Dense Lineable Criterion for Linear Dynamics Alexander Arbieto, Manuel Saavedra Abstract We study Li-Yorke chaos for sequences of continuous linear operators from an F F -space to a normed space. We introduce the D-phenomenon to establish a common dense lineable criterion that encompasses properties such as recurrence, universality, and Li-Yorke chaos. We show that in every infinite-dimensional separable complex Banach space, there exists a sequence of operators with a dense set of irregular vectors but without a dense irregular manifold, and we exhibit a recurrent operator whose set of recurrent vectors is not dense-lineable. This resolves in the negative a question posed by Grivaux et al. “Puesta la mano en el pecho, resuelvo por fin obedecer tus leyes.” — Ramiro Mendoza S., grandfather of the first author, who dedicated his life to poetry. Keywords: Dense lineability, Li-Yorke Chaos, Hypercyclicity, Recurrence, Quasi-rigidity. MSC : 47A16, 37B20, 37B02 1 Introduction Let X be a topological vector space (TVS). A subset A\subset X is said to be lineable if A\cup\{0\} contains an infinite-dimensional vector subspace of X . Moreover, A is called dense lineable if A\cup\{0\} contains a dense vector subspace of X . These concepts were introduced and extensively studied in [ 35 , 2 , 47 ] . For a comprehensive treatment of these and related notions, we refer to the works of L. Bernal et al. [ 8 , 9 ] . We introduce a motivating problem by recalling key aspects of quasi-rigid operators. Recently, the notion of recurrent operators was introduced by Costakis et al. [ 23 , 24 ] . Let X be an F -space. An operator T\in\mathcal{L}(X) is said to be recurrent if the set of recurrent vectors for T , denoted by \mathrm{Rec}(T) , is dense in X . Specifically, a vector x\in X is called recurrent for T if there exists a sequence (\omega_{n})_{n\in\mathbb{N}} with \omega_{n}\uparrow\infty such that \lim_{n}T^{\omega_{n}}x=x . Furthermore, a continuous linear operator T\in\mathcal{L}(X) is said to be quasi-rigid [ 33 , 46 ] if there exists a sequence (\theta_{n})_{n\in\mathbb{N}} with \theta_{n}\uparrow\infty such that the set \{x\in X:\lim_{n}T^{\theta_{n}}x=x\} is dense in X . Notably, if T is quasi-rigid, then \mathrm{Rec}(T) is dense lineable. Question 1.1 Let X be an infinite dimensional separable F -space, and let \Lambda\subset\mathcal{L}(X) be a countable collection of quasi-rigid operators. Is it true \bigcap_{T\in\Lambda}\mathrm{Rec}\left(T\right) is dense lineable? As a consequence of our main Theorem 4.3 , we can affirmatively answer our motivating problem. Moreover, we can assert the existence of a dense vector subspace E\subset X such that, for each T\in\Lambda , there exists a sequence (k_{n,T})_{n}\uparrow\infty satisfying \displaystyle E\subset\{x\in X:\lim_{n\rightarrow\infty}T^{k_{n,T}}x=x\},\quad\forall\;T\in\Lambda. Recall that a continuous linear operator T:X\to X is said to be hypercyclic if there exists a vector x\in X such that the set \{T^{n}x:n\in\mathbb{N}\} is dense in X . The set of such vectors is denoted by \mathrm{HC}(T) . On the other hand, a sequence of continuous linear operators (T_{n}:X\to Y)_{n\in\mathbb{N}} is called universal if there exists a vector x\in X such that the set \{T_{n}x:n\in\mathbb{N}\} is dense in Y . The set of such vectors is denoted by \mathrm{HC}((T_{n})_{n\in\mathbb{N}}) . A significant distinction arises between studying the hypercyclicity of a single operator T and the universality of a sequence (T_{n})_{n} . While the set of hypercyclic vectors \mathrm{HC}(T) is always dense in the hypercyclic case, the set \mathrm{HC}\left((T_{n})_{n}\right) may not necessarily be dense in the universal case. This difference is highlighted by Godefroy and Shapiro [ 29 , pp. 234] . Herrero [ 36 ] and Bourdon [ 20 ] demonstrated that if an operator T is hypercyclic on a complex Banach space, then the set of hypercyclic vectors \mathrm{HC}(T) is dense-lineable. For real Banach spaces, this result was