Sungyong Ahn (bio)
Abstract
The Internet of Things, as the object-oriented reconstruction of the traditional internet, is characterized by its smart objects freely inter-operating without being necessarily under human control. Re-building the internet’s information economy from the data captured by and communicated through these autonomous objects, the IoT operationalizes a sort of flat ontology, which recent realist philosophers suggest as a means to speculate about the world-making activities of nonhumans not necessarily correlated to human subjects. This paper examines the coincidence of recent interest in these nonhuman world-making processes drawn by two traditionally distinctive but now converging fields: computer engineering and philosophical ontology.
The wide distribution of microsensors, processors, and actuators into our environments during the last decade has changed the information economy of the Internet. According to Kevin Ashton, who coined the term Internet of Things (IoT), before the intervention of these small machines, the Internet was “almost wholly dependent on” the data “first captured and created by human beings—by typing, pressing a record button, taking a digital picture or scanning a bar code.” Replacing humans’ “limited time, attention and accuracy,” which are “not very good at capturing data about things in the real world,” the IoT has been developed as a platform for these things to expand their online presence by overcoming “the limitation of human-entered data” with thing-generated data, communicated in frequencies inaudible to humans (Ashton). Through their further miniaturization and attachment to various natural and technical objects, thing-generated data becomes extractable not only from smart appliances such as refrigerators and smartphones responding to users and environments, but from the territorial/migratory behaviors of animals (Gabrys) and physiological patterns of human organs (Parisi). Just as the digital remediates the incompatibility of analog signals through its binary codes (Bolter and Grusin), the attachment of these smart entities relocates objects from different contexts to the same communicational platform.
At the same time, actor-network theory (ANT) has been introduced in media studies as a critical tool to rethink the conventional boundaries of subject/object, human/nonhuman, cultural/natural, social/technological categories. Actor-network theory has taught us that these categories are not higher orders or contexts that define the legitimate places of things in hierarchies, and showed us how the categories can be resolved back into each thing’s way of influencing others or their mutual engagements. The IoT’s “new sensor/processor/actuator affiliations” (Crandall 83-4) expose hidden actor-networks of objects in our life world, once black-boxed by the habitual contexts of our uses of them as the only definitive typology of their use values. As these objects are now enrolled in a non-hierarchal communication structure of the IoT, the contexts of their human uses are also “unboxed” and their usefulness is re-measured in a digital network, not so much for their contribution to our self-imposed goals, but for the network’s prediction of human purposes.
Marx thought of the use value of commodities as realizable only through their consumption for human needs at “a terminal point” of exchange (Grundrisse 89), such as one’s non-smart home. But in the IoT and its domestic application called smart home, value is conversely concretized by the exchange of thing-generated signals between the smart objects, whose smartness is often advertised as the ability to detect the urgent needs of users even before the users recognize their own needs. John Law says that a black box, which refers to a “complex piece of equipment with contents that are mysterious to the user” (Merriam-Webster), can be reopened only by the appearance of “a stronger adversary, one better able to associate elements” (111). According to this “principle of symmetry,” the IoT would also unbox the previous contexts of the human uses of nonhuman beings, or their monopoly of the right to define the functionality of objects, since the IoT is more capable than humans of associating smart objects together into networks that address human needs. The human consumption of commodities “not only as a terminal point but also as an end-in-itself” was for Marx something easily put aside as “outside economics except in so far as it [what they reproduce namely living labor] reacts in turn upon the point of departure and initiates the whole process anew” in the labor market (Grundrisse 89). This reductionist interpretation of use value based entirely on human “needs as biologically given and the natural” (Dant 501) has been denaturalized by cultural critics such as Baudrillard, whose unboxing of human needs and desires out of “pure, natural, asocial” cocoons has relocated the concept of use value to “a system of relations of difference with other objects” (504). While this revisionist view of use value as “a fetishized social relation just as much as exchange-value” (504) still defines the social exclusively as human construction,1 the IoT—as one of the most advanced commodity forms today—pushes its users to agree with its “terms of use,” which suggest why humans should delegate their right to use objects for their needs to smart objects better at activating themselves in the most customized way to human needs. If outside economics in the Marxian sense has been preserved in domestic space for our inalienable right as tool users, this delegation of human right reopens and reconnects these spaces, renamed “smart homes,” to the economy of digital signals. Humans are the only smart beings whose access to this hidden economy is denied; other smart objects freely exchange queries and answers about their not-smart-enough human hosts.
This actor-network description of the IoT and its reversed user-object relation lead us to a “structure of ontological systems” characterized by the radical liquidation of any hierarchies among things: a world-view that recent realist philosophers, or speculative realists such as Quentin Meillassoux, Graham Harman, and Levi Bryant, call flat ontology.2 This paper instead examines this ontology of objects as something resonant with the recent media industry’s attempt to expand its domains even to the speculative realities of autonomous objects. Despite these philosophers’ inattentiveness to software businesses, the use of the term ontology in computer science as the protocol for machine-to-machine communication3 indicates a sort of commercial necessity for objects to be defined not by human use and access, but through their mutual nonhuman uses and inter-operations. For these philosophers, the autonomy of objects is required for “absolute truth” and reality to be redeemed from their subjective construction in anthropocentric “correlationism” (Meillassoux 5). But for the IoT, the autonomy of objects is required for problems inaccessible to humans to be managed instead by their environment-sensitive operations. The Internet of Things in this sense provides a starting point for a critical inquiry into the question Galloway once posed about “a coincidence between the structure of ontological systems and the structure of the most highly evolved technologies of post-Fordist capitalism” (347).
To re-contextualize this coincidence, this paper focuses on how the architectures of algorithmic systems have changed over the past few decades as programmers and users have delegated more control over a system’s operational environments to its algorithmic objects, which are better able to associate themselves into a more optimal collective state to respond to their environment. It then discusses two cases of algorithmic systems that concern this change: Herman Melville’s Bartleby, the Scrivener: A Story of Wall-street and the MIT Sensible City Lab’s Affective Intelligent Driving Agent. The justification for this unorthodox comparison of a literary work and a media application can be found in the story’s problematic character, Bartleby. In Marxist criticism of the last century, Bartleby has been understood as the “perfect exemplum” of dehumanized workers under industrial capitalism, whose existence as living labor has no other choice for realization than to participate in commodity exchange. The story restages this through its algorithmic distribution of “speculative-conditional” statements or the “logic of the ‘if…then’ statement” to define his possible uses in certain conditions of an office (Reed 258). Bartleby’s famous response, “I would prefer not to,” has been interpreted in this context as a gesture “to get out of circulation entirely” to the “space outside or beyond circulation,” never achievable “except, of course, through death” (266). However, what this reading of his gesture, delayed by one-and-half centuries, focuses not on his suicidal exit to the “humanity” outside commodity exchange, but on his sneaking into the edges of an employer’s algorithmic human resource distribution. Put differently, Bartleby’s withdrawal to the peripheries of commodity exchange is interpreted “in the era of computerized capitalism” as a gesture to nonhumanize himself as an office object not ontologically superior to other office supplies with which he persists in creating a secret network of nonhumans unseen by the employer (Galloway, “The Poverty” 362). Redefined as one of these objects whose inter-operations retrieve the office from the human employer’s exclusive use of nonhumans, Bartleby reminds us of the objects that prevail in recent smart offices. As I will discuss in the following section, these objects are the building blocks of today’s algorithmic culture, which construct both ontologies for those nonhumans and the most customized interface for humans under digital capitalism.
