The final check is a computer program that compares the paper's text with the text of every other paper already published on arXiv.Emphasis added. Excluded: books, paywalled papers, unpublished papers, lectures. The limits of this check are so obvious, that it's a wonder any significant cheating was discovered at all. But is this kind of thing at least useful for a snapshot?
We need much more such policing in Civil War history and social science generally. There is the problem of plagiarizing others and then of plagiarizing one's own work.
Academics have the luxury of blaming graduate assistants for their crimes but others are not so lucky.
And although an Allan Nevins could confidently steal Fremont research from an unpublished paper, one wonders at the mentality of a pop culture celebrity who thinks his theft from a recent issue of the New Yorker would go unnoticed. Hell, who among his TV fans reads at all?
Thieves. A friends' house was burgled in the night and the next day, he found the stolen stuff scattered around the garden, under bushes and in the driveway. It made no sense. He asked the police, "Why?" He received a look of scorn. "These aren't rocket scientists, you know."
Plagiarists: not rocket scientists. Not even garden variety scientists, really.