The Architecture of Work: A Three-Layer Structural Model

Executive Summary

Every productive activity in the economy, however specialized its vocabulary or elaborate its tooling, can be described at three levels of abstraction. At the deepest level sit a small, closed set of primitives: the fundamental reasoning operations the human mind performs, crossed with the fundamental kinds of substrate those operations act upon. At the middle level, certain combinations of reasoning and substrate recur so frequently across the economy that they constitute recognizable, nameable domain patterns — stable “shapes” of activity such as Evidentiary Adjudication, Synthesis and Design, or Diagnosis and Repair. At the surface level, each domain pattern is instantiated thousands of times over as sector-specific activities, each dressed in local terminology, local regulation, and local instruments, yet structurally identical to its siblings in other industries.

This paper formalizes that three-layer model. It defines the primitive axes, names and characterizes thirteen domain patterns, and enumerates for each pattern an extended catalog of real-world manifestations. For every manifestation, the paper identifies the same anatomical elements: the inputs the activity consumes, the criteria or methods that govern it, the working record it produces, the terminal output it delivers, and the degree to which its reasoning is conducted upon representations as opposed to requiring direct engagement with physical material or live human beings.

The model’s value is diagnostic and comparative. Once any activity is decomposed into its reasoning mode, its substrate, and its instruments, it can be compared rigorously with structurally similar activities in unrelated sectors. Practices, tools, and process designs proven in one manifestation of a pattern transfer readily to others, because the underlying structure is the same. The model also makes visible several properties that vary systematically across patterns — the explicitness of governing criteria, the observability of the substrate, the latency of feedback, the cost of error, and the physical embodiment of the work — and these properties turn out to predict a great deal about how each pattern’s activities can be organized, measured, divided, delegated, and re-instrumented.

1. Introduction and Motivation

Consider a nineteenth-century bank examiner at work. He sits in a side office of the bank under examination. To his left are the ledgers the bank’s clerks have carried in — the evidence. Before him is a checklist of things to verify, each with criteria for what constitutes a satisfactory finding — the rules. Under his hand is a digest in which he records observations, reasoning, conclusions, and recommendations — the working record. And at the end of the engagement he produces a report summarizing the whole — the judgment.

Modern bank supervision is vastly more elaborate: data extracts replace hand-carried ledgers, layered statutes and guidance replace the printed checklist, structured workpapers replace the digest, and a formal Report of Examination replaces the summary letter. Yet the shape of the activity is unchanged. Evidence flows in; codified criteria are applied; reasoning is documented; an authoritative conclusion flows out.

Once this shape is seen clearly, it is seen everywhere: in the auditor’s opinion, the pathologist’s report, the building inspector’s certificate, the judge’s ruling, the patent examiner’s office action. These activities share no industry, no regulator, no professional society — yet they are the same activity in a structural sense. They differ only in vocabulary, substrate, and instruments.

That observation raises the natural question: if Evidentiary Adjudication is one recurring shape, what are the others? How many are there? What are they made of? And what organizing framework relates a small number of universal shapes to the effectively unlimited variety of jobs, tasks, and institutional functions observed in the real economy?

This paper answers those questions with a three-layer model:

• Layer 1 — Primitives. A closed two-axis grid: fundamental reasoning modes crossed with fundamental substrates. This layer is universal and essentially fixed; new cells will rarely, if ever, appear.
• Layer 2 — Domain Patterns. Named, recurring combinations of reasoning and substrate — the economically common cells of the grid, of which this paper identifies thirteen. This layer is finite and stable over long horizons.
• Layer 3 — Sector Activities. The open-ended set of concrete instantiations of each pattern across industries. This layer grows and mutates constantly, but always by re-instantiating Layer 2 patterns with local materials.

The remainder of the paper proceeds top-down through the layers, then bottom-up through worked examples, and closes with a comparative analysis of the properties that distinguish the patterns from one another.

2. Layer 1: The Primitives

2.1 Axis A — Reasoning Modes

A reasoning mode is the characteristic cognitive operation an activity performs. Seven are sufficient to span the model; the seventh is a wrapper rather than a peer.

A1. Deduction (Subsumption). The classification of particular facts under general, pre-existing criteria. The rule exists before the case arrives; the work is determining which rule the facts satisfy and what conclusion the rule therefore compels. Deduction is the most explicit of the modes: its criteria can, in principle, be written down completely, and the quality of a deductive conclusion can be checked by anyone with access to the same facts and the same rules. Its characteristic failure is misclassification — applying the right rule to the wrong facts or the wrong rule to the right facts.

A2. Abduction (Inference to Best Explanation). The generation of hypotheses about hidden causes that would explain observed effects, followed by differential elimination among candidates. Unlike deduction, there is no rulebook that mechanically yields the answer; there is instead a space of possible explanations, a body of knowledge about which causes produce which symptoms, and a discipline of testing that progressively excludes candidates. Its characteristic instrument is the probe — any test that exposes otherwise hidden internal state. Its characteristic failure is premature closure: anchoring on an early hypothesis and ceasing to test.

A3. Constructive Search (Synthesis). The generation of a form that does not yet exist, satisfying a set of requirements and constraints simultaneously. Where deduction judges an existing thing against criteria, constructive search creates a thing to meet criteria. The work proceeds through cheap intermediate representations — sketches, models, drafts, prototypes — that allow candidate forms to be tested and discarded at low cost before commitment. Its characteristic failure is constraint violation discovered late, when the cost of revision has grown large.

A4. Procedural Execution. The faithful performance of a proven method within specified tolerances, with continuous monitoring for deviation. The intelligence in procedural execution is not in choosing what to do — the method is given — but in doing it consistently, detecting drift from tolerance, and knowing when a deviation is large enough to halt and escalate. Its characteristic artifacts are the process specification and the run record. Its characteristic failure is undetected drift.

A5. Optimization and Allocation. The arrangement of scarce resources — time, capacity, routes, attention, funds — across competing demands under constraint, seeking the best achievable configuration. Optimization presupposes that the options and constraints can be represented; its work is search over that representation, and its output is a plan, schedule, route, or allocation. Its characteristic failure is optimizing a stale representation of a changed reality.

A6. Theory-of-Mind. The modeling of other minds — their knowledge, preferences, intentions, and emotional states — in order to teach, persuade, comfort, negotiate with, perform for, or anticipate them. This is the least externalizable of the modes: its “rules” are heavily tacit, its evidence is behavioral and contextual, and its feedback arrives through the very relationship it operates on. It divides into cooperative applications (the counterparty wants the interaction to succeed) and adversarial applications (the counterparty is attempting to defeat the reasoner).

A7. Vigilance (a wrapper mode). A continuous monitoring loop — observe, compare to baseline, act on deviation — wrapped around any of the other modes. Vigilance is what converts a one-time judgment into stewardship, a one-time diagnosis into monitoring, a one-time plan into control. Its defining property is that value is produced by events prevented rather than outputs created, which makes its performance inherently difficult to observe.

2.2 Axis B — Substrates

A substrate is the kind of thing the reasoning operates upon. Six are sufficient.

