Semantic vs Episodic Memory | How Word Knowledge Becomes General

SEMANTIC MEMORY · EPISODIC MEMORY · WORD LEARNING · LEXICAL CONSOLIDATION · CONTEXT · SEMANTICISATION · GENERALISATION

Episodic memory preserves specific experiences: the story, lesson, conversation, place or event in which a word was encountered. Semantic memory preserves general knowledge that is less tied to one episode: what a word means, how it relates to other concepts and how it behaves across contexts.

New vocabulary often begins episodically. The learner remembers the sentence, teacher, picture or situation that carried the new form. With repetition and consolidation, useful regularities can become integrated into semantic memory so the word works outside the original episode.

This guide explains episodic word learning, semantic memory, consolidation, complementary learning systems, decontextualisation, context diversity, source memory, familiarity, semantic networks and lexical integration. Existing eduKateSG specialist pages on lexical consolidation and sleep remain untouched.

A word begins as something that happened. Mature vocabulary becomes something you know.

2. Episodic memory

Episodic memory preserves information tied to particular experiences: where, when and how a word was encountered. Early word learning often depends heavily on such event-specific traces.

For learning, the practical question is whether the word still needs the original episode as a cue. Early dependence is normal; mature lexical knowledge should increasingly survive when the story, image or sentence changes.

3. Semantic memory

Semantic memory stores generalised knowledge that is less tied to a single episode. Stable word meanings, category relations and familiar facts are examples of semantic knowledge.

For learning, the practical question is whether the word still needs the original episode as a cue. Early dependence is normal; mature lexical knowledge should increasingly survive when the story, image or sentence changes.

4. From episode to lexical knowledge

New words often begin as fragile episode-bound traces and gradually become integrated with prior knowledge. Repeated encounters help extract what remains stable across episodes.

For learning, the practical question is whether the word still needs the original episode as a cue. Early dependence is normal; mature lexical knowledge should increasingly survive when the story, image or sentence changes.

5. Complementary learning systems

Contemporary word-learning models distinguish rapid initial learning from slower consolidation and lexical integration. The systems cooperate rather than replace one another.

For learning, the practical question is whether the word still needs the original episode as a cue. Early dependence is normal; mature lexical knowledge should increasingly survive when the story, image or sentence changes.

6. Context binding

Remembering the sentence, teacher or story where a word appeared can help retrieval early on. Later, strong lexical knowledge should survive when those episodic details are absent.

For learning, the practical question is whether the word still needs the original episode as a cue. Early dependence is normal; mature lexical knowledge should increasingly survive when the story, image or sentence changes.

7. Decontextualisation

Semanticisation is the process by which knowledge becomes less dependent on the original event. For vocabulary, this means understanding a word beyond the first example.

For learning, the practical question is whether the word still needs the original episode as a cue. Early dependence is normal; mature lexical knowledge should increasingly survive when the story, image or sentence changes.

8. Repeated encounters

Multiple episodes allow learners to compare contexts and infer a stable semantic core. One encounter may be memorable but still too narrow.

For learning, the practical question is whether the word still needs the original episode as a cue. Early dependence is normal; mature lexical knowledge should increasingly survive when the story, image or sentence changes.

9. Context diversity

Different contexts can promote generalisation by varying surface details while preserving word meaning. Variation should increase only after initial form–meaning learning is stable enough.

For learning, the practical question is whether the word still needs the original episode as a cue. Early dependence is normal; mature lexical knowledge should increasingly survive when the story, image or sentence changes.

10. Sleep and consolidation

Sleep has been linked to lexical integration and consolidation. Offline processing can help newly learned forms interact with established lexical knowledge.

For learning, the practical question is whether the word still needs the original episode as a cue. Early dependence is normal; mature lexical knowledge should increasingly survive when the story, image or sentence changes.

11. Semantic networks

As a word becomes integrated, it acquires relations to synonyms, categories, collocations, morphology and concepts. Those links make retrieval less dependent on one episodic cue.

For learning, the practical question is whether the word still needs the original episode as a cue. Early dependence is normal; mature lexical knowledge should increasingly survive when the story, image or sentence changes.

12. Episodic advantage

Vivid episodes can create strong memory hooks. Personal relevance, emotion and distinctive context can improve recall without automatically producing full lexical depth.

This distinction should be tested by changing context. If the learner can preserve meaning and use across new episodes, semantic integration is stronger than simple recollection of the original encounter.

13. Episodic trap

A learner may remember the lesson story but not the word outside that story. Memorable context is useful only if the lexical target can later detach from it.

This distinction should be tested by changing context. If the learner can preserve meaning and use across new episodes, semantic integration is stronger than simple recollection of the original encounter.

14. Recollection and familiarity

A word can feel familiar without recollection of where it was learned. Conversely, a learner can remember a specific encounter yet still have weak general meaning knowledge.

This distinction should be tested by changing context. If the learner can preserve meaning and use across new episodes, semantic integration is stronger than simple recollection of the original encounter.

15. Recognition and recall

Recognition can rely partly on familiarity; productive recall often needs a stronger mapping from concept to form. Semantic integration improves flexible access.

This distinction should be tested by changing context. If the learner can preserve meaning and use across new episodes, semantic integration is stronger than simple recollection of the original encounter.

16. Expertise

Experts possess dense semantic networks in their domain, making technical terms easier to integrate. Novices often rely more on explicit episodes and definitions.

This distinction should be tested by changing context. If the learner can preserve meaning and use across new episodes, semantic integration is stronger than simple recollection of the original encounter.

17. Cross-language knowledge

Known concepts in another language can supply semantic structure for a new lexical form. The new target-language word still needs independent form, grammar and usage representation.

This distinction should be tested by changing context. If the learner can preserve meaning and use across new episodes, semantic integration is stronger than simple recollection of the original encounter.

18. Forgetting

Episodic details often fade faster than useful semantic knowledge. This is productive if the learner retains the general lexical structure rather than every learning event.

This distinction should be tested by changing context. If the learner can preserve meaning and use across new episodes, semantic integration is stronger than simple recollection of the original encounter.

19. Assessment

A strong vocabulary assessment should distinguish event memory from generalised knowledge. Repeating the original example tests something different from using the word in a new context.

This distinction should be tested by changing context. If the learner can preserve meaning and use across new episodes, semantic integration is stronger than simple recollection of the original encounter.

20. Teaching

Use memorable initial episodes, then deliberately vary examples and cues. The goal is to exploit episodic memory without trapping the word inside one event.

