Certain \(L\)-spaces under CH (Q1207237): Difference between revisions

HFC's are subspaces of \(2^{\omega_ 1}\) which capture much of the combinatorics involved in the study of \(L\)-spaces. With their duals, HFD's (which capture much of the combinatorics in the study of \(S\)- spaces) they have been studied both for the light they shed on \(S\) and \(L\) (e.g. Szenmiklóssy's proof that ``\(MA+\neg CH+\) there is an \(S\)- space'' used HFD's) and for their own sake. HFC's generalize to \(\text{HFC}_{\text{w}}\) (the ``w'' stands for ``weak'') and particularize to strong HFC's and strong \(\text{HFC}_{\text{w}}\)'s. This paper asks: if you start with an HFC (a) what kind of subspaces must it have? (b) must it be destroyed by some ccc forcing? Here are the technical definitions. A family \(M\) of finite functions from \(\omega_ 1\) to 2, \(M=\{\sigma_{\alpha,k}:\) \(\alpha<\kappa\), \(k<n<\omega\}\), is a regular \((\kappa,n)\)-matrix iff for \(\beta\neq\alpha\) and \(i,j < n\), \(\text{dom }\sigma_{\alpha,i}\cap\text{dom }\sigma_{\beta,j}=\emptyset\). \(M\) is an almost-\(n\) cover for \(Y\subset 2^{\omega_ 1}\) iff there is \(Y'\) a countable subset of \(Y\) so that for all \((x_ 0,\dots,x_{n-1})\in (Y\setminus Y')^ n\), there is some \(\alpha\) so that \(\sigma_{\alpha,k}\subset x_ k\) for all \(k\). A set \(X\subset 2^{\omega_ 1}\) is an HFC iff every regular \((\omega,1)\)- matrix \(M\) is an almost \(n\)-cover for \(X\); an \(\text{HFC}_{\text{w}}\) iff for every regular \((\omega_ 1,1)\)-matrix \(M\) there is a \(\alpha <\omega_ 1\) and an \((\alpha,1)\)-submatrix \(M'\) which is an almost \(n\)- cover for \(X\); a strong HFC iff for all \(n\), every regular \((\omega,n)\)- matrix \(M\) is an almost \(n\)-cover for \(X\); and a strong \(\text{HFC}_{\text{w}}\) iff for all \(n\) and every regular \((\omega_ 1,n)\)-matrix \(M\) there is \(\alpha <\omega_ 1\) and an \((\alpha,n)\)- submatrix \(M'\) which is an almost \(n\)-cover for \(X\). Clearly (strong) HFC's are (strong) \(\text{HFC}_{\text{w}}\)'s, and \(\text{HFC}_{\text{w}}\)'s are hereditarily Lindelöf. The main theorems are 1. Under CH, every HFC contains a strong HFC subspace. 2. Under stick (a weakening of CH), every HFC contains a strong \(\text{HFC}_{\text{w}}\). 3. Under stick, if \(X\) is a strong \(\text{HFC}_{\text{w}}\), then there is a ccc partial order forcing \(X\) to no longer be hereditarily Lindelöf. As a corollary to 3, under stick, if \(X\) is an HFC then there is a ccc partial order forcing \(X\) to no longer be hereditarily Lindelöf. On the other hand, 4. Under GCH, if \(X\) is an HFC, then there is a ccc partial order \(P\) forcing ``every ccc order forces \(X\) to be \(\text{HFC}_{\text{w}}\)''. So, for example, you could start with a model of GCH, construct an HFC \(L\)-space, force with the \(P\) of theorem 4, and then force \(MA + \neg CH\) in the usual way to get Abraham and Todorcevic's theorem: ``\(MA + \neg CH +\) there is an \(L\)-space'' is consistent. The proof of theorem 4 uses a generalization of the main notion (2- complicated) of the Abraham and Todorcevic proof. The proofs of theorems 1 and 2 use chains of elementary submodels.