This is an Accepted Manuscript of an article published by Taylor & Francis in Contemporary Music Review on 13 November 2013, available online:

When a score specifies a fixed instrumentation, it closes off certain sonic possibilities. By writing a piece for flute, for example, the set of possible sounds available is immediately reduced. While it is clear that this does not create a finite sonic resource–there are, after all, an infinite number of subtle variations available to the flautist–it does shape the resultant piece. Alternatively, when a score allows for an open instrumentation–where the choice of instruments is not fixed–the potential for a wider sonic resource becomes apparent. Neither of these situations is inherently better or worse: both provide their own constraints and freedoms. They do, however, have the potential to suggest different attitudes to working with sound.

One of my current interests as a composer is in the sonic properties of everyday materials and objects, and ways of using them as instruments. I am particularly interested in categories of materials, and the rich variety of sounds available when working with non-standard instrumentation. My aim as a composer when working with objects and materials is principally to explore this variety by allowing performers to experiment with the specific choice of instruments within a prescribed range. By doing so, I hope to investigate the quality of certain aspects of sound production within a musical context.

The approach outlined here focuses in particular on distributed decision-making, where performers are involved in selecting sounds, recognizing its potential for creating variety within given boundaries. It contrasts with centralized approaches to undertaking a detailed exploration of timbre in which a composer specifies the choice of materials and their playing techniques with precision. Typically such prescriptive approaches may include providing the dimensions of percussion instruments or non-standard objects, or the development of a notation method that carefully prescribes the control of actions (such as an extended stave notation or tablature). In James Dillon’s Windows and Canopies (1985) for example, a “large suspended crash cymbal (72 cm)” is specified. Some composers give more precise instructions however: in her piece dichroic seventeen (1998), Rebecca Saunders includes a tin can in the percussion part, with the requirement that it is “a large 1000 g empty coffee can (we used Dallmyr) without lid.” Largely, such approaches aim for reproducibility of a desired sound and/or action. They attempt to present the means to make very specific sounds, perhaps in order to replicate something the composer found through experimentation whilst composing the piece.

Using open instrumentation relocates this experimentation more explicitly to the rehearsal and performance of a piece. It requires the performers to make judgements with regards to the selection of sounds, normally within a set of relatively open constraints provided by the composer. For Brian Ferneyhough, using open instrumentation can be a way to focus on form. In his solo percussion piece Bone Alphabet (1991), the performer selects the instruments in what is otherwise a very prescriptive score. Ferneyhough (1991, pp. 443-444) comments

I opted not to choose the instruments myself; instead, I placed certain limitations on the choice available. For instance, no two adjacent instruments could belong to the same family of sonorities. … Not having particular sonorities in my mind as I composed (which actually involved a certain amount of disciplined renunciation) meant that I could concentrate entirely on the formal issues at hand.

In his earlier percussion duo Fanfare for Klaus Huber (1987), Ferneyhough approaches open instrumentation with a set of specific constraints. Here this is a way to create sonic variety, with independent choices by the two players creating unplanned correspondences. Ferneyhough (1991, p. 443) states that

two categories of timbre are defined: (1) sonorities capable of being grouped in reasonably homogeneous gamuts of high and low (like wood blocks, or tom-toms), which change every couple of measures, and (2) a whole series (over thirty, I think) of so-called ‘unique sounds’, each of which is to be played once only. The piece can be performed several times in the same concert … but each time a new set of unique sounds must be selected. Since each performer is instructed to select his instrumentarium without consulting the other player, it’s clear that all sorts of strange conjunctions could arise… I like the idea of totally unforeseen results.

A different use of distributed choice occurs in Mauricio Kagel’s Acustica (1968/1970), where the performers are asked to source and construct a set of found and invented instruments. These are documented in the score though a series of verbal descriptions, photographs, and diagrams, whilst allowing for variation within the specification provided. The modular score outlines specific actions to undertake using the instruments, with multiple instances of these actions potentially being realized by the performers. One group of actions involves adopting unconventional methods for eliciting sound from records by using different types of stylus (see Figure 1). Whilst there is a degree of variation created by the particular choice of stylus, the choice of record is left open. The actions and some of the materials are constrained, but by leaving open the medium to which they are applied Kagel invites the performers to consider the criteria for selecting the records.


Figure 1: Mauricio Kagel, Acustica (1968/1970), excerpt.

