Discuss and compare each of the following songs:

Machaut: Amours me fait desirer
Landini: Gram piant’agli ochi
Ciconia: Sus une fontayne
Anon: Ave Marie I say

In each case consider the structure of both the music and the text, and assess how far (and in what ways) the song seems to be typical of its date, country of origin and genre.


Crufts is the ultimate Dog show. Judges look at each dog presented and compare them to the Kennel Club Breed Standard. The breed standard is a detailed description of what each particular breed should be like in terms of height, colour, weight, etc. The final judging is for Best in Show and for this different breeds are compared. Instead of comparing dogs, the comparison here is between four different medieval songs, each of which is an example of a particular genre. Unlike the Best in Show of Crufts, these songs are not necessarily the ‘best’ in their genre. This essay aims to ‘judge’ each song with reference to their ‘breed standard’ and also to see how they compare with each other. The different ‘breeds’, or rather genres, are the ballade, the ballata, the virelai and the English carol.

The Ballade

Like most music of this time, the ballade basically consists of two parts: section A and section B. The basic form of the ballade, aab, relates directly to the earlier troubadour solo songs,[1] which was developed into a polyphonic work for two or three voices. By the time of Machaut (mid-14th century) ballades ‘usually have three stanzas of seven or eight lines, and the last line of each stanza is the same; it is a refrain.’[2] For each stanza there are two sections of music: A and B, with the A section being repeated, often with first and second time bars. A particular feature of the ballade is its use of musical rhyme. The end of the A section appears again at the end of the B section. The end of the B section also corresponds to the ‘refrain’ of the stanza. So the form now becomes: aa(b+C) – where C is the refrain. There was a further development of the ballade in which the B section was repeated – the refrain only appearing on the repeat. This is extended ballade form and has the structure: aab(b+C).

Machaut’s Amours me fait desirer follows the extended form of the ballade. However, Machaut chose this extended form for only five out of his 42 ballades.[3] The graphical analysis of the work (see Appendix B) clearly shows the musical rhyme that encompasses the whole of the first and second time bars. Looking to the text (see Appendix A), at first glance it appears that Machaut wrote a 12-line stanza. However, the line numbering shows that the short lines of 3 syllables are joined to their preceding lines. The total structure of each stanza of Amours me fait desirer is shown in Table 1.

Table 1

Musical rhyme is not the only feature of unity in the standard ballade. Another feature is provided through ‘subtle repetition of short melodic and rhythmic motifs.’[4] These identical (or near-identical) motifs are shown in the graphical analysis with matching colours indicating the different motifs.

Other developments of the ballade are: ‘elaborate melismas, especially on final syllables; frequent use of syncopation; setting for solo voice with accompanying parts which are clearly instrumental; extension of the second musical section to give the refrain greater prominence, preceded and closed by a cadence and reiterating material from the finish of the first musical section.’[5] How far does Amours me fait desirer compare with these developments?

‘elaborate melismas, especially on final syllables’: Long melismas only appear in the second time bars of both sections and on the penultimate syllable rather than the final.

‘frequent use of syncopation’: Almost every bar has some sort of ‘off-beat’ rhythm. Just a couple of examples are bar 3 (cantus) and bar 5 (triplum). (See Appendix C.)

‘extension of the second musical section to give the refrain greater prominence … reiterating material from the finish of the first musical section’: Machaut uses the extended form of the ballade in this piece so perhaps he found it unnecessary to extend the B section. The refrain in Amours me fait desirer is not given great prominence but it does (as already mentioned) use the music from the end of the A section.

‘setting for solo voice with accompanying parts which are clearly instrumental’: Machaut’s ballade is written for three voices: triplum, cantus and tenor, with only the cantus line having text. The triplum and cantus share the same range, but there is little crossing of parts – the cantus generally staying below the triplum. It is possible that Machaut intended this piece to be a solo accompanied song, since only one part has text. However, two recordings have interpreted this work differently. One[6] has words added to all three parts. The other[7] has the triplum and cantus with words and the tenor as an instrumental line. Certainly there are enough notes in both outer parts to facilitate the addition of words. Looking at the B section, though, the two upper voices would seem to match each other closely, with the tenor line providing more of a bass support. This lends credence to the second recording with two vocal lines and an instrumental tenor line. In the A section there is a closer tie between the triplum and tenor – perhaps indicating the future development of the triplum into a contratenor. However, this linkage is not particularly obvious and the three voices are more equal in the A section. Perhaps this is why the first recording chose to have all three voices texted.

