WTC II/20 in A minor - Prelude http://www-

personal.umich.edu/~siglind/wtc-ii-20.html

II/20.1.1 The prelude-type

The prelude in A minor consists of two halves which are of equal length and, at least upon
first glance, similar layout. This may remind one of the dance movements from Baroque
suites. The texture is deceptive: while there is consistent two-part structure with frequent
imitation, the voices are only rarely truly independent of one another.

With regard to the material presented, there are two significant motives which together
dominate almost the entire piece. Both come with a steady accompaniment which swaps
voices as the respective motive does, thus impeding any contrapuntal juxtaposition of the
two motives themselves.

This prelude can thus be described as motivically determined, or more exactly, determined
by two-part motives in the context of quasi-independent texture.

II/20.1.2 The overall design of the prelude

Apart from the final bars of both halves (see bars 16 and 32), all bars of the prelude are
taken up entirely by motivic material, thus not leaving any room for cadential formulas. In
addition, an extraordinarily high degree of chromaticism in the leading part of each motive
weakens the sense of harmonic progression.

In the absence of strong harmonic progressions with definite closures, the layout of the
piece is conveyed only by means of analogies in structure, created by a similar order in the
presentation of the motives. In the attempt to gain a preliminary understanding of the main
traits of this prelude, it is therefore meaningful to state all analogies.

bar 1 recurs in bar 8 and } both transposed, voices inverted; however, the
connection of bars 1
bars 2-5d recur in bars 9-12d } to 2 is different from that in bars 8 to 9)
bars 5-8d recur in bars 13-16d (transposed, voices inverted)
bars 5-8d recur in bars 25-27m (transposed and varied)
bars 4 and 5 recur in bars 30 and 31 } (both transposed, voices inverted and
varied; bar 30: motive and
} accompaniment in inversion, bar 31 in original)
bars 17-19 roughly correspond with bars 21-23
bar 16 corresponds with bar 32

The conclusions to be drawn for a first estimate of the layout are twofold:

The first half of the prelude consists of two sections: bars 1-8 and bars 9-16. The second
half falls equally in two; section III is divided in itself, while section IV has traces of a
recapitulation. This gives the picture of a ternary form in the structure of |: A A':||: B A":|

The correspondence of bar 32 with bar 16 supports the impression of binary form, given
already by the equal length of the two halves and the repeat signs.

II/20.1.3 Practical considerations for performers

1
The prelude displays a large variety of note values (eighth-notes, sixteenth-notes and
thirty-second-notes as well as frequent syncopated eighth-notes) in combination with a high
degree of chromaticism. Almost all of the apparent jumps in the thematic material
represent either different melodic levels or ornamentation (for details see below). The basic
character of the piece can therefore be interpreted as rather calm. The tempo is best taken
fairly slow; the chromatic sixteenth-notes should be given time to unfold their entire
emotional content, and thirty-second-notes ought not to appear as virtuoso.

The adequate articulation is continuous legato. This is only interrupted by phrasing (details
below) and in the few cases of cadential octave jumps (see lower voice, bars 16, 28 and
32).

Phrasing is a matter that needs to be pondered with particular care. At first glance one
would think this to be straight-forward, as one-bar motives and the imitative texture are
very obvious. Closer analysis reveals that there is frequent overlapping of, on the one hand,
the harmonically required resolution of a tied note at the end of one motive and, on the
other hand, the initial note of the subsequent motive. Such overlapping (in which an
interruption of the sound flow would therefore be wrong!) occurs in almost every bar; in all
cases, the turning from the end of one phrase into the beginning of a new one can only be
expressed by dynamics (which must therefore be very eloquent) and occasionally by a
change of tone color. The following table gives an overview of the phrasing in this prelude,
referring to the point after the downbeats in each bar.

("," = slight interruption of sound flow; "-" = interruption unwarranted)

b 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
ars
U , - - , - , , - , , - - , - -
L , - , , - , - - , - - , , - , -

B 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32
ars
U , , , , - , - - - , - - - - - ,
L , , , , , , , , - , , , , , -

In addition, there is a "breath" after the middle beat in U: bars 10, 14, 15, 20, 23, 24 and in
L: bars 3, 6, 7, 12, 19, 27.

