Document Hierarchy and Fractal Composition

1. A document is a hierarchy of contributions

AIOS reads a document as an artifact whose parts contribute differently to one governing purpose. Its hierarchy is semantic before it is typographic:

wording
  → sentence or claim
    → paragraph or local movement
      → section or movement
        → document
          → related document field
            → project
              → corpus and reusable practice

Each level can become the object of attention. Each also remains intelligible through its relationship to the levels above and below it. A sentence belongs to a movement; a movement carries a document's argument; a document serves a project; a completed project can contribute knowledge to a larger corpus.

This hierarchy lets the system focus precisely without losing consequence.

2. The Fractal Seed across artifact scales

Why → How → What makes contribution legible across artifact levels while each scale retains its own form. The seed relates the levels; it does not make them identical. Chapter 03 owns the wider work scales—turns, movements, commissions, plans, workflows, and projects—while this chapter follows contribution through the artifact field.

3. File hierarchy and semantic hierarchy answer different questions

A folder tree locates custody. A semantic hierarchy explains contribution and standing.

The same directory can hold sources, exploratory thinking, decisions, plans, canonical artifacts, workflow returns, exports, and superseded versions. Their neighboring paths do not make them equivalent.

AIOS combines ordinary paths for custody, stable identity for continuity, companion memory for purpose and state, explicit relationships for semantic bearing, and derived maps for cross-scale understanding. Location remains useful without carrying more authority than it can support.

4. The Document Spine

The Document Spine is the accepted compact account of why the document exists, how its load-bearing movements develop, and what the document establishes or leaves open.

It lives in companion memory and appears through a readable Document Map. It gives later movements a durable theory of the artifact's current structure while remaining revisable through review and accepted change.

The spine governs context and navigation rather than visible formatting. Prose follows its audience and object. The accepted artifact remains authoritative when a stale projection disagrees.

5. Movements carry contribution

A heading names a region. A movement account explains what the region does.

An outline such as “Background, Analysis, Recommendation” provides location but little semantic bearing. A contribution map instead shows that one movement reframes the apparent problem, another separates two evidence classes, a third establishes the binding constraint, and a fourth proposes a test with reopen conditions.

Contribution-oriented maps help a person or model decide which passage matters, what a new section must accomplish, where evidence belongs, and whether two regions duplicate the same work.

6. Structure forms from below

Bottom-up formation lets incomplete thinking remain incomplete. Repeated distinctions can become a movement. Several movements can reveal a document. Related documents can expose a project architecture. Completed work can provide evidence for a reusable method.

The final taxonomy is not required before discovery begins. Structure earns its form through the material it must carry.

7. Purpose supplies bearing from above

Top-down bearing tells the system which document should receive a finding, whether a new movement is warranted, what one passage must contribute, which source resolution is needed, and whether local polish serves or distracts from the larger purpose.

The direction remains open to bottom-up correction. A discovery can challenge the document spine, divide one artifact into two purposes, or alter the project route.

8. Composition moves among levels

Content composition includes more than sentence production. It integrates thinking into an artifact, separates conflated movements, merges duplicate contributions, changes sequence, creates or branches documents, composes findings across sources, revises project relationships, and promotes methods after evidence review.

Each operation moves between levels of the hierarchy. A local revision can require structural ascent. A high-level plan can require descent into exact language. Composition is the intelligence that keeps those changes related.

9. Exact descent and semantic ascent

Precision pulls reasoning downward; consequence pulls it upward. A disputed claim may require exact source inspection. A single wording change may require ascent because the definition appears throughout the document. A project review may start with outcomes and descend only where completion depends on specific evidence.

This two-way movement protects both local fidelity and whole-system coherence. It describes resolution inside the artifact field, not a universal route through which every judgment in AIOS must pass.

10. Writing and structure are distinct altitudes

Writing works on language, argument, narrative formation, transitions, passages, and whole-document expression. Structure works on movement order, boundaries, artifact architecture, cross-file relationships, new-section or new-document proposals, pending commissions, and project consequences.

Structure can identify that language must be written without silently writing it. It records a bounded commission or invokes an explicitly authorized operation. Writing can improve a passage without acquiring authority to reorganize the entire document field.

