VIEW THIS AS

Auto mode follows the Route Engine until you choose a viewpoint.

YOU ARE HERE

ROUTE CHECK

CONNECTED TO

WHAT NEXT

Use the canonical route for this room, or HELP if you are unsure.

Explaining to Learn and the Feynman Technique | How Self-Explanation Builds Usable Understanding

Three secondary students review open books and notes together in a classroom.

eduKateSG · Study & Learning Methods Hub · Explaining to Learn · 1 October 2026

The Feynman Technique is popular because it sounds almost too simple: choose a concept, explain it in plain language, notice where the explanation breaks, repair the gap and repeat. The useful part is not the celebrity label or the promise of effortless mastery. It is the self-explanation mechanism underneath: making the learner generate reasons, connections and structure instead of merely recognising familiar words.

This guide owns WEP106 beneath eduKateSG’s Study & Learning Methods Hub. It does not duplicate the existing Self-Explanation mechanism owner. Instead, it turns that mechanism into a practical study protocol and explains when the popular Feynman-style approach helps, when it wastes time, and how to combine it with retrieval, worked examples, spacing and transfer.

Research supports self-explanation as a useful learning technique, but not without conditions. A review on constraints on self-explanation notes that explaining can improve some learning outcomes while reducing others when prompts direct attention away from information that also matters. A separate study on self-explanation versus additional practice asks an important efficiency question: explanation can help, but study time still has opportunity cost. The right use is therefore selective and task-aware rather than universal.

Self-explanation is not just talking to yourself

Self-explanation is not just talking to yourself is mainly about generating an explanation that links new information to causes, principles, prior knowledge or a problem-solving step. The learner should be able to point to the extra work being done by the explanation. What relation became clearer? Which assumption was exposed? Which missing prerequisite became visible? If the explanation adds no new structure, the activity may be fluent rehearsal rather than useful elaboration.

A practical example is explaining why multiplying both sides of an equation by the same number preserves equality. The example matters because self-explanation works best when it changes what the learner notices. It can make a hidden rule visible, connect a new idea to prior knowledge or force a choice that passive review would avoid. This is why explanation is often especially valuable around worked examples and transferable principles.

The common failure is repeating the worked solution in different words without explaining why the move is valid. The failure usually feels productive because language is being generated. But verbal activity is not the same as learning. A learner can speak for several minutes while repeating surface features, relying on the source or using vague phrases that conceal the exact gap. Explanation earns its place when it changes later performance.

The acceptance test is that the learner adds a reason, connection or inference that was not explicitly supplied. That test should come after some support has been removed. Close the notes. Change the example. Ask for a reason rather than a definition. Require the learner to choose a method. If the explanation cannot survive those conditions, the learner may have created a good performance of understanding rather than durable understanding itself.

The Feynman Technique is best treated as a prompt structure

The Feynman Technique is best treated as a prompt structure is mainly about using simple explanation to expose gaps rather than as a magic method. The learner should be able to point to the extra work being done by the explanation. What relation became clearer? Which assumption was exposed? Which missing prerequisite became visible? If the explanation adds no new structure, the activity may be fluent rehearsal rather than useful elaboration.

A practical example is trying to explain photosynthesis to a younger learner and noticing that ‘energy’ is still vague. The example matters because self-explanation works best when it changes what the learner notices. It can make a hidden rule visible, connect a new idea to prior knowledge or force a choice that passive review would avoid. This is why explanation is often especially valuable around worked examples and transferable principles.

The common failure is believing simple language automatically means deep understanding. The failure usually feels productive because language is being generated. But verbal activity is not the same as learning. A learner can speak for several minutes while repeating surface features, relying on the source or using vague phrases that conceal the exact gap. Explanation earns its place when it changes later performance.

The acceptance test is that simplification reveals where the learner cannot yet explain the mechanism. That test should come after some support has been removed. Close the notes. Change the example. Ask for a reason rather than a definition. Require the learner to choose a method. If the explanation cannot survive those conditions, the learner may have created a good performance of understanding rather than durable understanding itself.

Explanation must follow the learning goal

Explanation must follow the learning goal is mainly about aligning the prompt with what the learner needs to know or transfer. The learner should be able to point to the extra work being done by the explanation. What relation became clearer? Which assumption was exposed? Which missing prerequisite became visible? If the explanation adds no new structure, the activity may be fluent rehearsal rather than useful elaboration.

A practical example is asking why a statistical test is appropriate rather than merely defining its name. The example matters because self-explanation works best when it changes what the learner notices. It can make a hidden rule visible, connect a new idea to prior knowledge or force a choice that passive review would avoid. This is why explanation is often especially valuable around worked examples and transferable principles.

The common failure is using ‘explain it simply’ for every type of task. The failure usually feels productive because language is being generated. But verbal activity is not the same as learning. A learner can speak for several minutes while repeating surface features, relying on the source or using vague phrases that conceal the exact gap. Explanation earns its place when it changes later performance.

The acceptance test is that the explanation targets the decision, principle or relation that must later be used. That test should come after some support has been removed. Close the notes. Change the example. Ask for a reason rather than a definition. Require the learner to choose a method. If the explanation cannot survive those conditions, the learner may have created a good performance of understanding rather than durable understanding itself.

Prior knowledge controls what can be explained

Prior knowledge controls what can be explained is mainly about recognising that explanation cannot substitute for missing foundations. The learner should be able to point to the extra work being done by the explanation. What relation became clearer? Which assumption was exposed? Which missing prerequisite became visible? If the explanation adds no new structure, the activity may be fluent rehearsal rather than useful elaboration.

A practical example is repairing fraction equivalence before asking a learner to explain algebraic rational expressions. The example matters because self-explanation works best when it changes what the learner notices. It can make a hidden rule visible, connect a new idea to prior knowledge or force a choice that passive review would avoid. This is why explanation is often especially valuable around worked examples and transferable principles.

