03.13.08 · modern-geometry / spectral-sequences

The telescope conjecture and its disproof

shipped3 tiersLean: none

Anchor (Master): Ravenel 1984 *Localization with respect to certain periodic homology theories* (Amer. J. Math. 106); Miller-Ravenel-Wilson 1977 *Periodic phenomena in the Adams-Novikov spectral sequence* (Ann. of Math. 106); Hopkins-Smith 1998 *Nilpotence and stable homotopy theory II* (Ann. of Math. 148); Burklund-Hahn-Levy-Schlank-Yuan 2023 *K-theoretic counterexamples to Ravenel's telescope conjecture* (arXiv:2310.17459)

Intuition Beginner

The homotopy groups of spheres are the deepest invariants in topology, and they hide repeating patterns. Some elements come in infinite families that march off to higher and higher dimensions in a regular rhythm. These periodic families are organised by a number called the height: height one is the simplest rhythm, height two is a more intricate one, and so on. The telescope conjecture is a question about whether two natural recipes for isolating a single rhythm produce the same answer.

The first recipe is mechanical. Take a small space that only carries the height- $n$ rhythm, find the self-map that shifts it by one beat, and apply that shift over and over forever. Stacking the copies end to end builds a long tube, the telescope, that captures every beat of the rhythm. The second recipe is more refined. There is a special measuring theory, Morava $K$ -theory at height $n$ , that is designed to detect exactly that rhythm and nothing else. Filtering a space through this theory isolates the same periodic information in a cleaner package.

For thirty years almost everyone expected these two recipes to agree. The conjecture said: the mechanical telescope and the refined Morava filtration give the same answer. It was checked and proved for the simplest rhythm. In 2023 a team of five mathematicians showed it is false for every rhythm from height two onward. The mechanical telescope sees strictly more than the refined theory does.

Visual Beginner

Picture a short strip of patterned wallpaper, the small finite space carrying one rhythm. The self-map is a rule that slides the pattern along by exactly one tile. Apply the slide again and again and glue each shifted copy onto the last: the strip grows into a very long roll, the telescope, that records the pattern forever. That is the first recipe drawn out as a picture.

The refined recipe is the tower on the right: the same space passed through Morava $K$ -theory, which keeps only the height- $n$ beat. The conjecture is the single arrow joining the roll to the tower, asking whether the two sides match. At the lowest height the arrow is an equality. From height two up, the crossed-out equals sign records the recent discovery: the roll carries extra information the tower throws away.

Worked example Beginner

Walk through the lowest height, where the conjecture is a theorem, to feel what the two recipes do.

Step 1. Start at height one, working one prime at a time, say the prime two. The small space carrying the height-one rhythm is the mod-two Moore space: a circle with a disc glued on by a degree-two map. It has just enough room for the simplest periodic family.

Step 2. The height-one self-map is multiplication by a class often written $v_{1}$ , which on this Moore space is a map that shifts dimension by eight. Applying it forever builds the telescope. What this telescope records is the image-of- $J$ pattern, the oldest known infinite family in the homotopy of spheres, repeating with period eight.

Step 3. The refined recipe filters the same Moore space through Morava $K$ -theory at height one, which at the prime two is a version of mod-two $K$ -theory. This also isolates the image-of- $J$ pattern.

Step 4. Compare. Miller and Mahowald computed both sides in the early 1980s and found they match exactly. The telescope of the height-one self-map and the Morava filtration give the same answer, at every prime. So at height one the two recipes agree, and the conjecture holds.

What this tells us: at the simplest height the mechanical and refined recipes capture the same image-of- $J$ rhythm, which is why everyone expected the agreement to continue. The surprise is that it stops being so at the next height.

Check your understanding Beginner

Formal definition Intermediate+

Fix a prime $p$ and work in the $p$ -local stable homotopy category. A finite $p$ -local spectrum $F$ has type $n$ when its Morava $K$ -theories satisfy $K (m)_{*} F = 0$ for $m < n$ and $K (n)_{*} F \neq = 0$ . By the Hopkins-Smith periodicity theorem 03.12.45, such an $F$ admits a $v_{n}$ -self-map, an essentially unique (up to iteration) map $$ f : \Sigma^d F \longrightarrow F $$ inducing an isomorphism on $K (n)_{*}$ and the zero map on $K (m)_{*}$ for $m \neq = n$ .

Definition (telescope). The telescope of $F$ is the homotopy colimit of the iterated self-map, $$ T(n) = v_n^{-1} F = \operatorname{hocolim}\left( F \xrightarrow{,f,} \Sigma^{-d} F \xrightarrow{,f,} \Sigma^{-2d} F \to \cdots \right). $$ Equivalently $T (n)$ is the Bousfield localisation $L_{T (n)} F$ at the homology theory represented by the telescope spectrum; the Bousfield class $⟨ T (n)⟩$ depends only on $n$ , not on the choice of type- $n$ complex $F$ or of self-map $f$ .