The prelude contains three ornaments; they occur in the final bars of the two halves
respectively - those two bars which, through the use of scalar passages and jumping
octaves, appear as virtuoso and thus distinctly different from the bars preceding them. In
bar 16, the lower voice is decorated with what appears in the score as a mordent and a trill.
The mordent, however, is followed by a typical suffix and thus given away as another note-
filling ornament. Both trills shake in sixty-fourth-notes (i.e. twice as fast as the fastest note
values in the piece, the regularly occurring thirty-second-notes). The one on the downbeat
of bar 16 is approached stepwise and therefore commences on the main note; it ends with
a suffix in slower values, as indicated by Bach. The other trill begins and ends regularly,
commencing on the upper note and concluding with a suffix in the speed of the shakes. As
both ornaments occur in the left hand and, what is more, against fast notes in the right
hand, they intensify the virtuoso impression which is created in the written-out notes and

2
further enhanced by the sudden appearance of the additional "middle voice". For easier
reference, the execution of this bar is spelled out in the following example (ex. 19):

The inverted mordent in bar 32, by contrast, is very simple: it consists only of the usual
three notes (A-G#-A) which should be played fast enough for the trill to end distinctly
before the B that launches the run in the lower voice.

3
II/20.1.4 What is happening in this prelude

The main motive of this prelude (M1) is introduced in bars 1-2d. Speaking precisely, the
lower-voice part of the motive commences on the second eighth-note; the downbeat note in
this bar (and similarly that in the upper voice of bar 17) serves as harmonic and metric
support but does not belong to the thematic material. It should therefore be played in a
neutral tone color and not too assertively.

M1a in the upper voice begins with an impetuous ascent (C to F). This is followed by a
descent in hidden two-part structure. The lower line, which is prevalent both because it
marks the four beats and because it is the one that leads the way at the end of the bar,
presents a falling chromatic line (C-B-Bb-A-G#) and ends in a diatonic closing gesture in
which two syncopations are followed by an accented falling fifth. The upper line within this
structure unfolds on and before the off-beat eighth-notes; taking up the F reached in the
initial ascent, this line also falls (F-E-D). The two falling steps are themselves preceded by
artificial leading notes (see D#-E and C#-D), thus adding further to the already high
chromatic content of the bar and giving the upper layer extra emotional emphasis.

M1b in the lower voice completes the picture with yet another falling chromatic line with
diatonic ending. The target note C, expected on the downbeat of bar 2, is delayed and thus
coincides with the beginning of the imitation of M1a (see above under phrasing). By welding
together the two phrases in such an inseparable way Bach achieves a particular stringency
in the progression from one motive to the other.

As the example shows, the three lines of M1 are conceived in parallel motion, presenting a
single gesture in what may be described as ornamented homophony (ex. 20):

The corresponding dynamic gesture is an overall diminuendo, triggered by a short but fairly
strong crescendo in the initial four-note ascent. While the decrease in tension is gradual and
unshaded in M1b, the leading M1a should be subtly shaded in such a way that the
predominant chromatic decline stands out against the backdrop of a softer secondary line
in the higher register.

M1 recurs without modifications in bars 2, 4, 5, 11, 13 and 25. With small variations it
further appears in bar 31 where the initial four-note ascent appears in parallel motion, and
in bars 8, 9, 26 and 30 where the ending of M1a is varied and that of M1b resolves
indirectly (see U bar 9: F(---)E, bar 31 E(---)F). Inversions of M1 occur in bars 17, 18 and 22,
while bar 21 contains a varied inversion and bar 20 sequences the first half of M1 with an
inverted and varied version of M1a. Very free modifications of M1 are heard in bars 19 and
23/24.

The dynamic shaping of the inversion poses a problem. The question whether one should
follow the natural temptation to express the general ascent in crescendo, or whether it is

4
preferable to retain the original tension design, is a very basic one. It not only influences
the shaping of the other motives but also determines whether the "B" section, which is
where most of the inverted statements occur, emerges as the dramatic climax of a prelude
in ternary form (if the rising lines are rendered in rising tension) or rather as an integrated
section in a prelude in binary form (if the original dynamic shaping is retained also for the
inversion of the motives).

