1 Introduction
 Throughout mathematics, the existence of various kinds of maximal sets can typically only be obtained by an appeal to the Axiom of Choice or one of its popular forms, such as Zorn’s Lemma. Under certain circumstances, it is possible though, to explicitly define such objects. The earliest result in this direction is probably due to Gödel who noted in [Reference Gödel15, p. 67] that in the constructible universe L, there is a 
 $\Delta ^1_2$
 well-order of the reals (see [Reference Jech18, Reference Schrittesser25] for a modern treatment). Using similar ideas, many other special sets of reals, such as Vitali sets, Hamel bases, or mad families, just to name a few, can be constructed in L in a
$\Delta ^1_2$
 well-order of the reals (see [Reference Jech18, Reference Schrittesser25] for a modern treatment). Using similar ideas, many other special sets of reals, such as Vitali sets, Hamel bases, or mad families, just to name a few, can be constructed in L in a 
 $\Delta ^1_2$
 way. This has become by now a standard set theoretic technique. In many cases, these results also give an optimal bound for the complexity of such a set. For example, a Vitali set cannot be Lebesgue measurable and in particular cannot have a
$\Delta ^1_2$
 way. This has become by now a standard set theoretic technique. In many cases, these results also give an optimal bound for the complexity of such a set. For example, a Vitali set cannot be Lebesgue measurable and in particular cannot have a 
 $\Sigma ^1_1$
 or
$\Sigma ^1_1$
 or 
 $\Pi ^1_1$
 definition. In other cases, one can get stronger results by constructing
$\Pi ^1_1$
 definition. In other cases, one can get stronger results by constructing 
 $\Pi ^1_1$
 witnesses. This is typically done using a coding technique, originally developed by Erdős, Kunen, and Mauldin in [Reference Erdős, Kunen and Mauldin8], later streamlined by Miller (see [Reference Miller21]) and further generalized by Vidnyánszky (see [Reference Vidnyánszky34]). For example, Miller showed that there are
$\Pi ^1_1$
 witnesses. This is typically done using a coding technique, originally developed by Erdős, Kunen, and Mauldin in [Reference Erdős, Kunen and Mauldin8], later streamlined by Miller (see [Reference Miller21]) and further generalized by Vidnyánszky (see [Reference Vidnyánszky34]). For example, Miller showed that there are 
 $\Pi ^1_1$
 Hamel bases and mad families in L. Other results of this type can be found, e.g., in [Reference Fischer, Friedman and Khomskii9], [Reference Fischer and Törnquist14], or [Reference Fischer and Schilhan11]. Since the assumption
$\Pi ^1_1$
 Hamel bases and mad families in L. Other results of this type can be found, e.g., in [Reference Fischer, Friedman and Khomskii9], [Reference Fischer and Törnquist14], or [Reference Fischer and Schilhan11]. Since the assumption 
 $V=L$
 is quite restrictive, it is interesting to know in what forcing extensions of L, definable witnesses for the above mentioned kinds of sets still exist. Various such results exist in the literature, e.g., in [Reference Brendle and Khomskii6], [Reference Fischer, Friedman and Zdomskyy10], [Reference Fischer, Schrittesser and Törnquist13], [Reference Schrittesser and Törnquist26], or [Reference Fischer and Schrittesser12].
$V=L$
 is quite restrictive, it is interesting to know in what forcing extensions of L, definable witnesses for the above mentioned kinds of sets still exist. Various such results exist in the literature, e.g., in [Reference Brendle and Khomskii6], [Reference Fischer, Friedman and Zdomskyy10], [Reference Fischer, Schrittesser and Törnquist13], [Reference Schrittesser and Törnquist26], or [Reference Fischer and Schrittesser12].
The starting observation for this paper is that almost all of these examples can be treated in the same framework, as maximal independent sets in hypergraphs.
Definition 1.1. A hypergraph E on a set X is a collection of finite non-empty subsets of X, i.e., 
 $E \subseteq [X]^{<\omega } \setminus \{ \emptyset \}$
. Whenever
$E \subseteq [X]^{<\omega } \setminus \{ \emptyset \}$
. Whenever 
 $Y \subseteq X$
, we say that Y is E-independent if
$Y \subseteq X$
, we say that Y is E-independent if 
 $[Y]^{<\omega } \cap E = \emptyset $
. Moreover, we say that Y is maximal E-independent if Y is maximal under inclusion in the collection of E-independent subsets of X.
$[Y]^{<\omega } \cap E = \emptyset $
. Moreover, we say that Y is maximal E-independent if Y is maximal under inclusion in the collection of E-independent subsets of X.
 Whenever X is a topological space, 
 $[X]^{<\omega }$
 is the disjoint sum of the spaces
$[X]^{<\omega }$
 is the disjoint sum of the spaces 
 $[X]^n$
 for
$[X]^n$
 for 
 $n \in \omega $
. Here, as usual,
$n \in \omega $
. Here, as usual, 
 $[X]^n$
, the set of subsets of X of size n becomes a topological space by identification with the quotient of
$[X]^n$
, the set of subsets of X of size n becomes a topological space by identification with the quotient of 
 $X^n$
 under the equivalence relation
$X^n$
 under the equivalence relation 
 $(x_0, \dots , x_{n-1}) \sim (y_0, \dots , y_{n-1})$
 iff
$(x_0, \dots , x_{n-1}) \sim (y_0, \dots , y_{n-1})$
 iff 
 $\{x_0, \dots , x_{n-1}\} = \{y_0, \dots , y_{n-1} \}$
. Whenever X is Polish,
$\{x_0, \dots , x_{n-1}\} = \{y_0, \dots , y_{n-1} \}$
. Whenever X is Polish, 
 $[X]^{<\omega }$
 is Polish as well and we can study its definable subsets. In particular, we can study definable hypergraphs on Polish spaces.
$[X]^{<\omega }$
 is Polish as well and we can study its definable subsets. In particular, we can study definable hypergraphs on Polish spaces.
The main result of this paper is the following theorem.
Theorem 1.2. After forcing with the 
 $\omega _2$
-length countable support iteration (csi) of Sacks or splitting forcing over L, every analytic hypergraph on a Polish space has a
$\omega _2$
-length countable support iteration (csi) of Sacks or splitting forcing over L, every analytic hypergraph on a Polish space has a 
 $\mathbf {\Delta }^1_2$
 maximal independent set.
$\mathbf {\Delta }^1_2$
 maximal independent set.
 This extends a result by Schrittesser [Reference Schrittesser25], who proved the above for Sacks forcing, which we denote by 
 $\mathbb {S}$
, and ordinary two-dimensional graphs (see also [Reference Schrittesser and Törnquist26]). For equivalence relations this was already known by Budinas [Reference Budinas7]. We will also prove the case of finite products but our main focus will be on the countable support iteration. Splitting forcing
$\mathbb {S}$
, and ordinary two-dimensional graphs (see also [Reference Schrittesser and Törnquist26]). For equivalence relations this was already known by Budinas [Reference Budinas7]. We will also prove the case of finite products but our main focus will be on the countable support iteration. Splitting forcing 
 $\mathbb {SP}$
 (Definition 4.1) is a less-known forcing notion that was originally introduced by Shelah in [Reference Shelah, Judah, Just and Woodin27] and has been studied in more detail recently [Reference Hein and Spinas17, Reference Laguzzi, Mildenberger and Stuber-Rousselle20, Reference Spinas29, Reference Spinas30]. Although it is very natural and gives a minimal way to add a splitting real (see more below), it has not been exploited a lot and to the best of our knowledge, there is no major set theoretic text treating it in more detail.
$\mathbb {SP}$
 (Definition 4.1) is a less-known forcing notion that was originally introduced by Shelah in [Reference Shelah, Judah, Just and Woodin27] and has been studied in more detail recently [Reference Hein and Spinas17, Reference Laguzzi, Mildenberger and Stuber-Rousselle20, Reference Spinas29, Reference Spinas30]. Although it is very natural and gives a minimal way to add a splitting real (see more below), it has not been exploited a lot and to the best of our knowledge, there is no major set theoretic text treating it in more detail.
Our three guiding examples for Theorem 1.2 will be ultrafilters, maximal independent families, and Hamel bases.
 Recall that an ultrafilter on 
 $\omega $
 is a maximal subset
$\omega $
 is a maximal subset 
 $\mathcal {U}$
 of
$\mathcal {U}$
 of 
 $\mathcal {P}(\omega )$
 with the strong finite intersection property, i.e., the property that for any
$\mathcal {P}(\omega )$
 with the strong finite intersection property, i.e., the property that for any 
 $\mathcal {A} \in [\mathcal {U}]^{<\omega }$
,
$\mathcal {A} \in [\mathcal {U}]^{<\omega }$
, 
 $\vert \bigcap \mathcal {A} \vert = \omega $
.Footnote 
1
 Thus, letting
$\vert \bigcap \mathcal {A} \vert = \omega $
.Footnote 
1
 Thus, letting 
 $E_u := \{ \mathcal {A} \in [\mathcal {P}(\omega )]^{<\omega } : \vert \bigcap \mathcal {A} \vert < \omega \}$
, an ultrafilter is a maximal
$E_u := \{ \mathcal {A} \in [\mathcal {P}(\omega )]^{<\omega } : \vert \bigcap \mathcal {A} \vert < \omega \}$
, an ultrafilter is a maximal 
 $E_u$
-independent set. In [Reference Schilhan23], we studied the projective definability of ultrafilters and introduced the cardinal invariant
$E_u$
-independent set. In [Reference Schilhan23], we studied the projective definability of ultrafilters and introduced the cardinal invariant 
 $\mathfrak {u}_B$
, which is the smallest size of a collection of Borel subsets of
$\mathfrak {u}_B$
, which is the smallest size of a collection of Borel subsets of 
 $\mathcal {P}(\omega )$
 whose union is an ultrafilter. If there is a
$\mathcal {P}(\omega )$
 whose union is an ultrafilter. If there is a 
 $\mathbf {\Sigma }^1_2$
 ultrafilter, then
$\mathbf {\Sigma }^1_2$
 ultrafilter, then 
 $\mathfrak {u}_B = \omega _1$
, since every
$\mathfrak {u}_B = \omega _1$
, since every 
 $\mathbf {\Sigma }^1_2$
 set is the union of
$\mathbf {\Sigma }^1_2$
 set is the union of 
 $\omega _1$
 many Borel sets. Recall that the classical ultrafilter number
$\omega _1$
 many Borel sets. Recall that the classical ultrafilter number 
 $\mathfrak {u}$
 is the smallest size of an ultrafilter base. We showed in [Reference Schilhan23], that
$\mathfrak {u}$
 is the smallest size of an ultrafilter base. We showed in [Reference Schilhan23], that 
 $\mathfrak {u}_B \leq \mathfrak {u}$
 and asked whether it is consistent that
$\mathfrak {u}_B \leq \mathfrak {u}$
 and asked whether it is consistent that 
 $\mathfrak {u}_B < \mathfrak {u}$
 or even whether a
$\mathfrak {u}_B < \mathfrak {u}$
 or even whether a 
 $\Delta ^1_2$
 ultrafilter can exist while
$\Delta ^1_2$
 ultrafilter can exist while 
 $\omega _1 <\mathfrak {u}$
. The difficulty is that we have to preserve a definition for an ultrafilter, while its interpretation in L must be destroyed. This has been achieved before for mad families (see [Reference Brendle and Khomskii6]).
$\omega _1 <\mathfrak {u}$
. The difficulty is that we have to preserve a definition for an ultrafilter, while its interpretation in L must be destroyed. This has been achieved before for mad families (see [Reference Brendle and Khomskii6]).
 An independent family is a subset 
 $\mathcal {I}$
 of
$\mathcal {I}$
 of 
 $\mathcal {P}(\omega )$
 so that for any disjoint
$\mathcal {P}(\omega )$
 so that for any disjoint 
 $\mathcal {A}_0, \mathcal {A}_1 \in [\mathcal {I}]^{<\omega }$
,
$\mathcal {A}_0, \mathcal {A}_1 \in [\mathcal {I}]^{<\omega }$
, 
 $\vert \bigcap _{x \in \mathcal {A}_0} x \cap \bigcap _{x \in \mathcal {A}_1} \omega \setminus x \vert = \omega $
. It is called maximal independent family if it is additionally maximal under inclusion. Thus, letting
$\vert \bigcap _{x \in \mathcal {A}_0} x \cap \bigcap _{x \in \mathcal {A}_1} \omega \setminus x \vert = \omega $
. It is called maximal independent family if it is additionally maximal under inclusion. Thus, letting 
 $E_i = \{ \mathcal {A}_0 \dot \cup \mathcal {A}_1 \in [\mathcal {P}(\omega )]^{<\omega } :\vert \bigcap _{x \in \mathcal {A}_0} x \cap \bigcap _{x \in \mathcal {A}_1} \omega \setminus x \vert < \omega \}$
, a maximal independent family is a maximal
$E_i = \{ \mathcal {A}_0 \dot \cup \mathcal {A}_1 \in [\mathcal {P}(\omega )]^{<\omega } :\vert \bigcap _{x \in \mathcal {A}_0} x \cap \bigcap _{x \in \mathcal {A}_1} \omega \setminus x \vert < \omega \}$
, a maximal independent family is a maximal 
 $E_i$
-independent set. The definability of maximal independent families was studied by Miller in [Reference Miller21], who showed that they cannot be analytic, and recently by Brendle, Fischer, and Khomskii in [Reference Brendle, Fischer and Khomskii5], where they introduced the invariant
$E_i$
-independent set. The definability of maximal independent families was studied by Miller in [Reference Miller21], who showed that they cannot be analytic, and recently by Brendle, Fischer, and Khomskii in [Reference Brendle, Fischer and Khomskii5], where they introduced the invariant 
 $\mathfrak {i}_B$
, the least size of a collection of Borel sets whose union is a maximal independent family. The classical independence number
$\mathfrak {i}_B$
, the least size of a collection of Borel sets whose union is a maximal independent family. The classical independence number 
 $\mathfrak {i}$
 is simply the smallest size of a maximal independent family. In [Reference Brendle, Fischer and Khomskii5], it was asked whether
$\mathfrak {i}$
 is simply the smallest size of a maximal independent family. In [Reference Brendle, Fischer and Khomskii5], it was asked whether 
 $\mathfrak {i}_B < \mathfrak {i}$
 is consistent and whether there can be a
$\mathfrak {i}_B < \mathfrak {i}$
 is consistent and whether there can be a 
 $\Pi ^1_1$
 maximal independent family while
$\Pi ^1_1$
 maximal independent family while 
 $\omega _1 < \mathfrak {i}$
. In the same article, it was shown that the existence of a
$\omega _1 < \mathfrak {i}$
. In the same article, it was shown that the existence of a 
 $\Delta ^1_2$
 maximal independent family is equivalent to that of a
$\Delta ^1_2$
 maximal independent family is equivalent to that of a 
 $\Pi ^1_1$
 such family. The difficulty in the problem is similar to that before.
$\Pi ^1_1$
 such family. The difficulty in the problem is similar to that before.
 A Hamel basis is a vector-space basis of 
 $\mathbb {R}$
 over the field of rationals
$\mathbb {R}$
 over the field of rationals 
 $\mathbb {Q}$
. Thus, letting
$\mathbb {Q}$
. Thus, letting 
 $E_h := \{ \mathcal {A} \in [\mathbb {R}]^{<\omega } : \mathcal {A} \text { is linearly dependent over } \mathbb {Q} \}$
, a Hamel basis is a maximal
$E_h := \{ \mathcal {A} \in [\mathbb {R}]^{<\omega } : \mathcal {A} \text { is linearly dependent over } \mathbb {Q} \}$
, a Hamel basis is a maximal 
 $E_h$
-independent set. A Hamel basis must be as large as the continuum itself. This is reflected in the fact that, when adding a real, every ground-model Hamel basis is destroyed. But still it makes sense to ask how many Borel sets are needed to get one. Miller, also in [Reference Miller21], showed that a Hamel basis can never be analytic. As before, we may ask whether there can be a
$E_h$
-independent set. A Hamel basis must be as large as the continuum itself. This is reflected in the fact that, when adding a real, every ground-model Hamel basis is destroyed. But still it makes sense to ask how many Borel sets are needed to get one. Miller, also in [Reference Miller21], showed that a Hamel basis can never be analytic. As before, we may ask whether there can be a 
 $\Delta ^1_2$
 Hamel basis while CH fails. Again, destroying ground-model Hamel bases seems to pose a major obstruction.
$\Delta ^1_2$
 Hamel basis while CH fails. Again, destroying ground-model Hamel bases seems to pose a major obstruction.
 The most natural way to increase 
 $\mathfrak {u}$
 and
$\mathfrak {u}$
 and 
 $\mathfrak {i}$
 is by iteratively adding splitting reals. Recall that for
$\mathfrak {i}$
 is by iteratively adding splitting reals. Recall that for 
 $x,y \in \mathcal {P}(\omega )$
, we say that x splits y iff
$x,y \in \mathcal {P}(\omega )$
, we say that x splits y iff 
 $ \vert x \cap y \vert = \omega $
 and
$ \vert x \cap y \vert = \omega $
 and 
 $\vert y \setminus x \vert = \omega $
. A real x is called splitting over V iff for every
$\vert y \setminus x \vert = \omega $
. A real x is called splitting over V iff for every 
 $y \in \mathcal {P}(\omega ) \cap V$
, x splits y. The classical forcing notions adding splitting reals are Cohen, Random, and Silver forcing and all forcings that add so-called dominating reals. It was shown though, in [Reference Schilhan23], that after forcing with any of these, a
$y \in \mathcal {P}(\omega ) \cap V$
, x splits y. The classical forcing notions adding splitting reals are Cohen, Random, and Silver forcing and all forcings that add so-called dominating reals. It was shown though, in [Reference Schilhan23], that after forcing with any of these, a 
 $\mathbf \Sigma ^1_2$
 definition with ground model parameters will not define an ultrafilter and the same argument can be applied to independent families. For this reason, we are going to use the forcing notion
$\mathbf \Sigma ^1_2$
 definition with ground model parameters will not define an ultrafilter and the same argument can be applied to independent families. For this reason, we are going to use the forcing notion 
 $\mathbb {SP}$
 that we mentioned above. As an immediate corollary of Theorem 1.2, we get the following.
$\mathbb {SP}$
 that we mentioned above. As an immediate corollary of Theorem 1.2, we get the following.
Theorem 1.3. It is consistent that 
 $\mathfrak {r} = \mathfrak {u} = \mathfrak {i} = \omega _2$
 while there is a
$\mathfrak {r} = \mathfrak {u} = \mathfrak {i} = \omega _2$
 while there is a 
 $\Delta ^1_2$
 ultrafilter, a
$\Delta ^1_2$
 ultrafilter, a 
 $\Pi ^1_1$
 maximal independent family, and a
$\Pi ^1_1$
 maximal independent family, and a 
 $\Delta ^1_2$
 Hamel basis. In particular, we get the consistency of
$\Delta ^1_2$
 Hamel basis. In particular, we get the consistency of 
 $\,\mathfrak {i}_B, \mathfrak {u}_B < \mathfrak {r}, \mathfrak {i}, \mathfrak {u}$
.
$\,\mathfrak {i}_B, \mathfrak {u}_B < \mathfrak {r}, \mathfrak {i}, \mathfrak {u}$
.
 Here, 
 $\mathfrak {r}$
 is the reaping number, the least size of a set
$\mathfrak {r}$
 is the reaping number, the least size of a set 
 $\mathcal {S} \subseteq \mathcal {P}(\omega )$
 so that there is no splitting real over
$\mathcal {S} \subseteq \mathcal {P}(\omega )$
 so that there is no splitting real over 
 $\mathcal {S}$
. This solves the abovementioned questions from [Reference Brendle, Fischer and Khomskii5, Reference Schilhan23]. Moreover, Theorem 1.2 gives a “black-box” way to get many results, saying that certain definable families exists in the Sacks model.
$\mathcal {S}$
. This solves the abovementioned questions from [Reference Brendle, Fischer and Khomskii5, Reference Schilhan23]. Moreover, Theorem 1.2 gives a “black-box” way to get many results, saying that certain definable families exists in the Sacks model.
 In [Reference Brendle, Fischer and Khomskii5], another cardinal invariant 
 $\mathfrak {i}_{cl}$
 is introduced, which is the smallest size of a collection of closed sets, whose union is a maximal independent family. Similarly, one can define a closed version of the ultrafilter number,
$\mathfrak {i}_{cl}$
 is introduced, which is the smallest size of a collection of closed sets, whose union is a maximal independent family. Similarly, one can define a closed version of the ultrafilter number, 
 $\mathfrak {u}_{cl}$
. Here, it is irrelevant whether we consider closed subsets of
$\mathfrak {u}_{cl}$
. Here, it is irrelevant whether we consider closed subsets of 
 $[\omega ]^\omega $
 or
$[\omega ]^\omega $
 or 
 $\mathcal {P}(\omega )$
, since every closed subset of
$\mathcal {P}(\omega )$
, since every closed subset of 
 $[\omega ]^\omega $
 with the strong finite intersection property is
$[\omega ]^\omega $
 with the strong finite intersection property is 
 $\sigma $
-compact (see Lemma 5.6). In the model of Theorem 1.3, we have that
$\sigma $
-compact (see Lemma 5.6). In the model of Theorem 1.3, we have that 
 $\mathfrak {i}_{cl} = \mathfrak {i}_B$
 and
$\mathfrak {i}_{cl} = \mathfrak {i}_B$
 and 
 $\mathfrak {u}_{cl} = \mathfrak {u}_B$
, further answering the questions of Brendle, Fischer, and Khomskii. On the other hand we show that
$\mathfrak {u}_{cl} = \mathfrak {u}_B$
, further answering the questions of Brendle, Fischer, and Khomskii. On the other hand we show that 
 $\mathfrak {d} \leq \mathfrak {i}_{cl}$
, mirroring Shelah’s result that
$\mathfrak {d} \leq \mathfrak {i}_{cl}$
, mirroring Shelah’s result that 
 $\mathfrak {d} \leq \mathfrak {i}$
 (see [Reference Vaughan32]). Here,
$\mathfrak {d} \leq \mathfrak {i}$
 (see [Reference Vaughan32]). Here, 
 $\mathfrak {d}$
 is the dominating number, the least size of a dominating family in
$\mathfrak {d}$
 is the dominating number, the least size of a dominating family in 
 $(\omega ^\omega , <^*)$
.
$(\omega ^\omega , <^*)$
.
Theorem 1.4. (ZFC) 
 $\mathfrak {d} \leq \mathfrak {i}_{cl}$
.
$\mathfrak {d} \leq \mathfrak {i}_{cl}$
.
The paper is organized as follows. In Section 2, we will consider basic results concerning iterations of tree forcings. This section is interesting in its own right and can be read independently from the rest. More specifically, we prove a version of continuous reading of names for countable support iterations that is widely applicable (Lemma 2.2). In Section 3, we prove our main combinatorial lemma (Main Lemmas 3.4 and 3.14) which is at the heart of Theorem 1.2. As for Section 2, Section 3 can be read independently of the rest, since our result is purely descriptive set theoretical. In Section 4, we introduce splitting and Sacks forcing and place it in bigger class of forcings to which we can apply the main lemma. This combines the results from Sections 2 and 3. In Section 4, we bring everything together and prove Theorems 1.2–1.4. We end with concluding remarks concerning the further outlook of our technique and pose some questions.
2 Tree forcing
 Let A be a fixed countable set, usually 
 $\omega $
 or
$\omega $
 or 
 $2$
.
$2$
. 
- 
(a) A tree T on A is a subset of  $A^{<\omega }$
 so that for every $A^{<\omega }$
 so that for every $t \in T$
 and $t \in T$
 and $n < \vert t \vert $
, $n < \vert t \vert $
, $t \restriction n \in T$
, where $t \restriction n \in T$
, where $\vert t \vert $
 denotes the length of t. For $\vert t \vert $
 denotes the length of t. For $s_0, s_1 \in A^{<\omega }$
, we write $s_0, s_1 \in A^{<\omega }$
, we write $s_0 \perp s_1$
 whenever $s_0 \perp s_1$
 whenever $s_0 \not \subseteq s_1$
 and $s_0 \not \subseteq s_1$
 and $s_1 \not \subseteq s_0$
. $s_1 \not \subseteq s_0$
.
- 
(b) T is perfect if for every  $t \in T$
 there are $t \in T$
 there are $s_0, s_1 \in T$
 so that $s_0, s_1 \in T$
 so that $s_0, s_1\supseteq t$
 and $s_0, s_1\supseteq t$
 and $s_0 \perp s_1$
. $s_0 \perp s_1$
.
- 
(c) A node  $t \in T$
 is called a splitting node, if there are $t \in T$
 is called a splitting node, if there are $i \neq j \in A$
 so that $i \neq j \in A$
 so that $t^{\frown } i, t^{\frown } j \in T$
. The set of splitting nodes in T is denoted $t^{\frown } i, t^{\frown } j \in T$
. The set of splitting nodes in T is denoted $\operatorname {\mathrm {split}}(T)$
. We define $\operatorname {\mathrm {split}}(T)$
. We define $\operatorname {\mathrm {split}}_n(T)$
 to be the set of $\operatorname {\mathrm {split}}_n(T)$
 to be the set of $t \in \operatorname {\mathrm {split}}(T)$
 such that there are exactly n splitting nodes below t in T. The finite subtree of T generated by $t \in \operatorname {\mathrm {split}}(T)$
 such that there are exactly n splitting nodes below t in T. The finite subtree of T generated by $\operatorname {\mathrm {split}}_n(T)$
 is denoted $\operatorname {\mathrm {split}}_n(T)$
 is denoted $\operatorname {\mathrm {split}}_{\leq n}(T)$
. $\operatorname {\mathrm {split}}_{\leq n}(T)$
.
- 
(d) For any  $t \in T$
 we define the restriction of T to t as $t \in T$
 we define the restriction of T to t as $T_t = \{ s \in T : s \not \perp t \}$
. $T_t = \{ s \in T : s \not \perp t \}$
.
- 
(e) The set of branches through T is denoted by  $[T] = \{ x \in A^\omega : \forall n \in \omega (x \restriction n \in T) \}$
. $[T] = \{ x \in A^\omega : \forall n \in \omega (x \restriction n \in T) \}$
.
- 
(f)  $A^{\omega }$
 carries a natural Polish topology generated by the clopen sets $A^{\omega }$
 carries a natural Polish topology generated by the clopen sets $[t] = \{ x \in A^\omega : t \subseteq x \}$
 for $[t] = \{ x \in A^\omega : t \subseteq x \}$
 for $t \in A^{<\omega }$
. Then $t \in A^{<\omega }$
. Then $[T]$
 is closed in $[T]$
 is closed in $A^\omega $
. $A^\omega $
.
- 
(g) Whenever  $X \subseteq A^{\omega }$
 is closed, there is a continuous retraction $X \subseteq A^{\omega }$
 is closed, there is a continuous retraction $\varphi \colon A^\omega \to X$
, i.e., $\varphi \colon A^\omega \to X$
, i.e., $\varphi "A^\omega = X$
 and $\varphi "A^\omega = X$
 and $\varphi \restriction X$
 is the identity. $\varphi \restriction X$
 is the identity.
- 
(h) A tree forcing is a collection  $\mathbb {P}$
 of perfect trees ordered by inclusion. $\mathbb {P}$
 of perfect trees ordered by inclusion.
- 
(i) By convention, all tree forcings are closed under restrictions, i.e., if  $T \in \mathbb {P}$
 and $T \in \mathbb {P}$
 and $t \in T$
, then $t \in T$
, then $T_t \in \mathbb {P}$
, and the trivial condition is $T_t \in \mathbb {P}$
, and the trivial condition is $A^{<\omega }$
. $A^{<\omega }$
.
- 
(j) The set  $\mathcal {T}$
 of perfect subtrees of $\mathcal {T}$
 of perfect subtrees of $A^{<\omega }$
 is a $A^{<\omega }$
 is a $G_\delta $
 subset of $G_\delta $
 subset of $\mathcal {P}(A^{<\omega }) \cong \mathcal {P}(\omega )$
, where we identify $\mathcal {P}(A^{<\omega }) \cong \mathcal {P}(\omega )$
, where we identify $A^{<\omega }$
 with $A^{<\omega }$
 with $\omega $
, and thus carries a natural Polish topology. It is not hard to see that it is homeomorphic to $\omega $
, and thus carries a natural Polish topology. It is not hard to see that it is homeomorphic to $\omega ^\omega $
, when $\omega ^\omega $
, when $\vert A \vert \geq 2$
. $\vert A \vert \geq 2$
.
- 
(k) Often times, we will use a bar above a variable, as in “  $\bar x$
,” to indicate that it denotes a sequence. In that case, we either write $\bar x$
,” to indicate that it denotes a sequence. In that case, we either write $x(\alpha )$
 or $x(\alpha )$
 or $x_\alpha $
 to denote the $x_\alpha $
 to denote the $\alpha $
-th element of that sequence, depending on the context. $\alpha $
-th element of that sequence, depending on the context.
- 
(l) Let  $\langle T_i : i < \alpha \rangle $
 be a sequence of trees where $\langle T_i : i < \alpha \rangle $
 be a sequence of trees where $\alpha $
 is an arbitrary ordinal. Then we write $\alpha $
 is an arbitrary ordinal. Then we write $\bigotimes _{i < \alpha } T_i$
 for the set of finite partial sequences $\bigotimes _{i < \alpha } T_i$
 for the set of finite partial sequences $\bar s$
 where $\bar s$
 where $\operatorname {\mathrm {dom}} \bar s \in [\alpha ]^{<\omega }$
 and for every $\operatorname {\mathrm {dom}} \bar s \in [\alpha ]^{<\omega }$
 and for every $i \in \operatorname {\mathrm {dom}} \bar s$
, $i \in \operatorname {\mathrm {dom}} \bar s$
, $s(i) \in T_i$
. $s(i) \in T_i$
.
- 
(m)  $(A^\omega )^\alpha $
 carries a topology generated by the sets $(A^\omega )^\alpha $
 carries a topology generated by the sets $[\bar s] = \{ \bar x \in (A^\omega )^\alpha : \forall i \in \operatorname {\mathrm {dom}} \bar s (x(i) \in [s(i)]) \}$
 for $[\bar s] = \{ \bar x \in (A^\omega )^\alpha : \forall i \in \operatorname {\mathrm {dom}} \bar s (x(i) \in [s(i)]) \}$
 for $\bar s \in \bigotimes _{i < \alpha } A^{<\omega }$
. $\bar s \in \bigotimes _{i < \alpha } A^{<\omega }$
.
- 
(n) Whenever  $X \subseteq (A^\omega )^\alpha $
 and $X \subseteq (A^\omega )^\alpha $
 and $C \subseteq \alpha $
, we define the projection of X to C as $C \subseteq \alpha $
, we define the projection of X to C as $X \restriction C = \{ \bar x \restriction C : \bar x \in X \}$
. $X \restriction C = \{ \bar x \restriction C : \bar x \in X \}$
.
Fact. Let 
 $\mathbb {P}$
 be a tree forcing and G a
$\mathbb {P}$
 be a tree forcing and G a 
 $\mathbb {P}$
-generic filter over V. Then
$\mathbb {P}$
-generic filter over V. Then 
 $\mathbb {P}$
 adds a real
$\mathbb {P}$
 adds a real 
 $ x_G := \bigcup \{ s \in A^{<\omega } : \forall T \in G (s \in T) \} \in A^\omega $
. Moreover,
$ x_G := \bigcup \{ s \in A^{<\omega } : \forall T \in G (s \in T) \} \in A^\omega $
. Moreover, 
 $V[G] = V[x_G]$
.
$V[G] = V[x_G]$
.
Definition 2.1. We say that 
 $(\mathbb {P}, \leq )$
 is Axiom A if there is a decreasing sequence of partial orders
$(\mathbb {P}, \leq )$
 is Axiom A if there is a decreasing sequence of partial orders 
 $\langle \leq _n : n \in \omega \rangle $
 refining
$\langle \leq _n : n \in \omega \rangle $
 refining 
 $\leq $
 on
$\leq $
 on 
 $\mathbb {P}$
 so that
$\mathbb {P}$
 so that
- 
(1) for any  $n \in \omega $
 and $n \in \omega $
 and $T, S \in \mathbb {P}$
, if $T, S \in \mathbb {P}$
, if $S \leq _n T$
, then $S \leq _n T$
, then $S \cap A^{<n} = T \cap A^{<n}$
, $S \cap A^{<n} = T \cap A^{<n}$
,
- 
(2) for any fusion sequence, i.e., a sequence  $\langle p_n : n \in \omega \rangle $
 where $\langle p_n : n \in \omega \rangle $
 where $p_{n+1} \leq _n p_n$
 for every n, $p_{n+1} \leq _n p_n$
 for every n, $p = \bigcap _{n \in \omega } p_n \in \mathbb {P}$
 and $p = \bigcap _{n \in \omega } p_n \in \mathbb {P}$
 and $p \leq _n p_n$
 for every n, and $p \leq _n p_n$
 for every n, and
- 
(3) for any maximal antichain  $D \subseteq \mathbb {P}$
, $D \subseteq \mathbb {P}$
, $p \in \mathbb {P}$
, $p \in \mathbb {P}$
, $n \in \omega $
, there is $n \in \omega $
, there is $q \leq _n p$
 so that $q \leq _n p$
 so that $\{ r \in D : r \not \perp q \}$
 is countable. $\{ r \in D : r \not \perp q \}$
 is countable.
Moreover we say that 
 $(\mathbb {P}, \leq )$
 is Axiom A with continuous reading of names (crn) if there is such a sequence of partial orders so that additionally,
$(\mathbb {P}, \leq )$
 is Axiom A with continuous reading of names (crn) if there is such a sequence of partial orders so that additionally, 
- 
(4) for every  $p \in \mathbb {P}$
, $p \in \mathbb {P}$
, $n \in \omega $
, and $n \in \omega $
, and $\dot y$
 a $\dot y$
 a $\mathbb {P}$
-name for an element of a Polish spaceFootnote 
2
 X, there is $\mathbb {P}$
-name for an element of a Polish spaceFootnote 
2
 X, there is $q \leq _n p$
 and a continuous function $q \leq _n p$
 and a continuous function $f \colon [q] \to X$
 so that $f \colon [q] \to X$
 so that $$ \begin{align*}q \Vdash \dot y [G] = f(x_G).\end{align*} $$ $$ \begin{align*}q \Vdash \dot y [G] = f(x_G).\end{align*} $$
Although (1) is typically not part of the definition of Axiom A, we include it for technical reasons. The only classical example that we are aware of, in which it is not clear whether (1)–(4) can be realized simultaneously, is Mathias forcing.
 Let 
 $\langle \mathbb {P}_\beta , \dot {\mathbb {Q}}_\beta : \beta < \alpha \rangle $
 be a countable support iteration of tree forcings that are Axiom A with crn, where for each
$\langle \mathbb {P}_\beta , \dot {\mathbb {Q}}_\beta : \beta < \alpha \rangle $
 be a countable support iteration of tree forcings that are Axiom A with crn, where for each 
 $\beta < \alpha $
,
$\beta < \alpha $
, 
 $$ \begin{align*}\Vdash_{\mathbb{P}_\beta} \text{"}\langle \dot{\leq}_{\beta,n} : n \in \omega \rangle \text{ witnesses that } \dot{\mathbb{Q}}_{\beta} \text{ is Axiom A with crn."}\end{align*} $$
$$ \begin{align*}\Vdash_{\mathbb{P}_\beta} \text{"}\langle \dot{\leq}_{\beta,n} : n \in \omega \rangle \text{ witnesses that } \dot{\mathbb{Q}}_{\beta} \text{ is Axiom A with crn."}\end{align*} $$
- 
(n) For each  $n \in \omega , a \subseteq \alpha $
, we define $n \in \omega , a \subseteq \alpha $
, we define $\leq _{n,a}$
 on $\leq _{n,a}$
 on $\mathbb {P}_\alpha $
, where $\mathbb {P}_\alpha $
, where $$ \begin{align*}\bar q \leq_{n,a} \bar p \leftrightarrow \big(\bar q \leq \bar p \wedge \forall \beta \in a ( \bar q \restriction \beta \Vdash_{\mathbb{P}_\beta} \dot q(\beta) \dot{\leq}_{\beta,n} \dot p(\beta))\big).\end{align*} $$ $$ \begin{align*}\bar q \leq_{n,a} \bar p \leftrightarrow \big(\bar q \leq \bar p \wedge \forall \beta \in a ( \bar q \restriction \beta \Vdash_{\mathbb{P}_\beta} \dot q(\beta) \dot{\leq}_{\beta,n} \dot p(\beta))\big).\end{align*} $$
- 
(o) The support of  $\bar p \in \mathbb {P}_\alpha $
 is the set $\bar p \in \mathbb {P}_\alpha $
 is the set . .
 Recall that a condition q is called a master condition over a model M if for any maximal antichain 
 $D \in M$
,
$D \in M$
, 
 $\{ p \in D : q \not \perp p \} \subseteq M$
. Equivalently, it means that for every generic filter G over V containing q, G is generic over M as well. Throughout this paper, when we say that M is elementary, we mean that it is elementary in a large enough model of the form
$\{ p \in D : q \not \perp p \} \subseteq M$
. Equivalently, it means that for every generic filter G over V containing q, G is generic over M as well. Throughout this paper, when we say that M is elementary, we mean that it is elementary in a large enough model of the form 
 $H(\theta )$
. Sometimes, we will say that M is a model of set theory or just that M is a model. In most generality, this just means that
$H(\theta )$
. Sometimes, we will say that M is a model of set theory or just that M is a model. In most generality, this just means that 
 $(M,\in )$
 satisfies a strong enough fragment of ZFC. But this is a way too general notion for our purposes. For instance, such M may not even be correct about what
$(M,\in )$
 satisfies a strong enough fragment of ZFC. But this is a way too general notion for our purposes. For instance, such M may not even be correct about what 
 $\omega $
 is. Thus, let us clarify that in all our instances this will mean that M is either elementary or an extension of an elementary model by a countable (in M) forcing. In particular, some basic absoluteness (e.g., for
$\omega $
 is. Thus, let us clarify that in all our instances this will mean that M is either elementary or an extension of an elementary model by a countable (in M) forcing. In particular, some basic absoluteness (e.g., for 
 $\Sigma ^1_1$
 or
$\Sigma ^1_1$
 or 
 $\Pi ^1_1$
 formulas) holds true between M and V, M is transitive below
$\Pi ^1_1$
 formulas) holds true between M and V, M is transitive below 
 $\omega _1$
, and
$\omega _1$
, and 
 $\omega _1$
 is computed correctly.
$\omega _1$
 is computed correctly.
Fact (Fusion Lemma, see, e.g., [Reference Baumgartner and Laver1, Lemmas 1.2 and 2.3]).
 If 
 $\langle a_n : n \in \omega \rangle $
 is
$\langle a_n : n \in \omega \rangle $
 is 
 $\subseteq $
-increasing,
$\subseteq $
-increasing, 
 $\langle \bar p_n : n \in \omega \rangle $
 is such that
$\langle \bar p_n : n \in \omega \rangle $
 is such that 
 $\forall n \in \omega (\bar p_{n+1} \leq _{n,a_n} \bar p_n)$
, and
$\forall n \in \omega (\bar p_{n+1} \leq _{n,a_n} \bar p_n)$
, and 
 $\bigcup _{n \in \omega } \operatorname {\mathrm {supp}}(\bar p_n) \subseteq \bigcup _{n \in \omega } a_n \subseteq \alpha $
, then there is a condition
$\bigcup _{n \in \omega } \operatorname {\mathrm {supp}}(\bar p_n) \subseteq \bigcup _{n \in \omega } a_n \subseteq \alpha $
, then there is a condition 
 $\bar p \in \mathbb {P}_\alpha $
 so that for every
$\bar p \in \mathbb {P}_\alpha $
 so that for every 
 $n \in \omega $
,
$n \in \omega $
, 
 $\bar p \leq _{n,a_n} \bar p_n$
; in fact, for every
$\bar p \leq _{n,a_n} \bar p_n$
; in fact, for every 
 $\beta < \alpha $
,
$\beta < \alpha $
, 
 $\bar p \restriction \beta \Vdash \dot p(\beta ) = \bigcap _{n \in \omega } \dot p_n(\beta )$
.
$\bar p \restriction \beta \Vdash \dot p(\beta ) = \bigcap _{n \in \omega } \dot p_n(\beta )$
.
 Moreover, let M be a countable elementary model, 
 $\bar p \in M \cap \mathbb {P}_\alpha $
,
$\bar p \in M \cap \mathbb {P}_\alpha $
, 
 $n \in \omega $
,
$n \in \omega $
, 
 $a \subseteq M\cap \alpha $
 finite, and
$a \subseteq M\cap \alpha $
 finite, and 
 $\langle \alpha _i : i \in \omega \rangle $
 a cofinal increasing sequence in
$\langle \alpha _i : i \in \omega \rangle $
 a cofinal increasing sequence in 
 $M\cap \alpha $
. Then there is
$M\cap \alpha $
. Then there is 
 $\bar q \leq _{n,a} \bar p$
 a master condition over M so that for every name
$\bar q \leq _{n,a} \bar p$
 a master condition over M so that for every name 
 $\dot y \in M$
 for an element of
$\dot y \in M$
 for an element of 
 $\omega ^\omega $
 and
$\omega ^\omega $
 and 
 $j \in \omega $
, there is
$j \in \omega $
, there is 
 $i \in \omega $
 so that below
$i \in \omega $
 so that below 
 $\bar q$
, the value of
$\bar q$
, the value of 
 $\dot y \restriction j$
 only depends on the
$\dot y \restriction j$
 only depends on the 
 $\mathbb {P}_{\alpha _i}$
-generic.
$\mathbb {P}_{\alpha _i}$
-generic.
- 
(p) For G a  $\mathbb {P}_\alpha $
-generic, we write $\mathbb {P}_\alpha $
-generic, we write $\bar x_G$
 for the generic element of $\bar x_G$
 for the generic element of $\prod _{\beta <\alpha }A^\omega $
 added by $\prod _{\beta <\alpha }A^\omega $
 added by $\mathbb {P}_\alpha $
. $\mathbb {P}_\alpha $
.
 Let us from now on assume that for each 
 $\beta < \alpha $
 and
$\beta < \alpha $
 and 
 $n \in \omega $
,
$n \in \omega $
, 
 $\mathbb {Q}_\beta $
 and
$\mathbb {Q}_\beta $
 and 
 $\leq _{\beta ,n}$
 are fixed analytic subsets of
$\leq _{\beta ,n}$
 are fixed analytic subsets of 
 $\mathcal {T}$
 and
$\mathcal {T}$
 and 
 $\mathcal {T}^2$
 respectively, coded in V. Although the theory that we develop below can be extended to a large extent to non-definable iterands, we will only focus on this case, since we need stronger results later on.
$\mathcal {T}^2$
 respectively, coded in V. Although the theory that we develop below can be extended to a large extent to non-definable iterands, we will only focus on this case, since we need stronger results later on.
Lemma 2.2. For any 
 $\bar p \in \mathbb {P}_\alpha $
, M a countable elementary model so that
$\bar p \in \mathbb {P}_\alpha $
, M a countable elementary model so that 
 ${\mathbb {P}_\alpha , \bar p \in M}$
, and
${\mathbb {P}_\alpha , \bar p \in M}$
, and 
 $n \in \omega , a \subseteq M\cap \alpha $
 finite, there is
$n \in \omega , a \subseteq M\cap \alpha $
 finite, there is 
 $\bar q \leq _{n,a} \bar p$
 a master condition over M and a closed set
$\bar q \leq _{n,a} \bar p$
 a master condition over M and a closed set 
 $[\bar q] \subseteq (A^\omega )^\alpha $
 so that
$[\bar q] \subseteq (A^\omega )^\alpha $
 so that 
- 
(1)  $\bar q \Vdash \bar x_G \in [\bar q]$
, $\bar q \Vdash \bar x_G \in [\bar q]$
,
 for every 
 $\beta < \alpha $
,
$\beta < \alpha $
,
- 
(2)  $\bar q \Vdash \dot q(\beta ) = \{s \in A^{<\omega } : \exists \bar z \in [\bar q] (\bar z \restriction \beta = \bar x_{G} \restriction \beta \wedge s \subseteq z(\beta )) \} $
, $\bar q \Vdash \dot q(\beta ) = \{s \in A^{<\omega } : \exists \bar z \in [\bar q] (\bar z \restriction \beta = \bar x_{G} \restriction \beta \wedge s \subseteq z(\beta )) \} $
,
- 
(3) the map sending  $\bar x \in [\bar q] \restriction \beta $
 to $\bar x \in [\bar q] \restriction \beta $
 to $\{s \in A^{<\omega } : \exists \bar z \in [\bar q] (\bar z \restriction \beta = \bar x \wedge s \subseteq z(\beta )) \}$
 is continuous and maps to $\{s \in A^{<\omega } : \exists \bar z \in [\bar q] (\bar z \restriction \beta = \bar x \wedge s \subseteq z(\beta )) \}$
 is continuous and maps to $\mathbb {Q}_\beta $
, $\mathbb {Q}_\beta $
,
- 
(4)  $[\bar q] \restriction \beta \subseteq (A^\omega )^\beta $
 is closed, $[\bar q] \restriction \beta \subseteq (A^\omega )^\beta $
 is closed,
 and for every name 
 $\dot y \in M$
 for an element of a Polish space X,
$\dot y \in M$
 for an element of a Polish space X,
- 
(5) there is a continuous function  $f \colon [\bar q] \to X$
 so that $f \colon [\bar q] \to X$
 so that $\bar q \Vdash \dot y = f(\bar x_G).$ $\bar q \Vdash \dot y = f(\bar x_G).$
- 
(q) We call such  $\bar q$
 as in Lemma 2.2 a good master condition over M. $\bar q$
 as in Lemma 2.2 a good master condition over M.
Before we prove Lemma 2.2, let us draw some consequences from the definition of a good master condition.
Lemma 2.3. Let 
 $\bar q \in \mathbb {P}_\alpha $
 be a good master condition over a model M and
$\bar q \in \mathbb {P}_\alpha $
 be a good master condition over a model M and 
 $\dot y \in M$
 a name for an element of a Polish space X.
$\dot y \in M$
 a name for an element of a Polish space X. 
- 
(i) Then  $[\bar q]$
 is unique, in fact it is the closure of $[\bar q]$
 is unique, in fact it is the closure of $\{ \bar x_G : G \ni \bar q \text { is generic over } V\}$
 in any forcing extension W of V where $\{ \bar x_G : G \ni \bar q \text { is generic over } V\}$
 in any forcing extension W of V where $\left (\beth _{\omega }(\vert \mathbb {P}_\alpha \vert )\right )^V$
 is countable. $\left (\beth _{\omega }(\vert \mathbb {P}_\alpha \vert )\right )^V$
 is countable.
- 
(ii) The continuous map  $f \colon [\bar q] \to X$
 given by (5) is unique, and $f \colon [\bar q] \to X$
 given by (5) is unique, and
- 
(iii) whenever  $Y \in M$
 is an analytic subset of X and $Y \in M$
 is an analytic subset of X and $\bar q \Vdash \dot y \in Y$
, then $\bar q \Vdash \dot y \in Y$
, then $f"[\bar q] \subseteq Y$
. $f"[\bar q] \subseteq Y$
.
Moreover, there is a countable set 
 $C \subseteq \alpha $
, not depending on
$C \subseteq \alpha $
, not depending on 
 $\dot y$
, so that
$\dot y$
, so that 
- 
(iv)  $[\bar q] \restriction C$
 is a closed subset of the Polish space $[\bar q] \restriction C$
 is a closed subset of the Polish space $(A^{\omega })^C$
 and $(A^{\omega })^C$
 and $[\bar q] = ([\bar q] \restriction C) \times (A^\omega )^{\alpha \setminus C}$
, $[\bar q] = ([\bar q] \restriction C) \times (A^\omega )^{\alpha \setminus C}$
,
- 
(v) for every  $\beta \in C$
, there is a continuous function $\beta \in C$
, there is a continuous function $g \colon [\bar q] \restriction (C \cap \beta )\to \mathbb {Q}_\beta $
, so that for every $g \colon [\bar q] \restriction (C \cap \beta )\to \mathbb {Q}_\beta $
, so that for every $\bar x \in [\bar q]$
, $\bar x \in [\bar q]$
, $$ \begin{align*}g(\bar x \restriction (C \cap \beta)) = \{s \in A^{<\omega} : \exists \bar z \in [\bar q] (\bar z \restriction \beta = \bar x \restriction \beta \wedge s \subseteq z(\beta)) \} ,\end{align*} $$ $$ \begin{align*}g(\bar x \restriction (C \cap \beta)) = \{s \in A^{<\omega} : \exists \bar z \in [\bar q] (\bar z \restriction \beta = \bar x \restriction \beta \wedge s \subseteq z(\beta)) \} ,\end{align*} $$
- 
(vi) there is a continuous function  $f \colon [\bar q] \restriction C \to X$
, so that $f \colon [\bar q] \restriction C \to X$
, so that $$ \begin{align*}\bar q \Vdash \dot y = f(\bar x_G \restriction C).\end{align*} $$ $$ \begin{align*}\bar q \Vdash \dot y = f(\bar x_G \restriction C).\end{align*} $$
Proof. Let us write, for every 
 $\beta < \alpha $
 and
$\beta < \alpha $
 and 
 $\bar x \in [\bar q] \restriction \beta $
,
$\bar x \in [\bar q] \restriction \beta $
, 
 $$ \begin{align*}T_{\bar x} := \{s \in A^{<\omega} : \exists \bar z \in [\bar q] (\bar z \restriction \beta = \bar x \wedge s \subseteq z(\beta)) \}.\end{align*} $$
$$ \begin{align*}T_{\bar x} := \{s \in A^{<\omega} : \exists \bar z \in [\bar q] (\bar z \restriction \beta = \bar x \wedge s \subseteq z(\beta)) \}.\end{align*} $$
 For (i), let W be an extension in which 
 $\left (\beth _{\omega }(\vert \mathbb {P}_\alpha \vert )\right )^V$
 is countable and let
$\left (\beth _{\omega }(\vert \mathbb {P}_\alpha \vert )\right )^V$
 is countable and let 
 $\bar s \in \bigotimes _{i < \alpha } A^{<\omega }$
 be arbitrary so that
$\bar s \in \bigotimes _{i < \alpha } A^{<\omega }$
 be arbitrary so that 
 $[\bar s] \cap [\bar q]$
 is non-empty. We claim that there is a generic G over V containing
$[\bar s] \cap [\bar q]$
 is non-empty. We claim that there is a generic G over V containing 
 $\bar q$
 so that
$\bar q$
 so that 
 $\bar x_G \in [\bar s]$
. This is shown by induction on
$\bar x_G \in [\bar s]$
. This is shown by induction on 
 $\max (\operatorname {\mathrm {dom}} (\bar s))$
. For
$\max (\operatorname {\mathrm {dom}} (\bar s))$
. For 
 $\bar s = \emptyset $
 the claim is obvious. Now assume
$\bar s = \emptyset $
 the claim is obvious. Now assume 
 $\max (\operatorname {\mathrm {dom}} (\bar s)) = \beta $
, for
$\max (\operatorname {\mathrm {dom}} (\bar s)) = \beta $
, for 
 $\beta < \alpha $
. Then, by (3),
$\beta < \alpha $
. Then, by (3), 
 $O := \{ \bar x \in [\bar q] : s(\beta ) \in T_{\bar x \restriction \beta }\}$
 is open and it is non-empty since
$O := \{ \bar x \in [\bar q] : s(\beta ) \in T_{\bar x \restriction \beta }\}$
 is open and it is non-empty since 
 $[\bar s] \cap [\bar q] \neq \emptyset $
. Applying the inductive hypothesis, there is a generic
$[\bar s] \cap [\bar q] \neq \emptyset $
. Applying the inductive hypothesis, there is a generic 
 $G \ni \bar q$
 so that
$G \ni \bar q$
 so that 
 $\bar x_G \in O$
. In
$\bar x_G \in O$
. In 
 $V[G \restriction \beta ]$
 we have, by (2), that
$V[G \restriction \beta ]$
 we have, by (2), that 
 $T_{\bar x_G \restriction \beta } = \dot q(\beta )[G]$
. Moreover, since
$T_{\bar x_G \restriction \beta } = \dot q(\beta )[G]$
. Moreover, since 
 $\bar x_G \in O$
, we have that
$\bar x_G \in O$
, we have that 
 $s(\beta ) \in \dot q(\beta )[G]$
. Then it is easy to force over
$s(\beta ) \in \dot q(\beta )[G]$
. Then it is easy to force over 
 $V[G \restriction \beta ]$
, to get a full
$V[G \restriction \beta ]$
, to get a full 
 $\mathbb {P}_\alpha $
 generic
$\mathbb {P}_\alpha $
 generic 
 $H \supseteq G \restriction \beta $
 containing
$H \supseteq G \restriction \beta $
 containing 
 $\bar q$
 so that
$\bar q$
 so that 
 $\bar x_{H} \restriction \beta = \bar x_G \restriction \beta $
 and
$\bar x_{H} \restriction \beta = \bar x_G \restriction \beta $
 and 
 $s(\beta ) \subseteq \bar x_H(\beta )$
. By (1), for every generic G over V containing
$s(\beta ) \subseteq \bar x_H(\beta )$
. By (1), for every generic G over V containing 
 $\bar q$
,
$\bar q$
, 
 $\bar x_G \in [\bar q]$
. Thus we have shown that the set of such
$\bar x_G \in [\bar q]$
. Thus we have shown that the set of such 
 $\bar x_G$
 is dense in
$\bar x_G$
 is dense in 
 $[\bar q]$
. Uniqueness follows from
$[\bar q]$
. Uniqueness follows from 
 $[\bar q]$
 being closed and the fact that if two closed sets coded in V agree in W, then they agree in V. This follows easily from
$[\bar q]$
 being closed and the fact that if two closed sets coded in V agree in W, then they agree in V. This follows easily from 
 $\mathbf \Pi ^1_1$
 absoluteness.
$\mathbf \Pi ^1_1$
 absoluteness.
 Now (ii) follows easily since any two continuous functions given by (5) have to agree on a dense set in an extension W and thus they agree in V. Again this is an easy consequence of 
 $\mathbf \Pi ^1_1$
 absoluteness.
$\mathbf \Pi ^1_1$
 absoluteness.
 For (iii), let us consider the analytic space 
 $Z = \{0\} \times X \cup \{1\} \times Y$
, which is the disjoint union of the spaces X and Y. Then there is a continuous surjection
$Z = \{0\} \times X \cup \{1\} \times Y$
, which is the disjoint union of the spaces X and Y. Then there is a continuous surjection 
 $F \colon \omega ^\omega \to Z$
 and by elementarity we can assume it is in M. Let us find in M a name
$F \colon \omega ^\omega \to Z$
 and by elementarity we can assume it is in M. Let us find in M a name 
 $\dot z$
 for an element of
$\dot z$
 for an element of 
 $\omega ^\omega $
 so that in
$\omega ^\omega $
 so that in 
 $V[G]$
, if
$V[G]$
, if 
 $\dot y[G] \in Y$
, then
$\dot y[G] \in Y$
, then 
 $F(\dot z[G]) = (1,\dot y[G])$
, and if
$F(\dot z[G]) = (1,\dot y[G])$
, and if 
 $\dot y[G] \notin Y$
, then
$\dot y[G] \notin Y$
, then 
 $F(\dot z[G]) = (0,\dot y[G])$
. By (5), there is a continuous function
$F(\dot z[G]) = (0,\dot y[G])$
. By (5), there is a continuous function 
 $g \colon [\bar q] \to \omega ^\omega $
 so that
$g \colon [\bar q] \to \omega ^\omega $
 so that 
 $\bar q \Vdash \dot z = g(\bar x_G)$
. Since
$\bar q \Vdash \dot z = g(\bar x_G)$
. Since 
 $\bar q \Vdash \dot y \in Y$
, we have that for any generic G containing
$\bar q \Vdash \dot y \in Y$
, we have that for any generic G containing 
 $\bar q$
,
$\bar q$
, 
 $F(g(\bar x_G)) = (1,f(\bar x_G))$
. By density, for every
$F(g(\bar x_G)) = (1,f(\bar x_G))$
. By density, for every 
 $\bar x \in [\bar q]$
,
$\bar x \in [\bar q]$
, 
 $F(g(\bar x)) = (1,f(\bar x))$
 and in particular
$F(g(\bar x)) = (1,f(\bar x))$
 and in particular 
 $f(\bar x) \in Y$
.
