Logos52
wiki / Syntheses / Are You Learning, or Just Using Techniques

Are You Learning, or Just Using Techniques

synthesis updated 2026-08-14

Are You Learning, or Just Using Techniques

Notes, maps, and highlights can look finished in an hour spent skipping comparison, reconstruction, or judgment. The productive sequence is strategy → thinking → knowledge → result. A technique is working when it triggers that thinking.

The thinking also sets the quality of the memory and whether that knowledge can be applied in the right ways. Judge the method by the effect it was chosen for, not by whether the steps looked clean.

Self-Regulation is the capacity that notices, mid-session, when the process has drifted off that sequence.

Form is not the technique

A method has two parts: the visible form, and the mental operation the form is meant to force. The form is easy to copy. The operation is the point.

The same passage can leave three maps that look comparable and three different operations underneath. One is a transcription of headings. One groups a little and still never compares. One groups, ranks, and explains why the arrows are there. The paper looks finished in every case.

FormWhat the form is meant to force
HighlightingNothing that finishes. A mark records a guess that a line matters, then stops. Evaluation of why it matters, and what it joins, has to happen somewhere else. Highlighting is not that evaluation. Isolated marks can pull attention onto isolated facts and make inference worse.
Mind mappingGroup, compare, rank, and explain relationships until a schema exists — an organised structure where a new fact has a place to attach.
Bear Hunter SystemThree encoding passes: write the questions the material must answer before opening it (Aim), answer them into a rough relational map while reading, then prune the map until it can be rebuilt without the source.
RetrievalReconstruct, discriminate, diagnose, and repair gaps. Recall or use from memory before the source is opened. Repeated testing has left about 61% recalled a week later against about 40% after the same time spent restudying.
Kolbs Experiential CycleTurn one attempt into a better rule for the next: do the thing, inspect what happened, form a principle, test that principle. The cycle-as-reflection-loop is the claim. The learning-styles inventory built beside it is a different and poorly supported idea, and the loop does not inherit that inventory’s authority.

When the form happens without the thinking, the technique becomes performance theater.

Doing the technique right is a weaker question than it seems. The same named method lands in different utility bands depending on the operation inside it. The better question is whether the technique is making the thinking happen at the level the result deserves.

The only row of the old five-layer model worth keeping is Execution: did the technique actually produce that thinking? Intent is cheap. The artifact is not an answer. Deep Processing is the home of the “what kind of thinking” question; this page is the test for whether a given hour produced it.

The shortcut

A learner can produce notes, maps, highlights, prompts, or answers while avoiding the comparison, reconstruction, uncertainty, or judgment that would actually change understanding. That is the danger, now with objects.

The default is lower-effort cognition. Given a free choice between two equally rewarded options, people pick the less demanding one. Effort is also valued, and can add value to what it produces — averse by default, not by nature, which is why a repair that makes the right effort attractive can hold. When a task turns uncertain, the search for a way out starts. Four pictures of that search: get the answer from somewhere else; look at how someone else is doing it; put it on paper without thinking; draw the arrow for the sake of the arrow.

Cognitive Load & What Mental Effort Is Trying to Cue is the vault’s home for effort as a signal rather than as a malfunction.

Eight common escapes:

  • looking up the answer instead of constructing it
  • copying the structure from someone else
  • asking a model to simplify before a first attempt
  • drawing arrows without explaining relationships
  • highlighting without evaluating
  • summarizing without organizing
  • repeating facts instead of building schema
  • switching techniques before one technique has been trained

The third is not the same kind of escape as the others. It is available in one keystroke, and it produces the most convincing artifact on the list. The harm is specifically the answer arriving before the attempt, not the tool. In a field experiment with nearly a thousand high-school students, unrestricted assistance raised practice scores (about 48% above the no-tool control) and left those students below the control once the tool was withdrawn (about 17% lower on the unaided exam). A version constrained to withhold answers and prompt for the student’s own attempt did not leave that deficit. The rule that follows is one line: make the attempt first, then let the model check it — never let it go first. The model can be a legitimate technique or a shortcut. The discriminator is the same one the rest of this page uses.

