Some Notes on the Compositional Process


Christopher Bailey


            Depending upon the piece,  attempting an analysis of one's own work can be a difficult process.  With Timelash, I have no programmatic or philosophical background to unveil:  Timelash is simply an abstract work.  I would love to talk about the sounds of the piece, however, this would be difficult in a situation where audio is precluded from the presentation.  Therefore, I am left with the possibility of recounting some of what I can explain and remember about the compositional choices made in the work's creation.




            For some, but not all, of my compositions, a certain work process has become more or less standard.  We'll discuss this process in relation to Timelash.  The process is as follows:


            1)         Generate one or more generalized, abstract pitch designs, often (but not always) with a serial structure of some kind.  Such a structure is 'open' enough that I might, in fact, use it for more than one work.


2)         Specifically arrange the structure in 1). This may include transposition/inversion/retrograde, etc., and/or combining, overlaying, or overlapping the structure upon itself.  A particular arrangement will be used for one specific work.


3)         'Parse' the structure created in 2) into smaller units, smaller groups of pitches.  The goal is to reach a point where I can switch my compositional thinking from an 'abstract design' mode, to what most people would think of as a 'composition' mode.  To do this, my intuition needs to be able to work with smaller groups of elements (notes or otherwise.)


4)         Generate any useful or need-to-know formal outlines:  large-scale contours of pitch, register, tempo, density, orchestration, etc.


5)         Do a rough 'improvisation/composition' sketch with the given materials. A kind of 'graphic score' is produced, with rhythm mostly inexact, many parameters (i.e. loudness) left unspecified, and so on.


6)         Separately and independently, generate some sort of abstract rhythmic structure. 


7)         Finally, the results of 5)  are 'squeezed into' the results of 6).  All necessary details (timbre, duration, loudness, etc.) are determined.



1)         Abstract pitch design


            Timelash employs a type of abstract pitch design, related to those developed by some American composers in the 60's and 70's,  the so-called 'all-partition array'.


For those who are unfamiliar with such structures, I provide a brief introduction. An array is a diagram of 'un-realized' counterpoint, which a composer can 'compose out' or, in a way,  'improvise upon'.  If a tone-row is thought of as a  'not yet fully [musically] realized' string of pitches[1], like so:


5 9 2 t 3 7 0 8 1 4 e      6 t 3 e 4 8 1 9 2 5 0

row T0                     row T1  


which can then be realized in a musical work:



then, an array is, analogously,  a 'not yet fully realized' counterpoint of such strings of pitches.  For example, here is a bit of 4-voice array used in the piano part of Timelash:



Here, the lines of counterpoint take effect as, or are realized in, different registers of the piano:



In another situation, array lines might be realized as different loudnesses, or different durations, or even different stereo positions (in an electronic piece, for example).  Here is an example with different durations:  short notes versus long notes, 2 lines:



            We can divide an array into 'blocks,' each of which obeys some rule.  For example, a block might be required to contain each of the 12 possible chromatic pitches once each:



       The 'partitioning' of each of these blocks is shown below the array:  the way the lines divide up the 12 tones, differently in each block.  In all of these examples (which are excerpts of larger structures), the lines themselves are sequences of tone-rows, in various transformations.  Timelash's underlying array is similar to a type of structure known as an 'all-partition-array'.[2]


All this is in the background of a piece.  It's not expected (at least not by me) that one would hear very much, or even any of this, directly.  But from my point of view, as a composer, an array acts as a vibrant creative stimulus:  each block suggesting different rhythmic formations or relationships between voices or registers, or loudness contours, or whatever may result from the particular way that the lines are realized as actual music. For example, in the first block above, the top note is like a 'drone' while a middle voice has a 'solo' of sorts, in the next block the 'solo' passes on down the lowest voice, then there are 'duets'  between the upper 2 voices,  and so on.  I like the way arrays constrain me as a composer, while at the same time allowing for a high degree of flexibility:  what actually happens in a piece is still hardly determined.



2)         A specific structure for a specific musical work


The complete pitch array actually used in Timelash is included at the end of this article.  Notice that there are two 4-line arrays, overlayed, both with 4 lines divided by register.  One array concerns the pitches in the instruments, the other, the pitches in the piano.    The two arrays are very similar:  the strings of pitches moving 'across' (through time) are the same, but they divide into blocks in different ways.  Because of this 'not quite equal' relationship, there can be complex,  'fuzzy' echos of materials between the piano and the instruments.





(Below are a few blocks of the source array used for those measures, showing, perhaps more clearly, the similarities between the instruments' array, and the piano's.  The boldfaced notes are the ones composed-out in the example:)




3)         Parse structure into smaller units.


            A structure like that shown directly above is normally too open for me as a composer--it is not constrained enough--it needs to be 'parsed' or broken into smaller units.


To explain how this happens in Timelash, let me digress for a moment.


Composing with these materials against some sort of 'drone', or constant harmonic entity, provides a sense of direction for gestures and phrases, a sense of moving away from and returning to a 'home base'.  Composing against a drone may also relate a work to European and non-European music I am interested in, such as North and South Indian classical music, whose gestures and phrases move away from and towards a 'drone' as well.


