composition that is in duple time on all levels of mensuration. The purposely varied detail-work
discloses the song’s literate origins: the first three lines of the poem are set to what are in essence three
repetitions of a single musical phrase, but each of them is subtly distinguished from the others. In the
original notation the music begins with rests, even though there are no accompanying parts, because
without bars the only way in which an initial upbeat could be indicated was by showing the silent part of
the hypothetical first measure. Like many duple-metered pieces of the time, it especially emphasizes
syncopes.
EX. 9-4 Guillaume de Machaut, Douce dame jolie (monophonic virelai)Eight—only eight—of Machaut’s thirty-nine virelais are polyphonic. Of these, six are in two voices
only, a texted “cantus” (“song” or “singer”) part and an untexted tenor. The nomenclature already suggests
that a tenor has been added to a “song,” or in other words, that the song existed as a monophonic
composition before it was made polyphonic by a lower accompanying voice. This is just the opposite
from the procedure we have observed in all the polyphonic genres of the twelfth and thirteenth centuries
—organum, clausula, even the homorhythmic conductus, which had no preexisting tenor but created its
own from scratch. Above all, adding a tenor to a preexisting cantus was the very antithesis of motet
composition. Starting “at the top” was a whole new concept of composing—within the literate tradition,
anyway (for we have always recognized the possibility, indeed the strong probability, that “accompanied
song” was a minstrel specialty at least from the time of the troubadours.) Two kinds of additional
evidence clinch the notion of “top-down” composition. One is the state of the musical sources. The virelai