03.13.07 · modern-geometry / spectral-sequences

Greek-letter elements in the stable homotopy of spheres

shipped3 tiersLean: nonepending prereqs

Anchor (Master): Ravenel 2004 (the Green Book) Ch. 5 (Greek-letter construction, chromatic resolution) and Ch. 4 ($BP_*BP$); Miller-Ravenel-Wilson 1977 (Ann. of Math. 106, 469-516); Smith 1970/Toda 1971 ($V(1)$ and the $\beta$-family); Adams 1966 *On the groups J(X) IV*; Hill-Hopkins-Ravenel 2016 *On the nonexistence of elements of Kervaire invariant one* (Ann. of Math. 184) for $\theta_j$

Intuition Beginner

The homotopy groups of spheres are among the hardest objects in mathematics to compute. Each one is a finite abelian group (away from a single copy of the integers), and the list of these groups looks, at first glance, like noise: no obvious pattern, numbers jumping around with no rhythm. The Greek-letter elements are the discovery that hidden inside this noise are clean, infinite, repeating families — and each family carries a Greek-letter name.

The first family is called the alpha family. Its members are spaced out by a fixed period, and they were the first sign that the chaos was organised. The second family, the beta family, sits one level deeper and repeats with a longer period. The third, the gamma family, is deeper still. Each new Greek letter records a new layer of periodicity that was invisible at the layers above it.

The picture to hold onto is a stack of transparent sheets, one per layer. The alpha sheet shows one repeating pattern. Lay the beta sheet on top and a second, slower pattern appears. The gamma sheet adds a third. Each layer is indexed by a whole number called the chromatic height, and the Greek letters march in step with that number: alpha at height one, beta at height two, gamma at height three.

Visual Beginner

Picture the integers along a horizontal line — these label the dimension of a homotopy group of spheres. Above the line, draw small marks wherever a known family member lands. The alpha marks appear at a steady spacing. The beta marks appear at a wider, slower spacing. The gamma marks are rarer still. Each family is a colour-coded comb, and the combs interleave without colliding.

The point of the picture is that each comb is generated by a single rule. Once you know the period of the alpha comb, every alpha mark is determined. The beta comb has its own period, the gamma comb its own. The whole infinite supply of family members comes from a finite recipe applied over and over, one recipe per height.

Worked example Beginner

Here is the alpha family at a small odd prime, say $p = 3$ , described without machinery.

Step 1. Fix the prime $3$ . We only track the part of each homotopy group that is a power of $3$ in size; the rest is handled prime by prime in the same way.

Step 2. The alpha family produces one element in each dimension of the form $4 k$ minus $1$ , that is, in dimensions $3, 7, 11, 15, \dots$ at this prime. So there is an alpha element in dimension $3$ , another in dimension $7$ , and so on up the ladder.

Step 3. The size of each alpha element is controlled by a counting rule. At the prime $3$ , the element in dimension $4 k - 1$ has order $3$ raised to a power that grows slowly with $k$ — it ticks up by one each time $k$ passes a multiple of $3$ . So most alpha elements have order exactly $3$ , but every third one is larger.

Step 4. Read off the first few. In dimension $3$ the alpha element has order $3$ . In dimension $7$ it has order $3$ . In dimension $11$ it has order $9$ , because $k = 3$ here and the counting rule ticks up. The pattern continues forever, one element per rung.

What this tells us: a single arithmetic rule, tied to ordinary number theory, predicts an infinite list of elements in the homotopy of spheres. The alpha family is the simplest such list, and the deeper Greek letters refine the same idea.

Check your understanding Beginner

Formal definition Intermediate+

Fix a prime $p$ and work $p$ -locally throughout. The input is the Adams-Novikov spectral sequence 03.12.38, whose $E_{2}$ -page is $Ext_{B P_{*} B P} (B P_{*}, B P_{*})$ , computed over the Hopf algebroid $(B P_{*}, B P_{*} B P)$ of the Brown-Peterson spectrum 03.13.05. Recall $B P_{*} = Z_{(p)} [v_{1}, v_{2}, \dots]$ with $∣ v_{n} ∣ = 2 (p^{n} - 1)$ , and the invariant prime ideals $I_{n} = (p, v_{1}, \dots, v_{n - 1})$ , each a sub-comodule of $B P_{*}$ .

