Summary

Give each concept that the emsdeavor depends on the full specification treatment, building bottom-up through the dependency graph so each layer has a proper home before the layer above it needs to reference it. For method concepts (layer 5), ground each concept in its home discipline(s) first — the semiotic-* convention ASSEMBLES from discipline knowledge, it doesn’t invent from scratch.

Motivation

The emsdeavor sits atop a layered conceptual stack (see the conceptual dependency map). Each layer contains concepts that need the specification treatment — the pattern completed for “specification”:

  1. concept page (what the thing is, in its home discipline)
  2. general specification (conformance requirements for the concept)
  3. semiotic convention (how the concept works in semiotic universes)
  4. semiotic conformance spec (requirements for the semiotic version)
  5. make skills where operational

Currently only “specification” has this full chain. Without it, agents conflate general concepts with their semiotic specializations, put files in wrong locations, and spend tokens searching for conventions that should be at deterministic paths.

The assembly principle

The semiotic-* specifications should be ASSEMBLIES of discipline knowledge, not standalone inventions. The build direction is:

discipline research → concepts/terms → general spec → semiotic convention

For example, semiotic-policy shouldn’t define governance from scratch. Governance has deep roots in sociology (governmentality.md, institution.md, horizontal-organization.md already exist), FOSS practice, and anarchist organizing theory. The general concept of “policy” should have coverage in those disciplines, and THEN semiotic-policy assembles what’s relevant to semiotic universes because it matches the formalism and/or philosophy.

This aligns with policy 003 (ground in discipline practice) and policy 006 (research produces texts). It also means we build reusable discipline knowledge that serves purposes beyond the emsdeavor.

The ASR flattening problem

The ASR directory is the worst case of concepts being flattened into one place. It currently contains content from at least 3 layers:

Layer 1-2 concepts misplaced in ASR concepts/: closure-pressure.md, composite-semioverse.md — mathematical objects that belong in mathematics/.

Layer 3 theory misplaced in ASR theory/: agent-skill-interaction.md, prompt-routing.md, work-unit-lifecycle.md, multi-agent-coordination.md, operational-closure.md, coordination-patterns.md, conditional-skill-invocation.md, questions-as-first-class-objects.md — these describe how agential semioverses work in general, not how THIS repository works specifically.

Layer 5 content misplaced in ASR: policy.md concept, planning-methodology.md, semantic-frontmatter.md — these belong to their respective semiotic-* specification directories.

Meanwhile the actual agential semioverse directory at mathematics/objects/universes/agential-semioverse/ has only 4 files.

Concept inventory with discipline homes

Structural concepts (layers 0-4)

LayerConceptHome disciplineExisting coverageChain status
0relationalityphilosophystubemsenn’s work
1semiotic universemathematicsmathematics/objects/universes/semiotic-universe/ (draft)concept only
2interactive semioversemathematicsmathematics/objects/universes/interactive-semioverse/ (draft)concept only
3agential semioversemathematics4 files at math/ + 8 misplaced in ASR theory/partial, scattered
3agentmathematics (+ philosophy, sociology)none at layer 3; referenced throughoutmissing
3toolmathematics (+ technology)none at layer 3missing
3normmathematics (+ sociology, philosophy)sociology has governmentality, institutionmissing at layer 3
3skillmathematics (+ technology)none at layer 3; only at ASR layer 4missing
4ASRtechnologytechnology/specifications/agential-semioverse-repository/partial, bloated

Method concepts (layer 5) with discipline sources

ConceptSemiotic conventionDiscipline home(s)Existing discipline coverageChain status
specificationtechnologytechnology/concepts/specification.md✅ complete
markdowntechnology (markup languages)❌ no concept pagepartial
endeavor✅ (v0.8)multiple (see derivation texts)derivation texts from TCCC, insurgent, disaster, military, medicine, FOSS, MUD, Starfleetpartial
project-managementbusiness, militarism, FOSSplanning-methodology.md misplaced in ASRpartial
versioningtechnology (software engineering)partial
changelogtechnology (software engineering)partial
policysociology (governance), FOSS, philosophysociology has governmentality.md, institution.md, horizontal-organization.md, anarchism.md, prefigurative-politics.mdpartial
triagemedicine (original concept), information science❌ no concept page in medicine/partial

Instance layer (layer 6)

ConceptStatus
emsemioversepersonal/projects/emsemioverse/index.md exists; IS the composition

Execution strategy

The dependency graph determines build order

The conceptual dependency map edges form a DAG. The optimal execution order is a topological sort — build each node only after its dependencies exist. This means bottom-up through the layers, with discipline grounding at each step.

Quickest wins first within each tier

Within each tier, prioritize by:

  1. How much existing content can be relocated (mechanical work)
  2. How much discipline coverage already exists (less research needed)
  3. How many dependents are unblocked

Execution phases

Phase 1: Decompress the ASR (highest leverage, mostly mechanical)

The ASR flattening is the single biggest source of misplaced content. Decompressing it is mostly relocation, not creation.

