1. Martina Leeker:
    Entfesselte technische Objekte
  2. Louis-Philippe Demers:
    Machine Performers

Louis-Philippe Demers:
Machine Performers: Neither Agentic nor Automatic


Machine performers could constitute an investigation of cultural codes of the theatre audience in order to look at a broader understanding of the machine itself. This paper investigates the perception of the robotic agent in the following four areas: the historical lineage of the uncanny valley, artificial intelligence (AI), anthropomorphism, causality and animacy. These four areas will be looked upon from the perspective of “body versus action” in an alternate way than science.


Robotic Art, Theatre, Embodiment, Artfificial Intelligence, Entertainment robotics.


Artistic languages and scientific concerns might revolve on very different orbits. While the HRI community strives to define functional models and theories grounded in the physical reality of the robotic agent, art and in particular Theatre, are more concerned about staging the unreal.

Theatre theorist Horakova entitled a paper: “Robots between Fictions and Facts”[1] and the seminal book from Reichardt is entitled: “Robots: Fact, Fiction and Prediction”[2]. Peformance Theory and Art Theory have embraced, in their own terms, oblique analysis of the Uncanny Valley, of the roles and representations of the Robot and of the living/non-living dichotomy manifested by the Robotic Media. The very paradox of the embodied object[1], the art robot, might be real in its incarnation of mechanics and control, but its perception departs from its sole capabilities of acting in the real world[3].

Machine or mechanical performers have been around for centuries. This development has been furthered by the current advents in the digital age conveyed by virtual characters on the silver screen, and the recent presence of unsurpassed humanoids [4] and geminoids [5]. This paper will put forward some communalities and differences between HRI and Theatre via the analysis of Machine Performers supported with case studies. Both fields have to deal with liveliness, presence (physical), embodiment, perception and identification, empathy and finally, bringing realism to the unreal.

In order to unfold the investigation of machine performers, the analysis is broken down into the inter-related constituents of a robot: its body (representation), its movements or behaviors (body in action) and its context (environment at large, from the stage to culture). Over the last twenty years, the author has been focusing this research on experimental robotics, which can relate these robots constituents to human perception in the following four areas: the historical lineage of the uncanny valley, artificial intelligence (AI), anthropomorphism, causality and animacy. These four areas will be looked upon from the perspective of “body versus action” in a different way than science. Mostly in the fields of Human- Machine Interfaces (HCI), Psychology, Artificial Intelligence and Engineering, the scientists do not consider the potentials from the context of Art and Theatre. On a practical as well as a theoretical level, this paper will present an overview of these topics and then pose a number of questions about of how these could be reformulated for a trans-disciplinary investigation between HRI and the Performing Arts. The cumulative analysis is synthesized in Table 1.

From the Gods to the Uncanny Valley

The Social Robot: a history of human representations.

Through the history of machines, the role of theatricality can be traced in order to augment the viewer’s reception of the machine performer. The aim is to also establish that the artistic renderings of the machines during certain periods can reflect the current social concerns and the current understanding of the human body.

Perhaps some answers can be found in the history of representations, models and simulations of the living body, by means of mechanical objects. This history is more than two millennia old. The media theorist David Tomas writes about modifications of the human-machine relationship as a “machine- based history of western body” [6]. Tomas often refers to the Cybernetics discourse, particularly to Norbert Wiener’s writing on a history of mirroring the human body with machines. Wiener traced the parallel histories of machines and human bodies, when he presented a history of automata that was divided into four stages [7]. These histories generated four models of the human body. Firstly, a mythic Golemic age that referred to the body as a malleable, magical, clay figure. Secondly, the age of clocks (17th and 18th centuries) where the body was seen as a clockwork mechanism. Thirdly, he considered that the age of steam (19th century) transferred the body into descartain scenario; a “glorified heat engine” which burnt combustible fuel instead of the glycerin from human muscles. Finally the last stage, Weiner identified the age of communication and control (the age of cybernetics); an age marked by a shift from power engineering into information and communication engineering; from “economy of energy” into the economy based on “the accurate reproduction of signal” that understands the body as an electronic system.

