WHAT BRIDGE PINS CAN TELL US OF THE ORIGINS OF THE HARPSICHORD
*Author’s note: This was a research paper I wrote for an undergraduate course in Archaeology way back in the year 2000. It might be totally out to lunch. But I’m proud of the work I put into it, so I have included it.
I've been a builder of musical instruments for a few years now, and one day in the course of my work my father wondered (aloud, of course) about a peculiarity of harpsichords, of which he is a player. He wanted an answer, and I wanted a paper for my archaeology course; that was the genesis of this essay. Now, harpsichords are not an instrument which many people are intimately familiar. First I will give a brief history of harpsichords, disregarding other instruments which are almost identical in mechanism and which differ in shape and size, such as the muselar, spinet, and virginal. I will also provide a description of relevant parts of its anatomy before I present the particular problem which confronted me.
History
Musical instruments with keyboards, sounded by the means of strings (as opposed to the pneumatic organ or hydraulic hydraulus of ancient Rome) jump into the historical record quite suddenly, and there is still great debate as to their origins, due to the ambiguity and scarcity of literary records. From 1388 we first find reference to an instrument called the “chekker”, "resembling an organ. but sounded by strings”{1}. This may have been a clavichord, which is sounded by means of metal tangents striking its string(s), instead of being plucked by plectra as harpsichords are.
The first accurate representation of a proto-harpsichord is from 1425, at the Staatliche Museum in Berlin, illustrated below.
It is exactly the same shape as modem harpsichords, rather like a grand piano (which is evolved directly from it) with square comers, but tiny, only about 50 cm long. A remarkable feature of this representation is that it also shows a plucked psaltry, which provides striking proof that the harpsichord is descended from that instrument. We can presume that the harpsichord is a direct descendant of the psaltry because they have the same sounding mechanism (plucking}, but the similarity of their form (the Staatlicht proto-harpsichord is precisely the shape of half the psaltry) and other details clinch it{2}.
Fifteenth-century representation and literary references of harpsichords occur almost exclusively north of the Alps, in France, Britain, Germany and Sweden; however, we have names of several Italian harpsichord makers who were active as early as 1419{3}, and nearly all surviving early harpsichords (the earliest dates from 1503{4}) are from Italy{5}. They were made very lightly and gracefully out of cypress or cedar, with relatively short, low-tension, thin strings. Their sound quality is unique and unmistakable. It seems that in the sixteenth century northern Italy, particularly Venice, was the world centre of harpsichord making, and that instruments were sent from there to all parts of Europe{6}. There was thought to be a greater degree of standardization in the Italian school (and that northern harpsichord makers tried vainly to catch up{7}), because modem historians have found almost identical instruments with construction dates spanning 300 years. The geographic origin of the harpsichord has not yet been firmly pinned down, but the suspicion is that they were invented in the region of Burgundy and that Italians quickly cornered the market with larger and more effective instruments.
By about 1565 many harpsichords were being made with a second set of strings, tuned in unison with the first, and soon after that, by about 1600, we find more and more instruments having a second set of strings tuned an octave above the first{8}. The seventeenth century saw the rise of one of the most important dynasties of harpsichord makers, the Ruckers family in Antwerp. They became the most important and prolific producers of instruments while the Italians languished, partly because they were capable of making instruments which sounded and looked better than any others, but also because they were not loathe to introduce changes in construction to cope with the changing nature of music in the baroque era{9}. They build larger, sturdier instruments, strung at a higher tension than the Italian instruments, among other technical differences. They popularized the second keyboard, shifted so that it sounded a fourth below the standard pitch. They also used three sets of strings, two of standard pitch and one an octave above, of which any combination of keyboard(s) or strings could be chosen by mechanical means, which greatly increased the instrument's rather limited dynamic and tonal range. Later instruments, particularly in Germany, got to be very large and occasionally had another set of strings at an octave below standard pitch, as well as a third keyboard. By 1800, however, the harpsichord had been rendered almost completely extinct by the newly invented piano, and was rarely heard until its revival in the mid twentieth century.