In this piece, somewhat arbitrarily, I decided upon this sonority:


EXAMPLE 7: Timelash drone-chord

as a kind of home 'quasi-tonic' around which the counterpoint would arrange itself .


The chord is allowed to ring in the piano, via the sostenuto pedal, through the entire work:  its notes are never dampened.  The damper pedal is used intermittently throughout to allow total resonance; there is a hierarchy of resonance in the piano part.


I also decided to use this sonority to 'parse' the array into smaller units.  In each block of the array, I looked for this chord's pitch-classes, and separated them out.  This chord has 8 notes, the 4 'leftover' notes are 0 4 8 9  (C E G#/Ab A).  The next example shows 2 such 'parsed' blocks of the array (compare with the 'unparsed' array at the end of the paper):



As noted in the example, this further parsing is usually more 'suggestion' than 'rule':  when it comes to the actual composing, I go with my ear.  Thus, in the final score, the pitches are rarely divided up in exactly this fashion. 


4)         Large-scale formal outlines


Generally speaking, at this point in composing a work, I may move tentatively from 'pre-composition' to 'composition,' and begin sculpting actual music out of the materials.  After these initial composing efforts, I often pull back to 'pre-composition' mode again, and map out changes in density, texture, climax/energy, and so forth for the entire work.


However, during the composition of Timelash, I don't recall making any maps of these parameters; rather, I had a set of 'goal-posts' or 'milestones' at various points in the materials that I would compose towards or away from.  For example, clearly the section at measures 188-217 was thought of as a kind of 'climax.'  There are other, sub-climactic areas of the piece, such as mm. 65 ff. and mm. 270 ff.  I no longer recall which were pre-composed, and which simply came about during the process of composition.


In short, the large-scale 'kinetic' structure for this piece was largely developed intuitively, in contrast to others of my works, where it is more thoroughly pre-composed.


5)         Initial composition/improvisation on materials.


            Step 5) in my process is where the 'real' composition begins to happen:  the abstract structures I've spoken of so far are now assembled into harmonies, real (sounding) lines assigned to instruments, and so on.  A 'graphic score' sketch of the whole piece is produced.  (Please see Example 10 below for a sample excerpt from the graphic-score sketch for Timelash).


            So far, my task in this paper has been relatively simple:  I have shown charts and diagrams, illustrating what was, for the most part, pre-composition.  But we still have barely discussed the most important aspects of the piece:  what makes the piece sound the way it does.  This is the most difficult step for me to discuss:  the composing-out of the materials is, for me, an intuitive, mysterious process.  I can perhaps make some 'guesses' as to how my intuition works, or about formations that itÕs drawn towards.


Some aspects of my sense of harmony, for example, can be found buried within the 'drone' chord.  I have noticed, recently, that the spacings of my harmonies tend to derive, apparently subconsciously, from overtone series spacings.


We can hear the drone-chord as a set of overtones:


EXAMPLE 7: Timelash drone-chord

Specifically, over a very low (and not actually sounding) F#, (3 octaves below the actual (sounding) F#), the chord represents overtones 8-20-24-25-26-30-34-44.  In simpler terms, we are interested only in the 'odd' partials (since any 'even' partial is simply an 'odd' partial transposed up by one or more octaves); we have the Root, 3rd,  5th,  9th, 13th, 15th, 17th, and 11th partials.  (Thus, essentially, none of the partials used are above 20.)  Obviously, some of these are up to a 1/4-tone out of tune in 12-tone-equal-temperament, but I believe my ear was drawn towards them nonetheless.


            Related to the drone-chord are a number of smaller pitch-groupings that tend to pop up frequently in Timelash,  and in much of my other music.  A few examples include:


1)         diatonic pitch-collections which have been turned 'hard, cold and brittle' through registration, rhythm and gesture, and through their juxtaposition against other, non-diatonic events.  For examples, at mm. 61 ff., a 'fanfare' on E-G#-A-C#-D, or at mm. 313 ff., an 'F# major 1st inversion triad,' with the A#/Bb in an extremely low register.

            2)         The jarring and beating of minor second intervals is the subject of much contemplation, occurring either in very small gestures:  msr. 98, 120, 125, to point out just a few examples; or in larger segments of music, as at 188 ff.

3)         similar collections,  3-note chromatic clusters,  with the pitches arranged between the instruments,  to beat against or 'rub against' one another:  mm. 18 ff., msr. 72, mm. 84 ff., mm. 138 ff., mm. 270 ff., to name just a few examples.   Note at 84, the un-dampened F# 'drone' poised against the F-G 'neighbors.'  The sound at this point is one that most clearly reminds me of Indian raga, especially those beautiful rags that include the upper and lower chromatic neighbors to the 'tonic'.  (for example, rag Purvi).