Definition (chromatic comodules). For each $n \geq 0$ set $$ N^n = BP_*/I_n, \qquad M^n = v_n^{-1} N^n = v_n^{-1}\bigl(BP_*/I_n\bigr), $$ with $N^{0} = B P_{*}$ and $M^{0} = v_{0}^{- 1} B P_{*} = p^{- 1} B P_{*} = Q \otimes B P_{*}$ . The short exact sequences of $B P_{*} B P$ -comodules $$ 0 \to N^n \to M^n \to N^{n+1} \to 0 $$ splice into the chromatic resolution $B P_{*} \to M^{0} \to M^{1} \to M^{2} \to \dots$ 03.13.06, whose derived spectral sequence (the chromatic spectral sequence) has $E_{1}^{n, s} = Ext_{B P_{*} B P}^{s} (B P_{*}, M^{n})$ and converges to $Ext_{B P_{*} B P}^{n + s} (B P_{*}, B P_{*})$ .

Definition (Greek-letter map). Each short exact sequence above induces a connecting homomorphism $$ \delta_n : \mathrm{Ext}^{s}{BPBP}(BP_, M^n) \longrightarrow \mathrm{Ext}^{s+1}{BPBP}(BP_, N^{n+1}), $$ and composing the $n$ successive connecting maps gives the $n$ -th Greek-letter map $$ \widetilde{\delta}^{(n)} : \mathrm{Ext}^{0}{BPBP}(BP_, M^n) \longrightarrow \mathrm{Ext}^{n}{BPBP}(BP_, BP_*). $$ The source $Ext^{0} (M^{n})$ consists of the $v_{n}$ -towers: $B P_{*} B P$ -comodule primitives in $M^{n}$ , which are spanned by classes of the form $v_{n}^{t} / (p^{i_{0}} v_{1}^{i_{1}} \dots v_{n - 1}^{i_{n - 1}})$ subject to invariance. The images $δ^{(n)} (v_{n}^{t} / \dots)$ are the Greek-letter elements of height $n$ .

Definition (the named families). Writing the Greek letter for the height:

$α$ -family ( $n = 1$ ): $α_{t / j} = δ^{(1)} (v_{1}^{t} / p^{j}) \in Ext^{1}$ , where $j$ is the largest power of $p$ for which $v_{1}^{t} / p^{j}$ is a primitive of $M^{1}$ .
$β$ -family ( $n = 2$ ): $β_{t / j, i} = δ^{(2)} (v_{2}^{t} / (p^{i}, v_{1}^{j})) \in Ext^{2}$ , with $β_{s} := β_{s /1, 1}$ .
$γ$ -family ( $n = 3$ ): $γ_{t / \dots} = δ^{(3)} (v_{3}^{t} / (p, v_{1}, v_{2})) \in Ext^{3}$ .

A Greek-letter element is a permanent cycle if it survives the Adams-Novikov differentials to define a non-zero class in $π_{*} S_{(p)}$ .

Counterexamples to common slips

Not every $v_{n}$ -tower class is a primitive: the divisibility datum $(i_{0}, \dots, i_{n - 1})$ is constrained by the comodule structure of $M^{n}$ , and a careless numerator like $v_{2} / v_{1}^{2}$ at $p = 2$ may fail invariance. The bookkeeping of which fractions are allowed is the technical heart of Miller-Ravenel-Wilson.
A Greek-letter element can be non-zero on the $E_{2}$ -page yet die before homotopy. Survival is a separate question from existence in $Ext$ ; the $α$ -family survives entirely (image of $J$ ), but higher families lose members to differentials.
The element $β_{1}$ is $β_{1/1, 1}$ , not $δ^{(2)} (v_{2})$ with empty denominator: the denominator $(p, v_{1})$ is forced because $v_{2}$ alone is not a primitive of $M^{2}$ .
Greek-letter elements live in the Adams-Novikov $E_{2}$ , that is $Ext_{B P_{*} B P}$ , not in the classical Adams $E_{2} = Ext_{A_{p}}$ . The two spectral sequences detect overlapping but distinct families.

Key theorem with proof Intermediate+

Theorem (Miller-Ravenel-Wilson; $α$ -family and the image of $J$ ). Fix a prime $p$ . The $α$ -family $$ \alpha_{t/j} = \widetilde{\delta}^{(1)}\bigl(v_1^{t}/p^{j}\bigr) \in \mathrm{Ext}^1_{BP_BP}(BP_, BP_*) $$ is a complete set of generators of $Ext^{1}$ above filtration zero, all of whose members are permanent cycles, and the resulting subgroup of $\pi_\mathbb{S}{(p)} $i s p r ec i se l y t h e$ p $- p r ima r y ima g eo f t h e$ J $- h o m o m or p hi s m [03.08.11] . T h eor d er o f$ \alpha{t/j} $in d im e n s i o n$ 2t(p-1) - 1 $i s$ p^j $, w h er e$ p^j $i s t h ee x a c tp o w er o f$ p $d i v i d in g$ t \cdot p $w h e n$ p $i so dd (an d a B er n o u l l i - n u mb er d e n o mina t or w h e n$ p = 2$).*

