Search

A specialization semilattice is a semilattice together with a coarser preorder satisfying a compatibility condition. We show that the category of specialization semilattices is isomorphic to the category of semilattices with a congruence, hence equivalent to the category of semilattice epimorphisms. Guided by the above example, we hint to an ``internal'' characterization of surjective homomorphisms between general relational systems., Comment: 8 pages

We present a definition for the sum of a sequence of combinatorial games. This sum coincides with the classical sum in the case of a converging sequence of real numbers and with the infinitary natural sum in the case of a sequence of ordinal numbers. We briefly discuss other possibilities, such as the string limit, some "magical" variants of Hackenbush, as well as "Dadaist" infinite sums, which allow transfinite runs, while still being loopfree., Comment: v1 was actually dealing with two distinct operations; in v2 I have fixed the issue. Moreover, I have fixed a few details, added some comments and discussed other possibilities

We show that a variety with J\'onsson terms $t_1, \dots, t_{n-1}$ has directed J\'onsson terms $d_1, \dots, d_{n-1}$, for the same value of the indices. Refined results are obtained for locally finite varieties., Comment: 15 pages

We show that the class of Contact join-semilattices, as introduced by T. Ivanova, is not finitely axiomatizable. On the other hand, a simple finite axiomatization exists for the class of those join semilattices with a weak contact relation which can be embedded into the reduct of a Boolean algebra (equivalently, a distributive lattice) with a weak contact relation., Comment: 8 pages

We characterize those join semilattices with a multicontact relation which can be embedded into (the reduct of) a Boolean algebra with multicontact given by overlap, equivalently, into a distributive lattice with additive multicontact, still equivalently, into the multicontact semilattice associated to some topological space. A similar characterization is presented for embeddability into Boolean algebras or distributive lattices with a not necessarily overlap nor additive multicontact., Comment: v2, various improvements

We provide optimal bounds for $\alpha (\beta \circ \gamma \circ \dots )$ in $4$-distributive varieties, as well as some further partial generalizations., Comment: 9 pages

We show that for a large class of varieties of algebras, the equational theory of the congruence lattices of the members is not finitely based., Comment: 18 pages

We devise exact conditions under which a join semilattice with a weak contact relation can be semilattice embedded into a Boolean algebra with an overlap contact relation, equivalently, into a distributive lattice with additive contact relation. A similar characterization is proved with respect to Boolean algebras and distributive lattices with weak contact, not necessarily additive, nor overlap., Comment: v3: noticed that former Condition (D2-) is pleonastic; added two new equivalent conditions in Theorem 3.2. We realized all this after the paper has been published: variations with respect to the published version are printed in a blue character. v2: solved a problem left open in v1; added a counterexample; a few fixes

We study contact posets and show that every contact poset can be embedded into a Boolean poset with overlap contact relation. Contact posets and (nonadditive) contact semilattices have the superamalgamation property, Fra\"\i ss\'e limits and model completion. Some results apply to event structures with binary conflict, as introduced in computer science., Comment: v2, some additions

We show that the theories of partially ordered sets, lattices, semilattices, Boolean algebras, Heyting algebras with a further coarser partial order, or a linearization, or an auxiliary relation have the strong amalgamation property, Fra\"\i ss\'e limits and, in many cases, an $ \omega$-categorical model completion with quantifier elimination. The same applies to Kronheimer and Penrose's causal spaces. On the other hand, Urquhart doubly ordered sets do not have the amalgamation property. Our main tool is the superamalgamation property (the strong amalgamation property is not enough), thus we provide further arguments suggesting the usefulness of the superamalgamation property also in pure model theory, not only in algebraic logic., Comment: v2, largely expanded

The theory of two binary relations has the strong amalgamation property when the first relation is assumed to be coarser than the second relation, and each relation satisfies a chosen set of properties from the following list: transitivity, reflexivity, symmetry, antireflexivity and antisymmetry. The amalgamation property is maintained when we add families of unary operations preserving all the relations. As a consequence, we get the existence of Fra\"\i ss\'e limits for classes of finite structures. The results fail, for general comparability conditions, when three or more binary relations are taken into account, or when we add an operation preserving just one relation., Comment: comes from the splitting of ArXiv2103.00563. arXiv admin note: text overlap with arXiv:2103.00563

