Volume 12, Number 2, May 2006
Copyright � 2006 Society for Music Theory

David H. Smyth*

Review of Stravinsky's Histoire du soldat: A Facsimile of the Sketches, edited by Maureen A. Carr (Middleton, Wisconsin: A-R Editions, 2005)


KEYWORDS: Stravinsky, Histoire du soldat, Carr, facsimile, sketches, Smyth

Received May 2006

[1] The appearance of a facsimile edition of Stravinsky's sketches for Histoire du soldat is cause for rejoicing. Editor Maureen Carr is to be congratulated for obtaining permission to publish sketches held in two Swiss collections (the Igor Stravinsky Collection at the Paul Sacher Stiftung in Basel, and the Rychenberg Stiftung at the Stadtbibliothek Winterthur); Chester Music of London also granted permission for the undertaking. In all, more than 250 sketch leaves appear in this large-format (approximately 9 by 12 inches) publication from A-R Editions. The editor has provided an introductory essay (pp. 3-19) and a pair of tables in which the sketches are identified insofar as possible, locating the corresponding passages in the 1987 reprint of the full score. There are also three short essays by Ramuz scholars Philippe Girard, Alain Rochat, and No�l Cordonier about the evolution of the text; each is provided in French and in English translation. The quality of the photographic reproductions is quite good, and scholars should be pleased that the price of the volume is not exorbitant.

[2] Carr's introductory essay explains that, like many of the works of Stravinsky's early period, Histoire du soldat has a complex history. She provides a brief account of the circumstances surrounding the collaboration between Stravinsky and Charles-Ferdinand Ramuz during the years of the composer's Swiss exile, a concise summary of the plot, excerpts from various reviews of performances, and tables listing dates inscribed in the score used at the first performance and publication information about later editions. A discussion of "Criteria for Establishing the Ordering of the Sketches" (pp. 10-11) and "The Musical Origins of Histoire du soldat" (pp. 11-12) provides a start toward the challenging task of untangling the web of source materials. Carr also discusses examples of rebarring in a series of sketches for the "March du soldat" (pp. 13-16) and explores questions of motivic linkage among the "march" movements (p. 17), thus providing some snapshots of the kinds of discoveries that await the careful reader of Stravinsky's sketches.

[3] Questions of the compositional chronology of Histoire remain vexing. As was previously noted by Robert Craft, important thematic ideas for the work originated in sketches made for a setting of "Antony and Cleopatra," a collaboration with André Gide that Stravinsky abandoned after several months of work in 1917.(1) Stravinsky apparently first introduced the subject of Histoire to Ramuz during the winter of that year. According to Richard Taruskin, the first clear reference to their collaboration comes in a letter from Stravinsky dated 28 February 1918.(2) Ramuz wrote a three-page synopsis of the story with the provisional title "Le soldat et le diable" on the very same day.(3) By 6 April, Stravinsky had completed a version of the Music to Scene One ("Petits airs au bord du ruisseau"). Composition progressed rapidly, and the work was premiered on 28 September 1918 in Lausanne.

[4] However, the first performance was (for a time) the last. Because of an outbreak of Spanish influenza, to which the composer and various members of the cast succumbed, only a single performance took place. Ramuz, an inveterate reviser, continued to make modifications to the text, and Stravinsky also took advantage of the opportunity to improve upon his score before the work was next performed (as a concert suite) on 20 July 1920 in London. During the same year, an abbreviated arrangement of the work (for clarinet, violin, and piano) was published, and another (for piano solo) followed in 1922. According to Craft, still more changes were made before the full score of the work was published in 1924 by J. W. Chester of London.(4)

