The Rite of Spring at 100

25Dionysos Monometrikos

Stephen Walsh

On an early visit to the Paul Sacher Stiftung in Basel, soon after it opened for research in the mid-1980s, I worked on the manuscript materials for the Symphonies of Wind Instruments and on the various sketchbooks that contained what seemed like random sketches for that work. It was, in fact, eventually assembled out of fragments partly intended for other works or for no particular work. A handful of sketches related to a project with the title “Cinq pièces monométriques”—“Five Monometric Pieces.” There was not much to it, certainly not the outline of an actual work. But I was intrigued by the concept and started looking at the Symphonies partly in this light. “Monometric” seemed to refer to a purely linear music based on a single-unit pulse without consistent metrical grouping at higher levels, and the interesting thing about it as far as the Symphonies of Wind Instruments was concerned was that it accommodated a kind of music in which the notated meter—that is, the barring—was entirely subject to the contour and accent pattern of the melodic line. This is the reverse of the situation in classical music, where the norm is to fit the melody and harmony to a regular meter and where exceptions—like the cross-rhythms in Schumann and Brahms—work specifically against the fixed meter.

You can see how the monometrics work in the Symphonies of Wind Instruments by comparing the 1920 version of the score with the revision Stravinsky made in 1947. For instance, the episode for three flutes soon after the start (compare R-4 in the original with R-6 in the revised version, Examples 25.1a and 25.1b) was not only retextured and rebarred in 1947, with its downbeat accents redistributed, but also extended by one measure of six eighth notes (one-sixth of its previous length), without any obvious damage to the balance of the phrase. Not only was the scheme of accents modified and to some extent regularized, but even the phrasing slurs were completely rethought in 1947 to make them coincide with the altered barring. There is a distinct feeling that barring had been a tiresome necessity for Stravinsky in ensemble works around 1920, forcing him to make decisions that were perhaps best left unmade. As is well known, much of the solo Piano-Rag-Music is unbarred: sketches for that are mixed up with the Symphonies material, and the piece can in part be understood as monometrics reconfigured as ragtime. In fact, in the draft fair copy of the Symphonies itself, the flute episode and one or two other passages are likewise unbarred.

Example 25.1. Symphonies of Wind Instruments: (a) 1920 version at R-4; (b) 1947 version at R-6. © Copyright 1926 by Hawkes & Son (London) Ltd. Reprinted by kind permission of Boosey & Hawkes Music Publishers Ltd.

There is, of course, a more famous and much earlier example of Stravinsky’s difficulties with bar lines. We all remember his observation in Expositions and Developments that when he composed the “Sacrificial Dance” in The Rite of Spring, he was able to play it but did not know how to write it down.¹ And although he does not say so, it seems clear that the chief difficulty lay in the precise location of the bar lines in music whose accents are distributed irrationally both within and between the polyphonic lines. There is an illustration of this in the barring changes Stravinsky made in the various editions the work went through after its first publication as a piano duet in 1913. And what these changes show, fairly clearly, is that the fixed, almost bullying characteristics of bar lines are alien to the varying weight and status of the music’s accents.

In other parts of the work, the same kind of problem arises from the polymetric nature of the music. The introduction to the ballet, for example, is strictly unbarrable but would obviously be unperformable without bar lines. The famous opening bassoon melody (barred strictly in in Anton Juszkiewicz’s collection) is rethought—or perhaps back-thought—as an unmetered piece of “drawn-out” (protiazhnaia) folksong, whose first bar line serves mainly to mark the entry of the accompanying horn (Example 25.2). Then, as the melodic strands proliferate, the bar lines become (rather as in modern editions of sixteenth-century polyphony) a more or less arbitrary compromise between the conflicting demands of the individual parts. The “Sacrificial Dance” and the “Evocation of the Ancestors,” which also had its barring problems, are somewhat different cases. In the “Evocation” (Figure 25.1), the opening chord groups are barred in seven quarter-note beats in the sketches,² as 4 + 3 + 4 + 2 quarter-note beats in the autograph full score and autograph four-hand piano score, and as 4 + 3 + 2 + 3 + 1 quarter-note beats in the published full score.³ The seven-beat measures in the sketch imply a floating accent, with fairly strong downbeat stresses, weakened by the sustained pedal bass, which moves only on the half-accent of the second measure. The piano reduction splits the sevens into 4 + 3s but shortens the second group by two attacks, removing the natural end downbeat and leaving a rather curious unstressed ending hanging. The printed full score then rebars this shortened version so as to make the final attack a downbeat.

