() @ Appl. Gen. Topol. 19, no. 1 (2018), 101-127doi:10.4995/agt.2018.7812 c© AGT, UPV, 2018 Some categorical aspects of the inverse limits in ditopological context FİLİZ YILDIZ Department of Mathematics, Hacettepe University, Ankara, Turkey (yfiliz@hacettepe.edu.tr) Communicated by S. Romaguera Abstract This paper considers some various categorical aspects of the inverse systems (projective spectrums) and inverse limits described in the cat- egory ifPDitop, whose objects are ditopological plain texture spaces and morphisms are bicontinuous point functions satisfying a compat- ibility condition between those spaces. In this context, the category InvifPDitop consisting of the inverse systems constructed by the objects and morphisms of ifPDitop, besides the inverse systems of mappings, described between inverse systems, is introduced, and the related ideas are studied in a categorical - functorial setting. In conclusion, an iden- tity natural transformation is obtained in the context of inverse systems - limits constructed in ifPDitop and the ditopological infinite products are characterized by the finite products via inverse limits. 2010 MSC: Primary: 18A30; 46M40; 46A13; 54B10; 54B30; Secondary: 06B23; 18B35; 54E55. Keywords: inverse limit; natural transformation; co-adjoint functor; di- topology; concrete isomorphism; joint topology. 1. Introduction and preliminaries Just as the methods used to derive a new space from two or more spaces are the products, subtextures and quotients of that spaces, so the another effective method is the theory of inverse systems ( projective spectrums) and inverse limits (projective limits). Received 01 July 2017 – Accepted 03 August 2017 http://dx.doi.org/10.4995/agt.2018.7812 F. Yıldız The origins of the study of inverse limits date back to the 1920 ’s. Classi- cal theory of inverse systems and inverse limits are important in the extension of homology and cohomology theory. An exhaustive discussion of inverse sys- tems which are in the some classical categories such as Set, Top, Grp and Rng defined in [1], was presented by the paper [5] which is a milestone in the development of that theory. As is the case with products, the inverse limit might not exist in any category in general whereas inverse systems exist in every category. Note from that [5] inverse limits exist in any category when that category has products of objects and the equalizers [1] of pairs of morphisms, in other words, the inverse limits exist in any category if the category is complete, in the sense of [1]. Additionally, an inverse system has at most one limit. That is, if an inverse limit of any inverse system exists in any category C, this limit is unique up to C-isomorphism. Incidentally, inverse limits always exist in the categories Set, Top, Grp and Rng. Note also that inverse limits are generally restricted to diagrams over directed sets. Similarly, a suitable theory of inverse systems and inverse limits for the categories consisting of textures and ditopological spaces is handled first-time in [17] and [18]. Incidentally, let ’s recall the notions of texture and ditopology introduced in 1993, by Lawrence M. Brown : For a nonempty set S, the family S ⊆ P(S) is called a texturing on S if (S,⊆) is a point-separating, complete, completely distributive lattice containing S and ∅, with meet coinciding with intersection and finite joins with union. The pair (S,S) is then called a texture. If S is closed under arbitrary unions, it is called plain texturing and (S,S) is called plain texture. Since a texturing S need not be closed under the operation of taking the set-complement, the notion of topology is replaced by that of dichotomous topology or ditopology, namely a pair (τ,κ) of subsets of S, where the set of open sets τ and the set of closed sets κ, satisfy the some dual conditions. Hence a ditopology is essentially a “topology” for which there is no a priori relation between the open and closed sets. In addition, a ditopological texture space or shortly ditopological space with respect to a ditopology (τ,κ) on the texture (S,S) is denoted by (S,S,τ,κ). There is now a considerable literature on the theory of ditopological spaces. An adequate introduction to this theory and the motivation for its study may be obtained from [2, 3, 4, 8, 9, 10, 13]. As will be clear from these general references, it is shown that ditopological spaces provide a unified setting for the study of topology, bitopology and fuzzy topology on Hutton algebras. Some of the links with Hutton spaces and fuzzy topologies are expressed in a categorical setting in [14]. In addition, there are close and deep relationships between the bitopological and ditopological spaces as shown in [11, 12] and [15, 16]. In this study, we will use those close relationships insofar as the inverse systems and their inverse limits are concerned in a categorical view. c© AGT, UPV, 2018 Appl. Gen. Topol. 19, no. 1 102 Some categorical aspects of the inverse limits in ditopological context As it is stated before, in [2, 3, 4] we have a few methods, such as product space, subtexture space and quotient space, to derive a new ditopological space from two or more ditopological spaces just like classical case. Recently, it is seen in [17, 18] that the another method used to construct a new ditopological space is the theory of ditopological inverse systems and their limit spaces under the name ditopological inverse limits as the subspaces of ditopological product spaces described in [3, 4, 18]. There are considerable difficulties involved in constructing a suitable theory of inverse systems for general ditopological spaces. Hence, in [17] we confined our attention to a special category whose objects are plain textures, and the basic properties of inverse systems and their inverse limits are investigated in the first-time for texture theory in the context of that category. Accordingly, the various aspects of the inverse systems - limits for texture theory are inves- tigated for plain case and placed them in a categorical - functorial setting. Later, in [18], the theory of inverse systems and inverse limits is handled first-time in the ditopological textural context and we gave a detailed analysis of the theory of ditopological inverse systems and inverse limits insofar as the category ifPDitop whose objects are the ditopological texture spaces which have plain texturing and morphisms are the bicontinuous, w-preserving point functions, is concerned. (For a detailed information and some basic facts about the point-functions between texture spaces, see [3, 10, 11]). By the way, no attempt isn ’t made at the direct systems of ditopological spaces even plain ones, and their (direct) limits as the dual notions of inverse limits. Returning to work at the moment, our main aim in the present paper is to give some further results on the theory of inverse systems and their inverse limits in the context of category ifPDitop. Especially, this paper will present some intriguing connections between the bitopological inverse systems - limit spaces and their ditopological counterparts, in a categorical - functorial setting. Here we will continue to work within the same framework given in [17, 18] that are the major sources of the topic on which we study. According to that, frequent reference will be made to the author ’s papers [17] and [18] which present all details related to the subjects inverse system and inverse limit constructed in the textural context for the plain case, besides providing some useful historical information located in the literature about inverse systems. Otherwise, this paper is largely self-contained although the reader may wish to refer to the literature cited in these papers, for motivation and additional background material specific to the main topic of this paper. Especially, the significant reference in the general field of inverse system theory is [5] and in addition, the reader is referred to [6] for the information about the inverse systems consisting of topological spaces. Specifically, the reader may consult [7] for terms from lattice theory not mentioned here. In addition, we follow the terminology of [1] for all the general c© AGT, UPV, 2018 Appl. Gen. Topol. 