arXiv subjects

Tarek Sayed Ahmed

Publications and source records attributed to Tarek Sayed Ahmed.

At least 19 recordsLinked to original sources

arXiv2020

Space and time via Topological and Tense cylindric algebras

Let $α$ be an arbritary ordinal, and $2<n<ω$. In \cite{3} accepted for publication in Quaestiones Mathematicae, we studied using algebraic logic, interpolation, amalgamation using $α$ many variables for topological logic with $α$ many variables briefly $\sf TopL_α$. This is a sequel to \cite{3}; the second part on modal cylindric algebras, where we study algebraically other properties of $\sf TopL_α$. Modal cylindric algebras are cylindric algebras of infinite dimension expanded with unary modalities inheriting their semantics from a unimodal logic $\sf L$ such as $\sf K5$ or $\sf S4$. Using the methodology of algebraic logic, we study topological (when $\sf L=S4$), in symbols $\sf TCA_α$. We study completeness and omitting types $\sf OTT$s for $\sf TopL_ω$ and $\sf TenL_ω$, by proving several representability results for locally finite such algebras. Furthermore, we study the notion of atom-canonicity for both ${\sf TCA}_{n}$ and ${\sf TenL}_n$, a well known persistence property in modal logic, in connection to $\sf OTT$ for ${\sf TopL}_n$ and ${\sf TeLCA}_n$, respectively. We study representability, omitting types, interpolation and complexity isssues (such as undecidability) for topological cylindric algebras. In a sequel to this paper, we introduce temporal cyindric algebras and point out the way how to amalgamate algebras of space (topological algebars) and algebras of time (temporal algebras) forming topological-temporal cylindric algebras that lend themselves to encompassing spacetime gemetries, in a purely algebraic manner.

math.LO↗

arXiv2020

Completely representable neat reducts

For an ordinal $α$, $\sf PEA_α$ denotes the class of polyadic equality algebras of dimension $α$. We show that for several classes of algebras that are reducts of $\PEA_ω$ whose signature contains all substitutions and finite cylindrifiers, if $\B$ is in such a class, and $\B$ is atomic, then for all $n<ω$, $\Nr_n\B$ is completely representable as a $\PEA_n$. Conversely, we show that for any $2<n<ω$, and any variety $\sf V$, between diagonal free cylindric algebras and quasipolyadic equality algebras of dimension $n$, the class of completely representable algebras in $\sf V$ is not elementary.

math.LO↗

arXiv2020

An infinite stratum of representability; some cylindric algebras are more representable than others

Let $2<n<m\leq ω$. Let $\CA_n$ denote the class of cylindric algebras of dimension $n$ and $\RCA_n$ denote the class of representable $\CA_n$s. We say that $\A\in \RCA_n$ is representable up to $m$ if $\Cm\At\A$ has an $m$-square representation. An $m$ square represenation is locally relativized represenation that is classical locally only on so called $m$-squares'. Roughly if we zoom in by a movable window to an $m$ square representation, there will become a point determinded and depending on $m$ where we mistake the $m$ square-representation for a genuine classical one. When we zoom out the non-representable part gets more exposed. For $2<n<m<l\leq ω$, an $l$ square represenation is $m$-square; the converse however is not true. The variety $\RCA_n$ is a limiting case coinciding with $\CA_n$s having $ω$-square representations. Let $\RCA_n^m$ be the class of algebras representable up to $m$. We show that $\RCA_n^{m+1}\subsetneq \bold \RCA_n^m$ for $m\geq n+2$.

math.LO↗

arXiv2019

A universal approach to Omitting types for various multimodal and quantifier logics

We intend to investigate the metalogical property of 'omitting types' for a wide variety of quantifier logics (that can also be seen as multimodal logics upon identifying existential quantifiers with modalities syntactically and semantically) exhibiting the essence of its abstract algebraic facet, namely, atom-canonicity, the last reflecting a well known persistence propery in modal logic. In the spirit of 'universal logic ', with this algebraic abstraction at hand, the omitting types theorem OTT will be studied for various reducts extensions and variants (possibly allowing formulas of infinite length) of first order logic. Our investigatons are algebraic, addressing (non) atom canoicity of varieties of algebra of relations. In the course of our investigations, both negative and positive results will be presented. For example, we show that for any countable theory $L_n$ theory $T$ that has quantifier elimination $< 2^ω$ many non-principal complete types can be omitted. Furthermore, the maximality (completeness) condition, if eliminated, leads to an independent statement from $\sf ZFC$ implied by Martin's axiom. $\sf OTT$s are approached for other algebraizable (in the classical Blok-Pigozzi sense) reformulations or/ and versions of $L_{ωω}$; they are shown to hold for some and fail for others. We use basic graph theory, finite combinatorics, combinatorial game theory orchestrated by algebraic logic.

