content/intuitionistic-logic/tableaux/tableaux.tex
1% Part: intuitionistic-logic2% Chapter: tableaux34\documentclass[../../../include/open-logic-chapter]{subfiles}56\begin{document}78\olchapter{int}{tab}{Intuitionistic \usetoken{P}{tableau}}910\begin{editorial}11 Draft chapter on prefixed tableaux for intuitionistic logic. Needs more12 examples, completeness proofs, and discussion of how one can find13 countermodels from unsuccessful searches for closed tableaux.14\end{editorial}1516\olimport{introduction}17\olimport{rules}18\olimport{proofs}19\olimport{soundness}20%\olimport{completeness}21%\olimport{countermodels}2223\OLEndChapterHook2425\end{document}26
content/intuitionistic-logic/tableaux/introduction.tex
1% Part: intuitionistic-logic2% Chapter: tableaux3% Section: introduction45\documentclass[../../../include/open-logic-section]{subfiles}67\begin{document}89\olfileid{int}{tab}{int}1011\olsection{Introduction}1213!!^{tableau}s are certain (downward-branching) trees of !!{signed14 formula}s, i.e., pairs consisting of a truth value sign ($\True$ or15$\False$) and !!a{sentence}16\[17\sFmla{\True}{!A} \text{ or } \sFmla{\False}{!A}.18\]19!!^a{tableau} begins with a number of \emph{assumptions}. Each further20!!{signed formula} is generated by applying one of the inference21rules. Some inference rules add one or more !!{signed formula}s to a22tip of the tree; others add two new tips, resulting in two branches.23Rules result in !!{signed formula}s where the !!{formula} is24less complex than that of the !!{signed formula} to which it was25applied. When a branch contains both $\sFmla{\True}{!A}$ and26$\sFmla{\False}{!A}$, we say the branch is \emph{closed}. If every27branch in !!a{tableau} is closed, the entire !!{tableau} is closed. A28closed !!{tableau} constitutes !!a{derivation} that shows that the set29of !!{signed formula}s which were used to begin the !!{tableau} are30unsatisfiable. This can be used to define a $\Proves$ relation:31$\Gamma \Proves !A$ iff there is some finite set~$\Gamma_0 = \{!B_1,32\dots, !B_n\} \subseteq \Gamma$ such that there is a closed33!!{tableau} for the assumptions34\[35\{\sFmla{\False}{!A}, \sFmla{\True}{!B_1}, \dots, \sFmla{\True}{!B_n}\}.36\]3738For intuitionistic logic, we have to both extend the notion of39!!{signed formula} and adjust the rules for the connectives. In40addition to a sign($\True$ or $\False$), !!{formula}s in modal41!!{tableau}s also have \emph{prefixes}~$\sigma$. The prefixes are42non-empty sequences of positive integers, i.e., $\sigma \in43(\PosInt)^* \setminus \{\emptyseq\}$. When we write such prefixes44without the surrounding $\tuple{\ }$, and separate the individual45!!{element}s by~$.$'s instead of $,$'s. If $\sigma$ is a prefix, then46$\sigma.n$ is $\sigma \concat \tuple{n}$; e.g., if $\sigma = 1.2.1$,47then $\sigma.3$ is $1.2.1.3$. So for instance,48\[49\sFmla{\True}{!A \lif (!B \lif !C)}[1.2]50\]51is a \emph{prefixed !!{signed formula}} (or just a \emph{prefixed52 !!{formula}} for short).5354Intuitively, the prefix names a world in a model that might satisfy55the !!{formula}s on a branch of !!a{tableau}, and if $\sigma$ names56some world, then $\sigma.n$ names a world accessible from (the world57named by)~$\sigma$.5859In intuitionistic models, the accessibility relation is reflexive and60transitive. In terms of prefixes, this means that $\sigma$ is61accessible from $\sigma$ itself, and so is any prefix that62extends~$\sigma$, i.e., any prefix of the form63$\sigma.n_1.\cdots.n_k$. Let's introduce the notation $\sigma.*$ to64indicate $\sigma$ itself and any extension of it. In other words, the65prefixes $\sigma.*$ are all and only the prefixes accessible66from~$\sigma$.6768\end{document}