Proof. The source $Ext^{0} (M^{1})$ is computed directly. By the Hopf-algebroid structure of $(B P_{*}, B P_{*} B P)$ , the comodule $M^{1} = v_{1}^{- 1} B P_{*} / p$ has primitives spanned by the classes $v_{1}^{t} / p^{j}$ with $j$ bounded by the invariance condition. The Morava-stabiliser computation of Miller-Ravenel-Wilson identifies this $Ext^{0}$ as a sum of cyclic groups $Z / p^{j (t)}$ , one for each $t$ , where $j (t) = 1 + ν_{p} (t)$ for $p$ odd and $j (t) = 2 + ν_{2} (t) + ν_{2} (t \cdot (Bernoulli factor))$ for $p = 2$ ; here $ν_{p}$ is the $p$ -adic valuation.

Applying the single connecting map $δ_{0}$ of the chromatic short exact sequence $0 \to B P_{*} \to M^{0} \to N^{1} \to 0$ followed by $δ_{1}$ of $0 \to N^{1} \to M^{1} \to N^{2} \to 0$ realises $δ^{(1)}$ as an isomorphism onto the subgroup of $Ext^{1}$ of positive stem. This is the chromatic computation of the first line of the Adams-Novikov $E_{2}$ .

Survival to homotopy is the theorem of Adams ^{[Adams 1966]}. The $J$ -homomorphism $J : π_{*} S O \to π_{*} S$ has image whose $p$ -primary part, in dimension $4 k - 1$ , is cyclic of order equal to the denominator of $B_{2 k} /4 k$ (von Staudt-Clausen). Adams proved every such class is detected, and the Adams conjecture (Quillen, Sullivan) shows the image of $J$ is a direct summand. Matching dimensions and orders against the $Ext^{1}$ computation identifies the $α$ -family with $im J$ class by class, and since $im J$ consists of permanent cycles, so does the $α$ -family. $□$

Bridge. This theorem builds toward the whole chromatic philosophy and appears again in every higher Greek-letter family. The foundational reason the $α$ -family is computable is exactly that height one is governed by $v_{1}$ -periodicity, and $v_{1}$ -periodicity is the same phenomenon that $K$ -theory and the image of $J$ detect — this is the central insight that the chromatic filtration generalises height by height. Putting these together, the connecting map $δ^{(1)}$ that produces $α_{t / j}$ is dual to the chromatic short exact sequences, and the same connecting-map machinery, iterated $n$ times, generalises to produce the $β$ - and $γ$ -families at heights two and three. The bridge is that the Bernoulli denominators controlling $im J$ are the height-one shadow of the deeper number-theoretic data — Eisenstein series and modular forms — that govern $v_{2}$ -periodicity, so the $α$ -family is the first visible rung of a ladder that the periodicity theorem [03.12.41-area] climbs through all heights.

Exercises Intermediate+

Exercise 3 (medium, symbolic).

Explain why $v_{2}$ by itself is not a primitive of $M^{2} = v_{2}^{- 1} B P_{*} / (p, v_{1})$ , so that the smallest $β$ -element must carry the denominator $(p, v_{1})$ .

Hint

A primitive must be killed by the reduced coaction. Apply the right unit $η_{R}$ to $v_{2}$ modulo $(p, v_{1})$ and check whether it differs from the left unit.

Answer

A class $x \in M^{2}$ is a comodule primitive when $ψ (x) = 1 \otimes x$ , equivalently $η_{R} (x) \equiv η_{L} (x)$ in $B P_{*} B P \otimes M^{2}$ . For $v_{2}$ the right unit satisfies $η_{R} (v_{2}) \equiv v_{2} + v_{1} t_{1}^{p} - v_{1}^{p} t_{1} (mod p)$ in $B P_{*} B P$ . Working in $M^{2}$ we have killed $v_{1}$ , so modulo $(p, v_{1})$ the correction terms $v_{1} t_{1}^{p}$ and $v_{1}^{p} t_{1}$ vanish and $v_{2}$ becomes primitive only after this reduction. The class $v_{2} / (p, v_{1})$ is the fraction whose numerator $v_{2}$ is primitive precisely because the denominator quotient $(p, v_{1})$ removes the obstruction; without dividing by $(p, v_{1})$ the element $v_{2}$ is not a cocycle representing a class in $Ext^{0} (M^{2})$ . Rubric: full credit for invoking the right-unit formula and identifying that reduction modulo $(p, v_{1})$ removes the coaction correction.