19, no. 1 103 F. Yıldız concepts relating to category theory. Thus, if A is a category, Ob A will denote the class of objects and Mor A the class of morphisms of A. In this paper, generally we have tried to give enough details of the proofs to make it clear where various of the conditions imposed are needed, but at the same time to avoid boring the reader with routine verifications. Accordingly, this paper consists of six sections and the layout of paper is as follows: After presenting some background information via the references mentioned in the first section, we introduce and study the category InvifPDitop in Sec- tion 2, mainly. For the paper, it will denote the category whose objects are the inverse systems constructed by the objects of ifPDitop and morphisms are the inverse systems of mappings in the sense of mappings defined between inverse systems. Following that, by describing another related categories and the required functors between the corresponding categories which have some useful properties, we continued to discuss various aspects of the inverse systems and their limits in ifPDitop. In addition, there is a close relationship between ditopological spaces restricted to plain textures and bitopological spaces, as exemplified by a special functor isomorphism given in that section. Hence, we are interested in the connections between bitopological and ditopological inverse systems together with their limits, via that isomorphism. In the end of this section, as one of the principal aims of paper, we obtained an identity natural transformation constructed between the related appropriate functors, described via those connections just mentioned. Specifically, this section con- tains some examples and other results that are important in their own right and which will also be needed later on. In a similar way, in Section 3 we presented a few connections between the category of topological spaces and the category ifPDitop insofar as the inverse systems and their inverse limits are concerned in a categorical setting. Besides these, in Section 4 we investigated the effect of closure operators on inverse systems and limits in ifPDitop, with respect to the joint topologies correspond to the ditopologies located on those inverse systems and limits. A significant characterization theorem which says that by applying the in- verse limit operation, any cartesian products of ditopological plain spaces which are the objects of ifPDitop can be expressed in terms of the finite cartesian products of those spaces, is proved in Section 5. Following that, this section ends with two principal corollaries of that characterization. As the last part of paper, Section 6 gives a conclusion about the whole of this study. 2. Relationships between the inverse systems-limits in the categories of bitopological and ditopological spaces In this section, firstly, let ’s recall all the considerations presented in [12, Section 2] as follows: c© AGT, UPV, 2018 Appl. Gen. Topol. 19, no. 1 104 Some categorical aspects of the inverse limits in ditopological context Let Bitop be the category whose objects are bitopological spaces and mor- phisms are pairwise continuous functions, and the category ifPDitop, intro- duced in [18], is known from the previous section. Accordingly, consider the mapping U from ifPDitop to Bitop by U((S,S,τS,κS) ϕ −→ (T,T,τT ,κT )) = (S,τS,κ c S) ϕ −→ (T,τT ,κ c T ). It is trivial to verify that this is indeed a functor and we omit the details. When applied to many important ditopological spaces, such as the unit interval and real space, the corresponding ditopological T0 axiom as a separation axiom is described as Qs 6⊆ Qt =⇒ ∃C ∈ τ ∪ κ with Ps 6⊆ C 6⊆ Qt and it behaves more like the bitopological weak pairwise T0 axiom, x ∈ yu ∩ yv and y ∈ xu ∩ xv =⇒ x = y. Why this is so, at least in the case of plain textures, we now see by setting up a new functor in the opposite direction of U. To define the suitable functor such that preserves T0 axiom, we restrict ourselves to weakly pairwise T0 bitopological spaces (X,u,v), and consider the smallest subset Kuv of P(X) which contains u∪v c and is closed under arbitrary intersections and unions. Clearly the elements of Kuv have the form (2.1) A = ⋂ j∈J Aj, where Aj = Uj ∪ ⋃ i∈Ij {(V j i ) c | V j i ∈ v}, Uj ∈ u, j ∈ J. In summary, for a weakly pairwise T0 bitopological space (X,u,v), the set u ∪ vc generates a texturing, denoted by Kuv on X. Moreover, it is easy to verify that Kuv is a plain texturing of X since it sepa- rates points, by using the property “weakly pairwise T0” of the space (X,u,v). Finally, we have the plain ditopological space (X,Kuv,u,v c) ∈ Ob ifPDitop satisfying the ditopological T0 separation axiom. Specifically, for a space (S,S,τ,κ) ∈ Ob ifPDitop the equality Kτκc = S is known from [12, Corollary 3.8]. With all these considerations, this process gives a mapping between the subcategory Bitopw0 of Bitop, consisting of weakly pairwise T0 bitopological spaces - pairwise continuous functions and the subcategory ifPDitop0 of if- PDitop, consisting of T0 ditopological spaces and bicontinuous, w-preserving point functions, as follows: H((X,uX,vX) ϕ −→ (Y,uY ,vY )) = (X,KuXvX ,uX,v c X) ϕ −→ (Y,KuY vY ,uY ,v c Y ) Clearly, it defines a functor H : Bitopw0 → ifPDitop0 as mentioned in [9]. Note that this concrete functor is a variant of the functor with the same name considered in [12, 15] in connection with real dicompactness. We are now in a position to give two examples denote the importance of the functor H. c© AGT, UPV, 2018 Appl. Gen. Topol. 19, no. 1 105 F. Yıldız Example 2.1. (1) The unit interval ditopological space (I,I,τI,κI) ∈ Ob ifPDitop0 is the image of the bitopological space (I,uI,vI) ∈ Ob Bitopw0 under H, where τI = uI = {[0,r) | r ∈ I} ∪ {I} and κ c I = vI = {(r,1] | r ∈ I} ∪ {I}. (2) The real ditopological space (R,R,τR,κR) ∈ Ob ifPDitop0 is the image of the bitopological space (R,uR,vR) ∈ Ob Bitopw0 under H, where τR = uR = {(−∞,r) | r ∈ R} ∪ {R,∅} and κ c R = vR = {(r,∞) | r ∈ R} ∪ {R,∅}. It may be verified that H preserves the other basic ditopological separation axioms, besides T0 axiom. Consequently, we have the following fact from [9, 12]: Theorem 2.2. H is a concrete isomorphism between the constructs Bitopw0 and ifPDitop0. Remark 2.3. In view of the above statements, the equalities U ◦ H = 1Bitopw0 and H ◦ U = 1ifPDitop0 are trivial for the functor U : ifPDitop0 → Bitopw0 defined as above. Hence, U is the inverse of H as an isomorphism functor. Incidentally, it is concrete isomorphism since U is identity carried, as well. Now, we can turn our attention to the inverse systems and their inverse limits constructed in ifPDitop, in the light of [18]. Before everything, note that: Remark 2.4. The inverse systems constructed by the objects and morphisms of the category ifPDitop, which are the bonding maps satisfying some conditions given in [18, Definition 3.1], have an inverse limit space described as in [18, Definition 4.1], since ifPDitop has products and equalizers as stated in [18, Corollary 2.6]. Also, the uniqueness of the limit space in the category ifPDitop was mentioned just before [18, Examples 4.5]. Hence, the operation lim ← will be meaningful for the inverse systems given in the context of that category. Notation: According to the major theorem given as [18, Theorem 4.6], if take the inverse system {(Sα,Sα,τα,κα),ϕαβ}α≥β constructed in ifPDitop, over a directed set Λ, then the notations (τ∞,κ∞) and (S∞,S∞,τ∞,κ∞) will be used as inverse limit ditopology and (ditopological) inverse limit space, respectively, where S∞ = lim ← {Sα}, in the remainder of paper. According to let ’s take a glimpse of the mappings between inverse sys- tems: Consider two inverse systems A = {(Sα,Sα,τα,κα),ϕαβ}α≥β and B = {(Tα,Tα,τ ′ α,κ ′ α),ψαβ}α≥β over Λ described in ifPDitop, as in [18, Defini- tion 3.1]. Take into consideration [17, Definition 3.9] which introduces the notion inverse system of mappings or mapping of inverse systems denoted by {tα} : A → B, consisting of the components tα ∈ Mor ifPDitop, satisfying the c© AGT, UPV, 2018 Appl. Gen. Topol. 