The common failure is forcing elaborate explanations when basic terms are still unfamiliar. The failure usually feels productive because language is being generated. But verbal activity is not the same as learning. A learner can speak for several minutes while repeating surface features, relying on the source or using vague phrases that conceal the exact gap. Explanation earns its place when it changes later performance.

The acceptance test is that the learner has enough knowledge to build connections. That test should come after some support has been removed. Close the notes. Change the example. Ask for a reason rather than a definition. Require the learner to choose a method. If the explanation cannot survive those conditions, the learner may have created a good performance of understanding rather than durable understanding itself.

Worked examples are fertile ground for self-explanation

Worked examples are fertile ground for self-explanation is mainly about making hidden expert decisions visible while studying a complete solution. The learner should be able to point to the extra work being done by the explanation. What relation became clearer? Which assumption was exposed? Which missing prerequisite became visible? If the explanation adds no new structure, the activity may be fluent rehearsal rather than useful elaboration.

A practical example is pausing after each line of a geometry proof to explain why it follows. The example matters because self-explanation works best when it changes what the learner notices. It can make a hidden rule visible, connect a new idea to prior knowledge or force a choice that passive review would avoid. This is why explanation is often especially valuable around worked examples and transferable principles.

The common failure is copying the example and calling that explanation. The failure usually feels productive because language is being generated. But verbal activity is not the same as learning. A learner can speak for several minutes while repeating surface features, relying on the source or using vague phrases that conceal the exact gap. Explanation earns its place when it changes later performance.

The acceptance test is that the learner can reconstruct the logic without the model answer. That test should come after some support has been removed. Close the notes. Change the example. Ask for a reason rather than a definition. Require the learner to choose a method. If the explanation cannot survive those conditions, the learner may have created a good performance of understanding rather than durable understanding itself.

Explaining why is different from recalling what

Explaining why is different from recalling what is mainly about using explanation to organise relations after basic facts are available. The learner should be able to point to the extra work being done by the explanation. What relation became clearer? Which assumption was exposed? Which missing prerequisite became visible? If the explanation adds no new structure, the activity may be fluent rehearsal rather than useful elaboration.

A practical example is recalling Newton’s second law and then explaining how force, mass and acceleration constrain each other. The example matters because self-explanation works best when it changes what the learner notices. It can make a hidden rule visible, connect a new idea to prior knowledge or force a choice that passive review would avoid. This is why explanation is often especially valuable around worked examples and transferable principles.

The common failure is using explanation to avoid memorising essential definitions. The failure usually feels productive because language is being generated. But verbal activity is not the same as learning. A learner can speak for several minutes while repeating surface features, relying on the source or using vague phrases that conceal the exact gap. Explanation earns its place when it changes later performance.

The acceptance test is that fact knowledge and relational understanding support each other. That test should come after some support has been removed. Close the notes. Change the example. Ask for a reason rather than a definition. Require the learner to choose a method. If the explanation cannot survive those conditions, the learner may have created a good performance of understanding rather than durable understanding itself.

Self-explanation can support transfer

Self-explanation can support transfer is mainly about drawing out principles that can survive surface changes. The learner should be able to point to the extra work being done by the explanation. What relation became clearer? Which assumption was exposed? Which missing prerequisite became visible? If the explanation adds no new structure, the activity may be fluent rehearsal rather than useful elaboration.

A practical example is explaining why a ratio method works before using it in a different context. The example matters because self-explanation works best when it changes what the learner notices. It can make a hidden rule visible, connect a new idea to prior knowledge or force a choice that passive review would avoid. This is why explanation is often especially valuable around worked examples and transferable principles.

The common failure is memorising the story details of one example. The failure usually feels productive because language is being generated. But verbal activity is not the same as learning. A learner can speak for several minutes while repeating surface features, relying on the source or using vague phrases that conceal the exact gap. Explanation earns its place when it changes later performance.

The acceptance test is that the learner recognises the principle when the surface features change. That test should come after some support has been removed. Close the notes. Change the example. Ask for a reason rather than a definition. Require the learner to choose a method. If the explanation cannot survive those conditions, the learner may have created a good performance of understanding rather than durable understanding itself.

Prompts can focus attention helpfully

Prompts can focus attention helpfully is mainly about directing the learner toward a useful feature. The learner should be able to point to the extra work being done by the explanation. What relation became clearer? Which assumption was exposed? Which missing prerequisite became visible? If the explanation adds no new structure, the activity may be fluent rehearsal rather than useful elaboration.

A practical example is asking ‘why is this step allowed?’ in a proof. The example matters because self-explanation works best when it changes what the learner notices. It can make a hidden rule visible, connect a new idea to prior knowledge or force a choice that passive review would avoid. This is why explanation is often especially valuable around worked examples and transferable principles.

The common failure is asking many prompts that scatter attention. The failure usually feels productive because language is being generated. But verbal activity is not the same as learning. A learner can speak for several minutes while repeating surface features, relying on the source or using vague phrases that conceal the exact gap. Explanation earns its place when it changes later performance.

The acceptance test is that the prompt points toward the structure that matters. That test should come after some support has been removed. Close the notes. Change the example. Ask for a reason rather than a definition. Require the learner to choose a method. If the explanation cannot survive those conditions, the learner may have created a good performance of understanding rather than durable understanding itself.

Prompts can also narrow attention too much

Prompts can also narrow attention too much is mainly about recognising that explanation is not always beneficial for every outcome. The learner should be able to point to the extra work being done by the explanation. What relation became clearer? Which assumption was exposed? Which missing prerequisite became visible? If the explanation adds no new structure, the activity may be fluent rehearsal rather than useful elaboration.

