Menu
BASICS
How Japanese schools work Japanese phrases for school Glossary of school termsWAYS TO LEARN
Free schools Education support centers Online schools Homeschooling in Japan International schools Correspondence high schoolsMORE
Neurodivergent children Where to get help Attendance and costs Learning Place Finder 日本語サイトへMy Child Sounds Rude in Japanese to Relatives: Politeness Is Overheard, Not Taught
The short answer, in six lines
- Politeness in Japanese is not mainly taught. It is watched. In a study of Japanese home recordings, only 4% of parents’ utterances were in the polite form, and only 2% of the children’s were — and yet the same researchers found children in Japan already switching correctly years before school.
- By three, children in Japan are already categorical about it. Asked what a child on screen would say to a shop assistant versus to family, three- to six-year-olds in Kyoto produced a 72% difference in polite-form use between the two situations. Three-year-olds already did it. This is not something Japanese children are drilled in.
- Where it comes from is the one thing a household abroad does not have. In the same corpus, parents used polite forms significantly more when speaking to other adults than to each other. The child learns the register by standing next to a parent who is talking to somebody else.
- So the shortage abroad is not politeness instruction. It is witnessed adult-to-adult Japanese. Your child has very probably never once seen you speak Japanese to a stranger.
- The error relatives actually notice is usually the opposite of the one you fear. Plain speech to a grandmother sounds like a child. Honorifics applied to your own family sounds like a foreigner — and that is the error children taught keigo as vocabulary reliably make.
- Japan’s own official framework has five categories, not the three in every English guide — and the two that matter for a child are the two that are almost never drilled.
About the research in this article. The numbers below come from two sources we read in the original: one open peer-reviewed study of Japanese children and their parents, and one official report of Japan’s Council for Cultural Affairs. Both are listed at the end. We have not found research that measures politeness acquisition in Japanese heritage children growing up abroad, so nothing here is presented as a measured effect for your situation. The study findings are used as a mechanism — evidence about how the register is normally acquired — and the practical sections are reasoning from that mechanism, clearly marked as such. Where we could not verify a number, we have left it out rather than softened it.
The two complaints, and why they are not the same problem
When parents say their child sounds rude in Japanese, they are almost always describing one of two quite different things, and the two have opposite causes.
The first is flatness. The child speaks to everyone in the plain form — the form they use with you, because you are the person they learned Japanese from. ありがとう rather than ありがとうございます. 見た rather than 見ました. Direct questions to a great-aunt with no softening at all. To a Japanese ear this does not sound hostile. It sounds young — like a four-year-old, which is roughly the last age at which it would have been unremarkable. Relatives generally find it endearing for about two visits and faintly worrying by the third.
The second is misdirection, and this one is much stranger to a Japanese listener. The child has picked up some honorific vocabulary — from a textbook, a Saturday school, an anime, a well-meaning parent — and applies it in the wrong direction. The clearest example, and the one Japan’s own official guidance singles out, is using respectful language about your own family when speaking to somebody outside it: saying that your father いらっしゃいます to a person who is not in your family. Grammatically flawless. Socially inside out.
Parents worry about the first and inadvertently cause the second. The whole of the rest of this article follows from understanding why.
What the research actually found, and why it is so useful
There is one open-access study we leaned on heavily here, because it happens to measure exactly the thing parents abroad need to know. Chang, Tatsumi, Hashimoto, Yatsushiro and Okada looked at Japanese politeness acquisition two ways: an experiment with children, and a very large analysis of recorded family speech.
The experiment showed children video clips and asked what the child on screen would say — once in a formal setting (a shop) and once in an informal one (family and peers). There were 57 children, from 15 three-year-olds up to 10 six-year-olds, from three Kyoto kindergartens. The result was not a gradual developmental curve. It was a cliff: a 72% difference in polite-form use between the formal and informal conditions, with most children, including the three-year-olds, being categorical — they did not hedge between the two, they used one form for one situation and the other for the other.
