By Luigi Lera, teacher and conductor
We have been asking ourselves for more than twenty years why our choral practice, despite being lively and most dedicated, has not been able to produce some sort of “resurrection” of early polyphony. There are of course many different reasons behind it. In this article, we will limit ourselves to exploring just one aspect: the extent to which the transcriptions that we are called upon to work with daily are to blame.
Why transcribe early music? Essentially, there is only one valid reason: so as to have the complete composition in front of you. Renaissance polyphony was always written and performed directly in individual parts, whether using printed sources or liturgical manuscripts. Thanks to transcription, before even beginning a piece, we know the chords used by the composer, how the parts have been arranged and how imitations have been structured. But beware, these requirements do not affect the standard chorister; they relate entirely to the work of the conductor. They may be of interest to any one of us who wishes to better understand the beauty of this music coming from a distant past, but it is not crucial for the simple purposes of keeping to time or perhaps finishing together.
Singers do not need to see the full score, any more than those performing a quartet, just as section violinists do not need to have the whole symphony in front of them. Many of us have most certainly sung the chorus of Nabucco or La Traviata reading individual parts without any problem at all. The only one who really needs a transcription is the conductor; and to be really precise, he only needs it when studying the piece. In fact, there is a whole series of examples of modern interpretational abuse, particularly with regard to dynamics and changes in tempo, which have been brought about directly by conductors having all of the parts in front of them. It may well be that the much-desired resurrection of authentic Renaissance practice indeed needs to be brought about by courageous efforts to put together a madrigal or a motet working exclusively from individual parts.
Beyond this first consideration, all other reasons in support of transcribing Renaissance music are simply not valid.
It may be thought that transcription is needed in order to obtain a clear and ordered graphical representation of the score: in some respects, this justification is plausible, even though sixteenth-century writing is among the most accurate and legible of all those from past centuries. In times gone by, transcription was used to effect a transposition, that is to make the tessitura suitable for the female vocal range. We shall refrain from commenting on this task in itself, because it sounds somewhat reasonable given the current state of choral practice. However, a healthy need has developed among the musicians to have the modal arrangement of the pieces, as it was put together by the composer, in front of them. Our knowledge of the composition techniques and chord sequences of the Renaissance is already poor enough, without needing to confuse it further with any anachronistic transposition. Modern music editors no longer consider it necessary to work on the score for the simple purpose of adjusting the pitch at the time of performing, just as no director would dream of asking them to do so.
It is said that transcription serves the purpose of eliminating the problem of music written in ancient clefs, and this is very true. On the other hand, however, there are those who uphold that only notation using C clefs allows polyphony to be read correctly. Both stances have their merits, even though at present it seems to be the more hard-line ideas having to make the greater concessions. As regards note reading, the modern clefs of G and (for tenors) G an octave down transmit, to all intents and purposes, the ancient clefs correctly. From this point of view, the C clefs were nothing but an unnecessary burden even in the time of Palestrina. It is, however, also true that the ancient clefs are essential for other purposes of no lesser importance: first and foremost for mode analysis according to the H. S. Powers method, a technique which directors and music scholars should master to perfection, and secondly for those who wish to read notes according to the medieval solmisation system. For the first of these purposes, the brief sample bar that is usually placed at the start of the score is more than enough; for the second, on the contrary, there are no real alternatives. For a variety of reasons, the complex mechanism of Renaissance mutation is not applicable to the modern G clef. He who claims so is unnecessarily deluding himself: in the best case scenario, he will not even succeed in grasping the complexity of the problems with which he is faced. Personally, I do not believe that a hexachordal reading involves the magical conquests it would seem to; but I am also convinced that there are other more hidden reasons to encourage its practice. It is always risky speculating about the future, but it is likely that when we succeed, by other means, in pinning down the concrete rules of Renaissance composition, our grandchildren will rediscover all the benefits of referring to the sounds in the same way as the polyphonic composers once did.
