content/many-valued-logic/three-valued-logics/three-valued-logics.tex
1% Part: many-valued-logic2% Chapter: three-valued-logics34\documentclass[../../../include/open-logic-chapter]{subfiles}56\begin{document}78\olchapter{mvl}{thr}{Three-valued Logics}910\olimport{introduction}1112\olimport{lukasiewicz}1314\olimport{kleene}1516\olimport{goedel}1718\olimport{multiple-designation}1920\OLEndChapterHook2122\end{document}
content/many-valued-logic/three-valued-logics/introduction.tex
1% Part: many-valued-logic2% Chapter: three-valued-logics3% Section: introduction45\documentclass[../../../include/open-logic-section]{subfiles}67\begin{document}89\olfileid{mvl}{thr}{int}1011\olsection{Introduction}1213If we just add one more value~$\Undef$ to $\True$ and $\False$, we get a14three-valued logic. Even though there is only one more truth value,15the possibilities for defining the truth-functions for $\lnot$,16$\land$, $\lor$, and $\lif$ are quite numerous. Then a logic might use17any combination of these truth functions, and you also have a choice18of making only $\True$ designated, or both $\True$ and~$\Undef$.1920We present here a selection of the most well-known three-valued21logics, their motivations, and some of their properties.2223\end{document}
content/many-valued-logic/three-valued-logics/lukasiewicz.tex
1% Part: many-valued-logic2% Chapter: three-valued-logics3% Section: lukasiewicz45\documentclass[../../../include/open-logic-section]{subfiles}67\begin{document}89\olfileid{mvl}{thr}{luk}1011\olsection{\L ukasiewicz logic}1213One of the first published, worked out proposals for a many-valued14logic is due to the Polish philosopher Jan \L ukasiewicz in 1921. \L15ukasiewicz was motivated by Aristotle's sea battle problem: It seems16that, \emph{today}, the sentence ``There will be a sea battle17tomorrow'' is neither true nor false: its truth value is not yet18settled. \L ukasiewicz proposed to introduce a third truth value,19to such ``future contingent'' sentences.20\begin{quote}21I can assume without contradiction that my presence in Warsaw at a22certain moment of next year, e.g., at noon on 21 December, is at the23present time determined neither positively nor negatively. Hence it is24possible, but not necessary, that I shall be present in Warsaw at the25given time. On this assumption the proposition ``I shall be in Warsaw26at noon on 21 December of next year,'' can at the present time be27neither true nor false. For if it were true now, my future presence in28Warsaw would have to be necessary, which is contradictory to the29assumption. If it were false now, on the other hand, my future30presence in Warsaw would have to be impossible, which is also31contradictory to the assumption. Therefore the proposition considered32is at the moment neither true nor false and must possess a third33value, different from ``0'' or falsity and ``1'' or truth. This value34we can designate by ``$\frac{1}{2}$.'' It represents ``the possible,''35and joins ``the true'' and ``the false'' as a third value.36\end{quote}37We will use $\Undef$ for \L ukasiewicz's third truth38value.\footnote{\L ukasiewicz here uses ``possible'' in a way that is39uncommon today, namely to mean possible but not necessary.}4041The truth functions for the connectives $\lnot$, $\land$, and $\lor$42are easy to determine on this interpretation: the negation of a future43contingent sentence is also a future contingent sentence, so44$\tf{\lnot}(\Undef) = \Undef$. If one conjunct of a45conjunction is undetermined and the other is true, the conjunction is46also undetermined---after all, depending on how the future contingent47conjunct turns out, the conjunction might turn out to be true, and it48might turn out to be false. So \[49 \tf{\land}(\True, \Undef) =50\tf{\land}(\Undef, \True) =51\Undef.52\]53If the other conjunct is false, however, it cannot54turn out true, so \[\tf{\land}(\False, \Undef) =55\tf{\land}(\False, \Undef) = \False.