The Clavisimbalum from the Manuscript of Henri Arnaut de Zwolle, c.
1440
By Carl Rennoldson (revised April 2013)
I have recently made a reconstruction of Arnaut’s clavisimbalum. Although very little
information about keyboard instruments exists from the 15th century, in written form or extant
examples, the manuscript of Henri Arnaut de Zwolle, c.1440 in the Biblioteque National in Paris
survives as a unique and useful document from this period. My reconstruction is not intended as
the definitive representation of the clavisimbalum, there are far too many ifs and buts for that to
be possible, but actually making the pieces that Arnaut describes and adhering as closely as
possible to the directions he gives perhaps helps to get nearer to his ideas and make changes and
adjustments where details are not fully explained or are left unclear in the manuscript.
The section of the manuscript that describes the Clavisimbalum shows a plan view of the
instrument and separate keyboard with drawings and descriptions of four devices for “attacking
the strings”. For the purpose of the reconstruction, only the description of the clavisimbalum and
the first device were taken from Arnaut’s manuscript.
------------------------------------
The first of Arnaut’s “devices for attacking the strings”, “the first and best”, as he says, is without
doubt a plucking mechanism (See fig.1). The three dimensional effect of Arnaut’s method of
drawing shows that it is fitted with a pivoted tongue in a similar way to the modern jack. In his
description of the construction of the clavisimbalum he seems to be describing an instrument
specifically for the installation of this first device or forpex as he calls it. His descriptions for the
assembly of this mechanism into the register (i.e. “by tilting” and “passing through notches in the
wrestplank”), and its attachment to the tails of the keys clearly indicate that an instrument with
this device was in his mind when he was making the drawings and descriptions. The
clavisimbalum fitted with this plucking device might therefore be termed a harpsichord as we
would understand it today.
Of over 31 illustrations shown in an article by Edmund A. Bowles 1 entitled, “A Checklist of
Fifteenth-century Representations of Stringed Keyboard Instruments”, fourteen show instruments
of harpsichord shape. Although somewhat smaller in overall size to the harpsichord we would
recognise today these instruments have many similarities to the clavisimbalum described by
Arnaut in his manuscript. All of the illustrations in Bowles’ article show small portable
instruments. Several appear to be played in the vertical position perhaps strung around the players
neck. Most show only the outline shape and some superficial decoration but three of the plates,
19, 20 and 27a (see figs.2,3,4) display some interesting details with a number of similarities to
Arnaut’s instrument and his mechanisms. From the information given in the manuscript the
instrument in plate 19, dated 1465-8, bears the closest resemblance to the clavisimbalum. It
shows a harpsichord in plan view which is small in comparison to the angel figure. Although the
overall size of the instrument has probably been reduced to facilitate the composition of the
sculpture the case sides appear to be quite thick and there is a raised section across the area of the
registers. The bridge, which is clearly shown in this plate, runs parallel to the bentside and is
positioned exactly as Arnaut describes it, (i.e. the arc forming the bridge line uses the same centre
as the arc of the bentside). Most interestingly, in the register gap, a row of quite thick marks
show what might be taken to indicate the tops of the plucking mechanisms. The first device that
Arnaut describes, when installed in its playing position, has a similar appearance to this thick line
between the strings. The compass of the instrument works out to 16 natural keys, calculated by
using the number of visible keys as a proportion of the distance between the key blocks. This is
somewhat less than Arnaut’s instrument which has 21 naturals, although the sculptor has, again,
probably reduced the number of his keys in order to make the task of carving simpler. The
number of slots in the register gap is coincidentally twenty-one the same number of natural keys
as in the Arnaut instrument. Whilst it can be said that the instrument in the sculpture is not a
precise representation of a harpsichord, the craftsman that produced this figure would certainly
be working from an instrument that he had seen. Perhaps the strongest indication that this may be
an Arnaut type instrument is the quite definitely horizontal nature of the marks in the register
gap. Comparing Bowles’ plate 19 with plate 20 dated 1468, is most intriguing. (Discounting the
artists optical illusion in the casework) The lines in the register gap on this instrument project
vertically from the soundboard and stop under what appears to be a jack rail that we might
recognize today. An artist will generally give a fair
representation of what he sees and the lines are clearly vertical not horizontal as would be the
case with Arnaut’s first device. It is, therefore, not out of the question to suggest that these lines
may indicate a different type of plucking mechanism and that before the end of the 15th century a
form of jack had been developed, perhaps using only the tongue and quilling, as shown in
Arnaut’s drawing of the first device. The harpsichord shown in Bowles’ plate 27a and dated
1475-90 is less clear than the other two but here also the position of the jack rail is definitely
below the level of the top edge of the case with the faintest, but certain, vertical lines projecting
from the soundboard. See fig.4.
