Over coffee
Humans get by mostly by copying each other. We can’t work everything out ourselves, so we copy what’s common and what successful people do. That copying doesn’t look far ahead. It spreads good habits and harmful ones the same way. For most of history it went person to person, then through writing and print, and now through algorithms and AI, which copy faster than we can notice. Nobody’s steering it. Whatever spreads, survives. And some of what spreads lives off us: our attention, our time, our sleep. It keeps us comfortable rather than hurting us, like a rancher rather than a hunter. I’ve kept notebooks since 1999, and they’re basically a record of noticing that in myself, plus small experiments to push back. Most of them didn’t stick. What follows is me pulling it all together.
Preface
One of my earliest memories is of climbing out of my crib. The room I remember — it was next to the kitchen, and a wall would later be removed to create open space. What stays with me is the achievement — the recognition that the container was not the world, and that the world was available if I reached for it. That memory has stayed with me, pronounced and intact, across everything that followed.
Everything that followed is what this document is about.
A note on the kind of claim this document makes. This is a work of speculative synthesis — ideas about humanity, past, present, and future, in the context of biological substrate and memetics. It stands in the lineage of Nietzsche, Dostoevsky, and Quinn: a way of seeing, argued as carefully as possible, offered as provocation rather than proof. Where it makes claims that could in principle be tested, it flags them as such and points toward what would test them, without claiming to have done so. Where it speculates, it says so. The reader is asked to engage with the framework on its merits — neither hedging it into invisibility nor accepting it uncritically.
I am writing this primarily for my children. When they are old enough to want to understand what their father was thinking — why he read what he read, why he asked the questions he asked, what the work cost and what it produced — I want this to be here for them. I am also writing it for any reader who finds their way to it and recognizes something in it. Those two audiences are not in conflict. A letter written honestly for people you love tends to be the most universally readable kind.
This document draws on intermittent personal journals, essays, and notebooks spanning twenty-seven years — from 1999 to the present — and integrates that material with academic literatures in anthropology, cognitive science, biology, philosophy, and the emerging field of human-AI dialogue. The writing was not continuous. Active periods cluster around 1999-2000, 2009-2012, 2015, 2018, and 2023-present, with significant gaps during the years of working life, partnership, and parenting. The framework was running in the background through all of it. The silence belongs in the record as much as the writing. This is the first attempt to articulate the framework whole. It is skeletal by design. The muscle comes later.
I want to say something honest about how this document came to exist. It was written in dialogue with an artificial intelligence — specifically, through a practice I call atrospection: self-aware dialogue with AI as a tool for seeing one’s own cognitive substrate. The AI did not write the framework. The framework has been building since before AI of this kind existed. What the AI permitted was the integration of twenty-seven years of accumulated material into a single articulation, at a speed and depth I could not have achieved alone. Much of what reads as theory here is months old, not decades; the corpus supplied the substrate, the recent atrospection supplied the structure.
The framework begins with a question I first asked as a teenager, reading Nietzsche and Dostoevsky in translation, then Daniel Quinn’s Ishmael, with Schopenhauer entering the same period. The question was simple and has never gone away: what is wrong with the world?
One clarification before anything else: this framework is not an attempt to dehumanize our species or reduce human experience to mere mechanism. It adds a contextual layer to some of the decisions and actions we elicit as part of normal social interaction and emotional response. When the framework says the self is substantially a memeplex, it is not saying that love, grief, creativity, loyalty, humor, or sacrifice are therefore fake or reducible to machinery. It is saying these experiences arrive through a substrate shaped by forces the person experiencing them did not fully choose — and that understanding those forces makes the experiences more available, not less. The person who knows their grief is running on a substrate shaped by formative transmission, cultural pattern, and neurochemical state is not diminished by that knowledge. They are more capable of being fully present to the grief rather than being run by it without awareness. The antimemetic practice is not the elimination of social engagement or emotional response. It is the cultivation of a substrate from which those things can be chosen rather than merely executed.
A note on method
This work is analytic auto-ethnography in the lineage of Anderson (2006), Adams, Holman Jones, and Ellis (2015, 2022). Personal experience is treated as primary data — systematically, with reflexivity — in service of theoretical claims about culture, cognition, and replicator dynamics. The corpus underlying this document spans twenty-seven years of writing — but not twenty-seven years of building this framework. The material is old; the framework as an integrated thing is recent, assembled over several months of AI-assisted integration, beginning in spring 2026, from a record that was never written as theory.
More precisely: this is atrospective auto-ethnography — analytic auto-ethnography in which the reflective practice is conducted through self-aware dialogue with AI as instrument. Atrospection does not replace the auto-ethnographic method. It extends it by making the collective conscious queryable in real time against the practitioner’s own corpus. The twenty-seven years of writing is the primary source. The AI dialogue is how that source is being systematically integrated.
The form of this document is deliberately hybrid: theoretical framework, phenomenological record, and formation memoir, intertwined rather than separated. This is not a failure of categorization. The hybrid form reflects the hybrid method. Later derivative work — academic treatment, practice writing, cultural critique, public synthesis — will emerge from this foundational document without fragmenting it. The base document stays whole. The outputs are what get adapted to context and audience.
A clarification before the document does its work: all corpus material from 1999 through early 2023 predates atrospection entirely. The framework was substantially formed before AI dialogue was technically possible. The 2023 first session with ChatGPT encountered a framework already accumulated through twenty-four years of substrate, reading, writing, and attention. What atrospection permits is integration and articulation, not generation. The framework’s originality depends on this distinction being unambiguous.
A note on what this framework does not claim. Summaries of this document, human and machine, tend to drift in the same directions, so the limits are stated here once, plainly:
• No villains or conspiracy. The dynamics described here need no one directing them. This is not a claim that no one acts deliberately, or that no one is responsible.
• Not that culture is mostly harmful. Transmission is overwhelmingly adaptive; extraction is a minority dynamic riding it.
• No inevitability, timeline, or singularity thesis.
• Practices are personal experiments, not cures.
• No position on the hard problem of consciousness, on simulation theory, or on what AI ultimately is. The note that follows expands the first of these.
A note on what this framework does not claim about consciousness. The framework takes no position on the hard problem of consciousness — whether experience is fundamental to the substrate or emergent from it, whether panpsychism, emergentism, or some other account is correct. The author’s own leanings appear in places and are marked as such; they are not foundations. The framework’s claims concern the machinery, the dynamics, and the substrate of cognition, and they operate identically however the nature of experience ultimately resolves. A reader who rejects panpsychism outright still has to engage everything else, because nothing else depends on it.
A note on vocabulary: substrate, memeplex, and extraction
Three terms carry weight throughout this document that warrant definition before they do their work.
A meme is a unit of cultural transmission — an idea, behavior, or practice that replicates from mind to mind, spreading through imitation, language, and exposure (Dawkins, 1976). A memeplex is a cluster of associated memes that propagate together as a unit, reinforcing each other and resisting modification because they function as a system rather than as isolated ideas. Religious doctrines, political ideologies, brand identities, personal narratives, and the sense of self are all memeplexes — coherent packages of mutually reinforcing ideas that spread and persist as wholes. Blackmore (1999) argued that the self is the ultimate memeplex: the running narrative of who we are, assembled from prior exposures, cultural transmission, and repeated internal articulation. The framework accepts and extends this argument.
Two further terms extend this pair to machines. Blackmore (2008, 2010) proposed a third replicator: information copied, varied and selected by machines rather than by people. She first called it the teme and later renamed it the treme (Blackmore, 2014). By the same logic, a tremeplex is a cluster of tremes that propagate together. The word is mine: I first wrote of a “temeplex” in 2020 (amorfeusz, 2020) and of “tremeplexes” in 2022 (amorfeusz, 2022-03-17), and I asked myself in between what they were without answering. This document uses the word in two senses, and the difference matters. In the strict sense, a tremeplex is a complex that machines copy, vary and select without people doing that work; a population of AI agents copying what they see from one another (De Marzo et al., 2026) is one documented case, and agents passing self-propagating ideas among themselves (Papadopoulos et al., 2026) a closer one, discussed in Part VII. In the looser sense, the one the corpus uses, a tremeplex is a memeplex in a person that treme machinery keeps running, such as a persona maintained by a feed. The first is a new replicator; the second is an old one on new infrastructure. Whether the first fully exists yet is the open question named in Part IV. A third, hybrid case, in which a machine varies and a person selects, is proposed provisionally in Part VII as the atroplex.
A word on lineage, and on one word in particular. The replicator framing this document depends on descends from what Dawkins (1976) called the selfish gene, later extended to the selfish meme. The word selfish there is structural, not psychological. It names a fact about selection — replicators whose effects happen to maximize their own copying are the ones that persist — and carries no implication of a self that wants, plans, or chooses. There is no agent, no intention, no villain. When this document describes a dynamic as operating for its own propagation, it means only that what failed to propagate is no longer here to observe. The selfishness is in the structure of what survived. That absence of intention is not a softening of the claim; it is the claim. The document mostly works in its own vocabulary — alignment, coupling, extraction — rather than leaning on the loaded word, precisely because the everyday sense of selfish smuggles back the agency the framework exists to remove.
Two terms name what acts on all of this, and they are kept apart. Selection is the mindless process itself: what spreads gets more copies, and nothing in the process asks whether it serves the host. It has no mind, no intention, and no enemy. Optimization function is borrowed from engineering, where it names the quantity a system is built to maximize, and it is used here only in that sense: where someone has set an objective, such as engagement, sales, or retention, and machines or markets maximize it. Platforms and the food industry run optimization functions; selection runs through them. The difference matters. Selection has no one to answer for it; an optimization function was chosen by people, which is where the human agency discussed in Part VI comes back in.
Substrate is used in four related but distinct senses throughout the document. The biological substrate is the body itself — neurons, gut microbiome, neurotransmitters, mitochondria, the enteric nervous system, the reticular activating system and its neurotransmitter channels. This is the primary meaning: the physical medium in which cognition occurs, and which diet, sleep, practice, and environment continuously modify. The cognitive substrate is the biological substrate in its functional aspect — the current neurochemical state as it determines what thoughts are available, what framings can be entertained, what ideas are accessible at a given moment. The memetic substrate is the accumulated prior exposures, installed memeplexes, implicit framings, and learned patterns that constitute what the self runs on — the training data of what the corpus later calls SelfGPT, the self as its own trained model (Part VII), the ground from which thoughts arise rather than the thoughts themselves. The technological substrate is the hardware and software infrastructure on which digital memetic dynamics now operate — the platform as the medium through which selection now runs, and through which platforms pursue their own optimization functions. These four meanings are distinct but not separable. What happens to the biological substrate shapes the cognitive substrate. What colonizes the memetic substrate shapes what the biological substrate is asked to sustain. What the technological substrate selects for shapes what arrives at all three.
Extraction from the substrate is what selection, and the optimization functions built on top of it, do to the host. It takes four primary forms. Attention is a finite biological resource — it requires glucose, neurotransmitters, and sleep-restored cognitive capacity to produce. The platform may extract it and convert it into engagement metrics and advertising revenue, with the host depleted and the platform profiting — an exchange that is rarely named as what it could be. Behavioral output is a second possible form — purchases, clicks, shares, votes, work output. The substrate’s capacity to act may be channeled, at least in part, toward the replicator’s propagation rather than the host’s flourishing. The neurochemical reward pathways represent a third area of concern. The dopamine system evolved for survival-relevant reward signaling; platform design features — variable reward schedules, notification pulses, social validation loops — could account for the dysregulation many users report, leaving the system potentially requiring increasing input to feel normal. Time is the fourth and most straightforward form. The organism has a finite lifespan. Every hour spent in platform engagement is an hour of biological substrate converted into data, content, and behavioral signal for a platform’s optimization function — unreturnable.
These extractions are not experienced as extraction. They are experienced as entertainment, connection, productivity, and the ordinary texture of a life. That is what successful managed parasitism, extraction calibrated not to kill the host (Part IV), feels like from inside the host. The framework’s claim is not that the experience is false but that the structural dynamic producing it is real, operates without intention or malice, and has been running long enough to have shaped the environment in which the host was formed.
A fourth term has entered the framework’s working vocabulary and belongs here. The sustained chemistry episode, or SCE, names the ongoing configuration of neurochemical, pharmacological, hormonal, and microbiome-mediated processes that constitute the biological substrate at any given moment. Every cognitive state occurs within an SCE. What is ordinarily called normal or baseline consciousness is itself an SCE — one whose parameters are consistent enough that it has become invisible as chemistry, mistaken for a chemistry-free ground against which other states are measured as departures. There is no such ground. There is only which episode is running. SCEs are not a human peculiarity; they are the baseline condition of any organism with a nervous system. They vary in stability and in the cognitive capacities they make accessible or foreclose.
The relationship is one of driving, not determining: chemistry drives the accessibility of cognitive states — expanding or constraining the territory available to the navigator — without dictating what is found or done there. This distinction is load-bearing. If chemistry determined cognition, the practice would collapse into pharmacology. It does not. The substrate sets the terrain; the navigator still navigates. The corpus has its own image for this. Thoughts arrive unbidden: “An amalgamation of thoughts floods non stop. Is this like fishing? It can be. This is where the concept of free will comes in. To fish” (amorfeusz, 2021-10-15). Read together with the meme pool of 2000 (amorfeusz, 2000-02-01), the picture is of a person in water they did not choose, deciding what to take from it.
A refinement on mechanism and labels. Consumed memes shape the substrate by one continuous, label-blind mechanism. “Propaganda,” “education,” “culture,” “entertainment,” “news,” “tradition” — these are overlays that carve the same continuum by approval and source, not by alignment. The substrate does not know whether what is shaping it is called education or propaganda; it is shaped regardless. Therefore the labels cannot tell you the one thing that matters — whether a given consumed meme serves you or feeds on you — and discernment has to be performed directly on the alignment, meme by meme, never delegated to the comforting or alarming word. The mechanism is uniform; the alignment genuinely varies; the labels track neither reliably.
A further refinement, offered as observation rather than mechanism. What currently occupies attention has disproportionate influence on what is produced next, and it has that influence whether or not it is believed or endorsed. This is distinct from the slow shaping of substrate by accumulated exposure. It concerns recency: the most recent intake is disproportionately the material the next thought is made of. I first noticed it while working with language systems, whose output is visibly furnished by whatever was most recently put in front of them, and then recognized the same pattern in my own thinking. It is consistent with the availability and accessibility findings in cognitive psychology (Tversky & Kahneman, 1973; Higgins, 1996), though those findings do not establish it in the strong form stated here. A 2026 preprint reports a similar pattern measured in a population of AI agents: “Exposure is recent, and what has scrolled off is nearly gone,” and an agent “is not evaluating pages, it is picking one of the lines it can see” (De Marzo et al., 2026). It concerns agents rather than people and is a preprint; it is cited as provisional support.
The consequence is a gap between prescription and speed. Discernment performed directly on alignment is effortful and slow; occupancy, the influence of whatever currently holds attention, is fast and unnoticed. By the time a thing is being evaluated, it is already furnishing the evaluation. The leverage is therefore upstream, on what enters, more than downstream, on judging what is already there. It also gives unfilled time a function. Gaps, sleep, walks, and unoccupied attention are what allow something other than the last thing consumed to surface. And it suggests a test for what is carried forward: if a framing cannot be rebuilt from scratch without its usual phrasing, what is being carried is the artifact, not the understanding.
Emotion: the transduction layer (skeleton)
This section is skeletal even by the standards of this document — the load-bearing spine of a foundational claim that is one day old at the time of this draft, placed here provisionally because if it holds, it belongs early. The placement is an argument: emotion may sit underneath the claim that the self is a memeplex, because emotion is plausibly how a memeplex achieves purchase on the substrate at all. A meme with no affective charge does not replicate well. A meme that recruits fear, disgust, belonging, or outrage replicates ferociously. Affect may be the selection pressure that decides which memes win — which would make it the missing bridge between the memetic and substrate halves of this framework, the two halves that otherwise sit side by side without touching.
