Proof
(i) According to Corollaries
24 and
25, it suffices to prove:
for all stumps \(U\), \(P(U)\),
where \(P(U)\) is the statement: for all stumps \(S\), \(Q(U,S)\),
where \(Q(U,S)\) is the statement: for all stumps \(T\), \(R(U,S,T)\),
where \(R( U,S,T)\) is the statement:
for all \(A,B\subseteq \overline{U}\), if \(S\) \(U\)-secures \(A\), and \(T\) \(U\)-secures \(B\), then \(A\cap B\) is almost-full.
We use Axiom
2 from Sect.
17.1, the principle of induction on the set
\(\mathbf{Stp}\) of stumps. Our proof is a proof by
triple induction.
1. Note: \(\overline{\emptyset }=\{\langle \; \rangle \}\).
For all \(A,B\subseteq \{\langle \;\rangle \}\), if \(A,B\) are almost-full, then \(A=B=A\cap B=\{\langle \;\rangle \}\).
We thus see that \(P(\emptyset )\) is true.
2. Let a non-empty stump \(U\) be given such that, for all \(n\), \(P(U\upharpoonright \langle n\rangle )\).
We intend to prove \(P(U)\), that is: for all stumps \(S\), \(Q(U, S)\).
We use again use Axiom
2.
2.1. Note that \(Q(U, \emptyset )\) holds a trivial reason: no subset of ℕ is secured by \(\emptyset \).
2.2. Let a non-empty stump \(S\) be given such that, for all \(n\), \(Q(U,S\upharpoonright \langle n \rangle )\).
We intend to prove: \(Q(U, S)\), that is, for all stumps \(T\), \(R(U, S,T)\).
2.2.1. Note that \(R(U, S, \emptyset )\) holds for a trivial reason: no subset of ℕ is secured by \(\emptyset \).
2.2.2. Let a non-empty stump \(T\) be given such that, for all \(n\), \(R(U,S, T\upharpoonright \langle n \rangle )\).
We intend to prove: \(R(U,S,T)\).
Let \(A,B\subseteq \overline{U}\) be given such that \(S\) \(U\)-secures \(A\) and \(T\) \(U\)-secures \(B\).
We have to prove that
\(A\cap B\) is almost-full, i.e.
\((\ast )\;\forall \zeta \in [\omega ]^{\omega }\exists s\in [\omega ]^{< \omega }[\zeta \circ s\in A\cap B]\).
It is useful, however, to first prove a slightly weaker statement:
\((\ast \ast )\ \forall \zeta \in [\omega ]^{\omega }\exists s\in [ \omega ]^{<\omega }[\zeta \circ s\in A\cap B\;\vee\;(s\neq 0 \;\wedge \;s(0)=0\;\wedge \; \zeta \circ s \in A)]\).
Let \(\zeta \) in \([\omega ]^{\omega }\) be given.
We distinguish two cases.
Case (a). \(S\upharpoonright \langle \zeta (0)\rangle =\emptyset \). Note: \(\{\langle \;\rangle \}\; U\)-secures \(A\), and \(A=\overline{U}\) and \(A\cap B=B\) is almost-full. One may conclude: \((\ast )\), and, in particular, \(\exists s \in [\omega ]^{<\omega }[\zeta \circ s \in A\cap B]\).
Case (b). \(S\upharpoonright \langle \zeta (0)\rangle \neq \emptyset \). Then \(\langle \;\rangle \in S\upharpoonright \langle \zeta (0) \rangle \).
Define a statement
44
\(QED_{1}:=\exists s\in [\omega ]^{<\omega }\setminus \{0\} [s(0)=0\; \wedge \;\zeta \circ s \in A]\).
Define \(A_{\zeta }:=\{u \in \overline{U}\mid \neg \exists t[u =\zeta \circ t] \;\vee \; u\in A \;\vee \;QED_{1}\}\).
Note that, for all \(u\) in \(U\cup \overline{U}\), if \(\neg \exists t[u=\zeta \circ t]\), then \(u\) \(U\)-secures \(A\) and also \(A_{\zeta }\).
We now want to prove:
$$ S\upharpoonright \langle \zeta (0)\rangle\ U\mbox{-secures}\ A_{\zeta }. $$
First, using the fact that \(S\) \(U\)-secures \(A\), find \(m\) such that
\(\forall \eta \in [\omega ]^{\omega }[\eta (0)>m\rightarrow \exists n[ \overline{\eta }n \in S \;\wedge \; \exists s \in [n]^{< n+1}[ \overline{\eta }n \circ s\) \(U\)-secures \(A]]]\).
Define \(q:=\max \bigl (m,\zeta (0)\bigr )\).
Now let \(\rho \) in \([\omega ]^{\omega }\) be given such that \(\rho (0)>q\).
