Note values exceeding time signature - Voices in Bach

[jacobrudduck]

Hi all. I'm struggling to get my head around this problem. My reading of rhythms has always been a bit of a weak spot for me (actually, that is an understatement).

Could somebody please explain why notes values written in the Goldberg Variations exceed the time signature? Let's look at the aria:



The first two measures in the bass have 6 beats written in, if I am not mistaken. This is double the time signature. Eh?

Wait, so this is because there are two voices going on. Those quarter rests above the notes, they signal that a new voice has begun. So...every one of those 'voices' is separate from the rest of the measure? Almost like a measure of their own? is this correct?

Variation 13:



Here we have an eighth note REST above the first dotted G. What difference does this make? What is that sixteenth note rest doing there? When I play the B sixteenth note do I count that as 2, or is the D half note on 2?

Could somebody please instruct me on how to count the bass measure. I know how to play it, but for the life of me I can't work out where the beats are, because they seem to exceed the time signature. Am I just not grasping basic rhythms or is there something else I am ignorant of?
I'll be very grateful to have some help on this, I am quite confuzled.

[iansinclair]

In the first example, your have three voices active in the first six measures.  The uppermost has very little to say -- two quarter rests and a quarter note.  The middle has a bit more to do; a quarter rest and a half note.  The lower voice has a dotted half.  Each voice has three beats.  In the seventh measure, one of those three voices drops (a coffee break, maybe?) and the upper of the two should have an eighth rest followed by the figure.

The second example has just three voiced; two in the bass stave.  The upper of those two has, in the first measure, an eighth rest followed by a sixteenth rest followed by the sixteenth note -- the whole making one beat (the sixteenth note is almost like a "pickup" to the half note following in that part, making three beats in total) while the dotted G and the three eighth notes following also make three beats.

And yes, your supposition that this is written that way because there are multiple voices going on is correct and each of those voices is separate from the others -- this is polyphony, after all!

[cliffy]

Hi Jacob,

I'm afraid you are mistaken, as there are only three beats written into your example from the Aria. The simple answer is that the Goldberg Variations is a polyphonic work, as is much of Bach's output and Baroque music in general. Think of it as though there were three different voices, each voice's note coming when it's written and being held for its written value.

For the Aria Left Hand:
Measure 1:
Beat 1: Play the G (this will be held for its full value, 3 beats)
Beat 2: While still holding the G, play the B (this will also be held for its full value)
Beat 3: While still holding the G and B, play the D
End of measure: Release notes

Measure 2:
Beat 1: Play the Fsharp
Beat 2: While still holding the Fsharp, play the A
Beat 3: While still holding the Fsharp and A, play the D
Ect.

For the variation:
Beat 1 (first sixteenth) Play the G
Beat 1 (second sixteenth): Keep holding the G
Beat 1 (third sixteenth): Keep holding the G
Beat 1 (fourth sixteenth): While still holding the G, play the B
Beat 2 (first eighth) : While still holding the G, release the B and play the D
Beat 2 (second eighth):While still holding the D, release the G and play it again
Beat 3 (first eighth): While still holding the D, release the G and play the Fsharp
Beat 3 (Second eighth): While still holding the D, release the Fsharp and play the E

I would have made pictures, but I don't yet have an editor for them on this phone. I hope this helps.

All the Best,
Cliffy

[jacobrudduck]

Wow, thank you for your replies. This has opened  an entire can of worms, I didn't know that polyphony music was actually written in any special way like this.

But I am still lost, how do I distinguish between the different voices? How am I reading them, down from up, or left to right? What groups them together? How do I know what is the uppermost, does that mean the highest on the lines? Do you see what I am trying to ask?

[iansinclair]

Wow, thank you for your replies. This has opened  an entire can of worms, I didn't know that polyphony music was actually written in any special way like this.

But I am still lost, how do I distinguish between the different voices? How am I reading them, down from up, or left to right? What groups them together? How do I know what is the uppermost, does that mean the highest on the lines? Do you see what I am trying to ask?

Fortunately for the sake of this discussion, both examples which you picked were notated properly.  This isn't always true... if the notation is done correctly, any given voice will always have the stems of the notes pointing either up or down, but never both from time to time.  You can compare the position of the stems of the notes on the upper stave of either of your examples -- which point either up or down to more or less fit on the stave -- with those on the lower stave, where one voice is always up and the other always down.  The upper stave has only one voice, so no confusion is possible.  The lower stave has two, so they have to be kept separate.

Unfortunately, not all publishers do this, to the consternation and dismay of people trying to follow things through!

As to which is uppermost, sometimes it is significant; more often it really isn't, and lines will be seen to move relative to each other, sometimes higher, sometimes lower.  There are some specific forms of early polyphony -- fourteenth and fifteenth century -- where this isn't true, and one line will form a ground or cantus firmus and always be either highest or lowest -- but these are, nowadays, pretty esoteric...

One might add, though, that from the standpoint of actually having to play one of these things on a piano or harpsichord, some rather interesting fingering or other technical problems can arise when one is trying to differentiate one specific voice from the others -- whether by volume or by subtle changes in the degree of legato -- and the relative position keeps changing!

But cheer up -- it's even worse on an organ, where it is a little unusual, but by no means unheard of, to have one of the voices on, say, the lower stave, played on one manual, and another on a different manual -- which leads to all sorts of interesting contortions...

[jacobrudduck]

Goodness, my brain. Sorry if I sound stupid, this is obviously way out of my depth. But in simple terms, how do I know which voice contains which notes? How do you look at that bass measuure in the variation and work out 'Okay, so voice 1 is playing these notes and voice 2 is playing these'?. I want to trust that I am holding the right notes for the correct duration.

[cliffy]

Well, polyphony vs monophony doesn't change the fact that each note lasts for its full written value. You usually won't have more than two voices on each staff, and they usually will have one voice with the stems up and the other down. In the end, it comes down to your experience (the more you read music the easier it gets), the edition (a good edition always makes sure the separate voices are easily distinguishable), and your instincts (which are developed through experience and listening to counterpoint). Try reading through, say, Bach's three part inventions, not sight-reading, but picking one voice and saying aloud all the notes it has from the first measure to the last. After that, do the other voices. After a couple of times doing this, you will be both a better reader and have an easy time disguingishing voices.

