Grothendieck Pairs in Profinite Groups

• Grothendieck pairs are embeddings of finitely generated residually finite groups where the induced map on profinite completions is an isomorphism despite the subgroup being proper.
• They expose a critical phenomenon in group theory where finite quotients fail to distinguish between distinct subgroup structures, highlighting limitations of profinite rigidity.
• Construction mechanisms often involve fiber-product techniques and homological conditions, such as the vanishing of H₂, with applications in arithmetic and hyperbolic groups.

Grothendieck pairs are, in their principal contemporary group-theoretic sense, embeddings of finitely generated residually finite groups

$\phi:A\hookrightarrow \Gamma$

for which the induced map on profinite completions

$$\widehat{\phi}:\widehat{A}\xrightarrow{\cong}\widehat{\Gamma}$$

is an isomorphism, even though $A$ may be a proper subgroup of $\Gamma$. In that situation, $A$ and $\Gamma$ have the same finite quotients, so the embedding becomes invisible to profinite completion. This is one of the sharpest possible failures of “recovering a group from its finite quotients,” and it lies at the center of the modern interaction between profinite rigidity, subgroup structure, and representation-theoretic extension phenomena (Jaikin-Zapirain et al., 2024, Bridson et al., 20 Jul 2025, Nyberg-Brodda, 8 Jun 2026).

1. Definition and basic phenomenon

The basic setup is an embedding

$\phi:A\hookrightarrow \Gamma$

between finitely generated residually finite groups such that

$\widehat{\phi}:\widehat{A}\to \widehat{\Gamma}$

is an isomorphism. One convention, used in the arithmetic-hyperbolic literature, reserves the term “Grothendieck pair” for the nontrivial case $A\neq \Gamma$. Another convention allows the embedding to be trivial and then calls the isomorphic case a trivial Grothendieck pair (Nyberg-Brodda, 8 Jun 2026, Jaikin-Zapirain et al., 2024).

The operational content of the definition is most transparent in terms of finite quotients. For an inclusion $P<\Gamma$ with $$\widehat{\phi}:\widehat{A}\xrightarrow{\cong}\widehat{\Gamma}$$0 finitely generated, proving that

$$\widehat{\phi}:\widehat{A}\xrightarrow{\cong}\widehat{\Gamma}$$1

amounts to showing two things: every finite quotient of $$\widehat{\phi}:\widehat{A}\xrightarrow{\cong}\widehat{\Gamma}$$2 is already detected on $$\widehat{\phi}:\widehat{A}\xrightarrow{\cong}\widehat{\Gamma}$$3, and every finite quotient of $$\widehat{\phi}:\widehat{A}\xrightarrow{\cong}\widehat{\Gamma}$$4 extends to a finite quotient of $$\widehat{\phi}:\widehat{A}\xrightarrow{\cong}\widehat{\Gamma}$$5. Thus Grothendieck pairs isolate a very specific form of profinite ambiguity: a proper subgroup can be algebraically quite different from the ambient group while remaining indistinguishable from it by all finite quotients (Bridson et al., 20 Jul 2025).

A related organizational notion is the strong profinite genus. For a finitely generated group $$\widehat{\phi}:\widehat{A}\xrightarrow{\cong}\widehat{\Gamma}$$6, its strong profinite genus consists of isomorphism classes of finitely generated groups $$\widehat{\phi}:\widehat{A}\xrightarrow{\cong}\widehat{\Gamma}$$7 admitting an embedding $$\widehat{\phi}:\widehat{A}\xrightarrow{\cong}\widehat{\Gamma}$$8 such that $$\widehat{\phi}:\widehat{A}\xrightarrow{\cong}\widehat{\Gamma}$$9 is an isomorphism. If $A$0, this embedding is a Grothendieck pair. In this language, Grothendieck pairs are the nontrivial elements of the strong profinite genus realized inside a fixed ambient group (Bridson et al., 20 Jul 2025).

2. Grothendieck’s original perspective and later rigidity notions

Grothendieck’s original viewpoint was not limited to finite quotients. The modern notes “Some remarks on Grothendieck pairs” emphasize that he introduced the subject through a representation-theoretic and categorical framework. A group $A$1 lies in Grothendieck’s class $A$2 if for every Grothendieck pair $A$3, the induced map

$A$4

is a bijection. The class $A$5 is closed under commensurability, inverse limits, and direct products, and it contains all nilpotent groups, groups of $A$6-points of affine group schemes of finite type, and compact Hausdorff groups. Consequently, Grothendieck pairs preserve not only profinite completion but also Bohr compactification and proalgebraic completion (Jaikin-Zapirain et al., 2024).

The same paper separates several rigidity notions that are often conflated.

