The Most Important Chant Manuscripts

I prepared the following article in response to Facebook comments from Corpus Christi Watershed challenging me to defend my ideas publicly, which I have already done on this website and my YouTube channel, as well as in various comments on Corpus Christi Watershed’s posts.  Their comments, with provocative and disrespectful language—for example, “indefensible,” “silly,” “absolutely bonkers,” “beyond far-fetched,” “fanciful”—can be read below the August 25 and 31, 2022, posts (assuming they haven't been removed by the time you see this).  See also the comments from July 5, 2022, and January 12, 2021.

At the Second Vatican Council, paragraph 117 of the Constitution on the Sacred Liturgy Sacrosanctum Concilium called for the preparation of “a more critical edition” of the chant books.  A little less than 60 years prior, Pope St. Pius X had ordered their thorough revision: “The melodies of the Church, so-called Gregorian, will be restored in their integrity and purity according to the testimony of the most ancient codices, in such a manner that due account is also taken of the legitimate tradition contained in the codices throughout the centuries, and of the practical usage of today’s liturgy” (Motu proprio Col nostro, April 25, 1904).  Although several revised, restored, or corrected editions have appeared, the more critical edition envisioned by Vatican II has yet to be promulgated nearly another 60 years later.  The Oxford Reference website furnishes a succinct definition of what constitutes a critical edition: “A scholarly edition that does not replicate the text of one document, but rather presents a corrected text, compiled from one or more source documents, and an apparatus recording editorial emendations and textual variants.”  My objective is not to explore the reasons why an official critical revision of the Roman Gradual is still not available, but rather to identify the most reliable sources for such a revision.

It is of vital importance to determine which manuscripts give rhythmic indications that accurately represent the authentic tradition.  I wish to reproduce a rather lengthy quotation from Fr. Jan Vollaerts, S.J., which explains the precise nature of the problem:

If it could be proved that the so-called rhythmic MSS were older than the non-rhythmic MSS, then it would be fairly obvious that the latter belonged to a place and time in which the rhythmic tradition had been lost.  Hence, without further proof, the non-rhythmic MSS would have less authority than the rhythmic MSS, and would certainly never be able to play a dominant part in the interpretation of those same rhythmic MSS.  However, there is no need for the argument of greater antiquity concerning the rhythmic MSS, for their superior value may be ascertained without establishing their greater antiquity.  The argument for this rests with the clear and unanimous pronouncements of the great medieval authors.  From the fifth to the twelfth centuries, it was their general teaching that rhythm was metrical, there being ‘breves’, ‘longae’ and ‘duplo longiores’.  Moreover, they tell us that the long and short durations of the sounds could be shown by the shape of the neums, and also that special letters indicated short and long sounds.  Hence it follows that the rhythmic MSS which have come down to us bear a truer representation of the melody than the non-rhythmic MSS.  Also, all the medieval rhythm theories agree only with the rhythmic MSS; indeed, the indications of ‘longa’ and ‘brevis’ fail in the non-rhythmic MSS.  The crumbling of the tradition [. . .] is clearly confirmed when we examine those neum-MSS which may show rhythmic indications but owing to increasing incompleteness and inaccuracy, have to be grouped in a secondary class.  These, in turn, run over into a third subsidiary grouping in which the original rhythmic indications (e.g. the episema) lose their significance and become pure graphic forms. [. . .] In these three classes of MSS it is possible to follow closely the gradual disappearance of the rhythmic tradition.  However, should one be tempted to rank the non-rhythmic MSS above the rhythmic MSS of the first and second classes above, a state of conflict will arise between the rules of the medieval theorists and the chant-books (neum-MSS).  In no way can this conflict be settled; the theorists prescribe metrical feet of ‘longae’ and ‘breves’ (2 : 1), while the non-rhythmic MSS omit the graphic indication of these different durations.  Nevertheless, there remains the possibility of the rhythmic MSS being related to the old non-rhythmic ones, in the same manner as a Hebrew text provided with vocal marks and accents is related to the same text lacking those marks and accents.  For centuries, the second text was by memory, read as the first.  Hence, both melody and rhythm may have been sung from the non-rhythmic MSS. (Rhythmic Proportions in Early Medieval Ecclesiastical Chant, 2nd ed., 1960, pp. 3–4)


Lest anyone think I have included the above quote merely as anti-Vatican edition or anti-Solesmes propaganda, here is another in the same vein from Dom Mocquereau:

