(M.J.E. / Composer Listings / Sorabji / Opus Clavicembalisticum Errors)
List of possible errors in
by Kaikhosru Shapurji Sorabji
Compiled by Michael Edwards
This page is mainly a list of misprints I believe I have found in the score
for Sorabji's massive piano work Opus Clavicembalisticum. However, I
would like to prefix this with a broader discussion of some of the
considerations that I believe a possible new edition of this work will need to
take into account and correct.
This follows now; the list of specific errors I've found follow, preceded
by a few explanatory notes about this list.
Some thoughts on how a new edition might be prepared
Some time after I started preparing this list of errors and put it on my
web site, Jonathan Powell, a pianist who belongs to the
Sorabji mailing list I
am also on wrote to me and asked me about the new, corrected edition of Opus
Clavicembalisticum that he appeared to believe I was preparing. I am not
preparing such an edition, and have no plans to prepare it, and do not believe I
would be even remotely qualified to do this - so it was apparent that Jonathan
Powell had somehow got a mistaken idea about what I was doing. In fact,
preparing the information given on this web page is the only project I am
engaged in related to this composition.
However, this incident did remind me of the fact that, if a new edition
were to be prepared, I do have certain ideas about how this might be
The Sorabji Archive issues copies of the work, it is clear to me
that a new edition would be good thing to have: not only are there many, perhaps
thousands, of misprints in the score, but I also believe there are many other
inconsistent and awkward features of the score's orthography, which cannot
exactly be called wrong, but which do (it seems to me) make the score
needlessly difficult to read and play from. Some of my suggestions will be
aimed at removing some of these difficulties, while preserving as far as
possible the composer's intentions, even while changing the way those intentions
I do not intend to prepare an edition of the score - not only because I do
not, as mentioned above, feel qualified to do so, but also because I cannot
spare sufficient time for the thousands of hours' work I believe it might take
to do the job properly. However, if anyone else one day prepares an edition, I
would like to put on record some of my ideas on how this might be done. Perhaps
any future editor of the score will come upon my page (I hope my list of errors
will be of use in preparing a new edition), and he or she might like to read
these suggestions and consider whether they might be adopted, although of course
the final decision on this will be up to the editor. (And I have it on good
authority from The Sorabji Archive that no-one is currently (as of October,
2002) preparing a new edition of the work, or planning to do so - so talk of a
possible new edition is, at this stage, purely speculative.)
Clearly an edited version of the work will not be, and cannot claim to be,
an urtext - an exact reproduction of the score as the composer wrote it. So,
owing to the difficulties and ambiguities in the original score, a performing
edition will have to quite frankly alter what is written, although I would hope
it would spare no effort to reflect the composer's intentions as closely as
Ideally, such an edition would have end-notes in which every change made
will be described, and the reasons for it explained. This would include not
only errors of the sort I have put on this page, but also other changes to the
style of notation intended to make the score easier to read.
It is these possible changes to the notation I want to discuss briefly
here, although I am at this time not going to propose a detailed list of such
changes, instance by instance. That would be a large part of the work of
preparing a new edition - a task I said above I could not undertake. (If anyone
who reads this page is considering doing an edition, however, I would be
interested to discuss what I say on this web page with them. My e-mail address
is given below.)
1. Change of octave signs adopted by Sorabji
This work makes far more use of extreme registers of the piano than most
piano works do, and many passages seem to cover the entire keyboard all at once.
To cope with this, the composer invented a special notation to indicate octave
transpositions of certain passages.
To indicate that a passage in the bass clef was to be played an octave
lower, he inserted after the bass clef the upper-case letter "I", with an arrow
head beneath the "I" pointing downwards. (The "I" probably represents a Roman
figure 1, indicating one octave.) To indicate that a passage in the treble clef
is to be played an octave higher, he again used the "I", but this time put an
upward-pointing arrow head above the "I". For some very high passages, he also
used the double upper-case letter "II", with an upward-pointing arrow head,
indicating that the passage should be played two octaves higher. (He didn't use
a corresponding "II" for bass passages to be played two octaves lower.)
While this system sounds quite clear and logical on the face of it, I
believe it would be worth considering reverting to the more traditional notation
used by most composers: namely to use "8va" signs or "15ma" signs for one-octave
and two-octave transpositions respectively, followed by dotted lines extending
forwards for the actual duration of the passage affected by the sign.
My reason for suggesting this stems from the fact that Sorabji's notation
is given, and then, following that, there is no ongoing visual reminder of the
fact that an octave or two-octave transposition is in force. When you are
playing highly complex music often notated on three or four staves, with maybe
two or three of those written in treble clef, it can at times be difficult to
remember which (if any) of those treble staves are transposed, and which played
at actual pitch. I don't think performers need any additional difficulties to
cope with in performing this music which can be avoided, and the dotted lines
associated with the traditional notation will ease the mental load the performer
has to bear.
Indeed, where, in treble staves, one-octave and two-octave transpositions
occur at the same time or in quick succession, I would recommend using a
different style of dotted line to extend the 8va or 15ma signs. A good system
would be to use a heavier type-face for the dotted line extending from the 15ma
sign than for that extending from the 8va sign.
2. Reduction of number of staves in some systems
As mentioned above, the music is often notated on three or four staves,
sometimes even five. Simply reading so many staves places an extra mental
burden on the player. While I concede that frequently the texture of the music
is so complex as to make this necessary, and cramming the densely-textured
polyphonic lines into fewer staves would make them very cluttered and very
difficult to read (and to distinguish the parts), I do believe at times the
parts could be condensed into fewer staves without loss of clarity, and I
believe this would make the music easier to read and to learn at the piano.
My own piano technique is not even remotely close to being able to perform
this music myself, but I have occasionally tried sight-reading passages at the
piano - and there are passages I find almost unperformable solely because of the
number of staves in use, sometimes containing parts that interlock with one
another. I say "solely because of the number of staves" because, in some cases,
I feel I could probably sight-read certain passages were it not for the
difficulty of piecing together the music on different staves that interlock with
one another - something I cannot do in real time while actually sight-reading.
If I were to rewrite some of these passages on fewer staves, their complexity
might be sufficiently reduced as to make the music more easily sight-readable
(with my technique, in slow motion).
