Relearning & Savings | Why Forgotten Words Return Faster

RELEARNING · SAVINGS · FORGOTTEN WORDS · MEMORY TRACE · VOCABULARY RECOVERY · RECOGNITION · RECALL · ACCESSIBILITY

Relearning is learning a word again after recall or use has weakened. Savings is the reduction in time, trials or repetitions needed for relearning compared with the original learning. The classic Ebbinghaus method used this difference to detect memory that ordinary recall could miss.

For vocabulary, savings explains a familiar experience: a word that seems forgotten often returns much faster than a completely new word. The learner may retain recognition, partial spelling, first sounds, meaning fragments, collocations or conceptual links even when free recall fails.

This guide explains relearning, savings, residual memory traces, accessibility, recognition, free recall, cueing, lexical attrition, error relapse, restoration, maintenance and accelerated reacquisition. It also explains why forgetting should not automatically be interpreted as complete erasure.

Forgotten is sometimes not gone. It is knowledge whose path back has become weak.

2. Relearning

Relearning means learning material again after performance has weakened or failed. The important question is not simply whether the word can be recalled now, but how much work is required to restore it.

For vocabulary teaching, the practical question is what remains accessible. A learner who cannot freely recall a word may still recognise it, complete it, explain part of it or relearn it quickly. Those residual signals should guide the restoration plan.

3. Savings

The classic savings measure compares the effort required for original learning with the effort required for relearning. If relearning takes fewer repetitions or less time, prior learning left a measurable residual effect even when recall looked poor.

For vocabulary teaching, the practical question is what remains accessible. A learner who cannot freely recall a word may still recognise it, complete it, explain part of it or relearn it quickly. Those residual signals should guide the restoration plan.

4. Recall failure is not erasure

A failed free-recall attempt can reflect poor accessibility rather than total loss of the underlying representation. Recognition, partial form knowledge, cueing or rapid relearning can reveal what remains.

For vocabulary teaching, the practical question is what remains accessible. A learner who cannot freely recall a word may still recognise it, complete it, explain part of it or relearn it quickly. Those residual signals should guide the restoration plan.

5. Residual lexical traces

A forgotten word can retain fragments: spelling familiarity, first sound, semantic category, collocation, translation or an episodic memory of where it was learned. Relearning can rebuild from these fragments.

For vocabulary teaching, the practical question is what remains accessible. A learner who cannot freely recall a word may still recognise it, complete it, explain part of it or relearn it quickly. Those residual signals should guide the restoration plan.

6. Recognition before recall

Recognition usually survives longer than free recall. A learner may know that a form has been seen before without being able to produce its meaning or use.

For vocabulary teaching, the practical question is what remains accessible. A learner who cannot freely recall a word may still recognise it, complete it, explain part of it or relearn it quickly. Those residual signals should guide the restoration plan.

7. Savings and vocabulary retention

Savings complements ordinary delayed tests. Retention asks what is still directly accessible; savings asks whether previous learning reduces the cost of rebuilding what is not directly accessible.

For vocabulary teaching, the practical question is what remains accessible. A learner who cannot freely recall a word may still recognise it, complete it, explain part of it or relearn it quickly. Those residual signals should guide the restoration plan.

8. Savings is graded

Relearning is not all-or-nothing. One learner may restore meaning immediately but need more practice for spelling or collocation. Different lexical components can show different amounts of savings.

For vocabulary teaching, the practical question is what remains accessible. A learner who cannot freely recall a word may still recognise it, complete it, explain part of it or relearn it quickly. Those residual signals should guide the restoration plan.

9. Spacing after relearning

A restored word should not immediately return to permanent daily review. Once relearning succeeds, spaced retrieval should test whether the repaired representation can survive again.

For vocabulary teaching, the practical question is what remains accessible. A learner who cannot freely recall a word may still recognise it, complete it, explain part of it or relearn it quickly. Those residual signals should guide the restoration plan.

10. Interference and relearning

Old competitors can reappear during relearning. A false friend, similar spelling or old wrong definition can be restored along with the target unless the contrast is explicitly repaired.

For vocabulary teaching, the practical question is what remains accessible. A learner who cannot freely recall a word may still recognise it, complete it, explain part of it or relearn it quickly. Those residual signals should guide the restoration plan.

11. Context reinstatement

Returning to an old context can make a forgotten word feel suddenly accessible. That can be useful diagnostically, but the word should eventually work without the reinstated context.

For vocabulary teaching, the practical question is what remains accessible. A learner who cannot freely recall a word may still recognise it, complete it, explain part of it or relearn it quickly. Those residual signals should guide the restoration plan.

12. Cued relearning

First-letter, translation, image, sentence-frame and category cues can expose residual memory. Strong cues should then be faded so relearning ends in independent access.

Relearning should end in independent delayed retrieval rather than permanent cue dependence. The purpose of restoration is to rebuild a usable lexical route, not merely make the item look familiar again.

13. Relearning technical vocabulary

Specialist vocabulary often returns quickly when a learner re-enters a professional or academic domain because concepts and networks remain partly intact even after active use declines.

Relearning should end in independent delayed retrieval rather than permanent cue dependence. The purpose of restoration is to rebuild a usable lexical route, not merely make the item look familiar again.

14. Relearning after language attrition

A second language that has not been used for years may show strong recognition and accelerated relearning. This does not mean every word remained fully stored in accessible form.

Relearning should end in independent delayed retrieval rather than permanent cue dependence. The purpose of restoration is to rebuild a usable lexical route, not merely make the item look familiar again.

15. Relearning and morphology

Known roots, prefixes and families can provide scaffolding when a forgotten derived form is relearned. Morphological structure can lower the restoration cost.

Relearning should end in independent delayed retrieval rather than permanent cue dependence. The purpose of restoration is to rebuild a usable lexical route, not merely make the item look familiar again.

16. Relearning multiword expressions

Idioms, collocations and phrasal verbs may lose one component while the overall meaning remains familiar. Relearning should restore the whole pattern, not only the headword.

Relearning should end in independent delayed retrieval rather than permanent cue dependence. The purpose of restoration is to rebuild a usable lexical route, not merely make the item look familiar again.

17. Relearning pronunciation

A learner may remember a printed word but lose oral access. Relearning can target sound-form recovery separately from semantic recovery.

Relearning should end in independent delayed retrieval rather than permanent cue dependence. The purpose of restoration is to rebuild a usable lexical route, not merely make the item look familiar again.

