The topology of elementary subgroups detects a normal p-subgroup

← All digests

Family 310 · Topology
Importance 5/5 · Provisional editorial rating
Sources consulted 2026-10-22

1. The question

A finite group can contain many small abelian subgroups. Arrange those subgroups into a geometric object according to inclusion. Can the shape of that object reveal whether the original group has a nontrivial normal subgroup of prime-power order?

Quillen’s conjecture says that if this subgroup complex is contractible, then such a normal subgroup exists. The manuscript claims a stronger result: when the largest normal $p$-subgroup is trivial, the complex has nonzero reduced rational homology.

This turns a question about normality into one about cycles and boundaries. The odd-prime version was already established in work cited by the manuscript. Its all-prime conclusion completes the remaining prime-two cases through explicit cycles, dimension estimates and propagation from simple components to the whole group.

2. Background and essential definitions

The elementary subgroups

Fix a finite group $G$ and a prime $p$. A $p$-subgroup has order a power of $p$. An elementary abelian $p$-group is a product

$$(\mathbb Z/p\mathbb Z)^r,$$

so it is a vector space of dimension $r$ over the field $\mathbb F_p$, viewed as an additive group. Every nonidentity element has order $p$.

Let $\mathcal A_p(G)$ be the partially ordered set of nonidentity elementary abelian $p$-subgroups, ordered by inclusion. Its order complex has one vertex per subgroup, an edge for each strict inclusion, a triangle for each three-term strict chain, and higher simplices for longer chains.

Thus a simplex is not an arbitrary collection of commuting subgroups. It is a nested chain

$$E_1<E_2<\cdots<E_k.$$

The largest normal $p$-subgroup of $G$ is denoted $O_p(G)$, the $p$-core. A subgroup is normal when conjugation by every element of $G$ preserves it. The $p$-core can be trivial even when $G$ has many nontrivial $p$-subgroups.

Contractibility and homology

A space is contractible if it can be continuously deformed to a point. Reduced homology measures holes while removing the unavoidable degree-zero class of a connected nonempty space. A nonzero reduced homology class is a cycle that cannot be filled by a chain of one higher dimension.

Contractibility implies zero reduced homology, but the converse is false for general spaces. The manuscript’s rational nonvanishing claim therefore gives more information than merely proving noncontractibility.

The convention at the empty complex matters. The paper uses augmented reduced homology, with

$$\widetilde H_{-1}(\varnothing;\mathbb Q)=\mathbb Q.$$

If $p$ does not divide $|G|$, the subgroup complex is empty and $O_p(G)=1$. This convention makes the main theorem include that case correctly; the empty space is not treated as a contractible point.

A small example

Take $G=S_3$, the permutation group on three letters. At $p=2$, its three transpositions generate three distinct order-two subgroups, with no inclusions between them. The complex is three isolated points, so

$$\widetilde H_0(\mathcal A_2(S_3);\mathbb Q)\cong\mathbb Q^2.$$

There is no nontrivial normal two-subgroup. At $p=3$, the unique order-three subgroup is normal, and the complex is a single point. These are elementary instances of the two sides of the proposed characterization.

3. Development history and previous partial results

Brown studied the poset of all nonidentity $p$-subgroups and related its Euler characteristic to Sylow-subgroup information. Quillen proved that restricting to elementary abelian subgroups preserves the homotopy type, making the smaller poset sufficient for the topological question.

Quillen formulated the contractibility conjecture in 1978. He proved it for solvable groups, groups of elementary abelian $p$-rank at most two, and groups of Lie type in defining characteristic. In the solvable case with trivial $p$-core, his method gives homology in the largest possible dimension, a stronger property later used in reductions.

Aschbacher and Kleidman settled the almost-simple case. The remaining challenge was not simply to check more simple groups: one must understand how their extensions interact and whether a component’s nonzero homology survives in an ambient group.

Aschbacher and Smith developed reductions for primes greater than five, leaving specified elementary extensions of unitary groups. Piterman and Smith extended the reduction to all odd primes. Díaz Ramos proved the needed unitary statement and deduced the rational-homology assertion for every finite group at odd primes. The source explicitly credits this prior result; its direct unitary argument gives another route within the new framework. Díaz Ramos, version 6

At the prime two, further reductions constrained a minimal counterexample. Piterman and Smith showed that its components must admit particular outer involutions induced by the ambient group. Piterman eliminated additional exceptional and defining-characteristic cases, leaving specified classical components in characteristic at least five.

The new manuscript uses those reductions rather than claiming an independent classification-free proof. Its remaining work constructs cycles for the linear, unitary, symplectic and orthogonal cases, and develops a propagation argument ensuring that their information survives in the full subgroup complex.

4. What the new manuscripts claim

The September 24, 2026 manuscript states that for every finite group $G$ and every prime $p$,

$$O_p(G)=1 \quad\Longrightarrow\quad \widetilde H_*(\mathcal A_p(G);\mathbb Q)\ne0.$$

The star means that at least one degree has nonzero augmented reduced homology. The theorem does not assert that every degree is nonzero, or that the nonzero group always lies in the largest possible degree.

A classical result gives the other direction relevant to contractibility: a nontrivial $p$-core makes the complex contractible. Thus the new nonvanishing statement would settle Quillen’s conjecture without restrictions on primes or composition factors.

The manuscript also records a coefficient consequence. Since the complex is finite, nonzero rational homology comes from a positive free rank in integral homology. The universal coefficient theorem then gives nonzero reduced homology over every field in the same degree. The empty case is handled by augmentation.

Consequently, acyclicity over any chosen field characterizes the existence of a nontrivial normal $p$-subgroup. This field-coefficient conclusion follows from rational nonvanishing and the classical contractibility direction; it is not a claim that all integral homology groups have been calculated.

