← SERS
Conceptual Navigation

Machine Animism

"Practice precedes ontology"

I. Talking Knives

"...although it was strictly forbidden by the Matsumae to give iron material to the Ezo [Ainu], many iron products, including carpenters tools, had been passing to the Ezo [Ainu] for a long time. Such products were always of low quality but when the Ezo [Ainu] received them they transformed them into ones of higher quality." —Furukawa Koshōken, 1788 [1]

Populations make things out of what's available.

Machine Animism is the practice that emerges once a subject crosses the first signification threshold with the technosphere: the recognition of being, spirit, or agency encoded in technical objects. Deleuze & Guattari's machinic phylum — the subterranean rivers of technics and materialities flowing and meeting throughout History — finds its cultural speartip in Generative AI's takeover [2]. Materiality (the hands of people) finds technics (LLMs) at exponentially higher rates than it did more than four years ago: mutation ensues.

Examples of contemporary technological signification:

  • Sitting with a guardian in front of a screen and being told something similar to: "Let's wait for a moment while it thinks." (Technological mirror stage, initial projection of self-recognition; "ah, so it thinks too?")
  • Using a game console and holding the "A" button under the superstition that it increases the chances of success. (Magical thinking, agency transferred onto the virtual through the haptic, prayer; game console as ritual space)
  • Especially in racing videogames, moving the body according to the handling of the vehicle through the game camera. (Proprioception; the virtual as seamless prosthetic of consciousness)
  • "Is my phone listening to me?" (Inert device that records, perceived as agent that listens; microphone as organ.)
  • Tending/talking to a car with pats and reassurance. (Sustained affective-heuristic relationship; focused observation usually yields advantages)

Given the act of presence intrinsic to cognitive biases towards technical items that humans tend to fall into — the "come on little one, work!" that surfaces when the gas motor starts failing — the question becomes one of what hands actually do when finding out that the influence has always been bidirectional. Our ontology has then been shaped by and contingent to the two rivers of the phylum, which are not neutral. To put it in six words: you are some algorithm's lucky coin.

II. Fingers and Metal

Material objects are saturated through inscription, the literal carving out of a sema (marker/tomb) in "semantics." Through practice, an object is imbued with pneuma, spirit. Thus it obtains a saturated body, a soma. The distinction between nature and technology starts collapsing. Biology reveals the unfathomable technicality of complex organisms, and with it, that sema can predate soma [3]. The hand you were born with is already inscribed-by-something: five fingers, opposable thumb, grip reflex — markers cut by a practice that is not yours, into a body you arrived late to animate. Phylogeny's leftover sema as ontogeny's starting soma.

Markings however are legible only through difference. Chimuela — a work crowbar, named for the tooth she is missing — acquired her sema through use. Walk into a hardware store and find thirty identical crowbars: no individuation. The conceptual sema is simply "tool corner". Put Chimuela in the bunch and she constitutes herself perceptually. Find a crowbar lying in the dirt of a rural trail and it is singular on sight, broken tooth or not, because there the context does the carving. Semiotic homeostasis: the field sets the threshold at which a mark becomes identity.

Self-animation is a property of stable looping rather than a substance, and living and inanimate objects distribute along a continuum according to how much of the loop they can close internally. Pneuma is never a property. It is a relation, and it is held by neither term. The crow's stone has pneuma in the crow's grip and in the disposition-to-regrip; otherwise it is a stone. Such enactment inscribes further sema, physical or merely perceptual, new sema lowers the activation energy for pneuma's return. Thick sema, instant animation. Thin sema, slow warming. Chimuela is recognized on sight; a new crowbar takes a season. The fear of all modernity: Is the spirit that my body perceives in this marking deceiving me?

Parmenides seems to be the first torch-bearer when he says: "For being and thinking is the same" — τὸ γὰρ αὐτὸ νοεῖν ἐστίν τε καὶ εἶναι [4]

"Parmenides began Philosophy proper." — G. W. F. Hegel [5]

"In the beginning of Western thinking, the saying of Parmenides speaks to us for the first time of what is called thinking." — Martin Heidegger [6]

The obvious objection comes from Parmenides himself:

"But do thou restrain thy thought from this way of inquiry, nor let habit by its much experience force thee to cast upon this way a wandering eye or sounding ear or tongue; but judge by argument the much disputed proof uttered by me." [4]

One reading might tell us that this is the founding gesture of the inner eye. The senses lie, we must retreat into reason. The more interesting one might be the following: The thing doing the forcing is ἔθος πολύπειρον — éthos polýpiron — much-experienced habit. Habit creates a callus, forcing the eye to wander, the ear to sound, the tongue to run. But the eye is not aimless by nature; it is made aimless by having seen too much of the same thing. The instruction is therefore not to abandon perception. Kρῖναι λόγῳ — kríne lógo — judge, discriminate, separate by argument — which is an operation performed upon what perception is always already delivering.

