Chapter Six: In the Shadows of Les Nuits d’été

When the second edition of these songs was published, in 1849, the word mélodie had become widely understood in its purely musical sense: the phrase “avec accompagnement de piano,” earlier useful, was now superfluous. Still, when he wrote to the publisher of the orchestral version, in 1856, Berlioz spoke of his “délicieuses romances.”3 CG 5:252. A leading French dictionary takes Berlioz’s Irish Melodies of 1830 as the “point of departure” for the new genre of the mélodie.4 Honegger, Dictionnaire de la musique, “mélodie.” The path leads, ten years later, to Les Nuits d’été.

“Ten years later” takes us to 1840, the year during which these small compositions were conceived, although we cannot date Les Nuits d’été with calendrical exactitude. In his correspondence of that year, Berlioz does not mention the collection. The autograph manuscripts of the songs are not dated. We know only that the titles “Absence” and “Le Spectre de la rose,” the eventual fourth and second numbers of the cycle, appeared in the Revue et Gazette musicale on Thursday, November 5, 1840, as items on the program of the concert, sponsored by that journal, to be given on the following Sunday, November 8, 1840. But when the same periodical appeared on Sunday morning, these two songs did not figure on the printed program. From the detailed reviews of this concert that appeared in the Revue et Gazette musicale itself, and in the elegant magazine L’Artiste, reviews that discuss every item on the printed program of November 8, we can assert with confidence that the songs were not performed.5 Revue et Gazette musicale (November 12, 1840); L’Artiste (November 1840): 316.

It has been suggested that they were not performed because they were not yet written, but that is not the case, as we shall see. Indeed, there is one piece of evidence that suggests that Berlioz had indeed begun to set a series of poems by Théophile Gautier as early as March 1840: an autograph fair copy of the eventual first song of the cycle, “Villanelle,” now preserved in Darmstadt, in the Hessische Landesbibliothek.6 NBE 25 [2nd ed.]:308. This manuscript is signed and carefully dated “Paris, 23 mars 1840.” It was reproduced some two and a half years later by the Allgemeine musikalische Zeitung, on November 16, 1842, in the journal’s intermittent series of facsimiles of composers’ manuscripts—with Berlioz’s signature, but without the date.7 Beilage no. 8 zur Allgemeinen musikalischen Zeitung (1842); Facsimile der Handschrift von H. Berlioz. Berlioz was not in the habit of putting precise dates on his autograph manuscripts; when he did so, it was sometimes after the fact, and inaccurately. The date on the Darmstadt manuscript of March 23, 1840, looks suspicious, as I shall explain, as though it were added, to commemorate something important, at the time the composer sent the manuscript to Carl Ferdinand Becker, editor of the Allgemeine musikalische Zeitung, in 1842.

Becker had good reason to reproduce Berlioz in November of that year: in the issue of the 16th, he printed an excerpt from Karl Gutzkow’s Briefe aus Paris, which had just appeared, apparently with a splash.8 Gutzkow, Briefe aus Paris; Dresch, Gutzkow et la jeune Allemagne. Gutzkow was of course one of the important members of the Junges Deutschland movement, one of the more faithful painters of the Parisian scene, along with his better-known contemporaries Ludwig Börne and Heinrich Heine, and one of Wagner’s associates in Dresden in the eighteen-forties. In this excerpt he writes of the dramatic contrast he finds between the mind and the music of Hektor Berlioz:

In 1842, Becker may have sent to Berlioz the paper he was to use to make the copy of “Villanelle” that appeared in November: three single sheets sewn together of a size and watermark, “GFJ,” elsewhere unknown in his œuvre. In March 1840, Berlioz himself would have had no reason to make the kind of careful fair copy of “Villanelle” that served the editor of the Allgemeine musikalische Zeitung. He had finished Roméo et Juliette in September 1839 and would not receive the commission for what became the Symphonie funèbre et triomphale until, at the earliest, April 1840. Between March 17 and April 12, 1840, Berlioz wrote nothing for the Journal des débats; and for the Revue et Gazette musicale he reviewed only the orchestral concerts at the Conservatoire (on March 12 and 29, and April 9). I therefore suppose that the Darmstadt manuscript was indeed prepared in the autumn of 1842, shortly before its facsimile reproduction, and that the date Berlioz inscribed on it, March 23, 1840, represents either the composer’s recollection of the moment at which he had composed the song—he would have had time to do so at the time—or, more likely, as I shall indicate below, his mis-recollection of the date of a performance. Be this as it may, Nuits d’été—Berlioz’s title, not Gautier’s, and thus a bit of evidence by no means too obvious to consider—may tell us something of the season of its main composition. From an article to be cited in the final section, below, we know that the cycle was in fact completed by the end of October 1840.

As for the publication: The Bibliographie de la France, primarily for literary publications, did have a regular section for artistic productions, including music, but Les Nuits d’été nowhere figures in that semi-official periodical. Berlioz’s publisher, Adolphe Catelin, whose prints do appear in the Bibliographie de la France in 1840 and 1841, seems neither to have registered nor to have advertised Berlioz’s newest collection. Catelin had had dealings with Berlioz since 1836, bringing out a second edition of the Neuf Mélodies as Mélodies irlandaises, and the full score and parts of the overture Le Roi Lear (both mentioned in the Journal des débats of March 28, 1840), and the piano reduction of the cantata Le Cinq Mai (mentioned in the Débats as well as in Le Constitutionnel of May 14, 1840). Some seven months later, in a letter of November 9, 1840, Berlioz specifically chastised Catelin for not advertising his works, saying that, after all, “the public cannot divine their existence”!10 CG 2:664.

