Consider the use of gesture recognition in interactive performance as a form of temporal linkage between the shifting intentions and actions of performers across time. Following Bergson, let us think of what is being linked here (past and present) as not static or point-like, but as a multiplicity of unfolding movements.
Formalising such links requires ‘chopping up’ complex interpenetrating movements which are distributed between people in space and time. But what then are these pieces we chop ourselves and others into? Generally we might refer to them as ‘intervals’ which record some part or aspect of a larger movement, both through temporal boundaries and selection of features. But if we think of these intervals as traces or echoes of lived movement, then we can think of their links too as partial, reductive traces of a larger multiplicity of potential or virtual linkage.
In this sense we can think of a concrete graph of recognised similarity relations over concretely defined intervals as a partial subset of a much, MUCH larger abstract graph that relates all possible intervals at all scales with all other possible intervals. So we might begin to rethink Bergson’s claims for the role of his ‘memory image’ in perception as describing a virtual space in which computable functions between time intervals form only a negligible subset.
Video credits: Tools that Propel by Sarah Levinsky (first dancer: Maria Evans; other dancers: see youtube description)
The unsolved hard problem of consciousness: the gulf between “internal consciousness” and physical observables, implies that no computational techniques based on physical sensing can yield certificate of intention.
Any choice of measure, or instrument of measurement, makes some aspect of experience legible, and others illegible.
Any sensor makes some feature of activity visible and all other features invisible.
Movement, as a temporal phenomenon, escapes synchronic representation.
Experience > perception > data. Examples abound from psychoacoustics and psychovision, and sensorimotor studies, gestalt, phenomenology.
All algorithms on Turing-equivalent machines (i.e. on any digital computer) suffer from absolute limits:
The space of computable functions is, in a strict sense, negligible as a subset of the space of all functions on the same domain and range. In particular, the space of functions mapping between finite discrete subsets of the reals is of measure zero in the space of all continuous (measurable) functions on the reals.
Undecidability puts a strong (fatal) bound on what can or cannot be amenable to any algorithm.
3 distinct interactive and participatory dance projects: Narcissus Reflected, RCO and BodyFail, in which various complex elements escape computation. On the level of the variability of audience involvement and participation, statistical evaluation of audience choices and decisions via their mobile phones, and their physical experiential actions and reactions. Finally, like in the case of BodyFail, it is precisely what escapes proper computation that constitutes the thematic itself of the piece: with non-habitual movements audience can make the computer system reach a crash.
What escapes computation in artistic performance is almost everything… From rich experiences, intentions, movement in continuous space and time, computation makes numerical representations, and these are incomplete snapshots of a larger story. Rather than the attempt to capture faithfully the phenomena in stake, computation should be a constrained frame that generates few possibilities and de-familiarizes the body.
What escapes computation in artistic performance is the whole conversation on politics of inclusion. Computation doesn’t like outliers nor special cases. It normalizes based on given data. Computation doesn’t engage in social conversation nor in what our future should look like. What escapes computation is the body, with skin, flesh and bones.
My body that needs to sleep, to eat and to have sex. My body that ages. My body that gets tired. My body that can only flex up to its limits. My body that is not isolated. My body that learns and grows. My body that exists in society for the years it has. My body that took time to become and that will take time to end. My body that beholds my experiences, feelings, fears, skills and weaknesses. My body is where artistic performance emerges. My body is so human with all its cells and their connections. That is what escapes computation.
As a provocation more than a personal statement, I would argue that if by interactive performance we mean technology-mediated performance involving interaction between human and digital technology of some kind, nothing escapes computation. In other words, interactive performance is, by definition, the act of performing computations.
Can we better understand the complex process of human music perception through a standardisation of the quantification of involuntary correspondences between motion and music?
Exploring the frequency-domain and time-domain links between sound and motion signals in a systematic manner might encourage international and interdisciplinary collaboration, boosting developments in the field and knowledge transferability.
As I began to write, Ananya Chatterjea’s  work immediately came to mind. Chatterjea questions the dominance of somatically informed, postmodern methods of ‘listening to the body’ as having ownership over embodied discourse within dance. In this she does not discredit or devalue such work, but instead articulates that to assume these practices and affiliated research universally speak to the whole of embodied experience both flattens such inquiry and omits others perspectives – particularly those that consider sociocultural, political, and economic ideologies to be inextricably entangled within our multilayered notions of embodiment. Considering Chatterjea’s arguments within the context of this discussion, I ask the following questions:
Does the somatic, postmodern practice of deconstructing embodied experience along spatial, temporal, kinesthetic, anatomical parameters neatly align with the limitations, goals, and aims of computational practices?
If so, are we aware that this neat coupling both flattens our work and excludes other embodied perspectives?
Is our flattening of embodiment akin to what Norman defines as the making of ideal lab spaces; a “clean-room type boxes with level floors, materially bounded to facilitate rigging and equipment maintenance, and the marking up of well delineated test areas?” [2, p.5] If so, what happens if we “diversify and extend the limits of our analytical apparatus?” [2, p.5]
My point here is not do devalue or dramatically change our practices, but to question how the repetition of a specific disciplinary coupling (namely somatically informed dance from postmodern lineages and movement sensing for interaction design) cultivates a flattened definition of ‘meaningful’ research and ‘pleasing’ aesthetics within the field. My aim is to better understand the gaps, limits, and assumptions we make about our practices – not necessarily to change them, but to better articulate their value and limitations beyond the scope of the field.
 Chatterjea, A., 2004. Butting out: Reading resistive choreographies through works by Jawole Willa Jo Zollar and Chandralekha. Wesleyan University Press.
 Sally-Jane Norman. 2015. Grappling With Movement Models: Performing Arts And Slippery Contexts. In Proceedings of the 2nd International Workshop on Movement and Computing. ACM, New York, NY, USA, 136-141.
Can machine bodies express the same ideas as human bodies? This question has been explored in my lab through embodied practice, study of theoretical underpinnings of computer science, and experimental quantification of robotic system capacity. The figure below, from an in progress work (https://arxiv.org/abs/1707.05365), shows the dearth in external complexity of extant robotic systems over the past 15 years (x-axis: internal complexity as measured by number of transistors in on-board CPU; y-axis: external complexity as measured by possible poses). As a result of this work, I do not think that machines can express the same breadth of behavior as natural systems.