Syntactic characterizations of various classes of locally presentable categories (Q5939609): Difference between revisions

The word `syntactic' in the title of this paper is to be understood in the context of the `Gabriel-Ulmer duality' between small finitely complete categories and locally finitely presentable categories: an example of the former being thought of as the `invariant form' of the (essentially algebraic) theory whose category of models is the corresponding example of the latter. The authors' concern is thus to relate properties of essentially algebraic categories, such as those of extensivity and regularity, to properties of the corresponding `syntactic categories'; they also consider the corresponding problems at the `algebraic' level, where the syntactic categories merely have finite products. As an example, they obtain a new and very simple answer to the question `When is a variety a topos?', asked by the reviewer more than 15 years ago [\textit{P. T. Johnstone}, Algebra Univers. 21, 198-212 (1985; Zbl 0599.18003)]: namely, if and only if the idempotent-completion of its syntactic category is coextensive. (The more detailed, conventionally `syntactic', answer provided by the reviewer in the paper cited may be extracted from this one).