More selected projects

Shaping a choreo-milieu to be in / with / within

This research is done in response to a personal urge to reconfigure the choreographic practise within a feminist, agential realistic framework. The process of investigating emerging choreographic approaches culminated in the invention of a new choreo-related term: the 'choreo-milieu'.  This practise will be free from choreography' s hauntological im/possibilities and will hopefully find a perceptive diffraction in a deep learning artefact: "Generative dance soloists improvising in the 2-dimensional screen space" constitute the first attempt of shaping and experiencing computationally the choreo-milieu. ​

produced by: Eirini Kalaitzidi

Choreography is a technology of the disciplined body, a mode of synthesis, a generative mode, a mode of power, a system of assigning perception and signification to dance. Choreography is an apparatus of capture, as André Lepecki mentions artfully (Lepecki, 2007), using a term that Deleuze and Guattari (Deleuze, 1987) introduced and which reflects a problematic relationship between choreography and dance, with the first one capturing the latter in order for it to exist.

The notion of choreography first appeared in 1589 in the version of an alternate but semiotically identical word: “orchesography”. “Orchesography” is the title of one of the most famous dance manuals of the late 16th century. Same as choreography, it is composed by two words: orchesis < [όρχηση], which means the art of the ancient greek chorus (dance), and graph < [γραφή] for writing (Lepecki, 2006). By the time orchesography was conceived, dance and writing became clung to one another within a notion that changed the relationalities between the subject that moves and the subject that writes. The divided existence of these two independent entities along with choreography as a mediating function seems like an ontological gap inside the deepest essence of choreography. A separation of entities hampers the dynamic relationship of choreography/dance, instead of interwinning the practises of knowing and becoming (Barad, 2007).

In response to that, this research is aiming to explore choreography’s insufficiency to fit in the feminist agential realistic scope, where human/non-human intra-actions (note 1) and mutually affective encounters emerge (Barad, 2007). It is also aiming to reconfigure choreography as a notion in the field of speculative resonances. In order to achieve that, this research attempts the invention of a new term which will be free from choreography’s hauntological im/possibilities (note 2) and which will hopefully find a perceptive diffraction (note 3) in a deep learning artefact.

Throughout the research, there is a persistent invocation of Karen Barad’s agential realism as an ideal feminist framework for new choreographic practises to emerge.

The tyrannies of choreography

“Dance, once it falls prey to a powerful apparatus of capture called ‘choreography’, loses many of its possibilities of becoming. Which is to say that dance loses its powers (puissances) as it is submitted to the power (pouvoir) of the choreographic” (note 4). (Lepecki, 2007)

In this case, the choreographic apparatus does not have Barad’s interpretation of an open-ended practise which would include intra-actions of humans and non-humans (Barad, 2007). Instead, this apparatus is a traditionally Western apparatus that defines dance by capturing it, creating a relationship which is seemingly equivalent to the antagonistic dualisms that Haraway describes: self/other, whole/part, agent/resource, maker/made, active/passive, God/man (Haraway 1991). Choreography’s origination in orchesography, reveals such masculinist dualisms (Lepecki, 2006), as described below.

In the manual “Orchesography”, a young lawyer, Capriol, arrives in Paris in order to visit his old master of computation, Thonoit Arbeau. Arbeau is a mathematician, a Jesuit priest and a dance master. Capriol asks his master to teach him the art of dance so that he may be able to perform well in society and instead of a physical training, Arbeau creates orchesography, as a technological binding of dancing and writing, as a pedagogical bond of the male couple, as a response to the request from the law. Arbeau writes the manual alone, in his empty chamber and destines it for Capriol to practise by himself, in an empty chamber as well (Lepecki, 2006).

Hence, choreography was created by a solitary male religious scientist for a solitary male lawyer who needed guidance with his bodily socialization, his heterosexual dancing and mating. And this guidance came in the form of writing and created the choreographic practise of accessing absent presences through a book, through solo reading and calling on a master.

