Shadows of shadows
David Toop
August 2020
Foreshadowing
In Oslo; in the physical world. Immersed as we were in discussions about the ethics, mechanics and feeling of collaborating with non-human entities (or processes, to be more accurate), it never crossed our minds that playing together, or even being human together, in a room would become impossible. So now, from the midst of the coronavirus pandemic, it becomes impossible to write about those experiences without a taste of their luxuriousness lingering in the mouth. By luxuriousness I mean not expensive fabrics, jewellery or foods but that simple act that we took for granted: to be in close proximity in a room together.
If my intention was to write in detail about Kim-Auto, the name given by the Goodbye Intuition research project to the algorithms that absorb sounds made by improvisers and return them in more or less unpredictable forms, that was also sabotaged by the pandemic. A handful of meetings, fascinating and enjoyable as they were, were insufficient for any deep insights into how Kim-Auto was working and what it might do. At most, I was able to shadow the shadower, taking notes as I did so. These initial observations were written up into a short text during one of the LAB and discussion sessions held at the Norwegian Academy of Music. I have chosen to include an adapted version of the notes within this essay because they formed a direct response to what I was hearing. In that sense they provide a closer, more immediate reflection than the more contextual approach into which they have been inserted here. Given the limitations imposed by the pandemic I have chosen to write in broad terms about certain ways in which improvisation is a multiplicity of simultaneous voices, always shadowed and reshaped by those non-human elements that are fundamental to the practice of working with sound. Hopefully these reflections may contribute something to our understanding of machine learning and its potentiality in relation to the practice of free improvisation.
Shadowing
Improvisation is an essentially collaborative practice. Contrary to that and central to its ethos is the high estimation given to individuality. Players bring individualistic, sometimes unique techniques, strategies, personalities and understandings of improvisation to the group, and yet for a group music to emerge, that individuality has to be tempered by the group’s needs. This may be why solo playing has flourished within free improvisation, for the freedom that it allows away from the group – a kind of holiday from responsibility – which is countered by the extra responsibility of maintaining momentum, shape and interest without the added stimulation and correctives of other players.
Experience demonstrates that for improvisation to remain dynamic and fresh its basic requirement is an integrated structure of three elements: the active individual player/listener, who works within the active group of player/listeners, which in turn works in relation to the inactive group of listeners (what we call the audience). Already there are omissions and difficulties with this triangular model. Is an audience inactive, or simply active in those invisible regions in which close listening, following and structuring operate with intensity and deep concentration? Is it even possible to generalise in this way? Some groups or players within groups deploy techniques of mirroring, shadowing, quick responsiveness – a highly alert, even perpetually agitated music that demands some degree of instrumental virtuosity. Others seem to ignore each other’s contributions, playing as if there is nobody else in the room. Which is more attentive to the audience? Players adhering to the first model may be too focussed on their own relationships to be aware of an audience. The second may be hyperconscious of environmental conditions and react to whatever responsiveness, positive or negative, emanates from the audience, the space or other fluctuations.
As an example (picked almost at random) of the relational mysteries that are typically atypical of improvisation, take a performance filmed at Kato, Berlin, in 1986, a duo of vocalist Phil Minton and percussionist Sven-Åke Johansson (available here:). Minton is a veteran of UK improvisation, active since the 1960s; Johansson is one of the pioneers of the music, noted not least for his drumming on Peter Brötzmann’s debut trio recording, For Adolph Sax (1967) and the landmark LP Machine Gun (1968). From this depth of experience alone, we can assume that their actions and reactions are intentional consequences of intensive listening. At the beginning, Minton is more active, a low-throat belching glossolalia; eyes closed, he pulls off the sponge foam microphone cover, presumably in search of some higher frequencies or even the touch of metal against his lips. Johansson’s touch on drums and cymbals is soft, sparse. They shadow each closely, never obvious. Their closeness is evident from a mutual observance of brief silences, a pausing between spasms. At 3’ 30” Minton becomes quiet, tender in keening and wailing. A split second earlier, Johansson had picked up two sheets of sandpaper. He commits to this instrument, the razor-sharp cut of it acting as a ritual marker to Minton’s lamentation. Tiny acoustic variations come into play as Johansson moves a few steps, back and forth, and as he rubs sandpaper on a cymbal the direction switches again. Minton is now the audible voice of interior nonsense. Johansson picks up a white cloth, folds it, flops it against cymbals, drums, but then as he returns to the cymbal the mood suddenly comes into focus, a deeply satisfying sound like flour dropping into a bowl, again a sudden microscopy not just of this specific auditory and visual event and its qualities but of the tactility and warmth it creates, an aura around itself, and Minton responds by inhaling as if physically drawing this warmth into his body, moving back close to the microphone to exhale inner mouth sounds of saliva and gums. The worlds of work, the body, materials, intricate spaces, consciousness and cultural practices are inferred as possibilities, as if in passing. All of these overlapping, non-linear sequences take place in the first five minutes of a twenty minutes performance of constant variation, ending with a kind of monologue by Johansson and a little song – “Time is running” – rather timidly performed, which makes the audience laugh, so drawing them with an act of generosity into the metaphorical firelight of performance and its fragility.
