Filtered Whitehead Theorem

• Filtered Whitehead theorem is a criterion where filtration data replaces isolated homotopy invariants to determine equivalences.
• It applies to non-compact spaces, persistent functors, and filtered CW complexes by adapting classical homotopical criteria using compact exhaustions and interleaving methods.
• The approach leverages inverse systems, derived limits, and filtered CW approximations to stabilize obstructions and rectify homotopy equivalences.

A filtered Whitehead theorem is a Whitehead-type criterion in which equivalence is detected not by the homotopy groups of a single space in isolation, but by homotopy invariants attached to a filtration. In current usage, the phrase covers several distinct but related settings: fine shape for non-compact spaces filtered by compacta, persistent spaces modeled as functors $\mathbb{R}^m \to S$, and filtered topological or stable-homotopy categories built from filtered CW complexes. In each case, the classical template is retained—homotopical data determine equivalence—but the invariants, the notion of morphism, and the required hypotheses are adapted to the filtered context (Melikhov, 2022, Lanari et al., 2020, Miller, 5 May 2025).

1. Classical template and the meaning of “filtered”

The classical Whitehead theorem for CW complexes states that a map between CW complexes is a homotopy equivalence if and only if it induces isomorphisms on all homotopy groups, including $\pi_1$ and $\pi_0$. The filtered variants replace this single-space criterion by one adapted to a filtration, and they do so in three materially different ways.

In fine shape, the filtration is intrinsic to a non-compact, locally compact space: one filters by compact subsets and assembles invariants as direct limits over compacta. The relevant homotopy groups are the Steenrod–Sitnikov groups, and the theorem identifies when a fine shape morphism is a fine shape equivalence (Melikhov, 2022).

In persistence theory, the filtered object is a functor $X:\mathbb{R}^m \to S$, with shifts and interleavings replacing ordinary equivalence. Here the Whitehead principle becomes quantitative: if a morphism induces interleavings on persistent homotopy groups, then the underlying persistent spaces are interleaved in the homotopy category, and after rectification they are homotopy interleaved with explicit constants (Lanari et al., 2020).

In filtered topology and filtered stable homotopy, the filtration is given levelwise by subspaces $X(r)$, together with structure maps $X(s)\subset X(r)$ for $s<r$. The paper on filtered topology and persistence in stable homotopy does not explicitly state a theorem named “Filtered Whitehead theorem,” but it gives the closest filtered Whitehead-type criteria through filtered CW approximation and through triangulated persistence notions such as $r$-isomorphisms and $r$-acyclic cones (Miller, 5 May 2025).

2. The fine-shape filtered Whitehead theorem

The most literal theorem bearing the name in the supplied literature is the Whitehead-type theorem in fine shape. Let $X$ and $\pi_1$0 be locally connected, finite-dimensional, locally compact, separable metrizable spaces whose Steenrod–Sitnikov $\pi_1$1 and $\pi_1$2 are trivial. A fine shape morphism $\pi_1$3 is a fine shape equivalence if and only if it induces isomorphisms

$\pi_1$4

for all $\pi_1$5. The paper also proves a dimension-bounded form: if $\pi_1$6 and $\pi_1$7 induces bijections on $\pi_1$8 for $\pi_1$9 and a surjection on $\pi_0$0, then $\pi_0$1 is a fine shape equivalence (Melikhov, 2022).

The notion of fine shape morphism is formulated through approaching maps between absolute retract models. For a metrizable space $\pi_0$2 with closed subsets $\pi_0$3, one chooses an absolute retract $\pi_0$4 and closed AR-subsets $\pi_0$5 with each $\pi_0$6 embedded as a closed $\pi_0$7-set. For corresponding data on $\pi_0$8, a map $\pi_0$9 is $X:\mathbb{R}^m \to S$0–$X:\mathbb{R}^m \to S$1-approaching if it sends sequences accumulating in $X:\mathbb{R}^m \to S$2 to sequences accumulating in $X:\mathbb{R}^m \to S$3. A fine shape morphism is such an approaching map, respecting the filtrations, taken up to approaching homotopy; a fine shape equivalence is an invertible morphism in this homotopy class.

For locally compact spaces, fine shape reduces to strong antishape. Operationally, the theory filters a non-compact space by its compact subsets and then forms appropriate pro- and ind-objects of homotopy and homology. If $X:\mathbb{R}^m \to S$4 is locally compact and metrizable, the Steenrod–Sitnikov invariants are defined by

$X:\mathbb{R}^m \to S$5

where the colimits range over compacta containing $X:\mathbb{R}^m \to S$6 in the pointed case. These $X:\mathbb{R}^m \to S$7 are groups for $X:\mathbb{R}^m \to S$8, abelian for $X:\mathbb{R}^m \to S$9, and the relative groups $X(r)$0 are groups for $X(r)$1, abelian for $X(r)$2. They satisfy long exact sequences and the usual module actions in a filtered form.

