\\ Researching and developing performance practice in late medieval chant [ENG]

One extremely sobering and in some ways stimulating thought for a researcher and performer of late medieval chant is the fact that no contemporary treatise, nor any study ever since, nor any recording ever made or concert sung provides the definitive answer to the question of how to perform chant from late medieval sources. This article is to be read as an interim report of the search for some more – and more detailed – answers to this basic question. However, as I will show, my quest is actually not just about reconstructing the performance practice of the plainsong of a bygone era, however detailed and painstakingly profound this research may be. It has more to do with the construction of a performance practice of fifteenth-century plainsong, based on genuine practice-as-research. In this article, apart from all the facts and observations related to notes and manuscripts, singing and singers, historicism and artistry, the final focus is on the kind of research we do, by doing research based on the doing.

Following an introduction to different artistic and scholarly aspects of the issue at stake (I), my aim is to clarify matters on a more disciplinary and practical level (II), and to illustrate how this kind of research requires interdisciplinary work that, to quote Alastair Williams , ‘not only translates across disciplines, but blurs their boundaries.’ As examples of this, I will focus on three very specific aspects in the performance of chant, placing them in a late medieval context: rhythm, memory and voice. Rhythm (III) is certainly a particularly thorny and highly controversial problem in the performance of chant from any period. The ability to memorize (IV) has been one of the major requisites for a good singer of chant. And the voice (V) is the most essential research tool of all. It is the voice that investigates; it is the voice that develops the sensuous knowledge we need to sing and perform. Returning to a more theoretical level of thinking, I will finally discuss some of my concerns regarding the institutional place of artistic research, before concluding with a look ahead to the challenges facing the research into and development of a performance practice of late medieval chant (VI).


It is not my intention to juggle with terminology, but the reader will notice that I use the terms chant and plainsong interchangeably. That is indeed what I consider them to be. However, I will not here discuss questions of terminology and definitions as John Caldwell has tackled the issue with considerable accuracy in the introduction to his essay on the interaction between plainsong and polyphony during the later Middle Ages.

Speaking of polyphony, this current research project actually arose from an artistic collaboration between my chant group Psallentes and the polyphonic ensemble Capilla Flamenca. The artistic director of this ensemble, Dirk Snellings, knew of my interest in late medieval chant, and encouraged me to found a professional ensemble specializing in the performance of plainsong from the age of polyphony. Both of us are convinced that polyphony benefits enormously from a contextual performance. By this we mean presenting polyphony (in this case mainly from the fifteenth and sixteenth centuries) in its natural environment. In monasteries and abbeys certainly but in collegiate churches too, plainsong was the rule, polyphony the exception. So, as polyphony in this period of history was very much based on, related to and connected with plainsong, it is all the more natural for modern performers to balance polyphony with chant, or vice versa. This is not to say that in performances today this strategy is unknown of or even particularly innovative. A quick discographical analysis shows us quite a respectable number of recordings, produced in recent decades and focusing on polyphony, that involve some performance of plainsong as well. Still, the majority of recordings featuring polyphony do not choose contextual plainsong performance. It is hard to say whether this is simply because of a lack of interest in chant on the part of the ensembles concerned, the difficulty of finding good chant sources compatible with the polyphony, or the uncertainty about or reservations (justified or not) about the performance practice of chant.

For a singer interested in chant performance practice at a specific period of time, several questions and problems arise. These include questions concerning language and vocal techniques, such as the proper pronunciation of Latin, voice use and pitch; performance practice issues such as tempo, rhythm and phrasing; contextual considerations such as the number and composition of the ensemble, the place and time of performance; and repertoire matters, such as the transmission of the old repertoire and the making of a new, regional differences within the repertoire, the use of simple polyphony and the interaction of chant and polyphony. It is a frighteningly complex field of investigation. Much work has been done already, although the vast majority of it concerns the oldest repertoire (from the ninth, tenth and eleventh centuries), found in the oldest manuscripts. This reflects the initial objective of chant scholars to restore plainsong to its ‘original state’, after long centuries of so-called mutilation. Until just a couple of decades ago, relatively few scholars were attracted to the plainsong of later periods. Moreover, those that were usually took a special interest in it primarily because of related polyphony.

In 1992, Thomas Forrest Kelly edited a book on plainsong in the age of polyphony. It was a landmark, and a clear sign that researchers finally had grown more interested in late medieval plainsong. Yet, apart from the contributions of Richard Sherr and John Caldwell, the essays in Kelly’s book do not represent research into the most concrete of performance practice questions such as the tempo of the singing, or what rhythm to sing in. For this, we need to turn to a much older landmark. In 1969, a doctoral degree was granted to Sister Thomas More (Mary Berry) at the University of Cambridge. Her unpublished doctoral dissertation is not very well known today. That is unfortunate, because her research has proved to be very valuable to performers, her main concern throughout being problems of rhythm. The chief materials from which Sister Thomas drew are primary sources, mainly manuscript and early printed service-books, as well as the writings of theorists. The conclusions she reached renew our understanding of later plainsong: ‘The picture that has emerged is complex, and that in itself is important: there were more ways than one of performing chant.’ I will return to this point presently, when I consider some ways of researching rhythmical problems.


