Digital Labor The Internet as Playground and Factory

transcribing audio, tagging images with descriptive text, or answering surveys on various topics. The products of this labor might serve many purposes ranging from spam generation to training machine learning algorithms that would eventually assume some of these human roles in the future. The payment amount per HIT ranges from one cent to several U.S. dollars, depending on the required time or dificulty of the task.

Amazon.com’s initial motivation to build Amazon Mechanical Turk (AMT) emerged after the failure of its artificial intelligence programs in the task of finding duplicate product pages on its retail website (Pontin 2007). After a series of futile and expensive attempts, the project engineers turned to humans to work behind computers within a streamlined web-based system. Later, AMT made this cognitive labor platform available to private contractors in return for a commission for each completed HIT. AMT’s digital workshop emulates artificial intelligence systems by replacing computing with human brainpower. Driven by what AMT calls “artificial artificial intelligence,” this sociotechnical system represents a crucial formation on a global scale as it facilitates the supply of cognitive labor needs of mainly Western information and communication technologies industries from a global workforce.² AMT explains the value of its labor market for the software industry as follows:

With Amazon Mechanical Turk, it may seem to your customers that your application is somehow using advanced artificial intelligence to accomplish tasks, but in reality it is the “Artificial Artificial Intelligence” of the Mechanical Turk workforce that is helping you efectively achieve your business objectives. 3

According to Panos Iperiotis, approximately half of the AMT workers—or Turkers, as some of them prefer to call themselves—are from the United States, and the other half are from 66 diferent countries. Most of the non-U.S. Turkers are from India, representing 33% of the AMT workforce (Iperiotis, 2010). Ross et al. have demonstrated that the demographics of AMT are becoming increasingly international, highlighted by an expanding group of young, male, Indian workers who make less than US\$10,000 per year. About a third of Indian workers reported that they partly rely on AMT “to make basic ends meet” (Ross et al., 2010).

Amazon.com branded its micropayment-based crowdsourcing platform as the Mechanical Turk, borrowing one of the names of the 18th-century Automaton Chess Player as shorthand for the relationship that the system establishes between the cognitive labor force and the seemingly automated complex tasks. In both cases, the performance of the workers who animate the artifice is obscured by the spectacle of the machine.

The idea of the chess-playing machine, which was realized by IBM’s Deep Blue computer in 1997 has been a key conceptual apparatus for imagining the automatization of the operations of the human mind since the Enlightenment era. This metaphor was also central for the idealization of cybernetic discourse as a universal system during the first half of the 20th century, embodied by the postwar symbol processors, which later became the architectural basis of the contemporary computer (Bowker 1993; Shannon 1950). In this chapter, I will study a neoliberal reincarnation of the chess-playing automaton, Amazon’s Mechanical Turk, in the light of its early-modern legacy of configuring the relationship between the division of cognitive labor and the automatic systems of computing and control. Particularly, in both cases, in varying degrees and methods, the labor performance of intellectual workers is an integral element of the disciplinary structure of the corresponding socioeconomic apparatus.

In the current configuration, this cognitive labor apparatus is situated within the neoliberal system of exception facilitated by the digital networks, taking advantage of legislative gray zones in the international labor regulations in order to maximize profits for multinational corporations (Ong 2006). Crowd-sourcing is one of the most significant elements of this configuration that expands the reach of the neoliberal economy through cognitive capitalism,⁴ in which immaterial labor plays a key, structural role.⁵ Not surprisingly, this configuration also embodies some of the conflicts whose seeds are placed during the early modern conceptualizations of the mechanization of industrial labor through division of cognitive labor. One of the most significant examples of this conceptualization was the chess-playing automaton that performed the insurmountable conflicts of the disciplining of the human mind for industrial production.

Docile Automata

Wolfgang von Kempelen’s Chess Player Automaton was constructed and presented in 1770 at the court of the Empress Maria Theresa of Austria and gave the

impression that the pipe-smoking Turk mannequin, controlled by a sophisticated mechanism under the cabinet, could play serious chess against human opponents (see Figure 5.2). However, the seemingly mechanical mind of the Turk was actually manipulated by Kempelen’s chess master assistant, who was hidden beneath the pseudo-mechanism. The Automaton Chess Player was exhibited for 84 years in Europe and the Americas and attracted many notable challengers and spectators, such as Napoleon Bonaparte, Charles Babbage, and Benjamin Franklin (Carroll 1975; Windisch 1784).

In his book, Discipline and Punish, Michel Foucault considers the 18th-century automata as models for human body and social order (Foucault 1977). Consequently, the mechanistic conception of human body needs to be read in two registers: the anatomicometaphysical register as constituted mainly through Cartesian mind/body duality, and the technico-political register that was constituted by empirical methods of the state for disciplining the operations of the body through the army, the school, and the hospital. In the context of these two registers, the 18th-century humanoid automata functions as a model, on the one hand, for submission and use and, on the other, for empirical analysis.

