Gt-Henselian Field Topologies

• Gt-henselian field topologies are refined field structures defined by an implicit function theorem, local homeomorphism of étale morphisms, and robust root-finding properties.
• They connect classical valuation topologies with the universal étale-open topology, providing new insights into fields lacking traditional valuation frameworks.
• Their abundance in large and countable fields leads to a rich lattice of incomparable topologies, raising significant open questions in model theory and algebraic geometry.

A gt-henselian (“generalized t-henselian”) field topology is a refinement of the notion of t-henselianity for topological fields. These topologies play a central role in the interaction between field arithmetic, valuation theory, algebraic geometry, and model theory, especially for fields devoid of classical valuation structures. Recent advances have clarified both their abundance and structural properties, emphasizing the diversity of gt-henselian topologies, their connection with large fields, the universal étale-open topology, and the prevailing open problems in the area (Walsberg, 2 Sep 2025, Johnson, 21 Aug 2025, Dittmann et al., 2022, Johnson et al., 2021, Johnson, 15 Apr 2025).

1. Definition and Characterizations

A field topology $\tau$ on a field $K$ is said to be gt-henselian if it satisfies any (and thus all) of the following equivalent conditions:

• Implicit Function Theorem for Polynomials: For any system of polynomials $$f_1(\bar x, \bar y), \ldots, f_m(\bar x, \bar y)\in K[x_1,\ldots,x_n, y_1, \ldots, y_m]$$ and any $(\bar p, \bar q)\in K^{n+m}$ with invertible Jacobian $$\left(\partial f_i/\partial y_j\right)(\bar p, \bar q)$$, there exist $\tau$-open neighborhoods $U\ni \bar p$ and $V\ni \bar q$ such that the common zero set in $U \times V$ is the graph of a $\tau$-continuous function $K$0.
• Étale Morphisms as Local Homeomorphisms: Any étale morphism $K$1 of $K$2-varieties induces a local homeomorphism $K$3 in the $K$4-topologies.
• Root-Finding Property: For every integer $K$5 and every $K$6-open neighborhood $K$7 of $K$8, there exists a $K$9-open $$f_1(\bar x, \bar y), \ldots, f_m(\bar x, \bar y)\in K[x_1,\ldots,x_n, y_1, \ldots, y_m]$$0 such that for any $$f_1(\bar x, \bar y), \ldots, f_m(\bar x, \bar y)\in K[x_1,\ldots,x_n, y_1, \ldots, y_m]$$1, the polynomial $$f_1(\bar x, \bar y), \ldots, f_m(\bar x, \bar y)\in K[x_1,\ldots,x_n, y_1, \ldots, y_m]$$2 has a simple root in $$f_1(\bar x, \bar y), \ldots, f_m(\bar x, \bar y)\in K[x_1,\ldots,x_n, y_1, \ldots, y_m]$$3.

A field topology is locally bounded if it admits a $$f_1(\bar x, \bar y), \ldots, f_m(\bar x, \bar y)\in K[x_1,\ldots,x_n, y_1, \ldots, y_m]$$4-neighborhood $$f_1(\bar x, \bar y), \ldots, f_m(\bar x, \bar y)\in K[x_1,\ldots,x_n, y_1, \ldots, y_m]$$5 of $$f_1(\bar x, \bar y), \ldots, f_m(\bar x, \bar y)\in K[x_1,\ldots,x_n, y_1, \ldots, y_m]$$6 which is $$f_1(\bar x, \bar y), \ldots, f_m(\bar x, \bar y)\in K[x_1,\ldots,x_n, y_1, \ldots, y_m]$$7-bounded, i.e., $$f_1(\bar x, \bar y), \ldots, f_m(\bar x, \bar y)\in K[x_1,\ldots,x_n, y_1, \ldots, y_m]$$8, $$f_1(\bar x, \bar y), \ldots, f_m(\bar x, \bar y)\in K[x_1,\ldots,x_n, y_1, \ldots, y_m]$$9 such that $(\bar p, \bar q)\in K^{n+m}$0 (Dittmann et al., 2022, Johnson, 15 Apr 2025).

Valuation topologies (V-topologies) are always t-henselian; in the non-discrete case, a field topology is t-henselian if and only if it is both a V-topology and gt-henselian (Dittmann et al., 2022).

2. Étale-Open Topology and Its Universal Property

The étale-open topology $(\bar p, \bar q)\in K^{n+m}$1 on the $(\bar p, \bar q)\in K^{n+m}$2-points of varieties is defined by taking as a basis the images of $(\bar p, \bar q)\in K^{n+m}$3-points of étale morphisms:

$(\bar p, \bar q)\in K^{n+m}$4

for $(\bar p, \bar q)\in K^{n+m}$5 étale. The canonical topologies induced on $(\bar p, \bar q)\in K^{n+m}$6 via affine embeddings are E-topologies.

For countable fields $(\bar p, \bar q)\in K^{n+m}$7, $(\bar p, \bar q)\in K^{n+m}$8 is the coarsest common refinement of all gt-henselian field topologies:

$(\bar p, \bar q)\in K^{n+m}$9

(Johnson, 21 Aug 2025). When $$\left(\partial f_i/\partial y_j\right)(\bar p, \bar q)$$0 is the fraction field of a quasi-excellent henselian local domain $$\left(\partial f_i/\partial y_j\right)(\bar p, \bar q)$$1, $$\left(\partial f_i/\partial y_j\right)(\bar p, \bar q)$$2 agrees with the $$\left(\partial f_i/\partial y_j\right)(\bar p, \bar q)$$3-adic topology (Dittmann et al., 2022, Johnson et al., 2021).

