REVIEW 2 major objections 3 minor 19 references
The third partial cohomology group and existence of extensions of semilattices of groups by groups
T0 review · 2 major / 3 minor · reviewed 2026-08-14 · deepseek-v4-flash
Pith's one-line read The paper proves that an abstract kernel $(A,G,\psi)$ for a semilattice of groups $A$ is realizable by an admissible extension exactly when $\mathrm{Obs}(A,G,\psi)$ is trivial in $H^3(G,C(A))$, and when realizable the equivalence classes…
desk verdict Careful nonabelian extension of the partial cohomology obstruction theory, but the classification theorem leans on unverified lemmas from an abelian-settings paper. 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 inverse monoid $\varsigma(A)$ of equivalence classes of isomorphisms between two-sided ideals of the semilattice of groups $A$, where two isomorphisms are identified if they differ by conjugation by an invertible multiplier of the common range. An abstract kernel is a unital partial homomorphism $\psi: G \to \varsigma(A)$; choosing representatives $\theta_g$ and multipliers $w_{g,h}$ that lift the compositions $\theta_g\theta_h$ to $\theta_{gh}$ produces the obstruction $\beta$ from the associativity identity (14). The restrictions of the $\theta_g$ to centers define a partial action of $G$ on $C(A)$, and $\beta$ is a partial $3$-cocycle for this partial module. This machinery converts the existential question 'does an extension exist?' into the cohomological question 'is $[\beta]=0$ in $H^3$?', and the passage from trivial $\beta$ to an extension uses the crossed-product construction by a twisted partial action.
What would settle it
Take $G=(\mathbb{Z}/2)^2$ and $A=A_e\cup A_f$ with $A_e\cong A_f\cong S_3$ and $e>f$ in the semilattice; let $\psi$ exchange the two fibers and act on them by automorphisms. For representatives $\theta_g$, compute $w_{g,h}$ from $\theta_g\theta_h(a)=w_{g,h}\theta_{gh}(a)w_{g,h}^{-1}$, derive $\beta$ from (14), and reduce $\beta$ modulo the partial coboundary. The theorem predicts that the class of $\beta$ in $H^3(G,C(A))$ is zero exactly when some admissible extension $U$ with $A \triangleleft U$ and $U/A\cong G$ exists; finding an extension with nonzero class, or proving none exists with zero class, would settle the claim.
Extended reading notes
Core claim
On the paper's own terms, for every abstract kernel $(A,G,\psi)$, the center $C(A)$ carries a partial $G$-module structure obtained by restricting any representative $\theta_g$ of $\psi(g)$ to the center of its domain. For any choices of representatives $\theta_g \in \psi(g)$ and multipliers $w_{g,h}$ satisfying $(\theta_g \theta_h)(a)=w_{g,h}\theta_{gh}(a)w_{g,h}^{-1}$, associativity forces a family $\beta(g,h,k)$ of central multipliers, and $\beta$ is a partial $3$-cocycle whose cohomology class $\mathrm{Obs}(A,G,\psi)\in H^3(G,C(A))$ is independent of the choices. The paper proves that $\mathrm{Obs}(A,G,\psi)$ is trivial if and only if there exists an admissible extension $A \to U \to G$ realizing $\psi$; in the trivial case, a normalized $2$-cocycle twists the chosen partial action into a second partial action whose crossed product is the desired extension. Theorem 2.22 then shows that the equivalence classes of admissible extensions realizing $\psi$ are in bijection with $H^2(G,C(A))$ by making $H^2$ act freely and transitively on those classes.
Load-bearing premise
The load-bearing premise is that every admissible extension of $A$ by $G$ is equivalent, as an extension, to a crossed product $A \ast_\Theta G$ built from a twisted partial action; if that classification from an earlier paper fails for some semilattice of groups, both the triviality criterion and the $H^2$ bijection could miss extensions.
Editorial extensions
If this is right
- A candidate abstract kernel is realizable precisely when one class $\mathrm{Obs}(A,G,\psi)$ in $H^3(G,C(A))$ is zero, so existence is decided by a single cohomology computation.
- Once one extension exists, all admissible extensions realizing the same kernel are obtained by twisting its partial action by normalized partial $2$-cocycles $v\in Z^2(G,C(A))$, with coboundaries giving equivalent extensions.
- The center $C(A)$ plays the role of the abelian kernel: $H^2(G,C(A))$ acts freely and transitively on equivalence classes of extensions, so the set of realizations is a torsor for this cohomology group.
- When $A$ is a single group, the construction collapses to the ordinary abstract-kernel classification, with obstructions in $H^3$ and extensions classified by $H^2$.