It is amazing how much difference these adjustments can make in a mindless performance, though, in fact, pianists will tend to play what they know rather than what they see and will probably think the groups as in an orchestral performance. But neither here nor in the “Sacrificial Dance” is there any issue of polymetrics. The problem lies within the monometrics. The opening section of the “Sacrificial Dance” (see Figure 25.2) is entirely based on varying distributions of a sixteenth-note unit and on varying weightings on individual chords (in the sketches the unit is the eighth note, and Stravinsky reverted to this when he revised the dance in 1943, but that was probably mainly for clarity). The variations are minute but say a lot about the music’s character. In the sketches (Figure 25.2a)⁴ and the published piano reduction (Figure 25.2b), Stravinsky starts with a five-unit measure followed by a three (indicated on Figure 25.2b by the arrow), whereas in the published orchestral score of 1921 and subsequent Édition Russe and Boosey & Hawkes reprints, this becomes 2 + 3 + 3 (indicated by the arrow on Figure 25.2c). But in the 1943 AMP revision he switched this around to a 3 + 2 + 3 (indicated by the arrow on Figure 25.2d), so that the extended upbeat chords that had ended the three-beat measures in the old score suddenly changed their relative position. The first one became a downbeat in the two-beat measure while the other remained an upbeat. At the same time he removed the contrast in the string parts between short single-unit chords and chords twice their length, that is, between sixteenths and eighths in the old score, which would have become eighths and quarters in 1943. In fact, in his own recorded performances he invariably ignored this distinction, as did Pierre Monteux—as far as one can tell, given the quality of the sound and the playing—in his 1929 recording, which presumably reflected his practice since the first performance. Other conductors, on the other hand, make the distinction very clearly, even sometimes exaggeratedly. It may seem a trivial point, but it is symptomatic of the rhythmic and articulative problems of a music so intricate in its metric detailing.

Figure 25.1. The Rite of Spring, Part II, “Evocation of the Ancestors”: (a) sketch for the opening (compare R-121:2 through R-122:1 in the printed score); (b) the opening in the autograph four-hand score; (c) the opening in the full score, R-121:4–8. © Copyright 1912, 1921 by Hawkes & Son (London) Ltd. Reprinted by kind permission of Boosey & Hawkes Music Publishers Ltd.

Figure 25.2. The Rite of Spring, Part II, “Sacrificial Dance”: (a) sketch for the opening (compare R-142:1 through R-143:4 in the printed score); (b) the opening in the autograph four-hand score; (c) the opening in the full score, R-142:2–5; (d) the opening in the 1943 revision. © Copyright 1912, 1921 by Hawkes & Son (London) Ltd. Reprinted by kind permission of Boosey & Hawkes Music Publishers Ltd.

This preoccupation with relative accents and groupings in a purely linear context—bearing in mind that the chords are not functional harmony in the conventional sense but vertical extensions of the line—is suggestive in all kinds of ways. To me it suggests not dance but song. It is almost as if there were a hidden text whose verbal accents were in some way operating on the orchestral rhythm. This is not such a fanciful idea, after all. In his next ballet, Svadebka (Les noces, or The Wedding), Stravinsky does introduce a text and does, precisely, play with its natural accents in order to create an intricate web of rhythmic implications, still essentially within a linear, monometric context.

