A universal Euler system for GSp(4)

David Loeffler; Sarah Livia Zerbes

arxiv: 2411.12576 · v2 · submitted 2024-11-19 · 🧮 math.NT

A universal Euler system for GSp(4)

David Loeffler , Sarah Livia Zerbes This is my paper

Pith reviewed 2026-05-23 08:26 UTC · model grok-4.3

classification 🧮 math.NT

keywords Euler systemsGSp(4)Galois representationsautomorphic formsmultiplicity onezeta integralsspin representations

0 comments

The pith

Different choices of local test data for GSp(4) Euler systems produce classes that are explicit multiples of one universal class.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

This paper refines an earlier construction of Euler systems for the four-dimensional spin Galois representations attached to automorphic forms on GSp(4). The earlier classes depended on arbitrary local test data. The authors apply multiplicity-one theorems to show that varying the test data only multiplies the class by a factor computed from local zeta integrals. All such classes therefore span the same one-dimensional space, yielding a canonical universal Euler system independent of those choices. This matters because it removes an artificial dependence from the construction and makes the system more suitable for arithmetic applications.

Core claim

In our earlier work, we constructed Euler systems for the spin Galois representations attached to automorphic forms on GSp(4). These classes depended on choices of local test data. Using multiplicity-one results, we show that all such classes lie in a one-dimensional space and are explicit multiples, given by local zeta-integrals, of a universal class that does not depend on the choice of test data.

What carries the argument

The universal Euler system class, obtained by showing via multiplicity-one results that classes for different test data are scalar multiples given by local zeta integrals.

If this is right

The Euler system is now independent of arbitrary local test data choices.
Normalization is given explicitly by factors from local zeta integrals.
The classes span a single one-dimensional space for use in arithmetic applications.
Comparisons with other invariants no longer require adjustments for specific test data.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

The universal class provides a fixed reference point for relating this Euler system to other constructions without local normalizations.
The same multiplicity technique could remove test-data dependence in Euler systems for additional groups.
Applications that rely on explicit classes become simpler once the universal representative is fixed.

Load-bearing premise

Multiplicity-one results for smooth representations of the relevant groups hold and can be applied directly to the Euler system classes constructed in the earlier paper.

What would settle it

A calculation showing that Euler system classes arising from two different choices of local test data are linearly independent over the coefficient field.

read the original abstract

In our earlier work with Christopher Skinner (J. Eur. Math. Soc 24 (2022), no. 2; DOI 10.4171/JEMS/1124; Arxiv 1706.00201), we constructed Euler systems for the 4-dimensional spin Galois representations corresponding to automorphic forms for GSp(4). This construction depended on various arbitrary choices of local test data. In this paper, we use multiplicity-one results for smooth representations to determine how these Euler system classes depend on the choice of test data, showing that all of these classes lie in a 1-dimensional space and are explicit multiples (given by local zeta-integrals) of a "universal" class independent of the choice of test data.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

This paper shows the Euler system classes from the 2022 work all lie in a 1-dimensional space and are explicit multiples of one universal class via local zeta integrals.

read the letter

The core result is that the authors have fixed the dependence on local test data from their earlier GSp(4) Euler system construction. Multiplicity-one theorems for the relevant smooth representations imply that every choice of test data produces a class that is a scalar multiple of a single fixed universal class, and the scalar is given by the ratio of the corresponding local zeta integrals. This removes an arbitrary choice that was left open in the 2022 paper with Skinner. The explicit identification is the new piece; the rest of the construction carries over unchanged. It is a straightforward cleanup that makes the Euler system more canonical and easier to apply downstream. The argument rests on the cited multiplicity-one results applying directly to the images of the earlier classes. The abstract states that this works, but the stress-test point about genericity and irreducibility of those specific classes is worth checking in the body: if the classes land outside the hypotheses of the multiplicity-one theorems, the dimension bound and the exact scalar formula would not follow. No other gaps are visible from the given material. This is a technical note for people already using or extending Euler systems for GSp(4) or similar groups. It is not a large conceptual advance, but the refinement is clean enough that the paper deserves referee time.

Referee Report

2 major / 2 minor

Summary. The paper extends the authors' 2022 JEMS construction of Euler systems for 4-dimensional spin Galois representations attached to GSp(4) automorphic forms. It applies multiplicity-one theorems for smooth representations of the relevant groups to show that the Euler system classes, which a priori depend on arbitrary choices of local test data, all lie in a 1-dimensional space and are explicit scalar multiples (given by ratios of local zeta integrals) of a single 'universal' class independent of those choices.