The Ballata

The ballata was the most popular form for secular music in Italy. ‘The preserved sources of Italian Trecento … music have left us approximately 175 madrigals and 25 cacce, with some 420 examples of the third form cultivated in the period, the ballata.’[8] Landini himself wrote 140 ballate, in comparison to only 1 caccia, 1 virelai and 10 madrigals. The name ballate may sound like ballade, but it is more like an Italian version of the French virelai. Like all formes fixe the ballata has two sections of music. ‘The first for the refrain (ripresa) and the third and fourth lines of the strophe (the volta); the second for the first and second lines of the strophe (the piedi) and sung twice, often with the first and second endings (the aperto and chiuso).’[9] The form of the ballata is AbbaA, where the capital letters indicate a refrain. Although the B section often has first and second time bars, Gram piant’agli ochi does not. The line lengths and rhyme schemes change from ballata to ballata, varying from eleven to seven syllables.[10] Table 2 shows the overall structure for Gram piant’agli ochi.

Table 2

According to Hoppin, ‘the two sections of the ballate tend to be nearly equal in length’ but the sections ‘lack the contrasts of mensuration’ and that melismas ‘are generally shorter, fewer in number, and much less florid than in the madrigal.’[11] Yudkin also states that ‘French music concentrated upon short melodic motifs and rhythmic intricacy, while Italian music in its heyday emphasized long flowing lines and vocal display … French three-part songs had a single texted upper voice supported by an instrumental tenor and contratenor. Italian songs usually had two texted upper parts.’[12] So, how far does Gram piant’agli ochi follow these characteristics?

‘the two sections of the ballate tend to be nearly equal in length’: This is true of Landini’s ballata. The A section is some 16 bars long and the B section 19 bars, so there is no great difference in length between the two.

‘lack of the contrasts of mensuration’: Contrast of mensuration allowed for syncopation. Syncopation does occur in Gram piant’agli ochi but it only appears within the space of a breve. This is, perhaps, mainly due to the failings of the Italian notational system, which did not allow for complex rhythms in the same way as the French notation did. However, Italian notation was useful for writing ornamented lines, which the French notation could not do. The later development of notation combined the advantages from each of these systems and music based on the new notational system became prominent in the fifteenth century.

melismas ‘are generally shorter, fewer in number, and much less florid than in the madrigal’: There are no madrigals among the four pieces discussed here. However, melismas generally occur on the penultimate syllable which ‘is extended by a fairly long melisma; in the complete ballata.’[13] Unlike the ballade, melismas in this piece are found at the end of each line of text, not just in the last line of each section.

‘French music concentrated upon short melodic motifs and rhythmic intricacy’: Even though Gram piant’agli ochi is an Italian piece, it does have some aspects of short melodic motifs. These are shown on the graphical analysis found in Appendix B. It is more based on the Italian style, which ‘emphasized long flowing lines and vocal display’.

‘French three-part songs had a single texted upper voice supported by an instrumental tenor and contratenor. Italian songs usually had two texted upper parts.’: The text underlay in this piece is a mixture of these two styles. Landini has text for the upper voice (cantus) and the lower voice (tenor), but leaves the middle voice (contratenor) as an instrumental line. (See Appendix C.) As with Amours me fait desirer, there are various ways of performing this piece and the recording from ASV[14] chooses to have the cantus line sung, with the two other parts being instrumental. Interestingly, the tenor line is played on a plucked instrument, even though there are several places with held notes.

Finally, another similarity with Amours me fait desirer is that Gram piant’agli ochi has a musical rhyme, even though this was by no means common in ballate of Landini. Landini also has his name attached to a particular style of cadence, the so-called ‘Landini’ cadence or ‘under-3rd’ cadence. This is where the cantus moves down from the seventh degree to the sixth before reaching the octave. Yudkin states that this became ‘a hallmark of Landini’s style’[15], although he was not the only composer of the time to use this. The cadences are marked on the graphical analysis. (See Appendix B.)