\]56The other values (if the arguments are settled truth values, $\True$57or $\False$, are like in classical logic.5859For the conditional, the situation is a little trickier. Suppose $q$60is a future contingent statement. If $p$ is false, then $p \lif q$61will be true, regardless of how $q$ turns out, so we should set62$\tf{\lif}(\False, \Undef) = \True$. And if $p$ is true, then63$q \lif p$ will be true, regardless of what $q$ turns out to be, so64$\tf{\lif}(\Undef, \True) = \True$. If $p$ is true, then $p65\lif q$ might turn out to be true or false, so $\tf{\lif}(\True,66\Undef) = \Undef$. Similarly, if $p$ is false, then $q67\lif p$ might turn out to be true or false, so68$\tf{\lif}(\Undef, \False) = \Undef$. This leaves the69case where $p$ and $q$ are both future contingents. On the basis of70the motivation, we should really assign $\Undef$ in this case.71However, this would make $!A \lif !A$ \emph{not} a tautology. \L72ukasiewicz had not trouble giving up $!A \lor \lnot !A$ and $\lnot(!A73\land \lnot !A)$, but balked at giving up $!A \lif !A$. So he74stipulated $\tf{\lif}(\Undef, \Undef) =75\True$.7677\begin{defn}\ollabel{def:lukasiewicz}78Three-valued \L ukasiewicz logic is defined using the matrix:79\begin{enumerate}80 \item The standard propositional language $\Lang L_0$ with81 $\lnot$, $\land$, $\lor$, $\lif$.82 \item The set of truth values $V = \{\True, \Undef, \False\}$.83 \item $\True$ is the only designated value, i.e., $V^+ = \{\True\}$.84 \item Truth functions are given by the following tables:85 \begin{center}86 \begin{tabular}{c|c} 87 $\tf{\lnot}$ & \\ 88 \hline 89 $\True$ & $\False$ \\ 90 $\Undef$ & $\Undef$ \\91 $\False$ & $\True$ 92 \end{tabular}93 \quad94 \begin{tabular}{c|ccc} 95 $\tf{\land}[\LogLuk[3]]$ & $\True$ & $\Undef$ & $\False$ \\ 96 \hline 97 $\True$ & $\True$ & $\Undef$ & $\False$ \\ 98 $\Undef$ & $\Undef$ & $\Undef$ & $\False$\\ 99 $\False$ & $\False$ & $\False$ & $\False$ 100 \end{tabular}101 \\[2ex]102 \begin{tabular}{c|ccc} 103 $\tf{\lor}[\LogLuk[3]]$ & $\True$ & $\Undef$ & $\False$ \\ 104 \hline 105 $\True$ & $\True$ & $\True$ & $\True$ \\ 106 $\Undef$ & $\True$ & $\Undef$ & $\Undef$ \\107 $\False$ & $\True$ & $\Undef$ & $\False$ 108 \end{tabular}109 \quad110 \begin{tabular}{c|ccc} 111 $\tf{\lif}[\LogLuk[3]]$ & $\True$ & $\Undef$ & $\False$ \\ 112 \hline 113 $\True$ & $\True$ & $\Undef$ & $\False$ \\ 114 $\Undef$ & $\True$ & $\True$ & $\Undef$ \\ 115 $\False$ & $\True$ & $\True$ & $\True$ 116 \end{tabular}117 \end{center} 118\end{enumerate}119\end{defn}120121As can easily be seen, any !!{formula}~$!A$ containing only $\lnot$,122$\land$, and $\lor$ will take the truth value~$\Undef$ if all123its !!{propositional variable}s are assigned~$\Undef$. So for124instance, the classical tautologies $p \lor \lnot p$ and $\lnot(p125\land \lnot p)$ are not tautologies in $\LogLuk[3]$, since $\pValue{v}(!A)126= \Undef$ whenever $\pAssign v(p) = \Undef$.127128On !!{valuation}s where $\pAssign v(p) = \True$ or $\False$, $\pValue129v(!A)$ will coincide with its classical truth value.130131\begin{prop}132 If $\pAssign v(p) \in \{\True, \False\}$ for all $p$ in~$!A$, then133 $\pValue v(!A)[\LogLuk[3]] = \pValue v(!A)[\LogCL]$.134\end{prop}135136\begin{prob}\label{mvl:thr:luk:prob:luk-iff} Suppose we define137 $\pValue v(!A \liff !B) = \pValue v((!A \lif !B) \land (!B \lif138 !A))$ in~$\LogLuk[3]$. What truth table would $\liff$ have?139\end{prob}140141Many classical tautologies \emph{are} also tautologies in \LogLuk[3],142e.g, $\lnot p \lif (p \lif q)$. Just like in classical logic, we can143use truth tables to verify this:144\begin{center}145\begin{tabular}{cc|cccccc}146 $p$ & $q$ & $\lnot$ & $p$ & $\lif$ & $\smash{(}p$ & $\lif$ & $q\smash{)}$ \\ \hline147 \True & \True & \False & \True & \True & \True & \True & \True \\148 \True & \Undef & \False & \True & \True & \True & \Undef & \Undef \\149 \True & \False & \False & \True & \True & \True & \False & \False \\150 \Undef & \True & \Undef & \Undef & \True & \Undef & \True & \True \\151 \Undef & \Undef & \Undef & \Undef & \True & \Undef & \True & \Undef \\152 \Undef & \False & \Undef & \Undef & \True & \Undef & \Undef & \False \\153 \False & \True & \True & \False & \True & \False & \True & \True \\154 \False & \Undef & \True & \False & \True & \False & \True & \Undef \\155 \False & \False & \True & \False & \True & \False & \True & \False \\ 156\end{tabular}157\end{center}158159\begin{prob}160 Show that the following are tautologies in \LogLuk[3]:161 \begin{enumerate}162 \item $p \lif (q \lif p)$163 \item\label{mvl:thr:luk:prob:luk-taut-2} $\lnot(p \land q) \liff (\lnot p \lor \lnot q)$164 \item\label{mvl:thr:luk:prob:luk-taut-3} $\lnot(p \lor q) \liff (\lnot p \land \lnot q)$165 \end{enumerate}166 (In \olref[mvl][thr][luk]{prob:luk-taut-2} and167 \olref[mvl][thr][luk]{prob:luk-taut-3}, take $!A \liff !B$ as an168 abbreviation for $(!A \lif !B) \land (!B \lif !A)$, or refer to your169 solution to \olref[mvl][thr][luk]{prob:luk-iff}.)170\end{prob}171172\begin{prob}173 Show that the following classical tautologies are not tautologies in~\LogLuk[3]:174 \begin{enumerate}175 \item $(\lnot p \land p) \lif q)$176 \item $((p \lif q) \lif p) \lif p$177 \item $(p \lif (p \lif q)) \lif (p \lif q)$178 \end{enumerate}179\end{prob}180181One might therefore perhaps think that although not all classical182tautologies are tautologies in~$\LogLuk[3]$, they should at least take183either the value~$\True$ or the value~$\Undef$ on every184!!{valuation}. This is not the case. A counterexample is given by185\[186 \lnot(p \lif \lnot p) \lor \lnot(\lnot p \lif p)187\]188which is $\False$ if $p$ is~$\Undef$.189190\begin{prob}191 Which of the following relations hold in \L ukasiewicz logic? Give a truth table for each.192 \begin{enumerate}193 \item $p, p \lif q \Entails q$194 \item $\lnot\lnot p \Entails p$195 \item $p \land q \Entails p$196 \item $p \Entails p \land p$197 \item $p \Entails p \lor q$198 \end{enumerate}199\end{prob}200201\L ukasiewicz hoped to build a logic of possibility on the basis of his202three-valued system, by introducing a one-place connective $\Diamond203!A$ (for ``$!A$ is possible'') and a corresponding $\Box !A$ (for ``$!A$204is necessary''):205\begin{center}206 \begin{tabular}{c|c} 207 $\tf{\Diamond}$ & \\ 208 \hline 209 $\True$ & $\True$ \\ 210 $\Undef$ & $\True$ \\211 $\False$ & $\False$ 212 \end{tabular}213 \quad214 \begin{tabular}{c|c} 215 $\tf{\Box}$ & \\ 216 \hline 217 $\True$ & $\True$ \\ 218 $\Undef$ & $\False$ \\219 $\False$ & $\False$ 220 \end{tabular}221\end{center}222In other words, $p$ is possible iff it is not already settled as false;223and $p$ is necessary iff it is already settled as true.224225\begin{prob}226 Show that $\Box p \liff \lnot\Diamond \lnot p$ and $\Diamond p \liff227 \lnot \Box \lnot p$ are tautologies in~\LogLuk[3], extended with the228 truth tables for $\Box$ and~$\Diamond$.229\end{prob}230231However, the shortcomings of this proposed modal logic soon became232evident: However things turn out, $p \land \lnot p$ can never turn out233to be true. So even if it is not now settled (and therefore234undetermined), it should count as impossible, i.e., $\lnot \Diamond(p235\land \lnot p)$ should be a tautology. However, if $\pAssign v(p) =236\Undef$, then $\pValue v(\lnot \Diamond(p \land \lnot p)) =237\Undef$. Although \L ukasiewicz was correct that two truth238values will not be enough to accommodate modal distinctions such as239possiblity and necessity, introducing a third truth value is also not240enough.241242\end{document}
content/many-valued-logic/three-valued-logics/kleene.tex
