Son-ne Trompette Eclatan-te
by Paul Zweifel
"Sound the loud trumpet!" So sing the Choeur des Gamins in Act I of Carmen as they march in accompanying the gar-de montan-te and then march out with the garde descendan-te. I have explicitly indicated the voiced final syllables (e.g. "-ne" and "-te"); such syllables, silent in spoken French, are frequently voiced in French poetry and song for reasons of scansion. The examples given above (and all other examples in this article) are also explicitly indicated in the scores I've checked--Schirmer, Kalmus, Belwyn Mills and Alkor-Bärenreiter. (The last-named score includes both the Opéra Comique version with dialogue and the Vienna version with Ernest Guiraud's recitatives.)
But strangely enough, many singers (one presumes with the acquiescence if not active collaboration of their conductors and/or vocal coaches) ignore both score and scansion, reverting to ordinary spoken French pronunciation. A particularly egregious example is Micaela's third-act aria in Carmen, "Je dis que rien ne m'épouvan-te," the third line of which is "Mais j'ai beau fai-re la vaillan-te." Kiri Te Kanawa,1 Licia Albanese2 and Janine Micheau3 sing "m'épouvante" and "vaillante" and I have often heard the same error made in live performances. Although it is admittedly presumptious of me to use the word "error" in referring to choices made by such great artists, I believe that I have justified this term in the discussion below. I would be delighted to hear any rebuttals that these ladies might wish to make. (Susanna Uher, the rising young American soprano, who has sung Micaela on a number of occasions, has informed me that she always sings "m'epouvan-te" and "vaillan-te" because "that's what's written in the score." Brava!)
In three other recordings4 one hears Mirella Freni, Faith Esham and Hilda Burke all sing "m'épouvan-te" and "vaillan-te" as written, so there seems not to be any consensus. Interestingly enough, in the three in which the Micaelas sing "m'épouvante" the Moralés's, in Act I, take liberty with the lines "L'oiseau s'envole, On s'en console" (sung after Micaela's hasty departure following Moralés clumsy sexual overtures). Bernard Planty in the Beecham recording and Hugh Thompson in the Reiner version both ignore the score and sing these lines without sounding the final e's. In the Hasselmans and De Burgos recordings, on the other hand, the Moralés's follow (or rather anticipate) Micaela's example and sing the lines as written, that is "…envo-le" and "…conso-le." Only in the Maazel recording do Micaela and Moralès differ--as already noted Esham sings her music as written, whereas Francois Le Doux drops the final e's in "envole" and "console." (George Cehanovsky sings the role in the Hasselmans CD while the singer's name is not listed in the De Burgos LP.)5
One can only speculate on the reason(s) that artists feel compelled to alter written scores. Perhaps they desire to sound "authentically French." But independent of what motivates the singers, here I would like to discuss the reasons they should in fact sing the parts as written (aside from avoiding insults to the composer's intelligence).
Consider the phrase "Je dis que rien me n'épouvan-te." If this were English poetry, it might be analyzed on the basis of metrical feet (three "amphibrachs.")6 However since the French language is considerably less stressed than English is is customary to analyze French poetry on the basis of syllables. As Will Crutchfield has put it in his analyis of the French and Italian versions of Don Carlo,7 "French verse makes great use of a certain poetic line that has nine syllables." Well, "Je dis..." has only eight syllables if the final e is mute, which is anothre good reason for sounding the final e.
A musical reason is the following: If the final e is mute, the phrase "Je dis.." ends as a dotted quarter note on the downbeat of a 9/8 measure. The voiced final e, an eighth note, would fall on the first part of the second-beat triplet, a more natural place (a weak beat) for a musical phrase to end. The usual practice of ending musical phrases on weak beats indicates that lyrical lines should normally end on unstressed syllables, so-called feminine endings.8 Muting the final e of Micaela's first line line masculanizes it! The preference for feminine over masculine endings in written lyrics emphasizes the flexibility of the French language with its plethora of final e's which in music (and poetry) may be either voiced or unvoiced. (Refer to the title of this article for a perfect example.) Since all words in spoken French are accented on the last syllable, it follows that only when a French phrase ends in an unvoiced e can it have a feminine ending. Poets, librettists and composers have taken advantage of this flexibility through the centuries, leading to a lyrical beauty in French poetry not fully appreciated, perhaps, by non-speakers of the language.
Italian has a similar poetic flexibility, not through the addition of extra syllables but through their removal, so-called "apocopation." Thus when necessary for scansion "amore" becomes "amor;" "andiamo" becomes "andiam;" "sono" becomes "son;" and so forth.9 One may even find both apocopated and normal versions of a word in the same aria. For example, in the last act of Rigoletto the Duke of Mantua sings "La donna è mobile…e di pensiero" at the beginning, but later, in the reprise, he sings "La donna è mobil…e di pensier", of course to different music. Gilda's aria "Caro nome" in the same opera is a treasure trove of apocopation: "core" becomes "cor;" palpitare" becomes "palpitar;" "amore" becomes "amor;" "sospiro" becomes "sospir;" "rammentare" becomes "rammentar;" "fine" becomes "fin;" and "pensiero" becomes "pensier." Note that this list includes both nouns and verbs, but that the only letters apocopated (ever) are e's and o's. Apocopation also occurs frequently in recitativi accompagnati (ariosi).
The text of recitativi secchi may also be apocopated. The need for this, which may not be obvious at first sight, comes from the stringent rules under which these recitativi are constructed. Originally, each line of a recitativo was supposed to contain either eleven or seven syllables.10 By the time of Mozart, this had changed to seven or five syllables. Apocopation was used to stay within the guidelines. For example, in the second scene of Don Giovanni, just after the death of the Commendatore, the Don and Leporello are trying to sort out what has happened. (The syllabization is indicated.)
