172
Views
0
CrossRef citations to date
0
Altmetric
Pure Mathematics

On characterizations of a Tr− space: Continuity, Domination, and Alliance

ORCID Icon, &
Article: 2366518 | Received 29 Mar 2024, Accepted 06 Jun 2024, Published online: 24 Jun 2024

ABSTRACT

This paper studies some properties of a topological space called Trspace. Specifically, we study the continuity of functions in this space, explore the most interesting sets, which are dominating sets, and alliances, and finally explore the existence of cut points in such a space. We study all these properties using the concept of ϕproximity.

MATHEMATICS SUBJECT CLASSIFICATION:

1. Introduction

Recently, Mayila et al. (Citation2023) constructed a topology called a Trtopology on a nation X using decision spaces as open sets and the concept of inclusiveness. The definition of a decision space is a modification of that suggested by Bossert (Bossert, Citation1998). According to Mayila et al. (Citation2023), a decision space in a nation X is defined as a triple-ordered collection (U,SU,PX(U)) consisting of a set U(X) of people, the set PX(U) of decision authority of U in X, which is a set-valued function that gives the right, power, or obligation to make a decision and the duty to answer for its success or failure, and the set SU of possible priority situations which may occur in the administrative section of U. The properties of decision spaces and topological operators were studied. Finally, the connectedness of Trspaces was developed and studied. All of these were studied using the concept of ϕproximity.

In this paper, we use the same approach Mayila et al. (Citation2023) used to define the continuity of functions on Trspaces and study the one-to-one correspondence between these spaces. We will further study the idea of dominance and alliance in an international system setting, with the key goal being how one nation sits within the other. The paper will reach its conclusion by studying the existence of cut points in a ϕconnected Trspace.

1.1. Preliminaries

This paper will use the following notations and terminologies. Unless stated otherwise, all preliminary results are taken from Mayila et al. (Citation2023).

Definition 1.1.

Mayila et al. (Citation2023, p. 2)

  1. A decision space refers to a person or commission who has the authority by law or ordinance to make a final decision in approving or disapproving a particular matter within its administrative area.

  2. A decision authority is the right, power, or obligation to make a decision and the duty to answer for its success or failure.

  3. A decision index is the strength of the decision authority.

This definition was put into a mathematical language by Mayila et al. (Citation2023). However, we will add a few lines to clarify the components of such a mathematical representation. Our presentation begins by formalizing the concept of a decision authority as defined in Definition 1.1 into a more rigorous mathematical language. A decision authority is a set of “IF [condition]-THEN [authority]” statements. The conditions are situations or facts, while the authority is a conclusion, a rule, or an action. Therefore, the value of an authority is a mapping of an observation: a situation. These mappings are words, symbols, and signs made by one or more people for the intention of being used by some set of people as a guide in decision-making (Baligh, Citation1990, p. 361). Mathematically, we define a decision authority mapping as follows:

Definition 1.2.

Let S be an ndimensional set of conditions or situations, and s=(s1,s2,,sn)S be a vector of facts, and a={a1,a2,}A be a set of prescriptively allowed values of an authority variable, one of which is to be made into a fact, where A is a set of sets of decision values. Then

  1. A decision authority mapping is a function PX:SA which maps an observation sS into an appropriate authority PX(a)A.

  2. A decision authority mapping PX:SA can be represented by

    (1.1) PX={(s,a) :a=PX(s),sS}.(1.1)

Note 1.2.

In EquationEquation (1.1), we observe the following;

  1. The first component s of every pair (s, a) in the set PX is a vector s=(s1,,sn) of values of facts (existing situations). The component values si of s are the components of the situation.

  2. The second component a=PX(s) is a set a={a1,a2,} of actions, one of which is to become a fact for the corresponding situation s. The values aj in set a are the authorities, one of which is to be made a rule (action) for a particular situation.

Example 1.

Consider a statement “If a spy is arrested, take him to high court”. This statement is a single-element authority mapping. It has one element in the domain and one in its range. Let s= a spy is arrested, and a= take him to high court. Then, by definition 1.2, a=PX(s), and PX={(s,a)}={(if s,do a)}. Therefore, the entire statement is a single element of the mapping PX.

Some authority mappings are complex as they may have a then segment made up of many elements connected by and or by or. These kinds of mappings will be represented using a collection of simple unit authority mappings, which we defined (see definition 1.2).

Example 2.

Consider a complex authority mapping which states, “If a public official commits a corruption offence, then either take him to jail for life or confiscate all his material properties, but not both.” In this statement, the element situation s is a public official commits a corruption offence, and the set, a, of prescriptively allowed values of an authority variable includes: take him to jail for life but do not confiscate all his material properties; do not take him to jail for life but confiscate all his material properties. One of these elements of a has to be made a fact. Let p,q,r be the propositions that p= a public official commits a corruption offence, q= take him to jail for life, and r= confiscate all his material properties. In mathematical logic, a given complex statement is represented as: (p(qr))(¬q¬r), and the simplification gives ((pr)¬q)((pq)¬r). The predicates of the latter representation are (pr,¬q) and (pq,¬r). Therefore, the decision authority mapping for a situation p is given by

PX(p)={do rif not q,do qif not r.

and the mapping is PX={((p,r),(¬q)),((p,q),(¬r))}.

The definition of a decision authority mapping gives the value of an action in the form of words. Since the same decision authority in the same situation can be exercised by different people sitting in different authority positions in the chain, we need to quantify the decision authority by assigning a numerical value to each element of a set of prescriptively allowed values of a decision variable. These numbers will be used as measures of the strength of the decision authority someone has in his or her administrative position over a particular situation. Mayila (Mayila et al., Citation2023) called these numerical values decision indices.

To capture this setting, we redefine, without distorting the original meaning, the decision authority exercised by some set, U, of people in their administrative positions in various situations. With the exception of 2U being a superset of a set U, the letters and symbols in the definition below carry the same meanings as in definition 1.2.

Definition 1.3.

A decision authority exercised by a set U of people in their administrative positions on a set S of situations is a mapping PX:2U×SA that assigns to every nonempty element (u,s)2U×S a set a (A) of actions or rules, one of which must be made a fact.

With this new definition, EquationEquation (1.1) gives PX={((u,s),a):a=PX(u,s),(u,s)2U×S}. Since a is a set of decision values, which are words; then, we assign a number to each element aj of a={a1,a2,} that shows the strength of someone to make a decision. These numbers, as noted earlier, are decision indices. Therefore, we define the decision index, denoted by ρX(u,s), of an element u2U in a situation sS as a strength or magnitude of PX(u,s). The decision index ρX(u,s) is non-negative, and for simplicity, we will restrict its range to a closed and bounded unit interval [0,1] and define it by ρX:2U×S[0,1]. If ρX(u,s)=0, u has no authority over the situation s. If ρX(u,s)=1, then u has the authority PX(u,s) which is final on a situation s, and if 0<ρX(u,s)<1, u has authority PX(u,s) which allows him or her to make decisions which are not final, implying that the decisions he or she makes have to be channelled to the higher-ups for approval.

We have laid a more than sufficient foundation for a decision authority. Since our purpose in this section is not to discuss decision-making processes and the execution of decision rules, we will not pursue this further. However, the discussion provides a setting towards defining a more rigorous mathematical representation of a decision space than it was given by Mayila (Mayila et al., Citation2023).

Definition 1.4.

Let U be a subset of X, PX(U) be a set of decision authorities of U, and SU be a set of possible situations in the administrative area of U. Mathematically, we define the following:

  1. A decision taken by u2U over a situation sSU using a decision authority PX(u,s) is a function

    (1.2) ϕ:2U×SU×PXD(1.2)

    which maps a collection (u,s,PX(u,s)) into a decision element d=ϕ(u,s,PX(u,s))D, where D is a set of decisions.