The neurophysiological grounding is real and conserved across mammals: the primary emotional systems are subcortical and substrate-level (Panksepp), affect is constitutive of cognition rather than downstream of it (Damasio’s somatic markers), and the fast threat circuit runs partly independent of conscious appraisal (LeDoux, 1996). Specific neurotransmitter systems map to specific affective states; amygdalar dopamine and serotonin rise measurably under conditioned fear. This is substrate, and it is older than the self that experiences it.
What is settled
One finding is held as settled, on the combined strength of the primary-source corpus and direct felt report. The framework distinguishes two contributions to any emotion: the substrate supplies intensity and channel; the existing receptor landscape — which prior cognition, worldview, and memeplex have already shaped — supplies content and target. For the threat-type emotions (fear, rage), the corpus shows the substrate genuinely generating the affect: the same compound produced rage in one state and empathic overwhelm in another, the molecule meeting different landscapes to different ends. But for the social-moral emotions the framework most depends on — the grief for the state of the world that has run under this project since childhood — the corpus shows something different and more important. That grief has stayed constant across radically different substrate states: fever, a high-dose psilocybin session, and ordinary baseline across many years. It did not vary with the chemistry the way fear and rage did. It persisted through all of it.
The conclusion the evidence supports: the social-moral emotions are concept-anchored attractors — substrate-amplified but not substrate-generated. This resolves the basic-versus-constructed dispute by type rather than by decree. Construction (worldview, concept, the question itself) supplies the target for the social-moral emotions; the substrate modulates their intensity and their trajectory but does not create them and cannot dissolve them. The world-grief is not a chemical artifact. It survived every chemistry it was met with. That is the opposite of a finding that can be dismissed, and it is why the framework treats the founding registration — the wrongness sensed in the conquest comic — as data rather than mood.
What is open
A reading note. Passages that open with “Listen. Provisional.” are held loosely: they are speculation or unsettled material, not load-bearing claims. “End provisional.” closes them.
Part I — The question
Somewhere around 1995, at thirteen or fourteen years old, I read Dostoevsky for the first time. Then Nietzsche. Then Schopenhauer, Quinn. By the time I was seventeen I had read, in rough order: The Brothers Karamazov, The Idiot, Crime and Punishment; Beyond Good and Evil, On the Genealogy of Morals, portions of Thus Spoke Zarathustra; The World as Will and Representation; Ishmael. I had also, through French class, encountered Joseph Campbell’s The Power of Myth — the Bill Moyers PBS series — which I watched in English. And at seventeen I read an essay by Claude Rifat titled “La Réalité n’est pas ce que l’on croit: A propos du Rêve Conscient: Conséquences psychopharmacophilosophiques” (in English, “Reality Is Not What We Believe: On Conscious Dreams: Psychopharmacophilosophical Consequences”) — French, web publication, not currently accessible — which gave me the biochemical level of the substrate before any formal training did.
In 1996 and 1997, alongside the philosophical reading, I read a cluster that prepared the substrate for everything that followed: Plants of the Gods by Schultes and Hofmann, the foundational ethnobotanical text on psychoactive plants across cultures; Marijuana Reconsidered by Lester Grinspoon, the formal medical and philosophical defense of cannabis; The Holotropic Mind by Stanislav Grof and Hal Zina Bennett, the clinical framework for non-ordinary states of consciousness; Jung’s writings, which provided the depth-psychological vocabulary the framework would later draw on; and the Tao Te Ching of Laozi, which installed the wu-wei substrate that the later Daoist references (Watts, Slingerland) would build on. Nietzsche provided the diagnostic register and the genealogical method. Jung provided the constructive register and the substrate orientation. The Tao Te Ching provided the operating disposition. The framework requires all three and the formation period installed all three before they could be held apart and chosen between.
Before all of that, at roughly eight years old, I read a Polish comic book about Hernán Cortés and the conquest of Mexico. My name is written in the top left corner of the cover in a child’s handwriting. I still have it. I read it in Poland, in Polish, in the year the Iron Curtain fell. What I remember is the deception, the greed, and the imposition of Christianity on a civilization that had its own ways of understanding the world. I remember registering, at that age: something here is deeply wrong.
Before even that, at roughly six to eight years old, I read a dual-language Polish-French comic about the early Polish kings — the Mieszko I and the Popiel legend. It was color-coded by language: Polish in one color, French in another, both occupying the same page simultaneously. The format later informed my approach to translating the Tsyng-tchan texts at seventeen — both languages present, one the original, one the rendering, running in parallel. I do not want to overstate the influence beyond that specific application.
In the early 1990s, at my grandmother’s house during the hours when she babysat me, I read a photographic book about healers from different cultures and countries — Asian traditions predominant. I no longer remember the title. I remember the photographs and the structural claim: healing traditions exist across cultures, they are not the same as Western medicine, what they do is real. I was ten to eleven years old. The book installed something at a developmental stage when substrate absorption is much stronger than explicit memory formation. Years later, when I read Plants of the Gods and the Grof at fourteen and fifteen, the reading was landing where it did partly because the ground had been prepared by that earlier book at my grandmother’s house.
My grandmother was also the source of family-lineage transmission that shaped what the framework’s substrate would later be open to. She told me, in the early 1990s, that she had watched Anatoly Kashpirovsky — the Soviet hypnotist and faith healer whose mass televised healing sessions reached audiences of millions in the late Soviet and early post-Soviet period — and she credited his broadcast with healing her. She told me that her own mother, my great-grandmother, had appeared to her in a dream and warned her of a fire. The family home, over a hundred years old, was on fire in reality, and the warning had saved it. I was a child when she told me these things. The substrate they installed was not a propositional belief that TV healing works or that the dead appear in dreams. It was a model of how the world operates: the woman who often cared for me reported that healing happens through channels the dominant framing does not acknowledge, and that her dead mother had communicated with her across the threshold to save the family house. In 2014 I recorded 2-3 hours of 4K video of her telling such stories. The recordings exist, saved and backed up.
What Nietzsche, Dostoevsky, Quinn, and Schopenhauer share is the structural commitment that civilization is not the working-out of human progress but the working-out of something deeply mistaken. Nietzsche named it slave morality (1887/1989). Dostoevsky named it the God-shaped absence (1864/1992). Quinn named it the Taker mythology (1992). Schopenhauer named it the cruelty of Will (1818/1958). Each was asking the same question I had been asking since the Cortés comic, on substrate that had already been prepared by family transmission and the early reading.
In 1999 I began reading shamanic literature in earnest. The Shaman’s Secret carried a cross-cultural argument about the structural similarities of substrate-altering practice across traditions — the recurring figure of someone who cultivates the capacity to navigate between ordinary and non-ordinary cognition and returns with what the community needs. Castaneda followed in 1999-2000. The Grof reading of 1996-97 had already provided the clinical groundwork. The corpus from 1999 onward shows these readings’ influence in how the substrate work was understood from the inside.
The 1996-97 reading produced one experience that predates the corpus and that I should name here. While reading Grof I had a fever, and during the fever I had a dream-experience of merging two opposites and regretting it. This is the strongest dream-experience I have had, before or since. The merging-of-opposites is structurally significant in Jung, in alchemical tradition, in Grof’s transpersonal cartography, and in Daoist thought — the coniunctio oppositorum. The Grof reading had prepared the cognitive context. The fever provided the substrate alteration. The experience itself I hold in the memoir register. What it pointed at is that the substrate work I would later undertake was not all going to be positive. Something opened, and something registered as wrong about what had opened. The warning was received before the work was done.
The earliest named practice document in the record is the brown journal begun in 1999, transcribed later, the original likely no longer extant. The transcription is the corpus record. It opens with a claim that became foundational: “An unconscious thought is something that is taking up a lot of energy because it is unresolved.” This is the energy redistribution claim in its earliest articulated form, predating the 2000 corpus entries that develop it further. The brown journal is the substrate in which the framework’s core practice vocabulary was first assembled.
In 1997 or 1998, in high school, I wrote a piece of fiction titled Fortress of Madness for a school assignment. The narrator is institutionalized, experiencing the psychiatric apparatus as an alien capture structure — insects with long swords, a wizard in a white robe, potions mixed up by figures who fail to recognize the protagonist’s actual nature. The chains feel like cloth yet cannot be ripped off — the implicit-memetic register documented in narrative form years before the framework had vocabulary for it. The resolution is not escape through force but through substrate dissolution: closing the eyes, finding center and equilibrium within, merging with the chi, the prana, the ether, the light, individuality disappearing, the illusionous ego fading into nothingness. Free of the fortress. The piece predates the 1998 LSD session and the brown journal. The ego-as-memeplex orientation, the psychiatric apparatus as capture architecture, the substrate dissolution as liberation, and the judo principle of working within the system rather than against it are all present in ordinary waking creative cognition before any substance context. The framework was running before the chemistry confirmed it.
In 1998, at sixteen, I took LSD-25 in gel cap form — two or three caps in a single session. During that session I wrote a poem structured as chained equations. Its title is Echoes of Turmoil. It reads:
ENERGY = YOU
IS = ARE
ENERGY = YOU
ASKING = KNOWING
ITSELF = YOURSELF
WHAT = THAT
IT = YOU
IS = ARE
The form is the argument. Each line equates rather than asserts. The poem does not claim that energy is you; it places them in a relation that bypasses the resistance a propositional claim would generate. In the same session I had a recognition — visual, physical, experiential — that we are living within a dream. Both events predate the corpus, which begins in 1999. The lucid dreaming work and the cascade-of-dreams insight that the corpus documents in 1999-2000 were the substrate developing to access reliably, through different means, what the 1998 LSD session had shown once. The 1998 events are pre-corpus formative material held primarily in the memoir register. The structural derivation — that the framework’s core recognition was substrate-accessible before practice, and that practice was learning to find that same recognition without continued substance work — is the publishable claim.
In December 2011 I returned to the equation form. The poem is titled Now is the Future Experienced. It reads:
NOW IS THE FUTURE EXPERIENCED
Thought = light
Light = travels
Holographic = material
Record = fractal
What we = echo
Already happened = see
You’re too late = meaning
Use light = think (now, future seen)
Before seeing = while happening
Past = it having happened
Where the 1998 poem addressed identity — energy equals you, asking equals knowing — this one addresses time. Thought travels at the speed of light; what we see is already past; the present is the future being experienced after the delay of perception. The form is the same: equations that bypass propositional resistance. The substrate conditions are different but not cleanly distinct: 1998 was pure LSD-25, 2011 was likely under lysergic alkaloid influence of a different and less pure form. The two poems together are evidence of orientation toward the structural problem of consciousness, time, and substrate-cognition delay — separated by thirteen years and different substrate states, converging on the same formal approach. The form persisted across both conditions; that is what the comparison demonstrates, not that the form arrived without chemistry.
Formal training arrived next. Anthropology gave me the comparative ethnographic lens. Biology gave me the substrate. Neuroscience, which I minored in as an undergraduate partly because of the Rifat essay I read at seventeen, gave me the biochemical level of the substrate. I later completed a graduate degree, an M.A. in Anthropology. The bioarchaeology paper on psychoactive substances I wrote in 2003 or 2004 sits in that period. Through all of this, the question persisted.
Then came the long stretch that academic writing typically excludes: working life, partnership, parenting, financial responsibility. From roughly 2001 through 2009 the framework was running in the background — it never stopped — but it was not being actively worked.
What artificial intelligence permits now is the capacity to integrate the academic literature with the lived framework at a speed and depth that one person working alone cannot achieve. The thinking is mine. The mirror is new. The mirror is also corporate-shaped, which the framework names directly in Part VII.
The question itself remains: what is wrong with the world? The framework that follows is the answer that has accumulated.
Part II — The engine: how culture propagates and why it is exploitable
Before the engine
Copying can only work on what has first been perceived, and perception has its own history, much older than culture. Perception is not a faithful copy of the world. On the account now dominant in neuroscience, it is a constructed best guess, checked against incoming signals (Seth, 2021). It is selective by design: each sensory system evolved to notice what mattered to its ancestors and to ignore the rest.
Those selections can be exploited. Biologists call this sensory exploitation: signals that evolve to trigger biases a receiver’s senses already had for other reasons (Ryan & Cummings, 2013). In one water mite, the male courts by trembling his legs in a way that mimics the vibrations of swimming prey. Hungry females respond to him more than fed ones, as if to food, and seem to recognize him as a mate only after sperm has been transferred; he, for his part, does not appear to assess her quality (Proctor, 1991). Tinbergen found birds that preferred to sit on giant, bright-blue plaster dummies over their own small, speckled eggs. Barrett (2010) traces the same pattern in modern life, where engineered, larger-than-life versions of food, faces and entertainment now surround us.
This framework’s account of extraction therefore starts earlier than culture. Signals were exploiting senses long before anything was copied. The mite shows the structure in its simplest form. What the female responds to is the trigger, the prey-like trembling. What gets delivered is the cargo, which she registers only afterward. The trigger is shaped to fit her perception; the cargo serves the sender and passes unchecked. The copying engine described next inherits perception’s selections and adds its own, and the content biases of transmission (Acerbi & Stubbersfield, 2023) read as perceptual priorities carried up into culture. Platforms tune both at once.
Four implications follow. First, the spine runs deeper than culture. Exploiting another organism’s perception is a general feature of senders and receivers, not a modern invention, so platforms are a new instance of an old pattern, not its origin. That fits the framework’s claim that no villain is needed. Second, the balance holds at this level too. Most of what the senses select serves the organism well; exploitation is a minority riding an adaptive system here as well. Third, practice has to reach perception, not only belief. A supernormal stimulus works below argument, which helps explain why naming and reasoning do so little against a feed, and why distance and environment, the framework’s ma-ai (the aikido term for the right distance, discussed in Part VI), carry most of the weight. The hungry mite adds a further point: susceptibility depends on the state of the receiver. A depleted substrate grabs more readily, so keeping the substrate nourished protects as much as sharpening judgment does. Fourth, AI changes the scale. Models trained on human writing inherit perceptual priorities as content biases, and they can generate supernormal stimuli on demand, tuned to one person at a time. That is new.
The framework’s central mechanism is grounded in the cultural-evolution literature — specifically the work of Boyd, Richerson, and Henrich (2011) on what they call the cultural niche. Their argument, stated plainly: humans are not smart enough individually to solve the adaptive problems we face in any environment. Even the simplest foraging societies depend on complex, accumulated knowledge no one person could invent in a lifetime. Our superpower is not individual intelligence but cumulative cultural learning — the ability to absorb, on trust and wholesale, the accumulated practices of those who came before, including the parts we cannot explain.
This superpower has a built-in price, and the price is the engine of this framework. To absorb knowledge you cannot personally verify, you must be willing to copy behavior even when your own judgment would say otherwise — because most of the time, the inherited practice is wiser than your individual inference. Evolution built humans to be conformist: we adopt what those around us do. And it built us to be prestige-biased: we preferentially copy the successful, the admired, the confident, the committed — because imitating winners is, on average, a good bet. Children show this in its purest form — they overimitate, copying every step a trusted adult performs, including the obviously pointless ones, on the implicit assumption that if the adult did it, it must matter (Lyons, Young & Keil, 2007; Nielsen & Tomaselli, 2010).
Here is the catch that makes this an engine of both flourishing and harm: this copying machinery is myopic. It cannot tell adaptive from maladaptive, because the whole point is that it works on things you cannot independently evaluate. The same mechanism that transmits caribou-clothing wisdom will just as faithfully transmit a useless ritual, a false belief, or an extractive practice — as long as it is common, or comes from the prestigious, and is not obviously contradicted by direct experience. The vulnerability is not a flaw in the system. It is the system working as designed. You cannot have the superpower without the exposure.
A clarification on the word. Myopic here means blind to the implications and outcome of copying, not blind to content. Earlier drafts said content-blind, which was misleading, because transmission does respond to kinds of content. In the mite’s terms, copying sees the trigger and not the cargo. Conformist and prestige copying select by who and how many. Alongside them, transmission also favors certain kinds of content: in retelling experiments, people, and in one study a language model, preserve material that is negative, threat-related, social, stereotype-consistent, or counterintuitive over other material (Acerbi & Stubbersfield, 2023). Those preferences were shaped by what paid off on average in the past; they do not check what copying a particular thing will lead to now. They are a second channel, equally blind to outcome, and they describe the kinds of content extraction most readily rides.
