When we attribute a quoted statement to a speaker, we commonly write X said. And sometimes, less commonly, we invert it: said X. Someone has been teaching that the latter form is Wrong.
I know that when I see students and interns mark it unnecessarily on proof.
Bryan Garner is down on inversions: "Awkward are most, though not all, grammatical inversions. ..." He finds them precious, suggesting "amateurish literary writing," and he quotes Roy Copperud as chiming in. And, to be sure, any device can quickly turn into an irritating tic.
But there are times when the inversion in journalism makes sense: "There are times when the inversion in journalism makes sense," said John E. McIntyre, the night content production manager at The Baltimore Sun.
Avoiding the inversion when the attribution includes an appositive phrase produces a limp little anticlimax: "There are times when the inversion in journalism makes sense," John E. McIntyre, the night content production manager at The Baltimore Sun, said.
That particular bit of awkwardness is apparently house style at The New Yorker.
I know why teachers convert guidelines into rules and editors proclaim ukases; it's simpler for writers to remember a rule than to apply judgment. But really, the principle that should underlie all guidelines, rules, and ukases is Don't annoy the reader.