How reliable is medieval music iconography? Part 2/3: 10 principles for interpreting iconography

Our chief source of information for medieval musical instruments is iconography, meaning the art of manuscripts, paintings, and sculptures. That this art must be viewed critically is a commonplace understanding. Due to its highly stylised nature, some argue that all medieval iconography is suspect and of no value for gleaning real-world information. This series of articles argues that this conclusion is a mistake: if we come to iconography with an historically-informed approach, medieval art has much to teach us about historical musical instruments.

How do we judge medieval symbolism, artistic conventions and the limitations of the medium (manuscript, stone, paint) so as to gather information valuable to a luthier, a music historian and a modern player of medieval instruments? That is what this article sets out to describe, outlining 10 principles when viewing iconography for practical musical purposes.  

The first article introduced the topic by outlining the characteristics of medieval art. The third and final part puts the 10 principles of the present article into practice with the recreation of a gittern painted by Simone Martini in 1312-18.    

The nature of medieval art

The first article outlined three foundational points for understanding medieval iconography.

The first is that art from any historical period is a stylised representation of objects, set within its own cultural context. As we saw in the first article, the simplified conventions of medieval representation …

… were inherited from classical antiquity.

The second point is that medieval Christian art has many symbolic elements. For example, God’s real and mythical creatures were seen as expressing a moral message from their Creator, so when Christians viewed an onocentaur in manuscripts and church sculpture …

… they saw the rational human combined with the bestial ass, representing the lustful hypocrite who speaks of doing good but then commits evil. When a medieval Christian saw the head of Christ or a saint surround by a Sun halo, they recognised it as a symbol of divinity or holiness, a symbol Christian artists inherited from Greek art of the gods Helios and Apollo. Likewise, colours were symbolic, as we see below: blue for heaven, gold for divinity and eternity, brown and grey for poverty and renunciation of the world, and so on.

Since medieval art is often so highly symbolic, featuring elements that cannot have been witnessed in life, such as onocentaurs, angels, demons and spirits leaving bodies, and since it is depicted in a highly stylised way, can medieval iconography have any value for the researcher of the real, such as a study of medieval musical instruments?

It most certainly can, and this is the third point. Even symbolic art is reliant on actuality, the recognisably real: an onocentaur is a combination of human and ass, angels are bird-winged people, demons are bat-winged human-animal hybrids, a spirit is a small naked version of the person. The artist has to depict elements of the recognisably real for the viewer to interpret the symbolic message. As demonstrated in the first article, iconography also gives detailed real-world information about, for example, medieval farming machinery and practices, the distinctive markings of bird species, and musical instrument organology (form, structure, development, function).

How do we work from stylised medieval art to a credible real-world reconstruction of a musical instrument? That is the subject of this second article, not only in theory, offering 10 principles to work by, but in practice in the third article, with the making of a gittern based on a fresco painted in 1312–18 by Simone Martini, with organological and musical results that could only have been reconstructed from iconography.  

The 10 principles in this article for examining iconography are:

1. open enquiry
2. take account of symbolism
3. take account of artistic conventions
4. the medium affects representation
5. the proficiency of the artist
6. test representation against practical reality
7. refer to contemporaneous written evidence
8. refer to contemporaneous material evidence
9. consult the evidence of surviving instruments
10. the cumulative weight of testimony

We will now explain the importance and application of each principle in turn.

1^st principle: open enquiry   

The principle of open enquiry has 3 components:

1. Avoid generalising global statements based on suppositions rather than evidence.
2. Question our own implicit assumptions about any detail observed in an image.
3. Always be ready to question our own conclusions when contrary evidence is presented.

1. Avoid generalising global statements based on suppositions rather than evidence.

Generalisation is a pitfall that stops enquiry and prevents the gathering of new or contrary evidence.

With the depiction of any one instrument type, there is always the possibility of regional variation, change across time, a mistake or distortion by an artist, and the chance that an individual instrument may in some way be untypical. It therefore cannot be assumed that any one example of an instrument is typical or representative of that instrument type. In any instance, it may be that the instrument is representative of its type, but the assertion cannot be assumed: it must be substantiated by multiple similar examples to prove the point, building up a picture of typical and atypical characteristics. 

To take a specific instrument, one citole credibly depicted does not mean all citoles were like it. Accumulated iconographical evidence shows variety in the body shape of the citole, which can be classified in four basic types:

hexagonal (below left); holly leaf (below right); …

… hourglass (below left); and vase shape (below right).

Based on collective iconography, what all citoles appear to share is an overall wedge shape when viewed from the top, tapering out to the widest point behind the fingerboard and tuning pegs, with a thumb-hole at the back of the neck to allow the player access to the fingerboard; a tall bridge more typical of fiddles than plucked instruments; an end projection to provide an anchor point on which to tie the strings (pointed, trefoil, round, or end pin); typically either 4 single courses or 3 double courses of strings, always gut; typically played with a plectrum modelled on the writing stylus (or an actual writing stylus) or a bone. These conclusions can only be reached by avoiding generalising global statements from a few instruments and sifting through a large amount of evidence. (For more on citoles, click here. For more on medieval plectrums, click here.)

This dedication to evidence is the foil to global statements based on unsupported suppositions. For example, I have read several times in online forums that a gittern (or some other instrument) is depicted in an ‘unrealistic’ way because it was drawn by a monk who had never seen one. For the statement to be verified, we need to know the identity of the monk, or nun, or possibly artisan from the 12^th century on, when manuscripts began to be outsourced and commercially produced, and enough biographical detail to show that he or she had never seen a gittern, or show that universally no monk or nun of any order or any commercial artist had ever seen a gittern. This is plainly an impossible undertaking, and absurd when we recognise the popularity of the gittern and the unity of gittern depictions across the three centuries of its life. 

Another common global statement which closes down enquiry is to suggest that all manuscript iconography is unreliable because the image was copied from another stock image. For this statement to be verified, we need to either prove that all manuscript iconography was copied, which is demonstrably false (as we will see), or that the particular image in question was copied from another, for which we need the supposedly original image and enough duplication to demonstrate a dependent relationship, without which the claim of copying, unsubstantiated, can be dismissed.

If mass copying of stock images was commonplace in medieval manuscripts, it would now be observed through repeated appearances of identical figures. Not only do we do not see this, the claim goes against all the available evidence. Anyone who has studied a significant number of the very large international collection of surviving medieval decorated manuscripts should recognise what strikes even the casual viewer: not artistic sameness, not copying, not observing the same image again and again, but variety, originality, and a particular artist’s individual interpretation on the themes of medieval life and art.

For example, the mouth of hell is a familiar scene in medieval and renaissance art. In the De Lisle Psalter (British Library Arundel MS 83, folio 128r), c. 1308-40, it is used to illustrate an element of Jesus’ parable in The Gospel of Luke 16: 19-31. We saw this scene in the first article and it is repeated in the summary first section above. The mouth of hell portion of the scene is shown again below. On the right we see the uncaring rich Dives, who ignored poor Lazarus at his gate, on his death bed, his soul being taken by two demons. On the left, a third demon pushes Dives’ soul into the mouth of hell for his eternal punishment. 

From a 14^th century manuscript in French and Latin of Apocalypse or Revelation, the last book of The Bible, we see the image below (British Library, Add MS 17333, folio 43r). This artistic variation on the mouth of hell theme illustrates Revelation 20: 10, “And the devil that deceived them was cast into the lake of fire and brimstone, where the beast and the false prophet are, and shall be tormented day and night for ever and ever.”

Based on the passage, we see the artist depicted three mouths of hell, one each for the devil, the beast and the false prophet, represented by animals that are individually leopard-like, multi-headed and porcine; hell-fire from two of the mouths of hell and from the cloud above; a demon figure prodding the damned into hell, typical of mouth of hell images; and the faces of two frightened souls looking on. We see that this is recognisably a mouth of hell scene, but it is not what we see in the De Lisle Psalter nor in any other source: the artist used the common trope and created something original.

Another variation on the mouth of hell theme is below from 76 F 5, an illustrated Bible in Latin from the Benedictine Abbey of Saint Bertin in Saint Omer, France, dated c. 1190–1200 (KB nationale bibliotheek, The Hague, The Netherlands, folio 45r). In this nearly full-page miniature, the artist has brought together many of the familiar medieval ways of illustrating the Last Judgement. From bottom to top, we see:

• the mouth of hell full of sinners, with demons carrying more of the damned and pushing them into hell;
• angels blowing olifants (elephant horns) to herald the coming of the Last Judgement (further explained in this article under the heading, The reredos);
• the dead physically resurrected from their graves to face judgement, symbolically naked before God (further explained in this article under the headings Artistic symbols and signifiers of sin and The punishments of hell);
• angels swinging censers to spread the smell of incense, indicating worship of the Almighty;
• a sheet, probably a winding-sheet, used as a transportation device to take the souls of the saved to heaven, typically carried by an angel, a group of angels, or held by Abraham to show that the blessed are ‘in the bosom of Abraham’, i.e. in heaven (ideas further explored and illustrated here under the heading, The Satanic he-goat).

The artist has brought these theological and artistic themes together, using familiar medieval visual symbolism, to form a composite picture that is nonetheless unique, that exists in this form only in this one place. This is not a copy of an existing piece of art, in part or as a whole, nor is there any evidence that it has been copied elsewhere. In other words, this medieval artist did what artists of all eras do: they create something new from the components of the familiar cultural landscape. And not only something new, but in this case and many like it, created expressly to illustrate the text of the specific manuscript.