B1. Records — representations of past states of affairs: ledgers, files, filings, images, specimens, testimony, logs, datasets. Records are inert, reproducible, and transportable; reasoning over them can occur anywhere the records can travel, at any time, by any number of parties in parallel.

B2. Matter — physical material and energy, including data treated as raw material to be transformed rather than interpreted. Matter obeys physics: acting on it requires presence (or actuated presence), consumes energy, and is rate-limited by physical process.

B3. Opaque Systems — functioning wholes with hidden internal state: machines, bodies, organizations, software in production, ecosystems. Their surfaces show symptoms; their interiors must be inferred or probed.

B4. The Unknown — structure not yet discovered by anyone: undescribed phenomena, unexplored territory, unanswered questions. Distinct from an opaque system in that no one possesses the answer; there is no repair manual to consult even in principle.

B5. Cooperative Minds — people engaged willingly: students, patients, customers, counterparties, audiences, colleagues. They have interests and interiority, but the interaction is fundamentally collaborative.

B6. Adversarial Minds — intelligences actively attempting to defeat the reasoner: attackers, fraudsters, opposing forces, evasive counterparties. The substrate adapts against whatever method is used on it, which forecloses purely static approaches.

2.3 Properties of the Grid

Crossing seven modes with six substrates yields a grid of possible activity-types. Not all cells are economically meaningful (procedural execution upon the unknown is nearly a contradiction), and the meaningful cells are not equally populated. Five analytical properties vary across the grid and will recur throughout this paper, because they govern how each cell’s activities can be organized, staffed, measured, and instrumented:

1. Codification — how completely the governing rules or methods can be written down. Highest for deduction and procedural execution; lowest for theory-of-mind.
2. Representational sufficiency — whether the reasoning can be conducted entirely upon representations (records, models, data) or requires direct engagement with matter or live persons. Highest for substrate B1; lowest for B2 and B5.
3. Feedback latency — how quickly the quality of the reasoning is revealed. Seconds in procedural execution; years in stewardship and formation.
4. Error cost and reversibility — the consequence of a wrong conclusion and whether it can be corrected downstream.
5. Adversarial pressure — whether the substrate adapts to defeat the method, forcing continuous method renewal.

These five properties are the model’s connective tissue: activities that share a cell share a property profile, and property profiles — not industry labels — determine which organizational designs, instruments, and quality regimes fit an activity.

3. Layer 2: The Thirteen Domain Patterns

A domain pattern is a named, recurring cell (or small cluster of cells) of the Layer 1 grid — a shape of activity that appears again and again across unrelated sectors because the underlying combination of reasoning and substrate is economically ubiquitous. Thirteen patterns cover the great majority of productive activity. For each, this section gives the grid placement, the canonical information flow, and the characteristic anatomy; Section 4 then enumerates real-world manifestations in depth.

Three structural remarks before the catalog.

Patterns are activities, not roles. A single role is almost always a bundle: the physician adjudicates (rendering a diagnosis against diagnostic criteria), then repairs (treating), then cares (attending to the patient as a person). The model classifies the activities within the role, and a role is analyzed as a weighted portfolio of patterns.

Each pattern separates a reasoning layer from an interface layer. In every pattern one can distinguish the cognitive core — the subsumption, the hypothesis search, the constraint solving, the scheduling — from the interface through which that core touches its substrate: the hands that carry ledgers, the probe that draws the blood sample, the vehicle that moves the freight, the voice that delivers the lesson. The two layers have very different property profiles. The reasoning layer operates on representations and inherits the portability of substrate B1; the interface layer inherits the physics of B2 or the relational demands of B5. Much of the historical evolution of every pattern consists of thickening the representational layer — more of the substrate captured as records, models, and telemetry — so that more of the reasoning can occur away from the substrate itself.

Each pattern carries a characteristic working record. The examiner’s digest, the engineer’s design file, the production run log, the diagnostic chart, the maintenance history, the shipment manifest, the lab notebook, the lesson plan and gradebook, the deal file, the project plan, the care plan, the score and script, the incident log. The working record is where the pattern’s reasoning is externalized — where the chain from input to conclusion is made inspectable, auditable, and transferable. Patterns differ enormously in how complete this externalization is, and that completeness is among the most consequential properties an activity possesses.

4. Layer 3: Manifestations of Each Pattern

This section enumerates real-world manifestations of each pattern. For each manifestation the same anatomy is identified: inputs, governing criteria or method, working record, terminal output, and the balance between reasoning-on-representations and direct engagement with the substrate.

4.1 Evidentiary Adjudication (Evidence → Rules → Reasoned Record → Judgment)

The archetype: a trained examiner applies codified criteria to gathered evidence, documents the reasoning, and renders an authoritative conclusion. Distinguishing marks: the criteria pre-exist the case; the evidence arrives as records or specimens (substrate B1); the conclusion is consequential and attributable; and the reasoning is expected to be reconstructible from the record. Because the entire activity operates on representations under explicit rules, it possesses the highest codification and representational sufficiency of any pattern.

1. Bank examination. Evidence: loan files, core-system extracts, board minutes, policies, call reports. Criteria: statute, regulation, supervisory guidance, rating frameworks. Record: workpapers linking each finding to evidence and criterion. Output: Report of Examination, ratings, supervisory actions. The evidence base is already substantially digital; the criteria are published; the workpaper discipline makes the reasoning chain fully explicit.

2. Financial statement audit. Evidence: the books, supporting documentation, confirmations, samples. Criteria: accounting standards and auditing standards. Record: audit workpapers tying each assertion to procedures performed and evidence obtained. Output: the audit opinion. Materiality thresholds and sampling rules render most of the judgment structure numeric and rule-expressible.

3. Tax examination. Evidence: returns, schedules, substantiating records. Criteria: the tax code and its regulations — among the most exhaustively codified rule systems in existence. Record: the revenue agent’s report. Output: assessment, refund, or no-change letter. Nearly the whole activity is document-in, document-out.

4. Judicial adjudication. Evidence: the admitted record — testimony, exhibits, stipulations. Criteria: statute, precedent, procedural rules. Record: the trial record and the court’s written opinion, which is explicitly a reasoned justification. Output: verdict, ruling, sentence. Fact-finding involves credibility assessment of live witnesses (a theory-of-mind intrusion), but the application of law to found facts is pure subsumption.

5. Patent examination. Evidence: the application and the prior art corpus. Criteria: novelty, non-obviousness, enablement, written description. Record: office actions and the prosecution history. Output: grant or final rejection. The prior-art search over a vast document corpus is the heart of the work; the entire exchange is conducted in writing.

6. Insurance claims adjudication. Evidence: the claim file — photographs, estimates, statements, policy documents. Criteria: policy terms, coverage law, valuation guides. Record: the adjuster’s file notes. Output: coverage decision and settlement amount. High-volume, low-severity claims are governed almost entirely by lookup-grade rules; complex claims add investigation (an abductive element).

