Parallel paths: historical-documentary and analytical contributions as a basis for the performance of Debussy’s Syrinx

by Luisa Curinga

1. Introduction
2. First path: historical-documentary investigation
3. Second path: the examination of analytical contributions
4. Conclusions

1. Introduction

Syrinx, a short and extraordinary piece, is one of the most important in flute literature. Its preeminent position in the historical repertoire for this instrument is essentially due to two factors: it is the first really significant piece for solo flute after the Sonata in A min composed by C. P. E. Bach exactly 150 years before (1763), and it is the first solo composition for the modern Böhm flute, perfected in 1847. The importance of Syrinx from a strictly musical point of view, is also testified by the great number of analyses dedicated to this piece. For a player – as in my case – Syrinx offers a valuable opportunity to compare the performing suggestions supplied by analytical contributions with other suggestions concerning the relations between the musical and the poetic text (Syrinx was conceived as incidental music), the comparison between the single existing autograph manuscript and the first printed edition, and the testimonies on particular performing aspects of famous historical flute-players who worked in contact with Debussy.
On the grounds of this double cognitive path – analytical and documentary – a methodological comparison between the two approaches would therefore seem appropriate, in order to verify if these parallel routes present any analogies or produce different results, and, above all, if they are able to provide, in the diversity of their contributions, a global understanding of the piece, which is able to have a concrete influence on the performing choices.

2. First path: historical-documentary investigation

Debussy conceived several theatrical projects which never were finished with Gabriel Mourey, the eclectic figure of playwright, novelist, poet and translator. [1] In 1912 Mourey, catching the recurring interest of Debussy for classical Greek mythology, asked the musician to compose the incidental music for his play in verse Psyché, but the project, probably conceived as a melodrama, was realized only in a minimal part; in fact, when on 1st December 1913 the work was due to be performed, Debussy had only written Syrinx [MOUREY 1913]. [2]
According to Debussy’s biographer Léon Vallas, Syrinx represents the last song of Pan before his death; on account of the great authoritativeness of Vallas, such an assertion has mainly been considered uncritically correct, with the result that a dramatic character has been attributed to the piece which has influenced the vision of some performers and some analysts [VALLAS 1926, 1927, 1932, 1944, 1958].
Actually, as already suspected by certain more thorough scholars and confirmed by an autograph manuscript kept in Brussels and published in fac-simile in 1992, Syrinx was to be played not on the occasion of the death of Pan – where the verses do not suggest any space for music – but in the first scene of the third act, a moment of intense lyricism in which the atmosphere is full of sensuousness and eroticism. [3]
In this scene the music accompanies the dialogue between two nymphs: a Naiad, who has never met Pan, is afraid of him and wants to escape, and another nymph, an Oread, who tries in vain to persuade the first of the miraculous and charming splendor of the God’s music, which provokes love and a sense of ardent Oneness with the universe in all those who hear it.
Only a few notes – that diffuse during a warm night, full of brilliant stars – are sufficient to change completely the attitude of the Naiad, who can not help but become inebriated and overcome, like her fellows, by the love of the God and who expresses her disturbance by describing in the meanwhile the voluptuous reactions of the other nymphs. [4]
When Debussy delivered the music for the performance, he gave instructions that it should be played off stage, behind the scenes; Pan was inside his cave and must remain invisible. The dedicatee of Syrinx, the flautist Louis Fleury, who also gave the première, jealously held onto the manuscript and made the piece famous by playing it often, both in France and abroad, always with great success. [5] Just to reproduce the conditions of the first performance, it seems that he always insisted on having a small curtain behind which to play Le flûte de Pan [BOPP 1983, 265]; this was in fact the title of the composition and when in 1927, after Fleury’s death, the publisher Jobert could at last print the piece, he preferred to call it Syrinx, since Le flûte de Pan was also the title of the first of the three Chansons de Bilitis, for voice and piano, and it was possible to cause confusion between the two. [VALLAS 1958, 360]
Jobert obtained the manuscript from Fleury’s widow and asked the famous flautist and teacher Marcel Moyse, who had surely given a private performance of the piece in presence of Debussy, to prepare it for publication. [WYE 1993, 68-69] The piece seems originally to have been without bar-lines, which were added in order to not frighten amateur flautists who wanted to approach the piece; Moyse also added some breathing marks, and perhaps some slurs.
The manuscript from which Jobert derived the first edition is unfortunately lost and the Brussels manuscript presents some differences with respect to the first edition, differences that regard breathing, dynamics and agogics, and are therefore very significant in connection with interpretation. For this reason they will now be examined in detail. [6]
In his scores for wind instruments, Debussy did not mark every breath, but only those musically most meaningful; in fact in the Br. ms. there are only three breathing marks, at the end of m. 2 and in the middle of mm. 4 and 14, while in the Jobert edition there are as many as eighteen breaths. Although it is a revision, the number of breaths that Moyse indicated could seem excessive, but there is an explanation: the flautist had a personal breathing problem contracted in his childhood, and so prepared the edition to match his own limits. The breaths he marked became conventional, given that all the several editions that followed one another over the years (until the Swedish one of 1992 which instead is based on the Br. ms.) from this point of view slavishly follow the Jobert version. Moreover it seems that Moyse refused to allow his name to appear as editor of the piece, declaring that he had only clarified some inconsistencies [WYE 1993, 68-69]; this may have induced subsequent editors and flute-players to think the breathing marks of the Jobert were directly by Debussy. Fortunately, even though many years later, Moyse revealed the real will of the composer: the breaths at the end of mm. 16 and 25 were added by Moyse because of his own personal difficulties, but Debussy disapproved them [WYE 1994, 5]. Therefore, although they are now in the performing praxis of the majority of flautists, they should be avoided.

