SINGING HORACE IN ANTIQUITY AND THE EARLY MIDDLE AGES

Stuart Lyons

doi:10.1017/S0261127921000085

SINGING HORACE IN ANTIQUITY AND THE EARLY MIDDLE AGES

Published online by Cambridge University Press: 23 February 2022

Stuart Lyons

Show author details

Stuart Lyons*: Affiliation:
London
*: Email: [email protected]

Article contents

Abstract
Horace’s Metres
Early Middle Ages
References

Rights & Permissions

Abstract

Horace (65–8 BC), the great lyric poet of the Augustan Age in Rome, composed over a hundred Odes. Scholarly understanding of their early medieval reception has been hampered by the insistence of classical philologists that he was a purely literary poet. Ancient sources and Horace’s own writings demonstrate that he was a performing artist who sang to the accompaniment of his lyre. His use of Alcaic, Sapphic and Asclepiad metres has musical implications. In manuscripts from the ninth to the twelfth centuries, forty-eight passages from the Odes are accompanied by musical notation. The Montpellier codex has notation for the Ode to Phyllis (Odes 4.11) which relates to Guido d’Arezzo’s ‘ut-re-mi’ mnemonic. The St Petersburg codex has settings which suggest various uses, in the schoolroom, abbey entertainments and goliardic performance. The surviving manuscripts were widely spread across Europe and supported a monastic and secular tradition of Horatian song.

Type: Research Article
Information: Early Music History , Volume 40 , October 2021 , pp. 167 - 205

DOI: https://doi.org/10.1017/S0261127921000085 [Opens in a new window]
Creative Commons: This is an Open Access article, distributed under the terms of the Creative Commons Attribution licence (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted re-use, distribution, and reproduction in any medium, provided the original work is properly cited.
Copyright: © The Author(s), 2022. Published by Cambridge University Press

Quintus Horatius Flaccus (‘Horace’) composed his Odes between 30 and 13 BC. For centuries, there was a general perception among classical scholars that he sang most of them in front of an audience to the accompaniment of his own lyre, following the practice of his models Sappho, Alcaeus, Pindar and other Greek lyric poets, whose musicianship is not in doubt.Footnote ¹ Around the turn of thirteenth century, the grammarian Huguccio of Pisa defined an Ode as a ‘cantus’ (‘chant’) or ‘laus’ (‘praise’) that was ‘cantabilis’ (‘capable of being sung’).Footnote ² In France, the music historian Jean-Benjamin de La Borde (1734–94) wrote that the Romans sang almost all their poetry and that the Odes of Horace were probably not recited but sung.Footnote ³ In Germany, the academician August Meineke (1790–1870) formulated the so-called Lex Meinekiana, in which he observed that the number of lines in nearly every Ode was divisible by four and that the quatrains implied a musical dimension.Footnote ⁴

The classical philologist Richard Heinze disagreed. Teaching in Leipzig in the 1920s, he asserted that Horace’s Odes were literary creations in song form.Footnote ⁵ This view became an orthodoxy among classical scholars which persists to the present day. In 1957 Eduard Fraenkel insisted that Horace was a non-musician.Footnote ⁶ Alessandro Barchiesi continues to propose that all Horace’s references to music are ‘at the level of theme, not of performance’ and that any notion of singing is a ‘fiction’.Footnote ⁷ Some take a similar position.Footnote ⁸ Others regard the issue as closed or unimportant.Footnote ⁹

The perpetuation of the orthodox viewpoint can be uncomfortable for musicians and musicologists, whose instincts pull them in another direction. In 2004, Sam Barrett observed in Early Music History: ‘Indications that this poetry was sung suggest a transformation in the reception of classical texts, for despite allusions to sung performance in the texts themselves, most philologists are convinced that, with occasional exceptions, classical Latin lyric poetry was not sung.’Footnote ¹⁰

Recently, T. P. Wiseman has argued that Horace composed Odes for sung performance in full view of the temples of Roman divinities.Footnote ¹¹ It is appropriate to reopen the discussion, so that music historians can consider alternative views of Horace and construct a solid conceptual basis for their research into Horace’s musical reception in the early Middle Ages.Footnote ¹²

Horace’s creative turning point came in about 38 BC, when Virgil introduced him to Maecenas, one of the richest men in Rome, an adviser of Octavian the future Augustus, and sponsor of a poetic circle. Maecenas invited Horace to join his group and their friendship lasted a lifetime. They travelled together.Footnote ¹³ They seem to have been joint survivors of a naval disaster off Cape Palinurus in 36 BC, during the war against Pompeius Sextus.Footnote ¹⁴ Shortly afterwards, Maecenas gave Horace a country estate in the Sabine Hills, an agreeable space for reflection, entertaining and writing. For himself, Maecenas built a splendid mansion on the Esquiline Hill, where he hosted parties in a sunken auditorium with trompe l’oeil wall paintings and an apse with cascading water to cool the air and enhance the acoustics.Footnote ¹⁵

Previously, Horace had composed hexameter satires to amuse his listeners.Footnote ¹⁶ Now in his mid-thirties, he began to experiment with the metres of the Greek lyric poets. The term ‘lyric’ did not have today’s connotations. The Latin adjective ‘lyricus’ was used to describe the poets of antiquity, who composed for the lyre and sang their own compositions. Under the patronage of Maecenas, Horace addressed Odes both to friends and to Rome’s distinguished citizens,Footnote ¹⁷ using Alcaic, Sapphic, Asclepiad and other metres.Footnote ¹⁸

Horace called the Odes ‘carmina’ (‘songs’). In them, he moulded the syntax and inflexions of Latin, with its complex case and verb endings, into the measures of Greece. What might have been a rhetorical strait-jacket became a kaleidoscope of jewel-like phrases. Horace’s metres articulate the Odes, give them their musicality, shape their architecture and make them memorable.Footnote ¹⁹ The first stanza of Odes 1.37, written on the death of Cleopatra, provides an example:Footnote ²⁰

The first two lines of the Alcaic quatrain have an identical rhythm. The long syllables of the third line slow the pace and build momentum like an incoming wave. The fourth line provides musical resolution with two opening dactyls before the tide recedes.Footnote ²¹

Between 30 BC and 23 BC, Horace composed eighty-eight Odes, which he then published in three volumes.Footnote ²² The collection was not enthusiastically received. Maecenas lost public esteem following a political scandal.Footnote ²³ Horace’s presentation of what may (if the hypothesis of orality is valid) have pleased a live audience did not translate into wide readership.Footnote ²⁴ Disappointed, Horace declared ‘non eadem est aetas, non mens’ (‘my age and inclination are not the same’) and reverted to writing agreeable hexameter verses.Footnote ²⁵

Horace’s fortunes revived with the Secular Games of 17 BC. Augustus, as head of the College responsible for sacred rituals, commissioned him to compose the hymn for the occasion.Footnote ²⁶ The Carmen Saeculare was performed on 3 June first on the Capitoline Hill and then on the Palatine by a mixed choir of twenty-seven boys and twenty-seven girls:

The biographer Suetonius, who had access to official archives, confirms that Horace composed the hymn.Footnote ²⁹ His evidence is supported by a carved inscription which reads: ‘Carmen composuit Q Hor..ius Flaccus.’Footnote ³⁰ A possible counterargument that Horace only composed the text is contradicted by Odes 4.6, in which Horace addresses one of the girls who participated in the choral performance:

Despite this evidence, Fraenkel maintained: ‘It is altogether unlikely that Horace, a non-musician, should have undertaken the arduous task of rehearsing and conducting the performance of a choir of amateurs.’Footnote ³² There is no evidential support for the proposition that Augustus’s quindecimviri authorised a hymn for a state occasion from a poet and choir who were not musically competent. Suetonius records that Augustus asked Horace to produce another ‘carmen’ to celebrate the victory of Drusus and Tiberius over the Germanic Vindelici tribe.Footnote ³³ Horace continued to compose Odes in honour of the imperial family.Footnote ³⁴ In 13 BC he brought out a fourth book of Odes at Augustus’s request.Footnote ³⁵

Most classical philologists today reject Fraenkel’s position and acknowledge that Horace composed and conducted the Carmen Saeculare. However, they regard this sung performance as an exception to the norm and insist that the remainder of Horace’s Odes were purely literary compositions. L. P. Wilkinson, however, suggested that the Carmen Saeculare may not have been the first hymn composed by Horace.Footnote ³⁶ The Ode to Mercury (Odes 1.10), dismissed by most critics as a ‘literary imitation’, was described by the early commentator Porphyrio as ‘hymnus’ (‘a hymn’).Footnote ³⁷ The Ode to Diana (Odes 1.21) also has the appearance of hymn:

This juxtaposition of Diana, Apollo and Latona appears to be a direct reference to the sculptural group in the Temple of Apollo, which Augustus dedicated on 9 October 28 BC after his three-day triumph that August.Footnote ³⁸ The ritualised structure of the Ode suggests that it was a choral hymn composed for performance either in the Temple itself or in the temple plaza.

