This paper was read originally at the Eleventh Annual Meeting of the Society held in New Haven, Connecticut, December 27-29, 1968 as part of a ROUND TABLE discussion concerning "Rehearsal Techniques and Historical Performance Practice." The other participants, whose papers were also included in SYMPOSIUM Volume 9, were DENIS STEVENS, Professor of Music at Columbia University and Artistic Director of the Accademia Monteverdiana; ALFRED MANN, Professor of Music at Rutgers University and Director of the Bethlehem Bach Choir; and FRANKLIN ZIMMERMAN, Professor of Music at the University of Pennsylvania and Conductor of the Concerto Soloists of Philadelphia. Mr. Zimmerman also served as Chairman of the Panel.

From the Roman de Fauvel to Ciconia and Lyonel Power is as long as from Gretry to Vaughan Williams and includes as many fluxes of musical style. So these remarks on a whole century of music must take the form of generalizations derived from our experience in rehearsing a substantial part of it and experiencing this part as it evolves in concert after concert toward the ideal of becoming second nature with our performers. Thirty performances of an anonymous composite mass and a dozen Machaut pieces (more joyfully received, by the by, in Oklahoma City than in New York City) give insights deeper than those gained in a quiet study or a classroom.

One comes quickly to the first generalization: this particular repertory presumed a splendid musicianship for its fourteenth-century players and singers and poses a challenge to the modern performer who decides to return to an era that has struck many observers, such as Stravinsky and Hindemith, as having remarkable parallels to our own musical era. Each participant must have individual strength, with agility and dexterity demanded as much as in any music. The elaborate rhythmic structures which subdivide the beat in two to seven and nine parts and syncopate these subdivisions and the combinations of meters require a quickness of musical spirit. Intonation is paramount in a music with complete chromatic availability that often leads to unexpected and, so to speak, abnormal relationships. Particularly needed is the finesse of ensemble playing where each part is highly articulated and likely to change from follower to leader and back again. These are aspects of complete musicianship demanded by the repertory that set it apart from some other genres of early music. There is little here for the amateur with more good will than skill and very much for the performer willing to take a challenge. While the circumstances of fourteenth-century music point to the highest social class of patronage in Avignon, Cyprus, Florence, Ferrara, etc., it is clear to me that its notes point to an aristocracy of the musical profession as well, an elite of composer and performer, markedly different from the sixteenth century, for instance, when much music was widely distributed and generally accessible to performers of modest ability. There is a popular side, music that may have surfaced from some indigenous subcultures, as in some dances and monophonic songs that do profit from a rough treatment in performance, but their number is small.

One wonders who these performers were, aside from the few who can be named by a letter or a poem—Machaut and his Lolita, Jacopo in contest with Giovanni da Cascia, blind Landini at the organetto. The instrumentalists may not have been legion, but they are called for, I estimate, in well more than a thousand surviving pieces that depend on one to three instrumental parts, sometimes astonishingly complicated. Regrettable, more so, is the fact that their instruments are also anonymous except for a small number of cases where a performance is described in detail, as in Il Solazzo.

The instrumentalists mainly served the solo singer, who was enjoying a kind of pre-bel-canto burst of florid song in the Italian repertory: Lorenzo's Ita se n'era star and Landini's Sia maladetta require as much agility and stamina as the average Italian coloratura aria of the seventeenth century for a prized professional singer with a lifetime of training. And in the French style the long lines of Machaut's ballade Je sui aussi ravi or Matheus de Perusio's Le grant desir demand sustained vocal control that is not generally considered an aspect of "medieval life." It seems obvious that this music could not be shouted or hummed or sung with a folk-singer's unpracticed equipment. It had to be sung, in the modern sense of the word. (On the occasion when this paper was read, a lively discussion followed, led by Prof. Howard Boatwright of Syracuse University, on the high degree of virtuosity achieved by many non-European singers quite unrepresentative of "bel canto." I would consider these equally trained and offering possible alternatives in timbre and vocal production to the student of medieval performance, although the truth of the matter seems irrevocably lost.)

