Volume 2022, Issue 1 1117881
Research Article
Open Access

A Note on Quotient Reflective Subcategories of O-REL

Muhammad Qasim

Corresponding Author

Muhammad Qasim

Department of Mathematics, School of Natural Sciences (SNS), National University of Sciences & Technology (NUST), H-12 Islamabad 44000, Pakistan nust.edu.pk

Search for more papers by this author
Ch. Muhammad Afaq Aslam

Ch. Muhammad Afaq Aslam

Department of Mathematics, School of Natural Sciences (SNS), National University of Sciences & Technology (NUST), H-12 Islamabad 44000, Pakistan nust.edu.pk

Search for more papers by this author
First published: 23 June 2022
Citations: 1
Academic Editor: Calogero Vetro

Abstract

In this paper, we examine the category of ordered-RELspaces. We show that it is a normalized and geometric topological category and give the characterization of local , local , and local T1 ordered-RELspaces. Furthermore, we characterize explicitly several notions of T0’s and T1 objects in O-REL and study their mutual relationship. Finally, it is shown that the category of T0’s (resp. T1) ordered-RELspaces are quotient reflective subcategories of O-REL.

1. Introduction

Many mathematical concepts were developed to describe certain structures of topology. The concepts of uniform convergences, uniform continuity, Cartesian closedness, completeness, and total boundedness do not exist in general topology. As a remedy, several approaches have been made to define these concepts in topology by mathematicians. For example, the concepts of uniform convergence in the sense of Kent [1] and Preuss [2], of set-convergence in the sense of Wyler [3], Tozzi [4] (which scrutinize filter convergence to bounded subset and generalizes classical point-convergence and supertopologies), of nearness by Bentely [5] and Herrlich [6] (particularly containing proximities and contiguities), and that of hullness by Čech [7] and Leseberg [8] containing the concepts of b-topologies and closures, respectively. In 2018, Leseberg [9] introduced a global concept which embeds the category of the above mentioned concepts into the category of RELspaces and RELmaps as subcategories. This construct, denoted by REL, forms thereby a topological category [9].

Classical separation axioms are very common and important ideas in general topology, and have many applications in all fields of mathematics. With the help of T0 reflection [10], characterizations of locally semi-simple morphisms are obtained in algebraic topology. Furthermore, lower separation axioms can be used in digital topology where they describe digital lines, and in image processing and computer graphs to construct cellular complexes [1113]. With having the understanding of T0 and T1 separation properties, several mathematicians have extended this idea to arbitrary topological categories [1418].

Classical separation axioms at some point p (locally) were generalized and have been inspected in [14], where the purpose was to describe the notion of strongly closed sets (resp., closed) in arbitrary set based topological categories [19]. Moreover, the notions of compactness [20], Hausdorffness [14], regular and normal objects [21], perfectness [20], and soberness [22] have been generalized by using the closed and strongly closed sets in some well-defined topological categories over sets [20, 2326]. Furthermore, the notion of closedness is suitable for the formation of closure operators [27] in several well-known topological categories [2830].

The salient objectives of this study are stated as follows:
  • (i)

    To define initial, final, discrete, and indiscrete objects in O-REL

  • (ii)

    To characterize local , local , and local T1 objects in O-REL and examine their mutual relationship

  • (iii)

    To give the characterization of , , and T1 objects in O-REL and examine their mutual relationship

  • (iv)

    To define several structures using ordered-RELspaces and discuss each of the T0 and T1 axioms there and examine their mutual relationship

  • (v)

    To examine the quotient-reflective properties of ordered-RELspaces.

2. Preliminaries

Recall [31, 32], a functor (the category of sets and functions) is called topological if
  • (i)

    is concrete

  • (ii)

    consists of small fibers

  • (iii)

    Every -source has a unique initial lift or every -sink has an unique final lift, i.e., if for every source (fj : X⟶(Xj, ηj))jI there exists an unique structure η on X such that g : (Y, ζ)⟶(X, η) is a morphism iff for each jI, fjg : (Y, ζ)⟶(Xj, ηj) is a morphism.