Time is running: the long shadow
Given these complexities, ambiguities and subtleties, why would it be considered necessary to build an algorithmic computer program that can improvise with human players by learning aspects of their playing style, responding and intervening on its own terms with a sound world that both shadows and extends its human collaborator? There are multiple histories attached to this – a history of inserting randomness into the playing situation, a history of collaborating with non-human entities, a history of autonomous machines which anticipate automated labour and AI but also speak to the desire of humans to play God and create life itself. As a way of unlocking at least some of these complexities and ambiguities, it may be useful to consider improvisation as a struggle within and beyond the self – struggles with habit, with learned techniques and routines, with the boundaries of socialisation, with limitations of intuition, imagination and instruments – and to examine ways in which this gravitational pull toward familiarity has been destabilised in different cultural settings and earlier periods of history.
Spirit shadows
Humans fear and are drawn to that which is not human; fear is articulated through imitation, glamourisation, control, exploitation, violence and extermination. The shadow world of non-human entities is a recurring theme throughout all we know of a history of listening and its representation in sound worlds, music and music technology.
From the deep shadows of caves humans looked (and listened) to those creatures who possessed all the qualities that slow, weak, two-legged, earthbound beings lacked. Somewhere between 43,000 and 35,000 years ago, bones from birds and mammals were fashioned into flutes, so the first sound-making technology of which we have any knowledge combined human ingenuity with raw materials taken from extra-human worlds of animals in death, hollow archives of the cries and songs once emitted from their throats. A supernatural world came into being, preserved on the walls of those caves, a domain of animal-human shamans and hunters, and since sound had mysterious qualities its efficacy as a signal from the spirits came to permeate all cultures in all ages.
Special sounds were devised and reserved for the voices of spirits, using unusual techniques and technologies. To give one example: in Gabon a recording was made in 1968 by French ethnomusicologist Pierre Sallée. He was recording the Bwiti ritual of the Mitsogho people. At one stage during initiatory rites of mourning a spirit named Ya Mwéï, the muddy monster, is invoked. “Mwéï is represented only by an unseen masked voice,” Sallée wrote, “a kind of raucous tiger’s roar uttered in a throaty voice by a hidden person whose pharynx has been irritated ahead of time by the consumption of a brew made of the leaves of a tiliacea.” At the end of this recording a dry drum beat sounds out what is understood by initiates to be the flicking of Mwéï’s tail.
Space in shadow; a shadow is a reflection of a different shape
Ya Mwéï is hidden but the disembodied voice manifests in shared air, in the form of an assemblage instrument of at least three parts: the human pharynx, the plant tea that acts as a throat irritant and the drum, in itself a hybrid of human hands, tree and animal skin. According to Sallée, Ya Mwéï is the “. . . tutelary spirit associated with the history of the tribe.” The spirit is a kind of archive who creates temporal and physical spaces in which transgressors may be swallowed in the same way that the spirit swallows the dead.
This is not dissimilar to how we might think about improvisation, always dealing with the archives of what has gone before, the building of a performance space and/as instruments, the necessity of negotiating remarkable or unremarkable problematic physical spaces in the moment, all of which act as non-human entities on the emergence of music. In 1995 I asked Pauline Oliveros to speak about her experiences of improvising in extremes of acoustic reverberation. “. . . these two words came together – deep listening,” she said, “because we had a very challenging space to create music in, when you have forty-five seconds of reverberation coming back at you. The sound is so well mirrored, so to speak, that it’s hard to tell direct sound from the reflected sound and that’s very challenging, so it puts you in the deep listening space.” For Oliveros, the behaviour of sound in such spaces gave rise to the idea that space itself could be otherworldly, a setting which facilitated what she called “. . . communion with other worlds.”
Improvising saxophonist John Butcher has worked frequently in such environments. I asked him to respond to some general enquiries into the role that reverberant spaces can play in a solo improvisation, including a question of whether such playing was a solo within liveliness or a duo between radically different forms of life. “I think Improvisation has to be about continual adjustment to the reality of a situation,” he replied.
