The short version

Speaking and reading are not the same skill, and only one of them comes from being around people

There is a reason this gap surprises parents, and it is worth naming because it removes a lot of unnecessary guilt.

Spoken language is acquired. A child who hears enough Japanese from enough people, in situations where Japanese is the only way to get what they want, will build a sound system, a vocabulary and a grammar without anybody teaching a single rule. It happens below the level of instruction. You did not teach your child the rules for particles any more than an English-speaking parent teaches the ordering of adjectives. It arrived.

Reading is not acquired. Writing systems are inventions. They are roughly five thousand years old, against a few hundred thousand years of speech, and no human being has ever picked one up by being in the room while other people used it. Every literate person on earth was taught, deliberately, over years, by somebody whose job that was. This is as true of English as it is of Japanese — the difference is that your child is being taught to read in English, five days a week, by a school, and nobody is doing the equivalent for Japanese.

So the asymmetry in your house is not a mystery. It is exactly what the inputs predict. Your child got the full, immersive, free version of the thing that comes free, and almost none of the deliberate, scheduled, expensive version of the thing that only comes from instruction.

What Japan’s own curriculum actually budgets for this

Most advice on this topic is written as though the job were small — a workbook here, an app there. It helps enormously to see the real size of what a Japanese elementary school spends on literacy, because it reframes the question from “why is my child failing at this” to “how much of this do we actually want, and what will it cost.”

Japanese elementary school runs six years. The standard number of class hours is set nationally. Here is what the national standard allocates to Japanese language alone, and what the total teaching time looks like.

Read the first column again. 1,461 hours of Japanese-language lessons — roughly 243 hours a year, close to five hours every single school week, for six consecutive years. And that is only the subject called Japanese. The 5,645 figure is the whole school day: maths, science, social studies, all of it delivered in Japanese, read from Japanese textbooks, written up in Japanese notebooks. A child in Japan is not doing 1,461 hours of literacy. They are doing 1,461 hours of explicit literacy instruction inside 5,645 hours of literacy practice.

Now put your own situation next to that. A Saturday supplementary school that runs three hours of Japanese a week for forty weeks delivers 120 hours a year. Two hours a week at home, kept up faithfully, delivers about 100. Those are good numbers. They are also, honestly, somewhere between a quarter and a half of the national pace — which is precisely why a child doing everything right at a Saturday school still ends up two or three grades behind the kanji schedule, and why their parents keep wondering what they are doing wrong.

They are not doing anything wrong. They are doing a fifth of the hours and getting more than a fifth of the result.

The three scripts, and which one is actually the problem

“Japanese reading” is a misleading phrase, because the three writing systems behave completely differently and only one of them is hard.

This is the single most useful thing to understand about your child’s situation. Kana is a solved problem. A nine-year-old who speaks Japanese fluently can learn to decode hiragana to a usable standard faster than most parents expect, because the hard half of reading — knowing what the word means once you have sounded it out — is already done. They are not learning a language. They are learning a cipher for a language they own.

Kanji is a different animal, and it is worth being precise about why. It is not that the characters are complicated to draw. It is that there are a great many of them, that each one typically has more than one reading, that the correct reading depends on the word it appears in, and that the whole thing is cumulative: Grade 4 text assumes Grades 1 to 3. You cannot skip to the interesting material, because the interesting material is written using everything that came before.

And critically: you cannot get kanji from listening. Everything else in Japanese — vocabulary, grammar, pitch, idiom, register — leaks into a child from the people around them. Kanji leaks from nowhere. It is the one component with no free channel, which is exactly why it is the component that is missing.

The literacy cliff: why reading so often stops at the same place

Families who get kana working usually report the same arc, and it is worth recognising because it looks like failure and is not.

First, things go well. The child decodes hiragana, slowly then quickly. They read a book aimed at four- and five-year-olds — written almost entirely in kana — and everyone is delighted. Then they pick up something written for a seven-year-old, and it falls apart inside a page. Not because the language is harder. Because the text has started using kanji, and every third or fourth content word is now a wall.

