Cue-Fading Schedule¶

Protocol — instantiates Encoding–Retrieval Context Alignment

Starts recall fully supported, then withdraws the props on a planned, evidence-gated ramp until the learner retrieves unaided in the conditions that count.

The Cue-Fading Schedule begins with a support cue strong enough to guarantee success — a prompt, hint, or template — and then removes it in graded steps, advancing to the next step only when unaided recall holds, until the target condition is met with no prop at all. Its defining move is treating the support cue as temporary scaffolding to be deliberately dismantled, and gating each removal on evidence — the exact opposite of leaving a helpful cue in place forever because it works. The finish line is fixed by the target retrieval context: recall must survive there, unaided, or the fade isn't done.

Example¶

A child is learning to brush his teeth as part of a morning routine. He starts fully supported — hand-over-hand guidance plus a spoken step-by-step — so he never practises the sequence wrong. The schedule then fades the support one level at a time: physical guidance gives way to a gesture, the gesture to a single verbal cue, the verbal cue to a small picture card taped to the mirror, and finally to nothing. Each level is removed only after the prior one is reliable, and a miss backs the schedule up a step rather than letting an error re-encode. The defined endpoint is not "does it in the clinic" but "brushes unprompted in his own bathroom, in the morning" — the context that actually counts.^[1]

How it works¶

You first define the independence endpoint in the target context, then build a prompt hierarchy from most to least support, set an advance criterion (how much unaided success unlocks the next fade) and a back-up rule (how fast to restore support on a miss), and fade on evidence rather than the calendar. What distinguishes the schedule from ordinary practice is this deliberate, monotonic withdrawal of a specific support cue, ending only when unaided recall meets criterion under conditions that model the real point of use.

Tuning parameters¶

Fade step size — large steps risk errors that re-encode failure; small steps are safe but slow.
Advance criterion — how many unaided successes before the next prop is pulled; strict criteria trade speed for durability.
Back-up rule — how quickly support is restored after a miss — the difference between an errorless ramp and an errorful one.
Prompt hierarchy — which support types, in what order (physical → gestural → verbal → visual), matched to the learner.
Endpoint fidelity — how closely the final, unaided step matches the real retrieval context it must survive.

When it helps, and when it misleads¶

The schedule is the right tool when durable independence is the goal and a helpful cue would otherwise become a permanent crutch that masks a gap. It misleads when faded too fast — errors slip in and re-encode the wrong response — or when it fades on a calendar instead of on evidence, declaring independence to hit a deadline. The classic misuse is pulling the last prop just before an assessment to demonstrate an independence that isn't real. The guard is to gate every step on evidence, back up on failure, and verify at the true endpoint — usually by handing that verification to a probe or test rather than trusting the schedule's own optimism. Note the boundary: fading assumes independence is wanted; when the stakes demand a permanent external route, the fade is the wrong bet (see Notes).

How it implements the components¶

The schedule realizes the independence-building side of the archetype — deliberately removing support toward unaided recall in the real context:

cue_fading_and_independence_plan — the graded withdrawal ramp, with its advance criteria and back-up rules, is this plan.
target_retrieval_context_model — the fade endpoint is defined and tested against a model of the target retrieval context; the schedule carries that model as its finish line, so "independent" means "independent there."

It does not generate the outcome signal that grades each step (retrieval_outcome_signal, produced by Interleaved Competitor Retrieval Test and Free-Recall-Then-Recognition Probe) or space retrievals over time for durability (Spaced Retrieval Scheduler); it withdraws support.

Instantiates: Encoding–Retrieval Context Alignment — the schedule is how the "grow independence from cues" strategy is actually executed.
Consumes: an outcome signal from a probe or test to gate each fade step, and a diagnostic cue set (from Cue-Diagnosticity Ablation Test) to know which cue to fade.
Sibling mechanisms: Environmental Retrieval Cue · Free-Recall-Then-Recognition Probe · Spaced Retrieval Scheduler · Context-Switch Recall Drill · Cue-Diagnosticity Ablation Test · Context Translation Card · Context Reinstatement Protocol · External Checklist or Job Aid · Interleaved Competitor Retrieval Test · Mnemonic Cue Pairing · Post-Event Re-Encoding Debrief · Representative-Environment Simulation · Scenario-Based Retrieval Test · Transfer-Appropriate Processing Rehearsal · Varied-Context Retrieval Practice

Notes¶

The Cue-Fading Schedule and the Environmental Retrieval Cue pull in opposite directions: one removes support to build independence, the other installs a permanent cue because memory should not be the sole control. They encode opposite bets about whether independence is even desirable. When a lapse is catastrophic and a permanent external route is the safer design, fading is the wrong mechanism — reach for External Checklist or Job Aid instead.

References¶

[1] Errorless learning and prompt fading — introduce material with enough support that early errors are prevented, then withdraw the prompts gradually. Preventing the error matters because a wrong response practised under a cue can itself be encoded and later retrieved. ↩