Audio Example 1, mm. 14-17:
A. with the breath at the end of m. 16 (Ed. Jobert)
B. without breath (ms. Br.)

Regarding the breath mark in m. 25, very important for its position between the end of animant peu à peu and the repeat, Moyse admitted: “Debussy me demanda de ne pas respirer à cet endroit, mais j’en étais incapable […] Pour moi c’est impossible”. [7] By avoiding this breath a continuity in tension is produced between the animando part and the repeat of the initial motif, a tension increased by the lengthening of the B flat.

Audio Example 2, mm. 22-26:
A. with the breath in m. 25 (Ed. Jobert)
B. without breath (ms. Br.)

Again according to Moyse, Debussy greatly appreciated the breaths at the end of mm. 28 and 29 [WYE 1994, 5].
The last important difference concerning the breath marks is in m. 31: in the Jobert there is a breath after the first beat, while in the Br. ms. the phrase is not interrupted and the two B flats are tied.

Audio Example 3, mm. 29-33:
A. with breath (Ed. Jobert)
B. without breath (ms. Br.)

From the above, it appears that Debussy desired a more relaxed, uninterrupted melodic movement, and a more fluid and less fragmentary performance. It should also be remembered that Fleury, the dedicatee of the piece, was famous for his ability to play very long phrases [8].
Other differences between the Jobert and the Br. ms. regard some dynamic suggestions. The Jobert lacks the diminuendo in m. 31 and the p at the beginning of mm. 5 and 17, which appears on the contrary in the Br. ms.; in Debussy’s orchestral and piano music, the sign) is frequent, and actually indicates a subito piano; it should therefore be interpreted in the same way in Syrinx.
The most important difference however concerns the accent on the first beat of the penultimate bar, which has always caused a lot of doubts among flautists, as it seems inconsistent with the tendency in smorzando of the last bars of the piece. When the reviser of a later edition, in 1968, asked Moyse about this accent, he replied that he believed it to be a mistake, and it should be a diminuendo [DEBUSSY 1992, 5]: in this form, in fact, it matches the Br. ms.

Audio Example 4, mm. 33-35:
A. with the accent in m. 35 (Ed. Jobert)
B. with diminuendo (ms. Br.)

An examination of the Br. ms. also supplies an important indication about the agogics: the accelerando which begins in m. 22 and which though not indicated in the Jobert edition is played by practically all flautists, and whose validity has been confirmed by various analysts, is definitively ratified by the autograph manuscript, where one can read en animant peu à peu. Moreover in the Jobert there is no trace of the pause sign at the end of m. 8, a very significant sign because in that silence, as marked in the Br. ms., the Oread had to say her line Tais-toi, contiens ta joie, écoute. [9]
Other useful contributions for the performance are the testimonies of the flautists who worked with Debussy, as well as those handed down through their students. Not only Moyse, but also Fleury commented on this matter, confirming Debussy’s preferences among other things for long-breathing phrases: “a lament was exactly what the composer had to express […]; he had a longbreathed phrase, he employs the lower octave, he indulges in no temperamental explosion, he confines himself to the severest and soberest expression of great mental suffering”. [FLEURY 1992, 390]
A testimony more specifically about performance comes to us from the German flautist Paul Krauß, who knew Debussy directly, being a member of the orchestra of the Paris Opera and flute-soloist in one of the Parisian Symphonic orchestras before the First World War. Through his student Joseph Bopp, he reports the importance Debussy gave to rhythmic accuracy in performing Syrinx; in particular it seems that Debussy was very demanding about the 32nd-notes in m. 13, often played very freely, almost as if they were 16th notes [BOPP 1983, 266]. This is perfectly in line with other similar assertions of musicians who worked with Debussy, regarding the attention he paid to the rhythmically precise performance of his music. For example, one can see the testimony of the piano-player Marguerite Long, or that of the conductor Ernest Ansermet who, according to Bopp, harshly reproached a flautist who took too many liberties in performing the solo in the Prélude à l’après-midi d’un faune [LONG 1960; BOPP 1983, 267].
Bopp also informs us about the metronome tempos that, as his teacher told him, Debussy would have liked: Très Modéré would correspond to , never faster. After m. 9, when the tempo becomes poco più mosso, the 8th notes should be equal to 92.

3. Second path: the examination of analytical contributions

The historical-documentary investigation made it possible to acquire important information about Syrinx’s literary referent, a valuable aid in placing it in the right atmosphere (nocturnal, sensual, lyric, mythological), and to obtain explanations about very concrete aspects such as the tendency of melodic line suggested by the breathing marks, dynamics and agogics. However, the very nature of this otherwise valid research makes it appear lacking in deep theoretical analysis, an aspect that is instead supplied by the analysts. I obviously do not intend to offer here a complete examination of the rich analytical literature on Syrinx, but I shall follow a cognitive path that, by basing itself on the analysts’ work, will attempt to shed new light on the piece, in particular from the viewpoint of the relationship between analysis and performance. [WHITMAN 1977, 68-104]
The main problem in such an investigation is that, of the numerous analyses of Syrinx, only the one by Ernestine Whitman, a researcher who takes on the double role of flute-player and analyst, sets out with the intention of correlating analysis and performance. The other analysts are not concerned with the possible repercussions of their work on performance, but it is nevertheless possible to obtain useful performing suggestions from these analyses, even if indirectly. The passage from theory to practice is, of course, in this case the result of the deductive activity of the performer, who reads and interprets the analyses.
In outlining a path which lends itself to the purpose of increasing our knowledge and aiding the performer, I shall make use of some essential points of reference, taken from a letter written by Debussy to Mourey on November 17th 1913, just a few days before the date of the performance of Psyché, with Syrinx inside.

