Blue Heron’s artistic director, Scott Metcalfe, has recently published an article in the Journal of the Alamire Foundation (Volume 14, Issue 2 / 2022) exploring the question of performing pitch for vocal music in the 15th and 16th centuries, a widely misunderstood topic that he has studied for years, testing his ideas in practice in Blue Heron concerts over two decades. The article is freely available on the website of Brepols Publishers as a Gold Open Access article under a Creative Commons CC 4.0: BY-NC license: https://www.brepolsonline.net/
Clefs, Pitch, and Transposition in Vocal Music before 1600
by Scott Metcalfe
Read the Article
Blue Heron is committed to performance that embodies the best of the early music movement, to learning as much as possible about historical practices and putting this knowledge at the service of vivid and imaginative performance in the present day. I cannot overemphasize the latter point. We are not singing this music for primarily historical reasons or to recreate anything, but because we love it now, because it opens our souls and moves us. The early music movement exists first because those of us in it love early music and second because we find that learning what we can about the way our musical ancestors experienced music helps us in our effort to make their music come to expressive life in our 21st-century performances.
There is general agreement among scholars of music before 1600 that parts with texts were normally sung, and that polyphony whose sources include texted parts was also sometimes played by purely instrumental ensembles. Beyond this there is not much about the performance of Renaissance repertoire that does not pose a question to be pondered by the present-day musician. Of particular interest to a vocal ensemble involved in the performance of music before 1600 are questions of ensemble size, voice type, performance pitch and transposition, instrumental participation, and regional pronunciations of Latin.
The notes below should be regarded as a work in progress, sketching out some of the issues and outlining some possible approaches. At the end are some thoughts about what sort of sound Blue Heron is aiming to create.
A choir in the fifteenth century could be as small as three or four men, or as large as a dozen or more. The top part of sacred polyphony was usually sung by adult male falsettists, but occasionally boys might replace them. Polyphony might be performed by soloists or by larger ensembles with more than one voice to a part; the latter possibility is a requirement for those works in which one line occasionally divides into two or more, as occurs, for example, in Du Fay’s Sanctus “Papale” (recorded on our CD) and many of his isorhythmic motets. There is no doubt that I am aware of that secular music (chansons, madrigals, and so on) was normally performed with just one singer on a part. The distinction between sacred and secular performance modes relates directly to the venue the music was heard in. Sacred music was heard predominantly in the public sphere of liturgy, often in very large spaces and on occasions calling for glorious effect: such venues and occasions would benefit from larger numbers of performers, including multiple performers on a given part. Secular music, on the other hand, was designed for private chambers and non-ecclesiastical entertainments; it emphasized individual virtuosity, both technical and expressive. Church music called for a voice that could carry in ecclesiastical spaces, whereas chamber music required a softer, more intimate and more personal voice. When sacred music was heard outside of the liturgy, as it often was, in smaller private chapels, or for private devotion, the style of performance and number of performers was probably the same as for secular music.1
The question of the relationship between our modern-day voice types and those of five centuries ago is a vexed and contested one, and involves an enormous breadth of factors such as: physical size, nutrition, and age of puberty; musical training; linguistic background; techniques of vocal production and aesthetics of sound. For the moment I will limit myself to saying that I find that American singers seem most suited to English music before 1600 by linguistic background (and possibly better suited than present-day English singers, whose accent and methods of vocal production are the products of much later historical developments in pronunciation and “choral” training), and beyond that, that the ranges within which our tenors, basses, and other broadly generalized voice types can sing comfortably, in a supple and text-oriented manner, seem to be more or less the same now as they were in the Renaissance. (A few more thoughts on this subject in the next section.) I do not wish to minimize, however, the difference, possibly quite large, between our trained voices and theirs. There are certain types of voices that we don’t encounter all that often in present-day North America, perhaps because of the way we bring up singers: the first example that comes to mind is the Franco-Flemish contratenor altus, which seems to have been a man we would now call a tenor with a rather large range and a particularly easy high end, who was comfortable singing for extended periods above middle C, up to a’ and beyond, but could also muster up a strong low c.2
The study of historical pitch standards, which has borne ample fruit for the seventeenth century and later,3 has been less successfully pursued before 1600, due mostly to the dearth of surviving instruments whose pitch might be measured. In the case of a cappella vocal music, it is often claimed, more or less off-handedly and on scant or no historical evidence, that in the absence of an instrument to establish a pitch, vocal ensembles simply chose a pitch out of the air, and that the result was a complete lack of vocal pitch standard across Europe. In its extreme formulation, this argument maintains that “A” could have been sung at any frequency whatsoever. Besides the lack of evidence in support of this view, there are serious objections to its plausibility.
