Un ensayo inédito de Bob Marvin

Hace un par de semanas recibí este mensaje de David Ohannesian: «Tengo un artículo de Bob de 1985, titulado Renaissance Styles on Recorders. ¿Sabes si ha sido publicado, quizá con un título diferente? Son 48 páginas sin contar los ejemplos musicales, así que imagino que debe de haber resultado demasiado largo como para que se lo publicaran».

No se equivocaba. Imagino que lo había buscado, sin éxito, en mi lista de artículos de Bob. Pero yo nunca había oído hablar de él, así que tenía mucha curiosidad por leerlo. David tuvo la amabilidad de enviármelo, con la ayuda de Beverly Simmons, que escaneó los tres documentos: el manuscrito original, que incluye los ejemplos musicales; una versión a máquina del artículo, mucho más legible pero sin los ejemplos; y la carta de Bob a David de 1985 que acompañaba el artículo.

En la carta, fechada el 1 de Trimilchi (de three milch, «mayo» en sajón antiguo, por el hecho de que en ese mes se ordeñaba a las vacas tres veces al día), Bob nos da un poco de contexto sobre el artículo:

La revista American Recorder me ha pedido un artículo, sin sospechar que acabo de leer «Guerra y paz» siete veces seguidas. Lo que les he enviado aparecerá con total seguridad expurgado, si es que aparece. He pensado que a ti y a los tuyos os gustaría echarle un vistazo al prolijo original.

Ha llevado un tiempo que el artículo vea la luz, pero me alegra que finalmente lo haya hecho. A la hora de editarlo, he intentado modificar el texto de Bob lo menos posible, usando la versión manuscrita como referencia para incorporar los ejemplos musicales. Primero creé una versión con OCR sobre la que trabajar en la versión web del documento que podéis encontrar a continuación. Aunque Bob no usó encabezados en el texto, la versión manuscrita incluye un índice de contenidos con números de página, que he añadido al texto original para facilitar la lectura.

Preparar la versión para web me ha llevado un tiempo que no tengo, pero, considerando las muchas horas que Bob trabajó para mí como constructor, es lo menos que podía hacer. Estoy seguro de que le hubiera encantado verlo publicado, y agradezco enormemente a David Ohannesian que lo haya puesto a mi disposición. Es una alegría poder compartirlo aquí, en un medio sin las lógicas limitaciones de una publicación impresa, y especialmente hoy, una fecha en la que Bob hubiera cumplido 78 años.

¡Feliz cumpleaños, Bob!

Renaissance Styles on Recorders

Introduction. Articulation and Tonguing

There’s little argument that instrument players (especially of recorders) in the 15th and 16th centuries imitated vocal styles. What follows here is an informal discussion of what those styles might have been and by what techniques and tricks recorder players can imitate them. I don’t mean to set definitive answers, but only to suggest possibilities consistent with historical evidence. And I’ll treat mostly those possibilities not much practiced today, which need extra consideration if our musical choices are to be based on aesthetics instead of habit and convenience.

A characteristic of syllabic (as opposed to melismatic) singing is the variety of sound from note to note. It’s a special effect in singing to directly repeat a note with the same sound. This variety can be vaguely analyzed into different components, variously practical to imitate on recorders.

Dynamics are not a recorder’s strong point. You can refinger, you can test the musical context’s tolerance for notes out of tune, and you can phrase and articulate for “softer” or “louder” impressions, but you usually can’t get close to the dynamic expression possible with voices and many other instruments. The same can be said of the different vowel sounds of voices. Some recorders have more uneveness of timbre from note to note, and even this random variation seems to imitate the variety of vowels in different syllables. Different fingerings give different timbres, and maybe the shape of a player’s mouth can affect the tone. But this imitation is weak compared to the possibilities of some other instruments.

The articulation of words, especially patterns of liaisons (the way syllables group and tie themselves together), goes much better on recorders. Languages vary in the prominence of the articulation of consonants. Some accents are slurred and rich in nuances of tonal modulation, while others are crisp, their syllables neatly defined. Old recordings of French popular songs are articulated to the point of patter, while a more modern style emphasizes modulations of vowel qualities. Recorders seem to need more variety of articulation than other instruments, which can get away with simple, unvaried articulation without sounding as stupid as recorders similarly played. Perhaps this comes from a recorder’s simpler initial transients, which need artificial animation.

In tonguing, the breath pressure behind the tongue at the moment of release makes the stress, while the “syllable” uttered characterizes the separation (or lack of it) between notes. Rhythm can be mostly independent of patterns of stress and separation (although some rhythms seem easier with certain tonguings — TR is easier short-long. This might be inate or just our habit). The standard tonguing consonants are T (or D), R and L. The chief difference between spoken T and D is whether the voice buzzes or not at the moment of breath release, but our practice is to make Ds softer. They certainly sound softer with their release of breath partially mashed by the voice. But every position and movement of the tongue and breath which can sound as D will be accepted as T if unvoiced (and vice versa). There’s an enormous variety of tongue positions and movements which will give much the same consonant sound, both in speech and instrument articulation. What all these Ts have in common is a complete stop of breath and voice (let it go through the nose and you have N). Through a recorder, the effect’s the same — complete momentary silence. It’s possible to make the moment very short with very light, quick tonguing, but the quality of the consonant becomes vague, almost a type of dental R.

The dental R, on the other hand, is a quick flip of the tongue which touches the palate only in passing. Its characteristic sound comes from the quick interruption and momentary compression of breath. It seems impossible to start a note with R. Among the various kinds of R, dental, uvular, and American, you can articulate with the latter two, but there’s not historical evidence for their use. Anglophones use types of dental R’s, usually spelled D or T, as in the British “veddy” or the American “Madduh.”

And there are different Ls. The American L is voiced (like our R, but with lower resonance behind the tongue). It partially blocks the throat, producing an articulative compression (and sometimes, as with R, an audible thud in the mouth) at the moment of blocking, with another articulation at release (when we feel we’ve uttered the L). This may be what Agricola was describing by his “tellellell…”; it corresponds to the later “did’ll” and “tootle,” and perhaps the Italian “ler” (such a long, voiced L is uncommon enough in Italian, I was told, that the sound might have been spelled with an “incomplete” R). The voiced L, like the R, can’t start a note, but only pulses an established sound. The L, more common in continental Europe, has a very brief voiced beginning. It has more of a dynamic flick of the tip of the tongue, with less blockage by its sides. It can start a note if the breath is carefully coordinated. If the tongue is curled back slightly, so the underside of its tip touches the palate, a last-minute breath can be used to blow it forward. This position helps the alteration L—R, where the tongue flocks forward (L) and backward (R) like a paintbrush against the palate. Combined with other consonants, it can sound much like the “blup” of a dental R.

The force or stress of R and L can come from diaphragmatic push, or from compressing the air behind the tongue at the moment of articulation, with a faster or harder movement. With a full breath, the brushing of L or R, or the thudding of lid’ll can be made so light as to gradually disappear. Or, with a light breath, they can be made hard and emphatic.

Part of articulation is the envelope of breath that starts or ends a note. A strong note can be started or ended gently and gradually and vice versa. Without changing fingers (practically only for a few fingerings) this is always a compromise with tuning, but we seem to hear the articulation much faster than the tuning, so there’s some leeway. Syllables starting with vowels, liquids, Ms or Ns, often need to be imitated by slow, sneake attacks, perhaps without any tongue. And imitating an Italian trillo by pulsing a note only with breath is quite effective.

