() @ Appl. Gen. Topol. 16, no. 1(2015), 53-64doi:10.4995/agt.2015.3141 © AGT, UPV, 2015 F-nodec spaces Lobna Dridi a,∗, Abdelwaheb Mhemdi b and Tarek Turki c a Department of Mathematics, Tunis Preparatory Engineering Institute. University of Tunis. 1089 Tunis, Tunisia. (lobna dridi 2006@yahoo.fr) b Higher Institute of applied sciences and technologies of Gafsa, bp 116, campus universitaire, 2112-Sidi Ahmed Zarroug Gafsa, Tunisia. (mhemdiabd@gmail.com) c Department of Mathematics, Faculty of Sciences of Tunis. University Tunis-El Manar. “Campus Universitaire” 2092 Tunis, Tunisia. (tarek turki@gmail.com) Abstract Following Van Douwen, a topological space is said to be nodec if it satisfies one of the following equivalent conditions: (i) every nowhere dense subset of X, is closed; (ii) every nowhere dense subset of X, is closed discrete; (iii) every subset containing a dense open subset is open. This paper deals with a characterization of topological spaces X such that F(X) is a nodec space for some covariant functor F from the category Top to itself. T0, ρ and FH functors are completely studied. Secondly, we characterize maps f given by a flow (X, f) in the category Set such that (X, P(f)) is nodec (resp., T0-nodec), where P(f) is a topology on X whose closed sets are precisely f-invariant sets. 2010 MSC: 54B30; 54D10; 54G12; 46M15. Keywords: Categories; functors; Nodec spaces; primal Space. 1. Introduction Recall that a topological space X is called submaximal if every dense subset of X is open. In [8] it was shown that, in a submaximal space without isolated points, every nowhere dense subset (that is the interior of its closure is empty), ∗Corresponding author. Received 25 June 2014 – Accepted 4 December 2014 http://dx.doi.org/10.4995/agt.2015.3141 L. Dridi, A. Mhemdi and T. Turki is closed and discrete. Hence, a space satisfying the later property is called nodec. Different equivalent conditions for a space to be submaximal are given in [1] and the ones for a space to be nodec in [14], [12] and [4]. Theorem 1.1. [4, Theorem 2.5] The following statements are equivalent: (1) X is nodec; (2) Each nowhere dense subset of X, is closed. (3) For each A ⊆ X, if ◦ A ⊆ A, then A is closed. (4) For each A ⊆ X, if A ⊆ ◦ ◦ A, then A is open. (5) For each A ⊆ X, A\A ⊆ ◦ A. (6) For each A ⊆ X, A = A ∪ ◦ A. (7) For each A ⊆ X, ◦ A = A ∩ ◦ ◦ A. On the other hand, the theory of category and functors play an enigmatic role in topology, specially the notion of reflective subcategory. So it is of importance to recall the standard notion of reflective subcategory A of B that is, a full subcategory such that the embedding A −→ B has a left adjoint F : B −→ A (called reflection). Further, recall that for all i = 0, 1, 2, 3, 3.1 2 the subcategory Topi of Ti-spaces is reflective in Top, the category of all topological spaces. T3 1 2 is also called the Tychonoff-reflection and will be denoted by ρ. In this paper the functors T0, ρ and FH (the functionally Hausdorff reflection) are studied. Some authors (see [3],[6], [11]) are interested in separation axioms using the theory of categories and functors as follows. Definition 1.2. Let C be a category and F, G two (covariant) functors from C to itself. (1) An object X of C is said to be a T(F,G)-object if G(F(X)) is isomorphic with F(X). (2) Let P be a topological property on