content/set-theory/story/story.tex
1% Chapter: Naive23\documentclass[../../../include/open-logic-chapter]{subfiles}45\begin{document}67\olchapter{sth}{story}{The Iterative Conception}89\olimport{extensionality}10\olimport{russells-paradox-again}11\olimport{predicativity}12\olimport{cumulative-approach}13\olimport{urelements}14\olimport{grundgesetze}1516\OLEndChapterHook1718\end{document}
content/set-theory/story/extensionality.tex
1\documentclass[../../../include/open-logic-section]{subfiles}23\begin{document}45\olfileid{sth}{story}{extensionality}6\olsection{Extensionality}78The very first thing to say is that sets are individuated by their9!!{element}s. More precisely:1011\begin{axiom}[Extensionality]12If sets $A$ and $B$ have the same !!{element}s, then $A$ and $B$ are13the same set.14\[15 \lforall[A][\lforall[B][(\lforall[x][(x \in A \liff x \in B)] \lif16 \eq[A][B])]]17\]18\end{axiom}1920We assumed this throughout \olref[sfr][][]{part}. But it bears21repeating. The Axiom of Extensionality expresses the basic idea that a22set is determined by its !!{element}s. (So sets might be contrasted with23\emph{concepts}, where precisely the same objects might fall under24many different concepts.) 2526Why embrace this principle? Well, it is plausible to say that any27denial of Extensionality is a decision to abandon anything which might28even be called \emph{set theory}. Set theory is no more nor less than29the theory of extensional collections. 3031The real challenge in \olref[sth][][]{part}, though, is to lay32down principles which tell us \emph{which sets exist}. And it turns33out that the only truly ``obvious'' answer to this question is34provably wrong.3536\end{document}
content/set-theory/story/russells-paradox-again.tex
1\documentclass[../../../include/open-logic-section]{subfiles}23\begin{document}45\olfileid{sth}{story}{rus}6\olsection{Russell's Paradox (again)}78In \olref[sfr][][]{part}, we worked with a na\"{i}ve set theory. But9according to a \emph{very} na\"{i}ve conception, sets are just the10extensions of predicates. This na\"ive thought would mandate the11following principle:1213\begin{defish}14 \emph{Na\"{i}ve Comprehension.} $\Setabs{x}{\phi(x)}$ exists for any formula $\phi$.15\end{defish}1617Tempting as this principle is, it is provably inconsistent. We saw this in \olref[sfr][set][rus]{sec}, but the result is so important, and so straightforward, that it's worth repeating. Verbatim.1819\begin{thm}[Russell's Paradox]20There is no set $R = \Setabs{x}{x \notin x}$21\end{thm}2223\begin{proof}24If $R = \Setabs{x}{x \notin x}$ exists, then25$R \in R$ iff $R \notin R$, which is a contradiction.26\end{proof}2728Russell discovered this result in June 1901. (He did not, though, put29the paradox in quite the form we just presented it, since he was30considering Frege's set theory, as outlined in \emph{Grundgesetze}. We31will return to this in \olref[blv]{sec}.) Russell wrote to32Frege on June 16, 1902, explaining the inconsistency in Frege's33system. For the correspondence, and a bit of background, see34\citet[pp.~124--8]{Heijenoort1967}. 3536It is worth emphasising that this two-line proof is a result of37\emph{pure logic}. Granted, we implicitly used a (non-logical?)\ axiom, Extensionality, in our notation $\Setabs{x}{x \notin x}$; for $\Setabs{x}{\phi(x)}$ is to be \emph{the unique} (by Extensionality) set of the $\phi$s, if one exists. But we can avoid even the hint of Extensionality, just by stating the result as follows:38\emph{there is no set whose members are exactly the non-self-membered39sets}. And this has nothing much to do with sets. As Russell himself observed, exactly similar reasoning40will lead you to conclude: \emph{no man shaves exactly the men who do41not shave themselves}. Or: \emph{no pug sniffs exactly the pugs which42don't sniff themselves}. And so on. Schematically, the shape of the43result is just: 44\[45\lnot \exists x \forall z(Rzx \liff \lnot R zz).46\]47And that's just a theorem (scheme) of first-order logic. Consequently,48we can't avoid Russell's Paradox just by tinkering with our set49theory; it arises before we even \emph{get} to set theory. If we're50going to use (classical) first-order logic, we simply have to51\emph{accept} that there is no set $R = \Setabs{x}{x\notin x}$. 5253The upshot is this. If you want to accept Na\"{i}ve Comprehension54whilst \emph{avoiding} inconsistency, you cannot just tinker with the55\emph{set theory}. Instead, you would have to overhaul your56\emph{logic}.5758Of course, set theories with non-classical logics have been presented.59But they are---to say the least---non-standard. The standard approach60to Russell's Paradox is to treat it as a straightforward non-existence61proof, and then to try to learn how to live with it. That is the62approach we will follow.6364\end{document}