Cliffy

[nyiregyhazi]

Sorry if this is a little blunt, but you need to use the most basic common sense in order to understand this properly. When I play this type of writing, I scarcely even register any rests of this kind, if at all- as  they don't tell you anything that you cannot derive by properly processing other pieces of important information. It saves time to ignore them. Start by looking at which notes are LINED UP to obvious beats in the other hand (and if locations of beats in the other hand are not yet obvious, they really do need to be, to understand this properly). Such a simple procedure should make most of these things painfully obvious. Likewise, in the second example, you need to use common sense. Start with the obvious alignments and then add up the duration of the notes that end the beat. logic ought to tell you that by subtracting those from the beat, you will deduce how long into that beat to wait before starting those notes (ie after three quarters of the beat, here, then you play the semiquaver on the fourth quarter)- meaning that the rests are little more than a tautology that provides no unique information beyond what can be fully derived from understanding notes in association with alignments. Rests that tell you when to release notes are significant and must be noted and taken precisely- but a fluent reader scarcely needs to notice the type seen here. Comparison to the other hand tells you literally all the same information in a very obvious form (and gets the brain organising every beat properly).

Stop trying to add up numbers for the whole bar and start seeing where in the bar every beat lies. Then use logic for the internal details. It's extremely simple when you have a pragmatic rather than a pedantic viewpoint based on adding up too many numbers over whole bars. If you can't spot beats, trying to do maths across a whole bar won't compensate. Notation is specifically designed to make new beats easy to spot- hence the tie in that example. Effortless reading is based on use of context- and is not possible until you see every beat location first and the fine details second.

[brogers70]

Sorry if this is a little blunt, .....

You're not sorry in the least; bluntness is your modus operandi. But with a little practice you could convey the same, helpful information in fewer words and with no bluntness at all.

[nyiregyhazi]

You're not sorry in the least; bluntness is your modus operandi. But with a little practice you could convey the same, helpful information in fewer words and with no bluntness at all.

bluntness is not my norm at all. however, it's necessary here as the poster is putting a kind of destructive humility and desire to bow down to what a rulebook only appears to be saying before his own common sense. I'm sure he already noticed that the notes are aligned and I'm not saying he has no sense. He just needs to TRUST his own intelligence and put that first and the rulebook second. you don't develop good reading skills without doing so. The overwhelming majority of sound reading comes from use of contextual clues- which actually make it far easier than trying to start with the reams of incidental rules on how to notate different parts. Remember that you only need to know all the advanced rules in order to NOTATE something. In order to merely READ that notation, it's a matter of knowing only the most basic rules and then using context in reference to individual beat locations and alignments.


Many composers don't even bother with the type of rests written there- because the meaning will be clear to anyone who observes alignments and is wise enough to trust their sense over some unnecessary desire to be trying to add up every note that they see, in reference to a rulebook. giving him the more advanced rules will only confuse all the more. The poster need merely throw out that mindset of adding up all the notes (rather than pinpointing beat locations) and he'll soon figure out how it really works for himself.

[tdawe]

bluntness is not my norm at all. however, it's necessary here as the poster is putting a kind of destructive humility and desire to bow down to what a rulebook only appears to be saying before his own common sense. I'm sure he already noticed that the notes are aligned and I'm not saying he has no sense. He just needs to TRUST his own intelligence and put that first and the rulebook second. you don't develop good reading skills without doing so. The overwhelming majority of sound reading comes from use of contextual clues- which actually make it far easier than trying to start with the reams of incidental rules on how to notate different parts. Remember that you only need to know all the advanced rules in order to NOTATE something. In order to merely READ that notation, it's a matter of knowing only the most basic rules and then using context in reference to individual beat locations and alignments.


Many composers don't even bother with the type of rests written there- because the meaning will be clear to anyone who observes alignments and is wise enough to trust their sense over some unnecessary desire to be trying to add up every note that they see, in reference to a rulebook. giving him the more advanced rules will only confuse all the more. The poster need merely throw out that mindset of adding up all the notes (rather than pinpointing beat locations) and he'll soon figure out how it really works for himself.

This is wrong... that he can't work out which voice is which is a serious problem. He's doing the right thing by addressing them now. If you can't identify the separate voices in a polyphonic work then you have no hope of playing it correctly.

[j_menz]

Goodness, my brain. Sorry if I sound stupid, this is obviously way out of my depth. But in simple terms, how do I know which voice contains which notes? How do you look at that bass measure in the variation and work out 'Okay, so voice 1 is playing these notes and voice 2 is playing these'?. I want to trust that I am holding the right notes for the correct duration.

It's not quite as complicated as you might think, and the edition you have helps.

Rule 1: The time signature is correct.

Rule 2: Notes (and rests) are held for their full value.

Rule 3: Don't think of chords, think of single note melody lines, of which there may be several.

Example 1 - First bass bar of the Aria:

Time signature is 3/4, so there are three crotchet beats to the bar.  The first note (G) {Voice 1} is a dotted minim, so is held for 3 crotchet beats (ie, the whole bar). But, there is a crotchet rest above it. That means that something has to happen after one crotchet - either another rest or a note. The fact that it is above the G indicates that whatever is going to happen is going to happen above the G.

So, after that one crotchet, you'll see that something does in fact happen - you get another note, a B minim {voice 2 - remember, no chords}, to be held for two crotchet beats. You're already one beat into the three allotted for the bar, so this also is held to the end of the bar.

But wait - there's another crotchet rest above it.  Apply the same logic, and lo and behold there is a D crotchet {voice 3}, which is held for the (one) remaining beat. So, on that last beat you are actually holding all 3 notes at once.

Example 2 - Bass line of Bar 1 of Variation 13:

Step 1 - Time signature - 3/4 still, so three crotchet beats.

First up, a dotted crotchet G (one and a half crotchet beats). Hold for that long (and no longer). But, there's a quaver rest above it, so look out for something after a half a crotchet beat. What happens then? Not a note this time, but another rest - for a semiquaver. So wait that 1/4 of a crotchet beat for something new, and there it is - a B semiquaver. At the end of that, we've waited the length of the two rests and the duration of the semiquaver, so we're up to the start of the second crotchet beat. The G in the bass isn't due to run out for another half a crotchet, but the B has run its course and something is needed to replace it. And what do we see - a minim D. That minim will last the rest of the bar. But in another half quaver beat, the bottom G will expire, so something needs to replace it, and so we get a series of three quavers - the G F# and E. One dotted crotchet plus three quavers = 3 crotchet beats and so we are at the end of the bar.  There were only two voices in (the bass of) this bar.

Homework:

Have a try at doing the same analysis on the bass of bars 2, 4 and 7 of the aria.