A practical example is focusing heavily on causal explanation while neglecting exact terminology that must also be remembered. The example matters because self-explanation works best when it changes what the learner notices. It can make a hidden rule visible, connect a new idea to prior knowledge or force a choice that passive review would avoid. This is why explanation is often especially valuable around worked examples and transferable principles.

The common failure is assuming more explanation can never hurt learning. The failure usually feels productive because language is being generated. But verbal activity is not the same as learning. A learner can speak for several minutes while repeating surface features, relying on the source or using vague phrases that conceal the exact gap. Explanation earns its place when it changes later performance.

The acceptance test is that the prompt does not crowd out other required information. That test should come after some support has been removed. Close the notes. Change the example. Ask for a reason rather than a definition. Require the learner to choose a method. If the explanation cannot survive those conditions, the learner may have created a good performance of understanding rather than durable understanding itself.

Explain the solution, not only the answer

Explain the solution, not only the answer is mainly about making method and decision criteria explicit. The learner should be able to point to the extra work being done by the explanation. What relation became clearer? Which assumption was exposed? Which missing prerequisite became visible? If the explanation adds no new structure, the activity may be fluent rehearsal rather than useful elaboration.

A practical example is explaining why a particular integration technique was chosen. The example matters because self-explanation works best when it changes what the learner notices. It can make a hidden rule visible, connect a new idea to prior knowledge or force a choice that passive review would avoid. This is why explanation is often especially valuable around worked examples and transferable principles.

The common failure is saying the answer is correct because the calculator confirms it. The failure usually feels productive because language is being generated. But verbal activity is not the same as learning. A learner can speak for several minutes while repeating surface features, relying on the source or using vague phrases that conceal the exact gap. Explanation earns its place when it changes later performance.

The acceptance test is that the learner can reproduce the decision without the final number. That test should come after some support has been removed. Close the notes. Change the example. Ask for a reason rather than a definition. Require the learner to choose a method. If the explanation cannot survive those conditions, the learner may have created a good performance of understanding rather than durable understanding itself.

Explain errors as well as correct work

Explain errors as well as correct work is mainly about using wrong turns to expose misconceptions. The learner should be able to point to the extra work being done by the explanation. What relation became clearer? Which assumption was exposed? Which missing prerequisite became visible? If the explanation adds no new structure, the activity may be fluent rehearsal rather than useful elaboration.

A practical example is explaining why cancelling terms across addition is invalid. The example matters because self-explanation works best when it changes what the learner notices. It can make a hidden rule visible, connect a new idea to prior knowledge or force a choice that passive review would avoid. This is why explanation is often especially valuable around worked examples and transferable principles.

The common failure is deleting the wrong step and moving on. The failure usually feels productive because language is being generated. But verbal activity is not the same as learning. A learner can speak for several minutes while repeating surface features, relying on the source or using vague phrases that conceal the exact gap. Explanation earns its place when it changes later performance.

The acceptance test is that the learner can name the tempting misconception and the repair. That test should come after some support has been removed. Close the notes. Change the example. Ask for a reason rather than a definition. Require the learner to choose a method. If the explanation cannot survive those conditions, the learner may have created a good performance of understanding rather than durable understanding itself.

Contrast explanations can sharpen categories

Contrast explanations can sharpen categories is mainly about explaining why two similar cases require different methods. The learner should be able to point to the extra work being done by the explanation. What relation became clearer? Which assumption was exposed? Which missing prerequisite became visible? If the explanation adds no new structure, the activity may be fluent rehearsal rather than useful elaboration.

A practical example is comparing when to use mean versus median. The example matters because self-explanation works best when it changes what the learner notices. It can make a hidden rule visible, connect a new idea to prior knowledge or force a choice that passive review would avoid. This is why explanation is often especially valuable around worked examples and transferable principles.

The common failure is studying each method in isolation only. The failure usually feels productive because language is being generated. But verbal activity is not the same as learning. A learner can speak for several minutes while repeating surface features, relying on the source or using vague phrases that conceal the exact gap. Explanation earns its place when it changes later performance.

The acceptance test is that the learner can discriminate between alternatives. That test should come after some support has been removed. Close the notes. Change the example. Ask for a reason rather than a definition. Require the learner to choose a method. If the explanation cannot survive those conditions, the learner may have created a good performance of understanding rather than durable understanding itself.

Teach-back can reveal hidden gaps

Teach-back can reveal hidden gaps is mainly about using another person or imagined audience as a diagnostic mirror. The learner should be able to point to the extra work being done by the explanation. What relation became clearer? Which assumption was exposed? Which missing prerequisite became visible? If the explanation adds no new structure, the activity may be fluent rehearsal rather than useful elaboration.

A practical example is explaining a concept aloud and noticing where examples replace definitions. The example matters because self-explanation works best when it changes what the learner notices. It can make a hidden rule visible, connect a new idea to prior knowledge or force a choice that passive review would avoid. This is why explanation is often especially valuable around worked examples and transferable principles.

The common failure is performing a memorised mini-lecture without checking coherence. The failure usually feels productive because language is being generated. But verbal activity is not the same as learning. A learner can speak for several minutes while repeating surface features, relying on the source or using vague phrases that conceal the exact gap. Explanation earns its place when it changes later performance.

The acceptance test is that the explanation survives interruption and follow-up questions. That test should come after some support has been removed. Close the notes. Change the example. Ask for a reason rather than a definition. Require the learner to choose a method. If the explanation cannot survive those conditions, the learner may have created a good performance of understanding rather than durable understanding itself.

Simple language is useful when it preserves precision

Simple language is useful when it preserves precision is mainly about reducing jargon without removing essential distinctions. The learner should be able to point to the extra work being done by the explanation. What relation became clearer? Which assumption was exposed? Which missing prerequisite became visible? If the explanation adds no new structure, the activity may be fluent rehearsal rather than useful elaboration.