In his Epilogue to Book 3, Horace describes his main achievement as ‘princeps Aeolium carmen ad Italos/deduxisse modos’.Footnote ³⁹ Enoch Powell rendered this curiously as ‘the leader in converting Aeolian poetry into Italian diction’.Footnote ⁴⁰ The literal and correct translation is ‘the first man to have translated Aeolian song into Italian modes’.Footnote ⁴¹ As Cicero explained, ‘in musicis, numeri et voces et modi’ (‘in music, there are numbers and voices and modes’).Footnote ⁴² The numbers are the rhythms or metres; the voices are the pitches or notes; and the modes represent the chosen musical scales with their particular intervallic relationships.

There is an eyewitness to support Horace’s musical claim. The poet Ovid (43 BC–c. AD 17), some twenty years Horace’s junior, reported that he saw and heard Horace giving a live performance in Rome. From his exile on the Black Sea, Ovid recalled the poets he knew and remarked on their different delivery methods:Footnote ⁴³

Ovid’s account provides first-hand evidence that Macer read his treatises aloud, Propertius recited his poems and Horace sang his Odes to the accompaniment of his own lyre. It is hard to argue that the references to the other poets are factual but that the description of Horace is metaphorical.Footnote ⁴⁹ Juvenal, the first-century poet, uses the same phrase as Ovid in describing Horace’s performance method, ‘feriat carmen’:Footnote ⁵⁰

Striking the strings, usually with a plectrum, was what musicians did in Augustan Rome.

The elegist Propertius, some years younger than Horace, was invited to join Maecenas’s circle. Writing of his lover Cynthia in about 24 BC, he confirms that singing songs, while striking the lyre-strings with a plectrum, was part of the accepted culture.Footnote ⁵² According to Sallust, Sempronia, the wife of Decimus Brutus, played the cithara and danced elegantly.Footnote ⁵³ Tiberius, Augustus’s stepson and the future emperor, composed a lyric song.Footnote ⁵⁴

Horace frequently claims to be a player of the lyre, the barbitos and the cithara.Footnote ⁵⁵ In Odes 1.31 he utters a prayer; he is apparently facing the statue of Apollo Citharoedus which stood in the plaza between the Temple of Apollo and Augustus’s mansion.Footnote ⁵⁶ ‘What is the bard’s request to Apollo on his dedication?’ he asks:Footnote ⁵⁷

Horace’s prayer is this: ‘Latoe, dones et precor integra/cum mente nec turpem senectam/degere nec cithara carentem’ (‘Son of Latona, grant, I pray also, that my mind continues unimpaired and that I do not spend my old age in shame or without the cithara’).Footnote ⁵⁸ His prayer over, Horace seems to turn towards his audience, barbitos in hand, as he begins the next Ode:Footnote ⁵⁹

In the Roman Odes, too, Horace creates musical adjacencies.Footnote ⁶² At the end of Odes 3.3, he senses that Juno’s oration on Rome’s destiny has run out of steam:Footnote ⁶³

At the beginning of the next Ode, Horace makes immediate amends:

The juxtaposition is clearly intentional. The context and contents demand oral presentation if Horace’s humour is to work.Footnote ⁶⁵

In Odes 4.9, Horace confirms that he is not just a ‘fidicen’ (‘a player of a stringed instrument’) but a composer: ‘non ante vulgatas per artis/verba loquor socianda chordis’ (‘Through arts not previously disseminated, I utter words to be married to strings’).Footnote ⁶⁶ In Epistles 2.2, he asks: ‘Could I be expected here, in the midst of the swell of business and the storms of the city, to weave words that will set the sound of the lyre in motion?’Footnote ⁶⁷ His musical awareness is confirmed by a joke about Augustus’s elderly cantor Tigellius: ‘si collibuisset, ab ovo/usqu[e] ad mala citaret “Io, Bacche!” modo summa/voce, mod[o] hac resonat quae chordis quattuor ima’ (‘If the mood had taken him, he would sing “Joy, Bacchus!” from hors d’oeuvre through to dessert at the topmost pitch, and then repeat it on the lowest note of the tetrachord’).Footnote ⁶⁸

After the successful performance of the Carmen Saeculare in 17 BC, Horace thanked his Muse for her inspiration and for his public recognition as ‘the minstrel of the Roman lyre’:Footnote ⁶⁹

In the period of his highest creativity, from 30 to 23 BC, Horace’s performances were generally given in private. He performed mainly at his sponsor’s entertainments in the Auditorium of Maecenas, honouring the great and good of Rome and his own friends with personally tailored ‘carmina’. Several reflected Epicurean philosophy; others touched on personal events, feelings and friendships, or supported the legitimacy of the Augustan regime. Some reflected Hellenistic influences.Footnote ⁷¹ In others, he entertains professional women for wine and song.Footnote ⁷² In the Ode to Numida (1.36), Horace describes ‘et ture et fidibus’Footnote ⁷³ (‘with incense and strings’) an ‘amystis’,Footnote ⁷⁴ a drinking contest imported from northern Greece. According to the third-century AD Athenaeus,Footnote ⁷⁵ the competition was inseparable from musical accompaniment. A girl played the pipe, a guest sang to the tune and the revellers competed to see who could take the longest draught without pausing for breath.

Horace’s Metres

While scholars have made marked progress in recent years in the interpretation of early Greek musical inscriptions and papyri,Footnote ⁷⁶ details of Roman musical performance are sketchy. There is no notation in the classical period for Horace’s Odes and the philological focus of Horatian scholars in the century after Heinze has contributed to his metres being regarded as imitative literary conceits, rather than fundamental elements in his performance art. The starting point for understanding Horace’s music has to be metre. He employed three principal metrical systems, Alcaic, Sapphic and Asclepiadean. Like his Greek lyric models, the metres relied on musical feet and rhythms and the placement of long and short syllables.

Alcaics

Horace uses the Alcaic metre in thirty-seven of his 103 Odes, including the six Roman Odes, and it is often associated with a philosophical theme. In Odes 2.14, Horace addresses Postumus:Footnote ⁷⁷

The Alcaic stanza is set in motion by two eleven-syllable lines, each of which has five syllables before the caesura. The third line adds weight and momentum, as all but the third and seventh syllables are long. The fourth line begins with two dactyls and resolves the earlier build-up of weighty syllables. The last two stanzas of the Ode to Postumus illustrate the mood, musicality and architecture of Horace’s metrical treatment:

The pace is measured, the sentiment is Epicurean and the lapidary conclusion creates a natural rallentando. These features are also well illustrated at the close of the Ode to Dellius:Footnote ⁸⁰

The consonantal and vocal harmony and the three successive elisions, one spilling over from the penultimate line, have a powerful musicality.

Sapphics

Horace derived his Sapphics from the songs of Sappho, using the Sapphic metre in twenty-five Odes. Many are hymn-like or offer gnomic advice. The metre is less fluid than Alcaics; typically, the stanzas are self-contained and end-stopped. The first three lines each have eleven syllables, which are identical in terms of quantities, although the beat can vary between monodic and choral treatments. The fourth line is a coda that closes off the stanza.

The Ode to Tranquillity (Odes 2.16), addressed to Grosphus, exemplifies the monodic pattern. There appear to be four beats in each of the first three lines and two beats in the fourth:

In the Ode to Tranquillity, as in most of Horace’s monodic Sapphic Odes, the first three lines have a caesura after the fifth syllable. By contrast, the Carmen Saeculare is a choral piece composed for performance at a religious ceremony. The quantities are the same as in the monodic Ode, but the rhythm is not.Footnote ⁸² The first line signals the difference:

Instead of the regular caesura after five syllables, there is a delayed caesura after six, marked here with a single vertical line. The same feature occurs twenty times in this seventy-six-line choral hymn. The effect is to have only three beats to the line, on the first, fifth and tenth syllables. The rhythmic variation is exemplified in the fourteenth stanza:Footnote ⁸³

This different accentual rhythm may have reflected liturgical practice. The distinction between the monodic and choral Sapphic styles is confirmed in Odes 4.6, when Horace breaks off from his opening theme and embarks on a rehearsal for his choir:Footnote ⁸⁴

Horace is conducting his choirboys and girls in rehearsal. His thumb is striking a surface, possibly the soundbox of his lyre, or perhaps he is using a plectrum. The last stanza of the Ode (‘nupta iam dices…’), discussed earlier, evidences a rhythmic change from hymn to narrative.

