Putting aside the notion of being historically authentic but embracing the idea of being historically informed, the aim is to arrive at a performable and historically justifiable arrangement of the problematic song, bryd one brere, from c. 1290–1320. This is the oldest surviving secular love song in the English language and so it is early music gold-dust, but it does have some severe holes: it is for two voices, but one voice is missing; and some of the roughly-written notation is difficult to decipher. What follows is not the only possible musical solution; but on this journey I’ll take you through the process step by step, so you can decide for yourself if you’re convinced. I’ll also delve a little into the background of the song, arguing that it is clearly influenced by the courtly love tradition of the troubadours and trouvères. The article starts with a video performance on voice and medieval harp.
Historically informed performance
Since the early music revival of the late 19th and early 20th century, the search has been for historical authenticity in musical performances and replicas of early instruments. Due to a great deal of research, looking at original manuscripts, music prints, contemporaneous accounts of performances and pedagogical material, close examination of players in iconography, and x-rays of historical instruments to find their inner construction, we have come a long way. There are still things we don’t know, in particular with medieval music: the further back in time we go, the more we have to rely on conjecture and guess-work due to the holes in the evidence, particularly when it comes to performance. So the by-word for musicians playing early music has been, from the second half of the 20th century, historically informed performance, or HIP for short. This is an important and practical distinction. The idea of authenticity constrains musical creativity to the academically justifiable, never going beyond the sometimes paper-thin walls of the definitively provable performance. With instrumental medieval music, this would usually mean never playing a note, as we rarely know which instruments were intended. Historically informed performance bestows a duty to refer to primary sources, to adhere where practicable [*] to historical conventions, to play copies of historical instruments, and then, informed with those tools, to be creative, and to fill in the gaps as best we can.
[*] If we adhere to historical conventions on performance regardless of practicability then, for example, a concert means playing privately to a small audience, if any audience at all; and, according to the Elizabeth Burwell lute tutor manuscript of c. 1660–1672, lutenists “will do well to play in a wainscot room where there is no furniture[;] if you can[,] let not the company exceed the number three or four for the noise of a mouse is a hindrance to that music.” I can’t see lutenists rushing to do that.
The song bryd one brere was written on a papal bull, an edict of the pope, being Pope Innocent III, dated 1199. The music and words are written roughly, in brownish ink, and upside down compared to the bull on the front. The bull was about 100 years old by the time the song was written out, c. 1290–1320 in the Priory of Saint James, near Exeter. It may seem extraordinary that a secular song should have been written on the back of a religious document from no less a person than the pope; more extraordinary still that it should have been written on a document already of some antiquity. However, this is to see both paper and papacy from a modern perspective: paper was precious, expensive, and not to be wasted; and the papacy was in dispute in England, with several clashes between King John and the pope in recent memory, as well as papal suppression of English Catholic reform movements.
It is clear that some of the music is missing. The top line for the first voice is a 5 line stave; next there is a 4 line stave drawn for a second voice, but left blank, and under that we have the words to be followed by the written first voice and the unwritten second voice; next is another 4 line stave with the rest of the first voice part; then there is a large space before the rest of the words, with ample room for another 4 or 5 line stave for the unwritten second voice, which was not added.
It is unambiguous that the scribe intended to write another voice part. Was this a song the monk knew, written out for others to learn? Or was this his own composition? If the latter, did his inspiration fail him or his duties call him away? We have no way of knowing. It may be that the anonymous monk did not write the lower voice part because it was familiar and he was working out the polyphonic accompaniment to a song that was originally monophonic. If this is the case, it would comply with the way polyphonic voices were added to plainsong in the 13th century, with the plainsong melody invariably in the lowest voice. If this is the method the scribe used then we are left only with the descant, the accompanying voice, with the original monophonic melody to be added on the second stave later, which unfortunately never happened.
Though incomplete, in writing bryd one brere this nameless brother left us the earliest extant English secular love song, and one of the oldest surviving English secular songs of any description (Mirie it is and Sumer is icumen in being the earliest).
The bryd one brere project
For reasons given above, and more that will emerge, it is impossible to sing this song in a historically authentic way, but an historically informed performance is certainly possible, and the process of arriving there is the subject of this article.
