In first-order set theory, the principle of complete reflection is the following.
$\tag{ CR} \forall \alpha\exists \beta>\alpha\forall\vec{x}\in V_\beta(\phi^\beta(\vec{x}) \leftrightarrow \phi(\vec{x}))$
whereLet $\phi\in \mathcal L_\in$ with free variables among$\alpha\not= 0$ be such that for every $\vec{x}$ and$\beta<\alpha$ there is $\beta<\gamma<\alpha$, where $\phi^\beta$$V_\gamma$ is the resultan elementary substructure of binding all the quantifiers in $\phi$ to$V_\alpha$. In other words, $V_\alpha$ is a limit of its $V_\beta$ elementary substructures.
It Then it is well-known that over ZFC minus replacement and infinity, CR implies replacement and infinity. I'm interested in the following principle:
$\tag{ CR*} \forall \alpha\exists \beta>\alpha\exists j\forall\vec{x}\in V_\beta(\phi^\beta(\vec{x}) \leftrightarrow \phi(\vec{j(x)}))$
Essentially, I want to know if CR* implies replacement over the other axioms of ZFC plus the claima simple result that every set is in some $V_\alpha$: that is, over ZFC - models replacement + $\forall x\exists \alpha(x\in V_\alpha)$.
However, since $j$ might be a proper class, theMy question is whether CR* implies first-order replacement in ZFC2: Let - second-order replacement +$\alpha\not= 0$ be such that for every $\forall x\exists \alpha(x\in V_\alpha)$.$\beta<\alpha$ there is (By ZFC2 I mean the axioms of ZFC with the axiom schemas of replacement$\beta<\gamma<\alpha$ and separation replaced by the corresponding second-order axioms together with the full comprehension schema (i.e.an elementary embedding from $\exists X\forall x(x\in X\leftrightarrow \phi)$, for any$V_\gamma$ to $\phi$ in the language))$V_\alpha$. Does it follow that $V_\alpha$ models replacement?