content/set-theory/story/predicativity.tex
1\documentclass[../../../include/open-logic-section]{subfiles}23\begin{document}45\olfileid{sth}{story}{predicative} 67\olsection{Predicative and Impredicative}89The Russell set, $R$, was defined via $\Setabs{x}{x \notin10x}$. Spelled out more fully, $R$ would be the set which contains all11and only those sets which are not non-self-membered. So in defining12$R$, we quantify over the domain which would contain $R$ (if it13existed).1415This is an \emph{impredicative} definition. More generally, we might16say that a definition is impredicative iff it quantifies over a domain17which contains the object that is being defined. 18 19In the wake of the paradoxes, Whitehead, Russell, Poincar\'{e} and20Weyl rejected such impredicative definitions as ``viciously21circular'':22\begin{quote}23 An analysis of the paradoxes to be avoided shows that they all24 result from a kind of vicious circle. The vicious circles in25 question arise from supposing that a collection of objects may26 contain members which can only be defined by means of the27 collection as a whole[\ldots. \textparagraph]2829 The principle which enables us to avoid illegitimate totalities30 may be stated as follows: `Whatever involves \emph{all} of a31 collection must not be one of the collection'; or, conversely:32 `If, provided a certain collection had a total, it would have33 members only definable in terms of that total, then the said34 collection has no total.' We shall call this the `vicious-circle35 principle,' because it enables us to avoid the vicious circles36 involved in the assumption of illegitimate totalities.37 \citep[p.~37]{WhiteheadRussell1910}38\end{quote}39If we follow them in rejecting the \emph{vicious-circle principle},40then we might attempt to replace the disastrous Na\"{i}ve41Comprehension Scheme (of \olref[sth][story][rus]{sec}) with something like this: 4243\begin{defish}44\emph{Predicative Comprehension.} For every formula $\phi$ quantifying only over sets: the set$^\prime$ $\Setabs{x}{\phi(x)}$ exists.45\end{defish}4647So long as sets$^{\prime}$ are not sets, no contradiction will ensue. 4849Unfortunately, Predicative Comprehension is not very50\emph{comprehensive}. After all, it introduces us to new entities,51sets$^\prime$. So we will have to consider formulas which quantify52over sets$^\prime$. If they always yield a set$^\prime$, then53Russell's paradox will arise again, just by considering the54set$^\prime$ of all non-self-membered sets$^\prime$. So, pursuing the55same thought, we must say that a formula quantifying over56sets$^\prime$ yields a corresponding set$^{\prime\prime}$. And then we57will need sets$^{\prime\prime\prime}$,58sets$^{\prime\prime\prime\prime}$, etc. To prevent a rash of primes,59it will be easier to think of these as sets$_0$, sets$_1$, sets$_2$,60sets$_3$, sets$_4$,\ldots. And this would give us a way into the61(simple) theory of types. 6263There are a few obvious objections against such a theory (though it is64not obvious that they are \emph{overwhelming} objections). In brief:65the resulting theory is cumbersome to use; it is profligate in66postulating different kinds of objects; and it is not clear, in the67end, that impredicative definitions are even \emph{all that bad}. 68 69To bring out the last point, consider this remark from70\citeauthor{Ramsey1925}:71\begin{quote}72 we may refer to a man as the tallest in a group, thus identifying73 him by means of a totality of which he is himself a member without74 there being any vicious circle. \citep{Ramsey1925}75\end{quote}76Ramsey's point is that ``the tallest man in the group'' \emph{is} an77impredicative definition; but it is obviously perfectly kosher. 7879One might respond that, in this case, we could pick out the tallest80person by \emph{predicative} means. For example, maybe we could just81point at the man in question. The objection against impredicative82definitions, then, would clearly need to be limited to entities which83can \emph{only} be picked out impredicatively. But even then, we would84need to hear more, about why such ``essential impredicativity'' would85be so bad.