Existing practises

Under the apparatus of capture, Western dance found its “liberation” through women who choreographed the “becoming minoritarian” (Lepecke, 2006). Isadora Duncan (1877-1927) choreographed based on her attachment to motherhood and to her children. Loie Fuller (1862-1928) choreographed based on her attachment to electricity and the technology of light. Martha Graham (1894-1991) choreographed based on woman’s strength and power that challenged the female image of the time (Franko, 2012). However, in these cases, the liberation from the masculinist apparatus is more of a driving force and a theme rather than a choreographic practise per se.

Currently though, the world of dance is raising more and more questions, compared to the mid and late 20th century, regarding the choreographic practises (Birringer, 2014). Artists are already experimenting with slightly differentiated approaches from what choreography was in its Western modernity version. In this direction, the incorporation of technological means in choreographic practises challenges the antagonistic dualisms: choreography/dance and choreographer/dancer, in highly intriguing ways, as Haraway would suggest (Haraway, 1991).

William Forsythe’s choreographic process creates conditions for events (Manning, 2008) and asks whether it is possible for choreography to generate autonomous expressions of its principals: choreographic objects, without the human bodies (Forsythe, 2009). He treats these objects as propositions (note 5) co-constituted by the environments they make possible (Manning, 2008). This approach is perfectly fitting into Barad’s agential realism where environments and bodies are intra-actively co-constituted (Barad, 2007).  In Forsythe’s multidisciplinary projects, “Synchronous Objects” and “Motion Bank”, computational means encourage the interpreter of the propositions to assemble and to manipulate the movement material or the memory of it. Specifically, “Synchronous Objects” follows a path from dance to data and from data to objects and finally transmits choreographic ideas without levying human bodies (Forsythe, 2009). Following “Synchronous Objects”, “Motion Bank” creates a bank of data on choreographic processes and by that Forsythe is focusing on the archiving potentials of choreography as computational composition (Forsythe, 2010-13). Scott deLahunta, who also worked in “Motion Bank”, promotes a vision of a new “dance literature”, exploring various non-linguistic forms of description and collateral knowledge relations drawn together by artists and researchers (deLahunta, 2013).

Similar to Forsythe’s approach, the collective of artists and designers “body > data > space” is placing the living body at the heart of the digital debate in ways that could be interpreted choreographically (body>data>space). The principle idea behind their practise is to forward connectivity between spaces and humans. In one of her interviews, the creative director Ghislaine Boddington describes the collective’s practise as a process of an internet of bodies, as a process of connectivity between the many things around us and ourselves. (Boddington, 2014)

Continuing with agential realistic versions of the choreographic, Monika Jaeckel made a performative lecture (“Turn around, make a round turn - Thinking around the transpositional options of movement” in transart-fest, 2013) with the intention to explore and define the space in orientation and relation to each other, to the audience and to the speaker. Although she does not practise choreography in her own work, she takes as a central reference for her discursive practise a choreographic piece from Laurent Chétouane (“Sacré sacre du Printemps”, 2012) and she treats the agent’s body of the lecture, the dancer, as a thing which is one among other essential contributors to the process of spacetimemattering of intra-action (Jaeckel, 2013).

“Cipher” is a video artwork made by the multimedia artist Katriona Beales. Beales is using a variation of a deep learning software to generate a virtual body with constantly changing surfaces (Beales, 2016). This dynamically and endlessly transforming body appertains to choreography as a spatiotemporal technique which offers spaces of appearance for bodies (note 6). In addition to that, the use of machine learning for the body to exist, unsettles the dualism of a discrete creator and a discrete creation. Who is generating the body? Beales, the computer or the body’s digital memories?

Bodies emerging in space can also be found in Deborah Hay’s “No Time To Fly” (Hay, 2011). Although she does not use high tech means, like deep learning, to generate bodies in space, she seeks for a continuous reinvention of the self as a body responding to the choreographic. “My vision of the dancer, through the intervention of performance as practise, is as a conscious flow of multiple perceptual occurrences unfolding continuously” she claims in her essay “Performance as Practise” (Goldman, 2007).