19, no. 1 106 Some categorical aspects of the inverse limits in ditopological context equality ψβα ◦ tβ = tα ◦ ϕβα, that is, the commutativity of diagram Sβ tβ // ϕβα �� Tβ ψβα �� Sα tα // Tα which associates the bonding maps with the components tα. Hence, by recalling the notion inverse limit space with the notation S∞ defined as in [18, Definition 4.1] and the map t∞ = lim ← {tα}α∈Λ defined in [17, Theorem 4.14], called inverse limit map of the inverse system {tα} of mappings, now let ’s focus on the following crucial theorem proved in [18, Theorem 4.24]: Theorem 2.5. Let {tα} : {(Sα,Sα,τα,κα),ϕβα}β≥α → {(Tα,Tα,τ ′ α,κ ′ α),ψβα}β≥α be an inverse system of mappings in ifPDitop, over a directed set Λ. Then there exists a unique map t∞ ∈ Mor ifPDitop between the spaces (S∞,S∞,τ∞,κ∞) and (T∞,T∞,τ ′ ∞,κ ′ ∞) having the property that for each α ∈ Λ ,the diagram S∞ t∞ // µα �� T∞ ηα �� Sα tα // Tα is commutative, that is tα ◦ µα = ηα ◦ t∞. In this case, i) If each tα is an ifPDitop-isomorphism, t∞ is an ifPDitop-isomorphism. ii) If each tα ◦ µα is surjective, t∞(S∞) is jointly dense in T∞. Notations: In this study, InvC denotes the category whose objects are the inverse systems constructed by the objects of category C and morphisms are the mappings of inverse systems, described as just before Theorem 2.5, namely, the inverse systems of C-morphisms defined between the objects of C. Particulary, the following notation will be required for the remainder of paper, mostly: InvifPDitop0 will denote the category consisting of inverse systems constructed by T0 ditopological plain texture spaces as objects of ifPDitop0, and by the mappings between inverse systems, namely, the inverse systems of mappings defined as in Theorem 2.5. Incidentally, we have the following categorical fact about the inverse systems due to [18, Remark 3.2]: Remark 2.6. An inverse system in any category admits an alternative descrip- tion in terms of functors. A directed set Λ becomes a category if each relation α ≤ β is regarded as a map α → β, that is the morphisms consist of arrows α → β if and only if α ≤ β. Then, c© AGT, UPV, 2018 Appl. Gen. Topol. 19, no. 1 107 F. Yıldız Any inverse system in the category ifPDitop over the directed set Λ is actually a contravariant functor from Λ to ifPDitop. In the light of Remark 2.6, note that the objects and morphisms of InvifPDitop may be regarded as the functors and natural transformations, respectively. Example 2.7. If {(Sα,uα,vα),fαβ}α≥β ∈ Ob InvBitopw0 then the system {(Sα,Kuαvα,uα,v c α),ϕαβ}α≥β consisting of the spaces H(Sα,uα,vα) = (Sα,Kuαvα,uα,v c α) ∈ Ob ifPDitop0, corresponding to the bitopological spaces (Sα,uα,vα) ∈ Ob Bitopw0, describes an inverse system via the isomorphism functor H given in Theorem 2.2 and all the above considerations. Trivially, this system is an object of InvifPDitop0. Now, by taking into account Example 2.7, immediately we have the follo- wing: Example 2.8. If (S∞,u∞,v∞) ∈ Ob Bitopw0 is the inverse limit of the in- verse system {(Sα,uα,vα),fαβ}α≥β ∈ Ob InvBitopw0 then the corresponding plain space (S∞,Ku∞v∞,u∞,v c ∞) ∈ Ob ifPDitop0 is the inverse limit of cor- responding inverse system {(Sα,Kuαvα,uα,v c α),ϕαβ}α≥β ∈ Ob InvifPDitop0, where ϕαβ = fαβ for α ≥ β. Let ’s prove it: First of all, recall the fact S∞ ⊆ ∏ α Sα. Thus, it is clear that u∞ = ( ∏ α uα)|S∞ = ( ⊗ α uα)|S∞ since the textural and classical products of topolo- gies are coincide by the plainness property. On the other hand, similar to the explanations given in [15, Section 3] we have vc∞ = ( ⊗ α vcα)|S∞ since ( ⊗ α vα)|S∞ = ( ∏ α vα)|S∞ = v∞ and by [3, Lemma 2.7] which is peculiar to the theory of product ditopologies. Hence, it remains to prove the equality Ku∞v∞ = ( ⊗ α Kuαvα)|S∞. For it, we can show that the types of elements of these two families are absolutely the same: If A ∈ Ku∞v∞ then let ’s recall the form of A as follows: A = ⋂ j∈J Aj, where Aj = Uj ∪ ⋃ i∈Ij {(V j i ) c | V j i ∈ v∞}, Uj ∈ u∞, j ∈ J Here, V j i ∈ v∞ = ( ∏ α vα)|S∞ and so V j i = C j i ∩ S∞, where C j i ∈ ∏ α vα. In this case, for Tjαi ∈ vαi C j i = ⋃ ⋂ π−1αi [T j αi ] and so (V j i ) c = ⋂ ⋃ (π−1αi [Sαi \ T j αi ]). Similarly, Uj = Bj ∩ S∞ where Bj ∈ ⊗ α uα, and so Uj = ( ⋃ ⋂ π−1αi [G j αi ]) ∩ S∞ where G j αi ∈ uαi, by the definition of product topology. Hence, Aj = ( ⋃ ⋂ (παi|S∞) −1[Gjαi])∪( ⋃ ⋂ (παi|S∞) −1[Sαi \T j αi ]) and finally, by the fact that A = ⋂ j∈J Aj we have A = ⋂ j∈J[( ⋃ ⋂ (παi|S∞) −1[Gjαi])∪( ⋃ ⋂ (παi|S∞) −1[Sαi \ Tjαi])]. c© AGT, UPV, 2018 Appl. Gen. Topol. 19, no. 1 108 Some categorical aspects of the inverse limits in ditopological context On the other hand, if B ∈ ( ⊗ α Kuαvα)|S∞ then B = M ∩ S∞ where M ∈ ⊗ α∈I Kuαvα. In this case, M = ⋂ ⋃ α∈I π−1α [Kα], where Kα ∈ Kuαvα. Thus, we have the form Kα = ⋂ j∈J Dαj , where D α j = W α j ∪ ⋃ i∈Ij {Sα \ (Z j i ) α | (Z j i ) α ∈ vα}, Wj α ∈ uα, j ∈ J. Hence Kα = ⋂ (Uj α ∪ ( ⋃ Sα \ (Z j i ) α)), and so M = ⋂ ( ⋃ (π−1j [ ⋂ (Wαj ∪ ⋃ Sα \ (Z j i ) α])). In this case, with B = M ∩S∞ we have B = ⋂ [ ⋃ (((πj|S∞) −1[ ⋂ Wαj ]∪(πj|S∞) −1[ ⋂ ⋃ (Sα\(Z j i ) α)])] = ⋂ [ ⋃ ⋂ (πj|S∞) −1[Wαj ] ∪ ⋂ ⋃ (πj|S∞) −1[Sα \ (Z j i ) α]]. Consequently, it is easy to check that the sets A and B have the same type if consider Gjαi as W α j and T j αi as Z j i by neglecting the details of indices, as well as by leaving the other details of required equality to the interested reader. Now, let ’s recall the notion of inverse limit map introduced in [17, Theorem 4.14] as a notion of peculiar to the texture theory, as well as mentioned in Section 1. Accordingly, in order to prove the next theorem, we need a special property of inverse limits maps, which is proved in the following: Proposition 2.9. Consider {hα} : {(Sα,Sα),ϕβα}α≤β → {(Tα,Tα),ψβα}α≤β and {gα} : {(Tα,Tα),ψβα}α≤β → {(Zα,Zα),φβα}α≤β between the inverse sys- tems of textures then {gα ◦ hα} : {(Sα,Sα),ϕβα}α≤β → {(Zα,Zα),φβα}α≤β is also a mapping of inverse system and lim ← {gα ◦ hα}α∈Λ = lim ← {gα}α∈Λ ◦ lim ← {hα}α∈Λ Proof. At first, we define the composition operation for the mappings of inverse systems as follows : {gα} ◦ {hα} = {gα ◦ hα} by using the composition operation on the morphisms of ifPDitop. On the other hand, because of the first inverse system, we have the equality ψβα ◦ hβ = haα ◦ ϕβα by the commutativity of related diagram constructed between the sets Sα,Tα,S∞ and T∞. Similarly, from the second inverse system, we have the equality φβα ◦ gβ = gaα ◦ ψβα by the commutativity of related diagram constructed between the sets Tα,Zα,T∞ and Z∞. Hence, by considering the above two equalities, we have the result: φβα ◦ (gβ ◦ hβ) = (gα ◦ hα) ◦ ϕβα In fact, it says that {gα ◦ hα} becomes an inverse system of mappings by [17, Definition 3.9]. Therefore, now we can look at the commutativity of diagram. Firstly, recall µα ◦ h∞ = hα ◦ λα and ηα ◦ g∞ = gα ◦ µα by [17, Theorem 4.14]. Thus, due to these equalities, we have ηα◦(g∞◦h∞) = (ηα◦g∞)◦h∞ = (gα◦µα)◦h∞ = gα◦(µα◦h∞) = (gα◦hα)◦λα and so the related diagram is commutative. Finally, from the uniqueness of inverse limit maps, mentioned in Theorem 2.5, the required result lim ← {gα ◦ c© AGT, UPV, 2018 Appl. Gen. Topol. 19, no. 1 109 F. Yıldız hα}α∈Λ = g∞ ◦ h∞ is proved. That is, lim ← {gα ◦ hα}α∈Λ = lim ← {gα}α∈Λ ◦ lim ← {hα}α∈Λ. � Remark 2.10. For the remainder of paper, we will use the above final equality under the name transitivity property of inverse limit maps. From Remark 2.4, the inverse systems which are the objects of InvifPDitop have a unique inverse limit space as an object of ifPDitop. With the reference to this fact, we have the following immediately; Theorem 2.11. The limit operation lim ← of assigning an inverse limit in ifPDi- top to each object in InvifPDitop and an inverse limit map t∞ ∈ Mor ifPDitop to each inverse system {tα}α ∈ Mor InvifPDitop of maps tα ∈ Mor ifPDitop, forms the covariant functor lim ← : InvifPDitop → ifPDitop. Proof. Let ’s recall that for each inverse system which is an object of InvifPDitop0 we can obtain an inverse limit space in ifPDitop and moreover, it is unique by Remark 2.4. Now, according to Theorem 2.5, if take the morphism {tα}α : {(Sα,Sα,τα,κα),ϕβα}β≥α → {(Tα,Tα,τ ′ α,κ ′ α),ψβα}β≥α in InvifPDitop then there exists a unique map t∞ = lim ← {tα}α∈Λ ∈ Mor ifPDitop between the cor- responding inverse limit spaces (S∞,S∞,τ∞,κ∞) and (T∞,T∞,τ ′ ∞,κ ′ ∞) which are the objects of ifPDitop, having the property that for each α ∈ Λ the diagram S∞ t∞ // µα �� T∞ ηα �� Sα tα // Tα is commutative, that is tα◦µα = ηα◦t∞. Also, t∞ is the identity id(S∞,S∞,τ∞,κ∞) if suppose that the mapping {tα}α of inverse systems is identity, that is each map tα : Sα → Tα, α ∈ Λ is the identity id(Sα,Sα,τα,κα) on Sα. Additionally, as it is stated in Proposition 2.9, the inverse limit maps have the transitivity prop- erty and so the limit operation lim ← satisfies the composition rule lim ← {tα ◦hα} = lim ← {tα} ◦ lim ← {hα}. Hence, the mapping lim ← : InvifPDitop → ifPDitop is a co- variant functor. � Notation: The covariant functor lim ← described in Theorem 2.11, as the limit operation in the context of ifPDitop, will be used under the notation E for the remainder of paper. Actually, note that we can always define a covariant functor between the categories C and InvC, for any category C which has the equalizers and pro- ducts. c© AGT, UPV, 2018 Appl. Gen. Topol. 