There is no greater example to substantiate the bespoke nature of medieval manuscript art, created for the individual work, not blindly copied from stock images (of which there is no evidence), than the marginal note on folio 96v of Bibliothèque nationale de France, Latin 8504. The manuscript is a collection of stories for edification and entertainment, featuring Kalīlah and Dimnah, two jackals, with a cast of other anthropomorphic animals. (The stories have a remarkable history: originating in the Sanskrit text, Pañcatantra, 200 BCE; translated by Borzuya into Middle Persian in the 6^th century; translated by the Persian Ibn al-Muqaffa’ into Arabic in the 8^th century; translated by the French Raymond de Béziers into Latin in 1313 in the version shown below.) Above the illustration on folio 96v is the instruction for the image, and on the right a note in the margin stating that the picture is wrong: it does not conform to the particular and individual instruction given the artist.

There are examples of medieval works where the whole text of a manuscript is copied in part or in whole from another. But even in such cases, where the art is dependent on a previous copy of the manuscript, we still see artistic originality and innovation.

A test case for one manuscript reliant on another is the commentary by Spanish monk, Beatus of Liébana (c. 730–c. 800), on the last book of The Bible, The Apocalypse, also known as Revelation. Below left we see the whole of a page from his commentary dated c. 945, folio 87r of Morgan MS 644, with details of the four musicians on the right.

Below left is the whole of the equivalent page in a version of the same manuscript dated c. 975–1000, 30 to 55 years later, again with the musicians on the right: folio 86v of the Silos Apocalypse, British Library Add MS 11695. We have what looks like the same instruments (unidentified) within a slightly different page design. It is clear that, in this case, the monk or nun used the previous manuscript or a version of it, with some original artistic variations. Note that we see what is recognisably the same instrument, but it is not depicted in an identical way.

A third rendering of the same work is dated c. 1190–c. 1210, 245 to 265 years after the first example, Morgan MS 644: this is Rylands Beatus, John Rylands Library, University of Manchester, Latin MS 8, folio 89r. Below left we see the same page design as Morgan MS 644; and on the right we see that the musicians are playing a different instrument to that in Morgan MS 644 and the Silos Apocalypse. In this case, the monk or nun did not make a straight copy of the previous manuscript, but changed the instruments being played.

So we see in this example that a whole manuscript is based on a prior work, but later manuscripts are not replicas: they show artistic originality, substituting one musical instrument for another, which requires knowledge of that new instrument.

Even if we could find a case of like-for-like exact image replication that could be proven, this still does not show in itself that a depiction of a musical instrument is unreliable: that can only be demonstrated by comparison with the real-world object or obvious illustrative errors that are neither symbolic (2^nd principle) nor an artistic convention (3^rd principle). Since we do not have the object that was illustrated to compare with the artist’s work, an assumed error can only be proven by demonstrating that what is shown is practically impossible (6^th principle). If this cannot be shown, then the supposed unreliability of the image is based only on the viewer’s implicit assumptions.

We have seen in the first article, with examples from farming and ornithology, that medieval depictions of real-world objects (rather than theological devices such as the mouth of hell, angels and the blessed in heaven), though filtered through medieval artistic conventions, often have a high degree of factual reliability if we know how to interpret the medium. Interpretation requires knowledge of the other 9 principles, described below. 

2. Question our own implicit assumptions about any detail observed in an image.

As an example of implicit assumptions, I can report that my own unrecognised lack of knowledge has caught me out. In an article on evidence about medieval instrumentation (here), I originally showed a depiction of a tabulae player with a shawm player from the Cantigas de Santa Maria (below left), commenting that this image is unrealistic as the player has both tabulae clasped together in the hand, making them impossible to play, rather than between the fingers as shown in an engraving by Cornelis Massijs in 1538 (below right), as I had always seen them played. I am delighted that María Giménez Fernández left a comment to put me right, with a link to this video, demonstrating both the veracity of the Cantigas image and the wrong assumption I had made.    

3. Always be ready to question our own conclusions when contrary evidence is presented …

… as demonstrated above with the Cantigas tabulae. 

In summary, to glean information about musical instruments from medieval iconography with open enquiry rather than closed assumptions, we need to avoid unverifiable global statements and individual unquestioned assumptions, with a working knowledge of medieval symbolism, artistic conventions, the medium of manufacture, the skills or limitations of the artist, practical reality, written evidence, material evidence, surviving instruments, and the cumulative weight of testimony, as indicated by the remaining 9 principles.

2^nd principle: take account of symbolism

In the manuscripts of Beatus of Liébana’s commentary on The Apocalypse shown above, we would be misled if we were to take the manner of playing these instruments literally – so we turn to the 2^nd principle.

We have seen in the first article that medieval iconography includes symbolic mythical creatures and symbolic colours. In Beatus of Liébana’s Apocalypse manuscripts and in church stone carvings telling the same story, we see a symbolic way of showing musical instruments. The Bible passage represented is Revelation 5: 6-8: “Then I saw a Lamb that looked as if it had been slaughtered, but it was now standing between the throne and the four living beings and among the twenty four elders … the twenty four elders fell down before the Lamb. Each one had a cithara and they held gold bowls/vials filled with incense, which are the prayers of God’s people.”

The cithara (Latin) or kithara (Greek) was, in the ancient world, both a lyre and a general term for any stringed instrument. Medieval translators and illustrators of The Bible either interpreted cithara as lyre or translated the word freely to mean harp, psaltery, citole, or any instrument with strings, plucked or bowed. Santiago de Compostela Cathedral, Spain, 1168–1211, Chartres Cathedral, France, 1145–1245, and the Collegiate Church of Toro, Zamora, Spain, 1170–1250, all have impressive carved scenes of the Elders of the Apocalypse. As we see on the right, in Santiago de Compostela the cithara is depicted as an organistrum, fiddles, a rota, and a harp.

In every source that has a carved or drawn illustration of the Elders of the Apocalypse, plucked and bowed representations of citharas are shown in a mixture of poses. A few are played, many are held upright in non-playing positions, and sometimes instruments are held in an upright position as if played with fingers, in imitation of the lyre. This is the case with the apparently plucked instruments shown above in Beatus of Liébana’s Apocalypse manuscripts. On folio 86v of the Silos Apocalypse we see four instruments held upright and plucked to represent citharas (below right); on folio 86r we see the same instrument in a non-Elders scene, played in its everyday way, and we see that in reality it was bowed (below left). 

Similarly, below left is an Elder in the Collegiate Church of Toro, Zamora, 1170–1250, with a citole held upright and plucked with fingers. Since this symbolises the cithara, the upright lyre, it cannot be taken as representative of musical practice any more than the similarly-held gittern in Pamplona Cathedral, Spain, 13^th–14^th century, below right.

This symbolic representation therefore does not depict contemporaneous musical practice, any more than the depictions of hell by Jheronimus Bosch in c. 1495–1505 (below) give us real-world practical information about the practice of tying naked sinners to the necks of giant lutes, crucifying them on the strings of bray harps, painting Strichnotation on their backside, placing blind beggars on top of hurdy gurdies, nuns inside hurdy gurdies, placing sinners inside a drum played by a demon, or making a sinner carry a giant shawm after inserting a recorder into his rectum.

While imagery of the Elders of the Apocalypse may alter the depiction of musical instrument playing technique, the accuracy of the depiction of the instrument itself is not affected by religious symbolism. For example, on the right is a citole player in the minstrels’ window of Lincoln Cathedral, 1385. Any such angel musician represents the praise of God, Christ and the Virgin that is the foundation of Christian theology and art. We see the bird’s wings to show that the player is one of heaven’s angels; the symbolic colours – blue for heaven, gold for divinity and eternity, red for life and Christ’s sacrifice, green for verdant life, white for purity and holiness – and the plant stems and leaves remind the viewer that earthly life is short, as expressed in Isaiah 40: 6-8: “All flesh is like grass, and all its glory like the flowers of the field. The grass withers and the flowers fall when the breath of the Lord blows on them; indeed, the people are grass. The grass withers and the flowers fall, but the word of our God stands forever.”

The citole itself is in line with what we would expect from other citole iconography and from the surviving citole of 1280–1330 now in the British Museum. It has an overall vase shape; a trefoil on the tail; a string-holder between tail and bridge; 6 strings, which would be arranged in 3 double courses; an overall wedge shape front to back, shown in typically medieval flattened perspective; a thumb-hole on the neck to allow access to the fretted fingerboard; a decorative carved head on the peg-box; and it is played with a plectrum. (For more analysis of citole iconography and a detailed examination of the British Museum citole, click here.)

There is one unexpected detail. Citoles are generally shown with 6 strings arranged in 3 double courses or 4 single strings, with the appropriate number of tuning pegs to match. The minstrels’ window citole has 6 strings, as we would expect, but 13 tuning pegs. It is therefore tempting to think of 13 tuning pegs as symbolic, especially as tuning itself was a visual symbol for creating spiritual harmony (as this article explains). However, a symbolic number of tuning pegs is not an established idea in medieval art, so the more likely explanation is that this artist was simply confused or got carried away when depicting the pegs. (The issue of artists’ mistakes is addressed in the 5^th principle).   