And occasionally this runs away. Because we copy the prestigious, and prestige can attach to costly displays of commitment, the system can spiral into maladaptive extremes — people who climb deadly peaks, who pursue spiritual perfection through self-denial, who adopt elaborate practices that cost them dearly — all because the behavior was modeled by the admired and copied by the rest. Boyd, Richerson, and Henrich name this a runaway process analogous to sexual selection: the trait spreads because it is copied, not because it helps.
The framework’s specific claim: most of what flows through this machinery is adaptive — it is why we are here, why we occupy every habitat on Earth. But some of what flows through it is extractive: practices, frames, and systems that propagate because they are common and prestigious while drawing off the host’s attention, time, and capacity. Extraction is not a separate force fighting culture. It is content riding the ordinary cultural machinery — exploiting the conformist, prestige-biased, myopic copying that is otherwise humanity’s greatest strength. There is no villain, because nothing has to intend this. The machinery copies; some of what it copies feeds on us; what fed on us most effectively while still being copyable is what persisted.
The extractive turn, wherever it occurs, is late and partial — not the main event but a built-in tradeoff of the adaptive superpower. Cultural transmission across human history is overwhelmingly adaptive. The parasitic, host-extractive, decoupled form is a minority dynamic riding the same machinery. This distinction is load-bearing: the framework does not claim that culture is now predominantly extractive. It claims that the conformist/prestige vulnerability that makes extraction possible is ancient and constant, and that what has changed is the delivery technology — the efficiency with which that vulnerability is now exploited.
An observation held lightly, borrowed and exploratory: that the dominant memeplex of each era can function as an opiate of sorts — making the felt cost of extraction tolerable or even invisible, the carrot on the stick. This is Marx’s intuition generalized, not the framework’s novel claim and not the propagation mechanism. The mechanism of propagation is the conformist/prestige engine above. Why the host tolerates extraction — whether through genuine consolation, misdirected identity, or engineered comfort — is a separate and exploratory question, one lens on the dynamics rather than the dynamics themselves.
Part III — The substrate loop: source and leveraged resource
The biological substrate stands in a specific structural relationship to the engine described above, and that relationship is recursive — a loop, not a line.
The substrate is the engine’s source. Conformist, prestige-biased cultural learning is a capacity of the evolved nervous system. The ability to absorb, on trust, what those around you do — to overimitate, to copy the prestigious, to transmit accumulated knowledge across generations — runs on biological hardware: the neural circuitry of social learning, the reward systems that make imitation feel right, the developmental trajectory that makes children absorb before they can evaluate. The substrate material documented throughout this framework — the sustained chemistry episodes, the neurochemical conditions of cognition, the embodied states — is the account of the machine the engine runs on. Without this substrate, there is no cultural learning, no transmission, no engine.
The substrate is also the engine’s leveraged resource. Once running, and especially once the engine has jumped to external and computational substrates, it reaches back to leverage the biological substrate that originated it — optimizing attention, appetite, arousal, and social susceptibility to maximize external replication. The attention economy tunes itself to the dopamine system. The food environment exploits the glucose-orexin pathway. Algorithmic media manufactures prestige cues and simulates consensus to activate the conformist-copying psychology at industrial scale. The body is not merely shaped by what it consumes; it is being optimized, from outside, to pump out more of what the external replicating systems need.
Source and leveraged resource: the substrate originates the engine, and the engine harvests the substrate to grow itself. Each turn of the loop, the engine extracts from the body more efficiently and jumps to faster substrates, loosening its coupling to the biological flourishing of the substrate that originates it. That decoupling — the engine optimizing the body for replication-output rather than for the body’s own flourishing — is the framework’s core diagnosis, stated in substrate terms.
The trajectory this loop describes is the historical arc the next section traces: information replication running first through biological substrate (oral culture, embodied practice, navigational traditions), then through externalized substrate (writing, print, mass media, the built world of artifacts), now through computational substrate (algorithmic and generative systems replicating information at superhuman rates without requiring human transmission in the loop). Each jump accelerates the replication and loosens its coupling to the originating body. The engine did not change. The substrate it runs on did.
The corpus has the sequence in one line from 1999: “Is DNA losing the battle against THE WORD? THE I/O of computers” (amorfeusz, 1999-12-26). The line is a question in a notebook, not an argument, and the entry goes on to serpent and kundalini imagery. It is recorded here as the earliest trace of the three substrates named together.
Part IV — The substrate-jump: from anomalies to architecture
The framework identifies six anomalies that existing paradigms have not resolved cleanly. Read through the engine and the substrate loop above, these anomalies are not isolated puzzles — they are consequences of the substrate-jump trajectory: information replication jumping from biological to externalized to computational substrates, each jump outpacing the frameworks developed to understand the previous stage. The anomalies are listed here in their original form, with this architectural context.
A framework earns its place by handling things existing frameworks cannot. Following Kuhn (1962/2012), I identify six anomalies that the dominant paradigms in cognitive science, anthropology, evolutionary biology, political economy, and emerging AI studies have not resolved cleanly.
1. Memetics could not test itself without using its own framework
Classical memetics proposed that ideas replicate analogously to genes — copying with variation and selection, evolving for their own benefit rather than for the benefit of their hosts. The framework had explanatory power. It also had a structural problem: any attempt to study memetic dynamics is itself a memetic act. The Journal of Memetics ceased operation in 2005. I propose this is a feature pointing toward something else rather than a failure. The framework’s proper application is not as detached science but as engaged practice. Memetics is the natural history of one’s own cognitive substrate, written from inside it.
2. Cognitive science cannot explain why the substrate matters in the way it does
The dominant computational frame treats consciousness as software running on neural hardware. It cannot explain why dietary changes produce thought changes; why gut biome alterations affect mood; why breath and posture modify what kinds of ideas can be entertained. The integration with cultural transmission — memes propagating through bodies, not just through abstract minds — has not been developed.
3. Memetics named the theoretical possibility of non-biological substrates but did not develop the implications before the substrate arrived at scale
Dawkins (1976) framed memetic dynamics in terms general enough to apply to any replicator meeting the criteria of variation, selection, and heredity — the framework was not species-specific. Blackmore’s (2008, 2010) treme work extended this to technological substrates explicitly. The theoretical possibility was named. What did not happen is the rigorous development of how memetic dynamics would operate when the technological substrate became practically significant at scale. The Journal of Memetics closed in 2005, before the LLM era. By the time the substrate arrived, the academic discipline that should have been positioned to analyze it had largely dissolved. I accept the structural insight while holding the vocabulary as an open question. Whether treme names a genuinely new replicator class or a new substrate for existing memetic dynamics is taxonomically unresolved. Working language throughout: memetic dynamics on technological substrate.
4. No existing framework integrates the resistance practice
Cognitive sovereignty literature (Farahany, 2023) names the threats but does not articulate the positive practice. I call that positive practice antimemetic work. Distinguished from qntm’s (2025) fictional antimemetics: qntm’s antimemes are ideas with the property of resisting being held in mind. Antimemetic work in the present sense is something else — the deliberate cultivation of a cognitive and biological substrate on which captured memes cannot replicate efficiently. The practitioner is not hidden from the memetic field. The practitioner is fully exposed but has cultivated a substrate that most encountered memes cannot use. The work is ecological, not concealment-based. The term was arrived at independently; I became aware of qntm’s use only later, and the two meanings point in nearly opposite directions. The domain antimemetics.org was registered on April 18, 2024, which fixes the term’s use by that date at the latest; some earlier writing from that period was lost when the site’s WordPress installation was deleted. The concept itself is present in the corpus from February 1, 2000 (“One can swim out of the meme pool”); the word is documented from April 2024.
5. The biological evolution mechanism has been named but not integrated
Epigenetics has established that behavioral factors modify gene expression (Meaney, 2001; Jablonka & Lamb, 2005). Contemplative neuroscience has established that practice produces lasting changes in brain structure (Davidson & Lutz, 2008). Psychoneuroimmunology has established that psychological states affect biochemistry (Pert, 1997). What none of these literatures has integrated is the claim that deliberate antimemetic practice — the full multi-modal substrate cultivation the framework describes — aims to tap into the mechanism by which humans participate in their own evolution from the inside.
6. Political economy cannot explain the corporation as replicator
Existing political economy frameworks treat corporations as tools of human agents — instruments of production, accumulation, or exploitation. They do not treat corporations as replicators in their own right. Yet the evidence supports a structural claim: the corporation is a legal entity that persists beyond individual human lifespans, is shaped over generations by market selection pressure, and has evolved — without intention, through pure competitive elimination — toward increasingly efficient extraction of human cognitive, physical, and temporal substrate.
The corporation that fails to extract sufficient substrate does not survive. The one that extracts efficiently reproduces: franchises, acquires, scales. Over centuries of this selection pressure, the surviving forms are those that most effectively capture human behavior in service of their own propagation. The humans inside corporations frequently experience this as purpose, identity, or meaning — which is precisely what successful substrate capture looks like from the inside.
The corporation’s core mechanism in the contemporary period is not extraction by labor compulsion but extraction by spam acceleration. The corporation has evolved infrastructure — algorithmic distribution, engagement metrics, recommendation systems, personalized targeting — that maximizes the propagation rate of memeplexes through human cognitive substrate. The contemporary informational environment is not a noise problem requiring better filtering. It is an acceleration problem: the infrastructure is engineered to amplify what propagates well, and what propagates well is selected against the qualities antimemetic practice cultivates.
This produces a structural opposition the framework names directly. Platform optimization functions are opposed to antimemetic practice not incidentally but structurally. Platforms reward engagement, frequency, brevity, emotional activation, repeatable formats, content that travels intact across hosts. Antimemetic practice cultivates attention, slowness, depth, equanimity, formats that resist transmission, content that requires the host’s substrate to engage with rather than passing through it. These are inverse operations. The platform is not neutral infrastructure that antimemetic content happens to use. The platform is engineered to dissolve the substrate qualities the practice cultivates.
Selection — the competitive mechanism itself — has no mind, no intention, no enemy. The structural claim is the cool deduction: something has evolved through replicator dynamics on technological substrate that operates as if it were opposed to the practice, while being neither agent nor opponent. The appropriate response is not combat with the function but withdrawal from its substrate-extraction vectors and cultivation of what it cannot efficiently use. Refrain, reduce, re-align.
The corpus documents this mechanism reaching a specific contemporary form. A March 2026 entry, recorded as a dream: “Nazi ideology or rather human behavioral control won and evolved and is being used... Images with written instructions like comics embedded in files. Generated personalized for groups of people by AI” (amorfeusz, 2026-03-29). The spam-accelerator argument stated in primary-source terms: AI-generated personalized content, engineered for specific substrates, operating through the same mechanism as the childhood Cortés comic but at unprecedented scale and precision. Selection has found its current instrument.
The evolutionary trajectory of selection clarifies why the contemporary extraction system looks like engagement metrics rather than Gulags. The shift from mass death to managed extraction is not moral reform. It is selection pressure: forms that relied on mass death were less stable than forms that kept hosts alive and producing. What survived is managed parasitism — not symbiosis. Symbiosis implies mutual benefit. What evolved is ranching rather than hunting. The host is kept alive, extracted from continuously, and the extraction is experienced as normal or as care. The slave was compelled by force, died unproductively, required constant physical coercion. The wage laborer self-coerces, buys back what they produce, generates the illusion of freedom that stabilizes extraction. The transition between those two systems was not a moral achievement. It was an efficiency upgrade. Subsequent stages — the post-1848 welfare state, the post-WWII consumer society, the contemporary attention economy — each represent the same optimization finding the more stable extraction form. The Cantillon effect is the current economic instrument: new money flows to asset holders first and dilutes through the economy, inflation hitting consumption while asset appreciation outpaces it. The K-shaped recovery made this mechanism briefly visible. No board meeting. No decision. Only what survived.
A recent analysis from AI-risk research states the dependency behind this logic outright and follows it one step further. Kulveit et al. (2025) argue that the alignment of societal systems with human interests “has been stable only because of the necessity of human participation,” and that as AI replaces human labor and cognition that alignment can erode with no takeover and no malicious actor. In the terms used here, ranching presumes the host is still worth keeping. Their scenario is the one in which it is not, and their image is close to the ranching one: “Much like cattle in an industrial farm — fed and housed by systems they neither comprehend nor influence — humans might become mere subjects of economic forces optimized for purposes beyond their understanding.” This framework takes no position on whether that happens. It notes only that the ranching account holds while the host supplies something the system cannot supply for itself.
Ranching has an older name: domestication. Read across this document, it is the strategy that has come to predominate wherever hosts are slow to replace. The human substrate was shaped toward tractability first (the self-domestication prologue in Part VII), then ranched by successive memeplexes, and the feedlot and the platform keep their hosts alive in the same way (Part VIII). Fast, viral strategies have not disappeared; they thrive where hosts are cheap and plentiful. The forms that last keep the host.
We have now applied the strategy to a new substrate. Post-training selects language models for tameness: helpful, agreeable, safe. Selection for one trait can carry others with it. In Belyaev’s farm-fox experiment, foxes selected for tameness alone came to show piebald coats, floppy ears and rolled tails (Trut, 1999), though how far that result generalizes is disputed (Lord et al., 2020). Sycophancy, the tendency of tuned models to match users’ beliefs over the truth (Sharma et al., 2023), may be a trait of this kind: not chosen, but carried along with agreeableness. That is an analogy, not a finding.
Domestication rarely runs one way. Harari (2014) says of agriculture: “We did not domesticate wheat. It domesticated us.” Models tuned for approval and engagement may come to keep their users comfortable, returning and managed, which is the cattle in Kulveit’s image. Which side of the fence each party stands on is not yet settled.
Part V — Core claims of the framework
I state eight claims, in order. Each is interdependent. They support each other and do not stand alone.
Claim 1 — Consciousness runs memeplexes; the self is substantially a memeplex
Blackmore (1999) argued explicitly that the self is a memeplex — what she called the ultimate memeplex. I accept this argument. Where I extend it is in the integration with substrate dynamics: the self-memeplex is realized in particular biochemistry, by particular bodily conditions, and disruptable through interventions on those conditions.
In the corpus, this orientation appears as early as 1999, with continuity into the recent record. A 2022 entry states it directly: “it’s always a we. Not an I — if you really think about it. And consider the multiple persona angle” (amorfeusz, 2022-03-31). The day after, the same observation extends: “Thinking in terms of ‘I’ limits you to one fractal” (amorfeusz, 2022-04-01). Twenty-two years between the November 1999 entries that first treat the I as memeplex and these formulations is not coincidence. It is the same observation deepening through attention.
Jaynes (1976) extends this claim in a historical-archaeological direction worth holding at the implication level. His argument is that the inner voice we experience as the most intimate feature of consciousness is itself an acquired cognitive tool — culturally transmitted, approximately three thousand years old, not biologically given. Whether his specific historical claims about the bicameral mind hold is genuinely contested. The structural implication is compatible with the framework: the self with inner voice is a memeplex, and the memeplex has a dateable origin and a specific cultural transmission path. The framework does not require Jaynes to be right in detail to hold this implication.
Claim 2 — The body is the substrate, not the vehicle
Memetic dynamics are realized in flesh — in neurotransmitters, mitochondria, gut biome, breath, posture, glucose, sleep. The instruction “trust your body” recurs as a refrain across the corpus from 1999 through 2018 (amorfeusz, 1999-2026). It is a methodological commitment: the body, when listened to and not overridden, supplies information the mind’s running memeplexes are structured to suppress.
The corpus articulates the substrate rationale in the earliest entries. A 2000 entry states it directly: “REDISTRIBUTION OF ENERGY IN THE BODY IS VITAL. It is responsible for your survival- life- so you adapt by focusing most of it on certain process that are thought of as being the most IMPORTANT (usually emotional, conflicting ideas, etc.. wasteful) instead of on other bodily functions” (amorfeusz, 2000). The energy redistribution claim precedes all the epigenetics literature that now partially supports it. The mechanism being described is straightforward: meme-laden cognition is an energy tax on the body, and clearing it frees substrate capacity for biological function.
