Designing a Medieval Lute

Note: I've recently written a new blog post that reviews some of the aspects of iconography and lute design that I've discussed here. Check it out: The Backside of an Angel



Today I'd like to talk about how I designed a recently completed 5 course fretless medieval lute. This instrument is the culmination of a couple of years of research, thinking, discussion and design with the client, Gus Denhard, of Seattle—a fascinating journey which (as always) I hope has yielded a playable and convincing musical instrument that the owner will get a lot of mileage out of.

Initial discussions between Gus and I focused on the basics: the size (and pitch) of the instrument, number of courses (5, instead of the usual 4), tuning (in fourths), and so on.  Gus didn’t have a specific design in mind, but he did have a firm idea of how he wanted to use the lute, which was to explore, with his musical collaborators, the connections between European and eastern/Arabic music of the Middle Ages.  The lute's a natural for this, since its direct ancestor, the 'ud, was brought to Europe by the Moors about a thousand years ago.   The modern 'ud, and what we now call the renaissance lute, are both descended from this early 'ud, and are therefore very close cousins.  My hope was that by judiciously combining certain features of both these instruments, I could perhaps find a way back to that common ancestor.

The main problem in designing and building an instrument like this is that no historical examples of such an early lute have survived.  Probably the earliest surviving European plucked string instrument, the gittern by Hans Ott of Nuremburg, dates to about the mid-15th century, and that's about as far back as we can get.  For anything earlier than that--say, one or two hundred years earlier, which is approximately the period Gus and I had in mind--we really need to rely on iconographic evidence.

Hans Ott giettern (photo courtesy of Wikimedia commons)

But which iconographic evidence?  Depictions of lutes abound from the middle ages--in altarpieces, manuscript illuminations, frescoes, wood carvings and so on, in a seemingly unending stream.  One piece of advice that proved invaluable came from Crawford Young, the early lute scholar and performer.  Actually, it was two pieces of advice: build the instrument lightly, he said (more on this later); and for a model, look at Sienese pictures from the 14th century.  He also narrowed the search further, by suggesting I consult Howard Mayer Brown's "Catalogus, A Corpus of Trecento Pictures with Musical Subject Matter," a series of articles published in the journal Imago Musicae in the mid-1980s.  What a resource!  I heartily recommend it to anyone interested in the subject: a near-exhaustive collection of pictures, indexed by artist and type of instrument depicted, all keyed to a list of thumbnail versions of the pictures.

I located the journal at the University of British Columbia library, spent a day up to my elbows in pictures, and emerged with one good depiction of the kind of lute that I thought would work.  Here it is, a Coronation of the Virgin by Andrea Di Bartolo, from the Ca d'Oro in Venice.

And here's a close up of the lute.  Why did I choose this particular picture as my model?  Well, there are a number of details that led me to believe it might offer a viable design.  First, it seemed to be of a size (in relation to the angel holding it) that could work for a lute with a string length of around 60cm, with the top course tuned in g', which was Gus's request.  It also had a ratio of neck length to body length that would allow for the equivalent of around 5 or 6 frets on the fingerboard, and 8 to the upper soundhole rosette (another of Gus's requests).  So the proportions seemed right, and indeed, when I put together a drawing using a compass and straightedge, the body outline conformed quite nicely to some simple geometry, using whole-number ratios to draw circles and arcs.  That in itself is a convincing detail.

But much else in the picture looked appropriate too.  The bridge had a nice shape, and looked a lot like the kind of bridge you might find on a modern 'ud.  The scratch plate in front of the bridge was another 'ud-like detail.  The soundhole rosettes are simple (in comparison to those that we see in renaissance lutes), but they're elegant, and their simplicity is more than balanced by the busyness of the rest of the design, including the soundboard edging (which looked to me like a kind of 'herringbone' binding), and the tilework inlay in the soundboard.

I found another example of this type of lute (not, incidentally, in H.M. Brown's article, but on flickr).  It appears in a fresco among a group of musical instruments (including a gittern) by Lippo Vanni, in the Church of San Lorenzo al Lago.  The larger image first:

And then the detail.  The picture resolution is not great, but you get the idea: two soundhole rosettes (both of simple design), bridge with scratch plate out front, tilework inlays, binding all around the top, and curving neck joint (rather than a blunt joint, as one finds on a renaissance lute).  All very similar to the image I used for my design, and confirming for me a lot of the features, both functional and decorative, that I decided to include.

So I had the model--now to design and build the actual lute.

One principle to keep in mind when working with an iconographic source is that a painting is not a photograph.  Parts will be exaggerated, and proportions distorted, to create an effect specific to the painting.  In a way, that's frustrating, because you can't just copy the image, but in another it's liberating, since you have to trust your own sense of proportion and your knowledge and experience of what will make a good, playable, working instrument.

That's nowhere more true than with the back side of the lute.  All a painting like this gives you is the plan view--it gives no information on the body, how deep it might be, what shape it has, how many ribs it has, or even whether it's built up from separate ribs at all (maybe it's carved from a single block of wood).  In designing the lute, these were the first big decisions that had to be made.

