Cohen Generic Structures with Functions
Extracted source text · arXiv 2310.11582
\documentclass[oneside]{amsart}
\usepackage{amsmath,amsfonts,amssymb}%
\usepackage[libertine]{newtxmath}
\usepackage[all]{xy}
\usepackage[usenames]{color}
\usepackage{setspace}
\usepackage{comment}
\usepackage{url}
\usepackage{enumitem}
%%% MATH
\usepackage{amsmath, amsthm, amsfonts, amssymb, stmaryrd}%
\usepackage{mathrsfs}
%%% BIBLIOGRAPHY
%\usepackage[numbers]{natbib} %% numbers is required.
%%% LINKS
%\usepackage[colorlinks,cite]
\usepackage{enumerate}
%LaTeX 3
\usepackage{xparse, expl3}
%
%%%%%%%%% REFERENCES %%%%%%%%
%
%\usepackage{url}
%\usepackage{hyperref}
%\hypersetup{
% colorlinks = true,
% citecolor=Blue,
% urlcolor=Blue,
% linkcolor=Green
%}
%%\usepackage{hypernat}
%
\usepackage[capitalize]{cleveref}
\crefname{lemma}{Lemma}{Lemmas}
\crefname{corollary}{Corollary}{Corollaries}
\crefname{theorem}{Theorem}{Theorems}
\crefname{equation}{Equation}{Equations}
\crefname{example}{Example}{Examples}
\crefname{section}{Section}{Sections}
\crefname{subsection}{Section}{Sections}
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%% BIBLIOGRAPHY %%%%%%%%
%\usepackage[numbers]{natbib} %% numbers is required.
%\usepackage[%
% minnames=1,maxnames=99,maxcitenames=4,
%% style=alphabetic,
%% doi=false,url=false,
% firstinits=true,hyperref,natbib,backend=bibtex]{biblatex}
%\bibliography{bibliography}
%%\addbibresource{bibliography.bib}
%\renewcommand\bibname{References}
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%%%%%%%%%% TURN ON OR OFF %%%%%%%%%%%%%%%%
%\renewcommand{\TBD}[1]{} \renewcommand{\fTBD}[1]{}
%\renewcommand{\PROBLEM}[1]{} \renewcommand{\fPROBLEM}[1]{}
%\renewcommand{\NATE}[1]{} \renewcommand{\fNATE}[1]{}
%\renewcommand{\LATER}[1]{} \renewcommand{\fLATER}[1]{}
%\renewcommand{\NA}[1]{#1}
%\definecolor{LightGray}{rgb}{.6,.6,.6}
%\usepackage{vruler} \setvruler[10pt][1][1][4][1][10pt][10pt][0pt][1.025\textheight]
%\setstretch{1.0}
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%NEW COMMAND%%%%%%%%%%%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
\include{macros}
\include{commenting}
\include{theorems}
\marginparwidth=1in
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%\date{----}
\begin{document}
\setlist[itemize]{itemsep=0.75em}
\title{Cohen Generic Structures with Functions}
\thanks{The second author's research has been supported by a grant from IPM (No. 1402030417).}
\begin{abstract}
Suppose $\LangBot\subseteq \Lang$ are languages where $\Lang \setminus \LangBot$ is relational. Additionally, let $\AgeK$ be a strong \Fraisse\ class in $\Lang$. We consider the partial ordering, under substructure, of those elements in $\AgeK$ whose reduct to $\LangBot$ are substructures of a fixed $\LangBot$-structure $\MBot$. In this paper, we establish that, under general conditions, this partial order satisfies the $|\MBot|$-chain condition.
Furthermore, under these conditions, we demonstrate that any generic for such a partial order satisfies the theory of the \Fraisse\ limit of $\AgeK$, provided $\MBot$ satisfies the theory of \Fraisse\ limit of its age.
We also provide general conditions that guarantee all such generics to be rigid, as well as conditions ensuring that these generics possess large automorphism groups.
\end{abstract}
\author{Nathanael Ackerman}
\address{Harvard University,
Cambridge, MA 02138, USA}
\email{nate@aleph0.net}
\author{Mohammad Golshani}
\address{School of Mathematics, Institute for Research in Fundamental Sciences (IPM), P.O. Box:
19395-5746, Tehran, Iran.}
\email{golshani.m@gmail.com}
\author{Mostafa Mirabi}
\address{Department of Mathematics and Computer Science, The Taft School, Watertown, CT 06795, USA}
\email{mmirabi@wesleyan.edu}
\subjclass[2010]{03C55, 03E35}%,\ 54C99,\ 03B60,\ 03C75}
\keywords{\Fraisse\ limit, Cohen forcing, Generic structures}%, Quantale, Potential Isomorphism \and Relativization}
\maketitle
%%%% %%%% %%%% %%%% %%%% %%%% %%%% %%%%
\section{Introduction}
%%%% %%%% %%%% %%%% %%%% %%%% %%%% %%%%
Given a strong \Fraisse\ class $\AgeK$ in a relational language $\Lang$, the concept of a Cohen generic structure for $\AgeK$ was introduced in \cite{Cohen} and \cite{Golshani}. Such a structure is obtained as a generic when $\AgeK$ is treated as a partial ordering whose order is the substructure relation. In \cite{Cohen} and \cite{Golshani}, they showed, among other things, that when $\Lang$ is a countable relational language, $\AgeK$ has only countably many elements up to isomorphism, and all elements of $\AgeK$ are finite, the partial ordering $\AgeK$ is ccc (countable chain condition) and hence preserves cardinals under forcing.
While the preservation of cardinals when forcing with strong \Fraisse\ classes in a relational language is an interesting phenomenon, the restriction to relational languages is a significant one. In particular, there are many strong \Fraisse\ classes which can be built on top of algebraic structures, and hence for which the results of \cite{Cohen} and \cite{Golshani} do not apply.
The following are two examples to keep in mind when considering uncountable generic objects for a strong \Fraisse\ class. Both examples make fundamental use of the function symbols in the language.
\begin{example}
\label{Example: Colored vector spaces}
Suppose $\Field$ is a finite field. Let $\LangBot_{\Field} = \{+\} \cup \{f_c\}_{c \in \Field} \cup \{v_0\}$ be the one sorted language where $v_0$ is a constant, $+$ is a binary function symbol and for each $c \in \Field$, $f_c$ is a unary function symbol. Let $\AgeKBot_{\Field}$ be the age in $\LangBot_{\Field}$ whose structures are finite dimensional vector spaces over $F$ where $v_0$ is interpreted as the $0$ object, $f_c$ is interpreted as ``multiplication by $c$'' (for $c \in \Field$) and $+$ is vector addition.
Let $\Lang_{\Field, n} = \LangBot_{\Field} \cup \{R\}$ where $R$ has arity $|\Field|^n$. We then let $\AgeK$ consist of $\Lang_{\Field, n}$ structures $\cA$ such that
\begin{itemize}
\item $\cA \rest[\LangBot_{\Field}]\in \AgeKBot_{\Field}$,
\item if $\cA \models R(a_0, \dots, a_{k-1})$ where $k = |\Field|^n$, then
\begin{itemize}
\item $\{a_0, \dots, a_{k-1}\}\rest[\LangBot_{\Field}] \in \AgeKBot_{\Field}$,
\item for all $i < j < k$, $a_i \neq a_j$,
\item if $\sigma$ is a permutation of $[k]$, then $\cA \models R(a_{\sigma(0)}, \dots, a_{\sigma(k-1)})$.
\end{itemize}
\end{itemize}
Elements of $\AgeK$ then consist of vector spaces with distinguished $n$-dimensional subspaces that satisfy the relation $R$.
These can be thought of as vector space analogs of $n$-ary hypergraphs and the \Fraisse\ limit of $\AgeK$ can be thought of as the infinite dimensional vector space analog of the generic $n$-ary hypergraph.
\end{example}
\begin{example}
\label{Example: Colored edges of trees}
Let $\LangBot_{\Tree} = \{p, r\}$ where $r$ is a constant and $p$ is a unary function symbol. Let $\AgeKBot_{\Tree}$ be the age over $\LangBot_{\Tree}$ consisting of those structures $\cA$ such that
\begin{itemize}
\item $\cA \models p(r) = r$,
\item $\cA \models (\forall a)\, p^{|\cA|}(a) = r$.
\end{itemize}
Elements of $\AgeKBot_{\Tree}$ can be thought of as trees with root $r$ and predecessor relation $p$.
Let $\Lang_{\Tree} = \LangBot_{\Tree} \cup \{R\}$ where $R$ is a binary relation. Let $\AgeK_{\Tree}$ consist of those $\Lang_{\Tree}$-structures $\cA$ such that
\begin{itemize}
\item $\cA \rest[\LangBot_{\Tree}] \in \AgeKBot_{\Tree}$,
\item $\cA \models R(a, b) \rightarrow p(a) = b$.
\end{itemize}
Elements of $\AgeK_{\Tree}$ can be thought of as finite trees where for each node we either color the edge between the node and its predecessor (with the relation $R$) or we don't.
A \Fraisse\ limit of $\AgeK_{\Tree}$ is then the countably branching tree where all edges of the tree are colored generically.
When we move to case of uncountable models whose age is $\AgeKBot_{\Tree}$ though a phenomenon emerges which wasn't present in \cref{Example: Colored vector spaces}. Specifically for every uncountable cardinal $\kappa$ there are $2^\kappa$ many models of the \Fraisse\ theory of $\AgeKBot_{\Tree}$ of size $\kappa$.
When defining generic structures we will want to fix a $\LangBot_{\Tree}$-structure $\MBot_{\Tree}$ and force with those elements of $\AgeK_{\Tree}$ whose reduct to $\LangBot_{\Tree}$ is a substructure of $\MBot_{\Tree}$. As such, the generic structure we end up with will depend on the choice of $\LangBot_{\Tree}$-structure we started with.
\end{example}
One of the main results of \cite{Cohen} and \cite{Golshani} is that when one forces with a strong \Fraisse\ class cardinals are preserved. The restriction in \cite{Cohen} and \cite{Golshani} to relational languages is necessary for their proofs as they rely on the $\Delta$-system lemma for sets, and hence require that every subset of a structure gives rise to a substructure.
In this paper, we will show how to weaken the assumption that the language is relational. For specificity we now describe our framework in the case where the language is countable, all elements of the age are finitely generated, and all there are (up to isomorphism) only countably many elements in the age. Note both \cref{Example: Colored vector spaces} and \cref{Example: Colored edges of trees} satisfy these conditions.
Suppose $\LangBot \subseteq \Lang$ are languages $\Lang \setminus \LangBot$ is relational. Also suppose $\AgeK$ is a strong \Fraisse\ class in $\Lang$ where $\AgeK\rest[\LangBot]$ is a strong \Fraisse\ class in $\LangBot$. Suppose $\MBot$ is any $\LangBot$-structure which satisfies the theory of the \Fraisse\ limit of $\AgeK\rest[\LangBot]$. We consider the partial ordering whose objects are elements of $\AgeK$ whose reduct to $\LangBot$ are substructures of $\MBot$ and where the ordering is substructure. By first proving an analog of the $\Delta$-system lemma we will show that this partial ordering is ccc. In a similar approach to \cite{Cohen}, we will also consider properties of this forcing when the age consists of non-finitely generated objects.
Both \cite{Cohen} and \cite{Golshani} also show an interesting phenomenon occurs with many Cohen generic structures. For example, \cite{Cohen} shows that there are several natural classes where the \Fraisse\ limits have many automorphisms but the Cohen generic structures for the classes, when the underlying sets are uncountable, are always rigid. We will show that this phenomenon extends to the case when there are function symbols in $\LangBot$. Specifically, we will give conditions which will ensure that the Cohen generic structure is rigid. We will also give some conditions which ensure the Cohen generic structure has a large set of automorphisms.
%%%% %%%% %%%% %%%% %%%% %%%%
\subsection{Notation}
%%%% %%%% %%%% %%%% %%%% %%%%
\iffalse
We assume we work in a background model $\Set$ of $\ZFC$. When we talk about an \emph{elementary substructure} of $\Set$ we mean a $\Sigma_n$-elementary substructure for some $n$ large enough to carry out the needed argument (note $\ZFC$ proves there are $\Sigma_n$-elementary substructures of $\Set$ for all $n$ and that the downward \LS\ theorem holds for such elementary substructures).\fTBD{Is it clear enough what we are saying? Do we want to give a reference? If so maybe Aki's "Higher Infinite"} We will always use $V$ to denote a transitive model of $\ZFC$.
\fi
We let $\ORD$ be the collection of ordinals. We let $\cofinal(\kappa)$ be the \emph{cofinality} of a cardinal $\kappa$. We let $\tc(X)$ be the \emph{transitive closure} of the set $X$. If $X$ is a set we let $\Perm(X)$ be the collection of \emph{permutations} of $X$. %We assume that for all language $\Lang$, they are represented so that $|\Lang| = |\tc(\Lang)|$.
A \defn{$\Delta$-system} is a collection of sets $\{S_i\}_{i \in \kappa}$ for which there is a set $S$ such that for all $i < j\in \kappa$, $S_i \cap S_j = S$.
By a language we mean a quadruple $\Lang = \{\Sort[\Lang], \Rel[\Lang], \Func[\Lang], \ar[\Lang]\}$ where $\Sort[\Lang]$ is the collection of sorts, $\Rel[\Lang]$ is the collection of relation symbols, $\Func[\Lang]$ is the collection of function symbols and $\ar[\Lang]\:\Rel[\Lang] \cup \Func[\Lang] \to \bigcup_{n \geq 1}\Sort[\Lang]^{n}$ is the function which takes a symbol and returns its arity. If $f \in \Func[\Lang]$ and $\ar[\Lang](f) = (\<S_i\>_{i \in [k]}, S_{k})$ then we say $f$ is a function symbol from $\<S_i\>_{i \in [k]}$ to $S_k$ and write $f\:\prod_{i \in [k]} S_i \to S_k$. The cardinality of a language is the cardinality of $\Sort[\Lang] \cup \Rel[\Lang] \cup \Func[\Lang]$. When $|\Sort[\Lang]| = 1$ we will omit mention of it.
If $\cM, \cN$ are $\Lang$-structure we write $\cM \subseteq \cN$ to denote that $\cM$ is a substructure of $\cN$. We will let $\UnderSet{\cM}$ denote the underlying set of $\cM$. If $\aa$ is a tuple we write $\aa \in \cM$ if every component of $\aa$ is an element of $\cM$, i.e. $\aa \in \UnderSet{\cM}^{|\aa|}$ where $|\aa|$ is the length of the tuple $\aa$. If $\LangBot \subseteq \Lang$ we let $\cM\rest[\LangBot]$ be the $\LangBot$-structure which is the restriction of $\cM$ to $\LangBot$.
We define the \defn{term definable closure} of a set $A \subseteq \cM$ to be the smallest subset of $\cM$ containing $A$ and closed under taking terms. We denote it by $\tdcl[\cM](A)$. If $\aa$ is a tuple which enumerates $A$, we then let $\tdcl[\cM](\aa) = \tdcl[\cM](A)$. We say $\cM$ is \defn{generated} by a set or tuple $X$ if $\tdcl[\cM](X) = \cM$. We say $\cM$ is \defn{$< \lambda$-generated} if there is a tuple of size $< \lambda$ which generates $\cM$. We say $\cM$ is \defn{finitely generated} if it is $< \w$-generated.
Suppose $\cM$ is an $\Lang$-structure and $\aa \in \cM$ is a tuple. We let $\qftp[\cM](\aa)(\xx)$ be the \defn{quantifier-free type} of $\aa$ in $\cM$ with variables $\xx$, i.e. the collection of all literals in $\xx$ true about tuples of $\aa$ in $\cM$. The following folklore result is immediate but important.
\begin{lemma}
Suppose $\cM, \cN$ are $\Lang$-structures where $\aa$ generates $\cM$ and $\bb$ generates $\cN$. If $\qftp[\cM](\aa)(\xx) = \qftp[\cN](\bb)(\xx)$ then $\cM \cong \cN$.
\end{lemma}
We say a subset $T \subseteq \Liw(\Lang)$ is \defn{complete} if for every $\varphi \in \Liw(\Lang)$, either $T \vdash \varphi$ or $T \vdash \neg \varphi$ where $\vdash$ is the standard provability relation for $\Liw(\Lang)$ (see \cite{Barwise}). Two $\Lang$-structures $\cM$ and $\cN$ are \defn{potentially isomorphic} if they satisfy the same sentences of $\Liw(\Lang)$.
It can be shown that $\cM$ and $\cN$ are potentially isomorphic if and only if they are isomorphic in some forcing extension of the universe (see \cite{Marker2016-MARLOI}).
%\TBD{Do we want to explain why it is called this?}
%Throughout this paper $\Lang$ will be a (not necessarily relational) language. All relations will have positive arity but we will allow functions to have arity $0$ (instead of allowing constant symbols in our languages). If $\cA$ and $\cB$ are $\Lang$-structures we write $\cA \subseteq \cB$ to denote the fact that $\cA$ is a substructure of $\cB$.