The first is that, in broad overview, the ranges expected from adult male singers in written-down music between 1400 and 1600 fall mostly into three general categories: cantus, extending from approximately c’ (middle C) to e”; tenor (including, for these most general purposes, the specialities of tenor altus or simply altus and contratenor), c or d to g’ or a’; and bassus, F or G to about c’. These correspond pretty well to the ranges of our modern categories of male falsettist, tenor, and bass or bass-baritone, and, allowing for some notable exceptions such as the high ranges sometimes employed by English boys and the occasional very low bass part, are remarkably stable across two centuries of music from all over the continent; they also correspond nicely and logically to ranges that can be written with a minimum of ledger lines in the soprano, tenor, and bass clefs (c-clef on the first line of the staff, c-clef on the fourth line, and F clef on the fourth line, respectively: c1, c4, and F4, for short). Now, if performers (a category which includes, without exception, every composer of music in this period) had been entirely random in their choice of performing pitch, we would expect instead a corresponding randomness in written ranges. That this is not the case suggests, again in broad overview, that singers in the fifteenth and sixteenth centuries fell out into voice types generally similar to those of singers today, with generally similar written ranges, and that the written ranges correspond to fairly stable pitch levels.
What these pitch levels may have been brings us to the next objection to the idea that the choice of pitch for a cappella music was completely haphazard. If the latter scenario were true for a cappella music before 1600 or so, we would expect to see a change in ranges in music that specifically requires voices to perform with instruments, as composers adapted from unfixed pitch to fixed pitch, but instead what we find is that the general stability of range continues unaltered into the seventeenth century with the advent of concerted music and general adoption of basso continuo. The earliest pitch standards Bruce Haynes has been able to establish are from the later sixteenth century, and whether in Italy, Germany, France, or England, they fall more or less within a semitone or two above and below a’=440 for music that combined voices and instruments: much more variety than exists in the mainstream musical world today, it’s true, but hardly random scatter.4 These data, combined with the continuity of written vocal ranges on the two sides of the basso continuo/concerted music divide, imply that pitch in what might be called the a cappella era did not vary wildly and haphazardly from place to place and time to time, but rather was more or less stable, with a’ lying somewhere within a tone or so of 440 hz. As for setting the pitch, what professional vocal ensemble nowadays lacks even one member with perfect pitch who may be called upon to act as a reference? In an era when professional singers working for the church made their living singing music largely without instrumental accompaniment, such skills may likely have been more common than today.
This is a topic worthy of its own book, of course.
Thus far I have been discussing pitch. This should not be confused with transposition. Establishing a pitch standard means deciding what absolute frequency we will call A. If we set a’ at 415, this does not mean transposing from a’ = 440 to g#’; it means that the pitch of 415 hz is what we call a’. (Pitch is a human choice, not an eternal verity, despite the modern tendency to regard a’ as somehow naturally and permanently fixed at 440.) Transposition, on the other hand, means looking at an A and singing or playing, for example, an E. Pitch and transposition are independent variables, although, since both lead eventually to the production of sound at certain frequencies, they are easily confused.
Nowhere is this confusion more apparent than with regard to the subject of chiavette, or high clefs. This refers to a system of clefs (normally with a treble clef or g2 on the top part, and a baritone clef, F3 or c5, on the lowest). Articles by Andrew Parrott, Patrizio Barbieri, and Andrew Johnstone, drawing extensively on primary sources, have established beyond doubt that in the sixteenth and seventeenth centuries such high clef systems implied a transposition downwards in performance of a fourth or a fifth relative to the prevailing local pitch standard; high clef systems were employed for vocal music by Franco-Flemish composers active in Italy from the mid-fifteenth century, and the principle of transposition is specified in print as early as 1543.5 As Barbieri and Johnstone show, high clefs came into use so that composers could avoid using ledger lines when writing music in certain modes. The transposition necessary to bring music written in high clefs back down into normal vocal range was effected by the performers at sight. The practice of transposing vocal music written in high clefs makes sense, of course, only if a pitch standard of some kind exists; otherwise the apparent pitch of the notes would be completely irrelevant.