extended by Bès [ 16 ] , while for more general settings, such as topological vector spaces, Wengenroth [ 52 ] further generalized the result. In 1999, L. Bernal [ 7 ] established two results concerning universal sequences and their connection to lineability and dense-lineability. Specifically, if X is a topological vector space, Y is a metrizable TVS, and (T_{n}:X\to Y)_{n\in\mathbb{N}} is hereditary hypercyclic, then \mathrm{HC}\left((T_{n})_{n}\right) is lineable. Moreover, if X,Y are separable metrizable spaces and (T_{n}:X\to Y)_{n\in\mathbb{N}} is densely hereditary hypercyclic (DHHC), then \mathrm{HC}\left((T_{n})_{n}\right) is dense-lineable. Recall that (T_{n}) is DHHC whenever every subsequence \mathrm{HC}((T_{n_{k}})_{k}) is dense. L. Bernal and Grosse-Erdmann established several equivalences related to the hypercyclicity criterion for a sequence of operators. Theorem 1.2 (Theorem 2.2 and Remark 2.3, [ 10 ] ) Let X and Y be two separable F -spaces, and let (T_{n})_{n}\subset\mathcal{L}(X,Y) . Then, the following assertions are equivalent: 1. (T_{n})_{n} satisfies the hypercyclicity criterion. 2. (T_{n})_{n} admits a densely hereditary hypercyclic subsequence. 3. For each N\in\mathbb{N} , the N -fold direct sum T_{n}\oplus\cdots\oplus T_{n} is densely hypercyclic. F. León and V. Müller [ 26 ] further developed the relationships between items (1) and (3) of the previous theorem, as well as additional implications, in the context of non-complete spaces. They also explored connections with dense-lineability. Recently, A. López and Q. Menet [ 42 ] provided a negative answer to an open problem proposed by S. Grivaux et al. [ 32 , Question 6.1] . The authors demonstrated that every infinite-dimensional separable complex Banach space admits a recurrent operator T for which the set of recurrent vectors, \mathrm{Rec}(T) , is not dense-lineable. An important notion in the study of linear dynamics is Li–Yorke chaos . This concept was originally introduced by Li and Yorke [ 41 ] in their seminal work on interval maps. Let (Z,d) be a metric space, and let f\colon Z\to Z be a continuous map. We say that a pair (p,q)\in Z\times Z is a Li–Yorke pair for f if: \liminf_{n\to\infty}d(f^{n}(p),f^{n}(q))=0\quad\text{and}\quad\limsup_{n\to\infty}d(f^{n}(p),f^{n}(q))>0. A subset \Gamma\subset Z is called a scrambled set for f if, for every pair of distinct points x,y\in\Gamma , the pair (x,y) is a Li–Yorke pair for f . The map f is said to be Li–Yorke chaotic if there exists an uncountable scrambled set for f . Bermúdez et al. [ 6 ] investigated Li-Yorke chaos for operators on Banach spaces. They proved that Li-Yorke chaos is equivalent to the existence of irregular vectors and established sufficient conditions under which the set of irregular vectors is dense-lineable. These results were later extended by Bernardes et al. to the setting of Fréchet spaces [ 11 ] . It is worth mentioning that Bernardes et al. [ 12 ] conducted a systematic study of Mean Li-Yorke Chaos, establishing criteria and sufficient conditions for the existence of a dense mean irregular manifold, followed by the work of [ 40 ] . Irregular vectors have become a central topic in linear dynamics. In [ 13 ] , Bernardes et al. studied distributional chaos in the context of Fréchet spaces. They introduced the Distributional Chaos Criterion , which was shown to be equivalent to the existence of a distributionally irregular vector. Moreover, they provided conditions under which the set of uniform distributionally irregular vectors is dense-lineable. Recently, Conejero et al. [ 22 ] explored distributional chaos for sequences of operators (T_{n})_{n}\in\mathcal{L}(X,Y) , where X and Y are Fréchet spaces. In this paper, motivated by the notion of universality, we introduce the concept of Li-Yorke chaos for sequences of continuous linear operators (T_{n}:X\to Y) , where X is an F -space and Y is a normed space, in Section 2. In Section 3, we introduce the notion of the D -phenomenon, study several properties, and review aspects of Furstenberg families. In Section 4, we establish a common dense-lineable criterion for sequences of operators. Additionally, we provide a negative answer to the open problem proposed by Grivaux et al. [ 32 , Problem 6.1] , and we also address [ 11 , Problem 1] for sequences of operators, providing a negative answer. 