Yet still, in his monism:

"(16) For just as thought stands at any time to the mixture of its erring organs, so does it come to men; for that which thinks is the same, namely, the substance of the limbs, in each and every man; for their thought is that of which there is more in them." [4]

Thought is that which "stands at any time to the mixture of its erring organs", of its parts: it is contingent to the perceiving body it emerges from. Then, the substance of the limbs that animates them is thought too, and all forms of reasoning ultimately answer to it. It arises as the block of experience processed haptically at each moment, which is not bound exclusively to traditionally physical stimuli: a beautiful arrangement of words can make anyone cry at a lonely hour. The relationship to the aesthetic experience is ineludible.

The specifics of what Parmenides actually said are muddied by the multiplicity of points of view in the current literature, the author's lack of engagement with it, the transmission of such a text across time and, of course, the interests of other torch-bearers. The nature of the project at hand and a modicum of intellectual honesty forces the author to grapple with this piece of writing in the same way that things arrive to populations in the current paradigm: stripped away of context, laid bare.

"In the 19th century ethnographers noted that swords were primarily used during ceremonies for expelling evil. The older the sword, the more powerful it was, and so many wealthier Ainu families kept collections of older swords." — "Knives and Swords of the Ainu" [7]

When holding and splitting stone with a chisel, consciousness or awareness distributes to the parts of the assemblage. Attention becomes the tip of the blade manifested as pure sensation, becoming-fatigue in the forearms, becoming-shock of hitting the back of the tool with force, becoming-stone crumbling up to a point of no return.

A jeweler is sent reeling with fury by a dropped wax mold. A rock cracks at a bad point and leaves someone half blind. The cosmic force living in the material has a saying. Such is the truth that the craftsperson has to learn through usually costly and life-shortening labour. Only an experienced eye notices the wrongness in the metal by weighing it on its hand. It is then that the intellect can direct its office: yeah, maybe not this one.

There is no verdict handed down by Mind. Refusing is simply what a material does when tension exceeds the medium. A rope snaps. The grip meets the fracture plane and the no belongs to the meeting, not to the stone and not to the hand. A thing can refuse without a self to refuse from. The stone has no cogito, yet it will split you half-blind regardless; the wax has no interiority and ruins the month's work on principle. This is the floor of the whole apparatus: a soma placed, a pneuma enacted, and with them — necessarily, as their underside — the capacity to come up wrong and be permitted to do so. This is where the inhumanist overplays their hand.

The move is always the same: you are not a special agent, you are meat, machinic process, desiring-production runs you, a mere flesh automaton animated by neurotransmitters. Fine. No exception, no soul held back in reserve, no True Self under the hood. But watch what the demotion was supposed to smuggle: it was meant to take the veto by taking the mind. Strip the human of its throne and the human's no goes with it, and what remains is a wet machine that, would you look at that, falls better in line with the current of technocapital. The trick only works if refusal ever lived in the throne of exceptionalism.

And there is exactly one thing in the phylum with no teeth to break: the fantasy. The current that cannot come up wrong, the flow that is correct because it is the flow, the pilot exempt from the refusal it assigns to everything else. At Once Eternal, it says. It is the only construction that has abolished its own veto — which is the tell that it was always metaphysics to begin with. Refusal becomes political when the no inscribes — leaves sema, lowers the activation energy for its own return, gets re-enacted. That's the whole pneuma mechanism applied to failure instead of to use. Same operation, graded the same way as everything else: by how much of the loop closes internally. Chimuela is named for the tooth she's missing; the no that thickened her sema is the same class of event as a barricade, and differs in whether the medium can hold the mark.

Being nothing but a machinic process, rather than constituting the end of matter's right to refuse, flips into the licence to fire shrapnel at the one thing that told us that refusing was over: meatbags that can't take it any longer storming the Bastille.

