J. S. Bach's "Well-Tempered Clavier": A New Approach .

2
Author(s): J. S. Bach and Peter Williams
Source: Early Music, Vol. 11, No. 3 (Jul., 1983), pp. 332-339
Published by: Oxford University Press
Stable URL: http://www.jstor.org/stable/3138020
Accessed: 05/10/2009 18:13

Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at
http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless
you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you
may use content in the JSTOR archive only for your personal, non-commercial use.

Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at
http://www.jstor.org/action/showPublisher?publisherCode=oup.

Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed
page of such transmission.

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact support@jstor.org.

Oxford University Press is collaborating with JSTOR to digitize, preserve and extend access to Early Music.

http://www.jstor.org
Peter Williams

J. S. Bach's Clavier
Well-tempered
A new approach 2

I Autograph manuscript of J. S. Bach's F major Prelude Bwv880.i('48',book 2) (London, British Library,Add.35021, f.8)

The Well-temperedClaviers 'method of composition'-
to quote from part 1 of this discussion (EM Jan 83
pp.46-52)-is a method based on the two main
problems that confront the composer of any piece of
music: how to plan the whole (strategy) and how to
compose the bar-by-bar detail (tactics). To generalize:
what was old fashioned about the '48' was its starting
point as a collection of preludes and fugues in various
keys; what was new was that the styles in which they
were composed embraced the newest idioms of the
day and in the process developed the idea both of what
a prelude is and what a fugue is far beyond the
achievements of any other composer. Sometimes the
prelude is built on a single motif; at other times it has
two or more sections, producing either a kind of
prelude-and-fugue within itself or a movement in