The Virelai

The virelai was never as popular as the other formes fixe. It is a French genre, like the ballade, but has the same basic structure as the Italian ballata. The virelai was ‘the typical vocal solo, often quite unaccompanied’[16] and indeed Machaut wrote 33 virelai of which 25 were for solo voice. Sus une fontayne by Ciconia is a solo-accompanied song. The text only appears in the upper voice (superious) leaving the other two parts as instrumental. The recording from Secular Music of Johannes Ciconia[17] has interpreted this work in a smooth and quiet manner. It has a texted vocal upper voice accompanied by 2 trumpet- or horn-type instruments on the lower parts. It also has a second vocal part, which takes its phrases selectively from the lower two parts, first following one line then jumping to the other line, particularly where the rhythm becomes more complex. It may be that there was an occasional word in the original manuscript that gave some indication that vocalisation was required. Whatever the reason, this recording does not seem to allow the sheer complexity and boisterous nature of the virelai to shine through.

The principal features of the virelai are:

Sus une fontayne has all these features. It is in typical virelai form with two sections of music, the second of which has first and second time bars. It also has great rhythmic complexity with constantly changing tactus. At the very beginning the three voices are in three different time signatures (to use modern terminology). The tenor is in modern 6/8. The contratenor is in 2/4 where a crotchet of its 2/4 equals a dotted crotchet of the tenor’s 6/8, and the superious is in 2/4 where a quaver of its 2/4 equals a quaver of the tenor’s 6/8. This arrangement changes throughout both sections. Looking at the contratenor in bar 14 we see the tactus change on every note! The cross rhythms and syncopations set up by the ever-changing tactus are relentless. Melismas occur on practically every syllable since there is only a small amount of text for such a long piece of music.

Table 3 shows some contrast in line length in terms of syllables.

Table 3

So this Sus une fontayne ‘is typical of the Ars Subtilior, both in its constantly changing mensural relationships and in its considerable use of chromaticism.’[18] However, it is not typical of its composer. Ciconia only wrote this one virelai and did not write in the high ‘mannered’ style of the French again. In this piece, though, he uses text and music of another composer: ‘Three passages in the virelai Sus un’ fontayne exactly quote both text and music from the beginnings of ballades by Philippus de Caserta, whom Ciconia must have encountered either at Avignon or in northern Italy.’[19] Such ‘borrowing’ was not uncommon in works of this period, but it was perhaps a little uncommon to use both words and music of another composer.

The English Carol

Ave Maria I say is strikingly different to the other three songs already discussed. The most obvious difference to the previous songs is that it is for only two voices. It also has a more obviously instrumental second part, since it does not have text underlaid and there are insufficient notes for the words. Another difference is the use of two languages. Ave Maria I say is one of three pieces in the CUL Add. 5943 manuscript[20] that have their first words in Latin with the remainder in English.

The basic structure is abb (in contrast to the ballade basic form of aab), which is similar to the virelai form of AbbaA. The A section is quite long with the B section being about half the length. However, repetition of the B section brings both sections to approximately the same length. The overall structure is shown in table 4.

Table 4

The piece is made up of motifs that are repeated. The motifs are shown on the graphical analysis in Appendix B. However, unlike motivic use in the other songs, here both voices are repeated. E.g. bar 3-4 later appears at bar 8-9 where both voices are the same. Also, unlike the previous songs, syncopation and rhythmic complexity are not found here. Even the harmony is simpler, being mainly fifths or octaves.

The ‘Landini’ cadence is used at the end of both section A and section B. The B section ‘Landini’ cadence being on the second degree of the scale. It is possible that some music has been missed out at the end. However, it is also fairly common to find pieces of this period ending on their ‘supertonic’, modality being far more flexible before the ‘rules’ of tonality of the late Renaissance.

So, Ave Maria I say is very much in contrast to the other songs discussed here. It is also not typical of the English carol. This is partly because it pre-dates the time of the proper English carol. The characteristics of the English carol are:[21] 6/3 harmony (6ths and 3rds in 2 part), having a ‘burden’ or refrain, and more than one stanza with the rhyming scheme of aaab. Ave Maria I say has none of these characteristics, but does have ‘repeated notes, simple and often angular melodic lines, and harmonic simplicity [which accounts] for its uniform style.’[22] So Ave Maria I say, along with the two other songs in the same style from CUL Add. 5943, shows the beginnings of the English carol.


In comparing four pieces of music that come from different genres we have seen that there are many similarities. All of the pieces discussed here have two sections of music, but the way those sections are repeated and how they are used varies from piece to piece. The virelai and the ballata have the same structure, with the carol having a truncated form of the same structure. The ballade and the ballata both have a musical rhyme, with the carol having matching cadences at the end of each section. The carol is the odd man out in being the only piece for two parts. It is unclear, though, in the other pieces just how many of the parts were sung and which were played, if any.

