Monotonicity of the Ultrafilter Number Function

Basic Information

• Paper ID: 2501.14988
• Title: Monotonicity of the Ultrafilter Number Function
• Author: Toshimichi Usuba (Waseda University)
• Classification: math.LO (Mathematical Logic)
• Publication Date: January 2025 (arXiv preprint v4, updated November 20, 2025)
• Paper Link: https://arxiv.org/abs/2501.14988

Abstract

This paper investigates whether the ultrafilter number function $\kappa \mapsto u(\kappa)$ is monotone on cardinals, that is, whether $u(\lambda) \leq u(\kappa)$ holds for all cardinals $\lambda < \kappa$. The author proves that monotonicity can fail, but such failure has large cardinal strength. On the other hand, the paper establishes many restrictions on the failure of monotonicity. For instance, if $\kappa$ is a singular cardinal with countable cofinality or a strong limit singular cardinal, then $u(\kappa) \leq u(\kappa^+)$ holds.

Research Background and Motivation

Core Problem

For a cardinal $\kappa$, the ultrafilter number $u(\kappa)$ is defined as the minimum cardinality of a uniform ultrafilter base on $\kappa$. The core problem studied in this paper is: Is the ultrafilter number function $\kappa \mapsto u(\kappa)$ monotone?

Importance of the Problem

1. Natural extension of cardinal invariant research: Traditional research focuses on comparing the ultrafilter number at a single cardinal with other cardinal invariants (such as the relationship between $u(\kappa)$ and $2^\kappa$). This paper studies ultrafilter numbers from the perspective of function behavior, opening a new viewpoint.
2. Contrast with GCH: Under the Generalized Continuum Hypothesis (GCH), $u(\kappa) = 2^\kappa = \kappa^+$, so the ultrafilter number function is strictly increasing. Whether monotonicity persists in non-GCH environments is a natural question.
3. Connection with set-theoretic topology: Hart and van Mill also posed similar questions in the context of set-theoretic topology (Question 63 in 12).

Limitations of Existing Methods

• For certain cardinal invariants (such as the dominating number $d(\kappa)$), failure of monotonicity is easy to construct, but these methods do not apply to the ultrafilter number.
• Raghavan and Shelah proved the consistency of $u(\omega_{\omega+1}) < 2^{\omega_{\omega+1}}$, but did not address monotonicity.
• Few results are known for the ultrafilter number at successor cardinals.

Research Motivation

This paper aims to systematically study the monotonicity of the ultrafilter number function, revealing the consistency of its failure, its large cardinal strength, and ZFC restrictions.

Core Contributions

The main contributions of this paper include:

1. Consistency results for failure of monotonicity (Theorem 1.2):
   • Under appropriate large cardinal assumptions, there exist cardinals $\kappa, \lambda$ such that $\lambda < \kappa$ but $u(\kappa) < u(\lambda)$
   • In particular, $u(\omega_{\omega+1}) < u(\omega_1)$ is consistent
   • There exists a singular cardinal $\kappa$ such that $u(\kappa^+) < u(\kappa)$ is consistent
2. ZFC restrictions on failure of monotonicity (Theorem 1.3):
   • $u(\omega) \leq u(\kappa)$ holds for all cardinals $\kappa$
   • If $\lambda$ is regular and $u(\kappa) < u(\lambda)$, then $\lambda^{+\omega} \leq \kappa$
   • If $\kappa$ is a singular cardinal with countable cofinality or a strong limit singular cardinal, then $u(\kappa) \leq u(\kappa^+)$
   • If $u(\kappa^+) < u(\kappa)$, then $\{\lambda < \kappa \mid \lambda \text{ is regular}, u(\lambda) > u(\kappa^+)\}$ is bounded in $\kappa$
3. New results on indecomposable ultrafilters (Theorem 1.4):
   • If $\kappa$ is singular, $U$ is a $\kappa^+$-decomposable and $\text{cf}(\kappa)$-decomposable ultrafilter, then $U$ is also $\kappa$-decomposable
   • Improves results of Kunen-Prikry and Lipparini
4. Consistency strength results (Theorems 1.5, 1.6):
   • "There exist $\lambda < \kappa$ such that $u(\kappa) < u(\lambda)$" is equiconsistent with "there exists a measurable cardinal"
   • If there exist $\lambda \leq \kappa$ such that $u(\kappa^+) < u(\lambda)$, then there exists an inner model with a proper class of strong cardinals

Methodology Details

Core Technical Framework

The methodology of this paper rests on three technical pillars:

1. Indecomposable Ultrafilter Theory

Definition: An ultrafilter $U$ is $\lambda$-decomposable if there exists a function $f: S \to \lambda$ such that for all $X \in [\lambda]^{<\lambda}$, $f^{-1}(X) \notin U$.