%If $\cM$ is a structure and $\aa$ is a finite substructure of $\cM$ we define $\Closure(\aa)$ to be the substructure of $\cM$ that results from closing $\aa$ under all function symbols in the language. If $\aa$ and $\bb$ are non-repeating tuples of the same length we let $\ClosureMap(\aa)[\bb]$ be the embedding from $\Closure(\aa)$ to $\Closure(\bb)$ extending the map that takes $\aa$ to $\bb$.
We will use $V$ to denote a model of Zermelo-Fraenkel set theory with the axiom of choice. For definitions with respect to forcing we refer the reader to \cite{MR756630}. In particular, we will use the convention of \cite{MR756630} that if $(P, \leq)$ is a partial ordering and $p \leq q$ then $p$ is \emph{stronger} than $q$.
%%%% %%%% %%%% %%%% %%%% %%%% %%%% %%%%
\section{\Fraisse\ classes}
%%%% %%%% %%%% %%%% %%%% %%%% %%%% %%%%
We now review basic notions of \Fraisse\ theory. See, for example, \cite[Section 7.1]{MR1221741} for more on countable \Fraisse\ theory.
\begin{definition}
An \defn{age} in a language $\Lang$ is a collection of $\Lang$-structures closed under isomorphism. If $\AgeK$ is an age in $\Lang$ then we define the following.
\begin{itemize}
\item For a cardinal $\kappa$ we let $\AgeK<\kappa>$ be the collection of elements of $\AgeK$ which are generated by a set of size $< \kappa$.
\item For $\LangBot \subseteq \Lang$ let $\AgeK\rest[\LangBot] = \{\cA \rest[\LangBot] \st \cA \in \AgeK\}$.
\end{itemize}
We say $\AgeK$ is \defn{$\kappa$-bounded} if $\AgeK = \AgeK_\kappa$.
If $\cM$ is a $\Lang$-structure we define the \defn{$\kappa$-generated age} of $\cM$ to be the collection of $\Lang$-structures isomorphic to a substructure of $\cM$ generated by a set of size $<\kappa$.
\end{definition}
\begin{definition}
Suppose $\AgeK$ is an age.
\begin{itemize}
\item We say $\AgeK$ has the \defn{hereditary property (HP)} if whenever $\cA_1 \in \AgeK$ and $\cA_0 \subseteq \cA_1$, then $\cA_0 \in \AgeK$.
\item We say $\AgeK$ has the \defn{joint embedding property (JEP)} if whenever $\cA_0, \cA_1 \in \AgeK$ there is an $\cA_2\in \AgeK$ and embeddings $i\:\cA_0 \to \cA_2$ and $j\:\cA_1\to \cA_2$.
\item Suppose $i_1\:\cA_0 \to \cA_1$, $i_2\:\cA_0 \to \cA_2$ are embeddings with $\cA_0, \cA_1, \cA_2 \in \AgeK$. An \defn{amalgam} of $(i_1, i_2)$ is a pair of embeddings $(j_1, j_2)$ such that, for some $\cA_3 \in \AgeK$, $j_1\:\cA_1\to \cA_3$, $j_2\:\cA_2 \to \cA_3$ and $j_1 \circ i_1 = j_2 \circ i_2$. If further $j_1``[\cA_1] \cap j_2``[\cA_2] = (j_1 \circ i_1)``[\cA_0]$, then we say $(j_1, j_2)$ is a \defn{disjoint amalgam} of $(i_1, i_2)$.
\item We say $\AgeK$ has the \defn{amalgamation property (AP)} if whenever $\cA_0, \cA_1, \cA_2 \in \AgeK$ and $i_1\:\cA_0 \to \cA_1$, $i_2\:\cA_0 \to \cA_2$ are embeddings, then $(i_1, i_2)$ has an amalgam.
\item We say $\AgeK$ has the \defn{strong amalgamation property (SAP)} if each such $(i_1, i_2)$ has a disjoint amalgam.
\item We say $\AgeK$ is \defn{closed under less than $\kappa$-unions (C$\kappa$U)} if for all $\gamma < \kappa$ and $(\cA_i)_{i \in \gamma} \subseteq \AgeK$ with $\cA_i \subseteq \cA_j$ for $i < j \in \gamma$ we have $\bigcup_{i \in \gamma} \cA_i \in \AgeK$.
\end{itemize}
A \defn{\Fraisse\ class} is an age which satisfies (HP), (JEP) and (AP). A \defn{strong \Fraisse\ class} is a \Fraisse\ class which satisfies (SAP).
\end{definition}
%%%%%%%%HERE%%%%%%
\begin{definition}
\label{Definition Fraisse limit}
Suppose $\AgeK$ is a \Fraisse\ class in $\Lang$. A \defn{\Fraisse\ limit} of $\AgeK$ is an $\Lang$ structure $\cM$ such that
\begin{itemize}
\item[(a)] For every $\cB \in \AgeK$ there is an embedding from $\cB$ into $\cM$,
\item[(b)] If $\cB \in \AgeK$ and $i, j\:\cB \to \cM$ are embeddings, then there is a $g \in \Aut(\cM)$ such that $i = g \circ j$.
\end{itemize}
\end{definition}
The following definition will be useful.
\begin{definition}
Suppose $\cM$ is an $\Lang$-structure and $A \subseteq \cM$. We define the \defn{group theoretic definable closure} of $A$ to be
\[
\gdcl[\cM](A) = \{b \in \cM \st (\forall g \in \Aut(\cM))\, g\rest[A] = \id \rightarrow g(b) = b\}.
\]
We say $\cM$ has \defn{trivial group theoretic definable closure} if for all finitely generated substructures $\cA\subseteq \cM$,\, $\gdcl[\cM](\UnderSet{\cA}) = \UnderSet{\cA}$.
\end{definition}
Note the following is a well known result (see \cite[Ch. 7.1]{MR1221741})
\begin{lemma}
\label{Fraisse limits exist and trivial group theoretic dcl}
Suppose $\Lang$ is a countable language, $\AgeK$ is an $\w$-bounded \Fraisse\ class in $\Lang$ and $\AgeK$ has only countably many elements up to isomorphism. Then there is a countable \Fraisse\ limit $\cM$ of $\AgeK$ which is unique up to isomorphism.
Furthermore, if $\Lang$ is relational, then $\AgeK$ is a strong \Fraisse\ class if and only if $\cM$ has trivial group theoretic definable closure.
\end{lemma}
\begin{definition}
\label{Definition Fraisse Theory}
Suppose $\AgeK$ is a \Fraisse\ class which is $\kappa$-bounded. We let $\ThFraisse_\kappa(\AgeK) \subseteq \Lik(\Lang)$ be the theory with the following axioms.
\begin{itemize}
\item[(a)] for all $\cA \in \AgeK$ and $\aa \in \cA$ with $\aa$ generating $\cA$
\[
(\exists \xx)\, \bigwedge\qftp[\cA](\aa)(\xx).
\]
\item[(b)] for all $\cA, \cB \in \AgeK$ with $\cA\subseteq \cB$, $\aa$ generating $\cM$ and $\bb$ generating $\cB$
\[
(\forall \xx)\, \bigg[ \bigwedge\qftp[\cA](\aa)(\xx) \rightarrow (\exists \yy)\, \bigwedge\qftp[\cB](\aa\bb)(\xx\yy)\bigg].
\]
%\fTBD{Shouldn't $\qftp[\cN](\aa\bb)(\xx\yy)$ be}
%\fTBD{$\qftp[\cN](\bb)(\yy)$}
\end{itemize}
We call $\ThFraisse_\kappa(\AgeK)$ the \defn{$\kappa$-\Fraisse\ theory} of $\AgeK$.
\end{definition}
Note that if $i\:\cA_0 \cong \cA_1$ and $\aa \in \cA_0$, then $\qftp[\cA_0](\aa)(\xx) = \qftp[\cA_1](i(\aa))(\xx)$. Therefore, as every $\kappa$-bounded \Fraisse\ class has only a set size collection of objects up to isomorphism, $\bigwedge \ThFraisse_\kappa(\AgeK_\kappa) \in \Lik(\Lang)$ for all $\kappa$.
The following is folklore. %\fTBD{Do we want to say more?}
\begin{lemma}
\label{Fraisse limits satisfy Fraisse theories}
If $\AgeK$ is $\kappa$-bounded and $\cM$ is a \Fraisse\ limit of $\AgeK$, then $\cM \models \ThFraisse_\kappa(\AgeK)$.
\end{lemma}
\begin{proof}
That $\cM$ satisfies the axioms in \cref{Definition Fraisse Theory} (a) follows from \cref{Definition Fraisse limit} (a). Now suppose $\cA, \cB, \aa, \bb$ are as in \cref{Definition Fraisse Theory} (b) and suppose $\cM \models \bigwedge\qftp[\cA](\aa)(\cc)$ for some tuple $\cc \in \cM$. There is then an embedding $i\:\cB \to \cM$ by \cref{Definition Fraisse limit} (a). But there is also an embedding $j\:\cc \to i(\aa)$ as $\aa$ has the same quantifier-free type in $\cA$ as $\cc$ does in $\cM$. Therefore, by \cref{Definition Fraisse limit} (b) there is an automorphism $g$ of $\cM$ taking $i(\aa)$ to $\cc$. Let $\dd = g(i(\bb))$. But then $\aa\bb$ has the same quantifier-free type in $\cB$ as $i(\aa\bb)$ does in $\cM$, which in turn has the same quantifier-free type in $\cM$ as $g(i(\aa\bb)) = \cc \dd$. Therefore $\cM$ satisfies the axioms in \cref{Definition Fraisse Theory} (b).
\end{proof}
The following is folklore.
\begin{proposition}
Suppose $\AgeK$ is $\w$-bounded. Then $\ThFraisse_\w(\AgeK)$ is complete for $\Liw(\Lang)$.
\end{proposition}
\begin{proof}
As $\ThFraisse_\w(\AgeK)$ is a set, it is complete for $\Liw(\Lang)$ if and only if $\bigwedge \ThFraisse_\w(\AgeK)$ is a Scott sentence in any model of set theory where it is countable. We can therefore assume without loss of generality that $\ThFraisse_\w(\AgeK)$ is countable.
In particular, we have $\AgeK$ has only countably many elements up to isomorphism. Therefore by \cref{Fraisse limits exist and trivial group theoretic dcl} $\AgeK$ has a countable \Fraisse\ limit $\cM$. But, by \cref{Fraisse limits satisfy Fraisse theories} we have $\cM \models \ThFraisse_\w(\AgeK)$.
Finally, a straightforward back and forth argument shows that any two countable models of $\ThFraisse_\w(\AgeK)$ are isomorphic. Therefore $\bigwedge \ThFraisse_\w(\AgeK)$ is a Scott sentence of $\cM$ and hence complete.
\end{proof}
In our context, we want to consider a pair of languages $\LangBot \subseteq \Lang$ where $\Lang\setminus \LangBot$ is relational. We then consider a strong \Fraisse\ class $\AgeK$ in $\Lang$ such that $\AgeK\rest[\LangBot]$ is a strong \Fraisse\ class in $\LangBot$. We then fix some large $\LangBot$-structure $\MBot$ and we will want to force with elements of $\AgeK$ whose restriction to $\LangBot$ is a substructure of $\MBot$. The following notation will be convenient.
\begin{definition}
Suppose $\LangBot \subseteq \Lang$ are languages, $\MBot$ is a $\LangBot$-structure and $\AgeK$ is an age over $\Lang$. We let
\[
\AgeK[\MBot] = \{\cA \in \AgeK \st \cA\rest[\LangBot] \subseteq \MBot\}.
\]
\end{definition}
There is though one issue with the above framework. Even though we have strong amalgamation in both $\AgeK$ and $\AgeK\rest[\LangBot]$ it is possible that given $\Lang$-structures $\cA \subseteq \cB$ and $\cA \subseteq \cC$ in $\AgeK[\MBot]$ the amalgamation of $\cB\rest[\LangBot]$ and $\cC\rest[\LangBot]$ over $\cA\rest[\LangBot]$, which is imposed by $\MBot$, is not compatible with any amalgamation of $\cB$ and $\cC$ over $\cA$. We therefore will need a stronger version of amalgamation which ensures every disjoint amalgam in $\AgeK\rest[\LangBot]$ can be expanded to a disjoint amalgam in $\AgeK$. We now make this precise.
\begin{definition}
Suppose
\begin{itemize}
\item $\LangBot \subseteq \Lang$,
\item $\AgeK$ is a strong \Fraisse\ class in $\Lang$ with $\AgeK\rest[\LangBot]$ a strong \Fraisse\ class in $\LangBot$.
\end{itemize}
We say $\AgeK$ has the \defn{extended strong amalgamation property over $\LangBot$} if, whenever
\begin{itemize}
\item $\cA, \cB, \cC \in \AgeK$ with $\cA \subseteq \cB$ and $\cA \subseteq \cC$ and $\UnderSet{\cB} \cap \UnderSet{\cC} = \UnderSet{\cA}$
\item $\DBot \in \AgeK\rest[\LangBot]$ with $\cB\rest[\LangBot] \subseteq \DBot$, $\cC\rest[\LangBot] \subseteq \DBot$ and where $\DBot$ is generated by $\UnderSet{\cB} \cup \UnderSet{\cC}$,
\end{itemize}
Then there is a $\cD \in \AgeK$ such that $\cB \subseteq \cD$, $\cC \subseteq \cD$ and $\cD \rest[\LangBot] = \DBot$.
\end{definition}
%%%% %%%% %%%% %%%% %%%% %%%% %%%% %%%%
\section{Forcings}
%%%% %%%% %%%% %%%% %%%% %%%% %%%% %%%%
We now introduce the forcing notions we are interested in.
\begin{definition}
Suppose
\begin{itemize}
\item $\LangBot \subseteq \Lang$,
\item $\Lang \setminus \LangBot$ is relational,
\item both $\AgeK$ and $\AgeK\rest[\LangBot]$ are strong \Fraisse\ classes,
\item $\MBot \models \ThFraisse_\kappa(\AgeK\rest[\LangBot])$,
\end{itemize}
We then let $\Fn<\MBot>[\AgeK](\kappa)$ be the partial ordering where
\begin{itemize}
\item the elements are structures in $\AgeK<\kappa>[\MBot]$
\item for $\cM, \cN \in \Fn<\MBot>[\AgeK](\kappa)$, $\cM \leq \cN$ if $\cN \subseteq \cM$.
\end{itemize}
\end{definition}
Note for $\cM, \cN \in \Fn<\MBot>[\AgeK](\kappa)$ we have $\cM$ is stronger than $\cN$ precisely when $\cN$ is a substructure of $\cM$, i.e. $\cN \subseteq \cM$.
%Note we have the following easy result about $\Fn<\MBot>[\AgeK](\kappa)$.
\begin{proposition}
\label{propclosure}
Suppose $\lambda \leq \cofinal(\kappa)$ and $\AgeK$ is a strong \Fraisse\ class with (C$\lambda$U). Then $\Fn<\MBot>[\AgeK](\kappa)$ is $\lambda$-closed.
\end{proposition}
\begin{proof}
Suppose $\<\cA_i\>_{i \in \gamma}$ is a decreasing sequence in $\Fn<\MBot>[\AgeK](\kappa)$ with $\gamma < \lambda$, i.e. for all $i \in \gamma$, $\cA_i$ is $<\kappa$-generated and for $i < j < \gamma$, $\cA_i \subseteq \cA_j$. Then, because $\AgeK$ has (C$\lambda$U), we know that $\bigcup_{i \in \gamma} \cA_i \in \AgeK$. Therefore, as $\gamma < \lambda \leq \cofinal(\kappa)$ we $\bigcup_{i \in \gamma} \cA_i \in \AgeK$ is $<\kappa$-generated. But then for all $j < \gamma$, $\cA_j \subseteq \bigcup_{i \in \gamma} \cA_i$. Therefore, as $\<\cA_i\>_{i \in \gamma}$ was arbitrary, $\Fn<\MBot>[\AgeK](\kappa)$ is $\lambda$-closed.
\end{proof}
Our main result will give conditions when $\Fn<\MBot>[\AgeK](\kappa)$ isn't just $\kappa$-closed but has the $\kappa^+$-chain condition. First though we will need to prove an analog of the $\Delta$-systems lemma which will play a crucial role in our result. For a proof of the $\Delta$-system lemma see, for example, \cite[Theorem II.1.5]{MR756630}.
\begin{lemma}[Term $\Delta$-Systems Lemma]
\label{Delta-Systems Lemma}
Suppose
\begin{itemize}
\item $\kappa^{<\lambda} = \kappa$
\item $|\Lang| \leq \kappa$,
\item $\cM$ is an $\Lang$-structure,
\item $B$ is a collection of distinct $<\lambda$-generated substructures of $\cM$ with $|B| = \kappa^+$.