Take the opening chorus, where the girls are plaiting Nastas'ina’s hair. The text runs as follows:

Chesu, pochesu Nastas'inu kosu,

Chesu, pochesu Timofeevny rusu,

A eshche pochesu,

A i kosu zapletu.⁵

[I am combing, I shall comb Nastas'ina’s braids,

I am combing, I shall comb Timofeevna’s blond hair,

and still I shall comb

and shall bind her plaits.] (Figure 25.3)

The score is barred initially as , and at first that is also how the music lies, with a loud downbeat chord on the first syllable of “chesu” and the half-measure marked by a (somewhat lighter) mezzo forte chord on the equivalent second syllable of “pochesu.” But, in fact, both “chesu” and “pochesu” would correctly be accented on the final syllable; this is precisely what happens in the next phrase, because Stravinsky honors the eleven syllables of the first line monometrically with a phrase of eleven eighth notes, so that the second line, to the same music, starts as an upbeat, and the metric accent now falls on the “su” in each case. However, the music has shifted back one eighth note, and the dynamic accent still falls on the “che” both times. The second line, on the other hand, has twelve syllables, but now Stravinsky engineers a downbeat start to line 3 by shortening the fourth measure from six to five eighth notes; hence the third line of the text starts with a downbeat and a dynamic accent on the weak “a” (a peculiar Russian cross between “and” and “but”). However, because of the foreshortening of the measure, this downbeat is probably heard as an accented upbeat, so that the foreshortening is delayed to the next measure, and the normally rather strong accent on the second syllable of “eshche” falls on the weakest possible second eighth note but with a dynamic accent supplied again by the mezzo forte.

The folk origins of these intricacies have been studied to a degree and with a precision far beyond my capacity by Richard Taruskin and Margarita Mazo. Here I am concerned only with the musical subtleties, particularly the way in which the text becomes involved in the process of metric and rhythmic variation. We are in the territory, needless to say, of the infamous “rejoicing discovery,” Stravinsky’s realization that the natural spoken accents of Russian folk poetry became mobile when sung.⁶ Taruskin has analyzed in great detail the consequences of this discovery in the songs Stravinsky composed just after The Rite of Spring and at the time that he was drafting Svadebka.⁷ In a number of these songs, Stravinsky exploits the variable accent and irregular meter in order to enrich a single melodic line of the monometric type. Taruskin suggests that such procedures were directly opposed to the kind of strict realism in which Stravinsky had been brought up: the realism of the kuchka (the Moguchaya Kuchka, the “Mighty Handful,” or the “Mighty Five” Russian composers, including Balakirev, Borodin, Cui, Mussorgsky, and Rimsky-Korsakov). He quotes César Cui on the duty of the vocal composer to respect the “natural” prosody of the text down to the smallest detail of enunciation and punctuation:

It is essential [Cui argued] that in quantity the music correspond to the dimensions of the poem, so that the music does not dangle on it like a gown on a hook, so that the text need not be artificially prolonged by repeating stanzas, verses, or individual words, and so that by such repetitions the artistic and elegant form of the poem be not distorted. It is essential that, in singing, the pronunciation of every word be suitably rendered, and that the phrasing of the text and the observance of its punctuation be correct. Besides that, the rhythm of the music and its meter must be in direct correspondence with the meter of the verse, the length of the musical phrase with the length of the text phrase, and, in fine, the music must in every way blend with the word so as to form with it one indissoluble, organic whole.⁸

And he cites Mussorgsky, specifically Boris Godunov, as a prime example of this philosophy in action.

But one can look at these matters from a slightly different angle. Mussorgsky did theorize about the virtues of exact declamation, and he studied the speech habits of the Russian peasantry when he was composing his opera Zhenit'ba (The marriage) directly from the text of Gogol’s play. The text setting in Zhenit'ba largely sticks to the natural accents. On the very first page of the vocal score, in the phrase “da takaia, nakonets, / skvernost' stanovitsia” (and in the end it’s such a horror), the words “nakonets” and “stanovitsia” are gently misaccentuated. Maybe the peasants at Shilovo had an odd way of talking; or, more likely, Mussorgsky was caricaturing the languid character and slovenly speech of Gogol’s antihero, Podkolesin. But there are songs by Mussorgsky that go well beyond realism in the sense of true declamation, for instance, “Darling Savishna,” in which a simpleton pleads with the village beauty to love him in an unbroken melodic flow of quarters and a quick meter; or in “Gathering Mushrooms,” again largely in rapid even quarter notes as the young mushroomer mutters to herself crossly about how she is going to poison her husband and his parents. This is more like psychological realism, with more than a trace of peasant musical dialect about it: Mussorgsky only occasionally misaccentuates (there are, in fact, examples in Boris as well: already, for example, in the opening chorus, where the bar lines by no means always correspond with the verbal accent); but he does frequently smooth out the strong tonic accent that Russian shares with English, and this is done as a musical and dramatic artifice. For instance, in “Darling Savishna,” the regular quarters are the even, manic speech of the intellectually disabled, raised to the level of an artistic image.

