Story Cards as Memory Scenes

The Tsumugu Encoding Dictionary carries two cards for every character. The Form card names the functional components — form, meaning, sound, empty — shows how they assemble into the glyph, and traces how the reading drifted. The Story card is a mnemonic scene built from those same components: a Chinese line with the parts marked, plus an English scene. The Form cards worked. The Story cards mostly did not, and the failure rate put the whole Story half in question.

I reviewed the first five hundred characters, the TOCFL A1 foundation. Reading a random sample confirmed three patterns.

The dead ones restate the Form. 領 (lǐng): the Form card ends “take the coat by the collar and the whole garment follows.” The Story said “one lift, and the whole coat follows.” Same beat, second slot. Cut.

The inert ones park a sound component as a bare glyph. 衙 (yá): the scene ran 行 around the edges and dropped 吾 inside as its written shape — “inside sits 吾.” The piece carries the reading and earns no picture, so it adds nothing to recall. Fix or drop.

The ones that work run a chain that settles into the meaning. 科 (kē): new grain (禾) rattles into the measure (斗); the strike-board levels it, and this grade goes to its bin. Each component acts on the next, and the category falls out of the leveling.

The split traced to a missing rule about what a Story card is for. So I rebuilt the rubric from the method I had already studied — the Modified Method of Loci, the rote-memorization technique in the iCanStudy course — and kept Outlier’s “Rules of Effective Memorization” as the secondary source. Outlier already powers the Form card and feeds the public synthesis at Rules of Effective Memorization.

One line of loci doctrine reorganized everything. Rote memorization is a last resort, layered on top of concepts. The method of loci does not help information that already has functional or conceptual relationships, and it should not replace a higher-order network when one exists. The Form card is that network — functional decomposition plus the sound-series and meaning-series that make most of a character predictable (How Chinese Characters Work holds the cluster). The Form owns the conceptual layer. The Story sits on top as the loci layer, and its one slot is spent only on the arbitrary residue the Form cannot make inevitable: binding the components into one drawable gestalt, and pinning the part of the sound the sound-series leaves over.

That is why 領 was dead. The Form had already made the collar-and-garment relation logical. A Story that says it again burns the single mnemonic slot on content the Form owns, and the spaced repetition that follows polishes an empty slot.

The pivot: Story cards become single-character loci scenes. The modification I work from drops the palace itself — no fixed place to walk through — and keeps the three rules for the elements of the scene I draw. Make them absurd — a wooden train is forgettable; a train with a forest growing inside it stays. Keep them spatially distinct — move through the scene so the pieces do not blur into one lump. Make them interdependent — chain each to the next so none can be recalled alone, and the connections multiply the routes back to any one of them. One element per piece of information, one connected scene, and the anchor is the drawing: the encoding happens in the act of representing it.

The reason sits in the forgetting curve. Connection and elaboration build a trace with many handles. Exaggeration and dual coding give it the salience that survives a single read. Reconstruction beats recall, which beats recognition — a scene the learner rebuilds from the components generates the answer instead of recognizing a cue.

Outlier adds two moves loci lacks. Recover a component’s ancient picture when the modern glyph shows nothing — 及 as a hand seizing a person, 禾 as a grain plant. And anchor the character to a spoken word the learner already owns, so the word’s sound and meaning feed the reconstruction.

This reframes the failures. 衙’s 吾 sat as a written shape because no rule told the author that a sound piece must become a picturable, salient element. 科 worked because grain and measure act on each other and the meaning rides the final move. And the literary, realistic register goes — that was an AI default, never chosen. The course names mundane elements as actively forgettable, so restraint here is a defect.

One honest caveat. Loci is demonstrated on ordered lists of twenty to a hundred and fifty arbitrary items. A single character with three components that already relate to one another is a different case. Rendering one glyph as a small loci micro-scene is my adaptation of the method, and I am naming it as one. The wager is that the same levers that hold a long list together — absurdity, spatial distinctness, interdependence, drawing — also bind three pieces into one glyph, applied only to the residue the Form leaves.

So the Story card now finishes the Form. The Form makes the character make sense; the Story makes the arbitrary remainder stick. 麋 (mí) shows the shape of the work: 米 reads as the rice the deer crops, which pins the sound, and the scene has to bake in “large” — an outsized deer — because that attribute is the meaning. Cut what the Form already owns, picture every component, weld the meaning into the last move, and spend the one exaggerated beat on the piece that carries the sound.