HCD.19:4.2 - Preserve what happened before and after help

Record the first answer with the reasoning, source use, or work trace necessary for the intended inference. Retain errors, omissions, justified uncertainty, and valid alternatives rather than silently replacing the answer with a corrected rendering.

Separate three support facts:

• what the arrangement promised;
• what was actually available to this reader;
• what the reader actually used, including a later teacher, peer, or AI cue.

When source contribution matters, ask the reader to identify what the material supplied and what they brought from prior knowledge. This report can help interpretation but is not by itself causal proof. Check it against the material and the response. The reader may not notice a familiar inference they supplied.

For a selected C.2.8 structural-recovery question, keep the correctly recovered relations and their attribution at these conditions. Familiar knowledge may enable decoding; correcting false or absent instruction from memory does not give the material credit for that correction. A response can correctly report what the material asserts without establishing that assertion’s subject truth. Compare claims and forms separately when a revised material changes content. One response is a bounded recovery observation, not a maximum extractable amount or a learning result.

After preserving the first response, provide the declared feedback. Record its content and timing relative to the response. Keep a teacher’s answer-bearing explanation distinct from a prompt that merely asks the reader to inspect an available criterion. Both may be appropriate support, but they permit different conclusions about independent work.

No universal number of readers or tasks establishes adequacy. Select evidence for the property and consequence at issue. A cold intended reader lacks the author’s private context; that condition does not erase the reader’s real expertise or turn an agent into a novice human.