18. Relearning spelling

Conversely, oral meaning can remain strong while orthographic detail fades. Relearning from dictation or meaning-to-spelling retrieval targets the weakened component.

Relearning should end in independent delayed retrieval rather than permanent cue dependence. The purpose of restoration is to rebuild a usable lexical route, not merely make the item look familiar again.

19. Metacognition

Learners often treat forgetting as proof that study failed. Savings shows a more accurate model: some learning can remain latent and make later rebuilding faster.

Relearning should end in independent delayed retrieval rather than permanent cue dependence. The purpose of restoration is to rebuild a usable lexical route, not merely make the item look familiar again.

20. Assessment

A useful relearning assessment records original learning trials, delayed performance and restoration trials. Even informal teachers can track whether forgotten items return faster than novel ones.

Relearning should end in independent delayed retrieval rather than permanent cue dependence. The purpose of restoration is to rebuild a usable lexical route, not merely make the item look familiar again.

21. Teaching

Do not automatically restart every forgotten word from zero. Probe what remains, rebuild the weakest component, then return the item to spaced independent retrieval.

Relearning should end in independent delayed retrieval rather than permanent cue dependence. The purpose of restoration is to rebuild a usable lexical route, not merely make the item look familiar again.

22. AI-era relearning

AI can instantly re-explain forgotten vocabulary, but useful relearning still requires the learner to reconstruct the form and meaning afterward. External restoration should be followed by internal recall.

Relearning should end in independent delayed retrieval rather than permanent cue dependence. The purpose of restoration is to rebuild a usable lexical route, not merely make the item look familiar again.

23. Relearning casebook — Cases 1–20

1. Travel vocabulary after five years

Residual mechanism: low recent use, strong old episodes. Likely state: recognition may survive while production fades. Restoration move: reactivate with travel contexts, then test without them.

Savings test: compare the number of successful study/retrieval trials needed now with the first learning period if records exist. Even an informal estimate can reveal whether the old learning still reduces cost.

Recognition test: present the form without giving the answer. If the item feels familiar or is correctly identified before meaning returns, some representation remains accessible.

Partial-information test: ask for first sound, syllable count, category, translation, collocation or context. Partial retrieval can identify which component survived.

Restoration criterion: after reactivate with travel contexts, then test without them, remove the cue and require independent retrieval after a short delay.

Future schedule: do not treat the restored word as brand new if savings is strong. Re-enter spaced review at an interval matched to current strength.

2. School science term after a long break

Residual mechanism: conceptual knowledge partly retained. Likely state: technical label may return quickly. Restoration move: reconnect term to concept and diagram.

Savings test: compare the number of successful study/retrieval trials needed now with the first learning period if records exist. Even an informal estimate can reveal whether the old learning still reduces cost.

Recognition test: present the form without giving the answer. If the item feels familiar or is correctly identified before meaning returns, some representation remains accessible.

Partial-information test: ask for first sound, syllable count, category, translation, collocation or context. Partial retrieval can identify which component survived.

Restoration criterion: after reconnect term to concept and diagram, remove the cue and require independent retrieval after a short delay.

Future schedule: do not treat the restored word as brand new if savings is strong. Re-enter spaced review at an interval matched to current strength.

3. Old foreign-language greeting

Residual mechanism: formulaic memory. Likely state: whole phrase may return before individual grammar. Restoration move: retrieve phrase, then analyse components only if useful.

Savings test: compare the number of successful study/retrieval trials needed now with the first learning period if records exist. Even an informal estimate can reveal whether the old learning still reduces cost.

Recognition test: present the form without giving the answer. If the item feels familiar or is correctly identified before meaning returns, some representation remains accessible.

Partial-information test: ask for first sound, syllable count, category, translation, collocation or context. Partial retrieval can identify which component survived.

Restoration criterion: after retrieve phrase, then analyse components only if useful, remove the cue and require independent retrieval after a short delay.

Future schedule: do not treat the restored word as brand new if savings is strong. Re-enter spaced review at an interval matched to current strength.

4. Forgotten irregular verb

Residual mechanism: form-specific decay. Likely state: meaning known, past form inaccessible. Restoration move: contrast base/past/participle and space retrieval.

Savings test: compare the number of successful study/retrieval trials needed now with the first learning period if records exist. Even an informal estimate can reveal whether the old learning still reduces cost.

Recognition test: present the form without giving the answer. If the item feels familiar or is correctly identified before meaning returns, some representation remains accessible.

Partial-information test: ask for first sound, syllable count, category, translation, collocation or context. Partial retrieval can identify which component survived.

Restoration criterion: after contrast base/past/participle and space retrieval, remove the cue and require independent retrieval after a short delay.

Future schedule: do not treat the restored word as brand new if savings is strong. Re-enter spaced review at an interval matched to current strength.

5. Forgotten spelling of familiar word

Residual mechanism: orthographic decay. Likely state: semantic/oral trace strong. Restoration move: spell from meaning and sound.

Savings test: compare the number of successful study/retrieval trials needed now with the first learning period if records exist. Even an informal estimate can reveal whether the old learning still reduces cost.

Recognition test: present the form without giving the answer. If the item feels familiar or is correctly identified before meaning returns, some representation remains accessible.

Partial-information test: ask for first sound, syllable count, category, translation, collocation or context. Partial retrieval can identify which component survived.

Restoration criterion: after spell from meaning and sound, remove the cue and require independent retrieval after a short delay.

Future schedule: do not treat the restored word as brand new if savings is strong. Re-enter spaced review at an interval matched to current strength.

6. Forgotten pronunciation of known print word

Residual mechanism: phonological decay. Likely state: orthographic/semantic trace strong. Restoration move: audio cue then spoken recall.

Savings test: compare the number of successful study/retrieval trials needed now with the first learning period if records exist. Even an informal estimate can reveal whether the old learning still reduces cost.

Recognition test: present the form without giving the answer. If the item feels familiar or is correctly identified before meaning returns, some representation remains accessible.

Partial-information test: ask for first sound, syllable count, category, translation, collocation or context. Partial retrieval can identify which component survived.

Restoration criterion: after audio cue then spoken recall, remove the cue and require independent retrieval after a short delay.

Future schedule: do not treat the restored word as brand new if savings is strong. Re-enter spaced review at an interval matched to current strength.

7. Forgotten collocation

Residual mechanism: phraseological decay. Likely state: individual words still known. Restoration move: restore whole phrase through contrastive examples.