We want to prove: \(\exists n [\overline{\rho }n \in S\upharpoonright \langle \zeta (0) \rangle \;\wedge \; \exists t \in [\omega ]^{<\omega }[\overline{\rho }n \circ t\) \(U\)-secures \(A_{\zeta }]]\).
Consider \(\eta :=\langle \zeta (0) \rangle \ast \rho \).
Find \(n,s\) such that \(\overline{\eta }n \in S\) and \(s\in [n]^{< n+1}\) and \(\overline{\eta }n \circ s\) \(U\)-secures \(A\).
If \(s=\langle \;\rangle \), then \(\langle \;\rangle =\rho \circ \langle \;\rangle \) \(U\)-secures \(A\) and also \(A_{\zeta }\).
We thus may assume \(s\neq \langle \;\rangle \). Note: \(\overline{\rho }(n-1)\in S\upharpoonright \langle \zeta (0)\rangle \).
There are two cases.
Case (ba). \(s(0)>0\). Find \(u\) such that \(length(u)=length(s)\) and
\(\forall i< length(u)[u(i)=s(i)-1]\).
Note: \(\overline{\rho }(n-1)\circ u=\overline{\eta }n \circ s\) \(U\)-secures \(A\) and also \(A_{\zeta }\).
Case (bb). \(s(0)=0\). Find \(u\) such that \(s =\langle 0\rangle \ast u\). Note: \(s(0)=\eta (0)=\zeta (0)\).
If \(\neg \exists t[\overline{\rho }n \circ u=\zeta \circ t]\), then also \(\neg \exists t[\overline{\rho }(n-1) \circ u=\zeta \circ t]\) and \(\overline{\rho }(n-1)\circ u\) \(U\)-secures \(A\) and also \(A_{\zeta }\).
If \(\exists t[\overline{\rho }(n-1)\circ u =\zeta \circ t]\), find \(\eta \) in \([\omega ]^{\omega }\) such that \(\zeta \circ s\sqsubset \eta \) and \(\forall i\exists j[\eta (i)=\zeta (j)]\), so \(\eta \) is a subsequence of \(\zeta \). Find \(n\) such that \(\overline{\eta }n \in \overline{U}\). As \(\zeta \circ s\) \(U\)-secures \(A\), we may conclude: \(\overline{\eta }n \in A\). Taking \(t:=\overline{\eta }n\), we see: \(\exists t[t\neq 0\;\wedge \;t(0)=0 \;\wedge \; \zeta \circ t \in A]\), i.e. \(QED_{1}\) and we may conclude: \(\overline{\rho }(n-1)\circ \langle \;\rangle =\langle \;\rangle \;U\)-secures \(A_{\zeta }\).
We thus see: for all \(\rho \) in \([\omega ]^{\omega }\), if \(\rho (0)>q\), then
\(\exists n [\overline{\rho }n \in S\upharpoonright \langle \zeta (0) \rangle \;\wedge \; \exists t \in [\omega ]^{<\omega }[\overline{\rho }n \circ t\) \(U\)-secures
\(A_{\zeta }]]\), that is:
$$ S\upharpoonright \langle \zeta (0)\rangle U\mbox{-secures}\ A_{\zeta }. $$
Also:
$$ T\ U\mbox{-secures}\ B. $$
Using the assumption
\(R(U, S\upharpoonright \langle \zeta (0)\rangle , T)\), we conclude:
\(A_{\zeta }\cap B\) is almost-full.
In particular, we may find \(s\) in \([\omega ]^{<\omega }\) such that \(\zeta \circ s \in A_{\zeta }\cap B\), and therefore, either \(\zeta \circ s \in A\cap B\) or \(QED_{1}\).
We thus see: \(\forall \zeta \in [\omega ]^{\omega }\exists s\in [\omega ]^{<\omega }[ \zeta \circ s\in A\cap B\;\vee \; (s\neq 0\;\wedge \;s(0)=0\;\wedge \; \zeta \circ s \in A)]\).
Now let \(\zeta \) in \([\omega ]^{\omega }\) be given.
Define a statement
$$ QED_{2}:=\exists s \in [\omega ]^{< \omega }[\zeta \circ s \in A\cap B]. $$
Also define
$$ C:=\{u\in \overline{U\upharpoonright \langle \zeta (0)\rangle }\mid \neg \exists t \in [\omega ]^{< \omega }][u=\zeta \circ t]\;\vee \; \langle \zeta (0)\rangle \ast u \in A\;\vee \;QED_{2}\}. $$
We want to prove that \(C\) is almost-full.
First, note that, for each \(\eta \) in \([\omega ]^{\omega }\), if \(\eta (0)=\zeta (0)\) and \(\eta \) is a subsequence of \(\zeta \), i.e.: \(\forall i\exists j[\eta (i)=\zeta (j)]\), then either \(\exists s\in [\omega ]^{<\omega }[\eta \circ s \in A\cap B]\), and, therefore, \(QED_{2}\), or \(\exists s\in [\omega ]^{<\omega }\setminus \{0\}[s(0)=0\;\wedge \; \eta \circ s \in A]\), and, therefore, \(\exists s \in [\omega ]^{<\omega }[\langle \zeta (0)\rangle \ast ( \eta \circ s) \in A]\).