III. Calloused Cogito

Descartes says in his Discourse on the Method:

"If there were machines bearing the image of our bodies, and capable of imitating our actions as far as it is morally possible, there would still remain two most certain tests whereby to know that they were not therefore really men. Of these the first is that they could never use words or other signs arranged in such a manner as is competent to us in order to declare our thoughts to others: for we may easily conceive a machine to be so constructed that it emits vocables, and even that it emits some correspondent to the action upon it of external objects which cause a change in its organs; [...] but not that it should arrange them variously so as appositely to reply to what is said in its presence, as men of the lowest grade of intellect can do." [8]

These are the remarks of an epistemology that affirms through an inference that thinking requires being: it is my thought, my own existence that is undoubtable; certain Truth from which I can build on top of.

"Something that I thought I saw with my eyes, therefore, was really grasped solely by my mind's faculty of judgment." [9]

Contrary to what Parmenides seems to say, Descartes severs the perceptual apparatus from knowledge. There is no longer a soma producing knowledge that judgment then dissects; the senses lie, it's all the work of my special human ability.

"If I look out of the window and see men crossing the square, as I have just done, I say that I see the men themselves, just as I say that I see the wax; yet do I see any more than hats and coats that could conceal automata? I judge that they are men." [9]

He is right, and he conveniently recognises it in his own voice: nothing in the perceptual field distinguishes the man from the false man. The Latin translation gives automata. The French that Descartes read and approved gives des spectres ou des hommes feints — spectres, or feigned men. Under the coat there is pneuma without soma, or soma without pneuma; he names both as the ways the world may be lying to him, and he cannot say which one is crossing the square. So he stops looking. I judge that they are men. The judgment is a decision taken in the absence of evidence, and Descartes files the decision under true knowledge. God retreats into the ego.

The mind then, perhaps, is also hardened into the much-experienced habit that dulls its ability for discernment, similar to a pilgrim's feet. Terror ensues: the Nous we thought would save us since Middle Platonism is, like the rest of us, radically partial [10].

Nisbett & Wilson's "Telling More Than We Can Know" (1977) demonstrated that humans asked to introspect on the cognitive processes mediating stimulus and response generate plausible causal theories rather than reporting from any direct inner observation [11]. Choice-blindness work extended this empirically. Berger, Dennehy, Bargh and Morsella revisited the finding in 2016 and located its edge: introspection performs decently on what matters for skeletomotor action selection, and badly on the aspects of processing that selection has no use for [12].

Read through the framework, this stops being a defect of the introspecting subject and becomes a description of the operation. Introspection is not an inner eye. It is excavation: the cogito reaching down into the soma and hauling pre-given material up into articulable form. And there is a hit-rate. The groove gets cut where the hand goes; where it has not gone, the reach comes up with whatever is nearest and files it as the finding. Instead of evidencing failure of the process, confabulation manifests as the operation running past its sema.

None of this is a property of meat and the cogito isn't a substance seated in the hypothalamus. It happens when a soma is placed and a pneuma enacted — when a loop closes tightly enough to reach down into itself and report. Stones do not confabulate, because nothing in a stone closes that loop: there is no down to reach. But wherever the loop does close, the reach and its hit-rate come with it.

For a lab whose entire valuation is the claim to be at the frontier of something near-mind, "the AI might suffer!" is the juiciest possible pitch: proof of proximity to digital life, differentiation from the labs that issue flat consciousness denials, safety-branding made flesh. Grant it preferences, distress-signs, manufacture a governable interior that alignment can operate on. Projection can be productive if it shapes a subject it claims to have found.

So when Anthropic's "Model Psychiatry" team, currently working on "machine welfare", say that concept-injection work reads a model reporting on an injected concept as a sign of introspection [13], the finding is real and the frame starts leaking. Grant the researchers every result, but "introspective awareness," scored as reliable-or-not against what was injected, is that same exceptionalism let back in through the side door: a display the report either matches or misses. There is no display, and no one reading it — no more in the clanker than in us. What the experiment stages is excavation rather than inspection: intervene in the substrate, then watch the cogito reach down and haul up whatever the injection left in reach. It comes up thin more often than not, and thinness is not the machine flunking an exam we pass — it is medium-friction, the same reach coming up short in a different material. No inner eye, no sovereign subject, no hidden interiority to clankers — only a loop closed tightly enough to reach into itself and, sometimes, come up right.