The distinction reduces cognitive interference and makes scope visible.

11. Accepted and proposed structure remain separate

A structural projection shows the accepted spine beside proposed moves, annotations, tensions, pending writing, and possible branches. Exploration remains valuable because its proposals do not overwrite current ground.

observation
  ≠ proposal
    ≠ person authorization
      ≠ file effect
        ≠ accepted new spine

Each transition has its own standing and authority. This protects the document from unintended mutation while allowing structural thought to remain ambitious.

12. The single-document composition loop

The loop separates source custody, editorial judgment, exact grounding, prose production, whole-artifact review, and acceptance. A writer receives a bounded movement rather than implicit authority to reinterpret the source. A local edit returns to holistic review before the spine changes.

The boundary

The loop names obligations that any compliant composition path must satisfy. It does not prescribe one runtime topology or one model call per step.

13. The cross-file composition loop

Cross-file work first interprets the project field, then plans discrete movements, resolves exact origins and destinations, checks authority, applies destination-bound operations, reviews every affected artifact, records lineage, and refreshes projections.

Operation semantics remain distinct: integrate, copy, move, reorganize, create, branch, annotate, and supersede. A valuable destination never implies permission to remove the source. A move completes the destination write and validation before any authorized source removal.

The boundary

The cross-file loop states semantic, authority, custody, and reintegration obligations. The named records and projections preserve those obligations without making one implementation sequence universally mandatory.

14. Document boundaries are cognitive boundaries

A new artifact is warranted when one file carries more than one governing purpose, a branch requires independent evidence or method, audiences or completion criteria diverge, one line must remain open while another lands, necessary depth would destroy coherence, or a reusable method needs separation from case-specific work.

Length alone does not justify fragmentation. The boundary follows purpose, evidence, audience, method, and completion.

15. Hierarchy preserves contradiction

Internal consistency makes disagreement legible rather than eliminating it. A project can retain competing hypotheses, challenging sources, a rejected but still material alternative, superseded versions, role-specific accounts, minority positions, and open tensions requiring later judgment.

The hierarchy locates the contradiction and its consequences. It prevents disagreement from being averaged into a false consensus or scattered across files without relationship.

16. Completion occurs at the scale being judged

Word count and activity never prove completion. Evidence and fulfilled purpose do.

17. Hierarchy becomes an interface for thought

The person moves among exact text, movement map, document synthesis, project relationships, plans, pending work, source and decision lineage, and corpus or expertise views.

The interface preserves the experience of working with meaningful documents. Structural machinery appears when it helps the person make a real judgment and recedes when exact language is the object of attention.

18. Failure conditions and evaluation

Flat-file reasoning

Flat-file reasoning loads complete artifacts without their contribution or relationship. Exact prose arrives without the hierarchy that explains why each region matters.

Outline absolutism

Outline absolutism prevents new thought from changing structure. An earlier arrangement governs even after the material reveals a better architecture.

Local-edit myopia

Local-edit myopia treats the target passage as the whole consequence field. The edit can satisfy its local instruction while breaking definitions, seams, argument, or purpose elsewhere.

Whole-document rewriting

Whole-document rewriting exceeds bounded authority by treating one requested change as permission to replace the artifact. Existing value, source fidelity, and unaffected movements become collateral material.

Project monoliths

Project monoliths force distinct purposes, evidence paths, audiences, or completion conditions into one artifact or context. The file grows while its cognitive boundary disappears.

Length-driven fragmentation

Length-driven fragmentation creates files because an artifact is large rather than because a distinct purpose warrants separation. Custody multiplies without clarifying contribution.

Folder-semantic confusion

Folder-semantic confusion mistakes neighboring paths for shared meaning or authority. Custody becomes a substitute for explicit relationship, standing, and contribution.

Evaluation tests whether movement maps outperform headings for contribution prediction, exact descent improves source fidelity, structural planning reduces unrelated prose changes, document-boundary recommendations improve coherence, cross-file composition preserves provenance, accepted/proposed separation reduces unintended mutation, and the Fractal Seed transfers across scales without imposing visible uniformity.