Moreover, a topological functor is called discrete (respectively, indiscrete) if it has a left (respectively, right) adjoint. In addition, a functor is called a normalized topological functor if constant objects, i.e., subterminals, have an unique structure, and said to be geometric functor if the discrete functor is left exact, i.e., it preserves finite limits [31, 32].

Let X be a non-empty, then is called a relative system for X, and it is denoted by REL(X). Moreover, REL(X) can be ordered by setting

iff for each , there exists such that .

Furthermore, we denote by and by .

Definition 1 (cf. [33]). Let Xϕ, then is called boundedness or B-set on X, if βX satisfies the following axioms:

  • (i)

    ϕβX

  • (ii)

    B2B1βX implies B2βX

  • (iii)

    aX implies {a} ∈ βX.

And for B-sets βX and βY a function g : XY is called bounded iff it satisfies;

(1)

By BOUND we denote the corresponding defined category.

Definition 2 (cf. [33]). The triple (X, βX, r) is called RELative space (shortly RELspace) if for the boundedness βX the function r: satisfies the following conditions:

  • (i)

    BβX and implies

  • (ii)

    {ϕ} ∉ r(B) for BβX

  • (iii)

    iff

  • (iv)

    aX implies {{a} × {a}} ∈ r({a}).

The RELspace (X, βX, r) is called ordered-RELspace provided that the following axiom holds:

  • (v)

    ϕB1BβX implies r(B1) ⊂ r(B).

Definition 3 (cf. [33]). Let (X, βX, r) and (Y, βY, v) be two RELspaces, then a bounded function g : XY is called RELative map (shortly RELmap) iff it satisfies the following condition:

(2)
where with (g × g)[R] = {(g × g)(a, c): (a, c) ∈ R} = {(g(a), g(c)): (a, c) ∈ R}. By O-REL, we denote the full subcategory of REL, whose objects are the ordered RELspaces. Note that OREL is a bireflective subcategory of REL [34].

Example 4. Let (X, TX) be a preuniform convergence space; then, the associated RELspace can be defined as follows:

(3)

Let PU-REL denotes the category, whose objects are triples and morphisms are RELmaps. Note that PUCONVPU-REL [9], where PUCONV is the category of preuniform convergence spaces and uniformly continuous maps as defined in [2].

Example 5. Let (X, βX, t) be a set-convergence space; then, the associated RELspace (X, βX, rt) can be defined by

(4)

, where and FIL(X) is the collection of all filters defined on X.

Let SET-REL denotes the category, whose objects are triples (X, βX, rt) and morphisms are RELmaps. Note that SETCONVSET-REL [9], where SETCONV is the category of set-convergence spaces and morphisms are b-continuous maps as defined in [3].

Example 6. Let (x, ζ) be prenearness space; then, the associated RELspace can be described as

(5)

Note that PNEARPN-REL [6, 9], where PNEAR is the category, whose objects are prenearness spaces and morphisms are nearness preserving maps as defined in [6], and PN-REL is the category of triples and morphisms are RELmaps.

Example 7. For a B-set βX, we put rb(ϕ)≔{ϕ}, and for BβX\{ϕ}, we set ; hence, (X, βX, rb) defines a RELspace, which is diagonal, meaning that for BβX\{ϕ} and , we can find xB such that .

Let Δ-REL be denote the corresponding defined full subcategory of REL; then, Δ-RELBOUND.

Remark 8. In this context, note that BORN, the full subcategory of BOUND, whose objects are the bornological spaces, then also has evidently a corresponding counterpart in REL.

Example 9. Let (X, βX, q) be b-topological space; then, the associated RELspace (X, βX, rq) is defined by

(6)

Note that b-TOPbTOP-REL [9], where bTOP-REL denotes the full subcategory of REL, whose objects are triples (X, βX, rq), and b-TOP denotes the category of b-topological spaces and b-continuous maps as defined in [9].

