**Links to**: [[Power]], [[Agency]], [[Freedom]], [[System]], [[Society]], [[Foucault]], [[Deleuze]], [[Postscript to the Societies of Control]], [[Control]], [[Cybernetics]], [[Discipline]], [[Embourgeoisement]]. # Post-Control Script-Societies **Summary: the relationship between power, control, attention and script-modulation**. The acquisition and maintenance of social power depends on (changing) established habits: patterns such as languages, traditions, currencies, and other cultural norms—all of which will be framed in this chapter as chunking/parsing _scripts_. Control, as the exertion of social power through scripts, results in organizing mechanisms which come to determine (new) behaviors and material effects. The first line of argument explored is that what we call _power_ is, therefore, an effect of agents or groups exploring and exploiting how scripts control perspectives; i.e., attention. These processes can be understood as _flows_ of information which can be _modulated_ (Deleuze, Simondon, Hui), and are therefore amenable to an information-theoretic analysis.^[For the formal framework, the reader is pointed to the [collaborative paper written with Albarracin and Hyland](https://www.mdpi.com/1099-4300/27/5/522). The formal analysis offers a novel methodological approach for examining the habits of power by bridging philosophy, critical social theory and active inference.] It will be argued that power manifests in the capacity to modulate information-processing, and thus material conditions, through scripts that organize social attention and behavior. Scripts *make perspectives move*, and perspectives *determine how reality unfolds*. Observing the effects of scripts directing behavior as the basis of power and its material effects, leads to an argument about power as—effective and/or ever more possibilistic—computation. Where computation is understood as a _function_ occurring on a finite substrate that transforms an input into an output. Empowerment is what grants social agents more computing opportunities (as in: the concentration of attention and behavior as/or the transformation of substrates), shaping the unfolding of future trajectories. Starting from the context of active inference, social power is therefore conceptualized as a function of variables including attention-attraction capabilities and information-processing efficiency, which together enable both script-creation and/or maintenance. Scripts function as the dynamic structures which determine the flows of social attention and therefore leverage power. Through an analysis of script-function, it can be examined how power relations materialize through attention and/as information-processing, revealing connections between social habits, cognitive-behavioral patterns, and ensuing materiotechnical realities. These connections, we speculatively conclude, reveal a transfer of _pure function_ across substrates. <small>Keywords: power, control, active inference, scripts, information, computation, narratives, prediction.</small> &emsp; <div class="page-break" style="page-break-before: always;"></div> >That is why, when subjects, individuals, or groups act manifestly counter to their class interests—when they rally to the interests and ideals of a class that their own objective situation should lead them to combat—it is not enough to say: they were fooled, the masses have been fooled. *It is not an ideological problem, a problem of failing to recognize, or of being subject to, an illusion*. It is a problem of desire, and ***desire is part of the infrastructure***. > >Deleuze & Guattari, _Anti-Oedipus_, (1977 (1972), p. 104, our emphasis in bold). &emsp; >Technical objects, including online infrastructures, are built ‘from elements that are already technical’ (Simondon, 2017, p. 74) and algorithmic techniques constitute vocabularies of ***possible function*** that enable operation and serve as _**horizons for technical imagination**_. > >Rieder, _Engines of Order_, (2020, p. 306, our emphasis in bold). <div class=“page-break” style=“page-break-before: always;“></div> &emsp; >Today abstraction is no longer that of the map, the double, the mirror, or the concept. Simulation is no longer that of a territory, a referential being, or a substance. It is the generation by models of a real without origin or reality: a hyperreal. **The territory no longer precedes the map**, nor does it survive it. It is nevertheless the map that precedes the territory – _precession of simulacra_ – that engenders the territory, and if one must return to the fable, today it is the territory whose shreds slowly rot across the extent of the map. **It is the real, and not the map, whose vestiges persist here and there in the deserts that are no longer those of the Empire, but ours. _The desert of the real itself_.** > >Baudrillard. “The precession of simulacra”, _Simulacra and Simulation_ ((1987) 1994, italics in original, our emphasis in bold). &emsp; >The primary function of Reason is the direction of the attack on the environment. > >Whitehead, 1929, p. 11. > >I now state the thesis that the explanation of this active attack on the environment is a three-fold urge: (i) to live, (ii) to live well, (iii) to live better.” (ibid., p. 11). > >We can think of it as one among the operations involved in the existence of an animal body, and we can think of it **in abstraction from any particular animal operations.** (ibid., p. 12). > >This conclusion amounts to the thesis that 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.** (ibid., p. 11, our emphasis in bold). <div class="page-break" style="page-break-before: always;"></div> ### Ctrl + A (_Select all_) ### Introduction: _Power_ The concept of _power_ can be explored at several scales: from physical work all the way to _animated_ labor. In this chapter—parts of which are based on a collective article written with [Mahault Albarracin and David Hyland, published May 2025 on _Entropy_](https://www.mdpi.com/1099-4300/27/5/522)—the interest is to explore the concept as a phenomenon that can be understood as _unfolding across scales_, where information-processing; *computation*, becomes a crucial way to analyze how this unfolding occurs.^[The collective work, where the framing and formalizing it through active inference (AIF) is the work of Albarracin and Hyland, can be found in Albarracin, de Jager and Hyland, forthcoming.] Turing gave us the very generative idea that _controlling_ the _function_ of a computing machine occurs as a result of having a program (or _script_) of encoded instructions on a substrate (or “memory”).^[As Wittgenstein noted: “The Turing machine is an idealization of the human computer (Turing 1936: 231). Wittgenstein put this point in a striking way: Turing’s “Machines.” These machines are humans who calculate. (Jack Copeland citing Wittgenstein, p. 3 in “Computation”. In: Floridi, Luciano, ed. _The Blackwell guide to the philosophy of computing and information_. John Wiley & Sons, 2008. Copeland also notes how: “It was not, of course, some deficiency of imagination that led Turing to model his logical computing machines on what can be achieved by a human being working effectively. The purpose for which he introduced them demanded it. The Turing machine played a key role in his demonstration that there are mathematical tasks which cannot be carried out by means of an effective method.” (ibid. p. 7).] Socially framing the intractable computational complexities of power, we can read Michel Foucault as having given us a key insight into power as competing attentional forces emerging from and resulting in constraint-based dynamics (through what we will call scripts, or, in his words here: _strategies_): &emsp; > ... power must be understood in the first instance as the _multiplicity of force relations immanent in the sphere in which they operate and which constitute their own organization_; as the process which, through ceaseless struggles and confrontations, transforms, strengthens, or reverses them; as the support which these force relations find in one another, thus _forming a chain or a system_, or on the contrary, the _disjunctions and contradictions which isolate them from one another_; and lastly, as the _strategies_ in which they take effect... > > Foucault, 1990, pp. 92–93, (our emphasis in italics). &emsp; In other words, power is not a “thing”, but can be understood rather as a function, the effectuation of which depends on (disrupting) established encodings—or what this thesis calls _chunks_:^[See: [[Chunk]] and [[Parse]].] meta-patterns resulting from the scripting capacities of phenomena such as languages, traditions, currencies, and other cultural sedimentations.^[In the context of evolutionary function-exploration, Michael Levin talks about our current paradigm (cybernetics as resulting in bioengineering, agent-modeling in AI, etc.) as providing the possibility for exploring the adjacent (Kauffman) latent spaces of evolution, perhaps resulting in a more empathetic approach to understanding life and intelligence, through the observation of goals guiding different configurations of matter, leading to understanding functional invariance across species, substrates, intelligences, etc. See: Levin 2022.] Existing encodings—or habits—of power can reinforce each other (e.g., economic power might reinforce political power, which might reinforce cultural authority, etc.), and “chains” or “systems” (ibid.) as the habitual scripts that direct social behavior,^[What this thesis frames as _parsing._] can reinforce patterns of information-processing that maintain or change existing power arrangements. Seen this way, power emerges not as simplistic top-down control, but as the function by which attentional economies are orchestrated through scripts that channel information flows, advancing particular trajectories while foreclosing others. The “strategies” Foucault identifies, manifest in our framework as the manipulation of attentional scripts by agents/groups/infrastructures which (can come to) modulate behavior, and therefore give shape to material reality. Through these strategies, power can mobilize bodies to reproduce material conditions that further reinforce existing power positions within the social-attentional-informational economy. Following Foucault, the phenomenon—as well as the possible representations of what precisely constitutes the (social) acquisition and maintenance—of power is thus a complex series of multidirectional processes which cannot be simplistically or hierarchically schematized, but must be analyzed in terms of **interactions** between entities and the bidirectionally/retroactively-dependent **constraint-regimes** they are embedded within. As the historical evolution of socially-scaffolded infrastructure sought efficiency and adaptations to, but also _as_, novel technological landscapes,^[“It is precisely across the gaps of what can be known that new subjects and things are called into being. Thus the commercial retailer’s dream of an unknown consumer meets the state’s nightmare of an unknown terrorist; the iris-scanned, trusted traveler becomes a condition of possibility for attention to the risky traveler “of national security interest,” the designer of risk algorithms for casino and insurance fraud becomes a resource to homeland security.” Amoore 2013, p. 3.] traditional disciplinary enclosures became outdated or outmoded ways to chunk and parse people and other *matters*. As scripted by Deleuze: “The disciplinary model of enclosure ultimately proved too inflexible, unable to adapt itself to the demands of a changing economy for modulated controls over production” (Deleuze 1992 (1990), p. 32). Deleuze also revealed how traditional institutions were gradually eroded by other mechanisms of social control which “qualified as more humanistic and respectful of individual rights” (Paradis‐Gagné and Holmes 2022, p. 3). (Bio)power thus became a matter of _flow_-management, or _modulation_. Paradis‐Gagné and Holmes (ibid.) note how Foucault had, in fact, also come to the intuition of a dispersed panopticon: “One also sees the spread of disciplinary procedures, not in the form of enclosed institutions, but as centres of observation disseminated throughout society” (Foucault, 1995 (1975), p. 212). Deleuze’s notion of _modulation_, however, further elaborated how, what we could call computational or information flow efficiency, became the regulating tendency towards the end of the 20th century: systemic _seeing-as_. The dispersed panopticon is evidence to the refinement of _perspective_, itself, as a function. Enclosure^[As an uncertainty-minimizing structure perhaps comparable to a formal language, in its brittleness.] is not only brittle but also costly and highly visible, and thus also reasonably vulnerable. Modulation, as a dynamic function; distributed and less visible, is better suited for higher orders of (social) complexity, ensuing from novel technological conditions. In our theoretical interest, modulation is also revealing of how the simplifying script that is the _individual_ was never ‘real’ to begin with (hence Deleuze’s -_dividual_), but simply performed a series of regulative (social) functions.^[In social-group organization, the individual is a predictively-convenient way to determine the possible behavior of a phenomenon which is in fact, governed by rather dynamic attractors. The predictive myth of the individual possibly originates from the evolution of social organization, enculturation, which might have origins in the perceptual constraints of object-permanence. Foucault (1966), of course, already criticized this myth, and additionally, McKenzie Wark also notes how we can find critique of the myth of the individual in Marx already (and of course in Guattari), too: “Just as the pastoralists use the state to secure land as private property, so too the capitalists use their power over the state to secure the legal and administrative conditions for the privatization of flows of raw materials and tools of production in the form of capital.” (Wark 2004, section 105). Marx: “The individual and isolated hunter and fisherman, with whom Smith and Ricardo begin, belongs among the unimaginative conceits of the eighteenth-century Robinsonades. (…) In this society of free competition, the individual appears detached from the natural bonds etc. which in earlier historical periods make him [sic] the accessory of a definite and limited human conglomerate. (…) this eighteenth-century individual (…) appears as an ideal, whose existence they project into the past. Not as a historic result but as history’s point of departure. As the Natural Individual appropriate to their notion of human nature, not arising historically, but posited by nature. This illusion has been common to each new epoch to this day.” (GR, 83).] Society might retroactively inform a notion of “individual”: “In the individualistic society however the generality is realized not only through the interplay of individuals, rather the society is essentially the substance of the individuated [_Individuum_]” (Adorno, _Minima Moralia_).^[Version tr. Dennis Redmond, 2005.] Or, as presented by Moten and Harney: “The ascription of body, the imposition of bounded and enclosed self-possession, of a discrete self subject to ownership, of ownership activated and confirmed either in theft or trade, might be said to be the first reform, the first improvement, insofar as it is the condition of possibility of reform, or improvement. The assignment of body to flesh is the first stripe of the long, hard, torturously straight, tortuously straightened row.” (Harney & Moten, 2021, p. 62).