WTC II/12 in F minor – Prelude

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

II/12. 1.1 The prelude type

Several features in this piece immediately capture the listener's attention. With regard to material, there are regular phrases which are very neatly contrasted; texture, melodic figures, rhythmic features and intensity all change drastically from one phrase to the other. With regard to structure, almost all phrases recur several times, so that a high degree of familiarization is possible in this piece, particularly since both halves are repeated.

We can thus state that this prelude is determined by material. If there is a reluctance to say "by motives", this results from the fact that one expects motives in Bach's compositions to be single-voiced units used within a polyphonic texture, with imitations, sequences and partial developments. The phrases here represent a variety of textures, although imitative play is not among them.


II/12.1.2 The overall design of the prelude

The harmonic progression in this piece is much more largely scaled than the design in four-bar phrases might suggest. Many of the phrases do not represent independent harmonic clauses. The composition encompasses mainly the familiar steps: a progressive step away from the tonic in the first half of the prelude, and the corresponding regression from the subdominant back to the tonic in the second half.


bars 0m-12m

tonic to relative major (F minor to Ab major)


bars 12m-24m

relative major key confirmed

bars 24m-28m

relative major key confirmed again


bars 28m-40m

modulation to the subdominant (Ab major to Bb minor)


bars 40m-48m

return to the tonic with imperfect cadence


bars 48m-60m

interrupted cadence in home key

bars 60m-62m

plagal cadence in home key

bars 62m-66m

home key confirmed

bars 66m-70m

home key confirmed again

In view of the repeated use of only a limited variety of material, it will come as a surprise that true structural analogies are scarce.

bars 4m- 8m

correspond with

bars 16m-20m (transposed)

bars 20m-28m

correspond with

bars 62m-70m (transposed and varied)


II/12.1.3 Practical considerations for performers

The rhythmic pattern in this prelude can be regarded as simple, being based on a very regular use of eighth-notes and sixteenth-notes. The pitch pattern, on the other hand, contains much variation in the different phrases: there are stepwise passages, many broken-chord figures and large intervals, the latter occurring both in the sixteenth-notes and in the eighth-notes.

The character must be determined as a borderline case. The basic features point to a rather lively character, but the abundant use of appoggiaturas indicates at the same time an emotional intensity which puts a restraint on the choice of tempo.

When making a decision regarding articulation, the various components of the material must be investigated very thoroughly:

non legato

with a neutral touch applies to all quarter-notes that represent a cadential bass line (see lower voice bars 1-4, 8-16, 28-32, 56-58); for all quarter-notes which must be distinguished from rhythmically interlocking, yet melodically more relevant note-groups in another voice (see e.g. in bars 20-24); and for all eighth-notes followed by rests in complementary rhythm patterns (see e.g. bars 4-8 and 52-56).

non legato

with a more singing touch is required for all quarter-notes and eighth-notes which form melodic lines


with high-intensity touch characterizes all appoggiatura/ resolution pairs; for details see bars 1-4, 9-16, 28, 29-32, 40, 57 and 70 in parallel voices; furthermore M: bar 34, U: bar 36 (where the appoggiatura grows out of a suspension), bars 37, 38, 42, 44, 45, 46 (in partial sequences), 49, 51, and L: bar 52.


also applies to the closing-formulas in bars 39/40 and 69/70.

quasi legato

with a light touch is appropriate for the sixteenth-notes in the toccata-style complementary patterns (see bars 4-8 etc.).

The score contains five ornaments, all of them mordents (see bars 40, 42, 45, 46 and 69). Three of them are approached in stepwise motion and therefore begin on the main note; a simple shake with three notes is sufficient here. The other two mordents are preceded by a large leap (bars 45, 46); they therefore commence on the upper neighbor note and contain a four-note shake.


II/12.1.4 What is happening in this prelude

As was already mentioned, the F minor prelude is built upon a small number of motives. These motives are not single-voiced units within a polyphonic texture but instead short homophonic phrases. Each is at first presented in its own unique texture but there is a vivid exchange and interplay of accompanying features later in the piece.


is presented in the first four bars. The lowest of its three voices features quarter-notes in cadential steps. The upper and middle voices run in parallels. Their core figure consists of a eighth-note upbeat, a downbeat-appoggiatura which repeats the pitches of the upbeat, and a stepwise resolution.