After Alan Turing’s pioneering works, robotic art and artificial intelligence emerged from the assumptions established by Cybernetics. The appearance of robotic art in the midst of the 1960’s cybernetic discourse is connected with an anti-mimetic shift in the history of humanlike- machines. As Tomas argues: “The cybernetic automaton’s mirroring of the human body was not established on the basis of conventional mimicry, as in the case of androids and their internal parts, so much as on a common understanding of the similarities that existed between the control mechanisms and communicational organizations of machine systems and living organisms”[6]. However, robotic art has very broad roots and a rich cultural history. This history references modern science-fiction as much as artificial creatures (either real or imaginary); from ancient artificial maidservant to mediaeval Golems to Homunculus of Renaissance to androids of the Enlightenment. Contemporary robotic art brings a new aesthetic dimension that prefers modeling of behavior over a representative form or a mimetic static object.

The Uncanny Valley conjecture then follows this long lineage of mechanical relationships, and it symbolizes the current state-of-the-art in technology alongside the cultural anxiety of transferred agencies [8, 9].


Machines are regarded as distinct entities from us. As much as we consider ourselves distinct from nature, machines are a physical rendering of abstractions and can also act as a tool for the comprehension of ourselves within the structure of the world [10]. It is significant that outcomes of this effort, embodied in different robots/machines, are typically exploited by theatrical means [11, 12] . This history is driven by the ongoing quest for a true genesis and the deeper understanding of the inner self in the environment.

The paradox of the robot can be found in the ambiguous status of artificial human-like (androids) creatures and their existence. This paradox is not only present in the case of fictitious artificial creatures but also in the case of the ‘real’ mechanical puppet or android. As performance and puppet theorist Mark Sussman argues, thaumaturgical strategies often intensify this trick during robots public performances. Sussman began from the assumption that:

“Certain pre-technological performances (…) can give us some insight into the tense metaphoric operations and interconnections of faith and scepticism, or belief and disbelief, in the staging of new technologies (…)”.

In his analysis of the staging of the Chess Player automaton by Wolfgang von Kempelen (1762), Sussman came to the following conclusion:

“The automatic thinking machine that concealed, in reality, a human person, can be seen as a model for how a spectator might reify, and deify, the hidden power at work in a new form of intelligent machinery”.

Sussman suggests that this visual proof is a demonstration of the level of control at a distance and that the transmission of human intelligence into inanimate body of the object extends the context of androids/automatons staging in general. But also, it demystifies and reenchants the performing object itself.

Figure 1. Le Proces (Kafka)

Figure 1: Le Proces(Kafka).

The machine performers depicted in Figure 1 are the main protagonists of the robotic performance adaptation of Le Procès, a novel by Franz Kafka (Kafka 1925,[13]). These robots are deliberately part zoomorphic (an arm, a hand) and part mechanomorphic (the lower body is a simulation platform structure). Utilizing existing mechanisms to construct life-like objects brings us back to the paradox of the quasi-living objects of the robot history. The signs of the machine design comprise both inert and living connotations about the performing objects. In parallel to human performers, we can ask, whether and how are these robot performers able to carry an alternate set of sign-systems of their bodies (shape, material) and their behaviors (actions/acting).

The Uncanny Valley: at the crossroad of Art and Science

In recent artificial intelligence discourses, social robots have attempted to embrace the human form endowing it with friendly appearances and behaviors as a privileged mode of intercommunication [14]. This addiction to the mimicry of humanoid form in the appearance of robots connects them with a long history that has been written in myths, legends and even in real experiments.

Figure2: AreaV5. 60 pairs of eyes gazing at the audience.

Figure 2: AreaV5. 60 pairs of eyes gazing at the audience.

This addiction includes in itself two challenging motives: the dream to create an artificial human being and the need to create helpers for ourselves. On the one hand, this is seen as an attempt to imitate a ‘Creator’, to make a creature in our own image or even to discover the secret of life. On the other hand, it may be an entirely practical ambition to make optimal or perfect servants of man. This second motif is often connected with utopian projections of an ideally ordered social system.

A specific issue has arisen about the increase in realism, both in virtual and physical agents. This research has led to a controversy about the acceptance of those agents by humans. Within this conjuncture, the Uncanny Valley was already described 30 years ago by roboticist Mori [15] and it has recently resurfaced and it lies at the centre of discussions, namely in the field of social robotics. Mori proposed a thesis that would create an asymptotical burden in the development of human-like robots. Mori posited that, as realism of a robot increased, dips or discontinuities in the relationship between affinity and realism would occur. First there would an increase in affinity but as the robot would approach a nearly human state, there would be a dramatic decrease in acceptance. Mori plotted his assertions as an apparent mathematical two-dimensional graph (see Figure 3). This displayed a strong dip - “the Uncanny Valley” - when the robot is a “near human” entity, and then slowly ascended towards affinity when the robot is a perfect replica of the healthy human body and its behavior.