  1. Audit each ASR theory/ file — classify as layer 3 (agential semioverse general) vs layer 4 (this repository specific)
  2. Relocate layer 3 content to mathematics/objects/universes/ agential-semioverse/
  3. Relocate layer 1-2 concepts (closure-pressure.md, composite-semioverse.md) to appropriate math directories
  4. Relocate layer 5 content (policy.md → semiotic-policy, planning-methodology.md → semiotic-project-management, semantic-frontmatter.md → semiotic-markdown)
  5. Verify the agential semioverse mathematical definition is sound after receiving the relocated content

Phase 2: Ground method concepts in their disciplines

For each layer 5 concept, check whether its home discipline(s) have adequate coverage. If not, the first step is discipline-level research and concept creation — NOT writing the semiotic-* spec.

ConceptWhat discipline coverage is needed
markdowntechnology concept page for markup languages
policysociology coverage of governance may already suffice; check FOSS governance in technology
triagemedicine concept page for medical triage; information science coverage
project-managementbusiness/PM schools; check militarism for planning doctrines
versioningtechnology concept for software versioning practices
changelogtechnology concept for change tracking practices

Endeavor already has extensive derivation texts grounding it across multiple disciplines.

Phase 3: Build general specifications

For each concept that now has discipline grounding:

  1. Create concept page in technology/concepts/ (or appropriate discipline) if it doesn’t exist
  2. Create {concept}-specification using make-specification — the general conformance scaffold, written FROM the discipline knowledge

Phase 4: Complete semiotic chains

For each concept with a general spec:

  1. Verify the existing semiotic-{concept} references the general spec
  2. Create semiotic-{concept}-specification using make-semiotic-specification
  3. Create make skills where operational

Phase 5: Verify the stack

Walk the dependency graph top-down and verify:

  1. Each edge in the conceptual dependency map has both endpoints properly chained
  2. The ASR contains only layer 4 content + operational infrastructure
  3. Each semiotic-* spec references its general spec and discipline sources

Where semiotic-* content needs to decompose

Some current semiotic-* content actually belongs in external disciplines. The semiotic-* layer should assemble FROM those disciplines, not contain the discipline knowledge itself.

Semiotic specContent that may belong elsewhereTarget discipline
semiotic-policyGeneral governance theory, policy lifecycle patternssociology (governance theory), FOSS governance practices
semiotic-project-managementGeneral PM methodology, cycle theorybusiness (PM schools), technology (FOSS PM)
semiotic-triageGeneral triage theory, priority sortingmedicine (triage origins), information science
semiotic-endeavorAlready well-decomposed via derivation textsmultiple (documented in X1-X3, X7)
semiotic-markdownGeneral markup language conventionstechnology (markup languages)
semiotic-versioningGeneral version semanticstechnology (semantic versioning)
semiotic-changelogGeneral change tracking conventionstechnology (software engineering)

Done when

  • Phase 1: ASR decompressed (layer 3 → math/, layer 5 → specs)
  • Phase 2: discipline coverage verified or created for each concept
  • Phase 3: general specifications exist for all method concepts
  • Phase 4: semiotic chains complete (semiotic-X-specification for each semiotic-X that needs one)
  • Phase 5: dependency graph walk passes — all edges have properly-chained endpoints
  • Conceptual dependency map updated with all new nodes
  • Plan 0033 log updated

Dependencies

  • specification-specification and make-specification skill (complete)
  • semiotic-specification-specification and make-semiotic-specification skill (complete)
  • Plan 0033 (active, provides the semiotic layer context)

Log

2026-03-08 — Created. Initially scoped to only Layer 5 method aspects. Revised to include the full conceptual stack (layers 0-6) after emsenn pointed out that structural concepts (semiotic universe, agential semioverse, agents, etc.) are what the method aspects carry and also need chains.

2026-03-08 — Second revision. emsenn identified the ASR flattening problem: the ASR directory has accumulated content from layers 1-2 (closure-pressure, composite-semioverse), layer 3 (8+ theory files about agential semioverses in general), and layer 5 (policy concept, planning methodology, semantic frontmatter). Meanwhile the actual agential semioverse directory has only 4 files. Agential semioverse reprioritized to tier 1.

2026-03-08 — Third revision. emsenn identified the assembly principle: semiotic-* specs should assemble FROM discipline knowledge, not invent standalone. Example: policy has roots in sociology (governmentality, institution, horizontal-organization already exist), FOSS, and anarchist organizing — semiotic-policy should reference those, not reinvent governance. Execution restructured into 5 phases: decompress ASR (mechanical) → ground in disciplines (research) → general specs → semiotic chains → verify stack. This builds bottom-up through the dependency graph, producing discipline knowledge as reusable byproduct. The dependency map edges determine build order; cross-cutting sources (X1-X7) identify where discipline coverage already exists.