3. OREL as a Normalized and Geometric Topological Category

Note that the forgetful functor , where is topological in the following sense:

Lemma 10. Let be a collection of RELspaces. A source is initial in REL iff

(7)
and for all ,
(8)

Proof. It is given in [34]. Consequently, since O-REL is a full and isomorphism-closed subcategory which is bireflective in REL, it is topological, too.

Lemma 11. Let be a collection of ordered-RELspaces. A sink is final in OREL iff

(9)
where , and for ,
(10)

Proof. It is easy to observe that is an ordered-RELspace and is a RELmap. Suppose that is a mapping. We show that g is a RELmap iff gfj is a RELmap. Necessity is obvious since the composition of two RELmaps is RELmap again.

Conversely, let be a RELmap.

Then, first, we show that g is a bounded map. Let ; it implies that g(fj(Bj)) = gfj(Bj) ∈ βY. For our own convenience, take fj(Bj) = B, and since fj is a RELmap, then and consequently, g is bounded.

Now, let and . By the Definition 3, we have g(fj(Bj)) = gfj(Bj) ∈ rY(g(fj(Bj))). On the other hand, fj is a RELmap; it follows that Take . Then, we have and subsequently, which shows g is a RELmap.

Lemma 12. Let Xϕ, and (X, βX, r) be an ordered-RELspace.

  • (i)

    A RELstructure (βX, r) is discrete iff , where and with rdis(ϕ)≔{ϕ}

  • (ii)

    A RELstructure (βX, r) is indiscrete iff , where if βXϕ with rid(ϕ)≔{ϕ}.

Proof. By applying Lemma 11, we get the desired result.

Remark 13. The topological functor , where is normalized since an unique RELstructure βX = {∅}, and r(∅) = {∅} exists whenever X = ∅ and a unique RELstructure βX = {∅, {a}}, r(∅) = {∅} and r({a}) = {∅, {(a, a)}} exists whenever X = {a}. Furthermore, the topological functor is geometric since the regular sub-object of a discrete RELspace is discrete, and finite product of discrete RELstructures is discrete again.

4. Local T0 and Local T1 Ordered-RELspaces

In this section, we define notions for T0 and T1 ordered-RELspaces at some point.

Let X be any set and pX. We define the wedge product of X at p as the two disjoint copies of X at p and denote it as XpX. For a point aXpX, we write it as a1 if a belongs to the first component of the wedge product; otherwise, we write a2 that is in the second component. Moreover, X2 is the cartesian product of X.

Definition 14 (cf. [14]).

  • (i)

    A mapping Ap : XpXX2 is said to be principalp-axis mapping provided that

(11)
  • (ii)

    A mapping Sp : XpXX2 is said to be skewedp-axis mapping provided that

(12)
  • (iii)

    A mapping ∇p : XpXX is said to be fold mapping atp provided that

(13)

Assume that is a topological functor, with and pZ.

Definition 15 (cf. [14]).

  • (i)

    X is at p provided that the initial lift of the -source is discrete

  • (ii)

    X is at p provided that the initial lift of the -source is discrete, where XpX is the wedge product in , i.e., the final lift of the -sink , where i1, i2 represent the canonical injections

  • (iii)

    X is T1 at p provided that the initial lift of the -source is discrete.

Remark 16.

  • (i)

    In TOP, and T0 at p (respectively, T1 at p) are equivalent to the classical T0 at p (respectively, the classical T1 at p), i.e., for each aX with ap, there exists a neighborhood Na of  ``a" not containing  ``p" or (respectively, and); there exists a neighborhood Np of  ``p" not containing  ``a" [35]

  • (ii)

    A topological space X is T0 (respectively T1) iff X is T0 (respectively T1) at p for each pX [35]

  • (iii)

    Let be a topological functor, and be a retract of X. Then, if X is or T1 at p, then X is at p but not conversely in general [36].

