Strong compactness and functions with the Menas property

6 minute read

Posted on

In this blog post I will travel back to the 70s and discuss about a result that appeared in (Menas, 1974). Briefly, the result says that we can always find fast functions for strongly compact cardinals which are limits of strongly compact cardinals, without the need to force one.

Before explaining what all this means, let me describe my motivation since this is a somewhat specialised post. First of all, the aforementioned paper is one of the most important ones in set theory, since among other results, it establishes the non-equivalence between strong compactness and supercompactness. One of the results in the article was recently brought to my attention by Arthur Apter and I was surprised that I could not find it in the modern literature. I believe it is quite beneficial to look at old results and recast them using modern tools, since it can make them more easily accessible to other researchers, without needing to trace the whole history. I am definitely not the first person who did this for Menas’ paper, since Joel Hamkins adapted Menas’ techniques in (Hamkins, 2000) to argue about the preservation of strong compactness in forcing extensions. In fact, part of this post is inspired by Hamkins’ approach.

To delve into the details now, recall that is -strongly compact if there is an elementary embedding with , , that satisfies the -covering property, meaning that for any that has size at most in , there is a set such that and .

Fortunately, these embeddings come with an ultrapower characterisation. If is a -complete fine ultrafilter on , the ultrapower embedding will have and As for the -covering property, one can see that the fineness of implies that is a cover for that has size less than in and using that, we can actually build a cover for any subset of of size at most .

The standard approach to preserve large cardinals in forcing extensions is to lift the corresponding elementary embeddings. Strong compactness embeddings are somewhat problematic when it comes to lifting arguments, since often the covering property is not enough. However, Menas noticed that things work better if we have a function such that we are able to pick for every a -strong compactness embedding for which can be quite large, namely larger than the size of the cover of . Not so surprisingly, Hamkins called this the Menas property. Let’s make this precise.

Definition

If is a strongly compact cardinal, then a function has the Menas property if for all there is -strong compactness embedding given by a fine ultrafilter on , such that .

What can we do with such a function? The point is that if we want to force a property at , we use a -iteration whose non-trivial stages are the closure points of the function . What happens then, is that the image of the iteration under will have a big jump, since is quite larger than . For someone familiar with lifting arguments, this gives the tail forcing a large degree of closure, so we can force to get a generic filter for it without seriously affecting the preservation argument. Back to the topic of the post, the result I am going to discuss about is the following.

Theorem (Menas, 1974)

If is strongly compact and a limit of strongly compact cardinals, then the function , where is the least strongly compact cardinal greater than , has the Menas propety.

Menas left open the question of the existence of such functions for arbitrary strongly compact cardinals, until Hamkins showed in (Hamkins, 2000) that we can always force one, using Woodin’s fast function forcing. I will give two proofs of Menas’ result, the first follows the original with the addition of some details, while the second uses the more modern framework of elementary embeddings.

Menas’ original proof

Pick an arbitrary and denote the club of closure points of by . Since is measurable, we can fix a normal ultrafilter on . Also, for each , let be a fine ultrafilter on . Since is normal and is a club, .

Now, we use the fact that we can compose these fine ultrafilters to define a fine ultrafilter on given by:

$$X\in U \leftrightarrow \{\alpha\in C\mid X\cap \mathcal{P}_{f(\alpha)}\lambda\in U_\alpha\}\in W.$$

(For a proof of this fact, see Theorem 22.19 (Kanamori, 2009)).

Let be the ultrapower embedding by with . We want to show that is the embedding that will satisfy the required property for .

For each , let . Since is -complete and fine, (consider sets in containing as a subset). It follows that . For each , let be the unique such that . Clearly, and we claim that the converse is also true. To see this, suppose and let , for some . We want to show that , where .

If we fix some , then for all , and so, . Also, it must be the case that , as otherwise either is not transitive or . By the definition of , it follows that . Hence,

$$X\cap D_\alpha=\{x\in D_\alpha \mid h(x)\in l(x)\}\in U_\alpha,$$

and in particular, . Since was arbitrary, it follows that and therefore, .

Now, it follows that and hence, .

As you can see, Menas’ original proof relies a lot on combinatorics and the particular properties of fine ultrafilters. In the next proof, I try to factor out these arguments and take advantage of the rich theory of elementary embeddings that we have today.

Modern proof

Firstly, although the Menas property refers only to ultrapower embeddings, we can actually relax this condition. Namely, if is a -strong compactness embedding with such that the cover of induced by the -covering property satisfies , then there is an ultrapower embedding that witnesses the Menas property. The construction is quite simple, since assuming we have such a , we use to define an ultrafilter on in the standard way:

$$X\in U \leftrightarrow s\in j(X).$$

Now, through a standard factoring argument it can be shown that (see Theorem 1.12 in (Hamkins, 2000)).

So, the proof boils down to constructing such a . To do this, pick an arbitrary and let be a -strong compactness embedding, with being the cover of induced by the -covering property. If in there is no strongly compact cardinal less or equal to , then we already know that is larger than and there is nothing to prove. Otherwise, let be the least strongly compact cardinal in above and below and let be a -strong compactness embedding with critical point , where is any ordinal greater than . This entails that there is such that and . Since in there is no strongly compact cardinal between and , by elementarity, in there is no strongly compact cardinal between and .

Diagram

We claim now that the embedding has the required property. It is easy to see that is a -strong compactness embedding with . Also, and has size less than in . As we said before, there is no strongly compact cardinal between and in , therefore as required.

References

  1. Menas, T. K. (1974). On strong compactness and supercompactness. Ann. Math. Logic, 7(327-359), 75.
  2. Hamkins, J. D. (2000). The lottery preparation. Ann. Pure Appl. Logic, 101(2-3), 103–146. Retrieved from https://doi.org/10.1016/S0168-0072(99)00010-X
  3. Kanamori, A. (2009). The higher infinite (Second, p. xxii+536). Berlin: Springer-Verlag.