Asclepiads

Horace’s third most frequent metrical system is Asclepiadean. His Asclepiads come in various forms and lengths. Five types of Asclepiad metre appear in thirty-four Odes but all have the choriamb as their basic building block: tūm-tí-tí-tūm, like an insistent drumbeat.Footnote ⁸⁷

Horace often chooses Asclepiads to create an agitated mood. The Ode to Pyrrha (Odes 1.5) rushes from stanza to stanza:

The combination of enjambment and the choriambic rhythm convey the anguish of a sore heart.

The Ode to Leuconoē (Odes 1.11) is composed in the Greater Asclepiad metre; it has sixteen syllables to the line. Each line begins and ends with a pair of syllables, which enclose three successive choriambs. Leuconoē needs firm guidance.Footnote ⁸⁸ The percussive beat suggests that Horace’s advice is presented not as a literary homily, but as live entertainment:

Horace’s debt to folk music is evident in the Asclepiadean love duet of Odes 3.9, derived from a traditional carmen amoebaeum (‘alternating song’).Footnote ⁸⁹ A similar tradition is apparent in the Ode to Neobulē (Odes 3.12), a unique example in Horace of the Ionic a minore metre.Footnote ⁹⁰ This Ode is generally discussed as a literary monologue;Footnote ⁹¹ the manuscript tradition divides the lines as follows:Footnote ⁹²

On being set out with two ionic feet to the line, the Ode’s rhythmic intensity becomes obvious:Footnote ⁹³

The starting point of Horace’s Ode is a fragment of Alcaeus.Footnote ⁹⁵ As with choriambs, there is no record in Greek lyric poetry of ionics being spoken rather than sung.Footnote ⁹⁶ The second-century AD grammarian Hephaistion of Alexandria noted that Alcaeus composed many lyric songs in the ionic metre.Footnote ⁹⁷ Horace’s Ode to Neobulē is consistent with performance art. It was acknowledged as a song in the early-eleventh-century goliardic songbook The Cambridge Songs,Footnote ⁹⁸ discussed at the end of this article.

Early Middle Ages

The continuation of a Greco-Roman song culture in Italy after the Augustan Age is confirmed by Suetonius. Tiberius, Nero and Domitian all promoted musical contests. The court of Hadrian had a vibrant musical culture. A third-century statue of ‘Orpheus Citharode’ in Ravenna is a visual reminder of the interest in musical performance.Footnote ⁹⁹ However, the collapse of Rome’s power in the third and fourth centuries, the spread of Christianity and its abhorrence of pagan writings, and invasions from northern Europe had a destructive effect on literary and musical culture. Treatises by Augustine of Hippo and Martianus Capella signalled a partial revival, which gathered pace in the sixth-century court of Theodoric the Great, where Boethius and Cassiodorus both served. The library at Vivarium, established by Cassiodorus, included an edition of Horace.Footnote ¹⁰⁰ Thereafter, there is a codicological lacuna. Manuscripts were copied but were then lost.Footnote ¹⁰¹

The re-emergence of Horatian song in the early Middle Ages was triggered by Charlemagne’s Admonitio Generalis 72 of AD 789. In this directive, Charlemagne called for the founding of new monastery and cathedral schools and for the instruction of boys in mathematics, music and Latin grammar. The Benedictine monasteries of northern France and Aquitaine were at the forefront of the development.Footnote ¹⁰² Selected Odes of Horace became a core element in Latin studies. Horace may have been a pagan but he was a model of grammatical purity.Footnote ¹⁰³ Nor had his rhythms been forgotten. When Paul the Deacon retired to Monte Cassino at about the time of Charlemagne’s edict, he composed a Hymn to St John the Baptist in Horatian Sapphics.Footnote ¹⁰⁴

Manuscripts of Horace’s work were recopied in Benedictine scriptoria. They had a limited lifespan, as they were in regular use. The earliest surviving manuscripts with musical notation date from the ninth century and most are from the eleventh and twelfth.Footnote ¹⁰⁵ Taken as a whole, they reveal that the Odes were chanted or sung in schoolrooms, abbeys and the wider community, over an area extending through Germany, Alsace Lorraine and northern France, to Brittany, Aquitaine and northern Italy.

Many of these manuscripts only came to public attention in 2002 as a result of the late Silvia Wälli’s study Melodien aus mittelalterlichen Horaz-Handschriften,Footnote ¹⁰⁶ which is still being absorbed by music historians. In 1930, Guido Adler was aware of only six melodies.Footnote ¹⁰⁷ Munk Olsen provided additional source material in the 1980s as a result of his wide-ranging study of early medieval Latin reception;Footnote ¹⁰⁸ he was not, however, aware of the important St Petersburg manuscript PET4 until his Addenda et corrigenda of 1989.Footnote ¹⁰⁹

Wälli identified twenty early manuscripts with varying degrees and standards of musical notation. They contain forty-eight notated passages, most of them marked with Aquitanian or Central-French neumes. These include the incipits of twenty-two Odes, the Carmen Saeculare and two of Horace’s early lyric Epodes, appearing as interlinear markings, footnotes or side-glosses. The most popular early medieval melodies, to judge from the frequency of their manuscript appearances, were the Ode to Tibullus (Odes 1.33) with seven appearances,Footnote ¹¹⁰ the prologue to Maecenas (Odes 1.1) with five, the Ode to Virgil (Odes 1.3), also with five, and the Bandusian Spring (Odes 3.13) with four.Footnote ¹¹¹

Working mainly from photocopies, Wälli related some of the melodies to early medieval hymns;Footnote ¹¹² others were from sources she was unable to identify. In Wälli’s view a purpose of the musical notation was to support orally transmitted melodies.Footnote ¹¹³ She believed it was unlikely that performance was restricted to schoolrooms or school audiences.Footnote ¹¹⁴

Five years after Wälli’s study, Jan Ziolkowski argued that a number of these passages were simple chants, recorded in footnotes or glosses and used as classroom aids by the schoolmaster to help students articulate and memorise their Latin.Footnote ¹¹⁵ Most of them were of no more than passing interest to music historians. The rise and fall of the notation reflected the sing-song voice of the teacher and sometimes revealed poor understanding of Horace’s metres.Footnote ¹¹⁶

The Montpellier manuscript M425 and the St Petersburg manuscript PET4 are of critical importance, both as musical records and in informing the debate on early medieval Horatian reception. They contain the most complete notated passages and should be of fundamental interest to historians and musicologists.Footnote ¹¹⁷

The Montpellier Codex M425

The Montpellier codex was copied in the middle of the eleventh century in Aquitaine, a European centre of musical excellence.Footnote ¹¹⁸ The book, which became part of the Pithou collection,Footnote ¹¹⁹ has pages 20 cm high and 14.4 cm wide. It contains a manuscript version of the complete works of Horace on folded quarto vellum sheets. On fols. 50v and 51r, and the first four lines of fol. 51v, Horace’s Sapphic Ode to Phyllis (Odes 4.11) is set out in Carolingian minuscule.Footnote ¹²⁰ (See Figures 1–2.) The text has been carefully formatted with incised horizontal lines to accommodate interlinear Aquitanian notation over the first twenty-four lines.Footnote ¹²¹ The first stanzas, set out below to reflect the manuscript layout and spelling, describe the welcome prepared by Horace for what Phyllis expects will be a pleasant evening with him on the Ides of April:

Figure 1 Montpellier M425: Odes 4.11 Est mihi nonum, rubric and lines 1–14

Figure 2 Montpellier M425: Odes 4.11 Est mihi nonum, lines 15–32

The Ode’s ending is significant both for its musical content and for Horace’s rhythmic treatment in this late composition:Footnote ¹²²

In 1851, a researcher using the pen-name ‘Théodore Nisard’ went to Montpellier to decipher the neumes over the M425 Ode to Phyllis.Footnote ¹²³ He had heard of a similarity between the melodic line of the Horatian Ode and the ‘ut-re-mi’ solmisation mnemonic of Guido d’Arezzo.Footnote ¹²⁴ Guido had explained his sol-fa method in his Epistola de ignoto cantu (‘Letter concerning Unknown Chant’)Footnote ¹²⁵ addressed to Monk Michael at Pomposa in about AD 1030.Footnote ¹²⁶ He set the words of Paul the Deacon’s Hymn to St John the BaptistFootnote ¹²⁷ to a ‘notissima symphonia’ (‘a very well-known melody’), placing alphabetical notes above Paul’s text.Footnote ¹²⁸ Each half-line began one note higher than its predecessor:

