content/set-theory/cardinals/cardinals.tex
1% Part: set-theory2% Chapter: cardinals34\documentclass[../../../include/open-logic-chapter]{subfiles}56\begin{document}78\olchapter{sth}{cardinals}{Cardinals}910\olimport{cp}11\olimport{cardsasords}12\olimport{milestone}13\olimport{classing}14\olimport{hp}1516\OLEndChapterHook1718\end{document}
content/set-theory/cardinals/cp.tex
1\documentclass[../../../include/open-logic-section]{subfiles}23\begin{document}45\olfileid{sth}{cardinals}{cp}6\olsection{Cantor's Principle}78Cast your mind back to \olref[ordinals][vn]{sec}. We were discussing9well-ordered sets, and suggested that it would be nice to have objects10which go proxy for well-orders. With this is mind, we introduced11ordinals, and then showed in12\olref[ordinals][ordtype]{ordtypesworklikeyouwant} that these13behave as we would want them to, i.e.:14\[15 \ordtype{A, <} = \ordtype{B, \lessdot} 16 \text{ iff } \tuple{A, <} \isomorphic \tuple{B, \lessdot}.17\]18Cast your mind back even further, to \olref[sfr][siz][equ]{sec}.19There, working na\"ively, we introduced the notion of the ``size'' of20a set. Specifically, we said that two sets are equinumerous,21$\cardeq{A}{B}$, just in case there is !!a{bijection} $f \colon A \to22B$. This is an intrinsically {simpler} notion than that of a23well-ordering: we are only interested in !!{bijection}s, and not (as24with order-isomorphisms) whether the !!{bijection}s ``preserve any25structure''.2627This all gives rise to an obvious thought. Just as we introduced28certain objects, \emph{ordinals}, to calibrate well-orders, we can29introduce certain objects, \emph{cardinals}, to calibrate size. That30is the aim of this chapter. 3132Before we say what these cardinals will be, we should lay down a33principle which they ought to satisfy. Writing $\card{X}$ for the34cardinality of the set $X$, we would want them to obey:35\[36 \card{A} = \card{B} \text{ iff } \cardeq{A}{B}.37\]38We'll call this \emph{Cantor's} Principle, since Cantor was probably39the first to have it very clearly in mind. (We'll say more about its40relationship to \emph{Hume's} Principle in \olref[hp]{sec}.) So41our aim is to define $\card{X}$, for each $X$, in such a way that it42delivers Cantor's Principle.4344\end{document}
content/set-theory/cardinals/cardsasords.tex
1\documentclass[../../../include/open-logic-section]{subfiles}23\begin{document}45\olfileid{sth}{cardinals}{cardsasords}6\olsection{Cardinals as Ordinals}78In fact, our theory of cardinals will just make (shameless) use of our9theory of ordinals. That is: we will just define cardinals as certain10specific ordinals. In particular, we will offer the following:1112\begin{defn}\ollabel{defcardinalasordinal}13If $A$ can be well-ordered, then $\card{A}$ is the least ordinal14$\gamma$ such that $\cardeq{A}{\gamma}$. For any ordinal $\gamma$, we15say that $\gamma$ is a \emph{cardinal} iff $\gamma = \card{\gamma}$.16\end{defn}1718We just used the phrase ``$A$ can be well-ordered''. As is almost19always the case in mathematics, the modal locution here is just a20hand-waving gloss on an existential claim: to say ``$A$ can be21well-ordered'' is just to say ``there is a relation which22well-orders~$A$''. 