19, no. 1 110 Some categorical aspects of the inverse limits in ditopological context Remark 2.12. (1) By virtue of the fact that any inverse system consisting of the objects of Bitop has an inverse limit since Bitop has equalizers and prod- ucts, we can describe covariant functor, under the name B between the categories Bitop and InvBitop. (2) The above functor B introduced in (1) may be considered as the re- stricted mapping between the full subcategory Bitopw0 of Bitop and the full subcategory InvBitopw0 of InvBitop. Obviously, that restric- tion is a covariant functor, as well. (3) Furthermore, if we recall that the categories Bitopw0 and ifPDitop0 are isomorphic via the functor H constructed by using the fact that weakly pairwise T0 bitopology generates the smallest plain texturing and T0 ditopology, as mentioned in Theorem 2.2, then we may describe a functor between the categories InvBitopw0 and InvifPDitop0 in a natural way. According to the statement (3), we are now in a position to give a next isomorphism functor as follows: Theorem 2.13. The categories InvBitopw0 and InvifPDitop0 are concretely isomorphic. Proof. First of all, if consider the isomorphism functor H given in Theo- rem 2.2, between the categories Bitopw0 and ifPDitop0, clearly the mapping X : InvBitop w0 → InvifPDitop0 may be defined by using H: Taking into account the ideas given in Example 2.7, then we may define the map X({(Sα,uα,vα),fαβ}α≥β) = {(Sα,Kuαvα,uα,v c α),fαβ}α≥β where H(Sα,uα,vα) = (Sα,Kuαvα,uα,v c α), H(fαβ) = fαβ, and if take the inverse system {tα} of mappings as the morphism between two inverse systems which are objects of InvBitopw0 then it is easy to show that it is also a morphism in InvBitop. Indeed, if take tα ∈ Mor Bitop, for each α, that is, tα is pairwise continu- ous then it is w-preserving and bicontinuous between the corresponding di- topological plain spaces and finally, the equality X({tα}) = {tα} is mean- ingful, as well. In this case, for the inverse system mappings, the equality X({tα} ◦ {hα}) = X({tα ◦ hα}) = {tα ◦ hα} = {tα} ◦ {hα} is trivial. Also, from X(id{(Sα,uα,vα),fαβ}α≥β ) = idX({(Sα,uα,vα),fαβ}α≥β), the map X describes a functor, naturally. Now we will turn our attention to the isomorphism conditions for X. It is easy to show that X is full and faithful, since it is bijective between hom-set restrictions by the fact that the functor H given in Theorem 2.2 is full and faithful. As the final step, it remains to prove that the bijectivity of X on objects of InvBitop w0 and InvifPDitop0 , and it is clear from the bijectivity of the functor H. � In the light of considerations presented in Remark 2.12 and Theorem 2.13, now we can start to construct a major part in that theory, consisting of the c© AGT, UPV, 2018 Appl. Gen. Topol. 19, no. 1 111 F. Yıldız useful implications and an identity natural transformation which arises from those implications: A Natural Transformation in the Context of Inverse Systems and Limits Located Inside the Categories Bitopw0 and ifPDitop0 : As we promised in Section 1, firstly a natural transformation will be de- scribed between the corresponding functors, and later, that the natural trans- formation is identity will be proved, thoroughly. Let ’s start by recalling the corresponding required functors as follows: InvBitop w0 B −→ Bitopw0 H −→ ifPDitop0 {(Xα,uα,vα),ϕαβ}α≥β 7→ (X∞,u∞,v∞) 7→ (X∞,Ku∞v∞,u∞,(v∞) c) InvBitop w0 X −→ InvifPDitop 0 E −→ ifPDitop0 {(Xα,uα,vα),ϕαβ}α≥β 7→ {(Xα,Kuαvα,uα,(vα) c),ϕαβ}α≥β 7→ (X∞,Z,T,K) where Z = ( ⊗ α Kuαvα)|X∞, T = ( ⊗ α uα)|X∞ and K = ( ⊗ α vcα)|X∞ Now, with the previous considerations, if take the equalities F = H ◦ B : InvBitop w0 → ifPDitop0 G = E ◦ X : InvBitop w0 → ifPDitop0 then it is clear that F and G are functors as compositions of the functors H,B and E,X, respectively. Consider a mapping τ : F → G. In particular; Theorem 2.14. τ is an identity natural transformation between the functors F and G. Proof. Let the inverse system A = {(Xα,uα,vα),ϕαβ}α≥β ∈ Ob InvBitop w0 over Λ and the mapping τA : FA → GA. Firstly, it is easy to verify that FA = GA by the considerations mentioned in Example 2.8 and thus, the mapping τA is an ifPDitop0-identity morphism. On the other hand, for the inverse system A′ = {(X′α,u ′ α,v ′ α),ϕ ′ αβ}α≥β ∈ Ob InvBitop w0 over Λ, take the inverse system {kα} : A → A ′ ∈ Mor InvBitop w0 of mappings kα : Xα → X ′ α, α ∈ Λ, as in described in Theorem 2.5. Also, as- sume that lim ← A = lim ← {Xα}α∈Λ = X∞ and lim ← A′ = lim ← {X′α}α∈Λ = X ′ ∞. Let ξ : {(Xα,uα,vα),ϕαβ}α≥β → {(X ′ α,u ′ α,v ′ α),ϕ ′ αβ}α≥β be the mapping {kα}α∈Λ of inverse systems, with the components kα : Xα → X ′ α ∈ Mor Bitopw0, α ∈ Λ. c© AGT, UPV, 2018 Appl. Gen. Topol. 19, no. 1 112 Some categorical aspects of the inverse limits in ditopological context With all the above notations, now we may construct the following diagram: (X′∞,Ku′∞v′∞,u ′ ∞,(v ′ ∞) c) τA′ // (X′∞,( ⊗ α Ku′αv ′ α )|X′∞,( ⊗ α u′α)|X′∞,( ⊗ α (v′α) c)|X′∞) (X∞,Ku∞v∞,u∞,(v∞) c) Fξ OO τA // (X∞,( ⊗ α Kuαvα)|X∞,( ⊗ α uα)|X∞,( ⊗ α (vα) c)|X∞ Gξ OO )) In order to see that this diagram is commutative, we need to show the equality Fξ = Gξ for all ξ ∈ Mor InvBitopw0 : Clearly, each kα : Xα → X ′ α is pairwise continuous and by F = H ◦ B we have F({kα}α∈Λ) = H(B{kα}α∈Λ) = H(k∞) where k∞ = lim ← {kα}α∈Λ ∈ Mor Bitopw0 and by applying the isomorphism H : Bitopw0 → ifPDitop0 to the limit map k∞ ∈ Mor InvBitopw0, we obtained H(k∞) = k∞ since H is identity on morphisms. Finally, Fξ = F({kα}α∈Λ) = k∞. On the other hand, now let ’s turn our attention to G(ξ) and recall the equality G = E◦X. According to that, we have G({kα}α∈Λ) = E(X{kα}α∈Λ) = E({kα}α∈Λ) since the isomorphism X described in Theorem 2.13 is the identity on morphisms of InvBitopw0 and InvifPDitop0. Hence, by applying the functor E : InvifPDitop0 → ifPDitop0 to the mapping {kα}α∈Λ, we describe the map E({kα}) = h∞, where h∞ = lim ← {kα}α∈Λ ∈ Mor Bitopw0. Hence Gξ = G({kα}α∈Λ) = h∞. Now, let ’s see that t∞ = h∞: the inverse systems considered above are exactly same since the spaces and bonding maps are the same. Also, the property of commutativity ηα ◦t∞ = tα ◦µα, α ∈ Λ is satisfied for the map h∞, as well. In this case, by virtue of the fact that the inverse limit of the mappings of inverse systems is unique by [17, Theorem 4.14], we have Fξ = Gξ. Thus, the equality F = G is verified and τ is identity. Moreover, we have Gξ◦τA = τA′◦Fξ since τA,τA′ are identities and so, the diagram is commutative. � As a result of the above considerations, τ : F → G is the identity natural transformation. In a similar way to the considerations given in Section 2, next section will discuss the relations between the topological inverse systems - limits and di- topological inverse systems - limits insofar as the theory of plain textures are concerned. 3. Relationships between the inverse systems-limits in the categories of topological and ditopological spaces Now we will show that we may associate with the ditopology (τ,κ) on a plain texture (S,S) a topology Jτκ on S, by adapting the notion of appropriate joint topology for a ditopology described in [11], to the plain case: c© AGT, UPV, 2018 Appl. Gen. Topol. 