Foundations for a Functional Model of Intelligence: A Literature Synthesis with Strongly-Typed Predictions and a Witness-Based Validation Protocol

We synthesize evidence from cognitive psychology, neuroscience, organizational science, philosophy of mind, artificial intelligence research, and science studies to ground claims underlying a Functional Model of Intelligence (FMI). The synthesis addresses twelve key propositions spanning knowledge production, collective coherence, conceptual representation, dual-process reasoning, meta-cognition, phase transitions, transmission barriers, perspective-taking, AI evaluation, coordination failures, recursive self-reference, and information-theoretic constraints. Each proposition finds substantial support in existing literature. This expanded version introduces three contributions beyond standard literature synthesis. First, we identify a reasoning style asymmetry between constraintanchored and empirically-anchored research traditions that has systematically impeded integration of these findings into unified frameworks, and we argue that this asymmetry is itself predicted by the FMI. Second, we derive twelve domainspecific empirical predictions from the FMI’s COAR constraint algebra—one per proposition—that differ from predictions generated by dominant frameworks in each domain, and we strongly type each prediction according to the derivation confidence and domain expertise required for operationalization. Third, we specify a witness-based validation protocol adapted from the Cognitive Near-Singularity (CNS) framework and apply it to validate one prediction within the paper itself. The strongly-typed prediction structure serves a dual purpose: it provides falsifiable empirical targets for each domain, and it identifies specific areas where domain specialists are needed to operationalize and test predictions that the constraint algebra derives but that require expertise beyond the framework’s author to implement. The paper thereby functions simultaneously as a scientific contribution and as a structured collaboration invitation.