moji. Get the app

Updated August 8, 2026

How to Remember Kanji: the Component Hook Method

The fastest way to remember a kanji is to stop treating it as a picture and start treating it as a small set of parts you already know. Name each component in a word, gloss it in a couple of words, then tie those glosses into one concrete image. In writing it looks like this: k1 (gloss) + k2 (gloss), one short sentence. The components are the hooks. The sentence is the thing your memory actually reaches for when the character shows up cold three weeks later.

That is the method behind the 5,300 hand-written mnemonics in Moji - Learn Japanese Daily for iPhone, and the rest of this page is the rulebook: which components are allowed to be hooks, what gets a mnemonic rewritten, and why kana-only words never get one at all.

What a component hook mnemonic looks like

Here is one from the catalog, on a word most beginners meet early:

猫 (ねこ, cat) 犭 (beast) + 苗 (seedlings), a small beast slinking through the seedling rows, hunting field mice.

Three things are doing work there. The breakdown names parts that genuinely appear in the character, so it also teaches you 犭 and 苗 for the next word that uses them. Each gloss is short enough to hold in your head. And the closing image is a scene, not a definition: it does not end on “so it means cat”, because the meaning is already printed on the card an inch above.

That last point is the rule that gets violated most. A mnemonic that opens by restating the headword and closes by restating the meaning has told you nothing you could not already see. The mnemonic’s whole job is the one thing the card cannot give you: a hook that brings the word back at recall time.

Not every kanji splits. Some are pictographs with nothing inside them, and those skip the breakdown entirely and go straight to the picture, one vivid line about what the shape looks like. Forcing a two-part split on a character that has no honest split is how false components get invented.

Why any component can be a hook, as long as the gloss is true

There is a common purist objection: many components in a kanji are phonetic, carrying sound rather than meaning, so building a meaning story on them is “wrong”. In practice that objection costs you more than it saves. A phonetic component is a perfectly good peg, and when its sound matches the word’s reading it is a bonus worth pointing out. 寺 (てら, temple) lends its じ reading to both 時 (とき, time) and 持つ (もつ, to hold), so a learner who notices 寺 gets the reading for free.

So the bar is not etymological purity. The bar is truth about the character:

The distinction is narrow on purpose. A story can be as invented as you like, because you are the only person who has to picture it. A gloss is a claim about Japanese, and the learner will believe it.

Why kana-only words never get a mnemonic

Moji’s catalog is 7,435 words, and 5,300 of them carry a mnemonic. The gap is not laziness. A mnemonic is a hook on written form, and a kana word has no form to hook: hiragana and katakana are phonetic scripts, so there are no components to name and no picture inside the shape.

What tends to fill that vacuum is a sound pun, and sound puns age badly. If the learner has to be talked into hearing the resemblance, the pun costs more attention than it returns, and it competes with the real thing they should be encoding, which is the word’s sound and its use. So kana-only vocabulary gets no mnemonic at all.

Some kana words do have a genuinely useful fact behind them: a loanword origin, a real etymology, a minimal pair worth not confusing. Those facts are real, so they go where facts go. Moji has 5,698 usage notes, written only when there is something true, non-obvious, and actionable to say, and deliberately left empty otherwise. A forced story for a kana word is worse than a true usage fact, and worse than silence. The same restraint drives the rest of the catalog, which you can read about in how we curate Japanese vocabulary.

Transparent loanwords and derived forms sit out too. If a word is taught through its base word, the base word’s mnemonic is already doing the job twice over.

How this compares to Remembering the Kanji and WaniKani

Both of those are serious systems, and both get real results. They differ from the component hook method mostly in what they ask of you and what unit they work on.

Heisig’s Remembering the Kanji assigns every kanji and primitive a unique keyword, then has you build a scene connecting the keyword to the primitives. Crucially, “the method requires the student to invent their own stories” (Wikipedia). Self-generated stories are genuinely stickier than borrowed ones, and that is the method’s edge. The cost is that the unit of study is the isolated character with an English keyword attached, and readings and vocabulary come later.

WaniKani teaches its own named radicals and supplies the mnemonics for you, describing radicals as “building blocks for learning kanji” used to “make mnemonics that allow you to memorize a kanji in seconds” (wanikani.com). It runs them through its own SRS on a fixed radical, then kanji, then vocabulary progression.

Moji’s mnemonics sit on the vocabulary word itself, next to the reading, the pitch accent, and a curated example sentence, so the hook and the word you will actually use arrive together rather than in separate phases. If you are weighing these up specifically on mnemonic quality, the mnemonic comparison page goes deeper, and there is a fuller WaniKani comparison too.

Where the mnemonic goes once the word sticks

A mnemonic is scaffolding. The first few times you meet 猫 you walk through the beast and the seedlings; by the tenth review the word arrives whole and the story has quietly fallen away. That is the intended outcome, not a failure of the mnemonic.

This is why the hook lives on a spaced repetition card rather than in a separate mnemonic course. Moji schedules reviews with FSRS-6 at a 90% target retention, so the card comes back at the point where the memory is starting to fade and the story still has something to do. Once stability is high, the interval stretches and the mnemonic stops being load-bearing. If you want the mechanics of that, see how Moji uses FSRS and memory bands.

One clarification, since the names are close: this is Moji - Learn Japanese Daily, an iPhone-only app (iOS 18.0 and up) by an independent developer, not the similarly named “Moji - Learn Japanese & Kanji”.

Try it on the first few hundred words

Flashcards with mnemonics and example sentences are on the free tier, so you can read a few hundred of these before deciding whether the method suits you. Start at the beginning of the catalog or take the placement test, and judge the mnemonics by the only test that matters: when the character turns up cold, does the image come back and hand you the meaning? Download it here: Moji is on the App Store.

Sources