math.LO↗

arXiv2019

Non elementary classes of relation and cylindric algebras

For any pair of ordinals $α<β$, $\sf CA_α$ denotes the class of cylindric algebras of dimension $α$, $\sf RCA_α$ denote the class of representable $\sf CA_α$s and $\sf Nr_αCA_β$ ($\sf Ra CA_β)$ denotes the class of $α$-neat reducts (relation algebra reducts) of $\sf CA_β$. We show that any class $\sf K$ such that $\sf RaCA_ω\subseteq \sf K\subseteq RaCA_5$, $\sf K$ is not elementary, i.e not definable in first order logic. Let $2<n<ω$. It is also shown that any class $\sf K$ such that $\sf Nr_nCA_ω\cap {\sf CRCA}_n\subseteq {\sf K}\subseteq \mathbf{S}_c\sf Nr_nCA_{n+3}$, where $\sf CRCA_n$ is the class of completely representable $\sf CA_n$s, and $\mathbf{S}_c$ denotes the operation of forming complete subalgebras, is proved not to be elementary. Finally, we show that any class $\sf K$ such that $\mathbf{S}_d\sf Ra CA_ω\subseteq {\sf K}\subseteq \mathbf{S}_c\sf RaCA_5$ is not elementary. It remains to be seen whether there exist elementary classes between $\sf RaCA_ω$ and $\mathbf{S}_d\sf RCA_ω$. In particular, for $m\geq n+3$, the classes $\sf Nr_nCA_m$, $\sf CRCA_n$, $\mathbf{S}_d\sf Nr_nCA_m$, where $\mathbf{S}_d$ is the operation of forming dense subalgebras are not first order definable.

math.LO↗

arXiv2018

Hilbert's tenth problem, Gödel's incompleteness, Halting problem, a unifying perspective

We formulate a property $P$ on a class of relations on the natural numbers, and formulate a general theorem on $P$, from which we get as corollaries the insolvability of Hilbert's tenth problem, Gödel's incompleteness theorem, and Turing's halting problem. By slightly strengthening the property $P$, we get Tarski's definability theorem, namely that truth is not first order definable. The property $P$ together with a "Cantor's diagonalization" process emphasizes that all the above theorems are a variation on a theme, that of self reference and diagonalization combined. We relate our results to self referential paradoxes, including a formalisation of the Liar paradox, and fixed point theorems. We also discuss the property $P$ for arbitrary rings. We give a survey on Hilbert's tenth problem for quadratic rings and for the rationals pointing the way to ongoing research in main stream mathematics involving recursion theory, definability in model theory, algebraic geometry and number theory.

math.LO↗

arXiv2018

Geometrical representation theorems for cylindric-type algebras

In this paper, we give new proofs of the celebrated Andréka-Resek-Thompson representability results of certain axiomatized cylindric-like algebras. Such representability results provide completeness theorems for variants of first order logic, that can also be viewed as multi-modal logics. The proofs herein are combinatorial and we also use some techniques from game theory.

math.LO↗

arXiv2016

Atom-canonicity in algebraic logic in connection to omitting types in modal fragments of L_{ω, ω}

Fix 2 n, Nr_n\CA_m(\subseteq CA_n) denotes the class of n-neat reducts of CA_m's. The existence of certain finite relation algebras and finite CA_n's lacking relativized complete representations is shown to imply that the omitting types theorem (OTT) fails for L_n with respect to clique guarded semantics (which is an equivalent formalism of its packed fragments), and for the multi-dimensional modal logic S5^n. Several such relation and cylindric algebras are explicitly exhibited using rainbow constructions and Monk-like algebras. Certain CA_n constructed to show non-atom canonicity of the variety S\Nr_n\CA_{n+3} are used to show that Vaught's theorem (VT) for L_{ω, ω}, looked upon as a special case of OTT for L_{ω, ω}, fails almost everywhere (a notion to be defined below) when restricted to L_n. That VT fails everywhere for L_n, which is stronger than failing almost everywhere as the name suggests, is reduced to the existence, for each n<m<ω, of a finite relation algebra R_m having a so-called m-1 strong blur, but R_m has no m-dimensional relational basis. VT for other modal fragments and expansions of L_n, like its guarded fragments, n-products of uni-modal logics like K^n, and first order definable expansions, is approached. It is shown that any multi-modal canonical logic L, such that $K^n\subseteq L\subseteq S5^n$, L cannot be axiomatized by canonical equations. In particular, L is not Sahlqvist. Elementary generation and di-completeness for L_n and its clique guarded fragments are proved. Positive omitting types theorems are proved for L_n with respect to standard semantics.