M2 is first presented in bar 3. It spans half a bar, beginning after the downbeat and ending
on the middle beat or, in case of a tied note, on the note immediately after it (see bar 3, L:
D, U: F). The motive bears a certain relationship to M1; see the combination of sixteenth-
notes + thirty-second-notes in the leading line against eighth-notes in the accompanying
voice, and the appearance of a hidden chromatic line with diatonic ending inside the
leading voice (bar 3, U: A-G#-G-F). It differs from M1 in that the accompanying voice is
diatonic (see bar 3, L: A-B-C#-D) and moves in contrary motion to the leading voice (U:
descending, L: ascending).

The dynamic profile of M2 is therefore much more complex than that of M1. M2a begins
with a written-out inverted mordent which propels a sudden major-sixth jump; this gesture
is best expressed in an emphatic crescendo. The subsequent chromatic descent provides
the corresponding relaxation, interrupted on an even softer level (no accent!) by the escape
note in the higher register. Against this highly emotional dynamic curve sounds the straight
yet gentle crescendo of the ascending eighth-notes in M2b.

M2 is sequenced in the second half of the bar (with an adjustment at the beginning of M2a
which is due to the resolution of the tied note). The motive further recurs unvaried in bars
10 and 12. A slight variation in bar 27 (first half) is followed by two sequences (see bars
27m-28m), two partial sequences (see bars 28m-29d) and a free development (see bar 29-
30d). The most interesting modification occurs in bars 6/7 and 14/15. Here the half-bar
sequence gives way to a half-bar imitation. In addition, the two components exchange small
features: from the first imitation onwards (see bars 6m and 14m), M2a begins in simplified
rhythm with an octave jump while the originally simple accompaniment M2b appears now
embellished with the initial inverted-mordent figure from M2a.

With regard to an overall dynamic shaping, the foremost issue in this prelude seems to be
the contrast between, on the one hand, M1: one-bar segments, set in dependent two-part
texture, always decreasing, (except possibly for the inversions in the B section, as
mentioned above) and, on the other hand, M2: half-bar segments, set in polyphonically
independent texture, with contrapuntal dynamics, expressing overall increases or
decreases depending on the ascent or descent in the sequences.

Beyond these contrasts created by the material itself, the following large-scale
developments of tension take place:

bars 1-3 describe an inverted curve consisting of the initial decrease (bar 1), its imitation
which might end in an even softer shade, and the overall increase in bar 3
(achieved through a combination of the double increase of M2b, the strong
impulses at the outset of M2a and its sequence, and the ascending direction of
the sequence).

bars 4-7 resemble an S-curve. Bars 4/5 present the same repeated (and perhaps
progressing) decrease as bars 1/2. Bar 6, despite its change from a mere

5
 contrapuntal to an imitative texture, corresponds with bar 3 in re-building the lost
tension, but the additional two statements of M2 in bar 7 are conceived in
descending sequence and thus present a new relaxation.

bars 11-16 consist of a compound pattern, both with regard to melodic intensity and
tension development. The decrease of bar 11 is answered with an increase in bar
12; after a probably slightly stronger beginning in bar 13 (owing to the right-hand
position and higher pitch range of the leading M1a here) another diminuendo is
picked up in a second crescendo (bar 14); the descending sequence in bar 15
causes a small loss of tension which is, however, made up for in the virtuoso trills
(left hand) and higher octave jump (right hand) at the beginning of bar 16. The
descending runs then conclude the first half of the prelude with a final increase.
In terms of intensity, this final bar should be played with a much lighter touch
since it represents a virtuoso pattern very different from the high emotional
content of the melodic motives.

bars 17-19, 20, 21-24 can be played, as was mentioned earlier, in two ways:

(a) An interpretation which emphasizes the ternary form of the prelude - and
therefore renders these bars as the beginning of the contrasting section - will
shape them as a curve with a strong crescendo in bars 17/18 and an equally
strong diminuendo in bar 19. The first half of bar 20 will prolong this decrease,
while the second half is ambiguous and thus serves as a link. Bars 21/22 follow
the design of bars 17/18, while bars 23/24 trace the pitch line in its ascents and
descents.

