Position Statement

This chapter establishes three interconnected phenomena that bridge individual motivational architecture to collective crisis: Confabulation, the output of I's Bayesian ground-resolution function when I's Hypothesis Space about M is impoverished, absent, or suppressed, taking the form of sincere but inaccurate causal narratives that misattribute Mimetically-driven desires to Appetitive or Intrinsic sources (Chapter 2, Confabulation; Chapter 4, Section 4.4.3); Projection, the outward extension of Confabulated I-resolution onto external Objects, Individuals, or agent-categories through Template availability; and the Scapegoat Mechanism, the collective-scale expression of the AIM Operational Loop (Chapter 2; Chapter 4, Section 4.4.9) under conditions of population-wide M-Hypothesis Space impoverishment, through which convergent Confabulated attributions propagate to produce coordinated targeting of Individuals or groups selected through prey-detection criteria while being framed through predator-attribution.

Dependency Statement

Chapter 11 assumes and builds upon:

From Chapter 2 (Definitions): The Three-Source Taxonomy (Appetites, Intrinsic Motivation, Mimetic Desire), Individual (natural person with A-I-M architecture), Object (target of motivated action), Confabulation (sincere but inaccurate causal narratives), Template (cognitively available category for social inference), Model (Individual from whom desires are acquired), Scapegoat (target of confabulated agent-based causation exhibiting Dual-Role Confabulation), Conscious Cognition, Imagination.

From Chapter 9 (Neural Architecture): Source Opacity (uniform across A, I, M: the destruction of source-tags during Common-Currency Integration), Decision Hub (vmPFC-ventral striatum valuation system), timing architecture (Preconscious Transmission within the Transmission Window, 60–340ms, precedes conscious awareness 300–500ms), I-Override (deliberate scrutiny of motivational signals before action), Conscious Cognition (the always-on Bayesian generative inference process operative during waking phenomenological experience, comprising a generative operation, Imagination, which constructs candidate Hypotheses within the representable Hypothesis Space $\mathcal{H}$ , and an evaluative operation, Ground-Resolution, which redistributes probability mass over $\mathcal{H}$ ), of which I-Override is Ground-Resolution run deliberately on the source-attribution special case and Confabulation is an output rather than an operation.

From Chapter 10 (The Evolutionary Mismatch Hypothesis) and Axiom 9: Agent-Based Causal Default (Axiom 9 (human causal cognition architecturally optimised for agent-based explanations), Four-Type Agent Taxonomy (predators, collaborators, rivals, prey), Evolutionary Mismatch Hypothesis (modern complexity exceeds ancestral cognitive capacity), Information Gap (derived from Axiom 9 and Axiom 4) true causes exceed computational tractability).

Derivation Preview

The chapter derives its claims from four established premises:

From Axioms 2+3 (Source Opacity): The destruction of source-tags during Common-Currency Integration makes accurate introspection of motivational source architecturally impossible.
From Chapter 9 (Timing Architecture): Preconscious Transmission (the Transmission Window, 60–340ms) precedes conscious awareness (300–500ms), ensuring that Mimetic acquisition cannot be witnessed by the acquiring Individual.
From Axiom 9 (Agent-Based Causal Default): Human causal-inference systems are optimised for identifying agents, not structural causes, producing systematic agent-attribution even when systemic factors are determinative.
From Axiom 9 and Axiom 4 (Information Gap): True causes in modern environments exceed tractability, creating a closure demand that Confabulation satisfies.
From Chapter 4, the two-operation structure of Conscious Cognition: Conscious Cognition is one always-on process comprising a generative operation, Imagination, and an evaluative operation, Ground-Resolution. I-Override is Ground-Resolution run deliberately on the source-attribution special case, with the interference term sustained and S expanded to admit the Mimetic Hypothesis. Confabulation is an output rather than an operation: a Hypothesis Ground-Resolution has returned that is sincere and wrong.
From the Confabulation Rationality Lemma iterated against an undischarged explanandum, the Confabulation Proliferation Corollary: where the M-Hypothesis is unavailable under Type I Exclusion and the wanting does not terminate on acquisition, ground-resolution generates an unbounded and non-convergent set of A-grounded and I-grounded attributions rather than a single stable misattribution, and the search terminates only on the agent-attribution Axiom 9 supplies.

Given these premises, Confabulation is entailed as the output Ground-Resolution returns on closing when an Individual attempts to explain M-driven desires under Source Opacity and Vocabulary Deficiency, Projection is entailed as the same operation closing on others through Template deployment, and the Scapegoat Mechanism is entailed as the collective completion when Information Gaps demand agent-based explanation.

Chapter Preview

The derivation follows four movements corresponding to scale transitions:

Movement A: Individual Confabulation (§11.1–11.6) The Information Gap → the Confabulation Rationality Lemma → the Confabulation Proliferation Corollary, its collective instance and the Exception-Handling Analogy → I-Override as interrogation site → Confabulation as closure mechanism → Epistemic Suspension as alternative → Confabulation in social interaction → Intractability Confabulation as agentless completion

Movement B: Projection (§11.7–11.9) Self-model Templates → Template inversion (contrast categories) → Guilt-priming (elevated Template activation) → Projection as Confabulation applied to others

Movement C: Collective Pathway (§11.10–11.18) Individual Confabulations become observable → Propagation through prestige-bias and social proof → Scapegoat Mechanism as highest-likelihood completion → Derivation of the Presumption Against Scapegoating

Movement D: Signatures and Implications (§11.19–11.23) Dual-Role Confabulation signature (diagnostic test) → Non-Malice Corollary (reframing conflict) → Moral Vocabulary Displacement Principle (normative implications) → Anticipating Objections

Empirical Foundations: Confabulation in Psychology and Neuroscience

The phenomena this chapter establishes (Source Opacity, Confabulation, and Projection) are not AIM innovations. They are extensions of well-documented findings in psychology and neuroscience. Humans routinely generate inaccurate explanations for their own behaviour, and this is a feature of normal cognition, not a pathology.

Nisbett & Wilson: Telling More Than We Can Know

In 1977, psychologists Richard Nisbett and Timothy Wilson published landmark research demonstrating that people have limited introspective access to the causes of their own behaviour. In a series of experiments, subjects were manipulated by factors they could not consciously perceive (word priming, positional effects (preferring items on the right side of a display), bystander presence) and then asked to explain their choices.

The consistent finding: subjects provided confident, detailed explanations for their behaviour that were demonstrably false. When told the actual cause (e.g., “You chose the rightmost item because of a position bias”), subjects denied the influence and maintained their original explanation. Introspection did not reveal the true cause; it generated a plausible story using culturally available explanations.

Nisbett and Wilson concluded that people “tell more than they can know”: they report explanations that feel like self-knowledge but are actually inferences constructed after the fact. The brain does not have direct access to the processes that generated the behaviour; it observes the behaviour and constructs a narrative.

This is the foundation for AIM’s Confabulation architecture: when asked “Why did I do that?” the brain generates an answer even when the actual cause is inaccessible. The answer is sincere (the Individual believes it), but it is constructed, not retrieved.

Gazzaniga: The Left-Brain Interpreter

Neuroscientist Michael Gazzaniga’s research with split-brain patients revealed the mechanism underlying Confabulation. In patients whose corpus callosum had been severed (preventing communication between brain hemispheres), the left hemisphere, which controls verbal reporting, would generate explanations for actions caused by the right hemisphere, even though it had no access to the right hemisphere’s inputs.

In one famous experiment, a patient’s right hemisphere (which processes visual information from the left visual field) was shown the word “walk.” The patient stood up and began walking. When asked why, the patient’s left hemisphere (which did not see the word) said “I’m going to get a Coke.” The explanation was immediate, confident, and entirely confabulated.

Gazzaniga called this the “left-brain interpreter”: a system that runs constantly, generating narratives to explain behaviour. The interpreter does not say “I don’t know”; it constructs a plausible story from available information. Critically, this is not pathology. The interpreter operates in healthy brains all the time, generating explanations for actions whose actual causes are outside conscious awareness.

The AIM Framework identifies this interpreter as the site where Confabulation occurs during I-Override. When the Individual asks “Why do I want this?” the interpreter responds, not with retrieved source-tags (which have been destroyed by Common-Currency Integration) but with constructed narrative using available Templates.

Confabulation as Routine Cognition

Early Confabulation research focused on clinical cases (brain damage, Korsakoff’s syndrome, dementia), where Confabulation is florid and obviously inaccurate. But more recent research establishes that Confabulation occurs in healthy populations as a routine feature of cognition.

Kopelman (1987) distinguishes “spontaneous confabulation” (unprompted false narratives, typically pathological) from “provoked confabulation” (generating explanations when asked, which occurs in normal individuals). Provoked Confabulation is “the routine generation of ill-grounded explanations” that serves to maintain a sense of competence and agency (Bortolotti, 2018). Healthy brains “fill in gaps, alter perceptions, and generate false Beliefs for creating meaning from confusing inputs.”

Confabulation is not a sign of brain damage or dishonesty. It is what brains do when asked to explain processes they cannot observe. The Individual is not lying; they are completing an Information Gap with the best-available narrative.

Self-Perception Theory: Inferring Internal States from Behaviour

In 1967, psychologist Daryl Bem proposed self-perception theory: individuals infer their own attitudes and preferences by observing their behaviour, not through direct introspective access. When internal states are weak or ambiguous, people answer “What do I feel?” by asking “What did I do?”

Example: A person who eats brown bread regularly might infer “I must like brown bread” by observing their behaviour, rather than accessing a pre-existing preference. The preference is constructed from behavioural observation, then treated as self-knowledge.

Self-perception theory explains why behaviour shapes subsequent attitudes: the Individual observes themselves acting, infers a motivation consistent with that action, and incorporates that motivation into their Self-Model. The process is identical to AIM’s Confabulation architecture, lacking direct access to motivational source, the Individual constructs an explanation from observable evidence.

The Introspection Illusion

Research by Emily Pronin and colleagues demonstrates the introspection illusion: people believe they have privileged access to their own mental states while treating others’ self-reports as unreliable. This creates systematic bias:

“I know why I did that” (confidence in own introspection)
“They don’t really know why they did that” (scepticism about others’ introspection)

The illusion persists even when people are informed about introspective limits. Individuals continue to feel that their own explanations are accurate self-knowledge, not constructions, even after learning that introspection is generally unreliable.

This explains why Confabulation is invisible to the confabulator. The narrative feels like discovery, not invention. The Individual cannot distinguish “I retrieved this information from memory” from “I generated this narrative just now,” because both processes produce the same phenomenology: a plausible explanation that achieves cognitive closure.

Dual-Process Theory: System 2 Explains System 1

Daniel Kahneman’s dual-process theory distinguishes System 1 (fast, automatic, unconscious) from System 2 (slow, deliberate, conscious). Most behaviour is driven by System 1 (pattern-matching, heuristics, emotional responses), but System 2 provides the explanations.

System 2 does not have access to System 1’s actual processes. It observes System 1’s outputs (decisions, behaviours, preferences) and constructs post-hoc rationalisations. System 2 is the narrator, not the originator; it tells a story about what System 1 did, using whatever explanatory frameworks are culturally available.

In AIM terms, Common-Currency Integration operates at System 1 speed (preconscious, automatic), and I-Override is System 2’s interrogation of System 1’s output. But I-Override cannot observe the integration process itself; it can only observe the wanting-signal that emerges. When I-Override asks “Why?” it is asking System 1 to explain itself, and System 1 cannot answer because the source-tags were destroyed during integration. System 2 therefore constructs an answer using available Templates.

This dual-process picture receives a sharper substrate-level and information-theoretic specification in Chapter 8. In AIM terms, the relevant distinction is not simply between fast and slow cognition in the abstract, but between a fast loop that delivers only the Decision Hub’s tag-stripped integrated output to Conscious Cognition and a slow loop in which Intrinsic Motivation performs prior-maintenance and Bayesian ground-attribution over that output. The post-hoc character of explanation is therefore not merely a psychological tendency but a direct consequence of the architecture.

Haidt: The Emotional Dog and Its Rational Tail

Jonathan Haidt’s social intuitionist model extends this architecture to moral judgment. In “The Emotional Dog and Its Rational Tail” (2001), Haidt argued that moral judgments are produced by rapid, automatic intuitions (System 1), and moral reasoning is the post-hoc construction of justifications (System 2). The reasoning does not cause the judgment; it explains and defends a judgment that was already made.

Haidt’s research showed that when moral intuitions conflict with reasoning (e.g., harmless taboo violations that feel wrong despite producing no victim), people maintain their intuition and search for new reasons. When one justification is defeated, they generate another. This pattern, conclusion-first, reasoning-second, matches AIM’s Confabulation architecture precisely.

The AIM Framework extends Haidt’s insight: moral reasoning about others’ motivations operates through the same post-hoc construction as reasoning about one’s own motivations. When the Individual asks “Why did they do that?” the brain generates an answer using available Templates, and the M-Template is absent from virtually everyone’s repertoire.

Mercier and Sperber: The Argumentative Theory of Reasoning

Hugo Mercier and Dan Sperber’s argumentative theory of reasoning (2011) proposes that human reasoning evolved not for truth-seeking but for argumentation, persuading others and evaluating others’ arguments. This explains why reasoning exhibits systematic biases: confirmation bias, motivated reasoning, and the failure to correct one’s own errors are not bugs but features of a system designed to win arguments rather than reach Truth.

On this account, the brain’s narrative-generation capacity evolved primarily as social technology: for coordinating with others, justifying actions to the group, and building coalitions. Truth-tracking is at best a secondary function, and one that operates only under specific social conditions (when one must defend one’s claims against critical interlocutors).

AIM incorporates this insight: the Confabulation system is not malfunctioning when it generates inaccurate explanations. It is functioning exactly as designed, generating persuasive narratives for social coordination. The problem is that social coordination through shared false Beliefs produces the Scapegoat Mechanism when those Beliefs converge on agent-based causation.

How AIM Extends the Literature

The psychological evidence establishes that:

Humans routinely generate inaccurate explanations for their behaviour
These explanations feel like accurate self-knowledge
Introspection does not provide reliable access to causal processes
Confabulation is sincere, not deceptive
The mechanism operates in healthy brains as routine cognition

AIM accepts these findings and adds:

Source-specific opacity: M is architecturally hidden while A and I have some (imperfect) phenomenological markers. This explains why certain desires are systematically confabulated while others are not.
The Vocabulary Gap: Even if introspection were possible, the M-Template is absent from most Individuals’ conceptual repertoires, ensuring that accurate M-attribution cannot occur.
Architectural mechanism and localisation: Common-Currency Integration in the Decision Hub destroys source-tags during valuation, leaving I's Bayesian ground-resolution function (the operation by which I infers source-grounds for the integrated wanting-signal) to infer the source from whatever Hypothesis Space it holds. This is why the source decomposition is unavailable, not merely difficult to access but architecturally absent from the conscious signal, and Confabulation is the named output of that resolution when the M-Hypothesis is missing from or suppressed within I's available candidates (Chapter 2, Confabulation; Chapter 4, Sections 4.4.0 and 4.4.3). The Bayesian updating itself is rational and content-agnostic; the pathology is the impoverishment of the Hypothesis Space on which it operates.
Projection as extension: The same Confabulation architecture applies to interpreting others, producing systematic M-blindness at population scale.
Collective pathway: Individual Confabulations propagate through social networks (prestige-bias, social proof), converge toward shared narratives (Confabulation Cascades), and complete as agent-based attributions (Scapegoat Mechanism) under Information Gap conditions.

AIM does not contradict the psychological literature; it explains why the phenomena occur, identifies which motivations are systematically confabulated (M-driven), and traces the collective consequences when entire populations confabulate the same Source Opacity (Scapegoat Mechanism, Mimetic crisis).

MOVEMENT A: INDIVIDUAL CONFABULATION

11.1 The Information Gap and Vocabulary Deficiency

Every Individual experiences wanting. The phenomenology is unified, I want X, regardless of whether X arose from Appetitive deficit (bodily need), Intrinsic engagement (self-endorsed activity), or Mimetic transmission (preconscious acquisition from Models). Chapter 9 established that Source Opacity is uniform across all three sources: Common-Currency Integration in the Decision Hub collapses unlike inputs into a single Common-Currency Scalar (the Scalar) before that signal reaches conscious awareness. By the time the Individual experiences wanting, the source-tag, the information specifying which motivational system generated the signal, has been destroyed. Under this destruction, I's Bayesian ground-resolution function, the operation by which I infers source-grounds for the integrated wanting-signal arriving at the Decision Hub, must operate over whatever Hypothesis Space about M it currently holds; this localisation is the canonical site of Confabulation and is developed across the remainder of this Movement, formalised in the Confabulation Rationality Lemma at §11.2, and threaded into the coupled-loop derivations of Projection and the Scapegoat Mechanism in Movements B and C (Chapter 2, Confabulation; Chapter 4, Sections 4.4.0 and 4.4.3).

Note (Appendix A cross-reference). The Confabulation architecture developed in this chapter is treated in Appendix A.6.3 as a load-bearing commitment whose weakening would constitute a failure mode of the fast valuation formalism ("preserved source tags downstream"). The appendix's formalism is compatible with external instrumented recovery of source information (which is what the framework's behavioural diagnostics exploit) but not with preservation of source tags within the Individual's introspective access. The two treatments are consistent: this chapter establishes the Confabulation architecture at the verbal level; the appendix exhibits it as a formal requirement of the fast valuation dynamics.

This creates an Information Gap: the Individual is asked (by social context, deliberate reflection, or decision-making demands) to explain why they want X, but the source decomposition that would answer it is not in the integrated signal. The gap is not a failure of introspection; no amount of introspective effort can read off that signal a decomposition it never encoded. The answer can only be inferred, from the Scalar's course under contrasts, and only by an Individual who holds the M-Hypothesis. The gap is structural.

The Vocabulary Deficiency

The Vocabulary Deficiency compounds this problem. Consider the available explanatory categories in everyday discourse:

Appetitive vocabulary is rich and inferentially transparent. “I’m hungry,” “I need sleep,” “I’m cold,” “I’m in pain.” These map reliably to bodily states that Individuals can observe: stomach contractions, fatigue, shivering, nociceptor activation. When someone says “I’m hungry,” they are usually correct because hunger announces itself through phenomenology accessible to introspection.

Intrinsic vocabulary is reasonably developed. “I’m curious,” “I enjoy this,” “It’s meaningful to me,” “I find it rewarding.” These map to engagement states that exist in real-time. An Individual knows if they are having fun. The experience itself is what is engaging. Because these types of experience can persist in private and (if they allow mastery) deepen with competence, the audience-removal test (Would I still want this if no one ever saw me do/possess it?) provides a reliable heuristic.

Mimetic vocabulary is effectively absent. The concept “what I am feeling is a reading of what those around me treat as valuable, not a shortfall in me, and it was integrated into my valuation of the Object before I became aware of wanting it” does not exist in vernacular explanatory frameworks. The word “envy” captures a narrow slice (wanting what a rival has), but it carries moral stigma and does not describe the architectural mechanism. “Keeping up with the Joneses” acknowledges the phenomenon colloquially but provides no actionable category for Source-tagging. The term “peer pressure” implies conscious awareness of external influence and volitional resistance: neither of which applies to preconscious Mimetic transmission.

The Vocabulary Gap Is M-Specific

The result: when the Individual attempts to explain a Mimetically-acquired desire, the correct answer is not among the available responses. The concept and vocabulary required to identify the source does not exist in their cognitive repertoire. This is not ignorance in the sense of lacking exposure to a concept they could acquire through ordinary experience; Mimetic transmission is preconscious (operating within the Transmission Window, 60–340ms), and its operation leaves no phenomenological residue accessible to introspection. The Individual cannot discover the Mimetic source through careful introspection, because there is no residue for introspection to find, and the signal that did arrive presents as a lack, which is what a deficit would also present as.

Appetitive and Intrinsic misattributions occur (confusing hunger for interest, or interest for social obligation), but these are inferentially correctable through ordinary experience. If someone misattributes hunger as aesthetic preference (“I’m buying this pastry because it’s artfully made”), eating the pastry will reveal the error through satiation (they may be unable to finish eating it because they are already full). If someone misattributes Intrinsic engagement as social duty (“I’m practising piano because my parents and teacher expect it”), removing the audience will reveal whether the activity remains rewarding. In this case it would mean observing whether the Individual practises or plays the piano at any point of their own volition, as opposed to being told to, or being held to a schedule for practice, lessons, or performance.

Mimetic misattribution, by contrast, is not corrected by ordinary experience: acting on a Mimetically-acquired desire (buying the house, pursuing the credential, adopting the status marker) does not end it, because Mimetic desires do not satiate, and without the M-Hypothesis the persistence is read as further A- or I-evidence. The contrasts that would reveal the lead (acquisition completed, audience removed) inform only an Individual who holds the M-Hypothesis.

Cross-Cultural Evidence for the Vocabulary Gap

The Vocabulary Gap is not a quirk of Western languages. Anthropological research confirms that while cultures develop rich vocabularies for bodily states (A) and for what activities are valued (I, though often conflated with social role), explicit vocabulary for “desires acquired through observation of others” is rare across human societies.

Some traditions approach M-recognition:

Buddhist tanha (craving, thirst) captures the endless quality of certain desires but does not distinguish Mimetic from Appetitive craving
Sanskrit spṛhā (longing for what another possesses) approaches envy but carries pejorative moral weight
Girard’s désir mimétique is the only systematic treatment, and it emerged from literary criticism in the twentieth century, not from vernacular experience

Every human society has developed vocabulary for hunger, thirst, cold, pain, fear, interest, joy, and meaning. Virtually none has developed vernacular vocabulary for “wanting that arises from observing another wanting.” The concept requires explicit theoretical construction; it does not emerge from ordinary introspection.

This cross-cultural absence is itself evidence for Source Opacity. If Mimetic transmission were introspectively accessible, at least some cultures would have developed rich vocabulary for it. The universal Vocabulary Gap suggests that the information required for M-recognition is architecturally unavailable to introspection.

The Critical Implication

Under these conditions (Source Opacity plus Vocabulary Deficiency) Confabulation is the expected outcome for M-driven desires. The Individual faces an Information Gap that cannot be closed through introspection and a Vocabulary Gap that precludes accurate response. What follows is not optional.

11.2 The Confabulation Rationality Lemma

The Confabulation Rationality Lemma follows from the architectural localisation of Confabulation at I's Bayesian ground-resolution step: Confabulation is not a failure of rationality, it is the correct operation of a content-agnostic updating process on defective input.

Bayesian updating is rational in the technical sense that it assigns posterior probability to Hypotheses in proportion to their prior probability weighted by the likelihood of the observed evidence under each Hypothesis. The operation is content-agnostic, it works identically regardless of whether the Hypothesis that best explains the evidence is present in or absent from the Hypothesis Space. When the best-fitting Hypothesis is present, the updating returns it with appropriate confidence. When the best-fitting Hypothesis is absent, the updating returns the most probable available Hypothesis with the same kind of confidence, because the updating operation has no mechanism for detecting its own incompleteness.

In the AIM architecture, I's Bayesian ground-resolution takes the tag-stripped integrated Scalar from the Decision Hub and returns a ground-attribution from its Hypothesis Space, A-ground, I-ground, or neither. Under Source Opacity, the integrated Scalar carries the three contributions only as their sum and does not encode their decomposition, so I has no access to the source-decomposition that would otherwise resolve the Salience directly. I must infer the ground from the available Hypotheses. If I's Hypothesis Space about M is well-populated and the M-Hypothesis has not been outweighed by an A-ground or I-ground attribution that the Individual and their reference group repeatedly restate, M-Salience can be recognised as M-Salience, and the resolution returns the correct ground-attribution. If I's Hypothesis Space about M is impoverished, absent, or suppressed (the Vocabulary Gap, treated in Chapter 10), the M-Hypothesis is simply not on the list of candidates the updating can return. I returns the most probable available candidate, which under almost all conditions is an A-ground or an I-ground, and it returns that candidate with the confidence appropriate to a well-functioning Bayesian update. That A-ground or I-ground completion is Comfortable Confabulation: a confabulated explanation that achieves cognitive closure while avoiding implications that would threaten the Individual's self-concept, Status, or group membership, and the architecture systematically selects it over an M-attribution.

Input is defective in two ways, and the Confabulation Rationality Lemma holds under either. The first is an incomplete Hypothesis Space, Type I Exclusion (Structural Exclusion): the true Hypothesis is not among the candidates the updating can return, so it cannot receive mass however much Evidence arrives. Where the true Hypothesis is the M-Hypothesis, this is the case set out above, and the exclusion is architectural rather than incidental.

The second requires no absence, and is Type II Exclusion (Prior Dominance). Where the true Hypothesis is among the candidates, the mass it can accumulate is governed by the Belief-state each observation meets, and that state is the accumulated product of everything encountered before it (Chapter 14, Section 14.9.2). Two mechanisms concentrate mass on a competitor. The first is order: an attribution that occupied the explanatory slot early is weighed as background by what arrives after it rather than as a rival.