In common with the vast majority of medieval music, there is no part written for an accompanying instrument, if indeed one was ever intended, though iconography regularly shows us that instruments accompanied voices. The bray harp was certainly played in England by the early 15th century, so bryd one brere, dated c. 1290–1320, is early for the instrument. Nevertheless, I gave myself some leeway as, to my ears, the song and the instrument go well together, and recorded it in an alternative video to the one above. You can see the bray harp performance here. In the video above, it is played on a harp typical of those played in Europe during the 13th and 14th centuries. The number of times harps are depicted in iconography and mentioned in literature are testament to its status, popularity and its cultural importance by assocation with the biblical King David. The depiction of King David in the Peterborough Psalter shown above right is a typical medieval portrayal.
No harp music from this period survives, and very little early harp music has survived in general. Enough clues exist from near contemporaneous sources to know the principles of both medieval music practice and medieval harping, so from this information it is possible to plausibly work through the ambiguities left by the bryd one brere scribe, and then to create a harp accompaniment for the song.
If you wish to read about the medieval square notation in which bryd is written and which I will explore below, Nigel Horne’s online article, The Written Notation of Medieval Music, is admirably clear. I am indebted to Christopher Page for his book, Voices and Instruments of the Middle Ages (London: J. M. Dent & Sons Ltd., 1987), for giving invaluable guidance on 12th and 13th century harp technique.
Many have created reconstructions of this song before me. That my solution is different to theirs, and that others’ solutions are different to each other, illustrates the difficulties of the text.
Interpreting the words
There are several issues to consider in bringing this song to a performable rendering. In the late 13th and early 14th century, spelling was regional rather than standard or national, and the language of bryd one brere text indicates the East Midlands or Norfolk, presumably reflecting the origin of the scribe. Singing in Middle English means precise regional pronunciation is somewhat arguable and, in any case, the vast majority of listeners would be excluded from understanding. I’d like the words to be accessible, so I have tried to steer a middle course keeping, as far as possible, to the syllables and metre of the original words. My lyrics are an amalgam of others’ translations from Middle English, with some compromises of my own to fit the notes, keep the rhythm, and retain the meaning.
Three chief issues arise.
The original word-play is impossible to recreate in modern English. To translate the similar words “bryd”/“brid” and “biryd” literally as “bird” and “woman” destroys the poetry, but there is no aural similarity between the words in modern English and so no good solution. I decided, therefore, in verse 1, line 3, to translate “blythful biryd” not literally as “blissful woman” but as “blissful bird” to retain the thematic connection with the opening line of “bryd/brid one brere“.
In verse 1, line 2, “kynd” means one’s innate characteristics, making the line difficult to both translate and scan while keeping true to the meaning. The meaning of the whole line, “kynd is come of love, love to crave” is ‘our nature, our human character, is derived from love, so we crave love’. It’s unclear to me whether “come of love” implies a religious meaning – we are all of God, who is love – or a secular meaning – it is in the human character to seek love. While it is true that a medieval writer would hardly have understood the modern distinction between the religious and the secular, there are no hints of religious sentiment anywhere else in the song, so I rendered the line in the emotional and biological sense. Though the connection with being born is not explicit in the original, “come”, as in “kynd is come of love”, can mean meeting or coming together, hence I arrived at “We are born of love [i.e. our nature comes from love just as two people join together], love to crave”. Would “We are made from love, love to crave” encapsulate it better? I’m not sure that any modern English line would do it justice.
The last issue is one of syllable sounds, diction, and of fitting the words to the notes. An e on the end of a word in c. 1300 was pronounced, so in speech brere was pronounced brē-reh not breer, and crave was crā-veh not craiv. To miss the second syllable is to miss a written note, but I couldn’t find a way of rendering the words faithfully while retaining the second syllable. So in the first verse, “brere” = “briar”, both being two syllable words, is not a problem; but my pronunciation of “crave” is more “crave-ah” and “grave” more “grave-ah” to make it more like two syllables.
In the original text it does not always seem, at first sight, to be clear which words go with which notes. In the first two thirds of the tune, words are written underneath the tune in a generally precise way. In the final third of the tune, the scribe dispensed with that idea and wrote the rest of the words for the first verse and the next two verses in continuous script further down the page. However, as we shall soon see, the last part of the music is written in such a way as to give us a good indication of which words go with which notes.