Foucault has often been criticized for ignoring the racial others in his historiography.⁶ Notably, his concept of docility displaces Orientalist traces by solely focusing on the European subject in a selective genealogy. This absence becomes more critical in the analysis of an automaton that carries significations of Oriental “other,” such as Kempelen’s chess-playing automaton. However, I believe that the trick of the chess-playing automaton involves more than just exchanging the enacted body of the European chess player with the represented body of the Turk, animated through its mechanical artifice. It also includes initial assumptions that were set up in the audience by the automaton’s chess performance that were crucial in influencing the public debates on the mechanized reason that provided the larger context for these performances. These initial assumptions are closely tied to Orientalist undercurrents that were exploited by Enlightenment discourse in order to configure the docile subject on the image of the Turk.⁷

The Orientalist assumptions that were active in Enlightenment automata were also efective in the cultural performance of Kempelen’s automaton. I will focus on the two main aspects of the afordance of the image of the Turk as a significant part of the main interface of the chess-playing automaton.⁸ The first critical aspect of the Turk’s performance is its liminal quality. This liminality created a bufer zone against the risk associated with the idea of the man-machine that most Enlightenment humanoid automata performed. That potential risk was often associated with instigations of libertinism, atheism, and insurrection in public due to the heretical understanding of a body without a soul (Vartanian 1960). Relegating this precarious role to an Oriental figure had, in fact, a long tradition with origins in medieval romance literature (Truitt 2004). The Oriental automata, through its association with liminal spaces and experiences in these literary accounts, conveyed surveillance, discipline, and enforcement of limits of morality.

The second aspect of the Turk’s performance is a particular form of docility that conveys the idea of the disciplined productive body, which played a salient role in the formation of the enlightened culture (Schafer 1999). The association of the Oriental with docility has its roots in medieval theology, where the Muslim subjects were considered as strict followers of religious code. Linking this association with the discourse of Oriental automata, Christian theology configured a particular discourse of Muslim as automaton (Biddick 2011). Furthermore, docility prefigures the hidden chess player’s performance of the intellectual labor on behalf of the Oriental automaton. This dual performance of docility highlights the question of the intellectual labor in the context of the epistemic renovation in 18th-century Europe (Foucault 1977). These two aspects of the Turk’s performance—docility and liminality—are crucial for grasping its function as a model of power for the idealization of a social order in the context of the large-scale processes of mechanization of labor in Europe in the 18th century.

The chess-playing automaton performed its role as a model of power in multiple layers, the first of which was the demonstration of knowledge as a tool of power. The hidden chess player was the open secret of Kempelen’s (and later Maelzel’s) shows (Carroll 1975). Kempelen admitted that his automaton was just a “happy deception” (Cook 1995). As Schafer notes, one of the roles of the Enlightenment automata was “to allow the selective entry by th[e] power to the inner workings of art and nature” (Schafer 1999: 135). In other words, this open secret was also a conceited wink of the guardians of knowledge and power, reminding the general public about their privilege and status.

The element of mystery in Kempelen’s performance can be considered within the system of representation of the natural philosophy, which perceived the whole of nature as a divine theater. The efects of this system of representation could be exploited to create a particular moral impression on its audience (Schafer 1983). Scottish philosopher Thomas Reid expounds this moral efect as follows: “Upon the theatre of nature we see innumerable efects, which requires an agent endowed with active

power; but the agent is behind the scene.”⁹ Kempelen’s automaton benefited from the assumptions within this theater as a significant representation of the technomythical idea of the mechanized mind, and was not just a machine but also provided the language that made it possible to explicate that myth (Beaune 1989). As in every technical medium, it carried the inscriptions of discursive traditions and formulations that defined its cultural system of significations. The Automaton Chess Player performed these inscribed notions that were formulated as technical puzzles, which have remained relevant throughout the history of the mechanization of the mind. The puzzle of the mechanical chess player, or the mechanized symbol processing, has been tackled by many notable scholars and practitioners, including Gottfried Leibniz, Edgar Allan Poe, Charles Babbage, Norbert Wiener, and Alan Turing.

In Europe in the second half of the 18th century, automata performed as a secure experimental apparatus for exploring impenetrable ontological liminalities in a more systemic way and most of the time simulated life in order to redefine it (Riskin 2003). Fueled by the mechanistic philosophy, humanoid automata transformed not only the cultural attitude toward living creatures but also machines, as they performed the idea that mechanisms were also living beings. The mutual relationship between the animation of machinery and the mechanization of life was explored through the experimental apparatus of humanoid and animal automata and was popularized through the debates instigated by their public exhibition in Europe.