7. Clinical diagnosis rendered against formal criteria. Evidence: history, examination findings, laboratory values, imaging. Criteria: published diagnostic criteria, staging systems, classification manuals. Record: the chart note, which documents findings and the reasoning to the diagnosis. Output: the diagnosis and its downstream entitlements (treatment authorization, disability status). The evidence-gathering interface is physical; the subsumption step operates entirely on the recorded findings.

8. Pathology and laboratory medicine. Evidence: the specimen, rendered into slides, images, and quantitative assay values. Criteria: histological classification systems, reference ranges. Record: the pathology report. Output: the tissue diagnosis. Once the specimen is converted to image and number, the reasoning proceeds wholly on representations.

9. Building and fire inspection. Evidence: the structure as observed, plans, prior permits. Criteria: building and fire codes — enumerated, versioned, published. Record: the inspection report with itemized deficiencies. Output: certificate of occupancy, correction order. The observation interface is physical presence; each individual check is binary and code-referenced.

10. Grading, credentialing, and admission decisions. Evidence: examinations, portfolios, transcripts, applications. Criteria: rubrics, cut scores, published requirements. Record: marked work, score sheets, committee notes. Output: the grade, the credential, the admission decision. Rubric-governed assessment is subsumption; holistic review mixes in tacit judgment, and institutions vary in how far they codify it.

11. Regulatory product approval. Evidence: submitted trial data, manufacturing information, labeling. Criteria: statutory approval standards and published guidance. Record: the reviewers’ memoranda — often thousands of pages of documented reasoning. Output: approval, complete-response letter, or denial. Entirely a documents-and-data activity conducted at scale.

12. Immigration, benefits, and licensing adjudication. Evidence: petitions, supporting documents, background checks. Criteria: eligibility rules in statute and regulation. Record: the decision file. Output: approval or denial with stated grounds. Among the highest-volume adjudicative activities in government, with rule systems designed for consistent application across millions of cases.

13. Quality acceptance and conformity assessment. Evidence: product samples, test results, process records. Criteria: specifications, tolerances, standards. Record: test reports and certificates. Output: accept/reject, certification mark. The criteria are numeric tolerances; the evidence is instrument-generated measurement.

14. Appraisal and valuation. Evidence: the subject asset’s characteristics, comparable transactions, income data. Criteria: professional valuation standards and accepted methodologies. Record: the workfile. Output: the value opinion. The comparable-selection and adjustment logic is semi-codified; the data substrate is records throughout.

15. Forensic examination. Evidence: physical traces rendered into analyzable form — prints, toxicology, ballistics, digital images. Criteria: validated comparison methodologies and match standards. Record: the examiner’s bench notes and report. Output: the forensic finding offered as expert conclusion. The collection interface is physical; the comparison step operates on captured representations.

4.2 Synthesis / Design (Requirements → Constraints → Iterated Candidates → Specified Artifact)

The inverse of adjudication: instead of judging an existing thing against criteria, the practitioner creates a thing to satisfy criteria. The work lives almost entirely in representational media — drawings, models, drafts, prototypes — whose whole purpose is to make candidate forms cheap to generate, evaluate, and discard. The terminal output is itself a representation: a specification complete enough that Transformation/Production (Pattern 3) can realize it.

1. Structural and civil engineering design. Requirements: loads, spans, site conditions, budget. Constraints: codes, material properties, constructability. Working medium: calculations, drawings, analysis models. Output: stamped construction documents. The candidate-evaluation loop — propose a member, check it against load cases — is explicit, numeric, and conducted entirely in the model.

2. Software development. Requirements: functional and performance specifications. Constraints: platform, interfaces, resources, security posture. Working medium: source code, which is simultaneously the design representation and the final artifact — a unique property among design fields, collapsing the design-to-production handoff entirely. Output: the deployable system. The evaluate-candidate loop (write, test, revise) runs in seconds.

3. Drug and molecule design. Requirements: target binding, efficacy, safety margins. Constraints: synthesizability, stability, pharmacokinetics. Working medium: molecular models, assay screens, candidate libraries. Output: the development candidate. The search space is astronomically large and the evaluation of candidates against constraints is the central cost of the entire industry.

4. Architecture and urban planning. Requirements: program, occupancy, experience. Constraints: zoning, codes, site, budget, climate. Working medium: drawings, massing models, renderings. Output: the design documents. Aesthetic and experiential requirements are tacit; the regulatory constraint layer is fully codified — the pattern thus spans the codification spectrum internally.

5. Contract and statute drafting. Requirements: the parties’ intended allocation of rights, duties, and risks (or the legislature’s policy aim). Constraints: existing law, precedent interpretation, enforceability. Working medium: successive drafts and markups. Output: the executed instrument or enacted text. The artifact is pure language; candidate provisions are tested mentally and adversarially against imagined future disputes.

6. Circuit, chip, and hardware design. Requirements: function, speed, power, cost. Constraints: physics, fabrication process rules, thermal envelopes. Working medium: schematics, hardware description languages, simulation. Output: the verified design file sent to fabrication. Design-rule checking — automated subsumption of the candidate against the constraint set — is already the field’s own internal quality regime.

7. Financial structuring. Requirements: the parties’ return, risk, tax, accounting, and regulatory objectives. Constraints: law, regulation, market convention. Working medium: term sheets, models, structure diagrams. Output: the executed structure and its documentation. Candidate structures are evaluated by running them through modeled scenarios.

8. Curriculum and instructional design. Requirements: target competencies and learner starting points. Constraints: time, prerequisites, assessment validity. Working medium: syllabi, lesson plans, item banks. Output: the deliverable course. Note the layering: this pattern designs the artifact that Pattern 8 (Formation) then delivers.

9. Industrial and product design. Requirements: user needs, ergonomics, brand. Constraints: manufacturability, materials, cost targets. Working medium: sketches, CAD, prototypes. Output: the production-ready specification. The prototype loop physically instantiates candidates precisely because some constraints (feel, fit) resist purely representational evaluation.

10. Process and plant design. Requirements: throughput, yield, purity. Constraints: thermodynamics, safety cases, environmental limits. Working medium: process flow diagrams, simulations, piping and instrumentation drawings. Output: the engineered process package. The design explicitly specifies the control loops that Pattern 3 will later execute.

11. Experiment and protocol design. Requirements: the question to be answered with statistical validity. Constraints: ethics, cost, feasibility, confounders. Working medium: protocols, power calculations, pre-registrations. Output: the executable protocol. Another cross-pattern feeder: this designs the instrument that Discovery (Pattern 7) runs.

12. Game, media, and narrative design. Requirements: intended audience experience. Constraints: medium, length, production budget. Working medium: scripts, storyboards, playable builds. Output: the master design document or finished work. The evaluation criterion — audience effect — is tacit and is tested by exposing candidates to real audiences.

4.3 Transformation / Production (Inputs → Process Spec → Controlled Execution → Conforming Output)

The application of a proven method to materials to make them more valuable. The intelligence is in the specification (inherited from Pattern 2) and in tolerance monitoring during execution; the hands-on work is governed by the spec. Its property profile is distinctive: maximal codification of method, maximal physical embodiment, near-instant feedback, and quality defined numerically as conformance.