We say that an idempotent term $t$ is an exact-$m$-majority term if $t$ evaluates to $a$, whenever the element $a$ occurs exactly $m$ times in the arguments of $t$, and all the other arguments are equal. If $m

If $X$ is a closure space with closure $K$, we consider the semilattice $(\mathcal P(X), \cup)$ endowed with further relations $ x \sqsubseteq y_1, y_2, \dots, y_n$ (a distinct $n+1$-ary relation for each $n \geq 1$), whose interpretation is $x \subseteq Ky_1 \cup Ky_2 \cup \dots \cup Ky_n $. We present axioms for such "multi-argument specialization semilattices" and show that this list of axioms is complete for substructures, namely, every model satisfying the axioms can be embedded into some structure originated by some closure space as in the previous sentence. We also provide a canonical embedding of a multi-argument specialization semilattice into (the reduct of) some closure semilattice., Comment: Similar to arXiv:2201.09083 and arXiv:2207.11745 but treats the case of $n$-ary "specializations"

A specialization semilattice is a structure which can be embedded into $(\mathcal P(X), \cup, \sqsubseteq )$, where $X$ is a topological space, $ x \sqsubseteq y$ means $x \subseteq Ky$, for $x,y \subseteq X$, and $K$ is closure in $X$. Specialization semilattices and posets appear as auxiliary structures in many disparate scientific fields, even unrelated to topology. In general, closure is not expressible in a specialization semilattice. On the other hand, we show that every specialization semilattice can be canonically embedded into a "principal" specialization semilattice in which closure can be actually defined., Comment: Treats the nonadditive case; the additive case is somewhat simpler and has been treated in arXiv:2201.09083 v2: added a few details

The superamalgamation property is a strong form of the amalgamation property which applies to ordered structures; it has found many applications in algebraic logic. We show that superamalgamation has some interest also from the pure model-theoretical point of view. Under a completion assumption, we prove that the superamalgamation property for some class of ordered structures implies strong amalgamation for classes with added operations, including isotone, idempotent, extensive, antitone and closure operations. Thus, for example, partially ordered sets, semilattices, lattices, Boolean algebras and Heyting algebras with an isotone extensive operation (or an operation as above) have the strong amalgamation property. The theory of join semilattices with a closure operation has model completion. The set of universal consequences of the theory of Boolean algebras (or posets, semilattices, distributive lattices) with a closure or isotone, etc., operation is decidable., Comment: v.2, many additions. V. 3 a few further improvements, some corrections

We provide conditions under which the union of two first-order theories has the amalgamation property., Comment: 10 pages

A specialization semilattice is a join semilattice together with a coarser preorder $ \sqsubseteq $ satisfying an appropriate compatibility condition. If $X$ is a topological space, then $(\mathcal P(X), \cup, \sqsubseteq )$ is a specialization semilattice, where $ x \sqsubseteq y$ if $x \subseteq Ky$, for $x,y \subseteq X$, and $K$ is closure. Specialization semilattices and posets appear as auxiliary structures in many disparate scientific fields, even unrelated to topology. For short, the notion is useful since it allows us to consider a relation of "being generated by" with no need to require the existence of an actual "closure" or "hull", which might be problematic in certain contexts. In a former work we showed that every specialization semilattice can be embedded into the specialization semilattice associated to a topological space as above. Here we describe the universal embedding of a specialization semilattice into an additive closure semilattice. We notice that a categorical argument guarantees the existence of universal embeddings in many parallel situations., Comment: In v2 we present a few more details. In comparison with the version submitted to the journal, the arxiv version v2 contains: a slightly expanded introduction; the added Remark 4.2 and an appendix

