ke'a/ce'u subscripting

["And Rosta" <a.rosta@lycos.co.uk>]

There is an ambiguity in

ko'a poi ko'e poi ke'a

as to whether ke'a is coindexed with ko'a or ke'a, and
the same ambiguity with

du'u du'u ce'u
ka ka ce'u

IIRC, the rule is that ke'a and ce'u belong to the localmost
candidate host bridi (i.e. to the innermost poi/du'u/ka). To
get them to attach to an outer bridi, one option is to use
a prenex and goi:

ko'a poi ke'a goi ko'i zo'u ko'e poi ko'i

du'u ce'u goi ko'i zo'u du'u ko'i
ka ce'u goi ko'i zo'u ka ko'i

But I also recall that there also exists a device whereby
scope can be shown by xi-subscripting. Can anybody explain
to me how it works?

--And.