  2. A decision space in a nation X is a triple ordered collection (U,SU,PX(U)) comprising the set U of people, the set SU of possible situations that may occur in the administrative section of U, and the set of decision authorities PX(U) of U in X on SU.

    If there is no confusion that may arise, the decision space (U,SU,PX(U)) will be denoted by U, and from now on, we will call the collection (X,SX,PX(X)) a nation space. For the sake of brevity and simplicity, (X,SX,PX(X)) will be denoted simply by X.

It is important to note that the decision authority mapping PX:2U×SUA is the heart of the decision space U.

Definition 1.5.

Mayila et al. (Citation2023, p. 3)

Let x be a point and U be a nonempty decision space in the nation X. Then

  1. xX is a direct member of U if x is contained directly in U (x is a leader). This person attends the decision-making processes in U and has the right to vote for a decision.

  2. xX is an indirect member of U if there exists at least one direct member yU who represents x in U. This person x has no right to vote for a decision even if s/he attends the meetings of U (ρU(x,s)=0 for all sSU). We will represent by yUx the phrase “y represents x”.

  3. U is a direct neighborhood of x if x is a direct member of U.

  4. U is an indirect neighborhood of x if x is an indirect member of U.

  5. U is a neighborhood of x if x is either a direct or indirect member of U.

Definition 1.6.

Mayila et al. (Citation2023, p. 3)

Let X be a nation and xX. Then

  1. X is said to be inclusive at x if there exists a decision space G in X that represents x.

  2. X is inclusive if it is inclusive at all points.

Lemma 1.7.

Mayila et al. (Citation2023, p. 3)

Every nonempty decision space in an inclusive nation X is the union of its one-point subsets composed by its direct members.

The following result is a consequence of Lemma 1.7, and its proof is easy to sketch.

Corollary 1.8.

Every nonempty subset of a decision space is a decision space.

The following proposition is an immediate consequence of the definitions 1.3, 1.4, lemma 1.7, and corollary 1.8.

Corollary 1.9.

If U is a decision space in a nation space X, then for each u2U, there exists at least one situation sSU such that ρX(u,s)>0.

Proof.

The proof here is very direct. Let u2U, and suppose in contrast that ρX(u,s)=0 for all sSU. This makes u a non-decision maker in U, and hence, by Lemma 1.7 and Corollary 1.8, this is a contradiction.□

Definition 1.10.

Mayila et al. (Citation2023, p. 4)

Let U and V be two distinct nonempty decision spaces in a Tr space (X,Tr). Then

  1. U is superior to V if V answers to U and we write VU.

  2. U and V are equivalent if UV and VU and we denote this by UV.

Definition 1.11.

(Maximal and Minimal element) Let S be an arbitrary ordered set. Then

  1. An element mS is said to be a maximal element of S if ms for all sS.

  2. Element lS is a minimal element of S if ls for all sS.

  3. A set S is bounded if it has both maximal and minimal elements.

The following proposition gives the boundedness of a Tr space:

Proposition 1.12.

Mayila et al. (Citation2023, p. 4)

The following statements hold for every Tr space (X,Tr);

  1. Every Tr space is linear.

  2. Every Tr space has a maximal and minimal decision space.

Corollary 1.13.

Mayila et al. (Citation2023, p. 5)

Every nonempty decision space in the space (X,Tr) has minimal and maximal elements.

Definition 1.14.

(ϕ-proximity)

Let U=(U,ρX(U),SEU) and V=(V,ρX(V),SFV) be two nonempty decision spaces in administrative sections E and F, respectively, in the space (X,Tr). Then

  1. The administrative intersection of U and V will be denoted by U ϕ V and defined by

    U ϕ V={xU  V:ρX{x}ρX(U)andρX{x}ρX(V)}.

  2. If UV (definition 1.10), we write

    1. U ϕ V=U

    2. U ϕ V=V

  3. If UV (definition 1.10), we write U ϕ V=U (or = V).

  4. U and V are said to be ϕnear to each other, denoted as in Naimpally and Peters (Citation2013: 67–68) by UδϕV if their administrative intersection is nonempty. Otherwise, U and V are ϕfar, adapting denotion of UδϕV.

  5. E and F are said to be ϕnear to each other if there exists a decision space in their union that answers to both.

Based on ϕ-proximity, Mayila (Mayila et al., Citation2023) defined the interior, closure, limit and boundary points of the subsets of the Trspace as follows:

Definition 1.15.

Mayila et al. (Citation2023, p. 5–6)

Let A be a subset of a Trspace (X,Tr). Then

  1. A point xA is a ϕ-interior point of A if there exists a decision space U in X such that x is a direct member of U and UA. The set of all ϕinterior points of A will be called ϕ-interior of A in X, and we shall denote it by intXϕA.

    It is worth noting that the properties of intXϕA mimic the well-known properties of interiors of sets under spatial or Cantorian set operations.

  2. A person xX is a ϕcontact point of A if every decision space Ux that directly contains x is administratively near to A. clXϕA will be used to denote the set of contact points of A in X under ϕproximity. Let {UαJ:|J|<} be an indexed family of all decision spaces in X, each of which contains x directly. Then, by definition, a person xX is a ϕcontact person of A only if

    UαϕAfor allαJ.

  3. A person xX is a ϕlimit point of A in X if, for every direct neighborhood Ux of x, Ux{x} is administratively near A. Let {UαJ:|J|<} be an indexed family of all decision spaces in X where each contains x directly. Then, by definition, a person xX is a ϕlimit point of A if

    (Uα{x}) ϕ Afor allαJ.

  4. A person xX is a ϕboundary point of A in X if every direct neighborhood of x is administratively near to both A and the complement of A in X.

    The set of all ϕboundary points of A in X is called the ϕboundary of A in X, and will be denoted by FrtXϕA.

Mayila (Mayila et al., Citation2023) then used this definition 1.15 and the concept of ϕproximity (see definition 1.14) and established various results regarding the properties of these new defined topological operators. The following are some of the results which were established, some of which will be used in our study. Their proofs are omitted here, unless it is a new preliminary result.

Theorem 1.16

Mayila et al. (Citation2023, p. 6)

A subset E of a Trspace X is inclusive if and only if intXϕE is nonempty.

Theorem 1.17

Mayila et al. (Citation2023, p. 7)

Let E be an administrative section of the space (X,Tr) and x be any point in X. If ρX (G)ρX{x} for every non-empty decision space G contained in E, then x is a ϕcontact point of E. The converse does not hold.

Corollary 1.18.

Mayila et al. (Citation2023, p. 7)

If x is a maximal element of the subset E of the Trspace (X,Tr), then x is a ϕfrontier point of E.

Corollary 1.19. 

Mayila et al. (Citation2023, p. 7)

If E1,E2 are two nonempty subsets of a subset E of the Trspace (X,Tr), then the intersection of their ϕclosures in E is nonempty.

The following theorem gives a topological meaning of a point being ϕnear to the set in a Trspace:

Theorem 1.20

Let x be a point and E be the subset of a space (X,Tr), respectively. Then, xδϕE if and only if xclXϕE.

Proof.

Suppose that xδϕE. Then, by definition 1.14, we have {x} ϕ E. Therefore, Ux ϕ E for each direct neighborhood Ux of xX. By definition of a ϕcontact point, we have xclXϕE. Conversely, suppose that x is a ϕcontact point of E, then for each direct neighborhood Ux of x, we get Ux ϕ E. If xUx ϕ E for each Ux, then ρX{x}ρXE. This implies that xδϕE. If xUx ϕ E, then there is an element yUx ϕ E with the property that ρX{y}ρXUx and ρX{y}ρXE. This further implies that xδϕy and yδϕE. By the property of transitivity, we have xδϕE.