the objects of C. An object X of C is said to be a T(F,P )-object if F(X) satisfies the property P . Consequently, some new separation axioms T(0,ρ), T(0,F H), T(ρ,F H) are in- troduced and characterized. Recently, in [2] a characterization of topological spaces X such that their compactification noted K(X) is a nodec space, is given. In [5], L. Dridi et al characterized topological spaces X such that F(X) is a submaximal space, for a given covariant functor F. The first section of this paper is devoted to the characterization of T0-nodec space. Second section studies the same problem using the functor ρ (resp., FH). © AGT, UPV, 2015 Appl. Gen. Topol. 16, no. 1 54 F -nodec spaces Finally, in the third section we are interested in the relation between nodec spaces and primal spaces. Some important results are given. 2. T0-nodec spaces First let us recall the T0-reflection of a topolgical space. Let X be a topo- logical space. We define the binary relation ∼ on X by x ∼ y if and only if {x} = {y}. Then ∼ is an equivalence relation on X and the resulting quotient space T0(X) := X/ ∼ is the T0-reflection of X. The canonical surjection µX : X −→ T0(X) is a quasihomeomorphism. ( A continuous map q : X −→ Y is said to be a quasihomeomorphism if U 7−→ q−1(U) (resp., C 7−→ q−1(C) ) defines a bijection O(Y ) −→ O(X) (resp., F(Y ) −→ F(X)), where O(X) (resp., F(X)) is the collection of all open sets (resp;, closed sets) of X )[7]. Before giving the main result of this section let us introduce some definitions, notations and remarks. Notations 2.1. [5, Notations 2.2] Let X be a topological space, a ∈ X and A ⊆ X. We denote by: (1) d0(a) := {x ∈ X : {x} = {a}} (2) d0(A) = ∪[d0(a); a ∈ A]. Remarks 2.2. [5, Remarks 2.3] Let X be a topological space and A be a subset of X. In [5] the following remarks are given. (i) d0(A) = µ −1 X (µX(A)). (ii) d0(d0(A)) = d0(A). (iii) A ⊆ d0(A) ⊆ A and consequently d0(A) = A. (iv) In particular if A is open (resp., closed ), then d0(A) = A. Definition 2.3. Let X be a topological space. X is called a T0-nodec space if its T0-reflection is a nodec space. Now we are in a position to give the characterization of T0-nodec space. Theorem 2.4. Let X be a topological space. Then the following statements are equivalent: (1) X is a T0-nodec space; (2) For any nowhere dense subset A of X, d0(A) is closed. (3) ∀A ⊆ X; if ◦ A ⊆ d0(A) =⇒ d0(A) = A. (4) ∀A ⊆ X; A\d0(A) ⊆ ◦ A. (5) ∀A ⊆ X; A = d0(A) ∪ ◦ A. Proof. (1) =⇒ (2) Let A be a nowhere dense subset of X. Then µX(A) ⊆ µX(A), since µX is a closed map. Suppose that there exist an open set U of T0(X) such that U ⊂ µX(A). So µ −1 X (U) is an open set and µ−1 X (U) ⊆ © AGT, UPV, 2015 Appl. Gen. Topol. 16, no. 1 55 L. Dridi, A. Mhemdi and T. Turki µ−1 X (µX(A)) ⊆ µ −1 X (µX(A)) = d0(A) = A, which contradict the fact that A is a nowhere dense subset of X. Then µX(A) is a nowhere dense subset of T0(X). Since the later is nodec, µX(A) is closed. So µ −1 X (µX(A)) = d0(A) is closed. (2) =⇒ (1) Let S be a nowhere dense subset of T0(X) and A = µ −1 X (S). Then d0(A) = A. Suppose that there exists x ∈ ◦ A. Thus there exists an open set U of X such that x ∈ U ⊂ A. So µX(U) is an open set containing µX(x), since µX is open. Moreover, µX(U) ⊆ µX(A) = µX(µ −1 X (S)) ⊆ µX(µ −1 X (S)) = S, by [5, Lemma 2.16]. Therefore µX(x) ∈ ◦ S, which contradicts the fact