\(\newcommand{\sem}[1]{[\![#1]\!]}\)

Semantics of Coalgebraic Logic#

(draft)

We first review the definitions and results that make sense for the general setting of coalgebraic logic.

Syntax and Semantics#

one-step syntax

\[\mathcal A\stackrel L \longrightarrow \mathcal A\]

syntax is the initial algebra

\[LI\longrightarrow I\]

one-step semantics

\[LP\stackrel \delta \longrightarrow PT\]

semantics wrt \(X\to TX\)

For the next result we assume that \(\mathcal X\) is a concrete category. Various generalisations to abstract categories are also possible.

Proposition: Formulas are invariant under bisimilarity.

Proof: Follows from the naturality of \(LP \to PT\).

The Logic Functor Induced by the Coalgebra Functor#

The definitions above are parametric in \(L\) and \(LP\to PT\).

On the other hand, from a coalgebraic point of view, \(T\) is given and we should ask how to derive \(L\) from \(T\).

If \(\mathcal A\) is a variety in the sense of universal algebra, then we can define a functor by its action on the free algebras. Moreover, if the variety has a presentation by operations of finite arity, then we can restrict attention to finitely generated free algebras and define \(L\) as

\[LFn=PTSFn,\]

where \(n\) ranges over finite sets.

(to be continued)