A practical example is explaining electrical resistance with ordinary language before returning to formal units and equations. The example matters because self-explanation works best when it changes what the learner notices. It can make a hidden rule visible, connect a new idea to prior knowledge or force a choice that passive review would avoid. This is why explanation is often especially valuable around worked examples and transferable principles.

The common failure is turning a precise concept into an inaccurate analogy. The failure usually feels productive because language is being generated. But verbal activity is not the same as learning. A learner can speak for several minutes while repeating surface features, relying on the source or using vague phrases that conceal the exact gap. Explanation earns its place when it changes later performance.

The acceptance test is that the simple version can be mapped back to the technical version. That test should come after some support has been removed. Close the notes. Change the example. Ask for a reason rather than a definition. Require the learner to choose a method. If the explanation cannot survive those conditions, the learner may have created a good performance of understanding rather than durable understanding itself.

Analogies need a return path

Analogies need a return path is mainly about using familiar relationships temporarily, then checking where the analogy breaks. The learner should be able to point to the extra work being done by the explanation. What relation became clearer? Which assumption was exposed? Which missing prerequisite became visible? If the explanation adds no new structure, the activity may be fluent rehearsal rather than useful elaboration.

A practical example is comparing electric current to flow while explaining why charge and water are not the same system. The example matters because self-explanation works best when it changes what the learner notices. It can make a hidden rule visible, connect a new idea to prior knowledge or force a choice that passive review would avoid. This is why explanation is often especially valuable around worked examples and transferable principles.

The common failure is letting the analogy become the concept. The failure usually feels productive because language is being generated. But verbal activity is not the same as learning. A learner can speak for several minutes while repeating surface features, relying on the source or using vague phrases that conceal the exact gap. Explanation earns its place when it changes later performance.

The acceptance test is that the learner can state both the helpful similarity and the limit. That test should come after some support has been removed. Close the notes. Change the example. Ask for a reason rather than a definition. Require the learner to choose a method. If the explanation cannot survive those conditions, the learner may have created a good performance of understanding rather than durable understanding itself.

Drawing can become explanation

Drawing can become explanation is mainly about using diagrams to express causal or structural relations. The learner should be able to point to the extra work being done by the explanation. What relation became clearer? Which assumption was exposed? Which missing prerequisite became visible? If the explanation adds no new structure, the activity may be fluent rehearsal rather than useful elaboration.

A practical example is drawing a labelled feedback loop and explaining each arrow. The example matters because self-explanation works best when it changes what the learner notices. It can make a hidden rule visible, connect a new idea to prior knowledge or force a choice that passive review would avoid. This is why explanation is often especially valuable around worked examples and transferable principles.

The common failure is copying a diagram without making relationships explicit. The failure usually feels productive because language is being generated. But verbal activity is not the same as learning. A learner can speak for several minutes while repeating surface features, relying on the source or using vague phrases that conceal the exact gap. Explanation earns its place when it changes later performance.

The acceptance test is that the representation carries an explanatory claim. That test should come after some support has been removed. Close the notes. Change the example. Ask for a reason rather than a definition. Require the learner to choose a method. If the explanation cannot survive those conditions, the learner may have created a good performance of understanding rather than durable understanding itself.

Examples need generalisation

Examples need generalisation is mainly about moving from a case to the rule it illustrates. The learner should be able to point to the extra work being done by the explanation. What relation became clearer? Which assumption was exposed? Which missing prerequisite became visible? If the explanation adds no new structure, the activity may be fluent rehearsal rather than useful elaboration.

A practical example is using one grammar sentence to explain a broader clause relationship. The example matters because self-explanation works best when it changes what the learner notices. It can make a hidden rule visible, connect a new idea to prior knowledge or force a choice that passive review would avoid. This is why explanation is often especially valuable around worked examples and transferable principles.

The common failure is remembering only the example sentence. The failure usually feels productive because language is being generated. But verbal activity is not the same as learning. A learner can speak for several minutes while repeating surface features, relying on the source or using vague phrases that conceal the exact gap. Explanation earns its place when it changes later performance.

The acceptance test is that the learner can produce a different valid example. That test should come after some support has been removed. Close the notes. Change the example. Ask for a reason rather than a definition. Require the learner to choose a method. If the explanation cannot survive those conditions, the learner may have created a good performance of understanding rather than durable understanding itself.

Explanation can be written, spoken or diagrammatic

Explanation can be written, spoken or diagrammatic is mainly about choosing a form that matches the task rather than a supposed learning style. The learner should be able to point to the extra work being done by the explanation. What relation became clearer? Which assumption was exposed? Which missing prerequisite became visible? If the explanation adds no new structure, the activity may be fluent rehearsal rather than useful elaboration.

A practical example is writing a proof explanation, speaking a pronunciation rationale or sketching a system. The example matters because self-explanation works best when it changes what the learner notices. It can make a hidden rule visible, connect a new idea to prior knowledge or force a choice that passive review would avoid. This is why explanation is often especially valuable around worked examples and transferable principles.

The common failure is believing spoken explanation is inherently superior. The failure usually feels productive because language is being generated. But verbal activity is not the same as learning. A learner can speak for several minutes while repeating surface features, relying on the source or using vague phrases that conceal the exact gap. Explanation earns its place when it changes later performance.

The acceptance test is that the form makes the target relations inspectable. That test should come after some support has been removed. Close the notes. Change the example. Ask for a reason rather than a definition. Require the learner to choose a method. If the explanation cannot survive those conditions, the learner may have created a good performance of understanding rather than durable understanding itself.

Retrieval and explanation are complementary

Retrieval and explanation are complementary is mainly about bringing knowledge to mind before elaborating it. The learner should be able to point to the extra work being done by the explanation. What relation became clearer? Which assumption was exposed? Which missing prerequisite became visible? If the explanation adds no new structure, the activity may be fluent rehearsal rather than useful elaboration.