Key connection (Lemma 5.1 and Corollary 5.2):

• If $\lambda < \kappa$ and $U$ is a $\lambda$-decomposable ultrafilter, then $u(\lambda) \leq \chi(U)$
• If $u(\kappa) < u(\lambda)$, then every uniform ultrafilter on $\kappa$ with character $u(\kappa)$ is $\lambda$-indecomposable

This establishes a direct connection between failure of monotonicity and existence of indecomposable ultrafilters.

2. Generalization of the Raghavan-Shelah Theorem

Theorem 3.1 (Core construction theorem): Let $\kappa, \mu$ be uncountable cardinals satisfying:

• $\text{cf}(\mu) < \kappa < \mu$
• $\nu^\kappa < \mu$ for all $\nu < \mu$
• $\kappa$ has a $\text{cf}(\mu)$-indecomposable uniform ultrafilter $U$

If a poset $P$ has $\text{cf}(\mu)$-c.c. and $|P| \leq \mu$, then $P$ forces $u(\kappa) \leq \mu$.

Proof strategy:

1. For each $P$-name $\dot{B} \subseteq \check{\kappa}$, use $\text{cf}(\mu)$-indecomposability to find $D \in U$ and $\gamma < \mu$ such that $\dot{B} \cap \check{D}$ is represented by $\gamma$-nice names
2. In the generic extension, extend $U$ to an ultrafilter $V$, proving that the family $\mathcal{G}$ generated by nice names is a base for $V$
3. Since $|\mathcal{G}| \leq \mu$, we have $u(\kappa) \leq \mu$

3. Application of PCF Theory at Singular Cardinals

Theorem 7.3 (Key technical theorem): Let $\kappa$ be singular and $U$ be an ultrafilter. If $U$ is $\kappa^+$-decomposable and $\text{cf}(\kappa)$-decomposable, then $U$ is also $\kappa$-decomposable.

Proof core:

1. By Kanamori's theorem, $U$ is $(\kappa, \kappa^+)$-regular
2. By Shelah's scale theory, there exists a scale of length $\kappa^+$: $\langle f_i \mid i < \kappa^+ \rangle$
3. Construct a function $h: S \to \kappa$; for each $s$, choose $h(s)$ larger than all $f_i(\xi_s)$ (for $i \in x_s$)
4. Prove that $h$ witnesses $\kappa$-decomposability

Construction of Models with Failure of Monotonicity

Proposition 4.4 (Basic construction): Let $\kappa$ be a cardinal, $\mu > \kappa$ a strong limit singular cardinal with $\omega_1 \leq \text{cf}(\mu) < \kappa$. Let $\lambda < \kappa$ satisfy $\text{cf}(\lambda) = \text{cf}(\mu)$. If $\kappa$ has a $\text{cf}(\lambda)$-indecomposable uniform ultrafilter, then $\text{Add}(\omega, \mu)$ forces $u(\kappa) < u(\lambda)$.

Application examples:

• Theorem 4.5: If $\kappa$ is measurable, $\mu > \kappa$ is strong limit singular with $\text{cf}(\mu) = \omega_1$, then $\text{Add}(\omega, \mu)$ forces $u(\kappa) < u(\omega_1)$
• Theorem 4.7: Under appropriate large cardinal assumptions, $u(\omega_{\omega+1}) < u(\omega_1)$ is consistent

Construction of Models with $u(\kappa) < u(\omega_\omega)$

Proposition 6.2 (Failure of diagonal reflection): Let $\kappa$ be a singular cardinal with countable cofinality, $\lambda < \kappa_0$ regular. If there exists a sequence $\langle \kappa_n, S_n \mid n < \omega \rangle$ satisfying specific non-reflecting stationary set conditions, then there is no $\lambda$-indecomposable uniform ultrafilter on $\kappa$.

Proposition 6.8 (Forcing construction): Under GCH, there exists a poset $P$ (inverse limit construction) adding a sequence satisfying the conditions of Proposition 6.2.

Theorem 6.11: If there is no $\omega_1$-indecomposable uniform ultrafilter on $\omega_\omega$, $\kappa$ is measurable, $\mu > \kappa$ is strong limit singular with $\text{cf}(\mu) = \omega_1$, then $\text{Add}(\omega, \mu)$ forces $u(\kappa) < u(\omega_\omega)$.

Experimental Setup

This paper is pure mathematical theoretical research and does not involve experiments or datasets. All results are established through rigorous mathematical proofs.