From left to right: “Synchronous Objects”, “Turn around, make a round turn - Thinking around the transpositional options of movement”, “Cipher”, “No Time To Fly”

While these practises are trying to disturb and unsettle the modernist logic within or close to the territory of choreography, the present research attempts the invention of a new term for a choreo-related practise. In this practise, the conventional positionality of the choreographer-master as author is no longer sustained as Birringer would encourage (Birringer, 2014) and all subjectivities are constantly created; not given or revealed. These subjectivities include all the bodies participating in the dance-emerging space: humans and non-humans, dance makers, dancers and spectators. In this practise, the significations of the moving bodies cannot be planned in advance. Rather, every encounter between bodies is open to intra-actions which inevitably screen out any particular disciplinary canonisation.


“Milieu” is the key word which will define the material semiotics of the new term. The milieu, as a scientific and philosophical notion, is the in-between two centres and brings with it a Newtonian fluidity, the fluidity of transitions and transmissions (Canguilhem, 2001). According to this research’s approach, milieu is the fluid space in-between bodies and body parts, constantly (re)generated from these bodies and body parts and unconditionally open to the environmental influential circumstances, as Lamarck designates the set of outside actions.

The new term invokes “milieu" to set the spatial stage for an ecology of bodily events and “choreo” to set the dancing materiality of bodies and movements. Choreo-milieu is a dance-emerging practise descending from choreography but founded on agential realism. It is calling dance artists to be in/with/within it rather than to apply it like a mediating method between them and a potential output.

Generative dance soloists improvising in the 2-dimensional screen space

Based on the idea of introducing this new dance-emerging practise, the artefact is an extension of me being-in-a-chore-milieu instead of the author choreographing. It is a first attempt of slipping into the space in-between the traditionally perceived as choreographer and the traditionally perceived as performer, while incorporating technology and therefore, slipping into the space in-between physical material and digital material.

“Generative dance soloists improvising in the 2-dimensional screen space” are computer generated sequences of movement, embodied by plain stick figures and collaged on top of photographical backgrounds where the soloists can be viewed within a transcendental and poetic context. The system of generating movement is trained after video recordings of one solo dancer  improvising in the humanly perceived 3-dimensional space for 30 minutes. At the core of this system is a deep recurrent neural network (RNN, specifically LSTM, note 7) trained on skeleton tracking data (x and y coordinates of 13 body points) and predicting new coordinates in sequence.

Tracking my body points coordinates throughout the 30 minutes of improvisation, using PoseNet inside electron and Max/Msp.

Compilation of the generated stick figure dancing on various backgrounds along with diagrams with the coordinates’ values through time.

Link to the choreo-milieu:

The primary dance material, the training dataset, is born while the solo human dancer is improvising. But since there always exists the bias of the same conscious human body/mind practising, computers are expected to perform better in accord to that. The term “improvisation” invokes associations with such related notions as extemporisation, invention and origination (Carter, 2000), notions that strongly relate to what generative models deliver in response to their training data. What we could not meet in a systematic preconceived choreographic process, we might discover in a deep learning generative process. We might find the stick figure flying, jumping in extreme heights, changing hip length, bending its waist by 270 degrees or equally, avoiding stagnation.

Writing on Baroque dance, Mark Franko notes how performance of choreography is first of all a performance centred on the display of a disciplined body performing the spectacle of its own capacity to be set into motion (Lepecki, 2006):

“Anyone who has studied baroque dance in the studio under the teacher’s watchful eye can testify that it allows little or no place for spontaneity. The royal body dancing was made to represent itself as if machined in the service of an exacting coordination between upper and lower limbs dictated by a strict musical frame. It was an early modern techno-body.”

In contrast to that, the emerging non-human techno-body of this research is free from disciplinary haunts. The seemingly author cannot apply any fatherly forces on the stick figure, apart from providing it with its training data. By the time human encounters the machine, they co-constitute the spaces in-between and we observe a choreo-milieu expanding. “It is not clear who makes and who is made in the relation between human and machine. It is not clear what is mind and what is body in machines that resolve into coding practises.” (Haraway, 1991). The hereon coding practises provide a creative predictive model (LSTM) with which the stick figure is predicting its next move based on its current one, for every single frame. This is how a sequence of movement is generated; this is how dance is generated in the fluidity of a choreo-milieu with no preconceived postures, temporality, body points’ encounters.