\end{itemize}
Then there is a $\cN \subseteq \cM$ and a collection $(\cN_i)_{i \in \kappa^+}$ with $\cN_i \in B$ for all $i \in\kappa^+$ and such that $(\forall i < j \in \kappa^+)\, \cN_i \cap \cN_j = \cN$
\end{lemma}
\begin{proof}
First, note that if $\cM^\circ$ is the smallest substructure of $\cM$ generated by $\bigcup B$ then $|\cM^\circ| \leq |B| \cdot (< \lambda) \cdot |\Lang|= \kappa^+$. We can therefore assume without loss of generality that $|\cM| = \kappa^+$ and therefore we can assume without loss of generality that the underlying set of $\cM$ is the cardinal $\kappa^+$.
Let $g\:B \to \cM^{<\lambda}$ be a map such that for all $b \in B$, $b$ is the substructure of $\cM$ generated by $g(b)$.
%
%As $\kappa^{<\lambda} = \kappa$ we can find a bijection $f\:\kappa^+ \to \cM^{<\lambda}$. We can further assume $f$ is such that there is a cofinal collection $(\gamma_i)_{i \in \kappa^+} \subseteq \kappa^+$ and an increasing sequence $(\cM_i)_{i \in \kappa^+}$ such that $|\cM_i| < \kappa^+$ for all $i \in \kappa^+$ and $f``[\gamma_i] = \cM_i^{<\gamma}$.
%
Let $X = \{\Lang, \Rel[\Lang], \Func[\Lang], \ar[\Lang]\} \cup \{B\} \cup \{g\} \cup \{\cM\}$.
As $\kappa^{<\lambda} = \kappa$ we have $\lambda \leq \kappa$.
Let $\theta > \kappa$ be a large enough regular cardinal, let $\mathcal{H}(\theta)$ be the class of hereditarily sets of cardinality less than $\theta$, and let
$<^*$ be a well-ordering of $\mathcal{H}(\theta)$. Consider the structure $\mathcal{H}=(\mathcal{H}(\theta), \in, <^*).$
Let $(W_i)_{i \in \kappa+1}$ be an increasing and continuous sequence of elementary substructures of $\mathcal{H}$ such that
\begin{itemize}
\item[(a)] $X \in W_0$,
\item[(b)] $|W_i| = \kappa$ for all $i \in \kappa$,
\item[(c)] $W_i^{<\lambda} \subseteq W_{i+1}$ for all $i \in \kappa$.
\item[(d)] if $x \in W_i \cap \ORD$ and $y \in \kappa^+$ with $y \in x$, then $y \in W_{i+1}$,
\item[(e)] $W_{\gamma} = \bigcup_{i \in \gamma} W_{i}$, for limit ordinal $\gamma$.
\end{itemize}
Note we can always find such a sequence as $\kappa^{<\lambda} = \kappa$. %Further, as $\kappa^{<\lambda} = \kappa$ we have as a consequence of \Konig's theorem that $\cofinal(\kappa) \geq \lambda$. Therefore for every $a \in V_{\kappa}^{<\lambda}$ there is an $i \in \kappa$ such that $a \in V_i$.
By (d) we have $(\kappa^+ \cap W_\kappa) \in \ORD$. Let this value be $\zeta$. As $|W_\kappa| = \kappa$ we have $\zeta < \kappa^+$. Therefore there must be some $B_0 \in B$ such that $B_0 \not \in W_\kappa$. Let $r = g(B_0) \cap \zeta$ and let $r^+$ be the substructure of $\cM$ generated by $r$. Note $r$ and hence $r^+$ are in $W_\kappa$. Note that if $<_{\cM}$ is the ordering on $\cM$ given from the fact that the underlying set of $\cM$ is $\kappa^+$, then $(\forall x \in B_0 \setminus r^+)\, x \not \in W_\kappa$ and hence $x >_{\cM} \zeta$.
Suppose $\alpha \in \zeta$. We then have
\[
\mathcal{H} \models (\exists B_0 \in B)\, \Bigl( r \subseteq \UnderSet{B_0} \,\And \,(\forall b \in \UnderSet{B_0} \setminus r^+)\, b >_{\cM} \alpha\Bigr).
\]
Therefore we have for each $\alpha \in \zeta$
\[
W_\kappa \models (\exists B_0 \in B)\, \Bigl(r \subseteq \UnderSet{B_0} \, \And \, (\forall b \in \UnderSet{B_0} \setminus r^+)\, b >_{\cM} \alpha\Bigr).
\]
But then we also have
\[
W_\kappa \models (\forall \alpha)(\exists B_0 \in B)\, \Bigl(r \subseteq \UnderSet{B_0} \,\And\, (\forall b \in \UnderSet{B_0} \setminus r^+)\, b >_{\cM} \alpha\Bigr).
\]
and so
\[
\mathcal{H} \models (\forall \alpha)(\exists B_0 \in B)\,\Bigl( r \subseteq \UnderSet{B_0} \,\And\, (\forall b \in \UnderSet{B_0} \setminus r^+)\, b >_{\cM} \alpha\Bigr).
\]
But as $|b| \leq \kappa$ for all $b \in B$ we can find a sequence $(b_i)_{i \in \kappa^+}$ with $b_i \in B$ for all $i \in \kappa^+$ and such that
\begin{itemize}
\item for all $i \in \kappa^+$, $r^+ \subseteq \UnderSet{b_i}$,
\item for all $i < j \in \kappa^+$, $\max (\UnderSet{b_i}\setminus r^+) < \min (\UnderSet{b_j} \setminus r^+)$.
\end{itemize}
Therefore $(b_i)_{i \in \kappa^+}$ is a $\Delta$-system.
\end{proof}
Note that even in the case when $\kappa = \lambda = \w$ this result need not follow from the ordinary $\Delta$-systems lemma. This is because, in this case, the $\Delta$-systems lemma only deals with finite subsets of $\w_1$. However, in \cref{Delta-Systems Lemma} the elements of $B$ might be countable when there is a countable language. In our context though this isn't a problem as each of our sets is finitely generated and there cannot be a finitely generated substructure of $\cM$ which is a subset of $W_\w$ but is not also contained in $W_\w$.
We now prove our main theorem.
\begin{theorem}
\label{Conditions which guarantee the chain condition}
Suppose
\begin{itemize}
\item $\kappa^{<\lambda} = \kappa$,
\item $\LangBot \subseteq \Lang$, $\Lang\setminus \LangBot$ is relational, and $|\Lang| \leq \kappa$,
\item $\AgeK$ is a $\lambda$-bounded strong \Fraisse\ class with (C$\lambda$U),
\item $\AgeK\rest[\LangBot]$ is a strong \Fraisse\ class,
\item $\AgeK$ has the extended strong amalgamation property over $\LangBot$,
\item $\MBot$ is an $\LangBot$-structure, and
\item for every $\LangBot$-structure $\cX \subseteq \MBot$ which is $< \lambda$-generated there are at most $\kappa$-many elements $\cN \in \AgeK$ with $\cN \rest[\LangBot] = \cX$.
\end{itemize}
Then any subset $Y$ of $\Fn<\MBot>[\AgeK](\lambda)$ of size $\kappa^+$ has a subset $Y_0 \subseteq Y$ of size $\kappa^+$ such that every two elements of $Y_0$ have a common lower bound.
In particular, this implies $\Fn<\MBot>[\AgeK](\lambda)$ has the $\kappa^+$-chain condition and, if $\kappa = \w_1$, $\Fn<\MBot>[\AgeK](\lambda)$ satisfies the Knaster condition.
\end{theorem}
\begin{proof}
Suppose, to get a contradiction, that $\{\cA_i\}_{i \in \kappa^+} \subseteq \Fn<\MBot>[\AgeK](\kappa)$ and for any $J \subseteq \kappa^+$ with $|J| = \kappa^+$ there are $x, y \in J$ such that $\cA_{x}$ and $\cA_y$ have no common extension in $\Fn<\MBot>[\AgeK](\lambda)$.
Let $\mbf{A} = \{\cA_i\rest[\LangBot]\}_{i \in \kappa^+}$. For each $i \in \kappa^+$ we have $\cA_i$ is $< \lambda$-generated as $\AgeK$ is $\lambda$-bounded. Therefore there are at most $\kappa$ many elements $\cN \in \AgeK$ with $\cN \rest[\LangBot] = \cA_i\rest[\LangBot]$. Therefore $\kappa^+ = |\{\cA_i\}_{i \in \kappa^+}| \leq |\mbf{A}| \cdot \kappa$ and hence $|\mbf{A}| = \kappa^+$.
But by \cref{Delta-Systems Lemma}, we can find a subset $I_0 \subseteq \kappa^+$ with $|I_0| = \kappa^+$ and a $\cX \subseteq \MBot$ such that $(\forall i < j \in I_0)\, \UnderSet{\cA_i} \cap \UnderSet{\cA_j} = \UnderSet{\cX}$. But then $\cX \in \AgeK\rest[\LangBot]$ and so is $< \lambda$-generated. Therefore there are at most $\kappa$-many elements $\cN \in \AgeK$ with $\cN \rest[\LangBot] = \cX$. But then there must be a subset $I_1 \subseteq I_0$ such that $|I_1| = \kappa^+$ and $(\forall i \in I_1)\, (\cA_i\rest[\UnderSet{\cX}])\rest[\LangBot] = \cX$.
For $i < j \in I_1$ let $\cB_{i, j}^-$ be the subset of $\MBot$ generated by $\UnderSet{\cA_i} \cup \UnderSet{\cA_j}$. As $\AgeK$ has the extended strong amalgamation property over $\LangBot$ there must be a $\cB_{i, j} \in \AgeK$ such that $\cA_i \subseteq \cB_{i, j}$, $\cA_j \subseteq \cB_{i, j}$ and $\cB_{i, j}\rest[\LangBot] = \cB_{i, j}^-$. Then $I_1$ contradicts our assumption.
\end{proof}
We end with a straightforward observation about the theory of generics for these partial orders.
\begin{proposition}
\label{Forcing satisfies Fraisse theory}
Suppose
\begin{itemize}
\item $\LangBot \subseteq \Lang$, $\Lang \setminus \LangBot$ is relational, and $|\Lang| \leq \kappa$,
\item $\AgeK$ and $\AgeK\rest[\LangBot]$ are strong \Fraisse\ classes,
\item $\AgeK$ is $\lambda$-bounded and satisfies the $(C\lambda U)$,
\item $\AgeK$ has the extended strong amalgamation property over $\LangBot$,
\item $\MBot$ is an $\LangBot$-structure which satisfies the $\ThFraisse_\kappa(\AgeK\rest[\LangBot])$,
\item $G$ is generic for $\Fn<\MBot>[\AgeK](\kappa)$ over $V$ and $\cG = \bigcup G$.
\end{itemize}
Then $V$ and $V[G]$ have the same sequences of elements of $V$ of length less than $\lambda$ and
$V[G] \models \text{``}\cG \models \ThFraisse_\kappa(\AgeK)$''.
\end{proposition}
\begin{proof}
As $\AgeK$ satisfies the $(C\lambda U)$, by Proposition \ref{propclosure}, the forcing notion
$\Fn<\MBot>[\AgeK](\kappa)$ is $\lambda$-closed, and hence $V$ and $V[G]$ have the same sequences of elements of $V$ of length less than $\lambda$.
For each $\cA \in \AgeK$ it is immediate that
\[
D_{\cA} = \{\cB \in \AgeK[\MBot]\st \cA \subseteq \cB\}
\]
is dense. Therefore for all $\cA \in \AgeK$ and $\aa \in \cA$ with $\aa$ generating $\cA$, $\cG \models (\exists \xx)\, \bigwedge\qftp[\cA](\aa)(\xx)$.
Further, as $\AgeK$ has the extended strong amalgamation property over $\LangBot$ we have that for any $\cA, \cB \in \AgeK[\MBot]$ with $\cA \subseteq \cB$ we have $D_{\cA, \cB}$ is dense, where
$D_{\cA, \cB}=D^1_{\cA, \cB} \cup D^2_{\cA, \cB}$, and
\[
D^1_{\cA, \cB} = \{\cC \in \AgeK[\MBot] \st \cA \subseteq \cC \text{ and }(\exists \cB^* \in \AgeK[\MBot])\, \cA \subseteq \cB^* \subseteq \cC\text{ and }(\cA, \cB) \cong (\cA, \cB^*)\}
\]
and
\[
D^2_{\cA, \cB} = \{\cC \in \AgeK[\MBot] \st \cA\rest[\LangBot] \subseteq \cC\rest[\LangBot] \text{ and }\cA \not \subseteq \cC\}.
\]
Therefore if $\aa$ generates $\cA$ and $\bb$ generates $\cB$
\[
\cG \models (\forall \xx) \bigg[\bigwedge\qftp[\cA](\aa)(\xx) \rightarrow (\exists \yy)\, \bigwedge\qftp[\cB](\aa\bb)(\xx\yy)\bigg].
\]
But, $V$ and $V[G]$ have the same sequences of elements of $V$ of length less than $\lambda$, and so in $V[G]$
\[
\cG \models (\forall \xx)\, \bigg[\bigwedge\qftp[\cA](\aa)(\xx) \rightarrow (\exists \yy)\, \bigwedge\qftp[\cB](\aa\bb)(\xx\yy)\bigg].
\]
\end{proof}
%%%% %%%% %%%% %%%% %%%% %%%% %%%% %%%%
\section{Rigidity}
%%%% %%%% %%%% %%%% %%%% %%%% %%%% %%%%
There are many situations when the generics generated by forcing with an age give rise to rigid structures, even when the underlying functional structure is not rigid. In this section, we give general conditions which will ensure such generic structures are rigid.
The conditions below which ensure rigidity of the generic are an abstraction of the techniques in \cite{Cohen} used to show rigidity of the uncountable generics for graphs, partial orders and linear orders. As we will see, the techniques in \cite{Cohen} make fundamental use of the fact that the languages are binary.
In order to understand the motivation for the following technical definitions it is worth looking at a high level summary of the proof of rigidity, which occurs in several stages. Recall the set up is that we have languages $\LangBot \subseteq \Lang$ where $\Lang \setminus \LangBot$ is relational. We also have a strong \Fraisse\ class $\AgeK$ in $\Lang$ which satisfies the extended strong amalgamation property over $\LangBot$. Finally we have an uncountable $\LangBot$-structure $\MBot$ whose age is $\AgeK\rest[\LangBot]$.
The proof occurs in several stages. First, consider pairs $(\cF, \cS)$ such that (a) $\cS$ is a sort of $\LangBot$ and where no non-identity automorphism of $\MBot$ is the identity on $\cS^{\MBot}$ and (b) $\cF$ is a subset $\MBot$ such that for two elements $x, y \in \cS^{\MBot}$ and any generic structure $\cG$ there is a tuple $\zz \in \cF$ such that $\qftp[\cG](x\zz) \neq \qftp[\cG](y\zz)$. We say such an $\cF$ has the \emph{splitting property} over $\AgeK$ and $\cS$ as we can always split the types of elements of $\cS$ using elements of $\cF$.
For such a pair any non-trivial automorphism $\cG$ does not fix $\cF$. Specifically, if $h$ was a non-trivial automorphism of $\cG$ that fixed $\cF$, then there must be distinct $x,y \in \cS^{\MBot}$ with $h(x) = y$. But then there would a $\zz \in \cF$ with $\qftp[\cG](x\zz) \neq \qftp[\cG](y\zz) = \qftp[\cG](h(x)h(\zz))$ getting us a contradiction. Our first assumption will be that there are many such pairs $(\cF_i, \cS_i)$ such that for distinct $i, j$, $\cF_i \cap \cF_j$ is the term closure of the empty set in $\MBot$.
Now assume to get a contradiction that $\name{h}$ is the name for an automorphism in the generic extension $V[G]$. For any such $\cF$ we can find an $s_{\cF} \in\cF$ and a $\cB_{\cF}$ in our generic such that $\cB_{\cF}$ forces $\name{h}(s_{\cF}) \neq s_{\cF}$. As there are many such $\cF$ we can find a large collection such that all of $\cB_{\cF}$ are isomorphic.
Using the analog of the $\Delta$-system lemma, we can therefore find a $\Delta$-system of elements $\cB_{\cF}$ such that for any such $\cF_0, \cF_1$ there is an isomorphism $i\:\cB_{\cF_0} \to \cB_{\cF_1}$ where $i(s_{\cF_0}) = s_{\cF_1}$ and $i(\name{h}(s_{\cF_0})) = \name{h}(s_{\cF_1})$.