The clavicytherium c.1480 in the collection at the Royal
College of Music, London has a plucking mechanism
similar to the modern jack with the tongue enclosed in a
small slot at the end of a thin strip of wood. It is guided
through the strings by little pins and fitted to the
keylever with a bracket arrangement. This instrument is
very similar in shape to Arnaut’s Clavisimbalum and
within the dated area of the illustrations in Bowles'
article. It has been established that the Clavicytherium is
largely original so the jack arrangement might therefore
be assumed to be 15th century. This might suggest that
for a time both Arnaut’s forpex and some form of the
later type jack were in use in similar instruments,
before the later type became standardized. Considering
the early date of clavicytheria, (Paulirinus refers to them
in 1460)2 it is not impossible that the development of
this instrument with its simpler plucking mechanism
prompted the development of the later type of jack. By employing this new mechanism the
number of moving parts and pinned joints would be reduced thereby improving reliability of
action whilst the overall size of the instrument remained small.
Before attempting the reconstruction of the clavisimbalum it must be accepted that some aspects
of the design have to remain conjectural based on iconographical evidence and deductions from
Arnaut’s text. The manuscript does not contain detailed information on how to build a complete
instrument. The fundamentals are given such as the over all shape, the position of the bridge, the
number of keys and the general principal of the plucking mechanisms etc. But, for example, the
exact nature of the plucking device guiding system and the keyboard housing arrangements are
not fully or clearly explained.
In the reconstruction of the Arnaut instrument the maker is guided into a natural order of
assembly by the nature of the design. For this instrument, as in the later Italian tradition, the base
board is the starting point. Once this is made, (in Arnaut’s instrument the inferior fundum), the
way to proceed is to make the wrestplank and all the frame components and lay them on the
board during assembly. Once assembled, the central section of Arnaut's instrument containing
the base board (inferior fundus), wrestplank and soundboard (fundo superiori) is about 40 mm
deep and surprisingly light in weight. It feels something akin to holding the sound-box of a
stringed instrument. Most of the illustrations in Bowles’ article show instruments with thick
outer casework. Arnaut’s only mention of side thickness is that between the superior and inferior
fundi (the soundboard and baseboard). This he says is equal to the line OT and works out to
about 18mm, which seems quite heavy. However if the top inside edge is beveled in a similar
manner to the liners of some later instruments the strength is retained whilst freeing up the
soundboard (See fig. 5). He gives no instructions as to the construction of the outer case work,
which is obviously needed in order to give stiffness to the inner frame, other than to say that they
extend above and below the top and bottom fundi. Unlike his clavichord drawing where there is a
clear indication of the side thickness Arnaut gives no similar details about the sides of the
clavisimbalum.
The schematic nature of the manuscript drawing and the fact that the keyboard is drawn
separately from the plan of the instrument denies us any clue as to whether there may have been
a key well. In Bowles’ article all but four of the harpsichords are depicted with cheeks to either
side of the keyboard. However Arnaut’s clavichord drawings show the keyboards in position and
no cheek pieces in evidence, an arrangement which is confirmed by the clavichord illustrations in
the Bowles article.