Then the corpus analysis, and this is the part that reframes everything. Across 311,933 parental utterances in home recordings, there were 11,936 polite tokens — only 4% of what parents said at home was in the polite form. The children themselves: 2%.
A Japanese child at home barely hears the polite form, and barely uses it — and is nonetheless fluent in the distinction before school starts
| What was measured | Result | |
|---|---|---|
| Polite forms in parents’ home speech | 311,933 parental utterances from Japanese home recordings | 4% were polite form |
| Polite forms in children’s home speech | The same recordings, the target children | 2% contained polite forms |
| Who parents were polite to | Utterances to other adults vs. between the parents | Significantly more polite to other adults |
| Parent and child, session by session | Correlation of polite-form proportion | 0.55 (medium) |
| Children’s performance when asked | 57 children aged 3–6, formal vs. informal scenario | 72% difference; most were categorical |
All five rows from Chang, Tatsumi, Hashimoto, Yatsushiro & Okada (2021), listed in the sources. Monolingual children in Japan, not heritage children abroad. The experimental task asked what a character would say, so it measures knowledge of the distinction rather than spontaneous use. We use these as evidence about the mechanism, not as a prediction about your child.
Put those two findings next to each other
A Japanese child at home hears the polite form in roughly one utterance in twenty-five, and produces it in one in fifty. That child nevertheless arrives at three years old able to switch cleanly between registers depending on who is being addressed.
Which means the register cannot be coming from being spoken to politely, and it cannot be coming from instruction, because nobody instructs a three-year-old in 敬語. It has to be coming from somewhere else, and the corpus points straight at it: parents were significantly more polite when talking to other adults than when talking to each other (the paper reports this at p=0.002), and the child’s own rate of polite speech tracked the parent’s, session by session, at a correlation of 0.55.
The child is not being taught. The child is watching their parent change — watching the same person who says ご飯、食べる？ to them turn to a delivery driver and say すみません、ここにお願いします. The information is not in the words. It is in the switch, and the switch is only visible if there is somebody present to switch for.
Now count how often your child has seen that switch
Here is the whole problem, and it takes one minute to confirm in your own house.
In the last month, how many times has your child watched you — or the Japanese-speaking parent — speak Japanese to an adult who is not a member of your household? Not on the phone to family, where the register stays warm and familial. To a stranger, or a near-stranger: a shop assistant, a colleague, a doctor, a neighbour, a parent at a Saturday school whom you do not know well.
For most families raising a child with Japanese abroad, the honest answer is zero, or close to it, and it has been zero for years. The parent’s Japanese has exactly one audience in that country: the children. Occasionally a spouse. That is it.
And this is the gap that produces both of the complaints at the top. The child never saw the plain form being traded for something else, so the plain form is simply what Japanese is — hence the flatness. And when politeness does finally arrive, it arrives as a vocabulary list rather than as a social move, so it gets applied to whatever seems important — hence the misdirection.
The error that actually sounds foreign
If you only fix one thing after reading this, fix this one, because it is the error that makes a relative quietly decide that your child has been raised somewhere else.
Japanese politeness is not a dial that runs from rough to refined. It is a direction. You raise the other person, and you lower your own side. And crucially, “your own side” is not just you. Japan’s Council for Cultural Affairs, in its 2007 official report on honorific usage, states the rule in so many words: your own side includes, for example, your own family — anyone you should recognise as being inside your circle. The counterpart rule is that your own side is not raised.
The report’s own worked example of what to avoid is almost exactly what a heritage child says. Speaking about your father to somebody outside the family and using respectful language about him — the equivalent of saying that your father いらっしゃいます — is given as a usage to steer away from.
Why do children abroad make this error so predictably? Because of how they got the words. A child who learns 敬語 as a list of fancier alternatives has no way to know that the fanciness is directional. いらっしゃる looks like a politer いる, so it must be the better one, so it is the one you use about people you love. The child is being respectful — sincerely, carefully respectful — and the output reads as though they have never belonged to a Japanese family.