The shortcut is dangerous because it often leaves evidence behind. Something got written. Something got highlighted. Something looks cleaner.

The visible artifact is not proof. The highest-quality-looking maps and notes have come from the condition that learned least, and the people who made them predicted the opposite. Fluency generated by the artifact is read as learning. The Shortcut Problem isolates that failure mode; this page is the diagnostic that decides whether the hour was that failure or not.

The goal’s order of thinking

Most serious learning goals are not recall goals. They ask for comparison, evaluation, application, transfer, discrimination, problem solving, and judgment under changed conditions. Lower-order work recognises and recalls discrete items. The second cannot be substituted for the first.

Lower-order techniques can support those goals. They cannot replace the thinking the goals require. Transfer from retrieval practice is real, smaller than same-format retention, and shrinks as the transfer distance grows. Practising the low-order form buys some of the high-order result and not all of it.

Repetition can strengthen memory. It cannot automatically create structure. Maintenance rehearsal produces little durable gain next to semantic processing. Familiarity can make content feel easier. It does not prove the learner can use it. Repeated restudy has raised confidence while lowering week-delayed recall.

If the goal is higher-order and the method only produces lower-order familiarity, the session is training the wrong thing. At volume that training collapses under forgetting: without a higher-order schema, memory fades fast enough that the load becomes unsustainable. Disorganised knowledge is the expert finding run in reverse — experts sort by deep structure, and that organisation predicts performance — so a result that never improves is often a structure that was never built, not a method that was never found.

The diagnostic that hands the next section its job: is the current method forcing the kind of thinking the outcome actually needs?

Misinterpreted effort

One common failure is quitting a technique too early. The learner tries a method, feels uncertain or unimpressed, and concludes the method does not work.

That reading has a name: misinterpreted effort. Perceived effort is taken as poor learning, which lowers the judged effectiveness of the strategy, which lowers the chance of choosing it again — even when the strategy is producing better retention than the easier alternative. The Learning Zone and the Reversion Response assigns that diagnostic here.

Two other explanations are often more likely. The method was starting to work, but progress was too subtle to notice yet. Or the method was not being executed with the intended thinking. The first is the learning/performance distinction: the conditions that most improve long-term learning routinely depress visible short-term performance, so the in-session signal is inverted. The second is the hypothesis itself.

The repair is not blind persistence. The repair is a clear theory of the technique, written before the session, not recalled during it — because the moment the theory is needed is the moment deliberation is failing. Four lines: what thinking it should trigger; what that thinking should feel like; what small signal should change first; what result should eventually improve.

Filled in on the highlighting row: the thinking it should trigger is evaluation of why a line matters and what it joins; that thinking is not the mark; the first signal is a question the mark cannot answer; the result that should improve is inference, not a tidier page. If the hour only produces marks, execution failed, and highlighting cannot be repaired into that evaluation. The evaluation has to move to a method that forces it.

The two explanations call for opposite next moves, so they need a discriminator. Subtle progress is settled by a delayed objective check — days later, not in the session, because in-session performance is the inverted signal. Failed execution is settled by an audit against the written theory: what the technique was supposed to make happen, set next to what the hour actually did. Without that split, the theory is advice that cannot be acted on. Without the theory, methods keep getting switched and the underlying skill never gets trained.

Marginal Gains is what to change instead of abandoning the method: the next improvement small enough to start.

Daily self-correction still needs a fast signal. The guiding question is whether the hour feels like the right kind of thinking — a process question, not a prediction of later recall. Six good signs: curiosity increases; questions come more easily; confusion becomes more specific; relationships become easier to see; the structure feels more navigable; the topic starts to feel usable instead of merely familiar. Two of those are in the source notes; the other four are the same kind of process cue. Judgment-of-learning findings indict predictions of future recall, not these process signs. Rising curiosity is more than a symptom: curiosity states enhance hippocampal-dependent learning, including of incidental material encountered while curious.