Don Giovanni: Le-po-rel-loo-ve se-i? (Seven syllables;
note the elision of the
Unlike the final e of French lyrics, it doesn't seem possible for singers to blithely ignore the apocopation of Italian lyrics. But just wait--someone is bound to come up with the idea, sooner or later.
So we have seen that both the French and Italian languages possess a flexibility which enhances their rhythmic integrity and greatly improves their beauty when sung. English has none of this flexibility, which perhaps explains why English-language operas, whatever their musical merits, are rarely noted for the beauty of their libretti. The exceptions are the operettas of Gilbert and Sullivan, but then Gilbert had already made his mark as a poet (albeit of light verse) long before he teamed up with Sullivan. But even Gilbert sometimes ran into trouble. For example, at the end of Patience, Reginald Bunthorne informs us that inasmuch as there is nobody left to be his bride:
In that case unprecedented single I must live and die.
(Note the masculine endings.)
And even the estimable librettists of Carmen, Henry Meilhac and Ludovic Halévy, had trouble with meter in another of their texts, Jacques Offenbach's comic opera La Périchole. Near the beginning of this work we hear the viceroy of Peru sing:
Je vais, je viens, je me vais faufi-le,
1. Welt der Oper. Decca CD 436 823-2.
2. Carmen. RCA Victor three CD set 7981-2-RG. Conducted by Fritz Reiner. Rïse Stevens, Carmen; Jan Peerce, Don José; Robert Merrill. Escamillo; Licia Albanese, Micaela. Robert Shaw, chorus director. (Date of recording not given but approximately 1956.) By the way, I have frequently heard the choeur des gamins sing "Sonne trompet-te eclatan-te" rather than the correct version as given in the score and in the title to this article. But not on this recording, as one would expect from the great Shaw.
3. Carmen. Angel LP SCL-3613. Conducted by Sir Thomas Beecham. Carmen, Victoria De Los Angeles; Don José, Nicolai Gedda. Escamillo, Ernest Blanc; Micaela, Janine Micheau.
4. Carmen. Eklipse Records Ltd. EKR CD6. Conducted by Louis Hasselmans: Rosa Ponselle, Carmen; René Maison, Don José; Ezio Pinza, Escamillo; Hilda Burke, Micaela. (1936). Carmen, Opéra Comique version. Angel LP SCL-3767. Conducted by Rafael Frühbeck De Burgos: Grace Bumbry, Carmen; Jon Vickers, Don José; Kostas Paskalis, Escamillo; Mirella Freni, Micaela. Carmen. RCA Columbia Pictures Home Video. Conducted by Loren Maazel: Julia Mignes-Johnson, Carmen; Placido Domingo, Don José; Ruggero Raimondo, Escamillo; Faith Esham, Micaela. This last recording uses spoken dialogue in place of Ernest Guiraud's recitativi, but is not the authentic opèra comique version.
5. The recordings listed in footnotes 2-4 were the only ones readily available to me. Actually only the two CD's are in my own collection; I found the LP's and the video in the university library. Although I own multiple recordings of many operas, I make it act of devotion to listen to no other Carmen than Rïse Stevens. The first time I heard Stevens as Carmen (in 1956) I knew that I would never want to hear another singer in the role. (I only bought the Eklipse Carmen to hear Ezio Pinza sing the Toreador Song.) Anyone who fails to appreciate the fact that Stevens is the only Carmen should listen to the Reiner recording cited above, especially the final scene. (It is possible that one needs a good working knowledge of French to appreciate this. Thanks to my parents for speaking French chez nous.)
6. An amphibrach consists of three syllables: weak, strong. weak.
7. Metropolitan Opera broadcast of March 19, 2005. First intermission. Available on the web at http://www.operainfo.org/intermissions
8. Poetry with feminine endings is notoriously difficult to write because polysyllabic rhyme schemes are often required. This is easier in French and Italian, with their stylized grammatical endings, than in English. (John Ciardi in the Translator's Note to The Purgatorio, Mentor Books, New York, 1957, points out how difficult English is to rhyme and uses this fact as an excuse for abandoning Dante's terza rima.) Note the easy rhymes in the Gamin's chorus: "Avec la gar-de mon-tan-te, Nous arrivons, nous arrivons, nous voilà. Son-ne trompette ecla-tan-te…" Similarly in Micaela's aria as cited in paragrph two above.
9. Even when speaking prose, native Italians, who have a very sensitive feeling for the cadences of their language, sometimes apocopate. This is one of the factors, along with the rapidity of Italian speech, which makes it difficult for non natives to understand spoken Italian, however well they may be able to read it and speak it themselves. (Here I am citing personal experience.)
10. The Oxford Illustrated History of Opera, Roger Parker, ed. Oxford University Press, New York, 1994. Page 27.
11. Why doesn't someone revive this wonderful opera? It boggles the mind that the Metropolitan can stage a piece of merde like The Merry Widow when it has ignored a wonderful work like La Perichole for so many years. Because of the extensive dialogue, this is an opera which should be performed in English translation; fortunately a splendid one exists, by Maurice Valency, who uses, incidentally, the identical construction on the word "incognito" as that in the French libretto. Perichole has a close connection with Gilbert and Sullivan, whose second operetta, Trial by Jury, was written as a curtain raiser to Perichole which was evidently not long enough to suit the London theater audiences. (The music to the first G&S operetta, Thespis, has been lost except for two fragments.) Trial by Jury actually opened, in 1875, on a triple bill with La Perichole and another piece, mercifully lost, the farce Cryptoconchoid Syphonostomata. (Gilbert and Sullivan, Ian Bradley, ed. Oxford Univ. Press, New York, 1996. pp. 3-4.)