Alleluia

(Latinized form of Heb. halleluyah: ‘praise God’; Gk. allēlouïa).

Chant of the Mass in the Western Church and of the Divine Liturgy in the Eastern Church.

I. Latin rite

II. Byzantine rite

BIBLIOGRAPHY

JAMES W. McKINNON (I), CHRISTIAN THODBERG (II)

Alleluia

I. Latin rite

1. Definition.

2. Origins and early history.

3. 8th-century Roman repertory.

4. Roman Easter Week Vespers.

5. 8th- and 9th-century Frankish repertory.

6. Later history.

7. Other Latin liturgies.

Alleluia, §I: Latin rite

1. Definition.

The alleluia of the Mass is a Proper chant sung during the Fore-Mass after the gradual (see Gradual (i)) on liturgical occasions associated with penitence and fasting (most notably during Lent), and on ones associated with sorrow (such as the Requiem Mass), when it may be replaced by the Tract. During Paschal Time, beginning with Low Sunday, the gradual is omitted and two alleluias are sung.

The alleluia is performed in a responsorial manner: first the word ‘alleluia’ is sung, concluding with an extended melismatic flourish – the Jubilus; then a verse (rarely, two or three verses) is chanted in a moderately elaborate setting; and finally the alleluia is repeated. Throughout much of the Middle Ages a cantor intoned the alleluia without its jubilus and the chorus answered with the entire alleluia; one or two cantors sang the verse and the chorus entered for the final word or two (usually concluding with a melisma echoing that of the jubilus); the chorus, finally, repeated the alleluia. Early sources fail to indicate such involvement by the chorus, but it might well be that the chorus performed at least the final repetition of the alleluia.

Alleluia, §I: Latin rite

2. Origins and early history.

The word ‘alleluia’ is superscribed over 20 psalms of the Hebrew Bible (Psalms cv–cvii, cxi–cxix, cxxxv–cxxxvi and cxlv–cl), and it was probably sung as a response to these psalms in the Temple at Jerusalem. It was much employed by early Christians also, who chanted it both as a response to the psalms and as an independent acclamation. As the Office developed in the second half of the 4th century, it became customary to sing ‘alleluia’ in a response to psalms other than those to which it was superscribed in the Bible (in the Latin and Greek numbering, Psalms civ–cvi, cx–cxviii, cxxxiv–cxxxv and cxlv–cl). The Rule of St Benedict (c530) gives precise instructions for this custom (chap.15), and the contemporary author of the pseudo-correspondence of Pope Damasus and St Jerome called for a similar practice in the Roman Office. The usage of the early Western Mass, however, appears to have been considerably more restrictive: ‘alleluia’ was sung as a response only to the alleluia psalms and was confined to Paschal Time.

Much early Western evidence for the psalmody of the Fore-Mass comes from the 700 or more extant sermons of Augustine of Hippo (d 430), in which the psalm that had been sung previously in the service is frequently mentioned. Augustine speaks of but a single psalm in the Fore-Mass, typically declaimed by a lector and responded to by the congregation with a selected verse of the psalm. If the psalm was one of those with ‘alleluia’ superscribed, then ‘alleluia’ was the response (Psalm cxvii, an alleluia psalm sometimes sung with its verse 24, ‘Haec dies’, as response, is an apparent exception). The totality of evidence suggests that this single psalm of the 4th-century Fore-Mass, whatever its response, is the direct ancestor of the ‘responsum’ of Ordo romanus I, that is, the early medieval gradual (McKinnon, 1996). A formal link with the alleluia of the Mass would require the regular singing of two psalms in the ancient Fore-Mass, the second of which would use an ‘alleluia’ response. Such a configuration first appeared in the early 5th-century liturgy of Jerusalem, as made known through the celebrated Armenian Lectionary (see Renoux). The Jerusalem format was observed at several other Eastern ecclesiastical centres, including Byzantium, in the succeeding centuries (see Martimort). At a relatively early date the Byzantine alleluia took on the familiar form of alleluia, verse (usually two), alleluia (see Thodberg).

There is no compelling evidence that the Eastern practice reached the Latin Churches in the centuries immediately following Augustine. Two items of literary evidence, however, are frequently cited in support of the early existence of the Western, and more particularly the Roman, alleluia: the wordless jubilus, described by several Church Fathers; and the letter of Pope Gregory I to Bishop John of Syracuse in 598, which speaks of singing ‘alleluia’ at Mass outside Paschal Time. The jubilus was a sort of wordless chant employed by farm workers as an aid to their labours. Patristic authors such as Augustine invoked it in their allegorical exposition of biblical words such as ‘jubilate’ and ‘jubilatio’, but they never spoke of it in connection with the singing of ‘alleluia’; the first author to do so was the 9th-century Amalarius of Metz, who applied the term ‘jubilus’ to various melismatic passages in the chant, including that of the alleluia (McKinnon, 1993 and 1996). Gregory's letter speaks only of singing ‘alleluia’ at Mass, not necessarily the alleluia chant of the Mass, that is, the medieval genre consisting of alleluia, verse, alleluia. The liturgical circumstances of the time suggest that he was probably referring simply to a fuller employment at Mass of the affixing of ‘alleluia’ to certain psalms, a custom that pervaded the contemporary Office (see Martimort; Hiley; and McKinnon, 1996).