Construction
Harpsichords are, by musicological definition, plucked chordophones. As you can imagine this means that they sound by means of strings which are plucked, like a guitar or harp. The mechanism of the harpsichord and its relatives the muselaar, spinet, and virginal is relatively simple at first sight. A key is depressed by the player. It pivots on a fulcrum, pushing the jack, a light wood stick of rectangular section which holds the plectrum, upwards, held in alignment by a vertical guide slot called the jack rail. The plectrum, a tiny spike of quill or delrin plastic (occasionally leather) is pushed into the string until it bends enough to let the string snap back down past it: a pluck. The strings can be made of brass, steel or iron, and on harpsichords run perpendicular to the keyboard. They are held in place at the far end of the instrument by being looped around sturdy hitch pins, run over a bridge of roughly triangular cross, section, which rests on a thin soundboard (reinforced on the underside by wooden bars) which covers the top of the interior of the instrument, past the jacks and plectra near the player, over the nut, which is similar in cross-section to the bridge but glued to the wrestplank, a solid bar of hard wood which runs parallel to the keyboard and also provides a secure fastening point for the near end of the strings, which are wrapped around iron tuning pegs inserted into that wrestplank. The plank of the case (or body) to the player's left, approximately parallel to the strings, is called the spine; the narrow far end of the instrument is called the tail. Running in a graceful curve from the tail towards the player on the right is the bentside, which terminates at the cheek, a short flat panel which ends at the right hand side of the keyboard. Early instruments, on account of their fragility, were usually stored inside a stronger case, and put out on a table to be played{10}. Flemish instruments never used separate outer cases because they were robust enough already, and soon Italian instruments followed suit, being permanently placed on a stand.
The particular detail which provides a startling amount of insight into the playing practice and tonal preferences of musicians (and their audiences) is this: as well as simply passing over the bridge and being held down against it by the angle produced by the bridge height, the string is also pulled at a sideways angle, against a thin metal pin inserted into the bridge. The bridge pin is an odd feature of the harpsichord; no other instrument uses it now, or ever did, even instruments with similar characteristics like harps or guitars. It is not at all necessary for structural reasons, the downward angle produced by the difference in height between the bridge top and the hitch pins is enough to hold the string firmly against the bridge, and the sideways angle allowed by the bridge pin is never great enough to increase the pressure. The sideways angle does allow the overall string length to be reduced while allowing the bridge to remain an adequate distance away from the edge of the instrument where the hitch pins are located (which is essential in order for the soundboard to vibrate properly), but this is only a matter of a few centimetres, totally insignificant considering the enormous lengths of wire which get used for the rest of the instrument. Moreover, the pins add an additional expense and complication to the construction of the harpsichord. They are also one of the main reasons structural failure in old harpsichords: they cause the strings to pull the bridge sideways along its short axis, causing it to come unglued from the soundboard and roll towards the bentside. Why would such a seemingly counterproductive device exist?
Theories
There were two schools of thought concerning the placement of bridge pins. The Italian builders put the pin snug up against the peak of the wooden bridge, while everyone else tended to place it several millimetres before it. Due to the varying angles of the bridge (it is curved along its length in a gentle S shape) to the strings, on a Flemish instrument the distance between bridge and bridge-pin varies from about 2 millimetres in the highest strings up to about 10 millimetres in the middle of the range. Bass strings are woven between two pins on either side of the bridge, effectively canceling out any tendency to roll the bridge, which means that the technique could easily have been applied to the higher strings but deliberately wasn't. There are a few possible reasons why this was done. First, I can testify from my own experience that musical instrument builders are notoriously conservative in their practice, or rather that they are forced to be so by their patrons who, generally lacking the knowledge and expertise of the builder, are mortally afraid that any innovation could only result in an inferior instrument. For example, respectable violin builders even in North America have been restricted to using the same European species of wood, picea abies for the soundboard and acer pseudoplatanus for the rest, for their instruments simply because that's all that was available to Andrea Amati, the first great violin builder{11}, nearly five hundred years ago. If Pinchas Zuckermann showed up for work one day with the wrong colour of violin he would likely be laughed out of a job. It is supremely possible that pinned bridges remained on Italian harpsichords until the present simply because a remote ancestral instrument required them in the fourteenth century. The way to find out if this is the case is to examine all known representations of possible ancestors of the harpsichord. Surprise, surprise, every single pre-1500 representation of psaltries show no bridge-pins, which could merely be an oversight on the part of the artist (it is a very small detail, and being a trained artist myself I can sympathize with their tendency towards less-than-photographic realism) except that they also show no sideways displacement of the string at the bridge or nut. This sideways deviation would have been obvious enough to show in many of the more detailed representations and would have indicated the presence of bridge-pins, but it isn't there, strongly suggesting thatthey weren't used. The only representation we have of bridge-pins in use (before we find them on actual surviving instruments) is a drawing by Henri Arnot de Zwolle, from between 1436 and 1454.