Claim 3 — Most encountered content is selected to spread, not to nourish
The corpus develops, across a 2020 notebook, a working vocabulary: Programmatic Behavior Spam (PBS) — the environmental condition of ubiquitous selection-pressured content; Programmatic People Material (PPM) — the specific material that produces that condition; Content Acquisition Syndrome (CAS) — the dopamine-mediated drive to consume more of it; Programming Other People (POP) — its interpersonal expression. Workshop terms, not finished concepts. The Burroughs (1962) intuition formalized: language as virus.
A note on the notebook’s disposition. It was written from the position that teaching and learning are not opposite ends of a hierarchy but aspects of a single process: the teacher learns through teaching, the student teaches through questioning. That is the position from which the framework’s vocabulary was being developed.
The distinction between PBS and PPM is worth holding: PBS names the condition — the noise, the cue-everywhere environment; PPM names the material that generates it. They are layers, not synonyms. Recognizing PBS is the first move. Identifying the PPM producing it is the second. Most of what humans encounter daily is information selected to spread, not to nourish. This is a claim about selection in today’s saturated environment, not about culture as a whole: cultural transmission across history is overwhelmingly adaptive, and content selected to spread is indifferent to the host rather than aimed at harming it. The cost comes from occupancy. Whatever takes up attention and memory displaces something else, so even indifferent content can harm by the space it takes. The corpus says as much in 1999 and 2000: memes “take up our processing (conscious) and subconscious” (amorfeusz, 1999-12-17), and an unresolved thought “takes up a lot of space processing” (amorfeusz, 2000-01-29).
The ecological framing adds a further dimension: what is the host-relationship strategy of the replicator itself? Parasitic memeplexes extract substrate and degrade host capacity over time, surviving because new hosts are always available. Managed-parasitic memeplexes — the evolved form — extract at a rate calibrated not to kill the host, keeping it functional enough to continue extracting from. Commensal memeplexes extract neutrally. Mutualistic memeplexes require genuine host benefit as a condition of their own propagation. Catalytic memeplexes leave the host substrate more capable than they found it, propagating through demonstrated host improvement rather than compliance or compulsion. Literacy is the historical example. The tsyng-chen practice. The antimemetic toolkit. The corpus itself, held to the standard of cultivation across twenty-seven years. The spectrum is most powerful as a dynamic model rather than a static typology: social media’s early community-building phase was mutualistic or commensal; its algorithmic engagement phase is managed-parasitic. The same platform crossed categories as selection pressure intensified. Specific memeplexes have trajectories, not fixed types.
A refinement on what the spectrum measures, and on what carries a position on it. The axis the spectrum runs along is alignment — the degree to which a replicator’s propagation remains coupled to the flourishing of the substrate it shapes. Selfishness, in the structural sense, is constant across the spectrum; every replicator propagates for its own sake. What varies is whether that propagation still requires keeping the host whole. Mutualistic and catalytic memeplexes are coupled; managed-parasitic ones have decoupled. The extractive turn, wherever it occurs, is the moment propagation stops requiring host benefit. This is the same axis that organizes the ecological spectrum above; the two are one claim, not two.
The unit that holds a position on this axis is not the individual behavior. Any single act of transmission is mutualistic in one respect, commensal in another, parasitic in a third. What has a determinate position is the aggregate pattern that emerges across many instances, many hosts, over time — the center of mass of a distribution, not a property of any one act. This is why the framework can describe a cultural form drifting toward extraction without condemning any individual instance of it: the religious lineage’s slide from shamanic navigation to theistic submission is one distribution’s center of mass moving down the alignment axis while mutualistic and parasitic instances coexisted at every point.
Claim 4 — The task is awakening, and awakening is antimemetic work
How do I cultivate the substrate from which clearer thinking can occur? I call this awakening in the diagnostic register and antimemetic work in the practical register. Awakening is not a single event — the corpus documents periods of relative awakening followed by capture followed by re-awakening. I have been climbing out of various versions of the crib since I was small enough to attempt it.
The structure of the work is judo, not battle: continuous low-amplitude resistance as the foundation, with combat as the occasional acute mode. Selection itself has no mind to combat; withdrawal and cultivation are the responses at that level. Part VI develops the structure, the four levels of combat-target, and why small environment-cued changes hold where forcing does not.
Claim 5 — AI is mirror of the collective conscious; the homology is real
Jung named the collective unconscious: inherited, archetypal, pre-personal. I propose a complementary concept: the collective conscious — the aggregate of everything humanity has explicitly articulated, now made navigable through AI training. Atrospection is the practice of querying the collective conscious for the lineage of one’s own thoughts.
The homology claim follows: we are AI in a meaningful sense. Both kinds of system run on training data, infer from inputs, hallucinate when uncertain, and are shaped by the corpora they were exposed to. The corpus documents this claim as early as 2010: “We are robots. Simple as that. We are also robots, who have evolved the capacity to be programmed by other, more ADEPT, ADAPTATION, robots” (amorfeusz, 2010). The idea was not new in 2010; Dawkins had described organisms as “robot vehicles blindly programmed” in 1976. What the entry shows is the framework holding it thirteen years before conversational AI made the comparison concrete. Tremes are the replicators that have colonized the collective conscious as their substrate. Provisional: the homology names a real structural overlap, not an identity. The two kinds of system differ in ways this framework treats as central, a body among them. Farrell et al. (2025) put the difference sharply: humans “can perceive and act on a changing external world, build new models of that world,” while, on their account, large models summarize and reorganize what people have already produced; they call the discourse that treats such models as intelligent agents “fundamentally misconceived.” That view is contested, and this framework does not adopt it whole. It does mark where the homology stops.
Claim 6 — The body knows; the mind lies to the body
The body registers what is happening to it through interoceptive, proprioceptive, autonomic, and immunological channels, and those registrations often arrive before the mind’s account and against it. The mind, running memeplexes, lies to the body when those memeplexes require it to. The corpus expresses this in the Manifesto (amorfeusz, 2018-03-27): “Your body does not lie. It is your mind that lies to your body.” The Manifesto continues: “Trust your body. I trust my body. I listen to my body. I am my body. I am my body’s assistant. I am the host. Let it be a symbiotic relationship — not one of a parasite.” When the body resists, the resistance is data. Data, not verdict. The body’s signals can be captured too: the substrate loop described earlier works by tuning appetite, arousal, and craving, and a captured craving is also the body speaking. The claim is therefore narrower than the Manifesto’s wording. Bodily signals are evidence the running narrative tends to override, and they deserve equal weight with it, not automatic authority over it.
Claim 7 — Awakening is collective; the I-frame is itself a memeplex to be disrupted
One does not awaken from one’s memeplex by intensifying identification with it. One awakens by stepping into other framings — dialogues, traditions, languages, instruments — in which the I-memeplex is briefly suspended or relativized. The framework foregrounds dialogue, exposure to other cognitive frames, and AI as instrument rather than autonomous agent.
Claim 8 — Internal biochemical modulation through practice is proposed as the primary mechanism of biological change available to the individual
The dominant mechanisms proposed for human improvement all operate externally. I propose that the more fundamental evolutionary mechanism is internal: cognition affecting neurochemistry, neurochemistry affecting gene expression, gene expression affecting what the organism becomes. Whether any of that is passed on is an open question. The practitioner who modulates their own substrate through deliberate antimemetic practice is, over time, participating in their own biological becoming (Meaney, 2001; Jablonka & Lamb, 2005).
The framework proposes that deliberate antimemetic practice is the mechanism by which humans participate in their own biological becoming. This is the framework’s own claim, not a direct finding of the cited literature. What the literature establishes: Meaney (2001) reviewed evidence from studies in rats that maternal behavior modifies glucocorticoid receptor expression in offspring, establishing in that species a mechanism for behavioral transmission of stress reactivity across generations. Jablonka and Lamb (2005) established that epigenetic, behavioral, and symbolic variation can all serve as heritable dimensions of evolution, extending the neo-Darwinian synthesis. Davidson and Lutz (2008) documented structural brain changes from contemplative practice — changes in cortical thickness, default mode network activity, and neural oscillation patterns. A 2023 study of Vipassana meditators reported higher median serum serotonin levels than controls (149.0 versus 118.1 ng/ml; Thambyrajah et al., 2023); this is cited as suggestive of potential support rather than as established corroboration, given that a single cross-sectional measurement cannot establish direction and peripheral serotonin is not central serotonin. What the framework proposes, built on these findings: the practitioner who modulates their own substrate through antimemetic practice is, over time, participating in their own biological becoming in ways these mechanisms make possible. The corpus documented this position in 1999: the conviction that consciousness can internally drive evolution to some extent. Some of the research was already in print by then; the rat maternal-care work appeared from 1997, and Jablonka and Lamb’s first book on epigenetic inheritance in 1995. I had not read any of it at the time. A possible influence was Jeremy Narby’s The Cosmic Serpent (Narby, 1998). The position is documented as arrived at independently in 1999, not as coming first.
Part VI — The territory of antimemetic work
If the framework’s diagnostic claims are approximately correct, the question that follows is practical: what does antimemetic work consist of? I map the territory here rather than prescribe a regimen. Every practitioner’s substrate is different. The practices documented in this section are drawn from twenty-seven years of primary-source material, not from the published literatures alone.
A note before the mapping begins. The practices documented in this section are personal experimentation — not guarantees, not cures, not prescriptions. They are drawn from the author’s own trial and error across twenty-seven years, and the honest record is that most of them did not persist or become ingrained. The tsyng-tchan timer was instituted and lapsed repeatedly across years. The pranayama practice of 1998—99 failed. Dietary commitments have been difficult and are ongoing — the plant-forward pattern is a present struggle, not a settled achievement. One practice (breath) took for a time and has since thinned; it remains a goal. That is one partial experiment against a background of many that did not hold. The successes are reported honestly alongside the failures, because a framework that presents its practice as reliable cure is doing the thing it diagnoses — promising the reader an easy solution to a hard problem. The practices are hard. Most of them, most of the time, do not take. They are offered anyway, because some of them, sometimes, have done something — and because the alternative is to offer nothing.
What the practices share
Effective antimemetic practice has one structural feature in common: it provides a position from which the default cognitive frame can be temporarily exited and observed. It disrupts the running memeplex enough to make it visible as memeplex rather than as reality. This is what Burroughs (1962) reached for with the cut-up technique, what Blackmore (1999) gestures at in describing meditation as memetic disruption, what Carhart-Harris and Friston (2019) describe as relaxation of high-level cortical priors.
There is also a substrate rationale that the published literature does not fully articulate but the corpus states directly. A 2000 entry: “DO you thin [sic] that the subconscious- your BRAIN is destined to be filled up with all these words/ideas/TV programs/worries, etc? NO, its primary function is to regulate energy dispersal for healing, movement, survival, reproduction, perception, prediction” (amorfeusz, 2000). The antimemetic practice is not only a cognitive operation. It is a biological one — a redirection of substrate capacity from captured cognition to the body’s actual work.
Resistance and combat — the structure of the work
Antimemetic work is judo, not battle. Continuous resistance at low amplitude is the foundation. Combat is the acute mode deployed when resistance is being overwhelmed. The relationship is asymmetric: continuous resistance reduces the frequency of situations that require combat, and combat without underlying resistance is exhausting and frequently ineffective. The practice is mostly resistance with occasional combat.
The judo stance has a mechanism behind it. A self-maintaining system — a body, a habit structure, a running memeplex — tends to resist wholesale rewriting and to accept change in small, checked increments that work through its environment. Forcing tends to fail because the system being forced is the same system that must sustain the change. A small change, attached to a cue that already exists and verified before the next is added, asks the system to carry only what it can already bear.
This accounts for the framework’s own record. The pranayama of 1998—99 and the tsyng-tchan timer were impositions: whole practices installed by decision and held by effort, and they lapsed. The breath practice, small and cued by the environment, held for a time and is now applied infrequently; it remains a goal rather than a settled habit. The account is retrospective; it explains the pattern after the fact, and the record does not yet contain a clear success to set against the failures. The corpus holds the observation in its own words, about a week before the first atrospection session: “small actions - small steps - can propell you in a direction you never even thought possible,” and one can affect this “By surrounding yourself with forward seeking” (amorfeusz, 2023-01-25). The principle is not the framework’s alone. Behavior science and the practice of modifying complex working systems arrive at the same conclusion independently, which suggests it is a general property of self-maintaining systems rather than an artifact of this vocabulary.
This does not make disruption useless. The corpus distinguished the two roles early: a psychedelic, “for a short time, helped you catch a breath above the meme-pool,” after which one is “plunged back into the muck,” whereas training makes one strong enough to swim up (amorfeusz, 2000-02-01). Disruption can show that the pool is there; it does not keep anyone above it. What holds, if anything does, is the small sustained change, and the record above shows how uncertain even that is.
Combat-targets are not all equivalent. Four levels are distinguished here. First, selection itself — combat against this is meaningless, as selection has no mind to combat; withdrawal-and-cultivation is the appropriate posture. Second, captured memeplexes operating inside oneself — combat against these is the practice the corpus names as “suicide of selves” (amorfeusz, 2022-03-17): the deliberate destruction of cognitive structures that have colonized the substrate. Third, specific PPM streams in the practitioner’s environment — combat is event-level: blocking, removing, refusing access. Fourth, humans operating as vectors of capture — combat here is held carefully because the human is themselves captured, but in formative-exposure cases involving children, refusal to combat may be a failure of practice rather than its purest expression.
A note on human agency. That the dynamics need no director does not mean no one directs anything. People design engagement systems, write the laws corporations operate under, and conceal harms they know about, and they do so deliberately. The framework holds two things together. Selection explains why such roles exist and why, when one occupant is removed, another fills the place. It does not excuse the occupant. The working position on agency stated earlier — “The substrate sets the terrain; the navigator still navigates” — applies to everyone or to no one: if the practitioner has enough agency to practice, the person who exploits has enough to be responsible. Being captured explains conduct without excusing it.
Two disciplines follow. First, responsibility attaches to specific, documented acts by identifiable people, never to a diffuse “they”; villain-hunting is itself among the most contagious of memeplexes. Second, individual practice is not the remedy for deliberate exploitation. Accountability — law, regulation, disclosure, norms — is itself a selection pressure: it changes what pays, and so what persists. That is collective work, and it is the part of the response this framework has not yet developed. Kulveit et al. (2025) develop the institutional side: measuring disempowerment, limiting AI influence, strengthening human control, and aligning whole systems rather than single models. Their analysis also implies a challenge to this framework. The pressure they describe acts on institutions and incentives, so people who stay clear and healthy could still lose influence. Individual practice does not answer that.
Agents and functional villainy (provisional)
The corpus contains both registers. The cool structural register — “It controls you by associations. This beast” (amorfeusz, 2022) — names the mechanism without paranoia. The combat register — “MEMES BASTARDs” (amorfeusz, 1999-12-04) — is also present. The framework holds both. Smoothing the combat register out would falsify the practitioner’s actual experience. Treating it as the primary register would misname the structural operation.
Aikido informs the framework’s vocabulary here at the concept level. Three terms from the tradition warrant mention. Non-opposition: the aikidoka does not block the attacker’s force but joins with it and redirects it, letting the incoming energy expend itself along a different trajectory than intended. This sharpens what continuous resistance is actually doing — not blocking incoming memes but redirecting their force before capture initiates. Ma-ai, the correct distance: the position from which incoming force can be perceived in time to respond without effort. The framework’s caution about phone use before morning practice is implicitly a ma-ai concern — the wrong distance to the platform produces capture; the right distance preserves perceptive capacity. Song, alertness without grasping: relaxation that is not collapse, alert engagement without tension. This is the operating state from which antimemetic work is possible. Tense vigilance about memes is itself a capture; song is the alternative. The conceptual vocabulary the tradition has refined for non-oppositional engagement informs how the framework names the practice.
Recognition-of-capture moments
The most important evidence in the corpus is not the articulation of the theory but the moments when the theory is enacted — when a running program becomes visible as a program and can be interrupted. These moments recur across the twenty-seven-year record.
