A four year old sharing biscuits between three friends is doing something more demanding than it looks. She is holding a quantity in mind, tracking who has been served, noticing when the shares come out uneven, and adjusting. No numerals appear anywhere. No worksheet exists. This is early mathematics, and it offers a clearer window into how her thinking is developing than almost anything else she will do that day.

Most parents recognise math only when it arrives in a textbook with a syllabus attached. The years before that get filed under preparation, or under play. The evidence points the other way. The informal mathematical thinking children do between the ages of three and seven is not a warm-up for the real thing. It is where a good deal of the underlying cognitive architecture is laid down.

Early math is not counting

The first misunderstanding is the most common one. A child who can recite numbers to a hundred is performing a memory feat, closer to singing a song than to doing arithmetic. Recitation is not number sense.

Number sense is a cluster of quieter abilities. Recognising three objects as three without counting them. Knowing the last number said while counting answers how many. Judging that one pile is larger at a glance. Matching one object to one word without skipping. Noticing a pattern repeats and predicting what follows.

A child can chant to a hundred and still be unsure whether seven is more than four without stopping to count. Another may know almost no number words and have an excellent feel for quantity. The second child is better placed, and parents routinely praise the first.

Spatial reasoning belongs here too and is the one most often left out. Picturing how objects fit, rotate and relate is a strong early signal, and it is trainable through nothing more exotic than blocks, puzzles and being talked to about position.

Why math loads the mind harder than most tasks

The link between early math and broader thinking runs through executive function, the set of mental controls that lets a person hold information in mind, resist an obvious wrong move, and change approach when the first one fails.

Watch a child solve a two step problem. They hold the first quantity in working memory while reading on for the second. They resist the operation the familiar keyword suggests, because the problem is built to punish that reflex. They notice partway through that the method is failing and switch without abandoning what is done.

That is working memory, inhibitory control and cognitive flexibility running at once. Reading comprehension calls on the same machinery, as does planning a task and holding back an impulse. Few childhood tasks demand all three at once, and fewer under a rule where nearly right counts as wrong.

That unforgiving quality is the source of both the difficulty and the value. Math gives immediate, unambiguous feedback about whether reasoning holds together. Little else in a young child’s life does.

Fluency is not the enemy of understanding

There is a long standing argument between drilling facts and teaching concepts, and it is largely a false choice.

The reason is capacity. A child sounding out every word cannot follow the story, because decoding has consumed the room the plot needed. The same happens in math. A child laboriously working out eight times seven has little left to hold the shape of the problem that required it.

Automatic recall of basic facts is not the opposite of understanding. It is what makes understanding affordable. The failure is stopping there, drilling fast recall of procedures a child cannot explain. The reverse failure is real too: a child who can discuss why a method works but cannot execute it quickly enough to reach the interesting part.

An honest caveat

The relationship runs in both directions. Children with stronger executive function find math easier, which leads to more practice and more success, which strengthens those controls further. Untangling which came first is hard, and household circumstances influence both.

The research supports a robust association and a strong case for early attention. It does not support the claim that math instruction manufactures intelligence. Any parent sold that promise should discount it.

Where it goes wrong

Math anxiety appears far earlier than most parents expect, well inside primary school, and it is cognitively expensive rather than merely unpleasant. Worry occupies working memory, the same narrow resource the problem needs. An anxious child is not simply unhappy while calculating. They are calculating with less capacity than a calm child of identical ability.

Then the loop closes. Difficulty produces anxiety, anxiety produces avoidance, and avoidance removes the practice that would have resolved the difficulty. By the time it reaches the report card, a child may have been quietly opting out for a year.

It is also catching. A parent who says “I was never a math person” in front of a child has handed over a licence to quit, and children accept it readily.

What helps at home

Ages three to six. Narrate quantity out loud during ordinary tasks. Compare amounts without counting them. Play board games with a numbered linear track. Use spatial words deliberately: behind, between, edge, corner, upside down.

Ages six to nine. Get basic facts to automatic, then stop drilling and start asking for the reasoning. Ask for an estimate before the calculation, which builds a sense of whether an answer is even plausible.

Ages nine and up. Protect the struggle. Stepping in to rescue a stuck child removes the exact moment where reasoning is being built. Wait longer than is comfortable. Mix problem types rather than doing twenty of the same kind.

At every age. Never announce your own math incompetence in front of your child.

What to watch for

Not marks. Marks lag the thing that matters by a year or more, and they are noisy. Watch for whether the child begins.

A child who reads a hard problem, does not flinch, and starts is telling you the architecture is holding. That habit, formed early around numbers, does not stay confined to numbers for very long.

Views expressed are personal

The author, Neelakantha Bhanu, is Founder & CEO, Bhanzu