1% Part: many-valued-logic2% Chapter: three-valued-logics3% Section: kleene45\documentclass[../../../include/open-logic-section]{subfiles}67\begin{document}89\olfileid{mvl}{thr}{skl}1011\olsection{Kleene logics}1213Stephen Kleene introduced two three-valued logics motivated by a logic14in which truth values are thought of the outcomes of computational15procedures: a procedure may yield $\True$ or $\False$, but it may also16fail to terminate. In that case the corresponding truth value is17undefined, represented by the truth value~$\Undef$.1819To compute the negation of a proposition~$!A$, you would first compute20the value of~$!A$, and then return the opposite of the result. If the21computation of $!A$ does not terminate, then the entire procedure does22not either: so the negation of $\Undef$ is $\Undef$.2324To compute a conjunction $!A \land !B$, there are two options: one can25first compute~$!A$, then $!B$, and then the result would be $\True$ if26the outcome of both is~$\True$, and $\False$ otherwise. If either27computation fails to halt, the entire procedure does as well. So in28this case, the if one conjunct is undefined, the conjunction is as29well. The same goes for disjunction.3031However, if we can evaluate $!A$ and $!B$ in parallel, we can do better.32Then, if one of the two procedures halts and returns $\False$, we can33stop, as the answer must be false. So in that case a conjunction with34one false conjunct is false, even if the other conjunct is undefined.35Similarly, when computing a disjunction in parallel, we can stop once36the procedure for one of the two disjuncts has returned true: then the37disjunction must be true. So in this case we can know what the outcome38of a compound claim is, even if one of the components is undefined. On39this interpretation, we might read $\Undef$ as ``unknown'' rather than40``undefined.''4142The two interpretations give rise to Kleene's strong and weak logic.43The conditional is defined as equivalent to $\lnot !A \lor !B$.4445\begin{defn}46\emph{Strong Kleene logic}~$\LogKs$ is defined using the matrix:47\begin{enumerate}48 \item The standard propositional language $\Lang L_0$ with49 $\lnot$, $\land$, $\lor$, $\lif$.50 \item The set of truth values $V = \{\True, \Undef, \False\}$.51 \item $\True$ is the only designated value, i.e., $V^+ = \{\True\}$.52 \item Truth functions are given by the following tables:53 \begin{center}54 \begin{tabular}{c|c} 55 $\tf{\lnot}$ & \\ 56 \hline 57 $\True$ & $\False$ \\ 58 $\Undef$ & $\Undef$ \\59 $\False$ & $\True$ 60 \end{tabular}61 \quad62 \begin{tabular}{c|ccc} 63 $\tf{\land}[\LogKs]$ & $\True$ & $\Undef$ & $\False$ \\ 64 \hline 65 $\True$ & $\True$ & $\Undef$ & $\False$ \\ 66 $\Undef$ & $\Undef$ & $\Undef$ & $\False$\\ 67 $\False$ & $\False$ & $\False$ & $\False$ 68 \end{tabular}69 \\[2ex]70 \begin{tabular}{c|ccc} 71 $\tf{\lor}[\LogKs]$ & $\True$ & $\Undef$ & $\False$ \\ 72 \hline 73 $\True$ & $\True$ & $\True$ & $\True$ \\ 74 $\Undef$ & $\True$ & $\Undef$ & $\Undef$ \\75 $\False$ & $\True$ & $\Undef$ & $\False$ 76 \end{tabular}77 \quad78 \begin{tabular}{c|ccc} 79 $\tf{\lif}[\LogKs]$ & $\True$ & $\Undef$ & $\False$ \\ 80 \hline 81 $\True$ & $\True$ & $\Undef$ & $\False$ \\ 82 $\Undef$ & $\True$ & $\Undef$ & $\Undef$ \\ 83 $\False$ & $\True$ & $\True$ & $\True$ 84 \end{tabular}85 \end{center} 86\end{enumerate}87\end{defn}8889\begin{defn}90\emph{Weak Kleene logic}~$\LogKw$ is defined using the matrix:91\begin{enumerate}92 \item The standard propositional language $\Lang L_0$ with93 $\lnot$, $\land$, $\lor$, $\lif$.94 \item The set of truth values $V = \{\True, \Undef, \False\}$.95 \item $\True$ is the only designated value, i.e., $V^+ = \{\True\}$.96 \item Truth functions are given by the following tables:97 \begin{center}98 \begin{tabular}{c|c} 99 $\tf{\lnot}$ & \\ 100 \hline 101 $\True$ & $\False$ \\ 102 $\Undef$ & $\Undef$ \\103 $\False$ & $\True$ 104 \end{tabular}105 \quad106 \begin{tabular}{c|ccc} 107 $\tf{\land}[\LogKw]$ & $\True$ & $\Undef$ & $\False$ \\ 108 \hline 109 $\True$ & $\True$ & $\Undef$ & $\False$ \\ 110 $\Undef$ & $\Undef$ & $\Undef$ & $\Undef$\\ 111 $\False$ & $\False$ & $\Undef$ & $\False$ 112 \end{tabular}113 \\[2ex]114 \begin{tabular}{c|ccc} 115 $\tf{\lor}[\LogKw]$ & $\True$ & $\Undef$ & $\False$ \\ 116 \hline 117 $\True$ & $\True$ & $\Undef$ & $\True$ \\ 118 $\Undef$ & $\Undef$ & $\Undef$ & $\Undef$ \\119 $\False$ & $\True$ & $\Undef$ & $\False$ 120 \end{tabular}121 \quad122 \begin{tabular}{c|ccc} 123 $\tf{\lif}[\LogKw]$ & $\True$ & $\Undef$ & $\False$ \\ 124 \hline 125 $\True$ & $\True$ & $\Undef$ & $\False$ \\ 126 $\Undef$ & $\Undef$ & $\Undef$ & $\Undef$ \\ 127 $\False$ & $\True$ & $\Undef$ & $\True$ 128 \end{tabular}129 \end{center} 130\end{enumerate}131\end{defn}132133\begin{prop}134 $\LogKs$ and $\LogKw$ have no tautologies.135\end{prop}136137\begin{proof}138 If $\pAssign v(p) = \Undef$ for all !!{propositional variable}s~$p$,139 then any formula~$!A$ will have truth value~$\pValue v(!A) =140 \Undef$, since141 \[142 \tf{\lnot}(\Undef) = \tf{\lor}(\Undef, \Undef) = \tf{\land}(\Undef,143 \Undef) = \tf{\lif}(\Undef, \Undef) = \Undef144 \]145 in both logics. As $\Undef \notin V^+$ for either $\LogKs$ or146 $\LogKw$, on this !!{valuation}, $!A$ will not be designated.147\end{proof}148149Although both weak and strong Kleene logic have no tautologies, they150have non-trivial consequence relations. 151152\begin{prob}153 Which of the following relations hold in (a) strong and154 (b) weak Kleene logic? Give a truth table for each.155 \begin{enumerate}156 \item $p, p \lif q \Entails q$157 \item $p \lor q, \lnot p \Entails q$158 \item $p \land q \Entails p$159 \item $p \Entails p \land p$160 \item $p \Entails p \lor q$161 \end{enumerate}162\end{prob}163164Dmitry Bochvar interpreted $\Undef$ as ``meaningless'' and attempted165to use it to solve paradoxes such as the Liar paradox by stipulating166that paradoxical sentences take the value~$\Undef$. He introduced a167logic which is essentially weak Kleene logic extended by additional168connectives, two of which are ``external negation'' and the ``is169undefined'' operator:170\begin{center}171 \begin{tabular}{c|c} 172 $\tf{\sim}$ & \\ 173 \hline 174 $\True$ & $\False$ \\ 175 $\Undef$ & $\True$ \\176 $\False$ & $\True$ 177 \end{tabular}178\quad179 \begin{tabular}{c|c} 180 $\tf{+}$ & \\ 181 \hline 182 $\True$ & $\False$ \\ 183 $\Undef$ & $\True$ \\184 $\False$ & $\False$ 185 \end{tabular}186\end{center}187188\begin{prob}189 Can you define $\sim$ in Bochvar's logic in terms of $\lnot$190 and~$+$, i.e., find !!a{formula} with only the !!{propositional191 variable}~$p$ and not involving $\sim$ which always takes the same192 truth value as~$\mathord{\sim}p$? Give a truth table to show you're193 right.194\end{prob}195196197\end{document}
content/many-valued-logic/three-valued-logics/goedel.tex
1% Part: many-valued-logic2% Chapter: three-valued-logics3% Section: goedel45\documentclass[../../../include/open-logic-section]{subfiles}67\begin{document}89\olfileid{mvl}{thr}{god}1011\olsection{G\"odel logics}1213Kurt G\"odel introduced a sequence of $n$-valued logics that each14contain all !!