After Nature; The Dynamic Automation of Technical Objects

Luciana Parisi/Texts/Essays/After Nature; The Dynamic Automation of Technical Objects.pdf

P. 1
7 AFTER NATURE The Dynamic Automation of Technical Objects L U C I A N A PA R I S I T he dominance of information technologies in contemporary culture has led to the widespread use of algorithmic processing, which has transformed our conception of automation. From smartphone apps to high-frequency trading, from the use of multiagent systems in design to social-media censorship and national surveillancecontrol centers, automation, it will be argued, coincides with the proliferation of inhuman functions of decision making. While industrial forms of automation aimed to reproduce physical movement mechanically, the automated systems of the information age have become intelligent agents, exhibiting a degree of autonomy from their programmers through complex algorithmic interactions. It has been argued that the information age reveals a condition in which the overlap between automation and autonomy has been triggered by the speed of algorithmic procedures of evaluation, control, and decision making. In the contemporary context of “big data,” rule-based formal structures of data processing work at volumes and speeds that exceed the capacities of individual human cognition and operate beyond human awareness (Galloway 2006, Terranova 2004, Fuller and Goffey 2012). From this standpoint, this chapter discusses the paradoxical overlapping between thought and automation by questioning the assumption that technoscience—and the contemporary convergence of biotechnology, nanotechnology, information technology, and cognitive science—cannot
P. 2
15 6 INO RGANIC RITES (and should not) constitute the ground for explaining the ontological consistency of being. This chapter questions the claim that it is now, more than ever, necessary to distinguish human life from human technoculture in order to keep the ontological autonomy of being separate from the history of technical objects. This argument is not new, and there have been many attempts at rearticulating the relation between human nature and technology. As such, this chapter will pay particular attention to Gilbert Simondon’s conception of the technical object because his theory has been central to contemporary debates about the question of technology in human culture (Deleuze 1993, 2001; Brian Massumi 2009; Bernard Stiegler 2009). On the Mode of Existence of Technical Objects discusses Gilbert Simondon’s proposition that technical objects are not opposed to human nature but are instead the result of the artificial action of human beings through which a mutual relation with machines is engendered. I want to take this proposition seriously and suggest that this rather difficult question about the mode of existence of the technical object involves a fundamental acknowledgment that artificiality can and does surpass and irreversibly modify the autonomy of being. While rejecting both the hylomorphism of form and matter and the mechanical view of the universe that runs according to the repetition of the same initial conditions, Simondon’s proposition seems to embrace the technoscientific revolution of the Enlightenment, for which machines were not simply tools but became embedded in culture and used as a means of governance. However, for Simondon, while technical objects are cybernetic organisms, they are also importantly imbued with a potential to aggregate and change over time. They are responsive to the environment and constantly probed by what he refers to as human creative action (Simondon 1958). Technical objects are not transcendent structures abstracted from materiality only to reveal once again the ultimate horizon or the desire to transcend human finitude. Simondon’s view of the ontogenetic capacity of technical objects to exist in a dynamic field of relational constituency, I will argue, contributes to the disentangling of the philosophy of technology from the question of human finitude. In particular, it suggests that this disentanglement is crucial to an articulation of the ontological modality of technical objects. Nevertheless, this chapter also addresses the limits of Simondon’s theory—showing that this ontological modality is attached to an energetic conception of continuity
P. 3
AFTE R NATU RE 157 between what being is and how being does, which risks overlooking the tendency of the artificial to become autonomous from a mere energetic plane insofar as artificiality is also, and importantly, a construction or a conceptual architecture of a mode of thinking. To tackle this question, I propose we revisit the concept of the digital vis-à-vis the notion of computation by looking at recent developments in information theory—for example, Gregory Chaitin’s (2006b)—that expand on Alan Turing’s discovery of the limit of computation (that is, the incomputable or randomness) to suggest that automated forms of algorithmic calculation are not simply absorbers of human creativity. Instead, algorithmic systems of decision making seem to suggest that technical objects are not one with human nature and, more importantly, are not simply simulations of human cognitive functions. In opposition to the hype around interactive interfaces and the affective usage of digital media, I contend that in order to challenge the critical theory of technology that maintains a separation between human nature and human production, it is necessary to acknowledge that this production involves the automation of logical reasoning and cannot be explained solely as a mutual relation between energy and information. The production involves the inhuman function of decision making, which may be understood in terms of an emerging— or rudimentary—autonomy of and for mechanized thinking. In what follows, I will first outline Simondon’s philosophy of technology in order to explain how it is underpinned by a strong antiautomation view. I will then clarify the relation between digitalization and computation so as to locate my critique of Simondon’s energetic view of machines within the context of the historical development of the mechanization of logic. TECHN ICA L OBJE C TS For Gilbert Simondon, the mode of existence of a technical object needs to be explained fundamentally in terms of culture, but he also argues that the technical object is culture’s constitutive vehicle for artificialization (1980, 47). This means that the technical object needs to be divorced from the critique of instrumental reason and at the same time from any substantialist and essentialist ontology. Instead, Simondon claims that
P. 4