(b) An interpretation which aims at underscoring the prelude's binary layout will
retain the dynamic decrease in M1. As the remote variations in bars 19, 23 and
24 are equally based on M1, the entire section thus presents itself as a chain of
retreating gestures on different levels of intensity - an interesting solution.

bars 25-29 recapitulate the order of material as presented in the first section, but not
its dynamic design. Consistent with the previous section, two options are open:

(a) The inversion of M1 can be taken to close the inverted curve already with bar
26. In this case, bars 27-29 would create a second, independent inverted curve.

(b) In an interpretation which retains the dynamic shape of M1 in its inversion,

the five bars appear as a single curve. Bars 25/26 follow the model of bars 1/2;
the bars presenting a variation of M2 are then conceived as an extended
decrease (bars 27/28) which is followed by an increase only in the developing bar
29 with its cadential-bass close.

bars 30-32 (a) describe, in the more virtuoso interpretation, a two-bar curve;
(b) within the second concept, they present a gentle relaxation.

6
WTC II/20 in A minor - Fugue
II/20.2.1 The subject

This subject spans two bars, with a substantial rest in its middle. Commencing on the
second beat of a four-four bar with the fifth degree of the A minor scale, it concludes on the
downbeat of bar 3 where it falls to the third degree.

Rhythm and pitch pattern are indeed intriguing in this phrase as both suggest two
subphrases related in symmetry.

Regarding the rhythmic pattern, the first half of the subject contains four quarter-notes
followed by a quarter-note rest, while the second half, viewed in reverse (i.e. beginning
from the final note and looking backwards) presents four eighth-notes and a eighth-note
rest.

Regarding the pitch pattern, one can discover an exclusive use of interval jumps (no
single step appears within the subject!) in almost complete symmetry: the first and last
jumps are formed by major thirds, the second and second-last intervals are perfect
fourths, and while the third leap represents a diminished seventh, the third-last leap
uses the complementing minor third. The point of connection between these two halves
of the phrase marks a striking gap, both rhythmically - the duration of the rest is longer
than any of the note values surrounding it - and in terms of pitch, with the interval that
expresses the highest degree of tension: the tritone (ex. 21):

Owing to this symmetrical correspondence of the two halves, we must regard the phrase as
an indivisible unit in which the rest, far from indicating phrasing, represents the moment of
highest tension. This view is supported by the harmonic background to the phrase, which
displays a simple cadence. The dominant (in Bach's harmonization of the subject
statements often appearing in the form of a ninth chord) falls on the downbeat of bar 2 and
resolves - through the rest and the jumping eighth-notes - only on the final note in bar 3
(ex. 22):

The dynamic design corresponding with these features consists of a powerful increase of
tension towards the G#, followed by a tension-sustaining rest and a gradual release
through the eighth-notes.

II/20.2.2 The statements of the subject

7
The subject appears eight times.

1 bars 1-3 L 5 bars 13-15 U

2 bars 3-5 M 6 bars 17-19 M
3 bars 6-8 U 7 bars 21-23 U
4 bars 9-11m L 8 bars 25-27m L

(ex. 23)

Apart from, on the one hand, the adjustment of the initial interval in the tonal answer and,
on the other hand, the enlarged jump in the major-mode statement (see bar 10: A-B =
minor seventh instead of diminished seventh), only small rhythmic modifications occur. In
bars 9, 17 and 25, the subject's initial note is shortened to a eighth-note. (At first glance,
the fourth and final statements seem to leave it open whether we are dealing with the
original of the subject, beginning with a third in which the gap is filled by a passing note, or
with the answer commencing with a step which enters a eighth-note late. Upon closer
inspection one discovers that the metric organization of the melodic lines preceding the
entry, together with the harmonic logic, determine the eighth-note on beat 4 as the end of
the preceding development.)

As will be shown below in more detail, all subject statements are separated by episodes.
Strettos, parallels, inversions or incomplete statements do not occur.

II/20.2.3 The counter-subjects

Bach invents two companions to the subject which appear with great regularity,
accompanying all except the initial and final entries.