17 Novembre ‘13
Mon cher Mourey,
Jusqu’à ce jour je n’ai pas encore trouvé ce qu’il faut…
Pour la raison, qu’un flute chantant sur l’horizon doit contenir tout de suite son emotion! Je veux dire, qu’on a pas le temps de s’y reprendre, à plusiere fois, et que: tout artifice devient grossier, la ligne du dessin mélodique ne pourrant compter sur aucune interruption de couleur, secourable. Dites moi, je vous prie, Très exactement, les vers après lesquels la musique intervient?
Après de nombreux essasis je crois qu’il faut s’en tenir à la seule flûte de Pan, sans autre accompagnement. C’est plus difficile, mais plus (logique – crossed in the autograph) dans la nature.

Claude Debussy [10]

The letter clearly expresses the nature of Debussy’s inspiration and intentions, which can be synthesized and schematized as follows:

- a flute singing on the horizon must at once contain all its emotion

- there is no time for repetitions

- any artifice becomes coarse
- Pan has to play alone, without any accompaniment, because it is more in the nature

- the line of the melodic contour cannot count on any interruption of color

According to Debussy, points 2, 3 and 4 are consequences, or rather extensions of the first. Which of these aspects have been covered by the analysts, who most probably did not know this letter, and how can their investigations help the performer? To know this, each point will need to be examined, being fully aware, though, that the different elements constituting a piece are, in the composer’s mind and, hopefully, in that of the performer too, an indivisible whole; here the elements have of course been split up for the sake of explanation.

1) What did Debussy mean when he asserted that a flute singing on the horizon, therefore in the distance and hidden – we know Pan was invisible in his cave and the sound had to spread, so to speak, without any body – must contain all its emotion? Probably that the flute has the task of representing all the emotions Pan had within him; and exactly because the expression is contained, it is impossible for it to unfold without any control. Translated into musical language this could mean avoiding great contrasts, an attitude reflected, in my opinion, in the dynamics marks. This aspect has been generally neglected by the analysts of Syrinx, with the almost sole exception of Whitman, who places side by side analysis and performance. [WHITMAN 1977, 68-104]
Her original work is all based on the hypothesis that Syrinx is constructed on two main motives, X and Y; X appears in the first two measures, Y in the last two; her analysis points out the subsequent and gradual mutations from X to Y, mutations that, according to Whitman, reflect the mythological tale of the change of the nymph Syrinx into a reed:

Fig. 1

Different intervals, registers, dynamics, and rhythmic figures are associated with each motive (fig.1), and in the performance any proposal of one of the two motives must be differentiated by highlighting its own features. As far as the dynamics are concerned, X, associated with the middle register, has a basic level of mf or p crescendo; Y, in the low register, is more static, with softer dynamics and smoother rhythms. Whitman suggests playing the parts characterized by X, which is more animated, warmly and with more vibrato than those dominated by Y.
Essentially Whitman recommends the performer to follow Debussy’s dynamic marks with great care; she underlines the importance of the subito p in measure 4, as it highlights the difference between measures 4 and 5. The figure in the first beat of mm. 4 and 5 is in fact identical with regard to pitch and rhythm, but while in m. 4 the crescendo mark is followed by a subito piano, in m. 5 the figure is always in crescendo; the tension does not decrease, but rather prepares for the dynamic increase of the ascending triplets which end the measure (see fig. 2 and audio example 5).

