WTC II/18 in G# minor - Prelude

    from Siglind Bruhn
J.S. Bach's Well-Tempered Clavier
In-depth Analysis and Interpretation

II/18.1.1 The prelude-type

The predominant texture in this prelude is homophonic; motives in one voice are supported by an accompaniment pattern in the other. Within this setting, there is hardly any inversion of voices or contrapuntal interplay between the melodic parts of the various motives. Apart from very few bars which display some contrapuntal work or imitation, the laws of polyphony do not apply here, and the term "motive" has to be understood as denoting a homophonic device.

This prelude carries two verbal indications which draw attention. The terms "piano" and "forte" (see bars 3 and 5) are an unusual sight in Bach's Well-Tempered Clavier, and they leave many interpreters wondering, especially since they are not ever taken up in the remaining forty-five bars of the piece. One way to read these dynamic indications is as a hint that this piece was written for performance on a keyboard instrument with registers (such as the harpsichord); but it is also possible to understand the two terms as a simple instruction conveying that the composer wished bars 3/4 to sound as an echo of bars 1/2. This second interpretation, as the more basic one, should definitely be taken very seriously, regardless of the particular instrument on which the piece is performed. (The first interpretation entails a register plan throughout the piece in order to convey what might have been Bach's idea beyond the very scarce indication of bars 3 and 5. This demands specialized knowledge about early keyboard instruments.)


II/18.1.2 The overall design of the prelude

Each of the initial four bars is harmonically conceived as i-V-i. These cadences are, however, not of structural relevance outside the motives they underline. The same holds true for the next cadence in bar 5. It contains all the necessary steps for a modulation and concludes this harmonic transition with the steps V-I on the down-beat of bar 6, but the conspicuous silence in the upper voices thwarts a feeling of harmonic resolution. Only in bars 7/8 does a similar cadence gain momentum, both because it comes with the more convincing double stop on the downbeat instead of the rests with which the preceding three bars commenced, and because it establishes the dominant of G# minor, a secondary key the listener of 18th century music is trained to anticipate. Structurally the most pertinent harmonic close in the prelude's first half occurs in bar 15m where the A# major chord (V of D# major) confirms the secondary key. Similarly in the second half of the piece, the V7 in bar 40 prepares the recurrence of the tonic.

The prelude consists of four sections; the first and third can be sub-divided:

Ia bars 1-8d i-v (G# minor to D# minor)
Ib bars 8-16d v (D# minor to A# major/D# major)

bars 16-24


dominant confirmed, Picardy-third ending
IIIa bars 25-36d V-VII (D# major to F# major)
IIIb bars 36-41d VII-i (F# major to D# major seventh/G# major)
IV bars 41-50 i-i tonic confirmed

There are three structural correspondences:

bars 3/4 recur in bars 41/42 (same key, varied)
bars 8-11m recur in bars 36-39m (transposed a fifth higher)
bars 11m-13m recur in bars 34/35 (transposed a fifth higher)


II/18.1.3 Practical considerations for performers

Since the rhythmic pattern is simple, consisting primarily of eighth-notes and sixteenth-notes, and the pitch pattern includes a great number of large intervals, the basic character of this prelude can be identified as rather lively. The articulation that corresponds with this character comprises sixteenth-notes in light legato touch and eighth-notes in non legato. The fact that Bach has found it necessary to link the consecutive eighth-notes in bars 9 and 10 (similarly bars 37 and 38) with explicit slurs supports this interpretation. Had he envisaged legato eighth-notes as a norm in this prelude, there would have been no need to indicate slurs. In both cases, the slurs mark characteristic appoggiatura-resolution combinations, i.e. pairs which should under all circumstances be rendered legato but which, without the slurs, might have been somewhat difficult to identify in the complex harmonic setting of these bars. The only other instance in which slurs are indicated probably also serves to avoid confusion in the different groups (see bars 44/45 beat 2: appoggiatura-resolution, i.e. inseparable; beat 4: harmonically supported intervals, to be detached). More pairs of appoggiatura-resolution do, however, exist - unmarked! - in the right-hand part of bars 16 (B-A#), 24 (G#-Fx and Cx-D#), 40 (E-D# and G#-Fx), 43 (E-D#), 50 (A#-G# and C#-B). Appoggiaturas written in the form of grace-notes do not, of course, pose a problem of articulation. The tempo of the prelude is flowing; fast enough to allow the listener to perceive swiftly moving quarter-notes, but not so rushed as to concert the sixteenth-notes into mere virtuoso runs. The many appoggiaturas which are distinguishing features in the prelude must retain some of their "sighing" quality. Ornaments in this piece include the grace-notes and two trills. All grace-notes appear as eighth-notes preceding quarter-notes; the main note value is split into equal halves, with the appoggiatura on the beat and the resolution following on the off-beat eighth-note (see bars 2, 4, 17, 31, 42). The trills appear on sixteenth-notes which form part of written-out ornamental patterns (see U: bar 18, M: bar 19). They are therefore rendered as suffixless mordents, beginning on the upper neighbor note and comprising four sixty-fourth-notes.


