Borel extractions of converging sequences in compact sets of Borel functions (Q2518765)

A result of \textit{J. Bourgain, D. H. Fremlin} and \textit{M. Talagrand} [Am. J. Math. 100, 845--886 (1978; Zbl 0413.54016)] says that if \(K\) is a pointwise compact set of Borel functions on a Polish space then, given any cluster point \(f\) of a sequence \((f_{n})_{n < \omega}\) in \(K\), one can extract a subsequence \((f_{n_{k}})_{k < \omega}\) converging to \(f\). The author improves on this result by proving that this extraction can be achieved in a ``Borel way''. Let \(X\) be a Polish space; when effective notions are considered then \(X\) is assumed to be recursively presented. We denote by \(\mathcal B (X)\) the set of Borel functions on \(X\), and \(\mathcal B (X)\) is endowed with the pointwise convergence topology. The main result of the paper is the following effective extension of the above-mentioned result: if \(K\) is a compact subset of \(\mathcal B (X)\), \((f_n )_{n < \omega}\) is a \(\Delta^{1}_{1}\) sequence in \(K\) and \(f\) is a \(\Delta^{1}_{1}\) cluster point of \((f_n )_{n < \omega}\) then there exists a \(\Delta^{1}_{1}\) subsequence of \((f_n )_{n < \omega}\) converging to \(f\). This result is closely related to the earlier work of the author in [Mathematika 34, 64--68 (1987; Zbl 0644.54012)], where he obtained that if \(K\) is a compact subset of \(\mathcal B (X)\) then every \(\Delta^{1}_{1}\) sequence in \(K\) has a converging \(\Delta^{1}_{1}\) subsequence, which is the effective analogue of a result of \textit{H. P. Rosenthal} [Am. J. Math. 99, 362--378 (1977; Zbl 0392.54009)], stating that every sequence \((f_n )_{n < \omega}\) in a compact subset \(K\) of \(\mathcal B (X)\) has a converging subsequence. The paper emphasizes the interest of effective methods in the treatment of classical descriptive problems. Accordingly, for the sake of non-specialist readers, the author gives a quick presentation of the main effective concepts and results used in his work. We also recall an effective descriptive set-theoretic result of the paper which plays an important role in the proof of the main result and is interesting in its own. Let \(\mathcal F\) be a filter on \(\omega\). We say that \(\mathcal F\) decides a set \(A \subseteq \mathcal \omega\) if either \(A \in \mathcal F\) or \(\omega \setminus A \in \mathcal F\). For a family \(\mathcal A\) of subsets of \(\omega\), \(\mathcal F\) decides \(\mathcal A\) if \(\mathcal F\) decides every \(A \in \mathcal A\). Theorem. For every family \(\mathcal A\) of \(\Pi^{1}_{1}\) subsets of \(\omega\) which is \(\Pi^{1}_{1}\) in the codes, there exists a filter basis \(\mathcal D \subseteq \mathcal A \cap \Delta^{1}_{1}\) such that \(\mathcal D\) generates a \(\Pi^{1}_{1}\) filter which decides \(\mathcal A\).