The inclusion of a static background behind the kinetic generative output is derived by a need to situate the non-human soloists in a human land (landscapes from Tinos and Kithira islands), where they have the exclusive non-human privilege of being able to dance. Through this collage, more spaces in-between are configured: the space in-between physical and digital, static and dynamic.

The choice of drawing a stick figure as the material body of the moving subjects aims at a simplicity and visual accessibility to them. Golan Levin used stick figures in his project “Ghost Pole Propagator II” as an “exceptionally expressive means of representing the human form. […] By eliminating unnecessary details and reducing visual elements to their absolute essence, the stick figure accomplishes a powerful, immediately apprehensible and possibly universal communication of human presence, attitude and gesture.” (Levin, 2016) In our project however, we face the stick figure more as a human-like cyborg rather than a representation of the human. The research did not focus on the materiality of the body, however, there lies an intention to maintain the skeleton tracking materiality (note 8) which reflects a human-computer interaction (Tuuri, 2017) and is consistent to a schema that has plasticity, the capacity to change by continuing to function as a total organism (Sparrow, 2014).


The notion of choreography and the choreographic carry with them a hauntological aura of solipsism and masculinity. The present research investigates how rethinking the relation between the author/choreographer and the reader/disciplined body would lead us to a feminist, agential realistic version of dance making. Varied approaches, computational and not, from notable dance artists constantly attempt to treat choreography as an environment of dynamically changing and responsive bodies. However, these practises inevitably fall back into the strata of the choreographic as long as there exists the onto historical material semiotics signification of the term. This research’s proposal is to invent a new term for the reconfigured dance emerging practise: Choreo-milieu is the spatial, temporal, material and meaningful “in-between” dancing subjectivities. Entering a choreo-milieu, being in/with/within it, is the diffractive alternative to the disciplinary practise of choreography.

“Generative dance soloists improvising in the 2-dimensional screen space” were created after this discursive route as a first material application of the developed theory. They are generated dance sequences originating from both a bodily improvising human and an algorithmically improvising computer. The intention is for them to emerge in the encounters of human and non-human, physical and digital, creator and creation. An additional intention is for the soloists to co-exist with one another and the occasional spectator, affectively, in the 2-dimensional screen space.

So now is a good time to wonder in what other techno-scientific forms and shapes could choreo-milieu occur. If it finds its essence in the (re)generation of encounters and becomings, how else could it emerge, except through generative deep learning practises? Since it is occurring now and continuously, is it considered to be an eternally live practise? What is its affinity to performance? Does it always need the initiative of a human to emerge? Does it ever have an ending?

While observing the sequences of movements, they seem quite vague and unearthly, most of the time resembling the state of floating, in such a poetic way that I am convinced that I witness and engage with an affective choreo-milieu.




1. That is to say that objects and agencies are always emerging while being with each other and mutually responding to each other, suggesting a new understanding of causality (Barad, 2007).

2. According to Karen Barad, hauntological im/possibilities are co-existing multiplicities of entangled relations of past-present-future-here-there which constitute the world phenomena. “What if the ghosts were encountered in the flesh, as iterative materialisations, continent and specific (agential) reconfiguring of spacetimematterings, spectral (re)workings without the presumption of erasure, the “past” repeatedly reconfigured not in the name of setting things right once and for all but in the continual reopening and unsettling of what might yet be, of what was, and what comes to be?” (Barad, 2010, p.261)

3. “As Haraway suggests, diffraction can serve as a useful counterpoint to reflection: both are optical phenomena, but whereas reflection is about mirroring and sameness, diffraction attends to patterns of difference.” (Barad, 2007, p.29)

4. Where puissances are a matter of becoming and pouvoir is a matter of State, according to Deleuze. (Boutang, 2004)

5. The way Forsythe is interpreting propositions is well described in Manning’s writings: “You don’t need a choreographer to move. You just need a proposition. Propositions are ontogenetic: they emerge as the gem of the occasion and persist on the nexus experience to take hold once more through new occasions of experience.” (Manning, 2008)

6. Space of appearance is Hannah Arendt’s notion.

7. LSTMs (long short-term memory models) are a special kind of RNN (recurrent neural networks), whereas RNN are models which allow operations over sequences of vectors: sequences in the input, in the output or in both. LSTMs are capable of learning long-term dependencies. They were introduced by Hochreiter & Schmidhuber (1997) and were refined and popularised by many people in following work. (Colah, 2015)

8. The skeleton tracking visualisation was already used by default in the first stage of the creative process, when reading the improvising human body’s points in space through another machine learning model, PoseNet.