With all the above in mind, Blue Heron chooses its pitch standard for any given work in this way: First, we assign voices to the various parts in such a way that best corresponds to what we know of the makeup of historical ensembles; usually this means male falsettists and/or female mezzos on cantus parts in c1 or soprano clef, high tenors on (tenor) altus parts in c3 or alto clef, lower tenors on tenor and contratenor parts in c4 or tenor clef, and basses on parts in F4 or bass clef, with female sopranos used in place of higher boys on treble-clef parts. (High-clef pieces are dealt with similarly, after transposition downwards.) Next we consider what little we know about regional variation in pitch—basically extrapolating backwards, with appropriate caution, from trends observable from the late sixteenth century onwards. Finally we try the piece out and see whether it would be more comfortable for our particular group of singers to adjust it a bit up or down. Using this method, we invariably end up within a semitone or two of A440.
Fifteenth-century sacred music was normally sung a cappella, and this remained the norm throughout the sixteenth century as well, but some sort of instrumental participation, most likely by the organ or some sort of brass instrument, seems to have been possible, especially on occasions of more pomp and splendor.6 By the later fifteenth century instrumental participation in vocal performances of sacred music seems to have been becoming more acceptable or desirable, or at least more frequent. With careful evaluation of primary sources we can begin to develop a nuanced view of the possibilities and one that takes heed of geographical and temporal differences, for practices varied from place to place and from occasion to occasion.7 In the case of Burgundian and German choirs there is, for example, a good deal of evidence for the use of instruments with voices in liturgical settings. Between about 1480 and 1520 such instrumental participation usually took the form of a single player within the vocal ensemble. For example, at a Mass in Toledo in 1502, “The singers of the King [Ferdinand] sang one part of the Mass, the singers of monseigneur [Philip the Fair] the other part; master Augustin played the cornet with the singers of monseigneur, which was good to hear with the singers” (from the chronicle of Antoine de Lalaing). This Augustin is Augustein Schubinger, a virtuoso from an Augsburg family of wind players, who spent a large part of his career in service to the Habsburgs. His performances with the singers of Philip the Fair and his father Maximilian I are documented at masses in the Netherlands, France, and Spain, as well as in German lands. We don’t know exactly what these single wind players may have played in a vocal ensemble, though, so our choices must be informed guesses. (For more on Maximilian’s choir, see my notes to our concert of June 15, 2005, at the Boston Early Music Festival.)
The participation of instruments with voices in liturgical polyphony took hold in Spain more vigorously than elsewhere on the continent; a bajón, used to double bass lines, was to be found in the closet of virtually every church in Spain from the early decades of the century onwards. The practice flourished on Mexican soil, where polyphony in the 1530s seems to have been accompanied by trumpets, shawms, flutes, drums, and bells, at least on occasions like Corpus Christi processions. Again, though, exactly how players participated in vocal polyphony is difficult to establish precisely.
In England, on the other hand, strictly vocal performance seems to have prevailed in the sixteenth century, at least until the later decades; this seems to have been generally true in Italian lands as well.
Until recently, the pronunciation of Latin by clerics (and thus church singers) was heavily influenced by the sound of the vernacular tongue, so that French Latin sounded distinct from English, Italian, or Spanish Latin. Blue Heron has experimented since its origins with regional, historical pronunciation of Latin, informed especially by Harold Copeman’s book Singing in Latin (Oxford, 1990) and Singing Early Music, edited by Timothy McGee (Bloomington and Indianapolis, 1996). It should be stressed that such an attempt is highly experimental, for definitive reconstruction of the sound of fifteenth-century Latin is, of course, impossible. Our hope has been that the unexpected sounds of “vernacular” Latin, as opposed to the bland, vaguely Italianate sounds of modern “Church Latin,” lend the music a particular, local flavor, draw attention to the texts, and make Latin sound more like a real language and less like a succession of attractive but not especially meaningful vowels. In latter years I have become somewhat skeptical that American singers are ever likely to produce a really convincing French Latin, any more than a French choir will ever sound really English, and this feeling has affected our choice of pronunciation for Franco-Flemish repertoire. English Latin, on the other hand, is truly a native tongue for Americans, for the sounds called for are just those American speakers normally use, and I think we can deliver a reasonably effective Spanish Latin as well.
