content/applied-modal-logic/epistemic-logic/epistemic-logic.tex
1% Part: applied-modal-logics2% Chapter: epistemic-logic34\documentclass[../../../include/open-logic-chapter]{subfiles}56\begin{document}78\olchapter{aml}{el}{Epistemic Logics}910\begin{editorial}11 This chapter covers the metatheory of epistemic logics. It is structured in a similar way to Aldo Antonelli's notes on classical basic modal logic, but has been rewritten by Audrey Yap in order to add material on bisimulation and dynamic epistemic logics.12\end{editorial}1314\olimport{introduction}15\olimport{language-epistemic-logic}16\olimport{relational-models}17\olimport{truth-at-w}18\olimport{properties-accessibility}19\olimport{bisimulations}20\olimport{public-announcement-logic-lang}21\olimport{public-announcement-logic-semantics}222324\OLEndPartHook2526\end{document}
content/applied-modal-logic/epistemic-logic/introduction.tex
1% Part: applied-modal-logics2% Chapter: epistemic-logic3% Section: introduction45\documentclass[../../../include/open-logic-section]{subfiles}67\begin{document}89\olfileid{aml}{el}{int}1011\olsection{Introduction}1213Just as modal logic deals with \emph{modal propositions} and the entailment relations among them, epistemic logic deals with \emph{epistemic propositions} and the entailment relations among them. Rather than interpreting the modal operators as representing possibility and necessity, the unary connectives are interpreted in epistemic or doxastic ways, to model knowledge and belief. For example, we might want to express claims like the following:1415\begin{enumerate}16\item Richard knows that Calgary is in Alberta.17\item Audrey thinks it is possible that a dog is on the couch.18\item Richard knows that Audrey knows that her class is on Tuesdays.19\item Everyone knows that a year has 12 months.20\end{enumerate}21Contemporary epistemic logic is often traced to Jaako Hintikka's \emph{Knowledge and Belief}, from 1962, and it was written at a time when possible worlds semantics were becoming increasingly more used in logic. In fact, epistemic logics use most of the same semantic tools as other modal logics, but will interpret them differently. The main change is in what we take the \emph{accessibility relation} to represent. In epistemic logics, they represent some form of \emph{epistemic possibility}. We'll see that the epistemic notion that we're modelling will affect the constraints that we want to place on the accessibility relation. And we'll also see what happens to correspondence theory when it is given an epistemic interpretation. You'll notice that the examples above mention two agents: Richard and Audrey, and the relationship between the things that each one knows. The epistemic logics we'll consider will be multi-agent logics, in which such things can be expressed. In contrast, a single-agent epistemic logic would only talk about what one individual knows or believes.2223\end{document}
content/applied-modal-logic/epistemic-logic/language-epistemic-logic.tex
1% Part: applied-modal-logics2% Chapter: epistemic-logic3% Section: language-epistemic-logic45\documentclass[../../../include/open-logic-section]{subfiles}67\begin{document}89\olfileid{aml}{el}{lan}1011\olsection{The Language of Epistemic Logic}1213\begin{defn}14Let $G$ be a set of agent-symbols. The basic language of multi-agent15epistemic logic contains16\begin{enumerate}17 \tagitem{prvFalse}{The propositional constant for !!{falsity}~$\lfalse$.}{}18 \tagitem{prvTrue}{The propositional constant for !!{truth}~$\ltrue$.}{}19 \item A !!{denumerable}s set of !!