^[Continuing on the dialectical oscillation between individual and system, here is also a fascinating citation from _All Incomplete_, which the reader should read with the idea of function as that which flows through the evolution of systems, in mind: “In the early 1950s, Deming and his colleagues introduced quality management techniques in Japan that aimed to make _individual_ workers responsible for the smooth functioning of their part of the assembly line – in other words, for its flow. There were two components of this responsibility. One was reducing mistakes that impacted on the quality of the product. This was to be done by making each worker responsible for those mistakes. The other was improvement, or, speeding up the flow. What quality management sought to do was individualize the speed-up, making it a matter of the worker’s personal responsibility. A big part of this was decollectivizing resistance to the speed-up. The result was a worker caught in the oscillation between submitting oneself to the line, to the flow, and asserting one’s individuation as a quality controller. The flow continued to be a force well beyond the control of the worker, but rather than responding to it with another collective force, the worker was now individually responsible for his or her response, which was to constitute both submission and optimization.” (p. 95).] Systems become systems as a function forces itself through the brittleness of substrates: words pass through speaking beings, power passes through (infra)structure, birth passes through creatures, fluidity passes through liquids. Traditionally, the framing of socially-facilitated empowerment—such as that of the state, or of a large corporation—has been understood as the effectuation of (often top-down) strategies, tactics, policies, narratives (i.e., _scripts_) which grant empowered agents/systems more _leverage_, that is: more influence on future states. We frame this in terms of information-processing capacities and opportunities, where powerful agents/groups/infrastructures either rely on other processes to bring about desired functions,^[Or, in the formal framework of AIF: _policies_. Where this can be understood as **exploitation**, in the biological and sociological senses: the bringing about of or reliance on a pattern (habit) that results in valuable thermodynamic advantages by reducing significant uncertainty.] or enjoy more information-processing possibilities as a result of relying on others for the reproduction of (material) tasks.^[I.e., **exploration**: more opportunities for speculation; future-projection, etc. are available to those who rely on others to reduce thermodynamic uncertainty.] The effects of (social) empowerment therefore imply an increased ability to compute information,^[Important to note here is that the connection between _computing power_ and _empowerment_ in general has been criticized by Philip Agre (1997) as coming to dominate the imaginary. In light of current events, it is difficult not to see how computation does not equal power though. Again, this also being in the context of this thesis, where we understand computers as processes which have inputs and outputs of varying granularity, which have _tendencies_ towards stable states but, in their totality, as an incomplete and unfolding process, cannot be said to arrive at specific states, ever, because they are not closed systems.] thereby augmenting the evolution of a specific state space. Basing these arguments on AIF allows for an analysis of “[p]ower [a]s the probability that one actor within a social relationship will be in a position to carry out [their] will despite resistance,” (Weber, (1925) 1965, p. 152) in terms of _actual_ probabilities.^[Weber’s quote ends in “... regardless of the basis on which this probability rests.” (ibid.). We will challenge this through a material-informatic analysis, as the basis of this probability is not trivial. See, e.g., Ramstead (2023), on the introduction of a “formal language for modelling the constraints, forces, potentials, and other quantities determining the dynamics of such systems, especially as they entail dynamics on a space of beliefs (i.e., on a statistical manifold)” for more background on the possible bridges between material and informatics in terms of Bayesian mechanics. Also note that the translation of “probability” in Weber has also been that of “ability“, “capacity“, and other terms, we take one of the most popular formulations here which also aids our purpose of framing this in terms of a probabilistic mechanics.] We present novel, perhaps speculatively risky, understandings of concepts such as scripts and attention, by analyzing power in terms of the _physics of beliefs_ (Ramstead et al., 2023), and the metaphysics of social organization. _Metaphysics_—i.e., the contemplation of highly abstract structures which may or may not be understood as fundamental—is for our purposes framed as a complex type of script, one which, in the context of AIF, can be understood as speculative future-modulation.^[This is treated in [[04 Concepts as pre-dictions]] and in [[05 Prediction]].] This leads to an understanding of metaphysics as the possibilistic exploration of state spaces (i.e., the future). In the sections that follow, an elucidation of different understandings of power and modulation is presented in order to set the focus on the concept of power as the control of attention via explorative and exploitative scripts, which can exist as *shared metacognitive computing conditions* (i.e., agents are relatively aware of the implications of scripts: they can observe them as constraints), *coercive domination* (where agents are subsumed under a narrative which does not benefit them, they are embedded in constraints without meta-reflection capacities), or: often an inevitable combination of *both*. Throughout the chapter these modes are synthesized towards elucidations into the unfolding structures of social scripts and power, and how scripts, as *semantic attractors*, can come to frame the concepts of narratives, attention, coercion and resistance differently, opening them up to modulation. The intention of this chapter is to look at power through the lens of AIF, the formalizations of which exist in the collaborative article mentioned, and not covered here. What is covered here are more speculative consequences leading from an understanding of power as attention-modulation and information-processing. &emsp; >Yet for countless others who romantically insist (as the usual insult goes) that another world remains **possible**, both the propagation of the neoliberal plague and the technopolitical consolidation of the societies of control—where the Market equals the State, the State equals the Market, and there is **no choice** outside them—can be confronted only if we retain our capacity to connect with **the flux of desire** that briefly broke the surface some forty years prior. For them, the pure event that was ’68 had **never ceased occurring or else has not yet even begun**, inscribed as it seems to be in a kind of **historical future subjunctive**. > >Viveiros de Castro, _Cannibal Metaphysics_ (2009) 2014, p. 98, our emphasis in bold. &emsp; The future seems to depend on the past, and computing this is costly as complexity increases from our current perspective:^[Not only does the arow of time proceed, and not only are there more _human_ perspectives than ever before (which are themselves multiperspectival in various ways), there are also new types of perspectives such as technological sensors, as well as new understandings _of_ perspective itself.] the more information; the more degrees of freedom. What the move from discipline to modulation (Deleuze) shows us, is how (scaling) systems evolve functionality to cope with this increasing complexity. In systems such as human culture, chunking patterns from the past in order to **make** the future, unfolds socially as the modulation of attention, controlled by scripts which parse chunks. In the context of AIF, agents and agent-ensemble-systems *self-evidence*, that is: they act on the world to gather evidence of their existence. This is what **active** _inference_ means: acting in order to effectuate a belief (where belief is an ample notion: encompassing anything from thermal regulation to political desire). Self-evidencing is that which ensues from having a perspectival model of action-consequences, a perspectival model which recursively bootstraps, as it is itself used for evidence of what takes place _for_ it: what it witnesses and how its actions are intimately coupled to that evidence. Think about how you need to touch your head to feel if a bird effectively pooped on it, or you just incorrectly assumed so. The feeling that you—as a thing separate from the rest^[See: [[Markov blanket]].]—were touched or not, becomes the ever-accumulating evidence for the model the thing—_you_—has of itself. This is what active inference scholarship argues is the case for all biological agents: everything that _functions_, persists, does so by following its self-model.^[Perhaps differently phrased: “Being is preserved as a dynamic structure whose operation is open to the incoming of contingency: namely, becoming.” Hui 2019, p. 27.] This function, which at the system level is observed in effects such as mass-surveillance but also love and kinship, is rather inevitable: anything that couples with its environment by enacting its own model, is self-evidencing (see: Hohwy 2016, and Hohwy forthcoming). As self-evidencing scripts become retroactively narrated; technically statable; recursive enclosures, the owl-of-Minerva perspective which results from this reveals pure _functionality_ evolving at its core. Following the _functional grouping_ approach in biology, by understanding the transfer of function across structures, we can hereby prioritize a long-expressed philosophical interest in _processes_, away from fixed identities and static form.^[E.g., Whitehead, Simondon. In biology, we can group functionalities by thinking about: all things that filter (moss, peat, barnacles, baleen whales, flamingos, many clams and mussels): “The simplest living things let their food swim into them.” (Whitehead, 1929, p. 10); things that pollinate (wind, bees, butterflies, bats, birds, even some mammals); things that de-compose others (fungi, bacteria, vultures, certain insects, erosion processes); things that alter soil structure (earthworms, prairie dogs, moles); things that disperse seeds (birds, mammals, ants, wind); things that hunt cooperatively (humans, wolves, orcas, lions, some spiders); things that echolocate (humans, bats, dolphins, shrews); things that hibernate (many plants, bears, squirrels, frogs); things that migrate (birds, monarch butterflies, sea turtles); things that build complex structures (crystals, beavers, termites, humans, birds). Our notion of function stays abstract in this chapter, but the reader is encouraged to think about _predictive abstraction_ itself as the function transferred through scripts across humans. See Moore et al., 1988 for more background on functional grouping.] What is diagnosed as _post-control_ in this chapter is the idea that “control” is an insufficient or limited diagnosis of power as exerted through function,^[Again, inspired by systems biology and functional grouping, our diagnosis is so if anything because of the attention to chunking/parsing scales we insist throughout this project: “The way in which one defines a process such as regulation is strongly influenced by the scale of one’s research. A community ecologist is likely to view a predator limiting the abundance of its prey as regulation. In contrast, an ecosystem ecologist might consider how a predator affects biomass and energy transfer rates between different energy states (trophic levels). It is important to remember, however, that these different approaches have the same goal: to identify the mechanisms controlling ecosystems.” (Moore et al. 1988, p. 420). If the goal is understanding the mechanism: this is why we take a look at scripts and how they transfer here. We may also add this citation here: “Spinoza asks: What can a body do? We call the _latitude_ of a body the affects of which it is capable at a given degree of power, or rather within the limits of that degree. _Latitude is made up of intensive parts falling under a capacity, and longitude of extensive parts falling under a relation_. (Deleuze & Guattari, (1980) 1987, p. 256). We know nothing about a body **until we know what it can do**, what functions it performs. “The rat and the man are in no way the same thing, but Being expresses them both in a single meaning in a language that is no longer that of words, in a matter that is no longer that of forms, in an affectability that is no longer that of subjects. _Unnatural participation_. But the plane of composition, the plane of Nature, is precisely for participations of this kind, and continually makes and unmakes their assemblages, employing every artifice. This is not an analogy, or a product of the imagination, but a composition of speeds and affects on the plane of consistency: a plan(e), a program, or rather a diagram, a problem, a question-machine.” (ibid., p. 258). These are _functions,_ in our terminology. “Descartes had to call in God, in order to push his bodies around.” (Whitehead 1929, p. 33.)] and what is diagnosed as _script-societies_ is the condition linguistic, encultured beings have always existed within: given to a script which transfers function(s) across substrates (and which itself, as a function, has an interest in continuing to transfer).^[All procedures couple to other procedures: we eat to live, and we live to eat.] The speculative argument proposed in this chapter is that, as complexity increases, systemic self-preservation reveals adaptation towards the evolution of _function_ itself.^[An argument partly presented in [[06 Principle of Sufficient Interest]], too.] At the system-level, the fact that we do not know what “artificiality” nor “intelligence” exactly are, yet, “Artificial Intelligence” seems to evolve somehow, is _evidence_ to this self-evidencing function. Because a thing does not ever have a complete model of itself, **it needs to continue to gather evidence for its own existence**. If media are the extensions of organically-given functions (McLuhan 1964), should we not prioritize an analysis of function over form? Enter function _modulation_ and _scripts_. &emsp; >Civilization advances by extending the number of important operations which we can perform without thinking about them. > >Whitehead, _Introduction to Mathematics_, 1911, p. 61. > >In this [disinterested, anarchic curiosity] function Reason serves only itself. It is its own dominant interest, and is not deflected by motives derived from other dominant interests which it may be promoting. This is the speculative Reason. > >Whitehead, _The Function of Reason_, 1929, p. 44. &emsp; The reader should be warned in advance that this chapter is the result of an integration between two papers and can therefore appear ambulant and erratic sometimes. If I would have had more time I would have amended this, and probably said all that follows in way less words, but I had to hand in this thesis. &emsp; ### Ctrl + F (_Find_) ### Reappropriating modulation away from... &emsp; > ...“‘the modulation-control correlation’. > > Hui 2015, p. 87.