A more subtle effect of experimenting with vernacular Latin pronunciation has been to reveal the profound influence of language on a composer’s melodic sense. The influence of the vernacular on the Latin compositions of French-speaking composers, for example, extends to a tendency to set Latin words with the rise in pitch on the last syllable characteristic of French speech; noticing this phenomenon has led me to consider how French melodies in general might operate, leading and lifting easily towards the ends of phrases in a way that English melodies do not, and this in turn has inspired a different sort of musical shaping in performance. I’m not sure that I would have come to this realization without spending several years working awkwardly away at French Latin. This sort of shift in musical perception is, of course, exactly the sort of mind-opening result we hope for in early music, and the reason we study historical performance practice at all.
What sound is Blue Heron aiming for?
Audience members often comment that they hear something different about the sound of Blue Heron, but they may not be sure exactly what it is. Some may wish to know about the “how”: what preparation we do in order to create a distinctive sound. What follows are a few thoughts, which were originally written down in answer to a listener’s question about our sound. They are not meant to offer an exhaustive explanation, but begin to address these questions.
First of all, we want to be sure that an attentive listener can understand every word. In order to achieve that, we work very hard on correct pronunciation (specific vowels and idiomatically pronounced consonants), correct and idiomatic word stress, and idiomatic stress, accent, and direction across larger verbal units (phrases and sentences). This applies to historical and regional Latin pronunciation as much as to the vulgar tongues. Besides having the virtue of adding verbal meaning to musical meaning, a highly varied and specific pronunciation with the widest possible palette of color and articulation is interesting and attractive on its own. Variety is an essential tool for compelling the listeners’ interest, in rhetorical music-making just as in rhetorical speech. (Classical rhetoricians stressed the importance of varietas and the idea was taken up by medieval and Renaissance music theorists.)
The emphasis on language and rhetoric informs how we understand and make decisions about the stresses and directions of musical phrases as well as verbal phrases; in vocal music, the two are inextricably linked.
Blend vs. complexity
I am not interested in “blend” per se. I am not aware of any evidence that “blend” was a priority for musicians in the past. I think that what people mean by it nowadays is that singers should modulate or moderate the individual qualities of their voices in order to create a homogeneous, smooth, or de-individualised sound. But there is no a priori reason to value homogeneity or smoothness over individuality and complexity; in fact, I think there are many reasons to prefer the latter. What I want to hear is a complex sound made up of all the fascinatingly individual sounds of particular singers who are singing at their best and using all the colors at their disposal, while at the same listening to and responding to the contributions of their colleagues. So I encourage everyone to use all the resources of their own individual voices.
One result of prioritising individual color over “blend” is that voice parts and sections sound different from one another, because they are sung by different people, and one of the wonderful things about voices is that they are all different from each other. Far from being a problem, this is a great virtue in polyphonic music, where each line is equally interesting on its own and no line is subordinate to any other, because the lines will be more easily distinguished from each other. It’s only a problem if you value “blend” above all—but “blend” is simply a quality, like volume, not an absolute good.
Same goes for balance. The sound of a polyphonic ensemble should be complex and multilayered. No one voice should predominate, necessarily, at any time.
The priority on heterogeneity and complexity extends to how musical lines are shaped. In polyphonic music, one should allow each line to follow its own logic. This may mean that one line is strong when another is weak, and so forth. But that’s how this music works: it is polyphony, not homophony.
