%! TEX root = FC.tex % vim: tw=50 % 11/02/2025 12PM There is a finite fragment $T_{\mathbb{L}}$ of $\zf$[!] that proves that all of the operations occuring in the definition of $\mathbb{L}$, i.e. $\Fml, X^{< \omega}, \models, D, \mathcal{D}$ are well-defined and \gls{abs}. Thus, if $M$ is a transitive model of $T_{\mathbb{L}}$, then \[ \forall \alpha \in \Ord \cap M, \mathbb{L}_\alpha \in M \] and thus $\mathbb{L}_{\alpha} \subseteq M$. So $\bigcup_{\alpha \in \Ord \cap M} \mathbb{L}_\alpha \subseteq M$. \subsubsection*{Axioms of ZF} \textbf{Structural axioms:} \begin{itemize} \item Extensionality: \[ \forall x, \forall y, (\forall w (w \in x \leftrightarrow w \in y) \to x = y) .\] \item Foundation: \[ \forall x(\exists v, (v \in x) \to \exists m, (m \in x \wedge \forall w, \neg (w \in m \wedge w \in x))) .\] \item Infinity: \[ \exists i, \exists e, (\forall v, \neg v \in e \wedge e \in i) \wedge \forall x, (x \in i \to \exists s, (s \in i \wedge \forall w, (w \in s \leftrightarrow w \in x \vee w = x))) .\] \end{itemize} \textbf{Functional axioms:} \begin{itemize} \item Pairing: \[ \forall x, \forall y, \exists p, \forall w, (w \in p \leftrightarrow w = x \vee w = y) .\] \item Union: \[ \forall x, \exists v, \forall w, (w \in v \leftrightarrow \exists z, (z \in x \wedge w \in z)) .\] \item Powerset: \[ \forall x, \exists p, \forall w, (w \in p \leftrightarrow \forall v, (v \in w \to v \in x)) .\] \item Separation $\varphi$: \[ \forall \ol{p}, \forall x, \exists s, \forall w, (w \in s \leftrightarrow w \in x \wedge \varphi(w, \ol{p})) .\] \item Replacement $\varphi$: \[ \forall \ol{p}, [\forall x, \forall y, \forall z (\varphi(x, y, \ol{p}) \wedge \varphi(x, z, \ol{p}) \to y = z)] \to \forall x, \exists r, \forall w, (w \in r \leftrightarrow \exists y, (y \in x \wedge \varphi(y, w, \ol{p}))) .\] \end{itemize} Now we check that these hold in $\mathbb{L}$. In Lecture 2, we proved Extensionality and Foundation in all transitive structures, so also in $\mathbb{L}$. Note that $\omega$ satisfies the condition of the axiom of infinity, so any $M$ transitive with $\omega \in M$ will satisfy the axiom of infinity. TODO Now do pairing and union. Since the definitions of pairs and unions \begin{align*} z &= \{x, y\} \\ z &= \bigcup x \end{align*} are \gls{abs} for transitive models, it's enough to show that \begin{align*} \forall x, y \in \mathbb{L}, &\{x,y\} \in \mathbb{L} \\ \forall x \in \mathbb{L} &\bigcup x \in \mathbb{L} \end{align*} If $x, y \in \mathbb{L}_\alpha$, $\varphi(w, x, y) \defeq w = x \vee w = y$, \begin{align*} D(\varphi, (x, y), L_\alpha) &= \{w \in L_\alpha : L_\alpha \models \varphi(w, x, y)\} \\ &= \{w \in L_\alpha : L_\alpha \models w = x \wedge w \} TODO \end{align*} Powerset axiom. \[ \forall x, \exists p, \ub{\forall w, (w \in p \leftrightarrow w \subseteq x)}_{*} \] The problem here is that $*$ is not obviously \gls{abs}. In particular, $z = \mathcal{P}(x)$ is not \gls{abs}. Consider $\mathbb{L}_{\omega + 1}$: we have $\omega \in \mathbb{L}_{\omega + 1}$ and $\mathcal{P} \cap \mathbb{L}_{\omega + 1}$ is countable. In $\mathbb{L}_{\omega + 2}$, we find \[ \{a \in \mathbb{L}_{\omega + 1} : a \subseteq \omega\} \] which is the best possible answer to the question ``what is the power set of $\omega$?'' that $\mathbb{L}_{\omega + 1}$ can give, but unlikely to be the correct answer. Consider instead $\mathcal{P}(\omega) \cap \mathbb{L} \eqdef P$ and define \[ \Omega \defeq \{\rho_{\mathbb{L}}(a) : a \in P\} .