Savings test: compare the number of successful study/retrieval trials needed now with the first learning period if records exist. Even an informal estimate can reveal whether the old learning still reduces cost.

Recognition test: present the form without giving the answer. If the item feels familiar or is correctly identified before meaning returns, some representation remains accessible.

Partial-information test: ask for first sound, syllable count, category, translation, collocation or context. Partial retrieval can identify which component survived.

Restoration criterion: after restore whole phrase through contrastive examples, remove the cue and require independent retrieval after a short delay.

Future schedule: do not treat the restored word as brand new if savings is strong. Re-enter spaced review at an interval matched to current strength.

8. Forgotten idiom

Residual mechanism: multiword semantic decay. Likely state: literal forms available. Restoration move: relearn conventional phrase-level meaning.

Savings test: compare the number of successful study/retrieval trials needed now with the first learning period if records exist. Even an informal estimate can reveal whether the old learning still reduces cost.

Recognition test: present the form without giving the answer. If the item feels familiar or is correctly identified before meaning returns, some representation remains accessible.

Partial-information test: ask for first sound, syllable count, category, translation, collocation or context. Partial retrieval can identify which component survived.

Restoration criterion: after relearn conventional phrase-level meaning, remove the cue and require independent retrieval after a short delay.

Future schedule: do not treat the restored word as brand new if savings is strong. Re-enter spaced review at an interval matched to current strength.

9. Forgotten phrasal verb particle

Residual mechanism: pattern decay. Likely state: verb and meaning partly known. Restoration move: retrieve complete verb+particle unit.

Savings test: compare the number of successful study/retrieval trials needed now with the first learning period if records exist. Even an informal estimate can reveal whether the old learning still reduces cost.

Recognition test: present the form without giving the answer. If the item feels familiar or is correctly identified before meaning returns, some representation remains accessible.

Partial-information test: ask for first sound, syllable count, category, translation, collocation or context. Partial retrieval can identify which component survived.

Restoration criterion: after retrieve complete verb+particle unit, remove the cue and require independent retrieval after a short delay.

Future schedule: do not treat the restored word as brand new if savings is strong. Re-enter spaced review at an interval matched to current strength.

10. Forgotten academic command word

Residual mechanism: instructional access decay. Likely state: school concept partly present. Restoration move: match command to operation, then perform it.

Savings test: compare the number of successful study/retrieval trials needed now with the first learning period if records exist. Even an informal estimate can reveal whether the old learning still reduces cost.

Recognition test: present the form without giving the answer. If the item feels familiar or is correctly identified before meaning returns, some representation remains accessible.

Partial-information test: ask for first sound, syllable count, category, translation, collocation or context. Partial retrieval can identify which component survived.

Restoration criterion: after match command to operation, then perform it, remove the cue and require independent retrieval after a short delay.

Future schedule: do not treat the restored word as brand new if savings is strong. Re-enter spaced review at an interval matched to current strength.

11. Forgotten legal term after career change

Residual mechanism: domain attrition. Likely state: semantic network dormant. Restoration move: reactivate using authentic case material.

Savings test: compare the number of successful study/retrieval trials needed now with the first learning period if records exist. Even an informal estimate can reveal whether the old learning still reduces cost.

Recognition test: present the form without giving the answer. If the item feels familiar or is correctly identified before meaning returns, some representation remains accessible.

Partial-information test: ask for first sound, syllable count, category, translation, collocation or context. Partial retrieval can identify which component survived.

Restoration criterion: after reactivate using authentic case material, remove the cue and require independent retrieval after a short delay.

Future schedule: do not treat the restored word as brand new if savings is strong. Re-enter spaced review at an interval matched to current strength.

12. Forgotten medical abbreviation

Residual mechanism: form compression lost. Likely state: concept may remain. Restoration move: re-expand acronym and reconnect to clinical context.

Savings test: compare the number of successful study/retrieval trials needed now with the first learning period if records exist. Even an informal estimate can reveal whether the old learning still reduces cost.

Recognition test: present the form without giving the answer. If the item feels familiar or is correctly identified before meaning returns, some representation remains accessible.

Partial-information test: ask for first sound, syllable count, category, translation, collocation or context. Partial retrieval can identify which component survived.

Restoration criterion: after re-expand acronym and reconnect to clinical context, remove the cue and require independent retrieval after a short delay.

Future schedule: do not treat the restored word as brand new if savings is strong. Re-enter spaced review at an interval matched to current strength.

13. Forgotten mathematics term

Residual mechanism: label loss. Likely state: procedure still possible. Restoration move: name the concept after solving example.

Savings test: compare the number of successful study/retrieval trials needed now with the first learning period if records exist. Even an informal estimate can reveal whether the old learning still reduces cost.

Recognition test: present the form without giving the answer. If the item feels familiar or is correctly identified before meaning returns, some representation remains accessible.

Partial-information test: ask for first sound, syllable count, category, translation, collocation or context. Partial retrieval can identify which component survived.

Restoration criterion: after name the concept after solving example, remove the cue and require independent retrieval after a short delay.

Future schedule: do not treat the restored word as brand new if savings is strong. Re-enter spaced review at an interval matched to current strength.

14. Forgotten history vocabulary

Residual mechanism: episode-rich, low reuse. Likely state: dates/events may cue term. Restoration move: use multiple new historical contexts.

Savings test: compare the number of successful study/retrieval trials needed now with the first learning period if records exist. Even an informal estimate can reveal whether the old learning still reduces cost.

Recognition test: present the form without giving the answer. If the item feels familiar or is correctly identified before meaning returns, some representation remains accessible.

Partial-information test: ask for first sound, syllable count, category, translation, collocation or context. Partial retrieval can identify which component survived.

Restoration criterion: after use multiple new historical contexts, remove the cue and require independent retrieval after a short delay.

Future schedule: do not treat the restored word as brand new if savings is strong. Re-enter spaced review at an interval matched to current strength.

15. Forgotten literary term

Residual mechanism: metalinguistic attrition. Likely state: examples may still feel familiar. Restoration move: identify device from fresh passages.

Savings test: compare the number of successful study/retrieval trials needed now with the first learning period if records exist. Even an informal estimate can reveal whether the old learning still reduces cost.

Recognition test: present the form without giving the answer. If the item feels familiar or is correctly identified before meaning returns, some representation remains accessible.

Partial-information test: ask for first sound, syllable count, category, translation, collocation or context. Partial retrieval can identify which component survived.

