1. Introduction
McKinsey and Tarski initiated the study of interior algebras in [
1], where the authors developed a structure called ’closure algebra’ using set-theoretic topology in a natural way. An interior algebra is a Boolean algebra enriched with an interior operator; such an algebra is an algebraic generalization of topological spaces. Essentially, an interior algebra is an algebraic structure that encodes the idea of the topological interior of a set (the interior of a subset
S of a topological space
X is the union of all subsets of
S that are open in
X). Their approach emphasized the interior operators, since the theory of closure operators [
2] can be easily translated into the language of interior operators (by replacing sets with their complements).
In general, a topology on a set X creates the mathematical framework for defining the concept of convergence and approximation, fundamental notions of mathematical analysis. A topology is a family of subsets of a given set that satisfy certain conditions: the empty set and the entire set are both included, arbitrary unions of subsets are included, and the intersection of any two subsets is also included. Every topology uniquely generates an interior operator, and every interior operator uniquely generates a topology. In a topological space, we define an interior operator as a function that maps any subset to the set of its interior points; we call it the topological interior operator. Conversely, given an interior operator i over the set of subsets , the set of its fixed points is a topology, and the associated topological interior operator is even i. Also, the topology generated by the interior operator is uniquely determined by the interior operator (it cannot be derived from other methods). Therefore, there is a bijection between the set of topologies on a non-empty set X and the set of interior operators on the set of subsets . Thus, it is enough to define an interior operator on the set to be able to define the other topological notions: open and closed set, neighbourhood, compact set, axioms of separation, separability, convergence, etc.
Domain theory is a mathematical framework to study specific partially ordered sets called domains [
3]. This theory is used in denotational semantics by formalizing the ideas of approximation and convergence [
4]. It is also closely related to topology.
Starting from these aspects, the authors of [
5] began to study a special class of interior operators
generated by a given interior operator
i and an ideal
I. The framework was given by the complete domains (which generalizes the set of subsets of a set endowed with the inclusion relation). Point sets are represented by minimal elements, and the fact that a set is represented as the union of its point subsets is ensured by the property that the set of minimal elements forms a basis. The role of the complement is played by the function
c. An interior operator is a kernel operator which preserves the infimum for finite sets. We were interested in the representation of the image
depending on the image
. Some properties of the operator
are studied; among them, two properties (denoted by
and
) were distinctive. Thus, if the ideal
I has the property
, then the image
has a simple representation
, showing a tight connection with the image
. Ideals with the property
are also called dominated ideals because each element of such an ideal is dominated by its image via the function
. Ideals with the property
are called Baire ideals because an important example is given by the ideal of meagre sets in a Baire space.
In this article, we present methods of constructing ideals with the properties and . Theorem 3 provides a method of constructing the ideals with the property . By customizing the functions f, g and h, and later the ideal J, several examples are presented. Next, we construct ideals with the property ; for this, we construct an interior operator by means of a topological interior operator on the set of minimal elements, and we present some topological results.
Many of the notations used in this article are taken from [
4].
For a non-empty set X, we denote the set of all subsets of X by . Let be a partially ordered set (poset), and let . We denote the upper set of A by , i.e., . Also, the set is the lower set of A, i.e., . For , the set is denoted by , and the set is denoted by . A subset A of X is an upper (lower) set if (). Also, denotes the upper bound set of A, i.e., , and denotes Sthe lower bounds set of A, i.e., .
If there exists the smallest element of X, it is denoted by ⊥ (called bottom). If there exists the biggest element of X, it is denoted by ⊤ (called top). A poset X is called pointed if it has the smallest element. If is a pointed poset, an element is called minimal if and . (or shortly, ) denotes the set of minimal elements of . If has the smallest element z, then z is called the supremum of A; it is denoted by . If has the biggest element z, then z is called the infimum of A and it is denoted by . For , we denote with , and with .
For a cardinal number , the set A is called -directed if with , . The 2-directed sets are called directed sets. It is obvious that a set is a directed set iff it is an -directed set for any finite cardinal . A set A is called an -ideal if it is a lower set and -directed; A is called a complete ideal if it is a lower set and -directed for any cardinal . If is a finite cardinal, an -ideal is simply called ideal. Let be the set of -ideals of , and be the set of ideals of . If , we denote by the family of directed sets in .