3. Redistributing notes on the staves
This could be related to the question of reducing the number of staves, but
often applies even in passages where I would not reduce the number of
Every note has to be played either by the right hand or the left hand; but,
frequently, the way the notes are distributed on the staves gives not the
slightest clue as to which notes belong to which hand; and working this out from
the notation as given is sometimes very difficult indeed.
Because of the highly varied and constantly changing textures, it would
probably be impossible to rigidly assign to a particular staff notes only played
by one particular hand. Nevertheless, I do feel that, by judicious adjustment
of the way notes are distributed to the staves, the notation could at least come
closer to making the assignment of notes to the hands clearer. Of course, one
cannot be absolute about this, because there will frequently be various ways of
assigning notes to the hands, and different performers may find different
schemes preferable - but I still feel the notation could correspond more closely
to the likely distribution of notes between the hands. Voices sometimes jump
from one staff to another quite arbitrarily, for no discernible reason, and I
can sometimes quite easily see ways in which this could be adjusted to conform
with this suggestion, without any loss of clarity - indeed, I believe, while
enhancing the clarity of the score.
I notice a peculiar mannerism Sorabji often uses in the score which I feel
is quite needlessly complicating, and it is one instance of what I am talking
about here. Imagine a passage written on three staves: the top staff is written
in the treble clef, the bottom staff in the bass clef; the middle staff is
currently empty, and may be either treble or bass clef (whichever clef is still
in force from a previous passage on that staff). A contrapuntal line low on the
treble staff may, for a few notes, descend just a few notes below middle C. At
the same time, a line high on the bass staff may ascend slightly above middle C
(temporarily crossing the part low on the treble clef just mentioned). What the
composer does sometimes is to briefly write both those parts on the middle staff
thus: the treble clef part, when it goes slightly lower, continues in bass clef
on the middle staff, and the higher notes of the bass clef part are written in
treble clef on that same middle staff. The treble clef notes which continue
from the bass-clef staff are low on the staff and have a treble clef placed
directly in front of them (positioned on the staff lower than a treble clef
would normally be), while higher up on the staff at the same time, the bass-clef
notes, continuing from the treble-clef staff, are notated, with a bass clef
directly in front of them (positioned on the staff higher than a bass clef would
normally be). Often the two clefs on the middle staff are directly lined up
vertically, and for several notes the staff is serving as both a treble and bass
The meaning of this is clear, in the sense that when you study it, there is
no doubt as to the meaning - but it strikes me as needlessly confusing notation,
and yet another avoidable obstacle to ease of reading the score. (The score is
already full enough of difficulties inherent in the music, without additional
ones resulting purely from ill-judged notational habits.) It also makes it very
difficult to work out how the notes on all the staves should be allotted to the
left and right hand at that point.
Idiosyncrasies of this sort would often be easily be avoided: for instance,
the passages of the sort I just mentioned often exceed the treble or bass staves
they start on only by a few notes, and keeping them on their respective staves
could be achieved by using only one or two leger lines, and the cross-overs to
the middle staff I just described could be completely and easily avoided.
Indeed, some idiosyncrasies of this sort look to me rather like ad-hoc
solutions to notational difficulties that arose, which Sorabji had not
anticipated when starting to write the system of staves and planning how to
distribute the notes on those staves. In effect, it looks as if he ran out of
space to accommodate a few notes here and there, and found a little bit of space
on a staff that was empty for the moment, and put the notes there, combining
treble and bass clefs in the way described above.
4. Regularization of stem directions
Normal notational practice specifies that when two voices share a staff,
the upper part should be written with note-stems pointing upwards, and the lower
part with stems pointing downwards. Sorabji frequently neglects this, and has
two parts on a staff both pointing up or down - often completely unnecessarily.
I would suggest this be corrected in a new edition.
If there are three or more voices on a staff, then of course two or more of
them have to point in the same direction, and this does not necessarily cause
difficulty in reading. But there might be times when I would want to reassign
the directions particular voices point. Due to changing vertical positions of
the voices with respect to each other, it might be necessary to change the stem
direction of a voice at a particular point. Sorabji sometimes does this, but I
sometimes feel the points at which such changes occur could be rearranged to
make the overall texture easier to read. Some needless changes of stem
direction in particular voices might also be removed.
5. Regularizing the notation of accidentals
The traditional rule for accidentals in music is that any accidental
continues to have force within the bar in which it occurs, in the same octave.
That is, if an F# occurs, all F's on the same line or space continue to be sharp
until the end of the bar, even if they don't have further sharps written in
front of them, unless cancelled by another sign. F's in higher or lower octaves
are not affected, and need their own sharp signs if they are to be sharp also.
(If the last note in the bar is an F, and it is tied into the next bar, then the
sharp sign continues to have force into the first note of the next bar, but no
This rule cannot be used in "Opus Clavicembalisticum" in any practical way,
because of the fact that this music is without metre. Although solid and dotted
bar-lines are used at long intervals to mark longer rhythmic or sectional
divisions, the music does not have discernible bars in the normal sense.
Without frequent bar-lines to act as delimiters, it is difficult to set a
precise and reasonably short scope that accidentals have force within. But, to
avoid ambiguity about accidentals, there must be a completely unambiguous
scope for accidentals.
To deal with this, Sorabji adopts a rule for accidentals that 20th-century
composers often use: he specifies that accidentals shall apply for one note only
- and he also adds the additional condition that when a note is immediately
repeated, the accidental continues to have force. This is presumably designed
to avoid the use of repeated accidental signs when a particular note is repeated
immediately several times. The exact rule given by Sorabji (to quote from the
score itself) goes thus: "NB. Accidentals hold good only for notes in front of
which they stand with the exception of repeated notes and tied notes."
I have problems with this rule, largely because the score shows strong
evidence that Sorabji did not actually follow this rule always in notating the
score. There are repeated patterns of notes which are obviously intended to be
played the same way several times; but when some of the notes in the pattern
have accidentals, these accidentals are not always repeated in succeeding
instances of the pattern. While the pattern as a whole is repeated, the
individual notes affected by accidentals are not immediately repeated, because
other notes come in between them, and the rule as stated by Sorabji seems
clearly to refer only to the immediate repetition of single notes.
It is my belief that the entire score needs to be revised with regard to
this rule, and many accidental signs need to be inserted to indicate necessary
inflections of notes and conform with the rule. I would also want to add
cautionary natural signs to indicate probable instances where an accidental is
not intended to continue to have force. While the rule might make these notes
natural anyway (because the note is not repeated immediately after the same note
with an accidental), such is the habit of musicians to consider accidentals to
have a scope longer than one note (or immediately repeated notes) that the
tendency to continue to read the note as inflected can be overpowering. As
things stand, Sorabji sometimes provides cautionary natural signs, and sometimes
doesn't, in a quite arbitrary pattern.
