---
title: Matet–Mathias Forcing
url: https://www.emergentmind.com/topics/matet-mathias-forcing
type: topic
---

# Matet–Mathias Forcing

Searching arXiv for the cited paper and closely related items.
arXiv_search query: "2507.15123"
Matet–Mathias forcing is the forcing notion $\mathbb{P}_{MM}(\mathcal{H})$ built from finite and infinite block-sequences in $FIN=[\omega]^{<\omega}\setminus\{\emptyset\}$ relative to a nonempty family $\mathcal{H}\subseteq FIN^{[\infty]}$. In the setting isolated in Section 3 of "There may be exactly $n$ $Q$-points" [2507.15123], conditions are pairs $\langle a,X\rangle$ consisting of a finite stem $a$ and an infinite remainder $X\in\mathcal{H}$ with the stem lying strictly below the remainder in the block-order. The forcing is used there in a restricted form, namely over Matet–adequate families, to obtain a length-$\omega_2$ countable support iteration that yields exactly two $Q$-points up to isomorphism in the final extension [2507.15123].

## 1. Block-sequences and the definition of $\mathbb{P}_{MM}(\mathcal{H})$

The ambient combinatorial space is the Milliken–Taylor space of finite “blocks.” One sets
$$
FIN=[\omega]^{<\omega}\setminus\{\emptyset\},
$$
and defines the block-order $<_b$ on $FIN$ by
$$
s<_b t \quad \text{iff} \quad \max(s)<\min(t).
$$
A finite block-sequence is any $a\in FIN^{[<\infty]}$, including the empty sequence $\emptyset$, and an infinite block-sequence is any $X\in FIN^{[\infty]}$. The notation $a\sqsubseteq b$ denotes the usual initial-segment relation on block-sequences, and $a^\frown b$ denotes concatenation when $a<b$ component-wise in the block-order [2507.15123].

Fix a nonempty family $\mathcal{H}\subseteq FIN^{[\infty]}$. The Matet–Mathias forcing relative to $\mathcal{H}$, denoted $\mathbb{P}_{MM}(\mathcal{H})$, consists of all pairs
$$
p=\langle a_p,X_p\rangle
$$
such that $a_p\in FIN^{[<\infty]}$, $X_p\in\mathcal{H}$, and $a_p<_b X_p$. Here $a_p<_b X_p$ means that every block in the finite sequence $a_p$ sits below the first block of $X_p$ in the $<_b$ order [2507.15123].

The order is defined by end-extension of the stem together with condensation of the remainder:
$$
\langle a_p,X_p\rangle \le_{\mathbb{P}} \langle a_q,X_q\rangle
$$
iff
1. $a_p\sqsupseteq a_q$,
2. $a_p\setminus a_q$ is a finite block-sequence lying inside $[X_q]$, and
3. $X_p\le X_q$ in the usual Milliken condensation order.

Equivalently, if one writes $a_p=b^\frown t$, then one requires $b=a_q$, $t\le X_q$, and also $X_p\le X_q$ [2507.15123]. In the unrestricted case $\mathcal{H}=FIN^{[\infty]}$, the notation is simply $\mathbb{P}_{MM}$.

This presentation makes the forcing bifurcated: a finite approximation is carried by the stem, while the infinite block-sequence controls admissible future extensions. A plausible implication is that much of the forcing’s behavior is determined not merely by the combinatorics of $FIN^{[\infty]}$ but by the closure and Ramsey properties of the chosen family $\mathcal{H}$.

