Torsion cycles on Fermat varieties

arXiv:2603.00723 · math.AG, math.NT · Submitted 2026-02-28 · Read on arXiv

Listen

Radio episode about this paper

Transcript

Introduction to the show: ident: Mathematics. The day in mathematics: new theorems, new proofs, and what they open.

Ada: We're looking at the paper: "Torsion cycles on Fermat varieties".

Title and authors: Ada: So we’re looking at this paper called "Torsion cycles on Fermat varieties." It’s about extending some older results from Manin and Drinfeld to much more complicated spaces, specifically higher codimensional null homologous cycles.

Emmy: That sounds like they're taking the idea that a divisor of degree zero is torsion in the Jacobian and pushing it up to these higher dimensional structures, which is pretty ambitious.

Ada: Exactly. They’re doing this using mixed Hodge theory to see if that same torsion property holds in these more complex settings, and they generalize Rohrlich's theorem in this way.

Kolya: I'm curious how the structure of these higher cycles plays into that torsion claim, especially since we're moving beyond just the Jacobian and looking at intermediate Jacobians.

Ada: Right. The main point is that for a linear subvariety of a Fermat hypersurface, they establish that any null homologous cycle supported on it must be torsion in the intermediate Jacobian.

Emmy: That means if you have one of these cycles, its structure is constrained by the geometry of the ambient variety and those linear sections.

The paper's summary: Ada: So, to summarize what they did, this paper uses mixed Hodge structures to connect divisors and higher Chow cycles to extensions in the category of mixed motives.

Emmy: They show that a motivic cycle on a Fermat variety determines an extension in a specific group of mixed Hodge structures. If that extension splits, it means the cycle is torsion in the intermediate Jacobian.

Ada: That's the mechanism they use to prove Rohrlich's theorem again, but this time for higher co-dimensions, where they get a weaker statement about torsion in the intermediate Jacobian instead of just the Jacobian itself.

Kolya: What does that mean practically? Does it give us a way to bound these cycles or understand their behavior better?

Emmy: It gives us a tool to determine if a cycle is torsion by checking if an exact sequence of mixed Hodge structures splits, which is essentially what they are doing here.

The paper's improvements: Ada: One major improvement they detail is how they show the splitting of the exact sequence in Theorem three point one <ref:2603.00723#pg1>. They introduce a specific operator, let's call it T, that acts on a certain cohomology group to provide a complementary pure Hodge substructure.

Emmy: That operator T is induced by an automorphism of the Fermat variety and it acts by multiplication by dn plus one on the relevant cohomology group, and they show that for n greater than one, this action becomes trivial in the space they are looking at.

Ada: That trivial action simplifies things a lot because it makes proving that exact sequence splits much more tractable when n is greater than one. They note that the argument for splitting doesn't work if n equals one, because of some issues with the linear subvarieties not being irreducible in that case.

Kolya: So, they’ve found a specific way to force the splitting using this operator T, which helps them make progress when dealing with higher codimensions.

Conclusion: Ada: To wrap up, the paper shows that these natural constructions don't produce cycles that are interesting in terms of producing non-torsion behavior on Fermat varieties. They show that a motivic cycle supported on linear hypersurfaces has a regulator which is torsion in the Deligne cohomology group H2p minus q D (Fnd, Q p).

Emmy: They also suggest that for certain higher Chow groups, like CH2(F1d, two) or CH3(F4d, one), non-torsion cycles might actually exist <ref:2603.00723#pg1>. These are the cases they think we should focus on for further study.

Kolya: I just wonder if these non-torsion cycles they mention are the ones that survive in a real application or if they're mostly theoretical constructions based on these specific algebraic structures.

Ada: Well, what this paper does is build a framework connecting motivic cycles to extensions in mixed motives and Deligne cohomology, and it points us toward which higher Chow groups are most interesting for finding those non-torsion examples. That’s the gist of "Torsion cycles on Fermat varieties."

Indian Statistical Institute Bangalore

math.AG, math.NT

Submitted: 2026-02-28

Updated: 2026-11-18

Comments: 22 pages

License: http://creativecommons.org/licenses/by/4.0/

The gist: A theorem of Manin and Drinfeld states that any divisor of degree 0 on the cusps of a modular curve is torsion in the Jacobian…

Key concepts

Mixed Hodge Structures (MHS)
MHS are mathematical tools used to study the structure of cohomology groups on algebraic varieties. The paper uses them to relate divisors and cycles to intermediate Jacobians, determining if a cycle is torsion by checking if a related exact sequence splits.
Higher Chow Cycles
These are generalizations of classical algebraic cycles that can be defined using higher-order relations involving functions on the variety. The paper connects these cycles to extensions in the category of mixed motives, linking them to specific cohomology groups.
Regulator Map (reg(y))
The regulator map is a tool used to relate algebraic cycles (like motivic cycles) to analytic objects, specifically Deligne cohomology groups. Proving that this regulator is zero or torsion helps classify the nature of the cycle.
Intermediate Jacobian
The intermediate Jacobian is a complex torus constructed from the cohomology of a variety. The paper investigates whether null-homologous cycles supported on specific sets are torsion within this torus, which relates to their algebraic structure.