Definition ( $K (n)$ -localisation). Let $K (n)$ be the height- $n$ Morava $K$ -theory, with coefficient ring $K (n)_{*} = F_{p} [v_{n}^{\pm 1}]$ and $∣ v_{n} ∣ = 2 (p^{n} - 1)$ . The $K (n)$ -localisation of $F$ is the Bousfield localisation $L_{K (n)} F$ at $K (n)$ 03.12.48.

The self-map $f$ becomes invertible after $K (n)$ -localisation, so the universal property of $L_{T (n)}$ supplies a natural comparison map $$ T(n) ;=; v_n^{-1}F \xrightarrow{;;c;;} L_{K(n)}F . $$

Definition (telescope conjecture). Ravenel's telescope conjecture asserts that $c$ is an equivalence for every type- $n$ finite complex $F$ — equivalently $⟨ T (n)⟩ = ⟨ K (n)⟩$ as Bousfield classes, equivalently the finite localisation $L_{n}^{f}$ and the ordinary chromatic localisation $L_{n}$ coincide. In words: $v_{n}$ -periodic homotopy is computed by $K (n)$ -local homotopy.

Counterexamples to common slips

The comparison map always exists and is always a $K (n)_{*}$ -equivalence; the content of the conjecture is whether it is an equivalence of spectra, not merely a $K (n)_{*}$ -iso. Failure means the two sides have the same $K (n)$ -homology but different homotopy types.
$T (n)$ depends only on the height $n$ through its Bousfield class, but it is not a finite spectrum: it is an infinite colimit, and its homotopy groups are the $v_{n}$ -periodic homotopy of $F$ .
The conjecture is a statement about localisation functors on the whole category, packaged as $L_{n}^{f} = L_{n}$ ; the per-complex formulation $T (n) ≃ L_{K (n)} F$ is the same assertion evaluated on type- $n$ objects.
Height zero is the rational case ( $T (0) = L_{K (0)} =$ rationalisation) and holds for degenerate reasons. The first genuine test is height one.

Key theorem with proof Intermediate+

Theorem (height-one telescope theorem; Miller, Mahowald). At every prime $p$ , the comparison map $c : T (1) \to L_{K (1)} F$ is an equivalence for every type- $1$ finite complex $F$ . The telescope conjecture holds at height $n = 1$ .

Proof. It suffices to treat one type- $1$ complex, the mod- $p$ Moore spectrum $V (0) = S / p$ , since the Bousfield class $⟨ T (1)⟩$ is independent of the chosen complex and self-map. At odd primes the $v_{1}$ -self-map on $V (0)$ has degree $2 (p - 1)$ ; at $p = 2$ one uses $V (0) \land V (0)$ or the degree-eight self-map, with the same conclusion.

The strategy is to compute both sides and match them. The $K (1)$ -local homotopy $π_{*} L_{K (1)} V (0)$ is accessible through the $K (1)$ -local Adams-Novikov spectral sequence 03.12.38, whose input is the continuous cohomology of the height-one Morava stabiliser group $S_{1} = Z_{p}^{\times}$ acting on the Lubin-Tate ring. This computes the image-of- $J$ pattern with its period $2 (p - 1) p^{k}$ refinements.

The telescope side is computed by a different resolution. Mahowald's $b o$ -resolution at $p = 2$ and Miller's localised Adams spectral sequence at odd primes ^{[Miller 1981]} ^{[Mahowald 1981]} present $π_{*} T (1) = v_{1}^{- 1} π_{*} V (0)$ as the homology of an explicit small complex. The output is, term for term, the same image-of- $J$ pattern.

The two spectral sequences converge to groups of the same size in each degree, and the comparison map $c$ induces a map between them that is the identity on the common $E_{\infty}$ -pattern. A map of spectra inducing an isomorphism on all homotopy groups is an equivalence. Hence $c$ is an equivalence and the conjecture holds at height one. $□$

Bridge. This height-one calculation builds toward the entire chromatic programme and is exactly the encouraging evidence that made the general conjecture so widely believed. The foundational reason the height-one case works is that the $v_{1}$ -periodic homotopy is small enough to compute on both sides by hand, and the answer is the classical image of $J$ 03.08.11 that predates chromatic language entirely. This is exactly the pattern that the chromatic spectral sequence 03.13.06 organises height by height: each Morava layer is meant to be the $K (n)$ -local homotopy, and the telescope conjecture is the claim that the mechanical $v_{n}$ -periodic homotopy generalises this agreement to every height. The conjecture is dual to the nilpotence theorem 03.12.45 in the chromatic dictionary — nilpotence controls what dies, periodicity controls what survives — and putting these together gives the central insight of the field: stable homotopy is filtered by height, and the open question for forty years was whether the filtration is detected by Morava $K$ -theory exactly or only up to the extra information the telescope retains. The bridge from height one to higher heights is the comparison map $c$ , and the surprise recorded below is that this bridge collapses from height two onward.

Exercises Intermediate+

Exercise 1 (easy, short-answer).

