Bounding 2D Functions by Products of 1D Functions

TL;DR

The paper investigates conditions under which 2D functions can be bounded by products of 1D functions, linking set-theoretic principles with large cardinal hypotheses.

Contribution

It demonstrates that certain bounding principles imply the existence of large cardinals under specific set-theoretic assumptions.

Findings

ZF + normal club filter + bounding principle implies existence of measurable cardinals.

In ZFC, the bounding principle for    is false.

The paper corrects a previous error in Welch's work and adjusts its consistency strength bounds.

Abstract

Given sets $X,Y$ and a regular cardinal $\mu$, let $\Phi(X,Y,\mu)$ be the statement that for any function $f : X \times Y \to \mu$, there are functions $g_1 : X \to \mu$ and $g_2 : Y \to \mu$ such that or all $(x,y) \in X \times Y$, $$f(x,y) \le \max \{ g_1(x), g_2(y) \}.$$ In ZFC, the statement $\Phi(\omega_1, \omega_1, \omega)$ is false. However, we show the theory ZF + ``the club filter on $\omega_1$ is normal'' + $\Phi(\omega_1, \omega_1, \omega)$ (which is implied by ZF + AD) implies that for every $\alpha < \omega_1$ there is a $\kappa \in (\alpha,\omega_1)$ such that in some inner model, $\kappa$ is measurable with Mitchell order $\ge \alpha$. There was an error in Welch's paper ``Characterizing Subsets of $\omega_1$ Constructible From a Real'', which he has retracted in a personal communication. Our paper originally referenced that paper. In this version of our paper, we are not using that result. Our consistency strength upper bound has changed accordingly.