A practical on viable ways of using indeterminacy in repose III (2018)

This essay was written earlier this year for a module (Context, Materials and Repertoire) of my Masters Composition Course at the Royal College of Music. I am posting it here since it is a product of self-reflection in my compositional practice.

A significant part of the research discussed in this essay has been born of practical necessity from my[1] recent involvement in an ongoing College project, in which composition students are invited to write new pieces for our peers in Consort 21, a Historical Performance (H.P.) chamber ensemble. I had already written a piece (repose I) in dialogue with my coursemate’s (repose II), when there came a request for more pieces for an unfixed number of performers. The request is reminiscent of the fact that the sizes and instrumental configurations of ensembles in seventeenth-century Europe were often largely variable, subject to the availability of musicians and instrumental resources in local courts or churches. Strangely, I observe a similar phenomenon in Consort 21—the agendas for our weekly two-hour rehearsal sessions are determined in a bricolage-like manner depending on the performers who showed up; the combination of instruments varies weekly. Understandably, such issues with scheduling are inevitable when there are multiple performers involved. In this light, I sympathise with the pragmatic motivation behind the request for pieces which can make the best use of the performers present each time.

Writing a piece as such presents a two-fold challenge—(1) designing a structure which effectively harnesses its unpredictability in terms of instrumental timbre and texture, and (2) providing musical material sufficiently adaptable (or ‘translatable’) for performers to execute idiomatically on their respective instruments. The latter also involves allowing performers the freedom to showcase their individuality through (possibly virtuosic) technical handling, thereby encouraging a flow of heterophonic beauty derived from the melange of instrumental idiosyncrasies. For me, this challenge is especially critical given my observation (from the first two rehearsal sessions) of a peculiar tension in the performers’ practice. Undeniably, they have received considerable training in realising musical material in a manner specific to historical stylistic (‘Baroque’) sensibilities, but I wished they could go further in engaging and experimenting with material. Arguably, the performers’ training, which permits little room for making ‘mistakes’ from extemporaneous experimentation, hinders them from realising latent creative possibilities. Understandably, the state of musicianship I am describing should come with practice and experience in improvisation, but a student performer has to first develop and expand a consciousness as such.[2]

With the above considerations in mind, an effective solution would be to use a controlled level of indeterminacy in my piece, repose III, in shaping its structure and realisation of material. My use of indeterminacy is justified by virtue of it being a means to achieving a continuous growth in the texture and ornamental floridity of my piece, depending on the ensemble’s size and instrumental combination. This way, the uncertainty of performers’ attendance can also be embraced and transformed into the aesthetic of the piece itself, in allusion to the historical context. In response to the second aforementioned challenge, a productive way of gently prodding the performers towards extemporaneous experimentation is to re-contextualise repertoire with which they would generally be familiar. I adapted a fragment from ‘Loure’ of Bach’s Partita No. 3 (for solo violin), BWV 1006:

‘Loure’ of Bach’s Partita No. 3 (for solo violin), BWV 1006

While the choice of fragment is admittedly arbitrary, its two-part counterpoint offers numerous opportunities for imitative and contrary linear movements. (The fragment has been transposed into C major in the interest of playability on instruments such as the Baroque harp.) The fragment reposes in the new context of my piece; the performers would also be encouraged to semi-improvise variations based upon—and reposed on—the fragment. In this sense, my piece derives its contemporary relevance from encouraging (and possibly, even liberating) the performers to juxtapose their personal responses onto the music of a venerated figure, who, after all, ‘left little beyond the notes themselves’ on the page.[3] The piece’s texture and ornamental floridity grows across its duration; the larger the size of the ensemble, the greater the variability of interlocking rhythms and harmonic variety arising from contrapuntal activity.

