Abstract
Given a continuum X, Cn(X) denotes the hyperspace of nonempty closed subsets of X with at most n components. A strong size level is a subset of the form σ−1(t), where σ is a strong size map for Cn(X) and t ∈ (0, 1]. In this paper, answering a question by Capulín-Pérez, Fuentes-Montes de Oca, Lara-Mejía and Orozco-Zitli, we prove that for each n ≥ 2, no strong size level for Cn(X) is irreducible.
1 Introduction
A continuum is a compact connected metric space with more than one point, a subcontinuum of X is a compact connected subspace of X, so subcontinua include one-point sets.
In this paper we use the following hyperspaces of X:
All these hyperspaces are considered with the Hausdorff metric [6: Theorem 3.1]. We write C(X) instead of C1(X). A map is a continuous function.
A continuum X is irreducible provided that there exist points p, q ∈ X such that no proper subcontinuum of X contains both p and q.
A Whitney map for Cn(X) is a map μ : Cn(X) → [0, 1] such that:
for each p ∈ X, μ({p}) = 0;
μ (X) = 1; and
if A, B ∈ Cn(X) and A⊊ B, then μ(A) < μ(B).
Sets of the form μ−1(t) where t ∈ (0, 1) are Whitney levels.
A strong size map for Cn(X) is a map σ : Cn(X) → [0, 1] such that:
σ(A) = 0 for each A ∈ Fn(X);
σ(X) = 1; and
if A, B ∈ Cn(X), A ⊊ B and B ∉ Fn(X), then σ(A) < σ(B).
Sets of the form σ−1(t) where t ∈ [0, 1) are strong size levels.
Notice that the restriction of a strong size map to C(X) is a Whitney map. Whitney levels and strong size levels are continua ([6: Theorem 19.9] and [3: Theorem 2.10]). Hosokawa [3] introduced and studied topological properties of strong size levels. Topological properties of Whitney levels have been studied widely. In [6: Chapter VIII]{in}, there is a very complete account of what was known on Whitney levels up to 1998.
A very important line of research about Whitney and strong size levels is to determine for which topological properties 𝓟 the following implications hold:
if X has property 𝓟, then Whitney (strong size) levels have property 𝓟, and
if Whitney (strong size) levels have property 𝓟, then X has property 𝓟.
Statements (a) and (b) have been studied for more than 40 years (see [6: Chapter VIII]). The study of strong size levels started with the paper [3] in 2011. Since then statements (a) and (b) have been studied for strong size levels in [1, 2, 3, 4, 5] and [7, 8, 9].
For the case that property 𝓟 is irreducibility and n = 1, in [4], it was shown that if X is a continuum having a Whitney level irreducible with respect to a finite subset, then X is also irreducible with respect to a finite set; and in [9], it was shown an irreducible, hereditarily decomposable continuum such that none of its proper strong size levels is irreducible. In [1], the authors asked whether there exists a continuum with irreducible strong size levels. In this paper we give a complete solution to this question for the case n ≥ 2 with the following theorem.
Theorem 1
Let X be a continuum and n ≥ 2. Then the strong size levels for Cn(X) are not irreducible.
2 Proof of Theorem 1
Given elements A, B ∈ 2X such that A ⊊ B, an order arc from A to B is a map α : [0, 1] → 2X such that α(0) = A, α(1) = B and if 0 ≤ s < t ≤ 1, then α (s) ⊊ α (t). The most important result about order arcs says that if A, B ∈ 2X, then there exists an order arc from A to B if and only if A ⊊ B and each component of B intersects A [6: Theorem 15.3].
Proof of Theorem 1
Take a strong size level 𝓐 = σ−1(t0), where σ : Cn(X) → [0, 1] is a strong size map and t0 ∈ [0, 1). We will use the following claim.
