Sentence Lab

Three random words. Students build one sentence that uses all of them — then defend it to the class.

Words are learned in syntactic environments

Lewis (1993) argues that vocabulary acquisition is inseparable from grammar: words are stored in mental lexicons alongside their syntactic environments, collocations, and typical sentence positions. Sentence Lab requires students to deploy words in a grammatical frame, encoding both lexical meaning and syntactic behaviour simultaneously.

Sentence construction forces syntactic decision-making

Swain (1985) predicts that output tasks — especially those requiring the learner to make syntactic choices — promote grammatical processing. When students must decide whether 'suspicious' modifies the subject or the object, they are doing the kind of syntactic reasoning that listening and reading tasks alone cannot produce.

Generation effect enhances retention

Slamecka & Graf (1978) demonstrated that words generated by the learner are remembered better than words provided by the experimenter. Sentence Lab's generation task produces exactly this effect: students who build their own sentence with a new word retain it more durably than students who read an example sentence.

Generates a noun, a common verb, and a basic adjective from everyday vocabulary. Students build a simple SVO sentence. Words are concrete and familiar — market, describe, colourful.

A2 learners can understand these words receptively but struggle to produce them in combination under time pressure. The three-word constraint prevents overlong attempts and forces decisions about article use, verb form, and adjective placement.

Generates a noun, an adverb, and a verb that invite complex sentence structures. Words like 'despite', 'gradually', 'achievement' push students toward subordinate clauses.

B1 students default to simple sentences when constructing under time pressure. Adverbs like 'gradually' or 'reluctantly' almost require a complex structure to use naturally — the word itself pushes students toward the complexity the level demands.

Generates semantically mismatched words that produce funny sentences when combined correctly. Grammar remains strictly level-appropriate even when the content is absurd.

The generation effect predicts that distinctive, memorable sentences are retained better. Goofball mode produces the most distinctive sentences possible — students remember 'The penguin reluctantly investigated the suspicious sandwich' for weeks.

Class Sentence Vote

After sharing all sentences, the class votes for their favourite — most grammatically sophisticated, most creative, or most amusing depending on the round's goal. The winning student reads it once more and explains their word-order decisions.

Add-a-Word Relay

Start with the three generated words in a sentence. Each student adds exactly one word to the sentence, passing it around the circle. The sentence grows into a complex, multi-clause structure. Read the final sentence aloud and analyse how it evolved.

Translation Workshop

After constructing their English sentences, students translate them into their L1. Then they swap translations with a partner, who translates back to English. Compare the back-translation to the original — mismatches reveal where words have different syntactic behaviour across languages.

Grammar Upgrade

Take a valid but simple sentence and challenge students to upgrade it using one grammar structure from their current unit — relative clause, conditional, passive voice. Compare the original and upgraded versions side by side.