The second is reaffirmation, and it operates on every cycle rather than once. The incumbent attribution is restated to the Individual from two directions, from the Belief Register's own prior state recalled from memory and re-entering as Evidence under the Self-Coupling Permanence Principle (Chapter 14, Section 14.5), and from the correlated attributions of other Individuals arriving through interpersonal coupling (Chapter 14, Section 14.6.2). Neither restatement is marked as one. Under Coupling-Induced Opacity the Individual holds no channel distinguishing endogenous Evidence from exogenous Evidence, so a Hypothesis repeated often registers as a Hypothesis widely corroborated, and the mass that Hypothesis accumulates tracks the count of its assertions rather than the weight of what supports them. Each increment of concentration lowers the Interference Term, and a lower Interference Term admits less movement from the next observation, so the concentration extinguishes the capacity that would have reversed it (Chapter 14, Section 14.9.3).

The Individual in this case holds the M-Hypothesis. The M-Hypothesis is formulable, is available to every cycle of ground-resolution, and is still never returned. The failure is not that the candidate list lacks the M-Hypothesis. It is that a competitor reached the explanatory slot first and is restated on every cycle, by the Belief Register itself and by the reference group whose attributions arrive as Evidence. Where that group repeatedly attributes wanting to Appetitive deficit or Intrinsic engagement, the M-Hypothesis can sit in the Hypothesis Space indefinitely without ever accumulating the mass that would let ground-resolution return it.

The Confabulation Rationality Lemma's phenomenological consequence is direct. Confabulation is confidently held because confidence is what correct Bayesian updating produces under good priors and strong evidence, and I has both. Confabulation is resistant to counter-evidence for a reason that differs by condition. Where the Hypothesis Space is incomplete, counter-evidence re-weights the existing incorrect candidates against each other without introducing the correct one. Where the M-Hypothesis is present but dominated, counter-evidence arrives at a Belief-state whose Interference Term has already fallen, and the movement that counter-evidence can produce falls with it. Confabulation is not detectable by introspection because introspection is itself an I-operation, it draws on the same Hypothesis Space, and it therefore returns the same Confabulated attribution.

The Confabulation Rationality Lemma has a normative consequence as well. Because Confabulation is a rational operation over a Hypothesis Space that does not yield the M-ground, Confabulation cannot be corrected by instructing the Individual to reason better, to reflect more carefully, or to be more honest with themselves. The updating operation is not the problem. The intervention architecturally available is the Knowledge Patch (the Patch), and the Patch operates on both defects. Adding the M-Hypothesis to I's available candidates answers the first. Holding the Interference Term above zero, so that the added Hypothesis can accumulate mass against an incumbent renewed on every cycle, answers the second. Neither operation substitutes for the other, and no amount of individual effort or introspective discipline substitutes for either (Chapter 2, Confabulation; Chapter 4, Section 4.4.3; Chapter 13 for the Confabulation Cascade treatment that iterates this Lemma at the population scale; Chapter 18 for the Knowledge Patch operationalisation; Appendix A, Section A.5.6, for the Confabulation Rationality Lemma in mechanism form and the Restriction-Induced Concentration Factor it defines).

A specific consequence of the Confabulation Rationality Lemma arises once a Referent has decohered. A Belief held in contradiction of an available decohered Referent of traced provenance is a Confabulation, sincere and wrong, whose completion drifts toward A-grounds and I-grounds and systematically away from any acknowledgement of M, and whose maintenance is therefore Mimetically driven by elimination. This result, Referent Abandonment, is derived in Chapter 14, Section 14.8.3.1.

11.2.1 The Confabulation Proliferation Corollary

The Confabulation Rationality Lemma establishes what ground-resolution returns on a single pass. A further consequence follows from what happens when the operation is run repeatedly against the same explanandum under Type I Exclusion, where the M-Hypothesis cannot be generated at all.

Two defects operate together. The first is that the correct Hypothesis cannot be generated: the Hypothesis Space does not contain the basis vector corresponding to the M-Hypothesis, so its prior is not small but undefined (Chapter 14, Section 14.3). The second is that the incorrect Hypotheses cannot be eliminated. An A-misattribution is corrected by consumption and an I-misattribution by the removal of an audience, but an M-misattribution is discharged by neither, because the demand does not terminate on acquisition (Section 11.1). Evidence therefore arrives continuously and eliminates nothing.

The Thinker's activation condition is novelty and prediction-mismatch (Chapter 4, Section 4.4.3). M-driven Salience presents a prediction-mismatch that resolution cannot remove: the Object is obtained and the demand persists, because the Salience was never a deficit and no consumption or repair closes it (the Salience-Deficit Conflation Principle, Chapter 4, Section 4.4.0). The Thinker is thus recruited by a mismatch it cannot discharge, and generation of candidates continues for as long as the recruitment does. The generation is Imagination, the operation that constructs a candidate Hypothesis from Knowledge held in the Memory Store when Ground-Resolution returns no adequate ground. What the Confabulation Proliferation Corollary describes is therefore that operation running against an explanandum that never supplies the closure which would release it.

The candidate set therefore expands rather than converging, and it expands laterally rather than deepening. Because Mimetic Desire attaches to Objects already carrying Appetitive or Intrinsic function (Axiom 5), a single underlying cause presents on each occasion in a different Object and a different context. What the Individual encounters is not one recurring problem admitting one explanation, but a succession of apparently distinct problems, each recruiting its own candidate. The diffuseness is not a feature of the cause. It is an artefact of the cause being read through the Objects it has attached itself to.

This is the Confabulation Proliferation Corollary: under Type I Exclusion, where the M-Hypothesis is unavailable, ground-resolution over a permanently undischarged explanandum produces an unbounded and non-convergent set of A-grounded and I-grounded attributions, rather than a single stable misattribution. The Confabulation Proliferation Corollary claims non-convergence and lateral branching; it claims no rate of expansion.

Two consequences follow. The first is that the search requires an exogenous terminator, and Axiom 9 supplies it: an agent-attribution delivers the cognitive closure that no Appetitive or Intrinsic candidate can, which is why the search characteristically terminates on a person. The Scapegoat Mechanism of Movement C is that termination at population scale. The second is that installation of the M-Hypothesis does not merely add one candidate. It collapses the proliferated set, because the members of that set were competing attempts at a single explanandum that the added Hypothesis resolves.

The Confabulation Proliferation Corollary carries a falsification condition. It predicts that the count of distinct explanatory candidates an Individual generates for a recurring dissatisfaction declines following AIM Vocabulary Acquisition, and that the decline is specific to dissatisfactions with a Mimetic component rather than general across all explanatory behaviour. Absence of that decline, or its appearance uniformly across Appetitive and Intrinsic explananda, would disconfirm the Confabulation Proliferation Corollary while leaving the Confabulation Rationality Lemma standing.

11.2.2 The Collective Instance

The Confabulation Proliferation Corollary describes an Individual whose ground-resolution runs against an explanandum it cannot discharge. The same structure appears wherever a community reasons about a class of phenomena whose generator lies outside the Hypotheses available to it, and the history of the natural sciences supplies its clearest instances. Kuhn (1962) characterised the period preceding a scientific revolution as one in which anomalies accumulate and are met by successive ad hoc modifications of the existing framework, each addressing a local difficulty and none resolving the class. That is proliferation under a missing basis, described from the outside.

Four instances exhibit it. Ptolemaic astronomy accounted for retrograde planetary motion by adding epicycles, deferents and equants, each addition answering a particular observational discrepancy and the set growing without converging; the collapse came only when the Earth's own motion was admitted and, in Kepler's elliptical formulation, the apparent motions resolved into consequences of a single geometry. Pre-Darwinian natural history treated biogeographic distribution, anatomical homology, the fossil succession, vestigial structures and the results of artificial breeding as separate problems with separate explanations. The anomalous precession of Mercury's perihelion drew a modified inverse-square law, an assumed solar oblateness, and a postulated intra-Mercurial planet that was searched for and never found. Disease before the microbial account was addressed through miasma, humoral imbalance, spontaneous generation and zymosis, with puerperal fever, cholera and surgical sepsis standing as distinct conditions requiring distinct causes.

In each case the addition of a single Hypothesis reduced the count of explanations rather than increasing it. Descent with modification did not supplement the separate explanations of homology and biogeography; it retired them, and the facts they had been assembled to cover became evidence for one proposition. This is the pattern Whewell (1840) named the consilience of inductions, and it is the signature of a basis vector having been supplied rather than a parameter having been fitted: the space acquires a dimension and the candidate set contracts.

The parallel is structural and its force is limited to that. Proliferation of independent explanations for a class of phenomena is evidence that a Hypothesis is missing; it is not evidence of which Hypothesis is missing, and no historical precedent can establish the content of a candidate. What the precedent establishes is that the Confabulation Proliferation Corollary describes a documented epistemic condition rather than a contrived one, and that the diffuseness of an explanandum is a poor guide to the multiplicity of its causes. Where a cause attaches parasitically to Objects that already carry Appetitive and Intrinsic function, the appearance of many separate problems is what a single cause looks like from inside a Hypothesis Space that cannot name it.

11.2.3 The Exception-Handling Analogy

A program raises an exception when execution reaches a state the program cannot handle, and a handler catches the exception so that execution continues. Where the fault is an untraced change elsewhere in the code that accumulates over time, the handler answers the symptom and leaves the change in place. The exception is raised again, another handler is written, and each handler makes the next diagnosis harder, which is technical debt. The Exception-Handling Analogy reads Confabulation through that structure.

In the architecture the untraced change is a cause that keeps running where no default completion reaches it. The cost of the A-Floor rises as the Mimetic Premium spreads into the goods that correct Appetitive deficits, which is the A-Floor Erosion Principle (Chapter 7, Section 7.5.6), and valuations pass between Individuals by Preconscious Transmission, before any Individual can evaluate what was passed (Axiom 7), and compound across a reference group as the Confabulation Cascade (Section 11.12). Neither cause has a location an Individual could point to, and both continue whatever any Individual believes about them. The exception is the open question the untraced change produces: a felt lack or felt threat whose cause the Hypothesis Space does not hold, which recruits the Thinker and, under Type I Exclusion, goes on recruiting the Thinker, as the Confabulation Proliferation Corollary establishes. The handler is the ground-attribution that Ground-Resolution returns. Where the Agent-Based Causal Default supplies the completion, the handler names an agent (Chapter 4, Section 4.3.9); in the Intractability Confabulation the handler names no agent at all (Section 11.6). In either shape the completion runs away from the Mimetic ground, the direction Comfortable Confabulation takes. Execution continues because the handler satisfies the demand for closure, and the search ends before a systemic candidate has been built (Chapter 4, Section 4.3.10).

The exception recurs because the handler resets the loop without touching the untraced change. Where the handler names an agent, discharge onto the target reconfigures the Object field the Observer reads, the distributed cause stays where it was, and the Confabulation Cascade accumulates again toward a new target (Section 11.12). Where the handler names no agent, each new failure to explain is absorbed as further Evidence that no explanation exists (Section 11.6). The technical debt is Type II Exclusion accumulating on the handler's attribution. The attribution is restated on every cycle, by the Belief Register's own prior state and by the reference group, each restatement lowers the Interference Term, and the attribution becomes harder to revise with every run of the handler.

The Exception-Handling Analogy licenses recognition of the pattern, not inference about its causes, and fails at three points. No one writes the handler: the target is the coincident output of parallel resolutions, and the selection has no author at either scale (Section 11.12). No stack trace exists: a program can log where a fault arose, but integration keeps no record of which source produced a valuation (Axiom 3), so I-Override infers the source-attribution that a debugger would read off the log (Section 11.3). And the Exception-Handling Analogy supplies no rate and no threshold: how fast the debt accumulates, and whether the accumulation turns terminal, are set by the cascade dynamics of Chapter 13 and the crisis dynamics of Chapter 17.

11.3 I-Override as Confabulation Site

Chapter 9 introduced I-Override as the cognitive operation by which a motivational signal is redirected from automatic Decision Hub processing to deliberative Intrinsically Motivated scrutiny. I-Override engages when the Decision Hub encounters novelty, ambiguity, or cognitive dissonance: situations where cached retrieval cannot complete action selection and deliberative evaluation is required. Through I-Override, the signal is interrogated against Intrinsic values and Appetitive needs before action proceeds. This is the architectural mechanism through which Source-Tagging, the inference of the source-attribution, can in principle be performed, routing around Source Opacity rather than restoring the destroyed tags. This redirection is Ground-Resolution run deliberately on the source-attribution special case, within Conscious Cognition, the always-on inference process operative during waking experience; I-Override is Conscious Cognition running on an expanded Hypothesis Space with the interference term deliberately held above zero, while Confabulation is the same engine running on a restricted subspace with the interference term collapsed. Subsection 11.3.1 develops this formally.

I-Override is the intervention point for escaping Mimetic capture, where deliberative scrutiny can, under the right conditions, identify M-driven desires and refuse to act on them. It is also where Confabulation occurs.

The Process of Confabulation

The process unfolds as follows:

Step 1: Signal Interruption. The wanting signal arrives at the Decision Hub. Instead of passing directly to action selection, the Individual deploys I-Override. This interruption is deliberate and effortful; it requires cognitive resources, attentional capacity, and Intrinsic Motivation to scrutinise rather than act.

Step 2: Interrogation. The Individual asks Why do I want this? This is not rhetorical. The question demands an answer because downstream processes (action planning, resource allocation, social justification) require a causal model. The brain cannot proceed efficiently without some explanation for the wanting.

Step 3: Confabulation Generation. Facing the Information Gap (true source unknown due to Source Opacity) and Vocabulary Deficiency (Mimetic category unavailable), Ground-Resolution, closing on a subspace that excludes H_M, returns a plausible narrative built from the available categories. These narratives are not fabrications in the sense of conscious deception; they are the engine’s best-effort completion over a Hypothesis Space that excludes H_M.

Examples:

“I want this expensive house in a prestigious suburb because shelter is a basic need and this location provides security for my family.” (M→A misattribution: Mimetic desire reframed as Appetitive necessity)
“I want this prestigious credential because I’m genuinely interested in the field and the knowledge will enable valuable work.” (M→I misattribution: Mimetic desire reframed as Intrinsic engagement)
“I want this car because it’s a sound investment and reliable transportation is essential.” (M→A via instrumental reasoning)

NB: The value placed on an Object may have varying subjective AIM Weights contributed to by each Source.

Step 4: Incorporation. The Confabulation is incorporated into the Self-Model. The Individual now believes the narrative they have generated. This Belief is sincere; there is no inner Truth hidden beneath a false surface. The Confabulation becomes the Individual’s understanding of their motivation.

Step 5: Action. The Individual proceeds to act on the desire, now confident that their action serves Appetitive necessity or Intrinsic value. The wanting is executed as though it were A- or I-driven, even though its actual source was M.

The Paradox of Deliberation

The very operation designed to escape automatic processing and perform accurate Source-Tagging instead produces systematically inaccurate narratives when applied to Mimetic desires. Deliberation does not correct the error; it elaborates it. I-Override applied to an M-driven desire produces not “this desire is Mimetic” but rather “here is a sophisticated A- or I-consistent explanation for this desire that I have carefully considered and now endorse.”

Individuals with high cognitive capacity, strong introspective habits, and explicit commitment to self-knowledge are not protected. In fact, they may be more vulnerable, because their I-Override generates more elaborate and internally coherent Confabulations. A sophisticated thinker can construct a philosophically rich narrative about authentic selfhood, meaningful work, and considered values: all of which rationalises a Mimetically-acquired desire for a prestigious position. The sophistication of the reasoning masks the absence of the M-Hypothesis that would let the reasoning return the Mimetic origin.

Why I-Override Fails Without AIM Vocabulary

The failure is not in the operation itself but in the inputs available to the operation. I-Override functions by interrogating the motivational signal using conceptual categories and heuristic tests. Without the AIM taxonomy (A, I, M as distinct sources), the Individual cannot ask the diagnostic questions:

Do I want this because my body needs it? (Appetitive test)
Would I still want this if no one could ever observe me having/doing it? (Intrinsic test)
Am I wanting this because I have observed a Model wanting it? (Mimetic test, but unavailable without the concept of M)

Lacking the third category, I-Override reduces to a two-option test: A or I? And because many Mimetic desires carry urgency (status threats activate alarm systems) and appear in domains relevant to identity and values (career, housing, relationships), they fail to map cleanly onto Appetitive need yet feel too important to dismiss. Conscious Cognition, operating on the restricted subspace {H_A, H_I}, resolves this ambiguity by generating an I-consistent narrative that incorporates the urgency: “This matters to me because it aligns with my core values.”

I-Override without AIM vocabulary is therefore structurally insufficient. It can successfully identify clear A-signals (hunger, pain, fatigue) and clear I-signals (activities that remain rewarding in private). But for M-signals (which constitute the majority of desires in High-Observability, positionally-scarce modern environments) I-Override becomes the production site of Confabulation rather than its corrective.

Chapter 18 establishes a further consequence of this finding. When AIM vocabulary is available (expanding the representable Hypothesis Space from {H_A, H_I} to {H_A, H_I, H_M}), even the default, non-deliberative Confabulation output is less wrong, because the Confabulation Rationality Lemma's projection onto a larger subspace is necessarily at least as close to Truth as the projection onto the smaller one (the Reduced Confabulation Error Corollary, Section 18.4.3). Deliberative deployment of the Two-Question Diagnostic reduces the error further. The passive benefit operates even in the absence of trained diagnostic capacity, which is why AIM Vocabulary Acquisition is architecturally prior to all other Knowledge Patch operations.

Confabulation Is Definitionally Inaccurate

If a causal inference about motivational source reaches an accurate result, it is not Confabulation. The term Confabulation applies exclusively to inaccurate outputs: sincere but false narratives. The I-Override process may sometimes produce accurate results (correctly identifying A-driven or I-driven desires), and it sometimes produces inaccurate results (misattributing M-driven desires). Only the inaccurate outputs are Confabulations.

The distinction prevents conceptual slippage. Confabulation is not any post-hoc explanation or any narrative generated during I-Override. It is specifically and only inaccurate post-hoc explanation. The process may be routine and architecturally constrained, but the label applies to outputs, not to the process itself.

11.3.1 I-Override Within Conscious Cognition

I-Override is Ground-Resolution run deliberately on the source-attribution special case. Conscious Cognition, the always-on inference process operative during the Individual’s waking phenomenological experience, comprises two coupled operations: a generative operation, Imagination, which constructs candidate Hypotheses and delivers them into the representable Hypothesis Space $\mathcal{H}$ without scoring them, and an evaluative operation, Ground-Resolution, which redistributes probability mass across Hypotheses in $\mathcal{H}$ according to evidence and the interference term. The outputs of Conscious Cognition are distinguished by two features: whether $\mathcal{H}$ contains the true generative Hypothesis, and the magnitude of the interference term sustained at the moment of output.

Under this architecture, Confabulation and I-Override are not separate faculties. They are two configurations of the single engine.

In the Confabulation configuration, probability mass has collapsed to near-unity on a Hypothesis either orthogonal to the true generative Hypothesis (Type I Exclusion) or dominant by prior over a true Hypothesis received as evidence (Type II Exclusion); the interference term has decohered to zero; the output is assigned to the Belief register; and action is taken within the timescale imposed by closure-demand before further Hypothesis search completes.
In the I-Override configuration, the interference term is deliberately held above zero (Epistemic Suspension); S has been expanded (most importantly through AIM Vocabulary Acquisition) to admit H_M and other structural alternatives; and the output is held provisionally in the under-evaluation register pending further evidence.

The same wanting-signal, arriving at the Decision Hub with its source-tag destroyed, can produce either configuration depending on whether the engine is permitted to close rapidly on its most-available Hypothesis or held open long enough for structural alternatives to compete. This is why the same operation is both the intervention point for escaping Mimetic capture, held open long enough for accurate source-attribution to complete, which is I-Override, and the site at which Confabulation is produced (the engine, if allowed to close rapidly on a subspace that excludes H_M, generates sophisticated A- or I-consistent narratives at full confidence). The difference between correction and elaboration is not a difference of faculty but a difference of configuration, determined by whether the interference term is held open and whether S has been expanded to contain the true Hypothesis direction.

This formulation carries through to the Sincerity Guarantee. The engine sincerely produces whatever Hypothesis probability mass has collapsed upon, because sincerity tracks the probability profile and the Interference Term at the moment of output, not the pathway by which the engine arrived at them. A Confabulated narrative produced during deliberative I-Override on a restricted subspace is sincerely believed for the same architectural reason a Confabulated narrative produced without deliberation is sincerely believed: the engine has done its updating work correctly given the inputs available to it, and probability mass has collapsed onto the output, so it is held as Belief.

11.4 Epistemic Suspension as Alternative

The architectural necessity of Confabulation is not absolute. An alternative exists: the Individual can respond to “Why do I want this?” with “I don’t know” and sustain that state of uncertainty without generating a closure narrative. This capacity, Epistemic Suspension, is the foundational skill required for I-Override to function correctly under Source Opacity conditions. The teachable, domain-general ability to sustain Epistemic Suspension is Epistemic Suspension Capacity.

Why “I Don’t Know” Was Once Impossible

Evolutionary pressures once demanded immediate causal explanations because action could not be deferred. If the rustling in the bushes might be a predator, the organism must decide (fight, flee, freeze) now, and that decision requires a causal model (predator vs. wind?). Indecision in ancestral environments carried survival costs.

Natural selection therefore favoured cognitive systems that generate closure rapidly, even if that closure is sometimes inaccurate. A false positive (mistaking wind for predator) was metabolically cheaper than a false negative (mistaking predator for wind).

This selection pressure produced a disposition toward rapid closure that persists in contemporary populations. Psychological research identifies this disposition as Need for Closure (NFC): the tendency to seek definite answers and experience discomfort with ambiguity. Individuals vary substantially in NFC, and this variation appears largely stable across situations with heritable components. NFC is, to a significant degree, a trait rather than a choice.

Why “I Don’t Know” Is Now Possible

Modern environments have largely removed the pressures that made rapid closure mandatory. An Individual deliberating whether to purchase a house, pursue a credential, or adopt a status marker faces no immediate survival threat. The decision can be deferred. Information can be gathered. Uncertainty can be tolerated for days, weeks, or indefinitely. The phrase “I don’t know why I want this” satisfies the architectural demand for a response (it is a complete sentence, it closes the interrogation, it provides an answer), while leaving the actual causal question open for investigation.

This environmental shift creates the opportunity for Epistemic Suspension that ancestral conditions precluded. The disposition toward rapid closure remains; it was selected for and is heritable. But the environmental pressure that made rapid closure necessary has relaxed, creating space for a compensatory capacity to develop. The Individual can now train tolerance for a discomfort that evolution built in but that modern conditions no longer require acting upon.

How “I Don’t Know” Functions as Closure

Conscious Cognition’s demand for causal explanation is not inherently a demand for accurate explanation; it is a demand for any explanation that allows downstream processes (threat response, resource allocation, social coordination) to proceed. In ancestral contexts, “I don’t know” was unstable because it left the organism in an alarm state without actionable direction. But in modern contexts, “I don’t know” can function as stable closure if the Individual:

Accepts uncertainty as tolerable, recognising that not knowing the source of a desire does not prevent functional decision-making. The Individual can still evaluate whether acting on the desire serves their long-term interests, even without knowing why the desire arose.
Treats “I don’t know” as informative; uncertainty itself is data. If the Individual cannot identify an Appetitive need or Intrinsic value that explains the wanting, that absence suggests the desire may be Mimetically transmitted, even if the Individual lacks the formal M-concept.
Defers action pending investigation: instead of acting immediately to resolve discomfort, the Individual treats uncertainty as a signal to gather more information: observing whether the desire persists when Models are removed, testing whether the Object satisfies after acquisition, consulting others who have pursued similar Objects.

Epistemic Suspension as Interference-Term Maintenance

Epistemic Suspension is the operation of holding the interference term of Conscious Cognition above zero. The interference term is the quantity that sustains residual probability mass on competing Hypotheses within S; when it decoheres to zero, probability collapses onto a single Hypothesis and the output is assigned to the Belief register. Confabulation is the name for that output when the collapse occurs on a Hypothesis that is either orthogonal to the true generative Hypothesis or dominant by prior over it. Epistemic Suspension prevents this collapse by maintaining the conditions under which multiple Hypotheses, including the true one once it has entered S, can continue to compete.

Three consequences follow directly.

First, Epistemic Suspension is the operation that makes the generative capacity of Conscious Cognition available for truth-discovery. The generative operation (Imagination) continuously constructs candidate Hypotheses, but under rapid closure those candidates are absorbed by the dominant Hypothesis before they can accumulate amplitude. When the interference term is held above zero, candidate Hypotheses can accumulate amplitude long enough to be tested. Epistemic Suspension therefore does not suppress generation; it preserves the conditions under which generation is consequential.

Second, Epistemic Suspension is compositional with AIM Vocabulary Acquisition. Vocabulary Acquisition expands S to admit H_M; Epistemic Suspension maintains the interference term long enough for mass to flow from the dominant prior toward H_M as diagnostic evidence accumulates. Vocabulary Acquisition without Epistemic Suspension delivers a Hypothesis direction that the engine collapses past before it can accumulate mass. Epistemic Suspension without Vocabulary Acquisition maintains doubt across a subspace that still does not contain H_M, so the sustained doubt distributes mass only across A and I directions and never reaches the true source.