[j_menz]

bluntness is not my norm at all. however, it's necessary here as the poster is putting a kind of destructive humility and desire to bow down to what a rulebook only appears to be saying before his own common sense. I'm sure he already noticed that the notes are aligned and I'm not saying he has no sense. He just needs to TRUST his own intelligence and put that first and the rulebook second. you don't develop good reading skills without doing so. The overwhelming majority of sound reading comes from use of contextual clues- which actually make it far easier than trying to start with the reams of incidental rules on how to notate different parts. Remember that you only need to know all the advanced rules in order to NOTATE something. In order to merely READ that notation, it's a matter of knowing only the most basic rules and then using context in reference to individual beat locations and alignments.


Many composers don't even bother with the type of rests written there- because the meaning will be clear to anyone who observes alignments and is wise enough to trust their sense over some unnecessary desire to be trying to add up every note that they see, in reference to a rulebook. giving him the more advanced rules will only confuse all the more. The poster need merely throw out that mindset of adding up all the notes (rather than pinpointing beat locations) and he'll soon figure out how it really works for himself.

There are two things I am cautious of regarding your method when applied to people starting out with reading polyphony.


Firstly, some assume that the beats will be evenly distributed physically across the bar, and this (as you will be aware) is often not the case.

Secondly, to read polyphony, one needs to read primarily horizontally, which is often rather different than the more usual vertical reading involved in piano music. Whilst your method doesn't preclude this, it also does not encourage it - and most people need all the encouragement they can get in this regard.

If the OP wasn't susceptible to both these issues, I doubt he'd have asked the question he did in the first place.

[nyiregyhazi]

There are two things I am cautious of regarding your method when applied to people starting out with reading polyphony.


Firstly, some assume that the beats will be evenly distributed physically across the bar, and this (as you will be aware) is often not the case.

Secondly, to read polyphony, one needs to read primarily horizontally, which is often rather different than the more usual vertical reading involved in piano music. Whilst your method doesn't preclude this, it also does not encourage it - and most people need all the encouragement they can get in this regard.

If the OP wasn't susceptible to both these issues, I doubt he'd have asked the question he did in the first place.

Being aware of vertical reference points doesn't preclude looking at music horizontally too. To use context you need both. I've noticed that it's common among younger students to miss these reference points and land the wrong notes together. Even when I practice hands separately, I'm looking at the other hand to at least some degree, for contextually relevant information. These shared arrival points need to be spotted without fail for any hope of fluency. If anything, if I'd actually place MORE importance on perception of alignments in counterpoint- where horizontal motion of one part easily gets carried away and causes the sync with another part to get missed. The alignment of beats is only the first step of the integrated thinking that counterpoint depends upon- but it's an absolutely vital step, both in reading and execution.

When it comes to understanding of rhythm, there should be not the slightest doubt in the mindin about the vertical location of each beat of the bar. I didn't either suggest or imply that horizontal distances give accurate indication of this. I said that finding obvious beats in one hand gives the location of beats in the other too (failing any outrageous misprint). This process comes first in practically oriented reading- which is not done by adding up numbers without considerations of context. Merely by appreciating this, the poster will figure it all for himself. Giving him the rules for notation of voices will probably baffle him as much as advanced rules for grammar would baffle someone who simply wants to be able to read english. That's why I told him straight up that he needs to trust his intelligence first and look for textbook rules second. Textbooks are technically accurate, but they neglect the methods of pragmatic reading and are rarely useful for conveying the real tricks.

[nyiregyhazi]

Just one additional point- foundations are always grounded in vertical reference points even in the most horizontal playing. something as simple as a strong beat is a purely vertical event. It doesn't carry into louder notes after but simply occurs as a  vertical event in a  horizontal chain. Would you expect someone to be able to organise a chain of rapid fire notes without these vertical reference points of stronger notes? Even the most horizontal singular line is organised around a foundation of vertical references. When a poster cannot see that this comes first in musical writing, it's clear that he's not set the foundation that he needs in order to progress to truly horizontal thought.

Recently I had many students playing bach's b flat major prelude and fugue from book one. A very common problem was fingers going out of sync. To fix that, they had to use the method of playing a beat plus a note , then pause. The reason this works is that it puts in place vertical reference points- where hands are being linked together. When you pass through, there's limited chance to notice and feel the association between hands. When they stop they're often surprised by associations that they never perceived before- because the hands were TOO independent and only used to running on their own steam. Before the illusion of independence can begin, you must set reference moments. Even the finest horizontal playing functions around a series of cast iron anchor points in which voices are contextualised to each other. It's impossible to do without this style of thinking in horizontal playing. The role of this in reading is merely the tip of the iceberg, when it comes to how important this is in execution.

[nyiregyhazi]

This is wrong... that he can't work out which voice is which is a serious problem. He's doing the right thing by addressing them now. If you can't identify the separate voices in a polyphonic work then you have no hope of playing it correctly.

you missed my point. my method TELLS you which voice is which and absolutely definitively. once you get to something like Bach's five voice fugues, it's actually the only approach that works. stems can only go up or down- which means that they do not provide enough information if three voices are on a stave. At this point, the rules about stem directions turn out to have been nothing but an incidental feature. if you placed all your understanding in this minor feature, you'll be absolutely lost when the stem directions are no longer fit to differentiate between voices. it's really a big mistake to base your understanding around such a superficial and limited feature.


 I scarcely pay attention to directions of stems because at best they tell you something you can figure out with reference to pieces of information which you ought to be processing anyway. And at worst they fail to divulge the necessary information altogether. Except in rare exceptions, note lengths and alignments tell you literally everything you need to know about rhythm and thus voice leading. These two features are what matter and everything else can be calculated from it. If you use this method, the stems can go in any direction you like but the voice leading will remain unmistakable. Also, when you think in voices, you don't always strictly need to count long notes - as long as you are clear that you MUST hold it until you either see it being replaced by another note or by a rest. I'm not saying to always think this way, but it's often far easier than trying to independently count out the length of 4 or 5 independent voices. if you can understand when a note finishes based on context, counterpoint is far less difficult to grasp. in the second example given there, for example, I wouldn't even think about the length of dotted crotchet in the lower voice. If I merely note that it's long then context tells me when it moves to the next note and the length sorts itself out. I'd look to length of the first note merely for confirmation- not as the primary piece of information from which I calculate when the next note is due. it's a process of working backwards AFTER pinpointing beat locations for the whole bar- not simplistically going from left to right a note at a time and adding them up. The latter is what textbooks imply, but it's not an intelligent or practical approach unless you start using your own logical processes.