Once these two purported indications for the early existence of the Roman alleluia are set aside, a much noted paradox in its history is resolved. Scholars have long been puzzled by the contradiction between the seemingly ancient establishment of the alleluia and traits of the medieval chant that point to its late development, most notably, its pronounced instability of liturgical assignment. This contradiction has been rationalized by the hypothesis that the alleluia was originally sung in the Fore-Mass as a melismatic chant (the jubilus) unrelated to a psalm or psalm verse, to which verses were added many centuries later (see Apel). However, all indications now point to a late date for the establishment of the genre: in addition to the alleluia's instability of assignment, the limited number of Roman alleluia melodies and the exploitation of melody types in an apparent effort to render the repertory adequate for liturgical needs (see below) must be taken into account.

There is, moreover, a plausible explanation for the late appearance of the alleluia: unlike other genres of the Roman Mass Proper, it did not originate as a complete psalm but was adopted from the Byzantine liturgy as a mature chant, that is, one having the form of alleluia, verse (or verses), alleluia. The central argument for this view is Thodberg's demonstration that the three Roman alleluias with Greek texts (Epi si kyrie, O kyrioc and Oti theos) are derived from Byzantine alleluias, both with respect to text and melody. The degree of concordance between Roman and Byzantine verse texts far exceeds that of any other item of the Mass Proper; nearly half of the early Roman repertory is involved, a figure that could hardly be accounted for by coincidence. And then there is the shared preference of the Byzantine and Roman alleluia for G mode and the avoidance of F. As for the date of adoption, some time during the period of the Greek-speaking popes (685–752) seems likely (i.e. later than Thodberg proposed), perhaps not too distant from the time when the four principal Marian feasts were adopted under Pope Sergius I (687–701).

Alleluia, §I: Latin rite

3. 8th-century Roman repertory.

An inventory of the 8th-century Roman repertory is shown in Table 1, the result of comparing the Roman graduals with their 9th-century Frankish counterparts (i.e. the unnotated graduals edited in Hesbert's Antiphonale missarum sextuplex [AMS], 1935/R) and their early notated Gregorian counterparts; this is a method that works with near precision for other items of the Mass Proper, but retains a number of uncertainties in the case of the alleluia because the instability of liturgical assignment is matched by a comparable melodic instability. (The appearance of a Roman verse text in the AMS graduals is no guarantee that the verse was sung to a melody transmitted from Rome.)

Among the more striking features of the Roman repertory is the small number of melodies used. Of the Latin ‘alleluias with unique melodies’ listed in Table 1, Beatus vir and Haec dies have the same melody, while Te decet hymnus and Venite exsultemus are the same for the alleluia and first half of the verse; two different melodies exist for Confitemini quoniam, one used for the day of the Greater Litany and the other for the vigils of Easter and Pentecost. There are, then, 14 different melodies in this Latin group; these together with the three melody types, Ostende, Dies sanctificatus and Excita, constitute a total of just 17. Of the three Byzantine-derived chants with Greek texts, Epi si kyrie, O kyrioc and Oti theos, each one is melodically related to a Latin textual cognate. O kyrioc, the least problematic, is related to the Roman Dominus regnavit decorem. The other two, however, illustrate the complexities of the early alleluia's history: Oti theos is not melodically related to its Roman textual cognate Quoniam Deus, which is set to the Dies sanctificatus melody type, but to the Gregorian Quoniam Deus; Epi si kyrie is truly enigmatic in that it has no Roman Latin counterpart, textual or musical, but is melodically related to the Gregorian In te speravi.

