“…2 Our rules (11), (12) and (13) were suggested by the definition of • given just below Remark 13 in [14]. As in [14], therefore, there is a resemblance between our proof and the classic reducibility argument: we have, for example, that for S f g to be in SN, it is necessary that for all x ∈ SN, S f g x ∈ SN, which resembles the condition for reducibility in [13, §6.1].…”