Examining the lyric: courtly love, the troubadours, and the church … and grave-digging
This being the first surviving secular love song in English, we have no previous English examples with which to compare it, but there are previous songs internationally, raising some tantalising questions about the influences on bryd one brere and what may have been lost before it in English secular song. The evidence points strongly in one direction and makes the papal bull that bryd was written on highly significant.
The troubadours were the poets and singers of southern France, the trouvères their northern counterparts. From the late 11th century to the mid 14th century, they developed and promoted romantic, courtly love, expressed in terms that over the considerable time of their influence became standard clichés. A typical courtly love song would express some or all of the ideas that: the object of love is perfect, beautiful, noble, virtuous, etc.; the singer is so happy when he sees her; his life would be complete and he would be forever happy serving her; but she is unattainable. These are exactly the themes of bryd one brere. Given that bryd was written in England during the period of troubadour/trouvère activity in France that spawned similar contemporaneous traditions in Italy, Spain, Germany, Galicia and Portugal, the lyrical similarity of theme cannot be coincidental.
In troubadour lyrics of this type, marriage is unspoken being, in reality, a feudal arrangement combining class, power and business. The women in such songs are most often out of reach, so love remains an unrequited ideal. Indeed, as troubadour/trouvère songs developed over time, the woman was rarely elevated to the level of an actual person: she was a type, category or formula, in just the manner we see in bryd one brere.
For the first millenium of Christianity, the Virgin Mary was very much a supporting player. There were beliefs about the Virgin Mary’s role as the new Eve, giving birth to Jesus in order to redeem Eve’s sin, with Christ as the new Adam, redeeming Adam’s sin, but it was not until the 12th and 13th centuries that there was a sophisticated Mariology, a special elevated place for the Virgin reflected in patterns of worship and the liturgical calendar. The huge expansion in Mary’s role was a response to the unwelcome influence of courtly love. The Catholic Church seems to have decided that it was not possible to successfully counter the popularity and influence of troubadour/trouvère lyrics and their sensual portrayal of women, so they co-opted and spiritualised it. The new Mariology was developed particularly by Benedictine monks, promoting Mary to the point where she was often more prominent than Christ/God.
An outline of the general characteristics of this new Mariology shows it to have been taken directly from courtly love lyrics, attempting to diffuse its sensual power by applying it to the Virgin. Desire for a courtly woman became desire for Mary; a description of a woman’s perfection matched by her physical beauty was transformed into descriptions of the Virgin’s physical beauty; the desire of the singer to be worthy of the virtuous woman became the desire of the believer to be worthy of virtuous Mary; just as the influence of the lady is superior and ennobling, so is Mary’s influence; troubadour lyrics had the lover bound in submissive service to his lady just as a knight is bound in service to his lord, and so the church developed the idea of the believer bound in service to Mary in the manner of a knight or lover.
The remoulding of Mary in the image of courtly love was a dramatic success, seen in all subsequent Mariological literature and song, practically established in the Catholic calendar through multiple Marian feasts by the end of the middle ages. An early English example of such courtly Marian imagery in song is Edi beo thu heuene quene, written anonymously in a manuscript in Llanthony Priory in Gloucestershire, dated between 1265 and the late 13th century, contemporaneous with the troubadours. A few lines from Edi show their unmistakable influence:
There is no maid of your complexion, fair and beautiful, fresh and bright.
Sweet lady, on me have compassion and have mercy on me, your knight.
Mother, full of noble virtue, maiden so patient, lady so wise.
I am in your love now bonded, and for you is all my desire.
Typical troubadour/trouvère themes could easily have been Marianised by the monk who wrote bryd one brere, just as they were by the monk who wrote Edi: Mary could have been the object of love – perfect, beautiful, noble, virtuous, etc.; the singer would be so happy when he sees her; his life would be complete and he would be forever happy serving her; and she would be attainable through prayer.
The Catholic appropriation of courtly love, diffusing it by absorbing it into Virgin worship, was a huge success in one respect, changing Catholic theology and practice for the following millennium. In another respect, Virgin worship clearly did not eclipse courtly love in every Catholic heart: the monk who wrote bryd one brere on the back of a papal bull evidently had the amorous tradition of courtly love in mind, not the Marianised vision promoted by his church.