1. Discrete manufacturing and machining. Inputs: stock material, components. Spec: drawings, tolerances, work instructions, routings. Record: travelers, inspection logs, statistical process control charts. Output: conforming parts. The entire modern quality apparatus of this manifestation consists of instrumenting the process so that deviation is detected in the data before it appears in the parts.

2. Continuous process production (refining, chemicals, pharmaceuticals, food processing). Inputs: feedstocks. Spec: the validated process — setpoints, ranges, sequences. Record: batch records and historian data streams. Output: in-spec product lots. The substrate is already operated principally through control rooms — that is, through its representational layer — with physical intervention as the exception.

3. Construction execution. Inputs: materials, the site. Spec: the construction documents from Pattern 2. Record: daily logs, inspection sign-offs, as-builts. Output: the built work. Uniquely among production manifestations, the “factory” is temporary and the product immovable, keeping physical embodiment high and repeatability low relative to plant-based production.

4. Agriculture and cultivation. Inputs: land, seed, water, nutrients. Spec: agronomic practice — planting windows, application rates, rotation. Record: field histories, yield maps, sensor telemetry. Output: the harvest. The controlling variables are increasingly captured as georeferenced data, thickening the representational layer over what was once purely embodied craft.

5. Surgical and procedural medicine (the procedure itself). Inputs: the patient prepared for intervention. Spec: the operative plan and established technique. Record: the operative note, instrument counts, monitoring traces. Output: the completed procedure. Distinct from the diagnosis that preceded it (Pattern 1/4) and the care that follows (Pattern 11); the procedure is spec-governed execution on the most unforgiving of substrates.

6. Back-office transaction processing (claims processing, loan servicing, payroll, settlements). Inputs: transaction instructions and supporting data. Spec: procedures, cutoffs, validation rules. Record: the transaction log itself. Output: completed, reconciled transactions. Here the “material” is data, the process is fully rule-expressed, and the substrate never leaves representational form — a production activity with the property profile of Pattern 1.

7. Food preparation at scale. Inputs: ingredients. Spec: recipes, portioning, safety temperatures. Record: prep logs, temperature logs. Output: served product. Codified method, embodied execution, immediate feedback.

8. Printing, rendering, and media replication. Inputs: the master file. Spec: output profiles and run parameters. Output: physical or digital copies. The clearest possible case of a fully specified transformation.

9. Energy generation operations. Inputs: fuel or flux. Spec: operating procedures and limits. Record: control-room logs and telemetry. Output: dispatched power. Operated almost wholly through instruments and setpoints.

10. Textile, apparel, and assembly work. Inputs: cut goods, components. Spec: patterns and assembly instructions. Output: finished goods. Historically the frontier case of decomposing a craft into separately specifiable, separately executable steps.

4.4 Diagnosis & Repair (Symptoms → Hypotheses → Probes → Root Cause → Intervention → Verified Restoration)

Abduction over an opaque system: infer the hidden cause of a malfunction, intervene, verify restoration. Distinct from Adjudication because the goal is restored function, not a verdict; distinct from Discovery because someone knows how the system is supposed to work — there is a reference model. The characteristic instrument is the probe; the characteristic knowledge asset is the fault model — the accumulated mapping from symptoms to causes.

1. Medical diagnosis and treatment (the investigative arc). Symptoms: presentation, history. Hypotheses: the differential. Probes: examinations, labs, imaging — each ordered because it discriminates among candidates. Intervention: therapy. Verification: follow-up. The fault model is the accumulated clinical literature; the probe results arrive as records, so the hypothesis-elimination loop runs largely on representations even though probing and treating are embodied.

2. Equipment and vehicle troubleshooting. Symptoms: fault codes, noises, performance loss. Probes: instrument readings, teardown inspection. The fault model is codified to an exceptional degree — service manuals literally publish symptom-to-cause trees — making the reasoning layer of this manifestation among the most externalized in the pattern.

3. Software debugging and incident response. Symptoms: errors, latency, failed assertions. Probes: logs, traces, breakpoints, replays. Intervention: the fix. Verification: tests and monitoring. The system, its symptoms, its probes, and its repair all exist in representational form; nothing in the loop is physical.

4. Network, grid, and infrastructure fault isolation. Symptoms: outages, telemetry anomalies. Probes: sectionalizing, signal tracing. The topology is fully mapped, so fault isolation is search over a known graph — abduction at its most structured.

5. Psychotherapy and behavioral intervention. Symptoms: distress, dysfunction as reported and observed. Hypotheses: formulations. Probes: the therapeutic conversation itself. The substrate is a mind (B5 folded into B3), the fault models are contested, and verification is slow — the low-codification pole of the pattern.

6. Organizational turnaround and management consulting diagnostics. Symptoms: margin erosion, attrition, missed targets. Probes: data analysis, interviews. The fault models are frameworks of varying rigor; evidence is a mix of records and testimony.

7. Root-cause analysis after failures (safety investigations, engineering failure analysis, mortality reviews). Symptoms: the accident. Probes: wreckage examination, recorder data, metallurgy, timeline reconstruction. Output: the probable-cause finding and corrective recommendations — a manifestation that terminates, unusually, in an adjudication-like reasoned public record rather than a repair.

8. Environmental and epidemiological investigation. Symptoms: contamination, disease clusters. Probes: sampling, tracing, case-control comparison. The substrate is a landscape or population; the reasoning is classic differential elimination over candidate sources.

9. Financial and operational remediation (troubled-institution workouts, audit-finding remediation). Symptoms: the cited deficiencies. Probes: process walkthroughs, data validation. Intervention: corrective action plans. Verification: validation testing and closure — the mirror image of the examination activity that produced the findings.

10. Household and building trades repair (plumbing, electrical, HVAC). Symptoms: leak, fault, failure. Probes: pressure tests, circuit tracing. High embodiment in both probe and intervention; well-developed informal fault models.

4.5 Custody / Stewardship (Entrusted Asset → Baseline → Monitoring → Preventive Action → Preserved Value)

A vigilance loop wrapped around an asset subject to entropy. Value is produced by preventing decay, loss, and unauthorized change — by non-events — which makes performance observable chiefly through the condition record and the absence of incident. The characteristic artifacts are the baseline (what good condition is), the monitoring stream, and the maintenance history.

1. Financial asset custody and safekeeping. Asset: securities, cash, collateral. Baseline: positions and entitlements of record. Monitoring: reconciliation — the continuous comparison of internal records against external truth. Action: break investigation and resolution. The entire activity operates on records; its core loop is automated comparison with human attention reserved for exceptions.

2. Infrastructure maintenance (roads, bridges, rail, pipelines, dams). Asset: the physical plant. Baseline: design condition and inspection ratings. Monitoring: scheduled inspection, increasingly continuous sensing. Action: preventive and corrective maintenance. Condition-based regimes — act when the telemetry says so, not when the calendar says so — represent the thickening of the representational layer over the physical asset.