$f(\bar x) \in Y$
.
 Now let us say that the support of a function 
 $g \colon [\bar q] \to X$
 is the smallest set
$g \colon [\bar q] \to X$
 is the smallest set 
 $C_g \subseteq \alpha $
 so that the value of
$C_g \subseteq \alpha $
 so that the value of 
 $g(\bar x)$
 only depends on
$g(\bar x)$
 only depends on 
 $\bar x \restriction C_g$
. The results of [Reference Bockstein3] imply that if g is continuous, then g has countable support. Note that for all
$\bar x \restriction C_g$
. The results of [Reference Bockstein3] imply that if g is continuous, then g has countable support. Note that for all 
 $\beta \notin \operatorname {\mathrm {supp}}(\bar q)$
, the map in (3) is constant on the set of generics and by continuity it is constant everywhere. Thus it has empty support. Let C be the union of
$\beta \notin \operatorname {\mathrm {supp}}(\bar q)$
, the map in (3) is constant on the set of generics and by continuity it is constant everywhere. Thus it has empty support. Let C be the union of 
 $\operatorname {\mathrm {supp}}(\bar q)$
 with all the countable supports given by instances of (3) and (5). Then C is a countable set. For (iv)–(vi), note that
$\operatorname {\mathrm {supp}}(\bar q)$
 with all the countable supports given by instances of (3) and (5). Then C is a countable set. For (iv)–(vi), note that 
 $[\bar q] \restriction C = \{ \bar y \in (A^\omega )^C : \bar y^{\frown } (\bar x \restriction \alpha \setminus C) \in [\bar q] \}$
 for
$[\bar q] \restriction C = \{ \bar y \in (A^\omega )^C : \bar y^{\frown } (\bar x \restriction \alpha \setminus C) \in [\bar q] \}$
 for 
 $\bar x \in [\bar q]$
 arbitrary, and recall that in a product, sections of closed sets are closed and continuous functions are coordinate-wise continuous.⊣
$\bar x \in [\bar q]$
 arbitrary, and recall that in a product, sections of closed sets are closed and continuous functions are coordinate-wise continuous.⊣
Proof of Lemma 2.2.
 Let us fix for each 
 $\beta < \alpha $
 a continuous surjection
$\beta < \alpha $
 a continuous surjection 
 $F_\beta \colon \omega ^\omega \to \mathbb {Q}_\beta $
. The proof is by induction on
$F_\beta \colon \omega ^\omega \to \mathbb {Q}_\beta $
. The proof is by induction on 
 $\alpha $
. If
$\alpha $
. If 
 $\alpha = \beta +1$
, then
$\alpha = \beta +1$
, then 
 $\mathbb {P}_\alpha = \mathbb {P}_\beta * \dot {\mathbb {Q}}_{\beta }$
. Let
$\mathbb {P}_\alpha = \mathbb {P}_\beta * \dot {\mathbb {Q}}_{\beta }$
. Let 
 $\bar q_0 \leq _{n,a} \bar p \restriction \beta $
 be a master condition over M and
$\bar q_0 \leq _{n,a} \bar p \restriction \beta $
 be a master condition over M and 
 $H \ni \bar q_0$
 a
$H \ni \bar q_0$
 a 
 $\mathbb {P}_\beta $
 generic over V. Then, applying a standard fusion argument using Axiom A with continuous reading of names in
$\mathbb {P}_\beta $
 generic over V. Then, applying a standard fusion argument using Axiom A with continuous reading of names in 
 $V[H]$
 to
$V[H]$
 to 
 $\mathbb {Q}_\beta $
, we find
$\mathbb {Q}_\beta $
, we find 
 $q(\beta ) \leq _{\beta ,n} p(\beta )$
 a master condition over
$q(\beta ) \leq _{\beta ,n} p(\beta )$
 a master condition over 
 $M[H]$
 (note that H is also M generic since
$M[H]$
 (note that H is also M generic since 
 $\bar q_0$
 is a master condition over M) so that for each name
$\bar q_0$
 is a master condition over M) so that for each name 
 $\dot y \in M[H]$
 for an element of a Polish space X there is a continuous function
$\dot y \in M[H]$
 for an element of a Polish space X there is a continuous function 
 $f \colon [q(\beta )] \to X$
 so that
$f \colon [q(\beta )] \to X$
 so that 
 $q(\beta ) \Vdash \dot y = f(\dot {x}_G)$
. Thus we find in V, a
$q(\beta ) \Vdash \dot y = f(\dot {x}_G)$
. Thus we find in V, a 
 $\mathbb {P}_\beta $
-name
$\mathbb {P}_\beta $
-name 
 $\dot q(\beta )$
 so that
$\dot q(\beta )$
 so that 
 $\bar q_0$
 forces that it is such a condition. Let
$\bar q_0$
 forces that it is such a condition. Let 
 $M^+ \ni M$
 be a countable elementary model containing
$M^+ \ni M$
 be a countable elementary model containing 
 $\dot q(\beta )$
 and
$\dot q(\beta )$
 and 
 $\bar q_0$
, and let
$\bar q_0$
, and let 
 $\bar q_{1/2} \leq _{n,a} \bar q_{0}$
 be a master condition over
$\bar q_{1/2} \leq _{n,a} \bar q_{0}$
 be a master condition over 
 $M^+$
. Again let
$M^+$
. Again let 
 $M^{++}\ni M^+$
 be a countable elementary model containing
$M^{++}\ni M^+$
 be a countable elementary model containing 
 $\bar q_{1/2}$
. By the induction hypothesis we find
$\bar q_{1/2}$
. By the induction hypothesis we find 
 $\bar q_1 \leq _{n,a} \bar q_{1/2}$
 a good master condition over
$\bar q_1 \leq _{n,a} \bar q_{1/2}$
 a good master condition over 
 $M^{++}$
. Finally, let
$M^{++}$
. Finally, let 
 $\bar q = \bar q_1 ^{\frown } \dot {q}(\beta )$
. Then
$\bar q = \bar q_1 ^{\frown } \dot {q}(\beta )$
. Then 
 $\bar q \leq _{n,a} \bar p$
 and
$\bar q \leq _{n,a} \bar p$
 and 
 $\bar q$
 is a master condition over M. Since
$\bar q$
 is a master condition over M. Since 
 $\dot q(\beta ) \in M^+\subseteq M^{++}$
, there is a continuous function
$\dot q(\beta ) \in M^+\subseteq M^{++}$
, there is a continuous function 
 $f \colon [\bar q_1] \to \omega ^\omega $
, so that
$f \colon [\bar q_1] \to \omega ^\omega $
, so that 
 $\bar q_1 \Vdash _\beta F_\beta (f(\bar x_{H})) = \dot q(\beta )$
. Here note that
$\bar q_1 \Vdash _\beta F_\beta (f(\bar x_{H})) = \dot q(\beta )$
. Here note that 
 $F_\beta $
 is in M by elementarity and we indeed find a name
$F_\beta $
 is in M by elementarity and we indeed find a name 
 $\dot z$
 in
$\dot z$
 in 
 $M^+$
 so that
$M^+$
 so that 
 $\bar q_0 \Vdash F_\beta (\dot z) = \dot q(\beta )$
. Let
$\bar q_0 \Vdash F_\beta (\dot z) = \dot q(\beta )$
. Let 
 $[ \bar q] = \{ \bar x \in (A^{\omega })^\alpha : \bar x \restriction \beta \in [\bar q_1] \wedge x(\beta ) \in [F_\beta (f(\bar x \restriction \beta ))] \}$
. Then
$[ \bar q] = \{ \bar x \in (A^{\omega })^\alpha : \bar x \restriction \beta \in [\bar q_1] \wedge x(\beta ) \in [F_\beta (f(\bar x \restriction \beta ))] \}$
. Then 
 $[\bar q]$
 is closed and (1)–(4) hold true. To see that
$[\bar q]$
 is closed and (1)–(4) hold true. To see that 
 $[\bar q]$
 is closed, note that the graph of a continuous function is always closed, when the codomain is a Hausdorff space. For (5), let
$[\bar q]$
 is closed, note that the graph of a continuous function is always closed, when the codomain is a Hausdorff space. For (5), let 
 $\dot y \in M$
 be a
$\dot y \in M$
 be a 
 $\mathbb {P}_\alpha $
-name for an element of a Polish space X. If
$\mathbb {P}_\alpha $
-name for an element of a Polish space X. If 
 $H \ni \bar q_1$
 is V-generic, then there is a continuous function
$H \ni \bar q_1$
 is V-generic, then there is a continuous function 
 $g \colon [q(\beta )] \to X$
 in
$g \colon [q(\beta )] \to X$
 in 
 $V[H]$
 so that
$V[H]$
 so that 
 $V[H] \models q(\beta ) \Vdash g(\dot x_G) = \dot y$
, where we view
$V[H] \models q(\beta ) \Vdash g(\dot x_G) = \dot y$
, where we view 
 $\dot y$
 as a
$\dot y$
 as a 
 $\mathbb {Q}_\beta $
-name in
$\mathbb {Q}_\beta $
-name in 
 $M[H]$
. Moreover there is a continuous retraction
$M[H]$
. Moreover there is a continuous retraction 
 $\varphi \colon A^\omega \to [q(\beta )]$
 in
$\varphi \colon A^\omega \to [q(\beta )]$
 in 
 $V[H]$
. Since
$V[H]$
. Since 
 $M^+$
 was chosen elementary enough, we find names
$M^+$
 was chosen elementary enough, we find names 
 $\dot g$
 and
$\dot g$
 and 
 $\dot \varphi $
 for g and
$\dot \varphi $
 for g and 
 $\varphi $
 in
$\varphi $
 in 
 $M^+$
. The function
$M^+$
. The function 
 $g \circ \varphi $
 is an element of the spaceFootnote 
3
$g \circ \varphi $
 is an element of the spaceFootnote 
3
 
 $C(A^\omega , X)$
, but this is not a Polish space when A is infinite, i.e., when
$C(A^\omega , X)$
, but this is not a Polish space when A is infinite, i.e., when 
 $A^\omega $
 is not compact. It is though, always a coanalytic space (consult, e.g., [Reference Kechris19, Proposition 2.6] to see how
$A^\omega $
 is not compact. It is though, always a coanalytic space (consult, e.g., [Reference Kechris19, Proposition 2.6] to see how 
 $C(A^\omega , X)$
 is a coanalytic subspace of a suitable Polish space). Thus there is an increasing sequence
$C(A^\omega , X)$
 is a coanalytic subspace of a suitable Polish space). Thus there is an increasing sequence 
 $\langle Y_\xi : \xi < \omega _1 \rangle $
 of analytic subspaces such that
$\langle Y_\xi : \xi < \omega _1 \rangle $
 of analytic subspaces such that 
 $\bigcup _{\xi < \omega _1} Y_\xi = C(A^\omega , X)$
 and the same equality holds in any
$\bigcup _{\xi < \omega _1} Y_\xi = C(A^\omega , X)$
 and the same equality holds in any 
 $\omega _1$
-preserving extension. Since
$\omega _1$
-preserving extension. Since 
 $\bar q_{1/2}$
 is a master condition over
$\bar q_{1/2}$
 is a master condition over 
 $M^+$
, we have that
$M^+$
, we have that 
 $\bar q_{1/2} \Vdash \dot g \circ \dot \varphi \in Y_\xi $
, where
$\bar q_{1/2} \Vdash \dot g \circ \dot \varphi \in Y_\xi $
, where 
 $\xi = M^+ \cap \omega _1$
. Since
$\xi = M^+ \cap \omega _1$
. Since 
 $\bar q_1$
 is a good master condition over
$\bar q_1$
 is a good master condition over 
 $M^{++}$
 and
$M^{++}$
 and 
 $Y_\xi \in M^{++}$
, by Lemma 2.3, there is a continuous function
$Y_\xi \in M^{++}$
, by Lemma 2.3, there is a continuous function 
 $g' \in V$
,
$g' \in V$
, 
 $g' \colon [\bar q_1] \to Y_\xi $
, so that
$g' \colon [\bar q_1] \to Y_\xi $
, so that 
 $\bar q_1 \Vdash g'(\bar x_H) = \dot g \circ \dot \varphi $
. Altogether we have that
$\bar q_1 \Vdash g'(\bar x_H) = \dot g \circ \dot \varphi $
. Altogether we have that 
 $\bar q \Vdash \dot y = g'(\bar x_G \restriction \beta )(x_G(\beta ))$
.
$\bar q \Vdash \dot y = g'(\bar x_G \restriction \beta )(x_G(\beta ))$
.
 For 
 $\alpha $
 limit, let
$\alpha $
 limit, let 
 $\langle \alpha _i : i \in \omega \rangle $
 be a strictly increasing sequence cofinal in
$\langle \alpha _i : i \in \omega \rangle $
 be a strictly increasing sequence cofinal in 
 $M \cap \alpha $
 and let
$M \cap \alpha $
 and let 
 $\bar q_0 \leq _{n,a} \bar p$
 be a master condition over M so that for every name
$\bar q_0 \leq _{n,a} \bar p$
 be a master condition over M so that for every name 
 $\dot y \in M$
 for an element of
$\dot y \in M$
 for an element of 
 $\omega ^\omega $
,
$\omega ^\omega $
, 
 $j \in \omega $
, the value of
$j \in \omega $
, the value of 
 $\dot y \restriction j$
 only depends on the generic restricted to
$\dot y \restriction j$
 only depends on the generic restricted to 
 $\mathbb {P}_{\alpha _i}$
 for some
$\mathbb {P}_{\alpha _i}$
 for some 
 $i \in \omega $
. Let us fix a “big” countable elementary model N, with
$i \in \omega $
. Let us fix a “big” countable elementary model N, with 
 $\bar q_0, M \in N$
. Let
$\bar q_0, M \in N$
. Let 
 $\langle a_i : i \in \omega \rangle $
 be an increasing sequence of finite subsets of
$\langle a_i : i \in \omega \rangle $
 be an increasing sequence of finite subsets of 
 $N \cap \alpha $
 so that
$N \cap \alpha $
 so that 
 $a_0 = a$
 and
$a_0 = a$
 and 
 $\bigcup _{i \in \omega } a_i = N \cap \alpha $
. Now inductively define sequences
$\bigcup _{i \in \omega } a_i = N \cap \alpha $
. Now inductively define sequences 
 $\langle M_i : i \in \omega \rangle $
,
$\langle M_i : i \in \omega \rangle $
, 
 $\langle \bar r_i : i \in \omega \rangle $
, initial segments lying in N, so that for every
$\langle \bar r_i : i \in \omega \rangle $
, initial segments lying in N, so that for every 
 $i \in \omega $
,
$i \in \omega $
,
- 
–  $M_0 = M$
, $M_0 = M$
, $\bar r_0 = \bar q_0 \restriction \alpha _0$
, $\bar r_0 = \bar q_0 \restriction \alpha _0$
,
- 
–  $M_{i+1} \ni \bar q_0$
 is a countable model, $M_{i+1} \ni \bar q_0$
 is a countable model,
- 
–  $M_{i}, \bar r_i, a_i \in M_{i+1}$
, $M_{i}, \bar r_i, a_i \in M_{i+1}$
,
- 
–  $\bar r_i$
 is a good $\bar r_i$
 is a good $\mathbb {P}_{\alpha _i}$
 master condition over $\mathbb {P}_{\alpha _i}$
 master condition over $M_i$
, $M_i$
,
- 
–  $r_{i+1} \leq _{n + i,a_i\cap \alpha _i} r_i ^{\frown } \bar q_0 \restriction [\alpha _i, \alpha _{i+1})$
. $r_{i+1} \leq _{n + i,a_i\cap \alpha _i} r_i ^{\frown } \bar q_0 \restriction [\alpha _i, \alpha _{i+1})$
.
 Define for each 
 $i \in \omega $
,
$i \in \omega $
, 
 $\bar q_i = \bar r_i^{\frown } \bar q_0 \restriction [\alpha _i, \alpha ) $
. Then
$\bar q_i = \bar r_i^{\frown } \bar q_0 \restriction [\alpha _i, \alpha ) $
. Then 
 $\langle \bar q_i : i \in \omega \rangle $
 is a fusion sequence in
$\langle \bar q_i : i \in \omega \rangle $
 is a fusion sequence in 
 $\mathbb {P}_\alpha $
 and we can find a condition
$\mathbb {P}_\alpha $
 and we can find a condition 
 $\bar q \leq _{n,a} \bar q_0 \leq _{n,a} \bar p$
, where for each
$\bar q \leq _{n,a} \bar q_0 \leq _{n,a} \bar p$
, where for each 
 $\beta < \alpha $
,
$\beta < \alpha $
, 
 $\bar q \restriction \beta \Vdash \dot q(\beta ) = \bigcap _{i \in \omega } \dot q_i(\beta )$
. Finally let
$\bar q \restriction \beta \Vdash \dot q(\beta ) = \bigcap _{i \in \omega } \dot q_i(\beta )$
. Finally let 
 $[\bar q] := \bigcap _{i \in \omega } ([\bar r_i] \times (A^\omega )^{[\alpha _i, \alpha )})$
. Then (1) is easy to check. For (5), we can assume without loss of generality that
$[\bar q] := \bigcap _{i \in \omega } ([\bar r_i] \times (A^\omega )^{[\alpha _i, \alpha )})$
. Then (1) is easy to check. For (5), we can assume without loss of generality that 
 $\dot y$
 is a name for an element of
$\dot y$
 is a name for an element of 
 $\omega ^\omega $
 since for any Polish space X, there is a continuous surjection from
$\omega ^\omega $
 since for any Polish space X, there is a continuous surjection from 
 $\omega ^\omega $
 to X. Now let
$\omega ^\omega $
 to X. Now let 
 $(i_j)_{j \in \omega }$
 be increasing so that
$(i_j)_{j \in \omega }$
 be increasing so that 
 $\dot y \restriction j$
 is determined on
$\dot y \restriction j$
 is determined on 
 $\mathbb {P}_{\alpha _{i_j}}$
 for every
$\mathbb {P}_{\alpha _{i_j}}$
 for every 
 $j \in \omega $
. Since
$j \in \omega $
. Since 
 $\bar r_{i_j}$
 is a good master condition over M, there is a continuous function
$\bar r_{i_j}$
 is a good master condition over M, there is a continuous function 
 $f_j \colon [\bar r_{i_j}] \to \omega ^j$
 so that
$f_j \colon [\bar r_{i_j}] \to \omega ^j$
 so that 
 $\bar r_{i_j} \Vdash \dot y \restriction j = f_j(\bar x_{G_{\alpha _{i_j}}})$
 for every
$\bar r_{i_j} \Vdash \dot y \restriction j = f_j(\bar x_{G_{\alpha _{i_j}}})$
 for every 
 $j \in \omega $
. It is easy to put these functions together to a continuous function
$j \in \omega $
. It is easy to put these functions together to a continuous function 
 $f \colon [\bar q] \to 2^\omega $
, so that
$f \colon [\bar q] \to 2^\omega $
, so that 
 $f(\bar x) \restriction j = f_j(\bar x \restriction \alpha _{i_j})$
. Then we obviously have that
$f(\bar x) \restriction j = f_j(\bar x \restriction \alpha _{i_j})$
. Then we obviously have that 
 $\bar q \Vdash \dot y = f(\bar x_G)$
.
$\bar q \Vdash \dot y = f(\bar x_G)$
.
 Now let us fix for each 
 $i \in \omega $
,
$i \in \omega $
, 
 $C_i \subseteq \alpha _i$
 a countable set as given by Lemma 2.3 applied to
$C_i \subseteq \alpha _i$
 a countable set as given by Lemma 2.3 applied to 
 $\bar r_i$
,
$\bar r_i$
, 
 $M_i$
, which by elementarity exists in N. Let
$M_i$
, which by elementarity exists in N. Let 
 $C = \bigcup _{i \in \omega } C_i$
. Then
$C = \bigcup _{i \in \omega } C_i$
. Then 
 $[\bar q] = [\bar q] \restriction C \times (A^\omega )^{\alpha \setminus C}$
 and
$[\bar q] = [\bar q] \restriction C \times (A^\omega )^{\alpha \setminus C}$
 and 
 $[\bar q] \restriction C$
 is closed. For every
$[\bar q] \restriction C$
 is closed. For every 
 $\beta \in \alpha \setminus C$
, the map given in (3) is constant and maps to
$\beta \in \alpha \setminus C$
, the map given in (3) is constant and maps to 
 $\mathbb {Q}_\beta $
, as
$\mathbb {Q}_\beta $
, as 
 $A^{<\omega }$
 is the trivial condition. Thus we may restrict our attention to
$A^{<\omega }$
 is the trivial condition. Thus we may restrict our attention to 
 $\beta \in C$
. Let us write
$\beta \in C$
. Let us write 
 $X_i = ([\bar r_i] \times (A^\omega )^{[\alpha _i, \alpha )}) \restriction C$
 for every
$X_i = ([\bar r_i] \times (A^\omega )^{[\alpha _i, \alpha )}) \restriction C$
 for every 
 $i \in \omega $
 and note that
$i \in \omega $
 and note that 
 $\bigcap _{i \in \omega } X_i = [\bar q] \restriction C$
. For every
$\bigcap _{i \in \omega } X_i = [\bar q] \restriction C$
. For every 
 $\beta \in C$
,
$\beta \in C$
, 
 $\bar x \in [\bar q] \restriction (C \cap \beta )$
, and
$\bar x \in [\bar q] \restriction (C \cap \beta )$
, and 
 $i \in \omega $
, we write
$i \in \omega $
, we write 
 $$ \begin{align*}T_{\bar x} := \{s \in A^{<\omega} : \exists \bar z \in [\bar q] \restriction C (\bar z \restriction \beta = \bar x \wedge s \subseteq z(\beta)) \}\end{align*} $$
$$ \begin{align*}T_{\bar x} := \{s \in A^{<\omega} : \exists \bar z \in [\bar q] \restriction C (\bar z \restriction \beta = \bar x \wedge s \subseteq z(\beta)) \}\end{align*} $$
and
 $$ \begin{align*}T_{\bar x}^i = \{s \in A^{<\omega} : \exists \bar z \in X_i (\bar z \restriction \beta = \bar x \wedge s \subseteq z(\beta))\}.\end{align*} $$
$$ \begin{align*}T_{\bar x}^i = \{s \in A^{<\omega} : \exists \bar z \in X_i (\bar z \restriction \beta = \bar x \wedge s \subseteq z(\beta))\}.\end{align*} $$
Claim 2.4. For every 
 $i \in \omega $
, where
$i \in \omega $
, where 
 $\beta \in a_i$
,
$\beta \in a_i$
, 
 $T^{i+1}_{\bar x} \leq _{\beta , i} T^i_{\bar x}$
. In particular,
$T^{i+1}_{\bar x} \leq _{\beta , i} T^i_{\bar x}$
. In particular, 
 $\bigcap _{i \in \omega } T^i_{\bar x} \in \mathbb {Q}_\beta $
.
$\bigcap _{i \in \omega } T^i_{\bar x} \in \mathbb {Q}_\beta $
.
Proof. If 
 $\alpha _{i+1} \leq \beta $
, then
$\alpha _{i+1} \leq \beta $
, then 
 $T^{i+1}_{\bar x} = T^i_{\bar x} = A^{<\omega }$
. Else consider a
$T^{i+1}_{\bar x} = T^i_{\bar x} = A^{<\omega }$
. Else consider a 
 $\mathbb {P}_{\alpha _{i+2}}$
-name for
$\mathbb {P}_{\alpha _{i+2}}$
-name for 
 $(T^{i+1}_{\bar y}, T^i_{\bar y}) \in \mathcal {T}^2$
, where
$(T^{i+1}_{\bar y}, T^i_{\bar y}) \in \mathcal {T}^2$
, where 
 $\bar y = \bar x_G \restriction (C\cap \beta )$
. Such a name exists in
$\bar y = \bar x_G \restriction (C\cap \beta )$
. Such a name exists in 
 $M_{i+2}$
 and
$M_{i+2}$
 and 
 $\beta \in a_i \subseteq M_{i+2}$
. Thus
$\beta \in a_i \subseteq M_{i+2}$
. Thus 
 $\leq _{\beta , i}\in M_{i+2}$
 and by Lemma 2.3, we have that for every
$\leq _{\beta , i}\in M_{i+2}$
 and by Lemma 2.3, we have that for every 
 $\bar y \in [\bar r_{i+2}] \restriction (C\cap \beta )$
,
$\bar y \in [\bar r_{i+2}] \restriction (C\cap \beta )$
, 
 $(T^{i+1}_{\bar y}, T^i_{\bar y}) \in \leq _{\beta ,i}$
, thus also for
$(T^{i+1}_{\bar y}, T^i_{\bar y}) \in \leq _{\beta ,i}$
, thus also for 
 $\bar y = \bar x$
. The rest follows from the fact that the statement, that for any fusion sequence in
$\bar y = \bar x$
. The rest follows from the fact that the statement, that for any fusion sequence in 
 $\mathbb {Q}_\beta $
, its intersection is in
$\mathbb {Q}_\beta $
, its intersection is in 
 $\mathbb {Q}_\beta $
, is
$\mathbb {Q}_\beta $
, is 
 $\mathbf {\Pi }^1_2$
 and thus absolute.⊣
$\mathbf {\Pi }^1_2$
 and thus absolute.⊣
Claim 2.5. For every 
 $\gamma $
,
$\gamma $
, 
 $\bigcap _{i \in \omega } (X_i \restriction \gamma ) = (\bigcap _{i \in \omega } X_i) \restriction \gamma $
.
$\bigcap _{i \in \omega } (X_i \restriction \gamma ) = (\bigcap _{i \in \omega } X_i) \restriction \gamma $
.
Proof. That 
 $\bigcap _{i \in \omega } (X_i \restriction \gamma ) \supseteq (\bigcap _{i \in \omega } X_i) \restriction \gamma $
 is obvious. Let us show by induction on
$\bigcap _{i \in \omega } (X_i \restriction \gamma ) \supseteq (\bigcap _{i \in \omega } X_i) \restriction \gamma $
 is obvious. Let us show by induction on 
 $\delta \in C$
, that for any
$\delta \in C$
, that for any 
 $\delta ' \in C \cap \delta $
,
$\delta ' \in C \cap \delta $
, 
 $$ \begin{align*}\bigcap_{i \in \omega} (X_i \restriction \delta') \subseteq \left(\bigcap_{i \in \omega} (X_i \restriction \delta)\right) \restriction \delta'.\end{align*} $$
$$ \begin{align*}\bigcap_{i \in \omega} (X_i \restriction \delta') \subseteq \left(\bigcap_{i \in \omega} (X_i \restriction \delta)\right) \restriction \delta'.\end{align*} $$
The base case 
 $\delta = \min C$
 is clear. For the limit case, let
$\delta = \min C$
 is clear. For the limit case, let 
 $\delta ' \in C \cap \delta $
 be given and let
$\delta ' \in C \cap \delta $
 be given and let 
 $(\delta _n)_{n \in \omega } $
 be increasing cofinal in
$(\delta _n)_{n \in \omega } $
 be increasing cofinal in 
 $(C \cap \delta ) \setminus \delta '$
. Whenever
$(C \cap \delta ) \setminus \delta '$
. Whenever 
 $\bar y \in \bigcap _{i \in \omega } (X_i \restriction \delta ')$
, by the inductive hypothesis, there is
$\bar y \in \bigcap _{i \in \omega } (X_i \restriction \delta ')$
, by the inductive hypothesis, there is 
 $\bar y_0 \in \bigcap _{i \in \omega } (X_i \restriction \delta _0)$
 extending
$\bar y_0 \in \bigcap _{i \in \omega } (X_i \restriction \delta _0)$
 extending 
 $\bar y$
. In particular, there is
$\bar y$
. In particular, there is 
 $\bar z_0 \in X_0 \restriction \delta $
 extending
$\bar z_0 \in X_0 \restriction \delta $
 extending 
 $\bar y_0$
. Next, there is
$\bar y_0$
. Next, there is 
 $\bar y_1 \in \bigcap _{i \in \omega } (X_i \restriction \delta _1)$
 extending
$\bar y_1 \in \bigcap _{i \in \omega } (X_i \restriction \delta _1)$
 extending 
 $\bar y_0$
 and
$\bar y_0$
 and 
 $\bar z_1 \in X_1 \restriction \delta $
 extending
$\bar z_1 \in X_1 \restriction \delta $
 extending 
 $\bar y_1$
. Continuing like this, we find a sequence
$\bar y_1$
. Continuing like this, we find a sequence 
 $\langle \bar z_n : n \in \omega \rangle $
 that converges to
$\langle \bar z_n : n \in \omega \rangle $
 that converges to 
 $\bar z \in (A^\omega )^{C \cap \delta }$
. Since
$\bar z \in (A^\omega )^{C \cap \delta }$
. Since 
 $\langle \bar z_n : n \geq m \rangle $
 is contained within the closed set
$\langle \bar z_n : n \geq m \rangle $
 is contained within the closed set 
 $X_m \restriction \delta $
 for each
$X_m \restriction \delta $
 for each 
 $m \in \omega $
,
$m \in \omega $
, 
 $\bar z \in \bigcap _{i \in \omega } (X_i \restriction \delta )$
. Since
$\bar z \in \bigcap _{i \in \omega } (X_i \restriction \delta )$
. Since 
 $\bar z \restriction \delta ' = \bar y$
, this proves the limit case. Now assume
$\bar z \restriction \delta ' = \bar y$
, this proves the limit case. Now assume 
 $\delta = \xi +1$
. Let
$\delta = \xi +1$
. Let 
 $\delta ' < \delta $
 be given and let
$\delta ' < \delta $
 be given and let 
 $\bar y \in \bigcap _{i \in \omega } (X_i \restriction \delta ')$
. Then there is
$\bar y \in \bigcap _{i \in \omega } (X_i \restriction \delta ')$
. Then there is 
 $\bar z \in \bigcap _{i \in \omega } (X_i \restriction \xi )$
 extending
$\bar z \in \bigcap _{i \in \omega } (X_i \restriction \xi )$
 extending 
 $\bar y$
 by the inductive hypothesis. Since
$\bar y$
 by the inductive hypothesis. Since 
 $\bigcap _{i \in \omega } T^i_{\bar z} \in \mathbb {Q}_\delta $
, there is
$\bigcap _{i \in \omega } T^i_{\bar z} \in \mathbb {Q}_\delta $
, there is 
 $u \in [\bigcap _{i \in \omega }T^i_{\bar z}]$
 and
$u \in [\bigcap _{i \in \omega }T^i_{\bar z}]$
 and 
 $\bar z {}^{\frown } u \in \bigcap _{i \in \omega } (X_i \restriction \delta )$
. To finish the proof apply the induction step one more time to
$\bar z {}^{\frown } u \in \bigcap _{i \in \omega } (X_i \restriction \delta )$
. To finish the proof apply the induction step one more time to 
 $\delta = \sup \{ \xi +1 : \xi \in C\}$
 and
$\delta = \sup \{ \xi +1 : \xi \in C\}$
 and 
 $\delta ' = \gamma $
.⊣
$\delta ' = \gamma $
.⊣
 Claim 2.5 shows that (4) holds as 
 $[\bar q] \restriction \beta = (\bigcap _{i \in \omega } X_i) \restriction \beta \times (A^\omega )^{\beta \setminus C} = \bigcap _{i \in \omega } (X_i \restriction \beta ) \times (A^\omega )^{\beta \setminus C}$
 and
$[\bar q] \restriction \beta = (\bigcap _{i \in \omega } X_i) \restriction \beta \times (A^\omega )^{\beta \setminus C} = \bigcap _{i \in \omega } (X_i \restriction \beta ) \times (A^\omega )^{\beta \setminus C}$
 and 
 $\bigcap _{i \in \omega } (X_i \restriction \beta )$
 is closed, being an intersection of closed sets.
$\bigcap _{i \in \omega } (X_i \restriction \beta )$
 is closed, being an intersection of closed sets.
Claim 2.6. 
 $T_{\bar x} = \bigcap _{i \in \omega } T^i_{\bar x}$
.
$T_{\bar x} = \bigcap _{i \in \omega } T^i_{\bar x}$
.
Proof. That 
 $T_{\bar x} \subseteq \bigcap _{i \in \omega } T^i_{\bar x}$
 is clear from the definitions. Thus let
$T_{\bar x} \subseteq \bigcap _{i \in \omega } T^i_{\bar x}$
 is clear from the definitions. Thus let 
 $s \in \bigcap _{i \in \omega } T^i_{\bar x}$
. As
$s \in \bigcap _{i \in \omega } T^i_{\bar x}$
. As 
 $\bigcap _{i \in \omega } T^i_{\bar x} \in \mathbb {Q}_\beta $
, there is
$\bigcap _{i \in \omega } T^i_{\bar x} \in \mathbb {Q}_\beta $
, there is 
 $y \in [\bigcap _{i \in \omega } T^i_{\bar x}]$
 with
$y \in [\bigcap _{i \in \omega } T^i_{\bar x}]$
 with 
 $s \subseteq y$
. In particular,
$s \subseteq y$
. In particular, 
 $\bar x {}^{\frown } y \in \bigcap _{i \in \omega } (X_i \restriction (\beta +1)) = (\bigcap _{i \in \omega } X_i )\restriction (\beta +1)$
. So there is
$\bar x {}^{\frown } y \in \bigcap _{i \in \omega } (X_i \restriction (\beta +1)) = (\bigcap _{i \in \omega } X_i )\restriction (\beta +1)$
. So there is 
 $\bar z \in \bigcap _{i \in \omega } X_i$
 with
$\bar z \in \bigcap _{i \in \omega } X_i$
 with 
 $\bar z \restriction \beta = \bar x$
 and
$\bar z \restriction \beta = \bar x$
 and 
 $z(\beta ) = y \supseteq s$
. Thus
$z(\beta ) = y \supseteq s$
. Thus 
 $s \in T_{\bar x}$
.⊣
$s \in T_{\bar x}$
.⊣
 Now (2) follows easily. For the continuity of 
 $\bar x \mapsto T_{\bar x}$
, let
$\bar x \mapsto T_{\bar x}$
, let 
 $t \in A^{<\omega }$
 be arbitrary and j large enough so that
$t \in A^{<\omega }$
 be arbitrary and j large enough so that 
 $\vert t \vert \leq j$
 and
$\vert t \vert \leq j$
 and 
 $\beta \in a_j$
. Then
$\beta \in a_j$
. Then 
 $\{ \bar x \in [\bar q] \restriction \beta : t \notin T_{\bar x} \} = \{ \bar x \in [\bar q] \restriction \beta : t \notin T^j_{\bar x} \}$
 and
$\{ \bar x \in [\bar q] \restriction \beta : t \notin T_{\bar x} \} = \{ \bar x \in [\bar q] \restriction \beta : t \notin T^j_{\bar x} \}$
 and 
 $\{ \bar x \in [\bar q] \restriction \beta : t \in T_{\bar x} \} = \{ \bar x \in [\bar q] \restriction \beta : t \in T^j_{\bar x} \}$
 which are both open.Footnote 
4
 Thus we have shown (3).⊣
$\{ \bar x \in [\bar q] \restriction \beta : t \in T_{\bar x} \} = \{ \bar x \in [\bar q] \restriction \beta : t \in T^j_{\bar x} \}$
 which are both open.Footnote 
4
 Thus we have shown (3).⊣
Lemma 2.7. Let 
 $C \subseteq \alpha $
 be countable and
$C \subseteq \alpha $
 be countable and 
 $X \subseteq (A^{\omega })^C$
 be a closed set so that for every
$X \subseteq (A^{\omega })^C$
 be a closed set so that for every 
 $\beta \in C$
 and
$\beta \in C$
 and 
 $\bar x \in X \restriction \beta $
,
$\bar x \in X \restriction \beta $
, 
 $$ \begin{align*}\{s \in A^{<\omega} : \exists \bar z \in X (\bar z \restriction \beta = \bar x \wedge s \subseteq z(\beta)) \} \in \mathbb{Q}_\beta.\end{align*} $$
$$ \begin{align*}\{s \in A^{<\omega} : \exists \bar z \in X (\bar z \restriction \beta = \bar x \wedge s \subseteq z(\beta)) \} \in \mathbb{Q}_\beta.\end{align*} $$
Let 
 $M \ni X$
 be countable elementary. Then there is a good master condition
$M \ni X$
 be countable elementary. Then there is a good master condition 
 $\bar r$
 over M so that
$\bar r$
 over M so that 
 $[\bar r] \restriction C \subseteq X$
.
$[\bar r] \restriction C \subseteq X$
.
Proof. It is easy to construct 
 $\bar q \in M$
 recursively so that
$\bar q \in M$
 recursively so that 
 $\bar q \Vdash \bar x_G \restriction C \in X$
. By Lemma 2.2, we can extend
$\bar q \Vdash \bar x_G \restriction C \in X$
. By Lemma 2.2, we can extend 
 $\bar q$
 to a good master condition
$\bar q$
 to a good master condition 
 $\bar r$
 over M. The unique continuous function
$\bar r$
 over M. The unique continuous function 
 $f \colon [\bar r] \to (A^{\omega })^C$
 so that for generic G,
$f \colon [\bar r] \to (A^{\omega })^C$
 so that for generic G, 
 $f(\bar x_G) = \bar x_G \restriction C$
, is so that
$f(\bar x_G) = \bar x_G \restriction C$
, is so that 
 $f(\bar x) = \bar x \restriction C$
 for every
$f(\bar x) = \bar x \restriction C$
 for every 
 $\bar x \in [\bar r]$
. Since f maps to X,
$\bar x \in [\bar r]$
. Since f maps to X, 
 $[\bar r] \restriction C \subseteq X$
.⊣
$[\bar r] \restriction C \subseteq X$
.⊣
3 The Main Lemma
3.1 Mutual Cohen genericity
 Let X be a Polish space and M a model of set theory with 
 $X \in M$
. Recall that
$X \in M$
. Recall that 
 $x \in X$
 is Cohen generic in X over M if for any open dense
$x \in X$
 is Cohen generic in X over M if for any open dense 
 $O \subseteq X$
, such that
$O \subseteq X$
, such that 
 $O \in M$
,
$O \in M$
, 
 $x \in O$
.
$x \in O$
.
 Let 
 $x_0, \dots , x_{n-1} \in X$
. Then we say that
$x_0, \dots , x_{n-1} \in X$
. Then we say that 
 $x_0, \dots ,x_{n-1}$
 are (X-)mutually Cohen generic (mCg) over M if
$x_0, \dots ,x_{n-1}$
 are (X-)mutually Cohen generic (mCg) over M if 
 $(y_0, \dots , y_{K-1})$
 is a Cohen generic real over M in the Polish space
$(y_0, \dots , y_{K-1})$
 is a Cohen generic real over M in the Polish space 
 $X^K$
, where
$X^K$
, where 
 $\langle y_i : i < K \rangle $
 is some, equivalently any, enumeration of
$\langle y_i : i < K \rangle $
 is some, equivalently any, enumeration of 
 $\{ x_0, \dots , x_{n-1} \}$
. In particular, we allow for repetition in the definition of mutual genericity.
$\{ x_0, \dots , x_{n-1} \}$
. In particular, we allow for repetition in the definition of mutual genericity.
Definition 3.1. Let 
 $\langle X_l : l < k \rangle \in M$
 be Polish spaces. Then we say that
$\langle X_l : l < k \rangle \in M$
 be Polish spaces. Then we say that 
 $\bar x_0, \dots , \bar x_{n-1} \in \prod _{l<k} X_l$
 are
$\bar x_0, \dots , \bar x_{n-1} \in \prod _{l<k} X_l$
 are 
 $\langle X_l : l < k \rangle $
-mutually Cohen generic (mCg) over M, if their components are mutually added Cohen generics, i.e.,
$\langle X_l : l < k \rangle $
-mutually Cohen generic (mCg) over M, if their components are mutually added Cohen generics, i.e., 
 $$ \begin{align*}(y_0^0, \dots, y_0^{K_0}, \dots,y_{k-1}^0, \dots, y_{k-1}^{K_{k-1}}) \text{ is Cohen generic in } \prod_{l<k }X_l^{K_l} \text{ over } M,\end{align*} $$
$$ \begin{align*}(y_0^0, \dots, y_0^{K_0}, \dots,y_{k-1}^0, \dots, y_{k-1}^{K_{k-1}}) \text{ is Cohen generic in } \prod_{l<k }X_l^{K_l} \text{ over } M,\end{align*} $$
where 
 $\langle y_l^{i} : i < K_l \rangle $
 is some, equivalently any, enumeration of
$\langle y_l^{i} : i < K_l \rangle $
 is some, equivalently any, enumeration of 
 $\{ x_i(l) : i<n \}$
 for each
$\{ x_i(l) : i<n \}$
 for each 
 $l< k$
.
$l< k$
.
Definition 3.2. Let X be a Polish space with a fixed countable basis 
 $\mathcal {B}$
. Then we define the forcing poset
$\mathcal {B}$
. Then we define the forcing poset 
 $\mathbb {C}(2^\omega , X)$
 consisting of functions
$\mathbb {C}(2^\omega , X)$
 consisting of functions 
 $h \colon 2^{\leq n} \to \mathcal {B} \setminus \{\emptyset \}$
 for some
$h \colon 2^{\leq n} \to \mathcal {B} \setminus \{\emptyset \}$
 for some 
 $n \in \omega $
 such that
$n \in \omega $
 such that 
 $\forall \sigma \subseteq \tau \in 2^{\leq n} (h(\sigma ) \supseteq h(\tau ))$
. The poset is ordered by function extension.
$\forall \sigma \subseteq \tau \in 2^{\leq n} (h(\sigma ) \supseteq h(\tau ))$
. The poset is ordered by function extension.
 The poset 
 $\mathbb {C}(2^\omega , X)$
 adds generically a continuous function
$\mathbb {C}(2^\omega , X)$
 adds generically a continuous function 
 $\chi \colon 2^\omega \to X$
, given by
$\chi \colon 2^\omega \to X$
, given by 
 $\chi (x) = y$
 where
$\chi (x) = y$
 where 
 $\bigcap _{n \in \omega } h(x \restriction n ) = \{ y\}$
 and
$\bigcap _{n \in \omega } h(x \restriction n ) = \{ y\}$
 and 
 $h = \bigcup G$
 for G the generic filter. This forcing will be used in this section several times to obtain ZFC results. Note for instance that if G is generic over M, then for any
$h = \bigcup G$
 for G the generic filter. This forcing will be used in this section several times to obtain ZFC results. Note for instance that if G is generic over M, then for any 
 $x \in 2^\omega $
,
$x \in 2^\omega $
, 
 $\chi (x)$
 is Cohen generic in X over M, and moreover, for any
$\chi (x)$
 is Cohen generic in X over M, and moreover, for any 
 $x_0, \dots , x_{n-1} \in 2^\omega $
,
$x_0, \dots , x_{n-1} \in 2^\omega $
, 
 $\chi (x_0), \dots , \chi (x_{n-1})$
 are X-mutually Cohen generic over M. Sometimes we will use
$\chi (x_0), \dots , \chi (x_{n-1})$
 are X-mutually Cohen generic over M. Sometimes we will use 
 $\mathbb {C}(2^\omega , X)$
 to force over a countable model a continuous function from a space homeomorphic to
$\mathbb {C}(2^\omega , X)$
 to force over a countable model a continuous function from a space homeomorphic to 
 $2^\omega $
, such as
$2^\omega $
, such as 
 $(2^\omega )^\alpha $
 for
$(2^\omega )^\alpha $
 for 
 $\alpha < \omega _1$
.
$\alpha < \omega _1$
.
Lemma 3.3. Let M be a model of set theory, 
 $K, n \in \omega $
,
$K, n \in \omega $
, 
 $X_j \in M$
 a Polish space for every
$X_j \in M$
 a Polish space for every 
 $j < n$
, and G a
$j < n$
, and G a 
 $\prod _{j < n} \mathbb {C}(2^\omega , X_j)$
-generic over M yielding
$\prod _{j < n} \mathbb {C}(2^\omega , X_j)$
-generic over M yielding 
 $\chi _j \colon 2^\omega \to X_j$
 for every
$\chi _j \colon 2^\omega \to X_j$
 for every 
 $j <n$
. Then, whenever
$j <n$
. Then, whenever 
 $\bar x$
 is Cohen generic in
$\bar x$
 is Cohen generic in 
 $(2^\omega )^K$
 over
$(2^\omega )^K$
 over 
 $M[G]$
 and
$M[G]$
 and 
 $u_0, \dots , u_{n-1} \in 2^\omega \cap M[\bar x]$
 are pairwise distinct,
$u_0, \dots , u_{n-1} \in 2^\omega \cap M[\bar x]$
 are pairwise distinct, 
 $$ \begin{align*}\bar x^{\frown} \langle \chi_j(u_i) : i < n, j < n \rangle\end{align*} $$
$$ \begin{align*}\bar x^{\frown} \langle \chi_j(u_i) : i < n, j < n \rangle\end{align*} $$
is Cohen generic in
 $$ \begin{align*}(2^\omega)^K \times \prod_{i < n} X_i\end{align*} $$
$$ \begin{align*}(2^\omega)^K \times \prod_{i < n} X_i\end{align*} $$
over M.
Proof. Since 
 $\bar x$
 is generic over M it suffices to show that
$\bar x$
 is generic over M it suffices to show that 
 $\langle \chi _j(u_i) : i < n, j \rangle $
 is generic over
$\langle \chi _j(u_i) : i < n, j \rangle $
 is generic over 
 $M[\bar x]$
. Let
$M[\bar x]$
. Let 
 $\dot O \in M$
 be a
$\dot O \in M$
 be a 
 $(2^{<\omega })^K$
-name for a dense open subset of
$(2^{<\omega })^K$
-name for a dense open subset of 
 $\prod _{j < n} (X_j)^n$
 and
$\prod _{j < n} (X_j)^n$
 and 
 $\dot u_i$
 a
$\dot u_i$
 a 
 $(2^{<\omega })^K$
-name for
$(2^{<\omega })^K$
-name for 
 $u_i$
,
$u_i$
, 
 $i < n$
, such that the trivial condition forces that the
$i < n$
, such that the trivial condition forces that the 
 $\dot u_i$
 are pairwise distinct. Then consider the set
$\dot u_i$
 are pairwise distinct. Then consider the set 
 $$ \begin{align*} D &:= \{ (\bar h,\bar s) \in \prod_{i<n} \mathbb{C}(2^\omega, X_i) \times (2^{<\omega})^K : \exists t_0, \dots, t_{n-1}\in 2^{<\omega} \\ &\quad( \forall i < n(\bar s \Vdash t_i \subseteq \dot u_i) \wedge \bar s \Vdash \prod_{i,j<n} h_{j}(t_i) \subseteq \dot O) \}. \end{align*} $$
$$ \begin{align*} D &:= \{ (\bar h,\bar s) \in \prod_{i<n} \mathbb{C}(2^\omega, X_i) \times (2^{<\omega})^K : \exists t_0, \dots, t_{n-1}\in 2^{<\omega} \\ &\quad( \forall i < n(\bar s \Vdash t_i \subseteq \dot u_i) \wedge \bar s \Vdash \prod_{i,j<n} h_{j}(t_i) \subseteq \dot O) \}. \end{align*} $$
 We claim that this set is dense in 
 $\prod _{i<n} \mathbb {C}(2^\omega , X_i) \times (2^{<\omega })^K$
 which finishes the proof. Namely let
$\prod _{i<n} \mathbb {C}(2^\omega , X_i) \times (2^{<\omega })^K$
 which finishes the proof. Namely let 
 $(\bar h, \bar s)$
 be arbitrary, wlog
$(\bar h, \bar s)$
 be arbitrary, wlog 
 $\operatorname {\mathrm {dom}} h_{j} = 2^{\leq n_0}$
 for every
$\operatorname {\mathrm {dom}} h_{j} = 2^{\leq n_0}$
 for every 
 $j < n$
. Then we can extend
$j < n$
. Then we can extend 
 $\bar s$
 to
$\bar s$
 to 
 $\bar s'$
 so that there are incompatible
$\bar s'$
 so that there are incompatible 
 $t_i$
, with
$t_i$
, with 
 $\vert t_i \vert \geq n_0$
, so that
$\vert t_i \vert \geq n_0$
, so that 
 $\bar s' \Vdash t_i \subseteq \dot u_i$
 and there are
$\bar s' \Vdash t_i \subseteq \dot u_i$
 and there are 
 $U_{i,j} \subseteq h_j(t_i \restriction n_0)$
 basic open subsets of
$U_{i,j} \subseteq h_j(t_i \restriction n_0)$
 basic open subsets of 
 $X_j$
 in M for every
$X_j$
 in M for every 
 $i < n$
 and
$i < n$
 and 
 $j<n$
, so that
$j<n$
, so that 
 $\bar s' \Vdash \prod _{i,j < n} U_{i,j} \subseteq \dot O$
. Then we can extend
$\bar s' \Vdash \prod _{i,j < n} U_{i,j} \subseteq \dot O$
. Then we can extend 
 $\bar h$
 to
$\bar h$
 to 
 $\bar h'$
 so that
$\bar h'$
 so that 
 $h^{\prime }_j(t_i) = U_{i,j}$
 for every
$h^{\prime }_j(t_i) = U_{i,j}$
 for every 
 $i,j < n$
. We see that
$i,j < n$
. We see that 
 $(\bar h', \bar s') \in D$
.⊣
$(\bar h', \bar s') \in D$
.⊣
3.2 Finite products
This subsection can be skipped entirely if one is only interested in the results for the countable support iteration. The lemma that we will prove below is relevant to finite products instead (see Theorem 5.5). It is very similar to Main Lemma 3.14 in the next subsection, but the proofs are completely different. Main Lemma 3.4 is based on a forcing-theoretic proof of the Halpern–Läuchli theorem that is commonly attributed to L. Harrington (see, e.g., [Reference Golshani16, Lemma 4.2.4] as a reference). On the other hand, Main Lemma 3.14 uses an inductive argument.
Main Lemma 3.4. Let 
 $k \in \omega $
 and
$k \in \omega $
 and 
 $E \subseteq [(2^\omega )^k]^{<\omega }$
 an analytic hypergraph on
$E \subseteq [(2^\omega )^k]^{<\omega }$
 an analytic hypergraph on 
 $(2^\omega )^k$
. Then there is a countable model M so that either
$(2^\omega )^k$
. Then there is a countable model M so that either
- 
(1) for any  $n \in \omega $
 and $n \in \omega $
 and $\bar x_0,\dots , \bar x_{n-1} \in (2^\omega )^k$
 that are $\bar x_0,\dots , \bar x_{n-1} \in (2^\omega )^k$
 that are $\langle 2^\omega : l < k\rangle $
-mCg over M, $\langle 2^\omega : l < k\rangle $
-mCg over M, $$ \begin{align*}\{\bar x_0,\dots, \bar x_{n-1}\} \text{ is } E \text{-independent}\end{align*} $$ $$ \begin{align*}\{\bar x_0,\dots, \bar x_{n-1}\} \text{ is } E \text{-independent}\end{align*} $$
 or for some 
 $N \in \omega $
,
$N \in \omega $
, 
- 
(2) there are  $\phi _0, \dots , \phi _{N-1} \colon (2^\omega )^k \to (2^\omega )^k$
 continuous, $\phi _0, \dots , \phi _{N-1} \colon (2^\omega )^k \to (2^\omega )^k$
 continuous, $\bar s \in \bigotimes _{l<k} 2^{<\omega }$
 so that for any $\bar s \in \bigotimes _{l<k} 2^{<\omega }$
 so that for any $n \in \omega $
 and $n \in \omega $
 and $\bar x_0,\dots , \bar x_{n-1} \in (2^\omega )^k \cap [\bar s]$
, that are $\bar x_0,\dots , \bar x_{n-1} \in (2^\omega )^k \cap [\bar s]$
, that are $\langle 2^\omega : l < k\rangle $
-mCg over M, $\langle 2^\omega : l < k\rangle $
-mCg over M, $$ \begin{align*}\{\phi_j(\bar x_i) : j < N, i<n\} \text{ is } E \text{-independent but } \{ \bar x_0\} \cup \{ \phi_j(\bar x_0) : j < N \} \in E.\end{align*} $$ $$ \begin{align*}\{\phi_j(\bar x_i) : j < N, i<n\} \text{ is } E \text{-independent but } \{ \bar x_0\} \cup \{ \phi_j(\bar x_0) : j < N \} \in E.\end{align*} $$
Remark 3.5. Note that 
 $N = 0$
 is possible in the second option. For example whenever
$N = 0$
 is possible in the second option. For example whenever 
 $[(2^\omega )^k]^1 \subseteq E$
, then
$[(2^\omega )^k]^1 \subseteq E$
, then 
 $\emptyset $
 is the only E-independent set. In this case the last line simplifies to “
$\emptyset $
 is the only E-independent set. In this case the last line simplifies to “
 $\{ \bar x_0 \} \in E$
.”
$\{ \bar x_0 \} \in E$
.”
Proof. Let 
 $\kappa = \beth _{2k -1}(\aleph _0)^+$
. Recall that by Erdős–Rado (see [Reference Jech18, Theorem 9.6]), for any
$\kappa = \beth _{2k -1}(\aleph _0)^+$
. Recall that by Erdős–Rado (see [Reference Jech18, Theorem 9.6]), for any 
 ${c \colon [\kappa ]^{2k} \to H(\omega )}$
, there is
${c \colon [\kappa ]^{2k} \to H(\omega )}$
, there is 
 $B \in [\kappa ]^{\aleph _1}$
 which is monochromatic for c, i.e.,
$B \in [\kappa ]^{\aleph _1}$
 which is monochromatic for c, i.e., 
 $c \restriction [B]^{2k}$
 is constant. Let
$c \restriction [B]^{2k}$
 is constant. Let 
 $\mathbb {Q}$
 be the forcing adding
$\mathbb {Q}$
 be the forcing adding 
 $\kappa $
 many Cohen reals
$\kappa $
 many Cohen reals 
 $$ \begin{align*}\langle z_{(l,\alpha)} : \alpha < \kappa\rangle \text{ in } 2^{\omega} \text{ for each } l<k\end{align*} $$
$$ \begin{align*}\langle z_{(l,\alpha)} : \alpha < \kappa\rangle \text{ in } 2^{\omega} \text{ for each } l<k\end{align*} $$
with finite conditions, i.e., 
 $\mathbb {Q} = \prod _\kappa ^{<\omega } (2^{<\omega })^k$
. We will use the notational convention that elements of
$\mathbb {Q} = \prod _\kappa ^{<\omega } (2^{<\omega })^k$
. We will use the notational convention that elements of 
 $[\kappa ]^d$
, for
$[\kappa ]^d$
, for 
 $d \in \omega $
, are sequences
$d \in \omega $
, are sequences 
 $\bar \alpha = (\alpha _0, \dots , \alpha _{d-1})$
 ordered increasingly. For any
$\bar \alpha = (\alpha _0, \dots , \alpha _{d-1})$
 ordered increasingly. For any 
 $\bar \alpha \in [\kappa ]^k$
 we define
$\bar \alpha \in [\kappa ]^k$
 we define 
 $\bar z_{\bar \alpha } := (z_{(0,\alpha _0)}, \dots , z_{(k-1,\alpha _{k-1})}) \in (2^\omega )^k$
.
