I want to avoid using {du'u} for the following reason.
a formulation like
> citkrfa: x1 is that which proposition x2 claims eats. ......F1
makes sense, but is not suitable for the interpretation of
fa ko'a citka ko'e === fi'o citkrfa ko'a citka ko'e ......S1
which is derived from what la_tsani stated.
F1 brings an ambiguity of interpretation that is shared with {fi'o citka be ko'e ko'a} or any BAI/{fi'o fe'u}-structure.
Formulation with {nu} or any other cmavo of NU that does not take {ce'u} as an argument will produce the same ambiguity.
sa'unai Accoding to F1, a statement
ko'a citkrfa lo du'u ko'a citka ko'e ......S2
fixes the proposition {ko'a citka ko'e}:
referents of {ko'a} and {ko'e} are fixed respectively.
Then, a statement {fi'o citkrfa ko'a citka ko'e} does not necessarily signify the same proposition as {ko'a citka ko'e} in S2.
The former signifies a proposition that ko'a who eats ko'e is involved in a proposition that zo'e eats ko'e.
An interpretation of ko'a!=zo'e makes sense when a tapeworm eats things eaten by the host, for example.
This ambiguity of interpretation comes from fixing the proposition in x2 of
{citkrfa}.
In order to make S1 always true, x2 of {citkrfa} should not be a proposition but an open sentence, which leaves one place be free for use in any other statement, and fixes referents of the other arguments to the same as the proposition intended.
Then, when {fi'o citkrfa ko'a} appears in a statement, we can have a consistent interpretation that {ko'a} occupies the free place of the open sentence, and this occupation brings a proposition intended.
The reasonable English translation of definition of {brodrfV} that satisfies
x1 brodrfV lo ka fV ce'u broda <=> broda fV x1
would be:
x1 brings a proposition by satisfying a formula stated in {ka}-clause.
mu'o