Related to this, I might also want to enharmonically respell some notes.
While in such chromatic music it may be unclear what enharmonic spelling would
be best or most correct anyway, there are instances where I feel the current
spelling is definitely wrong or awkward, and would benefit from respelling.
For instance, a passage written "A A# A" would probably be better written as
"A Bb A"; there are a few passages where double-flats or double-sharps are used
quite without logic of any sort; and, on the other hand, passages where natural
notes might, because of the context, be better written as double-accidentals.
(Some may disagree, but if I were sight-reading a passage at the piano, in a
context suggesting, for example, a tonality of G# minor, F# major, or similar
keys, and I came across a melodic line that went "G# G G#", I would actually
find this more difficult and disorienting to read than I would find
"G# Fx G#". Many of the wrong enharmonic spellings I feel Sorabji has used are
of this type, although usually more complexly than this simple example I made
Even if all these things are regularized, I still feel the sub-rule about
repeated notes not needing a separate accidental is problematical, and that
"repeated notes" needs to be defined much more closely. Here are some instances
of notes where it is not clear (without further definition) whether they are
"repeated notes" which are affected by the rule about accidentals. (I'm not
giving examples from the actual score, but making up the simplest examples
possible which illustrate the point being made.)
Suppose we have an F# on the lowest space of the treble staff; and now
consider certain notes that could follow it:
• The next note might come after a change to bass clef on the same staff,
and might be the F two leger lines above the staff. It doesn't have an
accidental in front of it, but it is the same F in pitch. Does this count as a
repetition of the first F? But it is written on a different position on the
staff. Such a note would need to be defined either inside or outside the scope
of "repeated notes".
• The next note is another F, in the same space on the same staff; but it
is in a different voice. Does it still count as a repeated note in spite of
this?; or do you count only repeated notes in the same voice?
• The next note in the same voice is the same F; but in between the two
F's, there is one or more other notes in another voice on the same staff. Is
the second F sharp or natural?
• The voice jumps to the next staff up, also a treble staff, and the next
note is the same F. Is it sharp or natural?
Or, for further examples, imagine that the first F# is the high F# three
leger lines above the treble staff:
• The note immediately following this is another F, written an octave
lower, but affected by an octave transposition sign. The note pitch is
repeated, but it is not at the same written pitch. Does it count as a repeated
note or not?
• The second F is also on three leger lines, but affected by an octave
transposition sign: it is written at the same pitch, but sounds in a higher
octave. Does this count as a repeated note?
Perhaps the repeated-note rule can be kept, but all doubtful cases such as
these should have the accidental sign repeated for them, or cautionary naturals
given if the sharp is not to continue. Definite rules should be established,
explained in notes, and followed, for how to deal with repeated notes in
Even so, providing definitions and rules to cover all these ambiguous
instances of "repeated" notes makes the definition of a repeated note more
difficult than one might expect. It might be better to remove the rule about
repeated notes altogether, and insist that all notes intended to be
sharp or flat, whether immediately repeated or not, be provided with their own
The attempt to exempt repeated notes (however they be defined) is
presumably intended to simplify notation by reducing the number of accidentals
that need to be notated; but it's beginning to look as if an unambiguous
definition of "repeated notes", because of the need to include or exclude
dubious cases such as those listed above, will be so complicated that it will in
itself place a mental burden on the performer that will vitiate the intention to
simplify matters. Simply making no exemption at all for repeated notes, and
requiring all inflected notes to have accidentals, may be the best way out of
this dilemma. Of course, a clear note should be placed at the start of the
score explaining the rule to be used, whichever one is actually adopted.
6. Clarifying rhythms
There are occasions where the rhythms are extremely complicated; but making
this worse is the fact that sometimes the rhythmic notation of passages is
unclear or incorrect, and the exact rhythmic arrangement of notes is sometimes
quite unclear. Or sometimes how one voice fits in rhythmically with others is
unclear. Or notes in two voices on different staves which cover the same
horizontal space within a system add up to a different total note value.
Some of these ambiguities seem to be caused simply by incorrect durations
being given to some notes, or vertical alignment of notes being incorrect; but,
in a few cases, the anomalies appear to be caused by the fact of triplet or
other "tuplet" signs being omitted from groups of notes that should have them.
In these cases, judiciously applied tuplet signs would remove the ambiguity,
but in extremely complex rhythms it is far from clear where the tuplet signs need
to be placed - the only thing that is clear is that they need to be applied
somewhere to make the various parts add up properly with each other.
Occasionally the vertical placement of notes in different parts, which should
indicate which notes are played simultaneously, is wrong or ambiguous, so that
does not always gives obvious clues about where the tuplet groups are meant to
be. (One or two instances of this are discussed in my list of errors given in
I believe that all these things need to be corrected and clarified in a
new, corrected edition of Opus Clavicembalisticum. In my list of errors which
now follows, I have suggested possible intepretations for a few passages of this
That covers the broad areas in which I believe changes could beneficially
be made in a new edition of the score. I might add to the above discussion as
and when further thoughts on this occur to me.
I will now move on to my detailed list of the errors I have so far found in
List of errors: Explanatory notes
This list of errors I've found in the score for Opus Clavicembalisticum by
Sorabji is intended to help in the event of a definitive edition one day being
produced of this work, if the editors are able to have access to this list.
Items were occasionally difficult to describe clearly in words - but if you
check them in the score itself as you read them, I think they should be
reasonably easy to follow. The list is designed to be read in conjunction with
the score itself, and will be meaningless to anyone who doesn't have a copy of
Please note: bars are numbered for each system of the score; dotted bar
lines are counted too, not just solid ones, and partial bars are counted as
Some of the anomalies or possible errors in this list result from grey
areas where it is not quite clear how one should interpret the rule given on the
first page: "Accidentals hold good only for notes in front of which they stand
with the exception of repeated notes and tied notes".