Figure3: The Uncanny Valley graph

Figure 3: The Uncanny Valley graph [15].

Research into the Uncanny Valley shifts our perception of the abilities of the mechanical performers both in artificial intelligence and in other cultural environments. In both settings, it forces the beholder of the robot to continuously draw lines between human and non-human traits. However, experimental theatre directors are also exploring this line, one that extends the notion of the fourth wall[2]. This is because one of the main aims of staging is to make the real out of the unreal. The uncanny Valley does not affect strongly the abilities of mechanical performers on the stage and in other cultural environments but it does shift our perception and empathy. The value lies in a breach of suspension of disbelief on the part of the audience; the moment where the agency of the machine performer is replaced by its sole automation.

What is the next step in the lineage? Would the supposed Uncanny Valley, the asymptotical burden in acceptation of the human-like robot, carry on to the machine performers on the stage? By bringing the robot onto the stage and away from the laboratory, how do shifts in perception affect the Uncanny Valley of the future?

By linking the Uncanny Valley to the experimental theatre stage, will anthropomorphism and anthropopathy play a major role in the perception of the non-human performer? Certainly, they both refer to the attribution of a human form, human characteristics, or human behavior to non-human things such as robots, computers and animals. However, anthropomorphism goes beyond the simple morphology of the perceived agent. For instance, the interaction designer Carl DiSalvo states that four categories of anthropomorphism can deal with what aspect of the human form is actually being imitated: the structural anthropomorphic, the gestural anthropomorphic, the anthropomorphic from the character and form of the aware anthropomorphic state [16].

By playing role between the agentic and the automatic, the robotic performer can explore some of the disturbing ambiguities associated with the machine’s uncanny lack of agency. The Uncanny Valley might lie where the perception start oscillating between the function (automatic) and the intention (agentic). In other words, the valley is where the perception oscillates between the inert/mechanical and the quasi-life qualities of the robot. In other words, shifting the anthropomorphism to level beyond morphology: what is seen as a pure functional entity and what has strong apparent autonomy and intentions?

The term uncanny, functions in both scientific and aesthetic significance. In the scientific view (AI discourse) the uncanny is seen as pseudo-natural perception without cultural codification. The task of the artistic and aesthetic level should be, to clarify cultural backgrounds of the robot- perception. The Uncanny Valley idea also chimes with Freud's writing, the narcissism of minor differences, where feuds between communities of adjoining territories ridicule each other’s. This territory is also depicted by Steve Dixon notion of ‘metallic camp’ [17]. Dixon argues that ‘robotic movement mimics and exaggerates but never achieves the human, just as camp movement mimics and exaggerates, but never achieves womanhood’, and that camp is an essential factor in anthropomorphic and zoomorphic robot performance.

Following Derrida's strategic valorization of the suppressed supplementary term of any binary, the dichotomy of the human versus the non-human lies in the grey area between these two realities. The place where the 'credibility gap' or the 'Uncanny Valley' occurs is not at the point furthest from the truth, but at the point closest to it. As observed by kinetic artist Kirbey, when a kinetic artwork is almost credible, it lacks credibility.

The interactive environment Area V5 (figure 2) deals with the Uncanny experience of the gaze of 60 pairs of disembodied eyes. It is a direct artistic comment on the role of the gaze in Social Robotics[18]. The design of the space and the unsettling embodiment from the sliced skulls was deliberately chosen to breach affinity. Audiences perceive the robots as both representative of the living and the non-living. The observed viewer experiences[3] lead to conclude that they voluntary engage in and out the suspension of disbelief (or the valley) without disengaging from the environment.

Artificial Intelligence and embodiment.

The role of the body in relation to the comprehension of human intelligence is now at the forefront of scientific research in many of the fields like psychology and neurobiology. This section brings along top-down and bottom- up approaches in Artificial Intelligence with the focus on the implication towards machine performers. The Cartesian and orderly juxtaposition of the brain, the body and its interaction with the environment will be challenged in the hopes to construct a new perspective of implementing behaviors into machine performers (embodied agent).

The turn of the embodiment.