The theorem is “filtered” because equivalence in fine shape is characterized by invariants computed from the compact-exhaustion filtration. In the compact case, strong shape Whitehead theorems do not require this filtration; in the non-compact case, the filtration by compacta is intrinsic (Melikhov, 2022).

3. Algebraic mechanism, derived limits, and proof strategy in fine shape

A central feature of the fine-shape theorem is the role of inverse systems, derived limits, and stabilization. When a compactum $X(r)$3 is represented as an inverse limit $X(r)$4 of compact polyhedra, Quigley’s short exact sequence gives

$X(r)$5

while Milnor’s short exact sequence gives

$X(r)$6

These sequences identify the $X(r)$7 and $X(r)$8 obstructions that must be controlled in order to pass from homotopy-group information to equivalence (Melikhov, 2022).

The cornerstone algebraic result is an ind-triviality theorem. If a direct sequence of groups

$X(r)$9

has trivial colimit, then it is trivial as an ind-group—meaning that for every $X(s)\subset X(r)$0 there exists $X(s)\subset X(r)$1 such that $X(s)\subset X(r)$2 is the zero map—provided the sequence has one of the prescribed forms built from $X(s)\subset X(r)$3 or $X(s)\subset X(r)$4 of inverse systems satisfying countability, finite generation, or Mittag–Leffler hypotheses. In the paper’s concise formulation,

$X(s)\subset X(r)$5

This algebra supplies the stabilization needed to kill derived-limit obstructions.

A byproduct is a characterization of vanishing Steenrod–Sitnikov homology for a locally compact separable metrizable space $X(s)\subset X(r)$6: $X(s)\subset X(r)$7 For an exhaustion $X(s)\subset X(r)$8 with $X(s)\subset X(r)$9 and $s<r$0, this becomes

$s<r$1

This criterion follows from Milnor’s exact sequence together with the algebraic lemma.

The geometric proof of the fine-shape Whitehead theorem proceeds by passing to filtered AR models, representing the morphism by a two-parameter $s<r$2-level map, and then analyzing relative mapping cylinders $s<r$3. Exact sequences for the mapping telescopes show that the relevant $s<r$4 and $s<r$5 terms have trivial colimit in the range required by the dimension bound. Local connectedness and trivial Čech $s<r$6 allow reduction to an “abelian, finitely generated” regime by coning off $s<r$7-skeleta. The algebraic lemma then yields ind-triviality, after which one kills skeleta inductively and constructs an inverse in fine shape by homotoping the mapping cylinders into the distinguished subspaces. The conclusion is that $s<r$8 and $s<r$9 are fine-shape homotopic to the identity, up to the shift in the telescope, which is itself fine-shape homotopic to the identity (Melikhov, 2022).

4. Persistent and interleaving-based filtered Whitehead theorems

In persistence theory, a filtered space is a functor $r$0, where $r$1, and shifts are defined by $r$2. An $r$3-interleaving between $r$4 and $r$5 consists of natural transformations $r$6 and $r$7 satisfying the standard interleaving identities. Three homotopy-invariant relaxations are compared: homotopy interleaving, interleaving in the homotopy category, and homotopy-commutative interleaving, with distances $r$8, $r$9, and $r$0 satisfying

$r$1

For $r$2, these distances are preserved under Quillen equivalence (Lanari et al., 2020).

The persistent Whitehead theorem is formulated in terms of persistent homotopy groups. For $r$3, $r$4, and $r$5,

$r$6

If there exists $r$7 in $r$8 that induces $r$9-interleavings in all persistent homotopy groups and in $X$0, then, under the stated dimension and cofibrancy hypotheses, $X$1 and $X$2 are $X$3-interleaved in $X$4. More precisely, if $X$5, $X$6 and $X$7 are projective cofibrant and pointwise $X$8-skeletal; if $X$9, $\pi_1$00 and $\pi_1$01 are persistent CW-complexes of dimension $\pi_1$02. Under these assumptions,

$\pi_1$03

A second theorem rectifies homotopy-commutative interleavings. If $\pi_1$04 and $\pi_1$05 are $\pi_1$06-homotopy-commutative interleaved, then they are $\pi_1$07-homotopy interleaved for every $\pi_1$08. Combining this rectification theorem with the persistent Whitehead theorem yields the filtered/persistent Whitehead corollary: under the same hypotheses, $\pi_1$09 and $\pi_1$10 are homotopy interleaved with parameter $\pi_1$11 for any constant $\pi_1$12. Equivalently,

$\pi_1$13

The proof uses the projective model structure on $\pi_1$14, discretization to $\pi_1$15, parity endofunctors, and the smothering property of the functor $\pi_1$16. On the Whitehead side, Jardine’s homotopy lifting lemma and a right lifting property up to controlled shift convert interleavings of persistent homotopy groups into interleavings in the homotopy category. On the rectification side, homotopy-commutative interleavings are strictified with constant $\pi_1$17, and the passage back to $\pi_1$18 yields the final $\pi_1$19 formulation (Lanari et al., 2020).