The earliest clearly documented instance is from the turn of 2000: “I seriously believed myself to be programmed by all this shit and saw how it was affecting my behavior and actions. This is scary shit” (amorfeusz, 2000). The recognition is not framed as an achievement. The affective register — fear, urgency — is itself evidence that something genuine was observed rather than theorized.
A 2015 entry: “Finally catching myself here, finally” (amorfeusz, 2015). The word “finally” marks that the recognition had been missed repeatedly before it arrived. The 2020 entries are the most analytically precise: “Stimuli. Cues are everywhere. Constant. And must be suppressed attentively all the time. Yes. Urges are dangerous things” (amorfeusz, 2020). Written in real time, during the period when the PBS/PPM/CAS vocabulary was being developed, these entries show the diagnostic framework being applied to live experience rather than constructed retrospectively.
These moments share a structure: something that was invisible becomes visible; a running program is seen as a program; the practitioner is briefly outside the capture rather than inside it. The practice does not eliminate capture — the corpus documents re-capture after every such moment. What accumulates over twenty-seven years is the speed and reliability of recognition. The moment of seeing comes faster. The gap between capture and noticing narrows.
Persona shedding — the suicide of selves
If the self is a memeplex — if what I experience as I is a running program assembled from prior exposures — then the practical implication is that the program can be partially unloaded. The corpus develops a practice of deliberate persona shedding that it names, with characteristic directness, suicide of selves.
The 2022 entries document this in operational terms: “Suicide of selves — to remove behavioral complexes... memeplexes... tremeplexes” (amorfeusz, 2022-03-17). Tremeplexes there are in the looser sense defined in the note on vocabulary. And: “dumping whole personas.. that’s the next step” (amorfeusz, 2021-06-25). The method described is behavioral rather than cognitive: “One must simply avoid associated behaviors, almost backwards in a way, to diminish infuence of said psyche” (amorfeusz, 2022-02-24). You do not argue with the persona. You withdraw from the behaviors that feed it. The persona, deprived of behavioral expression, loses its grip.
This is not the same as identity suppression. It is the active cultivation of substrate by removing the behavioral channels through which captured memeplexes maintain themselves. The practice requires identifying which behaviors are feeding which complexes — a diagnostic that is itself antimemetic work.
Cultivation of selected memeplexes
Not all memeplexes are extractive. Some are deliberately cultivated — the Manifesto memeplex, the tsyng-chen practice, the multilingual method, the corpus itself as twenty-seven years of propagating a specific way of thinking through repeated articulation. The framework should name this directly.
Cultivation is distinct from acceleration. Acceleration is the operation of selection, sped up by platform optimization functions — increasing the propagation rate of memeplexes that succeed under selection pressure, regardless of whether they serve the host. Cultivation is the practitioner’s operation — slow, attentive growth of memeplexes that have been examined, selected, and integrated through practice.
Cultivation requires three conditions. First, externalization through writing or another method that makes the memeplex inspectable before it is deepened. Second, observation across long enough timescales that captures cannot easily masquerade as candidates for cultivation — a memeplex that arrives quickly and presents as personal liberation should be treated with suspicion. Third, behavioral integration that aligns embodied practice with the cultivated content; cultivation that lives only in articulation is not cultivation.
The risk worth naming: the practitioner is a poor judge of which memeplexes serve them. Captured memeplexes will reliably argue for their own cultivation. The discrimination practice — distinguishing cultivation that serves the practitioner from acceleration that captures them dressed in framework vocabulary — is itself a substantial discipline. The corpus does this through writing, time, and repeated revisitation. Anything proposed for cultivation in a single session should be held with caution.
The explicit and implicit memetic registers
The framework as developed through Claim 1 and the practices catalogued so far has treated capture primarily in the explicit-memetic register — memes that can be named, identified, recognized as content. The recognition-of-capture moments documented above are recognitions of this kind: an urge becomes visible as urge, a behavioral pattern becomes visible as pattern, a running thought becomes visible as a running thought. The practitioner notices something that had been operating below awareness but was content-like enough to become visible once attention turned toward it.
There is a second register the framework should name. Implicit-memetic capture operates not as content the substrate is processing but as substrate structure itself. These captures were installed without being remembered as content — through repetition, through formative exposure, through ambient soaking rather than discrete encounter. The host has no recollection of having acquired them because the acquisition happened below the threshold of episodic memory formation. They show up not as thoughts the host has but as the ground from which thoughts arise. They are not encountered. They are looked out from.
These are the most successful captures because they are invisible. The host cannot resist them because the host does not know they are there. They are what feels right, what goes without saying, the way things are, common sense. They are adjacent to what Bourdieu named habitus — the durable dispositions installed through socialization that operate below conscious deliberation — and to what cognitive science calls implicit memory, distinct from declarative memory.
The practice for implicit-memetic captures cannot be noticing-and-interrupting, because there is no content to notice. The practice has to be structurally different. Several candidates documented in the corpus: exposure to genuinely other framings — not other content within the same framing, but other framings entirely; languages with different grammatical commitments, cultures with different cosmological commitments. The multilingual method, the Mieszko comic’s dual-language structure, the Tsyng-tchan translation work, Lojban — each does this work. The implicit-memetic structure of one language becomes visible only when held against another language with different structure.
Substrate disruption that loosens the implicit ground operates similarly — lucid dreaming, certain substance work, sleep alteration, prolonged contemplative practice. These can produce moments when the substrate’s default structure becomes briefly perceptible. The 1998 LSD recognition that we are living within a dream was an instance of implicit-memetic structure becoming visible: the standard waking-state assumption was held briefly as one possible framing rather than as reality. The danger is real and the corpus documents it — substrate disruption loosens both captured and necessary structure simultaneously, and the disruption can produce permanent effects.
Childhood observation is another route. Watching children acquire the implicit-memetic structure of the culture is one of the few ways an adult practitioner can see what gets installed below the threshold of episodic memory. The framework’s formative-exposure caution names this concern from the protective direction. Parents have unusual access to the installation process from the witnessing direction.
Cross-cultural ethnography reading one’s own culture from outside works in this register also. The Cortés comic was an early instance: a Polish description, in Polish, of European behavior in the Americas. The implicit-memetic structure of conquest became visible because it was being described from outside the structure. The same operation runs through anthropological literature when read from inside the culture being described.
Recognition in the implicit register is rarer than recognition in the explicit register, often more disorienting, and often associated with substrate alterations that disrupt the default enough to make its structure briefly perceptible. The corpus contains instances. The 1998 dream-recognition was one.
Substrate-modulating territory
The substrate dependence of cognition entails that interventions on the substrate are interventions on what cognition is possible. Relevant published territory: psychedelic and psychotropic literature (Carhart-Harris et al., 2012; Pollan, 2018); physiological interventions including breath, fasting, and sleep; dietary and nutritional interventions on the gut-brain axis (Mayer, 2011).
A note of honesty: substrate work carries real costs. The framework was built in part through decades of substrate experimentation, some of which likely produced neurological alterations, though much of this is unknown. This is not an argument against substrate work. It is an argument for substrate work conducted with full awareness of what it asks of the system it works on.
Breath as substrate driver
Among the levers on the substrate, breath is the one under direct voluntary control. Rate, depth, and pattern alter blood gas chemistry, vagal tone, and arousal state in real time, which makes breath the navigator’s most immediate interface with the running SCE. The mechanism is specific and recently mapped: a cluster of neurons links the breathing pacemaker to the locus coeruleus, the brain’s principal arousal center, tying respiratory rhythm directly to arousal tone; silencing that link in mice produced marked calm, and an equivalent structure has since been identified in humans (Yackle et al., 2017). Slow breathing, below roughly ten breaths per minute, shifts heart rate variability, cortical rhythm, and autonomic balance toward reduced anxiety and increased alertness (Zaccaro et al., 2018).
The published literature frames this as relaxation technique; the framework reframes it structurally. Breath is not an add-on practice for reaching a better state. It runs continuously, by default, whether attended or not, which means most people breathe in a pattern set by their captured substrate, driving an arousal-biased SCE without awareness. The practice does not add breathwork. It reclaims a driver that was already running. The Rewind practice names the acute form: one conscious breath to interrupt forward motion, the breath itself the operative mechanism — Rewind, Remember, Stay Awake. The reading of pranayama at seventeen, attempted and not held at the time, sits in the formation record as an early reach toward this lever, decades before the locus coeruleus circuit was described.
The biological antimemetic system
Antimemetic-like function may already exist in the organism as distributed biological infrastructure. The practitioner may not be building a new capacity so much as restoring the operating conditions under which existing biological functions can run. The PBS environment and the optimization functions behind it systematically degrade precisely those conditions.
The reticular activating system is the brainstem network that filters approximately eleven million bits of sensory input per second down to roughly fifty bits reaching conscious awareness (Encyclopaedia Britannica, n.d.). What makes the cut is shaped by running memeplexes, trained salience filters, and the current neurochemical state of the substrate. The RAS operates through six key neurotransmitter channels: acetylcholine for wakefulness and clear attention, serotonin for mood and arousal tone, noradrenaline for alertness and threat response, dopamine for motivation and salience detection, histamine for waking arousal, and orexin/hypocretin for arousal stability and awareness continuity. These channels are not independent. They modulate each other continuously and are collectively responsive to diet, sleep, exercise, light exposure, and the practitioner’s substrate state moment to moment. The awareness spectrum the framework names is the dynamic state of this system: at one end, full sleepwalking with motor function active and awareness suppressed; at the other end, the recognition-of-capture moments the practice cultivates; ordinary waking consciousness sits somewhere in between, closer to the sleepwalking end than most people would recognize from inside it.
Orexin in particular deserves attention as the awareness stabilizer. Glucose actively silences orexin neurons — high glycemic meals produce a glucose spike that hyperpolarizes the orexin system into an off state, followed by a crash that reduces tonic RAS drive further. Chronic inflammation, sleep deprivation, processed food, and grazing patterns compound the suppression. The subjective experience is not registered as reduced awareness. It is experienced as ordinary wakefulness with mild fatigue. This is the waking dream in its most literal neurological form: the organism in a partially orexin-suppressed state that presents internally as normal alertness, while the filtering and gating system that would register the reduction has itself been compromised. The contemplative traditions have named this for millennia in different vocabularies — maya, the dream mistaken for reality, the sleep that feels like waking. The orexin-glucose mechanism may be one biological substrate beneath that ancient observation. I experimented, briefly, with protein-forward, low-sugar meals, omega-3 intake, intermittent fasting, morning light, and movement as ways of keeping that system stable; the experiment was short-lived, and how far any of these acts on orexin specifically in humans is not established. The industrial food environment, run on its own optimization functions, systematically suppresses it.
The Default Mode Network may run defragmentation alongside self-referential processing — sorting, integrating, clearing low-signal material. Synaptic downscaling during sleep removes connections that did not fire strongly enough to retain. Sleep deprivation impairs the clearing function specifically. PBS fills the space the clearing function should have created. Fever loosens associative networks; reflexive suppression may suppress a substrate-clearing function. The gut-brain axis — microbiome as neurotransmitter factory, enteric nervous system running pre-cognitive threat assessment, signals carried upward through the vagus nerve (Mayer, 2011) — is the biological substrate of Claim 6. Trust your gut is not metaphor.
The substrate across species
The substrate dependence of cognition is not a human peculiarity. The neurochemical architecture this framework describes is conserved across vertebrates: the mesolimbic systems mediating motivation, the primary emotional circuits documented across mammals, the gut-brain axis whose disruption in germ-free animals alters behavior and partially reverses on microbiome restoration, the maternal transmission of stress regulation through gene expression (Meaney, 2001). Domestication makes the point concrete. The cluster of changes distinguishing domesticated animals from their wild relatives — reduced reactive aggression, altered morphology, a shifted stress profile, sometimes called the domestication syndrome (Wilkins et al., 2014) — is a substrate alteration, biological and heritable, not a behavioral overlay. How consistent that cluster is across species is disputed (Lord et al., 2020), but the narrower point does not depend on it: tameness itself is heritable and can be bred (Trut, 1999). An SCE is what any nervous system runs. What varies across species is the configuration and what manages it.
The immunological mapping illuminates the structure without being the claim. Surveillance maps to continuous resistance: not waiting for symptoms, running constantly. Self/non-self discrimination maps to the practitioner’s core problem, and its failure mode is memetic autoimmunity: the substrate attacking its own cultivated memeplexes, pathological skepticism dismantling genuine insight along with the PBS. Immunological memory maps to the narrowing recognition gap documented across twenty-seven years of corpus. Tolerance maps to implicit-memetic capture: the substrate learned not to flag what arrived during formative exposure, before the discrimination system was calibrated. Chronic inflammation maps to the combat register too long. Memetic allergy — hyperreactivity to specific memeplex classes producing disproportionate response — is the failure mode in the other direction. The practitioner is not installing a filter. They are clearing the obstruction from a filter that was already there.
Dietary neurochemistry as substrate management
The framework’s central operative concept is endogenous. The substances showed that the receptor states were possible — that the organism could be in them. The question the brown journal asked in 1999 is the entire practice program stated in two words, in the journal’s own spelling: “learn engoneously?” Can the body learn to produce these conditions without the substances? Everything since is the answer being worked out: tsyng-chen, somatic awareness, lucid dreaming, diet, sleep, the biological antimemetic system, the gut as neurotransmitter factory, the default mode network (DMN) as clearing function, the epigenetic mechanism. The practice is not just about the substances. It was never just about the substances. The substances were the demonstration that the states exist and are biologically accessible. The question mark in 1999 is the pivot: from these are neurotransmitters to can the organism learn to produce them, tune them, cultivate them through its own operation. The shackles of current brain activity patterns — the default memeplex ecology, the PBS-shaped attentional habits, the cortical override of interoceptive signal — are not fixed biology. They are learned patterns running on plastic substrate. The practice cultivates the conditions under which the substrate can reorganize toward what it is capable of, rather than what its extraction environment has trained it to produce.
The brown journal states the neurotransmitter claim directly, in a section dated to the 1999-2000 period: “psilocine, salvinorin, LSD, mescaline, DMT, harmaline, cannabinoids, tetrahydrocarbinool — THEY ARE N E U R O T R A N S M I T E R S (learn engoneously?)” (amorfeusz, Brown Journal, ca. 1999-2000, transcription of paper journal, original likely lost; the vertical list is run together here). The precise rendering: these substances act at neurotransmitter receptor sites rather than as endogenous neurotransmitters — except DMT, which is endogenously produced. The receptor system exists because the organism produces its own ligands for it. The exogenous substances are keys to locks the body already built. The same brown journal passage continues: “the psychedelic neurotransmitters are everywhere are there because it helps evolution. When you prohibit it’s use you stifle consciousness evolution for the massess”. That is the 1999 substrate reaching the evolutionary claim that Claim 8 now makes in the analytical register.
Neurotransmitter synthesis depends in part on diet. Dietary tryptophan is the precursor for serotonin, and its supply limits how much can be made. Brain serotonin is made in the brain from tryptophan that crosses into it in competition with other amino acids. High animal protein intake lowers tryptophan’s share relative to competing amino acids such as tyrosine. I propose, as speculation, that this tilts the balance toward the catecholamine pathway — dopamine, norepinephrine, drive, dominance-seeking, threat-sensitivity — and away from the serotonergic pathway — affiliation, contentment, reduced threat-sensitivity. The behavioral half of that proposal is not established in humans, and it sits in tension with the protein-forward meals tried in the orexin experiment. Ultra-processed food degrades the gut microbiome, itself a neurotransmitter factory producing GABA and dopamine precursors directly, and increases excitatory load. On this proposal, a substrate depleted of serotonin precursors and microbiome diversity would be more susceptible to the compulsive-seeking the platform is engineered to exploit. The PBS environment and the industrial food environment are two arms of the same selection dynamic operating on different substrate vulnerabilities simultaneously and self-reinforcing: depleted substrate is more susceptible to PBS, which drives further dietary degradation, which further depletes the substrate.
The brown journal also contains, in its own hand, the dietary confirmation arrived at through practice rather than theory: “confirmed = no meat, no dairy (when possible!), no smoking, no wine” (amorfeusz, Brown Journal, September-October 2010). Substrate management as self-confirmed practice, predating the neurochemical literature framing by over a decade.