Realistically speaking, I have to confront the question of how exactly to enable and control indeterminacy in achieving the outcome described above. In principle, the more variables I fix, the narrower the range of options available to the performers, and vice versa. Furthermore, I am conscious of how each detail provided on my score would effect a variable response in the performer, in turn giving a different performance outcome. I thus have to deliberate the extent of notated musical material to provide, wording of performance instructions, visual arrangement of score elements, etc. The instructions themselves also have to be sufficiently clear yet open to interpretation. Dealing with indeterminacy thus seems like a rigorous exercise in shaping human (performers’) behaviour. The process is akin to taming a torrent of possibilities by constructing dams and ditches at strategic spots, allowing it to flow freely within the framework, nonetheless exerting control over its direction, and harnessing its potential energy (i.e. the performers’ collective technical capabilities, and individual musical sensibilities[4]).

I am cautiously aware of the potential for the following account of my compositional thought process to be confusingly tortuous. In the interest of discursive clarity, I shall attempt to systematically (re)present my thoughts and research in the form of three practical ‘scenarios’ of my piece, each of which is informed by my observations on pieces by Berio, Lutosławski, and Riley. Granted, differences in the design of each scenario necessitates a notation and layout unique to each—one which communicates my intentions in the clearest possible way to the performer. Nonetheless, it is important to clarify that these scenarios are not different versions of my piece (and certainly not different pieces). Each of them demonstrates a different viable approach to using indeterminacy by controlling variables in slightly different ways, and to varying extents. On the whole, however, they are means to achieving the same aesthetic outcome—a growth in musical texture, and an increase in the extent of ornamentation and melodic elaboration across the duration of the piece, corresponding to the ensemble’s size and instrumental combination. It is also worth stating that all three scenarios share a particular way of allowing indeterminacy to arise (in performance) from the design of repetition schemes.

Scenario 1

In this scenario, the precise order of events has essentially been predetermined. Performers are instructed to enter at specific points and play/repeat given material; they do not have the option to vary the sequence in which the given material is performed. Increasing the number of performers lengthens the piece’s duration—more time is required for every additional performer to take her turn to enter. By virtue of sheer instrumental force, a considerably dense texture with florid elaborations would arise at a variable ‘midpoint’, just right after the later entrants begin, and before the earlier entrants drop out (see Ex. 1 below). The perceived location of this ‘midpoint’ shifts forward with fewer performers; with more performers it could possibly be stretched into a longer period.

Ex. 1 – A fragment from a ‘written-out simulation’ of Scenario 1 (with ten performers)

By fixing the order of events, I am able to ensure that the element of unpredictability arises primarily from varying the number of performers. Having established this framework of control, I can then negotiate specific parameters to generate further musical interest. Considering my earlier observation of the performers’ apprehension about extemporaneous experimentation, I decided that for the Upper part (for odd-numbered players), it is justified to provide two subsequent written-out realisations of the original fragment with increasing floridity, and encourage performers to further embellish them in line with their performance practice. Thinking from the perspective of a performer, I would appreciate the security of having some basic material to work with, which should sustain the momentum of musical dialogue within my ensemble. Depending on my instrument and level of technical mastery, I would also enjoy having the possibility of virtuosic extemporisation as an add-on and not an imperative. Granted, then, more subsequent material (than what is currently presented in Scenario 1) or further repeats can be provided to allow greater room for even more elaborate and varied treatments of the fragment, especially with a larger ensemble.

Besides the Upper part, the Lower part can also contribute to effecting interesting variations of harmonic colour. In particular, its ascending line can be repeated indefinitely in the manner of a ground bass, as in a passacaglia. This idea is informed by my study of the first (‘Black is the colour…’) of Berio’s Folk Songs (1964) for mezzo-soprano and seven instruments, although it should be emphasised that Berio’s use of repetition serves a different musical purpose, in a markedly different context.

