Competing-Case Spacing Rule¶
Scheduling rule — instantiates Post-Encoding Trace Stabilization
Schedules confusably-similar new items far enough apart that each stabilizes into its own distinct context before the next one can blur it.
When two new things are alike, learning the second one right after the first is the fastest way to lose both. Competing-Case Spacing Rule is a scheduling discipline that identifies which fresh traces are confusably similar and deliberately puts time and unlike material between them, so that each one sets into its own distinct context before its near-twin arrives to overwrite it. Its defining move is similarity-targeted separation: it does not quiet the environment in general, it reroutes the specific items that would blur into each other. The dial it turns is ordering and gap, not volume of input.
Example¶
A learner using a spaced-repetition deck to study a foreign language keeps confusing two words that look and sound alike — a classic pair of "false friends" that mean unrelated things. She had front-loaded them on the same evening, one card after the other, and ever since, each cue retrieves a smear of both. The problem is not that the words are hard; it is that they were encoded shoulder-to-shoulder, before either had bound to anything that told them apart.
She adopts a competing-case spacing rule. Her deck tags near-neighbors — same word family, similar spelling, overlapping meaning — and refuses to introduce a second member of a confusable cluster until the first has been reviewed successfully across several days and anchored to its own example sentence, image, and mnemonic. The two false friends are now separated by a week and by a wall of unrelated vocabulary, and each is introduced inside its own little scene rather than as an abstract pairing. The smear clears: each cue now lands on one word, because each word had room to become distinctly itself before its rival showed up.
How it works¶
- Flag the confusable set. A similarity monitor scans incoming items for the ones most likely to interfere — shared features, overlapping cues, adjacent meaning — and marks them as a competing cluster rather than treating all new items alike.
- Impose a minimum separation. No two members of a flagged cluster may be encoded within the same gap; the second waits until the first clears a stability check and until unrelated material has intervened.
- Bind each to a distinct anchor. As each item is introduced, it is tied to its own example, context, or discriminating feature, so the thing that separates it from its twin is encoded with it.
- Interleave, don't block. Once separated and anchored, the cluster's members can be interleaved for discrimination practice — but only after each has its own stable footing.
Tuning parameters¶
- Similarity threshold — how alike two items must be to count as competing. Set it loose and you space things that never would have interfered, wasting calendar; set it tight and true confusion pairs slip through together.
- Minimum gap — how much time and intervening material separates cluster members. Wider gaps protect better but slow how fast a similar batch can be learned.
- Anchor richness — how much distinct context each item is bound to on introduction. Richer anchors discriminate better but cost more effort per item.
- Release-to-interleave point — when separated items are allowed back together for contrast practice. Too early re-blurs them; too late forgoes the discrimination benefit.
When it helps, and when it misleads¶
Its strength is that it attacks interference at its source — the specific similar traces that overwrite one another — rather than paying the blanket cost of quieting everything. This is the everyday face of catastrophic interference, the tendency, first sharply characterized in connectionist networks, for new learning that overlaps old learning to wipe it out when the two are packed together.[n1] Spacing plus distinct anchoring is the cheap, targeted antidote.
Its failure mode is mis-clustering: if the similarity monitor flags the wrong pairs, the rule spaces apart items that would never have interfered while letting the real confusion pairs through side by side, and the learner pays the scheduling cost for no protection. The classic misuse is spacing on surface similarity (spelling, appearance) when the true interference is semantic, or vice versa. The guarding discipline is to let observed confusions feed back into the similarity model — when two items actually blur in practice, promote them to a competing cluster even if they looked unrelated on paper.
How it implements the components¶
competing_trace_monitor— the similarity scan that detects which fresh traces are confusably alike is exactly this monitor; it is what the whole rule keys on.integration_context_map— binding each item to its own distinct anchor on introduction is the context map that lets the trace join memory without losing its separate identity.
It does not lower the overall competing load or protect a single window by reducing all input — that broad interference_load_boundary over the fragility_window is Interference-Shielding Protocol's job; the spacing rule reorders confusable items in time, whereas the shield dampens everything around one trace regardless of similarity.
Related¶
- Instantiates: Post-Encoding Trace Stabilization — supplies the similarity-separation step that keeps competing traces from blurring during stabilization.
- Sibling mechanisms: Interference-Shielding Protocol · Spaced Integration Review · Delayed Retention Probe · Post-Learning Rest Window
Editorial Notes¶
Form Classification¶
Form family: Rule, Policy & Commitment
Rationale: Schedules confusably-similar new items far enough apart that each stabilizes into its own distinct context before the next one can blur it, making its operative form a standing rule, threshold, contractual commitment, or policy constraint governing future conduct.
Independent corroboration: The frozen evidence defines Competing-Case Spacing Rule as 'Schedules confusably-similar new items far enough apart that each stabilizes into its own distinct context before the next one can blur it', so its operative form is Rule, Policy & Commitment.
Review outcome: Independent reviewer agreement; high confidence.
Origin Attribution¶
Primary origin: Psychology
Origin pattern: Convergent development
Present-day reach: Multi-domain
Rationale: Experimental memory psychology established proactive and retroactive interference and showed that conceptually similar material competes strongly, motivating temporal separation and distinctive encoding contexts for confusable items.
Related originating lineages:
- Cognitive Science — Connectionist research on catastrophic interference supplies the computational account of overlapping new associations overwriting earlier representations.
- Education & Pedagogy — Instructional sequencing and interleaving practice translate interference evidence into teachable schedules and distinct contextual anchors.
Review resolution: Experimental psychology directly documents stronger proactive and retroactive interference among cognitively similar material. McCloskey and Cohen's connectionist work established the parallel catastrophic-interference problem in sequential neural-network learning. The mechanism synthesizes these with instructional scheduling, so psychology is primary and cognitive science and pedagogy are retained as formative convergent lineages.
Attribution caveat: Connectionist cognitive science independently exposed catastrophic sequential interference, but human-memory psychology is primary because the mechanism schedules human cases and vocabulary around observed proactive and retroactive interference.
Encyclopedia synthesis: The exact catalogued form synthesizes established practice rather than reproducing a single standard historical label.
Review outcome: Researched adjudication after independent review; high confidence.
Sources consulted:
- Bunt and Sanders: Effects of Cognitive Similarity on Proactive and Retroactive Interference
- McCloskey and Cohen: Catastrophic Interference in Connectionist Networks
Notes¶
[n1] Catastrophic interference (also catastrophic forgetting) — described by McCloskey and Cohen in connectionist networks, where learning new, overlapping associations rapidly erases previously learned ones. It is the machine-learning mirror of the retroactive interference long studied in human memory, and the reason separation-in-time is a load-bearing design choice for similar material. ↩