Oh, I completely understand what you are intending to do, and by all means I support your intention, but (to give you some examples) it is useless for someone who doesn't know Arabic/Persian/Urdu to see "(عدد)" (the word in each of those languages for "number") in a list of English aliases for a specific number such as 162 (Q290532). Similarly it is pointless for a monolingual Czech, Portuguese, or Romanian speaker to see "first alphabet in english" in a list of aliases for those languages, or for monolingual Quechua, Russian, or Slovak speakers to see "lizard" in a list of aliases for those languages. It simply will not be as meaningful for them. The main problem I had, and which I was in part trying to rectify via reverts, was simply that you were putting aliases intended for different languages all in the list of English aliases, and you were putting English aliases in other languages' alias lists.
In your own interest in finding consensus, you are welcome to ask on Wikidata:Project chat whether lowercase letters should be aliases for their uppercase equivalents and await responses from the community there, but please do not just add them by yourself like that.