Main Results

Consistency of Failure of Monotonicity

Complete statement of Theorem 1.2: Under appropriate large cardinal assumptions, the following statements are consistent:

1. There exist $\kappa, \lambda$ such that $\lambda < \kappa$ but $u(\kappa) < u(\lambda)$
2. $u(\omega_{\omega+1}) < u(\omega_1)$
3. There exists a singular cardinal $\kappa$ with cofinality $\omega_1$ such that $u(\kappa^+) < u(\kappa)$
4. There exists a regular cardinal $\kappa$ such that $u(\kappa^+) < u(\omega_1)$
5. There exists a cardinal $\kappa > \omega_\omega$ such that $u(\kappa) < u(\omega_\omega)$
6. There exists a cardinal $\kappa$ such that $u(\kappa^{+\omega_1}) < u(\omega_\omega)$

ZFC Restrictions on Monotonicity

Complete statement of Theorem 1.3:

1. Proposition 5.3: $u(\omega) \leq u(\kappa)$ holds for all cardinals $\kappa$
   • Proof: Otherwise, $\kappa$ has a $\omega$-decomposable ultrafilter, i.e., a $\sigma$-complete ultrafilter, so there exists a measurable cardinal $\lambda \leq \kappa$, but $u(\omega) \leq 2^\omega < \lambda \leq u(\kappa)$, contradiction.
2. Proposition 5.4: If $\lambda$ is regular, $\lambda < \kappa$, but $u(\kappa) < u(\lambda)$, then $\lambda^{+\omega} \leq \kappa$
   • Proof: By Proposition 5.4(1), $u(\lambda) \leq u(\lambda^+) \leq u(\lambda^{++}) \leq \cdots$
3. Corollary 7.5: If $\kappa$ is a singular cardinal with countable cofinality, then $u(\kappa) \leq u(\kappa^+)$
   • Proof: Using Theorem 7.3, if $u(\kappa^+) < u(\kappa)$, then there exists a non-$\sigma$-complete but $\kappa$-decomposable ultrafilter, contradiction.
4. Theorem 7.9: If $\kappa$ is a strong limit singular cardinal, then $u(\kappa) \leq u(\kappa^+)$
   • Proof: Using Proposition 7.8, if $u(\kappa^+) < u(\kappa)$, then $2^\kappa = \kappa^+$, but $u(\kappa) \leq 2^\kappa$, contradiction.
5. Theorem 7.15: If $\kappa$ is singular and $u(\kappa^+) < u(\kappa)$, then there exists $\lambda < \kappa$ such that for all regular $\mu \in (\lambda, \kappa)$, $u(\mu) \leq u(\kappa^+)$
   • Proof: Using Proposition 7.12, the ultrafilter on $\kappa^+$ with character $u(\kappa^+)$ is almost $<\kappa$-decomposable.

Consistency Strength

Theorem 8.4 (Equiconsistency): The following theories are equiconsistent:

1. ZFC + "there exists a measurable cardinal"
2. ZFC + "there exist $\lambda < \kappa$ such that $u(\kappa) < u(\lambda)$"
3. ZFC + "there exist a weakly inaccessible cardinal $\kappa$ and $\lambda < \kappa$ such that $u(\kappa) < u(\lambda)$"
4. ZFC + "there exist a singular cardinal $\kappa$ and $\lambda < \kappa$ such that $u(\kappa) < u(\lambda)$"
5. ZFC + "there exist $\lambda < \kappa$ such that $\kappa$ has a $\lambda$-indecomposable uniform ultrafilter"

Theorem 8.16 (Lower bound): If there exist $\lambda \leq \kappa$ such that $u(\kappa^+) < u(\lambda)$, then there exists an inner model with a proper class of strong cardinals.

Proof strategy:

• Using Corollary 8.13, both $\square(\kappa)$ and $\square(\kappa^+)$ fail
• Using Schimmerling's Theorem 8.8, there exists an inner model with a proper class of strong cardinals

Connection with the Square Principle

Corollary 8.13:

1. If $\kappa$ is regular and there exist $\lambda < \kappa$ such that $u(\kappa) < u(\lambda)$, then $\square(\kappa)$ fails
2. If there exist $\lambda \leq \kappa$ such that $u(\kappa^+) < u(\lambda)$, then $\square_\kappa$ fails

Proposition 8.14: If $\kappa$ is regular and there exist $\lambda \leq \kappa$ such that $u(\kappa^+) < u(\lambda)$, then both $\square(\kappa)$ and $\square(\kappa^+)$ fail.