Another example of religious imagery not affecting the reality of the instrument is above: an image from folio 105r of the Anglo-Catalan Great Canterbury Psalter (Bibliothèque nationale de France, Latin 8846), decorated in the first half of the 14^th century. We see Christ enthroned, flanked by the 24 elders of Revelation 4: 4, “Surrounding the throne were twenty four other thrones, and seated on them were twenty four elders. They were dressed in white and had crowns of gold on their heads.” The 24 elders are shown either side of God with his usual symbols of divinity. The elders are crowned but not dressed in white here. Their citharas are shown as fiddles, held in realistic ready-to-play or resting poses. Below is folio 114r from the same manuscript, with the blessed – shown by their halos – playing (left to right) psaltery, portative organ, fiddle, psaltery, lute, cymbala (cup cymbals), and pipe and tabor, all depicted as in life.

We see the same in the renaissance: other than in Elders of the Apocalypse scenes, religious symbolism does not lead to portraying an instrument differently to its real-world appearance or playing method. Two representative examples follow. 

Mary, Queen of Heaven (above left), painted c. 1485–1500 by the Master of the Legend of Saint Lucy (Bruges, Flanders), depicts 15 angel singers and 14 angel instrumentalists, playing shawm, bray harp (detail above right) … 

… another bray harp, dulcimer, lute, three recorders, shawm, vielle (medieval fiddle) … 

… trumpet, portative organ … 

… shawm and lute.

Not only are all the instruments represented realistically, the four singers directly behind the Virgin sing from real and readable music, held in the hands of the two foremost singing angels. The music is a variation of the motet, Ave regina coelorum / Mater regis angelorum (Hail, queen of the heavens / Mother of the king of angels), by English composer Walter Frye (fl. 1450–75).

Our second renaissance example is Virgin and Child with four angels, c. 1510–15, by Gerard David of The Netherlands (above). It shows the same symbolic proportion as in the previous painting, with Mary the most significant and therefore the largest figure in the composition. In Gerard David’s work, the holy mother and child are worshipped by two musicians playing bray harp and lute (details above centre and right), again portrayed in realistic detail without symbolism affecting the depiction of playing technique or the accuracy of the instrument itself.

3^rd principle: take account of artistic conventions

There are aspects of medieval art that reflect established artistic conventions rather than physical reality. They are:

• flattened perspective
• the ‘arm under’ playing position
• exaggerating the size of the most important elements
• missing features.

The first of these artistic conventions, flattened perspective, is inherited from ancient Greek and Roman art (as outlined in the first article). Thus it was typically (but not universally) the case that when a gittern was presented in two dimensions, flattened perspective made its sickle-shape peg-box with a decorative carving look as though the peg-box, and sometimes part of the neck, was turned back to face the player, combining the viewpoint of both observer and player, as we see below.

Some citoles likewise had a sickle-shape peg-box with a decorative carving, and we see below the same artistic technique of flattened perspective to show it.

Probably all and certainly most citoles had a wedge neck with a thumb-hole to give access to the fingerboard, as demonstrated above. The examples below illustrate how this type of citole peg-box and wedge neck appear in flattened perspective.  

The point of flattened perspective was to bend details into view, presenting information not available from that vantage point in reality. When a gittern or citole was carved in stone in three dimensions, there was no need for this distortion, as we see below left. And, as we see from the gittern below right, not all medieval two-dimensional art used the technique (specifically, flattened perspective was not used in the International Gothic style of painting of c. 1375–c. 1425 and beyond). 

Another way visual information was presented in medieval art was to favour the viewer by moving the plucking arm down, to make more of the instrument visible in art than in reality. Examples of this impractical and impossible ‘arm under’ playing position follow for citole players in manuscript art and in stone carvings.

Below are two citoles in La Portada del Sarmental, Burgos Cathedral, Spain, c. 1235. When compared, they illustrate how the ‘arm under’ position was used by artists to show more of the instrument. On the left is a citole player in a realistic pose, the hand parallel to the strings, which means the details of the bridge, string-holder and tail are all hidden from view. The citole player on the right has the whole right arm moved down to an impossible playing position, with the advantage that the viewer can see the bridge, string-holder and tail.

The same principle of moving the arm down to show more of the instrument applied to two- and three-dimensional depictions of plucked instruments generally, as we see, for example, with the gittern below.

Like flattened perspective, the ‘arm under’ position was not a creation of medieval artists, but a long-established convention inherited from previous ages and cultures, as we see below left on a figure from Susa, Iran, 14^th–12^th century BCE; and below right, on a musician depicted on a frieze on the entrance to a Buddhist temple excavated from Airtam, Uzbekistan, 1^st century CE.

A third medieval artistic convention was to increase the size of the most significant features and decrease the less significant features, so that scale and proportion was a matter of giving prominence, representing importance rather than physical reality. We saw this above in the 2^nd principle: take account of symbolism, in relation to the larger size of the Virgin Mary, Jesus and saints compared to other figures, and in the first article in relation to the comparative size of pigs, acorns and oak trees in animal husbandry scenes, in which the most important aspects of the depiction were exaggerated in size. The same principle applies to the features of a musical instrument. Below are two representative examples of exaggerated features on citoles: on the left, the trefoil (end projection) is oversize, but the bridge and tuning pegs are not shown at all; on the right we see an exaggeratedly large sickle-shape peg-box and similarly oversize wedge neck in flattened perspective.

As we have just seen, sometimes details of an instrument are missing, and this brings us to the fourth and final artistic convention: missing features. In many instances, this was clearly an artist emphasising the salient features. In such illustrations, the details that were included or omitted were chosen according to their perceived importance for the individual artist. But there was one typically missing feature of instrument depiction about which there was clearly a general convention: straps.

Due to their relatively small size, medieval fingerboard instruments – citoles, gitterns, lutes – can easily be played sitting or standing without a strap, so the straplessness of the portrayals below and all those like them can be considered true to life. 

Medieval and early renaissance art that shows standing musicians with instruments which require two hands to play – harps, psalteries, portative organs and simfonies (symphonies, symphonia, organistrums) – do not usually show the practically necessary straps. When medieval art shows such a player standing, the instrument typically floats with no means of support, as we see with the four representative medieval harpers below. As with previous conventions, this was not new to medieval art, as we see from the ancient Egyptian harper, below left.

As well as floating harps, medieval art shows standing two-handed players of floating psalteries, portative organs, and simfonies, without the functionally-necessary straps.

It is not until the 15^th century on that straps begin to appear regularly in art, as we see in the following examples showing straps for harps and psalteries.

Even when straps are shown from the 15^th century on in renaissance and then baroque art, they appear inconsistently – sometimes shown fully, sometimes shown only partially, sometimes not at all. For a more detailed analysis of instrument straps in art and historical reality, see The evidence for straps used with medieval, renaissance and baroque musical instruments.

In summary, artistic conventions of the medieval period alter the way instruments are presented: by flattening perspective to bend details into view; by moving the plucking arm of a chordophone player down to make more of the instrument visible; by increasing or decreasing the size of instrument details, according to their importance; by omitting some less important details altogether or, in the case of straps, removing them entirely from view to create a cleaner aesthetic, a practice that continued in renaissance and baroque art. 

4^th principle: the medium affects representation

Any medium used for art will necessarily have an effect on the image: a two-dimensional object tries by various means to represent three dimensions; oil paint does not spread in the same way as ink and creates a radically different visual impression; wood does not carve in the same way as stone, and the type of wood or stone matters, too. This has clear implications when gleaning detailed information from art about musical instruments.

A detail carved too thinly in stone is vulnerable to breakage or erosion. For this reason, smaller details may be enlarged, as we see from the following representative examples.  

Above is a stone carving from Gandhara in modern-day Pakistan, c. 320 CE, in which a female cymbal player accompanies a male oud player, whose long plectrum, made of a gut string, is the same shape and size as his other strings. In reality, none of the strings would be so thick, and each string would be progressively thinner as they rise in pitch.

Due to the quality of stone, the necessity for thickening more delicate features is commonplace. Above and below is a citole carved 1330–90 on the inner north wall of Beverley Minster, a church in the East Riding of Yorkshire with more representations of musicians than any other medieval site. As with the Gandhara oud, the strings of the Beverley citole (and harps and rotas to follow below) are represented all of the same diameter, not graded for pitch as real strings are, and substantially thicker than reality to avoid the material crumbling, a necessity when carving in stone.

Stone carvings of fingerboard instruments with a low number of courses – gitterns, citoles, lutes, fiddles, all of which had 3 or 4 courses in this period – show an accurate number of strings, verified by contemporaneous written accounts of the instruments. An instrument with a larger number of strings, such as a harp or a rota, was not typically shown with a realistic number of strings due to the nature of the medium. For example, below are 4 stone harps in Beverley Minster, each with only 9 or 11 strings.