Fig. 2

Audio example 5, mm. 4-5

Whitman complains about the fact that many flautists do not pay due attention to the crescendo and play the triplets in the second half of m. 5 with a diminuendo, whereas a crescendo leads to the upper octave (C flat in m. 6) in a much more convincing way than a decrescendo. [11] Moreover, the C flat must be louder than the E flat (as is written), so that the tonal shift towards B flat can be heard.
Whitman also suggests emphasizing the dynamic difference between mm. 17 and 19, preceded by two identical measures (mm. 16 and 18). The decrescendo in m. 17 leads to the D flat, the crescendo in m. 18 leads to the E flat, a note which, according to Whitman, sets up the transition back to X, and which Carol K. Baron defines as a startlingly fresh sound [BARON 1982, 124].
Almost all the analysts place the climax of the piece in m. 27. In the light of the results of the analysts’ work, I believe that the expressive peak should be characterized both timbrically as well as from the point of view of the tension, which is carried continuously from the previous measures en animant peu à peu, but the dynamics have the task of containing the emotion: in fact, even at this emotionally more intense moment, Debussy marks nothing more than a sober mf.
James Tenney too deals with dynamic aspect: among the several parameters necessary to obtain a temporal segmentation from a gestalt perceptual viewpoint, he includes intensity and timbre, and he investigates the importance of these two aspects in the structure of the piece. The conclusion is that in Syrinx (unlike, for example, in Varèse’s Density 21.5) intensity does not represent a perceptually structural element, and the piece in fact moves in a rather limited dynamic range, p through mf. [12] The control of the emotions can occur through other aspects, such as that of intervals. Almost all the analysts have noticed how the melody in Syrinx moves mainly by neighbor notes or by small skips: semitones, whole tones, minor thirds. In the first measure the whole-tone movement in the accented beats offsets the chromatic and fast movement in the unstressed beats.
When an intervallic expansion occurs, it is gradual and there are some distinctive intervals, such as the tritone, minor third, and minor sixth, whose return takes on a role of melodic support.
The analysis by Cogan and Escot examines Syrinx in terms of the spatial-linguistic transformation of a primary linguistic cell (B flat, E, D flat) [COGAN- ESCOT 1976, 92-101]. The two scholars identify a series of intervals – defined by the number of semitones they are composed of – and they measure the linguistic transformation on the basis of the appearance or the disappearance of these intervals. It is difficult to imagine that such an analysis, rather abstract, could influence the performance, except in the sense of general awareness, which is in any case useful for the performer. And yet it is actually possible to extract some particular performing suggestions from such highly theoretic analyses, although obviously again indirectly. Cogan and Escot believe that m. 3 is both a reminiscence of the phrase in mm. 1-4 and a point of connection between m. 3 and mm. 6-8. Figure 3 shows how m. 3, of which only the incipit has been reproduced in the first stave on account of its reminiscent character, is connected through the broken line to the C flat in m. 6 on the second stave. In fact, according to two theorists, the phrase contained in mm. 4-8 does not linearly descend from the B flat in m. 3, but linearly ascends from the B flat in m. 4. The linear ascent is subsequently carried on in two different registers: the arrival on the C flat in m. 8 has its starting point in the low B flat in mm. 4-5 and, in the other register, in the B flat in m. 3.
Of course a performer wishing to underline the reminiscent character of m. 3 could play it differently from the first measure by modifying, for example, the color of the sound, or keeping the dynamics at a slightly lower level. This would also allow the ascending line from mm. 4-5 to the C flat in m. 6 to be better highlighted.

Fig. 3

Cogan and Escot are not the only ones to recognize transformation as the main characteristic of Syrinx; even though with different sense and results, due to the diversity of their objectives and methodologies, Whitman, as already mentioned, Baron and Ulrich Mahlert are of the same opinion. Baron considers the transformation from the whole-tone scale beginning on C (Whole Tones 1) to the one beginning on C sharp (Whole Tones 2) to be fundamental in Syrinx; she also highlights the great importance in this passage of pentatonic conFig.tions, which substitute in a non-tonal way the classical modulating bridge.

Fig. 4 [13]

According to Mahlert [1986, 195], the arabesque in mm. 1-2 is transformed, in a way that is only seemingly casual, into that of mm. 13-14 (fig. 4), and the triplets figures too are in a certain way a derivation of the first two measures (fig. 5). Given that these figures come from all the sections of the piece [14], Mahlert observes that the intervallic transformation involves the whole development of the piece.

Fig. 5

Figure 5 shows how the triplet figure, which forms the basis for the three conclusive measures, measures considered as variations of a model X and referred to in the table with the signs X1, X2, X3, recalls the figure in m.16, that appears in section IV, but also presents a plurality of further and meaningful relations, especially concerning following moments, when both the formal variants (particularly the second) of the triplet figure are introduced. The forerunner of the triplet figure in m. 31, the original form of model X (where the mutations play on two elements: the appoggiatura and the change in the succession of sounds), can be traced to the end of section I, corresponding to the first line of the table, for this reason marked as Schluss. The sign Schluss in the third line corresponds instead to the end of section III. As regards the two final lines of the table, Mahlert explains how the two fragments pinpoint the reminiscences evoked by X3, whose transformational character they underline, where Anfang and Mitte represent the beginning and the middle of the musical connections. If the sign Anfang in the penultimate line needs no explanation, it should nevertheless be specified that the Mitte in the last line is referred to the connective role of the major ninth of m. 19 (D flat-B flat), between the two parts of section IV. Mahlert notes how this interval, rich in tension, is changed at the end of the piece into the more relaxing interval of a second (E flat - B flat). [15]

2) Debussy says that in Syrinx there is no time to propose further repetitions, and this impossibility to let oneself freely go, without worries of duration, is of course a limitation, a “containment”; the piece can be nothing else but short, being constrained by the temporal margins set by the poetic text. And yet, at least at first sight, Debussy’s assertion “There’s no time for repetition” can appear rather curious given that the piece, especially in the motivic analysis, seems to be made up of short incisi incessantly repeated and varied. This very repetitiveness is one of the aspects of Syrinx that has attracted researchers who deal with music in relationship with other cultural areas: semiologists, psychologists, computer scientists. [ANAGNOSTOPOULOU 1997; DELIÈGE 1987; LASKE 1984; NATTIEZ 1975; TENNEY-POLANSKY 1980] Being short, monodic, and built on a few elements simple to recognize and often repeated, the piece is in fact easy to schematize.
On a macroformal level, almost all the analysts agree in recognizing a tripartite division of the piece. Considerable differences are however to be found in the segmentation, given that to consider the same section as A1 or B implies, of course, completely different analytical interpretations. As shown in Table A, the main analyses of Syrinx present remarkable divergences in the formal segmentation of the piece; this is obviously due to the different approaches and to the various analytical methodologies used, but it also makes one think that probably the best thing would be to perform the piece as a single unbroken discourse, a long fluid phrase in which divisions are instrumental, useful to give the player a mental picture, but to be highlighted the least possible in the performance. [AUSTIN 1966, 7-15; BORRIS 1969, 173-174, DE NATALE 1996, 33-40 and 133-135; LARSON 1990, 1-15]
This is confirmed by Fleury, when he speaks about a ‘long-breathed’ phrase, and by the testimony of Moyse on the breath in m. 25, which Debussy asked him to avoid: therefore, even if where there is clearly a ripresa it must not lead to any fall in tension, and it seems legitimate to consider the piece as an unbroken melodic line.