content/intuitionistic-logic/tableaux/rules.tex
1% Part: intuitionistic-logic2% Chapter: tableaux3% Section: rules45\documentclass[../../../include/open-logic-section]{subfiles}67\begin{document}89\olfileid{int}{tab}{rul}1011\olsection{Rules for Intuitionistic Logic}1213The rules for the connectives $\land$ and $\lor$ are the same as for14regular propositional signed !!{tableau}s, just with prefixes added.15In each case, the rule applied to a signed !!{formula}16$\sFmla{S}{!A}[\sigma]$ produces new !!{formula}s that are also17prefixed by~$\sigma$. This should be intuitively clear: e.g., if $!A18\land !B$ is true at (a world named by)~$\sigma$, then $!A$ and $!B$19are true at~$\sigma$ (and not at any other world). We collect the20rules for $\land$ and $\lor$ in \olref{tab:prop-rules}.2122\begin{table}23 \[\def\arraystretch{3}\begin{array}{|c|c|}24 \hline25 \AxiomC{\sFmla{\True}{!A \land !B}[\sigma]}26 \RightLabel{\TRule{\True}{\land}}27 \UnaryInfC{\sFmla{\True}{!A}[\sigma]}28 \noLine29 \UnaryInfC{\sFmla{\True}{!B}[\sigma]}30 \DisplayProof31 &32 \AxiomC{\sFmla{\False}{!A \land !B}[\sigma]}33 \RightLabel{\TRule{\False}{\land}}34 \UnaryInfC{$\sFmla{\False}{!A}[\sigma] \quad \mid \quad35 \sFmla{\False}{!B}[\sigma]$}36 \DisplayProof37 \\[2ex]38 \hline39 \AxiomC{\sFmla{\True}{!A \lor !B}[\sigma]}40 \RightLabel{\TRule{\True}{\lor}}41 \UnaryInfC{$\sFmla{\True}{!A}[\sigma] \quad \mid \quad42 \sFmla{\True}{!B}[\sigma]$}43 \DisplayProof44 &45 \AxiomC{\sFmla{\False}{!A \lor !B}[\sigma]}46 \RightLabel{\TRule{\False}{\lor}}47 \UnaryInfC{\sFmla{\False}{!A}[\sigma]}48 \noLine49 \UnaryInfC{\sFmla{\False}{!B}[\sigma]}50 \DisplayProof51 \\[2ex]52 \hline53 \end{array}\]54 \caption{Prefixed !!{tableau} rules for $\land$ and $\lor$}55 \ollabel{tab:prop-rules}56\end{table}5758The closure condition is similar to that for ordinary !!{tableau}s,59although we require that not just the !!{formula}s, but also that the60prefixes must match. In fact, we can be somewhat more liberal: Since61in intuitionistic models, !!{formula}s, once true, remain true, it is62impossible that $!A$ is true at~$\sigma$ but false at any accessible63prefix~$\sigma.{*}$. So a branch is closed if it contains both64\[65\sFmla{\True}{!A}[\sigma] \quad\text{and}\quad \sFmla{\False}{!A}[\sigma.{*}]66\]67for some prefix $\sigma$ and !!{formula}~$!A$. Note that if the signs68are reversed, i.e., if it contains69\[70\sFmla{\False}{!A}[\sigma] \quad\text{and}\quad \sFmla{\True}{!A}[\sigma.{*}]71\]72the branch is closed only if $*$ is the empty sequence.7374In addition, a branch is closed if it contains~$\sFmla{\True}{\bot}[\sigma]$.7576The rules for setting up assumptions is also as for ordinary77!!{tableau}s, except that for assumptions we always use the78prefix~$1$. (It does not matter which prefix we use, as long as it's79the same for all assumptions.) So, e.g., we say that80\[81!B_1, \dots, !B_n \Proves !A82\]83iff there is a closed tableau for the assumptions84\[85\sFmla{\True}{!B_1}[1], \dots, \sFmla{\True}{!B_n}[1],86\sFmla{\False}{!A}[1].87\]8889For the conditional~$\lif$, the rules differ from the classical and90modal cases. The $\TRule{\lif}{\True}$ rule extends a branch91containing $\sFmla{\True}{!A \lif !B}[\sigma]$ by92$\sFmla{\True}{!A}[\sigma.