More open approaches are common in verbal scores, where the choice of sounds might be freer and contingent upon a different attitude to their selection. For example, in Michael Pisaro’s Only (2005-6), he specifies “an extremely long, very quiet tone”. There are many ways to realize this, and many possible instrumental choices would be suitable. Pisaro points to the attitude he wishes to instil in the person realizing the score, however, through his use of the indefinite article to specify the piece’s only performed sound:

…under those conditions to say “a sound” is a much more interesting way of approaching it, or was for me at least, than to write a whole-note B flat, which I could easily have done of course. But when you say “a sound” you might be saying “any sound”, but because it’s singular there’s also this pressure on the performer to think of what this “a” might be. “Any” kind of implies I’m going to find something and we’ll do it. “A” seems to say there’s nothing here and I have to put one sound in that place, and my experience with that so many times was I better feel good about that sound because I’m going to be doing that for the next five hours, you know in some of those early pieces. That’s all I have, that one thing. (Lely and Saunders, 2012, p. 321)

This obligates performers to find their own solutions to the instructions for sound making in the score, and to make a careful selection. There is no reproducible ideal to aim for: the attitude is specified, and a way to project that must be found. This is summarized by Joseph Kudirka’s observation about similar openness to sound selection in his own work:

An important distinction which isn’t spelled out often is that the “x equals anything” which is often the case […] isn’t a case of “x equals anything of everything which might be possible”. “x equals anything, but not everything”. So when there’s that openness there, and in my pieces, it’s not that everything would work. It’s not that every choice you could possibly make would work. Because there’s this importance that it’s taken into consideration and you find out what is going to work and there’s maybe an infinite number of possibilities, but an infinite set can still exclude things from itself. (Lely and Saunders, 2012, p. 241)

This engagement with sound selection by performers can also be seen in section III of Christian Wolff’s Burdocks (1970/1971), which asks that “Each player makes about 511 sounds, each sound different in some way.” This may be realized using many different sound sources, or deriving a wide variety of sounds from fewer sources. Howard Skempton notes in relation to this section of Burdocks that it “invites two very different approaches. One is strictly minimalist: for example, using a water-dropper and striking a glass, where the difference between sounds is guaranteed but almost certainly imperceptible. The other sees the piece as a call for inventiveness.” (personal communication, 2 July, 2012). The focus here is on finding differences, or allowing them to emerge; it is exploratory, and the performers are charged with a task. The performers are asked to source the sounds and/or the instruments, not the composer. It points to a distributed creativity, where multiplicity is harnessed to create sonic variety.

It is this approach to distributed sourcing of sonic resources that is at the heart of some of my recent music. Although most of my work has engaged with extended uses of conventional instruments and appropriation of everyday objects for sound-making purposes, it has only been in the past three years that I have begun to loosen up the precision with which such resources are specified. The first instance was in imperfections on the surface are occasionally apparent (2009), where my aim was to explore the difference between the sonic properties of surfaces when activated by dragging a standard cardboard coffee cup across them. The title is taken from Sol LeWitt’s statement Wall Drawings (1970) and is one of 13 pieces from a series I made from 2009-11 titled divisions that could be autonomous but that comprise the whole, also from a LeWitt text. All of the individual pieces take their title from artists’ statements, and all use a score format of multiple A4 landscape pages representing time windows of between 0’40”-1’20”, and containing specified actions (see Figure 2). There are two methods of producing sound specified in the piece–dragging and circling the cup–and the cup is used with either the rim or base in contact with the surface. Each performer is given five pages with actions, each lasting 1’20”, containing permutations of these actions, systematically spread across this duration such that no page is identical. They also each have five pages marked as silent. The pages are shuffled into a random order, so that the silences on the action pages and on the silent pages combine to space the sounds in an irregular and unpredictable manner. The result is a series of friction sounds that stop and start abruptly.


Figure 2: imperfections on the surface are occasionally apparent (2009), excerpt.

The 10 performers each use the same means of eliciting sounds from surfaces, but each performer must source five surfaces that are different to those of the other ensemble members. This allows for immediate comparison, and an ensemble of 50 different friction sounds. In practice, this is very hard to achieve, and the resultant negotiation of difference among the group is at the heart of the piece and what I wanted to explore. This is evident in the choices made by edges ensemble in a recent performance. Ensemble director Philip Thomas (personal communication, 3 January, 2012) explains how they prepared:

In terms of selecting the surfaces, firstly everyone turned up with a number of surfaces each. These we then listened to and discussed our favourite sounds and which sounds were too similar. Then we started discussing what ‘too’ similar meant, and this led us to discuss more in terms of how the speeds, pressure and articulation of the cup action affected the sound. We soon realized that we needed to adopt a very clear controlled cup movement so that the differences between the surfaces were more apparent and less the result of different types of movement. After a while it became like the Feldman ppp scenario–once you get accustomed to the sound of cups on surfaces you start hearing wildly different timbres within this.