Guido’s mnemonic used the first syllable of each half line of Sapphics: ut, re, mi, fa, sol, la. He pitched ‘ut’ on the note of C.Footnote ¹³⁰

Guido was a product of central Italy. Trained at Pomposa and now choirmaster in Arezzo, he taught his choristers with a monochord. In rehearsal, he pointed to the four fingers of his left hand and the spaces in between, explaining that notes on the same line or space were always on the same pitch. He set out his system in the Micrologus and subsequently transposed the lines and spaces onto a stave.Footnote ¹³¹ In his Antiphoner he coloured the line for C in yellow and the line for F in red, teaching his students to extend the musical range to a gamut of twenty-one notes, which he designated with alphabetical markings.Footnote ¹³²

Nisard did not delve into the relationship of M425 with Guido’s source music. He made the mistake of pitching the Horatian Ode’s first word ‘Est’ on D.Footnote ¹³³ His misinterpretation was widely accepted for the next 150 years, despite muddying the connection with Guido.Footnote ¹³⁴

The musical connection between Guido’s ‘ut-re-mi’ and the M425 setting of Horace’s Ode to Phyllis is unchallengeable if the first word of the Ode is pitched correctly. Nisard’s confusion may have arisen because the horizontal rope design above the incipit, together with the large opening E, interfered with the neumes. (See Figure 1 above.) The second stanza of the Ode, where the neumes are unimpeded, provides a better comparator. By matching Guido’s alphabetical notation to the Aquitanian neumes when pitched on C, it is possible to compare the Guidonian and Horatian versions, and decipher the M425 notation:Footnote ¹³⁵

Although Guido’s Epistola de ignoto cantu predated M425 by about two decades, the date when the Aquitanian scribe recorded the melody for the Ode to Phyllis is not evidence of when he first heard it. What we can deduce is that there are two surviving and closely related tunes of broadly similar date, one recorded in Guido’s Epistola and the other in M425. Guido’s version covered a single stanza of Sapphics. M425 carries notation for the first six of the nine Horatian stanzas, twenty-four lines in all.

There are three features in the Aquitanian notation that are particularly worthy of note. In the classical period, metrical analysis suggests that Horace, like Sappho, used only one note for each syllable of text. The early medieval setting in M425 is different. First, the treatment of the opening syllables of each Sapphic line varies as the stanza progresses; in the first line of every stanza the first syllable is sung on a single note but in lines 2 and 3 multiple notes are applied to each syllable.Footnote ¹³⁶ Second, additional notes are introduced above some syllables at the middle or end of a line;Footnote ¹³⁷ the underlying rhythm is Sapphic, but the ‘notissima symphonia’ develops with dramatic effect. Third, while the starting notes and elaborations in M425 are largely common to Guido d’Arezzo’s melody, there are two significant instances of artistic elaboration, which indicate that the melody was no mere adaptation of a Guidonian chant to help boys memorise their notes but performance art conceived and delivered to a high standard.

Horace’s Ode pivots architecturally on the word ‘quod’ (‘because’), the fourth word in the second line of the fifth stanza, after a pause at the midpoint of the nine-stanza Ode. Before this, Phyllis imagines Horace’s party is for her. Afterwards, she realises it is for Maecenas. Horace teases her, then he consoles her. In each of the other stanzas, the equivalent syllable to ‘quod’ is sung on the single note of G. Here, there is a pair of rising puncta (FG): ‘quo-ód’ (‘becau-aúse’), hesitation followed by revelation. This is the work of a scholar who understands the text and has the dramatic and musical skill to respond to it.

The scribe gives another example of his ability in the next stanza. Horace is describing Phyllis’s rival for the affections of Telephus:

Horace has used a delayed caesura after ‘lasciva’. He could have written:

Horace chose not to. A delayed caesura and the adjective ‘lasciva’ proved too tempting. The effect is to draw out the second syllable of ‘lasci-i-và’ (‘se-exy’) and to follow it with a nose-dive, a pause, a gulp and a squeak. The final syllable ‘-và’ falls from G by a fourth, perhaps even a fifth; after a brief hiatus, the first syllable of ‘ténet’ rises sharply to A, before a scandicus signals a three-note ascent, apparently starting on G. The scribe has offered a first-hand response to the text and validated the Ode to Phyllis as performance art.

Any attempt to transcribe the M425 music has inherent risks and uncertainties. Aquitanian neumes are diastematic and give no indication of either starting pitch or rhythm. Nevertheless, Guido’s use of alphabetic notation with a starting pitch of C, together with his selection of a well-known verse composition modelled on Horace’s Sapphic metre, allow a reasonable assumption of a musical line and note lengths to be made. The Sapphic metre suggests something approximating today’s 3/4 time in place of the common time assumed by Nisard (see Example 1).Footnote ¹³⁸

The St Petersburg codex PET4

The St Petersburg manuscript dates to the late twelfth century and originated in a Benedictine scriptorium at Corbie or St-Germain-des-Prés in Paris. It is both a pedagogical edition of the complete works of Horace and a document of musical record for the Odes, which appear in the first thirty-two of the eighty-seven folios. Towards the back of the codex, from folios 84v to 87v, there is a metrical commentary in the same hand, which underlines the book’s serious purpose as an aid to study and analysis. The pages average about 22.25 × 10 cm. The leaves have been ruled with a fine implement to create a frame for the principal text, which typically measures about 19.7 × 6.25 cm and leaves a distance from the frame to the edge of about 3.2 cm. There are horizontal ruled lines about 4 mm apart to regulate the text and neumes.Footnote ¹³⁹

The PET4 text is written in brown ink fading to sepia. The manuscript style is a carefully formed, slightly italicised, northern-French Carolingian minuscule, which supports the presumed provenance. There is interlinear musical notation in well-formed central-French neumes above the incipits of the first eleven Odes.Footnote ¹⁴⁰

The neumes are part of a planned project, drawn with confidant penmanship and including both simple notes and ligatures. None of the neumed passages proves an unbroken link with Horace or pre-Carolingian Horatian song but some demonstrate that in the early medieval period the Odes were sung in performance as well as chanted in the schoolroom. The notation is more than an aide-memoire; it is a permanent musical record, a virtual songbook, supported by a metrical commentary.

Three of the neumed passages, in Odes 1.1, 1.6 and 1.9, are reasonably clear and illustrate the diversity of the musical tradition on which the scribe has drawn.

Odes 1.1 to Maecenas

Odes 1.1, the prologue to Maecenas, is one of Horace’s compositions for which there is no credible case for sung performance in antiquity. It was a literary dedication composed in continuous Asclepiads and written shortly before the publication of the first three volumes of the Odes. In PET4, the limited heightening of the neumes produces a mundane musical line and this is also a feature of the four other neumed manuscripts of Odes 1.1 identified by Wälli.Footnote ¹⁴¹ (See Figure 3.) The version is rhythmically consistent with Horace’s Asclepiads but offers little of musical interest. It reflects a pedagogical tradition shared across northern Europe over three centuries.Footnote ¹⁴²

Figure 3 St Petersburg PET4: Odes 1.1 Maecenas atavis, rubric and lines 1–8

Odes 1.6 to Agrippa and 1.33 to Tibullus

The neumed incipit for the Ode to Agrippa (Odes 1.6) has an unusual purpose.Footnote ¹⁴³ The melody is shown above the first four lines but the tune ‘belongs’ to Odes 1.33, the Ode to Tibullus, also composed in Asclepiad quatrains.Footnote ¹⁴⁴ (See Figure 4.) This is evident from the fact that the melody, or a slight variation, invariably occurs, when it does occur, in association with Odes 1.33. It appears in that relationship twice in glosses to M425.Footnote ¹⁴⁵ It occurs similarly in four other neumed manuscripts.Footnote ¹⁴⁶ It is never, other than in PET4, linked with Odes 1.6.