3. Facilities and property management. Asset: buildings and grounds. Monitoring: rounds, work-order streams, building systems telemetry. Action: upkeep, repair dispatch. A composite steward: it monitors (this pattern), dispatches repair (Pattern 4), and serves occupants (Pattern 11).

4. Records, archives, and registries of record. Asset: authoritative information — titles, vital records, corporate registries, archives. Baseline: integrity and completeness of the record set. Monitoring: access control, audit trails, preservation checks. Action: conservation, migration, correction under controlled procedure. The asset is representation; stewardship of it is thus fully representational.

5. Data and system administration. Asset: databases, systems, digital estates. Monitoring: health metrics, backup verification, patch status. Action: tuning, patching, restoration drills. As with archives, substrate and representation coincide.

6. Trusteeship and fiduciary administration. Asset: the trust estate. Baseline: the governing instrument and duty of care. Monitoring: account review, distribution compliance. Record: fiduciary accountings. The rules are legal, the assets substantially financial, the loop record-based.

7. Conservation of land, ecosystems, and fisheries. Asset: the resource. Baseline: population and condition surveys. Monitoring: census, remote observation, quota tracking. Action: restoration, enforcement referrals. Long feedback horizons — the consequences of stewardship choices surface over decades.

8. Museum, art, and artifact conservation. Asset: the collection. Baseline: condition reports. Monitoring: environmental control, periodic examination. Action: minimal, reversible, exhaustively documented intervention — a stewardship culture with the strictest change-control discipline in existence.

9. Inventory and warehouse custody. Asset: goods held. Baseline: perpetual inventory records. Monitoring: cycle counts, shrinkage analysis. The loop is the reconciliation of physical count to record — identical in structure to securities custody, applied to matter.

10. Public-health surveillance of a population’s baseline. Asset: population health status. Monitoring: reportable-condition streams, sentinel indicators. Action: alerts and interventions. Stewardship of a statistical asset, conducted entirely through data streams.

4.6 Conveyance / Logistics (Origin → Routing → Movement with Integrity → Confirmed Delivery)

Value from changing location or possession without changing the thing. Two layers interlock: an optimization layer that plans routes and allocations over a network representation, and an execution layer that physically moves items while preserving integrity and chain of custody. The characteristic artifacts are the manifest, the tracking identifier, and the proof of delivery.

1. Freight, parcel, and postal networks. Items: shipments. Routing: network optimization over hubs and lanes. Integrity: sealing, condition monitoring, chain of custody. Confirmation: delivery scan. The planning layer is pure computation over the network model; the movement layer is embodied.

2. Payments, clearing, and settlement. Items: value and obligations. Routing: payment rails and netting arrangements. Integrity: authentication, finality rules. Confirmation: settlement. Conveyance whose cargo is pure record — the manifestation of this pattern with zero physical residue.

3. Utility distribution (power, water, gas). Items: the commodity as continuous flow. Routing: network flow management, dispatch. Integrity: pressure, voltage, quality standards. Managed from control centers through telemetry — movement of matter governed almost entirely at the representational layer.

4. Telecommunications and data transport. Items: packets, calls, streams. Routing: dynamic path selection. Integrity: error correction, encryption. The routing decisions themselves execute at machine timescales; humans steward the network (Pattern 5) rather than route the traffic.

5. Passenger transportation (airlines, transit, rail). Items: people — cargo with preferences and rights, importing Pattern 11 obligations into the movement. Routing: scheduling and capacity allocation. Confirmation: completed carriage.

6. Supply-chain orchestration and freight forwarding. Items: goods across parties and borders. Routing: multi-leg planning, customs sequencing. Record: the documentary chain — bills of lading, certificates, entries — a conveyance of paper alongside the conveyance of goods.

7. Medical logistics (blood, organs, specimens, pharmaceuticals). Items: perishable, condition-critical cargo. Integrity: cold chain and viability windows, continuously monitored and logged. The integrity record is as important as the delivery itself.

8. Warehouse fulfillment and last-mile delivery. Items: ordered goods. Routing: pick-path and route optimization. Confirmation: delivery evidence. The indoor segment operates in a fully mapped, structured environment; the last mile does not — a distinction with large consequences for how each segment can be instrumented.

9. Evidence and chain-of-custody transport. Items: legal and forensic materials. Integrity: sealed custody with attested handoffs. A manifestation in which the record of the movement carries more consequence than the movement’s efficiency.

4.7 Discovery / Inquiry (Question → Method → Investigation → Validated New Knowledge)

Reasoning against the unknown: no one possesses the answer, so there is no rulebook that determines it and no reference model to repair toward. The work alternates hypothesis generation (abduction) with designed tests (imported from Pattern 2) and disciplined validation. Its characteristic artifacts are the protocol, the notebook or evidence file, and the validated claim. Feedback latency is the longest of any pattern.

1. Basic scientific research. Question: open. Method: the experimental and observational canon. Record: lab notebooks, data archives, publications. Validation: replication and peer scrutiny (which is itself a Pattern 1 activity applied to research artifacts). The generation of candidate hypotheses and the design of discriminating experiments are the twin cognitive cores.

2. Applied R&D and product research. Question: can capability X be achieved within constraints Y? The discovery loop runs inside commercial constraint, feeding validated findings directly into Pattern 2 design.

3. Pharmaceutical discovery and clinical trials. Question: does the candidate work, safely? Method: screening cascades, then the phased trial apparatus — a discovery pipeline so proceduralized that its later stages resemble Pattern 3 execution wrapped around a Pattern 7 question.

4. Exploration and prospecting (minerals, energy, water). Question: what lies beneath? Probes: surveys, seismic imaging, drilling. The substrate is physically hidden; the reasoning runs on the imaging and sample data the probes return.

5. Investigative journalism. Question: what happened, that someone prefers concealed? Evidence: documents, sources, records. Validation: corroboration standards, editorial review. An inquiry conducted partly against adversarial concealment (a B6 tinge), documented in the story file.

6. Intelligence analysis. Question: an adversary’s capabilities and intentions. Evidence: collected reporting across sources. Record: the assessment with stated confidence. Discovery under active denial and deception — the pattern’s fully adversarial variant.

7. Market, user, and social research. Question: what do people want, believe, do? Method: surveys, interviews, experiments, behavioral data. The substrate is minds in aggregate, probed statistically.

8. Economic and policy research. Question: how does the system respond to intervention? Method: models, natural experiments, data analysis. Conducted almost entirely on recorded data; the laboratory is the dataset.

9. Archaeology, paleontology, and historical inquiry. Question: what was? Evidence: excavated and archival traces. The unknown here is the past — probed through its physical and documentary residue.

10. Mathematical and theoretical inquiry. Question: what follows? Method: proof and derivation. The purest case: substrate, probe, and product are all formal representations; nothing physical exists anywhere in the loop.

