First, some context. Ever since I was a high schooler, I have been fascinated with large numbers. As I have grown in mathematical maturity, I have become both disappointed and fascinated to see that the process of naming larger and larger numbers requires a sort of philosophical tradeoff.

For example, an ultrafinitist might reject $10^{10^{100}}$ as being a valid representation of a number. A finitist might reject numbers defined using ordinal collapsing functions with large cardinals. A person who believes only in the constructible universe might reject the same ordinal collapsing functions if cardinals incompatible with $V=L$ are used. In general, the more you are willing to accept philosophically, the "larger" a number you can name concisely.

The interesting thing about these tradeoffs is that some of these philosophical stances have concrete impacts on the "real" world. For example, there is a number $n$ for which new axioms in addition to those of ZFC would be needed to prove the value of $BB(n)$. So in a way, this "real world" value changes depending on our philosophical stance.

My question then is: what sorts of axioms might we accept to strengthen as much as possible the values of $BB(n)$ which we can prove? And more generally, is there a way we can evaluate whether certain statements independent of ZFC are "true" based on their implications for the "real world" value of $BB(n)$ (if there is even a "true" value for this number).

EDIT: useful reading for those not familiar with the phenomenon of independence of $BB(n)$ from ZFC:

- Adam Yedidia, Scott Aaronson,
*A Relatively Small Turing Machine Whose Behavior Is Independent of Set Theory*, Complex Systems**25**(4) (2016), journal, arXiv:1605.04343

EDIT: As requested, an example of a "real world" consequence of a set theoretical axiom would be how Con(ZFC) would prove the Turing machine in the above paper doesn’t halt, whereas the negation of Con(ZFC) proves it halts in finite time.

15more comments