On covers of quasivarieties of p-algebras

Zalán Gyenis · 2026

Read On covers of quasivarieties of p-algebras by Zalán Gyenis (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2603-14428.txt.

Introduction

Gyenis describes the covers of varieties of p-algebras inside the lattice of quasivarieties of p-algebras. Covers are the minimal steps upward in that lattice. The paper gives a full characterization of those covers. In particular, every variety of p-algebras has exactly one cover in the lattice of subquasivarieties. That single-cover fact settles a problem posed by Kowalski and Słomczyńska.

Catalog record

CREATORZalán Gyenis
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE30780
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2603.14428

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

On covers of quasivarieties of p-algebras Zalán Gyenis Abstract . This paper characterizes the covers of varieties of p-algebras in the lattice of quasivarieties of p-algebras. In particular, it is shown that every such variety has exactly one cover in the lattice of subquasivarieties. This answers a problem of Kowalski and Słomczyńska [ 6 ] . \urladdr 1991 Mathematics Subject Classification 06D15, 08B20, 06D20, 08B15 1. Overview This article characterizes covers of subvarieties in the lattice of subquasivarieties of p-algebras, settling Question 3 of Kowalski and Słomczyńska [ 6 ] . I follow the notation and terminology of [ 6 ] , but to make the paper self-contained, I recall below the most important facts from [ 6 ] that will be needed. Fundamentals of universal algebra (say, quasivarieties, subalgebras, etc.) are not recalled, but the readers might consult [ 2 , 4 ] . Recall that p-algebras are bounded distributive lattices endowed with a unary operation * satisfying the condition x\wedge y=0 iff x\leq y^{*} . It is known that p-algebras form a variety \mathsf{Pa} , and by a result of Lee [ 7 ] the lattice of subvarieties of \mathsf{Pa} is a chain \mathsf{Pa}_{-1}\subsetneq\mathsf{Pa}_{0}\subsetneq\mathsf{Pa}_{1}\subsetneq\cdots\subsetneq\mathsf{Pa} of order type \omega+1 , where \mathsf{Pa}_{-1} is the trivial variety, and \mathsf{Pa}_{k}=\mathbf{HSP}(\bar{\mathbf{B}}_{k}) for k\in\mathbb{N} . Here \bar{\mathbf{B}}_{k} is the k -atom Boolean algebra with an extra cover of the top element. In particular, \bar{\mathbf{B}}_{0} is the two-element Boolean algebra, and \mathsf{Pa}_{0} is the variety of Boolean algebras. \mathsf{Pa} is locally finite (finitely generated p-algebras are finite, see [ 3 ] ), and for any m>0 the variety \mathsf{Pa}_{m} is axiomatized by a single identity \bigvee_{i=1}^{m+1}(x_{i}\wedge\bigwedge_{j\neq i}x_{j}^{*})^{*}=1. (see [ 5 ] and Lemma 1.3 in [ 6 ] ). The main tool that is used in this paper is Priestley-duality for finite p-algebras. Objects dual to finite p-algebras are finite posets, and morphisms are order-preserving maps commuting with the maximal elements of the poset. Following [ 6 ] , these morphisms are called pp-morphisms. Let (P,\leq) be a finite poset, and for x\in P write M(x) for the set of maximal elements in P above x . (By the same token, M(P) is the set of all maximal elements of P ). Then h:(P,\leq)\to(Q,\leq) is a pp-morphism, if it preserves the ordering, and M(h(x))=h[M(x)] for every x\in P . For a finite p-algebra \mathbf{A}\in\mathsf{Pa} we