AKE principles for deeply ramified fields

Abstract: We study the model theory of deeply ramified fields of positive characteristic. Generalizing the perfect case treated in work by Jahnke and Kartas on the model theory of perfectoid fields, we obtain Ax-Kochen/Ershov principles for certain deeply ramified fields of positive characteristic and fixed degree of imperfection. Our results apply in particular to all deeply ramified henselian valued fields of rank 1.

• The paper introduces a variant of the AKE principle for deeply ramified fields, extending logical transfer to non-perfect fields with fixed degrees of imperfection.
• It applies rigorous model-theoretic methods using pointed valued fields in class Cₚ,ₑ to demonstrate elementary embeddings between value groups and residue rings.
• The results clarify the interplay between separable tameness and ramification, offering criteria for substructure, existential, and complete equivalence in positive characteristic.

AKE Principles for Deeply Ramified Fields: A Technical Overview

Introduction and Motivation

The paper "AKE principles for deeply ramified fields" (2603.29528) addresses the model theory of valued fields in positive characteristic, focusing on the extension of Ax–Kochen/Ershov (AKE) principles to a class of non-perfect, deeply ramified fields with fixed degree of imperfection. The work builds on and generalizes previous results for perfectoid fields and perfect deeply ramified fields, particularly those by Jahnke and Kartas, by eliminating the perfectness assumption while preserving strong transfer and completeness properties in logical structure. The theoretical interest is twofold: clarifying the model-theoretic landscape of non-perfect valued fields and demonstrating the robustness of AKE principles even under relaxation of classical perfection requirements.

Key Definitions and Structural Results

Central to the paper is the introduction of the class $C_{p,e}$ of pointed valued fields. A structure $(K,v,t)$ belongs to $C_{p,e}$ if:

• $K$ has characteristic $p>0$ and degree of imperfection $e \in \mathbb{N} \cup \{\infty\}$,
• $(K,v)$ is deeply ramified,
• $0 < vt < \infty$, and the coarsening of $v$ by inverting $t$ renders $(K,v,t)$0 separably tame.

Deeply ramified fields are characterized as those for which the completion is perfect, $(K,v,t)$1 is dense in its perfect hull, and various semi-perfectness criteria hold for quotients of the valuation ring.

A succession of technical lemmas ensures that the pointed structures are stable under elementary equivalence, and that the properties—especially separable tameness after coarsening—are first-order expressible and hence the class $(K,v,t)$2 is elementary. This provides solid ground for deploying model-theoretic tools.

The paper further clarifies the interplay of these properties with unramified and almost separably tame valuations. The latter are realized as fields that admit an $(K,v,t)$3-saturated elementary extension with a nontrivial separably tame coarsening.

The AKE-Type Results

Main Theorem: Relative AKE Principles

The central technical contribution is a variant of the AKE principle tailored to the class $(K,v,t)$4. The theorem demonstrates that, given separable extensions $(K,v,t)$5 and $(K,v,t)$6:

• If value groups embed existentially, and relative degrees of imperfection are preserved, then
• Elementary equivalence of $(K,v,t)$7 and $(K,v,t)$8 over $(K,v,t)$9 is equivalent to the elementary equivalence of their pointed value groups and residue rings modulo $C_{p,e}$0.

Symbolically:

$C_{p,e}$1

The result holds uniformly across all degrees of imperfection, including the infinite case, with the degree constraint relaxed to $C_{p,e}$2.

Consequences: Substructure, Existential, and Completeness Results

• Substructure Principle: For separable $C_{p,e}$3 with both in $C_{p,e}$4, $C_{p,e}$5 iff $C_{p,e}$6 and $C_{p,e}$7.
• Existential Closedness: $C_{p,e}$8 iff the corresponding value groups and residue rings embed existentially.
• Completeness Principle: $C_{p,e}$9 if their pointed value groups and residue rings are elementarily equivalent.

These capture both the classical $K$0, $K$1, and $K$2 principles, but in the context of the finer invariants (pointed value group and residue ring modulo a chosen $K$3) that react well to imperfection in positive characteristic.

Strong Applications and Explicit Constructions

The results yield several notable corollaries:

• Any deeply ramified algebraic extension $K$4 of the henselization $K$5 embeds elementarily into $K$6.
• For Artin-Schreier closures, $K$7.
• The embedding $K$8 is elementary, illustrating the transfer to higher rank value groups.

The paper also discusses examples demonstrating the strictness of class definitions (there exist deeply ramified fields in $K$9 that are not separably tame), as well as non-examples coming from ultraproducts, underlining the necessity of the technical restrictions imposed.

Implications for Model Theory and Valued Fields

From a model-theoretic perspective, the results show that even richly non-perfect fields, provided they are deeply ramified in the prescribed sense, satisfy strong logical transfer properties between field, value group, and residue invariants. The possibility of AKE-style principles in such generality indicates that the structure of positive characteristic valued fields is robust under conditions much weaker than previously recognized.

Practically, this facilitates quantifier reduction and transfer of elementary properties, suggesting that the complexity of definable sets and field-theoretic phenomena in these structures can be analyzed in terms of their lower-dimensional invariants.

The work also poses sharp open questions, notably whether the pointed version of AKE equivalence can be extended to full equivalence in the non-perfect case, which would have consequences for the model theory of fields with fixed imperfection degree.

Conclusion

This paper systematically extends the reach of the Ax–Kochen/Ershov principle to a new, wider class of valued fields in positive characteristic, removing the perfection assumption and working with deeply ramified henselian fields of fixed imperfection degree. The results tightly integrate valuation-theoretic, Galois-theoretic, and model-theoretic ideas, yielding strong, sharply formulated transfer principles and providing new insight into the logical geometry of valued fields. The established machinery offers substantial promise for future advances in the model theory of positive characteristic, potentially informing ongoing inquiries into decidability, definability, and the structure of higher-level ramification invariants.