Like many French publishers, Catelin kept changing addresses, and for a time he had more than one. From a plate number near to that of the Berlioz, it can nonetheless be established that Les Nuits d’été appeared no later than August 1841.11 The plate number of the Berlioz is Ad.C. 872. The plate number for a Menuet dans les bois by Alexis Roger is Ad.C. 841, listed in the Bibliographie de la France on August 28, 1841, and thus probably printed, given the bibliography’s slow-paced practice, in July. Indeed, though no advertisements of the Catelin edition have been found (when Simon Richault took over Berlioz’s publications from Catelin, in 1843, he did include Les Nuits d’été in an advertisement he placed in La France musicale on November 26, 1843), Berlioz’s cycle is mentioned in the feuilleton of the Allgemeine musikalische Zeitung of September 1, 1841. More important, a review of the collection appeared in Paris, in the capital’s leading music journal, as early as July 4, 1841. I shall say a word about the substance of this review below. Here let me note simply that the author was the pianist-composer Stephen Heller, who later became one of Berlioz’s closest friends. A copy of the Catelin edition, with an autograph dedication to Heller (“à M. St. Heller, témoignage d’amitié et d’une vive admiration pour son grand et noble talent. H Berlioz”), has been preserved,12 BnF, Musique, Rés. 1432 (27). and it is possible that, for his review, Heller worked from this very score. But it also is possible—since the publisher is not mentioned in Heller’s review—that in July 1841 he had only Berlioz’s autographs before his eyes.

In his authoritative catalogue of Berlioz’s autograph musical documents, D. Kern Holoman calls the Paris manuscripts “autograph fair copies.”5 NBE 25 [2nd ed.]:308–321. My examination suggests, from placement and penmanship, that “Reinschrift” or “fair copy” is a fair description of “Le Spectre de la rose” and “Barcarolle”; but that the other manuscripts are “Urschriften,” carefully corrected working copies with paste-overs for subsequent use by the publisher. And—the point is unrelated—whereas the manuscript of “Le Spectre de la rose” is headed “Andante un poco lento e dolce assai,” a later album leaf with the opening nine bars of song is marked “Adagio.”6 BnF, Musique, ms. 382. The slower marking may well represent Berlioz’s definitive conception of the tempo after having heard the song in performance.

Gautier was probably among those who came to Berlioz’s lodgings in Montmartre in 1835 for the celebration of his son Louis’ first birthday. Louis’ mother, Harriet Smithson Berlioz, was long admired by the young writer, and in his feuilletons that appeared regularly in Émile de Girardin’s new, inexpensive, and thus for the first time widely circulating newspaper, La Presse, Gautier—with cues and clues from the composer—regularly praised Berlioz’s concerts and compositions. When the popular song composer Hippolyte Monpou died suddenly in 1841, at the age of thirty-seven, Gautier praised him as “Le Berlioz de la ballade”; when the Société des Concerts du Conservatoire performed excerpts from La Damnation de Faust, on April 15, 1849, Gautier referred to the composer (in his column for La Presse of the following day) as “the only symphonist that we have in France”; and when Berlioz himself died, Gautier prepared a long and laudatory article that appeared on March 16, 1869, in the Journal official, and again in Gautier’s celebrated Histoire du Romantisme of 1874. For Gautier, Berlioz was “the most literary musician in existence.” More famously, he formed, with Victor Hugo and Eugene Delacroix, “the great trinity of French Romantic art.”1 Gautier, Correspondance générale, 1:2271; 3:298, 119. For Berlioz, Gautier was a poet of moonlight, melancholy, and, to paraphrase Henry James, a clear and undiluted strain in the minor key. In 1844, when he was negotiating a new contract with Alexandre Dujarier, owner of La Presse, Gautier went so far as to style himself the journal’s Jules Janin (drama critic), Étienne Delecluze (salon critic), and Hector Berlioz (music critic).2 Gautier, 2:195. In 1847, when Berlioz was engaged by the half-mad impresario Louis-Antoine Jullien as conductor of the concerts at the Drury Lane Theatre, in London, he requested a new ballet scenario from Gautier, whose excellent reputation in the genre had been earlier created by Giselle (1841). Gautier prepared a scenario, based on Goethe’s Wilhelm Meister’s Apprenticeship, but with the demise of Jullien’s enterprise, the project—for which Berlioz was not going to compose the music—came to naught. Apart from this non-venture, their only artistic collaboration occurred with Les Nuits d’été.

Is “collaboration” the applicable word? Gautier left Paris for Spain on May 5, 1840, and returned to the French capital five months later, on October 7. He and Berlioz could have spoken about the settings before Gautier’s departure, and I would assume they did. In August 1837, Victor Hugo wrote a note to Berlioz that introduces “some lovely verses which a young poet, my neighbor, has written for you and asked me to send along.” The neighbor in question was in all likelihood Gautier; the verses, some of those later set as Les Nuits d’été.3 CG 2:311–312. Katherine Kolb informed me that the letter is postmarked August 3, 1837, and that the Hugo scholar Jean Gaudon was certain that the “young poet” was Gautier. But apart from this letter, we have no document that attests to Berlioz and Gautier having worked actively together. As to Gautier’s way of working with musicians, however, documents there are. A number of his poems were written expressly for musical setting, including the “Barcarolle” that eventually became the finale of Berlioz’s cycle. This poem, drafted in 1834 for the composer Allyre Bureau, was several times revised by Gautier before its definitive publication in La Comédie de la mort, which Louis Desessart brought out in Paris in 1838—a collection to which Berlioz would have been attracted by the irony of the title, and from which he would make his own selections in 1840. The title of the volume applies to the opening, two-part poem, “La Vie dans la mort,” and “La Mort dans la vie.” This is followed by fifty-seven separate and unnumbered poems, of which Berlioz selected nos. [27], “Le Spectre de la rose”; [28], “Lamento. La Chanson du pêcheur”; [38], “Absence”; [44], “Lamento; [45], “Barcarolle”; and [56], “Villanelle rythmique.”