An algebraization of the notion of topology has been proposed more than seventy years ago in a classical paper by McKinsey and Tarski. However, in McKinsey and Tarski's setting the model theoretical notion of homomorphism does not correspond to the notion of continuity. We notice that the two notions correspond if instead we consider a preorder relation $ \sqsubseteq $ defined by $a \sqsubseteq b$ if $a$ is contained in the topological closure of $b$. A specialization poset is a partially ordered set endowed with a further coarser preorder relation $ \sqsubseteq $. We show that every specialization poset can be embedded in the specialization poset naturally associated to some topological space, where the order relation corresponds to set-theoretical inclusion. Specialization semilattices are defined in an analogous way and the corresponding embedding theorem is proved. The interest of these structures arises from the fact that they also occur in many rather disparate settings, even far removed from topology., Comment: 30 pages

As well-known, in a finitary algebraic structure the set $\Gamma$ of all the non-generators is the intersection of all the maximal proper substructures. In particular, $\Gamma$ is a substructure. We show that the corresponding statements hold for complete semilattices but fail for complete lattices, when as the notion of substructure we take complete subsemilattices and complete sublattices, respectively., Comment: v.2, some minor improvements, a few historical comments

A folklore argument shows that Frattini's characterization of non-generators works in the framework of algebraic partially ordered sets. We provide characterizations of non-generators in arbitrary partially ordered sets. The validity of some characterizations is equivalent to Zorn's Lemma, hence to the Axiom of choice. We notice that working on closure spaces or posets provides no essential improvement., Comment: 8 pages

A variety is a class of algebraic structures axiomatized by a set of equations. An equation is linear if there is at most one occurrence of an operation symbol on each side. We show that a variety axiomatized by linear equations has the strong amalgamation property. Suppose further that the language has no constant symbol and, for each equation, either one side is operation-free, or exactly the same variables appear on both sides. Then also the joint embedding property holds. Examples include most varieties defining classical Maltsev conditions. In a few special cases, the above properties are preserved when further unary operations appear in the equations., Comment: 24 pages

We devise a condition strictly between the existence of an $n$-ary and an $n{+}1$-ary near-unanimity term. We evaluate exactly the distributivity and modularity levels implied by such a condition., Comment: v. 2, a few fixes, a few additions

We consider the situation in which some class of structures has the Strong Amalgamation Property (SAP) with the further requirement that the amalgamating structure can be taken over the set theoretical union of (the images of) the domains of the structures to be amalgamated. We call this property SAPU. The main advantage of SAPU over SAP is that there are many preservation theorems showing that we can merge different theories with SAPU still obtaining a theory with SAPU, hence with SAP. In particular, we get SAPU for various theories with many binary relations, each relation satisfying any set of properties chosen among transitivity, reflexivity, symmetry, antireflexivity, antisymmetry. We may also add unary operations, possibly satisfying some coarseness, isotonicity and closure conditions. SAPU is not limited to relational theories: the varieties defining the most usual Maltsev conditions in universal algebra have SAPU. Other examples include bounded directoids, order algebras and various generalization., Comment: v.2 many improvements and additions; v.3 split the manuscript into two pieces; the part dealing with two comparable binary relations has been moved to another manuscript

The class of linearly ordered sets with one order preserving unary operation has the Strong Amalgamation Property (SAP). The class of linearly ordered sets with one strict order preserving unary operation has AP but not SAP. The class of linearly ordered sets with two order preserving unary operations has not AP. For every set $F$, the class of linearly ordered sets with an $F$-indexed family of automorphisms has SAP. Corresponding results are proved in the case of order reversing operations. Various subclasses of the above classes are considered and some model-theoretical consequences are presented., Comment: v2 minor fixes

Congruence modular and congruence distributive varieties can be characterized by the existence of sequences of Gumm and J\'onsson terms, respectively. Such sequences have variable lengths, in general. It is immediate from the above paragraph that there is a variety with Gumm terms but without J\'onsson terms. We prove the quite unexpected result that, on the other hand, if some variety has both kinds of terms, then the minimal lengths of the sequences differ at most by $1$. It follows that every $r$-modular congruence distributive variety is $r^2-r+2$-distributive., Comment: v2: a few improvements + an appendix with some additional proofs v3: a generalization in the appendix

