Avant-Garde and Comics
Serious Cartooning
by M. Todd Hignite

Robotic Performances at the Venice Biennial

by Philip Auslander

Sergei Shutovâs Abacus (2001) in the Russian pavilion of the 49th Venice Biennial (June 10-November 4, 2001) was a frequent subject of discussion during the press opening for the exhibition. Abacus consists of over 40 crouching figures draped in black, which face an open door and pray in numerous languages representing a multitude of faiths while making the reverential movements appropriate to prayer. Nearby video monitors display the texts of the prayers in their many alphabets. People at the opening talked of the ãperformanceä in the Russian pavilion; a colleague, knowing that performance is the main subject of my research and writing, asked me whether I considered the piece a performance. I blithely answered yes, realizing only later that I had taken a position I needed to consider further.

Sergei Shutov, Abacus, robotic performance, 2001.

The reason for both my colleagueâs question and my own desire to think more about it is that the figures performing in Abacus are not human beings-they are robots programmed by a computer to engage in dahvening movements accompanied by the recorded sounds of ecumenical prayer. Given that the figures are machines, not human beings, some might argue that the piece should be considered an animated sculptural installation, not a performance-it is described as an installation in the Biennial catalog. I prefer to think of it as a performance, however, not just because I believe that machines can perform but also because to view a piece such as Abacus as a performance by machines yields rich possibilities for its interpretation.

At the most basic level, the question ãCan machines perform?ä can only be answered in the affirmative. After all, the primary meaning of the verb ãto performä is simply ãto do.ä Inasmuch as machines (or human beings) do things, they perform. Moving from that basic level to the context of art practices, however, the definition of performance proves to be context-specific, not universal; it changes according to the particular aesthetic form and tradition under consideration. What it means to perform a piece of classical music is not the same as what it means to perform jazz, and neither musical definition of performance is applicable to the theatre or dance. One crucial area of difference is the assumed relationship of the performer to the text being performed: the relationship of a classical musician to the piece is not the same as that of a jazz musician to the music she performs, for instance, and the relationship of actors to dramas and dancers to choreography puts still other variables into play. One element common to most traditional definitions of performance, however, is an emphasis on the agency of the performer as the interpreter of the text and an artist who expresses something of her own through interpretation. The definitions of performance arising from the context of the traditional performing arts entail notions of interpretation and expression that exclude machines from being considered performers. At present, there probably is no machine capable of presenting a performance of the Kreutzer Sonata, Hamlet, or Swan Lake that would pass muster with the relevant audiences.

But since the robotic pieces I saw at the Biennial clearly do not belong to any of the traditional performing arts, it would make little sense to analyze them in terms of the definitions of performance emerging from that context. The context of performance art, a constellation of performance genres pioneered largely by visual artists rather than performing artists, seems the appropriate one in which to consider work that is presented as visual art and can be classified as sculptural installation. In the aesthetic context of performance art, the definitional picture is quite different than in the traditional performing arts, for performance art can involve a multiplicity of types of performance ranging from those found in conventional music, theatre, and dance to others in which expression and interpretation are much less important. Writing about one of the earliest genres of American performance art, the Happenings of the late 1950s and early 1960s, performance theorist Michael Kirby coined the term ãnonmatrixed performingä to describe the kind of performing he saw in many Happenings and to distinguish it from acting. In nonmatrixed performing, the performer does not represent anything other than herself, doing whatever sheâs doing, wherever, whenever, and in whatever situation sheâs doing it. As Kirby states in the introduction to his 1965 book Happenings, the only thing asked of the nonmatrixed performer is: ãthe execution of a generally simple and undemanding act. [ . . . ] The creation was done by the artist when he formulated the idea of the action. The performer merely embodies and makes concrete the idea.ä In keeping with the idea of the Happening as a form of visual art, the performer is used primarily as an element in a visual composition that unfolds in time and space.

Because nonmatrixed performing consists of the more or less mechanical execution of tasks designated by someone other than the performer and does not call upon the performer to interpret a text or be self-expressive, it is a kind of performance in which both human beings and machines can engage. The figures in Abacus are nonspecific, black-clad bodies (as opposed to representations of particular characters); because it is not necessary that they actually pray (whatever that may mean), only that they appear to do so, their actions may be treated as nonmatrixed even though the figures are representational. Human performers could execute the same movements as the robots simply by carrying out the artistâs instructions-the piece does not require that they enact characters who are praying or that they actually pray themselves.

Two other pieces on exhibit at the Biennial provide further examples. Max Dean and Raffaello DâAndreaâs The Table: Childhood (1984-2001) is described by its creators in the Biennial catalog as ãa fully autonomous robotic tableä capable of movement through a confined gallery space. I will paraphrase the action for which it is programmed from the catalog and render it as an instruction: ãSelect a viewer and attempt a relationship with that person.ä The robotic table does this by following a chosen person through the space and performing movements that are meant to be ingratiating. A human being clearly could perform the task for which the table is programmed as well. I am not saying, of course, that a hypothetical human performance of this piece would be similar to the performance of the robotic table, only that both robots and human beings are capable of giving performances that reflect the underlying programming. Expressed as a simple verbal instruction, this programming sounds very much like the score for a Fluxus performance, for instance, or a performance by Vito Acconci.

