For the predicate/subset bijection, and the nullary and unary logical operations on 2, see “2 as the Boolean subobject classifier”.
For the binary logical operations on 2 (“the Boolean logical connectives”), see “Elementary Boolean geometry and logic”.
Inputs:
Inputs are two predicate/subset pairs, A and B, on a given set X∈Set, as shown below, embedded as the right squares in 2×3 diagrams.
Also shown are two noteworthy special cases, which will be used in the sequel.
\begin{array}{}
\objA\times\objX & \xrightarrow{} & \objA & \xrightarrow{} & \bfone \\
\llap{i_\objA \times\objX} \downarrow & \mkern-1.8em{\raise2ex\hbox{$\lrcorner$}} & \llap{i_\objA} \downarrow & \mkern-1.8em{\raise2ex\hbox{$\lrcorner$}} & \downarrow\rlap\radjtop \\
\objX\times\objX & \xrightarrow[\textstyle \pi_1]{} & \objX & \xrightarrow[\textstyle \objA]{} & \bftwo \\
\end{array}
\begin{array}{}
\objX\times\objB & \xrightarrow{} & \objB & \xrightarrow{} & \bfone \\
\llap{\objX \times i_\objB} \downarrow & \mkern-1.8em{\raise2ex\hbox{$\lrcorner$}} & \llap{i_\objB} \downarrow & \mkern-1.8em{\raise2ex\hbox{$\lrcorner$}} & \downarrow\rlap\radjtop \\
\objX\times\objX & \xrightarrow[\textstyle \pi_2]{} & \objX & \xrightarrow[\textstyle \objB]{} & \bftwo \\
\end{array}
\begin{array}{}
\radjtop\times\bftwo & \xrightarrow{} & \bfone \\
\llap{\radjtop\times\bftwo} \downarrow & \mkern-1.8em{\raise2ex\hbox{$\lrcorner$}} & \downarrow\rlap\radjtop \\
\bftwo\times\bftwo & \xrightarrow[\textstyle \pi_1]{} & \bftwo \\
\end{array}
\begin{array}{}
\bftwo\times\radjtop & \xrightarrow{} & \bfone \\
\llap{\bftwo\times\radjtop} \downarrow & \mkern-1.8em{\raise2ex\hbox{$\lrcorner$}} & \downarrow\rlap\radjtop \\
\bftwo\times\bftwo & \xrightarrow[\textstyle \pi_2]{} & \bftwo \\
\end{array}
Internal:
(Intersection of subsets) and (conjunction of predicates)(the internal operations in \Set) \begin{array}{} &&& \eltx \in (\setA\cap\setB) & \xleftrightarrow{} & {\eltx\in\setA} \wedge {\eltx\in\setB} \\ \\ \begin{array}{l} \text{Intersection}\\ \text{of subsets}\\ \text{(internal)}\\ \end{array} &&& \boxed{\begin{array}{} \\ \setA \cap \setB\\ \\ \{ \eltx\in\setX \mid {{\eltx\in\setA} \wedge {\eltx\in\setB} \} }\\ \end{array} } & \longrightarrow & \boxed{ \begin{array}{} \pi_1^{-1}\objA \cap \pi_2^{-1}\objB \\ \setA\times\setX \cap \setX\times\setB\\ \setA \times \setB\\ \{ \langle\eltx,\elty\rangle \in \setX\times\setX \mid {{\eltx\in\setA} \wedge {\elty\in\setB} \} }\\ \end{array} } & \xrightarrow[]{} & \boxed { \begin{array}{} \pi_1^{-1}\radjtop \cap \pi_2^{-1}\radjtop \\ {\radjtop\times\bftwo} \cap {\bftwo\times\radjtop} \\ \radjtop\times\radjtop \\ \{ \langle \radjtop, \radjtop \rangle \} \\ \end{array} } & \xrightarrow[]{} & \bfone \\ &&& \Big\downarrow & \mkern-1.8em{\raise2ex\hbox{$\lrcorner$}} & \Big\downarrow & \mkern-1.8em{\raise2ex\hbox{$\lrcorner$}} & \Big\downarrow & \mkern-1.8em{\raise2ex\hbox{$\lrcorner$}} & \Big\downarrow\rlap\radjtop \\ & \eltx & \in & \setX & \xrightarrow[\textstyle \Delta]{} & \setX \times \setX & \xrightarrow[\textstyle \setA\times\setB]{} & \bftwo \times \bftwo & \xrightarrow[\textstyle \boxed{ \begin{array}{}\wedge\\ \pi_1\wedge\pi_2\\ \end{array} } ]{} & \bftwo & \ni & {{\setA_\eltx} \wedge {\setB_\eltx}} \\ &&& \Vert &&&& \Vert \\ &&& \setX && \xrightarrow[\textstyle \langle\functionA,\functionB\rangle]{} && \bftwo \times \bftwo \\ \begin{array}{l} \text{Conjunction}\\ \text{of predicates}\\ \text{(internal)}\\ \end{array} & \Bigg\Vert && \Bigg\Vert &&&&&& \Bigg\Vert && \Bigg\Vert \\ & \eltx & \in & \setX &&& \xrightarrow[\textstyle \; \boxed{\setA\wedge\setB} \; ]{} &&& \bftwo & \ni & {(\setA\wedge\setB)_\eltx} \\ \end{array}