{propositional variable}s: $\Obj20 p_0$, $\Obj p_1$, $\Obj p_2$, \dots21 \item The propositional connectives: \startycommalist22 \iftag{prvNot}{\ycomma $\lnot$ (negation)}{}%23 \iftag{prvAnd}{\ycomma $\land$ (conjunction)}{}%24 \iftag{prvOr}{\ycomma $\lor$ (disjunction)}{}%25 \iftag{prvIf}{\ycomma $\lif$ (!!{conditional})}{}%26 \iftag{prvIff}{\ycomma $\liff$ (!!{biconditional})}{}.27 \item The knowledge operator $\Knows_a$ where $a \in G$.28\end{enumerate}29\end{defn}3031If we are only concerned with the knowledge of a single agent in our32system, we can drop the reference to the set~$G$, and individual33agents. In that case, we only have the basic operator~$\Knows$.3435\begin{defn}36\emph{!!^{formula}s} of the epistemic language are inductively37 defined as follows:38\begin{enumerate}39\tagitem{prvFalse}{$\lfalse$ is an atomic !!{formula}.}{}4041\tagitem{prvTrue}{$\ltrue$ is an atomic !!{formula}.}{}4243\item Every propositional variable $\Obj p_i$ is an (atomic) !!{formula}.4445\tagitem{prvNot}{If $!A$ is !!a{formula}, then $\lnot !A$ is46 !!a{formula}.}{}4748\tagitem{prvAnd}{If $!A$ and $!B$ are !!{formula}s, then $(!A \land49 !B)$ is !!a{formula}.}{}5051\tagitem{prvOr}{If $!A$ and $!B$ are !!{formula}s, then $(!A \lor !B)$52 is !!a{formula}.}{}5354\tagitem{prvIf}{If $!A$ and $!B$ are !!{formula}s, then $(!A \lif !B)$55 is !!a{formula}.}{}5657\tagitem{prvIff}{If $!A$ and $!B$ are !!{formula}s, then $(!A \liff !B)$58 is !!a{formula}.}{}5960\item If $!A$ is !!a{formula} and $a \in G$, then $\Knows_a !A$ is61 !!a{formula}.6263\tagitem{limitClause}{Nothing else is !!a{formula}.}{}64\end{enumerate}65\end{defn}6667If !!a{formula}~$!A$ does not contain $\Knows_a$, we say it68is \emph{modal-free}.6970\begin{defn}71 While the $\Knows$ operator is intended to symbolize individual72 knowledge, $\EKnows$, often read as ``everybody knows,'' symbolizes73 group knowledge. Where $G' \subseteq G$, we define $\EKnows_{G'} !A$74 as an abbreviation for \[\bigwedge_{b \in G'} \Knows_b !A.\]75\end{defn}7677We can also define an even stronger sense of knowledge, namely78\emph{common knowledge} among a group of agents~$G$. When a piece of79information is common knowledge among a group of agents, it means that80for every combination of agents in that group, they all know that each81other knows that each other knows \dots ad infinitum. This is82significantly stronger than group knowledge, and it is easy to come up83with relational models in which !!a{formula} is group knowledge, but84not common knowledge. We will use $\CKnows_G !A$ to symbolize ``it is85common knowledge among~$G$ that~$!A$.''8687\end{document}
content/applied-modal-logic/epistemic-logic/relational-models.tex
1% Part: applied-modal-logics2% Chapter: epistemic-logic3% Section: relational-models45\documentclass[../../../include/open-logic-section]{subfiles}67\begin{document}89\olfileid{aml}{el}{rel}1011\olsection{Relational Models}1213The basic semantic concept for epistemic logics is the same as that of14ordinary modal logics. Relational models still consist of a set of15worlds, and an assignment that determines which !!{propositional16variable}s count as ``true'' at which worlds. And if we are only17dealing with a single agent, we have a single accessibility relation18as usual. However, if we have a multi-agent epistemic logic, then our19single accessibility relation becomes a set of accessibility20relations, one for each~$a$ in our set of agent symbols~$G$.2122A \emph{relational model} consists of a set of worlds, which are23related by binary accessibility relations---one for each24agent---together with an assignment which determines which25!!