II/18.1.4 What is happening in this prelude

Both the structure of the prelude and its dynamic layout are primarily influenced by the succession of motives. Orientation for a large-scale design is provided by several bass-note lines.

In tackling the latter issue first, one finds that the first three of the prelude's four sections are strung together by very consistent overall lines in the outer voices. In section I, the bass moves gingerly downwards (see from bar 5 onwards), with diatonic steps followed by more closely spaced chromatic steps over an entire octave, only to rise again in consecutive fourths. Similar processes, though even more pronounced, can be detected in section III where a more extended chromatic descent is allocated to the right hand. The second section is more static, displaying merely a few bass-note lines. Only in the fourth section are such lines of no consequence. The following example gives the relevant excerpt from the first three sections (ex. 60):

Let us now turn to the motives that characterize this prelude:


opens the piece with a two-bar unit. Accompanied by two metrically placed chords, a sixteenth-note line describe a double-curve in the treble (bar 1) before winding its way to the level of the lowest voice. There, the sixteenth-notes move in hidden two-part structure: a typical string of bass notes runs beneath a repeated pedal note. "New" treble voices present two groups of an upbeat + appoggiatura-resolution pattern in double thirds. The fact that the sixteenth-note run pretends to change voices results in the impression of three levels: there are the block chords, the sixteenth-notes, and a little melodic figure.
The dynamic shaping of this motive should take the illusion of these three levels into consideration. The block chords should sound fairly neutral in tone color, while the sixteenth-note line requires a certain degree of intensity, finding its climax on the leading-note Fx which captures the highest amount of tension. The two small melodic groups obviously contain two dynamic gestures, with an increase from the upbeat towards the appoggiatura and a subsequent decrease towards the resolution. Among the two groups, the first is slightly stronger in tension than the second, owing both to the nature of the chord (V7 versus a simple V), its position (inversion versus root position), and the melodic shape (ascent versus descent).
M1 recurs immediately (in piano as indicated by Bach); the melodic groups appear now in inverted voices. Variations of M1 can also be found in bars 16/17, 41/42 and 43/44; in all three cases, the accompanying block chords are replaced by melodic figures presenting a combination of the sixteenth-note component with the melodic component (see the appoggiatura-resolution in bars 16 and 43, and the written-out slur with a similar melodic gesture in bar 41).


is introduced in bars 5-7. In its accompaniment, it is related to M1 (see the left-hand figure which continues the hidden two-part structure established in the second bar of the previous motive). The melodic double notes are also taken over, together with their particular metric organization which avoids the strong beats. The double notes here describe what appears like an incomplete one-bar curve. Commencing with a written-out inverted mordent followed by a sudden jump, they embark on a descent which seems to prepare a conclusion on the next downbeat. This conclusion, however, is omitted twice and only appears at the end of the second sequence (where it is by now very much longed for), together with the harmonic resolution into the dominant key (see bar 8d).
Dynamically this motive may pose a trap. In sight-reading, one easily accents the sudden high notes (see bar 5 G#, bar 6 F#, bar 7 E#). But careful analysis reveals that these jumps represent escape notes, i.e. notes which are not part of the melodic substance, although they add extra charm to the line.
The basic thread to be shaped reads (for the upper part of the double notes): B-A#-B---A#-G-(A#). Within this line, the tension is higher at the outset (i.e. on beat 2) and relaxes towards the (twice omitted) final downbeat. On a larger scale, M2 and its two sequences also form a descending pattern and are therefore best rendered in diminuendo.
M2 recurs once - without its second sequence and, more importantly, without the final resolution (see bars 21/22). A free variation of the components continues through bar 23 and leads to the cadential close in bar 24.