15 8INO RGANIC RITES “culture must incorporate technical entities into its body of knowledge and its sense of values. Recognition of the modes of existence of technical objects should be the result of philosophical consideration” (1). The mode of existence of the technical object, however, is not to be thought in terms of prosthesis or artificial extension of an always already incomplete human nature. The reproducibility of the natural into a mechanized form is not Simondon’s concern—he offers instead a counterintuitive theory that the technical object is at its most concrete phase when it is able to behave as a natural object. In particular, it is the concretization of the technical object (defined by increasing levels of internal resonances and by the crystallization of its phases) that makes it independent from the external regulatory environment (laboratory, factory) in which it was ideated. By achieving an internal coherence in which its systemic functions become closed and organized, a technical object therefore becomes comparable to an object produced spontaneously—independent from the environment and defined by its capacity to have incorporated external dynamics within itself. Its concrete state thus shows conditions of operation that are regulatory and self-maintaining. In other words, by becoming concrete, the technical object loses its artificial character. The concretized technical object approximates the mode of existence of natural objects because, like a natural object, it tends toward internal coherence, closing the system of the causes and effects that operate circularly inside its boundaries. It thus incorporates and transforms parts of the natural world that intervene as a condition of its functioning (14). But what does it mean to conceive of the process of concretization of the technical object in terms of a tendency to achieve a natural state? It is evident that Simondon is here pointing to a new notion of functionality of the technical object. Far from being the incarnation of instrumental reason, Simondon argues for a dynamic function of the technical object primarily understood in terms of an energetics of time (9–10). Here the functionality of the technical object is explained in terms of critical phases of change in which the past and the present are coextensive rather than being defined by discrete and sequential units of time. The technical object is not, however, defined by an autonomous ontology or an autonomous being of time. Instead, the technical object incorporates and is at once part of the time of the living. Its process of concretization involves a transindividuation of different fields of potentialities that allows the
P. 5
AFTER NATURE159 technical object to become individuated as an ensemble of living energies (62). This is also how a technical object is able to acquire a natural state and thus cease to be the artifice of man. From this standpoint, to enter Simondon’s philosophy of technology is also to ask: What does it mean for artificial objects to become natural? To what extent does this view offer us a nonhumanistic or nonanthropomorphic approach of the technical that sees beyond the horizon of human finitude? I want to suggest that Simondon’s emphasis on the mode of existence and, thus, of being—of the what and the how of the technical object—provides us with an important point of departure from which to develop a new view of the digital and of digital automation. In particular, I will focus on how Simondon responds to and radically transforms one of the kernels of the digital in his original analysis of the relation between energy and information. Simondon’s view makes us realize the importance of an entropic conception of energy for understanding the mode of existence of technical objects (9, 27–31). However, while recognizing Simondon’s novel philosophical account of the technical machine developed in view of the historical shift from a thermodynamic to a cybernetic conception of technology, I want to add that a closer analysis of information theory may show us that the digital has more than simply a regulatory function for energy. Instead, information theory presents us with a challenging view of the mathematic theory of communication. It is well known that Claude Shannon (1949) conceived of entropy not as a threat to communication but rather as an element of surprise. Extending Shannon’s insight, I characterize this element of surprise in terms of entropy that is immanent to information. From my discussion, it may become evident that Simondon’s energetic view of technical objects is fundamentally an ontological study of technical objects, one concerned with the being of or the becoming of technical objects. However, I suggest that the theorization of technical objects in terms of energetic accumulators of being, involving an ontogenesis of being, cannot explain how thought becomes formalized as it is incorporated in technical objects. By drawing on Gregory Chaitin’s (2006a, 29–32) view of the relation between energy and information in terms of the tendency of information to increase in size (i.e., information is itself entropic or non–size compressible), I aim to disentangle the symbiosis of being and thought granted by the mutual relation between energy and information.
P. 6
16 0 INO RGANIC RITES Drawing on the central role that Alfred N. Whitehead (1929, 1967) assigns to capacities of abstraction, I will tend to view the automated processing of information in terms of a mode of thought that includes both physical and conceptual prehensions, or the speculative functions of elaboration and the becoming thought of physical data. I do so not to merge being and thought but rather to explain the process of becoming as the asymmetrical continuity between the physical and the conceptual. But, to clarify these distinct levels, I will first explore Simondon’s energetic philosophy. A N TIAUTOMATION It is well known that Simondon rejected the positivistic view of the universe predicated upon the physics of reversibility premised in classical mechanics and sustained by the figure of the Laplacian Demon and its ability to calculate infinite amounts of information without error (Longo 2008, 68–70). Here, error is seen as an external and arbitrary incident in opposition to the function and efficiency of the machine. According to Simondon (1980, 49–50), this deterministic view of the machine, defined by the constant return to its initial condition, is represented by the regulatory function of information in cybernetics. As Simondon notes, the technical ensemble of the twentieth century—thermodynamic energy—is replaced by information theory, a guarantor of stability (regulation and stabilization). The cybernetic machine augments the quantity of information, leading to the emergence of negentropy that surpasses or incorporates the degradation of energy. The cybernetic machine is opposed to disorder, the heat-death of the universe. Its aim is to stabilize the world. However, this clear-cut reaction against cybernetics may perhaps reveal a richer history of cybernetics to which Simondon’s own theory has implicitly contributed. Cybernetics (and, in large part, computation) is less an extension of positivism and more an attempt at constituting a formal language that could take into account the inevitable truth of a universe losing the equilibrium theorized by concepts of irreversibility and entropy that constitute the core of the nineteenth century’s second law of thermodynamics. Also, contrary to classical determinism, which argues that all results could be known in advance, cybernetics accounts for the
P. 7
