Reading Bach's Ideas, Part III
Published: 2000 Author: Jonathan Leathwood
III
Reading the Prelude to BWV 998
i The ideal stanza
The design of this prelude has frequently been remarked by guitarists: as so often in Bach’s preludes, the opening idea is visited in an array of keys, ritornello-like, defining five ‘strains’ or stanzas, each new strain being somewhat longer than the one before. As the table below shows, the sequence of keys visited describes a simple harmonic progression. Moreover, a count of the number of measures separating each return of the opening idea reveals a sequence based on ternary numbers (until the coda).
|
Stanza |
Measure |
Proportion |
Harmonic movement |
|
1 |
1 |
6 mm |
I to V |
|
2 |
6 |
9 mm |
V to vi |
|
3 |
14 |
12 mm |
vi to IV |
|
4 |
25 |
18 mm |
IV to I |
|
5 |
43 |
7 mm |
I to I |
And so to example 14. Borrowing a term from Schoenberg, we might describe this process of expanding lengths as ‘developing variation’. But what is being varied? In a way, this is not made clear until the end—or rather not until the coda begins (m. 42), for only then has the bass had its say in full. By then it is evident that if there are such things as modulating, expanding passacaglias, then this prelude is a model example. In example 14, the most complete statement is naturally to be found in the fourth and last strain, when the bass passes through five distinct phases:
Example 14
BWV 998, Prelude: bass plan[13]
In this sense the fourth strain represents the definitive unfolding of all five phases. Not that it is autonomous: like the previous strains it is necessarily an open structure, beginning as it does from a foreign key (IV). This raises the possibility of an intriguing compositional puzzle: to take the opening idea in the tonic and to essay all five phases in a single span. To do so would be to collapse the entire prelude into an ideal stanza: a closed structure, beginning and ending in the tonic; in effect, a complete piece.[14]
Example 15 takes up the challenge, and attempts to pass through all five phases of the bass just once. No doubt a condensed version of the prelude could have been realised in a number of different ways: but if example 15 is anything to go by, perhaps no solution would quite convince. A particular difficulty comes at the beginning of phase (iii): no sooner has A major (V) been approached then the music retreats back into the tonic of D.[15]To attempt the task of example 15 is not to suggest that Bach had before him his own ideal stanza which he then let out at the seams, and expanded into the finished work. Why then perpetrate such a fiction? As a matter of fact such an extrapolation proves its value, to the extent that it portrays our experience of structure. If each stanza outlines a facet of a submerged and fragmented ‘ideal’, viewed in a certain perspective, then what remains, after the piece is over, is a residue or trace which is a superimposition of all the stanzas taken together, a summation of all these different vantage points. The organisation of this prelude is prismatic: it is left to us to reblend the various approaches back into a whole.
Example 15
BWV 998, Prelude: an attempted ‘ideal stanza’
Example 15’s attempt, and inevitable failure, to notate this residue literally suggests that in the last resort no one statement can quite hold together, and for a simple reason—it lacks a climax. It is not for nothing that Bach has four stanzas, four commentaries, before the coda: only from such partial but graded statements is the sense of gaining a high point at all possible. For when at last, the finished picture emerges, we experience it not, surely, by an intellectual effort, but through an accumulation of feeling. This is achieved by Bach’s expansion of each stanza, brought to a head by the fermata of m. 40, and released by the unique sixteenth notes that celebrate its resolution. At such a moment, we are granted a powerfully direct impression of a single governing statement—no matter that it cannot be written down. An effort such as example 15 remains a metaphor: an icon not only of our intuition of form, but of the point of climax and resolution.
In the search for an ideal stanza, it is thought-provoking to consider the coda, since it is the only segment of the prelude which is tonally closed. Admittedly, it hardly represents the elusive ideal, but what it does show is its terminal poles: the opening tonic pedal and the disjunct bass movement of the final cadence—bass motion that has been painstakingly earned, snakes-and-ladders fashion, throughout the prelude.
ii Foreground & background
Bach’s rule for the bass line, then, is clear. Each successive stanza must review in order all previously stated material before adding anything new. Now segments of the bass line will occur in many different contexts, and simple pragmatism dictates that as he arrives at a new phase in the ground bass, Bach must at the same time look ahead and plan for its successful use in all subsequent stanzas.
What now of the upper voice? At this point it is startling to realise that Bach’s principle of use and reuse in fact applies to the complete two-voice framework. This is not apparent on the surface of the music simply because each time Bach arrives at an analogous or parallel passage (upper and lower voices together) he encounters—I should say devises—a number of technical problems. The rule of repetition of both voices together, then, is one that cannot be kept, but it is a rule nevertheless. For this very impossibility of a series of literal restatements (we shall shortly see why) is the source of the ‘variational’ character of the upper line. We miss the point if we only marvel at the unity of the bass line, and Bach’s rich inventiveness at finding new counterpoints to it: in fact Bach had to find new counterpoints, and when he didn’t have to, he generally didn’t. At the same time, each new invention is conscientiously related to the precedent of the previous stanzas—not least, but not just, through the persistence of the bass line.