Third, the “I don’t know” output is stable as closure because it is an honest report of the state of the Hypothesis Space. The engine returns an output (“I do not yet have a Hypothesis at near-unity probability on this source-question”) rather than collapsing onto the best available candidate. This satisfies the architectural closure-demand while leaving the substantive probability distribution open. The AIM taxonomy transforms “I don’t know” into “I don’t know yet, but I have a method for finding out” by supplying the Hypothesis directions against which diagnostic evidence can be weighed.

The Dispositional Challenge

The NFC research poses a challenge for Epistemic Suspension. If NFC is dispositional, high-NFC Individuals will experience “I don’t know” as architecturally uncomfortable: the state generates urgency, discomfort, and pressure toward resolution that does not arise from cultural training alone but from stable features of their cognitive architecture. Can such Individuals develop Epistemic Suspension, or is the capacity limited to those who start with low NFC?

The answer lies in distinguishing trait change from compensatory capacity. Epistemic Suspension does not require eliminating NFC, reducing the dispositional pressure toward closure. It requires developing the functional equivalent: the ability to sustain uncertainty despite the discomfort, achieving the same behavioural outcome (deferred closure, sustained inquiry) through a trained pathway rather than dispositional ease.

The Functional Equivalent Through Trained Tolerance

The mechanism operates as follows:

The discomfort remains. High-NFC Individuals will feel urgency for closure when facing “I don’t know.” This feeling is dispositional and may not diminish substantially.
Tolerance is trained. Through deliberate practice, the Individual learns to tolerate that discomfort rather than immediately act to eliminate it. The discomfort is acknowledged as real but not action-compelling: a signal that can be observed without being obeyed.
Reframing transforms the state. The intolerable version of uncertainty is “I don’t have an answer and cannot proceed.” The tolerable version is “I don’t have an answer yet, and I can hold this question open until sufficient information arrives.” The state is identical; the framing determines whether it compels immediate closure or permits sustained inquiry.
Neuroplasticity enables pathway formation. Repeated practice (deliberately tolerating uncertainty, observing the discomfort without acting on it, successfully sustaining open questions until resolution becomes possible) creates neural pathways that make the tolerance less effortful over time. The trait pressure persists, but the compensatory capacity strengthens.

Mindfulness training creates space between stimulus and response: the automatic reaction still fires, but a trained pause intervenes. Cognitive-behavioural approaches restructure automatic thoughts; the initial thought still arises, but an alternative pathway becomes available. Exposure therapy reduces phobic responses; the fear signal still generates, but habituation reduces its behavioural impact. Epistemic Suspension operates through the same mechanism: the NFC signal still generates, but trained tolerance reduces its closure-compelling force.

A Training Protocol Sketch

While Chapter 18 provides the full operationalisation of the Knowledge Patch, the basic elements of Epistemic Suspension training can be sketched here:

Stage 1: Recognition. The Individual learns to notice when they have generated a causal explanation for a desire. The question is not whether the explanation is correct but whether the Individual generated it rapidly, without investigation, to achieve cognitive closure.

Stage 2: Labelling. When a rapid explanation is noticed, the Individual labels it: “That was a closure move.” This labelling creates psychological distance; the explanation is no longer experienced as discovered Truth but as cognitive operation.

Stage 3: Holding. The Individual practises holding the question open after noticing the closure move. “I generated that explanation quickly. What if I don’t know the actual answer?” The discomfort is observed without immediate action.

Stage 4: Investigation. With the question held open, the Individual can now apply diagnostic tests: Does this desire pass the audience-removal test? Does it persist when Models are absent? Would satiation occur if the Object were acquired?

Stage 5: Iteration. The process becomes habitual through repetition. Each successful holding-and-investigation strengthens the neural pathway, making subsequent Epistemic Suspension less effortful.

The Difficulty Gradient

This framing implies a difficulty gradient. Low-NFC Individuals will find Epistemic Suspension relatively natural; the dispositional pressure toward closure is weak, and sustaining uncertainty requires little effort. High-NFC Individuals will find it harder; the dispositional pressure is strong, and sustaining uncertainty requires conscious deployment of trained capacity, especially under stress, fatigue, or time pressure when compensatory mechanisms are resource-depleted.

The gradient has implications for Knowledge Patch dissemination:

Training is necessary for most. Because NFC varies and skews toward closure-seeking in many populations, Epistemic Suspension will not emerge spontaneously for most Individuals. Deliberate practice is required.
Difficulty varies but capacity is achievable. High-NFC Individuals can develop functional Epistemic Suspension, but they will experience Epistemic Suspension as effortful skill deployment rather than natural state. This is acceptable; the behavioural outcome (sustained inquiry, deferred closure) is what matters, not the phenomenological ease of achieving it.
Stress degrades performance. Under conditions that elevate NFC (time pressure, threat, cognitive load), even trained Individuals may revert to premature closure. Institutional design should account for this; high-stakes decisions should not be made under conditions that maximally elevate NFC.

AIM Vocabulary as Compensatory Tool

Even if an Individual possesses Epistemic Suspension capacity, “I don’t know” is an unstable endpoint without conceptual vocabulary that makes the uncertainty actionable. Saying “I don’t know why I want this house” leaves the question open, but it provides no path forward. The Individual remains in an uncomfortable liminal state that closure pressure will eventually overwhelm.

The AIM taxonomy resolves this by providing the categories that transform “I don’t know” into “I don’t know whether this is A, I, or M, but I can test each Hypothesis”:

Appetitive test: If my survival were secure (housing, food, safety guaranteed), would I still feel urgency about acquiring this specific house? If no → A-driven.
Intrinsic test: If no one could ever observe me living here, would I still want this house? If yes → I-driven.
Mimetic test (by elimination): If neither Appetitive nor Intrinsic explanations hold, the desire is likely M-driven, acquired through observation of Models and amplified by status competition.

With AIM vocabulary, “I don’t know” becomes “I don’t know yet, but I have a method for finding out.” Uncertainty is no longer a dead end; it is the entry point for diagnostic investigation.

For high-NFC Individuals, this reframing may be essential. The open question (“Why do I want this?”) is closed not by generating a Confabulation but by generating an investigation plan. The closure achieved is methodological rather than substantive: “I don’t know the answer, but I know how to find out.” This satisfies closure demands sufficiently to tolerate the substantive uncertainty while the investigation proceeds.

The Knowledge Patch therefore serves dual functions: it provides the vocabulary required for accurate Source-tagging (its primary purpose), and it provides the structured inquiry that makes uncertainty tolerable for high-NFC Individuals (its secondary but practically essential function). Without the taxonomy, “I don’t know” is a dead end that closure pressure will not permit. With the taxonomy, “I don’t know yet” becomes an entry point for diagnostic investigation that closure pressure can accept.

Three Barriers to Widespread Adoption

Three barriers prevent widespread adoption of “I don’t know” as viable closure:

Dispositional variation: High-NFC Individuals experience uncertainty as aversive at a trait level. Training can develop compensatory capacity, but the training is effortful and the capacity degrades under stress. Many Individuals will find Confabulation (rapid closure through narrative generation) less costly than Epistemic Suspension (sustained discomfort requiring trained tolerance).
Vocabulary absence: Without the AIM taxonomy, “I don’t know” provides no actionable path forward, making it an uncomfortable and seemingly unproductive state. The Individual lacks the diagnostic tools that would transform uncertainty into structured investigation.
Social costs: Expressing uncertainty about one’s own motivations invites social sanctions: others interpret it as indecisiveness, lack of self-knowledge, or evasion. These costs make Confabulation (confident but inaccurate narrative) socially cheaper than honesty (uncertain acknowledgment of Information Gap).

The Knowledge Patch intervention (Chapter 18) addresses all three barriers: training Epistemic Suspension through deliberate practice to develop compensatory capacity even for high-NFC Individuals, disseminating AIM vocabulary to make uncertainty actionable, and creating social contexts where “I don’t know, let me investigate” is treated as epistemically virtuous rather than socially deficient.

11.5 Confabulation in Social Interaction

The preceding analysis focused on Confabulation as it applies to desires for Objects: "why do I want this house?" generates a Confabulated narrative because Source Opacity prevents access to the M-component of the desire. But Confabulation operates with equal force on a domain that the desire-focused analysis does not cover: social interactions with other Individuals.

When Individual A defers to Individual B (granting B greater credibility, directing disproportionate attention toward B, yielding positional priority to B), A's cognitive system requires a causal explanation for the deference, just as it requires a causal explanation for any motivated behaviour. The true cause may include a substantial PM-component: A defers partly because A has observed others deferring to B, and that social proof was integrated into A's valuation of B preconsciously, before A could evaluate whether B's expertise or character warranted the deference independently. But Source Opacity has destroyed the record of this PM-contribution in A's integrated valuation. The Information Gap is identical in structure to the desire case: the true source-composition of A's valuation of B is unavailable to introspection.

Confabulation fills the gap. A generates a narrative: "I defer because they clearly have deep expertise in this area" (I-sounding attribution), or "I defer because they've earned their position through years of hard work" (mixed A-I attribution), or "I defer because their track record speaks for itself" (evidence-sounding attribution that is actually social-proof-constituted). These narratives are sincere. A genuinely believes them. They are Confabulations in exactly the sense defined earlier in this chapter: sincere but inaccurate post-hoc explanations generated to fill an Information Gap created by Source Opacity.

The interaction-level domain matters because it is the site where PM-transfer occurs. The four channels through which Mimetic Premium flows between Individuals (attention, deference, narrative authority, positional concession) are all social-interaction behaviours. If Confabulation renders these transfers invisible to the transferring party, then PM extraction operates without constraint from the Individual's deliberative system. The Individual cannot deploy I-Override against a transfer they do not recognise as occurring. This is why PM flow is described as architecturally invisible (Chapter 8): the invisibility is not metaphorical but architectural, produced by Source Opacity operating at the interaction level and Confabulation providing the covering narrative that makes the transfer feel like "appropriate respect" or "earned recognition" rather than PM extraction.

The Projection analysis that follows (Movement B) inherits this extension. When Individuals project motivational Templates onto others, they project not only Templates for others' desires but Templates for others' social interactions. "They defer to him because he's brilliant" is Projected Confabulation: the observer attributes I-driven deference to the observed interaction because the I-Template is available and the M-Template (they defer because everyone defers, and the social proof was integrated preconsciously) is not.

11.6 The Intractability Confabulation

Movement A has so far developed Confabulation forms that complete as agent-based causation. There is a further Confabulation form that terminates causal inquiry without identifying any agent at all. The Individual, or more commonly the discipline, confronts a phenomenon whose true causes exceed the computational tractability of available causal-inference systems, and the Confabulation architecture produces the closure humans are too complex to be properly modelled. The narrative is sincere, it is phenomenologically indistinguishable from epistemic humility, and it achieves cognitive closure without generating a visible victim. These properties make it the Confabulation of choice in educated and professional contexts where agent-based attribution carries reputational risk.

The Intractability Confabulation is structurally identical to the agent-based completions already derived in this chapter. The Intractability Confabulation is exception handling in the sense of the Exception-Handling Analogy (Section 11.2.3), a handler that permits cognitive closure where no default completion reaches the root cause. The difference is only in the content of the handler. Where scapegoating discharges the explanatory demand by identifying an agent, the Intractability Confabulation discharges it by declaring that no explanation is available in principle. Both are generated by the same narrative-closure mechanism operating on the same Information Gap, and both produce the same architectural consequence, the root cause remains unaddressed while the Individual experiences the matter as settled.

The Intractability Confabulation is diagnostically distinguishable from genuine epistemic humility by a single structural feature. Genuine epistemic humility treats current explanatory failure as provisional, it leaves the prior on solvability undisturbed, and it leaves the investigative effort active. The Intractability Confabulation collapses the prior on solvability to zero. Once this collapse occurs, Content-Agnostic Processing (Axiom 8) ensures that every subsequent failure to explain is processed as further evidence for intractability, because the framework through which the failure is evaluated is now the prior that intractability holds. The Confabulation is therefore self-sustaining in the same manner as the Individual Confabulation Cascade developed in Chapter 13, with the additional property that the self-sustainment is invisible, because the Individual experiences the accumulating failures as confirmation of a claim they already accepted rather than as evidence they installed the claim too early.

The Intractability Confabulation is, in the end, falsifiable in the ordinary sense. Its empirical content is the prediction that no unifying general theory of human motivation can be constructed. The preceding chapters set out to construct one, from the nine axioms of Chapter 4, whose interaction across scales is derived to yield the behavioural, institutional, and civilisational phenomena. If that framework and the system it describes correctly map onto the human subject, then the prediction that humans are too complex to model is false, and the Intractability Confabulation is not merely structurally suspect but, on its central subject, empirically false.

MOVEMENT B: PROJECTION

11.7 Projection Through Template Availability

Confabulation operates at the individual level, explaining one’s own desires. Projection extends this operation outward, locating the cause of an internally generated Salience in an external Object, Individual, or agent-category. Projection is not a separate mechanism. It is Confabulated I-resolution running on the same architecture as the desire case (§11.2), with the Confabulated ground attributing the Salience to a property of the perceived Object rather than to M's continuous comparative-position tracking.

The Coupled-Loop Mechanism

Projection is the outward attribution of a Confabulated I-ground (an A-deficit or I-opportunity) onto an external Object, because the actual source of the Salience (M's continuous tracking of the Object's comparative position within the Individual's Reference Set) is not available to I's Hypothesis Space. The mechanism is a single iteration of the AIM Operational Loop (Chapter 2; Chapter 4, Section 4.4.9) operating with the perceived Object, often another Individual, as the salience-bearing target:

M flags the Object. M, the continuous conditional Object-assessment channel (Chapter 2, Mimetic Desire; Chapter 4, Section 4.4.7), assigns Salience to the perceived Object on the basis of its current conditional priors. Where the Object is another Individual, the Salience tracks the Object's position in the comparative-position field that M is continuously assessing (Individual-as-Object Principle, §8.10). The flag need not originate in any such comparative-position content at all: another Individual's directional valence toward the Object installs the flag directly, under the gate clause of the Minimum Installation Condition (Chapter 4, Section 4.4.7.1), whatever position the Object holds in the comparative-position field. Steps 2 through 4 of this loop run identically regardless of which route produced the flag, which is why the two are indistinguishable to the Individual once Confabulated.
The Decision Hub integrates and tag-strips. The Salience flag enters the Decision Hub's Common-Currency Integration alongside A-signals and I-signals, is summed into a single Scalar, and the source-tags are destroyed during integration (Chapter 2, Decision Hub and Source Opacity; Chapter 4, Section 4.4.0).
I performs Bayesian ground-resolution over its Hypothesis Space. The tag-stripped integrated Scalar arrives at I, and I infers the ground-attribution from the candidates in its Hypothesis Space (Chapter 2, Confabulation; Chapter 4, Section 4.4.3). Where I's Hypothesis Space about M is well-populated and the M-Hypothesis has not been outweighed by an A-ground or I-ground attribution that the Individual and their reference group repeatedly restate, the M-tracking ground is recognised and the resolution returns the correct attribution. Where I's Hypothesis Space about M is impoverished, absent, or suppressed, or where the M-Hypothesis is present but outweighed by that repeatedly restated attribution, the Confabulation Rationality Lemma (§11.2) applies and the resolution returns an A-ground or I-ground instead.
I returns a Confabulated I- or A-ground located in the external Object. The most probable available candidate locates the cause of the Salience in a property of the Object itself, the Object is threatening, the Object is contaminating, the Object is admirable, the Object is good. The Confabulation is returned with the confidence appropriate to a well-functioning Bayesian update and is incorporated into the Self-Model and into action selection.

The mechanism is content-agnostic. The Confabulated ground may be valenced positively (idealisation, admiration, deference) or negatively (vilification, contempt, blame) depending on the polarity of the Salience M assigns and on the local availability of valenced Templates. Across the chapter's catalogue of Projection instances, the same four-clause loop is operating with different Object classes and different valences:

Scapegoat attribution. M tracks rivalry-Salience on a vulnerable out-group Individual or sub-group, the Decision Hub integrates and tag-strips, I returns a Confabulated I-ground locating the cause of collective deficit in that Individual or group, and the attribution is returned with the confidence that makes targeting feel like defensive necessity (developed at the collective scale in §11.12).
Out-group vilification. M tracks comparative-position Salience on members of a contrast category, I returns a Confabulated I-ground locating the cause of the Salience in a stable trait of the out-group ("they are dishonest," "they are dangerous"), and the attribution is generalised across the category: a claim of a stable kind-trait carved across a whole category, which the Diversity Referent (Chapter 14, §14.8.3.1.2) marks as held against an available decohered Referent, since the trait's distribution across that category is non-modal and admits no such kind.
Idealisation. M tracks positive comparative-position Salience on a high-prestige Model, I returns a Confabulated I-ground locating the cause of the admiration in objective excellence of the Model rather than in M's prestige-biased tracking. The "objective excellence" attribution is doubly confabulated: Status is aggregated, tag-stripped Salience (the Salience-Status Identity, Chapter 2), and Valence-Neutral Status entails that the integrated Scalar carries no preserved merit or valence dimension (only comparative Salience), so I is reconstructing an evaluative axis the Decision Hub destroyed, not reading a merit property the signal still contains.
Moral panic. M tracks alarm-Salience on a behaviour or population at the population scale through the Confabulation Cascade (Chapter 13), each Individual's I returns a Confabulated I-ground locating the alarm in objective threat, and convergent Confabulations across the population produce shared moral mobilisation against the perceived source.

The AIM Definition of Projection

Projection is Confabulated I-resolution applied outward, the generation of sincere but inaccurate explanations for other Individuals' (or other Objects') states, located in properties of the external Object rather than in the M-tracking that supplied the Salience. Projection deploys cognitively available Templates drawn from the projecting Individual's Self-Model, cultural repertoire, and activated categories, with the critical constraint that Templates the Individual lacks cannot be projected regardless of their accuracy, ensuring that M-driven behaviour in others is systematically misattributed to A or I sources because the M-Template ("they want this because their M-channel tracked it from Models") is absent from nearly every Individual's conceptual vocabulary, exactly as Type I Exclusion operates on I's Hypothesis Space at the individual level.

Rejecting the Freudian Account

The classical psychoanalytic account frames Projection as a defence mechanism: the Individual represses uncomfortable Truths about themselves (unacceptable desires, moral failings, forbidden impulses), and these repressed contents are then “projected” onto others, allowing the Individual to condemn in others what they cannot acknowledge in themselves. This account requires:

Repression: the Individual must actively (though unconsciously) hide Truth from themselves.
Displacement: the repressed content must somehow “move” from self to other.
Motivated distortion: the Individual must be protecting themselves from uncomfortable self-knowledge.

AIM rejects this account because it invokes mechanisms that Source Opacity renders unnecessary. The Individual does not need to repress Mimetic desires; they are already unavailable to introspection due to preconscious transmission and Source Opacity. There is no hidden Truth being defended against; there is only an Information Gap that makes the Truth architecturally inaccessible to introspection.

The Template-Availability Account

Projection occurs not through repression-and-displacement but through Template deployment.

A Template is a cognitively available category for social inference: a schema that answers “Why did they do that?” When the Individual observes another’s behaviour and seeks to explain it, available Templates are retrieved and applied:

“They bought that car because they need reliable transportation” (A-Template)
“They chose that career because they find the work meaningful” (I-Template)
“They pursued that credential because they want to appear successful to others” (M-Template, but typically unavailable)

Only available Templates can be deployed. If the M-Template is absent from the Individual’s conceptual repertoire, it cannot be applied to others regardless of its accuracy. The Individual cannot attribute Mimetic motivation to others when they lack the concept of Mimetic motivation entirely.

This explains why Projection is systematic rather than idiosyncratic. Across populations, the M-Template is absent, so M-driven behaviour in others is systematically attributed to A or I sources, not because of motivated distortion but because of vocabulary constraint. The error is architectural, not defensive.

The Neural Basis of Template Deployment

The same neural systems that generate Self-Models generate other-models. Theory of mind, the capacity to attribute mental states to others, operates through simulation: the brain models others by running its own cognitive architecture with modified inputs. “What would I be thinking if I were in their situation?”

Neuroimaging research confirms substantial overlap between brain regions active during self-reflection and those active during mentalising about others. The medial prefrontal cortex, temporoparietal junction, and precuneus are engaged in both tasks. This overlap is not coincidental; it reflects the computational efficiency of using one’s own cognitive architecture as the default model for understanding others.

The implication for Projection: the Templates available for self-explanation are automatically available for other-explanation. If an Individual explains their own desires using only A and I categories, they will explain others’ desires using only A and I categories. The Vocabulary Gap that produces Confabulation at the individual level produces Projection at the social level through the same architectural pathway.

Mirror neuron research provides additional grounding. Mirror neurons fire both when an Individual performs an action and when they observe another performing the same action. This system enables rapid inference about others’ intentions, but the inference is constrained by what the observing brain can represent. If the observing brain cannot represent M-driven motivation, mirror neuron activity cannot produce M-attribution for observed behaviour.

Why the M-Template Is Specifically Absent

The Vocabulary Gap established in §11.1 explained why M-vocabulary is absent at the individual level. The same factors ensure its absence as a Template for social inference:

No phenomenological basis: The Individual has never introspectively observed Mimetic acquisition in themselves, so they have no experiential template to project onto others.
No cultural transmission: Vernacular explanatory frameworks lack M-vocabulary, so the Individual has not acquired the category through social learning.
Evolutionary recency: The conditions that make M-recognition useful (complex modern environments, positional goods, Mimetic Premium extraction) are evolutionarily recent. Ancestral environments did not select for M-Templates because M-recognition provided no survival advantage.
Inferential opacity: Even when the Individual observes another pursuing a visible status marker (luxury car, prestigious address), they cannot infer M-motivation because the behaviour is equally consistent with A-explanation (“they need reliable transportation”) or I-explanation (“they appreciate engineering excellence”). Without the M-Template, the M-consistent interpretation is not among the available completions.

The Systematicity of Projection

The Template-availability account predicts that Projection will be:

Universal: Occurring in all Individuals, not only those with specific psychological profiles
Automatic: Generated without deliberate effort or awareness
Systematic: Producing the same misattributions across populations due to shared Template constraints
Resistant to correction: Persisting even when the Individual is informed about their own Mimetic motivation, because the habit of deploying A-I Templates is deeply ingrained

These predictions match the empirical evidence. Social psychology research consistently finds that people attribute their own and others’ behaviour to stable dispositions (the fundamental attribution error) rather than situational factors, and that they do so automatically, universally, and with high resistance to correction. AIM extends this finding: the “dispositions” attributed are systematically biased toward A and I sources because the M-Template is unavailable.

11.8 Template Inversion: Explaining Different Others

Projection through Template availability explains why the Individual attributes M-driven behaviour to A or I sources in similar others: people whose Self-Models share Templates with the Individual’s own. But what about different others: people the Individual perceives as unlike themselves?

Contrast Categories

The Individual’s Self-Model generates not only positive Templates (“I am someone who values X”) but also contrast categories (“I am someone who does not value Y”). These contrast categories function as inverse Templates for explaining different others.

Example: An Individual whose Self-Model emphasises intellectual achievement may construct a contrast category for those they perceive as status-driven (“I pursue knowledge for its own sake; they pursue credentials for social climbing”). This contrast category attributes M-motivation to the other, but in a distorted form. The attribution is:

Stigmatised: M-motivation is framed as moral failing, not architectural necessity
Externalised: Located in the other, not recognised in oneself
Essentialised: Treated as stable character trait rather than situational response

The contrast category appears to recognise Mimetic motivation. But it is not the M-Template of AIM: a neutral, architectural category applicable to oneself and others equally. It is a stigmatised other-category that maintains the Individual’s Self-Model as non-Mimetic while attributing Mimetic (or quasi-Mimetic) motivation to those perceived as different.

The Double Standard

This creates a double standard in social inference:

Similar others: Behaviour attributed to A or I sources (positive Templates projected)
Different others: Behaviour attributed to stigmatised quasi-M sources (contrast categories projected)

The double standard is invisible to the Individual because both attributions use available Templates. The Individual is not consciously applying different standards; they are deploying whatever Template is cognitively available for the person in question.

The implication: even when an Individual appears to recognise M-motivation in others, they typically do so only for out-group members, in stigmatised form, without applying the same category to themselves or their in-group. This is not M-recognition in the AIM sense; it is prejudice using M-adjacent vocabulary.

True M-recognition requires:

Self-application: Recognising that one’s own desires may be Mimetically acquired
Universality: Applying the M-Template to all Individuals, not only stigmatised others
Architectural framing: Understanding M as cognitive architecture, not moral failing
Diagnostic use: Using M-recognition to guide investigation, not to condemn

The contrast-category version of “they are status-driven” satisfies none of these criteria. It is Projection in moralistic disguise, not Source-recognition.

11.8.1 The Self-Model Moral Boundary Principle

The preceding analysis established three results: Self-Model Inversion generates contrast categories for every positive self-attribution; these contrast categories are stigmatised, externalised, and essentialised; and moral vocabulary attaches to them, transforming descriptive labels into permission infrastructure for targeting.