that S is a nowhere dense subset of T0(X). Then ◦ A = ∅ and consequently A is a nowhere dense subset of X. Thus d0(A) = A is closed, by hypothesis. So S is a closed set of T0(X). (2) =⇒ (4) Let A be a subset of X. Since X is T0-nodec, µX(A)\µX(A) ⊂ ◦ µX(A), by Theorem 1.1. Then µ −1 X (µX(A)\µX(A)) = µ −1 X (µX(A))\µ −1 X (µX(A)) = d0(A)\d0(A) = A\d0(A) ⊆ µ −1 X ( ◦ µX(A)) = ◦ µ−1 X (µX(A)) = ◦ d0(A) = ◦ A. (4) =⇒ (5) Let A be a subset of X. As ◦ A ⊆ A and d0(A) ⊆ A, it is clear that d0(A) ∪ ◦ A ⊂ A. Conversely, A = d0(A) ∪ (A\d0(A)) ⊆ d0(A) ∪ ◦ A, by (4). Thus, the equality holds. (5) =⇒ (2) Let A be a nowhere dense subset of X. Then A = d0(A) ∪ ◦ A = d0(A). (2) =⇒ (3) Straightforward. (3) =⇒ (2) If A is a nowhere dense subset of X, then ◦ A = ∅ ⊆ d0(A). Thus d0(A) = A, by (3). � Remark 2.5. Clearly every nodec space is a T0-nodec space. The converse does not hold: Indeed, given a set X = {0, 1, 2} equipped with the topology τ = {∅, {2}, X}. We can easily see that T0(X) is a nodec space. However {0} is a nowhere dense subset of X and not closed. 3. ρ-nodec spaces and F H-nodec spaces Let X be a topological space, F a subset of X and x ∈ X. x and F are said to be completely separated if there exists a continuous map f : X −→ R such that f(x) = 0 and f(F) = {1}. Now, two distinct points x and y in X are called completely separated if x and {y} are completely separated. A space X is said to be completely regular if every closed subset F of X is completely separated from any point x not in F . Recall that a topological © AGT, UPV, 2015 Appl. Gen. Topol. 16, no. 1 56 F -nodec spaces space X is called a T1-space if each singleton of X is closed. A completely regular T1-space is called a Tychonoff space [13]. A functionally Hausdorff space is a topological space in which any two distinct points of this space are completely separated. Remark here that a Tychonoff space is a functionally Hausdorff space and consequently a Hausdorff space (T2-space). Now, for a given topological space X, we define the equivalence relation ∼ on X by x ∼ y if and only if f(x) = f(y) for all f ∈ C(X) (where C(X) design the family of all continuous maps from X to R). Let us denote by X/ ∼ the set of equivalence classes and let ρX : X −→ X/ ∼ be the canonical surjection map assigning to each point of X its equivalence class. Since every f in C(X) is constant on each equivalence class, we can define ρ(f) : X/ ∼−→ R by ρ(f)(ρX(x)) = f(x). One may illustrate this situation by the following commutative diagram. X ▽ ρX // X/ ∼ ρ(f) }}③ ③ ③③ ③ ③③ ③ R �� f ❄ ❄ ❄ ❄ ❄ ❄ ❄ ❄ Now, equip X/ ∼ with the topology whose closed sets are of the form ∩[ρ(fα) −1(Fα) : α ∈ I], where fα : X −→ R (resp., Fα) is a continuous map (resp., a closed subset of R). It is well known that, with this topology, X/ ∼ is a Tychonoff space (see for instance [15]) and its denoted by ρ(X). The construction of ρ(X) satisfies some categorical properties: For each Tychonoff space Y and each continuous map f : X −→ Y , there exists a unique continuous map f̃ : ρ(X) −→ Y such that f̃ ◦ ρX = f. We will say that ρ(X) is the ρ-reflection (or Tychonoff-reflection) of X. From the above properties, it is clear that ρ is a covariant functor from the category of topological spaces Top into the full subcategory Tych of Top whose objects are Tychonoff spaces. On the other hand the