So there were (and are) more ways than one of performing chant, as there are more ways than one of investigating late medieval chant performance practice. Scholarly research, to begin with, has brought to light all kinds of interesting facts and figures about the practices of chant singers in the later Middle Ages and Renaissance. Let me give a small and simple example. The critical edition of a fifteenth-century ordinal from Tongeren, Belgium, makes it clear that on Christmas Day the intonation of the responsory Verbum caro factum est is to be sung by the cantor. The first verse In principio should be sung by three canons facing the east, and after the repetendum the Gloria Patri should be sung by three vicars facing the west. This is valuable and fascinating information which a good scholar will take seriously, even at a basic level of investigation. He will check the rest of the ordinal for similar instructions, compare this instruction with information in other ordinals, and so on. The scholar will then pass on his findings to the research community, where an interested performer will pick up the information and use it in a concert situation. A good performer will then try to build on this idea, experimenting with locations, trying out positions, working with different combinations of singers, using other music in the same setting – always keeping his eyes and ears open to new and engaging perspectives.

It is no coincidence that in the above example, the three stages run parallel with the three perspectives in the almost classic trichotomy described by Henk Borgdorff. In his terms, the first situation could be seen as the interpretative perspective, which turns into the instrumental perspective when the performer enters the picture. Finally, things are looked at from the performative perspective. It is tempting to bundle these three perspectives into one artistic perspective or, at least, to bundle them into one research project or one research group.

In my approach to this issue I am following two paths. On the one hand, I simply want to gain substantial theoretical and practical knowledge about the performance practice of plainsong, and how this practice has or has not found its way into the manuscripts. On the other, my concern is to become more aware of the way in which chant in general, and particularly the chant of the fifteenth century can be approached by today’s voices and how it can find its way to the hearts and minds of today’s public. Since we are dealing with practice-as-research here, this means that my status as a performer and artist is a major factor in the research process. As already noted, there certainly are quite a few traditional musicological aspects in what I do, but the most typical and innovative part of it is that I ask the questions starting from an artistic viewpoint, and that I use my artistic practice as a research tool. All these boundary-blurring activities come together in the main objective, which is to determine the way in which square notation as used in fifteenth-century chant manuscripts provides a clue to performance practice. I want to see and experience how the manuscripts themselves can suggest answers to our performance-related questions, how certain features of these manuscripts can lead us singers to new sounds and perspectives, how our present-day training in chant or in the performance of chant can alter our understanding of the different historical sources. I will now give three examples to illustrate this.


The debates surrounding ‘research in the arts’ may be ferocious, but the debates surrounding chant rhythm are probably even more passionate. They certainly have a longer history. One of the major polemics, at least in recent musicological history, has been the Vollaerts-Solesmes controversy. The Dutch Jesuit Jan Vollaerts (amongst others), calling on evidence he found in texts by Carolingian theorists, claimed that plainsong is to be sung in strict time, using two values, in the ratio of two to one. The followers of the Solesmes school on the other hand, maintain that it all comes down to rhythmic nuances which are not susceptible to strict measurement. Most chant scholars studying the oldest repertoire seem quite convinced that Vollaerts is wrong. Moreover, in recordings of plainsong, the Solesmes school is audibly the winner. My referring to the ‘losers’, the camp of the so-called mensuralists like Vollaerts, is of importance for what is to be presented hereafter.

Central to my research project are late fifteenth and sixteenth century chant manuscripts from Ghent. Many of these are linked to the Abbey of Saint Bavo, where at the end of the fifteenth century the abbot Raphael de Mercatel had initiated a new liturgical dynamic. Especially noteworthy are a gradual in two volumes, in square notation (thick horizontal and hair-thin vertical strokes), written by the sub-prior Adrian Malins, with illustrations by the Master of Privileges of Ghent (Ms 14); an antiphonary in two volumes, probably written by Malins as well, in square notation (Ms 15); a psalter in which the notes to the notated hymns were later obliterated (Ms 73), and a notated processional in Messine notation (Ms 184). One of the most remarkable features of the antiphonary is that Malins employed a script that has some features in common with the mensural notation known from polyphonic sources. Thin lines were added to the large black notes – to the left of the note when in an ascending movement, to the right of the note in the descending case. It is difficult to say whether this is just the elegant mannerism of a copyist in the habit of writing polyphonic music, or whether it really is meant to be a rhythmical notation.