Beyond “Barcarolle,” written for Bureau, published in 1834 as “Le Pays inconnu” and in 1835 as “Mirage (Barcarolle),” other songs, too, were drafted prior to 1840, as perhaps Berlioz was aware. “Le Spectre de la rose” appeared on May 7, 1837, in the magazine Don Quixote; “Villanelle rythmique,” written for Xavier Boisselot, also appeared in 1837;4 According to a note on the autograph manuscript, BnF, Musique, ms. 4383. so, too, did “Lamento,” as “Sur la mer.”5 Poésies completes de Théophile Gautier, 1:lv-lxiii. Berlioz’s own selection from the complete publication includes two “pairs” of poems; and his final ordering, with many interstices, follows Gautier’s, except for the removal of the “Villanelle” from last position to first. It is almost as though Berlioz warmed to the idea of making musical settings as he read progressively through Gautier’s collection and began to compose, as he did the music for Les Troyens, with a “scene” near the end that moved him the most.

For Allyre Bureau, the first musician with whom Gautier collaborated, the poet drafted what he called his “chanson” in two ways, with and without refrain; he told the composer to use the version he thought was best for musical setting.6 Gautier, Correspondance générale, 1: 42–43. Gautier was more explicit with his friend François Bazin (winner of the Prix de Rome in 1840), to whom he sent a poem with the following advice: “Treat my poetry as you wish; if something [in the text] displeases you, I shall change it. I am sending it in two versions, with and without refrain. You may choose. And write to me if you have any other particular musical idea to which my poetry might be adapted.”7 Gautier, 1:248. When he sent a poem to Meyerbeer, in the spring of 1839, Gautier went so far as to provide a monstre—a schematic outline indicating the scansion of the text, the separate poetic feet, and the long and short syllables of each—something rarely done in French poetics, where analysis is based on syllable count and on the rhythm and intensity of the line as a whole. To the German composer, Gautier added: “If you find this pattern acceptable, I shall try to improve the verses while maintaining the present form. If you would prefer some other meter, please let me know. I have maintained a rigorous symmetry in these lines; if they are not yet worth much as poetry, they are, I think, appropriate for musical setting.”8 Gautier, 1:146.

From these examples—one could give more—it is clear that Gautier did not belong to that group of poets whom Berlioz considered completely lacking in musical sensibility. On one occasion, when he was assisting Louise Bertin in the preparation of her opera, La Esmeralda, mentioned in chapter 4, he wrote of the librettist, rather snarkily, that “[Victor] Hugo expects a great success. He judges music as do all the poets, which is to say that he is completely devoid of musical sensitivity.”9 CG 2:285. Gautier, on the contrary, was aware of what he referred to as the “double exigencies of poetry and music,” as he put it in a review of Roméo et Juliette in La Presse of December 11, 1839. He was willing to adapt his poetry to the needs of the musician. He seems to have preferred to do the adapting himself but would presumably have agreed with Berlioz that in some instances, “it is better to upset the progression of the poetry than it is to alter the musical continuity.”10 CG 2:183. Unlike Goethe, who preferred music that in no way challenged the supremacy of his verse, Gautier appears to have viewed the mélodie as a mutually creative venture. It is for this reason, no doubt, that so many composers, including Bizet, Fauré, and Duparc, found inspiration in the poems of La Comédie de la mort.

Of the fifty-seven poems in this collection, sixteen were sooner or later set by one or more composers during the nineteenth century. “Villanelle,” written for Boisselot, was set by Berlioz and at least twenty-three others; “Le Spectre de la rose” was set by eight others; “Sur les lagunes,” by twenty others; “Absence,” by sixteen others; “Au cimetière,” by only three others; “Barcarolle,” by eighteen others. Two further poems from La Comédie de la mort, “Romance” and “Les Papillons,” were set by eleven and twenty-one composers respectively.11 Spœlberch de Lovenjoul, Histoire des œuvres de Théophile Gautier, 1:123, 152–153, 155. These dry-as-dust statistics tell us that, except for “Au cimetière,” Berlioz chose poems that were already or soon became widely considered appropriate for musical setting. If we did not believe that his eventual song is a small miracle, we might be inclined to ask why, instead of setting “Au cimetière,” Berlioz did not set “Romance,” or even “Les Papillons”!

A partial answer is provided by our knowledge of Berlioz’s own thoughts about composing a mélodie, or romance, expressed succinctly in a letter to the editor of the Journal des jeunes personnes, a popular young women’s magazine, who had asked him to set a poem by the children’s book author Léon Guérin. Writing on October 10, 1834, to “Monsieur Duplessis” (I believe the man in question is Joseph Duplessy, editor of, among other things, a collection of writings by women),12 Trésor littéraire des jeunes personnes, 2nd ed., ed. Joseph Duplessy (Tours: A. Mame, 1862). he declined the editor’s request, saying that “the character of each couplet [of the poem] would require a different music—something that would make the dimensions of the piece simply incompatible with those of a work appropriate to your journal.” “Furthermore,” he adds, “I am at the moment so busy that I simply do not see how I could find an entire day and devote it exclusively to this little composition.” And he concludes with a credo: “Such things ought really to be improvised, so to speak, and when one fails [to find the appropriate setting] on first encountering the poem, in my opinion, one really ought to abandon the effort.”13 CG 2:203.

Here, then, is indirect evidence of Berlioz’s presumably immediate attraction to Gautier’s poetry. As for his title, perhaps he found “Les Nuits d’été” both euphonious and appropriate to the theme of melancholy longing that runs through the verse. “De belles nuits d’été” and other forms of that locution, including Le Songe d’une nuit d’été—Shakespeare’s play and Mendelssohn’s overture—are found on frequent occasion in the titles of the publications and in the pages of the press of the day. Considerations of euphony were obviously important to his choice of titles for the individual songs: Gautier’s “Villanelle rythmique” became Berlioz’s “Villanelle”; “Le Spectre de la rose” remained unchanged by the composer; what for Gautier was “Lamento. La Chanson du pêcheur” became for Berlioz “Lamento” (in the Geneva manuscript) and “Sur les lagunes. Lamento” (in the printed edition); “Absence” remained unchanged; Gautier’s second “Lamento” was entitled by the composer “Au cimetière. Clair de lune,” although the second part of Berlioz’s title seems to have been an afterthought. For the final song, Berlioz used Gautier’s title, “Barcarolle,” in the manuscript, but adopted “L’Île inconnue” for the printed edition. It is furthermore possible—the point is of no small significance—that some of these emendations (“Sur les lagunes”; “Au cimetière”; “L’Île inconnue”) were made by the composer with a view toward inspiring the scenic imagination of the artist who might eventually be charged with making title-page illustrations: these, as the reviews of the day make abundantly clear, were objects of appreciation equal in importance to that of the songs themselves. Unfortunately, Les Nuits d’été appeared from Catelin, and later from Richault, with neither engraved portraits nor lithographed vignettes. Those by Louis Boulanger and Barathier that grace the Boieldieu jeune edition of Berlioz’s Le Montagnard exilé (1823) and the Schlesinger edition of Neuf Mélodies (1830), for example, to say nothing of the later ones by Frédéric Sorrieu and Georges Staal that decorate the Richault editions of Berlioz’s La Captive (1849) and Sara la baigneuse (1850), provide a treat for the eye that some might have found more tempting than the music inside.

It is generally assumed that aside from “Absence,” orchestrated by Berlioz in 1843 for performance by his traveling companion, Marie Recio, the other five songs of Les Nuits d’été were orchestrated shortly prior to publication, in 1856. D’Ortigue’s remark allows for the possibility that one or two songs beyond “Absence” had been orchestrally “embellished” by 1852, if not before. When Berlioz penned a letter of candidacy for a chair at the Académie des Beaux-Arts, for example, on March 6, 1851, he mentioned the six songs for voice and piano of Les Nuits d’été, “several with orchestra.”1 CG 4:37. As for the thrust of d’Ortigue’s comment, that the voice and piano version has an inspirational freshness worth preserving, it is of course not to be underestimated, in spite of the historical appreciation with which Berlioz’s orchestration has been favored. Berlioz himself implied, as D. Kern Holoman puts it, “that the orchestration of a work occurred subsequent to what [he] believed to be its completion.”2 Holoman, The Creative Process, 173. In prompting d’Ortigue to write something about his collection of songs, Berlioz said that he wanted “only that their existence be known, that they are not shoddy goods, that [the composer] has in no way had sales in mind, and that these ‘petites compositions,’ which have nothing formally or stylistically in common with Schubert’s, require for proper execution singers and pianists—musicians—of consummate artistry.”3 CG 4:150–151. He would later commit the same notion to print:

Are we to believe Berlioz when he says of his songs that they were not commercially inspired? Some publishers at the time—Catelin was one—seem to have existed on such publications; their stable of composers was presumably well fed by such staples of the repertory. What other reasons might Berlioz have had for departing from his passion for the grandiose and taking up a genre in miniature? Beyond Ian Kemp’s suggestion that the work “must be regarded simply as a characteristic product of the Romantic temperament,”5 NBE 13:xi. there were surely specific reasons for Berlioz’s undertaking Les Nuits d’été for voice and piano in the spring of 1840, all of them no doubt overlapping.

(1) He wished to write something for a particular singer whom he admired. As in literature, where so many fictional characters are modelled on friends and acquaintances of the author, here, too—and especially in vocal music, characteristic performers may be “inscribed” into the characteristics of a score. The Catelin publication of 1841 is marked for mezzo-soprano or tenor, but Berlioz—who describes himself as a “second-rate baritone” in chapter 12 of the Mémoires (in fact he was a practiced vocal coach, as we know from, among other documents, his account of tutoring the young French tenor Victor-Hippolyte Delahaye, a potential successor to Gilbert Duprez)6 CG 2:699.—surely had a preference for the former: Marguerite, in La Damnation de Faust, is a mezzo; so, too, are his greatest heroines, Cassandra and Dido, in Les Troyens, to say nothing of Béatrice, in Béatrice et Bénédict. In July 1840, Berlioz criticized the hiring practices of the management of the Opéra by saying that if the theater continued to hire “only super-high sopranos, with no medium or lower registers, then I think it will be necessary to give up passionate scenes and dramatic music altogether.”7 Journal des débats (July 19, 1840). Gautier, too, favored the lower female voice, which he praised in his 1847 poem “Contralto,” probably written for his long-time mistress, the contralto Ernesta Grisi. The Paris autographs of Les Nuits d’été specify no vocal type, but the Darmstadt fair copy of “Villanelle” is carefully marked “Mezzo-Soprano.”

Berlioz’s ideal mezzo was Pauline Viardot, who was only nineteen years old in 1840 but who was even then recognized by connoisseurs as an artist. Meyerbeer would soon suggest that she be engaged by the Opéra, and Berlioz (who in 1838 called her a “diva manqué,” though more for her repertory than for her vocal resources) would later consider her one of the greatest artists in the history of music. The reigning mezzo at the Opéra at the time was Rosine Stoltz, who created the role of Ascanio in the 1838 production of Benvenuto Cellini (and who created a stir after 1840 as the mistress of the new director of the Opéra, Leon Pillet). It may be that Berlioz’s relations with Stoltz were on the wane at the time of Les Nuits d’été—and personal relationships, to say nothing of romantic attachments, were at the time (and continue to be) of no small consequence in the artistic world. Cornélie Falcon, who sang the first performance of the orchestral version of Berlioz’s La Captive, in 1834, was also a leading mezzo of the period and, according to Berlioz, the repository of the current Opéra director’s hopes for success.8 CG 3:635.