[jacobrudduck]

Thank you for the discussion! I am following this thread and digesting all the answers.

It's not quite as complicated as you might think, and the edition you have helps...

Thank you so much, that was EXACTLY the explanation I was looking for! Very easy to understand what those rests indicate, and how note values work in polyphony.

Also, when you think in voices, you don't always strictly need to count long notes - as long as you are clear that you MUST hold it until you either see it being replaced by another note or by a rest. I'm not saying to always think this way, but it's often far easier than trying to independently count out the length of 4 or 5 independent voices. if you can understand when a note finishes based on context, counterpoint is far less difficult to grasp. in the second example given there, for example, I wouldn't even think about the length of dotted crotchet in the lower voice. If I merely note that it's long then context tells me when it moves to the next note and the length sorts itself out. I'd look to length of the first note merely for confirmation- not as the primary piece of information from which I calculate when the next note is due. it's a process of working backwards AFTER pinpointing beat locations for the whole bar- not simplistically going from left to right a note at a time and adding them up. The latter is what textbooks imply, but it's not an intelligent or practical approach unless you start using your own logical processes.

This was too the most helpful advice. The bit in bold is really what I relate to here. I posted this thread because I am re-learning this piece a year later. I believe my reading of note values the first time was slightly inaccurate and so I completely lost confidence in whether I was actually doing it right. I went back to the sheets and started trying to add it all up and find where to count, but as you say this has just caused me a lot of problems.

nyiregyhazi, your 'blunt' comment.    Absolutely, no offense! I totally agree with your mentality: trust your instincts and use your musical intelligence. With the method j-menz has given me I feel I can combine the two now. My confidence was lacking (I am almost entirely self taught in reading music, so i never have any reassurance that I am doing it correctly) and I felt like I must have been doing *something* wrong, and wasn't understanding basic rhythm counting. I didn't stop to consider that it was a polyphonic work and had no idea it was written differently. Google has failed to provide me with information on how to read polyphony so I felt totally clueless. The three rules j_menz gave me certainly pointed me in the right direction.

Thanks guys!

[nystul]

This applies to more than just polyphonic works.  It is actually pretty common.  Any time you have multiple voices represented on the same staff that do things at different times, there is a possibility for this kind of confusion. 

[j_menz]

Any time you have multiple voices represented on the same staff that do things at different times

.... you have a polyphonic work.

[j_menz]

Just one additional point- foundations are always grounded in vertical reference points even in the most horizontal playing. something as simple as a strong beat is a purely vertical event. .

Every bar has vertical reference points that represent the beat. Each voice is calibrated to these reference points. I don't see the need to further calibrate them against each other, as if they are correct with the beat, they are correct with each other. If your students are getting out of alignment, they are playing one (or more) of the voices out of alignment with the beat - not a case of too much independence, but merely that one or more voices is rhythmically incorrect.

[nyiregyhazi]

Every bar has vertical reference points that represent the beat. Each voice is calibrated to these reference points. I don't see the need to further calibrate them against each other, as if they are correct with the beat, they are correct with each other. If your students are getting out of alignment, they are playing one (or more) of the voices out of alignment with the beat - not a case of too much independence, but merely that one or more voices is rhythmically incorrect.

Your logic doesn't work for a pianist. For five separate players, sure. All you have is beats to align notes to and the rest does itself. But a pianist cannot simultaneously run five lines as independent physical coordinations. That logic simply won't hold up. A pianist needs to be acutely aware of relationships- in order to integrate many things into a global coordination. It's a very different thing to the sum of its parts. The illusion of five independent lines owes a huge amount to how a pianist sees something in a global fashion. For a start, you cannot find a good fingering by practising voices separately, but only by looking at the relationship. Voices can be understood rhythmically on their own, but it doesn't mean that you're truly running five independent coordinations that only meet because they happen to share beats. Music is always vertical and horizontal- by the very nature of how we physically execute it. Only with multi track recording techniques can your logic stand up.

Also I'm not only talking of misreading. In the fugue it was a matter of subtle ongoing instability. When you truly settle into secure balance on regular reference points, you acquire the foundation that permits horizontal motion without that sense of skimming across notes vaguely. It also improves awareness of harmony to stop and listen to moments vertically, before going back to the bigger chain. When you see vertical points clearly, you can pass through them without interfering with horizontal movement, yet you still gain a sense of poise and certainty from seeing those relationships as you pass through. The students were were unstable because they'd never taken the time to to truly observe those relationships before skidding on forwards to the next bit. When observing relationships, you create better awareness of harmonic construction and set points of both mental and physical clarity. When this this hasn't been done, beats will never sound clearly or precisely as the organisation isn't there to to allow that.

[j_menz]

Your logic doesn't work for a pianist. For five separate players, sure. All you have is beats to align notes to and the rest does itself. But a pianist cannot simultaneously run five lines as independent physical coordinations. That logic simply won't hold up. A pianist needs to be acutely aware of relationships- in order to integrate many things into a global coordination. It's a very different thing to the sum of its parts. The illusion of five independent lines owes a huge amount to how a pianist sees something in a global fashion. For a start, you cannot find a good fingering by practising voices separately, but only by looking at the relationship. Voices can be understood rhythmically on their own, but it doesn't mean that you're truly running five independent coordinations that only meet because they happen to share beats. Music is always vertical and horizontal.

Whether you are one pianist, five pianists or a string quintet, five voices are five voices. Independent, but related - not just harmonically and temporally, but also in more complex fashions (questions, answers, inversions, retrogrades, double time, half time etc). In order to play it, you have to play each of the five voices on it's own merit, bearing in mind it's relationship to the other voices.

The job of the hands is to give a physical way of doing that, wherein lies the problems of co-ordination and fingering. These, however, are merely means to the end. It's not a case of five independent physical co-ordinations, It's one physical co-ordination giving effect to five independent voices.

And if you cannot mentally co-ordinate five independent voices, no amount of merely physical coordination is going to produce an acceptable result.

[nystul]

.... you have a polyphonic work.

It could be considered homophony if there is one primary melody and the other voices are supporting it, which applies to quite a lot of music.

[j_menz]

It could be considered homophony if there is one primary melody and the other voices are supporting it, which applies to quite a lot of music.

No, it's still polyphony. Don't make the assumption that if it ain't baroque it ain't polyphony.  There's a lot of it lurking out there.

