You do not understand it until you have taught it

What explaining a subject reveals about your grasp of it.

You do not understand it until you have taught it because using a skill correctly only proves that the hands know a sequence well enough to repeat it, while explaining that sequence to someone else, clearly enough that they can follow it without watching you do it first, proves it was understood and not simply memorised through enough repetition to stop needing conscious thought.

A shoelace tied in two seconds

Most adults can tie a shoelace in under two seconds without a conscious thought about which loop crosses which. Asked to describe the sequence out loud, clearly enough for someone who has never done it to produce a working knot, most of them stumble, contradict their previous sentence, or stop to watch their own hands mid-motion and narrate what they see. The ability to tie the lace thousands of times over a lifetime guarantees nothing about the ability to describe it, because one is stored as a practised motion in the hands and the other, if it exists at all, as an explicit sequence of steps in words.

The shoelace is a small, low-stakes version of the gap that appears when an experienced engineer is asked to explain a piece of judgement they use correctly every day.

Rules that work until someone asks why

Using a piece of knowledge only has to work in the situation in front of you, while explaining it has to work for a listener whose background and assumptions differ from yours and are not fully known in advance. A person who has absorbed a rule of thumb through repeated exposure, always constrain a sketch fully, always mate to a stable datum, can apply it perfectly well in familiar cases, because matching against past experience is enough there.

Asked why the rule holds, the same person must either reach for the underlying reason, which may never have been consciously assembled, or restate the rule in slightly different words, which an attentive listener will notice leaves the question unanswered. Explaining a routine procedure also turns up steps that were never consciously decided: the angle a tool is held at, the small resistance that signals a part is seated, the check performed automatically before moving on. None of it was written down, because none of it needed to be until someone else had to be told.

What writing a beginners' course exposed

The articles in this set each circle the same idea from a different angle: a CAD model is a chain of dependent decisions, a sketch's hidden freedom stays invisible until something moves it, a mate is a relationship between parts, a name carries information a stranger relies on before opening anything. None of these were obvious at first meeting. Writing a short course for people who had never touched CAD or simulation was the occasion for all of them, and the surprise lay in how many gaps the act of explaining opened in things that had felt solid beforehand.

Seen this way, teaching becomes one of the more reliable ways of finding out whether a subject has been mastered. Preparing to explain it, and working out which questions a newcomer will ask and whether a real answer exists for each, surfaces the parts running on habit far earlier and more cheaply than a failure would. A beginner's question that seems too simple to answer is often pointing at the one part of the explanation that was never built.

Fluent explanations can still be wrong

Teaching does not by itself guarantee correct understanding, since it is entirely possible to explain something fluently and confidently while explaining it wrong, and a listener with no independent check cannot tell the two apart. What teaching provides is exposure of a teacher's gaps to the teacher: the stumble, the reach for words that are not there, the question with no ready answer. Certifying that the material is correct still has to come from elsewhere, from testing the explanation against results, against a more experienced reviewer, or against the part working once it leaves the drawing.