But why would any ancestors or proto-harpsichords require pinned bridges either? One rather shaky explanation is the following: All fifteenth-century representations of harpsichords show instruments ranging in length from about 50 cm to about 100 cm, tiny by later standards. It is presumed by Sibyl Marcuse{12} that these instruments were small due to the difficulty of drawing longer lengths of iron wire owing to the limited technology of the time. Available iron wire was indeed poor until about 1500{13}, which is when we start to see the emergence of the full-sized Italian harpsichord. However, excellent brass wire had been available since at least late Roman times{14}, and while its lower tensile strength would have necessitated a lower pitch and made the tone of instruments somewhat less bright, one wonders why it would not have been used. I would suggest that it would be a useful experiment to reproduce a proto-harpsichord and to see if brass strings really did sound so bad on it as to be useless. I doubt it would sound so dramatically different as to affect the evolution of the instrument; longer strings on later instruments were normally brass anyway, which would suggest that the quality of iron wire should not have been the mechanism to place a restriction on the size of harpsichords. But our few available reliable references say that early harpsichords were just always strung in iron{15}.
If this is true, and it is by no means certain, then it suggests a reason for the use of bridge-pins: displacing the string towards the bentside of a harpsichord allows it to have a shorter overall length while retaining the same pitch and keeping the bridge an adequate distance away from the side of the instrument; while a matter of a few centimetres may not matter on a full-sized instrument, it could mean significantly extending the range of a small early harpsichord. If we factor in the inherent conservatism of the musical instrument building trade, we have possibly found the reason why pinned bridges are used to this very day. Admittedly, this conclusion is based on a fair bit of shaky evidence. I would be much more confident in a more concrete explanation.
Acoustic tests: the apparatus
The only other reason I can think of that anyone would add such an unusual feature to a musical instrument would be, of course, to change the way it sounds. Does a bridge without a pin sound different from one with? Does changing the distance between the bridge and the pin make any difference? Hypothetically, it should; here's the hypothesis. A plucked string can be said to vibrate in two dimensions, the horizontal and the vertical. Having a pin a distance away from the bridge allows the string to have a different effective length in each plane: horizontally, the string's active length is determined by the position of the pin, but vertically its length is determined by the position of the bridge, because the string is elastic enough to literally roll up and down the pin. So in the vertical plane on a typical Flemish harpsichord the c' string is about 10 millimetres longer than it is horizontally, which means that it will be a slightly lower pitch. Here's the neat part: only the vertical movement of the vibrating string has an effect on the soundboard, which means that we can only hear the note given by the vertical plane, but the horizontal movement affects the vertical by gradually transferring its energy into the verticalmotion of the string. This phenomenon is very well documented in the January 1979 issue of Scientific American, for pianos which use two or three strings tuned slightly apart instead of separate vertical/horizontal string lengths. Theoretically it should also apply to any other stringed instrument, and the results should show a similar pattern: as the difference between the two tones grows (i.e. the distance between the bridge and the pin) we should be able to notice an increase in sustain up to a point at which the two frequencies can no longer couple together, resulting in a sudden decrease in sustain. I would also expect to hear a corresponding change in the tone quality as the distance increases