State the three equivalent formulations of the telescope conjecture given in the formal definition, and explain why they say the same thing.

Hint

One is a statement about a single comparison map; one is about Bousfield classes; one is about two localisation functors.

Answer

(i) The comparison map $c : T (n) \to L_{K (n)} F$ is an equivalence for every type- $n$ finite $F$ . (ii) The Bousfield classes agree, $⟨ T (n)⟩ = ⟨ K (n)⟩$ . (iii) The finite localisation equals the ordinary localisation, $L_{n}^{f} = L_{n}$ . They coincide because $T (n)$ is the telescope summand of $L_{n}^{f}$ at height $n$ while $K (n)$ detects the height- $n$ layer of $L_{n}$ ; an equivalence of the per-height monochromatic layers for all heights $\leq n$ is the same as equality of the two localisation functors, and a single Bousfield-class equality at each height packages the comparison map being an equivalence. Rubric: full credit for stating all three and giving the layer-by-layer reason they match.

Exercise 2 (easy, multiple choice).

Why is the comparison map $c : T (n) \to L_{K (n)} F$ guaranteed to exist?

A. Because $T (n)$ and $L_{K (n)} F$ are both finite spectra B. Because the $v_{n}$ -self-map becomes invertible after $K (n)$ -localisation, so $L_{K (n)} F$ receives a map from the telescope by the universal property of localisation C. Because Morava $K$ -theory is a field D. Because the conjecture is true at height one

Hint

The self-map $f$ induces an isomorphism on $K (n)_{*}$ .

Answer

B. Since $f$ is a $K (n)_{*}$ -isomorphism, it becomes an equivalence in the $K (n)$ -local category, so $L_{K (n)} F$ already inverts $f$ . The universal property of the telescope localisation $L_{T (n)} = v_{n}^{- 1} (-)$ then produces a canonical map to any spectrum that inverts $f$ , giving $c$ . Its existence is automatic; whether it is an equivalence is the conjecture.

Exercise 3 (medium, short-answer).

Explain why $T (n)$ is independent, as a Bousfield class, of the choice of type- $n$ complex $F$ and of $v_{n}$ -self-map $f$ .

Hint

Use the asymptotic uniqueness clause of the Hopkins-Smith periodicity theorem and the thick-subcategory theorem.

Answer

The periodicity theorem 03.12.45 gives that any two $v_{n}$ -self-maps on the same complex agree after iteration, so the telescope is well defined up to the choice of complex. For two different type- $n$ complexes $F$ and $F^{'}$ , the thick-subcategory theorem says the type- $n$ complexes generate a single thick subcategory, so $F$ and $F^{'}$ build each other by finite cofibre sequences. The telescope construction carries finite cofibre sequences to cofibre sequences of telescopes, all of which are $v_{n}$ -periodic, and a Bousfield-class computation shows $⟨ T (n) F ⟩ = ⟨ T (n) F^{'} ⟩$ . Hence the class $⟨ T (n)⟩$ depends only on $n$ . Rubric: full credit for invoking both asymptotic uniqueness and the thick-subcategory generation.

Exercise 4 (medium, short-answer).

The comparison map $c$ is always a $K (n)_{*}$ -equivalence. Deduce that the fibre of $c$ is $K (n)$ -acyclic, and explain why this does not by itself settle the conjecture.

Hint

A $K (n)_{*}$ -equivalence has a fibre with vanishing $K (n)$ -homology. Acyclicity for one theory is weaker than contractibility.

Answer

Both $T (n)$ and $L_{K (n)} F$ have the same $K (n)$ -homology as $F$ , and $c$ induces the identity on $K (n)_{*}$ ; so the long exact sequence forces $K (n)_{*} (fib c) = 0$ , i.e. the fibre is $K (n)$ -acyclic. The conjecture would follow only if this fibre were contractible. But a $K (n)$ -acyclic spectrum can still carry substantial homotopy: $K (n)$ -homology does not detect all $v_{n}$ -periodic information. The disproof exhibits exactly such a fibre — non-contractible yet $K (n)$ -acyclic — living in the telescope but invisible to Morava $K$ -theory. Rubric: full credit for the acyclic-fibre deduction and the observation that $K (n)$ -acyclic does not imply contractible.

Exercise 5 (medium, short-answer).

Describe how the telescope conjecture would, if true, simplify the structure of $v_{n}$ -periodic homotopy and hence of $π_{*} S^{0}$ .

Hint

$K (n)$ -local homotopy is computable via the Morava stabiliser group; the telescope is not.

Answer

If $T (n) ≃ L_{K (n)} F$ , then $v_{n}$ -periodic homotopy would be computed by the $K (n)$ -local Adams-Novikov spectral sequence, whose $E_{2}$ -page is the continuous cohomology of the Morava stabiliser group $S_{n}$ acting on the Lubin-Tate ring — a finite, algebraically structured input. The chromatic reassembly of $π_{*} S^{0}$ from its layers would then be governed entirely by this $K (n)$ -local data at each height, making the periodic part of the stable stems as computable as group cohomology. The disproof means the telescope carries extra homotopy beyond the $K (n)$ -local prediction, so $v_{n}$ -periodic homotopy is strictly larger and the stable stems are correspondingly more intricate. Rubric: full credit for naming the $K (n)$ -local ANSS computability and the consequence that the true periodic homotopy is larger.