2324But there is a snag with \olref{defcardinalasordinal}. We would like25it to be the case that \emph{every} set has a size, i.e., that26$\card{A}$ exists for every~$A$. The definition we just gave, though,27begins with a conditional: ``\emph{If} $A$ can be28well-ordered\ldots''. If there is some set $A$ which cannot be29well-ordered, then our definition will simply fail to define an object~$\card{A}$.3031So, to use \olref{defcardinalasordinal}, we need a guarantee that32every set can be well-ordered. Sadly, though, this guarantee is33unavailable in~$\ZF$. So, if we want to use34\olref{defcardinalasordinal}, there is no alternative but to add a new35axiom, such as:36\begin{axiom}[Well-Ordering]37Every set can be well-ordered.38\end{axiom}39We will discuss whether the Well-Ordering Axiom is acceptable in40\olref[choice][]{chap}. From now on, though, we will simply help41ourselves to it. And, using it, it is quite straightforward to prove42that cardinals (as defined in \olref{defcardinalasordinal}) exist and43behave nicely:4445\begin{lem}\ollabel{lem:CardinalsExist}46For every set $A$:47\begin{enumerate}48 \item\ollabel{cardaexists} $\card{A}$ exists and is unique;49 \item\ollabel{cardaapprox} $\cardeq{\card{A}}{A}$;50 \item\ollabel{cardaidem} $\card{A}$ is a cardinal, i.e.,51 $\card{A} = \card{\card{A}}$;52\end{enumerate}53\end{lem}5455\begin{proof}56Fix $A$. By Well-Ordering, there is a well-ordering $\tuple{A, R}$. By57\olref[ordinals][ordtype]{thmOrdinalRepresentation}, $\tuple{A,58R}$ is isomorphic to a unique ordinal, $\beta$. So59$\cardeq{A}{\beta}$. By Transfinite Induction, there is a uniquely60least ordinal, $\gamma$, such that $\cardeq{A}{\gamma}$. So $\card{A}61= \gamma$, establishing \olref{cardaexists} and \olref{cardaapprox}.62To establish \olref{cardaidem}, note that if $\delta \in \gamma$ then63$\cardless{\delta}{A}$, by our choice of $\gamma$, so that also64$\cardless{\delta}{\gamma}$ since equinumerosity is an equivalence65relation (\olref[sfr][siz][equ]{equinumerosityisequi}). So $\gamma =66\card{\gamma}$. 67%So, for reductio, suppose that there is some ordinal $\delta \in \gamma$ such that $\cardeq{\gamma}{\delta}$. Then, $\cardeq{\cardeq{A}{\gamma}}{\delta}$ so that $\cardeq{A}{\delta}$ by \olref[sfr][set][equ]{equinumerosityisequi}, which contradicts the choice of $\gamma$ as the least ordinal such that $\cardeq{A}{\gamma}$. 68\end{proof}6970The next result guarantees Cantor's Principle, and more besides.71(Note that cardinals inherit their ordering from the ordinals, i.e.,72$\cardfont{a} < \cardfont{b}$ iff $\cardfont{a} \in \cardfont{b}$. In73formulating this, we will use Fraktur letters for objects we know to be74cardinals. This is fairly standard. A common alternative is to use75Greek letters, since cardinals are ordinals, but to choose them from76the middle of the alphabet, e.g.: $\kappa, \lambda$.):77\begin{lem}\ollabel{lem:CardinalsBehaveRight}78For any sets $A$ and $B$:79\begin{align*}80 \cardeq{A}{B} &\text{ iff } \card{A} = \card{B}\\81 \cardle{A}{B} &\text{ iff } \card{A} \leq \card{B}\\82 \cardless{A}{B}&\text{ iff } \card{A} < \card{B}83\end{align*}84\end{lem}8586\begin{proof}87We will prove the left-to-right direction of the second claim (the88other cases are similar, and left as an exercise). So, consider the89following diagram:90\begin{center}91 \begin{tikzpicture}92 \node (nodea) {$A$};93 \node[right = 6em of nodea] (nodeb) {$B$};94 \node[below = 2em of nodea] (nodecarda) {$\card{A}$};95 \node[below = 2em of nodeb] (nodecardb) {$\card{B}$};96 \draw[->] (nodea)--(nodeb);97 \draw[<->] (nodea)--(nodecarda);98 \draw[<->] (nodeb)--(nodecardb);99 \draw[->, dashed] (nodecarda)--(nodecardb);100\end{tikzpicture}101\end{center}102The double-headed arrows indicate !!