quotient space X/ ∼ which is denoted by FH(X) is a functionally Hausdorff space. The construction FH(X) satisfies some categorical properties: For each functionally Hausdorff space Y and each continuous map f : X −→ Y , there exists a unique continuous map f̃ : FH(X) −→ Y such that f̃◦ρX = f. We will say that FH(X) is the functionally Hausdorff-reflection of X (or the FH-reflection of X). Consequently, it is clear that FH is a covariant functor from the category of topological spaces Top into the full subcategory FunHaus of Top whose objects are functionally Hausdorff spaces. We need to introduce and recall some definitions, notations and results. Notations 3.1. [5, Notation 3.1] Let X be a topological space, a ∈ X and A a subset of X. We denote by: (1) dρ(a) := ∩[f −1(f({a})) : f ∈ C(X)]. © AGT, UPV, 2015 Appl. Gen. Topol. 16, no. 1 57 L. Dridi, A. Mhemdi and T. Turki (2) dρ(A) := ∪[dρ(a) : a ∈ A]. The following results are given in [5]. Proposition 3.2. [5, Proposition 3.2] Let X be a topological space, a ∈ X and A a subset of X. Then: (1) dρ(A) = ρ −1 X (ρX(A)). (2) dρ(a) is a closed subset of X. (3) A ⊆ dρ(A) ⊆ ∩[f −1(f(A)) : f ∈ C(X)]. (4) ∀f ∈ C(X), f(A) = f(dρ(A)). Now, we introduce the following definition. Definition 3.3. Let X be a topological space. X is called a ρ-nodec (resp., FH-nodec ) space if its ρ-reflection (resp., FH-reflection ) is a nodec space. In order to give a characterization of ρ-nodec spaces, we need to recall some elementary proporties which characterize Tychonoff spaces in terms of zero-set (resp., cozero-sets). Consider a topological space X and A ⊆ X. A is called a zero-set if there exists f ∈ C(X) such that A = f−1({0}). The complement of a zero-set is called a cozero-set. According to [15, Proposition 1.7], a space is Tychonoff if and only if the family of zero-sets of the space is a base for the closed sets (equivalently, the family of cozero-sets of the space is a base for the open sets). In [5] it is showen that a closed (resp., open) subset of ρ(X) is of the form ∩[ρ(f)−1({0}) : f ∈ H] (resp., ∪[ρ(f)−1(R⋆) : f ∈ H]) , where H is a collection of continuous maps from X to R. Recall that a subset V of a topological space X is called a functionally open subset of X (for short F-open ) if and only if dρ(V ) is open in X (see [5]). Now in order to characterize ρ-nodec spaces and FH-nodec spaces, we introduce the following definitions: Definition 3.4. Let X be a topological space and V a set of X. V is called a Functionally nowhere dense subset of X (for short F-nowhere dense ) if and only if for any F-open subset W of X, dρ(W) * dρ(V ). Definition 3.5. Let X be a topological space and A a nonempty subset of X. A is said to be a ρ-nowhere dense subset of X if for any nonzero continuous map f from X to R there exists a continuous map g from X to R satisfying fg 6= 0 and g(A) = {0}. Remark 3.6. V is an F-open set of X if and only if ρX(V ) is an open set of FH(X). Proposition 3.7. Let X be a topological space and A a subset of X. Then the following statements are equivalent: (i) A is a ρ-nowhere dense subset of X; (ii) ρX(A) is a nowhere dense subset of ρ(X). © AGT, UPV, 2015 Appl. Gen. Topol. 