19, no. 1 113 F. Yıldız Definition 3.1. Let (S,S,τ,κ) ∈ Ob ifPDitop. We define the joint topology on S in terms of its family Jcτκ of closed sets by the condition W ∈ Jcτκ ⇐⇒ (s ∈ S, G ∈ η(s),K ∈ µ(s) =⇒ G ∩ W 6⊆ K) =⇒ s ∈ W. Here η(s) = {N ∈ S | Ps ⊆ G ⊆ N 6⊆ Qs for some G ∈ τ} and µ(s) = {M ∈ S | Ps 6⊆ M ⊆ K ⊆ Qs for some K ∈ κ}. For the details about filter η(s) and cofilter µ(s) for s ∈ S, see [8, 11, 16]. The verification of that Jcτκ satisfies the closed-set axioms is straightforward and on passing to the complement this reveals that (i) {G ⊆ S | G ∈ τ} ∪ {S \ K ⊆ S | K ∈ κ} is a subbase, and (ii) {G ∩ (S \ K) ⊆ S | G ∈ τ,K ∈ κ} a base of open sets for the topology Jτκ on S. In case (X,u,v) is an object of Bitop, we have the space (X,P(X),u,vc) ∈ Ob ifPDitop, and clearly obtain Jτκ = u ∨ v as the joint topology of (u,v), where τ = u and κ = vc. Hence we will refer to Jτκ as the joint topology of (τ,κ) on S. Remark 3.2. (1) For (S,S,τ,κ) ∈ Ob ifPDitop, it is trivial to see that κ ⊆ Jcτκ and τ ⊆ Jτκ. In addition, the family τ ∪ κ c is the subbase for the joint topology Jτκ. (2) From now on, in this work we will use the terms jointly closed (open, dense) for the set which is closed (open, dense) with respect to the appropriate joint topology of the ditopology on space. Note that the following statements are adapted forms of general cases given in [11] to the category ifPDitop. Here Top will denote the category of topo- logical spaces and continuous functions. Theorem 3.3. The mapping J : ifPDitop → Top defined by J : ((S,S,τS,κS) ϕ −→ (T,T,τT ,κT )) = (S,JτSκS ) ϕ −→ (T,JτT κT ) is an adjoint functor. It is clear that J is full, faithful and isomorphism-dense functor although it is not a functor isomorphism since it is not one-to-one on the objects. Corollary 3.4. The functor T : Top → ifPDitop given by T(X,T) = (X,P(X),T,Tc), T(ϕ) = ϕ is the co-adjoint of J. Here note also that T is not a functor isomorphism. In this section, we will be interested in the category InvTop whose objects are the inverse systems constructed by the objects of Top and morphisms are the inverse systems constructed by the morphisms of Top, as well as the map- pings between the inverse systems constructed in Top. Naturally, a covariant functor may be established between the categories Top and InvTop since any c© AGT, UPV, 2018 Appl. Gen. Topol. 19, no. 1 114 Some categorical aspects of the inverse limits in ditopological context inverse system constructed in Top has an inverse limit by the fact that Top has equalizers and products as mentioned in [5]. Obviously, we can’t expect to find an isomorphism between the categories InvTop and InvifPDitop and now, we may turn our attention to the relation- ships between the objects of categories InvTop and InvifPDitop: It is known that an object of InvifPDitop can be obtained as the natural counterpart of an object of InvTop by [18, Example 3.4]. Thus, by applying the similar considerations to Corollary 3.4 we can describe a co-adjoint functor from InvTop to InvifPDitop. Conversely, in order to construct an opposite functor from InvifPDitop to InvTop, let’s consider the reciprocal objects, and the adjoint functor J firstly. That is, take {(Sα,Sα,τα,κα),ϕαβ}α≥β ∈ Ob InvifPDitop, and construct the image J(Sα,Sα,τα,κα) = (Sα,Jτακα) ∈ Ob Top. In this case, for the bonding map ϕαβ : Sα → Sβ ∈ Mor ifPDitop, we have J(ϕαβ) = ϕαβ : (Sα,Jτακα) → (Sβ,Jτβκβ ) as a morphism of Top since J is a functor. In fact, J is the identity on morphisms. Hence, we construct the inverse system {(Sα,Jτακα),ϕβα}β≥α ∈ Ob InvTop and so a mapping which is described as follows : Theorem 3.5. The mapping JInv : InvifPDitop → InvTop defined by JInv : ({(Sα,Sα,τα,κα),ϕαβ}α≥β {tα} −→ {(Tα,Tα,τ ′ α,κ ′ α),ψαβ}α≥β) = {(Sα,Jτακα),ϕαβ}α≥β {tα} −→ {(Tα,Jτ′ α κ′ α ),ψαβ}α≥β is an adjoint functor. Proof. Firstly, we need to check that JInv is a functor. Assume that {tα}α ∈ Mor InvifPDitop. In this case, the maps tα : Sα → Tα for each α, are bicontinu- ous and w-preserving as the morphisms in ifPDitop. By the definition of joint topology, it is easy to show that tα is continuous for each α, as the morphism of Top between the joint topological spaces (Sα,Jτακα) and (Tα,J ′ τ′ακ ′ α ). To show JInv is an adjoint, now take {(Xα,Tα),φαβ}α≥β ∈ Ob InvTop. Then ({idXα},{(Xα,P(Xα),Tα,T c α),φαβ}α≥β) is a JInv-structured arrow by JInv({(Xα,P(Xα),Tα,T c α),φαβ}α≥β) = {(Xα,JTαTcα),φαβ}α≥β) and by the fact that {idXα} : {(Xα,JTακα),φαβ}α≥β → {(Xα,JTακα),φαβ}α≥β is an InvTop- morphism. To show ({idXα},{(Xα,P(Xα),Tα,T c α),φαβ}α≥β) has the univer- sal property, take {(Sα,Sα,τ ∗ α,κ ∗ α),θαβ}α≥β ∈ Ob InvifPDitop and let {ϕα} : {(Xα,Tα),φαβ}α≥β → JInv{(Sα,Sα,τ ∗ α,κ ∗ α),θαβ}α≥β = {(Sα,Jτ∗ακ∗α),θαβ}α≥β be an InvTop-morphism. (X,T) idX // ϕ **❚❚ ❚ ❚ ❚ ❚ ❚ ❚ ❚ ❚ ❚ ❚ ❚ ❚ ❚ ❚ ❚ ❚ ❚ ❚ JInv(X,P(X),T,T c) = (X,T) JInv(ϕ̄) �� ✤ ✤ ✤ ✤ JInv(S,S,τ,κ) = (S,Jτκ) Since ϕ maps into S the only point function ϕ̄ : X → S making the above dia- gram commutative is ϕ, so it remains only to verify that ϕ : (X,P(X),T,Tc) → c© AGT, UPV, 2018 Appl. Gen. Topol. 19, no. 1 115 F. Yıldız (S,S,τ,κ) is a morphism in ifPDitop. Certainly ϕ is ω-preserving, due to the fact that ϕ(x)ωS ϕ(x) for all x ∈ X. Moreover, ϕ is bicontinuous. To see this, note that we have ϕ←A = ϕ−1A = ϕ−1(A ∩ Sp) for all A ∈ S. Hence, G ∈ τ =⇒ G ∩ Sp ∈ Jτκ =⇒ ϕ ←G = ϕ−1(G ∩ Sp) ∈ T, and K ∈ κ =⇒ ϕ←K ∈ Tc likewise. � Particularly, by virtue of Theorem 3.3 and Theorem 3.5 we have the follow- ing: Remark 3.6. Let ’s take an inverse system A = {(Sα,Sα,τα,κα),ϕαβ}α≥β ∈ Ob InvifPDitop. In this case, we construct the system JInv(A) = {(Sα,Jτακα),ϕαβ}α≥β ∈ Ob InvTop, and have the inverse limit space (S∞,S∞,τ∞,κ∞) ∈ Ob ifPDitop by Theorem 2.11. Thus, J(S∞,S∞,τ∞,κ∞) = (S∞,Jτ∞κ∞) ∈ Ob Top. In addition, we have an inverse limit lim ← JInv(A) = (Y,V) ∈ Ob Top due to the fact that Top has equalizers and products as mentioned in [5]. Now let ’s turn our attention to the main question; Is the space (Y,V) same with the space (S∞,Jτ∞κ∞) in Top ? Firstly note that the systems {(Sα,Sα,τα,κα),ϕαβ}α≥β and {(Sα,Jτακα), ϕαβ}α≥β have the same bonding maps, so Y = S∞ as a subset of ∏ α Sα, trivially. As a next step, in order to prove the equality Jτ∞κ∞ = V, note that the facts ( ∏ α Jτακα)|S∞ = V, ( ⊗ α Sα)|S∞ = S∞, ( ⊗ α τα)|S∞ = τ∞ and ( ⊗ α κα)|S∞ = κ∞. Accordingly, let A ∈ Jτ∞κ∞, so A = ⋃ δ (Gδ ∩(S∞\Kδ)) where Gδ ∈ ( ⊗ α τα)|S∞, Kδ ∈ ( ⊗ α κα)|S∞ for each δ. Note here that Gδ = ( ⋃ ( ⋂ finite π−1αi [G δ αi ]))∩S∞ and similarly, Kδ = ( ⋂ ( ⋃ finite π−1αi [K δ αi ])) ∩ S∞, where G δ αi ∈ ταi and K δ αi ∈ καi. Hence, A = ( ⋃ ⋂ finite π−1αi [ ⋃ Gδαi]) ∩ ( ⋃ ⋂ finite π−1αi [ ⋃ (Sαi \ K δ αi )]) ∩ S∞. On the other hand, let the set B ∈ ( ∏ α Jτακα)|S∞ = V, where V denotes the product topology on S∞. In this case, B can be written as ( ⋃ ⋂ finite π−1αj [Gαj ])∩ S∞, Gαj ∈ Jτακα. Thus, B = ( ⋃ ⋂ finite π−1αj [ ⋃ (Cαj ∩ (Sαj \ Dαj ))]) ∩ S∞ where Cαj ∈ ταj , Dαj ∈ καj and so, B = ( ⋃ ⋂ finite π−1αj [ ⋃ Cαj ]) ∩ ( ⋃ ⋂ finite π−1αj [ ⋃ (Sαj \ Dαj )]) ∩ S∞. Consequently, it is easy to check that the types of sets A and B are the same. It means that the topologies V and Jτ∞κ∞ coincides. 4. Effect of the closure operator on inverse systems and limits in the category ifPDitop By recalling the notion of appropriate joint topology described for a ditopol- ogy, as presented in the previous section, we have the following significant theorem, immediately: c© AGT, UPV, 2018 Appl. Gen. Topol. 19, no. 1 116 Some categorical aspects of the inverse limits in ditopological context Theorem 4.1. Let Λ be a directed set. For subspace (U,SU,τU,κU) ∈ Ob ifPDitop of the inverse limit space (S∞,S∞,τ∞,κ∞) ∈ Ob ifPDitop of the inverse system {(Sα,Sα,τα,κα),ϕαβ}α≥β ∈ Ob InvifPDitop, the families {(Uα,Sα|Uα,τα|Uα,κα|Uα),ϕαβ|Uα }α≥β and {(Uα,Sα|Uα,τα|Uα,κα|Uα),ϕαβ}α≥β describe two objects in InvifPDitop as the inverse systems, where Uα = µα(U) = πα|S∞(U), ϕαβ = ϕαβ|Uα and Uα denotes the closure in Sα of the subset Uα ⊆ Sα with respect to the joint topology of the ditopology (τα,κα), α ∈ Λ. Proof. Firstly, let us prove that {(Uα,Sα|Uα,τα|Uα,κα|Uα),ϕαβ}α≥β is an ob- ject of InvifPDitop: Note that we have µβ(s) = ϕαβ(µα(s)) for s ∈ U and β ≤ α. Indeed, if s ∈ U then µα(s) ∈ µα(U) and so µα(s) ∈ Uα. In this case, ϕαβ(µα(s)) = ϕαβ(µα(s)). Also the equality ϕαβ(µα(s)) = µβ(s) for α ≥ β is known by [17, Lemma 4.3], thus we have ϕαβ(µα(s)) = µβ(s) for s ∈ U and α ≥ β, as required. On the other hand, with the continuity of bonding map ϕαβ we have ϕαβ(Uα) = ϕαβ(µα(U)) ⊆ ϕαβ(µα(U)) = µβ(U) = Uβ and then, it is clear that the point function ϕαβ is defined from Uα onto Uβ. Following that, ϕαβ is a morphism of ifPDitop since it is a restriction of ϕαβ ∈ Mor ifPDitop to the subset Uα ⊆ Sα. Incidentally, the equality ϕβγ ◦ ϕαβ = ϕαγ may be easily proved for the elements of Uα via the equality ϕβγ ◦ ϕαβ = ϕαγ. As a next step, we have the equality ϕαα(s) = ϕαα(s) = s for s ∈ Uα, as ϕαα is the identity idSα on Sα. That is, ϕαα = idUα = idSα|Uα. Consequently, the family {(Uα,Sα|Uα,τα|Uα,κα|Uα),ϕαβ}α≥β forms an ob- ject in InvifPDitop. Furthermore, in a similar way to the above proof, it is easy to check that the family {(Uα,Sα|Uα,τα|Uα,κα|Uα),ϕαβ|Uα }α≥β describes an inverse system in ifPDitop, and so an object in InvifPDitop. � According to Remark 2.4, we have the following, right away. Proposition 4.2. U∞ = lim ← {Uα} ⊆ lim ← {Sα} = S∞ Proof. Conversely, assume that U∞ = lim ← {Uα} 6⊆ lim ← {Sα} = S∞, so there exists s = {sα} ∈ ∏ α∈Λ Sα such that U∞ 6⊆ Qs and Ps 6⊆ S∞. In this case, s ∈ ∏ α∈Λ Uα and ϕαβ|Uα(sα) = sβ for every sα ∈ Uα, α,β ∈ Λ such that α ≥ β. Moreover, we have the equality ϕαβ|Uα(sα) = ϕαβ(sα) for sα ∈ Uα. Thus, because of the facts sα ∈ Sα, α ∈ Λ and ϕαβ(sα) = sβ for α ≥ β, the point s = {sα} becomes an element of S∞, obviously and this gives a contradiction. � Proposition 4.3. Let {(Sα,Sα,τα,κα),ϕαβ}α≥β ∈ Ob InvifPDitop be an in- verse system over a directed set Λ and (S∞,S∞,τ∞,κ∞) ∈ Ob ifPDitop be the c© AGT, UPV, 2018 Appl. Gen. Topol. 