REVIEW 3 major objections 5 minor
From algebraic orthogonality to RAAG embedding obstructions in hierarchically hyperbolic groups
T0 review · 3 major / 5 minor · reviewed 2026-08-11 · deepseek-v4-flash
Pith's one-line read Every RAAG embedding into a hierarchically hyperbolic group in the paper's class $\Xi$ factors, after passing to powers, through a quasi-isometrically embedded RAAG generated by axial elements; RAAG embeddability then reduces to the…
desk verdict Genuine HHG extension of Kim–Koberda with detailed proofs and honest self-corrections, but the main theorems lean on an unproved black-box theorem from the authors' own earlier preprint. 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 expanded core graph $\widehat G_{\mathcal S}$ has as its vertices the axial directions carried by minimal unbounded domains: one vertex for each such domain when its hyperbolic space is a quasi-line, and one vertex per inequivalent axial direction otherwise; two vertices are adjacent when their domains are orthogonal. The mechanism that makes the graph relevant is the orthogonal decomposition property: high powers of any infinite-order element split into pairwise commuting axial factors, each fully supported on one active domain. Passing to a finite geometrically irredundant subcollection and applying the quoted undistorted-RAAG theorem produces the intermediate RAAG; the identification of axial directions with vertices of $\widehat G_{\mathcal S}$ is what embeds its extension graph into $\widehat G_{\mathcal S}$.
What would settle it
Test the engine directly: find a hierarchically hyperbolic structure in class $\Xi$ and a finite graph $\Lambda$ such that $A(\Lambda)$ embeds into $G$ but $\Lambda$ does not embed as an induced subgraph of the clique graph of the expanded core graph; that refutes the main obstruction. More decisively, look for a geometrically irredundant family of strongly fully supported axial elements on pairwise orthogonal domains whose sufficiently large powers fail to generate the prescribed RAAG, which would contradict the quoted undistorted-RAAG theorem the entire factorization relies on.
Extended reading notes
Core claim
The central claim is structural. For an HHG structure $(G,\mathcal S)$ in the class $\Xi$—or, more generally, for a group virtually admitting such a structure—every injective homomorphism $\phi: A(\Lambda)\to G$ can be modified by replacing each standard generator $v$ with $\phi(v)^N$ so that the resulting embedding lands in a subgroup $M\cong A(\Gamma)$ generated by strongly fully supported axial elements; in the $\Xi$ case $M$ is undistorted in $G$. The defining graph $\Gamma$ records orthogonality among the supporting domains of a geometrically irredundant collection of axial directions. The paper then proves that the extension graph $\Gamma^e$ embeds as an induced subgraph of the expanded core graph $\widehat G_{\mathcal S}$, so the classical clique-graph obstruction for RAAGs transfers verbatim: $\Lambda \le (\widehat G_{\mathcal S})^k$. When the structure has rank at most two, the defining graph of every intermediate RAAG is triangle-free, the stronger extension-graph criterion applies, and $A(\Lambda)\le G$ is equivalent to $\Lambda\le \widehat G_{\mathcal S}$. In the standard mapping class group structure, the expanded core graph is the disjointness graph of essential curves; in rich-family structures on a RAAG, it embeds into the extension graph and coincides with it when the rich family contains all singletons.
Load-bearing premise
The whole argument leans on a previously proven theorem, quoted without reproof, that sufficiently large powers of a geometrically irredundant collection of strongly fully supported axial elements generate exactly the right-angled Artin group prescribed by orthogonality among their supporting domains; if that theorem fails, the factorization and every embedding obstruction built on it fail as well.
Editorial extensions
If this is right
- Virtually compact special groups admit finite graphs $\Lambda_M$ of arbitrarily large girth such that $A(\Lambda_M)$ does not embed into $G$.
- For the standard mapping class group structure, the obstruction takes the form $A(\Lambda)\le G \Rightarrow \Lambda\le C(S)^k$, recovering the curve-graph clique obstruction for essential curves.