As a first instance, let us go to mm. 27–30. In m. 28 the music enters phase (ii) of our ‘ideal stanza’, and our first resort—the very simplest continuation—would naturally be to use again the material heard in mm. 3–5. The result is illustrated in example 16, line a. It fails because the sequential figure (x) leads to a leading note (the low C sharp) in m. 29 which is left hanging (until m. 35 at the earliest). Bach is ready with a solution: he simply avoids the C sharp altogether, and composes instead an ascending line over the original bass (line b). One problem remains: as the upper voice moves through an ascending fourth, a bare octave is struck in m. 29, and the bass’s E must be altered to C sharp (line c). This is the only instance of Bach’s altering a note apparently fixed in the bass plan so as to accomodate a variation in the treble.
Example 16
BWV 998, Prelude, mm. 27–30
Throughout the rest of the prelude, the problems are of two kinds, involving either changes of mode and non-parallel harmonic goals. We shall discuss each in turn. Note that in the musical examples below, line a will each time illustrate music notionally rejected by Bach: that is, it will notate a literal and latent restatement, in the governing tonality, of music heard previously; subsequent lines will illustrate Bach’s real choices—his remodelings and variations.
This is somewhat more and somewhat less than musical relic-hunting. Less because in the last resort, no-one can claim to understand anything about Bach’s actual thought-processes: the reader will by now have realised that the ‘Bach’ of my argument is nothing without the quotation marks. More because the rejected passages construed in the examples are not only ‘his’, they are mine and yours—in short, the work’s. In a word, they constitute a background to the music that we hear and play. They are the sum total of our expectations, our educated guesses. And by contradicting this background, by charting the unexpected, the music that Bach puts in its stead gains its vitality. Thus stated, this is Hans Keller’s contention that all good music relies on the friction between expected and unexpected—the listener provides the background (the expectations) and the composer supplies the foreground (the surprises).[16] But in this prelude, as we attempt to define these discarded continuations and bring them into awareness, we discover the most surprising thing of all. Whereas in later music especially, the background consists of the routine and the predictable, in this prelude it consists more of the intractable and the unworkable.
iii Changes of mode
MAJOR TO MINOR Not all music in the major mode survives literal translation into the minor. In this prelude, the difficulty is essentially a melodic one: what to do with the variable sixth degree of the scale. Bach’s solution is almost always to avoid it.
The simplest case of this is found in strain 3, when the composer has to restate the opening in the relative minor, B. It can be seen from example 17 that Bach was unhappy with the effect of G sharp in m. 15 and made a slight adjustment to the upper line (lines a and b). (G natural is of course impossible here, but G sharp is a borrowing from the major, and Bach evidently felt that in its harmonic context it introduced too heavy a major colouring.) The reader will note that this change is retained in strain 4 (m. 26) even though it is no longer needed (line c), and that the new motive thrown up by the change inspires the subsequent sequence (in mm. 30ff., not in the example). It is such particular and beautiful details as these that go to make up the sense of developing variation.
Example 17
BWV 998, Prelude, mm. 14–15
And here is a chance to induce something about Bach’s thought-process. For having discovered that his opening didn’t quite work in the minor mode, and having made an acceptable change—one which even persists in the next major-mode statement—what stopped him from returning to the opening and making the same change there, so as to ensure perfect consistency?
We must conclude that such a course would be contrary to Bach’s notion of craft. For to change the opening measures on the grounds of their slight unworkability in the minor would be to condemn their invention as defective. In this prelude, though, the fitness of the opening idea is not defined by its potential, but by its immediate rightness. On this view, any idea which is in itself attractive is eligible for elaboration. Hence the opening to this prelude must be worked out as it is, and if the constraints of tonal working subject it to variation, evolution and erosion, so be it: the idea itself must not be changed. After all, why else would Bach have written the simple fugue which follows this prelude, based as it is on the very plainest subject, which is nothing without its melodic shape, and yet which cannot be answered in the dominant without that shape being destroyed utterly?[17]
Example 18
BWV 998, Prelude, mm. 8–9
The next two examples again deal with the problematic scale degree 6. Example 18 is routine: line a demonstrates an unacceptable clash of C natural and C sharp within the same harmony. In example 19, line a, the D sharp in m. 18 interferes with the linear connection between the two upper D naturals: line b clarifies the connection with a harmonic reduction and Bach’s reworking is shown in line c.
Example 19
BWV 998, Prelude, mm. 17–19
This last example prompts an interesting question: if such a casual emendation was all that was needed in stanza 3, why did Bach go to so much more elaborate lengths in stanza 2, where phase (ii) has a quite new upper line (mm. 9–10)? The answer is supplied by the same considerations of voice-leading as applied in example 16. If we examine a transposition of mm. 4–5, duly altered as in stanza 3 and shown in example 20, we soon see that the upper voice’s D sharp in m. 10 is left hanging. This problem does not arise in stanza 3, since the next phase is recomposed (see section iv). Curiously, then, the more elaborate recomposition precedes the less elaborate one—a potent example of Bach’s, so to speak, passive or preordained variation of the material.
