I would still like to see a simple algebraic proof based on the fact that a cyclic H-module is flat, for a Heyting alg H. Andy's proof didn't seem all that easy to me. The background is that there is a notion of H-module and the pushout of a pair of Heyting algebra maps out of H is also the tensor product of H-modules. By an H-module I mean a sup semi-lattice that has unary ops x /\ - and x ==> - for each element x of H, and satisfying the obvious identities. Now for a ring R if every cyclic R-module is flat, so is every R-module. Here, every cyclic H-module is flat. Are they all? Michael ======================================