The clue explorer starts with three fictional tokens, narrows to E1 and E3 after Square, then reaches a unique match or an explicit contradiction depending on the selected clue version.
Use the worked example
Reveal Square in V1, then Blue: E1 is unique. Switch to V2 and reveal its Green replacement: E3 is unique. V3 adds Blue and Green together and leaves no compatible candidate. Version changes start again from all three tokens rather than retaining the earlier reduced set.
Explain the evidence for the answer
The activity follows the clue-and-reassessment approach used in ReadWriteThink's inference lesson. Here the possible identities are a closed three-token set: E1 is Blue Square, E2 Blue Circle, and E3 Green Square. Every exclusion must follow from a revealed clue about one of those tokens.
Before clues, all three candidates remain eligible. The identity is unknown because nothing distinguishes them. Unknown information is not an empty candidate set.
E1 and E3 both fit; E2 conflicts with C1 because it is a circle. Record both remaining IDs and the exclusion reason. The state is Insufficient clues, two candidates, not E1 selected as the answer.
V1 then adds C2: The token is blue. E1 fits both clues; E3 conflicts with C2. E2 is still excluded by C1 even though it matches Blue. The unique match is E1 within this declared bank.
Eligibility requires every active clue, not just the newest one. Showing E2 again after Blue would discard the Square condition rather than update the inference.
Independent V2 replaces C2 Blue with C2 Green. It does not add Green to V1. With C1 Square and the replacement C2 Green, E3 is the unique match.
All candidates should be recomputed from the original bank and the new active clues, not from V1's already reduced E1 set. Otherwise E3 would never return even though it is compatible with V2.
Independent V3 supplies C1 Square, C2 Blue and C3 Green together. No candidate fits all three because the declared color labels are exclusive.
The correct state is No compatible candidate, inconsistent supplied clues for this bank. It is not the same as the earlier two-candidate state, where more information could distinguish existing possibilities.
Do not select E1 because it fits two clues, invent a blue-green square outside the bank, or silently drop C3. Record which clues conflict and ask for source review. The activity cannot resolve the inconsistency by guessing.
Below the current clue summary, the table lists each candidate and the active clues it conflicts with. Enter every eligible ID in the answer field, separated by commas, then Check full set. ID order does not matter, but leaving out a compatible token changes the answer.
Use a separate spoken or written question for the reasoning: “Which clue excludes E2, and why is E3 still possible after Square?” The page checks the entered compatible set. It does not read that explanation, so a teacher can distinguish a lucky selection from an explained inference.
Type NONE to answer that no candidate fits. A blank answer remains unanswered. In V3, the explicit NONE response means that the complete clue set conflicts with this bank; it does not mean that more than one token remains possible. Reset hides clues but keeps the selected version.
Copy a prompt for this example
Make one HTML clue activity using exactly three fictional tokens: E1 Blue Square, E2 Blue Circle, and E3 Green Square. Display each token's ID, color, and shape in text. Before clues, all three remain possible.
Provide three version buttons and Reveal next clue. V1 reveals Square, then Blue: the first clue leaves E1 and E3, and both clues leave E1. V2 reveals Square, then Green, leaving E3. V3 reveals Square, Blue, Green; because these color labels are exclusive, no candidate fits the complete set.
Every reveal must recompute compatibility from the full original bank and every active clue. Show the active version, revealed clues, complete eligible-ID set, and a table explaining which active clues exclude each candidate. Distinguish insufficient clues, one compatible candidate, and an inconsistent clue set.
Add an answer field for comma-separated eligible IDs and an explicit NONE answer. Check accepts the complete correct set regardless of ID order, but blank remains unanswered. The comparison checks the set, not a learner's written reasoning.
Changing versions hides clues and clears the answer and feedback. Reset keeps the chosen version but likewise restores the unrevealed three-candidate state. Revealing another clue clears stale answer feedback.
Use a clear classroom card layout, keyboard-accessible controls, and readable stacked sections on a phone. Return one self-contained HTML page.
Choose the clue version that makes your discussion question visible, then ask students to cite the clue excluding each token. You can preview the prepared HTML in NOS beside the teacher matrix. Treat an empty unsubmitted answer differently from an explicit No compatible candidate response.