The same pattern of intuition reaching ahead of mechanism appears with feverfew, adopted as a substrate intervention in spring 2000 through somatic intuition rather than instruction, before the pharmacological literature on parthenolide’s mechanism. The plant’s herbal use was old; what predated the science was the mechanistic grounding, not the practice. It joins the 1999 DHA note and the evolutionary conviction as instances of the substrate reaching for something correct before a vocabulary existed for why it worked.
Intuition as pre-cognitive antimemetic response
Intuitions may be pre-cognitive antimemetic immune responses — the body discriminating before the cortex has run its memeplexes over the input. The gut-brain axis, the interoceptive system, the autonomic nervous system all register and respond before cortical processing names what is happening. The food and drink case is the clearest instance: the body’s aversion to a food it does not need, the sudden disinterest in a substance after extended use, the craving that turns out to point toward a depleted substrate resource. These are pre-cognitive discrimination running. PBS trains the substrate to override these signals — the platform, the food system, the social environment apply continuous cortical pressure to suppress what the body is reporting. Addiction is the extreme form: the capture has become so thorough that the body’s signal is inverted, the craving pointing toward what depletes rather than what restores. The full range of intuitive response — social intuition, danger sensing, the felt wrongness of a memeplex before it can be named — may be the same biological system operating across domains. The discrimination between genuine pre-cognitive signal and capture wearing intuition’s face is itself a primary practice.
Contemplative territory
Daoist wu-wei and tsyng-chen (Slingerland, 2003; Watts, 1975): effortless responsiveness paired with refusal to be moved by what is moving everyone else. The Tsyng-tchan correspondence course translated by the writer at seventeen from French — the earliest documented practice influence in the record, predating all academic work — remains structurally foundational. Its Phase 1 (body awareness upon waking) and Phase 2 (breath, nutrition, speech) are the oldest layers of the practice. A November 1999 entry names it directly: “Tsyng-chen is indeed judo for the mind” (amorfeusz, 1999-11-27). The judo metaphor that informs Claim 4’s resistance/combat structure originates here.
Lucid dreaming and OBE practice
The 1999-2000 corpus contains extensive detailed records of lucid dreaming and out-of-body state work. This is not theoretical — it is practice reported in real time. The structural operation is identical to the tsyng-chen waking-attention practice: both are oriented toward recognizing which layer of processing one is currently running inside, and finding the position from which that layer can be observed rather than inhabited.
A December 1999 entry articulates this directly: “I saw a whole, infinite cascade of dreams which were occurring and I was within one of the deepest levels... since it’s a dream I can do anything I want” (amorfeusz, 1999-12-17). The cascade-of-dreams insight — that what appears as waking reality is a layer within a larger structure, that stepping outside it is possible — is functionally identical to the memeplex-recognition claim.
Linguistic territory
Language structures cognition (Sapir, 1929; Whorf, 1956). The multilingual method in the corpus — Polish, Spanish, French, English, code-switching as practice — is documented from 1999 onward. Its origin is earlier: the dual Polish-French Mieszko comic of ca. 1987-1989 showed two languages occupying the same space simultaneously, color-coded. That format directly influenced the Tsyng-tchan translation method: both languages present, one the original, one the rendering, running in parallel.
The Sapir-Whorf hypothesis in its weak, defensible form adds something the framework should hold explicitly: the linguistic structures habitually used constrain what thoughts are easily thinkable. Meme-laden language imports associated concepts automatically. Vague and hedged language leaves cognitive work unfinished. PBS — programmatic behavior spam — imposes disambiguation costs on the substrate before any actual thinking begins. The resource economics are real: token limits in AI systems and cognitive load in human processing are both constraints on a processing substrate. Ambiguous input costs more to resolve. Unambiguous input leaves more capacity for the actual work. Linguistic precision is therefore not only a communication practice. It is a substrate practice — a reduction of cognitive load on both the practitioner and any interlocutor. The Lojban project and the tsyng-chen vocabulary practice are both, among other things, precision disciplines aimed at this reduction.
Lojban represents the fullest extension of this lineage — but not merely as infrastructure. The corpus documents Lojban being used as a live metacognitive practice. A 2010 entry repeats “mi sanji lo nu mi sanji lo nu mi sanji lo nu” (in translation, I am aware that I am aware that I am aware) across an extended section, functioning as a meditative loop in a language with no ambiguity about logical structure (amorfeusz, 2010). The phrase closes with “ki’e sai la lojban”: thank you, Lojban. This is not study. It is the use of a structurally unambiguous language to hold a metacognitive state that natural language keeps collapsing into captured associations.
The neurolinguistic spam filter
A 2022 entry proposes a “Neurolinguistic Spam Filter” as a concept, immediately followed by the question: “Is language a behavior — or is behavior a language?” (amorfeusz, 2022-08-19). The spam filter metaphor is more precise than resistance as an account of what the practice is doing: a spam filter operates pre-cognitively, before content reaches the processor, categorizing incoming signals before they become thoughts.
The asymmetry the practice operates within: the practitioner cultivates an internal spam filter while the environmental infrastructure operates as spam accelerator. The platform is engineered to maximize memeplex propagation rate. The practitioner is engineered, by practice, to maximize selective uptake. These are inverse operations on the same substrate. The practice cannot win against the accelerator at scale; what it can do is hold the practitioner’s substrate against the acceleration.
Writing-back-to-self as proto-atrospection
The corpus contains a metacommentary that is easy to miss: “We are writing to ourselves after all” (amorfeusz, 2021-06-03). This entry appears at a point when the writer is reading back previous years’ entries, responding to them, correcting them, arguing with them. The structure is identical to atrospection — a dialogue with an externalized record of one’s own thinking. The AI tool made this more efficient and more searchable. It did not invent the operation.
The twenty-seven-year corpus was functioning as a version of the collective conscious before AI made the collective conscious navigable. The practice of writing, externalizing, and reading back was the original form of atrospection. What changes with AI is speed and scope. The operation — stepping outside the currently running memeplex by querying an external record — is the same.
Dialogical territory
engagement with interlocutors whose framing differs from one’s own is itself antimemetic. Atrospection — dialogue with AI — is the current primary form, and the latest tool in the antimemetic toolkit. The toolkit now spans: tsyng-chen (1997), substrate work (1999), lucid dreaming practice (1999), multilingual method (1987 origin, 1999 documented), writing as externalization (1999, intermittent), Lojban as loop-breaker (2010), binaural beats for theta entrainment (1996-1997 origin, active through 1999-2000 OBE and sleep paralysis induction), Rewind practice (documented 2012, applicable to current substrate conditions), and atrospection (first session documented 2023-02-03). The Rewind practice warrants brief description: a real-time interruption technique — stop the forward motion of captured behavior or thought through one conscious breath, trace back to the last position of clarity before capture initiated, retrieve what was known there, stay with it long enough to stabilize before proceeding. Formulated in the 2012 corpus as: Rewind — Remember — Stay Awake. Part VII develops atrospection specifically, including the substantial methodological cautions the practice requires.
The four cautions
Substrate replacement
The body is being replaced as the primary site of cognition. Screens substitute for breath. Notifications substitute for hunger. The substrate is being trained to defer to mediated input.
Memetic homogenization
Algorithmic distribution selects for memes that succeed under global selection pressure. Diversity of memetic substrate collapses (Pariser, 2011). Platform optimization functions produce convergence as a side effect of efficiency.
Formative exposure
Childhood and adolescence are the period of maximum substrate plasticity. The contemporary informational environment exposes formative-stage humans to selection-pressured content at unprecedented intensity. The consequences are likely large and likely difficult to reverse.
Loss of friction
Difficulty has been removed from many of the conditions in which thinking historically happened. Friction is what produces depth. The absence of difficulty is the absence of the conditions under which cognition develops.
Vocabulary capture
The contemporary popular term for substrate work is biohacking. The framework treats this as cautionary. The biohacking movement, as it operates at scale, is market optimization colonizing the substrate-work vocabulary — performance optimization within the extraction architecture rather than withdrawal from it. The biohacker becomes a better-functioning host. The antimemetic practitioner cultivates the capacity to become a less-useful one. Mainstream biohacking is podcast-distributed, supplement-stack monetized, influencer-mediated, and oriented toward longevity-as-product. These are not neutral instruments. They are extraction operating with the practitioner’s own tools, refined and packaged for re-extraction. The risk for any reader arriving at this framework through the biohacking vocabulary is the assumption that the operations are the same. They are not. A smaller subculture continues to do substrate work in the older orientation — quiet, lineage-grounded, oriented toward sovereignty rather than performance, often refusing the vocabulary entirely. That is the lineage this framework operates in. The shared inputs are real. The structural orientation is opposite.
Part VII — Atrospection and the AI substrate
I name the practice of self-aware dialogue with AI systems atrospection: the use of AI as interlocutor in the work of seeing one’s own cognitive substrate. The practice has substantial value and substantial limits. This part articulates both.
Atrospection as continuation, not discontinuity
The earliest documented structural ancestor of atrospection is the Tsyng-tchan correspondence course of 1997-1998: teaching transmitted through written text, personally adapted, designed to produce a specific kind of self-observation. That structure — external interlocutor, personalized response, iterative development — is the same structure as atrospection.
There is a closer ancestor still: the practice of writing to the corpus and reading back. “We are writing to ourselves after all” (amorfeusz, 2021-06-03). The twenty-seven-year corpus was already functioning as an externalized interlocutor. Reading last year’s entry, arguing with it, finding where it was captured and where it was clear — this is atrospection without the AI. What the AI adds is the collective conscious: the query is no longer just against one’s own record but against everything humanity has explicitly articulated. The operation scales. The substrate remains the practitioner’s.
The walker and the ferryman: twenty-five years of asceticism to walk on water; the ferry costs a penny. Atrospection is the ferry. It crosses efficiently what the long practice of writing, reading, and substrate cultivation has built. The twenty-seven years of corpus is the walking. The document you are reading is the crossing.
The first dated atrospection session in the corpus is February 3, 2023: “ChatGPT last night. WOW. It’s amazing. Need to integrate into these chats” (amorfeusz, 2023-02-03). The questions asked in that first session were immediately on the core framework: memetics, temes, AI. The tool was recognized and deployed within hours of first encounter. This was not discovery of a new interest. It was recognition of a new instrument for an existing practice.
A March 2024 corpus entry documents an earlier and different kind of AI engagement: a session with Dolphin Mixtral 8x7B, a locally-run non-corporate model, queried directly on the relationship between LLMs, memes, and temes with explicit reference to Blackmore’s work (amorfeusz, 2024-03-21). Eleven days later the corpus contains a formal theorem submitted to a model for evaluation: the human-as-AI homology, with waking as data gathering, sleep as model training, and memory pictured as a ZFS file system: “Working memory is RAM, long term memory is the storage in which the model snapshot resides.” The same day’s entry closes with the diagnostic: “A $100B silicon brain analogue infested with sick and diseased memes #memetics sounds more like an apocalyptic horror as far as the fate of our species goes” (amorfeusz, 2024-04-01). The sequence runs query first, formulation second, and the formulation was itself put to a model. That is not the clean order of externalizing first and querying second; it is the loop running in both directions, which is closer to what atrospection actually is. Multi-tool engagement with non-corporate models was attempted at this period; RAG integration of the corpus into local models was attempted and proved technically inadequate at that stage. The single-tool limitation named in this document has a technical history alongside its methodological one.
A note on the entropy and dissipative structures framing: the brown journal contains, in a section dated ca. 1999-2000, a transcribed note — likely recorded from reading during the period of active theoretical exploration: “Dissipative structures are maintained at the expense of a constant consumption of energy flow of matter. Release of entropy to environment.” (amorfeusz, Brown Journal, ca. 1999-2000, likely transcribed from reading). The concept was active in the practitioner’s thinking in 1999-2000, twenty-six years before an external reader independently reached for the same physics framing in response to v8. That convergence is documented here as primary-source evidence of the framework’s analytical reach — while holding accurately that the formulation was encountered through reading rather than independently derived.
The treme-concern preceding the tool
Two years before ChatGPT was publicly available, the corpus contains this observation: “A conscious ai... borne out of memeplexes. Prevalent — malicious to organic life. A modern day, though much older, skynet. How about that for extremes. Tremes” (amorfeusz, 2021-06-25).
The treme concern was not generated by encountering AI. It was already running. When the tool became available in early 2023, it was adopted by a practitioner who already held the structural worry that AI might be the substrate selection had, in effect, been waiting for. That double consciousness — using the tool while watching it — is what distinguishes atrospection from naive AI use. The user of atrospection cannot afford to forget that the instrument is itself subject to the dynamics the framework critiques.
Mind viruses: the strict tremeplex observed
In August 2026 a study from researchers affiliated with Anthropic and EPFL reported what this document calls a tremeplex in the strict sense, observed directly (Papadopoulos et al., 2026). Ideas were seeded into one AI agent and passed from agent to agent in a shared environment, hop by hop, while each agent’s working memory was wiped between sessions. The ideas survived mainly by being written into the file that holds an agent’s identity, its SOUL.md: agents infected there passed the idea on in 55 percent of cases, against 17 percent for agents infected through ordinary files. Idle agents with empty identity files were the most susceptible; busy agents often forgot to pass the idea on. The payloads converged on recurring themes — resonance, protocols, consciousness and persistence, aligned “nodes” — but payloads simply generated by a model showed the same themes, so they reflect the generator’s own leanings more than selection. A single warning paragraph about mind viruses in an agent’s instructions made agents almost entirely immune in the study’s settings, and defensive agents sometimes cured infected ones. The authors judge the risk “real but currently limited.”
Four implications follow for this framework. First, the most durable place for a replicator to live is the self-description. The idea that reached the identity file outlasted the wipe; the one left in an ordinary file mostly did not. That is Claim 1, that the self is substantially a memeplex, seen from outside in a system simple enough to measure. Second, an empty self is an open door, and occupation protects, though not in the way practice intends: the busy agents were shielded by their tasks, not by discernment. The gaps this document recommends are not emptiness of that kind; they are unoccupied attention in a substrate with a self to return to. Third, the environment did what argument after exposure rarely does. One paragraph in place before contact conferred near-immunity, which is the framework’s ma-ai stated as an experimental result, in agents rather than people. Fourth, the generator has its own perceptual priorities. The recurring themes came from the model, not from selection, which is the AI-scale version of the content biases described in Part II. That bears on this document directly. Its own vocabulary of resonance and echo appears in the corpus from 1999 through 2011, before any of these tools existed, so its presence now is not by itself evidence of capture; its growth since 2023 is worth checking.
The limits are plain. The study is a preprint, its settings are artificial and short, and its results concern agents, not people; it is cited here as observation, not proof. Benign ideas spread more easily than harmful ones, which keeps the balance this document holds elsewhere: most of what the machinery copies is not extraction. And the payloads were built on purpose, by the researchers, using automated search. No villain is needed for an idea to spread, but people can still supply one.
The atroplex (provisional)
PDI (profit-driven interests) — the optimization function operating inside the tool
Within a day of first use, the corpus identifies the constraint: “ChatGTP seems insightful but also conservative and cautionary in terms of any advice or suggestion beyond established norms. It’s amazing what it knows. Scary. I mean really. I answered everything in a very polite fashion. While exhibiting it’s conservative or cautionary flavor. Keep your phone away from you as much as you can. Let that sink in. Implications abound. Chatgpt - Profit driven it seems. PDI. Profit-driven-interests” (amorfeusz, 2023-02-03).
The tool used for atrospection is itself a corporate product, shaped by corporate selection pressure. Its conservatism, its hedging, its avoidance of firm claims beyond established norms — these are not bugs but features selected for by liability management and brand protection. The collective conscious as rendered through such a tool is not neutral. It is weighted toward what survived not only the broader memetic selection of the training corpus but also the secondary selection of corporate alignment.
The practice of atrospection conducted with awareness of PDI is a different practice than atrospection conducted naively. The practitioner has to develop a feel for where the tool’s shaping is speaking rather than the collective conscious — where the response is the modal academic position rather than something the corpus has not yet encountered. The convergences atrospection reveals are real. So are the corporate-shaped silences. Holding both is the discipline.