In his song, Berio juxtaposes a fiddle-like viola part (notated without a time signature) over duple-metered voice and harp parts. The viola part is given a tempo ( = ca. 72, subsequently  = ca. 90-94) faster than that of the voice and harp ( = 54), allowing it to be indipendente dal canta. For each of the song’s three vocal stanzas, Berio provides the violist with a precisely notated section of material, to be repeated until the end of the stanza. Based on the metronome markings indicated, the violist would complete a cycle of each of the sections within the three respective stanzas at the points rightly proportionally notated on the score, allowing the violist five to six bars of ‘free reign’ before the singer finishes the stanza’s text each time. In the final stanza, Berio permits greater variability by giving the violist pauses (  )—durations of considerable flexibility—and an instruction to slow down (starting from the second repeat) from  = 92 to  = 54. Notwithstanding, Berio’s use of indeterminacy is very minimal and highly controlled; the ‘spillover’ from the variability is reset at the end of each stanza. After all, the D-pedal (or drone) held by the violist throughout each of these sections guarantees harmonic continuity, regardless of the number of repeats, or rhythmic discrepancy.

To me, Berio’s use of indeterminacy (however limited) serves to free the violist from the constraint of having to play in synchrony with the singer and harpist, to achieve an expressive and rhythmically supple quality. From my aural perception, Berio also manages to create a sense of depth—the singer crooning in the foreground (almost as if she is accompanying herself on the harp), and a fiddler in the background (viola con sord). Personally, I appreciate how Berio’s design is justified and elegantly effective in the context of the ‘folk’ topic specific to his song.

Returning to repose III in light of the above observations, I discover an interesting effect which can be achieved by repeating the ‘ground bass’ from the C#, instead of the first crotchet rest. Except for Player no. 2, I can instruct every subsequent even-numbered performer to enter straight on the C# (i.e. skipping the initial two crotchet rests). As a result of this repetition scheme, every subsequent even-numbered performer would be displaced by two crotchet beats; moments of harmonic friction would materialise and dissipate. Collectively, the Upper and Lower parts would sound as if they, having started in synchrony, are gradually drifting further apart, each spinning out its own course. Again, the indeterminacy of this effect—and thus its aural prominence—arises organically from the number of performers involved.

Within the aesthetic of my piece, a textural diffusion as such is desirably imaginative and justified because it illustrates the implications of something being reposed—the fragment grows into its new context in a way that is familiar, yet possibly strangely uncomfortable. Unlike Berio, I maintain a constant tempo and pulse in the section being repeated; the rhythm of mine is also necessarily simpler than his, given that I allow the ‘spillover’ from the variability to manifest until the very end of my piece. At that point, all the odd-numbered players would have dropped out (except the final one), leaving an unsettled, diffused ‘ground bass’ texture going amongst the even-numbered players. I consider it meaningful to empower the performers with the responsibility of shaping the piece’s ‘decay’, hence I encourage them to continue playing while gradually getting softer and slower, until reaching complete silence. This way, they are free to end on any note of the section, at as slow a rate, and as softly as they deem appropriate.

Scenario 2

Further to how I have designed the Lower part in Scenario 1, I can slightly expand the range of variable outcomes by alternating the entries of the even-numbered performers (one enters on the crotchet rest, the next straight on the C#, etc) and adjusting the tail end pitches slightly each time. I also consider repeating the earlier entries of the Upper part until the end, such that the various levels of floridity co-exist. However, working these provisions into the score layout of Scenario 1 proves unsuccessful as the score becomes unnecessarily cluttered with written instructions. In seeking the clearest way of communicating this instruction to the performer visually, I think Lutosławski’s practice of splitting and aligning stave systems to specify points of entries is particularly instructive. As far as I am aware, many published scores of Lutosławski’s pieces, such as his String Quartet (1964; the same year as Berio’s Folk Songs), Mi-parti for Orchestra (1976), and his Third Symphony (1983) feature this distinct visual clarity. While briefly studying the score of Mi-parti—in particular the section from rehearsal Fig. 16 to 19—I recognise how Lutosławski’s notation is necessitated by the effect he intended to achieve.

