The infinitesimal subgroup of interpretable groups in some dp-minimal valued fields

Yatir Halevi, Assaf Hasson, Ya’acov Peterzil · 2025

Read The infinitesimal subgroup of interpretable groups in some dp-minimal valued fields by Yatir Halevi, Assaf Hasson, Ya’acov Peterzil (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2502-08163.txt.

Introduction

Yatir Halevi, Assaf Hasson, and Ya'acov Peterzil wrote a 2025 math study on groups that can be described inside certain valued fields. The fields are of a spare kind where rank stays low, which keeps their logic tame. For each endless group of this sort, the authors build a small inner subgroup made of points very close to the identity.

The inner subgroup is stitched from four parts, one for each basic sort in the field. The team shows the parts commute, so the whole can be defined by a single type. They also prove the build respects products and subgroups. The work may help link rank with how such fields handle imagined elements.

Catalog record

CREATORYatir Halevi, Assaf Hasson, Ya’acov Peterzil
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE125 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2502-08163

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

The infinitesimal subgroup of interpretable groups in some dp-minimal valued fields Thanks: The first author was partially supported by ISF grants No. 555/21 and 290/19. The second author was supported by ISF grant No. 555/21. The third author was supported by ISF grant No. 290/19. Yatir Halevi, Assaf Hasson, Ya’acov Peterzil Abstract . We continue our local analysis of groups interpretable in various dp-minimal valued fields, as introduced in [ 8 ] . We associate with every infinite group G G interpretable in those fields an infinite type-definable infinitesimal subgroup ν ⁡ ( G ) \nu(G) , generated by the four infinitesimal subgroups ν D ​ ( G ) \nu_{D}(G) associated with the distinguished sorts K K , k , Γ \Gamma and K / 𝒪 K/{\mathcal{O}} . To show that ν ⁡ ( G ) \nu(G) is type-definable, we show that the resulting subgroups ν D ​ ( G ) \nu_{D}(G) commute with each other as D D ranges over the four distinguished sorts. We then study the basic properties of ν ⁡ ( G ) \nu(G) . Among others, we show that ν ⁡ ( G 1 × G 2 ) = ν ⁡ ( G 1 ) × ν ⁡ ( G 2 ) \nu(G_{1}\times G_{2})=\nu(G_{1})\times\nu(G_{2}) and that if G 1 ≤ G G_{1}\leq G is a definable subgroup then ν ⁡ ( G 1 ) \nu(G_{1}) is relatively definable in ν ⁡ ( G ) \nu(G) . We also discuss possible connections between dp ​ - ​ rk ​ ( ν ⁡ ( G ) ) \mathrm{dp\text{-}rk}(\nu(G)) and elimination of imaginaries. 1. Introduction We revisit the construction of infinitesimal subgroups of groups interpretable in V -minimal expansions of algebraically closed valued fields, T -convex expansions of power bounded o-minimal fields and in (analytic expansions of) p -adically closed fields. 1 1 1 See Remark 2.1 for possible expansions of p -adically closed fields. . We show that this construction, introduced in [ 5 ] and studied axiomatically in [ 8 ] , can be simplified, expanded and consolidated when restricted to the more concrete setting described above. To recall, in [ 5 , 8 , 6 ] , we introduced the construction, for a certain class of interpretable groups, of four infinitesimal subgroups . Each of these infinitesimal subgroups is associated with a different distinguished sort : the valued field K , the value group \Gamma , the residue field k (when infinite) or K/{\mathcal{O}} , the cosets of the valuation ring, and at least one of the four has positive \mathrm{dp\text{-}rk} . Below, these are denoted \nu_{D} , as D ranges over these distinguished sorts. Applications included, for example, the proof in [ 6 ] that a non-abelian definably simple group G is definably isomorphic to either a K -linear group or to a k -linear group. In [ 5 ] and in [ 8 ] infinitesimal subgroups were crucial in characterizing interpretable fields and studying dp-minimal interpretable groups in those classes of valued fields. The key observation underlying these results is the fact that every infinite group G has an infinite definable subset X\subseteq G which is almost strongly internal to at least one of the distinguished sorts. Namely, there is a distinguished sort D and a definable function f:X\to D^{n} (some n ) with finite fibres. For a fixed distinguished sort D , the almost D -rank of G , denoted by \mbox{aD-rk}(G) , is the maximal dp-rank of such a set X . The main result of the present paper is: Theorem 1 . Let {\mathcal{K}} be an expansion of a valued field that is either (i) V -minimal, (ii) power bounded T -convex or (iii) p -adically closed, and let G be an infinite interpretable group in {\mathcal{K}} . (1) To each distinguished sort D corresponds a type-definable subgroup \nu_{D}:=\nu_{D}(G) , with \mathrm{dp\text{-}rk}(\nu_{D}(G))=\mbox{aD-rk}(G) , that is almost strongly internal to D . (2) The group generated by the four \nu_{D}(\widehat{\mathcal{K}}) in a sufficiently saturated model \widehat{\mathcal{K}}\succ{\mathcal{K}} is infinite, type definable, and definably isomorphic to the direct product \nu_{K}(\widehat{\mathcal{K}})\times\nu_{\Gamma}(\widehat{\mathcal{K}})\times\nu_{K/{\mathcal{O}}}(\widehat{\mathcal{K}})\times\nu_{\textbf{k}}(\widehat{\mathcal{K}}). In particular, the four type-definable subgroups commute with each other. This theorem improves the results of [ 8 ] in two main aspects. First, it replaces the construction of the four infinitesimal subgroups with a single type-definable group that we denote \nu(G) . The subgroup \nu(G) encapsulates all the information captured previously by the four infinitesimal subgroups. Moreover, it associates an infinitesimal subgroup with any interpretable group, and not only with a distinguished subclass of such groups. Let us elaborate on the last point. In our previous work, the construction of the infinitesimal groups was carried out only after passing to a quotient of the interpretable group by a finite normal subgroup. In the present work, we eliminate the need for such quotients whenever D is other than K/{\mathcal{O}} in the p -adically closed case (Proposition 3.2 ). In all cases, we provide a uniform description of \nu_{D}(G) as the partial type generated by the collection of all definable sets XX^{-1} as X ranges over all subsets of G almost strongly internal to D of maximal dp-rank (Proposition 4.20 ). The rest of the paper is dedicated to studying the group \nu(G) . In [ 8 ] we asked (in slightly different terms), whether \mathrm{dp\text{-}rk}(\nu(G))=\mathrm{dp\text{-}rk}(G) . In Example 6.4 we show that this need not be true in the context of V -minimal analytic expansions of ACVF 0,0 . We then apply this example to give a short geometric proof of [ 10 , Theorem 1.1] , asserting that analytic expansions of ACVF 0,0 do not eliminate imaginaries down to the so-called geometric sorts of [ 9 ] . It may be worth noting that, assuming elimination of imaginaries down to 0-dimensional sorts (e.g., in pure ACVF 0,0 , RCVF or p -adically closed fields), partial positive solutions to the above question exist. This possible connection with questions of elimination of imaginaries (see also [ 10 , Question 1.5] ) remains unclear, and seems worthy of further investigation. In the last two sections of the paper we initiate a preliminary investigation of the correspondence G\mapsto\nu(G) . We show, among others: Theorem 2 . Let {\mathcal{K}} be an expansion of a valued field of characteristic 0 that is either (i) V -minimal, (ii) power bounded T -convex or (iii) p -adically closed and let G_{1},G_{2} be interpretable infinite groups. Then: (1) If G_{2}\leq G_{1} is a subgroup then \nu(G_{2}) is relatively definable in \nu(G_{1}) . Moreover, if \mathrm{dp\text{-}rk}(G_{1})=\mathrm{dp\text{-}rk}(G_{2}) then \nu(G_{1})=\nu(G_{2}) (Corollary 7.11 , Proposition 7.13 ). (2) \nu(G_{1}\times G_{2})=\nu(G_{1})\times\nu(G_{2}) (Proposition 8.8 ). (3) If f:G_{1}\to G_{2} is a definable surjective homomorphism with finite kernel then f maps \nu(G_{1}) onto \nu(G_{2}) (Proposition 4.22 ). The proof of the results above relies on proving corresponding results for the different distinguished sorts and in fact we have stronger results for the different sorts. For example, when D is not k in the V -minimal case or K/{\mathcal{O}} in the p -adically closed case then in (3), f induces an isomorphism between \nu_{D}(G_{1}) and \nu_{D}(G_{2}) and in (1), \nu_{D}(G_{2})=\nu_{D}(G_{1})\cap G_{2} . Acknowledgment We would like to thank Raf Cluckers, Deirdre Haskell and Moshe Kamensky for helpful discussions regarding

Read the complete text