Sullivan Conjecture

The proved Miller theorem that, for a finite group and finite-dimensional CW complex, the based mapping space from the group's classifying space is weakly contractible, equivalently constant maps give a weak equivalence from the target to the unbased mapping space.

Version

v1 · 2026-08-30 · History

Domain-specific #

2892

Origin domain

algebraic topology

Subdomain

mapping spaces

Aliases

Miller theorem, Sullivan's conjecture on maps from classifying spaces

Core Idea

The Sullivan Conjecture is now a theorem, proved in its mapping-space form by Haynes Miller in 1984. For a finite group \(G\) and a finite-dimensional CW complex \(X\), the based mapping space

\[ \operatorname{Map}_*(BG,X) \]

is weakly contractible: all of its homotopy groups vanish. Equivalently, the inclusion of constant maps

\[ c:X\longrightarrow \operatorname{Map}(BG,X) \]

is a weak homotopy equivalence. Here \(BG\) is the classifying space of \(G\), and the mapping spaces carry the relevant compactly generated/compact-open function-space topology. Miller's Annals paper is the primary proof source and fixes the theorem's historical and mathematical status.^[1]

The counterintuitive force is that \(BG\) is generally infinite-dimensional and topologically rich, yet it admits no nontrivial based mapping-space homotopy into a finite-dimensional target \(X\) under these hypotheses. The surviving named abstraction is therefore not an open prediction. It is a proved rigidity theorem for maps from finite-group classifying spaces, often called Miller's theorem.

Structural Signature

• Finite group: a group \(G\) with classifying space \(BG\).
• Classifying-space source: the source is \(BG\), not an arbitrary CW complex.
• Finite-dimensional target: a CW complex \(X\) of finite dimension under the theorem's standard hypotheses.
• Function-space object: the full space of continuous maps, not merely one homotopy set.
• Based form: \(\operatorname{Map}_*(BG,X)\) is assessed at the constant basepoint map.
• Weak contractibility: every homotopy group of the based mapping space vanishes.
• Unbased form: constant maps define \(X\to\operatorname{Map}(BG,X)\).
• Weak-equivalence conclusion: the unbased map is a weak homotopy equivalence.
• Trivial-action interpretation: \(\operatorname{Map}(BG,X)\) models homotopy fixed points for the trivial \(G\)-action.
• Status invariant: the named conjecture is proved; open or generalized variants must be labeled separately.

This role system locks the basic Miller theorem. Equivariant, \(p\)-complete, locally finite group, or generalized target variants may extend it, but they are not allowed to change the core statement silently.

What It Is Not

It is not an unresolved conjecture. Historical naming does not override proof status. It is not the claim that every continuous map \(BG\to X\) is literally constant as a point-set map; rather, based maps lie in a weakly contractible mapping space, and at the \(\pi_0\) level they are null-homotopic.

It is not a theorem for arbitrary infinite groups or arbitrary infinite-dimensional targets. It is not ordinary fixed-point theory for iterating a self-map. Homotopy fixed points arise from group actions and equivariant mapping spaces, not from convergence to a fixed state under repeated update. It is not the Sullivan conjecture in geometric topology concerning triangulations, nor any unrelated conjecture named for Dennis Sullivan.

Scope of Application

The theorem is foundational in unstable homotopy theory and the study of maps from classifying spaces. It controls mapping spaces \(BG\to X\), supports comparisons between fixed points and homotopy fixed points, and influenced Lannes's \(T\)-functor and equivariant generalizations. An authoritative mathematical encyclopedia states the finite-group, finite-CW-target version and its equivalent unbased form.^[n1]

The core node covers the basic trivial-action result. Generalized Sullivan conjectures can allow nontrivial group actions and compare ordinary fixed points \(X^G\) with homotopy fixed points \(X^{hG}\), often after \(p\)-completion and under additional finiteness or nilpotence conditions. Those results share lineage but have different hypotheses and conclusions. They should be treated as related variants, not evidence that the core identity is an umbrella with interchangeable formulas.

Clarity

Take \(G=\mathbb Z/p\). Then \(BG\) has nontrivial cohomology in arbitrarily high degrees, whereas \(X\) has finite dimension. The theorem says the based mapping space from this large source to \(X\) is weakly contractible. Consequently,

\[ [BG,X]_*=\pi_0\operatorname{Map}_*(BG,X) \]

contains only the constant homotopy class. The stronger mapping-space conclusion also kills higher homotopies among maps.

For unbased maps, each point \(x\in X\) determines a constant map \(BG\to X\). The theorem does not collapse all those constant maps to one component; instead it says the constant-map inclusion recovers \(X\) up to weak homotopy. This distinction prevents the common error of reading “weakly contractible based mapping space” as “unbased mapping space is a point.”

Manages Complexity

Function spaces out of \(BG\) are potentially enormous. The theorem replaces their homotopy-type computation with the target \(X\) in the unbased case and with a weak point in the based case. This converts a difficult infinite-dimensional mapping problem into a rigid structural conclusion.

It also separates ordinary maps from homotopy classes and higher mapping-space structure. Merely proving every based map null-homotopic establishes a \(\pi_0\) statement; Miller's theorem establishes all homotopy-group vanishing. That stronger organization matters in applications involving families of maps and homotopy fixed points.

Abstract Reasoning

Weak contractibility means

\[ \pi_k\operatorname{Map}_*(BG,X)=0\quad\text{for all }k\ge 0. \]

Using the mapping-space adjunction, a \(k\)-sphere in the based mapping space corresponds, under appropriate point-set conventions, to a based map from a smash product such as \(S^k\wedge BG\) into \(X\). The theorem simultaneously rules out this hierarchy of parametrized essential maps.