A practical example is attempting a concept from memory and then explaining the retrieved idea. The example matters because self-explanation works best when it changes what the learner notices. It can make a hidden rule visible, connect a new idea to prior knowledge or force a choice that passive review would avoid. This is why explanation is often especially valuable around worked examples and transferable principles.

The common failure is reading notes while explaining every line. The failure usually feels productive because language is being generated. But verbal activity is not the same as learning. A learner can speak for several minutes while repeating surface features, relying on the source or using vague phrases that conceal the exact gap. Explanation earns its place when it changes later performance.

The acceptance test is that the learner can explain with reduced support. That test should come after some support has been removed. Close the notes. Change the example. Ask for a reason rather than a definition. Require the learner to choose a method. If the explanation cannot survive those conditions, the learner may have created a good performance of understanding rather than durable understanding itself.

Spacing makes explanation a stronger test

Spacing makes explanation a stronger test is mainly about returning after delay so explanation reflects durable knowledge. The learner should be able to point to the extra work being done by the explanation. What relation became clearer? Which assumption was exposed? Which missing prerequisite became visible? If the explanation adds no new structure, the activity may be fluent rehearsal rather than useful elaboration.

A practical example is explaining a process again two days later without notes. The example matters because self-explanation works best when it changes what the learner notices. It can make a hidden rule visible, connect a new idea to prior knowledge or force a choice that passive review would avoid. This is why explanation is often especially valuable around worked examples and transferable principles.

The common failure is doing all explanations immediately after study. The failure usually feels productive because language is being generated. But verbal activity is not the same as learning. A learner can speak for several minutes while repeating surface features, relying on the source or using vague phrases that conceal the exact gap. Explanation earns its place when it changes later performance.

The acceptance test is that the learner can rebuild the reasoning after forgetting has begun. That test should come after some support has been removed. Close the notes. Change the example. Ask for a reason rather than a definition. Require the learner to choose a method. If the explanation cannot survive those conditions, the learner may have created a good performance of understanding rather than durable understanding itself.

Interleaving can strengthen explanatory choice

Interleaving can strengthen explanatory choice is mainly about mixing related problems so learners must explain which method fits. The learner should be able to point to the extra work being done by the explanation. What relation became clearer? Which assumption was exposed? Which missing prerequisite became visible? If the explanation adds no new structure, the activity may be fluent rehearsal rather than useful elaboration.

A practical example is explaining why a percentage-change problem differs from a ratio problem. The example matters because self-explanation works best when it changes what the learner notices. It can make a hidden rule visible, connect a new idea to prior knowledge or force a choice that passive review would avoid. This is why explanation is often especially valuable around worked examples and transferable principles.

The common failure is explaining methods only inside blocked practice where the choice is obvious. The failure usually feels productive because language is being generated. But verbal activity is not the same as learning. A learner can speak for several minutes while repeating surface features, relying on the source or using vague phrases that conceal the exact gap. Explanation earns its place when it changes later performance.

The acceptance test is that the learner can justify method selection. That test should come after some support has been removed. Close the notes. Change the example. Ask for a reason rather than a definition. Require the learner to choose a method. If the explanation cannot survive those conditions, the learner may have created a good performance of understanding rather than durable understanding itself.

Question stems should vary by domain

Question stems should vary by domain is mainly about matching explanation prompts to disciplinary reasoning. The learner should be able to point to the extra work being done by the explanation. What relation became clearer? Which assumption was exposed? Which missing prerequisite became visible? If the explanation adds no new structure, the activity may be fluent rehearsal rather than useful elaboration.

A practical example is asking ‘what evidence supports this interpretation?’ in literature and ‘why is this transformation valid?’ in mathematics. The example matters because self-explanation works best when it changes what the learner notices. It can make a hidden rule visible, connect a new idea to prior knowledge or force a choice that passive review would avoid. This is why explanation is often especially valuable around worked examples and transferable principles.

The common failure is using one generic ‘why?’ prompt everywhere. The failure usually feels productive because language is being generated. But verbal activity is not the same as learning. A learner can speak for several minutes while repeating surface features, relying on the source or using vague phrases that conceal the exact gap. Explanation earns its place when it changes later performance.

The acceptance test is that the question reflects the standards of the subject. That test should come after some support has been removed. Close the notes. Change the example. Ask for a reason rather than a definition. Require the learner to choose a method. If the explanation cannot survive those conditions, the learner may have created a good performance of understanding rather than durable understanding itself.

Novices need scaffolds

Novices need scaffolds is mainly about supporting explanation before expecting expert-level reasoning. The learner should be able to point to the extra work being done by the explanation. What relation became clearer? Which assumption was exposed? Which missing prerequisite became visible? If the explanation adds no new structure, the activity may be fluent rehearsal rather than useful elaboration.

A practical example is providing sentence starters such as ‘this step works because…’. The example matters because self-explanation works best when it changes what the learner notices. It can make a hidden rule visible, connect a new idea to prior knowledge or force a choice that passive review would avoid. This is why explanation is often especially valuable around worked examples and transferable principles.

The common failure is leaving a beginner with a blank page and calling struggle desirable. The failure usually feels productive because language is being generated. But verbal activity is not the same as learning. A learner can speak for several minutes while repeating surface features, relying on the source or using vague phrases that conceal the exact gap. Explanation earns its place when it changes later performance.

The acceptance test is that support reduces once explanatory structure becomes familiar. That test should come after some support has been removed. Close the notes. Change the example. Ask for a reason rather than a definition. Require the learner to choose a method. If the explanation cannot survive those conditions, the learner may have created a good performance of understanding rather than durable understanding itself.

