-Century Origin of the Gradual’, EMH, vii (1987), 91–106
B. Ribay: ‘Les graduels en IIA’, EG, xxii (1988), 43–107
R. Crocker: ‘Chants of the Roman Mass’, NOHM, ii (2/1990), 196–214
A.G. Martimort: Les lectures liturgiques et leurs livres (Turnhout, 1992)
D. Hiley: Western Plainchant: a Handbook (Oxford, 1993)
J. McKinnon: ‘Preface to the Study of the Alleluia’, EMH, xv (1996), 213–49
J. McKinnon: ‘Liturgical Psalmody in the Sermons of Saint Augustine’, Three Worlds of Medieval Chant: Comparative Studies of Greek, Latin and Slavonic Liturgical Music for Kenneth Levy, ed. P. Jeffery (forthcoming)