A poset is called a dcpo (directed complete partial order set) if any directed subset of X has supremum, and a poset is called complete if each subset of X has supremum and infimum.
Let be a pointed poset, be a non-empty family of subsets of X, and be a function. For , we say that z approximates x relative to , or z is essential part of x relative to , if for each set such that there exists and , there is such that . In this case, we denote . We say that x is -compact if it approximates itself relative to . For and , is denoted by , and we say z approximates x. Also, x is called compact if x is -compact. The set of -compact elements is denoted by , and the set of compact elements is denoted by . For , denotes the set . Then, for all .
A set is called a basis if there exists for all . A domain is called continuous if it has a basis, and it is called algebraic if it has a basis of compact elements.
Let be a function. denotes the set of fixed points of f, i.e., , and denotes the set . The function f is called isotone (antitone) if . The function composition is denoted by , and by , we mean for all . The function is called -Scott-continuous if f is isotone and such that , there exists and . Also, f is called Scott-continuous if f is -Scott-continuous.
2. Interior Operators on Complete Domains
We give the definition of the operator
introduced in [
5], and present some results; the results of this section are proved in [
5].
Definition 1. A poset is a complete domain if X is complete and is a basis.
Let be a complete domain, and be the function for all . To simplify the notations, we use instead of .
Proposition 1. The function c acts as a complement, and has the following properties:
for all ;
if , then ;
if , then ;
for all and ;
if , then ;
;
;
for all and .
A complete domain generalizes the family of subsets of a non-empty set, endowed with the inclusion relation. Point sets are represented by minimal elements. The fact that a set is represented as the union of its point subsets is ensured by the property that the set of minimal elements forms a basis. The role of the complement is played by the function
c. Moreover, any complete domain is an algebraic domain (see [
5]). Not every complete poset is a complete domain. A simple example is the set
with the partial order
. It is clear that
is a complete poset; however, it is not a complete domain.
Definition 2. A function is called an interior operator (on X) if
From the third property, if
, then we have
. In fact, the third property is equivalent to
and
for all
. Also, if
such that
, then
, namely
i is isotone. In conclusion, an interior operator is a kernel operator ([
3], p. 26) which preserves the infimum for finite sets. Since
, it results that
.
Let i be an interior operator, and let .
Definition 3. Let with , and let be a subset with at least two elements.
We say that x is -separated from y if such that and .
We say that x and y are i-separated if such that , , and .
We say that A is i-separated if for all with , x and y are i-separated.
If x and y are i-separated, then x is -separated from y, and y is -separated from x.
Proposition 2. Let the family of i-separated subsets of X, and be an isotone function such that (particularly, or ). Then, .
For each
, we define the set
Then, the function
is well defined:
Proposition 3. The function has the following properties:
is antitone;
for all ;
;
;
;
;
for all and .
These properties are proved in [
5] (in Propositions 12–18).
Because , if , then . If the reverse is true, we say that I has the property . Thus, , or . An ideal with the property is called a dominated ideal (relative to the operator i).
Proposition 4 ([5], Proposition 20). iff for all . We say that I has the property if , and denote . An ideal with the property is called a Baire ideal (relative to the operator i).
Proposition 5 ([5], Proposition 21). iff . If
, then
on
.
denotes the function
. In [
5], namely in Theorem 1 and Propositions 23–25, we proved the following properties of the function
.
Theorem 1. has the following properties:
is an interior operator on X;
;
;
;
If , then ;
If , then ;
If , then .
Thus, any interior operator
i on a complete domain
X together with any ideal
I generates a new interior operator
, closely related to the initial operator
i. An important property is provided by item 5, which expresses the image
as a function depending on the image
. More properties of
can be found in [
5].
3. Dominated Ideals: The -Ideals with Property
The dominated ideals relative to an interior operator i are ideals with the property , namely ideals with the elements for which .
Let be a complete domain, be a cardinal number, and . We denote and .
Remark 1. If A is a lower set, then and are lower sets.
If with , then .
Consequently, if A is a lower set, then , and for each transfinite cardinal with .
Lemma 1. Let i be an interior operator on X, and be two functions such that , h is isotone, and . Let be a lower set and . Then, for all with , there exist such that , and , where and for all , there exists an i-separated set such that .