In order to see if this could be the reason for the existence of bridge pins on a harpsichord I would have to create an apparatus with similar response characteristics to a typical harpsichord, but which would have a bridge that would allow me to place the string on different pins at different distances from the bridge, and examine the sound produced to see if the pins’ existence or placement had any effect. First I established certain physical dimensions and aspects of construction which I felt would be typical of a Flemish harpsichord, the most common and generally flexible type{16}:
String length (nut to bridge): 680 mm
Pitch: c', 261.6 hertz
Distance from hitch-pin rail to bridge: 100 mm
Distance from bridge to cut-offbar(under soundboard): 105 mm
Position of plucking-point: 155 mm from bridge
Thickness of soundboard: 2.6 mm
Thickness of string: 0.24 mm
Because the purpose of the test apparatus was only to test the variable of the bridge pin distances, I made it narrow and rectangular, with a single steel string. The bottom and sides are made of pine because they served no acoustic purpose, only structural. The ends, serving the purpose of hitch-pin rail and wrestplank, are thick slabs of sugar maple. In Europe beech was more normally used but again, since their purpose is structural, the material used is of little consequence. The soundboard on the other hand is of crucial significance, so it is made of a quartersawn piece of sitka spruce, which is very similar in all respects to the German spruce used in European instruments{17}. It was planed to a thickness of about 2.6 mm, slightly thicker in the centre. Likewise the cut-off bar, which determines the area of the soundboard under the direct influence of the bridge by being glued across the bottom of the soundboard, is made of spruce about 10mm wide. The plucking device consists of a single lightweight lever tipped with delrin (a type of plastic used for plectra because of its similarity to raven quill in use)
The bridge pins are polished tempered steel wire 1.1 mm thick, similar in flexibility to the slightly thicker brass wire used on most instruments. The bridge and nut were carefully made of European beech to the following dimensions to retain the same mass as a standard bridge and nut:
Acoustic tests: the results
Five different bridge-to-pin distances (Lp)were tested. First, the string was plucked running over the bridge with no contact to a pin. Then the string was plucked with a pin located right next to the bridge, an Lp distance of 0.5 mm, the radius of the pin wire. Then three tests weremade with Lp distances of 2.5 mm, 7.0 and 12.5 mm, respectively. They were digitally recorded with a stereo condenser microphone connected to the Soundblaster AW32 soundcard of a PC for analysis.
Our own ears are too subjective on their own to be entrusted with the task of accurately defining what difference the pin position makes. But it is still a useful procedure to perform a simple listening test, as there are some aspects of sound which are readily audible to the ear but which don't translate well into computer-generated waveforms. Besides, no harpsichord maker that before 1960 would have been able to perform computerized sound analysis at all.
First, it was felt that there was probably little difference between the sounds obtained from a pin next to the bridge and no pin at all. The sound is best described as slightly nasal, suggesting a lower component of the upper partials of the harmonic spectrum. I found it relatively pleasant. The sounds obtained from Lp 2.5mm and 7.0mm were noticeably (though not dramatically) different from Lp 0.5mm. The tones seemed more singing and less likely to ebb and flow in volume over their duration. The tone obtained from Lp 2.5mm is my personal favourite, though to be fair it was often indistinguishable from Lp 7 .0mm. At the position of Lp 12.5mm the tone definitely took a turn for the worse. It becomes quite noticeably harsh, twangy, and thin to the point of being false. While the overall range of tone is certainly quite subtle, a pattern is formed which corresponds quite well to our hypothesis.
The computer generated wave-forms, provided in Appendix 1, tell a similar story to our listening tests. As the pin is shifted away from the bridge an out-of-phase element (the horizontal motion of the string) becomes less and less prominent. The decay rate of the sound becomes much more consistent up to Lp 7.0mm. At Lp 12.5, however, we notice that the volume of the sound drops very sharply very soon after the pluck, resulting in a lower volume.