Figure 4 St Petersburg PET4: Odes 1.6 Scriberis Vario, rubric and lines 1–4

There is an obvious reason for the melody appearing above Odes 1.6 in the St Petersburg codex. The Ode is one of Horace’s Parade Odes.Footnote ¹⁴⁷ As this marks the first appearance of this Asclepiadean metre in PET4, a document of musical record, the scribe has chosen to showcase the melody in this location.Footnote ¹⁴⁸ Odes 1.6 is a ‘recusatio’ (‘refusal’) in which Horace turns down a request from Agrippa to honour his military achievements. The neumes are a rhythmic fit with the text, but an aesthetic misfit:

Once the music is transferred to its normal home of the Ode to Tibullus,Footnote ¹⁵¹ there is obvious artistic synergy:

The underlying choriambs are softened by a medieval song form. The racy subject matter as the Ode progresses, together with its wide dissemination, implies that the rendition was more closely related to goliardic performance than to monastic plainchant.Footnote ¹⁵³

Odes 1.9 Ode to Thaliarchus (Soracte) Footnote ¹⁵⁴

The Soracte Ode (Odes 1.9) is one of Horace’s best loved. It is the only one of his thirty-seven Odes in the Alcaic metre for which an early medieval setting survives. The neumes cover the entire first quatrain. The first line appears at the foot of fol. 3r below an introductory rubric.Footnote ¹⁵⁵ The second line follows at the top of fol. 3v, where the vertical space makes the heightening of the neumes exceptionally clear. (See Figure 5a–b.) The third and fourth lines also have clear markings. A tentative attempt to transcribe the music is shown in Example 2. The final melismas are problematical but are a valuable indication of the style of performance.

Figure 5 St Petersburg PET4, Odes 1.9 Vides ut alta: (a) rubric and line 1; (b) lines 2–4

Example 1 Transcription of Odes 4.11 Est mihi nonum from Montpellier M425. Source: Lyons, Horace’s Odes and the Mystery of Do-Re-Mi

Example 2 Transcription of Odes 1.9 Vides ut alta from St Petersburg PET4. Source: Lyons, Music in the Odes of Horace

Notwithstanding the challenges, the strophic composition makes it possible to deduce a musical line and rhythm for the entire Ode:

Horace’s Alcaics were modelled on Greek lyric poetry composed some six hundred years before his time. The interval between Horace’s Soracte Ode and PET4 was nearly twice as long. The underlying Alcaics can be discerned, but many syllables have been extended and there are significant musical elaborations. The diction and pace have changed, reflecting the development of vernacular Latin and popular song. The beauty of the musical line and the complex final melismas suggest that this was a concert piece performed by a practised singer. It is a jewel in the early medieval reception of Horace’s Odes.

The Benedictine manuscripts identified by Wälli do not hold all the clues to early medieval performance. There was a separate goliardic tradition. The Cambridge Songs, a collection of early medieval songs datable to AD 1039 and earlier, originated in Germany.Footnote ¹⁵⁸ It contains an un-neumed version of Horace’s Ode to Neobulē (Odes 3.12).Footnote ¹⁵⁹ The text has misspellings, misunderstandings and misplaced words.Footnote ¹⁶⁰ But the principal message is that, not far from the time when the M425 scribe was recording the Ode to Phyllis in Aquitaine, an English traveller to the Rhineland heard a Horatian Ode being performed in the secular repertoire. The channels of transmission were oral and the melodies were spread through performance.Footnote ¹⁶¹ The traveller brought the text of the Ode to Neobulē to the church of St Augustine in Canterbury, where it was copied with other songs in the anthology.

This evidence serves as a reminder that the old distinction between the classical and medieval periods, with the imagined ‘Dark Ages’ in between, is counterproductive.Footnote ¹⁶² History operates as a continuum. The Carolingian call for revival in the monastery schools could not have been implemented unless there was a literary and musical tradition worth reviving. The goliardic song tradition survived because it appealed to the popular imagination. The musical performers of Horace’s Odes in the early medieval period seem to have been prompted by a long history of Horatian song which began with Horace himself.

London

References

¹ M. Bowra, Greek Lyric Poetry from Alcman to Simonides (Oxford, 1971); G. Kirkwood, Early Greek Monody: The History of a Poetic Type (Ithaca, NY, 1974).

² Huguccio of Pisa (grammarian), Derivationes, 2 (Florence, 2004), pp. 862–3, noted by A.-Z. Rillon-Marne, ‘L’Ode au Moyen Âge: Une lyricité en sommeil?’, Camenae, 20 (2017), pp. 1–16, at p. 2.

³ J.-B. F. de La Borde, Essai sur la musique ancienne et moderne, i (Paris, 1780), p. 43: ‘On sait que les Romains … chantaient presque toutes leurs poésies. Probablement, les Odes d’Horace ne se déclamaient point, mais se chantaient.’

⁴ A. Meineke, Horaz, introduction to 2nd edn. (Berlin, 1854).

⁵ R. Heinze, ‘Die Horazische Ode’, Neue Jahrbücher, 51 (1923), pp. 153–68; translated in Horace: Odes and Epodes, ed. M. Laurie (Oxford, 2009), pp. 11–32.

⁶ E. Fraenkel, Horace (Oxford, 1957), pp. 403–4.

⁷ A. Barchiesi, ‘Carmina: Odes and Carmen Saeculare’, in S. Harrison (ed.), The Cambridge Companion to Horace (Cambridge, 2007), pp. 144–6.

⁸ R. Tarrant, ‘Lyricus Vates: Musical Settings of Horace’s Odes’, Yale Classical Studies, 36 (2012), pp. 72–93; S. Harrison, ‘Style and Poetic Texture’, in Harrison (ed.), The Cambridge Companion to Horace, pp. 262–73. S. Heyworth, ‘Sappho in Propertius?’, in T. Thorsen and S. Harrison (eds.), Roman Receptions of Sappho (Oxford, 2019), pp. 185–204, at p. 189, believes that the light Lesbian wine poured by a musician in Odes 1.17.21–2 is a metaphor for a book of poems in imitation of Sappho and Alcaeus; for the correct view see R. G. M. Nisbet and M. Hubbard, A Commentary on Horace Odes, Book 1 (Oxford, 1970), p. 225.

⁹ Counterarguments by N. Bonavia-Hunt, Horace the Minstrel – a Practical and Aesthetic Study of his Aeolic Verse (Kineton, 1969), failed to gain traction. This monograph, published by his family after his death, was belittled on its own dust jacket and in Quintin Hogg’s foreword.

¹⁰ S. Barrett, Review of S. Wälli, Melodien aus mittelalterlichen Horaz-Handschriften, Early Music History, 23 (2004), pp. 285–305, at p. 286.

¹¹ T. P. Wiseman, The House of Augustus (Princeton, 2019), pp. 141–5.

¹² See, for example, Wulf Arlt, ‘Hymnus und Ode: Horazvertonungen des Mittelalters’, in A. Haug, C. März and L. Welker (eds.), Der lateinische Hymnus im Mittelalter: Überlieferung – Ästhetik – Ausstrahlung, Monumenta Monodica Medii Aevi, Subsidia IV (Kassel, 2004), 133–6.

¹³ Horace, Satires 5.27–104.

¹⁴ S. Lyons, ‘Death at Sea’, Ad Familiares, 35 (Oct. 2008), pp. 3–4. Dio Cassius, Roman History, 49.1.

¹⁵ The Auditorium of Maecenas, illustrated and discussed in S. Lyons, Music in the Odes of Horace (Oxford, 2010), pp. 43–5 and p. 197 (colour plates 1 and 2). See also E. Carnabuci, Auditorium di Mecenate (Rome, 2005). In Epistles 1.19.39–49, Horace explains that his reluctance to perform in a crowded theatre is voluntary, invalidating Rossi’s assertion that he could not find an institutionally designated place for singing poetry in Rome; L. E. Rossi, ‘Orazio, un lirico greco senza musica’, Seminari Romani di Cultura Greca, 1 (Rome, 1998), pp. 163–81, at p. 167, translated as ‘Horace, a Greek Lyricist without Music’, in M. Lowrie (ed.), Horace: Odes and Epodes, Oxford Readings in Classical Studies (Oxford, 2009), pp. 356–77, at p. 359.

¹⁶ Horace’s first book of Satires was published in 35 BC.

¹⁷ R. Syme, The Augustan Aristocracy (Oxford, 1989), pp. 382–402. S. Lyons, Music in the Odes of Horace, pp. 55–62.

¹⁸ Derived from Alcaeus (621–560 BC), Sappho of Lesbos (c. 630–c. 570 BC), and Asclepiades of Samos (born c. 320 BC).

¹⁹ Mnemosynē (Greek Mνημoσύνη), the goddess of memory, was the mother of the Muses.

²⁰ In Latin verse, metre is determined not by stress but by ‘quantity’, i.e. whether a syllable is short or long. In this article, short syllables are marked with an accent (á) and long syllables with a horizontal stroke (ā). This change from normal textbook practice has been introduced because non-classicists consulted have said that forward-flowing markings make it easier to follow the rhythmic line. The double vertical lines in verses 1 and 2 indicate the ‘caesura’ (break), which shapes the metre and articulates the verse. The last syllable of ‘Saliaribus’ is naturally short, being followed by a vowel, but the metrical pause at the end of the line gives the short syllable the duration of a long.