The consideration of the ϕproximity of an element to a set in a Trspace (X,Tr) leads to the following theorem:

Theorem 1.21

Let p be a point and E,F be subsets of a Trspace (X,Tr,δϕ) endowed with ϕproximity δϕ. Then, the following statements hold:

  1. pδϕE implies E is nonempty.

  2. {p}  ϕ  E implies pδϕE.

  3. pδϕ(EF) if and only if pδϕE or pδϕF.

  4. If pδϕE and eδϕF for each eE, then pδϕF.

Proof.

  1. Given pδϕE, then {p} ϕ E. Then, there is an element y{p}E that is ϕnear to both {p} and E. This logically implies that E is nonempty. Otherwise, y would be ϕfar from E, contradicting the fact that y is ϕproximally (or administratively) contained in E.

  2. This follows from the definition 1.14.

  3. Suppose that pδϕ(EF). Then, for each direct neighborhood Up of pX, we have Up ϕ (EF). By elementary set theory, we have Up ϕ E or Up ϕ F for each Up. By definition of ϕcontact point, we get pclXϕE or pclXϕF. The latter statement further implies that pδϕE or pδϕF (see Theorem 1.20). Conversely, if pδϕE or pδϕF, then for each open neighborhood U of p, we have U ϕ E or U ϕ F. Using the distributive property of sets, we get U ϕ (EF) for each U. By theorem 1.20, xclXϕ(EF), and hence pδϕ(EF).

  4. Given that pδϕE and eδϕF for each eE. Then, by theorem 1.20, we have pclXϕE, and eclXϕF for each eE. The latter implies that EclXϕF so that clXϕEclXϕF. Therefore, pclXϕF, and by theorem 1.20, it follows that pδϕF.

Definition 1.22.

Mayila et al. (Citation2023, p. 8)

Let (X,Tr) be a Trtopological space. Then

  1. A ϕseparation of X is a pair U,V of spatially disjoint nonempty decision spaces in X such that UVXX and U ϕ V=.

  2. A space with a ϕseparation is said to be ϕseparated.

  3. A space without ϕseparation is said to be ϕconnected.

Since spatial nearness implies ϕnearness, then the following well-known result holds in ϕproximity.

Lemma 1.23.

Let Y be a subspace of (X,T). Then, Y is separated if and only if there exist two nonempty disjoint subsets A and B of Y whose union is Y, none of which contains the contact point of the other. The space Y is connected if such separation does not exist.

The statement and proof of Lemma 1.23 is found in Hocking and Young (Citation1961, pp. 15-16) and Munkers (Citation2000, pp. 148-149).

Theorem 1.24

Mayila et al. (Citation2023, p. 8)

Every subspace of the Trspace (X,Tr) is ϕconnected.

Corollary 1.25.

Mayila et al. (Citation2023, p. 8)

A subset E of a Trspace (X,Tr) is administratively connected if and only if the ϕinterior of E in X is nonempty.

Theorem 1.26 

Mayila et al. (Citation2023, p. 8)

Suppose that G is a nonempty decision space in (X,Tr) and that {Eα} is an indexed family of subsets of (X,Tr) such that EαϕG for each α. Then, the union E=G(αEα) is ϕconnected.

Corollary 1.27.

Mayila et al. (Citation2023, p. 9)

A Trspace (X,Tr) is ϕconnected.

Definition 1.28.

Mayila et al. (Citation2023, p. 10)

Let (X,Tr) be a Trspace. A chain of decision spaces, denoted by m-chain, is a finite sequence M=(G1,G2,G3,,Gn)Tr that satisfies the following axioms:

  1. Gk for all k{1,2,3,,n},

  2. Gk ϕ Gk+1 for all k=1,2,3,,n1,

  3. Every Gk=(Gk,ρX(Gk),SGk)M has only one immediate successor Gk+1M for every sSGk.

Definition 1.29.

Mayila et al. (Citation2023, p. 11)

Let M1={G11,,\allowbreakG1m} and M2={G21,,G2n} be m-chains from a point x to a point y in a Trspace X. The m-chain M1 is said to go strictly straight through M2 provided that

  1. each link G1j is strictly contained in some link G2k or lies within the two adjacent links G2k and G2k+1

  2. if G1j and G1l are both wholly contained in a set G2p for j < l, then for each integer k(j,l), the link G1k lies within G2p.

Theorem 1.30

Kinsey (Citation1993, p. 30)

The topological equivalence is an equivalence relation.

Definition 1.31.

Adams and Franzosa (Citation2008, p. 198)

Let X be a connected topological space. A cut set of X is a subset E of X such that X − E is disconnected. A cut point of a space X is a point pX such that the set {p} is a cut set of X. A cut set or cut point separates or disconnects X.

We have so far laid down a foundation for our study. We are now in a position to communicate the main findings of this paper.

2. Continuity and ϕcontinuity

If we are considering a function that is defined between two sets whose members are people, we need to be careful in defining the context of domain and co-domain. A person, as a point in a domain, is characterized by its attributes or by its state. The human being changes his mind, desires, needs, and any other states depending on the change in environments or situations. The function to a set of people may mean a change in the current state of a person into another state according to the requirements of that function. Since topology is a discipline that has been designed to formulate and treat continuity of functions (Singh, Citation2013, pp. 41–42; Latecki & Prokop, Citation1995; Kinsey, Citation1993, pp. 2–4; Alexandroff, Citation1961)), it is important to define the continuity of functions which show how points in two domains are interrelated. This will further enable us to study a one-to-one correspondence between the interacting nations and how one nation can sit within the other.

This section introduces the notion of continuous functions which preserve the ϕproximity of subsets of Trspaces.

Definition 2.1.

Let (X,Tr1,δϕ), (Y,Tr2,δϕ) be Trspaces endowed with ϕproximity. A function f:XY is said to be ϕcontinuous at a point p of X if for each subset E of X, pδϕE implies f(p)δϕf(E). The function f is ϕcontinuous if it is ϕcontinuous at every point of X.

Definition 2.1 can be extended further by replacing a single point p by another subset H of X. In this way, we say that the function f:XY is ϕcontinuous if for each subsets E,H of X, EδϕF implies f(E)δϕf(F) .

If f is bijective and both f and f−1 are ϕcontinuous, then f will be called a ϕhomeomorphism and the spaces (X,Tr1,δϕ) and (Y,Tr2,δϕ) are said to be proximally ϕhomeomorphic with respect to f.

Example 3.

The identity function id:(X,Tr,δϕ)(X,Tr,\allowbreakδϕ), given by id(x)=x for each xX, is ϕcontinuous. This is easy to demonstrate. If x is a point in X and E is a subspace of X with xδϕE, then id(x)δϕid(E) since id(x)=x and id(E)=E.

The following theorem gives some properties of ϕcontinuous functions between Trspaces.

Theorem 2.2

Let f:XY be a function defined on spaces (X,Tr1,δϕ), (Y,Tr2,δϕ). Then, the following statements are equivalent.

  1. For every point pX and subset E of X, pδϕE implies f(p)δϕf(E).

  2. For every subset E of X, f(clXϕE) is contained in clYϕ(f(E)).

  3. For each ϕclosed set C in Y, f1(C) is ϕclosedin X.

  4. For each decision space H in Y, f1(H) is a decision space in X.

Proof.

We show that (i)(ii)(iii)(iv)\allowbreak(i).