Exercise 4 (medium, short-answer).

State precisely the relationship between the $α$ -family and the image of $J$ , and explain why every $α$ -element is a permanent cycle while higher Greek-letter families need not be.

Hint

The image of $J$ is known to consist entirely of permanent cycles by Adams; what distinguishes height one from higher heights is that $Ext^{1}$ has no room above it for a differential to come from.

Answer

The $α$ -family is identified, class by class and dimension by dimension, with the $p$ -primary image of $J$ : in stem $4 k - 1$ both are cyclic of order the $p$ -part of the denominator of $B_{2 k} /4 k$ . Every $α$ -element is a permanent cycle because it lives in Adams-Novikov filtration one, and the only differentials into or out of filtration one in the relevant range are forced to vanish by sparseness, while Adams' theorem confirms the corresponding homotopy class is non-zero. Higher families sit in filtration two and above, where Adams-Novikov differentials $d_{r}$ can both originate and terminate; consequently a $β$ - or $γ$ -element may be hit by a differential or support one, killing it before it reaches homotopy. Rubric: full credit for the dimension-and-order matching plus the filtration-one sparseness argument.

Exercise 5 (medium, symbolic).

Using the iterated connecting map $δ^{(n)} = δ_{n - 1} \circ \dots \circ δ_{0}$ , explain why a Greek-letter element of height $n$ lands in $Ext^{n}$ and not in some other homological degree.

Hint

Each connecting homomorphism of a short exact sequence raises homological degree by exactly one. Count how many short exact sequences the chromatic resolution splices together to reach height $n$ .

Answer

The chromatic resolution $B P_{*} \to M^{0} \to M^{1} \to \dots$ is obtained by splicing the $n$ short exact sequences $0 \to N^{k} \to M^{k} \to N^{k + 1} \to 0$ for $k = 0, \dots, n - 1$ . The connecting homomorphism of each such sequence raises $Ext$ -degree by exactly one: $δ_{k} : Ext^{s} (M^{k} -input) \to Ext^{s + 1} (N^{k + 1})$ . Starting from a primitive in $Ext^{0} (M^{n})$ and applying the $n$ connecting maps in sequence raises the homological degree from $0$ to $n$ , landing in $Ext^{n} (B P_{*})$ . This is exactly why height $n$ produces $Ext^{n}$ -classes: $α$ in $Ext^{1}$ , $β$ in $Ext^{2}$ , $γ$ in $Ext^{3}$ . Rubric: full credit for the degree-raising count and the identification of the number of spliced sequences with the height.

Exercise 6 (hard, short-answer).

For $p \geq 5$ , Smith and Toda constructed the spectrum $V (1)$ and used it to prove the entire $β$ -family consists of permanent cycles. Sketch why a $v_{2}$ -self-map of a type- $2$ finite spectrum produces an infinite family of homotopy classes.

Hint

A $v_{2}$ -self-map $f : Σ^{d} V (1) \to V (1)$ induces an isomorphism on $K (2)_{*}$ . Iterating $f$ and composing with the inclusion of the bottom cell and projection to the top cell turns one map into an infinite tower.

Answer

For $p \geq 5$ the Smith-Toda spectrum $V (1) = S / (p, v_{1})$ is a type- $2$ finite spectrum admitting a $v_{2}$ -self-map $f : Σ^{2 (p^{2} - 1)} V (1) \to V (1)$ inducing an isomorphism on Morava $K (2)$ -homology and zero on $K (m)$ for $m \neq = 2$ . Iterating gives $f^{s} : Σ^{2 s (p^{2} - 1)} V (1) \to V (1)$ for every $s \geq 1$ . Pre-composing with the inclusion $ι : S^{2 s (p^{2} - 1)} \to Σ^{2 s (p^{2} - 1)} V (1)$ of the bottom cell and post-composing with the projection $π : V (1) \to S^{0}$ to the top cell (up to a degree shift) produces a non-zero homotopy class $π \circ f^{s} \circ ι \in π_{*} S_{(p)}$ for every $s$ . These classes are detected by $β_{s}$ in the Adams-Novikov $E_{2}$ , so the whole $β$ -family survives. The construction needs $p \geq 5$ because $V (1)$ exists as a ring spectrum with the required self-map only in that range; at $p = 2, 3$ the analysis is subtler and not every $β_{s}$ survives. Rubric: full credit for the iterate-and-cap construction plus the $p \geq 5$ existence caveat.

Exercise 7 (hard, short-answer).