Applying this to imitate words is not unambiguous. Our sense of the stress and liaison of syllables comes partly from their innate sounds, but also from their syntactical relations and the meaning of the speaker. Often there are contradictions between groupings of sound and sense. The 16th century German version of a Flemish song starts:

Ach Gott wem soll ich’s klagen

The sound of “Gott,” especially compared to its neighbors “Ach” and “Wem” would suggest a well-defined T for tonguing, followed by an R for the gentler beginning of “wem.” But in meaning “Gott” belongs more to “Ach” than to “wem.” One might say “Ach Gott” almost as one word, but to make a unit of “Gott wem” sounds strange. Likewise, there’s the possibility to use the “s” in “ich’s” to tie it to “klagen,” but that starts sounding like a German version of “Mairzy Dotes.” So I would tongue it thus:

Ach Gott wem soll ich’s klagen

Ach Gott wem soll ich’s kla- gen

According to sound, Sur le pont d’Avignon

should start TRT, and indeed it’s commonly sung with “sur le” almost one word. But more careful diction and attention to meaning would tie “le” more to “pont.” TTR. Two ways of singing, two tonguings.

This 15th century song begins unambiguously

L’amour de moi

Both the sounds and the sense suggest TRTR. But later, the same melody carries the words “Et aussi faict,” which is more ambiguous, tending toward TTRT if no elision is made between “et” and “aussi.” For such a simple song however, a consistent articulation might fit its repeating motifs better than an elaborate but confused variety of tonguing. Either way, with the meager expressive resources of recorders, some pattern of articulation is better than ungrouped blandness.

Few of the 16th century rules for text underlay have implications for tonguing. Setting the last syllable to the last note (some 15th century manuscripts have lines drawn between words and notes which indicate ultimate syllables or penultimate notes) suggests a definite tonguing at the end of a piece. The important rule is that the figures

carry only one syllable. This is contrary to baroque music, especially French, where the second note in ♩.♪♩ is often tied more to the following than to the preceeding note, to be tongued TTR (called “rebound” tonguing in modern band practice). So it’s


in the renaissance. The rule that a series of fast notes and the long note following them all get the same syllable was tempered, first in French songs where the long note could receive another syllable, and then more generally as fast notes could be grouped by two’s.

However, such rules are sometimes qualified with “rarely” or “almost always,” so occasional exceptions can be historically justified. But this tendency of not letting short notes stick out makes the fast tempi suggested by early 16th century evidence more compatible with flowing grace.


Renaissance tempo evidence pivots around Gafori, who, as a young man, wrote in the late 15th century that an undiminished semibreve got a human pulsebeat. He mentioned that older musicians applied this beat to a breve. Going earlier, an Italian doctor of the 15th century, Savonarola, tells that a heart should beat with a tempo of 4-6 minims per pulse. And a century or so earlier Verulus divides the year into months, days, hours, and finally little atoms of time to define the lengths of notes. His language is ambiguous, but one interpretation gives a tempo close to Savonarola’s and Gafori’s ancients. Savonarola used Italian mensural music to define the beating of a healthy heart, which suggests little variation in the tempo of such music. If Verulus’ tempo, also for Italian mensural notation, agrees, that suggests a long stability of tempo.

The transition of tempo in the 15th century, from ◻ [brevis] pulse to ○ [semibrevis] pulse is not always clear, especially in French music. Dufay seems to span the range of possible tempos. Whether the change in his tempo was gradual, or linked to notation (black vs. white) and minimum note values we don’t know. Some songs in the carefully copied Mellon MS of 15th century chansons leave off the diminishing stroke through the mensural sign found in other manuscripts. Does this imply a confusion of tempos in transition? The diminishing vertical stroke, 𝇉 or 𝇍, originally in the 15th century seemed to imply a quickening of tactus by 3/2. Doubling was indicated by “2,” where the same tactus was applied to the next bigger note value. Most l6th century sources say 𝇍 is double the speed of 𝇋, with the same tactus applied to ◻s instead of ○s. An early 17th century Dutch author accuses musicians of saying “half” when they mean just a little less, so 𝇍 was actually only 4/3 faster than 𝇋.

Were ancient pulses the same as ours? Evidence says pre-industrial hearts beat 60-80 times a minute. If lack of angst and weltschmerz calmed the antique breast, a smaller body probably compensated.

Various tempo citations are found throughout the 16th century, linked to pulses, grasscutting strokes, and walking. It seems clear that tempo slowed down, both in terms of the time a minim got, and the general impression of the music. In the early 17th century, Praetorius’ tempo for a church organist is an unambiguous half note = 42-1/2 / min. Mersenne says each piece of music should be preceded by two numbers, one giving the frequency of the first note (his two pitches are about a tone and a minor third above a=440, but don’t seem trustworthy. The other gives the length of pendulum beating the measure. His tempo is about half note = 60/min. These are roughly in agreement with somewhat later indications that half notes went as you would count “1, 2, 3, 4.” (One quirk of music history is that note times changed by a factor of 2 or more per century up to 1600, and then stopped changing). But Mersenne, in his theoretical section, also parrots Gafori’s tempo. This cautions us that some of the relations of tempo to pulse in the 16th century might only be lip service to authority. Zarlino and Stocker comment that new composers (supposedly Willaert, et. al.) use minims where older composers (Josquin, et.al.) used semibreves. While not specifically talking of tempo (it’s more in a context of counterpoint and treatment of dissonance), they suggest that the 16th century tempo change was related to style of composition. In addition to tendencies towards types of melody and counterpoint, frequency of fusae (♪) and syncopated minims might guide our choice of tempo.

Melodic accent and tactus

The range of tempo, both among contemporary pieces and within a piece, seems to broaden through the 16th century. Direct evidence for this is spare, but all sorts of things suggest it. Words and rhetoric became more important, and counterpoint allows more and more rhythmic liberty and variation of speed. That tempo was not much of an expressive variable before well into the 16th century is suggested by the strict tempo of Italian music implied by Savonarola, by the late 15th century mass movements (especially Agnus Dei) which repeat in diminution and by Glareanus’ recommendation that repeats be taken faster so not to bore the audience. Tactus was compared to the regular ticking of a clock in the early 16th century, but such regularity seems more and more out of place as the century progresses.

The notion that each piece of music has an intrinsic tempo to be discovered was weakened by an experiment in which obscure compositions were rewritten in similar note values and given to an experienced ensemble to find their tempos. The tempos they agreed upon bore little relation to the composers’ original intentions.

The relation between tactus and melody, especially in the 15th century, is unclear. Many simple rhythmic figures seem to have a strong intrinsic sense of pulse. So when they occur syncopated to the tactus, it’s hard not to give them full expressive treatment independent of the tactus, which becomes only a silent time beeper. Music of India is an example of this, where hands and feet keep the tactus, while the melody often gives no clue of its relation to that tactus. The relation between melodic accent and tactus is a major expressive factor, as in quantitative Latin poetry, with its tensions and resolutions of accent and quantity. When the same figure reappears in a piece syncopated, it should be possible to play the same way, independent of its relation to the tactus. A fuga in 𝇍 at a minim is an extreme example of this. Modern musicians tend to make syncopations recognizable without an audible tactus by playing more rigidly, in stricter rhythm than when the same melody is with the tactus. The ability to phrase and articulate independent of tactus (except as a timekeeper) is worth cultivating.

But how free from tactus should accent be? In some pieces, quite A duo from Trent 88, “Agwillare,” has repeated figures like

This passage certainly has a strong sense of its own beat, but it’s all syncopated. The first minim is on a semibrevis tactus, and the final brevis is not, the following minima rest with its dot of “completion,” shows this was the intention. The two voices relate quite simply to each other, so all sense of syncopation is quickly lost (unless a tactus is given visually).