Owing both to the melodic parallel and to the appoggiaturas, each of the one-bar patterns is high in emotional intensity. The entire four-bar motive dynamically represents a large curve. The tension increases until the downbeat of bar 3 where the subdominant harmony coincides with the sudden leap in the pitch level of the upper voice. The ensuing decrease in tension is expressed, in the lower voice by the chromatic transition to the dominant, and in the two higher voices by a gradual descent.


represents a toccata-pattern. The strong beats in the left hand are complemented by three sixteenth-notes in the right hand. The pitches of both voices interlock so as to give the impression of a single layer in broken-chord figures. There could thus hardly be a more obvious contrast between the textures of M1 and M2.

The melodic content of M2 is negligible: the left-hand notes owe their melodic value more to their metric position on strong beats than to the consistency of the line they build. This is not surprising since they are, after all, mainly part of broken chords. This lack of melodic expression is reflected in the somewhat subdued intensity of this motive and in its lack of direction.

Harmonically, these four bars make no progress but extend the C major chord reached in the middle of bar 4 in an ornamented fashion. The broken-chord figures seem to circle without a goal (see bars 5-6m which recur identically in bars 7-8m). Correspondingly, the dynamic tension is as if suspended. There is neither build-up nor relaxation.


appears in bars 20m-24m. The texture displays again three voices, each of them essentially independent of the other. What makes the interplay of the three voices nevertheless so simple is, on the one hand, the regular rhythm and, on the other hand, the one-bar sequential pattern. Each of these one-bar units contains a rising step in quarter-notes in the lower voice, a syncopated falling third ending on a weak beat in the middle voice, and a curve made up of rising and falling broken-chords in sixteenth-notes in the upper voice.

Harmonically, each middle beat represents a seventh-chord followed on the second eighth-note of the following bar by the resolution (see e.g. bars 20m-21m: Ab7-Db). The entire motive conveys the impression of a long release in tension.

The following four bars seem different at a first glance. They can, however, be regarded as a variant of M3. The complementary rhythm of the lower and middle voices appears in a different guise, and the broken-chord figures in the upper voice have given up any pretence to melodic shaping. This variant will be referred to as M3a. Harmonically, it is distinguished by a high content of accidentals (see particularly the Fb and Cb in bars 26/27). Although the melodic direction, as in M3, is descending, M3a thus expresses an increase until bar 27, followed only then by a relaxation.


is introduced in bars 40m-42m. Its obvious two-part texture hides a third layer, as the lower voice is in hidden two-part structure. The melody in the upper voice presents a two-bar unit, the longest unit in this prelude and also the most varied as it encompasses steps, skips, broken-chord and appoggiatura/resolution in a rhythmic pattern of eighth-notes, sixteenth-notes and ornamental notes in the mordent. The melodic part of the lower voice contrasts this melody with a descending line (see bars 40-42: C-Bb-Ab-G-F-Eb), while the harmonic background within the hidden two-part structure represents a D pedal followed by a broken G minor chord. The appropriate dynamic shape is that of a curve, in which the climax falls on the appoggiatura.

M4 does not recur later in the piece. Instead, it is sequenced immediately (see bars 42m-44m), continued in two varied partial sequences, and complemented by a tail in which we vaguely recognize inverted voices of the M4 pattern.

As has been shown, the first half of the prelude consists, strictly analyzed, of two sections (section I: bars 0m-12m, section II bars 12m-24m, extended to 28m). With regard to the material, the two sections begin similarly; they are distinguished only in their third segments. It is, however, important to notice that Bach composed the conclusion of the first section in such a way that it appears melodically threaded tightly into the beginning of section II. The entire first half of the prelude thus appears, despite its two cadential progressions, as one large entity.





bars 0m- 4m

F minor:: i-V

F minor: imperfect cadence



bars 4m- 8m


extension of imperfect cadence



bars 8m-12m


modulation to relative major



bars 12m-16m

Ab major: i-V

Ab major: imperfect cadence



bars 16m-20m


extension of imperfect cadence



bars 20m-24m


relative major confirmed



bars 24m-28m


relative major confirmed


The second half of the prelude begins and ends similarly to the first half (compare bars 28m-32m with bars 0m-4m, and bars 62m-70 with bars 20m-28). In-between, M4 distinguishes this half from the first, and components of the three earlier motives appear intermingled and developed.