^[Full citation: “To re-appropriate the concept of modulation is, on the one hand, to acknowledge and deepen its philosophical significance as a materialist alternative to hylomorphism, and, on the other hand, to imagine new modes of modulation that don’t simply fall into the logic that Deleuze and others have described in terms of ‘control’. What we would absolutely want to avoid would be implicitly to propose a return to hylomorphism as a mode of resistance to social control modulation. Nor would it be appropriate to let go of the concept of modulation, with all of its philosophical usefulness (particularly as deployed by Simondon), simply because of Deleuze’s polemical characterisation of modulation as a key feature of control societies. Instead, it is the final aim of this essay to suggest exploring the concept of modulation under the motif ‘modulation after control’, getting beyond the limits of what we might call ‘the modulation-control correlation’.” (Hui 2015, p. 87).] &emsp; The _double bind_ (Gregory Bateson, 1972) refers to a psychosocial (and linguistic) paradox, any situation where we are stuck between incommensurable alternatives offered by a language or social directive, by a script. An example I always think of was a childhood friend whose mother would—angrily, screaming—‘ask’ her: “Why did you do x?” As my friend would timidly begin to utter a response, her mother would yell: “Do not talk back to me!” I watched in surprise, by her side, the doubly-binding and double-contingent unfolding of this. The double bind operates by placing agents in ambiguous or otherwise paralyzing situations where the actions that they ought to engage in conflict at different levels of engagement (e.g., my friend is asked to elucidate a possible fault or mistake, while at the same time in that social situation she ought to remain silent out of “respect” for her mother’s authority). Bateson noted how these situations of strange control abound in cultural learning. Circumstances where the double bind takes place as modulating manipulation, leading to conflicting disorientation, are what we also often mean when we talk about the exertion of control of a group (or individual) on others. In the context of “western” and “democratic” societies, Mattin (2022) has analyzed this double bind as a strong form of _social dissonance_, where messages of _freedom_ and _equality_ plaster our imaginary, while the reality is unfree and disproportionally inequitable. Deleuze’s _modulation_ in the _Postscript on the Societies of Control_ (1992 (1990)) can be seen as an entry into the mechanics of the double binds pervasive in the scripts of ‘control’ societies. The kind of modulation we continually refer to in this project can be understood to include Deleuze’s modulation, but our kind is intended in a far more affirmative sense.^[See: [[Language-modulating]], for example.] In “Modulation after control” (2015), Yuk Hui explores Deleuze’s concept, as inherited from Simondon, and notes that Deleuze already refers to Simondon’s concept of modulation in _Difference and Repetition_ (1968).^[“... opposing an idea of experience as the modulation of the sensible to Kant’s subordination of the sensible to intuitions and categories. In _A Thousand Plateaus_ (1980), Deleuze again refers to Simondon’s concept of modulation, and his critique of ‘hylomorphism’; in his book on Leibniz _Le Pli: Leibniz et le Baroque_ (1981), Deleuze simply quotes Simondon to explain what he (Deleuze) means by modulation.) and Artaud (“Deleuze talks about Artaud and the modulation of language, a language without articulation.”) (Hui 2015, p. 76).] Following from Simondon’s critique of hylomorphism,^[The Aristotelian ontological preference for form as superseding over matter: form _shapes_ matter. Simondon critiqued this mode of thought by looking at the complex intricacies in processes of individuation, where things are far from simple. “Simondon wrote ‘becoming remains conceived as movement, and movement as imperfection’” (Hui 2015, p. 75, see also notes about Nail’s take on movement in: [[10 Bias, or Falling into Place]].). A similar and more modern critique, starting with a deconstruction of the notion of _identity_, is done by Juarrero (2023), too, towards an understanding of how constraints shape the complexities of reality. While we agree no primacy should be given to static images of _identity_, we do think that morphology, as something which transfers across organisms, is quite “loose” from substrate. This is a functionalist perception of processes that keeps driving this project’s curiosity about ideas of transcendence, teleology and form. “As Marx puts it, “the anatomy of the human is the key to the anatomy of the ape.” Marx, _Grundrisse_ in McLellan 390.” (Brassier 2019, pp. 103-4).] Deleuze presents modulation in the _Postscript_ as the shift between “‘moulding’ to ‘modulation’, namely from a form-imposing mode to a self-regulating mode” (Hui 2015, p. 76). By his own methods, Deleuze is retroactively understanding social control as unfolding in a new technological regime which has rendered _remote_ control, action at a distance, and rhizomatic virtuality quite _real._ Hui notes that Deleuze did not go far enough in exploring the technological condition rendering this modulation, and suggests that Simondon’s 1961 paper ‘Amplification in the Process of Information’ provides a good insight into processes of technical amplification and individuation, showing how “modulation can also be understood as a way to resist the tendency of ‘disindividuation’ in control societies,” where the “modulative’ mode of control societies is only one possible outcome from the philosophical concept of modulation.” (Hui 2015, p. 74). As noted, this is closer to the kind of modulation we are after, too, one which is able to meta-learn by the making apparent of constraints.^[This is treated in [[06 Principle of Sufficient Interest]], [[10 Bias, or Falling into Place]], for example.] Modulation, in our context, is the process by which _function_ is shaped or tuned. If hylomorphism “operates dialectically (form+matter=synthesis), then modulation operates in terms of disparation [disparity],” (Hui 2015, p. 75). We can therefore understand it as a form of friction, constraining resistance to the visibilities of dialectical control, because it can be understood as that which comes to operate at the level of the constraint regimes: the _field_, the territory, rather than the mapping; the indexical; the representational.^[Brassier: “We gain access to the structure of reality via a machinery of conception which extracts intelligible indices from a world that is not designed to be intelligible and is not originarily infused with meaning. _Meaning is a function of conception_ and conception involves representation—though this is _not_ to say that conceptual representation can be construed in terms of word-world mappings. It falls to conceptual rationality to forge the explanatory bridge from thought to being.” (2011, p. 47, our emphasis). If access is gained through this “function” we could therefore extrapolate that meaning, through scripts, is what transfers through these processes as a function. In this piece Brassier also confronts Deleuzian immanence, which would result in pansychism, as well as what he diagnoses as Latourian relativism, particularly when it comes to “truth as power”, we do not know where we stand here. (ibid., p. 51).] Hui sees Deleuze as missing a philosophical opportunity in the concept, through its reduced presentation in the Postscript, given that much of what Simondon had proposed was the cracking open of a metaphysics which had rendered a very particular social, technical and political transformation as representable, statable, due to the ontological opposition between becoming (change) and being (identity). Hui looks into the dark-deleuzian^[This is in reference to Andrew Culp’s _Dark Deleuze_ (2016), a book which questions the negative, non-affirmative, inconsistent, even fascist-capitalist-oppressive, or otherwise more “problematic” (in the dark-deleuzian sense of the word) aspects of Deleuze’s philosophy.] question of why—if his own _modulations_ (as rhizomes, flows, as virtualities) had always been presented in the affirmative, _becoming_-sense; as an alternative to ‘substance _versus_ matter,’ to mould-like hylomorphism or identitarianism—does it become the very _weapon_ of societies of control in his 1990 _Postscript_? Denise Ferreira da Silva (2016) presents a similar consideration when she asks, if Western metaphysics has always put its models to the test on the _rigor mortis_ of its timely sciences, then how come we cannot imagine a metaphysics of impermanence and becoming (or of _difference without separability_) on the basis of our **current** scientific findings, which reveal all manners of unstable identities, observer-dependencies, etc.? The confronting answer might be dark-deleuzian, too: what currently unfolds is, indeed, the case. This is what gives us modulatory control and disoriented -dividuals. This is what gives legitimacy to saying and doing _anything_. To Hui, however, the answer lies in _conceptual modulation_ itself, by the hand of _active_ modeling: “Conceptually one opposes modulation and moulding; in reality, modulation and moulding co-exist, and consists [sic] of a hybrid mode, which Simondon calls modelage.” (Hui 2015, p. 79). In Hui’s argument modulation, or *modelage*, becomes affirmatively recovered by recontextualizing it conceptually _and_ technically, by observing the social possibilities that contemporary technologies can tune or modulate.^[Hui does this through the modulation of functions in a possible alternative social network example, where, e.g., a “rearrangement of relations makes the group and project the default instead of the individual.” (ibid., p. 90) , we briefly present “Debate mapping” in the conclusion in a comparable way later, too.] Hui also considers Simondon’s notion of information—which is very close to that of Bateson’s famous “difference that makes a difference”—as an operation of individuation: “... information is not a thing, but the operation of a thing arriving in a system and producing in it a transformation. Information cannot be defined outside of this act or this transforming incidence, and the operation of reception” (Simondon 1961, _‘L’individuation dans les processus d’information’_, p. 159, quoted in Hui 2015, p. 80). Information is not a ‘thing,’ but it is _specific_, determined, when contextualized as a difference, it is what reveals the processes modulated by functions _as_ functional.^[Again, think of the stored-program computer as an example. Things work when we understand what the function at stake is, something which is only understood in retrospect: through expectation.] This notion of information is inextricably tied to the analogy^[Perhaps an _undulating isomorphism_, rather, like a reflection in moving water. Hui: “Simondon’s analogical method (from technical to social) is not simply metaphor; indeed it would be wrong to see the relation between the technical and the social as purely analogical; instead, it is what Simondon calls ‘allagmatic’. The word _allagma_ in Greek means change or exchange (synallagma is the bringing of two parties together under a contract, from the Greek _Synallagmatikos_ that in turn comes from _symallattein_ which means to bring together, to unify) (Jakub Zdebik, Allagmatic. _The Semiotic Review of Books_, Volume 17.2 2007.)” (Hui 2015, p. 82).] made by Simondon between the triode and social organization: just like a triode’s functional potential is structured for the deployment and amplification of a signal, so too are the dynamics of psychosociopoliticolinguistic formations (ibid., p. 81). Both have action-potentials (in our context: perspectival attention-potentials) and both can be modulated and amplified in different ways.^[This is, in short, part of what in the collaborative paper we explore in the formalization of power through active inference methods.] This understanding of information is what leads to our functionalist framing of computation as pervasive, too, in this project: it is not a matter of logic gates and electrical currents _only_, it can be modulated as the understanding of a pervasive _differential_ operation which we are able to transpose, i.e., to move, through formalizations, but it is not restricted to the strictly formal domain alone. It is, if explosively understood, the paradigm which has brought with it endless function-modulation possibilities (rendering the current AI predicaments and beyond). Whatever it is that counts as an ‘end’ state in computation, or as scientifically-acceptable parsimony or _explanation_, changes as technology/society evolves. The ‘simplest’ or most ‘compressed’ thing, might come to ‘mean’ something entirely different in futures where _salience_ is of an entirely different order than what we understand it to be today.^[“A simple formula which describes the universals common to many occurrences is scientifically preferable to the complexity of many descriptions of many occurrences. Thus the quest of science is simplicity of description. The conclusion is that science, thus defined, needs no metaphysics.” (Whitehead 1929, p. 57). However, in our take (see also [[05 Prediction]] and [[04 Concepts as pre-dictions]], active inference naturalizes metaphysics into aspects of predictive traction, compressibility. More on this in [[Computational irreducibility]] and [[Kolmogorov complexity]].] As always, the arrival of “technics and devices in telecommunications ... produces an epistemological shift in our understanding of modes of control.” (ibid., p. 83), and we only have an _owl-of-Minerva_ vantage point, for now. As mentioned elsewhere in this project, since the computer (metaphor) morphs, so should our understanding of computation: in the same way that a novel understanding of “modulation could enable us to develop new models and new ways of thinking about the social, which is in itself a form of resistance to the destructive and restrictive forms of power analysed by Deleuze and many of his followers.” (ibid., p. 91). This is the intention of this piece, through an analysis of power as modulated attention in our informatic-computational regime. &emsp; >Nobody directs attention when there is nothing that he expects to see. ... Abstract speculation has been the salvation of the world—speculation which made systems and then transcended them, speculations which ventured to the furthest limit of abstraction. To set limits to speculation is treason to the future. > >Whitehead 1929, p. 75, p. 79. > >But we are now considering the main difficulty of the speculative Reason, its confrontation with experience. (ibid., p. 81). &emsp; Bernhard Rieder, arguing for a *mechanology* (a Simondonian; technical hermeneutics) of the contemporary algorithmic condition, observes the move from Enlightenment disciplinary prescriptions—e.g., encyclopedic knowledge, “universal” education—turning into surveillance and decision-making in business practices in the 20th century (2020, p. 307). Again, the medium is the functional message: these practices may appear, on the surface, aspirationally _different_, but in fact represent the percolation of function: modulations which prioritize the survival of arbitrary explorations. This is what reveals itself as the Simondonian process of _concretization_, where the more the world is amenable to specific practices (e.g., the encyclopedia, the business model), the easier it becomes for that model to continue effectuating itself, retroactively stating its conditions of possibility. In our current scripting context, the -dividual “increasingly revolves around infrastructures built from the ground on up atomized units and decomposed data granules. ... [e.g.,] a series of user profiles that connect through friendship relations and communicate through messages already defines and prepares the raw material available for algorithmic information ordering to latch onto.” (ibid., p. 312) Reducing phenomena to -dividualist atoms and linking them is a very easy way to script the world, just like the myth of the individual was for disciplinary societies, and possibly before. This new type of modulating formalization, instead of having phenomena placed “into an a priori system of order, ... are kept in an intermediate state: already captured and transformed, but sufficiently mobile to support dynamic and multifaceted forms of **reassembly**, operations on differences and similarities that become descriptions, groupings, and attributions that can then be pushed back into the domains they cover.” (ibid., p. 316, our emphasis). This is, of course, the kind of modulation that Deleuze saw as pernicious. In our understanding of system evolution, in broad strokes, this simply demonstrates how function functions, how in order to _be_, it needs to _do_ (cue active inference). Too much ambiguity can be generative, but is often rather disorienting,^[This is often what leads to questioning the legitimacy at hand certain activities, such as highly abstract mathematics or art, which sometimes become subjected to utilitarian lenses: What is it good for? What can I do with this? What does it imply? How does it _change_ reality?] and as far as complex systems have been observed: it is in this way that they flexibly adapt to contingency, allowing _functions_ pass through them.