Other pieces have so much coloration that 𝇊 sounds like 𝇈

such as “Mon tres douce cuer” from Canonic:213, which begins and ends (with much the same in between) thus (the numbers tell how many minims in each note):

The mensura is 𝇊, putting a tactus on a semibrevis (= three minims). Accenting according tactus would emphasize the arrowed minims. But the figure including the third and eighth notes is so common in the form

it could be thought of as groups of three, two and three minims. The last six notes could be considered similarly

Accenting here according to tactus seems to introduce a nervous energy at odds with the rest of the piece. It’s not clear why this song was written in 𝇊 instead of 𝇋 or 𝇈, but the tension between the tactus in 𝇊 and the actual sound in 𝇋 is interesting.

However “Mon Vrai Desir” from Bologna 915 can profit from either interpretation. It starts (without text)

The cantus can be grouped

This is quite subtle and relates interestingly with the tenor, which has the same possibilities. But grouping according to the tactus, accenting the arrowed minims, gives a different effect, more vigorous and lively, especially the tenor. The alternation of the 𝇈 hemiolas with the 𝇊 patterns is also very interesting. Perhaps a combination of both types of accent would be best.

Do words give any help? Not often. Usually there are more notes than syllables, so there are different possibilities. “Mon vrai desir” starts without text, and in other songs it’s not clear where the text begins. And where the text underlay is clear, the normal accents of the syllables are often ill-coordinated with some strong intrinsic melodic accents.

From CAL 5943

Pange Lingua Gloriosa

we have a splendid example of early 15th century insensitivity to text.

And from Estocart’s 16th century setting of Psalm 42

Le grand le grand Dieu vivant

we have an unnatural accenting of “le,” which might signify “le seul,” but it’s not treated consistently in other voices. Old Roman marching songs let feet fall on normally unaccented syllables. We sing “God save our gracious Queen” and “My country, ’tis of thee,” both iambic (u|), as dactyls (|uu) without complaint.

Words can guide accents, but they can’t be depended on. Text underlay must often be chosen, sometimes arbitrarily, among many alternatives. An instrument should be ready and able to follow a singer’s choice.


Tuning can be expressive. The choice is usually between Pythagorean (close to 12-tone equal temperament) and harmonic (or just, or pure, of which meantone is a temperament). Pythagorean intervals are all concocted from combinations of pure fifths and octaves. Its imperfect consonances are less consonant than harmonic thirds and sixths, which have no beating. The difference between them is about a fifth of a semitone. The pythagorean ditone is this amount bigger than a pure major third. The semiditone is similarly smaller than a pure minor third. And so on. The scales look roughly like this



The Pythagorean semitone E-F, B-C is smaller than its harmonic equivalent and gives more contrast to whole tones. Thus the different species of fifth and fourth are more sharply defined by the position of their semitone. As just scales are more uniform than Pythagorean ones, so are their harmonies. Almost all notes in a just scale make good consonance together. The result is pure and stable, sometimes to the point of cloying monotony. There’s more of a continuous range of dissonance to consonance in a Pythagorean scale, giving a nicer balance and more variety. With just intervals, there’s a greater need to artificially introduce dissonance into the counterpoint.

All intervals produce difference tones in our ears, a lower sound often buzzing nastily with high recorders. A difference tone has, indeed, the frequency which is the difference between the generating frequencies. So the difference tone of a just interval has a just relation to both generating notes. They can be easily reckoned by finding intervals on the harmonic series, and simply subtracting the numbers of the harmonics. A pure major third is found between the fourth and fifth harmonics, so its difference tone is the first harmonic, two octaves beneath the lower note. The minor third is between the fifth and sixth harmonics, so its difference tone is two octaves and a fifth below the upper note. G and B give another G two octaves lower; G and B—flat give a low E-flat (a 6 minor triad with difference tone E-flat has the same notes as a major seventh chord of E-flat, which is the basis of the ambiguity of seventh major chords and their relative minors). The major sixth is between the third and fifth harmonics, with a difference tone, the second harmonic, a fifth beneath the lower note and a tenth below the upper. And the minor sixth is between the fifth and eighth harmonics, with the third harmonic as difference tone, an eleventh below the upper note and a major sixth beneath the lower.

The difference tones of Pythagorean imperfect consonances are shifted a semitone from those of pure thirds and a quarter tone from those of pure sixths. So a pythagorean G-B gives a dissonant G-sharp difference tone. That, with the beating overtones, make it a harsh discord. But G-B-flat gives a D difference tone, completing an inverted triad. Even though the overtones are beating, this pythagorean semiditone has a rich depth unmatched by the pure minor third. The pure major third is more consonant than the pure minor third, but the relation is reversed in pythagorean intervals, where dures is hard and mollis is soft.

The pythagorean tritone is fairly consonant, its difference tone close to a semiditone beneath the lower note.

Only from the end of the 15th century was there much academic effort to describe pure thirds and sixths. Before that, it was all Pythagorean, with occasional mention of tempering ditones, smaller to make them consonant. The close intervals in British singing were noted by the Romans. It’s an Englishman, Walter of Oddington, who gave one of the first references to using pure thirds. Perhaps this was a British specialty which spread to the continent in “la contenance Angloise.” It would seem that pure thirds and sixths were known and used by the end of the 14th century; their use increased through the 15th century, and were firmly entrenched in the 16th century (which equal temperament started making headway). Their official incorporation into theory was probably delayed by the far greater mathematical complexity they entail. But Pythagorean intervals still were known well into the 16th century. It seems likely that plainsong was pythagorean-based, as was much early training of musicians in the early 16th century. Increasingly through the 15th century, cadential and important imperfect consonances were tempered pure, until they became more the norm than the exception.

Pythagorean tuning is good for defining melodies and giving naturally varied harmony in relatively thin counterpoint. It emphasizes the “hard” quality of notes solmized “mi” and the “soft” side of “fa”s. In a Pythagorean first species fifth (re mi fa sol la), the third degree, fa, feels relatively stable and calm, but in the fourth species (ut re mi fa sol), the stable fa is the fourth degree, mi, feels unstable. A just “mi” would feel more stable than the “fa” between it and “sol.” Tritones are acceptable harmony. For the 14th century, it’s occasionally tempered Pythagorean.

Just tuning is good for rich, thick triadic counterpoint, where even melodic thirds and sixths have implications in harmony. Their sharps and flats tend to give a major/minor, cheerful/sad effect. The major third is calm and stable, both melodically and harmonically, while the minor third is less so. It’s useful for magnificent contrapuntal architecture and rich tonal tapestries, but for personal expression, it’s inferior to Pythagorean.

For late medieval-early renaissance music, it would be ideal to have the control to choose Pythagorean or pure intervals for their characteristic affects. The parallel thirds of Q 15’s Alleluja Katerina could be adjusted to give tensions and releases, good and bad notes and intervals:

There are clues to the use of melodic dieses (quartertones), from Marchetto at the beginning of the 14th century and finally from early 16th century Cardano. Marchetto is unclear and possibly confused himself (he may have had the pure tempering of thirds and sixths imperfectly in mind). Cardano was perhaps crazy, but clear. Without parotting Boethias, the standard authority on Greek music, he told of dividing melodic semitones into dieses as a nice ornament, adding that it’s most suitable for strings, less so for winds, and least for voices. This gives an image of diverging instrumental and vocal styles, although its application in most notated music seems dubious. Present day Adriatic music, rich in these ornaments, might be a remnant of lost Italian music (as evidentally was Gafori’s “false counterpoint” with its mournful movements in tones, fourths and tritones).

