% Part: propositional-logic % Chapter: propositional-logic % Section: preliminaries \documentclass[../../../include/open-logic-section]{subfiles} \begin{document} \olfileid{pl}{syn}{pre} \olsection{Preliminaries} \begin{thm}[\emph{Principle of induction on !!{formula}s}] \ollabel{thm:induction} If some property~$P$ holds for all the atomic !!{formula}s and is such that \begin{enumerate} \tagitem{prvNot}{it holds for $\lnot !A$ whenever it holds for~$!A$;}{} \tagitem{prvAnd}{it holds for $(!A \land !B)$ whenever it holds for $!A$ and~$!B$;}{} \tagitem{prvOr}{it holds for $(!A \lor !B)$ whenever it holds for $!A$ and~$!B$;}{} \tagitem{prvIf}{it holds for $(!A \lif !B)$ whenever it holds for $!A$ and~$!B$;}{} \tagitem{prvIff}{it holds for $(!A \liff !B)$ whenever it holds for $!A$ and~$!B$;}{} \end{enumerate} then $P$ holds for all !!{formula}s. \end{thm} \begin{proof} Let $S$ be the collection of all !!{formula}s with property~$P$. Clearly $S \subseteq \Frm[L_0]$. $S$~satisfies all the conditions of \olref[fml]{defn:formulas}: it contains all atomic !!{formula}s and is closed under the !!{operator}s. $\Frm[L_0]$ is the smallest such class, so $\Frm[L_0] \subseteq S$. So $\Frm[L_0] = S$, and every formula has property~$P$. \end{proof} \begin{prop}\ollabel{prop:balanced} Any !!{formula} in~$\Frm[L_0]$ is \emph{balanced}, in that it has as many left parentheses as right ones. \end{prop} \begin{prob} Prove \olref[pl][syn][pre]{prop:balanced} \end{prob} \begin{prop} \ollabel{prop:noinit} No proper initial segment of !!a{formula} is !!a{formula}. \end{prop} \begin{prob} Prove \olref[pl][syn][pre]{prop:noinit} \end{prob} \begin{prop}[Unique Readability] Any !!{formula}~$!A$ in $ \Frm[L_0]$ has exactly one parsing as one of the following \begin{enumerate} \tagitem{prvFalse}{$\lfalse$.}{} \tagitem{prvTrue}{$\ltrue$.}{} \item $\Obj p_n$ for some $\Obj p_n \in \PVar$. \tagitem{prvNot}{$\lnot !B$ for some !!{formula}~$!B$.}{} \tagitem{prvAnd}{$(!B \land !C)$ for some !!{formula}s $!B$ and~$!C$.}{} \tagitem{prvOr}{$(!B \lor !C)$ for some !!{formula}s $!B$ and~$!C$.}{} \tagitem{prvIf}{$(!B \lif !C)$ for some !!{formula}s $!B$ and~$!C$.}{} \tagitem{prvIff}{$(!B \liff !C)$ for some !!{formula}s $!B$ and~$!C$.}{} \end{enumerate} Moreover, this parsing is \emph{unique}. \end{prop} \begin{proof} By induction on $!A$. For instance, suppose that $!A$ has two distinct readings as $(!B \lif !C)$ and $(!B' \lif !C')$. Then $!B$ and $!B'$ must be the same (or else one would be a proper initial segment of the other); so if the two readings of $!A$ are distinct it must be because $!C$ and $!C'$ are distinct readings of the same sequence of symbols, which is impossible by the inductive hypothesis. \end{proof} \begin{defn}[Uniform Substitution] If $!A$ and $!B$ are !!{formula}s, and $\Obj p_i$ is a !!{propositional variable}, then $\Subst{!A}{!B}{\Obj p_i}$ denotes the result of replacing each occurrence of $\Obj p_i$ by an occurrence of $!B$ in $!A$; similarly, the simultaneous substitution of $\Obj p_1$, \dots,~$\Obj p_n$ by !!{formula}s $!B_1$, \dots,~$!B_n$ is denoted by $\SSubst{!A}{\subst{!B_1}{\Obj p_1},\dots,\subst{!B_n}{\Obj p_n}}$. \end{defn} \begin{prob} For each of the five !!{formula}s below determine whether the !!{formula} can be expressed as a substitution \( \Subst{!A}{!B}{\Obj p_i} \) where \( !A \) is (i) \( \Obj p_0 \); (ii) \( ( \lnot \Obj p_0 \land \Obj p_1) \); and (iii) \( ( ( \lnot \Obj p_0 \lif \Obj p_1 ) \land \Obj p_2 ) \). In each case specify the relevant substitution. \begin{enumerate} \item \( \Obj p_1 \) \item \( ( \lnot \Obj p_0 \land \Obj p_0 ) \) \item \( ( ( \Obj p_0 \lor \Obj p_1 ) \land \Obj p_2 ) \) \item \( \lnot ( ( \Obj p_0 \lif \Obj p_1 ) \land \Obj p_2 ) \) \item \( (( \lnot ( \Obj p_0 \lif \Obj p_1 ) \lif ( \Obj p_0 \lor \Obj p_1 )) \land \lnot ( \Obj p_0 \land \Obj p_1 )) \) \end{enumerate} \end{prob} \begin{prob} Give a mathematically rigorous definition of $\Subst{!A}{!B}{p}$ by induction. \end{prob} \end{document}