“So an improviser is especially able to make music for a space and not just import music into a space. But, before getting into that - something about the physical side of playing saxophone. I'm sure the degree of reverberation in a space affects how the reed (the actual source of the sound) vibrates in my mouth. So the nature of how the instrument plays is intimately connected to where it is. When an acoustic player says ‘this room's impossible, it's so dry’, it's not just because they're frustrated by what they hear. The instrument physically plays differently than in a resonant room. So perhaps the person/body/instrument/space is/are the true instrument and it's a mistake to think of them reductively.”
In February 2019, myself, John Butcher and Lucie Štepánková performed as a trio within one room of a Mario Merz exhibition in Pirelli HangarBicocca, Milan. The impressive reverberation and imposing dimensions of this vast ex-locomotive factory, combined with the suggestive impressions created by Merz’s igloos and neon numbers, all combined to affect the shape of our improvised performance, not least from our altered feeling of time and the cumulative resonance of atmosphere and acoustics.
Butcher’s first encounter with an extreme acoustic that asserted itself as a partner to improvisation, rather than a backdrop, was in the Oya Stone Museum, Japan. “At the time it felt like me and it,” he says. “It began as a struggle – but during the course of playing I realised I had to work with the space and try to make a piece for it. So I felt both the things you mentioned - initially a different-life-forms duo then a symbiotic solo. Also it was very cold – 30 metres underground (making the sax uncomfortable to hold, and breathing unpleasant) and the audience seemed a long way away. I didn't really enjoy it - but, a few months later, I heard the recording, which really had something happening, and got interested in this way of working.” The attraction of consciously working with extreme acoustics as a collaborator, he concludes, is because “. . . the music only exists in that form because the space existed in its particular form.”
The long shadow of intuition
As the examples given by John Butcher indicate, a relationship between his playing and the altering effect of an extreme acoustic could only have become possible with intuition. His conscious thoughts in the depths of the Oya Stone Museum led him to believe that the situation was unpromising yet his intuition led him a strategy that a retrospective auditioning of the recording showed to be positive.
From this point of view it seems a little bit premature to be saying goodbye to intuition. Maybe a trial separation would be better in the long term, making some space to ask big questions about what this process we call improvisation involves. Of course we know by now that it’s not free, that it grows from a continually growing base of experience, practice, empathy (which must be learned and cultivated), the layering of many histories, personal and collective psychology, a deep time of strata that can only be accessed in the moment by this other process that we call intuition. There is also a relationship to objects which is strange to say the least. The materials used by an improviser – voice, breath, muscles, joints, instruments, objects, processes, time, acoustics, contingency, resonance – seem to never end. They are perpetually in development and in a sense they can be thought of as living entities in formation of an assemblage. As John Butcher suggests, this is the real instrument. This raises all kinds of problems, of course, and those problems become even more complex when the materials possess certain qualities that we describe as artificial intelligence.
But it seems to me that all objects have intelligence in that they wish to do what they wish to do. Their properties are insistent. During a performance I tried to split a length of bamboo. Actually I didn’t want to split the bamboo as an act of violence. I wanted to hear the gradual and subtle sounds of matter changing its physical structure. The bamboo is immensely strong and supple so it resisted, threatening to damage my hands. In the end, the bamboo gave way because that is its nature also. It bends as an aspect of its own strength.
When the Goodbye Intuition research group gathered together in September 2018 I spoke about naughty children. I was thinking about the role of dynamism in improvisation, what comes back at you and how this constant dynamism is entrained within varying notions of coherence. In the improvising context, coherence can be read as a form of continuity – listening experiences that seems to connect in various ways or actions which persist and bind together as engaging or absorbing processes. Then I wrote in my notebook: “The machines don’t care about continuity.” This word – machines – is maybe regrettable, something nineteenth century about it; there has to be a better way to think about this. But also it’s not strictly true. The so-called machine does care about continuity in its unknowing way. It has a way of behaving, a rather pathetic need for energy, a process, all of these needs speak to continuity. This is closer to the naughty child, who can exhibit all of these characteristics yet may be wilful, wild and incomprehensible according to adult definitions. Often they will imitate adults in order to infuriate and disrupt.