The Self-Model is not self-generated. It is substantially constituted through M-contaminated comparison. In any relationship of internal mediation, the rival is architecturally embedded in the Subject's Self-Model as the defining reference point against which self-valuation is computed. The Individual is what the rival is not, even as (and precisely because) the rivalry produces convergence that makes the two increasingly similar. The insistence on difference is itself Confabulation, generated because “I am identical to my rival” does not achieve cognitive closure.

The chain from these results:

The Self-Model includes M-grounded elements constitutively dependent on comparison with rivals.
Self-Model Inversion generates contrast categories for every positive self-attribution.
These contrast categories are stigmatised, externalised, and essentialised, treating the inverse trait as stable character rather than situational response.
Moral vocabulary attaches to these contrast categories, transforming them from descriptive labels into permission infrastructure for targeting.
Therefore, “evil” is not an independent ethical discovery but an architecturally predictable output of Self-Model maintenance under rivalry conditions.

This is the Self-Model Moral Boundary Principle: moral labelling of others is architecturally a function of Self-Model Inversion operating on rival-constituted identity, making the content of “evil” definitionally relative to the labelling Individual's Self-Model.

Why evil is Individual-relative. Because the Self-Model is path-dependent (each Individual's Belief Register is the unique product of non-commutative information encounter), and because the rival constituting the Self-Model differs across Individuals, the contrast categories generated by Self-Model Inversion differ across Individuals. The scholar whose Self-Model centres on “I pursue knowledge for its own sake” generates the contrast category “status-seeking careerist” as evil. The entrepreneur whose Self-Model centres on “I build things that matter” generates “parasitic bureaucrat” as evil. The ascetic whose Self-Model centres on “I reject material excess” generates “greedy consumer” as evil. Each is deploying the same architectural operation on a different Self-Model substrate, producing a different moral taxonomy.

The double standard (similar others have behaviour attributed to A or I sources, while different others have behaviour attributed to stigmatised quasi-M sources) is therefore not a bias layered on top of moral reasoning. It is moral reasoning, at least as typically practised.

Why “evil” changes across epochs. The principle's Individual-level relativity scales to populations through Entanglement. Because Entanglement produces convergent Self-Models within reference groups (shared M-weighted valuations stabilised through prestige-biased transmission), populations converge not only on shared Self-Models but on shared contrast categories. When a population shares the Self-Model “we are hardworking, self-reliant, and honest,” it generates a shared contrast category “they are lazy, dependent, and dishonest.” This shared contrast category is then available as a ready-made Scapegoat Template. The Confabulation Cascade does not need to construct the accusation from scratch. The Self-Model has already generated it.

This explains a pattern that conventional moral philosophy has never accounted for: the content of “evil” undergoes wholesale revision across historical epochs, while the structural operation producing the label remains invariant.

Medieval Christendom's Self-Model (pious, faithful, obedient to divine order) generated the contrast category “heretic, blasphemer, poisoner of wells,” and applied it to Jews, who satisfied Prey-detection criteria (a minority, visible, reciprocity-exempt) while being framed through Predator-attribution (agents of cosmic evil, poisoners, ritual murderers). When plague struck fourteenth-century Europe, populations required causal explanation for incomprehensible mass death. The true causes (bacterial infection, trade-route transmission, urban sanitation failures) exceeded computational tractability. The Agent-Based Causal Default completed the causal chain with the Scapegoat Template already in place: Jews had poisoned the wells. The Belief was sincere. Participants were not cynically deploying a convenient narrative while privately acknowledging bacterial aetiology (which was unknown). They genuinely believed the causal account. The violence was experienced as a defensive public health measure, eliminating the source of contamination.

Hutu Power's Self-Model (hardworking cultivators, the true people of Rwanda, democratic majority) generated the contrast category “cunning, parasitic, secretly dominant,” and applied it to Tutsis. The radio propaganda of RTLM did not invent the accusation. It activated a contrast category the reference group's Self-Model had already produced. The speed of the genocide (approximately 200,000 perpetrators mobilised in 100 days) is intelligible only if the Scapegoat Template pre-existed the crisis, requiring activation rather than construction. Documented testimony from perpetrators shows genuine conviction that Tutsis posed existential threat, that the killings were defensive necessity, that failure to participate endangered Hutu survival (Des Forges, 1999; Straus, 2006). The phenomenology was not “I know this is wrong but I must comply.” The phenomenology was “We must defend ourselves against those who would destroy us.”

Post-WWI Germany's Self-Model (a great nation, undefeated in the field, betrayed from within) generated the contrast category “treacherous internal enemies, war profiteers, those who stabbed us in the back.” The Dolchstosslegende was not merely propaganda deployed by cynical nationalists (though it was that too). It was the architecturally predictable output of a Self-Model that could not accommodate systemic defeat without Confabulating agent-based causation, using contrast categories the Self-Model had already generated. The true causes of defeat (strategic overextension, resource exhaustion, American entry, civilian deprivation) were systemic, diffuse, and resistant to agent-based framing. They did not satisfy the cognitive demand for explanation. The Confabulation provided closure: Germany had been stabbed in the back by identifiable agents who could bear causal responsibility for incomprehensible national catastrophe.

In every case, the population experienced its moral taxonomy as discovery (“we have identified the evil among us”) rather than as production (“our Self-Model generated this label to maintain its own boundaries”). Source Opacity blocks introspective recognition that the moral label is M-generated, and Confabulation generates the narrative that the label tracks an objective property of the target rather than a structural property of the labeller.

Connection to the Scapegoat Mechanism. The Prey-selection criteria (vulnerability, visibility, reciprocity-exemption) identify who to target. The Self-Model's contrast category supplies the Predator-framing (they embody the evil we define ourselves against). The incoherence between selection and framing is guaranteed because the framing was never derived from observation of the target. It was derived from the Self-Model's need to maintain its own boundaries.

This is why the scapegoat is structurally helpless. Salience is computed entirely by the observing population under its own beliefs, over ambient configuration variables (appearance, group membership, bearing) the target did not choose and cannot alter; the target has no way to control the axes on which their Salience is received and calculated. What is Salient is the observer's own, not the observed's (the Salience-Emission Asymmetry and its refinement, Chapter 2), which is precisely why no conduct by the target can reliably dislodge a convergent hostile attribution once the population's beliefs have fixed the axes of judgment.

The Content-independence of the Scapegoat Mechanism is therefore not merely an empirical observation that different populations target different groups. It is an architectural entailment: the content of moral labelling is generated by Self-Model Inversion operating on path-dependent Self-Models, which means the content must vary across populations whose Self-Models are constituted against different rivals and different reference groups. The mechanism (Self-Model Inversion producing contrast categories, moral vocabulary attaching to those categories, Confabulation Cascade amplifying them to population scale, Scapegoat selection following Prey-detection criteria, Predator-framing drawing on the ready-made contrast category) is invariant. The content (who is “evil”) tracks the labelling population's Self-Model, not the target's behaviour.

11.9 Guilt-Priming and Elevated Template Activation

The Freudian account of Projection may be empirically wrong about mechanism (repression-and-displacement), but it captures a real phenomenon: people sometimes attribute to others precisely those characteristics they are most defensive about in themselves. AIM explains this phenomenon through guilt-priming and Template activation.

Guilt-Priming Mechanism

When an Individual experiences guilt, shame, or anxiety about a particular motivation or behaviour, the relevant Templates become chronically activated; readily available in working memory and more likely to be retrieved during social inference.

Example: An Individual who suspects (at some level) that their career ambition is Mimetically driven may experience guilt when this suspicion arises. This guilt activates the career-ambition Template, making it highly accessible. When this Individual then observes another’s career behaviour, the activated Template is retrieved rapidly: “They’re doing that for status reasons.”

The mechanism is not repression-and-displacement. The Individual is not hiding uncomfortable self-knowledge and projecting it onto others. Rather:

Guilt or anxiety activates a Template
The activated Template is more accessible in memory
The accessible Template is more likely to be deployed in social inference
The Individual attributes to others whatever Template is most accessible

This explains the Freudian observation (people attribute to others what they are defensive about in themselves) without requiring the Freudian mechanism (motivated repression).

Why Guilt-Priming Is Partial Rather Than Complete

Guilt-priming activates Templates, but it does not create them. The Individual can only project Templates they possess. If they lack the M-Template entirely, guilt about M-driven desires cannot produce M-projection; it produces projection of the available Template that best approximates the anxiety.

Example: An Individual feels uncomfortable urgency about acquiring a prestigious house but lacks the M-concept. Their guilt is diffuse; they cannot name what they are anxious about because they lack the category. When they observe a neighbour purchasing a similar house, they project an available Template: “They’re trying to keep up with the Joneses” (colloquial quasi-M) or “They must have come into money and are being irresponsible” (A-adjacent moral judgment).

The projection reflects the guilt without accurately naming its source. This is why guilt-priming produces Projection that feels meaningful to the projector but remains systematically inaccurate. The Individual is projecting their anxiety, not their insight.

The Emotional Fingerprint

Guilt-priming produces a characteristic emotional signature in Projection: the Individual’s attributions about others carry affective charge that exceeds what the observed behaviour warrants. If the Individual attributes career ambition to a colleague with unusual vehemence (“They’re obviously just climbing the ladder”), the intensity of the attribution may reveal more about the projector’s own anxiety than about the colleague’s actual motivation.

This emotional fingerprint provides a diagnostic opportunity. When an Individual notices that their attributions about others carry disproportionate emotional charge, this is evidence of guilt-priming, which in turn suggests that the attributed motivation may be present in the Individual themselves, even if they cannot consciously recognise it.

The diagnostic question: Why do I care so much about their motivation? If the answer involves moral indignation, contempt, or anxiety that exceeds what the situation warrants, guilt-priming is likely operative. The Individual’s strong reaction is not about the other; it is about themselves.

Projection does not remain private. When an Individual attributes M-motivation to others (whether through contrast categories or guilt-priming), they typically express these attributions socially: in conversation, gossip, and public commentary. These expressed attributions become evidence for other Individuals, who incorporate them into their own social inferences.

The pathway:

Individual A projects M-motivation onto Individual B (using stigmatised contrast category)
Individual A expresses this attribution to Individual C
Individual C, lacking independent evidence about B, treats A’s attribution as informative
Individual C updates their model of B to include the attributed M-motivation
Individual C expresses this updated model to others, propagating the attribution further

This is Confabulation contagion: inaccurate attributions propagate through social networks via the same prestige-bias and social-proof mechanisms that propagate accurate information. The attribution’s inaccuracy provides no barrier to transmission: in fact, emotionally charged attributions (which guilt-priming produces) may transmit more readily than neutral ones because they are more memorable and more socially engaging.

The collective consequence: populations converge on shared attributions about which groups are “really” M-driven, which are “authentically” I-driven, and which are “merely” A-driven. These shared attributions bear no necessary relationship to actual motivational profiles; they reflect the Templates available to the attributing population, the contrast categories that population has constructed, and the guilt-priming dynamics operative in influential members.

Movement C traces how this convergence produces the Scapegoat Mechanism.

MOVEMENT C: COLLECTIVE PATHWAY

11.10 From Individual Confabulation to Observable Expression

Movements A and B established individual-level dynamics: Confabulation as the output Ground-Resolution returns on closing an Information Gap, Projection as the same operation closing on others through Template deployment. Movement C traces the collective pathway; how individual Confabulations aggregate and propagate to produce population-scale convergence on shared (and systematically inaccurate) causal narratives.

The Observability Threshold

Individual Confabulations remain private until they cross the observability threshold: the point at which the Individual expresses their Confabulation in a form accessible to others. This expression can be:

Verbal: Stating the Confabulation in conversation, public speech, or writing
Behavioural: Acting on the Confabulation in ways that others can observe and interpret
Artefactual: Creating objects, institutions, or systems that embed the Confabulation

Once expressed, the Confabulation enters the social environment as evidence that other Individuals can observe. These observers do not know that the expression is Confabulation: it appears as confident testimony about motivation, which carries evidential weight in social inference.

Why Expression Is the Default

Confabulations are not typically held privately. Several factors push toward expression:

Social coordination demands: Others ask “Why did you do that?” and expect an answer. Failing to provide one invites negative inference (evasion, incompetence, untrustworthiness). The Confabulation provides a socially acceptable response.
Self-Model maintenance: Expressing the Confabulation reinforces its incorporation into the Self-Model. Each time the Individual tells the story of their motivation, the story becomes more deeply embedded in their self-understanding.
Coalition-building: Shared narratives create social bonds. When the Individual expresses a Confabulation that others recognise (“I want the best for my family” as explanation for status-seeking), they establish common ground that facilitates cooperation.
Status-signalling: Certain Confabulations carry prestige. Expressing I-consistent motivation (“I’m passionate about my work”) signals virtue; expressing A-consistent motivation (“I need security for my children”) signals responsible parenthood. The Confabulation serves social functions beyond self-explanation.

The result: Confabulations are routinely expressed, not held privately. The social environment is saturated with confident testimony about motivation: virtually all of which is Confabulated for M-driven desires because the M-Template is absent.

Debate as Institutional Confabulation-Laundering

The debate format illustrates a specific form of institutional Confabulation-laundering. When an M-led Individual expresses a provocative position within a debate, the audience confabulates I-motivation for the speaker because the debate format is culturally coded as I-activity (truth-seeking). The format itself functions as a Confabulation-generation device at the audience level; it supplies the I-consistent Template ("they are arguing this because they believe it to be true") that prevents M-recognition. This is the Camouflage Architecture operating through institutional design: the debate format is Layer 4 (Disciplinary Prior: "adversarial testing finds Truth") providing camouflage for the speaker's Source Opacity (Layer 1): the speaker cannot read the motivation from the inside, and the audience, lacking the M-Template, does not infer it from the outside. See Debate Exploitation Principle, Chapter 2.

11.11 Propagation Through Prestige-Bias and Social Proof

Once Confabulations are observable, they propagate through the same transmission mechanisms that propagate any socially available information: prestige-bias and social proof. Those transmission mechanisms are the Propagation Pathways: prestige-biased learning, social proof, and mimetic template transmission, through which individually generated Confabulations achieve collective uptake.

Prestige-Biased Transmission

Humans preferentially attend to, remember, and transmit information from high-prestige sources. This bias is adaptive: in ancestral environments, high-prestige individuals typically possessed useful knowledge or skills, so attending to them was informationally efficient.

In the context of Confabulation propagation:

High-prestige Individuals express their Confabulations
These Confabulations receive disproportionate attention due to prestige-bias
Observers incorporate the Confabulations into their own understanding (both of the prestigious Individual’s motivation and of motivation generally)
The Confabulations propagate to others who defer to the observers

Example: A successful entrepreneur explains their motivation as “building something meaningful” (I-consistent Confabulation for what may include substantial M-components). This explanation receives wide attention because the entrepreneur is prestigious. Observers incorporate the explanation into their understanding of entrepreneurial motivation. When asked about their own entrepreneurial ambitions, they deploy the same Template: “I want to build something meaningful.”

The prestige-biased transmission does not track accuracy: it tracks prestige. A Confabulation from a prestigious source propagates regardless of whether it accurately describes the prestigious Individual’s motivation. The transmission mechanism is indifferent to Truth.

Beyond prestige-bias, humans use social proof, the frequency of a Belief or behaviour in the population, as evidence of its validity. If many people express a particular Confabulation, observers infer that the Confabulation is likely accurate. This inference is often valid: widespread Beliefs frequently (though not always) reflect reality.

But in the context of Confabulation, social proof creates a pathological feedback loop:

Many Individuals express the same Confabulation (because they share the same Template constraints)
Observers note the frequency and infer validity
Observers adopt the Confabulation for themselves
The Confabulation becomes more frequent, providing stronger social proof
The cycle accelerates

This is the Confabulation Cascade: shared Template constraints produce shared Confabulations, which achieve social proof, which reinforces the Confabulations, which further constrains available Templates. The cascade converges toward population-wide consensus on narratives that were never accurate but that now appear obviously true because everyone believes them.

The Cascade Does Not Self-Correct

Standard epistemology assumes that false Beliefs eventually collide with reality and are corrected. This assumption fails for Confabulation Cascades because:

No phenomenological disconfirmation: Acting on M-driven desires does not reveal their Mimetic source. The house is purchased; the credential is obtained; the status marker is acquired. The behaviour “succeeds” in the sense of achieving its object, so the Confabulated motivation is not disconfirmed.
Social reinforcement: Others affirm the Confabulation. When the Individual says “I bought this house for my family’s security,” others nod approvingly. There is no social pressure to revise.
Absence of alternative Templates: Even if the Individual experienced vague dissatisfaction after acquisition (the house does not provide expected fulfilment), they lack the conceptual vocabulary to interpret this dissatisfaction as evidence of Mimetic origin. The dissatisfaction is attributed to other factors (“I’m stressed at work,” “the neighbourhood isn’t quite right”) rather than to Source misattribution.
Sunk-cost reinforcement: Having acted on the Confabulation, the Individual is motivated to maintain it. Revising the narrative requires acknowledging that the decision was not what it appeared: a psychologically costly admission. Confabulation persistence is the path of least resistance.

The Confabulation Cascade is therefore stable in the absence of external intervention. It will not self-correct through ordinary social epistemology. This is why the Knowledge Patch requires deliberate dissemination; accurate Source-recognition will not emerge spontaneously from social dynamics that actively reinforce Confabulation.

11.12 The Scapegoat Mechanism as Collective Completion

The Confabulation Cascade explains how populations converge on shared narratives about individual motivation. When populations experience shared felt deficit, the Confabulation Cascade completes as a Grievance Belief, the convergent identification of a Scapegoat as responsible for the collective condition. The remainder of §11.12 derives this completion as the coupled-loop's behaviour at the collective scale, then threads the chapter's existing empirical illustrations and historical examples into the four substeps the derivation specifies.

The Scapegoat Mechanism, A Coupled-Loop Derivation

The Scapegoat Mechanism is the collective-scale expression of the AIM Operational Loop (Chapter 2; Chapter 4, Section 4.4.9) operating under conditions of extreme M-Hypothesis Space impoverishment across a population of Individuals in internal mediation. It is the point at which the mechanism that produces individual Object Parasitism (Chapter 2, Object Parasitism; Chapter 4, Section 4.4.11), iterated across a reference group under rivalry-escalating conditions, generates a collective output that neither any single Individual nor any institutional structure has consciously designed. The derivation proceeds through four steps, each of which is an application of the coupled loop specified at Chapter 4, Section 4.4.9.

Step one, continuous M-assessment of Individuals as Objects

M operates continuously whenever the Individual is conscious, assessing the perceived Object field along the dimensions its conditional priors supply. Among the Objects M assesses are other Individuals, because by the Individual-as-Object Principle (Chapter 8, §8.10) every Individual is an Object in every other Individual's Decision Hub. Under internal mediation, the other Individuals in a reference group are assessed as Models whose valuations, pursuits, and comparative positions are continuously tracked. M does not decide which Objects matter, it assesses all Objects continuously, and it does so with greater Salience for Objects on which M's distinctive social-world input stream converges.

In a reference group of Individuals in internal mediation, each Individual is therefore a Model whose behaviour, pursuits, and valuations the others are continuously tracking. This is not a pathological state, it is the ordinary operation of M under conditions of sustained social proximity. The Individuals' mutual Model-tracking is what produces the coordination, cultural learning, and social cohesion the chapter has treated in its Positive Mimesis material, and it is what makes collective action possible.

Step two, iterated fast-path Confabulation Cascades

Within the reference group, each Individual's Decision Hub integrates their A-signals, M's Salience assessments, and I's ongoing resolutions into Common-Currency Scalars and tag-strips during integration (Chapter 2, Decision Hub and Source Opacity; Chapter 4, Section 4.4.0). Each Individual's I performs Bayesian ground-resolution on the integrated scalars under whatever Hypothesis Space it holds. Where I's Hypothesis Space about M is well-populated and the M-Hypothesis has not been outweighed by an A-ground or I-ground attribution the group repeatedly restates, M-Salience tracking the mutual comparative position of group members is recognised as such, and the ground-resolution proceeds accurately. Where I's Hypothesis Space about M is impoverished, the Confabulation Rationality Lemma (§11.2) applies, and I returns a Confabulated A-ground or I-ground. The Confabulated ground attributes the Salience to a property of the Object, typically another Individual, and the Decision Hub initiates acquisition, defensive, or approach behaviour on that ground.

The other Individuals in the reference group observe this behaviour. M's continuous conditional assessment integrates the observed behaviour into the social-world input stream for the next cycle. If the observed behaviour appears to value an Object, the Object carries higher Salience for the observers. If the observed behaviour appears to treat another Individual as threatening, the observers' M-assessment updates that Individual's Model-tracking data in the direction of threat-Salience. I's prior-maintenance function, running on the slow path, updates the observers' priors about M accordingly, and the next fast-path cycle runs with those updated priors.

This is the Confabulation Cascade at the interpersonal scale, treated formally in Chapter 13, Section 13.4. Each Individual's Confabulated ground re-enters the reference group as observed evidence, which updates the other Individuals' priors, which condition the next round of M-assessment and I-resolution. Under well-populated M-Hypothesis Spaces, the Confabulation Cascade is dampened by I's recognition that observed behaviour partially reflects the other Individual's M-tracking rather than genuine A-deficit or I-opportunity, and the loop stabilises on accurate ground-attributions. Under impoverished M-Hypothesis Spaces, the Confabulation Cascade is self-amplifying, because each Individual's Confabulated ground is observed by the others as genuine, and the others' Bayesian updating updates their priors accordingly.

Step three, convergent Confabulated attribution under collapsing differentiation

As the Confabulation Cascade iterates, each Individual's M-assessment detects increasing Salience on the comparative-position dimension, rivalry phenomenology intensifies (Rival-Constitution Principle, Chapter 8 §8.10.3), and differentiation between the Individuals' positions begins to collapse. The collapse is not a loss of actual differences, it is a compression of the differences that M's comparative-position templates can distinguish under the Confabulation Cascade's conditions. As differentiation collapses, the set of candidate Confabulated grounds available to each Individual's I narrows, because I's Bayesian updating operates over Hypotheses about why the Salience is as intense as it is, and the Hypothesis Space about M is by supposition impoverished.

Under these conditions, I's Bayesian updating in each Individual converges on the Confabulated attribution that locates the cause of the rivalry-Salience in an external agent. This is Projection (§11.7) operating at the collective scale. The attribution is Confabulated because the actual cause is the Confabulation Cascade itself, the iterated operation of M-assessment and I-resolution under impoverished priors, and the Confabulation Cascade is distributed across the reference group rather than locatable in any single Individual or external agent. But I has no Hypothesis for distributed Cascade-causation in an impoverished Hypothesis Space, so it returns the most probable available Hypothesis, which is agent-based causation located in a specific Individual or sub-group.

The convergence is convergent across Individuals. Each Individual's I runs the same Bayesian updating over a similar Hypothesis Space against similar observed evidence, and therefore each Individual's I returns a similar Confabulated attribution. The reference group's Individuals converge on the same external agent as the Confabulated cause. The convergence is not coordinated, it is the coincident output of similar Bayesian resolutions under similar conditions, which is why it operates without conscious collusion and why it is experienced by each Individual as the self-evident Truth of the situation rather than as a shared misattribution.

The convergent selection of the scapegoat therefore does not require any Individual to originate the selection. The selection is generated by the coupled loop operating at the collective scale, and it completes when the proportion of the reference group whose I-resolutions have converged on the same Confabulated target reaches a threshold sufficient for collective action. Girard's observation that scapegoat selection is convergent rather than arbitrary is, on this derivation, architecturally necessary rather than empirically contingent.

Step four, collective discharge and loop reset

The collective discharge of aggression against the scapegoat is a system-wide event at the level of the reference group. The discharge produces an immediate and substantial change in the comparative-position data M is tracking, because one Model has been removed from the Reference Set, the configuration of valuations across the remaining Individuals has shifted, and the observed behaviour of the group has undergone a phase transition from rivalrous to coordinated. M's continuous conditional assessment registers this phase transition as a reduction in comparative-position Salience, because the reference-group configuration that was generating the Salience has been disrupted.

I's prior-maintenance, running on the slow path, updates accordingly. The scapegoat, having been expelled or destroyed, is re-categorised in each Individual's priors about M as a Model for what-not-to-be, an anti-Model whose Salience is now negative rather than positive. The rivalry subsides because the M-assessment conditions that generated it have been substantively altered, not because the Confabulation Cascade has been corrected. I's Hypothesis Space about M has not been enriched, no new M-Hypothesis has been added, no Knowledge Patch has been applied. What has changed is the configuration of the Object field over which M's continuous assessment operates.

This accounts for the temporary character of the catharsis Girard observed, and it accounts for the centrality of ritual sacrifice in archaic religion, treated below in the historical-pattern subsections. The Confabulation Cascade's underlying conditions (the reference group's internal mediation, the population's impoverished M-Hypothesis Spaces, the continuous operation of M-assessment and I-resolution) remain in place. The loop has been reset by the disruption of the Object field, not by the correction of the ground-resolution failure that produced the Confabulation Cascade. M resumes its continuous assessment, I resumes its Bayesian resolution over the same impoverished priors, and the Confabulation Cascade begins accumulating again over time. The scapegoat is replaced, eventually, as the Confabulation Cascade's next iteration converges on a new target under the same architectural conditions.