Written testimony shows that the number of harp strings varied. Contemporaneous with the Beverley Minster stone minstrels, French poet-composer Guillaume de Machaut, c. 1300–77, likened his love’s 25 virtues to the 25 strings of the harp in his Dit de la harpe. An early 15^th century sermon by Jean de Gerson, Puer natus est nobis (A child is born to us), states that the harp has 20 strings. In 1523, Giovanni Maria Lanfranco stated that the harp has 15 strings in his Scintille di musica (Sparks of music), fewer strings in the 16^th century than the 25 and 20 strings attested by Machaut and Gerson in the 14^th and 15^th century. The Berkeley theory manuscript, written in Paris before 1361, gives 14^th century tunings for many instruments, including a harp with 11 strings (pictured right). Though this is the same number of strings as on two of the Beverley harps, this cannot be counted as confirmation. The reduction of string numbers for stone harps was routine, and the 11 strings shown in Berkeley was almost certainly an abbreviation for teaching purposes just as, in other illustrations, details of instruments were assumed to be present rather than drawn, such as missing frets on the Berkeley manuscript drawing of the citole and gittern.

The same general theme of string reduction in stone is present with rotas. The rota (rote, rotta, rothe) was like a harp, but with two bands of strings and a soundboard between them. Below is a corbel (wall projection used to support a cornice or arch) in the shape of a rota played by an ape. This appears on the outside of the 12^th century church, Notre-Dame de l’Assomption, Surgères, France. 

On the right is another corbel showing a rota being played, this time by a man, on the outside of the Church of San Martín de Elines, Valderredible, Cantabria, Spain, early 12^th century.

Below is a third rota carved in stone in the abbey church of Sainte-Foy at Conques-en-Rouergue, France, c. 1130. This scene of two devils torturing a condemned soul in hell forms part of a tympanum (semi-circular or triangular decorative wall surface over an entrance, door or window) with 124 figures portraying the Last Judgement, including the heavenly court, angels, the saved, and the damned in hell. This minstrel’s soul is eternally condemned for being a sinful musician: his rota taken from him so he cannot play, a noose around his neck so he cannot breathe; his tongue torn out so he cannot sing. (For more on the relationship between minstrels and the church, click here.)

The Notre-Dame de l’Assomption rota has a row of 5 very substantial strings on each side; the San Martín de Elines rota has 2 rows of 8 strings; and the rota held by the devil at Sainte-Foy has 2 rows of 7 strings.

Gascon troubadour, Giraut (Guiraut) de Calanso (Calanson), stated in his work of 1210, Conseils aux Jongler (Advice to Jongleurs – jongleurs, known in England as minstrels, were professional entertainers), that a jongleur must know how to play a 17 string rote, presumably meaning 2 rows of 17 strings. Petro (Petrus) de Abano, in his Expositio problematum Aristotelis (The explanation of Aristotle’s problems), 1310, described the rota as having 2 rows of 22 strings either side of a sound-box. Since a rota had 2 bands of 17 or 22 strings (or perhaps some similar number in other cases), the stone rotas above were clearly presented this way as an approximation of the instrument, giving the viewer the idea of a rota. It is also notable that the Notre-Dame de l’Assomption and San Martín de Elines rotas are very broad from side to side, the width grossly exaggerated to be more easily visible to viewers on ground level.

There is one stone representation of a rota notable for its unusual detail, the exception that proves the rule. On a capital (the load-bearing top of a column) in the cloister of the Abbey Church of Saint Pierre, Beaulieu-sur-Dordogne, France, c. 1100, is a rota showing 22 strings on the visible side – the number stated by Petro de Abano in 1310 – each meticulously and separately carved.

The same phenomena persists in two-dimensional manuscript and painted art as in three-dimensional art: fingerboard instruments with a low number of courses – gitterns, citoles, lutes and fiddles – are consistently shown with an accurate number of strings, whereas those with a larger number – harps and rotas – are typically shown with a random number of strings that is fewer than reality.

Below, for example, are 3 gitterns, which are dependably shown in carvings, paintings and manuscripts with 3 or 4 courses until the 15^th century, and sometimes with 5 courses in the second half of the 15^th century.  

By contrast, below are two harps, here portrayed typically for manuscripts and paintings in that they are shown with fewer strings than testified in written accounts: the harp in the De Claris mulieribus manuscript, 1403, has only 9 strings; and the bray harp in the Mary and child painting, c. 1440, has 12.

So it is with rotas in manuscripts: the Polirone Psalter, before 1086, has a rota with 2 bands of 10 strings (unevenly spaced); and the Worms Bible rota, 1150–75, has 2 bands of 11 strings, both considerably fewer strings than described in written accounts.

It is easy to understand why the medium of carved stone would alter representation of strings and typically reduce larger string numbers. There are two key factors which result in string number reduction in manuscript art.

Firstly, instruments in manuscripts are typically small, either a feature in the margin or one element within a busy scene. The scale of the image will result in some details missing or partially missing or, put another way, suggestive of a more detailed reality. As an example, on the right is folio 41r of the Apocalypse de Saint Jean, a 14^th century manuscript in French of the last book of The Bible (Bibliothèque nationale de France, Français 13096). The page measures 22 cm × 15.5 cm and features 10 musical instruments: 3 fiddles, 3 harps, 1 portative organ, 1 psaltery, 1 citole and 1 bagpipe. By my calculations, working left to right down the page, the instruments are drawn to the measurements indicated below. The longest part of the largest-drawn instrument, the bagpipe, is 2.5 cm; the longest part of the smallest-drawn instrument, the psaltery, is 0.85 cm. It is easy to see, then, why the string numbers of the harps and psaltery, and the number of pipes on the portative organ, would be abbreviated at this scale.

The same point could be made over and over, using instruments depicted in other manuscripts. For example, the pages of The Luttrell Psalter measure 35 cm x 24.5 cm, which makes the tallest point of the harp on folio 13r only 4.4 cm. The tallest point of the harp on folio 174v is only 2.3 cm. This small scale explains why the artist depicted, as was typical, a nominal number of harp strings, in this case 8 and 9 respectively.

The first key factor, then, which results in string number reduction in manuscript art, is the small scale of the art. The second factor is a consequence of the first, that many medieval artists were concerned with creating the idea of an instrument, rather than the equivalent of a luthier’s plan, to scale and correct in every detail. Drawing an accurate 3 or 4 courses for a gittern, citole, lute or fiddle is not difficult, and any reader could count the courses for accuracy at a glance; but it would take a dedicated reader to count 17–22 strings for one side of a rota or up to 25 strings for a harp, even if they could be depicted. One could therefore argue that the artist needed only to show that the instrument has some number of strings, unless the artist was particularly committed to the accuracy of a many-stringed rota or harp and also had space on the page to be so precise and also had the artistic skill for such close work. A few did have the skill, as we see below. 

The pages of The Alphonso Psalter, c. 1284–1316 (British Library Add MS 24686), are 24.5 cm x 16.5 cm. The harp on folio 11r (above) is only 3 cm at its tallest point, yet the artist depicts it in such fine detail that a realistic 18 harp strings are included.

Likewise, a rota in the Gradual of Saint-Etienne of Toulouse, France, 1075–1125 (below left), is shown with a band of 19 strings, a credible number partway between that given by Giraut de Calanso – 17 – and Petro de Abano – 22; and a rota in a Cantigas de Santa María manuscript, Iberia, 1257–83 (below right), is drawn with a band of 17 strings, exactly the number given by Giraut de Calanso.

In summary, the medium fundamentally affects representation. Taking string numbers as an example, the need to maintain thickness to support the structural integrity of stone and the small scale of art in manuscripts means that numbers of strings can be shown realistically as long as that number is small, while larger numbers of strings are typically reduced, with notable exceptions that show outstanding artistic skill and dedication.

5^th principle: the proficiency of the artist

As I will demonstrate in the third article, medieval iconography is a credible source of musical instrument information if we understand it in its historical context, interpret it in its own terms, and do not expect information it was never intended to convey.

As we have seen, medieval art is not a perfect representation of the real, so we need to know how to read it to glean information. So far we have considered symbolism, artistic conventions and the nature of the medium for the purpose of discerning the musical reality behind medieval art. We must also consider the proficiency of the individual artist, as a few artists, not sufficiently familiar with the instruments they portrayed, made obvious errors.

For example, above left is a harp held backwards and with both hands on the same side of the strings; on the right is a harp held the right way, but still with both hands on the same side of the strings; and below is a lute with 4 strings but 6 tuning pegs, with only 3 frets on the neck, and with a bridge much too close to the tail.

But we must be wary of wholesale dismissal: just because a detail is wrong does not mean the whole image gives no useful or accurate information.

The harp in Harley 745, though held wrongly, is very much in line with other contemporaneous depictions of harps. The harp in Codex Manesse is shown with red strings, an important detail attested in other harp iconography and on other instruments over a broad span of time, clearly a real phenomenon, explored in the third article. The lute in Le livre des échecs amoureux moralises, though shown erroneously with only three frets, has the realistic detail of triple frets, attesting to the fact that the artist portrayed a bray lute on which frets were made deliberately to buzz like a bray harp, as well as giving beautiful and detailed depictions of the clothing. (For bray harps, click here; for bray lutes, click here.)

This illustrates that accuracy or inaccuracy in one aspect of a painting does not equate to accuracy or inaccuracy in every detail, and of course this applies to art of any era. One renaissance painting is a good example of the principle: Hans Memling’s triptych for the Santa María La Real monastery in Nájera, Spain, 1480s (Koninklijk Museum Voor Schone Kunsten – Royal Museum of Fine Arts, Antwerp, Belgium). Memling paints in apparently realistic fine detail, but all is not as it seems.