The main problems seem to stem from two questions: does this rule apply to
repeated notes which belong to different voices?; is just one intervening note
between two repeated notes enough to break the scope of an accidental? If you
take things literally and assume the answers to these questions to be "yes",
"yes", respectively, some of the following comments will show that this itself
will lead to ambiguities, or at least strong ground to wonder if a misprint is
present. In other words, the composer appears at times to assume the answers to
the two questions above to be "no". But I will discuss the details of this on a
case-by-case basis in the proper place. See the discussion above for a
consideration of the general factors which might affect one's interpretation of
My copy of the score is an authorized photocopy done by The Sorabji Archive
of a master copy held by them, and includes handwritten corrections in the
composer's own hand. Some of the anomalies I note below result from the
composer's handwriting either being illegible, or being faint - in my copy, at
In the following list, I've marked each item with either an asterisk or a
question mark: the asterisk marks items I feel fairly sure are mistakes; the
question mark denotes items where I do not feel sure, but it seems uncertain
enough that it should at least be looked at more closely by an expert to decide.
In referring to musical notes, I do not have musical symbols available for
use in this list: therefore I use the symbol "#" for sharps (as in F#), the
letter "b" for flats (as in Bb), "x" for double-sharps (as in Fx), and "bb" for
double-flats (as in Bbb). A note-name without these symbols is always to be
taken as natural (as in D - that is, D natural). If I need to emphasize that a
note is natural, as against sharp or flat, I will say something like "D-nat.",
since I cannot even approximate a natural sign with any characters available to
E-mail me about this music.
Click here if you need an
explanation for the strange appearance of the e-mail address which will appear
when you click on the e-mail link, or if you don't know what you need to do to
make the e-mail address work properly.
List of errors
* p. 5, system 2, bar 1:
Error: In the second group of semiquavers, the last note should be
Bb, not Eb.
Reason: By analogy with the identical groups before and after.
* p. 6, system 2, bar 2, 3rd stave:
Error: In the second-last group of semiquavers, 3rd chord, the E
should probably be natural.
Reason: Eb doesn't make sense enharmonically, and the context
seems to suggest parallel 2nd-inversion major triads.
* p. 6, system 4, bar 1, 2nd stave:
Error: Two crotchets before the end of the bar, this stave should
have a treble clef inserted.
Reason: That would continue the pattern of the left hand
paralleling the right an octave lower.
* p. 6, system 4, bar 2, 2nd stave:
Error: In the second-last group of semiquavers, the last note Ab
should probably be Bb.
Reason: Considering that this voice follows the one above in
parallel 15ths, Bb would obviously fit in better.
? p. 7, system 1, bar 1, 3rd stave:
Error: In the second-last crotchet (the last group of 4
semiquavers), first chord, the G possibly should be natural, yet taking the
accidentals rule literally it would be read as a G#.
Reason: I don't feel sure of this, actually - but G-nat. seems to
fit the context better harmonically, and it seems uncertain enough to be looked
? p. 7, system 3, bar 2, 3rd stave:
Error: Although the Bs are not immediately repeated, and
therefore, according to the composer's note at the beginning of the score, not
governed by the same accidental, I presume they are all meant to be Bb. But the
Bb's in the 2nd and 4th septuplet groups do not have flat signs.
Reason: I feel slightly less than certain of this one, actually,
but it seems possible that the left hand is intended to be underpinned by Bb-
major harmony, with other elements introduced at the end of each septuplet.
? p. 7, system 4, 2nd stave:
Error: Left hand, first group of semiquavers, 4th note: Although
the E has no accidental, and therefore would be taken as natural, it seems
possible to me that it may be intended as an Eb.
Reason: Eb-minor harmony immediately before this suggests it; but
it could just as easily be an E-nat. to fit in with the following A-minor
? p. 7, system 4, 2nd stave:
Error: 4th crotchet chord in left hand should possibly include D#,
Reason: The context suggests major and minor triads, not
diminished ones, and D# would give a G#-minor triad. But I feel less sure of
this one, because a diminished triad did occur in the right hand a bit earlier.
II. Preludio Corale
? p. 8, system 4, bar 1, 2nd stave:
Error: Possibly the F should be sharp.
Reason: This would agree with the left hand, and it seems likely
that if F-nat. had been intended, it would be have been written as E#. But
maybe not; I feel less certain about this possible error.
* p. 8, system 4, at very end:
Error: In my copy of the score, the very end of the system is
clipped off by a tiny fraction of an inch, and therefore it is uncertain whether
a bar-line should appear here or not, and whether it should be dotted or solid.
Reason: The music's texture changes obviously at this point, so a
bar-line seems possible; the previous system's bar-line is also partially
clipped in my copy, although still visible at the top.
* p. 9, system 2, bar 1, 3rd stave:
Error: The low Bb octave halfway through the bar should be a
minim, not a crotchet.
Reason: Obviously it takes a minim to reach the next octave to
which it is tied.
? p. 10, system 3, bar 1, 3rd stave:
Error: The composer has handwritten in a correction, but it is not
quite clear what he intended by it.
Reason: The handwritten addition looks like a sharp sign preceded
by a letter which might be F - yet it is placed right next to the E in the bass
staff, and there is no reason to think that E should be sharp. Perhaps the "F#"
is referring to the B-minor triad in the staff above (2nd staff).
* p. 10, system 4, 3rd stave:
Error: At the very end of the stave, a crotchet rest is obviously
Reason: Compare with right hand.
? p. 12, system 3, bar 2, 2nd stave:
Error: In the middle of the bar, above the gap between the 3rd and
4th quintuplet groups in the bass staff, there are three chords for r.h., l.h.,
and r.h. respectively, and they all include the note D. It is unclear how many
of them are intended to be flat and how many natural.
Reason: Db appears immediately before and is not cancelled; but
the lower D-nat. would not be flat under the accidental rule given by the
composer, which suggests it might be natural, which in turn suggests that some
of the higher Ds might also be natural, in spite of the accidental rule.
* p. 12, system 4, bar 1, 3rd stave:
Error: At the very beginning of the bar, the first demisemiquaver
Db looks as if it may be intended as Eb.
Reason: This would fit the whole-tone progression of the first few
notes in the left hand, and avoid the anomaly of C# followed by Db, which
doesn't make sense enharmonically. Compare with the similar situation half a
bar later, where the same notes occur, although this time the second note is
written as D#, not Eb.
? p. 13, system 3, bar 1, 3rd stave:
Error: In the first semiquaver triplet, the 2nd and 3rd notes
consist of the repeated chord B-G. Possibly it should include a D, completing
Reason: Looking later in the figuration, where the chord recurs it
always includes the D. In the second ascending group of figuration, based on
different harmony, the chord is complete each time, and does not leave out a
note in the first triplet.