- For a RAAG $A(\Gamma)$ with a rich-family structure, the expanded core graph embeds into the extension graph $\Gamma^e$, and coincides with it when the rich family contains all singleton subgraphs.
- The classes $\Xi_{\mathrm{cc}}$ and $\Omega$ are preserved under finite-index restriction, finite direct products, and the standard relatively hyperbolic construction, so the obstructions persist for groups built from these operations.
- In rank at most two, $A(\Lambda)\le G$ if and only if $\Lambda$ is an induced subgraph of $\widehat G_{\mathcal S}$, making the criterion complete rather than merely obstructional.
Reading between the lines
- If the rank-two completeness persists in higher rank, or for wider families of defining graphs, the expanded core graph would become a full RAAG-subgroup invariant for the classes $\Xi$ and $\Omega$; the paper records this as an open direction.
- The existence of the purely algebraic class $\Omega$ suggests the quasi-isometric embedding conclusion is not needed for the obstruction itself; the same combinatorics may hold for HHG structures where undistortedness fails but decomposition and commutation still hold.
- A concrete testable prediction is that for graph products or right-angled Coxeter groups carrying natural HHG structures, the expanded core graph should embed into (or coincide with) the extension graph, giving explicit RAAG embedding obstructions that can be computed from the defining graph.
Signed reviews
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The paper introduces the expanded core graph for hierarchically hyperbolic group structures in two axiom classes, Ξ and Ω, and uses it to study embeddings of right-angled Artin groups. The main structural result, Theorem 4.8, states that after replacing the standard generators of an embedded RAAG by suitable positive powers, the embedding factors through an intermediate RAAG generated by strongly fully supported axial elements; for the class Ξ this intermediate RAAG is quasi-isometrically embedded. The authors then prove that the extension graph of this intermediate RAAG embeds into the expanded core graph, yielding a Kim–Koberda-type obstruction in the clique graph of the expanded core graph (Theorem 5.1) and a complete equivalence in rank at most two (Theorem 5.3). They identify the expanded core graph with the curve disjointness graph for mapping class groups and with the extension graph for natural rich-family structures on RAAGs, and derive chromatic obstructions. The paper also proves permanence properties for finite direct products and relatively hyperbolic constructions, and includes an appendix comparing the metric and hierarchical orthogonal stabilizers, with a counterexample to a lemma of Abbott–Behrstock.
Significance. If the main theorems are correct, the paper gives a genuinely unifying framework: it recovers the Kim–Koberda obstruction for RAAGs and mapping class groups as special cases of a single HHG mechanism, and it introduces a useful new combinatorial invariant, the expanded core graph. The paper is carefully written in many local respects: the structural lemmas in Section 3 are proved in detail, the examples in Section 5 are worked out, and Appendix B is candid about corrections to previous work and about the difference between metric and hierarchical stabilizers. The central caveat is that the engine of the whole paper, Proposition 2.10, is imported without proof from the authors' own preprint [OP26, Theorem 3.9], and the main theorems inherit any gap in that result. The paper also gives only a very compressed proof of the Ω-analogue of the extension-graph embedding, which is load-bearing for the mapping class group application.
major comments (3)
- [Section 2, Proposition 2.10] The paper's central factorization and all of its main consequences rest on [OP26, Theorem 3.9], quoted as Proposition 2.10, which is neither proved nor even sketched here. This theorem asserts that powers of a geometrically irredundant collection of fully supported axial elements generate the RAAG prescribed by orthogonality of supports, and that under a bounded-orbit hypothesis the embedding is quasi-isometric. It is used essentially in Proposition 3.13, Lemma 4.5, Theorem 4.8, Theorem 5.1, and Theorem 5.3. Because the present paper explicitly corrects other statements from the same predecessor line in Remark 3.10 and Appendix B, the reader cannot treat [OP26, Theorem 3.9] as already vetted by this text. Please provide a proof of Proposition 2.10 in this paper, or state the main theorems as conditional on that external result with a precise page-and-theorem reference and a clear indication of where the proof can be found.