\footnote{For more, see \citet{Linnebo2010}.}8687Admittedly, impredicative definitions are extremely bad news, if we88want our definitions to provide us with something like a recipe for89\emph{creating} an object. For, given an impredicative definition, one90would genuinely be caught in a vicious circle: to create the91impredicatively specified object, one would \emph{first} need to92create all the objects (including the impredicatively specified93object), since the impredicatively specified object is specified in94terms of all the objects; so one would need to create the95impredicatively specified object before one had created it itself. But96again, this is only a serious objection against ``essentially97impredicatively'' specified sets, if we think of sets as things that98we \emph{create}. And we (probably) don't.99100As such---for better or worse---the approach which became common does101not involve taking a hard line concerning (im)\-pre\-di\-ca\-tiv\-ity.102Rather, it involves what is now regarded as the cumulative-iterative103approach. In the end, this will allow us to stratify our sets into104``stages''---a \emph{bit} like the predicative approach stratifies105entities into sets$_0$, sets$_1$, sets$_2$, \ldots---but we will not106postulate any difference in kind between them. 107108\end{document}
content/set-theory/story/cumulative-approach.tex
1\documentclass[../../../include/open-logic-section]{subfiles}23\begin{document}45\olfileid{sth}{story}{approach}6\olsection{The Cumulative-Iterative Approach}78Here is a slightly fuller statement of how we will stratify sets into9stages:10\begin{quote}11 Sets are formed in \emph{stages}. For each stage $S$, there are12 certain stages which are \emph{before} $S$. At stage $S$, each13 collection consisting of sets formed at stages before $S$ is14 formed into a set. There are no sets other than the sets which are15 formed at stages. \citep[p.~323]{Shoenfield:AST}16\end{quote} 17This is a sketch of the \emph{cumulative-iterative conception of18set}. It will underpin the formal set theory that we present in19\olref[sth][][]{part}. 2021Let's explore this in a little more detail. As Shoenfield describes22the process, at every stage, we form new sets from the {sets} which23were available to us from earlier stages. So, on Shoenfield's picture,24at the initial stage, stage $0$, there are no \emph{earlier} stages,25and so \emph{a fortiori} there are no sets available to us from26earlier stages.\footnote{Why should we assume that there \emph{is} a27first stage? See the footnote to \stagesord{} in28\olref[z][story]{sec}.} So we form only one set: the set29with no !!{element}s $\emptyset$. At stage $1$, exactly one set is30available to us from earlier stages, so only one new set is31$\{\emptyset\}$. At stage $2$, two sets are available to us from32earlier stages, and we form two new sets $\{\{\emptyset\}\}$ and33$\{\emptyset, \{\emptyset\}\}$. At stage $3$, four sets are available34to us from earlier stages, so we form twelve new sets\ldots. As such,35the cumulative-iterative picture of the sets will look a bit like36this (with numbers indicating stages):37\begin{center}38 \begin{tikzpicture}[scale=0.6]39 \tikzset{cut_here/.style={densely dotted}}40 \draw (-4,6) -- (-1, 0)-- (2,6); 41 \draw[cut_here] (-5,8)--(-4,6);42 \draw[cut_here] (2,6)--(3,8);43 \draw(-1.5, 1)--(-0.5, 1);44 \draw(-2, 2)--(0, 2);45 \draw(-2.5, 3)--(0.5,3);46 \draw(-3, 4)--(1, 4);47 \draw(-3.5, 5)--(1.5, 5);48 \draw(-4, 6)--(2, 6);49 \node[label] at (0, 0) {\small 0};50 \node[label] at (0.5, 1) {\small 1};51 \node[label] at (1, 2) {\small 2};52 \node[label] at (1.5, 3) {\small 3};53 \node[label] at (2, 4) {\small 4};54 \node[label] at (2.5, 5) {\small 5};55 \node[label] at (3, 6) {\small 6};56 \end{tikzpicture}57\end{center}58So: why should we embrace this story? 5960One reason is that it is a nice, tractable story. Given the demise of61the most obvious story, i.e., Na\"ive Comprehension, we are in want of62something nice. 6364But the story is not \emph{just} nice. We have a good reason to65believe that any set theory based on this story will be66\emph{consistent}. Here is why. 6768Given the cumulative-iterative conception of set, we form sets at69stages; and their !!