Plain speech from a child to a grandmother, by contrast, is a small age error. It says: this child is younger than their years in Japanese. Misdirected honorifics say something categorically different, and much harder to unsay.
Five categories, not three — and the two that matter are not the two you are teaching
Almost every English-language explanation of keigo you will find divides it into three: respectful language, humble language, and polite language. That is the school model, and it is not wrong. But Japan’s own official framework since 2007 uses five, because the three-way split hides the distinctions that matter in practice. The 2007 report is explicit that the five-way division does not contradict the three-way one — it subdivides it.
The 2007 Council for Cultural Affairs report divides honorific language into five, and the two a child needs first sit at the bottom of the table
| Category | What it does | Examples | Priority for a child abroad |
|---|---|---|---|
| 尊敬語 | Raises the other person, or a third party, in describing what they do | いらっしゃる ・ おっしゃる ・ 召し上がる | Last — and the usual source of the misdirection error |
| 講譲語Ⅰ | Lowers your own action in describing it as directed at the person being raised | 伺う ・ 申し上げる ・ 差し上げる | Last |
| 講譲語Ⅱ | Describes your own actions courteously to the person you are talking to | 参る ・ 申す ・ いたす | Later |
| 丁寧語 | Addresses the listener courteously, whatever the content | です ・ ます | First, and almost everything |
| 美化語 | Beautifies the thing being named | お酒 ・ お料理 | Second, and nearly free |
Categories, definitions and examples from the Council for Cultural Affairs, Keigo no Shishin, 2 February 2007, listed in the sources. The priority column is our judgement for a child being raised abroad, not part of the report. Legend: ○ start here ・ △ useful later ・ × not yet.
Why that table is the practical heart of this
Look at where the two easy categories are. 丁寧語 is です and ます. It has no vocabulary to learn at all — it is a verb ending, and your child already knows every verb it attaches to. 美化語 is an お on the front of a noun. Between them they carry the overwhelming majority of what makes a child sound age-appropriate to a relative, and neither requires a single new word.
Now look at where the difficulty is. 尊敬語 and 講譲語Ⅰ are irregular, lexical, directional, and context-dependent. They are also the glamorous part — the part textbooks put in a box, the part Saturday schools test, the part a parent remembers struggling with. So that is what gets taught. And teaching those two first is precisely how you manufacture a child who honours their own father.
If your child says ありがとうございます and 見ました and お茶 instead of ありがとう and 見た and 茶, no relative on earth will describe them as rude. Nothing in that sentence is 尊敬語.
What “rude” is actually measuring when a relative says it
It is worth being precise, because the word rude sends parents in the wrong direction. Almost no Japanese relative believes a seven-year-old raised in Melbourne or Manchester is being disrespectful on purpose. What they are registering is one of three things, and it helps to know which.
The Japanese is coming out at an age younger than the child’s face. Plain forms to adults, no softening, direct imperatives. The relative’s inner reaction is protective rather than offended — and often expresses itself as speaking to your child in simpler Japanese than they can handle, which then makes the next visit worse.
Honorifics pointed the wrong way: at their own family, at themselves, at you. This is the one that reads as outsider rather than young, because no Japanese child of any age makes it. It is a specific by-product of learning keigo as vocabulary.
The small machinery of consideration — あの, すみません, ちょっと, a trailing んですけど, a rising hesitation before a request. Children who only ever speak Japanese to a parent who already knows what they want never develop any of it, because they have never once had to soften an approach.
These three have different fixes, and only the second is urgent. The first resolves itself once the register exists. The third needs other speakers, which is the subject of the next few sections.
Why teaching keigo directly tends to backfire
The instinctive response to all of this is a lesson: sit the child down, explain polite Japanese, practise some phrases for the grandparents. We would gently suggest not doing that as your first move, for three reasons.
The first is the one the research points at. The register is normally acquired by observing a model switch, not by being told the rule. A child who is told the rule learns a rule — something they can state and then fail to apply under time pressure, which is every real conversation.