2 Li-Yorke Chaos for Sequences of Operators In this section, X denotes an F -space (that is, a completely metrizable topological vector space), Y a normed space, and (T_{n}\colon X\to Y)_{n} a sequence of continuous linear operators. A key notion in the study of Li–Yorke chaos, which can naturally be extended to sequences of operators, is that of irregular vectors. Definition 2.1 We say that a vector x\in X is irregular for (T_{n})_{n} if \displaystyle\liminf_{n\to\infty}\|T_{n}(x)\|=0\quad\text{and}\quad\limsup_{n\to\infty}\|T_{n}(x)\|=\infty. (1) In [ 6 , 11 ] , it is established that, for a continuous linear operator T on a Banach or Fréchet space X , the existence of semi-irregular vectors for T guarantees the existence of irregular vectors. Specifically, a vector x\in X is called semi-irregular for T if it satisfies \liminf_{n\to\infty}\|{T^{n}x}\|=0 and \limsup_{n\to\infty}\|{T^{n}x}\|>0 . Example 2.2 Let X be a Banach space, and consider the sequence of operators (T_{n})_{n\in\mathbb{N}}\in\mathcal{L}(X) defined by T_{n}(x)=\begin{cases}0&\text{if }n\text{ is odd},\\ x&\text{if }n\text{ is even}.\end{cases} For every nonzero vector x\in X\setminus\{0\} , we observe that \displaystyle\liminf_{n\to\infty}\|T_{n}(x)\|=0\quad\text{and}\quad 0<\limsup_{n\to\infty}\|T_{n}(x)\|<\infty. From the above example, it follows that every nonzero vector is semi-irregular for (T_{n})_{n} , yet no irregular vectors exist. This observation motivates the following definition of Li–Yorke chaos based on the notion of irregular vectors: Definition 2.3 A sequence of operators (T_{n})_{n\in\mathbb{N}} is said to be Li–Yorke chaotic if there exists an uncountable set \Gamma\subset X such that, for every pair p,q\in\Gamma with p\neq q , the vector p-q is irregular for (T_{n})_{n} . Moreover, (T_{n})_{n} is called densely Li–Yorke chaotic if \Gamma is also dense in X . The first observation is that if a sequence of operators (T_{n})_{n} admits an irregular vector, then (T_{n})_{n} is Li–Yorke chaotic. Indeed, if x is an irregular vector for (T_{n})_{n} , it suffices to consider the uncountable set \Gamma=\mathbb{C}\cdot x . By construction, for every pair of distinct scalars \alpha,\beta\in\mathbb{C} , the vector \alpha x-\beta x=(\alpha-\beta)x is irregular. This classical argument establishes the Li–Yorke chaoticity of (T_{n})_{n} . Example 2.4 It is well known that every hypercyclic operator is Li–Yorke chaotic. Similarly, it can be shown, without much difficulty, that if a sequence of operators (T_{n})_{n} is universal, then (T_{n})_{n} is also Li–Yorke chaotic. Proposition 2.5 Let (T_{n})_{n\in\mathbb{N}}\subset\mathcal{L}(X,Y) . The following sets: \displaystyle A \displaystyle=\{x\in X:\inf_{n\in\mathbb{N}}\|T_{n}x\|=0\}, \displaystyle B \displaystyle=\{x\in X:\sup_{n\in\mathbb{N}}\|T_{n}x\|=\infty\}, are G_{\delta} -sets in X . In particular, the set of irregular vectors for (T_{n})_{n} is a G_{\delta} -set. Proof. To prove the result, we rewrite the sets A and B as countable intersections of open sets. Specifically, observe that: \displaystyle A \displaystyle=\bigcap_{k\in\mathbb{N}}\bigcup_{n\in\mathbb{N}}\{x\in X:\|T_{n}x\|<\tfrac{1}{k}\}, \displaystyle B \displaystyle=\bigcap_{k\in\mathbb{N}}\bigcup_{n\in\mathbb{N}}\{x\in X:\|T_{n}x\|>k\}. Since each set \{x\in X:\|T_{n}x\|<\tfrac{1}{k}\} and \{x\in X:\|T_{n}x\|>k\} is open in X , it follows that A and B are G_{\delta} -sets. Consequently, the set of irregular vectors for (T_{n})_{n} , which is the intersection A\cap B , is also a G_{\delta} -set. This completes the proof. ∎ Bernardes et al., in [ 11 , Theorem 