$\bar z_{\bar \alpha } := (z_{(0,\alpha _0)}, \dots , z_{(k-1,\alpha _{k-1})}) \in (2^\omega )^k$
.
 Let 
 $\dot {\mathcal {A}}$
 be a
$\dot {\mathcal {A}}$
 be a 
 $\mathbb {Q}$
-name for a maximal E-independent subset of
$\mathbb {Q}$
-name for a maximal E-independent subset of 
 $\{ \bar z_{\bar \alpha } : \bar \alpha \in [\kappa ]^k \}$
, reinterpreting E in the extension by
$\{ \bar z_{\bar \alpha } : \bar \alpha \in [\kappa ]^k \}$
, reinterpreting E in the extension by 
 $\mathbb {Q}$
. For any
$\mathbb {Q}$
. For any 
 $\bar \alpha \in [\kappa ]^k$
, we fix
$\bar \alpha \in [\kappa ]^k$
, we fix 
 $p_{\bar \alpha } \in \mathbb {Q}$
 so that either
$p_{\bar \alpha } \in \mathbb {Q}$
 so that either 

or
 $$ \begin{align} p_{\bar\alpha} \Vdash \bar z_{\bar \alpha} \not\in \dot{\mathcal{A}}. \end{align} $$
$$ \begin{align} p_{\bar\alpha} \Vdash \bar z_{\bar \alpha} \not\in \dot{\mathcal{A}}. \end{align} $$
In case (2) we additionally fix 
 $N_{\bar \alpha } < \omega $
 and
$N_{\bar \alpha } < \omega $
 and 
 $(\bar \beta ^i)_{i < N_{\bar \alpha }} = (\bar \beta ^i(\bar \alpha ))_{i < N_{\bar \alpha }}$
, and we assume that
$(\bar \beta ^i)_{i < N_{\bar \alpha }} = (\bar \beta ^i(\bar \alpha ))_{i < N_{\bar \alpha }}$
, and we assume that 
 $$ \begin{align*}p_{\bar\alpha} \Vdash \{ \bar z_{\bar \beta^i} : i < N_{\bar\alpha} \} \subseteq \dot{\mathcal{A}} \wedge \{ \bar z_{\bar \alpha} \} \cup \{ \bar z_{\bar \beta^i} : i < N_{\bar\alpha} \} \in E .\end{align*} $$
$$ \begin{align*}p_{\bar\alpha} \Vdash \{ \bar z_{\bar \beta^i} : i < N_{\bar\alpha} \} \subseteq \dot{\mathcal{A}} \wedge \{ \bar z_{\bar \alpha} \} \cup \{ \bar z_{\bar \beta^i} : i < N_{\bar\alpha} \} \in E .\end{align*} $$
We also define 
 $H_l(\bar \alpha ) = \{ \beta ^i_l : i < N_{\bar \alpha } \} \cup \{ \alpha _l \} \in [\kappa ]^{<\omega }$
 for each
$H_l(\bar \alpha ) = \{ \beta ^i_l : i < N_{\bar \alpha } \} \cup \{ \alpha _l \} \in [\kappa ]^{<\omega }$
 for each 
 $l < k$
.
$l < k$
.
 Now for 
 $\bar \alpha \in [\kappa ]^{2k}$
 we collect the following information:
$\bar \alpha \in [\kappa ]^{2k}$
 we collect the following information:
- 
(i) whether  $p_{\bar \alpha \restriction k } = p_{\alpha _0,\dots , \alpha _{k-1}}\Vdash \bar z_{\bar \alpha \restriction k} \in \dot {\mathcal {A}}$
 or not, $p_{\bar \alpha \restriction k } = p_{\alpha _0,\dots , \alpha _{k-1}}\Vdash \bar z_{\bar \alpha \restriction k} \in \dot {\mathcal {A}}$
 or not,
- 
(ii)  $\bar s = (p_{\bar \alpha \restriction k }(0,\alpha _0), \dots , p_{\bar \alpha \restriction k }(k-1,\alpha _{k-1})) \in (2^{<\omega })^k$
, $\bar s = (p_{\bar \alpha \restriction k }(0,\alpha _0), \dots , p_{\bar \alpha \restriction k }(k-1,\alpha _{k-1})) \in (2^{<\omega })^k$
,
- 
(iii) the relative position of the  $p_{\bar \gamma }$
 for $p_{\bar \gamma }$
 for $\bar \gamma \in \Gamma := \prod _{l<k} \{\alpha _{2l}, \alpha _{2l+1} \}$
 to each other. More precisely, consider $\bar \gamma \in \Gamma := \prod _{l<k} \{\alpha _{2l}, \alpha _{2l+1} \}$
 to each other. More precisely, consider $ \bigcup _{\bar \gamma \in \Gamma } \operatorname {\mathrm {dom}} p_{\bar \gamma } = \{0\} \times d_0 \cup \dots \cup \{k-1\} \times d_{k-1}$
 where $ \bigcup _{\bar \gamma \in \Gamma } \operatorname {\mathrm {dom}} p_{\bar \gamma } = \{0\} \times d_0 \cup \dots \cup \{k-1\} \times d_{k-1}$
 where $d_0,\dots , d_{k-1} \subseteq \kappa $
. Let $d_0,\dots , d_{k-1} \subseteq \kappa $
. Let $M_l = \vert d_l \vert $
 for $M_l = \vert d_l \vert $
 for $l<k$
 and for each $l<k$
 and for each $\bar j \in \prod _{l<k} \{2l, 2l+1 \}$
, collect a function $\bar j \in \prod _{l<k} \{2l, 2l+1 \}$
, collect a function $r_{\bar j}$
 with $r_{\bar j}$
 with $\operatorname {\mathrm {dom}} r_{\bar j} \subseteq \{0\} \times M_0 \cup \dots \cup \{k-1\} \times M_{k-1}$
 that is a copy of $\operatorname {\mathrm {dom}} r_{\bar j} \subseteq \{0\} \times M_0 \cup \dots \cup \{k-1\} \times M_{k-1}$
 that is a copy of $p_{\bar \gamma }$
, where $p_{\bar \gamma }$
, where $\bar \gamma = (\alpha _{j_0}, \dots , \alpha _{j_{k-1}})$
, $\bar \gamma = (\alpha _{j_0}, \dots , \alpha _{j_{k-1}})$
, $\bar j = (j_0, \dots , j_{k-1})$
. Namely, $\bar j = (j_0, \dots , j_{k-1})$
. Namely, $r_{\bar j}(l,m) = p_{\bar \gamma }(l,\beta )$
, whenever $r_{\bar j}(l,m) = p_{\bar \gamma }(l,\beta )$
, whenever $\beta $
 is the m-th element of $\beta $
 is the m-th element of $d_l$
. $d_l$
.
 In case 
 $p_{\bar \alpha \restriction k }\Vdash \bar z_{\bar \alpha \restriction k} \notin \dot {\mathcal {A}}$
 we additionally remember
$p_{\bar \alpha \restriction k }\Vdash \bar z_{\bar \alpha \restriction k} \notin \dot {\mathcal {A}}$
 we additionally remember
- 
(iv)  $N = N_{\bar \alpha \restriction k}$
, $N = N_{\bar \alpha \restriction k}$
,
- 
(v)  $N_l = \vert H_l(\bar \alpha \restriction k) \vert $
, for each $N_l = \vert H_l(\bar \alpha \restriction k) \vert $
, for each $l < k$
, $l < k$
,
- 
(vi)  $\bar b^i \in \prod _{l <k} N_l$
 so that $\bar b^i \in \prod _{l <k} N_l$
 so that $\beta ^i_l$
 is the $\beta ^i_l$
 is the $b^i_l$
-th element of $b^i_l$
-th element of $H_l(\bar \alpha \restriction k)$
, for each $H_l(\bar \alpha \restriction k)$
, for each $i<N$
, $i<N$
,
- 
(vii)  $\bar a \in \prod _{l <k} N_l$
 so that $\bar a \in \prod _{l <k} N_l$
 so that $\alpha _l$
 is the $\alpha _l$
 is the $a_l$
-th member of $a_l$
-th member of $H_l(\bar \alpha \restriction k)$
, $H_l(\bar \alpha \restriction k)$
,
- 
(viii) the partial function r with domain a subset of  $\bigcup _{l <k} \{l\} \times N_l$
, so that $\bigcup _{l <k} \{l\} \times N_l$
, so that $r(l,m)= t \in 2^{<\omega }$
 iff $r(l,m)= t \in 2^{<\omega }$
 iff $p_{\bar \alpha \restriction k}(l,\beta ) = t$
 where $p_{\bar \alpha \restriction k}(l,\beta ) = t$
 where $\beta $
 is the m-th element of $\beta $
 is the m-th element of $H_l(\bar \alpha \restriction k)$
. $H_l(\bar \alpha \restriction k)$
.
And finally we also remember
- 
(ix) for each pair  $\bar \gamma , \bar \delta \in \prod _{l<k} \{\alpha _{2l}, \alpha _{2l +1} \}$
, where $\bar \gamma , \bar \delta \in \prod _{l<k} \{\alpha _{2l}, \alpha _{2l +1} \}$
, where $\bar \gamma = (\alpha _{j_l})_{l < k}$
 and $\bar \gamma = (\alpha _{j_l})_{l < k}$
 and $\bar \delta = (\alpha _{j^{\prime }_l})_{l < k}$
, finite partial injections $\bar \delta = (\alpha _{j^{\prime }_l})_{l < k}$
, finite partial injections $e_{l, \bar j, \bar j'} \colon N_l \to N_l$
 so that $e_{l, \bar j, \bar j'} \colon N_l \to N_l$
 so that $e_{l, \bar j, \bar j'}(m) = m'$
 iff the m-th element of $e_{l, \bar j, \bar j'}(m) = m'$
 iff the m-th element of $H_l(\bar \gamma )$
 equals the $H_l(\bar \gamma )$
 equals the $m'$
-th element of $m'$
-th element of $H_l(\bar \delta )$
. $H_l(\bar \delta )$
.
 This information is finite and defines a coloring 
 $c \colon [\kappa ]^{2k} \to H(\omega )$
. Let
$c \colon [\kappa ]^{2k} \to H(\omega )$
. Let 
 $B \in [\kappa ]^{\omega _1}$
 be monochromatic for c. Let
$B \in [\kappa ]^{\omega _1}$
 be monochromatic for c. Let 
 $M \preccurlyeq H(\theta )$
 be countable for
$M \preccurlyeq H(\theta )$
 be countable for 
 $\theta $
 large enough so that
$\theta $
 large enough so that 
 $\kappa , c, B, \langle p_{\bar \alpha } : \bar \alpha \in [\kappa ]^k \rangle , E, \dot {\mathcal {A}} \in M$
.
$\kappa , c, B, \langle p_{\bar \alpha } : \bar \alpha \in [\kappa ]^k \rangle , E, \dot {\mathcal {A}} \in M$
.
Claim 3.6. If for every 
 $\bar \alpha \in [B]^k$
,
$\bar \alpha \in [B]^k$
, 
 $p_{\bar {\alpha }} \Vdash \bar z_{\bar \alpha } \in \dot {\mathcal {A}}$
, then (1) of the main lemma holds true.
$p_{\bar {\alpha }} \Vdash \bar z_{\bar \alpha } \in \dot {\mathcal {A}}$
, then (1) of the main lemma holds true.
Proof. Let 
 $\bar x_0, \dots , \bar x_{n-1}$
 be arbitrary mCg over M. Say
$\bar x_0, \dots , \bar x_{n-1}$
 be arbitrary mCg over M. Say 
 $\{x_i(l) : i < n\}$
 is enumerated by
$\{x_i(l) : i < n\}$
 is enumerated by 
 $\langle y^i_l : i < K_l \rangle $
 for every
$\langle y^i_l : i < K_l \rangle $
 for every 
 $l<k$
. Now find
$l<k$
. Now find 
 $$ \begin{align*}\alpha^0_0 < \dots < \alpha^{K_0 -1}_0 < \dots < \alpha^{0}_{k-1} < \dots < \alpha^{K_{k-1} -1}_{k-1}\end{align*} $$
$$ \begin{align*}\alpha^0_0 < \dots < \alpha^{K_0 -1}_0 < \dots < \alpha^{0}_{k-1} < \dots < \alpha^{K_{k-1} -1}_{k-1}\end{align*} $$
in 
 $M \cap B$
. Then there is a
$M \cap B$
. Then there is a 
 $\mathbb {Q}$
-generic G over M so that for any
$\mathbb {Q}$
-generic G over M so that for any 
 $\bar j \in \prod _{l < k} K_l$
,
$\bar j \in \prod _{l < k} K_l$
, 
 $$ \begin{align*}\bar z_{\bar \beta}[G] = (y^{j_0}_0, \dots, y^{j_{k-1}}_{k-1}),\end{align*} $$
$$ \begin{align*}\bar z_{\bar \beta}[G] = (y^{j_0}_0, \dots, y^{j_{k-1}}_{k-1}),\end{align*} $$
where 
 $\bar \beta = (\alpha ^{j_0}_0, \dots , \alpha ^{j_{k-1}}_{k-1})$
. In particular, for each
$\bar \beta = (\alpha ^{j_0}_0, \dots , \alpha ^{j_{k-1}}_{k-1})$
. In particular, for each 
 $i<n$
, there is
$i<n$
, there is 
 $\bar \beta _i \in [B \cap M]^k$
 so that
$\bar \beta _i \in [B \cap M]^k$
 so that 
 $\bar z_{\bar \beta _i}[G]= \bar x_{i}$
. Since
$\bar z_{\bar \beta _i}[G]= \bar x_{i}$
. Since 
 for every
 for every 
 $\bar \beta _i$
 we have that
$\bar \beta _i$
 we have that 
 $$ \begin{align*}M[G] \models \bar x_i \in \dot{\mathcal{A}}[G]\end{align*} $$
$$ \begin{align*}M[G] \models \bar x_i \in \dot{\mathcal{A}}[G]\end{align*} $$
for every 
 $i< n$
 and in particular
$i< n$
 and in particular 
 $$ \begin{align*}M[G] \models \{ \bar x_i : i < n \} \text{ is } E \text{-independent}.\end{align*} $$
$$ \begin{align*}M[G] \models \{ \bar x_i : i < n \} \text{ is } E \text{-independent}.\end{align*} $$
By absoluteness 
 $\{ \bar x_i : i < n \}$
 is indeed E-independent.⊣
$\{ \bar x_i : i < n \}$
 is indeed E-independent.⊣
 Assume from now on that 
 $p_{\bar \alpha }\Vdash \bar z_{\bar \alpha } \notin \dot {\mathcal {A}}$
 for every
$p_{\bar \alpha }\Vdash \bar z_{\bar \alpha } \notin \dot {\mathcal {A}}$
 for every 
 $\bar \alpha \in [B]^k$
. Then we may fix
$\bar \alpha \in [B]^k$
. Then we may fix 
 $\bar s$
, N,
$\bar s$
, N, 
 $(N_l)_{l<k}$
,
$(N_l)_{l<k}$
, 
 $\bar b^i$
 for
$\bar b^i$
 for 
 $i<N$
,
$i<N$
, 
 $\bar a$
, r and
$\bar a$
, r and 
 $e_{l,\bar j, \bar j'}$
 for all
$e_{l,\bar j, \bar j'}$
 for all 
 $l<k$
 and
$l<k$
 and 
 $\bar j,\bar j' \in \prod _{l'<k} \{2l', 2l'+1 \}$
 corresponding to the coloring on
$\bar j,\bar j' \in \prod _{l'<k} \{2l', 2l'+1 \}$
 corresponding to the coloring on 
 $[B]^{2k}$
.
$[B]^{2k}$
.
Claim 3.7. For any 
 $\bar \alpha \in [B]^{2k}$
 and
$\bar \alpha \in [B]^{2k}$
 and 
 $\bar \gamma , \bar \delta \in \prod _{l<k} \{\alpha _{2l}, \alpha _{2l +1} \}$
,
$\bar \gamma , \bar \delta \in \prod _{l<k} \{\alpha _{2l}, \alpha _{2l +1} \}$
, 
 $$ \begin{align*}p_{\bar\gamma} \restriction (\operatorname{\mathrm{dom}} p_{\bar\gamma} \cap \operatorname{\mathrm{dom}} p_{\bar\delta}) = p_{\bar\delta} \restriction (\operatorname{\mathrm{dom}} p_{\bar\gamma} \cap \operatorname{\mathrm{dom}} p_{\bar\delta}).\end{align*} $$
$$ \begin{align*}p_{\bar\gamma} \restriction (\operatorname{\mathrm{dom}} p_{\bar\gamma} \cap \operatorname{\mathrm{dom}} p_{\bar\delta}) = p_{\bar\delta} \restriction (\operatorname{\mathrm{dom}} p_{\bar\gamma} \cap \operatorname{\mathrm{dom}} p_{\bar\delta}).\end{align*} $$
Proof. Suppose not. By homogeneity we find a counterexample 
 $\bar \alpha $
,
$\bar \alpha $
, 
 $\bar \gamma $
,
$\bar \gamma $
, 
 $\bar \delta $
 where
$\bar \delta $
 where 
 $B \cap (\alpha _{2l'},\alpha _{2l' +1})$
 is non-empty for every
$B \cap (\alpha _{2l'},\alpha _{2l' +1})$
 is non-empty for every 
 $l' <k$
. So let
$l' <k$
. So let 
 $(l,\beta ) \in \operatorname {\mathrm {dom}} p_{\bar \gamma } \cap \operatorname {\mathrm {dom}} p_{\bar \delta }$
 such that
$(l,\beta ) \in \operatorname {\mathrm {dom}} p_{\bar \gamma } \cap \operatorname {\mathrm {dom}} p_{\bar \delta }$
 such that 
 $p_{\bar \gamma }(l,\beta ) = u \neq v = p_{\bar \delta }(l,\beta )$
. Let
$p_{\bar \gamma }(l,\beta ) = u \neq v = p_{\bar \delta }(l,\beta )$
. Let 
 $\bar \rho \in [B]^k$
 be such that for every
$\bar \rho \in [B]^k$
 be such that for every 
 $l' < k$
,
$l' < k$
, 
 $$ \begin{align*}\begin{cases} \rho_{l'} \in (\gamma_{l'}, \delta_{l'} ), & \text{if } \gamma_{l'} < \delta_{l'}, \\ \rho_{l'} \in (\delta_{l'}, \gamma_{l'} ), & \text{if } \delta_{l'} < \gamma_{l'}, \\ \rho_{l'} = \gamma_{l'}, & \text{if } \gamma_{l'} = \delta_{l'}. \end{cases}\end{align*} $$
$$ \begin{align*}\begin{cases} \rho_{l'} \in (\gamma_{l'}, \delta_{l'} ), & \text{if } \gamma_{l'} < \delta_{l'}, \\ \rho_{l'} \in (\delta_{l'}, \gamma_{l'} ), & \text{if } \delta_{l'} < \gamma_{l'}, \\ \rho_{l'} = \gamma_{l'}, & \text{if } \gamma_{l'} = \delta_{l'}. \end{cases}\end{align*} $$
Now note that 
 $\bar \rho $
’s relative position to
$\bar \rho $
’s relative position to 
 $\bar \gamma $
 is the same as that of
$\bar \gamma $
 is the same as that of 
 $\bar \delta $
 to
$\bar \delta $
 to 
 $\bar \gamma $
. More precisely, let
$\bar \gamma $
. More precisely, let 
 $\bar j, \bar j' \in \prod _{l'<k} \{2l', 2l'+1 \}$
 so that
$\bar j, \bar j' \in \prod _{l'<k} \{2l', 2l'+1 \}$
 so that 
 $\bar \gamma = (\alpha _{j_0}, \dots , \alpha _{j_{k-1}})$
 and
$\bar \gamma = (\alpha _{j_0}, \dots , \alpha _{j_{k-1}})$
 and 
 $\bar \delta = (\alpha _{j^{\prime }_0}, \dots , \alpha _{j^{\prime }_{k-1}})$
. Then there is
$\bar \delta = (\alpha _{j^{\prime }_0}, \dots , \alpha _{j^{\prime }_{k-1}})$
. Then there is 
 $\bar \beta \in [B]^{2k}$
 so that
$\bar \beta \in [B]^{2k}$
 so that 
 $\bar \gamma = (\beta _{j_0}, \dots , \beta _{j_{k-1}})$
 and
$\bar \gamma = (\beta _{j_0}, \dots , \beta _{j_{k-1}})$
 and 
 $\bar \rho = (\beta _{j^{\prime }_0}, \dots , \beta _{j^{\prime }_{k-1}})$
. Thus by homogeneity of
$\bar \rho = (\beta _{j^{\prime }_0}, \dots , \beta _{j^{\prime }_{k-1}})$
. Thus by homogeneity of 
 $[B]^{2k}$
 via c,
$[B]^{2k}$
 via c, 
 $p_{\bar \rho }(l,\beta ) = v$
. Similarly
$p_{\bar \rho }(l,\beta ) = v$
. Similarly 
 $\bar \delta $
 is in the same position relative to
$\bar \delta $
 is in the same position relative to 
 $\bar \rho $
 as to
$\bar \rho $
 as to 
 $\bar \gamma $
. Thus also
$\bar \gamma $
. Thus also 
 $p_{\bar \rho }(l,\beta ) = u$
 and we find that
$p_{\bar \rho }(l,\beta ) = u$
 and we find that 
 $v = u$
—we get a contradiction.⊣
$v = u$
—we get a contradiction.⊣
Claim 3.8. For any 
 $l<k$
 and
$l<k$
 and 
 $\bar j,\bar j' \in \prod _{l'<k} \{2l', 2l'+1 \}$
,
$\bar j,\bar j' \in \prod _{l'<k} \{2l', 2l'+1 \}$
, 
 $e_{l,\bar j, \bar j'}(m) = m$
 for every
$e_{l,\bar j, \bar j'}(m) = m$
 for every 
 $m \in \operatorname {\mathrm {dom}} e_{l,\bar j, \bar j'} $
.
$m \in \operatorname {\mathrm {dom}} e_{l,\bar j, \bar j'} $
.
Proof. Let 
 $\alpha _0 < \dots < \alpha _{2k} \in B$
 so that
$\alpha _0 < \dots < \alpha _{2k} \in B$
 so that 
 $(\alpha _{2l'},\alpha _{2l'+1}) \cap B \neq \emptyset $
 for every
$(\alpha _{2l'},\alpha _{2l'+1}) \cap B \neq \emptyset $
 for every 
 $l' <k$
. Consider
$l' <k$
. Consider 
 $\bar \gamma = (\alpha _{j_{l'}})_{l' < k}$
,
$\bar \gamma = (\alpha _{j_{l'}})_{l' < k}$
, 
 $\bar \delta = (\alpha _{j^{\prime }_{l'}})_{l' < k}$
 and again we find
$\bar \delta = (\alpha _{j^{\prime }_{l'}})_{l' < k}$
 and again we find 
 $\bar \rho \in [B]^k$
 so that
$\bar \rho \in [B]^k$
 so that 
 $\rho _{l'}$
 is between (possibly equal to)
$\rho _{l'}$
 is between (possibly equal to) 
 $\alpha _{j_{l'}}$
 and
$\alpha _{j_{l'}}$
 and 
 $\alpha _{j^{\prime }_{l'}}$
. If
$\alpha _{j^{\prime }_{l'}}$
. If 
 $e_{l,\bar j, \bar j'}(m) = m'$
, then if
$e_{l,\bar j, \bar j'}(m) = m'$
, then if 
 $\beta $
 is the m-th element of
$\beta $
 is the m-th element of 
 $H_l(\bar \gamma )$
, then
$H_l(\bar \gamma )$
, then 
 $\beta $
 is
$\beta $
 is 
 $m'$
-th element of
$m'$
-th element of 
 $H_l(\bar \delta )$
 as well as of
$H_l(\bar \delta )$
 as well as of 
 $H_l(\bar \rho )$
. But also
$H_l(\bar \rho )$
. But also 
 $\beta $
 is the m-th element of
$\beta $
 is the m-th element of 
 $H_l(\bar \rho )$
, thus
$H_l(\bar \rho )$
, thus 
 $m =m'$
.⊣
$m =m'$
.⊣
 Note that by the above claim 
 $e_{l,\bar j, \bar j'} = (e_{l,\bar j', \bar j})^{-1} = e_{l,\bar j', \bar j}$
 and the essential information given by
$e_{l,\bar j, \bar j'} = (e_{l,\bar j', \bar j})^{-1} = e_{l,\bar j', \bar j}$
 and the essential information given by 
 $e_{l,\bar j, \bar j'}$
 is its domain.
$e_{l,\bar j, \bar j'}$
 is its domain.
 Next let us introduce some notation. Whenever 
 $x,y \in 2^\omega $
, we write
$x,y \in 2^\omega $
, we write 
 $x < y$
 to say that x is lexicographically below y, i.e.,
$x < y$
 to say that x is lexicographically below y, i.e., 
 $x(n) < y(n)$
, where
$x(n) < y(n)$
, where 
 $n = \min \{m \in \omega : x(m) \neq x(m) \}$
. For any
$n = \min \{m \in \omega : x(m) \neq x(m) \}$
. For any 
 $g \in \{-1,0,1 \}^{k}$
 we naturally define a relation
$g \in \{-1,0,1 \}^{k}$
 we naturally define a relation 
 $\tilde R_g$
 between k-length sequences
$\tilde R_g$
 between k-length sequences 
 $\bar \nu $
 and
$\bar \nu $
 and 
 $\bar \mu $
, either of elements of
$\bar \mu $
, either of elements of 
 $2^\omega $
, or of ordinals
$2^\omega $
, or of ordinals 
 $<\kappa $
, as follows:
$<\kappa $
, as follows: 
 $$ \begin{align*}\bar \nu \tilde R_g \bar \mu \leftrightarrow \forall l<k \begin{cases} \nu_l < \mu_l, &\text{ if } g(l)=-1,\\ \nu_l = \mu_l, &\text{ if } g(l)=0,\\ \nu_l>\mu_l, &\text{ if } g(l)=1. \end{cases} \end{align*} $$
$$ \begin{align*}\bar \nu \tilde R_g \bar \mu \leftrightarrow \forall l<k \begin{cases} \nu_l < \mu_l, &\text{ if } g(l)=-1,\\ \nu_l = \mu_l, &\text{ if } g(l)=0,\\ \nu_l>\mu_l, &\text{ if } g(l)=1. \end{cases} \end{align*} $$
Further we write 
 $\bar \nu R_g \bar \mu $
 iff
$\bar \nu R_g \bar \mu $
 iff 
 $\bar \nu \tilde R_g \bar \mu $
 or
$\bar \nu \tilde R_g \bar \mu $
 or 
 $\bar \mu \tilde R_g \bar \nu $
. Enumerate
$\bar \mu \tilde R_g \bar \nu $
. Enumerate 
 $\{R_g : g \in \{-1,0,1 \}^{k} \}$
 without repetition as
$\{R_g : g \in \{-1,0,1 \}^{k} \}$
 without repetition as 
 $\langle R_i : i < K \rangle $
 (it is easy to see that
$\langle R_i : i < K \rangle $
 (it is easy to see that 
 $K = \frac {3^k +1}{2}$
). Note that for any
$K = \frac {3^k +1}{2}$
). Note that for any 
 $\bar \nu , \bar \mu $
 there is a unique
$\bar \nu , \bar \mu $
 there is a unique 
 $i <K$
 so that
$i <K$
 so that 
 $\bar \nu R_i \bar \mu $
. Now for each
$\bar \nu R_i \bar \mu $
. Now for each 
 $l < k$
 and
$l < k$
 and 
 $i < K$
, we let
$i < K$
, we let 
 $$ \begin{align*}I_{l,i} := \operatorname{\mathrm{dom}} e_{l,\bar j, \bar j'} \subseteq N_l,\end{align*} $$
$$ \begin{align*}I_{l,i} := \operatorname{\mathrm{dom}} e_{l,\bar j, \bar j'} \subseteq N_l,\end{align*} $$
where 
 $\bar j R_i \bar j'$
. By homogeneity of
$\bar j R_i \bar j'$
. By homogeneity of 
 $[B]^{2k}$
 and the observation that
$[B]^{2k}$
 and the observation that 
 $e_{l,\bar j, \bar j'} = e_{l,\bar j', \bar j}$
, we see that
$e_{l,\bar j, \bar j'} = e_{l,\bar j', \bar j}$
, we see that 
 $I_{l,i}$
 does not depend on the particular choice of
$I_{l,i}$
 does not depend on the particular choice of 
 $\bar j, \bar j'$
, such that
$\bar j, \bar j'$
, such that 
 $\bar j R_i \bar j'$
.
$\bar j R_i \bar j'$
.
 For each 
 $l< k$
 and
$l< k$
 and 
 $m < N_l$
, we define a relation
$m < N_l$
, we define a relation 
 $E_{l,m}$
 on
$E_{l,m}$
 on 
 $(2^\omega )^k$
 as follows:
$(2^\omega )^k$
 as follows: 
 $$ \begin{align*}\bar x E_{l,m} \bar y \leftrightarrow m \in I_{l,i} \text{ where } i \text{ is such that } \bar x R_i \bar y.\end{align*} $$
$$ \begin{align*}\bar x E_{l,m} \bar y \leftrightarrow m \in I_{l,i} \text{ where } i \text{ is such that } \bar x R_i \bar y.\end{align*} $$
Claim 3.9. 
 $E_{l,m}$
 is an equivalence relation.
$E_{l,m}$
 is an equivalence relation.
Proof. The reflexivity and symmetry of 
 $E_{l,m}$
 is obvious. Assume that
$E_{l,m}$
 is obvious. Assume that 
 $\bar x_0 E_{l,m} \bar x_1$
 and
$\bar x_0 E_{l,m} \bar x_1$
 and 
 $\bar x_1 E_{l,m} \bar x_2$
, and say
$\bar x_1 E_{l,m} \bar x_2$
, and say 
 $\bar x_0 R_{i_0} \bar x_1$
,
$\bar x_0 R_{i_0} \bar x_1$
, 
 $\bar x_1 R_{i_1} \bar x_2$
, and
$\bar x_1 R_{i_1} \bar x_2$
, and 
 $\bar x_0 R_{i_2} \bar x_2$
. Find
$\bar x_0 R_{i_2} \bar x_2$
. Find 
 $\bar \gamma ^0, \bar \gamma ^1, \bar \gamma ^2 \in [B]^k$
 so that
$\bar \gamma ^0, \bar \gamma ^1, \bar \gamma ^2 \in [B]^k$
 so that 
 $$ \begin{align*}\{\gamma^i_0 : i <3 \} < \dots < \{\gamma^i_{k-1} : i <3 \}\end{align*} $$
$$ \begin{align*}\{\gamma^i_0 : i <3 \} < \dots < \{\gamma^i_{k-1} : i <3 \}\end{align*} $$
and
 $$ \begin{align*}\bar \gamma^0 R_{i_0} \bar \gamma^1, \bar \gamma^1 R_{i_1} \bar \gamma^2, \bar \gamma^0 R_{i_2} \bar \gamma^2.\end{align*} $$
$$ \begin{align*}\bar \gamma^0 R_{i_0} \bar \gamma^1, \bar \gamma^1 R_{i_1} \bar \gamma^2, \bar \gamma^0 R_{i_2} \bar \gamma^2.\end{align*} $$
 If 
 $\beta $
 is the m-th element of
$\beta $
 is the m-th element of 
 $H_l(\bar \gamma ^0)$
, then
$H_l(\bar \gamma ^0)$
, then 
 $\beta $
 is also the m-th element of
$\beta $
 is also the m-th element of 
 $H_l(\bar \gamma ^1)$
, since we can find an appropriate
$H_l(\bar \gamma ^1)$
, since we can find an appropriate 
 $\bar \alpha \in [B]^{2k}$
 and
$\bar \alpha \in [B]^{2k}$
 and 
 $\bar j$
,
$\bar j$
, 
 $\bar j'$
 so that
$\bar j'$
 so that 
 $\bar \gamma ^0 = (\alpha _{j_l})_{l<k}$
 and
$\bar \gamma ^0 = (\alpha _{j_l})_{l<k}$
 and 
 $\bar \gamma ^1 = (\alpha _{j^{\prime }_l})_{l<k}$
,
$\bar \gamma ^1 = (\alpha _{j^{\prime }_l})_{l<k}$
, 
 $\bar j R_{i_0} \bar j'$
 and we have that
$\bar j R_{i_0} \bar j'$
 and we have that 
 $m \in I_{l,i_0}$
. Similarly
$m \in I_{l,i_0}$
. Similarly 
 $\beta $
 is the m-th element of
$\beta $
 is the m-th element of 
 $H_l(\bar \gamma ^2)$
.
$H_l(\bar \gamma ^2)$
.
 But now we find again 
 $\bar \alpha \in [B]^{2k}$
 and
$\bar \alpha \in [B]^{2k}$
 and 
 $\bar j$
,
$\bar j$
, 
 $\bar j'$
 so that
$\bar j'$
 so that 
 $\bar \gamma ^0 = (\alpha _{j_l})_{l<k}$
 and
$\bar \gamma ^0 = (\alpha _{j_l})_{l<k}$
 and 
 $\bar \gamma ^2 = (\alpha _{j^{\prime }_l})_{l<k}$
. Thus
$\bar \gamma ^2 = (\alpha _{j^{\prime }_l})_{l<k}$
. Thus 
 $m \in I_{l,i_2}$
, as
$m \in I_{l,i_2}$
, as 
 $e_{l,\bar j, \bar j'}(m) = m$
 and
$e_{l,\bar j, \bar j'}(m) = m$
 and 
 $\bar x_0 E_{l,m} \bar x_2$
.⊣
$\bar x_0 E_{l,m} \bar x_2$
.⊣
Claim 3.10. 
 $E_{l,m}$
 is smooth as witnessed by a continuous function, i.e., there is a continuous map
$E_{l,m}$
 is smooth as witnessed by a continuous function, i.e., there is a continuous map 
 $\varphi _{l,m} \colon (2^{\omega })^k \to 2^{\omega }$
 so that
$\varphi _{l,m} \colon (2^{\omega })^k \to 2^{\omega }$
 so that 
 $\bar x E_{l,m} \bar y$
 iff
$\bar x E_{l,m} \bar y$
 iff 
 $\varphi _{l,m}(\bar x) = \varphi _{l,m}(\bar y)$
.
$\varphi _{l,m}(\bar x) = \varphi _{l,m}(\bar y)$
.
Proof. We will check the following:
- 
(a) For every open  $O \subseteq (2^\omega )^k$
, the $O \subseteq (2^\omega )^k$
, the $E_{l,m}$
 saturation of O is Borel, $E_{l,m}$
 saturation of O is Borel,
- 
(b) every  $E_{l,m}$
 equivalence class is $E_{l,m}$
 equivalence class is $G_\delta $
. $G_\delta $
.
 By a theorem of Srivastava [Reference Srivastava31, Theorem 4.1.], (a) and (b) imply that 
 $E_{l,m}$
 is smooth, i.e., we can find
$E_{l,m}$
 is smooth, i.e., we can find 
 $\varphi _{l,m}$
 Borel.
$\varphi _{l,m}$
 Borel.
- 
(a) The  $E_{l,m}$
 saturation of O is the set $E_{l,m}$
 saturation of O is the set $\{\bar x : \exists \bar y \in O (\bar x E_{l,m} \bar y) \}$
. It suffices to check for each $\{\bar x : \exists \bar y \in O (\bar x E_{l,m} \bar y) \}$
. It suffices to check for each $g \in \{-1,0,1\}^k$
 that the set $g \in \{-1,0,1\}^k$
 that the set $X = \{\bar x : \exists \bar y \in O (\bar x \tilde R_g \bar y) \}$
 is Borel. Let $X = \{\bar x : \exists \bar y \in O (\bar x \tilde R_g \bar y) \}$
 is Borel. Let $\mathcal {S} = \{\bar \sigma \in (2^{<\omega })^k : [\sigma _0] \times \dots \times [\sigma _{k-1}] \subseteq O \}$
. Consider If $\mathcal {S} = \{\bar \sigma \in (2^{<\omega })^k : [\sigma _0] \times \dots \times [\sigma _{k-1}] \subseteq O \}$
. Consider If $$ \begin{align*}\varphi(\bar x) :\leftrightarrow \exists \bar \sigma \in \mathcal{S} \forall l' < k \begin{cases} x_{l'} <_{\operatorname{\mathrm{lex}}} {\sigma_{l'}}^{\frown} 0^\omega, &\text{ if } g(l') = {-1}, \\ x_{l'} \in [\sigma_{l'}], &\text{ if } g(l') = 0, \\ {\sigma_{l'}}^{\frown} 1^\omega <_{\operatorname{\mathrm{lex}}} x_{l'}, &\text{ if } g(l') = 1.\end{cases}\end{align*} $$ $$ \begin{align*}\varphi(\bar x) :\leftrightarrow \exists \bar \sigma \in \mathcal{S} \forall l' < k \begin{cases} x_{l'} <_{\operatorname{\mathrm{lex}}} {\sigma_{l'}}^{\frown} 0^\omega, &\text{ if } g(l') = {-1}, \\ x_{l'} \in [\sigma_{l'}], &\text{ if } g(l') = 0, \\ {\sigma_{l'}}^{\frown} 1^\omega <_{\operatorname{\mathrm{lex}}} x_{l'}, &\text{ if } g(l') = 1.\end{cases}\end{align*} $$ $\varphi (\bar x)$
 holds true then let $\varphi (\bar x)$
 holds true then let $\bar \sigma $
 witness this. We then see that there is $\bar \sigma $
 witness this. We then see that there is $\bar y \in [\sigma _0] \times \dots \times [\sigma _{k-1}]$
 with $\bar y \in [\sigma _0] \times \dots \times [\sigma _{k-1}]$
 with $\bar x \tilde R_g \bar y$
. On the other hand, if $\bar x \tilde R_g \bar y$
. On the other hand, if $\bar y \in O$
 is such that $\bar y \in O$
 is such that $\bar x R_g \bar y$
, then we find $\bar x R_g \bar y$
, then we find $\bar \sigma \in \mathcal {S}$
 defining a neighborhood of $\bar \sigma \in \mathcal {S}$
 defining a neighborhood of $\bar y$
 witnessing $\bar y$
 witnessing $\varphi (\bar x)$
. Thus X is defined by $\varphi (\bar x)$
. Thus X is defined by $\varphi $
 and is thus Borel. $\varphi $
 and is thus Borel.
- 
(b) Since finite unions of  $G_\delta $
’s are $G_\delta $
’s are $G_\delta $
 it suffices to check that $G_\delta $
 it suffices to check that $\{ \bar x : \bar x \tilde R_g \bar y \}$
 is $\{ \bar x : \bar x \tilde R_g \bar y \}$
 is $G_\delta $
 for every $G_\delta $
 for every $\bar y$
 and $\bar y$
 and $g \in \{-1,0,1\}^k$
. But this is obvious from the definition. $g \in \{-1,0,1\}^k$
. But this is obvious from the definition.
 Now note that given 
 $\varphi _{l,m}$
 Borel, we can find perfect
$\varphi _{l,m}$
 Borel, we can find perfect 
 $X_0,\dots ,X_{k-1} \subseteq 2^{\omega }$
 so that
$X_0,\dots ,X_{k-1} \subseteq 2^{\omega }$
 so that 
 $\varphi _{l,m}$
 is continuous on
$\varphi _{l,m}$
 is continuous on 
 $X_0 \times \dots \times X_{k-1}$
 (
$X_0 \times \dots \times X_{k-1}$
 (
 $\varphi _{l,m}$
 is continuous on a dense
$\varphi _{l,m}$
 is continuous on a dense 
 $G_\delta $
). But there is a
$G_\delta $
). But there is a 
 $<_{\operatorname {\mathrm {lex}}}$
 preserving homeomorphism from
$<_{\operatorname {\mathrm {lex}}}$
 preserving homeomorphism from 
 $X_l$
 to
$X_l$
 to 
 $2^{\omega }$
 for each
$2^{\omega }$
 for each 
 $l<k$
 so we may simply assume
$l<k$
 so we may simply assume 
 $X_l = 2^\omega $
.⊣
$X_l = 2^\omega $
.⊣
 Fix such 
 $\varphi _{l,m}$
 for every
$\varphi _{l,m}$
 for every 
 $l<k$
,
$l<k$
, 
 $m<N_l$
, so that
$m<N_l$
, so that 
 $\varphi _{l,a_l}(\bar x) = x_l$
 (note that
$\varphi _{l,a_l}(\bar x) = x_l$
 (note that 
 $\bar x E_{l,a_l} \bar y$
 iff
$\bar x E_{l,a_l} \bar y$
 iff 
 $x_l = y_l$
). Now let
$x_l = y_l$
). Now let 
 $M_0$
 be countable elementary, containing all relevant information and such that
$M_0$
 be countable elementary, containing all relevant information and such that 
 $\varphi _{l,m} \in M_0$
 for every
$\varphi _{l,m} \in M_0$
 for every 
 $l < k$
,
$l < k$
, 
 $m < N_l$
. Let
$m < N_l$
. Let 
 $\chi _{l,m} \colon 2^\omega \to [r(l,m)]$
 for
$\chi _{l,m} \colon 2^\omega \to [r(l,m)]$
 for 
 $l < k$
 and
$l < k$
 and 
 $m \neq a_l$
 be generic continuous functions over
$m \neq a_l$
 be generic continuous functions over 
 $M_0$
, i.e., the sequence
$M_0$
, i.e., the sequence 
 $(\chi _{l,m})_{l < k, m \in N_l \setminus \{a_l\} }$
 is
$(\chi _{l,m})_{l < k, m \in N_l \setminus \{a_l\} }$
 is 
 $\prod _{l < k, m \in N_l \setminus \{a_l\}} \mathbb {C}(2^\omega , [r(l,m)])$
 generic over
$\prod _{l < k, m \in N_l \setminus \{a_l\}} \mathbb {C}(2^\omega , [r(l,m)])$
 generic over 
 $M_0$
. Let us denote with M the generic extension of
$M_0$
. Let us denote with M the generic extension of 
 $M_0$
. Also let
$M_0$
. Also let 
 $\chi _{l,m}$
 for
$\chi _{l,m}$
 for 
 $m = a_l$
 be the identity. Finally we set
$m = a_l$
 be the identity. Finally we set 
 $$ \begin{align*}\phi_i(\bar x) = ((\chi_{l,b^i_l}\circ \varphi_{l,b^i_l})(\bar x))_{l < k}\end{align*} $$
$$ \begin{align*}\phi_i(\bar x) = ((\chi_{l,b^i_l}\circ \varphi_{l,b^i_l})(\bar x))_{l < k}\end{align*} $$
for each 
 $i < N$
.
$i < N$
.
Claim 3.11. (2) of the main lemma holds true with M, 
 $\bar s$
, and
$\bar s$
, and 
 $\phi _i$
,
$\phi _i$
, 
 $i <N$
, that we just defined.
$i <N$
, that we just defined.
Proof. Let 
 $\bar x_0, \dots , \bar x_{n-1} \in [\bar s]$
 be
$\bar x_0, \dots , \bar x_{n-1} \in [\bar s]$
 be 
 $\langle 2^\omega : l < k\rangle $
-mCg over M. Let us write
$\langle 2^\omega : l < k\rangle $
-mCg over M. Let us write 
 $\{ \bar x_i(l) : i <n \} = \{ y_l^i : i < K_l \}$
 for every
$\{ \bar x_i(l) : i <n \} = \{ y_l^i : i < K_l \}$
 for every 
 $l < k$
, where
$l < k$
, where 
 $y_l^0 <_{\operatorname {\mathrm {lex}}} \dots <_{\operatorname {\mathrm {lex}}} y_l^{K_l -1} $
. Now find
$y_l^0 <_{\operatorname {\mathrm {lex}}} \dots <_{\operatorname {\mathrm {lex}}} y_l^{K_l -1} $
. Now find 
 $$ \begin{align*}\alpha^0_0 < \dots < \alpha_0^{K_0 -1} < \dots < \alpha_{k-1}^0 < \dots < \alpha_{k-1}^{K_{k-1}-1}\end{align*} $$
$$ \begin{align*}\alpha^0_0 < \dots < \alpha_0^{K_0 -1} < \dots < \alpha_{k-1}^0 < \dots < \alpha_{k-1}^{K_{k-1}-1}\end{align*} $$
in 
 $B \cap M$
. For every
$B \cap M$
. For every 
 $\bar j \in \prod _{l < k} K_l$
, define
$\bar j \in \prod _{l < k} K_l$
, define 
 $\bar y_{\bar j} := (y_0^{j(0)}, \dots , y_{k-1}^{j(k-1)})$
 and
$\bar y_{\bar j} := (y_0^{j(0)}, \dots , y_{k-1}^{j(k-1)})$
 and 
 $\bar \alpha _{\bar j} := (\alpha _0^{j(0)}, \dots , \alpha _{k-1}^{j(k-1)})$
. Then, for each
$\bar \alpha _{\bar j} := (\alpha _0^{j(0)}, \dots , \alpha _{k-1}^{j(k-1)})$
. Then, for each 
 $i < n$
, we have
$i < n$
, we have 
 $\bar j_i \in \prod _{l < k} K_l$
 so that
$\bar j_i \in \prod _{l < k} K_l$
 so that 
 $\bar x_i = \bar y_{\bar j_i}$
. For each
$\bar x_i = \bar y_{\bar j_i}$
. For each 
 $i < n$
 define the function
$i < n$
 define the function 
 $g_i \colon \bigcup _{l < k} \{ l\} \times H_l(\bar \alpha _{\bar j_i}) \to 2^\omega $
, setting
$g_i \colon \bigcup _{l < k} \{ l\} \times H_l(\bar \alpha _{\bar j_i}) \to 2^\omega $
, setting 
 $$ \begin{align*}g_i(l,\beta) = \chi_{l,m}(\varphi_{l,m}(\bar x_{i})),\end{align*} $$
$$ \begin{align*}g_i(l,\beta) = \chi_{l,m}(\varphi_{l,m}(\bar x_{i})),\end{align*} $$
whenever 
 $\beta $
 is the m-th element of
$\beta $
 is the m-th element of 
 $H_l(\bar \alpha _{\bar j_i})$
.
$H_l(\bar \alpha _{\bar j_i})$
.
 Now we have that the 
 $g_i$
 agree on their common domain. Namely let
$g_i$
 agree on their common domain. Namely let 
 $i_0, i_1 < n$
 and
$i_0, i_1 < n$
 and 
 $(l,\beta ) \in \operatorname {\mathrm {dom}} g_{i_0} \cap \operatorname {\mathrm {dom}} g_{i_1}$
. Then if we set i to be so that
$(l,\beta ) \in \operatorname {\mathrm {dom}} g_{i_0} \cap \operatorname {\mathrm {dom}} g_{i_1}$
. Then if we set i to be so that 
 $\bar x_{i_0} R_i \bar x_{i_1}$
, we have that
$\bar x_{i_0} R_i \bar x_{i_1}$
, we have that 
 $m \in I_{l,i}$
, where
$m \in I_{l,i}$
, where 
 $\beta $
 is the m-th element of
$\beta $
 is the m-th element of 
 $H_{l}(\bar \alpha _{\bar j_{i_0}})$
 and of
$H_{l}(\bar \alpha _{\bar j_{i_0}})$
 and of 
 $H_{l}(\bar \alpha _{\bar j_{i_1}})$
. In particular
$H_{l}(\bar \alpha _{\bar j_{i_1}})$
. In particular 
 $\bar x_{i_0} E_{l,m} \bar x_{i_1}$
 and
$\bar x_{i_0} E_{l,m} \bar x_{i_1}$
 and 
 $\varphi _{l,m}(\bar x_{i_0}) = \varphi _{l,m}(\bar x_{i_1})$
 and thus
$\varphi _{l,m}(\bar x_{i_0}) = \varphi _{l,m}(\bar x_{i_1})$
 and thus 
 $$ \begin{align*}g_{i_0}(l,\beta) = \chi_{l,m}(\varphi_{l,m}(\bar x_{i_0})) = \chi_{l,m}(\varphi_{l,m}(\bar x_{i_1})) = g_{i_1}(l,\beta).\end{align*} $$
$$ \begin{align*}g_{i_0}(l,\beta) = \chi_{l,m}(\varphi_{l,m}(\bar x_{i_0})) = \chi_{l,m}(\varphi_{l,m}(\bar x_{i_1})) = g_{i_1}(l,\beta).\end{align*} $$
 Let 
 $g := \bigcup _{i < n} g_i$
. Then we see by Lemma 3.3 that g is Cohen generic in
$g := \bigcup _{i < n} g_i$
. Then we see by Lemma 3.3 that g is Cohen generic in 
 $\prod _{(l,\beta ) \in \operatorname {\mathrm {dom}} g} 2^\omega $
 over
$\prod _{(l,\beta ) \in \operatorname {\mathrm {dom}} g} 2^\omega $
 over 
 $M_0$
. Namely consider
$M_0$
. Namely consider 
 $K = \sum _{l < k} K_l$
 and
$K = \sum _{l < k} K_l$
 and 
 $(y^0_0, \dots , y_{k-1}^{K_{k-1}-1})$
 as a
$(y^0_0, \dots , y_{k-1}^{K_{k-1}-1})$
 as a 
 $(2^{<\omega })^K$
-generic over M. Then, if
$(2^{<\omega })^K$
-generic over M. Then, if 
 $\langle u_i : i < n'\rangle $
 enumerates
$\langle u_i : i < n'\rangle $
 enumerates 
 $\{ \varphi _{l,m}(\bar x_i) : i < n, l<k, m < N_l \}$
, we have that every value of g is contained in
$\{ \varphi _{l,m}(\bar x_i) : i < n, l<k, m < N_l \}$
, we have that every value of g is contained in 
 $\{ \chi _{l,m}(u_i) : i < n', l <k, m<N_l \}$
. Also note that by construction for every
$\{ \chi _{l,m}(u_i) : i < n', l <k, m<N_l \}$
. Also note that by construction for every 
 $i < n$
,
$i < n$
, 
 $p_{\bar \alpha _{\bar j_i}} \restriction \operatorname {\mathrm {dom}} g$
 is in the generic filter defined by g. Since
$p_{\bar \alpha _{\bar j_i}} \restriction \operatorname {\mathrm {dom}} g$
 is in the generic filter defined by g. Since 
 $\{ p_{\bar \alpha _{\bar j_i}} : i < n \}$
 is centered we can extend the generic filter of g to a
$\{ p_{\bar \alpha _{\bar j_i}} : i < n \}$
 is centered we can extend the generic filter of g to a 
 $\mathbb {Q}$
-generic G over
$\mathbb {Q}$
-generic G over 
 $M_0$
 so that
$M_0$
 so that 
 $p_{\bar \alpha _{j_i}} \in G$
 for every
$p_{\bar \alpha _{j_i}} \in G$
 for every 
 $i < n$
.
$i < n$
.
Now we have that
 $$ \begin{align*}\bar z_{\bar \alpha_{\bar j_i}}[G] = \bar x_i \text{ and } \bar z_{\bar \beta^{j}(\bar \alpha_{\bar j_i})}[G] = \phi_j(\bar x_{i})\end{align*} $$
$$ \begin{align*}\bar z_{\bar \alpha_{\bar j_i}}[G] = \bar x_i \text{ and } \bar z_{\bar \beta^{j}(\bar \alpha_{\bar j_i})}[G] = \phi_j(\bar x_{i})\end{align*} $$
for every 
 $i < n$
 and
$i < n$
 and 
 $j < N$
. Thus we get that
$j < N$
. Thus we get that 
 $$ \begin{align*}M_0[G] \models \bigcup_{i < n} \{ \phi_j(\bar x_i) : j < N \} \subseteq \dot{\mathcal{A}}[G] \wedge \{ \bar x_0 \} \cup \{\phi_j(\bar x_0) : j < N \} \in E.\end{align*} $$
$$ \begin{align*}M_0[G] \models \bigcup_{i < n} \{ \phi_j(\bar x_i) : j < N \} \subseteq \dot{\mathcal{A}}[G] \wedge \{ \bar x_0 \} \cup \{\phi_j(\bar x_0) : j < N \} \in E.\end{align*} $$
Again, by absoluteness, we get the required result.⊣
3.3 Infinite products
Definition 3.12. Let 
 $\langle X_i : i < \alpha \rangle \in M$
 be Polish spaces indexed by a countable ordinal
$\langle X_i : i < \alpha \rangle \in M$
 be Polish spaces indexed by a countable ordinal 
 $\alpha $
. Then we say that
$\alpha $
. Then we say that 
 $\bar x_0, \dots , \bar x_{n-1} \in \prod _{i<\alpha } X_i$
 are
$\bar x_0, \dots , \bar x_{n-1} \in \prod _{i<\alpha } X_i$
 are 
 $\langle X_i : i < \alpha \rangle $
-mutually Cohen generic (mCg) over M if there are
$\langle X_i : i < \alpha \rangle $
-mutually Cohen generic (mCg) over M if there are 
 $\xi _0 = 0 < \dots < \xi _k = \alpha $
 for some
$\xi _0 = 0 < \dots < \xi _k = \alpha $
 for some 
 $k \in \omega $
 so that
$k \in \omega $
 so that 
 $$ \begin{align*}\bar x_0, \dots, \bar x_{n-1} \text{ are } \langle Y_l : l < k\rangle\text{-mutually Cohen generic over } M,\end{align*} $$
$$ \begin{align*}\bar x_0, \dots, \bar x_{n-1} \text{ are } \langle Y_l : l < k\rangle\text{-mutually Cohen generic over } M,\end{align*} $$
where 
 $Y_l = \prod _{i \in [\xi _l, \xi _{l+1})} X_i$
 for every
$Y_l = \prod _{i \in [\xi _l, \xi _{l+1})} X_i$
 for every 
 $l <k$
 and we identify
$l <k$
 and we identify 
 $\bar x_i$
 with
$\bar x_i$
 with 
 $$ \begin{align*}(\bar x_i \restriction [\xi_0, \xi_1), \dots, \bar x_i \restriction [\xi_{k-1}, \xi_k)) \in \prod_{l <k} Y_l ,\end{align*} $$
$$ \begin{align*}(\bar x_i \restriction [\xi_0, \xi_1), \dots, \bar x_i \restriction [\xi_{k-1}, \xi_k)) \in \prod_{l <k} Y_l ,\end{align*} $$
for every 
 $i < n$
.
$i < n$
.
The identification of a sequence 
 $\bar x$
 with
$\bar x$
 with 
 $(\bar x \restriction [\xi _0, \xi _1), \dots , \bar x \restriction [\xi _{k-1}, \xi _k))$
 for a given
$(\bar x \restriction [\xi _0, \xi _1), \dots , \bar x \restriction [\xi _{k-1}, \xi _k))$
 for a given 
 $\langle \xi _l : l < k\rangle $
 will be implicitly made throughout the rest of the paper in order to reduce the notational load.
$\langle \xi _l : l < k\rangle $
 will be implicitly made throughout the rest of the paper in order to reduce the notational load.
 Note that whenever 
 $\bar x_0, \dots , \bar x_{n-1}$
 are
$\bar x_0, \dots , \bar x_{n-1}$
 are 
 $\langle X_i : i < \alpha \rangle $
-mCg over M and
$\langle X_i : i < \alpha \rangle $
-mCg over M and 
 $\beta \leq \alpha $
, then
$\beta \leq \alpha $
, then 
 ${\bar x_0 \restriction \beta }, \dots , {\bar x_{n-1} \restriction \beta }$
 are
${\bar x_0 \restriction \beta }, \dots , {\bar x_{n-1} \restriction \beta }$
 are 
 $\langle X_i : i < \beta \rangle $
-mCg over M. Also note that Definition 3.12 agrees with the notion of mCg for finite
$\langle X_i : i < \beta \rangle $
-mCg over M. Also note that Definition 3.12 agrees with the notion of mCg for finite 
 $\alpha $
.
