The headword is future optative, third person singular, of a(marta/nw (cf. alpha 1493, alpha 1494). This form is used by George of Pisidia (de expeditione Persica 1.137-8): after mentioning "the second Pharoah" he appends - "if one might not err in calling second him who truly was first into error" (ei) deu/tero/n tis ou)x a(marth/soi le/gwn | to\n w(s a)lhqw=s prw=ton ei)s a(marti/an). In classical Greek the future optative would not have been used in the protasis of a conditional statement unless it was a subordinate clause in indirect discourse (Smyth §2287, web address 1); e)a/n with the subjunctive would be a normal construction. The lexicographer seems to be criticizing George's hyper-Attic grammar, but unfortunately the form he supplies instead is future indicative rather than aorist subjunctive (which would be a(marth/sh|). Perhaps this is merely a textual corruption. In any case, all three forms would have been pronounced the same in Byzantine times.