{bijection}s, whose existence is103guaranteed by \olref{lem:CardinalsExist}. In assuming that104$\cardle{A}{B}$, there is !!a{injection} $A\to B$. Now,105chasing the arrows around from $\card{A}$ to $A$ to $B$ to $\card{B}$,106we obtain !!a{injection} $\card{A} \to \card{B}$ (the dashed arrow).107\end{proof}\noindent We can also use \olref{lem:CardinalsBehaveRight}108to re-prove Schr\"{o}der--Bernstein. This is the claim that if109$\cardle{A}{B}$ and $\cardle{B}{A}$ then $\cardeq{A}{B}$. We stated110this as \olref[sfr][siz][sb]{thm:schroder-bernstein}, but first proved111it---with some effort---in \olref[sfr][infinite][card-sb]{sec}.112Now consider:113114\begin{proof}[Re-proof of Schr\"oder-Bernstein]115If $\cardle{A}{B}$ and $\cardle{B}{A}$, then $\card{A} \leq \card{B}$116and $\card{B} \leq \card{A}$ by \olref{lem:CardinalsBehaveRight}. So117$\card{A} = \card{B}$ and $\cardeq{A}{B}$ by Trichotomy and118\olref{lem:CardinalsBehaveRight}.119\end{proof}120\noindent121Whilst this is a very simple proof, it implicitly relies on both122Replacement (to secure123\olref[ordinals][ordtype]{thmOrdinalRepresentation}) and on124Well-Ordering (to guarantee \olref{lem:CardinalsBehaveRight}). By125contrast, the proof of \olref[sfr][infinite][card-sb]{sec} was much126more self-standing (indeed, it can be carried out in~$\Zminus$).127128\end{document}
content/set-theory/cardinals/milestone.tex
1\documentclass[../../../include/open-logic-section]{subfiles}23\begin{document}45\olfileid{sth}{cardinals}{zfc} 6\olsection{$\ZFC$: A Milestone}78With the addition of Well-Ordering, we have reached the final9theoretical milestone. We now have all the axioms required for~$\ZFC$.10In detail:1112\begin{defn}13The theory $\ZFC$ has these axioms: Extensionality, Union, Pairs,14Powersets, Infinity, Foundation, Well-Ordering and all instances of15the Separation and Replacement schemes. Otherwise put, $\ZFC$ adds16Well-Ordering to~$\ZF$. 17\end{defn}1819$\ZFC$ stands for \emph{Zermelo--Fraenkel} set theory with20\emph{Choice}. Now this might seem slightly odd, since the axiom we21added was called ``Well-Ordering'', not ``Choice''. But, when we later22formulate {Choice}, it will turn out that Well-Ordering is equivalent23(modulo~$\ZF$) to Choice (see \olref[choice][woproblem]{thmwochoice}).24So which to take as our ``basic'' axiom is a matter of indifference.25And the name ``$\ZFC$'' is entirely standard in the literature. 2627\end{document}
content/set-theory/cardinals/classing.tex
1\documentclass[../../../include/open-logic-section]{subfiles}23\begin{document}45\olfileid{sth}{cardinals}{classing}6\olsection{Finite, \usetoken{S}{enumerable}, \usetoken{S}{nonenumerable}}78Now that we have been introduced to cardinals, it is worth spending a9little time talking about different varieties of cardinals;10specifically, finite, !!{enumerable}, and !!{nonenumerable} cardinals.1112Our first two results entail that the finite cardinals will be exactly13the finite ordinals, which we defined as our \emph{natural numbers}14back in \olref[z][infinity-again]{defnomega}: 1516\begin{prop}\ollabel{finitecardisoequal}17Let $n, m \in \omega$. Then $n = m$ iff $\cardeq{n}{m}$.18\end{prop}1920\begin{proof}21\emph{Left-to-right} is trivial. To prove \emph{right-to-left},22suppose $\cardeq{n}{m}$ although $n \neq m$. By Trichotomy, either $n23\in m$ or $m \in n$; suppose $n \in m$ without loss of generality.24Then $n \subsetneq m$ and there is !!a{bijection} $f \colon m \to n$,25so that $m$ is Dedekind infinite, contradicting26\olref[z][infinity-again]{naturalnumbersarentinfinite}.27\end{proof}2829\begin{cor}\ollabel{naturalsarecardinals}30If $n \in \omega$, then $n$ is a cardinal. 