This distinction should be tested by changing context. If the learner can preserve meaning and use across new episodes, semantic integration is stronger than simple recollection of the original encounter.

21. AI-era memory

AI can create unlimited example episodes. The learning value comes from comparing and retrieving across them, not from consuming many examples without forming stable semantic structure.

This distinction should be tested by changing context. If the learner can preserve meaning and use across new episodes, semantic integration is stronger than simple recollection of the original encounter.

22. Memory-system casebook — Cases 1–20

1. Learner remembers teacher’s funny story with the word

Memory profile: episodic-rich. Likely state: strong event cue, uncertain generalisation. Next move: reuse word in unrelated context.

Episode test: ask what details of the learning event are still remembered. Strong source recollection can support learning but should not be mistaken for complete lexical knowledge.

Semantic test: remove the original context and ask for definition, example and category relation. If knowledge collapses, the word remains episode-bound.

Form test: check spelling and pronunciation independently. Semantic knowledge can generalise while exact form still depends on an episodic cue.

Generalisation test: use the word in a new topic. Stable meaning across episodes is evidence of semanticisation.

Delay test: retest later. Episodic vividness and semantic durability can decay at different rates.

2. Learner forgets story but defines word correctly

Memory profile: semanticised. Likely state: episode faded, concept retained. Next move: test production and collocation.

Episode test: ask what details of the learning event are still remembered. Strong source recollection can support learning but should not be mistaken for complete lexical knowledge.

Semantic test: remove the original context and ask for definition, example and category relation. If knowledge collapses, the word remains episode-bound.

Form test: check spelling and pronunciation independently. Semantic knowledge can generalise while exact form still depends on an episodic cue.

Generalisation test: use the word in a new topic. Stable meaning across episodes is evidence of semanticisation.

Delay test: retest later. Episodic vividness and semantic durability can decay at different rates.

3. Learner recognises word only in original sentence

Memory profile: context-bound episodic. Likely state: weak decontextualisation. Next move: vary syntax and topic.

Episode test: ask what details of the learning event are still remembered. Strong source recollection can support learning but should not be mistaken for complete lexical knowledge.

Semantic test: remove the original context and ask for definition, example and category relation. If knowledge collapses, the word remains episode-bound.

Form test: check spelling and pronunciation independently. Semantic knowledge can generalise while exact form still depends on an episodic cue.

Generalisation test: use the word in a new topic. Stable meaning across episodes is evidence of semanticisation.

Delay test: retest later. Episodic vividness and semantic durability can decay at different rates.

4. Learner recognises word across many texts

Memory profile: semantic integration. Likely state: stable cross-context recognition. Next move: add productive retrieval.

Episode test: ask what details of the learning event are still remembered. Strong source recollection can support learning but should not be mistaken for complete lexical knowledge.

Semantic test: remove the original context and ask for definition, example and category relation. If knowledge collapses, the word remains episode-bound.

Form test: check spelling and pronunciation independently. Semantic knowledge can generalise while exact form still depends on an episodic cue.

Generalisation test: use the word in a new topic. Stable meaning across episodes is evidence of semanticisation.

Delay test: retest later. Episodic vividness and semantic durability can decay at different rates.

5. Learner remembers picture but not spelling

Memory profile: episodic visual cue. Likely state: form binding incomplete. Next move: spell from meaning after picture removed.

Episode test: ask what details of the learning event are still remembered. Strong source recollection can support learning but should not be mistaken for complete lexical knowledge.

Semantic test: remove the original context and ask for definition, example and category relation. If knowledge collapses, the word remains episode-bound.

Form test: check spelling and pronunciation independently. Semantic knowledge can generalise while exact form still depends on an episodic cue.

Generalisation test: use the word in a new topic. Stable meaning across episodes is evidence of semanticisation.

Delay test: retest later. Episodic vividness and semantic durability can decay at different rates.

6. Learner remembers spelling but not meaning

Memory profile: form trace. Likely state: semantic integration weak. Next move: rebuild concept across examples.

Episode test: ask what details of the learning event are still remembered. Strong source recollection can support learning but should not be mistaken for complete lexical knowledge.

Semantic test: remove the original context and ask for definition, example and category relation. If knowledge collapses, the word remains episode-bound.

Form test: check spelling and pronunciation independently. Semantic knowledge can generalise while exact form still depends on an episodic cue.

Generalisation test: use the word in a new topic. Stable meaning across episodes is evidence of semanticisation.

Delay test: retest later. Episodic vividness and semantic durability can decay at different rates.

7. Learner recalls exact flashcard location

Memory profile: spatial episodic cue. Likely state: memory tied to study event. Next move: shuffle cards and cues.

Episode test: ask what details of the learning event are still remembered. Strong source recollection can support learning but should not be mistaken for complete lexical knowledge.

Semantic test: remove the original context and ask for definition, example and category relation. If knowledge collapses, the word remains episode-bound.

Form test: check spelling and pronunciation independently. Semantic knowledge can generalise while exact form still depends on an episodic cue.

Generalisation test: use the word in a new topic. Stable meaning across episodes is evidence of semanticisation.

Delay test: retest later. Episodic vividness and semantic durability can decay at different rates.

8. Learner recalls word after seeing same colour highlight

Memory profile: perceptual episode cue. Likely state: surface dependence. Next move: remove formatting.

Episode test: ask what details of the learning event are still remembered. Strong source recollection can support learning but should not be mistaken for complete lexical knowledge.

Semantic test: remove the original context and ask for definition, example and category relation. If knowledge collapses, the word remains episode-bound.

Form test: check spelling and pronunciation independently. Semantic knowledge can generalise while exact form still depends on an episodic cue.

Generalisation test: use the word in a new topic. Stable meaning across episodes is evidence of semanticisation.

Delay test: retest later. Episodic vividness and semantic durability can decay at different rates.

9. Learner retrieves word from definition only

Memory profile: single semantic cue. Likely state: narrow access. Next move: add examples, category and contrast cues.

Episode test: ask what details of the learning event are still remembered. Strong source recollection can support learning but should not be mistaken for complete lexical knowledge.

Semantic test: remove the original context and ask for definition, example and category relation. If knowledge collapses, the word remains episode-bound.

Form test: check spelling and pronunciation independently. Semantic knowledge can generalise while exact form still depends on an episodic cue.

Generalisation test: use the word in a new topic. Stable meaning across episodes is evidence of semanticisation.