$\alpha $
.
Definition 3.13. We say that 
 $\bar x_0, \dots , \bar x_{n-1} \in \prod _{i<\alpha } X_i$
 are strongly
$\bar x_0, \dots , \bar x_{n-1} \in \prod _{i<\alpha } X_i$
 are strongly 
 $\langle X_i : i < \alpha \rangle $
-mCg over M if they are
$\langle X_i : i < \alpha \rangle $
-mCg over M if they are 
 $\langle X_i : i < \alpha \rangle $
-mCg over M and for any
$\langle X_i : i < \alpha \rangle $
-mCg over M and for any 
 $i,j < n$
, if
$i,j < n$
, if 
 $\xi = \min \{ \beta < \alpha : x_i(\beta ) \neq x_j(\beta ) \}$
, then
$\xi = \min \{ \beta < \alpha : x_i(\beta ) \neq x_j(\beta ) \}$
, then 
 $x_i(\beta ) \neq x_j(\beta )$
 for all
$x_i(\beta ) \neq x_j(\beta )$
 for all 
 $\beta \geq \xi $
.
$\beta \geq \xi $
.
Main Lemma 3.14. Let 
 $\alpha < \omega _1$
 and
$\alpha < \omega _1$
 and 
 $E \subseteq [(2^\omega )^\alpha ]^{<\omega }$
 be an analytic hypergraph. Then there is a countable model M,
$E \subseteq [(2^\omega )^\alpha ]^{<\omega }$
 be an analytic hypergraph. Then there is a countable model M, 
 $\alpha +1 \subseteq M$
, so that either
$\alpha +1 \subseteq M$
, so that either
- 
(1) for any  $n \in \omega $
 and $n \in \omega $
 and $\bar x_0,\dots , \bar x_{n-1} \in (2^\omega )^\alpha $
 that are strongly $\bar x_0,\dots , \bar x_{n-1} \in (2^\omega )^\alpha $
 that are strongly $\langle 2^\omega : i < \alpha \rangle $
-mCg over M, $\langle 2^\omega : i < \alpha \rangle $
-mCg over M, $$ \begin{align*}\{\bar x_0,\dots, \bar x_{n-1}\} \text{ is } E \text{-independent}\end{align*} $$ $$ \begin{align*}\{\bar x_0,\dots, \bar x_{n-1}\} \text{ is } E \text{-independent}\end{align*} $$
 or for some 
 $N\in \omega $
,
$N\in \omega $
, 
- 
(2) there are  $\phi _0, \dots , \phi _{N-1} \colon (2^\omega )^\alpha \to (2^\omega )^\alpha $
 continuous, $\phi _0, \dots , \phi _{N-1} \colon (2^\omega )^\alpha \to (2^\omega )^\alpha $
 continuous, $\bar s \in \bigotimes _{i<\alpha } 2^{<\omega }$
 so that for any $\bar s \in \bigotimes _{i<\alpha } 2^{<\omega }$
 so that for any $n \in \omega $
 and $n \in \omega $
 and $\bar x_0,\dots , \bar x_{n-1} \in (2^\omega )^\alpha \cap [\bar s]$
 that are strongly mCg over M, $\bar x_0,\dots , \bar x_{n-1} \in (2^\omega )^\alpha \cap [\bar s]$
 that are strongly mCg over M, $$ \begin{align*}\{\phi_j(\bar x_i) : j < N, i<n\} \text{ is } E \text{-independent but } \{ \bar x_0\} \cup \{ \phi_j(\bar x_0) : j < N \} \in E.\end{align*} $$ $$ \begin{align*}\{\phi_j(\bar x_i) : j < N, i<n\} \text{ is } E \text{-independent but } \{ \bar x_0\} \cup \{ \phi_j(\bar x_0) : j < N \} \in E.\end{align*} $$
Proof. We are going to show something slightly stronger. Let R be an analytic hypergraph on 
 $(2^\omega )^\alpha \times \omega $
, M a countable model with
$(2^\omega )^\alpha \times \omega $
, M a countable model with 
 $R \in M, \alpha +1 \subseteq M$
, and
$R \in M, \alpha +1 \subseteq M$
, and 
 $k \in \omega $
. Then consider the following two statements.
$k \in \omega $
. Then consider the following two statements. 
- 
(1) R, M, k : For any pairwise distinct  $\bar x_0, \dots , \bar x_{n-1}$
 that are strongly $\bar x_0, \dots , \bar x_{n-1}$
 that are strongly $\langle 2^\omega : i < \alpha \rangle $
-mCg over M, and any $\langle 2^\omega : i < \alpha \rangle $
-mCg over M, and any $k_0, \dots , k_{n-1} < k$
, $k_0, \dots , k_{n-1} < k$
, $$ \begin{align*}\{(\bar x_0{},k_0),\dots, (\bar x_{n-1}, k_{n-1})\} \text{ is } R \text{-independent}.\end{align*} $$ $$ \begin{align*}\{(\bar x_0{},k_0),\dots, (\bar x_{n-1}, k_{n-1})\} \text{ is } R \text{-independent}.\end{align*} $$
- 
(2) R, M, k : There is  $N \in \omega $
, there are $N \in \omega $
, there are $\phi _0, \dots , \phi _{N-1} \colon (2^\omega )^\alpha \to (2^\omega )^\alpha $
 continuous, such that for every $\phi _0, \dots , \phi _{N-1} \colon (2^\omega )^\alpha \to (2^\omega )^\alpha $
 continuous, such that for every $\bar x \in (2^\omega )^\alpha $
 and $\bar x \in (2^\omega )^\alpha $
 and $j_0 < j_1 < N$
, $j_0 < j_1 < N$
, $\phi _{j_0}(\bar x) \neq \phi _{j_1}(\bar x)$
 and $\phi _{j_0}(\bar x) \neq \phi _{j_1}(\bar x)$
 and $\phi _{j_0}(\bar x) \neq \bar x$
, there are $\phi _{j_0}(\bar x) \neq \bar x$
, there are $k_0, \dots , k_{N-1} \leq k$
 and $k_0, \dots , k_{N-1} \leq k$
 and $\bar s \in \bigotimes _{i<\alpha } 2^{<\omega }$
, so that for any pairwise distinct $\bar s \in \bigotimes _{i<\alpha } 2^{<\omega }$
, so that for any pairwise distinct $\bar x_0,\dots , \bar x_{n-1} \in (2^\omega )^\alpha \cap [\bar s]$
 that are strongly $\bar x_0,\dots , \bar x_{n-1} \in (2^\omega )^\alpha \cap [\bar s]$
 that are strongly $\langle 2^\omega : i < \alpha \rangle $
-mCg over M, $\langle 2^\omega : i < \alpha \rangle $
-mCg over M, $$ \begin{align*}\{(\phi_j(\bar x_i),k_{j}) : j < N, i<n\} \text{ is } R \text{-independent, but}\end{align*} $$
In fact, if $$ \begin{align*}\{(\phi_j(\bar x_i),k_{j}) : j < N, i<n\} \text{ is } R \text{-independent, but}\end{align*} $$
In fact, if $$ \begin{align*}\{ (\bar x_0, k)\} \cup \{(\phi_j(\bar x_0), k_{j}) : j < N \} \in R.\end{align*} $$ $$ \begin{align*}\{ (\bar x_0, k)\} \cup \{(\phi_j(\bar x_0), k_{j}) : j < N \} \in R.\end{align*} $$ $k> 0$
, $k> 0$
, $$ \begin{align*}\{ (\bar x_i, {k-1}) : i < n \} \cup \{(\phi_j(\bar x_i), k_{j}) : j < N, i<n\} \text{ is } R \text{-independent.}\end{align*} $$ $$ \begin{align*}\{ (\bar x_i, {k-1}) : i < n \} \cup \{(\phi_j(\bar x_i), k_{j}) : j < N, i<n\} \text{ is } R \text{-independent.}\end{align*} $$
 We are going to show by induction on 
 $\alpha $
 that for any
$\alpha $
 that for any 
 $R, M, k$
,
$R, M, k$
, 
 $(1)_{R,M,k}$
 implies that either
$(1)_{R,M,k}$
 implies that either 
 $(1)_{R,M,k+1}$
 or there is a countable model
$(1)_{R,M,k+1}$
 or there is a countable model 
 $M^+ \supseteq M$
 so that
$M^+ \supseteq M$
 so that 
 $(2)_{R,M^+,k}$
. From this we easily follow the statement of the main lemma. Namely, whenever E is a hypergraph on
$(2)_{R,M^+,k}$
. From this we easily follow the statement of the main lemma. Namely, whenever E is a hypergraph on 
 $(2^\omega )^\alpha $
, consider the hypergraph R on
$(2^\omega )^\alpha $
, consider the hypergraph R on 
 $(2^\omega )^\alpha \times \omega $
 where
$(2^\omega )^\alpha \times \omega $
 where 
 $\{ (\bar x_0, k_0), \dots , (\bar x_{n-1}, k_{n-1}) \} \in R$
 iff
$\{ (\bar x_0, k_0), \dots , (\bar x_{n-1}, k_{n-1}) \} \in R$
 iff 
 $\{ \bar x_0, \dots , \bar x_{n-1} \} \in R$
. Then, if M is an arbitrary countable elementary model with
$\{ \bar x_0, \dots , \bar x_{n-1} \} \in R$
. Then, if M is an arbitrary countable elementary model with 
 $R,\alpha \in M$
 and if
$R,\alpha \in M$
 and if 
 $k=0$
,
$k=0$
, 
 $(1)_{R,M,k}$
 holds vacuously true. Applying the claim we find
$(1)_{R,M,k}$
 holds vacuously true. Applying the claim we find 
 $M^+$
 so that either
$M^+$
 so that either 
 $(1)_{R,M,1}$
 or
$(1)_{R,M,1}$
 or 
 $(2)_{R,M^+,0}$
. The two options easily translate to the conclusion of the main lemma.
$(2)_{R,M^+,0}$
. The two options easily translate to the conclusion of the main lemma.
 Let us first consider the successor step. Assume that 
 $\alpha = \beta +1$
, R is an analytic hypergraph on
$\alpha = \beta +1$
, R is an analytic hypergraph on 
 $(2^\omega )^\alpha \times \omega $
, and M a countable model with
$(2^\omega )^\alpha \times \omega $
, and M a countable model with 
 $R \in M, \alpha +1 \subseteq M$
 so that
$R \in M, \alpha +1 \subseteq M$
 so that 
 $(1)_{R,M,k}$
 holds true for some given
$(1)_{R,M,k}$
 holds true for some given 
 $k \in \omega $
. Let
$k \in \omega $
. Let 
 $\mathbb {Q}$
 be the forcing adding mutual Cohen reals
$\mathbb {Q}$
 be the forcing adding mutual Cohen reals 
 $\langle z_{0,i,j}, z_{1,i,j} : i, j \in \omega \rangle $
 in
$\langle z_{0,i,j}, z_{1,i,j} : i, j \in \omega \rangle $
 in 
 $2^\omega $
. Then we define the hypergraph
$2^\omega $
. Then we define the hypergraph 
 $\tilde R$
 on
$\tilde R$
 on 
 $(2^\omega )^\beta \times \omega $
 where
$(2^\omega )^\beta \times \omega $
 where 
 $\{(\bar y_0, m_0), \dots , (\bar y_{n-1}, m_{n-1}) \} \in \tilde R \cap [(2^\omega )^\beta \times \omega ]^n$
 iff there is
$\{(\bar y_0, m_0), \dots , (\bar y_{n-1}, m_{n-1}) \} \in \tilde R \cap [(2^\omega )^\beta \times \omega ]^n$
 iff there is 
 $p \in \mathbb {Q}$
 and there are
$p \in \mathbb {Q}$
 and there are 
 $K_i \in \omega $
,
$K_i \in \omega $
, 
 $k_{i,0}, \dots , k_{i,K_i -1} < k$
 for every
$k_{i,0}, \dots , k_{i,K_i -1} < k$
 for every 
 $i < n$
, so that
$i < n$
, so that 
 $$ \begin{align*}p \Vdash_{\mathbb{Q}} \bigcup_{i<n} \{ (\bar y_i{}^{\frown} \dot z_{0,i,j}, k_{i,j}) : j < K_i \} \cup \{ (\bar y_i{}^{\frown} \dot z_{1,i,j}, k) : j < m_i \} \in R.\end{align*} $$
$$ \begin{align*}p \Vdash_{\mathbb{Q}} \bigcup_{i<n} \{ (\bar y_i{}^{\frown} \dot z_{0,i,j}, k_{i,j}) : j < K_i \} \cup \{ (\bar y_i{}^{\frown} \dot z_{1,i,j}, k) : j < m_i \} \in R.\end{align*} $$
 Then 
 $\tilde R$
 is analytic (see, e.g., [Reference Kechris19, Theorem 29.22]).
$\tilde R$
 is analytic (see, e.g., [Reference Kechris19, Theorem 29.22]).
Claim 3.15. 
 $(1)_{\tilde R, M, 1}$
 is satisfied.
$(1)_{\tilde R, M, 1}$
 is satisfied.
Proof. Suppose 
 $\bar y_0, \dots , \bar y_{n-1}$
 are pairwise distinct and strongly mCg over M, but
$\bar y_0, \dots , \bar y_{n-1}$
 are pairwise distinct and strongly mCg over M, but 
 $\{(\bar y_0{}, 0), \dots ,(\bar y_{n-1}, 0) \} \in \tilde R$
 as witnessed by
$\{(\bar y_0{}, 0), \dots ,(\bar y_{n-1}, 0) \} \in \tilde R$
 as witnessed by 
 $p \in \mathbb {Q}$
,
$p \in \mathbb {Q}$
, 
 $\langle K_i : i < n \rangle $
 and
$\langle K_i : i < n \rangle $
 and 
 $\langle k_{i,j} : i < n, j < K_i \rangle $
, each
$\langle k_{i,j} : i < n, j < K_i \rangle $
, each 
 $k_{i,j} < k$
. More precisely,
$k_{i,j} < k$
. More precisely, 
 $$\begin{align} p \Vdash_{\mathbb{Q}} \bigcup_{i<n} \{ (\bar y_i{}^{\frown} \dot z_{0,i,j}, k_{i,j}) : j < K_i \} \in R. \end{align}$$
$$\begin{align} p \Vdash_{\mathbb{Q}} \bigcup_{i<n} \{ (\bar y_i{}^{\frown} \dot z_{0,i,j}, k_{i,j}) : j < K_i \} \in R. \end{align}$$
 By absoluteness, 
 $(*_0)$
 is satisfied in
$(*_0)$
 is satisfied in 
 $M[\bar y_0, \dots , \bar y_{n-1}]$
. Thus, let
$M[\bar y_0, \dots , \bar y_{n-1}]$
. Thus, let 
 $\langle z_{0,i,j}, z_{1,i,j} : i, j \in \omega \rangle $
 be generic over
$\langle z_{0,i,j}, z_{1,i,j} : i, j \in \omega \rangle $
 be generic over 
 $M[\bar y_0, \dots , \bar y_{n-1}]$
 with p in the associated generic filter. Then the
$M[\bar y_0, \dots , \bar y_{n-1}]$
 with p in the associated generic filter. Then the 
 $\bar y_i{}^{\frown } z_{(0,i,j)}$
 for
$\bar y_i{}^{\frown } z_{(0,i,j)}$
 for 
 $i < n, j < K_i$
 are pairwise distinct and strongly mCg over M, but
$i < n, j < K_i$
 are pairwise distinct and strongly mCg over M, but 
 $$ \begin{align*}\bigcup_{i<n} \{ (\bar y_i{}^{\frown} z_{0,i,j}, k_{i,j}) : j < K_i \} \in R.\end{align*} $$
$$ \begin{align*}\bigcup_{i<n} \{ (\bar y_i{}^{\frown} z_{0,i,j}, k_{i,j}) : j < K_i \} \in R.\end{align*} $$
This poses a contradiction to 
 $(1)_{R,M,k}$
.⊣
$(1)_{R,M,k}$
.⊣
Claim 3.16. If 
 $(1)_{\tilde R, M, m}$
 is satisfied for every
$(1)_{\tilde R, M, m}$
 is satisfied for every 
 $m \in \omega $
, then also
$m \in \omega $
, then also 
 $(1)_{R,M,k+1}$
.
$(1)_{R,M,k+1}$
.
Proof. Let 
 $\bar x_0, \dots , \bar x_{n-1} \in (2^\omega )^\alpha $
 be pairwise distinct, strongly mCg over M and let
$\bar x_0, \dots , \bar x_{n-1} \in (2^\omega )^\alpha $
 be pairwise distinct, strongly mCg over M and let 
 $k_0, \dots , k_{n-1} \leq k$
. Then we may write
$k_0, \dots , k_{n-1} \leq k$
. Then we may write 
 $\{ (\bar x_0, k_0), \dots , (\bar x_{n-1}, k_{n-1}) \}$
 as
$\{ (\bar x_0, k_0), \dots , (\bar x_{n-1}, k_{n-1}) \}$
 as 
 $$\begin{align}\bigcup_{i < n'} \{ (\bar y_i{}^{\frown} z_{0,i,j}, k_{i,j}) : j < K_i \} \cup \{ (\bar y_i{}^{\frown} z_{1,i,j}, k) : j < m_i \},\end{align}$$
$$\begin{align}\bigcup_{i < n'} \{ (\bar y_i{}^{\frown} z_{0,i,j}, k_{i,j}) : j < K_i \} \cup \{ (\bar y_i{}^{\frown} z_{1,i,j}, k) : j < m_i \},\end{align}$$
for some pairwise distinct 
 $\bar y_0, \dots ,\bar y_{n'-1}$
,
$\bar y_0, \dots ,\bar y_{n'-1}$
, 
 $\langle K_i : i < n' \rangle $
,
$\langle K_i : i < n' \rangle $
, 
 $\langle k_{i,j} : i < n', j<K_i \rangle $
,
$\langle k_{i,j} : i < n', j<K_i \rangle $
, 
 $\langle m_i : i < n' \rangle $
 and
$\langle m_i : i < n' \rangle $
 and 
 $\langle z_{0,i,j} : i, j \in \omega \rangle $
,
$\langle z_{0,i,j} : i, j \in \omega \rangle $
, 
 $\langle z_{1,i,j} : i, j \in \omega \rangle 2^\omega $
-mutually Cohen generic over
$\langle z_{1,i,j} : i, j \in \omega \rangle 2^\omega $
-mutually Cohen generic over 
 $M[\bar y_0, \dots , \bar y_{n'-1}]$
. Letting
$M[\bar y_0, \dots , \bar y_{n'-1}]$
. Letting 
 $m = \max _{i < n'} m_i +1$
, we follow the R-independence of the set in
$m = \max _{i < n'} m_i +1$
, we follow the R-independence of the set in 
 $(*_1)$
 from
$(*_1)$
 from 
 $(1)_{\tilde R,M,m}$
.⊣
$(1)_{\tilde R,M,m}$
.⊣
Claim 3.17. If there is 
 $m \in \omega $
 so that
$m \in \omega $
 so that 
 $(1)_{\tilde R, M, m}$
 fails, then there is a countable model
$(1)_{\tilde R, M, m}$
 fails, then there is a countable model 
 $M^+ \supseteq M$
 so that
$M^+ \supseteq M$
 so that 
 $(2)_{R,M^+,k}$
.
$(2)_{R,M^+,k}$
.
Proof. Let 
 $m\geq 1$
 be least so that
$m\geq 1$
 be least so that 
 $(2)_{\tilde R, M_0, m}$
 for some countable model
$(2)_{\tilde R, M_0, m}$
 for some countable model 
 $M_0 \supseteq M$
. We know that such m exists, since from
$M_0 \supseteq M$
. We know that such m exists, since from 
 $(1)_{\tilde R, M, 1}$
 we follow that either
$(1)_{\tilde R, M, 1}$
 we follow that either 
 $(1)_{\tilde R, M, 2}$
 or
$(1)_{\tilde R, M, 2}$
 or 
 $(2)_{\tilde R, M_0,1}$
 for some
$(2)_{\tilde R, M_0,1}$
 for some 
 $M_0$
, then, if
$M_0$
, then, if 
 $(1)_{\tilde R, M, 2}$
, either
$(1)_{\tilde R, M, 2}$
, either 
 $(1)_{\tilde R, M, 3}$
 or
$(1)_{\tilde R, M, 3}$
 or 
 $(2)_{\tilde R, M_0, 2}$
 for some
$(2)_{\tilde R, M_0, 2}$
 for some 
 $M_0$
, and so on. Let
$M_0$
, and so on. Let 
 $\phi _0, \dots , \phi _{N-1}$
,
$\phi _0, \dots , \phi _{N-1}$
, 
 $m_0, \dots , m_{N-1} \leq m$
 and
$m_0, \dots , m_{N-1} \leq m$
 and 
 $\bar s \in \bigotimes _{i < \beta }(2^{<\omega })$
 witness
$\bar s \in \bigotimes _{i < \beta }(2^{<\omega })$
 witness 
 $(2)_{\tilde R, M_0, m}$
. Let
$(2)_{\tilde R, M_0, m}$
. Let 
 $M_1$
 be a countable elementary model such that
$M_1$
 be a countable elementary model such that 
 $\phi _0, \dots , \phi _{N-1}, M_0 \in M_1$
. Then we have that for any
$\phi _0, \dots , \phi _{N-1}, M_0 \in M_1$
. Then we have that for any 
 $\bar y$
 that is Cohen generic in
$\bar y$
 that is Cohen generic in 
 $(2^\omega )^\beta \cap [\bar s]$
 over
$(2^\omega )^\beta \cap [\bar s]$
 over 
 $M_1$
, in particular over
$M_1$
, in particular over 
 $M_0$
, that
$M_0$
, that 
 $$ \begin{align*}\{(\bar y, m) \} \cup \{ (\phi_j(\bar y), m_j) : j <N \} \in \tilde R,\end{align*} $$
$$ \begin{align*}\{(\bar y, m) \} \cup \{ (\phi_j(\bar y), m_j) : j <N \} \in \tilde R,\end{align*} $$
i.e., there is 
 $p \in \mathbb {Q}$
, there are
$p \in \mathbb {Q}$
, there are 
 $K_i \in \omega $
,
$K_i \in \omega $
, 
 $k_{i,0}, \dots , k_{i,K_i -1} < k$
 for every
$k_{i,0}, \dots , k_{i,K_i -1} < k$
 for every 
 $i \leq N$
, so that
$i \leq N$
, so that
 By extending 
 $\bar s$
, we can assume wlog that p,
$\bar s$
, we can assume wlog that p, 
 $\langle K_i : i \leq N \rangle $
,
$\langle K_i : i \leq N \rangle $
, 
 $\langle k_{i,j} : i \leq N, j < K_i \rangle $
 are the same for each
$\langle k_{i,j} : i \leq N, j < K_i \rangle $
 are the same for each 
 $\bar y \in [\bar s]$
 generic over
$\bar y \in [\bar s]$
 generic over 
 $M_1$
, since
$M_1$
, since 
 $(*_2)$
 can be forced over
$(*_2)$
 can be forced over 
 $M_1$
. Also, from the fact that
$M_1$
. Also, from the fact that 
 $\phi _j$
 is continuous for every
$\phi _j$
 is continuous for every 
 $j < N$
, that
$j < N$
, that 
 $\phi _j(\bar y) \neq \bar y$
 for every
$\phi _j(\bar y) \neq \bar y$
 for every 
 $j < N$
, and that
$j < N$
, and that 
 $\phi _{j_0}(\bar y) \neq \phi _{j_1}(\bar y)$
 for every
$\phi _{j_0}(\bar y) \neq \phi _{j_1}(\bar y)$
 for every 
 $j_0 < j_1 < N$
, we can assume wlog that for any
$j_0 < j_1 < N$
, we can assume wlog that for any 
 $\bar y_0, \bar y_1 \in [\bar s]$
 and
$\bar y_0, \bar y_1 \in [\bar s]$
 and 
 $j_0 < j_1 < N$
,
$j_0 < j_1 < N$
, 
 $$\begin{align} \phi_{j_0}(\bar y_0) \neq \bar y_1 \text{ and } \phi_{j_0}(\bar y_0) \neq \phi_{j_1}(\bar y_1). \end{align}$$
$$\begin{align} \phi_{j_0}(\bar y_0) \neq \bar y_1 \text{ and } \phi_{j_0}(\bar y_0) \neq \phi_{j_1}(\bar y_1). \end{align}$$
Let us force in a finite support product over 
 $M_1$
 continuous functions
$M_1$
 continuous functions 
 $\chi _{0,i,j} \colon (2^\omega )^\beta \to [p(0,i,j)]$
 and
$\chi _{0,i,j} \colon (2^\omega )^\beta \to [p(0,i,j)]$
 and 
 $\chi _{1,i,j} \colon (2^\omega )^\beta \to [p(1,i,j)]$
 for
$\chi _{1,i,j} \colon (2^\omega )^\beta \to [p(1,i,j)]$
 for 
 $i,j \in \omega $
 and write
$i,j \in \omega $
 and write 
 $M^+ = M_1[\langle \chi _{0,i,j}, \chi _{1,i,j} : i,j \in \omega \rangle ]$
. For every
$M^+ = M_1[\langle \chi _{0,i,j}, \chi _{1,i,j} : i,j \in \omega \rangle ]$
. For every 
 $i < N$
 and
$i < N$
 and 
 $j < K_i$
 and
$j < K_i$
 and 
 $\bar x \in (2^\omega )^\alpha $
, define
$\bar x \in (2^\omega )^\alpha $
, define 
 $$ \begin{align*}\phi_{0,i,j}(\bar x ) := \phi_i(\bar x \restriction \beta){}^{\frown} \chi_{0,i,j}(\phi_i(\bar x \restriction \beta)) \text{ and } k_{0,i,j} = k_{i,j}.\end{align*} $$
$$ \begin{align*}\phi_{0,i,j}(\bar x ) := \phi_i(\bar x \restriction \beta){}^{\frown} \chi_{0,i,j}(\phi_i(\bar x \restriction \beta)) \text{ and } k_{0,i,j} = k_{i,j}.\end{align*} $$
For every 
 $i< N$
 and
$i< N$
 and 
 $j < m_i$
 and
$j < m_i$
 and 
 $\bar x \in (2^\omega )^\alpha $
, define
$\bar x \in (2^\omega )^\alpha $
, define 
 $$ \begin{align*}\phi_{1,i,j}(\bar x ) := \phi_i(\bar x \restriction \beta){}^{\frown} \chi_{1,i,j}(\phi_i(\bar x \restriction \beta)) \text{ and } k_{1,i,j} = k.\end{align*} $$
$$ \begin{align*}\phi_{1,i,j}(\bar x ) := \phi_i(\bar x \restriction \beta){}^{\frown} \chi_{1,i,j}(\phi_i(\bar x \restriction \beta)) \text{ and } k_{1,i,j} = k.\end{align*} $$
For every 
 $j < K_{N}$
 and
$j < K_{N}$
 and 
 $\bar x \in (2^\omega )^\alpha $
, define
$\bar x \in (2^\omega )^\alpha $
, define 
 $$ \begin{align*}\phi_{0,N,j}(\bar x ) := \bar x \restriction \beta{}^{\frown} \chi_{0,N,j}(\bar x \restriction \beta) \text{ and } k_{0,N,j} = k_{N,j}.\end{align*} $$
$$ \begin{align*}\phi_{0,N,j}(\bar x ) := \bar x \restriction \beta{}^{\frown} \chi_{0,N,j}(\bar x \restriction \beta) \text{ and } k_{0,N,j} = k_{N,j}.\end{align*} $$
At last, define for every 
 $j < m-1$
 and
$j < m-1$
 and 
 $\bar x \in (2^\omega )^\alpha $
,
$\bar x \in (2^\omega )^\alpha $
, 
 $$ \begin{align*}\phi_{1,N,j}(\bar x ) := \bar x \restriction \beta{}^{\frown} \chi_{1,N,j}(\bar x \restriction \beta) \text{ and } k_{1,N,j} = k.\end{align*} $$
$$ \begin{align*}\phi_{1,N,j}(\bar x ) := \bar x \restriction \beta{}^{\frown} \chi_{1,N,j}(\bar x \restriction \beta) \text{ and } k_{1,N,j} = k.\end{align*} $$
 Let 
 $\bar t \in \bigotimes _{i < \alpha } 2^{<\alpha }$
 be
$\bar t \in \bigotimes _{i < \alpha } 2^{<\alpha }$
 be 
 $\bar s$
 with
$\bar s$
 with 
 $p(1,N,m-1)$
 added in coordinate
$p(1,N,m-1)$
 added in coordinate 
 $\beta $
. Now we have that for any
$\beta $
. Now we have that for any 
 $\bar x \in [\bar t]$
 that is Cohen generic in
$\bar x \in [\bar t]$
 that is Cohen generic in 
 $(2^\omega )^\alpha $
 over
$(2^\omega )^\alpha $
 over 
 $M^+$
,
$M^+$
, 
 $$ \begin{align*} \{ (\bar x, k) \} &\cup \{ (\phi_{0,i,j}(\bar x), k_{0,i,j}) : i \leq N, j < K_N \} \cup \{ (\phi_{1,i,j}(\bar x), k_{1,i,j}) : i < N, j < m_i\} \\ &\cup \{(\phi_{1,N,j}(\bar x), k_{1,N,j}) : j < m -1 \} \in R. \end{align*} $$
$$ \begin{align*} \{ (\bar x, k) \} &\cup \{ (\phi_{0,i,j}(\bar x), k_{0,i,j}) : i \leq N, j < K_N \} \cup \{ (\phi_{1,i,j}(\bar x), k_{1,i,j}) : i < N, j < m_i\} \\ &\cup \{(\phi_{1,N,j}(\bar x), k_{1,N,j}) : j < m -1 \} \in R. \end{align*} $$
 This follows from 
 $(*)_2$
 and applying Lemma 3.3 to see that the
$(*)_2$
 and applying Lemma 3.3 to see that the 
 $\chi _{0,i,j}(\phi _i(\bar x \restriction \beta ))$
,
$\chi _{0,i,j}(\phi _i(\bar x \restriction \beta ))$
, 
 $\chi _{1,i,j}(\phi _i(\bar x \restriction \beta ))$
,
$\chi _{1,i,j}(\phi _i(\bar x \restriction \beta ))$
, 
 $\chi _{0,N,j}(\bar x \restriction \beta )$
,
$\chi _{0,N,j}(\bar x \restriction \beta )$
, 
 $\chi _{1,N,j}(\bar x \restriction \beta )$
, and
$\chi _{1,N,j}(\bar x \restriction \beta )$
, and 
 $x(\beta )$
 are mutually Cohen generic over
$x(\beta )$
 are mutually Cohen generic over 
 $M_1[\bar x \restriction \beta ]$
. Moreover they correspond to the reals
$M_1[\bar x \restriction \beta ]$
. Moreover they correspond to the reals 
 $z_{0,i,j}$
,
$z_{0,i,j}$
, 
 $z_{1,i,j}$
 added by a
$z_{1,i,j}$
 added by a 
 $\mathbb {Q}$
-generic over
$\mathbb {Q}$
-generic over 
 $M_1[\bar x \restriction \beta ]$
, containing p in its generic filter. Also, remember that
$M_1[\bar x \restriction \beta ]$
, containing p in its generic filter. Also, remember that 
 $(*)_2$
 is absolute between models containing the relevant parameters, which
$(*)_2$
 is absolute between models containing the relevant parameters, which 
 $M_1[\bar y]$
 is, with
$M_1[\bar y]$
 is, with 
 $\bar y = {\bar x \restriction \beta }$
.
$\bar y = {\bar x \restriction \beta }$
.
 On the other hand, whenever 
 $\bar x_0, \dots , \bar x_{n-1} \in (2^\omega )^\alpha \cap [\bar t]$
 are pairwise distinct and strongly mCg over
$\bar x_0, \dots , \bar x_{n-1} \in (2^\omega )^\alpha \cap [\bar t]$
 are pairwise distinct and strongly mCg over 
 $M^+$
, letting
$M^+$
, letting 
 $\bar y_0, \dots , \bar y_{n'-1}$
 enumerate
$\bar y_0, \dots , \bar y_{n'-1}$
 enumerate 
 $\{\bar x_i \restriction \beta : i < n \}$
, we have that
$\{\bar x_i \restriction \beta : i < n \}$
, we have that 
 $$\begin{align} \{ (\bar y_i{},m-1) : i < n' \} \cup \{ (\phi_j(\bar y_i), m_j) : i < n', j < N \} \text{ is } \tilde R \text{-independent.} \end{align}$$
$$\begin{align} \{ (\bar y_i{},m-1) : i < n' \} \cup \{ (\phi_j(\bar y_i), m_j) : i < n', j < N \} \text{ is } \tilde R \text{-independent.} \end{align}$$
According to the definition of 
 $\tilde R$
,
$\tilde R$
, 
 $(*_4)$
 is saying, e.g., that whenever
$(*_4)$
 is saying, e.g., that whenever 
 $A \cup B \subseteq (2^\omega )^\alpha $
 is an arbitrary set of strongly mCg reals over
$A \cup B \subseteq (2^\omega )^\alpha $
 is an arbitrary set of strongly mCg reals over 
 $M_1$
, where
$M_1$
, where 
 $A \restriction \beta , B \restriction \beta \subseteq \{ \bar y_i, \phi _j(\bar y_i) : i < n', j < N \}$
 and in B,
$A \restriction \beta , B \restriction \beta \subseteq \{ \bar y_i, \phi _j(\bar y_i) : i < n', j < N \}$
 and in B, 
 $\bar y_i$
 is extended at most
$\bar y_i$
 is extended at most 
 $m-1$
 many times and
$m-1$
 many times and 
 $\phi _j(\bar y_i)$
 at most
$\phi _j(\bar y_i)$
 at most 
 $m_j$
 many times for every
$m_j$
 many times for every 
 $i < n'$
,
$i < n'$
, 
 $j < N$
, and, assuming for now that
$j < N$
, and, assuming for now that 
 $k> 0$
, if
$k> 0$
, if 
 $f \colon A \to k$
, then
$f \colon A \to k$
, then 
 $$ \begin{align*}\{ (\bar x{}, f(\bar x)) : \bar x \in A \} \cup (B \times \{k\}) \text{ is } R\text{-independent.}\end{align*} $$
$$ \begin{align*}\{ (\bar x{}, f(\bar x)) : \bar x \in A \} \cup (B \times \{k\}) \text{ is } R\text{-independent.}\end{align*} $$
As an example for such sets A and B, we have
 $$ \begin{align*}A = \{\phi_{0,i,j}(\bar x_l) : l < n, i \leq N, j < K_i \} \cup \{ \bar x_l : l < n' \} \text{ and}\end{align*} $$
$$ \begin{align*}A = \{\phi_{0,i,j}(\bar x_l) : l < n, i \leq N, j < K_i \} \cup \{ \bar x_l : l < n' \} \text{ and}\end{align*} $$
 $$ \begin{align*}B = \{ \phi_{1,i,j}(\bar x_l) : l < n, i < N, j < m_i \} \cup \{ \phi_{1,N,j}(\bar x_l) : l < n', j < m-1 \}.\end{align*} $$
$$ \begin{align*}B = \{ \phi_{1,i,j}(\bar x_l) : l < n, i < N, j < m_i \} \cup \{ \phi_{1,N,j}(\bar x_l) : l < n', j < m-1 \}.\end{align*} $$
Again, to see this we apply Lemma 3.3 to show that the relevant reals are mutually generic over the model 
 $M_1[\bar y_0, \dots , \bar y_{n'-1}]$
. Also, remember from the definition of
$M_1[\bar y_0, \dots , \bar y_{n'-1}]$
. Also, remember from the definition of 
 $\phi _{1,i,j}$
 for
$\phi _{1,i,j}$
 for 
 $i < N$
 and
$i < N$
 and 
 $j < m_i$
 that, if
$j < m_i$
 that, if 
 $\phi _i(\bar x_{l_0} \restriction \beta ) = \phi _i(\bar x_{l_1} \restriction \beta )$
, then also
$\phi _i(\bar x_{l_0} \restriction \beta ) = \phi _i(\bar x_{l_1} \restriction \beta )$
, then also 
 $\phi _{1,i,j}(\bar x_{l_0}) = \phi _{1,i,j}(\bar x_{l_1})$
, for all
$\phi _{1,i,j}(\bar x_{l_0}) = \phi _{1,i,j}(\bar x_{l_1})$
, for all 
 $l_0, l_1 < n$
. Equally, if
$l_0, l_1 < n$
. Equally, if 
 $\bar x_{l_0} \restriction \beta = \bar x_{l_1} \restriction \beta $
, then
$\bar x_{l_0} \restriction \beta = \bar x_{l_1} \restriction \beta $
, then 
 $\phi _{1,N,j}(\bar x_{l_0}) = \phi _{1,N,j}(\bar x_{l_1})$
 for every
$\phi _{1,N,j}(\bar x_{l_0}) = \phi _{1,N,j}(\bar x_{l_1})$
 for every 
 $j < m-1$
. Use
$j < m-1$
. Use 
 $(*_3)$
 to note that
$(*_3)$
 to note that 
 $\{ \bar y_i : i < n'\}$
,
$\{ \bar y_i : i < n'\}$
, 
 $\{\phi _0(\bar y_i) : i < n' \}, \dots $
,
$\{\phi _0(\bar y_i) : i < n' \}, \dots $
, 
 ${\{ \phi _{N-1}(\bar y_i) : i < n' \}}$
 are pairwise disjoint. From this we can follow that indeed, each
${\{ \phi _{N-1}(\bar y_i) : i < n' \}}$
 are pairwise disjoint. From this we can follow that indeed, each 
 $\bar y_i$
 is extended at most
$\bar y_i$
 is extended at most 
 $m-1$
 many times in B and
$m-1$
 many times in B and 
 $\phi _j(\bar y_i)$
 at most
$\phi _j(\bar y_i)$
 at most 
 $m_i$
 many times. In total, we get that
$m_i$
 many times. In total, we get that 
 $$ \begin{align*} &\{(\phi_{0,i,j}(\bar x_l), k_{0,i,j}) : l < n, i \leq N, j < K_i \} \cup \{ (\bar x_l,k-1) : l < n' \} \cup \\ &\{ (\phi_{1,i,j}(\bar x_l), k) : l < n, i < N, j < m_i \} \cup \{ (\phi_{1,N,j}(\bar x_l), k) : l < n', j < m-1 \} \\ &\qquad\qquad\qquad\qquad\qquad\qquad\qquad\qquad\qquad\qquad\qquad\qquad\text{ is } R\text{-independent}. \end{align*} $$
$$ \begin{align*} &\{(\phi_{0,i,j}(\bar x_l), k_{0,i,j}) : l < n, i \leq N, j < K_i \} \cup \{ (\bar x_l,k-1) : l < n' \} \cup \\ &\{ (\phi_{1,i,j}(\bar x_l), k) : l < n, i < N, j < m_i \} \cup \{ (\phi_{1,N,j}(\bar x_l), k) : l < n', j < m-1 \} \\ &\qquad\qquad\qquad\qquad\qquad\qquad\qquad\qquad\qquad\qquad\qquad\qquad\text{ is } R\text{-independent}. \end{align*} $$
 It is now easy to check that we have the witnesses required in the statement of 
 $(2)_{R,M^+,k}$
. For example,
$(2)_{R,M^+,k}$
. For example, 
 $\phi _{0,i,j}(\bar x) \neq \bar x$
 when
$\phi _{0,i,j}(\bar x) \neq \bar x$
 when 
 $i < N$
 follows from
$i < N$
 follows from 
 $\phi _i(\bar x) \neq \bar x$
. For the values
$\phi _i(\bar x) \neq \bar x$
. For the values 
 $\phi _{0,N,j}(\bar x)$
 we simply have that
$\phi _{0,N,j}(\bar x)$
 we simply have that 
 $\chi _{0,N,j}(\bar x \restriction \beta ) \neq x(\beta )$
, as the two values are mutually generic. Everything else is similar and consists only of a few case distinctions. Also, the continuity of the functions is clear.
$\chi _{0,N,j}(\bar x \restriction \beta ) \neq x(\beta )$
, as the two values are mutually generic. Everything else is similar and consists only of a few case distinctions. Also, the continuity of the functions is clear.
 If 
 $k = 0$
, then we can simply forget the set A above, since
$k = 0$
, then we can simply forget the set A above, since 
 $K_i$
 must be
$K_i$
 must be 
 $0$
 for every
$0$
 for every 
 $i \leq N$
. In this case we just get that
$i \leq N$
. In this case we just get that 
 $$ \begin{align*} \{ (\phi_{1,i,j}(\bar x_l), k) : l < n, i < N, j < m_i \} \cup \{ (\phi_{1,N,j}(\bar x_l), k) &: l < n', j < m-1 \} \\ &\quad\text{ is } R\text{-independent}, \end{align*} $$
$$ \begin{align*} \{ (\phi_{1,i,j}(\bar x_l), k) : l < n, i < N, j < m_i \} \cup \{ (\phi_{1,N,j}(\bar x_l), k) &: l < n', j < m-1 \} \\ &\quad\text{ is } R\text{-independent}, \end{align*} $$
which then yields 
 $(2)_{R,M^+,k}$
.⊣
$(2)_{R,M^+,k}$
.⊣
 This finishes the successor step. Now assume that 
 $\alpha $
 is a limit ordinal. We fix some arbitrary tree
$\alpha $
 is a limit ordinal. We fix some arbitrary tree 
 $T \subseteq \omega ^{<\omega }$
 such that for every
$T \subseteq \omega ^{<\omega }$
 such that for every 
 $t \in T$
,
$t \in T$
, 
 $\vert \{ n \in \omega : t^{\frown } n \in T\} \vert = \omega $
 and for any branches
$\vert \{ n \in \omega : t^{\frown } n \in T\} \vert = \omega $
 and for any branches 
 $x \neq y \in [T]$
, if
$x \neq y \in [T]$
, if 
 $d = \min \{ i \in \omega : x(i)\neq y(i) \}$
 then
$d = \min \{ i \in \omega : x(i)\neq y(i) \}$
 then 
 $x(j) \neq x(j)$
 for every
$x(j) \neq x(j)$
 for every 
 $j \geq d$
. We will use T only for notational purposes. For every sequence
$j \geq d$
. We will use T only for notational purposes. For every sequence 
 $\xi _0 < \dots < \xi _{k'} = \alpha $
, we let
$\xi _0 < \dots < \xi _{k'} = \alpha $
, we let 
 $\mathbb {Q}_{\xi _0, \dots , \xi _{k'}} = \left (\prod _{l < k'} (\bigotimes _{i \in [\xi _l, \xi _{l+1})} 2^{<\omega })^{<\omega }\right ) \times ( \bigotimes _{i \in [\xi _0, \alpha )} 2^{<\omega })^{<\omega }$
.
$\mathbb {Q}_{\xi _0, \dots , \xi _{k'}} = \left (\prod _{l < k'} (\bigotimes _{i \in [\xi _l, \xi _{l+1})} 2^{<\omega })^{<\omega }\right ) \times ( \bigotimes _{i \in [\xi _0, \alpha )} 2^{<\omega })^{<\omega }$
. 
 $\mathbb {Q}_{\xi _0, \dots , \xi _{k'}}$
 adds, in the natural way, reals
$\mathbb {Q}_{\xi _0, \dots , \xi _{k'}}$
 adds, in the natural way, reals 
 $\langle \bar z^0_{l, i} : l < k', i \in \omega \rangle $
 and
$\langle \bar z^0_{l, i} : l < k', i \in \omega \rangle $
 and 
 $\langle \bar z^0_i : i \in \omega \rangle $
, where
$\langle \bar z^0_i : i \in \omega \rangle $
, where 
 $\bar z^0_{l,i} \in (2^\omega )^{[\xi _l, \xi _{l+1})}$
 and
$\bar z^0_{l,i} \in (2^\omega )^{[\xi _l, \xi _{l+1})}$
 and 
 $\bar z^1_i \in (2^{\omega })^{[\xi _0, \alpha )}$
 for every
$\bar z^1_i \in (2^{\omega })^{[\xi _0, \alpha )}$
 for every 
 $l < k'$
,
$l < k'$
, 
 $i \in \omega $
. Whenever
$i \in \omega $
. Whenever 
 $t \in T \cap \omega ^{k'}$
, we write
$t \in T \cap \omega ^{k'}$
, we write 
 $\bar z^0_{t} = \bar z^0_{0,t(0)}{}^{\frown } \cdots {}^{\frown } \bar z^0_{k'-1, t(k'-1)}$
. Note that for generic
$\bar z^0_{t} = \bar z^0_{0,t(0)}{}^{\frown } \cdots {}^{\frown } \bar z^0_{k'-1, t(k'-1)}$
. Note that for generic 
 $\langle \bar z^0_{l,i} : i \in \omega , l < k' \rangle $
, the reals
$\langle \bar z^0_{l,i} : i \in \omega , l < k' \rangle $
, the reals 
 $\langle \bar z^0_t : t \in T \cap \omega ^{k'}\rangle $
 are strongly
$\langle \bar z^0_t : t \in T \cap \omega ^{k'}\rangle $
 are strongly 
 $\langle 2^\omega : i \in [\xi _0, \alpha )\rangle $
-mCg.
$\langle 2^\omega : i \in [\xi _0, \alpha )\rangle $
-mCg.
 Now, let us define for each 
 $\xi < \alpha $
 an analytic hypergraph
$\xi < \alpha $
 an analytic hypergraph 
 $R_\xi $
 on
$R_\xi $
 on 
 $(2^\omega )^\xi \times 2$
 so that
$(2^\omega )^\xi \times 2$
 so that 
 $\{(\bar y^0_i, 0) : i < n_0 \} \cup \{ (\bar y^1_i, 1) : i < n_1 \} \in R_\xi \cap [(2^\omega )^\xi \times 2]^{n_0+n_1}$
, where
$\{(\bar y^0_i, 0) : i < n_0 \} \cup \{ (\bar y^1_i, 1) : i < n_1 \} \in R_\xi \cap [(2^\omega )^\xi \times 2]^{n_0+n_1}$
, where 
 $\vert \{(\bar y^0_i, 0) : i < n_0 \}\vert = n_0$
 and
$\vert \{(\bar y^0_i, 0) : i < n_0 \}\vert = n_0$
 and 
 $\vert \{ (\bar y^1_i, 1) : i < n_1 \} \vert = n_1$
, iff there are
$\vert \{ (\bar y^1_i, 1) : i < n_1 \} \vert = n_1$
, iff there are 
 $\xi _0 = \xi < \dots < \xi _{k'} = \alpha $
,
$\xi _0 = \xi < \dots < \xi _{k'} = \alpha $
, 
 $(p,q) \in \mathbb {Q}_{\xi _0, \dots , \xi _{k'}}$
,
$(p,q) \in \mathbb {Q}_{\xi _0, \dots , \xi _{k'}}$
, 
 $K_i \in \omega $
,
$K_i \in \omega $
, 
 $k_{i,0}, \dots , k_{i,K_i -1}<k$
 and distinct
$k_{i,0}, \dots , k_{i,K_i -1}<k$
 and distinct 
 $t_{i,0}, \dots , t_{i,K_i -1} \in T \cap \omega ^{k'}$
 for every
$t_{i,0}, \dots , t_{i,K_i -1} \in T \cap \omega ^{k'}$
 for every 
 $i < n_0$
, so that
$i < n_0$
, so that 
 $t_{i_0,j_0}(0) \neq t_{i_1,j_1}(0)$
 for all
$t_{i_0,j_0}(0) \neq t_{i_1,j_1}(0)$
 for all 
 $i_0 < i_1 < n_0$
 and
$i_0 < i_1 < n_0$
 and 
 $j_0 < K_{i_0}, j_1 < K_{i_1}$
, and
$j_0 < K_{i_0}, j_1 < K_{i_1}$
, and 
 $$ \begin{align*}(p,q) \Vdash_{\mathbb{Q}_{\bar \xi}} \bigcup_{i < n_0} \{ (\bar y^0_i{}^{\frown} \bar z^0_{t_{i,j}}, k_{i,j}): j < K_i \} \cup \{ (\bar y^1_i{}^{\frown} \bar z^1_{i}, k): i < n_1 \} \in R.\end{align*} $$
$$ \begin{align*}(p,q) \Vdash_{\mathbb{Q}_{\bar \xi}} \bigcup_{i < n_0} \{ (\bar y^0_i{}^{\frown} \bar z^0_{t_{i,j}}, k_{i,j}): j < K_i \} \cup \{ (\bar y^1_i{}^{\frown} \bar z^1_{i}, k): i < n_1 \} \in R.\end{align*} $$
 Note that each 
 $R_\xi $
 can be defined within M. It should be clear, similar to the proof of Claim 3.15, that from
$R_\xi $
 can be defined within M. It should be clear, similar to the proof of Claim 3.15, that from 
 $(1)_{R,M,k}$
, we can show the following.
$(1)_{R,M,k}$
, we can show the following.
Claim 3.18. For every 
 $\xi < \alpha $
,
$\xi < \alpha $
, 
 $(1)_{R_\xi ,M,1}$
.
$(1)_{R_\xi ,M,1}$
.
Claim 3.19. Assume that for every 
 $\xi < \alpha $
,
$\xi < \alpha $
, 
 $(1)_{R_\xi ,M,2}$
. Then also
$(1)_{R_\xi ,M,2}$
. Then also 
 $(1)_{R,M,k+1}$
.
$(1)_{R,M,k+1}$
.
Proof. Let 
 $\bar x^0_0, \dots , \bar x^0_{n_0-1}, \bar x^1_0, \dots , \bar x^1_{n_1-1}$
 be pairwise distinct and strongly mCg over M and
$\bar x^0_0, \dots , \bar x^0_{n_0-1}, \bar x^1_0, \dots , \bar x^1_{n_1-1}$
 be pairwise distinct and strongly mCg over M and 
 $k_0, \dots , k_{n_0 -1} < k$
. Then there is
$k_0, \dots , k_{n_0 -1} < k$
. Then there is 
 $\xi < \alpha $
 large enough so that
$\xi < \alpha $
 large enough so that 
 $\bar x^0_0 \restriction \xi , \dots , \bar x^0_{n_0-1}\restriction \xi , \bar x^1_0\restriction \xi , \dots , \bar x^1_{n_1-1}\restriction \xi $
 are pairwise distinct and in particular,
$\bar x^0_0 \restriction \xi , \dots , \bar x^0_{n_0-1}\restriction \xi , \bar x^1_0\restriction \xi , \dots , \bar x^1_{n_1-1}\restriction \xi $
 are pairwise distinct and in particular, 
 $\bar x^0_0 \restriction [\xi ,\alpha ), \dots , \bar x^0_{n_0-1}\restriction [\xi , \alpha ) , \bar x^1_0\restriction [\xi , \alpha ), \dots , \bar x^1_{n_1-1}\restriction [\xi , \alpha )$
 are pairwise different in every coordinate. Let
$\bar x^0_0 \restriction [\xi ,\alpha ), \dots , \bar x^0_{n_0-1}\restriction [\xi , \alpha ) , \bar x^1_0\restriction [\xi , \alpha ), \dots , \bar x^1_{n_1-1}\restriction [\xi , \alpha )$
 are pairwise different in every coordinate. Let 
 $\xi _0 = \xi $
,
$\xi _0 = \xi $
, 
 $\xi _1 = \alpha $
,
$\xi _1 = \alpha $
, 
 $K_i = 1$
 for every
$K_i = 1$
 for every 
 $i < n_0$
 and
$i < n_0$
 and 
 $t_{0,0}, \dots , t_{n_0 -1,0} \in T \cap \omega ^1$
 pairwise distinct. Also, write
$t_{0,0}, \dots , t_{n_0 -1,0} \in T \cap \omega ^1$
 pairwise distinct. Also, write 
 $k_{0,0} = k_0, \ldots , k_{n_0-1,0} = k_{n_0 -1}$
. Then, from
$k_{0,0} = k_0, \ldots , k_{n_0-1,0} = k_{n_0 -1}$
. Then, from 
 $(1)_{R_\xi ,M,2}$
, we have that
$(1)_{R_\xi ,M,2}$
, we have that 

By absoluteness, this holds true in 
 $M[\langle \bar x^0_i \restriction \xi ,\bar x^1_j\restriction \xi : i < n_0, j < n_1\rangle ]$
 and we find that
$M[\langle \bar x^0_i \restriction \xi ,\bar x^1_j\restriction \xi : i < n_0, j < n_1\rangle ]$
 and we find that 
 $$ \begin{align*}\{(\bar x^0_i{}, k_i) : i < n_0 \} \cup \{ (\bar x^1_i, k) \} \text{ is } R\text{-independent},\end{align*} $$
$$ \begin{align*}\{(\bar x^0_i{}, k_i) : i < n_0 \} \cup \{ (\bar x^1_i, k) \} \text{ is } R\text{-independent},\end{align*} $$
as required.⊣
Claim 3.20. If there is 
 $\xi < \alpha $
 so that
$\xi < \alpha $
 so that 
 $(1)_{R_\xi ,M,2}\,$
 fails, then there is a countable model
$(1)_{R_\xi ,M,2}\,$
 fails, then there is a countable model 
 $M^+ \supseteq M$
 so that
$M^+ \supseteq M$
 so that 
 $(2)_{R,M^+,k}$
.
$(2)_{R,M^+,k}$
.