The discharge is therefore not the destruction of the Salience the Confabulation Cascade had accumulated but its conversion: into collective action against the scapegoat, and into a reconfigured Reference Set whose new comparative-position data is what M registers as reduced Salience. This is the same conservation that holds at every other stage of the M-channel's operation (the Salience-Status Identity and the Status-Premium Duality, Chapter 2; Chapter 8 §8.12); the accumulated quantity is never annihilated, only transformed, here into discharge and a re-armed loop, which is why the Confabulation Cascade resumes rather than terminates.

The derivation's structural consequences

Three consequences follow. First, the Scapegoat Mechanism is not a failure of any single component in the AIM architecture. Each component is operating within its specification, A signals deficits, M assesses the Object field continuously, the Decision Hub integrates and tag-strips, I performs Bayesian resolution over its Hypothesis Space, and the slow path updates priors from observed behaviour. The mechanism is the architectural output of the coupled loop operating at the collective scale under the specific condition of population-wide M-Hypothesis Space impoverishment.

Second, the Mechanism is unconscious at both scales. No single Individual's I-resolution represents the Confabulation Cascade's distributed causation as a Hypothesis, because the Hypothesis is not in the impoverished Hypothesis Space. No collective deliberation occurs, because the convergence on the scapegoat is the coincident output of parallel Bayesian resolutions, not a coordinated decision. The Individuals experience the scapegoat's guilt as self-evident and the collective action as righteous, because that is what their I-resolutions return.

Third, the Mechanism is recurrent rather than eliminable in the absence of intervention. The loop's reset does not modify the conditions that produced the Confabulation Cascade, and the subsequent iterations regenerate it. The architecturally available intervention is the Knowledge Patch (Chapter 18), which enriches I's Hypothesis Space about M so that Bayesian updating can return the distributed-Cascade Hypothesis rather than converging on agent-based Confabulated attribution. Subcriticality Maintenance, treated in Chapter 18, is the ongoing practice of keeping I's M-Hypothesis Space populated at the population scale, so that Cascades are dampened before they accumulate to scapegoat-generating intensity.

The remainder of §11.12 elaborates this derivation. The Collective Information Gap subsection redescribes Step two's conditions in terms of the cognitive demand for explanation that drives the Confabulation Cascade's iteration. The Highest-Likelihood Completion subsection redescribes Step three's convergent attribution in selection-and-framing terms (the prey-selection criteria and predator-attribution that the Dual-Role Confabulation signature in §11.19 will diagnose). Historical Pattern Recognition and the Dreyfus case study illustrate Step three across periods. The PM Production Cost Asymmetry subsection identifies an additional gradient that compounds the loop's attractiveness for the Confabulation Cascade's most active participants.

The Collective Information Gap

Chapter 10 established that human causal inference is optimised for detecting agents, not systems. When an Individual asks “Why am I experiencing this deficit?” the brain searches for an agent who is blocking, taking, or threatening. Systemic causes (diffuse institutional processes, emergent market dynamics, preconscious Mimetic transmission) exceed the computational capacity of the causal-inference system and are therefore structurally unavailable as default explanatory completions (Chapter 4, Section 4.3.9).

This Agent-Based Causal Default operates at collective as well as individual scale. When a population experiences shared felt deficit (economic strain, status anxiety, resource scarcity), the collective question “Why are we experiencing this?” demands an answer. But the collective question faces the same Information Gap as the individual question:

True causes may be systemically diffuse (distributed across thousands of institutional nodes)
True causes may be temporally distributed (accumulated over decades, not immediate)
True causes may involve M-dynamics that exceed computational tractability (Entanglement, Mimetic transmission, Object Parasitism)

The population’s causal-inference systems do not spontaneously generate these causes as default explanatory completions. But the demand for explanation persists; downstream processes (threat response, collective action, political mobilisation) require a causal model to proceed.

The Scapegoat as Highest-Likelihood Completion

Given the Agent-Based Causal Default and the Information Gap, the Scapegoat emerges as the highest-likelihood completion:

Agent-attribution is required: The cognitive architecture demands an agent-based answer
Accurate attribution is not the default: True systemic causes are structurally unavailable as default explanatory completions, so their formation is rare rather than impossible (Chapter 4, Sections 4.3.9 and 4.3.10)
Confabulation fills the gap: I's Bayesian ground-resolution function returns an agent-based narrative using available Templates, the M-Hypothesis being absent from the candidate list
Social proof selects the target: The agent-attribution that achieves widest social proof becomes the consensus explanation

This completion is the acute, crisis-mode form of the Agentic Precedence Principle (Chapter 10). The Agentic Precedence Principle establishes that agent-attribution is the standing default for any opaque causal domain; the Scapegoat Mechanism is what the default produces when the opacity is sudden, the suffering is mass, and the Information Gap is therefore at its widest and most urgent.

The paradigm case is epidemic disease before its systemic account existed. A plague presented exactly the conditions the Agentic Precedence Principle specifies (a large, fast, lethal harm of unseen cause) and the architecture returned an agent: the wrath of a god (the propitiation limb) or the malice of a poisoner (the scapegoating limb). The medieval blaming of Jews as well-poisoners during the Black Death is the completion run to its violent conclusion: agent-attribution under a maximal Information Gap selecting a prey-criterion target and framing it through predator-attribution. Germ theory is the systemic account that, once available and trusted, closed that gap and removed the completion's occasion; where the gap is reopened, by a novel pathogen, or by the refusal of an available account, the completion returns, which is why epidemic scapegoating recurs. The disease case is not one illustration among many but the cleanest, because it is the domain in which the systemic account arrived decisively enough to show the Mechanism switching off as the Information Gap closed.

The Scapegoat is therefore not selected randomly. Selection follows predictable criteria established by the Four-Type Agent Taxonomy (Chapter 10):

Predators: Agents perceived as threatening, dangerous, taking
Collaborators: Agents perceived as helpful, reciprocating, giving
Rivals: Agents perceived as competing for the same resources
Prey: Agents perceived as vulnerable, extractable, reciprocity-exempt

The Scapegoat is selected through prey-detection criteria: vulnerability, visibility, and reciprocity-exemption. Vulnerable populations can be targeted without triggering overwhelming retaliation. Visible populations are cognitively available as attribution targets. Reciprocity-exempt populations (those outside the in-group’s exchange networks) can be targeted without disrupting valuable cooperative relationships.

But the Scapegoat is framed through predator-attribution: the target is described as threatening, taking, blocking, causing the deficit. This framing is required because targeting a vulnerable population as vulnerable would be experienced as aggression. The predator-frame transforms aggression into defence.

This is the Dual-Role Confabulation: the Scapegoat is selected as prey but framed as predator. The selection and framing operate through different cognitive systems and point in opposite directions: a structural incoherence that is diagnostic of Confabulation.

Historical Pattern Recognition

The Scapegoat Mechanism is historical pattern. Across cultures and centuries, populations under stress have converged on Grievance Beliefs that exhibit the prey-selection/predator-framing structure:

Medieval Europe: Jewish communities were selected as Scapegoats during plague, famine, and economic strain. Selection followed prey-criteria: Jews were a numerical minority, legally vulnerable, and outside Christian reciprocity networks. Framing followed predator-attribution: Jews were accused of poisoning wells, spreading plague, controlling finance, and conspiring against Christendom.

Early 20th Century Germany: Jewish populations were again selected following military defeat, economic collapse, and national humiliation. Selection criteria were identical: numerical minority, legally targetable, outside dominant ethnic networks. Framing was predator-attribution: international conspiracy, financial manipulation, cultural corruption, existential threat to the German nation.

Contemporary Patterns: Immigration debates in multiple countries exhibit the same structure. Immigrant populations are selected through prey-criteria: legal precarity, linguistic isolation, outside established reciprocity networks. Framing is predator-attribution: “taking our jobs,” “threatening our culture,” “draining our resources.”

In each case, the diagnostic question reveals the Dual-Role Confabulation: If they are as powerful and threatening as described, why can we target them without expecting overwhelming retaliation? Genuine predators cannot be safely persecuted by populations lacking overwhelming military advantage. The fact that persecution proceeds reveals that the target was selected as prey regardless of predator-framing.

The Dreyfus Affair: A Worked Case Study

The Dreyfus Affair (1894–1906) provides a detailed case study of the Scapegoat Mechanism in operation.

Context: France in the 1890s experienced collective felt deficit: military humiliation from the Franco-Prussian War (1870–71), political instability, economic anxiety, and nationalist fervour seeking outlet. The question “Why is France diminished?” demanded an agent-based answer.

Selection (Prey-Criteria): Captain Alfred Dreyfus was a Jewish officer in the French Army: a member of a religious minority with historical vulnerability, visible due to his military position, and perceived as outside the Catholic-nationalist reciprocity network that defined “true” French identity.

Framing (Predator-Attribution): Dreyfus was accused of espionage, selling military secrets to Germany. The accusation framed him as an active threat to national security, a traitor whose actions explained French vulnerability. The predator-frame transformed what was actually targeting of a vulnerable minority into defence of the nation against internal enemies.

Evidence of Confabulation: The evidence against Dreyfus was fabricated and contradictory. When evidence of his innocence emerged, the military establishment doubled down rather than revising the attribution. This is the Bayesian compounding predicted by AIM: each Confabulation becomes a prior for subsequent interpretation, and disconfirming evidence is assigned low likelihood because it conflicts with the entrenched narrative.

Dual-Role Signature: Dreyfus was simultaneously described as a powerful agent of international conspiracy (predator-framing) and targeted precisely because he was a politically vulnerable Jewish officer who could be convicted without the resources to effectively resist (prey-selection). The incoherence is diagnostic.

Resolution: The Affair was eventually resolved not through spontaneous self-correction but through deliberate intervention; the work of Émile Zola, Georges Clemenceau, and others who applied external pressure to force re-examination of evidence. This confirms AIM’s prediction: Confabulation Cascades do not self-correct; they require intervention that disrupts the social proof dynamics maintaining the false attribution.

The PM Production Cost Asymmetry Compounds Scapegoating’s Attractiveness

The PM Production Cost Asymmetry compounds the Scapegoat Mechanism’s attractiveness. Scapegoating is not merely a cognitive error produced by the Information Gap; it is also an M-rational strategy for the scapegoater. Identifying and targeting a scapegoat generates PM through hatred at near-zero cost: the accuser need demonstrate no competence, produce no achievement, and sustain no I-led engagement. They need only point and condemn. The PM yield is immediate: the audience’s hostile engagement with the accusation registers as social Salience for the accuser, and Valence-Neutral Status ensures this Salience is neurologically equivalent to admiration-derived Status. This is why demagogues gravitate toward scapegoat narratives: the narratives generate PM more cheaply than any alternative Status-pathway, while Confabulation ensures the demagogue sincerely believes they are performing moral duty rather than extracting PM.

11.13 The Humanisation Null Hypothesis and the Architectural Cost of Dehumanisation

The scapegoat derivation raises a question the literature on dehumanisation has not successfully answered. If ordinary moral consideration of other persons rests on recognising their humanity, why is the revocation of that recognition so cheap, so fast, and so reliably producible under propaganda conditions? The standard answers (moral disengagement mechanisms, dispositional factors, propaganda effects) describe the phenomenon without explaining its architectural cost, and they leave the rapid transition from moral consideration to moral exclusion as a puzzle that must be filled in by contingent psychological or social variables.

The Humanisation Null Hypothesis (HNH) resolves the puzzle by reversing its polarity. The transition is cheap because nothing substantive is being removed. There is no humanisation layer in the Decision Hub that propaganda must deconstruct, because the Decision Hub does not perform a humanisation operation in the first place. Other Individuals appear in the Decision Hub as Objects, valued through A-I-M decomposition (Individual-as-Object Principle), and Source Opacity (Axiom 3) ensures that even if some upstream circuit did flag an Object as human, that flag would not survive common-currency integration into the scalar subjective-value signal. What produces the phenomenology we retrospectively call moral consideration is the Agent-Type classification (Axiom 9), which assigns that Individual-Object a position in a four-dimensional vector whose basis vectors are Collaborator-weight (C), Rival-weight (R), Predator-weight (X), and Prey-weight (Y), each of which carries its own distinct reciprocity structure (C generates reciprocity-of-cooperation, R generates reciprocity-of-contest, X generates reciprocity-of-threat-response, and Y generates reciprocity-of-extraction), with the moral consideration applied to the Object being the weighted combination of the four basis profiles under the current vector coordinates. The vector formalism and its three properties are specified in the subsection immediately following.

Moral consideration is not a binary output of the classifier and not a discrete four-state output. It is the continuous weighted sum of four basis-profile contents under the classifier’s current vector coordinates for the Individual-Object in question. Collaborator-weighted content carries cooperation, shared undertaking, solidarity, and fairness. Rival-weighted content carries recognition as a worthy opponent, honour-in-contest, and the conventions governing fair combat, violations of which (such as betrayal or striking from concealment) are experienced as moral violations in their own right. Predator-weighted content licences defensive force under specific constraints (proportionality, the prohibition on gratuitous cruelty even toward attackers, restraint once the threat is neutralised) and activates the community’s protective obligations toward those the Object threatens. Prey-weighted content carries the extraction-ethics profile visible in hunting and pastoral traditions (the prohibition on wastefulness, the requirement of clean kills, stewardship obligations, ritual acknowledgements offered to the animal taken, and the distinction between legitimate extraction and gratuitous cruelty).

The actual moral consideration applied to an Individual-Object is the vector-weighted blend of these four contents, and because most real Objects carry non-zero weight on more than one axis, the phenomenology typically reflects a blend rather than the pure content of any single basis profile. Prey-dominant vectors carry moral content, as do the other three. What varies across the vector’s coordinate space is the weighted structure of the moral consideration produced, not its presence or absence.

Dehumanisation, on this account, is a redistribution of weight across the Agent-Type vector, not a reclassification between discrete categories and not a stripping of moral consideration. The out-group Individual-Object’s vector coordinates are updated under the incoming observable evidence (propaganda, conflict behaviour, institutional signalling, prior-group dynamics, and the physical and contextual cues the classifier reads directly from the Object’s presentation), and weight flows from C (or R) toward X or Y. The A-I-M decomposition updates continuously with the vector, because the moral profile applied to the Object at any moment is the weighted sum of the four basis-profile contents under the current vector coordinates. As X-weight rises, PA is increasingly dominated by threat-response content, PI inverts toward the Object as a source of danger rather than engagement, and PM inverts as the Object accrues negative status signal as a threat to the community. As Y-weight rises, PA is increasingly dominated by extraction, PI collapses, and PM collapses or inverts as the Object moves below or outside the status-comparison space. The phenomenology changes because the weighted sum changes, not because a discrete switch has been thrown.

The humanisation-first view’s puzzle was why moral consideration could be suspended so easily. The vector reformulation’s answer is that moral consideration is never suspended, it is continuously recomputed, and the weights along the four axes move under evidence at rates that threat-detection and resource-detection circuits were selected to make fast in the X- and Y-directions and slow in the C- and R-directions. The moral content produced by an X- or Y-dominant vector is incompatible with the moral content produced by a C- or R-dominant vector, but incompatibility between weighted sums at different coordinates is not the same thing as the absence of moral consideration at any of them.

The asymmetry between dehumanisation and rehumanisation follows directly from the vector formalism. Weight flows toward X and Y fast because threat-detection and resource-detection circuits were selected to err on the side of false positives, which is why propaganda can redistribute substantial weight along those axes in weeks. Weight flows toward C and R slowly because accumulating reciprocity-evidence and contest-evidence under Source Opacity requires sustained behavioural exposure, which non-behavioural observable cues cannot substitute for, with Confabulation actively generating alternative explanations for the observed reciprocity throughout the accumulation process. This is why reconciliation takes generations and why the architectural cost of post-conflict recovery is systematically underestimated by frameworks that presuppose a humanisation baseline could be restored. What must be rebuilt is not a humanisation layer but the observable-evidence environment that drives the vector’s weight back toward C- and R-dominance along the specific axis the post-conflict relationship requires, and the rebuilding must proceed at the architecturally slow rate the C- and R-axes impose.

11.13.1 The Four-Way Agent-Type Vector

The preceding paragraph lists Collaborator, Rival, Predator, and Prey as four categories, and the remainder of the HNH section is written in language that at times treats these as discrete assignments, an Individual-Object "is" a Collaborator, "is reclassified as" Prey, and so on. That language is a presentational convenience and should not be read as the architectural claim. The architectural claim is that the Agent-Type classifier does not output a single category label. It outputs a position in a four-dimensional vector space whose basis vectors are Collaborator-weight (C), Rival-weight (R), Predator-weight (X), and Prey-weight (Y), and whose coordinates are the instantaneous weights the classifier assigns to each basis vector for a given Individual-Object given the observable evidence currently available. Observable evidence in this sense includes both behavioural evidence (what the Object has been observed to do, cooperate, contest, threaten, yield) and non-behavioural observable evidence (physical attributes such as size, speed, armament, visible group affiliation, proximity, apparent reproductive or resource value, and other perceptible properties that the classifier's ancestral circuitry was selected to read directly from the Object's presentation). Behavioural and non-behavioural evidence contribute jointly to the vector's coordinates, with their relative weights governed by the classifier's conditional priors under the current context.

The vector formalism is required because any real Individual-Object carries qualities that situate it differently along each of the four dimensions simultaneously. A trading partner from another community may carry substantial C-weight through sustained exchange, non-trivial R-weight through competition over shared markets, residual X-weight through the history of prior conflict and any physical indicators of coercive capacity, and a small but non-zero Y-weight through asymmetric bargaining position. A colleague may be largely C-weighted with an R-component that tracks competition for the same promotion. A pet dog carries the Collaborator-with-parent-child subtype (the Collaborator subtypes are derived at Chapter 9, Section 9.16) at high weight with negligible weights on the other three dimensions. An enemy combatant may carry high X-weight with a non-trivial R-component inherited from the pre-conflict relationship and residual C-weight from shared institutions or alliances that have not yet fully decohered. Discrete-category language approximates these positions by reporting the dominant dimension, and for many expository purposes the approximation is adequate, but the underlying architectural quantity is the full four-vector.

Five properties of the vector follow from the coupled-loop architecture already developed in this chapter.

First, the weights are continuous and real-valued, not Boolean. An Individual-Object's C-weight can be 0.8, its R-weight 0.15, its X-weight 0.04, and its Y-weight 0.01, and the moral profile applied to that Object is the profile-weighted sum of the four reciprocity contents rather than the content of any single profile in isolation. The reciprocity-of-cooperation, reciprocity-of-contest, reciprocity-of-threat-response, and reciprocity-of-extraction profiles are the four basis-profile contents, and the actual moral consideration applied to any given Individual-Object is their weighted combination under the current vector coordinates.

Second, the weights are dynamic and evidence-responsive. Observable evidence accumulating through encounter and interaction (behavioural evidence from what the Object does, together with non-behavioural evidence from the Object's physical and contextual presentation) redistributes probability mass across the four dimensions, with the redistribution rates governed by the same Bayesian updating operation M and I perform over all other Object-properties. The asymmetric rates noted later in this section (redistribution toward X and Y is fast because threat-detection and resource-detection err on the side of false positives, redistribution toward C and R is slow because it requires sustained evidence accumulated under Source Opacity conditions) apply to the rates at which weight flows along the four axes of the vector, not to transitions between discrete states. Non-behavioural observable evidence is particularly consequential for the speed of X- and Y-axis movement, because physical threat cues and vulnerability cues can drive rapid weight redistribution on the basis of a single exposure, whereas C- and R-axis movement requires the slower accumulation of behavioural evidence through repeated interaction.

Third, the vector is computed Individual-Object by Individual-Object, and its coordinates reflect the Individual's local evidence history for that specific Object rather than any population-level stereotype. This is why selective reclassification within slave-holding societies is architecturally predicted rather than anomalous (the master's vector for a specific household servant can carry high C-weight while the population-level stereotype-driven vector for the enslaved population remains Y-dominant), why first-contact classifications vary across specific encounters within the same contact event, and why rehumanisation proceeds Individual-by-Individual rather than at the population level. The population-level "stereotype" is a time-averaged and evidence-averaged vector inherited from cultural transmission, which conditions the priors under which each specific-Object vector is computed but does not override the specific-Object evidence that accumulates through direct encounter.

Fourth, when the vector is used to decide how much can be taken from an Individual-Object, a misreading of the weights produces an error whose direction is predictable in advance (Appendix A, Section A.11.5). The calculation is simple. The Individual estimates what the counterpart is worth to extract from, and then subtracts the expected cost of extracting it, chiefly the risk that the counterpart will resist or retaliate (Appendix A, Section A.11.1). That resistance cost is carried by the counterpart's standing as an agent: a counterpart read as a genuine agent, one that can flee, refuse, retaliate, or coordinate with others, carries a real cost in the estimate, and a counterpart read as a mere Object, an inert resource that cannot answer back, carries almost none. The estimate is only as good as the reading it rests on.

Because it is the Object reading that discounts the cost of resistance, the sign of any misreading follows from how much agenthood the reading assigns. Reading a counterpart as less capable of resistance than it truly is, sliding it toward the Object pole, drops a cost that belongs in the estimate and makes the estimate too high. The Individual believes more can be safely taken than the situation allows, and acts on that belief until the resistance the estimate had discounted actually arrives. Reading a counterpart as more adversarial than it truly is makes the opposite error: the estimate comes out too low, and the Individual holds back from an engagement that was in fact safe. The mistake is not random noise. It has a sign, set by which way the reading is wrong, and its size grows with the size of the misreading.

The upward error is the more consequential, and it is a familiar one. A bully treats a victim as someone who will not resist, presses further and further, and is eventually surprised when the victim fights back, finds allies, or reports the conduct. A ruler treats the peasantry as an endless resource rather than as people who can organise, extracts more and more on the assumption that they will not resist, and provokes the revolt the assumption had ruled out. A child treats the family cat as a toy that cannot object, pulls its tail once too often, and is scratched. In each case the capacity to resist was real all along; the error was in reading the counterpart as though it were not there.

Fifth, the vector is computed in two limbs of the architecture's loop rather than at a single locus, and the second property's asymmetric rates and the third's conditioning both follow from where each limb sits. The Observer's limb assigns the Individual-Object a salience weight on what its presentation supplies, upstream of the fast loop (Appendix A, Section A.2.1), and operates inside the Preconscious Transmission window rather than at deliberative speed (Section 4.4.8). The Thinker's limb carries the same vector as an Agent-Type hypothesis in the Hypothesis Space over which the Bayesian update operates, entering as a prior rather than as a change to the update's mechanics (Appendix A, Section A.5.6). The limbs communicate through the Decision Hub in one direction and the Belief Register in the other.

The loop runs in both directions, and which direction is running distinguishes a first encounter from a later one. Where the Individual holds no hypothesis for the Individual-Object, the Observer's read is conditioned by the inherited population-level template alone, and the Thinker's resolution of the resulting Scalar commits the first specific-Object vector. Where a hypothesis is already held, it conditions the read before the read is made.

A coordinate that presentation can move is moved in the Observer's limb, at the shortest time constant in the architecture; a coordinate requiring behavioural evidence accumulated across encounters is moved in the Thinker's, at the longest (Appendix A, Section A.5.6). X and Y are the coordinates presentation supplies; C and R are not. The rate asymmetry is therefore an implementation property of the loop as well as a selection result.

Judgments of aggressiveness and related traits formed on a hundred-millisecond exposure to an unfamiliar face correlate with the judgments the same perceivers make with no time limit, and further exposure raises confidence without altering the judgment (Willis & Todorov, 2006). Dominance judgments from facial glimpses are above chance from forty milliseconds and gain no accuracy from longer viewing (Rule, Adams, & Ambady, 2012). A threat judgment is recoverable from parametrically varied facial cues alone, decomposing onto dimensions of valence and dominance (Oosterhof & Todorov, 2008).

The remainder of the HNH section should be read with this vector architecture in mind. Phrases such as "classified as Prey" should be understood as "Y-dominant in the vector for this Individual-Object under the evidence available to this classifier at this time", and phrases such as "reclassification from Prey to Predator" should be understood as "redistribution of weight from Y toward X along the vector's axes, driven by updated observable evidence". Where the discrete-category language is retained in what follows, it is retained for readability and should be interpreted against the vector formalism specified here.

11.14 First-Contact Cases as Diagnostic Evidence for HNH

Empirical evidence for the Humanisation Null Hypothesis comes from first-contact encounters between human populations whose Agent-Type classifiers have no prior Collaborator or Rival assignment to apply. The first-contact scenario is diagnostically powerful because it strips away every candidate humanisation mechanism the standard dehumanisation literature relies on. There is no propaganda campaign, no sustained out-group vilification, no generations of cultural conditioning, no prior conflict history. There are simply two populations whose classifiers must assign a category on limited observational evidence.

If humanisation were a positive cognitive operation, first contact should produce provisional humanisation followed by either its maintenance or its revocation. The empirical record shows nothing of the sort. It shows immediate Agent-Type assignment on the basis of behavioural cues, technological capacity, and resource-competition signals, with the profile-appropriate moral consideration flowing downstream from the category assignment rather than from any species-recognition event. That is precisely the architectural sequence HNH predicts.

