A quotation from the preface to Donald Knuth’s The TeXBook (Amazon, archive.org):

[A] noteworthy characteristic of this manual is that it doesn’t always tell the truth. When certain concepts of TeX are introduced informally, general rules will be stated; afterwards you will find that the rules aren’t strictly true. […] The author feels that this technique of deliberate lying will actually make it easier for you to learn the ideas. Once you understand a simple but false rule, it will not be hard to supplement that rule with its exceptions.

This sentiment has stuck with me. I recently found its distant ancestor in Dante Alighieri’s Quaestio de aquae et terrae (c. 1320):

aliae duae [figurae] superiores falsae; et positae sunt, non quia sic sit, sed ut sentiat discens, ut ille dicit in primo Priorum.

[“The two preceding figures are false; and they are introduced, not because they are correct, but that the learner may understand, as he says in the first of the Priors.”]

This in turn refers to Aristotle’s Prior Analytics (I.xli), in which the philosopher writes:

We do not base our argument upon the reality of a particular example; we are doing the same as the geometrician who says that such-and-such a one-foot line or straight line or line without breadth exists when it does not, yet does not use his illustrations in the sense that he argues from them. […] On the contrary, we (I mean the student) use the setting out of terms as one uses sense-perception; we do not use them as though demonstration were impossible without these illustrations.

Now, Aristotle was simply saying that “examples are non-normative”: if I provide an illustration in support of a principle, and you dispute that particular illustration, the principle itself may yet be correct. (This differs from logic, where if you knock down one of my premises, my conclusion goes with it.)

Dante wants to present a particular concrete theory about the distribution of water on the earth. He does this by first presenting two competing hypotheses, neither of which is correct, solely for the purpose of knocking them down, before he reveals his own theory. Seeing the holes in the earlier versions helps the student understand the final version.

Knuth also wants to present a particular concrete system. He does this by first presenting an oversimplified version of the system — so oversimplified that it is, in fact, incorrect — to give the student the general gist of the system. Seeing a consistent general plan up front (even though it errs in some particulars) helps the student understand the final version.

In this Knuth differs from Einstein, who famously said, “Everything should be made as simple as possible, but no simpler.” (Well, sort of. In On the Method of Theoretical Physics (1933) he wrote: “[T]he supreme goal of all theory is to make the irreducible basic elements as simple and as few as possible without having to surrender the adequate representation of a single datum of experience.”) Knuth, on the other hand, says that you should make things simpler than possible, at least at first, because that’s the best way to plant the seed of understanding.

To put Knuth’s approach in the form of a grook:

When your topic is arcane
and on it you’d shed light,
First make it wronger ’til it’s plain,
Then weirder ’til it’s right.