1 The earliest known Western repertories of antiphons, understood as verses attached to liturgical psalms, are collected in the antiphonary of Bangor and the Old Hispanic orational of Verona. The antiphonary of Bangor is transmitted in Milan, Biblioteca Ambrosiana, MS C.5 inf., a manuscript dated 680–91 by Warren, Frederick Edward in The Antiphonary of Bangor, 2 vols. (London, 1893), 1:viiiGoogle Scholar. The Old Hispanic collection, an orational with antiphon incipits, is transmitted in Verona, Biblioteca capitolare, MS 89. Scholars agree that it was probably copied before 711, the date of the Muslim overthrow of the old Visigothic kingdom. A more precise date of 682–83 has been proposed by Manuel Diaz y Diaz in ‘La Fecha de implantación del oracional festivo visigótico’, Boletín arqueológico (Real Sociedad Arqueológica Tarraconense), 71–72 (1971–72), 215–43. The Expositio brevis antiquae liturgiae gallicanae, which Wilmart dated to the early eighth century, alludes to a repertory of antiphon-verses excerpted from a variety of biblical and ecclesiastical sources. Once thought settled, the dating of this source is now in doubt, as several scholars have placed the work in the middle or early part of the seventh century. That dating, however, remains within the broad framework of the author's assumption that antiphon-verses are an innovation roughly datable to the seventh century. See van der Mensbrugghe, A., ‘Pseudo-Germanus Reconsidered’, Studia Patristica, 5 (Texte und Arbeiten 80) (1962), 172–84Google Scholar; and Ekenberg, Anders, ‘Germanus oder Pseudo-Germanus?’ Archiv für Liturgiewissenschaft, 35 (1993), 135–9Google Scholar.

2 See the author's paper, ‘Antiphonal Psalmody in Western Monasticism from the Fourth to the Seventh Century’, read at the 15th Meeting of the International Musicological Society Study Group ‘Cantus Planus’, Vienna, 2011, forthcoming in the proceedings of the conference.

3 ‘Caeterum cum olim Antiphona cum Psalmo dicebatur vel Cantico, ad singulos Versiculos eamdem fuisse repetitam, nullus plane negabit.… Jam ergo tunc temporis coeperat in Ecclesiis Germaniae Antiphonarum repetitionem intermitti, imo ipsasmet etiam Antiphonas omitti brevitatis gratia: quae quidem omissio satis demonstrat, quod plurimum temporis postularet ex antiquo more Antiphonarum concentus, quam quod flagitat mos hodiernus, cum semel tantum vel bis Antiphonam canimus.’ Tommasi, Giuseppe Maria, Responsoria et Antiphonaria Romanae Ecclesiae (1686), rev. edn, ed. Vezzosi, Antonio Francisco (1749; repr. edn, Westmead, Farnborough, England, 1969), XXXIIIGoogle Scholar.

4 Gerbert, Martin, De Cantu et Musica Sacra a Prima Ecclesiae Aetate usque ad Praesens Tempus, 2 vols. (1774; repr. edn, Graz, 1968), 1:501–6Google Scholar.

5 Wagner, Peter, Einführung in die gregorianischen Melodien, pt. 1, Ursprung und Entwicklung der liturgischen Gesangsformen bis zum Ausgange des Mittelalters, 3rd edn (1911; repr. edn, Hildesheim, 1962), 144–5Google Scholar.

6 ‘Ein derartiges Verfahren verlängerte natürlich die Dauer eines Psalmes und noch mehr die einer ganzen Gebetstunde. Zuerst fiel darum die Wiederholung im Tagesoffizium fort, da man doch nicht den ganzen Tag nur mit Psalmodie ausfüllen konnte…. Von da aus drang der Gebrauch, die Antiphone nicht mehr nach allen Versen zu wiederholen, in die Matutin und die Laudes. Die Spuren der ursprünglichen Antiphonie verfolgen indessen das Offizium durch das ganze Mittelalter bis in die neuere Zeit hinein. Amalar erwähnt, wo er die Zusammensetzung des gewöhnlichen Nachtoffiziums beschreibt, sechs Antiphonen, die nach jedem Vers von den beiden Chören abwechselnd wiederholt wurden. Andererseits führte man noch später die Repetition nach jedem Vers gelegentlich wieder ein, wenn man die Dauer des Offiziums verlängern wollte, wie es die Cluniazenser mit den Antiphonen in der Matutin des hl. Martin machten. Viel später noch liebte man an Festtagen die Canticaantiphon in den Laudes und der Vesper dreimal zu wiederholen, also im ganzen viermal zu singen, vor dem Gloria Patri, vor und nach dem Sicut erat. Man hatte dafür die Bezeichnung triumphare oder triumphaliter canere, welche sowohl die dreimalige Wiederholung (ter fari) wie auch ihren festlichen Charakter schön andeuted. Schon der Bischof Durandus von Mende (13. Jahrh.) spricht von diesem Gebrauche als von einer eingebürgerten Sitte. Sie war ein Rest der alten Antiphonie.’ Ibid., 145–6. On the origin of the expression triumphare, see n. 26.

