4. Practical sources: changes of style about 1100.
To the extent that there were now alternatives between which the performers could freely choose, extempore polyphonic performance became that much more difficult. A way of relieving the difficulty was to fix some of the alternatives in written-down form. The earliest known practical sources of polyphony do in fact date from around the time of Guido (see Sources, ms, §§IV, 2 and VI; also RISM, B/IV/1, M. Gushee, V(3)1963, Rankin, VIII(1)1993, and Arlt, VIII(1)1993). In what is now called the Winchester Troper (early 11th century) numerous different versions of certain turns of phrase in the individual vox organalis parts are recorded expressly in the margin: this is some indication of the degree of freedom that had been gained meanwhile in polyphonic treatment, and also of the interest that each alternative aroused.
The early written sources are unfortunately difficult to decipher. They evidently assume the singer to be so well versed in the basic rules of polyphonic performance that a rendering in mostly staffless neumes would suffice. (Scribes did occasionally later in the Middle Ages resort to letter notation again; when they did so it was precisely in order to counter difficulties of reading that might arise.) Despite the uncertainties of deciphering these notations, it is possible, even in the earliest of practical sources, to determine certain characteristics that go beyond the teaching of Guido; they help to put the compulsoriness of the traditional rules into perspective.
Ex.7, from Alleluia, Angelus Domini in the Winchester Troper (f.164v; Holschneider, p.110; the vox organalis, in a different part of the manuscript from the vox principalis, is given in fig.1), is traditional in the parallel movement at the 4th below in its second phrase, and typically Guidonian in the close of its first phrase (at the asterisk). Similarly, the switch of lower pitch limit from g to f (first phrase) had already been authorized by Guido. By means of this switch the chant, which descends to f, does not have to cross the vox organalis; also, an occursus on to the final g is only possible via f (cf Guido’s example, CSM, iv, p.213). Among the new features that occur in this piece are the formation by the vox organalis of a 5th above the vox principalis at the beginning of the first and third phrases (assuming the transcription to be reliable). Evidently the tessitura of the vox organalis is regulated by the prevailing final; even in a chant of wider range the drone effect – which was still obviously much liked – is partially retained (according to Musica enchiriadis the vox organalis ought, by analogy with the melodic movement of the chant, to begin in unison on c: cf ex.1). As a result, the notion of a lower pitch limit in the strict sense scarcely applies any longer, so habitually is it exceeded (see for example the e at the end of the third phrase). Moreover, according to the movement of the chant, other notes, besides c, f or g, appear as organal holding-notes (as one might call them to distinguish them from the lower pitch limit, and also from the long-sustained notes of organum from the 12th century onwards): see for example the d in the second phrase of ex.7.
The principal requirements for the formulation of a ‘new organum’ about 1100 were by now fulfilled. Parallel movement and holding-notes were now so loosely applied that they could readily be replaced by a completely free use of intervals, including a free interchange between 4th and 5th. The 4th, which was for Guido the widest distance of ‘singing apart’, had evidently already been exceeded, and the two parts crossed as often as it seemed melodically or harmonically sensible for them to do so. In the light of Guido’s teaching on occursus the principle of contrary motion gradually emerged and took on significance; and in his extending of cadences there lay already the beginnings of an impulse to ornament the penultimate note, which after about 1100 became melismatic in character.
The new style of organum is evident as early as the latter part of the 11th century in one of the three Chartres fragments (F-CHRm 109). This fragment contains five two-voice pieces which can be accurately transcribed because they are notated on staff lines (fig.2 and ex.8, upper line). The principle of holding-notes is here completely abandoned. Even simple repetition of a note is avoided in the vox organalis, with the result that there is very little difference of melodic character between the two voices. With the exception of several parallel 3rds, which always converge onto a unison and function like a prolonged occursus (phrases 1–2, 5 and 7), contrary motion is prevalent, with the voices extending to a 6th apart (phrases 3 and 4) and occasionally as far as an octave apart (phrases 4 and 8). The vox organalis still tends, as in traditional practice, to lie below the vox principalis; but the two do nonetheless cross, as is natural when contrary motion is in force. In general the two voices seem to centre their movement on the final d. Caesuras (taken here as the points at which the two voices converge to unison, and in later sources as marked also by vertical strokes) occur not merely at each genuine distinctio in the chant but in practice at the end of each word of text. The price of emancipation from parallel movement and from drone effects is first and foremost sectionalization into small phrase units – the breaking-up of the chant into short harmonic progressions. The fact that almost all these progressions end on the final d means that the piece is, from the tonal point of view, remarkably homogeneous.
It is historically interesting to note that a second version of this piece has survived. It appears in a manuscript from the latter part of the 12th century, now in Oxford (GB-Ob Rawl.c.892, f.67v; see ex.8, lower line, and fig.3). The piece as it survives in Chartres 109 is evidently already a distinctive enough product to be worthy of preservation. For it is hard to imagine how, considering the great range of possibilities that had meanwhile evolved in polyphonic treatment, the two versions could so closely correspond simply by the application of analogous rules.
At the same time there are differences of detail that indicate that the first version was not thought to be absolutely definitive. Complete definitiveness is not found before the compositions of the Notre Dame repertory in the late 12th century: a repertory within which fixity had become a goal towards which composers might rightfully strive. In the first and sixth phrases there appear interchangeably an archaic initial 4th and a modern unison. Where Chartres lets the counterpoint expand to the octave Oxford on one occasion presents only a 5th, clearly preferring conjunct melodic movement (phrase 4, at the asterisk). Also, the author of the Oxford version is more concerned with contrary motion where Chartres has parallel 3rds. As the interchangeability of 4th and unison at the beginnings of phrases shows, the two versions were probably not far apart in time, despite certain differences. Rather, they are as the imprinting of two divergent stylistic tendencies upon the common basis of an established polyphonic solution.