5. Filtered CW approximation, $\pi_1$20-isomorphisms, and triangulated persistence

The paper on filtered topology and persistence in stable homotopy does not explicitly state a theorem named “Filtered Whitehead theorem.” Its closest statements are the filtered CW approximation lemma and the triangulated persistence criterion based on $\pi_1$21-acyclic cones (Miller, 5 May 2025).

A filtered space is given by a collection of pointed subspaces

$\pi_1$22

such that $\pi_1$23 for $\pi_1$24, together with lower and upper bounds $\pi_1$25 and $\pi_1$26. A morphism of filtered spaces is a family of maps

$\pi_1$27

commuting with the inclusion maps, and composition satisfies

$\pi_1$28

Filtered homotopies are defined levelwise through the filtered reduced cylinder, and the $\pi_1$29-th filtered homotopy group is

$\pi_1$30

A filtered CW complex is a filtered space whose level sets are sub-CW-complexes and whose inclusion maps are cellular inclusions. The filtered CW approximation lemma states that for any filtered space $\pi_1$31 there exists a filtered CW complex $\pi_1$32 and a map $\pi_1$33 of shift zero such that every level map

$\pi_1$34

is a weak equivalence and the squares with the filtration maps commute for all $\pi_1$35. The paper explicitly presents this as a levelwise Whitehead-type situation: by classical Whitehead, each $\pi_1$36 is a homotopy equivalence. The paper then states that the extra filtered commutativity allows one to assemble these levelwise homotopy inverses and homotopies into a filtered homotopy equivalence. Since the paper does not formalize this as a standalone theorem, a cautious formulation is that the lemma suggests a filtered analogue of Whitehead’s theorem for shift-zero maps between filtered CW complexes.

The same paper develops a persistence Spanier–Whitehead category $\pi_1$37 and a filtered stable homotopy category $\pi_1$38. In this setting, an object is $\pi_1$39-acyclic if $\pi_1$40, and a morphism is an $\pi_1$41-isomorphism if it fits into an exact triangle

$\pi_1$42

with $\pi_1$43. The paper states explicitly that an $\pi_1$44-isomorphism represents an isomorphism in the limit category $\pi_1$45. It also proves that $\pi_1$46 is a triangulated persistence category and that $\pi_1$47 satisfies the same TPC structure, with exact triangles whose cones are $\pi_1$48-acyclic. This is a Whitehead-type criterion in triangulated persistence form: equivalence is detected by the acyclicity of the cone, rather than by ordinary unstable homotopy groups.

6. Hypotheses, counterexamples, and scope

The filtered Whitehead principle is not hypothesis-free. In fine shape, local connectedness is essential in the non-compact case. The paper gives a comb-and-flea example: if $\pi_1$49 is the “comb-and-flea” set and $\pi_1$50 for a thin extra set $\pi_1$51, then $\pi_1$52, so $\pi_1$53 is not fine-shape equivalent to a point, yet $\pi_1$54 for all $\pi_1$55. After adding a small ray to obtain a locally compact space $\pi_1$56 homotopy equivalent to $\pi_1$57, one still has $\pi_1$58 for all $\pi_1$59, but $\pi_1$60 is not fine-shape equivalent to a point. This shows that isomorphisms on Steenrod–Sitnikov homotopy groups do not imply fine shape equivalence without local connectedness (Melikhov, 2022).

In the persistent setting, quantitative control has sharp limits. Rectification from homotopy-commutative interleavings to homotopy interleavings holds with multiplicative constant $\pi_1$61, but in spaces any inequality of the form $\pi_1$62 forces $\pi_1$63. Moreover, for $\pi_1$64 there is no constant $\pi_1$65 such that $\pi_1$66-homotopy-commutative interleaving implies $\pi_1$67-homotopy interleaving for all $\pi_1$68. The theorem therefore gives a one-parameter quantitative Whitehead principle, but not a uniform multiparameter rectification theorem (Lanari et al., 2020).

In filtered topology and filtered stable homotopy, the weighted Euler characteristic is a filtered homotopy invariant and induces the isomorphism

$\pi_1$69

through

$\pi_1$70

However, the paper also emphasizes that the weighted Euler polynomial detects only filtered homotopy type up to a coarse invariant and does not fully detect equivalence in the triangulated categories. The example comparing the torus $\pi_1$71 and $\pi_1$72 with trivial filtrations shows that stabilization can collapse distinctions that remain visible in the unstable filtered category (Miller, 5 May 2025).

Taken together, these results indicate that “Filtered Whitehead theorem” is best understood as a family of theorems and criteria rather than a single statement. In fine shape, it is an equivalence criterion for non-compact spaces filtered by compacta. In persistence, it is a quantitative theorem turning interleavings of persistent homotopy groups into interleavings of filtered objects. In filtered topology and stable homotopy, it appears through filtered CW approximation and triangulated persistence criteria based on acyclic cones. The common theme is that filtration data are not auxiliary: they determine both the correct invariants and the correct notion of equivalence.