Exercise 6 (hard, short-answer).

Sketch the logical shape of the Burklund-Hahn-Levy-Schlank-Yuan disproof: which invariant separates $T (n)$ from $L_{K (n)} F$ , and what computational input makes the separation visible.

Hint

The separating invariant is detected by algebraic $K$ -theory and the telescopic Picard group; the input is a trace-method computation.

Answer

The disproof builds a height- $n$ spectrum whose telescopic and $K (n)$ -local invariants differ, detected through the telescopic Picard group $Pic (Sp_{T (n)})$ versus $Pic (Sp_{K (n)})$ . The key computational input is the algebraic $K$ -theory of suitable cyclotomic/Lubin-Tate ring spectra, accessed by trace methods (topological cyclic homology and the cyclotomic-trace machinery). A redshift phenomenon raises a height- $(n - 1)$ input to a height- $n$ output, and the trace-method computation shows the resulting $K$ -theory spectrum carries $v_{n}$ -periodic classes that survive in the telescope but vanish $K (n)$ -locally. The presence of these classes makes the fibre of $c$ non-contractible at all heights $n \geq 2$ . Rubric: full credit for naming the telescopic Picard group / $K$ -theory invariant and the trace-method (TC / redshift) input.

Exercise 7 (hard, short-answer).

The conjecture is equivalent to $L_{n}^{f} = L_{n}$ . Explain what each functor is and why the failure of the telescope conjecture is the statement $L_{n}^{f} \neq = L_{n}$ for $n \geq 2$ .

Hint

$L_{n}^{f}$ is the finite localisation killing type- $(n + 1)$ finite complexes; $L_{n}$ is Bousfield localisation at $E (n)$ .

Answer

$L_{n}^{f}$ is the finite localisation away from the thick subcategory of type- $(n + 1)$ finite complexes — built from telescopes, so its monochromatic layers are the telescopes $T (m)$ for $m \leq n$ . The functor $L_{n}$ is Bousfield localisation at the Johnson-Wilson theory $E (n)$ , whose monochromatic layers are the $K (m)$ -localisations $L_{K (m)}$ . There is always a natural transformation $L_{n}^{f} \to L_{n}$ that is an equivalence layer by layer if and only if each $T (m) \to L_{K (m)}$ is. Since the height- $m$ comparison fails for $2 \leq m \leq n$ , the transformation $L_{n}^{f} \to L_{n}$ is not an equivalence, i.e. $L_{n}^{f} \neq = L_{n}$ . Rubric: full credit for identifying both functors by their defining localisations and tying the inequality to the per-layer comparison failure.

Advanced results Master

Theorem (Ravenel's telescope conjecture, 1984). For each $n \geq 1$ and each type- $n$ finite $p$ -local complex $F$ with $v_{n}$ -self-map $f$ , the comparison map $c : T (n) = v_{n}^{- 1} F \to L_{K (n)} F$ is an equivalence; equivalently $⟨ T (n)⟩ = ⟨ K (n)⟩$ and $L_{n}^{f} = L_{n}$ .

Ravenel posed this in his 1984 Localization paper ^{[Ravenel 1984]} (Conjectures 10.5 and 10.8), having reduced large parts of his chromatic programme to it. The Orange Book ^{[Ravenel 1992]} organises Chapters 6 through 8 around its consequences. Ravenel himself later expressed doubt, and several near-disproofs at the prime two were attempted in the 1990s and 2000s without reaching a definitive result.

Theorem (height one, true; Miller-Mahowald). The telescope conjecture holds at $n = 1$ for all primes. This is the content of the Key theorem above, due to Miller ^{[Miller 1981]} at odd primes and Mahowald ^{[Mahowald 1981]} at $p = 2$ ; the $v_{1}$ -periodic homotopy of the Moore spectrum is the image of $J$ and matches its $K (1)$ -localisation exactly.

Theorem (height zero, degenerate). At $n = 0$ the telescope and $K (0)$ -localisation are both rationalisation, so the comparison map is an equivalence. Height zero carries no genuine periodicity; $T (0) = L_{K (0)} = (-)_{Q}$ .

Theorem (disproof at heights $\geq 2$ ; Burklund-Hahn-Levy-Schlank-Yuan, 2023). For every prime $p$ and every height $n \geq 2$ , the comparison map $c : T (n) \to L_{K (n)} F$ is not an equivalence. The telescope conjecture is false: $v_{n}$ -periodic homotopy is strictly larger than $K (n)$ -local homotopy, and $L_{n}^{f} \neq = L_{n}$ .