(i)(ii): Assume that for every point p and every subset E of X, pδϕEf(p)δϕf(E). Then, for every decision space U covering p in X, we have UδϕE. Thus pclXϕE (see Theorem 1.20). A similar argument holds for f(p) being administratively near E. If V is a decision space in Y containing f(p), then Vδϕf(E), so that f(p)clYϕf(E). Therefore, by hypothesis, if pδϕE implies f(p)δϕf(E), then pclXϕE implies f(p)clYϕf(E). Hence f(clXϕE)clYϕf(E) for each EX.

(ii)(iii): Let C be ϕclosed in Y. To show that f1(C) is ϕclosed in X, it is enough to show that clXϕf1(C) is administratively contained in f1(C). By elementary set theory, f(f1(C))C. Since C is ϕclosed in Y, clYϕf(f1(C))C. By hypothesis (ii), clYϕf(f1(C) contains f(clXϕf1(C)), so that C contains f(clXϕf1(C)). By elementary set theory, we have that f1(C) contains clXϕf1(C), as desired.

(iii)(iv): This is obvious.

(iv)(i): Suppose that pδϕE, and G is a decision space that contains f(p) in Y. By hypothesis (iv), f1(G) is a decision space that contains p in X. Since pδϕE, then f1(G)δϕE. Therefore, f1(G)ϕE. Let q be an element of f1(G)ϕE. Then, f(q) is an element of Gϕf(E). Since G is an arbitrary decision space containing f(p) in Y, f(p)δϕf(E).

The following theorem states the construction of various ϕcontinuous functions from one Trspace to another.

Theorem 2.3

Let (X,Tr1,δϕ), (Y,Tr2,δϕ), and (Z,Tr2,\allowbreakδϕ) be Trspaces.

  1. If a function f:XY takes all points of X into a single point y0 of Y, then f is ϕcontinuous.

  2. If E is a subspace of X, the inclusion function i:EX is ϕcontinuous.

  3. If the functions f:XY and g:YZ are ϕcontinuous, then the composition function gf:XZ is ϕcontinuous.

  4. If the function f:XY is ϕcontinuous, and if E is a subspace of X, then f|E:EY is ϕcontinuous.

  5. Let the function f:XY be ϕcontinuous. If Z is a Trspace that contains Y as a subspace, then the function h:XZ is ϕcontinuous.

Proof.

  1. Let pX and E be any subset of X with the property that pδϕE. We show that f(p)δϕf(E). Since pδϕE, there is a point y of {p}E such that y{p} ϕ E and f(y)f({p}) ϕ f(E). Since f is constant, there is y0 in Y such that f(x)=y0 for every xX. This implies that y0{y0} ϕ {y0}. Therefore, f(p)δϕf(E).

  2. Suppose that pE and AE such that pδϕA. Since i(a)=a for each aE, then pδϕA implies i(p)δϕi(A).

  3. Let x be any point in X and E be a subset of X such that xδϕE. We show that gf(x)δϕgf(E). Since xδϕE and f is ϕcontinuous, then f(x)δϕf(E). Similarly, the ϕcontinuity of g implies that g(f(x))δϕg(f(E)), as desired.

  4. Note that the restricted function f|E:EY is a composition of the function f:XY and the inclusion function i:EY, both of which are ϕcontinuous. By (iii), f|E is ϕcontinuous

  5. Since Y is a subspace of Z, the inclusion function i:YZ is ϕcontinuous (refer to Theorem 2.3(ii)). But the function h is the composition of the two ϕcontinuous functions, i and f. That is, h=if:XZ. By Theorem 2.3(iii), h is ϕcontinuous.

Since a representation in decision-making is necessary, it is important to consider the continuity of a function over only a small set, such that the continuity of a function on such a small set will mean the continuity on the whole space. The following theorem establishes the continuity of a function on a small set, which implies continuity on the whole set.

Theorem 2.4

Let (X,Tr1,δϕ) and (Y,Tr2,δϕ) be Trspaces and h:XY be a function. Suppose that {Gα:α{1,,n}} is a collection of decision spaces in X such that α=1nGαXX and that the restriction function h|Gα:GαY is ϕcontinuous for each index α. Then, the function h:XY is ϕcontinuous.

Proof.

Let p be any point and E be any subset of X with the property that pδϕE. We need to show that h(p)δϕh(E). Let V be a decision space in Y that contains h(p). Since h1(V) is a subset of X, then by elementary set theory, we have

(2.1) h1(V) =h1(V)  X.(2.1)

Since α=1nGαXX, EquationEquation (2.1) gives

(2.2) h1(V)α=1nGαXh1(V)X.(2.2)

The simplification of EquationEquation (2.2) gives

(2.3) α=1nh|Gα1(V)Xh1(V).(2.3)

Therefore, for h(p)V, we have α=1nh|Gα1({h(p)})Xp. This implies that h|Gα1({h(p)})Xp for some fixed index α{1,,n}. Since pδϕE, then h|Gα1({h(p)})δϕE. By the ϕcontinuity of h|Gα:GαY, we have {h(p)}\allowbreakδϕh|Gα(E) or just

(2.4) h(p)δϕh|Gα(E).(2.4)

Since h|Gα is well defined on Gα, the fact that EquationEquation (2.4) holds implies that either E is properly contained in Gα or E is equal to Gα. In either case, h|Gα(E)=(h|Gα)|E(E)=h|E(E)=h(E) (refer to the definition of restriction of a function), and thus EquationEquation (2.4) gives h(p)δϕh(E).□

Theorem 2.5

(Continuous Invariants)

Let f:XY be a ϕcontinuous function between Trspaces (X,Tr1,δϕ) and (Y,Tr2,δϕ) endowed with ϕproximities.

  1. If X is inclusive, then f(X) is inclusive in Y.

  2. If X is ϕconnected, then the image f(X) is ϕconnected in Y.

Proof.

  1. Suppose that f(X) is not inclusive. Then, there exists at least one element of f(X) that is unrepresented in Y. Let y0 be such a point of f(X). Then, y0δϕG for each decision space G in Y. Since y0f(X), there is an element x0 of X such that y0=f(x0). Thus, y0δϕG implies f(x0)δϕG. Since f is ϕcontinuous, f1(G) is a decision space in X that is administratively far from x0 for each decision space GY. That is, x0δϕf1(G). This is a contradiction because, by hypothesis, X is inclusive, and therefore x0 cannot be unrepresented. So, f(X) is inclusive.

  2. Suppose, on the contrary, that f(X) is ϕseparated. Let G1,G2 be a ϕseparation of f(X) such that G1G2Yf(X) and G1δϕG2. By ϕcontinuity of f, we have that f1(G1) and f1(G2) are decision spaces which form a ϕseparation of X. This is a contradiction because it is given that X is ϕconnected.

Theorem 2.6

The Trspaces (X,Tr1,δϕ) and (Y,Tr2,δϕ) are ϕhomeomorphic if there exists a bijective function f:XY from X onto Y with the property that EδϕF in X if and only if f(E)δϕf(F) in Y, for all subsets E and F of X.

Proof.

To prove this, it is only enough to show that the function f:XY and its inverse f1:YX are both ϕcontinuous. Since it is given that EδϕFf(E)δϕf(F) for all subsets E,F of X, then the forward implication EδϕFf(E)δϕf(F) implies that f is ϕcontinuous (definition 2.1). Similarly, the backward implication f(E)δϕf(F)EδϕF implies that f−1 is, by the same definition 2.1, ϕcontinuous too. Hence, the conclusion.

In general topology, we have two kinds of properties: hereditary and topological properties. A topological property or topological invariant is a property of a space such that any space homeomorphic to this space has that property (see Adams and Franzosa (Citation2008, pp. 140–151), Croom (Citation2008, pp. 118–119)). Hereditary property, as the name suggests, is the property of a space that is inherited by all of its subspaces. It is too obvious that the ϕconnectedness is a topological property. We will show in the next theorem that the property of inclusiveness is also a topological property.