Delay test: retest later. Episodic vividness and semantic durability can decay at different rates.

10. Learner retrieves from several cue types

Memory profile: multi-route semantic memory. Likely state: flexible access. Next move: increase delay.

Episode test: ask what details of the learning event are still remembered. Strong source recollection can support learning but should not be mistaken for complete lexical knowledge.

Semantic test: remove the original context and ask for definition, example and category relation. If knowledge collapses, the word remains episode-bound.

Form test: check spelling and pronunciation independently. Semantic knowledge can generalise while exact form still depends on an episodic cue.

Generalisation test: use the word in a new topic. Stable meaning across episodes is evidence of semanticisation.

Delay test: retest later. Episodic vividness and semantic durability can decay at different rates.

11. Learner learns technical term during lab

Memory profile: episodic + domain. Likely state: concrete procedural anchor. Next move: revisit term in textbook explanation.

Episode test: ask what details of the learning event are still remembered. Strong source recollection can support learning but should not be mistaken for complete lexical knowledge.

Semantic test: remove the original context and ask for definition, example and category relation. If knowledge collapses, the word remains episode-bound.

Form test: check spelling and pronunciation independently. Semantic knowledge can generalise while exact form still depends on an episodic cue.

Generalisation test: use the word in a new topic. Stable meaning across episodes is evidence of semanticisation.

Delay test: retest later. Episodic vividness and semantic durability can decay at different rates.

12. Learner learns legal term from one case

Memory profile: episodic case binding. Likely state: risk of overgeneralising one instance. Next move: compare multiple cases.

Episode test: ask what details of the learning event are still remembered. Strong source recollection can support learning but should not be mistaken for complete lexical knowledge.

Semantic test: remove the original context and ask for definition, example and category relation. If knowledge collapses, the word remains episode-bound.

Form test: check spelling and pronunciation independently. Semantic knowledge can generalise while exact form still depends on an episodic cue.

Generalisation test: use the word in a new topic. Stable meaning across episodes is evidence of semanticisation.

Delay test: retest later. Episodic vividness and semantic durability can decay at different rates.

13. Learner learns emotion word from personal event

Memory profile: self-referential episode. Likely state: strong memory hook. Next move: generalise to other people and contexts.

Episode test: ask what details of the learning event are still remembered. Strong source recollection can support learning but should not be mistaken for complete lexical knowledge.

Semantic test: remove the original context and ask for definition, example and category relation. If knowledge collapses, the word remains episode-bound.

Form test: check spelling and pronunciation independently. Semantic knowledge can generalise while exact form still depends on an episodic cue.

Generalisation test: use the word in a new topic. Stable meaning across episodes is evidence of semanticisation.

Delay test: retest later. Episodic vividness and semantic durability can decay at different rates.

14. Learner learns travel word on trip

Memory profile: situated episode. Likely state: high context support. Next move: test after returning home.

Episode test: ask what details of the learning event are still remembered. Strong source recollection can support learning but should not be mistaken for complete lexical knowledge.

Semantic test: remove the original context and ask for definition, example and category relation. If knowledge collapses, the word remains episode-bound.

Form test: check spelling and pronunciation independently. Semantic knowledge can generalise while exact form still depends on an episodic cue.

Generalisation test: use the word in a new topic. Stable meaning across episodes is evidence of semanticisation.

Delay test: retest later. Episodic vividness and semantic durability can decay at different rates.

15. Learner learns food word while tasting it

Memory profile: multisensory episode. Likely state: strong perceptual encoding. Next move: test print and oral form later.

Episode test: ask what details of the learning event are still remembered. Strong source recollection can support learning but should not be mistaken for complete lexical knowledge.

Semantic test: remove the original context and ask for definition, example and category relation. If knowledge collapses, the word remains episode-bound.

Form test: check spelling and pronunciation independently. Semantic knowledge can generalise while exact form still depends on an episodic cue.

Generalisation test: use the word in a new topic. Stable meaning across episodes is evidence of semanticisation.

Delay test: retest later. Episodic vividness and semantic durability can decay at different rates.

16. Learner learns abstract word from analogy

Memory profile: constructed episode. Likely state: analogy supports concept. Next move: test without analogy.

Episode test: ask what details of the learning event are still remembered. Strong source recollection can support learning but should not be mistaken for complete lexical knowledge.

Semantic test: remove the original context and ask for definition, example and category relation. If knowledge collapses, the word remains episode-bound.

Form test: check spelling and pronunciation independently. Semantic knowledge can generalise while exact form still depends on an episodic cue.

Generalisation test: use the word in a new topic. Stable meaning across episodes is evidence of semanticisation.

Delay test: retest later. Episodic vividness and semantic durability can decay at different rates.

17. Learner learns word from dramatic novel scene

Memory profile: narrative episode. Likely state: emotionally distinctive context. Next move: use in expository context.

Episode test: ask what details of the learning event are still remembered. Strong source recollection can support learning but should not be mistaken for complete lexical knowledge.

Semantic test: remove the original context and ask for definition, example and category relation. If knowledge collapses, the word remains episode-bound.

Form test: check spelling and pronunciation independently. Semantic knowledge can generalise while exact form still depends on an episodic cue.

Generalisation test: use the word in a new topic. Stable meaning across episodes is evidence of semanticisation.

Delay test: retest later. Episodic vividness and semantic durability can decay at different rates.

18. Learner learns academic word from worksheet

Memory profile: instruction episode. Likely state: school cue dependence possible. Next move: move to authentic text.

Episode test: ask what details of the learning event are still remembered. Strong source recollection can support learning but should not be mistaken for complete lexical knowledge.

Semantic test: remove the original context and ask for definition, example and category relation. If knowledge collapses, the word remains episode-bound.

Form test: check spelling and pronunciation independently. Semantic knowledge can generalise while exact form still depends on an episodic cue.

Generalisation test: use the word in a new topic. Stable meaning across episodes is evidence of semanticisation.

Delay test: retest later. Episodic vividness and semantic durability can decay at different rates.

19. Learner hears word in several conversations

Memory profile: distributed episodes. Likely state: natural semanticisation. Next move: test meaning without context.

Episode test: ask what details of the learning event are still remembered. Strong source recollection can support learning but should not be mistaken for complete lexical knowledge.

Semantic test: remove the original context and ask for definition, example and category relation. If knowledge collapses, the word remains episode-bound.

Form test: check spelling and pronunciation independently. Semantic knowledge can generalise while exact form still depends on an episodic cue.