Significance. If the central claim holds, the result removes the dependence on auxiliary local data from the earlier Euler-system construction, yielding a more canonical object. This strengthens the potential utility of the Euler system for arithmetic applications such as Iwasawa theory or p-adic L-functions for GSp(4). The paper explicitly credits the 2022 construction and external multiplicity-one results rather than re-deriving them.

major comments (2)

[Abstract / §2] Abstract and the discussion of the embedding into smooth representations (likely §2 or §3): the argument requires that the images of the 2022 Euler-system classes under the local embedding satisfy the hypotheses of the cited multiplicity-one theorems (e.g., lying in an irreducible admissible representation and being Whittaker-generic). The manuscript does not appear to contain an explicit verification that these classes remain nonzero after localization or avoid reducible submodules; without this check the dimension bound and the explicit scalar identification via zeta integrals are not guaranteed.
[Introduction / comparison with 2022 construction] The dependence on the 2022 JEMS paper: while the multiplicity-one theorems are external, the manuscript must confirm that the specific test vectors used in the 2022 construction produce classes that are nonzero in the relevant local representation spaces; otherwise the 'universal' class could be zero for some choices of data, undermining the claim that all classes are multiples of a single nonzero object.

minor comments (2)

Notation for the local zeta integrals and the universal class should be introduced with a clear reference to the 2022 paper's notation to avoid confusion for readers.
[Abstract] The abstract mentions 'smooth representations of the relevant groups' but does not name the precise groups or the precise multiplicity-one theorems invoked; adding these citations in the introduction would improve readability.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the careful reading and for highlighting these points on the applicability of the multiplicity-one results. We agree that explicit checks are needed and will revise the manuscript accordingly.

read point-by-point responses

Referee: [Abstract / §2] Abstract and the discussion of the embedding into smooth representations (likely §2 or §3): the argument requires that the images of the 2022 Euler-system classes under the local embedding satisfy the hypotheses of the cited multiplicity-one theorems (e.g., lying in an irreducible admissible representation and being Whittaker-generic). The manuscript does not appear to contain an explicit verification that these classes remain nonzero after localization or avoid reducible submodules; without this check the dimension bound and the explicit scalar identification via zeta integrals are not guaranteed.

Authors: We agree that the manuscript should contain an explicit verification that the localized images of the 2022 classes satisfy the hypotheses (irreducible admissible, Whittaker-generic, nonzero). In the revision we will add a short paragraph in §2 that recalls the relevant non-vanishing statements from the 2022 JEMS paper (non-vanishing of the classes follows from the non-vanishing of the associated local zeta integrals at the chosen test vectors) and confirms that the images avoid reducible submodules by the genericity properties already established there. revision: yes
Referee: [Introduction / comparison with 2022 construction] The dependence on the 2022 JEMS paper: while the multiplicity-one theorems are external, the manuscript must confirm that the specific test vectors used in the 2022 construction produce classes that are nonzero in the relevant local representation spaces; otherwise the 'universal' class could be zero for some choices of data, undermining the claim that all classes are multiples of a single nonzero object.

Authors: We agree that non-vanishing for the concrete test vectors of the 2022 construction must be stated explicitly. The revision will add a sentence in the introduction (and a cross-reference in §2) noting that the 2022 construction produces nonzero classes precisely because the local integrals used to define the Euler system classes are nonzero at those vectors; this guarantees the universal class is nonzero. revision: yes

Circularity Check

0 steps flagged

External multiplicity-one theorems determine dependence on test data; prior construction cited but not load-bearing for new claim

full rationale

The paper's central step applies cited multiplicity-one results for smooth representations of GSp(4) and related groups to the Euler system classes from the 2022 JEMS paper (with Skinner). This determines that the classes span a 1-dimensional space and are related by explicit local zeta-integral scalars to a universal class. The multiplicity-one theorems are presented as external results (not derived here or in the authors' prior work), and no equation or definition in the abstract reduces the new statement to a tautology or self-fit. The citation to the 2022 construction is normal and does not make the dependence claim circular by construction.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The argument relies on multiplicity-one theorems from the representation theory of p-adic groups; these are treated as standard background rather than proved here. No free parameters or new entities are introduced in the abstract.

axioms (1)

domain assumption Multiplicity-one results hold for the smooth representations appearing in the local test data choices
Invoked to conclude that the Euler system classes span a 1-dimensional space.

pith-pipeline@v0.9.0 · 5648 in / 1325 out tokens · 26804 ms · 2026-05-23T08:26:56.075400+00:00 · methodology

A universal Euler system for GSp(4)

Core claim

What carries the argument

If this is right

Where Pith is reading between the lines

Load-bearing premise

What would settle it

discussion (0)