The example checks two independent parts of a turn: the speaker's exact wording and the listener's exact guess. Its original card targets envelope and forbids envelope, letter, mail, and paper in the clue. A completed bad-clue sample shows why a correct guess cannot erase a restriction violation. Editing either field clears the previous feedback before another check.
Announce the local word rule
For this activity, match the four declared words as whole words without regard to case. “Paper” is therefore flagged just like “paper.” Do not silently add synonyms or commercial rules. The phrase “written message” is allowed by this specific bank, even though its meaning relates to letters. If a learner proposes an inflected form or an ambiguous boundary, record the issue for teacher review instead of claiming the simple lexical checker resolves every language question.
The restriction applies to the speaker's clue. The listener is allowed to say envelope as the guess. Put target and forbidden bank on a separate speaker card for a live classroom turn. The example is an open practice review, so it displays the target and key; it cannot function as a secret listener packet. A collapsed answer panel would not change that distinction. Use physically separate cards when the listener must not see the target before guessing.
Work three useful cases
The model clue is “A flat container you can close to hold a written message.” With guess envelope, the word check reports no declared forbidden words and the guess matches. Replace the clue with “a paper pocket”: paper appears at character 3 in the one-based text position, while the envelope guess still matches. Now retain the model clue but change the guess to folder. The wording remains within the bank's restriction, yet the guess differs from the target. Each case suggests a different discussion.
For a folder guess, ask which feature would distinguish the intended object more clearly. An authored revision is “A flat container with a flap you close around a written message.” It adds a flap feature and avoids all four declared tokens. Keep the old clue alongside the revision on the teacher's turn record. The proposed change is a reasoned clarification, not a demonstrated improvement in a real listener's performance. A token checker cannot determine whether the full explanation teaches the meaning well.
Make revision the outcome
Rotate speaker, listener, and observer. The observer records exact clue, flagged word and position, exact guess, target comparison, revised clue, and reason for revision. A written untimed turn or a pass can replace spoken play. The teacher confirms the participation supports and the target meaning before use. Avoid using speed, penalties, or embarrassment to substitute for the actual explanation task; the useful evidence is what was said and how it could be made more specific.
An empty clue stays incomplete, and an empty guess stays not recorded. Do not award a compliant-clue result to a field with no text. The page reports these states explicitly, and Reset clears the active practice turn. It does not maintain a class ledger. End the round by asking the writer to explain what the revision adds and the listener to explain what they understood. Keep the lexical finding separate from that human review, especially when a correct guess arrived despite an overly broad or restricted clue.
Make your own working example
Build an original constrained-clue review page for card T1. Target envelope; forbidden clue words envelope, letter, mail, paper. Use labeled editable clue and listener-guess fields, Check, Load model, Load bad clue, and Reset buttons. Start with the authored model A flat container you can close to hold a written message. and guess envelope, showing both completed checks.
Check exact whole words case-insensitively and report matched tokens with one-based character positions. The bad clue a paper pocket must flag paper at character 3 while the envelope guess still matches. The model clue with guess folder has no listed tokens but a differing guess. Treat blank clue as incomplete and blank guess as not recorded. Do not auto-fill missing answers.
Keep lexical compliance and target comparison in separate visible outputs. Input edits clear old feedback. Reset clears both fields. Offer the proposed revised clue A flat container with a flap you close around a written message. as a teacher illustration, without claiming actual improved guessing. State that the open practice page reveals its target and is not a secret listener card or a semantic vocabulary grader.
Use one self-contained HTML file with no remote dependency. Keep the initial sample, output, and labeled controls near the top. Use at least 16px text, strong contrast, visible focus, and a stacked layout at 360px without horizontal overflow. Use explicit type=button click handlers and explicit DOM selectors; do not depend on native form submission. Clear stale feedback when task inputs change. Render entered text safely. Keep practice data local to the current page; do not imply saving, sending, or production integration. Return only the finished HTML.
Prepare the next class copy
After reviewing the turn, update the static instruction that defines whole-word matching so the next group uses the same declared restriction. In NOS, eligible static instructions can be revised in visual editing mode without asking AI to regenerate the HTML. Keep the teacher's source packet and actual learner work separately; use the reading copy to clarify the next activity.