English - AU
I'm saying it has less Bayesian probability. Quran uses it as jannat al-firdaws, which suggests the possibility that Muhammad or people around him weren't using the word in its secular meaning but as a proper noun. This doesn't strongly suggest it was borrowed in Arabic in this sense, but reduces the probability of the initial secular borrowing (which Persian only had).Why are we so confident that it couldn’t have come from Persian (in the secular sense)?
Again, it is the probability. If their common scripture was written in Aramaic, even if people knew Greek and used it for administrative or inscriptive purposes, it would have still been likelier for a person to borrow a religious term from Aramaic (because there was more intracommunal usage of this language on that topic).Also why do we assume only clergy would know Greek? There are Greco-Arabic and bilingual Greek-Arabic inscriptions and documents.