{*}]$ and $\sFmla{\False}{!B}[\sigma.{*}]$ on two93different branches. It can only be applied for a prefix~$\sigma.{*}$94which \emph{already} occurs on the branch in which it is applied.95Let's call such a prefix ``used'' (on the branch). (Since $\sigma.{*}$96includes $\sigma$ itself, the rule can always be applied by adding the97prefixed signed formulas $\sFmla{\True}{!A}[\sigma]$ and98$\sFmla{\False}{!B}[\sigma]$ on separate branches.)99100The $\TRule{\lif}{\False}$ rule extends a branch containing101$\sFmla{\False}{!A \lif !B}[\sigma]$ by both102$\sFmla{\True}{!A}[\sigma.n]$ and $\sFmla{\False}{!B}[\sigma.n]$ on103the same branch, with $\sigma.n$ a prefix new to the branch. 104105The rules for $\lnot$ are defined analogously (using the definition of106$\lnot !A$ as $!A \lif \lfalse$).107108The rules are given in \olref{tab:rules-lif-lnot}.109110\begin{table}111 \[\def\arraystretch{3}\begin{array}{|c|c|}112 \hline113 \AxiomC{\sFmla{\True}{\lnot !A}[\sigma]}114 \RightLabel{\TRule{\True}{\lnot}}115 \UnaryInfC{$\sFmla{\False}{!A}[\sigma.{*}]$}116 \DisplayProof117 &118 \AxiomC{\sFmla{\False}{\lnot !A}[\sigma]}119 \RightLabel{\TRule{\False}{\lnot}}120 \UnaryInfC{\sFmla{\True}{!A}[\sigma.n]}121 \DisplayProof122 \\[1ex]123 \text{$\sigma.{*}$ is used} & \text{$\sigma.n$ is new}\\124 \hline125 \AxiomC{\sFmla{\True}{!A \lif !B}[\sigma]}126 \RightLabel{\TRule{\True}{\lif}}127 \UnaryInfC{$\sFmla{\False}{!A}[\sigma.{*}] \quad \mid \quad128 \sFmla{\True}{!B}[\sigma.{*}]$}129 \DisplayProof130 &131 \AxiomC{\sFmla{\False}{!A \lif !B}[\sigma]}132 \RightLabel{\TRule{\False}{\lif}}133 \UnaryInfC{\sFmla{\True}{!A}[\sigma.n]}134 \noLine135 \UnaryInfC{\sFmla{\False}{!B}[\sigma.n]}136 \DisplayProof137 \\[1ex]138 \text{$\sigma.{*}$ is used} & \text{$\sigma.n$ is new}\\139 \hline140 \end{array}\]141 \caption{Prefixed !!{tableau} rules for $\lnot$ and $\lif$}142 \ollabel{tab:rules-lif-lnot}143\end{table}144145\end{document}
content/intuitionistic-logic/tableaux/proofs.tex
1% Part: intuitionistic-logic2% Chapter: tableaux3% Section: proofs45\documentclass[../../../include/open-logic-section]{subfiles}67\begin{document}89\olfileid{int}{tab}{prf}1011\olsection{\usetoken{P}{tableau} for Intuitionistic Logic}1213\begin{ex}14 We give a closed tableau that shows $(!A \land !B) \lif !C \Proves 15 !A \lif (!B \lif !C)$.16 \begin{oltableau}17 [\pFmla{\True}{(\formula{A} \land \formula{B}) \lif \formula{C}}{1},18 just =\TAss19 [\pFmla{\False}{\formula{A} \lif (\formula{B} \lif \formula{C})}{1},20 just =\TAss21 [\pFmla{\True}{\formula{A}}{1.1},22 just = {\TRule{\False}{\lif}[2]}23 [\pFmla{\False}{\formula{B} \lif \formula{C}}{1.1},24 just = {\TRule{\False}{\lif}[2]}25 [\pFmla{\True}{\formula{B}}{1.1.1},26 just = {\TRule{\False}{\lif}[4]}27 [\pFmla{\False}{\formula{C}}{1.1.1},28 just = {\TRule{\False}{\lif}[4]}29 [\pFmla{\False}{\formula{A} \land \formula{B}}{1.1.1},30 just = {\TRule{\True}{\lif}[1]}31 [\pFmla{\False}{\formula{A}}{1.1.1},32 just= {\TRule{\False}{\land}[4]}, close]33 [\pFmla{\False}{\formula{B}}{1.1.1},34 just= {\TRule{\False}{\land}[4]}, close]]35 [\pFmla{\True}{\formula{C}}{1.1.1},36 just = {\TRule{\True}{\lif}[1]}, close]]37 ]38 ]39 ]40 ]41 ]42 \end{oltableau}43\end{ex}44454647\begin{prob}48 Find closed intuitionistic !!{tableau}s to show the following:49 \begin{enumerate}50 \item $\Proves !A \lif (!B \lif !A)$51 \item $\Proves \lnot(!A \land \lnot !A)$52 \item $!A \lif (!B \lif !C) \Proves (!A \land !B) \lif !C$53 \item $\lnot !A \lor \lnot !B \Proves \lnot(!A \land !B)$54 \end{enumerate}55\end{prob}5657\end{document}
