Home/Library/Papers/A Formal Framework for Core and Deep Rationality

A Formal Framework for Core and Deep Rationality

This page anchors the project’s central distinction between rational skill and rational integrity.

Related Research Papers

The paper cluster behind Credencing

These Academia.edu papers support the public site as a connected research program: formal modeling, two-gap diagnosis, warranted uncertainty, institutional credencing, AI-mediated inquiry, scalar language, Bayesian audit methods, and motivated-cognition diagnosis.

Foundational modelA Formal Framework for Core and Deep RationalityDefines Objective Evidence (E0), Perceived Evidence (EP), Assigned Credence (CA), Deep Rationality (SD), Calculation Error (DE), and Core Irrationality (IC).Two-gap modelThe Two-Gap Theory of Irrationality: Evidence Perception, Belief Integrity, and Diagnostic RepairSeparates E0-EP evidence-perception failures from EP-CA belief-integrity failures, giving the framework its main diagnostic repair map.Confidence disciplineWarranted Uncertainty: The Discipline of Not Knowing in a Credence-First EpistemologyDefends confidence bands, warranted slack, and honest uncertainty so credencing does not collapse into false precision.Biased expert diagnosisIntelligence as Rationalization Engine: Deep Rationality, Core Rationality, and the Biased ExpertExplains how high Deep Rationality can reduce evidence-processing error while still defending an assigned credence the agent is motivated to keep.Institutional extensionInstitutional Credencing: Official Confidence, Evidence Pipelines, and Organizational Belief IntegrityExtends E0, EP, CA, and SD to organizations through evidence pipelines, public confidence, dissent channels, and update trails.AI-mediated credencingAI as Deep Rationality Prosthetic: Evidence Processing, Delegated Credence, and Core Rationality in AI-Mediated InquiryShows how AI can improve evidence processing while leaving final credence vulnerable to motivated uptake, false precision, and delegation errors.Conceptual bridgeCore RationalityIntroduces core rationality as evidence-proportionate credence, distinct from mastery of formal probability, logic, and decision theory.Credence-first entry pointEvidence-Proportioned Belief: A Credence-First Reframing of Faith and RationalityFrames belief as confidence scaled to evidence rather than a binary status or performance of conviction.Credence-first epistemologyCredence First: Against Plantinga's Warrant as an Epistemic UpgradeArgues that graded credencing is a better epistemic target than binary knowledge or warrant, especially when certainty can become miscalibrated.AI over-delegation warningHumanity as a Middle Child in the Age of AI: Patterns of Delegation and Epistemic AtrophyWarns that AI can weaken human judgment when people outsource inference without retaining responsibility for credence.Language and calibrationThe Gravity of Grammar: Binary Inertia and the Distortion of Epistemic CalibrationExplains how ordinary assertion can compress scalar confidence into binary belief-talk, distorting epistemic self-reporting.Linguistic repairBeyond the Binary: Linguistic Evolution and the Transition to Credence-Based DiscourseDevelops a scalar lexicon, Bayesian update protocol, and discussion norms for keeping belief revision explicit and socially usable.Applied Bayesian auditThe Miracle Audit: A Probabilistic Framework for Differentiating Anomalies from Supernatural InterventionsApplies probabilistic discipline to extraordinary claims by separating anomaly, causal link, evidential rigor, and burden of proof.Motivated cognitionThe Architecture of Explanatory Satisfaction and the Affective Veto Against Stochastic RealityConnects explanatory hunger, agency bias, and affective resistance to stochastic explanations with the site's Core Rationality (IC) diagnosis.Operator knowledgeFoundational Choices for the AutodidactDefends transferable reasoning operators: base rates, causal inference, measurement discipline, calibration, and feedback-rich practice.Methodological naturalismThe Asymptote of the Natural: Inductive Density and the A Posteriori Justification of Methodological NaturalismFrames naturalistic inquiry as a Bayesian and risk-management strategy grounded in projectibility, constraint, and historical evidential yield.Ontology and inductionThe Threshold of the Real: Inductive Density and the Emergence of the Human OnticExplains objecthood and category confidence as a function of predictive reliability and cognitive cost.Reification mechanismThe "Ontic Snap" as a Cognitive Phase TransitionModels the moment noisy input becomes a stable “thing,” helping diagnose premature closure and overconfident categories.Category humilityMereological Indigestion: Why We Cannot Divvy Up the BlobChallenges overconfident partitions of reality and supports the site’s caution about reifying useful categories.Pedagogical scaffoldingUseful Fictions Common in Philosophical PedagogyClarifies when simplified teaching models help inquiry and when they become overconfident substitutes for mature understanding.Attribution calibrationAgainst the Great Man: A Systems-Theoretic Critique of Individual AttributionUses a control coefficient to calibrate praise and blame in noisy systems where individual agency may be weak.

View full Academia.edu profile

Source Paper

This page is a simplified guide to the paper. The source document is linked below; the guide here keeps the main idea accessible and tied back to the site’s model.

Open on Academia.edu

Low-Register Abstract

This paper gives the site its basic model. It says that a person can go wrong in two different places: they may fail to see the evidence well, or they may see the evidence well enough and still assign the wrong level of confidence.

Deep Rationality (SD) is about evidence-reading skill. Core Rationality is about whether confidence stays honest once the evidence has been perceived. The model turns that distinction into variables, gaps, and archetypes that can be visualized in the interactive lab.

Central Thesis

The formal framework distinguishes Deep Rationality from Core Rationality. Deep Rationality concerns the tools that allow an agent to perceive evidential support accurately. Core Rationality concerns whether the agent assigns credence according to that perceived support.

The framework is designed to show why intelligence, education, and technical sophistication do not automatically produce epistemic integrity.

Formal Variables

The central variables are Objective Evidence (E0), Perceived Evidence (EP), Assigned Credence (CA), and Deep Rationality (SD). Calculation Error (DE) measures the Objective Evidence to Perceived Evidence gap (E0-EP). Core Irrationality (IC) measures the Perceived Evidence to Assigned Credence gap (EP-CA).

Public Use

The paper gives formal structure to the public model, while the site turns that structure into explanations, examples, exercises, and applications. The two should be read together: the paper supplies precision; the site supplies usability.