This question could be interpreted as computation failing to capture, categorise, and process the ‘whole’ of human movement, in all its multi-modal expressive, dynamic, and ecologically-entangled qualities. By failing thus, it cannot perhaps properly interact, dialogue or improvise with the performer in ‘a two-way, co-extensive process’ (Sutil 2015). But what if the ‘escape’ is not a ‘failure’ so much as a ‘difference’ of perception? What if we use the differences to create affordances for interaction? Calvert et al. discuss the lack of a ‘consistent, unambiguous way […] of representing human movement with a machine-readable ontology or grammar’ (Calvert et al. 2005: 9). But with Tools that Propel, a digital choreographic and improvisational system co-developed with Adam Russell, new creative potential was found not in solving the terms on which we create a shared ontological understanding of ephemeral, embodied concepts and relationships – movement, for example – but on using our inability (and refusal thereby) to reduce the ‘voluminousness of the experience’ to a grammar that would produce ‘instrumentalized action-reaction circuits’. (Massumi 2011: 46). Not striving for the greatest degree of accuracy in the categorisation of choreographic information digitally can be creatively productive – with Tools that Propel the discontinuous, the fragmented, the gaps, actually provide a source of defragmentation and continuity with the past: that is, the possibility for the past to recur in ways that surprise us in the interaction, make us look at ourselves and our movement anew, and offer possibilities of new ideas within them. So, what if instead we ask: what can we discover in our embodied understanding of ourselves when we interact with computation with an acceptance and embrace of the fact that it ‘perceives’ and ‘understands’ our movement differently from us?