The range of surfaces used in this performance included various simple materials, such as blocks of wood, metal trays, or types of paper, but also included more unusual objects such as a computer keyboard, wire cooling rack, bubble wrap, a polystyrene takeaway food box, a rubber hot water bottle, curtains, and a semi-pasted tile (see Figure 3). As well as the range of materials, each of the objects resonates in a different way: some are damped, often through connection to the table (which can be a problem), whilst others have a natural resonance. Texture too is important, with smooth, ridged, slatted, creased or woven surfaces all creating inconsistencies in the movement of the cup, as declared by the piece’s title.

Sample layout for Figures 3 and 6_Page_1

In addition to finding sonic richness, a distributed approach asks the players to really consider the sounds they make in relation to those made by others. The variety, on a moderate scale, is achieved by asking a group of people to each find a small number of sounds. This is perhaps likely to result in more idiosyncratic choices than a selection of 50 surfaces by one person.[1]

Figure 3: selected setups by members of edges ensemble in a performance of imperfections on the surface are occasionally apparent.

Asking performers to source instruments and sounds was also a central part of the process used in my piece for the Basel Sinfonietta, things whole and not whole (2011). My starting point for the piece was to create a cueing network modelled on research into the behaviour of flocking birds carried out by Craig Reynolds in the mid-1980s, which resulted in a computer simulation of flocking boids.[2] The movement of Reynolds’ boids was governed by three simple rules: a clumping force which keeps the flock together, a separation force which keeps them apart, and an ability to match velocity. The individual boids follow these rules, and the flock movement emerges from their interaction. In things whole and not whole, the interaction of the players is governed by rule-based cueing: players make sounds as soon as possible after those made by other players, choosing different reference players for each new sound they make.

This structure then needed to be articulated by sounds. I was very much aware of the imbalance of sound types across a standard symphony orchestra, with over half of the sounds being made by string instruments. I wanted to find a way to homogenize the palette without reducing the variety: it needed a more even spread of timbres. Initially I explored expanding the instrumental resource by allocating objects to each player alongside their regular instruments as I had done in a previous orchestra piece, geometria situs (2010), which went some way towards reducing the prominence of the string timbres.[3] I felt it needed to go beyond this however, so removed all the conventional orchestral instruments and began making a series of actions involving objects and materials. When making the score though, I was concerned about the arbitrary nature of my decision-making. Selecting pitches, playing techniques, and objects was somewhat unsystematic. I considered creating systems to make these decisions, but they too would be arbitrary. I needed a way to maximize the variety of sound selection without it being constrained by my own particular preferences and what I had to hand.

As with imperfections on the surface are occasionally apparent, distributed choice was the answer. There is a lot of potential in asking the 60 orchestral musicians to each select their sounds independently, so the final version asks them to do just that. Each player has to find 16 sounds, each of which should, as far as is possible, be different to those used by the other players. This results in 960 different sounds. I gave them some guidelines and rules to remove certain possibilities. I specified “noise sounds” and asked the players to “emphasize and prioritize noise components over pitch in each sound”, primarily to avoid focus on pitch patterning, which would have distorted the idea I had for the piece. I gave the performers the option to use any combination of their normal instruments and found materials, recognizing that some players might be resistant to abandoning something that was central to their identity as a musician (see Figure 4).

things whole and not whole complete (original Basel version)_Page_4

Figure 4: things whole and not whole (2011)

I attended the first rehearsal in Basel and was both surprised and pleased to see that virtually everyone had abandoned their instruments, and the stage was strewn with household debris and the occasional double bass or trombone. The decision by these players to use their instruments now felt like a conscious choice. There was a lot of variety: balloons, umbrellas, bottles, food, kitchen implements, sports equipment and building materials were abundant. In a break, one of the string players approached me and asked if it was acceptable to use a lump of uranium he had extracted from an open mine in Germany, and a Geiger counter. Suffice to say I would not have envisaged this as a possibility myself.

Asking the players to source sounds under guidance created the desired result. There was a lot of individual variety, yet a flock-like cohesiveness emerged: it produced localized differences, set within a global uniformity. It was a way to create the sonic resource I wanted, whilst reducing the potential for distortion through imposing the preferences of a single decision-maker. All choices made were subsumed into the whole, to the point where none had a functional prioritization over any other.