As a singer and leader of an ensemble, I chose the latter option as a working hypothesis (the manuscript being written in mensural notation), at least with the intention of extensively investigating this possibility. In different projects, theory was converted into practice. In each case the practice was part of the research, the manuscripts and the singing always raising questions and offering answers. Thanks to the upward-pointing lines on the left side of the note, this plainsong became a game of three lengths of notes: longa, brevis and semibrevis. For example: a normal podatus (two notes, the first low and the second higher) would be performed as brevis/longa, as would a clivis (two notes, the first high and the second lower). It was remarkable that, during the experimentation and rehearsal that we always fell back on a kind of tempus imperfectum (duple time). Moreover, we had an almost irresistible inclination to manipulate the supposedly intended rhythmical value of the ligatures in order to maintain the tactus (beat) of the imperfectum. In other words: an interplay between long, short and shorter notes was possible and even exciting, but difficult to maintain without some adjustment.

There is no evidence that the chant in this antiphonary was intended to be sung in a mensural way, but neither is there evidence to the contrary. Trying to find ways of answering questions not answered by hard evidence is, to quote Daniel Leech-Wilkinson, ‘endlessly fascinating, a battle of wits between the lack of evidence and one’s own ingenuity.’ Yet, it is not just about ingenuity: in the end, it is about persuasiveness. By rehearsing this chant in as it were a rhythmical notation, experimenting with it and performing it, my ensemble arrived at a logical and consistent artistic concept (in my view), one that could persuade and excite performers and listeners alike. This performance can lay claim to some validity: whether it has any historical validity will probably remain uncertain, but its artistic and musical validity is absolutely clear to us. What I am emphasising here is that our performance practice should not (only) be judged or measured by its demonstrable historical validity, but also by its demonstrable artistic validity and persuasiveness.

In singing this late fifteenth-century chant in a mensural way, the performance makes reference to a number of historical precedents. John Caldwell, in his already cited essay, has given an illuminating résumé of the attempts to put plainsong into a rhythmic straitjacket. From the writings of Carolingian and post-Carolingian theorists (the “two to one ratio”), through the famous treatise of Jerome of Moravia (whose interpretations of the rhythmical value of neumes and ligatures come remarkably close to the rhythms of mensural polyphony), to the anonymous Quatuor principalia musice (which states that all cantus planus should be performed in the fifth or the sixth mode, meaning in perfect longs or in breves), all seem to suggest that the mensuralist view like Vollaerts’ may after all be quite legitimate. All of these historical precedents, to be studied and experienced (with my ensemble) in practice and in detail, will nurture and improve this practice-based research project in the next few years.

There is one historical instruction that I should mention, because up till now it has been a great inspiration in our performance of late medieval plainsong. It concerns the second rule from Conrad von Zabern’s De modo bene cantandi (1474). Like so many medieval writers, he states that in plainsong all notes should be equal, but he goes on to say that the rhythmical performance of plainsong is a widespread abuse (‘una de communissimis abusionibus’). He complains that all too often singers lengthen the highest note, shortening the following note. This duality has been a starting point for genuine practice-as-research. The effort to seek equality between the notes while at the same time subtly going up to the highest ones (usually the word accent) and then slightly shortening the subsequent (less important) note has had a wonderfully expressive effect on our performance of plainsong.


The actual performance practice of plainsong in the late Middle Ages and Renaissance is beyond recall. As I indicated briefly at the beginning of this article, we may start from a desire to reconstruct the past – and this may have a highly stimulating effect on our whole enterprise – but we realise that we cannot but construct our performance practice. We want to be the construction workers of a chant performance practice for today’s voices and today’s public, using all the historical information we can get to lay out blueprints similar to those our ancestors used. The appearance of our “building” may turn out to be surprising, bewildering, or staggering (or plainly pointless or boring for that matter) but we know for certain that the foundations we are laying have stood the test of time. One such foundation is the singer’s memory.

It is a familiar image, known from illuminations of medieval manuscripts: a few singers have gathered around a book. At times one of the singers holds the book, at others the book is placed on a lectern. There are always more singers than books. The book is hardly more than a memoria technica, a mnemonic device where chant pieces are notated to remind the singer of a melody he already knows. The singers read over each other’s shoulders, they touch each other, they can hear each other’s breathing, they sing into each other’s voice and ears. Sooner or later this singing together results in their agreeing on how to sing this way or that. These agreements are sometimes made more explicit, but usually they are made without words. Nothing is written down – the valuable book remains untouched. All the details are to be filed in the singer’s memory. Even the slightest rhythmical subtlety is underpinned by the singer’s realization that the text has to be intelligible. The Magnificat antiphon at vespers is sung by heart, or at least the Magnificat itself is. The soloist who is pushed forward to deliver the lectio, has in his head a collection of melodic formulas, well chosen within the appropriate mode, and all intended to enhance the intelligible delivery of the text. Another singer will have developed an almost physical memory of the ornamentation possible in the solo verses of a gradual or a responsory.