Notion	Formulation	Role
Grothendieck pair	$A$7 with $A$8	Embedding-based profinite ambiguity
Left Grothendieck rigidity	No nontrivial Grothendieck pair with fixed source $A$9	Property of the source
Right Grothendieck rigidity	No nontrivial Grothendieck pair with fixed target $\Gamma$0	Property of the target
Profinite rigidity	Any finitely generated residually finite group with isomorphic profinite completion is isomorphic	Global completion rigidity

These notions are genuinely different. Left Grothendieck rigidity and right Grothendieck rigidity concern subgroup embeddings, whereas profinite rigidity compares arbitrary groups with the same profinite completion. The distinctions are substantial: finitely generated free groups, surface groups, and $\Gamma$1-arithmetic groups are left Grothendieck rigid; finitely generated residually finite LERF groups are right Grothendieck rigid; all ascending HNN extensions of finitely generated free groups are right Grothendieck rigid; but hyperbolic groups need not be right Grothendieck rigid, and there exist left Grothendieck rigid groups that are not right Grothendieck rigid. Whether there exists a right Grothendieck rigid group that is not left Grothendieck rigid remains open (Jaikin-Zapirain et al., 2024).

3. Construction mechanisms and the role of $\Gamma$2

The classical engine for constructing Grothendieck pairs is the fibre-product construction. Given an epimorphism $\Gamma$3, one forms

$\Gamma$4

If $\Gamma$5 is finitely generated and $\Gamma$6 finitely presented, then $\Gamma$7 is finitely generated by the $\Gamma$8-$\Gamma$9-$A$0 Lemma. A Platonov–Tavgen criterion states that if $A$1 and

$A$2

then the inclusion

$A$3

induces an isomorphism

$A$4

The quotient must therefore have no nontrivial finite quotients and trivial second homology, so $A$5 functions as a decisive obstruction detector in the construction of Grothendieck pairs (Bridson et al., 20 Jul 2025).

A more recent machine, used in the Kleinian and $A$6-orbifold setting, comes from Bridson–Reid. In the specialization recorded for cocompact lattices in $A$7, if $A$8 and

$A$9

then there exist uncountably many pairwise non-isomorphic groups $\Gamma$0 with embeddings

$\Gamma$1

inducing isomorphisms on profinite completions, and infinitely many of these $\Gamma$2 are finitely generated. Here again the vanishing of $\Gamma$3 is the critical hypothesis; the point of subsequent arithmetic work is often to verify or refute that vanishing for specific lattices (Nyberg-Brodda, 8 Jun 2026).

The relatively hyperbolic framework of Bridson–Reid and its later refinement also uses universal central extensions. If $\Gamma$4 is perfect with universal central extension

$\Gamma$5

then

$\Gamma$6

This allows one to replace quotients $\Gamma$7 by superperfect quotients $\Gamma$8, preserving the no-finite-quotients property while forcing the homological condition required for the Platonov–Tavgen criterion (Bridson et al., 20 Jul 2025).

4. Arithmetic and $\Gamma$9-orbifold realizations

A particularly explicit realization occurs in the arithmetic lattices associated to the Weeks manifold $\phi:A\hookrightarrow \Gamma$0. With $\phi:A\hookrightarrow \Gamma$1, $\phi:A\hookrightarrow \Gamma$2 the image of the norm-one units of a maximal order, and $\phi:A\hookrightarrow \Gamma$3 its normalizer in $\phi:A\hookrightarrow \Gamma$4, one has

$\phi:A\hookrightarrow \Gamma$5

together with

$\phi:A\hookrightarrow \Gamma$6

The most obvious attempt to construct Grothendieck pairs in $\phi:A\hookrightarrow \Gamma$7 is therefore to use $\phi:A\hookrightarrow \Gamma$8 and $\phi:A\hookrightarrow \Gamma$9 in the Bridson–Reid criterion. That route fails because

$\widehat{\phi}:\widehat{A}\to \widehat{\Gamma}$0

so the required homological vanishing is absent (Nyberg-Brodda, 8 Jun 2026).

The same work then supplies a workaround. It introduces two $\widehat{\phi}:\widehat{A}\to \widehat{\Gamma}$1-extension orbifold groups

$\widehat{\phi}:\widehat{A}\to \widehat{\Gamma}$2

sitting between $\widehat{\phi}:\widehat{A}\to \widehat{\Gamma}$3 and $\widehat{\phi}:\widehat{A}\to \widehat{\Gamma}$4. These satisfy

$\widehat{\phi}:\widehat{A}\to \widehat{\Gamma}$5

and, unlike $\widehat{\phi}:\widehat{A}\to \widehat{\Gamma}$6,

$\widehat{\phi}:\widehat{A}\to \widehat{\Gamma}$7

The Bridson–Reid theorem can therefore be applied with

$\widehat{\phi}:\widehat{A}\to \widehat{\Gamma}$8

yielding uncountably many non-isomorphic groups $\widehat{\phi}:\widehat{A}\to \widehat{\Gamma}$9 with embeddings

$A\neq \Gamma$0

that induce isomorphisms of profinite completions, with infinitely many $A\neq \Gamma$1 finitely generated. In this arithmetic setting, second homology literally decides whether the Grothendieck-pair machine can be run (Nyberg-Brodda, 8 Jun 2026).