AFTER NATU RE 161 centrality of variations, or the impossibility of foreseeing whether previous conditions could determine future outcomes. Similarly, we know that Simondon’s philosophy is influenced by Heisenberg’s Uncertainty Principle and by Luis De Broglie’s study of the wavelike behavior of particles, which reveal that the exact measurement of positions and momentum cannot be known.1 Isabelle Stengers (2003), for example, points out that in Simondon’s On the Mode of Existence there is a shift from a notion of causal indeterminism linked to the effects of measurement to Niels Bohr’s conception of indeterminism beyond measure—here conceived as being constitutive of physical reality itself. Bohr’s conception of indeterminism assigns a constitutive role to the energetic fluctuation of matter and explains that randomness is not an incident of measurement but is instead the very condition against which measurement could not but fail. Indeterminism is said to be the veritable condition of microphysical reality. It seems safe to argue that it is this new condition that allows Simondon to reconceive the technical object away from mechanistic tradition and move toward embracing an energetic understanding of matter. In Modes of Existence of Technical Objects, we witness a renewed notion of physicalism, no longer inscribed in the substantialist, atomistic, and monistic traditions in which the smallest part is the universal absolute principle of the individual and of the individuated universe. Instead, Simondon’s thermodynamic conception of matter in terms of potential energy, of scales and degradation, is rather a radicalization of microphysical indetermination—arguing that any object is primarily inexhaustible despite the tendency toward increasing degradation. But for Simondon, decay does not correspond to the inevitable finitude of Being or beings. Instead his view embraces the quantum state of matter, revealing that the modality of being is defined by critical phases of energy in which nothing is lost and the potential reserve is never scarce. A fundamental dynamism is here sustained by the constant process of transformation of energy for which there is no finite moment or arrest. This energetic view also characterizes his insistence on the intrinsic validity of technology for human culture. From this standpoint, what Simondon calls concretization does not simply mean that ideas become physically tangible in human culture (a sort of Aristotelian physicalism of tools). Instead, technical objects are here cast as processors of live energy
P. 8
16 2INO RGANIC RITES that becomes crystallized, and somehow objectified, and thus rendered a mode of existence in itself. But this is not to be confused with a notion of externalization or passive prosthetics of human faculties. Simondon’s argument against mechanization instead aims to keep the technical object alive, to bring back its energetic impulse. In terms of culture, this means that technical objects are living repositories of ideas, practices, forms, and innovations that are modulated by or in a transductive relation with humans. Ultimately humans and machines do not share the same ontological conditions of being. On the contrary, Simondon is more interested in matter’s self-organizing tendencies, involving the assemblage or temporal synthesis of heterogeneous elements that constitute the machine. Hence, the mode of existence, or being, of technology is not to be framed in terms of a recuperation of ontology or an appeal to an eternal state. Rather, Simondon’s reconceptualization not only of the mode of, but also the condition of, the technical object crucially reveals that the process of concretization is above all part of a larger ontogenetic process of being. As such, the technical object enjoys a condition beyond the world of the constructed and exists within the very energetic field of matter. CYBERN ETICS In this section I will home in on the ways in which Simondon’s philosophy can or cannot help us sustain the theorization of the inhuman condition of digital, algorithmic computation. Contrary to much of the critique against cybernetics, for Simondon, information is neither an input nor an output—it is not to be looked at in its finite state of individuated data. Information above all instead corresponds to “more than unity,” insofar as its discrete, individuated, and binary state is inseparable from the preindividual field of energetic potential through which information individuates a difference. In other words, information is understood in terms of topological formations whereby a tendency toward change, and not a binary state, defines information in terms of its dynamic form, or “la bonne forme,” as Simondon puts it (1980, 190).2 It is this aspect of Simondon’s philosophy that provides a sense of the neutralization of information that results in a rejection of its discrete state.
P. 9
AFTER NATU RE 163 To put forward his philosophy of technology, then, Simondon calls into question both Shannon’s mathematical model of communication and Norbert Weiner’s conception of information. For Simondon, a philosophy of technology cannot be founded on research into form, or the efficiency of form, in the transmission of information. Against the univocal concept of information and energy, according to which these are equivalent and already determined unities, Simondon points out that information is not a preexisting form that is then to be applied to matter. In a radical move, Simondon reconceptualizes information as the variability of forms, the addition of variation to a form, and the unpredictability of variation in form (13–17). It is clear that Shannon defines information neither as form nor as meaning. Information instead denotes the uncertainty of the outcomes and/or the necessity of noise in the transmission of a signal. But the cybernetician Norbert Weiner gives information a regulatory function, ensuring that the thermodynamic tendency of a system to run out of equilibrium is reduced, modulated, and put to work. Under this interpretation, information is the measure of order and is located in opposition to entropy, which is marked as the measure of uncertainty or thermodynamic decay. Following Simondon, one could argue that not only is information univocally related to energy while energy is locked within a process of irreversible degradation but also that cybernetics establishes an equivalence between humans and machines in its assertion that any system—living and nonliving—resolves the threat of entropy through the regulatory function of information. This regulatory function can both lead to a return to homeostasis and more notably, channel energy in order to overcome the threshold of decay and thus enter a new level of order. This is the notion of negentropy (the negative use of entropy that results in a modulation of energy in a transmission channel): information produces more information and becomes the motor of evolution toward a more complex level of order. Simondon rejects this functionalist and instrumental view of information as the regulator of energy’s tendency toward chaos and instead argues for the necessity of an evolutionary mode of being whereby information is always already in the process of formation. This is not simply, for Simondon, the claim that emphasis on the notion of “in-formation” is not an essentialist return to a fundamental analogy. This is instead an ontological claim of a totally different nature.