31\end{cor}3233\begin{proof}34Immediate.35\end{proof}36\noindent37It also follows that several reasonable notions of what it might mean38to describe a cardinal as ``finite'' or ``infinite'' coincide:39\begin{thm}\ollabel{generalinfinitycharacter}For any set $A$, the following are equivalent:40\begin{enumerate}41 \item\ollabel{card:notinomega} $\card{A} \notin \omega$, i.e.,42 $A$ is not a natural number;43 \item\ollabel{card:omegaplus} $\omega \leq \card{A}$;44 \item\ollabel{card:infinite} $A$ is Dedekind infinite.45\end{enumerate}46\end{thm}4748\begin{proof}49From \olref[ord-arithmetic][using-addition]{ordinfinitycharacter},50\olref[cardinals][cardsasords]{lem:CardinalsBehaveRight}, and51\olref{naturalsarecardinals}. 52\end{proof}5354This licenses the following \emph{definition} of some notions which we55used rather informally in \olref[sfr][][]{part}:5657\begin{defn}\ollabel{defnfinite}58We say that $A$ is \emph{finite} iff $\card{A}$ is a natural number,59i.e., $\card{A} \in \omega$. Otherwise, we say that $A$ is60\emph{infinite}.61\end{defn}62\noindent 63But note that this definition is presented against the background of64$\ZFC$. After all, we needed Well-Ordering to guarantee that every set65has a cardinality. And indeed, without Well-Ordering, there can be a66set which is neither finite nor Dedekind infinite. We will return to67this sort of issue in \olref[choice][]{chap}. For now, we continue to68rely upon Well-Ordering.6970Let us now turn from the finite cardinals to the infinite cardinals.71Here are two elementary points:7273\begin{cor}\ollabel{omegaisacardinal}74$\omega$ is the least infinite cardinal. 75\end{cor}7677\begin{proof}78$\omega$ is a cardinal, since $\omega$ is Dedekind infinite and if79$\cardeq{\omega}{n}$ for any $n \in \omega$ then $n$ would be Dedekind80infinite, contradicting81\olref[z][infinity-again]{naturalnumbersarentinfinite}. Now82$\omega$ is the least infinite cardinal by definition. 83\end{proof}8485\begin{cor}86Every infinite cardinal is a limit ordinal.87\end{cor}8889\begin{proof}90Let $\alpha$ be an infinite successor ordinal, so $\alpha = \beta91\ordplus 1$ for some $\beta$. By \olref{finitecardisoequal}, $\beta$92is also infinite, so $\cardeq{\beta}{\beta \ordplus 1}$ by93\olref[ord-arithmetic][using-addition]{ordinfinitycharacter}. Now94$\card{\beta} = \card{\beta\ordplus 1} = \card{\alpha}$ by95\olref[cardinals][cardsasords]{lem:CardinalsBehaveRight}, so96that $\alpha \neq \card{\alpha}$.97\end{proof}9899Now, as early as \olref[sfr][siz][enm-alt]{defn:enumerable}, we flagged we100can distinguish between !!{enumerable} and !!{nonenumerable} infinite101sets. That definition naturally leads to the following:102103\begin{prop}104$A$ is !!{enumerable} iff $\card{A} \leq \omega$, and $A$ is105!!{nonenumerable} iff $\omega < \card{A}$.106\end{prop}107108\begin{proof}109By Trichotomy, the two claims are equivalent, so it suffices to prove110that $A$ is !!{enumerable} iff $\card{A} \leq \omega$. For111\emph{right-to-left}: if $\card{A} \leq \omega$, then112$\cardle{A}{\omega}$ by113\olref[cardinals][cardsasords]{lem:CardinalsBehaveRight} and114\olref{omegaisacardinal}. For \emph{left-to-right}: suppose $A$ is115!!{enumerable}; then by \olref[sfr][siz][enm-alt]{defn:enumerable} there116are three possible cases:117\begin{enumerate}118 \item if $A = \emptyset$, then $\card{A} = 0 \in \omega$, by119 \olref{naturalsarecardinals} and120 \olref[cardinals][cardsasords]{lem:CardinalsBehaveRight}.121 \item if $\cardeq{n}{A}$, then $\card{A} = n \in \omega$, by122 \olref{naturalsarecardinals} and123 \olref[cardinals][cardsasords]{lem:CardinalsBehaveRight}.124 \item if $\cardeq{\omega}{A}$, then $\card{A} = \omega$, by \olref{omegaisacardinal}.125\end{enumerate}126So in all cases, $\card{A} \leq \omega$. 