The Table is a particularly interesting case because the performer, whether machine or human, is called upon to make choices-the exact shape of each iteration of the piece depends on which person the performer chooses to follow, that personâs reaction to being courted, and so on. In this respect, the robotic table contributes much more to defining the performance than Shutovâs praying figures. Even though these choices determine what happens in a given performance of the piece, they are not interpretive choices. As Kirby indicates, nonmatrixed performing frequently includes an element of indeterminacy, which he distinguishes from improvisation. For Kirby, improvisation demands that the performer make interpretive decisions on the fly, decisions that shape the performance in significant ways. Indeterminacy, by contrast, means that although the artist leaves certain aspects of the performance open-ended, the performerâs decisions are neither interpretive nor significant-any decision the performer makes within the parameters established by the artist will yield an iteration of the piece that reflects the artistâs intentions. Because a successful execution of the action that underlies The Table does not depend on factors like which spectator is selected and followed, exactly how the spectator reacts, and so on, that decision falls into the realm of indeterminacy rather than that of improvisation. It thus exemplifies the kind of noninterpretive decision that can be made either by a human being or a machine.

Whereas The Table, performed by a robot, could be performed by a human being, Nedko Solakovâs A Life (Black & White) (1999-2001), performed at the Biennial by human beings, could just as readily be undertaken by machines. The catalog description reads ãBlack and white paint; 2 workers/painters constantly repainting in black and white the space walls for the entire duration of the exhibition, day after day (following each other).ä (One painter paints the walls black; the other follows and paints the same walls white, and so on.) Because all three of these pieces are based in simple actions that can be executed effectively either by machines or human beings and because their impact does not depend on aesthetic decisions made by the performers (as Kirby indicates, the performersâ actions are only concretizations of the artistâs choices), it is not necessary that the actions be carried out by human beings in order that the piece be considered a performance. Whether the actions are undertaken by a person or a robot, the event is recognizable as a kind of performance belonging to the tradition of performance art.

Having argued that the events discussed here are characterized by nonmatrixed performing, a type of performance that can be undertaken by both human beings and machines, and that it doesnât matter to their constitution as performances within the tradition of performance art which type of performer is involved, I hasten to add that it does make a substantial difference to the interpretation of the pieces. Far from suggesting that human agency makes such pieces richer in meaning, I propose that machine performance, acknowledged as such, provides the deeper object of interpretation. The reason is simple: because machines generally can be viewed as surrogates for human beings or metaphors for human concerns (indeed, what else could they be?), a piece performed by a machine usually can be interpreted in the same way as the same piece would be if performed by a human being. (This is true even for an apparently nonanthropomorphic performance like The Table, which is subtitled ãChildhoodä because its creators see the object as reenacting the human developmental process.) The question of what it means to have the actions that define the piece performed by a machine provides for a further, enriched level of interpretation. In the Biennial catalog essay on Abacus, for example, Sergei Khripun reads Shutovâs piece as a conciliatory gesture ãbringing alienated and even confronting confessions together,ä while noting at the same time that religious belief both unites and divides people-ãthe paradox that humankind has faced throughout its history.ä It is interesting that this humanistic interpretation makes no reference to the fact that the piece is actually performed by machines; Khripun treats the robots as transparent representations of human beings whose own ontological status and materiality are irrelevant. I am not disputing Khripunâs interpretation; I am suggesting that acknowledging that the piece is performed by machines would add another layer of interpretation by inviting the critic to consider what it means to have machines serve as surrogates for human beings at prayer. What does the metaphoric use of machines signify in this particular context?

In his catalog essay on Solakovâs Life (Black and White), Daniel Kurjakovic suggests that the piece is ãan allegory of the abysmal, Sisyphus-like futility of human action. . . .ä This interpretation is certainly available when the piece is performed by human painters, but I would argue that it is equally available if the piece were to be performed by robotic painters. The futility of the robotsâ actions would almost inevitably be seen as metaphoric for human existence. Addressing the further question of what it means to deploy robots in the senseless task of continually repainting walls would open up other areas of interpretation that might enhance or expand this reading. Is it just as ironic, in an Existentialist sense, for a human being to program a robot to undertake futile and meaningless work as it is to assign that work to other human beings? Or does the use of a robot displace the futility of existence onto an entity that does not suffer from experiencing it, thus suggesting a path of liberation for human beings through a somewhat perverse use of technology? There are many possibilities for interpretation that arise from defining such a piece as a performance undertaken by machines and addressing directly both the ways in which the machines can be seen as metaphoric humans and the implications of using a machine in the particular context of the piece.