Oops, if only there were a "this looks like another article" warning button. "Are you sure you want to spam our readers?"
That could actually be done. The title of the self-referential post could be put through some Bayes analysis to list likely recent duplicates. Those could be presented to the creator of the self-ref post giving a chance to delete right away.