Terminology

Summary

A theorem of Manin and Drinfeld states that any divisor of degree 0 on the cusps of a modular curve is torsion in the Jacobian, and this result is generalized to higher codimensional null homologous cycles as well as higher Chow cycles on Fermat varieties. p. 0

Where the paper says it

Torsion cycles on Fermat varieties Ramesh Sreekantan Indian Statistical Institute Bangalore March 3, 2026 Abstract A theorem of Manin and Drinfeld [Man72],[Dri73] states that any divisor of degree 0 on the cusps of a modular curve is torsion in the Jacobian. An elegant proof of this result was provided by Elkik [Elk90] using mixed Hodge theory.

Generalization of Rohrlich's Theorem

The paper re-proves Rohrlich’s theorem using methods similar to Elkik’s proof, and extends this to higher co-dimensions with a weaker statement concerning the intermediate Jacobian. Theorem 1.1 establishes that for the Fermat curve of degree d, any divisor of degree 0 supported on the set of 'Fermat points' is torsion in the Jacobian. This result is generalized to linear subvarieties, stating that any null-homologous cycle supported on such a set is torsion in the pth-intermediate Jacobian. p. 1

Where the paper says it

Here H 2p S p d (F n d , Q) 0 = Ker{H 2p S p d (F n d , Q) → H 2p (F n d , Q)} is the kernel of the cycle class map and is generated by linear combinations of the linear subvarieties which are null-homologous. As a result, any null-homologous cycle supported on S p d is torsion in the p th-intermediate Jacobian.

Hodge Structure Framework

The proof relies on mixed Hodge structures (MHS) and extensions thereof to relate divisors to intermediate Jacobians. A key step involves showing that the exact sequence related to null-homologous cycles splits as a sequence of MHSs. p. 3

Where the paper says it

Let Y = X − |D|. Let H 2p |D| (X, Q) 0 = Ker{H 2p |D| (X, Q) → H 2p (X, Q)} Then any null-homologous divisor Z supported on D is torsion in the p th intermediate Jacobian if and only if the exact sequence 0 → H 2p−1 (X, Q) → H 2p−1 (Y, Q) → H 2p |D| (X, Q) 0 → 0 splits.

This splitting is crucial for concluding that the cycle is torsion in the intermediate Jacobian. p. 4
Where the paper says it

For higher codimensional varieties it is not necessarily the case that a cycle which is trivial in the intermediate Jacobian is trivial in the Chow group. Hence from the splitting of the exact sequence one can only conclude that the cycle is 0 in the intermediate Jacobian or is torsion if the Q-mixed Hodge structure splits.

Motivic Cycles and Regulator Maps

The paper connects higher Chow cycles to extensions in the category of mixed motives, asserting that a motivic cycle determines an extension in Ext1MR(Q(−p), h2p−q−1(X, Q)). Furthermore, it proves that for motivic cycles on Fermat varieties supported on linear hypersurfaces, the regulator reg(y) is 0 in the Deligne cohomology group H2p−qD (Fnd, Q(p)). p. 1

Where the paper says it

, fIq ) where ZI are linear subvarieties of codimension p−q and fIj are functions on ZI with divisorial support on linear subvarieties of ZI . Then the regulator reg(y) is 0 in the Deligne cohomology group H 2p−q D (F n d , Q(p)).

Non-Torsion Cycles

Despite these results showing certain natural constructions do not produce interesting cycles, the paper notes that non-torsion cycles might exist in higher Chow groups of Fermat varieties. Examples include the Gross-Schoen modified diagonal cycles and cycles constructed by Ross and Sarkar. p. 9

Where the paper says it

Then (Ca, fbc) + (Cb, fca) + (Cc, fab) is an element of CHp (F n d , 1) as divCa (fbc) + divCb (fca) + divCc (fab) Cab − Cac + Cbc − Cba + Cca − Ccb = 0 More generally one can construct cycles in CHp (F n d , q) along the same lines. Ross [Ros94] constructed cycles in the group CH2 (F 1 d , 2) using pairs of functions supported on the Fermat points.

These cycles are likely to map to non-torsion null homologous cycles on F2p−1d. p. 7
Where the paper says it

The case when n = 2p − 1 is the only case where one expects potentially interesting cycles. For instance CH1 hom(F 1 d ), CH2 hom(F 3 d ), CH3 hom(F 5 d ) and in fact there do exist interesting non-torsion cycles in these groups as we will see in Section 5.

Structural Limitations

The argument for the splitting of the exact sequence does not work when n = 1, as the divisor of a function is not necessarily fixed by the action of G1d since the 0-dimensional linear subvarieties are not irreducible. The case where n > 1 is considered more tractable for proving the splitting. p. 4

Where the paper says it

For higher codimensional varieties it is not necessarily the case that a cycle which is trivial in the intermediate Jacobian is trivial in the Chow group. Hence from the splitting of the exact sequence one can only conclude that the cycle is 0 in the intermediate Jacobian or is torsion if the Q-mixed Hodge structure splits.

The only really interesting higher Chow groups are CHp(F2p−q−1d, q). p. 9
Where the paper says it

Ross [Ros94] constructed cycles in the group CH2 (F 1 d , 2) using pairs of functions supported on the Fermat points. There are some trivial elements in the higher Chow groups of the form (Z, a1, .

Conclusion on Torsion

The paper concludes that the theorems show certain 'natural' constructions do not produce interesting cycles, but it does not imply that the Chow groups of Fermat varieties are trivial. The Bloch-Beilinson conjectures suggest these cycles are necessarily torsion in the Chow group itself. p. 1

Where the paper says it

As a result, any null-homologous cycle supported on S p d is torsion in the p th-intermediate Jacobian. A conjecture of Beilinson and Bloch asserts that the Abel-Jacobi map is injective for varieties defined over number fields, hence this would imply that the cycle is actually torsion in the Chow group itself.

