THE GREGORIAN RHYTHM WARS seem to have settled into something of a winter truce. In the most recent post in the series, now over a month old, Patrick Williams has proposed many questions that have so far gone unanswered. Since the principal combatants in this war are pushing proportionalism on the one hand and a non-Mocquereau version of the Vatican Edition rhythm on the other, I don’t want to step in and try to carry on the argument. My role in the series has been to represent the claims of the nuanced approaches taken by those who follow either the Classic Solesmes method or semiology, both of which I follow in my own practice. It doesn’t seem that any of us has changed any of our opinions about anything yet discussed!
In this post, I will indeed discuss Gregorian rhythm, but I’d like to take a different approach and describe an aspect of that rhythm that has almost nothing to do with the signs of length that have been the subject of dispute so far in the rhythm wars series. I think we may actually be able to find some common ground. I’m talking about the accentuation of Latin. In particular, I want to define a term that gets thrown around a lot: the tonic accent. Tonic is one of those overloaded words, and we should always define our terms carefully. I have heard many people use this term in a way that does not reflect its real meaning.
Why do I care about the Latin accent? I share the conviction of many other chant practitioners that Gregorian chant is intimately connected with the accent pattern of the Latin language. Indeed, I think singing Gregorian melodies in English is a complete waste of time, and while I know that opinion may irritate some people, I hope to make my reasons clear below.
How We Pronounce Latin • To begin with, let’s set music aside altogether and just think about Latin itself as a spoken language. Of course, with negligible exceptions, Latin is not spoken anywhere in the world as a living language. This means that when we decide how to pronounce it or speak it, we depend on historical and philological research to give us guidance in how the language was pronounced in the past. The pronunciation of Latin has changed over time and has also differed in different places. Consider the way we pronounce Latin phrases like habeas corpus that have come to take on a meaning in juridical English. That pronunciation differs greatly from the way we might pronounce it if we sing it in church. Perhaps you have attended a concert or a Mass where the Latin is pronounced according to the French tradition, with nasalized vowels and other French-language-like features. Nowadays, as a matter of unity, we generally try to stick to a “Roman” pronunciation of Latin when we sing in church, and this ecclesiastical or “Roman” Latin differs significantly from the reconstructed classical pronunciation you might have studied in school.
The Accent • One particular question of Latin pronunciation that is of particular interest to us as church musicians is what to do with the differentiation between accented and unaccented syllables. At some point in the classical era—that is, roughly the period encompassing the history of the Roman republic and empire—the Latin language developed its characteristic of a single accented syllable within the word. The accent always falls on either the second syllable or the third syllable from the end of the word. This placement depends to some extent on the older practice of vowel quantity, which distinguishes between long and short vowels. In any Latin dictionary you will see the vowel quantity marked, as it seems to have had a real bearing on pronunciation in the early centuries. If the second syllable from the end of the word is long, then the accent falls there, as in Salvátor, but if the second syllable from the end is short, then the accent gets bumped to the third syllable from the end, as in Dóminus. Over time, the quantity or length of the vowels became much less of an actual feature of spoken Latin, which is why our liturgical books do not mark vowel length like those Latin dictionaries.
Everyone agrees on that much. But when it comes to actually pronouncing the accent in Latin, we run into some controversy and difficulty. The problem is that different languages treat the phenomenon of accent differently, and the structure of our native speech often colors the way we want to pronounce other languages. This is one reason why it is so hard to lose one’s accent when learning a new language. Some languages rely on a heavier stress or weight on the accented syllable, while others rely on length. Let’s think about the accent in terms of three dimensions. We can bring out the character of the accented syllable by using:
- Durational stress. Pronouncing the accented syllable with more length than the others.
- Dynamic stress. Pronouncing the accented syllable with more volume than the other syllables.
- Pitch stress. Pronouncing the accented syllable at a different pitch from the others.