The original little ternary group M1/M2/M1 begins almost faithfully. M2 can be recognized only from its complementary-rhythm figure in the left hand, while the right hand retains the three-part texture here and even quotes appoggiaturas (as suspensions) from M1. In the third segment, however, even the familiar appoggiaturas cannot convince the listener to recognize a variation of M1.

The middle section of the prelude's second half is taken up by M4 and its development. After that, the "recapitulation" of M1 (see bars 48m-52m) occurs in the two lower voices, with several of the appoggiaturas omitted and the broken-chord figure from M3 as accompaniment. The subsequent M2 begins with the expected complementary rhythm, but the texture displays three parts and the broken chords are converted into step-wise motion. Furthermore, the second half of this four-bar motive is replaced by a cadential formula which seems to say: "Let us try once more whether we cannot get a better recapitulation."

The subsequent quotation of M1 begins similarly to bars 0/1 but diverts immediately thereafter, soon abandoning all motivic content in favor of a closing-formula which ends in a deceptive cadence (bar 60). The ensuing M2 is so essentially varied that it requires careful shading by the performer – in a color as close as possible to the original M2 – to help the listener's recognition, particularly since it is abandoned after only two bars.

Compared to these very strong modifications, the small variations in the recurrences of M3 (see bars 62m-66m) and M3a (see bars 66m-70) are easy to grasp even at first listening.

Despite the numerous modifications and even distortions of the original material -or perhaps because of just these irregularities – it seems crucial that the performer have in mind a clear picture of the structural layout of the prelude. The table below attempts to stress the similarities over the dissimilarities in order to facilitate a consistent presentation of the piece.

bars 0m-12m



bars 28m-40m




bars 40m-48m

M4 etc.


bars 12m-20m



bars 48m-56m




bars 56m-62m


bars 20m-28m



bars 62m-70




WTC II/12 in F minor – Fugue


II/12.2.1 The subject

The entire F minor fugue – and its subject most of all – expresses a dance-like character. This is achieved essentially by the persuasive rhythm and pitch pattern in the head of the subject: upbeat / triple repetition, upbeat triple repetition. These two immediately remembered initial bars are complemented by two bars of unobtrusive running notes. The end of the subject comes, as can be expected of such a regular phrase, after four bars. The only irregularity arises in the two possible endings Bach uses for this subject. There is the strong-beat ending (in a terminology taken from Greek poetic meter often called the "male" ending) in which the melodic line concludes with a return to the tonic harmony on the downbeat of the fourth bar (see e.g. in bar 32). Much more frequent, however, is the weak-beat (or "female") ending which extends this tonic through two more eighth-notes and thus achieves a more perfect complement to the subject's one-eighth-note upbeat (see bars 4, 8, 15, 28 etc.).

The divide between the two upbeat / triple repetition groups described above marks a comma in the phrase structure. In contrast, the final note of the sequence (the third E in bar 2) is simultaneously the beginning of the running notes. Therefore, the cut after the third F in the first bar remains the only one, and the subject thus consists of a short first subphrase and a much longer second one.

The rhythmic pattern, in the subject itself as well as in the fugue as a whole, is simple, with eighth-notes and sixteenth-notes clearly dominating. The most obvious features in the pitch pattern of this subject are, on the one hand, the jump and broken chord respectively between the first two upbeats and their subsequent downbeats (F-C = perfect fifth, Db-Bb-E = extract of C9 chord), and, on the other hand, the triple note repetitions. The remaining sixteenth-notes contain above all stepwise motion.

The harmonic layout of the subject is also straight-forward: the tonic on the downbeats of bars 1 and 4 surrounds the dominant in bars 2 and 3. The subdominant appears in a metrically weak position (bar 1, last eighth-note) and thus does not gain momentum. It marks, however, the only fast harmonic change in this subject (with only one eighth-note between its appearance and the change to the dominant) and thus endows the beginning of the second subphrase with a touch of urgency.

(ex. 70)


The dynamic development reflects these features. Within the initial short subphrase, the downbeat, prepared by the bouncing upbeat, is already clearly accented. This process is enhanced in the sequence where the first E is distinctly more prominent than the first F had been. While the note repetition in the first subphrase brings about an immediate release, the length of the second subphrase requires a very gradual diminuendo. (Any emotional "waves" in bar 3 should be carefully avoided in order not to blur the structure and character of this subject.)


