ZFC-, which is ZFC minus power set, is modelled by $ L_{\delta}$ where $\delta$ is an admissible ordinal larger than any least $\Sigma_{n}$-admissible ordinal for n a natural number. Can some provide more information? Does it have a name? What is its most natural omega sequence?
-
1$\begingroup$ If you only want some $\delta$ for which $L_\delta\models$ZFC-, then you can take $\delta=\omega_1$, or any infinite successor cardinal, for that matter. But probably you want to characterize the least $\delta$ such that $L_\delta\models$ZFC-? $\endgroup$– Joel David HamkinsCommented Jul 31, 2013 at 23:53
-
1$\begingroup$ Your ordinal $\delta$ is smaller than $\delta^1_2$, the supremum of the $\Delta^1_2$-definable pre-wellorder relations on $\mathbb{N}$, but larger of course than $\delta^1_1=\omega_1^{ck}$. $\endgroup$– Joel David HamkinsCommented Aug 1, 2013 at 2:00
-
$\begingroup$ Yes, I am interested in the least such ordinal. $\endgroup$– Frode Alfson BjørdalCommented Aug 1, 2013 at 8:16
-
$\begingroup$ Question: Is $L_{\beta_0}$ then also a model of $\Sigma_n$-$KP$ (that is, $KP$+Infinity+$\Sigma_n$-Collection+$\Sigma_n$-Separation? $\endgroup$– Thomas BenjaminCommented Feb 6, 2015 at 13:37
-
$\begingroup$ Yes, that would be so. $\endgroup$– Frode Alfson BjørdalCommented Feb 6, 2015 at 21:52
1 Answer
The least such ordinal $\beta$, often written $\beta_0$, for which $L_\beta$ is a $ZF^-$ is also characterised as the "ordinal of ramified analysis". This is because the ramified analytical hierarchy, which builds up cumulative second order number theoretic structures of the form $$\underline{P_\alpha}=( P_\alpha,\mathbb{N}, +, \times, \ldots)$$ with $P_\alpha\subseteq P_{\alpha +1}\subseteq \mathcal{P}(\mathbb{N})$ by looking at all sets of integers definable over $\underline{P_\alpha}$ by using instances of second order comprehensionto to obtain $P_{\alpha+1}$. (Unions are taken at limits). This hierarchy has height exactly $\beta_0$ (meaning $P_{\beta_0}=P_{\beta_0+1}$.) Alternatively put: $\underline{P_{\beta_0}}$ is then the least model of $Z_2$, or full second order comprehension (sometimes just abbreviated to "analysis", or ``second order number theory'').
${P_{\beta_0}}$ is then $\mathcal{P}(\mathbb{N}) \cap L_{\beta_0}$.
$\beta_0$ of course is $\Sigma_n$-admissible for all $n$. $L_{\beta_0+1}$ sees that $\beta_0$ is countable, and so I suppose one may define a particular $\omega$ sequence in $L_{\beta_0+1}$ cofinal in it by just taking the $<_L$- least such.
One can also paraphrase the assertion that $L_{\beta_0}$ is the least $\beta$ with $L_\beta$ a $ZF^-$ model, as being the least such that for no $n$ is the fine-structural projectum $\rho^n_\beta$ less than $\beta$. (But this is really only dressing up one concept in the jargon of another.)
$L_{\beta_0}$ is (I believe) the least level of the $L$ hierarchy whose reals form a model of $\Delta^0_4$-Determinacy (Martin - unpublished).
For the Ramified Analytical Hierarchy: see Boyd, Hensel, Putnam "An intrinsic characterisation of the ramified analytical hierarchy", JSL, late 60's I believe. This hierarchy is not much studied these days, but was of interest in the pre-Jensen fine-structural era of the analysis of levels of $L$ (the latter's work now surpassing it). For connections between models of subsystems of second order comprehension and constructible sets, see also S. Simpson's book "Subsystems of second order arithmetic".
-
$\begingroup$ Can you explain what work Jensen did and how it superseded the need to study the ramified analytical hierarchy? $\endgroup$ Commented May 14, 2019 at 6:38