A further complexity results from the considerable number of Roman alleluias that fail to appear in Gregorian sources. These chants can be divided into two overlapping categories: in the first (indicated in Table 1 by a single asterisk), the chants have textual concordances in the AMS graduals but different melodies in the early Gregorian sources (these Roman and Gregorian alleluias should probably be looked upon as essentially different chants that happen to use the same text); in the second (indicated by a double asterisk), even textual concordances in the AMS manuscripts are lacking. The first group, whose chants seem to indicate that the Franks substituted new melodies for the Roman originals, suggests a degree of difficulty in the transmission of the Roman alleluia repertory well beyond that encountered in other items of the Mass Proper (see below). The second group, that is, those chants entirely absent from Frankish sources, points to an interesting peculiarity of the Roman alleluia. All these chants, with the exception of Magnus sanctus Paulus, appear in the Roman post-Pentecostal cycle. In view of the Frankish practice of simply providing a list of alleluias for the season, it is surprising that there was a Roman post-Pentecostal cycle with fixed assignments. It is even more surprising that eight chants of the cycle were uniquely assigned – an apparent extravagance for a season in which one would expect most chants to be borrowed from other portions of the liturgical year. These uniquely assigned chants are Cantate Domino … cantate, Cantate Domino … quia, Laudate Dominum omnes, Laudate Dominum quoniam, Laetatus sum, Qui posuit fines, Qui sanat contritos and Quoniam Deus magnus, of which all except Qui posuit fall into either of the two categories in question, that is, they fail to appear in any Gregorian source with melodies related to those of the respective Roman chants, and Qui posuit itself appears only rarely. No fewer than six of the eight, moreover, are set to the Excita melody type. These two facts alone are sufficient to suggest that the Roman post-Pentecostal cycle was established after the mid-8th-century transmission of the Roman chant to the north, and that the cycle was hastily provided with chants, using especially the Excita melody type, and with texts of a similar nature, such as Cantate Domino and Laudate Dominum.

The instability of liturgical assignment of Gregorian alleluias has been much discussed, particularly the post-Pentecostal cycle, but also the Paschal Time alleluias, which, because of their supposed antiquity, might be expected to show some stability. The alleluias of the Advent-Christmas season, on the other hand, manifest a degree of stability roughly comparable to that of other items of the Mass Proper.

But these observations, though valuable, are inadequate because they fail to take into account the relationship with the Roman assignments. When Roman and Frankish assignments are compared, the overall degree of instability is even greater than previously suspected. Table 2 provides all alleluias for which there is both continuity of assignment between the Roman and Frankish manuscripts and stability of assignment within the Frankish books themselves. There are only 15 liturgical occasions involved, but these manifest a distinct pattern, or, more precisely, two distinct patterns, one each for the Christmas and Easter seasons. Every alleluia from Christmas Day to Epiphany without exception has the same Roman and Frankish assignment, while all the other alleluias of the Advent-Christmas season, that is, the Sundays of Advent and those after Epiphany, do not. For Paschal Time only eight dates manifest continuity between Rome and Francia, but these are the eight principal festivals of the season: the Easter Vigil, Easter Sunday, Easter Monday, Easter Saturday (the ancient clausum Paschae), the Greater Litany, Ascension, the Vigil of Pentecost and Pentecost Sunday. The remaining dates, like those of the Advent-Christmas season, are either ferias or numbered Sundays. It would appear, then, that only these 15 dates of the Roman Temporale had permanently assigned alleluias at the time of the mid-8th-century transmission to the north; the remainder of the repertory (i.e. the alleluias for all the Sundays and ferias, not just the post-Pentecostal Sundays) must have been conveyed in a list of some sort. Taking into account that this list of texts was probably transmitted without musical notation, it becomes clear why the Franks experienced difficulty in maintaining melodic continuity with the Roman alleluia. The alleluias of the 15 stable dates, on the other hand, display nearly total melodic continuity between the Roman and Gregorian books; the only exception is Laudate pueri, where the Gregorian chant retains the verse melody of the Excita type but substitutes a new alleluia melody.

A large proportion of the stable alleluias employ the three melody types. This provides at least a starting point for considering the difficult question of chronology within the Roman repertory (a subject that exceeds the scope of the present article). If some of the melody type alleluias, like the post-Pentecostal Excita chants, appear to be late additions to the repertory, then at least certain of their number, in view of their assignment to principal festivals, must have existed at a relatively early date.

Alleluia, §I: Latin rite

4. Roman Easter Week Vespers.

During Easter Week at Rome the principal clergy of the city gathered at the Lateran for a vesper service of particular splendour. On each day, between two and four alleluias with multiple verses were sung; some were borrowed from the Mass repertory, while others were unique to the vesper service and made use of a relatively simple common ‘vesper tone’. Table 3 gives the alleluias as they were distributed over the Week according to Ordo romanus XXVII. When the first verse is a Mass verse, as, for example, Pascha nostrum, the Proper Mass alleluia of that verse is sung; when the first verse is one of those set to the vesper tone, the melody of the Mass alleluia Dominus regnavit decorum is used. The vesper tone itself, according to Thodberg, is derived from the verse melody of Dominus regnavit decorem.

There is much else to recommend the special status of Dominus regnavit decorem: it precedes Pascha nostrum at Easter Sunday Vespers; it is one of the three Roman alleluias derived from a Byzantine model; and it appears no less than five times in the Roman Temporale, including Christmas Day itself and (in its Greek form) Easter Monday. It was, in fact, possibly the first Roman alleluia, originally assigned to Easter Sunday and later replaced by Pascha nostrum (see Bernard).