CS1 is launched, after an intermittent partial sequence of the subject's tail (see also
II/20.2.4) from the sixth eighth-note of bar 3 onwards. Its conspicuous feature is a five-
note plunge which occurs four times in descending sequences. While the first and
second of these five-thirty-second-note figures are separated neatly by an interval
jump and a rhythmic gap, this interruption is then softened (see bar 4, beats 1 and 2,
where the gaps in both pitch and rhythm are fleshed out). After a fourth plunge, an
ornamented ascent concludes CS1 on a relaxed note.
CS1 is thus independent from the subject in every possible respect. The dynamic
equivalent to the compositional details consists in a four-fold increase of tension - with
the climax falling each time on the lowest note -followed by a decrease in the gentle
upward movement.
CS1 experiences only one small modification in its six statements: in bar 21, the two
initial falling groups are softened when two additional ascending notes precede the
first group, and a step-wise link completely fills the rhythmic gap between the first two
plunges.

CS2 has a relationship to the subject quite different from that of CS1. Entering almost a bar
after the subject and undetermined in its ending (compare L: bars 7/8 with M: bar 11,
L: bars 14/15, U: bars 18/19 and M: bars 22/23), its most characteristic segment are its

8
 first four eighth-notes. These complement the subject's rhythm by filling the mid-
phrase rest in a way which ingeniously combines the three outstanding features from
the subject.

(1) The interval pattern in the CS2 segment sounds like a free imitation of the
subject's first half (see bars 6/7: S = E-C-F-G#, CS2 imitating F-B-D-G# );

(2) the rhythm in the CS2 segment is an anticipation of the subject's second half
(eighth-notes);

(3) the distinctive interval, used twice in the CS2 segment, is the tritone (see F-B and
D-G#); this is the interval in the subject which connects the two symmetrical
halves of the phrase.

If one wanted to go even further, one could discover a parallel in sixths between the
CS2 segment and the central portion of CS1 (see bar 7, CS1 in M: D...G#...B...E; CS2 in
L: F-B-D-G#).

Rather than completely independent, CS2 is thus a supporting and enhancing

complement to both the subject and its primary companion. Dynamically this is
expressed in a strong increase throughout the four-eighth-note segment, followed by a
relaxation which parallels that in the subject.

The following sketch displays the combination of the subject with its two counter-subjects,
showing the interplay of their respective phrase structure and dynamic design

(ex. 24):

II/20.2.4 The episodes

The fugue contains eight subject-free passages. As the two counter-subjects habitually
enter later than the subject, overlapping of episode material with subject statements occurs
regularly. The table below gives only those details which form part of definite episode
material, and which are therefore necessary for the interpretation of the fugue.

E1 bar 3 (M: second to fifth eighth-note)

E2 bars 5-6 (M + L: until fifth eighth-note)
E3 bars 8-9 (M: until bar 10 downbeat or fifth eighth-note)
E4 bars 11m-13 (U: until bar 13d, followed by link; L: until fifth eighth-note)
E5 bars 15-17m (L: until fifth eighth-note)

9
E6 bars 19-21 (M + L: until fifth eighth-note)
E7 bars 23-25m (M: until bar 26d)
E8 bars 27m-28

The episodes are built from a combination of, on the one hand, motives derived from
subject and counter-subject and, on the other hand, independent motives.

Ms is a motive derived from the subject. It consists of four eighth-notes which constitute a
free sequence of the subject's second half. First introduced in the minute link between
the initial subject entry and the answer, Ms displays a particularly determined gesture
in the interval pattern F#-D#-E. The final major-seventh leap recurs only once (see bar
13); it is elsewhere softened to the falling third in which the subject ended (see e.g. U:
bars 5/6, 8/9).
The dynamic gesture of this motive is decreasing, and as it usually appears as either a
sequence or an imitation of the subject's second half, it can be regarded as a kind of
dynamic extension.

Mcs acts as a counter-part to the above-described motive in that it imitates or sequences

the ending of CS1. Its ornamented ascent is first heard in E4 (see L: bars 11/12, M: bar
12, U: bars 12/13). Having in itself resolving tendency, this motive is first established as
a further extension to the subject entry (see bars 11-13, all three voices). Only where it
occurs segregated from the Ms context and builds its own sequential pattern does this
motive acquire an active attitude (see L: bars 19-21m). Its use in the final episode and,
like an anticipation (or a fragment of an otherwise omitted CS1), against the ending of
the final entry, confirms the basic gesture of relaxation (see U, M, L: bars 27/28).

