Sinha Spectral Sequence in Knot Theory

• Sinha Spectral Sequence is a computational tool based on a cosimplicial model of compactified configuration spaces and the Kontsevich operad to approximate the (co)homology of long knots.
• It employs explicit chain-level models using Fox–Neuwirth trees and graph complexes to compute higher differentials and reveal noncollapse phenomena.
• Its comparison with the Vassiliev spectral sequence highlights coefficient-sensitive behavior, operadic formality issues, and convergence challenges in low codimension.

Searching arXiv for the cited papers to ground the article in current metadata and source context. Search results found for the requested arXiv ids:

• (Moriya, 2023) — "Differentials of Sinha's spectral sequence for long knots in codimension one"
• (Marino, 29 May 2025) — "A Fox-Neuwirth Basis for the Sinha Spectral Sequence"
• (Marino et al., 23 Apr 2025) — "Non collapse of the Sinha spectral sequence for knots in R^3"
• (Moriya, 28 Sep 2025) — "Embedding calculus and Vassiliev spectral sequence"

These are the main sources I will use. The Sinha spectral sequence is the spectral sequence attached to Sinha’s cosimplicial model of compactified configuration spaces, constructed from the multiplicative structure on the Kontsevich operad and used to approximate the (co)homology of spaces of long knots, or long knots modulo immersions, depending on the model under consideration. In the homological formulation, taking singular or cellular chains on the cosimplicial space produces a bicomplex whose first page is $E^1_{p,q}=H_q(\Konts_m(p))$ and which converges to $H_{p+q}(\overline{\Emb}_m)$; in the cohomological formulation, normalized cochains on a cosimplicial model $X^\bullet$ give $E_1^{p,q}=H^q(X^p)/(degeneracies)$ and $E_2^{p,q}=H^p(\{E_1^{*,q},d_1\})$ (Marino, 29 May 2025). Recent work has emphasized three aspects of the subject: explicit chain-level models for the pages and differentials, combinatorial realizations via Fox–Neuwirth trees, and comparisons with the Vassiliev spectral sequence and operadic formality phenomena (Moriya, 28 Sep 2025).

1. Cosimplicial source and target of the construction

Fix $m\ge 2$. The starting point is the compactified configuration space model

$\Konts_m(n)\subset (S^{m-1})^{\binom n2},$

defined as the closure of the map

$\phi=(\phi_{ij})_{1\le i<j\le n}:\Conf_n(\mathbb{R}^m)\longrightarrow (S^{m-1})^{\binom n2},\qquad \phi_{ij}(x)=\frac{x_j-x_i}{\|x_j-x_i\|}.$

Sinha observed that the standard little-disks doubling maps

$d_i^S:\Konts_m(n)\longrightarrow \Konts_m(n+1)$

make $\Konts_m(\bullet)$ into a cosimplicial space, and that its totalization recovers the Taylor-tower models for

$H_{p+q}(\overline{\Emb}_m)$0

(Marino, 29 May 2025).

In a chain-level presentation, one forms the bicomplex

$H_{p+q}(\overline{\Emb}_m)$1

with vertical differential given by the chain boundary and horizontal differential given by the alternating sum of the cofaces. This yields the homology spectral sequence

$H_{p+q}(\overline{\Emb}_m)$2

A cohomological version is obtained from a cosimplicial space $H_{p+q}(\overline{\Emb}_m)$3 by passing to normalized cochains

$H_{p+q}(\overline{\Emb}_m)$4

Its total complex carries the filtration by simplicial degree $H_{p+q}(\overline{\Emb}_m)$5, and the resulting spectral sequence converges, under mild connectivity hypotheses, to the cohomology of the totalization (Moriya, 28 Sep 2025).

For $H_{p+q}(\overline{\Emb}_m)$6, Moriya also describes an equivalent punctured-knot model $H_{p+q}(\overline{\Emb}_m)$7 arising from Goodwillie–Weiss embedding calculus. There one assigns to a partition $H_{p+q}(\overline{\Emb}_m)$8 a space $H_{p+q}(\overline{\Emb}_m)$9 of embeddings of a punctured interval into $X^\bullet$0, and one has

$X^\bullet$1

A combinatorial functor from the partition poset to $X^\bullet$2 produces a cosimplicial replacement, and passing to normalized cochains recovers the Bousfield–Kan cohomology spectral sequence (Moriya, 28 Sep 2025).