math.LO↗

arXiv2015

A brief history of algebraic logic from neat embeddings to rainbow constructions

We take a long magical tour in algebraic logic, starting from classical results on neat embeddings due to Henkin, Monk and Tarski, all the way to recent results in algebraic logic using so--called rainbow constructions invented by Hirsch and Hodkinson. Highlighting the connections with graph theory, model theory, and finite combinatorics, this article aspires to present topics of broad interest in a way that is hopefully accessible to a large audience. The paper has a survey character but it contains new approaches to old ones. We aspire to make our survey fairly comprehensive, at least in so far as Tarskian algebraic logic, specifically, the theory of cylindric algebras, is concerned. Other topics, such as abstract algebraic logic, modal logic and the so--called (central) finitizability problem in algebraic logic will be dealt with; the last in some detail. Rainbow constructions are used to solve problems adressing classes of cylindric--like algebras consisting of algebras having a neat embedding property. The hitherto obtained results generalize seminal results of Hirsch and Hodkinson on non--atom canonicity, non--first order definabiity and non--finite axiomatizability, proved for classes of representable cylindric algebras of finite dimension$>2$. We show that such results remain valid for cylindric algebras possesing relativized {\it clique guarded} representations that are {\it only locally} well behaved. The paper is written in a way that makes it accessible to non--specialists curious about the state of the art in Tarskian algebraic logic. Reaching the boundaries of current research, the paper also aspires to be informative to the practitioner, and even more, stimulates her/him to carry on further research in main stream algebraic logic.

math.LO↗

arXiv2015

Finite relation algebras and omitting types in modal fragments of first order logic

Let 2<n\leq l<m< ω. Let L_n denote first order logic restricted to the first n variables. We show that the omitting types theorem fails dramatically for the n--variable fragments of first order logic with respect to clique guarded semantics, and for its packed n--variable fragments. Both are modal fragments of L_n. As a sample, we show that if there exists a finite relation algebra with a so--called strong l--blur, and no m--dimensional relational basis, then there exists a countable, atomic and complete L_n theory T and type Γ, such that Γis realizable in every so--called m--square model of T, but any witness isolating Γcannot use less than $l$ variables. An $m$--square model M of T gives a form of clique guarded semantics, where the parameter m, measures how locally well behaved M is. Every ordinary model is k--square for any n<k<ω, but the converse is not true. Any model M is ω--square, and the two notions are equivalent if M is countable. Such relation algebras are shown to exist for certain values of l and m like for n\leq l<ωand m=ω, and for l=n and m\geq n+3. The case l=n and m=ωgives that the omitting types theorem fails for L_n with respect to (usual) Tarskian semantics: There is an atomic countable L_n theory T for which the single non--principal type consisting of co--atoms cannot be omitted in any model M of T. For n<ω, positive results on omitting types are obained for L_n by imposing extra conditions on the theories and/or the types omitted. Positive and negative results on omitting types are obtained for infinitary variants and extensions of L_{ω, ω}.

math.LO↗

arXiv2015

Splitting methods in algebraic logic: Proving results on non-atom-canonicity, non-finite axiomatizability and non-first oder definability for cylindric and relation algebras

We deal with various splitting methods in algebraic logic. The word `splitting' refers to splitting some of the atoms in a given relation or cylindric algebra each into one or more subatoms obtaining a bigger algebra, where the number of subatoms obtained after splitting is adjusted for a certain combinatorial purpose. This number (of subatoms) can be an infinite cardinal. The idea originates with Leon Henkin. Splitting methods existing in a scattered form in the literature, possibly under different names, proved useful in obtaining (negative) results on non-atom canonicity, non-finite axiomatizability and non-first order definability for various classes of relation and cylindric algebras. In a unified framework, we give several known and new examples of each. Our framework covers Monk's splitting, Andréka's splitting, and, also, so-called blow up and blur constructions involving splitting (atoms) in finite Monk-like algebras and rainbow algebras.