Theorem 17. Let (X, βX, r) be ordered-RELspace and pX. Then, (X, βX, r) is at p if and only if for each aX with ap, the following holds:

  • (i)

    {a, p} ∉ βX

  • (ii)

  • (iii)

    or

  • (iv)

    or

Proof. Let (X, βX, r) be at p; we show the conditions (i) to (iv) are holding:

  • (i)

    Suppose that {a, p} ∈ βX for all aX with ap. Let U = {a1, a2} ∈ XpX, then since and for j = 1, 2, πjAp(U) = {a, p} ∈ βX (by the assumption), where πj : X2X for j =1,2 are projection maps. By Definitions 1 and 15 and Lemma 10, a contradiction, it follows {a, p} ∉ βX

  • (ii)

    Assume that and . Particularly, let and ; then, . By the assumption, and . Since (X, βX, r) is at p, it follows that , where is the discrete structure on XpX.

Similarly, for , we get , a contradiction to the discreteness of .

Thus, or .

  • (iii)

    Suppose that and . In particular, let and ; then, , and by the assumption and . Since (X, βX, r) is at p, we get that , where is the discrete structure on XpX

Similarly, for , we get , a contradiction.

Therefore, or .

  • (iv)

    Assume that and . Let and ; then, , , (by the assumption). Since (X, βX, r) is at p, it follows that , where is the discrete structure on XpX.

Similarly, for , we get , a contradiction.

Hence, or .

Conversely, suppose (i) to (iv) are holding.

Let be the initial structure induced by and , where is the product RELstructure on X2 and the discrete RELstructure on X.

We show that is the discrete REL structure on XpX, i.e., we show that and for , .

Let and ; if ∇pU = ∅, then U = ∅. Suppose ∇pU ≠ ∅. Then, we have ∇pU = {a} for some aX, and if a = p, then U = {p}; let ap; then, it further implies that U = {a1} or U = {a2} and U = {a1, a2}. By the assumption, πjApU = πjAp{a1, a2} = {a, p} ∉ βX (for j =1,2). Thus, U = {a1} and U = {a2}; subsequently, .

Now, implies B = {a1} and B = {a2}, and by Lemma 10, =

Suppose B = {a1}, then

; it follows that .

Since , we have the following possibilities of :

,

,

,

,

.

  • Case (i). Suppose . It follows that for all such that , . By Definition 2, . Similarly, . Therefore, holds

  • Case (ii). holds. The proof is similar to Case (i)

  • Case (iii). Let . It follows that for all such that , and , . By the assumption, we get . Similarly, . Thus, cannot be possible

  • Case (iv). Similar to Case (iii), we conclude that is not possible

  • Case (v). If . It follows that for all such that , and , for all implying . By the assumption, . Similarly, . Hence, is not possible.

Similarly, if B = {a2}, only Case (i) and Case (ii) are holding. By Lemma 12, is discrete.

Therefore, by Definition 15, (X, βX, r) is at p.

Theorem 18. Let (X, βX, r) be an ordered-RELspace and pX.

(X, βX, r) is T1 at p if and only if for any aX with ap, the following holds:

  • (i)

    {a, p} ∉ βX

  • (ii)

    and

  • (iii)

    and

  • (iv)

    and .

Proof. By following the same technique used in Theorem 17, and replacing the mapping Ap by the mapping Sp, we get the proof.

Theorem 19. All ordered-RELspaces are T0 at p.

Proof. Let (X, βX, r) be ordered-RELspace and pX. By Definition 15, we show that for each (where k = 1, 2) for some VβX and . ∇pU = ϕ implying U = ϕ. Suppose ∇pUϕ, it implies that ∇pU = {a} for some aX. If a = p, then ∇pU = {p} implying U = {p}.

Suppose ap, it follows that U = {a1}, {a2} or {a1, a2}. If U = {a1, a2}, then {a1, a2} ⊂ i1(V) for some VβX which shows that a2 should be in the first component of the wedge product XpX, a contradiction. In similar manner, for some VβX. Hence, . Thus, we must have U = {aj} for j = 1, 2 only and consequently, , the discrete RELstructure on XpX.

