Chance and Indeterminacy in Experimental Composition
Chance and indeterminacy are two different engines driving experimental composition, and confusing them blurs both the history and the listening experience. Chance — often called the aleatory — hands choices to dice, coins, charts, lotteries or random number generators. Indeterminacy leaves the choice to the performer at the moment of performance, bounded by rules the composer writes down in advance. The distinction is not pedantry. It decides where responsibility sits, what a score asks of the person standing in front of it, and whether a listener hears the result as composed or merely lucky.
The idea has a long pedigree. Mozart turned to dice for a musical entertainment in the 1780s. John Cage, working with percussion and chance operations in the late 1940s, made the procedure a philosophical programme, drawing I Ching hexagrams for Variations I and pitch and duration tables for Music of Changes. Pierre Boulez, at the same moment but from a different premise, wanted performers to decide inside a tightly controlled field, and Stockhausen pushed further, proposing a music in which the composer sets conditions and the performer supplies the content. By the 1960s the question was no longer whether chance could be used, but who should be trusted with it.
It matters here because chance music reaches most Australians indirectly: through a specialist mail-order catalogue of vinyl, cassette and archival releases, a weekly community radio slot in Melbourne, Brisbane or Canberra, a festival program staged in a converted warehouse in Fitzroy, or the second-hand copies that turn up on a Saturday morning among the trestle tables of a market. Each route carries a different idea of what an experimental record is for. What follows looks at the mechanics of chance and indeterminacy, then at what they mean for listeners, performers and collectors working under Australian conditions — including the legal questions that arise when a performance is composed as it happens.
Chance and indeterminacy are not the same thing
Stockhausen's categories in Towards a Cosmic Music, published in 1966, remain a useful map: composition, where the composer determines everything; a machine, where the mechanism determines the output; contingency, where the composer decides some elements and leaves others open; and pure chance, where only the rules are composed. Cage's aleatory music sits at the far end, the artist stepping aside so that taste cannot interfere. Boulez's Points and the third Piano Sonata work in the middle ground, letting the performer choose from materials fully specified in advance.
The differences are audible. In an aleatory piece the surprise is guaranteed by statistics: something arrives that was not predicted, because the procedure guarantees irregularity. In an indeterminate piece the surprise depends on skill, taste and the hour. One performs a work whose character is fixed; the other performs a work whose character is decided in the room. Feldman offers a third case, with nothing left to chance and no repeated patterns, yet music that can sound entirely improvised.
There is a social dimension too. Aleatory music is often read as democratic, since the dice are indifferent to who rolls them, and Cage made that indifference the point. Indeterminate music can be more hierarchical, because it assumes a performer able to realise a demanding brief. The argument resurfaces whenever a festival programmer commissions an ensemble to play a score that refuses to settle.
Dice, charts and the limits of randomness
The classic toolkit is small. Cage used the I Ching, sixty-four hexagrams built from broken and unbroken lines, to generate the sound events of Variations I; the hexagram names became movement titles. Stockhausen's Plus-Minus, written for a seven-player ensemble in the early 1960s, uses a probability table in which a cross means rare, a plus sign moderately likely and a minus sign frequent, so the score reads like a statistical weather report. Lutosławski's Adventus lets the conductor draw from pre-composed fragments in the final minutes of rehearsal, and Cage's Stop offers a pianist nineteen possible procedures with the instruction to stop after any length of time.
Electronic and computer-based practice widened the range and sharpened an old problem. A composer might push a hundred thousand samples through a constraint solver, place stochastic transients with a sequencer, or improvise with live coding in front of an audience. But a computer's random number generator is not truly random: seeded, it repeats, and even unseeded its sequence becomes reproducible once the machine has chosen. The claim that the machine frees the artist from taste needs qualifying, because the programmer has already decided what the procedure is allowed to decide.
For performers, the difference between chance and an honest attempt at unpredictability matters. A score saying "choose any pitch from this list" asks for a decision; a score saying "improvise freely" asks for a performance. Conservatoriums and community ensembles in Australia tend to treat the second as an excuse for unpreparedness, and rehearse the first like any other difficult repertoire. The improvised music cultures of Melbourne and Brisbane, where chance-based scores sit more comfortably, have absorbed that habit rather than fought it.
The score as an agreement
When a score delegates decisions, it becomes closer to a contract than a description. Cardew's Treatise, published in 1963, offers nearly two hundred pages of geometric shapes, curves and spirals to be traced with whatever line the player has, at whatever speed the system permits, with the explicit instruction that the result should not sound like what it looks like. Earle Brown's December 1952, written for three players, is read in four seconds and then played, with choices made once and for all by the people in the room.
