Cognitive Architecture as Relational Operations: A Database-Theoretic Reformulation of the Cognitive Action Model

We propose a database-theoretic reformulation of the Cognitive Action Model (CAM) that grounds personality differences in relational operations over two fundamental information structures: a Grand Premise table (general rules, beliefs, policies) and a Small Premise table (current situational states, observations). Action emerges from JOIN operations between these tables, with action readiness determined by result set cardinality. Individual differences emerge from four orthogonal architectural parameters operating at different levels of the cognitive system. First, **regime** (a universal system-wide switch) determines tuple cardinality: Single-row regime maintains one canonical row per logical entity in all tables, while Multi-row regime maintains multiple candidate rows per logical entity in all tables. This global switch cannot be set differently for different tables—the brain operates in one regime at a time. Second, **premise emphasis** determines traversal efficiency: Grand Premise emphasis (introversion) optimizes traversal within the Grand table, while Small Premise emphasis (extraversion) optimizes traversal within the Small table. Third, **integrity type** determines constraint enforcement: Entity Integrity emphasis (S/F functions) prioritizes context-specific accuracy, while Referential Integrity emphasis (N/T functions) prioritizes crossentity logical consistency. Fourth, **traversal mode** interacts with regime: Multi-row regime enables efficient row-to-row traversal within tables (versioning focus), while Single-row regime enables efficient table-to-table traversal across entities (sampling focus). The regime-cardinality mechanism explains action readiness and behavioral phenotypes. Single-row regime produces 1×1 JOINs with single executable results but low match probability, explaining Ptypes' "act-done-chill" pattern: sparse vector formation, immediate execution, idle periods. Multi-row regime produces N×M JOINs with multiple candidate results and high match probability, explaining Jtypes' "think-a-lot-yet-act-less" pattern: continuous candidate generation, deliberation overhead (search/sort/pinpoint), eventual commitment. The traversal mode explains functional differences: Multirow functions (Te, Fe, Ni, Si) exhibit "sticky" deep engagement with single contexts through row-to-row traversal, while Single-row functions (Ne, Se, Ti, Fi) exhibit "jumpy" broad sampling across contexts through table-to-table traversal. This framework generates falsifiable predictions about decision latencies, context-switching costs, cognitive load under different regime demands, and learning trajectories. Critically, it transforms personality typology from descriptive taxonomy to computational architecture specification with precise database semantic