- [Section 5, Proposition 5.8] The proof of Proposition 5.8 is a blanket reference to the proofs of Lemmas 4.5 and 4.7, Theorems 4.8, 5.1, and 5.3, with 'strongly fully supported' replaced by 'fully supported'. This is load-bearing for the mapping class group application and for the Ω-version of the rank-two criterion. Under the weaker Ω axioms one no longer has the pointwise triviality on orthogonal coordinates supplied by Lemma 3.3, so the proof of Lemma 4.5 does not formally transfer; one must instead use the bounded-orbit clause in Definition 2.9 and weak commutativity in Definition 5.4(3b) to carry out the construction of the intermediate RAAG and the extension-graph embedding. Please give the analogue of Lemma 4.5 for Ω in detail, or at least provide a precise statement of the modified claim and the places where the weaker axioms replace the strong-support axioms.
- [Sections 3.2 and 5.2.1] The paper does not prove that the class Ξ is nonempty beyond the examples in Proposition 3.9, and for the rich-family compact-special example the orthogonal decomposition and commutative properties are imported from [AB23, Proposition 3.10(2)]. Since Appendix B explicitly gives a counterexample to [AB23, Lemma 3.4] and discusses why the main results of [AB23] survive, the paper should clarify exactly which statements from [AB23] are being used and verify directly that those statements apply to the present strengthened F_U-stabilizers setting. This is not a fatal issue, but it is a point where the reader needs to be able to check that the axioms of Ξ are actually satisfied by the claimed examples.
minor comments (5)
- [Title and Abstract] The title and abstract contain typographical errors: 'OR THOGONALITY' should be 'ORTHOGONALITY', and the phrase 'er-graph' in the introduction appears to be a typo. Please proofread the text carefully.
- [Definition 4.1] The symbol G is used both for the ambient group and for the set of minimal unbounded domains in the definition of the expanded core graph. This is confusing in statements such as 'for every U∈G'; please use a different notation, for example script G or G_min.
- [Lemma 2.16] The sentence 'suitable positive powers of finitely many elements lie in any prescribed finite-index subgroup' is standard but should be justified, since the powers must be chosen for finitely many elements simultaneously; a short argument using the finite index would make the lemma self-contained.
- [Proposition 5.9] The verification that the standard HHG structure on MCG(S) belongs to Ω relies on facts about pure mapping classes summarized in [OP26, §7.1]. Please state the precise facts used, especially the description of active domains of pure mapping classes, so that the reader can verify Definition 5.4(2) without consulting another preprint.
- [Appendix B, Example B.2] The verification that the adjoined bounded domains B_{n,m} satisfy all axioms of an HHS, in particular the bounded geodesic image axiom and the realization axiom, is compressed. Please expand this verification, since the example is used to refute [AB23, Lemma 3.4] and to motivate the strengthened F_U-stabilizers property.
Circularity Check
Factorization rests on Proposition 2.10, imported from the authors' own [OP26, Theorem 3.9], and is not reproved here.
-
self citation load bearing
[Section 2.1, Proposition 2.10 (quoted as [OP26, Theorem 3.9]); used essentially in Section 3.3 Proposition 3.13, Section 4.2 Theorem 4.8, and Section 5.1 Theorem 5.3.]
"Proposition 2.10 ([OP26, Theorem 3.9]). Let g_1,...,g_m ∈ G be a geometrically irredundant collection of axial elements. ... Suppose that there exists a positive integer k_1 such that [g_i^{k_1}, g_j^{k_1}] = 1 whenever U_i ⊥ U_j. Then there exists D > 0 such that ... the assignment v_i ↦ g_i^{d k_1} extends to an injective homomorphism A(Γ) → G."