Questions of aesthetics
As you might guess, I’m not interested in “purity” or “detachment” or anything supposedly “angelically” disembodied, either. I don’t see any evidence in art or theory that those things were valued by musicians or any other artists of the past, either. Do angels in 15th-century art look disembodied? Do Dante or Petrarch or Michelangelo shy away from the intense expression of visceral emotion? I think that the notion that medieval or Renaissance music is meant to sound “pure” or “ethereal” is connected to a simplistic view of our musical and human past, a romantic or sentimental idea that the past was somehow simpler or more homogenous, or less passionate, or less expressive, or less advanced—as if human emotion was invented by the nineteenth century. It’s a ridiculous notion on the face of it, and has been thoroughly debunked in all the other arts, not to mention in any historical discipline, but to a certain extent we have all absorbed this sort of progressivist prejudice when it comes to music. (I suspect that the marketing of early music in the 1970s and 1980s is also responsible, along with, in the specific case of Renaissance polyphony, the domination of the CD market by English choral ensembles, which on the whole seem to prioritize cool competence, homogeneity, and a smooth surface.)
We make a deliberate effort to score music appropriately, endeavoring to have our ensemble reflect the way ensembles were constituted when the music was conceived. This does not necessarily mean an exact imitation of historical ensembles (even if we imagined that we could accomplish such a thing). In particular, since we are free from the old ecclesiastical prohibition against men and women singing together in church, we employ women on the upper lines of sacred music instead of boys. Indeed, musicians as eminent as William Byrd are known to have done so, with no apparent discomfort, when circumstances forced (or enabled) them to, and of course women and men have always sung secular music together.
I don’t think that our sound is “better” than anyone else’s, but it does deliberately reflect our priorities of complexity, specificity, and clarity. And I do think that these priorities are those of the music we perform.
© 2017 (revised version)
For more information, see:
our archive of notes, including the notes to our CD.
1 For the early fifteenth century see David Fallows, “Specific information on the ensembles for composed polyphony, 1400-1474,” in Stanley Boorman, ed., Studies in the performance of late medieval music (Cambridge, 1983), pp. 109-59. The article provides exactly what the title promises. There is no single study I am aware of that examines the question of the size of ensembles for vocal polyphony in Europe across the entire period we are concerned with. The interested reader might begin with information provided in Reynolds, Christopher A. “Sacred polyphony,” in Performance practice: music before 1600(New York and London, 1989), pp. 185-200, and the research cited therein; work by Fallows, Craig Wright, Richard Sherr, and Keith Polk may prove especially helpful. Additional data on the size of specific choirs may be found in the thoughtful and detailed introductions to the Renaissance Church Music series of Antico Edition, especially, for England in the early sixteenth century, Nick Sandon’s editions of the Peterhouse partbook repertory (RCM volumes 101-140). Whether the entire staff of a given choir sang any given piece of polyphony on any given occasion is another question entirely, and one for which we have precious few answers.
2 Rebecca Stewart has done some stimulating work on the subject of the voice types required for Josquin’s music. See, for a start, her early article “Voice types in Josquin’s music,”Tijdschrift van de Vereniging voor Nederlandse Muziekgeschiedenis25 (1985), pp. 97-189.
3 See Bruce Haynes’ comprehensive treatment of the subject, A history of performing pitch (Lanham, MD, and Oxford, 2002).
4 Haynes 2002, ch. 2.
5 Andrew Parrott, “Transposition in Monteverdi’s vespers of 1610: an ‘aberration’ defended,” Early music xii (1984), pp. 490-516; Patrizio Barbieri, “Chiavette and modal transposition in Italian practice (c. 1500-1837),” Recercare iii (1991), pp. 5-79; Andrew Johnstone, “‘High’ clefs in composition and performance,” Early music xxxiv (2006), pp. 29-53. The 1543 source is Silvestro Ganassi, Lettione seconda (Venice, 1543): see Barbieri, pp. 39-41.
6 Fallows lists a number of examples in “Specific information” on p. 127.
7 The Cathedral of Cambrai, for instance, permitted only all-vocal performance in the period 1475-1550, according to Craig Wright, “Performance practices at the Cathedral of Cambrai 1475-1550,” Musical quarterly lxiv (1978), pp. 295-328; on the other hand, English musicians at the Council of Constance in 1416 sang Vespers “with organs and prosunen [slide-trumpets] above which were tenor, discant and medius” and the Savoy court chapel in the years 1450-55 paid, alongside the singers and the organist, a single player of the tromba (see Fallows, “Specific information” on p. 127 for both these instances).