Proof. Let such that . For all with , we have . Hence, for all with , there is such that and . We denote with . Let and for all . Since every chain in has an upper bound, according to Zorn’s lemma, there is a maximal element in . Let . Since , then . We have , where and . We show that y and z satisfy the conditions of the lemma.
Let , and assume that . Since and , we have . Then, for all . Since , if then . Therefore, and , which contradicts the maximality of . Then, , and so . Since , we have , and so . Since , there exists such that and . Since , we have . Let . Since and , , and then . Moreover, ; since and I is a lower set, it results that . Then, . Also, from , it follows that , and from , it results that . Thus, for all . Therefore, and , which contradicts the maximality of . In conclusion, , i.e., . Then, we obtain , and so .
On the other hand, since , we obtain for all . It follows that for all , there exists with and . Let , and let K be a arbitrary set with . Since , there exists a surjective function . Then, , where and for all .
We have , and show that each is i-separated. Let ; then, for all , we have ; since , it results that . Let with . Then, there exist such that , and . From with and , it follows that . In conclusion, is i-separated. □
We recall that the ideals with the property are ideals with the elements for which .
Theorem 2. Let be a cardinal number with , be a lower set, and be two functions such that , h is isotone, , and for all i-separated sets . Then, .
Proof. Since J is a lower set and h is isotone, is a lower set. Then, is a lower set. Since , it results that is -directed. Consequently, is an -ideal. We prove that . I denotes the set , and we consider such that . According to Lemma 1, there exist such that , and , where and for all , there exists an i-separated set such that . Since implies that , we have . Let . Since and is an i-separated set, from the hypothesis, we obtain . Thus, , where and , i.e., . It follows that . In conclusion, . □
Theorem 3. Let be a cardinal number with , be a complete ideal, and be three functions such that
and ,
and ,
h is -Scott-continuous.
Then .
Proof. Let be an i-separated set. Then, , and since J is a complete ideal, we obtain . Since , . Since h is -Scott-continuous, . It results that . Because , it follows that , and so . Then, . The desired conclusion follows from Theorem 2. □
We present some properties of the -Scott-continuous functions, and some examples.
Remark 2. Let be a non-empty family of subsets of X, and be two functions such that f is -Scott-continuous. Then, f is -Scott-continuous for any function and any non-empty subset .
Proposition 6. Let be two non-empty families of subsets of X, and be three functions such that , f is -Scott-continuous, and g is -Scott-continuous. Then, is -Scott-continuous.
Proof. Since f and g are isotone, then is isotone. For , since f is -Scott-continuous, we have . Since and g is -Scott-continuous, we have . Then, . In conclusion, is -Scott-continuous. □
Proposition 7. Let be three functions such that , f is -Scott-continuous, and g is -Scott-continuous. Then, is -Scott-continuous.
Corollary 1. Let be a function such that and f is -Scott-continuous. Then, is -Scott-continuous for all .
Proposition 8. The interior operator i is -Scott-continuous.
Proof. Since i is an interior operator, i is isotone. Let an i-separated set, and . Since A is i-separated, for all there is such that , and for all with . Let . Since , then . Since for all with we have and , we obtain . Then, . It follows that , and so . Then, , and so . It follows that . Therefore, , i.e., . □
Proposition 9. The function is -Scott-continuous.
Proof. Since i is isotone and c is antitone, it follows that is isotone. Let and . Since , we have . Then, , and so . On the other hand, , and then . Then, . It follows that , namely , namely . In conclusion, is -Scott-continuous. □
Proposition 10. The functions , , , , are -Scott-continuous.
Proof. Since i is isotone and c is antitone, it follows that all these functions are isotone. Let and let . Then, (see the demonstration of Proposition 8).
We show that is -Scott-continuous.
.
Let . Since , . Then . It follows that , and then . From , it results that . It follows that and . On the other hand, for all , and so . Hence, . It follows that . In conclusion, , and so is -Scott-continuous.
In Proposition 7, we consider and . Since f is -Scott-continuous and g is -Scott-continuous, from Remark 2, it results that f is -Scott-continuous, and g is -Scott-continuous. Then, is -Scott-continuous.
