Volume XIV, No 8, April 1999

On Translating Verdi's "Il trovatore"

Mark Herman and Ronnie Apter

How does one translate Verdi's Il trovatore, an opera popular with audiences, but whose supposedly absurd libretto and vulgar music have been reviled by critics for nearly 150 years, an opera that everyone from Gilbert and Sullivan to the Marx Brothers has ridiculed? Are all the critics simply wrong? Yes. Is Verdi's music really so great as to overwhelm all defects in Cammarano and Bardare's libretto? Yes. Is Il trovatore the quintessentially silly Italian opera? No, but you'd never know that from many previous translations. The work is an action thriller, a prototype of a Hollywood Western, and it should be judged, and translated, accordingly.

1. Introduction

On January 19, 1853, Giuseppe Verdi's Il trovatore was performed for the first time at the Teatro Apollo in Rome. It was a huge success, becoming and remaining until about 1950 the most performed opera in the world. Even the most recent survey available shows Il trovatore to have had eight separate productions during the 1995-96 season among those opera companies affiliated with Opera America, making it number twenty in popularity (Ref. 1). Interestingly, the most popular opera during 1995-96, with 23 productions among Opera America affiliates, was la traviata, another Verdi opera which opened less than two months after Il trovatore. la traviata, a domestic drama with realistic characters - including ordinary people pretending to be gypsies at a masquerade - set in contemporary (mid-nineteenth century) Paris, did not have nearly the same initial success as Il trovatore, a work peopled by real gypsies among generally larger-than-life characters and set in the late middle ages (1409) against the backdrop of a civil war.

Because of the success of Il trovatore and a few of the turns of its plot, usually mangled at least to some extent by numerous atrocious translations, the work was and is seen by many English-speaking critics as the quintessentially silly Italian opera. As such, it was frequently parodied. Gilbert and Sullivan made fun of at least two of Il trovatore's plot devices in at least three of their operettas: baby switching in H.M.S. Pinafore and The Gondoliers, and witches in Ruddigore - although, contrary to what many people believe, there are no witches in Il trovatore, and the baby switching may or may not have actually occurred. And when the Marx Brothers spent A Night at the Opera, the opera was Il trovatore (Ref. 2).

In 1996, we were commissioned by Ricordi in Milan, Verdi's original publisher, now owned by a conglomerate, to produce a new performable English translation of Il trovatore for a forthcoming critical-edition dual-language piano-vocal score. We tried to approach the opera with open minds, or if anything, with a slight bias in its favor. The standard opinion, we suspected, could not be valid, because silly stuff could not have held the attention of audiences for so long. So what is it about Il trovatore that so many critics have misread? Musicologist Roger Parker offers a clue: "...if one trait can be singled out that best accounts for the opera's success, it is probably the sheer musical energy apparent in all the numbers. Time and again we find a relentless rhythmic propulsion in the accompaniment, and a tendency for the melodic lines to be forced into a restrictive compass, freeing themselves rarely but with consequent explosive power" (Ref. 3).

Movement, propulsion, action - no subtle psychological exploration of the characters' interior worlds - in fact, no subtlety at all. Il trovatore, unlike many other operas, resembles nothing so much as a Hollywood action thriller.

2. The Opera and its Plot

Salvadore Cammarano's Italian libretto for Il trovatore, completed by Leone Emanuele Bardare after Cammarano died in July 1852, is based on the Spanish play El trovador [The Troubadour] by Antonio García Gutiérrez. Both play and opera are set during the civil war caused by the historical Count Jaime de Urgel's pressing of his claim to the crown of Aragon. Urgel himself does not appear in either work, but the main fictional character, Manrico, one of Urgel's lieutenants, shares many of the count's characteristics, including those which led to Urgel being called Jaime el Desdichado, or Jaime the Hapless (Ref. 4).

The driving force of the plot of Il trovatore is the enmity between Manrico and the Count de Luna. De Luna has three reasons to hate Manrico: 1) they are on opposing sides in the civil war; 2) they are rivals for the hand of Leonora, who loves only Manrico; 3) Manrico's mother, the gypsy Azucena, murdered De Luna's baby brother fifteen years earlier (or so de Luna believes).