{formula}s valid in intuitionistic logic, and are15contained in classical logic. Here is the first interesting one:1617\begin{defn}\ollabel{defn:goedel}18\emph{$3$-valued G\"odel logic}~$\LogGod$ is defined using the matrix:19\begin{enumerate}20 \item The standard propositional language $\Lang L_0$ with21 $\lfalse$, $\lnot$, $\land$, $\lor$, $\lif$.22 \item The set of truth values $V = \{\True, \Undef, \False\}$.23 \item $\True$ is the only designated value, i.e., $V^+ = \{\True\}$.24 \item For $\lfalse$, we have $\tf{\lfalse} = \False$. Truth25 functions for the remaining connectives are given by the following26 tables:27 \begin{center}28 \begin{tabular}{c|c} 29 $\tf{\lnot}[\LogGod]$ & \\ 30 \hline 31 $\True$ & $\False$ \\ 32 $\Undef$ & $\False$ \\33 $\False$ & $\True$ 34 \end{tabular}35 \quad36 \begin{tabular}{c|ccc} 37 $\tf{\land}[\LogGod]$ & $\True$ & $\Undef$ & $\False$ \\ 38 \hline 39 $\True$ & $\True$ & $\Undef$ & $\False$ \\ 40 $\Undef$ & $\Undef$ & $\Undef$ & $\False$\\ 41 $\False$ & $\False$ & $\False$ & $\False$ 42 \end{tabular}43 \\[2ex]44 \begin{tabular}{c|ccc} 45 $\tf{\lor}[\LogGod]$ & $\True$ & $\Undef$ & $\False$ \\ 46 \hline 47 $\True$ & $\True$ & $\True$ & $\True$ \\ 48 $\Undef$ & $\True$ & $\Undef$ & $\Undef$ \\49 $\False$ & $\True$ & $\Undef$ & $\False$ 50 \end{tabular}51 \quad52 \begin{tabular}{c|ccc} 53 $\tf{\lif}[\LogGod]$ & $\True$ & $\Undef$ & $\False$ \\ 54 \hline 55 $\True$ & $\True$ & $\Undef$ & $\False$ \\ 56 $\Undef$ & $\True$ & $\True$ & $\False$ \\ 57 $\False$ & $\True$ & $\True$ & $\True$ 58 \end{tabular}59 \end{center} 60\end{enumerate}61\end{defn}6263You'll notice that the truth tables for $\land$ and~$\lor$ are the64same as in \L ukasiewicz and strong Kleene logic, but the truth tables65for $\lnot$ and~$\lif$ differ for each. In G\"odel logic,66$\tf{\lnot}(\Undef) = \False$. In contrast to \L ukasiewicz logic and67Kleene logic, $\tf{\lif}(\Undef, \False) = \False$; in contrast to68Kleene logic (but as in \L ukasiewicz logic), $\tf{\lif}(\Undef,69\Undef) = \True$.7071As the connection to intuitionistic logic alluded to above suggests,72$\LogGod[3]$ is close to intuitionistic logic. All intuitionistic73truths are tautologies in~$\LogGod[3]$, and many classical tautologies74that are not valid intuitionistically also fail to be tautologies75in~$\LogGod[3]$. For instance, the following are not tautologies:76\begin{align*}77 & p \lor \lnot p && (p \lif q) \lif (\lnot p \lor q) \\78 & \lnot\lnot p \lif p && \lnot(\lnot p \land \lnot q) \lif (p \lor q) \\79 & ((p \lif q) \lif p) \lif p && \lnot(p \lif q) \lif (p \land \lnot q)80\end{align*}81However, not every tautology of $\LogGod[3]$ is also82intuitionistically valid, e.g., $\lnot\lnot p \lor \lnot p$ or $(p83\lif q) \lor (q \lif p)$.8485\begin{prob}86 Give truth tables to show that the following are tautologies87 of~$\LogGod[3]$:88 \begin{align*}89 & \lnot\lnot p \lor \lnot p\\90 & (p \lif q) \lor (q \lif p) \\91 & \lnot(p \land q) \lif (\lnot p \lor \lnot q) \\92 & (p \lif q) \lor (q \lif r) \lor (r \lif s)93 \end{align*}94\end{prob}9596\begin{prob}97 Give truth tables that show that the following are not tautologies98 of~$\LogGod[3]$99 \begin{align*}100 & (p \lif q) \lif (\lnot p \lor q) \\101 & \lnot(\lnot p \land \lnot q) \lif (p \lor q) \\102 & ((p \lif q) \lif p) \lif p \\103 & \lnot(p \lif q) \lif (p \land \lnot q)104 \end{align*}105\end{prob}106107\begin{prob}108 Which of the following relations hold in G\"odel logic? Give a truth table for each.109 \begin{enumerate}110 \item $p, p \lif q \Entails q$111 \item $p \lor q, \lnot p \Entails q$112 \item $p \land q \Entails p$113 \item $p \Entails p \land p$114 \item $p \Entails p \lor q$115 \end{enumerate}116\end{prob}117118\end{document}
content/many-valued-logic/three-valued-logics/multiple-designation.tex
1% Part: many-valued-logic2% Chapter: three-valued-logics3% Section: multiple-designation45\documentclass[../../../include/open-logic-section]{subfiles}67\begin{document}89\olfileid{mvl}{thr}{mul}1011\olsection{Designating not just $\True$}1213So far the logics we've seen all had the set of designated truth14values $V^+ = \{\True\}$, i.e., something counts as true iff its truth15value is~$\True$. But one might also count something as true if it's16just not~$\False$. Then one would get a logic by stipulating in the17matrix, e.g., that $V^+ = \{\True, \Undef\}$.1819\begin{defn}20The \emph{logic of