P. 10
16 4INO RGANIC RITES He uses this notion to argue that the manner in which (or how) information becomes is constitutive of what it is. Thus not substance but quantic indeterminacy is the motor of the emergence or modification of form— information has no content, no structure, and no meaning, in itself; it is disparity, not univocity. But what does this mean, and how does this explain the mode of existence of technical objects as involving a becoming natural of the artificial? To answer this question, first we need to bear in mind that Simondon’s fundamental conception of self-differentiation or individuation is characterized in terms of a primary disparity between information and energy. The mode of existence of a technical object thus implies a resolution between two different orders of reality— information and energy—sustained by the operation of transduction. This also means that the disparity between information and energy cannot be understood in quantitative terms and is therefore free from any kind of stable formalization. Nevertheless, it may perhaps be reductive to argue that Simondon rejects all notions of quantification as a reduction of qualities to probabilities. He is known, on the contrary, to embrace the puzzles offered by the quantum physics of his time. Rather than being reductive, then, Simondon does not simply refuse but audaciously rearticulates the notion of quantity, quantification, and measuring from the standpoint of quantum physics. Drawing from De Broglie, Simondon uses the idea that a quantum leap coincides with the discharge of a measurable amount of energy but adds that such a leap also importantly coincides with a passing of a threshold to a qualitatively new level of existence. This inclination to explain the relation between energy and information in terms of what could be defined as the quality of a phenomenal experience also reveals that his indebtedness to De Broglie’s notion of indeterminacy is based on the observer’s measuring effect on the movement of quantic particles. Nevertheless, Simondon’s close reading of this scientific enterprise adds another level of understanding of indeterminacy. While he suggests that quantification is always laboring under a deficit of potential and that formalization always labors under a deficit of energy, he crucially argues for a notion of qualitative differentiation and energetic variation of form that defies the centrality of the subject and the human. Extending Bohr’s conception of indeterminacy as the fundamental condition of the microphysical fluctuations of matter, Simondon endorses a
P. 11
AFTER NATU RE 16 5 microphysical understanding of the technical object. This means that he rethinks the exchange between information and energy in terms of the qualitative variations of quantity. Here quantity is no longer described in terms of measure—as external observer or tool for measure—but becomes the manifestation of energetic variation. This means that there is no information without energy and that the relational exchange between them is precisely the process that ensures the constant transmutation of quantity into quality. The exchange indeed is not determined by a system of equivalence but rather of irreversible differentiation whereby information is always already affected by the energetic impulse. Here the energetics of a system are not defined by the system’s decay; rather, because of the ontogenetic force of the relation, information becomes a catalyst for novel ensembles of elements and individuals. In other words, one can suggest that for Simondon it is not measurement that involuntarily causes fluctuation but that it is information itself that becomes an energetic form of measure. This is a key point, as it places Simondon outside the phenomenology of experience and brings his work closer to a metaphysical investigation in the mode of the existence of technical objects—to the dynamic being of technical objects. In this sense, information ceases to be seen as the regulatory form of pure chance (otherwise described as entropy), and, going back to Simondon’s radical notion of la bonne forme, information is seen now as corresponding to a process of individuation or ontogenetic formation (Simondon 1992). Information seen in this way includes the potential field for difference that triggers and conditions the process by which individuation occurs (Simondon has in mind a form-taking informational activity that starts from preindividual energetic variations). But it is a mistake to compare this preindividual field to an amorphous pool of flows. The preindividual is not undifferentiated but is radically indeterminate; that is, it does not correspond to an ontological being but circumscribes the margins of indetermination that are always left to run with, or are consistent with, the informational operations of the technical object. One could, thus, argue that Simondon offers a new model of cybernetics that, although it is set in opposition to the mechanistic view, nonetheless embraces information by way of its proposal to understand measuring and quantification as defined primarily by the indeterminacy of physical reality.
P. 12
16 6 INO RGANIC RITES So, what does it mean to rethink information in microphysical terms? As opposed to a function set to reduce energy or to sacrifice and/or use qualitative variation, information is the expression of individuation and differentiation. In its transductive process of energy exchange, information is an event. In other words, information is subtracted from scientific objectivity and positivistic measurement in order to expose the levels of crystallization emerging from its relation with energy. One could also talk here of an affective relation to the extent that information is the affected recipient of an uncontainable spring of energy. What does it mean to com-penetrate quantity with quality or to suggest that the indeterminacy of physical reality is fundamental to understanding what appears to be the inhuman or inorganic dimension of technical objects? And what are the consequences of developing an energetic conception of information? Can such a microphysical conception of the technical object indicate a posthumous life in which it is possible to account for a continuous relation between the biological and the technical, human nature and human production, without falling back into binarisms and flat ontology? And finally, to what extent can this microphysical motor of machines coincide at all with the same level of order as technical systems? By merging the molecular—physical and biological—and the technical orders of function into one plane of continuous variation, information itself becomes neutralized and unexplained. By specifically addressing Simondon’s dynamic conception of information, Gilles Deleuze (1993) helps us clarify that this form-taking activity is not the result of an external relation nor the result of interactive communication between energy and information. On the contrary, this activity involves “at least two orders of magnitude, of two disparate scales of reality . . . it implies a fundamental difference, like a state of asymmetry. If it is nevertheless a system, it is only insofar as difference exists as potential energy, as a difference of potential distributed within certain limits” (246). Deleuze conceives of difference in terms of an internal principle of dedifferentiation through which he argues that quantity comprises a difference in itself that cannot be precisely measured and, thus, remains an intensive quantity. Individuation on this model reads as the organization of a solution, of a “resolution” (derived from two complementary yet asymmetric activities) for an objectively problematic system (a metastable system existing in a state of asymmetry). This means that, on the one