Generalisation test: use the word in a new topic. Stable meaning across episodes is evidence of semanticisation.

Delay test: retest later. Episodic vividness and semantic durability can decay at different rates.

20. Learner sees word repeatedly in one article

Memory profile: local repetition. Likely state: familiarity without diversity. Next move: add different sources.

Episode test: ask what details of the learning event are still remembered. Strong source recollection can support learning but should not be mistaken for complete lexical knowledge.

Semantic test: remove the original context and ask for definition, example and category relation. If knowledge collapses, the word remains episode-bound.

Form test: check spelling and pronunciation independently. Semantic knowledge can generalise while exact form still depends on an episodic cue.

Generalisation test: use the word in a new topic. Stable meaning across episodes is evidence of semanticisation.

Delay test: retest later. Episodic vividness and semantic durability can decay at different rates.

23. Memory-system casebook — Cases 21–40

21. Learner sees word in six topics

Memory profile: context diversity. Likely state: semantic core extraction. Next move: compare invariant meaning.

Invariant extraction: compare several encounters and identify what remains constant about the word’s meaning. Semantic memory grows when variable episode details are separated from stable lexical information.

Network test: connect the item to a category, contrast, collocation or morphological relative. Integrated semantic relations reduce dependence on source memory.

Source-memory test: deliberately ask where the word was learned. Losing the source is not necessarily a problem if general meaning and use remain strong.

Recognition versus recollection: note whether the word merely feels familiar or evokes a specific episode. Both can support performance, but neither alone proves deep lexical knowledge.

Transfer: generate a fresh example in a different mode—spoken if learned in print, written if learned orally.

22. Learner learns synonym pair together

Memory profile: semantic network. Likely state: relation-based memory. Next move: test independent use.

Invariant extraction: compare several encounters and identify what remains constant about the word’s meaning. Semantic memory grows when variable episode details are separated from stable lexical information.

Network test: connect the item to a category, contrast, collocation or morphological relative. Integrated semantic relations reduce dependence on source memory.

Source-memory test: deliberately ask where the word was learned. Losing the source is not necessarily a problem if general meaning and use remain strong.

Recognition versus recollection: note whether the word merely feels familiar or evokes a specific episode. Both can support performance, but neither alone proves deep lexical knowledge.

Transfer: generate a fresh example in a different mode—spoken if learned in print, written if learned orally.

23. Learner learns category hierarchy

Memory profile: semantic organisation. Likely state: structured integration. Next move: retrieve examples and superordinates.

Invariant extraction: compare several encounters and identify what remains constant about the word’s meaning. Semantic memory grows when variable episode details are separated from stable lexical information.

Network test: connect the item to a category, contrast, collocation or morphological relative. Integrated semantic relations reduce dependence on source memory.

Source-memory test: deliberately ask where the word was learned. Losing the source is not necessarily a problem if general meaning and use remain strong.

Recognition versus recollection: note whether the word merely feels familiar or evokes a specific episode. Both can support performance, but neither alone proves deep lexical knowledge.

Transfer: generate a fresh example in a different mode—spoken if learned in print, written if learned orally.

24. Learner learns collocation

Memory profile: phrase-level semantic integration. Likely state: partner cue strong. Next move: test alternate grammatical contexts.

Invariant extraction: compare several encounters and identify what remains constant about the word’s meaning. Semantic memory grows when variable episode details are separated from stable lexical information.

Network test: connect the item to a category, contrast, collocation or morphological relative. Integrated semantic relations reduce dependence on source memory.

Source-memory test: deliberately ask where the word was learned. Losing the source is not necessarily a problem if general meaning and use remain strong.

Recognition versus recollection: note whether the word merely feels familiar or evokes a specific episode. Both can support performance, but neither alone proves deep lexical knowledge.

Transfer: generate a fresh example in a different mode—spoken if learned in print, written if learned orally.

25. Learner learns morphology family

Memory profile: structured lexical network. Likely state: shared root helps. Next move: separate word-class meanings.

Invariant extraction: compare several encounters and identify what remains constant about the word’s meaning. Semantic memory grows when variable episode details are separated from stable lexical information.

Network test: connect the item to a category, contrast, collocation or morphological relative. Integrated semantic relations reduce dependence on source memory.

Source-memory test: deliberately ask where the word was learned. Losing the source is not necessarily a problem if general meaning and use remain strong.

Recognition versus recollection: note whether the word merely feels familiar or evokes a specific episode. Both can support performance, but neither alone proves deep lexical knowledge.

Transfer: generate a fresh example in a different mode—spoken if learned in print, written if learned orally.

26. Learner sleeps after new-word study

Memory profile: consolidation opportunity. Likely state: possible integration support. Next move: retest next day and later.

Invariant extraction: compare several encounters and identify what remains constant about the word’s meaning. Semantic memory grows when variable episode details are separated from stable lexical information.

Network test: connect the item to a category, contrast, collocation or morphological relative. Integrated semantic relations reduce dependence on source memory.

Source-memory test: deliberately ask where the word was learned. Losing the source is not necessarily a problem if general meaning and use remain strong.

Recognition versus recollection: note whether the word merely feels familiar or evokes a specific episode. Both can support performance, but neither alone proves deep lexical knowledge.

Transfer: generate a fresh example in a different mode—spoken if learned in print, written if learned orally.

27. Learner studies continuously without break

Memory profile: high recent activation. Likely state: less offline separation. Next move: add spaced session later.

Invariant extraction: compare several encounters and identify what remains constant about the word’s meaning. Semantic memory grows when variable episode details are separated from stable lexical information.

Network test: connect the item to a category, contrast, collocation or morphological relative. Integrated semantic relations reduce dependence on source memory.

Source-memory test: deliberately ask where the word was learned. Losing the source is not necessarily a problem if general meaning and use remain strong.

Recognition versus recollection: note whether the word merely feels familiar or evokes a specific episode. Both can support performance, but neither alone proves deep lexical knowledge.

Transfer: generate a fresh example in a different mode—spoken if learned in print, written if learned orally.

28. Learner learns word then many similar words

Memory profile: interference pressure. Likely state: new trace vulnerable. Next move: space and contrast.

Invariant extraction: compare several encounters and identify what remains constant about the word’s meaning. Semantic memory grows when variable episode details are separated from stable lexical information.

Network test: connect the item to a category, contrast, collocation or morphological relative. Integrated semantic relations reduce dependence on source memory.

Source-memory test: deliberately ask where the word was learned. Losing the source is not necessarily a problem if general meaning and use remain strong.

