How children's drawing develops, and what the stages don't tell you

Children's drawing changes in a broadly recognisable way as they get older, and several people have written that progression down as a set of stages. The description below is the standard one. The section at the end on what the stages do not tell you is at least as important as the stages themselves, and if you are here because you are worried about a particular child's drawings, read that part first.

Scribbling, which is not failed drawing

The earliest marks are not attempts at depiction that have gone wrong. They are a physical activity that happens to leave a trace. A young child with a crayon is enjoying the movement of the arm, the resistance of the surface, and the discovery that moving your hand changes the world. The marks are a byproduct of that pleasure.

This matters practically. Adults tend to respond to scribbles by asking what they are, and at this stage there is no answer — the question introduces the idea that a drawing is supposed to be of something, before the child has any use for it.

Controlled scribbling

Over time the marks become deliberate rather than ballistic. The child starts to repeat shapes on purpose, and the arm's whole-shoulder swings give way to movements the wrist and fingers can steer. Circles, loops, arcs, radiating lines and rough enclosures start appearing repeatedly. Rhoda Kellogg, who assembled and analysed a very large collection of young children's drawings, catalogued these recurring forms and described how they combine into more elaborate configurations. Whether the specific inventory she proposed is the right way to carve it up is debated; that these basic forms recur very widely is not seriously in dispute.

Named scribbling, where the name arrives afterwards

Then something changes that has nothing to do with the marks: the child begins to name them. Crucially, the naming usually comes after the drawing, and the name can change while you watch — the same tangle of loops is a dog, then a car, then Grandma.

This is a genuinely significant moment, but not because the drawing improved. It is the point at which the child treats a mark as standing for something else. That is the beginning of symbolism, and it arrives in drawing at roughly the time it is arriving in play and language too.

Early figures and the tadpole person

The first widely recognised figure is the "tadpole" person: a circle, two lines descending from it for legs, often arms attached to the head as well, sometimes facial features, and no separate torso. Children in many places produce it, and it is stable enough that it has been studied on its own for decades.

Why the torso is missing is genuinely unsettled. Proposals have included the limits of planning a multi-part figure, the head and legs being the parts that matter most to a child, and simple economy of effort. There are experiments in which small changes to the task get children to produce a torso, which suggests it is at least partly about the demands of the drawing task rather than about what the child knows a person looks like. A well-documented phenomenon with no settled explanation.

Drawing what you know, not what you see

Once children are drawing recognisable subjects, they draw them as they understand them rather than as they appear from a viewpoint. The results are consistent and delightful:

• Transparent houses, with the rooms and the people inside visible through the walls, because the people are part of what a house is.
• A table drawn with all four legs splayed outward, because the table has four legs and hiding two of them would be a lie.
• Important things drawn larger than unimportant things, regardless of actual size.
• A baseline along the bottom of the page and a strip of sky along the top, with everything standing on the line — a scheme for organising space that does not depend on a viewpoint.
• Multiple moments in one picture, or one object shown from two angles at once.

This is not a deficiency of observation. It is a different and internally consistent solution to the problem of getting a three-dimensional world onto a flat page, and it prioritises what is true about the object over what is visible from one spot.

The shift toward appearance, and the crisis

Later, often somewhere in the middle years of childhood, the goal changes. The child starts wanting the drawing to look like the thing looks — overlap, occlusion, a horizon, some attempt at perspective, proportions checked against reality.

This is where a great many children stop drawing. The ambition to produce a convincing likeness arrives before the technique for producing one, so the child becomes their own harshest critic, judges their work against an adult standard, and concludes they "can't draw". Viktor Lowenfeld, whose stage scheme runs from early scribbling through to adolescence, described this later period as one of increasing self-criticism, and anyone who teaches this age group recognises it.

The practical response people who work with children tend to converge on is to change what gets praised. Comment on the decisions rather than the resemblance — what is happening in the picture, why that colour, how the space was arranged — and avoid the two responses that do the most damage: "what is it?" and drawing a corrected version next to theirs.

What the stage schemes do not tell you

All of the above is a description of a general tendency, and it is regularly over-read. Five limitations, stated plainly:

The ages are wide and overlapping. Any age range attached to these stages spans years, and children move back and forth. A child who has drawn detailed figures for months will produce a scribble because scribbling is fun. Age bands in stage schemes are averages over groups, and they say very little about an individual.

The schemes describe, they do not predict. They were built by looking at large numbers of drawings and noticing what recurs — a legitimate and useful thing to do. It does not license using a drawing to forecast anything about a particular child.

Much of the classic work rests on small Western samples. The foundational studies were largely conducted with children in Western, school-attending settings, and cross-cultural research shows real variation — in what gets drawn, in how much drawing children do at all, in how much adult modelling and instruction they receive, and in the conventions available to copy. Children with more exposure to drawing produce different work at the same age. Treating one culture's developmental curve as human nature is a mistake with a long history.

Drawings are unreliable as assessments of an individual. The practice of reading a child's personality, intelligence or emotional state out of their drawings has been studied a great deal and has not held up well. Drawings are affected by mood, by the pen, by what was drawn last week, by what the child thinks you want, and by how tired they are. Interpreting a specific child's drawing as evidence about that child is not something this page is going to encourage.

The endpoint is a convention, not a summit. Framing the whole sequence as a climb toward visual realism smuggles in the assumption that realism is what drawing is for. Plenty of drawing traditions are not aiming at it, and a child who never makes that turn has not stalled.

What this does not claim

Nothing on this page says drawing makes children cleverer, more creative, or better at concentrating. Those claims get attached to art activities constantly and the evidence for the strong versions of them is thin. Children draw a great deal when given the means, they clearly enjoy it, and that is a sufficient reason to hand them something to draw with.

The same honesty applies to software. An app is a surface, not a programme. If a tablet is what is available — in a car, at a restaurant, in a room where you do not want a wall covered in crayon — then Flying Chalk is a surface with no ads, no account and nothing to buy, which is a low bar that a surprising number of children's apps fail to clear. Finger drawing works without a stylus, and clearing the board is one gesture. It will not teach anybody anything, and it is not supposed to.

Related guides

• How to lay out a chalkboard poster that people actually read
• Chalk art from the classroom to the street
• Drawing on an iPad with your finger vs an Apple Pencil