Claim 2
Let D, E ∈ 𝓐. Suppose that E = F ∪ G, G is a component of E, F ∩ G = ∅; L is a nonempty finite subset of F, with |L| equal to the number of components of F; α : [0, 1] → Cn(X) is an order arc from L to D such that F ∈ α([0, 1]) and q is a point of G. Then there exists a subcontinuum 𝓑 of 𝓐 such that {D, E} ⊂ 𝓑 and for each K ∈ 𝓑, either each component of K intersects L ∪ {q} or K = K0 ∪ {p}, where each component of K0intersects L, and p ∈ X.
Proof
Since E ∈ Cn(X) and G is component of E, we have that E has more components than F. Then |L| < n. Since α is an order arc, for each t ∈ (0, 1], there exists an order arc from L to α(t), then each component of α (t) has a point of L, so α(t) has at most |L| components. In particular, D has less components than E. Since α is an order arc from L to D, we have that L is a proper subset of D and each component of D intersects L. This implies that D is not a finite set. In particular, σ(D) = t0 > 0. Let tF ∈ [0, 1] be such that α(tF) = F. If E ⊂ D, since σ(E) = σ(D), we have that E = D, a contradiction. Therefore E is not contained in D.
We construct two subcontinua 𝓑1 and 𝓑2. In the case G = {q}, we define 𝓑1 = {F ∪ {q}} = {E}.
Now, suppose that {q} ⊊ G. Take an order arc β : [0, 1] → C(X), from {q} to G.
Given u ∈ [0, 1], β(u) ∪ α(0) = β(u) ∪ L ⊂ G ∪ F = E, so σ(β(u) ∪ α(0)) ≤ σ(E) = t0. On the other hand, β(u) ∪ α(1) = β(u) ∪ D. This implies that σ(β(u) ∪ α(1)) ≥ t0. This implies that there exists v(u) ∈ [0, 1] such that σ(β(u) ∪ α(v(u))) = t0.
Define g : [0, 1] → 𝓐 by g(u) = β(u) ∪ α(v(u)). First, we show that the definition of g does not depend on the choice of v(u). Suppose that v1, v2 ∈ [0, 1] satisfy that σ(β (u) ∪ α(v1)) = σ(β (u) ∪ α(v2)). We may assume that v1 < v2. Since α is an order arc, we have that α(v2) is not a finite set. Since β (u) ∪ α(v1) ⊂ β (u) ∪ α(v2) and σ(β (u) ∪ α(v1)) = σ(β (u) ∪ α(v2)), we conclude that β (u) ∪ α(v1) = β (u) ∪ α(v2). Therefore g(u) does not depend on the choice of v(u). We prove that g is continuous. Take a sequence
Define 𝓑1 = g([0, 1]). Then 𝓑1 is a subcontinuum of 𝓐. Given u ∈ [0, 1], {q} ⊂ β(u) and each component of α (v(u)) intersects L, so each component of g(u) intersects L ∪ {q}. Since β(1) = G, we have that g(1) = β (1) ∪ α (v(1)) = G ∪ α(v(1)). In the case that v(1) ≤ tF, we have that α(v(1)) ∪ G ⊂ α(tF) ∪ G = F ∪ G = E. Then g(1) ⊂ E. In the case that tF ≤ v(1), we have that E ⊂ g(1). We have shown that one of the sets E and g(1) is contained in the other. Since the value of σ on both sets is the equal to t0 and t0 > 0, we conclude that E = g(1) ∈ 𝓑1. Set E1 = g(0) = {q} ∪ α (v(0)).
This finishes the construction of 𝓑1.
Given p ∈ X, the set α(0) ∪ {p} = L ∪ {p} is finite, so σ(α(0) ∪ {p}) = 0. On the other hand, α (1) ∪ {p} contains the set D, so σ(α (1) ∪ {p}) ≥ t0. Thus there exists w(p) ∈ [0, 1] such that σ (α (w(p)) ∪ {p}) = t0. Define h : X → 𝓐 by h(p) = α(w(p)) ∪ {p}. Proceeding as we did with g, it is possible to prove that h does not depend on the choice of w(p) and that h is continuous.