7 Hanssens, Jean Michel, ed., Amalarii Episcopi Opera Liturgica Omnia, 3 vols. (Rome, 1948–50), 1:117 and 201Google Scholar.

8 ‘Antiphona dicitur vox reciproca. Antiphona inchoatur ab uno unius cori, et ad eius symphoniam psalmus cantatur per duos coros; ipsa enim, id est antiphona, coniunguntur simul duo cori.’ Ibid., 2:433.

9 Isidore's definition in the Etymologies, 6.19.7–8, written c.620, states that the antiphon was sung in dual-choir alternation, implying that it comprised a whole psalm or at least several verses. While it is conceivable that Isidore was referring to a single verse divided in half, it is more likely that he was simply adopting the usage of the preceding three centuries, in which ‘antiphon’ referred to a psalm performed antiphonally.

10 Davril, A. and Thibodeau, T.M., eds., Guillelmi Duranti Rationale Divinorum Officiorum, 3 vols., Corpus Christianorum, Continuatio Mediaevalis 140, 140A, 140B (Turnhout, 1998), 140A:26Google Scholar. The practice is mentioned in the general rubrics of the reform-breviary of Pius V. See Breviarium Romanum Ex Decreto Sacrosancti Concilii Tridentini restitutum, Pii V Pont. Max. iussu editum (Rome, 1568), 23Google Scholar.

11 ‘Quanto enim melior est anima corpore, tanto melior est cantus animae quam corporis. Igitur intendendum est quae sit antiphona animae. Videtur nobis virtus dilectionis esse quae coniungit opera duorum fratrum simul. Psalmi ad opera referuntur, antiphona ad illam dilectionem, qua unusquisque fratri suo porrigit suum opus. Verbi gratia, unus legit et discit doctrinam in scola, alter seminat in campo; tempore fructus doctor seminanti porrigit doctrinam, sator doctori panem.’ Hanssens, Amalarii Episcopi Opera, 2:433.

12 ‘Duobus coris alternatur antiphona, quoniam non potest minus esse caritas quam inter duos. Hanc vicissitudinem caritatis significant cantores, qui alternatim ex utraque parte antiphonas levant. Hos duos coros designaverunt pennae animalium invicem porrectae, quae vidit Iezechiel in figura adventus Christi et novi testamenti. Coniunctio duarum pennarum antiphona est, quae vicem tenet caritatis.’ Ibid.

13 Literally ‘raise’ (levant) the antiphon. A comparable usage may be found in Ordo Romanus IX (c.900): ‘Cumque a cantore versus ad invitatorium [i.e., introitum] elevatus fuerit, inchoatur processio’. Andrieu, Michel, ed., Les Ordines Romani du haut moyen age, 5 vols. (Louvain, 1931–61), 2:329Google Scholar.

14 Durand agrees with this reading: ‘Cantatur [antiphona] etiam a duobus choris alternatim, ad notandum mutuam dilectationem siue caritatem que in paucioribus quam duobus consistere nequit.’ Davril, and Thibodeau, , eds., Guillelmi Duranti Rationale, 140A:26Google Scholar. In an earlier study, I interpreted this passage to mean that each antiphon was sung entirely by one soloist and that the principle of reciprocity was realised by assigning that task to soloists from alternate choirs over the course of several psalms. I now regard that interpretation as too contrived. See Nowacki, Edward, ‘Antiphonal Psalmody in Christian Antiquity and Early Middle Ages’, in Essays on Medieval Music in Honor of David G. Hughes, ed. Boone, Graeme (Cambridge, MA, 1995), 287–315, at 314Google Scholar.