The Australian case is the paradigm example. The British colonists arriving in 1788 did not perform a humanisation operation on Aboriginal peoples and then strip it away. Their Agent-Type classifier processed the encountered population through the same circuits that processed the continent’s fauna, and the assignment it produced (Prey, in the resource-extraction sense, with occasional reclassification as Predator when resistance was mounted) was a direct output of the species-general machinery running on the available evidence. The terra nullius doctrine was not a legal fiction imposed over a prior moral recognition. It was the legal ratification of the Y-dominant vector the colonists’ architecture had already performed, and the paternalistic colonial frameworks that emerged over the 19th century (protectionism, missionary-led welfare, stewardship-adjacent administration) were the extraction-ethics moral profile applied to a human population Y-dominant in the Agent-Type vector (Section 11.13.1 for the formalism). They were not humanisation-lite. They were the profile-appropriate moral content being confabulated onto the classification the architecture had produced.

When Aboriginal people defended their territory, they were reclassified from Prey to Predator (visible in the frontier wars rhetoric about savages and raids), and the reclassification switched the moral profile from extraction-ethics to threat-response, producing the characteristic Predator-targeting response: elimination campaigns, punitive expeditions, and the Black War in Tasmania. The redistribution of vector weight from Y toward X is cheap because both profiles lie outside the Collaborator-Rival reciprocity structures where the cooperation and contest profiles architecturally reside. No humanisation layer was ever established and therefore none had to be removed between the two reclassifications. What changed was which of the four moral profiles was being applied, and the reciprocity content of each profile was always present in the architectural output.

Anthropology has long puzzled over the dramatic variation in the content of first-contact explanations. Ghosts, returning ancestors, gods, demons, punishment-sending spirits, beings from across the sea, the souls of the dead: the content varies so much across cultures that the literature has often treated cultural symbolic systems as the independent variable driving the explanatory diversity. This framing mislocates the explanandum. What varies across first-contact cases is not the architectural process but the self-model materials on which the architectural process draws. The process itself is invariant and species-general. Under causal ambiguity, the Agent-Based Causal Default (Axiom 9) classifies the encountered entity as an Agent of some kind. Source Opacity (Axiom 3) prevents the Individual from observing this classification directly. Confabulation (Axiom 4) generates a sincere post-hoc narrative explaining the encounter, and that narrative must be drawn from the categories the self-model already contains. When no prior Collaborator or Rival classification is available (which is the definitional condition of first contact), the only categories the self-model can supply are the ones culture has previously installed for interpreting encountered Agents in general. The content of the confabulation therefore varies with the cultural stock of Agent-Type categories, while the structural position of the confabulation within the encounter remains architecturally constant.

The categories selected are not random. They are the categories that come with behavioural affordances already specified (offer gifts, flee, fight, worship, avoid), because the classifier’s output must tell the Individual what to do next. Ghosts, ancestors, and gods are selected so often across first-contact cases not because cultures happen to share those categories but because those categories come pre-equipped with reciprocity rules the Individual can act on immediately. The anthropological red herring has been treating the cultural content as the causal driver when it is in fact the substrate on which a species-general architectural process runs.

The pattern is not culturally specific. Spanish contact with the peoples of the Americas from 1492 onwards followed the same architectural sequence, and the Valladolid controversy of 1550 is best read not as a debate over humanisation but as a debate over whether the encountered populations should be C-dominant in the Agent-Type vector or Rivals (with consequent obligations of reciprocity) or as Prey (legitimate targets of resource extraction and enslavement). The theological language of whether the Indians had souls was the culturally available vocabulary for resolving the classifier’s ambiguity, not a description of a humanisation operation being performed or withheld. Roman jurisprudence distinguished populus (a people with whom treaties and obligations could exist, meaning Collaborators or Rivals in AIM terms) from the barbarian category (outside the reciprocity space, processed as Predator or Prey depending on behaviour), and the mapping onto the Agent-Type taxonomy is direct. Japanese encounters with the Ainu, Chinese encounters with steppe peoples, Bantu expansions into southern Africa, and Polynesian settlement displacing earlier island populations all exhibit the same structural pattern because the architecture generating the pattern is species-general.

The reciprocal symmetry of first-contact encounters tells against the humanisation-first view. Aboriginal peoples encountering the British did not perform a humanisation operation on the colonists either. The initial assignments varied (ghosts of the dead, returning ancestors, Predators, occasional Collaborators through trade) and reclassification proceeded on the same architectural timetable. Both populations were running the same species-general classifier, and both produced their moral consideration or its absence as a downstream consequence of Agent-Type assignment. If humanisation were a cognitive operation, one would expect cultural asymmetries in its performance. The empirical record shows architectural symmetry instead, which is what HNH and SNPP jointly predict.

11.15 Cross-Species Reclassification and the Positive Polarity: Pets as Family Members

The first-contact case demonstrates HNH in the negative polarity, with reclassification from initial ambiguity toward Predator or Prey categories. The phenomenon of companion animals being treated as family members, and in some cases as children, demonstrates HNH in the positive polarity and provides a second diagnostic test for the species-general character of the Agent-Type classifier. The owner who describes a dog as a fur baby is not engaging in sentimental exaggeration. They are producing an accurate phenomenological report of where the Agent-Type classifier has landed for that Individual-Object under sustained behavioural evidence of reciprocity, dependency, and shared environment.

If moral consideration were contingent on humanisation, the fur baby phenomenon should be architecturally difficult, requiring some cognitive override that grants honorary human status to a non-human animal. The empirical record shows nothing of the sort. It shows effortless and spontaneous classification of companion animals as Collaborators with the parent-child subtype (Chapter 9, Section 9.16) that produces parental care phenomenology, with the A-I-M decomposition updating accordingly: PI becomes high (the pet is a genuine source of intrinsic engagement and care), PM attaches (pets carry status signal in human social networks), and PA becomes reciprocal (the pet is processed as a dependent whose bodily needs the owner is responsible for meeting). The architecture performs the classification because the architecture cannot tell that the Object it is processing is not human. It never performed a humanness check in the first place, which is exactly what SNPP predicts.

This also explains cross-cultural variation in pet bonding that has otherwise been difficult to account for. The same species can be food in one context and family in another, not because the species has changed but because the behavioural evidence available to the classifier has changed. Dogs in rural contexts where they function as working animals or food sources, dogs in suburban households where they function as companion animals, and dogs on commercial farms where they function as commodities are the same species processed through the same species-general classifier, but each context supplies different behavioural evidence and produces different Agent-Type classifications with different moral profiles. The Collaborator-with-parent-child-subtype classification applied to a suburban companion animal generates a protective-care-reciprocity profile that differs architecturally from the Collaborator-among-equals profile applied to adult human Collaborators, which is why the moral duties owed to pets feel phenomenologically similar to the duties owed to children and different from the duties owed to adult friends or colleagues. The Y-dominant vector applied to farmed dogs or working dogs generates the extraction-ethics profile with its stewardship-adjacent moral content (Section 11.13.1 for the formalism). The classification is not a judgement about the species. It is a judgement about the Individual-Object’s observed relational position within the Individual’s social field, and the moral content attaches to the classification.

The positive-polarity case reinforces the negative-polarity case because both proceed through the same architecture in opposite directions. Humans are reY-dominant in the Agent-Type vector in first-contact cases; non-human animals are reC-dominant in the Agent-Type vector with parent-child phenomenology in pet cases. Neither transition requires a humanisation operation, because neither transition encounters one. The symmetry is exactly what HNH and SNPP jointly predict, and the two principles accommodate cases from genocide to fur babies, structurally.

11.16 Cross-Species Sexual Behaviour and Object-Level Affordances

A further consequence of the species-general character of the Agent-Type classifier, and of the more general Object-valuation circuitry of the Decision Hub, is that the sexual-behaviour affordance is architecturally available under more than one Agent-Type classification and under Object-level conditions that do not involve Agent-Type classification at all. The affordance is triggered by A-I-M profile match, not by the Agent-Type label as such. Agent-Type classification matters because it reliably produces the right profiles in Objects that meet its conditions, but it is one route to the affordance rather than its sole architectural source. This observation accommodates three phenomena that conventional framings have treated as disconnected or anomalous: the archaeological record of human interbreeding with closely related hominin populations, the recurrent appearance of cross-species sexual behaviour across human societies, and the use of artefacts deliberately engineered to trigger the sexual-response system without any Agent-Type classification being involved.

Under Collaborator classification, cross-species sexual behaviour arises where sustained contact produces sufficient behavioural evidence to place the non-human Individual-Object into the Collaborator profile with the mating-pair subtype (Chapter 9, Section 9.16) activated. The behaviour carries the cooperation-reciprocity profile. Under Y-dominant vector, cross-species sexual behaviour arises where the non-human Individual-Object is processed as a resource and the affordance expressed is the reproductive or bodily-use extractive affordance that the Prey profile makes available (Section 11.13.1 for the formalism). The behaviour carries the extraction-ethics profile. Neither path requires any humanisation operation, because no such operation is architecturally available. Both paths are expected outputs of the classifier running its normal operation on the available behavioural evidence and under the prevailing power asymmetries between the populations involved.

Sex toys demonstrate the Object-level architecture at its cleanest. These are Objects that have never been classified as Agents and that are nonetheless processed by the Decision Hub as sexual-engagement Objects because they have been engineered to present the PA-component bodily cues that the sexual-response system is tuned to. The A-I-M profile is high-PA and high-PI with minimal or absent PM. PI is high because the activity is paradigmatically intrinsic: typically private, enjoyed in the doing, sustained for its own sake rather than instrumentally. Activities that remain stably valued when no external observer is present cannot be mimetically driven, because mimetic dynamics require at minimum an internalised model-observer in the reference set. PM is minimal because the privacy condition screens status-signalling out and the artefact typically carries no mimetic-desirability signal of its own. Individual-Objects are one subclass of Objects that present the configuration the response system is tuned to, but not the architecturally privileged subclass. What conventional framings treat as properly directed sexual behaviour is one pattern among several that the architecture can produce, and the others (cross-species Collaborator-path, cross-species Prey-path, and Object-level artefact-driven) are predicted rather than anomalous.

The archaeological record of human interbreeding with closely related hominin populations is compatible with multiple generative paths, and the archaeogenetic evidence constrains their plausible distribution. Modern non-African humans carry between one and four percent Neanderthal DNA, with additional Denisovan contributions documented across parts of Asia and Oceania. The reconstructed parental patterns are strongly asymmetric: Neanderthal males were the dominant paternal parent of the interspecies offspring that contributed to the modern genetic legacy, with Homo sapiens females as the maternal parent. This is inferred from mitochondrial and Y-chromosome inheritance patterns and from X-chromosome-linked introgression signals. The asymmetry is not compatible with symmetric Collaborator-path pair-bonding within mixed groups, which would produce interbreeding in both directions at comparable rates. It is compatible with, and most naturally explained by, generative paths in which Homo sapiens females were processed by Neanderthal males either through Prey-path extractive classifications in raiding or conflict contexts, or through asymmetric Collaborator-path dynamics in which sapiens females were integrated into Neanderthal groups following capture or exchange. Both sub-patterns are architecturally coherent, both are visible in analogous historical records of inter-group contact, and both track the power asymmetry that favoured Neanderthal-initiated extraction during the earliest phases of Eurasian contact. The framework predicts the asymmetric pattern rather than treating it as anomalous.

The contemporary rarity of cross-species sexual behaviour relative to the architectural availability of the affordance is produced by layered downstream mechanisms rather than by any species-gate in the classifier. Reproductive-waste selection disfavoured cross-species mating because such matings produce no offspring, sterile offspring, or offspring with reduced fitness. Pathogen-exposure selection disfavoured it because such contact is a vector for zoonotic transmission. These selection pressures produced the substrate on which aversive disgust responses were installed, and Confabulation generates sincere post-hoc narratives explaining the disgust in moral, religious, and cultural terms whose content varies across cultures but whose existence is architecturally predictable. The Object-level artefact case attracts a structurally similar but content-different confabulation architecture, typically framed around purity, self-restraint, or the proper direction of desire rather than around reproductive waste or pathogen exposure, because the selection pressures that shaped the cross-species aversion do not apply to artefacts.

The variability of the prohibitions across cultures is itself diagnostic. A universal human moral intuition would not require the elaborate legal, religious, and ritual architectures that actually appear in the historical record, and would not vary so systematically with the frequency of the relevant classifications and artefacts in the surrounding culture. The recurrence of the behaviour across every society with sufficient documentation, regardless of prohibition strength, confirms that it occurs at non-trivial rates wherever the architectural affordances are available. The prohibitions are confabulated elaborations of evolved aversions and cultural-purity frameworks operating against architectural affordances that the species-general Object-valuation circuitry continues to produce at their baseline rates. The argument here is confined to showing that the explanatory reach of the architecture extends to these cases as cleanly as it extends to the first-contact, pets, and slavery cases; the moral and legal status of any of these behaviours under contemporary human institutions is a separate question that the architecture does not attempt to settle.

11.17 Stable Prey Classification and the Architecture of Slavery

Slavery is the case where the Humanisation Null Hypothesis does its hardest explanatory work, because it is the phenomenon that humanisation-first accounts have most struggled to address without invoking either moral aberration or ideological false consciousness as causal drivers. HNH dissolves both moves. Slavery is not the suspension of a prior humanisation of the enslaved population. It is the stable classification of enslaved Individuals as Prey within the species-general Agent-Type taxonomy, with the institutional architecture that classification requires to remain stable under conditions of sustained contact.

The Y-dominant vector generated in slavery is architecturally distinct from the Y-dominant vectors that appear in genocide or first-contact cases (Section 11.13.1 for the formalism). It is resource-bearing Prey whose value lies in sustained extraction rather than in consumption or elimination. This is the most economically sophisticated form of Y-dominant vector the architecture can produce, because it requires the slaveholder to maintain behavioural conditions under which the enslaved Individual remains productive while the classification remains stable. The slaveholder must feed, shelter, and preserve the Prey, but must never permit behavioural evidence to accumulate that would trigger weight redistribution toward C or R. The institutional infrastructure that slave-holding societies develop is the behavioural-evidence management system the architecture requires. Slave codes, anti-literacy laws, assembly prohibitions, family-separation practices, obedience-focused religious instruction, and punitive responses to insubordination are not expressions of ideology. They are evidence-management mechanisms whose architectural function is to prevent accumulation of the reciprocity, intrinsic-engagement, and linguistic-capacity evidence that would otherwise push the classifier toward weight redistribution toward C.

The A-I-M decomposition of the enslaved Individual-Object under Y-dominant vector is diagnostic. PA is the extractive component: bodily labour and reproductive capacity as resources. PI is suppressed institutionally, which is why slave-holding societies so consistently prohibit literacy, craft specialisation, religious innovation, and self-directed activity. PM is inverted: the enslaved Individual carries negative status signal in the slaveholder’s social network, because the slaveholder gains status through ownership of Prey rather than through reciprocity with the enslaved Individual as Collaborator or Rival. The decomposition is stable precisely because the institutional architecture keeps it stable, and the institutional architecture persists because the decomposition it produces is economically productive.

This explains the persistence puzzle. Slavery endured for millennia across vastly different cultures not because humanisation was repeatedly suppressed but because no humanisation barrier was architecturally present and the Y-dominant vector was therefore cheap to maintain once the institutional infrastructure was in place. Abolition correspondingly required institutional reconstruction on the scale of the institutions that had maintained the original classification. What had to be built was not the restoration of a prior humanisation but the conditions under which redistribution of vector weight from Y toward C or R became architecturally possible: legal personhood, literacy, property rights, political participation, and the behavioural-evidence environments these conditions generate. The slow pace of post-abolition reclassification, and the persistence of residual Prey-classification artefacts in post-slavery societies, reflects the architectural cost of building new evidence environments against a prior stable classification.

The selective-reclassification phenomenon within slave-holding societies provides additional diagnostic support. Masters who developed sustained daily reciprocity with specific enslaved Individuals (household servants, skilled artisans, wet nurses) often reclassified those Individuals as Collaborators or quasi-family while maintaining Y-dominant vector for the rest of the enslaved population. The Agent-Type classifier operates Individual-by-Individual, not population-by-population, and behavioural evidence accumulated through sustained reciprocity with a specific Individual can drive reclassification for that Individual while the population-level classification remains stable. Humanisation-first accounts must invoke cognitive inconsistency or selective moral perception to explain this. HNH predicts it directly as a consequence of how the classifier operates.

The ideological architectures that slave-holding societies generate (biblical curses, natural slavery theories, civilisational hierarchies, scientific racism, paternalistic duty-of-care doctrines) are not the causal drivers of the practice but the confabulated elaboration of the extraction-ethics moral profile appropriate to Y-dominant vector. Source Opacity (Axiom 3) prevents the slaveholder from observing the Y-dominant vector directly. Confabulation (Axiom 4) produces a sincere post-hoc narrative that fills in the profile-appropriate moral content using culturally available categories. The paternalistic duty-of-care doctrines that appear so consistently across slave-holding societies (the master’s obligations to feed, shelter, and preserve the enslaved, the prohibitions on gratuitous cruelty, the ritual acknowledgements embedded in slave codes, the religious-instruction regimes framed as care for the enslaved soul) are not ideological bad faith or moral hypocrisy. They are the extraction-ethics moral profile operating at full architectural strength, generating the stewardship-adjacent moral content the profile requires. The moral consideration was present throughout the institution of slavery. It was simply the Prey-profile moral consideration rather than the Collaborator-profile moral consideration, and the two profiles are incompatible in ways that make the Prey-profile content look like the absence of moral consideration from within a Collaborator-profile perspective. The content of the confabulation varies dramatically across slave-holding societies because the self-model materials vary, but the existence of some profile-appropriate moral elaboration is architecturally necessary.

The same self-model-variation pattern identified in the first-contact case reappears in the slavery case in more elaborate form. The anthropological red herring of treating ideological content as causal driver reappears here as the historiographical red herring of treating racial ideology as the cause of slavery rather than as the extraction-ethics profile’s confabulated elaboration. Both are instances of the same architectural relationship between classification and confabulation: the content of the justification is substrate-dependent, but the existence of some profile-appropriate moral elaboration is architecturally necessary because Confabulation operates on any classification the self-model cannot directly observe.

The Diversity Referent (Chapter 14, §14.8.3.1.2) makes one clarification available, and it is needed because the substrate-dependence of the confabulated ideology can be misread as Referent-freedom. It is not: that the racial kind is elaborated differently across societies does not mean no Referent governs the kind-claim it posits. Human variation, in genes and in Self-Models alike, is non-modal, continuous, and Individual-specific, so it admits no such natural kind; the kind-claim scientific racism encodes is therefore held against an available decohered Referent and is Referent-abandoned, not Referent-free, and it is defeated by two independent channels: the genetic-empirical and the architecturally-entailed trajectory channel of Self-Model path-dependence. What the Y-dominant vector responds to is behavioural evidence read within an already-drawn boundary; the drawing of that boundary into a ranked natural kind is precisely what the Referent defeats, and the maintenance of the ranked kind, once the Referent is available, is Mimetically driven by elimination (Referent Abandonment).

The inversion HNH produces has a further consequence that the historiographical literature has typically treated as counter-intuitive but that the empirical record consistently supports. Slave-holding cultures do not precede slave-holding institutions. They follow them. The architectural sequence runs from classification to institution to culture, not from culture to institution to classification. The Agent-Type classifier performs a Y-dominant vector on available behavioural evidence, the institution of slavery emerges as the evidence-management infrastructure required to keep the classification stable, and the culture of slavery develops downstream as the confabulated justification layer that explains the institutional practice to the slave-holding population itself.

This predicts a specific pattern in the historical record. The elaborate racial ideologies, theological justifications, and civilisational hierarchies characteristic of slave-holding societies should appear as responses to institutional practice rather than as its preconditions, and they should develop their full sophistication only after the institution is already established. The pattern holds across cases. Roman legal theories of natural slavery elaborated an institution Rome had inherited rather than preceding and causing it. The racial ideologies of Atlantic slavery developed in the 17th and 18th centuries after the institutional infrastructure of Atlantic slavery was already established in the 16th. Islamic juridical frameworks for servile status developed within the legal tradition over subsequent centuries as confabulated elaboration of inherited practice. In no case does the historical record show a society developing an elaborate ideology of hereditary servile status and then implementing slavery to match. The institution always precedes the ideology, because the classification always precedes the institution.

This has a significant diagnostic consequence for abolition. If the humanisation-first view were correct, abolition would proceed primarily through moral persuasion, because changing cultural attitudes would cause the institutional practice to follow. The empirical record shows the opposite. Successful abolition movements proceeded primarily through institutional reconstruction (legal prohibition, economic disruption of slave-holding industries, political and military reconfiguration of the societies involved), and the cultural reclassification of the formerly enslaved population as Collaborators or Rivals followed the institutional changes by generations and in many respects remains incomplete. British abolition did not follow from prior cultural humanisation; racist ideologies remained entrenched in British culture long after 1833. American emancipation did not follow from prior cultural humanisation; the subsequent century of Jim Crow demonstrates that the classification persisted after the institution was legally dismantled. Abolition-as-cultural-change is architecturally backwards, because the culture is the confabulated justification layer and dismantling it without dismantling the institution leaves the classification intact and simply generates new confabulations. Abolition-as-institutional-reconstruction is what the architecture actually requires, because the institution is what stabilises the classification and the reclassification environment must be rebuilt through sustained behavioural-evidence infrastructure rather than through moral appeal.

11.18 Derivation of the Presumption Against Scapegoating

The Dual-Role Confabulation signature provides a diagnostic test for identifying confabulated scapegoating. But a diagnostic test presupposes a prior question: when should the test be applied? The answer requires establishing the Presumption Against Scapegoating: the epistemic default that a Grievance Belief, a Belief attributing a felt deficit to the conduct of an identifiable agent or agent-class (derived at Appendix A, Section A.11.8; prose derivation at Chapter 16, Section 16.3), should be presumed Confabulated until specific evidential tests are satisfied.

The Derivation

The Presumption follows from four premises established in this chapter and its dependencies:

Premise 1 (Source Opacity, Chapter 9): The destruction of source-tags during Common-Currency Integration makes accurate introspection of motivational source architecturally impossible. Individuals cannot reliably distinguish M-driven urgency from A-driven necessity through introspection alone.

Premise 2 (Agent-Based Causal Default, Axiom 9): Human causal cognition is architecturally optimised for detecting and modelling agents, defaulting to agent-based explanations when causal demands arise. When confronted with effects requiring explanation, the cognitive system searches for predators, rivals, collaborators, and prey, not for emergent systemic dynamics.

Premise 3 (Information Gap, derived from Axiom 9 and Axiom 4): True causes in modern environments (preconscious Mimetic transmission, diffuse institutional processes, complex feedback loops, Entanglement) exceed the computational tractability of available causal-inference systems.

Premise 4 (Confabulation Architecture, this chapter, §11.2 Confabulation Rationality Lemma): Given Premises 1–3, Confabulation is the expected output when Individuals attempt to explain felt deficits. I's Bayesian ground-resolution function returns agent-based explanations because the M-Hypothesis is absent from its candidate list, or is present but outweighed by an incumbent attribution the Belief Register re-emits as Evidence, and because systemic causes exceed the tractability of the available Hypothesis Space; these explanations achieve high subjective confidence because the Bayesian updating itself is content-agnostic and rational. The pathology is in the input to the updating, not in the updating.

Derivation: If agent-based causal attribution is the cognitive default (Premise 2), and true causes exceed tractability (Premise 3), and I's Bayesian ground-resolution function returns confident attributions over whatever Hypothesis Space it holds (Premise 4), then the base-rate probability that any given Grievance Belief is Confabulated is very high, not because humans are irrational, but because the conditions for accurate attribution (an M-Hypothesis on the candidate list and tractable systemic content) are rarely met.

This base-rate analysis establishes the Presumption:

The Presumption Stated

Presumption Against Scapegoating: Any Grievance Belief, any assessment that identifies an agent or agent-class as the primary causal factor for a felt deficit, must be presumed Confabulated until satisfaction of the Grievance Gateway (the Gateway; Origin Check, A-Check, I-Check) and, for grievances passing the Gateway, the Three-Test Battery (Immediacy, Mechanism, Authorship), because human causal-inference systems are architecturally constrained to produce agent-based explanations regardless of whether agents are actually responsible, and these confabulated explanations achieve high subjective confidence precisely because they are generated by well-functioning cognitive machinery operating outside its domain of competence.

Distinguishing the Presumption from Legal Innocence

The legal presumption of innocence operates as follows: the accused is presumed innocent until proven guilty beyond reasonable doubt. The prosecution bears the burden of proof. The justification is normative; it reflects a moral judgment that convicting the innocent is worse than acquitting the guilty.

The Presumption Against Scapegoating differs in three critical respects:

Feature	Legal Presumption of Innocence	Presumption Against Scapegoating
Justification	Normative: moral values about error distribution	Architectural: facts about cognitive malfunction
Status	A choice among alternatives	A necessity given the architecture
What it presumes	The accused did not do the act	The attribution is confabulated, not merely unproven

The architectural grounding transforms the Presumption from should to must. The legal presumption says: we choose to require proof before conviction. The Presumption Against Scapegoating says: absent proof satisfying specific tests, the attribution is almost certainly wrong because of how human causal cognition malfunctions under Information Gap conditions.