Define 𝓑2 = h(X). Then 𝓑2 is a subcontinuum of 𝓐. Since h(q) = α(w(q)) ∪ {q}, h does not depend on w(q) and E1 = {q} ∪ α(v(0)) ∈ 𝓐, we conclude that h(q) = E1. Therefore, E1 ∈ 𝓑2.
Fix a point p0 ∈ L. Then h(p0) = {p0} ∪ α(w(p0)) ⊂ α(w(p0)) ⊂ D. Since σ(h(p0)) = t0 = σ(D) and t0 > 0, we conclude that h(p0) = D. Hence D ∈ 𝓑2.
Define 𝓑 = 𝓑1 ∪ 𝓑2.
In the case that G = {q}, h(q) = α(w(q)) ∪ {q} = α(w(q)) ∪ G. Since F ∈ α([0, 1]), one of the sets F ∪ G or α(w(q)) ∪ G is contained in the other. Since σ takes the value t0 on both sets and t0 > 0, we conclude that F ∪ G = α(w(q)) ∪ G = h(q). Thus E ∈ 𝓑2, 𝓑 = 𝓑2 is a subcontinuum of 𝓐 containing the set {D, E} and each element of 𝓑 is of the form K = K0 ∪ {p}, where K0 ∈ α([0, 1]) and p ∈ X. Notice that each component of K0 intersects L.
In the case that G ≠ {q}, since E1 ∈ 𝓑1 ∩ 𝓑2, we have that 𝓑 is a subcontinuum of 𝓐 that contains the elements D and E and for each K ∈ 𝓑, if K ∈ 𝓑1 each component of K intersects L ∪ {q} and if K ∈ 𝓑2, then K = K0 ∪ {p} for some K0 ∈ α([0, 1]) and p ∈ X, so each component of K0 intersects L. This ends the proof of the claim.□
In order to prove that 𝓐 is not irreducible, take A, B ∈ 𝓐. Suppose that A1, …, Ar are the components of A and B1, …, Bs are the components of B. Fix points p1 ∈ A1, …, pr ∈ Ar and q1 ∈ B1, …, qs ∈ Bs. We consider two cases.
t 0 = 0. For each i ∈ {1, …, r}, Ai = {pi} and for each j ∈ {1, …, s}, Bj = {qj}. Fix two points p0 ≠ q0 in X ∖ (A ∪ B). Let 𝓒 = {{p1} ∪ E : E ∈ Fn−1(X)}. Then 𝓒 is a subcontinuum of 𝓐 such that A, {p1} ∈ 𝓒 and {p0, q0} ∉ 𝓒. Similarly, there exists a subcontinuum 𝓓 of 𝓐 such that B, {q1} ∈ 𝓓 and {p0, q0} ∉ 𝓓. Then 𝓒 ∪ 𝓓 ∪ F1(X) is a subcontinuum of 𝓐 containing {A, B} and not having the element {p0, q0}. Therefore, in this case, 𝓐 is not irreducible.
t 0 > 0. Taking an order arc from a one-point set to X, it is possible to show that there exists a connected element W in 𝓐. Since t0 < 1, the set X ∖ W is infinite. Fix a point z ∈ X ∖ (W ∪ {p1, …, pr} ∪ {q1, …, qs}). Using an order arc from {z} to X, it is possible to construct a subcontinuum M0 of X with the properties: M0 is non-degenerate, M0 ⊂ X ∖ (W ∪ {p1, …, pr} ∪ {q1, …, qs}) and σ(M0) < t0. Since σ(W ∪ M0) > t0 (n ≥ 2) using again an order arc from a one-point set to W it is possible to obtain a subcontinuum N0 of W such that the set T = M0 ∪ N0 belongs to 𝓐. Notice that M0 ∩ N0 = ∅, so T has a component (the continuum M0) not intersecting {p1, …, pr} ∪ {q1, …, qs}.
We are going to show that there exist a subcontinuum S of X and a subcontinuum 𝓒 of 𝓐 such that {A, S} ⊂ 𝓒 and T ∉ 𝓒. In the case that A is connected, set S = A and 𝓒 = {A}. Since all the components of A intersect {p1, …, pr}, by the previous paragraph, we infer that A ≠ T, so T ∉ 𝓒. Suppose then that A is not connected. Then r > 1.