15 Two Italian antiphonaries of the twelfth century, I-Rvat San Pietro B79 and I-Lc 601, expressly mark the place of the medial caesura. The former uses a majuscule letter, the latter a special sign in the shape of a cross. In addition, the practice is mentioned or implied in the rubrics of B79 as well as in the intrinsic syntactical and musical design of the majority of Office antiphons. See Nowacki, Edward, ‘The Performance of Office Antiphons in Twelfth-Century Rome’, in International Musicological Society Study Group “Cantus Planus”: Papers Read at the Third Meeting, Tihany, Hungary … 1988 (Budapest, 1990), 79–91Google Scholar.

16 In the ninth and tenth centuries, a period when the average repertory of Office antiphons doubled from roughly one thousand to roughly two thousand items, most performed only once a year, antiphons in all probability were not performed from memory but read. And in an era when monks and nuns did not possess personal breviaries, the likely alternative is that two soloists, representing the opposing choirs, sang the antiphons from a common antiphonary.

17 Tommasi (Responsoria, XXVIII–XXIX) believed the primitive method of antiphonal psalmody was to have the first choir sing the psalm verses and the opposing choir repeat the antiphon after every verse. But Amalar's description of both the antiphon and the psalm leaves little doubt that the practice of balanced, reciprocal alternation applied separately to each.

18 Hanssens, ed., Amalarii Episcopi Opera, 3:24.

19 Tommasi, Responsoria, XXXIII. See also Dyer, Joseph, ‘The Singing of Psalms in the Early-Medieval Office’, Speculum, 64 (1989), 535–78, at 541CrossRefGoogle Scholar; and Nowacki, ‘Performance of Office Antiphons’, 79, n. 2.

20 Precedent for referring to the antiphon as a verse may be found in Ordo Romanus I, whose Frankish redaction from the late eighth century explicitly describes the introit and communion as chants that are concluded with a final repetition of the ‘verse’, an unmistakable reference to the antiphon. See Andrieu, Ordines Romani, 2:83 and 105. Precedent for the preposition per introducing an expression of agency or manner may be found in Amalar's own usage in book 4, chapter 7 of the Liber officialis: ‘Antiphona inchoatur ab uno unius cori, et ad eius symphoniam psalmus cantatur per duos coros.’ Hanssens, ed., Amalarii Episcopi Opera, 2:433.

21 Perfect numbers are the sum of their positive divisors excluding themselves. Thus 6, whose positive divisors are 1, 2, 3 and 6, is the sum of 1, 2 and 3. (The second perfect number is 28.) The theory of perfect numbers was a standard subject in the quadrivium, transmitted, for example, in Boethius, De insititutione arithmetica, 1.20.

22 ‘Sunt namque nonnulli, qui … entrantes Ecclesiam, sine Antiphonis cursim et cum omni velocitate … divinis neglegenter assistunt laudibus.’ Migne, J.P., ed., Patrologiae Cursus Completus … Series Latina, 221 vols. (Paris, 1844–1904), 129:1399Google Scholar, quoted in Tommasi, Responsoria, XXXIII, and Wagner, Einführung, 145–6.

23 Migne, Patrologia Latina, 133:48.

24 Baroffio, Bonifacio G. and Kim, Soo Jung, eds., Biblioteca Apostolica Vaticana, Archivio S. Pietro B79, Musica Italiae Liturgica 1 (Rome, 1995), 20Google Scholar.

25 I-Rvat San Pietro B79, fols. 8, 10v, 14, 14v, 20v, 22, 23, 25, 30, 41v and 131v. The special manner of performing antiphons on Christmas, Epiphany and the feast of Sts Peter and Paul is described explicitly only for the first of the three, but is implied for the latter two by the rubric ad omnes antiphonas respondemus. On the interpretation of these rubrics, see Nowacki, ‘Performance of Office Antiphons’, 82–6.

26 Davril and Thibodeau, eds., Guillelmi Duranti Rationale, 140A:108. Gerbert calls the practice triumphare, claiming that the term derives from ter fari, and attributing the usage to the ancient church of Tours and to manuscripts of the fourteenth, fifteenth and sixteenth centuries at his monastery of Sankt Blasien. See Gerbert, De Cantu, 1:504.

27 In book 5, chapter 2, of the Rationale, Durand refers to the practice as follows: ‘Rursus, in quibusdam ecclesiis in fine antiphone fit neuma seu iubilus … Et fit neuma in unica et finali littera antiphone ad notandum quod laus Dei ineffabilis et incomprehensibilis est.’ Davril and Thibodeau, Guillelmi Duranti Rationale, 140A:27.