Proof. If 
 $(1)_{R_\xi , M, 2}$
 fails, then there is a countable model
$(1)_{R_\xi , M, 2}$
 fails, then there is a countable model 
 $M_0 \supseteq M$
 so that
$M_0 \supseteq M$
 so that 
 $(2)_{R_\xi , M, 1}$
 holds true as witnessed by
$(2)_{R_\xi , M, 1}$
 holds true as witnessed by 
 $\bar s \in \bigotimes _{i < \xi } 2^{<\omega }$
,
$\bar s \in \bigotimes _{i < \xi } 2^{<\omega }$
, 
 $\phi ^0_{0}, \dots , \phi ^0_{N_0 -1}, \phi ^1_{0}, \dots , \phi ^1_{N_1 -1} \colon (2^\omega )^\xi \to (2^\omega )^\xi $
 such that for any pairwise distinct
$\phi ^0_{0}, \dots , \phi ^0_{N_0 -1}, \phi ^1_{0}, \dots , \phi ^1_{N_1 -1} \colon (2^\omega )^\xi \to (2^\omega )^\xi $
 such that for any pairwise distinct 
 $\bar y_0, \dots , \bar y_{n-1} \in (2^\omega )^\xi \cap [\bar s]$
 that are strongly mCg over
$\bar y_0, \dots , \bar y_{n-1} \in (2^\omega )^\xi \cap [\bar s]$
 that are strongly mCg over 
 $M_0$
,
$M_0$
, 
 $$\begin{align} \{ (\bar y_i, 0) : i < n\} \cup \{(\phi^0_j(\bar y_i), 0) : i < n, j < N_0\} \cup \{ (\phi^1_j(\bar y_i), 1) : i < n, j < N_1\} \end{align}$$
$$\begin{align} \{ (\bar y_i, 0) : i < n\} \cup \{(\phi^0_j(\bar y_i), 0) : i < n, j < N_0\} \cup \{ (\phi^1_j(\bar y_i), 1) : i < n, j < N_1\} \end{align}$$
is 
 $R_\xi $
-independent, but
$R_\xi $
-independent, but 
 $$\begin{align} \{ (\bar y_0, 1) \} \cup \{ (\phi^0_j(\bar y_0), 0) : j < N_0\} \cup \{ (\phi^1_j(\bar y_0), 1) : j < N_1\} \in R_\xi. \end{align}$$
$$\begin{align} \{ (\bar y_0, 1) \} \cup \{ (\phi^0_j(\bar y_0), 0) : j < N_0\} \cup \{ (\phi^1_j(\bar y_0), 1) : j < N_1\} \in R_\xi. \end{align}$$
As before, we may pick 
 $M_1 \ni M_0$
 elementary containing all relevant information, assume that
$M_1 \ni M_0$
 elementary containing all relevant information, assume that 
 $(*_6)$
 is witnessed by fixed
$(*_6)$
 is witnessed by fixed 
 $\xi _0 = \xi < \dots < \xi _{k'} = \alpha $
,
$\xi _0 = \xi < \dots < \xi _{k'} = \alpha $
, 
 $(p,q) \in \mathbb {Q}_{\xi _0, \dots , \xi _{k'}}$
,
$(p,q) \in \mathbb {Q}_{\xi _0, \dots , \xi _{k'}}$
, 
 $K_0, \dots , K_{N_0-1}$
,
$K_0, \dots , K_{N_0-1}$
, 
 $k_{i,0}, \dots , k_{i, K_i -1}$
, and
$k_{i,0}, \dots , k_{i, K_i -1}$
, and 
 $t_{i,0}, \dots , t_{i, K_i-1} \in T \cap \omega ^{k'}$
 for every
$t_{i,0}, \dots , t_{i, K_i-1} \in T \cap \omega ^{k'}$
 for every 
 $i < N_0$
, so that for every generic
$i < N_0$
, so that for every generic 
 $\bar y_0 \in (2^\omega )^\xi \cap [\bar s]$
 over
$\bar y_0 \in (2^\omega )^\xi \cap [\bar s]$
 over 
 $M_1$
,
$M_1$
, 
 $$\begin{align} (p,q) \Vdash_{\mathbb{Q}_{\bar \xi}} \{(\bar y_0{}^{\frown} \bar z^1_{N_1}, k) \}\cup \bigcup_{i < N_0}&\{ (\phi^0_i(\bar y_0){}^{\frown} \bar z^0_{t_{i,j}}, k_{i,j}): j < K_i\} \cup \\ &\{ (\phi^1_j(\bar y_0){}^{\frown} \bar z^1_{j}, k): j < N_1 \} \in R.\nonumber \end{align}$$
$$\begin{align} (p,q) \Vdash_{\mathbb{Q}_{\bar \xi}} \{(\bar y_0{}^{\frown} \bar z^1_{N_1}, k) \}\cup \bigcup_{i < N_0}&\{ (\phi^0_i(\bar y_0){}^{\frown} \bar z^0_{t_{i,j}}, k_{i,j}): j < K_i\} \cup \\ &\{ (\phi^1_j(\bar y_0){}^{\frown} \bar z^1_{j}, k): j < N_1 \} \in R.\nonumber \end{align}$$
 As before, we may also assume that 
 $\bar y_0 \neq \phi ^{j_0}_{i_0}(\bar y_0) \neq \phi ^{j_1}_{i_1}(\bar y_1)$
 for every
$\bar y_0 \neq \phi ^{j_0}_{i_0}(\bar y_0) \neq \phi ^{j_1}_{i_1}(\bar y_1)$
 for every 
 $\bar y_0, \bar y_1 \in [\bar s]$
 and
$\bar y_0, \bar y_1 \in [\bar s]$
 and 
 $(j_0,i_1) \neq (j_1,i_1)$
. We let
$(j_0,i_1) \neq (j_1,i_1)$
. We let 
 $\bar s' = \bar s{}^{\frown } q(N_1)$
. Now we force continuous functions
$\bar s' = \bar s{}^{\frown } q(N_1)$
. Now we force continuous functions 
 $\chi ^0_{l,i} \colon (2^\omega )^\xi \to (2^{\omega })^{[\xi _l, \xi _{l+1})} \cap [p(l,i)]$
 and
$\chi ^0_{l,i} \colon (2^\omega )^\xi \to (2^{\omega })^{[\xi _l, \xi _{l+1})} \cap [p(l,i)]$
 and 
 $\chi ^1_{i} \colon (2^\omega )^\xi \to (2^{\omega })^{[\xi , \alpha )} \cap [q(i)]$
 over
$\chi ^1_{i} \colon (2^\omega )^\xi \to (2^{\omega })^{[\xi , \alpha )} \cap [q(i)]$
 over 
 $M_1$
 for every
$M_1$
 for every 
 $i \in \omega $
,
$i \in \omega $
, 
 $l < k'$
 and we let
$l < k'$
 and we let 
 $M^+ = M_1[\langle \chi ^0_{l,i}, \chi ^1_{i} : i \in \omega , l < k'\rangle ]$
. Finally we let
$M^+ = M_1[\langle \chi ^0_{l,i}, \chi ^1_{i} : i \in \omega , l < k'\rangle ]$
. Finally we let 
 $$ \begin{align*}\phi_{0,i,j}(\bar x) := \phi^0_i(\bar x \restriction \xi){}^{\frown} \chi_{0,t_{i,j}(0)}{}(\phi^0_i(\bar x \restriction \xi)){}^{\frown} \cdots{}^{\frown} \chi_{k'-1,t_{i,j}(k'-1)}(\phi^0_i(\bar x \restriction \xi))\end{align*} $$
$$ \begin{align*}\phi_{0,i,j}(\bar x) := \phi^0_i(\bar x \restriction \xi){}^{\frown} \chi_{0,t_{i,j}(0)}{}(\phi^0_i(\bar x \restriction \xi)){}^{\frown} \cdots{}^{\frown} \chi_{k'-1,t_{i,j}(k'-1)}(\phi^0_i(\bar x \restriction \xi))\end{align*} $$
for every 
 $i < N_0$
 and
$i < N_0$
 and 
 $j < K_i$
,
$j < K_i$
, 
 $\bar x \in (2^\omega )^\alpha $
, and
$\bar x \in (2^\omega )^\alpha $
, and 
 $$ \begin{align*}\phi_{1,i}(\bar x) := \phi^1_i(\bar x \restriction \xi){}^{\frown} \chi_{1,i}(\phi ^1_i(\bar x \restriction \xi))\end{align*} $$
$$ \begin{align*}\phi_{1,i}(\bar x) := \phi^1_i(\bar x \restriction \xi){}^{\frown} \chi_{1,i}(\phi ^1_i(\bar x \restriction \xi))\end{align*} $$
for every 
 $i < N_1$
,
$i < N_1$
, 
 $\bar x \in (2^\omega )^\alpha $
.
$\bar x \in (2^\omega )^\alpha $
.
 We get from 
 $(*_7)$
, and, as usual, applying Lemma 3.3, that for any
$(*_7)$
, and, as usual, applying Lemma 3.3, that for any 
 $\bar x \in (2^\omega )^\alpha \cap [\bar s']$
 which is generic over
$\bar x \in (2^\omega )^\alpha \cap [\bar s']$
 which is generic over 
 $M^+$
,
$M^+$
, 
 $$ \begin{align*}\{(\bar x, k)\} \cup \bigcup_{i < N_0} \{(\phi_{0,i,j}(\bar x), k_{i,j}) : j < K_i \} \cup \{(\phi_{1,i}(\bar x), k) : i < N_1 \} \in R.\end{align*} $$
$$ \begin{align*}\{(\bar x, k)\} \cup \bigcup_{i < N_0} \{(\phi_{0,i,j}(\bar x), k_{i,j}) : j < K_i \} \cup \{(\phi_{1,i}(\bar x), k) : i < N_1 \} \in R.\end{align*} $$
 On the other hand, whenever 
 $\bar x_0,\dots , \bar x_{n'-1}\in (2^{\omega })^\alpha \cap [\bar s']$
 are strongly mCg over
$\bar x_0,\dots , \bar x_{n'-1}\in (2^{\omega })^\alpha \cap [\bar s']$
 are strongly mCg over 
 $M^+$
, and letting
$M^+$
, and letting 
 $\bar y_0, \dots , \bar y_{n-1}$
 enumerate
$\bar y_0, \dots , \bar y_{n-1}$
 enumerate 
 $\{ \bar x_i \restriction \xi : i < n'\}$
, knowing that the set in
$\{ \bar x_i \restriction \xi : i < n'\}$
, knowing that the set in 
 $(*_5)$
 is
$(*_5)$
 is 
 $R_\xi $
-independent, we get that
$R_\xi $
-independent, we get that 
 $$ \begin{align*} \{ (\bar x_l,k-1) : l < n' \} \cup \bigcup_{i < N_0} &\{(\phi_{0,i,j}(\bar x_l),k_{i,j}) : j < K_i, l < n' \} \cup \\ &\quad\{(\phi_{1,i}(\bar x_l), k) : i < N_1, l < n' \} \text{ is } R\text{-independent}, \end{align*} $$
$$ \begin{align*} \{ (\bar x_l,k-1) : l < n' \} \cup \bigcup_{i < N_0} &\{(\phi_{0,i,j}(\bar x_l),k_{i,j}) : j < K_i, l < n' \} \cup \\ &\quad\{(\phi_{1,i}(\bar x_l), k) : i < N_1, l < n' \} \text{ is } R\text{-independent}, \end{align*} $$
 in case 
 $k>0$
. To see this, we let
$k>0$
. To see this, we let 
 $\eta _0< \dots < \eta _{k"}$
 be a partition refining
$\eta _0< \dots < \eta _{k"}$
 be a partition refining 
 $\xi _0 < \cdots \xi _{k'}$
 witnessing the mCg of
$\xi _0 < \cdots \xi _{k'}$
 witnessing the mCg of 
 $\bar x_0 \restriction [\xi , \alpha ), \dots , \bar x_{n'-1}\restriction [\xi , \alpha )$
 and we find appropriate
$\bar x_0 \restriction [\xi , \alpha ), \dots , \bar x_{n'-1}\restriction [\xi , \alpha )$
 and we find appropriate 
 $u_{0,0}, \dots , u_{0,L_0 -1}, \dots , u_{n-1,0 }, \dots , u_{n-1,L_{n-1}-1} \in T \cap \omega ^{k"}$
 and
$u_{0,0}, \dots , u_{0,L_0 -1}, \dots , u_{n-1,0 }, \dots , u_{n-1,L_{n-1}-1} \in T \cap \omega ^{k"}$
 and 
 $v_{i,j} \in T \cap \omega ^{k"}$
 for
$v_{i,j} \in T \cap \omega ^{k"}$
 for 
 $i < N_0, j < K_i$
 to interpret the above set in the form
$i < N_0, j < K_i$
 to interpret the above set in the form 
 $$ \begin{align*} \{ (\bar y_l{}^{\frown} \bar z^0_{u_{l,i}}, k-1) : l < n, i < L_i \} \cup \bigcup_{i < N_0} &\{(\phi^0_i(\bar y_l){}^{\frown} \bar z^0_{v_{i,j}}, k_{i,j}) : i < N_0, j < K_i, l < n \} \\ &\quad\cup \{(\phi^1_{i}(\bar y_l){}^{\frown} \bar z^1_i , k) : i < N_1, l < n \}, \end{align*} $$
$$ \begin{align*} \{ (\bar y_l{}^{\frown} \bar z^0_{u_{l,i}}, k-1) : l < n, i < L_i \} \cup \bigcup_{i < N_0} &\{(\phi^0_i(\bar y_l){}^{\frown} \bar z^0_{v_{i,j}}, k_{i,j}) : i < N_0, j < K_i, l < n \} \\ &\quad\cup \{(\phi^1_{i}(\bar y_l){}^{\frown} \bar z^1_i , k) : i < N_1, l < n \}, \end{align*} $$
for 
 $\mathbb {Q}_{\eta _0, \dots , \eta _{k"-1}}$
-generic
$\mathbb {Q}_{\eta _0, \dots , \eta _{k"-1}}$
-generic 
 $\langle \bar z ^0_{l,i}, \bar z^1_{i} : l < k", i \in \omega \rangle $
 over
$\langle \bar z ^0_{l,i}, \bar z^1_{i} : l < k", i \in \omega \rangle $
 over 
 $M_1[\bar y_0, \dots , \bar y_{n-1}]$
. We leave the details to the reader. In case
$M_1[\bar y_0, \dots , \bar y_{n-1}]$
. We leave the details to the reader. In case 
 $k= 0$
, all
$k= 0$
, all 
 $K_i$
 are
$K_i$
 are 
 $0$
 and we get that
$0$
 and we get that 
 $$ \begin{align*}\{(\phi_{1,i}(\bar x_l), k) : i < N_1, l < n' \} \text{ is } R\text{-independent}.\end{align*} $$
$$ \begin{align*}\{(\phi_{1,i}(\bar x_l), k) : i < N_1, l < n' \} \text{ is } R\text{-independent}.\end{align*} $$
Everything that remains, namely showing, e.g., that 
 $\bar x \neq \phi _{1,i}(\bar x)$
 is clear.⊣
$\bar x \neq \phi _{1,i}(\bar x)$
 is clear.⊣
 As a final note, let us observe that the case 
 $\alpha = 0$
 is trivial, since
$\alpha = 0$
 is trivial, since 
 $(2^\omega )^\alpha $
 has only one element.⊣
$(2^\omega )^\alpha $
 has only one element.⊣
Remark 3.21. If we replace “strong mCg” with “‘mCg” in the above lemma, then it already becomes false for 
 $\alpha = \omega $
. Namely consider the equivalence relation E on
$\alpha = \omega $
. Namely consider the equivalence relation E on 
 $(2^\omega )^\omega $
, where
$(2^\omega )^\omega $
, where 
 $\bar x E \bar y$
 if they eventually agree, i.e., if
$\bar x E \bar y$
 if they eventually agree, i.e., if 
 $\exists n \in \omega \forall m \geq n (x(n) = y(n))$
.Footnote 
5
 Then we can never be in case (1) since we can always find two distinct
$\exists n \in \omega \forall m \geq n (x(n) = y(n))$
.Footnote 
5
 Then we can never be in case (1) since we can always find two distinct 
 $\bar x$
 and
$\bar x$
 and 
 $\bar y$
 that are mCg and
$\bar y$
 that are mCg and 
 $\bar x E \bar y$
. On the other hand, in case (2) we get a continuous selector
$\bar x E \bar y$
. On the other hand, in case (2) we get a continuous selector 
 $\phi _0$
 for E (note that
$\phi _0$
 for E (note that 
 $N = 0$
 is not possible). More precisely we have that for any
$N = 0$
 is not possible). More precisely we have that for any 
 $\bar x$
,
$\bar x$
, 
 $\bar y$
 that are mCg,
$\bar y$
 that are mCg, 
 $\bar x E \phi _0(\bar x)$
 and
$\bar x E \phi _0(\bar x)$
 and 
 $\phi _0(\bar x) = \phi _0(\bar y)$
 iff
$\phi _0(\bar x) = \phi _0(\bar y)$
 iff 
 $\bar x E \bar y$
. But for arbitrary mCg
$\bar x E \bar y$
. But for arbitrary mCg 
 $\bar x$
 and
$\bar x$
 and 
 $\bar y$
 so that
$\bar y$
 so that 
 $\bar x \neg E \bar y$
, we easily find a sequence
$\bar x \neg E \bar y$
, we easily find a sequence 
 $\langle \bar x_n : n \in \omega \rangle $
 so that
$\langle \bar x_n : n \in \omega \rangle $
 so that 
 $\bar x$
 and
$\bar x$
 and 
 $\bar x_n$
 are mCg and
$\bar x_n$
 are mCg and 
 $\bar x E \bar x_n$
, but
$\bar x E \bar x_n$
, but 
 $\bar x_n \restriction n = \bar y \restriction n$
 for all n. In particular
$\bar x_n \restriction n = \bar y \restriction n$
 for all n. In particular 
 $\lim _{n \in \omega } \bar x_n = \bar y$
. Then
$\lim _{n \in \omega } \bar x_n = \bar y$
. Then 
 $\phi _0(\bar y) = \lim _{n\in \omega } \phi _0(\bar x_n) = \lim _{n\in \omega } \phi _0(\bar x) = \phi _0(\bar x)$
.
$\phi _0(\bar y) = \lim _{n\in \omega } \phi _0(\bar x_n) = \lim _{n\in \omega } \phi _0(\bar x) = \phi _0(\bar x)$
.
 The proofs of Main Lemmas 3.4 and 3.14 can be generalized to E that is 
 $\omega $
-universally Baire, in particular they also hold for coanalytic hypergraphs. The only assumptions on analytic sets that we used in the proofs are summarized below.
$\omega $
-universally Baire, in particular they also hold for coanalytic hypergraphs. The only assumptions on analytic sets that we used in the proofs are summarized below.
Proposition 3.22. Let 
 $\Gamma $
 be a pointclass closed under countable unions, countable intersections, and continuous preimages and assume that for every
$\Gamma $
 be a pointclass closed under countable unions, countable intersections, and continuous preimages and assume that for every 
 $A \in \Gamma \cap \mathcal {P}(\omega ^\omega )$
, there are formulas
$A \in \Gamma \cap \mathcal {P}(\omega ^\omega )$
, there are formulas 
 $\varphi $
,
$\varphi $
, 
 $\psi $
 (with parameters) in the language of set theory, such that for every countable elementary model M (with the relevant parameters) and G a generic over M for a finite support product of Cohen forcing,
$\psi $
 (with parameters) in the language of set theory, such that for every countable elementary model M (with the relevant parameters) and G a generic over M for a finite support product of Cohen forcing, 
- 
(1) for  $x \in M[G] \cap \omega ^\omega $
, $x \in M[G] \cap \omega ^\omega $
, $M[G] \models \varphi (x) \text { iff } x \in A \text { and } M[G] \models \psi (x) \text { iff } x \notin A,$ $M[G] \models \varphi (x) \text { iff } x \in A \text { and } M[G] \models \psi (x) \text { iff } x \notin A,$
- 
(2) for  $\dot x \in M[G]$
 a $\dot x \in M[G]$
 a $\mathbb {C}$
-name for a real, $\mathbb {C}$
-name for a real, $p \in \mathbb {C}$
, $p \in \mathbb {C}$
, $M[G] \models "p \Vdash \varphi (\dot x)"$
 iff $M[G] \models "p \Vdash \varphi (\dot x)"$
 iff $p \Vdash \varphi (\dot x)$
, $p \Vdash \varphi (\dot x)$
,
- 
(3) for  $\dot y$
 a $\dot y$
 a $\mathbb {C}$
-name for a real, $\mathbb {C}$
-name for a real, $p \in \mathbb {C}$
 and a continuous function $p \in \mathbb {C}$
 and a continuous function $f \colon \omega ^\omega \times \omega ^\omega \to \omega ^\omega $
, $f \colon \omega ^\omega \times \omega ^\omega \to \omega ^\omega $
, $\{ x \in \omega ^\omega : p \Vdash \varphi (f(x,\dot y)) \} \in \Gamma .$ $\{ x \in \omega ^\omega : p \Vdash \varphi (f(x,\dot y)) \} \in \Gamma .$
Then Main Lemmas 3.4 and 3.14 hold, where “analytic” is replaced by 
 $\Gamma $
.
$\Gamma $
.
Definition 3.23. For 
 $\bar x_0, \dots , \bar x_{n-1} \in \prod _{i<\alpha } X_i$
, we define
$\bar x_0, \dots , \bar x_{n-1} \in \prod _{i<\alpha } X_i$
, we define 
 $$ \begin{align*}\Delta(\bar x_0, \dots, \bar y_{n-1}) := \{ \Delta_{\bar x_i, \bar x_j} : i \neq j < n \} \cup \{ 0, \alpha\},\end{align*} $$
$$ \begin{align*}\Delta(\bar x_0, \dots, \bar y_{n-1}) := \{ \Delta_{\bar x_i, \bar x_j} : i \neq j < n \} \cup \{ 0, \alpha\},\end{align*} $$
where 
 $\Delta _{\bar x_i, \bar x_j} := \min \{ \xi < \alpha : x_i(\xi ) \neq x_j(\xi ) \}$
 if this exists and
$\Delta _{\bar x_i, \bar x_j} := \min \{ \xi < \alpha : x_i(\xi ) \neq x_j(\xi ) \}$
 if this exists and 
 $\Delta _{\bar x_i, \bar x_j} = \alpha $
 if
$\Delta _{\bar x_i, \bar x_j} = \alpha $
 if 
 $\bar x_i = \bar x_j$
.
$\bar x_i = \bar x_j$
.
Remark 3.24. Whenever 
 $\bar x_0, \dots , \bar x_{n-1}$
 are strongly mCg, then they are mCg as witnessed by the partition
$\bar x_0, \dots , \bar x_{n-1}$
 are strongly mCg, then they are mCg as witnessed by the partition 
 $\xi _0 < \dots < \xi _k$
, where
$\xi _0 < \dots < \xi _k$
, where 
 $\{\xi _0, \dots , \xi _k \} = \Delta (\bar x_0, \dots , \bar x_{n-1})$
.
$\{\xi _0, \dots , \xi _k \} = \Delta (\bar x_0, \dots , \bar x_{n-1})$
.
4 Sacks and splitting forcing
4.1 Splitting forcing
Definition 4.1. We say that 
 $S \subseteq 2^{<\omega }$
 is fat if there is
$S \subseteq 2^{<\omega }$
 is fat if there is 
 $m \in \omega $
 so that for all
$m \in \omega $
 so that for all 
 $n \geq m$
, there are
$n \geq m$
, there are 
 $s,t \in S$
 so that
$s,t \in S$
 so that 
 $s(n) =0$
 and
$s(n) =0$
 and 
 $t(n) =1$
. A tree T on
$t(n) =1$
. A tree T on 
 $2$
 is called splitting tree if for every
$2$
 is called splitting tree if for every 
 $s \in T$
,
$s \in T$
, 
 $T_s$
 is fat. We call splitting forcing the tree forcing
$T_s$
 is fat. We call splitting forcing the tree forcing 
 $\mathbb {SP}$
 consisting of splitting trees.
$\mathbb {SP}$
 consisting of splitting trees.
Note that for 
 $T \in \mathbb {SP}$
 and
$T \in \mathbb {SP}$
 and 
 $s \in T$
,
$s \in T$
, 
 $T_s$
 is again a splitting tree. Recall that
$T_s$
 is again a splitting tree. Recall that 
 $x \in 2^\omega $
 is called splitting over V, if for every
$x \in 2^\omega $
 is called splitting over V, if for every 
 $y \in 2^{\omega } \cap V$
,
$y \in 2^{\omega } \cap V$
, 
 $\{ n \in \omega : y(n) = x(n) = 1 \}$
 and
$\{ n \in \omega : y(n) = x(n) = 1 \}$
 and 
 $\{n \in \omega : x(n) = 1 \wedge y(n) =0 \}$
 are infinite. The following is easy to see.
$\{n \in \omega : x(n) = 1 \wedge y(n) =0 \}$
 are infinite. The following is easy to see.
Fact. Let G be 
 $\mathbb {SP}$
-generic over V. Then
$\mathbb {SP}$
-generic over V. Then 
 $x_G$
, the generic real added by
$x_G$
, the generic real added by 
 $\mathbb {SP}$
, is splitting over V.
$\mathbb {SP}$
, is splitting over V.
 Whenever S is fat let us write 
 $m(S)$
 for the minimal
$m(S)$
 for the minimal 
 $m \in \omega $
 witnessing this.
$m \in \omega $
 witnessing this.
Definition 4.2. Let 
 $S,T$
 be splitting trees and
$S,T$
 be splitting trees and 
 $n \in \omega $
. Then we write
$n \in \omega $
. Then we write 
 $S \leq _n T$
 iff
$S \leq _n T$
 iff 
 $S \leq T$
,
$S \leq T$
, 
 $\operatorname {\mathrm {split}}_{\leq n}(S) = \operatorname {\mathrm {split}}_{\leq n}(T)$
 and
$\operatorname {\mathrm {split}}_{\leq n}(S) = \operatorname {\mathrm {split}}_{\leq n}(T)$
 and 
 $\forall s \in \operatorname {\mathrm {split}}_{\leq n}(S) (m(S_s) = m(T_s))$
.
$\forall s \in \operatorname {\mathrm {split}}_{\leq n}(S) (m(S_s) = m(T_s))$
.
Proposition 4.3. The sequence 
 $\langle \leq _n : n \in \omega \rangle $
 witnesses that
$\langle \leq _n : n \in \omega \rangle $
 witnesses that 
 $\mathbb {SP}$
 has Axiom A with continuous reading of names.
$\mathbb {SP}$
 has Axiom A with continuous reading of names.
Proof. It is clear that 
 $\leq _{n}$
 is a partial order refining
$\leq _{n}$
 is a partial order refining 
 $\leq $
 and that
$\leq $
 and that 
 $\leq _{n+1} \subseteq \leq _n$
 for every
$\leq _{n+1} \subseteq \leq _n$
 for every 
 $n \in \omega $
. Let
$n \in \omega $
. Let 
 $\langle T_n : n \in \omega \rangle $
 be a fusion sequence in
$\langle T_n : n \in \omega \rangle $
 be a fusion sequence in 
 $\mathbb {SP}$
, i.e., for every n,
$\mathbb {SP}$
, i.e., for every n, 
 $T_{n+1} \leq _n T_n$
. Then we claim that
$T_{n+1} \leq _n T_n$
. Then we claim that 
 $T := \bigcap _{n \in \omega } T_n$
 is a splitting tree. More precisely, for
$T := \bigcap _{n \in \omega } T_n$
 is a splitting tree. More precisely, for 
 $s \in T$
, we claim that
$s \in T$
, we claim that 
 $m := m((T_{\vert s\vert })_s)$
 witnesses that
$m := m((T_{\vert s\vert })_s)$
 witnesses that 
 $T_s$
 is fat. To see this, let
$T_s$
 is fat. To see this, let 
 $n \geq m$
 be arbitrary and note that
$n \geq m$
 be arbitrary and note that 
 $n \geq m \geq \vert s \vert $
 must be the case. Then, since
$n \geq m \geq \vert s \vert $
 must be the case. Then, since 
 $\operatorname {\mathrm {split}}_{\leq n+1}(T_{n+1}) \subseteq T$
 we have that
$\operatorname {\mathrm {split}}_{\leq n+1}(T_{n+1}) \subseteq T$
 we have that 
 $s \in \operatorname {\mathrm {split}}_{\leq n+1}( T_{n+1})$
 and
$s \in \operatorname {\mathrm {split}}_{\leq n+1}( T_{n+1})$
 and 
 $m((T_{n+1})_s) = m$
. So find
$m((T_{n+1})_s) = m$
. So find 
 $t_0, t_1 \in T_{n+1}$
 so that
$t_0, t_1 \in T_{n+1}$
 so that 
 $t_0(n) = 0$
,
$t_0(n) = 0$
, 
 $t_1(n) = 1$
 and
$t_1(n) = 1$
 and 
 $\vert t_0 \vert = \vert t_1 \vert = n+1$
. But then
$\vert t_0 \vert = \vert t_1 \vert = n+1$
. But then 
 $t_0,t_1 \in T$
, because
$t_0,t_1 \in T$
, because 
 $t_0,t_1 \in \operatorname {\mathrm {split}}_{\leq n+1}(T_{n+1}) \subseteq T$
.
$t_0,t_1 \in \operatorname {\mathrm {split}}_{\leq n+1}(T_{n+1}) \subseteq T$
.
 Now let 
 $D \subseteq \mathbb {SP}$
 be open dense,
$D \subseteq \mathbb {SP}$
 be open dense, 
 $T \in \mathbb {SP}$
 and
$T \in \mathbb {SP}$
 and 
 $n \in \omega $
. We will show that there is
$n \in \omega $
. We will show that there is 
 $S \leq _n T$
 so that for every
$S \leq _n T$
 so that for every 
 $x \in [S]$
, there is
$x \in [S]$
, there is 
 $t \subseteq x$
, with
$t \subseteq x$
, with 
 $S_t \in D$
. This implies condition (3) in Definition 2.1.
$S_t \in D$
. This implies condition (3) in Definition 2.1.
Claim 4.4. Let S be a splitting tree. Then there is 
 $A \subseteq S$
 an antichain (seen as a subset of
$A \subseteq S$
 an antichain (seen as a subset of 
 $2^{<\omega }$
) so that for every
$2^{<\omega }$
) so that for every 
 $k \in \omega , j \in 2$
, if
$k \in \omega , j \in 2$
, if 
 $\exists s \in S (s(k) = j)$
, then
$\exists s \in S (s(k) = j)$
, then 
 $\exists t \in A (t(k) = j)$
.
$\exists t \in A (t(k) = j)$
.
Proof. Start with 
 $\{ s_i : i \in \omega \} \subseteq S$
 an arbitrary infinite antichain and let
$\{ s_i : i \in \omega \} \subseteq S$
 an arbitrary infinite antichain and let 
 $m_i := m(S_{s_i})$
 for every
$m_i := m(S_{s_i})$
 for every 
 $i \in \omega $
. Then find for each
$i \in \omega $
. Then find for each 
 $i \in \omega $
, a finite set
$i \in \omega $
, a finite set 
 $H_i \subseteq S_{s_i}$
 so that for all
$H_i \subseteq S_{s_i}$
 so that for all 
 $k \in [m_i, m_{i+1})$
, there are
$k \in [m_i, m_{i+1})$
, there are 
 $t_0,t_1 \in H_i$
, so that
$t_0,t_1 \in H_i$
, so that 
 $t_0(k) = 0$
 and
$t_0(k) = 0$
 and 
 $t_1(k) = 1$
. Moreover let
$t_1(k) = 1$
. Moreover let 
 $H \subseteq S$
 be finite so that for all
$H \subseteq S$
 be finite so that for all 
 $k \in [0,m_0)$
 and
$k \in [0,m_0)$
 and 
 $j \in 2$
, if
$j \in 2$
, if 
 $\exists s \in S (s(k) =j)$
, then
$\exists s \in S (s(k) =j)$
, then 
 $\exists t \in H (t(k) = j)$
. Then define
$\exists t \in H (t(k) = j)$
. Then define 
 $F_i = H_i \cup (H \cap S_{s_i})$
 for each
$F_i = H_i \cup (H \cap S_{s_i})$
 for each 
 $i \in \omega $
 and let
$i \in \omega $
 and let 
 $F_{-1} := H \setminus \bigcup _{i \in \omega } F_i$
. Since
$F_{-1} := H \setminus \bigcup _{i \in \omega } F_i$
. Since 
 $F_i$
 is finite for every
$F_i$
 is finite for every 
 $i \in \omega $
, it is easy to extend each of its elements to get a set
$i \in \omega $
, it is easy to extend each of its elements to get a set 
 $F_i'$
 that is additionally an antichain in
$F_i'$
 that is additionally an antichain in 
 $S_{s_i}$
. Also extend the elements of
$S_{s_i}$
. Also extend the elements of 
 $F_{-1}$
 to get an antichain
$F_{-1}$
 to get an antichain 
 $F_{-1}'$
 in S. It is easy to see that
$F_{-1}'$
 in S. It is easy to see that 
 $A := \bigcup _{i \in [-1,\omega )} F_i'$
 works.⊣
$A := \bigcup _{i \in [-1,\omega )} F_i'$
 works.⊣
 Now enumerate 
 $\operatorname {\mathrm {split}}_n(T)$
 as
$\operatorname {\mathrm {split}}_n(T)$
 as 
 $\langle \sigma _i : i < N \rangle $
,
$\langle \sigma _i : i < N \rangle $
, 
 $N := 2^n$
. For each
$N := 2^n$
. For each 
 $i < N$
, let
$i < N$
, let 
 $A_i \subseteq T_{\sigma _i}$
 be an antichain as in the claim applied to
$A_i \subseteq T_{\sigma _i}$
 be an antichain as in the claim applied to 
 $S = T_{\sigma _i}$
. For every
$S = T_{\sigma _i}$
. For every 
 $i < N$
 and
$i < N$
 and 
 $t \in A_i$
, let
$t \in A_i$
, let 
 $S^{t} \in D$
 be so that
$S^{t} \in D$
 be so that 
 $S^t \leq T_t$
. For every
$S^t \leq T_t$
. For every 
 $i <N$
 pick
$i <N$
 pick 
 $t_i \in A_i$
 arbitrarily and
$t_i \in A_i$
 arbitrarily and 
 $F_i \subseteq A_i$
 a finite set so that for every
$F_i \subseteq A_i$
 a finite set so that for every 
 $k \in [0, m(S^{t_i}))$
 and
$k \in [0, m(S^{t_i}))$
 and 
 $j \in 2$
, if
$j \in 2$
, if 
 $\exists s \in A_i (s(k)= j)$
, then
$\exists s \in A_i (s(k)= j)$
, then 
 $ \exists t \in F_i (t(k)=j)$
. Then we see that
$ \exists t \in F_i (t(k)=j)$
. Then we see that 
 $S := \bigcup _{i < N} (\bigcup _{t \in F_i} S^t \cup S^{t_i} )$
 works. We constructed S so that
$S := \bigcup _{i < N} (\bigcup _{t \in F_i} S^t \cup S^{t_i} )$
 works. We constructed S so that 
 $S \leq _n T$
. Moreover, whenever
$S \leq _n T$
. Moreover, whenever 
 $x \in [S]$
, then there is
$x \in [S]$
, then there is 
 $i < N$
 be so that
$i < N$
 be so that 
 $\sigma _i \subseteq x$
. Then
$\sigma _i \subseteq x$
. Then 
 $x \in [\bigcup _{t \in F_i} S_t \cup S_{t_i}]$
 and since
$x \in [\bigcup _{t \in F_i} S_t \cup S_{t_i}]$
 and since 
 $F_i$
 is finite, there is
$F_i$
 is finite, there is 
 $t \in F_i \cup \{t_i \}$
 so that
$t \in F_i \cup \{t_i \}$
 so that 
 $t \subseteq x$
. But then
$t \subseteq x$
. But then 
 $S_{t} \leq S^t \in D$
.
$S_{t} \leq S^t \in D$
.
 Finally, in order to show the continuous reading of names, let 
 $\dot y$
 be a name for an element of
$\dot y$
 be a name for an element of 
 $\omega ^\omega $
,
$\omega ^\omega $
, 
 $n \in \omega $
, and
$n \in \omega $
, and 
 $T \in \mathbb {SP}$
. It suffices to consider such names, since for every Polish space X, there is a continuous surjection
$T \in \mathbb {SP}$
. It suffices to consider such names, since for every Polish space X, there is a continuous surjection 
 $F \colon \omega ^\omega \to X$
. Then we have that for each
$F \colon \omega ^\omega \to X$
. Then we have that for each 
 $i \in \omega $
,
$i \in \omega $
, 
 $D_i := \{ S \in \mathbb {SP} : \exists s \in \omega ^{i}( S \Vdash \dot y \restriction i = s ) \}$
 is dense open. Let
$D_i := \{ S \in \mathbb {SP} : \exists s \in \omega ^{i}( S \Vdash \dot y \restriction i = s ) \}$
 is dense open. Let 
 $\langle T_i : i \in \omega \rangle $
 be so that
$\langle T_i : i \in \omega \rangle $
 be so that 
 $T_0 \leq _n T$
,
$T_0 \leq _n T$
, 
 $T_{i+1} \leq _{n+i} T_i$
 and for every
$T_{i+1} \leq _{n+i} T_i$
 and for every 
 $x \in [T_i]$
, there is
$x \in [T_i]$
, there is 
 $t \subseteq x$
 so that
$t \subseteq x$
 so that 
 $(T_i)_t \in D_i$
. Then
$(T_i)_t \in D_i$
. Then 
 $S = \bigcap _{i \in \omega } T_i \leq _n T$
. For every
$S = \bigcap _{i \in \omega } T_i \leq _n T$
. For every 
 $x \in [S]$
, define
$x \in [S]$
, define 
 $f(x) = \bigcup \{ s \in \omega ^{<\omega } : \exists t \subseteq x (S_t \Vdash s \subseteq \dot y) \}$
. Then
$f(x) = \bigcup \{ s \in \omega ^{<\omega } : \exists t \subseteq x (S_t \Vdash s \subseteq \dot y) \}$
. Then 
 $f \colon [S] \to \omega ^\omega $
 is continuous and
$f \colon [S] \to \omega ^\omega $
 is continuous and 
 $S \Vdash \dot y = f(x_G)$
.⊣
$S \Vdash \dot y = f(x_G)$
.⊣
Corollary 4.5. 
 $\mathbb {SP}$
 is proper and
$\mathbb {SP}$
 is proper and 
 $\omega ^\omega $
-bounding.
$\omega ^\omega $
-bounding.
4.2 Weighted tree forcing
In this subsection we define a class of forcings that we call weighted tree forcing. The definition is slightly ad-hoc, but simple enough to formulate and to check for various forcing notions. In earlier versions of this paper we proved many of the results only for splitting forcing, but we noted that similar combinatorial arguments apply more generally. The notion of a weight resulted directly from analysing the proof of Proposition 4.16 for splitting forcing. It also turned out to be particularly helpful in the next subsection.
Definition 4.6. Let T be a perfect tree. A weight on T is a map 
 $\rho \colon T \times T \to [T]^{<\omega }$
 so that
$\rho \colon T \times T \to [T]^{<\omega }$
 so that 
 $\rho (s,t) \subseteq T_s \setminus T_t$
 for all
$\rho (s,t) \subseteq T_s \setminus T_t$
 for all 
 $s,t \in T$
. Whenever
$s,t \in T$
. Whenever 
 $\rho _0, \rho _1$
 are weights on T we write
$\rho _0, \rho _1$
 are weights on T we write 
 $\rho _0 \subseteq \rho _1$
 to say that for all
$\rho _0 \subseteq \rho _1$
 to say that for all 
 $s,t \in T$
,
$s,t \in T$
, 
 $\rho _0(s,t) \subseteq \rho _1(s,t)$
.
$\rho _0(s,t) \subseteq \rho _1(s,t)$
.
 Note that if 
 $t \subseteq s$
 then
$t \subseteq s$
 then 
 $\rho (s,t) = \emptyset $
 must be the case.
$\rho (s,t) = \emptyset $
 must be the case.
Definition 4.7. Let T be a perfect tree, 
 $\rho $
 a weight on T, and S a tree. Then we write
$\rho $
 a weight on T, and S a tree. Then we write 
 $S \leq _\rho T$
 if
$S \leq _\rho T$
 if 
 $S \subseteq T$
 and there is a dense set of
$S \subseteq T$
 and there is a dense set of 
 $s_0 \in S$
 with an injective sequence
$s_0 \in S$
 with an injective sequence 
 $(s_n)_{n\in \omega }$
 in
$(s_n)_{n\in \omega }$
 in 
 $S_{s_0}$
 such that
$S_{s_0}$
 such that 
 $\forall n \in \omega (\rho (s_n, s_{n+1}) \subseteq S)$
.
$\forall n \in \omega (\rho (s_n, s_{n+1}) \subseteq S)$
.
Remark 4.8. Whenever 
 $\rho _0 \subseteq \rho _1$
, we have that
$\rho _0 \subseteq \rho _1$
, we have that 
 $S \leq _{\rho _1} T$
 implies
$S \leq _{\rho _1} T$
 implies 
 $S \leq _{\rho _0} T$
.
$S \leq _{\rho _0} T$
.
Definition 4.9. Let 
 $\mathbb {P}$
 be a tree forcing. Then we say that
$\mathbb {P}$
 be a tree forcing. Then we say that 
 $\mathbb {P}$
 is weighted if for any
$\mathbb {P}$
 is weighted if for any 
 $T \in \mathbb {P}$
 there is a weight
$T \in \mathbb {P}$
 there is a weight 
 $\rho $
 on T so that for any tree S, if
$\rho $
 on T so that for any tree S, if 
 $S \leq _\rho T$
 then
$S \leq _\rho T$
 then 
 $S \in \mathbb {P}$
.
$S \in \mathbb {P}$
.
Lemma 4.10. 
 $\mathbb {SP}$
 is weighted.
$\mathbb {SP}$
 is weighted.
Proof. Let 
 $T \in \mathbb {SP}$
. For any
$T \in \mathbb {SP}$
. For any 
 $s,t \in T$
 let
$s,t \in T$
 let 
 $\rho (s,t) \subseteq T_s \setminus T_t$
 be finite so that for any
$\rho (s,t) \subseteq T_s \setminus T_t$
 be finite so that for any 
 $k \in \omega $
 and
$k \in \omega $
 and 
 $i \in 2$
, if there is
$i \in 2$
, if there is 
 $r \in T_s$
 so that
$r \in T_s$
 so that 
 $r(k) = i$
 and there is no such
$r(k) = i$
 and there is no such 
 $r \in T_t$
, then there is such r in
$r \in T_t$
, then there is such r in 
 $\rho (s,t)$
. This is possible since
$\rho (s,t)$
. This is possible since 
 $T_t$
 is fat. Let us show that
$T_t$
 is fat. Let us show that 
 $\rho $
 works. Assume that
$\rho $
 works. Assume that 
 $S \leq _\rho T$
 and let
$S \leq _\rho T$
 and let 
 $s \in S$
 be arbitrary. Then there is
$s \in S$
 be arbitrary. Then there is 
 $s_0 \supseteq s$
 in S with a sequence
$s_0 \supseteq s$
 in S with a sequence 
 $(s_n)_{n\in \omega }$
 as in the definition of
$(s_n)_{n\in \omega }$
 as in the definition of 
 $\leq _\rho $
. Let
$\leq _\rho $
. Let 
 $k \geq m(T_{s_0})$
 and
$k \geq m(T_{s_0})$
 and 
 $i \in 2$
 and suppose there is no
$i \in 2$
 and suppose there is no 
 $r \in S_{s_0}$
 with
$r \in S_{s_0}$
 with 
 $r(k) = i$
. In particular this means that no such r is in
$r(k) = i$
. In particular this means that no such r is in 
 $\rho (s_n,s_{n+1})$
 for any
$\rho (s_n,s_{n+1})$
 for any 
 $n \in \omega $
, since
$n \in \omega $
, since 
 $\rho (s_n, s_{n+1}) \subseteq S_{s_0}$
. But then, using the definition of
$\rho (s_n, s_{n+1}) \subseteq S_{s_0}$
. But then, using the definition of 
 $\rho $
 and
$\rho $
 and 
 $m(T_{s_0})$
, we see inductively that for each
$m(T_{s_0})$
, we see inductively that for each 
 $n \in \omega $
 such r must be found in
$n \in \omega $
 such r must be found in 
 $T_{s_n}$
. Letting n large enough so that
$T_{s_n}$
. Letting n large enough so that 
 $k < \vert s_n \vert $
,
$k < \vert s_n \vert $
, 
 $s_n(k) = i$
 must be the case. But
$s_n(k) = i$
 must be the case. But 
 $s_n \in S_{s_0}$
, which is a contradiction.⊣
$s_n \in S_{s_0}$
, which is a contradiction.⊣
Definition 4.11. 
Sacks forcing is the tree forcing 
 $\mathbb {S}$
 consisting of all perfect subtrees of
$\mathbb {S}$
 consisting of all perfect subtrees of 
 $2^{<\omega }$
. It is well-known that it is Axiom A with continuous reading of names.
$2^{<\omega }$
. It is well-known that it is Axiom A with continuous reading of names.
Lemma 4.12. 
 $\mathbb {S}$
 is weighted.
$\mathbb {S}$
 is weighted.
Proof. Let 
 $T \in \mathbb {S}$
. For
$T \in \mathbb {S}$
. For 
 $s,t \in T$
, we let
$s,t \in T$
, we let 
 $\rho (s,t)$
 contain all
$\rho (s,t)$
 contain all 
 $r^{\frown } i \in T_s \setminus T_t$
 such that
$r^{\frown } i \in T_s \setminus T_t$
 such that 
 $r^{\frown } (1-i) \in T$
 and where
$r^{\frown } (1-i) \in T$
 and where 
 $\vert r \vert $
 is minimal with this property.⊣
$\vert r \vert $
 is minimal with this property.⊣
 Recall that for finite trees 
 $T_0$
,
$T_0$
, 
 $T_1$
 we say that
$T_1$
 we say that 
 $T_1$
 is an end-extension of
$T_1$
 is an end-extension of 
 $T_0$
, written as
$T_0$
, written as 
 $T_0 \sqsubset T_1$
, if
$T_0 \sqsubset T_1$
, if 
 $T_0 \subsetneq T_1$
 and for every
$T_0 \subsetneq T_1$
 and for every 
 $t \in T_1 \setminus T_0$
 there is a terminal node
$t \in T_1 \setminus T_0$
 there is a terminal node 
 $\sigma \in \operatorname {\mathrm {term}} (T_0)$
 so that
$\sigma \in \operatorname {\mathrm {term}} (T_0)$
 so that 
 $\sigma \subseteq t$
. A node
$\sigma \subseteq t$
. A node 
 $\sigma \in T_0$
 is called terminal if it has no proper extension in
$\sigma \in T_0$
 is called terminal if it has no proper extension in 
 $T_0$
.
$T_0$
.
Definition 4.13. Let T be a perfect tree, 
 $\rho $
 a weight on T, and
$\rho $
 a weight on T, and 
 $T_0, T_1$
 finite subtrees of T. Then we write
$T_0, T_1$
 finite subtrees of T. Then we write 
 $T_0 \lhd _\rho T_1$
 iff
$T_0 \lhd _\rho T_1$
 iff 
 $T_0 \sqsubset T_1$
 and
$T_0 \sqsubset T_1$
 and 
 $$\begin{align} \forall \sigma \in \operatorname{\mathrm{term}}(T_0) \exists N &\geq 2 \exists \langle s_i\rangle_{i < N} \in ((T_{1})_\sigma)^N \text{ injective } \nonumber\\ &\bigl( s_0 = \sigma \wedge s_{N-1} \in \operatorname{\mathrm{term}}(T_{1}) \wedge \forall i < N (\rho(s_i, s_{i+1}) \subseteq T_{1}) \bigr). \end{align}$$
$$\begin{align} \forall \sigma \in \operatorname{\mathrm{term}}(T_0) \exists N &\geq 2 \exists \langle s_i\rangle_{i < N} \in ((T_{1})_\sigma)^N \text{ injective } \nonumber\\ &\bigl( s_0 = \sigma \wedge s_{N-1} \in \operatorname{\mathrm{term}}(T_{1}) \wedge \forall i < N (\rho(s_i, s_{i+1}) \subseteq T_{1}) \bigr). \end{align}$$
Lemma 4.14. Let T be a perfect tree, 
 $\rho $
 be a weight on T, and
$\rho $
 be a weight on T, and 
 $\langle T_n : n \in \omega \rangle $
 be a sequence of finite subtrees of T so that
$\langle T_n : n \in \omega \rangle $
 be a sequence of finite subtrees of T so that 
 $T_n \lhd _\rho T_{n+1}$
 for every
$T_n \lhd _\rho T_{n+1}$
 for every 
 $n \in \omega $
. Then
$n \in \omega $
. Then 
 $\bigcup _{n \in \omega } T_n \leq _\rho T$
.
$\bigcup _{n \in \omega } T_n \leq _\rho T$
.
Proof. Let 
 $S := \bigcup _{n \in \omega } T_n$
. To see that
$S := \bigcup _{n \in \omega } T_n$
. To see that 
 $S \leq _\rho T$
 note that
$S \leq _\rho T$
 note that 
 $\bigcup _{n\in \omega } \operatorname {\mathrm {term}}(T_n)$
 is dense in S, in a very strong sense. Let
$\bigcup _{n\in \omega } \operatorname {\mathrm {term}}(T_n)$
 is dense in S, in a very strong sense. Let 
 $\sigma \in \operatorname {\mathrm {term}}(T_n)$
 for some
$\sigma \in \operatorname {\mathrm {term}}(T_n)$
 for some 
 $n \in \omega $
, then let
$n \in \omega $
, then let 
 $s_0, \dots , s_{N_0-1}$
 be as in
$s_0, \dots , s_{N_0-1}$
 be as in 
 $(*_0)$
 for
$(*_0)$
 for 
 $T_n,T_{n+1}$
. Since
$T_n,T_{n+1}$
. Since 
 $s_{N_0-1} \in \operatorname {\mathrm {term}}(T_{n+1})$
 we again find
$s_{N_0-1} \in \operatorname {\mathrm {term}}(T_{n+1})$
 we again find 
 $s_{N_0}, \dots , s_{N_1-1}$
 so that
$s_{N_0}, \dots , s_{N_1-1}$
 so that 
 $s_{N_0-1}, \dots , s_{N_1-1}$
 is as in
$s_{N_0-1}, \dots , s_{N_1-1}$
 is as in 
 $(*_0)$
 for
$(*_0)$
 for 
 $T_{n+1}, T_{n+2}$
. Continuing like this, we find a sequence
$T_{n+1}, T_{n+2}$
. Continuing like this, we find a sequence 
 $\langle s_i : i \in \omega \rangle $
 in S starting with
$\langle s_i : i \in \omega \rangle $
 in S starting with 
 $s_0 = \sigma $
 so that
$s_0 = \sigma $
 so that 
 $\rho (s_i, s_{i+1}) \subseteq S$
 for all
$\rho (s_i, s_{i+1}) \subseteq S$
 for all 
 $i \in \omega $
, as required.⊣
$i \in \omega $
, as required.⊣
Lemma 4.15. Let T be a perfect tree, 
 $\rho $
 a weight on T, and
$\rho $
 a weight on T, and 
 $T_0$
 a finite subtree of T. Moreover, let
$T_0$
 a finite subtree of T. Moreover, let 
 $k \in \omega $
 and
$k \in \omega $
 and 
 $D \subseteq (T)^k$
 be dense open. Then there is
$D \subseteq (T)^k$
 be dense open. Then there is 
 $T_1 \rhd _\rho T_0$
 so that
$T_1 \rhd _\rho T_0$
 so that 
 $$\begin{align} \forall \{\sigma_0, \dots, \sigma_{k -1}\} \in [\operatorname{\mathrm{term}}(T_{0})]^{k} \forall \sigma_0', \dots, \sigma_{k -1}' \in \operatorname{\mathrm{term}}(T_{1}) \nonumber\\ \left(\forall l < k (\sigma_l \subseteq \sigma_l') \rightarrow (\sigma_0', \dots, \sigma_{k -1}') \in D\right). \end{align}$$
$$\begin{align} \forall \{\sigma_0, \dots, \sigma_{k -1}\} \in [\operatorname{\mathrm{term}}(T_{0})]^{k} \forall \sigma_0', \dots, \sigma_{k -1}' \in \operatorname{\mathrm{term}}(T_{1}) \nonumber\\ \left(\forall l < k (\sigma_l \subseteq \sigma_l') \rightarrow (\sigma_0', \dots, \sigma_{k -1}') \in D\right). \end{align}$$
Proof. First let us enumerate 
 $\operatorname {\mathrm {term}}(T_0)$
 by
$\operatorname {\mathrm {term}}(T_0)$
 by 
 $\sigma _0, \dots , \sigma _{K-1}$
. We put
$\sigma _0, \dots , \sigma _{K-1}$
. We put 
 $s_0^l = \sigma _{l}$
 for each
$s_0^l = \sigma _{l}$
 for each 
 $l < K$
. Next find for each
$l < K$
. Next find for each 
 $l < K$
,
$l < K$
, 
 $s_1^l \in T, s_0^l \subsetneq s_1^l$
 above a splitting node in
$s_1^l \in T, s_0^l \subsetneq s_1^l$
 above a splitting node in 
 $T_{s_0^l}$
. Moreover we find
$T_{s_0^l}$
. Moreover we find 
 $s_2^l \in T_{s_0^l}$
 so that
$s_2^l \in T_{s_0^l}$
 so that 
 $s_2^l \perp s_1^l$
 and
$s_2^l \perp s_1^l$
 and 
 $s_2^l$
 is longer than any node appearing in
$s_2^l$
 is longer than any node appearing in 
 $\rho (s_0^l, s_1^l)$
. This is possible since we chose
$\rho (s_0^l, s_1^l)$
. This is possible since we chose 
 $s_1^l$
 to be above a splitting node in
$s_1^l$
 to be above a splitting node in 
 $T_{s_0^l}$
. For each
$T_{s_0^l}$
. For each 
 $l<K$
 we let
$l<K$
 we let 
 $\tilde {T}_{2}^l$
 be the tree generated by (i.e., the downwards closure of)
$\tilde {T}_{2}^l$
 be the tree generated by (i.e., the downwards closure of) 
 $\{s_1^l, s_2^l\} \cup \rho (s_0^l,s_1^l) \cup \rho (s_1^l,s_2^l)$
. Note that
$\{s_1^l, s_2^l\} \cup \rho (s_0^l,s_1^l) \cup \rho (s_1^l,s_2^l)$
. Note that 
 $s_2^l \in \operatorname {\mathrm {term}}(\tilde {T}_{2}^l)$
 as
$s_2^l \in \operatorname {\mathrm {term}}(\tilde {T}_{2}^l)$
 as 
 $\rho (s_1^l,s_2^l) \perp s_2^l$
.
$\rho (s_1^l,s_2^l) \perp s_2^l$
.
 Let us enumerate by 
 $(f_j)_{2\leq j < N}$
 all functions
$(f_j)_{2\leq j < N}$
 all functions 
 $f \colon K \to \{1,2\}$
 starting with
$f \colon K \to \{1,2\}$
 starting with 
 $f_2$
 the constant function mapping to
$f_2$
 the constant function mapping to 
 $1$
. We are going to construct recursively a sequence
$1$
. We are going to construct recursively a sequence 
 $\langle \tilde {T}_j^l : 2 \leq j \leq N \rangle $
 where
$\langle \tilde {T}_j^l : 2 \leq j \leq N \rangle $
 where 
 $\tilde {T}_j^l \sqsubseteq \tilde {T}_{j+1}^l$
, and
$\tilde {T}_j^l \sqsubseteq \tilde {T}_{j+1}^l$
, and 
 $\langle s_j^l : 2 \leq j \leq N \rangle $
 without repetitions, for each
$\langle s_j^l : 2 \leq j \leq N \rangle $
 without repetitions, for each 
 $l< K$
 such that at any step
$l< K$
 such that at any step 
 $j <N$
:
$j <N$
:
- 
(1) For every  $l < K$
, $l < K$
, $s_j^l \in \operatorname {\mathrm {term}}(\tilde {T}_{j}^l)$
 and $s_j^l \in \operatorname {\mathrm {term}}(\tilde {T}_{j}^l)$
 and $\begin {cases} s_2^l \subseteq s_j^l, & \text {if } f_j(l) = 1, \\ s_1^l \subseteq s_j^l, &\text {if } f_j(l) = 2. \end {cases}$ $\begin {cases} s_2^l \subseteq s_j^l, & \text {if } f_j(l) = 1, \\ s_1^l \subseteq s_j^l, &\text {if } f_j(l) = 2. \end {cases}$
- 
(2) For any  $\{l_i : i <k \} \in [K]^k$
 and $\{l_i : i <k \} \in [K]^k$
 and $(t_i)_{i<k}$
 where $(t_i)_{i<k}$
 where $t_i \in \operatorname {\mathrm {term}}(\tilde {T}_{j+1}^{l_i})$
 and $t_i \in \operatorname {\mathrm {term}}(\tilde {T}_{j+1}^{l_i})$
 and $\begin {cases} s_1^{l_i} \subseteq t_{i}, & \text {if } f_j(l_i) = 1, \\ s_1^{l_i} \perp t_{i}, &\text {if } f_j(l_i) = 2, \end {cases}$
 for every $\begin {cases} s_1^{l_i} \subseteq t_{i}, & \text {if } f_j(l_i) = 1, \\ s_1^{l_i} \perp t_{i}, &\text {if } f_j(l_i) = 2, \end {cases}$
 for every $i< k$
, $i< k$
, $(t_0, \dots , t_{k-1}) \in D$
. $(t_0, \dots , t_{k-1}) \in D$
.