If the choice of clefs really does not impact on the performance, many other graphic details do so very dramatically. Over a century of modern choral practice applied to the polyphonic repertory has made us quite certain of one important concept: the additional marks added by a reviser – legato, staccato, accents, dynamics and changes in tempo – do not help in any way to clarify the writer’s idea. On the contrary, they unfailingly lead to quite the opposite effect: they impose on the original performance practice, like a layer of transparent varnish, a whole series of modes of expression that belong to a more modern context. These additions end up distancing the performance of the piece from the original result. The notation of a polyphonic piece must contain only what was written, or what could have been written, by the composer; the sun has well and truly set on the days of Malipiero’s editions. Modern singers have the right to insist that the conductor, or the figure in charge of performing the piece, make an effort to transcribe the piece from scratch rather than subject them to a score that is spoiled by deliberate pseudo-interpretative markings.
The prerequisites for a good transcription, nonetheless, go well beyond the simple recommendation of not adding anything extra. What is then asked of those who draw up the score for polyphony? Above all, that all the notes are in place – but even satisfying this basic need is not as easy as it may seem. Renaissance writing concedes large margins of discretion even when reading the sounds. First, there is the issue of implied accidentals, those that should be placed above the note in a well-made edition; we know that some particular mechanisms of polyphonic technique, those related to the tritone or those typical of the cadence, imposed accidentals for some sounds without its being necessary to have this explicitly indicated in the parts. The reviser, from this point onwards, must have sound knowledge of all the harmonic conventions in use in the composer’s time. It may be said that, in this field, there are no absolute certainties, especially considering that even the Renaissance sources do not seem to be in complete agreement in this regard; nevertheless, the concreteness of the diametrically opposing viewpoint should be stressed. The general criteria for correctly positioning accidentals not only exist, but also leave very little room for personal interpretation. The days when it was thought that you could give modal character to a piece simply by flattening or lowering the leading-note that comes from a suspension are also over. The rules for placing implied accidentals are clearly defined and are also not difficult to understand – they just need to be taught correctly.
Another aspect the modern editor is called upon to address in making Renaissance writing truly legible is the placing of the text. Current printing techniques allow for a level of precision totally absent from sixteenth-century sources, when the text was merely positioned at the start of the musical phrase without any separation or spacing; indeed, a great deal of sacred music – both in manuscript and printed form – did no more than setting a title, leaving all the noted staves blank. The modern chorister, on the other hand, needs to be assisted syllable by syllable in the delicate task of matching the text with the notes. Sixteenth-century polyphony is too far from our musical culture to be able to rely, as was the case at the time, on a series of unspoken conventions. In this matter, we are still very far from having matured an informed approach: the conductor very seldom questions the positioning of the text as found in the score, just as it is true to say that rarely do his studies leave room for specific training on this matter.
Any one of us can easily verify how serious are the consequences of this state of affairs. If, for example, it is true that Kyrie eleison can range from four to seven syllables or that in gloria Dei Patris can range from six to eight, just think of how many times the melisma of the penultimate syllable has been sung in the wrong place, perhaps compromising the clarity and comprehensibility of an entire episode. The fundamental aspect for placing the text in Renaissance music is that the syllabication adapts naturally to the polyphonic phrasing: long syllables on long notes, short syllables on short notes, melismas on accented syllables, correct pronunciation of diphthongs and so on. Unfortunately, today, the only guarantee of having a good score is still the personal, or rather professional, skill of the transcriber.
None of these initial points actually carry sufficient weight to be accused of having strangled the rebirth of Renaissance polyphonic tradition single-handedly. Scores, keys, accidentals and the placing of texts are all important issues, but they make up no more than a sort of long introduction to our subject. In the next paragraphs we will explore important issues such as tempo, beats and values – we cannot stress enough how crucial these aspects are. It is one thing to place a syllable incorrectly or to add a sharp or a flat without good reason, but it is something quite different to completely misunderstand the tempo or the dynamics of an entire piece. It is said that transcribing polyphony is needed to “make understandable to the modern reader those aspects of notation from the past that today would prove to be too obscure for those who do not have a specific background in palaeography”. How valid is this reasoning?