We then come to what Annegret Fauser has called Berlioz’s larger “autobiographical project,”9 Fauser, “The Songs,” 124. to the young lady who became Berlioz’s supportive mistress, and, in 1854, the singer who became his second wife. That so little is known about Marie Recio, that “devoted and intelligent woman” who, though ungraciously excluded from the Mémoires, shared over twenty years of the composer’s life, “never for a day ceasing to lavish upon her husband the most tender and delicate attentions,”10 From the obituary in La France musicale (June 22, 1862). has been one of the real lacunae of modern Berlioz scholarship. Marie Recio appeared on the scene—of the Opéra—in 1841, after the composition of Les Nuits d’été. The archives indicate that she was hired at the Académie Royale de Musique on October 9, 1841; her contract was terminated on September 8, 1842, when she left Paris, with Berlioz, for Brussels. In those eleven months she sang in the house’s productions of Rossini’s Comte Ory and Donizetti’s La Favorite.

Until recently, her pre-operatic life was a mystery. We now know that before adopting the stage name of Recio, Marie-Geneviève Martin appeared in concert as Marie Willès, her adopted name clearly a Frenchified version of the name of the Spanish woman, Marie Sotera de Villas, who gave birth to her, in the Parisian suburb of Châtenay (now Châtenay-Malabry), in June 1814, as the consort of her father, a French military officer, Joseph Martin.11 Pascal Beyls, “A Surprising Discovery,” 39–52. (That other mezzo, Rosine Stolz, also appropriated for her stage name the maiden name of her mother.) As a singer, Marie Willès first appears in the press at the beginning of 1840, as the dedicatee of “Vive l’hiver,” a mélodie published in Le Ménestrel on January 26, 1840, by the then much in view composer-voice-teacher Giuseppe Concone. It is not clear how Marie Willès came to the attention of Concone, who settled in Paris in 1837, but perhaps she was among his first students in the French capital. (When Berlioz mentioned Concone’s new album in the Journal des débats of January 14, 1838, he noted that it had been recommended to him by “une grande musicienne,” but did not name the woman in question.) Marie Willès next appears in the press in reviews of a matinée musicale held in the concert rooms of the piano-making Richter brothers on March 26, 1840—precisely three days after the date that Berlioz inscribed on the “Darmstadt” manuscript of “Villanelle.” One month later, in his article for the Revue et Gazette musicale of April 26, 1840, Berlioz himself mentions “Mademoiselle Willès” as having caused a sensation at the concert given by Adolf Schimon, the Vienna-born pianist who was studying at the Conservatoire and acting as accompanist in the voice classes of Davide Banderali:

Berlioz—who, in 1840, it must be remembered, was troubled by a wife who was becoming increasingly dependent, isolated, prone to illness, sensitive, frustrated, resentful, and demanding, to employ the terms used by her biographer14 Raby, Fair Ophelia. —would continue to praise the artistic talents of Marie Willès, nowhere more fully than in his account of the concert she gave on February 9, 1841, which he reviewed at length in the Journal des débats of February 14:

Berlioz was obviously aware of the celebrity of Marie’s teacher, Davide Banderali, who had begun his tenure at the Conservatoire in 1828, when Berlioz was still a student there. And he knew of Giuseppe Concone, as we have seen, who may earlier have worked with Marie. Now, in 1841, infatuated with her, he would have presumably wished to help one of Banderali’s “best students,” as he referred to Marie, to obtain a contract at the Opéra, especially as he was on good terms with the recently appointed director, Léon Pillet. We have no evidence of any such preferential treatment, but Berlioz was on record as decrying the institution’s lack of mezzo-sopranos. It is in the announcement of Marie’s appointment to the Opéra, in Le Ménestrel of October 24, 1841, that we first discover the proof of her dual identity: “Mlle Marie Recio (Willèz) [sic], déjà connue du public des concerts, vient d’obtenir un engagement de deux ans à l’Opéra”—“Mademoiselle Marie Recio (Willèz), already known to the concert-going public, has just signed a two-year contract at the Opéra.” (We find further confirmation in a communication from Auguste Morel sent, after Berlioz’s death, to the editor of Le Ménestrel, in which he casually mentions that Berlioz married “Mademoiselle Willès […] after the death of his first wife.”)17 Le Ménestrel (October 31, 1869). In fact Marie Recio’s contract was for only one year, not two, but the identification is unmistakable.

The title page of the original edition of Les Nuits d’été is also perfectly clear: the songs are for mezzo-soprano or tenor. In his first reviews, Berlioz speaks of Marie as a soprano. But when he published La Belle Isabeau, shortely thereafter, and dedicated it explicitly to Marie Recio, he indicated that it was for mezzo-soprano. As Julian Rushton has demonstrated in his study of Berlioz’s understanding of the mezzo-soprano voice, the distinction between “soprano” and “mezzo-soprano” is not as clear-cut as we might like it to be.18 Rushton, “Berlioz and the Mezzo-Soprano,” 64–88. Still, the “pitch center of gravity” of “Villanelle”—to use Rushton’s ingenious measure of tessitura, and to take as an example the song that by the date on the “Darmstadt” manuscript must now be associated with Marie Recio—is higher than that of any other vocal number in Berlioz’s “mezzo-soprano” repertory. “Absence” lies high as well—and many singers transpose these songs down. But Marie presumably had no trouble singing “Villanelle” in its original key of A major. (The contemporary French soprano, Véronique Gens, sings it in A as well. The mezzo-sopranos Janet Baker and Anne Sofie von Otter, to mention only two, sing it in F.) We know, from a letter from 1843, that Marie sang “Absence,” and sang it well, in its original key of F-sharp major.19 CG 3:85. (Baker and Otter sing it respectively in E and E-flat.)

(2) He wished to write a work for a particular occasion. In the Revue et Gazette musicale of October 25, 1840, we find the following notice: “In recent days, we have been very happy to hear Wartel rehearsing six new mélodies that M. Berlioz has just composed on poems selected from the Théophile Gautier’s delightful volume La Comédie de la mort. We do not hesitate to say that never before, in this musical genre, has a composer achieved such originality or such profundity and grace of expression—a perfect marriage of melody and accompaniment, the one brought richly to life by the other.” This is the paragraph, clearly penned by the composer himself, that offers the proof I mentioned above that the six songs were in fact completed by October 1840.

In the subsequent issue of the magazine, dated incorrectly as “jeudi 4 novembre 1840”—in fact Thursday of that week fell on the 5th—we find the program of the concert sponsored by the Revue et Gazette musicale that was to take place on the following Sunday, November 8, 1840. The seventh of the ten items appears as “Absence, Le Spectre de la rose, mélodies de Berlioz, paroles de Monsieur Théophile Gautier, chantées par Monsieur Wartel et accompagnées par Monsieur Gustave Collignon.” But when the program was reprinted on the day of the concert, the items by Berlioz were nowhere to be found. It is probable that the tenor, Pierre-François Wartel, at the time singing demanding roles in the several works on the boards at the Opéra in that month, was simply unable to learn the songs to the satisfaction of the composer. Wartel, known for having introduced Schubert’s Lieder to French audiences, went on to a distinguished career in Paris and abroad as both singer and teacher. The scheduled pianist, Gustave Collignon, a first prize winner at the Conservatoire in 1837, is the man who accompanied Marie Willès in her début recital. He would pursue his career as a pianist, go into exile in September 1848, and become a mainstay of concert life in the American city of New Orleans: “Through the efforts of this one man, New Orleans experienced its first independent symphony orchestra, its first regular chamber and orchestral concert series, and its first systematic hearing of the greatest European instrumental works.”20 Baron, Concert Life, 77.

It is clear, from what we read in the Revue et Gazette musicale, that Berlioz had completed the songs and rehearsed them with Wartel in preparation of their performance at the magazine’s concert of November 8, 1840. If he had conceived the songs with only Wartel’s voice in mind, Berlioz would presumably not have published them “for mezzo-soprano or tenor.” (Wartel would indeed sing “Villanelle” and “Absence” on Marie Willès’s concert of February 9, 1841.)

Beyond performance at a concert sponsored by the journal with which he had been associated since 1834, Berlioz surely hoped for performances of the new mélodies at other venues as well, possibly at one of the soirées offered at the Palais Royal by the Duc d’Orléans. Berlioz had had contacts with the Orléans family since the early eighteen-thirties, and like other artists of the day, including Théophile Gautier, he was especially fond of this Duke, Ferdinand-Philippe, King Louis-Philippe’s eldest son and the heir to the throne. Indeed, as outlined in chapter 5, the Duc d’Orléans attended the performance of Berlioz’s major work of the year 1840, the Symphonie militaire soon to become funèbre et triomphale, and shortly thereafter accepted the dedication of the published score. There were, of course, numerous chamber music concerts in 1840 in the salons of the piano makers Érard, Herz, Pape, Pleyel, Bernhardt, Richter, Petzold, Couder, and others who had the requisite room and riches. Such soirées, though he liked the beverage, were not Berlioz’s usual cup of tea.

(3) He wished frankly to make some quick, cold cash. Berlioz had lifelong financial problems. To the dismay and embarrassment of his middle-class family, these problems were sometimes emphasized by those who wrote about the composer in the daily and weekly press. In 1841, Berlioz wrote to his sister that “la grande musique” was ruining him.21 CG 2:685. Perhaps he wished to recoup his losses with some potentially remunerative music that was “petite.” But despite the marketability of the romance or mélodie, the publisher of Les Nuits d’été, Adolphe Catelin, was in marketing neither inventive nor aggressive. He went out of business only two years later. Furthermore, the Berlioz whom we know from the letters and the Mémoires, while always fearful of imminent bankruptcy, was never really in the music business for profit: his regard for the sanctity of art—explicitly mentioned in Stephen Heller’s review of the song cycle, which appeared on July 4, 1841, and which we have cited—was well known, and genuine.

Indeed, in the other contemporary review of the new mélodies, the anonymous critic praises the “exquisite finesse and serenity” of Les Nuits d’été, but, as concerns marketability, expresses genuine doubts. Because the author of the review, “F,” has wrestled closely with the new score, we quote his review at length.

These reviews, positive or negative, were often personally or politically inspired; they were prepared on the nudging of those who knew the critic (or the subject of his notice), and were in a position to tender thanks for services rendered; they were motivated by envy, by competition, by clique-ism. No opinion expressed in nineteenth-century French newspaper print was unprimed: “critics’ ideologies, and more specifically their political attitudes and personal acquaintances played a singularly important role in aligning them ‘for’ and ‘against’ an artist.”25 Bellos, Balzac Criticism, 189. This is one of the overmastering messages we receive from Berlioz’s Correspondance générale and from the letters of, among others, Théophile Gautier. It is the message Karl Gutzkow conveyed to German readers in his Letters from Paris of 1842, when he said that it is impossible for a critic in Paris “to be completely independent, or, what is the same thing, to be completely honest.”26 Allgemeine musikalische Zeitung, 44 (November 16, 1842): col. 925. And it is a message of which one is constantly reminded in the France of today as well. Life and culture have been politicized in France to a degree difficult for Americans to imagine, or to accept, because Americans until recently have tended to believe in the myth of an “objective” press, and because in the United States the government’s contributions to culture have been minimal, while in France the great cultural philanthropist has from time immemorial been the government, the Maison du Roi, the Maison de l’Empereur, the administration, officialdom: indeed, it is not uncommon to see it remarked in the French press that “l’administration y fait toujours la pluie et le beau temps”—“it’s the administration that makes it rain or shine.” In the nineteenth century, it was not grants for the arts that the government was distributing, but rather subventions, indemnities, commissions, encouragements, administrative posts, and sinecures, as we have seen in previous chapters, that permitted a number of artists to work much, if not most, of the time for themselves. Alexandre Dumas, as I remarked here in the prologue, was librarian at the Palais Royal, as was Casimir Delavigne; Charles Nodier was librarian at the Arsenal; Alfred de Musset was librarian at the Ministry of the Interior; Berlioz was librarian at the Conservatoire; in 1868 Théophile Gautier became librarian for the Princess Mathilde. Filling out book orders and catalogue cards were for none of these men their primary occupation. Notices in the press, with the potential they offered of reaching “les dames du grand monde,” as Berlioz called them,27 CG 2:722. and others in high places, were crucial: the talked-about were on the road to success. We must thus read critical notices against a (loosely defined) political backdrop, and with political questions—the force of central authority and of those behind it—in mind.