Sidsel Endresen asked if there could be a music without time and this question hung in the air (still hangs in the air) as one of the most tantalising possibilities of all. What if these creatures – let’s call them creatures rather than machines or children – create a temporal reality that humans can’t fully understand yet they can work with it. Perhaps this is also true of improvisation. Its relation to time can be so complex, shaving down time into microparticles yet within the same process and moment operating in long arcs, no way to know the temporal subjectivities entangled within the group or audience. But time, of course, is a familiar recourse when automata, robots or synthetic beings are compared with human life. “Time is wrapped up in the android,” writes Gaby Wood in Living Dolls, “in a way that is directly opposed to the way in which we are wrapped up in time. Man is subject to time, to its inevitable march towards death, whereas the clockwork automaton merely marks time without falling prey to it.”
Speaking of time, I work often these days with younger players whose experience is much shorter than mine, their practice situated on the margins of improvisation. Their perception of what is going on or what this music means to them is not at all clear to me. Only that they want to do it. That mystery constitutes a valuable part of the relationship, an accepted unknowingness acting as a reminder of the impenetrable other that is other people, other entities, objects and the remote constellation of planets living inside us that we call self. The attraction, of course, is freshness and all that is unexpected and not yet weary. That has its dubious, potentially dark side, but on the other hand, improvisation situations that interest me the least are confederations of veterans with close to a century of shared knowledge, those struggles of avowed spontaneity and rule breaking that in practice have the attritional feeling of being stuck in the trenches of a particularly narrow form of humanism.
Having said that, the emotional temperature of improvisation can’t be discounted. It influences the form and direction of the music. Direction is a poor choice of words, implying as it does a singular, time’s-arrow approach to time, whereas a rising temperature creates thermals on which the music is lifted, circulates, spins through many configurations. A cool temperature allows consideration and care in the treatment of pace, space, the potentials within materials. Small volumes are given permission to relax and expand. In this sense, improvisation is a weather system, infinitely complex and contingent, sometimes luxuriant, sometimes hostile, even impossible.
So the question arises: does the computer, thinking-creature, possess this emotional temperature? Does it register excitement or calm? Children are supposedly naughty because they need to find the extent of themselves and their world, or because they ate too much sugar, or because they encountered some random frustration but lack the skills to negotiate it. I’m reminded of Jakob Johanne von Uexküll and his description of the Umwelt, the world within which a creature perceives itself as existing. The tick waits patiently (already an anthropomorphism because the tick is outside a time in which waiting or patience are possible), then a warm-blooded smelling material arrives within its scope of action. It leaps, attaches and feeds. This is a relationship, strange but basic.
Overshadowed
The difficult ethical questions are audible. Are we making something together? Do you trust me? Who does this belong to? Is the power moving around or is it stuck in one place? These can be heard as sound and form. These ethical questions become a little more knotted when a non-human entity is introduced. During one meeting I listened to Sidsel’s voice emerging from the piano – “hello, hello, hello” it sang – and I thought of a tiny Lionel Ritchie, trapped inside a piano. The piano is also a non-human entity, a thinking machine ancestor to the computer as well as being domestic furniture, so this thought of humans inhabiting furniture is a slightly spooky version of Uexküll’s tick, falling from a leaf into a warm body. It reminds me also of Edogawa Ranpo’s horrible short story about a man who wants to be close to a woman so he stitches himself into one of her chairs.
This is the uncanny, distasteful aspect. The productive aspect is to think about ecologies, of entities inhabiting the same space in a web of relationships, not all of them transparent or easily comprehensible but nevertheless interdependent. This seems to me to be the most productive starting point for moving towards an understanding of what is going on. I thought of certain moments, a duo of Andrea Neumann and KA, and then another duo of Sidsel and KA. In the first, Andrea played precise, delicate sounds, all of which were gradually overshadowed by KA’s crashing responses, louder and louder. Sidsel began with beautiful interactions with KA; as if incapable of maintaining a sensitive dialogue, KA regressed into crass interruptions, at which Sidsel retreated into silence. Afterwards she commented: “It’s a cold customer to work with.” At the time, I wrote: A kind of murmuring machine sometimes; then a vomiting machine.
Perhaps some of us are careful about judging other humans or non-human animals, less careful about judging a freezing cold cave and even less careful about judging an algorithm that responds to our sensitivity with nonsense. All of these responses are anthropomorphic; we look for humans, even when we know they are not there. As the coronavirus pandemic of 2020 continues into an uncertain future, we remain starved of familiar human contact and group playing opportunities. Any considered evaluation of an ‘improvising machine’ is made difficult by these extraordinary conditions (particularly given the backdrop of Artificial Intelligence that potentially threatens all of us) , and yet machines are part of the twenty-first century ecology of entities and conditions within which we work. Our obligation is to speak with them.