Not only is this song the earliest surviving English secular love song, showing clear retrospective continental influence, it includes the first extant appearance of a motif which appeared again and again in the English ballad tradition in the centuries that followed, that of the spurned or unrequited lover requesting that his/her grave be dug: “blythful biryd on me thu rewe / or greyth lef greith thu me my grave” (‘blissful woman, on me have pity / or dig, love, dig thou me my grave’).
To cite all the examples of this motif in the English song tradition would require an extensive article all of its own, such is its great regularly. It appears, for example, in versions of the most ubiquitous folk song of all, Barbara Allen, originally published three and a half centuries after bryd as a 17th century broadside ballad, first referred to by Samuel Pepys in his diary entry of 2nd January 1665, with new variants of the song still being heard by folk song collectors two and a half centuries later in the 1920s. The same or very similar lines appear in the hugely popular song, I once loved a lass, which first appeared as a broadside ballad in the 1670s as The Forlorn Lover. It was still collected from singers until the mid 20th century, with variants known as The False Bride, The Week Before Easter, I Courted a Wee Girl, Love is the Cause of my Mourning, The Lambs on the Green Hills, and so on. The original broadside of c. 1670 expressed the motif in typical fashion:
Interpreting the music: neumes
Medieval music notation evolved and different systems were used in different locations and at different times. The notation of German music theorist Franco of Cologne, described in his Ars Cantus Mensurabilis (The Art of Mensurable Music), written 1250–1280, was popular during the late 13th and 14th centuries, is the system used by the bryd one brere scribe, and is now known as square notation. In this system, staves of 3, 4, 5 or 6 lines were in use, 4 being the most popular. The pitch was understood by the clef, which could appear on any line. The advantage of a floating clef was that it enabled the scribe to keep the tune within the staff without the need for ledger lines on notes that would otherwise be above or below the staff. There were two clefs, either a C clef, as in bryd one brere, or an F clef, to indicate where C or F is located.
Interpreting the music: rhythmic modes
The music raises some questions, as a poor pen leaves us guessing at some of the writer’s musical intentions. This is of much greater significance in medieval square notation than modern music, as the shape of a neume or note indicates its value, so if the shape is difficult to discern we’re left to infer its value from the musical context. It is probably this problem of a poor pen, more than any other, which has led to different interpreters of the song arriving at different musical solutions. Seeking a way to resolve this by ascertaining what the musical context might be, I turned to medieval rhythmic modes.
First we need to understand proportion. Time signatures are written at the beginning of a piece of modern music to indication the number of beats in every bar: 4/4, 3/4, 6/8, etc. Not so in medieval notation, so we have to infer the rhythmic structure from the context. Until the mid-13th century, a long (see neumes above) was always worth two breves, as in today’s 2/4, 3/4 or 4/4 time signatures. A proportion of a long worth three breves – akin to today’s 6/4, 6/8 or 9/8, the equivalent of making a long dotted – was considered unwriteable, ultra mensuram, beyond measure, until the work of Franco of Cologne. In his Ars cantus mensurabilis, written shortly before bryd one brere, Franco was the first (in surviving records) to accept proportion tripla, triple proportion or triple time: indeed, he considered the proportion of 3 to 1 (long to breve) to be a perfect long and 2 to 1 an imperfect long. This added an extra layer of interpretation for medieval musicians playing or singing from the page. To interpret the music, one needs to see from the overall shape of the music whether there are 3 breves to 1 long, a perfect long, or 2 breves to 1 long, an imperfect long. During any given piece of music, the relationship remained perfect 3:1 or imperfect 2:1 throughout. When reading from the page, the perfect or imperfect proportion was implied rather than stated: it could only be understood from the musical context.
Being written c. 1290–1320, bryd one brere falls at the very end of the musical style known as ars antiqua, ancient art, denoting music of the high middle ages, roughly from 1170 to 1310. Most significantly, this includes the Notre Dame school of polyphony, composers working from c. 1160-70 to 1250, of whom only two names survive, Léonin and Pérotin. An anonymous member of that school wrote De mensurabili musica (The measurable music) in 1260, the first to describe the six rhythmic modes. These modes were not the equivalent of a modern time signature. All a modern time signature tells us is how many beats there are in a bar. In medieval notation, there are no bars and therefore no beats per bar: what we seek is an underlying rhythmic pulse. In church music, the mode was indicated visually by the way the notes were grouped using ligatures, indicating rhythmic accent. While the modes give an overall pulse, the rhythm could be changed for a particular phrase. Though there were six modes, in practice the first was used most often, with modes two, three and five also common. The fifth and sixth modes occured typically in only one polyphonic voice, the fifth in the lower voice and the sixth in the upper voice. The fourth mode was used rarely. They were as follows, together with rhythms in modern notation. The modern notation shows relationships between note values, not literal note values, so the first two modes could also be rendered minim crotchet and crotchet minim respectively.