That is the cliff. And the mechanics of it are brutal in a specific way:

There is one more piece, and it is the painful one. Around the same age, your child is becoming a genuinely good reader in English — chapter books, confidence, speed, identity as a reader. Japanese reading is now the thing that makes them feel stupid. Nobody keeps volunteering for that. The refusal that parents read as attitude is usually just a child protecting a self-image, and it is worth saying out loud to them: this is hard because it is hard, not because you are slow at it.

Four skills that people collapse into “reading and writing”

Here is where most family plans go wrong. “Reading and writing Japanese” is treated as one goal, so it gets one method — usually a workbook — and the workbook is built around the single most expensive skill in the set.

Look at skill 4 for a moment, because it is the one almost nobody plans for and it changes what is possible.

To type Japanese, you type the sounds of a word in Latin letters or kana, and the input method offers you the written forms to choose from. You never have to produce a character from memory. You only have to recognise the right one when you see it. That means a child who can speak Japanese and read kana can begin writing real Japanese — messages to a grandparent, captions, comments, a diary — almost immediately, years before they could write the same sentences by hand.

And because every typed word is a recognition exercise with instant feedback and a real reason to get it right, typing is not a shortcut away from kanji. It is one of the better ways in.

Which of the four is worth your time first

This depends almost entirely on one question: is your child likely to attend school in Japan? If yes, handwriting moves up sharply, because Japanese schools assess it directly and a child who cannot write by hand will be marked as behind regardless of how well they read. If no, handwriting is the lowest-return skill in the set and most families should stop treating it as the centre of the plan.

Here is an honest comparison of the three routes families actually choose. We run an online Japanese school, so take the last column with appropriate scepticism — which is why we have also marked where it is the wrong answer.

The row that matters most for most readers is the last-but-one. A plan that your child refuses is worth zero, whatever its theoretical merits, and the commonest cause of refusal is a plan built entirely out of handwriting drills for a child with no reason to handwrite.

The advantage your child has that an adult learner does not

Almost every kanji resource on the market was designed for adults learning Japanese from zero. That matters, because those learners face three problems at once for every character: a new meaning, a new sound, and a new symbol. Entire teaching systems exist to manage that triple load — mnemonics, radical breakdowns, elaborate story methods.

Your child has one of those problems, not three.

When your child meets 公園 for the first time, they are not learning what a park is, and they are not learning that it is called kōen. They have been to the park. They have said the word a thousand times. All they are learning is that this particular pair of shapes is the thing they already say. That is a labelling task, and labelling is far cheaper than acquisition.

This also tells you which order to do things in. Adult courses order kanji by stroke count, by frequency, or by an invented learning sequence. For your child, the right order is the order of the words they already say. Their own name. Family words. Food. The names of the things in your house. Numbers and days. Verbs they use daily. A child who can read fifty characters that appear in their own life is in a much stronger position than one who has drilled two hundred they have never needed.

It is also the reason spoken Japanese is not a side issue in a literacy plan. The spoken vocabulary is the fuel. A child whose speaking is quietly shrinking is losing the thing that makes reading cheap — which is why the families who get literacy working are rarely the ones who dropped speaking to focus on kanji.

What does not work, and why it keeps being recommended

None of the following are useless. Each is widely recommended, each has a real use, and each fails as the centre of a plan.

The thread running through all six: they treat kanji as a list to be memorised rather than as a property of text that a child meets repeatedly until it becomes familiar. Memorisation without encounter does not stick, and encounter without comprehension is just noise. You need both at once, which is the whole of the method below.

Read what you can already say

The method has one rule, and everything else is implementation: your child should read sentences whose meaning they would have understood instantly if you had said them out loud.

That single constraint does a lot of work. It keeps comprehension at 100%, which keeps reading a pleasure rather than a test. It puts every new character inside a word the child owns, which is the labelling task rather than the acquisition task. And it means progress is measured in text read, not characters drilled.