content/intuitionistic-logic/tableaux/soundness.tex
1% Part: intuitionistic-logic2% Chapter: tableaux3% Section: soundness45\documentclass[../../../include/open-logic-section]{subfiles}67\begin{document}89\olfileid{int}{tab}{sou}1011\olsection{Soundness for Intuitionistic \usetoken{P}{tableau}}1213\begin{explain}14 In order to show that intuitionistic !!{tableau}s are sound, we have to15 show that if16 \[17 \sFmla{\True}{!B_1}[1], \dots, \sFmla{\True}{!B_n}[1], \sFmla{\False}{!A}[1]18 \]19 has a closed !!{tableau} then $!B_1, \dots, !B_n \Entails !A$. It is20 easier to prove the contrapositive: if for some $\mModel{M}$ and21 world~$w$, $\mSat{M}{!B_i}[w]$ for all $i=1$, \dots,~$n$ but22 $\mSat{M}{!A}[w]$, then no !!{tableau} can close. Such a23 countermodel shows that the initial assumptions of the !!{tableau}24 are satisfiable. The strategy of the proof is to show that whenever25 all the prefixed !!{formula}s on !!a{tableau} branch are26 satisfiable, any application of a rule results in at least one27 extended branch that is also satisfiable. Since closed branches are28 unsatisfiable, any !!{tableau} for a satisfiable set of prefixed29 !!{formula}s must have at least one open branch.3031 In order to apply this strategy in the modal case, we have to extend32 our definition of ``satisfiable'' to relational and prefixes. With33 that in hand, however, the proof is straightforward.34\end{explain}3536\begin{defn}37 Let $P$ be some set of prefixes, i.e., $P \subseteq (\PosInt)^*38 \setminus \{\emptyseq\}$ and let $\mModel{M}$ be a model. A39 function~$f\colon P \to W$ is an \emph{interpretation of}~$P$40 in~$\mModel{M}$ if, whenever $\sigma$ and $\sigma.n$ are both41 in~$P$, then $Rf(\sigma)f(\sigma.n)$.4243 Relative to an interpretation of prefixes~$P$ we can define:44 \begin{enumerate}45 \item $\mModel{M}$ satisfies $\sFmla{\True}{!A}[\sigma]$ iff46 $\mSat{M}{!A}[f(\sigma)]$.47 \item $\mModel{M}$ satisfies $\sFmla{\False}{!A}[\sigma]$ iff48 $\mSat/{M}{!A}[f(\sigma)]$.49 \end{enumerate}50\end{defn}5152Note that since $R$ is reflexive and transitive and $\sigma.{*}$53denotes, $\sigma$, $\sigma.n_1$, $\sigma.n_1.n_2$, \dots, we also have54that $Rf(\sigma)f(\sigma.{*})$. 5556\begin{defn}57 Let $\Gamma$ be a set of prefixed !!{formula}s, and let $P(\Gamma)$58 be the set of prefixes that occur in it. If $f$ is an interpretation59 of~$P(\Gamma)$ in $\mModel{M}$, we say that $\mModel{M}$ satisfies60 $\Gamma$ with respect to~$f$, $\mSat{M}{\Gamma}[f]$, if $\mModel{M}$61 satisfies every prefixed !!{formula} in~$\Gamma$ with respect62 to~$f$. $\Gamma$ is \emph{satisfiable} iff there is a63 model~$\mModel{M}$ and interpretation~$f$ of $P(\Gamma)$ such that64 $\mSat{M}{\Gamma}[f]$.65\end{defn}6667\begin{prop}68 If $\Gamma$ contains both $\sFmla{\True}{!A}[\sigma]$ and69 $\sFmla{\False}{!A}[\sigma.{*}]$ for some !!{formula}~$!A$ and70 prefix~$\sigma$, or it contains $\sFmla{\True}{\lfalse}[\sigma]$, then $\Gamma$ is unsatisfiable.71\end{prop}7273\begin{proof}74 Since always $\mSat/{M}{\lfalse}[f(\sigma)]$, a $\Gamma$ that75 contains $\sFmla{\True}{\lfalse}$ is unsatisfiable.7677 There also cannot be a model~$\mModel{M}$ and interpretation~$f$ of78 $P(\Gamma)$ such that both If $\mSat{M}{!A}[f(\sigma)]$, then by79 \olref[sem][rel]{prop:true-monotonic}, since80 $Rf(\sigma)(\sigma.