[FOM] Cantor Bendixson and the Axiom of Replacement

[D.R. MacIver]

I do apologise - this ends up being much more trivial than I thought it was.

If you consider the map P(E) -> P(E) given by A -> { x : x is a limit point 
of A } then A is perfect iff it's a fixed point of this map. But if you 
consider P(X) ordered by reverse set inclusion (i.e. a <= b iff b \subseteq 
a ), this ends up being a monotonic map on a complete poset, and by 
tarski's fixed point theorem (which doesn't use the axiom of replacement 
anywhere) has a smallest fixed point. i.e. there exists a largest perfect 
set.

So it is indeed a theorem of Z set theory. Oh well.

David MacIver