Recognition versus recollection: note whether the word merely feels familiar or evokes a specific episode. Both can support performance, but neither alone proves deep lexical knowledge.

Transfer: generate a fresh example in a different mode—spoken if learned in print, written if learned orally.

29. Learner learns word then retrieves it next day

Memory profile: reactivated trace. Likely state: consolidation + retrieval. Next move: extend interval.

Invariant extraction: compare several encounters and identify what remains constant about the word’s meaning. Semantic memory grows when variable episode details are separated from stable lexical information.

Network test: connect the item to a category, contrast, collocation or morphological relative. Integrated semantic relations reduce dependence on source memory.

Source-memory test: deliberately ask where the word was learned. Losing the source is not necessarily a problem if general meaning and use remain strong.

Recognition versus recollection: note whether the word merely feels familiar or evokes a specific episode. Both can support performance, but neither alone proves deep lexical knowledge.

Transfer: generate a fresh example in a different mode—spoken if learned in print, written if learned orally.

30. Learner forgets learning source

Memory profile: source memory loss. Likely state: can coexist with preserved semantics. Next move: judge by lexical use.

Invariant extraction: compare several encounters and identify what remains constant about the word’s meaning. Semantic memory grows when variable episode details are separated from stable lexical information.

Network test: connect the item to a category, contrast, collocation or morphological relative. Integrated semantic relations reduce dependence on source memory.

Source-memory test: deliberately ask where the word was learned. Losing the source is not necessarily a problem if general meaning and use remain strong.

Recognition versus recollection: note whether the word merely feels familiar or evokes a specific episode. Both can support performance, but neither alone proves deep lexical knowledge.

Transfer: generate a fresh example in a different mode—spoken if learned in print, written if learned orally.

31. Learner remembers source but not word

Memory profile: episode survives, lexical mapping weak. Likely state: event stronger than target. Next move: re-encode form–meaning.

Invariant extraction: compare several encounters and identify what remains constant about the word’s meaning. Semantic memory grows when variable episode details are separated from stable lexical information.

Network test: connect the item to a category, contrast, collocation or morphological relative. Integrated semantic relations reduce dependence on source memory.

Source-memory test: deliberately ask where the word was learned. Losing the source is not necessarily a problem if general meaning and use remain strong.

Recognition versus recollection: note whether the word merely feels familiar or evokes a specific episode. Both can support performance, but neither alone proves deep lexical knowledge.

Transfer: generate a fresh example in a different mode—spoken if learned in print, written if learned orally.

32. Learner says ‘I know I’ve seen it’

Memory profile: familiarity. Likely state: partial lexical signal. Next move: test actual meaning.

Invariant extraction: compare several encounters and identify what remains constant about the word’s meaning. Semantic memory grows when variable episode details are separated from stable lexical information.

Network test: connect the item to a category, contrast, collocation or morphological relative. Integrated semantic relations reduce dependence on source memory.

Source-memory test: deliberately ask where the word was learned. Losing the source is not necessarily a problem if general meaning and use remain strong.

Recognition versus recollection: note whether the word merely feels familiar or evokes a specific episode. Both can support performance, but neither alone proves deep lexical knowledge.

Transfer: generate a fresh example in a different mode—spoken if learned in print, written if learned orally.

33. Learner experiences tip-of-the-tongue

Memory profile: semantic access without full form. Likely state: partial retrieval. Next move: use phonological retrieval practice.

Invariant extraction: compare several encounters and identify what remains constant about the word’s meaning. Semantic memory grows when variable episode details are separated from stable lexical information.

Network test: connect the item to a category, contrast, collocation or morphological relative. Integrated semantic relations reduce dependence on source memory.

Source-memory test: deliberately ask where the word was learned. Losing the source is not necessarily a problem if general meaning and use remain strong.

Recognition versus recollection: note whether the word merely feels familiar or evokes a specific episode. Both can support performance, but neither alone proves deep lexical knowledge.

Transfer: generate a fresh example in a different mode—spoken if learned in print, written if learned orally.

34. Learner recognises technical word from profession

Memory profile: semantic domain familiarity. Likely state: expert network. Next move: test novice explanation.

Invariant extraction: compare several encounters and identify what remains constant about the word’s meaning. Semantic memory grows when variable episode details are separated from stable lexical information.

Network test: connect the item to a category, contrast, collocation or morphological relative. Integrated semantic relations reduce dependence on source memory.

Source-memory test: deliberately ask where the word was learned. Losing the source is not necessarily a problem if general meaning and use remain strong.

Recognition versus recollection: note whether the word merely feels familiar or evokes a specific episode. Both can support performance, but neither alone proves deep lexical knowledge.

Transfer: generate a fresh example in a different mode—spoken if learned in print, written if learned orally.

35. Learner knows concept in L1

Memory profile: semantic concept pre-exists. Likely state: target form new. Next move: bind L2 form and usage.

Invariant extraction: compare several encounters and identify what remains constant about the word’s meaning. Semantic memory grows when variable episode details are separated from stable lexical information.

Network test: connect the item to a category, contrast, collocation or morphological relative. Integrated semantic relations reduce dependence on source memory.

Source-memory test: deliberately ask where the word was learned. Losing the source is not necessarily a problem if general meaning and use remain strong.

Recognition versus recollection: note whether the word merely feels familiar or evokes a specific episode. Both can support performance, but neither alone proves deep lexical knowledge.

Transfer: generate a fresh example in a different mode—spoken if learned in print, written if learned orally.

36. Learner knows translation but no collocations

Memory profile: semantic mapping narrow. Likely state: target-language lexical network incomplete. Next move: add authentic phrases.

Invariant extraction: compare several encounters and identify what remains constant about the word’s meaning. Semantic memory grows when variable episode details are separated from stable lexical information.

Network test: connect the item to a category, contrast, collocation or morphological relative. Integrated semantic relations reduce dependence on source memory.

Source-memory test: deliberately ask where the word was learned. Losing the source is not necessarily a problem if general meaning and use remain strong.

Recognition versus recollection: note whether the word merely feels familiar or evokes a specific episode. Both can support performance, but neither alone proves deep lexical knowledge.

Transfer: generate a fresh example in a different mode—spoken if learned in print, written if learned orally.

37. Learner remembers example but overapplies meaning

Memory profile: episodic overgeneralisation. Likely state: category boundary weak. Next move: use non-examples.