Thus the repertory is largely chamber music dependent on the solo voice, usually with instrumental polyphony in support. A multitude of Italian duets, madrigals and ballatas, and such pieces as the two triple ballades of Machaut, of which each voice part has text underlaid, may be entirely sung or performed with very light instrumental doubling or with an instrument replacing a voice. There is no question in my mind but that this music follows the chamber music ideal of one player or singer to a part, or at most two of very divergent families, for example a wind and a plucked string that do not clutter up each other's path. The idea of amplification by tripling or quadrupling a sonorous line seems quite foreign to the sound ideal, though at this stage in historical studies our notion of that sound ideal comes largely from the inside of the composition with very little exterior evidence to support it. There is just enough evidence from Machaut and Jacopo in words and from the Faenza codex in notes to tell us that a purely instrumental performance was at least the occasional thing and to permit us to add more than a thousand pieces to our instrumental repertory at one stroke.

While this generalization about the chamber aspect of fourteenth-century music—one or a very few to a part—seems fairly certain for the large secular repertory, I think it is also valid for much of the sacred repertory of mass movements and the intermediate repertory of motets with Latin or French texts. The parlato style of the upper voices in motets, glorias, credos, and other mass sections with tropes implies soloists or a diminutive ensemble at least in the main word-bearing sections, and only the most straightforward examples in conductus style seem suitable for a cappella performance. While such scholars as Frank Ll. Harrison and James McKennon have reached present conclusions that would permit only the organ in liturgical situations, I must rejoin "Not proven" on the evidence of compositions that challenge us to a performance, such as the Barcelona and Ivrea pieces of Hanna Harder-Staebelein's Fourteenth-Century Mass Music in France and the mass compositions of Richard Hoppin's edition of the Cypriot repertory. Perhaps there were two or three organists available, each playing a single line of the polyphony, or perhaps one organist with two left hands?

The isorhythmic motets pose particularly thorny problems of performance with their chattering upper parts and plodding tenors, and they have eluded our search for an exciting, compelling way of distribution and interpretation, although we have successfully performed a brilliant isorhythmic "Ite Missa Est" from Ivrea with two sopranos, two tenors, and two bass singers, recorders, vielle, bells, organ, and regal. One looks in vain for clues to the performance of these terribly complex pieces and must wait till 1436 for Manetti's description of the dedication of the cathedral in Florence and its performing forces with lutes, flutes, and trumpets when Dufay's isorhythmic Nuper rosarum flores was first heard.

Since the great majority of New York Pro Musica's concerts are done by the permanent ensemble of eleven or twelve performers, equally divided between singers and instrumentalists, we are attempting to exemplify the chamber music ideal of my main generalization. In our current repertory are some five hours of late medieval music, extending from anonymous works of the late thirteenth century to Lyonel Power and Dufay and including three composite masses drawn from the anonymous repertory, quite lengthy and of unusual musical interest, fully the equal, I believe, of the Machaut mass. They have a splendid breadth in contrast to the miniatures of the secular repertory.

Even within the limits of this diminutive ensemble, it is extremely difficult to find a full ensemble piece of the secular kind, say to end a concert with a Kehraus finale or to make a reasonably sumptuous moment of full participation. We have countered the problem with some of the popular-styled monophonic virelais of Machaut, treating them informally and returning to the popular dances from which they seem to stem. The refrains are for our small unison chorus and the verses are for soloists taking turns. The instrumentalists supply drones and heterophonic improvisations. There is also a small number of two-voice ballatas of the Italian repertory, of necessity very simply composed to allow so much unison participation, that also lend themselves to alternation between soloists and a very small chorus with instrumental byplay. But in the whole polyphonic work of Machaut I find only one secular piece, the triple ballade De triste cuer, that could enjoy a "small chorus"—our six singers with instrumental doublings. While it is polytextual in its three-voice polyphony, the voices unite most deliciously on the refrain text of the ballade, which they all of course have in common.

As centuries go in musical libraries, the fourteenth is fairly well fitted out with monumental editions. Two editions each of Machaut, Landini, and Jacopo are riches, for example, compared to one of Josquin with its single set of misprints. The need for performers to approach a text well armed, however eminent its scholarly genesis, is paramount, and it is often wise to retrace the paths of proofreader, engraver, scholar, and ancient copyist for oneself. In a music as remote as this it seems a terrible thing to perform a misprint with great relish and wonder.

