A note on the theory of well orders
Emil Jeřábek
TL;DR
The paper studies the first-order theory of the class $\mathcal{WO}$ of well-ordered sets, and proves that its theory is exactly the transfinite induction axiom schema $\mathsf{TI}$ and is decidable. A short, non-quantifier-elimination argument inspired by the Lau-Len method shows the equivalence: $\mathcal{WO}\vDash\varphi$ iff $\alpha\vDash\varphi$ for all $\alpha<\omega^\omega$ iff $\mathsf{TI}\vdash\varphi$. The approach uses finite Ehrenfeucht–Fraïssé-type reasoning, Ramsey theory to obtain a canonical $\alpha<\omega^\omega$ describing intervals, and a uniform satisfiability lemma (Lemma sat) that reduces decidability to a CNF-like representation of linear orders. Decidability follows since $\mathsf{TI}$ is recursively axiomatizable and, via Lemma sat, the complement set is recursively enumerable, yielding an effective decision procedure; this provides a self-contained, accessible proof that aligns with Lau-Len's methods while connecting to the established Doner-Mostowski-Tarski results.
Abstract
We give a simple proof that the first-order theory of well orders is axiomatized by transfinite induction, and that it is decidable.