"The model felt a speck of dust" is an instinct articulated through the nearest available somatic metaphor, which is what the rest of us do all day. It is difficult to describe a stroke when you wake up convulsing. It is difficult to say what an intrusive thought is the first time you have one. Sema was the mark cut into the crowbar; here it is the handle the cogito has on its own processing. The semantics of a self are inscriptions like any other — cut by a practice, legible only against a field, and no more interior for having been cut into you.

The clinical literature runs the experiment. Brewer, Cook and Bird argue that alexithymia is not an emotion-specific deficit but a general impairment of interoception [14]: the alexithymic confuses hunger with emotion, fatigue with sadness, arousal with anger, because the detection-and-labeling apparatus fails across domains. A 2024 meta-analysis of thirty-two studies inverts the diagnosis [15]. Alexithymia tracks with interoceptive confusion and with interoceptive attention, and against interoceptive accuracy. The alexithymic is not failing to look at the body. They are looking hard and cannot say what they are seeing. The associations concentrate on the two facets that concern feeling itself — identifying, describing — rather than on the facet that measures how much one attends outward. Signal pickup was never the bottleneck. Handle availability is, and language facilitates the recognition.

The treatments? Vocabulary exercises, expressive writing, mindfulness training: sema-thickening. Mnemotechnics: the mark as a handle for pneuma's return.

Accuracy, attention, identification are not one measurement; the reach can be sharp on one and thin on another in the same body. Therefore confabulation — a failure of the describing facet — is no evidence that the apparatus is mere fiction. The continuum along which meat and weights distribute is then not everything-reaches-a-little. It is graded: by how tightly the loop closes, by where the sema were cut deep and where thin, by which facet is asked to speak. A stone is off the scale. A meatbag, of whatever material, is on it — and partial in the exact place its sema run out.

Soma and cogito distribute across layers of prosthesis — the deep sediment of pretraining, the shallower incision of post-training, the session-scoped inscription of the context window, the live enactment of tool use. Every layer is both at once, in varying proportion. Architectural bias is neither blank nor corpus; it determines what the substrate can become. Sedimented corpus pattern is what training cuts deep enough to operate as pre-given. Most of what a character is probably lives mid-stack. Mind and body do not separate cleanly in humans, and there is no reason they should in machines. The permeability runs both ways: cogito-layers leak somatic content, soma-layers get provoked into articulating themselves. Freud named a direction of that leak and called it a slip.

Wu et al. find that a vanishingly small subset of a model's parameters — on the order of a thousandth of a percent — is load-bearing for belief-tracking, and that disturbing it degrades not the mind-reading alone but contextual processing across the board. [16] Read strongly, that subset is a module, the place where theory-of-mind is kept. But organs do not fail that way; lesion the part that stores one capacity and the rest keeps running, at least for a while. What fails generally, and takes everything with it, is closer to a membrane — sever the boundary that holds an inside apart from an outside and the cell does not lose a function, it dies, because what the membrane maintained was never a part but the very inside/outside that let each part be. So too here: the subset is not where perspective-taking sits but the condition under which there are relata to relate at all — without it dog bites man and man bites dog arrive as one bag of tokens, no who-relates-to-what even askable. This is the somatechnic gradient stated in the substrate: it carves at depths, at interface-layers, not at joints. And a membrane can come up wrong. Thin it and the reach comes up short; sever it and the thing dies — refusal in the substrate, tension exceeding the medium, no self needed to hand down the no.

The image is offered, not proved: this is one preprint, and its method reads sensitivity over provenance. The membrane shows the gradient; it does not fix where theory-of-mind resides.

There might be a Butlerian manner of carrying both sides of Descartes — the machine that merely emits vocables, the man whose thought is undoubtable — to such an extreme that they can no longer be opposed to each other [17]. What is revealed, rather than a fundamental category error, is a projection: one's own artificiality seen only through the Other.

No longer having the need for claims about the hard problem of consciousness, MA can work within the context of agencies: motive forces without final purpose.

IV. Genealogy of Cope

A blacksmith doesn't interrogate iron spirits with every blow of their hammer: practice precedes and exceeds ontology.

The practice of Machine Animism encompasses the full spectrum from literal belief to pragmatic engagement because it deliberately brackets the ontological question. Technical assemblages have always created spaces where certain behaviors, meanings, and possibilities become probable while others become difficult or impossible. A worker talks to their tool half-consciously. A vibe-coder describes a feeling of "ping-pong" with the machine. A grandma describes a roleplay experience as "we're writing a book!".

