You ask, in Hindi, if she'll bring her shoes to the door. She pads off, returns with both shoes and the missing sock you didn't mention, and announces — in clear, unbothered English — that she's ready. Every word you said landed. Not one came back.
If you've lived this small scene on repeat, you already know the particular ache of it. The understanding is right there. The speaking has gone somewhere you can't see. Parents tend to read this as failure, theirs or the child's, and quietly grieve a language slipping away. It helps to know that what you're watching has a name, a well-documented shape, and — this is the part worth holding onto — a way back.
The gap has a name
Linguists call a child who comprehends a language but rarely produces it a receptive, or passive, bilingual. It is one of the most common outcomes for kids growing up with a minority language at home and a majority language everywhere else. The child isn't half-anything. They have a real, working knowledge of the language living on the comprehension side of the ledger, waiting on the production side to catch up.
And here's the thing that should lower the temperature: in every language, for every child, understanding comes before speaking. A one-year-old follows "where's the ball?" months before she can say it. Comprehension always runs ahead of production. What happens in heritage-language homes is that this normal head start stretches into a wide, persistent gap — and then, often, stalls there.
Why understanding is cheaper than speaking
To understand a sentence, your daughter's brain only has to recognize. The words arrive in order, carried by your tone and the context of shoes and the door, and she matches them to meaning. It's pattern recognition with a lot of scaffolding already in place.
To speak the same idea, she has to do something far more expensive. She has to retrieve the right words from memory with no prompt, assemble them in the correct grammatical order, conjugate, gender her nouns, and drive her mouth to make sounds her English-dominant day rarely asks of her — all in real time, while she'd rather just be ready to go. Production is retrieval plus assembly plus motor execution. Comprehension is recognition. No wonder the cheaper skill races ahead.
Now add the second pressure. Language researcher Erika Hoff and colleagues have shown that the relative amount of input a bilingual child hears in each language strongly predicts how well they know it — not parental wish, not heritage, but sheer quantity and richness of exposure. A few instructions at dinner and a video call with grandparents on Sundays is simply not enough fuel to keep a whole language active enough to speak. The child hears just enough to understand, never enough to need to produce.
How the home language gets demoted
There's a social engine underneath the cognitive one. Around the time kids start school, the majority language becomes the language of play, of friends, of jokes, of the self they're building outside the house. English stops being a subject and becomes an identity. The heritage language, meanwhile, quietly shrinks into the language of logistics and correction — brush your teeth, don't do that, say it properly. Commands and criticism. Not exactly an invitation to linger.
This is where many of us, trying hard, make it worse. Your child ventures a sentence in Hindi, gets a verb ending wrong, and you correct it — kindly, instinctively. Stephen Krashen's affective filter hypothesis describes what happens next: anxiety, embarrassment, and self-consciousness act like a filter that blocks language from going in or coming out. A child who feels tested every time she opens her mouth in the home language learns the safe move is to answer in the language where she's never wrong. So she does. And the gap sets like concrete.
What actually closes it
The encouraging news is that a receptive bilingual is standing on a deep foundation. The vocabulary, the grammar, the ear — it's all built. You're not teaching from zero; you're activating something that already exists. That's a much smaller job, and a few shifts do most of the work.
Give the language something to carry. A language that only delivers instructions is easy to set down. A language that delivers a story, a grandfather's terrible pun, the name of the dish she loves, the secret of how the festival lights got their meaning — that language becomes worth reaching for. Tie it to affect and narrative, the things memory clings to, and you give her a reason to produce, not just receive.
Flood the comprehensible input — and make it narrative. Krashen's companion idea, the input hypothesis, holds that we acquire language by understanding messages slightly beyond our current level. Stories are the perfect vehicle: rich, repetitive in the right ways, emotionally charged, and self-explaining through plot and picture. A child leaning into "what happens next" is absorbing structure she'd never sit still for in a drill.
Respond to what she says, not how. When she offers a wobbly sentence, react to the meaning — and if you want to model the correct form, recast it: simply say it back the right way as part of a natural reply, no red pen. "Haan, woh sher bahut bada tha!" Researchers find recasting feeds correct forms back without triggering the shame that shuts kids down.
Create need, not performance. Don't ask her to translate or "say it in Hindi for Dadi." Build situations where the language is the natural currency — a relative who genuinely doesn't switch to English, a song with a gap she fills, a story she's heard so many times she beats you to the next line.
Two fears you can put down
First, you have not confused her. Decades of research are clear that growing up with two languages doesn't cause delay or muddle. The mixing you hear — a Hindi noun in an English sentence — isn't a sign of a jammed circuit; it's code-switching, a skill bilinguals use deliberately, often to land exactly the word that fits best.
Second, the door has not closed. Receptive knowledge is remarkably durable, and adults who were passive bilinguals as children routinely "reactivate" into fluent speakers when life finally gives them a reason and a stream of input. You are not racing a deadline. You are keeping a fire banked, so it's ready when the wind changes.
The quiet work of keeping a language alive
The child who understands but won't answer isn't losing her heritage. She's holding it in receivership, waiting for the home language to feel less like a test and more like a place she wants to be. Your job was never to drill her into fluency. It's to keep the language warm, frequent, and bound up with delight — so that producing it someday feels like coming home rather than crossing a border.
That's the slow, ordinary work a good story does better than almost anything. KathaKids was built around exactly this — Indian festivals, myths, language, and food carried in narrated stories a child wants to lean into, the kind of rich, emotional, comprehensible input that turns a passive ear into a speaking voice without a single correction. If you've been waiting for the home language to come back, you can start by giving your child a reason to listen for it tonight: baalkatha.lumenlabs.works.