Restoration criterion: after identify device from fresh passages, remove the cue and require independent retrieval after a short delay.

Future schedule: do not treat the restored word as brand new if savings is strong. Re-enter spaced review at an interval matched to current strength.

16. Forgotten childhood L2 word

Residual mechanism: early trace. Likely state: pronunciation may feel familiar. Restoration move: test recognition, translation and production separately.

Savings test: compare the number of successful study/retrieval trials needed now with the first learning period if records exist. Even an informal estimate can reveal whether the old learning still reduces cost.

Recognition test: present the form without giving the answer. If the item feels familiar or is correctly identified before meaning returns, some representation remains accessible.

Partial-information test: ask for first sound, syllable count, category, translation, collocation or context. Partial retrieval can identify which component survived.

Restoration criterion: after test recognition, translation and production separately, remove the cue and require independent retrieval after a short delay.

Future schedule: do not treat the restored word as brand new if savings is strong. Re-enter spaced review at an interval matched to current strength.

17. Forgotten word that instantly feels familiar

Residual mechanism: familiarity without recall. Likely state: residual lexical trace likely. Restoration move: use graded cues, then free retrieval.

Savings test: compare the number of successful study/retrieval trials needed now with the first learning period if records exist. Even an informal estimate can reveal whether the old learning still reduces cost.

Recognition test: present the form without giving the answer. If the item feels familiar or is correctly identified before meaning returns, some representation remains accessible.

Partial-information test: ask for first sound, syllable count, category, translation, collocation or context. Partial retrieval can identify which component survived.

Restoration criterion: after use graded cues, then free retrieval, remove the cue and require independent retrieval after a short delay.

Future schedule: do not treat the restored word as brand new if savings is strong. Re-enter spaced review at an interval matched to current strength.

18. Word remembered after first-letter cue

Residual mechanism: partial phonological access. Likely state: strong savings potential. Restoration move: fade first-letter cue on next trial.

Savings test: compare the number of successful study/retrieval trials needed now with the first learning period if records exist. Even an informal estimate can reveal whether the old learning still reduces cost.

Recognition test: present the form without giving the answer. If the item feels familiar or is correctly identified before meaning returns, some representation remains accessible.

Partial-information test: ask for first sound, syllable count, category, translation, collocation or context. Partial retrieval can identify which component survived.

Restoration criterion: after fade first-letter cue on next trial, remove the cue and require independent retrieval after a short delay.

Future schedule: do not treat the restored word as brand new if savings is strong. Re-enter spaced review at an interval matched to current strength.

19. Word remembered after translation cue

Residual mechanism: cross-language route intact. Likely state: L2 form weaker. Restoration move: switch to meaning-only cue after restoration.

Savings test: compare the number of successful study/retrieval trials needed now with the first learning period if records exist. Even an informal estimate can reveal whether the old learning still reduces cost.

Recognition test: present the form without giving the answer. If the item feels familiar or is correctly identified before meaning returns, some representation remains accessible.

Partial-information test: ask for first sound, syllable count, category, translation, collocation or context. Partial retrieval can identify which component survived.

Restoration criterion: after switch to meaning-only cue after restoration, remove the cue and require independent retrieval after a short delay.

Future schedule: do not treat the restored word as brand new if savings is strong. Re-enter spaced review at an interval matched to current strength.

20. Word remembered after picture cue

Residual mechanism: semantic-perceptual route intact. Likely state: form access weaker. Restoration move: remove image over later trials.

Savings test: compare the number of successful study/retrieval trials needed now with the first learning period if records exist. Even an informal estimate can reveal whether the old learning still reduces cost.

Recognition test: present the form without giving the answer. If the item feels familiar or is correctly identified before meaning returns, some representation remains accessible.

Partial-information test: ask for first sound, syllable count, category, translation, collocation or context. Partial retrieval can identify which component survived.

Restoration criterion: after remove image over later trials, remove the cue and require independent retrieval after a short delay.

Future schedule: do not treat the restored word as brand new if savings is strong. Re-enter spaced review at an interval matched to current strength.

24. Relearning casebook — Cases 21–40

21. Word remembered only in original textbook sentence

Residual mechanism: context-dependent access. Likely state: semanticisation weak. Restoration move: relearn in varied contexts.

Context-reinstatement test: use the old learning environment or cue once to see whether access improves. A large improvement signals cue dependence and should be followed by decontextualised retrieval.

Error-trace test: check whether an old wrong form or meaning returns with the target. Relearning can reactivate errors as well as correct knowledge.

Contrast test: if a competitor is involved, retrieve target and competitor alternately and explain the distinguishing feature.

Mode test: compare print and oral relearning. Savings can be strong in one representation and weak in another.

Generalisation: use the restored word in a fresh context so relearning produces flexible rather than nostalgic knowledge.

22. Word recovered after seeing old notebook

Residual mechanism: episodic reinstatement. Likely state: source context unlocks memory. Restoration move: test later without notebook.

Context-reinstatement test: use the old learning environment or cue once to see whether access improves. A large improvement signals cue dependence and should be followed by decontextualised retrieval.

Error-trace test: check whether an old wrong form or meaning returns with the target. Relearning can reactivate errors as well as correct knowledge.

Contrast test: if a competitor is involved, retrieve target and competitor alternately and explain the distinguishing feature.

Mode test: compare print and oral relearning. Savings can be strong in one representation and weak in another.

Generalisation: use the restored word in a fresh context so relearning produces flexible rather than nostalgic knowledge.

23. Word returns when old song is heard

Residual mechanism: auditory episodic cue. Likely state: phrase memory linked to music. Restoration move: extract and use word outside song.

Context-reinstatement test: use the old learning environment or cue once to see whether access improves. A large improvement signals cue dependence and should be followed by decontextualised retrieval.

Error-trace test: check whether an old wrong form or meaning returns with the target. Relearning can reactivate errors as well as correct knowledge.

Contrast test: if a competitor is involved, retrieve target and competitor alternately and explain the distinguishing feature.

Mode test: compare print and oral relearning. Savings can be strong in one representation and weak in another.

Generalisation: use the restored word in a fresh context so relearning produces flexible rather than nostalgic knowledge.

24. Word returns during conversation with old friend

Residual mechanism: social-context reinstatement. Likely state: interactional episode supports access. Restoration move: retest in neutral context.

Context-reinstatement test: use the old learning environment or cue once to see whether access improves. A large improvement signals cue dependence and should be followed by decontextualised retrieval.