The disproof ^{[Burklund-Hahn-Levy-Schlank-Yuan 2023]} proceeds by a chromatic redshift argument fed by trace-method computations in algebraic $K$ -theory. The separating invariant is the telescopic Picard group: $Pic (Sp_{T (n)})$ and $Pic (Sp_{K (n)})$ are shown to differ, exhibiting $v_{n}$ -periodic classes detected by the cyclotomic trace from the algebraic $K$ -theory of Lubin-Tate and cyclotomic ring spectra that are present telescopically but $K (n)$ -acyclic. The redshift principle promotes a height- $(n - 1)$ computation to a height- $n$ obstruction, and the trace methods (topological cyclic homology, the Dundas-Goodwillie-McCarthy and Lichtenbaum-Quillen circle of results) make the obstruction explicit.

Theorem (consequences for the stable stems). Because the telescope retains homotopy invisible to Morava $K$ -theory, the chromatic reassembly of $\pi_ S^0 $f r o mi t s m o n oc h r o ma t i c l a y er sc a r r i es m or e in f or ma t i o na t h e i g h t s$ \geq 2 $t han t h e$ K(n) $- l oc a l d a t aa l o n es u ppl i es . * T h ec h r o ma t i cco n v er g e n ce t h eor e m [03.12.45] s t i l l r e a sse mb l es t h e$ E(n) $- l oc a l t o w er; w ha t f ai l s i s t h e i d e n t i f i c a t i o n o f t h e t e l esco p i c l a y er s w i t h t h e$ K(n) $- l oc a l l a y er s, so$ v_n $- p er i o d i c f ami l i es in t h es t e m s a r e n o t f u l l y p r e d i c t e d b y$ K(n)$-local homotopy.

Synthesis. The telescope conjecture is the foundational reason the chromatic picture of stable homotopy was expected to be computable: it is exactly the claim that the mechanical $v_{n}$ -periodic homotopy, captured by the telescope, equals the algebraically structured $K (n)$ -local homotopy. The central insight of the chromatic programme is that $π_{*} S^{0}$ is filtered by height and reassembled from monochromatic layers, and the conjecture is dual to the nilpotence theorem in that dictionary — nilpotence governs vanishing, periodicity governs survival. Putting these together, the height-one theorem of Miller and Mahowald generalises the classical image of $J$ and made the agreement seem inevitable, while the 2023 disproof shows the bridge collapses: the telescope generalises to strictly more than $K (n)$ detects, and the finite localisation $L_{n}^{f}$ is genuinely finer than $L_{n}$ for $n \geq 2$ . The foundational reason the bridge fails is the redshift input from algebraic $K$ -theory, the same arithmetic that powers the cyclotomic trace; this is exactly where the telescopic Picard group separates from the $K (n)$ -local one. The bridge from the conjecture to its disproof is therefore the recognition that trace methods see periodic homotopy that Morava $K$ -theory cannot, and this pattern recurs throughout the modern interaction of chromatic homotopy theory with arithmetic.

Full proof set Master

Proposition (the comparison map is a $K (n)_{*}$ -equivalence). For any type- $n$ finite $F$ with $v_{n}$ -self-map $f$ , the comparison map $c : T (n) \to L_{K (n)} F$ induces an isomorphism on $K(n)_$.*

Proof. The self-map $f$ is a $K (n)_{*}$ -isomorphism by definition of a $v_{n}$ -self-map 03.12.45. Morava $K$ -theory $K (n)_{*}$ commutes with the sequential homotopy colimit defining $T (n) = hocolim (F f Σ^{- d} F \to \dots)$ , because $K (n)$ is a homology theory and homology commutes with filtered colimits. Therefore $$ K(n)* T(n) = \operatorname{colim}\left( K(n)F \xrightarrow{f_} K(n)*F \to \cdots \right) = K(n)F, $$ the colimit of isomorphisms. On the other side $K(n)_ L_{K(n)}F = K(n)*F $, s in ce$ K(n) $- l oc a l i s a t i o ni s a$ K(n)* $- e q u i v a l e n ce [03.12.48] . T h e ma p$ c $i sco m p a t ib l e w i t h t h es t r u c t u r e ma p s f r o m$ F $, so i t in d u ces t h e i d e n t i t y o n$ K(n)*F $. H e n ce$ c $i s a$ K(n)* $- i so m or p hi s m .$ \square$

Proposition (fibre of $c$ is $K (n)$ -acyclic but need not vanish). The fibre $C = fib (c)$ satisfies $K(n)_ C = 0 $. T h e t e l esco p eco nj ec t u r e a t h e i g h t$ n $i se q u i v a l e n tt o$ C \simeq 0 $f or a l l t y p e -$ nF$.*

Proof. From the cofibre sequence $C \to T (n) c L_{K (n)} F$ and the long exact sequence in $K (n)$ -homology, the previous proposition gives that $c_{*}$ is an isomorphism, so $K (n)_{*} C = 0$ . The map $c$ is an equivalence precisely when its fibre is contractible. Since the conjecture is the assertion that $c$ is an equivalence for all type- $n$ $F$ , it is equivalent to $C ≃ 0$ in every such case. The point of the disproof is that $K (n)_{*} C = 0$ does not force $C ≃ 0$ : a $K (n)$ -acyclic spectrum can carry non-vanishing homotopy detected by finer invariants. $□$

Proposition (height independence of the Bousfield class). The Bousfield class $⟨ T (n)⟩$ is independent of the type- $n$ complex $F$ and the self-map $f$ ; consequently the conjecture is a statement about $n$ alone.