Now, for , by Lemma 10, = . Since , we have the following possibilities of :

,

,

,

,

.

In particular, for . It follows that, for all such that , and (for k =1,2), implying . It follows a2 (respectively, a1) in the first (respectively, second) component of the wedge product XpX, a contradiction. Similarly, for and , we get a contradiction.

Therefore, . Consequently, by Definition 15(i) and Lemma 10, (X, βX, r) is T0 at p.

5. T0 and T1 Ordered-RELspaces

In this section, we define generically notions of T0 and T1 in ordered-RELspaces.

The characterization of T0 objects in categorical topology has been an important idea in a topological universe. Therefore, several attempts has been made such as in 1971 Brümmer [15], in 1973 Marny [18], in 1974 Hoffman [17], in 1977 Harvey [16], and in 1991 Baran [14] to discuss various approaches to generalize classical T0 object and examined the relationship between different forms of generalized T0 objects. One of the main purposes of generalization is to define Hausdorff objects in arbitrary topological categories. In 1991, Baran [14, 37] also generalizes the classical T1 objects of topology to topological categories [14, 37]. In abstract topological categories [21], T1 objects are used to define T3, T4, normal objects, regular, and completely regular. To characterize separation axioms, Baran’s approach was to use initial and final lifts and discreteness.

In 1991, Baran [14] used the generic element method of topos theory introduced by Johnstone [38], to define generic separation axioms, due to the fact that points does not make sense in topos theory. In general, the wedge product XpX at p can be replaced by X2ΔX2 at diagonal Δ. Any element (a, b) ∈ X2ΔX2 is written as (a, b)1 (resp., (a, b)1) if it lies in the first (resp., second) component of X2ΔX2. Clearly, (a, b)1 = (a, b)2, if and only if a = b.

Definition 20 (cf. [14]).

  • (i)

    A mapping A : X2ΔX2X3 is called principal axis mapping provided that

(14)
  • (ii)

    A mapping S : X2ΔX2X3 is called skewed axis mapping provided that

(15)
  • (iii)

    A mapping ∇:X2ΔX2X2 is called fold mapping provided that

(16)

Any element (a, b) ∈ X2ΔX2 is written as (a, b)1 (resp., (a, b)1) if it lies in the first (resp., second) component of X2ΔX2. Clearly, (a, b)1 = (a, b)2 if and only if a = b.

Now, we replace the point p by any generic point δ and define the following separation axioms.

Definition 21. Let be a topological functor, with .

  • (i)

    X is provided that the initial lift of the -source is discrete [14]

  • (ii)

    X is T0 provided that the initial lift of the -source is discrete, where is the final lift of the -sink [14, 39]

  • (iii)

    X is called T0 provided that X doesn’t contain an indiscrete subspace with at least two points [18, 40]

  • (iv)

    X is T1 provided that the initial lift of the -source is discrete [14].

Remark 22.

  • (i)

    In TOP, all the properties of being T0, and (respectively, T1) are equivalent to those classical ones which are T0 (respectively, T1), i.e., for each a, bX with ab, there exists a neighbourhood Na of  ``a" not containing  ``b" or (respectively and), there exists a neighbourhood Nb of  ``b" not containing  ``a" [14, 18, 40]

  • (ii)

    In any topological category, implies is T0 but not conversely in general. Also, each of the and T0 has no relation to a T0 [39]

  • (iii)

    Let be a topological functor, and be a retract of X. Then, if X is (respectively T1), then X is at p (respectively T1 at p) but not conversely in general [36].

Theorem 23. Let (X, βX, r) be an ordered-RELspace.

(X, βX, r) is iff for each a, bX with ab, the following holds:

  • (i)

    {a, b} ∉ βX

  • (ii)

    or

  • (iii)

    or

  • (iv)

    or .