{element}s must be objects which were available70\emph{already}. So, for any stage~$S$, we can form the set 71\[72 R_S = \Setabs{x}{x \notin x \text{ and $x$ was available before $S$}}73\]74The reasoning involved in proving Russell's Paradox will now establish75that $R_S$ itself is not available before stage $S$. And that's not a76contradiction. Moreover, if we embrace the cumulative-iterative77conception of set, then we shouldn't even have \emph{expected} to be78able to form the Russell set itself. For that would be the set of all79non-self-membered sets that ``will ever be available''. In short: the80fact that we (provably) can't form the Russell set isn't81\emph{surprising}, given the cumulative-iterative story; it's what we82would \emph{predict}.8384%In one sense, then, the cumulative-iterative conception of set yields a response to Russell's Paradox which is rather like the \emph{predicativist}'s response. After all, the predicativist said that the collection of all non-self-membered sets$_n$ is a set$_{n+1}$, and this set$_n$ will not be self-membered. (Indeed, most predicativists will treat it as \emph{ungrammatical} to try to ask whether a set is self-membered.)85%86%But there is an important difference between the cumulative-iterative approach and the predicativist approach: the cumulative-iterative approach treats all of our entities as being \emph{of the same kind}. The predicativist will presumably have an empty set$_0$ (i.e., a set$_0$ with no !!{element}s), and an empty set$_1$ (i.e., a set$_1$ with no !!{element}s), and these will be \emph{different entities}. But on the cumulative-iterative approach, there is just one empty set, $\emptyset$, and it will be available at every stage of the hierarchy. 8788%Indeed, the Axiom of Extensionality, stated at the start of this chapter, will hold true of the elements in this hierarchy (so that there can be \emph{only one} ``empty'' set). But I do not intend to use the cumulative-iterative conception to justify Extensionality (see \cite{Boolos1971}). Again: I suggest that we should accept Extensionality, just because we are interested in extensional collections. 8990\end{document}
content/set-theory/story/urelements.tex
1\documentclass[../../../include/open-logic-section]{subfiles}23\begin{document}45\olfileid{sth}{story}{urelements}6\olsection{Urelements or Not?}78In the next few chapters, we will try to extract axioms from the9cumulative-iterative conception of set. But, before going any further,10we need to say something more about \emph{urelements}. 1112The picture of \olref[approach]{sec} allowed us only to form new sets13from old \emph{sets}. However, we might want to allow that certain14\emph{non-sets}---cows, pigs, grains of sand, or whatever---can be15!!{element}s of sets. In that case, we would start with certain basic16elements, \emph{urelements}, and then say that at each stage $S$ we17would form ``all possible'' sets consisting of urelements or sets18formed at stages before $S$ (in any combination). The resulting19picture would look more like this:20\begin{center}21 \begin{tikzpicture}[scale=0.6]22 \tikzset{cut_here/.style={densely dotted}}23 \draw (-4,6) -- (-1, 0) -- (1,0) -- (4,6); 24 \draw[cut_here] (-5,8)--(-4,6);25 \draw[cut_here] (5,8)--(4,6);26 \draw(-1.5, 1)--(1.5, 1);27 \draw(-2, 2)--(2, 2);28 \draw(-2.5, 3)--(2.5,3);29 \draw(-3, 4)--(3, 4);30 \draw(-3.5, 5)--(3.5, 5);31 \draw(-4, 6)--(4, 6);32 \node[label] at (2, 0) {\small 0};33 \node[label] at (2.5, 1) {\small 1};34 \node[label] at (3, 2) {\small 2};35 \node[label] at (3.5, 3) {\small 3};36 \node[label] at (4, 4) {\small 4};37 \node[label] at (4.5, 5) {\small 5};38 \node[label] at (5, 6) {\small 6};39 \end{tikzpicture}40\end{center}41So now we have a decision to take: \emph{Should we allow urelements?}4243Philosophically, it makes sense to include urelements in our44theorising. The main reason for this is to make our set theory45\emph{applicable}. To illustrate the point, recall from46\olref[sfr][siz][]{chap} that we say that two sets $A$ and~$B$ have47the same size, i.e., $\cardeq{A}{B}$, iff there is a bijection between48them. Now, if the cows in the field and the pigs in the sty both form49sets, we can offer a set-theoretical treatment of the claim ``there50are as many cows as pigs''. But if we ban urelements, so that the cows51and the pigs do \emph{not} form sets, then that set-theoretical52treatment will be unavailable. Indeed, we will have no straightforward53ability to apply set theory to anything other than sets themselves.54(For more reasons to include urelements, see \citealt[pp.