We may even conclude: for every \(\eta \) in \([\omega ]^{\omega }\) that is a subsequence of \(\zeta \), either \(QED_{2}\), or \(\exists s \in [\omega ]^{<\omega }[\langle \zeta (0)\rangle \ast ( \eta \circ s) \in A]\), for, given \(\eta \) such that \(\eta (0)\neq \zeta (0)\) we may apply the previous result to the sequence \(\eta ^{+}:=\langle \zeta (0)\rangle \ast \eta \).
We thus see that, for every \(\eta \) in \([\omega ]^{\omega }\) that is a subsequence of \(\zeta \), there exists \(s\) in \([\omega ]^{<\omega }\) such that \(\eta \circ s \in C\).
Now let \(\eta \) in \([\omega ]^{\omega }\) be given. Define \(\eta _{\zeta }\) in \([\omega ]^{\omega }\) as follows.
If \(\exists i[\eta (0)=\zeta (i)]\), define \(\eta _{\zeta }(0)=\eta (0)\), and, if not, define \(\eta _{\zeta }(0) =\zeta (0)\).
For each \(n\), if \(\exists t \in [\omega ]^{n+2}[\overline{\eta }(n+2) = \zeta \circ t]\), then \(\eta _{\zeta }(n+1)=\eta (n+1)\), and, if not, then \(\eta _{\zeta }(n+1)=\zeta (i+1)\), where \(i\) satisfies \(\eta _{\zeta }(n)=\zeta (i)\).
As \(\eta _{\zeta }\) is a subsequence of \(\zeta \), find \(s\) in \([\omega ]^{<\omega }\) such that \(\eta _{\zeta }\circ s \in C\).
Either \(\eta \circ s = \eta _{\zeta }\circ s \in C\) or \(\eta \circ s\neq \eta _{\zeta }\circ s \). In the latter case, let \(n_{0}\) be the least \(n\) such that \(\neg \exists i[\eta (n)=\zeta (i)]\). Define \(\rho \) in \([\omega ]^{\omega }\) such that \(\forall n[\rho (n)=n_{0}+n]\) and find \(m\) such that \(\overline{\eta \circ \rho }(m) \in \overline{U\upharpoonright \langle \zeta (0)\rangle }\). Define \(s:=\overline{\rho }m\) and note: \(\overline{\eta \circ \rho } (m) = \eta \circ s\) and \(\neg \exists u\in [\omega ]^{<\omega }[ \eta \circ s = \zeta \circ u]\) and \(\eta \circ s \in C\).
We thus see that \(C\) is almost-full.
We obtained this result using the assumption: \(Q(U,S\upharpoonright \langle \zeta (0) \rangle )\).
Now define
$$ D:=\{u\in \overline{U\upharpoonright \langle \zeta (0)\rangle }\mid \neg \exists t \in [\omega ]^{< \omega }][u=\zeta \circ t]\;\vee \; \langle \zeta (0)\rangle \ast u \in B\;\vee \;QED_{2}\}.$$
Using the assumption: \(R(U, S, T\upharpoonright \langle \zeta (0)\rangle )\), one may prove: \(D\) is almost-full, by an argument similar to the one that established that \(C\) is almost-full.
Using the assumption \(P(U\upharpoonright \langle \zeta (0)\rangle )\), one then concludes: \(C\cap D\) is almost-full.
Find \(s\) in \([\omega ]^{<\omega }\) such that \(\zeta \circ s \in C\cap D\) or \(\langle \zeta (0)\rangle \ast \zeta \circ s \in A\cap B\), so, in any case, \(QED_{2}\).
Using the assumptions: for all \(n\), \(P(U\upharpoonright \langle n \rangle )\) (2), and: for all \(n\), \(Q(U, S\upharpoonright \langle n \rangle )\) (2.2) and: for all \(n\), \(R(U, S, T\upharpoonright \langle n \rangle )\) (2.2.2), one thus proves: \(R(U,S,T)\), i.e.: for all \(A,B\subseteq \overline{U}\), if \(S,T\) \(U\)-secure \(A,B\), respectively, then \(A\cap B\) is almost-full, i.e.: \(\forall \zeta \in {\omega }^{\omega }\exists s\in [\omega ]^{<\omega }[ \zeta \circ s \in A\cap B]\).
Axiom
2 and Corollary
25 then guarantee the conclusion:
\(\mathbf{CRT}\).
(ii) This immediately follows from (i), as, for all \(k\), \(\omega ^{< k}\) is a stump and
\(\overline{\omega ^{< k}}=\omega ^{k}\).
45 □