Our amateur choristers often do not shy away even from the St Gallen neumes of Gregorian chants or the obscurity of some twentieth-century writing. Is Renaissance notation really so different from ours as to justify transliteration into another system? Well, no, or at least not as regards its most basic mechanisms; most probably more than ninety per cent of the musical masterpieces from the sixteenth century are perfectly accessible to those who can read modern writing. There are of course differences, but the more substantial ones are not by any means difficult to take in because they all derive from one main source. Let us take a look at what that is.
The history of western music records, quite unambiguously, something happening exclusively to notation: the constant shifting of values in time. It can be said that this is something that simply happens and that’s that; there is no real reason for it, even though there are various different arguments which attempt to justify it in various historical contexts. It is a fundamental issue, so much so that any discussion about transcription should begin precisely from here.
It simply consists in this: over time the reference values used in the notation are gradually reduced. If one century thinks in terms longe and breves, the next thinks in terms of breves and semibreves; what comes next will be semibreves and minims and later still minims and crotchets. Pay close attention: all these changes concern only how the music is written, while it has no effect at all on the tempo. It could be said that every generation’s younger players are keen to cram their forebears’ music with smaller and smaller figures; with the passing of the decades these same composers start to age and the way they think of music becomes more austere. Getting old, they themselves are transformed into the defenders of the system in the face of the excesses of their grandchildren. The outcome of these two opposing and conflicting tendencies is precisely a slow shift towards smaller figures.
We could compare the phenomenon of figure shifting to the fairly constant increase in inflation, a similar but opposite mechanism that occurs regularly in the field of economics: it would not make sense to talk about a luxury car that costs 500 dollars, unless we point out that we are referring to the year 1925. The case of music is even easier, because no composer would normally use more than four or five different values within the same piece: the fact that Machaut uses breves in the same way that Palestrina uses semibreves, Monteverdi minims and Schumann crotchets should not cause problems to those who read their scores.
Nineteenth-century teaching knew nothing of this phenomenon; one cannot even blame the syllabus since it included only a rudimentary notion of all music prior to Classicism. Our school system has ended up rotating around only three binary tempos: two-two, two-four and two-eight. This is because these are the only binary measures that a musician of the late nineteenth century could realistically meet within the classical and Romantic repertoire. Everything else, it was said, is not really needed by the musician; and, if it were necessary, it could be easily transcribed. This naive and simplistic solution, used for more than a century, has made our generation stupid and lazy: indeed, using any base system other than the one they taught us at school should not cause any problems. Finding three crotchets or three minims or three quavers within a bar should not be any different from finding three longe or three breves or three semibreves: I have experienced this over many years, but anyone can test it for themselves by back-transcribing any solfeggio exercise. I never change the figures when I put polyphony into a musical score, even if I am dealing with a thirteenth-century motet: I believe that a true rebirth of Renaissance music cannot occur until we change our narrow minds. We have to convince ourselves that changing the figures of a polyphonic composition distances the original performance in exactly the same way as the added expressive dynamics and agogics: it is an operation that hides the composer’s idea, as under a coat of clear varnish, under an inevitably different rhythmic concept.
We must face at this point an important objection: if indeed the speed of the piece does not depend on the figures in use, it should be ever more evident that changing the figures is as irrelevant as changing the clefs or even changing the key. This argument has some semblance of stability, but only a semblance: for those who really read the figures, and each of us possesses a period or a genre or a style in which we feel particularly at ease, each system has a very precise connotation and does not sound at all the same if it is changed into different figures. The fact is, however, that reducing the figures is not only not a solution, but it is above all a conceptually wrong solution. We reduce the figures to make the performer play faster: the truth is that halving the figures is a solution that operates in a diametrically opposite direction compared to the logic of mensural notation itself. Every composer usually has a choice, within their rhythmic and expressive resources, between more than one tempo: this happens precisely because of the slowness with which figures change. For all musicians there are metric indications that belong to the classical and traditional spheres, more normal tempos and measures that let the imagination run wild. History shows that, keeping the same setting of the rhythm, composers chose the larger figures when they wanted to induce the players to go faster and smaller figures when they wanted them to go slower. In other words, they acted according to a logic that is exactly the opposite of that which leads to halving.
