And, no, there is no connection between “king” and “angel”. The word for angel (properly, ‘messenger’) is etymologically related to הלך, meaning “go, to go”. The actual root is understood to be לאך, which turns up most prominently in Ethiopic where it refers to sending a messenger (in the active) or ministering to somebody (in the passive). The Hebrew retains the aleph, keeping the word as מלאך but with no connection to the kingly מלך.

Of course, I can’t speak for what you refer to as folk-etymology and, in that regard, I’ve heard all manner of preposterous connections raised between completely different and unrelated words.

Another Hebrew example: the Aramaic abstract noun for kingship is מלכו (malkhu) which takes the feminine definate article מלכותא (malkhuta). In Hebrew, this was (mis)understood to be a masculine definate article and the absolute form was thus taken to be מלכות (malkhut). All abstract Hebrew nouns (I think) end in ות- and are feminine; pluralising as ויות-.

The assimilation of word forms happens in English too; for instance ‘penthouse’, which is an anglicization of ‘appendix’, now pluralised ‘penthouses’, as if from ‘houses’. Pluralisations are often lost too, just as with Hebrew: cherry was originally “cherries” as a singular word (Fr. cerise), but taken as plural; same with pea from pease (Fr. pois).

I certainly find it funny that a speaker of Modern Hebrew, in which nouns are pluralised by adding the suffix “-im”, might go to a clothing store to look at “jinzim”, or to a cake shop and ask for just one “browniz”.

There are scripts which have been used only to write a single language, like Cherokee, but in no case is this a necessary feature of the script, merely a historical contingency.

It’s also not really true, talking of necessity and contingency, that older always means more complex; in the Finno-Ugric languages, having lots of cases (15-25) is a fairly recent phenomenon. In the reconstructed versions, the case endings are separate postpositions, and only in the last millennium or so have they come to be irrevocably glued to the noun stems as a result of morphophonemic changes.

Maori verbs are another fascinating case. In Proto-Polynesian, the passive was formed by adding the invariable ending -ia to the final consonant of the active. However, this final consonant has now been lost in all active forms, leaving us (from a synchronic viewpoint) with a random and irregular lot of passive endings, -tia, -kia, -ria, and so on, each of which must be learned individually like so many Latin or Greek declensions. (The -tia ending is the “regular” one, in the sense that it is applied “by rule” to borrowed or otherwise novel verbs, and even to some verbs that once took other endings; therefore, it cannot be that the final consonant is still underlyingly present in the active, for there can hardly be an underlying /t/ in all foreign verbs!)