Registerial hybridity in an Early Byzantine official archive : making a case for historical diaglossia?