There's no faith required; the planner is guaranteed not to do that. The "normal" way is to create a composite type containing each of the columns you need, and then "unpack" it to separate columns. Horrible? Yeah, but it's possible.
Is it possible with scalar subqueries to perform anything other than a tree of correlation? With CROSS APPLY one can correlate a DAG of subqueries, e.g. a diamond where B and C depend on A, and D depends on B and C.