127\end{proof}128\noindent129Indeed, $\omega$ has a special place. Whilst there are many countable ordinals:130131\begin{cor}132$\omega$ is the only !!{enumerable} infinite cardinal.133\end{cor}134135\begin{proof}136Let $\cardfont{a}$ be !!a{enumerable} infinite cardinal. Since137$\cardfont{a}$ is infinite, $\omega \leq \cardfont{a}$. Since138$\cardfont{a}$ is !!a{enumerable} cardinal, $\cardfont{a} =139\card{\cardfont{a}} \leq \omega$. So $\cardfont{a} = \omega$ by140Trichotomy. \end{proof}141142Of course, there are infinitely many cardinals. So we might ask:143\emph{How many cardinals are there?} The following results show that144we might want to reconsider that question.145146\begin{prop}\ollabel{unioncardinalscardinal}147If every member of $X$ is a cardinal, then $\bigcup X$ is a cardinal.148\end{prop}149150\begin{proof}151It is easy to check that $\bigcup X$ is an ordinal. Let $\alpha \in152\bigcup X$ be an ordinal; then $\alpha \in \cardfont{b} \in X$ for153some cardinal $\cardfont{b}$. Since $\cardfont{b}$ is a cardinal,154$\cardless{\alpha}{\cardfont{b}}$. Since $\cardfont{b} \subseteq155\bigcup X$, we have $\cardle{\cardfont{b}}{\bigcup X}$, and so156$\cardneq{\alpha}{\bigcup X}$. Generalising, $\bigcup X$ is a157cardinal.158\end{proof} 159160\begin{thm}\ollabel{lem:NoLargestCardinal}161There is no largest cardinal.162\end{thm}163164\begin{proof}165For any cardinal $\cardfont{a}$, Cantor's Theorem166(\olref[sfr][siz][car]{thm:cantor}) and167\olref[cardinals][cardsasords]{lem:CardinalsExist} entail that168$\cardfont{a} < \card{\Pow{\cardfont{a}}}$.169\end{proof}170171\begin{thm}172The set of all cardinals does not exist.173\end{thm}174175\begin{proof}176For reductio, suppose $C = \Setabs{\cardfont{a}}{\cardfont{a} \text{177is a cardinal}}$. Now $\bigcup C$ is a cardinal by178\olref{unioncardinalscardinal}, so by \olref{lem:NoLargestCardinal}179there is a cardinal $\cardfont{b} > \bigcup C$. By definition180$\cardfont{b} \in C$, so $\cardfont{b} \subseteq \bigcup{C}$, so that181$\cardfont{b} \leq \bigcup C$, a contradiction.182\end{proof}183184You should compare this with both Russell's Paradox and Burali-Forti. 185186\end{document}
content/set-theory/cardinals/hp.tex
1\documentclass[../../../include/open-logic-section]{subfiles}23\begin{document}45\olfileid{sth}{cardinals}{hp}6\olsection{Appendix: Hume's Principle}78In \olref[cp]{sec}, we described Cantor's Principle. This was:9\begin{align*}10 \card{A} = \card{B} & \text{ iff } A \approx B.11\intertext{This is very similar to what is now called12\emph{Hume's Principle}, which says:}13 \fregenum{x} {F(x)} = \fregenum{x}{G(x)} & \text{ iff } F \sim G14\end{align*}15where `$F \sim G$' abbreviates that there are exactly as many $F$s as16$G$s, i.e., the $F$s can be put into a bijection with the $G$s, i.e.:17\begin{align*}18 \exists R(&\forall v\forall y(Rvy \lif (Fv \land Gy)) \land {}\\19 &\forall v(Fv \lif \lexists![y][Rvy]) \land {}\\20 &\forall y(Gy \lif \lexists![v][Rvy]))21\end{align*}22But there is a type-difference between Hume's Principle and Cantor's23Principle. In the statement of Cantor's Principle, the variables24``$A$'' and ``$B$'' are first-order terms which stand for \emph{sets}.25In the statement of Hume's Principle, ``$F$'', ``$G$'' and ``$R$'' are26\emph{not} first-order terms; rather, they are in \emph{predicate27position}. (Maybe they stand for \emph{properties}.) So we might gloss28Hume's Principle in English as: the number of $F$s is the number of29$G$s iff the $F$s are bijective with the~$G$s. This is called30\emph{Hume's Principle}, because Hume once wrote this:31\begin{quote}32 When two numbers are so combined as that the one has always an unit33 answering to every unit of the other, we pronounce them equal.34 \citep[Pt.III Bk.1 \S1]{Hume1740}35\end{quote}36And Hume's Principle was brought to contemporary mathematico-logical37prominence by \citet[\S63]{Frege1884}, who quoted this passage from38Hume, before (in effect) sketching (what we have called) Hume's39Principle. 