The Architecture of Agency Volume 8 Exemplars and the Axionic Criterion

Exemplars and the Axionic Criterion

How patterns travel, and which deserve to

This chapter is a review — it is readable but still changing.

A child does not learn to stand by consulting a rule. She watches, sways, falls, adjusts, and tries again, guided by the moving example of another body already balanced. No proposition passes between them. No one could state the proposition if asked. What passes is a pattern — a configuration of posture and correction that the child’s nervous system absorbs, rehearses, and eventually owns.

That process never retires. We like to imagine that adult competence graduates from imitation to calculation, that somewhere in adolescence we trade watching for reasoning. We do not. Adult agency continues to run on absorbed patterns rather than explicit computation. The surgeon’s hands, the negotiator’s timing, the parent’s judgment about when to intervene and when to let a child struggle — none of these were learned from manuals, and none could be fully written down. They were caught, not taught. Skills, norms, and values propagate through lived examples that can be perceived, compressed, and recomposed, which means cultural inheritance proceeds through embodied behavior at least as much as through instruction.

This is one mechanism through which developed agency acquires competence: internalizing action-patterns and deploying them flexibly under changing conditions. It is not a new definition of agency; the embedded model–evaluation–control loop remains the canonical core. Patterns as Players made the case that analyst-specified pattern types are useful units for tracking cultural selection, while people remain the choosing and responsible agents through whom transmission occurs. This chapter is about that transmission channel — how patterns move from body to body — and about the normative question the volume has been building toward: of all the patterns competing to move through us, which ones deserve to?

Exemplars as Behavioral Attractors

Open-ended environments demand judgment under variation. No rulebook anticipates every situation; the situations that matter most are precisely the ones the rulebook missed. What stabilizes action under novelty is not a longer rulebook but a reference pattern — a worked example of how someone handles the class of problem well.

An exemplar is an agent whose behavior functions as an attractor in behavioral space. Where Patterns as Players distinguished the three scales of attraction this volume tracks, this is the smallest and most intimate: not a Schelling point that coordinates strangers, not a black hole that captures a worldview, but a lived pattern that orients individual action. Exemplars shape what an agent notices, what she expects, what possibilities appear open to her. Around a genuine exemplar, the space of things one might do acquires a gradient.

An apprentice watches a master carpenter cut a joint. The lesson does not arrive as a proposition. It arrives as posture, timing, pressure, restraint. The pattern embeds itself and later re-emerges through the apprentice’s hands — not as a memory of the master, but as a way of being a carpenter. This is the ordinary miracle on which every craft, every profession, and every moral tradition depends.

An exemplar’s influence scales with three properties of the associated pattern. Coherence: the pattern hangs together — its parts reinforce rather than contradict one another, so that adopting one piece pulls the others along. Generality: it transfers — the carpenter’s restraint is legible to the surgeon, the master’s patience to the parent. Compressibility: it can be carried in a small package — a story, a maxim, a remembered gesture — and unpacked at the point of need. A pattern strong on all three axes can leap across centuries and cultures on the strength of a parable. A pattern weak on all three dies with its first host.

And exemplars persist through reinterpretation, not despite it. Agents selectively adopt, modify, and recombine aspects of a reference pattern as their conditions change. Over time this is evolutionary selection operating on action-patterns: variation in each new instantiation, selection by what continues to work, retention in the bodies and stories of those who carry it forward.

Lineages and the River

Cultural continuity is the persistence of shared reference patterns across generations. Lineages form around exemplars whose patterns remain legible enough to perceive, teachable enough to transmit, and adaptable enough to survive contact with new circumstances.

Religious figures are history’s highest-bandwidth exemplars. Their narratives encode dense bundles of behavior, motivation, and constraint — not a list of rules but a personality one can consult, a compressed answer to the question what would he do? These bundles propagate through repetition, reinterpretation, and partial instantiation across enormous populations and spans of time. No other transmission technology in the historical record comes close.

