REVIEW 1 major objections 3 minor 5 references
A conjecture on the tensor ideal for an elementary p-group generated by the restriction of a Steinberg module
T0 review · 1 major / 3 minor · reviewed 2026-08-06 · deepseek-v4-flash
Pith's one-line read The paper proves that tensoring the Steinberg module of an elementary abelian p-group with a broad class of representations yields restricted tilting modules, supporting a conjecture that would constrain the classification of tensor…
desk verdict Substantial evidence for the Steinberg tensor-ideal conjecture, but the r=2 classification of cyclic S-projectives has a real gap that undermines Theorem 3. read the letter →
The pith
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
The reading
What carries the argument
The load-bearing object is the restricted Steinberg module $S=T_{p^{r-1}-1}$, the simple $\mathrm{SL}_2$-module of highest weight $p^{r-1}-1$ regarded as an $E$-module through the fixed embedding $E < U(k) < \mathrm{SL}_2(k)$. The machinery converts 'is $S\otimes X$ tilting?' into support and extension data: the support variety of $S$ is a single point, so $S\otimes X$ is projective whenever $X$ avoids that point; when the point is present, the relevant extension group $\mathrm{Hom}_E(\Omega S,S)$ is identified with the endomorphism algebra of $S$ (Corollary 5.2.12), and the standard tensor-product formula $T_{i+pj}\simeq T_i\otimes T_j^{(1)}$ turns the resulting self-extensions into explicit tilting modules $T_{d p^{r-1}+p^i-1}$. Carlson modules $L_\zeta$ (defined by a short exact sequence $0\to L_\zeta\to \Omega^d 1\to 1\to 0$ attached to a cohomology class $\zeta$) are handled by a periodicity argument: because $S$ is periodic, the tensor product $S\otimes L_\zeta$ is determined by the image of $\zeta$ under an algebra map from the cohomology of $E$ into endomorphisms of $S$ (Lemma 6.2.4). A Gaussian-elimination lemma for extension matrices (Lemma 7.1.1) splits arbitrary allowed extensions into these building blocks.
What would settle it
For a fixed embedding of $E=C_p^r$ into $\mathrm{SL}_2$, take a module $X$ in the closure of the listed classes whose extension class lies outside the image of the map $\omega_{d,e}$ described in Proposition 5.2.8, and compute the indecomposable summands of $S\otimes X$; if any summand has dimension not equal to $\dim T_j$ for some $0\le j<p^r$, then $S\otimes X$ is not a restricted tilting module and Theorem 2 fails.
Extended reading notes
Core claim
The central claim is that a conjecture from the authors' previous paper reduces to a statement about one thick tensor ideal: inside the category $\mathcal T$ of restrictions of $\mathrm{SL}_2$-tilting modules to $E$, the ideal $\mathcal T_r = \langle T_j \mid j \ge p^{r-1}-1\rangle$ is generated by the Steinberg module $S = T_{p^{r-1}-1}$, and conjecturally this ideal is thick in the full category $\mathrm{Rep}\,E$. The paper proves the conjecture for the subcategory described in Theorem 2 and, when $r=2$, additionally proves in Theorem 8.2.1 that every cyclic direct summand of $S\otimes X$ is restricted tilting. The proof uses support varieties, showing that the support of $S$ is a single point; an explicit computation of the image of $S\otimes -$ on extension groups between uniserial modules (Proposition 5.2.8 and Corollary 5.2.12); and the standard tensor-product formula for $\mathrm{SL}_2$-tilting modules, which identifies the resulting extensions as tilting modules $T_{d p^{r-1}+p^i-1}$. For $r=2$, cyclic $S$-projective modules are classified as $T_{\ell p-1}$ for $1\le \ell \le p$.
Load-bearing premise
The equivalence between the conjecture and the statement that $S\otimes X$ lies in the ideal $\mathcal T_r$ depends on the imported classification of thick tensor ideals in restricted $\mathrm{SL}_2$-tilting modules and on the structure of that category from the authors' previous work; if those structural facts are wrong, proving that $S\otimes X$ is tilting would no longer be equivalent to proving the conjecture.
Editorial extensions
If this is right
- For every $X$ in the subcategory described in Theorem 2, $S\otimes X$ is a restricted tilting module, so its indecomposable summands are among $T_0,\dots,T_{p^r-1}$.
- Conjecture A holds for $p=r=2$ without appealing to the classification of indecomposable modules, giving a self-contained proof for $E=C_2\times C_2$.
- For $r=2$, every cyclic $S$-projective module lies in $\mathcal T_r$, so the cyclic case of Conjecture D is verified.