This situation was approached from the perspective of a single player in a piece composed at the same time as things whole and not whole. In 2010, I began working with percussionist Simon Limbrick on a solo piece that he wanted to be able to perform for 24 hours. My starting point was a piece I had begun in 2007, eventually titled with paper, in which the score is printed directly onto the instrument: any piece of paper. The actions described on each page elicit sound from the paper through touch, drawing, or cutting. The eventual piece I made for Simon, surfaces (2010-), extends this idea to any type of surface, and uses a series of 300 instruction cards to frame the performer’s activities (see Figure 5). Rather than specifying which surfaces to use in a performance, I defined actions which can be applied to any surface, and ask the performer to collect a wide variety of the required materials in advance. Here there is one player sourcing materials, but the repeated application of a large number of varied actions additionally transforms the initial resource into something else entirely: the processes defined in the piece create new surfaces.


Figure 5: surfaces (2011), selected score cards.

When writing the instructions, I decided to use the passive voice. This removes any sense of agency: it is not clear who should carry out the actions. This was a reference to Sol LeWitt, whose instructions for some of his Wall Drawings are presented in this manner. For example, in the certificate for Wall Drawing #960 (2001), he specifies that “A straight line about 18” (45.7 cm) is drawn”, before continuing to imply the passive voice in the remaining instructions. This avoids saying who should draw the line. LeWitt’s approach also demonstrates the richness possible when repeating a simple instruction whilst omitting certain structural constraints with regards to how the lines must be positioned on the wall, or any reference to the size of the wall. Similarly, in surfaces, by removing an active participant from the instructions, it opens up other ways of realizing the piece (by groups, through automation, recognizing external instances of the actions, and so on). The open specification of a surface also removes a specific context, allowing the scaling of each action relative to the chosen materials.

Sample layout for Figures 3 and 6_Page_2

The instructions vary in the level of detail they provide. For example, some actions involve just the specification of a surface, and a verb: “a [surface] is cut.” Here, the type of surface and the means of cutting it are left open to the performer. This allows for a number of different solutions, and engages the performer in assessing the sonic properties of available materials. In the course of a realization, new surfaces are also created through undertaking the actions, so this open specification enables these resultant materials to be used. Other instructions include nouns or additional clauses, normally specifying a constraint that refines the action (see Figure 6). Some actions require the use of other materials, specified in a similarly open way as “objects”, or with reference to their physical type, such as “liquid” or “granules”. The range of actions is diverse. It includes dropping, brushing, scoring, sawing, crumpling, attaching, grinding, pressing, pulping, stretching, tearing, folding, rubbing, and many others.

Figure 6: selected score cards from surfaces (2011), with documentation of their realization by Simon Limbrick.

While in this first realization of surfaces there was a single performer sourcing all of the materials, the instructions transform them through iterative processes. Surfaces are repeatedly cut, marked, torn, painted, and so on, to the point where all that remains is a pile of detritus. Any surfaces can be fed into the multiple processes that are defined in the score, with the performer’s selection and sequencing of materials and actions determining the outcome of the performance.

The three pieces presented here all deal with different approaches to the distributed sourcing of materials. All rely on choices by other people in realizing the pieces. This can of course be a precarious situation, given its susceptibility to abuse. In all three pieces however, the strategies I have adopted are reliant on different kinds of multiplicity, such that individual idiosyncrasies are averaged out by the overall mass to create an acceptably stable and predictable result.


[1]    A subsequent piece in this series, with the same material or still, to vary the material (2011), creates a similar situation but using bowed surfaces. It is scored for at least two players, each with a bow and nine different surfaces. At least one surface must be as close to identical as possible in appearance and sound for all players, and at least one surface must be different for all players.

[2]    See Craig Reynolds, ‘Boids: Background and Update’. Retrieved 25 June, 2012, from

[3]    geometria situs was commissioned by Südwestrundfunk and first performed by the SWR Sinfonieorchester at the Donaueschinger Musiktage on 17 October 2010. All the orchestral players need at least one auxiliary instrument that they must source independently. These are chosen from five specified categories: blown tube; bowed plastic cup; bowed polystyrene; bowed wood; and coffee cup on surface.


Dillon, J. (1985). Windows and Canopies. London: Edition Peters.

Ferneyhough, B. (1991). Interview with James Boros. In J. Boros & R. Toop (Eds.), Brian Ferneyhough – Collected Writings (pp. 431-446). Amsterdam: Harwood Academic Press.

Kagel, M. (1968/1970). Acustica. London: Universal Edition.

Kagel, M. (1975). Theatrum Instrumentorum. Köln: Kölnischer Kunstverein.

Lely, J. & Saunders, J. (2012). Word Events: Perspectives on Verbal Notation. New York: Continuum.

Reynolds, C. (1995). Boids: Background and Update. Craig Reynolds. Retrieved 25 June, 2012, from

Saunders, R. (1998). dichroic seventeen. London: Edition Peters.

Wolff, C. (1970/1971). Burdocks. New York: Peters Edition.