Theorem 2.7

The inclusive property of Trspaces is a topological property.

Proof.

Let (X,Tr1,δϕ) be an inclusive space and (X,Tr2,δϕ) be a Trspace that is ϕhomeomorphic to X. Let f:XY be a ϕhomeomorphism and y be a point in Y. Since f is bijective, then f1(y) is a point in X. Since X is inclusive (see definition 1.6), there is a decision space G in X such that GXf1(y). By ϕhomeomorphism of f, f(G) is a decision space in Y and f(G)Yy. Therefore, Y is inclusive, and hence inclusiveness is a topological property.

By Theorem 1.16, it can be easily shown that the inclusive property is a hereditary property. Because if X is a Trspace and E is its nonempty subspace, then for any point xE, there is a decision space GX such that GXx (X is inclusive). Under the subspace topology, we have GϕEEx.

3. ϕdominating sets and ϕalliance

One of the properties of nonempty open sets in Trspace (X,Tr) is the existence of a point in each such set whose decision index is greater than all in that set. We call such a point a maximal element (see Corollary 1.13). In characterizing the Trspace in the international systems, it is important to consider some special kinds of sets of the same stance as maximal elements. These sets are dominating sets and alliances. Their importance is essential in formulating the connection between one nation and another. In this section, we will explore these kinds of sets.

Definition 3.1.

Let X=(X,Tr,δϕ) be a Trspace and E a subset of X. A set E is called a ϕdominating set of X if for every point p outside E and every situation sSX, there is a direct point q in E such that ρX(p,s)ρX(q,s).

Definition 3.1 clearly asserts that if E is a ϕdominating set, then every point in X − E is ϕnear to some points in E. This definition considers the nearness of points outside E to E but does not talk about the nearness of points within E itself. In order to eliminate the necessity of points within E being ϕfar from each other, we need to extend our definition of ϕdominance so that the ϕnearness to E must consider all points in the entire space X.

Definition 3.2.

A subset E of a space (X,Tr,δϕ) is called a total ϕdominating set of X if for every point (p,s)X×SX, there is a direct point q in E such that ρX(p,s)ρX(q,s).

From these two definitions, it follows that the total ϕdominating set is the ϕdominating set. However, the converse does not necessarily hold. The two definitions lead to the following result:

Theorem 3.3

Every Trspace (X,Tr) has a total ϕdominating set.

Proof.

Since every Trspace has a maximal decision space (Proposition 1.12), let G be such a space. We claim that G is a total ϕdominating set of X. By definition of G, we have ρX(H)ρX(G) for all HTr. Since every point in X is represented by a decision space, it follows that for each point xX, there is an element gG such that ρX{x}ρX{g}.

Apart from the maximal decision space being the total ϕdominating set of X, the entire space X itself is a total ϕdominating set. This can be easily verified. In space X, there can be many sets with the property of ϕdominance. It is always more difficult to work on larger sets than smaller ones. We seek to identify the minimal one, for which no other subset has this property. The following theorem establishes this result:

Theorem 3.4

A maximal decision space in a Trspace (X,Tr) is a minimal total ϕdominating set in X.

Proof.

Let G be a maximal decision space in a Trspace (X,Tr). By theorem 3.3, G is a total ϕdominating set. Our task that remains is to show that G is a minimal total ϕdominating set. If G contains only one direct element, then G is minimal because a single-point set contains no nonempty proper subsets. If G has more than one direct element, we show that it is still a minimal total ϕdominating set. Suppose, on the contrary, that G is not a minimal total ϕdominating set of X and let G be a total ϕdominating set of X that is properly contained in G. Suppose further that there is a situation s in the administrative section of X that needs to be resolved. Since G is a maximal decision space in X, there must be an element x directly contained in G that resolves the situation s. If x is directly contained in G but outside G and since x is maximal in the m-chain formed due to situation s, it follows that there is no element y in G such that ρX{x}ρX{y}. This contradicts the assertion that G is a total ϕdominating set of X. Similarly, if x was a direct member of G, there would be an element p directly contained in GG such that ρX{x}>ρX{p}. This would violate the maximality of G (see definition 1.11) and its property of total ϕdominance. Therefore, it follows that in either case, G does not exist, and the result follows immediately.□

In the international system, no nation exists in isolation. Chances are that each nation is necessarily involved in interactions with others. Interactions may come with ϕdomination. A ϕdomination of one nation by another may mean a ϕdominating nation is sitting within a ϕdominated nation by having too much influence in the decision-making processes. This kind of relationship brings us to the topological concept of embedding one space into another.

Definition 3.5.

An injective function f:XY between Trspaces (X,Tr1,δϕ) and (X,Tr2,δϕ) is called a ϕembedding of the space X into the space Y if X is ϕhomeomorphic to its image f(X) as the subspace of Y.

Before we state the necessary and sufficient conditions for one nation to sit within and dominate another, it is important to note that the decision index of the dominating nation is always greater than or equal to that of the nation under dominance. The following theorem establishes the criteria for this dominance.

Theorem 3.6

Let f:XY be a ϕembedding of the space (X,Tr1,δϕ) into the space (X,Tr2,δϕ). The space X totally ϕdominates Y if and only if f(X) totally ϕdominates Y.

Proof.

Suppose that X totally ϕdominates Y. We show that f(X) totally ϕdominates Y. This is equivalent to showing that for each point yY, there is a point f(x)f(X) for some unique point xX such that ρY{f(x)}ρY{y}. Let y be an arbitrary point of Y. Since X totally ϕdominates Y, there is a point x directly contained in some decision space G in X such that ρX{x}ρY{y}. Since X is ϕembedded in Y, then X is topologically identical to its image f(X) in Y under ϕproximity (Definition 3.5). Therefore, x is ϕproximally equivalent to f(x)Gf(X) (see definition 1.10), and so ρX{x}ρY{f(x)}. This implies that ρY{f(x)}ρY{y}. Conversely, suppose that f(X) totally ϕdominates Y and that X does not totally ϕdominate Y. Then, there is at least one point yY such that yδϕG for all GTr1. Equivalently, yδϕx for all xX. Since X is ϕembedded in Y, then xX is uniquely mapped to f(x)f(X) with the property that {x}{f(x)}. Because f(X) totally ϕdominates Y, then for all yY, ρYf(X)ρY{y}. The ϕhomeomorphism and the bijective property between X and f(X) give ρXXρY{y}. This implies that yδϕX, or yδϕxX{x}. Therefore, yδϕx for some xX. This contradicts the assumption that yδϕx for each xX.□

In Cantor’s set theory setting, it is known that if X,Y,Z are topological spaces, and if X is embeddable in Y, and Y is embeddable in Z, then X is embeddable in Z (Croom Citation2008, p. 127). We will show that this result holds for ϕproximal Trspaces. This result will be needed in extending the ϕdominance of one space into the other.

Theorem 3.7

Let (X,Tr1,δϕ),(Y,Tr2,δϕ), and (Z,Tr3,\allowbreakδϕ) be Trspaces equipped with ϕproximities δϕ, δϕ, and δϕ, respectively. If X is ϕembeddable in Y, and Y is ϕembeddable in Z, then X is ϕembeddable in Z.

Proof.

