Bird on a briar (bryd one brere): constructing the missing voice

bryd one brere – bird on a briar – is the earliest surviving English secular love song with a complete lyric, dated c. 1290–1320. The music was written on the back of a papal bull with a poor pen, so interpreting the notation is problematic in parts. A previous article (available here) addressed interpretation of the music and the poetic meaning of the words.

This article addresses a second problem of interpretation: the song was clearly intended for two voices, but the primary voice is missing, leaving us only with the second voice, the polyphonic accompaniment. Using the principles of medieval English polyphony, author Ian Pittaway has constructed three possible versions of the lead voice, based on the gymel, contrary motion, and the mixolydian mode. While we cannot know if any one of these constructions was the intention of the composer, the exercise serves as an illustration of the principles of English polyphony at the turn of the 14^th century and an attempt to sing the song in the originally intended manner.

All three two-voice versions of bird on a briar are sung in a multi-tracked illustrative video by Ian Pittaway. In October 2019, all three versions were used in a concert performance by the early music ensemble Les Reverdies de Montréal, a video of which ends this article.

In the original Middle English, the imagery of bird on a briar creates wordplay between bryd/brid, meaning bird, and biryd, meaning woman. The bryd (bird) on a briar is equated metaphorically to a biryd (woman): handsome, beautiful and out of reach. We have lost the aural similarity in English between bryd and biryd due to the development of language. As we see below, we have also lost the lead voice of bryd one brere because it was never written down. There was an ecclesiastical practice of adding polyphonic accompaniment to existing monophonic melodies and, in music notation of the time, the cantus superior or lead voice was written underneath the cantus inferior or accompanying voice. In the case of bryd one brere, written on the back of a papal bull by a resident of Saint James’ Priory, near Exeter, the first line of music has the cantus inferior and a blank stave where the superior would have gone, and the second line has the cantus inferior and a space where the stave and neumes (note shapes) for the superior voice would have gone. These blanks would make sense practically if bryd one brere was a song already known to the scribe: if the point was to create a cantus inferior for a familiar monophonic song, then at the point of inspiration the already known melody could wait. Unfortunately, in this case the scribe never did go back to complete it, leaving us with accompaniment only.

Since there are contemporaneous English examples of the cantus superior being written on the top line (one is shown below), how can we be sure that the remaining voice is not the cantus superior? There are two major clues. The first is that the scribe was a monk and wrote on the back of a papal bull, so we would expect him to follow ecclesiastical practice, with the cantus inferior on the top line. The second is musical. After the first phrase there is a tonal shift that makes little musical sense in the early 14^th century modal system: it would make more sense if this was polyphonic accompaniment. Our task, then, is to work backwards and try to work out what it was accompanying.

I have constructed three versions of the lead voice in the place of the one that is missing. I have not reconstructed or restored, as the use of such words would suppose there were enough clues to discover the missing melody. There are, however, firm and established principles for medieval polyphony, and such can be used to create or construct a lead voice, using the type of methods medieval musicians used. As we will see below, the act of doing so revealed some important features of the accompanying voice, which may indicate elements of the missing primary melody.

The article, Performing medieval music. Part 2: Turning monophony into polyphony, presents 12 historically-informed methods of turning medieval monophonic music into polyphony, and explains the principles of medieval modes. For bird on a briar I have chosen two of these methods, the gymel and contrary motion, and based the third version on what is necessary for the leading voice to be in the mixolydian mode. The process of constructing the cantus superior had to take place in reverse: we have the accompaniment, so what might the lead voice have been?

The extant accompaniment is as follows at the original pitch. The scribe’s handwriting is problematic. My explanation for arriving at this musical solution is in the first article about this song. The polyphonic solutions that follow are transposed down a fifth to put them at a more singable pitch for most singers.

The following video presents my polyphonic versions of bird on a briar, with the cantus superior as a gymel, then based around contrary motion, and finally in the mixolydian mode. The rest of this article explains these solutions.

