Your child mixing Hindi and English in one sentence is not confusion. It is a grammar rule being followed.

A four-year-old in Pune says "Mummy, main park jaaunga and then we'll get ice cream" and a grandparent worries that the child has mastered neither language. A teacher in Chennai suggests that the family "stick to one language at home so he doesn't get confused". A parent quietly stops speaking Bhojpuri to a toddler, on the theory that it will crowd out the English the child will be tested in.

All three responses are common, well-meant, and unsupported by the research. India is one of the most multilingual societies on earth — the 2011 Census counted 121 languages and 22 scheduled ones, from nearly twenty thousand raw mother-tongue returns — and multilingual childhood is the norm here, not an experiment. It is worth knowing what the evidence actually says about it.

Two languages do not delay a child

The most persistent worry is that a second language slows the first. It does not. Bilingual children hit the standard milestones on the standard schedule: babbling, first words around a year, two-word combinations somewhere in the second year. A child who is genuinely late to talk is late for reasons that have nothing to do with how many languages are in the house — and the mistake that does real harm is attributing a genuine speech or hearing problem to bilingualism and waiting, instead of getting it assessed.

What is true, and gets misread as delay, is how vocabulary is distributed. A bilingual four-year-old will typically know fewer words in each language than a monolingual peer, because their waking hours are divided between two systems. Count the concepts the child can name in either language and the total is comparable. Assess only in the school language and the same child looks behind; assess across both and they do not. This distinction matters enormously in a country where a child is routinely tested in a language they hear for six hours a day and not the one they think in.

Code-switching is a skill, not a lapse

Mixing languages within a sentence looks like the strongest evidence of confusion. It is close to the opposite.

Code-switching is rule-governed. Speakers switch at grammatically permissible points and not at arbitrary ones — the sentence stays well-formed in both systems, which is why "main park jaaunga and then we'll get ice cream" sounds natural and a randomly chopped version of it does not. Doing that requires holding both grammars active and knowing where they can be joined. Young children switch for the ordinary reasons adults do: a word is more available, or more precise, or more appropriate to the person being addressed.

Bilingual children also work out very early who speaks what. Long before they can explain it, they will address one grandparent in one language and a neighbour in another, and adjust when a monolingual adult joins the room. That is social calibration, not muddle.

The one thing worth watching is not mixing but input volume. Children acquire a language roughly in proportion to how much interactive, responsive exposure they get in it. A language spoken to a child for a few minutes a day will produce understanding without speech — the very common Indian outcome where a child follows the grandparents perfectly and answers in English.

The question that predicts whether a child will speak a language is not whether the household keeps it separate from other languages. It is how many hours a week somebody actually talks to the child in it.

What the research does not settle

Popular writing on bilingualism leans heavily on the claim that it improves executive function — attention, task-switching, inhibiting distractions — and delays dementia by several years. This is the part to be careful about.

Early studies did report such advantages. Larger and preregistered replications have often failed to find them, and analyses of the literature have shown that studies reporting an advantage were more likely to be published than those finding nothing. The current honest position is that a general cognitive advantage from bilingualism is contested and, if it exists at all, is smaller and more situation-dependent than the popular version suggests. The dementia findings are observational and hard to separate from education, migration and lifestyle.

None of this is an argument against raising a child multilingually. It is an argument for the right reasons: access to family, literature, community and employment; the ability to be a full participant in more than one world. Those benefits are not contested and do not depend on a disputed effect size.

The gap that catches schools out

There is one finding from this literature that changes how a teacher should read a classroom, and it is not widely known outside education research.

Conversational fluency in a new language arrives fast — a child moved into an English-medium school will usually sound comfortable in playground English within about two years. The language needed for academic work is a different thing: decontextualised, abstract, dense with low-frequency vocabulary and complex sentence structure, and it takes far longer, commonly five to seven years, to reach the level of a peer who has been in that language all along.

The practical consequence is a child who sounds fluent and cannot follow a science chapter, and who is then judged as weak at science. This is one of the most consistent misdiagnoses in multilingual education. It is also part of the reasoning behind mother-tongue and home-language instruction in the early years — the position taken by NEP 2020, which recommends the home or local language as medium of instruction at least until Grade 5 — since concepts learned in a language a child already commands transfer to the second language later, while concepts never understood in the first place do not.

How a family language disappears

Left alone, a minority or heritage language tends to be lost within three generations: the grandparents speak it, the parents understand it, the children do not use it. The mechanism is unremarkable. Once the school and the peer group run in another language, the home language becomes the one with the least airtime, and children are ruthless about dropping what is not useful to them.

Families that keep a language tend to do the same few things: an adult who consistently uses it with the child, other speakers of the child's own age, a reason to use it that is not obligation — cousins, a grandparent who genuinely does not speak the other language, songs, stories, food, festivals — and reading material, which is where most Indian heritage languages struggle for want of published children's books.

Two things do not work. Screens do not teach language. Infants learn speech sounds from a live person who responds to them and largely fail to learn the same sounds from video of the same person, which is a striking result and a useful one for parents. And passive exposure is not input: a language spoken around a child rather than to them produces comprehension at best.

Why it matters for students and researchers

India is, from a linguistics standpoint, an extraordinary natural laboratory that is thinly studied. Most of the influential research on bilingualism comes from settings with two languages, one clearly dominant, and a migrant framing. That is not the Indian situation, where three or four languages routinely coexist without any of them being foreign, where the school language may be nobody's mother tongue, and where the sociolinguistic status of each language shifts by district.

The open questions are practical and local: how do children manage three or more languages simultaneously rather than two; what does code-switching look like in a stable multilingual community rather than an immigrant one; how should language assessment be designed so that it does not systematically underestimate multilingual children; and what actually happens to learning outcomes when the medium of instruction changes at Grade 5, which is now policy for a very large number of schools.

Those are exactly the emerging linguistic phenomena, methodologies and teaching frameworks that Emerging Trends in Languages (ISSN 3049-0693), a peer-reviewed journal launched in 2024, exists to publish. For students of linguistics, education and psychology, it is a field where a carefully designed study in one Indian district can say something that the international literature genuinely does not already know.

Frequently asked questions

Does learning two languages delay a child's speech?

No. Bilingual children reach babbling, first words and two-word combinations on the same schedule as monolingual children. A genuine delay has other causes and should be assessed rather than attributed to bilingualism.

Is it bad if my child mixes two languages in one sentence?

No. Code-switching follows grammatical rules and requires competence in both languages. Children also learn very early to match language to listener, which is a social skill rather than a sign of confusion.

Should we speak only one language at home?

There is no evidence that separating languages strictly is necessary. What predicts whether a child will speak a language is the amount of interactive, responsive exposure they get in it, not whether it is kept apart from other languages.

Does bilingualism make children smarter?

The claimed advantages in attention and task-switching are contested — early positive findings have often failed to replicate in larger studies. The reliable benefits of bilingualism are access to family, community, literature and opportunity, which do not depend on a cognitive effect.

Why does my child sound fluent in English but struggle with schoolwork?

Conversational fluency typically develops within about two years, while the academic language needed for textbooks and exams takes far longer — commonly five to seven years. A child can sound comfortable in a language and still lack the vocabulary and sentence structures that academic content requires.

Can screens help a child learn a language?

Very little. Infants learn speech sounds from a live person who responds to them and largely fail to learn from video of the same person. Language is acquired through interaction, not exposure.