Our second assumption about the age, which we call the \emph{amalgamation of distinct $2$-types property}, essentially this says that whenever we have $xy$ and $x'y'$ of the same quantifier-free type we can find an amalgamation such that $\qftp(xx') \neq \qftp(yy')$. By construction $\qftp[\cB_{\cF_0}](s_{\cF_0}\name{h}(s_{\cF_0})) = \qftp[\cB_{\cF_1}](s_{\cF_1}\name{h}(s_{\cF_1}))$. Therefore
our second assumption will allow us to find an amalgamation $\cC$ of $\cB_{\cF_0}$ and $\cB_{\cF_1}$ with $\qftp(s_{\cF_0} s_{\cF_1}) \neq \qftp(\name{h}(s_{\cF_0}) \name{h}(s_{\cF_1}))$. But any such $\cC$ must forces that $\name{h}$ is not an automorphism, contradicting the fact that $\cB_{\cF_0}$ forces that $\name{h}$ is an automorphism and $\cB_{\cF_0} \subseteq \cC$.
It is worth noting that in this argument we are focused on the quantifier-free type of pairs of elements. It is for this reason that in \cite{Cohen} all the examples of rigid structures where in binary languages. In particular, the techniques do not apply, for example, to the case of a language $\Lang$ with a single $n$-ary relation ($n > 2$) and the age of all finite $\Lang$-structure. It is though natural to ask if an uncountable generic for such a structure must be rigid? We will show by expanding our structure to add names for all $n$-tuples, something which fundamentally requires functions, that our techniques prove such generics are rigid. %\fTBD{Maybe put in intro to paper}
We now turn to the definitions we will need for our main result of this section.
\begin{definition}
\label{Amalgamation of distinct types property}
%\TBD{REMOVE:Suppose $\kappa^{<\lambda} = \kappa$, $|\LangBot| \leq \kappa$ and $\AgeK$ is a $\lambda$-bounded age in $\Lang$ such that for all $\cA \in \AgeK$ there are, up to isomorphisms which fix $\cA$, at most $\kappa$-many elements $\cB \in \AgeK$ with $\cA \subseteq \cB$.}
For a sort $\cS \in \LangBot$ we say an age $\AgeK$ in $\Lang$ has the \defn{amalgamation of distinct $2$-type property} (over $\LangBot$ and $\cS$) if whenever
\begin{itemize}
\item[(a)] $\cA, \cB, \cC \in \AgeK$,
\item[(b)] $\cA \subseteq \cB$, $\cA \subseteq \cC$, and $\UnderSet{\cB} \cap \UnderSet{\cC} = \UnderSet{\cA}$,
\item[(c)] $\alpha\:\cB \to \cC$ is an isomorphism with $\alpha\rest[\cA] = \id_{\cA}$,
\item[(d)] $x_0, x_1 \in \cS^{\cB} \setminus \cS^{\cA}$ are distinct elements with $\qftp[\cB](x_0) = \qftp[\cB](x_1)$,
\item[(e)] $\DBot \in \AgeK\rest[\LangBot]$ with $\cB\rest[\LangBot] \subseteq \DBot$ and $\cC\rest[\LangBot] \subseteq \DBot$,
\end{itemize}
then there is a $\cD \in \AgeK$ such that
\begin{itemize}
\item[(f)] $\cB \subseteq \cD$, $\cC \subseteq \cD$,
\item[(g)] $\cD\rest[\LangBot] = \DBot$,
\item[(h)] $\qftp[\cD](x_0\alpha(x_0)) \neq \qftp[\cD](x_1\alpha(x_1))$
\end{itemize}
Note when there is only one sort we will omit mention of it.
%\TBD{Note we don't need the cardinality assumptions in this definition}
\end{definition}
It is worth noting that an age having the amalgamation of distinct $2$-types is not enough to ensure uncountable generics are rigid, at least when $\LangBot$ is not the empty language.
\begin{example}
\label{Example amalgamation of distict 2-types doesn't imply rigid generic}
Consider the language $\LangBot = \{U, W\}$ where $U$ and $W$ are unary relations. Let $\AgeKBot$ be the age consisting of all finite $\LangBot$-structures $\cA$ such that $\{U^{\cA}, W^{\cA}\}$ is a partition of the underlying set. Let $\Lang = \LangBot \cup \{E\}$ and let $\AgeK$ be the collection of finite $\Lang$-structures such that if $E(x, y)$ holds, then $U(x) \leftrightarrow U(y)$.
Now suppose $\cA, \cB, \cC, \DBot, x_0, x_1, \alpha$ are in \cref{Amalgamation of distinct types property}. As $\qftp[\cB](x_0) = \qftp[\cB](x_1)$ we know that either $\cB \models U(x_0) \And U(x_1)$ or $\cB \models W(x_0) \And W(x_1)$. In the first case, we then have $\DBot \models U(\alpha(x_0)) \And U(\alpha(x_1))$. But as $\alpha(x_0), \alpha(x_1) \not \in \cB$ there is a $\cD \in \AgeK$ with $\cD \rest[\LangBot] = \DBot$, $\cB, \cC \subseteq \cD$ and $\cD \models E(x_0, \alpha(x_0)) \And\neg E(x_1, \alpha(x_1))$.
But if $\cB \models W(x_0) \And W(x_1)$, then we can also find such a $\cD$ as the only restriction on the edge relation $E$ is that there are no edges between elements satisfying $U$ and elements satisfying $W$. In particular this proves that $\AgeK$ has the amalgamation of distinct $2$-type property over $\LangBot$.
Now suppose $\MBot_{\kappa, \w}$ is the unique $\LangBot$-structure (up to isomorphism) where $|U^{\MBot_{\kappa, \w}}| = \kappa$ and $|W^{\MBot_{\kappa, \w}}| = \w$. Suppose $\cM$ is any $\Lang$-structure whose age is contained in $\AgeK$. If $g_0\:\cM\rest[U^{\cM}] \to \cM\rest[U^{\cM}]$ is an automorphism and $g_1\:\cM\rest[W^{\cM}] \to \cM \rest[W^{\cM}]$ is an automorphism, then $g_0 \cup g_1\:\cM \to\cM$ is an automorphism.
Let $G$ be a generic for $\Fn<\MBot_{\kappa, \w}>[\AgeK](\w)$ and let $\cG$ be the corresponding $\Lang$-structure. We then have that $\cR = (W^{\cG}, E^{\cG} \cap (W^{\cG} \times W^{\cG}))$ is isomorphic to the Rado graph. Let $I$ be the identity on $U^{\cG}$ and let $g$ be any automorphism of $\cR$ in $V[G]$. We then have that $I \cup g$ is an automorphism of $\cG$ and so $\cG$ is not rigid in $V[G]$.
%-----
%Let $\LangBot = \{U_n\}_{n \in \w} \cup \{f\} \cup \{E\}$ where $U$ and $V$ are sorts, $f\:U \to U$ and $E$ is a relation of arity $\<V, U\>$. Let $\AgeKBot$ be the collection of finite $\LangBot$-structures $\cA$ such that
%\begin{itemize}
%
%\item there is a unique element $r \in U^{\cA_0}$ such that
%\begin{itemize}
%\item $\cA_0 \models f(r) = r$,
%
%\item for all $a \in U^{\cA_0}$ there is an $n \in \w$ such that $\cA \models f^{n}(a) = r$.
%
%\end{itemize}
%
%\item $\cA_0 \models E(a, b) \And
%
%\end{itemize}
%
%
%Let $\Lang = \LangBot \cup \{R\}$ where $R$ has arity $\<U, U\>$\fTBD{Double check this is the right arity notation}.
%
%
%Let $\AgeK$ be the collection of $\Lang$-structures $\cA$ such that
%\begin{itemize}
%\item $\cA \rest[\LangBot] \in \Age_0$,
%
%\item $(V^{\cA}, R^{\cA})$ is a graph,
%
%\item $\cA \models (\forall x,y)\, R(x, y) \rightarrow g(x) = g(y)$.
%
%\end{itemize}
%
%\end{itemize}
%
%If $\cA$ is in $\AgeK$ then we can think of elements of $U^{\cA}$ as being names for equivalence classes in some equivalence relation (i.e. that of $g(x) = g(y)$). We then let $R$ be the edge relation of a graph which only can connect elements in the same equivalence class. Finally, there must be a unique $n \in \w$ and bijection $i\:U^{\cA} \to [n]$ such that $f^{i(a)}(a) = r$ but $f^{i(a) - 1}(a) \neq r$ (where here we say $f^{-1}(r)$ is undefined). In particular this will ensure that in any model whose age is contained in that of $\AgeK$ the size of $U$ is countable.
%
%To show $\AgeK$ has the amalgamation of distinct $2$-type property suppose
%\begin{itemize}
%\item $\cA, \cB, \cC \in \AgeK$,
%
%\item $x, y \in \cB$ with $\qftp[\cB](x) = \qftp[\cB](y)$,
%
%\item $x', y' \in \cC$,
%
%\item $\UnderSet{\cB} \cap \UnderSet{\cC} = \UnderSet{\cA}$,
%
%\item $x, y \not \in \UnderSet{\cA}$,
%
%\item and there is an isomorphism from $\cB$ to $\cC$ which takes $xy$ to $x'y'$ and fixes $\cA$.
%\end{itemize}
%
%In particular, this implies $\qftp[\cB](xy) = \qftp[\cC](x'y')$. Let $\DBot$ be the (unique) $\LangBot$-structure with $\UnderSet{\DBot} = \UnderSet{\cB} \cup \UnderSet{\cC}$ and $\cB\rest[\LangBot], \cC\rest[\LangBot] \subseteq \DBot$. We break into cases depending on the $\LangBot$-structure of $\qftp[{\cB\rest[\LangBot]}](xy)$. \nl
%
%\ul{Case $1$:} $f(x) = x$ and $f(y) = y$. \nl
%In this case we can find an amalgamation of $xy$ and $x'y'$ where $R_0(x,x')$ holds but $\neg R_0(y, y')$ holds. \nl\nl
%
%\ul{Case $2$:} $f(x) \neq x$, $f(y) \neq y$ and $f(x) = f(y)$. \nl
%If $f(x) = f(x')$ then $f(y) = f(y')$ and $x, y, x', y'$ are all in the same tree. We can then find amalgamation of $xy$ and $x'y'$ where $R_1(x,x')$ holds but $\neg R_1(y, y')$ holds. If $f(x) \neq f(x')$, then $f(y') \neq f(y)$ and so we can find an amalgamation where \TBD{NO!!!} $R_0(f(x)f(x')) \neq R_0(f(y)f(y'))$.\nl\nl
%
%\ul{Case $3$:} $f(x) \neq x$, $f(y) \neq y$ and $f(x) \neq f(y)$. \nl
%If $f(x) \neq f(x')$ and $f(y) \neq f(y')$, then we can find an amalgamation where $R_0(f(x)f(x'))$ and $\neg R_0(f(y)f(y'))$. If $f(x) \neq f(x')$ and $f(y) = f(y')$, then $\neg R_0(f(y)f(y'))$ holds and we can find an amalgamation where $R_0(f(x)f(x'))$ holds. Finally, if $f(x) = f(x')$ and $f(y) = f(y')$, then we can find an amalgamation where $R_1(xx')$ and $\neg R_1(yy')$ holds. \nl\nl
%
%Therefore $\AgeK$ has the amalgamation of distinct $2$-types property. \fTBD{Double check it is clear enough. In particular we never consider the base structure $\cA$ or explicitly mention $\DBot$}
%
%Now let $\cM_{0, \kappa}$ be the $\LangBot$-structure where there are $\kappa$-many elements $x$ with $f(x) = x$ and for each such $x$, $|f^{-1}(x)| = \w$. For $x \in \cM_{0, \kappa}$ with $f(x) = x$ let $F_x = f^{-1}(x)$. If $\cM$ is any structure which satisfies $\ThFraisse_\w(\AgeK)$ with $\cM \rest[\LangBot] = \cM_{0, \kappa}$ then $\Aut(\cM)$ is non-trivial. This is because for all $x \in \cM_{0, \kappa}$ with $f(x) = x$, $\cM\rest[F_x]$ has an automorphism, being essentially a copy of the Rado graph with edge relation $R_1$. But given any function $\alpha\: \{x \in \cM \st f(x) = x\}$ to $\Aut(\cM\rest[F_x])$ there is a unique $g \in \Aut(\cM)$ which is constant on $\{x \st f(x) = x\}$ and where $g \rest[F_x] = \alpha(x)$ for all $x$ with $f(x) = x$.
%
%
%----
%Let $\LangBot = \{f\}$ where $f$ is a unary function. Let $\Lang = \LangBot \cup \{R_0, R_1\}$. Let $\AgeK$ be the collection of structures such that
%\begin{itemize}
%
%\item for all $x$, $f^2(x) = f(x)$,
%
%\item $R_1(x, y)$ is a graph and if $R_1(x, y)$ then $f(x) = f(y) \not \in \{x, y\}$,
%
%\item $R_0(x, y)$ is a graph and whenever $R_0(x, y)$ holds $f(x) = x$ and $f(y) = y$.
%\end{itemize}
%
%If $\cA$ is in $\AgeK$ then $\cA$ can be thought of as a forest where all trees have height at most $1$ and $f$ is the predecessor function. We then can have an arbitrary graph $R_0$ between roots of trees and an arbitrary graph $R_1$ between elements of the same tree (but no $R_1$-edges between elements of different trees).
%
%To show $\AgeK$ has the amalgamation of distinct $2$-type property suppose $x, y \in \cB$ and $x', y' \in \cC$ where $\UnderSet{\cB} \cap \UnderSet{\cC} = \UnderSet{\cA}$, $x, y \not \in \UnderSet{\cA}$ and there is an isomorphism from $\cB$ to $\cC$ which takes $xy$ to $x'y'$. In particular, this implies $\qftp[\cB](xy) = \qftp[\cC](x'y')$. Let $\DBot$ be the (unique) $\LangBot$-structure with $\UnderSet{\DBot} = \UnderSet{\cB} \cup \UnderSet{\cC}$ and $\cB\rest[\LangBot], \cC\rest[\LangBot] \subseteq \DBot$. We break into cases depending on the $\LangBot$-structure of $\qftp[{\cB\rest[\LangBot]}](xy)$. \nl
%
%\ul{Case $1$:} $f(x) = x$ and $f(y) = y$. \nl
%In this case we can find an amalgamation of $xy$ and $x'y'$ where $R_0(x,x')$ holds but $\neg R_0(y, y')$ holds. \nl\nl
%
%\ul{Case $2$:} $f(x) \neq x$, $f(y) \neq y$ and $f(x) = f(y)$. \nl
%If $f(x) = f(x')$ then $f(y) = f(y')$ and $x, y, x', y'$ are all in the same tree. We can then find amalgamation of $xy$ and $x'y'$ where $R_1(x,x')$ holds but $\neg R_1(y, y')$ holds. If $f(x) \neq f(x')$, then $f(y') \neq f(y)$ and so we can find an amalgamation where \TBD{NO!!!} $R_0(f(x)f(x')) \neq R_0(f(y)f(y'))$.\nl\nl
%
%\ul{Case $3$:} $f(x) \neq x$, $f(y) \neq y$ and $f(x) \neq f(y)$. \nl
%If $f(x) \neq f(x')$ and $f(y) \neq f(y')$, then we can find an amalgamation where $R_0(f(x)f(x'))$ and $\neg R_0(f(y)f(y'))$. If $f(x) \neq f(x')$ and $f(y) = f(y')$, then $\neg R_0(f(y)f(y'))$ holds and we can find an amalgamation where $R_0(f(x)f(x'))$ holds. Finally, if $f(x) = f(x')$ and $f(y) = f(y')$, then we can find an amalgamation where $R_1(xx')$ and $\neg R_1(yy')$ holds. \nl\nl
%
%Therefore $\AgeK$ has the amalgamation of distinct $2$-types property. \fTBD{Double check it is clear enough. In particular we never consider the base structure $\cA$ or explicitly mention $\DBot$}
%
%Now let $\cM_{0, \kappa}$ be the $\LangBot$-structure where there are $\kappa$-many elements $x$ with $f(x) = x$ and for each such $x$, $|f^{-1}(x)| = \w$. For $x \in \cM_{0, \kappa}$ with $f(x) = x$ let $F_x = f^{-1}(x)$. If $\cM$ is any structure which satisfies $\ThFraisse_\w(\AgeK)$ with $\cM \rest[\LangBot] = \cM_{0, \kappa}$ then $\Aut(\cM)$ is non-trivial. This is because for all $x \in \cM_{0, \kappa}$ with $f(x) = x$, $\cM\rest[F_x]$ has an automorphism, being essentially a copy of the Rado graph with edge relation $R_1$. But given any function $\alpha\: \{x \in \cM \st f(x) = x\}$ to $\Aut(\cM\rest[F_x])$ there is a unique $g \in \Aut(\cM)$ which is constant on $\{x \st f(x) = x\}$ and where $g \rest[F_x] = \alpha(x)$ for all $x$ with $f(x) = x$.