1. From Determination to Agencies
As digital infrastructures become increasingly networked, media studies’ focus on their ability to re-organize media environments has also shifted. Affordance in media studies was once descriptive of media’s function to program “the possibilities in the world for how an agent (a person, animal, or machine) can interact with something” (Norman 18), but it seems more important now to study the way undetermined actions afforded to such agents can update and contribute to the functionality of a network. For instance, in 1986, Friedrich Kittler anticipated that the IF-THEN commands in computer languages would substitute for the symbolic order of human discourses as these “conditional jump instructions” would translate one’s free will into a cybernetic servo-mechanism. For Kittler, the IF-THEN command represents the computational logic of the early cybernetics, which analyzed human behaviors, including language, as “cruise missile”-like variables whose linear trajectories are conditioned by simple feedback loops executable in a linear manner (258). In contrast, what Katherine Hayles calls “a cognitive assemblage” describes how today’s technical infrastructures consist of many autonomous “technical cognizers” controlling the objects that behave like “highly mobile and flexible insurgents and ‘terrorists’” (132).4 Distributed in a swarm-like state, the modularity of these interoperable agents is designed to form an assemblage flexible enough not only to adapt to the changeable environments—like the US military “drone swarms” targeting actual human guerillas—but also to cultivate the things enmeshed by its environmental sensors into the nodes of a potential network. In Frans van der Helm’s media performance MeMachine, for instance, a human body in “a high-tech data suit outfitted with sensors” such as electrocardiography (ECG), electromyography (EMG), and electroencephalography (EEG) is transformed into an object as a source of manifold vital signals that organize these technical cognizers into a network (ARlab; Hayles 129-30). This change in the character of media affordance over recent decades, also coincident with the relocation of humans in digital infrastructures from outside-facing (facing the user) to inside-facing (as hosts for machine-readable signals), has been spurred not simply by the users and objects becoming too elusive to be caught in IF-THEN commands. It rather reflects the media industry’s need to disturb the traditional boundaries between human-subject and nonhuman-object in order to dispatch their machinic cognizers into a larger number of still-unexcavated human problems, and to market their possible solutions through the interoperations of these cognizers. As much as the ubiquitous dangers of guerilla-like intelligences have presented problems to which a military network can respond, smart clothes have conversely transformed wearers’ bodies into things full of ubiquitous problems manageable only through the network computing of microprocessors under their fabrics (Andrejevic). This twofold goal in engineering, namely to redefine problems to justify algorithmic solutions, has transformed human bodies and behaviors into research objects in the same communicational layer with smart objects and appliances. A peculiar commodity-form called IoT generalizes this engineering scenario even in our everyday practices in order to maximize its use value proportional to the number of ubiquitous risks properly manageable only by the networks of nonhuman cognizers.
The algorithm has been marketed as the commodification of efficient and automatized circuit-change technologies that can be applied to any goal-oriented processes from industrial production to domestic reproduction. Modeled as cybernetic servomechanisms, both human bodies and nonhuman objects were previously thought to be functional units that could be distributed most optimally by the discursive protocols or hard-wired circuits of IF-THEN logic, which controlled their sites of consumption and employment, such as a workplace for bodies to be exploited as the hosts of living labor and a house for objects to be used for reproducing labor power. As part of the IoT’s sensor/processor/actuator relations, on the other hand, they can now be placed in the same digital network, which affords their autonomous operations in swarm-like states rather than assigning them to predefined positions for the programmed goals. This design decision to give higher degrees of freedom to objects may entail inefficiencies in the case of simple goal-oriented processes, which were the most important tasks for the IF-THEN based systems, but its long-term advantage, the versatility of a network, is enough to compensate for these problems. In a typical IoT system such as a smart home, software applications newly added to the system usually reach their full functionality only after certain environmental parameters are detected as changing relative to the interoperation of smart objects. This necessary mapping period conversely promises more possible uses of the network in the long run insofar as more parameters are still assumed to hide in the environment, waiting to be detected by different combinations of smart objects for the applications marketed and purchased in the near future. From sequential computing to ubiquitous computing, the method of realizing the use value of an algorithmic system or of operationalizing the meaning of its efficiency has changed from hard-wiring to autonomous networking. This has also been paralleled by the changing understanding of the problems assumed to be embedded in the operational environments of algorithmic systems, from something re-constructible as a cruise-missile-like object in the linear reasoning of IF-THEN sequences to another that can be concretized only through its simultaneous and nonlinear interactions with distributed others.
In computational environments, this change can be described in terms of the shift from the correlationist modeling of early cybernetics to ubiquitous computing’s pan-correlationist modeling, which also distinguishes Hayles’s emphasis on technical agency from Kittler’s technological determinism. Galloway uses the term pan-correlationism to describe how Graham Harman’s object-oriented ontology (OOO) “democratizes” the concept of relation from its monopolized uses in the human construction of reality “by disseminating it to all entities” (“The Universe”). For Harman, a disciple of Bruno Latour, object-orientation is a philosophical method that restores the speculative autonomy of objects in the world represented by actor-network theory, in which the only definitive evidence for the presence of actors is their mobilizing or being mobilized by each other. For this redemptive mission, Harman instead takes as definitive for its becoming an object each entity’s withdrawal into its own core, which “contains unknown realities never touched by any or all of its relations.” By doing so he achieves two goals. First, he liberates objects from any correlationist others, both humans and nonhumans, who attempt to monopolize all the relations between objects for constructing and expanding their subjective reality or networks of technoscience, because “relations do not exhaust a thing” insofar as it always preserves hidden realities to withdraw further into. At the same time, insofar as relations conversely “rely on” the traces of the thing’s withdrawals, there always remain more relations to be further extracted between the objects that constantly withdraw from each other (Prince 132). Galloway chooses the term pan-correlationism to expose how vulnerable this endeavor of OOO (to cut all relations away from the speculative inner realities of objects) is to the ideological reprogramming of capitalist relationism. In his reframing, Harman’s assumption of the inexhaustible inside preserved for the objects’ further withdrawals ironically turns out to be what guarantees the inexhaustible correlations ever extractable from the exteriors of the objects as a result of their constant withdrawals from one another. In object-oriented ontology, there always remains “the sensual skin of exchange value” to be excavated from between any interacting objects (Galloway, “A Response”).
