After a few google tries, http://c2.com/cgi/wiki?DanaScott sadly only told me what I already knew:
"...Scott's key construction was a solution to the "domain equation" D corresponds to [D -> D] thus giving the first mathematical model of the type-free LambdaCalculus. This led to a general theory of solutions of recursive domain equations..." .
But the first try, googling for just "scott's construction" got me http://groups.msn.com/ScottsHornayHoneyz. "Home of the Honeyz", y'all.
Hm, this paper is kind of what I was looking for, although Paul Taylor is a little too bad-ass for my mortal understanding of category theory.