5. Direct products, relatively hyperbolic groups, and uncountable ambiguity

The broader picture developed in later notes is that Grothendieck pairs are the natural mechanism by which profinite rigidity collapses for direct products. Starting from a finitely presented, non-elementary, relatively hyperbolic group $A\neq \Gamma$2 with

$A\neq \Gamma$3

and infinitely many missing prime torsion orders, and a group $A\neq \Gamma$4 mapping onto a finite-index subgroup of $A\neq \Gamma$5, one can construct infinitely many quotients $A\neq \Gamma$6 such that each $A\neq \Gamma$7 has no nontrivial finite quotients and

$A\neq \Gamma$8

The associated fibre products

$A\neq \Gamma$9

then satisfy

$P<\Gamma$0

so each inclusion $P<\Gamma$1 is a Grothendieck pair. By varying the sequence of primes introduced into the quotient construction, one obtains uncountably many pairwise non-isomorphic subgroups $P<\Gamma$2 with the same profinite completion as the ambient product (Bridson et al., 20 Jul 2025).

Two examples emphasized in this setting are the Weeks manifold group $P<\Gamma$3 and the fundamental group $P<\Gamma$4 of the $P<\Gamma$5-fold cyclic branched cover of the figure-$P<\Gamma$6 knot. In both cases, the individual group $P<\Gamma$7 is absolutely profinitely rigid, but $P<\Gamma$8 admits uncountably many Grothendieck pairs. More strongly, for these examples, if $P<\Gamma$9 is finitely generated and residually finite with

$$\widehat{\phi}:\widehat{A}\xrightarrow{\cong}\widehat{\Gamma}$$00

then there exists an embedding

$$\widehat{\phi}:\widehat{A}\xrightarrow{\cong}\widehat{\Gamma}$$01

inducing that profinite isomorphism. Thus all profinite ambiguity for the square is internal: it is completely accounted for by Grothendieck pairs inside $$\widehat{\phi}:\widehat{A}\xrightarrow{\cong}\widehat{\Gamma}$$02 (Bridson et al., 20 Jul 2025).

This phenomenon resolves a common misconception. Absolute profinite rigidity of $$\widehat{\phi}:\widehat{A}\xrightarrow{\cong}\widehat{\Gamma}$$03 does not imply any comparable rigidity for $$\widehat{\phi}:\widehat{A}\xrightarrow{\cong}\widehat{\Gamma}$$04. The direct product can have a vast strong profinite genus even when the factors are rigid in isolation. In the examples above, direct products are precisely where the ambient hyperbolic or arithmetic rigidity breaks down (Bridson et al., 20 Jul 2025, Nyberg-Brodda, 8 Jun 2026).

6. Terminological heterogeneity in other fields

Although the group-theoretic meaning is now the dominant one in profinite rigidity, the phrase “Grothendieck pairs” is not uniform across mathematics. In algebraic combinatorics, “A bijective proof of the Cauchy identity for Grothendieck polynomials” uses “Grothendieck pairs” for pairs

$$\widehat{\phi}:\widehat{A}\xrightarrow{\cong}\widehat{\Gamma}$$05

of the same Young shape, where $$\widehat{\phi}:\widehat{A}\xrightarrow{\cong}\widehat{\Gamma}$$06 is a set-valued column-strict tableau and $$\widehat{\phi}:\widehat{A}\xrightarrow{\cong}\widehat{\Gamma}$$07 is a reverse plane partition. These pairs encode the summands of

$$\widehat{\phi}:\widehat{A}\xrightarrow{\cong}\widehat{\Gamma}$$08

and the paper constructs explicit algorithms $$\widehat{\phi}:\widehat{A}\xrightarrow{\cong}\widehat{\Gamma}$$09 and $$\widehat{\phi}:\widehat{A}\xrightarrow{\cong}\widehat{\Gamma}$$10 on such pairs to prove a finite Cauchy identity. This usage is entirely unrelated to profinite completion or subgroup embeddings (Numata, 2016).

By contrast, some papers associated with Grothendieck-type terminology do not define any notion called a Grothendieck pair at all. The Banach-space paper “1-Grothendieck $$\widehat{\phi}:\widehat{A}\xrightarrow{\cong}\widehat{\Gamma}$$11 spaces” studies the quantitative Grothendieck property and develops pairwise separation principles for sequences of Radon probability measures, but it does not introduce a pair notion under that name (Lechner, 2015). Likewise, the thesis “Hearts of t-structures which are Grothendieck or module categories” concerns torsion pairs and Grothendieck hearts; the phrase “Grothendieck pairs” does not appear formally there, although a natural nearby notion is a torsion pair $$\widehat{\phi}:\widehat{A}\xrightarrow{\cong}\widehat{\Gamma}$$12 whose HRS heart $$\widehat{\phi}:\widehat{A}\xrightarrow{\cong}\widehat{\Gamma}$$13 is a Grothendieck category (Parra, 2014).

The term is therefore domain-sensitive. In geometric group theory and profinite rigidity it denotes embeddings invisible to profinite completion; in algebraic combinatorics it denotes same-shape tableau pairs for Grothendieck polynomial identities; and in adjacent literatures the relevant “pair” structures may be present without that terminology.