Proof. Two $v_{n}$ -self-maps on a fixed $F$ agree after iteration by the asymptotic-uniqueness clause of the Hopkins-Smith periodicity theorem 03.12.45 ^{[Hopkins-Smith 1998]}, so their telescopes have equal Bousfield class. For two type- $n$ complexes $F$ and $F^{'}$ , the thick-subcategory theorem identifies the type- $\geq n$ finite complexes as a single thick subcategory generated by either one, so each is built from the other by finitely many cofibre sequences and retracts. The telescope functor $v_{n}^{- 1} (-)$ is exact and sends $v_{n}$ -equivalences to equivalences, so $⟨ T (n) F ⟩ = ⟨ T (n) F^{'} ⟩$ . The class depends only on $n$ . $□$

Proposition (height-one collapse to the image of $J$ ). At $n = 1$ , $\pi_ T(1) \otimes (\text{the type-1 complex } V(0)) $i s t h e$ v_1 $- p er i o d i c ima g eo f$ J $, an d e q u a l s$ \pi_* L_{K(1)}V(0)$.*

Proof. The $K (1)$ -local side is computed by the $K (1)$ -local Adams-Novikov spectral sequence 03.12.38, whose $E_{2}$ -page is the continuous cohomology $H_{c}^{*} (Z_{p}^{\times}; (E_{1})_{*} V (0))$ of the height-one Morava stabiliser group $S_{1} = Z_{p}^{\times}$ . This collapses to the image-of- $J$ pattern 03.08.11 with the $α$ -family periodicity. The telescope side is computed by Miller's localised Adams spectral sequence at odd primes and Mahowald's $b o$ -resolution at $p = 2$ ^{[Miller 1981]} ^{[Mahowald 1981]}, each presenting $v_{1}^{- 1} π_{*} V (0)$ as the homology of an explicit periodic complex with the same image-of- $J$ output. The comparison map carries one $E_{\infty}$ -pattern isomorphically to the other, so it is an isomorphism on homotopy, hence an equivalence. $□$

Proposition (disproof obstruction is non-nilpotent telescopically). For $n \geq 2$ , the fibre $C$ of $c$ carries a class detected by the cyclotomic trace from algebraic $K$ -theory that is $v_{n}$ -periodic in $T (n)$ but zero in $L_{K (n)} F$ ; hence $C \neq ≃ 0$ .

Proof sketch. The Burklund-Hahn-Levy-Schlank-Yuan argument ^{[Burklund-Hahn-Levy-Schlank-Yuan 2023]} constructs, from a height- $(n - 1)$ ring spectrum $R$ , the algebraic $K$ -theory $K (R)$ , which by the redshift principle has chromatic height $n$ . Trace methods compute $L_{T (n)} K (R)$ via topological cyclic homology and show it differs from $L_{K (n)} K (R)$ : the telescopic localisation retains classes coming from the cyclotomic structure (the $R P^{\infty}$ -type / Tate-construction input) that are $K (n)$ -acyclic. Transporting this difference along the comparison map for a suitable type- $n$ complex $F$ exhibits a non-zero class in $π_{*} C$ . The class is $v_{n}$ -periodic telescopically, so it is genuinely present in $T (n)$ , while its image in $L_{K (n)} F$ vanishes. Therefore $C$ is not contractible, and the comparison map is not an equivalence at any height $n \geq 2$ . $□$