- Equivalent admissible extensions induce the same abstract kernel, so the kernel is a well-defined invariant of an admissible extension class.
Reading between the lines
- Beyond the paper, the same associativity-of-representatives calculation should obstruct extensions of arbitrary inverse semigroups by groups: replace $C(A)$ by the center of the inverse semigroup and keep the same quotient monoid.
- Beyond the paper, for acting groups $G$ with vanishing third partial cohomology on the relevant center, the theorem becomes an automatic existence statement for every abstract kernel, so computing $H^3$ for such groups would turn the criterion into a supply of extensions.
- Beyond the paper, the free and transitive $H^2$ action suggests an explicit normal form for extension classes: fix the action part $\theta$ and vary only the multiplier by normalized $2$-cocycles, which would make the bijection constructive in concrete examples.
Signed reviews
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The paper develops an obstruction theory for extensions of semilattices of (not necessarily abelian) groups by groups, in the framework of partial group cohomology built by the authors in [8,9]. It introduces the notion of a partial abstract kernel (A,G,ψ), namely a unital partial homomorphism ψ:G→ς(A) into the inverse monoid of classes of isomorphisms between ideals of A, and shows that every admissible extension of A by G gives such a kernel (Proposition 2.3). For a given kernel it constructs, in Section 2.3, a partial 3-cocycle β (the obstruction) with values in the center C(A), proves that its cohomology class Obs(A,G,ψ)∈H³(G,C(A)) is independent of all choices (Lemmas 2.11–2.13), and proves in Theorem 2.15 that (A,G,ψ) admits an admissible extension if and only if Obs(A,G,ψ) is trivial. The final subsection proves Theorem 2.22: whenever admissible extensions realizing ψ exist, their equivalence classes are in one-to-one correspondence with H²(G,C(A)), by making H²(G,C(A)) act freely and transitively on the set of such classes. The center C(A) is canonically made into a partial G-module via Proposition 2.8, independently of the choice of representatives from each ψ(g).
Significance. If the cited framework from [8,9] is admitted, this is a substantial and natural extension of the classical Eilenberg–MacLane theory of abstract kernels to the partial setting with nonabelian semilattices of groups, and it gives the expected third-cohomology obstruction and second-cohomology classification. The paper's main strengths are the detailed internal proofs: Lemmas 2.11, 2.12, 2.13, 2.15, 2.20, and 2.21 are given full arguments, the normalization Lemma 2.14 is a useful technical tool, and Proposition 2.8 provides a canonical, representative-independent partial module structure on C(A). The results are concrete and checkable: the vanishing of an explicit 3-cocycle is a testable criterion, and the H² classification is a falsifiable statement about extension classes. The principal risk is the heavy reliance on prior results from the authors' papers [8] and [9], which are cited as black boxes at the two most load-bearing junctures of Theorem 2.22; this is where the referee has concentrated the major comments.
major comments (2)
- [§2.4, proof of Theorem 2.22] The freeness of the H²-action rests on the implication "A∗_ΘG and A∗_{vΘ}G equivalent as extensions of A by G ⇒ Θ is equivalent to vΘ", which is delegated to [9, Lemmas 4.1 and 4.4]; the transitivity step similarly uses [9, Lemma 4.7] to pass from equivalent twisted partial actions to equivalent extensions. Since [9] is titled "Partial cohomology of groups and extensions of semilattices of abelian groups" and the manuscript nowhere states the exact scope of these lemmas, the reader cannot verify from the present paper that they apply to the nonabelian semilattice A appearing in Theorem 2.22. This is load-bearing: if [9, Lemmas 4.1 and 4.4] depend on commutativity of A, the freeness argument is unsupported and the asserted one-to-one correspondence with H²(G,C(A)) may overcount the extension classes. Please either state explicitly, with the precise statements or with a remark, that the relevant lemmas of [9] (including Lemmas 4.1, 4.4, and 4.7) hold for arbitrary semilattices of groups, or give the short nonabelian proofs. The same scope question applies to the use of [8, Theorem 6.12] at the start of the proof: the reduction to crossed-product extensions must preserve equivalence classes of extensions, not merely objects, for the transitivity and freeness arguments to yield the bijection.