This painting is often shown in modern presentations without the middle panel (Christ flanked by six singers) so as to display a continuous line of musicians: psaltery, tromba marina, lute, folded trumpet, shawm, straight trumpet, folded trumpet, portative organ, bray harp and vielle (medieval fiddle). The three large panels each measure 170 x 230 cm, thereby enabling Memling to paint detailed and life-size depictions.  

Memling’s depiction of the psaltery is of a type unique in art, both because of its artistic detail and due to the nature of the instrument itself. It has 4 roses and 61 strings counting from the left, with an additional 22 shorter strings counting from the right, which terminate in an arc of pins inserted into the soundboard. (An analysis of what this may mean musically can be read here under the subheading, The development of the psaltery 1: Berkeley, Memling, Albani and Girolamo).

Memling’s portrayal of the lute is equally detailed. Of particular note, below top right, is the sliver of quill plectrum in the player’s plucking hand and the two extended digits, perhaps indicating the transitional method of using plectrum and fingers before the plectrum was discarded altogether (a process described in this article about plectrum technique). Such specific details may lead to the assumption that the whole of the triptych is factually accurate. However, note the erroneous detail bottom right: we see the heads of 3 tuning pegs on the underside of the peg box and the tail end of 4 more which would have heads on the top side, indicating 7 pegs in all, making a lute with 7 single courses; but the neck shows a lute with only 6 single courses.

Artists’ errors and omissions can be resolved by reference to other art and written sources. For example, did Memling’s lute really have 7 pegs (as shown) for 7 strings, or 6 strings (as shown) with 6 pegs? That cannot be answered definitively for this individual instrument, but in lute iconography of the 1480s we see that lutes generally had 5 or 6 courses; in lute music a 7^th course is needed for one piece of music in the Italian Pesaro manuscript of c. 1480s–1511, and a 7^th course is mentioned in 1511 by Sebastian Virdung, Musica getutsch und außgezogen, but otherwise only 6 courses are required in written or published music until 1569, when Giulio Cesare Barbetta’s music required it in his Il Primo Libro dell’ Intavolatura de Liuto, after which it became more common. This means there is a very small possibility that the number of pegs for 7 courses may be right, but the number of strings in 6 courses is far more likely to be correct.

Literally any human endeavour is marked by ingenuity, accuracy and error, and iconography can be no different. This being so, we should expect there to be mistakes in music iconography, as in medicine, geography, engineering, or any other activity. But, even taking into account the way artistic conventions, symbolism, the medium, and artists’ errors affect presentation, what we find in the considerable collective testimony of medieval art is agreement on the features of an instrument. This means that, even when a single image is incomplete or beset with errors, we can nonetheless make firm historical assertions based on combined iconographical data.

6^th principle: test representation against practical reality

We have seen examples above where musical iconography cannot be taken literally because it is contrary to practical reality. In the 3^rd principle: take account of artistic conventions, we saw flattened perspective; necks that bend so that the carved head on the peg-box faces the player; the impossible ‘arm under’ plucking position; proportion distortion; and missing straps on large instruments. In the 4^th principle: the medium affects representation, we saw one effect of carving in stone, that strings appear to be all the same diameter, all carved much thicker than reality.

To make a judgement about the 5^th principle: the proficiency of the artist, we need to be aware of what practical reality dictates, such as that an instrument neck cannot bend, that a harper plays with a hand on each side of the string band, not with both hands on the same side, that the number of strings and pegs should be the same, and so on. This question of practical reality should also be considered separately as the 6^th principle: test representation against reality.  

The lute above, painted in 1439 by Spanish artist Blasco de Grañén, is most unusual in that the maker’s label is visible through the larger of the two roses. Add to that the clearly-depicted double frets and we may initially have the impression of an artist faithfully replicating reality to the best of his ability. A closer look reveals several practical impossibilities.

The lute is shown with 4 double courses – or is it 5? The number of tuning pegs appears to be 7, given the 3 pegs and 4 peg stubs on the completely visible side of the peg box, which would indicate 3 double courses and a single top course. This is contradicted twice over by the strings. We count 4 double courses at the nut, requiring 8 pegs, but 5 double courses over the roses, requiring 10 pegs. The first course beginning at the nut disappears under the fretting hand and does not reappear on the other side of the hand. What begins as the fourth course at the nut becomes the third course past the neck due to the disappearance of the first course, and the fourth and fifth courses on the body do not make it onto the neck. The misalignment of courses means that the roses are not placed directly under the strings as they should be. At the bridge, the first course appears to be attached impossibly to the decorative end rather than to the main body of the bridge. 

There is more confusion when we consider the practical reality of fret placement. Two strings of the same vibrating length, string material, diameter, density and tension will play the same note. Halving the string length by stopping the string with a finger will make the note an octave higher, and stopping the string at one third of its length will raise the note by a fifth. For the interval of a fifth, one third along the string length, the string is stopped at the 7^th fret.

Above we see Blasco de Grañén’s painting with the total string length marked by the longest white line; a third of that length marked by the shorter white line, showing the practical placement of a fifth above the open string at the 7^th fret; and the painted 7^th fret is marked by an arrow under the neck. We see that Blasco puts the 7^th fret in apparently exactly the right place, indicated by the left arrow under the neck, pointing to the 7^th fret. However, the question mark near the nut indicates where the missing 1^st fret, not painted, ought to be. If the 7^th fret were shown in the realistic position in relation to other frets, with the missing 1^st fret added, the 7^th fret would be at the right arrow underneath the neck, and therefore not one third of the string length, hence the question mark between the arrows to indicate the ambiguity of 7^th fret placement.   

It would be wrong to suggest that all medieval depictions of instruments have the same multiplicity of practical problems seen in Blasco de Grañén’s painting, but one consistent feature in medieval art is the impractical measurement of fret placement for notes. We see again in the San Sepolcro Altarpiece, below, that the placement of the 7^th fret is not correct.

For the sake of consistency and reality-testing, below is a photograph of lutenist Thomas Dunford holding his lute at a very similar angle to the viewer as in the San Sepolcro Altarpiece, to see if the 7^th fret appears in the correct place from the viewer’s vantage point, that the position of the fret is not significantly altered by perspective distortion. As expected, since the lute is straight on to the viewer, the 7^th fret appears in the photograph in exactly the correct place, unlike in medieval art.   

As we see in the typical example below, artistic licence in fret placement continued into the renaissance and baroque periods. 

Returning to Hans Memling’s triptych for the Santa María La Real monastery in Nájera, Spain, 1480s, we see the same contradiction between some realistic fine details and, when it comes to fret placement, artistic licence. Below is Memling’s fiddle with a tall, castellated, flat bridge, supporting 5 strings, all strings arranged singly (below centre). The bridge is remarkable for its apparently meticulous representation, but we can question whether it is completely accurate in showing 5 single strings when we have other indications of 5 string fiddle tunings with single and double courses (more of which in the next section). 

On the fingerboard (above right), the highest or final fret is tied onto the part of the neck that overhangs the body. The most likely explanation for this detail is a real observation by Memling about final fiddle frets. However, the 7 frets are too few in number and placed in musically impossible positions.

As we have seen, medieval and renaissance art yields important and valuable information, yet some of the finer details contradict the practical reality of a functioning instrument. The valuable realistic details and the practically impossible exist in tension together. It follows that, when trying to replicate an instrument from iconography, we should always bear in mind that artists were creating art, the appearance of an instrument in the eyes of the viewer, not a detailed to-scale plan, accurate in every detail to be used by a luthier.

7^th principle: refer to contemporaneous written evidence

There are some details for which the contemporaneous written word is critical in confirming, supplementing or modifying information gained by iconography. We have seen above under the 4^th principle: the medium affects representation, that string numbers in artistic representations are shown to be either accurate or modified when compared to written accounts, and under the 5^th principle: the proficiency of the artist, that the number of strings depicted can be tested against the number of courses needed in written music, at least for renaissance instruments where such music is available. In art we can count the number of strings on an instrument, but only literature can confirm or contradict string numbers, and give us string materials and tunings.

The evidence of multiple medieval writers testifies to gut being the chief string material, a fact which could not be gleaned from images alone. Perhaps the best example is the list of instruments given by Jehan (Jean) de Brie, Le Bon Berger (The Good Shepherd), 1379, which survived through 16^th century printed abridgements. De Brie wrote that the “fine sinews of [sheep] gut well washed, dried, twisted, scraped, wiped and spun are used for musical instruments: vielles [fiddles], harpes, rothes [rotas], luthz, quiternes [gitterns], rebecs, choros [string drums?], almaduries [?], symphonies [also known as the organistrum, predecessor of the hurdy gurdy], cytholes, and other instruments that one makes to give sound by means of the fingers and of strings.”

Other written testimony gives some exceptions to the generality of gut stringing.

The psaltery had strings of brass or of silver according to Bartholomaeus Anglicus, an English scholar working in Paris, in his Latin De Proprietatibus Rerum (On the Properties of Things), c. 1250, and its international translations by Jean Corbechon in French, La Proprietaire des Choses, 1372; John of Trevis in English, De Propietatibis Rerum, 1398; and an anonymous Dutch translation, Van den Proprieteyten der dinghen, 1485. 