The last of these, the dimension of pitch, is precisely what we mean when we use the term “tonic accent.” A tonic accent is when we give accent to a syllable by using the pitch (i.e., high or low utterance) of our voices in speaking. We have abundant evidence from Latin authors and grammarians that the Latin accent was a tonic accent in which the accented syllable is spoken on a higher pitch than the syllables around it. At the same time, we know that classical Latin did not have a durational accent, since the accent is spoken in each word regardless of the length or brevity of the accented syllable. This is especially clear in poetry, where the pronunciation of the meter often does not align with the accented syllables. Perhaps you are familiar with the first line of the Aeneid:
Arma virumque cano, Troiae qui primus ab oris
In this line, arma, cano, and Troiae all have the same accentual pattern, with the accented syllable first and the unaccented syllable last. But in arma (as also in primus) the length is long-short, while in cano the length is short-long, and in Troiae it is long-long. So the accent is independent of the length of the syllable. See the setting of this line by Ludwig Senfl to understand how the longs and shorts of the prosody work:
In this setting, each measure corresponds to one foot of the classical epic meter, known as dactylic hexameter (six feet per line). Each foot (measure) contains syllables either in a pattern of long-long or a pattern of long-short-short. Notice that, as described above, the word accents do not necessarily correspond to the long notes.
That leaves dynamic stress. Some (mostly French) scholars have argued that the dynamic stress is light or non-existent in classical Latin, while other (mostly German and English-speaking) scholars have argued for a stronger dynamic stress. Given the difficulty of reconstructing speech patterns from millenia in the past, I doubt this question will ever be completely settled.
The Latin Accent and Liturgical Chant • What does all this have to do with Gregorian chant? It turns out that these issues of Latin accent (accented and unaccented syllables) and prosody (long and short syllables) have long played a role in the way chant melodies have changed over time. For instance, one primary reason that the chant melodies were drastically simplified in the decades following the council of Trent was that the musical scholars of the time objected to the way the Gregorian melodies seem to ignore the prosody of words. In a word like Dómine, the second syllable is undoubtedly short, which is why we pronounce it with the accent on the first syllable. But in the Introit to the Mass for the Dead, we get a setting of the word Dómine where the short syllable -mi- has more notes than either of the syllables around it:
This was seen as faulty Latin by reform-minded scholars of the sixteenth century and following. This view is exemplified by the great theorist Gioseffo Zarlino, writing in 1558:
We should also take care to accommodate the words of the text to the written notes in such a manner and with such rhythm that no barbarism is heard, such as when in a vocal piece a syllable that should be short is made long, or vice versa, a syllable that should be long is made short, something heard every day in innumerable compositions and really a shameful thing. This vice is found not only in mensural music but in plainsong as well, as is obvious to all who have judgment. Indeed there are few chants not full of similar barbarisms and in which repeatedly a length of time is given to the penultimate syllable of the words Dominus, Angelus, Filius, Miraculum, Gloria, and to many other syllables that pass quickly. It would be very commendable to correct this, and it would be very easy, for by a very small change the composition would be adjusted. Nor would its original form change by this, for the difficulty lies only in a ligature of many notes, placed over short syllables, which makes them inappropriately long, when a single note would have sufficed. (Le istituzione harmoniche, book 4, translated by Vered Cohen.)
Many post-Tridentine chant books followed precisely this approach, simply eliminating melismas on syllables whose quantity should be short according to classical prosody. Now when the restored editions started being produced at Solesmes and elsewhere in the late nineteenth century, the editors needed to justify and defend the Latinity of the restored Gregorian melodies. One way they did this was by appealing to the tonic accent. According to this theory, the tonic accent, rather than the quantity of the syllables, was the organizing rhythmic principle of Gregorian chant. This insight can be seen in the work of Dom Pothier and possibly stems from the teaching of Dom Guéranger. It is also a key feature of the work of Peter Wagner and Dom Ferretti, and it forms the backbone of Dom Mocquereau’s rhythmic theory, with most of the second volume of Le nombre musicale grégorien being devoted to the nature of the tonic accent in plainchant.
