Definition
The System 1 / System 2 framework — popularized by Daniel Kahneman in Thinking, Fast and Slow (2011) but with intellectual roots in earlier dual-process theories from Wason, Evans, Sloman, Stanovich, and Frederick — distinguishes two modes of cognitive processing differing on multiple dimensions. System 1 is fast, automatic, parallel, effortless, associative, affect-laden, and largely opaque to introspection; System 2 is slow, controlled, serial, effortful, rule-based, affect-neutral, and consciously accessible.
System 1 is computationally cheap and operates continuously: it generates impressions, intuitions, and immediate associations that flow into consciousness without permission. System 2 is computationally expensive — bounded by working memory capacity, dependent on dorsolateral PFC integrity, and fatigable. System 2's principal role is monitoring and overriding System 1 outputs when those outputs are likely to be wrong; absent active monitoring, System 1 outputs become behavior with minimal further processing.
The framework is heuristic rather than literally architectural. Critics (Keren and Schul, 2009; Osman) have argued that the binary cleanly maps onto neither neuroanatomy nor distinct mechanisms, and the systems are best understood as clusters of correlated properties rather than discrete modules. Kahneman himself in his later writing (Kahneman, Sibony, and Sunstein, Noise, 2021) acknowledged the framework's limitations while preserving its diagnostic utility. The empirical generalization survives even if the architectural simplification does not: rapid responses display systematic biases that slower, deliberate responses can correct, and the conditions that suppress slow processing are well-characterized.
Context
The System 1 / System 2 framework is the organizing principle of epistemic-manipulation taxonomy. Nearly every kompendium technique can be characterized in terms of how it interacts with the dual-process division:
- Cognitive overload techniques (EA-1.x) raise cognitive load until System 2 is unavailable, leaving System 1 unmonitored
- Affective techniques (EA-2.x) inject emotional content that biases System 1 outputs in directions favorable to the manipulator
- Schema techniques (EA-1.4 family) install or activate corrupted patterns that System 1 will subsequently retrieve as default responses
- Authority techniques (EA-2.4 family) trigger System 1 social-deference heuristics that System 2 must explicitly override to disregard
- Time pressure techniques (EA-1.3 family) deny the time required for System 2 engagement
The Decision Engine is structured around the assumption that System 2 is the scarce resource that must be protected, paced, and supplemented by external structure. Decisions are presented in a form that does not require holding more in mind than working memory permits; reasoning chains are externalized rather than internalized; pace is controlled to prevent System 2 exhaustion. The platform's epistemic stance is that System 1 outputs are valuable input but must always be validated by adequately-resourced System 2 review before being treated as decisions.
Interactive: Dual-Process Engagement
System 1 (fast, heuristic) and System 2 (slow, analytic) compete to handle each task. Task type biases which system fits naturally; cognitive load taxes System 2 specifically — at high load, System 1 wins by default regardless of task fit. The bars show response time and accuracy under current conditions.
0 = heuristic-friendly · 100 = strictly analytic
Working-memory pressure suppresses System 2 specifically
See Also
- Cognitive Load — Suppresses System 2
- Working Memory Limit — Bounds System 2
- Dorsolateral PFC — System 2 substrate
- Ventromedial PFC — System 1 substrate
- Mental Models — System 1 retrieves these
- Inhibitory Control — System 2 overrides System 1
- Kompendium — EA technique taxonomy
Connect & Contribute
Created by Tomáš Korcak (korczis) | Open Source under GHL