As “the structure of the most highly evolved technologies of post-Fordist capitalism” (Galloway, “The Poverty” 347), ubiquitous computing can be understood as what operationalizes this capitalist repurposing of speculative reality into the reservoir for ubiquitous correlations. In computer science, the object-orientation is a “computational logic” (Kowalski) that defines each algorithmic agent by its own “beliefs (what the agent knows), desires (what the agent wants) and intentions (what the agent is doing) at its core” (Jennings 288). Re-operationalized through ubiquitous computing, which affords more autonomous and unexpected encounters between these objects, the logic of object-orientation in turn redefines its shared environments as full of hidden data that can be extracted through any objects under interaction, transmitted as inputs to any others, and thus never fully bound by any attempt of linear modeling from a single object, but only concretizable through the constant relation-making between distributed objects. In this sense, the term pan-correlationism suggests how resonant the anti-correlationism of the speculative realists is, in fact, with the recent use of the term “correlation” by tech-savvy gurus such as Chris Anderson in Wired magazine. For Anderson, correlation is what manifests the end of “causation” as the human means for “crude approximations of the truth.” Correlation for him is anything that can be data-mined between any interactive objects except subject-object relations. If the recent software industry believes in the inexhaustible and ubiquitous problems always preserved for further excavation and commodification, the pan-correlationist modeling of reality promises this ever-exploitable future of the IoT.
For the formal architecture to accommodate this coincidence between the structures of ontological systems and the most advanced commodity form of today, we may need to first examine a programming paradigm called object-oriented programming. As Alan Kay, the architect of the early OOP language Smalltalk, writes, “Its semantics are a bit like having thousands and thousands of computers all hooked together by a very fast network” (70). OOP is characterized by its “behavioral building blocks,” objects that “have much in common with the monads of Leibniz” (70), as each object enfolds the definitions of its own constituents, data structures, and possible interoperations with others. Put differently, an object envelops its own “tiny ontology,” stating its selective exposures and responses to environments (Bogost 21). To build an algorithmic system for object-oriented programmers is thus to design the recursive inter-operations of these objects to replace cumbersome IF-THEN sequences in the obsolete procedural programming. In the source code of a program, the objects are distributed as autonomous behavioral units but held in a metastable state or in the “initial absence of interactive communication” until after its execution or compiling for “a subsequent communication between orders of magnitude and stabilization” (Simondon 304). The compiling of a source code begins as its exposure to an environment, namely, a set of user inputs or a database, triggers the response of the object assigned in the beginning, whose behavioral outputs in turn trigger further responses from the others until the intended set of states is singled out from, created in, or removed from their shared environment. Despite their seemingly autonomous becoming as a collective, the inter-operations of the objects during the compiling is designed as a sort of pre-established harmony, as a human programmer puts them in designated locations in a source code to make their environmental exposure happen in a predetermined order.
On the other hand, the recent achievements in the miniaturization of digital sensors/processors/actuators to a size attachable to any scale enables these purely algorithmic objects to be transplanted into natural/technical entities in the real world. In OOP, each object undergoes a process of individuation as its interactions with others gradually adjust the parameters in its data structure to niche values. The physical objects in ubiquitous computing undergo a similar process as they are enrolled in its sensor/processor/actuator affiliations. However, their new niches in the networks are not pre-established by human assumptions of harmony, but concretized through corporeal interactions with other sensor-augmented objects in environments. Just as a drone swarm constantly updates its flying formation using the aerodynamic data extracted from each drone’s interactions with others, the operational environment of an algorithmic system is no longer simply a metaphor for “human-entered data” but also for the ubiquity of data that can be extracted from any physically distributed interacting objects. These objects are virtually all re-locatable to a digital network from their own natural and technical contexts. And insofar as each of these contexts is what today’s commodification of ubiquitous computing advertises as the problem that can be more efficiently managed by its unboxing and tracing the objects’ relation-making in an actor-network-like manner, it is inevitable that the objects once stabilized in their own context will resume the individuation to find their new functional niches in the algorithmic cultures. For instance, one’s heart, muscles, and brain, already stabilized in a psychosomatic context, are now relocated to a digital network under a “smart cloth” outfitted with ECG, EMG, and EEG. Their resumed individuations to the digital niches are not based on the pre-established harmony between bodies and minds under one’s conscious or reflexive control. Rather, they can be harmonized further with other digital objects capable of sneaking under the cloth as new members of the affiliation, such as the Apple Watch or Fitbit. These gadgets, “better able to associate” the organic and machinic elements into more optimal states for different situations, such as workin out, sleeping, working, or shopping, begin to teach us what to do, much as the fitness app in your Apple Watch commands you to slow down or speed up. Human organs are no longer particular organs employed in a servomechanism but constantly re-individuate themselves for their temporal niches and uses within the nonhuman networks with which they are newly affiliated. From this changed use of human bodies, Bartleby’s gesture to disconnect himself from any capitalist uses of human beings by saying “I would prefer not to” do anything assigned by the human employer earns a new meaning. His gesture can be reinterpreted as a prophecy of recent smart objects and their withdrawals into the peripheries of human control.
2. Bartleby, the Scrivener
Melville’s narrator devotes the first quarter of this story of a Wall Street law office in the 1850s to describing the functional relations between his employees, which also define the end state for the rest of the story to restore after the disturbance caused by Bartleby. Nonhuman nicknames, Turkey, Nippers, and Ginger Nut, are “mutually conferred upon each other” because they are “deemed expressive of their respective persons or characters” (par. 6). It is Ginger Nut’s job to deliver “ginger-cake” to Turkey and Nippers, whose performances of “copying law papers” are complementary to each other because the former is reliable only in the morning whereas the latter works well only in the afternoon. For the lawyer confident in reorganizing their different responses, to make the office operational for his own goal is to distribute these workers along a procedural sequence: “it being morning, Turkey’s answer is couched in polite and tranquil terms, but Nippers replies in ill-tempered ones … to repeat a previous sentence, Nippers’s ugly mood was on duty, and Turkey’s off” (par. 45); to repeat in an object-oriented pseudo-code, IF it is morning THEN call Turkey or ELSE call Nipper. The lawyer’s “doctrine of assumptions” is applied everywhere in the office and enables him to predict how the actors will respond in certain conditions and to mobilize their conditional responses “to enlist the smallest suffrage in [his] behalf” (par. 155, 46). Following a pre-established harmony assumed by the lawyer, Ginger Nut claims his contribution by circulating ginger-cake, which in turn demonstrates its functionality through “probable effects upon the human constitution” of Turkey and Nippers (par. 52), whose functions as scriveners alternate in the morning and afternoon. However, the lawyer’s confidence in mobilizing these switching circuits faces a crisis in Bartleby, a new scrivener. As “more a man of preferences than assumptions,” his becoming a meticulous actor in the office is defined at first by his highly selective response of “prefer[ing] not to do” any tasks other than transcribing law papers “at the usual rate of four cent a folio” (par. 83). In the middle of the story, Bartleby begins to narrow this response further to the extent of preferring not to answer any queries from the lawyer and finally ceasing to produce any readable texts. At this point, the lawyer (as a system builder) has the following conversation with Bartleby:
“Now what sort of business would you like to engage in? Would you like to re-engage in copying for some one?”