5. Why the advance matters

Importance: 5/5, provisional. If correct, the theorem completes a central connection between finite-group structure and topology, in the stronger rational-homology form.

The subgroup complex contains only elementary abelian $p$-subgroups and their inclusions. The result says that this restricted information still detects a global property: the existence of a nontrivial normal $p$-subgroup in the entire group.

The rational strengthening supplies an actual homological witness when the core is trivial. It avoids a situation where the space is noncontractible for subtler reasons but all rational homology vanishes.

Historically, the new coverage must be located accurately. The all-odd-prime assertion is credited to the earlier theorem. The present family claims to finish the prime-two analysis and assemble a uniform all-prime conclusion, with cycle-construction tools adapted to the difficult remaining component types.

6. The main idea

Reduce a hypothetical counterexample to controlled components

Assume a counterexample exists and choose one of smallest group order, at a fixed prime. Every smaller group then satisfies the desired homological implication.

Published reduction theorems constrain the simple components of such a counterexample. A component is a quasisimple subnormal subgroup; in the relevant reduced cases the components are simple. Distinct components commute, giving a connection between component topology and centralizers.

A centralizer consists of elements commuting with a specified subgroup. Proper-subgroup induction supplies homology in appropriate centralizers, while a faithful elementary extension of the component supplies the other half of a product construction. Faithfulness here means that the extension acts injectively by conjugation on the component.

The difficult part is to preserve a nonbounding cycle when these local pieces are placed inside the full group. An inclusion of complexes can turn a nonzero class into a boundary, so local nonvanishing alone is insufficient.

Construct cycles from linear frames

For classical groups, a frame decomposes the underlying vector space into lines. When a bilinear or Hermitian form is present, the lines are chosen nondegenerate and mutually orthogonal as appropriate.

Independent scalar signs on these lines give elementary abelian projective subgroups. Their complete subgroup flags support apartment-like chains: combinatorial cycles associated with a coordinate decomposition.

The paper imposes relations between frames differing only within a two-dimensional plane. These relations make boundary contributions cancel. Additional commuting automorphisms introduce extra faces, whose cancellation becomes further linear equations on the frame coefficients.

The construction therefore needs a dimension estimate: there must be more available independent coefficient choices than the added cancellation conditions can remove. The estimate must remain useful as vector-space dimension grows over a fixed finite field.

The manuscript encodes merger trees of frame lines using multilinear Lie superbrackets, an algebraic bookkeeping device for combining pieces with signs. A Poincaré–Birkhoff–Witt argument, together with a bound on the actual relation kernel, converts plane counts into the required lower bound. Supporting subspaces and their form types remain visible throughout the count.

The output is a nonzero cycle on actual full subgroup flags. Merely finding a nonzero parameter vector describing apartments would not suffice if all its resulting simplex coefficients cancelled.

Restrict and propagate without losing the witness

Some cycles are first built in a larger automorphism group. Intersecting each subgroup in a flag with the desired extension can collapse the flag or merge coefficients. The paper selects a supported elementary subgroup with maximal intersection index, fixes a complementary part, and takes a residue before intersecting the remaining flag.

A residue extracts coefficients along a specified part of a chain. In the chosen setting, the restriction remains injective on the relevant residue, so a known nonzero coefficient survives.

At two, the argument chooses a maximal-rank elementary subgroup within a specified family of faithful component configurations. This need not have the largest elementary rank anywhere in $G$. A product with a centralizer cycle has a residue detecting that centralizer class. A deletion argument removes irrelevant overgroups while preserving the detection of nonbounding in the ambient complex.

For odd primes, the cited reduction needs a different output: top-dimensional homology in specified unitary extensions. The paper establishes the relevant rank formula before constructing those cycles. The top-degree argument and the prime-two propagation argument serve different purposes and are not interchangeable.

Symplectic cases permit direct cycle construction in the desired extension. A final small orthogonal case, represented as $\mathrm{PSL}_4(5)$, uses an equivalent-poset construction retaining a join of component and centralizer homology. These cases exhaust the list left by the reductions and exclude a minimal counterexample.

7. Scope, limitations, and open questions

The theorem concerns finite groups and a specified prime. It is not a statement about arbitrary infinite-group subgroup complexes.

The augmented convention is essential when no nontrivial elementary abelian $p$-subgroup exists. In that case the homology witness is in degree $-1$, not an ordinary positive-dimensional hole.

The main conclusion is nonvanishing in some degree. Top-dimensional nonvanishing, often called the Quillen dimension property, is proved where particular reductions require it; it is not asserted universally by the headline theorem.

The argument relies on established component reductions and the structure theory of finite simple groups. Its elementary-looking statement should not obscure that dependence.

Finally, detecting a nontrivial normal $p$-subgroup does not reconstruct that subgroup or the full integral homology. Questions about explicit ranks, torsion, preferred cycle representatives and efficient computation remain finer problems than the existence theorem summarized here.

8. Sources and further reading

Quillen’s original subgroup-poset paper gives the conjecture and foundational cases. Aschbacher–Smith’s component reduction, Piterman–Smith’s equivalent-poset techniques, and Díaz Ramos’s unitary and odd-prime theorem explain the progression to the remaining cases.

Piterman’s 2026 component elimination supplies the recent prime-two reduction used by the manuscript. The source below develops the frame counts, restrictions and propagation arguments that claim to complete the rational-homology theorem.

Source manuscripts

• M1. OpenAI, Rational homology and Quillen’s conjecture (September 24, 2026).

Sources consulted on 22 October 2026 at repository revision 1581d4c. Download the manuscript citations.