Table A

It is interesting to note how the segmentations of Tenney are the most different from the others; his division is in fact based on temporal perceptual gestalt-units, in which above all the “sections” – units on the higher hierarchical level – but also the “segments” – on a lower hierarchical level – are often not at the beginning of the measures, but occur for example on a levare. Certainly they are perceptually significant points, but quite dissimilar from the divisions of the other analysts (see Figure 6 which shows the first twenty measures of Syrinx in the segmentation of Tenney).

Fig. 6

If one accepts Tenney’s conclusions, one must take into consideration that the segmentation of a piece is often, from an aesthesic point of view, completely different from the common formal divisions. Nattiez points out how all the analysts agree to isolate the first eight bars, explaining the reason for this choice with the convergence of the criteria used: “tempo omogeneo; linea di battuta; omogeneità dovuta, se non a una definita ‘tonalità’, almeno alla nota polare Si bem; sviluppo di un piccolo numero di temi paradigmatici” [NATTIEZ 1975, it. trans. p. 101]. This uniformity of results (with the sole exception of Larson, who makes the first part of the piece end at m. 15), also coincides with the indications of the score in its original scenic aspect too, even if Syrinx was performed in this version probably only the first time.
In the theatrical version, in fact, the acting intervenes at the end of m. 8, with the Oread who, without any accompaniment, says her lines Tais-toi, contiens ta joie, écoute; there the music really breaks off, and the only sure formal division is actually that at the end of m.8. In this case the conclusions of the historian and the analyst agree, and I think it is absolutely acceptable to make a short pause at that point, as the pause mark in the Br. ms. suggests.

3) The need for naturalness has been dealt with with great sensibility by Ulrich Mahlert: he considers the arabesque the paradigm of Debussy’s way of composing, and Syrinx one of the most significant examples of his modus operandi [1986, 182]. Debussy mentions it several times in his writings: with reference to Bach, he exalts it in opposition to the by then stiff and academically encrusted classical schemes [DEBUSSY 1987, 34-35, 65-66, 228-230, 245-247]. Thanks to the arabesque, which has an open structure and derives from a decorative, ornamental and Fig.tive conception, that of Jugendstil, music can overcome the mechanical stiffness and achieve that “free” art, the result of the spiritual correspondence between art and nature, at which the French composer aimed.
The arabesque, intended as a force of nature, has its own organic development, which, according to Mahlert, is exemplarily illustrated in the two first measures of Syrinx: the force of gravity, which is a passive force, is, in a subtly proportioned way, balanced by a corresponding ascending and active force (Figure 7). [16]

Fig. 7

Mahlert also notes how, as a moment of a melodic arabesque, the return of the initial note in the second measure is a sign of the balancing of proportions between the two complementary forces.
I believe it is interesting to express in performance this sense of the two forces, active and passive, but how is this really possible? Mahlert offers no suggestions, but I think that the passive parts, those with the tendency to fall downwards, could be played gently, trying to not to stress or highlight any note, remaining always strictly in time and supporting the natural timbre of the flute when it descends towards the low register, in order to express this sense of gravitational attraction. This suggestion can be applied very well to mm. 2, 13 and 14.
Mahlert underlines in each of the five sections into which he has divided the piece (see note 14), the melodic contour, the arabesque line (fig. 8), an arabesque which fades following organic laws, to the extent that the last section with its summing-up character and wider intervallic ambitus, has no true close and ends in an indeterminate way.

Fig. 8

I suggest that this tendency to fade should be reflected in the performance; and Whitman too, while starting from different assumptions from those of Mahert, proposes something similar when she suggests playing the last two measures, made up of motive Y with a static character, adopting softer dynamics, in diminuendo, with a fairly ‘closed’ sound and no vibrato.
The arabesque is for Mahlert not the only sign of the naturalness of Syrinx; he also mentions two other elements: the descending fourth and the pentatonic scales. The descending fourth, which is very evident in m.8 (E flat-B flat), as well as at other points ( for example in m. 4 : B flat- F; B-F dur), is par excellence the “sound of nature”, with its organic structure, it is the Naturlaut of Mahler’s 1st Symphony, the divine and sacred melody of Pan, coming from cut reeds and intended to imitate the birds singing, it is by definition Ursprunge, a symbol of Nature’s cry; the fourth in m. 8 is anticipated by the enharmonic change in m. 5, silent, secret, mysterious, that is not part of an enharmonic process. In fact the C flat appears for the first time suspended in time and in sound, since the figure that contains it lies between two rests. [MAHLERT 1986, 193]