Alleluia, §I: Latin rite

5. 8th- and 9th-century Frankish repertory.

The list of 8th- and 9th-century Frankish additions to the Roman repertory shown in Table 4 gives all those verse texts appearing in the AMS manuscripts that are absent from the Roman graduals, a total of some 50. They are divided into melodic categories on the basis that each text has been traced through a representative selection of earlier Gregorian sources and the results checked against the compilations of Schlager (1965, 1968 and 1987). Additionally, a comparison of the Roman melodies with the Gregorian provides a group of chants (marked in Table 4 with an asterisk) for which a text that had also been used at Rome is set to a different Gregorian melody. (Because of the scale of the task, the results given here must be considered as approximate and provisional. There are, moreover, complexities that are difficult to reconcile with the format of this table: In te Domine, for example, which is given in brackets, has a Roman melodic counterpart in the Greek Epi si kyrie; and Laudate pueri, which is omitted, has a new Gregorian alleluia while retaining the Excita melody type for its verse.)

Perhaps the most striking feature of the Frankish repertory is the large number of new melodies. No fewer than 30 are indicated in the two right-hand columns of Table 4, while the alleluias ‘with multiple melodies’ must account for several additional new examples, even if the verses involved are sometimes set both to melodies represented by chants in the right-hand columns and to the three melody types. Finally, to the alleluias indicated in the AMS manuscripts should be added several that appear in the earliest notated sources. It would seem fair to estimate that the 8th- and 9th-century Franks at least trebled the number of melodies transmitted to them in the original Roman fund. A notable trait of the Frankish contribution is the preference for D mode in contrast to the Roman preference for G mode. The difference is nicely exemplified by Dominus regnavit decorem (ex.1). The Roman chant is in G and the Frankish one in D, even though the two are apparently related: they have the same overall melodic contour, their syllabic and melismatic passages are similarly apportioned, and they share numerous pitches at key points in both alleluia and verse.

Another striking feature of the Frankish alleluia repertory is the considerable degree of melodic instability, which far exceeds that of any other item of the Mass Proper. The alleluias in Table 4 labelled ‘with new melodies (unstable)’ manifest melodic variants from manuscript to manuscript that go well beyond those defined by Hughes (JAMS, 1987), and the alleluias ‘with multiple melodies’ can be said to display a still more radical degree of melodic instability. Finally, the verses ‘not found in medieval notated sources’ are those that appear in the unnotated AMS manuscripts but not in notated Gregorian sources; seemingly representative of alleluias that were sung at one Carolingian centre or another but failed to make their way into the Gregorian repertory, they thus exemplify a further kind of instability.

The musical characteristic of the alleluia that is of greatest interest is the tendency of the genre to display certain types of melodic repetition, most notably the re-use of the jubilus at the end of the verse, the imitation of the opening melodic gesture of the alleluia at the beginning of the verse, and the repetition of a melodic figure within a melismatic passage, whether in the jubilus or the verse. Peter Wagner (Einführung, iii, 1921) noticed this trait and appreciated its value in determining the relative age of alleluias – the later the chant the more likely it is to display such melodic patterning. This broadly valid proposition is useful in determining chronological layers in later alleluias, but a fundamental methodological difficulty arises when it is applied to determine the most ancient examples of the genre; this is generally attempted by matching the melodic traits of an alleluia with its liturgical assignment, on the assumed basis that important festivals such as Christmas and Easter will have older chants. Thus in the alleluias sung on such occasions, little evidence of melodic repetition might be expected.

There is a problem, however, in trying (as scholars have done) to match Gregorian melodies with events in Roman liturgical history. But even if Roman melodies were involved in the analysis, the results would still remain inconclusive. Comparing the jubilus of a Roman alleluia to the end of its verse has little point, because the verses (at least as they appear in the so-called Old Roman graduals) do not have extended melismas but only cadential figures of lesser length that nearly always duplicate the cadential figure at the end of the jubilus. As regards the internal repetition of melodic figures, only Pascha nostrum and Adorabo display this trait within their verses, obviously an insufficient finding for the construction of a meaningful chronology; and no Roman alleluia displays internal repetition in the jubilus of its opening alleluia, even if several do so in a second, greatly extended alleluia appearing after the verse in the Old Roman manuscripts.

If Gregorian melodic characteristics were to be applied to the history of the Roman alleluia, the results would still disappoint: for although seven of the alleluia melodies transmitted from Rome (Adorabo, Beatus vir, Dominus regnavit, Qui sanat, Venite exsultemus, and the Dies and Ostende melody types) fail, in their Gregorian versions, to display a melodic correspondence between the jubilus of the alleluia and the end of the verse, an almost equal number (six: Gaudete justi, Jubilate Deo, Pascha nostrum, Spiritus Domini, Te decet hymnus and the Excita melody type) do display such a correspondence. There is little to differentiate the two groups liturgically: if, for example, a chant such as Dies sanctificatus, with the supposedly ancient trait of a jubilus that fails to correspond with the end of its verse, appears on a venerable festival like Christmas, then Spiritus Domini, with the supposedly late trait of such correspondence, appears on the even more venerable festival of Pentecost. A final factor rendering such comparisons dubious is the real possibility that in a final revision of the Roman alleluia repertory liturgical assignments were reordered.