Connections Master

Bousfield localisation and Morava $K$ -theory 03.12.48. The entire statement lives in the language of Bousfield localisation: $T (n)$ is the localisation $L_{T (n)}$ at the telescope homology theory, and $L_{K (n)}$ is localisation at Morava $K$ -theory. The telescope conjecture is the equality of two Bousfield classes, $⟨ T (n)⟩ = ⟨ K (n)⟩$ . Everything about the comparison map, the chromatic tower, and the $L_{n}^{f}$ versus $L_{n}$ formulation is read off from the localisation framework of 03.12.48; without it the conjecture cannot even be phrased.
Nilpotence and periodicity 03.12.45. The Hopkins-Smith periodicity theorem supplies the $v_{n}$ -self-map whose telescope defines $T (n)$ , and asymptotic uniqueness is what makes $T (n)$ depend only on $n$ . The nilpotence theorem is the dual half of the dictionary: nilpotence detects what vanishes, periodicity what survives. The telescope conjecture is exactly the question of whether periodicity, mechanically captured, agrees with the Morava- $K (n)$ prediction, so it sits directly on top of the periodicity theorem and the chromatic convergence theorem proved there.
Adams-Novikov spectral sequence 03.12.38. Both sides of the comparison are computed by spectral sequences. The $K (n)$ -local side uses the $K (n)$ -local Adams-Novikov spectral sequence, whose $E_{2}$ -page is continuous cohomology of the Morava stabiliser group acting on the Lubin-Tate ring; the telescope side uses localised Adams-type spectral sequences. The height-one agreement and the higher-height failure are statements about whether these two spectral sequences converge to the same homotopy, making the ANSS the principal computational instrument for testing the conjecture.
Chromatic spectral sequence 03.13.06. The chromatic spectral sequence organises the Adams-Novikov $E_{2}$ -page into height layers, and the telescope conjecture is the topological counterpart: it asks whether the topological height- $n$ layer (the telescope) matches the $K (n)$ -local layer that the chromatic machinery isolates algebraically. The conjecture's failure means the topological chromatic filtration by telescopes is strictly finer than the $E (n)$ -local filtration, refining the picture that the chromatic spectral sequence draws on the $E_{2}$ -page.
Image of $J$ 03.08.11. The height-one telescope is the image of $J$ , the oldest infinite family in the stable stems. The Miller-Mahowald theorem is precisely the statement that the $v_{1}$ -periodic telescope of the Moore spectrum reproduces the image of $J$ and agrees with its $K (1)$ -localisation. The image of $J$ is therefore the concrete prototype for the whole conjecture, and the reason the agreement was expected to persist to all heights.

Historical & philosophical context Master

Ravenel introduced the telescope conjecture in his 1984 paper Localization with respect to certain periodic homology theories (Amer. J. Math. 106) ^{[Ravenel 1984]}, as part of a sweeping reorganisation of stable homotopy theory around the chromatic filtration. The conjecture was attractive because it promised that the mysterious $v_{n}$ -periodic homotopy — built mechanically by inverting a self-map — would be computed by the algebraically tractable $K (n)$ -local homotopy, whose input is the cohomology of the Morava stabiliser group. Ravenel reduced significant portions of his programme to it and built Chapters 6 through 8 of his 1992 Orange Book, Nilpotence and Periodicity in Stable Homotopy Theory (Annals of Math. Studies 128) ^{[Ravenel 1992]}, around its consequences.

The height-one case was settled affirmatively almost immediately. Miller's 1981 On relations between Adams spectral sequences (J. Pure Appl. Algebra 20) ^{[Miller 1981]} handled odd primes through a localised Adams spectral sequence, while Mahowald's 1981 $b o$ -resolution (Pacific J. Math. 92) ^{[Mahowald 1981]} covered the prime two; both identified the $v_{1}$ -periodic homotopy of the Moore spectrum with the image of $J$ . This early success cemented the expectation that the conjecture held in general. Yet Ravenel himself grew skeptical, and through the 1990s and 2000s a sequence of attempted disproofs at the prime two, notably by Mahowald, Ravenel, and Shick, produced strong evidence against the conjecture without a complete proof of failure.

The matter was resolved in 2023 by Robert Burklund, Jeremy Hahn, Ishan Levy, Tomer Schlank, and Allen Yuan in K-theoretic counterexamples to Ravenel's telescope conjecture (arXiv:2310.17459) ^{[Burklund-Hahn-Levy-Schlank-Yuan 2023]}, who disproved it at every height $n \geq 2$ and every prime. The decisive new ingredient was the influx of trace methods and the redshift philosophy: algebraic $K$ -theory raises chromatic height by one, and topological cyclic homology computes the relevant telescopic localisations. The telescopic Picard group was shown to differ from its $K (n)$ -local counterpart, exhibiting periodic homotopy detected by the cyclotomic trace but invisible to Morava $K$ -theory. Philosophically the result marks a shift: the periodic structure of the stable stems is not governed by Morava $K$ -theory alone but is entangled with the arithmetic of algebraic $K$ -theory, so the chromatic picture is richer, and harder, than the conjecture had promised.

Bibliography Master

@article{Ravenel1984,
  author  = {Ravenel, Douglas C.},
  title   = {Localization with respect to certain periodic homology theories},
  journal = {Amer. J. Math.},
  volume  = {106},
  number  = {2},
  year    = {1984},
  pages   = {351--414}
}

@article{Miller1981,
  author  = {Miller, Haynes R.},
  title   = {On relations between {Adams} spectral sequences, with an application to the stable homotopy of a {Moore} space},
  journal = {J. Pure Appl. Algebra},
  volume  = {20},
  number  = {3},
  year    = {1981},
  pages   = {287--312}
}

@article{Mahowald1981,
  author  = {Mahowald, Mark},
  title   = {bo-resolutions},
  journal = {Pacific J. Math.},
  volume  = {92},
  number  = {2},
  year    = {1981},
  pages   = {365--383}
}