4) Regarding the melodic line we have already spoken of the arabesque and of intervals. It is more difficult to associate the melodic line to timbre, because definitions like “the importance of the color of the sound” risk being generic, and being tailored to fit more than one piece or, worse still, more than one composer. [17] The difficulty of using a verbal language to describe subjective concepts like that of color is undeniable, since they really and exclusively belong to spheres of hearing and emotion; all the same it is equally undeniable that timbre and color are among the main structural elements in Syrinx, as Debussy himself testifies in his letter to Mourey.
Analysts have been generally more concerned about defining entities connected with formal structures, intervals, scales, but some of them have actually dealt with the problem of color. We have already seen how Whitman offers precise instructions on the subject, for example when she says that motives X and Y, which have different characters, should be differentiated in timbre, or when she suggests playing the apex in m. 27 with a warm and very vibrato sound, or the two last measures with a closed timbre and without vibrato. Tenney and Polansky too deal with the problem of timbre, but even though they are well aware of the importance of this element, they are quite familiar with the difficulty of using its features concretely in an analysis. In a comparison of several pieces including Syrinx and Varèse’s Density 21.5, they consider four parameters (see Tab B).

  Duration Pitch Intensity Timbre
Debussy   1.0  1.5  2.0  0.0
Varèse   1.0 0.67 6.0 20.0

Table B

 It was possible for them to attribute measurable values to the first three parameters (duration, pitch, intensity): an input weight of 1.0 implies a time-unit of one-tenth of a second, a unit-interval of one semitone, or of one dynamic-level difference (e.g. between mf and f). As regards timbre, though, of which even the definition escapes (Tenney writes about “one more parameter, which we call ‘timbre’”) [TENNEY-POLANSKY 1980, 219], the values cannot be specified and not even approximated in order to be processed by the computer program. For this reason Tenney informs us that in the diagram the parameter timbre has been only used in a very primitive way, with scale values of either 0 or 1, and only to represent the “key-clicks” in mm. 24-28 of Varèse’s Density 21.5.
Also Marco De Natale is well aware of the importance of the sound color in Syrinx, but he touches on the problem rather than tackling it; he limits himself to making us note, with an observation that is only seemingly banal, how the piece could not be proposed on any other instrument, such as the piano. De Natale also believes that timbre in Syrinx becomes a sort of theme, and that it is closely linked to the sense of movement, given that the softness of sound counterbalances the rhythmic throbbing. Unfortunately, though, these interesting observations are not further studied in depth.

4. Conclusions

We have seen the way in which the analysts have covered the main aspects of Syrinx and how their analytical findings can aid the performer. Generally the topics dealt with, deduced from the letter of Debussy, also cover the set of musical problems tackled in the analyses, except for the problem of tonality: some analysts (Austin, Seraphin, Larson, De Natale) have in fact attempted to find the tonality or the tonalities of the piece, or at least to search for one or more tonal centers, while others (Mahlert, Baron, Cogan) consider the piece to be free from tonality, and suggest that any affinities are only temporary and seeming, underlining how Debussy carefully avoids all tonal resolutions.
At the beginning of the this article I wondered if the two methodological paths would lead to similar results or not; although in some cases significant coincidences have been verified, generally the analogies are only a few. In fact the main results of the historical-documentary investigation concern breaths, dynamics, agogics, and the relationship between the music and the poetic text: the examination of analytical literature reveals how the analysts have barely touched upon these questions.
Breaths, essential for marking the rhythm of musical speech, in that they underline moments of continuity or discontinuity, have not been an object of the theorists’ attention. The only one who writes about them is Nattiez, who cites all the breaths in the Jobert wrongly calling them “pauses” and asserts they are fundamental for the segmentation of the piece [NATTIEZ 1975, 340; it. trans. p. 100]. Actually we have seen that the breaths marked in the Jobert derive, at least in some cases, from Moyse’s own needs; therefore to rely on them could lead one astray.
We have seen that only Whitman has dealt widely with the dynamic aspect, and with regard to the agogics only De Natale notices how considerable differences can be found between the various performance of the piece. [18] Finally, not knowing the real relation between the music and the poetic text has led in same cases to misrepresentations which have influenced the vision of the piece; some analysts (for example Whitman) have been inspired, at least as a starting point, by the myth of the nymph Syrinx’s mutation into a reed, when in reality the title Syrinx certainly was not by Debussy and appears for the first time in the Jobert edition; others (for example Mahlert) have seen in some compositional structures, such as chromaticism, a reflection of the last mournful melody of the dying Pan, but we have seen in the first part of this paper that Syrinx is definitely not played on this occasion and has, on the contrary, a lyrical and sensuous character.
Furthermore, not knowing the real authorship of the breathing marks in the Jobert edition, has led people to assume them as conventional and the work of Debussy, as in the case of Nattiez.
The analytical investigation has, on the other hand, touched on structural aspects and has developed in depth problems that the documentary research is not able to tackle, succeeding in covering, in a more or less convincing way, all the elements Debussy suggested were fundamental in Syrinx.
If on the one hand we could lament an overall lack of interest in performing problems on the part of the analysts, on the other it should be born in mind that in most cases the objectives they set – how the piece “works”, what the peculiar characteristics of its style are and so on – are different from those set by the performers. And it is also necessary to make another consideration: if the overall knowledge of the piece and its structure is essential for an enlightened performance, the choice between the various performing possibilities is actually up to the single player, that is, to his/her sensibility and technical and expressive skill. After all, the double approach – historical-documentary and analytical – seems able to produce interesting results, for the very reason that it is supported by different and often “complementary” competences; the two kinds of survey do not overlap, but rather complement each other, leading to a global knowledge of the piece and, not uncommonly, to valuable performing suggestions.