\end{example}
In \cref{Example amalgamation of distict 2-types doesn't imply rigid generic} we were able to find generics which were not rigid because the underlying structure had a countable set whose structure was independent of the structure of the rest of the model. The next property which we will need guarantees that can't happen.
\begin{definition}
Suppose $\Lang$ is a language with sort $\cS$ and $\AgeK$ is an age over $\Lang$. We say $\cS$ \defn{pins} $\AgeK$ if whenever
\begin{itemize}
\item $\cA, \cB \in \AgeK$,
\item $i_0, i_1\:\cA \to \cB$ are isomorphisms,
\item $i_0 \rest[\cS^{\cA}] = i_1 \rest[\cS^{\cA}]$
\end{itemize}
then $i_0 = i_1$.
\end{definition}
In particular if $\cS$ pins $\AgeK$ and $\cM$ is any structure whose age is contained in $\AgeK$ then any non-trivial automorphism of $\cM$ must move an element of $\cS^\cM$. The following is an important example of this.
\begin{example}
\label{Exaple of names for sequences}
Let $\LangBot_{n} = \{U, W\} \cup \{f\} \cup \{\pi_i\}_{i \in [n]}$ where $U$ and $W$ are the sorts, $f\:U^n \to W$ and for each $i \in [n]$, $\pi_i\:W \to U$. Let $\AgeKBot_{s, n}$ be the collection of structures such that
\begin{itemize}
\item $f$ is a bijection,
\item for all $v \in W$, $f(\pi_0(v), \dots, \pi_{n-1}(v)) = v$.
\end{itemize}
In this example we can think of elements of $W$ as \emph{names} for sequences of elements of $U$ of length $n$.
If $\MBot$ is a $\LangBot$-structure with age $\AgeKBot_{s, n}$ then any non-trivial automorphism of $\MBot$ must be non-trivial on $W^{\MBot}$. Therefore $W$ pins $\AgeKBot_{s, n}$.
\end{example}
Fundamental to the proof of the rigidity of generics is the fact that we can find many disjoint sets $\cF$ such that, whenever there is an automorphism $h$ of the structure, there is some element $s$ such that the quantifier-free type of $s$ over $\cF$ is different than the quantifier-free type of $h(s)$ over $\cF$. This will ensure that any non-trivial automorphism moves some element of $\cF$. We now make precise this property.
\begin{definition}
\label{Secondary rigidity property}
Suppose $\cS$ pins $\AgeK$. We say $\cF \subseteq \MBot$ has the \defn{splitting property} (over $\AgeK$ and $\cS$ and $\MBot$) if whenever $\cA \in \AgeK[\MBot]$ and $x, y \in \cA^{\cS}$ there is a $\cB \in \AgeK[\MBot]$ with $\cA \subseteq \cB$ and a tuple $\zz \in \cB \cap \cF$ such that $\qftp[\cB](x \zz) \neq \qftp[\cB](y \zz)$.
\end{definition}
Note that if $\LangBot$ is the one sorted language with no relations and $\MBot$ is an infinite $\LangBot$-structure then every infinite subset $\cF\subseteq \MBot$ (of sufficient size) has the splitting property over any non-trivial age. As such this is a property which only is needed in the proofs when the language $\LangBot$ is non-trivial.
\begin{proposition}
\label{Main rigidity result}
Suppose $V$ is a model of $\ZFC$ in which the following hold.
\begin{itemize}
\item $\kappa$ and $\lambda$ are cardinals with $\kappa^{<\lambda} = \kappa$ and $n \in \w$,
\item $\LangBot \subseteq \Lang$ with $|\LangBot| \leq \kappa$ and $\Lang \setminus \LangBot$ relational.
\item $\AgeK$ and $\AgeK\rest[\LangBot]$ are $\lambda$-bounded strong \Fraisse\ classes and $\AgeK$ has extended strong amalgamation over $\LangBot$,
\item for all $\cA \in \AgeK$ there are, up to isomorphism which fix $\cA$, at most $\kappa$-many elements $\cB \in \AgeK$ with $\cA \subseteq \cB$.
\item $\AgeK\rest[\LangBot]$ is a strong \Fraisse\ class.
\item $\MBot\models \ThFraisse_\lambda(\AgeK\rest[\LangBot])$
%
%$\MBot$ is a $\LangBot$-structure with $\AgeK<\lambda>[\MBot] = \AgeK\rest[\LangBot]$,
\item there is a collection of sorts $\{\cS_i\st i \in \kappa^+\}$, a collection of sorts $\{\cT_i\st i \in \kappa^+\}$, and a collection of sets $\{\cF_i \st i \in \kappa^+\}$ such that for all $i \in \kappa^*$
\begin{itemize}
\item $\cF_i \subseteq \cS_i^{\MBot}$ and $\cF_i \in V$
\item $\AgeK$ has the amalgamation of distinct $2$-type property over $\LangBot$ and $\cS_i$,
\item $\cT_i$ pins $\AgeK$
\item $\cF_i$ has splitting property over $\AgeK$ and $\cT_i$,
\end{itemize}
and $(\forall i < j \in \kappa)\, \UnderSet{\cF_i} \cap \UnderSet{\cF_j} = \UnderSet{\tdcl[\MBot](\emptyset)}$
\end{itemize}
Then whenever $G$ is a generic for $\Fn<\MBot>[\AgeK\rest[\LangBot]](\lambda)$ over $V$, $\bigcup G$ is rigid in $V[G]$.
\end{proposition}
\begin{proof}
Let $\cG = \bigcup G$. Suppose, to get a contradiction, this result fails. Then there must be some name $\name{h}$ for $\Fn<\MBot>[\AgeK](\lambda)$ in $V$ and $\cA \in G$ such that
\[
\cA \forces[\Fn<\MBot>[\AgeK](\lambda)] \name{h}\text{ is a non-trivial automorphism of }\cG.
\]
%\begin{claim}
%\label{Main rigidity result: Claim 1}
%Suppose $\cA, \cB, \cC, \alpha, \xx_0, \xx_1$ are as in \cref{Amalgamation of distinct types property}. Then there is a $\cD \in \cK$ which is an amalgamation of $\cB$ and $\cC$ over $\cA$\fTBD{Make sure "an amalgamation" is defined} and a $\yy \in \cD \setminus \cB$ such that $\qftp[\cD](\xx_0\yy) \neq \qftp(\xx_1\yy)$.
%\end{claim}
%
%\begin{proof}
%Suppose not. As $\AgeK$ has the amalgamation of distinct $2n$-types over $\LangBot$ we have for some $\cD$ that $\qftp[\cD](\xx_0\alpha(\xx_0)) \neq \qftp[\cD](\xx_1\alpha(\xx_0))$. We then also have for all $\yy \in \cD\setminus \cB$ of length $n$ that $\qftp[\cD](\xx_0\yy) = \qftp(\xx_1\yy\cA)$ and in particular that $\qftp[\cD](\xx_0\alpha(\xx_0)) = \qftp(\xx_1\alpha(\xx_0))$.
%
%But $\qftp[\cD](\xx_0\xx_1) = \qftp[\cD](\alpha(\xx_0)\alpha(\xx_1))$. Therefore for all $\yy \in \cD$ of length $n$ we have $\qftp[\cD](\yy\alpha(\xx_0)) = \qftp(\yy \alpha(\xx_0))$ and in particular $\qftp[\cD](\xx_1\alpha(\xx_0)) = \qftp[\cD](\xx_1 \alpha(\xx_1))$. But then $\qftp[\cD](\xx_0\alpha(\xx_0)) = \qftp[\cD](\xx_1\alpha(\xx_1))$ getting us our contradiction.
%\end{proof}
We now show that for each $\cF_i \in V$ the automorphism given by $\name{h}^G$ is not the identity on those $\cF_i$.
\begin{claim}
\label{Main rigidity result: Claim 1}
Suppose $\cT$ pins $\AgeK$, $\cF \subseteq \MBot$ with $\cF \in V$ and $\cF$ has splitting property over $\MBot$ and $\cT$. Then $V[G] \models \name{h}^G\rest[{\cF}] \neq \id$
\end{claim}
\begin{proof}
Suppose to get a contradiction that $\name{h}^{G}\rest[{\cF}] = \id$. Let $s, t \in \cT^{\MBot}$ be distinct with $\name{h}^{G}(s) = t$.
Let $\cB \in G$ such that $\cA \subseteq \cB$ and $\cB \forces[\Fn<\MBot>[\AgeK](\lambda)] \name{h}(s) = t$. Let $\cA^- = \cA\rest[\LangBot]$.
Let $E$ be the collection of $\cE \in \AgeK[\MBot]$ such that
\begin{itemize}
\item $\cB \subseteq \cE$,
\item $s, t \in \cE$,
\item $(\exists \zz_{\cE} \in \cE \cap \cF)\,
\qftp[\cE](s\zz_{\cE}) \neq \qftp[\cE](t\zz_{\cE})\}$.
\end{itemize}
Then $E$ is dense below $\cB$ because $\cF$ has the splitting property with respect to $\MBot$ and $\cT$ and $s, t \in \cT^{\MBot}$. Therefore there must be some $\cD \in G \cap E$. But then $\qftp[\cG](s\zz_{\cD}) = \qftp[\cG](\name{h}^G(s)\name{h}^G(\zz_{\cD})) = \qftp[\cE](t\zz_{\cD})$, contradicting the fact that $\cD \in E$.
\end{proof}
By the claim, for each $i \in \kappa^+$, there are $\cB_i \in \AgeK$, $s_i \in \MBot$ and $t_i \in \MBot$ such that $\cA \subseteq \cB_i$, $s_i \in \cF_i$ and
\[
\cB_i \forces[\Fn<\MBot>[\AgeK](\lambda)] \name{h}(s_i) = t_i\text{ and }t_i \neq s_i.
\]
Further as the $\cF_i$ are disjoint outside $\tdcl[\MBot](\emptyset)$ and $\tdcl[\MBot](\emptyset)$ is fixed by any automorphism of $\MBot$ we must have $s_i \neq s_j$ for $i < j \in \kappa^+$.
But then by \cref{Delta-Systems Lemma} we can find an $I \subseteq \kappa^+$ with $|I| = \kappa^+$ such that for all $i < j \in I$, $\cA, \cB_i, \cB_j$ satisfy \cref{Amalgamation of distinct types property} (b). Further, as there are at most $\kappa$-many extensions of $\cA$ up to isomorphism we can assume $I$ was chosen so that \cref{Amalgamation of distinct types property} (c) holds as well. Let $\alpha_{i, j}$ be an isomorphism from $\cB_i$ to $\cB_j$ guaranteed by \cref{Amalgamation of distinct types property} (c). As each $\cB_i$ must have cardinality $\leq \kappa$, we can further assume without loss of generality that $\alpha_{i, j}(s_i) = s_j$ and $\alpha_{i, j}(t_i) = t_j$ for all $i, j \in I$.
For $i, j \in I$ let $\DBot_{i, j}$ be the substructure of $\MBot$ generated by $\UnderSet{\cB_i} \cup \UnderSet{\cB_j}$. Then \cref{Amalgamation of distinct types property} (d) and (e) hold. Therefore there must be a $\cD \in \AgeK[\MBot]$ such that $\cD\rest[\LangBot] = \DBot$, $\cB_i \subseteq \cD$, $\cB_j \subseteq \cD$ and $\qftp[\cD](s_is_j) \neq \qftp[\cD](t_it_j)$.
But then
\[
\cD \forces[\Fn<\MBot>[\AgeK](\lambda)] \name{h}\text{ is not an isomorphism }
\]
getting us our contradiction.
\end{proof}
We now give some examples of how to apply \cref{Main rigidity result}
\begin{example}
\label{Example orientation of 2-dimensional subspace}
Suppose $\Field$ is a finite field. Let $\LangBot_{\Field}$ and $\AgeKBot_{\Field}$ be the language and age in \cref{Example: Colored vector spaces} associated to finite vector spaces over $\Field$.
For $\kappa$ a cardinal let $\MBot_{\Field, \kappa}$ be the $\LangBot_{\Field}$-structure with age $\AgeKBot_{\Field}$ which is a $\kappa$-dimensional $\Field$-vector space. Note $\MBot_{\Field, \kappa} \models \ThFraisse_\w(\AgeKBot_{\Field})$ whenever $\kappa$ is infinite.
Let $\Lang = \LangBot_{\Field} \cup \{R\}$ where $R$ is a binary relation. Let $\AgeK$ be the age in $\Lang$ such that for all $\cA \in \AgeK$
\begin{itemize}
\item $\cA\rest[\LangBot_{\Field}] \in \AgeKBot_{\Field}$,
\item For every embedding $i\:\MBot_{\Field, 2} \to \cA\rest[\LangBot_{\Field}]$ there is a unique pair of independent elements $a, b \in \MBot_{\Field, 2}$ such that $\cA \models R(i(a), i(b))$.
\end{itemize}
We can think of $\AgeK$ as the age of vector spaces over $\Field$ where we have, for ever $2$-dimensional subspace, a distinguished orientation.
Note it is straightforward to check that both $\AgeKBot_{\Field}$ and $\AgeK$ are strong \Fraisse\ classes and that $\AgeK$ has the extended strong amalgamation property over $\LangBot_{\Field}$.
To show that $\AgeK$ has the amalgamation of distinct $2$-types property we now let $\cA, \cB, \cC, \DBot, \alpha, x_0, x_1, \cS$ be as in \cref{Amalgamation of distinct types property} where $\cS$ is the unique sort and $\LangBot = \LangBot_{\Field}$. Let $W_0$ be the vector space generated by $x_0\alpha(x_0)$ and $W_1$ be the vector space generated by $x_1\alpha(x_1)$. Note that $W_0 \cap \cB$ consists of the $1$-dimensional vector space containing $x_0$ and similarly $W_1 \cap \cB$ consists of the $1$-dimensional vector space containing $x_1$. Therefore $R$ does not holds of any tuple in $W_0 \cap \cB$ nor of any tuple in $W_1 \cap \cB$. Therefore there must be some $\cD \in \AgeK[\cD^-]$ where $\cD \models R(x_0\alpha(x_0))$ and $\cD \models \neg R(x_1\alpha(x_1))$. Such a $\cD$ witnesses the amalgamation of distinct $2$-types property.
Now let $(\cF_i)_{i \in \kappa^+}$ be a collection of infinite dimensional subspaces of $\MBot_{F, \kappa^+}$ such that for $i < j \in \kappa^+$, $\cF_i \cap \cF_j = \{v_0^{\MBot_{F, \kappa^+}}\}$. It is immediate that each $\cF_i$ has the splitting property for amalgamation property (with respect to $\cS$).
Therefore, by \cref{Main rigidity result}, if $G$ is a generic for $\Fn<\MBot_{\Field, \kappa^+}>[\AgeK](\w)$ over $V$, $\bigcup G$ is rigid in $V[G]$. In other words if we choose the orientations of $2$-dimensional subspaces generically the result is a rigid structure.
\end{example}
Note \cref{Example orientation of 2-dimensional subspace} was only done for $2$-dimensional subspaces. This is because the conditions which go into \cref{Main rigidity result} only talk about pairs of elements. The following example gives us a way, sometimes, to get around this limitation. In particular this will allow us to show if $R$ is a $n$-ary relation (with $n \geq 5$), $\AgeK_R$ is the age consisting of all finite structures in the language $\{R\}$, and $\NBot$ is the $\kappa$-sized structure in the empty language than any generic for $\Fn<\NBot>[\AgeK_R](\w)$ gives rise to a rigid structure. This generalizes Theorem 4 of \cite{Cohen} to the higher arity case.
\begin{example}
\label{Example of 5-ary directed hypergraphs}
Let $n \geq 5$ and $\LangBot_{s, n}$ and $\AgeKBot_{s, n}$ be as in \cref{Exaple of names for sequences}. Let $\Lang_{s,n} = \LangBot_{s, n} \cup \{R\}$ where $R$ is a relation of arity $U^n$. Let $\AgeK<s,n>$ be the age in $\Lang_{s, n}$ consisting of all elements whose restriction to $\LangBot_{s, n}$ is in $\AgeKBot_{s, n}$. Elements of $\AgeK_{s, n}$ can be thought of as arbitrary structures in the language $\{R\}$ along with names for each sequence of length $n$.
It is immediate that $\AgeK<s, n>$ is a strong \Fraisse\ class as $\AgeKBot_{s, n}$ is a strong \Fraisse\ class. It is also immediate that $\AgeK<s,n>$ has the extended strong amalgamation property over $\Lang_{s,n}$. Also, note that the sort $W$ pins $\AgeKBot_{s, n}$.
If $\MBot_{s, n}$ is a $\LangBot_{s, n}$-structure such that $\MBot_{s, n} \models \ThFraisse_\w(\AgeKBot_{s, n})$ then $\MBot$ is completely determined, up to isomorphism, by its cardinality. For $\kappa$ an infinite cardinal let $\MBot_{\kappa} \models \ThFraisse_\w(\AgeKBot_{s, n})$ be the unique model (up to isomorphism) where $|\MBot_{\kappa}| = \kappa$.