While it is true that the graphic form of the rhythmic signs varied according to the different schools of writing, as was the case with the melodic signs, yet, in spite of this freedom and variety of form, it is easy to discover a primitive and universal rhythmic tradition which affirms itself with a wealth of evidence and an authority equal to that which reveals the unity of the melodic tradition.  Nevertheless, we must admit that the primitive tradition as regards the figuration of rhythm was maintained with less constancy than the melodic tradition.  Already, the manuscripts of the tenth and eleventh centuries reveal inequalities in the manner of representing, with greater or less fidelity and with forms more or less perfect, the characteristics of the original rhythm. [. . .] The decline in the notation of rhythm becomes more evident and already foreshadows the approaching decadence of Gregorian chant, a decadence which was brought about principally by the neglect of the rhythmic signs, the understanding of which was slowly vanishing. [. . .] Other representatives of the same schools of writing [St. Gall and Metz] failed to cling with equal fidelity to the rhythmic tradition; far from it, indeed, for while they used the same signs, they wrote them in a haphazard way and with little understanding of their meaning.  Yet, even so, these precious fragments of a dying tradition bear witness to the existence and vitality of the tradition itself and bring valuable evidence to the work of restoration. [. . .] There are a great number of manuscripts which have preserved little of the rhythmic indications; others which have kept absolutely nothing.  But it would be an error to cite these manuscripts as bearing witness against the general rhythmic tradition, particularly in view of the positive testimony of the rhythmic manuscripts of the different categories.  In regard to the vital question of rhythm, they are silent, and these non-rhythmic codices can be considered—in relation to the rhythmic codices—exactly as we should consider a text lacking all punctuation and accents in contrast with a text carefully accentuated and punctuated.  Thus we have the primitive Hebrew texts of the Bible, for example, contrasted with the same texts adorned with dots to indicate the vowels or Massoretic accents which define and establish the punctuation, the accentuation, the expression and even the sense itself.  On the one hand we have precision, on the other uncertainty; here perfection, there imperfection; but nowhere is there a contradiction. (“Le nombre musical grégorien,” A Study of Gregorian Musical Rhythm, vol. 1, part 1, 1932, pp. 168–170)

Readers interested in learning more about the distinction between rhythmic and non-rhythmic manuscripts and the decline of the rhythmic tradition may wish to refer to The Rhythmic Tradition in the Manuscripts by Dom Mocquereau and Dom Gajard.


Several of the oldest extant manuscripts have long been considered a model group (to use the words of Fr. Vollaerts) for the restoration of Gregorian chant: for the Mass, the three manuscripts included in the triplex editions, known as C (St. Gall Cantatorium 359), L (Laon Gradual 239), and E (Einsiedeln Gradual 121), along with Chartres 47, Bamberg 6, and St. Gall 339; and H (Hartker Antiphoner, St. Gall 390/391) for the Divine Office.  Although St. Gall 339 is held in slightly lower esteem today than in previous decades, and Chartres shows evidence of stylized notational conventions that are less precise than the other manuscripts of the model group, the reliability of these sources is acknowledged by nearly all chant scholars over the last century.  Fr. Vollaerts remarked that “To plead their value would be akin to forcing open an unlocked door” (Rhythmic Proportions, p. 7).  No serious musicologist would dismiss codex C, L, E, or H’s importance, which is recognized by the foremost expositors of the Solesmes, mensuralist (measured or proportional rhythm), and semiological schools of interpretation.  Along with Chartres, Mont Renaud, and the few surviving fragments from Nonantola, they are unquestionably the oldest extant manuscripts from about A.D. 920 to 1000 and, as records of how chant was sung in the tenth century, have an inherent authority (here this word is used in the sense of the Latin auctoritas) that is not dependent upon the judgment of anyone—neither Pope, musicologist, choirmaster, nor cantor.


In the oldest manuscripts, short and long values are indicated by the shape of the notes themselves and whether they are connected (cursive) or separated (non-cursive).  Some of the length indications are indeed reproduced in later manuscripts with what Dom Cardine called the neumatic break or cut, where the punctum mora (augmentation dot) is typically added in the Solesmes editions, filling in the blank space of the Vatican edition, which Jeffrey Ostrowski has written about in “The Rhythm of the Vatican Edition.”  The Vatican edition and the diastematic manuscripts, however, are utterly irrelevant to the discussion of the original rhythm.  In most cases, it is impossible to notate chant sung in equalist or accentualist rhythm with adiastematic (staffless) neumes in a way that would agree with the oldest manuscripts.  The Solesmes editions reproduce a number the long markings of the ancient manuscripts by means of the horizontal episema, but they omit many others, treating the normal syllabic value as short and indivisible, and often marking only the first note of a group that ought to be entirely long.