{propositional variable}s are true at which worlds.2627\begin{defn}28 A \emph{model} for the multi-agent epistemic language is a triple29 $\mModel{M} = \tuple{W, R, V}$, where30 \begin{enumerate}31 \item $W$ is a nonempty set of ``worlds,''32 \item For each $a \in G$, ${R}_a$ is a binary accessibility relation33 on~$W$, and34 \item $V$ is a function assigning to each !!{propositional35 variable}~$p$ a set $V(p)$ of possible worlds.36 \end{enumerate}37 When $R_a ww'$ holds, we say that $w'$ is \emph{accessible by a38 from}~$w$. When $w \in V(p)$ we say $p$ is \emph{true at}~$w$.39\end{defn}4041The mechanics are just like the mechanics for normal modal logic, just42with more accessibility relations added in. For a given agent, we will43generally interpret their accessibility relation as representing44something about their informational states. For example, we often45treat $R_a ww'$, as expressing that $w'$ is consistent with~$a$'s46information at~$w$. Or to put it another way, at~$w$, they cannot tell47the difference between world $w$ and world~$w'$. 4849\end{document}
content/applied-modal-logic/epistemic-logic/truth-at-w.tex
1% Part: applied-modal-logics2% Chapter: epistemic-logic3% Section: truth-at-w45\documentclass[../../../include/open-logic-section]{subfiles}67\begin{document}89\olfileid{aml}{el}{trw}1011\olsection{Truth at a World}1213Just as with normal modal logic, every epistemic model determines which !!{formula}s count as true at14which worlds in it. We use the same notation ``model $\mModel{M}$ makes15!!{formula}~$!A$ true at world~$w$'' for the basic notion of relational16semantics. The relation is defined inductively and is identical to the normal modal case for all non-modal operators.1718\begin{defn}\ollabel{defn:mmodels}19 \emph{Truth of !!a{formula}~$!A$ at~$w$} in a~$\mModel M$, in symbols:20 $\mSat{M}{!A}[w]$, is defined inductively as follows:21 \begin{enumerate}22 \tagitem{prvFalse}{\indcase{!A}{\lfalse}{Never $\mSat{M}{\lfalse}[w]$}.}{}23 \tagitem{prvTrue}{\indcase{!A}{\ltrue}{Always $\mSat{M}{\ltrue}[w]$}.}{}24 \item $\mSat{M}{p}[w]$ iff $w \in V(p)$25 \tagitem{prvNot}{\indcase{!A}{\lnot !B}{$\mSat{M}{\indfrm}[w]$ iff26 $\mSat/{M}{!B}[w]$}.}{}27 \tagitem{prvAnd}{\indcase{!A}{(!B \land !C)}{$\mSat{M}{\indfrm}[w]$ iff28 $\mSat{M}{!B}[w]$ and $\mSat{M}{!C}[w]$}.}{}29 \tagitem{prvOr}{\indcase{!A}{(!B \lor !C)}{$\mSat{M}{\indfrm}[w]$ iff30 $\mSat{M}{!B}[w]$ or $\mSat{M}{!C}[w]$} (or both).}{}31 \tagitem{prvIf}{\indcase{!A}{(!B \lif !C)}{$\mSat{M}{\indfrm}[w]$ iff32 $\mSat/{M}{!B}[w]$ or $\mSat{M}{!C}[w]$}.}{}33 \tagitem{prvIff}{\indcase{!A}{(!B \liff !C)}{$\mSat{M}{\indfrm}[w]$ iff34 either both $\mSat{M}{!B}[w]$ and $\mSat{M}{!C}[w]$ or35 neither $\mSat{M}{!B}[w]$ nor $\mSat{M}{!C}[w]$}.}{}36 \item\ollabel{defn:sub:mmodels-box}37 \indcase{!A}{\Knows_a !B}{$\mSat{M}{\indfrm}[w]$ iff38 $\mSat{M}{!B}[w']$ for all $w' \in W$ with $R_a ww'$}39 \end{enumerate} 40\end{defn}4142Here's where we need to think about restrictions on our accessibility43relations, though. After all, by clause~\olref{defn:sub:mmodels-box},44!!a{formula}~$\Knows_a !B$ is true at~$w$ whenever there are no~$w'$45with $R_a ww'$. This is the same clause as in normal modal logic; when46a world has no successors, all $\Box$-formulas are vacuously true47there. This seems extremely counterintuitive if we think about48$\Knows$ as representing \emph{knowledge}. After all, we tend to think49that there are \emph{no} circumstances under which an agent might know50both $!A$ and $\lnot !A$ at the same time.5152One solution is to ensure that our accessibility relation in epistemic53logic will always be \emph{reflexive}. This roughly corresponds to the54idea that the actual world is consistent with an agent's information.55In fact, epistemic logics typically use S5, but others might use56weaker systems depending on what