We apply the Claim 2 to F = A1 ∪ ⋯ ∪ Ar−1, G = Ar, E = A, L = {p1, …, pr−1} and q = pr. Taking an order arc from F to X, it is possible to find an element S1 ∈ 𝓐 such that each component of S1 intersects F. Since each component of F intersects L, it is possible to find an order arc α : [0, 1] → Cn(X) such that α(0) = L, α (1) = S1 and F ∈ α([0, 1]). Then there exists a subcontinuum 𝓑 of 𝓐 such that {S1, A} ⊂ 𝓑 and for each K ∈ 𝓑, either each component of K intersects {p1, …, pr} or K = K0 ∪ {p}, where each component of K0 intersects L and p ∈ X.
Since the component M0 of T is not degenerate and it does not intersect {p1, …, pr}, we have that T ∉ 𝓑. Since each component of S1 intersects {p1, …, pr−1}, we obtain that S1 has at most r − 1 components. Suppose that D1, …, Dk are the components of S1. Then we can choose points pl1 ∈ D1, …, plk ∈ Dk.
We have shown that A can be connected by a continuum contained in 𝓐 ∖ {T} with an element S1 ∈ 𝓐 having less components than A. In the case that S1 is connected, let S = S1 and 𝓒 = 𝓑. In the case that S1 is not connected, we can repeat the procedure to obtain an element S2 ∈ 𝓐 and a continuum 𝓑1 contained in 𝓐 ∖ {T} such that {S1, S2} ⊂ 𝓑1, S2 has less components than S1 and each component of S2 intersects {pl1}, …, plk}. If S2 is connected, let S = S2 and 𝓒 = 𝓑1 ∪ 𝓑2. In the case that S2 is not connected, we repeat this process until we obtain a connected set Sm = S and a continuum 𝓒 = 𝓑1 ∪ ⋯ ∪ 𝓑m contained in 𝓐 ∖ {T} such that {A, S} ⊂ 𝓒.
In a similar way, there exist a connected element S0 ∈ 𝓐 and a continuum 𝓒0 contained in 𝓐 ∖ {T} such that {S0, B} ∈ 𝓒0. Define
Since σ| C(X) is a Whitney map for C(X), we have that σ| C(X))−1(t0) is a Whitney level for C(X) and then this set is connected. Thus 𝓒* is a subcontinuum of 𝓐. Since T has two components, T ∉ 𝓒*. Therefore 𝓒* is a proper subcontinuum of 𝓐 containing A and B. Hence 𝓐 is not irreducible. This finishes the proof of Theorem 1.
The work of Dr. Martínez-de-la-Vega is part of her research project during her sabbatical (academic year 2023–2024) supported by the grant PASPA of program DGAPA of UNAM (Universidad Nacional Autónoma de México). This paper was partially supported by the project “Teoría de Continuos e Hiperespacios, dos”(AI-S-15492) of CONACyT and the project “Sistemas Dinámicos discretos y Teoría de Continuos” PAPIIT-IN105624.
Acknowledgement
We wish to thank the referee for his/her careful reading of the paper and suggestions.