This murder, though it occurs long before the curtain goes up, is the work's key incident. Azucena's mother (Manrico's grandmother) had intruded into the baby's bedroom in de Luna's father's castle. When the little boy subsequently became sick, the mother was accused of bewitching him and burned at the stake. In revenge, Azucena kidnapped the boy and burned him to death on the very spot where her mother had died. However, at the time, Azucena had her own infant son in tow, and, in her confusion, might have murdered her own son instead of the count's. If so, Manrico is in fact de Luna's younger brother whom Azucena has raised as her own son. Azucena tells Manrico that this is indeed the case, although she then recants her story. At the end of the opera, after de Luna has killed Manrico, she gloats and tells de Luna that he has killed his own brother, and that therefore the revenge of her mother's death is now complete.

But the script is indefinite, and the opera can be played in at least three different ways: 1) Azucena has indeed killed her own son and raised the count's son as her own; 2) Azucena killed the count's son, and, horrified at what she had done, changed the victim in her telling of the story in order to at least partially expiate her guilt; later, she again lies about the victim in order to make the count believe he has killed his own brother; 3) Azucena is so wracked with guilt and her wits so confused that she does not know which of the two babies she murdered, and therefore no one else can ever know either. The opera suggests that 1) above is correct, in that Manrico, when he has a chance to kill de Luna, is prevented from doing so by a "grido dal cielo" [cry from heaven]. However, this suggestion is countered by the fact that nothing impedes de Luna's killing of Manrico.

By the end of the opera, de Luna has killed Manrico (whom he now believes is his brother), Leonora has committed suicide by poison, and there is every expectation that Azucena will be burned at the stake as soon as the curtain comes down. And, of course, historically, Urgel's forces lost the war.

3. Translating Il trovatore

A given of opera translation is that, except in rare extreme circumstances, the music must not be changed to accommodate the new words. But many translators also believe that the translation must re-create the rhyme scheme of the original. However, because English is a relatively rhyme-poor language, rhymes, especially multi-syllabic (feminine) rhymes cannot be used in the same way that they are used in a language like Italian. Even assuming that the English grammar has not been stilted beyond recognition to allow for rhyme words at the ends of lines, and that the rhyme words themselves are not trivially silly, too much rhyme (and again especially feminine rhyme) can, in English, have a comic effect in the manner of Gilbert and Sullivan, imparting totally the wrong tone to a serious opera. But a bad translation, and bad rhymes in particular, can do even more damage to an opera like Il trovatore: they can slow down or stop the action.

3.1. Example 1

Consider a section of the long aria that constitutes the first scene of the opera. Ferrando, a captain in de Luna's guard, is telling the soldiers about the events leading to the death of Azucena's mother at the stake, Azucena's revenge, and her subsequent escape. Verdi achieves the forward motion in this particular aria largely by means of tempo changes. Each verse starts out in a straightforward 4/4 at 88 beats per minute, switching to a more rapid 3/4 at 112 beats per minute as the events described become more exciting. After some bridging recitative, an agitated coda describing the men's fear of witches and witchcraft is in 3/4 at a breakneck 216 beats per minute.

The lyrics of concern here are from the end of the second verse (3/4, 112 beats per minute). The original Italian text is given in the left column below, with rhyme words in bold (the final rhyme 'ancor' rhymes to lines that come later). In the right column is William Weaver's literal non-rhyming non-singing translation (Ref. 5, p. 88), with some re-wording to make it follow the Italian meaning more closely on a line-by-line basis (Weaver followed a slightly different division of the lyrics into individual lines):

The original Italian text is given in the left column below, with rhyme words in bold (the final rhyme 'ancor' rhymes to lines that come later). In the right column is William Weaver's literal non-rhyming non-singing translation (Ref. 5, p. 88), with some re-wording to make it follow the Italian meaning more closely on a line-by-line basis (Weaver followed a slightly different division of the lyrics into individual lines):

Notice that, in the Italian, these 20 lines rhyme as:

a', a', b', b', c', d', c', x, x, d', x, x, x, x, x, e, x, e, x, e,

where a', b', c', and d' are feminine rhymes, while e is a masculine (one syllable) rhyme. Further, as indicated, seven lines are repeats of one of three other lines: line 8 is repeated in line 9; line 13 is repeated in lines 14, 15, 17, and 19; and line 16 is repeated in lines 18 and 20.