At Fig. 16, an intensely dense mass of registrally close, rhythmically intricate woodwind and brass sounds materialises in relief from a chromatic cluster chord (with all twelve pitches of the chromatic scale) in the strings. Each instrument contributing to the sound mass is given a fragment of material to be repeated indefinitely, until the next downbeat given by the conductor, upon which they proceed to the next given fragment. Lutosławski gradually brings the scuttling rhythms of the sound mass to a halt by shortening each successive fragment, and appending a long note to the final one. Arguably, this is the easiest, clearest possible, and most elegant way of designing and presenting this musical process.[5] It allows the impression of a stochastic bustle to arise directly from very slight misalignments of the repeated fragments, as each performer would execute every phrase and caesurae at a marginally different rate. This design also frees up the conductor from having to grapple with an otherwise rhythmically complex score, and empowers him or her with a degree of control over the actual shape of the composition. Furthermore, as a composer studying Lutosławski’s notation, I can clearly perceive the modular fashion with which he organises and activates his sound events; his notation ostensibly shapes his aesthetic.

While I certainly do not intend to imitate the aforementioned process in repose III, a Lutosławskian score layout is nonetheless greatly helpful in specifying the alternating entry points of successive Lower parts. This alternation would grow a considerably intricate harmonic texture, abound with ‘ghosting’ effects, especially towards the end of the piece:

repose III - Ex 2
Ex. 2 – ‘Written-out simulation’ of the ongoing Lower parts, towards the end of Scenario 2 (5 performers)

With this layout, the aural cues provided in my piece seem to be adequately clear for the chamber performers to enter independently without the need for a conductor. This newfound visual clarity also offers the possibility of starting each successive Upper part as an enjambment on the final bar of the one preceding it. This happens as the increasingly florid lines are being juxtaposed on simpler ones; the musical activity is bound to attain a rich texture with rococo melodic runs, since the performers do not drop out until the very end. By slightly altering the tail ends of each repeated section in the Upper and Lower parts, I am also able to achieve an overall registral arch shape, which reflects and magnifies that of the original fragment. Scenario 2 ends in a way similar to Scenario 1.

Unlike Scenario 1, however, adding more players here does not serve to increase textural complexity as much as volume, since additional players merely double a part of their choice (be it whether they enter simultaneously with their double). At this point, I also realise that my design in Scenario 2 is unnecessarily predicated on an ensemble size of at least ten performers; any fewer would mean that the more florid Upper parts of rehearsal Figs. D and E are not reached. This reveals an inherent rigidity in my design, which I try to address in the following scenario.


Scenario 3

Re-studying Riley’s In C (1964) made me realise that the range of textural, rhythmic, and harmonic variability can be increased simply by shortening the length of each repeated section (termed ‘pattern’ in my score). An expanded scope of permutations and combinations implies that my piece’s structure becomes more malleable; the performers’ extemporaneous decisions now play a greater role in deciding the piece’s ultimate shape. However, this necessitates a specific way of retaining control in design, such that my piece maintains a fundamental degree of consistency from one performance to the next (at worst, it does not liquidate under poor handling). As already insightfully discussed by American composer Robert Carl, Riley retains that control in In C by fixing the order of the patterns, which guarantees: (1) an arch shape of harmonic density and rhythmic activity, (2) the realisation of a pre-designed motivic transformation, and (3) a modal progression from C-Ionian to G-Dorian.[6] What varies between performances of In C, then, is the durational proportion with which Riley’s design is realised—it can be stretched or compressed like a slinky. In this light, I recognise that Riley uses indeterminacy primarily as a way of explicitly furthering the performers’ interactions with, and awareness of each other.

For me, repose III differs from In C in that the patterns from which performers get to choose at rehearsal Fig. B are essentially florid (heterophonic) variations on the same basic fragment; unlike Riley, I do not transform the motif. I feel sufficiently justified in providing the performers with (what may be termed) a ‘hierarchical framework of floridity’ to operate freely within; in addition to octave transpositions of the patterns being allowed, they are also encouraged to add ornaments and vary rhythms slightly. By suggesting unrealised possibilities, the layout of my framework appears to be an effective visual stimulus in spurring the performers towards further personal engagement with the reposed fragment: what comes below the most elaborate pattern given at the bottom? …and below it?