Nineteenth-century teaching made us blind as regards this fundamental truth: reduced figures inevitably lead to fatally choosing less smooth tempos. What would motorists say if someone told them to lower the gear on the motorway to go faster? Basically, the reasoning goes, when travelling in second gear the engine is revved up so much that it seems like a Ferrari!
Why would Beethoven write the Scherzo of the Ninth Symphony in three-four time if it was really equivalent to twelve-eight (and nine-eight) time? Why would Schumann have used the same solution in the Finale of his Piano Concerto, a piece with which many conductors still make fools of themselves today?
Why does the Viennese Waltz insist on being in “one in a bar three-four time” rather than in a more comfortable six-eight time? Simply because smaller figures, even when they would have considerably reduced the number of bar lines, invariably led performers to play too slowly. The only way to encourage them to go faster was to use larger symbols. (see fig. 1)
Figure 1 – Beethoven in difficulties: unorthodox, but he gets the message across
As if the wretched practice of reducing figures were not enough, the definition of the rhythm is affected by a second delicate question, which concerns the arrangement of the bar lines. We continue to repeat that any added sign distances the piece from how it was originally performed: how is it then possible that we find confining the fluidity of Renaissance polyphonic phrasing into a veritable cage of recurrent bar lines? The bar dates back to the seventeenth century. It became necessary during the Baroque period because it conveys meanings that are only justified by that style. It serves the purpose of differentiating the accents that articulate the tempo, dividing them into strong and weak and consequently distributing the various harmonic functions within the musical period: none of this is feasible for Renaissance music, where the time count proceeds in absolutely undifferentiated units. We would like to say that those who deal with early music have by now a certain awareness of this issue; however, the way in which the transcribers seek to meet these new requirements is still totally unsatisfactory. In polyphonic arrangements, you often find a sort of subtle premise where the editor has distanced himself from the measures that he himself has used, trying to push all the responsibility for any poor outcome onto the singers; it clearly states that the bars had been added only to aid the singers and that it really must not influence the rhythm of the piece. Aiding the singers: could this be a good justification in support of all the transcriptions into bars? It is also true that, reading between the lines, we have been forced to open the door to many other compromises; let’s try to look at the issue more closely.
Adding bar lines perhaps helps the singers too much: normal polyphonic scores never contain more than an average of four figures between one bar line and another. A very low average compared to all other living music repertoires, whether Baroque, Classical, or Romantic. It is said that this is inevitable because, while a professional – any third-year student – can read the Inventions by Bach with twelve notes per bar, an amateur chorister would have some difficulty with such long bars. But is it really true that the members of our amateur choirs have such a low skill level? Their repertoire can easily include Bach’s Masses, Vivaldi’s Gloria, Mozart’s Vespers and lots of other music that is divided into very complex bars. Of course not: only with early polyphony – notoriously dead music – are the amateur choirs and their conductors literally insulted by having four figures per bar.
The choice to put a cage around the polyphonic repertoire in bars of two single times is actually the fruit of a much deeper motivation. So deep that the philologist himself, if he is really aware of it, refuses to acknowledge it; and it is a reason that can never be blamed on the performers, because it is, by definition, to do with music theory. The fact is that by having duple time bars, the modern philologist believes, or would it be better to say deludes himself, that he is conforming his transcription to the statements of sixteenth-century music theory.
The question is now very specific, but not enough to close the door in the face of those who have only a smattering of music theory. Our research must start, once again, from the phenomenon of changing figures over time. Sixteenth-century theorists were never completely aware of the existence of this constant shift, and least of all of its importance: as often happens to those who focus too much on purely academic matters, they were simply a few decades behind.
