The Three Sets Lemma is the following Lemma:
Lemma: Let $f(x)$ be a function from $X$ to $X$ where $f(x)$ has no fixed points. Then there exists a partition of $X$ into three disjoint sets $X_1$, $X_2$, $X_3$ such that for any $i$, with $1 \leq i \leq 3$, we have $X_i \cap f(X_i)= \emptyset$.
The Three Sets Lemma is a theorem of ZFC. See e.g. Hagen von Eitzen's answer to 3-set-lemma in (naive) set theory. However, the Three Sets Lemma is not a theorem of ZF, although it is, much weaker than choice, and is in fact even weaker than the Boolean Prime Ideal theorem. See this other Mathoverflow question and the answer by godelian.
Let $S(k)$ be the following statement: "Let $f(x)$ be a function from $X$ to $X$ where $f(x)$ has no fixed points. Then there exists a partition of $X$ into $k$ disjoint sets $X_1, X_2, \dotsc X_k$ such that for any $i$, with $1 \leq i \leq k$, we have $X_i \cap f(X_i)= \emptyset$.
Note that $S(3)$ is then the Three Sets Lemma (and $S(1)$ and $S(2)$ are both just false).
Question: Is there any $k$ such that $S(k)$ is a theorem of ZF?
I suspect that answer is no.
I also suspect that the even weaker statement where we swap the order of the quantifiers (so first choose an $f(x)$ and then are allowed to choose $k$) is not a theorem of ZF.
Note that there is this other weaker variant of the Three Sets Lemma which is a theorem of ZF, but it requires the assumption that the set $X$ is well-ordered.