Versions Ancient and Modern
Gregorian chant for the Propers of the Roman Mass have been used in the twentieth century according to a standard edition of the Graduale issued by the Vatican in 1908; this is known as the Vatican Graduale. Designed as a practical edition for liturgical use, it contained no critical report accounting for its melodic versions, even though these had been prepared on the basis of knowledge and good judgment from the materials produced by research at Solesmes.
The melodic versions in the Vatican Graduale were used consistently in the subsequent Solesmes publications of the Graduale and the Liber usualis. Neither did these use any critical report; if any changes in musical text were made during the twentieth century on the basis of the continuing research at Solesmes, these could not be determined except by personal experience in using the books. (I am not aware of changes, at least through the Graduale triplex of 1979.)
None the less, awareness of the versions in medieval chant books, and comparison of these with the Vatican Graduale (and the Solesmes Graduale) progressed steadily through the century. Many interesting and important aspects of the relationship among these sources have come to light. I list some of these here, then go on to discuss them and assess their impact on our approach to the problem of versions.
Against the background of accusations that the Vatican Graduale was severely, perhaps arbitrarily, limited in its choice of source materials, or that it was expedient and presumptive in its determination of readings, the Vatican Graduale has turned out to be not such a bad report on medieval chant books: its selection of readings is a good representation of medieval sources, and the readings are good melodically. It is a perfectly acceptable source of Gregorian melody, provided only that we read it not as a critical edition, but simply as received custom, another chant book, derived from preceding chant books and from traditional practice in much the same way that medieval chant books were derived. As a physical book between covers, it is more comprehensive and easier to see and use than the manuscript facsimiles. That, however, does not diminish the value of the facsimiles, which are the life's blood of continuing chant scholarship. And the scholarship has uncovered relatively few new medieval sources that might substantially alter the melodic repertory as a whole.
Furthermore, the scholarship continues to confirm the overall stability of this melodic repertory. What is new is the acknowledgement of the peculiar features of such melodic variants as appear, and their impact on our use of the Vatican versions.
The variant readings are very numerous. At the same time, they are in nature slight, to a degree that they were called by one expert, speaking in a precise and technical sense, “trivial.” Unlike the kind of variant usually encountered in literature, they seem almost all to be acceptable, not mistakes. These qualities seem paradoxical relative to one another, and understanding how they can all be true at once shows the way to a new approach to versions.
Summarizing a long, complex discussion, Gregorian variants do not lend themselves to the construction of a stemma of filiation, which is the most sophisticated kind of analysis performed by textual criticism. This is because the usual type of Gregorian variant cannot be identified as a scribal mistake. This makes even simple collation not indicative of which manuscript source is closer to the archetype. The usual procedures of textual criticism, regarded as scientific or at least objective, do not in this case permit conclusions that a given text is good, in the sense of representing the composer's intent. So these usual procedures do not get us to the original. Beyond that, however, we are less and less sure that there was in fact an original, or, if there was, whether we are interested in finding it.
We are thrown back on the first, and best, meaning of “textual criticism,” namely that it is criticism, judgment, concerning the text in hand. When we are interested in a particular text that we find in a historical document, we exercise our personal judgment concerning its value. We may say that it is a good text; we mean—even if we do not say—that it is good for our purposes. It is the purposes that are often not clearly apparent, for scholars have tended to presume that their own purposes were obvious and that everyone would share in those purposes. For example, much Gregorian research in the twentieth century has presumed that the purpose was to determine how they sang chant at Rome circa ad 700. This was not obvious, however, around ad 1900, and was not shared by at least some of the Vatican authorities; they had to be persuaded, apparently by highly placed friends of scholars such as Dom Mocquereau and Dr. Wagner. It required a papal pronouncement (presumably prepared by the Sacred Congregation of Rites), to encourage merely the propriety of consulting medieval chant books as part of the preparation of new editions. And it was a near thing. In other words, for some liturgists, documentation of how they sang chant in ad 700 was irrelevant to a new edition.
It is easy to imagine other purposes of chant research: for instance, we might want to determine how chant that had the current reputation of being “Gregorian” might be sung. We might want to determine how chant was sung at Solesmes in ad 1935. Some current scholars want to determine how various singers sang chant circa ad 1200; they have access to many chant books from that time that give exact indication of pitch, making it relatively easy to determine alternate ways of how the melodies go—even with the problems of the ubiquitous variants. Of course, the problem of deciding which way is good still requires personal judgment.
My own interest, guided by historical intuition, is in the way the scribes of L and C sang Gregorian chant. The fact that they included marks of interpretation in their chant books makes it possible to gain access to qualities of musical performance other than just the pitches. Actually, their notation gives less access to the exact intervals between pitches, and in that respect the use of L and C requires careful collation with later diastemmatic manuscripts. In the process of collation, the melodic versions of the later manuscripts can be modified to conform better to the pitch notation of L and C (in ways that will be exemplified presently). This indirect and sometimes intricate process cannot be done, however, in such a way as to satisfy requirements of scientific research, for its hypotheses cannot be confirmed. I find, none the less, that the procedure produces results that are good for my purposes. The results are embodied in the versions I sing.