exactly they want the $\Knows_a$57relation to represent.5859\begin{figure}60 \begin{center}61 \begin{tikzpicture}[modal]62 \node[world] (w1) [label={[align=right]right:\mTrue{p}\\ \mFalse{q}}]{$w_1$}; 63 \node[world] (w2) [label={[align=right]right:\mTrue{p}\\ \mTrue{q}},64 above right=of w1]{$w_2$}; 65 \node[world] (w3) [label={[align=right]right:\mFalse{p}\\ \mFalse{q}},66 below right=of w1] {$w_3$};67 \draw[<->] (w1) to node {$a$} (w2);68 \draw[->,loop left] (w1) to node {$a,b$} (w1);69 \draw[<->] (w1) to [swap] node {$b$} (w3);70 \draw[->,loop above] (w2) to node {$a,b$} (w2) ;71 \draw[->,loop below] (w3) to node {$a,b$} (w3) ;72 \end{tikzpicture}73 \end{center}74 \caption{A simple epistemic model.}75 \ollabel{fig:simple}76\end{figure}7778\begin{prob}79 Consider which of the following hold in80 \olref[aml][el][trw]{fig:simple}:81 \begin{enumerate}82 \item $\mSat{M}{\lnot q}[w_1]$;83 \item $\mSat{M}{\Knows_a \lnot q}[w_1]$;84 \item $\mSat{M}{\Knows_b \lnot q}[w_1]$;85 \item $\mSat{M}{\Knows_b q \lor \Knows_b \lnot q}[w_2]$;86 \item $\mSat{M}{\Knows_a( \Knows_b q \lor \Knows_b \lnot q)}[w_2]$;87 \item $\mSat{M}{\EKnows_{\{a,b\}} \lnot q}[w_3]$;88 \end{enumerate}89\end{prob}9091Now that we have given our basic definition of truth at a world, the92other semantic concepts from normal modal logic, such as modal93validity and entailment, simply carry over, applied to this new way of94thinking about the interpretation for the modal operators.9596We are now also in a position to give truth conditions for the common97knowledge operator~$\CKnows_G$. Recall from \olref[sfr][rel][ops]{sec}98that the \emph{transitive closure}~$R^+$ of a relation~$R$ is defined99as100\begin{align*}101 R^+ &= \bigcup_{ n \in \mathbb{N}} R^n, 102\intertext{where}103 R^0 & = R \text{ and}\\ 104 R^{n+1} & = \Setabs{\tuple{x, z}}{\exists y (R^n xy \land Ryz)}.105\end{align*}106Then, where $G$ is a group of agents, we define $R_G = ( \bigcup_{b107\in G} R_b )^+$ to be the transitive closure of the union of all108agents' accessibility relations.109110\begin{defn}111If $G' \subseteq G$, we let $\mSat{M}{\CKnows_{G'} !A}[ w]$ iff for112every $w'$ such that $R_{G'} w w'$, $\mSat{M}{!A}[w']$.113\end{defn}114115\end{document}
content/applied-modal-logic/epistemic-logic/properties-accessibility.tex
1% Part: applied-modal-logics2% Chapter: epistemic-logic3% Section: properties-accessibility45\documentclass[../../../include/open-logic-section]{subfiles}67\begin{document}89\olfileid{aml}{el}{acc}1011\olsection{Accessibility Relations and Epistemic Principles}1213Given what we already know about frame correspondence in normal modal14logics, we might want to see what the characteristic !!{formula}s look15like given epistemic interpretations. We have already said that16epistemic logics are typically interpreted in S5. So let's take a look17at how various epistemic principles are represented, and consider how18they correspond to various frame conditions.1920Recall from normal modal logic, that different modal !!{formula}s21characterized different properties of accessibility relations. This22table picks out a few that correspond to particular epistemic23principles.2425\begin{table}[t]26 \begin{tabular}{| p{.48\textwidth} || p{.48\textwidth} |}27 \hline28 {\emph{If $R$ is \dots}} & {\emph{then \dots is true in~$\mModel{M}$:}} \\29 \hline \hline30 31 & $\Knows (p \lif q) \lif (\Knows p \lif \Knows q)$ 32 \hfill \newline{(Closure)} \\33 \hline34 \emph{reflexive}: $\forall w Rww$ 35 & $\Knows p \lif p$ \hfill \newline{(Veridicality)} \\36 \hline37 \emph{transitive}: \newline38 $\forall u \forall v \forall w ((Ruv \land Rvw) \lif Ruw)$ & 39 $\Knows p \lif \Knows \Knows p$ \hfill 40 \newline{(Positive Introspection)} \\41 \hline 42 \emph{euclidean}: \newline43 $\forall w \forall u \forall v ((Rwu \land Rwv) \lif Ruv)$ & 44 $\lnot \Knows p \lif \Knows \neg \Knows p$ \hfill 45 \newline{(Negative Introspection)} \\46 \hline47 \end{tabular}48 \caption{Four epistemic principles.