[nyiregyhazi]

Whether you are one pianist, five pianists or a string quintet, five voices are five voices. Independent, but related - not just harmonically and temporally, but also in more complex fashions (questions, answers, inversions, retrogrades, double time, half time etc). In order to play it, you have to play each of the five voices on it's own merit, bearing in mind it's relationship to the other voices.
y
The job of the hands is to give a physical way of doing that, wherein lies the problems of co-ordination and fingering. These, however, are merely means to the end. It's not a case of five independent physical co-ordinations, It's one physical co-ordination giving effect to five independent voices.
2
And if you cannot mentally co-ordinate five independent voices, no amount of merely physical coordination is going to produce an acceptable result.

You're only coming at this from one end and one- which is that popularly stressed to pianists. but that should never contradict the necessity of the vertical viewpoints, which you argued against. there's no necessity to choose. the argument you are giving becomes deeply antimusical if taken to the extreme of not associating parts vertically, not to mention utterly impractical, I'll illustrate exactly why.


picture a pianist given a five voice fugue. they have a score which only shows a single part with no cues or vertical references. they can play it as many times as they like. then the same with other parts. how many pianists could then sit down an execute the fugue by memory? Glenn gould perhaps, but few others. a pianist who finds a fingering that works has thought about associations or he could never have done that- so even then it was done internally, even if the hard way. he'd probably have much preferred to see all the parts alongside each other for context. The better the musician, the more attention he will place in other parts even when playing through a single part in isolation. Other than convenience of execution there's a VITAL musical reason. a chromaticism such as a minor ninth exists solely in context. you need to know which two parts create it and also which of those two parts is the non harmonic note. that will come not merely from those two but from looking at all parts- ie. it's inherently interdependent and never truly independent altogether. by exaggerating the popular idealised argument so far, you are actually arguing for something spectacularly unmusical- not for anything profound, as you seem to believe. profundity comes from seeing the big picture, not from stressing independence at the expense of how the musical whole stems from appreciation of relationships.


The rationale seems to be that pianists are bad at doing different layers, therefore let's only argue for the independence side of things (despite the fact that a string quartet that learns parts separately and then merely plays metronomically and without context in the whole would be a very poor quartet). but nobody has to choose a one sided view. What I saw from each of three students in that fugue was lack of rhythmic stability and grounding due to lack of clear vertical reference points to feel horizontal movement between. There's more than one way to go wrong. with versatile balanced thinking, the horizontal aids the vertical and vice versa. Also, from the point of view of pragmatic reality, even few great Bach players can really do 5 things at once. they learn to fake it by shifting the attention around skillfully so as to create the illusion. Whether you look at it musically or pragmatically, in terms of demands, there's simply no such thing as true independence in a musical and technically successful performance. it's a balance between associations and independence and demands versatile thinking from many viewpoints.

[nyiregyhazi]

by the way, people often forget quite how grounded good voice playing is in the dirty thing referred to as the vertical. what makes good horizontal playing? a few things, but an essential foundation among those is the ability to create distinction between voices via different dynamic levels ie voicing. Voicing is inherently a vertical issue. all the horizontal thinking in the world will not produce anything but a monotonous din unless a pianist has mastery of vertical voicing skills (and this is exactly what is lacking in so many pianists who can spin notes out dryly- the ability to voice). it's just totally irrational to portray these things in terms of horizontal is good and vertical is bad, when everything is so comprehensively interrelated. Vertical thinking is exactly where you learn the basic skills of individuating different voices. Frequently players who are "too vertical" are the very ones who need to stop and listen explore how to actually make differentiation on that vertical level. Their weakness is that they haven't explored vertical issues anywhere near well enough and instead hit chords with a monotonous single level. and none of the vertical explorations they need to start getting involved with precludes them from complementing that by playing individual voices too. Horizontal issues determine the desired voicing. Vertical issues determine whether you can execute that, or whether the parts turn into a homogenised lump.

[iansinclair]

OK -- I'll go way out on a limb and try a completely different metaphor: team sports.  Which need not be competitive, but it might be easier to envision in whatever team sport you envision -- doesn't matter, cricket, Rugby, soccer (football), football, baseball.  Whatever.

One may liken each individual player to one voice in polyphonic music.  If you follow just that player, you find that he or she is sometimes doing something really important, and sometimes something not so important.  But --if each and every player on the team is not aware of what all the other players are doing, the team will be just a jumble of folks running around on a field, and not a team, and you won't have the game.

Or vocal music: each part in a polyphonic composition is, of itself, interesting and sometimes very difficult -- but if each part is not paying attention to every other part, you have a sorry mess.

And so it is in individual playing of polyphonic music: each part is important to varying degrees at varying times all by itself (the "horizontal" aspect) but at the same time must be completely aware of and together with all the other parts (the "vertical" aspect).  You can't separate them...

[j_menz]

I haven't meant to suggest that the voices in a polyphonic piece are musically unrelated. They are deeply interrelated (and I rather like Ian's sporting team metaphor). The relationship is harmonic (which corresponds to the vertical), rhythmic (likewise), but also more complex - a voice can at a point be an answer to a question posed in another voice, or it can be an echo or development of an idea originating in another voice. The interplay is vital to the operation of the work.

To take N's example of playing a single voice of a five voice fugue in isolation - of course it lacks the perspective of those interrelationships. I don't memorise anyway, so I can't really say what would happen if one were to memorise each voice separately. I suspect part of the process would be to try and form those relationships on the way, but it's outside my field of experience.

In reading your post, though N, I am struck that you appear to think through your hands - from the movement to the music, if you like. I think I do it the other way round, so I suppose it is inevitable that we will see things differently here.

[nyiregyhazi]

I haven't meant to suggest that the voices in a polyphonic piece are musically unrelated. They are deeply interrelated (and I rather like Ian's sporting team metaphor). The relationship is harmonic (which corresponds to the vertical), rhythmic (likewise), but also more complex - a voice can at a point be an answer to a question posed in another voice, or it can be an echo or development of an idea originating in another voice. The interplay is vital to the operation of the work.

To take N's example of playing a single voice of a five voice fugue in isolation - of course it lacks the perspective of those interrelationships. I don't memorise anyway, so I can't really say y what would happen if one were to memorise each voice separately. I suspect part of the process would be to try and form those relationships on the way, but it's outside my field of experience.