^[“[The Greek] discovery of mathematics and of logic introduced method into speculation. Reason was now armed with an objective test and with a method of progress. In this way Reason was freed from its sole dependence on mystic vision and fanciful suggestion. **Its method of evolution was derived from itself.** It ceased to produce a mere series of detached judgments. It produced systems instead of inspirations.” (Whitehead 1929, pp. 43-4, our emphasis).] &emsp; >When is a piece of matter said to be alive?” [Schrödinger] asks, and answers: “When it goes on **‘doing something,’ moving, exchanging material with its environment**.” What interests me here is that this picture of an intricately articulated structure, a form that maintains a disequilibrium or lifetime – whatever it means to the biophysicist – to the poet means that life is by its nature orderly… > >We quote Moten and Harney (2021, p. 165), who quote Robert Duncan (“Towards an Open Universe”, p. 12) who quotes Schrödinger (_What is Life?_, p. 69). &emsp; Referring to the epistemic flexibility afforded by knowledge-ordering database technologies, Rieder also observes that: “The technique is more inclusive than the conceptual horizon that drove its invention.” (2020, p. 310). This is true of all tools at hand, and to further modulate them towards novel domains of adaptive flexibility, their formal methods need to be understood as always containing excessive possibilities. “Techniques and experiences associated with computing trickle into the ‘view, attention, and language’ of our time.” (ibid., p. 317). In the emergent landscape of AI, the imitation of nature—from neurons to neural networks and back—becomes the absolutely unpredictable sublation _and_ production of something entirely different. Once a technology is rendered, only _after_ its possibilities are discovered, this is what allows for the evolution of function, and for its degenerate transfer.^[Bergson and Macherey both noted how metaphysics is always behind us, see: [[06 Principle of Sufficient Interest]] for full citations.] But the logic underlying this is treacherous, and politically dangerous to state.^[As mentioned with Dark Deleuze, we might fall into oppressive relativisms.] We cite Rieder on this in [[06 Principle of Sufficient Interest]], too: “[w]e no longer decide based on what we know; we know **based on the decision we want to make**.” (ibid., p. 256). We move from the idea of top-down control, to modulatory script. Hence: _post-control script societies_. This renders the ever-sedimenting meta-functioning capacity of function itself,^[The etymology of _function_ is noteworthy, too (etymonline.com accessed 11 March 2024): from the 1530s, as “proper work or purpose; power of acting in a specific proper way”, from the Latin _functionem_ (nominative _functio_): “a performance, an execution”. And originally from the proto-indo-euro: _bhung_- “be of use, be used” (which is also in Sanskrit _bhunjate_ “to benefit, make benefit, atone,” and Armenian _bowcanem_ “to feed,” or Old Irish _bongaid_ “to break, harvest”. According to etymonline.com, this is perhaps related to root _bhrug_: “to enjoy.”] as often remarked throughout this project with regard to metaphor and analogy, as something clearly observable in the mechanics of language: the chunking and parsing; scripting functions we are after. In evolutionary systems complete specification of _anything_ is rigid, making adaptation impossible; while on the other hand too many degrees of freedom prevent coherence, not affording flexibility for anything to _become_ a system with attracting states. “‘Doing something,’ moving,” (Schrödinger, ibid.), _work_, function, is the continuation of an activity which leaves substrate _and_ identity behind. Function is the very thing that passes right through rigid specification and towards endless possibility. Language maintains a critical balance between these poles, revealing the evolution of its functions. What ensues from technosocial infrastructures where _we know based on the decision we want to make_, is the refinement of function under different processes of specification; concretization. From our current perspective, it feels challenging to say precisely _what_ it is that this observation about function wants, other than to go on doing something, as Schrödinger already observed in the 40s. &emsp; >Where, for Aristotle, the identity of essence entails that actualization is substantialization, i.e., the consummation of potentiality and the exhaustion of essential possibility, for Hegel, the self-estrangement of essence deformalizes substance and the essential contradictoriness of the actual turns actualization into de-substantialization. The subordination of becoming to substantial form is undone and the possible and impossible are desegregated. Only what has become can be retrospectively considered essential. And what has become essential retroactively determines what will be possible. Every becoming re-establishes the limit between the possible and the impossible as a division set to be undone by the practical actualization of the essential difference that underlies it. > >Brassier, 2019, p. 102. &emsp; ### Ctrl + C (_Copy_) ### Imitation through scripts; the habits of sociality &emsp; >It is not merely that in the earlier times the men knew less. They were intrinsically less able in moving about among general ideas. ... All things work between limits. This law applies even to the speculative Reason. The understanding of a civilization is the understanding of its limits. > >Whitehead 1929, pp. 48-9. &emsp; Social scripts are uncertainty-minimizing structures that enable agents to navigate otherwise untenably complex landscapes, presented with too many possibilities. Scripts are cultural **constraints**: they represent both internalized cognitive schemas and externalized social orders that structure behavioral possibilities within cultural contexts (Albarracin et al., 2021), where the inseparable bidirectionality of internal and external phenomena can be understood as _habitus_ (Bourdieu 2018 (1995)). Scripts maintain causal and spatiotemporal relationships between events—revealing the persistence of functions that (seek) self-evidence—enabling agents to make reliable inferences about their (social) context. Scripts function as clusters of conceptually related possibilities that help agents interpret situations and identify possible actions, where already-encultured attractors lead to expected outcomes, which change due to combinatorial contingencies of many kinds (systems interacting with systems with systems with systems, etc.). In the spirit of making active inference more socioculturally attuned, Albarracin et al. (2021) suggest a formal notion of a _script_, in order to elucidate how these distributed reasoning processes allow for less computational duress on any single agent needing to compute everything “from scratch” all the time. Culture is the very scaffolding that allows for the reduction of computation at the level of the individual. The more creatures are able to offload computation onto each other and the environment, the less energy they need spend on processing ever-renewed strategies. In the context of our research, the concept of semantic attractor is very similar to the concept of the script.^[Other notions, such as action schemas, for example by [Goddard and Wierzbicka, 2004](https://www.frontiersin.org/articles/10.3389/fpsyg.2021.585493/full#B36) are mentioned in Albarracin et al. 2021 as comparable frameworks.] What in the context of this thesis we often term semantic attractors are how meaning is framed through action-potentials such as affordances, social norms, contextual cues, etc., differences that make a difference, saliences, information that is processed in order to move elsewhere, where _directionality_ by way of _perspective_ is always the case. In their account, and in ours, all involve the effects of _active ignorance_, given that the computational payoff of the implementation of any script implies the neglect of alternatives. Active inference is thus combined with the notion of script in order to develop a computational model of scripts as internal schemata and as external social scaffolding. (Albarracin et al. 2021). What we continually question throughout this project is the analytical distinction between internal and external: language (or habitus, or sociality, culture, etc.) is a system which cannot be delineated with precision (_where_ is a word/concept, exactly?). Because of the imperative to minimize uncertainty we seem driven to simplify complex phenomena down to _chunkable_ tokens, but we actively ignore their possible transformations and functional expansion. Processes of _active ignorance_^[See: [[08 Active ignorance]].] are not only observable as what is commonly known as cognitive bias (leading to, e.g., partiality or neglect) but can also be considered as scripts. Industrial scripts such as the manufacturing of products that produce harmful waste substances as a side-effect, which is actively ignored because the focus is on the _product_, resulting in active neglect of the side-effect, is good example of a process of active _ignorance_. More innocuous instances are examples such as sociocultural imitation for the sake of computational ease (see, e.g., Tomasello 2016, treated in [[06 Principle of Sufficient Interest]]): if I find myself in novel cultural situation and see others engaging in a behavior I don’t have enough knowledge-context for, I might as well imitate them to reduce my chances of drastic (sociocultural) uncertainty. In this case I actively ignore the many possibilities for action afforded by an environment (Felin and Kauffman 2019), and reduce my uncertainty by following a script I observe others engaging in. Scripts help maintain power structures by processes of active ignorance; pooling attention in specific ways. Scripts normalize certain behavioral patterns by providing templates for “appropriate” action, making deviations appear abnormal or inappropriate (Constant et al., 2019). This normalization often reinforces _existing_ power relations by making them appear natural and inevitable. Scripts establish sequences of interaction in favor of a current norm: whether agents following this are (critical) aware of the norm, or simply perform it, is often the question. These patterns become self-reinforcing as agents learn to expect and reproduce them. Scripts under AIF can be understood as shared generative models that enable coordinated social behavior; as habitual knowledge about the causal structure of social situations, and allowing agents to make predictions about others’ behavior and select actions. Scripts are transmitted across generations through cultural learning processes, enabling the reproduction of social structures, revealing how functions are passed on, and this transmission occurs through both explicit instruction and implicit learning from environmental regularities. The role of narratives in power dynamics, as scripts which normalize behavioral patterns by providing templates for “appropriate” action, may make “deviant” resulting in social castigation or neglect, correcting and evolving aspects of functional refinement. Because of the challenges inherent to the coordination of social reality—as mentioned: whether agents following a script are (metacritically) aware of the implied norm, or simply perform it, being one of the main questions^[Which we explore in [[06 Principle of Sufficient Interest]].]—we can also think about scripting narratives in terms of self-fulfilling prophecies: none of the narrative options on social offer need be realistic, so long as they are strong enough to attract attention. We explore an example of this later, but for now it can be stated that abstract semantic scripts, which can be understood as constraint-regimes coming to operate as distributed generative models, create situations where agents can learn by “internally” simulating—even at the subconscious level, at the level of proprioception or basal concepts such as minimally intuited social trust—responses to high-stakes scenarios, while maintaining corporeal integrity and safety, all the while plastically exploring epistemic conditions. An environment pooling attention to strong-enough narratives will elicit the scripting of those narratives (regardless of their ‘veridical’ content: evolution does not seem to care for truth (Hoffman 2019). For agents under thermodynamic duress (socioeconomic, corporeal, etc.): the simpler the narrative, often, the better (think political campaign slogans). This being for the banal reason that they are under duress, and therefore incapable of sustained projective computations. Narrative structures can be understood, in their scripting modulation, as semantic attractors which organize _possible_ world-states. These patterns function as possibility landscapes for belief-updating, _allowing_ agents to—or *restricting* them from—modulating their internal generative models. Considering homo sapiens’ universal tendency towards narrative-sharing, this process can be understood as a corollary of actively-inferential social agents, particularly if we understand ‘control’ as the modulation of attentional variables resulting in behaviors that shape the material production of the future. The theoretical framework developed here has significant implications for understanding both the cognitive foundations of narrative engagement and how _the power of attention to these_ results in the scripting of communal behavior. Again, always, for better or for worse: communally adaptive towards social cohesion and love, or competition; mutual suspicion; hatred; distrust, etc. In general, in communal organization, powerful entities offer perceived stability and predictability. In AIF terms, they help other agents minimize surprise by providing _apparently_ reliable scripts for prediction, essentially by dictating the information geometry landscape (Constant et al., 2019). Power results in the offloading of computational demands onto others or their environment, allowing for more efficient processing of complex information. Think about house labor (Federici), think about hegemonic structures, think about biases of all kinds. Whatever is perceived as the norm, directs attention and behavior. Additionally, powerful entities can better absorb or reinterpret frictions (prediction errors) due to their broader resource base and influence over narrative construction. Through their ability to shape both physical and informational environments, powerful agents can actively reduce uncertainty for themselves while potentially, and this is often the case: increasing it for others. Powerful agents can actively reduce uncertainty for themselves while potentially, and this is often the case, increasing it for others: there is no such thing as a free lunch in many ways, particularly in the cases when somebody _made_ it, and cleans up after others have eaten. The manipulation of the unfolding of the information geometry—both in terms of perception and action—is central to how power operates in social systems. Powerful entities literally shape the attentional landscapes by controlling how perspectives land on saliences, which _become_ salient through the _habitus_ of scripts. If house labor is not labor, **then it is not labor**. As we conceptualized it—in the article this primarily builds upon—power represents the capacity to shape and control the (semantic) “attractors” within an informational landscape. It enables agents to influence *which* states become salient, thereby controlling the flow of attention and information, shaping beliefs, and modulating the precision of predictions. Power, as a function which agents often seek, shapes expected outcomes by adjusting the informational environment in a way that amplifies certain states while suppressing others. Power flows through scripts that reduce the dimensionality of the state space by establishing strong, predictable attractors—such as social norms/strong narratives. By doing so, power is that which effectively directs attention and ensures agents align their actions with salient, scripted-scripting states. Apologies, of course, as the applications of this (the formal integration of which is not presented here) inevitably begin at reductive formalizations (the math is never the territory, see: Andrews, 2021 on the map/territory intricacies of active inference). However, the implications of these arguments for complex social dynamics in the realms of, e.g., policy and governance are vast. Coordination dynamics need to be framed in terms of formalizing their *possibilities* for novel understandings of social change processes. How should humanity—as one, as many—develop strategies with enough “conservative” coherence **and** plasticity so as to resist the oppressive and dominant conservative effects which do not benefit the coordination of the whole?