Rhythmic freedom

How free were people with rhythm in the renaissance? Towards the end, it seems quite free. Loys Bourgeois cites French use of notes inegales at mid century. Counterpoint in the 16th century is increasingly permissive of rubato and expressive distortions of time. Musical rhythm seems to follow the fancy of words and ideas instead of a tactus. But earlier? There’s little music in the world which limits itself to simple and rigid proportions in note values. It would be surprising if renaissance music did completely. A good argument can be made for ambiguity between duple and triple rhythms in the 17th and 18th centuries,

representing two versions of the same rhythmically flexible group. The use of coloration in the early 16th century seems comparable, semibrevis + minim became the same as dotted minim + semiminim,

but earlier in the 17th century, it divided the 𝇈 into three parts, not four, suggesting an ambiguity and variety of actual rhythm. The control of the amount of inequality in such cases can do much to make a song suave or vivacious.

National monophonic songs

Generations of using Palestrina to teach counterpoint may have dimmed our appreciation of renaissance melody. One way to revive it is through the essentially monophonic songs which seem to have characterized the music of various nations at different times. As Glinka said, “A nation creates music — the composer only arranges it.” Finding a different personality and developing a special feel for each type of song is a good exercise in style. It may be more of an exercise in imagination since we have so little evidence of how music sounded, but sensitive and reasoned guesses are better than a uniform display of unspecific musicality. And with a little style, these songs are beautiful and exciting.

The late 15th century repertory of French songs found in the Bibliotheque Nationale, Paris (f.fr. 9346, the Bayeux Chansonnier, and f.fr. 12744. There are no facsimiles, only editions by Gerold and by Paris and Gevaert) gives the themes of much early 16th century polyphony. As the soul of French baroque music can be found in the simpler songs and dances which lent it a style of melody, the stylistic spirit of polyphonic settings of these 15th century songs is more essentially evident in their pure, monphonic forms. Here are two from Bayeux — “Helas moy coeur n’est pas a moy” and “Fleur de gay te donner moi Joye.”

“Helas” is simple and effective. The humble singer refers to himself in low notes and then to his exalted love in a more celestial tessiture. The misery of the opening is contrasted to “Il est a vous” which can sound pure and high, and to “doulce amie” which can trip along in sweet grace. The

needs a gentle tonguing. T lid’ll L or T rid’ll L. The descending “amour” (echoing “coeur”),

suggests an ambiguity in the inequality, a liberty to be expressively exploited. Exaggerating the ◻ ○ [brevis + semibrevis] inequality at the beginning gives an appropriate tension, while minimizing it at “Il est” makes that calmer and gracefuller. A break before “n’est pas” emphasizes its negativeness, as well as the descending fourth. The text underlay is fairly obvious, and the song begs a high sense of drama. To achieve this on a recorder needs a real virtuosity of breath, attack and timing.

“Fleur de gayte” is just a great melody (less ornamented, but more rhythmically subtle in the 12744 version), a bit carried away with itself and less coordinated with its text (or perhaps only a subtler relation). To realize the melismatic character of the descending figure, a tonguing such as


is needed. The ♩s give good exercise in controlling rhythm, less or more than their official value for different effects. The half note + ♩♩ of “allegrement” can be played lightly

exaggerating their inequality (although it could be argued to minimize the inequality to emphasize the calm relief of “allegrement,” but equalizing the ◻ ○ [brevis + semibrevis] can take care of that). The same figure at “longuement” needs stretching out towards


in between are a nice contrast if bien pointees. Again, the actual inequality between half note and quarter note in the descending figure will do much to establish a mood, if only of an abstract character. The semibrevis rest at the start is perhaps an error, since it makes an odd number of semibreves. The 12744 version has only the minim rest. A tempo of a ◻ [brevis] per slow pulse (60–65/min.) seems fast enough yet expressively practical.

I find these fit the later descriptions of French baroque music in their strong sense of regular but often unexpected movement, benefitting from a sense of inspired and animated order. The French baroque precision of nuance isn’t out of place here, along with sentiment which, within the limits of taste, feels no constraint. The basic graceful neatness can be strained but not deformed.

Italian frottole of the early 16th century are a different matter. Minkoff have published facsimiles of the Bossinensis lute versions, from which we have an exhortation to declare a smoldering passion so it may flare into a consuming flame, “Scopri lingua el cieco ardore.” While the words suggest passionate abandon, action instead of considered sentiment, the melody lacks the sweeping gestures of French songs. It’s as if the Italians poured violent emotion into neutral melodies, while the French emotion showed itself by restraining highly charged ones.

The repeated notes on final syllables (lingua, ardore, cresciata, core, etc.) are typical in Italian music. Refingering the second note can accentuate the first by making the second weaker. On a G flauto, this is quite easy for the As, “lingua” and “ardore”

([all holes open] is A is normally [all holes open except for no. 2], or [ø123456] with renaissance bores), but other notes (D, E) give odd timbres with the necessary leakings or forked fingerings. The cadential Gs (“muta”) can be muted by leaking the middle finger. The final “muuuuuuuu” could be aspirated tonguelessly to express an inarticulate dumbness, the decrescendo on “ta” capped with a loud and emphatic end. The beginning is so syllabic it could sound quite rhetorical, notes lengthened and shortened according to the words “Ardore” should be broadened, so “Scopri lingua” can contrast with it. “Sco” lengthened (as fits its vocal sound) and “pri” and “ling” shortened to give a long-short-short-long pattern (although shortening “sco” to give Brunelli’s “migliore” lombard rhythm in the following century is perhaps more exciting “discovering”). The figure at “hormai non star” seems typical in frottole, a syncopation sounding as a triplet filled in with fast notes. One general device to establish a rhythmic grouping is to distort the rhythm within it. In this case, lengthening the minima “hor” and quickening the ♩♩ “moi” give this feeling, increase the excitement, establish a pattern to imitate at “star” (dwelling a little before rushing the ♩♩♩s), and give contrast to the declamatory style of the beginning.

Frottole can use rather definite, emphatic articulation, as long as notes are tied with Ls and Rs. It’s hard to imagine too much exaggeration and contrast of texture and feeling. A pulse-per-◻ [brevis] tempo is quite fast, but possible. Modern lutenists, who haven’t applied themselves much to syllabic articulation, will have some trouble treating their part as two independent voices, insted of plunking chords. And the ornamented three-voice Giustiniani of Petrucci’s Libro VI are very fast at this tempo (although no more than the ultimate speed of late 16th century diminutions). About 60 ◻s [semibreves] per minute would seem tops.

Clemens non Papa arranged the Souterliedekens around mid 16th century (Dutch psalm tunes, facsimile by Editions Culture & Civilisation, Buxelles). They’re three voice contrafacta of popular songs, which appear also in Het Antwerps Liedboek published by the Vereniging Voor Nederlandse Muziekgeschiedenis, Amsterdam. Some of the songs have ties to France, Germany and England, going back to the 15th century. Most seem to be ornamented with extra ♩s (e.g., “Pastyme with good companye” gains some notes here as “Waer soe machse zyn die midich heeft verhuecht” or “De ma tristesse”), sometimes in a syllabic pattern on the same pitch. “Troostliker Troost” contains the “Wilhelmus” signature, with marvelous contrasts and responses between phrases, which can be heightened by adjustments of attack and rhythm. The original songs probably had a pulse-per-breve tempo, but these elaborate settings suggest something slower, (although we should be wary of letting tempo compensate for lack of skill to keep a piece clear and graceful at higher speeds). With the range of time and place represented by these songs, tempos from 60 ◻’s [brevis]/minute to 100 ◻’s/minute seem reasonable (although many ♪s or syncopated half notes might suggest even a slower speed).