[5] Even this cursory account of the genesis of Histoire is not without its problems. Unfortunately, most sources contain contradictions, errors, and omissions. For example, Craft's account of the revisions has appeared in two slightly varied versions.(5) In the latter, he states that "Stravinsky began revising the score soon after this unique performance [the premiere on 28 September 1918], until only the Scene by the Brook and the Tango remained unchanged." He further informs us that both the Lausanne premiere and the 1920 London performance of the concert suite were conducted by Ansermet from a full score copied by Stravinsky's wife Catherine, into which numerous corrections were later added.(6) Presumably, this is the "conductor's score" mentioned by Carr, currently located at the Stadtbibliothek Winterthur, from which she provides her table of the dates "as written by Stravinsky"(p. 4). A comparison of the table to the corresponding facsimile pages confirms that these dates correspond exactly to the ones on the sketch leaves--down to the occasional inclusion of both Old Style and New Style dating (the Gregorian and Julian calendars were 13 days out of phase at this time, and Stravinsky sometimes used both). Many of the dates are accompanied by the composer's signature, some with the designation "Morges," the location of the family dwelling at the time the sketches were completed. This leaves us wondering whether Catherine's copy of the score was annotated with dates from the sketches by the composer, or whether Craft is incorrect about the hand in which this score is written. Craft also refers to another score, requested by Stravinsky for a performance in Paris (7 November 1923), which had been "used by Scherchen and myself. . . [and] has many more corrections than the one you sent me." Craft says that score was in use in Leipzig at that time.(7) Carr writes "The copyist's manuscript of 1920, with annotations by Stravinsky and others, is missing. This manuscript was used as a source for the Chester editions of 1987 and 1924" (p. 19, note 68). It is unclear, in short, exactly how many manuscript copies of the score exist, and who wrote them.

[6] This is but one of the many areas in which Stravinsky scholars will require additional documentary evidence in order to use the composer's sketches to their fullest. In this regard, Carr's book represents both a great step forward and a frustrating failure. On the one hand, it provides access to a large body of primary source material that has never before been readily available. Although Carr admits the possibility that some sketches may have become separated from the ones reproduced here (p. 9), she claims the edition contains "all the existing sketches for Histoire du soldat" (p. 17). On the other hand, issues of chronology and interpretation are only minimally addressed. There is no thorough account of the physical properties of the sketches. The dimensions of each sketch leaf are given, but there is no description of paper types, the physical makeup of the two bound sketchbooks, and exact locations of the various pages within Stravinsky's other bound sketchbooks are sometimes vague. Regrettably, there is no concordance with the catalogs of either the Paul Sacher Stiftung or the Rychenberg Stiftung. Those who wish to consult the original documents will need to use those finding aids on site.

[7] Let us make a brief tour of the sketches in order to clarify what this facsimile edition might enable us to learn. The sketches are divided into eight groups of varying sizes, and as Carr explains (pp. 10-11), it is very difficult to establish a strictly chronological ordering among them. However, some general conclusions could have been drawn. Some of the sketches in Group One are surely among the earliest, because they include instrumentation indications that reflect a different ensemble than that used in Histoire. These are assumed (by Craft and Carr) to belong to Stravinsky's work on "Antony and Cleopatra," and would thus date from 1917. Others, however, clearly relate to much later efforts. For example, Figure 1.30 is a short score version of the Little Chorale--the very last movement to be completed (the date in the "conductor's score" is 23 September 1918, a mere five days before the premiere). Throughout this group we find early versions of a number of the work's most prominent and memorable themes: the chromatic melody from Rehearsal 10 of the Soldier's March makes several appearances (in Figure 1.1, it is scored for three trumpets; Figures 1.5 and 1.10 include less fully developed versions), and the quintessential rhythmic motive from the Music to Scene One (five bars after Rehearsal 4) appears in Figure 1.13, and so on. There are significant bits of the Tango, the Ragtime, and the Waltz, as well as recognizable bits of the Triumphal March of the Devil, the Devil's Dance, and the Royal March. It is possible that some of these sketches may relate to revisions made after the first performance. For example, Figures 1.35 through 1.41 relate to the Triumphant March of the Devil, the instrumentation of which Stravinsky was still revising for performances in 1923.(8) What is crucial to know about this group is that all of these sketch leaves (excepting the last four) are on the same type of paper, all of uniform (or very nearly uniform) size, with six holes punched in the margin. They were probably kept in a binder, or in several binders. It is highly unlikely that the order in which they are presented here is the order in which they were written. Placing them in chronological order may require some very careful reconstruction.