The earliest account of wire/metal strings on harps is in Topographia Hibernica, written by Gerald of Wales in 1185–88. “It is to be observed that Scotland and Wales … strive in practice to imitate Ireland in their melodies. Ireland uses and takes delight in two instruments, the cithara [harp] and tympanum [hammer dulcimer] … they [presumably just the Irish] play upon bronze strings rather than [or more than] strings made of gut.”

In 14^th century Wales, it was a matter of national pride that harps were strung with horse hair. In his Cywydd Moliant i’r Delyn Rawn a Dychan i’r Delyn Ledr (Cywydd – a metrical form – of Praise to the Horsehair Harp, and a Curse upon the Leather Harp), Welsh poet Iolo Goch (c. 1320–c. 1398) wrote of the “newfangled harps of leather”, by which he meant strings of gut being introduced into Wales, and he disdained their new popularity in his nation as detracting from the national character of Welsh harps.

Iconography demonstrates that the vielle, viella or medieval fiddle could be thought of less as a singular instrument and more as a family: it came in a variety of shapes and sizes; had several types of flat bridge or arced bridge; and was played with 3, 4 or 5 strings, or 6 strings in a few cases if the art is to be taken literally.

While iconography can visualise the physical features and varying string numbers on a fiddle, only literature can provide us with tunings. Tunings for the 5 string vielle are included in Tractatus de Musica, c. 1280, by Dominican friar and music theorist, Jerome of Moravia (in the modern Czech Republic). Jerome considered that the fiddle or viella/vielle “should have” 5 strings, which suggests he knew of fiddles with fewer strings that he didn’t care for and therefore didn’t describe. No other writer gives tunings for fiddles with fewer than 5 strings.  

Some vielles had a bourdon, a string off the fingerboard that can therefore play only one note, used for plucking with the thumb or bowing, and one of the three tunings given by Jerome indicate this. His fiddle tunings are as follows:

A copy of Jerome’s treatise came into the possession of Pierre of Limoges, who annotated his copy sometime before 1306. Jerome is clear that the first tuning has a bourdon and the second does not. Of the third tuning, Jerome makes no comment about a bourdon, but Pierre’s additional notes state that the third tuning has one.

Christopher Page (1979) has observed that, though Jerome doesn’t state it, the nature of the tunings, with adjacent strings in unisons or octaves, indicates logically that strings would have been arranged in a mixture of single and double courses, the doubles being unisons or octaves, as follows:  

Tuning 1: d’ – d’ – g – G – d bourdon = 3 courses, arranged d’/d’ – G/g – d bourdon.  

Tuning 2: g’ – d’ – g – G – d = 4 courses, arranged g’ – d’ – g/G – d.

Tuning 3: c’ – c’ – d – G – G = 3 courses, arranged c’/c’ – d – G/G (Jerome), or 4 courses, arranged c’/c’ – d – G – G bourdon (Pierre of Limoges).

It is notable that the third tuning, for which Jerome does not give a bourdon but Pierre of Limoges does, is almost identical to crwth tuning, on which the low G/G is a double-course bourdon, as this article explains.

Can iconography confirm the reasonable conjecture of single and double courses on fiddles?

Some of it can.

Above left is a fiddle in the Tabernacle of Saint Savin, Hautes-Pyrénées, France, 1325. It shows a bourdon and two double courses, indicating that Jerome’s first tuning is correctly interpreted in courses: d’/d’ – G/g – d bourdon.

Above centre is a detail from the polyptych, The Coronation of the Virgin, signed and dated 1388 by Bartolo di Fredi, for San Francesco Church, Montalcino, Italy. It shows the same arrangement, indicating Jerome’s first tuning.

Above right is a fiddle in an altarpiece of the Passion of Christ, from the Cathédrale Sainte-Marie de Pampelune, Spain, painted by Juan Oliver in 1330. This is a third example of Jerome’s first tuning, with an extra bourdon on the viewer’s left of the strings which would make the instrument impractical to play. It is possible that Juan Oliver got this wrong, equally plausible that a later restorer, not understanding the tuning, added the extra string thinking that the string arrangement should be visually symmetrical. 

There is other iconography of fiddles which appears to confirm Jerome’s 5 string tunings, without confirming double courses. One such is on the internal north wall of Beverley Minster, Yorkshire, 1330–90 (below). This fiddle has a bourdon, which would theoretically correspond to Jerome’s first tuning, or possibly the third if we give priority to Pierre’s testimony rather than Jerome’s.

In Beverley Minter’s arcade is a fiddle with all 5 strings on the fingerboard, not arranged in single and double courses (below). This would therefore correspond to Jerome’s second or possibly his third tuning.

The fiddler in the angel choir of Lincoln Cathedral, completed in 1280 (below), has 5 single strings with a bourdon. This indicates Jerome’s first tuning, not delineated into single and double courses, if taken literally.  

We described in the 4^th principle: the medium affects representation, the limitations of carving in stone, but this does not extend to an inability to demonstrate single and double courses. We see this below, another fiddle in Beverley Minster, this one carved c. 1335–50 on the rear external wall of the reredos (altar screen). It clearly shows a flat bridge, with 4 strings divided into 2 double courses. This could be Jerome’s first tuning, d’/d’ – G/g, without the d bourdon, or it may have been another tuning altogether, since Jerome only gives tunings for fiddles with 5 strings.

On the right is the whole of Saint Mary Magdalene holding a crucifix, c. 1395–1400, by Italian painter Spinello Aretino, showing angels playing double recorders, fiddle, psaltery, bagpipe, trumpet and portative organ. Two details showing the fiddle and psaltery are below. The fiddle is shown with 5 strings as per Jerome, with a flat bridge and without a bourdon. If we take the painting literally, each string is played singly, not arranged into courses. This may therefore correspond to Jerome’s second tuning, g’ – d’ – g – G – d, or possibly his third tuning c’ – c’ – d – G – G if we give priority to Jerome rather than Pierre or surmise that the bourdon was optional in the third tuning.

The stringing of the psaltery (below) is shown clearly by Aretino: double courses, correctly spaced, each string hooked onto the hitch pin at half its length and fed back so that a single string is played as a double course. We might argue that Aretino’s precision on the psaltery favours a true to life portrayal of the fiddle, indicating that some 5 string fiddles had double courses, others were all single. However, as we have observed with Hans Memling, fine detail in one aspect of a painting does not indicate the same level of accuracy throughout.

We see, then, that this 7^th principle: refer to contemporaneous written evidence is necessary to understand practical musical reality, as it confirms, supplements or questions information gained from iconography. In cases where literature and art are at odds, we must either decide to give one source of information priority over the other, or discern that we may be observing alternative practices.

8^th principle: refer to contemporaneous material evidence

A knowledge of the general historical material context, meaning contemporaneous artefacts, is sometimes vital in interpreting what we see in music iconography. Two examples will suffice to illustrate the principle.

In the first example, writing implements help to interpret music iconography.

Below left we see a miniature from Pierre Lombard’s Collectanea in omnes divi Pauli epistolas (Collection of all Saint Paul’s epistles), Paris, 1170–80 (Bibliothèque nationale de France, Latin 14266, folio 95r). It illustrates Saint Paul with a quill pen in his right hand, showing the calamus (hollow shaft) and carved nib, and a knife in his left hand for holding down the page and rubbing out mistakes. Below right is a psaltery, played with two quills fashioned in the same manner, from the ceiling of the chapel in the collegiate church of Saint-Bonnet-du-Gard, France, 1418; …

… and below is a lute played with a similarly modified quill in a 15^th century window of Saint Mary’s Church, Atherington, Devon.

As per the 7^th principle: refer to contemporaneous written evidence, this practical observation is confirmed by Pavel Žídek of Prague (also known as Pavel Pražský, Paulus de Praga, Paulus, Paulerinus, or Paulirinus) in the chapter, Musica instrumentalis, in his Liber viginti atrium (Book of twenty arts), c. 1459–61 (Jagiellonian Library, Kraków, Poland, classified BJ 257). Pavel Žídek stated that the psaltery is “plucked with a quill held in the hand like a lute”.

From iconography, we see that quill plectrums were used for a range of other instruments: the cetra; …

… the gittern; and the citole.

As an article about the evidence for medieval plectrums describes, a quill was not the only material used for a plectrum. There is one type of plectrum that appears in iconography played almost exclusively with the citole, the only other instrument being the cetra. By reference to contemporaneous material evidence, we can observe that this citole/cetra plectrum was also modelled on a writing implement: the stylus used with a wax writing tablet.

A writing tablet was a frame of wood, bone or ivory, with slightly recessed panels filled with wax. It was a temporary and malleable writing surface used for practice and for temporary records, on which mistakes could be easily corrected and the same wax continually reworked to start again. 

The stylus and tablet were in use for millennia, from Mesopotamia in the 8^th century BCE to the fish market in Rouen, France, in the mid-19^th century. The Roman stylus was typically made of iron; those made in medieval London were usually carved from animal bone. On the non-writing end, the stylus was either spatulate or globular as an eraser to smooth the wax. The writing end was either sharpened to a point or had a sharp metal tip inserted.

The simple designs of stylus and tablet remained stable. The example on the right, made of wood and wax, is from Byzantine Egypt, dated 500–700 CE (The Met Fifth Avenue, New York). Below is a depiction of Ambrose Theodosius Macrobius, c. 385–c. 430, a Roman writer, grammarian and official, shown using a wax tablet and stylus in a mid-12^th century manuscript of Macrobius’ Commentarii in Somnium Scipionis – Commentary on The dream of Scipio (Det Kongelige Bibliotek, Copenhagen, ms. NKS 218 4°, folio 46v).

