Click here to return to the main course web page.
Let \(\mathcal{L}\) be a finite relational language. A class \(\mathfrak{F}\) of finite \(\mathcal{L}\)-structures is a Fraïssé class if it is nonempty, hereditary (closed under taking substructures), closed under isomorphism, and has the amalgamation property: if \(f:\mathcal{A} \to \mathcal{B}\) and \(g:\mathcal{A} \to \mathcal{C}\) are embeddings between elements of \(\mathfrak{F}\), there exists a \(\mathcal{D}\) in \(\mathfrak{F}\) and embeddings \(\bar f: \mathcal{B} \to \mathcal{D}\) and \(\bar g : \mathcal{C} \to \mathcal{D}\) such that \(\bar f \circ f = \bar g \circ g\). If \(\mathcal{M}\) is an \(\mathcal{L}\)-structure and every finite partial automorphism of \(\mathcal{M}\) extends to an automorphism, then we say that \(\mathcal{M}\) is ultrahomogeneous. A countable ultrahomogenous structure is called a Fraïssé strucure. (The terminology ultrahomogeneous is common in Fraïsse theory. Model theorists define homogeneous to mean something stronger than being ultrahomogeneous and equivalent in the class of countable structures.)