Let f:XY be a ϕembedding of X into Y and g:YZ be a ϕembedding of Y into Z. By definition 3.5, the restrictions f:Xf(X) and g:Yg(Y) on ranges of f and g are both ϕhomeomorphisms. We claim that a composition function gf:XZ is a ϕembedding of X into Z, and the restriction gf|g(f(X)):Xg(f(X)) defined by

(3.1) h=gf|g(f(X))=ig(f(X))1gif(X)f(3.1)

is a ϕhomeomorphism. In this EquationEquation (3.1), the functions ig(f(X)) and if(X) are inclusions. We show that h and h−1 are both ϕcontinuous. Since h=ig(f(X))1gif(X)f is a composition of ϕcontinuous functions, it is ϕcontinuous. Similarly, h1=f1if(X)1g1ig(f(X)) is also ϕcontinuous (see Theorem 2.3 and the fact that fʹ and gʹ are ϕhomeomorphisms). We next show that h and h−1 are inverses of each other.

hh1=ig(f(X))1gif(X)ff1if(X)1g1ig(f(X))=ig(f(X))1gif(X)(ff1)if(X)1g1ig(f(X))=ig(f(X))1gif(X)if(X)if(X)1g1ig(f(X))=ig(f(X))1gif(X)(if(X)if(X)1)g1ig(f(X))=ig(f(X))1gif(X)if(X)1g1ig(f(X))=ig(f(X))1gif(X)1g1ig(f(X))=ig(f(X))1gif(X)g1ig(f(X))=ig(f(X))1gg1ig(f(X))=ig(f(X))1(gg1)ig(f(X))=ig(f(X))1ig(Y)ig(f(X))=ig(f(X))1ig(f(X))=ig(f(X)),

and

h1h=f1if(X)1g1ig(f(X))ig(f(X))1gif(X)f=f1if(X)1g1(ig(f(X))ig(f(X))1)gif(X)f=f1if(X)1g1ig(f(X))1gif(X)f=f1if(X)1g1ig(f(X))gif(X)f=f1if(X)1g1gif(X)f=f1if(X)1(g1g)if(X)f=f1if(X)1iYif(X)f=f1if(X)1if(X)f=f1if(X)f=f1f=iX.

Therefore, h and h−1 are inverses of each other, and therefore the function h=ig(f(X))1gif(X)f:Xgf(X) is a ϕhomeomorphism. This implies that the function gf:XZ is a ϕembedding of X into Z.□

The result of the theorem 3.7 can be generalized to any finite number of topological spaces. Technically speaking, the homeomorphisms established between the embedded space and the subspace of the embedding space are constructed using the properties of composite functions and their bijective behaviour. Suppose that (Xi)1in is a finite sequence of topological spaces with the property that there is an embedding fi:XiXi+1 of Xi into Xi+1 for each i=1,2,,n1. Then, X1 is embeddable in Xn. The homeomorphism between X1 and the subspace of Xn is a restriction of a function fn1fn2f2f1:X1Xn on its range and it is given by Fn2:X1fn1(fn2(fn3((f3(f2(f1(X1))))))) where

(3.2) Fn2=ifn1fn2f2f1(X1)1fn1ifn2fn3f2f1(X1)1fn2ifn3f2f1(X1)1fn3ifn4f2f1(X1)1if2f1(X1)1f2if1(X1)f1,(n3),(3.2)

and the function fi:Xifi(Xi) is a homeomorphism between embedded space Xi and the subspace fi(Xi) of the embedding space Xi+1. EquationEquation (3.2), for instance, yields EquationEquation (3.1) when n = 3, with f1=f and f2=g. This result is very powerful, and with it at hand, we can picture the dominating m-chain that goes through m-chains of decision spaces in nations. Suppose that Xi totally ϕdominates Xi+1 for i=1,2,,n1 and fi:XiXi+1 is a ϕembedding of Xi into Xi+1, then by Theorem 3.6, fi(Xi) totally ϕdominates Xi+1. This further implies that fi(Xi) totally ϕdominates fi+1(Xi+1). Since X1 is ϕhomeomorphic to f1(X1), then the sequence Mfi=(fi(Xi):i=1,,n1) forms a total ϕdominating m-chain that goes straight (see definition 1.29) through the sequence (Xi:i=2,,n) and that is induced by ϕembeddings fi . There is another m-chain that goes strictly straight through Mfi and that is given by subspaces of fi(Xi) for each Xi which are in one-to-one correspondence to the space X1. Since X1f1(X1)f2f1(X1)f3f2f1(X1)fn2fn3f2f1(X1)fn1fn2f2f1(X1) and for each k1, fkfk1f2f1(X1)fk(Xk)Xk+1, then the chain M=(f1(X1),f2f1(X1),f3f2f1(X1),,fn2fn3f2f1(X1),fn1fn2f2f1(X1)) is an m-chain that goes strictly straight through Mfi if and only if fkfk1f2f1(X1) ϕ fk+1fk+2f2f1(X1) for each k. Because X1 is ϕhomeomorphic to f1(X1), and f1(X1) is ϕhomeomorphic to f2f1(X1) and so on, then by Theorem 1.30, all elements in M are pairwise ϕhomeomorphic, and, therefore, at any finite step k3, there is always a ϕhomeomorphism Fk2:X1fk1fk2f2f1(X1) (see EquationEquation (3.2)) that completes the diagram f1:X1f1(X1),if2f1(X1)1f2if1(X1):f1(X1)f2f1(X1),if3f2f1(X1)1f3if2f1(X1):f2f1(X1)f3f2f1(X1),,ifk1fk2f2f1(X1)1fk1ifk2fk3f2f1(X1):fk2fk3f2f1(X1)fk1fk2f2f1(X1).

It may happen that nations join forces for a common cause or situation. This may lead to the formation of an alliance. An alliance is generally thought of as a treaty or formal agreement between two or more parties made in order to unite for a common situation (see Haynes and Hedetniemi (Citation2001, p. 52), Snyder (Citation1990, p. 104)). This can be simply put as the union of allying spaces. This union will be assumed to be disjoint: a condition that eliminates the necessity of one point being directly contained in more than one Trspaces. To study an alliance topologically on the basis of ϕproximity, we need to develop the topology of such a disjoint union. Now, let X=i=1nXi be the union of an indexed family of the set {(Xi,Tri)} of Trspaces. Since each Xi is a subset of X, there is an inclusion function

fi:Xij=1nXj

defined by fi(x)=(x,i). The coordinate i in (x, i) identifies with the function fi. Our task here is to build an open set (or a decision space) in X from Xi. By the ϕcontinuity of fi (Theorem 2.3), a subset G of X is open in X if and only if fi1(G) is open in Xi for each index i. Since fi1(G)X, then fi1(G)=fi1(G)ϕXi=GϕXi for each i. This implies that a subset G of X is open in X if and only if the administrative intersection of G with each Xi is an element of Tri. We will now show, in the proposition 3.8, that the collection of sets of this kind in X constitutes a representative topology on X.

Proposition 3.8.

Let {(Xi,Tri)|i=1,2,,n} be an indexed family of distinct Trspaces and X=i=1nXi. Define the set T by T={GX|GϕXiTrifor all i}. Then, the collection T is a Trtopology on X.

Proof.

It is obvious that and the whole set X are in T, because if G= or G = X, then for each i, ϕXi=Tri or XϕXi=XiTri, respectively (Tri is a Trtopology on Xi). Likewise, if G1,G2,G3, are elements of T, then GjϕXiTri for each i. Since Tri is a topology, then jJGjϕXi=(jJGj)ϕXiTri. By definition, jJGjT. To show that T is closed under finite intersection, we suppose that G1,,GmT, then by definition, GkϕXiTri for every index i, 1km. This implies that ϕ,k=1m(GkϕXi)=(ϕ,k=1mGk)ϕXiTri. It follows by definition that ϕ,k=1mGkT.□

We are now in a position to define what we have been postponing: the ϕalliance.

Definition 3.9.

