Boundary comparison matrix

Reading the matrix

Coverage codes describe source-supported field scope:

• S — strong/direct coverage.
• P — partial, contextual, or supporting coverage.
• W — weak or normally outside the field's central scope.
• ? — source set is insufficient for a reliable judgment.

These codes do not measure quality, frequency, or outcome superiority. A field can cover a mechanism without implementing it in the same way as FE.

Abbreviations:

• SE: systems engineering.
• ME: method and situational method engineering.
• RE: requirements engineering.
• KE: knowledge and ontology engineering.
• DA: decision analysis.
• CY: cybernetics.
• QE: quality engineering and quality management.
• OL: organizational learning and knowledge management.
• AO: agent orchestration and multi-agent systems engineering.
• DSR: design science research.

Frozen-dimension matrix

Steelman comparator profiles

Systems engineering

Systems engineering is the strongest lifecycle and integration comparator. It explicitly handles complex human, hardware, software, and conceptual systems; integrates stakeholder needs, requirements, architecture, design, verification, validation, transitions, operation, and retirement; and includes decision analysis, configuration, information management, risk, measurement, and tailoring. Its process standards are method-neutral, so a framework-focused profile can exist inside SE without creating a new discipline. [EV-FEB-001–007, EV-FEB-018]

Strongest subsumption claim: an analytical framework can be treated as a conceptual system or enabling system, and FE can be a tailored lifecycle profile for that class of system.

Meaningful residual: general SE does not make analytical frameworks its default unit of analysis or supply the repository's specific genome, identity/capability, and human-agent handoff conventions.

Method and situational method engineering

Method engineering directly targets construction of methods and supporting tools. Situational method engineering supplies reusable fragments or chunks, metamodels, assembly, tailoring to project context, and quality evaluation. SPEM demonstrates that method content and process can be represented in a standard metamodel. [EV-FEB-008–010]

Strongest subsumption claim: most of FE's construction, decomposition, adaptation, and context-fit claims are method-engineering claims with “framework” used as the artifact label.

Meaningful residual: FE includes descriptive, diagnostic, taxonomic, and decision frameworks that are not always executable methods. It also binds evidence, confidence, contradiction, provenance, and machine handoff more tightly than the reviewed ME sources require.

Requirements engineering

Requirements engineering covers elicitation, analysis, specification, validation, lifecycle information, traceability, and change for stakeholder needs and requirements. [EV-FEB-011]

Strongest subsumption claim: FE's purpose/scope, constraints, acceptance criteria, traceability, and change controls can be represented as requirements-engineering work products.

Meaningful residual: requirements are inputs and constraints, not the entire framework artifact or its explanatory and procedural semantics.

Knowledge and ontology engineering

Knowledge and ontology engineering cover explicit conceptualization, semantics, constraints, interoperability, inference, and tradeoffs among representational profiles. PROV-O directly covers interoperable provenance and domain specialization. [EV-FEB-012–013]

Strongest subsumption claim: framework identity, concepts, relationships, lineage, and machine-readable semantics can be modeled using existing knowledge-representation and provenance methods.

Meaningful residual: a semantically valid knowledge model does not by itself establish operational procedure, decision utility, empirical validity, or organizational adoption.

Decision analysis

Decision analysis supplies principled problem framing, alternatives, uncertainty, preferences, value, tradeoffs, sensitivity, and decision quality. It distinguishes normative guidance from descriptive behavior. [EV-FEB-016–018]

Strongest subsumption claim: FE frameworks intended to improve decisions must ultimately be evaluated as decision interventions, including whether they improve decision quality net of elicitation and analysis burden.

Meaningful residual: many frameworks organize explanation, diagnosis, learning, or coordination without directly choosing among alternatives.

Cybernetics

Cybernetics supplies feedback, communication, control, regulation, adaptation, and self-organization across technical, biological, and social systems. [EV-FEB-023]

Strongest subsumption claim: FE's observe–evaluate–adapt loop and framework-evolution language instantiate established feedback and control ideas.

Meaningful residual: cybernetics does not prescribe the repository's artifact schema, evidence ledger, lifecycle information, or research governance.

Quality engineering and quality management

Quality disciplines supply measurement, experimental design, process characterization, comparison, monitoring, reliability, controlled improvement, customer focus, corrective action, and continual improvement. [EV-FEB-019–020]

Strongest subsumption claim: FE evaluation, defect classification, validation gates, and improvement cycles are applications of established quality and statistical engineering.

Meaningful residual: QE needs a product/process definition; it does not make the semantic structure of analytical frameworks its primary object.

Organizational learning and knowledge management

Knowledge management covers establishing, maintaining, reviewing, and improving organizational knowledge systems. Double-loop learning adds the revision of governing assumptions and policies, not only corrective action within them. [EV-FEB-021–022]

Strongest subsumption claim: FE's durable repositories, contradiction retention, retrospectives, governance updates, and evolution are mechanisms of organizational knowledge and learning.

Meaningful residual: OL/KM do not necessarily provide an engineering grammar for individual framework artifacts or matched artifact evaluation.

Agent orchestration and multi-agent systems engineering

FIPA standardizes agent identities, services, communication, and lifecycle. Gaia models multi-agent systems as computational organizations with roles, interactions, and organizational rules. Current NIST work shows that agent identity, security, and interoperability remain active standards concerns. [EV-FEB-024–026]

Strongest subsumption claim: FE's provider manifests, role boundaries, handoffs, registry discovery, and execution records apply established agent lifecycle and organizational-design concerns.

Meaningful residual: these sources do not define the analytical framework as the artifact under study or validate LLM-based research execution.

Design science research

Design science builds and evaluates artifacts to create useful knowledge. FEDS explicitly requires evidence of artifact utility, quality, and efficacy, allows artificial and naturalistic evaluation, and asks why an artifact works, when, and with what side effects. [EV-FEB-014–015]

Strongest subsumption claim: FE frameworks are design artifacts, and FE's research program should be judged as a domain-specific DSR program.

Meaningful residual: DSR is artifact-generic and does not supply FE's particular framework grammar, provenance conventions, or operating profile.

Mechanism-level subsumption test

Full-subsumption search

Individual-field full subsumption

Not established. Systems engineering comes closest, particularly because it can include conceptual systems and full lifecycle processes, but it does not make cross-domain analytical frameworks the default unit of analysis. Method engineering comes closest at the construction and tailoring level, but its canonical literature is method- and software-oriented. Design science comes closest at artifact build/evaluate logic, but remains artifact-generic.

Collective functional subsumption

Substantially supported. Every FE mechanism identified in the current repository has a plausible established home in one or more adjacent fields. The literature review found no clearly non-subsumed causal mechanism. FE's remaining candidate contribution is the deliberate integration of those mechanisms around a neglected unit of analysis and a portable record model.

Consequence

The result does not justify terminating the research program because an integrated profile can still be useful. It does require stopping any unqualified claim that FE is already a distinct discipline. The burden of proof moves from “the pieces are assembled” to “the integration produces measurably better net outcomes than matched alternatives.”

Institutional maturity comparison

Boundary judgment

Confidence update for FEH-001

• Distinct-discipline interpretation: Low → Very Low.
• Coherent integrated-profile interpretation: unassessed → Moderate.
• Incremental utility over matched alternatives: remains Very Low.
• Generalizability beyond the repository: remains Low.

The update is driven by strong overlap evidence and the absence of non-subsumed mechanisms. Confidence does not move higher for the profile because literature and executable infrastructure cannot substitute for comparative outcome evidence.