The Story-Telling Ape described what this looks like from the outside: legend colonizes fact, the historical kernel disappears under the accumulating bailey of story, and the Jesus of history becomes unrecoverable beneath the Jesus of faith. That account is true, and it is the descriptive half of the picture. But it invites a mistaken inference — that because the legend outran the man, the pattern is therefore fraudulent, a mere artifact of ravenous imagination. The functional question is separate from the historical one. Whether the kernel survives excavation tells us nothing about whether the pattern works: whether the bundle of behaviors the tradition transmits — the forgiveness under provocation, the dignity extended downward, the refusal to return violence — supports the agency of those who instantiate it. Mythogenesis explains how the pattern got its vehicle. It does not evaluate the cargo. That evaluation needs a different instrument, and this chapter builds it.

A pattern that survives centuries does so the way a river survives: by flowing, branching, eroding obstacles, and reshaping itself while retaining a recognizable course. The Danube is not the same water, the same banks, or even quite the same route it was a thousand years ago, yet it is unmistakably the same river. So with a living tradition. Persistence of this kind is not inertia. It records successful transmission under changing conditions, but not one verdict on why: agents may find a pattern useful or worth its cost, while family inheritance, institutional support, coercion, path dependence, and chance also keep traditions flowing.

Reading “Lord” as Pattern Centrality

Within a lineage, compressed phrases do diagnostic work. When a believer says “Jesus is Lord,” a metaphysical reading takes this as a claim about the governance of the cosmos, and the argument immediately becomes one about evidence for supernatural facts — an argument that has been stuck for centuries because the two sides are not discussing the same kind of proposition.

Read functionally, the phrase is a diagnostic of pattern centrality. “Lord” designates salience and gravitational pull within a shared pattern space. The phrase marks a convergence point: it announces that this behavioral pattern serves as the common reference against which action gets evaluated within this lineage — the exemplar of exemplars, the pattern to consult when patterns conflict. That centrality arose through adoption, repetition, reinterpretation, institutions, and inheritance across two thousand years of agents. Its continuation demonstrates durable transmission, not that every carrier found the pattern agency-supporting or freely chose it. Whether it now supports agency is a separate question for the Criterion this chapter is about.

This reading is neither endorsement nor debunking. It does not require the metaphysics, and it does not require the sneer. It takes the phrase seriously as a report about the structure of a cultural pattern space — which is, I think, more respect than either the apologist or the village atheist typically manages.

Coordination Through Attraction

Notice what exemplars do not do as exemplars. They do not command or enforce; institutions built around them may. Exemplars constrain behavior through attraction: agents orient toward patterns that appear to support stable coordination, reduce destructive conflict, and preserve their agency across the situations they actually face. Where participation remains interpretive and revisable, nobody instantiates the carpenter or the Christ by rote; each agent can take what fits, adapt what almost fits, and drop what does not. Attraction and usefulness can sustain a pattern, but so can authority, dependency, and inherited default.

This yields something remarkable: coordination without an enforcer. When many agents orbit overlapping reference patterns, their expectations about one another align — not perfectly, but enough. The handshake, the queue, the honest weight, the kept promise: these run on shared patterns, not on surveillance. Enforcement exists, but as a backstop for the margins, not as the engine. A society that had to police every norm it depends on would spend everything on police. The rhyme with the political argument of Volume 7 is not accidental — the conversion of coercion makes the same case at the level of institutions that this chapter makes at the level of culture: orders that run on attraction are cheaper, more adaptive, and more durable than orders that run on force.

Plurality of exemplars is a further, underrated asset. Different domains sustain different reference patterns — the scientist’s exemplars are not the soldier’s, the mother’s are not the merchant’s — with partial overlap providing resilience under environmental change. When one lineage degrades or meets conditions it cannot handle, agents can shift weight to another without the whole coordination fabric tearing. A culture that enforces a single exemplar everywhere has built a pattern monoculture, and monocultures fail the way monocultures always fail: all at once, to the first pathogen adapted to them. What that looks like in practice — communities sealed around one pattern, treating every alternative as contamination — is the subject of Gated Realities.

Distributed Instantiation

Where does a cultural pattern live? Not in any text, though texts carry compressed versions of it. Not in any institution, though institutions scaffold it. It lives in its instantiations: collective participation takes the form of distributed instantiation of a pattern across agents, each realizing it under local constraints, each contributing marginally to its evolution.

