Fraïssé sequences: category-theoretic approach to universal homogeneous structures (Q400425): Difference between revisions

For a set \(\mathfrak{K}\) of finite (or, at least, finitely generated) algebraic systems relative to a first order signature, if \(\mathfrak{K}\) has the amalgamation property, has only countably many isomorphic types, and has the joint embedding property, then there exists a countable algebraic system \(\mathbf{U}\), the \textit{Fraïssé limit} of \(\mathfrak{K}\), that can be represented as the union of a chain of algebraic systems from \(\mathfrak{K}\), contains isomorphic copies of all algebraic systems in \(\mathfrak{K}\) and has the property that every isomorphism between subsystems of \(\mathbf{U}\) which are in \(\mathfrak{K}\) extends to an automorphism of \(\mathbf{U}\) (strong homogeneity property). Moreover, \(\mathbf{U}\) is unique up to isomorphism. The above Fraïssé theory was later on generalized by considering uncountable sets of algebraic systems and by imposing cardinality conditions. In addition, it was assumed that the set of algebraic systems \(\mathfrak{K}\) was closed under unions of chains of length less than the size of the universal homogeneous algebraic system. As the author of the paper under review says: ``One of the objectives of this work is to relax this assumption and to make the theory general enough for capturing new cases and obtaining new examples of universal homogeneous objects.'' Specifically, the author, extends the theory of Fraïssé-Jónsson limits to the framework of category theory. The key point of his approach is dealing with sequences, Fraïssé sequences, instead of (co-)limits. After defining Fraïssé sequences, the author, shows its existence and states its properties. Following this, and motivated by later applications, he describes a way of dealing with discontinuous Fraïssé sequences and defines the concept of an amalgamation structure, specializing the result on cofinality of a Fraïssé sequence in the category of sequences. Next, the author, provides examples of classes of categories with Fraïssé sequences (categories of functors, monoids, and some categories of trees). Moreover, the author, presents an example of a category with many incomparable Fraïssé sequences. Afterwards, the author, describes a general construction on a given category, first used by D. Scott to provide models of untyped \(\lambda\)-calculus. Concretely, he assigns to a category another category, the \textit{category of embedding-projection} (or, what is equivalent, \textit{section-retraction}) \textit{pairs}. Such a category is suitable for applications to the theory of Valdivia compacta and Banach spaces with Markushevich bases. We point out that, the author, also works out a variant of Fraïssé-Jónsson theory replacing full continuity of sequences by semi-continuous sequences of section-retraction pairs. Finally, we point out that the definition of subcategory given by the author, on p. 1758, is not correct (consider the category \(\mathbf{C}\) with \(\mathrm{Ob}(\mathbf{C}) = \{c\}\), \(\mathrm{Mor}(\mathbf{C}) = \{f,g\}\), and where the composition is defined by the following equations: \(f\circ f = f\), \(f\circ g = g\circ f = g\circ g = g\). Let \(\mathbf{D}\) be the category with \(\mathrm{Ob}(\mathbf{D}) = \{c\}\), \(\mathrm{Mor}(\mathbf{D}) = \{g\}\), and where the composition is defined by the following equation: \(g\circ g = g\). According to the definition given by the author \(\mathbf{D}\) should be a subcategory of \(\mathbf{C}\). However, it is obvious that \(\mathbf{D}\) is not a subcategory of \(\mathbf{C}\)), that there are some typographical errors on pp. 1767, 1769, 1777, 1778, 1780, 1793, that what the author calls, on p. 1779, a ``semigroup'' is usually called a ``monoid'', and that it would be convenient to add to the Bibliography of the paper, at least, the following items: \textit{P. C. Baayen} [Universal morphisms. Amsterdam: Mathematisch Centrum (1964; Zbl 0139.40001)], and \textit{W. W. Comfort} and \textit{S. Negrepontis} [The theory of ultrafilters. Die Grundlehren der mathematischen Wissenschaften. Band 211. Berlin-Heidelberg-New York: Springer-Verlag (1974; Zbl 0298.02004)].