? p. 13, system 3, bar 1:
Error: There is a series of semiquaver triplets based on 3rd-
inversion G-dom.-7 chords, rising octave by octave, written as if for the two
hands playing alternate triplets. (No notes are specifically assigned to either
hand, but it looks as if this was intended to be done, except possibly for the
final triplet, which would be inconveniently high for the left hand.) The chord
is split into two portions each time, with just the F being played first, then
the G-major triad. Sometimes the F is played once and the triad twice within
the triplet, and sometimes the F twice and the triad once. But these two
patterns alternate irregularly, as if one or more triplets may have been written
If you look at the second half of the same bar, you have a similar pattern
of ascending triplets, this time based on an Eb-minor-7th chord, with a similar
alternating pattern of either the single note or the chord being repeated, as
Reason: It looks possible at least that the two types of
triplet pattern just described are intended to alternate; but they don't do so
regularly in either half of the bar; moreover, the pattern, irregular as it is,
is not the same in the two halves of the bar. Perhaps it was so intended by the
composer; but irregularities of this sort at least make one suspect a misprint.
? p. 13, system 3, 3rd stave:
Error: In the last of the crotchet chords with fermatas, following
the rule about accidentals, the G should be sharp. But G-nat. seems likely to
have been intended.
Reason: The context seems to suggest G-nat., continuing the
pattern of major or minor triads (not augmented).
? p. 13, system 3, 3rd stave:
Error: In the last chord before the upward quaver run, left hand,
maybe the F should be sharp, as suggested by the rule about accidentals and
Reason: My doubt stems from the fact that, although the passage
seems to include cautionary repetitions of accidentals, it lacks one here.
Harmonically, either way seems about equally plausible; and there do not seem to
be any contrapuntal clues that would give preference one way or the other.
* p. 13, system 4, bar 2, 3rd stave:
Error: Probably the first note (in octaves) of the last group of
semiquavers should be Eb.
Reason: This avoids repeated notes, which do not occur elsewhere
in the downward run.
* p. 17, system 1, bar 1, 2nd stave:
Error: In the 3rd-last tremolo crotchet group, the D should be Db.
Reason: This would continue the pattern through that whole tremolo
passage. The tremolo obviously has the right hand continue what the left hand
was doing in the previous crotchet tremolo, and the D was Db there.
* p. 17, system 1, bar 1, 3rd stave:
Error: In the very last chord, perhaps a C# should be included,
added to the notes already there.
Reason: The preceding tremolo's notes all seem to be tied into the
chord, and only the C# is missing. The omission seems more likely to be
unintentional than deliberate. (The doubled A a third lower, however, should
not be taken for the missing C#'s notehead wrongly dropped into the A position.
In reality, one A is Ab and the other Ab, in spite of the lack of accidentals in
front of the doubled A, because the preceding tremolo has all its notes tied
into this chord, and that includes both an Ab and an A-nat.)
III. Fuga I
* p. 29, system 2, bar 1, 4th stave:
Error: In the lowest voice, the 3rd long note should have a
natural sign in the lower octave.
Reason: This would agree with the natural sign in the upper
octave, and a false relation doesn't seem to be intended here. This would make
the motif agree with the bass line at the very top of this same page, which
would eliminate the possibility of considering both notes of the octave to be
? p. 35, system 4, bar 2, 3rd stave:
Error: In the second semiquaver group, maybe the 2nd
note, B, should be a D of some kind (natural or sharp), or perhaps an E or F.
Reason: My only reason for suggesting this is that it would make
the shape of the semiquaver group better fit the surrounding ones. The B
appearing in two octaves strikes me as a slight anomaly.
V. Fuga a Due Soggetti
* p. 58, system 3 at end, all staves:
Error: This system is quite illegible owing to dozens of
handwritten markings entered by the composer correcting what is printed, and
rendering both the print and the handwritten additions mostly illegible.
Reason: Obvious: it cannot be read. Does a clearer, correct
version of this passage exist?
VI. Interludium Primum (Thema cum XLIX Variationibus)
* p. 74, system 2, bar 2, 3rd stave:
Error: Near the beginning of the bar, the flat sign is missing
from the top Bb.
Reason: The bottom B is Bb, and the whole context suggests that a
false relation is not intended. While it is possible that both notes of the
octave were intended to be B-nat. this seems unlikely, for two reasons: a
wrongly omitted flat sign seems more likely than a wrongly inserted one; and an
examination of the corresponding notes in further figures of the same type
suggests Bb because that would make a chromatic scale rising upwards through
successive recurrences of the figure.
* p. 86, system 3, bars 1-2, 2nd stave:
Error: Clearly there should be a treble clef just before the bar-
Reason: The following passage seems nonsense if read in the bass
clef. Further confirmation of this is given by a motif which continues from
there into the top stave, which is treble. If that isn't reason enough, the
passage would be unplayable if the bass clef continued, because the two hands
would seriously clash with each other physically.
* p. 95, system 3, bar 1, 2nd stave:
Error: In the group of three notes immediately after the second
treble clef and "I" symbol, perhaps the quaver should be a semiquaver.
Reason: This would fit in with the left-hand pattern in that
passage, and make all groups the same in duration. Also, only this way would
the left-hand part add up to the same note value as the right-hand tremolo it is
set against. (How on earth does one play those tremolos? They look unplayable
to me - and my hands are reasonably large, too.)
* p. 96, system 3, bar 1, 2nd stave:
Error: The very first chord is a cluster. It is unclear whether
the sharp and flat sign both appearing before the E are a misprint, or whether
it is intended to convey that both E# and Eb should be played. Or perhaps the
sharp sign really belongs to the F, although it appears in the E space.
Reason: In a case like this, where two E's occur with different
inflections, I would expect the note E to be written twice to make that clear,
and to put each accidental before its note - or at least to have both the
accidentals together before the duplicated note head.
VII. Cadenza I
* p. 101, system 1, 3rd stave:
Error: The tied chord near the end of that stave should be Bb Eb
G, not Bb D G as written.
Reason: It is tied from a chord in the 2nd stave, which is an Eb-
VIII. Fuga a Tre Soggetti
* p. 109, system 4, bar 3, 3rd stave:
Error: Almost half-way through the bar, the lowest note is written
as a 10th, F-A. It probably should be an octave, A-A.
