there is the further problem of guessing exactly which notes of the tenor coincide with which notes of the
melismatic voice at moments of change in the slower-moving part. For assured transcription or
performance of this music, then, we would need to hear it sung by its “native” singers, and that is
something we will never hear.
That has not stopped imaginative early-music performers from conjecturing a performance practice
for this music, often very persuasively. As Leo Treitler, a specialist in early notation and polyphony, has
justly observed, “the question for us is not ‘how must they have sung this music?’ but rather ‘how can we
sing it?”’ In seeking an answer to “our” question, Prof. Treitler goes on, “it may be that analysis and
performance can teach us what exact methods have so far withheld about the problems confronting the
musicians of the twelfth century.”^6
EX. 5-6 Transcription of the beginning of Jubilemus, exultemus (Fig. 5-2)But whatever we learn from our own analyses and our own performances must obviously go far
beyond the evidence of the sources into a realm where only artists dare tread, not historians. The
speculative or conjectural performance of early music to delight modern audiences has provided an arena
where artists and historians have been collaborating fruitfully, sometimes within the heads of a new breed
of scholarly or musicologically-minded performer. But such performers know best of all that the historian
cannot always helpfully advise the artist, and that the artist’s successes, though they may convince an
audience that includes the historian, still cannot provide the latter with evidence.
Not all the polyphonic pieces in the St. Martial manuscripts present modern performers with such
difficulties. Along with sustained-note organum settings like the one in Fig. 5-2, the St. Martial sources
contain numerous versus and hymn settings in discant style, in which the two frequently crossing voices, if
not precisely note-againstnote, are at least rhythmically similar. In such settings it is not often possible to
identify a preexisting tune or cantus firmus in either part; thus there is nothing in them to distinguish a vox
principalis from a vox organalis. In such cases the two parts were in all likelihood conceived as a pair.
To say this is not necessarily to imply that the two-part texture was actually conceived as a unit, even
if it was composed in one sitting. One voice might have been written first and then treated as an ersatz
cantus firmus for the second; some theoretical discussions seem to imply as much. But in some settings the
two voices are so intricately (and playfully!) interrelated that simultaneous conception of the whole
texture seems a virtual certainty. One such is given in Ex. 5-7.