External plus internal:
(Intersection of subsets) and (conjunction of predicates)(the external operations on \Set, plus projections/evaluations to the internal Boolean algebra \bftwo) \begin{array}{lcccccccccccc} \text{Intersection of subsets (external)} && \langle\setA,\setB\rangle & \in & \Sub(\setX) \times \Sub(\setX) && \xrightarrow{\textstyle \boxed\cap} && \Sub(\setX) & \ni & \setA\cap\setB \\ && \wr && \Vert\wr &&&& \Vert\wr && \wr \\ && \langle\functionA,\functionB\rangle & \in & \hom \setX \Set \bftwo \times \hom \setX \Set \bftwo & \xrightarrow[\textstyle \times]{} & \hom {\setX\times\setX} \Set {\bftwo\times\bftwo} & \xrightarrow[\textstyle \hom \Delta \Set \wedge]{} & \hom \setX \Set \bftwo & \ni & \functionA \wedge \functionB \\ \text{Conjunction of predicates (external)} && \Vert && \Vert &&&& \Vert && \Vert \\ && \langle\functionA,\functionB\rangle & \in & \hom \setX \Set \bftwo \times \hom \setX \Set \bftwo && \xrightarrow[\textstyle \boxed\wedge]{} && \hom \setX \Set \bftwo & \ni & \functionA \wedge \functionB \\ && \mapsdownto && \Big\downarrow \rlap{\mkern-2.1em E_\eltx\times E_\eltx} &&&& \Big\downarrow \rlap{E_\eltx} && \mapsdownto \\ \text{Conjunction in $\bftwo$ (internal)} && \langle \functionA_\eltx,\functionB_\eltx \rangle & \in & \bftwo\times\bftwo && \xrightarrow[\textstyle \boxed\wedge]{} && \bftwo & \ni & \functionA_\eltx \wedge \functionB_\eltx \\ \end{array}
The above shows how, given a function in \Set from \bftwo\times\bftwo to \bftwo (an internal binary operation on \bftwo),
that function generates binary operations on \hom \objX \Set \bftwo and \Sub(X), for arbitrary sets \objX\in\Set (which in fact are natural with respect to \setX\in\Set\op),
a process we call extension.
The process may be reversed.
First, binary operations on \hom \objX \Set \bftwo and \Sub(X) are in bijection, using the predicate/subset bijection.
Second, any binary operation on \hom \objX \Set \bftwo yields a binary operation on \bftwo,
by applying the operation to the specific predicates \pi_1, \pi_2 : \bftwo\times\bftwo \to \bftwo, a process we call restriction.
If we restrict to binary operations on \hom \objX \Set \bftwo and \Sub(X) which satisfy the categorical condition of naturality,
then these two processes are inverse to each other.
For example, evaluating (the construction of conjunction for predicates) on (the specific predicates \pi_1, \pi_2 : \bftwo\times\bftwo \to \bftwo), we have: \begin{array}{} \objA,\objB & \xmapsto{} & \objA,\objB,\objA,\objB & \xmapsto{} & \objA,\objB & \xmapsto{} & \objA\wedge\objB \\ \bftwo\times\bftwo && \xrightarrow{\textstyle \text{identity}} && \bftwo\times\bftwo \\ \Vert &&&& \Vert \\ \bftwo\times\bftwo & \xrightarrow[\textstyle \Delta]{} & \bftwo\times\bftwo\times\bftwo\times\bftwo & \xrightarrow[\textstyle \pi_1 \times \pi_2]{} & \bftwo\times\bftwo & \xrightarrow[\textstyle \wedge]{} & \bftwo \\ \Vert &&&&&& \Vert \\ \bftwo\times\bftwo &&& \xrightarrow[\textstyle \pi_1\wedge\pi_2]{} &&& \bftwo\times\bftwo \\ \end{array} thus \boxed{\pi_1\wedge\pi_2 = \wedge}, thus extension followed by restricting gives the original function.