Natalia Macfarren (1827-1916) was a famous British translator of the Victorian era, but her work is still reprinted today. Her translation of this passage - with rhyme words again in bold type - is (Ref. 6, pp 9-11):

Macfarren follows the Italian rhyme scheme exactly. She repeats all lines as in the Italian except for line 9, which does not repeat line 8. And she uses minimum resyllabification. That is, a glance at the music and its word underlay would verify that, almost without variation, Macfarren's English syllables correspond to the original Italian syllables one-for-one.

Of the many things wrong with Macfarren's translation, we will consider here only how it stops the movement dead. Without knowing anything about the music, or how the lyrics fit it, it is easy to see that two phrases catch the lyrics up and bring everything to a screeching halt. They are "she lives secluded" and "murd'rously to bind him." The first is a totally superfluous thought, thrown in only for the sake of a rhyme. The second includes the gratuitous tri-syllabic action-stopping adverb "murd'rously," again thrown in only as filler to place the rhyme words "bind him" at the end of the line.

If it is indeed the difficulty of rhyming in English that is the cause of the awfulness of Macfarren's translation - and we admit that the difficulty of rhyming is at least a partial cause - how can this difficulty be overcome? One suggested solution is to dispense with rhyme altogether. This sometimes works. But new words to music originally written for rhymed lyrics must usually either be rhymed or achieve sonic closure in some other way if they are to fulfill the expectations aroused by the music.

It turns out that "achieving sonic closure in some other way" is the key to solving the problem. Two alternative means of sonic closure are 1) rhyming, but according to a different scheme than in the original, and 2) using something other than true rhyme, such as off-rhyme or half-rhyme. Other means, not illustrated here, are the use of alliteration, assonance, and consonance. And two ways of obtaining more flexibility for the English words that lead up to the rhyme words are to not repeat as in the original and to re-syllabify, that is, have more or fewer musical notes per syllable than in the original. Of course, all the resulting lyrics must still fit the music.

Here is our translation-in-progress of the Italian text cited above, incorporating all of the ideas mentioned in the preceding paragraph. Since the translation has not yet been through the editorial process, it is still subject to change.

The rhyme scheme of this translation is:

a'*, a'*, b', b', c'*, c'*, c'*, x, x, c'*, c'*, x, x, x, x, d, x, d, x, d,

where asterisks indicate half or off-rhymes. This may be compared with the rhyme scheme of the original:

a', a', b', b', c', d', c', x, x, d', x, x, x, x, x, e, x, e, x, e.

In the English translation, the two sets of off-rhymes (the a's and c's) are:

a: tried her, fire - "fire" is stretched into two syllables: fie-er

c: malice, vanished, perished, ashes - the vowel sounds of "malice," "vanished," and "ashes" are more or less the same, although the consonant sounds are completely different. The "- ished" of "perished" is identical to that of "vanished," although "perished" has nothing else in common with the other three c words.

Our translation does not repeat as in the original: line 9 is only a partial repeat of line 8; lines 14 and 15 repeat line 13, but lines 17 and 19 have a partial change of wording; lines 18 and 20 do repeat line 16 as in the original. And a glance at the music and the verbal underlay would indicate that we have re-syllabified considerably. Despite, or rather, because of the changes, our translation can be sung to the original music and still maintain the momentum. However, this is apparent only when the words are actually sung rather than read - libretto lyrics are not the same as spoken poetry. Readers with access to the music or a recording of Il trovatore are advised to mentally fit our translation to the music and "hear" the result for themselves.