- 
(3) For every  $l < K$
, $l < K$
, $\rho (s_j^l, s_{j+1}^l ) \subseteq \tilde {T}_{j+1}^{l}$
. $\rho (s_j^l, s_{j+1}^l ) \subseteq \tilde {T}_{j+1}^{l}$
.
 Note that 
 $(1)$
 holds true at the initial step
$(1)$
 holds true at the initial step 
 $j=2$
 since
$j=2$
 since 
 $f_2(l) = 1$
,
$f_2(l) = 1$
, 
 $s_2^l \subseteq s_2^l$
, and
$s_2^l \subseteq s_2^l$
, and 
 $s_2^l \in \operatorname {\mathrm {term}}(\tilde {T}_2^l)$
 for each
$s_2^l \in \operatorname {\mathrm {term}}(\tilde {T}_2^l)$
 for each 
 $l < K$
. Now suppose that for some
$l < K$
. Now suppose that for some 
 $j < N$
 we have constructed
$j < N$
 we have constructed 
 $\tilde {T}_j^l$
 and
$\tilde {T}_j^l$
 and 
 $s_j^l$
 for each l with
$s_j^l$
 for each l with 
 $(1)$
 holding true. Then we proceed as follows. Let
$(1)$
 holding true. Then we proceed as follows. Let 
 $\{ t^l_i : i < N_l\}$
 enumerate
$\{ t^l_i : i < N_l\}$
 enumerate 
 $\{ t : t\in \operatorname {\mathrm {term}}(\tilde {T}_j^l) \wedge s_1^l \subseteq t \text { if } f_j(l) = 1 \wedge s_1^l \perp t \text { if } f_j(l) = 2\} $
 for each
$\{ t : t\in \operatorname {\mathrm {term}}(\tilde {T}_j^l) \wedge s_1^l \subseteq t \text { if } f_j(l) = 1 \wedge s_1^l \perp t \text { if } f_j(l) = 2\} $
 for each 
 $l <K$
. Now it is simple to find
$l <K$
. Now it is simple to find 
 $r^l_i \in T$
,
$r^l_i \in T$
, 
 $t^l_i \subseteq r^l_i$
 for each
$t^l_i \subseteq r^l_i$
 for each 
 $i <N_l, l< K$
 so that
$i <N_l, l< K$
 so that 
 $[\{r^l_i : i < N_l, l< K \}]^k \subseteq D$
.
$[\{r^l_i : i < N_l, l< K \}]^k \subseteq D$
.
 Let 
 $R_l$
 be the tree generated by
$R_l$
 be the tree generated by 
 $\tilde {T}_{j}^l$
 and
$\tilde {T}_{j}^l$
 and 
 $\{r^l_i : i < N_l\}$
 for each
$\{r^l_i : i < N_l\}$
 for each 
 $l <K$
. It is easy to see that
$l <K$
. It is easy to see that 
 $\tilde {T}_{j}^l \sqsubseteq R_l$
 since we only extended elements from
$\tilde {T}_{j}^l \sqsubseteq R_l$
 since we only extended elements from 
 $\operatorname {\mathrm {term}}(\tilde {T}_j^l)$
 (namely the
$\operatorname {\mathrm {term}}(\tilde {T}_j^l)$
 (namely the 
 $t_i^l$
’s). Note that it is still the case that
$t_i^l$
’s). Note that it is still the case that 
 $s_j^l \in \operatorname {\mathrm {term}} (R_l)$
 since
$s_j^l \in \operatorname {\mathrm {term}} (R_l)$
 since 
 $s_j^l \perp t_i^l$
 for all
$s_j^l \perp t_i^l$
 for all 
 $i < N_l$
. Next we choose
$i < N_l$
. Next we choose 
 $s_{j+1}^l$
 extending an element of
$s_{j+1}^l$
 extending an element of 
 $\operatorname {\mathrm {term}} (R_l)$
, distinct from all previous choices and so that
$\operatorname {\mathrm {term}} (R_l)$
, distinct from all previous choices and so that 
 $s_2^l \subseteq s_j^l \text { if } f_{j+1}(l) = 1$
 and
$s_2^l \subseteq s_j^l \text { if } f_{j+1}(l) = 1$
 and 
 $s_1^l \subseteq s_j^l \text { if } f_{j+1}(l) = 2.$
$s_1^l \subseteq s_j^l \text { if } f_{j+1}(l) = 2.$
 Taking 
 $\tilde {T}_{j+1}^l$
 to be the tree generated by
$\tilde {T}_{j+1}^l$
 to be the tree generated by 
 $R_l \cup \{ s_{j+1}^l \} \cup \rho (s_j^l, s_{j+1}^{l})$
 gives the next step of the construction. Again
$R_l \cup \{ s_{j+1}^l \} \cup \rho (s_j^l, s_{j+1}^{l})$
 gives the next step of the construction. Again 
 $R_l \sqsubseteq \tilde {T}_{j+1}^l$
, as we only extended terminal nodes of
$R_l \sqsubseteq \tilde {T}_{j+1}^l$
, as we only extended terminal nodes of 
 $R_l$
. Then
$R_l$
. Then 
 $(3)$
 obviously holds true and
$(3)$
 obviously holds true and 
 $s^l_{j +1} \in \operatorname {\mathrm {term}}(\tilde {T}_{j+1}^l)$
 since
$s^l_{j +1} \in \operatorname {\mathrm {term}}(\tilde {T}_{j+1}^l)$
 since 
 $\rho (s_j^l, s_{j+1}^{l}) \perp s^l_{j +1}$
. It follows from the construction that
$\rho (s_j^l, s_{j+1}^{l}) \perp s^l_{j +1}$
. It follows from the construction that 
 $(2)$
 holds true for each
$(2)$
 holds true for each 
 $\tilde {T}_{j+1}^l$
 replaced by
$\tilde {T}_{j+1}^l$
 replaced by 
 $R_l$
. Since
$R_l$
. Since 
 $R_l \sqsubseteq \tilde {T}_{j+1}^l$
 we easily see that
$R_l \sqsubseteq \tilde {T}_{j+1}^l$
 we easily see that 
 $(2)$
 is satisfied.
$(2)$
 is satisfied.
 Finally we put 
 $T_{1} = \bigcup _{l<K} \tilde T_{N}^l$
. It is clear that
$T_{1} = \bigcup _{l<K} \tilde T_{N}^l$
. It is clear that 
 $(*_0)$
 is true, in particular that
$(*_0)$
 is true, in particular that 
 $T_0 \lhd _\rho T_1$
. For
$T_0 \lhd _\rho T_1$
. For 
 $(*_1)$
 let
$(*_1)$
 let 
 $\{l_i : i<k \} \in [K]^k$
 be arbitrary and assume that
$\{l_i : i<k \} \in [K]^k$
 be arbitrary and assume that 
 $t_i \in \operatorname {\mathrm {term}} (\tilde T_{N}^{l_i})$
 for each
$t_i \in \operatorname {\mathrm {term}} (\tilde T_{N}^{l_i})$
 for each 
 $i < k$
. Let
$i < k$
. Let 
 $f \colon K \to \{1,2\}$
 be so that for each
$f \colon K \to \{1,2\}$
 be so that for each 
 $i < k$
 if
$i < k$
 if 
 $s_1^{l_i} \subseteq t_i$
 then
$s_1^{l_i} \subseteq t_i$
 then 
 $f(l_i) = 1$
, and if
$f(l_i) = 1$
, and if 
 $s_1^{l_i} \perp t_i$
 then
$s_1^{l_i} \perp t_i$
 then 
 $f(l_i) = 2$
. Then there is
$f(l_i) = 2$
. Then there is 
 $j \in [2,N)$
 so that
$j \in [2,N)$
 so that 
 $f_j = f$
. Clause
$f_j = f$
. Clause 
 $(2)$
 ensured that for initial segments
$(2)$
 ensured that for initial segments 
 $t_i' \subseteq t_i$
 where
$t_i' \subseteq t_i$
 where 
 $t_i' \in \operatorname {\mathrm {term}}(\tilde {T}_{j+1}^{l_i})$
,
$t_i' \in \operatorname {\mathrm {term}}(\tilde {T}_{j+1}^{l_i})$
, 
 $(t_0',\dots , t_{k-1}' ) \in D$
. In particular
$(t_0',\dots , t_{k-1}' ) \in D$
. In particular 
 $(t_0,\dots , t_{k-1}) \in D$
 which proves
$(t_0,\dots , t_{k-1}) \in D$
 which proves 
 $(*_1)$
.⊣
$(*_1)$
.⊣
Proposition 4.16. Let M be a countable model of set theory, 
 $R_l \in M$
 a perfect tree, and
$R_l \in M$
 a perfect tree, and 
 $\rho _l$
 a weight on
$\rho _l$
 a weight on 
 $R_l$
 for every
$R_l$
 for every 
 $l < k \in \omega $
. Then there is
$l < k \in \omega $
. Then there is 
 $S_l \leq _{\rho _l} R_l$
 for every
$S_l \leq _{\rho _l} R_l$
 for every 
 $l < k$
 so that any
$l < k$
 so that any 
 $\bar x_0, \dots , \bar x_{n-1} \in \prod _{l < k}[S_l]$
 are
$\bar x_0, \dots , \bar x_{n-1} \in \prod _{l < k}[S_l]$
 are 
 $\langle [R_l] : l < k\rangle $
-mutually Cohen generic over M.
$\langle [R_l] : l < k\rangle $
-mutually Cohen generic over M.
Proof. Let 
 $T := \{ \emptyset \} \cup \{ \langle l \rangle ^{\frown } s : s \in R_l, l<k \}$
 be the disjoint sum of the trees
$T := \{ \emptyset \} \cup \{ \langle l \rangle ^{\frown } s : s \in R_l, l<k \}$
 be the disjoint sum of the trees 
 $R_l$
 for
$R_l$
 for 
 $l < k$
. Also let
$l < k$
. Also let 
 $\rho $
 be a weight on T extending arbitrarily the weights
$\rho $
 be a weight on T extending arbitrarily the weights 
 $\rho _l$
 defined on the copy of
$\rho _l$
 defined on the copy of 
 $R_l$
 in T. As M is countable, let
$R_l$
 in T. As M is countable, let 
 $(D_n,k_n)_{n\in \omega }$
 enumerate all pairs
$(D_n,k_n)_{n\in \omega }$
 enumerate all pairs 
 $(D,m) \in M$
, such that D is a dense open subset of
$(D,m) \in M$
, such that D is a dense open subset of 
 $T^m$
 and
$T^m$
 and 
 $m \in \omega \setminus \{0\}$
, infinitely often. Let us find a sequence
$m \in \omega \setminus \{0\}$
, infinitely often. Let us find a sequence 
 $(T_n)_{n\in \omega }$
 of finite subtrees of T, such that for each
$(T_n)_{n\in \omega }$
 of finite subtrees of T, such that for each 
 $n\in \omega $
,
$n\in \omega $
, 
 $T_n \lhd _\rho T_{n+1}$
 and
$T_n \lhd _\rho T_{n+1}$
 and 
 $$\begin{align} \forall \{\sigma_0, \dots, \sigma_{k_n -1}\} \in [\operatorname{\mathrm{term}}(T_{n})]^{k_n} \forall \sigma_0', \dots, \sigma_{k_n -1}' \in \operatorname{\mathrm{term}}(T_{n+1}) \nonumber\\ [\forall l < k (\sigma_l \subseteq \sigma_l') \rightarrow (\sigma_0', \dots, \sigma_{k_n -1}') \in D_n]. \end{align}$$
$$\begin{align} \forall \{\sigma_0, \dots, \sigma_{k_n -1}\} \in [\operatorname{\mathrm{term}}(T_{n})]^{k_n} \forall \sigma_0', \dots, \sigma_{k_n -1}' \in \operatorname{\mathrm{term}}(T_{n+1}) \nonumber\\ [\forall l < k (\sigma_l \subseteq \sigma_l') \rightarrow (\sigma_0', \dots, \sigma_{k_n -1}') \in D_n]. \end{align}$$
 We start with 
 $T_0 = k^{<2} = \{ \emptyset \} \cup \{ \langle l \rangle : l<k \}$
 and then apply Lemma 4.15 recursively. Let
$T_0 = k^{<2} = \{ \emptyset \} \cup \{ \langle l \rangle : l<k \}$
 and then apply Lemma 4.15 recursively. Let 
 $S := \bigcup _{n \in \omega } T_n$
. Then we have that
$S := \bigcup _{n \in \omega } T_n$
. Then we have that 
 $S \leq _\rho T$
.
$S \leq _\rho T$
.
Claim 4.17. For any 
 $m \in \omega $
 and distinct
$m \in \omega $
 and distinct 
 $x_0, \dots , x_{m-1} \in [S]$
,
$x_0, \dots , x_{m-1} \in [S]$
, 
 $(x_0, \dots , x_{m-1})$
 is
$(x_0, \dots , x_{m-1})$
 is 
 $T^m$
-generic over M.
$T^m$
-generic over M.
Proof. Let 
 $D \subseteq T^m$
 be open dense with
$D \subseteq T^m$
 be open dense with 
 $D \in M$
. Then there is a large enough
$D \in M$
. Then there is a large enough 
 $n \in \omega $
 with
$n \in \omega $
 with 
 $(D_n, k_n) = (D,m)$
 and
$(D_n, k_n) = (D,m)$
 and 
 $\sigma _0, \dots , \sigma _{m-1} \in \operatorname {\mathrm {term}}(T_n)$
 distinct such that
$\sigma _0, \dots , \sigma _{m-1} \in \operatorname {\mathrm {term}}(T_n)$
 distinct such that 
 $\sigma _0 \subseteq x_0, \dots , \sigma _{m-1} \subseteq x_{m-1}$
. Then there are unique
$\sigma _0 \subseteq x_0, \dots , \sigma _{m-1} \subseteq x_{m-1}$
. Then there are unique 
 $\sigma ^{\prime }_0, \dots , \sigma ^{\prime }_{m-1} \in \operatorname {\mathrm {term}}(T_{n+1})$
 such that
$\sigma ^{\prime }_0, \dots , \sigma ^{\prime }_{m-1} \in \operatorname {\mathrm {term}}(T_{n+1})$
 such that 
 $\sigma ^{\prime }_0 \subseteq x_0, \dots , \sigma ^{\prime }_{m-1} \subseteq x_{m-1}$
. By
$\sigma ^{\prime }_0 \subseteq x_0, \dots , \sigma ^{\prime }_{m-1} \subseteq x_{m-1}$
. By 
 $(*_1)$
,
$(*_1)$
, 
 $(\sigma ^{\prime }_0, \dots , \sigma ^{\prime }_{m-1}) \in D$
.⊣
$(\sigma ^{\prime }_0, \dots , \sigma ^{\prime }_{m-1}) \in D$
.⊣
 Finally let 
 $S_l = \{ s : \langle l \rangle ^{\frown } s \in S \}$
 and note that
$S_l = \{ s : \langle l \rangle ^{\frown } s \in S \}$
 and note that 
 $S_l \leq _{\rho _l} R_l$
 for every
$S_l \leq _{\rho _l} R_l$
 for every 
 $l < k$
. The above claim clearly implies the statement of the proposition.
$l < k$
. The above claim clearly implies the statement of the proposition.
Remark 4.18. Proposition 4.16 implies directly the main result of [Reference Spinas30]. A modification of the above construction for splitting forcing can be used to show that for 
 $T \in M$
, we can in fact find a master condition
$T \in M$
, we can in fact find a master condition 
 $S \leq T$
 so that for any distinct
$S \leq T$
 so that for any distinct 
 $x_0,\dots , x_{n-1} \in [S]$
,
$x_0,\dots , x_{n-1} \in [S]$
, 
 $(x_0, \dots , x_{n-1})$
 is
$(x_0, \dots , x_{n-1})$
 is 
 $\mathbb {SP}^n$
-generic over M. In that case
$\mathbb {SP}^n$
-generic over M. In that case 
 $(S,\dots , S) \in \mathbb {SP}^n$
 is an
$(S,\dots , S) \in \mathbb {SP}^n$
 is an 
 $\mathbb {SP}^n$
-master condition over M. We won’t provide a proof of this since our only application is Corollary 4.20, which seems to be implicit in [Reference Spinas30]. The analogous statement for Sacks forcing is a standard fusion argument.
$\mathbb {SP}^n$
-master condition over M. We won’t provide a proof of this since our only application is Corollary 4.20, which seems to be implicit in [Reference Spinas30]. The analogous statement for Sacks forcing is a standard fusion argument.
Corollary 4.19. Let 
 $\,\mathbb {P}$
 be a weighted tree forcing with continuous reading of names. Then
$\,\mathbb {P}$
 be a weighted tree forcing with continuous reading of names. Then 
 $\mathbb {P}$
 adds a minimal real. In fact for any
$\mathbb {P}$
 adds a minimal real. In fact for any 
 $\mathbb {P}$
-generic G, if
$\mathbb {P}$
-generic G, if 
 $y \in 2^\omega \cap V[G] \setminus V$
, then there is a continuous map
$y \in 2^\omega \cap V[G] \setminus V$
, then there is a continuous map 
 $f \colon 2^\omega \to A^\omega $
 in V so that
$f \colon 2^\omega \to A^\omega $
 in V so that 
 $x_G = f(y)$
.
$x_G = f(y)$
.
Proof. Using the continuous reading of names let 
 $T \in G$
 be so that there is a continuous map
$T \in G$
 be so that there is a continuous map 
 $g \colon [T] \to 2^\omega $
 with
$g \colon [T] \to 2^\omega $
 with 
 $T \Vdash \dot y = g(x_G)$
. It is easy to see from the definition, that in any weighted tree forcing, the set of finitely branching trees is dense.Footnote 
6
 Thus, let us assume that
$T \Vdash \dot y = g(x_G)$
. It is easy to see from the definition, that in any weighted tree forcing, the set of finitely branching trees is dense.Footnote 
6
 Thus, let us assume that 
 $[T]$
 is compact. Moreover let M be countable elementary with
$[T]$
 is compact. Moreover let M be countable elementary with 
 $g, T \in M$
. Now let
$g, T \in M$
. Now let 
 $S \leq T$
 be so that any
$S \leq T$
 be so that any 
 $x_0,x_1 \in [S]$
 are
$x_0,x_1 \in [S]$
 are 
 $[T]$
-mCg over M.
$[T]$
-mCg over M.
 Suppose that there are 
 $x_0 \neq x_1 \in [S]$
, with
$x_0 \neq x_1 \in [S]$
, with 
 $g(x_0) = g(x_1)$
. Then there must be
$g(x_0) = g(x_1)$
. Then there must be 
 $s \subseteq x_0$
 and
$s \subseteq x_0$
 and 
 $t \subseteq x_1$
, so that
$t \subseteq x_1$
, so that 
 $M \models (s,t) \Vdash _{T^2} g(\dot c_0) = g(\dot c_1)$
, where
$M \models (s,t) \Vdash _{T^2} g(\dot c_0) = g(\dot c_1)$
, where 
 $\dot c_0$
,
$\dot c_0$
, 
 $\dot c_1$
 are names for the generic branches added by
$\dot c_1$
 are names for the generic branches added by 
 $T^2$
. But then note that for any
$T^2$
. But then note that for any 
 $x \in S_{t}$
, since x and
$x \in S_{t}$
, since x and 
 $x_0$
 are mCg and
$x_0$
 are mCg and 
 $s \subseteq x_0$
,
$s \subseteq x_0$
, 
 $t \subseteq x$
, we have that
$t \subseteq x$
, we have that 
 $g(x) = g(x_0)$
. In particular g is constant on
$g(x) = g(x_0)$
. In particular g is constant on 
 $S_t$
 and
$S_t$
 and 
 $S_t \Vdash g(x_G) = g(\check x_0) \in V$
.
$S_t \Vdash g(x_G) = g(\check x_0) \in V$
.
 On the other hand, if g is injective on 
 $[S]$
, then
$[S]$
, then 
 $g^{-1}$
 is continuous as
$g^{-1}$
 is continuous as 
 $[S]$
 is compact and it is easy to extend
$[S]$
 is compact and it is easy to extend 
 $g^{-1}$
 to a continuous function
$g^{-1}$
 to a continuous function 
 $f \colon A^\omega \to 2^\omega $
.⊣
$f \colon A^\omega \to 2^\omega $
.⊣
Corollary 4.20. 
 $V^{\mathbb {SP}}$
 is a minimal extension of V, i.e., whenever W is a model of ZFC so that
$V^{\mathbb {SP}}$
 is a minimal extension of V, i.e., whenever W is a model of ZFC so that 
 $V \subseteq W \subseteq V^{\mathbb {SP}}$
, then
$V \subseteq W \subseteq V^{\mathbb {SP}}$
, then 
 $W = V$
 or
$W = V$
 or 
 $W = V^{\mathbb {SP}}$
.
$W = V^{\mathbb {SP}}$
.
Proof. Let G be an 
 $\mathbb {SP}$
-generic filter over V. It suffices to show that if
$\mathbb {SP}$
-generic filter over V. It suffices to show that if 
 $\langle \alpha _\xi : \xi < \delta \rangle \in W \setminus V$
 is an increasing sequence of ordinals, then
$\langle \alpha _\xi : \xi < \delta \rangle \in W \setminus V$
 is an increasing sequence of ordinals, then 
 $x_G \in W$
 (see also [Reference Jech18, Theorem 13.28]). So let
$x_G \in W$
 (see also [Reference Jech18, Theorem 13.28]). So let 
 $ \langle \dot \alpha _\xi : \xi < \delta \rangle $
 be a name for such a sequence of ordinals and
$ \langle \dot \alpha _\xi : \xi < \delta \rangle $
 be a name for such a sequence of ordinals and 
 $T \in \mathbb {SP}$
 be such that
$T \in \mathbb {SP}$
 be such that 
 $T \Vdash \langle \dot \alpha _\xi : \xi < \delta \rangle \notin V$
. Note that this is in fact equivalent to saying that
$T \Vdash \langle \dot \alpha _\xi : \xi < \delta \rangle \notin V$
. Note that this is in fact equivalent to saying that 
 $(T,T) \Vdash _{\mathbb {SP}^2} \langle \dot \alpha _\xi [\dot x_0] : \xi < \delta \rangle \neq \langle \dot \alpha _\xi [\dot x_1] : \xi < \delta \rangle $
, where
$(T,T) \Vdash _{\mathbb {SP}^2} \langle \dot \alpha _\xi [\dot x_0] : \xi < \delta \rangle \neq \langle \dot \alpha _\xi [\dot x_1] : \xi < \delta \rangle $
, where 
 $\dot x_0, \dot x_1$
 are names for the generic reals added by
$\dot x_0, \dot x_1$
 are names for the generic reals added by 
 $\mathbb {SP}^2$
. Let M be a countable elementary model so that
$\mathbb {SP}^2$
. Let M be a countable elementary model so that 
 $T, \langle \dot \alpha _\xi : \xi < \delta \rangle \in M$
 and let
$T, \langle \dot \alpha _\xi : \xi < \delta \rangle \in M$
 and let 
 $T' \leq T$
 be a master condition over M as in Remark 4.18. Then also
$T' \leq T$
 be a master condition over M as in Remark 4.18. Then also 
 $T' \Vdash \langle \dot \alpha _\xi : \xi \in \delta \cap M \rangle \notin V$
. Namely, suppose towards a contradiction that there are
$T' \Vdash \langle \dot \alpha _\xi : \xi \in \delta \cap M \rangle \notin V$
. Namely, suppose towards a contradiction that there are 
 $x_0, x_1 \in [T']$
 generic over V so that
$x_0, x_1 \in [T']$
 generic over V so that 
 $\langle \dot \alpha _\xi [x_0] : \xi \in \delta \cap M \rangle = \langle \dot \alpha _\xi [x_1] : \xi \in \delta \cap M \rangle $
, then
$\langle \dot \alpha _\xi [x_0] : \xi \in \delta \cap M \rangle = \langle \dot \alpha _\xi [x_1] : \xi \in \delta \cap M \rangle $
, then 
 $(x_0,x_1)$
 is
$(x_0,x_1)$
 is 
 $\mathbb {SP}^2$
-generic over M and
$\mathbb {SP}^2$
-generic over M and 
 $M[x_0][x_1] \models \langle \dot \alpha _\xi [x_0] : \xi < \delta \rangle = \langle \dot \alpha _\xi [x_1] : \xi < \delta \rangle $
 which yields a contradiction to the sufficient elementarity of M. Since
$M[x_0][x_1] \models \langle \dot \alpha _\xi [x_0] : \xi < \delta \rangle = \langle \dot \alpha _\xi [x_1] : \xi < \delta \rangle $
 which yields a contradiction to the sufficient elementarity of M. Since 
 $T' \Vdash \langle \dot \alpha _\xi : \xi \in \delta \cap M \rangle \subseteq M$
 we can view
$T' \Vdash \langle \dot \alpha _\xi : \xi \in \delta \cap M \rangle \subseteq M$
 we can view 
 $\langle \dot \alpha _\xi : \xi \in \delta \cap M \rangle $
 as a name for a real, for M is countable. Back in W, we can define
$\langle \dot \alpha _\xi : \xi \in \delta \cap M \rangle $
 as a name for a real, for M is countable. Back in W, we can define 
 $\langle \alpha _\xi : \xi \in \delta \cap M \rangle $
 since
$\langle \alpha _\xi : \xi \in \delta \cap M \rangle $
 since 
 $M \in V \subseteq W$
. But then, applying Corollary 4.19, we find that
$M \in V \subseteq W$
. But then, applying Corollary 4.19, we find that 
 $x_G \in W$
.⊣
$x_G \in W$
.⊣
4.3 The countable support iteration
 Recall that for any perfect subtree T of 
 $2^{<\omega }$
,
$2^{<\omega }$
, 
 $\operatorname {\mathrm {split}}(T)$
 is order-isomorphic to
$\operatorname {\mathrm {split}}(T)$
 is order-isomorphic to 
 $2^{<\omega }$
 in a canonical way, via a map
$2^{<\omega }$
 in a canonical way, via a map 
 $\eta _T \colon \operatorname {\mathrm {split}}(T) \to 2^{<\omega }$
. This map induces a homeomorphism
$\eta _T \colon \operatorname {\mathrm {split}}(T) \to 2^{<\omega }$
. This map induces a homeomorphism 
 $\tilde \eta _T \colon [T] \to 2^\omega $
 and note that the value of
$\tilde \eta _T \colon [T] \to 2^\omega $
 and note that the value of 
 $\tilde \eta _T(x)$
 depends continuously on T and x. Whenever
$\tilde \eta _T(x)$
 depends continuously on T and x. Whenever 
 $\rho $
 is a weight on T,
$\rho $
 is a weight on T, 
 $\eta _T$
 also induces a weight
$\eta _T$
 also induces a weight 
 $\tilde \rho $
 on
$\tilde \rho $
 on 
 $2^{<\omega }$
, so that whenever
$2^{<\omega }$
, so that whenever 
 $S \leq _{\tilde \rho } 2^{<\omega }$
, then
$S \leq _{\tilde \rho } 2^{<\omega }$
, then 
 $\eta _T^{-1}(S)$
 generates a tree
$\eta _T^{-1}(S)$
 generates a tree 
 $S'$
 with
$S'$
 with 
 $S' \leq _{\rho } T$
.
$S' \leq _{\rho } T$
.
 Let 
 $\langle \mathbb {P}_\beta , \dot {\mathbb {Q}}_\beta : \beta < \lambda \rangle $
 be a countable support iteration where for each
$\langle \mathbb {P}_\beta , \dot {\mathbb {Q}}_\beta : \beta < \lambda \rangle $
 be a countable support iteration where for each 
 $\beta < \lambda $
,
$\beta < \lambda $
, 
 $\Vdash _{\mathbb {P}_\beta } \dot {\mathbb {Q}}_\beta \in \{ \mathbb {SP},\mathbb {S}\}$
. We fix in this section a
$\Vdash _{\mathbb {P}_\beta } \dot {\mathbb {Q}}_\beta \in \{ \mathbb {SP},\mathbb {S}\}$
. We fix in this section a 
 $\mathbb {P}_\lambda $
 name
$\mathbb {P}_\lambda $
 name 
 $\dot y$
 for an element of a Polish space X, a good master condition
$\dot y$
 for an element of a Polish space X, a good master condition 
 $\bar p \in \mathbb {P_\lambda }$
 over a countable model
$\bar p \in \mathbb {P_\lambda }$
 over a countable model 
 $M_0$
, where
$M_0$
, where 
 $\dot y, X \in M_0$
, and let
$\dot y, X \in M_0$
, and let 
 $C\subseteq \lambda $
 be a countable set as in Lemma 2.3. For every
$C\subseteq \lambda $
 be a countable set as in Lemma 2.3. For every 
 $\beta \in C$
 and
$\beta \in C$
 and 
 $\bar y \in [\bar p]\restriction (C \cap \beta )$
, let us write
$\bar y \in [\bar p]\restriction (C \cap \beta )$
, let us write 
 $$ \begin{align*}T_{\bar y} = \{s \in 2^{<\omega} : \exists \bar x \in [\bar p] \left[ \bar x \restriction (C \cap\beta) = \bar y \wedge s \subseteq x(\beta)\right] \}.\end{align*} $$
$$ \begin{align*}T_{\bar y} = \{s \in 2^{<\omega} : \exists \bar x \in [\bar p] \left[ \bar x \restriction (C \cap\beta) = \bar y \wedge s \subseteq x(\beta)\right] \}.\end{align*} $$
According to Lemma 2.3, the map 
 $\bar y \mapsto T_{\bar y}$
 is a continuous function from
$\bar y \mapsto T_{\bar y}$
 is a continuous function from 
 ${[\bar p]\restriction (C \cap \beta )}$
 to
${[\bar p]\restriction (C \cap \beta )}$
 to 
 $\mathcal {T}$
. Let
$\mathcal {T}$
. Let 
 $\alpha := \operatorname {\mathrm {otp}}(C) < \omega _1$
 as witnessed by an order-isomorphism
$\alpha := \operatorname {\mathrm {otp}}(C) < \omega _1$
 as witnessed by an order-isomorphism 
 $\iota \colon \alpha \to C$
. Then we define the homeomorphism
$\iota \colon \alpha \to C$
. Then we define the homeomorphism 
 $\Phi \colon [\bar p] \restriction C \to (2^\omega )^\alpha $
 so that for every
$\Phi \colon [\bar p] \restriction C \to (2^\omega )^\alpha $
 so that for every 
 $\bar y \in [\bar p] \restriction C$
 and every
$\bar y \in [\bar p] \restriction C$
 and every 
 $\delta < \alpha $
,
$\delta < \alpha $
, 
 $$ \begin{align*}\Phi(\bar y) \restriction (\delta + 1) = \Phi(\bar y) \restriction \delta^{\frown} \tilde \eta_{T_{\bar y \restriction \iota(\delta)}}(y(\iota(\delta))).\end{align*} $$
$$ \begin{align*}\Phi(\bar y) \restriction (\delta + 1) = \Phi(\bar y) \restriction \delta^{\frown} \tilde \eta_{T_{\bar y \restriction \iota(\delta)}}(y(\iota(\delta))).\end{align*} $$
 Note that for 
 $\mathbb {P} \in \{ \mathbb {SP}, \mathbb {S}\}$
, the map sending
$\mathbb {P} \in \{ \mathbb {SP}, \mathbb {S}\}$
, the map sending 
 $T \in \mathbb {P}$
 to the weight
$T \in \mathbb {P}$
 to the weight 
 $\rho _T$
 defined in Lemma 4.10 or Lemma 4.12 is a Borel function from
$\rho _T$
 defined in Lemma 4.10 or Lemma 4.12 is a Borel function from 
 $\mathbb {P}$
 to the Polish space of partial functions from
$\mathbb {P}$
 to the Polish space of partial functions from 
 $(2^{<\omega })^2$
 to
$(2^{<\omega })^2$
 to 
 $[2^{<\omega }]^{<\omega }$
. Thus for
$[2^{<\omega }]^{<\omega }$
. Thus for 
 $\beta \in C$
 and
$\beta \in C$
 and 
 $\bar x \in [\bar p] \restriction (C \cap \beta )$
, letting
$\bar x \in [\bar p] \restriction (C \cap \beta )$
, letting 
 $\rho _{\bar x} := \rho _{T_{\bar x}}$
, we get that
$\rho _{\bar x} := \rho _{T_{\bar x}}$
, we get that 
 $\bar x \mapsto \rho _{\bar x}$
 is a Borel function on
$\bar x \mapsto \rho _{\bar x}$
 is a Borel function on 
 $[\bar p] \restriction (C \cap \beta )$
. For each
$[\bar p] \restriction (C \cap \beta )$
. For each 
 $\delta < \alpha $
 and
$\delta < \alpha $
 and 
 $\bar y \in (2^\omega )^\delta $
, we may then define
$\bar y \in (2^\omega )^\delta $
, we may then define 
 $\tilde \rho _{\bar y}$
 a weight on
$\tilde \rho _{\bar y}$
 a weight on 
 $2^{<\omega }$
, induced by
$2^{<\omega }$
, induced by 
 $\rho _{\bar x}$
 and
$\rho _{\bar x}$
 and 
 $\eta _{T_{\bar x}}$
, where
$\eta _{T_{\bar x}}$
, where 
 $\bar x = \Phi ^{-1}(\bar y^{\frown } \bar z) \restriction \beta $
 for arbitrary, equivalently for every,
$\bar x = \Phi ^{-1}(\bar y^{\frown } \bar z) \restriction \beta $
 for arbitrary, equivalently for every, 
 $\bar z \in (2^\omega )^{\alpha \setminus \delta }$
. The map sending
$\bar z \in (2^\omega )^{\alpha \setminus \delta }$
. The map sending 
 $\bar y \in (2^\omega )^\delta $
 to
$\bar y \in (2^\omega )^\delta $
 to 
 $\tilde \rho _{\bar y}$
 is then Borel as well.
$\tilde \rho _{\bar y}$
 is then Borel as well.
Lemma 4.21. Let 
 $M_1$
 be a countable elementary model with
$M_1$
 be a countable elementary model with 
 $M_0, \bar p, \mathbb {P}_\lambda \in M_1$
 and let
$M_0, \bar p, \mathbb {P}_\lambda \in M_1$
 and let 
 $\bar s \in \bigotimes _{i < \alpha } 2^{<\omega }$
. Then there is
$\bar s \in \bigotimes _{i < \alpha } 2^{<\omega }$
. Then there is 
 $\bar q \leq \bar p$
, a good master condition over
$\bar q \leq \bar p$
, a good master condition over 
 $M_0$
, so that
$M_0$
, so that 
 $$ \begin{align*} \forall \bar x_0, \dots, \bar x_{n-1} \in [\bar q] \big(\Phi(\bar x_0 \restriction C), \dots, \Phi&(\bar x_{n-1} \restriction C) \in (2^\omega)^\alpha \cap [\bar s] \\ & \text{ are strongly }\langle 2^\omega : i < \alpha \rangle \text{-mCg } \text{ over } M_1\big). \end{align*} $$
$$ \begin{align*} \forall \bar x_0, \dots, \bar x_{n-1} \in [\bar q] \big(\Phi(\bar x_0 \restriction C), \dots, \Phi&(\bar x_{n-1} \restriction C) \in (2^\omega)^\alpha \cap [\bar s] \\ & \text{ are strongly }\langle 2^\omega : i < \alpha \rangle \text{-mCg } \text{ over } M_1\big). \end{align*} $$
Moreover 
 $[\bar q] \restriction C$
 is a closed subset of
$[\bar q] \restriction C$
 is a closed subset of 
 $[\bar p] \restriction C$
 and
$[\bar p] \restriction C$
 and 
 $[\bar q] = ([\bar q] \restriction C) \times (2^\omega )^{\lambda \setminus C}$
 (cf. Lemma 2.3).
$[\bar q] = ([\bar q] \restriction C) \times (2^\omega )^{\lambda \setminus C}$
 (cf. Lemma 2.3).
Proof. We can assume without loss of generality that 
 $\bar s = \emptyset $
, i.e.,
$\bar s = \emptyset $
, i.e., 
 $[\bar s] = (2^\omega )^\alpha $
. It will be obvious that this assumption is inessential. Next, let us introduce some notation. For any
$[\bar s] = (2^\omega )^\alpha $
. It will be obvious that this assumption is inessential. Next, let us introduce some notation. For any 
 $\delta \leq \alpha $
 and
$\delta \leq \alpha $
 and 
 $\bar y_0, \dots , \bar y_{n-1} \in (2^\omega )^\delta $
, recall from Definition 3.23 that
$\bar y_0, \dots , \bar y_{n-1} \in (2^\omega )^\delta $
, recall from Definition 3.23 that 
 $$ \begin{align*}\Delta(\bar y_0, \dots, \bar y_{n-1}) := \{ \Delta_{\bar y_i, \bar y_j} : i \neq j < n \} \cup \{ 0, \delta\}.\end{align*} $$
$$ \begin{align*}\Delta(\bar y_0, \dots, \bar y_{n-1}) := \{ \Delta_{\bar y_i, \bar y_j} : i \neq j < n \} \cup \{ 0, \delta\}.\end{align*} $$
Let us write
 $$ \begin{align*}\operatorname{\mathrm{tp}}(\bar y_0, \dots, \bar y_{n-1}) := (\langle \xi_l : l \leq k \rangle, \langle K_l : l < k \rangle, \langle U_i : i < n \rangle),\end{align*} $$
$$ \begin{align*}\operatorname{\mathrm{tp}}(\bar y_0, \dots, \bar y_{n-1}) := (\langle \xi_l : l \leq k \rangle, \langle K_l : l < k \rangle, \langle U_i : i < n \rangle),\end{align*} $$
where 
 $\{ \xi _0 < \dots < \xi _k\} = \Delta (\bar y_0, \dots , \bar y_{n-1})$
,
$\{ \xi _0 < \dots < \xi _k\} = \Delta (\bar y_0, \dots , \bar y_{n-1})$
, 
 $K_l = \vert \{ \bar y_i \restriction [\xi _l, \xi _{l+1}) : i < n \} \vert $
 for every
$K_l = \vert \{ \bar y_i \restriction [\xi _l, \xi _{l+1}) : i < n \} \vert $
 for every 
 $l < k$
 and
$l < k$
 and 
 $\langle U_i : i < n \rangle $
 are the clopen subsets of
$\langle U_i : i < n \rangle $
 are the clopen subsets of 
 $(2^\omega )^\delta $
 of the form
$(2^\omega )^\delta $
 of the form 
 $U_i = [\bar s_i]$
 for
$U_i = [\bar s_i]$
 for 
 $\bar s_i \in \bigotimes _{\xi < \delta } 2^{<\omega }$
 with
$\bar s_i \in \bigotimes _{\xi < \delta } 2^{<\omega }$
 with 
 $\operatorname {\mathrm {dom}}(\bar s_i) = \{ \Delta _{\bar y_i, \bar y_j} : j < n, \bar y_j \neq \bar y_i \}$
 and
$\operatorname {\mathrm {dom}}(\bar s_i) = \{ \Delta _{\bar y_i, \bar y_j} : j < n, \bar y_j \neq \bar y_i \}$
 and 
 $\bar s_i$
 minimal in the order of
$\bar s_i$
 minimal in the order of 
 $\bigotimes _{\xi < \delta } 2^{<\omega }$
 so that
$\bigotimes _{\xi < \delta } 2^{<\omega }$
 so that 
 $$ \begin{align*}\bar y_i \in [\bar s_i] \text{ and } \forall j < n (\bar y_j \neq \bar y_i \rightarrow \bar y_j \notin [\bar s_i]),\end{align*} $$
$$ \begin{align*}\bar y_i \in [\bar s_i] \text{ and } \forall j < n (\bar y_j \neq \bar y_i \rightarrow \bar y_j \notin [\bar s_i]),\end{align*} $$
for every 
 $i < n$
.
$i < n$
.
 Note that for any 
 $\delta _0 \leq \delta $
, if
$\delta _0 \leq \delta $
, if 
 $$ \begin{align*}\operatorname{\mathrm{tp}}(\bar y_0 \restriction \delta_0, \dots, \bar y_{n-1} \restriction \delta_0) := (\langle \eta_l : l \leq k' \rangle, \langle M_l : l < k' \rangle, \langle V_i : i < n \rangle),\end{align*} $$
$$ \begin{align*}\operatorname{\mathrm{tp}}(\bar y_0 \restriction \delta_0, \dots, \bar y_{n-1} \restriction \delta_0) := (\langle \eta_l : l \leq k' \rangle, \langle M_l : l < k' \rangle, \langle V_i : i < n \rangle),\end{align*} $$
then 
 $V_i = U_i \restriction \delta _0$
 for every
$V_i = U_i \restriction \delta _0$
 for every 
 $i < n$
. Moreover, for any
$i < n$
. Moreover, for any 
 $\bar y^{\prime }_0, \dots , \bar y^{\prime }_{n-1} \in (2^\omega )^\delta $
 with
$\bar y^{\prime }_0, \dots , \bar y^{\prime }_{n-1} \in (2^\omega )^\delta $
 with 
 $$ \begin{align*}\operatorname{\mathrm{tp}}(\bar y^{\prime}_0, \dots, \bar y^{\prime}_{n-1}) = (\langle \xi_l : l \leq k \rangle, \langle K_l : l < k \rangle, \langle U_i : i < n \rangle),\end{align*} $$
$$ \begin{align*}\operatorname{\mathrm{tp}}(\bar y^{\prime}_0, \dots, \bar y^{\prime}_{n-1}) = (\langle \xi_l : l \leq k \rangle, \langle K_l : l < k \rangle, \langle U_i : i < n \rangle),\end{align*} $$
we have that
 $$ \begin{align*}\operatorname{\mathrm{tp}}(\bar y^{\prime}_0 \restriction \delta_0, \dots, \bar y^{\prime}_{n-1} \restriction \delta_0) = (\langle \eta_l : l \leq k' \rangle, \langle M_l : l < k' \rangle, \langle V_i : i < n \rangle).\end{align*} $$
$$ \begin{align*}\operatorname{\mathrm{tp}}(\bar y^{\prime}_0 \restriction \delta_0, \dots, \bar y^{\prime}_{n-1} \restriction \delta_0) = (\langle \eta_l : l \leq k' \rangle, \langle M_l : l < k' \rangle, \langle V_i : i < n \rangle).\end{align*} $$
 Any 
 $\bar y_0, \dots , \bar y_{n-1}$
, with
$\bar y_0, \dots , \bar y_{n-1}$
, with 
 $\operatorname {\mathrm {tp}}(\bar y_0, \dots , \bar y_{n-1}) := (\langle \xi _l : l \leq k \rangle , \langle K_l : l < k \rangle , \langle U_i : i < n \rangle )$
, that are
$\operatorname {\mathrm {tp}}(\bar y_0, \dots , \bar y_{n-1}) := (\langle \xi _l : l \leq k \rangle , \langle K_l : l < k \rangle , \langle U_i : i < n \rangle )$
, that are 
 $\langle 2^\omega : i < \delta \rangle $
-mutually Cohen generic over
$\langle 2^\omega : i < \delta \rangle $
-mutually Cohen generic over 
 $M_1$
 as witnessed by
$M_1$
 as witnessed by 
 $\xi _0 < \dots < \xi _k$
, induce a
$\xi _0 < \dots < \xi _k$
, induce a 
 $\prod _{l < k} (\bigotimes _{\xi \in [\xi _l,\xi _{l+1})} 2^{<\omega })^{K_l}$
-generic and vice versa. Thus whenever
$\prod _{l < k} (\bigotimes _{\xi \in [\xi _l,\xi _{l+1})} 2^{<\omega })^{K_l}$
-generic and vice versa. Thus whenever 
 $\tau $
 is a
$\tau $
 is a 
 $\prod _{l < k} (\bigotimes _{\xi \in [\xi _l,\xi _{l+1})} 2^{<\omega })^{K_l}$
-name, we may write
$\prod _{l < k} (\bigotimes _{\xi \in [\xi _l,\xi _{l+1})} 2^{<\omega })^{K_l}$
-name, we may write 
 $\tau [\bar y_0, \dots , \bar y_{n-1}]$
 for the evaluation of
$\tau [\bar y_0, \dots , \bar y_{n-1}]$
 for the evaluation of 
 $\tau $
 via the induced generic. It will not matter in what particular way we define the
$\tau $
 via the induced generic. It will not matter in what particular way we define the 
 $\prod _{l < k} (\bigotimes _{\xi \in [\xi _l,\xi _{l+1})} 2^{<\omega })^{K_l}$
-generic from given
$\prod _{l < k} (\bigotimes _{\xi \in [\xi _l,\xi _{l+1})} 2^{<\omega })^{K_l}$
-generic from given 
 $\bar y_0, \dots , \bar y_{n-1}$
. We may stipulate for instance, that the generic induced by
$\bar y_0, \dots , \bar y_{n-1}$
. We may stipulate for instance, that the generic induced by 
 $\bar y_0, \dots , \bar y_{n-1}$
 is
$\bar y_0, \dots , \bar y_{n-1}$
 is 
 $\langle \bar z_{l,j} : l < k, j < K_l \rangle $
, where for each fixed
$\langle \bar z_{l,j} : l < k, j < K_l \rangle $
, where for each fixed 
 $l < k$
,
$l < k$
, 
 $\langle \bar z_{l,j} : l < k, j < K_l \rangle $
 enumerates
$\langle \bar z_{l,j} : l < k, j < K_l \rangle $
 enumerates 
 $\{ \bar y_i \restriction [\xi _l, \xi _{l+1}) : i < n \}$
 in lexicographic order.
$\{ \bar y_i \restriction [\xi _l, \xi _{l+1}) : i < n \}$
 in lexicographic order.
 Let us get to the bulk of the proof. We will define a finite support iteration 
 $\langle \mathbb {R}_{\delta }, \dot {\mathbb {S}}_\delta : \delta \leq \alpha \rangle $
 in
$\langle \mathbb {R}_{\delta }, \dot {\mathbb {S}}_\delta : \delta \leq \alpha \rangle $
 in 
 $M_1$
, together with, for each
$M_1$
, together with, for each 
 $\delta \leq \alpha $
, an
$\delta \leq \alpha $
, an 
 $\mathbb {R}_{\delta }$
-name
$\mathbb {R}_{\delta }$
-name 
 $\dot X_\delta $
 for a closed subspace of
$\dot X_\delta $
 for a closed subspace of 
 $(2^{\omega })^\delta $
, where
$(2^{\omega })^\delta $
, where 
 $\Vdash _{\mathbb {R}_{\delta _1}} \dot X_{\delta _0} = \dot X_{\delta _1} \restriction \delta _0$
 for every
$\Vdash _{\mathbb {R}_{\delta _1}} \dot X_{\delta _0} = \dot X_{\delta _1} \restriction \delta _0$
 for every 
 $\delta _0 < \delta _1 \leq \alpha $
. This uniquely determines the limit steps of the construction. Additionally we will make the following inductive assumptions
$\delta _0 < \delta _1 \leq \alpha $
. This uniquely determines the limit steps of the construction. Additionally we will make the following inductive assumptions 
 $(1)_\delta $
 and
$(1)_\delta $
 and 
 $(2)_\delta $
 for all
$(2)_\delta $
 for all 
 $\delta \leq \alpha $
 and any
$\delta \leq \alpha $
 and any 
 $\mathbb {R}_\delta $
-generic G. Let
$\mathbb {R}_\delta $
-generic G. Let 
 $\bar y_0, \dots , \bar y_{n-1} \in \dot X_{\delta }[G]$
 be arbitrary and
$\bar y_0, \dots , \bar y_{n-1} \in \dot X_{\delta }[G]$
 be arbitrary and 
 $\operatorname {\mathrm {tp}}(\bar y_0, \dots , \bar y_{n-1}) = (\langle \xi _l : l \leq k \rangle , \langle K_l : l < k \rangle , \langle U_i : i < n \rangle )$
. Then
$\operatorname {\mathrm {tp}}(\bar y_0, \dots , \bar y_{n-1}) = (\langle \xi _l : l \leq k \rangle , \langle K_l : l < k \rangle , \langle U_i : i < n \rangle )$
. Then
- 
(1) δ  $\bar y_0, \dots , \bar y_{n-1}$
 are strongly $\bar y_0, \dots , \bar y_{n-1}$
 are strongly $\langle 2^\omega : i < \delta \rangle $
-mCg over $\langle 2^\omega : i < \delta \rangle $
-mCg over $M_1$
 and $M_1$
 and
- 
(2) δ for any  $\prod _{l < k} (\bigotimes _{\xi \in [\xi _l,\xi _{l+1})} 2^{<\omega })^{K_l}$
-name $\prod _{l < k} (\bigotimes _{\xi \in [\xi _l,\xi _{l+1})} 2^{<\omega })^{K_l}$
-name $\dot D \in M_1$
 for an open dense subset of a countable poset $\dot D \in M_1$
 for an open dense subset of a countable poset $\mathbb {Q} \in M_1$
, is open dense in $\mathbb {Q} \in M_1$
, is open dense in $$ \begin{align*} & \bigcap \Big\{ \dot D[\bar y^{\prime}_0, \dots, \bar y^{\prime}_{n-1}] : \bar y^{\prime}_0, \dots, \bar y^{\prime}_{n-1} \in X_\delta, \\ &\qquad\qquad\qquad\operatorname{\mathrm{tp}}(\bar y^{\prime}_0, \dots, \bar y^{\prime}_{n-1}) = (\langle \xi_l : l \leq k \rangle, \langle K_l : l < k \rangle, \langle U_i : i < n \rangle) \Big\} \end{align*} $$ $$ \begin{align*} & \bigcap \Big\{ \dot D[\bar y^{\prime}_0, \dots, \bar y^{\prime}_{n-1}] : \bar y^{\prime}_0, \dots, \bar y^{\prime}_{n-1} \in X_\delta, \\ &\qquad\qquad\qquad\operatorname{\mathrm{tp}}(\bar y^{\prime}_0, \dots, \bar y^{\prime}_{n-1}) = (\langle \xi_l : l \leq k \rangle, \langle K_l : l < k \rangle, \langle U_i : i < n \rangle) \Big\} \end{align*} $$ $\mathbb {Q}$
. $\mathbb {Q}$
.
 Having defined 
 $\mathbb {R}_\delta $
 and
$\mathbb {R}_\delta $
 and 
 $\dot {X}_\delta $
, for
$\dot {X}_\delta $
, for 
 $\delta < \alpha $
, we proceed as follows. Fix for now G an
$\delta < \alpha $
, we proceed as follows. Fix for now G an 
 $\mathbb {R}_\delta $
-generic over
$\mathbb {R}_\delta $
-generic over 
 $M_1$
 and
$M_1$
 and 
 $X_\delta := \dot {X}_\delta [G]$
. Then we define a forcing
$X_\delta := \dot {X}_\delta [G]$
. Then we define a forcing 
 $\mathbb {S}_\delta \in M_1[G]$
 which generically adds a continuous map
$\mathbb {S}_\delta \in M_1[G]$
 which generically adds a continuous map 
 $F \colon X_\delta \to \mathcal {T}$
, so that for each
$F \colon X_\delta \to \mathcal {T}$
, so that for each 
 $\bar y \in X_\delta $
,
$\bar y \in X_\delta $
, 
 $ S_{\bar y} := F(\bar y) \leq _{\tilde \rho _{\bar y}} 2^{<\omega }$
. In
$ S_{\bar y} := F(\bar y) \leq _{\tilde \rho _{\bar y}} 2^{<\omega }$
. In 
 $M_1[G][F]$
, we then define
$M_1[G][F]$
, we then define 
 $X_{\delta +1} \subseteq (2^\omega )^{\delta +1}$
 to be
$X_{\delta +1} \subseteq (2^\omega )^{\delta +1}$
 to be 
 $\{ \bar y^{\frown } z : \bar y \in X_\delta , z \in [S_{\bar y}] \}$
. The definition of
$\{ \bar y^{\frown } z : \bar y \in X_\delta , z \in [S_{\bar y}] \}$
. The definition of 
 $\mathbb {S}_\delta $
 is as follows.
$\mathbb {S}_\delta $
 is as follows.
 Work in 
 $M_1[G]$
. Since the map
$M_1[G]$
. Since the map 
 $\bar y \in (2^\omega )^\delta \mapsto \tilde \rho _{\bar y}$
 is Borel and an element of
$\bar y \in (2^\omega )^\delta \mapsto \tilde \rho _{\bar y}$
 is Borel and an element of 
 $M_1$
 and by
$M_1$
 and by 
 $(1)_\delta $
 any
$(1)_\delta $
 any 
 $\bar y \in X_\delta $
 is Cohen generic over
$\bar y \in X_\delta $
 is Cohen generic over 
 $M_1$
, it is continuous on
$M_1$
, it is continuous on 
 $X_\delta $
. Since
$X_\delta $
. Since 
 $X_\delta $
 is compact we find a single weight
$X_\delta $
 is compact we find a single weight 
 $\tilde \rho $
 on
$\tilde \rho $
 on 
 $2^{<\omega }$
, so that
$2^{<\omega }$
, so that 
 $\tilde \rho _{\bar y} \subseteq \tilde \rho $
 for every
$\tilde \rho _{\bar y} \subseteq \tilde \rho $
 for every 
 $\bar y \in X_\delta $
. Let
$\bar y \in X_\delta $
. Let 
 $\{ O_s : s \in 2^{<\omega } \}$
 be a basis of
$\{ O_s : s \in 2^{<\omega } \}$
 be a basis of 
 $X_\delta $
 so that
$X_\delta $
 so that 
 $O_s \subseteq O_t$
 for
$O_s \subseteq O_t$
 for 
 $t \subseteq s$
 and
$t \subseteq s$
 and 
 $O_s \cap O_t = \emptyset $
 for
$O_s \cap O_t = \emptyset $
 for 
 $s \perp t$
. This is possible since
$s \perp t$
. This is possible since 
 $X_\delta $
 is homeomorphic to
$X_\delta $
 is homeomorphic to 
 $2^\omega $
. Let
$2^\omega $
. Let 
 $\mathcal {FT}$
 be the set of finite subtrees of
$\mathcal {FT}$
 be the set of finite subtrees of 
 $2^{<\omega }$
. Then
$2^{<\omega }$
. Then 
 $\mathbb {S}_\delta $
 consists of functions
$\mathbb {S}_\delta $
 consists of functions 
 $h \colon 2^{\leq n} \to \mathcal {FT}$
, for some
$h \colon 2^{\leq n} \to \mathcal {FT}$
, for some 
 $n \in \omega $
, so that for every
$n \in \omega $
, so that for every 
 $s \subseteq t \in 2^{\leq n}$
,
$s \subseteq t \in 2^{\leq n}$
, 
 $ (h(s) \unlhd _{\tilde \rho } h(t))$
. The extension relation is defined by function extension. Note that
$ (h(s) \unlhd _{\tilde \rho } h(t))$
. The extension relation is defined by function extension. Note that 
 $\mathbb {S}_\delta $
 is indeed a forcing poset with trivial condition
$\mathbb {S}_\delta $
 is indeed a forcing poset with trivial condition 
 $\emptyset $
.
$\emptyset $
.