Example 20
BWV 998, Prelude, mm. 9–10
Bach does not invariably avoid the problematic scale degree 6. Example 21 provides a counter-example to show how a sharpened sixth rejected by Bach in one circumstance (line a) could be accepted in another. The G sharp in line b is acceptable in three ways: firstly, the tonality is in flux; secondly, the G sharp is supported with its own bass note; and thirdly, its chromatic colour is prepared by A sharp in in the previous bar.
Example 21
BWV 998, Prelude, mm. 13–14
MINOR TO MAJOR Translating from minor to major does not usually present melodic problems, but the different collection of chords offered by the two systems can throw up difficulties when translating sequences. As we have seen, phase three of our ideal stanza is a three-bar modulating sequence: in stanzas 2 and 3 (of course, the sequence is not yet presented in stanza 1) the respective sequences begin from a minor tonality. Not until stanza 4 is the sequence put to the test in major. The result is shown in example 22: the sequence becomes problematic in the second measure when a diminished chord is encountered and we are left wondering what to do with the upper line’s G. (Another potential problem is that these measures virtually duplicate mm. 12–13 at the same pitch, but in fact Bach did not always forbid himself such straightforward repetitions.) Bach’s solution is again to avoid the false note, and so a new sequence is composed against the background of the old one.
Example 22
BWV 998, Prelude, mm. 30–33
iv Non-parallel harmonic goals
Perhaps the most fascinating set of compositional alternatives occurs in m. 19. At this point Bach must use again the modulating sequence, on the model of mm. 12–14. Trying this for ourselves, as in example 23, we discover that the sequence breaks down, since it is bound to reach either C sharp diminished (which is no goal at all) or C major (which isn’t either, since it is not a direct key-relation of the tonic, D).
Example 23
BWV 998, Prelude, mm. 19ff., recomposed
The solution appears to be to quit the sequence a measure earlier than the model (see line b of the example). It is not hard to see why Bach rejects this solution too (after all, everything is easier to see after Bach has shown it to us): to curtail the sequence so abruptly seems to admit its defects without compensating for them—the background (the impracticable) replaces, indeed defeats the foreground (the inventive). (Some, too would press the claims of the proportions of each stanza, for if the sequence were reduced to two bars, this third stanza would total only ten bars, breaking the ternary-based sequence shown in the table above.) Bach, to be sure, admits the pressing harmonic need to break the sequence after two repetitions, but he transforms bare necessity into rich invention by stating the bass in augmentation, and so composing a quite new upper line.
v The tonic minor
One more feature of the bass plan invites comment: the transformation of mm. 23–25 into mm. 36–38. In the latter section, as the tightly packed motivic content of the upper voice liquidates into instrumental figuration, the bass restates mm. 23–25 but in the minor key. Given this restatement, the performer must find some way of comparing in sound the two bass descents. There are many ways of shaping them: one could choose to underline the similarity by playing them in the same way, or one might try to confront equivalent but opposing inflections, such as a diminuendo in mm. 23–25 versus a crescendo in mm. 36–38 (example 24).
Example 24
BWV 998, Prelude, mm. 23–25 and 36–38
The goal of the second descent, of course, is the G minor harmony—a key relation not of D major but of D minor. The consequence is a truly startling progression, in which the Neapolitan harmony (Eb of m. 39 is wrenched upwards by a semitone to the secondary dominant (E seventh) in m. 40. Functionally speaking, they each represent the same chord in two guises, the first essentially from the minor domain, the second from the major. Example 25 clarifies this equivalence. The shock of m. 40 is made even greater by the unprepared seventh in the bass—the most dissonant spacing of a seventh chord available—and the appoggiatura which for a moment creates a quartal sonority.[18]
Example 25
- In example 14, the slurs are not intended to show phrasing, but, loosely, movement and dependence. [back]
- See the chapter entitled ‘The Ideal Ritornello’ in Laurence Dreyfus’s deeply insightful book, Bach and the Patterns of Invention (Cambridge [Massachusetts] and London, 1996: Harvard University Press).[back]
- The choices governing the treble voice need not detain us at this point: later it will become clear why, for example, I could not merely transpose the upper line in mm. 11–14 in phase (ii), but had to content myself with a more indirect derivation.[back]
- Hans Keller, Essays on Music, ed. Christopher Wintle (Cambridge 1994: Cambridge University Press). See particularly ‘Towards a Theory of Music’, and ‘Mozart’s Wrong Key Signature’.[back]
- In a future article I plan to show how Bach elaborates this fugue so as to balance these two incompatible imperatives: the melodic shape of the subject versus the tonal constraints of the answer.[back]
- quartal: made up of fourths—in this case, the bass’ D is part of the quartal chord, transferred registrally down by three octaves.[back]
©2000 Jonathan Leathwood