consists of two contrapuntally conceived voices. The lower part (M3a) exactly spans a single bar (see bar 8, sequenced in bars 9 and 10); beginning with simple sixteenth-notes, it then breaks into jumps which describe two parallel lines descending mostly in chromatic steps. The right-hand part (M3b), by contrast, begins after beat 2; it contains an ornamented ascent which reaches into the next bar with an appoggiatura-resolution (see bars 8/9: E#-Fx-G#).
While the right-hand part displays a little crescendo during the ascent, with a climax on the appoggiatura and a relaxation in the resolution, there is a strong feeling of continuous decrease in the left-hand part; this extends throughout the first half of bar 11 which is composed as a partial sequence of the preceding bar.
M3 recurs, in inverted voices, in bars 36-38. It is followed by a longer extension here which leads right up to the end of this section in bar 40.

x The conclusion of section I is not so much a motive but a sequencing half-bar figure (see bar 11, second half, etc.) which counters the preceding protracted descent in the bass with a straight ascent (bar 11: D#-E#-Fx, bar 12: G#-A#-B#, C#-D#-E#, bar 13: F#-G#-A#, B), and the preceding gradual diminuendo with a much more forceful crescendo, before leading to the cadential close in bar 15.

In the course of the three sections which follow the statement of the main material, there are various interesting developments to be observed.

Section II contains an imitative interplay of two voices in the right hand which are remotely reminiscent of M2. These are accompanied by the pattern in hidden two-part structure as introduced in M1 (compare bars 18-20 left hand with bar 2 left hand).

Section IIIa is dedicated exclusively to the development of the main motive. The sixteenth-note figure from M1 appears first in its two-bar version with only slight variation (see right-hand part: bars 25/26 and 27/28); this is set against a remote relative of M3a in the left hand. The first half of the sixteenth-note figure then dominates two further bars (see bars 29-30), while the second half with its accompaniment pattern in hidden two-part structure is taken up by the lower part (see bar 31) where it is soon varied (bars 32/33). The melodic group of upbeat + appoggiatura-resolution is quoted once (see bar 31), already stripped of the ornament its second half, and the following bars bring free variations which carry reminiscences of M2. The section ends with a short quotation of the sequencing half-bar figure which had before prepared the close of section I (see bars 34/35) - here again with a powerful ascent in the bass.

Finally in the fourth section, M1 experiences a gradual liquidation: bars 44/45 still display a similar accompaniment and a variation of the upbeat + appoggiatura-resolution; bar 46 recalls the initial sixteenth-note figure very freely; and bars 47-49 take up a hidden two-part structure pattern together with the off-beat rhythms heard - differently - in M2. The prelude's final bar takes up the closing features of bar 24, thus insinuating a binary relationship between the two repeated halves which Bach has otherwise taken much care to blur.

The following table attempts to show the material in the four sections, together with the more conspicuous dynamic developments. (var = varied; dev't = development)

Section I



x + close
diminuendo crescendo
Section II bars
extension + close

Section III




Section IV bars 41-43
M1 var. +
extended 46-49
+ close


WTC II/18 in G# minor - Fugue

II/18.2.1 The subject

Beginning on the downbeat of a six-eight bar, the subject extends over four bars, ending on the downbeat of bar 5 where B, the third of the G# minor chord, concludes the first complete cadence.

In this subject, the phrase structure is perhaps the most striking feature, particularly since, as will be seen later, this structure serves as a model for several other components of the primary material in this fugue. What is so remarkable is that the second half of the subject is an almost exact sequence of the first half - an unusually simple pattern for the main thematic idea of a polyphonic composition. (The final note, in this light, seems almost like an aborted attempt to add yet more sequences.) This structure, in conjunction with the extraordinary rhythmic simplicity, might cause the lulling effect of an on-going perpetuum mobile if one subject statement were to follow another without interruption. Bach avoids this by separating, after the initial entry pair of subject and answer, all further statements by contrasting secondary material.

The rhythmic pattern invites the following four observations:

The subject itself consists exclusively of eighth-notes - an extremely simple rhythmic pattern.
Outside the subject, quarter-notes and dotted quarter-notes - note values which can be expected in any six-eight time - appear regularly; so far the rhythm could still be described as simple.
Apart from the even patterns, syncopations also become a frequent feature. There are first those long tied notes which can be identified as the beginning of large-scale closing-formulas (see e.g. bars 5/6, 7/8). Besides these, several rhythmically more decisive syncopations occur. Metric accents shift from the middle beat to the second (see e.g. bars 9, 10) or to the third eighth-note (see e.g. bars 17, 18) and from the downbeat to the fifth (see e.g. bars 45, 46) or to the sixth eighth-note (see e.g. bars 15/16, 17/18).
Finally, there are even sixteenth-notes: only scarcely on the first two pages of the score (bars 39 and 62) but more frequently from bar 71 onwards.