There is a subliminal example of this kind of artificial realism in Stravinsky too. It comes at the start of Petrushka, and it may well be the first example in his music of free monometrics, of meter controlled by the length of a figure built up from a single unit. The opening material of the ballet is derived from the cries of street vendors (Example 25.3a). The flute theme is supposed to be the cry of a coal, or charcoal, vendor, and the cello theme at R-1 is that of an apple merchant. Both melodies are strictly unbarred, though the ostinato figure—and of course the ordinary requirements of the situation—impose on them a strict meter, which, in fact, they simply ignore. The cries are in a variant of the form that Messiaen later called accent-désinence—an initial figure repeated more or less ad libitum, followed by a final flourish (Stravinsky usually lacks Messiaen’s anacrouse—his upbeat figure). Repetitions of the initial figure might vary in either number or length. In the charcoal theme, it comes three times, then twice with the flourish, and then several more times, slightly varied but always cutting across the triple meter. The apple cry concentrates more on the désinence, whose natural meter is actually duple. When it comes around a second time it starts on a different beat of the measure and at a different point in the charcoal cry, which it accompanies both times (Example 25.3b). If Stravinsky had stopped to think about it, or if the music had been much trickier, he might already have struggled to write it down, as he later said about the “Sacrificial Dance.” In fact, he did apparently have trouble with the next passage (Example 25.3c). In its original version the “Song of the Volochobniki” stomps along in duple meter against septuplet eighth notes in the upper woodwind and all kinds of textural undergrowth in the middle of the orchestra, the whole thing still nominally in time. In 1946 he renotated the septuplets and simplified the scoring at this point, but it still looks a bit of a muddle.

Example 25.3. Petrushka, 1911 version, Tableau 1: (a) “The Shrovetide Fair,” (a) m. 1 to R-1:6; (b) “The Shrovetide Fair,” R-2:4–8; (c) “Song of the Volochobniki,” R-3:1–5. © Copyright 1912 by Hawkes & Son (London) Ltd. Revised version © Copyright 1948 by Hawkes & Son (London) Ltd. Reprinted by kind permission of Boosey & Hawkes Music Publishers Ltd.

Example 25.4. The Rite of Spring, Part I, “Ritual of Abduction”: (a) R-37:3–5; (b) R-40:2–5. © Copyright 1912, 1921 by Hawkes & Son (London) Ltd. Reprinted by kind permission of Boosey & Hawkes Music Publishers Ltd.

Which brings us back to The Rite of Spring. The opening of the second sequence in the ballet, the “Igra umykaniia,” or “Ritual of Abduction,” is itself a form of abduction. The woodwind motive (actually derived from another folksong in Juszkiewicz) is nearly identical with the coal vendor’s cry in Petrushka, and at the same pitch (Example 25.4a). Later in the dance, the first horn bellows out the same desperate cry in fourths, at the original pitch, but in inversion. One would like to see a production of The Rite with street vendors; it would be an effect worthy of Pirandello. But the musical point here is that this figure again flows across the supposed meter—a meter that is, in fact, a pure abstraction: none of the orchestral sections are playing in at the start. A few measures later, the figure comes back, is curtailed, and then is varied with a backward shift of one eighth (Example 25.4b). Needless to say, the meter remains unchanged, though exactly how it maintains its hold is a good question, since most of the orchestra is still either ignoring it or obeying it only in spasms.