Those signs help steer before formal results appear. They do not replace the test. Spaced Interleaved Retrieval — recall before checking, gaps that widen, topics and formats mixed inside a session — and later objective performance still have to test the knowledge. The subjective signal is wrong in a direction. The smoothest sessions are the ones most in need of an objective check, not least.

How to use this page

Use this page when a method looks correct but results are not improving.

Interleaved retrieval is valuable because it makes shortcutting harder. It removes the cue that tells which method applies, so the discrimination has to be made rather than inherited from the block. Plain review allows recognition. Highlighting allows decoration. Copying allows false clarity. Mixed retrieval forces reconstruct, choose, compare, and discriminate. A meta-analysis of interleaving puts the overall effect around g = 0.42, strongest for inductive category learning and weaker or absent for some material types.

That is why mixed retrieval is an entry point to the right level of thinking, not an entry point to the material. The benefit depends on having enough of each category encoded to tell them apart. For a genuine novice, block until the categories can be told apart, then interleave. Skip the structure-building pass on material that already has a frame; scaffolding that helps a beginner can hurt someone who already holds the schema.

It exposes the level of thinking that has to be reached, then trains that level through repeated attempts. An unresolved attempt is a debt. Close it in the same session, or what gets encoded is the guess. Failed attempts help when the correction arrives.

WPW — whole, then parts, then whole again, taught back with every source closed — is the objective test for whether a topic is structured rather than merely familiar.

The seven questions are the retrospective instrument. Four other pages send readers here for them.

  1. What result was wanted?
  2. What kind of thinking should create that result?
  3. What thinking did the technique actually trigger?
  4. Where was a shortcut taken?
  5. Did the method get dropped before the skill had time to develop?
  6. What subjective signal should be watched next time?
  7. What objective check will eventually confirm it?

If the strategy does not trigger the intended thinking, it will not create the intended knowledge. Same activity time, different mental operation, different outcome. Question 3 is that fact asked of one hour.

The Technique Is Only as Good as the Thinking It Produces is the short diagnostic of the same thesis; this page is the synthesis.

The case against the page is honest. Process signs can be flattered. This is a diagnostic, not a method, and interleaved retrieval is the wrong first move on material that has not yet been told apart. The seven questions take longer than switching techniques, which is the point. The price is writing the theory before the session and running a delayed check. Quit if two cycles of the seven questions produce no change in the Execution answer. The checkable expectation is a delayed test that is not a recognition check.

The locked-in state

The target is organised usability: the topic feels less confusing; the main structure feels stable; details have places to attach; movement between whole and part is possible; explaining, comparing, and applying are possible; the knowledge can be topped up later through retrieval.

That is the “locked in” feeling — a felt-sense name, not a construct. Familiarity is a bad test, which is why restudy that feels easier proves nothing. Familiarity is also a real destination. The end state does feel familiar and comfortable. Both are true. The test and the destination are not the same job.

It does not mean everything is remembered forever. Forgetting is the default. It means a structure exists that can survive forgetting and be rebuilt. Material that appears completely forgotten is relearned faster than it was learned the first time, so the structure persists below the threshold of recall. What decays is access, not the trace. Topping it up later is that mechanism, not a reassurance.

Knowledge Mastery: From Recognition to Usable Knowledge is the levelled, objective counterpart to that feeling: retrieve, relate, evaluate, use across contexts.

The goal is to make each technique produce the right thinking. Tidy output was never the evidence. A structure that can be rebuilt after it has been forgotten is.

Open Questions

  • Which of the techniques in use is easiest to fake?
  • What does the correct thinking for Bear Hunter System feel like in the first five minutes?
  • What does the correct thinking for Spaced Interleaved Retrieval feel like when interleaving is working?
  • Which current learning goal is higher-order but still being trained with lower-order methods?
  • What signal would show that a technique is producing organised usability?

Sources