{*})$, $\mSat{M}{!A}[f(\sigma)]$. So we cannot81 have both $\mSat{M}{!A}[f(\sigma)]$ and82 $\mSat/{M}{!A}[f(\sigma.{*})]$.83\end{proof}8485\begin{thm}[Soundness]86 \ollabel{thm:tableau-soundness}87 If $\Gamma$ has a closed !!{tableau}, $\Gamma$ is unsatisfiable.88\end{thm}8990\begin{proof}91We call a branch of !!a{tableau} satisfiable iff the set of92!!{signed formula}s on it is satisfiable, and let's call !!a{tableau}93satisfiable if it contains at least one satisfiable branch.9495We show the following: Extending a satisfiable !!{tableau} by one of96the rules of inference always results in a satisfiable !!{tableau}.97This will prove the theorem: any closed !!{tableau} results by98applying rules of inference to the !!{tableau} consisting only of99assumptions from~$\Gamma$. So if $\Gamma$ were satisfiable, any100!!{tableau} for it would be satisfiable. A closed !!{tableau},101however, is clearly not satisfiable, since all its branches are closed102and closed branches are unsatisfiable.103104Suppose we have a satisfiable !!{tableau}, i.e., !!a{tableau} with at105least one satisfiable branch. Applying a rule of inference either adds106!!{signed formula}s to a branch, or splits a branch in two. If the107!!{tableau} has a satisfiable branch which is not extended by the rule108application in question, it remains a satisfiable branch in the109extended !!{tableau}, so the extended tableau is satisfiable. So we110only have to consider the case where a rule is applied to a111satisfiable branch.112113Let $\Gamma$ be the set of !!{signed formula}s on that branch, and let114$\sFmla{S}{!A}[\sigma] \in \Gamma$ be the !!{signed formula} to which115the rule is applied. If the rule does not result in a split branch, we116have to show that the extended branch, i.e., $\Gamma$ together with117the conclusions of the rule, is still satisfiable. If the rule results118in split branch, we have to show that at least one of the two119resulting branches is satisfiable.120\begin{enumerate}121\item The branch is expanded by applying $\TRule{\True}{\lnot}$ to122 $\sFmla{\True}{\lnot !B}[\sigma] \in \Gamma$. Then the extended123 branch contains the !!{signed formula}s $\Gamma \cup124 \{\sFmla{\False}{!B}[\sigma.{*}]\}$. Suppose $\mSat{M}{\Gamma}[f]$.125 In particular, $\mSat{M}{\lnot !B}[f(\sigma)]$. Thus,126 $\mSat/{M}{!B}[w]$ for any $w$ such that $Rf(\sigma)w$, and that127 includes $f(\sigma.{*})$. So, $\mModel{M}$ satisfies128 $\sFmla{\False}{!B}[\sigma.{*}]$ with respect to~$f$.129\item The branch is expanded by applying $\TRule{\False}{\lnot}$ to130 $\sFmla{\False}{\lnot !B}[\sigma] \in \Gamma$: Exercise.131\item The branch is expanded by applying $\TRule{\True}{\land}$ to132 $\sFmla{\True}{!B \land !C}[\sigma] \in \Gamma$, which results in133 two new !!{signed formula}s on the branch:134 $\sFmla{\True}{!B}[\sigma]$ and $\sFmla{\True}{!C}[\sigma]$. Suppose135 $\mSat{M}{\Gamma}[f]$, in particular $\mSat{M}{!B \land136 !C}[f(\sigma)]$. Then $\mSat{M}{!B}[f(\sigma)]$ and137 $\mSat{M}{!C}[f(\sigma)]$. This means that $\mModel{M}$ satisfies138 both $\sFmla{\True}{!B}[\sigma]$ and $\sFmla{\True}{!C}[\sigma]$139 with respect to~$f$.140\item The branch is expanded by applying $\TRule{\False}{\lor}$ to141 $\sFmla{\False}{!B \lor !C} \in \Gamma$: Exercise.142\item The branch is expanded by applying $\TRule{\False}{\lif}$ to143 $\sFmla{\False}{!B \lif !C}[\sigma] \in \Gamma$: This results in two144 new !!