Consistency theory for SM-methods (Q5906299): Difference between revisions

A \({\mathcal V}\)-SM-method (briefly SM-method or \({\mathcal V}\)-method) is defined by the authors as follows. Let \({\mathcal V}\) be a double sequence space with a (linear limit) functional \({\mathcal V}\)-\(\lim:{\mathcal V}\to\mathbb{K}\) \((\mathbb{C}\) or \(\mathbb{R})\) and let \({\mathcal A}=(A^{(\nu)})\) be a sequence of infinite matrices (with entries in \(\mathbb{K}\)). Then the application domain of \({\mathcal A}\) is defined by \(\Omega_{\mathcal A}:=\bigcap_\nu\omega_{A^{(\nu)}}\), and the domain of \({\mathcal A}\) with respect to \({\mathcal V}\) by \({\mathcal V}_{\mathcal A}:=\{x\in\Omega_{\mathcal A}\mid(\sum_ka^{(\nu)}_{nk}x_k)_{n\nu}\in{\mathcal V}\}\) where \(\omega_{A^{(\nu)}}\) is the application domain of \(A^{(\nu)}\), i.e. \(\omega_{A^{(\nu)}}=\{x\in\omega\mid\sum_ka_{nk}^{(\nu)}x_k\) converges\}, \(\omega\) being the space of scalar (real or complex) sequences. Choosing \({\mathcal V}\) one obtains different SM-methods. We recall some double sequence spaces considered in this paper. \(\Omega\) being the set of all double sequences, \({\mathcal M}\) and \({\mathcal C}\) are the subsets of \(\Omega\) defined as follows: \((x_{n\nu})_n\in m\) respectively \((x_{n\nu})_n\in c\). \(\widehat{\mathcal C}\) is the subset of \({\mathcal M}\) whose elements \(x_{n\nu}\) verify the following condition: there exists \(a_x\in\mathbb{K}\) such that \((\limsup_n|x_{n\nu}-a_x|)_\nu\in c_0\). \((m,c,c_0\) are known notations for resepective sets of simple scalar sequences.) The authors consider the set \(W_{\mathcal A}\) of all members of the domain \(\widehat{\mathcal C}_{\mathcal A}\) being weakly sectionally convergent. They also introduce a bigger subspace \(V_{\mathcal A}\) suitable for consistency considerations. Although citation of results from this paper is technically difficult, we cite some of them. Theorem 4.1. Let \({\mathcal A}\) be a \(\widehat{\mathcal C}\)-SM-method with \({\mathcal C}_{c{\mathcal A}}\) containing the space of finitely non-zero sequences \(\varphi\). Both \(W_{\mathcal A}\) and \(V_{\mathcal A}\) have the absolute strong pointwise oscillating property. (The statement on \(W_{\mathcal A}\) remains true under the weaker assumption that \(\varphi\subset\widehat{\mathcal C}_{\mathcal A}\).) The proof of this theorem is developed in 9 pages (!) and is based on some results from an earlier paper of the first author and the second one, together with D. J. Fleming; see \textit{J. Boos}, \textit{D. J. Fleming}, and \textit{T. Leiger} [Sequence spaces with oscillating properties, J. Math. Anal. Appl. 200, No. 3, 519-537 (1996; Zbl 0885.46006)]. As an immediate corollary the authors get the following theorem of Mazur-Orlicz type in case of \(V_{\mathcal A}\). Theorem 4.2. Let \({\mathcal A}\) be a \(\widehat{\mathcal C}\)-SM-method with \(\varphi\subset{\mathcal C}_{c{\mathcal A}}\) and \(Y\) be a sequence space including \(\varphi\) and having the signed pointwise gliding hump property. Then \(Y\cap V_{\mathcal A}\subset F\Rightarrow Y\cap V_{\mathcal A}\subset W_F\) holds for any \(L_\varphi\)-space \(F\), in particular, for any separable FK-space \(F\). Here an FK-space is a K-space with a Fréchet topology. (Any linear subspace \(E\) of the linear space of all scalar sequences \(\omega\) with a locally convex topology \(\tau\) is called a K-space if the inclusion map \(i:(E,\tau)\to\omega\) is continuous, where \(\omega\) has the topology of coordinatewise convergence.) A sequence space \(E\) containing \(\varphi\) has the signed pointwise gliding hump property if for each \(x\in E\) and every block sequence \((y^{(j)})\) with \(\sup_j\|y^{(j)}\|_{bv}<\infty\) there exists a subsequence \((y^{(j_k)})\) of \((y^{(j)})\) and a sequence \((h_k)\) in \(\{1,-1\}\) such that \(yx\in E\) where \(y:=\sum_kh_ky^{(j_k)}\). In fact the question that gave an impulse to this paper is the question ``whether one may use gliding hump techniques for the development of consistency theorems on the base of theorems of Mazur-Orlicz type proved in several papers of the first and second author using factor sequence methods ... and the third author in his very early paper [Math. Z. 56, 134-151 (1952; Zbl 0046.06402)]''.