“No; I would prefer not to make any change.”
“Would you like a clerkship in a dry-goods store?”
“[…] I would prefer not to take a clerkship,” he rejoined, as if to settle that little item at once.
“How would a bar-tender’s business suit you? There is no trying of the eyesight in that.” “I would not like it at all; though, as I said before, I am not particular.”
[…]
“Well then, would you like to travel through the country collecting bills for the merchants? That would improve your health.”
“No, I would prefer to be doing something else.”
“How then would going as a companion to Europe, to entertain some young gentleman with your conversation,—how would that suit you?”
“Not at all. It does not strike me that there is anything definite about that. I like to be stationary. But I am not particular.” (par. 197-209)
ANT’s principle of symmetry states that “all the elements that go to make up a heterogeneous network, whether these elements are devices, natural forces, or social groups,” can make themselves present as actors only “by influencing the structure of the network in a noticeable and individual way” (Law 124-6). Conversely, the same principle implies that any actors withdrawing from their current network should enroll in another that is “better able to associate elements” (111), unless they prefer not to return any noticeable responses and thus not to be present any longer to others. Bartleby’s preference not to do something else expresses his fatigue over being this kind of element unable to be present at all if not assigned to a new functional niche in the office or in an outside labor market according to the lawyer’s “doctrine of assumptions.”5 Bartleby’s strategy to respond to the queries by saying that he prefers not to suggests the minimum that an actor should do to stay in a current state. As a dehumanized object stuck within the algorithmic human resource management, Bartleby’s gesture to postpone his assignment to particular uses thus unboxes the apparently seamless commodity exchange in the labor market. The lawyer’s subsequent and never-ending IF-THEN questions, “would you like to re-engage in …? Well then, would you …? How then would …?,” reveal the maximum that the employer needs to do to black-box again the formal symmetry of the capitalist uses of human beings.
The lawyer’s efforts to find a new niche for Bartleby, however, always turn out to be undertaken too late, after Bartleby has already declared his preference not to do that work. And when Bartleby is proved not to be handled by the servo-mechanical “logic of the ‘if…then’ statement,” the lawyer discovers a secret network of nonhumans in which Bartleby’s withdrawal finds the smallest niche for his presence: a “bachelor’s hall” that “Bartleby has been making” with things hidden at the peripheries of the lawyer’s attention, such as “a blanket” under his desk rolled away, “blacking box and brush” under the empty grate, “a tin basin, with soap and a ragged towel” on a chair, “a few crumbs of ginger-nuts and a morsel of cheese” in a newspaper (par. 88). Shortly after Bartleby declares his presence in the office despite his refusal to accept any of the new positions the lawyer recommends, these objects, once supposedly governed under the lawyer’s “doctrine of assumptions,” appear to converge instead upon an alternative network. In this flat network, each thing’s presence is concretized not through the lawyer’s monopoly of (non)human resources, but through their mutual engagement at the peripheries of capitalist resource distribution. Contrary to the traditional interpretations of Bartleby’s gesture as a suicidal disconnect from any social ties, what he really achieves through his withdrawal is not the redemption of humanity “through death” (Reed 266) but the retrieval of social ties among nonhumans from capital’s dehumanizing correlationism, which defines every object, including human labor, as exchange or use value to preserve or increase capital. The withdrawal of objects into their inner realities “never touched by any or all of [their] relations” is enough for these objects to be present without necessarily being engaged in the businesses of others (Harman, Prince 132); at the same time, for Harman, this withdrawal also suffices to enable the ubiquitous distances between these objects to be filled with finer-grained relations as “the joints and glue that hold the universe together” (Guerilla Metaphysics 20). Bartleby’s disappearance into the peripheries of commodity exchange likewise finds a hidden society of nonhumans in which his presence in the world stands on an equal footing with everything else. Through the lawyer’s lost confidence in his assumption as to the possible uses of Bartleby, Melville’s story dramatizes a conflict between two ontologies: the correlationist modeling of reality through a human employer in the center as the avatar of old capitalism, and the pan-correlationist through the distributed nonhumans and their mutual engagements. However, there are also things his story fails to anticipate, such as how vulnerable these nonhumans are to the finer-grained resource management algorithms under advanced capitalism, and how the new avatar of capitalism will, 150 years later, appear in the form of these distributed nonhumans called smart objects.
3. Ontic Principle of Ubiquitous Computing
Mark Weiser defines ubiquitous computing as the withdrawl of microprocessors from the center of users’ attention towards the peripheries, where they are more correlated with other microprocessor-augmented things such as smart appliances (Weiser and Brown). Once they stop competing for human attentions to be chosen as indispensable units dragged to a narrow Graphical User Interface (GUI), the devices become more functional to each other in their exchange of the data secretly extracted from humans. Information technologies before ubiquitous computing such as “pagers, cellphones, newservices, the World-Wide-Web, email, TV, and radio” were designed to “bombard us frenetically” to draw our attention and claim their increasing presence in a human-centered network (79). Like Turkey claiming his functionality to the lawyer even in the afternoon when he malfunctions by asking “if his services in the morning were useful, how indispensible, then, in the afternoon?” (par. 6), these machines once appealed for their usefulness to human users who monopolized what ANT calls the “Obligatory Passage Point (OPP)” of the network, through which all actors must pass to be assigned and specified as the actors appropriate for the goal of a system (Callon 205-6). On the other hand, objects in the Ubiquitous Computing (UC) era prefer not to respond to queries from users. Like Bartleby, they stay “calm” in the periphery of attention and maintain the slightest presence in “a confederacy of ‘smart’ objects,” which “whisper information to one another in inaudible frequencies,” not in order to reoccupy the center of attention at the most timely moment for our needs (as Weiser’s original design intended), but to “conspire to sell us products” in a timely manner (Andrejevic 113-4).