332 EARLY MUSIC JULY 1983

binary form as polished and highly organized as any
sonata first movement by Haydn. Both in its style and
in what may be called its weight, the prelude usually
differs from the fugue, often giving the appearance of
being the bigger.(This is also characteristicof some of
the organ music. I think it quite out of style to register
organ fugues, as many players register them, loud and
powerful, as if they were the main fare to which the
preludeserved only as horsd'oeuvre; in olderpraeludia-
that is, works in several sections-the fugue is rarelyif
ever the climax of the work as a whole.)
If one looks forwardfrom such earliercomposers as
Bohm or J. C. F. Fischer to book 1 .of the '48', rather
than backwardsfrom Beethoven, it becomes obvious
that the player ought to bring out these differences
between prelude and fugue and to forget the idea,
characteristic of later keyboard music, that pairs of
pieces have an integrated identity or an emotional
uniformity.Clearexamplesof such differences between
prelude and fugue are the E flat major Prelude and
Fugue of book 1, which are so different in length, in
form and even in numberof voices that ironing out the
differences can hardly seem feasible to anybody. But
such differences are characteristic of almost all the
'48', more than seems to be usually thought. A good
starting point for the player is to assume these
differences, until persuaded otherwise by careful
reasoning.
To judge by similar motifs found in the preludes of
Fischer and Gottlieb Muffat, the opening of the E flat
majorPreludeof book 1 is grandand spacious, and the
section that follows is in the traditionalalla brevestyle.
By 1720, Bach was already sufficiently a master of
such alla brevewriting to be able to mislead the ear as
to the number of beats in each bar-a detail that can
be brought out by the phrasing (ex.la):
Ex.l E flat major Prelude Bwv852.i ('48', book 1)
(a)
10
. A I 4 I t
i
j
r~~rr!
JI
?i j a
FlbFf Fr V
4 14
4 i? i? t
There then follows a third section combining the two
motifs heard so far (ex.1b). So systematic-even
doctrinaire--is the working out of this combination
that the prelude becomes very difficult for the fingers
to get around: not for the first or only time in his life
(b)
from tfirst section
9iJ r
ro r
from second section
I I
was Bach more concerned with the mind than the
fingers. I think this prelude one of the hardest pieces
to play in the entire harpsichordrepertoryif a constant
tempo is to be kept between the sections, which is
what this traditionalcounterpoint requires (the whole
idea of motivic combination of the kind here demon-
strated is that the motifs are combined at the same
tempo, with diminutions really diminishing the note-
values).
By comparison with this, the E flat major Fugue is
wonderfully light, though also difficult to play. Its
delicate lines, broken chords, pretty cadences and
open textures-all of which make it totally different
from the prelude- need to be made clear to the player
and listener. Aftersuch a prelude, then, one could well
play such a fugue in a rather affected and 'posed'
manner, making it charming, even capricious. It may
seem odd to think of a Bach fugue as something
capricious, but for a composer to allow his first codetta
(alreadythere at the end of the subject) to dominate a
fugue as it does here is exceptional, even-who
knows?-ironical (ex.2):
Ex.2 E flat major Fugue Bwv852.ii ('48', book 1)
II t JI
r i XbbbCC ^~~~~~~~codetta motif
codetta motif
g
My point here is that although in theory the tempo,
touch, registration, phrasing and general style of
playing are all open to the playerto decide, in practice
EARLY MUSIC JULY 1983 333
an awareness of what the composer was aiming at in
the contrast between these two movements narrows
down those decisions. And what Bach was aiming at
can be understood only by knowing what music of his
and of other composers led to this particularpiece-it
should be appreciatedby the (thinking)playerthat the
fugue is much more 'modern'than the prelude-and
then by playing it on an instrumentof the kind that one
may reasonably imagine him to have had.
To take another example. In book 1 both the D
minor and the G majorpreludes exploit a triplet figure
(ex.3):
Ex.3
(a) D minor Prelude Bwv851.i('48', book 1)
(b) G major Prelude BWV860.i('48', book 1)
24 zAm'^fe
^rtfff^~~~~~etc
Once you have learnt to get your fingers to play these
notes, neither prelude is particularly difficult or
uncomfortable to play: both fall very idiomatically
under the natural shape of the hand. They are also
lively and dashing, so it could be that these character-
istics will best be expressed by a very lively tempo,
even by one that a pianist might regard as too fast.
Indeed, it is possible to reason that the triplets invite a
tempo as fast as possible, for the harmony is slow
moving enough to make sense at such a tempo and the
figuration avoids the little awkwardnesses of many
another passage in the '48'. But quite apart from the
question of tempo, these triplet preludes offer the
thinking player good examples of the way the com-
poser's mind worked.Although the chief motif of both
preludes is a triplet figure, that in the G majorPrelude
begins on the beat, while that in the D minorbegins off
or rather before it (ex.4):
Ex.4
(a) D minor Prelude (b) G major Prelude
, L,. I I _I I on a fixed motif, a good starting point is to see the
Similarly. wh-Ile te lft hd in te D m
Similarly, while the left hand in the D minor Prelude
plays a repeated note, the left hand in the G major is
more broken up (though it too is a pointdorgue)(ex.5):
334 EARLY MUSIC JULY 1983
Ex.5
(a) D minor Prelude (b) G major Prelude
Now these differences between the two kinds of triplet
are so direct and clear- so obvious that they stare at