²¹ The verse translations are adapted from S. Lyons, The Fleeting Years: Odes of Horace from the Augustan Age of Rome. A New Verse Translation with Introduction and Notes (Stoke-on-Trent, 1996) and Horace’s Odes and the Mystery of Do-Re-Mi (Oxford, 2007). The Horatian text here and elsewhere is taken from Horatii Opera, ed. E. C. Wickham (Oxford, 1901), unless otherwise stated.

²² E. J. Kenney (ed.), The Cambridge History of Classical Literature, ii (Cambridge, 1982), p. 12: ‘In general it may be taken for granted that throughout antiquity books were written to be read aloud, and that even private reading often took on some of the characteristics of a modulated declamation. It might be said without undue exaggeration that a book of poetry or artistic prose was not simply a text in the modern sense but something like a score for public or private performance.’

²³ Licinius Murena, half-brother of Maecenas’s wife Terentia, was implicated in an attempted coup d’état against Augustus. Maecenas is believed to have warned her that Murena’s life was in danger. Murena was executed, allegedly while trying to escape, and Maecenas fell from power following his indiscretion.

²⁴ Horace describes the transition from oral to written presentation in Epistles 1.19.32–4, published in 20 BC.

²⁵ Horace, Epistles 1.1.17.

²⁶ Martha W. Hoffman, ‘The College of Quindecimviri (Sacris Faciundis) in 17BC’, American Journal of Philology, 73 (1952), pp. 289–94, notes that Augustus, in the name of ‘Imp. Caesar’, was ‘magister’ and head of the list; his close associate Marcus Agrippa was second.

²⁷ The metre is Sapphic. The vertical line indicates a delayed caesura after the sixth syllable; the normal fifth-syllable caesura does not occur, for reasons discussed later in this article.

²⁸ Horace, Carmen Saeculare, 1–4.

²⁹ Suetonius (c. AD 69–122), Life of Horace, 2: ‘Augustus … Saeculare Carmen componendum iniunxerit’ (‘Augustus enjoined him to compose the Centennial Hymn’).

³⁰ Corpus Inscriptionum Latinarum, 6.32323. There is no mention in the inscription of a separate musical composer; contrast the comedies of Terence, where it is recorded that Flaccus the son of Claudius ‘modos fecit’ (‘created the music’); see title pages in Terenti Comoediae, ed. R. Kauer and W. Lindsay (Oxford, 1926).

³¹ Horace, Odes 4.6.41–4. Nisbet and Hubbard, A Commentary on Horace’s Odes, Book 1, pp. 15 and 349, argue that the term ‘vates’ (‘seer’), the oldest Latin term for ‘poet’, was ‘ennobled’ in the Augustan age and implied inspiration by the Muses. P. T. Wiseman, Unwritten Rome (Exeter, 2008), p. 44: ‘Prophecy was primarily an oral mode; the words used of it were “song” and “singing”, cantare and canere.’

³² Fraenkel, Horace, pp. 403–4.

³³ Suetonius, Life of Horace, 2. The carmen celebrating the victory of Drusus and Tiberius survives as Odes 4.4.

³⁴ Odes 4.1 for Paullus, the bridegroom of Augustus’s niece Marcia; Odes 4.2 to Iullus; and Odes 4.5, 4.14 and 4.15 in honour of Augustus.

³⁵ Suetonius, Life of Horace, 2.

³⁶ L. P. Wilkinson, Horace and his Lyric Poetry (Cambridge, 1945), p. 14, n. 3, suggests that Odes 1.10 and 1.21 are hymns.

³⁷ Nisbet and Hubbard, A Commentary on Horace’s Odes, Book 1, pp. 125–7. Mercury was Horace’s guardian god, the protector of men of commerce (‘Mercuriales’) and the father of the lyre.

³⁸ Propertius, 2.31.15–16, who was present at the ceremony, confirms that the cult statue of Apollo stood between Latona and Diana: ‘deinde inter matrem deus ipse interque sororem/Pythius in longa carmina veste sonat’ (‘Between his mother and sister, the Pythian god himself sounds his songs, clad in a long garment’).

³⁹ Odes 3.30.13–14.

⁴⁰ J. E. Powell speech archive (Horatian Society, London, 21 July 1992), pp. 43–51.

⁴¹ Odes 3.30.14–15. Aeolian refers to coastal Asia Minor.

⁴² Cicero, De Oratore, 1.42.147. Horace, Epistles, 2.2.143–4, also distinguishes metre from mode: ‘ac non verba sequi fidibus modulanda Latinis,/sed verae numerosque modosque ediscere vitae’ (‘not to hunt for words that can be set to Latin strings, but to learn the metres and modes of a true life’)

⁴³ Ovid, Tristia, 4.10.41–52.

⁴⁴ Aemilius Macer (died 16 BC). His two works were derived from the Ornithogonia of the Greek poet Boios and the Theriaca of Nicander of Colophon.

⁴⁵ Sextus Propertius (died c. 15 BC), author of four books of Elegies.

⁴⁶ The identities of Ponticus and Bassus have been subjects of speculation. Propertius addressed Elegies 1.7 and 1.9 to Ponticus and 1.4 to Bassus. Horace refers to a friend named Bassus in Odes 1.36.14.

⁴⁷ Ausonian lyre: ‘Italian’, or more specifically ‘middle or lower Italian’.

⁴⁸ Virgil and Tibullus both died in 19 BC.

⁴⁹ R. Tarrant, ‘Ancient Receptions of Horace’, in Harrison (ed.), The Cambridge Companion to Horace, pp. 277–90, at pp. 277–8 and n. 3, and S. Harrison, Horace: Odes Book 2 (Cambridge, 2017), p. 19. Both writers quote only the first line of the couplet (‘et tenuit nostras numerosus Horatius aures’) and omit the second (‘dum ferit Ausonia carmina culta lyra’).

⁵⁰ Juvenal, Satires, 7.53–5 and 62.

⁵¹ The same verb that Horace uses in Odes 3.30.14, quoted above.

⁵² Propertius, 2.3.19–20.

⁵³ Sallust, Bellum Catalinae, 25.2.

⁵⁴ Suetonius, Life of Tiberius, 70: ‘composuit et carmen lyricum, cuius est titulus “Conquestio de Morte L. Caesaris”’ (‘He also composed a lyric song, whose title was “Lament on the death of Lucius Caesar”’).

⁵⁵ For Horace’s references to the lyre and other instruments in the Odes, see S. Lyons, Music in the Odes of Horace, pp. 91–2. The ‘barbitos’ was a long-stringed baritone lyre about 50 per cent longer than the traditional instrument; see J. G. Landels, Music in Ancient Greece and Rome (London and New York, 1999), pp. 66–7. The cithara was a stringed instrument placed on the knee or a flat surface and preferred by professional citharists or citharodes.

⁵⁶ Propertius, 2.31.5–6, admired the statue at the opening ceremony: ‘hic equidem Phoebo visus mihi pulchrior ipso/marmoreus tacita carmen hiare lyra’ (‘Indeed, he seemed to me more beautiful than Phoebus himself, a marble statue, the mouth open in song to a silent lyre’).

⁵⁷ Odes 1.31.1–2.

⁵⁸ Odes 1.31.17–20. Nisbet and Hubbard, A Commentary on Horace’s Odes, Book 1, pp. 347–8, argue that the Ode was delivered in front of the god’s statue at the festival of the Meditrinalia on 11 October 28 BC, two days after the consecration of the temple. This view is supported by T. P. Wiseman, The House of Augustus, p. 144.

⁵⁹ Odes 1.32, 1–4 and 13–16.

⁶⁰ ‘Barbitos’, a longer instrument, is here translated as ‘Grecian lyre’ for the purposes of rhyme.

⁶¹ The tortoiseshell was the sound box of the lyre. Compare Odes 3.11.3–4: ‘tuque testudo resonare septem/callida nervis’ (‘and you, tortoiseshell, practised at resounding with seven strings’).

⁶² The Roman Odes is a name commonly given to Odes 3.1 to 3.6 inclusive.

⁶³ Odes 3.3.69–72.

⁶⁴ Horace gives a similar instruction to his ‘Musa procax’ (‘wayward Muse’) in Odes 2.1.40: ‘quaēré módōs lévíōré plēctrō’ (‘choose modes with a lighter plectrum’).

⁶⁵ Odes 1.24 is a funerary lament which has similar musical implications: ‘praecipe lugubris/cantus, Melpomene, cui liquidam pater/vocem cum cithara dedit’ (‘Teach us, Melpomene, your mournful songs with voice limpid and high that with the cithara your father gave.’)