Similar difficulties are involved in any attempt to determine chronology by applying Wagner's approach to the 8th- and 9th-century Frankish repertory itself. The trait of melodic correspondence between jubilus and verse ending is virtually universal among the Frankish additions to the Roman repertory; the number of exceptions is too slight to form a basis for determining chronological layers. Similarly, the traits of repeating melodic figures within melismas and especially the imitation of the beginning of the alleluia at the beginning of the verse are distinctly minority occurrences. The striking melodic repetitions of a Frankish chant such as Alleluia, Justus ut palma are by no means typical. What can be said of such characteristics in the context of the 8th- and 9th-century Frankish alleluia is that they are encountered frequently enough, especially the customary repetition of the jubilus at the end of the verse, to differentiate the alleluia melodically from other genres of the Mass Proper. They are significant, too, as harbingers of a dramatically increased usage of melodic repetition in the alleluia after the 9th century.

Alleluia, §I: Latin rite

6. Later history.

The expansion in the 8th- and 9th-century Frankish alleluia repertory continued throughout the 10th and 11th centuries, confined primarily, however, to southern France (especially Aquitania) and central and southern Italy (especially Benevento); northern Italy, northern France, England and German-speaking regions retained the Frankish repertory with comparatively modest additions. The 410 pre-1100 alleluia melodies and about 600 verse texts listed by Schlager (1968) represent a complex mixture, with many texts being set to different melodies and a considerable number of melodies (beyond the original melody types) being used for several texts.

As for the melodic traits of the new alleluias, the correspondence of jubilus and verse ending is virtually universal, while the other forms of repetition are distinctly more common than before. About a third of the new chants imitate the beginning of the alleluia at the start of the verse, and a comparable number have some sort of melodic repetition within the verse; as many as half display melodic repetition within the jubilus, frequently creating an AAB pattern. Beneventan alleluias (the majority of the 10th- and 11th-century chants are unique to their own region) manifest these traits slightly more often than Aquitanian alleluias, while the latter display to a far greater degree than the Beneventan chants another progressive characteristic, namely, a generous range – at least an octave with the occasional filling in by upward and downward scalic and sequential passages.

The alleluias of the period, however, can by no means be described as extravagant or manneristic. They manage for the most part to integrate the new traits into a style of classic restraint; the very beautiful chant Veni Sancte Spiritus (ex.2) is not uncharacteristic. Here, the expected correspondence of jubilus (the entire alleluia in fact) and verse ending are present, and the opening of the verse imitates the beginning of the alleluia; a descending melismatic figure over the word ‘amoris’, spanning a 7th and echoing the jubilus of the alleluia, is repeated. There is no melodic repetition within the jubilus as it appears in most manuscripts, but some sources, including F-CHRm 520, create the common AAB pattern by duplicating a portion of the jubilus (that labelled ‘A’ in ex.2).

After 1100 the alleluia repertory continued to expand; before the end of the medieval period the number of melodies had more than doubled and there was a comparable increase in texts. As before, many melodies were used with more than one text and many texts with more than one melody, making the overall number of alleluias in the repertory difficult to calculate. Moreover, stylistic characteristics that might be considered extravagant in comparison to the pre-1100 chants made their appearance. This was particularly true in southern Germany between the 14th and 16th centuries and also in Bohemia, areas that had previously contented themselves with what was essentially a 9th-century Frankish repertory (see Schlager, 1987). Musically these new alleluias (for examples, see Hiley) are characterized by a greatly expanded range, frequent leaps of as much as an octave, extended melismas with scalic and sequential passages, and sometimes an abrupt juxtaposition of syllabic and melismatic phrases. The texts, most of which are non-biblical, are frequently rhymed and often interspersed with tropes.

Many of the new alleluias were composed to accommodate the later medieval intensification of devotion to the Virgin Mary, and an even greater number celebrated local saints. As a result much of the new repertory was regionally confined. But at the same time the core repertory remained intact and provided the chants for much of the universally observed liturgical festivals; many of the new melodies were even composed in classical Gregorian style, while new texts continued to be set to favourite traditional melodies such as those of Dies sanctificatus, Justus ut palma and Veni Sancte Spiritus.