With Les Nuits d’été, Berlioz’s detractors were in a sense asked to reconsider. Just as they had been asked to reconsider in 1830, with the publication of the Neuf Mélodies, which François-Joseph Fétis, Berlioz’s first important critic, aware of the composer’s principled independence and yet presciently alarmed by the composer’s inclination to grandiosity, greeted with encouragement:

The anonymous reviewer for La Quotidienne, whom we have quoted at length, opened his review with comments on the romance, whose commercial value, he recognized, was sadly opposite to its artistic merit. He went on—and this is crucial—to speak prophetically of the world’s greatest master of the Lied:

In his later review, Joseph d’Ortigue opened his remarks on Les Nuits d’été with irony:

Berlioz’s song is through-composed. Parsing the melodic phrases is no simple matter, as it is, by contrast, in the contemporary setting of the poem by Félicien David (published in 1840 and listed in the Bibliographie de la France on January 2, 1841), where Gautier’s first four lines are set to two four-bar phrases of patently obvious symmetry. In the Berlioz, those four lines encompass fifteen bars: saying that they comprise phrase units of 4 + 4 + 7 fails to do justice to Berlioz’s always originally inflected rhythmic imagination. David does write ad libitum at the head of the refrain, he does set “Ah, sans amour, s’en aller sur la mer” in five bars rather than four, and he does choose an appropriately minor key, in this case G minor, for his setting as a whole. (The setting of the same poem by Charles Lenepveu, of 1870, is in F minor; the setting by Charles Gounod, of 1872, is in E minor.) Berlioz’s “Lamento,” too, is in a minor key: following on the heels of “Le Spectre de la rose,” in D major, G minor here can sound like the second member of a logical progression. But the sequence of the keys of the six songs—A, D, G minor, F-sharp, D, F—is hardly suggestive of a largescale tonal design. The unity of the “cycle,” such as it is, turns upon the poetic narrative and imagery, the vocal style that revolves around melodic recitative and lyrical arioso, and the presence of repeated musical gestures.

“Sur les lagunes” is built—this is its most outstanding feature—upon a recurring motive, a kind of ostinato, even a leitmotif, if I may use the term loosely: a musical response to Gautier’s haunting refrain, “Ah, sans amour, s’en aller sur la mer.” The leitmotif occurs some fourteen times in the piano part; though it is echoed in the vocal line, it occurs there literally only twice—at “pleure, pleure” (bars 57–58), one of the few words of the original poem repeated by the composer, and at the final sigh, “Ah,” another Berliozian textual addition and a dramatic touch that is both fitting and lovely. The linear half-step motif is initially presented (bars 1–2; 5–6) as a sustaining of the fifth degree of G minor briefly inflected by the flattened sixth. In the course of the song, the sustained D functions variously as the fifth, the third (major and minor) and the root of a triad. In bars 13–14, the D figures in an inversion of the harmony we now know as the “Tristan chord,” although in this case Berlioz is concerned less by vertical simultaneities than by changing colors produced by the process of reiteration. Indeed, we have here one of a privileged group of movements, or moments, controlled by a systematically repeated gesture that by its very strictness calls attention to all that is free and imaginative in Berlioz. Such an expressive scenario is used to sublime effect in the septet in act 4 of Les Troyens, which is linked to the love duet that follows, in G-flat major, by the insistent oscillation of the fifth and flattened sixth degrees of F major. It was used by Berlioz as early as 1834, for the tolling of the evening bells in the “Pilgrims’ March” of Harold en Italie; it was used again in the “Funeral March” of the dramatic symphony, Roméo et Juliette; and it was used in what is the true forerunner of the compositional procedure employed in “Sur les lagunes,” the Offertorium of the Requiem, which won for Berlioz the “priceless approval” of Robert Schumann: “This Offertorium,” said Schumann to the composer, “surpasses everything.”4 Mémoires, 532.

The choral psalmody in the Requiem, which Berlioz called a “chorus of souls in Purgatory,”5 NBE 9:156. surely had a direct bearing on the significance, for the composer, of the ostinato in “Sur les lagunes.” The significance would be identical to that of Juliet’s “Convoi funèbre,” whose ostinato Berlioz also called a psalmody: it is a spiritual significance, suggested by the “heavens” and “angels” of Gautier’s poem, which adds the deity to the typically Romantic constellation of love, night, and death, that motivates this lamentation on the soul of the departed lover.

At the end of “Sur les lagunes” we are left poised on the dominant, longing until “infinity” for our “belle amie.” Only on one other occasion did Berlioz end a movement with a similarly unresolved harmony, namely, in “La Harpe eolienne—Souvenirs,” the single instrumental movement of Lélio ou Le Retour a la vie, the sequel to the Symphonie fantastique, as revised in 1855.6 NBE 7:68. The poetic, or dramatic, circumstances there are similar, and there, as in the song, Berlioz reworked the ending with particular care. Indeed, for “Sur les lagunes,” there are even slight differences between the corrected autograph and the first edition, suggesting further adjustments to the ending in proof.