The gymel solution

A gymel, from the Latin, cantus gemellus, is twin song, two part polyphony in which the second voice tracks the first almost entirely in thirds or sixths, often moving in parallel. This definition of a gymel as thirds and sixths is from the 15^th century music theorist Guilielmus Monachus (about whom nothing is known, not even his nationality – English or Italian?), though the practice is evident in England from the beginning of the 13^th century.

In the medieval period, polyphonic accompaniment moved from perfect consonance at the beginning of a piece to unresolved dissonance and back to perfect consonance at the end. Perfectly consonant intervals, those considered stable for beginnings and endings, were unison, octave, fifth and fourth. In the 14^th century, seconds, sixths and sevenths were considered dissonant or unstable, and so could not be used to conclude a resolving cadence nor in constant parallel movement. Thirds were considered imperfect consonances, not as sweet as unisons, fifths, fourths and octaves, and not stable enough for resolution or parallel movement. Two countries were the exception to this international rule: in England and Scotland. The gymel, then, is a particularly English and Scottish form: instead of following the movement from consonance to dissonance and back to consonance, it is founded on the repeated English/Scottish consonance of thirds.     

There were two kinds of gymel: the first is based around almost constant parallel thirds; and the second uses thirds as a foundational interval, but with freer motion between the voices.

The first type is represented by the hymn to Saint Magnus, Nobilis humilis, in the Codex Upsalensis, circa 1280, written out by the monks of Saint Magnus Cathedral, Kirkwall, Orkney. In this type of gymel, the voices move almost exclusively in parallel thirds, as we see and hear below. 

The second kind of gymel is more complex than parallel thirds. An example is Edi beo þu heuene quene, a Middle English song in praise of the Virgin Mary, written in a manuscript found in Llanthony Priory in Gloucestershire, now classified as Corpus Christi College Oxford 59, dated between 1265 and the late 13^th century. Edi ends cadences with unisons but otherwise employs thirds almost exclusively, but not often in parallel. Instead, the accompanying voice is simpler than the main melody, with a smaller range based around only 3 notes, and the 2 voices move closer and further apart, from thirds to unisons, and from thirds to fifths and back to thirds, but always with thirds as the foundational interval between the voices. It is shown on the right in modern notation and can be heard played by Kathryn Wheeler playing both parts on vielle by clicking on the soundfile below. (For a full article about this song, click here.)

 

A non-parallel gymel more musically complex than Edi is the mysterious and beautiful one verse, two voice fragment, Foweles in þe frith (Birds in the wood) of c. 1270, from the Douce 139 manuscript, now in the Bodleian Library, Oxford. Foweles begins and resolves on the unison, combining the gymel with effective use of contrary motion, two voices moving in opposite directions. This means the voices cross and interweave on the basis of constantly reaching for the interval of a third between the voices, the passing notes between landing on fifths, unisons, and seconds. (An article about Foweles in þe frith is available here.)

My lead gymel voice for bird on a briar, then, is based on the more complex technique used in Foweles: I start and end on the unison, aiming constantly for the interval of a third between voices, with some parallel thirds, contrary motion, and voice crossings, producing some unisons and passing seconds while reaching for the thirds.     

The contrary motion solution

In the medieval church, organum was the name given to any form of polyphonic accompaniment, i.e. different parts moving at the same time, such as parallel fifths, octaves or twelfths, or vocal drones. The accompanying line was extemporised rather than written down, known as free organum. The writing down of organum began at the end of the 10^th century and beginning of the 11^th century with the advent of tropers. Tropers were books of tropes (from the Greek, τρόπος, tropos, a turn or a change), additions of new music to pre-existing chants in which organum was written down, but the plainsong it was written to supplement was not included. Then, in the late 11^th or early 12^th century, there were stylistic shifts: in Ad Organum Faciendum, c. 1100 (now in the Ambrosian Library, Milan), the organum voice was written above plainsong rather than below (a practice I have followed, since the bryd one brere scribe did); and there was an increasing emphasis on contrary motion of voices, resulting in voice crossings. This is seen in secular polyphonic music, too, as noted in Foweles in þe frith above.