In reading your post, though N, I am struck that you appear to think through your hands - from the movement to the music, if you like. I think I do it the other way round, so I suppose it is inevitable that we will see things differently here.

how could I have any such movement, if it were not derived directly from the music- particularly given what I said about voicing? It's simply that I'm aware of different perspectives- including the fact that the movement used is specifically what generates the result (and that independent thinking cannot contribute to a musical or technical execution without interdependent awareness) it's selected based on musical intention but movement is what generates what actually gets to be heard. there's a whole lot of stuff I never realised at first but which I have gradually begun to understand through a combination of rational logic and self observation.


I'm quite sure you use the same processes but haven't noticed that you are doing them. can you actually read a fugue that is notated on multiple staves and calculate your own fingerings and distributions? perhaps so, but assuming not, you are doing all the same interrelationships in your reading and using them as a very direct means of finding the physical actions (and even a pianist who can read such notation fluently is assembling relationships inside the head, if not seeing them so transparently on the page) the only real test of a pianist who has full capability of independence is to play any single lines from a fugue fully by memory. if you can't do so, your thinking is massively based on interdependence. of course, ideally we should know every voice so well as to be able to do that- but that doesn't mean it's a shameful thing to do a massive amount based more relationships than on true independence. it's just a matter of whether the listener hears proper distinction between voices- not whether you can literally be multitasking five things at once.


given that you now agree with interdependence (where you previously said that happening to land on the same beat should be enough, without any sense of making mental or physical associations between voices) the only difference appears to be that I'm acutely aware that there is a process of assembling relationships between parts (both in sound and in means of execution) whereas you do it more intuitively.

[j_menz]

the only real test of a pianist who has full capability of independence is to play any single lines from a fugue fully by memory. if you can't do so, your thinking is massively based on interdependence.

Since I don't do anything from memory, I can't say much about this. I can't play the whole thing together from memory either, so that failing doesn't speak to independence.

I don't especially like reading more than two staves, though I'm not sure that that's because of a greater difficulty in putting it together or just not liking having more places to look. When I see a score, I see the individual lines anyway, so apart from the convenience of having them already sorted geographically* and all in one easy to see place, I'm not sure it makes any difference.  No doubt I have to sort out the mechanics (fingering and movement), but I'm not generally aware of doing so.

I still think we see things differently and so risk speaking at cross purposes. Whereas you see movement as being what you do to produce a preconceived sound, I see that preconceived sound as what  produces the required movement.

* I did recently read through a fugue where the "middle" voice was notated on the bass clef and the "upper" voice was notated on the treble. The two voices frequently went above or below each other, so there could be leger lines up on the bass and leger lines down on the treble at the same time. I have to say I found it something of a headache.

[nyiregyhazi]

Since I don't do anything from memory, I can't say much about this. I can't play the whole thing together from memory either, so that failing doesn't speak to independence.

I don't especially like reading more than two staves, though I'm not sure that that's because of a greater difficulty in putting it together or just not liking having more places to look. When I see a score, I see the individual lines anyway, so apart from the convenience of having them already sorted geographically* and all in one easy to see place, I'm not sure it makes any difference.  No doubt I have to sort out the mechanics (fingering and movement), but I'm not generally aware of doing so.

I still think we see things differently and so risk speaking at cross purposes. Whereas you see movement as being what you do to produce a preconceived sound, I see that preconceived sound as what  produces the required movement.

no I don't. I see both. neither truth contradicts the other and I absolutely despise false polarisation. there's no need to ignore either half of the picture. The fact I refuse to ignore either half is precisely what I am saying - just as I refuse to ignore the role of verticality in fugues simply because horizontal aspects are a huge issue.


consider that unless you always take fingering from what is written in, you are constantly making associations. Nobody would use the same fingerings for a single part that they will apply to it when they have two parts in that hand. this alone is proof that you're doing all the same things but simply not doing it consciously. personally, what I've found is that the more I understand what I used to do only intuitively, the better I become at avoiding the sloppy moments that were part and parcel of my former intuitive version, and the better my intuitive playing goes even without those conscious thought processes. my execution of such things as Bach fugues at first sight remains unspectacular but has made a marked improvement.

[j_menz]

consider that unless you always take fingering from what is written in, you are constantly making associations. Nobody would use the same fingerings for a single part that they will apply to it when they have two parts in that hand. this alone is proof that you're doing all the same things but simply not doing it consciously. personally, what I've found is that the more I understand what I used to do only intuitively, the better I become at avoiding the sloppy moments that were part and parcel of my former intuitive version, and the better my intuitive playing goes even without those conscious thought processes. my execution of such things as Bach fugues at first sight remains unspectacular but has made a marked improvement.

I never even notice notated fingerings in most works, and that includes fugues. I sometimes look at them consciously if my intuition fails me.

It may well be that I do what you do but unconsciously. Obviously, my fingers have to get to the notes - I'm just not thinking about it.

I don't think I have any Bach fugues left that I could do at first (or second or third) sight, but I find my approach works on the rather large collection of other composers fugues I am reading my way through.

I still find that when you talk about "verticality" in these works you talk about it in relation to fingering and movement. Not about the musical relation between the voices.

I;'d be interested to know if you have read through or played any of the Liszt or Busoni transcriptions of Bach's organ fugues; if so, did you perceive verticality operating differently there?

[nyiregyhazi]

I still find that when you talk about "verticality" in these works you talk about it in relation to fingering and movement. Not about the musical relation between the voices.

I;'d be interested to know if you have read through or played any of the Liszt or Busoni transcriptions of Bach's organ fugues; if so, did you perceive verticality operating differently there?

? I've spoken various times about intervals, voicing and harmony. the issue of fingering is the simplest definitive proof that it's impossible not to involve constantly associated thinking, but I'm speaking as much about musical issues.


Liszt and Busoni are interesting ones- as vertical issues are much less likely to be missed due to the thickness of chords. you can't accidentally run your fingers on to the next without really noticing the texture you just played (or at least, it's less likely). with a three part fugue, it's actually much more likely that you fail to really notice your vertical reference points before skidding on. This is both a musical issue and one of stability, physically speaking. With the score in front of me, the notes can go straight from eyes to the fingers- but when playing from memory you start to realise that you're not really noticing the details before passing on. this doesn't happen as much in thicker textures, as you tend to need to be more aware of vertical issues simply in order to execute them at all. In thinner writing, it's easier to miss the associations mentally, yet still get through the notes physically, despite all the mental holes- at least when the score is there. take it away, and you will start to feel that you don't even know know what the fingers are really doing or where they are headed, unless you are clear on many vertical references (none of which are ever reached outside of horizontal context or in any way taken entirely out of that context).