- [§2.4, definition of the action in Theorem 2.22] The sentence "In view of Lemma 2.19 this is well defined" is not sufficient for the action of H²(G,C(A)) on the set of equivalence classes of admissible extensions. Lemma 2.19 proves only that vΘ is a twisted partial action when v∈NZ²(G,C(A)). For the action of the cohomology class [v] on the equivalence class of A∗_ΘG to be well defined one must also establish: (a) independence of the representative of [v], i.e., if v′=v·δ¹ε then A∗_{v′Θ}G and A∗_{vΘ}G are equivalent extensions; and (b) compatibility with equivalence of the base twisted partial action, i.e., if Θ≡Θ′ then A∗_{vΘ}G and A∗_{vΘ′}G are equivalent. Point (a) follows from Lemma 2.20 only in combination with a statement that twisting by v preserves equivalence of twisted partial actions, which is not proved or cited explicitly. Please add the missing argument for both points, or give a precise citation that covers them.
minor comments (3)
- [Lemma 2.14] In the displayed computation of the proof, "av′_{1,1} = v_{1,1}θ_1(θ_1(a)v^{−1}_{1,1}) = a" appears to contain θ_1(a) where θ^{-1}_1(a) is intended, and the final equality uses both the centrality of v_{1,1} and the fact that θ_1∈ψ(1)=[id_A] acts trivially on central multipliers; please expand this one-line justification so the step is unambiguous.
- [Theorem 2.15, proof of the "if" direction] The normalization step "since the class of β does not depend on the choice of (θ,w), we may also take θ_1=id_A and w_{1,1}=id_A" is compressed; a sentence explaining how Lemmas 2.12, 2.13, and 2.14 jointly justify the simultaneous normalization (and the subsequent conclusion w′_{1,g}=w′_{g,1}=id_{D_g} from (TPA6)) would improve readability.
- [Throughout] The typography needs a final pass: the abstract and Introduction contain broken words such as "semila ttices" and "F APES P", and the final addresses contain "Insituto"; these are purely presentational but should be corrected.
Circularity Check
No circularity: Theorem 2.15 and Theorem 2.22 are new consequences built on prior independent machinery; no prediction or classification step reduces to its own input by construction.
full rationale
The paper's central claims are not assumed in their inputs. The obstruction in Section 2.3 is constructed by choosing representatives θ_g ∈ ψ(g), proving the existence of multipliers w_{g,h} from the relation θ_gθ_h ∼ θ_gh id_{D_{h^{-1}}}, and then deriving the 3-cocycle condition in Lemma 2.11; the well-definedness of Obs(A,G,ψ) is proved in Lemmas 2.12 and 2.13, which are independent of the existence of an extension. The 'only if' direction of Theorem 2.15 uses the existence of an extension to produce a twisted partial action satisfying (TPA6), and the 'if' direction constructs such an action from a coboundary and invokes [8, Proposition 5.15] to build the extension; neither direction quotes Theorem 2.15 itself. Theorem 2.22 defines a genuine H^2(G,C(A))-action on extension classes and proves transitivity and freeness using previous lemmas; even if the cited [9, Lemmas 4.1/4.4] are later found not to cover nonabelian A, that would be a correctness gap in an imported lemma, not a circular reduction of the theorem to its own conclusion. The self-citations to [7,8,9] are load-bearing, but they are prior published results with stated assumptions that do not include the H^3 obstruction or the H^2 classification, so under the review rules they count as independent support rather than circularity.
Assumptions & free parameters
assumptions (5)
- domain assumption A is a semilattice of groups
- domain assumption The partial cohomology groups H^2(G,C(A)) and H^3(G,C(A)) are defined as in [7] and [9]
- domain assumption Every admissible extension of A by G is equivalent to a crossed product A ∗_Θ G for some twisted partial action Θ
- domain assumption The construction of the induced twisted partial action from an order-preserving transversal (from [8, Proposition 6.11]) is valid
- domain assumption Inverse semigroup multiplier facts from [8, Remarks 5.2 and 5.3]
Cite this review
Pith. "Pith review of The third partial cohomology group and existence of extensions of semilattices of groups by groups." pith.science (2026). https://pith.science/paper/MFHMIWJX
@misc{pith2026190900650,
author = {Pith},
title = {Pith review of: The third partial cohomology group and existence of extensions of semilattices of groups by groups},
year = {2026},
howpublished = {\url{https://pith.science/paper/MFHMIWJX}},
note = {Machine review of arXiv:1909.00650}
}
read the original abstract
We introduce the concept of a partial abstract kernel associated to a group G and a semilattice of groups A and relate the partial cohomology group H^3(G,C(A)) with the obstructions to the existence of admissible extensions of A by G which realize the given abstract kernel. We also show that if such extensions exist then they are classified by H^2(G,C(A)).
