eipcp Publications A Thousand Machines

New Robot Revolution, Multi-Agency and the Machinic

Review of Gerald Raunig’s A Thousand Machines (semiotext(e), 2010, 120 pp)

Samuel Gerald Collins

Samuel Gerald Collins


[Excerpt from Samuel Gerald Collins' review. The full text can be found at http://www.irma-international.org/article/new-robot-revolution-multi-agency/52095/]


The challenge is to consider distributed multiagency, where an agent’s actions and cognitions are distributed across other agents, a continuously expanding, even stochastic, open system.  What are needed are conceptual tools for thinking about multiagent systems in their broadest social and cultural contexts. 

Gerald Raunig’s A Thousand Machines suggests just architecture of ideas.  A cultural critic focusing on art and activism, he is nevertheless uniquely positioned to theorize on the machinic.  Really, his work is a bit of a machine itself, taking the “machinic” from Deleuze’s and Guattari’s Anti-Oedipus and Thousand Plateaus and mobilizing it for art and life.  Famously opaque and tantalizing, Deleuze and Guattari's “nomadology” is meant to resist easy assimilation into commodified popular culture.  Raunig’s task is to take the machinic and make it applicable to art and activism—to the kinds of emergent socialities he finds with European anarchists and autonomous movements. 

Here, he elaborates on the chiasmatic appearances of the machinic in Deleuze and Guattari, which has (as a war machine) an emancipatory and an oppressive dimension.  But this ambiguity is key to the machine.  As he reminds us, the term had two valences in the Latin:

On the one hand there was the military use as an apparatus for besieging, conquering or defending cities, in other words as a war machine, while on the other hand it was also used as a comprehensive term for the machinery of the theater.  (36)

Ultimately, the “machinic” contains the germ of both meanings, tracking between power, resistance and spectacle: “In both cases the application of the term both holds the technical meaning of apparatuses, frames devices as well as the psychosocial meaning of trick, artifice, deception” (36-37).  Raunig spends the course of this slim volume unraveling (and tying) the strings of this etymology together.    

Why should multiagent systems researchers pay attention to this?  The same qualities he attributes to German autonomen are readily evident in the mediated social actions of everyday life.  Consider the coordination of a Friday happy hour downtown—5 friends embark from their homes and offices in 5 different parts of the city without specifying in advance where they will meet. Along the way, ample text messaging, perhaps some locational applications that help locate friends, together with information on traffic and timetables enable their meeting to emerge from the (loosely) coordinated actions of a variety of agents (Urry 2007).  A machine?   Certainly the ambiguity of the “machine” includes at once technology, but moves beyond it to include thinking, action, strategy, speech—the machine at the collision of people, machines, communications, social networking. 

What this understanding moves beyond is our sense (no more pronounced than in anthropology), where technologies are thought to merely extend our techne (τέχνη ), as it were—writing extending our memory, the lever extending our strength.   As Raunig points out (32),

It is no longer a matter of confronting man and machine to estimate possible or impossible correspondences,   extensions and substitutions of the one or the other, or ever new relationships of similarity and metaphorical relations between humans and machines, but rather of concatenations, of how man becomes a piece with the machine or with other things in order to constitute a machine.  The “other things” may be animals, tools, other people, statements, signs or desires, but they only become a machine in the process of exchange, not in the paradigm of substitution.

I think this is the big difference between the machinc Raunig elaborates and some of the more cyborg/cybernetic theorizing of previous decades.  That is, he offers a succinct account of the move from substitution (I am becoming a machine) to networked exchanges, where agents swap information, communication and characteristics in variously coupled exchanges along lines of affinity, rather than filiation and genealogy.

Ultimately, this is not a closed system of exchanges, in at least two senses.  First, machinic assemblages can incorporate any number of additional agents and agencies, like adding additional nodes in an infinitely scalable network. 

The machine is not limited to managing and striating entities closed off to one another, but opens up to other machines and, together with them, moves machinic assemblages.  It consists of machines and penetrates several structures simultaneously.  It depends on external elements in order to exist at all.  It implies a complementareity not only with the human being that fabricates it, allows it to function or destroys it, but also exists in a relationship of difference and exchange with other virtual and actual machines. (33)

We could say that the image is one of montage (an image that Raunig uses to describe machinic theater), but that would belie the extent to which the machine is open, accreting additional relationships, machines and affiliations.

The second chief characteristic is the centrality of emergence, where every “catcactenation” of agency generates the conditions for new, novel assemblages irreducible to constituent elements, “power of invention, in the capacity for change, in the creation of other worlds” (57-58).  This is the essential meaning of “lines of flight,” the way that the machinic extends beyond into new forms of actualizations.  An assemblage of elements by no means exhausts the potentialities for different formations and forms of action.  Rather, the machinic assemblage opens up countless other virtual opportunities for actualization. 

But where do we find these machines?  Here, Raunig gives us two areas to consider.  First, the theater, where he finds intimations of the machinic in some of the experimental theater of the early Soviet Union, in particular the “Theater of Attractions” of 1923-1924.  Through the  efforts of its activists writers, directors and actors, including the work of a young Sergei Eisenstein, the theater built assemblages out of autonomous elements of theater, including both scenery and audience.   Not, in other words, the reduction of one element to another, but the interest in producing an expanding, emergent assemblage of agencies. 