In their withdrawal to the peripheries where they awaken each other to avoid awakening users, devices spend most of their time performing the minimum for their enrollment as sensors, namely scrivening “unmodulated digital data” from their operational environments. As actuators influencing others, the devices also undergo a constant and creative process of individuation to find the functional niches within their temporal inter-operations (Clemens and Nash). Unlike the lawyer’s algorithmic IF-THEN instructions, ubiquitous computing as the global intelligence of these distributed devices is capable of and patient with performing this task of never-ending resource distribution to transduce all non-particular objects in metastable states into each individual in situ. It does this by calculating the optimal way to weave the devices’ autonomous and oft-conflicting operations and goals into a collective that can be mobilized for a problem at the system level. In this sense, what become ubiquitous in the UC era are not only the symmetrical edges of the network for the smart devices’ horizontal communications but their asymmetries to a collective intelligence that seems to replace successfully the lawyer in Melville’s story and intervene in the stabilization of conflicting devices into the reciprocal and modular functions of the network. As Galloway writes, “no arbiter impedes” these symmetrical individuations of smart objects into the actors influencing each other along the edges of a network, but for their autonomous responses to each other to be gradually realigned in the most efficient and harmonious way to reach collective goals, a sort of “ultimate mystical medium” is still assumed to operate as an invisible hand (“A Response”). In the Personal Computing (PC) era, human users performed this arbiter by monopolizing the obligatory passage point represented by intuitive user interfaces that enabled them to design the harmonious interoperations of algorithmic agents for their conscious goals. In the UC era, the mystical arbiters are rather omnipresent in the form of microprocessors that may attach anywhere, more ethereal in infiltration into every edge of the networks; the operations of these smart arbiters are as immanent as marketplaces that have also infiltrated into every corner of our lives, augmented by the so-called smart applications. Harmony is no longer pre-established by the assumptions of human designers but, like the flying formation of a drone swarm, must be constantly gathered and updated from lots of minute discrepancies between each object’s expectation of its environment and its actual operation within the data extracted from others.
Whitehead’s philosophy of actual entities provides another ontological model of “the universe of things” in the UC era that, as Steven Shaviro notes, supplements the “countless tiny vacuums” between objects mutually withdrawing in Harman’s object-oriented ontology with “a finely articulated plenum” of data left by each object’s becoming (Shaviro 39). For Galloway, in my interpretation, the coincidence between OOO and OOP implies the former’s vulnerability to ideological reprogramming; the “unknown realities” preserved and inexhaustible inside each object ironically promise the constant extraction of correlations from any objects under interactions, whose exteriors, or “sensual skin[s] of exchange value” (Galloway, “A Response”), are thus able to create ever-regenerative inter-objective realities. As a critical approach to demystify these worldly relations supposedly waiting to be excavated for ubicomp solutions, we can examine how Whitehead brings the problematic realities that OOO hides inside each object back to the platform for inter-objective communications. Whereas an object for Harman preserves its speculative presence through constant withdrawals, an actual entity for Whitehead lives only for its process of becoming called “concrescence.” Through this process, an entity prehends the universe as a “multifold datum” left by the already finished becoming of all others until the process is completed with the satisfaction of its “subjective aim” and turned into just another datum for the genesis of others (Whitehead 19, 185). The resources for creations are not hidden inside but scattered all over the world, revealed to be a large data set accumulated from the finished processes of concrescences and given for further data-mining by new actors to come. In this respect, the speculative presence of actual entities in Whitehead provides a philosophical analogy for the algorithmic objects in a source code, which also live only during their exposure to shared environments for processing input data and are then left just as what they processed, namely the changed state of these environments for others to process further. However, in that each actual entity’s concrescence is not determined by any others but performed according to its own assumption on the “harmony” between its “subjective forms” and the objective data it feels (27), the source code as the nexus of objects in this analogy should not be based in the hardwired electromagnetic circuits of personal computers. Rather, the technical incarnation of Whiteheadian actual entities is found in the smart objects in the Internet of Things as they constantly re-individuate themselves within their data-intensive environments without predefined orders. Besides this structural similarity, Whitehead’s “secularization of the concept of God” (207) as no other than one of these entities provides another rationale for the appropriateness of the analogy. Contrary to those whose concrescences are temporal and short-lived, the Whiteheadian God is characterized by its never-completed concrescence. This God’s subjective aim is “the ultimate unity” between the entire multiplicity of actual entities it senses and its conceptual prehension of their ideal harmony “in such a perfect system” (346), and this is inevitably an ever-delayed goal insofar as God cannot determine the courses of other entities’ becoming but only induces them to adjust their subjective aims. Taking the position of this global but not omnipotent agent in the analogy, the aim of an ubicomp system—the algorithmic calculation of the optimal way for the smart objects to inter-operate for systemic goals—is also a never-completed process that must be constantly updated from each object’s actual operation without any pre-given harmony.
This secularized understanding of God is decisive in order to preserve the symmetry that Harman’s OOO sees in a flat ontology. Shaviro emphasizes that “all actual entities in the universe stand on the same ontological footing,” and even God for Whitehead has “no special ontological privileges” over the most trivial entities “in spite of” the asymmetrical “gradations of importance, and diversities of function,” among entities (Shaviro 29). However, in the emerging universe of things called IoT, these gradations of importance and functionality relocated and persisting in a flat ontology are in fact what make Whiteheadian philosophy a better analogy for the recent smart environments than Harman’s, and also make media studies’ recent interest in Whitehead (Gabrys; Parisi; Hansen) more coincident with “the structure of the most highly evolved technologies” of today. The particular entity standing at the apex of these gradations was once called God, but now reappears in the form of ubiquitous computing, and its never-ending concresence as a global intelligence intervenes in all the other entities’ temporal concresence as the nodes of its network. Rather than assigning each entity one-by-one to a specific place already prepared—what Melville’s lawyer attempted but failed—ubiquitous computing encourages the entities to find their own bachelor’s hall within the multifold data transmitted from the actual world by letting them interact according to their preferred responses to environments and in turn enabling their data structures to be coupled optimally to each niche in the ubiquitous thing-generated data. It is not in spite but because of these asymmetrical interventions of omnipresent microprocessors that all other less important but still functional entities are relocated and “rethingified” upon a flat and symmetrical platform of smart objects (Gabrys 192).