math.LO↗

arXiv2015

A solution to the finitizability problem for quantifier logics with equality

We consider countable so-called rich subsemigroups of (ωω,\circ); each such semigroup $T$ gives a variety CPEA_T that is axiomatizable by a finite schema of equations taken in a countable subsignature of that of ω-dimensional cylindric-polyadic algebras with equality where substitutions are restricted to maps in T. It is shown that for any such T, A\in CPEA_T iff A is representable as a concrete set algebra of ω-ary relations. The operations in the signature are set-theoretically interpreted like in polyadic equality set algebras, but such operations are relativized to a union of cartesian spaces that are not necessarily disjoint. This is a form of guarding semantics. We show that CPEA_T is canonical and atom-canonical. Imposing an extra condition on T, we prove that atomic algebras in CPEA_T are completely representable and that CPEA_T has the super amalgamation property. If T is rich and {\it finitely represented}, it is shown that CPEA_T is term definitionally equivalent to a finitely axiomatizable Sahlqvist variety. Such semigroups exist. This can be regarded as a solution to the central finitizability problem in algebraic logic for first order logic {\it with equality} if we do not insist on full fledged commutativity of quantifiers. The finite dimensional case is approached from the view point of guarded and clique guarded (relativized) semantics of fragments of first order logic using finitely many variables. Both positive and negative results are presented.

math.LO↗

arXiv2015

Problems on neat embeddings solved by rainbow constructions and Monk algebras

This paper is a survey of recent results and methods in (Tarskian) algebraic logic. We focus on cylindric algebras. Fix 2<n<ω. Rainbow constructions are used to solve problems on classes consisting of algebras having a neat embedding property substantially generalizing seminal results of Hodkinson as well as Hirsch and Hodkinson on atom-canonicity and complete representations, respectively. For proving non-atom-canonicity of infinitely many varieties approximating the variety of representable algebras of dimension n, so-called blow up and blur constructions are used. Rainbow constructions are compared to constructions using Monk-like algebras and cases where both constructions work are given. When splitting methods fail. rainbow constructions are used to show that diagonal free varieties of representable diagonal free algebras of finite dimension n, do no admit universal axiomatizations containing only finitely many variables. Notions of representability, like complete, weak and strong are lifted from atom structures to atomic algebras and investigated in terms of neat embedding properties. The classical results of Monk and Maddux on non-finite axiomatizability of the classes of representable relation and cylindric algebras of finite dimension n are reproved using also a blow up and blur construction. Applications to n-variable fragments of first order logic are given. The main results of the paper are summarized in tabular form at the end of the paper.

math.LO↗

arXiv2014

Atom-canonicity and complete representations for cylindric-like algebras, and omitting types for the clque guarded fragment of first order logic

Fix a finite ordinal n>2. We show that there exists an atomic, simple and countable representable CA_n, such that its minimal completion is outside SNr_nCA_{n+3}. Hence, for any finite k\geq 3, the variety SNr_nCA_{n+k} is not atom-canonical, so that the variety of CA_n's having n+k-flat representations is not atom-canonical, too. We show, for finite k\geq 3, that S_cNr_nCA_{n+k} is not elementary, hence the class of CA_n's having complete n+3-smooth representations is not elementary. We obtain analogous results by replacing flat and smooth, respectively, by (the weaker notion of) square; this give a stronger result in both cases and here we can allow k to be infinite. Our results are proved using rainbow constructions for CA's. We lift the negative result on atom-canonicity to the transfinite. We also show that for any ordinal α\geq ω, for any finite k\geq 1, and for any r\in ω, there exists an atomic algebra A_r\in SNr_\alphaCA_{α+k}\sim SNr_nCA_{α+k+1}, such that Π_{r/U} A_r\in RCA_α where U is any non--principal ultrafilter on ω. Reaping the harvest of our algebraic results we investigate a plethora of omitting types theorems for variants of first logic including its finite variable fragments and its packed fragment.

math.LO↗

arXiv2014

Algebraic analysis of temporal and topological finite variable fragments, using cylindric modal algebras

