The Dyers Hand and Other Essays

In comparison with the ordinary translater, the translator of a libretto is much more strictly bound in some respects and much freer in others. Since the music is so infinitely more important than the text, the translator must start with the premise that his translation must demand no change of musical intervals or rhythms in order to fit it. This law is absolute for arias and ensembles; in recitative, occasions may arise when the drop­ping or addition of a note is justified, but they are very rare. The translator of a libretto, therefore, has to produce a version which is rhythmically identical, not with the verse prosody of the original as it would be spoken, but with the musical prosody as it is sung. The difficulty in achieving this lies in the fact that musical prosody is both quantitative, like Greek and Latin verse, and accentual like English and German. In a quantitative prosody, syllables are either long or short and one long syllable is regarded as being equal in length to two short syllables; in an accentual prosody like our own, the length of the syllables is ignored—metrically, they are re­garded as all being equal in length—and the distinction is between accented and unaccented syllables. This means that the rhythmical value of the trisyllabic feet and the dissyllabic feet are the reverse in a quantitative prosody from what they are in an accentual. Thus

A quantitative dactyl or anapaest is in 4/4 or 2/4 time. (March time.)

A quantitative trochee or iamb is in 3/4 or 6/8 time. (Waltz time.)

An accentual dactyl or anapaest is in waltz time. An accentual trochee or iamb in march time.

But in music both quantity and accent count:

A 2/4 bar made up of a half note followed by two quarter notes is, quantitatively, a dactyl but, accentually, a bacchic.

A musical triplet jf) is, quantitatively, a tribrach but, accentually, a dactyl.

To add to the translators' troubles, the felt tempo of the spoken word and of musical notes is utterly different. If, timing myself with a stop watch, I recite, first the most rapid piece of verse I can think of—The Nightmare Son from Iolanthe, let us say—and then the slowest verse I can think of—Tennyson's Tears, idle tears—I find that the proportional difference be­tween the time taken in each case to recite the same number of syllables is, at most, 2,-1, and much of this difference is attributable, not to the change in speed of uttering the syllables but to the pauses in speaking which I make at the caesuras in the slow piece. Further, the two tempi at which I speak them both lie in what is in music the faster half of the tempo range. The tempo which in speaking verse is felt to be an adagio is felt in music as an allegretto. The consequence of this differ­ence is that, when a composer sets verses to a slow tempo, verse dactyls and anapaests turn into molossoi, its trochees and iambs into spondees. The line Now thank -we all our God is iambic when spoken but spondaic when sung.

This means that it is not enough for the translator to read the verses of the libretto, scan them, and produce a prosodic copy in English for, when he then matches his copy against the score, he will often find that the musical distortion of the spoken rhythm which sounded possible in the original tongue sounds impossible in English. This is particularly liable to happen when translating from Italian because, even when speaking, an Italian has a far greater license in prolonging or shortening the length of his syllables than an Englishman.

Two Examples

O In Leporello's aria at the beginning of Don Giovanni occurs the line Ma mi par che venga gente (But it seems to me that people are coming).

To begin with, we decided that Leporello must say something else. He is on guard outside the house where Don Giovanni is raping or trying to rape Donna Anna. Da Ponte's line suggests that a crowd of strangers are about to come on stage; actually, it will only be Don Giovanni pursued by Donna Anna and some time will elapse before the Commendatore enters. Our first at­tempt was

What was that? There's trouble brewing.

Spoken, che venga gente and there's trouble brewing sound more or less metrically equivalent, but the phrase is set to three eighth notes and two quarter notes, so that gente which, when spoken, is a trochee becomes a spondee. But brewing, because of the lack of consonants between the syllables, sounds distorted as a spondee, so we had to revise the line to

What was that? We're in for trouble,