Technical Summary: Tame Fundamental Groups of Rigid Spaces

1. Problem Statement and Motivation

The paper addresses the structural properties of the étale fundamental group for rigid-analytic spaces over non-archimedean fields. While the étale cohomology groups of smooth, proper rigid spaces are known to be finite (Scholze), the étale fundamental group $\pi_1(X)$ is generally not topologically finitely generated. A primary obstruction is wild ramification, particularly along the "boundary at infinity" or the special fibre of formal models. For instance, the étale fundamental group of the affinoid unit disc is not finitely generated because it captures wild ramification phenomena analogous to those in the affine line over a finite field.

To overcome this, the authors introduce the tame étale fundamental group $\pi_1^t(X/K)$ for a rigid space $X$ over a non-archimedean field $K$. The central problem is to define a notion of tameness that is robust enough to ensure finiteness properties (finite generation and finite presentation) while remaining compatible with the geometry of rigid spaces. The authors note that a standard "absolute" definition of tameness (based on valuations of residue fields) fails to yield finiteness because it does not account for the "boundary at infinity" in the special fibre.

2. Methodology and Definitions

2.1 Relative Tameness

The core innovation is the definition of tameness relative to the base field $K$.

• Definition: A finite étale morphism $f: Y \to X$ of rigid spaces is tame relative to $K$ if, for every point $y \in Y$, the extension of valued fields $(k(y), k(y)^+)/(k(f(y)), k(f(y))^+)$ is tamely ramified with respect to every valuation on $k(y)$ whose valuation ring $V$ satisfies $K^\circ \subseteq V \subseteq k(y)^+$.
• Geometric Interpretation: This condition effectively tests tameness not just on the space $X$ itself, but on its universal vertical compactification (or Huber's universal compactification). The authors utilize the concept of root triples $(x^\circ, k(x)^+, x^\succ)$ to characterize the points of this compactification, where $x^\circ$ is a point without vertical generalizations, $k(x)^+$ is a valuation subring, and $x^\succ$ is a center on the base.

2.2 Reduction to Logarithmic Geometry

The proof strategy relies on reducing the problem to the special fibre of a formal model via logarithmic geometry:

1. Descent and Uniformization: Using results by Gabber and Temkin, the authors reduce the general case to the situation where $K$ is algebraically closed and $X$ admits a strictly semistable formal model $\mathfrak{X}$.
2. Logarithmic Correspondence: They establish an equivalence between the category of finite Kummer étale covers of the log formal scheme $\mathfrak{X}^{\log}$ (endowed with the standard log structure induced by the generic fibre) and the category of finite étale covers of the rigid generic fibre $X$ that are tame relative to $K$.
3. Special Fibre Analysis: The problem is then transferred to the special fibre $X_k$ (a log scheme over the residue field $k$). The authors prove that the tame fundamental group of the rigid space is isomorphic to the tame fundamental group of the special fibre log scheme:
   $$\pi_1^t(X/K) \simeq \pi_1^t(X_k^{\log}/k)$$
4. Log Schemes of Type (Vdiv): A significant technical hurdle is that the log structures arising from semistable reduction over general valuation rings are not necessarily fs (fine and saturated). The authors work within the framework of log schemes of type (Vdiv) (finitely generated over a divisible valuative submonoid), extending the theory of Kummer étale fundamental groups beyond the fs case (building on their companion paper [AHLS24]).

2.3 Finiteness Techniques

• Formal Gluing: To prove finite generation for log schemes with non-constant log structures, the authors employ a "formal gluing" argument along the log stratification. This reduces the problem to the case of locally constant log structures.
• Surgery and Descent: For finite presentation, they use a "surgical" procedure (decomposing the log scheme into strata closures) combined with descent arguments (Van Kampen formulas for Galois categories) to relate the fundamental group to that of a smooth variety with a projective snc (simple normal crossings) compactification.

3. Key Contributions and Results

3.1 Finite Generation

Theorem B: Let $K$ be a non-archimedean field with a topologically finitely generated tame Galois group (e.g., algebraically closed or a local field). If $X$ is a non-empty, connected, quasi-compact, and quasi-separated (qcqs) rigid space over $K$, then $\pi_1^t(X/K)$ is topologically finitely generated.

• Significance: This resolves the finiteness issue for a broad class of rigid spaces, provided one restricts to tame covers. The proof relies on the reduction to the case of semistable formal models and the finite generation of the tame fundamental group of log schemes of type (Vdiv).

3.2 Finite Presentation

Theorem C: Let $K$ be algebraically closed. If $X$ is the rigid generic fibre of a strictly semistable formal scheme $\mathfrak{X}$ such that the smooth locus of the special fibre $\mathfrak{X}_k$ admits a projective snc compactification, then $\pi_1^t(X/K)$ is topologically finitely presented.

• Significance: This provides a condition under which the group is not just generated by a finite set but has a finite presentation. The result depends on the finite presentation of the tame fundamental group of log schemes with specific compactification properties.

3.3 Comparison Theorems

The paper establishes rigorous comparison theorems between different notions of tameness:

• Absolute vs. Relative: It clarifies that absolute tameness is insufficient for finiteness, while relative tameness (testing against the base field's valuation ring) is the correct notion.
• Rigid vs. Logarithmic: It proves that for strictly semistable formal schemes, the tame fundamental group of the rigid generic fibre is isomorphic to the tame fundamental group of the special fibre log scheme (Corollary 5.7.5).
• Scheme vs. Adic: It extends comparison results from schemes to adic spaces, showing that tameness conditions on the universal compactification correspond to tameness on the underlying scheme's special fibre under cdh uniformisation properties.

3.4 Examples and Computations

The authors compute $\pi_1^t(X/K)$ for several standard examples:

• Unit Disc: $\pi_1^t(\mathbb{D}^1/K) = 1$ (trivial).
• Polydiscs and Tori: For $X = \mathbb{D}^n \times (\mathbb{G}_m)^m$, $\pi_1^t(X/K) \simeq \widehat{\mathbb{Z}}'(1)^m$.
• Annulus: $\pi_1^t(\text{Annulus}/K) \simeq \widehat{\mathbb{Z}}'(1)$.
• Punctured Disc: $\pi_1^t(\mathbb{D}^1 \setminus \{0\}/K) \simeq \widehat{\mathbb{Z}}'(1)$.
• Standard Semistable Schemes: For $X = \text{Spa}(K\langle T_1, \dots, T_n \rangle / (\pi - T_1 \cdots T_n))$, $\pi_1^t(X/K) \simeq \widehat{\mathbb{Z}}'(1)^{n-1}$.

4. Significance and Scope

The paper claims that the tame fundamental group $\pi_1^t(X/K)$ is a "reasonably well-behaved" invariant for rigid spaces, specifically in terms of finite generation. The authors emphasize that this invariant is crucial for establishing further structural properties, such as a Künneth formula and invariance under algebraically closed base field extensions, which are the subjects of their upcoming work.

The work is modest in its scope regarding the general case of finite presentation; it explicitly states that finite presentation remains an open question in general, holding only under specific geometric assumptions (strictly semistable models with snc compactifications). The paper does not claim to solve the finiteness of the full étale fundamental group (which includes wild ramification) but rather isolates the tame part to demonstrate that the "bad" behavior of rigid spaces is entirely attributable to wild ramification.

The methodology relies heavily on extending logarithmic geometry to non-fs monoids (type Vdiv) and utilizing the "vertical compactification" of adic spaces, bridging the gap between the geometry of rigid spaces and the combinatorics of log schemes.