paradox}~$\LogLP$ is defined using the matrix:21\begin{enumerate}22 \item The standard propositional language $\Lang L_0$ with23 $\lnot$, $\land$, $\lor$, $\lif$.24 \item The set of truth values $V = \{\True, \Undef, \False\}$.25 \item $\True$ and $\Undef$ are designated, i.e., $V^+ = \{\True, \Undef\}$.26 \item Truth functions are the same as in strong Kleene logic.27\end{enumerate}28\end{defn}2930\begin{defn}31Halld\'en's \emph{logic of nonsense}~$\LogHal$ is defined using the matrix:32\begin{enumerate}33 \item The standard propositional language $\Lang L_0$ with34 $\lnot$, $\land$, $\lor$, $\lif$ and a $1$-place connective~$+$.35 \item The set of truth values $V = \{\True, \Undef, \False\}$.36 \item $\True$ and $\Undef$ are designated, i.e., $V^+ = \{\True, \Undef\}$.37 \item Truth functions are the same as weak Kleene logic, plus the38 ``is meaningless'' operator:39 \begin{center}40 \begin{tabular}{c|c} 41 $\tf{+}$ & \\ 42 \hline 43 $\True$ & $\False$ \\ 44 $\Undef$ & $\True$ \\45 $\False$ & $\False$ 46 \end{tabular}47 \end{center}48\end{enumerate}49\end{defn}5051By contrast to the Kleene logics with which they share truth tables,52these \emph{do} have tautologies. 5354\begin{prop}\ollabel{prop:LP-taut-CL} The tautologies of $\LogLP$ are55 the same as the tautologies of classical propositional logic.56\end{prop}5758\begin{proof}59 By \olref[syn][sub]{prop:mvl-cl}, if $\Entails[\LogLP] !A$ then60 $\Entails[\LogCL] !A$. To show the reverse, we show that if there is61 !!a{valuation}~$\pAssign v\colon \PVar \to \{\False, \True,62 \Undef\}$ such that $\pValue v(!A)[\LogKs] = \False$ then there is63 !!a{valuation}~$\pAssign {v'}\colon \PVar \to \{\False, \True\}$64 such that $\pValue {v'}(!A)[\LogCL] = \False$. This establishes the65 result for $\LogLP$, since $\LogKs$ and $\LogLP$ have the same66 characteristic truth functions, and $\False$ is the only truth value67 of $\LogLP$ that is not designated (that is the only difference68 between $\LogLP$ and~$\LogKs$). Thus, if $\Entails/[\LogLP] !A$, for69 some !!{valuation}~$\pAssign v$, $\pValue v(!A)[\LogLP] = \pValue70 v[\LogKs](!A) = \False$. By the claim we're proving,71 $\pValue{v'}[\LogCL](!A) = \False$, i.e., $\Entails/[\LogCL] !A$.72 73 To establish the claim, we first define $\pAssign {v'}$ as74 \[75 \pAssign {v'}(p) = 76 \begin{cases}77 \True & \text{if } \pAssign {v}(p) \in \{\True, \Undef\}\\78 \False & \text{otherwise}79 \end{cases}80 \]81 We now show by induction on $!A$ that (a)~if $\pValue v(!A)[\LogKs]82 = \False$ then $\pValue {v'}(!A)[\LogCL] = \False$, and (b)~if83 $\pValue v(!A)[\LogKs] = \True$ then $\pValue {v'}(!A)[\LogCL] =84 \True$85 \begin{enumerate}86 \item Induction basis: $!A \ident p$. By87 \olref[syn][val]{defn:pValue}, $\pValue v(!A)[\LogKs]88 = \pAssign v(p) = \pValue {v'}(!A)[\LogCL]$, which implies both (a)89 and~(b).90 91 For the induction step, consider the cases:92 \item $!A \ident \lnot !B$. 93 \begin{enumerate}94 \item Suppose $\pValue v(\lnot!B)[\LogKs] = \False$. By the95 definition of $\tf{\lnot}[\LogKs]$, $\pValue v(!B)[\LogKs] =96 \True$. By inductive hypothesis, case~(b), we get $\pValue97 {v'}(!B)[\LogCL] = \True$, so $\pValue {v'}(\lnot !B)[\LogCL] = \False$.98 \item Suppose $\pValue v(\lnot!B)[\LogKs] = \True$. By the99 definition of $\tf{\lnot}[\LogKs]$, $\pValue v(!B)[\LogKs] =100 \False$. By inductive hypothesis, case~(a), we get $\pValue101 {v'}(!B)[\LogCL] = \False$, so $\pValue {v'}(\lnot !B)[\LogCL] = \True$.102 \end{enumerate}103104 \item $!A \ident (!B \land !C)$.105 \begin{enumerate}106 \item Suppose $\pValue v(!B \land !C)[\LogKs] = \False$. By the107 definition of $\tf{\land}[\LogKs]$, $\pValue v(!B)[\LogKs] =108 \False$ or $\pValue v(!B)[\LogKs] = \False$. By inductive109 hypothesis, case~(a), we get $\pValue {v'}(!B)[\LogCL] = \False$110 or $\pValue {v'}(!C)[\LogCL] = \False$, so $\pValue {v'}(!B \land111 !C)[\LogCL] = \False$.112 \item Suppose $\pValue v(!B \land !C)[\LogKs] = \True$. By the113 definition of $\tf{\land}[\LogKs]$, $\pValue v(!B)[\LogKs] =114 \True$ and $\pValue v(!B)[\LogKs] = \True$. By inductive115 hypothesis, case~(b), we get $\pValue {v'}(!B)[\LogCL] = \True$116 and $\pValue {v'}(!C)[\LogCL] = \True$, so $\pValue {v'}(!B \land117 !C)[\LogCL] = \True$.118 \end{enumerate}119 \end{enumerate}120121 The other two cases are similar, and left as exercises. Alternatively,122 the proof above establishes the result for all !!