The short and long note values are to be sung in 1:2 proportion according to the medieval writers.  Aribo, writing in the 1070s, laments the demise of proportional rhythm, and it should not surprise us to observe rhythmic alterations in the manuscripts of the eleventh century and later.  The same phenomenon is apparent in countless hymns, chorales, and metrical psalms of Protestant origin: an equalization of the note values, which is of course a corruption of the original rhythm, with a note of double value at the end of each phrase, sometimes followed by an additional beat of rest.  St. Anne (“O God, Our Help in Ages Past”), Nun danket alle Gott (“Now Thank We All Our God”), O Haupt voll Blut und Wunden/Passion Chorale (“O Sacred Head”), and Old Hundredth (“All People That on Earth Do Dwell”) are a few of these tunes that are generally known to English-speaking Catholics.  It would be preposterous to claim that the isometric version is the most authentic and that the original rhythmic version merely represents “nuances” of a particular congregation, yet similar if not identical claims are unashamedly made concerning the rhythm of chant!  The absurdity of such a proposition is easily demonstrated by having two or more voices attempt to sing both (or several) versions simultaneously.  Likewise, if we attempt to sing chant simultaneously in proportional, equalist, accentualist, Solesmes, and semiological rhythm, the resultant cacophony will dispel any notion that the rhythmic disparities among the various schools of interpretation are little more than a matter of an extra fermata or ornamental note here or there.


Despite the evidence for the loss of the authentic traditional rhythm, it would be a mistake to imagine that either the rhythm or its notation degenerated simultaneously and uniformly across Europe in the eleventh century.  As Fr. Vollaerts asserts, the age of a manuscript is not in and of itself a guarantee of accuracy.  In modern music, a newer edition may correct errors of previous editions.  It should not be unthinkable that later chant manuscripts sometimes corrected earlier ones, but it would also be unreasonable to assume that such was necessarily the case most of the time.  Bear in mind that Gregorian chant was written down for the use of people who already had the repertory memorized—a repertory that had probably been handed down as an oral tradition.  An alternative theory proposes that neumatic manuscripts from earlier centuries either disintegrated or were lost, discarded, or destroyed.  Who can say when a lost manuscript from the ninth century might be rediscovered?


I have compared the estimated dates for thirteen manuscripts in several scholarly works and editions.  Below are the ranges of probable dates.

C = St. Gall Cantatorium, Codex Sangallensis 359, 922–926; contains only graduals, alleluias, and tracts

L = Laon Gradual, Codex Laudunensis 239, ca. 930; the municipal library site still dates it to the 9th cent.

E = Einsiedeln Gradual, Codex Einsidlensis 121, 960–996

Bam = Bamberg 6, 966-1000

G 339 = St. Gall Gradual & Sacramentary, Codex Sangallensis 339, 980–1000

MR = Mont Renaud Gradual, 10th cent.

Ch = Chartres 47, 10th cent.; destroyed in 1944

Non = Nonantola Antiphonale Missarum, 10th cent.; fragments

H = Hartker Antiphoner, Codex Sangallensis 390/391, 986–1011; for the Divine Office, not the Mass

Ang = Angelica 123, 10th or early 11th cent.

Bv 33 = Benevento 33, 10th or early 11th cent.

Mp = Antiphonary or Tonary of St. Benigne, Montpellier H 159, 11th cent.; digraphic alphabetical notation

Bv 34 = Benevento 34, 11th or 12th cent.; diastematic

The normal and easiest way to determine an approximate date for liturgical books (codices) from the Middle Ages is to check which feast days are included; a text from 950 would not have a Mass for a saint who was not publicly venerated until 955, for example.  (Note that canonizations continued to be handled at the diocesan level in some places until as late as the mid-twelfth century.)


In closing, I challenge any reader to produce a single shred of evidence from the rhythmic manuscript era, roughly 920 through the early 11th century, in support of a non-proportional interpretation—“nuanced,” equalist, or otherwise—of the normal long values, i.e., notes indicated as long by their shape or by separation from the following notes.  In the words of Dom Gregory Murray, “As we have seen, Dom Mocquereau admits that there were mensuralists during the Gregorian centuries; it would be interesting if clear evidence could be cited to show that during the same period there were some who were not mensuralists” (“Gregorian Rhythm in the Gregorian Centuries: The Literary Evidence,” Downside Review, Summer 1957, p. 235).  Study the oldest sources for yourself to determine whether the normal syllabic value is long and divisible or short and indivisible, whether the chant moves primarily in long or short notes, and whether or not it has a steady beat.  You may be surprised at what you discover.