The Rite of Spring has many different ways of denying the kind of regular meter that one might have supposed necessary for dancing. The “Ritual of Abduction” ignores its supposed meter; the “Mysterious Circles of the Young Girls” simply tailors the meter to its own mobile phrase lengths: 6 + 7 + 4 + 5, with a two-beat extension. It would be easy enough to redesign this melody so that it fits into a regular quadruple or sextuple meter, which is probably what Rimsky-Korsakov or, say, Vaughan Williams would have done. But for Stravinsky it was the additive aspect of the tune that gave it its rhythmic character by throwing it in and out of synchrony with the regular eighth-note groupings in the cellos and basses. The first episode of the “Sacrificial Dance” is an intriguing variation on this pattern. A single-unit chord (1) and a unit rest (0) alternate irregularly in a series of binary patterns: 11010, 1110, 1010, 110, 1110, 10110, and so on (Example 25.5). Other figures supervene, but essentially the episode is controlled by this irregular scheme of binaries. Barring is not strictly a problem here, since the texture is homophonic and the silences are unanimous. But the score is absurdly hard to read because Stravinsky beams the groups across the bar lines and makes no clear distinction between the eighth rests and the articulative sixteenth rests that are rhythmically part of the sounding chords. Many of the bar lines could simply be done away with. At bottom, this is the purest imaginable application of a monometric principle, with a single-unit pulse, varied group lengths, and no variations of pitch, several years before Stravinsky seems to have consciously experimented with the idea in his wartime sketches.

One might expect that Neoclassicism, with its regulated metric structures, would kill off a method of this kind. But of course these regulated structures are more symbol than process, and there are actually countless instances of monometric themes in Stravinsky’s work post-1920. One of the sketches for the “Cinq pièces monométriques” (see Example 25.6a) turns up as the fugato theme in the first movement of the Octet, nicely arpeggiated in C major but barred in eighth-note beats as 1 + 3 [= 4] + 6 + 3 + 4 + 4 + 3 (see Example 25.6b). The main first-movement theme of the Piano Sonata looks like a monometric line that has been tidied up to resemble a Bach fugue subject. The second movement (“Eclogue I”) of Duo Concertant is unbarred for its first two pages, presumably in acknowledgment of the improvisatory character of the kazachok dance upon which it is modeled. The main theme of the Symphony in C (Example 25.7) is essentially a meterless eighth-note melody forced into what looks like, but is not, a tonal phrase structure. It would be hard to imagine a better example of a habit of thought formed in one musical style turning up in disguise in another.

This is perhaps to say no more than that Stravinsky was consistent in his background method, regardless of changes in style, which is a commonplace of Stravinsky criticism these days. But there is one way in which he converted the single-unit idea that has less to do with style and more to do with process. One example of this is well known: in the Symphonies of Wind Instruments, from the time when Stravinsky first put it together from the loose sketches, the tempi were regulated proportionally in multiples of thirty-six, so that in theory a single underlying value controlled the music from first note to last. But similar procedures can also be found in the Neoclassical period. They are partly a symptom of Stravinsky’s notorious fear of the performer-interpreter, but there are artistic considerations at work here as well.

Example 25.6a. “Cinq pièces monométriques,” sketch. Igor Stravinsky Collection, Paul Sacher Foundation, Basel.

Example 25.6b. Octet at R-19. © Copyright 1924 by Hawkes & Son (London) Ltd. Revised version © Copyright 1952 by Hawkes & Son (London) Ltd. Reprinted by kind permission of Boosey & Hawkes Music Publishers Ltd.