Wolfgang von Kempelen’s Chess Player Automaton formulated the question of the mechanized life with a unique emphasis: Can the mind exist without the body? To this question, it gave two answers simultaneously: yes and not yet. The actual answer was not yet, as the automaton was indeed controlled by a human operator. However, the deceptive yes response was still valuable as a philosophical game¹⁰ for grappling with the ideas that were later made technically possible and implemented systematically, such as self-regulating mechanisms.

In contrast to other automata of the 18th century, the Turk’s apparatus did not act as mere clockwork; instead, it gave the impression of a self-regulating system that could counter external actions within the symbolic logic of chess.¹¹ As historian of technology Otto Mayr (1970) suggests, in contrast to the idea of clockwork universe, which was the political universe of autocratic feudalism, the mechanical, political, and economic ideas of selfregulating systems influenced the Enlightenment ideas of liberal subjects and democracy. This association is partly constructed as a result of the rationalization of the socioeconomic life through industrialization, where subjects self-regulate according to their rational economic interests. Philosophical arguments for such conflict-free social systems had already been provided by various scholars, the most famous of whom was Gottfried Leibniz.

The preceding century had seen Leibniz’s proposal of a universal symbolic language or algebra of thought. Since the expansion of the commerce in Leibniz’s time, there had been a search for a universal language that would allow European traders to establish a sustainable communication with the people in the new colonies. Leibniz’s universal language could be manipulated by a logical calculation framework, calculus ratiocinator (ca. 1680), which was a precursor model of modern computing (Wiener 1948). Leibniz suggested that the mind is a spiritual automaton that operates involuntarily based on a predetermined set of laws. “The operation of spiritual automata, that is of souls, is not mechanical, but it contains in the highest degree all that is beautiful in mechanism” (Leibniz 2005: 365). However, the automaton/self-moving soul does not eliminate agency for Leibniz, because symbols and the symbolic systems of language play a constitutive role for reasoning. Based on this principle, Leibniz proposed calculus ratiocinator as an ultimate solution for all conflicts between the people of the world. This perspective finds its expression in Leibniz’s Machiavellian motto calculemus (let us calculate!): “if controversies were to arise there would be no more need of disputation between two philosophers than between two accountants. For it would sufice to take their pencils in their hands, and say to each other: Calculemus.”¹²

Chess is a perfect example for computable symbolic systems; consequently, when the Turk spoke the language of the symbolic via chess, it entered “the world of the machine” (Lacan 1991: 47). But that machine denoted a particular type of subjectivity because of the nature of the actors and their limited set of behaviors that are strictly defined within a set of rules in the game of chess. As Deleuze and Guattari state:

Chess is a game of State, or of the court: the emperor of China played it. Chess pieces are coded; they have an internal nature and intrinsic properties from which their movements, situations, and confrontations derive. They have qualities; a knight remains a knight, a pawn a pawn, a bishop a bishop. Each is like a subject of the statement endowed with a relative power, and these relative powers combine in a subject of enunciation, that is, the chess player or the game’s form of interiority…. Within their milieu of interiority, chess pieces entertain biunivocal relations with one another, and with the adversary’s pieces: their functioning is structural…. Chess is indeed a war, but an institutionalized, regulated, coded war. (Deleuze and Guattari 2004: 352)

Consequently, an automatized chessboard represents the ideal Enlightenment universe, where the subjects and their possible actions are coded according to the regulations informed by the power structure of the society. Each subject is endowed with a relative power, and they cannot go beyond the roles for which they qualify. Particularly, when these intrinsic properties are abstracted into geometric functions and when combined with the functions of other subjects, they have the potential to exhibit numerous but finite possibilities for a final outcome. This is another reason for mechanized chess being a model for imagining a society whose coded subjects articulate a plurality of results. Thus, the chess-playing Turk embodied an integration of the selfregulating liberal subject with the mechanical docility of the Oriental, performed within the coded socioeconomic universe of the game of chess.

In the mid-19th century, during the Turk’s tour in the Americas, Edgar Alan Poe took this seemingly conflicting performance for his argument of why an automaton chess player would be impossible. In an editorial published in the Richmond-based Southern Literary Messenger, he surmised that “[n]o one move in chess necessarily follows upon any one other. From no particular disposition of the men at one period of a game can we predicate their disposition at a diferent period” (Poe 2009: 1). This perceived paradox was mainly due to the assumption Poe had for the possible operational principle of the automaton, which was mainly based on the linear mechanism paradigm, where the interaction between the mechanism and the environment is not a relevant factor during its operation. Poe explicitly based his argument on a comparison of the performance of the chessplaying automaton with Charles Babbage’s calculating machine and concluded that:

There is then no analogy whatever between the operations of the Chess-Player, and those of the calculating machine of Mr. Babbage, and if we choose to call the former a pure machine we must be prepared to admit that it is, beyond all comparison, the most wonderful of the inventions of mankind. (Poe, 1910: 84)