2. First pages and algebraic descriptions

The first page is controlled by the cohomology or homology of configuration spaces. Because each $X^\bullet$3 in Moriya’s codimension-one setting, one finds

$X^\bullet$4

In characteristic $X^\bullet$5 or $X^\bullet$6, the $X^\bullet$7-page is generated by the Arnold classes

$X^\bullet$8

with relations

$X^\bullet$9

(Moriya, 2023).

For $E_1^{p,q}=H^q(X^p)/(degeneracies)$0, one has

$E_1^{p,q}=H^q(X^p)/(degeneracies)$1

modulo the relations

$E_1^{p,q}=H^q(X^p)/(degeneracies)$2

where $E_1^{p,q}=H^q(X^p)/(degeneracies)$3. In the compactified Sinha model, the normalized cohomology is identified with the subquotient of this algebra consisting of normalized monomials in the $E_1^{p,q}=H^q(X^p)/(degeneracies)$4 (Moriya, 28 Sep 2025).

The $E_1^{p,q}=H^q(X^p)/(degeneracies)$5-differential is induced by the alternating sum of simplicial cofaces. In the cohomological indexing,

$E_1^{p,q}=H^q(X^p)/(degeneracies)$6

Turchin’s graph-complex description identifies $E_1^{p,q}=H^q(X^p)/(degeneracies)$7 with the cohomology of a complex spanned by Feynman graphs on $E_1^{p,q}=H^q(X^p)/(degeneracies)$8 labeled vertices, with cohomological degree $E_1^{p,q}=H^q(X^p)/(degeneracies)$9. In that model the next non-trivial differential has bidegree $E_2^{p,q}=H^p(\{E_1^{*,q},d_1\})$0 and can be written combinatorially as

$E_2^{p,q}=H^p(\{E_1^{*,q},d_1\})$1

sending a graph $E_2^{p,q}=H^p(\{E_1^{*,q},d_1\})$2 to the sum of all contractions of a single 3-valent subgraph with one 3-cycle; in practice it vanishes on the diagonal $E_2^{p,q}=H^p(\{E_1^{*,q},d_1\})$3 by degree reasons (Moriya, 28 Sep 2025).

A recurring theme is the distinction between rational and positive-characteristic behavior. In favorable cases, such as $E_2^{p,q}=H^p(\{E_1^{*,q},d_1\})$4 with $E_2^{p,q}=H^p(\{E_1^{*,q},d_1\})$5-coefficients, the sequence collapses at $E_2^{p,q}=H^p(\{E_1^{*,q},d_1\})$6. By contrast, explicit nonzero higher differentials occur in characteristic $E_2^{p,q}=H^p(\{E_1^{*,q},d_1\})$7 and $E_2^{p,q}=H^p(\{E_1^{*,q},d_1\})$8 in low-dimensional settings, so collapse cannot be treated as coefficient-independent.

3. Fox–Neuwirth trees and combinatorial models

A major simplification is provided by the Fox–Neuwirth basis. A Fox–Neuwirth tree of height $E_2^{p,q}=H^p(\{E_1^{*,q},d_1\})$9 on $m\ge 2$0 leaves may be given by

$m\ge 2$1

with compatibility

$m\ge 2$2

Geometrically, $m\ge 2$3 encodes a stratum $m\ge 2$4 determined by successive equalities in the first $m\ge 2$5 coordinates and inequalities in the $m\ge 2$6st coordinate. One shows that $m\ge 2$7 is a ball of dimension $m\ge 2$8, and that the one-point compactification admits a $m\ge 2$9-equivariant regular CW decomposition whose open cells are these strata (Marino, 29 May 2025).

Blagojević–Ziegler’s finite regular CW complex $\Konts_m(n)\subset (S^{m-1})^{\binom n2},$0 gives one cell

$\Konts_m(n)\subset (S^{m-1})^{\binom n2},$1

in dual dimension $\Konts_m(n)\subset (S^{m-1})^{\binom n2},$2. The cellular chain complex is

$\Konts_m(n)\subset (S^{m-1})^{\binom n2},$3

Dualizing gives a cochain model for $\Konts_m(n)\subset (S^{m-1})^{\binom n2},$4. The paper “A Fox-Neuwirth Basis for the Sinha Spectral Sequence” states that this presentation is dramatically smaller than, for example, the McClure–Smith or good-operad model, and that it yields a combinatorial presentation of the Sinha spectral sequence for all dimensions $\Konts_m(n)\subset (S^{m-1})^{\binom n2},$5 and all coefficients (Marino, 29 May 2025).