M1 is the first independent motive - independent insofar as neither its pitch nor its rhythm
derive immediately from the primary material (while other links might well exist, as we
shall see). This motive is introduced in E2 (see L: bar 5-5m, 5m-6d). It consists of an
ascending tetrachord in thirty-second-notes, followed after a rest by the descending
complement. The symmetrical design, the number of notes (4 + 4) and the tension-
sustaining rest bestow this motive with a certain structural relationship to the subject.
Like the subject, M1 describes a dynamic curve which finds its climax on the note
before the rest.
Apart from recurring in the same shape, M1 spawns two variations:

M1a is the contracted version of the motive; the pitch remains intact but the rhythm
appears deprived of its interrupting rest. This variant first occurs in E3 (see M, U, M:
bars 9/10).

M1b is a more remote relative. It features the initial four-note ascent followed, without
any rhythmic interruption, by a descent beginning above the climax and extended to an
eight-note group (see bar 16, U: E-B). This variant bridges the gap between motivic and
non-motivic material, i.e. between M1 and the scales which appear, in both ascending
and descending direction, in bars (13 U), 16 (U), 17 (L) and 25 (L, M, U), 26 (U, M).

M2 appears exclusively in E6 where it is heard in a pattern with imitation and varied

sequence. Its characteristic feature is the opening five-note descent reminiscent of CS1,
which is followed by a broken seventh chord that resolves on the next strong beat (see
U/M: bars 19/20). In the sequence, which is also imitated, the opening figure is varied
and the broken chord simplified to a mere seventh jump (see U/M: bars 20/21).

10
 Like M1, this motive also describes a complete dynamic curve. The climax here falls on
the root of the seventh chord or interval, i.e. on the lowest note of the motive.

The use of episode material establishes a number of relationships among the subject-free
passages. (Subdivisions within some episodes are made on the basis of material, not as a
result of harmonic closes.)

recurs in material compare with

E1 E4b Ms only bar 13 L bar 2 L
E2 E3a, E5a, E7b Ms + M1 bars 8-9d bars 5-6d,15-16d,24-
25d
E3b E5b M1a imit. bars 16/17 bars 9/10
E4a E8 Ms + Mcs bars 27/28 bars 11-13d

It was already mentioned that E6, due to its combination of M2 and the ascending
sequences of Mcs, is in a category apart. Outstanding for other reasons is E7 which consists
of three segments. E7a (see bars 23-24d) is conceived as an extension to the preceding
subject entry. This becomes evident when one discovers that all three voices perform one-
bar sequences of the previous bar (see in bars 22/23 and 23/24 the subject's second half in
U; the entire CS2 in M; and the final two-thirds of CS1 in L). E7b follows with a transposition
of E2, and E7c rounds this episode off with scales which are not even abandoned with the
advent of the subsequent subject entry, but substitute the counter-subjects.

II/20.2.5 Character, tempo, articulation, ornament realization

Both the powerful tension created in subject and counter-subject and the large variety of
rhythmic values used in the thematic material of the fugue indicate a rather calm basic
character. The tempo should be stately - slow enough to allow for the listener to perceive
the syncopations in CS1 - but not dragging. The movement from one quarter-note in the
subject to the next must express dramatic tension of the highest order. (Counting eighth-
note beats, desirable as it may appear in light of the rhythmic complexity in the piece, is
thus definitely detrimental to the character of the subject.) The relative tempo of the
prelude to the fugue may be simple and direct, equaling a quarter-note in the prelude with
a one quarter-note in the fugue. (Approximate metronome settings: 60 for all beats in
prelude and fugue.)

The articulation of this fugue constitutes an interesting case. While one generally expects
all melodic notes in calm character to be played legato, the jumping pattern of the subject
might cause some headache had Bach not given a very obvious hint - obvious at least to
the performers of his time, while it may seem a somewhat crooked argumentation for us.
This is how we can approach the case: If Bach assumed that performers of his time would
interpret the subject as a pattern of consecutive jumps which must be detached, there
would have been no need to write the wedges on the eighth-notes. He obviously trusted
that musicians would be guided, by the high degree of tension expressed in this particular
arrangement of intervals, to choose legato or at least hardly detached non legato. He
marked the eighth-notes because he wished them to be energetically separated.