Enthusiasm and sincerity (unpsychoanalyzed) are two keys to German expression (and a link with the English). Renaissance German songs don’t seem to need French fussiness or Italian fire, but only clear, direct sentiment (but not necessarily crude or brutal). Schlick’s Tabulaturen Etlicher Lob of 1512 (facsimile by Zentralantiquariat, Leipzig) has a dozen lovely songs with lute tablature (modern transcription available). Kristian Gerwig, Hennef, has a facsimile of Forster’s Ein aussbund Schoner Teutscher Liedlein, 30 typical songs of the early 16th century, from which we have “von Edler Art.”

A small break after “Art” can help define the two phrases “von-Art,” “auch-zart” as well as prepare for a change in mood from the nobleness of the beginning to the more personal sensuality of “zart,” which can be a very sweet descent. Broad notes, firm articulation (TTRT), and regular rhythm help set the mood for “von edler Art,” and increasing flexibility in rhythm and soft attacks can do the same for “auch-zart.” The following pattern of ○ [semibrevis] and ◻[brevis] is typical of German songs and can use special recognition of some sort instead of just repeating the four-note figure. I find articulating in pairs (TTRT or TRTR) while dwelling slightly on the first ○ [semibrevis] at the expense of the second helpful. The inconsolate end of the song can benefit from minimizing the inequality of dotted half note ♩ and half note ♩♩, while “Hilff mir” becomes more urgent with a sharp articulation. The last phrase could be tongued

Hilff mir ich hab nit tro stest mehr

Hilff mir ich hab nit tro stest mehr

These songs go well in slow breves, a minimal pulse of 60/minutes or so.

English lute songs of the end of the 16th century delight in variety, moods changing with whim and fancy, full of wit and surprises. It’s difficult to guess how much symbolism and double meanings were put into them, but such opportunities abound.

It was a lover and his lasse

can represent the lover springing up and his lass gently reclining, which can be emphasized by a contrast in tempo and sharpness of rhythm and articulation, changing at “lover” by dwelling on the first syllable and softening the second (a more emphatic “lover,” short-long, doesn’t go so well with the other voices). The English took the feeling and freedom from transalpine madrigals and, infusing them with osalpine Gallic wit, refined and concentrated them.

Rosseter’s “Laura” may be a sister of Plutarch’s. The meter of the first part is ambiguous. In both words and music there’s an iambic tendency:

When Laura smiles her sight revives both night and day

But the lute starts more definitive dactyls, |uu |uu. This ambiguity disappears in the second part, only to return at the end where

and untamed dispair

seems more natural. To play up this feature you can start “When” as a separate pickup note, with a gentle TR for “Laura,” broadening the first note of “smiles” (the lute lets you do anything you want with the smiling ♪s). “Both night and day” make two pairs, to be tongued TRTR(L). “And” can come in very gently and “her speech” can flow with a slight speeding up, being sure not to break up “musicke” by taking “sicke” as a pickup in imitation of the later “the” and “of.” The lute will let you dwell on “paire,” “wounds” and “and” as long as you like. Leaking and alternate fingerings let you let these notes die away in a truly pitiful way. However, the other verses don’t have such negative images, and Laura’s praise is undiverted down dismal paths.

Expression of words

Expression of words in the 16th century seemed to start with attuning the music specifically to whatever words were being sung at the moment, with little regard for their context in overall meaning. This developed into something close to trickery where a mood would be established by a word, only to be contradicted by an innocent negative or qualifier.

Lasso’s “Quis sequitur me” seems to treat the text piecemeal, phrase by phrase, expressing each in turn in contrasting sections. We start out following a march-like fuga “Qui sequiturme,” but stumble in falling thirds and fourths at “non ambulat” to end up in a dark melisma “in tenebris.” “Sed habebit” shows hope in rising thirds, “lumen vitae” shows the light in grand, rising gestures. “Dicit Dominus” can enter with ponderous authority, tempered in a gracious melisma. Over interpretation? perhaps, but such possibilities are rare in earlier music, growing through the l6th century. More examples can be found in the following presentation of 15th and 16th century polyphony.

Fifteenh-century polyphony

“Helas je n’se decouvrir ma volente” is from the Barenreiter facsimile of Ms. VIII 24 Biblioteca del Monasterio [de] El Escorial of the early mid 15th century. There seems to be a clef confusion. It’s unlikely that the cantus starts an eleventh below the tenor. The clefs

seem more likely. In the first phrase of the cantus there are more notes (even counting ♩. ♪ pairs as one) than syllables, so the words are not a clear guide to melodic organization. Grouping according to the mensura (semibrevis = ♩♩♩) gives many units of ♩♩.♪, emphasizing several ascending figures seemingly at odds with the sentimental text. Taking the first four notes as “Helas” and the next 14 notes as perfect semibreves, feeling the beat on the dotted minims and the semibreves gives a more flowing effect and emphasizes descending figures. The same can be said for the next phrases, although “bel” seems intentionally aligned with the undotted minim rather than the following dotted minim, and the leap up of the fourth can sound quite “bel.” The last phrase “Ne me lesse la bouche ouvrir” also has more notes than syllables (counting ♪ pairs as one), so again, the words aren’t much help. But one possible grouping is

which seems to fit “Ne me lesse” better than accenting according to tactus. Here, as well as at the pattern ♩♪♪♩♪♪♩ a strict division of ♩s into ♪s seems rigid and stifling. Either minimizing the inequality towards a triple division of an imperfect semibrevis, or exaggerating it in the direction of

seems more graceful. Grouping the tenor according to the tactus accents many ♩s, giving a bounce ill-suited to the despair of “Helas.” Taking the

seems better.

Compared to other songs moving mainly in semibreves, this is a fast mensura, and ever at a slow pulse for

it takes some softening of attacks and rhythms to keep it a sentimental song and not a bouncy dance. ♩.♪ rhythms can be softened either by minimizing the inequality or by making the ♪ very soft. The triplets ♩.♪♩ can be tongued TRL, with a gentle T for such syllables as “He,” “Ne,” etc. Cadences ♩♩♪♪• can be tongued TRLRL(T).

“Vous soye la ties bien venue” is from the Forni (Bologna) facsimile of Bologna Biblioteca Universitaria MS 221G (also in Trent 87, a3). The composer didn’t keep his melody in meter, since the last note is the second ● of a perfect ◾. So we may be excused from accenting the fourth note, which would give the same ascending bounce we avoided in “Helas.” “Vous soies la tres” suggests a grouping ●●|●●|◾, but “De tous honorec” suits ●|●●●|◾. What follows can be grouped into three perfect ◾’s, an imperfect ◾, and a perfect ◾, although the melody seems to suit one perfect ◾, an imperfect ◾, and three perfect ◾’s, due to the upward leap of a third in the ♩’s at “venue” and the following ligature. I feel a strong melodic temptation to group “halte princesse” in imperfect ◾’s, ●●|●●|◾. The figure ●♩●♩● should sound like two perfect ●’s, with as sure a feeling of ♩♩♩|♩♩♩|● as if the whole piece were in 𝇊. You can tongue these five notes T(L)RLRL. As always, rhythmic distortion will help define groupings, while a strict proportion will sound stiff and “syncopated” in a modern sense. Making the ♩’s shorter and quicker than half the ●’s will let you play them softer without sounding so out of tune. At “En cest amoreuse votre [?]” a separation can be made between the first two notes to help define them better and to emphasize the elision between “cest” and “amoreuse.” This can be symmetrically reflected by a small break after the first ligature “mo,” with that ligature slightly short—long, to be echoed in the second ligature “se.” At this point something special has been composed into the counterpoint, and perhaps a special interpretation was meant to accompany it. The tenor’s syncopation ●.♩♩●♩◾ more graceful if played perfect, imperfect, perfect, ●.|♩♩|●♩|◾, with a small break after the first perfect group. The cantus can join in, contrasting light, lithesome ♩’s to the sensuous semibrevis preceeding. The noble praise of this song, without much lamentation, can be served by well-defined perfect ◾’s, the ◾● pairs perhaps exaggerated in inequality to give motion, but with a fluid rubato, especially in ♩’s and at the cadential figures ●.♩♩♩◾. The ligatures at the end of the cantus can give a sense of lift and elevation of the heart (tempered, of course, by a certain sweet sadness).