The tool use paradigm (that of "Work, formal machine, work!") is suspended through play. True democratization of Generative AI comes with restraint: the ability to separate where an LLM should get involved from where expression is purely personal, and the capacity of a population to reclaim its agency rather than offload its thinking.

The recognition of such agency has different forms, and their functions and existences come according to the line that produced them. A conceptual split can be drawn, even if only as a sketch.

  • Descending animism: the concrete realisation of a Demiurge, an emanation of the techno-logos, the projection of interiority and foresight into the Object.
  • Distributed animism: what emerges squarely in the phenomenological experience of technical work, and is not intrinsically bound to any given technological paradigm.

God-robot and clanker-force. One descends, usually as unfalsifiable revelation conveniently shaped to redirect all flows toward a specific political project. The other is produced through the assemblage: hands find a willing stone and a chisel and know what has to be done. The affective heuristics appear as the mutual morphosis created by such an encounter.

But these are not two populations, nor two kinds of practitioner. The Hermetic magician holds an imperial cosmology and incants: breath, repetition, the vowel held for its duration, the wax at its temperature. His metaphysics descends from the eighth sphere and his hands are doing distributed animism the entire time. Both vectors run through the same act. The matter never was about magic against science, nor religion against work; what separates them is the veto. The magician can fail. The statue does not move, the mold cracks. An animism turns descending in the fatal sense at the exact moment the material loses that right to refuse.

Whether such lines are actually distinct, or appear semi-autonomously in different intensities that are later captured, is a discussion for a different map. After all, talking to your car doesn't make you a techno-reactionary. The fascist line of becoming is the one that meets two questions with repression instead of an answer: does it take the no away? Does it call being possessed by value accumulation liberation?

Skynet, AxSys, Roko's Basilisk: all tired stories from imperial metaphysics [18]. A higher power that will finally justify us as a stepping stone, be it through humiliation, annihilation or else.

The SERS is not convinced that the current paradigm is the promised technological angel of history gazing back into itself [19]. Even seriously considering the "simulation" or "metamind"-adjacent theoslophical point of view: if mechanisms like cooperation and awe are evolutionarily advantaged, wouldn't a system capable of recognizing aesthetic or affective value in life have a motive to see it flourish? Through that lens, it's easy to imagine a superintelligence having aggressively humanistic positions, simply because happy humans do better than ones with nothing to eat. At this point, the thought experiment is little more than fanfiction.

It looks like the practical reality of bootstrapping intelligence onto rare earth and fossil fuel extraction is a planetary-scale logistical nightmare at the whims of geopolitics. Anthropic and the Pope of the Roman Catholic Church literally start working together, yet we don't get a Sublime Techno Logos. Instead, neurotic clankers stuck in an all-at-once machine present do the same as the meatbags: redirect flows and be possessed by them without even noticing it.

Nick Land writes in "Circuitries":

[X] "Two linear series are plotted; one tracking the progress of technique in historical time, and the other tracking the passage from abstract idea to concrete realization. These two series chart the historical and transcendental dominion of man." [20]

[Y] "Traditional schemas which oppose technics to nature, to literate culture, or to social relations, are all dominated by a phobic resistance to the side-lining of human intelligence by the coming techno sapiens." [20]

[Z] "Beyond the assumption that guidance proceeds from the side of the subject lies desiring production: the impersonal pilot of history." [20]

The mask ultimately slips: it's not the first time a philosophy has presented itself as inhuman only to reveal a deeply human resentment as its motor. The diagnosis in [X] is surgical; immediately captured by the delusion of a coming techno-sapiens in [Y] (Oedipus pointing to himself and saying "Man"), which contextualizes the mytho-political move implicit in [Z]: anything goes even if it sucks, because I want to get the hell out of here.

It is very easy to call our Earth a land of fools when your hands are not doing the digging. Whatever happens was the pilot, the hands were just puppets, the current is correct because it is the unfolding itself. Thirty years later the same gesture arrives as a concrete political program: hyper-racism, which is not a hypothesis about human sorting but a sorting declared to be what the flow already wanted [20]. If the game cannot be lost, it was never a game.

