Paul de la Ville, 2011
Available online with audio at: http://www.logarithmic.net/dance_music/
Music is correct when it sounds good.
What follows are some tools and material that might help you get started, or might suggest ways to proceed if you have something that is not quite working. Some are fairly standard, others are my own diagonal take on things.
Speech has a loosely rhythmic pattern of emphasis to it, called prosody. We interpret this as a hierachy, it lets us hear which are the key words and which words are modifiers attached as prefixes or suffixes, and this helps parse the sentence. The hierachy even descends down into words, which themselves have strong and weak syllables.
English tends to use prefixes more often than suffixes above word level. Linguists call this the Nuclear Stress Rule, apparently (Hayes, 1995, p.368). This rule applies to many but not all languages.
In music, we tighten this loose rhythm of speech into a much stricter rhythmic division of time.
I'm going to assume you are familiar with modern musical notation, and with time signatures and the within-bar empahsis structures they imply. Barlines as such weren't used consistently by Arbeau or Playford or their contemporaries. Vertical bars more often denoted sections. The idea that putting a dot after a note always made it a bit longer was also a new and exciting development, which the English called "pricking".
| A beat has a span of time before it, and a span of time after it. |
(We're going to be fighting against musical notation a bit here. Musical notation ties the beat to the span of time after it, which is what you care about when playing it on a melodic instrument, and especially in polyphony, but here we'll be looking mostly at the association between beats and the spans of time before them.)
| By dividing these spans of time into two or three sections by inserting new beats, we can elaborate this single beat. Just like in English, we will mostly be dividing the span of time before the beat. |  |
We can also divide the span of time after a beat to produce a suffix, but this seens to be much rarer in period music.
Further division is possible. We can make all kinds of strange divisions by twos and threes.
 |
| After dividing a span of time, you don't need to put a beat in every break point. But if you then divide a span of time that doesn't have a beat after it, it will feel like a suffix on the beat before it. If a span of time doesn't have a beat before or after it, don't divide it further. |  |
However, it's fine for the sounding of a beat to slosh over. This lets us construct medieval-style triple time rhythms.
 |
| It also lets us construct the pricked rhythms that then English are so fond of. |  |
So far we've considered short spans of times, but usually we'll be thinking about dividing phrases that extend over multiple bars.
The root of the prosodic hierachy will almost always be the last beat of the phrase. When we translate Renaissance music to modern musical notation, this will either be at the start or in the middle of the last bar of the phrase. The last bar will show you how the prosodic hierachy aligns with the bar-lines, whether the key beats are in the middle or at the start of bars.
You can help communicate the prosodic hierachy by emphasizing the more important beats in two ways:
(Conversely, you can produce an energetic rhythm by making sure the more important beats have a short sounding time.)
 | I'll mention one final trick to make an ending extra strong and funky, which is to decorate its immediate prefix beats in a way that feels suffixy. A sterotyped form of this is common in pavane and galliard music. There are also a few examples in the English Dancing Master, such as the endings of Stingo and Scotch Cap. |
Music isn't just a succession of notes, it has a structure, pieces that combine together and reach a conclusion. This is the basis from which we will build melodies.
( Dance movements follow this structure too, they're not simply a succession of actions, and you can start to see where you can put in extra actions to enhance the structure. You can put in some extra florishes to conclude more strongly, or you can weaken a beat by putting in some moves after it, making it a part of a larger phrase. For example two five-step galliards can be linked to form an eleven-step. )
| Arbeau begins and ends Orchesography with a marching beat. This corresponds to a left foot-fall then a right foot-fall. The right foot-fall is emphasized by fancy stuff before it and a long sounding time after it. |  |
 | The actual steps of his final dance, the buffens, have a complementary rhythm |
| Bransles can be straight crotchets, or a little more complicated. The gavotte and pavanes also draw from these. |  |
 | The morris dance has some extra twiddles. |
| The gay bransle is in 3/4. Arbeau also says that a good musician can freely convert music between duple and triple times. |  |
 | The galliard and tourdion have a characteristic rhythm. |
| Variations on the galliard rhythm. |  |
Arbeau sometimes explicitly simplified melodies to aid his exposition on dancing. It's not clear how much he has simplified other melodies. He is writing down a largely unwritten tradition, in which the line between composition and improvisation is entirely blurred.
 | Basic medieval. |
| Pricked rhythms. |  |
The above can be mixed together freely. If you stick to just the basic medieval half-bars, you'll obtain a more medieval sound. The dances from the Gresley manuscript, an earlier source for English dances, only use these.