The Kervaire-invariant elements $θ_{j} \in π_{2^{j + 1} - 2} S$ are described as " $β$ -family-adjacent." Explain the sense in which they are related to the $β$ -family, and state the Hill-Hopkins-Ravenel theorem that resolves their existence.

Hint

At $p = 2$ the $θ_{j}$ are detected by $h_{j}^{2}$ in the classical Adams spectral sequence; the analogue of the $β$ -family at the prime $2$ overlaps the Kervaire classes. Hill-Hopkins-Ravenel settled which $θ_{j}$ exist.

Answer

At the prime $2$ the Kervaire-invariant elements $θ_{j}$ are the classes detected by $h_{j}^{2}$ on the second line of the classical Adams spectral sequence, and the second line is exactly where the $β$ -family lives in the Adams-Novikov picture — so the $θ_{j}$ are the $2$ -primary, filtration-two cousins of the odd-primary $β$ -elements, hence " $β$ -family-adjacent." The existence question asks for which $j$ a permanent cycle $θ_{j}$ exists, defining a manifold of Kervaire invariant one in dimension $2^{j + 1} - 2$ . Classical work established $θ_{j}$ for $j \leq 5$ ; Hill-Hopkins-Ravenel ^{[Hill-Hopkins-Ravenel 2016]} proved $θ_{j}$ does not exist for $j \geq 7$ , using a $256$ -periodic equivariant detecting spectrum, leaving only $j = 6$ (dimension $126$ ) open at the time. Rubric: full credit for the $h_{j}^{2}$ / second-line adjacency and the correct statement of the $j \geq 7$ non-existence result.

Advanced results Master

Theorem (chromatic computation of $Ext^{1}$ ; Miller-Ravenel-Wilson 1977). For every prime $p$ , the positive-stem part of $\mathrm{Ext}^1_{BP_BP}(BP_, BP_) $i s t h e$ \alpha $- f ami l y, a d i r ec t s u m o f cy c l i c$ p $- g r o u p s$ \mathbb{Z}/p^{j(t)} $in d e x e d b y$ t \geq 1 $, w h er e$ j(t) = 1 + \nu_p(t) $f or$ p $o dd . E v er y c l a ss i s a p er man e n t cy c l e an d t h e t o t a l i t y i s t h e$ p $- p r ima r y ima g eo f$ J$.*

This is the height-one base case of the chromatic program. The identification with $im J$ ties the homotopy of spheres to the arithmetic of Bernoulli numbers and to algebraic $K$ -theory 03.08.11, because the denominators of $B_{2 k} /4 k$ are exactly the orders of the $K$ -theoretic $J$ -classes.

Theorem ( $β$ -family at $Ext^{2}$ ). For $p \geq 5$ the elements $β_{s} = β_{s /1, 1}$ for $s \geq 1$ are non-zero in $\mathrm{Ext}^2_{BP_BP}(BP_, BP_) $an d a r e a l l p er man e n t cy c l es, d e t ec t in g anin f ini t e$ v_2 $- p er i o d i c f ami l y in$ \pi_*\mathbb{S}_{(p)} $. T h e f i r s t m e mb er$ \beta_1 $l i es in s t e m$ 2p^2 - 2p - 2$.*

The proof of survival uses the Smith-Toda spectrum $V (1)$ and its $v_{2}$ -self-map, as in Exercise 6. At $p = 2$ and $p = 3$ the $β$ -family is more delicate: $V (1)$ either fails to be a ring spectrum or fails to admit the self-map, and only part of the family survives. The general $β_{s / j, i}$ with larger denominators fill out the rest of $Ext^{2}$ subject to the Miller-Ravenel-Wilson divisibility constraints.

Theorem ( $γ$ -family at $Ext^{3}$ ). For $p \geq 7$ the $γ$ -family $γ_{s} = δ^{(3)} (v_{3}^{s} / (p, v_{1}, v_{2}))$ gives non-zero $v_{3}$ -periodic classes in $\mathrm{Ext}^3_{BP_BP}(BP_, BP_) $, in f ini t e l y man y o f w hi c h s u r v i v e t o$ \pi_*\mathbb{S}_{(p)}$ (Ravenel-Wilson).*

The $γ$ -family is the deepest classically constructed family. Its survival analysis requires the Smith-Toda spectrum $V (2) = S / (p, v_{1}, v_{2})$ and a $v_{3}$ -self-map, which exist only for sufficiently large $p$ . The pattern $V (0), V (1), V (2), \dots$ of Smith-Toda spectra realises successive quotients $S / I_{n}$ , and the existence of $V (n)$ becomes harder as $n$ grows — a manifestation of the same obstruction theory the periodicity theorem [03.12.41-area] organises.