A. Mitschke showed that a variety with an $m$-ary near-unanimity term has J\'onsson terms $t_0, \dots, t _{2m-4} $ witnessing congruence distributivity. We show that Mitschke's result is sharp. We also evaluate the best possible number of Day terms witnessing congruence modularity. More generally, we characterize exactly the best bounds for many congruence identities satisfied by varieties with an $m$-ary near-unanimity term., Comment: v2 a few improvements, some simplifications, some exemplifications v3 minor fixes, a few improvements, some more notions introduced

We provide a Maltsev characterization of congruence distributive varieties by showing that a variety $\mathcal {V}$ is congruence distributive if and only if the congruence identity $\alpha \cap (\beta \circ \gamma \circ \beta ) \subseteq \alpha \beta \circ \gamma \circ \alpha \beta \circ \gamma \dots $ ($k$ factors) holds in $\mathcal {V}$, for some natural number $k$., Comment: some minor modifications + an appendix with some computations

We solve some problems about relative lengths of Maltsev conditions, in particular, we give an affirmative answer to a classical problem raised by A. Day more than fifty years ago. In detail, both congruence distributive and congruence modular varieties admit Maltsev characterizations by means of the existence of a finite but variable number of appropriate terms. A. Day showed that from J\'onsson terms $t_0, \dots, t_n$ witnessing congruence distributivity it is possible to construct terms $u_0, \dots, u _{2n-1} $ witnessing congruence modularity. We show that Day's result about the number of such terms is sharp when $n$ is even. We also deal with other kinds of terms, such as alvin, Gumm, Pixley, directed, specular, mixed and defective. All the results hold when restricted to locally finite varieties. We introduce some families of congruence distributive varieties and characterize many congruence identities they satisfy., Comment: v2 - v4 some improvements, corrections, simplifications; v5, v6 expanded the introduction; v7 corrected a formula

Let $\alpha$, $\beta$, $\gamma, \dots$ $\Theta$, $\Psi, \dots$ $R$, $S$, $T, \dots$ be variables for, respectively, congruences, tolerances and reflexive admissible relations. Let juxtaposition denote intersection. We show that if the identity $\alpha(\beta \circ \Theta) \subseteq \alpha \beta \circ \alpha \Theta \circ \alpha \beta $ holds in a variety $\mathcal {V}$, then $\mathcal {V}$ has a majority term, equivalently, $\mathcal {V}$ satisfies $ \alpha (\beta \circ \gamma) \subseteq \alpha \beta \circ \alpha \gamma $. The result is unexpected, since in the displayed identity we have one more factor on the right and, moreover, if we let $\Theta$ be a congruence, we get a condition equivalent to $3$-distributivity, which is well-known to be strictly weaker than the existence of a majority term. The above result is optimal in many senses, for example, we show that slight variations on the displayed identity, such as $ R (S \circ \gamma) \subseteq R S \circ R \gamma \circ R S$ or $R(S \circ T) \subseteq R S \circ RT \circ RT \circ RS$ hold in every $3$-distributive variety. Similar identities are valid even in varieties with $2$ Gumm terms, with no distributivity assumption. We also discuss relation identities in $n$-permutable varieties and present a few remarks about implication algebras., Comment: v2, entirely rewritten, the main theorems of v1 are now corollaries of more general results, v3, expanded the introduction, some further additions

Let surreal numbers be defined by means of sign sequences. We give a proof that if $S < T$ are sets of surreals, then there is some surreal $w$ such that $S < w < T$. The classical proof is simplified by observing that, for every set $S$ of surreals, there exists a surreal $s$ such that, for every surreal $w$, we have $S

We introduce the notion of a nest-representable tolerance and show that some results from our former paper "From congruence identities to tolerance identities" [CT] can be extended to this more general setting.