 | Sarabands and jigs can have a very simple rhythm, and focus more on chord structures (we'll get to this later). |
The options in 4/4 time are much as in Arbeau's Orchesography, albeit with more use of semiquavers.
| Double bransle. Straightforward division by two. |  |
 | Single bransle. First division is by three, then by two. |
| The Scottish bransles have a larger scale division by three. |  |
The "mixed" or "cut" bransles are made up of pieces of more basic bransles. On the larger scale, they do not have a strict division of time. Where this happens the prosodic hierachy can be indicated by almost-repeats of more regular sections and by characteristic phrases.
 | The pavane demonstrates how rhythm and melody can separate two short phrases or string them together. |
Galliards are preferably done in pairs, first on the left, then on the right. I don't know of any further rules for them.
The prototypical Playford dance is in two sections, each section being repeated. There are however many variation on the number of sections, and whether they are repeated.
Sections are overwhelmingly of four or eight bars, indicating construction by repeated division by two, much like the double bransle.
I usually proceed by division, fleshing out the major beats first, then filling in details.
Repetition and development of patterns holds a piece together. You don't want to be introducing new ideas all the time. Pick an idea and use it in multiple places. It produces a good effect to progressively reveal an idea over the course of a piece, first presenting a version with less divisions.
Expect to try out a lot of different things before finding something good. Pick an environment that allows rapid experimentation. Possibly I am weird, but I dislike what-you-see-is-what-you get musical notation programs -- too slow, too awkward! I like a text editor and command-line based ABC-notation tools. But whatever works for you. Pencil, paper and a recorder perhaps.
Draw your targets after you've finished the shooting. Let it be what it wants to be.
Pick a skeleton and make divisions on it to produce a complete rhythm. You can use the half-bar rhythmic material above for inspiration.
Perhaps pick a one or two bar rhythm and use it several times.
A melody gains a first piece of coherency by being made up of a collection of pitches that are related.
To keep things simple let's say we are using the pitches F G A Bb C D E. This will be easy to play on a soprano recorder.
We might occasionally throw in a B natural or an F sharp for spice.
A melody is a journey. The journey ends on a particular note, called the tonic. Often the melody will also begin on the tonic, go somewhere else, then come back to it.
The mode that a melody is in refers to several things. Firstly, it defines a collection of pitches to be used. Secondly it defines the tonic, and the somewhere else that you go before returning to the tonic. The somewhere else is called the dominant.
With the pitch collection we are using, our modern major scale would have tonic F and dominant C. This has a happy or military sound. The modern minor scale would have tonic D and dominant A. The minor has a sad or melancholic sound.
We can obtain a different feel by using more exotic modes. For example Dorian would have tonic G and dominant D. Dorian is a popular mode for dance music. The whole tone step up to the tonic F -> G gives a stereotypical medieval coming-to-rest sound.
So far we've only seen modes where the dominant is a fifth above the tonic. These are called "authentic" modes. The "plagal" modes have the dominant a third above the tonic. For example hypodorian would have tonic G but dominant Bb.
Note that I'm simplifying a rather complex story here. A comedy of errors meant that medieval names for modes were different from the ancient Greek names, and the way they were used changed throughout period.
C G# E C F C# A F Bb F# D Bb Eb B G Eb G# E C G# C# A F C# F# D Bb F# B G Eb B E C G# E A F C# A D Bb F# D G Eb B G C G# E C
It is well known that falling by a perfect fifth produces a feeling of forward movement. The return from the dominant to the tonic in the authentic modes makes use of this.
The plagal modes use falling by a third, which also to me produces a feeling of forward movement.
Can we assign motion to all of the intervals? I think we can, and I've come to believe that the proper use of this is what makes certain melodies particularly compelling.
The progressive motion of perfect fifths and major thirds suggests to me the map of the territory shown to the right. Moving across this map, we fall by major thirds. Moving down it, we fall by perfect fifths. This layout is similar to the keys on a piano accordion.
From this map, we can see how other intervals can be composed from perfect fifths and major thirds (give or take an octave leap).
The octave, 12 semitones, frequency times 2. This is neutral.
The perfect fifth, 7 semitones, frequency times 3/2. Likes to fall.
Examples: f-Bb, c-F, g-c, d-G, a-d, e-A,
The perfect fourth, 5 semitones, frequency times 4/3. An octave minus a perfect fifth, therefore likes to rise.