you from the page- as to suggest that the composer
was deliberatelygiving them distinct characters.Such
difference-within-similarityis exactly what one would
expect of a composer engaged in what certain German
writers earlier this century called Figurenlehre: that is,
composing by motifs or figurae, and thus creating
complete and even spacious movements out of tiny
melodic cells. Once it is realized that this is how most
of the preludes and fugues of the '48' were composed,
a whole world of interpretation is opened up. The
player will want to begin to make the composer's
technique clear through his playing. Forexample, the
G major Prelude's triplets begin with each beat, and
n2l"
this can be shown by markingthe groups thus, on the
beat. So what may appearto be an even, pianistic line
becomes broken up into clear triplets (ex.6):
Ex.6 G major Prelude
17
Everything in the background to the music of Bach
supportsthis interpretation.I am thinking particularly
of the kind of harpsichord(see part 1 of this article,EM
Jan 83 p.51) and the kinds of fingering technique
known to have been customary. Bach's keyboard
fingering is an involved subject deserving more than
just the most circumspect summaryhere; but one can
certainly say that in this relatively early music based
right-handfingers 234 as carryingmost weight and to
Pl
think not of a flexible, changing hand but of one that is
constantly shifting position, like a violinists. Perhaps
more important than individual fingering technique,
however, is the correct overall attitude to such music
as this. Forinstance, I do not believe that ars est celare
artemwas a dictum for musicians of Bach's period; on
the contrary, phrasing should show how a piece is
constructed, just as in Scarlattithe technical difficul-
ties ought not to be ironed out by fast fingers but
related directly to some musical effect. Smooth, even
playing would be an anachronism for the '48'; articu-
lating in an idiomatic way can show the original
process of the composition in the mind of the com-
poser, the kind of process he would introduce his
pupils to (quite probably by means of these very
pieces). In the case of the G major Prelude, I must
admit that I for one would be inclined to exaggerate
the effect, so that for example the final bars of the
prelude would be particularlycut up and thus, in my
opinion, become even more exciting.
I say 'in my opinion' because I cannot refer to any
objective source and say that even J. G. Walther,who
was writinghis book on figural composition soon after
he and Bach took up their new jobs in Weimar, says
anything so clear-cut as 'when you are playing a
prelude built on triplet figures, be careful to point out
the phrasing of them in the way you play'. Musical
evidence is never as direct as that, and instead we must
take what hints we can. Everyplayerhas an idea about
such pieces and in his enthusiasm for this or that
interpretationwill somewhat exaggerateit fromtime to
time;the question is, do the direct and indirect sources
justify the interpretation?In the case of phrasing or
articulation based on an understanding of figurae, I
think they do. As you play, show the figurae,that is,
how the music has been built.
Yet we all assume that some kind of absolute
governs music: that whatever your theory or your
technique, you could not play the G major Prelude
faster than a certain speed, nor phrase it and cut it up
beyond a certain level. But comparisons among the
pieces of the '48' are very useful and even crucial:
many a piece becomes clearer when we look at it in
relation to another. A good example is the pair of
preludes in C majorand C minor in book 1. How these
differ from each other in very specific ways, just as
they are very similar in others, can be made clear by
the player.Both preludes are built on a motif or pattern
of eight semiquavers: the first is divided between the
two hands, the second played by both hands at the
same time (ex.7):
Ex.7
(a) C major Prelude Bwv846.i ('48', book 1)
s .('y5 I Fugue of book 2 would not only be more ponderous if
(b) C minor Prelude Bwv847.i('48', book 1)
bbb' t X
Of course, one is major, the other minor; one is
predominantlyhigher up the keyboardthan the other;
and so on. But more subtly, the C major motif has its
eight notes grouped as 5 + 3 while the C minor has its
eight grouped as 4 + 4 (ex.8):
Ex.8
(a) C major Prelude (b) C minor Prelude
i Il l L IllI l
Such similar-but-differentcharactersof two pieces are
again hardly accidental: from the appearance on the
page there is nothing to say that these two preludes do
not have the same tempo, since both are concerned
with patternsof eight semiquavers in a 4/4 barof slow-
moving harmonies. Each could be seen as comple-
menting the other. What is more, these factors are
objective and can be seen by all; they are not a
question of my opinion. But to play them at the same
tempo means playing either the C majorfaster or the C
minor slower than many think desirable. Think of
them as a pair of models or exemplaof how a whole
movement can be woven from motifs of eight semi-
quavers; then it is helping the listener to understand
them if they are indeed played at the same tempo. They
are not concert pieces intended to provide a passive
listener with a 'fulfilling experience' (to use 20th-
century jargon). Ratherthey help him to ponder the
nature of music and how it behaves; concert per-
formance is, more or less, irrelevantto intimate music
of this kind.
One could add, nevertheless, that J. S. Bach was a
good deal more consistent and deliberate about his
notation than many players today seem to think.
Without giving a catalogue of his time signatures, it
can certainly be claimed that the 4/4 notation of these
two preludes says more than it might appear.Through-
