Logic Workshop

The talk will begin by discussing the basic definitions and general goals behind the theory of Borel equivalence relations. We focus on the Friedman-Stanley jumps $=^{+n}$, for $n=1,2,...$ and $n=\omega$. These Borel equivalence relations represent the notions of being classifiable using invariants which are countable sets of reals, countable sets of countable sets of reals, and so on. We consider the problem of constructing a Borel reduction from $=^{+n}$ to some other equivalence relation.

For $n=1$ the situation is well understood and there are many such results. For example: Marker proved that for a first order theory with an uncountable type space, its isomorphism relation is above $=^{+1}$; Larson and Zapletal characterized the analytic equivalence relations above $=^{+1}$ as those which are 'unpinned' in the Solovay extension.

In this talk we present a new technique for proving that an equivalence relation is above $=^{+n}$, when $n>1$, based on Baire-category methods. As corollaries, we conclude that $=^{+\omega}$ is 'regular' (answering a question of Clemens), and that $=^{+n}$ is 'in the spectrum of the meager ideal' (extending a result of Kanovei, Sabok, and Zapletal for $n=1$).

Expansions of the ordered additive group of the reals (or more generally definably complete expansions of ordered Abelian groups) of finite dp-rank are a class of reasonably well-behaved ordered structures that generalize the class of o-minimal structures. In this talk I will give a survey of ongoing work with John Goodrick on exploring the properties of definable sets in this class of structures.

In recent work with Luca Motto Ros we prove that under analytic determinacy there exists an analytic relation that is not class-wise Borel embeddable into any orbit equivalence relation. The result builds on an unpublished result of Becker from 2001 and fits in the area of invariant descriptive set theory. I will mainly discuss our result and how it is related to a major conjecture in the field known as the '$E_1$ conjecture'.

Polish spaces, the objects of study of descriptive set theory, are completely metrizable topological spaces that have a countable dense subset. For example, the reals - the first Polish space in the world. We will look at 'sigma-ideals' of compact subsets of a Polish space. Think of a sigma-ideal as being a collection of 'small' compact sets, under some notion of smallness -- so for example, your Polish space could be the interval $[0,1]$ and your sigma-ideal could be the collection of all its compact sets of Lebesgue measure $0$. The descriptive-set-theoretic study of these objects yields rich results for the following reason. If you look at the collection of all the compact subsets of a Polish space, that too, topologized and metrized in a natural way, turns out to be a Polish space. This means that you can look at your sigma-ideal of compact sets in two places: in the original space, say $E$, and in the `hyperspace' of all compact sets of $E$. In this talk we will deal with sigma-ideals that can be represented in a very nice way inside this hyperspace, and we will examine the behaviour of so-called G-delta subsets of $E$ with respect to this representation.

Recall that a structure is pseudofinite if every sentence satisfied by that structure has a finite model - equivalently, if the structure is elementarily equivalent to an ultraproduct of finite structures. In this talk, I will present some work in progress from two independent projects around pseudofinite rings and pseudofinite modules: one is joint work with Alex Van Abel, the other is work of my PhD student Roberto Torres. These two projects are linked by the important role played by the class of von Neumann regular rings.

The relationship between left distributivity and very large cardinal embeddings was discovered in the 1980s but remains, in many ways, mysterious. In the late 1980s Richard Laver showed that the closure of a single elementary embedding under the application operation generates a free left distributive algebra and demonstrated the linearity of a particular ordering on terms of the free left distributive algebra (given the existence of such embeddings). Patrick Dehornoy later used the braid group on infinitely many generators to show the linearity of that ordering relation within ZFC. (The consistency strength of other related theorems is still unknown). David Larue subsequently extended that work to demonstrate braid group representations of the free left distributive algebra on $n$ generators, for any natural number $n$. Still elusive was an algebra of embeddings isomorphic to a free left distributive algebra on more than one generator. We present an inverse limit construction of such a free, two-generated left distributive algebra of embeddings from a slightly stronger large cardinal assumption than the one used by Laver. (Joint work with Andrew Brooke-Taylor and Scott Cramer.)

We identify several examples of distal metric structures and examine several consequences of distality, such as the existence of distal cell decompositions, in each. These results include joint work with Itaï Ben Yaacov and with Diego Bejarano. One class of examples starts with finding a metric structure whose automorphism group is the group of increasing homeomorphisms of the unit interval. We will discuss some properties of this structure and extrapolate to other models of its theory, which we call 'dual linear continua.' Another source of examples includes real closed metric valued fields. These give rise to a notion of ordered metric structure, providing a viewpoint to study o-minimality in continuous logic.

We consider a computability-theoretic version of the ultraproduct construction for an infinite uniformly computable sequence of structures, where the role of an ultrafilter is played by an infinite set of natural numbers that cannot be split into two infinite subsets by any computably enumerable set. For computable structures, effective powers preserve only the first-order sentences of lower levels of quantifier complexity. Additional decidability of the structure increases preservation of the fragments of its theory in an effective power, so that a structure with a computable elementary diagram is elementarily equivalent to its effective power. We will present a number of recent collaborative results.

We say that a Turing degree is high in some context if it can always compute a correct answer given an input for which this is possible. When no correct answer is possible, however, what might such a degree do? We explore the possibilities in the context of computable structure theory. This is joint work with Wesley Calvert and Dan Turetsky.

Interest in transseries and Hardy fields comes from several fields, including asymptotic analysis, dynamical systems, and model theory of the real numbers. The first-order theory of (logarithmic-exponential) transseries and maximal Hardy fields is completely axiomatized by the theory of closed H-fields, which is model complete, as Aschenbrenner, Van den Dries, and Van der Hoeven have shown in a long series of works. I will describe my extension of this model completeness to tame pairs of closed H-fields, in order to better understand large closed H-fields, such as maximal Hardy fields, hyperseries, or surreal numbers. Time permitting, I may mention ongoing work on differential-algebraic dimension in transserial tame pairs.

This talk will focus on preference and utility theory in the context of o-minimal expansions of the ordered field of real numbers. We give a complete description of all preferences that can be defined in such a structure. We also raise questions about incomplete preferences in this context.