Error-trace test: check whether an old wrong form or meaning returns with the target. Relearning can reactivate errors as well as correct knowledge.

Contrast test: if a competitor is involved, retrieve target and competitor alternately and explain the distinguishing feature.

Mode test: compare print and oral relearning. Savings can be strong in one representation and weak in another.

Generalisation: use the restored word in a fresh context so relearning produces flexible rather than nostalgic knowledge.

25. Word returns in old workplace

Residual mechanism: environment reinstatement. Likely state: place-dependent retrieval. Restoration move: use term away from workplace.

Context-reinstatement test: use the old learning environment or cue once to see whether access improves. A large improvement signals cue dependence and should be followed by decontextualised retrieval.

Error-trace test: check whether an old wrong form or meaning returns with the target. Relearning can reactivate errors as well as correct knowledge.

Contrast test: if a competitor is involved, retrieve target and competitor alternately and explain the distinguishing feature.

Mode test: compare print and oral relearning. Savings can be strong in one representation and weak in another.

Generalisation: use the restored word in a fresh context so relearning produces flexible rather than nostalgic knowledge.

26. Forgotten word relearned in one exposure

Residual mechanism: large savings. Likely state: strong residual memory. Restoration move: extend spacing quickly.

Context-reinstatement test: use the old learning environment or cue once to see whether access improves. A large improvement signals cue dependence and should be followed by decontextualised retrieval.

Error-trace test: check whether an old wrong form or meaning returns with the target. Relearning can reactivate errors as well as correct knowledge.

Contrast test: if a competitor is involved, retrieve target and competitor alternately and explain the distinguishing feature.

Mode test: compare print and oral relearning. Savings can be strong in one representation and weak in another.

Generalisation: use the restored word in a fresh context so relearning produces flexible rather than nostalgic knowledge.

27. Forgotten word needs half original trials

Residual mechanism: moderate savings. Likely state: partial trace. Restoration move: restore then use spaced retrieval.

Context-reinstatement test: use the old learning environment or cue once to see whether access improves. A large improvement signals cue dependence and should be followed by decontextualised retrieval.

Error-trace test: check whether an old wrong form or meaning returns with the target. Relearning can reactivate errors as well as correct knowledge.

Contrast test: if a competitor is involved, retrieve target and competitor alternately and explain the distinguishing feature.

Mode test: compare print and oral relearning. Savings can be strong in one representation and weak in another.

Generalisation: use the restored word in a fresh context so relearning produces flexible rather than nostalgic knowledge.

28. Forgotten word needs same trials as first time

Residual mechanism: little observable savings. Likely state: weak residual trace or changed conditions. Restoration move: re-encode carefully and inspect interference.

Context-reinstatement test: use the old learning environment or cue once to see whether access improves. A large improvement signals cue dependence and should be followed by decontextualised retrieval.

Error-trace test: check whether an old wrong form or meaning returns with the target. Relearning can reactivate errors as well as correct knowledge.

Contrast test: if a competitor is involved, retrieve target and competitor alternately and explain the distinguishing feature.

Mode test: compare print and oral relearning. Savings can be strong in one representation and weak in another.

Generalisation: use the restored word in a fresh context so relearning produces flexible rather than nostalgic knowledge.

29. Forgotten word takes longer than first time

Residual mechanism: negative transfer/interference. Likely state: old wrong mapping may compete. Restoration move: diagnose competitor before more practice.

Context-reinstatement test: use the old learning environment or cue once to see whether access improves. A large improvement signals cue dependence and should be followed by decontextualised retrieval.

Error-trace test: check whether an old wrong form or meaning returns with the target. Relearning can reactivate errors as well as correct knowledge.

Contrast test: if a competitor is involved, retrieve target and competitor alternately and explain the distinguishing feature.

Mode test: compare print and oral relearning. Savings can be strong in one representation and weak in another.

Generalisation: use the restored word in a fresh context so relearning produces flexible rather than nostalgic knowledge.

30. False friend error returns during relearning

Residual mechanism: old competing mapping. Likely state: interference survives. Restoration move: contrast meanings explicitly.

Context-reinstatement test: use the old learning environment or cue once to see whether access improves. A large improvement signals cue dependence and should be followed by decontextualised retrieval.

Error-trace test: check whether an old wrong form or meaning returns with the target. Relearning can reactivate errors as well as correct knowledge.

Contrast test: if a competitor is involved, retrieve target and competitor alternately and explain the distinguishing feature.

Mode test: compare print and oral relearning. Savings can be strong in one representation and weak in another.

Generalisation: use the restored word in a fresh context so relearning produces flexible rather than nostalgic knowledge.

31. Old misspelling returns

Residual mechanism: error trace. Likely state: form competition. Restoration move: use errorless spelling retrieval plus feedback.

Context-reinstatement test: use the old learning environment or cue once to see whether access improves. A large improvement signals cue dependence and should be followed by decontextualised retrieval.

Error-trace test: check whether an old wrong form or meaning returns with the target. Relearning can reactivate errors as well as correct knowledge.

Contrast test: if a competitor is involved, retrieve target and competitor alternately and explain the distinguishing feature.

Mode test: compare print and oral relearning. Savings can be strong in one representation and weak in another.

Generalisation: use the restored word in a fresh context so relearning produces flexible rather than nostalgic knowledge.

32. Old wrong pronunciation returns

Residual mechanism: phonological habit. Likely state: articulatory pattern entrenched. Restoration move: contrast old/new sound with recording.

Context-reinstatement test: use the old learning environment or cue once to see whether access improves. A large improvement signals cue dependence and should be followed by decontextualised retrieval.

Error-trace test: check whether an old wrong form or meaning returns with the target. Relearning can reactivate errors as well as correct knowledge.

Contrast test: if a competitor is involved, retrieve target and competitor alternately and explain the distinguishing feature.

Mode test: compare print and oral relearning. Savings can be strong in one representation and weak in another.

Generalisation: use the restored word in a fresh context so relearning produces flexible rather than nostalgic knowledge.

33. Old wrong collocation returns

Residual mechanism: phrase habit. Likely state: competition at phrase level. Restoration move: retrieve target collocation against wrong neighbour.

Context-reinstatement test: use the old learning environment or cue once to see whether access improves. A large improvement signals cue dependence and should be followed by decontextualised retrieval.

Error-trace test: check whether an old wrong form or meaning returns with the target. Relearning can reactivate errors as well as correct knowledge.