Proof. Suppose (X, βX, r) is , we show that conditions (i) to (iv) are holding.

  • (i)

    Suppose that {a, b} ∈ βX for each a,b∈ X, ab. Let U = {((a, b)1, (a, b)2)} ∈ X2ΔX2. Note that and π1A(U) = {a} ∈ βX. By the assumption, πkA(U) = πkA{((a, b)1, (a, b)2)} = {a, b} ∈ βX, where πk : X3X2 (for k =2,3) are projection maps. By Definitions 1 and 15 and Lemma 10, it leads to a contradiction, it follows that {a, b} ∉ βX

  • (ii)

    Suppose that and . Let and , then . By Definition 2, π1A{{((a, b)1, (a, b)2)}} = {{(π1A(a, b)1, π1A(a, b)2)}} = {{(a, a)}} ∈ r({a}) and by the assumption, π2A{{((a, b)1, (a, b)2)}} = {{(b, a)}} ∈ r({b}) and π3A{{((a, b)1, (a, b)2)}} = {{(a, b)}} ∈ r({a}). Since (X, βX, r) is , we conclude , where is the discrete structure on X2ΔX2

Similarly, for , we get , a contradiction.

Therefore, or .

  • (iii)

    Suppose that and . In particular, let and , then . By Definition 2, π1A{{((a, b)2, (a, b)1)}} = {{(a, a)}} ∈ r({a}) and by the assumption, π2A{{((a, b)2, (a, b)1)}} = {{(π2A(a, b)2, π2A(a, b)1)}} = {{(a, b)}} ∈ r({b}) and π3A{{((a, b)2, (a, b)1)}} = {{(π3A(a, b)2, π3A(a, b)1)}} = {{(b, a)}} ∈ r({a}). Since (X, βX, r) is it follows that , where is the discrete structure on X2ΔX2

Similarly, for , we get , a contradiction.

Thus, or .

  • (iv)

    Suppose that and . Let , and ; then, , and by Definition 2, π1A{{(((a, b)1, (a, b)1), ((a, b)2, (a, b)2)}} = {{(a, a)}} ∈ r({a}). By the assumption, π2A{{(((a, b)1, (a, b)1), ((a, b)2, (a, b)2)}} = {{(b, b), (a, a)}} ∈ r({b}) and π3A{(((a, b)1, (a, b)1), ((a, b)2, (a, b)2)}} = {{(a, a), (b, b)}} ∈ r({a}). Since (X, βX, r) is , we conclude , where is the discrete structure on X2ΔX2.

Similarly, for , we get , a contradiction to the discreteness of .

Hence, or .

Conversely, suppose (i) to (iv) are holding.

Let be the initial structure induced by and , where is the product RELstructure on X3 and the discrete RELstructure on X2.

We show that is the discrete RELstructure on X2ΔX2, i.e, and for .

Let and . If ∇U = ϕ, then U = ϕ. Suppose ∇Uϕ, then it follows that ∇U = {(a, b)} for some (a, b) ∈ X2. If a = b, then U = {(b, b)}. Next, let ab; then, we have U = {(a, b)1} or U = {(a, b)2} or U = {(a, b)1, (a, b)2} and π1AU = π1A{(a, b)1, (a, b)2} = {π1A(a, b)1, π1A(a, b)2} = {a, a}, and by the assumption, we get πkA{(a, b)1, (a, b)2} = {a, b} ∉ βX, (for k =2,3). Thus, U = {(a, b)1} or U = {(a, b)2}, and subsequently, .

Now, implies B = {(a, b)1} and B = {(a, b)2}, and by Lemma 10, .

Suppose B = U = {(a, b)1}, thensince , we have the following possibilities:

,

,

,

,

.
  • Case (i). If . It follows that for all and , , and by the Definition 2, we get .

  • In a similar way, and .

  • Thus, holds

  • Case (ii). holds. The proof is similar to Case (i)

  • Case (iii). . It follows that, for all such that . And , , and by Definition 2

  • Similarly, by the assumption and .