We study the validity of congruence inclusions of the form $ \alpha ( \beta \circ \alpha \gamma \circ \beta \circ \dotsc \circ \alpha \gamma \circ \beta ) \subseteq \alpha \beta \circ \alpha \gamma \circ \alpha \beta \circ \dots$ in congruence modular varieties, with an appropriate number of terms on each side and where juxtaposition denotes intersection. Two different methods using Day and Gumm terms are merged in order to obtain the so far best bounds. We introduce and study other related identities, possibly involving tolerances and admissible relations. We also slightly improve a result by A. Day, to the effect that if $n$ is even, then every variety with $n+2$ J\'onsson terms has $2n+1$ Day terms., Comment: v2: some additions; v3: only added a note about further developments, the rest of the manuscript remains unchanged

For every $n$, we evaluate the smallest $k$ such that the congruence inclusion $\alpha (\beta \circ_n \gamma ) \subseteq \alpha \beta \circ_{k} \alpha \gamma $ holds in a variety of reducts of lattices introduced by K. Baker. We also study varieties with a near-unanimity term and discuss identities dealing with reflexive and admissible relations., Comment: v3, v4 improvements and simplifications v5, v6 minor fixes

We present some identities dealing with reflexive and admissible relations and which, through a variety, are equivalent to congruence modularity.

We show that a variety $\mathcal V$ is congruence distributive if and only if there is some $h$ such that the inclusion (1) $\Theta \cap ( \sigma \circ \sigma ) \subseteq ( \Theta \cap \sigma ) \circ ( \Theta \cap \sigma ) \circ \dots $ ($h$ factors) holds in every algebra in $\mathcal V$, for every tolerance $\Theta$ and every U-admissible relation $\sigma$. By a U-admissible relation we mean a binary relation which is the set-theoretical union of a set of reflexive and admissible relations. For any fixed $h$, a Maltsev-type characterization is given for the inclusion (1). It is an open problem whether (1) is still equivalent to congruence distributivity when $\Theta$ is assumed to be a $U$-admissible relation, rather than a tolerance. In both cases many equivalent formulations for (1) are presented. The results suggest that it might be interesting to study the structure of the set of U-admissible relations on an algebra, as well as identities dealing with such relations., Comment: v.3, some improvements

We study an infinite countable iteration of the natural product between ordinals. We present an "effective" way to compute this countable natural product, in the non trivial cases the result depends only on the natural sum of the degrees of the factors, where the degree of a nonzero ordinal is the largest exponent in its Cantor normal form representation. Thus we are able to lift former results about infinitary sums to infinitary products. Finally, we provide an order-theoretical characterization of the infinite natural product, this characterization merges in a nontrivial way a theorem by Carruth describing the natural product of two ordinals and a known description of the ordinal product of a possibly infinite sequence of ordinals., Comment: v.2 added something

Suppose throughout that $\mathcal V$ is a congruence distributive variety. If $m \geq 1$, let $ J _{ \mathcal V} (m) $ be the smallest natural number $k$ such that the congruence identity $\alpha ( \beta \circ \gamma \circ \beta \dots ) \subseteq \alpha \beta \circ \alpha \gamma \circ \alpha \beta \circ \dots $ holds in $\mathcal V$, with $m$ occurrences of $ \circ$ on the left and $k$ occurrences of $\circ$ on the right. We show that if $ J _{ \mathcal V} (m) =k$, then $ J _{ \mathcal V} (m \ell ) \leq k \ell $, for every natural number $\ell$. The key to the proof is an identity which, through a variety, is equivalent to the above congruence identity, but involves also reflexive and admissible relations. If $ J _{ \mathcal V} (1)=2 $, that is, $\mathcal V$ is $3$-distributive, then $ J _{ \mathcal V} (m) \leq m $, for every $m \geq 3$ (actually, a more general result is presented which holds even in nondistributive varieties). If $\mathcal V$ is $m$-modular, that is, congruence modularity of $\mathcal V$ is witnessed by $m+1$ Day terms, then $ J _{ \mathcal V} (2) \leq J _{ \mathcal V} (1) + 2m^2-2m -1 $. Various problems are stated at various places., Comment: v. 4, added something