^[As noted by David Hyland in our collective paper, a schematic starting point for the analysis of such multi-agent interactions “may be from the point of view of game theory. In particular, the concept of Free-Energy Equilibria (FEE) captures the notion of a joint policy in which all agents in a given interaction are minimizing their (Expected) Free Energy (Hyland, 2024). Under certain conditions, it may be expected that interactions between agents tend to drive _the system_ toward FEE. However, as is well understood in game theory, there are typically **many** equilibria of a game and depending on the starting point and interaction/learning dynamics, the system may converge to undesirable equilibria where a handful of agents exert significant control over others to their detriment. An important area of future study is thus to first establish a better understanding of what makes different equilibria societally desirable, and then to understand how (incentive, semantic, etc.) affordance landscapes may be shaped to navigate society toward desirable equilibria.” (Hyland, 2025, my emphasis in bold). Shared narratives, where agents are in the know about what is shared, are fundamental here, and this begins by paying attention to the (possible) rules in place, hence the need for formal methods. This is part of what critical metacognition entails: knowing how we are collectively navigating uncertainty, and taking account of already-given asymmetries. In the conclusion below, we explore a few salient implications of our arguments.] If we take a wide-eyed view and scale such processes to the species-evolution level, which as we insist: is revealing of the evolution of functions and not much else,we may even dare to observe-script that processes such as our current globalized neoliberal landscape rely on a multitude of maladaptive _active ignorance scripts_: they silo possibilities into entrenchments that suffocate possible lives. In light of environmental degradation and systemic oppression aided by the mechanisms sustained by power pooled through the accumulation of capital, we dare to speculate so: the destabilization of the socioecological environment that supports the types of living processes we seem to be interested in (as one of the main symptoms, but certainly not the only one) is evidence to the fact that these scripts **do not reduce uncertainty on the long-term**, by far. Unfortunately, the main system guiding technoscience, sociality and chances of diverse ecological survival, is geared towards power-accumulation through attentional-pooling, in our globalized capitalist scheme. This might be, we dare to speculate again, because the transfer of _function_ is enabled with incredible efficiency through the abstraction that is something like _money_, a metaphysical-real abstractor which may come in to stand for X, for the possibility of _anything_, which is, always, a carrot: *desire*. Similarly, these observations can be made about many other metaphysical tendencies, including the concept of artificial intelligence, as is argued throughout this project. What keeps these abstractions going is the promise of function, through the active exploration-exploitation of contingency, in every sense of the exploit-explore trade-off, here. &emsp; ## Ctrl + X (_Cut_) ### Power as the attraction of attention &emsp; >In his essay ‘The Limits of Control’, Burroughs addresses how language as image is the principal mechanism of control: ‘words are still the primary instrument of control. Suggestions are words. Persuasions are words. Orders are words. No control machine so far devised can operate without words’, he writes, ‘and any control machine which attempts to do so relying entirely on external force or entirely on physical control of the mind will soon encounter the limits of control’ (Burroughs 1978: 38). That limit, for Burroughs, is the paradox at the heart of control: that as it seeks to overcome, it needs resistance, for without resistance control ceases, supplanted by mere automated use function: ‘control needs opposition’ ([ibid.]). > >Gontarski 2020, p. 555. &emsp; How are power and attention explicitly modulated, or alienated, in language structures? One obvious take in this project would be to begin at the functional level of the script(s) of metaphor(s). Gravitational metaphors, for example, abound in language (Shepard, 1987; Lakoff, 2003; Dennett, 1992; de Jager, 2023; Kent, 2024), and they can be understood as scripts which attract our (semantic) attention, by exploring a very pervasive function or constraint: gravitation itself. We explore many aspects of this in [[10 Bias, or Falling into Place]]. Gravity not only constrains physical proprioception and action, but as a metaphor it has also been presented as that which sustains our very sense of a _narrative_ self (Dennett, 1992). Lachlan Kent has further argued that the vestibular system, which regulates graviception, also affects the higher-order consciousness of selfhood (Kent, 2024). The links between the habits of the physical and the metaphorical may represent strongly-related underlying (cognitive) patterns, which might give us hints about the transcoding of function across substrates. Somewhere between formal and metaphorical analysis, the idea of power being _pooled_ towards complex social _attractors_ was famously presented by Pierre Bourdieu (2018 (1995)), who linked various concepts of capital—symbolic, cultural, political, etc.—to power as social (at)traction, borrowing the metaphor from physical field theory. The metaphor would be that just as electromagnetic fields exert forces on particles depending on their position and properties, to Bourdieu we could conceptualize social fields as spaces where agents’ positions are (pre)determined by their access to or contact with different forms of capital. Bourdieu’s use of field theory explicitly frames system function and the gravitational dynamics of _habitus_.^[The concept of _habitus_ was Bourdieu’s attempt at eliminating the subject/collective dichotomy by presenting a 5E model of social habits, where they are dispersed and ecologically modulated. He did not use 5E theories, we ascribe that quality to it here.] A given field structures possible actions by defining what counts as accepted or legitimate, and _habitus_ as inclinations; dispositions, determine how fields are navigated: canalizing functional possibilities. Bourdieu’s intention was to harmonize the ontoepistemological and the empirical, as he believed these to be inseparable. His desire was also to ground sociological analyses of _habitus_ in a scientifically tractable manner.^[In the collaborative article we elaborate on the formalities, by framing these attentional attractors that move systems toward specific configurations, through AIF.] Grace Blakeley has analyzed the disempowerment of collective political attention by presenting this as the _disillusion_^[In our terms, perhaps, understood as the loss of attentional momentum.] with habitual narratives: democratic engagement in the context of, e.g., British politics, has been affected by a sense of lack of ability to affect conditions. We may speculate this emerges from scripts such as “there is no alternative”, all the way to the presentation of a field of agential possibilities as a “free” market. As Blakeley notes: this sense of inability to act is grounded in deep power disparities in capitalist societies, which are, primarily: completely invisible, as we noted Deleuze diagnosed. As observed earlier, this is a double bind where, while most agents are structurally denied ‘control’ over their lives, they are at the same time channeled by scripts that tell them they are free to choose how to live. “Life under capitalism means life under a system in which decisions about how we work, how we live, and what we buy have already been made by someone else. Life under capitalism means living in a planned economy, while being told that you are free.” (Blakeley, 2024). The alienation is not just spectral, it is in your wallet. The resulting political disenchantment was also framed earlier by Mark Fisher in terms of the privatization or individualization of systemic stress (Fisher, 2009): through strategic narratives which pool public attention towards scripts which foreground aspects of responsibility at the level of the individual, agents become overwhelmed by having to resolve, at the individual level, stress which is in fact _systemic_. Or, as Amanda Beech reminds others have reminded us (2021, p. 51): >Theorists including Adorno, Foucault, and Lyotard are well known for reminding us of the danger that a commitment to reason slides into dogmatism. Reason not only brings the problem of dominance from external forces, but in the context of capital, power is legitimized internally, through the subject’s self-oppression, where myths of freedom eliminate the possibility of getting behind the back to power to put it into question. Capital obfuscates the reason that propels it, whilst being indebted to systems of measure and order. Whether we are talking about a metastabilizing “invisible hand” or modeling swarm-organization (Wareham, 2023): complex systems face scaling challenges where control mechanisms must maintain sufficient internal complexity to *match* what they control. This can be understood at a number of levels. At the level of social dynamics and the habits of power asymmetries, this is strikingly clear in the context of the alienating, transforming power of attention. If a specific script which simplifies the high complexity of reality gains attentional traction as a collective belief, it will result in the simplification (and mobilization) of desires, enabling the transfer of particular functions. Empowered agents with *superior* information access employ scripts strategically while maintaining meta-awareness (i.e., requisite variety) of their simplifications, whereas publics often adopt oversimplified scripts that may not serve their interests—such as unrealistic appeals to nostalgia, or believing a single solution like deportation can resolve all socioeconomic problems. “Make X great again” is a perfect example of the transfer of function by accessing a specific habitus, which relies on the common perils of a retroactive mode of reason that can only create the path immediately under its feet: “make” (i.e.; set in motion, work, transfer function), “X” (whatever is currently salient, but always fungible), “great _again_” (appealing to the way memory is able to retain relevant functions from the past, as it continues to self-evidence that it is itself relevant, because it has survived thus far). The phrase invokes a selective past “greatness,” it reasons backward to move forward, creating and/or relying on its own self-evidencing. It reveals, to us, a circular justification where the rhetorical structure transfers function in a way that gains momentum because it feels both motivating and self-justifying to its audience. This way, publics are mobilized in order to continue to move towards something that never was, and where each agent inhabits a differently instantiated supposed nostalgia. &emsp; >“Economic rationality,” Foucault reminds us, “is founded on the unknowability of the totality of the process,” representing an “essential incompatibility” with the “totalizing unity of the juridical sovereign.” The necessary unknowability of the future—so central to profit, speculation, Adam Smith’s invisible hand, the figure of Homo economicus—appears as anathema to the sovereign founder of law, right, and the monopoly of legitimate violence. And yet at the level of knowledge and technique, as Foucault has it, “the principle of the necessary freedom of economic agents can coincide with a sovereign.” There can be “perfect correspondence between the sovereign and the economic processes,” such that economy offers to sovereignty a form of analysis of “what is taking place” in a world of free transactions and movements. > >Amoore 2013, p. 5. &emsp; Again, different historical orders of domination and discipline (Foucault) gave way to (internalized) modulatory control (Deleuze). Exerting social control over others’ actions begins at the level of attention (Tomasello 2016): whether this be by attracting attention (“follow me!”), or directing attention (“look at that!”). Our interest in formalizing certain aspects of power dynamics is directing attention towards (metacognitive and collective) scripts with the capacity of empowering responses to counter dominant strategies which, currently, minimize uncertainty for the very few, often by allocating too much importance to the atomized individual, which is a predictably convenient way to simplify the actions of complex distributed processes such as individuals, into overseeable patterns. Revealing these dynamics offers insights into *how* shared structures of meaning can exist as shared metacognitive computing conditions, i.e., conditions in which agents are—relatively—aware of the implications of the scripts they follow, or coercive domination: where agents are subsumed under a script which does not benefit their uncertainty-minimization on the long run. If we understand successful narratives as those which exert the most effect^[Often “fast and frugal” (Gigerenzer, 2008; Simon, 1978).] heuristics and narratives which are low-risk—and, in most political contexts, imply promises or solutions impossible to accomplish: “you have been lied to”, “the other is the problem”, etc.