Reason: An octave appears to be intended here. The F shown is a
note that doesn't even exist on most pianos, and a 10th interval makes no
contrapuntal sense here, either, in a passage where this voice is otherwise
composed of octaves (and the odd single note).
IX. Interludium Alterum (Toccata. Adagio. Passacaglia cum LXXXI Variationibus)
* p. 142, system 2, bar 1, 1st stave:
Error: The treble clef should have an "I" octave transposition
indicator after it.
Reason: This would maintain continuity from the same voice in the
preceding bar, and prevent a clash between the two hands in the same register.
* p. 142, system 2, bar 2, 2nd stave:
Error: In the third quintuplet, the first note (the lower of the
double note) should perhaps be E, not D.
Reason: This would maintain the octaves in the first note of each
quintuplet, which seem to form a voice of their own, even though they are not
given separate note stems.
* p. 142, system 2, bar 2, 2nd stave:
Error: In the 6th-last group of semiquavers (the groups taken
according to the beaming), the last semiquaver is a double note G-F#. Below
that there is a composer's correction which is not quite legible: I can't tell
whether he has written F# or G#.
Reason: Clearly this passage is ambiguous in intent. F# certainly
seems more likely, given the pattern of octaves alternating with unisons.
? p. 150, system 2, bar 2, 3rd stave:
Error: In the third of those chorale-like chords, perhaps all the
C's should be natural, even though one of them is by implication sharp (because
it is an immediately repetition of a C#). If it is really intended to be A
major in the left hand, and F major in the right, cautionary naturals in the
right hand would make this unambiguous.
Reason: The context seems to suggest to me a slight preference for
the left hand being A minor, but the edge is slight, and A major would seem
almost as likely. At any rate, my feeling is that the ambiguity is sufficient
to suggest a cautionary accidental, whichever it should be.
Passacaglia cum LXXXI Variationibus
* p. 157, system 2, bar 1, 2nd stave:
Error: At the very beginning of the system, the rhythmic
distribution is not clear. Are the three demisemiquavers a triplet, or just an
irregular group of three (non-triplet) demisemiquavers?
Reason: In neither case would the notes add up to the crotchet
value they are set against in the bottom voice. Possible readings would
include: make the three notes a semiquaver triplet; insert a demisemiquaver rest
above the crotchet in the left hand; insert a semiquaver rest and make the three
notes a demisemiquaver triplet.
* p. 157, system 3, bar 1, 3rd stave:
Error: About half-way along the system, there is a quaver rest and
three notes bracketed with the indication "4=3". But they are aligned wrongly
with respect to the right-hand notes.
Reason: The 1st note should come half-way between the r.h. B and
A, the 2nd half-way between the first two r.h. G#'s, and the 3rd half-way
between the r.h. B-G# and G-nat.
* p. 157, system 4, bar 1, 2nd stave:
Error: Of the first 4 B's, it is not quite clear how many of them
should be Bb.
Error: If the rule about accidentals governing repeated notes
applies to repeated notes in a different voice, then the first four should be
flat, counting the grace-note too - but then what about the 5th note, an octave
higher, which looks like it's meant to be the higher octave of the same note?
Taking it literally, you would have an augmented octave going from the grace-
note to the higher note; but cautionary accidentals would remove any doubt.
* p. 157, system 4:
Error: The rhythmic distribution of notes in this system is
clearly wrong, although it's difficult to say exactly what is wrong.
Discussion: The rhythmic distribution of this system is very
difficult to decipher, and this is the most complex area of contention I have so
far found in the score. I realize the rhythm is inherently intricate, but, even
allowing for that, I could only make sense of it by assuming that a couple of
triplet indications were omitted or wrongly placed, and that one septuplet
indication is wrong. My putative modifications to this passage would be, going
in order from left to right in the system:
1. The septuplet "rest-B-C-D-F-G-B" should be a sextuplet, with the rest
belonging to the previous rhythmic division.
2. C-C#-C# semiquavers: insert triplet sign.
3. In the lowest voice, starting underneath the G# dotted minim, the group
"quaver rest, demisemiquaver sextuplet (B-A-G-F-E-Eb, occupying the time of a
semiquaver)" should be bracketed and a triplet sign added. It matches the
right-hand group "semiquaver rest, B, B (two demisemiquavers), C-F#-G
(semiquaver)", which already has a triplet sign. Of course it might be possible
that neither of these groups should have a triplet; but in that case
of all notes from the G# dotted minim onwards don't add up to a dotted minim,
even though they are matched exactly against that dotted minim, and it would be
impossible to make the notes and rests add up properly without postulating all
sorts of additional triplet groups without any warrant. This seems the solution
that makes the fewest postulations.
One problem remains: that is the quintuplet which forms the last rest and
notes on the 2nd stave, which looks as if it has been dropped in there out of
the blue, and is difficult to relate to all the other notes (rhythmically, that
is, although it clearly belongs to the top voice in terms of melodic shape). I
think we have to assume that the top voice here temporarily splits into two
voices which then merge again after the quintuplet. This quintuplet occupies
the time of one quaver, and it therefore seems clear to me (after much study)
that the semiquaver rest which begins the quintuplet should be positioned a bit
earlier, so that it is lined up directly underneath the last A-C#-G in the top
stave. The semiquaver rest in the top stave coming after the A-C#-G should
therefore be positioned half-way between the Eb and E in the quintuplet.
I realize I'm being a bit speculative here: but these assumptions I've made
are the simplest way (making as few assumptions as possible) I can make rhythmic
sense of the passage. I wonder what someone more expert than myself thinks
* p. 167, system 1, bar 1, 2nd stave:
Error: In the 4th note of the voice in chords (containing the
passacaglia theme), probably the top F should be sharp.
Reason: It would agree with the lower octave in that voice, and
also conform to the passacaglia theme.
? p. 167, systems 2 and 3, 1st stave:
Error: There are two C#'s which go above the top note of the
Reason: The note in question is not found on a standard piano.
Could there be some mistake here? - or did Sorabji write the composition for a
piano with extended compass? (I believe they exist, although I've never seen
nor heard one.)
* p. 171, system 2, bar 2, 1st stave:
Error: There seem to be a number of anomalies here. The second
chord looks wrong to me somehow, although I can't quite say why I think that.
Is it possible the accidentals are before the wrong notes?