Theorem (general height- $n$ family). The $n$ -th Greek-letter map $\widetilde{\delta}^{(n)}: \mathrm{Ext}^0(M^n) \to \mathrm{Ext}^n(BP_) $p r o d u ces ah e i g h t -$ n $f ami l y$ \alpha^{(n)}_{t/\ldots} $f or e v er y$ n $, f r o m t h e$ v_n $- t o w er s in$ M^n $. T h ec h r o ma t i cs p ec t r a l se q u e n ce a sse mb l es t h ese ma p s, w i t h$ E_1^{n,s} = \mathrm{Ext}^s(M^n) $, in t o a co m p u t a t i o n o f t h e f u l l A d am s - N o v ik o v$ E_2$-page.*

The uniform construction is the structural payoff: every line of the Adams-Novikov $E_{2}$ is, in principle, the image of a $v_{n}$ -tower under an iterated connecting map. The $α, β, γ$ names are the cases $n = 1, 2, 3$ ; beyond gamma the families are written $α^{(n)}$ or with double-indexed Greek letters, and explicit computation becomes formidable.

Theorem (Kervaire-invariant non-existence; Hill-Hopkins-Ravenel 2016). The Kervaire-invariant elements $θ_{j} \in π_{2^{j + 1} - 2} S$ , the $2$ -primary $β$ -family-adjacent classes detected by $h_{j}^{2}$ , do not exist for $j \geq 7$ . Classes $θ_{j}$ exist for $j \leq 5$ ; the case $j = 6$ (dimension $126$ ) is the sole remaining unsettled dimension.

This is the headline modern result downstream of the Greek-letter circle of ideas: a periodicity argument, run on a carefully built $256$ -periodic equivariant spectrum, kills an entire would-be infinite family. It demonstrates that the question of which Greek-letter and Greek-letter-adjacent classes survive is genuinely hard and genuinely deep.

Synthesis. The Greek-letter construction is the central insight that the apparent randomness of the stable stems is organised, height by height, by the chromatic filtration, and putting these together gives the structural picture the Green Book is built around. The foundational reason the families exist is exactly that the chromatic resolution splices short exact sequences whose iterated connecting maps carry $v_{n}$ -towers from $Ext^{0} (M^{n})$ up to $Ext^{n} (B P_{*})$ ; this is dual to the $L_{n}$ -localisation tower on the topological side. The $α$ -family generalises the image of $J$ , the $β$ -family generalises that to $v_{2}$ -periodicity through the Smith-Toda spectra, and the bridge is the recognition that survival to homotopy is governed by the same $v_{n}$ -self-map periodicity that the Hopkins-Smith periodicity theorem [03.12.41-area] makes precise. The whole edifice — chromatic resolution, Greek-letter maps, Smith-Toda spectra, periodicity, and the Kervaire-invariant resolution — is one connected story: the homotopy of spheres is a sum of periodic pieces, one per chromatic height, and the Greek letters are their names.

Full proof set Master

Proposition (degree of the height- $n$ Greek-letter element). Let $x = v_{n}^{t} / (p^{i_{0}} v_{1}^{i_{1}} \dots v_{n - 1}^{i_{n - 1}}) \in Ext^{0} (M^{n})$ be a $v_{n}$ -tower primitive. Then its image $\widetilde{\delta}^{(n)}(x) \in \mathrm{Ext}^n_{BP_BP}(BP_, BP_)$ lies in stem* $$ \bigl| \widetilde{\delta}^{(n)}(x) \bigr| = 2t(p^n - 1) - 2\sum_{k=1}^{n-1} i_k (p^k - 1) - n, $$ where the internal degree of the numerator is $∣ v_{n}^{t} ∣ = 2 t (p^{n} - 1)$ , each $v_{k}^{i_{k}}$ in the denominator subtracts $∣ v_{k}^{i_{k}} ∣ = 2 i_{k} (p^{k} - 1)$ , and the $n$ connecting homomorphisms each lower total degree by one.

Proof. The total degree of a class in $Ext_{B P_{*} B P}^{s} (B P_{*}, B P_{*})$ is measured by the stem $t - s$ , where $t$ is the internal degree and $s$ the homological degree; the stem is what records the homotopy dimension.