The “ontic principle” of speculative realism often promotes itself as a democratic principle for nonhumans in opposition to “the vertical ontologies of ontotheology or a humanism” that “trace back and relate all beings to either God, humans, language, culture or any of the other princes.” It suggests “a flat ontology, one made exclusively of unique, singular individuals, differing in spatio-temporal scale but not ontological status” (Bryant, “The Ontic Principle” 268-9). The Internet of Things as the commodification of ubiquitous computing seems to operationalize this principle by liberating digital objects from their previous obligation to pass through the mediations of human princes. However, its blatant attempts to diversify the problems that can be detected and marketed along the networks of these liberated objects reveal the ideological undertone of the societal metaphors for objects such as Bryant’s “democracy of objects” (The Democracy), applied to today’s media systems without mention of the primary asymmetry of the network. The ubiquity of smart objects and their autonomous operations translate and integrate each singular reality they locally perceive into a larger data set as a shared environment in which all of them are interoperable no matter how different their narrow world views are from one another. These ubiquitous symmetries for the ubiquitous accumulation of sharable data are, however, also asymmetrically engaged in the quasi-theological individuation of a global intelligence. As the following section will exemplify by examining a scenario involving a smart navigation system, Bartleby’s gesture to non-humanize himself as one of many office objects in a flat and invisible network does not mean his or its liberation from capitalist resource management. This instead forms the condition for a global intelligence system to emerge from its asymmetrical interventions in each symmetrical edge of the network.
4. AIDA: A New Bachelor’s Hall
AIDA (Affective Intelligent Driving Agent) is an in-dash navigation system developed by MIT’s SENSEable City Lab, Media Lab, and the Volkswagen’s Electronics Research Lab. Equipped with several projectors that display a 3D map on the dashboard, AIDA visualizes the most efficient route to a destination as a solution to the possible need of a registered driver (see fig. 1).
Fig. 1 AIDI’s dashboard display. MIT Sensable City Lab. “AIDA 2.0.” Youtube, uploaded by MIT SENSEable City Lab, 11 May, 2011, https://www.youtube.com/watch?v=zKLAeq1m2TY
Unlike non-smart systems, “AIDA analyses the driver’s mobility patterns, keeping track of common routes and destinations” to “identify the set of goals the driver would like to achieve” (MIT Sensible City Lab). To provide the driver with the most customized niche not only in a vehicle but in the traffic networks and points of interest (POI) in neighboring areas, AIDA interfaces sensor-generated driver data concerning her implicit needs with the data “pertaining to various aspects of the city including traffic, seasonal information, environmental conditions, commercial offerings, and events” (Lorenzo et al.).
An ordinary object-oriented navigation simply receives the data packets from outside sources, such as GPS and traffic information, in order to remodel its surroundings with algorithmic objects such as a street, intersection, or geo-tagged landmarks, whose functional relations as nodes in a graph individuate the shortest route on the map to solve the problem of the “human-entered” destination. On the other hand, as a prophet-like agent smart enough to direct the driver to where she must go to fulfill her current need, what AIDA should individuate foremost is not the shortest route but the most urgent problem of the driver, which has yet to surface but is lurking in the peripheries of her attention as the ubiquitous symptoms filling the car. AIDA’s ubiquitous computing individuates the problem preemptively and puts it in a navigable form that would be solved progressively as she drives the car along the route to a spot it recommended on the map. As an IoT system counterinsurgent to this guerilla-like problem—namely, a human driver demoted to a host of machine-readable vital signals—AIDA populates not only the interior network of the Audi full of interconnected sensors for facial expressions, voices, galvanic skin response, braking/acceleration pressures, seat position, and steering (see fig. 2), but “a multitude of tags, sensors, locationing devices, telecommunications networks, online social networks, and other pervasive networks … proliferating in cities,” as well as the driver’s social networks.
Fig 2. A network of sensor-augmented things in AIDA 1.0. The AIDA Project (Affective, Intelligent Driving Agent), MIT SENSEable City Lab and Personal Robots Group of Media Lab, 2009, http://senseable.mit.edu/aida/. Accessed 10 Dec. 2019.
Suppose that Bartleby, hired as a test driver, found his new bachelor’s hall in this Audi. After the first week, long enough for the sensor network to be trained to relate the behavioral signals collected from the distributed body parts to his current affective state, AIDA would begin to figure out his “home and work location” and “be able to direct” him to the grocery store that he is likely to prefer. After a month, AIDA could detect his hunger from the signals collected and analyzed through the “historical behavioral collector (HBC)” and “historical route collector (HRC),” and then recommend the restaurant rated highest by Yellow Pages users with similar social networking service (SNS) profiles (Lorenzo et al.). Bartleby may find that he is aware of his hunger only several minutes after the distributed symptoms were identified by AIDA, but may not seriously care about this delay even though it is always long enough to pass the restaurant most customized to his taste. However, after he learns that his too-human consciousness is, as Hayles warns, always behind the non-conscious responses of his body by at least several hundred milliseconds, so-called “missing half-second,” long enough to be hijacked by other non-conscious cues from “the advent of affective capitalism and computational media” (Hayles 191; Hansen 190), even hunger would become a crisis that should be preempted by AIDA and immediately visualized as a red route to the restaurant on the map. He is now responsible for eliminating this route by driving his Audi corner to corner according to AIDA’s instructions (see fig. 3).
Fig 3. AIDA recommending a restaurant. MIT Sensable City Lab. MIT SENSEable City Lab and Personal Robots Group of Media Lab, “AIDA 2.0.” Youtube, uploaded by MIT SENSEable City Lab, 11 May, 2011, https://www.youtube.com/watch?v=zKLAeq1m2TY
What has occurred in this local network of smart sensors after a month of test driving is Bartleby’s individuation as a registered driver. But, on the other side of the interface, where his current physiological and behavioral states are ceaselessly translated into the red route heading somewhere to be resolved, his individuation appears to have been driven entirely by another individuation of AIDA into a prophet-like intelligence, ceaselessly weaving a flat ontology out of many different types of sensor data—such as GPS data, a city’s Points of Interest and their rankings in Yellow Page, lots of geotagged images of the city, the driver’s and his neighbors’ social network profiles, and his historical route and behavior data—by folding them into the pathway he draws (Lury and Day 30).
After these reciprocal individuations, Bartleby on the day of his public demonstration would see something reminiscent of the compulsive questions of the lawyer in Melville’s story haunting the dashboard, tuned up for the maximum functionality of AIDA. On the way to the destination that AIDA would already have predicted from his route histories, Bartleby would encounter many small pop-up windows and tags on the map referring to places for entertainment, social events, and other sensor-augmented commodities, claiming to concretize his unknown desires distributed across his facial expressions, voice, galvanic responses, butt position, accelerating and braking foot pressures (see fig. 4).