Experts may need fewer prompts

Experts may need fewer prompts is mainly about avoiding unnecessary verbalisation that slows fluent performance. The learner should be able to point to the extra work being done by the explanation. What relation became clearer? Which assumption was exposed? Which missing prerequisite became visible? If the explanation adds no new structure, the activity may be fluent rehearsal rather than useful elaboration.

A practical example is asking an expert to explain only unusual decisions or errors. The example matters because self-explanation works best when it changes what the learner notices. It can make a hidden rule visible, connect a new idea to prior knowledge or force a choice that passive review would avoid. This is why explanation is often especially valuable around worked examples and transferable principles.

The common failure is forcing a detailed explanation of every routine step. The failure usually feels productive because language is being generated. But verbal activity is not the same as learning. A learner can speak for several minutes while repeating surface features, relying on the source or using vague phrases that conceal the exact gap. Explanation earns its place when it changes later performance.

The acceptance test is that explanation is used where it adds diagnosis or transfer. That test should come after some support has been removed. Close the notes. Change the example. Ask for a reason rather than a definition. Require the learner to choose a method. If the explanation cannot survive those conditions, the learner may have created a good performance of understanding rather than durable understanding itself.

Time cost matters

Time cost matters is mainly about comparing explanation with alternative uses of study time. The learner should be able to point to the extra work being done by the explanation. What relation became clearer? Which assumption was exposed? Which missing prerequisite became visible? If the explanation adds no new structure, the activity may be fluent rehearsal rather than useful elaboration.

A practical example is self-explaining a difficult worked example rather than explaining every easy fact. The example matters because self-explanation works best when it changes what the learner notices. It can make a hidden rule visible, connect a new idea to prior knowledge or force a choice that passive review would avoid. This is why explanation is often especially valuable around worked examples and transferable principles.

The common failure is adding explanation to everything until practice volume collapses. The failure usually feels productive because language is being generated. But verbal activity is not the same as learning. A learner can speak for several minutes while repeating surface features, relying on the source or using vague phrases that conceal the exact gap. Explanation earns its place when it changes later performance.

The acceptance test is that the learner uses explanation where it has the highest value. That test should come after some support has been removed. Close the notes. Change the example. Ask for a reason rather than a definition. Require the learner to choose a method. If the explanation cannot survive those conditions, the learner may have created a good performance of understanding rather than durable understanding itself.

Metacognition improves when explanations are checked

Metacognition improves when explanations are checked is mainly about comparing perceived understanding with actual performance. The learner should be able to point to the extra work being done by the explanation. What relation became clearer? Which assumption was exposed? Which missing prerequisite became visible? If the explanation adds no new structure, the activity may be fluent rehearsal rather than useful elaboration.

A practical example is explaining a concept, then answering a changed question without notes. The example matters because self-explanation works best when it changes what the learner notices. It can make a hidden rule visible, connect a new idea to prior knowledge or force a choice that passive review would avoid. This is why explanation is often especially valuable around worked examples and transferable principles.

The common failure is feeling fluent after a smooth explanation and assuming mastery. The failure usually feels productive because language is being generated. But verbal activity is not the same as learning. A learner can speak for several minutes while repeating surface features, relying on the source or using vague phrases that conceal the exact gap. Explanation earns its place when it changes later performance.

The acceptance test is that confidence is calibrated against performance. That test should come after some support has been removed. Close the notes. Change the example. Ask for a reason rather than a definition. Require the learner to choose a method. If the explanation cannot survive those conditions, the learner may have created a good performance of understanding rather than durable understanding itself.

Peer explanation can help both speaker and listener

Peer explanation can help both speaker and listener is mainly about using dialogue to expose assumptions and questions. The learner should be able to point to the extra work being done by the explanation. What relation became clearer? Which assumption was exposed? Which missing prerequisite became visible? If the explanation adds no new structure, the activity may be fluent rehearsal rather than useful elaboration.

A practical example is one student explaining while the partner challenges an unclear step. The example matters because self-explanation works best when it changes what the learner notices. It can make a hidden rule visible, connect a new idea to prior knowledge or force a choice that passive review would avoid. This is why explanation is often especially valuable around worked examples and transferable principles.

The common failure is dividing work so one student speaks and the other passively agrees. The failure usually feels productive because language is being generated. But verbal activity is not the same as learning. A learner can speak for several minutes while repeating surface features, relying on the source or using vague phrases that conceal the exact gap. Explanation earns its place when it changes later performance.

The acceptance test is that both participants must inspect the reasoning. That test should come after some support has been removed. Close the notes. Change the example. Ask for a reason rather than a definition. Require the learner to choose a method. If the explanation cannot survive those conditions, the learner may have created a good performance of understanding rather than durable understanding itself.

AI can be used as an explanation partner carefully

AI can be used as an explanation partner carefully is mainly about using a model to ask questions or critique a learner’s explanation without outsourcing the reasoning. The learner should be able to point to the extra work being done by the explanation. What relation became clearer? Which assumption was exposed? Which missing prerequisite became visible? If the explanation adds no new structure, the activity may be fluent rehearsal rather than useful elaboration.

A practical example is having AI identify an unclear claim after the learner first explains independently. The example matters because self-explanation works best when it changes what the learner notices. It can make a hidden rule visible, connect a new idea to prior knowledge or force a choice that passive review would avoid. This is why explanation is often especially valuable around worked examples and transferable principles.

The common failure is asking AI to generate the explanation and then rereading it. The failure usually feels productive because language is being generated. But verbal activity is not the same as learning. A learner can speak for several minutes while repeating surface features, relying on the source or using vague phrases that conceal the exact gap. Explanation earns its place when it changes later performance.

The acceptance test is that the learner still produces and verifies the reasoning. That test should come after some support has been removed. Close the notes. Change the example. Ask for a reason rather than a definition. Require the learner to choose a method. If the explanation cannot survive those conditions, the learner may have created a good performance of understanding rather than durable understanding itself.