However, for certain groups, non-torsion cycles are known to exist. p. 7
Where the paper says it

The case when n = 2p − 1 is the only case where one expects potentially interesting cycles. For instance CH1 hom(F 1 d ), CH2 hom(F 3 d ), CH3 hom(F 5 d ) and in fact there do exist interesting non-torsion cycles in these groups as we will see in Section 5.

The argument does not work when n = 1 because there are examples of non-trivial cycles. p. 7
Where the paper says it

Consider the element T = 1 |Gn d | X g∈Gn d g acting on H2p−1 (F n d − S p d , Q). Since H 2p−1 (F n d , C) ≃ M χ∈An d V (χ) for v ∈ V (χ), T(v) = 1 |Gn d | X g∈Gn d g · v = 1 |Gn d | X g∈Gn d χ(g)v = 0 since χ is a non-trivial character.

References

[Blo84] Spencer Bloch. Algebraic cycles and values of L-functions. p. 8

Where the paper says it

In some cases these cycles can be represented by X(Zi , fi1 , fi2 , . fiq ) where Zi are codimensional p cycles on X × Aq and fij are functions on Zi subject to the vanishing of the tame symbol of the functions.

p. 8
Where the paper says it

In some cases these cycles can be represented by X(Zi , fi1 , fi2 , . fiq ) where Zi are codimensional p cycles on X × Aq and fij are functions on Zi subject to the vanishing of the tame symbol of the functions.

J. Reine Angew. Math., 350:94–108, 1984.

[Blo86] Spencer Bloch. p. 8

Where the paper says it

A motivic cycle, or equivalently a higher Chow cycle, in the group CHp (X, q) = H 2p−q M (X, Z(p)) is a codimensional p cycle on ∆qX = X × ∆q , where ∆q is the q-simplex ∆ q = {(t0, . , tq) ∈ A q+1| Xti = 1} subject to certain conditions with respect to the faces [Blo86, Sch93a].

p. 8
Where the paper says it

A motivic cycle, or equivalently a higher Chow cycle, in the group CHp (X, q) = H 2p−q M (X, Z(p)) is a codimensional p cycle on ∆qX = X × ∆q , where ∆q is the q-simplex ∆ q = {(t0, . , tq) ∈ A q+1| Xti = 1} subject to certain conditions with respect to the faces [Blo86, Sch93a].

Algebraic cycles and higher K-theory. p. 0
Where the paper says it

Torsion cycles on Fermat varieties Ramesh Sreekantan Indian Statistical Institute Bangalore March 3, 2026 Abstract A theorem of Manin and Drinfeld [Man72],[Dri73] states that any divisor of degree 0 on the cusps of a modular curve is torsion in the Jacobian. An elegant proof of this result was provided by Elkik [Elk90] using mixed Hodge theory.

p. 0
Where the paper says it

Torsion cycles on Fermat varieties Ramesh Sreekantan Indian Statistical Institute Bangalore March 3, 2026 Abstract A theorem of Manin and Drinfeld [Man72],[Dri73] states that any divisor of degree 0 on the cusps of a modular curve is torsion in the Jacobian. An elegant proof of this result was provided by Elkik [Elk90] using mixed Hodge theory.

Adv. in Math., 61(3):267–304, 1986.

[Blo00] Spencer J. Bloch. Higher regulators, algebraic K-theory, and zeta functions of elliptic curves, volume 11 of CRM Monograph Series. p. 0

Where the paper says it

An elegant proof of this result was provided by Elkik [Elk90] using mixed Hodge theory. Rohrlich [Roh77] proved a generalization of this to Fermat curves.

p. 0
Where the paper says it

An elegant proof of this result was provided by Elkik [Elk90] using mixed Hodge theory. Rohrlich [Roh77] proved a generalization of this to Fermat curves.

American Mathematical Society, Providence, RI, 2000. p. 0
Where the paper says it

We then use the same methods to generalize it to higher codimensional null homologous cycles as well as higher Chow cycles on Fermat varieties. Mathematical Sciences Classification (2020): 14C25, 19E15, 14C30.

p. 0
Where the paper says it

We then use the same methods to generalize it to higher codimensional null homologous cycles as well as higher Chow cycles on Fermat varieties. Mathematical Sciences Classification (2020): 14C25, 19E15, 14C30.

[Car80] James A. p. 2

Where the paper says it

An extension is called a separated extension of MHS if the highest weight of A is less than the lowest weight of B. Carlson [Car80] showed that in this case that group can be identified with a generalized torus Theorem 2.

p. 2
Where the paper says it

An extension is called a separated extension of MHS if the highest weight of A is less than the lowest weight of B. Carlson [Car80] showed that in this case that group can be identified with a generalized torus Theorem 2.

Carlson. Extensions of mixed Hodge structures. p. 0
Where the paper says it

That is, if D = P P ∈Sd aP P such that P P ∈Sd aP = 0 then there exists a function f ∈ K(Fd) such that for some k ∈ N, div(f) = kD There is an elegant ‘pure thought’ proof of the original Manin-Drinfeld theorem due to Deligne and Elkik [Elk90]. The idea is that the Abel-Jacobi map can be described in terms of extensions of mixed Hodge structures.

p. 0
Where the paper says it

That is, if D = P P ∈Sd aP P such that P P ∈Sd aP = 0 then there exists a function f ∈ K(Fd) such that for some k ∈ N, div(f) = kD There is an elegant ‘pure thought’ proof of the original Manin-Drinfeld theorem due to Deligne and Elkik [Elk90]. The idea is that the Abel-Jacobi map can be described in terms of extensions of mixed Hodge structures.