P. 13
AFTER NATU RE 167 hand, difference coincides with internal resonance, “the most primitive mode of communication between realities of different orders” (Simondon 1992, 298), and on the other hand, it is information that, in its turn, establishes communication between two disparate levels—one defined by a form already contained in the receiver, the other by the signal carried from the outside. The relation between information and energy, however, never involves a simple breakdown. Instead it entails a temporal resolution—a crystallization or concretization in the form of critical phases of change. The relation therefore is also a self-differentiation, always a more than one, so that the relation is both superimposed above the terms and simultaneous to itself. Nevertheless, the relation as individuated is again multiple because it is “multiphasic”; it is a “phase of becoming that will lead to new operations” (Deleuze 2001, 49). From this standpoint, one can understand the digital object as a mode of existence of a metastable system, which is itself a structure (not yet a synthesis) of the heterogeneous. At this point, it is interesting to note how Deleuze and Deleuze and Guattari follow Simondon in theorizing the technical object in terms of machinic assemblages and are therefore motivated to ask to what extent, if at all, the machinic assemblage embraces a nuanced conception of information. One could argue that it is possible that a machinic assemblage involves the capacity of the technical object to behave as a metastable system in which the artificial is no longer built out of what exists in nature. If so, the machinic—but not the mechanical or the technical—becomes a natural object only to the extent that it is placed within the larger field of metastable systems in which the relation between information and energy primarily leads to—or becomes the causal efficacy for—the concretization of preindividual potentialities. In this case, information ceases to coincide with a quantification of energy and becomes the subjective form—in Whiteheadian terms—of energetic force. Entropy can then be thought as productive and not simply regulated by information; its power is amplified rather than repressed. Far from being the measure of (deterministic) chaos, here entropy marks the margin of indetermination of any system of communication. It becomes the condition for communication rather than the moment of its deconstruction. Entropy is indeed no longer put in contrast to information, insofar as quantity cannot be divorced from difference. Similarly, noise cannot be completely silenced or expelled from
P. 14
16 8INO RGANIC RITES the signal. Entropy is thus amplified and distributed across both communication channels and the algorithmic infrastructure of the communication apparatus itself. But what is entropy in the digital? Do digital systems not operate precisely by transforming energy into information, modulating variation through an immanent mode of calculating potentialities, ergo, at base by eliminating friction? Like Simondon, Deleuze and Deleuze and Guattari reject the automatism of communication, and at the same time they propose a new notion of information and of the technical machine in which energy is the motor of computation (Deleuze and Guattari 1983, 1996). Defying the deductive method of formalism upon which computation is based, the machinic is envisioned as a praxis defined by the encounter, the affective and transductive production of the absolutely novel. The disparity between information and energy is, thus, resolved by means of reverse causality. Entropy is not the effect of measuring the incident emerging in communication but is of another order all together that constitutes information. My initial question about what it means for the technical object to become naturalized through concretization needs to be readdressed. In line with Simondon, the machinic is not defined by an ontological equivalence with the natural. Instead, the machinic explains the primary process of difference defined by the superiority of the relation, the transductive (and not deductive or inductive) operation of the encounter. Taken in this way, the machinic is precisely this continuous process of transduction between energy and information, not simply a naturalization of the technical. One can locate this point in Guattari’s (1995) discussion of the emergence, within the postmedia age, of machinic heterogeneities understood as living or live machines and no longer as a subset of technology. This is a big departure from Laplace’s or Turing’s universal, mechanical models in that the machine is ontogenesis, a mode of existence or a modification of being, an irreversible transformation of nature. However, if the machine acquires the status of a living being, one must reject the specific function of information because it, thus, can only be subsumed in the field of energy and cannot exist on its own. From this perspective, it is impossible to surpass the incumbent paradox that has afflicted the critique of digital technologies: While technical objects cannot be reduced to their information function because their processes of individuation coincide
P. 15
AFTER NATU RE 169 with becoming alive, digital objects, with their heavy reliance on discrete states and formal axiomatics, can never achieve such a state. What we are left with is the assumption that the becoming alive—or concretization or naturalization—of technical objects involves not a posthumous life or the seeping of inhuman, inorganic functions into the biological order of the organism but a tendency to incorporate the inhuman into the organic order without recognizing the information capacities of technical objects to be more than organic. COMPUTATION In this section, I return to theories of information that have tackled the question of entropy in the context of algorithmic machines, and I propose that the digital function of information needs to be contextualized in the history of mechanized logic. I revisit these kernels of cybernetics and information theory, from both the computational model of quantification and from the problem of entropy in systems of prediction and control, in order to develop a view about the specific function of information in automated systems. As already mentioned, I suggest that the interactive paradigm, which more clearly sits in second-order cybernetics, has become the dominant form of mediatic communication today by way of its expansion of interactive algorithms. This paradigm specifically involves: 1. An organization of data no longer originally inputted into the system but rather derived from the outside via the social and personal use of software and the capacities of algorithms to interact with one another in the computational architecture of digital media 2. A computational model of prediction based on an open architecture of axioms, whereby new rules can be added as a result of the adaptive response of algorithms to the environment 3. A conception and use of the incomputable, whereby real numbers or infinities of data are no longer incalculable but are part of the data landscape that algorithms transform into patterns that then become quantified as randomness, complexities, or Omega infinity algorithms
P. 16
