Nils

Worth adding just how much of chemical naming is this kind of debris, because it reframes the whole subject for people learning it.

The systematic scheme is genuinely systematic once you are past the first few members, and it can name any structure unambiguously. But it was imposed on a field that had already been naming things for a century, and the most common substances were the ones named earliest and therefore the least systematically.

That produces exactly the frustrating pattern students meet: the compounds you encounter first have arbitrary names, and the rules you are being taught only start working on compounds you rarely see. The system looks unreliable when in fact you are being shown its exceptions first, in order of how common they are.

Wood, air, fat and butter is a genuinely delightful sequence. That is going straight into the next conversation I have about this.

There is a rule, and the short version is that the elision is specific to mono-, not general.

The naming conventions permit dropping the final vowel of a prefix when it meets the vowel starting the element name, but in practice this is applied to mono- before an o, and essentially nowhere else. So monoxide is standard, and so is the pattern in tetroxide, where the a of tetra- disappears.

Di- has no vowel to spare in the same way. Its i is the entire prefix apart from the d, so dropping it destroys the word rather than smoothing it. That is the structural difference between the two cases, and it is why no rule was ever written that would produce "diodine".

The other half of the answer is frequency. Carbon monoxide is a substance ordinary people talk about, so its name got worn smooth by use. Diiodine is a term almost nobody says out loud, so nothing wore it down.