Feynman-style rewriting should return to source checking

Feynman-style rewriting should return to source checking is mainly about using simplification as a diagnostic, then verifying against authoritative material. The learner should be able to point to the extra work being done by the explanation. What relation became clearer? Which assumption was exposed? Which missing prerequisite became visible? If the explanation adds no new structure, the activity may be fluent rehearsal rather than useful elaboration.

A practical example is checking a simplified explanation of mitosis against the textbook after writing it from memory. The example matters because self-explanation works best when it changes what the learner notices. It can make a hidden rule visible, connect a new idea to prior knowledge or force a choice that passive review would avoid. This is why explanation is often especially valuable around worked examples and transferable principles.

The common failure is treating the learner’s simple version as final truth. The failure usually feels productive because language is being generated. But verbal activity is not the same as learning. A learner can speak for several minutes while repeating surface features, relying on the source or using vague phrases that conceal the exact gap. Explanation earns its place when it changes later performance.

The acceptance test is that the explanation is both understandable and accurate. That test should come after some support has been removed. Close the notes. Change the example. Ask for a reason rather than a definition. Require the learner to choose a method. If the explanation cannot survive those conditions, the learner may have created a good performance of understanding rather than durable understanding itself.

Explanation should end in use

Explanation should end in use is mainly about testing whether understanding changes performance. The learner should be able to point to the extra work being done by the explanation. What relation became clearer? Which assumption was exposed? Which missing prerequisite became visible? If the explanation adds no new structure, the activity may be fluent rehearsal rather than useful elaboration.

A practical example is solving a new problem after explaining the governing principle. The example matters because self-explanation works best when it changes what the learner notices. It can make a hidden rule visible, connect a new idea to prior knowledge or force a choice that passive review would avoid. This is why explanation is often especially valuable around worked examples and transferable principles.

The common failure is stopping when the explanation sounds polished. The failure usually feels productive because language is being generated. But verbal activity is not the same as learning. A learner can speak for several minutes while repeating surface features, relying on the source or using vague phrases that conceal the exact gap. Explanation earns its place when it changes later performance.

The acceptance test is that the learner can act on the explanation. That test should come after some support has been removed. Close the notes. Change the example. Ask for a reason rather than a definition. Require the learner to choose a method. If the explanation cannot survive those conditions, the learner may have created a good performance of understanding rather than durable understanding itself.

The method should be abandoned when it no longer helps

The method should be abandoned when it no longer helps is mainly about recognising diminishing returns. The learner should be able to point to the extra work being done by the explanation. What relation became clearer? Which assumption was exposed? Which missing prerequisite became visible? If the explanation adds no new structure, the activity may be fluent rehearsal rather than useful elaboration.

A practical example is moving from explanation to timed practice once the decision structure is stable. The example matters because self-explanation works best when it changes what the learner notices. It can make a hidden rule visible, connect a new idea to prior knowledge or force a choice that passive review would avoid. This is why explanation is often especially valuable around worked examples and transferable principles.

The common failure is explaining familiar facts repeatedly because the technique is fashionable. The failure usually feels productive because language is being generated. But verbal activity is not the same as learning. A learner can speak for several minutes while repeating surface features, relying on the source or using vague phrases that conceal the exact gap. Explanation earns its place when it changes later performance.

The acceptance test is that study time shifts to the next bottleneck. That test should come after some support has been removed. Close the notes. Change the example. Ask for a reason rather than a definition. Require the learner to choose a method. If the explanation cannot survive those conditions, the learner may have created a good performance of understanding rather than durable understanding itself.

Good explanation exposes uncertainty

Good explanation exposes uncertainty is mainly about allowing the learner to mark what is not yet understood. The learner should be able to point to the extra work being done by the explanation. What relation became clearer? Which assumption was exposed? Which missing prerequisite became visible? If the explanation adds no new structure, the activity may be fluent rehearsal rather than useful elaboration.

A practical example is saying ‘I can explain the first two steps but not why this assumption is needed’. The example matters because self-explanation works best when it changes what the learner notices. It can make a hidden rule visible, connect a new idea to prior knowledge or force a choice that passive review would avoid. This is why explanation is often especially valuable around worked examples and transferable principles.

The common failure is hiding gaps with vague words such as ‘basically’ or ‘somehow’. The failure usually feels productive because language is being generated. But verbal activity is not the same as learning. A learner can speak for several minutes while repeating surface features, relying on the source or using vague phrases that conceal the exact gap. Explanation earns its place when it changes later performance.

The acceptance test is that uncertainty becomes a precise study target. That test should come after some support has been removed. Close the notes. Change the example. Ask for a reason rather than a definition. Require the learner to choose a method. If the explanation cannot survive those conditions, the learner may have created a good performance of understanding rather than durable understanding itself.

The final test is independent reconstruction

The final test is independent reconstruction is mainly about rebuilding the idea without notes, examples or a prepared script. The learner should be able to point to the extra work being done by the explanation. What relation became clearer? Which assumption was exposed? Which missing prerequisite became visible? If the explanation adds no new structure, the activity may be fluent rehearsal rather than useful elaboration.

A practical example is explaining the concept to a blank page after delay and then solving a changed task. The example matters because self-explanation works best when it changes what the learner notices. It can make a hidden rule visible, connect a new idea to prior knowledge or force a choice that passive review would avoid. This is why explanation is often especially valuable around worked examples and transferable principles.

The common failure is reciting a rehearsed paragraph. The failure usually feels productive because language is being generated. But verbal activity is not the same as learning. A learner can speak for several minutes while repeating surface features, relying on the source or using vague phrases that conceal the exact gap. Explanation earns its place when it changes later performance.

The acceptance test is that the learner can regenerate the structure flexibly. That test should come after some support has been removed. Close the notes. Change the example. Ask for a reason rather than a definition. Require the learner to choose a method. If the explanation cannot survive those conditions, the learner may have created a good performance of understanding rather than durable understanding itself.