As a result of this train of thought, the subject's four quarter-notes are very long - just
slightly detached - and all short notes values legato. The eighth-note pattern in the
subject's second half and, by extension, in Ms should be played non legato in a dramatic,

11
not softly detached way. In CS2, the articulation of the eighth-notes is at the discretion of
each interpreter, although it should be mentioned that playing them in a soft approach not
only emphasizes their relationship to the subject's beginning but also makes it easier for
the listener to distinguish this second counter-subject among the numerous patterns of
detached eighth-notes. (Performers who follow this suggestion would thus take the initial
four eighth-notes of CS2 very connected - hardly detached - but distinctly separate the
remaining ones where the jumps no longer carry melodic importance.) Finally, where only
one high-tension interval appears (as in U: bar 20) or where stepwise motion prevails in
eighth-notes (as in M: bar 21), unbroken legato is recommended.

The fugue contains one kind of ornament only: the trill at the end of CS1 and, consequently,
in Mcs. This trill begins on the upper neighboring note. Its shakes are usually taken in thirty-
second-notes. This is acceptable, but choosing twice the speed might be preferable for
anybody who can manage. The faster trills not only add drama and brilliance to the work,
but also avoid the problem of fusing ornamental notes with melodically essential ones (e.g.
the thirty-second-notes in CS1, M1 and M2). The impasse arising then is that one has a
choice between either a somewhat moderate trill motion (which seems inappropriate in so
suspenseful a composition) or great speed in the melodic material (which inevitably looses
many a listener, not to speak of performers). For those who can bring themselves to
overcome the psychological barrier of thinking in sixty-fourth-notes, the trill consists, after
the written-out two-thirty-second-note prefix, of twelve fast notes, followed by the suffix in
the rhythm notated by Bach.

There are two exceptions in the execution of the trill. In M: bar 27, the ornament begins on
the main note and thus with a thirty-second-note before launching its shakes. In bar 28, a
compound ornament beginning with an inverted mordent is indicated; the trill thus
commences, in full speed, with the lower neighbor note.

II/20/2/6 The design of the fugue

As has been shown above, Bach's design of the episodes divides the fugue into two halves.
The first half ends with a cadence in the home key on the downbeat of bar 13. The second
of the slightly uneven halves literally consists of two sections: section II (bars 13-211)
contains two subject statements in the upper and middle voices, and section III (bars 212-
28) follows with two statements in the upper and lower voices. The relationship among the
subject entries in this fugue (U: bars 13-15 see M: bars 3-5; L: bars 25-27 see L: bars 9-11)
stresses the design in two halves over that in three sections. The second half is only slightly
more extended, as a table of the corresponding episodes displays:

E1 E4b
E2 E3 E5 (E6, E7a) E7b (E7c)
E4a E8

For a sketch showing the design of the fugue in A minor see ex. 25.

12
II/20.2.7 The overall dynamic outline of the fugue

The first section begins with a majestic increase from the first to the second subject
statement. The following episode, conceived in descending sequences, reduces the level of
tension which is then picked up all the more forcefully by the third subject entry in the
upper voice. E3 brings another gradual decrease. Interpreters who regards the redundant
lower-voice statement as a "false fourth-voice entry" - a possible view - should aim at
slightly surpassing the dynamic level of the previous entry. Performers who do not share
this concept should render the entry in slightly lesser intensity, as is fitting for a redundant
one. The final episode of section I (E4) provides the dynamic closure.

Section II begins with an upper-voice entry in the highest possible register. The following
relaxing episode and much less exposed middle-voice statement suggest a gradual
decrease through this portion (bars 13-19d). E6, the final episode of this section, just like
that of the previous one, propels the tension upwards rather than preparing a close. Section
III, thus closely linked to the preceding section, commences with a reiteration of the upper-
voice statement in the highest register possible on eighteenth-century keyboards - thus
repeating if not surpassing the climax in bar 13. Again, the following episode together with
the lower-voice entry, which recalls features of the redundant entry from section I, present
a decrease in tension which is, this time, completed with a final cadence.

13