Discussion: I suppose the Cb throws me, where I would have
expected a B-nat. Also, the top E is not to be found on any pianos that I know
of (unless there are ones that go higher than the usual top C). Is it possible
the chord is intended to be C-Fb-Ab-C? This would be more easily stretchable in
a position where there would be little time to spread a 10th. This would also
seem to fit better with the first chord (C-Ab-C). The third chord (tied) is
C-C-E-Cb-E, which doesn't match the preceding, either. Perhaps it is meant to
My reasoning on all these points is tenuous; but the number of ambiguities
suggests that perhaps this entire passage needs to be checked with the
manuscript, and/or referred to Sorabji experts. It definitely seems wrong to
Also, unconnected with this, but in the same location, the dots seem to be
missing from the portion of the third chord whose stem points downwards. (It
seems reasonable to at least speculate that the entire chord is meant to have
the same duration.)
* p. 171, system 2, bar 2, 3rd stave:
Error: In the very first chord, the top note E is notated as a
minim, stemmed as the top note of a large chord, all of whose other notes have
dots. However, it and the E an octave below that should probably be stemmed
sepately from the rest of the chord, and turned into an octave crotchet. Also,
the lower E is marked as E#, almost certainly wrongly.
Reason: The E octaves in question are part of the passacaglia
theme, and the E at this point in the theme is a crotchet. This is the reason I
suggest the lower E should be divested of its sharp sign.
* p. 175, system 3, bar 1, 2nd stave:
Error: In the second-last quaver of the bar, there is an A in the
sextuplet which belongs to that voice that goes up and down all the time. It
should probably be a B.
Reason: Compare with the surrounding figures in that same voice.
* p. 176, system 2, bar 1, 1st stave:
Error: In the upward-running figuration at the beginning of the
system, there are notes that are repeated with a single note intervening. The
first occurrences of some of these notes have flat signs, and probably the
repetitions are intended to be flat also - but a strict interpretation of the
accidental rule would preclude this, and lead to results that are probably
Reason: Other repeated notes have the same inflection - namely,
natural - and there is no reason to think the flat notes should be treated any
differently. There is, in my opinion, ample evidence in this score that Sorabji
has not himself strictly obeyed the rule he gave about the scope of accidentals.
* p. 176, system 2, bar 1, 4th stave:
Error: There are missing sharp signs for some of the F's near the
beginning of the system.
Reason: I assume that all the F's at the beginning are meant to be
sharp; the context strongly suggests this. But a literal interpretation of the
accidentals rule (having a scope of one note only, except when a note is
repeated, which I presume to mean "immediately repeated") would make some of the
F's natural. Perhaps a few sharp signs need to be put in front of F#'s that are
not immediately repeated notes.
* p. 177, system 2, bar 2, 3rd stave:
Error: The top note of the minim chord should probably be A.
Reason: There are three reasons why Ab is almost certainly wrong:
Ab doesn't fit in with the passacaglia theme; it doesn't agree with the bottom
note of the same chord; and it jibs enharmonically with the following G# in the
same voice (on the next system): i.e., B Ab G# in a single voice just doesn't
make sense enharmonically.
* p. 180, system 1, left margin:
Error: The composer's handwritten correction is not quite legible
in my photocopy. If it is legible in the master copy, perhaps it needs to be
traced into it a bit more heavily. (I presume it is F#-A-nat., though; but F#-
A# seems possible.) (There are a number of other passages too where the
composer's corrections are illegible, or at least uncertain.)
* p. 192, system 2, 2nd stave:
Error: In the second hemidemisemiquaver run in the right hand, the
first chord should probably contain a D#, not an E#.
Reason: A G#-minor triad would fit in with the overall pattern of
1st-inversion minor triads.
* p. 193, system 1, 3rd stave:
Error: In the final hemidemisemiquaver run, in the right hand, the
4th chord should possibly be an A-major triad, thus requiring a C# (not C-nat.).
Reason: All the surrounding right-hand triads are also major.
X. Cadenza II
XI. Fuga a Quattro Soggetti
XII. Coda Stretta
* p. 245, system 1, bar 1, 3rd stave:
Error: There is a handwritten composer's correction near the
beginning of the bar which is not very clear (at least in my copy). If this is
legible in the master copy, perhaps it should be filled in a bit heavier for
Discussion: It does look as if the unclear handwritten correction
might be a treble clef and a bass clef immediately beneath it, indicating that
from that point the stave has shared clefs applying to different parts (a device
found elsewhere in this work also). However, if the stave is considered from
that point to have shared clefs, the passage might make sense only if these two
clefs are swapped in position. The minim and semibreve in the lowest voice
would then be read in the treble clef (E and E, although I also suggest the
minim should be an F, to agree with the lower octave on the next stave down);
and the upper voice (two voices for the last crotchet in the bar) would be read
in the bass clef.
However, if this interpretation is correct, it takes a bit of guesswork to
determine how long the stave has shared clefs, and at what point a single clef
is resumed. But I would suggest that the shared clefs end at the next bar-line,
after which that stave is treble clef only. (Treble clef from that point on
makes the best sense of the intermittent semiquavers that follow on that stave,
especially where those figures change momentarily to the next stave down.
* p. 245, system 1, bar 1, 3rd stave:
Error: The minim E near the beginning of the system, whose stem is
joined to a note on the stave below, should almost certainly be an F.
Reason: See the previous item, where I discuss my reasons for
assuming that this note is an E in treble clef, not a G in bass clef. If we
accept this (and nothing else makes any sense here), changing the E to F would
agree with the F in the lower stave that the note is joined to by a single stem.
Preserving the octave fits in contrapuntally with the preceding and following
octaves in the same voice.
* p. 249, system 1, beginning of 1st stave:
Error: The treble clef should be followed by the "I" which
indicates that the entire stave is played an octave higher.
Reason: This would fit in with the preceding and following staves,
and nowhere is there a "loco" or "I" to change the octave transposition status
of the stave. Moreover, the keyboard layout and melodic shape of motifs make
better sense by assuming the "I" sign to be present.
* p. 249, system 1, beginning of 1st stave:
Error: Probably the ties into the first chord are in error.
Reason: There is nothing for them to be tied from in the preceding
* p. 250, system 1, 4th stave:
Error: In the tied chord in square semibreves, in parentheses, the
A should be removed.
Reason: The chord is tied from a chord in the 3rd stave, which
does not include that A.