28 McKinnon, James, The Advent Project (Berkeley, CA, 2000), 118–22, 196–97CrossRefGoogle Scholar. On the general institutional history of Roman basilical monasteries in the seventh century, see Ferrari, Guy, Early Roman Monasteries: Notes for the History of the Monasteries and Convents at Rome from the V through the X Century, Studi di Antichità Cristiana 23 (Vatican City, 1957), esp. 365–407Google Scholar.

29 On the date of Ordo Romanus I and its Frankish redactions, see Andrieu, Ordines Romani, 2:38–51.

30 'Tunc illi, elevantes per ordinem, vadunt ante altare; statuuntur per ordinem acies duae tantum parafonistae quidem hinc inde a foris, infantes ab utroque latere infra per ordinem. Et mox incipit prior scolae antiphonam ad introitum, quorum vocem diaconi dum audierint, continuo intrant ad pontificem in secretarium. Tunc pontifex elevans se dat manum dexteram archidiacono et sinistram secundo, vel qui fuerit in ordine; et illi, osculatis manibus ipsius, procedunt cum ipso sustentantes eum.… Tunc peraccedens, antequam veniat ad scolam, dividuntur cereostata, ad dexteram et ad sinistram et pertransit pontifex in caput scolae et inclinat caput ad altare, surgens et faciens crucem in fronte sua, et dat pacem uni episcopo de ebdomadariis et archipresbitero et diaconibus omnibus. Et respiciens ad priorem scolae annuit ei ut dicat Gloriam; et prior scolae inclinat se pontifici et inponit. Quartus vero scolae praecedit ante pontificem, ut ponat oratorium ante altare; et accedens pontifex orat super ipsum usque ad repetitionem versus. Nam diaconi surgunt quando dicit: Sicut erat, ut salutent altaris latera.… Scola vero, finita antiphona, inponit Kyrieleison. Ibid., 2:81–4.

31 Joseph Dyer has interpreted this passage to mean that ‘a member or members of the schola cantorum sang psalm verses to which the entire schola responded with a refrain, a musical form found in the earliest graduals’. That interpretation, effectively conflating responsorial and antiphonal psalmody, represents a widely held view. It is not sustained, however, in my opinion, by the text here under consideration. See Joseph Dyer, review of The Advent Project: The Later-Seventh-Century Creation of the Roman Mass Proper, by McKinnon, James, Early Music History, 20 (2001), 279–309, at 296Google Scholar.

32 ‘Nam, mox ut pontifex coeperit in senatorio communicare, statim scola incipit antiphonam ad communionem et psallunt usquedum communicato omni populo, annuat pontifex ut dicant Gloria patri; et tunc repetito versu quiescunt.’ Andrieu, ed., Ordines Romani, 2:105.

33 See Jeffery, Peter, ‘The Introduction of Psalmody into the Roman Mass by Pope Celestine I (422–432)’, Archiv für Liturgiewissenschaft, 26 (1984), 147–65Google Scholar; and Dyer, Joseph, ‘Psalmi ante sacrificium and the Origin of the Introit’, Plainsong and Medieval Music, 20 (2011), 91–121CrossRefGoogle Scholar.

34 ‘Hic multa constituta fecit et constituit ut psalmi David CL ante sacrificium psalli antephanatim ex omnibus, quod ante non fiebat, nisi tantum epistula beati Pauli recitabatur et sanctum evangelium.’ Duchesne, Louis and Vogel, Cyril, eds., Le Liber Pontificalis: Texte, introduction et commentaire, 3 vols. (Paris, 1886–1957), 1:88–9Google Scholar.

35 Jeffery, ‘Introduction of Psalmody’, 163.

36 ‘Hoc officium addidit Caelestinus ille, ut in Gestis pontificalibus continetur: “Constituit ut psalmi David centum quinquaginta ante sacrificium psallerentur antiphonatim ex omnibus…” Quod nos ita intellegimus, ut ex omnibus psalmis excerperet antiphonas quae psallerentur in officio missae.’ Liber officialis, 3.5.2–3, Hanssens, ed., Amalarii Episcopi Opera, 2:272.

