In Lectures in Logic and Set Theory(Vol.2) by George Tourlakis, there is a metatheorem and its proof (fraction) as follows:
What confuses me is why the proof consider the cases that A[e1, . . . , en] is a nonlogical axiom, since "we did not add any new nonlogical axioms" and nonlogical axioms in the original language cannot contain constant symbols.
Any help is appreciated!