* p. 251, system 1, 3rd stave:
Error: The second chord on that stave is notated as F-A. Probably
the notes should be F#-A#.
Reason: The C#-major context seems to require this: C#-E#-G# to
F-nat.-A-nat. just doesn't make sense. Compare with the passage two
demisemiquavers later, which goes E-G#-B to A-C#. And similarly in many of the
following demisemiquaver groups.
* p. 251, system 2, bar 2, 5th stave:
Error: The minim chord seems to consist merely of the accumulated
notes tied over from the previous 3 chords. But if you look carefully, a low C
and F have been added (the C with a tie leading into it, even though that tie
comes from an E). It is not clear whether these are meant to be struck at that
point, or whether it is a wrongly printed ties-only chord. It seems to me that
the C and F should be deleted, and an E (the one-leger-line E) inserted.
Reason: The C and F make no sense harmonically, and there are no
fingers to spare to play them at that point anyway, owing to the left hand being
busy at that moment in other staves.
* p. 252, system 2, bar 1, 4th stave:
Error: A treble clef should be inserted immediately before the
first crotchet chord in that stave (and two tied chords before a bass clef
Reason: The harmonic sense and keyboard layout precludes the notes
being read in the bass clef.
* p. 252, system 2, bar 1, 3rd and 4th staves:
Error: About three quarters of the way through the bar, in the
last chord of the second group of quintuplet demisemiquavers, it seems to me
that the G's should be natural, making a G-major triad; but the repeated-
accidental rule would make the G's flat, which doesn't seem to make enharmonic
Reason: If they had been intended, I would have expected them to
be written as F#'s, making an ordinary B-minor triad.
? p. 252, system 2, bar 2, 5th stave:
Error: Perhaps the very last chord in the left hand should include
B, not B#.
Reason: The context seems to suggest an ending in G# minor (plus a
few other notes), not G# major. But of course this opinion is far from
conclusive: it could, I suppose, be a modern version of a "Tierce de Picardie".
What a stunning conclusion the work has!
This list was compiled from an earlier list I e-mailed to The Sorabji
Archive's curator, Alistair Hinton, in April, 1999. I have revised it slightly
to correct inaccuracies I myself incorporated in my reasoning; but the list is
far from complete, since I have discovered many further errors in the score
which I have not yet included. I hope to do this as time permits - so anyone
who happens on this list who is interested might like to come back every now and
then and check for further entries.
But I have to say that I think, in compiling the above list, I have begun a
task I cannot possibly hope to complete. I lack the expertise to finish it, and
ultimately I will probably lack the time, too. I believe the few hundred errors
I have found only scratch the surface, and that there are likely to be
thousands of misprints in this score. I could probably spend hundreds of hours
discovering and documenting errors in this score without exhausting them.
I hope this list, by some good fortune, comes to the notice of Sorabji
experts, or someone involved in the production of a definitive edition of Opus
Clavicembalisticum. If any such people read this, please e-mail me if you think
the above list will be of assistance in preparing the edition, or if you think,
on the above showing, I have the capacity to be of assistance in any other way -
perhaps doing more error-hunting, or perhaps even proof-reading of a new
engraving of this work.
While ultimately time will probably defeat any endeavours on my part to
compile a list even coming close to complete, the time I can dedicate to this
task is flexible, and could depend on how useful my work is. In other words, if
I had reliable advice that what I was doing was valuable, I could probably
dedicate more time to it than if I thought I was only duplicating discoveries
others more expert than myself have already made, or could make more easily or
quickly, which is a bit the feeling I currently have. (If any Sorabji experts
or editors of his music really feel I have nothing useful to contribute, and
that I am merely duplicating discoveries others have already made, I would
appreciate being told this so that I do not waste time searching the score for
But I would very much appreciate feedback or suggestions from anyone who
reads this who is expert in Sorabji, or involved in producing definitive
editions of his music.
Monday, 10 April, 2000.
E-mail me about this music.
Click here if you need an
explanation for the strange appearance of the e-mail address which will appear
when you click on the e-mail link, or if you don't know what you need to do to
make the e-mail address work properly.
NOTE ABOUT THE FUTURE OF THIS PAGE - Tuesday, 5 June, 2001:
This page is incomplete, by the very nature of what it attempts to do: to
add to a list of errors in a musical score progressively as I find them. Please
for an explanation about the setback that is responsible for my not completing
this page, and why it may never be completed, or why it may happen only very
slowly. Not that this makes any real difference to someone who would like to
see the page complete - but I feel I should at least explain the situation.
I will leave this page here, incomplete as it is, in case the information
already written is useful to some readers researching a topic they may have
difficulty finding information about on the Internet.
Given that the score for Opus Clavicembalisticum obviously contains
many hundreds or even thousands of errors, any page attempting to list them will
by its nature be an ongoing project with no definite end. Without doubt, the
above list of mistakes does no more than scratch the surface. While I do not
feel able at present to continue searching the score for errors, it is possible
I may later be able to devote more time to it. Feedback from musicians or
scholars which confirms that I am not doing this work uselessly in a wilderness,
but that it is, or will be, really useful, might motivate me to devote more
time to it. But at the moment this work, or anything I add to my web site, is
having to struggle against problems of depression and poor concentration that
seem to be afflicting me at present.
One last thing: I have managed, since I wrote most of this page, to acquire
a copy of John Ogdon's recording of this work, although my opportunities to
listen to it have been limited, owing to not possessing suitable equipment to
play the recording on. Before listening to this, I thought that hearing Odgon
play passages with ambiguous points in them would help me resolve them, help me
decide on a correct reading. In fact, I did not find it at all helpful in the
majority of cases. This was largely because the music is so complexly-textured,
so chromatic, and (often) so fast, that it was quite impossible when following
in the score while listening, to tell whether Ogdon played (for example) a G#
instead of a G. In the end I just have to rely on my own musical sense in
deciding which reading I think most likely to be correct. In many cases, I feel
completely confident of my claims, they seem so obvioius; but, in other cases, I
admit that the reading I've arrived at is little more than a guess.
Introduction - Front page, which leads to Contents
Web Site of Michael Edwards - Contents
Kaikhosru Shapurji Sorabji
Introduction to Kaikhosru Shapurji Sorabji
Errors in Opus Clavicembalisticum (this page)
This page created on Tuesday, 11 April, 2000;
last modified on Friday, 4 October, 2002.