In Journ´ees de G´eometrie Alg´ebrique d’Angers, Juillet 1979/Algebraic Geometry, Angers, 1979, pages 107–127. Sijthoff & Noordhoff, Alphen aan den Rijn—Germantown, Md., 1980.

[Del71] Pierre Deligne. p. 2

Where the paper says it

1 Mixed Hodge structures. According to Deligne [Del71], a mixed Hodge structure (MHS) is an Abelian group HZ with two filtrations.

p. 2
Where the paper says it

1 Mixed Hodge structures. According to Deligne [Del71], a mixed Hodge structure (MHS) is an Abelian group HZ with two filtrations.

Th´eorie de Hodge. p. 0
Where the paper says it

Torsion cycles on Fermat varieties Ramesh Sreekantan Indian Statistical Institute Bangalore March 3, 2026 Abstract A theorem of Manin and Drinfeld [Man72],[Dri73] states that any divisor of degree 0 on the cusps of a modular curve is torsion in the Jacobian. An elegant proof of this result was provided by Elkik [Elk90] using mixed Hodge theory.

p. 0
Where the paper says it

Torsion cycles on Fermat varieties Ramesh Sreekantan Indian Statistical Institute Bangalore March 3, 2026 Abstract A theorem of Manin and Drinfeld [Man72],[Dri73] states that any divisor of degree 0 on the cusps of a modular curve is torsion in the Jacobian. An elegant proof of this result was provided by Elkik [Elk90] using mixed Hodge theory.

I. In Actes du Congr´es International des Math´ematiciens (Nice, 1970), Tome 1, pages 425–430. p. 0
Where the paper says it

Further, if X is a modular curve, S the set of cusps 1 arXiv:2603. 00723v1 [math.

Gauthier-Villars Editeur, ´Paris, 1971.

[Dri73] V. G. Drinfeld. Two theorems on modular curves. p. 0

Where the paper says it

A well known theorem of Manin and Drinfeld [Man72, Dri73] states that any divisor of degree 0 supported on the cusps of a modular curve is torsion in the Jacobian. This was generalized to Fermat curves by Rohrlich [Roh77].

p. 0
Where the paper says it

A well known theorem of Manin and Drinfeld [Man72, Dri73] states that any divisor of degree 0 supported on the cusps of a modular curve is torsion in the Jacobian. This was generalized to Fermat curves by Rohrlich [Roh77].

Funkcional. Anal. i Priloˇzen., 7(2):83–84, 1973.

[Elk90] R. Elkik. Le th´eor´eme de manin-drinfel’d. p. 0

Where the paper says it

Torsion cycles on Fermat varieties Ramesh Sreekantan Indian Statistical Institute Bangalore March 3, 2026 Abstract A theorem of Manin and Drinfeld [Man72],[Dri73] states that any divisor of degree 0 on the cusps of a modular curve is torsion in the Jacobian. An elegant proof of this result was provided by Elkik [Elk90] using mixed Hodge theory.

p. 0
Where the paper says it

Torsion cycles on Fermat varieties Ramesh Sreekantan Indian Statistical Institute Bangalore March 3, 2026 Abstract A theorem of Manin and Drinfeld [Man72],[Dri73] states that any divisor of degree 0 on the cusps of a modular curve is torsion in the Jacobian. An elegant proof of this result was provided by Elkik [Elk90] using mixed Hodge theory.

Number 183, pages 59–67, 1990 S´eminaire sur les Pinceaux de Courbes Elliptiques (Paris, 1988). p. 1
Where the paper says it

As a result, any null-homologous cycle supported on S p d is torsion in the p th-intermediate Jacobian. A conjecture of Beilinson and Bloch asserts that the Abel-Jacobi map is injective for varieties defined over number fields, hence this would imply that the cycle is actually torsion in the Chow group itself.

p. 1
Where the paper says it

As a result, any null-homologous cycle supported on S p d is torsion in the p th-intermediate Jacobian. A conjecture of Beilinson and Bloch asserts that the Abel-Jacobi map is injective for varieties defined over number fields, hence this would imply that the cycle is actually torsion in the Chow group itself.

[Esk19] Payman Eskandari. Algebraic cycles and the mixed Hodge structure on the fundamental group of a punctured curve. p. 0

Where the paper says it

Torsion cycles on Fermat varieties Ramesh Sreekantan Indian Statistical Institute Bangalore March 3, 2026 Abstract A theorem of Manin and Drinfeld [Man72],[Dri73] states that any divisor of degree 0 on the cusps of a modular curve is torsion in the Jacobian. An elegant proof of this result was provided by Elkik [Elk90] using mixed Hodge theory.

p. 0
Where the paper says it

Torsion cycles on Fermat varieties Ramesh Sreekantan Indian Statistical Institute Bangalore March 3, 2026 Abstract A theorem of Manin and Drinfeld [Man72],[Dri73] states that any divisor of degree 0 on the cusps of a modular curve is torsion in the Jacobian. An elegant proof of this result was provided by Elkik [Elk90] using mixed Hodge theory.

Math. Ann., 375(3-4):1665–1719, 2019.

