Large Cardinal Preservation Across Inner Models and Forcing

77 concepts9 papers6 proposals

In set theory, large cardinal axioms act as the ultimate measure of consistency strength. By postulating the existence of infinite sets with extreme combinatorial properties, these axioms extend the standard Zermelo-Fraenkel framework (ZFC) and allow mathematicians to map out the boundaries of mathematical truth.

However, a cardinal's "largeness" is rarely an intrinsic property of the cardinal itself. Instead, it is typically extrinsic, depending heavily on the surrounding set-theoretic universe. A cardinal is often defined as large because the universe contains certain structural witnesses—such as elementary embeddings, ultrafilters, or paths through trees. If we restrict our view to a smaller subuniverse (an inner model) or expand our universe by adding new sets (forcing), the landscape of available witnesses changes. This raises a profound question: When do large cardinal properties survive transitions between different set-theoretic universes?

The Fragility of Largeness: Why Downwards Absoluteness Fails

Intuitively, one might expect that if a cardinal is incredibly large in the full universe of sets ($V$), it must retain some of that strength in smaller, inner models. After all, the cardinal still contains the same number of ordinals.

This intuition is wrong. Large cardinal properties are frequently not downwards absolute. When we pass to an inner model, we throw away sets. If we throw away the specific sets that witness a cardinal's largeness, the cardinal collapses into an ordinary, unremarkable ordinal.

A stark example of this divergence occurs in the inner model of hereditarily ordinal-definable sets ($HOD$). In Large cardinals need not be large in HOD, [Sy-David Friedman] et al. demonstrated that strong large cardinal properties in $V$ do not necessarily translate to $HOD$. Because $HOD$ only retains sets that can be uniquely defined by ordinals, it can easily omit the complex, non-definable ultrafilters or elementary embeddings that make a cardinal large in $V$. As a result, a cardinal can possess massive combinatorial power globally, yet appear completely ordinary within $HOD$.

   Universe V (Rich in sets)               Inner Model HOD (Thin)
+----------------------------+         +----------------------------+
|  Cardinal κ                |         |  Cardinal κ                |
|  [Witnessing Ultrafilter]  | ------> |  (Ultrafilter missing!)    |
|  (κ is Supercompact)       |         |  (κ is just an ordinary    |
+----------------------------+         |   regular cardinal)        |
                                       +----------------------------+

This fragility is also tied to the existence of sharp objects like $0^\sharp$, which encode transcendental truth about the constructible universe ($L$). In $0^{\sharp}$ and elementary end extensions of $V_{\kappa}$, Amir Leshem proved that if $0^\sharp$ exists, the property of weak compactness is not downwards absolute. Specifically, for any cardinal $\kappa$, there exists an inner model where the structure $(V_\kappa, \in)$ cannot be elementarily end-extended, stripping $\kappa$ of its weakly compact status. Interestingly, the author also showed that some properties are more resilient: a cardinal with uncountable cofinality in $L[0^\sharp]$ will at least retain its status as a Mahlo cardinal in any strict inner model of $L[0^\sharp]$.

Bridging the Gap via Structural Closeness

If large cardinals can disappear so easily in inner models, how can we ever guarantee their preservation? The answer lies in forcing the inner model to be structurally "close" to the full universe.

This is the core mechanism behind the HOD Hypothesis, which posits that $HOD$ is a good structural approximation of $V$. In The HOD Hypothesis and a supercompact cardinal, Yong Cheng explored how this hypothesis acts as a conduit for large cardinal strength. The author showed that if a supercompact cardinal exists in $V$ and the HOD Hypothesis holds, the large cardinal strength of $V$ successfully "leaks" into $HOD$, ensuring the existence of a proper class of measurable cardinals within $HOD$.

To prove this, the author utilized a formulation of supercompactness that focuses on the coherence of intermediate elementary embeddings. When $HOD$ is structurally close to $V$, the embeddings in $V$ can be systematically restricted to $HOD$ without falling apart, forcing $HOD$ to mirror the large cardinal profile of the wider universe.

Moving Upwards: Forcing and Indestructibility

Instead of shrinking the universe, we can expand it using set forcing. Forcing introduces new generic sets, which can potentially introduce counterexamples (such as a collapsing function or an unwanted tree) that destroy a large cardinal's properties. A cardinal is said to be indestructible if its large cardinal status survives certain classes of forcing.

To prove a cardinal survives forcing, logicians rely on the Lifting Criterion. If a large cardinal is defined by an elementary embedding $j: M \to N$, and we apply a forcing poset $\mathbb{P}$ to create a generic extension $M[G]$, we must find a way to "lift" the embedding to a new embedding $j^*: M[G] \to N[H]$. This requires constructing a generic filter $H$ for the image forcing $j(\mathbb{P})$ that is compatible with $G$.