4041You should note the structural similarity between Hume's Principle and42Basic Law~V. We formulated this in \olref[story][blv]{sec} as43follows:44\[45 \fregeext{x}{F(x)} = \fregeext{x}{G(x)} \text{iff } \lforall[x][(F(x) \liff G(x))].46\]47And, at this point, some commentary and comparison might help. 4849There are two ways to take a principle like Hume's Principle or Basic50Law~V: \emph{predicatively} or \emph{impredicatively} (recall51\olref[story][predicative]{sec}). On the impredicative reading of52Basic Law~V, for each~$F$, the object $\fregeext{x}{F(x)}$ falls53within the domain of quantification that we used in formulating Basic54Law~V itself. Similarly, on the impredicative reading of Hume's55Principle, for each~$F$, the object $\fregenum{x}{F(x)}$ falls within56the domain of quantification that we used in formulating Hume's57Principle. By contrast, on the \emph{predicative} understanding, the58objects $\fregeext{x}{F(x)}$ and~$\fregenum{x}{F(x)}$ would be59entities from some \emph{different} domain. 6061Now, if we read Basic Law~V impredicatively, it leads to62inconsistency, via Na\"ive Comprehension (for the details, see63\olref[story][blv]{sec}). Much like Na\"ive Comprehension, it can be64rendered consistent by reading it \emph{predicatively}. But it65probably will not do everything that we wanted it to. 6667Hume's Principle, however, \emph{can} consistently be read68impredicatively. And, read thus, it is quite powerful.6970To illustrate: consider the predicate ``$x \neq x$'', which obviously71nothing satisfies. Hume's Principle now yields an object $\# x( x\neq72x)$. We might treat this as the number~$0$. Now, on the73\emph{impredicative} understanding---but \emph{only} on the74impredicative understanding---this entity $0$ falls within our75original domain of quantification. So we can sensibly apply Hume's76Principle with the predicate ``$x = 0$'' to obtain an object $\#x (x =770)$. We might treat this as the number~$1$. Moreover, Hume's Principle78entails that $0 \neq 1$, since there cannot be a bijection from the79non-self-identical objects to the objects identical with $0$ (there80are none of the former, but one of the latter). Now, working81impredicatively again, $1$~falls within our original domain of82quantification. So we can sensibly apply Hume's Principle with the83predicate ``$(x = 0 \lor x = 1)$'' to obtain an object $\#x(x = 0 \lor84x = 1)$. We might treat this as the number~$2$, and we can show that85$0\neq 2$ and $1 \neq 2$ and so on. 8687In short, taken impredicatively, Hume's Principle entails that there88are \emph{infinitely many objects}. And this has encouraged89\emph{neo-Fregean logicists} to take Hume's Principle as the90foundation for arithmetic. 9192Frege \emph{himself}, though, did not take Hume's Principle as his93foundation for arithmetic. Instead, Frege proved Hume's Principle from94an explicit definition: $\fregenum{x}{F(x)}$ is defined as the extension of95the concept $F \sim \Phi$. In modern terms, we might attempt to render96this as $\fregenum{x}{F(x)} = \Setabs{G}{F \sim G}$; but this will pull us97back into the problems of Na\"ive Comprehension.9899\end{document}