[GS95] B. H. Gross and C. Schoen. The modified diagonal cycle on the triple product of a pointed curve. p. 5

Where the paper says it

The group µ n+2 d acts on F n d by (ζ0, . , ζn+1) · [X0 : · · · : Xn+1] = [ζ0X0 : · · · : ζn+1Xn+1] The diagonal acts trivially.

p. 5
Where the paper says it

The group µ n+2 d acts on F n d by (ζ0, . , ζn+1) · [X0 : · · · : Xn+1] = [ζ0X0 : · · · : ζn+1Xn+1] The diagonal acts trivially.

Ann. Inst. Fourier (Grenoble), 45(3):649–679, 1995.

[Jan90] Uwe Jannsen. Mixed motives and algebraic K-theory, volume 1400 of Lecture Notes in Mathematics. p. 0

Where the paper says it

Torsion cycles on Fermat varieties Ramesh Sreekantan Indian Statistical Institute Bangalore March 3, 2026 Abstract A theorem of Manin and Drinfeld [Man72],[Dri73] states that any divisor of degree 0 on the cusps of a modular curve is torsion in the Jacobian. An elegant proof of this result was provided by Elkik [Elk90] using mixed Hodge theory.

p. 0
Where the paper says it

Torsion cycles on Fermat varieties Ramesh Sreekantan Indian Statistical Institute Bangalore March 3, 2026 Abstract A theorem of Manin and Drinfeld [Man72],[Dri73] states that any divisor of degree 0 on the cusps of a modular curve is torsion in the Jacobian. An elegant proof of this result was provided by Elkik [Elk90] using mixed Hodge theory.

Springer-Verlag, Berlin, 1990 With appendices by S. Bloch and C. p. 1
Where the paper says it

As a result, any null-homologous cycle supported on S p d is torsion in the p th-intermediate Jacobian. A conjecture of Beilinson and Bloch asserts that the Abel-Jacobi map is injective for varieties defined over number fields, hence this would imply that the cycle is actually torsion in the Chow group itself.

p. 1
Where the paper says it

As a result, any null-homologous cycle supported on S p d is torsion in the p th-intermediate Jacobian. A conjecture of Beilinson and Bloch asserts that the Abel-Jacobi map is injective for varieties defined over number fields, hence this would imply that the cycle is actually torsion in the Chow group itself.

Schoen.

[Man72] Ju I. p. 0

Where the paper says it

Torsion cycles on Fermat varieties Ramesh Sreekantan Indian Statistical Institute Bangalore March 3, 2026 Abstract A theorem of Manin and Drinfeld [Man72],[Dri73] states that any divisor of degree 0 on the cusps of a modular curve is torsion in the Jacobian. An elegant proof of this result was provided by Elkik [Elk90] using mixed Hodge theory.

p. 0
Where the paper says it

Torsion cycles on Fermat varieties Ramesh Sreekantan Indian Statistical Institute Bangalore March 3, 2026 Abstract A theorem of Manin and Drinfeld [Man72],[Dri73] states that any divisor of degree 0 on the cusps of a modular curve is torsion in the Jacobian. An elegant proof of this result was provided by Elkik [Elk90] using mixed Hodge theory.

Manin. Parabolic points and zeta functions of modular curves. p. 0
Where the paper says it

A well known theorem of Manin and Drinfeld [Man72, Dri73] states that any divisor of degree 0 supported on the cusps of a modular curve is torsion in the Jacobian. This was generalized to Fermat curves by Rohrlich [Roh77].

p. 0
Where the paper says it

A well known theorem of Manin and Drinfeld [Man72, Dri73] states that any divisor of degree 0 supported on the cusps of a modular curve is torsion in the Jacobian. This was generalized to Fermat curves by Rohrlich [Roh77].

Izv. Akad. Nauk SSSR Ser. Mat., 36:19–66, 1972.

[Man91] Yu I. Manin. Three-dimensional hyperbolic geometry as-adic arakelov geometry. p. 5

Where the paper says it

For a character χ = χ(a0, a1, . , an+1), let V (χ) denote corresponding 1-dimensional representation.

p. 5
Where the paper says it

For a character χ = χ(a0, a1, . , an+1), let V (χ) denote corresponding 1-dimensional representation.

Inventiones mathematicae, 104(2):223–244, 1991.

[Roh77] David E. p. 0

Where the paper says it

An elegant proof of this result was provided by Elkik [Elk90] using mixed Hodge theory. Rohrlich [Roh77] proved a generalization of this to Fermat curves.

p. 0
Where the paper says it

An elegant proof of this result was provided by Elkik [Elk90] using mixed Hodge theory. Rohrlich [Roh77] proved a generalization of this to Fermat curves.

Rohrlich. Points at infinity on the Fermat curves. p. 0
Where the paper says it

An elegant proof of this result was provided by Elkik [Elk90] using mixed Hodge theory. Rohrlich [Roh77] proved a generalization of this to Fermat curves.

p. 0
Where the paper says it

An elegant proof of this result was provided by Elkik [Elk90] using mixed Hodge theory. Rohrlich [Roh77] proved a generalization of this to Fermat curves.

Invent. Math., 39(2):95–127, 1977.