Conclusions
I think it is fair to say that it has been demonstrated that the distance of bridge-pin from the bridge on a typical harpsichord has a significant effect on the sound of the instrument. At the right distance from the bridge, a pin does not specifically enhance the sustaining power of a note, but it allows the decay in volume to be much more evenly gradual, critically important for an instrument with no dynamic range (the sound volume of a harpsichord is constant; it cannot be changed by hitting the keys harder or softer like a piano). It is obvious why all harpsichords built north of the alps use pins placed at a distance from the bridge. However, Italian harpsichords retain the pin even though as far as we can tell it should make no difference to the sound, being located right next to the bridge. It is this inconsistency which can tell us something about the origins of the harpsichord.
We have seen that there are two possible reasons for the existence of bridge pins. First, the possibility that iron wire was so precious and short in length that bridge pins were a means of extending the range slightly: this is not a very tenable conclusion. The second possibility is far more likely, because the proper employment of bridge pins affected the sound of an instrument so advantageously. This strengthens the case that the harpsichord was invented in northern Europe rather than in Italy. Surviving instruments consistently show that in Italy the sound of a bridge with a pin right next to it (the same as no pin at all) was preferred, while in the north it is evident that the sound of a bridge with pins a certain distance away was preferred. The only excuse for the pins on an Italian harpsichord is that they, along with the corresponding sideways string angle, are a vestigial remnant carried over by tradition from the earlier northern proto-harpsichord.
There is still ample scope for further inquiry into the early history of the harpsichord. I have found no hard evidence indicating what type of wire was used before about 1600, a potential variable which could greatly affect the validity of the findings in this essay. It should also be possible to create a more accurate testing apparatus than my own, which was lacking in bass response and may not have been reflecting the acoustic tendencies of full sized harpsichords as well as it could have. It would also definitely be possible to perform much more sophisticated listening tests and computerized sound analysis.
I can't wait to get started!
Footnotes
Marcuse, 243.
Ibid, 260
Hubbard, 2.
Hellwig, in Ripin, 29.
Hubbard, 2.
Ibid, 4.
Ibid, 3.
New Grove, 17.
Sachs, 376.
Marcuse, 266.
Sachs, 358.
Marcuse, 260.
Goodway/Odell, 41.
Ibid, 22.
New Grove, 14.
All dimensions and aspects of construction are averaged from several specimens described in Frank Hubbard’s “Three centuries of harpsichord making”, as well as the instrument by William Post Ross In the possession of John Sankey, Ottawa, Canada.
Hoadley, 153.
Bibliography
Bessaraboff, Nicholas. Ancient European musical instruments. New York: October House, 1964.
Buchner, Alexander. Musical instruments through the ages. London: Batchworth press, 1961.
Galpin, Francis W. Old English instruments of music. London: Methuen & co., 1932.
Good way, Martha and Odell, J .S. The metallurgy of 17h- and 18th-century music wire. Stuyvesant: Pendragon press, 1987.
Hellwig, Friedemann. "The single-strung Italian harpsichord", in Ripin, Edwin M., ed. Keyboard instruments: studies in keyboard organology 1500-1800. New York: Dover publications, 1977.
Hoadley, R. Bruce. Identifying wood. Newtown: Taunton Press, 1998.
Hubbard, Frank. Three centuries of harpsichord making. Cambridge: Harvard University Press, 1967.
Marcuse, Sibyl. A survey of musical instruments. New York: Harper&Row, 1975.
Russell, Raymond. The harpsichord and clavichord. London: Faber and Faber, 1959.
Sachs, Curt. The history of musical instruments. New York: W.W. Norton & co., 1968.
Sadie, Stanley, ed. The new grove dictionary of musical instruments: early keyboard instruments. London: Macmillan press, 1989.
Seltzer, Leon E., ed. The Columbia Lippincott gazetteer of the world. Morningside Heights: Columbia University Press, 1952.
Weinreich, Gabriel. "The coupled motions of piano strings", in Scientific American, v. 240, no. 1, pg. 118.