The face posets $\Konts_m(n)\subset (S^{m-1})^{\binom n2},$6 assemble into a semicosimplicial poset. Via the nerve construction one obtains

$\Konts_m(n)\subset (S^{m-1})^{\binom n2},$7

and the same McClure–Smith prescription giving $\Konts_m(n)\subset (S^{m-1})^{\binom n2},$8 a cosimplicial structure transports to this barycentric subdivision model. On a chain of trees one has

$\Konts_m(n)\subset (S^{m-1})^{\binom n2},$9

where $\phi=(\phi_{ij})_{1\le i<j\le n}:\Conf_n(\mathbb{R}^m)\longrightarrow (S^{m-1})^{\binom n2},\qquad \phi_{ij}(x)=\frac{x_j-x_i}{\|x_j-x_i\|}.$0 is obtained by inserting a tiny fork at leaf $\phi=(\phi_{ij})_{1\le i<j\le n}:\Conf_n(\mathbb{R}^m)\longrightarrow (S^{m-1})^{\binom n2},\qquad \phi_{ij}(x)=\frac{x_j-x_i}{\|x_j-x_i\|}.$1; degeneracies collapse leaves. The associated spectral sequence coincides with Sinha’s from page one onward (Marino, 29 May 2025).

For $\phi=(\phi_{ij})_{1\le i<j\le n}:\Conf_n(\mathbb{R}^m)\longrightarrow (S^{m-1})^{\binom n2},\qquad \phi_{ij}(x)=\frac{x_j-x_i}{\|x_j-x_i\|}.$2, Marino and Salvatore refine this further by homotopy transfer from the barycentric bicomplex to a multicomplex

$\phi=(\phi_{ij})_{1\le i<j\le n}:\Conf_n(\mathbb{R}^m)\longrightarrow (S^{m-1})^{\binom n2},\qquad \phi_{ij}(x)=\frac{x_j-x_i}{\|x_j-x_i\|}.$3

with

$\phi=(\phi_{ij})_{1\le i<j\le n}:\Conf_n(\mathbb{R}^m)\longrightarrow (S^{m-1})^{\binom n2},\qquad \phi_{ij}(x)=\frac{x_j-x_i}{\|x_j-x_i\|}.$4

Here $\phi=(\phi_{ij})_{1\le i<j\le n}:\Conf_n(\mathbb{R}^m)\longrightarrow (S^{m-1})^{\binom n2},\qquad \phi_{ij}(x)=\frac{x_j-x_i}{\|x_j-x_i\|}.$5 is the usual cellular differential, $\phi=(\phi_{ij})_{1\le i<j\le n}:\Conf_n(\mathbb{R}^m)\longrightarrow (S^{m-1})^{\binom n2},\qquad \phi_{ij}(x)=\frac{x_j-x_i}{\|x_j-x_i\|}.$6 is the doubling-point horizontally operator induced by cofaces, and $\phi=(\phi_{ij})_{1\le i<j\le n}:\Conf_n(\mathbb{R}^m)\longrightarrow (S^{m-1})^{\binom n2},\qquad \phi_{ij}(x)=\frac{x_j-x_i}{\|x_j-x_i\|}.$7 are higher horizontal maps coming from simultaneous doublings of $\phi=(\phi_{ij})_{1\le i<j\le n}:\Conf_n(\mathbb{R}^m)\longrightarrow (S^{m-1})^{\binom n2},\qquad \phi_{ij}(x)=\frac{x_j-x_i}{\|x_j-x_i\|}.$8 or $\phi=(\phi_{ij})_{1\le i<j\le n}:\Conf_n(\mathbb{R}^m)\longrightarrow (S^{m-1})^{\binom n2},\qquad \phi_{ij}(x)=\frac{x_j-x_i}{\|x_j-x_i\|}.$9 points. This produces an explicit model for the first three pages of the mod-$d_i^S:\Konts_m(n)\longrightarrow \Konts_m(n+1)$0 spectral sequence in $d_i^S:\Konts_m(n)\longrightarrow \Konts_m(n+1)$1 (Marino et al., 23 Apr 2025).