}49 \ollabel{tab:four}50 \end{table} 5152Veridicality, corresponding to the $T$ axiom, is often treated as the53most uncontroversial of these principles, as it represents that claim54that if !!a{formula} is known, then it must be true. Closure, as well55as Positive and Negative Introspection are much more contested. 5657Closure, corresponding to the $K$ axiom, represents the idea that an58agent's knowledge is closed under implication. This might seem59plausible to us in some cases. For instance, I might know that if I am60in Victoria, then I am on Vancouver Island. Barring odd skeptical61scenarios, I do know that I am in Victoria, and this should also62suggest that I know I am on Vancouver Island. So in this case, the63logical closure of my knowledge might seem relatively intuitive. On64the other hand, we do not always think through the consequences of our65knowledge, and so this might lead to less intuitive results in other66cases.6768Positive Introspection, sometimes known as the KK-principle, is69sometimes articulated as the statement that if I know something, then70I know that I know. It is the epistemic counterpart of the 4 axiom.71Correspondingly, negative introspection is articulated as the72statement that if I \emph{don't} know something, then I know that I73don't know it, which is the counterpart of the 5 axiom. Both of these74seem to admit of relatively ordinary counterexamples, in which I am75unsure whether or not I know something that I do in fact know. 767778\end{document}
content/applied-modal-logic/epistemic-logic/bisimulations.tex
1% Part: applied-modal-logics2% Chapter: epistemic-logic3% Section: bisimulations45\documentclass[../../../include/open-logic-section]{subfiles}67\begin{document}89\olfileid{aml}{el}{bsd}1011\olsection{Bisimulations}1213One remaining question that we might have about the expressive power14of our epistemic language has to do with the relationship between15models and the formulas that hold in them. We have seen from our frame16correspondence results that when certain formulas are valid in a17frame, they will also ensure that those frames satisfy certain18properties. But does our modal language, for example, allow us to19distinguish between a world at which there is a reflexive arrow, and20an infinite chain of worlds, each of which leads to the next? That is,21is there any formula $A$ that might hold at only one of these two22worlds?2324Bisimulation is a relationship that we can define between relational25models to say that they have effectively the same structure. And as we26will see, it will capture a sense of equivalence between models that27can be captured in our epistemic language.2829\begin{defn}[Bisimulation]30Let $M_1 = \tuple{W_1, R_1, V_1}$ and $M_2 = \tuple{W_2, R_2, V_2}$ be31two relational models. And let $\mathcal{R} \subseteq W_1 \times W_2$32be a binary relation. We say that $\mathcal{R}$ is a33\emph{bisimulation} when for every $\tuple{w_1, w_2} \in \mathcal{R}$,34we have:3536\begin{enumerate}37 \item $w_1 \in V_1(p)$ iff $w_2 \in V_2(p)$ for all38 !!