- Any module whose support avoids the distinguished point $\lambda$ automatically satisfies $S\otimes X\in \mathcal T$, by Remark 4.2.8.
- If the full conjecture holds, every $S$-projective module is self-dual, admits a graded lift, has a filtration with subquotients $S$, is the restriction of an $\mathrm{SL}_2$-module, and satisfies the dimension bounds $t=2^\ell$, $d=\ell'2^\ell p^{r-1}$ from Lemma 3.1.9.
Reading between the lines
- A natural next test is to isolate modules whose support is exactly the distinguished point $\lambda$: Theorem 2 already handles modules avoiding $\lambda$ and several classes supported on it, so the remaining obstruction to the full conjecture is concentrated in a one-point support class that is still computationally wild in general.
- The splitting mechanism suggests a quantitative bound: for the allowed classes, the number of indecomposable summands of $S\otimes X$ should be controlled by the rank of an extension matrix; for small $p$ and $r$ one could test computationally whether the bound depends only on $p$ and $r$ rather than on $\dim X$.
- If the conjecture is correct, $\mathrm{Rep}\,E$ becomes a positive-characteristic tensor category whose stable category has only one non-projective thick tensor ideal above projectives; that rigidity is precisely what classification programmes for incompressible symmetric tensor categories need.
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The paper studies Conjecture 1 of Coulembier–Flake [CF24], which predicts that for an elementary abelian p-group E and S the restriction of the largest non-projective Steinberg SL2-module, the tensor product S⊗X is a restricted SL2-tilting module for every finite-dimensional E-representation X. The authors establish Theorem 2, verifying the conjecture for a substantial subcategory of Rep E: modules of Loewy length at most two, Carlson modules, extensions of quotients of certain uniserial p-dimensional modules, modules whose support avoids a specified point, restrictions of simple/standard/costandard SL2-modules, and Frobenius twists of restricted tilting modules. They also prove Theorem 3: for r=2 every cyclic direct summand of S⊗X is restricted tilting. The paper contains several reformulations of the conjecture (Conjectures A–G) and gives examples showing that natural nearby statements fail, illustrating the sharpness of the conjecture.
Significance. The conjecture, if true, would identify a surprisingly small thick tensor ideal in Rep E and has implications for the classification of incompressible symmetric tensor categories in positive characteristic. The paper does not prove the conjecture, but it supplies the most extensive evidence to date and reduces the question to a tractable-looking statement about S-projective modules. The main positive results are stated with precise hypotheses and are supported by detailed arguments; the paper is not machine-checked, but many proofs are explicit. The counterexamples in Sections 3.4–3.5 are a useful sanity check on the conjecture's boundaries.
major comments (1)
- [Proposition 8.2.4, first paragraph] The assertion that a cyclic S-projective module X with a proper quotient T_{ℓp−1} also has ∆_{ℓp} as a quotient is not justified by the cited Lemma 8.2.2(b). Lemma 8.2.2(b) identifies ∆_{ℓp} as the unique non-split extension 0→1→∆_{ℓp}→T_{ℓp−1}→0; it does not imply that a given surjection f:X→T_{ℓp−1} lifts to ∆_{ℓp}. Such a lift is equivalent to the vanishing of f^*(cl) in Ext^1_E(X,1), and cyclicity alone does not force this vanishing. Since this is the base step of the induction that produces all quotients of X, the classification in Proposition 8.2.4 (and hence Theorem 8.2.1 / Theorem 3) rests on an unproved lifting statement. Please supply an argument using S-projectivity or modify the induction.
minor comments (3)
- [Section 3.4, Proposition 3.4.2] The proof is described as 'straightforward but tedious computations'; since this proposition is used in Corollary 3.4.5 and Remark 3.4.7, the computations should be supplied in an appendix or a supplementary file.
- [Section 3.1, Lemma 3.1.2 and [Cou18]] The equivalence of Conjectures A–G depends on Lemma 3.1.2 from [CF24] and on the classification of thick tensor ideals in Tilt SL2 from [Cou18]; please state the imported statements explicitly and clarify the status of [CF24], which is cited as an arXiv preprint.
- [Throughout] There are several typographical errors, including 'cateogry' in Section 2.1, 'LemmLemma' in the proof of Lemma 4.1.5, and 'of of' in Lemma 3.1.9(5); Remark 3.1.4(3) also uses the notation TC_q without definition, and this should be clarified.