For example, in Tretyakov’s 1923 play, Can you Hear, Moscow?, its montage of human and object “so inflamed especially the youthful audience and the extras that the actors playing the bourgeois were attacked even during the play; following the conclusion of Can you Hear, Moscow?, the mobilized audience purportedly poured into the streets in tumultuous scenes, singling and “wildly flailing against the shop windows” (55).   Now part of the theater machine, the audience moves outside in an expansive assemblage that threatens to engulf the city. 

The second site of the machinic draws from what Deleuze and Guattari have described as the “war machine,” not in its more molar incarnation in the striated war machine of the state, but in the sense of an undisciplined (in the Foucauldian sense) groups of activists expanding across the landscape, like the PublicTheatreCaravan’s protests against the G8 summit in Genoa, or similar efforts to transgress state-inscribed boundaries through a performative direct action.  

To maintain that the Caravan—as I wrote at the time—operates on a line of flight, offensively as a war machine, does not at all mean attributing a special form of violence to it.  On the contrary, the war machine points beyond the discourse of violence and terror, it is the machine that seeks to escape the violence of the state apparatus, the order of representation.  (58)

The Caravan is made up of any number of variously autonomous human agents operating through affinity rather than through vertical control.   Theorizing exactly who these machinic agents are takes up the next couple of section in Raunig’s essay.  Instead of agencies built on stable identities (as citizens, as activists, as techniques), he focuses on a “precariate” made up of human agents that would have been described in preceding decades as a lumpenproletariat, what for Marx amounts to that wild, ultimately ignored, amalgam of diverse, archaic and chaotic elements that are unable to come to any realization of class identity.

Unlike the image of the sleeping giant of the proletariat, which must be awakened through class consciousness and a political party, the precariat is a monster that knows no sleep.  There is no teleological movement here from sleeping to class consciousness; there is neither the empiricism of the class itself nor the political innovation of a class for itself, but rather a constant coming, questioning struggling.  (104)

That is, the machinic combines in intersecting trajectories that transcend official, state-sanctioned orders and identities, releasing agencies in a recombinant cascade resembling nothing so much as Rodney Brooks’s and Anita Flynn’s model for multi-agent planetary exploration using thousands of what they then were calling “gnat robots”:

But how should the gnats be spread out over the surface?  On Mars they could blow in the wind.  Alternately they could locomote themselves by hopping.  They could use solar cells to collect energy and store it in a silicon spring.  After a certain level of compression, the sensor fires and the spring is let go, and the robot goes flying.  Wherever it lands, it checks for the desired compound.  If it finds it, it anchors itself and puts up its corner reflector.  Otherwise, the robot continues its stochastic search of the planet, for years perhaps.  (Brooks and Flynn 1989: 484)

Of course, given Raunig’s oeuvre, it is not surprising that he concentrates on human elements of the machinic assemblage.   My one wish for this engaging essay is that it had taken on the question of the non-human, since it is the “machine” side of the machinic that enables the kinds of political (dis)organization he outlines.

When we examine contemporary protests against global capital, one of the most striking differences between this and the civil rights era Gladwell lionizies is most certainly the prominence of information and communication technologies, from the “Indymedia” centers that disseminate, connect and focus affinity on particular sites of protest and direct action, to the hand-helds that facilitate anything from on-the-fly organizing to documenting police brutality.  Consider Juris’s description from the 2000 protests against the World Bank at Prague:

The emerging terrain of resistance was thus divided into various color-coded zones, each providing a space for diverse kinds of political expression enacted through distinct embodied performances.  Different styles and different forms of direct action reflected divergent political logic, inscribing contrasting political meanings on the urban and mass-media landscapes.   [ . . .] The overall diversity-of-tactics strategy would reorder a networking logic on the theatrical plane as emerging networks norms and forms were transposed onto urban space and expressed through horizontal coordination among distinct protest performances. (Juris 2008: 131) 

We would find in this an example of swarming behavior.  And that’s the point.  With the help of cell phones with internet capability, autonomen and other activists can (loosely) couple their behaviors based on a dynamically changing environment that includes other technologies and other activists.  It’s the self-conscious application of a distributed, multi-agent systems thinking to political action.   To go back to Gladwell’s attack—I think it would be a mistake to concentrate one’s attention on the ways in which multiagent systems simulate or mimic human lives—when what we’re really dealing with are novel forms of behavior that exist in the interstices of other forms of human action.  That is, those strong ties whose loss Gladwell bemoans still exist, but exist alongside more multiagent systems that mediate between human and non-human agents through weak ties.

But what’s clear from Raunig’s work here is that multiagent systems research can only assume greater and greater importance as more and more agents proliferate in our lives.  It is not too much to say that the task of “simulation” will gradually bleed into that of actual description.  For example, a multiagent simulation of automobile traffic that models decision-making on the part of simulated drivers becomes a description of the way multiagent-agent systems combine with human agents to minimize accidents and traffic conditions.   Or, more than a description, it becomes illustrative of emergent properties for multiagent systems that might suggest new machinic assemblages towards new ends.  Here, the political implications for MAS lie with the evocations of a new existence altogether.