First compute the internal degree of the representing fraction $x$ . The Brown-Peterson generators have $∣ v_{k} ∣ = 2 (p^{k} - 1)$ 03.13.05. The numerator $v_{n}^{t}$ contributes $2 t (p^{n} - 1)$ . Each denominator factor $v_{k}^{i_{k}}$ contributes $- 2 i_{k} (p^{k} - 1)$ , since dividing lowers internal degree; the factor $p^{i_{0}}$ has degree zero and contributes nothing. So the internal degree of $x$ is $$ \deg_{\mathrm{int}}(x) = 2t(p^n - 1) - 2\sum_{k=1}^{n-1} i_k(p^k - 1). $$ The class $x$ sits in $Ext^{0}$ , so at this stage the stem equals the internal degree.

Now apply $δ^{(n)} = δ_{n - 1} \circ \dots \circ δ_{0}$ . A connecting homomorphism of a short exact sequence of comodules preserves internal degree and raises homological degree by one. Raising homological degree by one while fixing internal degree lowers the stem $t - s$ by one. Iterating $n$ times lowers the stem by $n$ while keeping the internal degree fixed at $de g_{int} (x)$ . Therefore $$ \bigl|\widetilde{\delta}^{(n)}(x)\bigr| = \deg_{\mathrm{int}}(x) - n = 2t(p^n - 1) - 2\sum_{k=1}^{n-1} i_k(p^k - 1) - n, $$ as claimed. $□$

Corollary ( $β_{1}$ ). Taking $n = 2$ , $t = 1$ , $i_{0} = i_{1} = 1$ gives the stem of $β_{1} = β_{1/1, 1}$ : $$ |\beta_1| = 2(p^2 - 1) - 2(p - 1) - 2 = 2p^2 - 2p - 2, $$ recovering $∣ β_{1} ∣ = 10$ at $p = 3$ and $∣ β_{1} ∣ = 38$ at $p = 5$ , consistent with Exercise 2.

Proof. Substitute $n = 2$ , $t = 1$ into the proposition. The sum $\sum_{k = 1}^{1} i_{k} (p^{k} - 1) = i_{1} (p - 1) = (p - 1)$ . The internal degree is $2 (p^{2} - 1) - 2 (p - 1)$ , and subtracting $n = 2$ gives $2 (p^{2} - 1) - 2 (p - 1) - 2 = 2 p^{2} - 2 - 2 p + 2 - 2 = 2 p^{2} - 2 p - 2$ . Evaluating: $p = 3$ gives $18 - 6 - 2 = 10$ ; $p = 5$ gives $50 - 10 - 2 = 38$ . $□$

Connections Master

The chromatic spectral sequence 03.13.06 is the engine that produces every Greek-letter family: its $E_{1}$ -page $Ext^{s} (M^{n})$ supplies the $v_{n}$ -towers, and its connecting differentials are the Greek-letter maps $δ^{(n)}$ that send those towers into the Adams-Novikov $E_{2}$ . Without the chromatic resolution there is no construction; this unit is the catalogue of what that machine outputs at each height.
The Brown-Peterson spectrum and its Hopf algebroid $B P_{*} B P$ 03.13.05 provide the coefficient data: the generators $v_{n}$ with $∣ v_{n} ∣ = 2 (p^{n} - 1)$ set every degree computed here, and the right-unit formula governs which fractions are comodule primitives. The divisibility bookkeeping of Miller-Ravenel-Wilson is entirely a computation inside $(B P_{*}, B P_{*} B P)$ .
The image of $J$ 03.08.11 is the height-one Greek-letter family made concrete: the $α$ -family equals $im J$ class by class, tying the homotopy of spheres to Bernoulli-number denominators and to algebraic $K$ -theory. This is the one Greek-letter family that was understood long before the chromatic language existed.
The Adams-Novikov spectral sequence 03.12.38 is the ambient $E_{2}$ -page in which all Greek-letter elements live; survival to homotopy is a question about its differentials. The lateral comparison with the classical Adams spectral sequence — where the Kervaire classes $θ_{j}$ are detected by $h_{j}^{2}$ — shows the same families seen through two filtrations.

Historical & philosophical context Master

The Greek-letter notation was introduced by Miller, Ravenel, and Wilson in their 1977 Annals paper ^{[Miller-Ravenel-Wilson 1977]}, which crystallised a decade of work on periodic phenomena in stable homotopy. The $α$ -family had a longer prehistory: it is the image of the $J$ -homomorphism, computed by Adams in his 1966 series of papers ^{[Adams 1966]}, where the link to Bernoulli numbers and the denominators of $B_{2 k} /4 k$ first appeared. The $β$ -family emerged from Smith's and Toda's construction of the spectra $V (n)$ around 1970 ^{[Smith 1970]}, realising the quotients $S / I_{n}$ and their self-maps. What Miller-Ravenel-Wilson added was the organising principle: a single chromatic resolution whose iterated connecting maps produce every family uniformly, one Greek letter per height, transforming a collection of ad hoc constructions into a systematic theory.