[1] Gabriel Mourey (1865-1943) translated among others Edgar Allan Poe and Algernon Charles Swinburne. Moreover he was a contributor to the “Revue wagnérienne” and acted as intermediary in the encounter between D’Annunzio and Debussy for Le Martyre de Saint-Sébastien. The theatrical projects never finished in cooperation with Debussy are: L’embarquement pour ailleurs, “commentaire symphonique”, 1891; Histoire de Tristan, “drame lirique”, 1907-09; Huon de Bordeaux, 1909; Le Chat botté, by La Fontaine, 1909; Le marchand de rêves, 1909.
[2] The play deals with the myth of Psyche in the version of Apuleio’s Metamorfosi; in the third act the story of the death of Pan, based on Plutarch, is also introduced.
[3] With regard to this aspect, see SERAPHIN 1964 and BOPP 1983, 265-267.
The Brussels manuscript is owned privately, by Mrs. Hollanders de Ouderaen, and is published in fac-simile in DEBUSSY 1992. In the Swedish edition, unfortunately still little known, the editor A. J. Chapelon gives no information about the provenience of the manuscript. He only says that it is probably a copy of the lost manuscript source of the Jobert edition. Therefore we can deduce that the Brussels autograph copy was not the source of the Jobert.
[4] Unlike Syrinx, the atmosphere in the Prèlude à l’après-midi d’un faune was that of a sunny afternoon.
[5] It’s possible, even if we have no documents which testify it, that Fleury had some sort of rights over the piece, given that he appears to be the only public performer of Syrinx until the publication of the Jobert edition. It is also possible that Debussy decided to give the piece to his friend Fleury, and therefore did not insist on an immediate publication.
[6] From now on I refer to the Brussels manuscript as the Br. ms..
[7] “I remember Debussy asking me not to breathe here, but I couldn’t do it. […] For me it’s not possible”. [WYE 1994, 5]
[8] A. L. Chapelon, introduction to DEBUSSY 1992.
[9] “Keep silent, contain your joy, listen”. These verses are not the only ones mentioned in the Br. ms.; there are also those which immediately precede the entry of the music: les astres et les dieux? / Mais voici que Pan de sa flûte recommence à jouer. [“the stars and the gods? But here is Pan beginning to play his flute again”].
[10] “Dear Mourey, so far I have not found what is needed… because a flute singing on the horizon must at once contain its emotion! That is, there is no time for repetitions, and exaggerated artificialness will coarsen the expression since the line or melodic pattern cannot rely on any interruption of color. Please tell me, very precisely, after what lines the music starts. After several attempts I think that one has to stick to the Pan flute alone without any accompaniment. This is more difficult but more (logical /crossed out in the autograph) in the nature”. Letter by Claude Debussy to Gabriel Mourey of November, 17th 1913. The original autograph is kept in Frederick R. Koch Collection, Beinecke Rare Books and Manuscript Library, Yale University. A fac-simile is published in DEBUSSY, 1992.
[11] With regard to the cadence figure in mm. 6-8, Whitman notices how important the length of the notes is, in order to establish the static nature of the Y material; therefore flautists must not rush these measures, as they often do.
[12] Tenney carries out an analysis of Syrinx from the point of view of temporal gestaltic perception, using a computer analysis program written by Polansky. [1980]
[13] Legenda:

[14] Mahlert divides Syrinx into five sections:

I section mm. 1-2
II section mm. 3-8
III section mm. 9-15
IV section mm. 16-25
V section mm. 26-35

[15] This kind of analysis is inspired by that of Nicolas Ruwet [1983] who superimposes similar melodic lines to recognize any derivations. It is interesting to note how in the diagram of Mahlert, unlike Ruwet who considers the derivations chronologically throughout the development of a piece, transformations are not consecutive, but are inserted apart from their temporal succession. In confirmation of this, it is interesting to observe that Mahlert believes the last phrase to be linked to the beginning and to the middle of Syrinx, thus transforming the time flow, so that the piece is perceived in a sort of state of temporal suspension.
[16] Similar arabesques are often found at the beginning of other compositions by Debussy; the philosopher Jankelevitch also mentions them, considering them a phenomenon of “geotropism”. With regard to the use of the arabesque in Syrinx he says: Dans Syrinx, pour flûte seule, la cantilène ravisseuse plane, tournoie, enroule ses fantasques triolets, puis fond en piqué du haut des airs comme pour capturer une proie.. / [“In Syrinx for solo flute, the greedy lament glides, circles, twins in fanciful triplets and then , from the height of the air, nosedives as to capture a prey”]. [JANKELEVITCH 1949, 91]. The tendency of the melody to wander and oscillate has been also noticed by Marco De Natale, who, concerning the melody of Syrinx, speaks about a sliding, fluttering movement, on the boundary with improvisation [DE NATALE, 1996, 34 and 37]. On the question of the fluidity of the arabesque in Debussy, see also Spampinato [2000, 33-55].
[17] This definition, referring to Syrinx, is by Austin [1966, 9].
[18] De Natale poses the question, suggesting the utility of a performing analysis, but he supplies no answers; moreover the books in which he deals with Syrinx make up a teaching method, and the problem of agogics is introduced in a questionnaire for students.