-
Communicated by David Buhagiar
References
[1] Capulín, F.—Lara, M. A.—Orozco-Zitli, F.: Sequential decreasing strong size properties, Math. Slovaca 68(5) (2018), 1141–1148.10.1515/ms-2017-0176Suche in Google Scholar
[2] Capulín F.—Fuentes-Montes De Oca, A.—Lara-Mejía, M. A.—Orozco-Zitli, F.: Increasing strong size properties and strong size block properties, Topology Appl. 283 (2020), Art. ID 107339.10.1016/j.topol.2020.107339Suche in Google Scholar
[3] Hosokawa, H.: Strong size levels of Cn(X), Houston J. Math. 37(3) (2011), 955–965.Suche in Google Scholar
[4] Illanes, A.: Irreducible Whitney levels with respect to finite and countable subsets, An. Inst. Mat. Univ. Nac. Autónoma México 26 (1986), 59–64.Suche in Google Scholar
[5] Illanes, A.—Martínez-De-La-Vega, V.—Martínez-Montejano, J.: Problems on hyperspaces of continua, some answers, Topology Appl. 310 (2022), Art. ID 108006.10.1016/j.topol.2022.108006Suche in Google Scholar
[6] Illanes, A.—Nadler, Jr. S. B.: Hyperspaces, Fundamentals and Recent Advances. Monographs and Textbooks in Pure and Applied Math., Vol. 216, Marcel Dekker, Inc. New York and Basel, 1999.Suche in Google Scholar
[7] Macías, S.—Piceno, C.: Strong size properties, Glas. Mat. Ser. III 48(68) (2013), 103–114.10.3336/gm.48.1.10Suche in Google Scholar
[8] Macías, S.—Piceno, C.: More on strong size properties, Glas. Mat. Ser. III 50(70) (2015), 467–488.10.3336/gm.50.2.14Suche in Google Scholar
[9] Paredes-Rivas, L.—Pellicer-Covarrubias, P.: On strong size levels, Topology Appl. 160(13) (2013), 1816–1828.10.1016/j.topol.2013.07.010Suche in Google Scholar
© 2024 Mathematical Institute Slovak Academy of Sciences
This work is licensed under the Creative Commons Attribution 4.0 International License.
Artikel in diesem Heft
- Right algebras in Sup and the topological representation of semi-unital and semi-integral quantales, revisited
- Topological representation of some lattices
- Exact-m-majority terms
- Polynomials whose coefficients are generalized Leonardo numbers
- A study on error bounds for Newton-type inequalities in conformable fractional integrals
- Improved conditions for the distributivity of the product for σ-algebras with respect to the intersection
- Close-to-convex functions associated with a rational function
- Complete monotonicity for a ratio of finitely many gamma functions
- Class of bounds of the generalized Volterra functions
- Some new uniqueness and Ulam–Hyers type stability results for nonlinear fractional neutral hybrid differential equations with time-varying lags
- Existence of positive solutions to a class of boundary value problems with derivative dependence on the half-line
- Solving Fredholm integro-differential equations involving integral condition: A new numerical method
- Bounds of some divergence measures on time scales via Abel–Gontscharoff interpolation
- Weighted 1MP and MP1 inverses for operators
- Non-commutative effect algebras, L-algebras, and local duality
- Operator Bohr-type inequalities
- Some results for weighted Bergman space operators via Berezin symbols
- Lower separation axioms in bitopogenous spaces
- Fisher information in order statistics and their concomitants for Cambanis bivariate distribution
- Irreducibility of strong size levels
Artikel in diesem Heft
- Right algebras in Sup and the topological representation of semi-unital and semi-integral quantales, revisited
- Topological representation of some lattices
- Exact-m-majority terms
- Polynomials whose coefficients are generalized Leonardo numbers
- A study on error bounds for Newton-type inequalities in conformable fractional integrals
- Improved conditions for the distributivity of the product for σ-algebras with respect to the intersection
- Close-to-convex functions associated with a rational function
- Complete monotonicity for a ratio of finitely many gamma functions
- Class of bounds of the generalized Volterra functions
- Some new uniqueness and Ulam–Hyers type stability results for nonlinear fractional neutral hybrid differential equations with time-varying lags
- Existence of positive solutions to a class of boundary value problems with derivative dependence on the half-line
- Solving Fredholm integro-differential equations involving integral condition: A new numerical method
- Bounds of some divergence measures on time scales via Abel–Gontscharoff interpolation
- Weighted 1MP and MP1 inverses for operators
- Non-commutative effect algebras, L-algebras, and local duality
- Operator Bohr-type inequalities
- Some results for weighted Bergman space operators via Berezin symbols
- Lower separation axioms in bitopogenous spaces
- Fisher information in order statistics and their concomitants for Cambanis bivariate distribution
- Irreducibility of strong size levels