{propositional variable}s~$p$.39 \item For all agents $a \in A$ and worlds $v_1 \in W_1$, if40 $R_{1_a} w_1 v_1$ then there is some $v_2 \in W_2$ such that41 $R_{2_a} w_2 v_2$, and $\tuple{v_1, v_2} \in42 \mathcal{R}$.43 \item For all agents $a \in A$ and worlds $v_2 \in W_2$, if44 $R_{2_a} w_2 v_2$ then there is some $v_1 \in W_1$ such that45 $R_{1_a} w_1 v_1$, and $\tuple{v_1, v_2} \in46 \mathcal{R}$.47\end{enumerate}4849When there is a bisimulation between $M_1$ and $M_2$ that links worlds50$w_1$ and $w_2$, we can also write $\tuple{M_1, w_1} \leftrightarroweq51\tuple{M_2, w_2}$, and call $\tuple{M_1, w_1}$ and $\tuple{M_2, w_2}$52\emph{bisimilar}.53\end{defn}5455The different clauses in the bisimulation relation ensure different56things. Clause 1 ensures that bisimilar worlds will satisfy the same57modal-free formulas, since it ensures agreement on all58!!{propositional variable}s. The other two clauses, sometimes referred59to as ``forth'' and ``back,'' respectively, ensure that the60accessibility relations will have the same structure.6162\begin{thm}63If $\tuple{M_1, w_1} \leftrightarroweq \tuple{M_2, w_2}$, then for64every !!{formula}~$!A$, we have that $\mSat{M_1}{!A}[w_1]$ iff65$\mSat{M_2}{!A}[w_2]$.66\end{thm}6768\begin{figure}69 \begin{center}70 \begin{tikzpicture}[modal]71 \node[world] (w1) {$w_1$}; 72 \node[world] (w2) [above left=of w1]{$w_2$}; 73 \node[world] (w3) [above right=of w1] {$w_3$};74 \draw[<->] (w1) to node {$a$} (w2);75 \draw[->,loop below] (w1) to node {$a$} (w1);76 \draw[<->] (w1) to [swap] node {$a$} (w3);77 \draw[->,loop above] (w2) to node {$a$} (w2) ;78 \draw[->,loop above] (w3) to node {$a$} (w3) ;79 80 \node[world](v1)[right of=w1, xshift=1.5in]{$v_1$};81 \node[world](v2)[above of=v1]{$v_2$};82 \draw[<->] (v1) to node {$a$} (v2);83 \draw[->,loop below] (v1) to node {$a$} (v1);84 \draw[->,loop above] (v2) to node {$a$} (v2) ;8586 \draw[-,dotted] (w1) to (v1) ;87 \draw[-,dotted,bend left=45] (w2) to (v2) ;88 \draw[-,dotted,bend left=30] (w3) to (v2) ;89 \end{tikzpicture}90 \end{center}91 \caption{Two bisimilar models.}92 \ollabel{fig:bisimilar}93\end{figure}9495Even though the two models pictured in \olref{fig:bisimilar} aren't96quite the same as each other, there is a bisimulation linking worlds97$w_1$ and~$v_1$. This bisimulation will also link both $w_2$ and $w_3$98to~$v_2$, with the idea being that there is nothing expressible in our99modal language that can really distinguish between them. The situation100would be different if $w_2$ and~$w_3$ satisfied different101!!{propositional variable}s, however. 102103\end{document}
content/applied-modal-logic/epistemic-logic/public-announcement-logic-lang.tex
1% Part: applied-modal-logics2% Chapter: epistemic-logic3% Section: language-epistemic-logic45\documentclass[../../../include/open-logic-section]{subfiles}67\begin{document}89\olfileid{aml}{el}{pal}1011\olsection{Public Announcement Logic}1213Dynamic epistemic logics allow us to represent the ways in which14agents' knowledge changes over time, or as they gain new information.15Many of these represent changes in knowlege using informational16\emph{events} or \emph{updates}. The most basic kind of update is a17public announcement in which some formula is truthfully announced and18all of the agents witness this taking place together. To do this, we19expand the language as follows2021\begin{defn}22Let $G$ be a set of agent-symbols. The basic language of multi-agent23epistemic logic with public announcements contains24\begin{enumerate}25 \tagitem{prvFalse}{The propositional constant for !!{falsity}~$\lfalse$.}{}26 \tagitem{prvTrue}{The propositional constant for !!{truth}~$\ltrue$.}{}27 \item A !!{denumerable}s set of !!{propositional variable}s: $\Obj28 p_0$, $\Obj p_1$, $\Obj p_2$, \dots29 \item The propositional connectives: \startycommalist30 \iftag{prvNot}{\ycomma $\lnot$ (negation)}{}%31 \iftag{prvAnd}{\ycomma $\land$ (conjunction)}{}%32 \iftag{prvOr}{\ycomma $\lor$ (disjunction)}{}%33 \iftag{prvIf}{\ycomma $\lif$ (!!