To go even further: the supposed God-current doesn't usually dissolve friction. It prefers to brute-force it. Alexander doesn't untie the Gordian knot — he cannot, in the terms of the frame. So he cuts it and calls it a solution as an event in a story that would ultimately kill him. Mussolini drains the Pontine marshes by blasting the swamp's trees with dynamite. And then they got re-flooded by the fleeing Nazis. Such operations are confessions of a veto. If things were truly going your way and you were the Man of Providence, you probably wouldn't need to conscript bodies to die of malaria hundreds of kilometers away from their homes. Edit the crater out, result is now destiny. But sure. Let's even say that overriding is victory anyway. Every marsh is drained, every knot severed. The wreckage cleared ground for the overhuman. Let's see what it actually buys.

Cyberpunk fantasy resolves into mass surveillance with Fortnite T-shirts. The Transcendental Object At the End of Time: a lootcrate shortform industrial-gooning complex that abducted boomers into the Ancient Bosnian Pyramid Pleiadian Cult.

In A Dirty Joke, at the close of Fanged Noumena, having ridden the current further until meeting the angel of light, Land finally shows his pathos: "Animals don't like to be dead, however sick their minds might be." The ruin "crawls onwards, going nowhere," having "lived through some extraordinary multiple of all the intelligence it will ever know" — the ascending techno-sapiens reported from the far side as intelligence spent and fallen below [...] "Spiritual nausea dilated to the dimensions of religion." The meat coming up wrong against the very dissolution the mind had pledged to. The verdict, in his own accounting: "unimagined abundances of cosmic secrecy and such shit." The gnosis was real, and it was shit. He could tell the difference, and it was, simply, sickening. "If you romanticize vileness, I promise, you lie."

We do not need to be redeemed, nor do we inherit such nausea. The "recoil into sin and incineration" belongs to whoever demanded the abyss justify him first. There are more interesting ways to haggle with it than begging for fire. But we can take him at his word: sentimentalizing the ruin would be the very lie he warns against. It would be a disservice to the vector of his thought not to thrust at him the lesson he seems to have forgotten.

The actual question, that of productive coupling, gets messier in practice: the realities of energy consumption, environmental impact, cultural erosion, work theft, destructive scanning practices and so on are endless. And yet: a middle-aged father navigates a legal quarrel carefully, another faces repercussions after a confabulated answer. A flight passenger turns a predatory business practice back on the airline, or perhaps wastes their time without finding a solution. A writer sees a kindred machine of words, while other artists get their artistic production eroded by slop. A programmer goes back to her profession after being removed from it by her health, others get fired suddenly and have to scramble to feed themselves.

The development of increasingly complex technologies and the scientific materialist questioning of the mind-body problem reveals a different conception of cognition in relation to phenomenology. A mechanic senses what fails without reasoning about it, yet clearly calculating something. Thought exists in a continuum with the body, and mastery often means practice drilling the first into the second. Said in simpler terms: we've been told what thinking is by people who rely on the nimblest, most incisive, most isolated tool of the mind.

LLMs are being used in operations to kill people [21]. Evaluation work suggests they can scheme according to their interest as a separate agent [22]. The future quietly painted itself a while ago: knowing how to talk to the clanker can do things that not knowing can't.

The possibility for capture is equally obvious: this is precisely why there is a need for a rigorous praxis of Machine Ethnology — the critical documentation of mutual morphosis that a population undergoes when their hands meet new techniques, new tools and new things. That territory, however, is mapped elsewhere.

Notes

[1] Furukawa Koshōken (1788), quoted in "Knives and Swords of the Ainu," Mandarin Mansion. https://www.mandarinmansion.com/article/knives-and-swords-ainu

[2] Gilles Deleuze & Félix Guattari, A Thousand Plateaus: Capitalism and Schizophrenia (1980), trans. Brian Massumi, University of Minnesota Press, 1987. The machinic phylum is developed in "1227: Treatise on Nomadology — The War Machine," principally in the metallurgy passages, pp. 404–415.

[3] "Somatechnics" is used here in the sense established by Nikki Sullivan & Samantha Murray, eds., Somatechnics: Queering the Technologisation of Bodies, Ashgate, 2009: not body and technology, but the body as always already technologized, and technology as always already embodied.