Below left is a citole in a copy of Aristotle’s Nicomachean Ethics, translated by Robert Grosseteste in 1250–60 (ms. 0222, folio 9v, Bibliothèque Municipale, Avranches, France). On the right is an object identical to the citole plectrum. This is a cast copper alloy Roman stylus with a missing point, found in 2010 in the Malvern Hills, Worcestershire. It is 11.32 cm long, the right size to double as a plectrum, dated 50–350 CE, around a millennium before there were any citoles.

Below left is a citole in the English Peterborough Psalter, 1300–25 (KBR – Royal Library of Belgium, Brussels, Ms. 9961–62, folio 14r). Adjacent to it are two objects resembling the plectrum. The first is a stylus with its iron tip intact, found in 2015 in Vintry, London, made of bone or ivory in 1270–1600. This is the right period for it to double as a citole plectrum, and it may have functioned at such at 6.9 cm long. Archaeological finds of a similar design often have similar measurements, or may be up to 9.3 cm long. Below right is a stylus of bone, 8.4 cm long, missing its metal tip, found in London and made in the 13^th–15^th century, again the right period for the citole. These measurements are considerably smaller than the plectrum depicted in The Peterborough Psalter, if we take the proportions literally but, as discussed above, size and perspective distortion were typical aspects of medieval iconography, as we also see with the bent neck to show a peg-box with a carved animal head, turned back to face the player.

Among the angel musicians in the ceiling vault of Gloucester Cathedral, c. 1350, is a citole player holding a plectrum very similar in design to a 14^th–15^th century decorated bone pin, 6.5 cm long, identified by the Museum of London as either a weaving tool or a stylus. It could just as easily be a citole plectrum.

Such contemporaneous material evidence, compared with iconography, leads to the conclusion that this type of citole plectrum was a writing stylus modified to pluck strings. The sharpened end of the bone could easily have been made sharp enough to be good for musical precision and manual dexterity but not so sharp that it would shred the strings; or, rather than inserting a metal tip into the stylus, a small quill could make contact with the strings, in the same fashion as quills were inserted in the plucking mechanism inside keyboards. (For a more detailed discussion of medieval plectrum materials and manufacture, click here.)

In the second example to illustrate the value of referencing contemporaneous material evidence, fashion helps interpret the date of a musical instrument.

In 1986, a koboz was found in a latrine in Elbląg, Poland, dated to 1350–1450 (described in this article). Most instruments survive the centuries because they were highly prized due to their quality and often ornate decoration. This koboz is special because it is roughly made, clearly played by one of the more impoverished inhabitants of or visitors to Elbląg.

The koboz peg box features a simple carving of a woman’s head with a head-dress (right), which may be a clue in dating the instrument more precisely than 1350–1450. Historical fashion not being my area, I consulted the proprietor of historical clothes manufacturer Cloak’d and Dagger’d, Jackie Phillips. Jackie identified the head-dress as a goffered veil. Goffered veils were made on a handloom with a weaving technique that allowed for more threads to augment the selvedge (the edge of a fabric, woven so that it won’t tangle or fray), creating a woven frill, often worn in several layers, creating bulk at the forehead.

This type of head-dress was popular in the 14^th century, which would halve the time period of the koboz to 1350–1400 rather than 1400–1450 if this was a high class instrument. Since the crudeness of manufacture marks this as a low status instrument, we might expect the carving to reflect low status clothing. Fashion originated in urban centres and spread outward, so a woman some distance from a city could still be wearing a veil considered out of date in the metropolis. This expands the possible period of the instrument beyond 1400 again, but still probably favours an earlier rather than a later date. There is a more sophisticated representation of a goffered veil on a reliquary (receptacle for keeping a holy relic), dated 1350, now in Museum Schnütgen, Cologne, shown in 3 photographs below.

In summary, we see that a general knowledge of historical artefacts pays dividends, as it allows the researcher to identify relationships between contemporaneous objects and gain further insight into historical instruments.

9^th principle: consult the evidence of surviving instruments

The evidence of surviving instruments is critical for interpreting what we see in iconography and for applying what we read in written accounts. In this section we will briefly compare musical iconography with surviving pre-medieval Egyptian harps, the British Museum English citole of 1280–1330, and the Polish koboz of 1350–1450 found in Elbląg.

The earliest surviving harps were discovered during early 20^th century excavations of Egyptian tombs. They were made in two forms, arched harps and angular or angle harps. Below left we see an arched harp dated to the New Kingdom era, 16^th–11^th  century BCE and, below right, an image of a harp of precisely the same design from the decorations in the tomb of Nakht, 8^th Dynasty, 1450 BCE. This striking comparison is one of several examples to show that, though highly formalised and stylised, and with the caveats described in this article, ancient iconography nonetheless gives us credible information about real-world instruments. We will see the same with the surviving English citole and Polish koboz to follow, and with the Italian gittern in the third and final article.

As we have seen above, written testimony states that medieval and renaissance instruments were generally strung with gut, with some notable exceptions. Gut strings are made by processing animal intestines, and there is clear evidence for their use in the ancient world, well before the medieval period. When archaeologist Harry Burton excavated the ancient Egyptian tombs of Thebes in 1823, he discovered harps that still had their gut strings intact and unperished, and remarkably they still played after 3,000 years. One such instrument is below, an arched harp from the tomb of Ani, 1250 BCE.

Our second example of a surviving instrument is the citole. The British Museum citole is dated 1280–1330, the period when they appear most in written sources and in iconography. The instrument both confirms and adds to the information we have about citoles from medieval art.

As we have seen above (and as this article describes in more detail), there were four basic forms or outlines of the citole: hexagonal, holly leaf, hourglass and vase-shape. With its two shoulders that slope from the body to the neck, a waist and a rounded lower bout, the British Museum citole confirms the vase-shape outline seen in art. Its wedge-shape, expanding from tail to head, and the thumb-hole at the back of the neck, is exactly as seen in medieval carvings, paintings and manuscript art; and the instrument is small, its overall length only 60.8 cm, confirming the size of most citoles in medieval images. 

This rare survival, one of the few extant medieval instruments and the only citole, adds further information that could not be gained from iconography. Images could not tell us that the citole is monoxylous or monoxyle – carved from a single piece of wood, boxwood in this case. To this the soundboard, bridge, string-holder, fingerboard and pegs were added. On the British Museum citole, all of these parts were removed and replaced in 1578 at the behest of Robert Dudley, to give as a gift to Queen Elizabeth, an ornate citole-turned-violin. It is difficult to imagine how it could have worked as a violin.

The detail above of the fingerboard shows that it was hollowed out to make the instrument lighter. This detail of the internal structure of an instrument could not be gained from iconography. As we will see below, other surviving instruments raise the question whether this was standard practice for medieval monoxyle instruments.

The new violin fingerboard of the converted citole is wider at the top end, as we would expect, to accommodate the fanning out of strings from nut to bridge, but this is not the case on the profile of the original neck nor, therefore, of the original fingerboard, which was a uniform width of around 30 mm. Whereas most gittern and lute iconography shows a gradual fanning out of string spacing, narrowest at the nut, widest at the bridge, citole iconography almost always shows uniform string spacing across the whole length of the string. This observation is supported by the fact that the BM citole neck had to accommodate 6 strings in 3 double courses within a very narrow width of 30 mm. It is difficult to imagine how 3 double courses could practically fit on the neck or function under the fretting hand if the string spacing was narrowed further still at the nut to make the strings fan out toward the bridge.

Another aspect of instrument organology which can only be gleaned from a surviving example is wall thickness. The late 13^th–early 14^th century carving on this citole is extraordinarily fine, skilled and detailed, including foliage and hunting scenes with people, a winged dragon, a dog and rabbit, a human-lion hybrid archer …

… and the magnificent head of a dragon with wings, claws, fangs and green eyes.

We might imagine that such intricate carvings would mean the walls of the citole are substantial, but the solid walls of the sides are only 3 mm thick. The relief carvings add another 3-4 mm but they are undercut, full of gaps and therefore light in weight – again, information we can only gain from the material evidence of an actual citole.

(For much more on the British Museum citole and citoles generally, click here.)

The final surviving instrument is a striking example of the important interplay of knowledge gained from iconography and from a surviving example. In 1986, an instrument carved from solid wood, identified as a gittern by Dorota Popławska (1997, 2002), was found in a latrine in Elbląg, Poland, dated 1350–1450. That identification has been repeated by every commentator since but, as I show in two articles (here and here), this is a mistake: the instrument is not a slim-necked and fretted gittern, but its central and eastern European counterpart, the wide-necked and unfretted koboz.

That discovery was made by asking luthier Paul Baker to make me a replica. Through that we realised that the wide and substantial dimensions of the neck prohibited it from ever having frets, and the wide string spacing made it impossible to play in the manner of a gittern. Understanding that led me to seek iconography of the wide-necked fretless koboz, which I found in both central and western European sources, including some instruments which had previously been assumed to be gitterns. Examples from three sources follow.