write \delta(\mathbf{A}) for its dual poset; and for a finite poset P we write \varepsilon(P) for its dual p-algebra. Elements of \delta(\mathbf{A}) are the join-irreducible elements of \mathbf{A} ordered by the converse of the order inherited from \mathbf{A} ; while elements of \varepsilon(P) are the upsets of the poset. For finite \mathbf{A},\mathbf{B}\in\mathsf{Pa} we have \mathbf{A}\subseteq\mathbf{B} if and only if there is a surjective pp-morphism \delta(\mathbf{A})\twoheadrightarrow\delta(\mathbf{B}) . Conversely, for finite posets P and Q , there is a surjective pp-morphism P\twoheadrightarrow Q if and only if \varepsilon(Q)\subseteq\varepsilon(P) . (Here and later on \subseteq between algebras means subalgebra). Recall (see [ 5 ] ) what the dual posets of \bar{\mathbf{B}}_{m} look like: \delta(\bar{\mathbf{B}}_{0}) \delta(\bar{\mathbf{B}}_{1}) \delta(\bar{\mathbf{B}}_{2}) \delta(\bar{\mathbf{B}}_{m}) \ldots m points Figure 1. The dual posets \delta(\bar{\mathbf{B}}_{m}) of the p-algebras \bar{\mathbf{B}}_{m} generating \mathsf{Pa}_{m} . The following lemma is used frequently. Lemma 1.1 (Lemma 1.7 in [ 6 ] ) . Let \mathcal{P} be a family of finite posets, and let X be a finite poset. Then \varepsilon(X)\in\mathbf{Q}\big(\{\varepsilon(P):P\in\mathcal{P}\}\big) if and only if for some P_{1},\ldots,P_{n}\in\mathcal{P} there is a surjective pp-morphism P_{1}\uplus\cdots\uplus P_{n}\twoheadrightarrow X . Considering the lattice of quasivarieties of p-algebras, it is known that the interval [\mathsf{Pa}_{-1},\mathsf{Pa}_{2}] is a 4 -element chain. [ 6 ] lists 3 finite posets (see Figure 2 below) and claims that \mathbf{Q}(\varepsilon(P)) for any of these three posets is a cover of \mathsf{Pa}_{2} , 1 1 1 We believe that only one of these posets generates a cover of \mathsf{Pa}_{2} ; see the discussion towards the end of this paper. and [ 6 ] asks the question: Question 3. Are there any other covers of \mathsf{Pa}_{2} ? More generally, what are the covers of \mathsf{Pa}_{m} ? This paper answers both questions. Theorem 2.15 proves that for every m\geq 2 the variety \mathsf{Pa}_{m} has exactly one cover in the lattice of subquasivarieties of \mathsf{Pa} . We give a complete characterization of the posets generating these covers (Theorem 2.14 ), and in the last section we discuss and illustrate the case m=2 . 2. Subvariety covers of \mathsf{Pa}_{m} for m\geq 2 Lemma 2.1 . If \mathsf{Pa}_{m}\prec\mathsf{K} is a cover, then \mathsf{K}=\mathbf{Q}(\mathbf{A}) for a finite \mathbf{A}\in\mathsf{Pa} . Proof . Let \mathbf{A},\mathbf{B}\in\mathsf{Pa} be arbitrary non-trivial p-algebras. Claim 1: There are homomorphisms h:\mathbf{A}\to\mathbf{B} and k:\mathbf{B}\to\mathbf{A} . To see this consider the two-element \bar{\mathbf{B}}_{0}\in\mathsf{Pa} (which is the two-element Boolean algebra). By Glivenko’s theorem (see [ 1 ] ) there is a homomorphism from \mathbf{A} into a non-trivial Boolean algebra, and from any non-trivial Boolean algebra there is a homomorphism into the two-element Boolean algebra. Composition of homomorphisms gives a homomorphism \mathbf{A}\to\bar{\mathbf{B}}_{0} . Also, \bar{\mathbf{B}}_{0} is a subalgebra of \mathbf{B} , therefore the composition \mathbf{A}\to\bar{\mathbf{B}}_{0}\hookrightarrow\mathbf{B} gives h:\mathbf{A}\to\mathbf{B} . The other case is analogous. Claim 2: \mathbf{Q}(\mathbf{A},\mathbf{B})=\mathbf{Q}(\mathbf{A}\times\mathbf{B}) . That \mathbf{A}\times\mathbf{B}\in\mathbf{Q}(\mathbf{A},\mathbf{B}) and thus \mathbf{Q}(\mathbf{A}\times\mathbf{B})\subseteq\mathbf{Q}(\mathbf{A},\mathbf{B}) is straightforward. For the converse inclusion define f:\mathbf{A}\to\mathbf{A}\times\mathbf{B} by f(a)=(a,h(a)) . Then f is a homomorphism and as it is the identity on the first coordinate, f is injective. Therefore, \mathbf{A}\in\mathbf{IS}(\mathbf{A}\times\mathbf{B}) . Similarly, \mathbf{B}\in\mathbf{IS}(\mathbf{A}\times\mathbf{B}) . Consequently, \mathbf{Q}(\mathbf{A},\mathbf{B})\subseteq\mathbf{Q}(\mathbf{A}\times\mathbf{B}) . To complete the proof, let \mathsf{K}=\mathbf{Q}(\{\mathbf{A}_{i}:i\in I\}) be a cover of \mathsf{Pa}_{m} . Then there is \mathbf{A}_{i}\notin\mathsf{Pa}_{m} . As \mathsf{Pa}_{m} is axiomatized by a single identity \mathbf{ib}_{m} , it follows that \mathbf{A}_{i}\nvDash\mathbf{ib}_{m} , and there are finitely many elements a_{1},\ldots,a_{n}\in\mathbf{A}_{i} such that in the generated subalgebra \mathbf{A}=\mathbf{Sg}^{\mathbf{A}_{i}}(\{a_{1},\ldots,a_{n}\}) we also have \mathbf{A}\nvDash\mathbf{ib}_{m} , that is, \mathbf{A}\notin\mathsf{Pa}_{m} . The class \mathsf{Pa} is locally finite, ensuring \mathbf{A} is finite. As \mathsf{Pa}_{m} is generated by \bar{\mathbf{B}}_{m} , it follows that \mathsf{Pa}_{m}=\mathbf{Q}(\bar{\mathbf{B}}_{m})\subseteq\mathbf{Q}(\bar{\mathbf{B}}_{m},\mathbf{A})\subseteq\mathsf{K}\,. As \mathsf{K} is a cover, \mathsf{K}=\mathbf{Q}(\bar{\mathbf{B}}_{m},\mathbf{A}) . By claim 2 above, \mathbf{Q}(\bar{\mathbf{B}}_{m},\mathbf{A})=\mathbf{Q}(\bar{\mathbf{B}}_{m}\times\mathbf{A}) . But \bar{\mathbf{B}}_{m}\times\mathbf{A} is finite, concluding the proof. \blacksquare Lemma 2.2 . For a finite poset X , the following are equivalent: (1) \varepsilon(X)\in\mathsf{Pa}_{m} ; (2) every x\in X satisfies |M(x)|\leq m . Proof . (1)\Rightarrow(2) Since \mathsf{Pa}_{m}=\mathbf{Q}(\bar{\mathbf{B}}_{m})=\mathbf{Q}(\varepsilon(B_{m})) , where B_{m} is the poset with one bottom element and m maximal elements, Lemma 1.1 yields a surjective pp-morphism from a finite disjoint union of copies of B_{m} onto X . Every point of B_{m} has at most m maximal elements above it, so by preservation of maxima the same holds in X . (2)\Rightarrow(1) For each x\in X , choose a surjection from the set of maxima of B_{m} onto M(x) , and map the bottom of B_{m} to x . This gives a pp-morphism f_{x}:B_{m}\to X . Taking the disjoint union of all these maps yields a surjective pp-morphism \biguplus_{x\in X}B_{m}\twoheadrightarrow X . By Lemma 1.1 , \varepsilon(X)\in\mathbf{Q}(\varepsilon(B_{m}))=\mathsf{Pa}_{m} . \blacksquare Lemma 2.3 . For a finite poset X , the following are equivalent: (1) \mathsf{Pa}_{m}\subseteq\mathbf{Q}(\varepsilon(X)) ; (2) there is a surjective pp-morphism g:X\twoheadrightarrow\delta(\bar{\mathbf{B}}_{m}) . Proof . By Lemma 1.1 , \mathsf{Pa}_{m}\subseteq\mathbf{Q}(\varepsilon(X)) if and only if there is a surjective pp-morphism f:\biguplus_{i}X\twoheadrightarrow\delta(\bar{\mathbf{B}}_{m}) from finitely many copies of X onto \delta(\bar{\mathbf{B}}_{m}) . From this (2)\Rightarrow(1) is immediate. For (1)\Rightarrow(2) Write \bot for the bottom element of \delta(\bar{\mathbf{B}}_{m}) , and let x\in\biguplus_{i}X such that f(x)=\bot . Then f

Read the complete text