out the '48', the time signatures are important and
deserve attention. Did not one of his pupils, J. P.
Kirnberger,note in his Kunstdes reinenSatzes,2 (1776-
9) that his teacher's signatures implied a style of
playing as well as tempo? For example, the F major
EARLY MUSIC JULY 1983 335
written in 6/8 instead of 6/16, but the passing notes
would be overemphasized and the piece as a whole
would be less light.
I do not think that it is in connection with tempos
for the two opening preludes that absolutes come in,
for it could not be claimed from objective evidence
that the C major Prelude interpreted as I suggest was
too fast. Only modern convention might make it seem
so to some players-convention, custom, familiarity,
memory of previous performances and so on. But
where absolutes might come in is in the matter of
phrasing. According to the theory of figural composi-
tion, the player could separate the 5 + 3 notes of the C
major motif, slurring them (as subtly as possible)
according to the groupings shown in ex.8. This would
point out clearly how the composer constructed his
motif. After all, only our own century has seen a
separation between player and composer: in the age of
which we speak, they were the same person (or, rather,
every player was a composer and vice versa). I do not
suppose there is a single harpsichordist today, either
student or professional, who has ever composed a
work based onfigurae; so how can he begin to see the
necessity of conveying the construction of a piece? Yet
playing the whole C major Prelude with its 5 + 3 too
marked would get on one's nerves; that is to say, it
would offend against an absolute idea of what is
musical and would produce something that is un-
musical. So it must be done subtly. But keeping a
balance between these various factors-absolutes,
historical knowledge, instrumental technique-is what
makes performance of old music far more complicated
than is still normally assumed. Or, to put it another
way, an ill-kept balance is the reason why most
performances of old music today are arbitrary, con-
jectural, opinionated (in the negative sense) and of
merely ephemeral value (if that). By emphasizing one
or other of these factors we can easily neglect the
others: stressing historical knowledge offends against
the absolutes and may produce a performance that will
irritate or bore, while stressing the absolutes demotes
historical knowledge and makes us forget that people
have not always been irritated and bored by the same
things. Besides, all this passes over a few questions
about who if anybody was expected to listen to the '48'
in the first place. These are very broad issues, and in
pursuing them we should do well to turn our attention
to 'style' and all that that means.
Of all the difficult words used by musicians, especially
336 EARLY MUSIC JULY 1983
those who like to think they work from instinct and
innate musicianship rather than from thought or
intellect, 'style' is one of the most difficult. To play in
'style' or to judge that another player is 'hopelessly out
of style' are phrases that constantly crop up in musical
discussion, and we all suppose ourselves to know
roughly what is meant by them. Perhaps in general
terms we do, for there is usually a measure of
agreement on what 'stylish' playing is, whatever the
music concerned. But for something as individualistic
as Bach's 48 Preludes and Fugues, each one of which is
a kind of checkerboard of the traditional and the brand
new, 'style' is a most complicated matter and one over
which, judging by the various performances that can
be heard of them, there is massive disagreement. That
this disagreement comes from players allowing them-
selves personal opinions too early on in the process of
becoming acquainted with the music I think very
likely; for there are several crucial elements in the total
picture of the '48' which are not open to opinion.
In the broadest sense, the '48' obviously supplies
pieces in different styles; so would any collection of
music, and the differences between say the F major
and F minor preludes of book 2 are clear to anybody
(ex.9):
Ex.9
(a) F major Prelude Bwv880.i('48', book 2)
(b) F minor Prelude wv881.i 48, book 2)
('48', book 2)
(b) F minor Prelude BWV881.i
'
rX r 1 r * j .X;
What is perhaps less clear is that these two pieces
represent two distinct historical styles of keyboard
music. The first is an old-fashioned piece of figural
composition based on a motif that produces a contin-
uous line, and hence the total fabric of the piece, from
first note to last (literally). This is exactly the kind of
music aimed at but of course never so well achieved by
the lesser keyboard composers of Bach's period and
area, organists like J. G. Walther, Georg B6hm, J. C. F.
Fischer,Vetter,Buttstedt,Kuhnau,Zachowand Pachel- phrases and so produce as neat an example of the
bel. This was in the days before composers or perform- Haydnesquequestion-and-answerphraseologyas any-
ers made much use of octave scales or smooth, even thing written by Haydn himself. After bars 9-19 see
lines; if a scale-like line does emerge in a prelude of furtherplay with these two ideas, a new theme enters,
the '48' it is in fact made up of many such motifs and built on old French harmonies whose Frenchness
should not be looked at as if it were a smooth scale of could be made clear by a bit of sustained playing in the
the kind Mozartmight give to the piano when it enters Couperin manner (ex.12):
during a concerto. I think it totally anachronistic to Ex.12 F minor Prelude
play the opening of the F major Prelude (illus.l)
smoothly; much more likely is something cut up and
broken, something clearly resulting from an endless
series of figurae(ex.10): etc
'
Ex.10 F major Prelude
/rbb 7 ; h. \ .r
r
rI