The second is the direction error. Explicit teaching, in the absence of lived examples, nearly always teaches the words before the geometry. If a child does not yet feel the inside and outside of their own family as a line, respectful vocabulary has nothing to attach to and lands wherever affection points.
The third is motivational, and it is the quiet one. A child who is corrected for rudeness in Japanese learns that Japanese is a language in which they are routinely in the wrong. We have written at length elsewhere about what happens to a child’s willingness to speak a heritage language once speaking it becomes a test — see our piece on a teenager who refuses to speak Japanese. Politeness correction is an unusually effective way to start that process early, because the child cannot tell which sentences are safe.
The polite register is learned by watching a fluent speaker change for a listener — which is the one thing a household abroad cannot stage. TOMODACHI JAPAN runs small online classes with children in Japan and a teacher who is not a family member, which puts your child in a room where the plain form is not the default and somebody other than a parent is being addressed. The school guide is free to read.
What does supply the register, in order of how much it gives you
If the register comes from witnessed switching and from being addressed by adults who are not your parents, then the useful question is which arrangements actually produce those two things. They are not the arrangements most families buy first.
| One-to-one private tutor | Saturday supplementary school | Small online class with peers in Japan | |
|---|---|---|---|
| An adult who is not family addresses the child | Yes | Yes | Yes |
| The child watches that adult address somebody else | Never — the child is the only listener | Constantly | Yes, every lesson |
| The child must handle being misunderstood | Rarely — a paid tutor works hard to understand | Sometimes | Yes |
| Peer-to-peer Japanese with no English fallback | None | Depends on the cohort — often English in the corridor | Yes, if the peers are in Japan |
| Available where no Japanese community exists | Yes | No | Yes |
| Where it is weak | Excellent for reading, writing and explicit grammar | Strong but needs a local community and a Saturday | No playground, fixed time slots, and a time difference that can put lessons at an unkind hour |
General pattern as at October 2026, written from the mechanism described above rather than from a trial. Legend: ○ supplies it ・ △ partly ・ × does not. The highlighted column is the kind of provision our own school runs, and its weaknesses are stated in the last row.
The counter-intuitive bit about tutors
A private tutor is the first thing most families buy when they notice a politeness problem, and for this particular problem it is close to the least effective purchase available. The reason is in the second row of the table: a one-to-one lesson contains no witnessed switching, because the child is the only person in the room to switch for. The tutor speaks politely to the child, constantly — which teaches the child that polite Japanese is what adults aim at children, the exact inverse of the social fact.
There is a second, subtler problem. A tutor is a paid, patient, maximally charitable listener. So is a parent. A child who speaks Japanese only to people who are working hard to understand them never has to repair an utterance, and never discovers that the softening machinery in item 03 above has a function. This is not an argument against tutors — they are the best thing available for reading, writing and explicit grammar. It is an argument against expecting a tutor to fix the thing relatives are noticing.
Three audiences you can manufacture, even with no Japanese community nearby
The honest constraint is that most families abroad cannot produce real Japanese strangers on demand. But the mechanism does not require strangers. It requires your child watching Japanese being used with somebody who is not them. That you can arrange, and all three of these are free.
One: do not take the call in another room. When you ring a Japanese bank, a school in Japan, a city office, a hotel, a relative’s care home — anything administrative — take it where your child can hear it, and do not lower your voice. Twenty minutes of you being formal with a call centre is worth more to this specific problem than a month of lessons, because it is the only time your child will ever hear you be deferential. Most parents instinctively do the opposite and step out of the room so as not to be overheard.
Two: hand your child a third party on the same call. Grandparent calls are usually two-way and intimate, which keeps the register exactly where it already is. Change the shape instead. Ask the grandparent to have a neighbour, a friend, a carer or an older cousin say hello for two minutes. That person is outside your family, and the child now has a reason to shift — and, just as importantly, will hear the grandparent shift when addressing them, which is the demonstration.