## 2. Matet–adequate families

Eisworth isolated the axioms on $\mathcal{H}\subseteq FIN^{[\infty]}$ under which the forcing acquires its principal regularity features. Such a family is called Matet–adequate if it satisfies four requirements [2507.15123].

| Property | Requirement |
|---|---|
| finite-change | If $X\in\mathcal{H}$ and $Y$ differs from $X$ in only finitely many blocks, then $Y\in\mathcal{H}$ |
| upwards $\le^*$ | If $X\in\mathcal{H}$ and $Y\ge^* X$, then $Y\in\mathcal{H}$ |
| diagonal intersection | Every descending sequence $X_0\ge X_1\ge\cdots$ in $\mathcal{H}$ has $X\in\mathcal{H}$ with $X\le^* X_n$ for all $n$ |
| homogeneity | For every $X\in\mathcal{H}$ and every $2$-coloring $c:[X]\to 2$, there exists $Y\in\mathcal{H}$ with $Y\le X$ and $c\!\upharpoonright\![Y]$ constant |

Here $Y\ge^* X$ means $\exists n\, (Y/n\le X)$. The adequacy axioms therefore combine finite robustness, eventual upward closure, countable diagonal closure, and a Ramsey-type homogeneity principle [2507.15123].

Lemma 2.3, attributed to Eisworth and Mildenberger, strengthens this package in two ways. First, for each $n<\omega$ and any $r$-coloring $c:[X]^{[n]}\to r$, there is $Y\in\mathcal{H}$, $Y\le X$, on which $c$ is constant. Second, diagonal intersections may be chosen so that every block of the diagonal lies “far enough down” in each $X_n$ [2507.15123]. This suggests that Matet–adequacy is designed to support fusion arguments requiring both finite-dimensional Ramsey canonization and quantitative control over where diagonal blocks occur.

Two typical examples are recorded. The first is the full space $\mathcal{H}=FIN^{[\infty]}$, which yields the unrestricted forcing $\mathbb{P}_{MM}$. The second is the family
$$
FIN^{[\infty]}(\langle \mathcal{U}_0,\mathcal{U}_1\rangle),
$$
defined from a pair of Ramsey ultrafilters $\langle \mathcal{U}_0,\mathcal{U}_1\rangle$ by requiring infinite block-sequences to admit condensations hitting given sets in $\mathcal{U}_0$ on minima and $\mathcal{U}_1$ on maxima; Mildenberger shows that this family is Matet–adequate [2507.15123].

## 3. Pure decision, the $h$-Laver property, and preservation

When $\mathcal{H}$ is Matet–adequate, $\mathbb{P}_{MM}(\mathcal{H})$ satisfies a pure decision theorem. Theorem 3.1, due to Calderón–Di Prisco–Mijares, states that for every condition $p\in\mathbb{P}_{MM}(\mathcal{H})$ and every sentence $\varphi$ in the forcing language, there is an extension $q\le p$ which decides $\varphi$ [2507.15123].

A second structural fact is a Laver-type bounding property. Let
$$
h(n)=|\{a\in FIN^{[<\infty]}: a\le TOP\!\upharpoonright\!(n+1)\}|,
$$
where $TOP$ is the “top” block-sequence $\langle\{n\}:n<\omega\rangle$. If $\mathcal{H}$ is Matet–adequate and
$$
p\Vdash \text{``}\dot g\in\omega^\omega \text{ and } \forall n\, \dot g(n)<f(n)\text{''},
$$
then there are an extension $q\le p$ and a ground-model function $H$ with $|H(n)|\le h(n)$ such that
$$
q\Vdash \forall n\, [\,\dot g(n)\in H(n)\,].
$$
This is Lemma 3.2, the $h$-Laver property [2507.15123].

The proof sketch in the source is explicitly fusion-based. Pure decision first produces one value of $\dot g(0)$ on a decidable condition; then one enumerates all finite stems of length $1$, decides $\dot g(1)$ on each, thins out by homogeneity to make the choice independent of the stem, and continues inductively. A final diagonal intersection in $\mathcal{H}$ unifies the fusion into one condition [2507.15123].