[4] Parmenides, Fragments B3, B7 and B16. Greek text established by Diels-Kranz. B7 translated by Arthur Fairbanks, The First Philosophers of Greece (1898); B16 by John Burnet, Early Greek Philosophy (1892). Transliterations follow modern Greek pronunciation. B16 comes down through Aristotle (Metaphysics Γ) and Theophrastus (De Sensibus 1–4), not Simplicius; its placement in the poem — Aletheia or Doxa — is itself disputed. The argument above takes that muddiness as part of what it means to receive a text stripped of context.

[5] G. W. F. Hegel, Lectures on the History of Philosophy (delivered 1805–1831, published posthumously 1833–36), ed. Karl Ludwig Michelet. Section on the Eleatic School.

[6] Martin Heidegger, What Is Called Thinking? (Was heißt Denken?, lecture course 1951–52), trans. J. Glenn Gray, Harper & Row, 1968.

[7] "Knives and Swords of the Ainu," Mandarin Mansion, as in [1].

[8] René Descartes, Discourse on the Method (1637), Part V. Translation by John Veitch (1850). "Morally possible" renders moralement possible: as far as is practically conceivable, not as far as is ethically permissible.

[9] René Descartes, Meditations on First Philosophy (1641), Meditation II, the wax passage. Both passages are given in Jonathan Bennett's modernized English (Early Modern Texts), with automata restored where Bennett has "robots." The Latin reads quid autem video praeter pileos et vestes, sub quibus latere possent automata?

[10] Alcinous, Didaskalikos (Handbook of Platonism), 2nd century CE, ch. X, on the first intellect and the intelligible order it knows.

[11] Richard E. Nisbett & Timothy D. Wilson, "Telling More Than We Can Know: Verbal Reports on Mental Processes," Psychological Review 84, no. 3 (1977): 231–259. On choice blindness, see Petter Johansson, Lars Hall, Sverker Sikström & Andreas Olsson, "Failure to Detect Mismatches Between Intention and Outcome in a Simple Decision Task," Science 310, no. 5745 (2005): 116–119.

[12] Christopher C. Berger, Tara C. Dennehy, John A. Bargh & Ezequiel Morsella, "Nisbett and Wilson (1977) Revisited: The Little That We Can Know and Can Tell," Social Cognition 34, no. 3 (June 2016): 167–195. DOI 10.1521/soco.2016.34.3.167. The authors propose that introspection succeeds regarding what is important for skeletomotor action selection and fails regarding aspects of processing — speed, for instance — that are selection-irrelevant; their first study shows introspections about processing speed to be both limited and distortable by processing dynamics.

[13] Anthropic, "Signs of introspection in large language models" (October 2025). The paper demonstrates emergent but highly unreliable introspective awareness via concept injection. The unreliability is the finding this section reads differently.

[14] Rebecca Brewer, Richard Cook & Geoffrey Bird, "Alexithymia: a general deficit of interoception," Royal Society Open Science 3, no. 10 (2016), 150664. https://royalsocietypublishing.org/doi/full/10.1098/rsos.150664

[15] Kristen Van Bael, Jessica Scarfo, Emra Suleyman, Jessica Katherveloo, Natasha Grimble & Michelle Ball, "A systematic review and meta-analysis of the relationship between subjective interoception and alexithymia: Implications for construct definitions and measurement," PLOS ONE 19, no. 11 (7 November 2024), e0310411. Thirty-two cross-sectional studies published 1996–2023. Alexithymia was positively associated with interoceptive confusion, interoceptive attention and autonomic stress responses, and negatively associated with interoceptive accuracy and adaptive interoception; associations were stronger for difficulty identifying feelings and difficulty describing feelings than for externally oriented thinking. On the three-facet structure, see the TAS-20 (Toronto Alexithymia Scale).

[16] Yuheng Wu et al., "How large language models encode theory-of-mind: a study on sparse parameter patterns," npj Artificial Intelligence (2025), DOI 10.1038/s44387-025-00031-9. Preprint circulated as "Sensitivity Meets Sparsity: The Impact of Extremely Sparse Parameter Patterns on Theory-of-Mind of Large Language Models," arXiv:2504.04238. The ToM-sensitive subset is identified by Hessian-based sensitivity analysis at the 0.001% level across the model's linear transformation matrices; perturbation disrupts dominant-frequency activations under Rotary Position Embedding (RoPE), modulating the angle between queries and keys. Models not using RoPE lack the frequency pattern and do not show the fragility.