4.8 Transmission / Formation (Capability Model → Curriculum → Guided Practice → Assessed Competence)

Replicating capability in another mind. The practitioner maintains a running model of the learner’s current state, exposes the gap to target competence, and scaffolds practice across it. The characteristic artifacts are the curriculum (designed under Pattern 2), the practice record, and the assessment (a Pattern 1 activity embedded at the end). Feedback is slow and the substrate is a cooperative mind — the pattern’s codification is correspondingly low at its core, even where its materials are highly structured.

1. Classroom instruction. Learner state: revealed through questioning and work product. Scaffolding: explanation, worked examples, graded practice. Record: the gradebook and portfolio. The content is fully codified; the moment-to-moment diagnosis of a particular mind’s misunderstanding is not.

2. Apprenticeship and trades training. Capability: embodied skill. Scaffolding: demonstration, supervised performance, progressive responsibility. The knowledge transmitted is substantially tacit — resident in hands and judgment rather than text — which is precisely why the pattern here requires co-presence.

3. Clinical training and residency. Scaffolding: graduated autonomy under supervision — see one, do one, teach one — with the supervisor adjudicating readiness at each step. Formation interleaved with real Pattern 1/4/11 work on real patients.

4. Examiner commissioning and professional formation (bank examiners, auditors, actuaries, attorneys). Capability: the judgment of the field. Method: rotations, casework under review, capstone examination. The formal syllabus is codified; the judgment being formed — what a seasoned examiner senses in a loan file — is the tacit residue the syllabus cannot hold.

5. Corporate onboarding and workforce training. Capability: role proficiency. Method: structured modules plus supervised ramp. The module layer is fully explicit and endlessly repeatable; the coaching layer is not.

6. Athletic and performance coaching. Capability: physical and expressive skill. Scaffolding: drill design, cueing, feedback on attempts. The coach’s core act is perceiving the delta between the attempt and the model and selecting the cue that closes it.

7. Flight, simulation, and safety-critical training. Method: syllabus of maneuvers, simulated abnormal conditions, check rides. Notable for moving dangerous practice into synthetic environments — practice conducted against a representation of the world rather than the world.

8. Coaching, mentoring, and advisory formation. Capability: judgment and self-direction. Method: dialogue, reflection, modeled thinking. The least structured manifestation; almost wholly tacit.

9. Parenting and early formation. The unpriced base case of the entire pattern: continuous, relational, unspecialized capability transfer.

10. Religious and civic formation. Capability: identity, norms, practice. Method: instruction, ritual, community participation. Transmission of meaning as well as skill.

4.9 Negotiated Exchange (Divergent Interests → Preference Discovery → Bargaining → Binding Terms)

Value created by finding mutually acceptable terms where none existed. The cognitive core is strategic theory-of-mind: inferring the counterparty’s true preferences, reservation points, and constraints from behavior and signal, while managing one’s own disclosure and commitment. The characteristic artifacts are the term sheet, the negotiation record, and the executed agreement (drafted under Pattern 2).

1. Enterprise and complex sales. Interests: the buyer’s problem and budget versus the seller’s price and terms. Discovery: needs analysis, stakeholder mapping. Terms: the contract. The discovery phase is increasingly conducted through recorded interactions and data trails, giving the strategic layer a growing representational base.

2. Procurement and sourcing. The mirror image: specification, competitive solicitation, evaluation (a Pattern 1 step against stated criteria), then negotiation of final terms.

3. Mergers, acquisitions, and investment dealmaking. Interests: price, risk allocation, control. Discovery: diligence (a Pattern 1 investigation feeding the negotiation). Terms: purchase agreements dense with negotiated risk allocation. The valuation models on which positions rest are fully representational; reading the counterparty across the table is not.

4. Labor and collective bargaining. Parties: institutional, repeat-play, with internal constituencies. The negotiation is two-level — across the table and within each side — doubling the theory-of-mind burden.

5. Diplomacy and treaty-making. Interests: national. Instruments: communiqués, drafts, linkage across issues. The longest time horizons and highest ambiguity tolerance in the pattern.

6. Litigation settlement and plea negotiation. Interests: certainty versus vindication, resources versus exposure. Conducted in the shadow of Pattern 1 — each side’s position is a forecast of what adjudication would yield, making the negotiation substantially a dispute about a predictable distribution.

7. Commercial contracting and vendor terms. Repeat, semi-standardized bargaining over price, service levels, and risk. The standardized clauses are codified; the deviations are where the negotiation lives.

8. Real estate transactions. Discovery: inspection and appraisal records feeding positions. Terms: price and contingencies. A negotiation between parties who each hold recorded evidence about the same asset.

9. Insurance and reinsurance placement. Interests: risk transfer versus premium. The bargaining operates over actuarial representations of the risk — models negotiating with models, through people.

10. Hostage, crisis, and high-stakes negotiation. The extreme case: preferences opaque, stakes absolute, rapport itself the principal instrument.

4.10 Coordination / Orchestration (Shared Objective → Plan → Sequencing/Allocation → Synchronized Outcome)

The work is the alignment of others’ work: decomposing an objective into tasks, allocating them across actors and time, tracking state, and replanning as reality diverges from plan. The cognitive core is scheduling and allocation over a dependency structure; the relational layer manages the humans attached to the tasks. The characteristic artifacts are the plan, the status representation, and the change log.

1. Project and program management. Objective: the deliverable. Plan: work breakdown, dependencies, critical path. Status: the tracked plan — a live representation of many actors’ progress. The replanning loop is continuous computation over that representation, interleaved with persuasion and escalation among the actors.

2. Air traffic control. Objective: safe, efficient flow. Plan: clearances and sequencing. Status: surveillance picture. The most time-critical manifestation — allocation decisions issued and revised in seconds over a fully instrumented representation of the traffic.

3. General contracting and construction management. Objective: the built work. Plan: trade sequencing under physical dependency (the slab before the frame). Status: schedules, submittals, daily reports. Orchestration where the dependency structure is dictated by physics.

4. Operating-room, clinic, and hospital scheduling. Objective: throughput under safety. Allocation: rooms, teams, equipment across cases of uncertain duration. Continuous reallocation as cases run long or emergencies preempt.

5. Incident command and emergency management. Objective: containment and restoration. Plan: the incident action plan, refreshed each operational period. A coordination doctrine explicitly designed (Pattern 2) to be instantiable in hours across strangers.

6. Manufacturing production planning and scheduling. Objective: demand fulfillment. Allocation: lines, labor, materials across orders. Conducted almost entirely within planning systems — orchestration over a model of the plant.

7. Software release and platform engineering coordination. Objective: the release. Plan: dependency-ordered integration, gated promotion. The tracked state and the work itself share a medium; coordination artifacts and product artifacts live in the same systems.

8. Event, broadcast, and film production. Objective: the show. Plan: call sheets, run-of-show, shot lists. Real-time cueing during execution — orchestration compressed to the second.

9. Military command and staff planning. Objective: the mission. Plan: orders, synchronization matrices, contingencies. Coordination under an adversary (B6) who is actively disrupting the plan — hence the doctrine that plans survive only until contact.