Let Y={(Xi,Tri):1in} be a collection of Trspaces and X={(Xj,Trj):1jm} be a subcollection of Y, where 2mn. Suppose that Si is a set of priority situations for each space Xi. Then, a ϕalliance of spaces (Xj,Trj) is a union X=j=1mXj of Trspaces such that there exists an element sj=1mSj and a decision space (G,ρXG,{s})X with the property that GδϕXj for each index j.

Definition 3.9 deduces that every space in a ϕalliance is ϕnear to other members of the ϕalliance. This ϕproximity between spaces forming an alliance suggests the following result: the result which states the ϕconnectedness of the ϕalliance.

Theorem 3.10

A ϕalliance is ϕconnected with respect to a common priority situation.

Before we provide the proof of Theorem 3.10, it is important to note that the ϕconnectedness of a ϕalliance is built on the essence of a set of common priority situation(s). These situations may be many in number; however, only one or a few situations among these many can cause the formation of a ϕalliance.

Proof.

Proof of Theorem 3.10

Let X=i=1nXi be a ϕalliance of Trspaces Xi,i=1,,n, and let si=1nSi be a common situation upon which the ϕalliance is made. Suppose that X is ϕseparated with respect to s. Then, there exist two disjoint nonempty decision spaces G1=(G1,ρXG1,{s}) and G2=(G2,ρXG2,{s}) such that G1G2Xi=1nXi and G1δϕG2 (refer Definition 1.22). Since X is a ϕalliance, there is a decision space GX for s such that GδϕXi for each index i. The ϕconnectedness of G,G1, and G2 (see Theorem 1.24 and Corollary 1.25) implies that G either lies entirely in G1 or in G2 but not in both. If G entirely lies in G1, then GδϕG2. Since G2 is nonempty, there are some spaces Xji=1nXi for some index j{1,2,,n} such that G2XXj. Because G2 is ϕfar from G, then all Xj’s represented by G2 will be ϕfar from G. This contradicts the fact that G is ϕnear to all Xi’s.

The definition of ϕalliance requires a common priority situation that brings together distinct Trspaces. It is obvious that there might be some other common situations, but not all can cause nations to ally. The ϕalliance formed in a single common situation is weakly ϕconnected. To have strong or complete ϕconnectedness, the ϕalliance must be ϕconnected for each priority situation common to all such spaces. Hence, the following theorem states the condition for an alliance to be strongly ϕconnected.

Theorem 3.11

The ϕalliance is ϕconnected if and only if it is ϕconnected with respect to each common priority situation.

Proof.

We know by Theorem 3.10 that any ϕalliance is ϕconnected with respect to a common priority situation. If the ϕalliance X=i=1nXi is ϕconnected, then X has no ϕseparation, and so for any common situation si=1nSi, X is ϕconnected. Conversely, suppose that X is ϕconnected with respect to every element smi=1nSi. Then, for each sm, there is a decision space GmX such that GmϕXi for each i. This further implies that Gmϕ(i=1nXi). Since each Xi is ϕconnected (Corollary 1.27), it follows by Theorem 1.26 that X=i=1nXi is ϕconnected.

With these results, we give one living example of a ϕalliance and show that such an alliance is ϕconnected.

Example 4.

Let us consider the political and military alliance, the North Atlantic Treaty Organization (NATO), the alliance formed primarily to keep and maintain peace and security in the North Atlantic area. The members of the alliance are resolved to unite their efforts for collective defence and the preservation of peace and security. Therefore, the members have constructed a number of principles to run the alliance. Among those principles, we only consider two principles, Articles 5 and 9 of the North Atlantic Treaty (see (NATO, Citation1949)). For the purpose of self-containment in this work, we briefly state them below:

Article 5

The Parties agree that an armed attack against one or more of them in Europe or North America shall be considered an attack against them all and consequently they agree that, if such an armed attack occurs, each of them, in exercise of the right of individual or collective self-defence recognised by Article 51 of the Charter of the United Nations, will assist the Party or Parties so attacked by taking forthwith, individually and in concert with the other Parties, such action as it deems necessary, including the use of armed force, to restore and maintain the security of the North Atlantic area.

Article 9

The Parties hereby establish a Council, on which each of them shall be represented, to consider matters concerning the implementation of this Treaty. The Council shall be so organised as to be able to meet promptly at any time. The Council shall set up such subsidiary bodies as may be necessary; in particular it shall establish immediately a defence committee which shall recommend measures for the implementation of Articles 3 and 5.

For Article 3 and other articles not stated here, refer to (NATO, Citation1949). We claim that this alliance is ϕalliance and we show that it is ϕconnected with respect to a priority situation (collective defence) upon which the alliance is formed. Before this, we first need to understand the embodiment of Articles 5 and 9. According to definition 1.2, Article 5 has two decision authority mappings:

  1. If one or more parties of the alliance are armed attacked, then this attack is considered an armed attack against all.

    Let s= one or more parties are armed attacked, and a= it is an attack against all. By definition 1.2,

    (3.3) a=PC(s).(3.3)

  2. If an armed attack occurs, then each member of the alliance will assist the party or parties attacked to restore and maintain the security of the North Atlantic area. Let b= each member of the alliance will assist the party or parties attacked to restore and maintain the security of the North Atlantic area, then the decision authority is

    (3.4) b=PC(a).(3.4)

    Combining EquationEquations (3.3) and (Equation3.4) gives

    (3.5) b=PC(PC(s)).(3.5)

    Therefore, by definition 1.2, EquationEquation (3.5) mathematically summarizes Article 5 as follows:

    (3.6) PC={(s,a),(a,b)}.(3.6)

Following the decision authority stipulated in EquationEquation (3.6), Article 9 establishes a decision space called a Council (also known as the North Atlantic Council), on which every member of the alliance is represented, and that which oversees the political and military process relating to security issues affecting the whole alliance by exercising the decision authority identified in Article 5. Let this council be denoted by C=(C,PC(C),{s}), and the alliance be denoted by N and defined as N=i=1nXi, where Xi is a member country (in our case, a Trspace) of N. Then, we can now show that this alliance is a ϕalliance and is ϕconnected with respect to s.

  • On N is a ϕalliance: This clearly follows consequently from Articles 5 and 6, and the definition 3.9. The ϕalliance forming situation is s, and the council C is a decision space that is proximally ϕnear to every member Xi of N.

  • On ϕconnectedness of N: This result follows from Theorems 1.26 and 3.10. Since each Xi is a Trspace, it is ϕconnected (Corollary 1.27). Therefore, N=i=1nXi is ϕconnected with respect to s because there is a nonempty decision space CN such that XiδϕC for each i.

Mayila et al. (Citation2023, p. 9) established the ϕconnectedness of a Trspace (X,Tr), and we have just shown that the ϕalliance is also ϕconnected. Now, let us recall the concept of a cut point or a cut set in definition 1.31. A connected space can be made disconnected if some points are removed from it. Since a Trspace is ϕconnected, we need to identify the points or sets whose removal or deletion from X produces a ϕseparated subspace.

Theorem 3.12

Let (X,Tr,δϕ) be a Trspace endowed with ϕproximity δϕ and let E be a nonempty subset of X. Moreover, suppose that intXϕEUXEU and intXϕ(XE)VXXEV, where U=EFrtXϕE and V=(XE)FrtXϕE. Then, a ϕboundary FrtXϕE is a cut set of X.

Proof.