“Fugir non posso” (also from BU 2216) is more Italian. It doesn’t soar and flow, oozing French decadence. It’s busier, and more diatonic, with more variety, short choppy phrases contrasting with longer, graceful melismas. Some of the same need for articulate precision and definition of notes found in the Bayeux MS songs is here (but in cruder, more brutish melodies according to the French, more directly animated and vital according to the Italians). Pictures of small-sized instruments, an Italian predilection for agile duos, the sound of Italian harpsichords all suggest an Italian renaissance musical environment of high, well-defined sounds, rich in themselves. Northerns were busy exploring deeper voices and concocting lusher textures, which probably sounded like smarmy sauces to Italians.

Nervous distress, bordering on the frantic, seems to characterize the beginning of “Fugir,” with its hockety imitations. The articulation seems clear, “fugir non posso,” TRTR(L)R and “Daltro dolze volt” TRT(L)RL(T) (I prefer “Dolze volt” LRT for its softer “Dol”). Control of the ends of notes (♩’s at “tro” and “ze”) will help define the figure

([the first symbol] is Italian for dotted ●). A big change of melodic mood comes at “Pero chel cor” and especially at the melisma, which should sound like senaria imperfecta (♩♩♩|♩♩♩) instead of syncopations of s. perfecta (♩♩|♩♩|♩♩). It can be tongued


The tonguing at “Pero chel cor tu may,” after TR for “Pero” is a problem. To hard “C”s followed by a “T” suggest more separation, but the syntactical linking of “chel” to “cor,” as well as the latter’s sentimental softness might permit TRTRTR. The tenor has interesting syncopation there, which could be grouped

at “ferito,” the last three ♩’s taken as quick pickups to the following ◾, and the first four notes echoing the rising third at ”tu may.” “Io toy ferito” can be an abrupt, stabbing imperfecta in the cantus, with the descending “si dolziment” figures softer (although I prefer an ironic interpretation with them sharp, as seems to fit the last ascending “si dolzement” with its ornamented cadence). The choice between sweet sincerity and irony continues at “Daltro lizadro e graciosso aspeto,” with its jumpy syncopations. At the end, I prefer not to accent any fusae (♪’s), pairing them with preceeding ♩’s to make quick triplets. If this song was meant to be as sweet as some of its words suggest, despite the business (●= 60/minute seems a minimal tempo) and jumpiness of the melody, there would be quite a virtuosity of delicate expression implied. “Daltro lizadro”’s melodic imitation of the cruel “Io foy ferito” suggests a sweetness barbed with irony.

Sixteenh-century polyphony

Taking the diminished ◻’s of Diuitis’ mass movements of the early 16th century at even a slow pulse beat makes for fast music. Feeling the movement in perfect and imperfect ◻’s and ○’s helps give the calmer flow of a slower melody ornamented. “Crucifixus” can start with a perfect ◻, followed by two perfect ○’s. The tenor’s second and third line syncopations following minims rests can be felt in imperfect ◻’s. The cantus also anticipates the start of “Pleni” with similar syncopations in its second line, which can be accented to imitate “Pleni sunt celi.” Other places give more opportunity to avoid accenting minims followed by ○’s by phrasing according to the melody’s intrinsic character and not in relation to the tactus. Tonguing helps too:

TRLR for dotted minim + ♩♩♩, TRL for dotted minim + ♩+ minim, TLRL for minim +♩♩+ semibrevis, and for dotted minim +♪♪+ minim Tlid’ll L (or Trid’ll L), so there will be no sharp break, only perhaps a gently dying of breath before the ♪’s. the difference between the sadness of “crucifixus” and the positive glory of “Pleni” can be effected by attack and rhythm instead of tempo. Tonguing can be sharpened and more space left between notes in “Pleni,” leaving a more soothing effect in “Crucifixus”. In “Pleni,” short nots can be made shorter and long notes longer, exaggerating inequalities, while “Crucifixus” can begin with the minim lengthened, taking some time from the preceeding dotted brevis. The following ♩’s can be broadened at the expense of their preceeding dotted minim and minim, and so on.

“Entlaubet is der Walde” is a sad song from Rhau’s Bicinia of 1545. The sentiment of this song and its busy counterpoint make it difficult to keep a 𝇍◻ tactus. But in 1547 Glareanus was still saying this was preferable, although a brevis tactus could be used while learning a piece. Grouping mostly in breves and dotted semibreves, gentle attacks, and smoothing out inequalities can still give a rather thrilling sadness at ◻ = 60/minute. In each voice, the initial theme is presented without rhythmic counterpoint. The actual rhythm can be completely ad libitum. The pairs

whatever ratio you want. To get more control over these divisions a metronome can be set to perfect semibreves (dotted semibrevis = 80/minute) and the first phrase played to that beat (making sure “de” comes before the click). This way, it’s relatively easy to modify the inequality of the semibrevis + minim pair until they divide the dotted semibrevis equally, and the semibrevis +♩♩ of the second phrase becomes an even minims triplet. Other semibrevis + minim pairs can be likewise smoothed, especially at cadences. As in “cruxifixus,” dotted minim + semiminim can be broadened to slow the minims a little, and similar tonguing can be used. A decision must be made at the cantus’ “macht,” whether the “C” minim is the end of “leiden,” making a descending third, or whether it’s the beginning of “macht,” which gives “macht” a mightier feeling than it had as a minim pickup, and places it between the tenor’s “macht” and “ein.”

Certon’s “Contentez Vous” (from Bicinia, siue Cantiones Suavissimae, 1530, but also in Rhau) has similar problems of tempo. With a large repertoire of French songs, it’s poised in the early mid 16th century, at a time of tempo transition. I find its melodic structure faithfuller to earlier habits than to new madrigali and motetti. Taken with a 𝇍◻ pulse (or quick 𝇍○ tactus), it’s busy and breathless, especially the alto, with its phrases ending on dangling minims. But a fast tempo avoids the maudlin smarm, uncharacteristic of Gallic wit, which dominates serious performances at more “reasonable” speeds. It’s possible that these songs were virtuosic balancing acts between pathos and patter, a light velocity helping the French keep the taint of naive sincerity at a healthy distance. The foe of smooth speed is the patter of regular minims. Group them with articulation, rhythm and accent, to let patterns of ○’s and ◻’s come through, and speed seems more plausible. The counterpoint seems somewhat elastic, allowing the tenor a little extra time at cadences ending in minims with the next phrase beginning tight on its heels. The accord between melody and text helps, letting the song find its own rhythms and tempi, instead of feeling as tied to a regular tactus as music of previous generations. “Par l’heureuse fortune” is fitted into accents in different ways, yet they all seem proper. The second “fortune” in the superius and the first in the alto have a long accent on “for” minim +♩♩ and dotted minim +♩ respectively) while the other time, the length and accent seem to be on “tune.” “Par l’heureuse” appears grouped once as minim|2 minims|minim and once as minim|3 minims, that is

Awareness of such patterns can help alleviate the patter and hint at pathos.

In both these songs, a fast tempo still allows expression of the text, both in articulation and general sentiment. An instrumentalist can become an actor and play not notes or phrases, but feelings. There is here a tension with the business of the melody and counterpoint which later music resolved in favor of verbal expression. But these songs can have a sweeping majesty unavailable at slower speeds. While historical evidence for such tempos is ambiguous, if not tenuous, I find it a possibility with great aesthetic appeal. But it takes some practice and shifting of gears to try.