Although of course I cannot justify this interpretation

from any source, musical or theoretical, it does make
a/ ; ij-4 JJdoJ -~
sense in the context because the passage is preceded
and followed by clearer and more straightforward
patterns.
If such stylistic observations on this prelude are
correct, Bach is not concerned with the continuity
characteristic of Mozart or the dynamism typical of
Beethoven but with a learned composer's taste for
figural contrast. That is, he is playing with motifs and
Of course, notation will not show the subtleties of figures, contrasting one with another, running one
performancein such cases. Fortunately,very often the into another, creating fluency from subtle contrast. To
more modern or galant the style of a movement in the markthese contrasts in the course of performingsuch
'48', the more obvious the details of performance. a piece is to be true to his intentions, in principle at
Nobody would play the F minor Prelude and treat the least. He has alreadyprovidedthe piece with continuity
little phrases as anything but appoggiaturas:appoggiare and a sense of musical logic from phrase to phrase;
(s'appuyer)is to lean, and you lean on the first beats. and since there is no question of one phrase being
How you do it and to what extent may depend on louderthan another-that is a form of contrastfamiliar
circumstances, the instrument,your own taste. Butthe to us from later music-the player creates light and
extraordinarilyfertile brain of this composer never shade by subtler means of touch and articulation. If
rests with an idea, and afterthe initial appoggiaturasof the style of phrases suggests it, you play now strictlya
the F minor Prelude we pass immediately to a little tempo,now freely and more rubato.I call J. S. Bach a
figure (also very galant in shape) that has nothing 'learned composer';but of course he can write a pretty
leaning or affettuosoabout it (ex.11): tune as well, and the whole F minor Prelude is an
Ex.11 F minor Prelude inspired melody, delicate and charming. It is a parti-
4 cularly useful piece to take as a means of coming to
grips with his compositional procedure and his con-
stant allusions of'style', suggesting as it does the kind
of questions the playershould ask himself. Is the piece
& ?b
JI C m
JI traditional?Is itgalant?Is it antique? Is it Italianate, or
French?What kind of figure or motif is being used?
What kind of touch does it suggest?
Notice how these two phrases relate to each other The player takes hints from where he can. For
through their very differences: both are four-bar example, the composer provided the F major Prelude