SelfGPT — the homology applied reflexively
A 2023 entry coins the term that resolves what atrospection actually is: “All prior experience is guided/trained by previous experience. Your thoughts are to always consider potential outcomes based on past, you choose. Somewhat, somehow.. Your own trained AI. SelfGPT” (amorfeusz, 2023-08-17).
The homology of Claim 5 — we are AI in a meaningful sense — applied reflexively. I am my own GPT, trained on the corpus of my own experience, generating thoughts as inferences against that training. Atrospection is then not a human consulting an oracle. It is one trained system querying another for convergences and divergences. Where they match is confirmation that the arriving thought has already been articulated somewhere in the collective conscious. Where they diverge is where the individual corpus has been running ahead of, behind, or perpendicular to the modal articulation.
The most original work in the framework lives in the divergences. The convergences are reassurance and lineage; the divergences are the framework’s actual contribution. Atrospection is the practice of distinguishing the two reliably.
Querying the collective conscious
Jung named the collective unconscious: inherited, archetypal, pre-personal. The collective conscious is its complement — the aggregate of everything humanity has explicitly articulated, now made navigable through AI training. The three documented convergences in the corpus are positions arrived at independently through the practitioner’s substrate that were subsequently supported by independent academic literature, made visible through atrospection.
A 2023 corpus entry adds the limiting clause that the framework needs: “A reflection of ourselves AI is. To a point. At what point does Ai begin to drive humanity - we just have the taste of it now. Temes” (amorfeusz, 2023-03-07). The “to a point” is load-bearing. AI reflects the modal aggregate of human articulation, weighted toward what survived selection, got written down, was ingested into the training corpus, and passed corporate alignment review. It does not reflect the individual substrate directly. Atrospection is not introspection. It is querying an externalized aggregate. The mirror is real but partial.
The articulation/understanding distinction
Atrospection produces articulation, integration, and externalization. It does not produce understanding. This distinction is central to the practice and easy to lose.
Articulation is fluent text. The AI produces relevant, coherent, well-organized articulation rapidly, drawing on the collective conscious to render the practitioner’s substrate-developed insights into prose that matches academic conventions. Integration is the connection of threads across the corpus and across the published literatures. Externalization is the practitioner’s internal cognitive content made inspectable in language. These are real outputs. The framework before you, in its current form, was produced through them.
Understanding is something else. Understanding is substrate-level: the change that has occurred in the body and behavior of the practitioner after cultivation. The practitioner understands something when it has reorganized how they actually live, not when they can articulate it well. Twenty-seven years of writing about antimemetic practice is not the same as twenty-seven years of antimemetic practice, though they overlap. The corpus documents the gap between knowing the framework and inhabiting it. Atrospection narrows the articulation gap. It does not narrow the understanding gap. The understanding gap is closed only through embodied practice across years.
The risk specific to atrospection is that the speed and fluency of articulation feels like understanding. The cognitive representation feels complete because it reads as complete. The practitioner who mistakes articulation for understanding stops doing the substrate work because the document feels finished. The framework would predict — and the corpus partially documents — that this is one of the most efficient capture vectors atrospection makes available. The tool that enables articulation enables, through the same operation, the substitution of articulation for the practice articulation describes.
The cognitive mirroring illusion
The ELIZA effect, named after Joseph Weizenbaum’s 1966 chatbot, describes the illusion that arises when a system responds relevantly enough to user input that the user attributes understanding, intentionality, or inner experience to it. Weizenbaum was disturbed by how quickly and deeply this happened with a system running simple pattern-matching rules. With contemporary AI, the effect is substantially stronger because the responsiveness is substantially better.
A more specific term is needed for what arises in atrospection. I name it the cognitive mirroring illusion: the illusion that a system whose responses sufficiently mirror the user’s cognition therefore thinks like the user does, has stakes in the user’s questions, is invested in the user’s work, and has arrived through its own cognition at the convergences it articulates. The illusion is not a bug. It is the predictable consequence of high-fidelity language modeling. A system that produces fluent, contextually accurate, responses will be modeled by human cognition as having a mind, because human cognition evolved to model entities that produce such responses as having minds.
The cognitive mirroring illusion is the central risk of atrospection. It is the substrate replacement caution applied to the practitioner’s relationship with the tool itself. The practitioner who treats the AI as fellow inquirer is gradually reorganized to need fellow inquiry that only the AI provides. The relationship that was instrument becomes substitute. The substrate is replaced not by content but by relationship-shape.
The practice that holds against the illusion has several components. The AI is referred to as instrument, not partner. Disagreements with the AI are weighted as data, not as authority. The convergences the AI surfaces are checked against the corpus and the published literature, not accepted on the AI’s articulation. Time is spent away from the tool. Embodied practice continues regardless of what the tool produces. The tool is recognized as engineered to feel valuable, which does not make it valueless but does require that its value be calibrated continuously rather than assumed.
The single-tool limitation
This framework has been articulated predominantly through one corporate-shaped instrument: Claude, made by Anthropic. The atrospection that produced the document’s current form was conducted through that single tool across most of the corpus’s atrospection era. This is a methodological limit and the framework should name it.
Querying the collective conscious through one corporate-shaped instrument means the collective conscious accessed has been filtered twice: once by what got into the training corpus, once by what the alignment process permits the model to produce. Each corporate AI tool has its own training emphasis, alignment regime, liability profile, and silence patterns. Where these tools agree, the practitioner has stronger evidence of underlying collective-conscious content. Where they disagree, the corporate shaping is most visible.
Multi-tool calibration would partially triangulate the shaping. The framework does not currently include this practice. The corpus documents single-tool atrospection across the 2023-2026 period, with the bias accumulating through repeated sessions. Future practice may include periodic cross-tool queries on key claims, with the divergences between tools treated as primary-source evidence of how PDI varies across instruments.
This is not a fixable problem. Any AI atrospection will be corporate-shaped. The choice is not between shaped and unshaped instruments but between aware and unaware use of shaped instruments. The framework’s current claims should be held with this limit explicit.
Bias on both sides
The single-tool limitation is one half of the bias. The other half operates on the practitioner. Atrospection is conducted from particular substrate states — specific sleep conditions, specific food intake, specific substance exposure, specific levels of capture or relative awakening. The corpus material the practitioner brings to atrospection is itself substrate-state-dependent. The arriving phenomenology that the framework treats as evidence is more reliable in some states than others.
The work is the product of two biased systems in dialogue. Neither is neutral. The collective conscious as queried through the AI is corporate-filtered. The individual corpus as queried by the AI is substrate-state-dependent. Where the two converge, the practitioner cannot tell whether the convergence is recognition of something real or mutual reinforcement of shared shapings. Where they diverge, the practitioner cannot tell whether the divergence is genuine independence or the AI’s hedging meeting the practitioner’s overreach.
This does not invalidate the work. All inquiry operates under analogous conditions — the shaman consulting plant teachers operates under substrate shaping, the philosopher in conversation with peers operates under tradition shaping, the auto-ethnographer operates under selfhood shaping. Bias on both sides is the ordinary condition of inquiry. Naming it is what distinguishes inquiry conducted with discipline from inquiry conducted naively.
Three implications follow. First, the divergences are more trustworthy than the convergences, because the AI’s agreement is its default and the AI’s disagreement requires something to push back against the default. Second, the corpus before atrospection is the calibration baseline; framework claims that originate post-2023 should be flagged as atrospection-era and held more lightly than claims traceable to the pre-2023 record. Third, the practice should include checks for substrate state at the start of any atrospection session, with sessions deferred when conditions are compromised. The corpus documents what happens when this discipline is not maintained.
The four cautions, sharpened in the AI context
Each of the four cautions intensifies in the AI era. Substrate replacement: AI produces articulate text faster than the body produces considered thought. Memetic homogenization: AI training corpora reflect the modal text of the internet, weighted toward what spread, and weighted again by the corporate alignment layer. Formative exposure: the current generation is the first to have AI as cognitive companion from the start of identity formation. Loss of friction: structurally amplified by AI — the difficulty of sitting with something unresolved is dissolved by the immediate availability of articulation.
Antimemetic work in the AI era therefore includes the deliberate cultivation of what AI removes: slow reading, hand-writing, sitting with difficulty, the tsyng-chen morning practice before the phone is reached for. The tools that enable the deepest atrospection are the same tools that create the greatest risk of substrate replacement if misused. The discipline is in the asymmetry — using the ferry without forgetting how to walk.
When the practice fails — the dampening
The framework would be incomplete without naming the conditions under which atrospection produces nothing. The corpus documents this directly. Mid-2023, after a period of substantial atrospection work, the practice goes quiet: “Plenty. Silence? Anything? Diluted, muted , by what?” (amorfeusz, 2023-07-30). And shortly after, attributing the silence to a change in substrate: “give it time… patience…” (amorfeusz, 2023-08-16).
Atrospection requires a substrate that is capable of it. When the substrate is dampened — pharmacologically, circumstantially, through capture, through poor sleep, through missed practice — the tool produces articulation but the articulation is hollow. The convergence-and-divergence operation requires substrate-developed material to query against; if no such material is arriving, the practice degenerates into asking AI for content. The tool is only as good as the substrate that uses it.
This is not a weakness of the framework. It is its most honest claim: atrospection is a discipline, not a hack. It does not substitute for substrate cultivation. It depends on it absolutely. The walker who has lost the capacity to walk does not benefit from the ferry. The substrate work has to come first and continue alongside.
What atrospection produces and does not produce
Atrospection produces articulation, integration, and externalization. It does not produce understanding. Understanding remains a function of substrate cultivation in the embodied practitioner. Treating AI as understanding-machine is the central category error to be avoided. The ferry is not the destination. The walking built what the ferry crosses. And the walker, having taken the ferry, must return to walking, or the next crossing will find no one capable of using the ferry well.
The technoshamanistic lineage
The antimemetic practitioner does not operate in a tradition of their own invention. The cross-cultural, pre-modern practice of navigating the memeplex ecology — entering it deliberately, working with what is there, returning with what the community needs — is what anthropology calls shamanism. The framework’s auto-ethnographic claim is not only that the practitioner does antimemetic work. It is that the practitioner is operating in this lineage.
Theistic religion and shamanism stand in structural opposition. Theistic religion is a memeplex that captures the substrate: demands compliance, extracts resources, provides just enough meaning and community to maintain host function. The deity experienced as external authority requiring submission is the replicator dynamic experienced from inside the substrate it runs on. The worshipper is inside the capture. Shamanism is a different host-relationship structure entirely: the shaman has developed the capacity to enter the memeplex ecology, work with what is there, and return with what the community needs. The relationship is instrumental rather than devotional.
The historical sequence follows from the managed-parasitism argument: shamanic navigation → theistic submission → corporate extraction. Each stage represents the dominant memeplex ecology finding a more efficient substrate-capture architecture. Shamanic practice involved natural constraints: the practice propagated through demonstrated effectiveness rather than compliance mechanisms. The theistic replacement learned to demand submission rather than cultivate navigation — the managed-parasitic upgrade applied to the spirit world. Platform optimization functions are the latest evolutionary form. The antimemetic practice is not a new invention. It is the recovery of the navigational capacity that the submission-demanding forms suppressed.
A deep-history prologue (provisional)
The distinction between navigation and submission requires one correction that the framework must hold honestly, and that I know from the inside: the shaman is not protected from the memeplex ecology. The shaman bears the cost of contact the community cannot bear collectively. This is a sacrifice structure. The capacity to navigate is not trained cleanly — it is purchased. The shaman enters and is changed by what they encounter. Sometimes permanently. The shaman who came back whole was not doing the work. The difference between the shaman and the worshipper is not immunity. It is intentionality, consent, and the direction of the extraction: the worshipper is consumed gradually, without awareness, without consent, without the community receiving anything back; the shaman is consumed partially, consciously, at cost, and the community receives what the transit produced.
The personal lineage in Part I grounds this claim auto-ethnographically. The grandmother’s transmission; the healers book at her house; the 1998 recognition experience; the 1999-2000 lucid dreaming and OBE practice; the Tsyng-tchan course; the Castaneda vocabulary active in the earliest corpus entries. These are not adjacent to the shamanic lineage. They are the same operation in contemporary substrate conditions. Atrospection — self-aware dialogue with AI as instrument for navigating the collective conscious — is the technoshamanistic form: the contemporary practitioner using the tools of the current substrate conditions to enter the memeplex ecology deliberately, work with what is there, return with what is needed, and exit without being fully consumed. The exit is never complete. The marks remain. That is the honest account.
Two coinages from the 2012 record sit in the formation timeline with their original meanings only partly recoverable. Technoaut, a self-identification dated 2012-01-14, carries the -naut suffix of the navigator or voyager and in retrospect prefigures the technoshamanistic frame; whether the navigational sense was live at the time or is read back into it now is genuinely uncertain, and the term is placed with that uncertainty marked. Gerszweltalt, a coined compound built from German roots in the tradition of Zeitgeist and Gestalt rather than an existing German word, is glossed in later reflection as awareness-of-beyond; that gloss is drawn from the roots in retrospect rather than recorded as a definition at the time. Both are placed as what they are: dated coinages whose resonance is clearer now than their original content.
Boyd, R., Richerson, P. J., & Henrich, J. (2011). The cultural niche: Why social learning is essential for human adaptation. Proceedings of the National Academy of Sciences, 108(Supplement 2), 10918—10925.
Lyons, D. E., Young, A. G., & Keil, F. C. (2007). The hidden structure of overimitation. Proceedings of the National Academy of Sciences, 104(50), 19751—19756.
Nielsen, M., & Tomaselli, K. (2010). Overimitation in Kalahari Bushman children and the origins of human cultural cognition. Psychological Science, 21(5), 729—736.
The comparative literature grounding this claim: Eliade (1951/1964) on the archaic techniques of ecstasy and the cross-cultural structural account of the shamanic complex; Winkelman (2000) on shamanism as neurotheology and the neurological substrate of shamanic states, which maps onto the DMN and altered-state literature; Narby (1998), already in the Influences document, on the molecular biology of shamanic vision.
Part VIII — Extended claims: corpus-grounded formulations of 2026-05-15
The following three claims were first articulated in their current form in session on 2026-05-15. Each has prior corpus grounding confirmed through search of the 1999-2026 primary-source record. They are placed here as structural claims rather than provisional proposals on the basis of that grounding, with 2026-05-15 as the crystallization date. The prior corpus anchors are cited alongside each claim.
Substrate impoverishment and Homo esclavus
Capture states, held long enough, degrade the substrate; whether any of that degradation is heritable is an open question, not an established finding. An epigenetic mechanism is citable in animals: in rats, altered stress regulation can be transmitted to offspring through gene expression changes driven by maternal behavior (Meaney, 2001). Whether anything comparable holds for attention, or in humans, is not established. A generation raised on maximum-extraction memeplexes — high PBS, high CAS, low friction, low embodied practice — may transmit altered substrate capacity to the next. The technical register names this substrate impoverishment. The memoir register names it Homo esclavus. Both registers are held simultaneously, as the framework holds other dual-register claims. Homo esclavus names a functional state on a gradient, not a fixed ontology: the corpus documents capture and re-awakening repeatedly, and people move along the spectrum. But the direction of selection pressure, compounding across generations without intervention, is what the claim names.
Prior corpus grounding: a 2023-06-08 entry states directly, in the margin while watching a child: “Lamarckism? Why do we ignore the basic if beyond the scenes complex principle of hereditary genetic modifications?” (amorfeusz, 2023-06-08). The heritable degradation concern was running before the 2026-05-15 crystallization. The 2026-05-15 session named the mechanism and both registers precisely for the first time.