Three: narrate yourself in two directions. At the end of a shop interaction in your own country, say in Japanese what you would have said in Japanese: the すみません before the question, the ありがとうございます afterwards. It is artificial and your child will find it odd for about a week. It is also the cheapest way to show the switch dozens of times in a month.
What to say to the relative, rather than to the child
There is one conversation that helps more than any amount of practice, and it is not with your child. It is with the relative who is going to be on the receiving end.
Relatives do two things, from kindness, that make this worse. They simplify their Japanese to the level they think the child is at, which removes the model the child needs. And they answer questions the child was asked, or finish the child’s sentences, which removes the need to produce anything.
So the request is small and specific: please speak to our child in the Japanese you would use with a child in Japan of the same age, including the polite bits, and if they say something blunt, please do not correct it — just answer in the form you would have used. That last clause is the whole of it. A relative who reliably recasts — who hears ありがとう and replies naturally with a sentence containing ありがとうございます in it — supplies the model without ever creating a test. Children correct themselves towards what they hear far more reliably than towards what they are told.
Where a grandparent is the main Japanese relationship your child has, the way those calls are designed matters a great deal on its own, and we have written separately about it in our piece on a child who cannot really talk to their Japanese grandparents.
If your child is already a teenager
The calculation changes, in both directions. A teenager can be told the rule and will understand it immediately — the inside-and-outside geometry is not conceptually hard for a fourteen-year-old, and ten minutes of explanation will stop the direction error for good. That is a genuinely quick win and worth doing straight away.
What does not work at that age is practice imposed from above. A teenager who is self-conscious about sounding wrong in Japanese will reduce their Japanese output rather than increase their accuracy, which is the opposite of what you want. Give the explanation once, clearly, as information rather than correction, and then put the effort into arrangements where somebody other than you is the listener.
Where this fits with everything else you are worried about
Politeness is one slice of a broader pattern, and it is worth seeing the shape of the whole thing rather than treating each symptom separately.
A child raised with Japanese in a household abroad typically ends up fluent in one register — the warm, familial, plain register they use with a parent — and thin or absent in the others: the register children use with each other, the register used with adults outside the family, and the written and school register. We laid that map out in detail in our article on a child with no Japanese-speaking peers, which deliberately left politeness aside for this article. The flatness relatives notice is the third of those registers being missing. The direction error is what happens when you try to fill it with vocabulary.
Seen that way, the politeness problem is good news of a sort. It is not a sign that the Japanese is failing. It is a sign that the Japanese is working in exactly the one setting it has ever been used in.
Who this article is not right for
- A child who is not producing much Japanese at all. If your child understands but will not speak, politeness is the wrong thing to work on — the register problem is downstream. Start with a child who understands Japanese but will not speak it.
- A household where nobody speaks Japanese to the child daily. Witnessed switching requires somebody to do the switching. If the Japanese-speaking parent is rarely present, the first task is securing ordinary household Japanese, not the polite register.
- A family moving to Japan within the year. Politeness will be resolved by immersion faster than by anything you can do now, and your attention is better spent on reading and on classroom Japanese. See moving back to Japan with a child whose Japanese is behind.
- A child under about four. Nothing described here is a problem yet. Plain speech from a three-year-old is plain speech from a three-year-old everywhere in Japan.
- A family where the relative is genuinely upset rather than amused. That is a relationship conversation, and it will go better held directly than routed through your child’s grammar.
What it costs if nothing changes
We want to be careful here, because this is where articles like this one usually start inventing consequences. We did not find research measuring what happens to heritage speakers of Japanese who never acquire the polite register, and so we are not going to give you a figure or a prediction.
What can be said is structural. The register that is missing is not a decorative one. It is the register in which a Japanese adult applies for work, registers at a city office, speaks to a doctor, talks to a colleague, or meets their partner’s parents. A person can be entirely fluent in family Japanese and still be unable to do any of those things without sounding like a child, and the gap is usually discovered at the worst possible moment — at about nineteen, in a situation that matters, when it is no longer a thing a parent can help with.