If $$\left(\partial f_i/\partial y_j\right)(\bar p, \bar q)$$4 is non-large (e.g., a number field), $$\left(\partial f_i/\partial y_j\right)(\bar p, \bar q)$$5 is discrete. In contrast, for real closed, $$\left(\partial f_i/\partial y_j\right)(\bar p, \bar q)$$6-adic, or fields like $$\left(\partial f_i/\partial y_j\right)(\bar p, \bar q)$$7 for $$\left(\partial f_i/\partial y_j\right)(\bar p, \bar q)$$8, the étale-open topology coincides with classical analytic or $$\left(\partial f_i/\partial y_j\right)(\bar p, \bar q)$$9-adic topologies (Johnson et al., 2021).

3. Abundance and Construction of Gt-Henselian Topologies

On fields of characteristic zero and infinite transcendence degree, if there exists any gt-henselian topology, there are $\tau$0 pairwise incomparable gt-henselian field topologies, constructed via derivations. For a given gt-henselian topology $\tau$1 and a set $\tau$2 of derivations,

$\tau$3

where basic $\tau$4-neighborhoods of $\tau$5 refine $\tau$6 by additional continuity conditions for the derivations. If $\tau$7 is large and countable, there are $\tau$8 incomparable gt-henselian topologies, and $\tau$9 second-countable ones (Walsberg, 2 Sep 2025).

A locally bounded field topology is gt-henselian if and only if certain monic polynomial families have dense roots, or equivalently, if the root-finding axiom of gt-henselianity holds, or if all étale morphisms are open. In saturated fields, every locally bounded gt-henselian topology is locally equivalent to some $U\ni \bar p$0-adic topology for a henselian local ring $U\ni \bar p$1 (Dittmann et al., 2022).

4. Structural and Model-Theoretic Criteria

Let $U\ni \bar p$2 be a field with a locally bounded field topology. Then:

• If $U\ni \bar p$3 is an NIP field with $U\ni \bar p$4 or finite dp-rank, every definable field topology is gt-henselian (Johnson, 15 Apr 2025).
• Any $U\ni \bar p$5-adic topology ($U\ni \bar p$6 henselian local) is gt-henselian.
• For finite dp-rank $U\ni \bar p$7, definable field topologies are $U\ni \bar p$8-topologies (topologies of finite breadth).
• In dp-minimal fields, every definable field topology is a V-topology, so gt-henselianity coincides with t-henselianity.

The Generalized Henselianity Conjecture for NIP integral domains states that NIP domains are henselian local if and only if all definable field topologies on NIP fields are gt-henselian (Johnson, 15 Apr 2025).

5. Comparison with Canonical and V-Topologies

In dp-finite unstable fields, the canonical topology is defined via “heavy” definable sets—sets of full dp-rank. The group of infinitesimals $U\ni \bar p$9 is defined as:

$V\ni \bar q$0

with $V\ni \bar q$1 being the group of multiplicative infinitesimals. The canonical topology is a field topology, and conjecturally coincides with a V-topology, i.e., comes from a valuation (Johnson, 2019).

Key relationships:

• If the canonical topology is a V-topology, then the field is henselian with respect to the corresponding valuation.
• The classification of dp-finite fields (Shelah conjecture) depends on this V-topology conjecture: unstable dp-finite fields then admit invariant nontrivial henselian valuations.

For fields like $V\ni \bar q$2 with $V\ni \bar q$3, the $V\ni \bar q$4-adic (and hence étale-open) topology is gt-henselian but not a V-topology (Johnson et al., 2021).

6. Special Classes, Examples, and Pathologies

Typical examples of gt-henselian topologies:

• Classical valuation topologies on henselian fields (V-topologies): both gt-henselian and t-henselian.
• The order topology on real closed fields: gt-henselian, not V-topological (except $V\ni \bar q$5).
• $V\ni \bar q$6-adic topologies for henselian local domains: locally bounded gt-henselian topologies, sometimes coinciding with the étale-open topology if $V\ni \bar q$7 is quasi-excellent.
• Derivation-induced topologies: may be highly non-canonical, not V-topological, and in general not locally bounded if the derivation set is infinite (Walsberg, 2 Sep 2025).

Negative results:

• For pseudo-algebraically closed (PAC) fields, the étale-open topology can never be induced by a field topology; thus, while such fields may admit gt-henselian field topologies, these never coincide with the universal geometric (étale-open) topology (Dittmann et al., 2022).

7. Open Problems and Future Directions

Major open problems and research directions include:

• Determining for which large fields $V\ni \bar q$8 the étale-open topology $V\ni \bar q$9 is itself a field topology (Walsberg, 2 Sep 2025).
• Whether the intersection of all gt-henselian topologies on any large field coincides with the étale-open topology, generalizing results from the countable case (Johnson, 21 Aug 2025, Walsberg, 2 Sep 2025).
• Classifying definable gt-henselian topologies in central model-theoretic classes (NIP, dp-minimal, dp-finite).
• Understanding the lattice and independence properties of gt-henselian topologies on a given field; for countable, large fields, the space of such topologies is extremely rich (Walsberg, 2 Sep 2025).
• Investigating the structure and role of $U \times V$0-completeness and the translation between gt-henselian and $U \times V$1-adic topologies in saturated fields (Dittmann et al., 2022).
• Elucidating the relationship between dp-finite topologies, canonical field topologies, and classical valuation theory (Johnson, 2019).

The explicit connection of gt-henselian field topologies to local boundedness, Galois-algebraic and definability considerations, and their geometric and model-theoretic implications situates them as a central object in modern field theory and algebraic geometry.