Rhythmic modes can be helpful in giving shape to a secular melody. It seems clear to me that bryd one brere was written in the first rhythmic mode, and seeing it this way was extremely helpful in puzzling my way through some of its written ambiguiities.
However, this method has some controversy and is not without problems. In early music debates there has been a great deal of heat over whether these modes apply only to the polyphonic ecclesiastical music for which they were explicitly intended or also to secular music of the period. In other words, it is a question of whether the rhythmic modes were only part of the musical life of the church or a reflection of music-making generally. The weakness of their global application is that some secular melodies cannot be made to fit any of the rhythmic modes. But there is no sound reason why the modes should apply in this either/or fashion, either globally or not at all. Many secular melodies, notated before clear rhythm was written, do make musical sense when a rhythmic mode is applied. It is true that there is no clear evidence for their application outside the church, but then all the music theorists were ecclesiastical, so this is to be expected. The question, then, is not whether they were always or never used in secular music, but whether they were sometimes used.
To me, the answer is clear and positive. The notation for bryd one brere was written on the back of a papal bull, so the scribe was clearly familiar with ecclesiastical music practice (and I have already raised the possibility that it was set out in the manner of a polyphonic piece based on plainsong). Unless we are to argue that all secular music was universally without a rhythmic pulse – an argument that is musically odd and for which there is no evidence – then some melodies outside the church’s jurisdiction must have used some of the same rhythmic patterns used by the church. When using them, there is no reason why secular players would have conceived of them as the church’s rhythms, as the church did not and could not have had a monopoly on musical rhythm. It is my judgement, therefore, that whether or not the brother who wrote out bryd one brere was making an original composition or adding to a secular song, he was using what the church thought of as the first rhythmic mode, and what a non-ecclesiastical musician may have conceived in other terms. This pulse is clearly present at the beginning of the song.
Interpreting the music: ligatures
In modern notation, quavers or shorter notes are joined by a line or ligature to group them, to help the player successfully interpret the rhythm of the music. In a similar but not identical way, neumes may be joined in groups to indicate a melisma (plural melismata), a single sung syllable lasting two or more notes. This is indicated by neumes touching, close together, or notes placed vertically joined by a line on the right side to indicate rising notes. Knowing that joined or grouped notes are sung on the same syllable is a great help for twinning words with notes of music in the lines of bryd one brere that are written at the bottom of the page, rather than directly underneath the stave.
And so, bearing all this in mind, we run into some problems of interpretation. In the first line, for example, we see the notes on the word “come”, in the line “kynd is come of love”. Are they intended to be 2 unsingable Fs on the same syllable, or F E F, or F E G, or F G? The latter sounds best to me, and I interpret the possible middle note rather as a line indicating an upward movement, but that this and some other scribal intentions are not entirely clear is the reason you’ll hear a range of versions.
I have taken the first rhythmic mode as my fundamental key to working out the difficult to read note values, and thereby to interpret the overall shape of the melody. This is my foundation stone and what marks out my solution from most others I’ve heard, which tend to lack a foundational rhythmic pulse. What my solution has in common with others is that every interpretation I have heard has included some semi-breve (modern quaver) triplets in the melody: I can find no other way of retaining the metre.
So these are the principles on which I’ve worked, and here is my own solution to the melody in modern notation, set at the same pitch as the original manuscript.
The next task is accompaniment. As is the norm for music of the period, no instrumentation is indicated, but there is clear and widespread evidence that singers were accompanied by instruments, including self-accompanying solo harpers. No contemporaneous harp music survives, but we do have indications of harp style, thanks to the thorough historical research of Christopher Page.