Invariant extraction: compare several encounters and identify what remains constant about the word’s meaning. Semantic memory grows when variable episode details are separated from stable lexical information.

Network test: connect the item to a category, contrast, collocation or morphological relative. Integrated semantic relations reduce dependence on source memory.

Source-memory test: deliberately ask where the word was learned. Losing the source is not necessarily a problem if general meaning and use remain strong.

Recognition versus recollection: note whether the word merely feels familiar or evokes a specific episode. Both can support performance, but neither alone proves deep lexical knowledge.

Transfer: generate a fresh example in a different mode—spoken if learned in print, written if learned orally.

38. Learner learns several senses separately

Memory profile: multiple episodic traces. Likely state: possible fragmented polysemy. Next move: connect senses conceptually.

Invariant extraction: compare several encounters and identify what remains constant about the word’s meaning. Semantic memory grows when variable episode details are separated from stable lexical information.

Network test: connect the item to a category, contrast, collocation or morphological relative. Integrated semantic relations reduce dependence on source memory.

Source-memory test: deliberately ask where the word was learned. Losing the source is not necessarily a problem if general meaning and use remain strong.

Recognition versus recollection: note whether the word merely feels familiar or evokes a specific episode. Both can support performance, but neither alone proves deep lexical knowledge.

Transfer: generate a fresh example in a different mode—spoken if learned in print, written if learned orally.

39. Learner integrates metaphorical and literal senses

Memory profile: semantic network enrichment. Likely state: deeper lexical representation. Next move: test context selection.

Invariant extraction: compare several encounters and identify what remains constant about the word’s meaning. Semantic memory grows when variable episode details are separated from stable lexical information.

Network test: connect the item to a category, contrast, collocation or morphological relative. Integrated semantic relations reduce dependence on source memory.

Source-memory test: deliberately ask where the word was learned. Losing the source is not necessarily a problem if general meaning and use remain strong.

Recognition versus recollection: note whether the word merely feels familiar or evokes a specific episode. Both can support performance, but neither alone proves deep lexical knowledge.

Transfer: generate a fresh example in a different mode—spoken if learned in print, written if learned orally.

40. Learner remembers word only with image

Memory profile: image-dependent cue. Likely state: semantic route not autonomous. Next move: fade picture.

Invariant extraction: compare several encounters and identify what remains constant about the word’s meaning. Semantic memory grows when variable episode details are separated from stable lexical information.

Network test: connect the item to a category, contrast, collocation or morphological relative. Integrated semantic relations reduce dependence on source memory.

Source-memory test: deliberately ask where the word was learned. Losing the source is not necessarily a problem if general meaning and use remain strong.

Recognition versus recollection: note whether the word merely feels familiar or evokes a specific episode. Both can support performance, but neither alone proves deep lexical knowledge.

Transfer: generate a fresh example in a different mode—spoken if learned in print, written if learned orally.

24. Memory-system casebook — Cases 41–60

41. Learner remembers only with first letter

Memory profile: phonological cue dependence. Likely state: partial form access. Next move: fade cue.

Cue fading: remove image, teacher, topic and location cues gradually. The lexical representation should become retrievable from meaning and language context rather than from one event.

Usage integration: add grammar, collocation and register information. Semantic memory for a word includes more than an abstract definition; it includes how the lexical item participates in the language system.

Competing-memory check: identify older meanings, L1 equivalents or near-synonyms that may be interfering. Integration sometimes requires reorganising prior knowledge, not simply adding a new trace.

Consolidation check: compare immediate performance with performance after sleep or a longer interval. Stable later access suggests the new representation has become less dependent on recent activation.

Mastery criterion: treat the word as semantically integrated when it can be interpreted, retrieved and adapted without reconstructing the original learning episode.

42. Learner remembers only after teacher prompt

Memory profile: social/context cue dependence. Likely state: low independence. Next move: self-test alone.

Cue fading: remove image, teacher, topic and location cues gradually. The lexical representation should become retrievable from meaning and language context rather than from one event.

Usage integration: add grammar, collocation and register information. Semantic memory for a word includes more than an abstract definition; it includes how the lexical item participates in the language system.

Competing-memory check: identify older meanings, L1 equivalents or near-synonyms that may be interfering. Integration sometimes requires reorganising prior knowledge, not simply adding a new trace.

Consolidation check: compare immediate performance with performance after sleep or a longer interval. Stable later access suggests the new representation has become less dependent on recent activation.

Mastery criterion: treat the word as semantically integrated when it can be interpreted, retrieved and adapted without reconstructing the original learning episode.

43. Learner explains word in own words

Memory profile: generative semantic encoding. Likely state: stronger conceptualisation. Next move: retest after delay.

Cue fading: remove image, teacher, topic and location cues gradually. The lexical representation should become retrievable from meaning and language context rather than from one event.

Usage integration: add grammar, collocation and register information. Semantic memory for a word includes more than an abstract definition; it includes how the lexical item participates in the language system.

Competing-memory check: identify older meanings, L1 equivalents or near-synonyms that may be interfering. Integration sometimes requires reorganising prior knowledge, not simply adding a new trace.

Consolidation check: compare immediate performance with performance after sleep or a longer interval. Stable later access suggests the new representation has become less dependent on recent activation.

Mastery criterion: treat the word as semantically integrated when it can be interpreted, retrieved and adapted without reconstructing the original learning episode.

44. Learner compares word with near-synonym

Memory profile: contrastive semantic encoding. Likely state: boundary clarified. Next move: use both in new examples.

Cue fading: remove image, teacher, topic and location cues gradually. The lexical representation should become retrievable from meaning and language context rather than from one event.

Usage integration: add grammar, collocation and register information. Semantic memory for a word includes more than an abstract definition; it includes how the lexical item participates in the language system.

Competing-memory check: identify older meanings, L1 equivalents or near-synonyms that may be interfering. Integration sometimes requires reorganising prior knowledge, not simply adding a new trace.

Consolidation check: compare immediate performance with performance after sleep or a longer interval. Stable later access suggests the new representation has become less dependent on recent activation.

Mastery criterion: treat the word as semantically integrated when it can be interpreted, retrieved and adapted without reconstructing the original learning episode.

45. Learner produces novel example

Memory profile: semantic transfer. Likely state: generalised knowledge. Next move: vary genre.

Cue fading: remove image, teacher, topic and location cues gradually. The lexical representation should become retrievable from meaning and language context rather than from one event.

Usage integration: add grammar, collocation and register information. Semantic memory for a word includes more than an abstract definition; it includes how the lexical item participates in the language system.