Alleluia, §I: Latin rite

7. Other Latin liturgies.

There is no evidence that an alleluia was sung in the Fore-Mass of Gallican liturgical centres before the mid-8th-century transmission of the Roman chant to the Carolingian realm. The single existing description of a particular Fore-Mass from the period – that of Caesarius of Arles (d 542) for Pentecost Sunday – cites the responsorial singing of Psalm l but does not mention an alleluia. And similarly the 7th-century Expositio antiquae gallicanae liturgiae of Pseudo-Germanus states that ‘only the response’ is sung between the Benedicite and the Gospel.

In the Mozarabic rite an alleluia, called the laudes, was sung after the Gospel as early as the 7th century (see Mozarabic chant, §4(viii)); the acts of the Fourth Council of Toledo (633) state: ‘For the laudes follow the gospel because of the glory of Christ, which is proclaimed through that very gospel’. It is not certain, however, how much of the repertory of some 100 laudes appearing in the 11th-century chant manuscripts can be dated to the earlier period. The Mozarabic laudes have the same responsorial format observed in the Mass alleluias of most Western and Eastern rites, that is, melismatic alleluia, neumatic verse and melismatic alleluia.

Similarly, the alleluia of the Ambrosian rite is available only in sources dating from the 12th and 13th centuries (see Bailey; see also Ambrosian chant, §7(i)). In the limited extent of its repertory it bears a certain resemblance to its Roman counterpart. The manuscripts give just ten alleluia melodies (some of which are related to each other) and only 52 verses, which are themselves set to ten melody types matching the ten alleluia melodies. The alleluia melodies are notable for the complex development of their jubili: in addition to the original form of the melody there is an extended one called ‘prima melodia’ and a still more extended one called ‘secunda melodia’. This allowed the chant to be sung in a number of variations upon the basic responsorial format, employing different groups of singers. There are elements here that recall the less elaborate scheme given in the Old Roman graduals, that is, alleluia, verse, and an extended form of the alleluia.

While the alleluia of the Beneventan rite is known only from 11th- and 12th-century manuscripts, liturgical evidence suggests that the Beneventan repertory flourished in the 8th century (see Kelly; see also Beneventan chant, §3). It is probable that the entire body of this chant has not been preserved, but what remnants there are nevertheless create the impression that the 8th-century Beneventan alleluia repertory was particularly limited. Some 15 texts are set to the same alleluia and verse melody, while two others, Posuisti and Resurrexit tamquam dormiens, have unique melodies. The common format of melismatic alleluia followed by neumatic verse is employed, and although the repeat of the alleluia is not indicated in the sources, it was probably observed in practice. About a quarter of the Beneventan verse texts appear additionally at Rome (Pascha nostrum, Specie tua, Spiritus Domini and Tu es Petrus), and more than half in the Frankish sources, numbers not easily explained by coincidence. A comparative study of Latin alleluia verse texts might well shed light on the relationships among the various Western chant dialects.

Alleluia

II. Byzantine rite

1. The Byzantine ‘allēlouïarion’.

The allēlouïa takes its place in the Divine Liturgy, the Mass of the Orthodox Church, just before the recitation of the Gospel and immediately after the reading from the apostolos (the Epistle). The allēlouïa in fact consists of the singing of the word ‘allēlouïa’ followed by two or three psalm verses (stichoi), the latter known collectively as the allēlouïarion. The following instructions taken from the 12th–13th-century euchologion ET-MSsc 1020 illustrate this:

PSALTĒS: Allēlouïa, a Psalm of David.
DEACON: Attention
And the allēlouïa is sung.
Psaltēs [sings] the allēlouïarion.
People [sing] the allēlouïa.

The first allēlouïa is presumably sung by the psaltēs (the soloist peculiar to the Byzantine liturgy), and is then repeated by the people.

In present-day practice only the word ‘Allēlouïa’ is sung, but the liturgical books still contain the psalm verses belonging to the classical period. Of great musicological interest, however, is the medieval form reflected in the euchologion quoted above. A cycle of 59 allēlouïaria covering the Church year is found in medieval manuscripts. They are divided into six ēchoi (modes; since the tritos, or F modes, do not occur, only six of the ordinary eight ēchoi are represented). Unlike the Roman tradition there are only six melodies in the Byzantine rite associated with the word ‘allēlouïa’, one for each ēchos. Thus, for example, all allēlouïaria of the ēchos prōtos (first authentic mode) have one common allēlouïa melody.

By way of illustration, the allēlouïa of the ēchos plagios prōtos is given in ex.3 as it appears in the early psaltikon GR-P 221 (dating from 1177), together with the related melody of Alleluia, Dominus regnavit from the Roman Gradual. The similarity between the two examples to some extent springs from a direct relationship between the two melodies, and also from the common character of the mode. (See below for the relationship between Byzantine and Roman chant.) Whereas the Roman alleluia continues with the celebrated jubilus ending, the latter is not found in the Byzantine form.