The unbased equivalence follows from the evaluation fibration intuition: the fiber of evaluation at the basepoint is the based mapping space. If the fiber is weakly contractible, evaluation and the constant-map section identify the weak homotopy type of \(\operatorname{Map}(BG,X)\) with \(X\). This explanation records structure without pretending that every technical point-set hypothesis is automatic.

Knowledge Transfer

The transferable pattern is a structured infinite source becomes mapping-rigid against a finite-dimensional target. It suggests checking source cohomology, target finiteness, and function-space topology rather than judging difficulty from the existence of individual point-set maps.

Literal transfer remains confined to homotopy theory. Claims that a social or computational space is “too complex to map” are metaphors, not instances. The named theorem requires classifying spaces, CW targets, based mapping spaces, and weak homotopy groups.

Examples

1. Cyclic prime group: \(\operatorname{Map}_*(B\mathbb Z/p,X)\) is weakly contractible for finite-dimensional CW \(X\).
2. Finite group: the same basic conclusion holds for \(BG\) with \(G\) finite.
3. Unbased mapping space: constant maps give \(X\simeq_w\operatorname{Map}(BG,X)\).
4. Component consequence: every based map \(BG\to X\) is null-homotopic.
5. Higher-family consequence: parametrized families representing positive homotopy groups of the based mapping space are null.
6. Boundary failure: replacing \(X\) by an arbitrary infinite-dimensional space is not licensed by the core theorem.

Structural Tensions

• Historical name vs. present status. “Conjecture” persists after proof. Diagnostic: label the Miller result as a theorem and date the proof.
• Based vs. unbased mapping spaces. The former is weakly contractible; the latter has the weak type of \(X\). Diagnostic: identify the evaluation basepoint and constant-map section.
• Null homotopy vs. literal constancy. A map can be nonconstant pointwise yet null-homotopic. Diagnostic: state whether equality, homotopy class, or mapping-space homotopy type is asserted.
• Basic theorem vs. generalized family. Equivariant and completed variants change hypotheses. Diagnostic: lock finite \(G\), finite-dimensional \(X\), and the mapping-space conclusion before invoking the core node.
• Autonomous theorem vs. generic Constraint. Constraint language alone does not entail classifying spaces, weak contractibility, or constant-map equivalence. Diagnostic: subtract the parent and require the complete finite-group mapping-rigidity package.

Structural–Framed Character

The structural core is a rigidity result reducing a function space to its constant-map locus up to weak equivalence. The frame is algebraic topology: classifying spaces, CW dimension, based maps, mapping-space topology, homotopy groups, and group actions.

It remains domain-specific. The theorem is too typed to recur literally across unrelated domains, even though “large source, rigid maps” can inspire analogy.

Structural Core vs. Domain Accent

Structural core: a source-target class under exact finiteness conditions, a space of admissible maps, a distinguished constant-map inclusion, and vanishing of every based deformation class.

Domain accent: finite groups, \(BG\), finite-dimensional CW complexes, compact-open function spaces, weak homotopy equivalence, and homotopy fixed points.

Instantiates / Related Primes

The Sullivan Conjecture compositionally presupposes Constraint in the precise sense that it establishes a severe admissibility result for homotopy classes of maps from \(BG\) into \(X\): the based mapping space has only the weak type of a point. Fixed Point is related through the homotopy-fixed-point reformulation, but the accepted prime's iterative self-map identity is not literal enough for a direct parent. Constraint is therefore the minimal accepted-899 endpoint, while the theorem-specific residual remains explicit.

Relationships to Other Abstractions

Current abstraction Sullivan Conjecture Domain-specific

Parents (1) — more general patterns this builds on

• Sullivan Conjecture presupposes Constraint Prime

  The Sullivan Conjecture compositionally presupposes Constraint in the precise sense that it establishes a severe admissibility result for homotopy classes of maps from \(BG\) into \(X\): the based mapping space has only the weak type of.

Hierarchy path (1) — routes to 1 parentless root

• Sullivan Conjecture → Constraint

Neighborhood in Abstraction Space

Sullivan Conjecture sits in a sparse region of the domain-specific corpus (79^th percentile for distinctiveness): few abstractions share its structure, so a faithful description tends to retrieve it precisely.

Family — Topological Groups & Homotopy Actions (11 abstractions)

Nearest neighbors

• Whitehead Theorem — 0.85
• Loop Group — 0.83
• Classifying space — 0.83
• Eilenberg–MacLane space — 0.81
• Direct Sum of Topological Groups — 0.81

Computed from structural-signature embeddings · 2026-09-08

Not to Be Confused With

• An open conjecture: the basic Miller form was proved in 1984.
• Sullivan's fixed-point conjecture variants: equivariant and \(p\)-complete generalizations with added hypotheses.
• Ordinary fixed point: a state unchanged by one self-map.
• Literal absence of maps: continuous maps exist; the conclusion concerns their homotopy type.
• Contractible vs. weakly contractible: the theorem guarantees weak contractibility in its standard statement.
• Other Sullivan conjectures: unrelated problems sharing the mathematician's name.

Notes

[n1] “Sullivan Conjecture,” Encyclopedia of Mathematics, statement of the finite-group mapping-space theorem and equivalent constant-map weak equivalence. ↩

References

[1] Haynes Miller, “The Sullivan Conjecture on Maps from Classifying Spaces,” Annals of Mathematics 120 (1984), 39–87, DOI: 10.2307/2007071. registry ↩