developments that took place before there were any means of documenting them, an answer to this
question is nevertheless suggested by recent research into the practices of more recent, in some cases still
active, oral traditions of church music. Nicholas Temperley, investigating the history of what has
sometimes been called “the Old Way of Singing” in English parish churches of the seventeenth century and
New England Congregational churches of the eighteenth, and the “surge songs” of black churches in the
American south, noted a pattern.^7 Musically unlettered or semilettered congregations that sing without
professional direction over long periods of time tend to develop a characteristic style: “the tempo
becomes extremely slow, the sense of rhythm is weakened; extraneous pitches appear, sometimes
coinciding with those of the hymn tune, sometimes inserted between them.” Wesley Berg, a Canadian
scholar working with Mennonite communities in Western Canada, has corroborated the process by direct
observation.^8
What both scholars describe is the transformation, over time, of simple syllabic melodies into ornate,
melismatic ones. (And the point about rhythmic weakening jibes tellingly with the notorious nonmetrical
rhythm of the chant, about which little is known and about which, therefore, many strong opinions are
maintained.) In New England, the process was thought to be one of corruption. Professional singing
masters, armed with notated hymnbooks, sought to counteract the tendency by training their congregations
to be not only literate but literal-minded in their attitude toward written texts. In a wholly oral age, when
alternative methods of transmission were not available, the process of transformation was more likely
seen as desirable, since it produced an ever more artistic, “skilled” product. In the context of the evolving
Christian liturgy, degrees of melismatic elaboration served as a means of differentiating types of chants as
well as liturgical occasions on the basis of their relative “solemnity.” As we will see in the next chapter,
moreover, there is evidence that the Gregorian chant itself continued to develop melismatic
embellishments in parts of Europe where a relatively fluid oral culture seems to have continued, perhaps
for centuries, after the Franks had begun relying on notation as a fixative.
It used to be thought that the large amount of shared material within chant families reflected a
“patchwork” process of composition, called centonization (after the Latin cento, “quilt”). Peter Wagner,
one of the pioneering historians of early Christian music, compared centonized chants to articles of
jewelry in which prized gems have been selected to receive “a splendid mounting, an ingenious
combination, and a tasteful arrangement.”^9 Today, scholars prefer a different analogy or model: instead of
a fund of individual memorized formulas from which chants are assembled on the basis of artistic
ingenuity and taste, one imagines a process of elaboration from a repertory of simple prototypes for
various liturgical genres and classes.
The shared formulas found in the Graduals we have been comparing, for example, are found only in
Graduals. Another type of chant that is comparably formulaic in its melodic content is the Tract, a long,
sometimes highly melismatic psalm setting that is sung in place of the Alleluia during penitential seasons
such as Advent and Lent, when the joyous ejaculation alleluia—Hebrew for “Praise God!”—is
suppressed. Tracts come in two mutually exclusive formula-families, and their characteristic turns are not
found in any other chant genre.^10
A fund of shared melodic turns characterizing the chants of a given functional type, or those proper to
a certain category of ritual observance, is exactly how the term mode is defined in its earliest usages. The
concept of mode as formula-family is still prevalent in the Greek Orthodox (Byzantine) church, where the
liturgical singing follows what is known as the oktoechos, an eight-week cycle of formulaic “modes”
(echoi in Greek).
Our more recent concept of mode, based on that of a scale, and defined mainly in terms of its final
note, fits the Gregorian repertory poorly. (We have already seen, in fact, that Gregorian psalm tones often
have a variety of potential final notes, the differentiae—see Ex. 1-8.) The concept of mode as a function