Listen at the door some evening. You speak Hindi to your daughter and she answers — haltingly, maybe, but she answers. Then she turns to her little brother to argue about whose turn it is on the tablet, and the whole exchange happens in English. Every syllable. It's as if a switch flipped the moment the conversation stopped including you.
It did. And once you understand why, the sibling switch stops looking like a failure of your efforts and starts looking like exactly what it is: two children speaking the language of their shared world. The question worth asking isn't how to force them back. It's how to widen the world their heritage language belongs to — especially for the younger one, who has more riding on this than anyone in the house realizes.
Sibling Talk Is Peer Talk
Sociolinguists have long observed that multilingual people don't hold their languages in one undifferentiated pool. Languages get indexed to domains — the sociolinguist Joshua Fishman famously framed it as the question of who speaks what language to whom, and when. A child sorting out two or three languages is doing exactly this kind of mapping, constantly and mostly unconsciously: Hindi is what happens with Mummy, Tamil with Paati on the phone, English at school.
And siblings? Siblings are each other's first peers. Their relationship is built out of peer material — games, jokes, alliances, grievances, the entire private economy of childhood. Nearly all of that material lives at school and on the playground, which means it lives in English. When your daughter switches languages to argue with her brother, she isn't rejecting anything. She's converging on the language in which their shared life actually happens, the one that already contains the game's rules, the joke's setup, the insult's precise sting. It's linguistic economy, not disloyalty.
The Pattern Researchers Keep Finding
This switch is so common it shows up as a statistical signature. The linguist Annick De Houwer surveyed thousands of multilingual families in Flanders and found that hearing a language at home was no guarantee of speaking it — in roughly a quarter of homes where parents used a heritage language, the children did not actively speak it themselves. Input is necessary; it isn't sufficient. A child needs reasons to produce the language, not just chances to absorb it.
Sibling dynamics turn out to be one of the strongest forces tipping that balance. Researchers who study bilingual families — Sarah Shin's work with Korean American children is a good example — have documented a recurring sequence: the home runs largely in the heritage language until the firstborn starts school. Within months, English comes home in the firstborn's mouth, and the sibling channel flips first. The younger child then grows up in a measurably different house than the older one did. The firstborn got years of undiluted parent-language input; the second-born arrives to find English already circulating at child height, spoken by the person they most want to imitate.
This is the birth-order effect nobody warns you about. When parents notice that the younger child's Hindi or Telugu is weaker than the older one's was at the same age, they often blame themselves for slacking off. Usually they didn't slack. The linguistic ecosystem changed, and the older sibling — lovable, fluent, endlessly present — is the one who changed it.
What the Switch Doesn't Mean
It helps to be precise about the damage, because it's easy to overestimate. Sibling English does not erase what you're building. Comprehension keeps growing as long as you keep talking; the receptive foundation — the part that lets a child follow a grandmother's story or pick the language back up as a teenager — is laid by input, and your input continues. The switch also says nothing about identity. Children who chatter to each other in English can still feel deeply, unselfconsciously Indian.
What the switch genuinely costs is production practice for the younger child. Speaking a language is a skill assembled through use, and the sibling channel is where a huge share of a small child's total talking happens. If that entire channel runs in English, the younger one may end up understanding everything and saying little — fluent ears, shy tongue.
Why Policing Sibling Talk Backfires
The instinctive fix — "Speak Hindi to your sister!" — is the one intervention that reliably makes things worse. Research on family language policy keeps arriving at the same conclusion: enforcement without motivation breeds resistance. When the heritage language becomes the language of correction, the thing children get scolded about, it accumulates negative feeling, and children avoid what feels bad. Worse, the sibling relationship is theirs — one of the few territories in a managed childhood that parents don't run. Commandeering the language of that territory turns your mother tongue into a tax on their intimacy. They will pay the tax in your presence and evade it everywhere else.
Change the Waters, Not the Fish
What works instead is quieter: stop trying to regulate which language the children use, and start engineering situations where the heritage language is simply the natural medium — where switching to English would make the activity worse, not easier.
A video call where Nani's jokes only land in Hindi. A bedtime story that exists in Tamil and nowhere else. Cooking that comes with its own vocabulary — tadka, atta, thoda sa — that English has no words for. A song game in the car. Festival preparations with their own scripts and phrases. The rule of thumb: attach the language to things children want, never to things they owe.
Here is the part parents rarely expect. When siblings share an activity that genuinely lives in the heritage language — a story world with characters they both love, a song they both know, a running joke from a myth — fragments of that language start leaking into their private register on their own. A character's name. A punchline. A dramatic line delivered in Hindi because it's funnier that way. Their sentences will still be mostly English, seamed with borrowed words, and that mixing is not a defeat. It's a doorway. It means the language has gotten into the shared world, which is the only place sibling speech was ever going to come from.
Give the Younger One Their Own Doorway
The younger child needs one more thing: a firsthand relationship with the language, not a hand-me-down one. Barbara Zurer Pearson's research on bilingual children points to a plain truth — the amount of input a child personally receives in a language predicts what they build in it. So carve out dyad time that belongs to the younger one alone. Ten minutes of one-on-one talk in the mother tongue. Their own story at bedtime, not just the tail end of the older one's. Their own turn on the call to India, with the older sibling out of the room, so there's no fluent interpreter to hide behind.
And occasionally, let the younger one be the knower. Teach them a word, a song, a story the older sibling hasn't heard, and watch what it does to their willingness to speak. A language you can be the expert in is a language worth opening your mouth for.
The Current Moves Slowly Both Ways
Language shift in immigrant families rarely happens by decision. It happens by default — a thousand small conveniences, each individually harmless, all leaning the same direction. Reversing it works the same way in miniature: not one dramatic house rule, but a slow re-weighting of what the language is for, until it's the medium of enough beloved things that it keeps showing up on its own.
That's the idea behind KathaKids, which gives Indian-origin kids a shared world worth talking about — mythology told for young listeners, festivals with their whys intact, food and language woven into stories rather than drills. When both siblings know the same tales and the same characters, the heritage language stops being homework and becomes material: something to quote, argue over, and giggle about in the back seat, in whatever glorious mix of languages they choose. If you'd like that world within reach at bedtime, you can find it at baalkatha.lumenlabs.works.