10] , established six equivalent conditions characterizing when an operator T\colon X\to X , defined on a separable Fréchet space X , is densely Li–Yorke chaotic. Among these equivalences is the density of the set of irregular (or semi-irregular) vectors. This result motivates us to extend such characterizations to the context of sequences of operators, as formalized in the following theorem: Theorem 2.6 Let X be a separable F -space, and let (T_{n}\colon X\to Y)_{n\in\mathbb{N}} be a sequence of continuous linear operators. Then the following assertions are equivalent: 1. (T_{n})_{n} is densely Li–Yorke chaotic. 2. (T_{n})_{n} admits a dense set of irregular vectors. 3. (T_{n})_{n} admits a residual set of irregular vectors. Remark 2.7 In [ 11 , Theorem 10] , a key implication is established: if T admits a residual set of irregular vectors, then T is densely Li–Yorke chaotic. Their proof relies on Zorn’s Lemma and the construction of a D -subspace, where D:=\mathbb{Q} or D:=\mathbb{Q}+i\mathbb{Q} , consisting (up to the zero vector) of irregular vectors for T . Our approach to proving the implication (3)\implies(1) in Theorem 2.6 differs from theirs. We will rely on Mycielski’s Theorem [ 44 , 50 ] , which has been widely used in the study of nonlinear dynamical systems, particularly in the context of compact spaces [ 28 , 17 , 49 , 39 , 37 , 38 ] . In the setting of linear dynamics, it has been applied by Zhen Jiang and Jian Li [ 40 ] . It is worth noting that these works involve Li-Yorke chaos. Beyond the context of chaoticity, M. Saavedra and M. Stadlbauer employed this theorem to establish an equivalence for quasi-rigid operators on separable F -spaces [ 46 , Theorem 3.3] . Theorem 2.8 (Mycielski Theorem) Suppose that X is a separable complete metric space without isolated points, and that for every n\in\mathbb{N} , the set \mathcal{R}_{n} is residual in the product space X^{n} . Then there is a Mycielski set \mathcal{K} in X such that \displaystyle(x_{1},x_{2},\cdots,x_{n})\in\mathcal{R}_{n} for each n\in\mathbb{N} and any pairwise different n points x_{1},x_{2},\ldots,x_{n} in \mathcal{K} . A set \mathcal{K} is referred to as a Mycielski set if the intersection of \mathcal{K} and any nonempty open set U contains a Cantor set. Proof of Theorem 2.6 . Assume that (1) holds. Then, there exists a dense uncountable set \Gamma\subset X such that, for any p,q\in\Gamma with p\neq q , the vector p-q is irregular. Fix q_{0}\in\Gamma and define the set D:=\{p-q_{0}:p\in\Gamma\setminus\{q_{0}\}\} . Since \Gamma is dense and uncountable, D is a dense set of irregular vectors. This proves (2). To show that (2) implies (3), observe that the set of irregular vectors is a G_{\delta} -set by Proposition 2.5 . Since this set is dense, it follows that it is residual, thereby establishing (3). Finally, assume that (3) holds, that is, the set of irregular vectors is residual. Then, there exists a sequence of dense open sets (U_{\ell})_{\ell\in\mathbb{N}} in X such that the set of irregular vectors is \bigcap_{\ell}U_{\ell} . For each \ell\in\mathbb{N} , define the dense open set W_{\ell}\subset X\times X by W_{\ell}:=\{(p,q)\in X\times X:p-q\in U_{\ell}\}. The intersection \bigcap_{\ell}W_{\ell} is residual in X\times X . By Mycielski’s Theorem, there exists a Mycielski set \mathcal{K}\subset X such that \mathcal{K}\times\mathcal{K}\setminus\Delta\subset\bigcap_{\ell}W_{\ell}, where \Delta denotes the diagonal in X\times X . Let \Gamma:=\mathcal{K} . Clearly, \Gamma defined in this way is an uncountable dense set of X such that, for any distinct p,q\in\Gamma , we have p-q\in\bigcap_{\ell}U_{\ell} . In other words, p-q is an irregular vector for the sequence (T_{n})_{n} . This completes the proof. ∎ Remark 2.9 The implications (1)\Rightarrow(2) and (2)\Rightarrow(3) in the previous statement hold without requiring the space X to be separable. Definition 2.10 A sequence of operators (T_{n}\colon X\to Y)_{n\in\mathbb{N}} is said to satisfy the Li-Yorke Chaotic Criterion (LYCC) if there exist a subset X_{0}\subset X and a sequence of indices (n_{k})_{k} such that the following properties hold: 1. The sequence (T_{n_{k}}x)_{k} converges to zero for every x\in X_{0} . 