How much heritage language is enough?

For the early years, King and Mackey in The Bilingual Edge recommend a target of roughly 80% of home interaction in the heritage language, which they call the "hot-house" approach. The reason is the mechanism: the majority language seeps in from everywhere else, so the home has to carry the other one. The app counts this as a share of the child's waking hours.

What the hot-house target is

Two languages are never on equal footing outside a mixed-language or heritage-language home. The community language arrives through the street, school, other children and screens, with no effort from anyone. The book's answer is to tilt the home hard the other way: about 80 percent of child interaction in the weaker language and 20 percent in the majority language.

Two notes on scope. The wording is about the early years, not a lifelong rule. And the book speaks of home interaction; the app measures heritage-language hours as a share of the child's waking hours. That measure is the app's choice. The audit says so under its 80% result: King and Mackey state the line as a share of the child's interaction, and hours out of waking hours is an approximation.

The book also speaks to a family where both parents speak the heritage language. Its advice there is to make the heritage language the everyday language between at least one parent and the child, without worrying that this will hold back the majority language. The majority language seeps in from everywhere else regardless, so it rarely needs protecting at home.

How to count it

The Family Language Audit takes the hours per week each caregiver spends with the child and the language that caregiver uses. A caregiver logged as heritage-only counts in full. A caregiver who switches between languages counts at half weight. A community-only caregiver counts as zero for the heritage language. The total is divided by the child's waking hours per week.

Those waking hours are a working estimate in the app's code, by age band, and not from a cited source (UNVERIFIED). For ages 3 to 5 it uses 11.5 hours a day, or 80.5 a week. Treat the result as a rough measure, not a grade. If the hours you log add up to more than the child's waking hours, the audit warns that some rows probably overlap; it does not change your numbers. With three or more languages it does not apply the 80% line as a verdict: it gives an even-share rule of thumb and a per-language breakdown.

An illustration, not data

ILLUSTRATION built from the app's arithmetic, not data about real families and not a recommendation for every family. Take a child aged 4. At 80.5 waking hours a week, 80% is 64.4 hours of heritage language. I ran three invented weeks through the audit code:

• Week A: 40 hours with heritage-only caregivers (for example a parent and a grandparent). Share of waking hours: about 50%. Short of 64.4 by 24.4 hours.
• Week B: the same 40 hours, plus 20 hours with a nanny who switches between languages (counted as 10). Total counted: 50 hours, about 62%.
• Week C: 50 heritage-only hours, plus the same 20 switching hours (counted as 10). Total counted: 60 hours, about 75%. Still below the line.

What this shows is the size of the target. Even week C falls short. The count is useful because it shows where the hours are, so you can decide what to change. The numbers in these weeks are invented for the arithmetic; only the way the audit counts them (full weight, half weight, zero) comes from the app. Your own week will look different, and that is the point of logging it.

What to do this week

1. Log one real week in the audit, including the caregivers you tend to forget.
2. Read your share against the 80% line. If you are below it, look for hours with a person who speaks only the heritage language.
3. Count reading. King and Mackey report a study in which how often children were read to in a language predicted vocabulary in that language, even beyond total exposure.

What is still unknown

• A lower line. Some studies associate roughly 20% exposure with a child being willing to speak a language. That figure is CONTESTED and has its own article.
• Whether the 80% target changes as the child gets older. The book frames it for the early years, and no source read for this site gives a later figure.