A countable-support symmetric iteration separating PP from AC

This paper constructs a transitive symmetric model of ZF + DC + PP + AC_wo + ¬AC by performing an Ord-length countable-support symmetric iteration over a Cohen seed model, utilizing a diagonal-lift scheme and Ryan–Smith localization to separate the partition principle from the axiom of choice while preserving dependent choice and the non-well-orderability of a specific set.

The Gist

Imagine the universe of mathematics as a giant, infinite library. For a long time, librarians (mathematicians) believed that every single book in this library could be neatly arranged on a shelf in a perfect, unbroken line from start to finish. This belief is called the Axiom of Choice (AC). It's the idea that if you have a pile of boxes, you can always pick one item from each box to make a new collection, even if the boxes are infinite and you can't see inside them all at once.

However, some mathematicians wondered: Is this "perfect ordering" actually a rule of the universe, or is it just a tool we invented?

This paper, written by Frank Trevor Gilson, is like a master architect building a new, alternative library to prove that the "perfect ordering" isn't strictly necessary. In this new library, you can still do almost everything you need, but you cannot arrange every single book in a perfect line.

Here is the story of how they built this library, explained through simple analogies.

1. The Starting Point: The "Seed" Library

The architects started with a standard, well-ordered library (the "Ground Model"). But they wanted to introduce a specific chaos: a giant pile of Cohen Reals. Think of these as a million identical-looking, unlabelled jars of paint.

• In the old library, you could easily say, "This is Jar #1, that's Jar #2."
• In the new library, the architects used a special "symmetry filter" (a magical rule) that makes all the jars look exactly the same to the librarian. You can swap any two jars, and the library looks unchanged.
• Result: You have a pile of jars (a set), but you cannot number them or order them. The "Axiom of Choice" is broken here.

2. The Problem: The "Partition Principle" (PP)

There is a famous rule called the Partition Principle (PP). It sounds like this:

"If you can pour water from a big bucket (Set A) into a smaller cup (Set B) so that the cup is full, you should be able to pour water back from the cup to the bucket to fill it up."

In math terms: If there is a surjection (a way to cover Set B using Set A), there should be an injection (a way to fit Set B inside Set A).

• In a normal library with the Axiom of Choice, this is obvious.
• The big question was: Does PP hold even if we don't have the Axiom of Choice?

For decades, no one knew if you could build a library where PP is true, but the Axiom of Choice is false. This paper says: Yes, we can.

3. The Construction: The "Infinite Construction Crew"

To build this library, the architects didn't just add one shelf; they built an infinite construction crew that works forever (an "Ord-length iteration").

Think of the library as a skyscraper being built floor by floor, but the building goes up infinitely high.

• The Goal: On every floor, they need to fix specific "leaks" in the logic.
• The Leaks: Sometimes, they find a situation where a big bucket can pour into a cup, but they can't pour back.
• The Fix: They use a special "Package Forcing." Imagine a construction crew that arrives with a magic toolkit.
  • If they see a "leak" where a surjection exists, the toolkit instantly builds a "back-pouring" mechanism (an injection).
  • They do this for every possible leak they can find, ensuring that the Partition Principle (PP) holds everywhere.

4. The Trick: "Orbit Symmetrization"

Here is the clever part. If they just added these tools, they might accidentally fix the "Jar" problem too, making the jars orderable again (which would bring back the Axiom of Choice). They didn't want that!

So, they used a technique called Orbit Symmetrization.

• Imagine the construction crew is wearing blindfolds and gloves. When they add a new tool to fix a leak, they add it in a way that is "symmetric."
• If you swap two jars, the new tool moves with them. It doesn't care which jar is which; it only cares about the relationship between them.
• This ensures that while they fix the "pouring" rules (PP), they never accidentally give the librarian a way to number the jars (AC). The jars remain a chaotic, unorderable pile.

5. The "Ryan-Smith" Shortcut

The architects didn't have to check every single book in the infinite library. They used a clever shortcut discovered by Ryan and Smith.