Related Work

Classical Research on Ultrafilter Numbers

1. Brendle and Shelah 5: Proved $u(\kappa) \geq \kappa^+$ and $\text{cf}(u(\kappa)) > \omega$
2. Garti and Shelah 8: Studied ultrafilter numbers at singular cardinals
3. Raghavan and Shelah 21: Proved the consistency of $u(\omega_{\omega+1}) < 2^{\omega_{\omega+1}}$ under large cardinal assumptions

Indecomposable Ultrafilters

1. Prikry and Silver 20: Existence of $\kappa$-indecomposable uniform ultrafilters implies stationary reflection and has large cardinal strength
2. Kunen and Prikry 17: Studied descendingly incomplete ultrafilters
3. Lipparini 19: Characterization of almost $<\kappa$-decomposable ultrafilters
4. Ben-David and Magidor 2: Construction of indecomposable ultrafilters on $\omega_{\omega+1}$ under GCH

Square Principle and Large Cardinals

1. Schimmerling 22: Failure of the square principle implies inner models with strong cardinals
2. Lambie-Hanson and Rinot 18, Inamdar and Rinot 14: Connection between indecomposable ultrafilters and stationary reflection

PCF Theory

1. Shelah 23: Foundations of PCF theory, existence of scales
2. Kanamori 15: Weakly normal filters and regularity results

Innovations of This Paper

• First systematic study of monotonicity of the ultrafilter number function
• Establishes direct connection between failure of monotonicity and indecomposable ultrafilters
• Improves Kunen-Prikry and Lipparini's results on decomposability
• Determines the precise consistency strength of failure of monotonicity

Conclusions and Discussion

Main Conclusions

1. Monotonicity can fail: Under appropriate large cardinal assumptions, the ultrafilter number function can be non-monotone, and even $u(\omega_{\omega+1}) < u(\omega_1)$ is possible
2. Failure has large cardinal strength: Failure of monotonicity is equiconsistent with the existence of a measurable cardinal; if $u(\kappa^+) < u(\lambda)$ for some $\lambda \leq \kappa$, then there exists an inner model with a proper class of strong cardinals
3. ZFC restrictions are ubiquitous:
   • $u(\omega)$ is always minimal
   • Monotonicity always holds at regular cardinals
   • Monotonicity holds at countable cofinality or strong limit singular cardinals
   • When monotonicity fails, there must be sufficient distance between failure points
4. Indecomposability is key: Failure of monotonicity is equivalent to existence of indecomposable ultrafilters, which is the core insight of this paper

Limitations

1. Necessity of large cardinal assumptions: Although it is proved that failure of monotonicity requires large cardinals, constructing models with failure also requires large cardinal assumptions. The precise strength of these assumptions (particularly for $u(\omega_{\omega+1}) < u(\omega_1)$) remains unclear
2. Some cases unresolved:
   • Does there exist $\lambda < \kappa$ such that $2^{<\lambda} = \lambda$ and $u(\kappa) < u(\lambda)$?
   • Can monotonicity fail at three or more cardinals?
3. Technical limitations:
   • For the case of countable cofinality, special constructions are needed (Propositions 6.2-6.8)
   • Cannot directly replace $\omega_1$ with $\omega$ in some results

Future Directions

The author poses the following in Problem 8.17:

1. Precise consistency strength: What is the precise consistency strength of $u(\kappa^+) < u(\lambda)$ for some $\lambda \leq \kappa$? Currently known to require an inner model with a proper class of strong cardinals, but the upper bound may be weaker.
2. Case of $2^{<\lambda} = \lambda$: Is it consistent that there exist $\lambda < \kappa$ such that $2^{<\lambda} = \lambda$ and $u(\kappa) < u(\lambda)$?
3. Multiple point failures: Do there exist three cardinals $\kappa_0 < \kappa_1 < \kappa_2$ such that:
   • $u(\kappa_2) < u(\kappa_1) < u(\kappa_0)$?
   • $u(\kappa_1) < u(\kappa_2) < u(\kappa_0)$?
   • $u(\kappa_2) < u(\kappa_0) < u(\kappa_1)$?
4. Other cardinal invariants: Are there other cardinal invariant functions $k(\kappa)$ whose failure of monotonicity also has large cardinal strength?