4. Explicit higher differentials

The first explicit higher differentials in codimension one were computed by Moriya for the cohomology spectral sequence of the space of long knots modulo immersions in codimension one, over a field of characteristic $d_i^S:\Konts_m(n)\longrightarrow \Konts_m(n+1)$2 or $d_i^S:\Konts_m(n)\longrightarrow \Konts_m(n+1)$3. In characteristic $d_i^S:\Konts_m(n)\longrightarrow \Konts_m(n+1)$4, on the $d_i^S:\Konts_m(n)\longrightarrow \Konts_m(n+1)$5-page for $d_i^S:\Konts_m(n)\longrightarrow \Konts_m(n+1)$6 one considers

$d_i^S:\Konts_m(n)\longrightarrow \Konts_m(n+1)$7

This class is a $d_i^S:\Konts_m(n)\longrightarrow \Konts_m(n+1)$8-cycle but not a boundary. By lifting $d_i^S:\Konts_m(n)\longrightarrow \Konts_m(n+1)$9 to an explicit chain-level representative in the total complex of a triple complex built from fat-diagonal Thom spaces, Moriya constructs graph-labeled chains indexed by the three 2-edge graphs

$\Konts_m(\bullet)$0

and proves that

$\Konts_m(\bullet)$1

At chain level, each $\Konts_m(\bullet)$2 is built from a condensed embedding $\Konts_m(\bullet)$3 together with two edge-contractions $\Konts_m(\bullet)$4, and the calculation is organized through the total differential

$\Konts_m(\bullet)$5

on the triple complex (Moriya, 2023).

In characteristic $\Konts_m(\bullet)$6, Moriya studies $\Konts_m(\bullet)$7, $\Konts_m(\bullet)$8, $\Konts_m(\bullet)$9, and the class

$H_{p+q}(\overline{\Emb}_m)$00

associated to the four three-edge graphs

$H_{p+q}(\overline{\Emb}_m)$01

One checks that $H_{p+q}(\overline{\Emb}_m)$02, constructs an explicit chain $H_{p+q}(\overline{\Emb}_m)$03 in the truncated total complex, and proves

$H_{p+q}(\overline{\Emb}_m)$04

These results provide explicit obstructions already on the first few pages, even though convergence in codimension one remains unclear (Moriya, 2023).

A distinct noncollapse result is obtained for knots in $H_{p+q}(\overline{\Emb}_m)$05 over $H_{p+q}(\overline{\Emb}_m)$06. Marino and Salvatore give an explicit description up to the third page of the Sinha homology mod $H_{p+q}(\overline{\Emb}_m)$07 spectral sequence. They show that $H_{p+q}(\overline{\Emb}_m)$08 by dimension reasons, so $H_{p+q}(\overline{\Emb}_m)$09, but then detect a nontrivial third-page differential

$H_{p+q}(\overline{\Emb}_m)$10

where $H_{p+q}(\overline{\Emb}_m)$11 is the first nontrivial Vassiliev class. The computation uses the multicomplex operators $H_{p+q}(\overline{\Emb}_m)$12 and a computer-assisted calculation in MATLAB starting from a $H_{p+q}(\overline{\Emb}_m)$13-dimensional Vassiliev cycle $H_{p+q}(\overline{\Emb}_m)$14 given by a sum of $H_{p+q}(\overline{\Emb}_m)$15 trees (Marino et al., 23 Apr 2025).

5. Duality, fat diagonals, and chain-level realizations

A central technical development is Moriya’s point-set-level duality between punctured-knot models and fat-diagonal Thom-space models. For each partition $H_{p+q}(\overline{\Emb}_m)$16 of $H_{p+q}(\overline{\Emb}_m)$17, one constructs a pointed space $H_{p+q}(\overline{\Emb}_m)$18 which is essentially the Thom space of a tubular neighborhood of the fat diagonal in $H_{p+q}(\overline{\Emb}_m)$19. Explicit collapsing maps and Čech covers show that the resulting Thom-space cosimplicial spectrum is Atiyah-dual to the punctured-knot cosimplicial spectrum (Moriya, 2023).

Passing to chains gives a zig-zag of quasi-isomorphisms

$H_{p+q}(\overline{\Emb}_m)$20

which identifies Sinha’s spectral sequence with the homology spectral sequence of the triple complex

$H_{p+q}(\overline{\Emb}_m)$21

In that complex, the total differential splits as

$H_{p+q}(\overline{\Emb}_m)$22

where $H_{p+q}(\overline{\Emb}_m)$23 is induced by merging adjacent pieces in the partition, $H_{p+q}(\overline{\Emb}_m)$24 is the singular-chain differential, and $H_{p+q}(\overline{\Emb}_m)$25 is the Čech differential coming from the Thom-space coverings. Filtering by $H_{p+q}(\overline{\Emb}_m)$26, or equivalently by $H_{p+q}(\overline{\Emb}_m)$27, recovers Sinha’s spectral sequence with differential

$H_{p+q}(\overline{\Emb}_m)$28

in the standard grading convention (Moriya, 2023).