2. There exists a bounded sequence (a_{n})_{n}\subset\overline{\mathrm{span}(X_{0})} such that (T_{n}a_{n})_{n} is unbounded. Definition 2.11 We say that a sequence of operators (T_{n}\colon X\to Y)_{n\in\mathbb{N}} satisfies the dense Li–Yorke chaos criterion if there exists a dense subset X_{0}\subseteq X such that X_{0} satisfies properties (1) and (2) of Definition 2.10 . Remark 2.12 In Theorem 2.15 , we will prove that the Li–Yorke Chaotic Criterion (LYCC) is equivalent to Li–Yorke chaos for sequences of operators. These results are natural adaptations of Definition 7 and Theorem 8 in [ 6 ] to the sequential setting. Remark 2.13 In [ 11 ] , a (dense) Li–Yorke Chaos criterion is established for an operator T\colon X\to X , where X is a Fréchet space. Condition (1) in 2.10 is given by the requirement that T^{n}x has a subsequence converging to zero for every x\in X_{0} . They prove that the (dense) Li–Yorke Chaos criterion is equivalent to (dense) Li–Yorke chaos, respectively. Remark 2.14 In Theorem 2.17 , we characterize the dense Li-Yorke Chaos criterion. However, in general, this criterion is not equivalent to densely Li–Yorke chaotic, as we will show in Theorem 4.2 . This represents a significant departure from the results obtained in [ 11 ] . Theorem 2.15 A sequence of operators (T_{n}\colon X\to Y)_{n\in\mathbb{N}} is Li–Yorke chaotic if and only if it satisfies the Li–Yorke Chaotic Criterion (LYCC). Proof. If (T_{n})_{n\in\mathbb{N}} is Li-Yorke chaotic, then there exists an irregular vector q\in X for (T_{n})_{n} . In this case, it suffices to take X_{0}=\{q\} . Conversely, if some vector in X_{0} is irregular for (T_{n})_{n} , the conclusion follows immediately. Otherwise, suppose that no vector in X_{0} is irregular. Then, for every u\in\mathrm{span}(X_{0}) , we have: \displaystyle\lim_{k\to\infty}\|T_{n_{k}}u\|=0\quad\text{and}\quad\sup_{n\in\mathbb{N}}\|T_{n}u\|<\infty. By applying an inductive construction, we obtain a subsequence (\theta_{k})_{k} of (n_{k})_{k} , a sequence (m_{k})_{k}\subset\mathbb{N} and a sequence of vectors \{u_{k}\}_{k}\subset\mathrm{span}(X_{0}) satisfying the following conditions for all k\in\mathbb{N} : • u_{k}\in B(0,2^{-k}) , • \|T_{\theta_{k}}u_{i}\|<2^{-i}k^{-1} for all 1\leq i\leq k , • For k>1 , \|T_{m_{k}}u_{k}\|>k+\sum_{i=1}^{k-1}\sup_{n\in\mathbb{N}}\|T_{n}u_{i}\|. Inductively, we can find a strictly increasing sequence of positive integers (i_{\ell})_{\ell} such that: • i_{\ell+1}>\max\{m_{i_{\ell}},\theta_{i_{\ell}}\} , and • For each j\in\{1,\ldots,i_{\ell}\} , \displaystyle\sup_{x\in B(0,2^{-i_{\ell+1}})}\|{T_{j}x}\|<2^{-\ell}. Certainly, p:=\sum_{\ell\in\mathbb{N}}u_{i_{\ell}}\in X . We claim that the vector p is irregular for (T_{n})_{n} . We observe that \displaystyle\|{T_{m_{i_{t}}}p}\| \displaystyle\geq\|{T_{m_{i_{t}}}u_{i_{t}}}\|-\sum_{\ellt}\|{T_{m_{i_{t}}}u_{i_{\ell}}}\| \displaystyle>i_{t}+\sum_{st}2^{-\ell} \displaystyle>i_{t}-1\xrightarrow[t\rightarrow\infty]{}\infty On the other hand, \displaystyle\|{T_{\theta_{i_{t}}}p}\| \displaystyle\leq\sum_{\ell\leq t}\|{T_{\theta_{i_{t}}}u_{i_{\ell}}}\|+\sum_{\ell>t}\|{T_{\theta_{i_{t}}}u_{i_{\ell}}}\| \displaystyle\leq\frac{1}{i_{t}}\sum_{\ell\leq t}2^{-i_{\ell}}+\sum_{\ell>t}2^{-\ell} \displaystyle\leq\frac{1}{i_{t}}+2^{-t}\xrightarrow[t\rightarrow\infty]{}0. This completes the proof. ∎ Proposition 2.16 If (T_{n}:X\rightarrow Y)_{n} satisfies the Dense Li-Yorke Chaotic Criterion, then it is densely Li-Yorke chaotic. Proof. It suffices to show that the set of irregular vectors is dense, as guaranteed by Theorem 2.6 . Assume, for contradiction, that the set of irregular vectors for (T_{n})_{n} , denoted by \mathcal{I} , is not dense in X . Then, there exists a non-empty open set U\subset X such that