The pitch pattern is equally ambiguous. The string of consecutive jumps at the very beginning of the subject (see e.g. in bars 1/2 A#-D#-D#-G#) contrasts with the high percentage of chromaticism throughout the fugue. (For details see below, the description of the different components of the material.)

The harmonic background implied in the unaccompanied subject is very simple; it consists of a simple progression interspersed, at the beginning of each subphrase, with ornamenting steps (ex. 61):

Yet Bach's choices for the actual harmonization of this subject in the course of the fugue cover a wide range of alternatives. Apart from the final V-i there is not one chord in the phrase which is not at some instance in the fugue substituted by another cadential step. This gives the subject an uncanny versatility.

The dynamic shaping of the subject brings up the question of the relative melodic value of the eighth-notes: are all eighth-notes on the same level of intensity, or can some be seen as secondary? While this question certainly enters way into the sphere of individual interpretation, it is important that it be answered in detail as it influences many aspects, including that of conveying the phrase structure.

A possible concept is shown below. One can regard (a) the sudden upward jumps in bars 1 and 3 (fifth eighth-note) as escape notes and (b) the smooth wavy lines in bars 2 and 4 as ornaments of a simpler background. In this option, the underlying melodic line is perceived as consisting of a pattern which allows both a singing style and a subtle phrasing before the sequence. The dynamic development then encompasses two soft curves, the second of which is slightly more outreaching than the first.

(ex. 62)



II/18.2.2 The statements of the subject

The main subject - which later turns out to be one of two subjects in this fugue - appears in twelve statements.

1. bars 1- 5 U 7. bars 55- 59 L
2. bars 5- 9 M 8. bars 97-101 L
3. bars 13-17 L 9. bars 103-107 U
4. bars 19-23 M 10. bars 111-115 M
5. bars 33-37 L 11. bars 125-129 M
6. bars 45-49 U 12. bars 135-139 U

(ex. 63)

The subject receives a real answer and suffers no modification whatsoever.

From bar 61 onward, Bach introduces a second subject. It begins on the third eighth-note of the bar with a five-step chromatic descent, followed by a slightly more straight-forward ascent back to the initial pitch, and concluded by a keynote / leading-note / keynote formula in the traditional rhythmic pattern.

Subject 2 thus contrasts in several ways with subject 1:

Its rhythmic pattern is varied. The outstanding rhythmic features are the short - long, short - long of the chromatic segment and the dotted notes with syncopation in the closing-formula segment. Furthermore, the ornament on the second-last note adds faster movement.
Its pitch pattern is restricted to very small intervals: only one whole-tone step in an overwhelming surrounding of semitones.
The phrase is inseparable, containing neither sequences nor other features which would permit subphrasing.

What subject 2 shares with subject 1 is the four-bar length and the fact that all entries are separated by episodes.

Despite its entry sixty bars into the fugue, subject 2 appears almost as frequently as subject 1: altogether nine times.

1. bars 61-65 U 6. bars 103-107 M
2. bars 66-70 M 7. bars 111-115 U
3. bars 71-75 L 8. bars 125-128 L
4. bars 79-83 U 9. bars 135-139 M
5. bars 97-101 M      

Two of the statements of subject 2 are slightly changed in shape: in bar 83, the final resolution is reached only after a suspension and subsequent ornamentation of the leading-note, whereas in bar 128 subject 2 already breaks off on this leading-note.


II/18.2.3 The counter-subjects

Both subjects are accompanied by their own companions - but both companions prove to be not very faithful. Two further counter-subjects to subject 1 recur only once and in considerably altered shape; they thus exert no real impact on the polyphonic design of the fugue.


is introduced against the second entry of S1 (see bar 5 B# to bar 9d D#). Its structure displays a striking similarity to the subject it accompanies: it consists of two subphrases, the second of which is a sequence of the first. Each subphrase features the keynote / leading-note / keynote formula (which Bach also uses in subject 2), together with a chromatic upbeat. The pitch pattern of CS1 thus encompasses only semitones.
Dynamic shaping in this counter-subject is not problematic. In each subphrase, the eighth-note upbeat prepares the climax which falls on the syncopation and is resolved throughout the formula.
CS1 recurs only once (see bars 19-23).


is presented together with S2 (see L bars 61-65). Like its leader it is unphrased. Apart from one syncopation with subsequent written-out inverted mordent, CS2 features only eighth-notes and a not very distinct melodic pattern.
This counter-subject recurs twice: in bars 66-70, where the beginning is completely changed, it only imitates the model from the last eighth-note of bar 67 onwards; in bars 71-75, by contrast, only the first bar is modified in such a way that the characteristic syncopation followed by an inverted mordent now appears twice instead of just once.