Competing-memory check: identify older meanings, L1 equivalents or near-synonyms that may be interfering. Integration sometimes requires reorganising prior knowledge, not simply adding a new trace.

Consolidation check: compare immediate performance with performance after sleep or a longer interval. Stable later access suggests the new representation has become less dependent on recent activation.

Mastery criterion: treat the word as semantically integrated when it can be interpreted, retrieved and adapted without reconstructing the original learning episode.

46. Learner uses word correctly months later

Memory profile: durable semantic integration. Likely state: strong long-term result. Next move: reduce deliberate review.

Cue fading: remove image, teacher, topic and location cues gradually. The lexical representation should become retrievable from meaning and language context rather than from one event.

Usage integration: add grammar, collocation and register information. Semantic memory for a word includes more than an abstract definition; it includes how the lexical item participates in the language system.

Competing-memory check: identify older meanings, L1 equivalents or near-synonyms that may be interfering. Integration sometimes requires reorganising prior knowledge, not simply adding a new trace.

Consolidation check: compare immediate performance with performance after sleep or a longer interval. Stable later access suggests the new representation has become less dependent on recent activation.

Mastery criterion: treat the word as semantically integrated when it can be interpreted, retrieved and adapted without reconstructing the original learning episode.

47. Learner relearns forgotten word quickly

Memory profile: residual trace. Likely state: partial memory survived. Next move: exploit savings.

Cue fading: remove image, teacher, topic and location cues gradually. The lexical representation should become retrievable from meaning and language context rather than from one event.

Usage integration: add grammar, collocation and register information. Semantic memory for a word includes more than an abstract definition; it includes how the lexical item participates in the language system.

Competing-memory check: identify older meanings, L1 equivalents or near-synonyms that may be interfering. Integration sometimes requires reorganising prior knowledge, not simply adding a new trace.

Consolidation check: compare immediate performance with performance after sleep or a longer interval. Stable later access suggests the new representation has become less dependent on recent activation.

Mastery criterion: treat the word as semantically integrated when it can be interpreted, retrieved and adapted without reconstructing the original learning episode.

48. Learner remembers definition but not pronunciation

Memory profile: semantic > phonological memory. Likely state: representations uneven. Next move: oral retrieval.

Cue fading: remove image, teacher, topic and location cues gradually. The lexical representation should become retrievable from meaning and language context rather than from one event.

Usage integration: add grammar, collocation and register information. Semantic memory for a word includes more than an abstract definition; it includes how the lexical item participates in the language system.

Competing-memory check: identify older meanings, L1 equivalents or near-synonyms that may be interfering. Integration sometimes requires reorganising prior knowledge, not simply adding a new trace.

Consolidation check: compare immediate performance with performance after sleep or a longer interval. Stable later access suggests the new representation has become less dependent on recent activation.

Mastery criterion: treat the word as semantically integrated when it can be interpreted, retrieved and adapted without reconstructing the original learning episode.

49. Learner remembers pronunciation but not spelling

Memory profile: phonological > orthographic memory. Likely state: print mapping weak. Next move: dictation/reading.

Cue fading: remove image, teacher, topic and location cues gradually. The lexical representation should become retrievable from meaning and language context rather than from one event.

Usage integration: add grammar, collocation and register information. Semantic memory for a word includes more than an abstract definition; it includes how the lexical item participates in the language system.

Competing-memory check: identify older meanings, L1 equivalents or near-synonyms that may be interfering. Integration sometimes requires reorganising prior knowledge, not simply adding a new trace.

Consolidation check: compare immediate performance with performance after sleep or a longer interval. Stable later access suggests the new representation has become less dependent on recent activation.

Mastery criterion: treat the word as semantically integrated when it can be interpreted, retrieved and adapted without reconstructing the original learning episode.

50. Learner remembers collocation but not isolated definition

Memory profile: chunk-supported memory. Likely state: phrase route strong. Next move: extract core meaning too.

Cue fading: remove image, teacher, topic and location cues gradually. The lexical representation should become retrievable from meaning and language context rather than from one event.

Usage integration: add grammar, collocation and register information. Semantic memory for a word includes more than an abstract definition; it includes how the lexical item participates in the language system.

Competing-memory check: identify older meanings, L1 equivalents or near-synonyms that may be interfering. Integration sometimes requires reorganising prior knowledge, not simply adding a new trace.

Consolidation check: compare immediate performance with performance after sleep or a longer interval. Stable later access suggests the new representation has become less dependent on recent activation.

Mastery criterion: treat the word as semantically integrated when it can be interpreted, retrieved and adapted without reconstructing the original learning episode.

51. Learner remembers idiom meaning as whole

Memory profile: multiword semantic unit. Likely state: component words insufficient. Next move: test phrase recognition and use.

Cue fading: remove image, teacher, topic and location cues gradually. The lexical representation should become retrievable from meaning and language context rather than from one event.

Usage integration: add grammar, collocation and register information. Semantic memory for a word includes more than an abstract definition; it includes how the lexical item participates in the language system.

Competing-memory check: identify older meanings, L1 equivalents or near-synonyms that may be interfering. Integration sometimes requires reorganising prior knowledge, not simply adding a new trace.

Consolidation check: compare immediate performance with performance after sleep or a longer interval. Stable later access suggests the new representation has become less dependent on recent activation.

Mastery criterion: treat the word as semantically integrated when it can be interpreted, retrieved and adapted without reconstructing the original learning episode.

52. Learner learns word from video captions

Memory profile: multimodal episode. Likely state: sound-print-context bundle. Next move: later separate modes.

Cue fading: remove image, teacher, topic and location cues gradually. The lexical representation should become retrievable from meaning and language context rather than from one event.

Usage integration: add grammar, collocation and register information. Semantic memory for a word includes more than an abstract definition; it includes how the lexical item participates in the language system.

Competing-memory check: identify older meanings, L1 equivalents or near-synonyms that may be interfering. Integration sometimes requires reorganising prior knowledge, not simply adding a new trace.

Consolidation check: compare immediate performance with performance after sleep or a longer interval. Stable later access suggests the new representation has become less dependent on recent activation.

Mastery criterion: treat the word as semantically integrated when it can be interpreted, retrieved and adapted without reconstructing the original learning episode.

53. Learner learns word from game

Memory profile: goal-rich episode. Likely state: context highly engaging. Next move: test outside game.