170 INO RGANIC RITES The dominance of the interactive paradigm thus requires a new level of critique that can no longer hold the algorithmic regime of communication in opposition to the qualitative form of variation of the living, the subject, and thought. A return to computational theories reveals that incomputables or randomness—noncompressible quantities of information—are part of the computational tendency to elaborate information and establish a functional order of integration of data at the limit of the incomputable. Not only are incomputable quantities of data at the epicenter of communication, but also they unravel the importance of another order of entropy. Here entropy—the measure of disorder/chaos/noise in a system, or simply uncertainty—is no longer, or not, exclusively resolved through a negentropic process of transduction whereby the information becomes modulated as if it were energy; the problem of algorithmic randomness in the computational processing of increasing quantities of data instead denotes the tendency of information itself to increase in size—defining a fundamental complexity corresponding to patternless information. Gregory Chaitin’s (2006a) algorithmic information theory shows that the incomputable is not a probability known in advance that the program is expected to perform: Incomputables are maximally unknowable information quantities—infinities—that are nonetheless detected within computational processing as incomputables. These are at once discrete and infinite quantities of information. On this interpretation, computation does not resemble the Laplacian Universal Mechanism that repeats preset conditions, since incomputables (or randomness emerging in algorithmic sequencing) define computation as an incomplete affair. Likewise, such incompleteness ceases to mark the limit of information in terms of measure, as it involves the discovery of algorithmic entropy, that is, the capacity of information to increase in size/quantity. In short, in the algorithmic age, entropy cannot exclusively be explained by recourse to the function of primary energetic order. Whereas the function of energetic fields cannot of course be denied, what happens in the computational processing of vast amounts of data is indeed the emergence of patternless information. This defines the tendency of information as increasing in size complexity and reveals the function of information as implicating an algorithmic elaboration of primitive data. As such, in the order of iteration of the algorithmic function, this continuous compression of data leads to
P. 17
AFTE R NATU RE 17 1 the emergence of Omega, or patternless information—of which only few ciphers can be decoded with rational numbers. This form of information complexity, however, must not be confused with an absolute incalculability that breaks down the ceaseless patterning of data into information. This entropic or size complexity of information can instead be used to explain how information constitutes its own domain of complexity, a domain emerging from the function of integration of data and involving an algorithmic metaelaboration for which randomness cannot be reduced to algorithms yet emerges from its own function (Chaitin 2006b, 74–81). The size complexity of algorithmic information as described above points to an even more alien (or nonreducible to the energetic ground) complexity rather than an intensive notion of information. Drawing on Simondon, we saw that Deleuze argues for forms of energetic variation against the univocal relation between energy and information. Rather than opposing discreteness and continuity, or the digital part and the differential dynamics of the whole, the incomputable exposes incompleteness and complexity both at the core of discreteness and in models of quantification. On this view, one can suggest that the extension of microphysics to the technological field problematically overlooks the algorithmic order of the function of information and its own levels of complexity (that is, the incomputable character of the discrete). One could therefore argue that information entropy is central to what could be theorized as third-order cybernetics, which also explains anew the persistence of an inhuman form of elaboration of data that cannot be explained in terms of its physical causality. But this new centrality of the algorithmic randomness/entropy embedded into computational interactive systems implies that it is no longer possible to overlook either the discrete nature of the digital or the possibility of algorithmic randomness in digital sequencing. It is therefore also impossible to exclude the mechanization (or the becoming inhuman) of the function of reason by either remaining attached to the primacy of the being of the sensible (or of affective potentiality) or by returning to a cognitivist being of the intelligible—a formalism that imposes a logic of symbol manipulations on data. What I am proposing here is neither a re-vision of computation attached to the ontology nor the ontogenesis of being. Computation, within the context of algorithmic machines, instead points to an ontological process
P. 18
172INO RGANIC RITES by which an inhuman thought emerges from the complex elaboration of data and patterns. The notion of information entropy in this instance enables a new understanding of the nature of the digital itself as expressed from within a dynamic conception of computation, already envisioned in the nonlinear form of interactive algorithms and parallel distributed systems. But this dynamism stems not from an energetics of information but rather from within the quantitative nature of information itself by way of the computational problem of incomputable and incompressible randomness. To revisit the question of the digital from that problem implies a retheorization of the meaning of information. The quantitative attribute of information denotes not simply a mechanical view of complex computational dynamics but points to the problem of incompressible quantities within algorithmic systems: a potentiality that is properly quantitative and involves the entropic character of quantities, or the irreversible tendency of information to increase in size. Despite its limits, Chaitin’s work on the problem of the incomputable produces a new notion of both the digital and information—now defined by incomputable dynamics (irreversible change) rather than a relation of exchange with energy, which is not to dismiss the epistemological and ontological importance of such a relation. I intend to unpack the layers of dynamism in both the microphysical and the computational order. In other words, far from extending the microphysics of energy to information, I revisit the digital in order to find a way to articulate a computational dynamism proper to the strata of encoded information. Simondon’s recuperation of the technical, the technical machine, and, to some extent, reason from the cognitivist and representational understandings of technology constitutes an important step in the formulation of an inhuman form of information that bestows the tendencies to the artificial. However, I argue that, while this microphysical conception of the technical object has now become central to the dominance of an energetic view of interaction in the computational paradigm, further scientific and philosophical investigation into the algorithmic nature of computation can help us challenge the mechanical view of technical objects more directly. Doing additional research would, I think, expose the irreconcilable, perhaps nonresolvable, schism defined by the incomputable dimension of quantification rather than a mutual relation between information and energy. This is not simply to delimit our engagement with the
P. 19
