How to raise a bilingual baby (ages 0 to 2)

For a baby, the sources read for this site point to live, shared activity: you talking, singing and reading with your baby in the heritage language, in person. In The Bilingual Edge, King and Mackey conclude that joint adult-child activities do much more for a baby's language learning than baby videos of any sort. No source read here gives a number of minutes that a baby needs, so treat the plan below as a starting point.

Why a person and not a screen

King and Mackey's own conclusion for babies and toddlers is that joint adult-child activities do much more to promote language learning than baby videos of any sort, and that the idea of videos teaching infants language is not supported by research. This is their stated position.

Their main evidence is a study by Kuhl and colleagues, as the book describes it. Nine-month-old babies watched high-quality videos of a native Chinese speaker. They learned no Chinese sounds, and did no better than babies with no Chinese exposure at all. Only the group with a live, interactive speaker learned anything. Three limits apply. This is the book's account, not the original paper. It is one study. And it measured sounds at nine months, not words. It compared video with a live person; it did not test live video calls.

King and Mackey also quote the American Academy of Pediatrics as recommending no television at all under age two. They source it to aap.org, and the book is a 2006-era text, so the current wording was not checked for this site. Your pediatrician is the right person to ask about screen time.

What to do in an ordinary day

Everything below is done with a person present. The app's plan for this age rests on three activities.

• Talk through routines. Narrate meals, baths, dressing and walks in the heritage language. No lesson is needed. In the app, talk gets the largest share of minutes at this age.
• Sing short songs and rhymes. The app's content for this age suggests short, repetitive songs with simple actions such as clapping or rocking, and says repetition matters more than variety. That is app guidance with no cited source (UNVERIFIED).
• Read board books, and re-read them. One object per page works, and the same book again is fine. In a study King and Mackey describe (Patterson, 2002), how often 64 two-year-olds were read to in each of their two languages predicted their vocabulary in that language, even more than total exposure. Those children were two, so the finding does not directly cover younger babies.

What the plan generator gives a baby

ILLUSTRATION from the app's real code, not data about real families. I ran the plan generator for a 1-year-old with Spanish at home, English in the community, one Spanish-speaking parent, 30 minutes a day, a conversational goal and some existing exposure. It returned 17 minutes of talk, 8 of songs, 3 of reading and 2 of a daily ritual. These add up to 30. Structured activities such as workbook pages get no time at this age.

Two cautions. First, how the app splits minutes by age is its own design choice, with no cited source (UNVERIFIED). It gives reading the smallest share for babies, and you do not have to follow that. Second, 30 minutes is a deliberate block, not the whole day. The app assumes a baby is awake 11 hours a day, which is also an unsourced estimate, and 30 of those 660 minutes is about 4.5%. King and Mackey's 80% target for the early years concerns home interaction overall, and the Family Language Audit counts every caregiver's hours.

What is still unknown

• How much live heritage-language time a baby needs. The 80% target is for the early years, and the sources read give no figure specific to babies. The 20% figure is CONTESTED and has its own article.
• Whether a live video call with a relative helps a baby. The Kuhl study, as reported, compared video with a live person. The sources read do not test calls.
• Whether the Kuhl result holds for other ages, sounds or languages. The sources read contain only that one study, secondhand.