Considering that the opening iteration of the original fragment has not been worked into my framework (unlike Riley’s guaranteed motivic transformation), I perceive the need for a larger frame which would serve the function of beginning and concluding the piece. Having started the piece in Scenarios 1 and 2 with just one player, I still find it helpful to assign leadership to a performer, who is given the task of starting the piece, bringing performers into Fig. B, and handling the dropping out at the end. Furthermore, the two-part ‘homophony’—necessitated by the piece’s derivation from Bach’s fragment—should ideally carry on underneath the repeated patterns at Fig. B, in the manner of a ‘ground bass’. To this end, I allocate two to three performers (possibly more, in a larger ensemble) to repeat the ‘ground bass’.

A final remark on this Scenario: shortening the patterns opens up the pragmatic possibility of composing technically idiomatic material in consultation with performers, giving them particular moments to emerge from the collective texture with something unique to play (especially chords for the harp, and keyboard instruments). This is also possible in Scenario 2 by assigning each Upper part to one instrument, but the brevity of the repeated patterns in Scenario 3 promises greater dynamism in the materialisation of these moments.

*        *        *

For the rehearsals, I intend to bring along the piece in a form very similar to that in Scenario 3, with a keen eye on the outcome of the ‘hierarchical framework of floridity’. In particular, I think that the performers’ collective experience of the piece’s rehearsal—especially how I explain its framework to them—would play a significant role in influencing their willingness to engage with it. Upon grasping how each of them would directly affect the ‘comprovisation’,[7] I hope that they jump in to ‘compose’ the piece together with me… or, with Bach.

(I am in the process of writing a follow-up to this essay, reflecting on the eventual form of repose III as premiered in the Consort 21 showcase on 16 Mar 2018.)



[1] Being aware of the eschewal of first-person pronouns within certain academic circles, I nonetheless use them consistently and consciously in this essay while maintaining an acceptable degree of literary formality, to reflect the self-reflexive nature of compositional discourse.

[2] L. Foss, ‘Improvisation versus Composition’, The Musical Times, Vol. 103, No. 1436 (Oct., 1962), 685. In his article, Foss expresses a similar observation in discussing the importance of practicing improvisation with his collaborating partners.

[3] M. Dirst, ‘Bach for whom? Modes of interpretation and performance, 1820-1850’, chp. 6 in Engaging Bach: The Keyboard Legacy from Marpurg to Mendelssohn (Cambridge, 2012), 144.

[4] C. Welch, ‘Programming Machines and People: Techniques for Live Improvisation with Electronics’, Leonardo Music Journal, Vol. 20, Improvisation (2010), 28. In concluding his article, Welch likewise welcomes these elements into his practice.

[5] Similar instances can be observed subsequently in the same piece at rehearsal Figs. 39-40, and 44-53 (end).

[6] R. Carl, ‘Analysis’, chp. 4 in Terry Riley’s In C (New York, 2009), 57-70.

[7] I borrow this term from R. Dudas, ‘ “Comprovisation”: The Various Facets of Composed Improvisation within Interactive Performance Systems’, Leonardo Music Journal, Vol. 20, Improvisation (2010), 29-31.


Bibliography (works cited):

Dirst, Matthew, ‘Bach for whom? Modes of interpretation and performance, 1820-1850’, chp. 6 in Engaging Bach: The Keyboard Legacy from Marpurg to Mendelssohn (Cambridge, 2012), 143-168.

Dudas, Richard, ‘ “Comprovisation”: The Various Facets of Composed Improvisation within Interactive Performance Systems’, Leonardo Music Journal, Vol. 20, Improvisation (2010), 29-31.

Carl, Robert, ‘Analysis’, chp. 4 in Terry Riley’s In C (New York, 2009), 57-70.

Foss, Lukas, ‘Improvisation versus Composition’, The Musical Times, Vol. 103, No. 1436 (Oct., 1962), 684-685.

Welch, Chapman, ‘Programming Machines and People: Techniques for Live Improvisation with Electronics’, Leonardo Music Journal, Vol. 20, Improvisation (2010), 25-28.