{formula}s only123 containing $\lnot$ and~$\land$. One may now appeal to the facts that in124 both $\LogKs$ and $\LogCL$, for any $\pAssign v$, $\pValue v(!B \lor125 !C) = \pValue v(\lnot(\lnot!B \land \lnot !C))$ and $\pValue v(!B \lif126 !C) = \pValue v(\lnot(!B \land \lnot !C))$.127\end{proof}128129\begin{prob}130 Complete the proof \olref[mvl][thr][mul]{prop:LP-taut-CL}, i.e.,131 establish (a) and~(b) for the cases where $!A \ident (!B \lor !C)$132 and $!A \ident (!B \lif !C)$.133\end{prob}134135\begin{prob}136Prove that every classical tautology is a tautology in~$\LogHal$.137\end{prob}138139Although they have the same tautologies as classical logic, their140consequence relations are different. $\LogLP$, for instance, is141\emph{paraconsistent} in that $\lnot p, p \Entails/ q$, and so the142principle of explosion $\lnot !A, !A \Entails !B$ does not hold in143general. (It holds for some cases of $!A$ and $!B$, e.g., if $!B$~is a144tautology.)145146\begin{prob}147 Which of the following relations hold in (a)~$\LogLP$148 and in (b)~$\LogHal$? Give a truth table for each.149 \begin{enumerate}150 \item $p, p \lif q \Entails q$151 \item $\lnot q, p \lif q \Entails \lnot p$152 \item $p \lor q, \lnot p \Entails q$153 \item $\lnot p, p \Entails q$154 \item $p \Entails p \lor q$155 \item $p \lif q, q\lif r \Entails p \lif r$156 \end{enumerate}157\end{prob}158159What if you make $\Undef$ designated in $\LogLuk[3]$?160161\begin{defn}162 The logic \emph{3-valued R-Mingle}~$\LogRM[3]$ is defined using the matrix:163 \begin{enumerate}164 \item The standard propositional language $\Lang L_0$ with165 $\lfalse$, $\lnot$, $\land$, $\lor$, $\lif$.166 \item The set of truth values $V = \{\True, \Undef, \False\}$.167 \item $\True$ and $\Undef$ are designated, i.e., $V^+ = \{\True, \Undef\}$.168 \item Truth functions are the same as \L ukasiewicz logic~$\LogLuk[3]$.169 \end{enumerate}170\end{defn}171172\begin{prob}173 Which of the following relations hold in $\LogRM[3]$?174 \begin{enumerate}175 \item $p, p \lif q \Entails q$176 \item $p \lor q, \lnot p \Entails q$177 \item $\lnot p, p \Entails q$178 \item $p \Entails p \lor q$179 \end{enumerate}180\end{prob}181182Different truth tables can sometimes generate the same logic183(entailment relation) just by changing the designated values.184E.g., this happens if in G\"odel logic we take $V^+ = \{\True,185\Undef\}$ instead of~$\{\True\}$.186187\begin{prop}\ollabel{prop:gl-udes}188 The matrix with $V = \{\False, \Undef,\True\}$, $V^+=\{\True,189 \Undef\}$, and the truth functions of $3$-valued G\"odel logic190 defines classical logic.191\end{prop}192193\begin{proof}194 Exercise. 195\end{proof}196197\begin{prob}198 Prove \olref[mvl][thr][mul]{prop:gl-udes} by showing that for the199 logic~$\Log L$ defined just like G\"odel logic but with200 $V^+=\{\True,\Undef\}$, if $\Gamma \Entails/[\Log L] !B$ then201 $\Gamma \Entails/[\LogCL] !B$. Use the ideas of202 \olref[mvl][thr][mul]{prop:LP-taut-CL}, except instead of proving203 properties (a) and~(b), show that $\pValue v(!A)[\LogGod] = \False$204 iff $\pValue {v'}(!A)[\LogCL] = \False$ (and hence that $\pValue205 v(!A)[\LogGod] \in \{\True,\Undef\}$ iff $\pValue {v'}(!A)[\LogCL] =206 \True$). Explain why this establishes the proposition.207\end{prob}208209\end{document}