I'm sorry for the improper use of notation and lack of terminology, but I know very little about meters.

There is this famous hymn in Catholic liturgy called Veni Creator [Spiritus]. AFAIK, it follows a very established meter, and the Gregorian melody reflects so. The fourth stanza reads:

Accende lumen sensibus,
infunde amorem cordibus,
infirma nostri corporis
virtute firmans perpeti.

The final e in infunde is sometimes rendered in italics, since —I think— the meter demands an apheresis/elision. However, it sounds natural to me to make a diphthong from e+a, even if it is not a valid diphthong in Latin. (FWIW, in Spanish /ea/ is always a hiatus, but it still sounds natural to this native Spanish speaker to pronounce something like this as a diphthong in a Spanish song.)

In this specific song, I think I have heard different renderings, including an /ea/ diphthong and even breaking the long last syllable in amorem which in other verses would take two notes, to give space for an additional syllable (in.fun.de.a.mo.rem instead of in.fun.da.mo.re.em).

Now, my question is: has de "illegal diphthong instead of elision" some kind of historical/technical support (even if only in ecclesiastical or medieval Latin), or is it just a XX/XXI century, "illiterate" mistake?

2 Answers 2


I think hardly anyone uses reconstructed Classical Latin pronunciation for plainchant, so I'm not sure you need the "even" in "even if only in ecclesiastical or medieval Latin".

It has been difficult for me to find information online about the use and realization of elision in Gregorian chant. The circumstances in which it occurs do not appear to always be the same as in Classical Latin poetry; because of this, I think that resources about the pronunciation of elision in Classical Latin may not be applicable to the pronunciation of apparently "elided" syllables in chant.

For example, the website Gregorian Chant Hymns* shows that it occurs in this hymn not only in the line "Infunde amorem cordibus" ...

enter image description here

But also affecting a vowel between two consonants in "digitus paternae dexterae":

enter image description here

*The music on the website is noted to be sourced from the Parish Book of Chant (Church Music Association of America, 2012, page 208).

That said, surprisingly to me, I found a source saying that a word-final vowel or a word-final "m" along with the preceding vowel can be elided before a following vowel in chant; elision of "m" in this context is a well-known feature of Classical Latin but not as far as I know a practice observed in normal speech by any modern "Ecclesiastical Latin" speakers (Chapter X, "The Hymns", §106 Elision., in A New School of Gregorian Chant by Dominicus Johner, 1906, translated by H. S. Butterfield):

If a word ends with a vowel or the consonant m, and the following word begins with a vowel or the consonant h, the vowel (also the m) of the first word is omitted (elided). Therefore, e.g., in the hymn for the feast of the Sacred Heart: ille‿abstulerat, marisque‿et, illa‿amoris, percussum‿ad, regnumque‿in omne‿est.

(page 90-91)

However, in the second English edition of the same book, the section on elision seems to have been revised to say something quite different:

If there occurs in any line a syllable in excess of the usual number in the metre (such syllables are printed in italics in the Antiphonale) -- it must nevertheless be sung (thus elision is no longer permissible)

(page 112, Chapter IX, "The Hymns", §113 "Secondary syllables" in the hymns, in A New School of Gregorian Chant, second English edition, translated from the third German edition by W. A. Hofler, 1914).

The passage goes on to say that, if not separated from the preceding syllable by a hyphen or similar mark, the syllable printed in italics should be sung with the same pitch as the preceding note, giving the following example:

enter image description here

To me, this description implies that in this pronunciation, not only is the vowel not lost entirely, it isn't even glided with the following vowel to form a [e̯a] diphthong in the way that you describe (which is a natural feature of vowel + vowel combinations between words in Spanish, called synaloepha).

And I found a modern resource that seems to say that elision is optional:

Hypermetric Syllables in the Hymns
LU 127: “According to a decree of the S. C. of Rites, dated May 14th 1915, hypermetric or redundant syllables in the Hymns may be elided, if this method of interpretation be considered easier or more fitting. “Two methods are therefore allowed: “either (a) the pronunciation of the hypermetric syllable, by giving it the separate note allotted to it in the notation, according to the rules indicated in the official edition of the Roman Antiphonary. “or (b) the suppression of the hypermetric syllable by elision, thus keeping the ordinary melodic formula.”

(Ad Libitum Variants in the chant for the Divine Office - DRAFT, A survey by Jonathan Kadar-Kallen, November 7th, 2016; original underlining replaced by me with italicization)

You can indeed find the passage that Kadar-Kallen quotes above on page 141 of the English version of the Liber Usualis (1957).

But if this is accurate, the acceptability of elision doesn't seem to have been known in 1921 to H. T. Henry, the author of an article "Office Hymns of St. Jeanne D'arc", published in The Ecclesiastical Review, Volume 64, which cites Johner 1914's statement that elision is no longer permissible in the singing of hymns (page 492).

As far as I know, the realization of "infunde" isn't related to the realization of the "rem" in "amorem". The Gregorian Chant Hymns website shows "rem" here as being pronounced as a "clivis", two descending notes. This might sound a bit like two syllables, but metrically it has to be counted as one syllable to keep to the 8 8 8 8 meter.

  • Thank you! I had been thinking of this question for a long period now. At some point I thought of including the partiture too, but I forgot. The first time I learned about the Veni Creator, it was sung as in.fun.de.a.mo.rem and di.gi.tus.pa.ter.næ thus breaking the clivis (a term I didn't know until now, thank you BTW.) The partiture is why I assumed that rendering was wrong.
    – Rafael
    May 13, 2022 at 17:22

The main modern source I can think of on this topic is Allen's Vox Latina chapter 4, which takes most of its evidence from Sturtevant's Elision and Hiatus in Latin Prose and Verse, so that's what this answer is based off. But note that it's focused primarily on the Classical and pre-Classical period; evidence from later eras is harder to come by.

The gist is that various types of evidence all point to complete loss of the preceding vowel in Classical Latin: if the preceding vowel turned into a semivowel we might expect it to make position, which it doesn't, and if it were pronounced as an extra-short vowel it would be difficult to deal with four elisions in a single proceleusmatic* which appears in Plautus.

* A rare type of metrical foot consisting of four shorts in the span of three shorts. I had to look it up for this answer.

We also see puns relying on the final vowel being lost, like a joke in Plautus where a character can't tell whether someone is sleeping cum catellō or cum catellā due to a following vowel, and grammarians prescribe that this vowel perit and necesse est excludī. This continues well past the fourth century, but how well this reflects the actual speech of later times is an open question (the grammarians tended to rely more on their predecessors than on how people actually spoke at the time).

On the flipside, Quintilian insists that the vowel shouldn't be lost when the elision happens across a final m, as in Vergil's multum ille. This doesn't seem to have been the case in poetry but Sturtevant suggests it may have been standard in prose.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.