[j_menz]

What I was getting at was that in Liszt and Busoni (and others, too, of course) sometimes a voice or voices are doubled (or tripled and other fun things).

The relationship between the doubled notes is purely vertical, and seems to me a rather different relationship than the relationship between the notes of two voices would be, even if the actual notes were identical (and employed the same fingering and even had the same relative volume).

I use it to illustrate that although the fingers may be doing identical things, and the sound may be, at that moment, the same, the relationship is actually different.

I appreciate that I'm taking a moment in isolation, and that that may be somewhat unrealistic.

Also, as I have stated earlier - if you take the score away, I'm completely lost regardless.

[nyiregyhazi]

What I was getting at was that in Liszt and Busoni (and others, too, of course) sometimes a voice or voices are doubled (or tripled and other fun things).

The relationship between the doubled notes is purely vertical, and seems to me a rather different relationship than the relationship between the notes of two voices would be, even if the actual notes were identical (and employed the same fingering and even had the same relative volume).

I use it to illustrate that although the fingers may be doing identical things, and the sound may be, at that moment, the same, the relationship is actually different.

I appreciate that I'm taking a moment in isolation, and that that may be somewhat unrealistic.

Also, as I have stated earlier - if you take the score away, I'm completely lost regardless.

although I might play one of the doubled voices with more intensity, I wouldn't see these issues in the same light as comparisons between independent voices. it's more about pianistic sonority and colouring and I wouldn't do that terribly analytically. in terms of counterpoint, it's just a single unit. The thing I find interesting in such writing is how much easier it is to organise mentally due to the necessity to be vertically aware. a run of single note semi quavers can easily run away and there can be virtually no awareness of the interesting arrival points. Simply put that same run into octaves, and it becomes almost impossible not to develop a serious awareness of relationships between that voice and others. This is because the whole arm travels to each note. when the fingers play notes under the hand, you don't always observe what they are really doing in terms of intervals. With octaves, the internal mental processes get transformed as a matter of necessity. learning to do the same mental processes in a regular Bach fugue (rather than running things off physically without deep awareness of associations between voices) creates much more sense of musical structure and solidity.


one thing that springs to mind is that vertical playing is often associated with lack of voicing in a chord. That doesn't mean it will be. Useful exploration of a vertical moment should be a snapshot of a single moment in a horizontal path- not a monotone lump without context. the thing like about doing a beat plus a note and then stopping is that everything is done in association. Every observation of a vertical moment is approached horizontally- so it's a means of proper cross-referncing, rather than any kind of polarised viewpoint.

[j_menz]

Much as you say we must be seeing the same thing going on, I can assure you we don't.

You speak of chord voicing in a single note multi voice polyphonic work as if there was such a thing.  I simply don't see that there is.

The volume, and every other attribute, of each note at a single point is determined exclusively by its position in and in relation to the voice wherein it occurs. Nothing else is needed.

Certainly the voices interrelate, but that is a matter of harmony and melodic relationship.  It is nothing like voicing a chord, where the relationship is purely vertical.

[nyiregyhazi]

Much as you say we must be seeing the same thing going on, I can assure you we don't.

You speak of chord voicing in a single note multi voice polyphonic work as if there was such a thing.  I simply don't see that there is.

The volume, and every other attribute, of each note at a single point is determined exclusively by its position in and in relation to the voice wherein it occurs. Nothing else is needed.

Certainly the voices interrelate, but that is a matter of harmony and melodic relationship.  It is nothing like voicing a chord, where the relationship is purely vertical.

sorry, but I find this point plain bizarre. have you not stopped to consider where "voicing" comes from as a word? it comes from the idea of distinguishing separate voices. for some reason you seem to think it inherently refers to experimenting with totally decontextualised layers? it doesn't. you're creating a totally false polarisation again. a chord (which is a perfectly accurate description of multiple notes sounding together in- whether they are approached homophonically or in counterpoint) in a fugue needs to be voiced to reflect horizontal movement of voices. that why it's called voicing of chords. if you cannot voice vertically in reference to horizontal context, neither can you maintain any musical logic in individual parts- the sound of which will be limited by inability to select their dynamic level, when there are other notes involved.



when you deny the vertical aspects you're completely contradicting your point about musical relationships which you also appear to have dropped again? how can there be relationships if everything is done exclusively in reference to that voice and none other? this is a total contradiction to what you have said elsewhere. interval and harmonic relationships between intervals can only be observed vertically and by comparison to other parts. otherwise they are simply random. you don't have to choose any single perspective and noticing one important thing in no way means you cannot notice another important thing. if you want to throw out this key aspect of musicianship (which would destroy the musical value of suspensions and chromaticisms) because of dogma about the equally important horizontal aspects, you really don't need to. you're coming from the point of view that if A is true B cannot be. In this case not only do A and B not contradict each other, but both are absolutely essential ingredients.

[j_menz]

No. Consider a simple example. Two voices, the first doubled at the sixth.  The dynamics of the first voice, in totality, are determined by the horizontal working of that voice ( crescendo/decrescendo, pulse etc). The same for the second voice. If the dynamics are correct horizontally, they are correct vertically as well.

In the first voice, however, the relationship between the two notes is balanced vertically to give effect to the sound to be produced.

The two notes of the first voice act like a chord, but the relationship between that chord and the other voice is quite different.

[nyiregyhazi]

No. Consider a simple example. Two voices, the first doubled at the sixth.  The dynamics of the first voice, in totality, are determined by the horizontal working of that voice ( crescendo/decrescendo, pulse etc). The same for the second voice. If the dynamics are correct horizontally, they are correct vertically as well.

In the first voice, however, the relationship between the two notes is balanced vertically to give effect to the sound to be produced.

The two notes of the first voice act like a chord, but the relationship between that chord and the other voice is quite different.

do you not actually understand what false polarisation is? the relationship between the two voices is vertical. do you think it doesn't matter if the dynamics are associates? purely horizontal thinking would mean that you can start the voices in any old way. also, unless you feel that there's literally only one possible shaping for any given phrase, true independence could easily have one part getting carried away with a huge crescendo while the other is fading out.


This is really getting barmy. I have absolutely no idea by what is so baffling about the idea that the importance of horizontal and elements and of vertical elements are both true. there's nothing complex about this utterly irrefutable truth and the only arguments you are giving merely illustrate that horizontal elements are an issue. that in no way serves as a counterproof to the truth that BOTH are big issues.

