A.5 - Open-Ended FPF Kernel and Extension Layering

Status. Informative. This section defines no dedicated “module” subsystem. Enforceable boundary discipline lives in A.6.0 U.Signature and A.6.1 U.Mechanism, with guard‑rails in E.5.3 (Unidirectional Dependency) and E.10 (LEX‑BUNDLE stratification).

A.5:1 - Problem frame

FPF helps practitioners combine shared conceptual grounds and reusable methods with suitable domain methods. FPF can support that use across changing disciplines only if it:

(i) remains stable and self‑consistent over multi‑decade timespans;
(ii) invites, rather than resists, the continual influx of new disciplinary knowledge; and
(iii) allows multiple, even competing, explanatory lenses to coexist without forcing a “winner‑takes‑all” unification.

Historically, grand “total” ontologies—Aristotle’s Categories, Carnap’s Logical Construction of the World, Bunge’s TOE—failed precisely because each tried to embed every domain’s primitives directly into a single monolith. Once the monolith cracked under domain pressure, the whole edifice became unmaintainable.

A.5:2 - Problem

If FPF were to let domain‑specific primitives creep into its Kernel, two pathologies would follow:

Pathology	Manifestation	Breach of Constitution
Kernel Bloat	Adding domain-specific root U-kinds or local type vocabularies to the Kernel increases its size and review burden.	Violates C-5 Ontological Parsimony; erodes P-1 Cognitive Elegance.
Conceptual Gridlock	Conflicting axioms (deterministic thermodynamics vs. indeterministic econ‑metrics) must fight for space in the same namespace.	Breaks C‑3 Cross‑Scale Consistency; triggers chronic DRR deadlock.

A minimal, extensible design is therefore mandatory.

A.5:3 - Forces

Force	Tension
Stability vs. Evolvability	Stable core needed for trust ↔ constant domain innovation needed for relevance.
Universality vs. Specificity	Single kernel language ↔ rich disciplinary idioms.
Parsimony vs. Coverage	Few primitives keep reasoning elegant ↔ framework must still model energy budgets, epistemic uncertainty, agentic goals.

A.5:4 - Solution

FPF’s modularity is declarative, not “callable”: pattern texts publish law‑governed declarations (vocabulary + laws + applicability) that can be reused and specialised.

To keep the Kernel open‑ended, use the following boundary rules:

Kernel minimality (C‑5). Domain knowledge stays outside the Kernel by default; it enters as extension vocabularies and laws.
Boundary packaging via U.Signature (A.6.0). For reusable declaration bundles admitted as signatures under A.6.0, expose actual declaration dependencies in an explicit SignatureManifest (imports, provides).
Dependency vs specialisation are separate relations. imports forms a dependency DAG constrained by E.5.3; refinement/extension (⊑, ⊑⁺) is expressed separately (for mechanism declarations, see A.6.1:4.8; use C.29 when a mathematical morphism is claimed) and should not be conflated with imports.
Registry references stay references. Bridge ids, policy‑ids, and edition‑ids (Part F) are registry identifiers: they are cited/pinned where needed, not treated as exported symbols in provides.