Fig 4. AIDA advertising a social event. MIT SENSEable City Lab and Personal Robots Group of Media Lab, “AIDA 2.0.” Youtube, uploaded by MIT SENSEable City Lab, 11 May, 2011, https://www.youtube.com/watch?v=zKLAeq1m2TY
Just like the lawyer in the story, AIDA asks, “would you like to …? Well then, would you …? How then would …?” Contrary to the lawyer who failed to keep enumerating all possible niches for Bartleby due to his too-human managerial skill, AIDA’s recommendation is ever-extendable. Only this global intelligence can access the problem called ubiquity, and it does not ask him to share his preferences out loud. Bartleby may already find himself in his most customized bachelor’s hall, which eliminates any possible disturbances even before they actually occur. Now he needs to accept the “terms of use” for AIDA, but what pushes him to agree with these terms—which describe his new human condition as the host of digitalized vital signals—is the dormancy of problems whose symptoms are too widely distributed to be cognized by any single object except AIDA.
Conclusion
In critical theories after 9/11, these sorts of omnipresent agents and their asymmetrical interventions in the life of local actors have been justified by the ubiquity of problems. The latency of these problems in the peripheries of each actor’s narrow attention has justified the local actors’ commitments to a collective intelligence that preempts problems before they actually occur. Hardt and Negri claim that “the gray zone of war and peace,” in permanent danger of insurgency and terrorism, justifies the “total mobilization of social forces” for the preemptive strike of a military power that is “in asymmetrical conflicts” over unpredictable “guerrilla attacks” (13, 51-2). Massumi also writes that civilian life in this “crisis-prone environment” falls “onto a continuum with war” in which a preemptive power’s intervention should be “as ubiquitously irruptible as the indiscriminate threats it seeks to counter” (27-8). Not necessarily based on the wide spaces reminiscent of battlefields, or necessarily generalizing these military environments to the scales of human bodies to be covered by wearable devices and home/office for smart appliances, the Internet of Things invents a novel strategy for its market penetration from this tension between the insurgency of ubiquitous problems and the counterinsurgency of an intelligent system. In Melville’s fiction, Bartleby’s symmetry-breaking insurgency was never preventable by the lawyer’s linear management programmed in IF-THEN statements. But, as I re-fictionalized through Bartleby in AIDA, this human inaccessibility to the problem called ubiquity is also the justification for the humans’ participation in the IoT as the non-humanized hosts of vital signals. For its becoming as a collective intelligence from the concrescence of these vital signals with other thing-generated data, the ubiquity of lurking problems should be advertised as the reason why it is time for humans to relinquish their right to the uses of objects that they have unfairly held for a long time and why it is time to hand it right over to the IoT, which can use them more preemptively to maintain a space always customized to our needs.
Footnotes
1. Fetishism in this Marxian context has been consistently defined as the mis-imposed value of the object-in-itself, which can be analyzed as the social relations congealed around the object. Arjun Appadurai instead takes fetishism as his methodology for “a corrective to the tendency to excessively sociologize transactions in things” (5). But even in his “methodological fetishism,” the values of objects are subject to a multitude of local contexts of symbolic transactions, despite their irreducibility to the global capitalist economy.
2. For “flat ontology” in their speculative realism see Bryant, “The Ontic Principle”; Bogost, Alien Phenomenology. For Galloway’s criticism of flat ontology as the “structure of ontological systems” in recent software companies, see “The Poverty,” p. 347.
3. In “The Role of Common Ontology,” Gruber suggests ontology as an engineering term for “knowledge-level protocols” between AI systems, each of which is distinguished by its own “symbolic-level” of representation of environments. The role of ontology is not to organize a single globally shared theory of the environment. It rather aims to provide languages for an output of a system to be translated into the input for another to maximize the inter-operability of and communicability between the systems.
4. These objects and object-like users may be modeled best as the actor in the term actor-network, not an “individual atom” hyphenated to a network in a deterministic way, but a “circulating entity” that draws lots of hyphens to “hook up with” each other for both specifying its local agency and organizing global structure (Latour 17-8). For these actors, “a substrata: something upon which something else ‘runs’ or ‘operates’” is no longer a proper metaphor for infrastructures; rather, technological infrastructures are installed as communication protocols for these circulating entities to make their “local practices afforded by a larger-scale technology” into the modular functions “which can then be used in a natural, ready-to-hand fashion” by others (Star and Ruhleder 113-4).
5. It is noteworthy that David Kuebrich relates the lawyer’s “doctrine of assumption” about the niche positions for each actor in his design of the fully operational office to “the larger culture that there is no inherent contradiction between the dedicated pursuit of self-interest, even when it involves the exploitation of others, and devotion to traditional Christian values” (396). According to him, the doctrine “exemplifies the values and attitudes of the Protestant entrepreneur who fused his Christian faith,” such as the faith in the “Starvation and wretchedness … by Heavenly appointment,” with “emerging economic practices in such a way as to legitimate inequality and class privilege” (383, 386).
Works Cited
- Anderson, Chris. “The End of Theory: the Data Deluge Makes the Scientific Method Obsolete.” Wired, 23 Jun. 2008, www.wired.com/2008/06/pb-theory. Accessed 4 May 2019.
- Andrejevic, Mark. “Nothing Comes Between Me and My CPU: Smart Clothes and ‘Ubiquitous’ Computing.” Theory, Culture & Society, vol. 22, no. 3, 2005, pp. 101–119. SAGE, doi:10.1177/0263276405053723.
- Appadurai, Arjun. “Introduction: Commodities and the Politics of Value.” The Social Life of Things: Commodities in Cultural Perspective, edited by Arjun Appadurai, Cambridge UP, 1986, pp. 3–63.
- ARlab. “MeMachine: Bio-technology, Privacy and Transparency.” ARlab, 13 Aug. 2013. arlab.kabk.nl/project/memachine-bio-technology-privacy-and-transparency.
- Ashton, Kevin. “That ‘Internet of Things’ Thing,” RFID Journal, 22 Jun. 2009, www.rfidjournal.com/articles/view?4986. Accessed 4 May 2019.
- Baudrillard, Jean. For a Critique of the Political Economy of the Sign. Translated by Charles Levin, Telos Press, 1981.
- Bogost, Ian. Alien Phenomenology, or, What It’s Like to be a Thing. U of Minnesota P, 2012.
- Bolter, Jay David, and Richard Grusin. Remediation: Understanding New Media. MIT P, 1999.
- Bryant, Levy. The Democracy of Objects. Open Humanities Press, 2012.
- —. “The Ontic Principle: Outline of an Object-Oriented Ontology.” The Speculative Turn: Continental Materialism and Realism, edited by Levy Bryant, et al., re.press, 2011, pp. 261–278.
- Callon, Michel. “Some Elements of Sociology of Translation: Domestication of the Scallops and the Fishermen of the Saint Brieuc Bay.” Power, Action, and Belief: A New Sociology of Knowledge, edited by John Law, Routledge, 1986, pp. 205–6.