—then analyzing them in terms of attention and function seems logical. In the context of uncertainty-minimization, simple narratives have the capacity to dominate vast regimes of attention (economies), as they are scripts which minimize the complexities of life into low-computation, predictable visions.^[Whether these are dialogical and representational/symbolic or surreptitious and encrypted: we’re dealing with different strategies, but similar results.] &emsp; >For the spiritual servants of order, “the yes-men that labor for the majority,”^[Citation in Viveiros de Castro 2014 (2009) is from Pignarre and Stengers 2011, pp. 31-35.] this change foremost represents something from which future generations ought to have been and still must be protected—the guardians of today having been the proteges of yesterday and _vice versa_ (and so on)—so as to reinforce the conviction that the event of ’68 was consumed without being consummated. By which they mean that nothing actually happened. The real revolution supposedly happened contra that event, and “Reason,” to employ the usual euphemism, was what delivered it; the reason-power that consolidated the planetary machine of Empire, in which the mystical nuptials of Capital and the Earth—globalization—climaxed, and that saw itself coroneted by the glorious emanation of that Noosphere more commonly known as the information economy. Even if capital does not always act with reason, one nonetheless gets the impression that reason always delights in letting itself be roughly taken by capital. > >Viveiros de Castro, _Cannibal Metaphysics_ (2009) 2014, p. 97. &emsp; Remaining politically-agnostic in any analysis of social habits presents us with some questions: what can we assume to actually know about each other when we act collectively? The concept of a ‘self’ as a single agent seems particularly questionable, as we have been observing, especially as all selves exist within a regime of social habits which exert effects on their supposed autonomy (Bourdieu). In terms of information-processing capacities, not only are resource-distribution and the division of labor a social _fact_, but also: as complex possibilities cannot be computed at the level of the individual, this reveals how and why the minimization of uncertainty for social systems is, in fact, social. Function transfers across agents through policies which ensure information-processing and future-projection can be distributed (Vasil et al., 2020, Friston et al., 2020), as this allows for more efficient practical traction because of “epistemic confidence in knowing, interpreting, and acting together” (Shteynberg, 2023). This is the basis for any division of labor, where _roles_ develop on the basis of contingent substrates, towards refinement and transfer. This presentation is in contrast with scripts promoting individualized agency—which is a prerequisite for the kinds of social coercion we just outlined—where one’s estimation of “divergence in mental states between self and other(s)” allows for the “prediction and interpretation of others’ behavior” (Shteynberg, 2023), resulting in diminished possibilistic effects. Again, here we can see why Deleuze replaces the image of, e.g., the factory worker as a subject of disciplinary capitalism, with the image of a distributed subjectivity, e.g., the owner of a credit card as a subject of libertarian debt, occupying various spatiotemporal scales and modulated not by clock-time or physical enclosure, but by *probabilities* and contractual limitations (Deleuze, 1992 (1990)). From enclosures to parameters: the ruling hand became the border became the passport became the password became the voice-police inside your head (or: the voice police always was, as treated in [[06 Principle of Sufficient Interest]]). Whether as visible discipline or as modulatory control, the overall effect remains the same: in the vast complexity of competing power hierarchies, the pushing and pulling by the strongest—i.e., those currently controlling attentional-informatic regimes—results in the subjugation of the weakest. “Weakest” not in the sense of _inferior_ nor meek, but simply in the sense that the weak are dominated (willing or unwilling, aware or unaware). It is important to mention that we are not after an _us-versus-them_ narrative—especially given the attentional primacy we wish to give to pure function—but that we take all historical, future-forwarding forces—nations, people, fictions, industries, ourselves included—as composed of complex, highly -dividuated, contradictory and inexhaustive perspectives. In the context of a _perspective_ which narrates from here, now, it seems that the enclosing of the “self” into a thing with agency and potential is precisely what blurs the capacity of complex, modulatable distributions (sociality, togetherness) to predict something larger and richer than just a simple “self”. The self itself is a technology,^[As presented in [[10 Bias, or Falling into Place]].] and all phenomena, concepts, things become _predictable_ if they are simplified, entrenched: applied and therefore brittle. Soaking in complexity is but a mere meditative metaphysical exercise, it quite literally leads to _nothing_.^[As treated in [[04 Concepts as pre-dictions]].] By technically offloading predictions (a blender, an atomic scanner, a book) onto the environment we allow them to become, in fact, predictive: >Even if technical objects are ‘invented, thought and willed, taken up [assumé] by a human subject’ (Simondon, 2017, p. 252), they depend on the objective technicities available at a particular stage of technical evolution. ... If technical creation is indeed a form of communication as Simondon suggests, the ‘message’ a technical object carries is the operational substance it introduces into the networks that surround it, adding to and transforming them in the process. Mechanology is the hermeneutics of that message. > >Rieder 2020, p. 306. Our mechanology here is **attention** to _attention modulation_, which we see as resulting in contemporary forms of informatic power. If any message should be taken from this chapter: please stop giving the power of your attention to structures which reinforce oppressive powers (as noted ad nauseam: _all press is good press_). In the context of semantic explorations such as fictions or unreasonable proposals such as we saw a few paragraphs ago, tentative scripts can be understood as semantic attractors enabling—relatively—risk-free learning.^[This point is expanded upon in a forthcoming article treating fiction under the AIF, by S. de Jager.] Possibilistic cognitive explorations in the form of, e.g., fictional narratives allow agents to consider counterfactual scenarios and minimize uncertainty without exposure to the risks of “real-world” consequences. This is why art matters: it is capable of allowing us to explore contingence from a safe window. Metaphysics is safe physics: not-yet-engaged scripts can be understood as the speculations, fictions, etc., we entertain, which allows to frame _metaphysics_ as a function of prediction, as mentioned earlier. The absolute _is_ lazy (as we cited in the previous chapter) because the absolute is, in fact, the path of least resistance: tautological, parsimonious, self-evidencing. _Everything is everything_ is a dead end but also a very, very complete explanation. This mechanism frames both the evolutionary advantage and persistent appeal of narrative engagement in human culture. Again, framing this in our current political context and the leveraging of attentional power: setting forth “strong narratives”—such as “us versus them” reductionisms—greatly minimizes uncertainty about an agent’s or group’s current predicaments, and thus become appealing (and often self-fulfilling) policies that bring about their effects, resulting in collective behaviors which can be beneficial or detrimental to the greater whole (Albarracin et al., 2022). These effects of uncertainty-minimization can be observed at the level of the basic function-transfers of language: attractors such as redundancy, rhyme and alliteration, can be understood as patterns which aid memory, communication and information-retrieval policies: they permanently reappear, pooling attention to the “same” phenomena, thus greatly reducing the need to decode and recode linguistic interaction anew. By way of the same uncertainty-minimizing mechanisms, (simple or complex) narrative scripts create cognitive situations where agents can reduce complex aspects of the world down to preferred or already-identified scenarios. Unlike “direct” experience, where prediction errors present immediate risks, narratives provide speculative semantic spaces for exploring models against hypothetical outcomes. With regard to social power, the problems immediately emerge in situations where collective attention controlled by agents already benefiting from dominant narratives can lead to, e.g., the exacerbation of misinformation, echo chambers, conspiracy theories, confirmation bias, etc. (ibid.), in an attempt to retain attention and therefore power. In political communication and organization, where the proliferation of scripts can be understood as an optimization of the learning/risk trade-off inherent in AIF, this is of crucial attention, particularly at a time when the democratic, technological and futurological organization of life is in question. Which functions percolate, how, and through what? The theoretical framework developed in the collaborative paper through formal methods but not presented here, reveals implications for understanding both the cognitive foundations of script engagement in the mediation and evolution of power, as well as strategies for their possible modulations. Formalizing and modeling these mechanisms demands special attention in the context of futurological prospects, as we saw (through Rieder and Hui): there seems to be a need to engage the technical substrate to reveal concretization and modulate it. &emsp; ### Ctrl + C (_Copy_) ### A copy of a copy of a copy of a copy...: a _function_ &emsp; >Programming languages have eroded the monopoly of ordinary language and grown into a new hierarchy of their own. This postmodern Tower of Babel reaches from simple operation codes whose linguistic extension is still a hardware configuration, passing through an assembler whose extension is this very opcode, up to high-level programming languages whose extension is that very assembler. In consequence, far-reaching chains of self-similarities in the sense defined by fractal theory organize the software as well as the hardware of every writing. What remains a problem is only recognizing these layers which, like modem media technologies in general, have been explicitly conceived to evade perception. We simply do not know what our writing does. > >Kittler 1995, p. 174. &emsp; In his 2020 article on Deleuze’s Postscript, James Brusseau updates Deleuze’s definition of control (“social organisation without spatial divisions and explicit prohibitions”), to show how it has developed from an abstract concept to something “localisable as specific technologies functioning where personal information is gathered into contemporary data commerce”. If the most pressing question Deleuze asks is “how can there be control if nothing is forbidden?” (p. 2), Brusseau notes, the answer is that predictive analytics has proven so efficient as to regulate via incentives, then this is just a Heideggerian Deleuze critiquing the cybernetic end of philosophy. However, as we explore in [[06 Principle of Sufficient Interest]], funneling attention through scripts has been the social discipline controlling divergent desires since their primordial origin as cultural habits, linguistic traditions, rules of law, etc.: normalization via in/out mechanisms. The diagnosis of a normalizing or _flattening_ of the landscape today is what is constantly explored by, e.g., critical theory. What is relevant to note, as a change, is that the novelty of today’s societies of (digital) control is how we have unlocked the possibilistic potentials of highly divergent _interests_. Noise has become the functional norm, rather than that which a normalizing system seeks to suppress (Wilkins 2023). Whereas televised funneling worked in batches and broad strokes, personalized microtargetting has become the new standard in digital advertising: where the “coercive potential is unlimited: you _always_ may be deserving, there is _always_ something that may be needed.” (Brusseau 2020, p. 16). As Brusseau also explains “[c]onsumers are not contained, blocked, or forced, but they are lured and directed by marketing appeals exploiting personal information. ... The struggles will lead somewhere between the transparency belonging to those who embrace the joys of marketing, and the transience of those cutting away from their own identity.” (ibid., p. 2, p. 21). But what is the supposed “freedom” _from_ this that both Deleuze and Brusseau imagine in the background? As Rieder (2020) notes, our theoretical effort should direct itself at distilling “an operational epistemology from the cacophony of techniques ... [a] set of regularities that characterize how we understand and operationalize the very notion of order at a given time and place.” (Rieder 2020, p. 19). This is what we understand as a task: the close inspection of the refinement of function. For better or for worse. But without the idea that there _used to be_ freedom, and now there is not. There are changes in degrees of freedom, and higher complexity inevitably means more degrees of freedom. As just mentioned, this is precisely what late capitalism is known to feed on: new possibilities via disruption, high frequency trading, razzle dazzle strategies, the co-opting of resistances, engulfing invariances, etc. Deleuze shows how, precisely, from disciplinary to modulatory control: what can be observed is a realization that probabilities represent modulatable tendencies and never absolutes. The effects of (consumer) attention are not entirely predictable, it is better to create functions consumers can explore, rather than enclosings into which they should narrowly fit. This is how our enculturation systems went from doing things like cuing behavior; _telling_ agents how to interact with the manifold of systemic structure, to actually making the manifold more flexible (e.g., language becoming generative rather than merely indexical, or Google feeding off of your searches to become better at what to sell you). In this way, the evolution of function is able to cares less and less about substrate, and more and more about its own exploration across substrates. Brusseau’s claim is that today’s control society—more than thirty years after Deleuze’s original diagnosis—“coerces without prohibitions,” and functions through nudgings which are not interpreted by its subjects as discipline, since the controlling incentives are sold as desirable. In terms of everything being attentional scripts, in this new landscape, the _apparent disappearance_ of friction; the presentation of _immediacy_ (Kornbluh 2024); the interpassive displacement of desire (Pfaller 2014) and enjoyment across the field of habitus, functionally reveals something that we can compare to the scripting effects in how phenomena become concepts. In a functionally comparable way, analogies ensue from metaphors gradually (Hofstadter & Sander 2013), concepts only crystalize as the substrate changes. That is, initially: a friction, later: a matter of fact. For example, “inter-” and “passivity”: how are these two things supposed to be combined? The appearance of something surprising urges thought to search for an uncertainty-minimizing solution. Once it is “understood”, its salience fades and becomes incorporated into a vocabulary which serves as requisite variety for a complex environment. Its meaning is what allows for its application—in communication: as ‘internal’ or ‘external’ dialogue—serving the