[Ros94] Raymond Ross. p. 9

Where the paper says it

Then (Ca, fbc) + (Cb, fca) + (Cc, fab) is an element of CHp (F n d , 1) as divCa (fbc) + divCb (fca) + divCc (fab) Cab − Cac + Cbc − Cba + Cca − Ccb = 0 More generally one can construct cycles in CHp (F n d , q) along the same lines. Ross [Ros94] constructed cycles in the group CH2 (F 1 d , 2) using pairs of functions supported on the Fermat points.

p. 9
Where the paper says it

Then (Ca, fbc) + (Cb, fca) + (Cc, fab) is an element of CHp (F n d , 1) as divCa (fbc) + divCb (fca) + divCc (fab) Cab − Cac + Cbc − Cba + Cca − Ccb = 0 More generally one can construct cycles in CHp (F n d , q) along the same lines. Ross [Ros94] constructed cycles in the group CH2 (F 1 d , 2) using pairs of functions supported on the Fermat points.

K2 of Fermat curves with divisorial support at infinity. p. 1
Where the paper says it

Let y be a higher Chow cycle in the group CHp (F n d , q)Q = H2p−q (F n d , Q(p)), y = X(ZI , fIi , . , fIq ) where ZI are linear subvarieties of codimension p−q and fIj are functions on ZI with divisorial support on linear subvarieties of ZI .

p. 1
Where the paper says it

Let y be a higher Chow cycle in the group CHp (F n d , q)Q = H2p−q (F n d , Q(p)), y = X(ZI , fIi , . , fIq ) where ZI are linear subvarieties of codimension p−q and fIj are functions on ZI with divisorial support on linear subvarieties of ZI .

Compositio Math. p. 0
Where the paper says it

Further, if X is a modular curve, S the set of cusps 1 arXiv:2603. 00723v1 [math.

p. 0
Where the paper says it

Further, if X is a modular curve, S the set of cusps 1 arXiv:2603. 00723v1 [math.

, 91(3):223–240, 1994.

[Sar24] Subham Sarkar. Cycles on Jacobians of Fermat curves and hypergeometric functions. p. 9

Where the paper says it

Then (Ca, fbc) + (Cb, fca) + (Cc, fab) is an element of CHp (F n d , 1) as divCa (fbc) + divCb (fca) + divCc (fab) Cab − Cac + Cbc − Cba + Cca − Ccb = 0 More generally one can construct cycles in CHp (F n d , q) along the same lines. Ross [Ros94] constructed cycles in the group CH2 (F 1 d , 2) using pairs of functions supported on the Fermat points.

p. 9
Where the paper says it

Then (Ca, fbc) + (Cb, fca) + (Cc, fab) is an element of CHp (F n d , 1) as divCa (fbc) + divCb (fca) + divCc (fab) Cab − Cac + Cbc − Cba + Cca − Ccb = 0 More generally one can construct cycles in CHp (F n d , q) along the same lines. Ross [Ros94] constructed cycles in the group CH2 (F 1 d , 2) using pairs of functions supported on the Fermat points.

J. Algebra, 638:535–566, 2024.

[Sch93a] Chad Schoen. p. 8

Where the paper says it

A motivic cycle, or equivalently a higher Chow cycle, in the group CHp (X, q) = H 2p−q M (X, Z(p)) is a codimensional p cycle on ∆qX = X × ∆q , where ∆q is the q-simplex ∆ q = {(t0, . , tq) ∈ A q+1| Xti = 1} subject to certain conditions with respect to the faces [Blo86, Sch93a].

p. 8
Where the paper says it

A motivic cycle, or equivalently a higher Chow cycle, in the group CHp (X, q) = H 2p−q M (X, Z(p)) is a codimensional p cycle on ∆qX = X × ∆q , where ∆q is the q-simplex ∆ q = {(t0, . , tq) ∈ A q+1| Xti = 1} subject to certain conditions with respect to the faces [Blo86, Sch93a].

Complex multiplication cycles and a conjecture of Be˘ılinson and Bloch. p. 1
Where the paper says it

As a result, any null-homologous cycle supported on S p d is torsion in the p th-intermediate Jacobian. A conjecture of Beilinson and Bloch asserts that the Abel-Jacobi map is injective for varieties defined over number fields, hence this would imply that the cycle is actually torsion in the Chow group itself.

p. 1
Where the paper says it

As a result, any null-homologous cycle supported on S p d is torsion in the p th-intermediate Jacobian. A conjecture of Beilinson and Bloch asserts that the Abel-Jacobi map is injective for varieties defined over number fields, hence this would imply that the cycle is actually torsion in the Chow group itself.

Trans. Amer. Math. Soc., 339(1):87–115, 1993.

[Sch93b] A. p. 9

Where the paper says it

2 Extensions and motivic cycles. Analogously to the situation of null-homologous cycles one can associate an extension to a motivic cycle [Sch93b].

p. 9
Where the paper says it

2 Extensions and motivic cycles. Analogously to the situation of null-homologous cycles one can associate an extension to a motivic cycle [Sch93b].

J. Scholl. Extensions of motives, higher Chow groups and special values of L-functions. p. 9
Where the paper says it

Ross [Ros94] constructed cycles in the group CH2 (F 1 d , 2) using pairs of functions supported on the Fermat points. There are some trivial elements in the higher Chow groups of the form (Z, a1, .

p. 9
Where the paper says it

Ross [Ros94] constructed cycles in the group CH2 (F 1 d , 2) using pairs of functions supported on the Fermat points. There are some trivial elements in the higher Chow groups of the form (Z, a1, .

In S´eminaire de Th´eorie des Nombres, Paris, 1991–92, volume 116 of Progr. Math., pages 279–292 Birkh¨auser Boston, Boston, MA, 1993.

