It is probable, however, that some or all of them, possibly in varying proportion depending on the region of the Nabataean Kingdom where they lived, spoke Arabic.
Nabataean א in دوسرا (dwsrʾ) does not signal [aː]; it would seem that *ay# collapsed to something like [æː].
Scribes must have felt that this sound was closer to א when the spelling conventions of Nabataean were fixed.
In Greek transcription, this sound was felt to be closer to an e-class vowel, yielding Δουσαρης.
The reconstructed text of the inscription is as follows:[5] In JSNab 17, All Arabic triptotes terminate in w regardless of their syntactic position or whether they are defined.