Cue fading: remove image, teacher, topic and location cues gradually. The lexical representation should become retrievable from meaning and language context rather than from one event.

Usage integration: add grammar, collocation and register information. Semantic memory for a word includes more than an abstract definition; it includes how the lexical item participates in the language system.

Competing-memory check: identify older meanings, L1 equivalents or near-synonyms that may be interfering. Integration sometimes requires reorganising prior knowledge, not simply adding a new trace.

Consolidation check: compare immediate performance with performance after sleep or a longer interval. Stable later access suggests the new representation has become less dependent on recent activation.

Mastery criterion: treat the word as semantically integrated when it can be interpreted, retrieved and adapted without reconstructing the original learning episode.

54. Learner learns word from AI examples

Memory profile: many synthetic episodes. Likely state: quantity without guaranteed integration. Next move: compare meanings and retrieve independently.

Cue fading: remove image, teacher, topic and location cues gradually. The lexical representation should become retrievable from meaning and language context rather than from one event.

Usage integration: add grammar, collocation and register information. Semantic memory for a word includes more than an abstract definition; it includes how the lexical item participates in the language system.

Competing-memory check: identify older meanings, L1 equivalents or near-synonyms that may be interfering. Integration sometimes requires reorganising prior knowledge, not simply adding a new trace.

Consolidation check: compare immediate performance with performance after sleep or a longer interval. Stable later access suggests the new representation has become less dependent on recent activation.

Mastery criterion: treat the word as semantically integrated when it can be interpreted, retrieved and adapted without reconstructing the original learning episode.

55. Learner generates own AI-assisted contrast set

Memory profile: active semantic organisation. Likely state: better potential integration. Next move: verify and delayed test.

Cue fading: remove image, teacher, topic and location cues gradually. The lexical representation should become retrievable from meaning and language context rather than from one event.

Usage integration: add grammar, collocation and register information. Semantic memory for a word includes more than an abstract definition; it includes how the lexical item participates in the language system.

Competing-memory check: identify older meanings, L1 equivalents or near-synonyms that may be interfering. Integration sometimes requires reorganising prior knowledge, not simply adding a new trace.

Consolidation check: compare immediate performance with performance after sleep or a longer interval. Stable later access suggests the new representation has become less dependent on recent activation.

Mastery criterion: treat the word as semantically integrated when it can be interpreted, retrieved and adapted without reconstructing the original learning episode.

56. Learner recalls word during exam but not later

Memory profile: context reinstatement. Likely state: exam cue support possible. Next move: test after context change.

Cue fading: remove image, teacher, topic and location cues gradually. The lexical representation should become retrievable from meaning and language context rather than from one event.

Usage integration: add grammar, collocation and register information. Semantic memory for a word includes more than an abstract definition; it includes how the lexical item participates in the language system.

Competing-memory check: identify older meanings, L1 equivalents or near-synonyms that may be interfering. Integration sometimes requires reorganising prior knowledge, not simply adding a new trace.

Consolidation check: compare immediate performance with performance after sleep or a longer interval. Stable later access suggests the new representation has become less dependent on recent activation.

Mastery criterion: treat the word as semantically integrated when it can be interpreted, retrieved and adapted without reconstructing the original learning episode.

57. Learner recalls at same desk only

Memory profile: environment cue dependence. Likely state: episodic context effect. Next move: vary study location.

Cue fading: remove image, teacher, topic and location cues gradually. The lexical representation should become retrievable from meaning and language context rather than from one event.

Usage integration: add grammar, collocation and register information. Semantic memory for a word includes more than an abstract definition; it includes how the lexical item participates in the language system.

Competing-memory check: identify older meanings, L1 equivalents or near-synonyms that may be interfering. Integration sometimes requires reorganising prior knowledge, not simply adding a new trace.

Consolidation check: compare immediate performance with performance after sleep or a longer interval. Stable later access suggests the new representation has become less dependent on recent activation.

Mastery criterion: treat the word as semantically integrated when it can be interpreted, retrieved and adapted without reconstructing the original learning episode.

58. Learner recalls after seeing topic heading

Memory profile: topic cue dependence. Likely state: partial semantic route. Next move: retrieve without heading.

Cue fading: remove image, teacher, topic and location cues gradually. The lexical representation should become retrievable from meaning and language context rather than from one event.

Usage integration: add grammar, collocation and register information. Semantic memory for a word includes more than an abstract definition; it includes how the lexical item participates in the language system.

Competing-memory check: identify older meanings, L1 equivalents or near-synonyms that may be interfering. Integration sometimes requires reorganising prior knowledge, not simply adding a new trace.

Consolidation check: compare immediate performance with performance after sleep or a longer interval. Stable later access suggests the new representation has become less dependent on recent activation.

Mastery criterion: treat the word as semantically integrated when it can be interpreted, retrieved and adapted without reconstructing the original learning episode.

59. Learner recognises word instantly in any context

Memory profile: semanticised lexical familiarity. Likely state: strong integration. Next move: focus on productive precision.

Cue fading: remove image, teacher, topic and location cues gradually. The lexical representation should become retrievable from meaning and language context rather than from one event.

Usage integration: add grammar, collocation and register information. Semantic memory for a word includes more than an abstract definition; it includes how the lexical item participates in the language system.

Competing-memory check: identify older meanings, L1 equivalents or near-synonyms that may be interfering. Integration sometimes requires reorganising prior knowledge, not simply adding a new trace.

Consolidation check: compare immediate performance with performance after sleep or a longer interval. Stable later access suggests the new representation has become less dependent on recent activation.

Mastery criterion: treat the word as semantically integrated when it can be interpreted, retrieved and adapted without reconstructing the original learning episode.

60. Learner can define, collocate and adapt register

Memory profile: mature lexical knowledge. Likely state: semantic + usage integration. Next move: maintain through authentic use.

Cue fading: remove image, teacher, topic and location cues gradually. The lexical representation should become retrievable from meaning and language context rather than from one event.

Usage integration: add grammar, collocation and register information. Semantic memory for a word includes more than an abstract definition; it includes how the lexical item participates in the language system.

Competing-memory check: identify older meanings, L1 equivalents or near-synonyms that may be interfering. Integration sometimes requires reorganising prior knowledge, not simply adding a new trace.

Consolidation check: compare immediate performance with performance after sleep or a longer interval. Stable later access suggests the new representation has become less dependent on recent activation.

Mastery criterion: treat the word as semantically integrated when it can be interpreted, retrieved and adapted without reconstructing the original learning episode.