Examples: F-Bb, C-F, G-c, D-G, A-d, E-A
The major third, 4 semitones, frequency times 5/4. Ding-dong, someone's at the door. Likes to fall.
Examples: E-C, A-F, d-Bb
The minor third, 3 semitones, frequency times 6/5. Most interesting and equivocal of intervals, composed of a perfect fifth and a major third in opposition. There are some schools of music instruction in which this is the very first interval that children are exposed to. Can you believe it? I am inclined to believe that the fifth wins, and therefore the minor third likes to fall.
Examples: G-E, c-A, F-D, Bb-G
The major sixth, 9 semitones, frequency times 5/3. An octave minus a minor third, therefore likes to rise.
Examples: G-e, C-A, F-d, Bb-g
The minor sixth, 8 semitones, frequency times 8/5. An octave minus a major third, therefore likes to rise.
Examples: E-c, A-f, D-Bb
The major second, 2 semitones, frequency times 9/8. The melodic workhorse. Two perfect fifths (minus an octave). Therefore, likes to fall.
Examples: c-Bb, A-G, G-F, E-D, D-C
The minor second, 1 semitone, frequency times 16/15. The powers of the fifth and third combine to make the minor second strongly like to rise. The classic ending interval.
Examples: E-F, A-Bb
The tritone, 6 semitones, is neutral and can be used for a disquieting or comic effect.
Example: E-Bb
With these intervals in mind, how can we use them?
Here are some ideas. Use them to structure a whole piece of music, and phrases and sub-phrases within it.
When going from a prefix to its root, simply use a progressive interval.
| The bransle of Poitou features a division by three, both prefixes progressing to the root. In fact it sketches a "chord", a beast we will meet in due course. |  |
We first introduce the tonic, then we go somewhere else against the progressive tendency, then we return to the tonic.
 |
If we bracket the root pitch, we produce a feeling of coming to rest which is very satisfactory. One prefix progresses to the root, another progresses from it. The root is not too hot, not too cold, just right.
 |
A common pattern is to construct a larger phrase from two nearly identical sub-phrases, the sub-phrases only differing in their ending. The ending of each sub-phrase places the sub-sub-phrases in a different light, with the first sub-phrase being less satisfying.
The end of the first will commonly proceed to the end of the second.
See for examples the first Scottish bransle, and the first section of Grimstock. This is also used in saltarellos, estampies, and rottas.
 |
 |
Arbeau's music is often well suited to playing on a three-hole pipe. Certain intervals are particularly easy to play on a three hole pipe, just by blowing harder without changing fingering. These are the fifths F-c G-d A-e Bb-f, and perfect fourths c-f, d-g, e-a.
Choose a tonic note.
Our melodic skeleton will consist of:
1. The tonic
2. A pitch that proceeds from the tonic
3. A pitch the proceeds to the tonic
4. The tonic again
Play the skeleton to ensure it produces a satisfying feeling.
Divide this skeleton further, inserting prefixes that progress to their parent note where possible while maintaining a flowing melody line.
The melody line should have an overall rising and falling pattern.
A body of musical theory based on chords developed in the late 16th and early 17th century, and continues to dominate music to this day.
A chord is a triad of notes, a root note, a note a third above the root, and a note a perfect fifth above the root. (We can also shift any of these notes by an octave, for example to move the root to a perfect fourth above the third note. This is called inversion.)
There is a hidden structure of chords, associated with bars, or half bars, or sometimes shorter, with the actual notes of the music being drawn from these hidden chords.
If you like, this is about the progressive tendencies not of single pitches but of triads of pitches.
The rise of chord-based music saw the death of the modes, and the dominance of the major and minor scales. Early editions of Playford's book on music theory, An Introduction to the Skill of Musick, describe the Greek modes. Later editions say that only the major and minor are required (Herissone, 2000). Much traditional English folk music has remained modal, however (Williamson, 1977).
| If we name each chord by its root note, chord structures will look much like the melodic structures we've just been making. For example: F B C F. We introduce the key with the initial F chord (FAC), which progresses to the B chord (BDF) by falling a fifth, we then have a C chord (CEG) which is a fifth above the final F chord. The final F chord is bracketed by the B and C chords, producing a feeling of coming to rest. This is a very satisfying structure, and is terribly over-used. |  |
While the music in the English Dancing Master is monophonic, it has a chordal sensibility. Chords are often sketched out one note at a time within a bar or half-bar. Just sketching out the root and the third is enough to give a chordal feeling.