Reference graph
Works this paper leans on
-
[9]
Partial cohomology of groups and extensions of semilattices of abelian groups
Dokuchaev, M., and Khrypchenko, M. Partial cohomology of groups and extensions of semilattices of abelian groups. J. Pure Appl. Algebra 222 (2018), 2897–2930
work page 2018
-
[8]
Twisted partial actions and extensions of semilat- tices of groups by groups
Dokuchaev, M., and Khrypchenko, M. Twisted partial actions and extensions of semilat- tices of groups by groups. Int. J. Algebra Comput. 27 , 7 (2017), 887–933
work page 2017
-
[1]
Batista, E., Mortari, A. D. M., and Teixeira, M. M. Cohomology for partial actions of Hopf algebras. J. Algebra 528 (2019), 339–380
work page 2019
-
[2]
The algebraic theory of semigroups , vol
Clifford, A., and Preston, G. The algebraic theory of semigroups , vol. 2 of Math. Surveys and Monographs 7 . Amer. Math. Soc., Providence, Rhode Island, 1967
work page 1967
-
[3]
Recent developments around partial actions
Dokuchaev, M. Recent developments around partial actions. S˜ ao Paulo J. Math. Sci. 13 , 1 (2019), 195–247
work page 2019
-
[4]
Dokuchaev, M., and Exel, R. Associativity of crossed products by partial actions, enve lop- ing actions and partial representations. Trans. Amer. Math. Soc. 357 , 5 (2005), 1931–1952
work page 2005
-
[5]
Dokuchaev, M., Exel, R., and Sim ´on, J. J. Crossed products by twisted partial actions and graded algebras. J. Algebra 320 , 8 (2008), 3278–3310
work page 2008
-
[6]
Dokuchaev, M., Exel, R., and Sim ´on, J. J. Globalization of twisted partial actions. Trans. Amer. Math. Soc. 362 , 8 (2010), 4137–4160
work page 2010
Show all 19 references
-
[7]
Partial cohomology of groups
Dokuchaev, M., and Khrypchenko, M. Partial cohomology of groups. J. Algebra 427 (2015), 142–182
2015
-
[10]
Partial projective representations and partial actions
Dokuchaev, M., and Novikov, B. Partial projective representations and partial actions. J. Pure Appl. Algebra 214 , 3 (2010), 251–268
2010
-
[11]
Partial Galois cohomology and related ho- momorphisms
Dokuchaev, M., Paques, A., and Pinedo, H. Partial Galois cohomology and related ho- momorphisms. Quarterly J. Math. 70 , 2 (2019), 737–766
2019
-
[12]
Partial generalized crossed prod- ucts and a seven-term exact sequence
Dokuchaev, M., Paques, A., Pinedo, H., and Rocha, I. Partial generalized crossed prod- ucts and a seven-term exact sequence. arXiv:1908.05820
1908 arXiv
-
[13]
Schur’s theory for partial projective representations
Dokuchaev, M., and Sambonet, N. Schur’s theory for partial projective representations. Israel J. Math. 232 , 1 (2019), 373–399
2019
-
[14]
Twisted partial actions: a classification of regular C∗ -algebraic bundles
Exel, R. Twisted partial actions: a classification of regular C∗ -algebraic bundles. Proc. London Math. Soc. 74 , 3 (1997), 417–443
1997
-
[15]
Noncommutative boundaries and the ideal structure of reduced crossed products
Kennedy, M., and Schafhauser, C. Noncommutative boundaries and the ideal structure of reduced crossed products. Duke Math. J. 168 , 17 (2019), 3215–3260
2019
-
[16]
Cohomology of inverse semigroups
Lausch, H. Cohomology of inverse semigroups. J. Algebra 35 (1975), 273–303
1975
-
[17]
Lawson, M. V. Inverse semigroups. The theory of partial symmetries . W orld Scientific, Singapore-New Jersey-London-Hong Kong, 1998
1998
-
[18]
Homology
Maclane, S. Homology. Springer-Verlag, Berlin-Guttingen-Heidelberg, 1963
1963
-
[19]
Inverse semigroups
Petrich, M. Inverse semigroups . Pure and Applied Mathematics (New York). John Wiley & Sons, Inc., New York, 1984. A Wiley-Interscience Publicat ion. 22 MIKHAILO DOKUCHAEV, MYKOLA KHRYPCHENKO, AND MAYUMI MAKU TA Insituto de Matem ´atica e Estat ´ıstica, Universidade de S ˜ao P...
1984
Reviewed August 14, 2026 · model on record in the stance chip above.
Discussion (0). Continue with ORCID to comment.