10. Examination scoping and engagement management. Objective: supervisory coverage of an institution. Allocation: examiner-hours across risk areas, sequenced access to bank personnel and records. The supervisory instance of the pattern, wrapped around a core of Pattern 1 work.

4.11 Care / Sustenance (Dependent Person → Assessed Needs → Continuous Attention → Maintained Wellbeing)

Like stewardship, but the asset is a person, and the relationship is itself part of the product. The practitioner runs a continuous, empathic assessment loop — noticing state changes, anticipating needs, responding — in which presence and trust are instruments, not overhead. Codification is lowest here of all patterns at the relational core, even as the surrounding documentation (care plans, charts) is highly structured.

1. Nursing and bedside care. Needs: assessed continuously through observation, monitoring, and the relationship. Record: the chart — vitals, assessments, interventions. The monitoring layer is heavily instrumented and representational; the noticing, comforting, and advocating are irreducibly relational.

2. Eldercare and long-term care. Needs: daily living, dignity, safety, companionship. Long-horizon, low-event care in which consistency of presence is the substance of quality.

3. Childcare and early education. Needs: safety, development, attachment. Care fused with Formation (Pattern 8) — the assessment loop tracks both wellbeing and growth.

4. Hospitality and guest services. Needs: comfort and experience, anticipated before expressed. The commercial manifestation: the anticipation loop applied to guests, with service standards codifying what can be codified and staff judgment covering the rest.

5. Social work and case management. Needs: stability across housing, benefits, safety, health. The caseworker orchestrates services (importing Pattern 10) while sustaining the relationship that keeps the client engaged.

6. Chaplaincy, counseling presence, and bereavement support. Needs: meaning and accompaniment. The manifestation where presence is not an instrument of the work but its entirety.

7. Disability support and personal assistance. Needs: negotiated daily enablement. Care under the direction of the person cared for — the assessment loop is collaborative by design.

8. Customer support and success (the relational tier). Needs: resolution and confidence. The transactional tier of this work is rule-governed lookup (Pattern 1/3 in character); the escalated tier — the distressed customer, the ambiguous failure — is care proper.

9. Veterinary and animal husbandry care. Needs: read without language, through observation and handling. The assessment loop with the report channel removed.

10. Palliative and hospice care. Needs: comfort, dignity, family support at the end of life. Care with the restoration objective removed — pure sustenance, the pattern in its most distilled form.

4.12 Performance / Expression (Intent → Craft → Rendered Experience → Audience Effect)

The output is an experience, produced and consumed in the same act (or captured for replay). The practitioner exercises real-time expressive control — of voice, body, medium — calibrated against the audience’s live response. The evaluation criterion is effect on minds, which is tacit, and the feedback channel is the audience itself.

1. Musical performance. Intent: the interpretation. Craft: technique over instrument or voice. The score is fully codified; everything the performance adds to the score is not.

2. Theater, oratory, and public speaking. Craft: delivery, timing, presence. The same text renders differently to every room — the calibration loop against the audience is the work.

3. Trial advocacy. Intent: persuasion of a factfinder. Craft: narrative, examination, argument — performance in service of a Pattern 1 proceeding, aimed at the adjudicator’s mind.

4. Teaching-as-performance and keynote delivery. The expressive layer of Formation: attention held, material animated, room read.

5. Broadcasting, commentary, and live media. Craft: continuous fluent rendering under time pressure, calibrated to an unseen audience through delayed signals.

6. Competitive athletic performance. Craft: embodied excellence rendered for spectators, inside a rule frame (adjudicated by Pattern 1 officiating).

7. Fine art, authorship, and composition. The asynchronous variant: the experience is rendered into a durable artifact and the audience effect arrives after the fact. The craft loop runs against an internal model of the audience rather than a live one.

8. Preaching and liturgical leadership. Intent: meaning and exhortation. Performance fused with Formation and Care within a community.

9. Brand storytelling and campaign creative. Intent: attitude change at population scale. Rendered through media, measured through response — the manifestation with the most instrumented audience-effect signal.

10. Stand-up, improvisation, and interactive entertainment. The tightest calibration loop in the pattern: audience response arrives in milliseconds and reshapes the performance continuously.

4.13 Protection / Enforcement (Threat Model → Vigilance → Detection → Interdiction → Deterrence)

Value from bad things not happening, produced against an adversary who adapts. The pattern combines a vigilance loop (Pattern 5’s structure) with abductive detection (is this anomaly hostile?) and allocation of protective resources — all under adversarial pressure, which continuously erodes any static method. The characteristic artifacts are the threat model, the detection log, and the incident record.

1. Cybersecurity operations. Threats: intrusion, exfiltration, ransom. Vigilance: telemetry across the estate. Detection: anomaly triage in the security operations center. Interdiction: containment and eviction. The entire engagement — attack and defense alike — occurs in representational space; nothing physical need move.

2. Fraud detection and interdiction. Threats: payment fraud, account takeover, application fraud. Vigilance: transaction streams scored in real time. Detection: flagged anomalies adjudicated by analysts (a Pattern 1 step embedded in the loop). The adversary probes the rules and adapts, forcing perpetual model renewal.

3. Anti-money-laundering and sanctions surveillance. Threats: illicit flows through legitimate channels. Vigilance: monitoring scenarios over transaction data. Record: alerts, case files, suspicious activity reports — a protection loop whose output is itself evidence feeding public-side adjudication and enforcement.

4. Policing and public safety patrol. Threats: crime and disorder. Vigilance: presence and observation. Interdiction: response and arrest. Deterrence through visible capacity. High embodiment, high discretion, governed by rules of engagement.

5. Military defense. Threats: state adversaries. Vigilance: surveillance and warning. The apex adversarial manifestation, spanning every substrate at once.

6. Physical security and protective services. Threats: intrusion, targeted harm. Vigilance: perimeters, screening, close protection. Layered defenses designed (Pattern 2) so that detection at each layer buys interdiction time.

7. Border, customs, and screening operations. Threats: prohibited entry and cargo. Detection: inspection and targeting — allocation of scarce scrutiny across a flow, informed by risk models over the flow’s records.

8. Content moderation and platform integrity. Threats: abuse, manipulation, prohibited material at scale. Vigilance: automated screening over the content stream with human adjudication of the hard residue — the newest large-scale manifestation, and one conducted entirely on representations.

9. Lifeguarding, fire watch, and safety observation. Threats: accident rather than adversary — the pattern’s non-adversarial boundary case, where the “opponent” is hazard and inattention. Pure vigilance with embodied interdiction.

10. Counterintelligence and insider-threat programs. Threats: the adversary already inside. Detection: behavioral and access anomalies against baseline. The manifestation with the most delicate ratio of surveillance to trust.

5. Composition: Workflows, Roles, and Institutions

The patterns are atoms; real work is molecular. Three composition rules govern how Layer 2 patterns assemble into the jobs and institutions observed at Layer 3.