[8] The Group Two sketches are equally fascinating--and equally mysterious. Most of these sketches were written on the backs and in the margins of heavy sheets of varying sizes cut from an old Italian ledger or account book. Again, there are bits from various movements, and several leaves are dated: the transition from the Tango to the Waltz (Figure 2.21, double-dated 16-3 July 1918), the ending of the Little Concert (Figure 2.30, dated 10 August 1918), and the end of the Triumphant March of the Devil (Figure 2.39, dated 26 August 1918). Miscellaneous sketches for the narration during the reprise of the Soldier's March, the Devil's Song, and a six-page draft of the Devil's Dance are not dated. The ordering of these loose leaves is again impossible to determine absolutely without more careful study.

[9] Group Three consists of sketches from a small vellum-bound sketchbook, on the cover of which Stravinsky wrote "Soldat croquis 1918." The leaves are of a uniform paper type, and of nearly uniform size. In several cases, Stravinsky's writing continues across an opening--suggesting that the writing took place after the leaves were bound. But the sketchbook may have been rebound at some point: Figure 3.35 (the last leaf on which musical sketches appear) is written upside down, while Figure 3.36 (the verso of the same leaf) is right side up. The latter leaf is dated 1 May 1918, and bears the fascinating inscription in Italian "parent of my first aria" above an early version of the "Princess Motive." Stravinsky himself noted a resemblance between this theme and the Dies irae, although he said the chant was not on his mind when he initially invented the theme--and the sketches corroborate his claim: the earlier sketch version of the theme is indeed less closely related to the chant melody than the revision.(9) Here again, a detailed description of the stitching, the gatherings, and the condition of the covers would be welcome. Occasionally, it is impossible to read some of the fainter pencil entries on some of the facsimile pages. Figure 3.7 is an example: Carr finds what she believes to be possible early sketches for the Great Chorale here, but I cannot make out enough notes to confirm her hypothesis. The notation I can read looks like that used in the Tango (as in Figure 3.35). Among the most interesting entries are several leaves devoted to work on the Great Chorale (Figures 3.28 through 3.33). Here one can study some primitive versions of the melody and bass lines, multiple attempts to fashion the movement's cadences, and some fascinating changes in the details of scoring and meter.(10) However, as we shall see, Craft is incorrect in claiming that the entire chorale is to be found here.(11)

[10] Group Four includes sketches from another bound sketchbook, the so-called "Winterthur Sketchbook," now located at the Rychenberg Stiftung. Once again, only the barest of physical description is supplied, leaving us to wonder about many aspects of the book's condition. Carr has discovered that a series of pages was at one point removed (they were presented to Stravinsky's patroness Eugenia Errazuriz on 6 February 1920). These leaves are now at the Paul Sacher Stiftung in Basel. In Carr's edition, they appear in facsimile in their rightful place, reunited with the Winterthur Sketchbook. The sketchbook begins with a sketches and a draft of the Music to Scene One, the end of which is dated 6 April 1918. This is the earliest ending date for any of the movements of the work, but clearly these sketches are more highly developed than the primitive ones in Group One. The next section of the (reconstructed) sketchbook consists of sketches for the Ragtime and Waltz. The end of the Ragtime is dated 15 August 1918 (Figure 4a.14); the end of the Waltz bears no date. Another set of sketches for the Great Chorale follows. Here, Stravinsky includes timings for the spoken texts, as well as several important revisions to the central phrases, and (for the first time), several drafts of the final phrases. A few details remain to be changed before the published version is reached, but clearly these sketches post-date those of the "Croquis" book. The next batch of sketches develops the Royal March--which is initially titled "Pasa-doble" (Figure 4.28). Another lengthy run of pages contains sketches and drafts for the Soldier's March, the end of which is dated 25 June 1918. This would suggest that the March was completed before the Ragtime, which precedes it in this sketchbook. Together with the evidence that various groups of pages are separately paginated (in what I believe to be Stravinsky's hand), this strongly implies that the Winterthur Sketchbook was rebound after its original use, or that its components were first gathered for binding only after they were written. Undated sketches and drafts for the Tango follow, after which the sketchbook concludes with a few jottings relating to the Music to Scene Two and a portrait sketch of Ren� Auberjonois dated 1917 (Figure 4.85). This last is conclusive evidence that the ordering of the contents of the book as it now exists has little to do with the chronology of their creation.