A further example from the same period is an untitled piece dated c. 1250–1275, a few decades before bird on a briar, from folio 9 of English manuscript Harley 978, now in the British Library. (In the secular example below, we see the cantus superior on the top line and the cantus inferior on the second line, contrary to ecclasiastical practice and, purely on musical grounds, contrary to bryd one brere.)

The music in the manuscript and video above is one of three two-voice polyphonic pieces on folios 8v–9r of Harley 978. From these, several important points can be gleaned, representative of two-voice English secular polyphony of this era:

• All beginnings of phrases and cadences are consonant, either unisons, octaves, fifths, or fourths.
• Within each phrase there can be consonant intervals and dissonances or imperfect consonances.
• The interval relationship between the two moving voices is a mixture of contrary motion starting from the same note, or from a fifth apart, or from an octave apart; and groups of notes in parallel octaves and parallel fifths. This movement between voices is the key point for understanding how the parts work together.
• The rhythm of the two parts can have note for note equivalence. Where they do not have rhythmic equivalence, the cantus inferior usually has fewer notes than the cantus superior; or the two voices take turns to have the greater number of notes while the other voice moves more slowly.
• Both parts share a small number of basic rhythmic sequences.

Upon the principles of these near-contemporaneous English polyphonic pieces, I fashioned a cantus superior for bird on a briar, with the following similar characteristics:

• All beginnings of phrases and ends of resolving cadences are consonant.
• Within each phrase there are consonant and dissonant intervals.
• The interval relationship between the two moving voices is a mixture of contrary motion starting from the same note, or from a sixth or third apart (since this is English); and groups of parallel unisons and parallel thirds.
• The rhythm of the two parts sometimes has note for note equivalence; at other times the cantus inferior (taken from the original manuscript) has fewer notes than the cantus superior; or the two voices take turns to have the greater number of notes while the other voice moves more slowly.
• Both parts share a small number of basic rhythmic sequences.

The mixolydian solution

The medieval church stipulated eight modes in which their music was to be composed and sung, structured around natural notes. The dorian mode, for example, begins and ends on D, with the melody clustering around the tenor or reciting note of A (a fifth higher than D); and the mixolydian mode begins and ends on G, with the melody clustering around the tenor or reciting note of D (a fifth higher than G). (You can read more about medieval modes here.)

Each mode, when used in chant, had its own intonation figures, clusters of notes which were used to mark out the special character of the mode. When modes were used in secular music, they lacked these intonation figures. The reciting note in secular music was kept or dispensed with. This raises a question of judgement: how much of an ecclesiastical mode can be removed before we cease to think of it as modal? Does the ambit alone, D to D for dorian, G to G for mixolydian, and so on, still retain a modal character? And, though ecclesiastical modes were part of the cultural air everyone breathed, some secular tunes were clearly outside the modal system altogether.

It is a moot point, then, whether medieval modes can be used to offer a third solution to the problem of the missing voice. The accompanying voice in the original music begins on d’ and ends on c’ (without sharps or flats, as we see in the manuscript). There were modes that began on one note and ended on another – plagal modes – but no mode begins on d’ and ends on c’. Furthermore, the reciting note of the mode is the tonal centre of the melody, remaining the same throughout, and the tonal centre of bird on a briar audibly and visibly shifts after the first phrase, which we would not expect to find in a mode.

There are other factors which nevertheless make a modal solution theoretically possible. Firstly, even ecclesiastical chant slips between modes occasionally, so absolute musical rigidity is not necessary. Secondly, the extant voice is the cantus inferior, so the central question is not about the surviving voice beginning on d’ and ending on c’, but whether the lost notes which d’ and c’ originally accompanied can credibly reveal a modal origin.

