Pattern Recognition vs. Explicit Rules

Novices follow explicit rules. Experts pattern-match. This distinction is not about effort — it is about the cognitive architecture that develops through experience. The novice checks the checklist consciously. The expert scans the situation holistically and arrives at the same answer in a fraction of the time, by a route they cannot fully trace. The implication for knowledge extraction: if you only capture the rules, you have captured the novice model. The expert model is built from patterns, and patterns require a different capture technique.

1. The Dreyfus Model The Dreyfus skill acquisition model describes five stages: novice, advanced beginner, competent, proficient, expert. At the novice level, all behavior is rule-governed. At the expert level, rules are invisible — the expert sees the situation holistically and acts on pattern recognition. Extraction techniques must be calibrated to the expert level, not the novice level.
2. Externalizing Patterns Pattern recognition cannot be extracted through rule enumeration. It must be externalized through scenario work: present the expert with cases, ask them to classify, sort, or respond — then probe their reasoning. "Why did you classify this as high-risk?" is more productive than "what makes a deal high-risk?" The case forces the pattern to the surface.
3. Anomaly Detection Experts are extraordinarily sensitive to anomalies — deviations from the expected pattern that novices miss entirely. One of the most effective extraction techniques is showing experts cases with deliberate anomalies and asking them to spot what is wrong. The anomalies they catch, and the ones they miss, define the boundary of their expertise.
4. The Knowledge Map The endpoint of this module is not a process document. It is a knowledge map: a structured representation of what the expert knows, what they do not know they know, and what conditions trigger which patterns. The map is the prerequisite for Module 2's extraction techniques.