⁶⁶ Odes 4.9.3–4. Internal evidence in the Ode suggests that it was written between 16 BC, when the recipient Marcus Lollius was replaced as a legate in Gaul, and 13 BC, when Book 4 was published.

⁶⁷ Epistles 2.2.84–6: ‘hic ego rerum/fluctibus in mediis et tempestatibus urbis/verba lyrae motura sonum conectere digner?’

⁶⁸ Satires 1.3.6–8.

⁶⁹ Odes 4.3.17–24. Compare also Odes 3.4, 1–4 and 21–8, addressed to the muse Calliope.

⁷⁰ The allusion is to Pindar, ‘the Swan of Dirce’ and a model for Horace in several Odes in Book 4; see Odes 4.2.1–32 (compare Odes 2.20.9–16).

⁷¹ O. Murray, ‘Symposium and Genre in the Poetry of Horace’, Journal of Roman Studies, 75 (1985), pp. 39–50. G. Davis, ‘Wine and the Symposium’, in Harrison (ed.), The Cambridge Companion to Horace, pp. 207–20. Horace’s ‘sympotic’ Odes include 1.27, 2.5 and 2.11.

⁷² Odes 1.17, 3.28 and 4.11.

⁷³ Odes 1.36.1.

⁷⁴ Greek ἄμυστις.

⁷⁵ Athenaeus, Deipnosophistai, 11.783e.

⁷⁶ A. Barker, Greek Musical Writings (Cambridge, 1984); M. L. West, Ancient Greek Music (Oxford, 1994); E. Pöhlman and M. L. West, Documents of Ancient Greek Music (Oxford, 2001); A. D’Angour, ‘Ancient Greek Music: Now We Finally Know What It Sounded Like’ (www.theconversation.com, 31.07.2018).

⁷⁷ Identified with the Postumus of Propertius, book 3, poem 12; see Nisbet and Hubbard, A Commentary on Horace Odes, Book 2 (Oxford, 1978), p. 223.

⁷⁸ The image of ‘placens uxor’ (‘pleasing wife’) is probably derived from Lucretius, De Rerum Natura, 3.894–6. Postumus’s wife is believed to have been Aelia Galla, noted for her propriety; see J. Wallis, ‘Marriage and the Elegiac Woman in Propertius 3.12’, Ramus, 40/2 (2011), pp. 106–29.

⁷⁹ Literally, pure unmixed wine.

⁸⁰ Odes 2.3.25–8.

⁸¹ ‘Otium’ is the peacefulness of leisure, the opposite of ‘negotium’ (‘business’).

⁸² Noted by A. D’Angour, ‘Horace’s “Victory Odes”’, Bulletin of the Institute of Classical Studies (London, 2012), pp. 57–71.

⁸³ Carmen Saeculare, 53–6.

⁸⁴ Odes 4.6.25–36. The choral rhythm is signalled by the delayed caesura in line 3 of this extract.

⁸⁵ Thalia, the daughter of Zeus and Mnemosynē, was the eighth-born of the Muses.

⁸⁶ Epithet of Apollo (Greek Ἀγυιεύς), guardian deity of streets and highways (ἄγυιαι), here in the vocative case.

⁸⁷ J. Halporn, M. Ostwald and T. Rosenmeyer, The Metres of Greek and Latin Poetry (Indianapolis, 1994), pp. 29–31, note that in Greek quantitative verse there is no surviving example of choriambs or ionics being spoken rather than sung.

⁸⁸ The Greek name Λευκoνόη (‘white–minded’) suggests empty-headedness or naïveté.

⁸⁹ Translated in Lyons, Horace’s Odes and the Mystery of Do-Re-Mi, pp. 145–6.

⁹⁰ As her name suggests, Neobulē (Greek Nεoβoύλη) has new-fangled ideas.

⁹¹ K. Quinn, Horace – The Odes (London, 1980), p. 266. S. Thom, ‘Staking a Claim for Lyric Achievement: Horace, “Odes” 3.7–12’, Acta Classica, 53 (Classical Association of South Africa, 2010), pp. 69–93, at p. 86, describes the Ode as a ‘stock literary situation’.

⁹² Horatii Opera, ed. Wickham.

⁹³ Nisbet and Hubbard, A Commentary on Horace’s Odes, Book 1, p. 166, offer further possibilities for metrical layout.

⁹⁴ Literally, ‘the beauty of Hebrus from Lipara’.

⁹⁵ Alcaeus, fragment 10B, in Greek Lyric 1, Sappho and Alcaeus, ed. D. A. Campbell, Loeb Classical Library (Cambridge, MA, 1990), p. 242: ἔμε δείλαν, ἔμε παίσαν κακoτάτων πεδέχoισαν.

⁹⁶ Halporn, Ostwald and Rosenmeyer, The Metres of Greek and Latin Poetry, pp. 29–31.

⁹⁷ Hephaistion, ‘Enchiridion de Metris’, fr. 135, in Greek Lyric 1, Sappho and Alcaeus, p. 151.

⁹⁸ See below, nn. 158–9.

⁹⁹ Lyons, Music in the Odes of Horace, pp. 93–6.

¹⁰⁰ J. W. Thompson, The Medieval Library (repr. New York, 1957), p. 40.

¹⁰¹ Horace’s literary survival was largely due to just two surviving fontes; see Lyons, Music in the Odes of Horace, pp. 96–9.

¹⁰² J. Grier, The Musical World of a Medieval Monk (Cambridge, 2006). Adémar de Chabannes (c. AD 989–1034), Chronicles 2.8. For library collections, see Thompson, The Medieval Library.

¹⁰³ R. Glaber (AD 985–1047), Historiarum libri quinque, ed. and trans. John France (Oxford, 1989), 2.23, writes of demons assuming the guise of Virgil, Horace and Juvenal to tempt the grammarian Vilgard, who was later found guilty of heresy. An edition of Horace was, however, held at Cluny; see B. Gillingham, The Social Background to Secular Medieval Latin Song, Musicological Studies, LX/3 (Ottawa, 1998), p. 119. See also H. Fichtenau, Living in the Tenth Century (Chicago, 1991), p. 288.

¹⁰⁴ Lyons, Music in the Odes of Horace, pp. 110–11.

¹⁰⁵ The Weissenberg manuscript, now in the Vatican, MS Reg. lat. 1703, dates to the first half of the ninth century and has notation over lines 5 to 8 of Ode 1.3. The text has corrections in the hand of Walafrid Strabo, tutor to Charles the Bald until AD 838 and Abbot of Reichenau to his death in AD 849.

¹⁰⁶ Kassel, 2002.

¹⁰⁷ G. Adler, Handbuch der Musikgeschichte (Berlin, 1930), p. 60.

¹⁰⁸ B. Munk Olsen, L’Étude des auteurs classiques latins aux XIe et XIIe siècles, 1 (Paris, 1982).

¹⁰⁹ B. Munk Olsen, Addenda et corrigenda (Paris, 1989), pp. 68–9.

¹¹⁰ Including the association with Ode 1.6 in the St Petersburg manuscript, described later in this article.

¹¹¹ For a detailed Table of Neumed Early Medieval Manuscripts of Horace, see Lyons, Music in the Odes of Horace, p. 136.

¹¹² For example, the melody for Ode 1.15 shown in a footnote to the Paris manuscript 7979 is related to a hymn tune for Prudentius’s Inventor rutili sung on the lighting of the Easter fire and recorded in Laon manuscript 263.

¹¹³ Wälli, Melodien aus mittelalterlichen Horaz-Handschriften, p. 345.

¹¹⁴ Ibid ., p. 354.

¹¹⁵ J. M. Ziolkowski, Nota Bene: Reading Classics and Writing Melodies in the Early Middle Ages (Turnhout, 2007). See also Suzanne Reynolds, Medieval Reading: Grammar, Rhetoric, and the Classical Text (Cambridge 1996), p. 14, and ‘Glossing Horace: Using the Classics in the Medieval Classroom’, in C. Chavannes-Mazel and M. Smith (eds.), Medieval Manuscripts of the Latin Classics: Production and Use (Leiden, 1993), pp. 103–17.

¹¹⁶ For example, the formulaic jingles for Odes 1.5, 3.9, 3.12 and 3.13 in schoolmasters’ footnotes in the Paris manuscript PA7979.

¹¹⁷ I thank the staff at the Faculté de Médecine library in Montpellier and the National Library of Russia in St Petersburg for enabling me to inspect these manuscripts. I also thank James Grier, Professor of Music History at the University of Western Ontario, for his advice and encouragement during my research.