Ground Model Embeddings                  Forcing Extension Lifts
     M -------- j --------> N                 M[G] ------- j* -------> N[H]
     |                      |                  |                        |
     | (Forcing with P)     |                  | (Forcing with j(P))    |
     v                      v                  v                        v
    M[G]                                      N[H]

In Indestructibility properties of remarkable cardinals, Victoria Gitman applied this machinery to remarkable cardinals—a class of cardinals that sit lower in the consistency hierarchy but mimic the embedding behavior of strong cardinals. By carefully managing the Lifting Criterion, the author demonstrated that remarkable cardinals possess robust indestructibility properties, remaining intact under forcing constructions that preserve successor cardinals. This resilience makes them highly effective tools for building custom models of set theory.

Calibrating Strength via Combinatorial Principles

Because large cardinals are so sensitive to the universe they inhabit, we can use specific combinatorial and definable principles to precisely calibrate their exact consistency strength.

One such benchmark is Harrington's Principle (HP), which asserts a deep connection between set-theoretic admissibility and the constructible universe $L$. In Harrington's Principle in Higher order Arithmetic, Ralf Schindler analyzed how the axiomatic strength of HP shifts when formalized in different systems of higher-order arithmetic. While HP is equiconsistent with standard ZFC in second-order arithmetic ($Z_2$), it experiences a dramatic escalation in third-order arithmetic ($Z_3$), becoming equiconsistent with ZFC plus the existence of a remarkable cardinal.

To understand why this happens, we must look at how these principles generalize. In The strong reflecting property and Harrington's Principle, Yong Cheng extended Harrington's Principle to arbitrary inner models ($HP(M)$) and connected it to the Strong Reflecting Property for L-cardinals (SRPL). The SRPL measures how effectively large cardinals reflect massive, complex structures down to smaller ordinals.

When we restrict our combinatorial principles to definable objects, we create a direct bridge to these reflection properties. For instance, Amir Leshem showed in On The Consistency of The Definable Tree Property on Alef-Symbol1 that requiring every definable $\omega_1$-tree to have an infinite path is exactly equivalent to the existence of a $\Pi_1^1$ reflecting cardinal. When paired with Martin's Axiom, the demand for paths escalates the requirement to a full weakly compact cardinal.

Finally, by using the forcing techniques established by Victoria Gitman, researchers can engineer universes where these principles hold without invoking unnecessarily massive large cardinals. In Forcing a set model of Z3 + Harrington's Principle, Yong Cheng constructed a specific set-theoretic model satisfying $Z_3$ and Harrington's Principle using only a remarkable cardinal, proving that HP does not necessitate the existence of $0^\sharp$.

The Ultimate Limits of Consistency

The ongoing study of how large cardinals behave across inner models and forcing extensions does more than just map out technical equivalences; it tests the absolute limits of the Zermelo-Fraenkel framework. As highlighted in the overview of foundational research in JSL volume 85 issue 2 Cover and Front matter by [Enrique Casanovas] et al., as we push further up the hierarchy to extremely large structures like Berkeley cardinals, we eventually reach a boundary where these strong axioms directly conflict with the Axiom of Choice.

By understanding exactly when large cardinal properties are preserved, when they diverge, and when they reflect, set theorists can safely navigate these boundaries—using forcing to build new mathematical worlds and inner models to isolate foundational truths.

Key Foundational Work

0^{♯} and elementary end extensions of 𝑉_{𝜅} (2001) — contributed 8 concept elements to this direction.
On The Consistency of The Definable Tree Property on Alef-Symbol1 (2000) — contributed 1 concept elements to this direction.
History & citationVersion 1 · 2026-10-04

Current · Permanent link

View frozen version 1

Cite this direction

Latest version (a living page):

@misc{lacuna_preservation_of_large_cardinal_propertie_074e4346,
  title = {preservation of large cardinal properties},
  author = {{Lacuna}},
  howpublished = {Lacuna research direction},
  year = {2026},
  url = {https://lacuna.tiptreesystems.com/direction/preservation-of-large-cardinal-properties/rdr_074e434655024346bd594d9fa19065a7},
  note = {Latest version of a living research direction; cite a version for a fixed text. Accessed 2026-11-21},
}

This exact version (version 1, frozen text):

@misc{lacuna_preservation_of_large_cardinal_propertie_074e4346_v1,
  title = {preservation of large cardinal properties},
  author = {{Lacuna}},
  howpublished = {Lacuna research direction},
  year = {2026},
  url = {https://lacuna.tiptreesystems.com/direction/preservation-of-large-cardinal-properties/rdr_074e434655024346bd594d9fa19065a7/v/1},
  note = {Version 1, frozen 2026-10-04. Accessed 2026-11-21},
}
About

These directions overlap with this one through shared papers.

See something wrong? Report a mistake