~vi, 24,5550--1]{Potter2004}.)5657Mathematically, however, it is quite rare to allow urelements. In58part, this is because it is \emph{very slightly} easier to formulate59set theory without urelements. But, occasionally, one finds more60interesting justifications for excluding urelement from set theory:61\begin{quote}62 In accordance with the belief that set theory is the foundation of63 mathematics, we should be able to capture all of mathematics by64 just talking about sets, so our variable should not range over65 objects like cows and pigs. 66 %But if $C$ is a cow, $\{C\}$ is a set, but not a legitimate mathematical object. 67 \citep[p.~8]{Kunen1980}68\end{quote}69So: a focus on applicability would suggest \emph{including}70urelements; a focus on a reductive foundational goal (reducing71mathematics to pure set theory) might suggest \emph{excluding} them.72Mild laziness, too, points in the direction of excluding urelements. 7374We will follow the laziest path. Partly, though, there is a75pedagogical justification. Our aim is to introduce you to the elements76of set theory that you would need in order to get started on the77philosophy of set theory. And most of that philosophical literature78discusses set theories formulated \emph{without} urelements. So this79book will, perhaps, be of more use, if it hews fairly closely to that80literature.8182\end{document}
content/set-theory/story/grundgesetze.tex
1\documentclass[../../../include/open-logic-section]{subfiles}23\begin{document}45\olfileid{sth}{story}{blv}67\olsection{Appendix: Frege's Basic Law V}89In \olref[rus]{sec}, we explained that Russell's formulated his10paradox as a problem for the system Frege outlined in his11\emph{Grundgesetze}. Frege's system did not include a direct12formulation of Na\"{i}ve Comprehension. So, in this appendix, we will13very briefly explain what Frege's system \emph{did} include, and how14it relates to Na\"ive Comprehension and how it relates to Russell's15Paradox.1617Frege's system is \emph{second-order}, and was designed to formulate18the notion of an \emph{extension of a concept}.\footnote{Strictly19speaking, Frege attempts to formalize a more general notion: the20``value-range'' of a function. Extensions of concepts are a special21case of the more general notion. See \citet[pp.\ 8--9]{Heck2012} for22the details.} Using notation inspired by Frege, we will write23$\fregeext{x}{F(x)}$ for \emph{the extension of the concept~$F$}. This24is a device which takes a \emph{predicate}, ``$F$'', and turns it into25a (first-order) \emph{term}, ``$\fregeext{x}{F(x)}$''. Using this26device, Frege offered the following \emph{definition} of membership:27\[28 a \in b =_\text{df} \exists G(b = \fregeext{x}{G(x)} \land Ga)29\]30roughly: $a \in b$ iff $a$ falls under a concept whose extension is31$b$. (Note that the quantifier ``$\exists G$'' is second-order.) Frege32also maintained the following principle, known as \emph{Basic Law V}: 33$$\fregeext{x}{F(x)} = \fregeext{x}{G(x)} \liff \forall x (Fx \liff Gx)$$34roughly: concepts have identical extensions iff they are coextensive. (Again, both ``$F$'' and ``$G$'' are in predicate position.) Now a simple principle connects membership with property-satisfaction:3536\begin{lem}[in \emph{Grundgesetze}]\ollabel{lem:Fregeextensions}37$\forall F \forall a(a \in \fregeext{x}{F(x)} \liff Fa)$38\end{lem}3940\begin{proof} 41Fix $F$ and $a$. Now $a \in \fregeext{x}{F(x)}$ iff $\exists G(\fregeext{x}{F(x)}42= \fregeext{x}{G(x)} \land Ga)$ (by the definition of membership) iff43$\exists G(\forall x(Fx \liff Gx) \land Ga)$ (by Basic Law V) iff $Fa$44(by elementary second-order logic).45\end{proof}4647And this yields Na\"ive Comprehension almost immediately:4849\begin{lem}[in \emph{Grundgesetze}.]50$\forall F \exists s \forall a (a \in s \liff Fa)$51\end{lem}5253\begin{proof}54Fix $F$; now \olref{lem:Fregeextensions} yields $\forall a (a \in55\fregeext{x}{F(x)} \liff Fa)$; so $\exists s\forall a(a \in s \liff Fa)$ by56existential generalisation. The result follows since $F$ was57arbitrary.58\end{proof}5960Russell's Paradox follows by taking $F$ as given by $\forall x(Fx \liff x \notin x)$. 6162\end{document}