In the Dartmouth Artificial Intelligence Project Proposal of 1956, the community convened under the assumption that "...every aspect of learning or any other feature of intelligence can in principle be so precisely described that a machine can be made to simulate it”. This idea dominated until the mid-1980s as a challenging approach. Recently the interest has increased in the notion of "embodiment" claiming that intelligent behavior is not only a matter of computation, but it requires a body, a complete organism that interacts with the real world. As a consequence, many researchers have shifted their attention away from the central brain (the computer) towards embodiment (the robots). This is seen by many AI theorists as a transfer from a top-down towards bottom-up approach [19, 20]. The far-reaching and often surprising implications of embodiment have often been used in relation to a literal meaning that intelligence actually requires a body. However there are deeper and more important consequences, concerned with the interaction of the brain, the body and the environment. Intelligence is not an “all-or-nothing” phenomena, but an incremental and interdependent gradual set of attributes that have emerged through the process of evolution [20]. Often the results of embodied AI start from models of physical agent behaviors without a complete base. For example, in “Stumpy” (figure 4), a robot that explores natural gaits and locomotion, the behavioral model was “outsourced” into a physical construction, where the apparent jumping actions emerge from the interaction of the agent within the physical world [21]. This was made with minimal computational efforts and representational models.

Figure4: Photograph and Schematic of Stumpy (courtesy AI Lab U. of Zurich)

Figure 4: Photograph and Schematic of Stumpy (courtesy AI Lab U. of Zurich)

Acting without thinking/modelling.

As Rolf Pfeifer in his book “ How the Body Shapes the Way We Think” suggests, AI should continue to focus on non hierarchical links between the brain (the computer) and the body as an eloquent distribution of muscular control mechanisms and cognition. He suggests that walking and the manipulation of objects reveal themselves to be a combination of the materials (tissues, bones, flexibility, sensors) and the system of distributed processing between system between the body and the brain.

On a deeper level, how does a hand grab a glass or an object? In this case, the sole capacity of the anatomy and morphology of the forearm and the hand enables them to adapt to all different shapes. Currently, the perpetual paradox of AI is very much at stake: being highly contested between simulation and modelisation, the grasping hand demonstrates that this kind of intelligence does not reside in the intelligence of the brain, neither of all the memorized forms of glasses, nor even from the cognitive level of the object “glass”.

“Outsourcing” behavioral and emotional models into physical constructions is along the line of the creative process of Kinetic Art. Apparent actions emerge from the interaction of the agent within the physical world with minimal computational efforts and representational models. Can machine performers behaviors and emotions be implemented without a complete computational model?

Figure5: The Tiller Girls as 12 Autonoumous Stumpies.

Figure 5: The Tiller Girls as 12 Autonoumous Stumpies.

Machine performers can express emotions due to their intrinsic materials and the very complex dynamics of their structure in motion. Such paradigm is similar to the psychophysical relation found in theatre acting methods where behavior and emotions are inherently physically grounded. For instance, the walking table of figure 6 manages to navigate even under a deliberate poor gait. The behaviour is a collaboration of the unstable equilibrium of the construction and the staging. The introduction of a latent failure in the gait not only creates a poetic moment but also gives a supplementary spark of life to the object, as it is similarly proposed for social robots. Acting methods also propose opposite stances to be taken by actors: presence or absence. The presence calls upon the performer’s experience to dwell into his/her experience to deliver the character, absence requires an abnegation of the self to produce a pure rendering of the directors’ directives and scripts. The beggar of figure 6 had no experience of misery neither of being poor. Its shape was a square box (symbol of a chest) that could rock over a hinge (body language of imploring). The beggar performer lean towards absence while the table is rooted more in presence via the physicality of its shape. This situation gravitates around the cheap design paradigm[20] where an ecological niche is being exploited thru an ecological balance among morphology, sensing, control and finally dramaturgy.

Figure6: A begging machine (left) and a limping table (right)

Figure 6: A begging machine (left) and a limping table (right)

The above situation is far from the mechanomorphic attempts of traditional robots from the top-down cybernetic wave. Perhaps the concept of embodiment inside of the theatre environment is underestimated by AI researchers where the relation between basic human movement and robotic movement can be compared simultaneously not only on a literal level but also on a metaphorical level. Being situated, theses agents can also empower intangible contributions from the cultural context, the suspension of disbelief and the attribution of intention towards any outside physical objects acting upon the world. From the audience standpoint, where our perception departs from the simple function (mechanical or programmed) towards the intention or emotions (self motivated complex agent)? Perhaps this collaboration will help to derive a form of the bottom-up “synthetic methodologies” [20] for the machine performers on the stage.