II/12.2.2 The statements of the subject

The F minor fugue encompasses only nine subject statements.


bars 0 - 4



bars 40 - 44



bars 4 - 8



bars 50 - 54



bars 11-15



bars 71 - 75



bars 24-28



bars 74 - 78



bars 28-32


(ex. 71)


Apart from the interval adjustment in the answer (as always in fugues whose subject begins with a fifth, this interval is converted into a fourth to retain the tonality) and the above-mentioned weak and strong ("female" and "male") endings of the phrase, the subject remains unchanged. No inversions occur, and the only overlap of consecutive statements (see bars 74/75) is so minimal that it is not perceived as a stretto entry.


II/12.2.3 The counter subjects

Bach has not invented any counter-subject for this fugue.


II/12.2.4 The episodes

We can identify six subject-free passages.


bars 8m - 11


bars 44m - 50


bars 15m - 24


bars 54 - 71


bars 32 - 40


bars 78 - 85

In view of their very distinct material, four of these episodes can be subdivided:


E2a bars 15m-16,

E2b bars 17-24;


E3a bar 32,

E3b bars 33-40;


E4a bars 44m-46m,

E4b bars 46m-50;


E5a bars 54-65,

E5b bars 66-71.

Material from the subject appears in two versions in the episodes. On the one hand, episodes can begin as prolongations of the subject ending. E2a is conceived as an extension of the preceding entry: the lower voice continues in descending sequences while the upper voice repeats the step Db-C (see bars 14-16). Similarly, E3a features a one-bar imitation of the final subject bar (see M: bar 31, U: bar 32), while E4a contains a sequence of the final two bars of the subject (see bars 42m-44m: L, sequenced in 44m-46m). On the other hand, the subject head with extending sequences appears as episode material in E5a (see bars 56-58 and 62-64: L).

Genuine episode material in this fugue includes one sequence model and one small motive. The sequence model is introduced in bars 17/18. It consists of a leading voice (here U), a supporting voice with a triple-note repetition (here M), and broken-chord figures (L). E2b consists of this model with two descending sequences, one faithful (bars 19/20) and the other varied and extended (bars 21-24). Structurally corresponding is E3b where the upper and middle voices are inverted and the extension of the second sequence is slightly modified. Another analogous episode is E6, while E5b is shortened in its second sequence.

The motive (M1) is first presented in bars 47/48 (see L: Db to F). It is sequenced in bars 48-50 (L) and imitated in bars 50/51; it also appears as slightly shortened parallels in bars 47/48 and 49/50 (U), and finally in bars 54/55 (U) and 55/56 (L).

In the absence of a counter-subject, the various lines that accompany the subject entries must also be regarded as secondary material. Here, as in E1, we can observe many more sixteenth-note figures. Some of them are sequenced or imitated (see e.g. L: bars 24/25, 25/26 followed by M1; L: bar 29 / U: bar 30 followed by M1var in L; U: 51, 52, 53; L: bars 75, 76). Yet as they do not create any obvious patterns on a larger scale, they need not be considered in detail.

The relationships between the episodes of this fugue are manifold and very revealing for the structural layout. Most conspicuous are the four larger portions which correspond and are distinguished from their surroundings also by texture, due to the three-part model:













All of them have harmonically concluding endings – two of them more open in character (see bars 24 and 71), the other two with explicit closing-formulas (see bars 39/40 and 84/85). E2b and E3b are prepared by analogous episode segments extending a preceding entry: E2a and E3b. Finally, E5a consists of two corresponding halves (compare bars 54-59 with bars 60-65).

This leaves two episodes – E1 and E4 – which are not so distinct in their material and structure. Both of them serve as links between two subject statements within a section.

Dynamically, E5a is the only episode which represents truly active increases in the ascending sequences of the subject head. (One might have a point in arguing that these make up for the missing third entry which balance with section I would require). E1 and E4 begin with a slight increase before giving way to relaxation, while all other episodes in this fugue represent decreasing tendency. Within these others, it seems a good idea to set the four above-mentioned corresponding blocks with their distinct sequence models apart from the remainder of the fugue by choosing a different color. (In such a case, having the second manual of a harpsichord in mind may help; the equivalent on our pianos is to depress the soft pedal for these segments. This gives an effect which is both very convincing in terms of the structure of this fugue and very much in keeping with Baroque performance practice. Very exact pedaling of the left foot is, of course, essential.)