Therefore, is not possible.

Case (iv). Similar to Case (iii), we conclude that is not possible.
  • Case (v). If . It follows that, for all such that and implies . By Definition 2, .

Similarly, by the assumption, and .

Hence, is not possible.

Similarly, if B= {(a, b)2} only Case (i) and Case (ii) are holding. By Lemma 12, is discrete. Therefore, by Definition 21 (i), (X, βX, r) is .

Theorem 24. Let (X, βX, r) be an ordered-RELspace.

(X, βX, r) is T0 iff for each a, bX with ab, each of the following conditions are satisfied:

  • (i)

    {a, b} ∉ βX

  • (ii)

    or

  • (iii)

    or

  • (iv)

    or .

Proof. Let (X, βX, r) be T0, {a, b} ∈ βX and and and and and and .

Let U = {a, b}. Note that (U, βU, rU) is the subspace of (X, βX, r), where (βU, rU) is the initial lift of the ordered-RELsystem induced by the inclusion map i : SU and for any SU, SβU, whenever i(S) = SβU and for any , whenever .

By the assumption, i(U) = U = {a, b} ∈ βU and by Definition 1, we get .

Now, for any let . By Definition 2, i({{(a, a)}}) = {{(a, a)}} ∈ r({a}). By the assumption, implying that and .

Similarly, for , it follows that and .

Now, if then by the assumption, and .

And for then by the assumption, and .

Therefore, and , which is a contradiction by Lemma 12. Thus (i) − (iv) are holding.

Conversely, suppose that for all a, bX with ab, conditions (i) − (iv) are holding. We show that the initial structure (βU, rU) is not an indiscrete ordered-RELstructure on U. Let U = {a, b} ⊂ X. By the assumption, {a, b} ∉ βX and or and or and or . Thus, (U, βU, r) is not an indiscrete ordered-RELsubspace of (X, βX, r). Hence, by Definition 21 (iii), (X, βX, r) is T0.

Theorem 25. Let (X, βX, r) be an ordered-RELspace. Then, (X, βX, r) is T1 iff for all a, bX with ab, the following holds:

  • (i)

    {a, b} ∉ βX

  • (ii)

    and

  • (iii)

    and

  • (iv)

    and .

Proof. Similarly, using Theorem 23, and replacing mapping A by the mapping S, we obtain the proof.

Theorem 26. All ordered-RELspaces are T0.

Proof. Let (X, βX, r) be an ordered-RELspace. By Definition 21, we show that for any , Uik(V) (where k = 1, 2) for some and . If ∇U = ϕ implies U = ϕ. Suppose ∇Uϕ, hence ∇U = {(a, b)} for some (a, b) ∈ X2.

Suppose ab, it follows that U = {(a, b)1}or{(a, b)2}or{(a, b)1, (a, b)2}. If U = {(a, b)1, (a, b)2}, then {(a, b)1, (a, b)2} ⊂ i1(V) for some , which shows that (a, b)2 must be in the first component of X2ΔX2, a contradiction. Similarly, , for . Hence, U = {{(a, b)j}} for j = 1, 2. Consequently, , the discrete ordered-RELstructure on X2ΔX2.

Now, for , and by Lemma 10, . But gives the following possibilities:

,

,

,

,

.

In particular, for . Then, it follows, for all , and consequently, (for k =1,2). As a result, (a, b)2 (respectively, (a, b)1) is in the first (respectively, second) component of the wedge product X2ΔX2 which leads to a contradiction. Similarly, for and we get a contradiction.

Hence, . Thus, by Lemma 10 and Definition 21, (X, βX, r) is T0.

Remark 27. Let X be an ordered-RELspace.

  • (i)

    By Theorems 17 and 23, X is iff X is at p, for each pX

  • (ii)

    By Theorems 18 and 25, X is T1 iff X is T1 at p, for each pX

  • (iii)

    By Theorems 19 and 26, X is T0 iff X is T0 at p, for each pX

  • (iv)

    By Theorems 2326, T1T0T0 but the converse does not hold in general.