 Given H, an 
 $\mathbb {S}_\delta $
-generic over
$\mathbb {S}_\delta $
-generic over 
 $M_1[G]$
, we let
$M_1[G]$
, we let 
 $F \colon X_\delta \to \mathcal {T}$
 be defined as
$F \colon X_\delta \to \mathcal {T}$
 be defined as 
 $$ \begin{align*}F(\bar y) := \bigcup_{\substack{s \in 2^{<\omega}, \bar y \in O_s \\ h \in H}} h(s).\end{align*} $$
$$ \begin{align*}F(\bar y) := \bigcup_{\substack{s \in 2^{<\omega}, \bar y \in O_s \\ h \in H}} h(s).\end{align*} $$
Claim 4.22. For every 
 $\bar y \in X_\delta $
,
$\bar y \in X_\delta $
, 
 $F(\bar y) = S_{\bar y} \leq _{\tilde \rho } 2^{<\omega }$
, in particular
$F(\bar y) = S_{\bar y} \leq _{\tilde \rho } 2^{<\omega }$
, in particular 
 $S_{\bar y} \leq _{\tilde \rho _{\bar y}} 2^{<\omega }$
. For any
$S_{\bar y} \leq _{\tilde \rho _{\bar y}} 2^{<\omega }$
. For any 
 $\bar y_0, \bar y_1 \in X_\delta $
,
$\bar y_0, \bar y_1 \in X_\delta $
, 
 $[S_{\bar y_0}] \cap [S_{\bar y_1}] \neq \emptyset $
. Any
$[S_{\bar y_0}] \cap [S_{\bar y_1}] \neq \emptyset $
. Any 
 $z_0, \dots , z_{n-1} \in \bigcup _{\bar y \in X_\delta }[S_{\bar y}]$
 are
$z_0, \dots , z_{n-1} \in \bigcup _{\bar y \in X_\delta }[S_{\bar y}]$
 are 
 $2^\omega $
-mutually Cohen generic over
$2^\omega $
-mutually Cohen generic over 
 $M_1[G]$
. And for any countable poset
$M_1[G]$
. And for any countable poset 
 $\mathbb {Q} \in M_1$
, any
$\mathbb {Q} \in M_1$
, any 
 $m \in \omega $
 and any dense open
$m \in \omega $
 and any dense open 
 $E \subseteq (2^{<\omega })^n \times \mathbb {Q}$
 in
$E \subseteq (2^{<\omega })^n \times \mathbb {Q}$
 in 
 $M_1[G]$
, there is
$M_1[G]$
, there is 
 $r \in \mathbb {Q}$
 and
$r \in \mathbb {Q}$
 and 
 $m_0 \geq m$
 so that for any
$m_0 \geq m$
 so that for any 
 $z_0, \dots , z_{n-1} \in \bigcup _{\bar y \in X_\delta }[S_{\bar y}]$
 where
$z_0, \dots , z_{n-1} \in \bigcup _{\bar y \in X_\delta }[S_{\bar y}]$
 where 
 $z_0 \restriction m, \dots , z_{n-1} \restriction m$
 are pairwise distinct,
$z_0 \restriction m, \dots , z_{n-1} \restriction m$
 are pairwise distinct, 
 $( (z_0 \restriction m_0, \dots , z_{n-1} \restriction m_0), r) \in E$
.
$( (z_0 \restriction m_0, \dots , z_{n-1} \restriction m_0), r) \in E$
.
Proof. We will make a genericity argument over 
 $M_1[G]$
. Let
$M_1[G]$
. Let 
 $h \in \mathbb {S}_\delta $
 be arbitrary. Then it is easy to find
$h \in \mathbb {S}_\delta $
 be arbitrary. Then it is easy to find 
 $h' \leq h$
, say with
$h' \leq h$
, say with 
 $\operatorname {\mathrm {dom}}(h') = 2^{\leq a_0}$
, so that for every
$\operatorname {\mathrm {dom}}(h') = 2^{\leq a_0}$
, so that for every 
 $s \in 2^{a_0}$
 and every
$s \in 2^{a_0}$
 and every 
 $t \in \operatorname {\mathrm {term}}(h(s))$
,
$t \in \operatorname {\mathrm {term}}(h(s))$
, 
 $\vert t \vert \geq m$
. For the first claim, it suffices through Lemma 4.14 to find
$\vert t \vert \geq m$
. For the first claim, it suffices through Lemma 4.14 to find 
 $h" \leq h'$
, say with
$h" \leq h'$
, say with 
 $\operatorname {\mathrm {dom}}(h") = 2^{\leq a_1}$
,
$\operatorname {\mathrm {dom}}(h") = 2^{\leq a_1}$
, 
 $a_0 < a_1$
, so that for every
$a_0 < a_1$
, so that for every 
 $s \in 2^{a_0}$
 and
$s \in 2^{a_0}$
 and 
 $t \in 2^{a_1}$
, with
$t \in 2^{a_1}$
, with 
 $s \subseteq t$
,
$s \subseteq t$
, 
 $h"(s) \lhd _{\tilde \rho } h"(t)$
. Finding
$h"(s) \lhd _{\tilde \rho } h"(t)$
. Finding 
 $h"$
 so that additionally
$h"$
 so that additionally 
 $\operatorname {\mathrm {term}}(h"(t_0)) \cap \operatorname {\mathrm {term}}(h"(t_1)) = \emptyset $
 for every
$\operatorname {\mathrm {term}}(h"(t_0)) \cap \operatorname {\mathrm {term}}(h"(t_1)) = \emptyset $
 for every 
 $t_0 \neq t_1 \in 2^{a_1}$
 proves the second claim. For the last two claims, given a fixed dense open subset
$t_0 \neq t_1 \in 2^{a_1}$
 proves the second claim. For the last two claims, given a fixed dense open subset 
 $E \subseteq (2^{<\omega })^n \times \mathbb {Q}$
 in
$E \subseteq (2^{<\omega })^n \times \mathbb {Q}$
 in 
 $M_1[G]$
, it suffices to find
$M_1[G]$
, it suffices to find 
 $r \in \mathbb {Q}$
 and to ensure that for any pairwise distinct
$r \in \mathbb {Q}$
 and to ensure that for any pairwise distinct 
 $s_0, \dots , s_{n-1} \in \bigcup _{s \in 2^{a_0}} \operatorname {\mathrm {term}}(h"(s))$
 and
$s_0, \dots , s_{n-1} \in \bigcup _{s \in 2^{a_0}} \operatorname {\mathrm {term}}(h"(s))$
 and 
 $t_0 \supseteq s_0, \dots , t_{n-1} \supseteq s_{n-1}$
 with
$t_0 \supseteq s_0, \dots , t_{n-1} \supseteq s_{n-1}$
 with 
 $t_0, \dots , t_{n-1} \in \bigcup _{t \in 2^{a_1}} \operatorname {\mathrm {term}}(h"(t))$
,
$t_0, \dots , t_{n-1} \in \bigcup _{t \in 2^{a_1}} \operatorname {\mathrm {term}}(h"(t))$
, 
 ${((t_0, \dots , t_{n-1}), r) \in E}$
. Then we may put
${((t_0, \dots , t_{n-1}), r) \in E}$
. Then we may put 
 $m_0 = \max \{ \vert t \vert : t \in \bigcup _{s \in 2^{a_1}} \operatorname {\mathrm {term}}(h"(s)) \}$
. We may also assume wlog that
$m_0 = \max \{ \vert t \vert : t \in \bigcup _{s \in 2^{a_1}} \operatorname {\mathrm {term}}(h"(s)) \}$
. We may also assume wlog that 
 $\mathbb {Q} = 2^{<\omega }$
.
$\mathbb {Q} = 2^{<\omega }$
.
 To find such 
 $h"$
 we apply Lemma 4.15 as in the proof of Proposition 4.16. More precisely, for every
$h"$
 we apply Lemma 4.15 as in the proof of Proposition 4.16. More precisely, for every 
 $s \in 2^{a_0}$
, we find
$s \in 2^{a_0}$
, we find 
 $T^0_s, T^1_s \rhd _{\tilde \rho } h'(s)$
, and we find
$T^0_s, T^1_s \rhd _{\tilde \rho } h'(s)$
, and we find 
 $T \subseteq 2^{<\omega }$
 finite, so that for any pairwise distinct
$T \subseteq 2^{<\omega }$
 finite, so that for any pairwise distinct 
 $s_0, \dots , s_{n-1} \in \bigcup _{s \in 2^{a_0}} \operatorname {\mathrm {term}}(h'(s))$
, any
$s_0, \dots , s_{n-1} \in \bigcup _{s \in 2^{a_0}} \operatorname {\mathrm {term}}(h'(s))$
, any 
 $t_0 \supseteq s_0, \dots , t_{n-1} \supseteq s_{n-1}$
 with
$t_0 \supseteq s_0, \dots , t_{n-1} \supseteq s_{n-1}$
 with 
 $t_0, \dots , t_{n-1} \in \bigcup _{s \in 2^{a_0}, i \in 2} \operatorname {\mathrm {term}}(T^i_s)$
 and any
$t_0, \dots , t_{n-1} \in \bigcup _{s \in 2^{a_0}, i \in 2} \operatorname {\mathrm {term}}(T^i_s)$
 and any 
 $\sigma \in \operatorname {\mathrm {term}}(T)$
,
$\sigma \in \operatorname {\mathrm {term}}(T)$
, 
 $((t_0, \dots , t_{n-1}), \sigma ) \in E$
 and
$((t_0, \dots , t_{n-1}), \sigma ) \in E$
 and 
 $\operatorname {\mathrm {term}} (T_s^i) \cap \operatorname {\mathrm {term}} (T_{t}^j) = \emptyset $
 for every
$\operatorname {\mathrm {term}} (T_s^i) \cap \operatorname {\mathrm {term}} (T_{t}^j) = \emptyset $
 for every 
 $i, j \in 2$
,
$i, j \in 2$
, 
 $s, t \in 2^{a_0}$
. Then simply define
$s, t \in 2^{a_0}$
. Then simply define 
 $h" \leq h'$
 with
$h" \leq h'$
 with 
 $\operatorname {\mathrm {dom}}(h") = 2^{a_0 +1}$
, where
$\operatorname {\mathrm {dom}}(h") = 2^{a_0 +1}$
, where 
 $h"(s^{\frown } i) = T_s^i$
 for
$h"(s^{\frown } i) = T_s^i$
 for 
 $s \in 2^{a_0}$
,
$s \in 2^{a_0}$
, 
 $i \in 2$
.⊣
$i \in 2$
.⊣
 The function F is obviously continuous and 
 $X_{\delta +1 }$
 is a closed subset of
$X_{\delta +1 }$
 is a closed subset of 
 $(2^\omega )^{\delta +1}$
, with
$(2^\omega )^{\delta +1}$
, with 
 $X_{\delta +1} \restriction \delta _0 = (X_{\delta +1} \restriction \delta ) \restriction \delta _0 = X_\delta \restriction \delta _0 = X_{\delta _0}$
 for every
$X_{\delta +1} \restriction \delta _0 = (X_{\delta +1} \restriction \delta ) \restriction \delta _0 = X_\delta \restriction \delta _0 = X_{\delta _0}$
 for every 
 $\delta _0 < \delta + 1$
.
$\delta _0 < \delta + 1$
.
Proof of 
 $(1)_{\delta +1}$
 and
$(1)_{\delta +1}$
 and 
 $(2)_{\delta +1}$
.
$(2)_{\delta +1}$
.
 Let G be 
 $\mathbb {R}_{\delta +1}$
 generic over
$\mathbb {R}_{\delta +1}$
 generic over 
 $M_1$
 and
$M_1$
 and 
 $\bar y_0, \dots , \bar y_{n-1} \in \dot X_{\delta +1}[G] = X_{\delta +1}$
 be arbitrary. By the inductive assumption we have that
$\bar y_0, \dots , \bar y_{n-1} \in \dot X_{\delta +1}[G] = X_{\delta +1}$
 be arbitrary. By the inductive assumption we have that 
 $\bar y_0 \restriction \delta , \dots , \bar y_{n-1} \restriction \delta $
 are strongly
$\bar y_0 \restriction \delta , \dots , \bar y_{n-1} \restriction \delta $
 are strongly 
 $\langle 2^\omega : i < \delta \rangle $
-mCg over
$\langle 2^\omega : i < \delta \rangle $
-mCg over 
 $M_1$
. By the above claim, whenever
$M_1$
. By the above claim, whenever 
 $\bar y_i \restriction \delta \neq \bar y_j \restriction \delta $
, then
$\bar y_i \restriction \delta \neq \bar y_j \restriction \delta $
, then 
 $\bar y_i(\delta ) \neq \bar y_j(\delta )$
. Thus, for
$\bar y_i(\delta ) \neq \bar y_j(\delta )$
. Thus, for 
 $(1)_{\delta +1}$
, we only need to show that
$(1)_{\delta +1}$
, we only need to show that 
 $\bar y_0, \dots , \bar y_{n-1}$
 are mCg. Let
$\bar y_0, \dots , \bar y_{n-1}$
 are mCg. Let 
 $\operatorname {\mathrm {tp}}(\bar y_0, \dots , \bar y_{n-1}) = (\langle \xi _{l} : l < k \rangle ,\langle K_{l} : l < k \rangle , \langle U_i : i < n \rangle )$
,
$\operatorname {\mathrm {tp}}(\bar y_0, \dots , \bar y_{n-1}) = (\langle \xi _{l} : l < k \rangle ,\langle K_{l} : l < k \rangle , \langle U_i : i < n \rangle )$
, 
 $\operatorname {\mathrm {tp}}(\bar y_0 \restriction \delta , \dots , \bar y_{n-1} \restriction \delta ) = (\langle \eta _l : l \leq k' \rangle , \langle M_l : l < k \rangle , \langle U_i\restriction \delta : i < n \rangle )$
, and
$\operatorname {\mathrm {tp}}(\bar y_0 \restriction \delta , \dots , \bar y_{n-1} \restriction \delta ) = (\langle \eta _l : l \leq k' \rangle , \langle M_l : l < k \rangle , \langle U_i\restriction \delta : i < n \rangle )$
, and 
 $n' = \vert \{ y_i(\delta ) : i < n \} \vert = K_{k-1}$
. Then we may view a dense open subset of
$n' = \vert \{ y_i(\delta ) : i < n \} \vert = K_{k-1}$
. Then we may view a dense open subset of 
 $\prod _{l < k} (\bigotimes _{\xi \in [\xi _l,\xi _{l+1})} 2^{<\omega })^{K_l}$
 as a
$\prod _{l < k} (\bigotimes _{\xi \in [\xi _l,\xi _{l+1})} 2^{<\omega })^{K_l}$
 as a 
 $\prod _{l < k'} (\bigotimes _{\xi \in [\eta _l,\eta _{l+1})} 2^{<\omega })^{M_l}$
-name for a dense open subset of
$\prod _{l < k'} (\bigotimes _{\xi \in [\eta _l,\eta _{l+1})} 2^{<\omega })^{M_l}$
-name for a dense open subset of 
 $(2^{<\omega })^{n'}$
. To this end, let
$(2^{<\omega })^{n'}$
. To this end, let 
 $\dot D \in M_1$
 be a
$\dot D \in M_1$
 be a 
 $\prod _{l < k'} (\bigotimes _{\xi \in [\eta _l,\eta _{l+1})} 2^{<\omega })^{M_l}$
 name for a dense open subset of
$\prod _{l < k'} (\bigotimes _{\xi \in [\eta _l,\eta _{l+1})} 2^{<\omega })^{M_l}$
 name for a dense open subset of 
 $(2^{<\omega })^{n'}$
. Then we have, by
$(2^{<\omega })^{n'}$
. Then we have, by 
 $(2)_\delta $
, that
$(2)_\delta $
, that 
 $$ \begin{align*} \tilde D = \bigcap &\Big\{ \dot D[\bar y^{\prime}_0, \dots, \bar y^{\prime}_{n-1}] : \bar y^{\prime}_0, \dots, \bar y^{\prime}_{n-1} \in X_\delta, \\ &\qquad \operatorname{\mathrm{tp}}(\bar y^{\prime}_0, \dots, \bar y^{\prime}_{n-1}) = (\langle \eta_l : l \leq k' \rangle, \langle M_l : l < k' \rangle, \langle U_i \restriction \delta : i < n \rangle) \Big\} \end{align*} $$
$$ \begin{align*} \tilde D = \bigcap &\Big\{ \dot D[\bar y^{\prime}_0, \dots, \bar y^{\prime}_{n-1}] : \bar y^{\prime}_0, \dots, \bar y^{\prime}_{n-1} \in X_\delta, \\ &\qquad \operatorname{\mathrm{tp}}(\bar y^{\prime}_0, \dots, \bar y^{\prime}_{n-1}) = (\langle \eta_l : l \leq k' \rangle, \langle M_l : l < k' \rangle, \langle U_i \restriction \delta : i < n \rangle) \Big\} \end{align*} $$
is a dense open subset of 
 $(2^{<\omega })^{n'}$
 and
$(2^{<\omega })^{n'}$
 and 
 $\tilde D \in M_1[G\restriction \delta ]$
. By the above claim,
$\tilde D \in M_1[G\restriction \delta ]$
. By the above claim, 
 $y_0(\delta ), \dots , y_{n-1}(\delta )$
 are mCg over
$y_0(\delta ), \dots , y_{n-1}(\delta )$
 are mCg over 
 $M_1[G\restriction \delta ]$
 in
$M_1[G\restriction \delta ]$
 in 
 $2^\omega $
. Altogether, this shows that
$2^\omega $
. Altogether, this shows that 
 $\bar y_0, \dots , \bar y_{n-1}$
 are
$\bar y_0, \dots , \bar y_{n-1}$
 are 
 $\langle 2^\omega : i < \delta +1 \rangle $
-mCg over
$\langle 2^\omega : i < \delta +1 \rangle $
-mCg over 
 $M_1$
.
$M_1$
.
 For 
 $(2)_{\delta +1}$
, let
$(2)_{\delta +1}$
, let 
 $\dot D \in M_1$
 now be a
$\dot D \in M_1$
 now be a 
 $\prod _{l < k} (\bigotimes _{\xi \in [\xi _l,\xi _{l+1})} 2^{<\omega })^{K_l}$
-name for a dense open subset of
$\prod _{l < k} (\bigotimes _{\xi \in [\xi _l,\xi _{l+1})} 2^{<\omega })^{K_l}$
-name for a dense open subset of 
 $\mathbb {Q}$
. Consider a name
$\mathbb {Q}$
. Consider a name 
 $\dot E$
 in
$\dot E$
 in 
 $M_1$
 for the dense open subset of
$M_1$
 for the dense open subset of 
 ${(2^{<\omega })^{n'} \times \mathbb {Q}}$
, where for any
${(2^{<\omega })^{n'} \times \mathbb {Q}}$
, where for any 
 $\bar y^{\prime }_0, \dots , \bar y^{\prime }_{n-1} \in X_{\delta }$
, with
$\bar y^{\prime }_0, \dots , \bar y^{\prime }_{n-1} \in X_{\delta }$
, with 
 $\operatorname {\mathrm {tp}}(\bar y^{\prime }_0, \dots , \bar y^{\prime }_{n-1}) = (\langle \eta _l : l \leq k' \rangle , \langle M_l : l < k \rangle , \langle U_i\restriction \delta : i < n \rangle )$
,
$\operatorname {\mathrm {tp}}(\bar y^{\prime }_0, \dots , \bar y^{\prime }_{n-1}) = (\langle \eta _l : l \leq k' \rangle , \langle M_l : l < k \rangle , \langle U_i\restriction \delta : i < n \rangle )$
, 
 $$ \begin{align*} \dot E[\bar y^{\prime}_0 , \dots, \bar y^{\prime}_{n-1}] &= \{ (\bar t,r) : M_1[\bar y^{\prime}_0, \dots, \bar y^{\prime}_{n-1}] \models \\ &\qquad\bar t \Vdash r \in \dot D[\bar y^{\prime}_0, \dots, \bar y^{\prime}_{n-1}][\dot z_0, \dots, \dot z_{n'-1}]\}, \end{align*} $$
$$ \begin{align*} \dot E[\bar y^{\prime}_0 , \dots, \bar y^{\prime}_{n-1}] &= \{ (\bar t,r) : M_1[\bar y^{\prime}_0, \dots, \bar y^{\prime}_{n-1}] \models \\ &\qquad\bar t \Vdash r \in \dot D[\bar y^{\prime}_0, \dots, \bar y^{\prime}_{n-1}][\dot z_0, \dots, \dot z_{n'-1}]\}, \end{align*} $$
 where 
 $(\dot z_0, \dots , \dot z_{n'-1})$
 is a name for the
$(\dot z_0, \dots , \dot z_{n'-1})$
 is a name for the 
 $(2^{<\omega })^{n'}$
-generic. By
$(2^{<\omega })^{n'}$
-generic. By 
 $(2)_\delta $
, we have that
$(2)_\delta $
, we have that 
 $$ \begin{align*} \tilde E = \bigcap &\Big\{ \dot E[\bar y^{\prime}_0, \dots, \bar y^{\prime}_{n-1}] : \bar y^{\prime}_0, \dots, \bar y^{\prime}_{n-1} \in X_{\delta}, \\ &\qquad \operatorname{\mathrm{tp}}(\bar y^{\prime}_0, \dots, \bar y^{\prime}_{n-1}) = (\langle \eta_l : l \leq k' \rangle, \langle M_l : l < k \rangle, \langle U_i\restriction \delta : i < n \rangle) \Big\} \end{align*} $$
$$ \begin{align*} \tilde E = \bigcap &\Big\{ \dot E[\bar y^{\prime}_0, \dots, \bar y^{\prime}_{n-1}] : \bar y^{\prime}_0, \dots, \bar y^{\prime}_{n-1} \in X_{\delta}, \\ &\qquad \operatorname{\mathrm{tp}}(\bar y^{\prime}_0, \dots, \bar y^{\prime}_{n-1}) = (\langle \eta_l : l \leq k' \rangle, \langle M_l : l < k \rangle, \langle U_i\restriction \delta : i < n \rangle) \Big\} \end{align*} $$
is a dense open subset of 
 $(2^{<\omega })^{n'} \times \mathbb {Q}$
 and
$(2^{<\omega })^{n'} \times \mathbb {Q}$
 and 
 $\tilde E \in M_1[G\restriction \delta ]$
. Let
$\tilde E \in M_1[G\restriction \delta ]$
. Let 
 $m \in \omega $
 be large enough so that for any
$m \in \omega $
 be large enough so that for any 
 $i,j < n$
, if
$i,j < n$
, if 
 $U_i \neq U_j$
, then
$U_i \neq U_j$
, then 
 $\forall \bar y^{\prime }_i \in U_i \cap X_{\delta +1}, \bar y^{\prime }_j \in U_j \cap X_{\delta +1} (y^{\prime }_i(\delta ) \restriction m \neq y^{\prime }_j(\delta )\restriction m)$
. To see that such m exists, note that if
$\forall \bar y^{\prime }_i \in U_i \cap X_{\delta +1}, \bar y^{\prime }_j \in U_j \cap X_{\delta +1} (y^{\prime }_i(\delta ) \restriction m \neq y^{\prime }_j(\delta )\restriction m)$
. To see that such m exists, note that if 
 $U_i \neq U_j$
, then
$U_i \neq U_j$
, then 
 $U_i \cap X_{\delta +1}$
 and
$U_i \cap X_{\delta +1}$
 and 
 $U_j \cap X_{\delta +1}$
 are disjoint compact subsets of
$U_j \cap X_{\delta +1}$
 are disjoint compact subsets of 
 $X_{\delta +1}$
. By the claim, there is
$X_{\delta +1}$
. By the claim, there is 
 $r \in \mathbb {Q}$
 and
$r \in \mathbb {Q}$
 and 
 $m_0 \geq m$
 so that for any
$m_0 \geq m$
 so that for any 
 $z_0, \dots , z_{n'-1} \in \bigcup _{\bar y \in X_\delta }[S_{\bar y}]$
, if
$z_0, \dots , z_{n'-1} \in \bigcup _{\bar y \in X_\delta }[S_{\bar y}]$
, if 
 $z_0 \restriction m, \dots , z_{n'-1} \restriction m$
 are pairwise different, then
$z_0 \restriction m, \dots , z_{n'-1} \restriction m$
 are pairwise different, then 
 $((z_0 \restriction m_0, \dots , z_{n'-1} \restriction m_0), r ) \in \tilde E$
. Altogether we find that
$((z_0 \restriction m_0, \dots , z_{n'-1} \restriction m_0), r ) \in \tilde E$
. Altogether we find that 
 $$ \begin{align*} r \in \bigcap \Big\{ \dot D[\bar y^{\prime}_0, \dots, \bar y^{\prime}_{n-1}] &: \bar y^{\prime}_0, \dots, \bar y^{\prime}_{n-1} \in X_{\delta+1}, \\ &\operatorname{\mathrm{tp}}(\bar y^{\prime}_0, \dots, \bar y^{\prime}_{n-1}) = (\langle \xi_l : l \leq k \rangle, \langle K_l : l < k \rangle, \langle U_i : i < n \rangle) \Big\}. \end{align*} $$
$$ \begin{align*} r \in \bigcap \Big\{ \dot D[\bar y^{\prime}_0, \dots, \bar y^{\prime}_{n-1}] &: \bar y^{\prime}_0, \dots, \bar y^{\prime}_{n-1} \in X_{\delta+1}, \\ &\operatorname{\mathrm{tp}}(\bar y^{\prime}_0, \dots, \bar y^{\prime}_{n-1}) = (\langle \xi_l : l \leq k \rangle, \langle K_l : l < k \rangle, \langle U_i : i < n \rangle) \Big\}. \end{align*} $$
Of course the same argument can be carried out below any condition in 
 $\mathbb {Q}$
, showing that this set is dense. That it is open is also clear since it is the intersection of open subsets of a partial order.⊣
$\mathbb {Q}$
, showing that this set is dense. That it is open is also clear since it is the intersection of open subsets of a partial order.⊣
 Now let 
 $\delta \leq \alpha $
 be a limit ordinal.
$\delta \leq \alpha $
 be a limit ordinal.
Proof of 
 $(1)_\delta $
 and
$(1)_\delta $
 and 
 $(2)_{\delta }$
.
$(2)_{\delta }$
.
 Let G be 
 $\mathbb {R}_{\delta }$
-generic over
$\mathbb {R}_{\delta }$
-generic over 
 $M_1$
,
$M_1$
, 
 ${\bar y_0, \dots , \bar y_{n-1} \in \dot X_\delta [G] = X_\delta}$
, this time wlog pairwise distinct, and
${\bar y_0, \dots , \bar y_{n-1} \in \dot X_\delta [G] = X_\delta}$
, this time wlog pairwise distinct, and 
 $\operatorname {\mathrm {tp}}(\bar y_0, \dots , \bar y_{n-1}) = (\langle \xi _l : l \leq k \rangle , \langle K_l : l < k \rangle , \langle U_i : i < n \rangle )$
. We will make a genericity argument over
$\operatorname {\mathrm {tp}}(\bar y_0, \dots , \bar y_{n-1}) = (\langle \xi _l : l \leq k \rangle , \langle K_l : l < k \rangle , \langle U_i : i < n \rangle )$
. We will make a genericity argument over 
 $M_1$
 to show
$M_1$
 to show 
 $(1)_\delta $
 and
$(1)_\delta $
 and 
 $(2)_\delta $
. To this end, let
$(2)_\delta $
. To this end, let 
 $D_0 \subseteq \prod _{l < k} (\bigotimes _{\xi \in [\xi _l,\xi _{l+1})} 2^{<\omega })^{K_l}$
 be dense open,
$D_0 \subseteq \prod _{l < k} (\bigotimes _{\xi \in [\xi _l,\xi _{l+1})} 2^{<\omega })^{K_l}$
 be dense open, 
 $D_0 \in M_1$
, and let
$D_0 \in M_1$
, and let 
 $\dot D_1 \in M_1$
 be a
$\dot D_1 \in M_1$
 be a 
 $\prod _{l < k} (\bigotimes _{\xi \in [\xi _l,\xi _{l+1})} 2^{<\omega })^{K_l}$
-name for a dense open subset of
$\prod _{l < k} (\bigotimes _{\xi \in [\xi _l,\xi _{l+1})} 2^{<\omega })^{K_l}$
-name for a dense open subset of 
 $\mathbb {Q}$
. Then consider the dense open subset
$\mathbb {Q}$
. Then consider the dense open subset 
 $D_2 \subseteq \prod _{l < k} (\bigotimes _{\xi \in [\xi _l,\xi _{l+1})} 2^{<\omega })^{K_l} \times \mathbb {Q}$
 in
$D_2 \subseteq \prod _{l < k} (\bigotimes _{\xi \in [\xi _l,\xi _{l+1})} 2^{<\omega })^{K_l} \times \mathbb {Q}$
 in 
 $M_1$
, where
$M_1$
, where 
 $$ \begin{align*}D_2 = \{ (r_0, r_1) : r_0 \in D_0 \wedge r_0 \Vdash r_1 \in \dot D_1 \}.\end{align*} $$
$$ \begin{align*}D_2 = \{ (r_0, r_1) : r_0 \in D_0 \wedge r_0 \Vdash r_1 \in \dot D_1 \}.\end{align*} $$
Also let 
 $\bar h_0 \in G$
 be an arbitrary condition so that
$\bar h_0 \in G$
 be an arbitrary condition so that 
 $$ \begin{align*}\bar h_0 \Vdash \forall i < n (U_i \cap \dot X_\delta \neq \emptyset).\end{align*} $$
$$ \begin{align*}\bar h_0 \Vdash \forall i < n (U_i \cap \dot X_\delta \neq \emptyset).\end{align*} $$
Then there is 
 $\delta _0 < \delta $
 so that
$\delta _0 < \delta $
 so that 
 $\operatorname {\mathrm {supp}}(\bar h_0), \xi _{k-1}+1 \subseteq \delta _0$
. We may equally well view
$\operatorname {\mathrm {supp}}(\bar h_0), \xi _{k-1}+1 \subseteq \delta _0$
. We may equally well view 
 $D_2$
 as a
$D_2$
 as a 
 $\prod _{l < k-1} (\bigotimes _{\xi \in [\xi _l,\xi _{l+1})} 2^{<\omega })^{K_l} \times (\bigotimes _{\xi \in [\xi _{k-1},\delta _0)} 2^{<\omega })^{K_{k-1}}$
-name
$\prod _{l < k-1} (\bigotimes _{\xi \in [\xi _l,\xi _{l+1})} 2^{<\omega })^{K_l} \times (\bigotimes _{\xi \in [\xi _{k-1},\delta _0)} 2^{<\omega })^{K_{k-1}}$
-name 
 $\dot E \in M_1$
 for a dense open subset
$\dot E \in M_1$
 for a dense open subset 
 $$ \begin{align*}E \subseteq \left(\bigotimes_{\xi \in [\delta_0,\xi_k)} 2^{<\omega}\right)^{K_{k-1}} \times \mathbb{Q}= \left(\bigotimes_{\xi \in [\delta_0,\delta)} 2^{<\omega}\right)^{n} \times \mathbb{Q}.\end{align*} $$
$$ \begin{align*}E \subseteq \left(\bigotimes_{\xi \in [\delta_0,\xi_k)} 2^{<\omega}\right)^{K_{k-1}} \times \mathbb{Q}= \left(\bigotimes_{\xi \in [\delta_0,\delta)} 2^{<\omega}\right)^{n} \times \mathbb{Q}.\end{align*} $$
 We follow again from 
 $(2)_{\delta _0}$
 that the set
$(2)_{\delta _0}$
 that the set 
 $\tilde E \in M_1[G \cap \mathbb {R}_{\delta _0}]$
, where
$\tilde E \in M_1[G \cap \mathbb {R}_{\delta _0}]$
, where 
 $$ \begin{align*} \tilde E &= \bigcap \Big\{ \dot E[\bar y^{\prime}_0, \dots, \bar y^{\prime}_{n-1}] : \bar y^{\prime}_0, \dots, \bar y^{\prime}_{n-1} \in X_{\delta_0}, \\ &\quad\operatorname{\mathrm{tp}}(\bar y^{\prime}_0, \dots, \bar y^{\prime}_{n-1}) = (\langle \xi_0 < \dots < \xi_{k-1} < \delta_0 \rangle, \langle K_l : l < k \rangle, \langle U_i \restriction \delta_0 : i < n \rangle) \Big\} \end{align*} $$
$$ \begin{align*} \tilde E &= \bigcap \Big\{ \dot E[\bar y^{\prime}_0, \dots, \bar y^{\prime}_{n-1}] : \bar y^{\prime}_0, \dots, \bar y^{\prime}_{n-1} \in X_{\delta_0}, \\ &\quad\operatorname{\mathrm{tp}}(\bar y^{\prime}_0, \dots, \bar y^{\prime}_{n-1}) = (\langle \xi_0 < \dots < \xi_{k-1} < \delta_0 \rangle, \langle K_l : l < k \rangle, \langle U_i \restriction \delta_0 : i < n \rangle) \Big\} \end{align*} $$
is dense open. Let 
 $((\bar t_0, \dots , \bar t_{n-1}), r) \in \tilde E$
 be arbitrary and
$((\bar t_0, \dots , \bar t_{n-1}), r) \in \tilde E$
 be arbitrary and 
 $\bar h_1 \in G \cap \mathbb {R}_{\delta _0}$
,
$\bar h_1 \in G \cap \mathbb {R}_{\delta _0}$
, 
 $\bar h_1 \leq \bar h_0$
, so that
$\bar h_1 \leq \bar h_0$
, so that 
 $\bar h_1 \Vdash ((\bar t_0, \dots , \bar t_{n-1}), r) \in \tilde E$
.
$\bar h_1 \Vdash ((\bar t_0, \dots , \bar t_{n-1}), r) \in \tilde E$
.
 Let us show by induction on 
 $\xi \in [\delta _0, \delta )$
,
$\xi \in [\delta _0, \delta )$
, 
 $\xi> \sup \left ( \bigcup _{i < n} \operatorname {\mathrm {dom}}(\bar t_i) \right )$
, that there is a condition
$\xi> \sup \left ( \bigcup _{i < n} \operatorname {\mathrm {dom}}(\bar t_i) \right )$
, that there is a condition 
 $\bar h_2 \in \mathbb {R}_\xi , \bar h_2 \leq \bar h_1$
, so that
$\bar h_2 \in \mathbb {R}_\xi , \bar h_2 \leq \bar h_1$
, so that 
 $$ \begin{align*} \bar h_2 \Vdash \forall \bar y^{\prime}_0, \dots, \bar y^{\prime}_{n-1} \in \dot X_\delta \big(\operatorname{\mathrm{tp}}(\bar y^{\prime}_0, \dots, \bar y^{\prime}_{n-1}) &= (\langle \xi_l : l \leq k \rangle, \langle K_l : l < k \rangle, \langle U_i : i < n \rangle) \\ &\quad\rightarrow \bar y^{\prime}_0 \in [\bar t_0] \wedge \dots \wedge \bar y^{\prime}_{n-1} \in [\bar t_{n-1}] \big) \end{align*} $$
$$ \begin{align*} \bar h_2 \Vdash \forall \bar y^{\prime}_0, \dots, \bar y^{\prime}_{n-1} \in \dot X_\delta \big(\operatorname{\mathrm{tp}}(\bar y^{\prime}_0, \dots, \bar y^{\prime}_{n-1}) &= (\langle \xi_l : l \leq k \rangle, \langle K_l : l < k \rangle, \langle U_i : i < n \rangle) \\ &\quad\rightarrow \bar y^{\prime}_0 \in [\bar t_0] \wedge \dots \wedge \bar y^{\prime}_{n-1} \in [\bar t_{n-1}] \big) \end{align*} $$
and in particular, if 
 $\bar h_2 \in G$
, then for all
$\bar h_2 \in G$
, then for all 
 $\bar y^{\prime }_0, \dots , \bar y^{\prime }_{n-1} \in X_\delta $
 with
$\bar y^{\prime }_0, \dots , \bar y^{\prime }_{n-1} \in X_\delta $
 with 
 $\operatorname {\mathrm {tp}}(\bar y^{\prime }_0, \dots , \bar y^{\prime }_{n-1}) = (\langle \xi _l : l \leq k \rangle , \langle K_l : l < k \rangle , \langle U_i : i < n \rangle )$
, the generic corresponding to
$\operatorname {\mathrm {tp}}(\bar y^{\prime }_0, \dots , \bar y^{\prime }_{n-1}) = (\langle \xi _l : l \leq k \rangle , \langle K_l : l < k \rangle , \langle U_i : i < n \rangle )$
, the generic corresponding to 
 $\bar y^{\prime }_0, \dots , \bar y^{\prime }_{n-1}$
 hits
$\bar y^{\prime }_0, \dots , \bar y^{\prime }_{n-1}$
 hits 
 $D_0$
, and
$D_0$
, and 
 $r \in \dot D_1[\bar y^{\prime }_0, \dots , \bar y^{\prime }_{n-1}]$
. Since
$r \in \dot D_1[\bar y^{\prime }_0, \dots , \bar y^{\prime }_{n-1}]$
. Since 
 $\bar h_0 \in G$
 was arbitrary, genericity finishes the argument.
$\bar h_0 \in G$
 was arbitrary, genericity finishes the argument.
 The limit step of the induction follows directly from the earlier steps since if 
 $\operatorname {\mathrm {dom}} (\bar t_i) \subseteq \xi $
, with
$\operatorname {\mathrm {dom}} (\bar t_i) \subseteq \xi $
, with 
 $\xi $
 limit, then there is
$\xi $
 limit, then there is 
 $\eta < \xi $
 so that
$\eta < \xi $
 so that 
 $\operatorname {\mathrm {dom}} (\bar t_i) \subseteq \eta $
. So let us consider step
$\operatorname {\mathrm {dom}} (\bar t_i) \subseteq \eta $
. So let us consider step 
 $\xi +1$
. Then there is, by the inductive assumption,
$\xi +1$
. Then there is, by the inductive assumption, 
 $\bar h^{\prime }_2 \in \mathbb {R}_\xi $
,
$\bar h^{\prime }_2 \in \mathbb {R}_\xi $
, 
 $\bar h^{\prime }_2 \leq \bar h_1$
, so that
$\bar h^{\prime }_2 \leq \bar h_1$
, so that 
 $$ \begin{align*} \bar h^{\prime}_2 \Vdash \forall \bar y^{\prime}_0, \dots, \bar y^{\prime}_{n-1} \in \dot X_\delta \big(\operatorname{\mathrm{tp}}(\bar y^{\prime}_0, \dots, \bar y^{\prime}_{n-1}) &= (\langle \xi_l : l \leq k \rangle, \langle K_l : l < k \rangle, \langle U_i : i < n \rangle) \\ &\rightarrow ( \bar y^{\prime}_0 \in [\bar t_0 \restriction \xi] \wedge \dots \wedge \bar y^{\prime}_{n-1} \in [\bar t_{n-1} \restriction \xi] \big). \end{align*} $$
$$ \begin{align*} \bar h^{\prime}_2 \Vdash \forall \bar y^{\prime}_0, \dots, \bar y^{\prime}_{n-1} \in \dot X_\delta \big(\operatorname{\mathrm{tp}}(\bar y^{\prime}_0, \dots, \bar y^{\prime}_{n-1}) &= (\langle \xi_l : l \leq k \rangle, \langle K_l : l < k \rangle, \langle U_i : i < n \rangle) \\ &\rightarrow ( \bar y^{\prime}_0 \in [\bar t_0 \restriction \xi] \wedge \dots \wedge \bar y^{\prime}_{n-1} \in [\bar t_{n-1} \restriction \xi] \big). \end{align*} $$
 Now extend 
 $\bar h^{\prime }_2$
 to
$\bar h^{\prime }_2$
 to 
 $\bar h^{\prime \prime }_2$
 in
$\bar h^{\prime \prime }_2$
 in 
 $\mathbb {R}_\xi $
, so that there is
$\mathbb {R}_\xi $
, so that there is 
 $m \in \omega $
 such that for every
$m \in \omega $
 such that for every 
 $s \in 2^{m}$
 and every
$s \in 2^{m}$
 and every 
 $i < n$
, either
$i < n$
, either 
 $\bar h^{\prime \prime }_2 \Vdash \dot O_{s} \subseteq U_i \restriction \xi $
 or
$\bar h^{\prime \prime }_2 \Vdash \dot O_{s} \subseteq U_i \restriction \xi $
 or 
 $\bar h^{\prime \prime }_2 \Vdash \dot O_{s} \cap (U_i \restriction \xi ) = \emptyset $
, where
$\bar h^{\prime \prime }_2 \Vdash \dot O_{s} \cap (U_i \restriction \xi ) = \emptyset $
, where 
 $\langle \dot O_s : s \in 2^{<\omega } \rangle $
 is a name for the base of
$\langle \dot O_s : s \in 2^{<\omega } \rangle $
 is a name for the base of 
 $\dot X_\xi $
 used to define
$\dot X_\xi $
 used to define 
 $\dot {\mathbb {S}}_\xi $
. The reason why this is possible is that in any extension by
$\dot {\mathbb {S}}_\xi $
. The reason why this is possible is that in any extension by 
 $\mathbb {R}_\xi $
 and for every
$\mathbb {R}_\xi $
 and for every 
 $i < n$
, by compactness of
$i < n$
, by compactness of 
 $X_\xi \cap (U_i \restriction \xi )$
, there is a finite set
$X_\xi \cap (U_i \restriction \xi )$
, there is a finite set 
 $a \subseteq 2^{<\omega }$
 so that
$a \subseteq 2^{<\omega }$
 so that 
 $X_\xi \cap (U_i \restriction \xi ) = \bigcup _{s \in a} O_s$
. Let us define
$X_\xi \cap (U_i \restriction \xi ) = \bigcup _{s \in a} O_s$
. Let us define 
 $h \colon 2^{\leq m} \to \mathcal {FT}$
, where
$h \colon 2^{\leq m} \to \mathcal {FT}$
, where 
 $$ \begin{align*}h(s) =\begin{cases} \emptyset, & \text{if } \forall i < n ( \bar h^{\prime\prime}_2 \Vdash \dot O_{s} \cap U_i \restriction \xi = \emptyset), \\ \{ t \in 2^{<\omega} : t \subseteq t_i(\xi) \}, & \text{if } \bar h^{\prime\prime}_2 \Vdash \dot O_{s} \subseteq U_i \restriction \xi \text{ and } i < n. \end{cases}\end{align*} $$
$$ \begin{align*}h(s) =\begin{cases} \emptyset, & \text{if } \forall i < n ( \bar h^{\prime\prime}_2 \Vdash \dot O_{s} \cap U_i \restriction \xi = \emptyset), \\ \{ t \in 2^{<\omega} : t \subseteq t_i(\xi) \}, & \text{if } \bar h^{\prime\prime}_2 \Vdash \dot O_{s} \subseteq U_i \restriction \xi \text{ and } i < n. \end{cases}\end{align*} $$
 Note that h is well-defined as 
 $(U_i \restriction \xi ) \cap (U_j \restriction \xi ) = \emptyset $
 for every
$(U_i \restriction \xi ) \cap (U_j \restriction \xi ) = \emptyset $
 for every 
 $i \neq j < n$
. Since
$i \neq j < n$
. Since 
 $\emptyset \unlhd _{\rho } T$
 and
$\emptyset \unlhd _{\rho } T$
 and 
 $T \unlhd _{\rho } T$
 for any weight
$T \unlhd _{\rho } T$
 for any weight 
 $\rho $
 and any finite tree T, we have that
$\rho $
 and any finite tree T, we have that 
 $\bar h^{\prime \prime }_2 \Vdash h \in \dot {\mathbb {S}}_\xi $
 and
$\bar h^{\prime \prime }_2 \Vdash h \in \dot {\mathbb {S}}_\xi $
 and 
 $\bar h_2 = \bar h^{\prime \prime }_2{}^{\frown } h \in \mathbb {R}_{\xi +1}$
 is as required.⊣
$\bar h_2 = \bar h^{\prime \prime }_2{}^{\frown } h \in \mathbb {R}_{\xi +1}$
 is as required.⊣
 This finishes the definition of 
 $\mathbb {R}_\alpha $
 and
$\mathbb {R}_\alpha $
 and 
 $\dot X_\alpha $
. Finally let G be
$\dot X_\alpha $
. Finally let G be 
 $\mathbb {R}_\alpha $
-generic over
$\mathbb {R}_\alpha $
-generic over 
 $M_1$
 and
$M_1$
 and 
 $X_\alpha = \dot X_\alpha [G]$
. Now let us define
$X_\alpha = \dot X_\alpha [G]$
. Now let us define 
 $\bar q \leq \bar p$
 recursively so that for every
$\bar q \leq \bar p$
 recursively so that for every 
 $\delta \leq \alpha $
,
$\delta \leq \alpha $
, 
 $$ \begin{align*}\forall \bar x \in [\bar q] (\Phi(\bar x \restriction C) \restriction \delta \in X_\alpha \restriction \delta).\end{align*} $$
$$ \begin{align*}\forall \bar x \in [\bar q] (\Phi(\bar x \restriction C) \restriction \delta \in X_\alpha \restriction \delta).\end{align*} $$
 If 
 $\beta \notin C$
 we let
$\beta \notin C$
 we let 
 $\dot q(\beta )$
 be a name for the trivial condition
$\dot q(\beta )$
 be a name for the trivial condition 
 $2^{<\omega }$
, say, e.g.,
$2^{<\omega }$
, say, e.g., 
 $\dot q(\beta ) = \dot p(\beta )$
. If
$\dot q(\beta ) = \dot p(\beta )$
. If 
 $\beta \in C$
, say
$\beta \in C$
, say 
 $\beta = \iota (\delta )$
, we define
$\beta = \iota (\delta )$
, we define 
 $\dot q(\beta )$
 to be a name for the tree generated by
$\dot q(\beta )$
 to be a name for the tree generated by 
 $$ \begin{align*}\eta^{-1}_{T_{\bar x_{G} \restriction (C \cap \beta) }} (S_{\bar y}),\end{align*} $$
$$ \begin{align*}\eta^{-1}_{T_{\bar x_{G} \restriction (C \cap \beta) }} (S_{\bar y}),\end{align*} $$
where 
 $\bar x_{G}$
 is the generic sequence added by
$\bar x_{G}$
 is the generic sequence added by 
 $\mathbb {P}_\lambda $
 and
$\mathbb {P}_\lambda $
 and 
 $\bar y = \Phi (\bar x_G \restriction C) \restriction \delta $
. This ensures that
$\bar y = \Phi (\bar x_G \restriction C) \restriction \delta $
. This ensures that 
 $\bar q \restriction \beta \Vdash \dot q(\beta ) \in \mathbb {Q}_\beta \wedge \dot q(\beta ) \leq \dot p(\beta )$
. Inductively we see that
$\bar q \restriction \beta \Vdash \dot q(\beta ) \in \mathbb {Q}_\beta \wedge \dot q(\beta ) \leq \dot p(\beta )$
. Inductively we see that 
 $\bar q \restriction \beta ^{\frown } \bar p \restriction (\lambda \setminus \beta ) \Vdash \Phi (\bar x_G \restriction C) \restriction \delta \in X_\alpha \restriction \delta $
. Having defined
$\bar q \restriction \beta ^{\frown } \bar p \restriction (\lambda \setminus \beta ) \Vdash \Phi (\bar x_G \restriction C) \restriction \delta \in X_\alpha \restriction \delta $
. Having defined 
 $\bar q$
, it is also easy to check that it is a good master condition over
$\bar q$
, it is also easy to check that it is a good master condition over 
 $M_0$
, with
$M_0$
, with 
 $[\bar q] = \Phi ^{-1}(X_\alpha ) \times (2^{\omega })^{\lambda \setminus C}$
. Since for every
$[\bar q] = \Phi ^{-1}(X_\alpha ) \times (2^{\omega })^{\lambda \setminus C}$
. Since for every 
 $\bar x \in [\bar q]$
,
$\bar x \in [\bar q]$
, 
 $\Phi (\bar x \restriction C) \in X_\alpha $
 and by
$\Phi (\bar x \restriction C) \in X_\alpha $
 and by 
 $(1)_\alpha $
,
$(1)_\alpha $
, 
 $\bar q$
 is as required.⊣
$\bar q$
 is as required.⊣
Proposition 4.23. Let 
 $E \subseteq [X]^{<\omega } $
 be an analytic hypergraph on X, say E is the projection of a closed set
$E \subseteq [X]^{<\omega } $
 be an analytic hypergraph on X, say E is the projection of a closed set 
 $F \subseteq [X]^{<\omega } \times \omega ^\omega $
, and let
$F \subseteq [X]^{<\omega } \times \omega ^\omega $
, and let 
 $f \colon [\bar p] \restriction C \to X$
 be continuous so that
$f \colon [\bar p] \restriction C \to X$
 be continuous so that 
 $\bar p \Vdash \dot y = f(\bar x_G \restriction C)$
 (cf. Lemma 2.3). Then there is a good master condition
$\bar p \Vdash \dot y = f(\bar x_G \restriction C)$
 (cf. Lemma 2.3). Then there is a good master condition 
 $\bar q \leq \bar p$
, with
$\bar q \leq \bar p$
, with 
 $[\bar q] \restriction C$
 a closed subset of
$[\bar q] \restriction C$
 a closed subset of 
 $[\bar p] \restriction C$
 and
$[\bar p] \restriction C$
 and 
 $[\bar q] = ([\bar q] \restriction C) \times (2^\omega )^{\lambda \setminus C}$
, a compact E-independent set
$[\bar q] = ([\bar q] \restriction C) \times (2^\omega )^{\lambda \setminus C}$
, a compact E-independent set 
 $Y \subseteq X$
,
$Y \subseteq X$
, 
 $N \in \omega $
 and continuous functions
$N \in \omega $
 and continuous functions 
 $\phi \colon [\bar q] \restriction C \to [Y]^{<N}$
,
$\phi \colon [\bar q] \restriction C \to [Y]^{<N}$
, 
 $w \colon [\bar q] \restriction C \to \omega ^\omega $
, so that
$w \colon [\bar q] \restriction C \to \omega ^\omega $
, so that
- 
(i) either  $f"([\bar q] \restriction C) \subseteq Y$
, thus $f"([\bar q] \restriction C) \subseteq Y$
, thus $\bar q \Vdash \dot y \in Y$
, $\bar q \Vdash \dot y \in Y$
,
- 
(ii) or  $\forall \bar x \in [\bar q]\restriction C ( (\phi (\bar x) \cup \{ f(\bar x) \}, w(\bar x)) \in F)$
, thus $\forall \bar x \in [\bar q]\restriction C ( (\phi (\bar x) \cup \{ f(\bar x) \}, w(\bar x)) \in F)$
, thus $\bar q \Vdash {\{\dot y\} \cup Y}$
 is not $\bar q \Vdash {\{\dot y\} \cup Y}$
 is not $E\text {-independent}$
. $E\text {-independent}$
.
Proof. On 
 $(2^\omega )^\alpha $
 let us define the analytic hypergraph
$(2^\omega )^\alpha $
 let us define the analytic hypergraph 
 $\tilde E$
, where
$\tilde E$
, where 
 $$ \begin{align*}\{ \bar y_0, \dots, \bar y_{n-1} \} \in \tilde E \leftrightarrow \{ f(\Phi^{-1}(\bar y_0), \dots,f(\Phi^{-1}(\bar y_{n-1})) \} \in E.\end{align*} $$
$$ \begin{align*}\{ \bar y_0, \dots, \bar y_{n-1} \} \in \tilde E \leftrightarrow \{ f(\Phi^{-1}(\bar y_0), \dots,f(\Phi^{-1}(\bar y_{n-1})) \} \in E.\end{align*} $$
By Main Lemma 3.14, there is a countable model M and 
 $\bar s \in \bigotimes _{i<\alpha } 2^{<\omega }$
 so that either
$\bar s \in \bigotimes _{i<\alpha } 2^{<\omega }$
 so that either 
- 
(1) for any  $\bar y_0,\dots , \bar y_{n-1} \in (2^\omega )^\alpha \cap [\bar s]$
 that are strongly $\bar y_0,\dots , \bar y_{n-1} \in (2^\omega )^\alpha \cap [\bar s]$
 that are strongly $\langle 2^\omega : i < \alpha \rangle $
-mCg over M, $\langle 2^\omega : i < \alpha \rangle $
-mCg over M, $\{\bar y_0,\dots , \bar y_{n-1}\} \text { is } E\text {-independent}$
, $\{\bar y_0,\dots , \bar y_{n-1}\} \text { is } E\text {-independent}$
,
 or for some 
 $N\in \omega $
,
$N\in \omega $
, 
- 
(2) there are  $\phi _0, \dots , \phi _{N-1} \colon (2^\omega )^\alpha \to (2^\omega )^\alpha $
 continuous so that for any $\phi _0, \dots , \phi _{N-1} \colon (2^\omega )^\alpha \to (2^\omega )^\alpha $
 continuous so that for any $\bar y_0,\dots , \bar y_{n-1} \in (2^\omega )^\alpha \cap [\bar s]$
 that are strongly mCg over M, $\bar y_0,\dots , \bar y_{n-1} \in (2^\omega )^\alpha \cap [\bar s]$
 that are strongly mCg over M, $\{\phi _j(\bar y_i) : j < N, i<n\}$
 is E-independent but $\{\phi _j(\bar y_i) : j < N, i<n\}$
 is E-independent but $\{ \bar y_0\} \cup \{ \phi _j(\bar y_0) : j < N \} \in E.$ $\{ \bar y_0\} \cup \{ \phi _j(\bar y_0) : j < N \} \in E.$
 Let 
 $M_1$
 be a countable elementary model with
$M_1$
 be a countable elementary model with 
 $M_0, M, \bar p, \mathbb {P}_\lambda \in M_{1}$
 and apply Lemma 4.21 to get the condition
$M_0, M, \bar p, \mathbb {P}_\lambda \in M_{1}$
 and apply Lemma 4.21 to get the condition 
 $\bar q \leq \bar p$
. In case (1), let
$\bar q \leq \bar p$
. In case (1), let 
 $Y := f"([\bar q] \restriction C)$
. Then (i) is satisfied. To see that Y is E-independent let
$Y := f"([\bar q] \restriction C)$
. Then (i) is satisfied. To see that Y is E-independent let 
 $\bar x_0, \dots , \bar x_{n-1} \in [\bar q]$
 be arbitrary and suppose that
$\bar x_0, \dots , \bar x_{n-1} \in [\bar q]$
 be arbitrary and suppose that 
 $\{{f(\bar x_0 \restriction C)}, \dots ,f(\bar x_{n-1} \restriction C) \} \in E$
. By definition of
$\{{f(\bar x_0 \restriction C)}, \dots ,f(\bar x_{n-1} \restriction C) \} \in E$
. By definition of 
 $\tilde E$
 this implies that
$\tilde E$
 this implies that 
 $\{ {\Phi (\bar x_0 \restriction C)}, \dots , \Phi (\bar x_{n-1} \restriction C) \} \in \tilde E$
 but this is a contradiction to (1) and the conclusion of Lemma 4.21. In case (2), by elementarity, the
$\{ {\Phi (\bar x_0 \restriction C)}, \dots , \Phi (\bar x_{n-1} \restriction C) \} \in \tilde E$
 but this is a contradiction to (1) and the conclusion of Lemma 4.21. In case (2), by elementarity, the 
 $\phi _j$
 are in
$\phi _j$
 are in 
 $M_1$
 and there is a continuous function
$M_1$
 and there is a continuous function 
 $\tilde w \in M_1$
, with domain some dense
$\tilde w \in M_1$
, with domain some dense 
 $G_\delta $
 subset of
$G_\delta $
 subset of 
 $(2^\omega )^\alpha $
, so that
$(2^\omega )^\alpha $
, so that 
 $\bar s \Vdash (\{f(\bar z), \phi _{j}(\bar z ) : j < N \}, \tilde w(\bar z)) \in F$
, where
$\bar s \Vdash (\{f(\bar z), \phi _{j}(\bar z ) : j < N \}, \tilde w(\bar z)) \in F$
, where 
 $\bar z$
 is a name for the Cohen generic. Let
$\bar z$
 is a name for the Cohen generic. Let 
 $\phi (\bar x) = \{ f(\Phi ^{-1}(\phi _j(\Phi (\bar x)))) : j < N \}$
,
$\phi (\bar x) = \{ f(\Phi ^{-1}(\phi _j(\Phi (\bar x)))) : j < N \}$
, 
 $w(\bar x) = \tilde w(\Phi (\bar x))$
 for
$w(\bar x) = \tilde w(\Phi (\bar x))$
 for 
 $\bar x \in [\bar q] \restriction C$
 and
$\bar x \in [\bar q] \restriction C$
 and 
 $Y := \bigcup _{\bar x \in [\bar q] \restriction C} \phi (\bar x)$
. Since
$Y := \bigcup _{\bar x \in [\bar q] \restriction C} \phi (\bar x)$
. Since 
 $\Phi (\bar x)$
 is generic over
$\Phi (\bar x)$
 is generic over 
 $M_1$
, we indeed have that
$M_1$
, we indeed have that 
 $(\phi (\bar x), w(\bar x)) \in F$
 for every
$(\phi (\bar x), w(\bar x)) \in F$
 for every 
 $\bar x \in [\bar q] \restriction C$
. Seeing that Y is E-independent is as before.⊣
$\bar x \in [\bar q] \restriction C$
. Seeing that Y is E-independent is as before.⊣
5 Main results and applications
Theorem 5.1. (V=L) Let 
 $\mathbb {P}$
 be a countable support iteration of Sacks or splitting forcing of arbitrary length. Let X be a Polish space and
$\mathbb {P}$
 be a countable support iteration of Sacks or splitting forcing of arbitrary length. Let X be a Polish space and 
 $E \subseteq [X]^{<\omega } $
 be an analytic hypergraph. Then there is a
$E \subseteq [X]^{<\omega } $
 be an analytic hypergraph. Then there is a 
 $\mathbf {\Delta }^1_2$
 maximal E-independent set in
$\mathbf {\Delta }^1_2$
 maximal E-independent set in 
 $V^{\mathbb {P}}$
. If
$V^{\mathbb {P}}$
. If 
 $X = 2^\omega $
 or
$X = 2^\omega $
 or 
 $X = \omega ^\omega $
,
$X = \omega ^\omega $
, 
 $r \in 2^\omega $
, and E is
$r \in 2^\omega $
, and E is 
 $\Sigma ^1_1(r)$
, then we can find a
$\Sigma ^1_1(r)$
, then we can find a 
 $\Delta ^1_2(r)$
 such set.
