Calibrate. Seed. Train. Deploy.
Training Minds | Essay 5
The book I am writing is organized around four words. I want to explain what they mean, because they are doing more work than they appear to be.
Calibrate.
Before any learning can happen across a flat surface — a screen, a book, a photograph, a flashcard — the child has to understand that flat surfaces reliably represent three-dimensional reality. This understanding is not automatic. It is built. And until it is built, the entire two-dimensional layer of the world — every picture book, every learning app, every foam letter in the bath — is producing considerably less learning than the parent thinks it is.
This is what the research on the video deficit is actually telling us — not that screens are bad, but that there is a representational bridge between flat images and three-dimensional reality that has to be constructed before two-dimensional input becomes fully usable. [Essay 2 covers this in detail.] Calibrate is the phase where you build that bridge — deliberately, starting at twelve months, using tools that are free and always available. The payoff runs across every domain that follows.
Seed.
Seed is the phase before a child is ready to learn something — the phase where the concept is being introduced ambiently, before any active training begins. Foam letters in the bath. Animal figurines on the kitchen floor. A color sorting tray on the shelf. The child is not being taught. You are stocking the environment.
Seed is running on all ten domains simultaneously, all the time. You are not choosing what to teach. You are putting signal in the environments your child already occupies — the bath, the changing table, the high chair, the floor — and waiting to see what the brain reaches for. The parent’s job in Seed is to stock shelves and step back.
Train.
Train begins when the child tells you it begins. Not when the calendar says. Not when the parenting book suggests. When the child picks up the object and brings it to you.
That gesture — the foam T held up in the bath, the figurine carried across the room and presented, the book demanded for the fifteenth time in a week — is not a preference. It is a readiness signal. The brain has accumulated enough associative density around a concept that it is now reaching for the label. That is the announcement that Train can begin.
The goal of Train is a self-running loop: a tool the child can operate independently, that labels concepts automatically, closes its own feedback cycles, and runs at the child’s pace without requiring the parent to be present. Not because the parent is unimportant — but because the loop that runs fifty times while you are on a work call is producing more learning than the lesson you deliver in twenty minutes of focused attention.
Deploy.
Deploy is not a phase you administer. It is a phase you recognize.
Your child encounters the concept in a novel context — not in the training environment, not with any of the usual cues — and identifies it correctly. The letter on the parking lot asphalt. The animal at the zoo that matches the figurine on the kitchen floor. The color named for the first time without anyone prompting. That is Deploy. That is generalization. That is the whole point.
These four phases are not a schedule. They are a map of what the brain is doing at each stage of learning a new concept. The parent who understands the map knows what to do and what to stop doing at every phase. The parent who does not understand it spends a lot of time wondering why the toy they bought isn’t working — when the toy is fine, and the phase is wrong.
The map does not tell you which domain to teach first, or when. That is what the Interest Index is for.
Training Minds is a Substack about categorical learning, the Signal Stack, and what the research actually says about how children ages 12–36 months build knowledge. Written for analytically minded parents who want frameworks and evidence, not parenting philosophy.
Next essay: what the Interest Index is, why it works, and what to do when your child shows no interest in a domain you expected them to reach for.
— Sandra