• They proved that if you fix the "pouring" rules for a specific, manageable type of book (let's call them "Seed Books"), and you have a few other small rules (like "Choice for well-ordered sets"), then automatically, the pouring rules work for every book in the library.
• This allowed them to focus their infinite construction crew on just a few specific tasks, knowing the rest would fall into place.

6. The Final Result: A "Just-Right" Library

When the infinite construction is finished, they open the doors to the new library (Model M). They check the rules:

1. Can we do math? Yes! (ZF is satisfied).
2. Can we make infinite choices step-by-step? Yes! (Dependent Choice, or DC, is satisfied).
3. Does the Partition Principle hold? Yes! If you can cover a set, you can fit it inside.
4. Can we order everything? NO! The pile of "Cohen Jars" remains unorderable. The Axiom of Choice is false.

The Big Takeaway

This paper is a tour de force of mathematical engineering. It proves that the Partition Principle (PP) is a weaker rule than the Axiom of Choice (AC).

• Before this, we didn't know if PP could exist without AC.
• Now we know: You can have a universe where "if you can cover it, you can fit it" is true, but "you can pick one from every set" is false.

It's like discovering a universe where you can perfectly match socks to shoes (PP), but you cannot organize the entire sock drawer into a numbered list (AC). The architects built this universe to show that the two rules are distinct, separating a long-standing mystery in the foundations of mathematics.

Technical Summary

1. Problem Statement

The central problem addressed is the independence of the Partition Principle (PP) from the Axiom of Choice (AC) within the framework of Zermelo-Fraenkel set theory (ZF).

• The Partition Principle (PP): States that if there is a surjection $A \twoheadrightarrow B$, then there is an injection $B \hookrightarrow A$.
• The Context: In ZFC, PP is trivial because every surjection has a right inverse (a choice function). However, in ZF, it is a long-standing open question whether PP implies AC.
• The Goal: To construct a model of ZF where PP holds, but AC fails (specifically, where the Axiom of Choice for well-orderable sets, ACWO, holds, but full AC fails). This would demonstrate that PP does not imply AC.

2. Methodology

The author employs a sophisticated Ord-length countable-support symmetric iteration starting from a Cohen symmetric seed model. The construction is modular, combining several advanced set-theoretic techniques:

A. The Seed Model ($N$)

• Base: The construction begins with a ground model $V \models \text{ZFC}$.
• Forcing: A Cohen symmetric extension is created using $\text{Add}(\omega, \omega_1)$ (adding $\omega_1$ Cohen reals).
• Symmetry: The group of automorphisms is $G = \text{Sym}(\omega_1)$, acting by permuting the indices of the Cohen reals. The filter of subgroups is generated by pointwise stabilizers of countable sets ($\text{Fix}(E)$ for $E \in [\omega_1]^{\leq \omega}$).
• Resulting Model ($N$): This seed model $N$ satisfies $\text{ZF} + \text{DC}$ (Dependent Choice) but fails AC because the set of Cohen reals $A = \{c_\alpha : \alpha < \omega_1\}$ is not well-orderable. Crucially, $N$ satisfies $\text{SVC}(A^\omega)$ (the Strong Axiom of Choice for $A^\omega$), meaning every set can be mapped onto by a product of $A^\omega$ and an ordinal.

B. The Iteration Strategy

The main construction is an iteration of length $\text{Ord}$ (the class of all ordinals) over the seed model $N$.

• Countable Support: The iteration uses countable support, ensuring that the Dependent Choice (DC) principle is preserved throughout the process.
• Package Forcing: At successor stages, the iteration forces "packages" designed to satisfy specific local principles:
  1. Localized PP-splitting ($Q_f$): For surjections $f: Y \twoheadrightarrow X$ where $X, Y \subseteq T$ (with $T = \mathcal{P}(A^\omega)$), the forcing adds a right inverse (section) for $f$. This forces the localized version of PP, denoted $\text{PP}_{\text{split}} \restriction T$.
  2. ACWO Forcing ($R_f$): For surjections $f: S \times \eta \twoheadrightarrow \lambda$ (where $S = A^\omega$ and $\lambda < \aleph^*(S)$), the forcing adds a right inverse. This ensures the Axiom of Choice for well-orderable sets (ACWO).
• Orbit-Symmetrization: To maintain symmetry and prevent the collapse of the non-well-orderable set $A$, the forcing is not applied to individual functions but to their orbits under the group action. The forcing conditions are countable partial sections of the entire orbit of a function.