Annotated Bibliography:

1.     Canguilhem, G. (2001). The Living and Its Milieu. Grey Room, translated by John Savage. Vol 3, pp. 7-31.
In his lecture and later publication “The living and its milieu”, Georges Canguilhem analyses the notion of milieu in a scientific and philosophical way. He firstly introduces the term as a universal means of experiencing life and thinking of life.

Following Newton’s problematics on this matter, Canguilhem elicits an absolute meaning of the term: the “between” two centres. “The fluid (relative to ether) is the intermediary between two bodies; it is their milieu; and to the extent that it penetrates these bodies, they are situated within it.” Passing through perceptions of Lamarck, Buffon, Comte, Darwin, Ritter, Humboldt, Watson, Weiss and correspondingly through the disciplines of biology, geography, mathematics, psychology, physics, all adding to the meaning of milieu, Canguilhem mentions that the behavioural, the geographical and the physical milieu coincide with each other.

Canguilhem’s living thing (man) is becoming a “creator of the geographical configuration” which allows us to think of it (him) as a geographical and spatial factor; a relation that can be directly translated into dance terms. In every choreographic approach there is a need to define the space along with the form and the temporality of the bodily behaviours. The milieu as the world of a man’s mental and physical perception, is the space between bodies and body parts; it is the space “between” that exists even within the living body and is also (re)generated from this body.

2.    Sparrow, T. (2014). Plastic bodies: Rebuilding sensation after phenomenology. Open Humanity Press.

In his book “Plastic bodies: Rebuilding sensation after phenomenology”, Tom Sparrows responds to the transcendental phenomenological approach to the body with a post-phenomenological approach that views the body as plastic. The writer indicates a need to rethink the essence of the materiality of the body. In this process, he analysis Merleau-Ponty’s and Levinas’ thesis on embodiment and goes a step further. There is no given or a priori spatial body. There is the body that acquires its temporal space and reforms this space “within the general realm of extension”. But in addition to its spatial dimension, the body has material properties that need to be distinguished from their functionality: the future of the flesh. “How can we conceive of matter without reverting to mechanism? In order to properly distinguish matter from mechanism, we will call this post-phenomenological materiality plasticity. Plasticity is thus defined as that which comes after the flesh.“ It is a generalized disposition of material bodies.

Sensation is the key to this process of re-materialization of the body after the preceded de-materialization which accompanied phenomenology’s attempt to de-objectify the body. The body, as the subject, is produced by a set of perspectives provided from the environment of aesthetic stimuli and is sent back to this environment of stimuli. “The body’s senses hold a power of their own” that determine its schema and its plasticity.

There is a clear distinction between plasticity and elasticity. Plasticity inhibits the reversibility of the body, therefore the absolute healing of the injured body, therefore the return to the integrity of an original form. But, there is a vitality and a generative aspect to this plasticity. The plastic material “retains the trail of its deformation” in a way that gives birth to new forms, new schemata. Plasticity is directly linked to the schema of the body. Catherine Malabou, who writes the foreword of the book, defines plasticity as the capacity to change by continuing to function as a total organism. She also claims that plasticity entails a detonation of form, an explosiveness that “transforms nature into freedom”.

3.    Tuuri, K., & Parviainen, J., & Pirhonen, A. (2017). Who controls who? Embodied control within human-technology choreographies. Oxford University Press.

In this paper, the researchers explore the embodiment relations that emerge between current technologies and the moving body. They treat movements as “lived experiences that constitute humans as embodied beings” and they focus on the intentional (or not) choreographies humans engage with in everyday life. Unlike the dance field, the choreographies of this sort do not follow any choreographer’s instructions - rather, they are led dynamically by human and non-human agents.

