There’s one thing we hear from bilingual families over and over again: “My child understands everything I say. They just answer me in English.”
It can be confusing, especially when the understanding is clearly there. You ask a question in Spanish and they answer it correctly in English. Their grandparents speak to them in Mandarin and they follow the conversation without a problem. They know what’s being said. So why won’t they say it back?
There’s actually a pretty big difference between the language a child can understand and the language they can produce.
Researchers usually talk about this in terms of receptive and expressive language. Receptive language is what a child can understand; expressive language is what they can actually retrieve and use themselves. The two develop together, but not necessarily at the same pace, and the gap can be especially noticeable for children growing up with a heritage language.
Think about a child whose parent asks, “¿Quieres la pelota roja o la azul?”
To understand the question, the child has to recognize enough of the words and sentence to know that they’re choosing between the red ball and the blue one. But to answer in Spanish, they have to retrieve the Spanish word themselves and produce it quickly enough to keep the conversation going.
Or they can just say “red.”
And if “red” works every time, they get a lot more practice understanding Spanish than retrieving it.
That distinction matters because research on bilingual development increasingly suggests that it isn’t enough to look only at how much of a language a child hears. Researchers also look at how children use each language and who they use it with. A child can have quite a lot of exposure to a language without having nearly as many opportunities where they actually need to produce it.
Some of the most interesting research on language development goes even further and looks at what happens during back-and-forth interaction.
In a now well-known study from the University of Washington, Patricia Kuhl and her colleagues exposed nine-month-old American babies to Mandarin Chinese. The babies had twelve sessions with the language. Some interacted with Mandarin speakers in person, while others experienced Mandarin through recorded audiovisual or audio-only material.
The babies who interacted with the Mandarin speakers learned to distinguish Mandarin speech sounds. The babies who received the recorded exposure did not show the same learning.
Of course, babies learning to distinguish speech sounds are very different from a five-year-old who understands Mandarin but answers Mom in English. But the study raised an idea that has shaped a lot of language research since: the amount of language entering a child’s ears isn’t the whole story. The social interaction happening around that language matters too.
More recent research has looked at conversational technology from a similar angle. Ying Xu and colleagues have studied what happens when children don’t just listen to a story, but have a back-and-forth conversation about it with a conversational agent. When the agent asked questions, listened to children’s answers and responded to what they said, children showed stronger story comprehension and vocabulary learning than children who experienced the same story without that dialogue.
It’s an interesting difference because the child may be hearing similar language in both cases. What changes is that at some point, they have to take a turn.
And taking a turn is practice.
Not practice in the “How do you say apple in Spanish?” sense. It’s practice finding a word because somebody is waiting for your answer.
That distinction has made us think a lot about another thing we hear from parents: their child can speak the language, but often doesn’t want to when asked.
One of our mothers told us this about her five-year-old daughter:
“I’ve never heard her speak so much Mandarin as she does with Phoenyx. When she’s home and I say ‘say something in Mandarin,’ she doesn’t want to. But with Phoenyx, it was fun.”
There’s something really interesting buried in that sentence.
Her daughter already knew some Mandarin. Phoenyx hadn’t suddenly put all of those words into her head. The difference was that she was willing to try using them.
We’ve heard similar things from other families. Kids who normally answer their parents in English will give Phoenyx a Mandarin or Spanish word. They’ll start a sentence in one language and finish it in another. They’ll repeat something Phoenyx said and change one part of it. Some answers are tiny. Some aren’t grammatically right.
But they’re attempts to speak.
And that has made us much less interested in whether every sentence is correct and much more interested in whether a child is trying to produce the language at all.
There’s an important difference between asking a child to demonstrate a language and giving them a reason to use it.
“Say something in Mandarin” is surprisingly hard. There are thousands of possible things to say, no real reason to say any of them, and suddenly everyone is listening to see what you know.
But imagine a child is drawing and someone asks whether the next star should be 红色 or 蓝色. Now there’s a reason to answer. The point isn’t to prove that they know a color word. They just really want the star to be blue.
This is also why we don’t think mixing languages is necessarily something that needs to be stopped immediately. If a child says one Spanish word in an otherwise English sentence, they still retrieved and used a Spanish word. If they answer a Mandarin question with two Mandarin words and then switch to English, they still produced something they might not have produced yesterday.
The perfect sentence can come later.
For parents, this can mean looking for slightly different signs of progress. Instead of only asking, “Can my child speak Spanish?”, it can be useful to notice smaller things. Are Spanish words beginning to appear without being requested? Is your child answering with two words where they used to answer with one? Are they willing to try even when they don’t know the whole sentence? Do familiar phrases come out a little faster now?
And most importantly: are there places in their life where the language is something they actually use?
That might be talking to a grandparent, playing with another child, ordering something at a restaurant, telling a story, playing pretend, or talking with Phoenyx. The specific activity matters less than the fact that there is something on the other side of the language that the child actually wants.
So when a child understands everything and answers in English, it doesn’t necessarily mean all that exposure has failed.
There may simply be a much bigger difference between knowing enough to understand and having enough practice to speak than it appears from the outside.
And sometimes the first sign that the gap is closing isn’t a beautiful sentence.
It’s one word they didn’t have to say in Spanish, but did.