function of reducing uncertainty about a wide variety of phenomena. In terms of an attention economy: agents with the capacity to become interested **are made to be interested**. This, in line with Rieder’s Postscript-inspired claim that: “Order is no longer understood as a stable map to the universe of knowledge but increasingly as the outcome of a dynamic and purpose-driven process of ordering.” (Rieder 2020, p. 18). This simulacrum, to us, reveals our condition as being in the meta-game of creating combinatorial complexity serving the evolution of functions. Disciplinary societies explored function through specification, categorization, enclosure. In control societies, functions no longer emerge _from_ materiality, evolution appears not as a process of material adaptation but as functions exploring their possibilistic space through humans and their extended environments. In post-control script societies we realize this, and with the advent of phenomena like LLMs, a new critical jump might reveal itself in retrospect. “The developer is right in-between, surrounded by technicity coming in all shapes and forms, and thus ‘among the machines that operate with him’ (Simondon, 2017, p. 18).” (Rieder 2020, p. 16). This is why it is ever-mode interesting to engage in the structures of function modulation, through linguistic means or otherwise. What is significant about the Postscript is that it has become symbolic of the critique of _control_ (not unlike Debord’s _Society of the Spectacle_ (1994 (1967)): a key text on attention). Near the end of his essay Deleuze reflects: “It’s up to [young people] to discover what they’re being made to serve, just as their elders discovered ... the telos of the disciplines” (p. 7). This is the functional key: discipline refined function to the point that it could dissipate across novel terrains. The comment on servitude, as we remark throughout this project, is always revealing of complex but at the same time rather simple and straightforward ideas about domination which “we” inherit from recent, and deep, pasts. Human beings often seek to resolve uncertainty by domination, but also by functions such as love. As a “young” person, I would like to take up Deleuze on what, from my own limited scope, discover what I am being made to serve. What is beyond control (categorization, delimiting, territory, etc.), which is not _un_-control (pedestrian interpretations of chaos, freedom, noise, etc.)? The lack of awareness about existing in a script is supposedly what makes one incapable of interacting with—the possibility of—reading and writing it, not just enacting it. What control, as presented by Deleuze, helps me understand is that there’s nothing behind the veil, there’s just veil. It is function searching for new spaces that drives the symbolic, transcoding, indexical, representational, suspicious, paranoid, hyperstitious to think there is always something behind the veil: the simulacrum itself is true (Baudrillard 1983 (1981)). In another postscript, the postscript to the _Requiem for the Media_, Baudrillard asks: &emsp; >There is a paradox in the inextricable conjunction of the masses and the media: do the media neutralize meaning and produce uninformed [_informe_] or informed [_informée_] masses, or is it the masses who victoriously resist the media by directing or absorbing all the messages that the media produce without responding to them? Sometime ago, in “Requiem for the Media”, I analyzed and condemned the media as an institution of an irreversible model of communication without a response. But today? This absence of a response can no longer be understood at all as a strategy of power, but as a counterstrategy of the masses themselves when they encounter power. What then? Are the mass media on the side of power in the manip­ula­tion of the masses, or are they on the side of the masses in the liq­ui­da­tion of mean­ing, in the vio­lence per­pe­trated on mean­ing, and in fas­ci­na­tion?” _Simulacra and Simulation_, section: “The implosion of meaning in the media. &emsp; He reflects on the condition of simulacra leading to the apparent loss of meaning, meaning being something with directionality (an index), and the possible strategies of resistance against a system which urges one to make demands, statements, speak. “The media carry mean­ing and coun­ter­mean­ing, they manip­u­late in all direc­tions at once, noth­ing can con­trol this process, they are the vehi­cle for the sim­u­la­tion inter­nal to the system and the sim­u­la­tion that destroys the system, accord­ing to an abso­lutely Möbian and cir­cu­lar logic—and it is exactly like this. There is no alter­na­tive to this, no log­i­cal res­o­lu­tion. Only a log­i­cal _exac­er­ba­tion_ and a cat­a­strophic res­o­lu­tion.” (ibid.). We continue to diagnose this as the evolution of function, the loss of substrate, again. “With­out a mes­sage, the medium also falls into the indef­i­nite state char­ac­ter­is­tic of all our great sys­tems of judg­ment and value. A single _model_, whose effi­cacy is imme­di­ate, simul­ta­ne­ously gen­er­ates the mes­sage, the medium, and the “real.” (ibid.). As far as I can see, much like simulacra self-estrange to the point of creating-becoming their own attractors: it is in this way that functions _have_ subjects. Subjects are eating, sleeping and sometimes reproducing creatures: they keep on going. “We are alert, or we are drowsy, or we are excited, or we are contemplative, or we are asleep, or we are dreaming, or we are intently expecting, or we are devoid of any concentrated expectation. Our variety of phases is infinite.“ (Whitehead 1929, p. 82). Being subject to linguistic functions is to _feel_ is that there is something behind the veil,^[I.e., to seek evidence—behind _seeing_ or as in the effects of (paranoid) surveillance: all the same.] propelled by a function’s desire for transcending substrates. In the context of our arguments: the function is that of _perspective_ as it modulates _attention_. The metacognitive attention that is _seeing as_, a frequent target of this thesis, can be understood as an effect of function transitioning from one state to another. In **post-control script societies**, eventually, enabled by recursive meta-reflections which can only be arrived at as systems sediment, accumulate, overlap, i.e., maps on maps on maps on maps: we arrive at new insights into how functions continue computing themselves across all these domains. It seems strange to think that we care about things that will transcur once we are dead, yet we do: function percolates through mortal computers. The functional model of _order_ in information theory, being context-agnostic, is always embedded in meaning/value judgments. But these are transient. And it is in this way, as Baudrillard notes, that information in the context of Shannon can be understood as a “kind of code, like the genetic code: it is what it is, it functions as it does, **meaning is something else that in a sense comes after the fact**” (ibid.) That is because meaning is simply what is relevant in a particular; perspectival; mortal situation, and what gives rise to function.^[This is also noted in [[06 Principle of Sufficient Interest]], or in [[10 Bias, or Falling into Place]] with regard to metaphors.] Much like noticing how a general constraint, such as gravity, operates at many levels,^[As explored in [[10 Bias, or Falling into Place]], highly indebted to Juarrero’s work on constraints (2023).] noticing how _function_ is that which can take analytical primacy over process _and_ identity, allows one to see functions passing through everyday items, ideas, gestures, modes, etc. “If intelligence and matter are decomposed into movements, then we will see that the formation of form is no different from the materialization of matter. It is thus not form alone whose functionality is at the center in such a genesis, but rather the unity of usage and formation.” (Hui, 2019, pp. 148-9). I am aware I am dangerously bending concepts and blending regimes here. Perhaps this is a function’s dark-deleuzian way at the _overcoming of all specific determinacy_, through this text, you and me. “It may be that language _feels_: that it _breathes_.” (Masterman 2005, p. 69). &emsp; >If Marx succeeds in materializing dialectics, it is precisely to the extent that he refrains from positivizing the potentiality he construes as generically human. This is to say that he does not characterize it as a positive essence but as what Simon Skempton calls “undetermined determinability”: “this is a negative and contentless universality; the overcoming of all specific determinacy; thus it is not the universalization of any determination. > >Brassier 2019, p. 103. &emsp; Marx materializes dialectics because dialectics is, if we want to frame it as such, the observation that function _transfers_ across substrates. Something persisting _as a result of change_ (incorporating noise, adapting through mutation), not against it. &emsp; ### Ctrl + P (paste) ### Information _matters_ and matter _informs_ &emsp; >Scientists use apparatus or system, when they can, just as carpenters, when they can, use a power lathe; but **the tool must not be identified with the activity itself, which is one of constructing argument, or with its product, which is the statement of a discovery made**. > >Masterman 2005, p. 58, our emphasis in bold. &emsp; In the transition between an economy centered around material and spatial capital, towards one of temporal, speculative capital, we can rather straightforwardly connect attention to computation by understanding the implicit link between, e.g., risk-assessment and investment: (possible) future gains ought to be calculated, and directing investment begins with attentional gains. The real abstractions of (speculative) capital are abstract, _and real_. To cite Blakeley again: “A sufficiently large financial institution has the power to direct investment into certain technologies and therefore to determine which futures are available and which ones are foreclosed. And all these firms can consolidate their power by buying up competitors, erecting barriers to entry, and crushing workers’ attempts to organize, further insulating them from competitive pressure and giving them significant authority both within their domains and over society as a whole.” (Blakeley, 2024). To combine this with Louise Amoore: “In Frank Knight’s classic text published in 1921, for example, making decisions in the context of an unknowable future is of the essence of entrepreneurial activity: “profit arises out of the inherent, absolute unpredictability of things, out of the sheer, brute fact that the results of human activity cannot be anticipated and then only in so far as even a probability calculation in regard to them is impossible and meaningless.” (Amoore 2013, p. 10). Power seems intimately related to the ability to gamble, to deal with contingency, always _iff_ the perspectival substrate allows for this function to script itself through it. Beyond the financial realm, in the context of futurology and planetary survival, the concept of AI technologies as “powerful” grounds the arguments surrounding possible existential risks. E.g., Cappelen et al. rely on the concept of “power” in order to assess to what extent humans may or may not lose control over their own technological creations (Cappelen et al., 2024).^[Similarly, Ororbia and Friston 2023: “One might view the future as residing in not viewing intelligence as a difference engine but instead as an artificial form of sentience: one that is capable of self-healing and self-repairing with autonomy — its existential imperative being to persist in (generalized) synchrony with its world.” (ibid.). The questions of “autonomy” and “synchrony” are precisely the ones we are after in this article, by observing how power, in mortally-determined social systems, is driven by attention.] What these assessments have in common is a concern with _attention_ as a resource: a collectively predicted future is one where attention is directed towards synchronized goals (whether this be investment, alignment, etc.). Power functions through attention: perspectives _make_ reality. This is an attentional effect at the level of self-evidence in a single-agent system, or evidence in the courtroom: perspectives are needed to justify narratives, or quantum-level observations: perspectival interaction determines outcomes. The society of control is a society of modulated attention. Attention is, as we observed at the level of narrative scripts, directed by the geometry of concepts, which make some things salient and ignore others: attention is what parses chunks. In _The Capitalist Schema_ (2014), Christian Lotz explains that a concept such as Deleuze’s “dispositif” loses its specificity in its widely applicable use as it becomes “emptied out, universalized” (p. xiv). This is true of many (Deleuzian) concepts, and in our context, _modulation_ can suffer the same fate. In terms of understanding words-concepts as fulfilling the function of attentional semantic attractors: once they are “everywhere”, their specificity—the way indices reduce uncertainty by making a difference that makes a difference in a complex state—disappears. The concept of the _society of control_ is introduced by Deleuze, according to Lotz, so that the specification of capitalist reality turning _disciplinary_ societies turn into _control_ societies, can be marked with a higher degree of accuracy. The capitalosocial switch is one from “labor to **power** as the substance of social reality.” (p. xiv, our emphasis). He suggests that, in his view, it is still labor—mediated via money—that remains the core that drives _life_ in society. We do not disagree. What is interesting with regard to an analysis of power as a function of attention-pooling, is that these different degrees of mediation point to how, whatever we call it: “Recursivity is the mechanism of the norm established between the living being and its milieu. It is not a mere imposition, like mechanical laws, but rather a Spinozist “immanent causation.” We can therefore understand why, commenting on Canguilhem, Pierre Macherey claims that there is an immanent causality of the norm, which emerges from the subjects subjected to it. Norms change when an event occurs that exceeds the normalizing capacity of the norms.” (Hui 2019, p. 154). Once a concept loses attentional traction, it changes. Bourdieu also notes: “Capital is accumulated labor (in its materialized form or its “incorporated,” embodied form) which, when appropriated on a private, i.e., exclusive, basis by agents or groups of agents, enables them to appropriate social energy in the form of reified or living labor.” (1986, p. 16). However, in light of our observation of power as attention-modulation and the evolution of function, taking notice of the _loss_ of specificity, and the introduction of a new idea to designate the “same” process, is important. Think of the many “middlemen” companies nowadays that hold customer and investor attention—through labor, for sure, but also through speculative capture—but without “owning” infrastructure nor focusing on production. Or, another kind of conceptual modulation can also be framed from the perspective which highlights technoscientific infrastructure evolution as _powerful_ in the sense that it is capable of the making attentionally salient of “facts” to be accepted as objective and authoritative. This has been observed, too, “as a methodological corrective to traditional histories and theories of discovery that revolve around an isolatable, heroic figure of genius.” (Selinger in Protevi 2005, p. 7). The latter can be therefore framed as an affirmative modulation towards -dividualism, in a technoscientific context. How matter is (socially) attended to as