A highly intriguing fact about these two main counter-subjects of the G# minor fugue is that CS1 is very obviously related to subject 2 and, upon closer inspection, CS2 is equally related to subject 1. Note, besides the coincidences in pitch, the similarities of the rhythmic pattern: mainly short - long, short - long in CS1 and S2 versus predominant eighth-note motion in CS2 and S1. The following sketches point out the likeness.

(ex. 64a)

(ex. 64b)

One thus discovers a fugue with two main thematic ideas. At the outset of the composition, one dominates and is accompanied by the other; later, the second idea dominates and is accompanied by a variant of the first, until they eventually meet each other in their full individual shapes from bar 97 onwards.

The two subordinate counter-subjects of S1 appear only in bars 13-17 and 55-59. The variation is, however, so significant, that it is preferable to denote particular components of the two lines rather than to speak of two melodic entities. This procedure also serves to draw the performer's attention - once again - to the fact of shared essential characteristics.

The first feature to be recognized is the keynote / leading-note / keynote formula. It appears twice (see bars 13/14: M and bars 14/15: U; bars 55/56: M and bars 56/57: M). The typical rhythmic structure is retained but partly diminished and metrically shifted. This feature is reminiscent of CS1 and S2.
Secondly, there is a strong element of simple eighth-note motion (compare bars 13/14 U with bars 55/56 U and bars 14/15 M with bars 56/57 U). This trait recalls S1 and CS2.



II/18.2.4 The episodes

The fugue encompasses fifteen subject-free passages.

E1 bars 9-12 E9 bars 75-79d
E2 bars 17-18 E10 bars 83-96
E3 bars 23-32 E11 bars 101-102
E4 bars 37-44 E12 bars 107-110
E5 bars 49-54 E13 bars 115-124
E6 bars 59-61d E14 bars 129-134
E7 bars 65-66d E15 bars 139-143
E8 bars 70-71d    

Among the material used in the episodes are components from the subjects and counter-subjects, several independent motives as well as traditional closing-formulas. Yet the line between motives derived from primary material and those invented exclusively for use in episodes is sometimes blurred, particularly where later variations of originally independent motives end up resembling segments of a subject or counter-subject. Moreover, several of the motives are related among themselves, and the designation of one as a "modification" or as "new" seems often arbitrary. The list given below, while aiming at the highest degree of transparency possible in so complex a piece, tries to mention further links wherever applicable.

S1 Material from the main subject appears in two cases.
* In E4, the final eight eighth-notes of S1 are quoted (see U: bars 37-39d) and imitated (see M: bars 39-41d). Both quotes are accompanied by the second subphrase of CS1 (see M: bars 37-39d; U: bars 39-41d). (This first half of the fourth episode, bars 37-41d, will further be referred to as E4A; a cadential close in B major supports a structural caesura on the downbeat of bar 41.)
* In E14, only the final six eighth-notes of S1 appear, but all the more frequently (see L: bars 128/129, M: bars 129/130, U: 130/131, M: bars 132/133, L: 134/135). The ending of CS1 is considerably shortened and dropped after only one occurrence (see U: bars 129/130).

S2 Material from the second subject appears in two versions: as a chromatic descent (taken from the initial six notes of S2) and as an ascent of one whole and three half steps (derived from the five rising notes of S2).
* E13 features the chromatic six-note descent (see M: bars 115/116, L: bars 117/118).
* E10 makes ample use of a variant (with abridged first note) of the five-note ascent (see U: bars 83/84, M: bars 84/85; L: bars 85/86, 86/87, 87/88, 88/89).

CS2 appears twice as an extending sequence (see in E7: L bars 65/66, and in E8 varied: U bars 70/71). Later on, the entire second half of CS2 (i.e. its final twelve eighth-notes) is quoted in E13 (see U: bars 123-125). This recurrence is accompanied by the shared ending of CS1 and S2 (see L: third eighth-note bar 123 to 125d). As a simple closing-formula, this shared ending appears additionally at the close of episodes, particularly in EE3E (see U: bars 31-33) and E6 (see U: bars 59-61) as well as varied in E5 (see U: bars 53-55).

