The Pipelining Strategy

Binding a character to a spoken word already owned is cheaper than taking form, sound, and meaning in as three unknowns. The three-session stagger manufactures that precondition: the spoken word both ways first, then the character bound to a pair already overlearned.

Why the default card fails, and the ladder

Native literacy already runs on the asymmetry. A child learning to write commands the spoken form, so brute repetition can carry them to literacy. Analysis of internal structure is a foreign-learner tool, not a native prerequisite.

The default foreign-learner flashcard throws that advantage away. Front “self; from”, back 自. Form, meaning, and sound all hit memory in one session, while grammar, tones, and vocabulary compete. Three Attributes of a Character is the bundle this schedule staggers. Rules of Effective Memorization is the domain-general version of one new link at a time.

Session 1 is the spoken word, one direction. A pronunciation–meaning pair sits on a spoken-word card. No written form anywhere on it.

Session 2 is the same pair reversed. The forward link already exists, so the backward association costs less than learning the pair fresh. That is a direction, not a measured ratio.

Session 3 attaches the form. Front “dào — way, path”, back 道. The shape locks onto a pair already held, instead of arriving as a third stranger.

Which direction goes first can vary. Meaning→pronunciation or pronunciation→meaning can lead. The invariant is the order: one direction, its reverse, then the form.

One session a day is the usual rhythm. The ladder is the construct, not the calendar.

Overlap, not a slower batch

Each session a new cohort enters. A cohort is a small set of characters that climb the three-session ladder together. House tripwire: 3–5 characters, not a finding. Each cohort spends three sessions inside. In steady state one cohort still graduates every session.

Overlap is what keeps throughput constant. Session 2 is B new plus A reversed. Session 3 is C new plus B reversed plus A forms.

A worked rotation:

• A = 道 那 但
• B = 前 很 性
• C = 情 物 等

A common misreading turns this into five characters per three sessions. That slowdown is sequential batching: one cohort finishes its whole ladder before the next one starts. Staggered entry is what avoids the penalty.

One card per pronunciation–meaning pairing. 道 dào and 道 dǎo get separate cards; 那 nà and 那 nèi too. A three-character cohort can yield five session-1 cards. Merged cards are a legitimate personal variant.

What the stagger buys

By the form session two of the three chunks are already owned. Across thousands of characters that per-item saving compounds.

The price is three sessions of attention per cohort before the form is even on a card. The benefit in the same breath: the form binds to an overlearned pair.

This schedule governs intake only. After the form session it has nothing further to say about the item. The group leaves the rotation and survives only as ordinary review cards. Spaced Interleaved Retrieval is the other system.

Session budget and the deferred variant

House tripwire, not a finding: 3–5 new characters inside a 20-item session. The rest of the session is common words built from the new characters or from known ones. 3 + 17 is a typical split, also a house default.

One or two high-frequency host words per new character give the written shape a real sentence to live in. Isolated-character study is the thing those host words refuse.

The wait before the form session can stretch. Three consecutive sessions is the compact default. Parking a character’s writing for a month or two after the spoken pair is overlearned still keeps the cheap-attachment mechanism. That span is a house tripwire. The practical ceiling is open.

After intake, and the breaks

Spaced Interleaved Retrieval governs everything after. SIR is retrieval spread over time and mixed across topics and forms. Spacing, interleaving, and ongoing reconstruction belong there. This stagger never touches review scheduling.

A study session is the active cohorts’ intake cards plus every review SIR has scheduled. Intake cards never substitute for due reviews.

Bundled cards — form, sound, and meaning while all three are unfamiliar — triple the new material per repetition. That “triple” is this page’s cost-structure claim, labelled, not a measured finding. It is the default the stagger replaces.

Imitating native rote copying fails without the spoken-word precondition.

Sequential batching drops intake to one graduating cohort every three sessions and buys nothing.

Reading sample glosses as a meanings curriculum is a separate decision, owned by Meaning Trees and Original Meanings.

The IME Method is production recall after intake. It is not this schedule.

Early skepticism is expected. A few weeks of running the stagger is the honest window before a verdict.

The schedule does not choose senses. It does not retrieve missing forms. It does not space reviews. 3–5 and 20 are unmeasured house defaults. If form-session cards are still three unfamiliar chunks, the spoken pair was never overlearned — new cohorts wait, and the existing ones reverse longer. What “working” looks like: after a few weeks, one cohort still exits each session, and form-session cards feel like attaching a shape to a word already held, not like a third new item.

The same asymmetry is now a stagger that holds in steady state. Spoken first, both ways, then the form. One cohort still finishes each session. Review is the other system.

Links into the knowledge base

• How Chinese Characters Work — the hub: the cluster this schedule finishes.
• Three Attributes of a Character — the bundle this schedule staggers.
• Spaced Interleaved Retrieval — review after a cohort exits. This page stops at intake.
• Meaning Trees and Original Meanings — which senses deserve cards.
• The IME Method — production recall after intake.
• Rules of Effective Memorization — domain-general rules the schedule instantiates.

Open Questions

One card per pairing versus merged cards for a character with two readings — which holds in practice?

How far the form session can be deferred after the spoken pair is overlearned, past the house “month or two.”

How a 20-item intake budget sits next to a large SIR backlog on the same day.

Sources

Share, David L. “Phonological recoding and self-teaching: sine qua non of reading acquisition.” Cognition 55 (1995): 151–218. Phonological-primary, orthographic-secondary — a new written form attaches more cheaply to a word already held by sound.

L2 character work showing oral associations developing before form associations, e.g. the open-access study at PMC10313221. Spoken-first is not a house preference; it is the cheaper order.

Paired-associate / incremental vocabulary (ERIC ED550749 and kin): a new link onto an existing pair costs less than three new nodes at once. Backward recall after a forward link is the same direction this ladder uses in session 2.