They certainly can. That the accompanying voice begins on d’ and ends on c’ conforms to what we might expect of polyphonic accompaniment in the mixolydian mode. If we imagine the missing voice begins on g, the first note of the mixolydian mode, then the existing d’ above starts on a consonant fifth; and if we imagine the missing voice ends an octave above the starting note, on g’, this also makes a consonant resolving fifth with the existing c’ below.

My starting point for the missing mixolydian voice was the first note of g, the final note of g’ and the tenor (reciting note around which the melody is based) of d’. The result of emphasising the tenor while composing has two important consequences. Firstly, the shift in the tonal centre of the extant cantus inferior, apparent when sung on its own, disappears when sung against the mixolydian cantus superior, and what in isolation sounds like a tonal shift in the second voice becomes a function of singing against the mixolydian cantus superior. Secondly, the application of mixolydian and polyphonic principles, giving me the placement of the first note, final note and reciting note against the existing accompanying voice, results in a gravitation towards English thirds between the voices. If the missing cantus superior of bird on a briar was composed in the mixolydian mode, this strongly suggests it was originally a gymel.    

bird on a briar and historically informed performance

This has been an exercise in historically informed performance or HIP, a modern phrase which describes being informed about historical music theory and playing practice, using this as a springboard for our own music-making. To be historically informed requires humility and flexibility because the frontiers of knowledge, and the boundaries between certainty and conjecture, are ever-shifting. This necessitates change and growth, making any claim to full ‘authenticity’ or absolute ‘truth’ an illusion, which is why the idea of ‘historical authenticity’ in music has fallen from use. Historically informed performance is based on working with our understanding of the historical evidence and that, of course, changes, grows and develops.

I am aware, in the light of HIP, that two aspects of this article may seem surprising. They need not be.

The first is that I chose in the first article on the song and in this article to sing the words in modern English, not the original Middle English. My reasoning is simple and, I argue, historically informed: the 13^th –14^th century audience read and heard the song in a language that was theirs, that they understood without effort or translation. I would like my listeners to be in the same position as the first audience. Others may wish to sing it in Middle English, and I hope it goes without saying that this, too, has its obvious merits. (The issue of singing in the vernacular of the listening audience is further discussed in this article under the heading, Language.)

The second is that this article is about far more than simply playing the original medieval music from the manuscript: it is encouraging composition. For a great deal of medieval music, this is normal, since monophonic music cannot be played straight from the page, requiring more than the manuscript’s single line melody in performance (as this article explains in detail). It was required of medieval musicians, both ecclesiastical and secular, that they improvise to given principles. Such improvisation is, of course, in-the-moment composition, not written on the page. For the performer of medieval music, then, the scribe’s work is only the starting point for performance, the bare bones which the skilled musician must flesh out. Due to circumstance, this is especially the case with the incomplete remains of bird on a briar, since we have to use medieval polyphonic principles to devise the missing lead voice.

Of my three solutions above, I am most convinced by the mixolydian version, since it makes most sense of the surviving voice: the mixolydian mode against the existing voice creates consonant intervals at the beginning of phrases and the ends of resolving cadences; the mixolydian cantus superior resolves and removes the tonal shift in the cantus inferior; and the mode’s reciting note creates repeating thirds against the surviving voice, suggesting the song was originally a mixolydian gymel. 

While none of my solutions can be claimed as ‘authentic’, I hope I have shown them to be historically informed, based on creative principles derived from the musical evidence. If you decide to use this article as material for your own solution, I’d be delighted if you’d let me know the results.

Postscript: Les Reverdies de Montréal

After the publication of this article, I was delighted to be contacted by Bruno Cognyl-Fournier of Les Reverdies de Montréal, asking if they could use my two-voice arrangements of bryd one brere in a workshop and concert. It took place in October 2019, and their concert performance can be heard by clicking on the picture below. You’ll notice they have a slightly different interpretation of the neume rhythm to mine, and they sing and play: the surviving second voice in unison; mixolydian; gymel; contrary motion; gymel. Thank you to Bruno Cognyl-Fournier for permission to use his recording.

 

© Ian Pittaway. Not to be reproduced in any form without permission. All rights reserved.