Sequencing. Most end-to-end processes are chains of distinct patterns with handoffs between them. A full episode of medical care runs Adjudication (diagnosis against criteria) → Diagnosis & Repair (investigation and treatment) → Transformation (the procedure) → Care (recovery) → Stewardship (ongoing monitoring). A capital project runs Discovery (feasibility) → Design → Negotiated Exchange (contracting) → Coordination (construction management) → Transformation (the build) → Adjudication (inspection and acceptance) → Stewardship (operations and maintenance). The handoff points are where each pattern’s terminal artifact — the diagnosis, the specification, the contract, the plan, the certificate — becomes the next pattern’s input. The completeness and fidelity of these artifacts determine how cleanly the chain can be divided among different actors, different organizations, and different instruments.

Bundling. A role is a weighted portfolio of patterns performed by one person because the substrate or the economics favor co-location of the activities. The bedside nurse bundles Care, Stewardship (monitoring), Procedural Execution (medication administration), and Adjudication (assessment against protocols). The bank examiner bundles Adjudication at the core with Coordination (engagement management), Negotiated Exchange (scoping and findings discussions), and Performance (the exit meeting). Bundles are historical and economic artifacts, not logical necessities; they re-form whenever the relative costs of the constituent activities shift, and the model’s decomposition makes visible exactly which strands of a bundle share a property profile and which do not.

Embedding. Patterns nest inside one another. Every pattern embeds micro-adjudications (is this evidence sufficient? is this deviation in tolerance?). Protection embeds Adjudication at its detection-triage step. Formation embeds Adjudication at assessment. Negotiation embeds Adjudication in diligence and runs in the shadow of forecasted adjudication. Design embeds thousands of candidate evaluations, each a subsumption of the candidate under the constraint set. This is why Evidentiary Adjudication, the anchor pattern, is best understood not merely as one cell among thirteen but as the most widely embedded operation in the entire economy: wherever criteria meet evidence inside any other pattern, a small adjudication is running.

Institutions, finally, are standing compositions: a bank is Adjudication (underwriting), Custody, Conveyance (payments), Negotiated Exchange, Protection (fraud and financial-crime controls), and Stewardship, bound by Coordination. A hospital, a court system, an airline, a university each decompose the same way. The institutional chart rarely matches the pattern decomposition — which is precisely why structurally identical activities inside different institutions so often fail to learn from one another.

6. Comparative Properties Across the Patterns

The five properties introduced in Section 2.3 vary systematically across the thirteen patterns. The comparison below states each pattern’s typical profile; individual manifestations deviate, and the deviations are themselves informative.

Several regularities deserve emphasis.

Codification and representational sufficiency travel together, and Adjudication sits at their joint maximum. Where the rules can be fully written and the evidence arrives as records, the entire activity exists in inspectable, transportable, reproducible form. Patterns 1, 3, 6 (planning layer), and 10 cluster at this pole. Patterns 8, 9, 11, and 12 — the theory-of-mind family — cluster at the opposite pole, where the operative knowledge resists writing down and the substrate must be met in person.

Every pattern contains an internal gradient. Within Adjudication, high-volume eligibility determinations sit near the codified extreme while credibility findings sit far from it. Within Care, vital-sign monitoring is fully instrumented while comforting is not. Within Design, code compliance checking is mechanical while aesthetic judgment is not. Analyzing an activity therefore means locating it not just in a pattern but along its pattern’s internal gradient — from the rule-dense, record-borne core to the tacit, embodied, relational periphery.

The historical drift of every pattern runs in one direction: toward thicker representation. Ledgers became data extracts; the patient became the chart plus the monitor stream; the plant became the historian database; the shipment became the tracking record; the battlefield became the surveillance picture. Each pattern has spent its modern history converting more of its substrate into records and more of its method into explicit criteria — moving activity, cell by cell, toward the high-codification, high-representation corner of the grid. The patterns differ in how far this drift has proceeded and in where it meets hard limits: physics in Production and Conveyance, co-presence in Care and Formation, adaptive opposition in Protection, and the genuinely unwritten in Discovery.

Vigilance-based patterns share a measurement pathology. Stewardship and Protection produce non-events, so their performance is visible only in the condition record and the incident log — which places extraordinary weight on the completeness of those records and creates chronic underinvestment wherever the records are thin.

Adversarial pressure is the great destabilizer of method. Every other pattern can accumulate method monotonically; Protection cannot, because its substrate studies its methods and routes around them. Any manifestation of any pattern that acquires an adversarial element — gamed criteria in Adjudication, diverted cargo in Conveyance, manipulated findings in Discovery — inherits this instability in proportion.

7. Applying the Model

The model is used bottom-up, in four steps.

Step 1 — Decompose to activities. Take the role, process, or institution and break it into constituent activities at the level where each has a single dominant reasoning mode. “Physician” is too coarse; “renders diagnosis against criteria,” “selects discriminating test,” “performs procedure,” “comforts family” are right-sized.

Step 2 — Classify each activity on the grid. Ask two questions: What is the cognitive operation? (judging against criteria, explaining symptoms, generating a form, following a method, allocating resources, modeling a mind) and What does it operate on? (records, matter, an opaque system, the unknown, cooperative minds, adversarial minds). The activity snaps into a Layer 2 pattern.

Step 3 — Profile the activity on the five properties. How completely are its rules written? How fully does its substrate exist as records? How fast does feedback arrive? What does error cost, and can it be reversed? Does anything adapt against it? Then locate it on its pattern’s internal gradient, from codified core to tacit periphery.

Step 4 — Compare across sectors. Find the activity’s structural siblings — the other manifestations of its pattern with matching property profiles — and examine how they are instrumented, divided, staffed, recorded, and quality-controlled. Because the structure is the same, their solutions transfer. The custody reconciliation loop illuminates warehouse cycle counting; the aviation checkride illuminates examiner commissioning; the manufacturing control chart illuminates transaction-processing quality; the pathology workflow illuminates any activity that converts a physical substrate into images and numbers before judging it.

Two further uses follow. Interface analysis: the handoff artifacts between patterns in a workflow (the specification, the diagnosis, the plan, the certificate) are the natural seams of any process; making them complete and explicit is what allows the chain to be re-divided among new actors and new instruments without loss. Bundle analysis: wherever a role bundles activities with sharply different property profiles — a highly codified, record-borne strand alongside a tacit, relational strand — the bundle is under latent tension, and the model predicts where it will eventually re-form.

8. Conclusion

The nineteenth-century examiner’s desk — evidence to the left, criteria before him, digest under his hand, report at the end — turns out to be one tile in a small mosaic. Beneath the entire economy’s surface variety lie a handful of reasoning modes acting on a handful of substrates; where the combinations are economically common, they crystallize into thirteen recurring patterns; and every sector’s activities, however specialized their dress, are re-instantiations of those patterns with local vocabulary and local instruments.

The model’s discipline is to see every activity three times: as a primitive combination (what operation, on what substrate), as a pattern instance (which recurring shape, where on its internal gradient), and as a sector practice (which local rules, records, and tools). Seen this way, the economy becomes legible as a system of structural families whose members can learn from one another across industry lines — and whose long, shared drift toward explicit criteria and richer representation of their substrates is the single most consequential trend running through them all.