The sections that Arnaut terms the, “lateral sides”, extend above the soundboard and below the
base board. His base board (inferior fundi) is therefore set above the level of the lateral sides
making the bottom an active component of the resonating chamber. This is a feature which may
have been carried over into later larger instruments as a number of other harpsichords are
constructed in this way. The early, c.1550 (possibly 1535 according to Barnes) Italian
harpsichord in the RCM collection has its bottom boards set 7mm above the lower edge of the
sides as does another anonymous Neapolitan harpsichord of 1537 in the Castello Sforzesco in
Milan and the anonymous Neapolitan harpsichord in the Museum of Fine Art, Boston. The
clavicytherium in the RCM collection also has a similar arrangement to the Arnaut harpsichord
with a sound box contained within the outer framework. The sides of the RCM clavicytherium
vary between 4.8 - 5.7mm in thickness. For the reconstruction, sides were made, as Arnaut says,
extending above and below the soundboard about 10mm thick following the shape shown in plate
20 by Bowles (see fig.5). In his manuscript Arnaut mentions a chest covering the area of the
mechanism. He states that this goes down between the strings on the soundboard side and down
to the semitones (sharps) on the keyboard side. However none of the illustrations given by
Bowles shows any instrument with exactly this feature. Arnaut distinguishes between the part of
the instrument behind what would be the jack rail area and that which extends from the housing
of the mechanism to the extremity of the keys. He indicates a different dimension for these two
areas. The plates, mentioned above, from the Bowles’ article show the sides of the instrument
raised at each end of the keyboard and register section. Considering the slightly later date of the
illustrations than the manuscript it is possible that this apparently structurally unnecessary section
is a remnant of the feature described by Arnaut, before the later modification to the jack rail that
the plates seem to show.
The rather strange arrangement of the keyboard with doglegs and miss alignments must be due to
faulty drawing technique rather than a representation of exactly how the keys were made (see
fig.6). There seems to be no practical reason why the keys should be formed in this way and it is
difficult to imagine that Arnaut could draw such an incongruous layout when the remainder of the
drawings in his manuscript are so clear. These apparent inaccuracies almost suggest that the
drawing was made by someone unaccustomed to the subtleties of marking out a keyboard, where,
to quote Hubbard, “a degree of tempering” is required. The keyboards of Arnaut’s clavichords
however are drawn accurately with each section of the key tails at right angles to the length of
the instrument and coinciding with the string and tangents position. These are then joined by
variously angled lines to the conventionally drawn keyboard. It seems inconceivable that Arnaut
or his assistant!! was incapable of doing the same for the clavisimbalum.
The first of Arnaut’s devices, or forpex as he calls it, is a rigid unit (See fig.1). The two main
components of this assembly, the forpex and a kapsel-like pusher, appear to be riveted together
where it might be expected that the joint would pivot. Used as shown in the manuscript, the base
of the pusher will rub on the keylever top when the key is depressed. The slot in the key tail
would also need to be elongated to allow for the quite large to and fro movement as the key is
pressed and released. Perhaps this is why Arnaut says it can be lined with brass to prevent wear.
For the reconstruction the joint between pusher and forpex was made to pivot with a brass pin.
(see fig.7) Whilst Arnaut gives details of the forpices and pivots for the keyboard he does not
say anything about guiding the keys at their tail end. The method of fitting the forpices into
the wrestplank guides them to a certain extent but there is still enough lateral movement to cause
the pluck to be unreliable. If there is no other method of guidance the key and forpex assembly
will not work with any precision. As with the later type of harpsichord it appears that it would
have been necessary for Arnaut's forpices and keys to have some sort of rack type guidance
system. In his description of the fourth device he says that the key must strike something hard to
allow the top piece to jump towards the strings. This suggests that he used some kind of guide
and stop system to provide accurate movement of the key levers. In the drawing of the fourth
device a reduced area is shown at the tail end of the key which may have been intended to locate
with a rack slot. For the reconstruction a rack of this type was made for the key tails and a similar
slotted arrangement fitted to the belly rail for the forpices. They would then be pivoted at one end
on the iron wire fitted to the wrest plank and guided at the other. For practical purposes the iron
wire that Arnaut speaks of must have been something more substantial than wire as we
understand it today. The forpices pivots around this “wire” which needs to be rigid enough to
support the device with some precision. For the reconstruction a 6mm diameter iron rod was
used, fitted into a small hollow made in the face of the wrestplank (see fig.8).
The layout on the underside of the wrestplank where the keys pivot is not made clear in Arnaut’s
manuscript. At one point he says that the fundum inferiorem goes up to the line BC and at
another that the wrest plank is thinner in that area. It cannot be both. If as his description says the
keys rock on a semicircular profile, “balance rail”, this is presumably fitted to a baseboard in the
keyboard area which is in turn fitted to the sides of the instrument. In the absence of a removable
key frame about 15mm would be enough space above the balance pins to enable easy removal of
the keylevers without need to thin the wrestplank. For the reconstruction this method was used
leaving a removable section of the baseboard beneath the key tails to allow fitting of the pins
through the feet of the forpices.