[11] Group Five consists of five leaves of uniform size, which Carr informs us were once bound together. The cover and second leaf bear no musical sketches. The remaining leaves contain references to motives from the Music to Scene Two interspersed with unidentified fragments. Some look to me a bit like the cimbalom notation Stravinsky employed in Renard and the Ragtime for Eleven Instruments, and indeed, the abbreviation "Cimb" appears twice in Figure 5.10. These sketches do not look like they relate directly to the Ragtime for Eleven, but one should remember that in fact, that work was completely sketched (although not orchestrated) by 21 March 1918--that is, before work on Histoire had progressed beyond its earliest phase.(12)

[12] Group Six consists of miscellaneous early sketches from what is known as Sketchbook III (there are ten numbered sketchbooks in the Igor Stravinsky collection at the Sacher Stiftung). The four pages of sketches reproduced in Carr's edition are all early ones for the Music to Scene Two; all are in keyboard texture, and interestingly, the first is explicitly labeled "piano" in Russian (Figure 6.1). These tentative sketches, and a listing of instruments in Figure 1.6 (seven players, but with piano listed instead of bassoon!), suggest that Stravinsky considered an alternative to the famous Histoire ensemble.(13)

[13] Group Seven consists of two sketches from leaves filed with Sketchbook IV. Carr is unable to identify the first (as am I); she suggests that the "Valse" fragment in Figure 7.2 might or might not be associated with the Waltz in Histoire. I tend to think it is not.

[14] The final ten sketches (Group Eight) are somewhat easier to identify. Part of the first is nearly identical to material in Figure 5.7, and both clearly relate to the Music to Scene Two. However, the sketch immediately above this is marked for clarinetto piccolo in E-flat, and probably has nothing whatever to do with Histoire. Other ideas here are tentative experiments that are difficult to identify with any degree of certainty, although most of Carr's attempts seem plausible. The entries she does not identify are of considerable interest: some may reflect paths not taken, while others could be jottings related to other works. Ideally, one would want to consider the content of pages before and after the ones reproduced here, for it is the contextual relationships that are often most valuable in assessing his entries in these sketchbooks, which tend to be quite discontinuous. In general, Stravinsky seems to have used these books to set down initial ideas; when he was ready to do more sustained sketching, he often moved to a different format.

[15] As noted above, Carr provides two tables in which the contents of the sketch leaves are identified. The first is organized by Group, and follows the order in which the facsimile pages are laid out. The second is arranged to correlate with the 1987 edition of the full score, so that one can readily locate all of the sketches that correspond with any given movement in the work. Both tables include translations of most (but not all) of Stravinsky's notations in various languages. Unfortunately, a number of the instrumental designations are not mentioned--and they can be important. For example, in Figure 8.1, the sketch Carr believes may be one of the earliest versions of the opening of the Music to Scene Two is marked "cl. La" (clarinet in A), whereas the published version is for B-flat clarinet; the Russian expressions at the bottom of this leaf are not translated. Here and elsewhere, designations for instruments that do not belong to the Histoire ensemble (piccolo, clarinet in E-flat, horns, cimbalom) are sometimes passed over. There are a number of trivial errors in the book: for example, on p. 11, Erik Satie's name is misspelled, and on p. 47, the French term for snare drum is misspelled. A more significant error is the wrong date at the end of Musical Example 1 (p. 9), which should read 12 May (not 12 March).