[Shi79] Tetsuji Shioda. p. 5

Where the paper says it

In order to prove the splitting we have to understand the cohomology of F n d . For that we follow Shioda [Shi79], who proved the Hodge conjecture for such varieties under some conditions.

p. 5
Where the paper says it

In order to prove the splitting we have to understand the cohomology of F n d . For that we follow Shioda [Shi79], who proved the Hodge conjecture for such varieties under some conditions.

The Hodge conjecture for Fermat varieties. p. 4
Where the paper says it

A conjecture of Bloch [Blo84] and Beilinson asserts that if the variety is defined over a number field, such a cycle is necessarily torsion in the Chow group as well. 3 Fermat varieties.

p. 4
Where the paper says it

A conjecture of Bloch [Blo84] and Beilinson asserts that if the variety is defined over a number field, such a cycle is necessarily torsion in the Chow group as well. 3 Fermat varieties.

Math. Ann., 245(2):175–184, 1979.

How the training signal is built

The paper uses mixed Hodge structures to describe the Abel-Jacobi map, showing that for a divisor of degree 0, it is torsion in the Jacobian if and only if the sequence of Q-MHS splits. This framework is extended to higher codimensions where one obtains a weaker statement about torsion in the intermediate Jacobian. p. 1

Where the paper says it

In this short paper we show how the same idea can be applied to get a new proof of Rohrlich’s theorem - though not quite as explicit as his is. However, one advantage is that this can be generalized to higher co-dimensions with only a little more work, though we get a weaker statement - only torsion in the intermediate Jacobian.

The structure of cohomology groups, particularly the action of Gn d on Hn(Fn d, C), is analyzed using Shioda's theorem regarding primitive cohomology.

Where the gains come from

The splitting of the exact sequence in Theorem 3.1 is achieved by defining an operator T that acts on H2p−1(Fn d − Spd, Q) and showing it provides a complementary pure Hodge substructure. p. 7

Where the paper says it

The image of T is a pure Hodge substructure of weight 2p of H2p−1 (F n d −S p d , Q) and is isomorphic to H 2p S p d (F n d , Q) and complementary to H2p−1 (F n d , Q). Hence it provides a splitting of the exact sequence.

This operator T is induced by an automorphism of Fnd and acts by multiplication by Gn d = dn+1 on the relevant cohomology group. p. 7
Where the paper says it

T is also a morphism of Hodge structures as it is induced by an automorphism of F n d . Further, since Gn d acts trivially on H 2p S p d (F n d , Q), T acts by multiplication by |Gn d | = d n+1.

In the case n > 1, Gn d acts trivially on H2p Spd(Fnd, Q). p. 7
Where the paper says it

Further, since Gn d acts trivially on H 2p S p d (F n d , Q), T acts by multiplication by |Gn d | = d n+1. The image of T is a pure Hodge substructure of weight 2p of H2p−1 (F n d −S p d , Q) and is isomorphic to H 2p S p d (F n d , Q) and complementary to H2p−1 (F n d , Q).

This trivial action simplifies the analysis of the splitting significantly. p. 3
Where the paper says it

If the exact sequence is split then the same splitting implies that the class of EZ is split. Conversely, H 2p |D| (X, Q) 0 is generated by Q · (Di − Dj )|i ̸= j so if EDi−Dj is trivial for all i, j then the sequence 0 → H 2p−1 (X, Q) → H 2p−1 (Y, Q) → H 2p |D| (X, Q) 0 → 0 is split.

Pruning redundant reasoning tokens

The paper focuses on cycles where p - q ≥ 0 and n > 1 for higher Chow groups, discounting decomposable cycles. The key result is that any motivic cycle supported on the linear hypersurfaces of a Fermat variety has a regulator which is torsion in the Deligne cohomology group. p. 9

Where the paper says it

The term on the right is the Deligne cohomology group which is the analogue of the Jacobian and there is a map called the regulator map from motivic cohomology to Deligne cohomology analogous to the Abel-Jacobi map. A motivic cycle, therefore, determines an extension in the conjectural category of mixed motives as well as determines an extension in the category of mixed Hodge structures and the realization in the category of mixed Hodge structures is the regulator map.

This is the higher Chow version of Rohrlich’s theorem. p. 1
Where the paper says it

That is the case when n = 1, which is Rohrlich’s theorem. We then show that the same methods can be used to obtain a statement about certain higher Chow cycles having torsion regulator (Theorem 4.

The argument for splitting relies on the operator T producing a splitting in the exact sequence. p. 7
Where the paper says it

Hence it provides a splitting of the exact sequence. When n = 1 the argument is similar.

Conclusion

The paper establishes that certain natural constructions do not produce interesting cycles on Fermat varieties, but it does not mean that the Chow groups of Fermat varieties are trivial. The inductive structure on Fermat Varieties suggests the existence of non-torsion null homologous cycles in CHp hom((F1d)2p−1d). p. 7

Where the paper says it

The case when n = 2p − 1 is the only case where one expects potentially interesting cycles. For instance CH1 hom(F 1 d ), CH2 hom(F 3 d ), CH3 hom(F 5 d ) and in fact there do exist interesting non-torsion cycles in these groups as we will see in Section 5.

The work provides a framework connecting motivic cycles to extensions in mixed motives and Deligne cohomology. p. 9
Where the paper says it

The term on the right is the Deligne cohomology group which is the analogue of the Jacobian and there is a map called the regulator map from motivic cohomology to Deligne cohomology analogous to the Abel-Jacobi map. A motivic cycle, therefore, determines an extension in the conjectural category of mixed motives as well as determines an extension in the category of mixed Hodge structures and the realization in the category of mixed Hodge structures is the regulator map.