With Lasso’s “Oculus non vidit” at late mid-century (also from Cantiones Suavissimae) we have music truly coordinated with words. (The 12 instrumental ricercari which accompany these 12 vocal motetti are written in almost half the vocal note values, suggesting a tempo about half that of earlier music, with a 𝇍 semibrevis getting a moderately slow tactus of 60/minute or so). It can sound good with almost no sense of regular tactus, the voices following in response to each other. The counterpoint allows great rhythmic freedom and rubato, which seem to be suggested in the melodies. The ascending line at “ascendit,” beginning in the superius, can start with slow dotted pairs, accelerating to emphasize the ascent, broadening the arrival on the “F” ♩ . In the contratenor, the minims before “ascendit” and “Deus” might be treated as passing notes if it weren’t for the words and the superius’ example, which make those two minims the ends of small phrases. The same can be said for the minims which end “diligent” before “illum” in both voices. “Oculus” can begin with no audible tongue, the swelling, perhaps quite gradually by changing fingerings (if, fingered as the Xth note of a recorder, the left middle finger can leak to allow soft blowing; if taken a tone lower, as the IXth note, the middle finger can leak again, or you can overblow the IInd notes, leaking the thumb). “Cu” and “lus” belong together as minim-semibrevis, to a lesser extend, “non” and “vi.” So the opening could be tongued

Oculus non vi dit

Oculus non vi dit

“Qui” sounds good separated from “Diligunt,” giving contrast to the fluid cadence of “Deus” and emphasis to a serious “diligunt.”

“Oanim’ accecata” is from Razzi’s Laudi Spirituali of 1563 (Forni facsimile), a large collection of simple songs to religious texts for two, three and four voices. It continues an earlier tradition of laude going back into the 14th century, in which groups of amateur singers met regularly to sing laude, some of which were traditional melodies, some specially composed. There are many early 15th century laude in BU 2216, some in simple settings, some ornamented and counterpointed into higher art. What art there is in most laudi (sometime “la laudea” became “il laude”) seems to lie in artless sincerity, the expression of genuine religious sentiment. Taken just as simple notes and simply rhymes, nothing could sound so musically stupid. But simple notes and rhymes united by meaning can be wonderful, especially if the expression of that meaning transforms these simple notes into subtle, responsive phrases. Music united with drama usually yields pathos, but we can lift that melodramatic union to the sublime with attentive enthusiasm. A trancelike fervor is probably most suitable to laudi, so loosen any constrictive clothing and get out your crying towel. Your face should glow and your eyes should glaze. Let the notes speak a story. “O anima” starts out innocently enough, perhaps even spiritedly, but the words soon blindly descend into an abyss, from which they try to climb up and out, only to find the wretched way leading down again, to a deceiving demon who wants our spirit for its own. The neutral theme returns with “Pero,” however, and gives hope in a long, pious cadence, which then turns suddenly to Christ, accepting Him, who wants to wonderfully uplift you. But woe, woe, woe is you without the love of God. “O anima” can start off without tongue, but it doesn’t seem as important as at “Oculus.” The two cadences of the first section can be very slow and sad, with livelier beginnings of their phrases. The next cadence, of the pious mind, can be broad and “pious.” Perhaps any itch to raise the “F” to a mi should be resisted here, since it would smack of a frivolous, sensuous, almost prurient mode. Keeping the “F” fa, as written, perhaps would have had more serious connotations. “Vota t’a Christo” can be as lively as a Shaker dance, while “cerca di fardite” can suggest the sweet rewards of faith with a slow, tender cadence. No need to rush through the “Ohime”s and the semibrevis can get a decrescendo by refingering. Although the final phrase delivers a rather negative message, I like taking it somewhat neutrally, to contrast the emotion of “Ohime” with a lofty, preaching admonition. Breath trilli could be made on sad cadences.

Cornelio Verdoncq (not Ian, who wrote very odd, awkward, sometimes charming counterpoint) wrote “Prest est mon mal” as a marvellous fuga a2, which can be also a duo with an alto, or a3 all together. This is also from Cantiones Suavissimae. Melody follows words, which form a series of contrasts, each part somewhat separate, the fuga is nicely put together, so the contrasting moods overlap very little, allowing much expression in each voice. The tempo seems to belong to the end of the century, with slow minims and semiminims only a little faster than their modern sense. The tempo can be completely elastic, moods expressed by speed, now slowly languishing, now brightly proclaiming total contentment. Perhaps there’s an actual scenario behind this song, but on the surface, it’s about the paradoxical bittersweet of a tender passion. It’s the recital of the swings of mood which characterize an ambivalent amour. The descending figure of “mon mal” can be slow and sad, to be partly relieved by far-off hope, and magnificently cured to a calm “sainement.” That can be followed by a lugubrious “longaissant,” slow in speed and attack. “Je me deteste,” two pairs of syllables, can be sharply, almost brutally tongued TRTR, but not rushed. “Et me vas blandissant” can flow caressingly, pleasingly with well-defined, but gentle articulation. Exaggerating the inequality between minims and semiminims, while keeping them light and easy, helps express pleasure. The final “blandissant” cadence can be very graceful, accelerating smoothly at the end to anticipate “Tout est a moy,” which is best imitated TRTR. But that emphasizes the descending third, and relates to the negative phrases “Je me…” and “Et rien…” more than TTTR. After the straightforward “Tout est a moy”s, “Et rien je ne possed” can be tortured into pairs of notes, TRTRTLR, perhaps making “sede” quicker (which conflicts with a space between “rien” and “je” in the other voice, but with “sede” on the same pitch, it seems worth it). Where “je me deteste” can be brutal, “et rine…” falls in dispairing, enervated notes lacking energy for vigorous complaint.

“Dime Robadora” is from T.H. Jensen’s facsimile of the 1556 “cancionero de Upsala,” which begins with duos and proceeds to three and four voice songs, two of which (including “Dime”) appear a2. The duos often seem more “authentic” in the sense of having a more special flavor and relation to the Spanish words. In many pieces (including “Dime”), voices seem to be added in a more northern style. The duos are mostly in 𝇋, while the trios and quartets are mostly in 𝇍 (those of duple time, that is), often with much the same range of note values. At the end are eight duos on the Modes, in quite different style, imitating Josquin in a peculiar meandering fashion all their own. There’s a temptation to see different layers of musical culture here, with the duos and some melodies representing the most pristine and primitive, perhaps going back into the 15th century. These melodies go well with the texts, but without a very expressive relation. Especially in the 𝇍3 songs, there’s a lot of syncopation, both from intrinsic melodic patterns, and from the accents of words being at odds to the meter. It’s easy to find the haughty, noble, melancholy Spanish stereotype here.

“Dime”’s original duo is between the upper voices, where a late 15th century tempo of 𝇍3 dotted brevis = slow pulse works well. But the third voice sounds frantically active at that tempo. A comfortable tempo for the bass gives quite a different feel to the top duet. It could be that the song’s character was changed in the addition of the third voice. Or it could be the job of that voice to somehow fit into the original duet. Again, we’re faced with minimizing jumpiness and we can use the same tricks, but with one valuable lesson in the figure

(also appearing colored in the bass. The coloration here only indicates imperfect breves; the accompanying black notes make groups of perfect breves. To otherwise notate such rhythms would need an elaborate system of dots of transportation). The first “que” in the bass seems to be an error and should be the end of “Robadora,” imitating the duet melody. This is fortunate, since it puts the unaccented “a” on the low minim, making it possible to play “Robadora”

(as in the duet, where it’s

with ambiguous text underlay following). This pattern can be smoothed out and apparently slowed down by making the first minim extra long and the second minim extra short and light, more as a pickup to the next semibrevis. Similarly at “olvidando,” the first ♩ can be broadened at the expense of the second. Elsewhere in the book, there are similar rhythms that lend themselves to the same treatment. TRLRL seems the best tonguing. The text underlay at “Que ganas” is clear for all voices, giving accents sometimes with the meter, sometimes not, but repeating the pattern

which applies to “que muera” also. At “Yo siempre,” the accent is on “siem,” leaving “yo” sounding like an upbeat.

