When to introduce a second language

The honest answer to "when should we start" is that the question contains a hidden assumption worth examining first: that there is a window, and that missing it costs something permanent.

There is an age effect. It is real, it is measurable, and it is smaller and less absolute than the way it usually gets repeated.

Two paths, not a deadline

The Canadian Paediatric Society divides early bilingualism into two kinds.

Simultaneous bilingualism is learning two languages from birth, or before the age of three. Both languages develop together from the beginning.

Sequential bilingualism is learning a second language after age three, once the first is established.

That age three line is a description of two different paths. It is not a gate that shuts. A child who starts at four is on the sequential path, which is an ordinary and well-described route rather than a consolation prize.

What the simultaneous path looks like

Children learning two languages from the start reach early communication milestones at similar times to children learning one. The Canadian Paediatric Society is explicit that exposure to two languages does not cause a language delay.

What parents notice is that vocabulary in each individual language may look smaller than a monolingual peer's at a given moment, because the child's words are distributed across two systems. Counting across both languages gives a fairer picture.

The other thing parents notice is code-mixing, covered on does bilingualism delay speech?.

What the sequential path looks like

This one has a shape worth knowing in advance, because two of its stages are commonly mistaken for problems.

1. A silent period. The Canadian Paediatric Society describes a silent period lasting roughly three to six months during sequential bilingualism, when a child stops using their first language in the new environment while still engaging socially. The child is listening and participating without producing much of the new language.
2. Fast early progress. Early language skills develop within six months to two years of exposure.
3. A much longer road to full competency. It may take five to seven years to reach oral language competency similar to a monolingual child of the same age.

Those last two figures are the pair that prevents most unnecessary worry. A child doing well after a year is doing well. That child is also still years from finished, and that is normal rather than a sign something went wrong.

What the critical period research actually says

This is where popular summaries drift furthest from the evidence.

A large study of English speakers published in the journal Cognition found evidence of a critical period for second language acquisition, with an advantage for learners who begin well before adolescence.

Read carefully, that is a statement about likelihood of native-like proficiency across large populations. It is not a claim that learning becomes hard after a certain birthday, and it is not a prediction about your particular child. Adults learn second languages to a high standard routinely.

So earlier is generally easier, particularly for accent and for the intuitive feel of grammar. Later is not a closed door.

The variable that matters more than age

Across the sources, one thing comes up more often than age: the amount and richness of exposure.

The Canadian Paediatric Society's recommendations are about frequency and variety. Speak the language consistently. Read in it. Tell stories in it. Play in it. Sing in it. Provide exposure through varied contexts.

That is a useful corrective, because age is fixed and exposure is not. A four year old hearing a language across many hours and many kinds of activity is in a better position than a two year old hearing it for an hour a week.

It also explains why an immersion setting works differently from a class. In immersion the language carries the actual business of the day: what is for lunch, whose turn it is, where the shoes go. In a class the language is the subject rather than the medium. Immersion versus language classes works through that difference.

What a licensed setting is already required to do

Worth knowing, because it applies to every licensed centre in British Columbia whether or not it offers a second language.

Schedule G of the Child Care Licensing Regulation requires a program of activities that encourages language development, including modelling of good language and listening skills, opportunities to develop receptive and expressive language, and activities that encourage communication.

The same schedule requires an emotional development program that provides an atmosphere in which children can feel proud of their cultural heritage and in which cultural sharing is encouraged.

Those are minimum standards rather than a language program. They are a reasonable thing to ask about on a tour, in any centre.

If your child already seems behind

Do not use a delay as a reason to drop a language. The evidence does not support that move.

The Canadian Paediatric Society states that children with a speech or language disorder will not become more delayed by hearing or learning two languages. Research summarised by the American Speech-Language-Hearing Association found preschoolers with language impairment in bilingual immersion did as well as peers in monolingual programs, and that maintaining the home language is best clinical practice.

If you are worried about speech, the useful step is an assessment by a speech-language pathologist, ideally one who can assess across both languages. Dropping a language removes something valuable without addressing the concern.

What the pressure to start early gets wrong

The critical period finding gets repeated in a way that produces anxiety rather than action, so it is worth naming the failure mode.

A parent hears that earlier is better, concludes they are already behind, and either rushes into an arrangement that does not suit the family or gives up on the idea entirely. Both responses come from treating a gradual population-level effect as a personal deadline.

The finding is that the likelihood of native-like proficiency declines gradually with age of first exposure. It is not that a door closes. Children starting at four, five or eight learn second languages, and the sequential path the Canadian Paediatric Society describes is the ordinary route by which most of the world's bilingual people got there.

Useful response: start when you can, with an arrangement you can keep. Unhelpful response: panic about a birthday.

What changes, and what does not, with age

	Starting before three	Starting after three
Name used by the Canadian Paediatric Society	Simultaneous bilingualism	Sequential bilingualism
Relationship to the first language	Both develop together	Second builds on an established first
Silent period	Not described as a stage	Roughly three to six months
Early skills appear	Alongside first-language milestones	Within six months to two years of exposure
Time to competency similar to a monolingual peer	Not framed this way	May take five to seven years
Likelihood of native-like proficiency	Higher, per the critical period research	Still high, and declining gradually with age

What does not change across that table is more important than what does. Exposure has to be frequent and varied in both columns. The home language needs protecting in both columns. And in neither column does a second language cause a delay.

Two things that get confused with age

Accent and competence are not the same thing. The critical period research is strongest on native-like pronunciation. A person can speak a language precisely, fluently and with a large vocabulary while carrying an accent. Treating an accent as the measure of success sets a standard that has little to do with whether someone can use a language well.

Speed of early progress is not the same as depth. Young children pick up conversational fluency in a way that looks miraculous to adults, and it can hide how far there is still to go. The Canadian Paediatric Society's five to seven year figure exists precisely because the early fluency is real and partial at the same time.

Both confusions push families toward the same mistake, which is judging the project after a year and concluding either that it is finished or that it has failed.

A practical way to decide

If your child is under three and you have a second language in the household or in reach, use it. The simultaneous path is available and costs nothing.

If your child is three to five, the sequential path is well described and works. Prioritise quantity and variety of exposure, and expect the shape described above, including the silent period.

Whatever the age, protect the home language. That is the thing most easily lost and hardest to rebuild, and language strategies at home covers how.

Next, does bilingualism delay speech?.