To return to the analogy of Crufts, there are many ‘breeds’ which do not make it to the ‘Best in Show’ ring. Likewise, there are other genres, which have not been included in this study, genres such as the rondeau and the madrigal – the former becoming more favoured in the fifteenth century and the latter losing its appeal during the fourteenth century. An interesting study would be to follow the development of each of these genres from their beginnings to their demise. This would surely show how all these genres came from one style of a two-sectioned piece of monophonic music through to the polyphonic music we have seen here. It would also show that these different types of secular song continued to be developed throughout the fifteenth century and have influenced music down the ages.


Brown, Howard M, Music in the Renaissance
Prentice Hall, New Jersey, 1976

Hoppin, Richard H, Medieval Music
W W Norton & Company, New York, 1978

Hughs, Dom Anselm & Abraham, Gerald, The New Oxford History of Music,
Vol III – Ars Nova & the Renaissance 1300-1540
Oxford University Press, London, 1964

Knighton, Tess, & Fallows, David, Companion to Medieval & Renaissance Music
J M Dent & Sons Ltd, London, 1992.

Chapter 20 – Polyphonic Song

by David Fallows Ed. LaRue, Jan, Aspects of Medieval and Renaissance Music:
A Birthday Offering to Gustave Reese

Oxford University Press, London, 1967

The Performance of Medieval Music

by Gilbert Reaney

Ed. Sadie, Stanley, Grove Dictionary of Music and Musicians
Macmillan Pub. Ltd., London, 1980

Volume 2 and Volume 4

Seay, Albert, Music in the Medieval World (2nd edition)
Prentice-Hall Inc, New Jersey, 1975

Yudkin, Jeremy, Music in Medieval Europe
Prentice-Hall History of Music Series
Prentice-Hall Int’l, London, 1989

Further Reading

Ed. Brown, Howard Mayer & Sadie, Stanley, Performance Practice
Music before 1600 (The New Grove Handbooks in Music
The Macmillan Press, London, 1990

Caldwell, John, Medieval Music
Hutchinson & Co Ltd, London, 1978

Grout, Donald J & Palisca, Claude V, A History of Western Music (4th edition)
J M Dent & Sons Ltd, London, 1993

Ed. McKinnon, James, Antiquity and the Middle Ages –
from Ancient Greece to the 15th Century
The Macmillan Press, London, 1990

Reese, Gustave, Music in the Middle Ages
J M Dent & Sons Ltd, London, 1941

Ed. Sternfeld, F W, Music for the Middle Ages to the Renaissance
Weidenfeld & Nicholson, London, 1973

Internet site of interest: http://www.medieval.org


  1. Knighton & Fellows, page 123
  2. Yudkin, page 503
  3. Hoppin, page 425
  4. Yudkin, page 503
  5. Grove, Vol. II, page 77 (ballade)
  6. Medieval Music, sung by The Hilliard Ensemble, The Sunday Times, The Music Collection, ST4, Conifer Records, 1994
  7. The Mirror of Narcissus, Songs by Guillaume de Machaut, Hyperion CDA 66087, 1983
  8. Seay, page 157
  9. Ibid., page 158
  10. New Oxford History of Music (NOHM), page 65
  11. Hoppin, page 447
  12. Yudkin, page 539
  13. Seay, page 158
  14. A Medieval Banquet, hosted by St George’s Canzona, ASV (Quicksilva), CD QS 6131, 1994
  15. See Yudkin, page 542-3 for a more detailed analysis of Gram piant’agli ochi.
  16. LaRue, page 717 (Reaney)
  17. Secular Music of Johannes Ciconia, Homage to Johannes Ciconia, Project Ars Nova, NA048CD, 1992
  18. Grove, Vol. IV, page 392 (Ciconia)
  19. Ibid.
  20. CUL = Cambridge University Library. The Add. 5943 give the manuscript reference number.
    This manuscript is the largest collection of English songs in one manuscript. The manuscript dates from the late 14th century although there are no dates attached to the songs and only one song has the composer’s name written in.
  21. Characteristics taken from NOHM, page 122 and Brown, page 21
  22. NOHM, page 122

© 19 May 1999, Lez Bullwer