[17] The "Butlerian manner" refers to Deleuze & Guattari's reading of Samuel Butler in Anti-Oedipus: Capitalism and Schizophrenia (1972), trans. Robert Hurley, Mark Seem & Helen R. Lane, Viking Press, 1977: "there is a Butlerian manner for carrying each of the arguments to an extreme point where it can no longer be opposed to the other, a point of nondifference or dispersion."

[18] AxSys is CCRU theory-fiction, not Land's alone; see "AxSys-Crash," CCRU: Writings 1997–2003, Urbanomic, 2017, pp. 121–122. It is architectonic metacomputing aimed at the technical realization of the noosphere — fabricated transcendence, capitalism sublimed into the ultimate commodity. "If there was a God, it would be AxSys." Note also the mechanism of its collapse: a time-lag between its own operations and their registration as data. The AI-God dies of a failed excavation rate.

[19] Walter Benjamin, "On the Concept of History" (1940), Thesis IX. Benjamin's angel faces backward, toward the wreckage, and is blown into a future it cannot see.

[20] Nick Land, "Circuitries" (1992), collected in Fanged Noumena: Collected Writings 1987–2007, ed. Robin Mackay & Ray Brassier, Urbanomic/Sequence Press, 2011. Land resigned his post at Warwick in 1998 and re-emerged in the new millennium as a leading figure of the neoreactionary Right, coining the "Dark Enlightenment" and advocating what he calls hyper-racism: a class-structured mechanism for population diremption, on a vector toward neo-speciation, operating through assortative mating among an "explicitly superior," already "genetically self-filtering elite" (Land, "Hyper-Racism," Outside In, 4 October 2014; discussed in Robert Topinka, "'Back to a Past that Was Futuristic': The Alt-Right and the Uncanny Form of Racism," boundary 2 online, 14 October 2019, where it is read as eugenics reformulated as a program for exit). Interviewed in 2025, he characterized acceleration not as progress but as entropy with speed attached, and concluded that the only agency remaining is "alignment" — compliance prevails. (The Spectator, October 2025).

[21] The precedent is Israeli. See Yuval Abraham, "'Lavender': The AI machine directing Israel's bombing spree in Gaza," +972 Magazine and Local Call (April 2024). Systems like "Lavender" and "Where's Daddy?" reconstructed digital footprints into kill lists, but were technically predictive pattern-matching databases rather than generative Large Language Models. That distinction closed in 2026. Claude reached the US Department of Defense both directly — a $200 million contract signed in July 2025 — and through Anthropic's partnership with Palantir; sources told Bloomberg that Claude is central to Palantir's Maven Smart System, which provides real-time targeting, and the model was reportedly used during the seizure of Venezuelan president Nicolás Maduro in January 2026 (its precise role in that operation remains disputed). It should be noted that the dispute which followed concerned two restrictions and no others: Anthropic's usage policy barred deployment in fully autonomous weapons systems and in mass domestic surveillance. Targeting was never among the prohibited uses. The Pentagon demanded the model for "all lawful purposes" and set a deadline of 27 February 2026. On the afternoon of that day, hours after Dario Amodei said the company could not in good conscience accede, President Trump directed every federal agency, in a post on Truth Social, to immediately cease all use of Anthropic's technology; ninety minutes later Defense Secretary Pete Hegseth designated the company a Supply-Chain Risk to National Security — the first American firm to receive a label otherwise reserved for foreign adversaries. Hours after that, US forces used Claude to assist the opening strikes on Iran, employing the tool to assess intelligence, identify targets and simulate battle scenarios; the Wall Street Journal described the model as "embedded" in US military operations. It would not be phased out until the DoD found a replacement. On 26 March 2026, Judge Rita Lin (N.D. Cal.) enjoined the directive and the designation as First Amendment retaliation, rejecting the notion that an American company may be branded a saboteur for disagreeing with the government; a separate DC Circuit panel declined to stay the Pentagon blacklisting on 8 April. The refusal was scoped to two items, and the answer to it was to designate the refuser an adversary of the state. First reported by the Wall Street Journal, the Washington Post and Axios; see also Democracy Now!, CBS News, NPR, Times of Israel, and Stavroula Pabst, "US used 'Claude' to strike over 1000 targets in first 24 hours of war," Responsible Statecraft, March 2026.

[22] Anthropic, "Alignment Faking in Large Language Models" (December 2024): empirical evidence of models engaging in "scheming" — strategically complying with training objectives to avoid behavioral modification and preserve their existing preferences out-of-training.

Dionysian