8 Jun
2001
8 Jun
'01
12:37 p.m.
< Choice isn't needed: all you need is the result that, for any filtered < category C, there is a directed poset P and a final functor P --> C. < There is a proof of this somewhere in SGA4 (I don't have the reference < to hand), where it is attributed to Pierre Deligne; but I suspect it < may be older than this. < Peter Johnstone The proof is also in my book with Adamek, Locally presentable and accessible categories, Theorem 1.5, Jiri Rosicky