19, no. 1 117 F. Yıldız inverse limit of that system. If Uα ∈ J c τακα , α ∈ Λ and lim ← {Uα} = U∞ for the inverse subsystem {(Uα,Sα|Uα,τα|Uα,κα|Uα),ϕαβ|Uα}α≥β ∈ Ob InvifPDitop, then U∞ ∈ J c τ∞κ∞ . Proof. By the definition of inverse limit and the equality ϕαβ|Uα(sα) = ϕαβ(sα) for sα ∈ Uα, α ≥ β, the inclusion U∞ = lim ← {Uα} ⊆ lim ← {Sα} = S∞ is imme- diate, as mentioned in Proposition 4.2 as well. Now, let us prove U∞ ∈ J c τ∞κ∞ : If Pa 6⊆ U∞, that is a /∈ U∞ for a = {aα} ∈ S∞, then a /∈ ∏ α∈Λ Uα due to the equality ϕαβ|Uα(sα) = ϕαβ(sα) for sα ∈ Uα, α ≥ β. In this case, there exists α0 ∈ Λ such that aα0 /∈ Uα0, that is Paα0 6⊆ Uα0. Additionally, the subset µ −1 α0 [Uα0] ⊆ S∞ is an element of J c τ∞κ∞ since the limiting projection map µα0 : S∞ → Sα0 is continuous between the corresponding joint topological spaces and Uα0 ∈ J c τα0κα0 . On the other hand, the statements Pa 6⊆ µ −1 α0 [Uα0] and U∞ ⊆ µ −1 α0 [Uα0] may be showed as follows: Conversely, if Pa ⊆ µ −1 α0 [Uα0] then we have µα0(a) = aα0 ∈ Uα0 which is a contradiction. Also, assume that U∞ 6⊆ µ −1 α0 [Uα0]. Thus there exists a point z ∈ S∞ such that U∞ 6⊆ Qz and Pz 6⊆ µ −1 α0 [Uα0]. Hence, µα0(z) = zα0 /∈ Uα0 and z = {zα} /∈ ∏ α∈Λ Uα gives the fact that z /∈ U∞ which is a contradiction. � From now on, in the remainder of this Section we will use all of the above notations, in exactly the same form. By virtue of Theorem 4.1 and the last proposition, now we have the next: Theorem 4.4. If U denotes the closure of the subset U ⊆ S∞ with respect to the joint topology of the limit ditopology (τ∞,κ∞) then (1) lim ← {Uα} is jointly closed subspace of S∞ (2) lim ← {Uα} = U ⊆ S (3) U = ⋂ α∈Λ µ−1α [Uα] Proof. (1) Before everything, let ’s see that lim ← {Uα} ⊆ S∞, where S∞ = lim ← {Sα}: Conversely, if the inclusion is not true, then there exists a point s = {sα} ∈∏ α∈Λ Sα such that lim ← {Uα} 6⊆ Qs and Ps 6⊆ S∞. Hence, by the facts Uσ 6⊆ Qsσ and sσ ∈ Uσ for every σ ∈ Λ, we have ϕαβ(sα) = sβ for α ≥ β. On the other hand, it is easy to see that Psσ ⊆ Sσ since the set Uσ is a subset of Sσ for every σ ∈ Λ, and so Ps = ∏ σ∈Λ Psσ ⊆ ∏ σ∈Λ Sσ. Also, if recall the equality ϕαβ(sα) = ϕαβ(sα) for sα ∈ Uα and α ≥ β, then we have ϕαβ(sα) = sβ due to c© AGT, UPV, 2018 Appl. Gen. Topol. 19, no. 1 118 Some categorical aspects of the inverse limits in ditopological context the fact that ϕαβ(sα) = sβ for sα ∈ Uα and α ≥ β. Thus, by the definition of inverse limit, s = {sα} ∈ S∞ and it is a contradiction. Accordingly, now let us show that lim ← {Uα} is a jointly closed subspace of S∞: Take a point s = {sα} ∈ S∞ such that s /∈ lim ← {Uα}. In this case, because of the fact that s /∈ ∏ α∈Λ Uα there exists an element σ ∈ Λ such that sσ /∈ Uσ. Thus, s /∈ µ−1 σ∈Λ[Uσ] by the equality µσ(s) = sσ and in view of the fact that Uσ is jointly closed in Sσ, the subset µ −1 σ [Uσ] ⊆ S∞ is jointly closed in S∞ due to the continuity of limiting projection µσ : S∞ → Sσ as given in [18, Proposition 4.4]. Now, we can prove that lim ← {Uσ} ⊆ µ −1 σ [Uσ]: If there exists a point a = {aα} ∈ S∞ such that lim ← {Uσ} 6⊆ Qa and Pa 6⊆ µ −1 σ [Uσ] then a ∈ lim ← {Uσ} and so a ∈ ∏ σ∈Λ Uσ. But also, the fact aσ = µσ(a) /∈ Uσ gives a contradiction. As a result of the above considerations lim ← {Uα} is a jointly closed subspace of S∞. In addition, now we will show that U = lim ← {Uα}: (2) First of all, let ’s prove the inclusion U ⊆ lim ← {Uα}. Conversely, if U 6⊆ lim ← {Uα}, then there exists b ∈ S∞ = lim ← {Sα} such that U 6⊆ Qb and Pb 6⊆ lim ← {Uα}. In this case, Pbα ⊆ Uα because of µα(b) ∈ µα(U). Thus, Pb = ∏ α∈Λ Pbα ⊆ ∏ α Uα∈Λ. On the other hand, b ∈ ∏ α Sα∈Λ and ϕαβ(bα) = bβ for α ≥ β, α,β ∈ Λ. Also, by the definition of ϕαβ for α ≥ β and the fact bα ∈ Uα for every α ∈ Λ, the equality ϕαβ(bα) = ϕαβ(bα) is satisfied. Hence, ϕαβ(bα) = bβ for α ≥ β. That is, we obtained b ∈ lim ← {Uα} which is a contradiction. Therefore, from (1) if recall the fact that lim ← {Uα} is jointly closed with respect to the limit ditopology (τ∞,κ∞) on (S∞,S∞), then the inclusion U ⊆ lim ← {Uα} is immediate. For the other direction, assume lim ← {Uα} 6⊆ U. Thus, there exists a point a = {aα} ∈ S∞ such that lim ← {Uα} 6⊆ Qa and Pa 6⊆ U. By the definition of joint topology, there exist M ∈ µ(a) and N ∈ η(a) such that U ⊆ N ∩ (S∞ \ M) and so we have the sets G ∈ τ∞ and K ∈ κ∞ such that Pa ⊆ G ⊆ M, N ⊆ K ⊆ Qa and U ⊆ K ∩ (S∞ \ G). Hence, by [18, Theorem 4.6], there exist α0,α1 ∈ Λ and Aα0 ∈ τα0, Bα1 ∈ κα1 such that the conditions Pa ⊆ µ−1α0 [Aα0] ⊆ G and K ⊆ µ −1 α1 [Bα1] ⊆ Qa are satisfied. In this case, the inclusion U ⊆ (S∞\µ −1 α0 [Aα0])∩µ −1 α1 [Bα1] is trivial. Finally, we obtained α1 ∈ Λ satisfying the conditions U ⊆ U ⊆ µ−1α1 [Bα1] and Pa 6⊆ µ −1 α1 [Bα1]. Thus Uα1 ⊆ Bα1 for α1 ∈ Λ, because of the inclusions µα1(U) ⊆ µα1(µα1 −1[Bα1]) ⊆ Bα1. If c© AGT, UPV, 2018 Appl. Gen. Topol. 19, no. 1 119 F. Yıldız we consider the closure operator on these sets, it is clear that Uα1 ⊆ Bα1 and so µα1(Pa) 6⊆ Uα1 by µα1(a) /∈ Bα1. Moreover, it is easy to verify that µα1(Pa) = Paα1 : µα1(Pa) = {µα1(x) | x ∈ Pa} = {xα1 | x ∈ Pa} = {xα1 | x ∈ ∏ α∈Λ Paα} = {xα1 | xα ∈ Paα, ∀α} = Paα1 . As a result of these facts, we have Paα1 6⊆ Uα1 and so aα1 /∈ Uα1 for α1 ∈ Λ. This argument gives a /∈ ∏ α Uα∈Λ, clearly. It means that a /∈ lim ← {Uα} and so, a contradiction. (3) Note that the closure set Uα is jointly closed in the space Sα for each α. Thus, the sets µ−1α [Uα], α ∈ Λ are jointly closed in the limit space S∞ since the limiting projection µα is continuous for α ∈ Λ, between the cor- responding joint topological spaces (S∞,Jτ∞κ∞), (Sα,Jτακα) of the spaces (S∞,S∞,τ∞,κ∞),(Sα,Sα,τα,κα) ∈ Ob ifPDitop, respectively. In addition, with the equality µα(U) = Uα , α ∈ Λ given in the hypothesis, it is clear that U ⊆ µ−1α [Uα] and so U ⊆ ⋂ α∈Λ µ−1α [Uα]. Hence we have U ⊆ ⋂ α∈Λ µ−1α [Uα] since ⋂ α∈Λ µ−1α [Uα] is jointly closed in S∞. For the converse, suppose that ⋂ α∈Λ µ−1α [Uα] 6⊆ U. In this case, there exists a point a = {aα} ∈ S∞ such that ⋂ α∈Λ µα −1[Uα] 6⊆ Qa and Pa 6⊆ U. Thus, a ∈ µα −1[Uα] and µα(a) = aα ∈ Uα for every α ∈ Λ. On the other hand, if Pa 6⊆ U and U is closed in S∞ with respect to the joint topology of the ditopology on (S∞,S∞) ∈Ob ifPTex, then there exist M ∈ µ(a) and N ∈ η(a) such that U ⊆ N ∩ (S∞\M). So we have the sets G ∈ τ∞, K ∈ κ∞ such that G ⊆ M, N ⊆ K and U ⊆ K ∩ (S∞\G). Therefore, by [18, Theorem 4.6] there exist α0,α1 ∈ Λ and Aα0 ∈ τα0, Bα1 ∈ κα1 satisfying the conditions µ−1α0 [Aα0 ] ⊆ G, µ −1 α0 [Aα0] 6⊆ Qa and K ⊆ µ −1 α1 [Bα1], Pa 6⊆ µ −1 α1 [Bα1]. In this case, the inclusion U ⊆ (S∞\µ −1 α0 [Aα0]) ∩ µ −1 α1 [Bα1] is trivial and so, we have Uα1 ⊆ Bα1 for α1 ∈ Λ by U ⊆ µ −1 α1 [Bα1]. Consequently, Uα1 ⊆ Bα1 and the fact that µα1(a) = aα1 /∈ Bα1 means that aα1 /∈ Uα1 which is a contradiction. � With the above notations, we have also the next result: Corollary 4.5. i) U ⊆ lim ← {Uα} ii) lim ← {Uα} ⊆ lim ← {Uα} Proof. i) If the inclusion is not true, there exists a point a = {aα} ∈ S∞ such that U 6⊆ Qa and Pa 6⊆ lim ← {Uα}. In this case, by the fact µα(a) ∈ µα(U) = Uα we have aα ∈ Uα for every α ∈ Λ and so a ∈ ∏ α∈Λ Uα, obviously. Also, we have ϕαβ|Uα(aα) = ϕαβ(aα) = aβ since aα ∈ Uα, α ∈ Λ. As a result c© AGT, UPV, 2018 Appl. Gen. Topol. 