U\cap\mathcal{I}=\emptyset . By hypothesis, there exists a dense set X_{0}\subset X and a sequence of indices (n_{k})_{k} satisfying conditions (1) and (2) in Definition 2.10 . By the density of X_{0} , there exists a vector q\in U\cap X_{0} . Since q is not irregular, it follows that T_{n_{k}}q\xrightarrow[k\to\infty]{}0\quad\text{and}\quad M:=\sup_{n}\|{T_{n}q}\|<\infty. In the proof of Theorem 2.15 , we construct an irregular vector p\in X , a subsequence (\theta_{i_{t}}) of (n_{k})_{k} , and a sequence (m_{i_{t}}) . For any non-zero complex number \lambda , we claim that the vector q+\lambda p is irregular. This follows from the following properties: T_{\theta_{i_{t}}}(q+\lambda p)\xrightarrow[t\to\infty]{}0, and \|{T_{m_{i_{t}}}(q+\lambda p)}\|\geq|{\lambda}|(i_{t}-1)-M\xrightarrow[t\to\infty]{}\infty. Clearly, there exists some \lambda_{0}\neq 0 such that q+\lambda_{0}p\in U . This contradicts the assumption that U\cap\mathcal{I}=\emptyset . ∎ Theorem 2.17 Let X be a separable F -space, Y a normed space, and (T_{n}:X\rightarrow Y)_{n}\subset\mathcal{L}(X,Y) . Then, the following statements are equivalent: 1. (T_{n})_{n} satisfies the Dense Li-Yorke Chaotic Criterion. 2. For each m\in\mathbb{N} , the m -fold direct sum T_{n}\oplus\cdots\oplus T_{n}:X^{m}\longrightarrow Y^{m} is densely Li-Yorke chaotic. Proof. If (T_{n})_{n} satisfies the Dense Li-Yorke Chaotic Criterion, it is clear that for each m\in\mathbb{N} , the m -fold direct sum T_{n}\oplus\cdots\oplus T_{n}:X^{m}\to Y^{m} also satisfies the criterion. According to Proposition 2.16 , this implies (2). To show that (2) implies (1), it suffices to prove the existence of a sequence (n_{k})_{k} and a dense subset X_{0}\subset X such that T_{n_{k}}x\to 0 for every x\in X_{0} . For each m\in\mathbb{N} , define \mathcal{R}_{m}\subset X^{m} as: \mathcal{R}_{m}:=\{(x_{1},\ldots,x_{m})\in X^{m}:\exists\theta_{\ell}\uparrow\infty\;\text{such that}\;\lim_{\ell}T_{\theta_{\ell}}x_{i}=0,\forall i\in\{1,\ldots,m\}\}. Clearly, \mathcal{R}_{m} is a G_{\delta} -set and is dense since it contains the set of irregular vectors. By Mycielski’s Theorem, there exists a Mycielski set \mathcal{K}\subset X such that \mathcal{K}^{m}\subset\mathcal{R}_{m} for every m\in\mathbb{N} . Since X is separable, there exists a countable dense set X_{0}:=\{q_{i}\}_{i\in\mathbb{N}}\subset\mathcal{K} . We can then choose a strictly increasing sequence of positive integers (n_{k})_{k} such that \lim_{k}T_{n_{k}}x=0 for every x\in X_{0} . ∎ We say that T:X\to X admits a dense irregular manifold if there exists a dense subspace where every non-zero vector is irregular for T ; in other words, if the set of irregular vectors is dense-lineable. In [ 6 , 11 , 40 ] , sufficient conditions for the existence of a dense irregular manifold are established. Question 2.18 ( [ 11 ] ) Does dense Li–Yorke chaos imply the existence of a dense irregular manifold for operators on Fréchet (or Banach) spaces? It is natural to ask whether a sequence of operators (T_{n}:X\to X)_{n} that is densely Li-Yorke chaotic necessarily admits a dense irregular manifold. We answer this question in the negative for every infinite-dimensional separable complex Banach space X in Theorem 4.2 . 3 D-phenomenon and Furstenberg family In linear dynamics, there are several behaviors of interest, such as recurrence, Li-Yorke chaos, hypercyclicity, among others. Clearly, seeking a dense-lineable criterion for each of these aspects individually is not particularly motivating. In this section, we introduce the notion of the D -phenomenon , which allows us to study the aforementioned behaviors from a unified perspective. Moreover, it enables us to establish a common dense-lineable criterion for a countable family of D -phenomena, as we will see in the next section. Definition 3.1 Let X and Y be topological vector spaces. We say that \Psi is a D-phenomenon from X to Y if, for each x\in X , there exists a non-empty set \Psi(x) consisting of open subsets of Y