\] (reminder: $\rho_{\mathbb{L}}(a)$ is the least $\alpha$ such that $a \in \mathbb{L}_{\alpha + 1}$) By Replacement, $\Omega$ is a set of ordinals, so find $\alpha > \Omega$. Then $P \subseteq \mathbb{L}_\alpha$. Therefore $P = \{a \in \mathbb{L}_\alpha : a \subseteq \omega\} \in \mathbb{L}_{\alpha + 1}$, so $P \in \mathbb{L}$. Separation: \[ \forall \ol{p}, \forall x, \exists s, \forall w, (w \in s \leftrightarrow \ub{w \in x \wedge \varphi(w, \ol{p})}_{\varphi'(w, x, \ol{p})}) \] If $x \in \mathbb{L}_\alpha$, then \begin{align*} D(\varphi', x, L_\alpha) &\defeq \{w \in L_\alpha : L_\alpha \models \varphi'(w, x, \ol{p})\} \\ &= \{w \in L_\alpha : L_\alpha \models w \in x \wedge \varphi(w, \ol{p})\} \\ &\stackrel{?}{=} \{w \in \ub{\mathbb{L}}_{\text{not a problem}} : \ub{\mathbb{L}}_{\text{this is a problem}} \models w \in x \wedge \varphi(w, \ol{p})\} \end{align*} If $\varphi$ is not \gls{abs} between $L_\alpha$ and $L$, this won't work. \nameref{lrt} to the rescue: $\forall \varphi, \forall \alpha, \exists \theta > \alpha$ such that $\varphi$ is \gls{abs} between $L_\theta$ and $L$. Thus: form \begin{align*} D(\varphi', x, L_\theta) &= \{w \in L_\theta : L_\theta \models w \in x \wedge \varphi(w, \ol{p})\} \\ &\stackrel{\text{\gls{abs}}}{=} \{w \in L_\theta : L \models w \in x \wedge \varphi(w, \ol{p})\} \end{align*} Replacement: This will be on \es{2}. The proof is a combination of the ideas from power set and separation. \begin{fccoro}[Minimality] Assuming: - $T$ is a transitive model of $\zf$ Then: for all $\alpha \in T \cap \Ord$, $\mathbb{L}_\alpha \subseteq T$. Axiom of TODO. \end{fccoro} \begin{remark*} Remark on the Axiom of Choice. G\"odel (1938): $\Con(\zf) \to \Con(\zfc)$. Note first that everything we did so far only needed $\zf$ in the universe. We will sketch that $\mathbb{L} \models \ac$. In fact a strong version of $\ac$ known as GLOBAL CHOICE: there is an absolutely definable bijective operation between $\mathbb{L}$ and $\Ord$. Sketch: Recursive construction of bijections $\pi_\alpha : \mathbb{L}_\alpha \to \eta_\alpha$ for some ordinal $\eta_\alpha$, and such that for $\beta < \alpha$ we have $\pi_\alpha |_{\mathbb{L}_\beta} = \pi_\beta$. If $\lambda$ is a limit and $\pi_\alpha$ is defined for $\alpha < \lambda$, then let \[ \pi_\lambda(x) \defeq \pi_\alpha(x) \] if $x \in \mathbb{L}_\alpha$. Suppose $\alpha = \beta + 1$ and $\pi_\beta$ is given by $\pi_\beta : \mathbb{L}_\beta \to \eta_\beta$. Consider $\Fml \times L_\beta^{<\omega}$. Well-order it in order type $\eta_\beta'$ via the induced $\pi_\beta$ well-order. Then if $x \in L_\alpha$, say \[ \pi_\alpha(x) \defeq \begin{cases} \pi_\beta(x) & \text{if $x \in L_\beta$} \\ \eta_\beta \xi & \text{if $\xi$ is the ordinal corresponding to the least $(\varphi, \ol{p})$ such that $x = D(\varphi, \ol{p}, TODO)$} \end{cases} \] \end{remark*}