¹¹⁸ It is wrongly dated as tenth century in the Catalogue général des manuscrits des bibliothèques publiques, 1 (Paris, 1849).

¹¹⁹ For the provenance, see Lyons, Music in the Odes of Horace, p. 101.

¹²⁰ Carolingian minuscule script was widely used in the Germanic and Frankish regions between AD 800 and 1200.

¹²¹ Aquitanian neumes are diastematic; the accuracy of the reading depends on the correct identification of the starting pitch.

¹²² Odes 4.11, published in 13 BC.

¹²³ Théodore Nisard was the nom-de-plume of Abbé Théodule Éléazar Xavier Normand (1812–88), a Paris-based church organist who had been born in Quaregnon outside Mons. The Bibliothèque nationale de France lists him as a musical biographer and compiler of plainchant studies. Grove’s Dictionary of Music and Musicians (1900 edition) describes him as having ‘a controversial turn of mind’.

¹²⁴ Guido d’Arezzo (AD 991/2–post 1033), choirmaster. The connection was reported by J.-B. Laborde, Essai sur la musique ancienne et moderne, i (Paris, 1780), p. 43.

¹²⁵ For text and translation, see Lyons, Music in the Odes of Horace, pp. 108–11.

¹²⁶ The letter’s date falls between August 1028, when Guido had an audience with Pope John XIX in Rome, and October 1032, when the Pope died; see Lyons, Music in the Odes of Horace, pp. 108–10. Guido had been a novice at Pomposa and spent his early life there; see L. Angeloni, Sopra la via, le opere, ed il sapere di Guido d’Arezzo, restauratore della scienza e dell’arte musica: dissertazione (Paris, 1811); G. Billanovich, Pomposa monasterium modo in Italia primum (Padua, 1994); A. Samaritani, Presenza monastica ed ecclesiale di Pomposa nell’Italia centrosettentrionale, secoli X–XIV (Ferrara, 1996).

¹²⁷ As noted above, Paul the Deacon’s ‘Hymn’ was a poem in Horatian Sapphics. There was no contemporary musical setting.

¹²⁸ Guido’s lower case ‘a’ signifies high A, the note immediately above G.

¹²⁹ Literally: ‘In order that your servants may with loosened chords (i.e. full voices) re-echo the wonders of your deeds, absolve the guilt of the polluted lip, St John.’

¹³⁰ For reasons why Guido may have hidden his original source, see Lyons, Music in the Odes of Horace, pp. 126–31.

¹³¹ Guidonis Aretini Micrologus, ed. J. Smits van Waesberghe ([Rome], 1955).

¹³² Lyons, Music in the Odes of Horace, pp. 106–11.

¹³³ T. Nisard, Musique des Odes d’Horace, reproduced in Horace (Orellius edition, Zurich, 1852), ii, pp. 924–35.

¹³⁴ Wälli, Melodien aus mittelalterlichen Horaz-Handschriften, pp. 158–9 and 283–6, followed Nisard’s opening pitch. B. Gillingham, Secular Medieval Latin Song: An Anthology, Musicological Studies, LX/1 (Ottawa, 1993), p. 14, started on middle E.

¹³⁵ I am grateful to Richard Andrewes, former head of music at Cambridge University Library, for identifying the correct starting pitch for the M425 neumes.

¹³⁶ In line 2, the first syllable has a quilisma, a three-note liquescent ascent of a third (EFG: ri-í-i), before falling back to the punctum on the next syllable (E: -dèt). In line 3, the first syllable has a torculus, an ascent from the first note followed by a descent to the same note (GAG: vi-í-ìn-); the melody then falls back by a third to the following punctum (E: -ctà). (The acute and grave markings show rises and falls, not quantities.)

¹³⁷ In line 1, ‘crines’ has a pair of rising puncta on the first syllable before falling back on the second (DE D: cri-í-nès); in later stanzas this is shown as a pes, but the effect is the same. In line 2, in the middle syllable of ‘argento’, the clivis shows a fall of a third before a punctum reverts to the starting note (D EC D: argé-èn-to). In the last word of line 2, two puncta on the first syllable of ‘castis’ maintain the same pitch before a descent of a third (FFD: ca-as-tìs). In line 3, the penultimate syllable of ‘immolato’ has a pes (AF GA A: immòlá-á-to). In line 4, the first syllable of ‘spargier’ has a clivis (GF DC: spa-àr-gìèr), while the two puncta over the first syllable of ‘agno’ are at the same pitch before a descent on the final syllable (EED: a-ag-nò).

¹³⁸ I am indebted to Iain Kerr for his help with this setting. A recording by Signum Records, issued on 1 February 2007, ref. SIGCD098, features Christopher Gabbitas (baritone) and David Miller (lute). See also YouTube, ‘Paul the Deacon & Guido d’Arezzo: Ut queant laxis’.

¹³⁹ For further details of the style and provenance, see Lyons, Music in the Odes of Horace, pp. 153–60.

¹⁴⁰ Of the four other notated compositions in the same codex, the neumes for Ode 4.2 and Epode 1 are in the same hand and ink as the first eleven Odes, but the notation for Ode 3.28 and 4.8 is by a different and later hand using a lighter ink.

¹⁴¹ For a comparison with other manuscript versions of Ode 1.1, see Lyons, Music in the Odes of Horace, pp. 160–3.

¹⁴² See above, n. 104.

¹⁴³ In line 1, ‘vario’ refers to the poet Varius.

¹⁴⁴ Nisbet and Hubbard, A Commentary on Horace’s Odes, Book 1, p. 83, refer to the metre as ‘Second Asclepiad’; Quinn, Horace – The Odes, p. 133, calls it ‘Asclepiad (c)’.

¹⁴⁵ At the top of fol. 1r and as a sideways gloss down the right-hand margin of fol. 16r.

¹⁴⁶ For details, see Lyons, Music in the Odes of Horace, pp. 165–7.

¹⁴⁷ The name traditionally given to Odes 1 to 9, where Horace showcased his metrical virtuosity with a different metre for each Ode. In PET4, the scribe also treats Odes 1.10 and 1.11 as Parade Odes.

¹⁴⁸ An ancillary reason may have been that the Ode to Tibullus was regarded as unsuitable for young boys of school age.

¹⁴⁹ Odes 1.6.1–2. Homer was said by some commentators to have been a native of Maeonia, an ancient name for Lydia.

¹⁵⁰ The rise and fall of the melody and its elaborations can be seen in Figure 4. A final ‘m’ (for the accusative singular or genitive plural) is often indicated in PET4 by a tick over the preceding vowel (usually a ‘u’) and is marked here with square brackets.

¹⁵¹ Odes 1.33.1–4.

¹⁵² Vocative of Albius, the first name of Horace’s friend, the elegiac poet Tibullus.

¹⁵³ See translation in Lyons, Horace’s Odes and the Mystery of Do-Re-Mi, p. 92.

¹⁵⁴ Mount Soracte is about 55 km north of Rome.

¹⁵⁵ ‘Thaliarchum libertum ad latiorem vitam hortat’ (‘He encourages the freedman Thaliarchus to adopt a more relaxed lifestyle’).

¹⁵⁶ Thaliarchus may be the name of a freedman, as suggested in the rubric, or a title signifying the maître d’hôtel (συμπoσίαρχoς) at a Roman symposium.

¹⁵⁷ The words ‘permitte divis cetera’ echo Sappho, Brothers Song, 13–14: τάδ’ ἄλλα | πάντα δαιμόνεσσιν ἐπιτρόπωμεν (‘let us entrust all other things to the gods!’).

¹⁵⁸ The Cambridge Songs: A Goliard’s Song Book of the Eleventh Century, ed. K. Breul (Cambridge, 1915).

¹⁵⁹ Fols. 441v(a) and 442v(b).

¹⁶⁰ These include ‘ex animali’ for ‘exanimari’ (line 3), a misplaced ‘ntar ebri’ for ‘nitor hebri’ (line 6), ‘belloro fonte’ for ‘Bellerophonte’ (line 8), ‘victus pede’ for ‘pede victus’ (line 9), ‘iaculare’ for ‘iaculari’ (line 11), ‘celere alto’ for ‘celer arto’ (line 11), and ‘frugi’ (with a full stop) followed by ‘tectum’ for ‘fruticeto’ (line 12).

¹⁶¹ L. Treitler, With Voice and Pen: Coming to Know Medieval Song and How It Was Made (Oxford, 2003), pp. 237–8.

¹⁶² E. R. Curtius, Europäische Literatur und Lateinisches Mittelalter (Bern, 1948), translated by W. R. Trask as European Literature and the Latin Middle Ages (New York, 1953).