such that the following holds: for any p=\sum_{j=1}^{m}\alpha_{j}x_{j}\in X with \alpha_{i}\neq 0 and for any V\in\Psi(p) , there exist non-empty open sets W_{i}\in\Psi(x_{i}) such that \sum_{j=1}^{m}\alpha_{j}W_{i}\subset V. For the m -fold product X^{m} , we define \Psi(x_{1},\ldots,x_{m}):=\left\{\prod_{i=1}^{m}A_{i}\subset X^{m}:A_{i}\in\Psi(x_{i})\right\}. Remark 3.2 The choice of the term ”D-phenomenon” refers to a dynamic phenomenon. Throughout this section, the connection to dynamic aspects, such as recurrence, hypercyclicity, and others, will become more apparent. Example 3.3 The following are key examples of D-phenomena in linear dynamics. • Let X be a topological vector space (TVS). For each x\in X , define \Psi^{\mathrm{Rec}}(x):=\left\{\text{all open neighborhoods of }x\text{ in }X\right\}. (2) • Let X and Y be topological vector spaces. For each x\in X , define \Psi^{\mathrm{HC}}(x):=\left\{\text{all non-empty open subsets of }Y\right\}. (3) • Let X,Y be topological vector spaces, and let h\colon X\to Y be a continuous linear operator. For each x\in X , define \Psi^{h}(x):=\left\{\text{all open neighborhoods of }h(x)\text{ in }Y\right\}. (4) In all cases, one can verify that \Psi defines a D -phenomenon . Proposition 3.4 Let X and Y be topological vector spaces, and let \{\Psi_{i}\}_{i\in J} be a family of D -phenomena from X to Y . Define the union D -phenomenon \Psi_{J} as \Psi_{J}(x):=\bigcup_{i\in J}\Psi_{i}(x)=\left\{V\subseteq Y:V\in\Psi_{i}(x)\ \text{for some}\ i\in J\right\}. Then, \Psi_{J} is a D -phenomenon from X to Y . Proof. Let p=\sum_{j=1}^{m}\alpha_{j}x_{j} , where \alpha_{j} are non-zero complex numbers and x_{j}\in X . Fix an open set V\in\Psi_{J}(p) . By definition of \Psi_{J} , there exists i\in J such that V\in\Psi_{i}(p) . Since \Psi_{i} is a D -phenomenon, there exist non-empty open sets W_{j}\in\Psi_{i}(x_{j}) for each j=1,\ldots,m , satisfying: \sum_{j=1}^{m}\alpha_{j}W_{j}\subseteq V. The key observation is that W_{j}\in\Psi_{J}(x_{j}) for all j , as \Psi_{J} contains all sets from each \Psi_{i} . Thus, \Psi_{J} inherits the D -phenomenon property from the family \{\Psi_{i}\}_{i\in J} . ∎ We observe that the D -phenomena presented in ( 2 ) and ( 4 ) of Example 3.3 are associated with single operators. In contrast, the phenomenon in ( 3 ) corresponds to a union of D -phenomena (as in Proposition 3.4 ), where each constituent D -phenomenon is linked to an operator in \mathcal{L}(X,Y) . This raises a natural question: Is every D -phenomenon expressible as a union of operator-associated D -phenomena? The following result addresses this inquiry. Proposition 3.5 Let h:X\rightarrow Y be a surjective operator between two TVSs X and Y . For each x\in X , define \Psi(x):=\left\{h^{-1}(U)\subseteq X:U\text{ is an open neighborhood of }h(x)\text{ in }Y\right\}. (5) Then, \Psi is a D -phenomenon from X to X . Proof. Let p=\sum_{i=1}^{m}\alpha_{i}x_{i} with \alpha_{i}\neq 0 and x_{i}\in X . Fix V=h^{-1}(U)\in\Psi(p) , where U\subseteq Y is open and h(p)\in U . Since h is linear and continuous, we have: h(p)=\sum_{i=1}^{m}\alpha_{i}h(x_{i})\in U. By the continuity of vector addition and scalar multiplication in Y , there exist open neighborhoods W_{i}\subseteq Y of h(x_{i}) such that \sum_{i=1}^{m}\alpha_{i}W_{i}\subseteq U . For each i , define Z_{i}:=h^{-1}(W_{i})\in\Psi(x_{i}) . Then: \sum_{i=1}^{m}\alpha_{i}Z_{i}\subseteq h^{-1}(U)=V. Thus, \Psi satisfies the D -phenomenon property. ∎ Corollary 3.6 Let X be an infinite-dimensional Banach space and let h\in\mathcal{L}(X) be a surjective bounded operator that is not injective. Then, the D-phenomenon \Psi in ( 5 ) for h is not a union of D-phenomena associated with continuous operators. Proof. We can notice that every D-phenomenon generated by the union of a family of D-phenomena associated with continuous operators admits bounded open sets. In contrast, every open set in \Psi(0) is unbounded ∎ The following result provides a characterization of how