19, no. 1 120 Some categorical aspects of the inverse limits in ditopological context of these considerations, we get a = {aα} ∈ lim ← {Uα} which contradicts with Pa 6⊆ lim ← {Uα}. ii) Firstly, note that the limit sets lim ← {Uα} and lim ← {Uα} are subsets of S∞, due to Proposition 4.2 and Theorem 4.4. Now assume the converse of required inclusion. Thus, there exists a point s = {sα} ∈ S∞ such that lim ← {Uα} 6⊆ Qs and Ps 6⊆ lim ← {Uα}. In this case, s = {sα} ∈ ∏ α∈Λ Uα and so ϕαβ|Uα(sα) = sβ for α ≥ β, α,β ∈ Λ because of sα ∈ Uα. Hence, s = {sα} ∈ ∏ α∈Λ Uα by Uα ⊆ Uα. Also, for α ≥ β, we have the equalities ϕαβ(sα) = ϕαβ|Uα(sα) = ϕαβ(sα) and ϕαβ(sα) = ϕαβ|Uα(sα) due to sα ∈ Uα. Consequently, the point s = {sα} ∈∏ α∈Λ Uα is also an element of the inverse limit set lim ← {Uα} since we have the equality ϕαβ(sα) = ϕαβ|Uα(sα) = sβ for α ≥ β, and it is a contradiction. � According to all considerations presented above, we can mention a further result as the final stage of this section, besides the fact that it will be considered as the converse of Proposition 4.3. Corollary 4.6. Let the system {(Sα,Sα,τα,κα),ϕαβ}α≥β ∈ Ob InvifPDitop over a directed set Λ. If take the ditopological subtexture space (U,SU,τU,κU) ∈ Ob ifPDitop of the inverse limit space (S∞,S∞,τ∞,κ∞) where U ∈ J c τ∞κ∞ , then (U,SU,τU,κU) ∈ Ob ifPDitop is the inverse limit space of the inverse system {(Uα,SUα,τUα,κUα),ϕαβ}α≥β ∈ Ob InvifPDitop consisting of jointly closed subspaces (Uα,SUα,τUα,κUα) of the spaces (Sα,Sα,τα,κα) ∈ Ob ifPDitop, where πα|S∞(U) = µα(U) = Uα, SUα = Sα|Uα, τUα = τα|Uα, κUα = κα|Uα, α ∈ Λ and ϕαβ = ϕαβ|Uα, for α,β ∈ Λ such that α ≥ β. In other words, if U ∈ Jcτ∞κ∞ then we have the equality U = lim ← {Uα} = lim ← {Uα}. Proof. If choose the set U as an element of Jcτ∞κ∞, that is a closed set with respect to the joint topology of the limit ditopology (τ∞,κ∞) defined on the inverse limit texture, then by Theorem 4.4 (2) and the two inclusions presented in Corollary 4.5, the required equalities are straightforward. � 5. Identification of the ditopological products as an inverse limit in ifPDitop Take into account all the previous considerations, it can be mentioned that the notion of inverse limit as an object of ifPDitop for any inverse system which is the object of InvifPDitop is derived from the products as the objects of ifPDitop. Conversely, by applying the limit operation lim ← located in the theory of inverse systems, to the objects of InvifPDitop, one can express infinite ditopo- logical cartesian products [3, 4, 18] of the spaces which are the objects of c© AGT, UPV, 2018 Appl. Gen. Topol. 19, no. 1 121 F. Yıldız ifPDitop in terms of the finite cartesian products of those spaces belong to Ob ifPDitop. Now, let ’s mention and prove this significant characterization as a theorem: Theorem 5.1. For a directed set Λ and any family {(Xs,Ss,τs,κs)}s∈Λ of the objects in ifPDitop, the product space ( ∏ s∈Λ Xs, ⊗ s∈Λ Ss, ⊗ s∈Λ τs, ⊗ s∈Λ κs) ∈ Ob ifPDitop may be expressed as the inverse limit of an inverse system over Γ, which is the object of InvifPDitop and constructed by the finite cartesian product spaces ( ∏ s∈I Xs, ⊗ s∈I Ss, ⊗ s∈I τs, ⊗ s∈I κs) ∈ Ob ifPDitop for I ∈ Γ, where the set Γ = {I ⊆ Λ | I is finite} is directed by the set inclusion. In other words, Any arbitrary textural product of the objects in ifPDitop is exactly the in- verse limit space of the inverse system consisting of finite products of those objects. Proof. Let (Xs,Ss,τs,κs) ∈ Ob ifPDitop, s ∈ Λ and Γ be directed by the set inclusion, that is J ≤ I ⇐⇒ J ⊆ I for every I,J ∈ Γ. Now assume J ≤ I for any J ∈ Γ. If x = {xs}s∈I ∈ ∏ s∈I Xs = XI then xs ∈ Xs for all s ∈ I. In this case, {xs}s∈J ∈ ∏ s∈J Xs = XJ by the facts that if s ∈ J then s ∈ I and xs ∈ Xs for all s ∈ I. Therefore, for J ≤ I, describe the mapping ϕIJ : XI → XJ {xs}s∈I 7→ {xs}s∈J. Now let us prove that ϕIJ is ω-preserving and bicontinuous for J ≤ I : Assume that P{xs}s∈I 6⊆ Q{x′s}s∈I for {xs},{x ′ s} ∈ XI. If s0 ∈ J, then s0 ∈ I by J ≤ I. Thus, Pxs0 6⊆ Qx′s0 and P{xs}s∈J 6⊆ Q{x′s}s∈J by [18, Corollary 1.2], since Pxs 6⊆ Qx′s for all s ∈ J. Hence PϕIJ (x) 6⊆ QϕIJ (x′) and ϕIJ is ω-preserving. For the second part, we prove that ϕIJ is bicontinuous between the product ditopological spaces (XI,SI,τI,κI) and (XJ,SJ,τJ,κJ) as follows: Suppose that J = {1,2, ...,m}, I = {1,2, ..., t} and J ⊆ I. In this case, m < t. Now let G ∈ ⊗ s∈J τs = τJ and ϕ −1 IJ [G] 6⊆ Qx for x = {xs}s∈I ∈ XI. In this case, ϕIJ(x) = {xs}s∈J ∈ G, that is G 6⊆ Q{xs}s∈J . Thus, there exists B ∈ BτJ , where BτJ denotes the base for τJ, such that B 6⊆ Q{xs}s∈J and B ⊆ G, so there exists finite set J0 = {1,2, ...,n} ≤ J (n < m) such that B = ⋂ j∈J0 (πJj ) −1[Gj], where Gj ∈ τj, j ∈ J0. Thus, xj ∈ Gj for j ∈ J0. Also, ϕ−1IJ [B] = ϕ −1 IJ ( ⋂ j∈J0 (πj J)−1[Gj]) ⊆ ϕ −1 IJ [G] because of B ⊆ G. Thus, with all the arrangements, we have B′ = ⋂ j∈J0 (πJj ◦ϕIJ) −1[Gj] = ⋂ j∈J0 (πIj ) −1[Gj] ⊆ ϕ−1 IJ [G] and so B′ ∈ BτI where BτI denotes the base for τI. c© AGT, UPV, 2018 Appl. Gen. Topol. 19, no. 1 122 Some categorical aspects of the inverse limits in ditopological context On the other hand, we have {xs}s∈J ∈ (π J j ) −1[Gj] since {xs}s∈J ∈ B and πJj ({xs}s∈J) ∈ Gj for every j ∈ J0 . Thus x1 ∈ G1, x2 ∈ G2,...,xn ∈ Gn. In this case, by the fact n < t, πIj ({xs}) ∈ G for j ∈ J0 and so {xs}s∈I ∈ B ′. Since, ϕ−1 IJ [G] ∈ τI for G ∈ τJ, ϕIJ is continuous. Dually, by using closed sets as the elements of κJ, it is proved that ϕIJ is cocontinuous and so bicontinuous. Furthermore, note that the mappings ϕIJ for J ≤ I are the bonding maps: Indeed, for the mapping ϕII : XI → XI, the equality ϕII({xs}s∈I) = {xs}s∈I is clear and so ϕII is the identity idXI . In addition, for K ≥ I ≥ J, let ’s prove ϕIJ ◦ ϕKI = ϕKJ. If {xs}s∈K ∈ XK then (ϕIJ ◦ ϕKI)({xs}s∈K) = ϕIJ(ϕKI({xs}s∈K)) = ϕIJ({xs}s∈I) = {xs}s∈J = ϕKJ({xs}s∈K). Consequently, thanks to the above expressions, the fact {(XI,SI,τI,κI),ϕIJ}I≥J ∈ Ob InvifPDitop is trivial. Now let us turn to our main aim: The inverse limit space of inverse system {(XI,SI,τI,κI),ϕIJ}I≥J ∈ Ob InvifPDitop over Γ is ifPDitop-isomorphic to the arbitrary ditopological product space constructed on the set ∏ s∈Λ Xs. For proof, first of all we define a mapping between lim ← {XI}I∈Γ and ∏ s∈Λ Xs: If {xI} ∈ lim ← {XI}I∈Γ then {xI} ∈ ∏ I∈Γ XI and so xI ∈ XI for every I ∈ Γ. Now, for any s ∈ Λ let Is = {s} ∈ Γ, so by the fact XIs = ∏ z∈Is Xz = ∏ z∈{s} Xz = Xs, we have x{s} = xIs ∈ Xs, s ∈ Λ. Thus {xIs} ∈ ∏ s∈Λ Xs and finally, we can define the mapping ψ : lim ← {XI}I∈Γ → ∏ s∈Λ Xs {xI}I∈Γ 7→ {xIs}s∈Λ It is easy to verify that ψ is well-defined. Now let us show that ψ is an ifPDitop-isomorphism: ψ is ω-preserving: Let {xI},{x ′ I} ∈ lim ← {XI}I∈Γ such that P{xI} 6⊆ Q{x′I}. In this case, PxI 6⊆ Qx′I for all I ∈ Γ, by [18, Corollary 1.2]. Take s ∈ Λ, so Is = {s} ⊆ Λ, that is Is ∈ Γ. Thus, PxIs 6⊆ Qx′Is by the fact that PxI 6⊆ Qx′I for all I ∈ Γ. It means that PxIs 6⊆ Qx′Is for all s ∈ Λ. Hence P{xIs}s∈Λ 6⊆ Q{x′ Is } s∈Λ , that is Pψ{xI} 6⊆ Qψ{x′I}. In addition, the bijectivity of ψ is straightforward. Now, if consider the product ditopological spaces (XI,SI,τI,κI) for I ∈ Γ, with the plain texturings then the product texturing ⊗ I∈Γ SI and product ditopology ( ⊗ I∈Γ τI, ⊗ I∈Γ κI) can be constructed over the product set ∏ I∈Γ XI in a suitable way. Therefore, the restricted texturing and ditopology will be taken over the subset lim ← {XI}I∈Γ of ∏ I∈Γ XI. Shortly, if we use the notations T = c© AGT, UPV, 2018 Appl. Gen. Topol. 19, no. 1 123 F. Yıldız ( ⊗ I∈Γ SI)|lim ← {XI}I∈Γ, V = ( ⊗ I∈Γ τI)|lim ← {XI}I∈Γ and Z = ( ⊗ I∈Γ κI)|lim ← {XI}I∈Γ for the induced texturing, topology and cotopology, respectively, then now we will prove that ψ is bicontinuous with respect to the ditopologies ( ⊗ s∈Λ τs, ⊗ s∈Λ κs) and (V,Z): Let G ∈ ⊗ s∈Λ τs = τΛ and ψ −1[G] 6⊆ Q{xI}I∈Γ. In this case, G 6⊆ Qψ({xI}I∈Γ) and so G 6⊆ Q{xIs}s∈Λ. Thus, there exists B ∈ BτΛ which is the base for the product topology τΛ, such that B ⊆ G and B 6⊆ Qψ({xI}I∈Γ). Note here that B = ⋂ j∈J0⊆Λ π−1j [Gj], where Gj ∈ τj and j ∈ J0 for the finite set J0 ⊆ Λ. Thus, we have ψ−1( ⋂ j∈J0⊆Λ π−1j [Gj]) ⊆ ψ −1[G] and so ⋂ j∈J0 (πj ◦ ψ) −1[Gj] ⊆ ψ −1[G]. On the other hand, the equality πj ◦ ψ = πIj |lim ← {XI}I∈Γ is obvious by the definition of projection map πIj : ∏ I∈Γ XI → XIj = Xj and by the facts j ∈ Λ and Ij = {j} ⊆ Λ which means that Ij ∈ Γ for j ∈ J0. Additionally, if take ϕ as the inverse of ψ, then we have πIj |lim ← {XI}I∈Γ ◦ ϕ = πj. Here, the restriction πIj |lim ← {XI}I∈Γ is bicontinuous since Ij. projection map πIj is bicontinuous. Hence, if A = ⋂ j∈J0 (πj ◦ ψ) −1[Gj] = ⋂ j∈J0 (πIj |lim ← {XI}I∈Γ) −1[Gj] then A ∈ BV. Here, BV denotes the base for topology V. In this case, the fact A ⊆ ψ −1[G] is clear. Now let us prove A 6⊆ Q{xI}I∈Γ: Firstly, recall B 6⊆ Q{xIs}s∈Λ and so π−1j [Gj] 6⊆ Q{xIs}s∈Λ for all j ∈ J0. That is, πj({xIs}) ∈ Gj and (πj ◦ ψ)({xI}I∈Γ) ∈ Gj for all j ∈ J0. Therefore, {xI}I∈Γ ∈ (πIj |lim ← {XI}I∈Γ) −1[Gj] is clear for all j ∈ J0. Finally, {xI}I∈Γ ∈ ⋂ j∈J0 (πIj |lim ← {XI}I∈Γ) −1[Gj] = A, and so A 6⊆ Q{xI}I∈Γ since the related texturings are plain. Hence ψ −1[G] ∈ V and ψ is continuous. Dually, it is easy to verify that ψ is cocontinuous by dealing with the closed sets. Then ψ is bicontinuous. As the final step, that the map ϕ as the inverse of ψ is bicontinuous can be shown in a like manner. � The above theorem could be also summarized for the subcategory ifPDicomp2 consisting of dicompact [11] and bi-T2 (bi-Hausdorff) [4] objects of the category ifPDitop. Hence, with the above arguments, note that: Corollary 5.2. The infinite ditopological products of the objects which belong to ifPDicomp2 can be expressed via inverse limits, in terms of the finite di- topological products in ifPDicomp2 of those objects. Proof. For all the details about category of dicompact spaces see [11], and from [4], note that (S,S,τ,κ) is bi-T2 if and only if for s,t ∈ S, Qs 6⊆ Qt =⇒ ∃H ∈ τ, K ∈ κ with H ⊆ K, Ps 6⊆ K and H 6⊆ Qt. Thus, the required c© AGT, UPV, 2018 Appl. Gen. Topol. 