Contrast test: if a competitor is involved, retrieve target and competitor alternately and explain the distinguishing feature.

Mode test: compare print and oral relearning. Savings can be strong in one representation and weak in another.

Generalisation: use the restored word in a fresh context so relearning produces flexible rather than nostalgic knowledge.

34. Technical word relearned through morphology

Residual mechanism: structural savings. Likely state: root family remains. Restoration move: reconstruct derived meaning from known morphemes.

Context-reinstatement test: use the old learning environment or cue once to see whether access improves. A large improvement signals cue dependence and should be followed by decontextualised retrieval.

Error-trace test: check whether an old wrong form or meaning returns with the target. Relearning can reactivate errors as well as correct knowledge.

Contrast test: if a competitor is involved, retrieve target and competitor alternately and explain the distinguishing feature.

Mode test: compare print and oral relearning. Savings can be strong in one representation and weak in another.

Generalisation: use the restored word in a fresh context so relearning produces flexible rather than nostalgic knowledge.

35. New word related to forgotten family

Residual mechanism: family reactivation. Likely state: old network supports learning. Restoration move: map base, derivatives and new item.

Context-reinstatement test: use the old learning environment or cue once to see whether access improves. A large improvement signals cue dependence and should be followed by decontextualised retrieval.

Error-trace test: check whether an old wrong form or meaning returns with the target. Relearning can reactivate errors as well as correct knowledge.

Contrast test: if a competitor is involved, retrieve target and competitor alternately and explain the distinguishing feature.

Mode test: compare print and oral relearning. Savings can be strong in one representation and weak in another.

Generalisation: use the restored word in a fresh context so relearning produces flexible rather than nostalgic knowledge.

36. Relearning several similar synonyms

Residual mechanism: semantic competition. Likely state: boundaries blur. Restoration move: restore one at a time then contrast.

Context-reinstatement test: use the old learning environment or cue once to see whether access improves. A large improvement signals cue dependence and should be followed by decontextualised retrieval.

Error-trace test: check whether an old wrong form or meaning returns with the target. Relearning can reactivate errors as well as correct knowledge.

Contrast test: if a competitor is involved, retrieve target and competitor alternately and explain the distinguishing feature.

Mode test: compare print and oral relearning. Savings can be strong in one representation and weak in another.

Generalisation: use the restored word in a fresh context so relearning produces flexible rather than nostalgic knowledge.

37. Relearning antonym pair

Residual mechanism: relational support. Likely state: each can cue the other. Restoration move: test both independent and paired recall.

Context-reinstatement test: use the old learning environment or cue once to see whether access improves. A large improvement signals cue dependence and should be followed by decontextualised retrieval.

Error-trace test: check whether an old wrong form or meaning returns with the target. Relearning can reactivate errors as well as correct knowledge.

Contrast test: if a competitor is involved, retrieve target and competitor alternately and explain the distinguishing feature.

Mode test: compare print and oral relearning. Savings can be strong in one representation and weak in another.

Generalisation: use the restored word in a fresh context so relearning produces flexible rather than nostalgic knowledge.

38. Relearning category members

Residual mechanism: semantic grouping. Likely state: category cue helps but competition rises. Restoration move: alternate category cue with item-specific cues.

Context-reinstatement test: use the old learning environment or cue once to see whether access improves. A large improvement signals cue dependence and should be followed by decontextualised retrieval.

Error-trace test: check whether an old wrong form or meaning returns with the target. Relearning can reactivate errors as well as correct knowledge.

Contrast test: if a competitor is involved, retrieve target and competitor alternately and explain the distinguishing feature.

Mode test: compare print and oral relearning. Savings can be strong in one representation and weak in another.

Generalisation: use the restored word in a fresh context so relearning produces flexible rather than nostalgic knowledge.

39. Relearning with flashcards

Residual mechanism: strong item cue. Likely state: efficient initial restoration. Restoration move: fade card prompt into free production.

Context-reinstatement test: use the old learning environment or cue once to see whether access improves. A large improvement signals cue dependence and should be followed by decontextualised retrieval.

Error-trace test: check whether an old wrong form or meaning returns with the target. Relearning can reactivate errors as well as correct knowledge.

Contrast test: if a competitor is involved, retrieve target and competitor alternately and explain the distinguishing feature.

Mode test: compare print and oral relearning. Savings can be strong in one representation and weak in another.

Generalisation: use the restored word in a fresh context so relearning produces flexible rather than nostalgic knowledge.

40. Relearning through reading

Residual mechanism: context-rich restoration. Likely state: meaning inferred before form recall. Restoration move: follow with productive retrieval.

Context-reinstatement test: use the old learning environment or cue once to see whether access improves. A large improvement signals cue dependence and should be followed by decontextualised retrieval.

Error-trace test: check whether an old wrong form or meaning returns with the target. Relearning can reactivate errors as well as correct knowledge.

Contrast test: if a competitor is involved, retrieve target and competitor alternately and explain the distinguishing feature.

Mode test: compare print and oral relearning. Savings can be strong in one representation and weak in another.

Generalisation: use the restored word in a fresh context so relearning produces flexible rather than nostalgic knowledge.

25. Relearning casebook — Cases 41–60

41. Relearning through listening

Residual mechanism: phonological restoration. Likely state: spoken form reactivates. Restoration move: add spelling if print use matters.

Efficiency test: record how much support is actually needed. Relearning should exploit surviving structure rather than repeat the entire original lesson automatically.

Depth test: determine whether restoration has recovered only a rough meaning or also grammar, collocation, spelling, pronunciation and register.

Delay test: retest after several days. Rapid same-session recovery can reflect renewed activation rather than durable restoration.

Maintenance decision: decide whether the item’s future value justifies productive maintenance. Some vocabulary can return to receptive-only status.

Transfer: use the same savings diagnostic on another apparently forgotten word and compare restoration cost.

42. Relearning through writing

Residual mechanism: productive restoration. Likely state: form and grammar exposed. Restoration move: include later oral test.

Efficiency test: record how much support is actually needed. Relearning should exploit surviving structure rather than repeat the entire original lesson automatically.

Depth test: determine whether restoration has recovered only a rough meaning or also grammar, collocation, spelling, pronunciation and register.

Delay test: retest after several days. Rapid same-session recovery can reflect renewed activation rather than durable restoration.

Maintenance decision: decide whether the item’s future value justifies productive maintenance. Some vocabulary can return to receptive-only status.