Even Friedrich Ludwig and Leo Schrade are often in unanimity with errors in their editions of Machaut's works, and it helps the performer of early music to cultivate strong musicological and philosophical friendships in his cause. A case in point is our experience with Machaut's triple ballade Sanz cuer m'en voi which both respected scholars published as a 3-voice canon with entrances at the distance of one measure, full of delicious and praiseworthy dissonances and partwriting highly prophetic of things to come six centuries later, much enjoyed by us in rehearsal. The triple-ballad text is in fact three consecutive stanzas of a ballade that tells a tale, in Stanza 1 the lover beseeching his lady, in 2 she consoling him, in 3 he telling her he is consoled. A philological friend, Prof. Lewis Gordon of Brown University, first pointed out to me as we were working on the translation that the lover's reply begins in the Ludwig-Schrade version before the lady has a chance to tell him what she thinks. Two inquiries of friendly musicologists led us to Gilbert Reaney's published solution to this canon with the voices entering at the distance of two measures with their stanzas in the right order. An unusually beautiful composition emerges but, alas, without the neo-Hindemithian reckless counterpoint of the Ludwig-Schrade versions. Intending no dishonor to these great scholars, one learns the lesson that there is still much room for informed, creative insight and fresher views about this repertory, still far from completely reclaimed.

Certain insights will arrive simply through the exigencies of performance committed to a group of living, lively musicians. On first glance through the collected works of Machaut or Landini, one might conclude the performers should be all male, all tenors ranging up and down a sixth from middle C. This is the narrow range mentioned in elementary history texts as characteristic of the fourteenth-century formes fixes. Were there no sopranos, no high instruments? Was Machaut's Peronne a tenor when she asked him to send her his compositions? Did no one speak or sing in a deep bass voice then, and were bagpipes, psalteries, and rebecs all tenors? In consequence of these questions, we not only transpose entire compositions up and down at will but also have decided to treat the octave location of particular voices very freely. A case in point is the forementioned triple ballad De triste cuer, whose all-tenor score is impacted in mid-range and has been so recorded by a trio of intertwined tenors in Brussels. We have in our performances lifted the cantus an octave, preserved the location of the remaining voices, and entrusted a viol player to provide a simple, improvised, plucked part to assure that the integral lowest notes are preserved, thus avoiding inversions that might be produced by our relocation. In this version nearly all of the editor's accidentals are ignored and the composer's accidentals are allowed to have indefinite value, to the point where the reprise in the second section has the same sound as its appearance in the first, even though accidentals were not repeated. It is our practice commonly to ignore all editorial accidentals for this repertory for the first few rehearsals and slowly to restore them as aural testing requires. The performers who live with this music exclusively will refuse to play some, and others will turn out to be "bad notes" on our reconstructed instruments.

There is comfort in the fact that no pure, invariable text need be sought for these challenging pieces, and the critical variants give leeway in matters of accidentals, real and feigned, the underlay of the literary text, the ornamentation, the distribution and even the number of parts. A case in point is Machaut's rondeau Rose, liz, printemps. It comes in two manuscripts as a four-voice composition: text underlaid to one voice with three instrumental voices, one higher than the singing voice. In a third source there is a second contratenor part, so labelled (making it a five-voice composition?). The instrumental cantus fails to appear, however, although there is a blank staff with room for it. What is intended? A performance for four with two contratenors, or are the two contratenors mutually exclusive? Does the presence of the second contratenor preclude the performance of the missing cantus by some fourteenth-century musician who remembers it from another source? Or is it the only five-voice secular piece of Machaut? The temptation to perform it all is irresistible, and the result is not nearly so problematical as De toute fleurs or the Ludwig-Schrade solution of the above-mentioned three-voice canon. A complete performance of this rondeau runs between seven and eight minutes. Pressed by the realities of the modern concert situation, we sometimes reluctantly truncate this fixed form to an ABaAB and a reasonable four minutes and 35 seconds.

Surely here is the principal difference between our experience in rehearsing fourteenth-century music and the activity flourishing in many academic environments. The scholar may ponder a thorny problem of performance practice and leave his indecision buried in a footnote, and a collegium musicum may postpone indefinitely its annual concert and go into consultation. We must proceed by a shorter path, going almost immediately from the printed musicological sum of information—the score, the editorial emendations, the critical variants—to the sounding public performance. The very pressure of this commitment leads to practical solutions which, I hazard, are not so remote from the solutions that sounded six centuries earlier from a remarkable body of performers who shared with us the same commitment.