This chain-level translation has two consequences. First, it makes higher differentials concrete: each differential can be tracked graphically by collapsing edges or sliding strands. Second, it supplies a geometric bridge between cosimplicial compactified configuration spaces and the combinatorics of graphs and trees. A plausible implication is that the analytic and combinatorial descriptions are not competing models but different presentations of the same filtration data.

6. Relation to the Vassiliev spectral sequence, formality, and collapse phenomena

The relation with Vassiliev theory is now highly structured. Moriya proves that the Vassiliev spectral sequence and the Sinha spectral sequence have isomorphic $H_{p+q}(\overline{\Emb}_m)$29-pages if the coefficient ring is a field, and that under the degree shift

$H_{p+q}(\overline{\Emb}_m)$30

there is a canonical isomorphism

$H_{p+q}(\overline{\Emb}_m)$31

In particular,

$H_{p+q}(\overline{\Emb}_m)$32

The comparison proceeds by replacing the punctured-knot model with a Thom-space model, comparing normalized cochains, and then resolving the Thom-space model by a simplicial space built from fat diagonals so as to match Vassiliev’s simplicial resolution of the discriminant (Moriya, 28 Sep 2025).

This comparison yields consequences for finite-type invariants. The diagonal of $H_{p+q}(\overline{\Emb}_m)$33 satisfies

$H_{p+q}(\overline{\Emb}_m)$34

while the space of weight systems of degree $H_{p+q}(\overline{\Emb}_m)$35 is identified with the classical $H_{p+q}(\overline{\Emb}_m)$36-page in bidegree $H_{p+q}(\overline{\Emb}_m)$37. Moriya’s criterion states that over a PID $H_{p+q}(\overline{\Emb}_m)$38 one has

$H_{p+q}(\overline{\Emb}_m)$39

if and only if the Sinha spectral sequence on the diagonal bidegree $H_{p+q}(\overline{\Emb}_m)$40 degenerates at $H_{p+q}(\overline{\Emb}_m)$41 for all $H_{p+q}(\overline{\Emb}_m)$42 (Moriya, 28 Sep 2025).

Formality results sharply separate rational from positive-characteristic behavior. In codimension at least $H_{p+q}(\overline{\Emb}_m)$43, operadic formality implies that the Sinha spectral sequence collapses at $H_{p+q}(\overline{\Emb}_m)$44 over $H_{p+q}(\overline{\Emb}_m)$45. For knots in $H_{p+q}(\overline{\Emb}_m)$46, Moriya states that the Sinha sequence for knots modulo immersions collapses at $H_{p+q}(\overline{\Emb}_m)$47 over any field by $H_{p+q}(\overline{\Emb}_m)$48-completed GT-action arguments, and that injectivity of the map from knots modulo immersions to real knots on $H_{p+q}(\overline{\Emb}_m)$49 implies degeneration for real knots when $H_{p+q}(\overline{\Emb}_m)$50 (Moriya, 28 Sep 2025). By contrast, Moriya’s codimension-one calculations show that the nonzero $H_{p+q}(\overline{\Emb}_m)$51 in characteristic $H_{p+q}(\overline{\Emb}_m)$52 implies planar non-formality of the standard map

$H_{p+q}(\overline{\Emb}_m)$53

over $H_{p+q}(\overline{\Emb}_m)$54, and the nonzero $H_{p+q}(\overline{\Emb}_m)$55 in characteristic $H_{p+q}(\overline{\Emb}_m)$56 reproves that $H_{p+q}(\overline{\Emb}_m)$57 is not formal as a planar operad over $H_{p+q}(\overline{\Emb}_m)$58 (Moriya, 2023).

Two open-ended issues therefore remain central. One is convergence in low codimension: Goodwillie–Weiss embedding calculus is known to converge only in codimension at least $H_{p+q}(\overline{\Emb}_m)$59, so convergence for $H_{p+q}(\overline{\Emb}_m)$60 remains open, and for $H_{p+q}(\overline{\Emb}_m)$61 convergence beyond the diagonal is also stated to be open (Moriya, 2023). The other is collapse: over $H_{p+q}(\overline{\Emb}_m)$62 and in high codimension, collapse at $H_{p+q}(\overline{\Emb}_m)$63 is standard, whereas over finite fields explicit higher differentials occur. This suggests that the Sinha spectral sequence should be regarded not merely as a bookkeeping device for known invariants, but as a setting in which coefficient-sensitive and operad-sensitive phenomena become visible at the chain level.