Poe’s resistance to the idea of mechanized thinking may also be related to a particular kind of predicament that concerned his intellectual labor. The very possibility of chess automaton as a “pure machine” must have posed an uncanny prospect for Poe as an intellectual worker. The designer of the actual calculating machine, Babbage, had already implicated the particular kind of skill and labor that Poe uses for his intellectual work as part of the mechanization and division of the cognitive labor system. Charles Babbage specifically used newspapers as an example

of a manufactory in which the division of labor, both mentally and bodily, is admirably illustrated, and in which also the efect of the domestic economy is well exemplified. It is scarcely imagined, by the thousands who read that paper in various quarters of the globe, what a scene of organized activity the factory presents during the whole night, or what a quantity of talent and mechanical skill put in action for their amusement and information (Babbage 1963: 216).

Following Adam Smith’s analysis in The Wealth of Nations, Babbage thought that the process of division of mental labor would serve for the eventual goal of transferring the functions of the human cognitive labor to the operations of a machine. Thus, for Poe, chess-playing automaton, with its allusions to such a “manufactory of information” processing, must have posed an uncanny puzzle also for its implications about the exchange value of his previously irreplaceable intellectual labor. In one of his later speculative narratives, Poe depicted von Kempelen as an alchemist who transforms lead into gold, which results in an enormous reduction of the value of gold and an inflation in the price of lead in international markets.¹³ This narrative could be read as an allusion to the expected decline in the exchange value of the intellectual labor as an unavoidable outcome of the mechanization of reason through a division of cognitive labor.

The imminent threat for the privileged labor position of the enlightened subject as a result of the division of cognitive labor was already under way in the 19th century. Industrial capitalism’s premise was that any of the roles in the socioeconomic chessboard could be played by anyone when these functions are regulated into smaller units through industrial organization.

This social program as an instance of the expansion of the Cartesian mechanistic universe is mainly guided by the flow of human thinking into computational organizations and apparatuses, including the state. In this volume, Jonathan Beller’s essay emphasizes this point through a close reading of Villem Flusser’s works. Particularly, the scientific discourse of the Enlightenment led the trend toward the encoding of human thought into numerical representations. My historical analysis considers the Chess Playing Automaton as one of the behavioral prototypes of this trend inhabiting the associated tensions of the emergent apparatuses that later realize these initially imaginary encodings into commodified social and cultural programs such as cameras or computers. In fact, Flusser finds similarities between photographers’ attempts to find possibilities within the program of the camera and the chess players’ pursuit of finding new possibilities in the program of chess (Flusser 2000: 27–30). In both instances, humans and apparatuses merge into a unity, which explains how human functionaries of apparatuses both control them and are controlled by them. These apparatuses, as a codified set of social relations, are integrated into other socioeconomic apparatuses ranging from industrial production regimes to disciplinary apparatuses of the state. By emphasizing this particular definition of apparatus, Jonathan Beller guides us through a potential and common risk of ignoring the socioeconomic ground of their emergence.

There is an immense similarity between Flusser’s use of apparatus and that of Michel Foucault’s, which denotes strategic constellations of tangible and intangible tools, institutions, and discourses that are inscribed into politics of knowledge and power. Apparatus/Dispositif, according to Foucault, is “a set of strategies of the relations of supporting, and supported by certain types of knowledge.”¹⁴ Foucault uses the term apparatus in order to move beyond discourses to include material, behavioral, and institutional elements for describing formations of structures of knowledge. The term frequently appears in relation to his studies on governmentality from the mid-1970s.

Foucault uses the prison as an example of an apparatus by emphasizing its optic attributes that are configured based on Bentham’s Panopticon architecture. The prison in this view is simultaneously a technical medium designed for seeing without being seen and consequently a tool for subjectification by internalization of surveillance. As Agamben succinctly describes, the apparatus “is first of all a machine that produces subjectifications, and only as such it is also a machine of governance” (Agamben 2009: 20). As a result, the term apparatus provides a very useful vantage point for studying mechanization of mind with an integrated focus on its technical, industrial mediations such as the division of cognitive labor as a disciplinary formation and its constitutive socioeconomic conditions.

Consequently, the analysis of the evolution of the industrial cognitive labor apparatus from its imaginary Enlightenment conceptual prototype to large-scale computable social systems in postindustrial capitalism needs to take into account the neoliberal grounds of its emergence.

The transfer of the functionary role of the cognitive labor apparatus from the privileged labor of the Enlightened subject to unqualified crowds of the neoliberal cognitive capitalism is primarily enabled by the digital networks of the 21st century. This transfer further extends the efect of the foreclosure of the semiosis for the cognitive worker by microdivision of cognitive tasks and its distribution across cultural, temporal, and geographical zones.