Transfer: use the same savings diagnostic on another apparently forgotten word and compare restoration cost.

43. Relearning through speaking

Residual mechanism: fast-access restoration. Likely state: pronunciation and retrieval practised. Restoration move: add written form later.

Efficiency test: record how much support is actually needed. Relearning should exploit surviving structure rather than repeat the entire original lesson automatically.

Depth test: determine whether restoration has recovered only a rough meaning or also grammar, collocation, spelling, pronunciation and register.

Delay test: retest after several days. Rapid same-session recovery can reflect renewed activation rather than durable restoration.

Maintenance decision: decide whether the item’s future value justifies productive maintenance. Some vocabulary can return to receptive-only status.

Transfer: use the same savings diagnostic on another apparently forgotten word and compare restoration cost.

44. Relearning through translation

Residual mechanism: concept bridge. Likely state: fast semantic route. Restoration move: avoid permanent translation dependence.

Efficiency test: record how much support is actually needed. Relearning should exploit surviving structure rather than repeat the entire original lesson automatically.

Depth test: determine whether restoration has recovered only a rough meaning or also grammar, collocation, spelling, pronunciation and register.

Delay test: retest after several days. Rapid same-session recovery can reflect renewed activation rather than durable restoration.

Maintenance decision: decide whether the item’s future value justifies productive maintenance. Some vocabulary can return to receptive-only status.

Transfer: use the same savings diagnostic on another apparently forgotten word and compare restoration cost.

45. Relearning with definition

Residual mechanism: semantic route. Likely state: precise but possibly narrow. Restoration move: add example and collocation.

Efficiency test: record how much support is actually needed. Relearning should exploit surviving structure rather than repeat the entire original lesson automatically.

Depth test: determine whether restoration has recovered only a rough meaning or also grammar, collocation, spelling, pronunciation and register.

Delay test: retest after several days. Rapid same-session recovery can reflect renewed activation rather than durable restoration.

Maintenance decision: decide whether the item’s future value justifies productive maintenance. Some vocabulary can return to receptive-only status.

Transfer: use the same savings diagnostic on another apparently forgotten word and compare restoration cost.

46. Relearning with example sentence

Residual mechanism: context route. Likely state: usage pattern restored. Restoration move: later change sentence.

Efficiency test: record how much support is actually needed. Relearning should exploit surviving structure rather than repeat the entire original lesson automatically.

Depth test: determine whether restoration has recovered only a rough meaning or also grammar, collocation, spelling, pronunciation and register.

Delay test: retest after several days. Rapid same-session recovery can reflect renewed activation rather than durable restoration.

Maintenance decision: decide whether the item’s future value justifies productive maintenance. Some vocabulary can return to receptive-only status.

Transfer: use the same savings diagnostic on another apparently forgotten word and compare restoration cost.

47. Relearning with image

Residual mechanism: perceptual route. Likely state: strong for concrete nouns. Restoration move: add language-only retrieval.

Efficiency test: record how much support is actually needed. Relearning should exploit surviving structure rather than repeat the entire original lesson automatically.

Depth test: determine whether restoration has recovered only a rough meaning or also grammar, collocation, spelling, pronunciation and register.

Delay test: retest after several days. Rapid same-session recovery can reflect renewed activation rather than durable restoration.

Maintenance decision: decide whether the item’s future value justifies productive maintenance. Some vocabulary can return to receptive-only status.

Transfer: use the same savings diagnostic on another apparently forgotten word and compare restoration cost.

48. Relearning abstract vocabulary

Residual mechanism: conceptual route. Likely state: image support limited. Restoration move: use analogy, contrast and examples.

Efficiency test: record how much support is actually needed. Relearning should exploit surviving structure rather than repeat the entire original lesson automatically.

Depth test: determine whether restoration has recovered only a rough meaning or also grammar, collocation, spelling, pronunciation and register.

Delay test: retest after several days. Rapid same-session recovery can reflect renewed activation rather than durable restoration.

Maintenance decision: decide whether the item’s future value justifies productive maintenance. Some vocabulary can return to receptive-only status.

Transfer: use the same savings diagnostic on another apparently forgotten word and compare restoration cost.

49. Relearning multi-sense core word

Residual mechanism: sense network. Likely state: one sense may cue others. Restoration move: test each sense separately.

Efficiency test: record how much support is actually needed. Relearning should exploit surviving structure rather than repeat the entire original lesson automatically.

Depth test: determine whether restoration has recovered only a rough meaning or also grammar, collocation, spelling, pronunciation and register.

Delay test: retest after several days. Rapid same-session recovery can reflect renewed activation rather than durable restoration.

Maintenance decision: decide whether the item’s future value justifies productive maintenance. Some vocabulary can return to receptive-only status.

Transfer: use the same savings diagnostic on another apparently forgotten word and compare restoration cost.

50. Relearning rare specialist word

Residual mechanism: low natural frequency. Likely state: maintenance cost high. Restoration move: decide receptive vs productive goal.

Efficiency test: record how much support is actually needed. Relearning should exploit surviving structure rather than repeat the entire original lesson automatically.

Depth test: determine whether restoration has recovered only a rough meaning or also grammar, collocation, spelling, pronunciation and register.

Delay test: retest after several days. Rapid same-session recovery can reflect renewed activation rather than durable restoration.

Maintenance decision: decide whether the item’s future value justifies productive maintenance. Some vocabulary can return to receptive-only status.

Transfer: use the same savings diagnostic on another apparently forgotten word and compare restoration cost.

51. Relearning common word

Residual mechanism: high natural recurrence. Likely state: environment should help maintain. Restoration move: find why natural exposure did not suffice.

Efficiency test: record how much support is actually needed. Relearning should exploit surviving structure rather than repeat the entire original lesson automatically.

Depth test: determine whether restoration has recovered only a rough meaning or also grammar, collocation, spelling, pronunciation and register.

Delay test: retest after several days. Rapid same-session recovery can reflect renewed activation rather than durable restoration.

Maintenance decision: decide whether the item’s future value justifies productive maintenance. Some vocabulary can return to receptive-only status.

Transfer: use the same savings diagnostic on another apparently forgotten word and compare restoration cost.

52. Relearning after exam cramming

Residual mechanism: weak long-term trace. Likely state: some rapid restoration possible. Restoration move: space later reviews.

Efficiency test: record how much support is actually needed. Relearning should exploit surviving structure rather than repeat the entire original lesson automatically.

