CIGTI — Cognitive Integration–Generativity Taxonomy of Intelligence

CIGTI — the Cognitive Integration–Generativity Taxonomy of Intelligence — is a formal conceptual theory that represents intelligence as an evidence-bounded, non-scalar, multidimensional cognitive property rather than a unitary rank. The framework distinguishes Developed Cognitive Capacity (DCC), Cognitive Integration, evaluative domain scope, characteristic cognitive operations, and three co-instantiable person-level dimensions: Specialized Intelligence (D1), Synthetic Intelligence (D2), and Generative Intelligence (D3). DCC describes the developed configuration of cognitive resources, representations, operations, regulatory capacities, and composite repertoires arising through the interaction of baseline architecture, exposure, learning, training, practice, and experience. CIGTI further distinguishes Feature Standing, DCC Configuration Standing, and Target-Relative DCC Sufficiency from scope-specific coordinated enactment. D1 represents specialized structural understanding and integrated cognition within an explicit Evaluable Domain Scope (EDS). D2 represents genuine higher-order synthesis across materially distinct constituent domains while preserving constituent adequacy, structural integration, reliability, and person-specific attribution. D3 represents attributable intellectual generation evaluated through a declared Generative Claim Object, a Novelty Reference Frame, structural qualification conditions, and generative attribution. Version 4.3 additionally formalizes the Diagnostic Classification Hypothesis (HYPO), a non-classificatory and evidence-bounded diagnostic estimate for contexts in which available evidence is insufficient for qualified dimensional classification. It also develops an adaptive diagnostic architecture, Expression–Structure Separation, alternative representational routes, evidence-sensitive updating, and model-agnostic AI-assisted assessment. CIGTI is currently presented as a formal conceptual and empirically testable research architecture. Its constructs, classes, thresholds, and evidence architecture have not yet undergone psychometric or clinical validation. The framework is explicitly designed to remain open to operational testing, adversarial case analysis, empirical revision, simplification, or falsification.