Logic group

NOTES: the speaker will be online.
I will give a talk on several recent results regarding ultrahomogeneous graphs. A relational structure R is ultrahomogeneous if every isomorphism of finite induced substructures of R extends to an automorphism of R. We classify the finite vertex-colored oriented ultrahomogeneous graphs and the finite vertex-colored undirected graphs. The classifications comprise new general methods which govern how graphs can be combined or extended to create new ultrahomogeneous graphs. Further, we extend a classic theorem of Cameron (every 5-tuple regular graph is already ultrahomogeneous) to the setting of colored graphs.

This is joint work with Sofia Brenner, Eda Kaja, Thomas Schneider, and Pascal Schweitzer.

Fix a sequence of countable structures $\mathcal{M}_n$, for $n$ in $ℕ$, in a given first-order language $\mathcal{L}$. The reduced product of $(\mathcal{M}_n)$, denoted $\prod_n\mathcal{M}_n/\mathrm{Fin}$ is the $\mathcal{L}$-structure obtained by quotienting the product of the $\mathcal{M}_n$ by the equivalence relation of 'eventual equality'. This is similar to the ultraproduct construction, yet its theory is fairly less understood.

We consider the following question: if two reduced products $\prod_n\mathcal{M}_n/\mathrm{Fin}$ and $\prod_n\mathcal{N}_n/\mathrm{Fin}$ are isomorphic, what can be said about relations between structures we started with? Of course, one can obtain isomorphisms by taking fiberwise isomorphisms, and/or shuffling the indexes, but is that all? We discuss how, in specific interesting cases (such as fields, or certain graphs), answers to this question depend on the set theoretic ambient.

The concept of a monster model in model theory is a helpful tool for avoiding cumbersome bookkeeping and notation. Instead of needing to repeatedly take elementary extensions to realise types, or find automorphisms, one can simply say that there is a model M of a theory such that any 'small' model of that theory embeds into M, and any 'small' partial isomorphism in M extends to an automorphism. While usually 'small' means 'cardinality less than M' (and M is then taken to be 'big enough'), sometimes there are special cases in which 'small' means 'set-sized'. In particular, the surreal numbers acts as a monster model for the theory of dense linear orders, real-closed fields, and more.

I will introduce the concept of a monster model and its uses, expand somewhat on the hidden set-theoretic baggage associated with these objects, and show off the monstrosity of the surreal numbers. I will then show how, without the axiom of choice, the surreal numbers may no longer be a monster, using a simple symmetric extension argument.