16, no. 1 58 F -nodec spaces Proof. (i) =⇒ (ii) Let A be a ρ-nowhere dense subset of X. According to the universal property of ρ(X), each continuous map g : ρ(X) −→ R may be written as g = ρ(f) with f = g◦ρX. Now suppose that there exists a nonzero continuous map f from X to R such that ρ(f)−1(R⋆) ⊂ ρX(A). Then there exists a continuous map g such that fg 6= 0 and g(A) = {0}. Thus ρX(A) ⊂ ρ(g) −1{0} and consequently ρX(A) ⊂ ρ(g) −1{0}. So ρ(f)−1(R⋆) ⊂ ρX(A) ⊂ ρ(g) −1{0}. Therefore ρ(f)−1(R⋆)∩ρ(g)−1(R⋆) = ∅ which contradict the fact that fg 6= 0. (ii) =⇒ (i) Let A be a subset of X such that ρX(A) is a nowhere dense subset of ρ(X). Then for any nonzero continuous map f from X to R, ρ(f)−1(R⋆) * ρX(A). Thus there exists x ∈ X such that ρ(x) ∈ ρ(f) −1(R⋆) and ρ(x) /∈ ρX(A). Therefore there exists a nonzero continuous map g such that ρ(x) ∈ ρ(g)−1(R⋆) and ρ(g)−1(R⋆) ∩ ρX(A) = ∅. So g(A) = {0}. Hence A is a ρ-nowhere dense subset of X. � Now, we are in a position to give the characterization of ρ-nodec spaces. Theorem 3.8. Let X be a topological space. Then the following statements are equivalent: (i) X is a ρ-nodec space; (ii) For any ρ-nowhere dense set A of X, we have dρ(A) is an intersection of zero sets of X. Proof. (i) =⇒ (ii) Let A be a ρ-nowhere dense subset of X. According to the Proposition 3.7 ρX(A) is a nowhere dense subset of ρ(X). Since X is a ρ-nodec space, then ρX(A) is a closed set of ρ(X) and thus ρX(A) = ⋂ [ρ(f)−1{0} : f ∈ H] (where H is a subfamily of C(X) ). So that ρ−1 X (ρX(A)) = ⋂ [ρ−1 X (ρ(f)−1{0}) : f ∈ H]. Therefore dρ(A) = ⋂ [f−1{0} : f ∈ H] is an intersection of zero sets of X. (ii) =⇒ (i) Conversely, let A a subset of X such that ρX(A) is a nowhere dense subset of ρ(X). Then, by Proposition 3.7, A is a ρ-nowhere dense subset of X and consequently dρ(A) is an intersection of zero sets of X. Hence there exists a subfamily {fi : i ∈ I} of C(X) satisfying ρ −1 X (ρX(A)) = ⋂ [f−1i {0} : i ∈ I]. Then: ρX(A) = ρX( ⋂ [f−1i {0} : i ∈ I]) = ρX( ⋂ [ρ−1X (ρ(fi) −1{0}) : i ∈ I]) = ρX(ρ −1 X ( ⋂ [ρ(fi) −1{0} : i ∈ I])) = ⋂ [ρ(fi) −1{0} : i ∈ I] Finally, ρX(A) is a closed subset of ρ(X). � Theorem 3.9. Let X be a topological space. Then the following statements are equivalent: (i) X is FH-nodec. (ii) For any F-nowhere dense subset A of X, dρ(A) is closed. Proof. (i) =⇒ (ii) Let A be an F-nowhere dense subset of X. First let us show that ρX (A) is a nowhere dense subset of FH(X). © AGT, UPV, 2015 Appl. Gen. Topol. 16, no. 1 59 L. Dridi, A. Mhemdi and T. Turki Suppose that ◦ ρX (A) 6= ∅. Then there exist ρX (U) an open subset of FH(X) such that: ∅ 6= ρX (U) ⊂ ρX (A). Thus ρ −1 X (ρX (U)) ⊂ ρ −1 X (ρX (A)) ⊂ ρ−1X (ρX (A)) = dρ (A). So dρ (U) = ρ −1 X (ρX (U)) is open in X that is U is an F- open subset and dρ (U) ⊂ dρ (A), which contradict the fact that A is F-nowhere dense. Hence ρX (A) is a nowhere dense subset of FH(X) and consequently a closed set of FH(X). Therefore dρ(A) = ρ −1 X (ρX (A)) is closed. (ii) =⇒ (i) Let ρX (A) be a nowhere dense subset of FH (X), where A is a subset of X. We prove that A is F-nowhere dense in X. Suppose that there exists V an F-open subset of X, such that dρ (V ) ⊂ dρ (A). So ρX (dρ (V )) ⊂ ρX(dρ (A)) ⊂ ρX (dρ (A)). Thus ρX (V ) ⊂ ρX (A). V is an F-open set of X, that is dρ (V ) an open subset of X, and thus ρX (V ) is open in FH (X). Then ◦ ρX (A) 6= ∅, which contradict the fact that ρX (A) is nowhere dense in FH (X). Hence A is F-nowhere dense in X. By (ii), dρ (A) is closed in X and consequently ρX (A) is closed in FH (X). Therefore