From pure decision together with this Laver-type fusion, properness and preservation of $\omega_1$ follow “in the usual way,” and preservation of $\omega_2$ in the iteration follows by standard CH-arguments [2507.15123]. In this presentation, the forcing’s preservation theory is not stated abstractly but emerges from the interaction of decidability, homogeneity, and diagonal closure inside Matet–adequate families.

## 4. The length-$\omega_2$ countable-support iteration

The paper uses Matet–Mathias forcing restricted to a Matet–adequate family to produce exactly two $Q$-points. The iteration begins in a ground model $V\models CH$ with two nonisomorphic Ramsey ultrafilters $\mathcal{U}^0_0,\mathcal{U}^0_1$ [2507.15123].

By induction on $\alpha<\omega_2$, one defines a countable-support iteration
$$
P_\alpha,\quad Q_\alpha,
$$
where
$$
Q_\alpha=\mathbb{P}_{MM}\big(FIN^{[\infty]}(\langle \mathcal{U}^\alpha_0,\mathcal{U}^\alpha_1\rangle)\big).
$$
At successor stage $\alpha+1$, one forces with $Q_\alpha$ over $P_\alpha$, thereby adjoining a generic infinite block-sequence
$$
\eta^\alpha\in FIN^{[\infty]}(\langle \mathcal{U}^\alpha_0,\mathcal{U}^\alpha_1\rangle).
$$
Its two projections
$$
(\eta^\alpha)_{\min}=\{\min(\text{block}):\text{block}\in\eta^\alpha\},\qquad
(\eta^\alpha)_{\max}=\{\max(\text{block}):\text{block}\in\eta^\alpha\}
$$
form pseudo-intersections of the ultrafilters $\mathcal{U}^\alpha_0$ and $\mathcal{U}^\alpha_1$, respectively. Under $CH$ these are then extended to new Ramsey ultrafilters
$$
\mathcal{U}^{\alpha+1}_0\supseteq (\mathcal{U}^\alpha_0\cup (\eta^\alpha)_{\min}),\qquad
\mathcal{U}^{\alpha+1}_1\supseteq (\mathcal{U}^\alpha_1\cup (\eta^\alpha)_{\max})
$$
[2507.15123].

At limits of countable cofinality, one diagonal-intersects the prior ultrafilters. At limits of uncountable cofinality, no new reals are added, so the union of the towers remains Ramsey [2507.15123]. By standard facts about iterations of proper pure-decision forcings with fusion,
$$
P_{\omega_2}=\bigcup_{\alpha<\omega_2} P_\alpha
$$
is proper and preserves all cardinals $\le\omega_2$; moreover, it has the $\omega_2$-chain-condition, with a citation in the source to Abraham’s theorem [2507.15123].

Within the paper’s broader program, this iteration is presented as an alternative route for the case $n=2$: the abstract states that the statement for $n=2$ can be obtained by a length-$\omega_2$ countable support iteration of Matet–Mathias forcing restricted to a Matet–adequate family, while the full paper proves consistency of “There are exactly $n$ $Q$-points up to isomorphism” for any finite $n$ and describes this restricted iteration specifically for the case of two $Q$-points [2507.15123].

## 5. Interaction with $Q$-points

A $Q$-point is defined in the paper as an ultrafilter $E$ on $\omega$ such that for every function $f:\omega\to\omega$ there is $X\in E$ on which $f$ is finite-to-one; equivalently, every countable subset of $E$ has a pseudo-intersection in $E$ [2507.15123]. The forcing construction is calibrated so that two particular ultrafilters survive as $Q$-points while all others are ruled out.

Theorem 4.2 states that in $V^{P_{\omega_2}}$ there are exactly two $Q$-points up to isomorphism, namely
$$
U^\infty_0=\bigcup_{\alpha<\omega_2}\mathcal{U}^\alpha_0,\qquad
U^\infty_1=\bigcup_{\alpha<\omega_2}\mathcal{U}^\alpha_1
$$
[2507.15123].