The cheerful half of that is how small the fix is while the child is still at home. 丁寧語 is a verb ending. 美化語 is one syllable in front of a noun. What is scarce is not the grammar. It is a listener who is not you.
A 90-day plan
- Week 1 — audit, do not teach. Count the times in a week your child hears the Japanese-speaking parent address an adult outside the household. Write the number down. For most families it is zero, and knowing that is the point of the exercise.
- Week 1 — fix the direction error, once. If your child honours their own family, explain it one time, plainly: we lower our own side and raise theirs, including about Mum, Dad and grandparents. Do not practise it, do not test it, do not raise it again.
- Weeks 2–4 — stop leaving the room. Every administrative call in Japanese happens within earshot. Aim for four in the month.
- Weeks 2–4 — ask one relative for the recast favour. Not a lesson, not corrections. Answer in the form you would have used.
- Weeks 4–8 — put one non-family adult in the child’s week. A class, a club, a small group — anything where somebody who is not a relative addresses your child in Japanese, and where your child can see that person address others.
- Weeks 4–8 — add a third party to one call a fortnight. A neighbour, a cousin, a friend of the grandparents, for two minutes.
- Weeks 8–12 — look for the right signal, not accuracy. You are not looking for correct 敬語. You are looking for variation: any sign that your child sounds different with one listener than with another. That is the register arriving. The accuracy follows it, and never precedes it.
- Throughout — do not correct politeness in the moment. Recast and move on. The cost of correction is output, and output is the only raw material you have.
Frequently asked questions
Sources used for this article
- Chang, F., Tatsumi, T., Hashimoto, H., Yatsushiro, M. & Okada, N., The Role of Parental Input in the Early Acquisition of Japanese Politeness Distinctions, Collabra: Psychology 7(1), 2021, article 18989, DOI 10.1525/collabra.18989. Open access. Source of the 72% difference between formal and informal conditions across 57 children aged 3–6 from three Kyoto kindergartens; of the 311,933 parental utterances containing 11,936 polite tokens (4%) and the children’s 2%; of the finding that utterances to other adults contained significantly more polite forms than those between the parents; and of the 0.55 correlation between parent and child polite-form proportions per recording session. Important limits: these are monolingual children in Japan, not heritage children abroad; the experimental task asked what a character would say rather than recording spontaneous use; and the authors themselves note that home recordings may have limited their ability to observe politeness distinctions in child speech, since the role-play contexts in which children display them most clearly were not captured. We use these findings as evidence about how the register is normally acquired, not as a measured effect for heritage speakers.
- Council for Cultural Affairs (文化審議会), 敬語の指針 (Guidelines on Honorific Language), report of 2 February 2007, Agency for Cultural Affairs. Source of the five-category framework (尊敬語 / 講譲語Ⅰ / 講譲語Ⅱ / 丁寧語 / 美化語), the definitions and examples in the table above, the report’s own statement that the five-way division subdivides rather than contradicts the conventional three-way one, the rule that one’s own side is not raised, the statement that one’s own side includes one’s own family, and the worked example of using respectful language about one’s own father to an outsider as a usage to avoid.
- Nijiiro, Only child with no Japanese-speaking peers. The register map referred to above, which set politeness aside for separate treatment.
All sources were opened and read in the original. Where a number appears in this article it comes from one of the two primary sources above. Claims that we could not verify in a primary source — in particular any figure linking heritage politeness acquisition to later outcomes — have been left out entirely rather than hedged. Last checked 7 October 2026.
Polite Japanese is acquired by watching a fluent speaker change for somebody else, and by being addressed by adults outside the family — neither of which a household abroad can produce on its own. TOMODACHI JAPAN runs small online classes with children in Japan and teachers who are not relatives, grouped by what each child can actually say rather than by age. The school guide is free to read.
Published: 7 October 2026 / Last updated: 7 October 2026