Independent motives - motives apparently invented specifically for the use in episodes and not derived from primary material - are introduced in the first two subject-free passages. The same few motives, abundantly varied, determine all fifteen episodes. (Among the independent motives, an anachronistic relationship could be established between M1, M3 and CS2 on the one hand, and M2, M4 and S2 on the other. Yet this seems more an intellectual game than a valuable contribution to a better understanding of the fugue.)


is introduced in bars 9/10 (U). Its outstanding feature is a beat-2 syncopation which is followed by what reminds one of degrees 6-7-8 of the melodic minor scale. M1 is sequenced one whole tone lower (see bars 10/11) and recurs in bars 49/50 and 50/51 (M).

M1a, a variant of this motive, retains the rhythm as well as the interval structure of the ascent, but displaces the syncopation one whole tone lower and ties the final note into the opening of the subsequent sequence. M1a appears in E13 (see U: bars 118/119, 119/120, 120/121; imitated in M: bars 119/120, 120/121).


features a one-bar pattern in short-long, short-long rhythm. A three-note diatonic descent is preceded by a leading-note. This motive forms the counter-point to M1 in E1 (see M: bars 9/10 and 10/11) and E5 (U: bars 49/50 and 50/51).

In M2a, the leading-note beginning is abandoned and the motive appears straightened to a single diatonic descent, still in short-long, short-long pattern. As such it can be found in E7 (see U: bars 65/66), E8 (see M: bars 70/71), in double length in E9 (see L: bars 75-77).



is introduced in E2. It is characterized by an eighth-note upbeat to a beat-3 syncopation; the interval of this upbeat is generally a fourth but is occasionally adjusted later in the piece. The syncopation gives way to two eighth-notes which have exactly the same interval structure (minor third down, whole tone up) as those in M1. (One could thus regard all the following cases as further relatives of M1. In light of the great number of variants this might, however, cause confusion rather than add clarity.)
In its original appearance, M3 ends on a suspension (see M: bars 17/18); only its sequence brings the expected resolution (see M: bars 18/19). In the same combination, M3 recurs in E4 (see L: bars 41-43, U: bars 43-45).

M3a features a small variation after the syncopation: the pitches of the two eighth-notes now form an inverted-mordent figure. This variant can be found in E9 (see L: bars 77-79); in E12 (see M: bars 107-109 and U: bars 109-111); in E13 (see L: bars 115-117 and M: bars 117-119) and, with a separated sequence, in E11 (see L: bars 101/102 and 102/103). It is worth mentioning that this variant bears a certain resemblance to the second bar of CS2. This becomes particularly apparent at the beginning of E9 where the final note is diverted and the motive thus completely merges with the CS2-segment (see U: bars 75/76 and M: bars 76/77).

M3b is a variant in which the final note is displaced. It occurs in E3 (see M: bars 23/24, U: bars 24/25, L: bars 25/26, U: bars 26/27).




is introduced as a contrapuntal line to M3; and in the same way as M3 was related to M1, M4 is close to M2: it shares the same rhythm as well as the diatonic scale segment which, here, is ascending. One may therefore hardly be surprised to find that it also shares the same kind of "straightened" variant.
M4 in its original shape appears in E2 (see L: bars 17/18, sequenced in bars 18/19) and in E12 (see L: bars 107/108 and 110/111).

In M4a, the motive is straightened to a single diatonic ascent, retaining the short - long, short - long pattern. As such it can be found in E3 (see U: bars 23/24), in E10 (see L: bars 89/90, 90/91; with an extra note: bars 91/92); in E12 (see L: bars 108/109, 109/110); in E13 (see U: bars 115/116).

The fourth motive also engenders a variant in which the final note is displaced: M4b occurs in E3 (see M: bars 25/26), in E4 (see M: bars 41/42) and in E13 (see U: bars 116/117).




is instituted as a two-bar diatonic descent in dotted quarter-notes. It appears in E2 (see U: bars 17-19d), in E3 (see L: bars 23-25d), in E4 with a shortened initial note (see M: bars 43-45), and in E9 (see M: bars 77-79; with a parallel in U).

Where this two-bar descent is fleshed out with additional eighth-notes, a chromatic line results. Due to the context in which it is set by the other voices, this line appears clearly as a variation of M5; see E12 (U: bars 107-109, imitated in M: bars 109-111). In its chromatic content, this variant M5a bears a close resemblance to subject 2; the rhythm, however, is slightly but essentially different (see the beginning on a strong beat and the interruption in the middle).