[j_menz]

purely horizontal thinking would mean that you can start the voices in any old way. also, unless you feel that there's literally only one possible shaping for any given phrase, true independence could easily have one part getting carried away with a huge crescendo while the other is fading out.

Not at all. But what determines how you start one voice is as much a matter of how you started another voice of which it is a restatement/answer/echo. That most often will be not vertical, but a way back in the score. 

There are many ways of shaping a given phrase, but that shaping must relate to how you have shaped the corresponding parts of other voices elsewhere, not what those other voices are doing at the present time. 

And it is true, you could have a crescendo in one voice and a diminuendo in another - I don't see why you would think that to be surprising.

[iansinclair]

j_menz, your patience is astounding.  Hang in there... and keep trying to explain.  You're doing just fine.

That said, however, it seems to me upon looking at this thread that there is something sadly lacking in at least some music education: an exposure to counterpoint.  Not as a strictly keyboard exercise.  And perhaps it is that something which explains why some folks find Bach and company  so deadly dull or, perhaps worse, simply mystifying.

May I make a suggestion to all and sundry?  Find a really good recording of the St. Matthew Passion, or some similar work, and listen to it carefully.  Actually, almost any vocal work of the era would do, but the St. Matthew Passion is a pretty decent place to begin.  Listen to how the various voices intertwine, sometimes following one another, sometimes going off on their own... but always making one marvelous unity.  Perhaps even get a copy of the score, and see really analyse what is happening.  You could do the same sort of thing with the Monteverdi Vespers of the Virgin Mary, if you wanted to.

It might be worth the effort...

[j_menz]

Thanks Ian - I thought everyone else had switched off ages ago. 

Excellent advice on the listening, too. I'd merely add that doing it with a score is especially beneficial - particularly for those who haven't yet learnt to listen to multiple voices at work.

Additionally, if Bach and the baroque really aren't to your taste, the later symphonies of Tchaikovsky provide an alternative, albeit the voices are not as immediately apparent.

[ajspiano]

If any one note from a horizontal line sounds at the same time as a note from any other voice those notes have a vertical relationship. That's just the way it is, period.

However, in my experience, while studying fugues or just counterpoint in general - Failure to execute the vertical relationship successfully (eg. voice entry at right time, or dynamic control between voices sounding at the same time) has near 100% of the time been due to not processing the voice individually and horizontally in my mind. So long as I think "that note comes in time with this other note in the other voice" or "this note louder than that note" it fails.

It works when I process the beat only, and i feel where the entry is in relation to the beat, not the other voice. The dynamics and timing can be described can be describe vertically, but that's not how I deal with it when playing it.

We even rationally decide perhaps on dynamics with reference to a vertical context (upper voice vs lower voice).. only the thing is, its not on a note by note basis using each individual chord. its always based on the full line. There's practically zero vertical processing for me.

improvising/writing it is another matter though..

[ajspiano]

there is something sadly lacking in at least some music education: an exposure to counterpoint.  Not as a strictly keyboard exercise.

Not that I would intend to know what other people experience, but I can certainly attest to a very limited exposure just as a piano student.

I personally reached diploma level material (with contralpuntal elements ofcourse) under the guide of a teacher without ever hearing the word counterpoint or there being any real discussion of voices, and I don't remember ever being shown a bach invention, let alone studying one.

I was given a P+F, f minor book 2 I think it was, very late in the piece but didnt follow it through - and I didn't get it because the teacher thought I should study counterpoint. It was just "here's some pieces you can do for the exam, which one do you like the sound of"

....

I was probably Miyagi'ed to a degree though. I did begin learning how to write in 4 part vocal harmony under that teacher. Still, that was a very vertical experience..

Strikes me as quite odd really, because I'm fairly sure my teacher had a theory Amus, so she must have known plenty enough about it. I personally introduce it to students ASAP.

[j_menz]

We even rationally decide perhaps on dynamics with reference to a vertical context (upper voice vs lower voice).. only the thing is, its not on a note by note basis using each individual chord. its always based on the full line. There's practically zero vertical processing for me.

I'd only add that it may be based on another voice, but at a different time.

For example, in a fugue it's usual to structure the first entry of each of the voices in the same way, both to reflect that they are the same subject, but also to help the listener get to grips with them before you later start to play around with them. You make it harder for the listener to understand "this is the subject - listen to what I can do with it" if you play it three different ways the first three times they hear it.

[ajspiano]

I'd only add that it may be based on another voice, but at a different time.

For sure, I meant that we consider the entire line in an individual voice as a whole, and make decisions about about lines in other voices based on that. You could argue its vertical, but on whole horizontal lines, never on isolated vertical notes/sounds.

The only truly vertical consideration I give is to fingering, which gets left behind mentally very quickly.. and these days I don't have to actually consciously set an exact one anyway in most cases. So I'll just start horizontal. I'm not sure you can argue that the vertical doesnt matter so far as that though. That's part of the learning that you have to go through to be able to process it the right way musically perhaps?

[j_menz]

fingering ... I'm not sure you can argue that the vertical doesnt matter so far as that though.

Nor would I try. You need to be able to hit the notes (and for reading that takes some practice to get a feel for it). But you do need to know what to do with them when you do hit them.

[ajspiano]

But you do need to know what to do with them when you do hit them.

Indeed, which is while all and sundry should learn all 15 inventions (and perhaps some of bach's other works that might qualify as 2 part inventions) before tackling 3 and in turn more voices.

..as we keep saying to people here.

[j_menz]

Indeed, which is while all and sundry should learn all 15 inventions (and perhaps some of bach's other works that might qualify as 2 part inventions) before tackling 3 and in turn more voices.

..as we keep saying to people here.

Apropos of something close to that, I'd invite people also to have a listen to this:

The interplay of the voices is perhaps given a different perspective when played by two people.

This should also finally shut up that school of thought that says Bach should be all mono volume detache unmusical crap. I won't be holding my breath, tough.

[ajspiano]

I'd invite people also to have a listen to this:

The interplay of the voices is perhaps given a different perspective when played by two people.

Cool.

I have an ongoing fantasy to find someone willing to try and improvise this kind of thing. Where you define a motif and counter motif, and then just mimic each other and develop it together on the fly. I'm not sure you could do anything all that complex without running into problems, but I wonder if you could set certain constraints that mean that the ending of a particular subject defines the beginning point of the following response absolutely. So that one musician can predict the most of the others improvisational reactions to the music, and choose the appropriate variation of a counter-motif for their expectation of the other performer.

I use such a concept in my own playing, but I'd love to see if its possible to do between two separate minds.