The proof is by contradiction. If some ultrafilter $E$ in the extension were a $Q$-point not isomorphic to $U^\infty_0$ or $U^\infty_1$, then by standard closure it would already appear in some intermediate model $V^{P_\delta}$. The tail of the iteration above $\delta$ factors as
$$
P_{\omega_2}\cong P_\delta * (\mathbb{P}_{MM}(\mathcal{H}) * R),
$$
where
$$
\mathcal{H}=FIN^{[\infty]}(\langle \mathcal{U}^\delta_0,\mathcal{U}^\delta_1\rangle)
$$
and $R$ is the remainder, which has the Laver property. The argument then shows, exactly as in Proposition 4.4, that $\mathbb{P}_{MM}(\mathcal{H})*R$ destroys the $Q$-point property of any ultrafilter not isomorphic to one of the two diagonally-built Ramsey ultrafilters [2507.15123].

The significance of this factorization is methodological: the decisive combinatorics are concentrated in one restricted Matet–Mathias stage together with a tail forcing having a Laver-type bounding property. This suggests that the role of $\mathbb{P}_{MM}(\mathcal{H})$ is not only generative, via the block generic $\eta^\alpha$, but also eliminative, via a structured obstruction to the persistence of extraneous $Q$-points.

## 6. Local combinatorial lemmas and their role

Three lemmas provide the local mechanism by which the forcing closes off unwanted $Q$-points. They are formulated for $\mathbb{P}_{MM}(\mathcal{H})$ and are assembled in the proof of the main theorem for $n=2$ [2507.15123].

Lemma 4.6 is a finite-approximation decision statement. Given $p=\langle a,X\rangle$ and an $\mathbb{P}_{MM}(\mathcal{H})$-name $\dot C\subseteq\omega$, there is $X^*\le X$ in $\mathcal{H}$ such that for every finite block-sequence $b\le X^*$, the condition
$$
\langle a^\frown b, X^*/b\rangle
$$
decides the intersection $C\cap [0,\max\,\mathrm{last}(b)]$ as some finite set $C_b$, and moreover these $C_b$ do not “flip” on refinements of the same height [2507.15123]. This is a localized form of pure decision, adapted to finite initial segments of the relevant subset of $\omega$.

Lemma 4.7 is an interval-partitioning statement. If
$$
\langle a,X\rangle\Vdash \text{``}|C\cap (\max \eta(i-1),\max \eta(i)]|\le h(i)\text{ for all }i\text{,''}
$$
then one can thin out to force the existence of an interval partition
$$
0=k_0<k_1<\cdots
$$
such that whenever $C$ meets $[k_{i-1},k_i)$, the generic real $\eta$ must meet one of the three adjacent intervals [2507.15123]. The forcing thereby converts cardinality bounds indexed by generic block maxima into a geometric constraint on interval interaction.

Lemma 4.8 is the closing-off lemma. Given any partition $\{[k_i,k_{i+1})\}$ of $\omega$ and a $Q$-point $W$ not isomorphic to $U_0$ or $U_1$, then for any $X\in\mathcal{H}$ one can find $Y\le X$ and a set $w\in W$ such that whenever $Y$ hits $[k_{i-1},k_i)$, the set $w$ avoids all three adjacent intervals
$$
\{[k_{i-2},k_{i-1}),[k_{i-1},k_i),[k_i,k_{i+1})\}
$$
[2507.15123].

Putting these together exactly as in Proposition 4.4, the paper concludes that in the final extension any candidate ultrafilter $E$ distinct from the two designated towers is forced not to have the $Q$-point property [2507.15123]. A common misunderstanding would be to view the result as a mere preservation argument for two pre-existing ultrafilters. The local lemmas show that the construction is equally a destruction argument: the restricted Matet–Mathias forcing is used to engineer interval configurations incompatible with the defining pseudo-intersection behavior of other $Q$-points.

Source: https://www.emergentmind.com/topics/matet-mathias-forcing