Three of the episodes are distinguished from the remaining ones by the fact that they feature motives which appear nowhere else in the fugue. As these three episodes are at the same time the longest subject-free passages, this fact certainly deserves attention:

in E3 a syncopation followed by five eighth-notes,
see L: bars 26/27, M: bars 27/28, L: bars 28/29;
in E10

two consecutive syncopations with upbeat and subsequent mordent,
see U: bars 85-87d, M: bars 86-88d, U: bars 87-89d, M: bars 88-90;
and the figure characterized by a "slide" in sixteenth-notes,|
see U: bars 90/91, M: bars 91/92, L: bars 92/93;

in E13

a three-note ascent with artificial leading-notes,
see U: bars 119/120, M: bars 119/120, U: bars 120/121, M: bars 120/121,
U: bars 121/122

The role played by each of the fifteen episodes in the dynamic development of the fugue can be described as follows:

Two episodes (E6 and E15) are nothing but elaborate cadential closes; i.e. they extend the preceding subject statements without any color contrast and bring forth an immediate and complete relaxation.
Four further episodes (E4A, E7, E8, and E14) evolve from the subject entries preceding them by way of imitation or sequence; they, too, do not bring forth a color contrast but just a gradual relaxation.
The three long episodes (E3, E10 and E13) describe self-contained curves. They all begin with an increase of tension (see bars 23-26, bars 83-86 and bars 115-121 respectively) followed by an extended decrease and a final relaxation underpinned by a cadential close.
Most of the episodes (E1, E2, E4B, E5, E9, E11, E12) are short and display independent material. Four of them (E1, E2, E5, E9) are determined by descending sequences, while the others (particularly E11 and E1E2) display rising lines. Besides their individual increases or decreases, all these episodes establish a color contrast to the surrounding subject statements.


II/18.2.5 Character, tempo, articulation, ornament realization

In this fugue, the main subject with its simple rhythm and restricted melodic evidence can be dangerously deceiving, misleading performers into too swift and light-paced a mood. The complexity of the overall rhythmic pattern and the high content of chromatic lines and altered notes indicate in fact a rather calm basic character. The appropriate tempo should express the compound meter in very moderate half-bar pulse.

The relative tempo of the prelude to the fugue sets the larger beats into proportion, thus providing enough variation on the surface with sixteenth-notes in one, triplet eighth-notes in the other piece.

a quarter-note (a beat)
in the prelude
corresponds with
a dotted quarter-note (half a bar)
in the fugue
(Approximate metronome settings: 76 for the quarter-notes in the prelude and the dotted quarter-notes in the fugue.)

The articulation corresponding with this character demands legato for all melodic notes. Exceptions occur in cadential-bass patterns and consecutive jumps (see e.g. L: bars 28, 40/41, 54/55, 112/113, 114, 136/137, 142/143). Due, however, to the high density in this piece of relatively short motives, the legato touch is very frequently interrupted by phrasing which, in almost all cases in this fugue, comes with both dynamic shaping and a cut in the sound flow.

The fugue features several ornaments which require detailed planning. Among the primary material, subject 2 is decorated with a cadential trill before the final note. This trill, as it is approached stepwise, commences on the main note. For the speed of its shakes, sixteenth-notes seem most reasonable (a choice which, one has to be aware, gives all written-out sixteenth-note groups in this piece the value of ornaments). The trill ends in a suffix which uses the major sixth of the current scale. Bach indicates the trill clearly in bar 64. In bar 69, he obviously added it later and would, for all we know, have expected it equally in bars 74 (L), 100 (M), 106 (M), 114 (U) and 138 (M). Only in bars 128 where the subject breaks off early, and in bar 82 where the resolution of the leading-note is delayed, is a trill not necessary. (In bar 82, an ornamental figure actually appears in a complementary place in the middle voice.) The two remaining trills appear outside the thematic material. In bars 30 and 60, the lower voice carries whole-bar ornaments which both begin on the main note and end with a suffix.