The two presumptions also differ in how they end. The legal presumption is discharged when the trier of fact is sure, and a verdict of guilt then licenses every member of the community to be certain of the Grievance Belief it settles (Chapter 16, Sections 16.6 and 16.8). The Presumption Against Scapegoating is discharged only by the tests that follow, never by conviction.

The Inverse Confidence Principle

The Presumption includes a counter-intuitive element absent from the legal presumption: confidence is inversely correlated with accuracy in confabulated contexts.

The legal presumption treats confident testimony as stronger evidence. The Presumption Against Scapegoating recognises that, where the Information Gap is operative, confident testimony is evidence of Bayesian compounding on corrupted data. The more people are certain about the agent-attribution, the more likely the Confabulation Cascade has achieved Collective Fixed Point (Chapter 13).

This inverts standard evidential intuitions:

“Everyone knows they’re responsible” → elevated probability of collective Confabulation
“The pattern is so obvious” → elevated probability that visibility rather than causation is driving the attribution
“I’ve thought about this carefully” → elevated probability of elaborated Confabulation

The Grievance Gateway: Pre-Screening Filter

Before the Three-Test Battery is applied, the Individual must first pass the Grievance Gateway, a pre-screening filter that tests whether the Grievance Belief is grounded in a traced deficit rather than socially copied.

Step 1: Origin Check. "Whose deficit is this, and how did it reach me?" The Origin Check passes where the Individual has traced access to an Appetitive or Intrinsic deficit: they sustained it themselves, or they can identify a specific sufferer and a specific harm to that sufferer's A or I provision, reaching them by a traceable route rather than as circulating account. It fails where the deficit is known only as expressed Belief, absorbed from a social group, a news cycle, a feed, or a community narrative, with no route back to a sufferer and a harm. The Origin Check screens provenance rather than proximity, which is what separates the witness from the audience: a Grievance Belief held on another's behalf passes where its route is traceable and fails where it is not.

Step 2: A-Check. "Does my body need the thing being denied?" This tests whether the grievance has an Appetitive foundation.

Step 3: I-Check. "Would I still care about this if nobody knew?" This tests whether the grievance has Intrinsic grounding.

Only grievances passing the Origin Check (Step 1) and at least one of the A-Check or I-Check (Steps 2/3) proceed to the Three-Test Battery. A grievance that passes the Origin Check but fails both the A-Check and the I-Check is likely M-driven and should be redirected to the Two-Question Diagnostic rather than the scapegoat pathway.

What the Grievance Gateway settles is not whether the holder may target the agent, because targeting is not what a passing grievance licenses. It settles whether the holder has a traced deficit whose correction the field-correction obligation places on them. Where the Gateway passes, what follows is Cascading Support and Cascading Consequences (Chapter 16, Sections 16.15 and 16.16) and nothing further: restoration of the sufferer's Appetitive and Intrinsic provision, and correction across the whole generating field, which includes provision to the aggressor, or to the aggressor's population, where a genuine deficit of theirs loaded the loop, since the Appetitive branch closes by supply and no node is excluded from the obligation, the perpetrator included. The purpose is to lower the loop rather than to settle the account. Nothing available undoes the harm already done; what provision reaches is the harm not yet done. Where the Gateway fails, the holder has no traced deficit to correct, and Epistemic Suspension governs.

The Three-Test Battery: Rebuttal Standard

The Presumption is rebuttable, but only through satisfaction of the Grievance Gateway (above) and the following specific evidential tests that establish the Grievance Belief is tractable rather than confabulated. The Three-Test Battery provides this standard:

Test 1: Immediacy Are these people literally and immediately blocking me from accessing what I need for my body right now?

This tests whether the grievance represents acute A-threat or diffuse felt deficit. The key terms (literally, immediately, right now) block temporal displacement (“they will eventually cause harm”), metaphorical extension (“they’re blocking me symbolically”), and speculative projection (“if this continues...”).

If NO → The grievance is not an acute A-threat requiring emergency defensive action. Proceed to systemic analysis.

Test 2: Mechanism What is the clear and obvious mechanism that they are using, and can only be used by them, to stop me from getting what I need?

This tests tractability (can the mechanism be specified concretely?) and exclusivity (is the mechanism uniquely attributable to the accused?). “Armed guards at the entrance” is tractable; “they somehow make resources scarce” is not. “Only they control the guards” is exclusive; “they participate in the same market as everyone else” is not.

If cannot specify → The mechanism gap reveals the Information Gap. The causal-inference system has exceeded its computational capacity and is completing a Confabulation to achieve closure.

Test 3: Authorship Did they personally design and build the system to block access to what I need?

This tests whether the accused occupy the origin point of causal chains or merely visible nodes within emergent systems. Most M-captured systems (housing markets, credential hierarchies, prestige economies) are emergent properties of distributed Mimetic dynamics. No individual or identifiable group designed them to exclude.

If NO → Targeting the accused addresses a symptom (visible beneficiary) rather than a cause (system architecture or emergent dynamics).

The Anti-Conflation Safeguard

When Test 3 yields NO, a predictable evasion occurs: “Well, they didn’t design it, but they’re benefiting from it, so they’re still responsible.” This is the Beneficiary-Cause Conflation, substituting outcome description for causal explanation. The brain achieves cognitive closure by identifying visible beneficiaries as responsible-as-if-architects.

The safeguard question blocks this move:

Is benefiting from a system the same thing as designing it? Could someone benefit from a system they didn’t create, didn’t choose, and would change if they could?

Once the Individual acknowledges that benefit ≠ authorship, the beneficiary-as-cause move is blocked, and causal inquiry must either continue backward (who did design this?) or acknowledge emergence (no one designed it: it’s an emergent property of distributed dynamics).

When the Presumption Is Rebutted

The Presumption Against Scapegoating does not deny that attributions are sometimes accurate. When all three tests are satisfied, the attribution may be accurate:

Example: An employee’s final paycheck is withheld and security physically prevents building access to collect personal belongings.

Grievance Gateway Origin Check: YES, this is my personal experience, not a copied narrative
Grievance Gateway A-Check: YES, the paycheck is needed for rent and food
Grievance Gateway I-Check: YES, I would still care if nobody knew
Gateway: PASSED. Proceed to Three-Test Battery.
Q1 Immediacy: YES; security guards physically prevent entry; paycheck and items are inside
Q2 Mechanism: YES, armed security, badge deactivation, locks changed; only they control these access points
Q3 Authorship: YES: management specifically implemented this policy after termination; HR documented the decision

Outcome: The Grievance Belief is tractable and accurate. The tractable action is recovery of the pay and the belongings through legal action or a labour board complaint.

The Three-Test Battery distinguishes tractable from confabulated attributions. When the tests pass, the Presumption is rebutted: the attribution is accurate, and the agent is established as the gate on the Object. When the tests fail (as they do for the vast majority of felt deficits in complex modern societies), the Presumption holds, and cognitive resources should be redirected toward systemic analysis. The Practical Redirect follows the Three-Test Battery in either case, the question after any diagnostic being what would actually solve this. Where the tests fail the answer is systemic, and attention moves from who to punish toward which structural levers would reduce A-harm, restore Intrinsic Motivation, or damp Mimetic Desire in the system. Where the tests pass the same question returns the tractable action: in the case above, recovery of the withheld pay and the belongings, and not punishment of the managers who withheld them, which answers nothing the question asked.

The answer the question returns is a structural lever in either case, never a person, and which lever is fixed by the substrate. Chapter 16, Section 16.13 partitions it: the Appetitive branch closes by supply, the Intrinsic branch by reinstatement in kind, and the Mimetic branch by neither, a comparative position being an Object that cannot be furnished twice. What a cleared Battery adds is the forward limb. Provision to the aggrieved, and to the field around the author, is live either way, each field deformed around its node requiring correction whether or not an agent gated the Object, which is Cascading Support (Chapter 16, Section 16.15). Where an agent is tractable, the question also asks what would prevent recurrence, and that runs through provision again wherever the agent's own Appetitive or Intrinsic deficits loaded the loop, since closing those branches at source removes the drive. The Mimetic residue is what remains, addressed by Cascading Consequences (Chapter 16, Section 16.16) on the levers and channels that loaded the stake, floor-preserving throughout, so provision is not suspended for the party the Three-Test Battery cleared against. Containment is the limiting case and narrower than that residue, reserved for a node the Eliminative Structure cannot dissolve at any branch.

Connection to Dual-Role Confabulation

The Presumption Against Scapegoating and the Dual-Role Confabulation signature operate as complementary diagnostics:

The Presumption establishes the epistemic default: presume confabulated until tests are satisfied
The Grievance Gateway provides the upstream filter: Origin Check, A-Check, I-Check, screening whether the grievance is grounded in a traced deficit before heavier diagnostic machinery is applied
The Three-Test Battery provides the rebuttal standard: Immediacy, Mechanism, Authorship
The Dual-Role Signature provides independent confirmation: if the target is selected as prey but framed as predator, Confabulation is near-certain regardless of how the Three-Test Battery is answered

An attribution that fails the Three-Test Battery and exhibits the Dual-Role Signature is confabulated with extremely high probability. The diagnostics reinforce each other: either alone provides strong evidence of Confabulation; together they are decisive.

Every element of the apparatus tests a structural property of the causal chain a Grievance Belief asserts, traceability, mechanism-specificity, temporal immediacy, personal authorship, selection-framing coincidence, and none tests whether the felt deficit is real, whether the grievance is sympathetic, or what the attribution says. That is the form the Belief Multipolarity Principle requires of any intervention in the Subcriticality Maintenance set: specifiable on a rate, a threshold, or a structural property of the arrangement, and never on Belief content, because a body converging on which attributions to credit would have to obtain the agreement that only coupling supplies in a domain where no Referent is available to settle it (Chapter 2; Chapter 13, Section 13.8.5).

MOVEMENT D: SIGNATURES AND IMPLICATIONS

11.19 The Diagnostic Signature of Dual-Role Confabulation

The analysis of the Scapegoat Mechanism yields a diagnostic criterion that can be applied at both individual and population scales:

Dual-Role Confabulation: When a Grievance Belief exhibits prey-selection criteria (vulnerability, visibility, reciprocity-exemption) combined with predator-framing (threatening, taking, blocking), the attribution is confabulated with near-certainty.

The architecture beneath it is stated at Chapter 4, Section 4.4.7: prey-selection and predator-framing are the two aspects under which the Observer's signed channel reads one Individual-Object, and the Decision Hub receives one signed scalar carrying neither the aspect nor the coordinate that produced it.

Why the Signature Is Diagnostic

Genuine predator-identification exhibits predator-selection coinciding with predator-framing; the agent is targeted because they are threatening, not despite being vulnerable. The signature is diagnostic because it tracks an architectural impossibility:

In genuine threat scenarios:

The agent exhibits predator-characteristics (capability, hostile intent)
These characteristics trigger predator-detection systems
Selection occurs because of the predator-characteristics
Framing matches selection; the threat is real

In confabulated scapegoating:

The agent exhibits prey-characteristics (vulnerability, reciprocity-exemption)
These characteristics trigger prey-detection systems for target selection
But the Confabulation requires predator-framing for justification
Framing contradicts selection; the threat is constructed

A single diagnostic question can reveal the Dual-Role Signature:

If they are as powerful and threatening as described, why can we target them without expecting overwhelming retaliation?

Genuine predators cannot be safely targeted without superior force. If a target can be persecuted, expelled, or eliminated by a group lacking overwhelming military advantage, then the target was selected through prey-detection, not predator-detection; regardless of how the targeting is framed.

Application at Individual Scale

The Dual-Role Signature can be self-applied when an Individual notices a strong Grievance Belief about their own felt deficit:

Step 1: Identify the attribution. “I believe X is responsible for my deficit.”

Step 2: Check prey-selection criteria. Is X vulnerable relative to me? Is X outside my reciprocity network? Is X visible and identifiable?

Step 3: Check predator-framing. Am I describing X as threatening, taking, blocking, powerful?

Step 4: Test coherence. If Steps 2 and 3 both yield YES, the Dual-Role Signature is present. The attribution is likely confabulated.

Step 5: Redirect. Apply the Three-Test Battery. If the Three-Test Battery fails, redirect cognitive resources from agent-targeting to systemic analysis.

Application at Population Scale

The Dual-Role Signature is especially valuable at population scale, where Confabulation Cascades can achieve overwhelming social proof. When a society converges on a designated enemy, the signature provides a check:

Step 1: Identify the collective attribution. “Our society believes Group X is responsible for our collective problems.”

Step 2: Check prey-selection criteria. Is Group X a numerical minority? Does Group X have reduced political power, legal vulnerability, or limited capacity for retaliation? Is Group X outside the dominant reciprocity network?

Step 3: Check predator-framing. Is Group X described as powerful, threatening, conspiratorial, responsible for large-scale harms?

Step 4: Test coherence. If Steps 2 and 3 both yield YES, the Dual-Role Signature is present at population scale. The collective attribution is almost certainly confabulated.

The signature’s power is that it operates independently of the specific content of the accusation. It does not require evaluating whether the accused actually did what they are accused of doing. The structural incoherence between selection and framing is diagnostic regardless of content, which makes it applicable across cultures, periods, and specific grievances.

11.20 Pseudo-Empathy and the Empathy-Mimesis Antagonism Principle

Empathy is Intrinsic. It is process-rewarding, it persists without an audience, and it arises from the relatedness need Chapter 7 classifies under I. Its characteristic function is the registration of another Individual's state and the restoration of their access. Nothing in the architecture makes care mimetic, and the Confabulation pathway developed in this chapter does not begin by doubting it.

The vulnerability is downstream of the care, not inside it. When empathy activates, the Individual asks why the other is suffering, and that question opens the Information Gap of Section 11.6: the true causes of distress in a complex social system exceed what the causal-inference systems can compute. The agent-based causal default resolves the gap as it resolves every other, by supplying an agent. Systemic answers require abstract reasoning the threat-detection circuits cannot perform; agent-based answers activate machinery built for exactly this. So the answer that arrives is that someone is responsible.

Empathy and that attribution then fuse. The Individual does not experience feeling for the sufferer and separately identifying a perpetrator; the two arrive as one state, in which caring about the victim and knowing who did it are the same thought. The fusion converts the care into targeting permission. Compassion supplies the moral licence, because action against the named party is experienced as protection rather than aggression. It supplies the coalition signal, because expressed concern demonstrates membership among those who share the attribution. And it supplies intensity, because the more compassion is felt for the sufferer, the more legitimate action against the accused becomes. This is Pseudo-Empathy, and the conversion is the Empathy-Mimesis Antagonism Principle: unmanaged Mimetic Desire structurally prevents empathy from discharging its Intrinsic function by converting it into targeting permission for scapegoat operations.

Pseudo-Empathy has a signature, and it separates cleanly from care by what each tracks. Pseudo-Empathy intensifies with the vividness of the perpetrator-narrative rather than with proximity to the sufferer. It requires the attribution to hold: remove the identified agent and the motivational force dissipates or turns on whoever removed it, because questioning the attribution is experienced as attacking the victims. And it produces action directed at the accused rather than restoration for the suffering, which is the diagnostic that matters, since the Intrinsic function was restoration and the output delivers none of it.

None of this is insincere. The Individual believes they are moved by compassion, and by Axiom 4 the belief is the honest output of the architecture. What has happened is that a real Intrinsic signal was routed through an attribution the Individual could not inspect, and the routing completed before deliberation could reach it.

The conversion is conditional, and the conditions are identifiable. It requires the Information Gap to be operating, so suffering with a visible and proximate cause is far less vulnerable than diffuse suffering that must be inferred. It requires M-salience to be elevated, so the mechanism weakens where status concerns are quiet and the A-Floor is secured. And it requires a categorical boundary, so it dissolves where the Individual recognises the accused as subject to the same Source-Opaque processes they are. The principle therefore does not hold that empathy is compromised, but that under these three conditions its capture is architecturally expected, which is why I-Override is a necessary condition for empathy to do what it is for.

11.21 The Non-Malice Corollary

The Confabulation architecture established in this chapter yields a corollary:

Non-Malice Corollary: Most Grievance Beliefs that generate conflict are confabulated rather than accurate, which means that most conflict arises from cognitive architecture operating under Information Gap conditions rather than from malicious intent, deliberate deception, or accurate perception of genuine threat.

The Reframing

This corollary reframes the default interpretation of conflict. Standard interpretation:

“They are targeting us because they hate us”
“They believe lies about us because they want to harm us”
“Their accusations reveal their malice”

AIM-informed interpretation:

“They are targeting us because their causal-inference systems generated an agent-based completion for their felt deficit, and we were selected through prey-detection criteria”
“They believe inaccurate narratives about us because the Information Gap prevents accurate attribution and Confabulation fills the gap”
“Their accusations reveal their cognitive architecture under Information Gap conditions, not their character”

The Non-Malice Corollary does not claim that malice never exists. Some Grievance Beliefs are accurate. Some targeting is genuinely malicious. The corollary claims that malice is not the default explanation; architectural Confabulation is. The burden of proof shifts: where previously malice was assumed and innocence required proof, now Confabulation is assumed and malice requires proof.

Implications for Response

The Non-Malice Corollary transforms appropriate response to being targeted:

Standard response (malice-assumption):

Counter-targeting (“they are attacking us, so we must attack them”)
Moral condemnation (“they are evil for targeting us”)
Defensive escalation (“we must demonstrate strength to deter further attacks”)

AIM-informed response (Confabulation-assumption):

Systemic analysis (“what Information Gap are they trying to close?”)
Compassionate diagnosis (“they are confused, not evil”)
Intervention design (“how can we disrupt the Confabulation Cascade?”)

The shift from malice-assumption to Confabulation-assumption does not require passivity. Defending against targeting remains appropriate: the Scapegoat’s safety matters regardless of whether the targeting is malicious or confabulated. But the interpretation of the targeting shifts, which changes downstream responses.

Counter-targeting based on malice-assumption is itself a Grievance Belief that faces the same Information Gap. When Group A targets Group B, and Group B responds by targeting Group A (“they are evil aggressors”), Group B has generated its own Confabulation, attributing malice where architectural Confabulation is the more parsimonious explanation. This produces the reciprocal Confabulation spiral: each group’s targeting confirms the other group’s predator-attribution, escalating conflict through feedback dynamics that neither group can perceive because both lack the AIM vocabulary to recognise Confabulation.

The Non-Malice Corollary interrupts this spiral by reframing the initial targeting as cognitive architecture rather than moral failing. The reframe does not excuse the targeting; it explains it in terms that permit de-escalation rather than requiring escalation.

Edge Case: Genuine Malice

The Non-Malice Corollary does not claim that all targeting is confabulated. Some Individuals and groups do target others through accurate perception and deliberate choice. How does one distinguish genuine malice from architectural Confabulation?

The Three-Test Battery provides the distinction:

If the targeting passes all three tests (Immediacy, Mechanism, Authorship), the attribution may be accurate and the targeting may be genuinely malicious
If the targeting fails any test, Confabulation is the more likely explanation regardless of how the targeting feels to either party

The Dual-Role Signature provides additional evidence:

If the targeting exhibits prey-selection with predator-framing, Confabulation is near-certain
If the targeting exhibits predator-selection with predator-framing (targeting a genuinely threatening agent because they are threatening), the targeting may be defensive rather than scapegoating, though even here, the possibility of Confabulation should be entertained

The practical implication: genuine malice exists but is rarer than it appears. Most conflict that feels like malice, most targeting that seems deliberate and hostile, is actually architectural Confabulation that would not occur if the targeting population had accurate causal models. Interventions that provide accurate models are therefore more promising than interventions that assume malice and attempt deterrence.

"Only a Joke"

A joke told at another Individual's expense is a downward attribution: the telling lowers the target's computed position in the field of every Individual present, and by the Extraction-Direction Asymmetry (Chapter 10, Section 10.3.5) a downward attribution extracts more Mimetic Premium per act than an upward one. The teller's integrated valuation of the telling therefore carries two contributions. The first is the Thinker's valuation of the resolution, which is Humour (Chapter 7, Section 7.5.7). The second is Mimetic: the Observer's reading of the teller's own gain in standing, returned through the teller's executed action and through the responses of those present (Chapter 4, Section 4.4.9.1). Integration strips the source tags, so the teller feels one amusement and holds no phenomenological property by which the Mimetic contribution could be told from the Thinker's.

When the targeted Individual objects to being made fun of, the teller's Thinker is recruited to resolve what the amusement was. By the Confabulation Rationality Lemma (Section 11.2), Ground-Resolution returns the best-supported ground the Hypothesis Space holds, and that ground is an I ground: the amusement is attributed to the joke and the joke to its cleverness. Where the Hypothesis that the amusement was Mimetic is absent from the Hypothesis Space, the attribution follows under Type I Exclusion. Where it is present, probability mass has already concentrated on the joke's cleverness, because the amusement was felt and resolved before any objection arrived, which is the order-of-encounter mechanism of Type II Exclusion. "It was only a joke" is therefore a sincere report. The teller does not know that the joke was a positional act, because the only evidence the teller holds about the amusement is a Scalar from which the positional contribution has been removed.

The account does more than report the teller's state. Offered to the target, "it was only a joke" is Evidence that the telling was not a positional act, and the teller's insistence that the target accept the account carries its own valuation, which is Mimetic. The Mutual-Regard Equilibrium (Chapter 10, Section 10.3.5) establishes that the Mimetic Premium is the differential between mutual deferences. A target who accepts the account defers: the target concedes that the lowering was not a lowering and that the objection was a misreading, and the differential the telling opened is held rather than contested. The Mimetic Premium the joke extracted is thereby secured by the target's own concession, and the teller experiences the concession as the target coming to see the joke correctly.

A target who declines to defer meets the reciprocal structure of the Non-Malice Corollary. The teller, holding the I ground sincerely, experiences the refusal as a failure to perceive what the joke plainly contained, and the characteristic reply, that the target cannot take a joke, is a second downward attribution: the target's standing is lowered for the objection as it was lowered by the joke. The target, for the target's part, reads the insistence as malice, which is the reading the Non-Malice Corollary displaces. The lowering is real, and the intent the target attributes to it is not. The AIM-informed response names what occurred: the joke was a positional act, the target's standing was lowered, and the teller did not know it. None of this requires the target to defer or the teller to be condemned.

11.22 The Moral Vocabulary Displacement Principle

A sophisticated objection to the Presumption Against Scapegoating arises: even if the agent didn’t design the system, don’t they have a moral obligation to change it? Doesn’t justice require accountability?

The Moral Vocabulary Displacement Principle addresses this:

Moral Vocabulary Displacement Principle: In the context of a Grievance Belief, moral vocabulary (“they are responsible,” “they have obligations,” “justice requires accountability”) functions not as legitimate ethical inquiry but as a justification layer: where the attribution is confabulated, for targeting an agent who did not author the harm; where the attribution is accurate, for the retributive surplus beyond repair. Engagement with moral framing on its own terms strengthens the mechanism it serves in either case; redirection to practical effectiveness is the only response that supplies neither justification.

The Moral Vocabulary Displacement Principle Established

The principle establishes three claims:

Claim 1: Moral vocabulary is not parallel to scapegoating; it is constitutive of scapegoating.

The moral question is not a separate inquiry that happens to accompany Scapegoat narratives; it is the mechanism by which Scapegoat narratives achieve permission. The accusation “they are responsible” transforms aggression into righteousness. Without the moral frame, targeting would be experienced as unprovoked attack. With the moral frame, targeting is experienced as justified response to wrongdoing.

Claim 2: Engaging moral questions on their own terms strengthens the mechanism.

When someone asks “but don’t they have a moral responsibility?” and receives the response “perhaps, but...”, the “perhaps” concedes the frame. The Scapegoat Mechanism has received partial permission. The conversation has been pulled into evaluating whether the target deserves targeting, which is precisely the evaluative frame that permits targeting. Conceding authorship where the Three-Test Battery has established it does not concede that frame: authorship is a causal finding, and desert is a claim about what the author is owed.

Claim 3: AIM’s practical focus is not strategically preferable but theoretically mandatory.

If moral vocabulary is scapegoat infrastructure, then the only non-scapegoating frame is practical: Would targeting them restore your access? This question bypasses the permission-seeking structure of moral vocabulary entirely. It does not ask whether targeting is deserved: it asks whether targeting would work.

The Sacred-Secular Continuity

Historically, religious vocabulary served the justification function. The victim was not merely killed but sacrificed, which sanctified the violence. The Scapegoat carried the community’s sins; expulsion or death restored sacred order. Religious framing transformed collective violence into cosmic duty.

In secular contexts, moral vocabulary has inherited this function entirely:

Sacred Scapegoating	Moral Scapegoating
“The gods demand sacrifice”	“Justice demands accountability”
“The victim is polluted/cursed”	“They are morally responsible”
“Sacrifice restores cosmic order”	“Accountability restores justice”
“We perform sacred duty”	“We fulfil moral obligation”

The structure is identical. The moral framing sanctifies the targeting. It transforms aggression into righteousness. This is why moral vocabulary cannot rebut the Presumption Against Scapegoating; it is itself part of the Scapegoat Mechanism’s immune system, not an independent check on the Scapegoat Mechanism.