Arnaut gives no indication of the plucking material other than the triangular cornu mentioned on
the second device. This has been translated to mean horn. The shape given in the manuscript
would produce a rigid plectrum unlike the flexible traditional quill of the later jack. In other early
keyboard instruments there is evidence to suggest that metal plectra were fitted and it is possible
that the clavisimbalum was similarly voiced. The RCM clavicytherium has traces of brass plectra
reported by Hipkins in 18853. The Theeuwes claviorgan 1579 in the Victoria and Albert Museum
has two jacks with iron plectra. O’Brien4 says that these appear never to have been disturbed so
can be assumed to be original and that, “an instrument fitted thus would be better equipped to
compete with the louder organ”. For the reconstruction Delrin was used. It is proposed to try thin
brass plectra at a later date.
The position of the bridge in Arnaut’s drawing shows it close to the sides of the instrument
in the bass and treble and makes no allowance for the liner to support the soundboard. This would
have had a deadening effect on the tone as the bridge would effectively be held at each end. For
the reconstruction sufficient room was allowed around each end of the bridge for it to work more
efficiently. This was achieved by placing Arnauts’ plan within a slightly larger casework. Using
Arnaut’s geometrical instructions for marking out the plan of the clavisimbalum (See manuscript
text below) and, as the basis for the octave span, Stewart Pollens suggestion for the width of the
clavisimbalum5, (i.e. the width of 21 naturals from the keyboard of the 1581 Hans Ruckers
virginal, 502mm) gives an Arnaut ‘unit’ of 62.75mm. The resulting scale of the clavisimbalum is
short, c”= 235mm (Italian average, 350mm for iron, 270mm for brass), and the scaling plan is not
of Pythagorean proportion, the bass strings being considerably shortened. This would indicate a
higher pitch and after some experimenting it was found that the instrument needed to be about a
4th higher than a=415Hz to produce a satisfactory tone especially in the short bass strings.
Referring to the RCM clavicytherium in his article in the New Grove Denzil Wraight suggests
that these small instruments were about a 5th higher than the later 16th century harpsichords6.
Arnaut mentions either brass or iron strings so clearly the string material was given no particular
consideration, however the need for two different materials is demonstrated in the tables below.
Above c” iron had to be used as the tension in this section was higher than brass would allow.
Tuning was made a 4th higher than a’=440Hz. giving a=554Hz.
THE STRING LENGTH, FREQUENCY AND TENSION TABLES
The table below shows the note, length and frequency in Hz. for each string.Pitch : a’ = 554Hz.
Note Length Frequency Note Length Frequency Note Length Frequency
B 686 155.4 b 396 310.4 b 244 621.8
C 650 164.7 c' 380 329.4 c" 235 658.8
C# 615 174.5 c# 365 349.0 c# 225 697.9
d 586 184.9 d 350 369.7 d 216 739.4
e* 558 195.9 e* 336 391.7 e* 208 783.4
e 533 207.5 e 323 415.0 e 200 830.0
f 510 219.8 f 310 439.7 f 192 879.4
f# 488 232.9 f# 298 465.8 f# 185 931.6
g 468 246.8 g 286 493.5 g 177 987.0
g# 449 261.4 g# 275 522.9 g# 171 1046.0
a 431 277.0 a 265 554.0 a 164 1108.0
b* 412 293.4 b* 254 586.9
T= (FDL)2 ᴨk .
9.81
F=frequency, D=diameter in metres, L=length in metres, k=density of string material
Using the formula above to calculate the string tension, a graph of tension curve was produced.
English Brass and English Iron from Malcolm Rose were used for stringing. Iron, Average
density 7769 kg/m3. Brass, Average density 8536 kg/m3.
The table below shows the note, length and tension in kg. for each string. Change of material and
wire diameter shown in brackets. Iron is required above c”. B = brass, I = iron.