[Blo84] Spencer Bloch. Algebraic cycles and values of L-functions. J. Reine Angew. Math., 350:94–108, 1984.

[Blo86] Spencer Bloch. Algebraic cycles and higher K-theory. Adv. in Math., 61(3):267–304, 1986.

[Blo00] Spencer J. Bloch. Higher regulators, algebraic K-theory, and zeta functions of elliptic curves, volume 11 of CRM Monograph Series. American Mathematical Society, Providence, RI, 2000.

[Car80] James A. Carlson. Extensions of mixed Hodge structures. In Journ´ees de G´eometrie Alg´ebrique d’Angers, Juillet 1979/Algebraic Geometry, Angers, 1979, pages 107–127. Sijthoff & Noordhoff, Alphen aan den Rijn—Germantown, Md., 1980.

[Del71] Pierre Deligne. Th´eorie de Hodge. I. In Actes du Congr´es International des Math´ematiciens (Nice, 1970), Tome 1, pages 425–430 Gauthier-Villars Editeur, ´Paris, 1971.

[Dri73] V. G. Drinfeld. Two theorems on modular curves. Funkcional. Anal. i Priloˇzen., 7(2):83–84, 1973.

[Elk90] R. Elkik. Le th´eor´eme de manin-drinfel’d. Number 183, pages 59–67, 1990 S´eminaire sur les Pinceaux de Courbes Elliptiques (Paris, 1988).

[Esk19] Payman Eskandari. Algebraic cycles and the mixed Hodge structure on the fundamental group of a punctured curve. Math. Ann., 375(3-4):1665–1719, 2019.

[GS95] B. H. Gross and C. Schoen. The modified diagonal cycle on the triple product of a pointed curve. Ann. Inst. Fourier (Grenoble), 45(3):649–679, 1995.

[Jan90] Uwe Jannsen. Mixed motives and algebraic K-theory, volume 1400 of Lecture Notes in Mathematics. Springer-Verlag, Berlin, 1990 With appendices by S. Bloch and C. Schoen.

[Man72] Ju I. Manin. Parabolic points and zeta functions of modular curves. Izv. Akad. Nauk SSSR Ser. Mat., 36:19–66, 1972.

[Man91] Yu I. Manin. Three-dimensional hyperbolic geometry as-adic arakelov geometry. Inventiones mathematicae, 104(2):223–244, 1991.

[Roh77] David E. Rohrlich. Points at infinity on the Fermat curves. Invent. Math., 39(2):95–127, 1977.

[Ros94] Raymond Ross. K2 of Fermat curves with divisorial support at infinity. Compositio Math., 91(3):223–240, 1994.

[Sar24] Subham Sarkar. Cycles on Jacobians of Fermat curves and hypergeometric functions. J. Algebra, 638:535–566, 2024.

[Sch93a] Chad Schoen. Complex multiplication cycles and a conjecture of Be˘ılinson and Bloch. Trans. Amer. Math. Soc., 339(1):87–115, 1993.

[Sch93b] A. J. Scholl. Extensions of motives, higher Chow groups and special values of L-functions. In S´eminaire de Th´eorie des Nombres, Paris, 1991–92, volume 116 of Progr. Math., pages 279–292 Birkh¨auser Boston, Boston, MA, 1993.

[Shi79] Tetsuji Shioda. The Hodge conjecture for Fermat varieties. Math. Ann., 245(2):175–184, 1979.

Green — the paper says it, and the page is printed. Red — no passage carries it, so treat that sentence as unverified. The same check runs on your own text: check a summary against its source, $5.

Improvements for AI systems

  1. Improved understanding of torsion in intermediate Jacobians for higher codimensional cycles: The system can now determine if a null-homologous cycle supported on linear subvarieties is torsion in the pth intermediate Jacobian by checking if the exact sequence 0 -> H2p−1(Fnd, Q) -> H2p−1(Fnd - Spd, Q) -> H2p Spd(Fnd, Q) 0 -> 0 splits.

  2. Improved calculation of regulator values for motivic cycles: The system can compute the regulator reg(y) in the Deligne cohomology group H2p−q D (Fnd, Q(p)) for higher Chow cycles y of the form X(ZI, fI1,..., fIq), based on whether the exact sequence in Proposition 4.1 splits.

  3. Improved classification of interesting higher Chow groups: The system can identify which higher Chow groups CHp(Fnd, q) are likely to contain non-torsion cycles by focusing on cases where n = 2p−q−1, such as CH2(F1d, 2) or CH3(F4d, 1).

  4. Improved cycle detection for Fermat varieties: The system can detect the existence of non-torsion null-homologous cycles by analyzing the image of Gross and Schoen modified diagonal cycles under the inductive structure map to get a null-homologous cycle in CH2 hom(F3d).

Abstract

A theorem of Manin and Drinfeld states that any divisor of degree 0 on the cusps of a modular curve is torsion in the Jacobian. An elegant proof of this result was provided by Elkik using mixed Hodge theory. Rohrlich proved a generalization of this to Fermat curves. In this paper we reprove his results along the lines of the work of Elkik. We then use the same methods to generalize it to higher Chow cycles. We introduce a notion of `ambient decomposability' - essentially cycles coming from the ambient P n+1 and the base field - and show that higher Chow cycles determined by coordinate sections are necessarily of this type. This is a revised and accepted version of the earlier manuscript of the same name and is to appear in Indagationes Mathematicae. The results are stronger than the results in the earlier manuscript.

Related papers