II/12.2.5 Character, tempo, articulation, ornament realization

The thematic jumps and broken chords as well as the simple rhythm pattern clearly indicate rather lively basic character. The tempo should be fast enough to create the effect of a true 2/4 meter – with only one strong beat in each bar – and to guarantee the necessary light character in the sixteenth-notes.

The relative tempo of the prelude to the fugue poses a question for the interpreter. A direct proportion of quarter-note = quarter-note (and therefore: bar = bar) is possible but may easily sound dull. A complex proportion, while more difficult to imagine, gives more satisfactory results for each of the two pieces. This translation implies using a note value that does not exist, the triplet eighth-note, for correspondence:

one triplet eighth-note

corresponds with

one eighth-note

in prelude tempo

in the fugue

(Approximate metronome settings: prelude beats = 66, fugue beats = 100.)

The appropriate articulation includes non legato for the eighth-notes and a crisp legato for the sixteenth-notes. The non legato may sound merrily bouncing in the subject itself, slightly less aggressive in the secondary material and particularly gentle in the sequence models.

The only ornament suggested in the score appears, in parentheses, on the first down-beat note of the first subject entry. If it is played, it has to be added consistently in all other complete entries (which is generally no problem except for the middle-voice entries in bars 28/29 and 74/75 where playing an inverted mordent would be somewhat awkward – although it can certainly be done). An ornament which is not indicated even in brackets but which almost all performers will intuitively add is the cadential trill on the dotted G in bar 84. It commences on the upper neighboring note and stops after a mordent of two shakes.


II/12.2.6 The design of the fugue

The structural layout of this fugue is very clear, as can already be seen from the outline and function of the episodes. The four corresponding episode segments each conclude a section. As they end with different degrees of melodic conclusiveness, an impression that sections I+II and sections III+IV form two larger blocks inside a binary design arises. This concept is supported by the observation that the first and third sections contain bridging episodes between their entries, while the second and fourth sections do not. The following table attempts to visualize the design.

section I:

subject in U, M
E1 (linking)
subject in L
E2a (short opening)
E2b (sequence model;
no closing-formula)

section III:

subject in L
E4 (linking)
subject in M
E5a (long opening)
E5b (sequence model;
no closing-formula)

section II:

subject in U, M
E3a (short opening)
E3b (sequence model;

section IV:

subject in U, M
- -
E6 (sequence model;

The harmonic plan of the fugue supports the binary layout in that both the first and third sections set out from the tonic. The developments inside each half of the fugue, however, are different:


Section I contains three statements in the field of the tonic, after which E2 modulates to Eb major (the dominant of the tonic relative Ab major).


The second section is in major mode, with its two entries in Ab and Eb major respectively; E3a then modulates to C minor, and E3b closes in C major (the dominant).


The two statements of the third section are both on the tonic, after which E5 ends on an F7 chord (the dominant of the subdominant Bb minor).


The fourth section features entries on the subdominant and tonic and thus closes the circle.

For a sketch showing the design of the fugue in F minor, see ex.72.



II/12.2.7 The overall dynamic outline of the fugue

In the absence of counter-subjects, strettos and other tension-enhancing features, the dynamic development from one entry to the next is not the foremost intent in this fugue. In section I, the tension rises throughout the three entries – the dynamic curve in the bridging episodes constituting no interruption but just suspension – and then diminishes gradually through the concluding episode E2. Correspondingly in section III, the tension increases through two entries (suspended during E4) and again in the sequences of the subject head, before it subsides gradually throughout the concluding E5. In both cases, the harmonic ending on what is in fact the dominant of the subsequent section beginning creates further release and thus mollifies the beginning of sections II and IV. The reduced texture supports this, as the entry in bars 24-28 appears manifestly (and that in bars 71-75 practically) in two-part setting. In both cases, the final subject statement is placed in the middle voice which is the most subdued of the three possible positions; any growth in tension that might be brought forward by the increase in texture is thus annihilated, and sections II and IV remain much less outgoing than sections I and III. The concluding episodes E3 and E6 end both halves of the fugue on a soft note.