A learner can remember that allocate appeared in a sentence about a project budget and still miss it in “allocate two hours for revision.” The spelling is familiar. The second setting is not. For an adult learner around B1 or B2, a vocabulary item becomes useful when it survives a change of subject, grammar and neighbouring words, then remains available when the learner has to read, write or speak without the original example nearby.
That makes varied context a lesson-design decision. After the first clear encounter, the same word should return in several understandable sentences that make the learner retrieve its meaning again. Four copies of one example train four copies of one route.
What context changes
A word in memory carries more than a translation. It develops links to meanings, sounds, grammatical patterns and the words that often sit beside it. A narrow example can establish one strong link. It may also leave the learner dependent on that example.
Ferreira and Ellis taught English words to second-language learners and tested semantic decisions as well as reading aloud. Their study found effects of contextual diversity on how newly learned words were processed, though the effects differed across tasks. The 2016 paper is useful because it does not reduce word knowledge to one translation test.
Frances, Martin and Duñabeitia compared incidental learning in a native and a foreign language. Their Scientific Reports experiment showed that varied contexts can support learning, with results shaped by whether participants were working in their first or foreign language. Context diversity has a cost: each new sentence asks the learner to interpret the word again.
That cost is the point, within limits.
Start with a stable meaning
Imagine a 32-year-old project coordinator in Taichung who knows allocate only as 分配. A first example should make the intended meaning easy to verify: “The team allocated NT$18,700 to translation.” The number and object constrain the interpretation. A short Chinese gloss can settle any remaining doubt.
Later encounters can widen the frame:
- “Allocate two hours for the final review.”
- “The school allocated three rooms to the evening course.”
- “We need to allocate staff according to demand.”
- “How much storage has the app allocated to offline files?”
These sentences vary the object and subject while keeping the core sense recoverable, and they reveal patterns that a bilingual entry alone may hide: allocate something to something, allocate something for a purpose, and the noun allocation.
Learners should make a choice each time. A gap, a paraphrase match or a question such as “What was assigned, and where?” forces a fresh retrieval. Copying all four sentences into a notebook can look industrious while leaving the meaning untouched.
Diversity can arrive too early
Context does not rescue a word that the learner cannot infer. Peter Prince compared context learning with translations and found that proficiency affected which route helped. His Modern Language Journal study cautions against treating context as automatically superior for every learner.
Recent reviews find the same boundary. Studies use “contextual diversity” to mean different things, from distinct sentences to separate documents and semantic neighbourhoods. A 2026 systematic review and meta-analysis found a small L2 evidence base compared with first-language research and called attention to proficiency as a possible moderator. It would be reckless to turn that literature into a promise that any four random sentences will produce durable vocabulary.
For an A2 learner, four dense news sentences may create four comprehension failures. One plain definition, one translation and two controlled examples can do more. Proper nouns, unknown idioms and tangled clauses add difficulty that has nothing to do with the target word.
Keep the meaning constant before widening it
Polysemous words need stricter sequencing. Consider issue. A newspaper can cover an issue, a bank can issue a card, and a magazine has an October issue. Presenting all three on day one asks the learner to build three meanings while also learning the form.
Teach one sense until the learner can retrieve it from two or three sentences. Add another sense later and mark the difference directly. This fits the argument in our article on giving a familiar word a second meaning lesson: recognition of the form can conceal a missing meaning.
Closely related words create another problem. Economic and economical share letters and occupy similar grammatical slots, so varied examples may blur them if both remain unstable. Separate them first, then compare them. Our piece on separating similar English words before comparison explains why the order matters.
A compact classroom sequence
One useful sequence takes about 17 minutes for six target words. The teacher gives a definition and one transparent sentence, then checks the meaning with a forced choice. Learners meet each word in a second sentence from the same topic. After a short intervening task, they see a third sentence from a different topic and explain what the word contributes there.
Production comes last. Each learner chooses a setting from work, study or daily life and writes one sentence. A partner checks two things: whether the core meaning survived and whether the surrounding grammar sounds possible. The teacher collects only the sentences that expose uncertainty. Those become tomorrow's retrieval items.
This sequence has an admitted gap. A learner who attends once a week may forget the word before the wider context arrives. The practical answer is a brief review outside class, though the research does not specify one universal interval.
Vocabulary software can follow the same logic. Avoid showing the identical card front and example at every review. Keep the prompt stable for the first successful retrieval, then rotate among three vetted sentences, each different enough to require interpretation and plain enough that an unknown idiom does not become the real test. Random corpus lines are a poor substitute because fragments often omit the very information a B1 reader needs. Prepared lines come first. Independent corpus work comes later.
What to check after a week
Translation alone will miss the benefit you are trying to build. Use one familiar sentence, one new sentence and one short production prompt. If the learner understands allocate in the original budget example yet fails with time, rooms and staff, the representation remains tied to one route.
Six words are enough for this check. Put one of them in an unfamiliar but readable sentence on Friday, such as “The clinic allocated the first 23 appointments to returning patients.” Ask for a paraphrase, then ask the learner to use the word in a Monday-morning scheduling message. The change of context is the test.