{signed formula}s on the branch: $\sFmla{\True}{!B}[\sigma.n]$ and145 $\sFmla{\False}{!C}[\sigma.n]$, where $\sigma.n$ is a146 new prefix on the branch, i.e., $\sigma.n \notin P(\Gamma)$. 147 148 Since $\Gamma$ is satisfiable, there is a $\mModel{M}$ and149 interpretation~$f$ of $P(\Gamma)$ such that $\mSat{M}{\Gamma}[f]$, in150 particular $\mSat/{M}{!B \lif !C}[f(\sigma)]$. We have to show that151 $\Gamma \cup \{\sFmla{\False}{!B \lif !C}[\sigma.n]\}$ is152 satisfiable. To do this, we define an interpretation of $P(\Gamma)153 \cup \{\sigma.n\}$ as follows:154 155 Since $\mSat/{M}{!B \lif !C}[f(\sigma)]$, there is a $w \in W$ such156 that $Rf(\sigma)w$ such that $\mSat{M}{!B}[w]$ and157 $\mSat/{M}{!C}[w]$. Let $f'$ be like $f$, except that $f'(\sigma.n)158 = w$. Since $f'(\sigma) = f(\sigma)$ and $Rf(\sigma)w$, we have159 $Rf'(\sigma)f'(\sigma.n)$, so $f'$ is an interpretation of160 $P(\Gamma) \cup \{\sigma.n\}$. Obviously161 $\mSat{M}{!B}[f'(\sigma.n)]$ and $\mSat/{M}{!C}[f'(\sigma.n)]$.162 Since $f(\sigma') = f'(\sigma')$ for all prefixes $\sigma' \in163 P(\Gamma)$, $\mSat{M}{\Gamma}[f']$. So, $\mModel{M}, f'$ satisfies164 $\Gamma \cup \{\sFmla{\False}{!B \lif !C}[\sigma.n]\}$165\end{enumerate}166Now let's consider the possible inferences with two premises.167\begin{enumerate}168\item The branch is expanded by applying $\TRule{\False}{\land}$ to169 $\sFmla{\False}{!B \land !C}[\sigma] \in \Gamma$, which results in170 two branches, a left one continuing through171 $\sFmla{\False}{!B}[\sigma]$ and a right one through172 $\sFmla{\False}{!C}[\sigma]$. Suppose $\mSat{M}{\Gamma}[f]$, in173 particular $\mSat/{M}{!B \land !C}[f(\sigma)]$. Then174 $\mSat/{M}{!B}[f(\sigma)]$ or $\mSat/{M}{!C}[f(\sigma)]$. In the175 former case, $\mModel{M}, f$ satisfies $\sFmla{\False}{!B}[\sigma]$,176 i.e., the left branch is satisfiable. In the latter, $\mModel{M}, f$177 satisfies $\sFmla{\False}{!C}[\sigma]$, i.e., the right branch is178 satisfiable.179\item The branch is expanded by applying $\TRule{\True}{\lor}$ to180 $\sFmla{\True}{!B \lor !C}[\sigma] \in \Gamma$: Exercise.181\item The branch is expanded by applying $\TRule{\True}{\lif}$ to182 $\sFmla{\True}{!B \lif !C}[\sigma] \in \Gamma$: Exercise.183\end{enumerate}184\end{proof}185186\begin{prob}187Complete the proof of \olref[int][tab][sou]{thm:tableau-soundness}.188\end{prob}189190\begin{cor}191\ollabel{cor:entailment-soundness}192If $\Gamma \Proves !A$ then $\Gamma \Entails !A$.193\end{cor}194195\begin{proof}196 If $\Gamma \Proves !A$ then for some $!B_1$, \dots, $!B_n \in197 \Gamma$, $\Delta = \{\sFmla{\False}{!A}[1], \sFmla{\True}{!B_1}[1],198 \dots, \sFmla{\True}{!B_n}[1]\}$ has a closed !!{tableau}. We want199 to show that $\Gamma \Entails !A$. Suppose not, so for some200 $\mModel{M}$ and $w$, $\mSat{M}{!B_i}[w]$ for $i=1$, \dots,~$n$, but201 $\mSat/{M}{!A}[w]$. Let $f(1) = w$; then $f$ is an interpretation202 of~$P(\Delta)$ into~$\mModel{M}$, and $\mModel{M}$ satisfies~$\Delta$ with203 respect to~$f$. But by \olref{thm:tableau-soundness}, $\Delta$ is204 unsatisfiable since it has a closed !!{tableau}, a205 contradiction. So we must have $\Gamma \Proves !A$ after all.206\end{proof}207208\begin{cor}209\ollabel{cor:weak-soundness}210If $\Proves !A$ then $!A$ is true in all models.211\end{cor}212213\end{document}