Explanation should distinguish cause from description

Explanation should distinguish cause from description is mainly about requiring the learner to state what makes an outcome happen rather than merely what is observed. The learner should be able to point to the extra work being done by the explanation. What relation became clearer? Which assumption was exposed? Which missing prerequisite became visible? If the explanation adds no new structure, the activity may be fluent rehearsal rather than useful elaboration.

A practical example is explaining why increasing resistance lowers current under fixed voltage rather than saying the current becomes smaller. The example matters because self-explanation works best when it changes what the learner notices. It can make a hidden rule visible, connect a new idea to prior knowledge or force a choice that passive review would avoid. This is why explanation is often especially valuable around worked examples and transferable principles.

The common failure is using restated observations as if they were causal accounts. The failure usually feels productive because language is being generated. But verbal activity is not the same as learning. A learner can speak for several minutes while repeating surface features, relying on the source or using vague phrases that conceal the exact gap. Explanation earns its place when it changes later performance.

The acceptance test is that the explanation identifies the relation that produces the result. That test should come after some support has been removed. Close the notes. Change the example. Ask for a reason rather than a definition. Require the learner to choose a method. If the explanation cannot survive those conditions, the learner may have created a good performance of understanding rather than durable understanding itself.

Counterexamples protect against overgeneralisation

Counterexamples protect against overgeneralisation is mainly about testing whether the learner knows where a rule stops applying. The learner should be able to point to the extra work being done by the explanation. What relation became clearer? Which assumption was exposed? Which missing prerequisite became visible? If the explanation adds no new structure, the activity may be fluent rehearsal rather than useful elaboration.

A practical example is giving a case that looks similar but violates an assumption and asking why the usual method fails. The example matters because self-explanation works best when it changes what the learner notices. It can make a hidden rule visible, connect a new idea to prior knowledge or force a choice that passive review would avoid. This is why explanation is often especially valuable around worked examples and transferable principles.

The common failure is collecting only examples where the preferred explanation works. The failure usually feels productive because language is being generated. But verbal activity is not the same as learning. A learner can speak for several minutes while repeating surface features, relying on the source or using vague phrases that conceal the exact gap. Explanation earns its place when it changes later performance.

The acceptance test is that the learner can state the boundary conditions of the idea. That test should come after some support has been removed. Close the notes. Change the example. Ask for a reason rather than a definition. Require the learner to choose a method. If the explanation cannot survive those conditions, the learner may have created a good performance of understanding rather than durable understanding itself.

A practical Feynman-style study loop

  • Choose one concept narrow enough to explain in five to ten minutes.
  • Close the source and explain the concept from memory in ordinary language.
  • Mark every vague phrase, unexplained step, missing example and point where you rely on jargon.
  • Return to the source only for those gaps.
  • Rewrite the explanation more precisely, not merely more simply.
  • Add one worked example and one counterexample or failure case.
  • After a delay, explain again without the previous script.
  • Finish with a changed task that requires using the idea rather than describing it.

When explaining is worth the time

Explanation is especially valuable when the learner must understand why a method works, discriminate between similar methods, transfer a principle or repair a misconception. It can be less efficient for material that mainly requires exact memorisation, already-fluent routine procedures or facts whose explanatory structure is weak. The learner should allocate effort according to the bottleneck.

This is why the technique should not replace retrieval practice or problem solving. A learner can explain the formula for compound interest beautifully and still be slow applying it under exam conditions. Another learner can calculate quickly but fail when the problem is phrased differently. The study system needs both explanation and performance.

Frequently asked questions

Is the Feynman Technique scientifically proven?

The popular branded sequence is better understood as an application of established learning processes such as self-explanation, retrieval, simplification and gap detection. Evidence supports self-explanation under many conditions, but not a universal claim that one fixed Feynman routine is always best.

Do I have to explain to another person?

No. Explaining to a blank page, a voice recorder or an imagined learner can work. Another person becomes useful when their questions expose assumptions you did not notice.

Should I use simple language only?

Use simple language to test understanding, then reconnect it to precise technical vocabulary. A simplified explanation that becomes inaccurate is not an improvement.

Can I use the technique for Mathematics?

Yes, especially to explain why a method is valid, how to choose between methods and where common errors come from. It should be followed by independent problem solving.

Can I use it for vocabulary?

Yes, by explaining meaning, contrasts, collocations and examples in your own words. But exact spelling, pronunciation and usage still need direct practice and retrieval.

Can I use AI as the audience?

Yes, cautiously. Explain first from your own knowledge, then ask the AI to question unclear points or generate a counterexample. Verify factual corrections against reliable sources.

How long should one explanation take?

Long enough to expose the structure, short enough that explanation does not consume all practice time. Five to fifteen minutes can be enough for a bounded concept, but task complexity matters.

Why does explaining sometimes make me feel worse?

Because gaps become visible. That is often useful. A smooth reread can create familiarity without revealing missing understanding.

What if I cannot explain the topic simply?

Locate the first word or step you cannot unpack. Repair that dependency rather than repeatedly attempting the whole explanation.

How do I know the method worked?

Test after delay and under changed conditions. If you can explain the principle and use it in a new task with less support, the technique has contributed useful learning.

The quiet conclusion: explanation is a diagnostic instrument

The strongest version of the Feynman Technique is not a performance of simplicity. It is a disciplined attempt to make understanding inspectable. Every vague phrase is a clue. Every missing link is a study target. Every counterexample checks whether the learner has mistaken a familiar case for the general rule.

Used this way, explaining becomes part of a larger learning system: retrieve, explain, verify, practise, vary and return after delay. The technique earns its place not because the explanation sounds elegant, but because the learner becomes more capable when the notes are gone and the problem has changed.