AUSTIN W. W. (1966), Music in the Twentieth Century, Norton, New York, pp. 7-15.

ANAGNOSTOPOULOU C. (1997), “Lexical cohesion in linguistic and musical discours”, in Third triennal ESCOM Conference: Proceedings, ESCOM, Liège, pp. 655-660 (contiene un’analisi semantica di Syrinx)

BARON C. K. (1982), “Varèse’s explications of Debussy’s Syrinx in Density 21.5: a secret model revealed”, The Music Review, XLIII/2, pp. 121-134.

BOPP J. (1983), “Syrinx von Claude Debussy”, Tibia, 1, pp. 265-267.

BORRIS S. (1969), “Claude Debussy, "Syrinx" für Soloflöte”, Musik und Bildung, 4, pp. 173-174.

COGAN R. - ESCOT P. O. (1976), Sonic Design. The Nature of Sound and Music, Prentice-Hall, Englewood Cliffs, N. J.

DEBUSSY C. (1987), Monsieur Croche et autres écrits, Gallimard, Paris; ed. orig. 1921.

DEBUSSY C. (1992), La flûte de Pan ou Syrinx, Autographus Musicus, Stockholm.

DELIÈGE I. (1987), “Le parallélisme, support d’une analyse auditive de la musique: vers un modèle de parcours cognitifs de l’information musicale”, Analyse Musicale, 6, pp. 73-78 (contiene un’analisi di Syrinx, dal punto di vista dei rapporti tra livelli di informazione e memorizzazione).

DE NATALE M. (1996), Analisi musicale in itinere. Questionari e soluzioni per lo sviluppo dell’intuizione analitica, Ricordi, Milano.

FLEURY L. (1922), “The flute and his power of expression”, Music and Letters, 3, pp. 383-393.

JANKELEVITCH V. (1949), Debussy et le mystère, Ed. de la Baconniere, Neuchâtel; trad. it.: Debussy e il mistero, Il Mulino, Bologna, 1989.

LARSON J. A. (1990), Flute without accompaniment: works from Debussy: Syrinx (1913) to Varèse: Density 21.5 (1936), DMA doc., Univ. of Maryland.

LASKE O. (1984), “Keith: a rule system for making music-analytical discoveries”, in Baroni M. - Callegari L. (cur.), Atti del convegno Musical Grammars and Computer analysis, Olschki, Firenze, pp. 165-199 (contiene un’analisi di Syrinx ottenuta tramite un sistema computazionale di intelligenza artificiale).

LONG M. (1960), Avec Claude Debussy au piano, René Julliard, Paris.

MAHLERT U. (1986), “Die Göttliche Arabesque: zu Debussys Syrinx”, Archiv für Musikwissenschaft, XLIII/3, pp. 181-200.

MOUREY G. (1913), Psyché, poème dramatique en trois actes, Mercure de France, Paris.

NATTIEZ J.-J. (1975), Fondements d’une sémiologie de la musique, Union Générale d’Editions, Paris, pp. 330-354; trad. ital. leggermente modificata con il titolo: “Syrinx di Debussy: un’analisi paradigmatica”, in Marconi L. - Stefani G. (cur.), Il senso in musica, CLUEB, Bologna, 1987, pp. 93-137.

RUWET N. (1983), "Metodi di analisi in musicologia", in N. Ruwet, Linguaggio, musica, poesia, Einaudi, Torino, pp. 86-119.

SERAPHIN H. (1964), Debussys Kammermusikwerke der mittleren Schaffenszeit, Bärenreiter, Kassel.

SPAMPINATO F. (2000), "La poetica dell'acqua in Debussy", Diastema, 1 (14), pp. 33-55.

TENNEY J. - POLANSKY L. (1980), “Temporal gestalt perception in music”, Journal of Music Theory, XXIV/2, pp. 205-241.

VALLAS L. (1926), Claude Debussy, Librairie Plon, Paris.

VALLAS L. (1927), Les idées de Claude Debussy, musicien français, Ed. Musicales de la Librairie de France, Paris.

VALLAS L. (1932), Claude Debussy et son temps, Librairie Felix Alcan, Paris.

VALLAS L. (1944), Achille-Claude Debussy, Presses Universitaires de France, Paris; trad. it.: Guanda, Modena, 1952.

VALLAS L. (1958), Claude Debussy et son temps, nuova ed. con pubblicazione delle lettere, Albin Michel, Paris.

WHITMAN E. (1977), Analysis and performance critique of Debussy’s flute works, DMA diss., Univ. of Wisconsin.

WYE T. (1993), Marcel Moyse, an extraordinary man, Winzer Press, Iowa.

WYE T. (1994), note esplicative a C. Debussy, Syrinx or la Flûte de Pan, Novello, London.

< previous | index | next >

Editor: Egidio Pozzi
Scientific Committee
: Mario Baroni, Rossana Dalmonte, Susanna Pasticci, Giorgio Sanguinetti
Editorial Staff: Alessandro Bratus, Antonio Cascelli, Fabio Cifariello Ciardi, Catello Gallotti, Giuseppe Sellari
Webmaster & webdesigner: Giacomo Sciommeri