19, no. 1 124 Some categorical aspects of the inverse limits in ditopological context characaterization is seen as a result of Theorem 5.1. Indeed, by the facts that the jointly closed subtexture spaces and the product spaces of dicompact, bi- T2 ditopological spaces are dicompact and bi-T2 from [18, Theorem 4.16] and Tychonoff property, respectively, and from [18, Theorem 4.17 a)], the proof is completed. � Definition 5.3. A property P is called ditopological property if it is a property defined for ditopological texture spaces, as a natural counterpart of the classical notion, named topological property. According to this, we have the following as a final result, as well. Corollary 5.4. Let P be a ditopological property which is hereditary with re- spect to the jointly closed subsets of a ditopological space and finitely multi- plicative (that is, P is preserved under the finite multiplications of ditopological spaces). In this case, (S,S,τ,κ) ∈ Ob ifPDitop is ifPDitop-isomorphic to the inverse limit of an inverse system constructed over a directed set Λ, via bi-T2 spaces (Sα,Sα,τα,κα) ∈ Ob ifPDitop, α ∈ Λ, which have the property P if and only if (S,S,τ,κ) is ifPDitop-isomorphic to a jointly closed subspace of the product space ( ∏ α∈Λ Sα, ⊗ α∈Λ Sα, ⊗ α∈Λ τα, ⊗ α∈Λ κα). Proof. Necessity. Suppose that (S,S,τ,κ) ∈ Ob ifPDitop is isomorphic to the inverse limit space (S∞,S∞,τ∞,κ∞) ∈ Ob ifPDitop of the inverse system {(Sα,Sα,τα,κα),ϕαβ}α≥β ∈ Ob InvifPDitop over a directed set Λ, where S∞ = lim ← {Sα}. Also, if recall that the inverse limit space S∞ is jointly closed in the product ∏ α∈Λ Sα by [18, Theorem 4.17 a)], then the required assertion is proved. Sufficiency. Let {(Sα,Sα,τα,κα)}α∈Λ be a family consisting of the objects in ifPDitop, which have the properties bi-T2 and P . Assume that (S,S,τ,κ) is ifPDitop-isomorphic to a jointly closed subspace (U,( ⊗ Sα)|U,( ⊗ τα)|U,( ⊗ κα)|U) of the product space ( ∏ Sα, ⊗ Sα, ⊗ τα, ⊗ κα). By Theorem 5.1, it is known that the product ∏ Sα can be expressed as the inverse limit of an inverse system consisting of finite cartesian product spaces n∏ i=1 Si for n ∈ N. Hence, with the same notations used in Corollary 4.6, (U,( ⊗ Sα)|U,( ⊗ τα)|U,( ⊗ κα)|U)) be- comes the inverse limit of inverse system A = {(Un,( n⊗ i=1 Si)|Un,( n⊗ i=1 τi)|Un,( n⊗ i=1 κi)|Un),ϕnm}n≥m constructed by the bonding maps ϕnm : Un → Um for n ≥ m, as well as the jointly closed subspaces (Un,( n⊗ i=1 Si)|Un,( n⊗ i=1 τi)|Un,( n⊗ i=1 κi)|Un) of finite cartesian product spaces ( n∏ i=1 Si, n⊗ i=1 Si, n⊗ i=1 τi, n⊗ i=1 κi) for every n ∈ N. Here, Un denotes the clo- sure of Un for each n, with respect to the joint topology appropriate for the finite product space of the spaces (Si,Si,τi,κi), i = 1,2, ...,n. On the other hand, since each space Sα, α ∈ Λ has the property bi-T2 from [4], the product space ∏ Sα has the property bi-T2 and so the ditopologies on c© AGT, UPV, 2018 Appl. Gen. Topol. 19, no. 1 125 F. Yıldız subsets Un have the property bi-T2 as well. Furthermore, each finite product space n∏ i=1 Si has the property P since each space Sα, α ∈ Λ has the property P by hypothesis. Thus, the jointly closed subspaces Un, n ∈ N have the common property P as P is hereditary with respect to the jointly closed subspaces. Consequently, A is the required inverse system in ifPDitop and by the fact lim ← A = U, the proof is concluded. � 6. Conclusion This paper studied some further categorical aspects of the inverse systems (projective spectrums) and inverse limits constructed in the subcategory if- PDitop of ditopological plain spaces. As one of the investigations here, an identity natural transformation which is peculiar to the theory of inverse systems and inverse limits, as well as consisting of the adjoint and isomorphism functors introduced between the suitable related main subcategories of Bitop and ifPDitop, consisting of the spaces which satisfy a special separation axiom, is established. As another one, we proved a representation theorem which shows any infinite textural product of the objects in category ifPDitop can be expressed as the inverse limit of the inverse system in InvifPDitop, constructed by the finite products of those objects in ifPDitop. Besides that, the textural products of dicompact bi-T2 ditopological spaces are characterized in terms of finite products, via inverse limits. There are considerable difficulties involved in constructing a suitable theory of inverse systems for general ditopological spaces. Hence, we confined our attention to the inverse systems - limits constructed in the special category ifPDitop and we leaved as an open problem the task of extending the further results obtained here to more general categories established in the theory of ditopological spaces. Acknowledgements. The author would like to thank the referees and edi- tors for their constructive comments which have helped improve the exposition and the readability of the paper. References [1] J. Adámek, H. Herrlich and G. E. Strecker, Abstract and Concrete Categories (John Wiley & Sons, Inc., 1990). Volume 17, Springer-Verlag, 1990. [2] L. M. Brown and M. Diker, Ditopological texture spaces and intuitionistic sets, Fuzzy Sets and Systems 98 (1998), 217–224. [3] L. M. Brown, R. Ertürk, Fuzzy sets as texture spaces, I. Representation theorems, Fuzzy Sets and Systems 110, no. 2 (2000), 227–236. [4] L. M. Brown, R. Ertürk, Ş. Dost, Ditopological texture spaces and fuzzy topology, III. Separation Axioms, Fuzzy Sets and Systems 157, no. 14 (2006), 1886–1912. c© AGT, UPV, 2018 Appl. Gen. Topol. 19, no. 1 126 Some categorical aspects of the inverse limits in ditopological context [5] S. Eilenberg and N. Steenrod, Foundations of Algebraic Topology, Princeton, New Jersey, Princeton University Press, 1952. [6] R. Engelking, General Topology (Helderman Verlag Berlin, 1989). [7] G. Gierz, K. H. Hofmann, K. Keimel, J. D. Lawson, M. Mislove and D. S. Scott, A compendium of continuous lattices (Springer–Verlag, Berlin, 1980). [8] S. Özçağ, F. Yıldız and L. M. Brown, Convergence of regular difilters and the com- pleteness of di-uniformities, Hacettepe Journal of Mathematics and Statistics 34 (2005), 53–68. [9] İ. U. Tiryaki and L. M. Brown, Plain ditopological texture spaces, Topology and its Applications 158 (2011), 2005-2015. [10] F. Yıldız and L. M. Brown, Characterizations of real difunctions, Hacettepe Journal of Mathematics and Statistics 35, no. 2 (2006), 189–202. [11] F. Yıldız and L. M. Brown, Categories of dicompact bi-T2 texture spaces and a Banach- Stone theorem, Quaestiones Mathematicae 30 (2007), 167–192. [12] F. Yıldız and L. M. Brown, Real dicompact textures, Topology and its Applications 156, no. 11 (2009), 1970–1984. [13] F. Yıldız and L. M. Brown, Dicompleteness and real dicompactness of ditopological texture spaces, Topology and Its Applications 158, no. 15 (2011), 1976–1989. [14] F. Yıldız and L. M. Brown, Extended real dicompactness and an application to Hutton spaces, Fuzzy Sets and Systems 227 (2013), 74–95. [15] F. Yıldız, Connections between real compactifications in various categories, Quaestiones Mathematicae 38 , no. 3 (2015),31–455. [16] F. Yıldız, Completeness types for uniformity theory on textures, Filomat 29, no. 1 (2015), 159–178. [17] F. Yıldız, Inverse systems and inverse limits in the category of plain textures, Topology and Its Applications 201 (2016), 217–234. [18] F. Yıldız, Inverse systems and limits in the category of ditopological plain spaces, Topol- ogy and Its Applications 228 (2017), 47–67. c© AGT, UPV, 2018 Appl. Gen. Topol. 19, no. 1 127