In-Depth Evaluation

Strengths

1. Pioneering work: First systematic study of the overall behavior of the ultrafilter number function, opening new research directions
2. Technical depth:
   • Skillfully combines forcing, large cardinals, PCF theory, and indecomposable ultrafilter theory
   • Theorem 7.3 improves the classical Kunen-Prikry result and has independent value
   • Construction of $u(\kappa) < u(\omega_\omega)$ models (Section 6) has high technical difficulty
3. Comprehensive results:
   • Proves both consistency results and ZFC restrictions
   • Provides both upper bounds (Theorem 8.16) and lower bounds (Theorem 8.4)
   • Covers regular, singular, countable cofinality, and uncountable cofinality cases
4. Clear structure:
   • Well-organized paper, progressing from simple to complex
   • Section 3's generalized Raghavan-Shelah theorem provides a unified framework for subsequent constructions
   • Section 7 systematically develops indecomposable ultrafilter theory with independent significance
5. Broad connections: Links ultrafilter numbers to the square principle, stationary reflection, and inner model theory

Weaknesses

1. Completeness of some proofs:
   • The proof of Lemma 4.2, though technically strong, is acknowledged by the author in Remark 4.3 to be unnecessary if one could prove $\text{cf}(u(\kappa)) \neq \text{cf}(\kappa)$ (a Garti-Shelah open problem)
   • Some results depend on "appropriate large cardinal assumptions" whose precise form is not explicitly stated
2. Complexity of constructions:
   • Section 6's construction of models with $u(\kappa) < u(\omega_\omega)$ involves complex inverse limit forcing and stationary set operations, with lower readability
   • Condition (3) in Proposition 6.2 is quite technical, making intuitive understanding difficult
3. Many open problems: Although significant progress is made, Problem 8.17 shows that many fundamental questions remain unresolved
4. Limited applications: As pure theoretical research, applications to other areas of mathematics are not yet apparent

Impact

1. Theoretical contribution:
   • Provides new perspective for cardinal invariant research (function behavior rather than point values)
   • Technical results like Theorem 7.3 will be cited in subsequent research
   • Establishes profound connection between failure of monotonicity and large cardinals
2. Methodological value:
   • Demonstrates how to synthesize forcing, large cardinals, and PCF theory
   • The diagonal construction technique in Proposition 6.2 may apply to other problems
3. Future research directions:
   • Problem 8.17 provides clear research directions
   • May inspire research on monotonicity of other cardinal invariant functions
   • Connections with the square principle and inner model theory may yield new results
4. Reproducibility: As pure mathematical proofs, results are completely verifiable, though specific implementations of some constructions (such as Section 6) require deep set-theoretic background

Applicable Scenarios

1. Set theory research:
   • Cardinal invariant theory
   • Large cardinal theory
   • Forcing theory
   • PCF theory
2. Related fields:
   • Set-theoretic topology (applications of ultrafilters in Stone-Čech compactifications)
   • Model theory (ultraproduct constructions)
   • Combinatorial set theory (infinite combinatorics)
3. Educational value:
   • Demonstrates comprehensive application of multiple modern set-theoretic techniques
   • Can serve as a case study for advanced set theory courses

Key References

2 S. Ben-David, M. Magidor. The weak □∗ is really weaker than the full □. J. Symb. Log. 51(4), 1029–1033 (1986).

5 J. Brendle, S. Shelah, Ultrafilters on ω–their ideals and their cardinal characteristics. Trans. Amer. Math. Soc. 351 (1999), no. 7, 2643–2674.

6 H.-D. Donder. Regularity of ultrafilters and the core model. Israel J. Math. 63(1988), no.3, 289–322.

15 A. Kanamori. Weakly normal filters and irregular ultrafilters. Trans. Amer. Math. Soc. 220(1976), 393–399.

17 K. Kunen, K. Prikry, On descendingly incomplete ultrafilters. J. Symbolic Logic 36 (1971), 650–652.

19 P. Lipparini, Decomposable ultrafilters and possible cofinalities. Notre Dame J. Form. Log. 49 (2008), no. 3, 307–312.

20 K. L. Prikry. On descendingly complete ultrafilters. In Cambridge Summer School in Mathematical Logic (1971), Lecture notes in Mathematics, 337, 459–488.

21 D. Raghavan, S. Shelah. A small ultrafilter number at smaller cardinals. Arch. Math. Logic 59(2020), no.3-4, 325–334.

22 E. Schimmerling. Coherent sequences and threads. Adv. Math. 216(2007), no.1, 89–117.

23 S. Shelah. Cardinal arithmetic. Oxford Logic Guides, 29. Oxford University Press, 1994.

Overall Assessment: This is a high-quality set theory research paper that pioneering studies the monotonicity problem of the ultrafilter number function. The paper has high technical depth, comprehensive results, both consistency constructions and ZFC restrictions, both upper and lower bounds. Technical results like Theorem 7.3 have independent value. Although some constructions are quite complex and several open problems remain, overall this is an important contribution to cardinal invariant theory that will have lasting impact on subsequent research.