That transfer of responsibility reaches beyond the studio. Under the Copyright Act 1968, the author of a sound recording is generally the owner of the instrument on which the work was performed, with special provisions for films, published editions and broadcasts. For a recording of a fixed work the attribution is obvious: the composer wrote it, the soloist played it. For a recording of an indeterminate work, where the notes were chosen at the moment of performance, the question of who composed the performance, and who is owed the performance royalty, becomes genuinely unsettled. APRA AMCOS and PPCA collect those rights from named parties.
The same question shadows the movement of repertoire. If a piece changes every time it is played, what exactly is licensed, reproduced and sold? A conservatorium may buy a score once and expect to mount it; a festival may want a recording for broadcast; a label may want the take on vinyl. Australian institutions treat these as administrative categories rather than artistic ones, but the categories were built with fixed works in mind, and the gap between a document's language and a piece's behaviour is where curators, archivists and copyright lawyers work most carefully.
Tape, radio and the documented accident
Before the score there was the tape. Musique concrète gave composers material to manipulate directly, and the BBC Radiophonic Workshop gave the technique a household name in Australia, with ABC school broadcasts carrying British radiophonic techniques into classrooms where children heard oscillators and filters producing sounds nobody had named. Tape loops make chance physical: once a loop is closed, the tape has decided when the next event arrives, and the performer negotiates with a machine keeping its own time. Cage's distinction between the accidental and the intent runs through this territory, where an error on a reel can be the most meaningful event in the piece.
Sampling inherits the ambiguity, and it is where the law has done the most work. Australia has no general sampling exception; the flexible dealing provisions of the Copyright Act, introduced through the Digital Republic Agreement of 2013, give some room for transformative use, and the reference group on copyright and artificial intelligence reported in 2025 in favour of a statutory text and data mining exception with transparency obligations, a proposal with clear consequences for anyone training models on large public archives of experimental music. Chance and machine learning meet here: a generative system that assembles fragments is making decisions, and someone is answerable for them.
Print has its own role, since chance procedures need documenting: the charts, the tables, photographs of dice rolls, notes telling a future performer what was intended. A specialist journal makes the case for treating that printed layer as part of the work rather than packaging, and there is something apt about it, because when the music is partly an accident the booklet becomes the only surviving evidence of the method — and that makes the art of the booklet a strange kind of documentation, part score and part confession. For collectors browsing a catalogue of vinyl, cassettes, books and limited editions, that paper object is often where the story of a release begins, which is why editions with careful sleeve design feel like more than ephemera.
Chance in Australian studios, shops and radio
Australian experimental music has its own institutions. Room 40, run from Brisbane by Lawrence English, has built a catalogue around long-form improvisation, feedback systems and chance operations, issuing records in editions small enough to feel like correspondence. Elsewhere Records in Melbourne curates archival and contemporary electronic releases, and the city's OTHERS festival has made a habit of commissioning work that resists a fixed form. In Sydney, 2FB has given experimental and improvised music regular hours for decades, while 3RRR in Melbourne, 4EB in Brisbane and FFRFM in Canberra carry the same material to listeners who will never enter a gallery.
The physical infrastructure is equally specific. Record hunting tends to run through Saturday markets, with trestle tables at Flemington in Melbourne and Carriageworks at Eveleigh in Sydney, and through shops trading in used sleeves rather than new stock. Record Store Day Australia gives the ritual a fixed date in April, and specialist online catalogues fill gaps local shops cannot. Importing a rare edition from overseas means Australia's parallel import restrictions, which limit bringing in copies of recordings already available here at a commercial price, and GST on low-value imported goods, so the cheapest copy is often not the one you buy.
Australian labels releasing archives of experimental music navigate copyright rules routinely, from the statutory licensing of sound recordings to the ownership questions raised by indeterminate works. The National Film and Sound Archive in Canberra holds broadcast and music material that chance-based works have passed through, and its public playback sessions have turned odd preserved fragments into common property. For anyone buying a limited edition online, the Australian Consumer Law guarantees apply to the transaction as they do to any other purchase, even when the object is a one-off pressing of a work that will never sound the same way twice.
Chance and indeterminacy describe two different kinds of trust. Aleatory music trusts a procedure; indeterminate music trusts a person. Both remove taste from at least part of the process, and both make the listener more active, because the ear has to supply the judgement the score withheld. That is why these pieces resist the tidy summaries that reduce them to a technique or a quirk.
What stays with you should be this: the accident is a relationship between a rule and the person applying it. The dice, the chart, the loop and the performer each produce something that could not otherwise have existed. The composed part is the permission, the constraint and the decision to let something else finish the sentence. Chance, in other words, is not the opposite of composition. It is composition's way of making room.