We study what we call topological cylindric algebras and tense cylindric algebras defined for every ordinal $α$. The former are cylindric algebras of dimension $α$ expanded with $\sf S4$ modalities indexed by $α$. The semantics of representable topological algebras is induced by the interior operation relative to a topology defined on their bases. Tense cylindric algebras are cylindric algebras expanded by the modalities $F$(future) and $P$ (past) algebraising predicate temporal logic. We show for both tense and topological cylindric algebras of finite dimension $n>2$ that infinitely many varieties containing and including the variety of representable algebras of dimension $n$ are not atom canonical. We show that any class containing the class of completely representable algebras having a weak neat embedding property is not elementary. From these two results we draw the same conclusion on omitting types for finite variable fragments of predicate topologic and temporal logic. We show that the usual version of the omitting types theorem restricted to such fragments when the number of variables is $>2$ fails dramatically even if we considerably broaden the class of models permitted to omit a single non principal type in countable atomic theories, namely, the non-principal type consting of co atoms.

math.LO↗

arXiv2014

Amalgamation, interpolation and congruence extension properties in topological cylindric algebras

Topological cylindric algebras of dimension α, αany ordinal are cylindric algebras with dimension αexpanded with αS4 modalities. The S4 modalities in representable algebras are induced by a topology on the base of the representation of its cylindric reduct, that is not necessarily an Alexandrov topolgy. For α>2, the class of representable algebras is a variety that is not axiomatized by a finite schema, and in fact all complexity results on representations for cylindric algebras, proved by Andreka (concerning number of variables needed for axiomatizations) Hodkinson (on Sahlqvist axiomatizations and canonicity) and others, transfer to the topological addition, by implementing a very simple procedure of `discretely topologizing a cylindric algebra' Given a cylindric algebra of dimension α, one adds αmany interior identity operations, the latter algebra is representable as a topological cylindric algebra if and only if the former is; the representation induced by the discrete topology. In this paper we investigate amalgamation properties for various classes of topological cylindric algebras of all dimensions. We recover, in the topological context, all of the results proved by Andreka, Comer Madarasz, Nemeti, Pigozzi, Sain, Sayed Ahmed, Sagi, Shelah, Simon, and others for cylindric algebras and much more.

math.LO↗

arXiv2014

Algebraisable versions of predicate topological logic

Motivated by questions like: which spatial structures may be characterized by means of modal logic, what is the logic of space, how to encode in modal logic different geometric relations, topological logic provides a framework for studying the confluence of the topological semantics for $\sf S4$ modalities, based on topological spaces rather than Kripke frames. Following research initiated by Sgro, and further pursued algebraically by Georgescu, we prove an interpolation theorem and an omitting types theorem for various extensions of predicate topological logic and Chang's modal logic. Our proof is algebraic addressing expansions of cylindric algebras using interior operators and boxes, respectively. Then we proceed like is done in abstract algebraic logic by studing algebraisable extensions of both logics; obtaining a plethora of results on the amalgamation property for various subclasses of their algebraic counterparts, which are varieties. Notions like atom-canonicity and complete representations are approached for finite dimensional topological cylindric algebras. The logical consequences of our algebraic results are carefully worked out for infinitary extensions of Chang's predicate modal logic and finite versions thereof, by restricting to $n$ variables, $n$ finite, viewed as a propositional multi-dimensional modal logic, and $n$ products of bimodal whose frames are of the form $(U, U\times U, R)$ where $R$ is a pre-order, endowed with diagonal constants.

math.LO↗

arXiv2014

Dedekind completions, neat embeddings and omitting types

Let n be finite >2. We show that any class between S\Nr_n\CA_{n+3} and RCA_n is not atom canonical, and any class containing the class of completely representable algebras and contained in S_c\Nr_n\CA_{n+3} is not elementary. We show that there is no finite variable universal axiomatization of many diagonal free reducts of representable cylindric algebras of dimension n, like the varieties of representable diagonal-free cylindric algebras and Halmos' polyadic algebras (without equality). We apply our hitherto obtained algebraic results to show that the omitting types theorem fails for finite variable fragments of first order logic with and without equality, having n variables, even if we count in severely relativized models as candidates for omitting single non-principle types. Finally, we show that for many cylindric-like algebras, like diagonal free cylindric algebras and Halmos' polyadic algebras with and without equality the class of strongly representable atom structures of finite dimension >2 is not elementary.

math.LO↗