Although, during the pre-compositional stage of composing, I am interested in abstract, so-called 'set-theoretic' pitch manipulations, once actual sounds in a work begin to take shape, I am more interested in exactly where (registrally) the pitches are, how loud, in what instrument, and how they relate to other pitches, acoustically (i.e.: consonance vs. beating, etc.) or otherwise, than I am in abstract issues of questionable perceptibility.



6)         Develop an abstract rhythm structure


What I generated so far in the compositional process was a 'graphic-score sketch' of the composition, its harmonies and gestures laid out in a vague way, not precisely rhythmicized.  Now I want to be able to 'squeeze' those ideas into an arbitrary rhythmic 'schema' that comes from elsewhere.


In Timelash the 'schema' is a 26-bar rhythm of 32nd-notes which loops through the piece until the end.[3]


EXAMPLE 9: The Rhythm-Schema for Timelash




What follows is an example sketch of part of the piece, followed by the same excerpt, 'squeezed' into the rhythm-schema, in final score form.



EXAMPLE 10:  Graphic-score sketch





The most interesting thing about this way of working is the 'transaction' that happens between what the sketch implies, rhythmically, and what the rhythmic schema 'requires.'  Compare the 'chord' at the end of the sketch, with the way that it is composed out in the final score.  What is written as a simple 'chord' or vertical sonority in the sketch, is transformed into a sort of arpeggiated sonority, because we need to 'use up' rhythm attack-points from the schema.  Similarly, you will notice how the clarinetÕs A-C-Eb line is 'lined up' with the pianoÕs E-C-Eb in the sketch; in the final score, their unison is 'staggered,' 'fuzzified,' again, to 'use up' attack points.  Of course, as well, a staggered, arpeggiated chord or dyad is (usually) more interesting than a simple vertical chord.


Another possibility that I allow myself in composing out the final rhythm is to 'break up' a measure of the schema into 2 measures; in effect, one measure is 'paused,' and it then continues in the next measure.  The following is an example of 1 measure of the schema being stretched in this way to 3 measures of actual music:


EXAMPLE 12: 'paused' Schema






            Previous to composing Timelash, my 'style', in terms of rhythm, had been to use lots of complex, nested tuplets, as with much music of the so-called 'New Complexity'.  I deliberately chose the rhythmic 'schema' for Timelash, with it's relentless grid of 32nd notes and no tuplets; I wanted to prove to myself that I could achieve a level of rhythmic complexity, varying from the inflected, expressive, 'floating' rhythms that I had worked with before, to the energetic, precise rhythms that one can create in a strictly pulsed time field.  The former is usually achieved by taking a series of consecutive attack points and distributing them around the ensemble, as in Example 12; the way that the different instruments' timbres attack automatically insures a kind of 'floatiness' to the sounding rhythm, which is very much intended.  On the other hand, the latter situation is fairly easily achieved in a steadily pulsed time field, following examples set by Stravinsky, the minimalists, the 'Bang on a Can' composers, and so on.




This, then, concludes the overview of the compositional process I used in composing Timelash.  Although there are "rules" that I followed as a composer, we can see that there was still a great deal of room for freedom.  I don't think that this way of working is any more constrained than the tonal system---the constraints are, in fact, simpler here; certainly, they are easier to describe.


            In several other works of mine, Sand, interactive computer-music, or Trio for piano, clarinet and 'cello, this same composing process was followed more or less exactly.  In a work such as Motet  for 6 mezzo-soprano, or The Stuffed Ones, for computer music, the rhythm was written out,  but intuitively created.  (There was no 'schema'.)  In The Quiet Play of Lights, a trio for piano, violin, and 'cello, I decided to stop at step 4),   and simply present to the players, in effect, my "graphic score sketch" and see what they'd do with it.  Still other works (6 Songs on Poems of John Monroe), I experiment with non-serial 'abstract pitch designs'.  Finally, other works of mine do not concern themselves with pitch structures at all, except perhaps in a [purposefully] primitive fashion; for example, my purely musique concrete works Ow, My Head and Duude.




EXAMPLE 12:  Timelash, complete pitch array

(upper 4 lines are vln., clar., 'cello, lower 4 lines are piano)

 (All measure numbers are approximate:  in the actual music, piano and instrumental arrays don't always line up exactly,  and boundaries between array blocks are fuzzy and often slightly violated for musical reasons.)

(Lines are made up of sequences of 'classic' row forms P, I, RI, and R of a 12-tone series: <0291ab576384>)












[1] I am using a numerical notation for pitch classes here,  from 0-11, where 10 and 11 are "t" and "e" respectively. Hence C is 0,  C# is 1,   . . .  Bb/A# is "t",  B is "e".

[2] The constraints for an 'all-partition array' include:

            1)         The lines are classic forms of a tone-row (Prime, Retrograde, Inverstion, RetroInv)

            2)         Each  block contains all 12 tones once each.

            4)         'All-partition' thus means that all of the possible partitionings are used, once each.

[3] The source of this schema is a series of attack-points taken from another work.  I no longer remember what the piece was,  but the schema, and the music I make with it in Timelash is so far abstracted from that piece,  that I don't think it matters.  You'll certainly not remember the other work when you hear Timelash.