To show that $\AgeK<s, n>$ has the amalgamation of distinct $2$-types property we now let $\cA, \cB, \cC, \DBot, \alpha, x_0, x_1, \cS$ be as in \cref{Amalgamation of distinct types property} where $\cS = W$ and $\LangBot = \LangBot_{s, n}$. For $i \in [n]$ let $y_i = \pi_i(x_0)$ and $z_i = \pi_i(x_1)$. There must be some $\ell \in [n]$ such that $y_\ell \neq z_\ell$. As $n \geq 5$ we can assume without loss of generality that $\ell > 3$. There also must be some $k_0, k_1 \in [n]$ such that $\alpha(y_{k_0}), \alpha(z_{k_1}) \not \in \cB$ as otherwise $\alpha(x_{0}), \alpha(y_{1}) \in \cB$. This then implies that $y_{k_0}, z_{k_1} \not \in \cC$.
Define
\[
a = f(\alpha(y_{k_0}), \alpha(y_{k_1}), y_{k_0}, y_{k_1}, \dots, y_\ell, \dots, y_{n-1})
\]
and define
\[
b = f(\alpha(z_{k_0}), \alpha(z_{k_1}), z_{k_0}, z_{k_1}, \dots, z_\ell, \dots, z_{n-1}).
\]
Note that $a \neq b$ as $y_\ell \neq z_\ell$. Further $a \not \in \{x_0, x_1\}$ as $\alpha(y_{k_0}) \not \in \cB$ and $a \not \in\{\alpha(x_0),\alpha(x_1)\}$ as $y_{k_0} \not \in \cC$. We have similar conclusions about $b$.
Therefore there is some $\cD \in \AgeK<s, n>[\DBot]$ with $\cB \subseteq \cD$, $\cC \subseteq \cD$ and
\[
\cD \models R(\alpha(y_{k_0}), \alpha(y_{k_1}), y_{k_0}, y_{k_1}, \dots, y_\ell, \dots, y_{n-1}) \And\neg R(\alpha(z_{k_0}), \alpha(z_{k_1}), z_{k_0}, z_{k_1}, \dots, z_\ell, \dots, z_{n-1}).
\]
Therefore $\AgeK<s, n>$ has the amalgamation of distinct $2$-types property.
Note a more detailed analysis of which elements of $a$ and $b$ were in $\cA$ would have shown that $\AgeK<s,4>$ and $\AgeK<s, 3>$ had amalgamation of distinct $2$-types property. However such an analysis would not have added much to the example. In contrast, $\AgeK<s, 2>$ does not have amalgamation of distinct $2$-types property. To see this consider $x_0 = f(y_0, y_1)$ and $x_1 = f(y_0, z_1)$ with $y_1, z_1 \in \cA$. Then $\qftp[\cD](y_0, y_1, \alpha(y_0), \alpha(y_1)) = \qftp[\cD](y_0, y_1, \alpha(y_0), y_1)$ which is completely determined by $\qftp[\cD](y_0, \alpha(y_0))$, $\cB$ and $\cC$. But we similarly have $\qftp[\cD](y_0, z_1, \alpha(y_0), \alpha(z_1))= \qftp[\cD](y_0, z_1, \alpha(y_0), z_1)$ which is also completely determined by $\qftp[\cD](y_0, \alpha(y_0))$, $\cB$ and $\cC$.
Let $P$ be a partition of $U^{\MBot_{s, \kappa^+}}$ into countable infinite sets. For $p \in P$ let $\cF_p = \{f(x_0, \dots, x_{n-1}) \st x_0 \in p\And x_1, \dots, x_{n-1} \in U^{\MBot_{s, \kappa^+}}\}$. Now let $\cA \in \AgeK<s, n>[\MBot_{s\kappa^+}]$ and let $a_0, a_1 \in W^{\cA}$ be distinct. As $\cA$ is finite, given any $p \in P$ we can find an $b \in \cF_p$ such that $\pi_0(b) \not \in \cA$. We can therefore find a $\cB \in \AgeK<s, n>[\MBot_{s, \kappa^+}]$ with $\cA \subseteq \cB$ such that
\[
\cB \models R(\pi_0(b), \pi_1(a_0), \dots, \pi_{n-1}(a_{0})) \And \neg R(\pi_0(b), \pi_1(a_1), \dots, \pi_{n-1}(a_{1})).
\]
Therefore $\cF_p$ has the splitting property over $\AgeK<s, n>$ and $W$. Further, for $p_0, p_1 \in P$ distinct, we have $\cF_{p_0} \cap \cF_{p_1} = \emptyset$. Therefore, as $|P| = \kappa^+$, we can apply \cref{Main rigidity result} to get whenever $G$ is a generic for $\Fn<\MBot_{s, \kappa^+}>[\AgeK<s, n>](\w)$ over $V$, $\bigcup G$ is rigid in $V[G]$.
Finally, suppose $\LangBot_* = \{U\}$ where $U$ is a sort and $\Lang_* = \LangBot_* \cup \{R\}$ where $R$ is a relation of type $U^n$. If $\MBot$ is an $\LangBot_*$ structure then there is a unique expansion of $\MBot$ to a $\LangBot_{s, n}$-structure whose age is in $\AgeKBot_{s, n}$. Similarly if $\cM$ is a $\Lang_*$-structure then there is a unique expansion of $\cM$ to a $\Lang_{n, s}$-structure whose age is in $\AgeK<n, s>$. Therefore if $\AgeK_R$ is the age consisting of all $\Lang_*$-structures and $\NBot_{\kappa^*}$ is the unique $\LangBot_*$-structure of size $\kappa^+$, whenever $G$ is a generic for $\Fn<\NBot_{\kappa^+}>[\AgeK_R](\w)$ over $V$, $\bigcup G$ is rigid in $V[G]$.
\end{example}
In \cref{Example of 5-ary directed hypergraphs}, we showed that forcing with arbitrary structures in a relation of arity $\geq 5$ results in a rigid structure in the generic extension. This required us to have access to function symbols to name the tuples and so, even though the structures we are interested in, i.e. $\Lang_*$-structures, are in a relational language, we needed to be able to deal with function symbols to show the rigidity (at least if we wanted to use these techniques).
Note though that we dealt with arbitrary $\Lang_*$-structures and the techniques we applied do not immediately generalize to deal with symmetric $n$-hypergraphs. There are two natural approaches one might try to generalized these techniques. First one could have names for all tuples of length $n$ and then simply require that whenever the relation holds it must hold of all permutations of its arguments. This however would cause a problem as if $a, b$ were names of permutations of the same tuple there would be no way to distinguish them in the structure and amalgamation of distinct $2$-types property would fail.
The second natural approach one might try is to provide elements which ``name'' a set of size $n$, instead of sequences. However in this case there is no obvious way to recover the elements of the set from the name using just function symbols. This is because if $y$ is a name for $\{x_0, \dots, x_{n-1}\}$ and $x_0 = t(y)$ for some term, then any automorphism of the structure which fixes $y$ also fixes $x_0$. This therefore adds structure to the set of elements $\{x_0, \dots, x_{n-1}\}$.
We therefore have the following conjecture.
\begin{conjecture}
Suppose $n \geq 2$. Let $\LangBot$ be the empty language and let $\Lang_n = \{R\}$ where $R$ is a relation of arity $n$. Let $\AgeK<sym,n>$ consist of all those finite $\Lang_n$-structures where $R$ is symmetric, i.e. which satisfy
\[
\bigwedge_{\tau \in \Sym{[n]}} (\forall x_0, \dots, x_{n-1})\, R(x_0, \dots, x_{n-1}) \leftrightarrow R(x_{\tau(0)}, \dots, x_{\tau(n-1)}).
\]
Then for any uncountable cardinal $\kappa$, if $\MBot_{\kappa}$ is the unique $\LangBot$-structure of size $\kappa$ then whenever $G$ is a generic for $\Fn<\MBot_{\kappa}>[\AgeK<{sym, n}>](\w)$ over $V$, $\bigcup G$ is rigid in $V[G]$.
\end{conjecture}
%It is now worth giving an example of how this applies.
%
%\begin{example}
%Suppose $\LangBot = \{f\}$ where $f$ is a unary function. Let $\Lang\setminus \LangBot = \{\leq\}$ where $\leq$ is a binary relation. Let $\AgeK$ be the collection of structures where
%\begin{itemize}
%\item there is a unique element $r$ such that $f(r) = r$,
%
%\item for all $x$ there is an $n \in \w$ such that $f^n(x) = r$,
%
%\item $\leq$ is a partial order (and in particular $x \leq y$ and $y \leq x$ implies $x = y$).
%\item if $f^n(x) = r$ but $f^{n}(y) \neq r$ then $x \not \leq y$ and $y \not \leq x$,
%
%\item if $f^{n+1}(x) = f^{n+1}(y) = r$ and $f^n(x) \neq r$ and $f^n(y) \neq r$ then either $x \leq y$ or $y \leq x$,
%
%\item if $x \leq y$ then $f(x) \leq f(y)$.
%
%\item for all $x$, $\leq$ is a linear ordering on $f^{-1}(x) \setminus \{r\}$.
%
%\end{itemize}
%
%We can think of elements of $\AgeK\rest[\LangBot]$ as trees, where $f$ is the predecessor function. The ordering $\leq$ is such that it is a linear ordering on each level of the tree and where the predecessor function preserves the ordering.
%
%For $\DBot \in \AgeK\rest[\LangBot]$ let $h(x)$ be the least such that $f^{h(x)}(x) = r$ for $x \neq r$ and $h(r) = 0$. Suppose $\DBot \in \AgeK\rest[\LangBot]$. Let $m(\DBot)$ be the elements $x$ of $\DBot$ with $\DBot \models f^{-1}(x) = \emptyset$.
%
%Suppose $\cD \in \AgeK$. For $x, y \in m(\cD\rest[\LangBot])$ let $x \preceq^{\cD} y$ if either
%\begin{itemize}
%\item $h(x) = h(y)$ and $x \leq y$,
%
%\item $h(x) < h(y)$ and $x < f^{h(y) - h(x)}(y)$, or
%
%\item $h(x) > h(y)$ and $f^{h(x) - h(y)}(x) < y$
%\end{itemize}
%So $x \preceq^{\cD} y$ if, after applying $f$ a sufficient number of times to get $x$ and $y$ to be on the same level, the image of $x$ is less than the image of $y$ (note they can't be equal as $x, y \in m(\cD\rest[\LangBot])$).
%
%It is straightforward to check that for all $\cD \in \AgeK$, we have $\preceq^{\cD}$ is a linear order on $m(\cD\rest[\LangBot])$. It is also straightforward to check that for any linear ordering $\preceq$ on $m(\DBot)$ there is a unique $\cD \in \AgeK$ with $\cD\rest[\LangBot] = \DBot$ and $\preceq^{\cD}$ equal to $\preceq$.
%
%
%Now suppose we have $\cA \subseteq \cB \in \AgeK$ and we have distinct $x, y \in \cB$ such that $\qftp(x) = \qftp(y)$. We must have for some $n$ that $f(x) \neq r$ and $f(y) \neq r$ but $f^{n+1}(x) = f^{n+1}(y) = r$. Therefore we also must have either $x \leq y$ or $y \leq x$. Let $x^*, y^* \in m(\cB\rest[\LangBot])$ such that for some $k_x, k_y$, $f^{k_x}(x^*) = x$ and $f^{k_y}(y^*) = y$. Note $x \leq y$ if and only if $x^* \preceq^{\cB\rest[\LangBot]} y^*$.
%
%Now suppose $\cB^*$ is a copy $\cB$ which agrees with $\cB$ on $\cA$ and where $\UnderSet{\cB} \cap \UnderSet{\cB^*} = \UnderSet{\cA}$. Let $\alpha\:\cB \to \cB^*$ be an isomorphism which is the identity on $\cA$.
%
%Note that there is a unique amalgamation $\cC_0$ of $\cB\rest[\LangBot]$ and $\cB^*\rest[\LangBot]$ with $\UnderSet{\cC_0} = \UnderSet{\cB} \cup \UnderSet{\cB^*}$.
%
%Let $\sqsubseteq$ be an ordering on $m(\cC_0)$ such that
%\begin{itemize}
%\item[(a)] $\preceq^{\cB}$ is contained in $\sqsubseteq$ and $\preceq^{\cB^*}$ is contained in $\sqsubseteq$,
%
%\item[(b)] For $z_0, z_1 \in m(\cB\rest[\LangBot])$, $z_0 \sqsubseteq \alpha(z_1)$ if and only if $z_0 \preceq^{\cB} z_1$.
%\end{itemize}
%
%If we imagine $\sqsubseteq$ as a subset of $\Rationals$ then $\sqsubseteq$ is such that for all $z_0$, $|z_0- \alpha(z_0)|$ is smaller than $|z_0 - z_1|/2$ for any $z_0, z_1 \in \cB$.
%
%Now let $\sqsubseteq$ further be such that $x^* \sqsubseteq \alpha(x^*)$ and $\alpha(y^*) \sqsubseteq y^*$. Let $\cC \in \AgeK$ be such that $\preceq^{\cC}$ is $\sqsubseteq$. As $x, y \not \in \cA$ we have $\alpha(x) \neq x$ and $\alpha(y) \neq y$. Therefore we must have $x \leq \alpha(x)$ and $\alpha(y) \leq y$.
%
%Therefore, by \cref{Main rigidity result}, we have that if $\MBot$ is any uncountable $\LangBot$-structure satisfying the \Fraisse\ theory of $\AgeK$ and $G$ is a generic over $V$ for $\Fn<\MBot>[\AgeK](\w)$
%\[
%V[G] \models \bigcup G\text{ is rigid}.
%\]
%
%\end{example}
%%%% %%%% %%%% %%%% %%%% %%%% %%%% %%%%
\section{Non-Rigidity}
%%%% %%%% %%%% %%%% %%%% %%%% %%%% %%%%
%\TBD{To Mohammad: Nate and Mostafa would like it if there was an easy way to generalize this result so that the underlying structure $\MBot$ doesn't have to be trivial. We didn't see an immediate way to do this, but don't necessarily think it will be hard. We also don't think it is strictly necessary}
While \cref{Main rigidity result} gives us several examples of ages such that any generic over an uncountable structure is rigid, there are also many examples where the generics have large automorphism groups. The following proposition gives us a collection of such ages.
\begin{definition}
Suppose $\Lang$ is a relational language and let $\Phi$ be the collection of atomic formulas. If $\cM$ is an $\Lang$-structure and $a, b \in \cM$ say $\cM \models a \equiv b$ if
\[
\cM \models (\forall \cc)\bigwedge_{\psi(x,\yy) \in \Phi} \psi(a, \cc) \leftrightarrow \psi(b, \cc).
\]
\end{definition}
Note the following is immediate.
\begin{lemma}
Suppose $\Lang$ is a relational language and $\cM$ is an $\Lang$-structure. Suppose $A \subseteq\cM$ is an $\equiv^{\cM}$-equivalence class. Then for any bijection $f\:A \to A$ the map $f \cup \id_{\cM \setminus A}$ is an automorphism of $\cM$.
\end{lemma}
\begin{corollary}
\label{Conditions when S_k is in the automorphism group}
Suppose $\Lang$ is a relational language and $\cM$ is an $\Lang$-structure. If $\cM$ has an $\equiv^{\cM}$-equivalence class of size $\kappa$ then there is a subgroup of $\Aut(\cM)$ isomorphic to $\Perm(\kappa)$.
\end{corollary}
\begin{proposition}
Suppose
\begin{itemize}
\item $\Lang$ is a countable relational language,
\item $\AgeK$ is a strong \Fraisse\ class which is $\w$-bounded and which has at most $\w$-many types up to isomorphism.
\item $\cM$ is the \Fraisse\ limit of $\AgeK$ and $\equiv^{\cM}$ is not equality.
\item for $\kappa$ an infinite cardinal, $\cN_\kappa$ is the unique structure on $\kappa$ in the empty language.
\end{itemize}
Then if $G$ is any generic over $V$ for $\Fn<\cN_\kappa>[\AgeK](\w)$ we have $\Perm(\kappa)$ is a subgroup of $\Aut(\bigcup G)$.
\end{proposition}
\begin{proof}
By \cref{Fraisse limits exist and trivial group theoretic dcl} and the fact that $\AgeK$ is a strong \Fraisse\ class, $\cM$ has trivial group theoretic definable closure. Now suppose to get a contradiction $\equiv^{\cM}$ had a finite equivalence class $\{a_i\}_{i \in [n]}$ where $n > 1$. Because being in the same $\equiv^{\cM}$-equivalence class is preserved under automorphisms, if $g \in \Aut(\cM)$ and $g(a_0, \dots, a_{n-2}) = (a_0, \dots, a_{n-2})$ we must also have $g(a_{n-1}) = a_{n-1}$. However this would contradict the fact that $\cM$ has trivial group theoretic definable closure. Therefore every $\equiv^{\cM}$-equivalence class must either have size $1$ or be infinite. Furthermore, as $\equiv^{\cM}$ is not equality it must have an infinite equivalence class.