{conditional})}{}%34 \item The knowledge operator $\Knows_a$ where $a \in G$.35 \item The public announcement operator $[!B]$ where $!B$ is !!a{formula}.36\end{enumerate}37\end{defn}3839The public announcement operator functions as a box operator, and our40inductive definition of the language is given accordingly:4142\begin{defn}43 \emph{!!^{formula}s} of the epistemic language are inductively44 defined as follows:45 \begin{enumerate}46 \tagitem{prvFalse}{$\lfalse$ is an atomic !!{formula}.}{}4748 \tagitem{prvTrue}{$\ltrue$ is an atomic !!{formula}.}{}4950 \item Every propositional variable $\Obj p_i$ is an (atomic)51 !!{formula}.5253 \tagitem{prvNot}{If $!A$ is !!a{formula}, then $\lnot !A$ is54 !!a{formula}.}{}5556 \tagitem{prvAnd}{If $!A$ and $!B$ are !!{formula}s, then $(!A \land57 !B)$ is !!a{formula}.}{}5859 \tagitem{prvOr}{If $!A$ and $!B$ are !!{formula}s, then $(!A \lor !B)$60 is !!a{formula}.}{}6162 \tagitem{prvIf}{If $!A$ and $!B$ are !!{formula}s, then $(!A \lif !B)$63 is !!a{formula}.}{}6465 \tagitem{prvIff}{If $!A$ and $!B$ are !!{formula}s, then $(!A \liff !B)$66 is !!a{formula}.}{}6768 \item If $!A$ is !!a{formula} and $a \in G$, then $\Knows_a !A$ is69 !!a{formula}.70 71 \item If $!A$ and $!B$ are !!{formula}s, then $[!A] !B$ is !!a{formula}.7273 \tagitem{limitClause}{Nothing else is !!a{formula}.}{}74\end{enumerate}75\end{defn}7677The intended reading of the !!{formula} $[!A] !B$ is ``After $!A$ is78truthfully announced, $!B$~holds. It will sometimes also be useful to79talk about common knowledge in the context of public announcements, so80the language may also include the common knowledge operator~$\CKnows_G81!A$.8283\end{document}
content/applied-modal-logic/epistemic-logic/public-announcement-logic-semantics.tex
1% Part: applied-modal-logics2% Chapter: epistemic-logic3% Section: public-announcement-logic-semantics45\documentclass[../../../include/open-logic-section]{subfiles}67\begin{document}89\olfileid{aml}{el}{psm}1011\olsection{Semantics of Public Announcement Logic}1213Relational models for public announcement logics are the same as they14were in epistemic logics. However, the semantics for the public15announcement operator are something new.1617\begin{defn}\ollabel{defn:mmodels} \emph{Truth of !!a{formula}~$!A$18 at~$w$} in a~$\mModel M = \tuple{W, R, V}$, in symbols:19 $\mSat{M}{!A}[w]$, is defined inductively as follows:20 \begin{enumerate}21 \tagitem{prvFalse}{\indcase{!A}{\lfalse}{Never $\mSat{M}{\lfalse}[w]$}.}{}22 \tagitem{prvTrue}{\indcase{!A}{\ltrue}{Always $\mSat{M}{\ltrue}[w]$}.}{}23 \item $\mSat{M}{p}[w]$ iff $w \in V(p)$24 \tagitem{prvNot}{\indcase{!A}{\lnot !B}{$\mSat{M}{\indfrm}[w]$ iff25 $\mSat/{M}{!B}[w]$}.}{}26 \tagitem{prvAnd}{\indcase{!A}{(!B \land !C)}{$\mSat{M}{\indfrm}[w]$ iff27 $\mSat{M}{!B}[w]$ and $\mSat{M}{!C}[w]$}.}{}28 \tagitem{prvOr}{\indcase{!A}{(!B \lor !C)}{$\mSat{M}{\indfrm}[w]$ iff29 $\mSat{M}{!B}[w]$ or $\mSat{M}{!C}[w]$} (or both).}{}30 \tagitem{prvIf}{\indcase{!A}{(!B \lif !C)}{$\mSat{M}{\indfrm}[w]$ iff31 $\mSat/{M}{!B}[w]$ or $\mSat{M}{!C}[w]$}.}{}32 \tagitem{prvIff}{\indcase{!A}{(!B \liff !C)}{$\mSat{M}{\indfrm}[w]$ iff33 either both $\mSat{M}{!B}[w]$ and $\mSat{M}{!C}[w]$ or34 neither $\mSat{M}{!B}[w]$ nor $\mSat{M}{!C}[w]$}.