These singers share an ability to communicate with each other, they can sing with one another as a unit. They trust each other’s voices as they trust each other’s memories, in a co-operative effort where everyone’s input is essential. This is a unique kind of ensemble singing. In being together and in singing together, they develop the conventions of their practices. These, as said before, might be beyond recall, but the circumstances can certainly be revisited. In this way, today’s singers can develop new conventions, which can eventually tell us more (and certainly quite different things) about the historical conventions than any academic treatise or study could. Is this not a fascinating research question: how can the practising and performing of plainsong, in circumstances similar to those of the past (such as singing from the lectern and relying heavily on the singer’s memory) lead us to new ways of singing plainsong, and at the same time aid the research into the lost conventions of late medieval plainsong performance?


The principal tool in researching and developing a performance practice of late medieval plainsong is the voice. As suggested earlier, it all comes down to the singing itself. The scrutiny of the evidence is done with today’s voices; it is with these voices that research results resound.

Much has been said about the use of the singing voice in the Middle Ages. A reference to the Instituta partum, an early thirteenth-century Cistercian source will here suffice. According to the anonymous writer, the psalms in plainsong should be sung ‘with an even voice, at a steady tempo that is not excessively drawn out, but at a moderate pitch, not too quickly, but with a full, virile, lively and precise voice’. Is there any other way to try to carry out these instructions than by researching them through practice? How can one sing in a virile manner, and yet hold back on the richness of the voice in the interests of the group’s overall blend? The singing of polyphony can accommodate quite a lot of personal character and richness in the voice, but in group singing of plainsong, the singer will have to be aware of his/her own sound and take great care to blend in continuously with the other voices.

Another voice-related matter is the pronunciation of the Latin. It is a very complicated matter which needs special attention in rehearsal and performance. It seems logical that performers should use a pronunciation in accordance with the provenance and period of the manuscripts in question. But what Latin should this be? Erasmus lamented the absence of an international pronunciation, and poked fun at contemporary ways of speaking Latin. He also described how the French pronounced Latin, with striking features such as the vernacular ‘u’. In our performances, we have mainly been using the so-called Franco-Flemish Latin, a mixture of different pronunciations closely resembling the French accent, but without nasalization. This sometimes has had a startling effect on listeners who are accustomed to singing or listening to chant in the more Italianate Roman pronunciation. Singing Latin with this Franco-Flemish pronunciation has helped us to streamline our vocal-technical efforts. For example, the use of the ‘u’ ([y], as in the French volume), has repercussions on the consonants surrounding it, making them smaller and lighter. Consequently, the use of Franco-Flemish Latin considerably changes the enunciation and prosody of our singing, with serious consequences for the overall performance. Working with the Franco-Flemish Latin is helping us to develop a smoothly elegant, more fluid style of singing late medieval plainsong. Starting from a historically ‘more correct’ position, the artistic concept evolved by way of a research method led by the practice of rehearsal and performance. Up till now, this research would seem to indicate that an historically oriented pronunciation of Latin enhances the artistic validity of my ensemble’s performance practice.


Perhaps in trying to illustrate why the practice of late medieval plainsong is also the researching of it, a picture may have emerged of a research project that is vulnerable and has difficulty in explaining what its methods are. I am not too worried about this – after all, one must be able to deal with uncertainty (in the sense of a lack of absolute certainty). This is a relatively new style of research, based on the experience and conduct of the artist-practitioner, and ‘researchers must have the courage to come to terms with the diffuseness and uncertainty’ of it. Consequently, it is my sincere hope that artistic practice-as-research be given time to settle within the more traditional institutional research environment, while yet maintaining its unique characteristics.

In this preliminary report, I have tried to focus on examples that show how in my practice-as-research project, tacit knowledge can be revealed and articulated. To my mind, the key to success is interdisciplinary work, related always to more than one branch of knowledge or to various methods of research. Some challenging problems still have to be tackled, in my own work and in the field of practice-as-research in general. But to quote Henk Borgdorff once again, ‘it is a challenge we can meet.’

One challenge is to find the right balance between words and music. Discursive reporting of data is one thing, but it is quite another matter to try to explain with words what should be illustrated and demonstrated with the singing voice. In footnotes four and ten respectively, I refer to the Psallentes website and discography. This is not intended as a kind of clandestine advertising, but rather as an attempt to offer an opportunity to complement the reading with listening. As with the adage about pictures, one minute spent listening is worth more than a thousand words. It is simply essential that what the writer attempts to describe in words should have a musical pendant. After all, whatever we may say or write about music, does not music have to have the last word?

Hendrik Vanden Abeele
This text was previously published in the Dutch Journal of Music Theory, February 2007.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

Blog at WordPress.com.

Up ↑

%d bloggers like this: