Failure of weak approximation in adjoint groups
Chayansudha Biswas · 2026
Read Failure of weak approximation in adjoint groups by Chayansudha Biswas (2026) 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-2604-14420.txt.
Introduction
Biswas settles a long-open question about weak approximation in adjoint groups in the negative. Platonov conjectured in 1991 that adjoint groups are rational varieties over arbitrary infinite fields, which would imply it. Merkurjev disproved the rationality half in 1996, leaving that question open until now. Biswas closes that gap, showing it fails for adjoint groups.
Catalog record
| CREATOR | Chayansudha Biswas |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 21748 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2604.14420
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Failure of weak approximation in adjoint groups Chayansudha Biswas Abstract . Platonov in 1991 conjectured that adjoint groups are rational as varieties over arbitrary infinite fields, and as a consequence have weak approximation. The rationality part of the conjecture was disproved by Merkurjev in 1996, but the question about weak approximation remained open. We settle this in the negative. 1991 Mathematics Subject Classification 14G12, 11E57, 20G15, 14E08 Introduction A generalization of the Chinese Remainder Theorem is the property of weak approximation of a field: if S is any finite set of inequivalent discrete valuations on a field F , then the image of F is dense in \prod_{v\in S}F_{v} under the diagonal embedding, where F_{v} denotes the completion of F with respect to a valuation v\in S [ 15 , Theorem 1.4] . The topology considered here is the product adic topology. If X is a variety defined over the field F , then one can ask if the image of X(F) is dense in \prod_{v\in S}X(F_{v}) under the diagonal embedding. If this is the case, then we say that X has weak approximation with respect to S . We say that X has weak approximation if it has weak approximation with respect to every finite set of discrete valuations of F . All rational varieties have weak approximation, and hence failure of weak approximation is an obstruction to rationality. Weak approximation in the setting of algebraic groups is classical and well understood in many cases. For instance, it is well established that all semisimple simply connected and semisimple adjoint groups over global fields have weak approximation. The same statement is known to be false (due to work by Platonov in [ 14 ] ) over arbitrary valued fields in the simply connected case, motivating a focus on adjoint groups instead. Platonov conjectured all semisimple adjoint groups over arbitrary fields are rational as varieties, and therefore have weak approximation [ 15 , §7.3] . The rationality part of the conjecture was disproved by Merkurjev in [ 12 ] , where examples were given of adjoint groups that fail rationality. More examples of nonrational adjoint groups can be found in [ 7 ] , [ 1 ] and [ 2 ] . The obstruction to rationality in these examples is the failure of universal R -triviality: for a group G defined over a field F , we say two points x,y\in G(F) are R -equivalent if there is a rational map f\colon\mathbb{A}_{F}^{1}\dashrightarrow G defined at 0 and 1 such that f(0)=x and f(1)=y . This is an equivalence relation on G(F) and we denote by G(F)/R the set of R -equivalence classes. We say the group G is universally R -trivial if G(E)/R=1 for all field extensions E of F . This notion of R -equivalence was introduced by Manin in [ 11 ] . Rationality of (the underlying variety of) a group implies universal R -triviality, and therefore Merkurjev constructed adjoint groups that fail universal R -triviality to obtain nonrational adjoint groups. It was still not known, however, if there exist adjoint groups failing weak approximation. Over arbitrary fields, weak approximation has been shown for many cases of adjoint groups by Thăńg in [ 18 ] , but the question remained open in general. To the author’s knowledge, no further progress has been made towards the conjecture since. In the present article, we will settle the problem by constructing adjoint groups that fail weak approximation (see Theorem 11 , and see Example 12 for an explicit example). For every global field F of characteristic different from 2 in which -1 is a square, we will give a recipe to construct infinitely many adjoint groups, defined over a suitable overfield of F , that fail weak approximation. To construct our counterexamples, we fix such a field F , and using a criterion of Merkurjev we first construct an adjoint group G of type D_{n} over F such that G(F)/R=1 but G(E)/R\neq 1 for some extension E over F . The main trick is to slightly modify the situation above to obtain a field K over F which admits a discrete valuation v such that G(K)/R=1 and G(K_{v})/R\neq 1 . We then use a result of Raghunathan to show that every K_{v} -rational point g of G has a v -adic neighborhood U_{g} contained in the R -equivalence class of that point. In particular, we consider g\in G(K_{v}) not in the R -equivalence class of 1 . Weak approximation in G would imply that U_{g} has nonempty intersection with G(K) . But G(K)/R=1 implies that points in this intersection are R -equivalent to the identity, which gives us a contradiction. 1. Constructing adjoint groups failing universal R -triviality The aim of this section is the construction of adjoint groups that fail universal R -triviality. This is done in Proposition 2 , using a criterion due to Merkurjev. Proposition 2 will later be used in §2 to prove the main theorem. The following proposition is the main technical ingredient: Proposition 1 . Let F be any global field with \operatorname{char}F\neq 2 in which -1 is a square. Then there is a quaternion algebra Q over F and an element d\in~F^{\times}\setminus F^{\times 2} such that d\in-\operatorname{Nrd}(Q) and d\notin\operatorname{Nrd}(Q^{+}) . Here Q^{+} denotes the subspace of Q consisting of the pure quaternions, and \operatorname{Nrd}\colon Q\to F is the reduced norm. For F and Q as in the proposition, let A=M_{n}(Q) , for n\geq 3 an odd integer. Then the hypothesis of [ 12 , Proposition 11] is satisfied, and thus there exists an involution \sigma on A with \operatorname{disc}(\sigma)=d\cdot F^{\times 2} . In the proof of this proposition it is shown that the hypothesis of [ 12 , Theorem 2] is satisfied, and hence the F -group G=\mathbf{PSim}_{+}(A,\sigma) is an adjoint group of type D_{n} failing universal R -triviality. (See [ 1 , p.2] for a definition of this group.) Failure of universal R -triviality means that there exists a field E over F such that G(E)/R\neq 1 . The field E is constructed from an iterated finite tower of fields over F that are either function fields of quadrics or function fields of Severi-Brauer varieties (see [ 12 , Theorem 2] ). This ensures that E is a separably generated extension over F of finite transcendence degree. We will use this fact to prove in the next section that G fails weak approximation over a suitable overfield of F . We summarize this in the following proposition: Proposition 2 . For any global field F with \operatorname{char}F\neq 2 in which -1 is a square, there is an adjoint F -group G of classical type D_{n} , where n is an odd integer \geq 3 , and a separably generated extension E over F of finite transcendence degree such that G(E)/R\neq 1 . Remark . One checks, by following Merkurjev’s construction, that the transcendence degree of E over F can be bounded above by 3n+4 in the above notation. The remainder of this section is devoted to the proof of Proposition 1 . Lemma 3 . Let F be any global field of characteristic different from 2 , and denote by \Omega_{F} the set of