The existence of no more than six short melodies in association with the word ‘allēlouïa’ throws the main interest on to the melodies of the allēlouïarion – the verses, from the Psalter, of the biblical canticles that follow. For medieval Byzantine music these are to be found in the psaltikon, the book belonging to the office of the psaltēs and presenting a special genre, perhaps the oldest one, of the psaltikon style (see Kontakion). These verses consist of between two and ten lines. The structure of the line is so stable that the verse (stichos) is without question the Byzantine form most strongly influenced by Centonization (as defined by Ferretti). The structure is determined by constantly recurring motifs, formulae and cadences.

As an example of a stichos, the first line of the allēlouïarion for Christmas Day (Psalm xix, ‘The heavens declare the glory of God’) is given in ex.4 in two versions: the two main traditions, the so-called ‘short’ and ‘long’ psaltikon traditions represented respectively by I-Rvat gr.345 and Ashb. 64. The modal characteristics stand out clearly. In theory the melodies belong to the ēchos prōtos. In practice the ‘short’ form may better be described as an ēchos plagios prōtos melody with a fixed B, whereas the ‘long’ form after its first few notes turns out to be a transposition one degree of the scale higher as compared with its counterpart. The F is particularly remarkable. Consequently the occurrence of the ēchos plagios prōtos in the allēlouïarion cycle seems to be a fiction. An analysis reveals only three or four modal groups, a fact that partly explains the weak position of the oktōēchos system. (For the modal system of the psaltikon see Kontakion.)

2. The Byzantine ‘allēlouïa’ and Roman alleluia.

From both the musicological and liturgical points of view, the Byzantine allēlouïarion cycle attracts attention because its relatively late musical evidence may prove that some of the alleluias of the Roman Mass were based on Byzantine models, with the so-called Old Roman tradition as an intermediate stage. The cases in point are the alleluias Dominus regnavit (ex.3 above), In te Domine and Quoniam Deus magnus. Ex.5 demonstrates the three stages of development for the Alleluia, Quoniam Deus magnus: (1) the allēlouïarion (ēchos tetartos); (2) the Old Roman version (from I-Rvat lat.5319), in which the original Greek text is transliterated into Latin letters; (3) the equivalent from the ordinary Roman Gradual.

This example shows that since the transition from the Byzantine to the Old Roman form and the Gregorian transformation of the Old Roman melody between the 6th and 8th centuries, the melodies have developed according to their own stylistic rules until their codification in 13th- and 14th-century manuscripts. Nevertheless, in general the similarities remain.

Alleluia

BIBLIOGRAPHY

latin rite

P. Wagner: Einführung in die gregorianischen Melodien, iii: Gregorianische Formenlehre: eine choralische Stilkunde (Leipzig, 1921/R), 397–417

R.-J. Hesbert: Antiphonale missarum sextuplex [AMS] (Brussels, 1935/R)

P. Blanchard: ‘La correspondance apocryphe du pape S. Damase et de S. Jérome sur le psautier et le chant de “l'alléluia”’, Ephemerides liturgicae, lxiii (1949), 376–88

L. Brou: ‘L'alleluia dans la liturgie mozarabe’, AnM, vi (1951), 3–90

E. Wellesz: ‘Gregory the Great’s Letter on the Alleluia’, AnnM, ii (1954), 7–26

H. Husmann: ‘Zum Grossaufbau des Ambrosianischen Alleluia’, AnM, xii (1957), 17–33

H. Husmann: ‘Alleluia, Sequenz und Prosa im alt-spanischen Choral’, Miscelánea en homenaje a Monseñor Higinio Anglés (Barcelona, 1958–61), 407–15

A. Wellnitz: Die Alleluia-Melodien der Handschrift St. Gallen 359 (diss., U. of Bonn, 1960)

E. Gerson-Kiwi: ‘Halleluia and Jubilus in Hebrew-Oriental Chant’, Festschrift Heinrich Besseler, ed. E. Klemm, (Leipzig, 1961), 43–9

H. Husmann: ‘Studien zur geschichtlichen Stellung der Liturgie Kopenhagens’, DAM, ii (1962), 3–58

B. Stäblein: ‘Das sogenannte aquitanische Alleluia Dies sanctificatus und seine Sequenz’, Hans Albrecht in memoriam, ed. W. Brennecke and H. Haase (Kassel, 1962), 22–6

B. Stäblein: ‘Der Tropus Dies sanctificatus zum Alleluia Dies sanctificatus’, SMw, xxv (1962), 504–15

W. Wiora: ‘Jubilare sine verbis’, In memoriam Jacques Handschin, ed. H. Anglés and others (Strasbourg, 1962), 39–65

B. Stäblein: ‘Zwei Textierungen des Alleluia Christus resurgens in St. Emmeram, Regensburg’, Organicae voces: Festschrift Joseph Smits van Waesberghe angeboten anlässlich seines 60. Geburtstag, ed. P. Fischer (Amsterdam, 1963), 157–67