[16] Still, let the rejoicing continue. Carr's edition provides an opportunity to become acquainted with the complexities of Stravinsky's creative process, and to spend many hours at the composer's elbow (as it were). With some patience, one can compare and contrast numerous versions of the intricately interwoven themes and motives of Histoire, and thus gain a greater appreciation for the work's special integrity. Along with Carr's sketch tables, the essays of the literary scholars provide a basic understanding of the fabulously complex process of rewriting and revision both Ramuz and Stravinsky undertook as the work grew. Of course, access to the original documents will still be necessary in order to answer many of the harder questions, if only because no facsimile can convey the full essence of an original document. Differences in paper type, shades of coloration in the writing, erasures, and many other details jump out at one when the original document is viewed, whereas these features may not be visible at all in a facsimile. Even so, Stravinsky fans should celebrate the appearance of this publication. We hope it will be only one among many more, and that its appearance will foster close study and deeper appreciation for the composer's works and his working methods.

List of Works Cited

Carr, Maureen, editor, 2005. Stravinsky's Histoire du soldat: A Facsimile of the Sketches. Middleton, Wisconsin: A-R Editions.

Craft, Robert, 1980. "Histoire du soldat: The Musical Revisions, the Sketches, and the Evolution of the Libretto." Musical Quarterly 66: 321-328.

Craft, Robert, 1985. "Histoire du soldat: The Musical Revisions, the Sketches, and the Evolution of the Libretto." Appendix A in Stravinsky: Selected Correspondence, vol. III. New York: Alfred A. Knopf. [The musical examples here are typeset, whereas in the previous publication, many appeared in reduced facsimile.]

Stravinsky, Igor, 1962. Igor Stravinsky: an Autobiography. New York and London: W. W. Norton and Company. [originally published 1936]

Stravinsky, Igor, and Robert Craft, 1981. Expositions and Developments. Berkeley and Los Angeles: University of California Press. [originally published 1960]

Taruskin, Richard, 1996. Stravinsky and the Russian Traditions: A Biography of the Works Through Mavra. Berkeley and Los Angeles: University of California Press.

White, Eric Walter, 1979. Stravinsky: The Composer and His Works. Berkeley and Los Angeles: University of California Press.

David H. Smyth
Louisiana State University
School of Music
Baton Rouge, LA 70803


1. Craft (1980) and Craft (1985).
Return to text

2.Taruskin (1996), p. 1294, note 69.
Return to text

3. Alain Rochat in Carr (2005), p. 32. One wonders whether the three-page scenario listed in the provisional catalog of the Igor Stravinsky Collection at the Sacher Stiftung may be a copy of this same document, and whether some of the annotations may be in Stravinsky's hand.
Return to text

4. Craft (1985), p. 461, note 1.
Return to text

5. Craft (1980) and Craft (1985).
Return to text

6. Craft (1985), p. 461.
Return to text

7. Ibid.
Return to text

8. Craft (1985), p. 461.
Return to text

9. Stravinsky (1981), p. 92.
Return to text

10. My colleague Donald Traut and I are at work on a study of the analytical implications of these and other sketches for the Great Chorale, which both of us have studied at the Paul Sacher Foundation in Basel.
Return to text

11. Craft (1985), p. 469.
Return to text

12. White (1979), p. 275. The instrumental score of the Ragtime for Eleven is inscribed 11 November 1918, at 11:00 AM--the precise moment of the signing of the Armistice.
Return to text

13. See Stravinsky (1962), pp. 71-72 for the composer's fascinating remarks regarding the impossibility of including a piano in this ensemble, and compare Taruskin's cranky insistence that no jazz band worthy of the name could possibly do without one (Taruskin 1996, p. 1302).
Return to text

Prepared by Brent Yorgason, Managing Editor
Updated 24 November 2006