This means that at the time in which Verdelot and Janequin wrote music using semibreves, theorists were still talking about breves, and in the days when Marenzio and Wert were using minims, the theorists were still talking about semibreves. Renaissance musical theory, in other words, was never really contemporary to the compositions it dealt with. On this decisive point, the modern philologist absolutely needs to know what to do: only his ability to read the harmonic and contrapuntal weave can help him distance himself from the theoretical statements. If he does not have the courage to do this, he will end up giving the entire transcription double the value of that intended by the composer. If the piece is set up in semibreves, he will place a bar line after each breve; if the piece is set up in minims, he will put a bar line after each semibreve. At best, its layout will cause the performer to believe that Renaissance music makes use of some kind of binary movement: this is a concept which is in no way suitable for the style of the era. Renaissance music thrives on single time, at most liable to a binary subdivision; the falsely applied binary division referred to the upper value can only horribly cage the changeable vitality of the phrasing to the point of making it completely unrecognisable.
Our argument now rests upon the full complement of common figurations that pass themselves off as true and appropriate rhythm marks for polyphonic compositions: dotted notes, suspensions, cadence markings and so on. Those who will feel this does not constitute substantial proof need only uncover the equally many instances in the works of the early Renaissance, the period in which the theorists speak of cut common time alla breve: within the chanson and the frottola, which adopt that time signature, it is not uncommon to come across sections containing an odd number of semibreves.
This is easily established by observing the last beat of the transcriptions, in which the final note sometimes lands on the downbeat and sometimes seems to land on the upbeat; in these pieces, which modern editors obstinately pigeonhole as alla breve, the only legitimate reference point can be nothing other than the semibreve. Further, take note that in many of these cases the editor may seek to disguise the evidence, by placing in some well-concealed location a single bar containing three semibreves to ensure that everything tallies when we reach the concluding cadence.
Fig. 2 – Ambiguity in Arcadelt: which is correct?
Further powerful evidence, originating from the following decade, can be found in the madrigals of Verdelot and Arcadelt. It is well known that the aesthetic conventions of the madrigal adopted the requirement that the music must always be different in every episode, that is in every phrase of the text: however, this was not the case with the so-called proto-madrigal, in which entire phrases might easily be repeated for a different part of the text. In the music of both of these authors it is not unusual to come across sections repeated note by note, clearly identical with regard to musical substance, for which the modern bar lines fall in different places: the notes which the first time landed on the first beat of the bar land on the second beat when repeated, and those which the first time around landed second end up in the repetition landing on the first.
The essence of these observations ought to be clear: even in these cases bar lines drawn alla breve illegitimately group together two semibreves which actually have the same effect on the metre of the rhythm. This results in an absurd transcription, which inevitably contradicts itself when one seeks to make sense of the contrapuntal arc; a section which will sound still more illogical the harder the conductor tries to faithfully adopt a gestural pattern based on the beat (see figure 2).
The picture would be incomplete without mentioning the last bad habit of transcribers, that of modifying the time signature to accommodate the durations that they themselves have modified and the bar lines that they themselves have positioned. Following this path it is truly impossible to imagine an authentic resurrection of the entire musical tradition of a full century. Most of the defects found in modern transcriptions lie solely in the fact that the editors reserve for themselves the right to interfere on a whim with time signatures, note values and the beat: modifying every step of the way the relationship between these three variables means in the end that when the performance is unsatisfactory, no one knows which of the three parameters to blame for it. Transcribing the Déploration sur la mort de Ockeghem (Josquin’s famous motet written in the final years of the fifteenth century), Smijers maintained the original time signature and figures but inserted a series of bars so close together as to actually suggest a performance alla minima: during a famous final of the ‘Guido d’Arezzo International Polyphonic Contest’ in the early nineteen-eighties, this piece which should take not much more than two minutes was stretched out to the incredible total of nearly twenty minutes. Not that today, more than two decades later, professional recordings do any better: none of them manages to be less than six minutes. In this instance, we are told that the fault lies in the note durations which were not adequately reduced: rather it lies in the patchy education of an entire generation of performers who do not possess even the elementary ability to count alla breve.
