Graph models of untyped \lambda -calculus
The untyped
Graph models never satisfy the
Definition
A general definition of what a graph model is can be given through Scott-Engeler algebras (SE algebras). An SE algebra consists of a set
i : \mathcal{P}_\text{fin}(D) \times D \to D
An SE algebra is extensional just if
The map
The
(P(D), \subseteq)
Each function will be mapped to the set of its 'finite approximants' of the form
f \cdot x \triangleq
\{ y \mid \exists S \subseteq_\text{fin} x.\ i(S, y) \in f \}
Examples
\mathcal{P}\omega - Scott-Engeler models are essentially graph models where
i is an inclusion.
References
- A review article:
@article{plotkin_1993,
title = {Set-theoretical and other elementary models of the λ-calculus},
volume = {121},
doi = {10.1016/0304-3975(93)90094-A},
pages = {351--409},
number = {1},
journaltitle = {Theoretical Computer Science},
author = {Plotkin, Gordon D.},
date = {1993},
}
- A review article on webbed models of untyped
\lambda -calculus, which include graph models, is
@article{berline_2000,
title = {From computation to foundations via functions and application: The λ-calculus and its webbed models},
volume = {249},
doi = {10.1016/S0304-3975(00)00057-8},
pages = {81--161},
number = {1},
journaltitle = {Theoretical Computer Science},
author = {Berline, Chantal},
date = {2000},
}