Starting by describing the prevailing concept of using the body and its physical movement with an instrumental intention (user to device control-scheme), the paper turns its attention to the phenomenological understanding of human - computer interaction (HCI)  in a choreographic context:

“How discrete movements belong to a bigger whole of both spontaneous and orchestrated/arranged movements […] how these choreographies differ across time and places […] how movement and space are lived through by subjects (individual choreographies), and among subjects belonging a cultural community (social choreographies).”

Analysing these ways of “choreographing” as an assistive method of approaching HCI, the paper distinguishes three types of embodied control in relation to interaction design: a) The instrumental control, where the body (its gestures and movements) is the instrument in control. b) The experiential control, where the body is experiencing both push and pull effects, both being controlled and being in control. c) The infrastructural control, where the body is controlled by the material and immaterial infrastructures of its lived space.


Additional References:

Arbeau, T. (1967) [1589]. Orchesography. Translated by Mary Steward Evans, edited by         Julia Sutton. Mineola: Dover Publications.

Barad, K. (2007). Meeting the Universe Halfway: Quantum Physics and the Entanglement of Matter and Meaning. Duke University Press Books.

Barad, K. (2010). Quantum Entanglements and Hauntological Relations of Inheritance Dis/continuities, SpaceTime Enfoldings and Justice to Come. Edinburgh University Press.

Beales, K. (2016). Cipher. Retrieved from:

Birringer, J. (2014). What Score? Pre-choreography and Post-choreography. International Journal of Performance Arts and Digital Media, 9:1, 7-13

Boddington, G. (2014). body/data/space. Retrieved from: body>data>space, since 1990s. Retrieved from:

Boutang, P. A. (2004). L’Abécédaire de Gilles Deleuze. France: Editions Montparnasse.

Butler, J. (1990). Gender Trouble: Feminism and the Subversion of Identity. New York: Routledge.

Carter, L. C. (2000). Improvisation in Dance. The Journal of Aesthetics and Art Criticism. Vol. 58, No. 2, Improvisation in the Arts.

Colah. (2015). Understanding LSTM networks. Retrieved from:

deLahunta, S. (2013). Publishing choreographic ideas: Discourse from practice, in E. Boxberger, G. Wittmann (eds), German Dance Education Biennale Proceedings, München: E-podium.

Deleuze Gilles, and Félix Guattari. (1987). A Thousand Plateaus: Capitalism and Schizophrenia. Translated by Brian Massumi. London: Athlone Press.

Forsythe, W. et al. (2009). Synchronous Objects. Ohio State University - Advanced Computing Center for the Arts and Design (ACCAD). Retrieved from:

Forsythe, W. (2010-13). Motion Bank. Retrieved from:

Franko, M. (2012). Martha Graham in Love and War: The life in the work. New York: Oxford University Press

Goldman, D. (2007). Dance Composes Philosophy Composes Dance, Series on New Choreography, Part II, TDR - The Drama Review, Summer 2007, Vol.51(2), pp.157-170.

Haraway, D. (1991). Simians, Cyborgs and Women: The Reinvention of Nature. New York: Routledge: 149-181.

Hay, D. (2011). No Time To Fly. Retrieved from:

Jaeckel, M., & Matsuyama Y.  (2013). Turn around, make a round turn - Thinking around the transpositional options of movement. transartfest – trans-what Symposium, Berlin

Lepecki, A. (2006). Exhausting Dance: Performance and the Politics of Movement. New York: Routledge.

Lepecki, A. (2007). Dance Composes Philosophy Composes Dance, Series on New Choreography, Part II, TDR - The Drama Review, Summer 2007, Vol.51(2), pp.119-123.

Levin, G. (2016). Ghost Pole Propagator II. Retrieved from:

Manning, E. (2008). Propositions for the verge: William Forsythe’s Choreographic Objects, Inflexions: A Journal for Research Creation, no. 2, Rhythmic Nexus, 1–32.

Rosenberger, R., & Verbeek, P. P. C. C. (2015). A field guide to postphenomenology. In R. Rosenberger, & P-P. Verbeek (Eds.), Postphenomenological Investigations: Essays on Human-Technology Relations (pp. 9-41). (Postphenomenology and the Philosophy of Technology). Lexington Books.