This is the engine that converts orthogonality of supports into a RAAG subgroup. It is quoted verbatim from the authors' own earlier preprint [OP26, Theorem 3.9] and is not reproved in the present paper. Proposition 3.13 applies it directly to obtain undistorted RAAGs from strongly fully supported elements; Lemma 4.5, Theorem 4.8, Theorem 5.3, and their corollaries all inherit the conclusion from that one citation. Since the same predecessor work is explicitly corrected in Remark 3.10 and Appendix B, the reader cannot treat [OP26, Theorem 3.9] as independently certified by this text. The central factorization is therefore load-bearing on a self-citation.
full rationale
No fitted-parameter circularity or definitional reduction was found. The expanded core graph is a new construction, and Lemma 4.4's identification of its vertices with axial directions is proved in the paper from the strengthened F_U stabilizers property. The rank-two converse builds A(Λ) from vertices of Gp using that identification and Proposition 3.13, so it is a genuine theorem under the stated axioms. The single most load-bearing input, however, is Proposition 2.10, imported from the authors' own previous preprint [OP26] and used essentially in Proposition 3.13, Lemma 4.5, Theorem 4.8, and Theorem 5.3. The present paper does not reproduce the proof, and its own corrections to [OP26] in Remark 3.10 and Appendix B make the safety of that import less automatic. This is self-citation load-bearing rather than a by-construction reduction, so the score is 4 rather than 6-8.
Assumptions & free parameters
assumptions (8)
- standard math HHS axioms from [BHS19] as packaged in Definition 2.1, including nesting, orthogonality, transversality, consistency, realization, and the distance formula.
- standard math Acylindricity of G_U acting on C_U [BHS17, Corollary 14.4] and Osin's alternatives for acylindrically hyperbolic actions [Osi16, Theorem 1.1].
- domain assumption Proposition 2.10, quoted from [OP26, Theorem 3.9]: geometrically irredundant strongly fully supported axial elements with commuting orthogonal powers generate an injective, and under bounded orbits quasi-isometrically embedded, RAAG.
- ad hoc to paper Axioms of the class Xi (Definitions 3.1 and 3.7): FU stabilizers property, orthogonal decomposition property, commutative property.
- ad hoc to paper Axioms of the class Omega (Definition 5.4): supply, decomposition, weak commutativity, and commutation detection for fully supported axial elements.
- domain assumption Standard mapping class group facts: Nielsen-Thurston decomposition, pure mapping classes, and the standard HHG structure on the marking complex [BHS19, Theorem 11.1].
- domain assumption Rich-family factor systems on compact special groups and clean container properties [BHS17, ABD21].
- standard math Erdos's construction of finite graphs with high girth and chromatic number [Erd59] and [KK14, Lemma 3.2] on chromatic number of clique graphs.
Cite this review
Pith. "Pith review of From algebraic orthogonality to RAAG embedding obstructions in hierarchically hyperbolic groups." pith.science (2026). https://pith.science/paper/24GWKACM
@misc{pith2026260809535,
author = {Pith},
title = {Pith review of: From algebraic orthogonality to RAAG embedding obstructions in hierarchically hyperbolic groups},
year = {2026},
howpublished = {\url{https://pith.science/paper/24GWKACM}},
note = {Machine review of arXiv:2608.09535}
}
abstract
We introduce the $\textit{expanded core graph}$, which records minimal unbounded domains and their axial directions, for two classes of hierarchically hyperbolic group structures modeled on compact special groups and mapping class groups. Our main structural result shows that every embedding of a right-angled Artin group, after replacing its standard generators by positive powers, factors through an intermediate RAAG generated by suitably supported axial elements; for the class modeled on compact special groups, this intermediate RAAG is quasi-isometrically embedded. We show that its extension graph embeds into the expanded core graph. This yields a Kim--Koberda-type obstruction to RAAG embeddings and a complete embedding criterion when the rank is at most two. For the standard HHG structure on a mapping class group, the expanded core graph is the disjointness graph of essential curves, while for natural rich-family structures on a RAAG it recovers the extension graph. We also establish permanence results under finite direct products and the standard relatively hyperbolic construction.
Reviewed August 11, 2026 · model on record in the stance chip above.
Discussion (0). Continue with ORCID to comment.