37 Peter Jeffery, citing Arnobius the Younger (mid-fifth century) as one of the earliest Western writers to use the word antiphona, adopts the common assumption that the word denotes a short verse. Hence, in his translation he inserts the indefinite article, an: ‘To these things the Holy Spirit responds in [an] antiphon’ (brackets original). But the Latin, ‘Ad haec respondit in antiphona Spiritus Sanctus’, may also be translated, ‘To these things the Holy Spirit responded in reply.’ We have no evidence at this time of any kind of responsive verse called an antiphon. Writers used the word in the non-technical sense of ‘answer’ or ‘reply’ and to refer to whole liturgical chants performed antiphonally. See Jeffery, Peter, ‘Monastic Reading and the Emerging Roman Chant Repertory’, in Western Plainchant in the First Millennium: Studies in the Medieval Liturgy and Its Music, ed. Gallagher, Sean, Haar, James, Nadas, John and Striplin, Timothy (Burlington, VT, 2003), 45–103, at 68Google Scholar.

38 Joseph Dyer, citing contemporary sources that use sacrificium for the Mass as a whole, along with architectural evidence, argues that the psalmody sung before the sacrifice in the sixth century must have been part of an entrance ritual. This conclusion differs from mine only in assuming that antiphonatim implies something ‘refrain-oriented’. The alternative that I propose, psalms sung in dual-choir alternation without refrain, would serve the ceremonial purpose just as well. See Dyer, ‘Psalmi ante sacrificium’, 104.

39 Wagner, Einführung, 1:145; Dyer, ‘The Singing of Psalms’, 568.

40 Davril and Thibodeau, eds., Guillelmi Duranti Rationale, 140:VIII.

41 Liber de officiis ecclesiasticis (1395), De canonum observantia liber (1397) and Tractatus de psalterio observando (1400). Dates, discussion and complete editions in Mohlberg, Cunibert, Radulph de Rivo: Der letzte Vertreter der altrömischen Liturgie, 2 vols. (Louvain, 1911, and Münster, Westfalen, 1915)Google Scholar.

42 Thibodeau, Timothy M., ‘From Durand of Mende to St. Thomas More: Lessons Learned from Medieval Liturgy’, in Ritual, Text, and Law: Studies in Medieval Canon Law and Liturgy Presented to Roger E. Reynolds, ed. Cushing, Kathleen G. and Gyug, Richard F. (Burlington, VT, 2004), 83–94, at 84Google Scholar.

43 Davril and Thibodeau, eds., Guillelmi Duranti Rationale, 140A:26.

44 Ibid., 27.

45 Radulph de Rivo is perhaps best known to chant scholars as the writer who reports the final demise of the Old Roman liturgy by decree of Pope Nicholas III (1277–80). See his De canonum observantia liber, prop. 22, in Mohlberg, Radulph de Rivo, 2:128. See also Van Dijk, Stephen J.P. and Walker, Joan Hazelden, The Origins of the Modern Roman Liturgy: The Liturgy of the Papal Court and the Franciscan Order in the Thirteenth Century (Westminster, MD, 1960), 411Google Scholar.

46 On the details of Radulph's career, see Mohlberg, Radulph de Rivo, 1:3–62.

47 For a modern reprint of the Paris edition, see Hittorp, Melchior, ed., De Catholicae Ecclesiae Divinis Officiis ac Ministeriis (1610; repr. edn, Westmead, Farnborough, England, 1970)Google Scholar. On Radulph's posthumous influence, see Mohlberg, Radulph de Rivo, 1:202–4.

48 Mohlberg, Radulph de Rivo, 1:185.

49 Ibid., 1:78–86.

50 Ibid., 1:179–80, and De canonum observantia liber, prop. 2–7, ibid., 2:38–50.

51 Tractatus de psalterio observando, cap. 19, ibid., 2:245.

52 De canonum observantia liber, prop. 10, ibid., 2:67.

53 Taft, Robert, The Liturgy of the Hours in East and West: The Origins of the Divine Office and Its Meaning for Today (Collegeville, MN, 1986), esp. 139Google Scholar. In later work, Taft has reiterated this view with emphasis. See Taft, Robert, ‘Christian Liturgical Psalmody: Origins, Development, Decomposition, Collapse’, in Psalms in Community: Jewish and Christian Textual, Liturgical, and Artistic Traditions, ed. Attridge, Harold W. and Fassler, Margot E. (Atlanta, 2003), 7–23, at 19Google Scholar.

54 Bäumer, Suitbert, Histoire du bréviaire, trans. Biron, Réginald, 2 vols. (1904; repr. edn, Rome, 1967), esp. 1:173-5Google Scholar.

55 Gelineau, Joseph, Chant et Musique dans le Culte Chrétien (Paris, 1962)Google Scholar, and as translated in Gelineau, Joseph, Voices and Instruments in Christian Worship, trans. Howell, Clifford (Collegeville, MN, 1964)Google Scholar.