But by \cref{Forcing satisfies Fraisse theory} we know that $\bigcup \cG$ satisfies the \Fraisse\ theory of $\AgeK$ and so is potentially isomorphic to $\cM$. Therefore there must be some $\cA \in \AgeK$ and $a \in \cA$ such that
\[
\cA \forces[\Fn<\cN_\kappa>[\AgeK](\w)] a\text{ is in an infinite }\equiv\text{-equivalence class}.
\]
But then for all $\alpha \in \kappa$, the collection of $\cB$ such that $\cA \subseteq \cB$ and
\[
\cB \forces[\Fn<\cN_\kappa>[\AgeK](\w)] (\exists \beta > \alpha)\, a \equiv \beta.
\]
is dense. Therefore in $\bigcup G$ we have that $a$ is in an $\equiv$-equivalence class of size $\kappa$.
But, by \cref{Conditions when S_k is in the automorphism group}, we then have that $\Perm(\kappa)$ is a subgroup of $\Aut(\bigcup G)$ as desired.
\end{proof}
%We will end with a conjecture on when generics are rigid, at least in the case of relational languages.
%
%\begin{conjecture}
%Suppose
%\begin{itemize}
%\item $\Lang$ is a countable relational language,
%
%\item $\AgeK$ is a strong \Fraisse\ class which is $\w$-bounded and which has at most $\w$-many types up to isomorphism.
%
%\item $\cM$ is the \Fraisse\ limit.
%\end{itemize}
%
%Suppose $\kappa$ is a cardinal in $V$, $\cN_\kappa$ is the unique structure in the empty language on $\kappa$ and $G$ is generic for $\Fn<\cN_\kappa>[\AgeK](\w)$. Then precisely one of the following holds
%\begin{itemize}
%\item[(a)] $\bigcup G$ is rigid,
%
%\item[(b)] $\equiv^{\cM}$ is not equality on $\cM$,
%\end{itemize}
%Further in (b) $\equiv^{\cM}$ has an infinite equivalence class and $\Perm(\kappa)$ is a subgroup of $\Aut(\bigcup G)$.
%\end{conjecture}
\bibliographystyle{amsnomr}
\bibliography{bibliography}
\end{document}
%MACROS:
%Local
%\renewcommand{\labelitemi}{$\circ$}
\renewcommand{\labelitemii}{$\circ$}
%\renewcommand{\labelitemiii}{$\circ$}
%\renewcommand{\labelitemiv}{$\circ$}
\def\xx{{\EM{\mbf{x}}}}
\def\yy{{\EM{\mbf{y}}}}
\def\zz{{\EM{\mbf{z}}}}
\def\Aut{{\mathrm{{Aut}}}}
\def\Lempty{\EM{\mc{L}}}
\def\Lzero#1{\EM{\Lempty_{#1, 0}}}
\def\Lomega#1{\EM{\Lempty_{#1, \w}}}
\def\Linfty#1{\EM{\Lempty_{#1, \infty}}}
\def\Lkappalambda#1#2{{\EM{\Lempty_{#1, #2}}}}
\def\Lzz{\Lzero{0}}
\def\Lkw{\Lkappalambda{\kappa}{\w}}
\def\Lkg{\Lkappalambda{\kappa}{\gamma}}
\def\Lww{\Lomega{\w}}
\def\Lczz{\Lzz^c}
\def\Lcww{\Lww^c}
\def\abLcww{\EM{\ol{\mc{L}}^c_{\w, \w}}}
\def\Lclasww{\Lww^*}
\def\Lwow{\Lomega{\w_1}}
\def\Liw{\Lomega{\infty}}
\def\Lii{{\EM{\Lkappalambda{\infty}{\infty}}}}
\def\Lik{{\EM{\Lkappalambda{\infty}{\kappa}}}}
\def\Lgk{{\EM{\Lkappalambda{\gamma}{\kappa}}}}
\def\Lkappa#1{{\EM{\Lempty_{#1}}}}
\def\Lz{\EM{\Lkappa{0}}}
\def\Lw{\EM{\Lkappa{\w}}}
\def\Lk{\EM{\Lkappa{\kappa}}}
\def\Lg{\EM{\Lkappa{\gamma}}}
\def\Li{\EM{\Lkappa{\infty}}}
\def\Liwdu{\Liw^{\bigvee}}
\def\ZeroJump{\EM{\mbf{0'}}}
\newcommand{\Rest}{{\restriction}}
\newcommand\rest[1][]{\EM{\!\restriction_{#1}}}
\newcommand\pars{\,\cdot\,}
\def\TC{\mathcal{TC}}
\def\FormApp{\text{FA}}
\newcommand{\defn}[1]{{\bf{#1}}}
\newcommand{\defas}{\EM{:=}}
\newcommand{\dashedarrow}[1]
{%
\EM{\stackrel{#1}{\dashrightarrow}}
}
%Specific to this document
\def\ZF{\text{ZF}}
\def\ZFC{\text{ZFC}}
\def\Set{\text{SET}}
\def\Powerset{\mathcal{P}}
\newcommand{\Lowenheim}{\text{L\"{o}wenheim}}
\newcommand{\LS}{\text{\Lowenheim-Skolem}}
\newcommand{\Lindstrom}{\text{Lindstr\"{o}m}}
\newcommand{\Los}{\text{\L{}o\'{s}}}
\def\Models{\text{Mod}}
\DeclareMathOperator{\colim}{colim}
%\def\limit{\text{Lim}}
\def\ProbMeas{\EM{\mathcal{P}_1}}
\def\extent#1{\EM{[|#1|]}}
\DeclareMathOperator{\ORD}{ORD}
%\DeclareMathOperator{\sup}{sup}
%\DeclareMathOperator{\inf}{inf}
\DeclareMathOperator{\dom}{dom}
\DeclareMathOperator{\range}{ran}
\DeclareMathOperator{\obj}{obj}
\DeclareMathOperator{\Obj}{Obj}
\DeclareMathOperator{\morph}{mor}
\DeclareMathOperator{\Morph}{Mor}
\DeclareMathOperator{\cod}{cod}
\DeclareMathOperator{\Graph}{graph}
\DeclareMathOperator{\codom}{codom}
%\DeclareMathOperator{\graph}{graph}
\DeclareMathOperator{\rank}{rank}
\DeclareMathOperator{\qrank}{qrank}
\DeclareMathOperator{\aqrank}{aqrank}
%To Clean Up
\newcommand{\Bool}{\EM{\mathfrak{B}}}
\DeclareDocumentCommand{\forces}{d[]}
{
\IfNoValueTF{#1}
{
\EM{\Vdash}
}
{
\EM{\Vdash_{#1}}
}
}
\DeclareMathOperator{\dist}{\EM{d}}
\DeclareMathOperator{\Th}{Th}
\DeclareMathOperator{\Str}{Str}
\DeclareDocumentCommand{\dual}{d()}
{
\IfNoValueTF{#1}
{
\EM{\hat{\ }}
}
{
\EM{\hat{#1}}
}
}
\DeclareDocumentCommand{\SizeAS}{d()}
{
\IfNoValueTF{#1}
{
\EM{|\cdot|}
}
{
\EM{|#1|}
}
}
\DeclareDocumentCommand{\compcK}{d[]}
{
\IfNoValueTF{#1}
{
\EM{\mathbb{K}}
}
{
\EM{\mathbb{K}[#1]}
}
}
\DeclareDocumentCommand{\AgeK}{d<> d[] d()}
{
\IfNoValueTF{#1}
{
\IfNoValueTF{#2}
{
\IfNoValueTF{#3}
{
\EM{\mbf{K}}%
}
{
\EM{\mbf{K}^{#3}}%
}
}
{
\IfNoValueTF{#3}
{
\EM{\mbf{K}[#2]}%
}
{
\EM{\mbf{K}^{#3}[#2]}%
}
}
}
{
\IfNoValueTF{#2}
{
\IfNoValueTF{#3}
{
\EM{\mbf{K}_{#1}}%
}
{
\EM{\mbf{K}^{#3}_{#1}}%
}
}
{
\IfNoValueTF{#3}
{
\EM{\mbf{K}_{#1}[#2]}%
}
{
\EM{\mbf{K}_{#1}^{#3}[#2]}%
}
}
}
}
\DeclareDocumentCommand{\AgeKBot}{}
{
\AgeK(-)
}
\DeclareDocumentCommand{\LangBot}{}
{
\Lang^-%
}
\DeclareDocumentCommand{\MBot}{}
{
\cM^-%
}
\DeclareDocumentCommand{\NBot}{}
{
\cN^-%
}
\DeclareDocumentCommand{\DBot}{}
{
\cD^-%
}
\DeclareDocumentCommand{\Rel}{d[]}
{
\IfNoValueTF{#1}
{
\EM{\mathcal{R}}
}
{
\EM{\mathcal{R}_{#1}}
}
}
\DeclareDocumentCommand{\Func}{d[]}
{
\IfNoValueTF{#1}
{
\EM{\mathcal{F}}
}
{
\EM{\mathcal{F}_{#1}}
}
}
\DeclareDocumentCommand{\Sort}{d[]}
{
\IfNoValueTF{#1}
{
\EM{\mathcal{S}}
}
{
\EM{\mathcal{S}_{#1}}
}
}
\DeclareDocumentCommand{\arity}{}{\textbf{ar}}
\DeclareDocumentCommand{\ar}{d[]}
{
\IfNoValueTF{#1}
{
\EM{\arity}
}
{
\EM{\arity_{#1}}
}
}
\DeclareDocumentCommand{\Fn}{d<> d[] d()}
{
\IfNoValueTF{#3}
{
\EM{\textrm{Fn}(#1, #2)}
}
{
\EM{\textrm{Fn}(#1, #2, #3)}
}
}
\DeclareDocumentCommand{\UnderSet}{m}
{
\EM{\mbf{set}(#1)}
}
\DeclareMathOperator{\Tree}{Tr}
\DeclareMathOperator{\Sym}{Sym}
\DeclareDocumentCommand{\Perm}{d()}
{
\EM{\Sym(#1)}
}
\DeclareDocumentCommand{\name}{m}
{
\EM{\dot{#1}}
}
\newcommand{\compleq}{\EM{\mathbb{\leq}}}
\newcommand{\fraisse}{\textrm{Fra\"iss\'e}}
\newcommand{\Fraisse}{\textrm{Fra\"iss\'e}}
\newcommand{\FRAISSE}{\textrm{FRA\"ISS\'E}}
\newcommand{\ThFraisse}{\text{Fr}}
\newcommand{\cl}{\EM{\textrm{cl}}}
\newcommand{\cofinal}{\EM{\textrm{cf}}}
\newcommand{\tc}{\EM{\textrm{tc}}}
\newcommand{\len}{\EM{\textrm{len}}}
\newcommand{\Field}{\mbb{F}}
\DeclareDocumentCommand{\tdcl}{d[] d()}
{
\IfNoValueTF{#1}
{
\EM{\textbf{tcl}(#2)}
}
{
\EM{\textbf{tcl}_{#1}(#2)}
}
}
\DeclareDocumentCommand{\gdcl}{d[] d()}
{
\IfNoValueTF{#1}
{
\EM{\textrm{gcl}(#2)}
}
{
\EM{\textrm{gcl}_{#1}(#2)}
}
}
\DeclareDocumentCommand{\qftp}{d[] d()}
{
\IfNoValueTF{#1}
{
\EM{\quantfreetp(#2)}
}
{
\EM{\quantfreetp_{#1}(#2)}
}
}
\DeclareMathOperator{\quantfreetp}{qtp}
\DeclareDocumentCommand{\Closure}{d<> d()}
{
\IfNoValueTF{#1}
{
\EM{\textrm{cl}(#2)}
}
{
\EM{\textrm{cl}_{#1}(#2)}
}
}
\DeclareDocumentCommand{\ClosureMap}{d[] d()}
{
\EM{\textrm{clMap}(#1, #2)}
}
\DeclareDocumentCommand{\clsim}{}
{
\EM{\sim}
}
\DeclareDocumentCommand{\HP}{}
{
\textrm{(HP)}
}
\DeclareDocumentCommand{\CHP}{d()}
{
\IfNoValueTF{#1}
{
\textrm{(CHP)}
}
{
\textrm{(\EM{#1}-CHP)}
}
}
\DeclareDocumentCommand{\JEP}{}
{
\textrm{(JEP)}
}
\DeclareDocumentCommand{\CJEP}{d()}
{
\IfNoValueTF{#1}
{
\textrm{(CJEP)}
}
{
\textrm{(\EM{#1}-CJEP)}
}
}
\DeclareDocumentCommand{\AP}{}
{
\textrm{(AP)}
}
\DeclareDocumentCommand{\SAP}{}
{
\textrm{(SAP)}
}
\makeatletter
\newcommand{\dotminus}{\mathbin{\text{\@dotminus}}}
\newcommand{\@dotminus}{%
\ooalign{\hidewidth\raise1ex\hbox{.}\hidewidth\cr$\m@th-$\cr}%
}
\DeclareMathOperator{\Lang}{\mathscr{L}}
\DeclareMathOperator{\id}{id}
\DeclareMathOperator{\den}{den}
\DeclareMathOperator{\incl}{in}
\DeclareMathOperator{\tb}{tb}
\DeclareMathOperator{\CompactIndex}{CI}
\DeclareDocumentCommand{\quadiff}{}
{
\EM{\quad\text{ if and only if }\quad}
}
\DeclareDocumentCommand{\Realize}{m}
{
\EM{\|#1\|}
}
\DeclareMathOperator*{\supleq}{sup^{\leq}}
\DeclareMathOperator*{\infleq}{inf^{\leq}}
\DeclareMathOperator{\RegFr}{Rf}
\DeclareMathOperator{\Scott}{sr}
\DeclareMathOperator{\Ext}{Ext}
\DeclareMathOperator{\FinSpec}{FinSpec}
\DeclareMathOperator{\CompSpec}{CompSpec}
\def\aa{{\EM{\mbf{a}}}}
\def\bb{{\EM{\mbf{b}}}}
\def\cc{{\EM{\mbf{c}}}}
\def\dd{{\EM{\mbf{d}}}}
\def\ee{{\EM{\mbf{e}}}}
\def\ff{{\EM{\mbf{f}}}}
\def\qq{{\EM{\mbf{q}}}}
\def\ss{{\EM{\mbf{s}}}}
\def\gg{{\EM{\mbf{g}}}}
\def\cA{{\EM{\mathcal{A}}}}
\def\cB{{\EM{\mathcal{B}}}}
\def\cC{{\EM{\mathcal{C}}}}
\def\cD{{\EM{\mathcal{D}}}}
\def\cE{{\EM{\mathcal{E}}}}
\def\cF{{\EM{\mathcal{F}}}}
\def\cG{{\EM{\mathcal{G}}}}
\def\cK{{\EM{\mathcal{K}}}}
\def\cM{{\EM{\mathcal{M}}}}
\def\cN{{\EM{\mathcal{N}}}}
\def\cO{{\EM{\mathcal{O}}}}
\def\cP{{\EM{\mathcal{P}}}}
\def\cQ{{\EM{\mathcal{Q}}}}
\def\cR{{\EM{\mathcal{R}}}}
\def\cS{{\EM{\mathcal{S}}}}
\def\cT{{\EM{\mathcal{T}}}}
\def\cW{{\EM{\mathcal{W}}}}
\def\cX{{\EM{\mathcal{X}}}}
\def\cY{{\EM{\mathcal{Y}}}}
%- - - -
\def\w{\EM{\omega}}
\def\Naturals{{\EM{{\mbb{N}}}}}
\def\Nats{\Naturals}
\def\Integers{{\EM{{\mbb{Z}}}}}
\def\Rationals{{\EM{{\mbb{Q}}}}}
\def\PosRationals{\EM{\Rationals^{+}}}
\def\NNRationals{\EM{\Rationals^{\geq 0}}}
\def\Reals{\EM{\mbb{R}}}
\def\RealsQuantale{\EM{\mbf{[0,\infty]}}}
\def\^{\EM{{}^{\And}}}
\def\Or{\EM{\vee}}
\def\And{\EM{\wedge}}
\def\<{\EM{\langle}}
\def\>{\EM{\rangle}}
\def\qu#1{``#1''}
\def\indent{\hspace*{2em}}
\def\newp{\newline\indent}
\def\nl{\newline}
\def\nnll{\newline\newline}
\def\EM#1{\ensuremath{#1}}
\def\mbb#1{\EM{\mathbb{#1}}}
\def\mbf#1{\EM{\mathop{\pmb{#1}}}}
\def\mc#1{\EM{\mathcal{#1}}}
\def\mf#1{\EM{\mathfrak{#1}}}
%\ol:Overline \tol:Text Overline
\def\ol#1{\EM{\overline{#1}}}
\def\tol#1{\EM{\overline{\text{#1}}}}
\def\ul#1{\underline{#1}}
\def\st{\,:\,}
\def\:{\colon}
%For uniform monic/epic arrows
\def\monicrightarrow{\rightarrowtail}
\def\mrightarrow{\rightarrowtail}
\def\monicleftarrow{\leftarrowtail}
\def\mleftarrow{\leftarrowtail}
%
\def\epirightarrow{\twoheadrightarrow}
\def\erightarrow{\twoheadrightarrow}
\def\epileftarrow{\twoheadleftarrow}
\def\eleftarrow{\twoheadleftarrow}
\providecommand{\dotdiv}{% Don't redefine it if available
\mathbin{% We want a binary operation
\vphantom{+}% The same height as a plus or minus
\text{% Change size in sub/superscripts
\mathsurround=0pt % To be on the safe side
\ooalign{% Superimpose the two symbols
\noalign{\kern-.35ex}% but the dot is raised a bit
\hidewidth$\smash{\cdot}$\hidewidth\cr % Dot
\noalign{\kern.35ex}% Backup for vertical alignment
$-$\cr % Minus
}%
}%
}%
}
%- - - -
\DeclareDocumentCommand{\RightJustify}{m}{\hspace*{\fill}\mbox{#1}\penalty-9999\relax}
%
%COMMENTING
%NEEDS:
% PACKAGES: xparse, color
% DOCUMENTS: macros.tex
%These are the margin counter functions
\newcounter{margincounter}
\DeclareDocumentCommand{\displaycounter}{}
{{\arabic{margincounter}}}
\DeclareDocumentCommand{\incdisplaycounter}{}
{{\stepcounter{margincounter}\arabic{margincounter}}}
%DECLARECOMMENT
%This is a general template for declaring comment like commands.