FH (X) is nodec. � 4. Alexandroff Topology According to Kennisson, a flow in a category C is a couple (X, f), where X is an object of C and f : X −→ X is a morphism, called the iterator (see [9] and [10]). Recall that the topology P(f) defined on a flow (X, f) of the category Set, is a topology such that closed sets are exactly those A which are f-invariant (i.e., f(A) ⊆ A) and consequently open sets are those which are f−1-invariant. It is clearly seen that for any subset A of X, the topological closure A is exactly ∪[fn(A) : n ∈ N]. In particular for any point x ∈ X, {x} = Of (x) = {fn(x) : n ∈ N} called the orbit of x by f. One can see easily that the family {Vf (x) : x ∈ X} is a basis of open sets of P(f), where Vf (x) := {y ∈ X : fn(y) = x, for some integer n}. Clearly, P(f) is an Alexandroff topology on X. An element x of X is said to be a periodic point if fn(x) = x for some positive integer n. A characterization of maps f such that (X, P(f)) is nodec, is given by the following result. Proposition 4.1. Let (X, f) be a flow in Set. Then the following statements are equivalent. (i) (X, P(f)) is a nodec space; (ii) ∀ x ∈ X we have x is either a periodic point or f(x) is a fixed point. Proof. (i) =⇒ (ii). Let x ∈ X. If x is not a periodic point, then x /∈ {f(x)}. Suppose that there exists y ∈ ◦ {f(x)}, then Vf (y) ⊆ ◦ {f(x)} ⊆ {f(x)}. Hence y = fn(f(x)) = fn+1(x) for some integer n so that x ∈ Vf (y) ⊆ {f(x)}, which © AGT, UPV, 2015 Appl. Gen. Topol. 16, no. 1 60 F -nodec spaces contradict the fact that x is not a periodic point. Thus ◦ {f(x)} = ∅. Since (X, P(f)) is a nodec space, we get {f(x)} is closed and consequently f(x) is a fixed point. (ii) =⇒ (i). Let A be a subset of X such that ◦ A = ∅ and x ∈ A. Then we have f−1({x}) * {x} ⊂ A, otherwise {x} is an open set. Let y ∈ f−1({x})\{x}, it is clear that y is not a periodic point. Indeed, if y is a periodic point, then {y} = {f(y)}. But f(y) ∈ {x}, so we get y ∈ {x}, a contradiction. Thus f(y) is a fixed point. Since ◦ A = ∅, then we have {x} is not open which means that f−1({x}) \ {x} 6= ∅. Let z ∈ f−1({x}) \ {x}, it is clear that z is not a periodic point. Hence f(z) = x is a fixed point and consequently {x} is a closed subset. Therefore A is a closed set. � Example 4.2. Consider the map f: N −→ N n 7−→ n + 1 where N is the set of all natural numbers including 0. Let p be a positive integer and n ∈ N, then we have fp(n) = n + p. Hence n is not a periodic point and f(n) is not a fixed point for every n ∈ N. Now, consider the topological space (N, P(f)) and set A = 2N + 1. Since every open subset of (N, P(f)) must contain 0, then ◦ A = ∅. However, A is not closed (A = N \ {0}). Remark 4.3. Let (X, f) be a flow in Set, we equip X with the topology P(f). Then for every a ∈ X we have d0(a) is closed if and only if a is a periodic point. Indeed, suppose that d0(a) is a closed subset of (X, P(f)), that is d0(a) is a f-invariant set. On the other hand, we know that a ∈ d0(a) then f(a) ∈ f(d0(a)) ⊆ d0(a) and consequently {a} = {f(a)}. Hence a ∈ {f(a)} which implies that a = fn(f(a)) = fn+1(a) for some n in N. Therefore a is a periodic point. For the converse, suppose that a is a periodic point which means that fn(a) = a for some positive integer n. Let x ∈ d0(a), then {x} = {a} and so x ∈ {a}. Hence d0(a) ⊆ {a}. For the reverse inclusion, let