$\Delta ^1_2(r)$
 such set.
Proof. We will only concentrate on the case 
 $X = 2^\omega $
 since the rest follows easily from the fact that there is a Borel isomorphism from
$X = 2^\omega $
 since the rest follows easily from the fact that there is a Borel isomorphism from 
 $2^\omega $
 to any uncountable Polish space X, and if
$2^\omega $
 to any uncountable Polish space X, and if 
 $X = \omega ^\omega $
 that isomorphism is (lightface)
$X = \omega ^\omega $
 that isomorphism is (lightface) 
 $\Delta ^1_1$
. If X is countable, then the statement is trivial. Also, let us only consider splitting forcing. The proof for Sacks forcing is the same.
$\Delta ^1_1$
. If X is countable, then the statement is trivial. Also, let us only consider splitting forcing. The proof for Sacks forcing is the same.
 First let us mention some well-known facts and introduce some notation. Recall that a set 
 $Y \subseteq 2^\omega $
 is
$Y \subseteq 2^\omega $
 is 
 $\Sigma ^1_2(x)$
-definable if and only if it is
$\Sigma ^1_2(x)$
-definable if and only if it is 
 $\Sigma _1(x)$
-definable over
$\Sigma _1(x)$
-definable over 
 $H(\omega _1)$
 (see, e.g., [Reference Jech18, Lemma 25.25]). Also recall that there is a
$H(\omega _1)$
 (see, e.g., [Reference Jech18, Lemma 25.25]). Also recall that there is a 
 $\Sigma ^1_1$
 set
$\Sigma ^1_1$
 set 
 $A \subseteq 2^{\omega } \times 2^{\omega }$
 that is universal for analytic sets, i.e., for every analytic
$A \subseteq 2^{\omega } \times 2^{\omega }$
 that is universal for analytic sets, i.e., for every analytic 
 $B \subseteq 2^{\omega }$
, there is some
$B \subseteq 2^{\omega }$
, there is some 
 $x\in 2^\omega $
 so that
$x\in 2^\omega $
 so that 
 $B = A_x$
, where
$B = A_x$
, where 
 $A_x = \{ y \in 2^\omega : (x,y) \in A \}$
. In the same way, there is a universal
$A_x = \{ y \in 2^\omega : (x,y) \in A \}$
. In the same way, there is a universal 
 $\Pi ^0_1$
 set
$\Pi ^0_1$
 set 
 $F \subseteq 2^\omega \times [2^\omega ]^{<\omega } \times \omega ^\omega $
 [Reference Kechris19, Theorems 22.3 and 26.1]. For any
$F \subseteq 2^\omega \times [2^\omega ]^{<\omega } \times \omega ^\omega $
 [Reference Kechris19, Theorems 22.3 and 26.1]. For any 
 $x \in 2^\omega $
, let
$x \in 2^\omega $
, let 
 $E_x$
 be the analytic hypergraph on
$E_x$
 be the analytic hypergraph on 
 $2^\omega $
 consisting of
$2^\omega $
 consisting of 
 $a \in [2^{\omega }]^{<\omega } \setminus \{\emptyset \}$
 so that there is
$a \in [2^{\omega }]^{<\omega } \setminus \{\emptyset \}$
 so that there is 
 $b \in [A_x]^{<\omega }$
 with
$b \in [A_x]^{<\omega }$
 with 
 $a \cup b \in E$
. Then there is
$a \cup b \in E$
. Then there is 
 $y \in 2^\omega $
 so that
$y \in 2^\omega $
 so that 
 $E_x$
 is the projection of
$E_x$
 is the projection of 
 $F_y$
. Moreover, it is standard to note, from the way A and F are defined, that for every x,
$F_y$
. Moreover, it is standard to note, from the way A and F are defined, that for every x, 
 $y= e(x,r)$
 for some fixed recursive function e. Whenever
$y= e(x,r)$
 for some fixed recursive function e. Whenever 
 $\alpha < \omega _1$
 and
$\alpha < \omega _1$
 and 
 $Z \subseteq (2^\omega )^\alpha $
 is closed, it can be coded naturally by the set
$Z \subseteq (2^\omega )^\alpha $
 is closed, it can be coded naturally by the set 
 $S \subseteq \bigotimes _{i < \alpha } 2^{<\omega }$
, where
$S \subseteq \bigotimes _{i < \alpha } 2^{<\omega }$
, where 
 $$ \begin{align*}S = \{ (\bar x \restriction a) \restriction n : \bar x \in Z, a \in [\alpha]^{<\omega}, n \in \omega\}\end{align*} $$
$$ \begin{align*}S = \{ (\bar x \restriction a) \restriction n : \bar x \in Z, a \in [\alpha]^{<\omega}, n \in \omega\}\end{align*} $$
and we write 
 $Z = Z_S$
. Similarly, any continuous function
$Z = Z_S$
. Similarly, any continuous function 
 $f \colon Z \to \omega ^\omega $
 can be coded by a function
$f \colon Z \to \omega ^\omega $
 can be coded by a function 
 $\zeta \colon S \to \omega ^{<\omega }$
, where
$\zeta \colon S \to \omega ^{<\omega }$
, where 
 $$ \begin{align*}f(\bar x) = \bigcup_{\bar s \in S, \bar x \in [\bar s]} \zeta(\bar s)\end{align*} $$
$$ \begin{align*}f(\bar x) = \bigcup_{\bar s \in S, \bar x \in [\bar s]} \zeta(\bar s)\end{align*} $$
and we write 
 $f = f_\zeta $
. For any
$f = f_\zeta $
. For any 
 $\beta < \alpha $
 and
$\beta < \alpha $
 and 
 $\bar x \in Z \restriction \beta $
, let us write
$\bar x \in Z \restriction \beta $
, let us write 
 $T_{\bar x, Z} = \{ s \in 2^{<\omega } : \exists \bar z \in Z (\bar z \restriction \delta = \bar x \wedge s\subseteq z(\delta ))\}$
. The set
$T_{\bar x, Z} = \{ s \in 2^{<\omega } : \exists \bar z \in Z (\bar z \restriction \delta = \bar x \wedge s\subseteq z(\delta ))\}$
. The set 
 $\Psi _0$
 of pairs
$\Psi _0$
 of pairs 
 $(\alpha ,S)$
, where S codes a closed set
$(\alpha ,S)$
, where S codes a closed set 
 $Z \subseteq (2^\omega )^\alpha $
 so that for every
$Z \subseteq (2^\omega )^\alpha $
 so that for every 
 $\beta < \alpha $
 and
$\beta < \alpha $
 and 
 $\bar x \in Z \restriction \beta $
,
$\bar x \in Z \restriction \beta $
, 
 $T_{\bar x,Z} \in \mathbb {SP}$
 is then
$T_{\bar x,Z} \in \mathbb {SP}$
 is then 
 $\Delta _1$
 over
$\Delta _1$
 over 
 $H(\omega _1)$
. This follows since the set of such S is
$H(\omega _1)$
. This follows since the set of such S is 
 $\Pi ^1_1$
, seen as a subset of
$\Pi ^1_1$
, seen as a subset of 
 $\mathcal {P}(\bigotimes _{i < \alpha } 2^{<\omega })$
, uniformly on
$\mathcal {P}(\bigotimes _{i < \alpha } 2^{<\omega })$
, uniformly on 
 $\alpha $
. Similarly, the set
$\alpha $
. Similarly, the set 
 $\Psi _1$
 of triples
$\Psi _1$
 of triples 
 $(\alpha , S, \zeta )$
, where
$(\alpha , S, \zeta )$
, where 
 $(\alpha , S) \in \Psi _0$
 and
$(\alpha , S) \in \Psi _0$
 and 
 $\zeta $
 codes a continuous function
$\zeta $
 codes a continuous function 
 $f \colon Z_S \to \omega ^\omega $
, is
$f \colon Z_S \to \omega ^\omega $
, is 
 $\Delta _1$
.
$\Delta _1$
.
 Now let 
 $\langle \alpha _\xi , S_\xi , \zeta _\xi : \xi < \omega _1 \rangle $
 be a
$\langle \alpha _\xi , S_\xi , \zeta _\xi : \xi < \omega _1 \rangle $
 be a 
 $\Delta _1$
-definable enumeration of all triples
$\Delta _1$
-definable enumeration of all triples 
 $(\alpha , S, \zeta )\in \Psi _1$
. This is possible since we assume
$(\alpha , S, \zeta )\in \Psi _1$
. This is possible since we assume 
 $V=L$
 (cf. [Reference Jech18, Theorem 25.26]). Let us recursively construct a sequence
$V=L$
 (cf. [Reference Jech18, Theorem 25.26]). Let us recursively construct a sequence 
 $\langle x_\xi , y_\xi , T_\xi , \bar \eta _\xi , \theta _\xi : \xi < \omega _1 \rangle $
, where for each
$\langle x_\xi , y_\xi , T_\xi , \bar \eta _\xi , \theta _\xi : \xi < \omega _1 \rangle $
, where for each 
 $\xi < \omega _1$
,
$\xi < \omega _1$
, 
- 
(1)  $\bigcup _{\xi ' < \xi } A_{x_{\xi '}} = A_{y_\xi }$
 and $\bigcup _{\xi ' < \xi } A_{x_{\xi '}} = A_{y_\xi }$
 and $A_{y_\xi } \cup A_{x_\xi }$
 is E-independent, $A_{y_\xi } \cup A_{x_\xi }$
 is E-independent,
- 
(2)  $\bar \eta _\xi = \langle \eta _{\xi ,j} : j <N \rangle $
 for some $\bar \eta _\xi = \langle \eta _{\xi ,j} : j <N \rangle $
 for some $N \in \omega $
, $N \in \omega $
,
- 
(3)  $T_\xi \subseteq S_\xi $
, $T_\xi \subseteq S_\xi $
, $(\alpha _\xi , T_\xi , \eta _{\xi ,j}) \in \Psi _1$
 for every $(\alpha _\xi , T_\xi , \eta _{\xi ,j}) \in \Psi _1$
 for every $j < N$
 and $j < N$
 and $(\alpha _\xi , T_\xi , \theta _{\xi }) \in \Psi _1$
, $(\alpha _\xi , T_\xi , \theta _{\xi }) \in \Psi _1$
,
- 
(4) either  $\forall \bar x \in Z_{T_\xi } (f_{\zeta _\xi }(\bar x) \in A_{x_\xi })$
 or $\forall \bar x \in Z_{T_\xi } (f_{\zeta _\xi }(\bar x) \in A_{x_\xi })$
 or $\forall \bar x \in Z_{T_\xi } \big (\forall n < N (f_{\eta _{\xi ,n}}(\bar x) \in A_{x_\xi }) \wedge (\{ f_{\eta _{\xi ,n}}(\bar x), f_{\zeta _\xi }(\bar x): n < N\},f_{\theta _\xi }(\bar x) ) \in F_{e(y_\xi ,r)}\big )$
, $\forall \bar x \in Z_{T_\xi } \big (\forall n < N (f_{\eta _{\xi ,n}}(\bar x) \in A_{x_\xi }) \wedge (\{ f_{\eta _{\xi ,n}}(\bar x), f_{\zeta _\xi }(\bar x): n < N\},f_{\theta _\xi }(\bar x) ) \in F_{e(y_\xi ,r)}\big )$
,
and 
 $(x_\xi , y_\xi , T_\xi , \bar \eta _\xi , \theta _\xi )$
 is
$(x_\xi , y_\xi , T_\xi , \bar \eta _\xi , \theta _\xi )$
 is 
 $<_L$
-least such that (1)–(4), where
$<_L$
-least such that (1)–(4), where 
 $<_L$
 is the
$<_L$
 is the 
 $\Delta _1$
-good global well-order of L. That
$\Delta _1$
-good global well-order of L. That 
 $<_L$
 is
$<_L$
 is 
 $\Delta _1$
-good means that for every
$\Delta _1$
-good means that for every 
 $z \in L$
, the set
$z \in L$
, the set 
 $\{ z' : z' <_L z \}$
 is
$\{ z' : z' <_L z \}$
 is 
 $\Delta _1(z)$
 uniformly on the parameter z. In particular, quantifying over this set does not increase the complexity of a
$\Delta _1(z)$
 uniformly on the parameter z. In particular, quantifying over this set does not increase the complexity of a 
 $\Sigma _n$
-formula. Note that (1)–(4) are all
$\Sigma _n$
-formula. Note that (1)–(4) are all 
 $\Delta _1(r)$
 in the given variables. For example, the second part of (1) is uniformly
$\Delta _1(r)$
 in the given variables. For example, the second part of (1) is uniformly 
 $\Pi ^1_1(r)$
 in the variables
$\Pi ^1_1(r)$
 in the variables 
 $x_\xi , y_\xi $
, similarly for (4).
$x_\xi , y_\xi $
, similarly for (4).
Claim 5.2. For every 
 $\xi < \omega _1$
,
$\xi < \omega _1$
, 
 $(x_\xi , y_\xi , T_\xi , \bar \eta _\xi , \theta _\xi )$
 exists.
$(x_\xi , y_\xi , T_\xi , \bar \eta _\xi , \theta _\xi )$
 exists.
Proof. Assume we succeeded in constructing the sequence up to 
 $\xi $
. Then there is
$\xi $
. Then there is 
 $y_\xi $
 so that
$y_\xi $
 so that 
 $\bigcup _{\xi ' < \xi } A_{x_{\xi '}} = A_{y_\xi }$
. By Lemma 2.7, there is a good master condition
$\bigcup _{\xi ' < \xi } A_{x_{\xi '}} = A_{y_\xi }$
. By Lemma 2.7, there is a good master condition 
 $\bar r \in \mathbb {P}_{\alpha _\xi }$
 so that
$\bar r \in \mathbb {P}_{\alpha _\xi }$
 so that 
 $[\bar r] \subseteq Z_{S_\xi }$
, where
$[\bar r] \subseteq Z_{S_\xi }$
, where 
 $\mathbb {P}_{\alpha _\xi }$
 is the
$\mathbb {P}_{\alpha _\xi }$
 is the 
 $\alpha _\xi $
-long csi of splitting forcing. Then
$\alpha _\xi $
-long csi of splitting forcing. Then 
 $f_{\zeta _\xi }$
 corresponds to a
$f_{\zeta _\xi }$
 corresponds to a 
 $\mathbb {P}_{\alpha _\xi }$
-name
$\mathbb {P}_{\alpha _\xi }$
-name 
 $\dot y$
 so that
$\dot y$
 so that 
 $\bar r \Vdash \dot y = f_{\zeta _\xi }(\bar x_G)$
. Let
$\bar r \Vdash \dot y = f_{\zeta _\xi }(\bar x_G)$
. Let 
 $M_0$
 be a countable elementary model with
$M_0$
 be a countable elementary model with 
 $\dot y,\mathbb {P}_{\alpha _\xi }, \bar r \in M_0$
 and
$\dot y,\mathbb {P}_{\alpha _\xi }, \bar r \in M_0$
 and 
 $\bar p \leq \bar r$
 a good master condition over
$\bar p \leq \bar r$
 a good master condition over 
 $M_0$
. Let
$M_0$
. Let 
 $C := \alpha _\xi $
 and consider the results of the last section. By Proposition 4.23 applied to
$C := \alpha _\xi $
 and consider the results of the last section. By Proposition 4.23 applied to 
 $E_{y_\xi }$
 and
$E_{y_\xi }$
 and 
 $X = 2^\omega $
, there is
$X = 2^\omega $
, there is 
 $\bar q \leq \bar p$
, a compact
$\bar q \leq \bar p$
, a compact 
 $E_{y_\xi }$
 independent set
$E_{y_\xi }$
 independent set 
 $Y_\xi $
,
$Y_\xi $
, 
 $N \in \omega $
 and continuous functions
$N \in \omega $
 and continuous functions 
 $\phi \colon [\bar q] \to [Y_\xi ]^{<N}$
,
$\phi \colon [\bar q] \to [Y_\xi ]^{<N}$
, 
 $w \colon [\bar q] \to \omega ^\omega $
 such that
$w \colon [\bar q] \to \omega ^\omega $
 such that 
- 
(i) either  $f_{\zeta _\xi }"([\bar q]) \subseteq Y_\xi $
, $f_{\zeta _\xi }"([\bar q]) \subseteq Y_\xi $
,
- 
(ii) or  $\forall \bar x \in [\bar q] ( (\phi (\bar x) \cup \{ f_{\zeta _\xi }(\bar x) \}, w(\bar x)) \in F)$
. $\forall \bar x \in [\bar q] ( (\phi (\bar x) \cup \{ f_{\zeta _\xi }(\bar x) \}, w(\bar x)) \in F)$
.
 Let 
 $x_\xi $
,
$x_\xi $
, 
 $T_\xi $
,
$T_\xi $
, 
 $\bar \eta _\xi = \langle \eta _{\xi ,j} : j < N\rangle $
 and
$\bar \eta _\xi = \langle \eta _{\xi ,j} : j < N\rangle $
 and 
 $\theta _\xi $
 be such that
$\theta _\xi $
 be such that 
 $A_{x_\xi } = Y_\xi $
,
$A_{x_\xi } = Y_\xi $
, 
 $Z_{T_\xi } = [\bar q]$
,
$Z_{T_\xi } = [\bar q]$
, 
 $\{ f_{\eta _{\xi ,j}}(\bar x) : j < N \} = \phi (\bar x)$
 for every
$\{ f_{\eta _{\xi ,j}}(\bar x) : j < N \} = \phi (\bar x)$
 for every 
 $\bar x \in [\bar q]$
, and
$\bar x \in [\bar q]$
, and 
 $f_{\theta _\xi } = w$
. Then
$f_{\theta _\xi } = w$
. Then 
 $(x_\xi , y_\xi , T_\xi , \bar \eta _\xi , \theta _\xi )$
 is as required.⊣
$(x_\xi , y_\xi , T_\xi , \bar \eta _\xi , \theta _\xi )$
 is as required.⊣
 Let 
 $Y = \bigcup _{\xi < \omega _1} A_{x_\xi }$
. Then Y is
$Y = \bigcup _{\xi < \omega _1} A_{x_\xi }$
. Then Y is 
 $\Sigma _1(r)$
-definable over
$\Sigma _1(r)$
-definable over 
 $H(\omega _1)$
, namely
$H(\omega _1)$
, namely 
 $x \in Y$
 iff there is a sequence
$x \in Y$
 iff there is a sequence 
 $\langle x_\xi , y_\xi , T_\xi , \bar \eta _\xi , \theta _\xi : \xi \leq \alpha < \omega _1 \rangle $
 so that for every
$\langle x_\xi , y_\xi , T_\xi , \bar \eta _\xi , \theta _\xi : \xi \leq \alpha < \omega _1 \rangle $
 so that for every 
 $\xi \leq \alpha $
, (1)–(4), for every
$\xi \leq \alpha $
, (1)–(4), for every 
 $(x,y, T, \bar \eta , \theta ) <_L (x_\xi ,y_\xi , T_\xi , \bar \eta _\xi , \theta _\xi )$
, not (1)–(4), and
$(x,y, T, \bar \eta , \theta ) <_L (x_\xi ,y_\xi , T_\xi , \bar \eta _\xi , \theta _\xi )$
, not (1)–(4), and 
 $x \in A_{x_\alpha }$
.
$x \in A_{x_\alpha }$
.
Claim 5.3. In 
 $V^{\mathbb {P}}$
, the reinterpretation of Y is maximal E-independent.
$V^{\mathbb {P}}$
, the reinterpretation of Y is maximal E-independent.
Proof. Let 
 $\bar p \in \mathbb {P}$
 and
$\bar p \in \mathbb {P}$
 and 
 $\dot y \in M_0$
 be a
$\dot y \in M_0$
 be a 
 $\mathbb {P}$
-name for an element of
$\mathbb {P}$
-name for an element of 
 $2^\omega $
,
$2^\omega $
, 
 $M_0 \ni \mathbb {P}, \bar p$
 a countable elementary model. Then let
$M_0 \ni \mathbb {P}, \bar p$
 a countable elementary model. Then let 
 $\bar q \leq \bar p$
 be a good master condition over
$\bar q \leq \bar p$
 be a good master condition over 
 $M_0$
 and C countable,
$M_0$
 and C countable, 
 $f \colon [\bar q] \restriction C \to 2^\omega $
 continuous according to Lemma 2.3. Now
$f \colon [\bar q] \restriction C \to 2^\omega $
 continuous according to Lemma 2.3. Now 
 $(2^\omega )^{C}$
 is canonically homeomorphic to
$(2^\omega )^{C}$
 is canonically homeomorphic to 
 $(2^\omega )^\alpha $
,
$(2^\omega )^\alpha $
, 
 $\alpha = \operatorname {\mathrm {otp}}(C)$
, via the map
$\alpha = \operatorname {\mathrm {otp}}(C)$
, via the map 
 $\Phi \colon (2^\omega )^{C} \to (2^\omega )^\alpha $
. Then we find some
$\Phi \colon (2^\omega )^{C} \to (2^\omega )^\alpha $
. Then we find some 
 $\xi < \omega _1$
 so that
$\xi < \omega _1$
 so that 
 $\alpha _\xi = \alpha $
,
$\alpha _\xi = \alpha $
, 
 $\Phi "([\bar q] \restriction C) = Z_{S_\xi }$
, and
$\Phi "([\bar q] \restriction C) = Z_{S_\xi }$
, and 
 $f_{\zeta _\xi } \circ \Phi = f$
. On the other hand,
$f_{\zeta _\xi } \circ \Phi = f$
. On the other hand, 
 $\Phi ^{-1} (Z_{T_\xi })$
 is a subset of
$\Phi ^{-1} (Z_{T_\xi })$
 is a subset of 
 $[\bar q] \restriction C$
 conforming to the assumptions of Lemma 2.7. Thus we get
$[\bar q] \restriction C$
 conforming to the assumptions of Lemma 2.7. Thus we get 
 $\bar r \leq \bar q$
 so that
$\bar r \leq \bar q$
 so that 
 $[\bar r] \restriction C \subseteq \Phi ^{-1} (Z_{T_\xi })$
. According to (4), either
$[\bar r] \restriction C \subseteq \Phi ^{-1} (Z_{T_\xi })$
. According to (4), either 
 $\bar r \Vdash \dot y \in A_{x_{\xi }}$
 or
$\bar r \Vdash \dot y \in A_{x_{\xi }}$
 or 
 $\bar r \Vdash \{ \dot y\} \cup A_{x_\xi } \cup A_{y_\xi }$
 is not E-independent. Thus we cannot have that
$\bar r \Vdash \{ \dot y\} \cup A_{x_\xi } \cup A_{y_\xi }$
 is not E-independent. Thus we cannot have that 
 $\bar p \Vdash \dot y \notin Y \wedge \{\dot y\} \cup Y$
 is E-independent. This finishes the proof of the claim, as
$\bar p \Vdash \dot y \notin Y \wedge \{\dot y\} \cup Y$
 is E-independent. This finishes the proof of the claim, as 
 $\bar p$
 and
$\bar p$
 and 
 $\dot y$
 were arbitrary.⊣
$\dot y$
 were arbitrary.⊣
 To see that Y is 
 $\Delta ^1_2(r)$
 in
$\Delta ^1_2(r)$
 in 
 $V^{\mathbb {P}}$
 it suffices to observe that any
$V^{\mathbb {P}}$
 it suffices to observe that any 
 $\Sigma ^1_2(r)$
 set that is maximal E-independent is already
$\Sigma ^1_2(r)$
 set that is maximal E-independent is already 
 $\Pi ^1_2(r)$
.⊣
$\Pi ^1_2(r)$
.⊣
A priori, Theorem 5.1 only works for hypergraphs that are defined in the ground model. But note that there is a universal analytic hypergraph on 
 $2^\omega \times 2^\omega $
, whereby we can follow the more general statement of Theorem 1.2.
$2^\omega \times 2^\omega $
, whereby we can follow the more general statement of Theorem 1.2.
Theorem 5.4. After forcing with the 
 $\omega _2$
-length countable support iteration of
$\omega _2$
-length countable support iteration of 
 $\mathbb {SP}$
 over L, there is a
$\mathbb {SP}$
 over L, there is a 
 $\Delta ^1_2$
 ultrafilter, a
$\Delta ^1_2$
 ultrafilter, a 
 $\Pi ^1_1$
 maximal independent family, and a
$\Pi ^1_1$
 maximal independent family, and a 
 $\Delta ^1_2$
 Hamel basis, and in particular,
$\Delta ^1_2$
 Hamel basis, and in particular, 
 $\mathfrak {i}_{B} = \mathfrak {i}_{cl} = \mathfrak {u}_B = \mathfrak {u}_{cl} = \omega _1 < \mathfrak {r} = \mathfrak {i} = \mathfrak {u} = \omega _2$
.
$\mathfrak {i}_{B} = \mathfrak {i}_{cl} = \mathfrak {u}_B = \mathfrak {u}_{cl} = \omega _1 < \mathfrak {r} = \mathfrak {i} = \mathfrak {u} = \omega _2$
.
Proof. Apply Theorem 5.1 to 
 $E_u$
,
$E_u$
, 
 $E_i$
, and
$E_i$
, and 
 $E_h$
 from the introduction. To see that
$E_h$
 from the introduction. To see that 
 $\mathfrak {i}_{cl} = \mathfrak {u}_{cl} = \omega _1$
 note that every analytic set is the union of
$\mathfrak {i}_{cl} = \mathfrak {u}_{cl} = \omega _1$
 note that every analytic set is the union of 
 $\mathfrak {d}$
 many compact sets and that
$\mathfrak {d}$
 many compact sets and that 
 $\mathfrak {d} = \omega _1$
, since
$\mathfrak {d} = \omega _1$
, since 
 $\mathbb {SP}$
 is
$\mathbb {SP}$
 is 
 $\omega ^\omega $
-bounding.⊣
$\omega ^\omega $
-bounding.⊣
Theorem 5.5. (V=L) Let 
 $\mathbb {P}$
 be either Sacks or splitting forcing and
$\mathbb {P}$
 be either Sacks or splitting forcing and 
 $k \in \omega $
. Let X be a Polish space and
$k \in \omega $
. Let X be a Polish space and 
 $E \subseteq [X]^{<\omega }$
 be an analytic hypergraph. Then there is a
$E \subseteq [X]^{<\omega }$
 be an analytic hypergraph. Then there is a 
 $\mathbf {\Delta }^1_2$
 maximal E-independent set in
$\mathbf {\Delta }^1_2$
 maximal E-independent set in 
 $V^{\mathbb {P}^k}$
.
$V^{\mathbb {P}^k}$
.
Proof. This is similar to the proof of Theorem 5.1, using Main Lemma 3.4 and Proposition 4.16 to get an analogue of Proposition 4.23.⊣
Lastly, we are going to prove Theorem 1.4.
Lemma 5.6. Let 
 $X \subseteq [\omega ]^\omega $
 be closed so that
$X \subseteq [\omega ]^\omega $
 be closed so that 
 $\forall x,y \in X ( \vert x \cap y \vert = \omega )$
. Then X is
$\forall x,y \in X ( \vert x \cap y \vert = \omega )$
. Then X is 
 $\sigma $
-compact.
$\sigma $
-compact.
Proof. If not, then by Hurewicz’s Theorem (see [Reference Kechris19, Theorem 7.10]), there is a superperfect tree 
 $T \subseteq \omega ^{<\omega }$
 so that
$T \subseteq \omega ^{<\omega }$
 so that 
 $[T] \subseteq X$
, identifying elements of
$[T] \subseteq X$
, identifying elements of 
 $[\omega ]^\omega $
 with their increasing enumeration, as usual. But then it is easy to recursively construct increasing sequences
$[\omega ]^\omega $
 with their increasing enumeration, as usual. But then it is easy to recursively construct increasing sequences 
 $\langle s_n : n \in \omega \rangle $
,
$\langle s_n : n \in \omega \rangle $
, 
 $\langle t_n : n \in \omega \rangle $
 in T so that
$\langle t_n : n \in \omega \rangle $
 in T so that 
 $s_0 = t_0 = \operatorname {stem}(T)$
, for every
$s_0 = t_0 = \operatorname {stem}(T)$
, for every 
 $n \in \omega $
,
$n \in \omega $
, 
 $t_n$
 and
$t_n$
 and 
 $s_n$
 are infinite-splitting nodes in T, and
$s_n$
 are infinite-splitting nodes in T, and 
 ${s_{2n+1}(\vert s_{2n} \vert )> t_{2n+1}(\vert t_{2n+1} \vert -1)}$
,
${s_{2n+1}(\vert s_{2n} \vert )> t_{2n+1}(\vert t_{2n+1} \vert -1)}$
, 
 $t_{2n+2}(\vert t_{2n} \vert )> s_{2n+1}(\vert s_{2n+1} \vert -1)$
. Then, letting
$t_{2n+2}(\vert t_{2n} \vert )> s_{2n+1}(\vert s_{2n+1} \vert -1)$
. Then, letting 
 $x = \bigcup _{n \in \omega } s_n$
 and
$x = \bigcup _{n \in \omega } s_n$
 and 
 $y = \bigcup _{n \in \omega } t_n$
,
$y = \bigcup _{n \in \omega } t_n$
, 
 $x \cap y \subseteq \vert s_0 \vert $
, viewing
$x \cap y \subseteq \vert s_0 \vert $
, viewing 
 $x,y$
 as elements of
$x,y$
 as elements of 
 $[\omega ]^\omega $
. This contradicts that
$[\omega ]^\omega $
. This contradicts that 
 $x,y \in X$
.⊣
$x,y \in X$
.⊣
 The proof of Theorem 1.4 is a modification of Shelah’s proof that 
 $\mathfrak {d} \leq \mathfrak {i}$
.
$\mathfrak {d} \leq \mathfrak {i}$
.
Proof of Theorem 1.4.
 Let 
 $\langle C_\alpha : \alpha < \kappa \rangle $
 be compact independent families so that
$\langle C_\alpha : \alpha < \kappa \rangle $
 be compact independent families so that 
 $ \mathcal {I} = \bigcup _{\alpha < \kappa } C_\alpha $
 is maximal independent and
$ \mathcal {I} = \bigcup _{\alpha < \kappa } C_\alpha $
 is maximal independent and 
 $\kappa < \mathfrak d$
 and assume without loss of generality that
$\kappa < \mathfrak d$
 and assume without loss of generality that 
 $\{C_\alpha : \alpha < \kappa \}$
 is closed under finite unions. Here, we will identify elements of
$\{C_\alpha : \alpha < \kappa \}$
 is closed under finite unions. Here, we will identify elements of 
 $[\omega ]^\omega $
 with their characteristic function in
$[\omega ]^\omega $
 with their characteristic function in 
 $2^\omega $
 at several places and it should always be clear from context which representation we consider at the moment.
$2^\omega $
 at several places and it should always be clear from context which representation we consider at the moment.
Claim 5.7. There are 
 $\langle x_n : n \in \omega \rangle $
 pairwise distinct in
$\langle x_n : n \in \omega \rangle $
 pairwise distinct in 
 $\mathcal {I}$
 so that
$\mathcal {I}$
 so that 
 $\{x_n : n \in \omega \} \cap C_\alpha $
 is finite for every
$\{x_n : n \in \omega \} \cap C_\alpha $
 is finite for every 
 $\alpha < \kappa $
.
$\alpha < \kappa $
.
Proof. The closure of 
 $\mathcal {I}$
 is not independent. Thus there is
$\mathcal {I}$
 is not independent. Thus there is 
 $x \in \bar {\mathcal {I}} \setminus \mathcal {I}$
. Now we pick
$x \in \bar {\mathcal {I}} \setminus \mathcal {I}$
. Now we pick 
 $\langle x_n : n \in \omega \rangle \subseteq \mathcal {I}$
 converging to x. Since
$\langle x_n : n \in \omega \rangle \subseteq \mathcal {I}$
 converging to x. Since 
 $C_\alpha $
 is closed, whenever for infinitely many n,
$C_\alpha $
 is closed, whenever for infinitely many n, 
 $x_n \in C_\alpha $
, then also
$x_n \in C_\alpha $
, then also 
 $x \in C_\alpha $
 which is impossible.⊣
$x \in C_\alpha $
 which is impossible.⊣
 Fix a sequence 
 $\langle x_n : n \in \omega \rangle $
 as above. And let
$\langle x_n : n \in \omega \rangle $
 as above. And let 
 $a_\alpha = \{ n \in \omega : x_n \in C_\alpha \} \in [\omega ]^{<\omega }$
. We will say that x is a Boolean combination of a set
$a_\alpha = \{ n \in \omega : x_n \in C_\alpha \} \in [\omega ]^{<\omega }$
. We will say that x is a Boolean combination of a set 
 $X \subseteq [\omega ]^\omega $
, if there are finite disjoint
$X \subseteq [\omega ]^\omega $
, if there are finite disjoint 
 $Y,Z \subseteq X$
 so that
$Y,Z \subseteq X$
 so that 
 $x = (\bigcap _{y \in Y} y) \cap (\bigcap _{z \in Z} \omega \setminus z)$
.
$x = (\bigcap _{y \in Y} y) \cap (\bigcap _{z \in Z} \omega \setminus z)$
.
Claim 5.8. For any 
 $\alpha < \kappa $
 there is
$\alpha < \kappa $
 there is 
 $f_{\alpha } \colon \omega \to \omega $
 so that for any
$f_{\alpha } \colon \omega \to \omega $
 so that for any 
 $K \in [C_\alpha \setminus \{x_n : n \in a_\alpha \} ]^{<\omega }$
, for all but finitely many
$K \in [C_\alpha \setminus \{x_n : n \in a_\alpha \} ]^{<\omega }$
, for all but finitely many 
 $k \in \omega $
, and for any Boolean combination x of
$k \in \omega $
, and for any Boolean combination x of 
 $K \cup \{x_0, \dots , x_k \}$
,
$K \cup \{x_0, \dots , x_k \}$
, 
 $x \cap [k,f_{\alpha }(k)) \neq \emptyset $
.
$x \cap [k,f_{\alpha }(k)) \neq \emptyset $
.
Proof. We define 
 $f_{\alpha }(k)$
 as follows. For every
$f_{\alpha }(k)$
 as follows. For every 
 $l \leq k$
, we define a collection of basic open subsets of
$l \leq k$
, we define a collection of basic open subsets of 
 $(2^\omega )^{l}$
,
$(2^\omega )^{l}$
, 
 $ \mathcal {O}_{0,l} := \{[\bar s] : \bar s \in (2^{<\omega })^l \wedge \forall i < l (\vert s_i\vert> k) \wedge (\exists i < l, n \in a_\alpha (s_i \subseteq x_n) \vee \exists i<j<l (s_i \not \perp s_j))\}$
. Further we call any
$ \mathcal {O}_{0,l} := \{[\bar s] : \bar s \in (2^{<\omega })^l \wedge \forall i < l (\vert s_i\vert> k) \wedge (\exists i < l, n \in a_\alpha (s_i \subseteq x_n) \vee \exists i<j<l (s_i \not \perp s_j))\}$
. Further we call any 
 $[\bar s] \notin \mathcal {O}_{0,l}$
 good if for any
$[\bar s] \notin \mathcal {O}_{0,l}$
 good if for any 
 $F,G \subseteq l$
 with
$F,G \subseteq l$
 with 
 $F \cap G = \emptyset $
 and for any Boolean combination x of
$F \cap G = \emptyset $
 and for any Boolean combination x of 
 $\{ x_0,\dots x_k \}$
, there is
$\{ x_0,\dots x_k \}$
, there is 
 $ k'> k$
 so that for every
$ k'> k$
 so that for every 
 $i \in F$
,
$i \in F$
, 
 $s_i(k') = 1$
, for every
$s_i(k') = 1$
, for every 
 $ i \in G$
,
$ i \in G$
, 
 $s_i(k')=0$
, and
$s_i(k')=0$
, and 
 $x(k')=1$
. Let
$x(k')=1$
. Let 
 $\mathcal {O}_{1,l}$
 be the collection of all good
$\mathcal {O}_{1,l}$
 be the collection of all good 
 $[\bar s]$
. We see that
$[\bar s]$
. We see that 
 $\bigcup _{l \leq k}(\mathcal {O}_{0,l} \cup \mathcal {O}_{1,l})$
 is an open cover of
$\bigcup _{l \leq k}(\mathcal {O}_{0,l} \cup \mathcal {O}_{1,l})$
 is an open cover of 
 $C_\alpha \cup (C_\alpha )^2 \cup \dots \cup (C_\alpha )^k$
. Thus it has a finite subcover
$C_\alpha \cup (C_\alpha )^2 \cup \dots \cup (C_\alpha )^k$
. Thus it has a finite subcover 
 $\mathcal {O}'$
. Now let
$\mathcal {O}'$
. Now let 
 $f_{\alpha }(k) := \max \{ \vert t \vert : \exists [\bar s] \in \mathcal {O}' \exists i < k (t = s_i) \}$
.
$f_{\alpha }(k) := \max \{ \vert t \vert : \exists [\bar s] \in \mathcal {O}' \exists i < k (t = s_i) \}$
.
 Now we want to show that 
 $f_{\alpha }$
 is as required. Let
$f_{\alpha }$
 is as required. Let 
 $(y_0,\dots ,y_{l-1}) \in (C_\alpha \setminus \{ x_n : n \in a_\alpha \} )^{l}$
 be arbitrary,
$(y_0,\dots ,y_{l-1}) \in (C_\alpha \setminus \{ x_n : n \in a_\alpha \} )^{l}$
 be arbitrary, 
 $y_0, \dots , y_{l-1}$
 pairwise distinct, and
$y_0, \dots , y_{l-1}$
 pairwise distinct, and 
 $k \geq l$
 so that
$k \geq l$
 so that 
 $y_i \restriction k \neq x_n \restriction k$
 for all
$y_i \restriction k \neq x_n \restriction k$
 for all 
 $i < l$
,
$i < l$
, 
 $n \in a_\alpha $
 and
$n \in a_\alpha $
 and 
 $y_i \restriction k \neq y_j \restriction k$
 for all
$y_i \restriction k \neq y_j \restriction k$
 for all 
 $i < j < l$
. In the definition of
$i < j < l$
. In the definition of 
 $f_{\alpha }(k)$
, we have the finite cover
$f_{\alpha }(k)$
, we have the finite cover 
 $\mathcal {O}'$
 of
$\mathcal {O}'$
 of 
 $(C_\alpha )^l$
 and thus
$(C_\alpha )^l$
 and thus 
 $(y_0,\dots ,y_{l-1}) \in [\bar s]$
 for some
$(y_0,\dots ,y_{l-1}) \in [\bar s]$
 for some 
 $[\bar s] \in \mathcal {O}'$
. We see that
$[\bar s] \in \mathcal {O}'$
. We see that 
 $[\bar s] \in \mathcal {O}_{0,l}$
 is impossible as we chose k large enough so that for no
$[\bar s] \in \mathcal {O}_{0,l}$
 is impossible as we chose k large enough so that for no 
 $i < l, n \in a_\alpha $
,
$i < l, n \in a_\alpha $
, 
 $s_i \subseteq x_n$
, and for every
$s_i \subseteq x_n$
, and for every 
 $i < j<l$
,
$i < j<l$
, 
 $s_i \perp s_j$
. Thus
$s_i \perp s_j$
. Thus 
 $[\bar s] \in \mathcal {O}_{1,l}$
. But then, by the definition of
$[\bar s] \in \mathcal {O}_{1,l}$
. But then, by the definition of 
 $\mathcal {O}_{1,l}$
,
$\mathcal {O}_{1,l}$
, 
 $f_{\alpha }(k)$
 is as required.⊣
$f_{\alpha }(k)$
 is as required.⊣
 As 
 $\kappa < \mathfrak d$
 we find
$\kappa < \mathfrak d$
 we find 
 $f \in \omega ^\omega $
 so that f is unbounded over
$f \in \omega ^\omega $
 so that f is unbounded over 
 $\{f_{\alpha } : \alpha < \kappa \}$
. Let
$\{f_{\alpha } : \alpha < \kappa \}$
. Let 
 $x_n^0 := x_n$
 and
$x_n^0 := x_n$
 and 
 $x_n^1 := \omega \setminus x_n$
 for every
$x_n^1 := \omega \setminus x_n$
 for every 
 $n \in \omega $
. For any
$n \in \omega $
. For any 
 $g \in 2^\omega $
 and
$g \in 2^\omega $
 and 
 $n \in \omega $
 we define
$n \in \omega $
 we define 
 $y_{n,g} := \bigcap _{m\leq n} x_m^{g(m)}$
. Further define
$y_{n,g} := \bigcap _{m\leq n} x_m^{g(m)}$
. Further define 
 $y_g = \bigcup _{n \in \omega } y_{n,g} \cap f(n)$
. Note
$y_g = \bigcup _{n \in \omega } y_{n,g} \cap f(n)$
. Note 
 $y_{n,g} \subseteq y_{m,g}$
 for
$y_{n,g} \subseteq y_{m,g}$
 for 
 $m\leq n$
 and that
$m\leq n$
 and that 
 $y_g \subseteq ^* y_{n,g}$
 for all
$y_g \subseteq ^* y_{n,g}$
 for all 
 $n \in \omega $
.
$n \in \omega $
.
Claim 5.9. For any 
 $g \in 2^\omega $
,
$g \in 2^\omega $
, 
 $y_g$
 has infinite intersection with any Boolean combination of
$y_g$
 has infinite intersection with any Boolean combination of 
 $\bigcup _{\alpha < \kappa } C_\alpha \setminus \{x_n : n \in \omega \}$
.
$\bigcup _{\alpha < \kappa } C_\alpha \setminus \{x_n : n \in \omega \}$
.
Proof. Let 
 $\{y_0,\dots ,y_{l-1}\} \in [C_\alpha \setminus \{x_n : n \in a_\alpha \}]^l$
 for some
$\{y_0,\dots ,y_{l-1}\} \in [C_\alpha \setminus \{x_n : n \in a_\alpha \}]^l$
 for some 
 $l \in \omega $
,
$l \in \omega $
, 
 $\alpha < \kappa $
 be arbitrary. Here, recall that
$\alpha < \kappa $
 be arbitrary. Here, recall that 
 $\{C_\alpha : \alpha < \kappa \}$
 is closed under finite unions. We have that there is some
$\{C_\alpha : \alpha < \kappa \}$
 is closed under finite unions. We have that there is some 
 $k_0 \in \omega $
 so that for every
$k_0 \in \omega $
 so that for every 
 $k \geq k_0$
, any Boolean combination y of
$k \geq k_0$
, any Boolean combination y of 
 $\{y_0,\dots ,y_{l-1}\}$
 and x of
$\{y_0,\dots ,y_{l-1}\}$
 and x of 
 $\{x_n : n \leq k \}$
,
$\{x_n : n \leq k \}$
, 
 $x\cap y \cap [k,f_{\alpha }(k)) \neq \emptyset $
. Let y be an arbitrary Boolean combination of
$x\cap y \cap [k,f_{\alpha }(k)) \neq \emptyset $
. Let y be an arbitrary Boolean combination of 
 $\{y_0,\dots ,y_{l-1}\}$
 and
$\{y_0,\dots ,y_{l-1}\}$
 and 
 $m \in \omega $
. Then there is
$m \in \omega $
. Then there is 
 $k> m,k_0$
 so that
$k> m,k_0$
 so that 
 $f(k)> f_{\alpha }(k)$
. But then we have that
$f(k)> f_{\alpha }(k)$
. But then we have that 
 $y_{k,g}$
 is a Boolean combination of
$y_{k,g}$
 is a Boolean combination of 
 $\{x_0, \dots , x_{k} \}$
 and thus
$\{x_0, \dots , x_{k} \}$
 and thus 
 $y_{k,g} \cap y \cap [k,f(k)) \neq \emptyset $
. In particular, this shows that
$y_{k,g} \cap y \cap [k,f(k)) \neq \emptyset $
. In particular, this shows that 
 $y \cap y_g \not \subseteq m$
 and unfixing m,
$y \cap y_g \not \subseteq m$
 and unfixing m, 
 $\vert y \cap y_g \vert = \omega $
.⊣
$\vert y \cap y_g \vert = \omega $
.⊣
 Now let 
 $Q_0,Q_1$
 be disjoint countable dense subsets of
$Q_0,Q_1$
 be disjoint countable dense subsets of 
 $2^{\omega }$
. We see that
$2^{\omega }$
. We see that 
 $\vert y_g \cap y_h\vert < \omega $
 for
$\vert y_g \cap y_h\vert < \omega $
 for 
 $h \neq g \in 2^\omega $
. Thus the family
$h \neq g \in 2^\omega $
. Thus the family 
 $\{ y_g : g \in Q_0 \cup Q_1 \}$
 is countable almost disjoint and we can find
$\{ y_g : g \in Q_0 \cup Q_1 \}$
 is countable almost disjoint and we can find 
 $y^{\prime }_g =^* y_g$
, for every
$y^{\prime }_g =^* y_g$
, for every 
 $g \in Q_0 \cup Q_1$
, so that
$g \in Q_0 \cup Q_1$
, so that 
 $\{y^{\prime }_g : g \in Q_0 \cup Q_1 \}$
 is pairwise disjoint. Let
$\{y^{\prime }_g : g \in Q_0 \cup Q_1 \}$
 is pairwise disjoint. Let 
 $y = \bigcup _{g \in Q_0} y^{\prime }_g$
. We claim that any Boolean combination x of sets in
$y = \bigcup _{g \in Q_0} y^{\prime }_g$
. We claim that any Boolean combination x of sets in 
 $\mathcal {I}$
 has infinite intersection with y and
$\mathcal {I}$
 has infinite intersection with y and 
 $\omega \setminus y$
. To see this, assume without loss of generality that x is of the form
$\omega \setminus y$
. To see this, assume without loss of generality that x is of the form 
 $\tilde x \cap x_0^{g(0)} \cap \dots \cap x_k^{g(k)}$
, where
$\tilde x \cap x_0^{g(0)} \cap \dots \cap x_k^{g(k)}$
, where 
 $\tilde x$
 is a Boolean combination of sets in
$\tilde x$
 is a Boolean combination of sets in 
 $\mathcal {I} \setminus \{ x_n : n \in \omega \}$
 and
$\mathcal {I} \setminus \{ x_n : n \in \omega \}$
 and 
 $g \in 2^\omega $
. As
$g \in 2^\omega $
. As 
 $Q_0$
 is dense there is some
$Q_0$
 is dense there is some 
 $h \in Q_0$
 such that
$h \in Q_0$
 such that 
 $h \restriction (k+1) = g \restriction (k+1)$
. Thus we have that
$h \restriction (k+1) = g \restriction (k+1)$
. Thus we have that 
 $y^{\prime }_h \subseteq ^* x_0^{g(0)} \cap \dots \cap x_k^{g(k)}$
 but also
$y^{\prime }_h \subseteq ^* x_0^{g(0)} \cap \dots \cap x_k^{g(k)}$
 but also 
 $y^{\prime }_h \cap \tilde x$
 is infinite by the claim above. In particular we have that
$y^{\prime }_h \cap \tilde x$
 is infinite by the claim above. In particular we have that 
 $y \cap x$
 is infinite. The complement of y is handled by replacing
$y \cap x$
 is infinite. The complement of y is handled by replacing 
 $Q_0$
 with
$Q_0$
 with 
 $Q_1$
. We now have a contradiction to
$Q_1$
. We now have a contradiction to 
 $\mathcal {I}$
 being maximal.⊣
$\mathcal {I}$
 being maximal.⊣
6 Concluding remarks
 Our focus in this paper was on Sacks and splitting forcing but it is clear that the method presented is more general. We mostly used that our forcing has Axiom A with continuous reading of names and that it is a weighted tree forcing (Definition 4.9), both in a definable way. For instance, the more general versions of splitting forcing given by Shelah in [Reference Shelah, Judah, Just and Woodin27] fall into this class. It would be interesting to know for what other tree forcings Theorem 5.1 holds true. In [Reference Schrittesser and Törnquist26], the authors showed that after adding a single Miller real over L, every (two-dimensional) graph on a Polish space has a 
 $\mathbf \Delta ^1_2$
 maximal independent set. It is very plausible that this can be extended to the countable support iteration. One line of attack might be to use a similar method to ours, where Cohen genericity is replaced by other kinds of genericity. For instance, the following was shown by Spinas in [Reference Spinas28] (compare this with Proposition 4.16):
$\mathbf \Delta ^1_2$
 maximal independent set. It is very plausible that this can be extended to the countable support iteration. One line of attack might be to use a similar method to ours, where Cohen genericity is replaced by other kinds of genericity. For instance, the following was shown by Spinas in [Reference Spinas28] (compare this with Proposition 4.16):
Fact. Let M be a countable model, then there is a superperfect tree T so that for any 
 $x \neq y \in [T]$
,
$x \neq y \in [T]$
, 
 $(x,y)$
 is
$(x,y)$
 is 
 $\mathbb {M}^2$
 generic over M, where
$\mathbb {M}^2$
 generic over M, where 
 $\mathbb {M}$
 denotes Miller forcing.
$\mathbb {M}$
 denotes Miller forcing.
 On the other hand, it is impossible to have that any three 
 $x,y,z \in [T]$
 are mutually generic. This follows from a fact due to Velickovic and Woodin (see [Reference Velickovic and Hugh Woodin33, Theorem 1]) that there is a Borel function
$x,y,z \in [T]$
 are mutually generic. This follows from a fact due to Velickovic and Woodin (see [Reference Velickovic and Hugh Woodin33, Theorem 1]) that there is a Borel function 
 $h \colon (\omega ^\omega )^3 \to 2^\omega $
, such that for any superperfect T,
$h \colon (\omega ^\omega )^3 \to 2^\omega $
, such that for any superperfect T, 
 $h"([T]^3) = 2^\omega $
. Also,
$h"([T]^3) = 2^\omega $
. Also, 
 $\mathbb {M}^3$
 always adds a Cohen real (see, e.g., the last paragraph in [Reference Brendle4]). This means that Theorem 5.5 can’t hold for Miller forcing and
$\mathbb {M}^3$
 always adds a Cohen real (see, e.g., the last paragraph in [Reference Brendle4]). This means that Theorem 5.5 can’t hold for Miller forcing and 
 $k \geq 3$
, even for just equivalence relations. Namely, after adding a Cohen real, there can’t be any
$k \geq 3$
, even for just equivalence relations. Namely, after adding a Cohen real, there can’t be any 
 $E_0$
-transversal that is definable with parameters from the ground model. This doesn’t rule out though that the iteration might work. Let us ask the following question.
$E_0$
-transversal that is definable with parameters from the ground model. This doesn’t rule out though that the iteration might work. Let us ask the following question.
Question 1. Does Theorem 5.1 hold true for Miller forcing?
 A positive result would yield a model in which 
 $\mathfrak {i}_{B} < \mathfrak {i}_{cl}$
, as per
$\mathfrak {i}_{B} < \mathfrak {i}_{cl}$
, as per 
 $\mathfrak {d} \leq \mathfrak {i}_{cl}$
. No result of this kind has been obtained so far.
$\mathfrak {d} \leq \mathfrak {i}_{cl}$
. No result of this kind has been obtained so far.
 Another common way to iterate Sacks or splitting forcing is to use the countable support product. The argument in Remark 3.21 can be used to see that no definable 
 $E_1$
-transversals can exist in an extension by a (uncountably long) countable support product of Sacks or splitting forcing. This raises the question for which hypergraph Theorem 5.1 applies to countable support products.
$E_1$
-transversals can exist in an extension by a (uncountably long) countable support product of Sacks or splitting forcing. This raises the question for which hypergraph Theorem 5.1 applies to countable support products.
Question 2. Is there a nice characterization of hypergraphs for which Theorem 5.1 holds when using countable support products?
An interesting application of our method that appears in the authors thesis, is that P-points exist after iterating splitting forcing over a model of CH. To the best of our knowledge this is different to any other method of P-point existence in the literature.
 Other applications are related to questions about families of reals and the existence of a well-order of the continuum. For example, it is known that after adding 
 $\omega _1$
 many Sacks reals, there is no well-order of the reals in
$\omega _1$
 many Sacks reals, there is no well-order of the reals in 
 $L(\mathbb {R})$
. On the other hand, if we start with
$L(\mathbb {R})$
. On the other hand, if we start with 
 $V = L$
, a
$V = L$
, a 
 $\Delta ^1_2$
 Hamel base exists in the extension. In particular, a Hamel base will exist in
$\Delta ^1_2$
 Hamel base exists in the extension. In particular, a Hamel base will exist in 
 $L(\mathbb {R})$
. This gives a new solution to a question by Pincus and Prikry [Reference Pincus and Prikry22], which asks whether a Hamel basis can exist without a well-order of the reals. This has only been solved recently (see [Reference Beriashvili, Schindler, Wu and Yu2]). Our results solve this problem not just for Hamel bases but for a big class of families of reals. In an upcoming paper [Reference Schilhan24] we will consider further applications to questions related to the Axiom of Choice.
$L(\mathbb {R})$
. This gives a new solution to a question by Pincus and Prikry [Reference Pincus and Prikry22], which asks whether a Hamel basis can exist without a well-order of the reals. This has only been solved recently (see [Reference Beriashvili, Schindler, Wu and Yu2]). Our results solve this problem not just for Hamel bases but for a big class of families of reals. In an upcoming paper [Reference Schilhan24] we will consider further applications to questions related to the Axiom of Choice.
7 Acknowledgments
I would like to thank the Austrian Science Fund, FWF, for generous support through START Project Y1012-N35. I would also like to thank Jörg Brendle, David Chodounský, and David Schrittesser for many fruitful discussions.
 
 





 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