%It displays both:
% \XXX: Inline command.
% \fXXX: Margin command
% use \fXXX versions of these macros to put additional explanation into a footnote.
% The idea is that we don't want to interrupt the flow of the paper or make it
% impossible to read because there are a bunch of comments.
%
\DeclareDocumentCommand{\DeclareComment}{m m m o d()}{%
%
% Variables
% #1: This is the commenting variables name.
% Note this should *not* have a backslash in front of it.
% #2: This is the commenting color.
% #3: This is the inline version symbol. It will be displayed as (#3)
% #4: This is the fTBD version symbol to occur after the counter. It will be displayed as (counter#4)
% #5: This is text that should start each comment
%
%This does the following:
% It defines the \#1 commenting command which appears in the text
% It defines the \f#1 commenting command which appears in the margin (and has a counter)
% It defines the \Hide#1 command which turns off all comment of this type.
% It does this by redefining them to do nothing.
% It defines the \Show#1 command which turns on all comment of this type.
% It does this by redefining them to the original values.
%
%Note it does this by first defining Hide/Show and then calling show.
%
%
%TODO:
%Note this still can't take paragraphs. For some reason adding +in front of the m when we are declaring the commands (below) doesn't make them long.
%
%Make it so that we can hide the environment
%Declare the hide version
\expandafter\DeclareDocumentCommand\csname Hide#1\endcsname {}
{%
%Clear the inline version
\expandafter\DeclareDocumentCommand\csname #1\endcsname {+m} {}
%Clear the margin version
\expandafter\DeclareDocumentCommand\csname f#1\endcsname {+m} {}
%Clear the environment version and make it a comment
%Note right now this just turns it off but doesn't hide the text
\expandafter\DeclareDocumentEnvironment{e#1} {} {} {}
}
\expandafter\DeclareDocumentCommand\csname Show#1\endcsname {}
{
%Show the inline version
\expandafter\DeclareDocumentCommand\csname #1\endcsname {+m}
{%
%Set color
\textcolor{#2}
%Text
{
%The maker in front of the text
{\tiny \bf (#3)}
% %Starting text (if there is any)
\IfValueT{#5}
{%
#5
}
%The text
####1
}
}
%Show the margin version
\expandafter\DeclareDocumentCommand\csname f#1\endcsname {+m}
{%
%Check if there is an extra character to add
\IfValueTF{#4}
{%Extra character
%Set color
\textcolor{#2}
%Set the counter in the text
{\text{$\,^{(\incdisplaycounter{#4})}$}}
%Set the counter in the margin
\marginpar{\tiny\textcolor{#2}{
%Counter
{\text{\tiny $(\displaycounter{#4})$}}
%Starting text (if there is any)
\text{\IfValueT{#5}{#5}
%Text
####1}}}
}
{%No extra character
%Set color
\textcolor{#2}
%Set the counter in the text
{$\,^{(\incdisplaycounter)}$}
%Set the counter in the margin
\marginpar{\tiny\textcolor{#2}{
%Counter
{\tiny $(\displaycounter)$}
%Starting text (if there is any)
\text{\IfValueT{#5}{#5}
%Text
####1}}}
}
}
%Declare the environment version
\expandafter\DeclareDocumentEnvironment{e#1} {}
{%BEGIN
\textcolor{#2}
%begin the color group
\bgroup
%Starting text (if there is any)
\IfValueT{#5}
{%
#5
}
}
{%END
\egroup
}
}
%Call Show to define everything
\csname Show#1\endcsname
}
%COMMENTING COMMANDS
%These are the commands
%Colors
\definecolor{NAColor}{rgb}{1.0,0.0,0.0}
\definecolor{ProblemColor}{rgb}{0.7,0.1,0.7}
\definecolor{TBDColor}{rgb}{0.0,0.0,0.8}
\definecolor{MathColor}{rgb}{0.0,0.4,0.1}
\definecolor{NateColor}{rgb}{0.0,0.5,1.0}
\definecolor{MostafaColor}{rgb}{1.0,0.0,1.0}
\definecolor{RefColor}{rgb}{1.0,0.0,1.0}
\definecolor{LaterColor}{rgb}{1.0,0.0,1.0}
%% NA: needs attention (rough writing whose correctness needs to be verified)
%% TBD: instructions for how to fix a gap ("Describe the propagation by ...")
%% PROBLEM: bug or missing crucial bit
%% CITE: this is references to place that the current result is generalizing
\DeclareComment{NA}{NAColor}{\EM{\star}}
\DeclareComment{TBD}{TBDColor}{\EM{!}}
\DeclareComment{PROBLEM}{ProblemColor}{\EM{!!}}(\bf)
\DeclareComment{MATH}{MathColor}{\EM{\#}}
\DeclareComment{NATE}{NateColor}{\EM{\square}}(N:)
\DeclareComment{MOSTAFA}{MostafaColor}{\EM{\square}}(M:)
\DeclareComment{REF}{RefColor}{\EM{(r)}}(Ref:)
\DeclareComment{LATER}{LaterColor}{\EM{(L)}}(Later:)
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%Hide these commands
%\HideNA
%\HideTBD
%\HidePROBLEM
%\HideLATER
%\HideMATH
%\HideNATE
%\HideLEAH
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%THEOREMS
%NEEDS:
% PACKAGES: amsthm, setspace, xparse, l3tl
% DOCUMENTS: programming.tex
%- - - - - - - -
%DECLARE COUNTER
% This creates a counter if one doesn't exist and does nothing if one does.
% This is to avoid errors when you don't know if a counter already exists.
%
\DeclareDocumentCommand{\DeclareCounter}{m}%
% Variables
% #1: This is the name of the counter to declare.
%
{%
% Does the counter already exist?
% If yes then it does nothing, if no it creates it.
\ifcsname c@#1\endcsname%
%
\else%
\newcounter{#1}%
\fi%
}
%PROOFS
%This is the QED symbol that proof of uses.
% By default it is the standard qed
\DeclareDocumentCommand{\MyQED}{}{\qed}
\DeclareDocumentEnvironment{proof}{d() D[]{\MyQED} d<>}
% Variables
% #1: This is the text to add after the word proof but before the period.
% If it exists then a space is added after proof. So [of Thm 1] would read
% Proof of Thm1
% and not
% Proofof Thm 1
% #2: This is the QED symbol to use.
% #3: This is the subscript to put on the QED symbol. If it is empty nothing is put on the symbol.
%
%TODO:
%Sometimes extra space is added between "Proof." and the text and I am unsure why.
% I think it has something to do with the \RightJustify
% Currently I am not putting the brackets on the begin environment on a new line and using negative space and that seems to work. However I still don't know why the problem is happening.
%
{%BEGIN
%This blank line is to make sure the proof starts on a new line.
%
\noindent\IfNoValueTF{#1}
{\emph{Proof.\!\!}}
{\emph{Proof\ #1.\ }}
}
{%END
%Check and see if there a subscript to add.
\IfValueTF{#3}
{%
%Note we need \ExplSyntaxOff to get the subscript
\ExplSyntaxOff
\RightJustify{\EM{#2_{#3}}}
\vspace{1ex}
}
{
\RightJustify{\EM{#2}}
\vspace{1ex}
}
}
%PROOF WITH LABELS
%This is a collection of proof environments which automatically adds the number of the result to the proof box.
% We have multiple "proof types" so that when we have nested proof environments the end{proof} references the right theorem. i.e. in nested proof environments each proof should be of a different "type"
%
% This does it in the following way:
% ProofLabelcOUntEr: This is a counter which keeps track of all theorem-like environments which have been called.
%
% ProofLabel: This does the following:
% -Adds one to ProofLabelcOUntEr
% -Creates a label from ProofLabelcOUntEr.
%
% In particular the begin{theorem} environment is written to always call this (unless it is a *ed version)
%
% ProofRef, ProofCref: This takes an argument which is the "proof type" and returns reference to the most recent theorem of the of that type.
%
% Proof: This sets the current "proof type" reference to the most recent theorem and then calls the proof environment.
%
%
%This is the counter which keeps track of what the most recent theorem was.
\DeclareCounter{ProofLabelcOUntEr}
%This creates the label.
%
%This should be called in the theorem environment
\DeclareDocumentCommand{\ProofLabel}{}{%
%
% Increment the counter
\addtocounter{ProofLabelcOUntEr}{1}
%Create the label
\label{cUrrEntProoflAbEl\arabic{ProofLabelcOUntEr}}
}
%This references the most current proof environment. It uses \ref
\DeclareDocumentCommand{\ProofRef}{D<>{1}}
% Variables
% #1: This is the "proof type"
{%
\ref{cUrrEntProoflAbEl\arabic{ProofcOUntEr#1}}
}
%This references the most current proof environment. It uses \cref
\DeclareDocumentCommand{\ProofCref}{D<>{1}}
% Variables
% #1: This is the "proof type"
{%
\cref{cUrrEntProoflAbEl\arabic{ProofcOUntEr#1}}
}
\DeclareDocumentEnvironment{Proof}{d() D[]{\MyQED} D<>{1}}
% Variables
% #1: This is the text to add after the word proof but before the period.
% If it exists then a space is added after proof. So [of Thm 1] would read
% Proof of Thm1
% and not
% Proofof Thm1
% #2: This is the QED symbol to use.
% #3: This is the type of proof
% These should be different for different nested proof environments.
%
% Note this always assumes the proof is of the most recent theorem which was written.
%
{%BEGIN
% This creates a counter of the appropriate "proof type"
% If this already exists then nothing is done.
\DeclareCounter{ProofcOUntEr#3}%
%
% This sets the counter of this proof type to the most recently called theorem.
\setcounter{ProofcOUntEr#3}{\value{ProofLabelcOUntEr}}%
%
%
% Call the begin proof environment
\begin{proof}(#1)[#2]<\ProofRef<#3>>%
}
{%END
\end{proof}
}
%THEOREM LIKE OBJECTS
\ExplSyntaxOn
%This is the sectioning depth at which theorems (and related objects) get reset to 1
\def\TheoremDepth{section}
%DECLARETHEOREM
%This is a general template for declaring theorem like environments.
%
\DeclareDocumentCommand{\DeclareTheorem}{m o m o}{%
%
% Variables
% #1: Environment's name
% #2: Sibling counter. Note this should have vArIAblE after the name when used.
% #3: Label
% #4: Parent counter
%
% This declares a theorem style environment \begin{#1}\end{#1} which is numbered and \begin{#1*}\end{#1*} which isn't.
%
% This first calls \newtheorem on #1vArIAblE and #vArIAblE* and then creates the environments using them.
%
%If 5 is empty then also call ProofLabel (to set a proof label). Otherwise it calls SubproofLabel (to set a subproof lable)
%Declare #1vArIAblE newtheorems
\IfNoValueTF{#4}
{%
\IfNoValueTF{#2}
{%
\newtheorem{#1vArIAblE}{#3}
}
{%
\newtheorem{#1vArIAblE}[#2vArIAblE]{#3}
}
}
{%
\newtheorem{#1vArIAblE}{#3}[#4]%
}
\newtheorem*{#1vArIAblE*}{#3}
%Declare the #1 environment.
\DeclareDocumentEnvironment{#1}{o o}
% #1: This is an argument to be passed to \begin{#1vArIAblE}
% #2: This is the spacing.
{%BEGIN
%Add spacing if there is a second argument
\IfValueT{##2}%
{
\begin{spacing}{##2}
}
%Begin the theorem environment
%See if there is anything to pass to it.
\IfValueTF{##1}
{
\begin{#1vArIAblE}[##1]
}
{
\begin{#1vArIAblE}
}
%
% Create a label for this theorem from ProofLabelcOUntEr.
\ProofLabel
}
{%END
%End spacing if there is a second argument
\IfValueT{##2}%
{
\end{spacing}{##2}
}
%End the theorem environment
\end{#1vArIAblE}
}
%Declare the #1 environment.
\DeclareDocumentEnvironment{#1*}{o o}
% #1: This is an argument to be passed to \begin{#1vArIAblE}
% #2: This is the spacing.
{%BEGIN
%Add spacing if there is a second argument
\IfValueT{##2}%
{
\begin{spacing}{##2}
}
%Begin the theorem environment
%See if there is anything to pass to it.
\IfValueTF{##1}
{
\begin{#1vArIAblE*}[##1]
}
{
\begin{#1vArIAblE*}
}
}
{%END
%End spacing if there is a second argument
\IfValueT{##2}%
{
\end{spacing}{##2}
}
%End the theorem environment
\end{#1vArIAblE*}
}
}
\ExplSyntaxOff
%We now use the \DeclareTheorem
%Theorem like objects
\theoremstyle{plain}
%Uses proofs
\DeclareTheorem{theorem}{Theorem}[\TheoremDepth]
\DeclareTheorem{proposition}[theorem]{Proposition}
\DeclareTheorem{lemma}[theorem]{Lemma}
\DeclareTheorem{corollary}[theorem]{Corollary}
%Uses subproos
\DeclareTheorem{claim}[theorem]{Claim}
\DeclareTheorem{remark}[theorem]{Remark}
\DeclareTheorem{example}[theorem]{Example}
%Definition like objects
\theoremstyle{definition}
\DeclareTheorem{definition}[theorem]{Definition}
\DeclareTheorem{axiom}[theorem]{Axiom}
\DeclareTheorem{convention}[theorem]{Convention}
\DeclareTheorem{conjecture}[theorem]{Conjecture}
%Remark like objects
\theoremstyle{remark}
\DeclareTheorem{question}[theorem]{Question}
\providecommand{\bysame}{\leavevmode\hbox to3em{\hrulefill}\thinspace}
\providecommand{\MR}{\relax\ifhmode\unskip\space\fi MR }
% \MRhref is called by the amsart/book/proc definition of \MR.
\providecommand{\MRhref}[2]{%
\href{http://www.ams.org/mathscinet-getitem?mr=#1}{#2}
}
\providecommand{\href}[2]{#2}
\begin{thebibliography}{Kun83}
\bibitem[Bar75]{Barwise}
J.~Barwise, \emph{Admissible sets and structures}, Springer-Verlag, Berlin-New
York, 1975, An approach to definability theory, Perspectives in Mathematical
Logic.
\bibitem[Gol19]{Golshani}
M.~Golshani, \emph{Fraïssé limit via forcing}, arXiv e-prints (2019).
\bibitem[Hod93]{MR1221741}
W.~Hodges, \emph{Model theory}, Encyclopedia of Mathematics and its
Applications, vol.~42, Cambridge University Press, Cambridge, 1993.
\bibitem[Kos23]{Cohen}
Z.~Kostana, \emph{Cohen-like first order structures}, Ann. Pure Appl. Logic
\textbf{174} (2023), no.~1, Paper No. 103172, 17.
\bibitem[Kun83]{MR756630}
K.~Kunen, \emph{Set theory}, Studies in Logic and the Foundations of
Mathematics, vol. 102, North-Holland Publishing Co., Amsterdam, 1983, An
introduction to independence proofs, Reprint of the 1980 original.
\bibitem[Mar16]{Marker2016-MARLOI}
D.~Marker, \emph{Lectures on infinitary model theory}, Cambridge University
Press, New York, NY, USA, 2016.
\end{thebibliography}