The Tonic Accent in Gregorian Melodies • According to the principle of tonic accent in spoken Latin, the accented syllable should be pronounced higher than the syllables around it. And in Gregorian chant, it is often the case that the accented syllable is sung on a higher note than the syllables around it. This feature is so common that most of us who sing chant probably encounter it every week without thinking of it. For one thing, it’s a feature of all the modern psalm tones and is the basis of the liturgical recitative tones used to sing the prayers and readings at Mass. According to the proponents of the tonic accent theory, it is also a feature of the more complex melodies. Consider the beautiful communion chant for the fourth Sunday of Lent:
Notice how many of the words fit the pattern where the accented syllable is higher either than what precedes it or what follows it, and often both. This is especially noticeable at passages like “ejus in idipsum” or “tribus, tribus, Domini.” But you can find it on other words too: aedificatur, ascenderunt, nomini. It is even possible that the first word, Jerusalem, is set with the accent following the Hebrew pronunciation, with the tonic accent falling on the final syllable.
The Tonic Accent and Rhythmic Approaches to Chant • Now for both Pothier and for Mocquereau, but in different ways, attention to the tonic accent forms the basis for the entire approach to Gregorian rhythm. The idea here is that the accented syllable is normatively high, and as a result is not heavy and not necessarily lengthened but sung with an energetic emphasis that governs the dynamic shape of the word. Mocquereau extends the idea a bit further (although still along the lines established by Pothier) to say that the proper character of the accent (high and with a light dynamic stress) belongs more to the upbeat than to the downbeat. So perhaps we can find some common ground here between our various favored rhythmic approaches. Ostrowski often posts about Dom Mocquereau’s rhythmic signs, but he tends to focus only on the episemata, which are rather marginal to the theory being discussed here. I would say that the theory of the ictus and the accent are much more central to understanding the Classic Solesmes method than the additions of the rhythmic signs. To the extent that he finds this style of singing plodding, I would suggest that it is probably because the singers are paying too much attention to the signs and not enough to the tonic and melodic accent, which is the real governing principle of the method.
As for the rhythmic proposals offered by Williams, I would suggest that what I have outlined here shows that Mocquereau’s and Pothier’s approaches are based on something rather different and separate from the earliest neumes that form the basis of the proportionalist rhythmic systems. This accounts for what Williams sees as Mocquereau’s rather light editorial touch with adding episemata and so forth. Mocquereau’s deep conviction (drawing on the older Solesmes tradition of Pothier) of the tonic accent as the organizing principle of Gregorian rhythm and melody must have made the mensuralist approach completely unacceptable to him from the start. I have examined, at Solesmes, a letter from the musicologist Hugo Riemann to Mocquereau, in which, after trying to convince Mocquereau of a basically mensuralist approach to chant rhythm, he writes: “But all this is well known to you; I understand well that for you there are limits to respect, which I do not know.” I think those limits are precisely the Solesmes-derived tradition of rhythm based on the tonic accent. I’m sure I won’t convince anyone of the merits of the Classic Solesmes approach with this post, but perhaps we can at least be clear on what we mean when we say “tonic accent.”
Anima vocis et seminarium musices • For the proponents of Mocquereau’s methods, the tonic accent is the unifying characteristic of the word. In fact, we can go further and quote the fifth-century author Martianus Cappella in saying that the accent is the “soul of the voice and the seedbed of music.” I find this thought to be a worthy object of our extended meditation as church musicians who sing in Latin. Whether or not you follow the various rhythmic systems discussed on this blog, perhaps you share my observation, based on the daily singing of chant, of the marvelous way in which the words fit to the neumes, with especial attention to the tonic accent as the “soul of the voice.”
Much modern scholarship has moved away from the study of the tonic accent in chant. You won’t find anything about it in the magisterial books by David Hiley. Some scholars, including Terrence Bailey, have dismissed the whole notion as Romantic wishful thinking. It is interesting that this feature of the chant—so obvious to Pothier, Mocquereau, Riemann, and Wagner in spite of all their differences on the question of chant rhythm—has become a complete non-starter within the scholarly discussion of chant. Fashions change! The next time you sing a chant, think about how well it conforms to the idea of the tonic accent, with high melodic points falling on accented syllables. Indeed, I think this aspect of the chant is intimately and inextricably linked to the Latin language. By comparison, chanting in English lacks that “soul of the voice.” Does this accord with your experience as practitioners?