Chopin and Schumann essayed unresolved endings in certain multisectional works (the Prelude in F major from op. 28, for example, and “Bittendes Kind” from Kinderszenen), and in this case, as in those, the technique of avoiding closure—in a sense the most obvious of all representational practices of the Romantic era in music—signals the composer’s larger structural conception. I earlier suggested, on the basis of differences among the Paris and Geneva manuscripts of Les Nuits d’été, that “Sur les lagunes” was separately composed and subsequently incorporated into the six-song collection published in 1841. It is on the one hand logical to assume that the decision to end this song on the dominant was a part of the larger determination to publish the six songs in what became their final, progressive arrangement as—there is no other word for it—a “cycle.” On the other hand, this ending on the dominant is not as expectant, or anticipatory, as others. The seventh degree of the final chord is not present. The chord, with the third in the tenor register, has a life of its own. Its “yes and no” quality, which I hear as a forerunner of the exquisite closing of the slow movement of Mahler’s Fourth Symphony, is as expressive as it is vague; its message could be one of irony, or of hope, or of despair.

Indistinct and ambiguous though it may be, there is nonetheless a message here, which filters through the works of artists as diverse as Carl Wilhelm Kolbe the Younger, E. T. A. Hoffmann, Lord Byron, Eugène Delacroix, Casimir Delavigne, and Gaetano Donizetti. Let me be precise. That lovely, incantatory line, the line that encapsulates the gloom of life in the shadow of death and thus a quintessential element of Gautier’s collection, “Ah, sans amour, s’en aller sur la mer,” did not simply come to Gautier from the recesses of his poetic imagination. It is in fact a translation of the first two lines of what is apparently an old Venetian verse:

It is obvious, then, that the subject—the arrogant old doge who is executed for breaking with his patrician class and favoring “the people”; the aged aristocrat who is married to the young and beautiful daughter of one of his former comrades-in-arms—was in the air. The interrelationships and anxieties of influence among these iterations of the legend cannot long concern us here. For Byron and Delacroix, the high tragedy of the political actor is preeminent, and the youthful wife remains faithful. Indeed, Balzac, in Les Martyrs ignorés of 1837, has a character congratulate Byron for having her do so: “Lord Byron had a great idea in having her remain faithful to her husband.”8 Balzac, “Les Martyrs ignorés,” in Études philosophiques, 208. For Delavigne, the youthful wife succumbs to a young lover, then later, before his execution, asks her husband’s forgiveness. For Bidera and Donizetti, it is simply not clear whether the young lovers consummate their guilty passion.9 Gossett, “Music at the Théâtre-Italien,” 347. For Hoffmann, the love intrigue takes precedence: the wife of the doge returns the affections of her youthful admirer but remains faithful to the old man until his demise. Then, in a kind of pre-Wagnerian “Liebestod,” the lovers are united—only to have their gondola swallowed up by a storm that arises in the aftermath of the execution of the doge. The lovers are drowned in revenge by what is, of course, the traditionally metaphorical wife of the doge—the sea itself.

It is earlier, while riding with the doge in the channel before St. Mark’s Square, that the dogaressa hears “the notes of a soft male voice, gliding along the waves of the sea” and singing the verses we earlier learned were inscribed on the picture frame, “Ah! senza amare, andare sul mare”—“Ah, sans amour, s’en aller sur la mer.” In Hoffmann, these verses crystallize the young woman’s emotions: upon hearing them, she feels the anguish of the lovelessness of her marriage to the doge and the temptation of lust for her handsome young admirer. More broadly, for the reviewer of the French translation, these verses capture “the risk of the unequal marriage of youth and old age, the inevitable misery of the couple, the sadness of the bride, the apprehension of the groom.”10 Journal des débats (January 27, 1830).

We know that Théophile Gautier was familiar with these verses from Hoffmann’s tale. In his novel Les Deux Étoiles, serialized in La Presse (from September 20 to October 15, 1848), he sets them down: “La chanson vénitienne, dans son admirable mélancolie, dit qu’il est triste de s’en aller sur la mer sans amour”—“The Venetian chanson, it its admirable melancholy, tells us how sad it is, without love, to set out upon the sea.”11 Gautier, Les Deux Étoiles, La Presse (October 5, 1848). In 1851 the novel appeared as Partie carrée. We know that Berlioz devoured Hoffmann when the tales first appeared in French in 1829—the impact of Hoffmann is a major focus of Francesca Brittan’s Music and Fantasy—and that he was friendly with Pauline Richard, one of Hoffmann’s translators.12 CG 1:293, 301, 348. We know, furthermore, that Berlioz was familiar with the play by Byron. He might well have seen the Delacroix. And it would have been difficult for him not to have seen or known the Delavigne, as it saturated the theatrical press in 1829 and 1830. He reviewed the Donizetti in Le Rénovateur of March 29, 1835, and said of the subject that it was too well known to require retelling. The sea, finally, had always been the scene of his dreams: he qualified the Mémoires themselves as “le livre de loch de [son] pénible voyage”—the log, or log-book, of his arduous journey. We may safely assume, then, that Berlioz took up these verses—not only a tale of a doge and a dogaressa but an encapsulation of stressfully conflicting emotions to which at the time he was acutely sensitive—with a full awareness of their historical, poetic, and musical resonance.

In “Sur les lagunes,” with its musically open-ended conclusion and its ostinato so effortlessly demonstrating “the complex character and function” of only one note, achieving a myriad of expressive possibilities with a miraculous economy of means, Berlioz, in what I have called a privileged moment, reveals “something of the inner process of Romanticism itself.”13 Primmer, The Berlioz Style, 97. He begins with a text that calls itself a romance. But instead of doing with it what a “normal” composer might have done, and despite what might have been a fleeting reason for turning his attention to song in the spring and summer of 1840, namely a desire to show a certain public that in fact he was a “normal” composer, he rather applies to that text, as he does to the others of the collection, each in its own way, a compositional technique of both obvious premeditation and apparent spontaneity that leads us now to view putatively “normal” works in the shadows of Les Nuits d’été.