In terms of prosodic structure, these chord sketches will often feel like suffixes on the emphasized beat of the bar or half-bar.
 |
 |
If you want a flowing melody line without so many leaps of a third, you can interpolate between the chord notes on weak beats (short, deep in the prosodic hierarchy).
| The pricked rhythm is particularly good for this. |  |
 | The "medieval" rhythms are pretty good at this too, but sound less medieval if used in this way. |
Two melody lines played simultaneously is counterpoint. We may wish to invent two wholly new melody lines, or we may wish to add a second melody line to an existing melody. More simply, we may wish to compose a melody to be accompanied by a drone.
You should be able to guess how I approach this: look for a progressive tendency. Counterpoint means we have intervals not just between notes separated in time, but intervals between notes played simultaneously. Such intervals have a level of consonance or dissonance. The obvious progressive tendency is to move from dissonance to consonance. (This is in addition to all the rules we have already seen about rhythm and melodic progression.)
My friend Nathan Hurst thinks that the level of dissonance is simply the period of the combined waveform, and I think this rule works quite well. For equal values of the lower note, we get a hierachy from most consonant to most dissonant as follows:
1 Unison, octave 2 Perfect fifth 3 Perfect fourth, major sixth 4 Major third 5 Minor third, minor sixth 8 Major second, major seventh 9 Minor seventh 15 Minor second
( The numbers are the denominators of the frequency ratios we saw earlier. )
Note that the frequency of the lower note is also important. The lower the note is the more dissonant it is, by this system. For example, if you are writing below an existing melody and need to make your lower note quite low in order to produce a major sixth, this of itself reduces the consonance somewhat.
Oddly, music theorists of the renaissance considered the perfect fourth to be a dissonance. This is strange because it sounds almost identical to a perfect fifth. Note that it truly is more dissonant than the perfect fifth, and can function as the dissonance in a progression from dissonance to consonance where the consonance if a perfect fifth or octave. You certainly can't treat a perfect fourth the same as a perfect fifth.
Other than this, our table is in accord with Renaissance music theory. For example, Thomas Morley (1771, p.76) in his book "A Plain and Easy Introduction to Practical Music" first published in 1597, divides the intervals into perfect concords, imperfect concords, and discords:
Consonance is not good per se, nor is dissonance bad. Counterpoint requires the use of a range of levels of consonance.
As with melody, we are always thinking about progress within the context of the prosodic hierachy, not just of beats that are next to each other.
It's not always going to be possible to give a progressive feeling to consonance levels while also keeping melodic concerns in mind, and vice versa. It's a balancing act.
The patterns discussed with regard to melody (progressing, going and returning, above and below, variant endings) can also be applied to the changing level of consonance. We'll generally start and end a phrase on a consonant interval, usually a unison, fifth, or octave.
Counterpoint is a large topic. There are plenty of texts out there to study, which will make perfectly clear to you how inadequate your understanding is when compared to Palestrina or Bach. More productive might be to pick apart your favourite counterpoint music and work out what makes it tick. Look for repetitions and elaborations of rhythms, melodic fragments, and structures of consonance. Progressively remove divisions to examine the larger scale structure.
With these ideas about rhythm and melody in mind, we can see some of the broad sweep of musical development. The change from triple to duple time during the renaissance, then increased diversity as we move into the common practice period (but with 4/4 time remaining popular). The rise of suffixy prosody. The switch from monophony and counterpoint to chord-based functional harmony. The corresponding loss of modal music and the dominance of the major and minor keys.
We can also see the structure and language of existing music. If we encounter medieval music where the rhythm is uncertain, such as troubadour songs or the Cantigas de Santa Maria, we can use understanding of rhythm and melodic intervals, and the grammatical structure of the text and within-word syllable stress rules, to assign a rhythm that is pleasing and perhaps correct. And finally, we can construct our own melodies in appropriate style.
Finally, I'd like to suggest a worthwhile project would be to write counterpoint lines for period dance melodies. We have many modern three and four part harmonic arrangements of period dance music, but very few counterpoint arrangements.
Hayes, B. (1995) Metrical Stress Theory. The University of Chicago Press.
Herissone, R. (2000) Music Theory in Sevententh-Century England. Oxford University Press.
Morley, T. (1771) A Plain and Easy Introduction to Practical Music. (First published 1597)
Williamson, R. (1977) The Penny Whistle Book. Oak Publications, NY.