}{}35 \item\ollabel{defn:sub:mmodels-box}36 \indcase{!A}{\Knows_a !B}{$\mSat{M}{\indfrm}[w]$ iff37 $\mSat{M}{!B}[w']$ for all $w' \in W$ with $R_a ww'$}38 \item\ollabel{defn:sub:mmodels-pal}39 \indcase{!A}{[!B] !C}{$\mSat{M}{\indfrm}[w]$ iff40 $\mSat{M}{!B}[w]$ implies $\mSat{M \mid !B}{!C}[w]$}4142 43 Where $\mModel M \mid !B = \tuple{W', R', V'}$ is defined as44 follows:45 \begin{enumerate}46 \item $W' = \Setabs{ u \in W}{\mSat{M}{!B}[u]}$. So the worlds of47 $\mModel M \mid !B$ are the worlds in~$\mModel M$ at which $!B$48 holds.49 \item $R'_a = R_a \cap (W' \times W')$. Each agent's accessibility50 relation is simply restricted to the worlds that remain in~$W'$.51 \item $V'(p) = \Setabs{ u \in W'}{u \in V(p) }$. Similarly, the52 propositional valuations at worlds remain the same, representing53 the idea that informational events will not change the truth54 value of propositional variables.55 \end{enumerate}56 57 \end{enumerate} 58\end{defn}5960What is distinctive, then, about public announcement logics, is that61the truth of !!a{formula} at~$\mModel M$ can sometimes only be decided62by referring to a model other than~$\mModel M$ itself.6364Notice also that our semantics treats the announcement operator as a65$\Box$~operator, and so if !!a{formula}~$!A$ cannot be truthfully66announced at a world, then $[!A]B$ will hold there trivially, just as67all $\Box$ !!{formula}s hold at endpoints. 6869\begin{figure}70 \begin{center}71 \begin{tikzpicture}[modal]72 \node[world] (w1) [label={below left:$p, \lnot q$}]{$w_1$}; 73 \node[world] (w2) [left=of w1, label={left:$\lnot p, \lnot q$}]{$w_2$}; 74 \node[world] (w3) [above=of w1, label={right:$p, q$}] {$w_3$};75 \draw[<->] (w1) to [swap] node {$b$} (w2);76 \draw[->,loop below] (w1) to node {$a, b$} (w1);77 \draw[<->] (w1) to [swap] node {$a$} (w3);78 \draw[->,loop above] (w2) to node {$a, b$} (w2) ;79 \draw[->,loop above] (w3) to node {$a, b$} (w3) ;80 81 \node[world](v1)[right of=w1, xshift=1.25in, label={right:$p, \lnot q$}]{$w'_1$};82 \node[world](v3)[above of=v1,label={right:$p,q$}]{$w'_3$};83 \draw[<->] (v1) to node {$a$} (v3);84 \draw[->,loop below] (v1) to node {$a,b$} (v1);85 \draw[->,loop above] (v3) to node {$a,b$} (v3) ;86 87 \node(m)[below=of w1]{$\mModel M$};88 \node(m')[below=of v1]{$\mModel M \mid p$}; 8990 \draw[-,dotted] (w1) to node {\footnotesize announcement of $p$}(v1) ;91 92 \end{tikzpicture}93 \end{center}94 \caption{Before and after the public announcement of $p$.}95 \ollabel{fig:announcement-example}96\end{figure}9798We can see the public announcement of !!a{formula} as shrinking a99model, or restricting it to the worlds at which the !!{formula} was100true. \olref{fig:announcement-example} gives an example of the effects101of publicly announcing~$p$. One notable thing about that model is that102agent~$b$ learns that~$p$ as a result of the announcement, while103agent~$a$ does not (since $a$ already knew that~$p$ was true).104105More formally, we have $\mSat{M}{\lnot \Knows_b p}[w_1]$ but $\mSat{M106\mid p}{\Knows_b p}[w'_1]$. This implies that $\mSat{M}{[p] \Knows_b107p}[w_1]$. But we have some even stronger claims that we can make about108the result of the announcement. In fact, it is the case that109$\mSat{M}{[p]\CKnows_{\{a,b\}} p}[w_1]$. In other words, after $p$ is110announced, it becomes \emph{common knowledge}.111112We might wonder, though, whether this holds in the general case, and113whether a truthful announcement of~$!A$ will \emph{always} result in114$!A$ becoming common knowledge. It may be surprising that the answer115is no. And in fact, it is possible to truthfully announce !!{formula}s116that will no longer be true once they are announced. For example,117consider the effects of announcing $p \land \lnot \Knows_b p$ at~$w_1$118in \olref{fig:announcement-example}. In fact, $\mModel M \mid p$ and119$\mModel M \mid (p \land \lnot \Knows_b p)$ are the same model.120However, as we have already noted, $\mSat{M \mid p}{\Knows_b121p}[w'_1]$. Therefore, $\mSat{M \mid (p \land \lnot \Knows_b p)}{\lnot122(p \land \lnot \Knows_b p)}[w'_1]$, so this is !!a{formula} that123becomes false once it has been announced. 124125\end{document}