Handwriting: when it matters, and when it does not

Handwriting deserves its own section because it absorbs more family effort than anything else and the decision about it should be deliberate rather than inherited.

It matters a great deal if your child is going to school in Japan. Japanese schools test writing directly, kanji dictation is routine, and a child who reads well but cannot write will be assessed as behind. If a move is likely within a few years, handwriting is not optional and the Saturday-school route exists precisely for this.

It matters much less otherwise. Japanese adults living in Japan write by hand less every year; most text is typed. For a child whose Japanese life will be conducted with family, with friends and online, handwriting is closer to a cultural skill than a functional one — genuinely worth having, not worth sacrificing reading for.

A reasonable middle path for families not returning: keep handwriting alive at a small, fixed scale — their name, family names, numbers, dates, the thirty or so characters that appear in their own life — and spend the rest of the budget on recognition and reading volume. This keeps the door open without letting the most expensive skill dictate the plan.

If you are thinking about moving to Japan

This changes the calculation more than anything else in the article, so it is worth being blunt about the numbers.

Japan’s education ministry counts children in its public schools who need Japanese-language support. As of 1 May 2025 that figure was 84,759 — and 11,446 of them hold Japanese nationality. That second number is the one that matters here. Those are, in large part, children of returning families: Japanese citizens, often fluent speakers, who arrived without the literacy the classroom assumes.

The outcomes attached to that group are sobering. Among students identified as needing Japanese-language support, the rate going on to university or equivalent is 41.2%, against 75.0% for students overall; the high-school continuation rate is 91.5% against 98.9%. Those gaps are not about intelligence or about spoken Japanese. They are about reading and writing the language the entire curriculum is delivered in.

So if a return is plausible, treat literacy as the priority it is, start earlier than feels necessary, and put handwriting back in the plan. If you want the school-side view of that transition rather than the language-side view, we have separate guides on a returnee child struggling at Japanese school and on how Japanese public elementary school actually works.

Where an online Japanese school fits, and where it does not

We should be straightforward about this, because the honest answer is narrower than the marketing answer.

The reason a child abroad stops speaking Japanese is almost never ability; it is that everyone they speak Japanese with also speaks their other language, so there is never a moment when Japanese is the only way through. We have written about that mechanism at length in the case of a child who understands but will not speak, and the same logic applies between siblings who have switched to English with each other. The fix is people who cannot switch.

That is what TOMODACHI JAPAN is built for: small online classes grouped by what a child can actually say, alongside peers in Japan who do not speak English. It keeps the spoken base growing, and the spoken base is the fuel that makes reading cheap for a heritage child.

What it is not is a kanji curriculum. If your child speaks comfortably and the only gap is literacy — if what you actually need is 1,026 characters on a schedule with assessment attached — a supplementary school or a literacy-focused tutor is the better tool, and we would rather say so than sell you the wrong thing.

Who this approach is not right for

How to tell whether it is working

Grade levels are the wrong instrument here; they will tell you your child is behind for years, which is accurate, useless and demoralising. Use measures that move on the timescale you are actually working on.

A realistic twelve-week plan

This assumes a child who speaks Japanese comfortably, reads little or no kanji, and is not about to move to Japan. Twenty minutes a day, five days a week — about twenty hours over the quarter. Small, and far more than most families manage, because it is sized to survive a bad week.

If you want the wider picture — speaking, siblings, grandparents, mixed-language households — the rest of this series is collected on our hub, Keeping Japanese alive when you live abroad.

Questions parents ask

The sources behind the numbers in this guide

Every word they can say is a word they will read almost for free

Reading is cheap for a child who already owns the vocabulary, and expensive for one who does not — which is why the spoken base is worth protecting while you work on literacy. TOMODACHI JAPAN puts children abroad in classes of eight with peers in Japan who do not speak English, grouped by what each child can actually say. The pilot opens in December 2026 and the school guide is free.

Visit TOMODACHI JAPAN