C. Technical Innovations

• Diagonal-Lift/Diagonal-Cancellation: A novel machinery is introduced to define the limit filters at limit stages.
  • Diagonal Lifts: Automorphisms that act as the identity on a specific set of "protected" coordinates (package coordinates) while acting as a permutation on the rest.
  • Diagonal Cancellation: Used to ensure that the stabilizers of the generic objects (the sections added) belong to the filter, making them hereditarily symmetric.
  • Modified Limit Filter ($\tilde{F}^*_\lambda$): The author defines a new limit filter that includes subgroups generated by these diagonal lifts. This is crucial for proving that the final model satisfies $\text{ZF}$ and that the set $A$ remains non-well-orderable.
• Bookkeeping: A class-length bookkeeping scheme ensures that every relevant surjection instance appearing in the final model is eventually addressed by the forcing at some successor stage.

3. Key Contributions

1. Separation of PP and AC: The primary contribution is the construction of a model $M$ satisfying $\text{ZF} + \text{DC} + \text{PP} + \text{ACWO} + \neg \text{AC}$. This proves that PP does not imply AC, even when restricted to well-orderable choice.
2. Ryan–Smith Localization: The paper utilizes and refines the Ryan–Smith Localization Theorem, which states that under $\text{SVC}^+(T)$, the global PP is equivalent to the conjunction of localized PP ($\text{PP} \restriction T$) and ACWO. The construction explicitly forces these two components.
3. Diagonal-Cancellation Machinery: The development of diagonal-lift/diagonal-cancellation automorphisms provides a robust method for handling limit stages in symmetric iterations where standard support arguments fail (specifically at limits of cofinality $\omega$). This ensures the preservation of $\text{DC}$ and the non-well-orderability of the seed set.
4. Preservation of SVC: The paper proves that the property $\text{SVC}(S)$ (where $S = A^\omega$) persists through the entire Ord-length iteration, which is a non-trivial result for class-length symmetric iterations.

4. Main Results

The final symmetric model $M$ satisfies the following:

• $\text{ZF} + \text{DC}$: The model is a valid model of ZF with Dependent Choice.
• $\neg \text{AC}$: The set of Cohen reals $A$ is not well-orderable in $M$.
• $\text{ACWO}$: The Axiom of Choice holds for all well-orderable families of sets.
• $\text{PP}$: The Partition Principle holds globally.
  • Derivation: The iteration forces $\text{PP}_{\text{split}} \restriction T$ (every surjection between subsets of $T$ splits) and $\text{ACWO}$. Combined with the persistence of $\text{SVC}^+(T)$, the Ryan–Smith theorem upgrades these to global PP.
• Consequences:
  • The Ordering Principle holds (every set can be linearly ordered).
  • The Kinna–Wagner Principles ($\text{KWP}_1$ and $\text{KWP}_2$) hold.

5. Significance

• Resolution of a Major Open Problem: This paper provides a definitive negative answer to the question of whether PP implies AC. It establishes that PP is strictly weaker than AC in the context of ZF + DC.
• Advancement of Symmetric Iteration Theory: The techniques developed, particularly the diagonal-cancellation method for limit filters and the handling of Ord-length iterations with countable support, represent a significant technical advancement. These tools may be applicable to other problems involving the separation of choice principles or the construction of models with specific combinatorial properties.
• Refinement of Localization: The work demonstrates the power of the Ryan–Smith localization framework, showing how global choice principles can be reduced to localized splitting principles over a specific parameter set, provided the "seed" model satisfies strong structural properties like SVC.

In summary, Gilson constructs a complex, class-length symmetric iteration that carefully balances the addition of choice functions (to satisfy PP and ACWO) with the preservation of symmetry (to maintain $\neg$AC), successfully separating the Partition Principle from the full Axiom of Choice.