Harps at this time were strung diatonically, i.e. without a row of sharp/flat strings as on the triple harp and without the levers on modern harps to achieve sharps and flats. If the overall pitch of a song needed to be changed to suit the singer’s voice, requiring some retuning, or to add occasional musica ficta (sharps or flats, which the church considered contrary to music in most cases, but which were added to modes by secular players), a harp would be retuned to the appropriate scordatura (non-standard tuning). Retuning a harp was itself a performance where the harpist showed the audience their skill. (For more on musica ficta and its role in the modes, see The mysteries of the medieval fiddle: lifting the veil on the vielle.) Another musical practice used then and lost now was to have an accompanied vocal performance of a song followed immediately by an instrumental version of the same, as I do in the video.
12th and 13th century sources indicate that either hand of a harpist could play the upper line while the other played the lower. There is a such a large degree of variety seen in iconography that it was clearly not the case that being right-handed or left-handed was a factor. One hand either doubled the vocal line or added harmonies to the vocal line. These harmonies could potentially have been either organum or a second polyphonic line. Organum was a second vocal (or instrumental) line, one form of which moved in a parallel octave, fourth or fifth. Clerical vocal organum often used parallel fifths until the close of the 13th century, so it seems reasonable to think this was also used in secular contexts, and that a harper may reproduce this. Polyphonic music is two or more independent lines of music that fit harmonically together. Unlike modern music, where we have four distinct vocal ranges (soprano, alto, tenor and bass, or SATB for short), the pitch of different polyphonic voices was similar until the late middle ages, often to the extent of crossing over each other in pitch during a piece of music.
For the other hand, a harpist had three options, the first being a low drone on one note or alternating between two notes. The lower strings of a harp were known as bourdons in 13th century Anglo-Norman (the English language of those conquered by the Normans), the same or similar word for a bagpipe drone, for the unstopped low plucked string of a medieval fiddle, and for the drone pipe on a 13th century organ, indicating that the lowest harp strings were designated to play a drone underneath the melody. Otherwise, the second hand played a polyphonic second line to the first, or organum.
And so, to these principles, I arrived at my harp accompaniment for the work of an anonymous English monk around the turn of the 14th century. Since I cannot sing at the pitch of the modern notation given, derived from the manuscript, I have pitched it with my first note at g’, and therefore lowered the b and b’ strings a semitone to retain the correct relationship between notes.
When I play the melody instrumentally, the left hand is playing the pulse of the first rhythmic mode, long breve, as a two note bourdon drone on g and f, which would be d and c at the original pitch. In the last part of the third line and the first part of the fourth line of the lyric there are some deviations from that rhythm on the left hand, without ever deviating from that pulse. The long breve pulse continues during the sung verses.
For the first half of the verse, in the first 2 lines of the lyric, “Bird on a briar, bird, bird on a briar, / We are born of love, love to crave”, the right and left hand together continue the rhythmic pulse underneath the melody, with the right hand playing a parallel fifth and parallel octave above the two note bourdon.
In the next line, the harp notes are a parallel third down from the melody on “Blissful bird, on me have”, followed by a parallel fifth down on the melismatic “pity”. The last line follows with the harp playing a parallel fourth down from the melody on “Or dig, love, dig”, identical to a fifth up then dropped an octave. On the final musical phrase, “thou me my grave”, I start a third down from the melody and fall on the harp parallel to the melody, but when the melody rises again, the harp continues to fall, ending with a harmony a fourth down from the melody, being a fifth up then dropped an octave, to comply with the dictum of medieval music that the consonant harmony on the last note of a cadence is a unison, an octave or a fifth.
Summary and conclusion
I have followed the contemporaneous practice of a two note bourdon; some parallel fifths, fourths and thirds; and one instance of the two voices crossing over in pitch, as we sometimes find in polyphony (for example, in Pérotin’s Dum siggilum, c. 1200; and Exiit sermo from the ms. Magnus Liber Organi, 13th century). Harmonising in thirds is unusual in medieval music, with two exceptions: their increased use in English music; and the constant use of parallel thirds (and sixths) in the musical form known as the gymel.
Of course, we can’t possibly know whether my musical solutions were the precise intentions of the medieval monk, whether the surviving melody was actually intended to be the polyphonic accompaniment, or whether my harp part bears any relation to an intended musical accompaniment, if any. It would therefore be disingenuous to call the arrangement authentic, but I hope I have at least shown it to be historically informed, based on known and documented performance practices, enabling me to get to an approximation of the music self-accompanying singer-harpists may have made at the turn of the 14th century.