Morley’s “Flora,” from his Canzonets of 1595 is a free translation and condensation of Anerio’s four-voice “Flori morir debbio,” or as Oliphant put it “from the Madrigals of Felice Anerio, which he has dished up wholesale in his Canzonets for two voices.” Oliphant, of the early 19th century Madrigal Society and organizer of the British Library’s music collection had a pithily negative opinion of Morley: “…neither his original compositions nor his adaptations from the Italian do him much credit. Some of the latter especially are execrable.” “Flora” has charms which seem to have escaped Oliphant’s notice. Morley’s translation for his duet has introduced more direct contrast between phrases, even compared to his translation for Anerio’s four-voice original. The melody is changed, too, to suit alterations of mood.

The initial phrase, all alone, can be as free as free can be, truly tormented (with perhaps a trillo on “ment,” and elsewhere, too), with a drawn-out cadence. A relatively light, contentedly skipping “and yet” can follow, only to return to the original “Flora” lament, and its dance-like response. Further tactless rhythmic liberty can lead to “Lo then”s, which truly catch our attention by their tentative rises, answering completely out of rhythm, then descending more definitely to point the way to a death more lusty, perhaps, than dusty. After complaining he may not get the pleasure of her treasure, our hero gets a rise from Flora as he dies, and the high “G” of “dying” can be climaxed with an “exclaim” (a sfp, which can be done easily on the XIIth note by leaking the thumb and forefinger). “Flora gentile and faire” can be calm and sweetly tender. Strictly speaking, the 3/2 semibrevis + minim should have the same time as a 𝇋 semibrevis. In any case, it should be unhurried and noble. “Alas hath staine me” can be a brutal contrast with hard, sharp rhythms and tonguings and trilli on cadential minims. The last “alas…” can be made light and mischievously gay, revealing the nature of this “death” which leaves a smile on the face. I prefer most of the ♩.♪♩ figures (“must I content,” “then I dye,” “And dying thus”) tongued TRL, but “that thy beauties” goes well TTR, lightly and pleasantly contrasting with the previous lack of pleasure.

Didactic use of duos. French articulation

It’s acknowledged that the 16th century tradition of duets had a pedagogical side but perhaps they were mostly used to learn style. Work in quartets and even trios tends to favor simplistic rigidity, while duets encourage a freer style which can be more easily applied to larger ensembles than developed there.

Hotteterre’s pointee tonguing can be used to imitate the syllabi patterns of ♩’s in many mid-16th century French songs. To take a song apart, examining the liaisons of syllables and then to find tongue-tying combinations of Ts, Rs, and Ls to imitate them down to the last detail may seem excessive. But to someone used to it, it could be second nature, no bother or fuss at all. It seems worth practicing, to be rejected, perhaps, from aesthetic or historical considerations, but not merely for its novelty or difficulty.

From Alamire’s facsimile of Attaignant’s 26 Chansons of 1534, here are two versions of “Martin menoit son porceau,” which in the Claudin version can be tongued

Martin menoit son porceau au marche avec Alix

Mar tin me noit son por ceau au mar che avec Alix

This same articulation is imitated at “pria Martin de faire le peche de ung sur lautre,” but “lautre” is a melisma, to be tongued LRLR, etc. “Et Martin luy” is TTRT(L) and the middle of “demande” is a melisma as is “y.” “Serre Martin” is TRTR and “nostre porceau mentraine” can go TRTRTRTRTRLRT(L) (the text underlay isn’t clear. Perhaps “mentraine” is repeated. If “Por” is a melisma, its ♩’s should go LRLR).

In Jannequin’s “Martin,” the same can be done. Most of the ♩’s are pointees. But “Et ce tient droit” and “Et Martin suche” go TTRT, while “sa” (fourth line) is melismatic. The articulation of “le pore eut peur et alix se cria” isn’t unambiguous, but could go TRTLRLRLTR. The second “mentraine” has just enough notes, if “serre…” repeats on the second “G” ♩ (eleventh note from the last), and this can show the articulation for the first and third “mentraine,” where “train” gets the extra notes. The while “serre-mentraine” could be tongued TRTRTRTRTRLRT(L).

Certon’s “Cest trop parle de Bach,” also from 26 Chansons, has several places needing notes pointees: “Pour vider potz, flacons, brocz, and Chopines” (“Pour vider” might be TTR instead of TRT); “Je boy a toy or ca haul sons le temps;” and “sans souspirer regardes faiz je mines.”

I suspect that although busy, these songs could go quite fast (up to semibrevis = 120/minutes) with care, delicacy and precision.

Gero’s French songs (here from Cantiones Suavissimae) have less patter and seem less suited to speed than the “Martin”s with their repeated notes (in Gero’s original 1541 Madrigali and Canzoni, there are a few humorous patter songs). Perhaps we could allow ourselves a comfortable semibrevis = 100/minutes. Taking all the ♩’s of the alto’s “Quande je boys” pointees seems appropriate. But dotted minim + semiminim at “me” should be TR. “Quand n’ay maille ne dernier” and “belle fille a mon coucher” might go TRLRLRTR. The bassus has melismas at “tourne” and “Je n’en.”

should be T lid’ll L or Trid’ll L.

For “Au Joly son du chansonet” I’d use TLRLTRTR or just TRTRTRTR. The melismas of “nuicte,” “souhaite” and “Amye” can be LRLR. “Va si te marie” is perhaps TTRTR(L)R (more so than “vas-y te marie” should be). There’s a melisma at the end of the third “Va si…”, so it could be tongued TTRTLRLR, etc. The second “Jamais je ne me marieray” also has extra notes at the end, so it can go TRTRTRTLRL. When such detail becomes facile, more regular tonguings sound dull in comparison.


Most of us are authentic in our approach to renaissance music, but to differing authorities. The authority can be the influences of our peers, or evidences of past practices. It can broaden into eclecticism or it can focus upon one teacher. But even the non-choice of unconsciously following what seems natural is a type of authenticity. We are all authentic in our gross interpretation of clefs and the pitches and rhythms of notes and the order of reading notes. But at a point we part company from past practices as suggested by historical evidence. Do these divergences come from aesthetic considerations, unconscious habits, social and personal ease, or what? Much historical evidence is ambiguous, but the range of style in today’s renaissance music is much narrower than the possibilities of historical authenticity. We can at least try various possibilities, but to do so often needs a considerable extending of techniques and attitudes, and much work and time.

The concept “music” can include instruments and playing style as well as composition. Much of the world’s music, African mbiras and New Guinean flutes, for example, seems to be an inseparable fusion of these elements. In other circumstances the strong character of instruments such as jews harps and bells (and to an extent, “Hammond” organs and steel bands) seems to dominate over the other qualities (perhaps this is less absolute strength of character than unusual timbre). And some national and personal styles thickly color whatever melodies they interpret, as do certain crooners and “gypsy” violinists. But in the renaissance was there a wide range of the relative importance of these elements? Did instrumentation and playing style change as much as with style of composition as they have in the recent history of popular music? In matters of taste and style that survive, clothing, architecture, art, and portrayal, there seems to have been almost as rapid a changeover from generation to generation as we enjoy now.

Bob Marvin
Eustis, Maine
Spring, 1985