H. Husmann: ‘Die Oster- und Pfingstalleluia der Kopenhagener Liturgie und ihre historischen Beziehungen’, DAM, iv (1964–5), 3–62

K. Schlager: Thematischer Katalog der ältesten Alleluia-Melodien aus Handschriften des 10. und 11. Jahrhunderts (Munich, 1965)

C. Thodberg: Der byzantinische Alleluiarionzyklus: Studien im kurzen Psaltikonstil, MMB, Subsidia, viii (1966)

K. Schlager: ‘Anmerkungen zu den zweiten Alleluia-Versen’, AMw, xxiv (1967), 199–219

K. Schlager: ‘Ein beneventanisches Alleluia und seine Prosula’, Festschrift Bruno Stäblein zum 70. Geburtstag, ed. M. Ruhnke (Kassel, 1967), 217–25

W. Apel: Gregorian Chant (Bloomington, IN, 1968, 2/1990), 375–92

K. Schlager, ed.: Alleluia-Melodien, i: Bis 1100, MMMA, vii (1968)

L. Treitler: ‘On the Structure of the Alleluia Melisma: a Western Tendency in Western Chant’, Studies in Music History: Essays for Oliver Strunk, ed. H.S. Powers (Princeton, NJ, 1968), 59–72

A. Burt: The Alleluias in the Manuscript Paris, Bibliothèque Nationale, lat.903 (diss., Catholic U. of America, 1969)

A. Renoux: Le codex arménien Jérusalem 121 (Turnhout, 1969–71)

M. Huglo: ‘Les listes alléluiatiques dans les témoins du graduel grégorien’, Speculum musicae artis: Festgabe für Heinrich Husmann, ed. H. Becker and R. Gerlach (Munich, 1970), 219–27

K. Levy: ‘The Italian Neophytes' Chants’, JAMS, xxiii (1970), 181–227

A.G. Martimort: ‘Origine et signification de l'alléluia de la messe romaine’, Kyriakon: Festschrift Johannes Quasten, ed. P. Granfield and J. Jungmann (Münster, 1970), ii, 811–34

B. Stäblein: ‘Die altrömischen Vespern der Osterwoche’, Die Gesänge des altrömischen Graduale Vat. lat. 5319, ed. M. Landwehr-Melnicki, MMMA, ii (1970), 84–140

H.G. Hammer: Die Allelujagesänge in der Choralüberlieferrung der Abtei Altenberg (Cologne, 1971)

E. Jammers: Das Alleluia in der Gregorianischen Messe (Munster, 1973)

H. Avenary: ‘Reflections on the Origins of the Alleluia-Jubilus’, Orbis musicae, vi (1978), 34–42

K. Schlager: ‘Choraltextierung und Melodieverständnis im frühen und späten Mittelalter’, Ut mens concordet voci: Festschrift Eugène Cardine, ed. J.B. Göschl (St Ottilien, 1980), 314–37

J. Claire: ‘Les formules centons des alléluia anciens’, EG, xx (1981), 3–4 [with music exx.]; xxi (1986), 23–45

T. Bailey: The Ambrosian Alleluias (Egham, 1983)

J. Claire: ‘Aux origines de l'alléluia’, Orbis musicae, ix (1986–7), 17–59

D.G. Hughes: ‘Evidence for the Traditional View of the Transmission of Gregorian Chant’, JAMS, xl (1987), 377–404

K. Schlager, ed: Alleluia-Melodien, ii: Ab 1100, MMMA, viii (1987)

J.W. McKinnon: ‘The Patristic Jubilus and the Alleluia of the Mass’, Cantus Planus III: Tihány 1988, 61–70

T. Kelly: The Beneventan Chant (Cambridge, 1989)

R. Crocker: ‘Chants of the Roman Mass’, NOHM, ii (2/1990), 214–22

P. Bernard: ‘Les alléluia mélismatiques dans le chant romain: recherches sur la genèse de l'alléluia de la messe romaine’, Rivista internazionale di musica sacra, xii (1991), 286–362

D. Hiley: Western Plainsong: a Handbook (Oxford, 1993), 130–39

J.W. McKinnon: ‘Preface to the Study of the Alleluia’, EMH, xv (1996), 213–49

byzantine rite

E. Wellesz: ‘Gregory the Great’s Letter on the Alleluia’, AnnM, ii (1954), 7–26

C. Høeg, ed.: Contacariam ashburnhamense, MMB, Principale, iv (1956)

C. Thodberg: Der byzantinische Alleluiarionzyklus: Studien im kurzen Psaltikonstil, MMB, Subsidia, viii (1966)

G. Hintze: Das byzantinische Prokeimenon-Repertoire (Hamburg, 1973)