I took to heart Crawford Young's advice to build it lightly, and that meant, first, building a ribbed back, rather than carving the body from a solid block.  A simple aesthetic seemed right to me, so I went with 9 ribs (about the smallest number that's practical), and kept the longitudinal section basically the same as one half of the plan view (in other words, the body is pretty much semicircular in cross section.)

With the body designed, I turned my attention to the soundboard, and its many features--bridge, scratch plate, rose designs, tiles and binding.  All of these things needed their proportions re-imagined to match each other, and to match the scale of the whole instrument.

The first element that I tackled was the herringbone edge binding.  In fact, I think that was the very first thing I made of the entire lute, before the body or anything else.  It became a minor obsession, because I needed to make sure of two things: one, that I could manufacture the binding (actually two identical  bindings, one for each side of the belly), in a width that was right for the lute's proportions; and two, that I could bend it to fit around the edge of the belly, without it falling to pieces under the stress and heat of the bending iron.  If I could figure a way to do those things I'd be fine, but if I couldn't, I was pretty sure I'd need to rethink my whole concept of the lute.

The binding turned out well, and so I felt free to build.  Most of the other features I figured out as I went along, and as they emerged in the building process.  The bridge design, for instance, was a product of three influences: the overall shape came from the painting, the width derived from the string spacings that Gus had provided, and the height and other dimensions were adapted from my observations of modern 'ud bridges.  The roses needed a bit of redesigning to better balance the ratio of negative to positive space; they also needed to be repositioned, slightly, to accommodate the body shape I'd created, and the internal barring system I'd worked out (you can read about the barring system here.)  Once I had the sizes and positions of the bridge and roses, I could then size, position and glue the scratch plate.  And once that was in place, I could create and position the four tiles that really give the lute its characteristic look.

The tiles were fun to design, and to make.  The photo I had of the painting was not of a sufficient resolution that I could see very well what their actual design was, so I was left to come up with something that would create the impression of the tiles, while fitting in with the proportions of the rest of the lute.  Once again, it can be a bit of a hassle not to have things spelled out for you, but it can also be quite liberating.

 The same principle applies to the peg box, which is not detailed in the painting, and is therefore free to be imagined.  I decided on slightly curvy peg box, in homage to the elegantly slim peg boxes on modern 'uds; I also designed a simple, square-ish peg head to go along with it.

There's one last feature of the design that I'd like to mention, and that's the smoothly curved joint between the body and the neck.  You see this type of joint everywhere in the iconography of the period, so we know it was commonly used, but figuring out why it was used, or why it continued to be used for so long, is a bit problematic.  It gives the whole lute a very organic shape (rather like a bulb of garlic), and to me, at least, it's pretty easy to see its evolutionary origins first in instruments made from gourds, and then, later on, in instruments whose neck and body are carved from a single piece of wood.  This is the case with the Hans Ott gittern, and, one suspects, most other gitterns of the period; it's also easy to imagine that up until a certain point, lutes were made this way too.

However, somewhere along the line,  someone decided to try for a different sound, something a bit more refined, a bit more responsive, and rather than carving the back from a single piece, built it from separate ribs--and attached the neck after it was made.  A different kind of lute was born.  But why preserve that old-style neck joint?  Isn't it incredibly difficult to bend the rib in two different directions, and still get a good, close fit between them?

The answer to that question (as I found when I built the back) is that, yes, it is difficult, but not incredibly so.  In fact, I was a little surprised at how (relatively) easy it was.  The rib joints looked good, as did, indeed, the whole back, and the whole instrument.  Like much else about this lute, it represented a technical challenge that I was eager to accept.  (This is the way that I push myself to keep developing and refining my craft skills and why, even if I'm never asked to build another one of these lutes, it was well worth my while to do so this time around.)

And why did this kind of joint keep being used long after it was convenient to do so?  In my opinion, two reasons.  First, tradition is a very powerful thing: if the joint has always been made this way, why should it be changed?  Second, the joint itself feels very slick, and, I believe, lends itself to single-line playing up the neck and onto the upper part of the soundboard.  The joint, in other words, fits the music that was (and will be) played on it.

I hope this account has given you some idea of the kinds of choices I made to design and build this lute.  Most of the choices are, fundamentally, pragmatic ones, and have ultimately to do with the playability of the instrument: the width of the stringband dictates the size of the bridge and width of the nut, which both in turn dictate the dimensions of the fingerboard and neck, which then affect the size and contours of the body, which then, in turn, dictate the size and proportions of the decorative features of the soundboard.  Everything is of a piece, and all the features are based upon the body, hands, and technique of the player.

That player's name is Gus Denhard, and he drove up from Seattle about a month ago to pick up his new lute at my workshop.  His initial response was quite positive, and I'm looking forward to hearing from him in his musical journeys with this lute.  My impression throughout the design and building process was that the music and repertoire for this instrument were not already known, but were to be discovered, and for me that was an exciting feeling.  Without the spark that Gus provided, I may never have built an instrument like this; his encouragement and enthusiasm allowed me to follow my instincts, and try new techniques and designs.  He's promised to send me some musical postcards from his journey, and as he does, I hope to share them with you.