Performative vs Interpretive.

Performers in the traditional performing arts such as music, dance and theatre are generally thought to have both technical skills and interpretive skills, where the latter skills are regarded as specific human skills. Auslander highlights the ‘grey’ area between these with examples from the performing arts such as the practiced routines of orchestral musicians, and the famous early 20th-century Tiller Girls’ (figure 7) synchronized chorus-line dance, in which human performers are ‘called upon to exercise their technical skills but not their interpretive skills[9].

“Auslander exposes indeterminacies in this binary thinking in the traditional performing arts. In contrast, Auslander draws upon performance theorist Michael Kirby’s notion of ‘nonmatrixed performing’, in which a performer does not feign or present any role and is simply being himself or herself, carrying out tasks, to assert that robot performances can indeed be placed within the continuum of performance art. Auslander discusses examples of performance art in which there is no difference in overall artistic intention whether tasks are carried out by human performer or robot performer, and where the actions of a human or a robot can be regarded equally as art performances (p. 98).” [8]

The robotic Tiller Girls (figure 5) are empowering the AI embodiment while exploring what would happen if the physical world were shifted from the lab onto of the stage.

Figure7: The Tiller Girls.

Figure 7: The Tiller Girls.

Between the Agentic and the Automatic:

As the movement of the machine is one of the most prominent factors for the perception of its agency, this section investigates the human intrinsic mechanisms of perception of motion and the attribution of causality (from the audience experience standpoint).

Perception of Causality and Animacy

The perception of Animacy, Causality and Motion was an important field of research uncovered by Psychology and Neuro-Biology, which began in the 1900s by Albert Michotte [22]. At the time, scientific evidences were being accumulated about very simple displays (visual cues) and how they give rise to surprisingly high-level percepts.

The awareness of those fields were expanded by the Heider and Simmel [23] through the method of testing animated perceptual experiments with different audiences. In Michotte’s concept of “functional relations” wherein one perceives properties in visual cues that are found in an objective environment, he posits that one can not locate judgment in neither the actual events nor in their retinal reception. Heider and Simmel proved that the functional relations are primarily perceptual but that the interpretations are highly personalized and individual (see Figure 3).

Clearly animacy cannot be separated from the concepts of embodiment nor from the bodies of primal mechanism. The results tend to show that the perception of animacy and causality are innately connected in the human. Perhaps as Scholl and Tremoulet both posit, the pathways of animacy and causal modular processing could be dissociated from high-level cognitive judgments (2000).

Figure8: Various frames from a sequence of visual display (Heider and Simmel 1944)
	where participants anthropomorphizes through individual narratives.

Figure 8: Various frames from a sequence of visual display (Heider and Simmel 1944) where participants anthropomorphizes through individual narratives.

Animacy for the stage.

This research into causality and animacy could prove to be very valuable for designer of machine performers in the future. Only by gaining a deeper understanding into how primal and visceral human perception functions, can artists invent more fictional and more factual arrays of movement for the machine performers on the stage. Therefore, studies into causality and animacy will also extend the understanding of the agentic (intentional) and the automatic (functional) forms of robots on the stage. The main question is: at what point does a machine performer graduate from an automata into an agent?

The movement (or perceptible change of state) of an object can be seen in part as its objective nature, while its perception can be its subjective counterpart. Consequently, a rather abstract inert shape can become fluid, organic and eventually anthropomorphic, by the sole means of contextualization and movement. In figure 9, a simple motor mounted on springs creates a rich range of chaotic movement, staging this object in a cage anthropomorphises its essence resulting with the viewers perceiving it as an untamed miserable entity in La Cour des Miracles[13, 24]. Without an immense degree of computation, the behaviour is carried out by a juxtaposition of this social mise-en-scène and the inherent complex dynamic characteristics of the structure.