AFTE R NATU RE 173 complexity of computation to the mere level of information but to zoom in to this informational level in order to amplify the levels of dynamic change and thus propose that the two orders of randomness or entropy— both at the informational and energetic levels— point to a computational process in which one order cannot be collapsed onto the other. IN HUMA N THOUG HT Information randomness indeed cannot be resolved by energetic randomness but instead needs to be explained and unpacked in terms of a real mode of information production, one where the precondition is not being organically bound. I borrow this idea from Whitehead’s discussion of the function of reason, which is set in contrast from both pure and practical reason and from deductive and inductive methods—he insists that the function of reason remains speculative (Whitehead 1929). This nondeductive understanding of reason and ultimately of a function of thought that can draw consequences from the physical prehension of reality, sustains my view that the artificial dimension of thinking encapsulated in the computational order of information also involves a speculative functioning that goes beyond the ontological grounding of being in the energetic impulse of living. In a sort of reversal between life and thought, being and reason, Whitehead insists that thinking is a function for living. In other words, for Whitehead, reasoning involves the conceptual elaboration of physical data, or the envisaging of the consequences of such a process, which precisely requires placing reason after nature. Instead of a bifurcation between the complexity of nature and the complexity of thought, Whitehead insists that the function of reason is to elucidate, explain, and unpack the reasoning behind living and thus beyond the art of surviving. “Reason is a factor in experience which directs and criticizes the urge towards the attainment of an end realized in imagination but not in fact” (1929, 16). The finality of reason, however, is not part of a grand plan toward a preestablished schema. Instead, it is embedded in the actual occasion of which reason has a constructive function. This means that while Whitehead recognizes that all thinking emerges from the biophysical constraints of the living, he also argues that the function of reason is to
P. 20
174INO RGANIC RITES elucidate and evaluate the causes through which these can be transcended. For Whitehead, the function of reason has almost a metaconceptual task, to the extent that it transcends immediate fact by relating systems of ideas and generalizing observations. In other words, the function of reason involves a passage leading from the awareness of what is observed in (sense) experience—also defined in terms of a physical prehension—to the conceptual understanding of this observation, which leads to complex related entities as presented by mathematical thought (1920, 13–15). In particular, for Whitehead, mathematical thought enlightens all thought; mathematics is an essential element in the history of thought. Whitehead’s process philosophy thus leaves an important place to mathematical logic and to the importance of separating mere matter of fact from the purely abstract conditions that they indicate. The function of reason is not determined by the direct apprehension of experience but is rather a function of abstraction of the particular entities involved and crucially involves the elaboration of the general conditions of the observations that they are expressible without having to make reference to particular relations or to particular relata occurring in that particular occasion of experience. For Whitehead, the rational attainment of this condition of generality ensures that these hold for an indefinite variety of other occasions (1967, 24–25). With mathematical logic, the concreteness of entities and relations has become a variable and absolutely abstract propositional function. This does not mean, however, that pure mathematics is imposed on concrete occasions. Instead, the function of reason is precisely to explain the process by which logic or the establishment of rules is attained through a process of abstraction in which the varieties of sense awareness are conceptually prehended. Whitehead’s view on the speculative function of reason thus can help us push Simondon’s philosophy of technical objects in another direction. As I have suggested, Simondon’s reflections on the mode of existence of technical objects inscribes the relation between energy and information into a physicalist-oriented schema concerned with the material—or even biophysical—order for which information patterns are fundamentally propelled by a preexisting energetic field of dynamic or highly perturbed level variations. In other words, this relation is defined in terms of a transductive modulation by which information ultimately coincides with informed energy. For Whitehead, instead, process involves various levels of abstraction, including physical, conceptual,
P. 21
AFTE R NATU RE 175 and metaconceptual or generalizing systems of relations that highlight the importance of a rational articulation of data. While for Simondon, the conceptual, ideal, or abstract configuration of a technical object needs to become concrete in order to be an effective or bonne forme of a culture, society, and mentality, Whitehead instead argues that indetermination, or not knowing, is the condition by which conceptual generalization—or determination—can be achieved through distinct degrees of prehension. From this standpoint, Whitehead’s insistence on the central function of reason can support a view of technology that involves not simply a mechanical function of organization whereby information is defined by the algorithmic pattern that encodes it. His view allows us to shed light on the importance of the logic by which technology functions. This means that Whitehead’s processual ontology could allow us to explain the emergence of computational logic not in terms of predetermined axioms but according to an emerging abstraction by which the energetic field cannot simply be the grantor of the ordering of information into patterns and the logical abstraction of patterns into a series of determinations. While it is true to say that algorithms are above all encoders of complexity and that Turing’s Universal Machine was devised to compress infinite variations into finite instructions to accomplish determinate tasks, this algorithmic procedure cannot be isolated from the problem of the incomputable nor from the entropic tendency of information to increase quantitatively and add volume to patterns. From the standpoint of algorithmic information theory, computational logic has thus become ampliative, which means that results are not directly deducted from inputs (that is, they do not reproduce premises) but instead add more information to the starting set of instructions. In other words, the algorithmic elaboration of data also involves a growth of information. However, since the incomputable is the condition of information, the fallacy of deductive logic is not simply an error in the system expressing the limit of computation; instead, the emergence of a nondeductive logic of interactive, distributive, and learning algorithms implies the partial or continuous determination of randomness. Indeed, this nondeductive logic has its own function of reason that, one can suggest, implies a process of algorithmic determination of randomness. This can also be theorized as the advance of an alien thought in the form of complex algorithmic evolutive patterns that continuously— albeit partially—incorporate randomness.