In Proposition 7, we take and . Since f is -Scott-continuous, f is -Scott-continuous. Since g is -Scott-continuous, it results that is -Scott-continuous.
In Proposition 7, we take and . Since f is -Scott-continuous, f is -Scott-continuous. Since g is -Scott-continuous, it results that is -Scott-continuous.
In Proposition 7, we take and . Since f is -Scott-continuous, and g is -Scott-continuous, it results that is -Scott-continuous. □
Theorem 4. Let be a cardinal number with , and be a complete ideal.
Then, .
Proof. We consider and in Theorem 3. It results that , , , and h is -Scott-continuous. Since implies that , it follows that and ; thus, we obtain . Then, . □
Corollary 2. Let be a cardinal number with . Then, .
Proof. Just consider in Theorem 4. □
Theorem 5. Let be a cardinal number with , and be a complete ideal such that . If , then .
Proof. Let . Then, and . According to Proposition 10, function h is -Scott-continuous, and so h is -Scott-continuous. If , we take . If , we take . If , we take . If , we take . If , we take . In all these situations, we have and . Consequently, according to Theorem 3, it follows that . □
Corollary 3. Let be a cardinal number with , and .
Then, .
Proof. Just consider in Theorem 5. □
4. Baire Ideals: The -Ideals with Property
The Baire ideals relative to an interior operator i are ideals with the property , namely ideals for which .
Let be a complete domain, and be a topology. For , denotes the interior of A relative to . Let be the function defined by .
Proposition 11. The function i is an interior operator.
Proof. We prove the following:
- 1.
,
- 2.
,
- 3.
implies ,
- 4.
for all ,
- 5.
.
1. .
2. For all . Then, , and so .
3. implies ; it follows that and Thus, .
4. Let . implies , meaning that . Let such that . Since , then . Also, ; it follows that . Then, . Therefore, . Consequently, . Thus, .
5. If , then and . This means that . Therefore, . Then, , and so . □
For the following results, let be a cardinal number, an -ideal relative to inclusion, and .
Proposition 12. .
Proof. Let such that . For each , there exists such that . Since is -directed relative to inclusion, . Then, for all . Hence, I is -directed. Now, let and . Let . Since and , there exists such that . Since , it results that . Thus, , and so . Then, . Consequently, I is a lower set, and so I is an -ideal. □
Proposition 13. iff .
Proof. If and , then exists such that . Then, for all , we have , implying and because and . Hence, , and so .
The reverse of this implication is obvious. □
Corollary 4. iff .
Proof. We assume and . Then, there is such that . Let . Since and , it results that . Then, , and so . From the hypothesis, ; hence, . Then, ; contradiction!
Now, let us assume that and . Thus, there is such that and . Then, . However, , and from the hypothesis, we obtain . Then, ; a contradiction! □
The following result is a consequence of Proposition 5 (claiming iff ).
Corollary 5. iff .
Let us consider ideals
with the property
. The Baire spaces have this property relative to the ideal of meagre sets [
6]; according to Baire’s theorem (see [
6], p. 296), each complete metric space and each compact Hausdorff space is a Baire space. A meagre set (also called first-category set) is a subset of a topological space that is small (negligible) in a precise sense. A subset
is
-nowhere-dense if its closure has empty interior; a subset
is said to be of the
-first-category if it is the union of a countable collection of
-nowhere-dense sets. From Corollary 5 and Baire’s theorem, we obtain the following result.
Theorem 6. Let τ be a topology on such that is a compact Hausdorff space, or there is d a metric on such that is a complete metric space and τ is the metric topology. Let be the ideal of τ-first-category sets. Then, .
A more interesting case is when can be organized as an algebraic domain.
Theorem 7. Let ⊑ be a partial order on such that is an algebraic domain, τ be the density topology, be a cardinal number, and be the ideal of all α-unions of τ-nowhere-dense subsets of . Then, .
Proof. We consider the Scott topology
on
, and the essential topology
on
(see [
7]).
Let
be the smallest common refinement of Scott topology and essential topology; this
is called the density topology on
, and it is generated by the basis
(see [
7]). According to [
7], since
is an algebraic domain, we have that the union of every family of non-empty
-nowhere dense sets is not a
-open set. Therefore,
, and so
. □