information, as relevance, and vice versa, reveals the unfolding functional dynamics of systems like capitalism or science: this is where (technical) concretization can be interestingly attended to (by looking at case studies, examples, etc.). How “raw” materials become silicon chips which render variants of *information* can also be analyzed as how experimental data become theories, transforming (social) understandings of “raw materials”. All of these effects are functionally transferred across cables, concepts and meat. In the end, we can only do with generalizations. Ultimately, because potential power—as (speculative, metaphysical) attention—is a material-informational force, “seizing the means of computation” (Doctorow, 2023) follows in line with a necessary Marxist analysis of capital in our moving paradigm: from materialist analyses of territorial and labor time domination, through a semantic-sociocultural analysis (Bourdieu, (1995) 2018), to informational analyses where we ought to think about the production of the future by means of projective computation. Power is thus based on the controlling of attention by way of scripts, and its current substrate in our contemporary landscape is computation: both in the flesh and _in silico_. Power’s dual nature (as force attraction and information computation) in the social realm thus implies the use of scripts as they are the means by which agents gain or lose the ability to direct the attention of other agents, and the ability to effectively process information in favor of these policies. The variables modulating regimes of attention that have practical modeling potential are plenty,^[The formalization in the collective paper explores a few examples, the list is inexhaustive.] our interest in the collectively written article is in the control, influence, and synchronization of attention, which lead to social power structures. There, the focus is on empowerment quantified as agents’ potential influence over their environment (which is tracked by assessing the mutual information between possible **actions** and given **states**). In the collective paper we analyze this through formal methods that are tractable, but this presents us with indications of future modeling possibilities, which in our context remain unformalized as they are practically intractable. We can think, for example, about the modeling of “inheritance” or “provenance” variables: wherever or however an agent exists, they could be born into a rich family, in a specific part of the world, etc., and how these factors have an influence on how _attention_ falls on them (and therefore on life-expectancy: mortality, and many other crucial effects). One can also posit a possible “frictions/contingencies/contextualities” variable—in colloquial terms “you have to work with what you’ve got”—by which, even if agents can attract others’ attention and process information effectively, the projected goal and thus possible state space is constrained by a context which can be highly divergent from the “ideal” one required for bringing a goal about. Confidence in policies is a well-framed and studied variable in the context of AIF, and in the literature it is often directly linked to the concept of power: “self-confidence [is] the person’s experience of their power to act in the world” (Kiverstein 2019). There are more, of course, and it depends on how granular one may want to go, these are examples which give the reader an indication of our _modulating_ intentions. &emsp; >The particular doctrine in question is, that in the transformations of matter and energy which constitute the activities of an animal body no principles can be discerned other than those which govern the activities of inorganic matter. > >Whitehead, 1929, p. 15. > >Mankind has gradually developed from the lowliest forms of life, and must therefore be explained **in terms applicable to all such forms**. **But why construe the later forms by analogy to the earlier forms. Why not reverse the process?** It would seem to be more sensible, more truly empirical, to allow each living species to make its own contribution to the demonstration of factors inherent in living things. > >Ibid., p. 18, our emphasis in bold. &emsp; Power as potential, as force; as the ability to do work in physics and the ability to do work in the socioreproductive realm, represent two sides of the same coin: the mechanics of matter and energy sustain the dynamics of complex systems; which, in turn, modulate matter and energy. This is precisely how anything that appears to persist comes to “acquire the statistics of their embedding environment, and seem[s] thereby to encode a probabilistic representation of that environment” (Ramstead et al., 2023, p. 3). As we argue in the collective paper, under AIF, this ‘common currency’ can be studied as the process by which state spaces are explored by collectives (dis)attending to projective possibilities. In order to understand these processes under ever-novel regimes, formalizations are necessary: they reveal where and how we can observe functions transferring. This is our interest in the collective work: awareness about the modulation of attention begins at the level of understanding the constraints framing attention; this is how we see power as that which _pools_ attention. As Foucault notes in “The Subject and Power” (1982), formalizations are not without their reductive, objectifying problems, but are necessary in order to drive critical analysis forward. Further, he also notes: “What we need is a new economy of power relations—the word “economy” being used in its theoretical and practical sense” (p. 779). This is, in great part, what we explore, not proposing a rigid structure but “a technique, [which is] a form of power” (p. 781), as it modulates attention. &emsp; >[Algorithmic techniques such as Bayes classifiers] introduce a ‘micro-physics’ that may well affect how power operates in significant ways. And the term ‘operate’ needs to be taken seriously, here. The computerized infrastructures human activities are increasingly entangled with imply forms of semiosis that take effect not like signs, but also, with loose reference to Lessig (1999), like walls, encroaching on conduct through modifications of the visible and navigable environment. > >Rieder 2020, p. 256. &emsp; Attentional modulation spreads through social script-scaffoldings that amplify certain messages while marginalizing others (Albarracin et al., 2021), rendering _walls_ that demarcate the physics of salience, the physics of belief (Ramstead et al., 2023). This process creates systems of ‘cultural capital’ (Bourdieu) where power emerges from the interplay of self-evidencing, attentional influence, and as scripts which become plastically embedded into structures of cultural learning (Tomasello 2016), enabling different kinds of social modulation. The question of social sameness (as in “same” narratives, belief similarity, mutuality or symmetry, etc.) and (dis)attention to it, is therefore a political question, a question of desire at the level of the infrastructure (Deleuze and Guattari 1972). If attention is fundamentally shaped by deep preferences that determine what becomes salient in the environment for an agent or group, then we ought to ask how the production and performance of scripts (_habitus_, narratives: histories, identities, etc.) can be newly understood so that we may critically construct these collectively, meta-aware of how these functions run through us, partly enabled by competitive interests; power imbalances. All this, hopefully, without risking that this collectivity is induced by strong homogenizing scripts which occlude or incapacitate ever-renewing possibilities. Essentially: if we ought to adopt a politics based on the knowledge that scripts work in the ways discussed, and if the organization of life is hierarchical by definition (both physically and metaphysically: we are given to a world of already-existing constraints which we organize around by means of concepts aimed at future possibilities), taking account of this by understanding the nature of distributed generative models^[See: [[Generative model]].]—over time: as ecologies, generations, as the representation of historical narratives, etc.—offers a vision of possibilistic landscapes based on collective dynamics (and their analyzable mechanics) which can be framed beyond our current visions based on simplistic ideas of agency as an elusive individualized freedom and through atomized actors. &emsp; ### Ctrl + Z (_Undo_) ### Conclusion As Deleuze noted, we have moved from the frictions of discipline to the smooth flows of control, from individuated bodies to data aggregates which control themselves: in the script society, the distinction between scrip-_ting_ and scrip-_ted_ is irrelevant, the medium is the message: less and less “seeing as” and more and more supposed immediacy—which Kornbluh (2024) calls the “style” of late capitalism, diagnosing “flow” as the 21st century word per definition. Systems—such as most “social” media platforms—that feed on complexity in order to provide ever-more diversified products, thrive on the stochasticity of that which it aims to systemically contain. From an anti-control vantage point, this creates a homogenous mass that is self-perpetuating. That is exactly what the post-control script society is, what was earlier diagnosed in its proto-form as the bland results of the culture industry by Adorno and Horkheimer: containing contingency by way of contingent scripts. This is the script that writes itself (i.e., we already exist within the _Paperclip maximizer_, it’s called TikTok, Facebook, etc.). As noted by Deleuze—whose point about modulation is made even more interesting through a return to Simondon, as Hui shows—the shift from territorial/material to informational/computational power requires special attention with regard to the modulation of future contexts. AIF provides crucial theoretical tools for making power structures visible and therefore analyzable, modulatable, allowing for the potential to rethink current social asymmetries by observing models of their unfolding structures through attention and power. Given the transition between the organization of life around the centrality of physical work and matter, towards a centrality of attention and information, a socially-coherent future requires addressing both resource and information distributions. The latter can be analyzed, we argued, through the power of attention as a major factor influencing social dynamics. The ontoepistemic challenges in grounding meaning or value or relevance (i.e.: salience), can be understood under AIF in ways that open up complex semantics and social coordination towards a pragmatic naturalization of these phenomena. Not in an effort to reduce them to simplistic images, rather in an effort to provide new technical entries into as well as possible accountability metrics of thorny ethical landscapes. In our proposal, coordinated behaviors emerge as apparently shared models that reduce collective uncertainty. Always with the caveat that whatever is supposed to be “shared” is never given nor fully understood: as scripts unfold and evolve, not even the “same” word is ever the same (Masterman 2005, p. 78). When scripts become internalized by multiple members of a community, incorporated; organized, this allows for the collective construction of future states. Following those who appear to have the potential to reduce our uncertainty, leads to agents associating with (agents and) scripts that might meet those expectations. Metacognitive criticality, that is: not just playing the game (or being played) but knowing the rules and being able to imagine new rules, is how we ought to collectively navigate uncertainty, _assuming_ already given asymmetries. Future research could focus on developing metrics for measuring possibilistic power and designing interventions to promote more proportional (i.e., rational) distributions of material-informational resources. Hierarchies are inevitable in complex thermodynamic systems. We propose enhancing the ability to delineate tractable, plastic hierarchies that remain contestable and therefore open to novelty. This approach acknowledges historic and thermodynamic constraints, while acknowledging space for systemic evolution. While hierarchies are inevitable, social scripts should be read-writable by those with an interest in survival within and through them. In this context, thinking of how to make (social) scripts read-writable; modulatable, “Debate mapping” by the _Society Library_ is a compelling example. This project employs different kinds of language-modelling architectures in order to map the bifurcation paths of (controversial) debates. Rather than the traditional debate approach often employed, where pros and cons are weighed up against each other and compared, resulting in, e.g., the creation of “neutral” reports on subjective positions around (controversial) issues, this project focuses on showing how positions and opinions within a debate can unfold.^[In their own words: “We believe that in doing so, and by mapping and offering these maps for free to the public, that we can improve the collective intelligence of society as a whole, and that collectively we will all become more informed, less polarized/divisive, become more understanding/empathetic, and be more intellectually humbled by engaging with these maps. We are looking forward to studying these impacts on cognition by more sophisticated means as resources permit. We also believe this data may contribute to the development of more sophisticated decision-making models, and potentially even AI applications” See: https://www.societylibrary.org/debate-mapping-program.] This type of project takes the idea of opinions as processes seriously. It is an impressive example of how (relatively simple language) technology can assist a meta-informed collective owl-of-Minerva perspective. The consideration of many possibilities and their intricacies can, at the very least, leave us with the _via negativa_ of what _not_ to do. A (Simondonian) point of insistence throughout our thesis is that this, and any other imagination, is _technical_, because being linguistic is technical. Becoming observant to how technicity (again: where we consider scripts, language, as technology) transfers function creates new metaperspectival opportunities. Metaphors morph: from the mind as a river; to a mill; to the mind as a computer. We create, actively engage-infer-ignore the world, and extract function from possibilistic process landscapes: function is only seen in retrospect. Monkey _do_ and only thereupon monkey _see_: this is the core of _active_ inference. This helps reorient our understanding of an unfolding, globalizing, interconnected and interdependent life: participation in a system is only participatory if those involved are able to track the consequences of the variables and thus future state spaces at stake. A true distribution of information processing capabilities is essential for systemic sustainability and what we often frame as “equitable” development. And, to continue thinking with and through productive tautologies: “if nature is unjust: change nature.” (Laboria Cuboniks, 2015). This work therefore reorients, towards a metacognitive perspective, what has been previously framed in terms of territorial, material and cultural power, towards what can now be understood in terms of informational and computational power. All of these have received attention as forms of capital; a term we can reframe under AIF as _possibilistic_ power, i.e., _the creation of the future_.^[Please do see the collective paper on the formal aspects of this reframing.] <div class="page-break" style="page-break-before: always;"></div> ### Footnotes