Size/Note Length Tension Size/Note Length Tension Size/Note Length Tension
Br(0.27)B 686 2.26 b 396 2.18 b 244 2.27
C 650 2.28 c 380 2.27 I(0.19)c" 235 2.15
C# 615 2.29 c# 365 2.35 c# 225 2.21
d 586 2.34 d 350 2.42 d 216 2.29
e* 558 2.38 B(0.21)e* 336 2.08 e* 208 2.38
Br(0.25)e 533 2.10 e 323 2.16 e 200 2.47
f 510 2.15 f 310 2.24 f 192 2.56
f# 488 2.21 f# 298 2.23 f# 185 2.67
g 468 2.28 g 286 2.40 g 177 2.74
g# 449 2.35 B(0.19)g# 275 2.04 g# 171 2.87
a 431 2.44 a 265 2.13 a 164 2.97
Br0.23)b* 412 2.11 b* 254 2.19
27.19Kg 26.69Kg 27.79Kg
CONCLUSION
When the reconstruction is played the bottom can clearly be felt to vibrate so if this was the
original intention it certainly works. Although many points in the Arnaut reconstruction remain
conjectural once the nature of the missing information had been established the mechanism
worked very well. Making the forpex in exactly the way that Arnaut depicts with a rigid,
apparently riveted, pusher rod (kapsel) would have made it impossible to fit it through the
register gap that he describes (one unit). Pivoting the pusher on the forpex allows it to drop
through the gap of one unit width, but only just. The general impression of the action suggests
that wear would very quickly become a problem. With so many joints and bearing surfaces the
instrument would soon develop a rattle. The main area of wear would most likely be the key tail
where the end of the pusher passes through. Arnaut says to line this with thin brass. This was
possibly a brass plate around the oblong slot for the feet of the forpexes, laid on top of the key it
would take the main force when the key is depressed. This area is the most unsatisfactory part of
the action. Arnaut states clearly that little cotter pin pass through the foot of the forpex beneath
the keys and this produces an imprecise connection. A better solution would be to pin the joint
through the key. This would integrate the three components, key, pusher and forpex making a
smoother connecting rod action eliminating any slack with it’s potential to rattle. The other,
pinned, joint and the rack guide works well but once the action is assembled there is no provision
for adjusting the height of the plectrum other than the back touch. This makes voicing an
awkward operation and accurate drilling and assembly essential in order to position the plectrum
in just the right place beneath the string.
Arnaut describes the keys as rocking on an iron wire against pins passing through the keys with a
semicircular piece beneath. This is very similar to the clavichord of c.1620 owned by Dr. Roger
Mirrey of London7. In this instrument the balance pins are also in a straight line but instead of
an iron wire twisted gut is used against the pins.
The light weight of the action was not enough on its own to return the keys and in order for them
to balance it was necessary to add a small lead weight to the tails of the naturals. In addition to
this the friction in the system with all the moving parts caused the action to catch up and jamb.
So, after fitting all the components as accurately as possible it was necessary to go through
everything again introducing a little more clearance. In addition all the touching surfaces were
rubbed with a soft lead pencil as a lubricant, this improved the whole action. In the manuscript
Arnaut says that the first device can also be made of brass which would remove the key balance
problem. All this may make the clavisimbalum sound ‘difficult’ but this is not the case. Once the
teething troubles were overcome the action was smooth and light. The instrument has a clear
voice, crisp and bright with plenty of volume. The buzz on the plectrums return without damping
does not intrude as one might have expected. Overall it makes a viable keyboard instrument well
able to hold its own with other early instruments.
THE MANUSCRIPT TEXT
For ease of reference during the reconstruction the following working outline was extracted from
the translation of the manuscript by Lewis Jones and the book, “The Early Pianoforte” by
Stewart Pollens.
The principal device for attacking the string:
The first jack system and the best. This is the jack which is suspended from the wrest plank by
means of an iron wire passing in the slits of the register, and its tail crosses the bottom and
extremity of the key by passing through a mortise, and it is pierced by a hole under the key and
we give here two diagrams of this jack, so that one can see the two faces and it should be made of
good wood or better , in brass.
The bent finger of the front part of the jack is engaged in a vertical slot formed in the register; an
iron wire crosses the finger and the slot (which acts as a guide to the jack) and allows it to pivot.
This method (a) guides the jack well and prevents lateral displacement. (b) when one removes the
pin for regulating, it prevents it falling into the housing of the mechanism.
Construction of the Clavisimbalum: (See fig.6 for Arnaut’s lettering system)
For the construction of the clavisimbalum trace the line BC of a length equal to the width we
would give to the instrument and divide the line into 8 equal parts. The line AB is equivalent to 4
of these units, the line CD to 13, the line DE to one and the line EF equally to one. Then trace a
straight line FA and having placed the point G at its center, then draw the straight line CG; next
on the line CG take 8 parts going from C to H [the line CG is first divided into 9 parts, Arnaut
means 8 of these parts] and trace the segment of a circle that passes through the three points
FHA, [by construction this produces an arc of 945mm radius] and starting from the center of this
circle describe the bridge EIK and the segments of the circle neighboring the bridge are not
equidistant, because the upper part of the bridge must be wider/thicker and higher than the lower.