Corollary 28. Let be in PUREL. Then the following statements are equivalent.

  • (i)

    is

  • (ii)

    is , where is the category of pre-uniform convergence spaces and uniformly continuous maps

  • (iii)

    For each a, bX with , and for all , or , and .

Proof. By applying Example 4, Theorem 23, and Theorem 3.1.10 of [41].

Corollary 29. Let be in PUREL. Then, the following statements are equivalent:

  • (i)

    is T1

  • (ii)

    is T1PUCONV, where T1PUCONV is the category of T1 pre-uniform convergence spaces and uniformly continuous maps

  • (iii)

    For all a, bX with , and for all , and , and .

Proof. This follows from Example 4, Theorem 25, and Theorem 3.2.4 of [41].

6. Quotient-Reflective Subcategories of the Category of Ordered-RELspaces

Definition 30 (cf. [42]). Given a topological functor , and a full and isomorphism-closed subcategory of , we say that is

  • (i)

    Epireflective in and closed if and only if is closed under the formation of products and extremal subobjects (i.e., subspaces)

  • (ii)

    Quotient-reflective in if and only if is epireflective and is closed under finer structures (i.e., if , , , and id : AB is a -morphism, then ).

Theorem 31.

  • (i)

    Any O-REL, T0O-REL and T1O-REL is a quotient-reflective subcategory of O-REL

  • (ii)

    O-REL is a normalized topological construct

Proof. (i) Suppose OREL and . It can be easily verified that is a full and isomorphism-closed subcategory of O-REL and closed under finer structures. It remains to show that X is closed under extremal subobjects and closed under the formation of products.

Let AX and (βA, rA) denotes the sub O-REL structure on A, induced by the inclusion map i : AX. We show that (A, βA, rA) is O-REL space. Suppose that for any a,b ∈ A with ab, {a, b} ∈ βA, then by the inclusion map i({a, b}) = {i(a), i(b)} = {a, b} ∈ βX, a contradiction by Theorem 23.Thus, {a, b} ∉ βA.

Now, suppose and . It follows that, for all such that {(a, b)} ⊂ R, and by the inclusion map i{(a, b)} ⊂ i(R) implying {(a, b)} ⊂ R. It follows that , a contradiction by Theorem 23. Similarly, by the same argument , a contradiction. Therefore, or .

In similar way, or , and or . Hence, X is closed under extremal subobjects.

Next, suppose that X = ∏kIXk, where are the O-REL structures on Xk induced by projection map πk : XkX for all kI, i.e., . We show that (X, βX, rX) is a O-REL space. Let {a, b} ∈ βX for any a, bX with ab. Then, , a contradiction by Theorem 23. Thus, {a, b} ∉ βX.

Now, suppose and . It follows implies {(a, b)} ⊂ R. Then, there is kI for which akbkXk, and πk{(a, b)} ⊂ πkR implying {(πka, πkb)} = {(ak, bk)} ⊂ πkR. It follows that , a contradiction by Theorem 23. By the same process, , a contradiction. Hence, or . In similar way, or , and or . Hence, X is closed under the formation of products.

Therefore, the category O-REL is a quotient-reflective subcategory of O-REL.

Analogous to the above argument, setting OREL or T1OREL, the proof can be easily followed by using Theorem 24 or Theorem 25, respectively.

(ii) By the Theorem 26 and Remark 13, O-REL and O-REL are isomorphic categories and thus O-REL is normalized

Conflicts of Interest

The authors declare that they have no conflicts of interest.

Acknowledgments

We would like to pay our gratitude to Prof. Dieter Leseberg for his continuous support during the preparation of this article and specially providing the final structure of O-REL.

    Data Availability

    Data sharing is not applicable to this article as no datasets were generated or analyzed during the current study.

      The full text of this article hosted at iucr.org is unavailable due to technical difficulties.