Figure9: Untamed Machine in La Cour des Miracles

Figure 9: Untamed Machine in La Cour des Miracles

Movement is seen as a sign of life and in order to fully understand causality of motion and perception, further research has to be considered when designing the robotic agent. How does our understanding of animacy affect our prediction and our assignment of “judgments” to the movements? When a human acts and when a human observes the same action performed by machines, how do they relate and respond? Like the mirror neurons[4], machines are often built to mimic our behavior as though the viewers were carrying the actions themselves. The author suggests that performing machines feel conspecific for the viewers and this aspect is part of their major appeal for humans. Can a social relation between machines be based on our understanding of animacy and how can adaptation to the theatre stage change their relations to each other? How does social interactions between machines on stage translate into social interactions of the audience? For example, will the audience feel excluded, indifferent or fascinated? One of the contributions from the Theater would be to formulate a new way of gathering information from the above questions.

Point Light Animation.

Animacy and causality are important aspects to explore in relation to the emotional reaction of the audience. Point Light Animation technique (PLA)[5] investigates the isolation of movement from other visual cues such as the morphology. Surprising results from these experiences demonstrate that we are able to identify action with a very reduced set of information as well as basic emotional reactions. Does the Uncanny Valley dip resonate within the agentic vs. automatic qualities of machine performers under the PLA? How can machine performers be built to explore these questions?

Anthropomorphism and Theatre: the realism of the unreal

Anthropomorphism within the history of machines, design of objects and theatre semiotics includes a broad array of abstract and representative mimicries of human behavior. From the beginnings of Greek automata to biomechanics, the enhancement via anthropomorphism has played a major perceptual role. The anthropomorphic character inside its context of arts is an important factor to consider.

Anthropomorphism entails attributing humanlike emotional states, behavioral characteristics, or humanlike forms to nonhuman agents. In 6th century B.C., Xenophanes was the first to use “anthropomorphism” when describing the similarities between religious agents and their believers (Lesher 1992). As Dennett (1987) also confirms, the audience often attributed intentions to mechanical characters or agents in relation to their predominant belief systems.

The close association between the performing machine’s visual image and the human image raises difficult questions not only for artistic practices but also for scientific disciplines. According to Cary Wolfe (Wolfe 2003) anthropocentrism or specieism both reflect the priority of visual reception into the human sensorium and only performances that engage other senses can be considered truly posthumanist. Wolfe suggests that our senses should be extended into a broader range not necessarily confined to the human bodily sensorium.

Today, this attribution raises questions about the level of anthropomorphism needed in robots [25]. It also raises discussions in relation to the act of projecting intentions on performing machines and question if this is an inevitable reflex or not (Duffy 2005). When comparing attributions in the field of AI to staged robots, the fictional potentials of the stage and the robots in this environment, have always and always will allow the audience to have more associative attributes rather than literal ones. Normally a literal interpretation by the audience is related to the goal oriented bottom-up approach of AI (Pfeifer 2007) however, complex behavior could emerge from robot morphologies that bear no direct resemblance to zoomorphic entities. This does allow for more free association by the audience.

The author suggests that the definition of anthropomorphism by behavioral scientist Nicholas Epley is more suitable for performing machines. He defines it as a process of inference about unobservable characteristics of a nonhuman agent, rather than descriptive reports of a nonhuman agent’s observable or imagined behavior [26]. Similarly, as American historian Lewis Mummford writes, the machine is a mythical construction. A machine is not only a complex tool but also a social apparatus. It is not only constituted of material pieces but also of immaterial elements, of a mentality and of a belief in a goal or an effect [27].

In relation to these shifting definitions of anthropomorphism, machine performers and puppets share the essential characteristics of being inert entities that are “animated” and “brought to life” in the front of an audience. When Steve Tellis (1992) writes about puppet anthropomorphism, he suggests that the verisimilitude in mimicking human behavior often creates a superficial sense of realism. He further suggests that the illusion of life is better supported from movements exclusive to the puppets morphology. A comparable argument can also be raised in relation to sculptural movement. In the “Morphology of Movement”, kinetic artist George Rickey traces the history of verisimilitude in art and argues that when the artists attempts to abstract and stylize form from reality, they are often more successful [28]. He further suggests that awkwardness and failure to achieve verisimilitude permitted objects to evolve into an artwork. In his terms, kinetic art cannot be served by a direct imitation of nature but by recognition of it laws, awareness of its analogies and a response to the vast repertory of its movement through the environment. Therefore, the interpretation of robots as performers, or staged robots, involves an act of suspension of disbelief as a first and constitutive condition of theatrical reality. The puppet as the machine performer take on their metaphorical connotations because they inherently provokes the process of double-vision, creating doubt as to their ontological status: “What is the nature of its being?”