II/18.2.6 The design of the fugue

The structural layout of the G# minor fugue is evident from both the entering order of the two subjects and the design and material of the episodes. Yet while the facts concerning the number and extension of sections are quickly explained, it is intriguing to observe Bach's detailed and (in this fugue particularly) artful use of secondary means for creating balance. The fugue encompasses five sections. Subject 1 determines the first two sections, in both cases concluding with a redundant entry (see section I: U M L M and section II: L U L) and a cadential close (bars 31-33d and bars 59-61d respectively). In section III, subject 2 repeats the entering order of the fugue's exposition but ends with the redundant statement in another voice (U M L U). This section, too, closes with a perfect cadence (bars 96-97d). Sections IV and V are then dedicated to the juxtaposition of the two subjects. The cadential close in bars 123-125d divides the sections. It is worth noticing that sections II to V all commence with the three-part ensemble reduced by one resting voice. (The upper voice rests in bars 34m-37, 97-102 and 125/126, the middle voice in bars 61-66.) Bach's mastery of balance is apparent particularly in the subtle play of analogy and symmetry in the episodes. The following summary focuses only on the structurally relevant details.

(1) There are only two episodes conceived entirely as cadential closes: the final episode E15 and E6. The structural importance of the latter is underpinned by two facts: (a) it ends without the overlap of resolution note and new phrase-beginning found in all other episodes (see resolution bar 61 beat 1, beginning of subsequent phrase bar 61 beat 2); (b) it precedes the first statement of subject 2. A falling apart of the fugue into two entirely detached trunks is, however, avoided by the fact that this mid-composition close harmonically represents an imperfect cadence.

With regard to their material, the five episodes before this major structural caesura are basically laid out in the shape of a curve: E1 corresponds with E5, E2 with E4B. This seems to indicate a large-scale bond of the first sixty-one bars of the fugue.


E3 stands out among these episodes for three reasons: (a) it is the longest of the episodes; (b) it ends with a perfect cadence complete with upper-voice closing-formula; (c) it features an extension with a motive, the occurrence of which is restricted to this episode alone.


The seven episodes following the structural caesura in bar 61 display an almost perfect parallel layout. Most obvious is the analogy of, on the one hand, E9 and E12 and, on the other hand, E10 and E13. This leaves the two very short (and structurally identical) episodes E7 and E8 which balance E11. (E11, as a matter of fact, contains two halves of which the second is a varied transposition of the first - one fifth up. E8, interestingly, is a varied transposition of E7 - one fifth down and with inverted voices.)

(5) E10 and E13 stand out from among these seven episodes in the same way as did E3 among the initial group: (a) both of them are considerably longer than any other episode; (b) both end with special closing features, both times in the lower voice (see the long dominant pedal in bars 93-95, and the closing-formula in bars 123-125); (c) both feature an extension with motives, the occurrence of which is restricted to the episode alone.
(6) E14 is closely related to E4a and would fulfil a structurally analogous task.

The diagram below attempts to visualize these correspondences. For a sketch showing the design of the fugue in G# minor, see ex. 65.


E1: M1 + M2
        E5: M1 + M2
  + extension          

+ extension
    E2: M3 + M4 + M5         E4b: M3 + M4b + M5
      E3: M3b + M4a + M5       E4a: Ms1 + Mcs1
        + extension with "E3-motive"      
E6: cadential close


E7: M2a + CS2-sequence E11: M3a

E8: M2a + CS2-sequence

      E12: M3a + M4/4a + M5a
      E9: M2a + M3a + M5           E13: Ms2 + M3a + M4a

E10: Ms2
          + extension with "E13-motive"
+ extension with "E10-motive"
E14: Ms1 , E15: extended cadential close



II/18.2.7 The overall dynamic outline of the fugue

In the absence of any tension-enhancing features such as strettos or parallel entries, minor/major contrasts, or varying contrapuntal density, large-scale developments of tension seem less indicated than careful contrasts of color between subject-determined bars and episodes. Within each section, the dynamic level of the consecutive subject entries increases slightly. This is due to different means in each case:

In I the increase is engendered by the growing number of voices but weakened by the receding tendency of the linking episodes E1 and E2.
In II the less powerful increase of the ensemble is supported by the use of primary material in E4a (interpreted by some analysts as a substitute for a fourth entry), but is again counter-balanced by the decreasing gesture of E5.
In III the entries of the second subject's exposition are strung together by the two short episodes E7 and E8 which evolve from the subject statements by way of sequence and thus enable a longer stretch of a single color, uninterrupted by secondary material. The increase is, however, discontinued in E9 before the redundant entry.
In IV an "exposition" of the subject juxtaposition is rendered similarly stringent by the two linking episodes with increasing tendency (E11 and E12). The only major-mode statement of subject 1 in the entire fugue adds further luster here and converts this statement pair into the climax of the composition.
In V

the use of S1-material in the linking episode avoids a color contrast between the two statement pairs, but the descending motion of E14 counter-acts any dynamic increase.