We need to show that FrtXϕE ϕseparates X. It suffices by definition 1.22 to show that there exist two disjoint nonempty decision spaces G1 and G2 in X such that G1G2XXFrtXϕE and G1δϕG2. We claim that such decision spaces are intXϕEU and intXϕ(XE)V. Indeed, these sets are decision spaces in X. This follows from Lemma 1.7, and they are all nonempty because intXϕEUXEU and intXϕ(XE)VXXEV. To show that the sets intXϕEU and intXϕ(XE)V form a ϕseparation of XFrtXϕE, it is a basic condition, as Lemma 1.23 postulates, to show that neither of the two contains a ϕcontact point (see definition 1.15) of the other. Specifically, we will show that

  1. (intXϕEU) ϕ FrtXϕE=

  2. (intXϕ(XE)V) ϕ FrtXϕE=

  3. (intXϕEU) ϕ (intXϕ(XE)V)= and

  4. (intXϕEU)  (intXϕ(XE)V)XXFrtXϕE

Now

  1. Suppose, in contrast, that (intXϕEU)ϕFrtXϕE, and let x(intXϕEU)ϕFrtXϕE. Then, by definition 1.14, x(intXϕEU)FrtXϕE with a property that ρX{x}ρX(intXϕEU) and ρX{x}ρX(FrtXϕE). If xintXϕEU and Ux is any direct neighborhood of x, then UxϕFrtXϕE. This means that xFrtXϕE. This is impossible because xFrtXϕE would mean xEFrtXϕE=U: the elements which are removed from intXϕE. So, xintXϕEU. If xFrtXϕE, we have that either xEFrtXϕE or x(XE)FrtXϕE. Suppose xEFrtXϕE, then xδϕ(intXϕEEFrtXϕE), contradicting the fact that ρX{x}ρX(intXϕEU). Therefore, x cannot be in EFrtXϕE. Similarly, if x(XE)FrtXϕE=V, x would be ϕfar from intXϕ(XE)V. This also contradicts the fact that xδϕ(intXϕ(XE)V). Therefore, x(XE)FrtXϕE. Since in either case there is no point in X that is ϕnear to both intXϕEU and FrtXϕE, it follows that (intXϕEU)δϕ(intXϕ(XE)V).

  2. On (intXϕ(XE)V)ϕFrtXϕE=: This result easily follows from (i) above because FrtXϕE=FrtXϕ(XE). By definition,

    (3.7) FrtXϕE=clXϕEclXϕ(XE).(3.7)

    Replacing E by X − E, EquationEquation (3.7) gives FrtXϕ(XE) =clXϕ(XE)  clXϕ(E), and the result follows immediately. Therefore,

    (3.8) (intXϕ(XE)V) ϕ FrtXϕE=(intXϕ(XE)V) ϕ FrtXϕ(XE)=.(3.8)

  3. On (intXϕEU) ϕ (intXϕ(XE)V)=: This is very obvious. There is no point in X that is ϕnear to both sets intXϕEU and intXϕ(XE)V, because if they were, such points would be contained either in U or V: the set of deleted points.

  4. On (intXϕEU)  (intXϕ(XE)V)XXFrtXϕE: Since it is given that (intXϕEU)XEU and intXϕ(XE)V)XXEV, then

    (intXϕEU)(intXϕ(XE)V)X(EU)(XEV)=(EEFrtXϕE)(XE(XE)FrtXϕE)=[E(Ec(FrtXϕE)c)][(XE)((XE)c(FrtXϕE)c)]=[E(FrtXϕE)c)][(XE)(FrtXϕE)c)]=[E(XE)](FrtXϕE)c=X(FrtXϕE)c=XFrtXϕE.

Since (intXϕEU)  (intXϕ(XE)V)XXFrtXϕE and (intXϕEU)δϕ(intXϕ(XE)V), then the spaces (intXϕEU) and (intXϕ(XE)V) form a ϕseparation of XFrtXϕE, and thus FrtXϕE is a cut set of X.

The result of Theorem 3.12 applies to a number of cases. In an m-chain M={G1,Gn} (see definition 1.28), it can be easily shown that a set GkϕGk+1 is a cut set of M={G1,G2,,Gk,Gk+1,,Gn1,Gn} for all k. This is because every point of GkϕGk+1 is proximally ϕnear to both Gk and Gk+1. The omission of such points results in holes in the links Gk and Gk+1. This ϕseparates the chain and produces two ϕseparated components: C1={G1,G2,,Gk1,GkGkϕGk+1} and C2={Gk+1GkϕGk+1,Gk+2,,Gn1,Gn}. In the case of ϕalliance, the omission of a decision space in the alliance that is ϕnear to all member spaces results in the ϕdisconnectedness of such a ϕalliance.

4. Conclusion

In this paper, we have used ϕproximity to define continuous functions on Trspaces. The ϕcontinuity so defined preserves the ϕnearness between individual points or between a point and a set. We then presented various results concerning ϕcontinuous functions on Trspaces. This includes the fact that ϕcontinuous functions preserve ϕconnectedness and the inclusive property of Trspaces. We then studied the special sets in an international system: ϕdominating sets and ϕalliance. Using a concept of ϕproximity and ϕcontinuity on Trspaces, we showed that a space X totally ϕdominates space Y if and only if there exists a ϕembedding f:XY such that the image f(X) of X under f totally ϕdominates Y. This result is very important as it states when and how one nation sits within the other and exercises its influence, decision authority, and dominance. We concluded our study by defining what we called a ϕalliance on Trspaces, developing a Trtopology on it, establishing its ϕconnectedness, and finally identifying the cut sets (or cut points) in a ϕconnected space X, whose removal ϕseparates the remaining subspaces.

Acknowledgements

We acknowledge the contribution from every single individual, including our fellow colleagues from our respective departments, who in one way or another contributed to the development and improvement of this work.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This research did not receive a grant from any funding agency.

References

  • Adams, C., & Franzosa, R. (2008). Introduction to Topology: Pure and applied. Pearson Education, Inc.
  • Alexandroff, P. (1961). Elementary concepts of Topology. A translation of (Alexandroff, 1932) by A.E. Farley. Dover Publications, Inc.
  • Baligh, H. H. (1990). Decision rule theory and its use in the analysis of organization’s performance. Organization Science, 1(4), 360–374. https://doi.org/10.1287/orsc.1.4.360
  • Bossert, T. (1998). Analyzing the decentralization of health systems in developing countries: Decision space, innovation and performance. Social Science & Medicine, 47(10), 1513–1527. https://doi.org/10.1016/S0277-9536(98)00234-2
  • Croom, F. H. (2008). Principles of Topology. Cengage Learning.
  • Haynes, T. W., & Hedetniemi, S. T. (2001). Alliances and related domination parameters. In T. W. Haynes, S. T. Hedetniemi, & M. A. Henning (Eds.), Structures of domination in graphs. Developments in Mathematics (Vol. 66). Springer. https://doi.org/10.1007/978-3-030-58892-2
  • Hocking, J. G., & Young, G. S. (1961). Topology. Addison-Wesley Publishing Company, Inc. U.S.A.
  • Kinsey, L. C. (1993). Topology of surfaces. Springer-Verlag.
  • Latecki, L., & Prokop, F. (1995). Semi-proximity continuous functions in digital images. Pattern Recognition Letters, 16(11), 1175–1187. https://doi.org/10.1016/0167-8655(95)00069-S
  • Mayila, S., Mpimbo, M., Rugeihyamu, S., & Liu, L. (2023). On a nation as a topological space. Research in Mathematics, 10(1), 2187020. https://doi.org/10.1080/27684830.2023.2187020
  • Munkres, J. R. (2000). Topology (Second Edition). Prentice Hall, Inc.
  • Naimpally, S., & Peters, J. (2013). TOPOLOGY with APPLICATIONS. Topological spaces via Near and Far. World Scientific Publishing Co. Pte. Ltd. Singapore.
  • NATO. (1949). The North Atlantic treaty. NATO.
  • Singh, T. B. (2013). Elements of Topology. CRC Press, Taylor & Francis Group.
  • Snyder, G. H. (1990). Alliance theory: A neorealist first cut. Journal of International Affairs, 44(1), 103–123.