Declarative, Procedural, and Conditional Knowledge

Mastering any skill divides into three jobs: knowing what to do, executing it without conscious thought, and judging when to deploy it. Each job runs on a different knowledge type — declarative for the what, procedural for the how, conditional for the when and why — and each type has a different best teacher. Another person can only ever teach you the first one. Explanation, however good, delivers declarative knowledge and nothing else; the other two are built exclusively by your own experience. Knowing what to do is roughly 10% of mastery. The remaining 90% is constructed through practice, which is why a learner can understand a technique completely and still be unable to perform it.

Use the triad as a diagnostic: before choosing how to practice or test anything, classify which type the work actually demands and which type is failing.

Where the boundaries sit

• Declarative is the foundation. Facts, concepts, relationships — anything you can explain and understand. Every other knowledge type builds on top of it, and content-heavy subjects (history, biology, medicine, theory of any kind) are almost entirely declarative.
• Procedural means automatic. Driving, native-sounding pronunciation, fluent code — execution so practiced it runs without thinking about the steps. It develops slowly, only through doing, and once it locks in, the declarative scaffolding underneath can even fade away.
• Applying knowledge is still declarative. Using disease facts to diagnose a patient and draft a treatment plan runs on conscious selection, reasoning, and sequencing at every step — declarative work end to end, however applied it feels. Procedural begins only where conscious control ends.
• Conditional develops for free. Knowing when and why to deploy what you know is strictly a subset of declarative knowledge — you can explain it on demand. No dedicated techniques exist for it; building declarative and procedural knowledge well across varied contexts, challenge types, and time pressures produces it as a by-product.

Experience teaches the other two

• Move at working understanding. Engage the theory deeply — notes, analogies, connections to what you already know, critique of the technique itself — then attempt the skill as soon as you know enough to discover your first mistakes. Certainty before the first attempt is impossible, and chasing it through more clarifying questions just postpones the only teacher that works.
• Fail fast, fail safe. Find mistakes early, when each one costs nothing. A botched study session six months before an exam is a free data point; the same discovery in the final week concentrates all the risk into time you no longer have. Small early errors pre-empt the large late ones.
• Hunt your five to seven tendencies. A skill may offer 300–400 possible mistakes, but any individual makes only about five to seven of them. Practice exists to surface your personal subset; trying to pre-empt all 400 with what-if questions overwhelms working memory and stalls the start.
• Run the loop. Try, reflect against your declarative model, refine, retry. Strong theory converts a failed attempt into a targeted, answerable question instead of a vague plea for help — and cycling the loop across different contexts and pressures is what grows the conditional layer.
• Borrow outside eyes. Some tendencies are habits trained in another domain that leak in invisibly; you cannot see them from the inside, while an observer spots them in minutes.

Diagnosing the real bottleneck

• Suspect declarative first. Even in procedural-leaning subjects like maths, most people’s bottleneck is approach choice — which method, in what order, for this context — and approach choice is declarative. The learner who solves familiar practice problems but collapses when the problem arrives in context has a conceptual gap that more practice volume cannot fill.
• Test with a complex challenge. Attempt something genuinely hard, then classify the error. Wrong approach, strategy, logic, or argument structure: declarative gap. Right approach with botched execution — correct method, wrong arithmetic; sound architecture, syntax bugs; coherent argument, clumsy prose: procedural gap.
• Match practice to the diagnosis. Retrieval methods split along the same line, so the diagnosis decides which methods earn your time and where the gaps you find actually live.
• Declarative compounds in value. Pure execution is increasingly cheap to automate; the approach, strategy, and judgment layer is the part of expertise that stays expensive.

Links into the system

Supplies the knowledge-type definitions behind the Interleaving Table and the type variable in Spaced Interleaved Retrieval; the declarative-bottleneck diagnosis is applied in depth by Interleaving for Complex Problem Solving, the declarative side is graded by Knowledge Mastery From Recognition to Usable Knowledge, and the theory-to-practice pacing lives in Rapid Skill Acquisition and Kolbs Experiential Cycle. The reason procedural knowledge runs automatically and resists verbal report is that proceduralized skill executes outside the deliberate workspace; the cognitive mechanism, and its model-side mirror, is in Global Workspace and J-space.