And at the point where the line YA traced with scriber cuts the line CG will be the center of the
principal rose indicated by M. And starting from this center describe a circle the radius of which
is one unit and this circle will be the principal rose. And two thirds of the diameter of the
principal rose will be the diameter of the two other middle sized roses, as for the little roses one
can make them as one pleases and even if one wishes without any openwork decoration; and one
should place between the roses, and equidistantly bars to the number of 4; and these bars must be
stuck first to the soundboard that is after, by means of points, they have first been fixed in their
respective positions on the lateral sides. Likewise the bottom board will be extended up to the
line BC; as for the soundboard it will be extended only as far as the housing for the mechanism,
that is to say up to the line NO; NOPQ is thus the housing of the mechanism, PQCB the width of
the wrest plank; and the line RS will be the wire in the middle of the wrest plank , over which
will pass the strings; and in its lower part that is to say at the side of the line CB, the register will
be thinner so that the pivots can be placed there, likewise the line BS which is equal to half a unit
represents the distance between the two fundos, and the line BQ is the width or height of all the
lateral sides up to the housing of the mechanism; the line OS is the width or height of the lateral
sides after {depuis} the housing of the mechanism up to the extremity of the keys; and on the
adjacent part of the sides it is necessary to construct a square box the fore part of which goes
down between the strings, while the rear part descends as far as the semitones. As for the
thickness of the wrest plank this will be only the distance between the fundo inferiori and the
fundum superiorum. Likewise the thickness of the lateral sides situated between the two fundos
will be equal to the line OT. And the exterior lateral sides, in addition to the dimension that has
just been mentioned, will be extended below the fundum inferiorum, and this because the keys
will be situated below the fundum inferiorum. And the jacks having been put in place by a tilting
movement will be placed in their housing and they will be connected to the wrestplank by an
iron wire, and they will be suspended from this wire by a certain slot contrived for the purpose so
that one can take them out very easily. Likewise the line VX drawn beneath the fundum
inferiorem will be a wire of iron or of fairly thick brass, and the key will turn on this when
ascending and descending , and opposite this said line at the part of the fundi inferioris which
goes from the front to the housing of the mechanism, one will fix little pins which cross the keys
[these pins are the key lever balance pins], and then the board overlapping the lower part of the
keys will have a piece of semicircular profile stuck to it [this is the balance rail] so that the
movement of the keys is not hampered by the fundos. Next the keys will have oblong slits in
their extremities; the feet of the jacks will penetrate through these and through the fundum of the
housing of the mechanism, and these feet are pierced with a hole [the jack is then secured to the
key tail with a pin through this hole] and these oblong slits may be garnished with a thin sheet of
brass so that they do not wear in use.
----------------------------------------------------------------
Notes
1 Bowles, page 11.
2. Wraight page 180. Ref. to. “manuscript treatise of Paulus Paulirinus of Prague ( c1460)”
3 O’Brien, page 28.
4. ibid page 28.
5. Pollans, page 19.
6. Wraight, page 181 – 182.
7. Barnes, see bibliography.
BIBLIOGRAPHY
Barnes, John, Anonymous Fretted Clavichord c.1620. Drawing 1988. From Sheila Barnes,
Edinburgh.
Hubbard, Frank., Three Centuries of Harpsichord Making, Cambridge, 1965.
Jones, Lewis., Translation of the manuscript of “Henri Arnaut de Zwolle” c1440.
(unpublished document).
Le Cerf, G. And Labande, E. Les Traites d ‘Henri-Arnaut de Zwolle et divers anonymes, Paris
1932. Kassal, 1972.
O’Brien, G., Ruckers, A Harpsichord and Virginal Building Tradition, Cambridge, 1990.
Pollens, Stewart., The Early Pianoforte, Cambridge, 1995
Rose, M. and Law, D., A Handbook of Historical Stringing Practice for Keyboard
Instruments, 1671-1856. Malcolm Rose and David Law, 1991
Ripin, E. M., Editor, Keyboard Instruments, Studies in Keyboard Organology, 1500-1800.