bounded sets propagate through D -phenomena, revealing that the local existence of bounded open sets implies their ubiquity: Proposition 3.7 Let X and Y be topological vector spaces, and \Psi a D -phenomenon from X to Y . The following are equivalent: (i) \Psi(x) contains a bounded open set for every x\in X ; (ii) \Psi(q) contains a bounded open set for some q\in X . Proof. The implication (i) \Rightarrow (ii) is immediate. For the converse, suppose there exists q\in X with a bounded open set U\in\Psi(q) . Let x\in X be arbitrary. Decompose q=x+(q-x) and observe that: By the D -phenomenon property, there exist open sets A\in\Psi(x) and B\in\Psi(q-x) satisfying A+B\subseteq U . Clearly, this implies that A is a bounded open set. Thus \Psi(x) contains the bounded open set A . ∎ Definition 3.8 A D -phenomenon \Psi from X to Y is called bounded if \Psi(0) contains a bounded open set. Example 3.9 Let X be a Banach space. For each x\in X , define \Psi(x):=\left\{B_{X}\left(x,\tfrac{1}{n}\right):n\in\mathbb{N}\right\}, where B_{X}(x,\tfrac{1}{n}) denotes the open ball of radius 1/n centered at x . Then \Psi is a bounded D -phenomenon, refining the structure introduced in ( 2 ). This example reveals natural structural relationships between D -phenomena. To formalize this, we define a partial order on the class of D -phenomena from X to Y : • For D -phenomena \Psi_{1} and \Psi_{2} , we write \Psi_{1}\prec\Psi_{2} if for every x\in X and every V\in\Psi_{1}(x) , there exists W\in\Psi_{2}(x) with W\subseteq V . • We write \Psi_{1}\sim\Psi_{2} if both \Psi_{1}\prec\Psi_{2} and \Psi_{2}\prec\Psi_{1} . A D -phenomenon \Psi is called minimal if, whenever \Psi_{1}\prec\Psi for some D -phenomenon \Psi_{1} , it follows that \Psi_{1}\sim\Psi . Proposition 3.10 Let X and Y be topological vector spaces, where Y has a bounded neighborhood basis. For every h\in\mathcal{L}(X,Y) , the D -phenomenon \Psi^{h} (defined in ( 4 )) is minimal among all bounded D -phenomena under the partial order \prec . Proof. To establish the minimality of \Psi^{h} , let \Psi be any bounded D -phenomenon satisfying \Psi\prec\Psi^{h} . We will prove that \Psi^{h}\prec\Psi . Since \Psi\prec\Psi^{h} , we observe that for every x\in X , each open set in \Psi(x) contains the vector h(x) . Fix q\in X and consider V\in\Psi^{h}(q) . Define M:=V-h(q) , which is an open neighborhood of 0 in Y . Since \Psi is a bounded D -phenomenon, there exists a bounded open set K\in\Psi(0) , and for some m\in\mathbb{N} we have that K\subseteq m\cdot M . On the other hand, since m\cdot 0=0 , there exists U_{1}\in\Psi(0) with m\cdot U_{1}\subseteq K , implying that U_{1}\subseteq M . Writing 0=q+(-q) , the D -phenomenon guarantees the existence of two open sets W\in\Psi(q) and U_{2}\in\Psi(-q) such that W+U_{2}\subseteq U_{1} . Observing that U_{2}+h(q) is an open neighborhood of 0 , we obtain: W+(U_{2}+h(q))\subseteq M+h(q)=V. Thus, W\subseteq V , proving that \Psi^{h}\prec\Psi . Since \Psi was arbitrary, the minimality of \Psi^{h} follows. ∎ {forest} Figure 1: Diagram of the partial order of D-phenomena Up to this point, all studied D -phenomena have required that every open set V\subset Y in \Psi(0) contains the zero vector. The following proposition establishes a fundamentally different behavior, which is intrinsically linked to Li–Yorke chaos. Proposition 3.11 Let X be a topological vector space and Y be a normed space. For each x\in X , define \displaystyle\Psi^{\mathrm{LY}}(x):=\{B(0,r),Y\setminus B[0,s]:r,s>0\} (6) Then \Psi^{\mathrm{LY}} is a D-phenomenon from X to Y . Proof. Let p=\sum_{i=1}^{t}\alpha_{i}x_{i} be any fixed vector in X with \alpha_{i}\neq 0 . We distinguish two cases based on the open set in \Psi(p) : Case 1: Consider an open set B(0,r)\in\Psi(p) for some r>0 . Clearly, there exists n\in\mathbb{N} such that \sum_{i}|\alpha_{i}|s+1 . For i\in\{2,\ldots,t\} , choose the open sets W_{i}:=B(0,\frac{1}{n}) in \Psi(x_{i}) such that n\in\mathbb{N} and \sum_{i=2}^{t}|\alpha_{i}|