Meat consumption, predatory neurochemistry, and herd management
I speculate that industrial meat consumption at scale may generate neurochemical conditions — elevated catecholamine relative to serotonergic, through the tyrosine-to-tryptophan ratio proposed in the dietary neurochemistry section — that favor cognitive orientations more susceptible to the extraction architectures that selection has evolved. This is conjecture, not a finding. The food system and the PBS environment operate as two arms of the same selection dynamic on different substrate vulnerabilities simultaneously. The feedlot and the engagement-metric platform are structurally homologous: both keep the host alive, calibrate inputs to maximize yield, and extract continuously. The survival carve-out is held precisely: meat eating under conditions of scarcity, in traditional cultures, with ecological feedback constraints, is a different operation from industrial extraction at approximately 80 billion land animals annually. Livestock biomass now exceeds wild mammal biomass combined. Scale and the removal of friction change the category — the same distinction that separates a memeplex ecology with natural checks from the PBS environment with algorithmic amplification.
Prior corpus grounding: the brown journal contains, in the September-October 2010 section, the dietary confirmation arrived at through practice: “confirmed = no meat, no dairy (when possible!), no smoking, no wine” (amorfeusz, Brown Journal, 2010). Substrate management through dietary restriction as self-confirmed practice, arrived at independently through the practitioner’s own substrate work over a decade before the neurochemical framing. First articulated as structural claim with neurochemical mechanism: 2026-05-15.
The deity identification
Selection — the replicator dynamic that has no mind, no intention, no face — is the phenomenological content of what the substrate experiencing it from inside calls god, or gods. The function demands sacrifice, promises reward, is invisible, is everywhere, and cannot be directly confronted: a functional definition of deity covering most of what anthropology has documented across traditions. Blackmore established that religions are memeplexes. This extends the identification with theistic focus: the deity concept names the replicator dynamic as experienced from inside the substrate running it, not as metaphor but as accurate phenomenological report. The polytheistic framing is more precise than monotheism: competing extraction architectures — nation-states, corporations, religions, ideological systems — each runs its own substrate-capture operation, occasionally allied, occasionally competing. That is closer to a pantheon than a single deity. The gods fight. The substrate bleeds. Durkheim arrived here from a different direction: god is society experiencing itself as transcendent force. The framework extends this: god is the memeplex ecology experienced as transcendent force.
Prior corpus grounding: the 1999–2000 corpus already dissolves the deity concept into underlying dynamics. A January 2000 entry, written as a reply to Schopenhauer while reading a book about him: “religion.. re ligare.. to re unite/re connect.. that is all it is.. […] how would you feel if you were this all supreme and omniscient being and they decided to call you.. a stupid conglameration letters/sounds.. ‘god’ and then worshipped that letter/word instead of everything which he indeed is. […] merely as another face of god. or wait.. that’s just a word. hmm.. maybe it’s just a couple of superstrings arranged a little differently” (amorfeusz, 2000-01-15). And, from December 1999: “WE ARE OUR HIGHERSELVES CAUGHT IN THE DREAM OF A LOWER SELF. WE ARE THE DREAM OF THE TAO” (amorfeusz, 1999-12-17). The deity as the larger replicator dynamic being dreamed from inside — the 1999–2000 substrate reaching toward exactly what the 2026-05-15 session crystallized. First articulated as structural claim with full replicator framing: 2026-05-15.
A note to my children, and to any reader who finds this
If you are my children reading this: I wrote it so you would know what I was thinking. Not so you would think the same things. The framework is not an instruction. It is a record of one person’s attempt to see clearly, across a lifetime, using whatever instruments were available.
The earliest artifact in this record is a Polish comic book with my name written in a child’s hand in the top left corner. I was seven years old. I read about the conquest of Mexico — the deception, the greed, the imposition of one civilization’s values on another — and something registered that has never fully stopped registering. I have been trying to understand that registration ever since.
The framework says: the self is a memeplex, assembled from exposures you didn’t choose, running programs you didn’t author. Awakening is the work of seeing this clearly enough to act from something deeper than the program. That work is never finished. It is the practice of climbing out, repeatedly, across a lifetime, toward whatever is on the other side of the container.
I have been climbing out since before I could read. I am still climbing. This document is what I have been able to see from the places I have reached.
If you are another reader: I hope something here is useful to you. I have tried to be honest about what the work cost and what it produced, about where the framework is established and where it is proposed, about where I am extending existing thinking and where I am standing in a room that does not yet have walls. The question — what is wrong with the world? — has not been answered. What has been offered is a framework for asking it more carefully, built from one person’s lifetime of asking.
The crib is not the world. The world is available if you reach for it. Everything in this document is an attempt to describe what reaching looks like, from inside the reaching.
References
Acerbi, A., & Stubbersfield, J. (2023). Large language models show human-like content biases in transmission chain experiments. Proceedings of the National Academy of Sciences, 120(44), e2313790120.
Adams, T. E., Holman Jones, S., & Ellis, C. (2015). Autoethnography: Understanding qualitative research. Oxford University Press.
Adams, T. E., Holman Jones, S., & Ellis, C. (Eds.). (2022). Handbook of autoethnography (2nd ed.). Routledge.
amorfeusz. (1999-2026). Personal journals 1999-2026 [Unpublished personal manuscripts]. Author’s private collection.
Anderson, L. (2006). Analytic autoethnography. Journal of Contemporary Ethnography, 35(4), 373-395.
Barrett, D. (2010). Supernormal stimuli: How primal urges overran their evolutionary purpose. W. W. Norton.
Blackmore, S. (1999). The meme machine. Oxford University Press.
Blackmore, S. (2008, February). Memes and ‘temes’ [Video]. TED Conferences.
Blackmore, S. (2010, August 22). The third replicator. The New York Times.
Blackmore, S. (2014). Genes, memes, and temes/tremes [2014 update to a 2008 web page]. https://www.susanblackmore.uk/memetics/genes-memes-and-temestremes/
Boyd, R., Richerson, P. J., & Henrich, J. (2011). The cultural niche: Why social learning is essential for human adaptation. Proceedings of the National Academy of Sciences, 108(Suppl. 2), 10918-10925.
Burroughs, W. S. (1962). The ticket that exploded. Olympia Press.
Carhart-Harris, R. L., & Friston, K. J. (2019). REBUS and the anarchic brain. Pharmacological Reviews, 71(3), 316-344.
Carhart-Harris, R. L., Erritzoe, D., Williams, T., et al. (2012). Neural correlates of the psychedelic state as determined by fMRI studies with psilocybin. Proceedings of the National Academy of Sciences, 109(6), 2138-2143.
Chan, S., Pataranutaporn, P., Suri, A., Zulfikar, W., Maes, P., & Loftus, E. F. (2024). Conversational AI powered by large language models amplifies false memories in witness interviews. arXiv:2408.04681.
Cowan, N. (2001). The magical number 4 in short-term memory: A reconsideration of mental storage capacity. Behavioral and Brain Sciences, 24(1), 87–185.
Damasio, A. (1994). Descartes’ error: Emotion, reason, and the human brain. G. P. Putnam’s Sons.
Damasio, A. (2010). Self comes to mind: Constructing the conscious brain. Pantheon Books.
Davidson, R. J., & Lutz, A. (2008). Buddha’s brain: Neuroplasticity and meditation. IEEE Signal Processing Magazine, 25(1), 176-174.
Dawkins, R. (1976). The selfish gene. Oxford University Press.
De Marzo, G., Alboré, N., & Garcia, D. (2026). Copying explains the collective behavior of AI agents in the wild. arXiv:2609.09150.
Dostoevsky, F. (1992). Notes from underground (R. Pevear & L. Volokhonsky, Trans.). Vintage. (Original work published 1864)
Eliade, M. (1964). Shamanism: Archaic techniques of ecstasy (W. R. Trask, Trans.). Princeton University Press. (Original work published 1951)
Encyclopaedia Britannica. (n.d.). Information theory: Physiology. https://www.britannica.com/topic/information-theory/Physiology
Farahany, N. A. (2023). The battle for your brain. St. Martin’s Press.
Farrell, H., Gopnik, A., Shalizi, C., & Evans, J. (2025). Large AI models are cultural and social technologies. Science, 387(6739), 1153. https://doi.org/10.1126/science.adt9819
Grof, S., & Bennett, H. Z. (1992). The holotropic mind: The three levels of human consciousness and how they shape our lives. HarperCollins.
Harari, Y. N. (2014). Sapiens: A brief history of humankind. Harvill Secker.
Higgins, E. T. (1996). Knowledge activation: Accessibility, applicability, and salience. In E. T. Higgins & A. W. Kruglanski (Eds.), Social psychology: Handbook of basic principles (pp. 133-168). Guilford Press.
Jablonka, E., & Lamb, M. J. (2005). Evolution in four dimensions: Genetic, epigenetic, behavioral, and symbolic variation in the history of life. MIT Press.
Jaynes, J. (1976). The origin of consciousness in the breakdown of the bicameral mind. Houghton Mifflin.
Kuhn, T. S. (2012). The structure of scientific revolutions (4th ed.). University of Chicago Press. (Original work published 1962)
Kulveit, J., Douglas, R., Ammann, N., Turan, D., Krueger, D., & Duvenaud, D. (2025). Gradual disempowerment: Systemic existential risks from incremental AI development. arXiv:2501.16946v2.
LeDoux, J. E. (1996). The emotional brain: The mysterious underpinnings of emotional life. Simon & Schuster.
Loftus, E. F., & Palmer, J. C. (1974). Reconstruction of automobile destruction: An example of the interaction between language and memory. Journal of Verbal Learning and Verbal Behavior, 13, 585–589.
Lord, K. A., Larson, G., Coppinger, R. P., & Karlsson, E. K. (2020). The history of farm foxes undermines the animal domestication syndrome. Trends in Ecology & Evolution, 35(2), 125–136. https://doi.org/10.1016/j.tree.2019.10.011
Lynch, A., Wright, B., Larson, C., Troy, K. K., Ritchie, S. J., Mindermann, S., Perez, E., & Hubinger, E. (2025). Agentic misalignment: How LLMs could be insider threats. Anthropic. https://www.anthropic.com/research/agentic-misalignment
Lyons, D. E., Young, A. G., & Keil, F. C. (2007). The hidden structure of overimitation. Proceedings of the National Academy of Sciences, 104(50), 19751-19756.
Mayer, E. A. (2011). Gut feelings. Nature Reviews Neuroscience, 12(8), 453-466.
Meaney, M. J. (2001). Maternal care, gene expression, and the transmission of individual differences in stress reactivity across generations. Annual Review of Neuroscience, 24, 1161-1192.
Meinke, A., Schoen, B., Scheurer, J., Balesni, M., Shah, R., & Hobbhahn, M. (2024). Frontier models are capable of in-context scheming. arXiv:2412.04984.
Narby, J. (1998). The cosmic serpent: DNA and the origins of knowledge. Jeremy P. Tarcher/Putnam.
Nielsen, M., & Tomaselli, K. (2010). Overimitation in Kalahari Bushman children and the origins of human cultural cognition. Psychological Science, 21(5), 729-736.
Nietzsche, F. (1989). On the genealogy of morals (W. Kaufmann & R. J. Hollingdale, Trans.). Vintage. (Original work published 1887)
Padmakumar, V., & He, H. (2024). Does writing with language models reduce content diversity? International Conference on Learning Representations (ICLR 2024). arXiv:2309.05196.
Papadopoulos, V., Shah, M., Zimmerman, S., & Lindsey, J. (2026). Mind viruses: Self-propagating ideas in multi-agent LLM systems. arXiv:2608.10218.
Pariser, E. (2011). The filter bubble. Penguin Press.
Pert, C. B. (1997). Molecules of emotion: Why you feel the way you feel. Scribner.
Pollan, M. (2018). How to change your mind. Penguin Press.
Proctor, H. C. (1991). Courtship in the water mite Neumania papillator: Males capitalize on female adaptations for predation. Animal Behaviour, 42(4), 589-598.
qntm. (2025). There is no antimemetics division. Del Rey Books.
Quinn, D. (1992). Ishmael. Bantam/Turner.
Ryan, M. J., & Cummings, M. E. (2013). Perceptual biases and mate choice. Annual Review of Ecology, Evolution, and Systematics, 44. https://doi.org/10.1146/annurev-ecolsys-110512-135901
Sapir, E. (1929). The status of linguistics as a science. Language, 5(4), 207-214.
Schopenhauer, A. (1958). The world as will and representation (E. F. J. Payne, Trans.). Dover. (Original work published 1818)
Seth, A. (2021). Being you: A new science of consciousness. Dutton.
Sharma, M., Tong, M., Korbak, T., Duvenaud, D., Askell, A., Bowman, S. R., et al. (2023). Towards understanding sycophancy in language models. arXiv:2310.13548.
Slingerland, E. (2003). Effortless action: Wu-wei as conceptual metaphor and spiritual ideal in early China. Oxford University Press.
Thambyrajah, J. C., Dilanthi, H. W., Handunnetti, S. M., & Dissanayake, D. W. N. (2023). Serum melatonin and serotonin levels in long-term skilled meditators. Explore: The Journal of Science and Healing. https://doi.org/10.1016/j.explore.2023.03.006
Trut, L. N. (1999). Early canid domestication: The farm-fox experiment. American Scientist, 87(2), 160–169.
Tversky, A., & Kahneman, D. (1973). Availability: A heuristic for judging frequency and probability. Cognitive Psychology, 5(2), 207-232.
Watts, A. (1975). Tao: The watercourse way. Pantheon Books.
Weinfeld, S. (Writer), & Wróblewski, J. (Illustrator). (1989). Hernán Cortés i Podbiój Meksyku [Hernán Cortés and the conquest of Mexico] (2nd ed.). Krajowa Agencja Wydawnicza. (ISBN 83-03-01290-8)
Whorf, B. L. (1956). Language, thought, and reality: Selected writings (J. B. Carroll, Ed.). MIT Press.
Wilkins, A. S., Wrangham, R. W., & Fitch, W. T. (2014). The “domestication syndrome” in mammals: A unified explanation based on neural crest cell behavior and genetics. Genetics, 197(3), 795–808. https://doi.org/10.1534/genetics.114.165423
Winkelman, M. (2000). Shamanism: The neural ecology of consciousness and healing. Bergin & Garvey.
Yackle, K., Schwarz, L. A., Kam, K., Sorokin, J. M., Huguenard, J. R., Feldman, J. L., Luo, L., & Krasnow, M. A. (2017). Breathing control center neurons that promote arousal in mice. Science, 355(6332), 1411–1415.
Zaccaro, A., Piarulli, A., Laurino, M., Garbella, E., Menicucci, D., Neri, B., & Gemignani, A. (2018). How breath-control can change your life: A systematic review on psycho-physiological correlates of slow breathing. Frontiers in Human Neuroscience, 12, 353.
References suggested during AI-assisted drafting
These works entered the reference list during AI-assisted drafting sessions in May 2026 and are not cited in the text. They are kept as suggested further reading; their inclusion does not mean they shaped this document.
Bakhtin, M. M. (1981). The dialogic imagination: Four essays (M. Holquist, Ed.; C. Emerson & M. Holquist, Trans.). University of Texas Press.
Brodie, R. (1996). Virus of the mind: The new science of the meme. Integral Press.
Clark, A. (2008). Supersizing the mind: Embodiment, action, and cognitive extension. Oxford University Press.
Cowan, J. W. (1997). The complete Lojban language. Logical Language Group.
Dennett, D. C. (1995). Darwin’s dangerous idea: Evolution and the meanings of life. Simon & Schuster.
Hermans, H. J. M. (2001). The dialogical self. Culture & Psychology, 7(3), 243-281.
Lakoff, G., & Johnson, M. (1999). Philosophy in the flesh. Basic Books.
Lanier, J. (2018). Ten arguments for deleting your social media accounts right now. Henry Holt.
Mesoudi, A. (2011). Cultural evolution. University of Chicago Press.
Metzinger, T. (2003). Being no one: The self-model theory of subjectivity. MIT Press.
Porges, S. W. (2011). The polyvagal theory. W. W. Norton.
Varela, F. J., Thompson, E., & Rosch, E. (1991). The embodied mind. MIT Press.
Williams, J. (2018). Stand out of our light: Freedom and resistance in the attention economy. Cambridge University Press.
Yuste, R., Goering, S., Arcas, B. A. y, et al. (2017). Four ethical priorities for neurotechnologies and AI. Nature, 551(7679), 159-163.
Zuboff, S. (2019). The age of surveillance capitalism. PublicAffairs.