Building independence relations in abstract elementary classes

TL;DR

This paper develops methods to construct forking-like independence notions in tame abstract elementary classes (AECs) with amalgamation, showing that categoricity in high cardinals implies stability, categoricity, and the existence of a good frame, advancing the understanding of classification theory in AECs.

Contribution

It introduces new techniques to build forking-like notions in tame AECs and proves categoricity in high cardinals leads to stability, categoricity, and global independence, extending classical model theory results.

Findings

High-level categoricity implies stability and categoricity in tame AECs.

Existence of a good frame for singleton types in categorical high cardinals.

Under certain set-theoretic assumptions, categoricity in some high cardinals implies categoricity in all high cardinals.

Abstract

We study general methods to build forking-like notions in the framework of tame abstract elementary classes (AECs) with amalgamation. We show that whenever such classes are categorical in a high-enough cardinal, they admit a good frame: a forking-like notion for types of singleton elements. $\mathbf{Theorem}$ (Superstability from categoricity) Let $K$ be a $(<\kappa)$-tame AEC with amalgamation. If $\kappa = \beth_\kappa > \text{LS} (K)$ and $K$ is categorical in a $\lambda > \kappa$, then: * $K$ is stable in all cardinals $\ge \kappa$. * $K$ is categorical in $\kappa$. * There is a type-full good $\lambda$-frame with underlying class $K_\lambda$. Under more locality conditions, we prove that the frame extends to a global independence notion (for types of arbitrary length). $\mathbf{Theorem}$ (A global independence notion from categoricity) Let $K$ be a densely type-local, fully tame and type short AEC with amalgamation. If $K$ is categorical in unboundedly many cardinals, then there exists $\lambda \ge \text{LS} (K)$ such that $K_{\ge \lambda}$ admits a global independence relation with the properties of forking in a superstable first-order theory. As an application, we deduce (modulo an unproven claim of Shelah) that Shelah's eventual categoricity conjecture for AECs (without assuming categoricity in a successor cardinal) follows from the weak generalized continuum hypothesis and a large cardinal axiom. $\textbf{Corollary}$ Assume $2^{\lambda} < 2^{\lambda^+}$ for all cardinals $\lambda$, as well as an unpublished claim of Shelah. If there exists a proper class of strongly compact cardinals, then any AEC categorical in some high-enough cardinal is categorical in all high-enough cardinals.