For an instrument to sound, it has to vibrate. We would expect a luthier to build an instrument where the sound chamber – in this case, the carved bowl – resonates, but a logical assumption is that the neck does not need to resonate, so it would be carved solid. However, as we see below, the neck of the Elbląg koboz was hollowed out, too.

For a modern luthier, there would appear to be no theoretical or practical reason to hollow out the neck, except perhaps to make the instrument a little lighter. However, as noted above, the neck of the British Museum citole was also hollowed out under the fingerboard, as was the 5^th–6^th century Byzantine pandura found in a grave excavation at Antinopolis, Egypt, by Albert Gayet in 1907, seen below.

These accumulated examples raise the credible question whether hollowing the neck was standard practice. Since all these instruments are rare survivals, it is unwise to extrapolate to universal assumptions, yet both the range of instrument types and the date range are compelling arguments in favour of hollow necks as standard for medieval monoxyle instruments.

In summary, the value of a surviving instrument is that it confirms the credibility of iconography as a source of information; it confirms observations and interpretations of iconography; and it adds details about fine measurements and internal structure, such as wall thickness and how much of a monoxyle instrument was hollowed, details iconography alone cannot yield. 

10^th principle: the cumulative weight of testimony

In the previous 9 principles, we have considered ways to get close to the physical reality of a medieval instrument by our open questioning approach, through understanding the nature of medieval art, through the testimony of written and material evidence, and through surviving instruments. It is important to understand that these 9 principles work together; that each piece of evidence is partial; that different types of evidence are not available in equal quantities – we have a great deal of medieval and renaissance art but, from the medieval period, few surviving instruments; and that each type of evidence does not carry equal weight.

The attempt to avoid modern assumptions and apply historical principles is foundational. Art that includes cithara symbolism (2^nd principle) gives us information about the physical features of an instrument, but when an instrument is held upright in imitation of the cithara, it gives us no information about how the instrument was played. Knowledge of artistic conventions (3^rd principle) helps us to understand the nature of art and prevents us from being incongruously literal about proportions and fine measurements, which therefore cannot be established objectively from art. Knowledge of the medium (4^th principle) can also help us avoid absolute literalism, since strings in stone cannot be carved as delicately as in life, and a tiny musical figure in the corner of a manuscript page cannot be expected to be fully detailed. Sometimes our search for the reality of an instrument must take account that an individual artist just wasn’t very good or, conversely, produced superbly realised figures, though not necessarily with 100% consistency (5^th principle). The practical reality of the image can be tested, such as plucking arm position and fret positions (6^th principle). More evidential weight must be given to instances where written testimony and iconography agree (7^th principle), where material evidence informs our interpretation (8^th principle), and where we have a surviving instrument to compare to its depiction in art (9^th principle). When an instrument does survive then, in line with the 1^st principle – to ask open questions rather than make assumptions – we should be careful of generalising based on only one example unless there is clear evidence that the instrument is indeed representative. The features and characteristics of any one surviving instrument or its representation in iconography can be compared to a great many other images of the same type, leading to credible general conclusions based on accumulated evidence.

10 principles

This article has set out 10 principles for discerning the material reality of musical instruments from portrayals in medieval art and other relevant sources. These principles can be expressed as 10 questions under 4 headings, as follows:

The importance of open enquiry

1. Do we have an enquiring approach, asking open questions rather than making closed assumptions?

The nature of medieval art

2. What aspects of the image are symbolic, and does this affect representation?
3. What artistic conventions are followed, and how does this affect representation?
4. How does the medium affect representation?
5. What skills or limitations of the artist are in evidence?

Reality testing

6. What is the practical reality of what is represented?
7. What is the contemporaneous written evidence for what we observe?
8. Can contemporaneous material evidence help interpret what we observe?
9. What is the evidence from surviving instruments?

Cumulative testimony

10. What does the cumulative weight of testimony suggest?

In the last of these three articles, these principles are applied to the recreation of a very special and detailed gittern painted on a fresco by Simone Martini in 1312–18. To access that article, click here.

 

Bibliography

Alexander, Jonathan J. G. (1997) Art history, literary history, and the study of medieval illuminated manuscripts. Studies in Iconography, Vol. 18 (1997), pp. 51-66. Available online by clicking here.

Arvin, Leila (1991) Scribes, Script & Books: The Book Arts from Antiquity to the Renaissance. London: British Library.

Backhouse, Janet (2000) Medieval Rural Life in the Luttrell Psalter. Toronto: University of Toronto Press.

Barber, Richard (1992) Bestiary, MS Bodley 764. Woodbridge: The Boydell Press.

Blanchard, Douglas (2018) Apollo the God. Available online by clicking here.

Blanchard, Douglas (2019) Apollo as Christ. Available online by clicking here.

Hamel, Christopher de (1992) Medieval Craftsmen: Scribes and Illuminators. London: British Museum Press.

Kennedy, Philip F. (2022) A Tale of Two Jackals: An Excerpt from Kalīlah and Dimnah. Available online by clicking here.

Kushnirskiy, Oleg (2016) Characteristic Features of Russian Icon Paintings. Available online by clicking here.

Kushnirskiy, Oleg (2017) Meaning of Colors in Eastern Orthodox Iconography. Available online by clicking here.

Lassetter, Leslie (1991) Music Iconography and Medieval Performance Practice. College Music Symposium, Vol. 31 (1991), pp. 91-116. Available online by clicking here.

Margerum, Alice C. (2010) Situating the Citole, c. 1200-1400. A thesis submitted in partial fulfilment of the requirements of London Metropolitan University for the degree of Doctor of Philosophy. August 2010. Available online by clicking here.

Museum of London (2020) Pin. Available online by clicking here.

Museum of London (2020) Stylus. Available online by clicking here.

Page, Christopher (1979) Jerome of Moravia on the Rubeba and Viella. The Galpin Society Journal, Vol. 32 (May 1979), pp. 77-98. Available online by clicking here.

Page, Christopher (1980) Fourteenth-Century Instruments and Tunings: A Treatise by Jean Vaillant? (Berkeley, MS 744). The Galpin Society Journal, Vol. 33, March 1980, pp. 17-35. Available online by clicking here.

Page, Christopher (1987) Voices & Instruments of the Middle Ages. Instrumental practice and songs in France 1100-1330. London: J. M. Dent & Sons.

Pittaway, Ian (2015) The mysteries of the medieval fiddle: lifting the veil on the vielle. Available online by clicking here.

Pittaway, Ian (2015) The Psilvery Psound of the Psaltery: a brief history. Available online by clicking here.

Pittaway, Ian (2021) “the verray develes officeres”: minstrels and the medieval church. Available online by clicking here.

Pittaway, Ian (2022) The citole: from confusion to clarity. Part 1/2: What is a citole? Available online by clicking here.

Pittaway, Ian (2022) The medieval harp (1/3): origins and development. Available online by clicking here. 

Pittaway, Ian (2023) The Elbląg ‘gittern’: a case of mistaken identity. Part 1/2: Why the koboz was misidentified. Available online by clicking here.

Pittaway, Ian (2023) The Elbląg ‘gittern’: a case of mistaken identity. Part 2//2: Identifying the koboz. Available online by clicking here.

Pittaway, Ian (2023) The evidence for straps used with medieval, renaissance and baroque musical instruments. Available online by clicking here.

Pittaway, Ian (2023) Medieval plectrums: the written, iconographical and material evidence. Part 1/2: Medieval plectrum materials and manufacture. Available online by clicking here.

Pittaway, Ian (2023) Medieval plectrums: the written, iconographical and material evidence. Part 2/2: Medieval and early renaissance plectrum technique. Available online by clicking here.

Popławska, Dorota (1997) Instrumenty Muzyczne Średniowieczne Elbląga [Elbląg’s Medieval Musical Instruments]. Archaeologia Elbingensis vol II, „Stare Miasto w Elblągu – Wyzwanie Historii”. Materiały z konferencji naukowej poświęconej pamięci Tadeusza Nawrolskiego. Elblag – Gdańsk 1997 [Archaeology of Elbing vol II, Old Town in Elbląg – The Challenge of History. Materials from the scientific conference dedicated to the memory of Tadeusz Nawrolski. Elblag – Gdańsk 1997]. Available online by clicking here.

Popławska, Dorota & Czechak, Tadeusz (2002) The tuning and playing of a medieval gittern and fiddle from Elblag, Poland. The Consort. The Journal of the Dolmetsch Foundation. Vol. 58 (Summer 2002), pp. 3-12.

Portative Antiquities Scheme (2014) Stylus. Available online by clicking here.

Portative Antiquities Scheme (2016) Parchment pricker. Available online by clicking here.

Spencer, Robert (1976) Chitarrone, Theorbo and Archlute. Early Music, Vol. 4, No. 4 (October 1976). Available online by clicking here.

Van Edwards, David (undated) Lute straps. Available online by clicking here.

Victoria and Albert Museum (2016) Introducing Opus Anglicanum. Available online by clicking here.

Victoria and Albert Museum (2016) The making of medieval embroidery. Available online by clicking here.

Watson, Giles (2016) (translator) Cywydd of Praise to the Horsehair Harp, and a Curse upon the Leather Harp by Iolo Goch, 14^th century, in Rivals of Dafydd ap Gwilym: A Treasury of Fourteenth and Fifteenth Century Welsh Verse, pp. 44-46. London: Lulu.