x ∈ {a} then x = fm(a) for some m in N. One can easily see that fpn(a) = a for every p ∈ N, thus we get a = fnm(a) = fnm−m+m(a) = fnm−m(fm(a)) = fnm−m(x). It follows that a ∈ {x} and so {x} = {a}. Therefore x ∈ d0(a). Finally d0(a) = {a} is closed. It is easily to see that every submaximal space (resp., T0-submaximal space) is a nodec space (resp., T0-Nodec space). The converse does not hold. as showen in the following examples. Example 4.4. given a set X = {a, b, c} equipped with the trivial topology. Clearly X is a nodec space which is not T0 and consequently not submaximal. © AGT, UPV, 2015 Appl. Gen. Topol. 16, no. 1 61 L. Dridi, A. Mhemdi and T. Turki Example 4.5. Let L be an infinite set, a, b /∈ L and X = L ∪ {a, b}. We equip X with the topology whose open sets are ∅, X, L, co-finite subset of L and co- finite subset of X containing {a, b}. Then {a} = {b} and T0(X) = L∪{µX(a)}. Thus T0(X) is nodec. Indeed, it is straightforward to see that the nowhere dense sets are finite subsets of T0(X), which are closed. However, X is not T0-submaximal, since for an infinite subset of L such that L \ S is infinite, we have S \ S is not closed. Remark 4.6. A primal nodec space is not always a submaximal space as shows the following example. Let X = {a, b, c} and f : X −→ X the map defined by:    f(a) = b f(b) = c f(c) = a. It is clear that (X, P(f)) is nodec but not submaximal, since f2 6= f. The following Proposition shows that for a given primal space (X, P(f)), there is an equivalence between T0-submaximal and T0-nodec. Proposition 4.7. Let (X, f) be a flow in Set. Then the following statements are equivalent. (i) (X, P(f)) is a T0-submaximal space; (ii) (X, P(f)) is a T0-nodec space; (iii) f(x) is a periodic point for every x ∈ X. Proof. (i) =⇒ (ii). Straightforward. (ii) =⇒ (iii). Let x ∈ X. Suppose that f(x) is not a periodic point, then x is not a periodic point and x 6= f(x). For the set A = {f(x)}, we have ◦ A = ∅. Indeed, if there exists y ∈ ◦ A ⊆ A then y = fn(x) for some positive integer n. Hence x ∈ V f (y) ⊆ ◦ A ⊆ A = {f(x)} which implies that f(x) is a periodic point, a contradiction. Therefore ◦ A = ∅. By Theorem 1.1 we get d0(A) is closed and finally, by Remark 4.3, we conclude that f(x) is a periodic point. (iii) =⇒ (i). First, remark that if x is not a periodic point then {x} is open. In fact, if {x} is not open, then there exists y ∈ f−1({x}) \ {x} which implies that x = f(y) is a periodic point. Now, let A be a dense subset of X. Then every non periodic point of X belongs to A which means that all points of Ac are periodic points. Since [d0(A)] c ⊆ Ac, then all points of [d0(A)] c are also periodic points. Let x ∈ [d0(A)] c and y ∈ {x}. So y is a periodic point and {y} = {x}. Therefore y ∈ [d0(A)] c and consequently {x} ⊆ [d0(A)] c for each x ∈ [d0(A)] c. Since (X, P(f)) is an Alexandroff space, then [d0(A)] c is a closed subset of X and [d0(A)] is open. � © AGT, UPV, 2015 Appl. Gen. Topol. 