This process need not have a headquarters. There is no committee that maintains carpentry, and no single office owns every instantiation of the Sermon on the Mount. Pattern survival depends on continued retention or replication under selective pressure. As environments shift, instantiations diverge where local conditions differ, converge where a variant propagates better under the operative filters, and attenuate where transmission fails. The same dynamic can govern cultural norms, ethical frameworks, and institutional practices — anything transmitted by example, inheritance, enforcement, or sustained use.

Distributed instantiation explains both the resilience and the drift of traditions. Resilience, because a pattern spread across a million bodies cannot be killed by destroying any one of them, or any thousand. Drift, because a million local reinterpretations under a million local constraints will carry the pattern somewhere its founders would not recognize. The river again: no single obstacle stops it, and no stretch of it holds still.

The Axionic Criterion

Everything so far has been descriptive: how patterns transmit, why some persist. The question the whole volume needs answered is normative: which patterns deserve to propagate? Selection alone cannot answer it — Patterns as Players established that the environment rewards performance, not virtue, and history is full of patterns that spread magnificently while hollowing out everyone who carried them.

Here is the test. For a pattern to merit propagation through a population of agents, it must demonstrate agency preservation under ongoing instantiation. Concretely, three questions:

  1. Does the pattern support coherent choice under constraint? Agents who instantiate it should become more capable of judgment in hard situations, not less — the pattern should function like the carpenter’s absorbed skill, extending what its carriers can do.
  2. Does it reduce reliance on coercive stabilization? A pattern that can only persist by punishing defectors, silencing critics, or walling off alternatives is confessing that it cannot hold agents by attraction. The police budget of a pattern is a measure of its failure.
  3. Does it scale across agents without collapsing autonomy? Some patterns work beautifully at small scale and become machines for obedience at large scale. The test is whether the millionth instantiation is still a choosing agent or has become a component.

Call this the Axionic Criterion. It is the volume’s instrument, and the chapters that follow will apply it repeatedly — to ideologies, to institutions, to the weaponized discourse patterns of Part III and the capture systems of Part IV.

Two clarifications about its standing. First, the Criterion is not a smuggled objective value. Volume 5 argued that value has no agent-independent existence — Phosphorism locates all valuing in valuers — and the Criterion respects that: it does not rank patterns against a cosmic good. Agency is not one value among others; it is the precondition of anything being valued at all. But that is a descriptive fact, and a descriptive fact compels no one’s rankings by itself. What the Criterion adds is a choice: this book adopts reciprocal agency preservation as an explicit commitment, and then asks whether a pattern preserves the very capacity that makes valuing possible. A pattern that consumes the agency of its hosts dissolves the ground on which those hosts could call anything good — self-undermining in the same way, and for the same reason, as the ethical systems the viability criterion rules out in Volume 5, rather than wrong by cosmic decree.

Second, the Criterion is not administered by a tribunal. The evaluation occurs locally and continuously, in the lives of instantiating agents. Every time someone finds that a tradition helps them choose well under pressure, the pattern passes a trial for that agent under those conditions; every time someone notices that their framework survives in them only through fear of leaving it, the pattern fails one. Passing can support voluntary propagation, and failure can motivate disengagement, reinterpretation, or abandonment, but neither outcome mechanically determines cultural survival. When The Geometry of Culture maps the curved landscape of cultural evolution and confronts the relativism built into it — no neutral vantage, no fixed fitness function — agency preservation will emerge as the leading candidate for a cross-cultural standard: a measure that stays legible across frames because every frame needs agents to hold it, and one this book chooses to privilege rather than reads off the terrain. Selection will not enforce it for us — coercion and lock-in can keep agency-consuming patterns alive — which is why the standard has to be applied, not awaited. The Criterion is that standard made operational: three questions any agent, or any accountable institution, can ask about a pattern reaching for their hands.

Civilization is not run from anywhere. It is a fabric of shared action-patterns instantiated by billions of embedded agents, coordinated by attraction to exemplars whose influence rests on coherence, compressibility, and sustained performance. The patterns worth keeping are the ones that carry agency forward — like a river shaping its course, like a craft refined through generations of hands.