To prove the other direction, that first restricting, then extending, gives the original, is more complex.
A proof of that, in greater generality, is at my discussion of the identities-are-free (Yoneda) lemma.
Pullback preserves the Boolean structure,
e.g., given a function in \Set, \functionf:\setX\to\setY,
completing the diagram at left below in either of the two evident ways yields isomorphic results
\begin{array}{}
& \rightarrow & \functionf^{-1}\objBp \\
& \searrow && \searrow \\
&& \functionf^{-1}\objB & \rightarrow & \objX\\
\\
& \objB\cap\objBp & \rightarrow & \objBp\\
&& \searrow && \searrow & \Bigg\downarrow\rlap\functionf \\
&&& \objB & \rightarrow & \objY \\
\end{array} \mkern1em for we have
\mkern1em \begin{array}{}
\eltx \in (\functionf^{-1}\objB \cap \functionf^{-1}\objB') & \xleftrightarrow{\textstyle \cap} & \eltx\in\functionf^{-1}\objB \wedge \eltx\in\functionf^{-1}\objB' \\
\\
\eltx\in\functionf^{-1}(\objB\cap\objB') && \updownarrow \rlap{\functionf^{-1}} \\
\llap{\functionf^{-1}} \updownarrow \\
\eltx\functionf \in \objB\cap\objB' & \xleftrightarrow[\textstyle \cap]{} & \eltx\functionf\in\objB \wedge \eltx\functionf \in \objB' \\
\end{array} \mkern1em or
\mkern1em \begin{array}{}
(\functionf^{-1}\objB \wedge \functionf^{-1}\objB')_\eltx & \xlongequal{\textstyle \wedge} & (\functionf^{-1}\objB)_\eltx \wedge (\functionf^{-1}\objB')_\eltx \\
\\
\big(\functionf^{-1}(\objB\wedge\objB')\big)_\eltx && \Vert \rlap{\;\functionf^{-1}} \\
\llap{\functionf^{-1}} \Vert \\
(\objB\wedge\objB')_{\eltx\functionf} & \xlongequal[\textstyle \wedge]{} & \objB_{\eltx\functionf} \wedge {\objB'}_{\eltx\functionf} \\
\end{array}.
Moving up several levels of generality and abstraction, there are the following bijections
(in the three examples of natural transformations, we show their components at \setX\in\Set\op):
Description | Bijections | Example | ||
---|---|---|---|---|
Internal subsets of \bftwo\times\bftwo | \Sub(\bftwo\times\bftwo) | \{\langle \radjtop,\radjtop \rangle\} \subset \bftwo\times\bftwo | ||
predicate/subset bijection | \Vert\wr | \wr | ||
Internal predicates on \bftwo\times\bftwo | \hom {\bftwo\times\bftwo} \Set \bftwo | \bftwo\times\bftwo \xrightarrow{\textstyle\wedge} \bftwo | ||
Yoneda bijection | \Vert\wr | \wr | ||
Natural transformations between indicated functors | \hom {\hom - \Set {\bftwo\times\bftwo}} {[\Set\op,\Set]} {\hom - \Set \bftwo} | \hom \setX \Set {\bftwo\times\bftwo} \xrightarrow{\textstyle \hom \setX \Set \wedge} \hom \setX \Set \bftwo | ||
defn. of internal product in \Set | \Vert\wr | \wr | ||
External operations on predicates | \hom {{\hom - \Set \bftwo} \times {\hom - \Set \bftwo}} {[\Set\op,\Set]} {\hom - \Set \bftwo} | {\hom \setX \Set \bftwo} \times {\hom \setX \Set \bftwo} \xrightarrow{\textstyle \wedge} {\hom \setX \Set \bftwo} | ||
predicate/subset bijection | \Vert\wr | \wr | ||
External operations on subsets | \hom {\Sub \times \Sub} {[\Set\op,\Set]} \Sub | {\Sub(\setX)} \times {\Sub(\setX)} \xrightarrow{\textstyle\cap} {\Sub(\setX)} |
No comments:
Post a Comment