The Practical Frame as the Ethical Frame

The practical frame is the ethical frame, properly understood.

If someone genuinely wants to address injustice (to actually improve conditions for those experiencing deficit) then they must ask: what would work? Persisting with targeting that won’t work is not ethical seriousness; it is Scapegoat Mechanism capture masquerading as ethical seriousness.

The practical redirect is not amoral pragmatism displacing ethics. It is what ethics requires once we understand that moral vocabulary in scapegoat contexts functions as targeting permission. True ethical concern for those experiencing deficit demands the practical question: what would actually restore their access?

Scapegoat-Based Approach	Systemic Approach
Requires moral transformation of targets	Requires only institutional redesign
Depends on targets’ voluntary compliance	Depends on policy implementation
Triggers defensive counter-escalation	Bypasses interpersonal confrontation
Leaves system architecture intact	Changes the rules of the game
Produces temporary catharsis at best	Addresses actual causal dynamics

Systemic solutions work regardless of anyone’s moral stance because they change the structure within which Individuals operate, not the Individuals themselves. The moral framing, by contrast, is pseudo-ethical: it performs righteousness while pursuing action that won’t help anyone except by providing temporary catharsis through targeting.

The Absolute Redirect

The redirect is absolute, and it is absolute whether the Three-Test Battery fails or passes. Where the Three-Test Battery fails, the moral frame is seeking permission to target an agent who did not author the harm. Where the Three-Test Battery passes, the attribution is accurate and the agent did author the harm, and the moral frame of desert then seeks something else: the retributive surplus, the cost imposed beyond what reparation of the harm requires. The Punitive Inversion Principle holds for an accurate attribution exactly as for a confabulated one, because punishment moves the couplings it reaches in the direction that raises the Loop Gain whoever authored the harm (Chapter 16, Section 16.9). Punishment is an input to the loop and not a coupling in it, and only the couplings enter the stability condition (Appendix A, Section A.5.1). The Revenge Cascade holds for it too. A retribution discharged against a true author is a Harm inflicted on an agent who sits in a field of their own, so the avenged wrong, entering the fields anchored to its target through the network Evidence coupling (Appendix A, Section A.5.7), becomes the ignition of the next loaded loop, and the discharge adds a Harm to the field rather than subtracting one (Chapter 16, Section 16.12). The two cases run alike by Truth-Value Orthogonality: the accuracy of a Grievance Belief is an axis the loop does not read. At the mutual-consistency fixed point of the Interactional Confirmation Loop, the distance of either belief from its truth is unconstrained (Appendix A, Section A.11.8).

The redirect therefore concedes authorship where the Three-Test Battery has established it and declines desert in every case. Authorship is a causal finding, and the Three-Test Battery settles it. Desert is the claim that the author is owed a cost beyond repair, and nothing the Three-Test Battery tests bears on that claim.

When moral vocabulary appears: “They have a responsibility,” “They’re benefiting unjustly,” “Justice requires accountability,” “They deserve to pay,” “They owe me an apology.”

Recognise: Where the Three-Test Battery has failed, this is the justification layer of the Scapegoat Mechanism seeking permission to target. Where the Three-Test Battery has passed, it is the demand for the retributive surplus seeking permission to impose it.

Redirect, where the Three-Test Battery has failed: “Notice what’s happening: the moral question is seeking permission to target. The diagnostic has already established that targeting won’t restore your access. The question isn’t who is responsible; it’s what would actually solve the problem. What would?”

Redirect, where the Three-Test Battery has passed: “Yes, they did this, and the diagnostic confirms it. The question now is what repairs the harm. Punishing them will not repair it: punishment raises the gain of the loop you are both in, and it spreads the harm to people around them who did nothing to you. What gets you back what was taken, and what stops it happening again? No one will punish them on your behalf. If you want retribution, you will have to take it yourself, and then you will stand where they stand now.”

Where the Three-Test Battery has passed, the choice the redirect puts belongs to the harmed Individual and to no one else, and the architecture fixes both of its outcomes. The choice is the harmed Individual’s because the design target the architecture fixes for a justice system is a state that declines the role of punisher (Chapter 16, Section 16.9), so a retribution the harmed Individual chooses is a retribution the harmed Individual performs. Pressing the demand for punishment drives the Loop Gain up and produces a net increase in Appetitive and Intrinsic harm across the network, whatever the truth of the attribution. Turning to repair is the other course. The Appetitive and Intrinsic harm is repaired by reparation, which Cascading Support places ahead of any consequence for the author (Chapter 16, Section 16.15). The Trauma is repaired by the harmed Individual’s own Bayesian updating, which no other party can run on the harmed Individual’s behalf, because the Belief Register is written only by the Thinker (Chapter 4, Section 4.4.8) and the recall of an Individual’s own prior state is the one slow-layer coupling that cannot be reduced (Appendix A, Section A.5.6), and which reparation enables by restoring the stability that updating requires (Chapter 16, Section 16.12). Punishment does neither. Harming the author works no Self-Model toward repair, the harmed Individual’s included.

Where the gate is still closed, opening it is the first repair. The legal process that recovers withheld pay, and the defensive act that ends an assault in progress (Chapter 17, Section 17.15), are repair and not punishment.

None of this leaves the author without consequence. The author remains accountable for the consequences of the act, and those consequences are reparation of the harm done and prevention of recurrence, which Cascading Consequences directs at the levers and channels that made the harm possible, floor-preserving throughout (Chapter 16, Sections 16.8 and 16.16). What the redirect removes is the surplus, the one part of the response that repairs nothing. The consequences differ from the ones that feel right, because the retributive impulse is what feels right, and the retributive impulse is a failure mode of the architecture that presents as obligation (Chapter 16, Section 16.12).

This redirect:

Names the function: the moral question is seeking permission, to target or to impose the surplus
Exposes the mechanism: moral vocabulary → targeting justification, or surplus justification
Returns to practical: what repairs the harm, and what prevents recurrence?
Concedes authorship where the Three-Test Battery establishes it, and never concedes desert

11.23 Anticipating Objections

This section addresses likely objections to the Confabulation architecture and the Presumption Against Scapegoating.

Objection 1: “The ‘Sometimes Accurate’ Problem”

Grievance Beliefs are sometimes correct. If the Presumption treats all such attributions as confabulated, won’t it produce false negatives, failing to identify genuine threats?

Response: The Presumption is rebuttable, not absolute. The Three-Test Battery provides the rebuttal standard. When an attribution passes all three tests (Immediacy, Mechanism, Authorship), the Presumption is rebutted and the attribution stands. What a rebuttal licenses is action on the gate, not punishment of the agent (Section 11.22).

The Presumption’s function is to shift the default, not to prohibit accurate attribution. Without the Presumption, the default is to assume agent-attribution is accurate (because it feels accurate due to high subjective confidence). This default produces systematic false positives; the Scapegoat Mechanism is historical pattern, not theoretical speculation.

With the Presumption, the default shifts: assume confabulated until tests are satisfied. This may produce some false negatives (failing to identify genuine threats), but the base-rate analysis suggests false negatives will be far less common than the false positives prevented. In modern complex societies, the vast majority of Grievance Beliefs are confabulated because true causes are systemically diffuse.

The question is not “can we achieve perfect accuracy?” but “which error distribution is preferable?” Given the historical catastrophes produced by false positives (genocides, pogroms, mass persecution), and the relative rarity of situations where agent-attribution is both accurate and urgent, the Presumption’s error distribution is preferable.

Objection 2: “The Paralysis Problem”

Doesn’t the Presumption paralyse legitimate accountability? If we must presume all agent-attributions confabulated, we can never hold anyone responsible for anything.

Response: The Presumption does not prohibit accountability; it requires that accountability satisfy evidential standards before proceeding.

When the Three-Test Battery passes, accountability is appropriate. The employee whose paycheck is withheld can pursue legal remedies. The Individual whose property is stolen can seek restitution. The population subjected to military aggression can defend itself.

What the Presumption prohibits is presumptive accountability: holding agents responsible based on felt grievance without evidential support. This prohibition is not paralysis; it is Epistemic Hygiene, the class of practices maintaining the conditions under which a Belief's Referent condition can be determined and its certainty held accordingly (Chapter 20, Section 20.4.5). The alternative, permitting accountability based on felt grievance alone, is precisely the mechanism that produces scapegoating.

The objection also conflates two different accountability structures:

Tractable accountability: Agent A performed specific action X that caused specific harm Y through identifiable mechanism Z. This is what the Three-Test Battery tests for.
Diffuse accountability: Agent A benefits from system S that produces outcome O that I experience as harmful. This is what the Presumption identifies as likely confabulated.

The Presumption does not prohibit tractable accountability. It prohibits treating diffuse accountability as though it were tractable, which is the cognitive operation that produces scapegoating.

Objection 3: “The Elite Protection Problem”

Could the Presumption be weaponised to protect genuinely harmful actors? If powerful groups cause harm through diffuse systems, doesn’t the Presumption let them off the hook?

Response: First, the Presumption does not claim that diffuse systems are harmless.

First, the Presumption does not claim that diffuse systems are harmless. It claims that a Grievance Belief for diffuse systemic harm is likely confabulated. The harm is real; the attribution is inaccurate.

Second, the Presumption redirects attention from agent-targeting to systemic intervention. If housing unaffordability is caused by emergent market dynamics rather than identifiable villains, the Presumption does not say “therefore housing unaffordability is acceptable.” It says “therefore targeting identifiable villains won’t solve the problem; systemic intervention is required.”

Third, the alternative is worse. If we permit Grievance Beliefs without evidential standards, the targets will not be powerful elites (who can defend themselves) but vulnerable populations (who cannot). The Scapegoat Mechanism’s prey-selection criteria ensure that confabulated attribution targets the vulnerable, not the powerful. The Presumption therefore protects the vulnerable more than it protects elites.

Fourth, where elites do satisfy the Three-Test Battery (where they literally and immediately block access through identifiable mechanisms they personally designed) the Presumption is rebutted, and the Practical Redirect returns the tractable action on the gate they built, with Cascading Consequences directed at the levers that built it. The Presumption does not provide blanket immunity; it requires evidential standards.

The concern is legitimate: the Presumption could be misused by powerful actors seeking to deflect accountability. The safeguard is that the Presumption applies equally to all agent-attributions, including attributions that target critics of elites. If an elite claims “protesters are threatening social order,” this attribution faces the same Presumption, and will likely fail the Three-Test Battery.

Objection 4: “The Self-Scapegoating Problem”

The chapter focuses on scapegoating others. But people also scapegoat themselves; “I am the problem,” “I deserve what’s happening to me.” Does the framework address this?

Response: Self-scapegoating operates through the same architecture with the same diagnostic tools.

When an Individual attributes their felt deficit to their own failings (“I’m experiencing career difficulty because I’m fundamentally inadequate”), this attribution faces the Three-Test Battery:

Immediacy: Am I literally and immediately blocking myself from accessing what I need right now? (Usually no; the blocking is diffuse and temporal)
Mechanism: What clear, exclusive mechanism am I using to block my own access? (Usually cannot specify beyond vague character traits)
Authorship: Did I personally design the system that produces my deficit? (Usually no; career systems, credential hierarchies, and labour markets are emergent)

Self-scapegoating typically fails all three tests, revealing it as confabulated attribution. The Information Gap that prevents accurate attribution to others also prevents accurate self-attribution. The brain’s demand for agent-based causation can be satisfied by targeting oneself as readily as by targeting others.

The Dual-Role Signature also applies: self-scapegoating often exhibits prey-selection (I am vulnerable, I am accessible, I am unable to escape myself) combined with predator-framing (I am sabotaging myself, I am responsible for my failures, I am the author of my suffering). This incoherence is diagnostic of Confabulation.

The practical redirect applies equally: Would targeting myself solve the problem? Would self-condemnation restore my access? If no, the self-attribution is non-functional regardless of its accuracy, and cognitive resources should redirect to systemic analysis (what structural factors actually produce my deficit?) and practical intervention (what would actually improve my situation?). After a fall in standing, the conditions under which the demand is met by the self are derived at Chapter 16, Section 16.12.

Objection 5: “The Institutional Scapegoating Problem”

The chapter discusses individual and population-scale scapegoating. But institutions also scapegoat; organisations blame individuals for systemic failures, bureaucracies target visible cases to demonstrate action. Does the framework address this?

Response: Institutional scapegoating operates through the same mechanisms at organisational scale.

When an institution faces collective felt deficit (declining performance, public criticism, resource constraint), the institution’s decision-makers face the same Information Gap. True causes may be systemically diffuse (accumulated organisational dysfunction, emergent market conditions, distributed policy failures). But the institution’s stakeholders demand agent-based explanation; “Who is responsible?”

The institution’s response typically follows Scapegoat Mechanism dynamics:

Selection through prey-criteria: Low-status employees, temporary workers, or recent hires who cannot effectively resist
Framing through predator-attribution: “Their poor performance/misconduct/negligence caused the problem”
Dual-Role Signature: The same Individual is too powerless to resist termination yet somehow powerful enough to have caused organisational-scale deficit

Institutional scapegoating serves organisational functions parallel to individual scapegoating:

Cognitive closure: The question “why did this happen?” is answered
Stakeholder satisfaction: External critics receive an agent to blame
System preservation: The institutional structure remains intact; only the Individual is removed
Catharsis: Collective anxiety is discharged through the targeting

The Three-Test Battery applies to institutional attributions:

Did this Individual literally and immediately cause the deficit through actions occurring right now?
What clear, exclusive mechanism did this Individual use: a mechanism only they controlled?
Did this Individual personally design the system that produced the deficit?

When institutions fail these tests (as they typically do), the attribution is revealed as confabulated, and attention should redirect to systemic factors: What organisational structures enabled the problem? What policies would prevent recurrence? What institutional redesign is required?

Objection 6: “The Partial Causation Problem”

Scapegoats need not be entirely innocent. What if the target actually contributed to the problem, just not to the extent attributed? The Dual-Role Signature might still be present even when the target bears some genuine responsibility.

Response: The Presumption Against Scapegoating does not require that targets be entirely innocent.

The Presumption Against Scapegoating does not require that targets be entirely innocent. It addresses proportionality and causation: the Scapegoat bears blame vastly exceeding their actual causal contribution to the deficit, and often bears blame for systemic dynamics they did not create and cannot control.

An Individual may be a node in a causal network while still being scapegoated. Consider:

A bank teller who followed standard procedures may be blamed for a fraud they did not design
A low-level soldier who followed orders may be blamed for atrocities planned by high command
An immigrant worker who accepts available employment may be blamed for wage depression caused by labour market structure

In each case, the target is not entirely disconnected from the causal chain. But the attribution assigns them primary causal responsibility when their actual contribution is minor, conditional on larger structural factors, and not unique to them.

The Three-Test Battery accommodates this:

Immediacy: Their contribution may not be immediate even if it exists
Mechanism: The mechanism may not be exclusive to them: others participate in the same causal network
Authorship: They did not design the system; they merely participate in it

The Dual-Role Signature remains diagnostic: if they are selected through prey-criteria while being framed as primary cause, the attribution is disproportionate regardless of partial contribution.

The practical redirect addresses partial contribution: “Suppose they do bear some responsibility. Would removing them solve the problem? If others would immediately fill the same role in the same system, targeting this Individual is addressing a symptom, not a cause.”

Chapter Closing

Establishment Summary

This chapter has established:

Confabulation as the routine, architecturally-necessitated generation of inaccurate causal narratives when Individuals attempt to explain M-driven desires under Source Opacity and Vocabulary Deficiency conditions
Projection as Confabulation applied to others: systematic misattribution of M-driven behaviour to A or I sources due to Template constraints
The Scapegoat Mechanism as the collective pathway through which individual Confabulations propagate via prestige-bias and social proof, converging on Grievance Beliefs that identify targets selected through prey-criteria while framed through predator-attribution
The Presumption Against Scapegoating with its Three-Test Battery (Immediacy, Mechanism, Authorship) as the epistemic default required by the architecture
Dual-Role Confabulation as the diagnostic criterion that reveals confabulated scapegoating regardless of specific content
The Non-Malice Corollary reframing conflict as typically arising from architectural Confabulation rather than deliberate malice
The Moral Vocabulary Displacement Principle establishing that moral framing functions as permission infrastructure for targeting rather than independent ethical inquiry

Axiomatic Recap

This chapter’s conclusions are grounded in:

Axioms 2+3 (Source Opacity): The destruction of source-tags during Common-Currency Integration
Chapter 9 (Timing Architecture): Preconscious Transmission preceding conscious awareness
Axiom 9 (Agent-Based Causal Default): Human causal cognition architecturally optimised for agent-based explanations
Axiom 9 + Axiom 4 (Information Gap): True systemic causes exceeding computational tractability

Forward Connection

Chapter 13 (Bayesian Belief Dynamics) will formalise how Confabulations compound through iterative updating, producing the Confabulation Cascade at population scale and establishing the mathematical symmetry that makes crisis architecture identical to prevention architecture.

Chapter 18 (The Knowledge Patch) will operationalise the interventions sketched here (Epistemic Suspension, the Two-Question Diagnostic, and the Presumption Against Scapegoating) as the vocabulary dissemination that exploits the Bayesian symmetry to maintain subcriticality.

The order-effects-from-Bayes derivation that propagates the Confabulation Cascade at the population level is established in Chapter 14 (Quantum Epistemology), Section 14.8.

What Would Falsify This?

The claims in this chapter would require significant revision if:

Introspective accuracy for M-attribution: If Individuals could reliably identify Mimetically-acquired desires through ordinary introspection without AIM vocabulary (demonstrated by high correlation between self-reported M-attribution and behavioural signatures of Mimetic origin (audience-sensitivity, Model-dependence, non-satiation)) then the Confabulation architecture would be disconfirmed. The standard for reliability is the attribution-shuffling benchmark used across the intervention chapters: self-reported attributions must exceed the accuracy achievable when attributions are shuffled within the same case set, that is, they must carry information about the true cause rather than matching it at the rate independent generation would produce (Chapter 18, What Would Falsify This?).
Self-correcting Confabulation Cascades: If populations that converge on Grievance Beliefs spontaneously revise those Beliefs when disconfirming evidence emerges, without external intervention providing alternative explanatory frameworks, then the Confabulation Cascade’s resistance to self-correction would be disconfirmed.
Scapegoat selection without prey-criteria: If historical analysis of scapegoating episodes reveals targets selected through predator-criteria (capability, threat potential, retaliatory capacity) rather than prey-criteria (vulnerability, visibility, reciprocity-exemption), then the Dual-Role Confabulation signature would be disconfirmed.
Moral vocabulary independent of targeting permission: If moral vocabulary in scapegoat contexts functions to restrain targeting rather than to permit it (demonstrated by higher rates of de-escalation when moral framing is introduced) then the Moral Vocabulary Displacement Principle would be disconfirmed.

These criteria are not already established by the phenomena this chapter explains; they remain open empirical questions.

[DIAGRAM PLACEHOLDER: The Confabulation-to-Scapegoat Pathway] Illustrates the progression from individual Information Gap through Confabulation, Projection, prestige-biased transmission, social proof, and collective convergence to Scapegoat Mechanism activation.

Diagram should show:

Individual level: Information Gap → Confabulation → Self-Model incorporation
Social level: Expression → Prestige-biased transmission → Social proof
Collective level: Confabulation Cascade → Agent-based completion → Scapegoat selection
Diagnostic overlay: Three-Test Battery and Dual-Role Signature as intervention points

[DIAGRAM PLACEHOLDER: The Three-Test Battery Decision Tree] Illustrates the sequential application of Immediacy, Mechanism, and Authorship tests with outcomes at each node.

Diagram should show:

Entry: Grievance Belief detected
Q1 (Immediacy): YES → proceed to Q2; NO → Presumption holds, redirect to systemic analysis
Q2 (Mechanism): YES → proceed to Q3; NO → Presumption holds, mechanism gap reveals Information Gap
Q3 (Authorship): YES → Presumption rebutted; Practical Redirect returns the tractable action on the gate; NO → apply Anti-Conflation Safeguard
Q3b (Anti-Conflation): Benefit ≠ Authorship acknowledged → redirect to systemic analysis

[DIAGRAM PLACEHOLDER: The Diagnostic Signature of Dual-Role Confabulation] Illustrates the structural incoherence between prey-selection criteria and predator-framing.

Diagram should show:

Left column: Selection criteria (vulnerability, visibility, reciprocity-exemption) → Prey-detection systems
Right column: Framing criteria (threatening, taking, blocking) → Predator-attribution systems
Center: Scapegoat as intersection point with visible incoherence
Diagnostic question: “If powerful enough to cause this harm, why targetable without overwhelming retaliation?”

References

Bortolotti, L. (2018). Stranger than fiction: Costs and benefits of everyday confabulation. Review of Philosophy and Psychology, 9(2), 227–249.

Des Forges, A. (1999). Leave None to Tell the Story: Genocide in Rwanda. Human Rights Watch.

Kopelman, M. D. (1987). Two types of confabulation. Journal of Neurology, Neurosurgery, and Psychiatry, 50(11), 1482–1487.

Kuhn, T. S. (1962). The Structure of Scientific Revolutions. University of Chicago Press.

Oosterhof, N. N., & Todorov, A. (2008). The functional basis of face evaluation. Proceedings of the National Academy of Sciences, 105(32), 11087-11092.

Rule, N. O., Adams, R. B., & Ambady, N. (2012). Perceptions of dominance following glimpses of faces and bodies. Perception, 41(6), 687-706.

Straus, S. (2006). The Order of Genocide: Race, Power, and War in Rwanda. Cornell University Press.

Whewell, W. (1840). The Philosophy of the Inductive Sciences, Founded Upon Their History. John W. Parker.

Willis, J., & Todorov, A. (2006). First impressions: Making up your mind after a 100-ms exposure to a face. Psychological Science, 17(7), 592-598.

CONFABULATION, PROJECTION, AND THE SCAPEGOAT MECHANISM

Position Statement

Dependency Statement

Derivation Preview

Chapter Preview

Empirical Foundations: Confabulation in Psychology and Neuroscience

Nisbett & Wilson: Telling More Than We Can Know

Gazzaniga: The Left-Brain Interpreter

Confabulation as Routine Cognition

Self-Perception Theory: Inferring Internal States from Behaviour

The Introspection Illusion

Dual-Process Theory: System 2 Explains System 1

Haidt: The Emotional Dog and Its Rational Tail

Mercier and Sperber: The Argumentative Theory of Reasoning

How AIM Extends the Literature

MOVEMENT A: INDIVIDUAL CONFABULATION

11.1 The Information Gap and Vocabulary Deficiency

The Vocabulary Deficiency

The Vocabulary Gap Is M-Specific

Cross-Cultural Evidence for the Vocabulary Gap

The Critical Implication

11.2 The Confabulation Rationality Lemma

11.2.1 The Confabulation Proliferation Corollary

11.2.2 The Collective Instance

11.2.3 The Exception-Handling Analogy

11.3 I-Override as Confabulation Site

The Process of Confabulation

The Paradox of Deliberation

Why I-Override Fails Without AIM Vocabulary

Confabulation Is Definitionally Inaccurate

11.3.1 I-Override Within Conscious Cognition

11.4 Epistemic Suspension as Alternative

Why “I Don’t Know” Was Once Impossible

Why “I Don’t Know” Is Now Possible

How “I Don’t Know” Functions as Closure

Epistemic Suspension as Interference-Term Maintenance

The Dispositional Challenge

The Functional Equivalent Through Trained Tolerance

A Training Protocol Sketch

The Difficulty Gradient

AIM Vocabulary as Compensatory Tool

Three Barriers to Widespread Adoption

11.5 Confabulation in Social Interaction

11.6 The Intractability Confabulation

MOVEMENT B: PROJECTION

11.7 Projection Through Template Availability

The Coupled-Loop Mechanism

The AIM Definition of Projection

Rejecting the Freudian Account

The Template-Availability Account

The Neural Basis of Template Deployment

Why the M-Template Is Specifically Absent

The Systematicity of Projection

11.8 Template Inversion: Explaining Different Others

Contrast Categories

The Double Standard

11.8.1 The Self-Model Moral Boundary Principle

11.9 Guilt-Priming and Elevated Template Activation

Guilt-Priming Mechanism

Why Guilt-Priming Is Partial Rather Than Complete

The Emotional Fingerprint

Projection as Social Contagion Vector

MOVEMENT C: COLLECTIVE PATHWAY

11.10 From Individual Confabulation to Observable Expression

The Observability Threshold

Why Expression Is the Default

Debate as Institutional Confabulation-Laundering

11.11 Propagation Through Prestige-Bias and Social Proof

Prestige-Biased Transmission

Social Proof

The Cascade Does Not Self-Correct

11.12 The Scapegoat Mechanism as Collective Completion

The Scapegoat Mechanism, A Coupled-Loop Derivation

Step one, continuous M-assessment of Individuals as Objects

Step two, iterated fast-path Confabulation Cascades

Step three, convergent Confabulated attribution under collapsing differentiation

Step four, collective discharge and loop reset

The derivation's structural consequences

The Collective Information Gap

The Scapegoat as Highest-Likelihood Completion