EARLY MUSIC JULY 1983 337

with phrase marksthat confirm my impression that it
should be somewhat cut up (ex.10). Phrase marksare
not plentiful in the'48' (oranywhereelse in J. S. Bach's
keyboard music), and their occasional appearance,
when absolutely authentic as here, is always crucial;
organists,for example,would do well to ask themselves
why only one variation in the C minor Passacaglia
BWV582has phrase marks. In the F minor Prelude,
there are no phrase marksbecause the phrasingof the
appoggiaturas is obvious. This is conjecture on my
part, of course; but what is one to make of pieces in
which the notation offers no clue to the problem of
phrasing,however it is that one understandsthat word
'phrasing'?What about those pieces in which our
being awarethat the composer was engaging in figural
composition only serves to make the interpretationyet
more difficult? For example, what to make of the
phrasing of the F major Prelude of book 1 (ex.13)?By
'phrasing' I mean not merely the pianists notion of
legato segments into which a paragraphis divided, but
something broader:the markingof a phrase structure
not necessarily by legato but by timing, spacing,
phrase-clarifying,that is, by the way each subdivision
of the whole is made clear. In the case of this F major
Prelude, the left-hand line shows that the prelude's
first two bars are two long, single phrases; this is
confirmed by the prelude being notated in 12/8 and
the left-handphrasebeing playeddetache(not staccato).
However,the next two bars (3-4) cut up the phrase to
half that length, while in bars 5-6 the phrase length is
halved again. The whole opening section therefore,
has a kind of cumulative effect of diminishing phrase
lengths: two whole bars,then four half bars,then eight
quarterbars.
As a final example, take the E minor Fugue of book
1: a striking,even eccentric piece, with nothing that I
Ex.13 F major Prelude Bwv856.i('48', book 1)
(t b kT :
338 EARLY MUSIC JULY 1983
can hear in common with Fischer,Kuhnau,Buxtehude,
de Grigny, Frescobaldi or any other composer with
whom Bach was familiar by 1720. The nearest thing
seems to be those minor German composers who
worked with spinning chromatic fugue subjects-say
Bruhns,now alas known by so few extant compositions
but once perhaps of major importance to J. S. Bach.
The E minorFugue is aperpetuummobileof 41 bars, and
one can imagine the Czernys of a later generation
finding it a marvellous exercise for this-or-thatfinger-
ing, at a continuous speed, in two neat, sharply
defined parts (ex.14):
Ex.14 E minor Fugue Bwv855.ii('48', book 1)
0'.: ..c.....
... .Ol
But a regular Czernian interpretationcannot be right
and would be responsible for the fugue's rather low
reputationamong even quite enthusiastic '48'pianists.
Suppose instead we take as our startingpoint the 17th-
century harpsichordand then look at the notes afresh.
Do not put away the volumes of Buxtehude, because
they will teach you that in non-stop fugues it is vital to
understand what semiquaver figures the composer is
playing with. Thus the opening four semiquavers are a
brokenchord (ex.15a)while the next eight semiquavers
could be plain sostenuto/legato (ex.15b)or paired off
(ex.15c):
Ex.15 E minor Fugue
(a) (b)
detache _
.,
9- h
l"i -w
etc
'r - r'r r3 1r T
(c) hardly makes sense if played straight. Here a conven-
tional rallentando,such as one mighthave in Beethoven
lo 0 /, ^, I 1
or Schumann, is hopelessly anachronistic. But if the
final entry is phrased very markedly,then the cadence
becomes strong, logical and absolutely final (ex.18):
I preferthe last interpretation:first, because it accords
Ex.18 E minor Fugue
with phrase marks sometimes found in instrumental
parts(of roughly similarnature)in cantata movements; 38

second, because it brings out the chromatic fall of the K 'J IJ+",
I .I I
line more effectively, by alluding to the traditionof the
falling chromatic 4th, the passus duriusculusof the
theory books; third, because it contrasts with the
opening arpeggio. In arriving at this phrasing, one
would naturally and necessarily come to adopt a
slower tempo, because of the sheer amount of detail
being introduced into the semiquaver lines. The
second bar of the subject is even more promising, for
9sf:
1- i
though it looks plain enough, it actually incorporates
two three-note semitone phrases on different parts of
the beat (ex.16): It would be startling-perhaps too startling-if such
slurring were very marked. But apart from anything
Ex.16 E minor Fugue else, the '48' is a collection of at least 96 final
cadences, and there is room for many kinds of final.
f- Everycomposer and historically minded playerknows
C)
_
bo.homodowd - just how vital a good cadence is. The more one works
with this fugue, the more it can be seen to be a crucial
Again, that can be brought out by cutting up the clue to J.S. Bach'smind as he workedon the '48',for its
phrasing. The whole subject and its answer now use of motif and phrasing, even to its final cadence,
become a wonderful battle between phrases and should alert one to the many issues at stake. A thinking
differently articulated lines: a true 'counterpoint by performercan ask for no more.
articulation'(ex.17):
Ex.17 E minor Fugue

1 ~~~~~~~~~~~~~~~f; -
EARLY MUSIC
In my opinion, that is what this fugue is about. One can October 1983
then see why it has two conspicuously eccentric
details: first, the passage in octaves when the two JEAN-PHILIPPE RAMEAU(1683-1764)
hands come together, which is pretty odd in a fugue; Guesteditor:Neal Zaslaw
second, the amazing abrupt final cadence, which

EARLY MUSIC JULY 1983 339