"A person may licitly perform an action that he foresees will produce a good effect and a bad effect provided that four conditions are verified at one and the same time:Ficarrotta suggested that the justification or grounds for (3) -- something in the neighborhood of Kantian respect for a might make the majority of cases of 'collateral damage' morally impermissible -- even though the PDE is often invoked to justify such military actions.
(1) that the action in itself from its very object be good or at least indifferent;
(2) that the good effect and not the evil effect be intended;
(3) that the good effect be not produced by means of the evil effect;
(4) that there be a proportionately grave reason for permitting the evil effect” (1949, p. 43).
I'm interested in something else about the PDE; specifically, it seems like the Knobe effect shows that (2) is untenable. Roughly, the Knobe effect is: people judge bad side-effects to be intentionally caused, though people do not judge good side-effects to be intentional. The problem for (2) is obvious: if there's an evil side-effect of an action, then that side-effect will be judged to be intended. (Ficarrotta's own formulation of (2) perhaps makes the problem even more perspicuous: "Evil consequences are foreseen, but not intended." If the folk's attributions of intentionality are accepted, then there won't be any foreseen evil consequences that are not intentional.)
The New Catholic Encyclopedia version of (2) (again, this is in the Stanford Online Encyclopedia entry) actually reflects the Knobe idea: such actions are called 'indirectly voluntary', i.e., intentional, but somehow in a second-class sort of way.
And finally, a google search reveals that the PDE and the Knobe effect are briefly dealt with in a footnote to this paper by Jen Wright and John Bergson, which was recently posted to, and discussed on, the Experimental Philosophy blog. They make the good point that there is a further experimental question to ask the folk: do ascriptions of intentionality track blameworthiness or responsibility? Someone in a PDE situation appears to be responsible for the evil side-effect, even though she is not to be considered blameworthy on that account (at least, if the PDE is right).