What of the madrigals of Marenzio? These were written much later, in the 1590s, with the uncut common time symbol, and therefore they are set up in minims; yet in modern editions these are frequently transcribed in bars of a semibreve while leaving unchanged the time signature, so that their meter ends up tragically turning into a very up-to-date four-four time. Avoid creating even the illusion of a duple measure? And how! In this instance the unwary conductor risks falling into the trap of a fictitious quadruple meter. Each one of us, in our personal library of recordings-that-should-never-have-been-made, certainly possesses many examples that illustrate the consequences of this wretched coincidence (see figures 3 and 4).
Figure 3 – Marenzio’s trap
There is one last consideration, one which is generally disregarded, which I consider decisive in making a good transcription. It is its general and overall layout: how many notes should go on a single line, where to place the D.S., how to organize page layout. I have seen too many editions that might have suggested interesting solutions but were unusable because of their unsuitable format, mile-long spacing between notes, or asymmetric layouts of parallel phrases. Here too there are no guarantees, and the quality of the finished product depends entirely on the skill of the editor; but those who undertake the training of a conductor should seek to inculcate at least a minimum of awareness of these very delicate problems. It is frequently upon mistaken details such as these that the ultimate success or failure of a transcription depends.
Figure 4 – Setting a record for Josquin: one note per beat (transcribed by Smijers)
It is already time to draw our conclusions and we have not yet spoken of the positive aspects. So far in this work we have limited ourselves to the pars destruens regarding the bad habits we inherited from the past: we have taken up all of our space distancing ourselves from almost all of the proposed solutions coming from the two centuries in which reviving early polyphony was unsuccessfully attempted. There has not been enough time to emphasise that, really, transcription is beautiful: perhaps because even today the art of transcription has not yet begun to show all of its most surprising potential. We should come to see transcription as a wonderful opportunity to be exploited to the full: a good edition can shed light on the formal structure of an entire piece, it can breathe new life into the phrasing or bring a greater understanding of the various facets of the composition; it can even communicate original and exciting interpretative suggestions to the performer.
Without doubt, the coming decades will provide us with both time to return to these discussions and material to discuss. For the time being we need to do our part, and our part is to understand this fundamental concept: it is useless to look for a shortcut if the shortcut turns out to be more arduous and impassable than the main road. Explaining to singers and conductors the meanings of the symbols used in mensural notation within their original contexts is a process so much simpler and easier than attempting to reconstruct a theory of performance guided only by some improbable graphic transliteration. And it is above all the only process by which we can legitimately expect artistically valid and lasting results. The current state of our transcriptions much too often brings about polyphonic performances that are dismal and asphyxiating; it comes as no surprise to learn that for singers – and for their audiences – a villanelle by Banchieri or a balletto by Gastoldi are still today much more gratifying than an absolute masterpiece such as a motet by Palestrina or a madrigal by De Rore. We must start expecting our teachers to show us how to read the written music of past centuries, not only at school, but also at the moment of actual performance.
Bear in mind that this wider issue is applicable not only to early polyphonic music: the need to make out the meaning hidden within the notation applies to all periods of the history of music.
When Handel or Corelli write Largo or Grave and then expand the note values, it is not the same as when they conserve or reduce them. On the contrary, by simply bringing up this matter in this way our thinking is already pointing in the wrong direction: since music first began to be written, it is the notes which should speak first. The indications of tempo can at best further clarify what the notes have already said. In my personal library of recordings-that-should-never-have-been-made the examples are not all from the Renaissance, and they contain performances by some of the most highly regarded names on the international scene: even in the best families one may study for a lifetime bowing techniques or fingering, extracting them with loving care from the original sources, and then slip up wretchedly when it comes to deciphering the movement of an Andante. The resurrection of all of the music of the past will be a dream forever unfulfilled until the issues touched upon here are included, with a much higher profile than is the case today, in the subjects studied by the future generations of performers.
Translated from the Italian by Mirella Biagi, Italy/UK; Hayley Smith, Italy/UK; Anthony Litchfield, USA.
Edited by Mirella Biagi and Gillian Forlivesi Heywood, Italy/UK