P. 22
176 INO RGANIC RITES Nevertheless, in relation to my suggestion, one could easily argue that the function of reason in computational processing is strongly limited by the lack of a metaconceptual form able to accomplish anything else than a faster mechanization of decision making. My suggestion therefore may work ultimately to reinforce rather than challenge the fundamental problem of automation that Simondon took pains to avoid. In other words, while Simondon’s theory of a technical object may help us unearth the inhuman field of energetic potential that challenges the instrumental view of technology as an extension or a prosthetic of human skills, it could appear to some that the form of automation that I am proposing only, and at best, manages to reproduce the mechanistic assumptions of the formulation of the inhuman in the context of a software culture. Nevertheless, it is possible to recognize that the automated processing of information does not fulfill Whitehead’s model of the function of reason, insofar as algorithmic computation cannot achieve the order of abstraction that encompasses an infinite infinity of variations without representing any particularity. In other words, this kind of processing seems to be too attached to the operations of classification and labeling of information, however complex these may be. On the other hand, however, it is also problematic to dismiss the specificity of algorithmic information and the emergence of a computational formalism that moves from indetermination to the determination of randomness—or the incomputable. My point is that this kind of formalism could instead serve us to explain the complexity of the technical object in terms of the algorithmic procedures that run it and according to a form of logic that no longer obeys its initial premises or axiomatic preestablished truths. This is a form of abstraction that can neither be ascribed to a bland mechanism nor to the force of energetic impulse. Neither mechanicism nor vitalism thus will suffice to explain how important it is to radicalize both informational and energetic dynamisms because these have both become defined in terms of their margins of indetermination. This is true not only of the living but also, significantly, of thought—both concrete and abstract, both practical and theoretical, and in both its natural and artificial dimensions. From this standpoint, it is crucial that algorithms cease to be representational forms of energy and become hosts of the order of information entropy that must be addressed from within their computational matrices. The algorithmic information
P. 23
AFTE R NATU RE 177 theory’s conception of entropy can serve as a point of departure for theorizing a third-order cybernetics in which the technical object needs to be understood in terms of a function of reasoning that is inorganic and nonbiologically bound to any particular organism. How we understand this functioning, however, involves not falling back into the representational framework of cognitivism promoted in classic AI discussions. To conclude, I want to reiterate that my reflection on the advance of the inhuman thought of technical systems is inspired by the condition of randomness at the core of the complex function of elaboration in the algorithmic processing of data, which are embedded in socialities of all sorts within distributive agglomerations of systems. The task ahead, then, is to elaborate further the distinct orders of function of energetic and information systems without grounding them in a specific field (either energetic or mechanical). The task, therefore, is to analyze the process by which algorithmic classification and evaluation of data involves functions of abstraction and how these functions map the emergence of automated reasoning and its inhuman capacities of thinking after nature. NOTES 1. 2. This is also to say that Simondon does not reject but reevaluates the statistical calculus in the mathematical formulation of the principle of indetermination insofar as this calculation can be useful as a diagnostic tool of the effective reality of the quantic behavior of particles and the knowledge or the measure of it. A parallelism between what is and what is possible to know allows Simondon to propose a new notion of information as being always already attached to the reality of the preindividual. Simondon’s critique of the concept of information proposed by Norbert Weiner’s cybernetics was also a critique of the notion of totality of form proposed by Gestalt, which, contrary to the hylomorphic schema, based on the analogy between matter and form, pointed to a fundamental equilibrium of the form. However, for Simondon the concept of information had to account for an internal dynamics that sustained the process by which the unity of individuation could rather encompass the variations of form, the crucial variation emerging from the tension between energy, matter, and information. Information is thus important because it can resolve the disparity between matter and energy insofar as it is here posed as another level of condition for achieving the integrity of complexity of form. However, as I discuss in this chapter, Simondon’s notion of information is profoundly attached to the microphysical explanation that precondition information (and matter) is ontologically dependent on the highly vibrating consistency of the energetic field.
P. 24
178 INO RGANIC RITES WORKS CITED Chaitin, Gregory J. 2006a. Meta Maths: The Quest for Omega. London: Atlantic. ——. 2006b. “The Limits of Reason.” Scientific American 294 (3): 74–81. Deleuze, Gilles. 1966. “Review of Gilbert Simondon’s ‘L’individu et sa genèse physic-biologique.’ ” Pli 12 (2001): 43–49. ——. 1993. Difference and Repetition. New York: Columbia University Press. Deleuze, Gilles, and Felix Guattari. 1983. Anti-Oedipus: Capitalism and Schizophrenia. Minneapolis: University of Minnesota Press. ——. 1996. What Is Philosophy? New York: Columbia University Press. Fuller, Matthew, and Andy Goffey. 2012. Evil Media. Cambridge, Mass.: MIT Press. Galloway, Alexander R. 2006. Protocol: How Control Exists After Decentralization. Cambridge, Mass.: MIT Press. Guattari, Felix. 1995. Chaosmosis: An Ethico-Aesthetic Paradigm. Bloomington: Indiana University Press. Longo, Giuseppe. 2008. “Laplace, Turing, and the ‘Imitation Game’ Impossible Geometry: Randomness, Determinism, and Programs in Turing’s Test.” In Parsing the Turing Test, ed. R. Epstein et al., 67–83. New York: Springer. Massumi, Brian, Arne De Boever, Alex Murray, and Jon Roffe. 2009. “On Gilbert Simondon: Technical Mentality, Revisited.” Parrhesia 7:36–45. Shannon, Claude E., and Warren Weaver. 1949. The Mathematical Theory of Communication. Urbana: The University of Illinois Press. Simondon, Gilbert. 1958. Du mode d’existence des objets techniques. Paris: Aubier, Editions Montaigne. ——. 1980. On the Mode of Existence of Technical Objects. Trans. Ninian Mellamphy. Preface by John Hart. University of Western Ontario. https://english.duke.edu/uploads/assets /Simondon_MEOT_part_1.pdf. ——. 1992. “The Genesis of the Individual.” In Incorporations, ed. J. Crary and S. Kwinter, 297–319. New York: Zone. ——. 2005. L’invention dans les techniques: Cours et conférences. Paris: Seuil. Stengers, Isabelle. 2003. Cosmopolitiques II. Paris: La Découverte. Stiegler, Bernard. 2009. “The Theater of Individuation: Phase-Shift and Resolution in Simondon and Heidegger.” Parrhesia 7:46–57. Terranova, Tiziana. 2004. Network Culture: Politics for the Information Age. London: Pluto. Whitehead, Alfred N. 1920. The Concept of Nature. Cambridge: Cambridge University Press. ——. 1929. The Function of Reason. Princeton, N.J.: Princeton University Press. ——. 1967. Science and the Modern World. New York: The Free Press.