Grammar comes second

Nobody has ever explained a rule of grammar to a two-year-old, and every two-year-old is learning grammar. Michael Tomasello's Constructing a Language is the book that lays out how, from decades of recordings of children talking.

The order is the interesting part. Children do not start with categories like noun and verb. They start with whole utterances they have heard, used as one piece: "gimme that", "where's the ball".

Then come patterns organized around one word, with a slot in them: "more juice", "more milk", and only much later an abstract "more X" that will take anything. The rule emerges last, by analogy across hundreds of examples, and the child could not state it if you asked. Tomasello's name for this is a usage-based theory: the grammar is built out of the uses.

The same thing is visible one level down, in the sounds. Kuhl (2004) reviews how infants learn which sounds their language uses from the speech around them, by statistical learning and by interaction, with no instruction at any stage.

The two skills Tomasello says the whole edifice rests on are reading what somebody means and finding patterns. Neither is a grammar lesson.

This is the reverse of most language courses, which give you the rule first and hope you will find the sentences later. I am not against rules.

Adults can use them, and a rule you have read is a real fact about the language. But a rule you have read is a fact, and a rule you have met two hundred times in sentences is an instinct, and only the second one is there when you open your mouth.

Which is why the way in is stories. A story is hundreds of sentences in the shapes the language actually takes, and the shapes recur because the story is about something.

Read enough of them and "more X" builds itself. Aloglot reads them to you, with the words lit as they are spoken and the English under every sentence, and when you do want the rule, tapping a word explains why it takes the shape it does, right there in the sentence where you met it.

Sources

• Tomasello (2003). Constructing a Language: A Usage-Based Theory of Language Acquisition. Harvard University Press.
• Kuhl (2004). Early language acquisition: cracking the speech code. Nature Reviews Neuroscience, 5, 831–843.