3.2 Example 2

Near the end of Act III, just as a wedding ceremony between Manrico and Leonora is about to be performed, Manrico learns that Count de Luna has captured Azucena and is about to burn her at the stake. Manrico rushes out, totally against Leonora's wishes, to rescue his mother, as he had previously rushed out in Act II, totally against Azucena's wishes, to rescue Leonora. However, while his rescue of Leonora was successful, his attempt at rescuing Azucena will prove to be spectacularly unsuccessful.

Verdi propels the music forward by setting the tempo in a moderately fast 3/4 at 100 beats per minute, and by means of a rhythmic figure in the accompaniment starting on the second half of the first beat of each measure. This rhythmic figure, which Verdi used many times in his operas, can easily be described as "vulgar" when played loudly by the orchestra, as it usually is, in accordance with corrupt scores. However, Verdi's original dynamic marking, as the new critical edition reveals, was piano (soft).

Below, in the same format as for Example 1, is what Manrico sings as he rushes away from Leonora to rescue Azucena. Again, William Weaver's literal translation has been somewhat re-worded because he followed a different division of the lyrics into individual lines (Ref. 5, p. 139):

In the Italian, these 20 lines rhyme:

x, a', x, b, x, a', x, b, x, c', x, d, x, c', x, d, x, d, x, d,

where rhymes a and c are feminine, while b and d are masculine. Further, lines 17 through 20 are repeats, with minor variations of lines 15 and 16.

Here is Natalia Macfarren's translation (Ref. 6, pp. 170-72):

Once again we will consider only the problem of how Macfarren's translation, which follows the Italian rhyme scheme exactly, slows down the action. As before, that particular fault is due to irrelevant thoughts thrown in for the sake of rhyme words, namely "though my heart break," which gives the audience a momentary pause, and "I, who implor'd thee," which stops everything dead. "Implor'd thee" to what?

Here is our own translation-in-progress:

Our translation follows the rhyme scheme:

x, a'*, x, b, x, a'*, x, b, c'*, d'*, c'*, e, c'*, d'*, x, e, x, e, x, e,

with no true feminine rhymes, but with three feminine off-rhymes (as opposed to two feminine true rhymes in the original). These are:

a: fiber, ember - with the same second unaccented syllable;

c: suffer, mother - with the same vowel sounds;

d: stay here, save you - with the same first accented vowel.

As before, because of the changes, our translation can be sung to the original music and still maintain the momentum. This is only apparent when the words are sung rather than read.

4. Conclusion

Opera lyrics are meant to be sung, and must not only be singable, but worth singing. There is no opera without words, just as there is no opera without music. A bad translation is an insult to the creators of the work, the performers, and the audience. It can do more harm than merely make a good work seem bad, a serious work seem at times comic or silly, and plot details or character motivations seem incomprehensible. A bad translation can change the very nature of a work and cause it to be judged by irrelevant standards. For Il trovatore, the question to be asked of a translation, in addition to the usual ones for operas - does it sing? does it play on the stage? does it present the actual plot? does it maintain the characterizations? - is the all important one: does it get in the way of the action, or does it let the opera rush ahead?

5. References

1. Opera America, Washington, D.C. 1996 Annual Field Report, p. 11.

2. The Marx Brothers, A Night at the Opera, Metro-Goldwyn-Mayer, 1935.

3. Parker, Roger. Il trovatore, in The New Grove Dictionary of Opera, edited by Stanley Sadie. Grove's Dictionaries of Music Inc., New York, 1992. Vol. 4, p. 827.

4. Morgan, Christopher. Don Carlos and Company. Julia MacRae, 1994. Reprint, New York: Oxford University Press, 1996, pp. 33-34.

5. Weaver, William, translator. Seven Verdi Librettos. New York: W. W. Norton & Company, 1975.

6. Macfarren, Natalia, translator. Il trovatore. New York: G. Schirmer, 1898.

Mark Herman is a freelance translator in Shepherd, Michigan, and Ronnie Apter teaches at Central Michigan University in Mt. Pleasant, Michigan. This article is based on their presentation at the 39th Annual Conference of the American Translators Association, Hilton Head, South Carolina, on 4-8 November 1998. Their e-mail address is:herman.apter@sensible-net.com. We are indebted to them for granting permission to publish their article in the Gotham.