Dover Publications Inc. 1971, 1977.
Wraight, D., Early Keyboard Instruments. Grove Musical Instrument Series, Macmillan,
1980. Contributions with other authors.
Sources of Bowles’ illustrations;
Plate 19; Manchester Cathedral, Carved Musician-Angel (wooden corbel at nave ceiling), Date.
1465-68.
Plate 20; Paris, Bibliotheque Nationale, MS fr. 331, fol, 145v. Guillaume Fillastre, Istoire de la
conqueste du noble et riche thoison d’or. Date. 1468.
Plate 27a; New York, Pierpont Morgan Library, MS M834, fol.25. Book of Hours (Fouquet
Workshop). Date 1475-90.
For more details of the clavisimbalum see my website
www.harpsi.com
The Clavisimbalum from the Manuscript of Henri Arnaut de Zwolle, c. 1440
An update by Carl Rennoldson
Soon after the appearance of the above article in FoMRHI Comm.1765 it was pointed out by Jeremy
Montagu (Bull.106. p.3) that the roof of Manchester Cathedral had been extensively reconstructed in
recent times, the implication being that some of the carvings may not be original fifteenth-century work,
therefore undermining the association of the instrument depicted in the carving (Bowles, plate 19) with the
clavisimbalum described and drawn by Arnaut de Zwolle in his manuscript. The history of the Cathedral,
certainly over the past 200 years, has been one of continual restoration and improvement. In 1815 an effort
to 'beautify' the cathedral resulted in much of the interior being 'damaged' by over-enthusiastic
improvements when fear that the roof might collapse prompted a full-scale restoration.* At this time,
undoubtedly, much new work would have been introduced. In 1882 the building was again extensively
renovated and in the twentieth century, work has involved restoration after war damage. However in 1815
the harpsichord would still have been a familiar keyboard instrument. With plenty of examples around of
how the instrument should appear, the question is; from where did the carver derive the information to put
in place this subterfuge for future generations?
Whilst it is possible that the carving is later work (the addition of restorers work is known to have been
included and encouraged in modern restorations, e.g. at York Minster and Windsor Castle), the exact
arrangement of the harpsichord at Manchester is so similar in appearance to the plucking mechanism and
bridge layout of the Arnaut reconstruction, and to other fifteenth-century illustrations reproduced by
Bowles, that it is difficult to imagine that a carver could just think it up, without having access to the
manuscript or other information about early instruments. It is of course also possible, if the carvings at
Manchester are new, that replicas of badly damaged or worm-eaten originals were made.
The unusual arrangement of the keyboard may not have been made quite clear in Comm. 1765. I was
referring not to the cutouts at each end of the keyboard (which were probably to allow the key tails to fit
between the wrestplank supports), but to the division of the keys. The slight offset is most clearly seen in
the treble. The point being made was that the layout, as shown in the manuscript, was unlikely to be how
the keyboard actually was, but rather results from an error in the drawing technique. In fact it is quite easy
to reproduce this arrangement. See the diagram below.
Arnaut’s key divisions are all of equal width
B C a
First key division aligned Misaligned keys
Lay out a keyboard using three parallel lines, one for the tail ends, one in front of the sharps and one in
front of the naturals. Mark off the naturals along the front section and mark the corresponding naturals and
sharps at the tails. Starting with the first bass key division aligned, as Arnaut does, draw lines equidistant
and at right angles from each tail division down to the bottom of the diagram to meet the naturals. The
discrepancy begins to appear immediately, in the same places as in the manuscript.
The discrepancy results from the problem of dividing up the octave. The octave C to B has seven naturals
of equal width. Section C to E has a total of five notes and F to B seven notes. The notes BC and EF are
divided by a line at right angles to the front of the keyboard. This makes the tails of the F to B section a
little narrower than the tails of the C to B section. As all the sharps are usually made the same width, for
the sake of appearance, the difference is absorbed by a slight variation in the width of the natural tails
between the sharps. This in turn means that the lines joining the backs of the keys to the tails cannot be
parallel. On the Arnaut drawing this method has not been followed, and all the key divisions are shown
parallel, resulting in the misalignment. This suggests that the misalignment was probably just a drawing
error and not an accurate depiction of a clavisimbalum keyboard.
*The History of Manchester Cathedral, www.dws.ndirect.co.uk/mc1.htm