Philosophically, the Greek-letter program embodies a shift from computation to structure. Before the chromatic viewpoint, the homotopy groups of spheres were a table of numbers to be extended dimension by dimension. After it, the table became a superposition of periodic layers, each with its own arithmetic — $K$ -theory and Bernoulli numbers at height one, modular forms and elliptic curves at height two. Ravenel's Green Book ^{[Ravenel 2004]} is the canonical synthesis. The culmination of this structural turn is the Hill-Hopkins-Ravenel resolution of the Kervaire invariant problem ^{[Hill-Hopkins-Ravenel 2016]}: a periodicity argument, descended from the same circle of ideas, settling a fifty-year-old geometric question about framed manifolds by showing that an entire would-be family of homotopy elements simply does not exist.

Bibliography Master

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

@book{Ravenel2004,
  author    = {Ravenel, Douglas C.},
  title     = {Complex Cobordism and Stable Homotopy Groups of Spheres},
  edition   = {2nd},
  publisher = {AMS Chelsea Publishing},
  year      = {2004},
  note      = {The Green Book; Ch. 4 (BP and BP_*BP), Ch. 5 (chromatic spectral sequence, Greek-letter elements)},
}

@article{Adams1966,
  author  = {Adams, J. F.},
  title   = {On the groups {$J(X)$}, {IV}},
  journal = {Topology},
  volume  = {5},
  pages   = {21--71},
  year    = {1966},
}

@article{Smith1970,
  author  = {Smith, Larry},
  title   = {On realizing complex bordism modules: applications to the stable homotopy of spheres},
  journal = {American Journal of Mathematics},
  volume  = {92},
  pages   = {793--856},
  year    = {1970},
}

@article{HillHopkinsRavenel2016,
  author  = {Hill, Michael A. and Hopkins, Michael J. and Ravenel, Douglas C.},
  title   = {On the nonexistence of elements of {Kervaire} invariant one},
  journal = {Annals of Mathematics},
  volume  = {184},
  number  = {1},
  pages   = {1--262},
  year    = {2016},
}

Prerequisites

03.13.06 pending
03.13.05 pending

Tier anchors

beginner: named infinite families of elements in the homotopy of spheres, one Greek letter per chromatic height — informal opening of Ravenel 2004 *Complex Cobordism and Stable Homotopy Groups of Spheres* (2nd ed., AMS Chelsea) Ch. 1 'overview'; Hatcher *Algebraic Topology* §4.L survey of stable stems
intermediate: Miller-Ravenel-Wilson 1977 *Periodic phenomena in the Adams-Novikov spectral sequence* (Ann. of Math. 106), §3-§5; Ravenel 2004 *Complex Cobordism and Stable Homotopy Groups of Spheres* (the Green Book) Ch. 5 §5.1-§5.3; Adams 1966 *On the groups J(X) IV* (Topology 5) for the image of J
master: Ravenel 2004 (the Green Book) Ch. 5 (Greek-letter construction, chromatic resolution) and Ch. 4 ($BP_*BP$); Miller-Ravenel-Wilson 1977 (Ann. of Math. 106, 469-516); Smith 1970/Toda 1971 ($V(1)$ and the $\beta$-family); Adams 1966 *On the groups J(X) IV*; Hill-Hopkins-Ravenel 2016 *On the nonexistence of elements of Kervaire invariant one* (Ann. of Math. 184) for $\theta_j$

References

Ravenel — Complex Cobordism and Stable Homotopy Groups of Spheres · 2nd ed., AMS Chelsea 2004 (the Green Book); Ch. 5 the chromatic spectral sequence and Greek-letter elements, Ch. 4 BP and BP_*BP
Miller-Ravenel-Wilson — Periodic phenomena in the Adams-Novikov spectral sequence · Ann. of Math. 106 (1977), 469-516 — the originator paper: chromatic resolution, alpha/beta/gamma families, divisibility bookkeeping
Adams — On the groups J(X) IV · Topology 5 (1966), 21-71 — image of J, the height-1 alpha-family, Bernoulli-number denominators
Smith — On realizing complex bordism modules · Amer. J. Math. 92 (1970), 793-856; Toda 1971 — the Smith-Toda spectrum V(1) and the beta-family for p >= 5
Hill-Hopkins-Ravenel — On the nonexistence of elements of Kervaire invariant one · Ann. of Math. 184 (2016), 1-262 — theta_j does not exist for j >= 7; Kervaire-invariant elements beta-family-adjacent

Estimated time

beginner: 20m
intermediate: 50m
master: 85m