krtisfranks asked me to come back to this. A few months ago, I claimed
that the rewrites in terms of "me da poi" do not work. Let me try again to
explain why.

Wuzzy wrote:
> > lo gerku cu poi'i ke'a jersi lo rebla be ke'a
> > "Dogs are such that they chase their own tail."
> With “me”, “da” and “poi”
> “lo gerku cu me da poi ke'a jersi lo rebla be ke'a”

The problem in your paraphrase has not much to do with poi'i, but with
the differences between da and lo, which are not equivalent. da is a
singular variable, and lo is a referring plural expression.

We can see how the kind of paraphrase you propose breaks even without
poi'i.

Let J be the (distributive) predicate "to chase one's own tail". Let G be
"to [be a] dog". Let <= denote the among relation.

lo gerku is the constant g.

J(g) would denote "lo gerku cu jersi lo rebla be vo'a".

If you replace jersi with "me da poi jersi", your logic changes to:

[∃x : J(x)] g <= x

which is false for example when g is plural, and is not the same kind of
statement anyway. It is an existential statement, while the original
phrase makes a claim about certain referents.

So you can't rewrite "lo poi'i" as "da poi" either. lo is plural, da
is singular. lo is a constant, da is a variable and creates scope.
They are not interchangeable.

poi'i (see also: http://mw.lojban.org/index.php?title=new_voi) is very
useful and allows for more flexibility (see the examples on the page I
linked).

Again, while you can try to come up with paraphrases (I already showed
"ckaji be lo ka ce'u goi ko'a zo'u", which is equivalent but way too
long), it's like trying to prove, say, poi useless by showing that it's
redundant to je.