Depth test: determine whether restoration has recovered only a rough meaning or also grammar, collocation, spelling, pronunciation and register.

Delay test: retest after several days. Rapid same-session recovery can reflect renewed activation rather than durable restoration.

Maintenance decision: decide whether the item’s future value justifies productive maintenance. Some vocabulary can return to receptive-only status.

Transfer: use the same savings diagnostic on another apparently forgotten word and compare restoration cost.

53. Relearning after spaced study

Residual mechanism: stronger residual trace expected. Likely state: faster restoration. Restoration move: compare trials with crammed items.

Efficiency test: record how much support is actually needed. Relearning should exploit surviving structure rather than repeat the entire original lesson automatically.

Depth test: determine whether restoration has recovered only a rough meaning or also grammar, collocation, spelling, pronunciation and register.

Delay test: retest after several days. Rapid same-session recovery can reflect renewed activation rather than durable restoration.

Maintenance decision: decide whether the item’s future value justifies productive maintenance. Some vocabulary can return to receptive-only status.

Transfer: use the same savings diagnostic on another apparently forgotten word and compare restoration cost.

54. Relearning after sleep

Residual mechanism: consolidation may help. Likely state: integration stronger. Restoration move: test long-delay savings.

Efficiency test: record how much support is actually needed. Relearning should exploit surviving structure rather than repeat the entire original lesson automatically.

Depth test: determine whether restoration has recovered only a rough meaning or also grammar, collocation, spelling, pronunciation and register.

Delay test: retest after several days. Rapid same-session recovery can reflect renewed activation rather than durable restoration.

Maintenance decision: decide whether the item’s future value justifies productive maintenance. Some vocabulary can return to receptive-only status.

Transfer: use the same savings diagnostic on another apparently forgotten word and compare restoration cost.

55. Relearning after interference-heavy study

Residual mechanism: competitors remain active. Likely state: restoration unstable. Restoration move: separate and discriminate.

Efficiency test: record how much support is actually needed. Relearning should exploit surviving structure rather than repeat the entire original lesson automatically.

Depth test: determine whether restoration has recovered only a rough meaning or also grammar, collocation, spelling, pronunciation and register.

Delay test: retest after several days. Rapid same-session recovery can reflect renewed activation rather than durable restoration.

Maintenance decision: decide whether the item’s future value justifies productive maintenance. Some vocabulary can return to receptive-only status.

Transfer: use the same savings diagnostic on another apparently forgotten word and compare restoration cost.

56. Relearning in same context

Residual mechanism: context support high. Likely state: may overestimate independence. Restoration move: later switch context.

Efficiency test: record how much support is actually needed. Relearning should exploit surviving structure rather than repeat the entire original lesson automatically.

Depth test: determine whether restoration has recovered only a rough meaning or also grammar, collocation, spelling, pronunciation and register.

Delay test: retest after several days. Rapid same-session recovery can reflect renewed activation rather than durable restoration.

Maintenance decision: decide whether the item’s future value justifies productive maintenance. Some vocabulary can return to receptive-only status.

Transfer: use the same savings diagnostic on another apparently forgotten word and compare restoration cost.

57. Relearning in varied contexts

Residual mechanism: multiple routes. Likely state: generalisation stronger. Restoration move: maintain stable core meaning.

Efficiency test: record how much support is actually needed. Relearning should exploit surviving structure rather than repeat the entire original lesson automatically.

Depth test: determine whether restoration has recovered only a rough meaning or also grammar, collocation, spelling, pronunciation and register.

Delay test: retest after several days. Rapid same-session recovery can reflect renewed activation rather than durable restoration.

Maintenance decision: decide whether the item’s future value justifies productive maintenance. Some vocabulary can return to receptive-only status.

Transfer: use the same savings diagnostic on another apparently forgotten word and compare restoration cost.

58. Relearning under time pressure

Residual mechanism: access demand high. Likely state: latent knowledge may look lost. Restoration move: restore untimed then speed up.

Efficiency test: record how much support is actually needed. Relearning should exploit surviving structure rather than repeat the entire original lesson automatically.

Depth test: determine whether restoration has recovered only a rough meaning or also grammar, collocation, spelling, pronunciation and register.

Delay test: retest after several days. Rapid same-session recovery can reflect renewed activation rather than durable restoration.

Maintenance decision: decide whether the item’s future value justifies productive maintenance. Some vocabulary can return to receptive-only status.

Transfer: use the same savings diagnostic on another apparently forgotten word and compare restoration cost.

59. Relearning with AI explanation

Residual mechanism: external reconstruction. Likely state: understanding restored rapidly. Restoration move: close AI and retrieve independently.

Efficiency test: record how much support is actually needed. Relearning should exploit surviving structure rather than repeat the entire original lesson automatically.

Depth test: determine whether restoration has recovered only a rough meaning or also grammar, collocation, spelling, pronunciation and register.

Delay test: retest after several days. Rapid same-session recovery can reflect renewed activation rather than durable restoration.

Maintenance decision: decide whether the item’s future value justifies productive maintenance. Some vocabulary can return to receptive-only status.

Transfer: use the same savings diagnostic on another apparently forgotten word and compare restoration cost.

60. Relearning after many years

Residual mechanism: long-term savings test. Likely state: partial traces may surprise learner. Restoration move: measure recognition and learning rate rather than assuming zero.

Efficiency test: record how much support is actually needed. Relearning should exploit surviving structure rather than repeat the entire original lesson automatically.

Depth test: determine whether restoration has recovered only a rough meaning or also grammar, collocation, spelling, pronunciation and register.

Delay test: retest after several days. Rapid same-session recovery can reflect renewed activation rather than durable restoration.

Maintenance decision: decide whether the item’s future value justifies productive maintenance. Some vocabulary can return to receptive-only status.

Transfer: use the same savings diagnostic on another apparently forgotten word and compare restoration cost.

26. A practical savings protocol

1. Travel vocabulary after five years

Probe before reteaching: test recognition, partial form, category, translation, collocation and original context before showing the answer. The goal is to discover surviving traces rather than assume zero knowledge.

Minimal restoration: give only enough support to restore a correct response. In this case, begin with reactivate with travel contexts, then test without them. If the answer returns, reduce support immediately so the learner reconstructs the route.

Measure savings: compare restoration effort with a truly novel item of similar difficulty. Faster recovery is evidence that the earlier learning left a functional residue.