scientific article; zbMATH DE number 3886100

From MaRDI portal
Publication:3220729





[For the entire collection see Zbl 0546.00024.] This excellent survey article addresses the question of viewing a topos as a generalized space. Since its birth in SGA 4 and the subsequent pioneering work of Lawvere and Tierney, the scope of topos theory has greatly expanded and it is distinguished by its generality and the diversity of its examples and applications. However, as a result of the represenation theorems outlined in this article, the notion of Grothendieck topos can be described in terms of locales and localic groups and groupoids. After an outline of the necessary preliminaries on geometric morphisms and geometric theories, the article proceeds to the two main open covering theorems. The first one is an improvement by André Joyal of Barr's theorem. It states that if \({\mathcal E}\) is a Grothendieck topos, there exists a connected, locally connected geometric morphism \({\mathcal F}\to^{f}{\mathcal E}\), where \({\mathcal F}\) is localic over \({\mathcal S}\) (the base topos of sets). The second major result is Peter Freyd's theorem that given a Grothendieck topos \({\mathcal E}\), there is a connected, atomic geometric morphism \({\mathcal F}\to^{f}{\mathcal E}\), where \({\mathcal F}\) is localic over \({\mathcal C}(G)\), the topos of continuous G-sets, where G is a topological group. The author also briefly discusses the result of the reviewer, which states that every Grothendieck topos \({\mathcal E}\) admits a hyperconnected geometric morphism \({\mathcal F}\to^{f}{\mathcal E}\) where \({\mathcal F}\) is an étendue, i.e. \({\mathcal F}\) is locally localic. The covering theorems of Joyal and Freyd lead to representation theorems for Grothendieck topoi. Joyal and Tierney have shown that every Grothendieck topos is equivalent to a topos of \({\mathfrak G}\)-sheaves, where \({\mathfrak G}\) is a localic groupoid. On the other hand, from Freyd's result it follows that a Grothendieck topos is equivalent to an exponential variety (full subcategory closed under limits, colimits, and power objects) in a topos localic over continuous G-sets, where G is a topological group. These results provide some justification to the view of toposes as generalized spaces. The author is to be commended for compiling these results about the basic nature of Grothendieck toposes into this highly readable article. Survey articles such as this one serve the valuable function of making results from diverse sources more readily accessible as well as providing a perspective on the interrelationships of these results.











This page was built for publication:

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3220729)