16, no. 1 62 F -nodec spaces Remarks 4.8. Let (X, f) be a flow in Set. (1) (X, P(f)) is FH-nodec and ρ-nodec, since P(f) is an Alexandroff topol- ogy and for every a ∈ X, dρ(a) is a closed subset of X. (2) Let A be a subset of X and denote B = X \ dρ(A). Then dρ(B) = B. Indeed suppose that there exists x ∈ dρ(B) ∩ dρ(A). Thus there exist a ∈ A and b ∈ B such that f(x) = f(a) = f(b), for all f ∈ C(X). Therefore b ∈ dρ(A) which contradict the fact that B = X \ dρ(A). On the other hand, dρ(B) == ∪[dρ(b) : b ∈ B] is a closed set of (X, P(f)), since the latter is Alexandroff. Hence, dρ(A) is an open subset of X which implies that (X, P(f)) is FH-submaximal and ρ-submaximal. As example of primal space which is T0-Nodec but not nodec, we give the following: Example 4.9. Consider the map f : N −→ N defined by:    f(0) = 0 f(n) = n + 1 if n /∈ 3N f(n) = n − 1 if n ∈ 3N \ {0}. Let n ∈ N \ {0}. − If n ≡ 0[3] then n − 1 /∈ 3N. Hence f2(n) = f(f(n)) = f(n − 1) = n − 1 + 1 = n, that is n is a periodic point. − If n ≡ 2[3] then n + 1 ≡ 0[3]. Hence f2(n) = f(f(n)) = f(n + 1) = n + 1 − 1 = n, that is n is a periodic point. − If n ≡ 1[3] then f(n) = n + 1 ≡ 2[3] that is f(n) is a periodic point. Therefore f(n) is a periodic point for every n ∈ N. Now, consider the topological space (N, P(f)). We can easily check that for each n ∈ N we have f(n) ≡ 0[3] or f(n) ≡ 2[3]. Then {n} is open if and only if n = 0 or n ≡ 1[3]. Let A ⊆ N such that ◦ A = ∅, then a is a periodic point for each a ∈ A. Therefore d0(A) is a closed subset. Therefore, (N, P(f)) is a T0-Nodec space which is not a nodec space. Acknowledgements. The authors thank the referee for his useful comments and suggestions to improve the presentation and the mathematical content of this paper. © AGT, UPV, 2015 Appl. Gen. Topol. 16, no. 1 63 L. Dridi, A. Mhemdi and T. Turki References [1] A. V. Arhangel’skii and P. J. Collins, On submaximal spaces, Topology Appl. 64 (1995), 219–241. [2] K. Belaid and L. Dridi, I-spaces, nodec spaces and compactifications, Topology Appl. 161 (2014), 196–205. [3] K. Belaid, O. Echi and S. Lazaar, T(α,β)-spaces and the Wallman compactification, Int. J. Math. Math. Sc. 68 (2004), 3717–3735. [4] G. Bezhanishvili, L. Esakia and D. Gabelaia, Modal Logics of submaximal and nodec spaces, collection of Essays Dedicated to Dick De Jongh on occasion of his 65th Birthday, J. van Benthem, F. Veltman, A. Troelstra, A. Visser, editors. (2004), pp. 1–13. [5] L. Dridi, S. Lazaar and T. Turki, F-Door Spaces and F-Submaximal Spaces, Appl. Gen. Topol. 14, no. 1 (2013), 97–113. [6] O. Echi and S. Lazaar, Reflective subcategories, Tychonoff spaces, and spectral spaces, Top. Proc. 34 (2009), 307–319. [7] A. Grothendieck and J. Dieudonné, Eléments de géométrie algébrique I: le langage des schemas, Inst. Hautes Etudes Sci. Publ. Math. no. 4, 1960. [8] E. Hewitt, A problem of a set theoretic topology, Duke Mat. J. 10 (1943), 309–333. [9] J. F. Kennisson, The cyclic spectrum of a boolean flow, Theory Appl. Categ. 10 (2002), 392–409. [10] J. F. Kennisson, Spectra of finitely generated boolean flows, Theory Appl. Categ. 16 (2006), 434–459. [11] S. Lazaar, On functionally Hausdorff Spaces, Missouri J. Math. Sci. 1 (2013), 88–97. [12] O. Njastad, On some classes of nearly open sets, Pacific J. Math. 15 (1965), 961–970. [13] J. W. Tukey, Convergence and uniformity in topology, Annals of Mathematics Studies, no. 2. Princeton University Press, (1940) Princeton, N. J. [14] E. K. van Douwen, Applications of maximal topologies, Topology Appl. 51 (1993), 125– 240. [15] R. C. Walker, The Stone-Cech compactification, Ergebnisse der Mathematik und ihrer Grenzgebiete, Band 83. New York-Berlin: Springer-Verlag, (1974). © AGT, UPV, 2015 Appl. Gen. Topol. 16, no. 1 64