@article{MillerRavenelWilson1977,
  author  = {Miller, Haynes R. and Ravenel, Douglas C. and Wilson, W. Stephen},
  title   = {Periodic phenomena in the {Adams-Novikov} spectral sequence},
  journal = {Ann. of Math. (2)},
  volume  = {106},
  number  = {3},
  year    = {1977},
  pages   = {469--516}
}

@article{HopkinsSmith1998,
  author  = {Hopkins, Michael J. and Smith, Jeffrey H.},
  title   = {Nilpotence and stable homotopy theory {II}},
  journal = {Ann. of Math. (2)},
  volume  = {148},
  number  = {1},
  year    = {1998},
  pages   = {1--49}
}

@article{BHLSY2023,
  author  = {Burklund, Robert and Hahn, Jeremy and Levy, Ishan and Schlank, Tomer M. and Yuan, Allen},
  title   = {K-theoretic counterexamples to {Ravenel's} telescope conjecture},
  journal = {arXiv preprint arXiv:2310.17459},
  year    = {2023}
}

@book{Ravenel1992,
  author    = {Ravenel, Douglas C.},
  title     = {Nilpotence and Periodicity in Stable Homotopy Theory},
  series    = {Annals of Mathematics Studies},
  volume    = {128},
  publisher = {Princeton University Press},
  year      = {1992}
}

@incollection{BarthelBeaudry2020,
  author    = {Barthel, Tobias and Beaudry, Agn\`es},
  title     = {Chromatic structures in stable homotopy theory},
  booktitle = {Handbook of Homotopy Theory},
  editor    = {Miller, Haynes},
  publisher = {CRC Press},
  year      = {2020}
}

Prerequisites

03.12.48
03.12.45
03.12.38

Tier anchors

beginner: the two-ways-to-isolate-periodicity picture — informal opening of Ravenel 1992 *Nilpotence and Periodicity in Stable Homotopy Theory* (Annals of Math. Studies 128) Ch. 1, and the survey framing in Barthel-Beaudry 2020 *Handbook of Homotopy Theory* (chromatic homotopy chapter) §1
intermediate: Ravenel 1984 *Localization with respect to certain periodic homology theories* (Amer. J. Math. 106) §10 (Conjectures 10.5, 10.8); Miller 1981 *On relations between Adams spectral sequences* (J. Pure Appl. Algebra 20) — the height-1 telescope theorem; Mahowald 1981 *bo-resolutions* (Pacific J. Math. 92)
master: Ravenel 1984 *Localization with respect to certain periodic homology theories* (Amer. J. Math. 106); Miller-Ravenel-Wilson 1977 *Periodic phenomena in the Adams-Novikov spectral sequence* (Ann. of Math. 106); Hopkins-Smith 1998 *Nilpotence and stable homotopy theory II* (Ann. of Math. 148); Burklund-Hahn-Levy-Schlank-Yuan 2023 *K-theoretic counterexamples to Ravenel's telescope conjecture* (arXiv:2310.17459)

References

Ravenel — Localization with respect to certain periodic homology theories · Amer. J. Math. 106 (1984), 351-414 — §10 states the telescope conjecture (Conjecture 10.5 / 10.8): the localisation telescope agrees with K(n)-localisation
Miller — On relations between Adams spectral sequences, with an application to the stable homotopy of a Moore space · J. Pure Appl. Algebra 20 (1981), 287-312 — the height-1 telescope theorem (v_1-periodic case), TRUE at all primes
Mahowald — bo-resolutions · Pacific J. Math. 92 (1981), 365-383 — v_1-periodic homotopy via the Adams bo-resolution, partner to Miller's height-1 result at p=2
Miller-Ravenel-Wilson — Periodic phenomena in the Adams-Novikov spectral sequence · Ann. of Math. 106 (1977), 469-516 — chromatic resolution and the v_n-periodic families that the telescope is meant to capture
Hopkins-Smith — Nilpotence and stable homotopy theory II · Ann. of Math. 148 (1998), 1-49 — periodicity theorem: existence and asymptotic uniqueness of the v_n-self-map whose telescope defines T(n)
Burklund-Hahn-Levy-Schlank-Yuan — K-theoretic counterexamples to Ravenel's telescope conjecture · arXiv:2310.17459 (2023) — disproof at all heights n>=2: T(n) and L_{K(n)} genuinely differ, via algebraic K-theory of cyclotomic data and the telescopic Picard group
Ravenel — Nilpotence and Periodicity in Stable Homotopy Theory · Annals of Mathematics Studies 128, Princeton (1992) — the 'Orange Book'; Ch. 6-8 organise the chromatic picture around the telescope conjecture
Barthel-Beaudry — Chromatic structures in stable homotopy theory · Handbook of Homotopy Theory (ed. Miller), CRC Press (2020) — survey statement of the telescope conjecture and the finite vs ordinary localisation distinction L_n^f vs L_n

Estimated time

beginner: 18m
intermediate: 50m
master: 85m