By sharing these ontological[6] interrogations [29, 30] raised by puppet theorists and by exploring the paradox of the quasi-living, machine performers force to define a set of new ontological states that could become guidelines, in artistic and scientific domains, for both researchers and educationalists in the future.

Kinetic art, usually mechanomorphic, feeds on continuous transformation and participation of the viewer. Shapes of figure 10 were created by a set of discrete manipulators [] where theses geometries are asked to perform to an audience. Beyond the aesthetic of the hypnotic organic movements of these machines, audiences readily address the intent. This weak or shifting anthropomorphism is here an advantage as it frees the “sign from the signified”. It enables a multiplicity of readings from a simple starting shape: an array of cubes.

Figure 10: The Deus Ex Machina character in Devolution. Mechanomorphic (left) and
	Zoomorphic in motion (right).

Figure 10: The Deus Ex Machina character in Devolution. Mechanomorphic (left) and Zoomorphic in motion (right).


Since the concept of environment in AI is limited to the physical world, it does not include the social and the aesthetic potentials. Therefore, by combining AI with Theatre, new question will be raised about how a presentation of an experiment from an AI lab can differ from a theatre presentation of the same machine. Researchers from these disciplines operate from different perspectives; art can become the “new” experimental environment for science because it the world does not only consists of physical attributes but also of intangible realities. Hence, machine performers could constitute an investigation of cultural codes of the audience in order to look at a broader understanding of the machine itself. Perhaps the, theatre directors would start to attribute the character of the actor to the character of the machine or as performance theorist Erika Fischer-Lichte says: from the “state of being” towards the “state of becoming”[31] .


The author wishes to point out the support of Le Musee de la Civilisation (Quebec), the Australian Dance Theatre, Musee d’Art Contemporain de Montreal, Canada and Quebec Arts Council, Ex-Machina/Robert Lepage and the SUG scheme from NTU.


  • J. Horakova, Kelemen J., "Robots between Fictions and Facts," in Intl Symposium on Computational Intelligence and Informatics, 2006, pp. 21-39.
  • J. Reichardt, Robots: Fact, Fiction, and Prediction: Viking Penguin, 1978.
  • G. Skogerson, "Embodying robotic art: cybernetic cinematics," IEEE MULTIMEDIA, pp. 4-7, 2001.
  • D. Ferber, "The Man Who Mistook his Girlfriend for a Robot," Popular Science, 2003.
  • H. Ishiguro, "Interactive humanoids and androids as ideal interfaces for humans," Proceedings of the 11th international conference on Intelligent user interfaces, pp. 2-9, 2006.
  • D. Tomas, "Feebback and Cybernetic: Reimaging the Body in the Age of Cybernetics," in Cyberspace/Cyberbodies/Cyberpunk Cultures of Technological Embodiment, ed London: Sage Publications, 1995, pp. 21-43.
  • N. Weiner, Cybernetics: or Control and Communication in the Animal and the Machine. New York, New York: John Wiley, 1948.
  • Y. Sone, "Realism of the unreal: the Japanese robot and the performance of representation," Visual Communication, vol. 7, pp. 345-362, 2008.
  • P. Auslander, "Humanoid Boogie: Reflections on Robotic Performance," in Staging Philopshy: Intersections of Theatre, Performance, and Philosophy, D. Z. S. David Krasner, Ed., ed: U. of Michigan Press, 2006, pp. 87-103.
  • P. Demarinis, "Mechanization Takes Command: Verses from Giedion's Bible," in Our of Control, ed Linz, Austria: Ars Electronica, 1990.
  • M. J. Gorman, "Between the Demonic and the Miracoulus: Athanasius Kircher and the Baroque culture of machines," in Encyclopedia of Athanasisu Kircher, D. Stolzenberg, Ed., ed Stanford, California: Stanford University Librairies, 2001, pp. 59-70.
  • M. Sussman, "Performing the Intelligent Machine: Deception and Enchantment in the Life of the Automation Chess Player," The Drama Review, vol. 43, pp. 81-96, 1999.
  • L.-P. Demers and B. Vorn, Es: das Wesen der Maschine: European Media Arts Festival, 2002.
  • T. Fong, et al., "A survey of socially interactive robots," Robotics and Autonomous Systems, vol. 42, pp. 143-166, 2003.
  • M. Mori, "Bukimi no tani [the uncanny valley]," Energy, vol. 7, pp. 33–35, 1970.
  • C. DiSalvo, et al. (2005, Imitating the Human Form: Four Kinds of Anthropomorphic Form. Available:
  • S. Dixon, "Metal Performance Humanizing Robots, Returning to Nature, and Camping About," TDR/The Drama Review, vol. 48, pp. 15-46, 2004.
  • C. Breazeal, "Toward sociable robots," Robotics and Autonomous Systems, vol. 42, pp. 167-175, 2003.
  • R. A. Brooks, Flesh and Machines: How Robots Will Change Us: Pantheon Books, 2002.
  • R. Pfeifer, et al., How the body shapes the way we think: a new view of intelligence: MIT Press, 2007.
  • I. Fumiya, et al., "Design and Control of a Pendulum Driven Hopping Robot," Intl. Conference on Intelligent Robots and Systems (IROS 02), 2002.
  • A. Michotte, The Perception of causality. London, UK: Methuen, 1963.
  • F. Heider and M. Simmel, "An experimental study of apparent behaviour," American Journal of Psychology, vol. 57, pp. 243-259, 1944.
  • L.-P. Demers and J. Horakova, "Antrhopocentricism and the Staging of Robots," in Transdisciplinary Digital Art. Sound, Vision and the New Screen. vol. 7, S. Gibson, et al., Eds., ed: Springer , 2008, pp. 434-450.
  • B. R. Duffy, "Anthropomorphism and the social robot," Robotics and Autonomous Systems, vol. 42, pp. 177-190, 2003.
  • N. Epley, et al., "On Seeing Human: A Three-Factor Theory of Anthropomorphism," Psychological Review - NY-, vol. 114, p. 864, 2007.
  • L. Mumford, The Myth of the Machine: technics and human development. New York, New York: HBJ, 1967.
  • G. W. Rickey, "The Morphology of Movement: A Study of Kinetic Art," Art Journal, vol. 22, pp. 220-231, 1963.
  • C. Emmeche, The Garden in the Machine: The Emerging Science of Artificial Life. Princeton, New Jersy: Princeton University Press, 1994.
  • T. Gruber, "A Translation Approach to Portable Ontologies," Knowledge Acquisition, vol. 5, pp. 199-220, 1993.
  • E. Fischer-Lichte, Ästhetik des Performativen. Frankfurt am Main, 2004.