Circularity Check
No significant circularity: the main theorems are independent verifications of the conjecture, and the imported prior classifications are external structural results rather than the conjecture itself.
full rationale
I examined the derivation chain. The paper's target, Conjecture 1, is imported from the authors' earlier paper [CF24], but it is used as a conjecture to be verified, not as an assumption of the proofs. The main new results (Theorem 2, Theorem 3, and their corollaries) prove that S tensor X is restricted tilting for specified classes of X by direct module-theoretic arguments. The reformulations in Section 3 rely on Lemma 3.1.2 from [CF24] and on the classification of thick tensor ideals in Tilt SL2 from [Cou18]; these are prior parameter-free structural theorems whose assumptions do not include the target conjecture, so they constitute independent imported support rather than circularity. Lemma 3.1.5 is an equivalence between reformulations, not a reduction of the conjecture to itself. The contrast examples in Sections 3.4 and 3.5 are independent computations showing that lookalike modules do not satisfy the analogous property. No fitted parameter is renamed as a prediction, and no defining equation is equivalent to the claimed conclusion by construction. The flagged assertion in Proposition 8.2.4, concerning the lift from a quotient T_{ell p - 1} to a quotient Delta_{ell p}, is a possible proof gap or missing justification, but it is not a circular dependence: it does not assume the conclusion it is meant to establish. Accordingly, the paper exhibits no significant circularity.
Assumptions & free parameters
assumptions (4)
- domain assumption Support variety theory for modules of elementary abelian p-groups, as summarized in [Ben17]; in particular equations (2.1)-(2.4) and the characterization of projective and periodic modules by support.
- standard math Donkin's tensor product formula (2.5) and standard facts about SL2 tilting modules, for example from [Jan03] and [STWZ23].
- domain assumption Classification of thick tensor ideals in Tilt SL2 over an algebraically closed field [Cou18], used to assert T_r is generated by S = T_{p^{r-1}-1}.
- domain assumption The pseudo-tensor subcategory structure of T and Lemma 3.1.2 from [CF24], which identifies the indecomposable modules in T as restricted tilting modules T_i for i < p^r.
Cite this review
Pith. "Pith review of A conjecture on the tensor ideal for an elementary p-group generated by the restriction of a Steinberg module." pith.science (2026). https://pith.science/paper/CADI2ZDM
@misc{pith2026250702722,
author = {Pith},
title = {Pith review of: A conjecture on the tensor ideal for an elementary p-group generated by the restriction of a Steinberg module},
year = {2026},
howpublished = {\url{https://pith.science/paper/CADI2ZDM}},
note = {Machine review of arXiv:2507.02722}
}
abstract
In previous work (Coulembier--Flake 2024), the authors conjectured that the tensor product of an arbitrary finite-dimensional modular representation of an elementary abelian $p$-group with the biggest non-projective restricted Steinberg $SL_2$-module is a restricted tilting module. We showed that the validity of the conjecture would have interesting implications in the theory of tensor categories in positive characteristic, in particular, with respect to the classification of incompressible symmetric tensor categories, which is the subject of arguably the main open conjecture in the area. We present here some evidence for the conjecture to hold.
Reference graph
Works this paper leans on
-
[1]
[BCR97] D. J. Benson, J. F. Carlson, and J. Rickard, Thick subcategories of the stable module category , Fund. Math. 153 (1997), no. 1, 59–80. [BE19] D. Benson and P. Etingof, Symmetric tensor categories in characteristic 2 , Adv. Math. 351 (2019), 967–
work page 1997
-
[561]
Coulembier, Tensor ideals, Deligne categories and invariant theory , Selecta Math
[Cou18] K. Coulembier, Tensor ideals, Deligne categories and invariant theory , Selecta Math. (N.S.) 24 (2018), no. 5, 4659–4710. [Cou21] , Monoidal abelian envelopes , Compos. Math. 157 (2021), no. 7, 1584–1609. [Cou24] , Inductive systems of the symmetric group, polynomial functors and tensor categories (2024), available at arXiv:2406.00892. [CPW98] J. ...
arXiv 2018
- [2003]
-
[2006]
New trends and methods, Second printing of the 1984 original. [Ben17] , Representations of elementary abelian p-groups and vector bundles , Cambridge Tracts in Math- ematics, vol. 208, Cambridge University Press, Cambridge,
work page 1984
-
[2017]
[BEO23] D. Benson, P. Etingof, and V. Ostrik, New incompressible symmetric tensor categories in positive char- acteristic, Duke Math. J. 172 (2023), no. 1, 105–200. [BIK11] D. J. Benson, S. B. Iyengar, and H. Krause, Stratifying modular representations of finite groups , Ann. of Math. (2) 174 (2011), no. 3, 1643–1684. [BIKP18] D. Benson, S. B. Iyengar, H....
arXiv 2023
Reviewed August 6, 2026 · model on record in the stance chip above.
Discussion (0). Continue with ORCID to comment.