We note that if a sequence of real numbers converges to some limit, then the sequence of the corresponding strings in the surreal $+,-$ sign expansion representation converges, for a natural notion of string convergence, to the string corresponding to the limit, modulo an infinitesimal. The corresponding statement would be obviously false if we were considering, as strings, decimal or binary representations, instead. The string limit of a possibly transfinite sequence of surreal numbers is always defined and, when considering increasing sequences of ordinals, corresponds to taking the supremum. A transfinite sum can be defined using the string limit and this sum agrees with the representation of a surreal number in Conway normal form., Comment: Merger of version 1 (by Paolo Lipparini) and of arXiv:1210.5675v1 (by Istv\'an Mez\"o) with further additions

We study a transfinite iteration of the ordinal Hessenberg natural sum obtained by taking suprema at limit stages and show that such an iterated natural sum differs from the more usual transfinite ordinal sum only for a finite number of iteration steps. The iterated natural sum of a sequence of ordinals can be obtained as a "mixed sum" (in an order-theoretical sense) of the ordinals in the sequence, in fact, it is the largest mixed sum which satisfies a finiteness condition, relative to the ordering of the sequence. We introduce other infinite natural sums which are invariant under permutations and show that they all coincide in the countable case. Finally, in the last section we use the above infinitary natural sums in order to provide a definition of size for a well-founded tree, together with an order-theoretical characterization in the countable case. The proof of this order-theoretical characterization is mostly independent from the rest of this paper., Comment: v2 some improvements

We give conditions under which a product of topological spaces satisfies some local property. The conditions are necessary and sufficient when the corresponding global property is preserved under finite products. Further examples include local sequential compactness, local Lindel\"ofness, the local Menger property., Comment: 14 pages

The class $\mathfrak C $ relative to countably compact topological spaces and the class $\mathfrak P$ relative to pseudocompact spaces introduced by Z. Frol\'ik are naturally generalized relative to every topological property. We provide a characterization of such generalized Frol\'ik classes in the broad case of properties defined in terms of filter convergence. If a class of spaces can be defined in terms of filter convergence, then the same is true for its Frol\'ik class., Comment: 8 pages

As far as algebraic properties are concerned, the usual addition on the class of ordinal numbers is not really well behaved; for example, it is not commutative, nor left cancellative etc. In a few cases, the natural Hessemberg sum is a better alternative, since it shares most of the usual properties of the addition on the naturals. A countably infinite version of the natural sum has been used in a recent paper by V\"a\"an\"anen and Wang, with applications to infinitary logics. We provide an order theoretical characterization of this operation. We show that this countable natural sum differs from the more usual infinite ordinal sum only for an initial finite "head" and agrees on the remaining infinite "tail". We show how to evaluate the countable natural sum just by computing a finite natural sum. Various kinds of infinite mixed sums of ordinals are discussed., Comment: v3 added a remark connected with surreal numbers

Local versions of measurability have been around for a long time. Roughly, one splits the notion of $\mu $-completeness into pieces, and asks for a uniform ultrafilter over $\mu $ satisfying just some piece of $\mu $-completeness. Analogue local versions of weak compactness are harder to come by, since weak compactness cannot be defined by using a single ultrafilter. We deal with the problem by restricting just to a subset $P$ of all the partitions of $\mu $ into $<\mu $ classes and asking for some ultrafilter $D$ over $\mu $ such that no partition in $P$ disproves the $\mu $-completeness of $D$. By making $P$ vary in appropriate classes, one gets both measurability and weak compactness, as well as possible intermediate notions of "weak measurability". We systematize the above procedures and combine them to obtain variants of measurability which are at the same time weaker and local. Of particular interest is the fact that the notions thus obtained admit equivalent formulations through topological, model theoretical, combinatorial and Boolean algebraic conditions. We also hint a connection with Kat{\v{e}}tov order on filters.