I. Multiple Ontologies of the Machine Performers.

Theatre / Semiotics
History AI Uncanney Valley Anthropomorphism Animacy and Causality
Body / Enrichment Understanding of internal mechanisms Outsourcing of intellignece Quest for the "perfect double" Stzarting from the body. The predisposition into the capabilities of the agent from its shape Subtracting the body from the shape
Performance (performative actions) / Stage
Co-presence of audience
Using theatricaility Implementation of natural actions The uncanny effect of verismilitude of movements Cognitive dissonance in non-anthropomorphic agents.
The predisposition into the cpabilities of the agent from its movements
Passage from pure movement to judgement, i.e., automatic to the agent
Specification of the situation
Utopian Social Concerns
Imitation and the Fake
Staging Realism
Context is included in intelligence Japanese glorify their machines as westerner express anxiety. Theatre stage is outside normal lab contexts. Belief systems of today
The predisposition into the capabilities of the agent from its context
Isolated in experiments the field but in our case, we shift it to the theatre stage
Subject / Object
Agentic / Automatic
From combined fragments of history Life as it could be?
Dividing line of the human and non-human Belief systems shifting interpretation of the object Function vs. Intention vs. Aestehtics/td>

  1. The subtitle refers to the monographic work Neither Bachelor Nor Brides about the hybrid machines of Rebecca Horn
  2. In a proscenium theatre, the term fourth wall refers to the imaginary invisible wall between the stage (universe of the play) and the audience. It was made explicit by Denis Diderot and spread in nineteenth century theatre with the advent of theatrical realism.
  3. Empirical and qualitative observations were made in the museum by the author (personal archive).
  4. Psychologist Susan Blackmore attempts to constitute memetics as a science by discussing its empirical and analytic potential
  5. Psychologist Gunnar Johansson (1973) devised an experiment, the Point Light animation (PL animation), to study the information carried by biological movements.
  6. As regarded by the field of Artificial Life, an ontology defines how the world in which the agent lives is constructed, how this world is perceived by the agent and how the agent may act upon this world by combining AI with Tipulators Devolution. Mechanomorphic (left) and Zoomorphic in motion (right).to breach affinity.