- Clemens, Justin, and Adam Nash. “Being and Media: Digital Ontology after the Event of the End of Media.” The Fibreculture Journal, vol. 24, 2015. twentyfour.fibreculturejournal.org/2015/06/03/fcj-173-being-and-media-digital-ontologyafter-the-event-of-the-end-of-media. Accessed 4 May 2019.
- Crandall, Jordan. “The Geospatialization of Calculative Operations: Tracking, Sensing and Megacities.” Theory, Culture & Society, vol. 27, no. 6, 2010, pp. 83–4. SAGE, doi:10.1177/0263276410382027.
- Dant, Tim. “Fetishism and the Social Value of Objects.” The Sociological Review, vol. 44, no. 3, 1996, pp. 495–516. Wiley, doi:10.1111/j.1467–954X.1996.tb00434.x.
- Gabrys, Jennifer. Program Earth: Environmental Sensing Technology and the Making of a Computational Planet. U of Minnesota P, 2016.
- —. “Re-thingifying the Internet of Things.” Sustainable Media: Critical Approaches to Media and Environment, edited by Nicole Starosielski and Janet Walker, Routledge, 2016, pp. 180–195.
- Galloway, Alexander. “The Poverty of Philosophy: Realism and Post-Fordism.” Critical Inquiry, vol. 39, Winter 2013, pp. 347–366. JSTOR, doi:10.1086/668529.
- —. “A Response to Graham Harman’s ‘Marginalia on Radical Thinking.’” An und für sich, 3 June 2012. itself.wordpress.com/2012/06/03/a-response-to-graham-harmans-marginaliaon-radical-thinking. Accessed 4 May 2019.
- —. “The Universe of Things.” 29 Nov. 2014. cultureandcommunication.org/galloway/the-universe-of-things. Accessed 4 May 2019.
- Gruber, Thomas. “The Role of Common Ontology in Achieving Sharable, Reusable Knowledge Bases.” Principles of Knowledge Representation and Reasoning: Proceedings of the Second International Conference, edited by James Allen, et al., Morgan Kaufmann, 1991, pp. 601–602.
- Hansen, Mark. Feed-Forward: On the Future of Twenty-first-century Media. U of Chicago P, 2015.
- Hardt, Michael, and Antonio Negri. Multitude: War and Democracy in the Age of Empire. Penguin, 2004.
- Harman, Graham. Guerilla Metaphysics: Phenomenology and the Carpentry of Things. Open Court, 2005.
- —. Prince of Networks: Bruno Latour and Metaphysics. re.press, 2009.
- Hayles, Katherine. Unthought: The Power of the Cognitive Nonconscious. U of Chicago P, 2017.
- Jennings, Nicholas. “On Agent-based Software Engineering.” Artificial Intelligence, vol. 117, 2000, pp. 277–296. ScienceDirect, doi:10.1016/S0004–3702(99)00107–1.
- Kay, Alan. “The Early History Of Smalltalk.” ACM SIGPLAN Notices, vol. 28, no. 3, 1993, 69–95. ACM Digitial Library, doi:10.1145/155360.155364.
- Kittler, Friedrich. Gramophone, Film, Typewriter. Translated by Geoffrey Winthrop-Young and Michael Wutz, Stanford UP, 1999.
- Kowalski, Bob. “Computational Logic in an Object-Oriented World.” Reasoning, Action and Interaction in AI Theories and Systems. Lecture Notes in Computer Science, edited by Oliviero Stock and Marco Schaerf, Springer, 2006, pp. 59–82.
- Kuebrich, David. “Melville’s Doctrine of Assumptions: The Hidden Ideology of Capitalist Production in ‘Bartleby.’” The New England Quarterly, vol. 69, no. 3, 1996, pp. 381–405. JSTOR, doi:10.2307/366781.
- Latour, Bruno. “On Recalling ANT.” Actor Network Theory and After, edited by John Law, Blackwell Publisher, 1999, pp. 15–25.
- Law, John. “Technology and Heterogeneous Engineering: The Case of Portuguese Expansion.” The Social Construction of Technological Systems, edited by Wiebe Bijker, et al., MIT P, 2012, pp. 105–128.
- Lorenzo, Giusy Di, et al. “An Affective Intelligent Driving Agent: Driver’s Trajectory and Activities Prediction.” Vehicular Technology Conference, IEEE, 20–23 Sep. 2009, Anchorage, AK.
- Lury, Celia, and Sophie Day. “Algorithmic Personalization as a Mode of Individuation.” Theory, Culture & Society, vol. 36, no. 2, 2019, pp.17–37. SAGE, doi:10.1177/0263276418818888.
- Marx, Karl. Grundrisse: Foundation of the Critique of Political Economy. 1857. Translated by Martin Nicolaus, Penguin Books, 1993.
- Massumi, Brian. Ontopower: War, Powers, and the State of Perception. Duke UP, 2015.
- Meillassoux, Quentin. After Finitude: An Essay on the Necessity of Contingency. Translated by Ray Brassier, Continuum, 2008.
- Melville, Herman. Bartleby, the Scrivener: A Story of Wall-street. 1853. www.bartleby.com/129/. Accessed 4 May 2019.
- MIT Sensible City Lab. “MIT Researchers Develop Affective Intelligent Driving Agent.” MIT Sensible City Lab, 29 Oct. 2009. senseable.mit.edu/aida/downloads/AIDA_Press_Release.doc. Accessed 4 May 2019. Norman, Donald. The Psychology of Everyday Things. Basic Books, 1988.
- Parisi, Luciana. “Technoecologies of Sensation.” Deleuze/Guattari & Ecology, edited by Bernd Herzogenrath, Palgrave Macmillan, 2009.
- Reed, Naomi. “The Specter of Wall Street: ‘Bartleby, the Scrivener’ and the Language of Commodities.” American Literature, vol. 76, no. 2, 2004, pp. 247–273. Duke University Press, doi:10.1215/00029831–76-2–247.
- Shaviro, Steven. The Universe of Things: On Speculative Realism. U of Minnesota P, 2014.
- Simondon, Gilbert. “The Genesis of the Individual.” Incorporations, edited by Jonathan Crary and Sanford Kwinter, Zoon Books, 1992, pp. 297–319.
- Star, Susan Leigh, and Karen Ruhleder. “Steps Toward an Ecology of Infrastructure: Design and Access for Large Information Spaces.” Information Systems Research, vol. 7, no. 1, 1996, pp. 111–134. JSTOR, doi:10.1287/isre.7.1.111.
- Weiser, Mark, and John Seely Brown. “The Coming Age of Calm Technology.” Beyond Calculation, edited by Peter Denning and Robert Metcalfe, Springer, 1997, pp. 75–85.
- Whitehead, Alfred North. Process and Reality: An Essay in Cosmology. The Free Press, 1978.