In Oedipus Rex, for instance, there is plenty of evidence that he wanted to fix the music’s motion, just as he wanted to fix the characters behind masks and their expressivity behind the Latin text. In Dialogues and a Diary he writes about “the geometry of tragedy” and the “‘fatal development’ that, for me, is the meaning of the play.”⁹ One aspect of this tragic geometry is expressed, if that is the right word, through a network of precise tempo relationships. Several scenes are at least partly governed by proportional tempi, but one scene in particular, the scene of the Messenger, in which Oedipus at last realizes the truth, is entirely controlled in this way. The base tempo is ♪ = 96 for the entrance of the Messenger and his announcement of the death of King Polybus; this becomes ♪ = 192 (eighth note equals sixteenth note) for the chorus’s realization that Polybus was not Oedipus’s father and for the following section, in which the Messenger tells of the finding of the child Oedipus on a mountainside. Next, the Shepherd sings his haunting ranz des vaches (an Alpine horn melody played by Swiss herdsmen when driving their cattle to or from pasture): it would have been better to stay silent, he warns, and the tempo switches via ♩. = 63 to ♩ = 63—the closest mark on the old metronome scale to the strictly accurate 64, a third of the previous value. This is a true piece of metric modulation, with the eighth-note value slowed down by a controlled process of equivalence. Oedipus’s “Nonne monstrum” then modulates back to ♪ = 192, but with a metric modulation from to , after which the chorus proclaims his crime, in tempo but with quarter notes and triplet quarters (= 96 and = 63). Finally, light dawns: “Lux facta est,” the quarter note now equal to three of the preceding eighths. These underlying values are maintained through many changes of meter and very different figurations, so there is little surface sense of rigidity. The dramatic idea of bondage, the inescapability of fate, is latent. We may sense it, but we are unlikely to hear it, not least because few conductors actually observe these modulations with any precision, if at all. Even Stravinsky does not do so, unwilling, perhaps, to be trapped—like Oedipus—by his own past.

Example 25.7. Symphony in C, first movement at R-5, melody. Copyright © 1940 by Schott Music GmbH & Co. Kg, Mainz, Germany, worldwide rights except the United Kingdom, Ireland, Australia, Canada, South Africa, and all so-called reversionary rights territories where copyright is held jointly by Schott Music GmbH & Co. KG, Mainz, Germany, and Chester Music Ltd., London. Copyright © renewed. All rights reserved.

What, one may ask, is the connection between metric modulation and monometrics? The answer might be that both suggest metric thinking based on unit values. This is blatantly the case with the Symphonies of Wind Instruments, more discreetly so with the more regularly barred Oedipus Rex. In the scene just discussed, the listener is most conscious of the eighth-note unit in the Messenger’s “Reppereram in monte puerum,” with its unsettled alternation of duple and triple meter. But Stravinsky ensures eighth-note consciousness elsewhere by a persistent kind of mechanical figuration. This is already prominent in the very first scene of the opera-oratorio, where the controlling value is ♪ = 150, and the schema is spoiled only by an unmeasured meno mosso at Oedipus’s “Uxoris frater mittitur.” In the Messenger scene, Stravinsky draws attention to Oedipus’s arrogant stupidity by the mindless eighth-note arpeggios that accompany his last desperate piece of bluster in “Nonne monstrum.” But the underlying tragic geometry embraces everyone, Messenger, Shepherd, and People, as well as the doomed king himself. Should one see the Chosen One of The Rite of Spring in the same light? She, too, is the prisoner of a mechanistic fate that drives her remorselessly to her death. But so too are her co-tribalists, the Old and the Wise, the Painted Girls, and the rest. It’s a grim thought that the beautiful impersonality of folk music should lend itself to the disasters of fate. But there’s the twentieth century for you.

Notes

1. Stravinsky and Craft, Expositions and Developments, 141.

2. Stravinsky, The Rite of Spring: Sketches, 73.

3. Compare Pieter C. van den Toorn’s essay in this volume.

4. Stravinsky, The Rite of Spring: Sketches, 86.

5. The most recent (2005) Chester Music edition of Svadebka employs a pragmatic approach to transliteration that deviates from the practices employed in this book, so spelling in Figure 25.3 does not match the present discussion.

6. Stravinsky and Craft, Expositions and Developments, 121.

7. See Taruskin, Stravinsky and the Russian Traditions, 2:1145–1236, passim.

8. Ibid., 2:1200, translated from César Cui, “Neskol'ko slov o sovremennykh opernykh formakh” [A few words about contemporary opera forms] (1889), in Izbrannye stat'i, 406–408.

9. Stravinsky and Craft, Dialogues and a Diary, 24.