If programmers thought like this, we would all still be using ENIAC.
The issue has to do with the violation of the terms of the contract. If the publisher wanted something unoriginal, then they would make an explicit request to publish a previous work. (In other words, the publisher would say to the author, "We want to publish your previous work X. Can we negotiate a price?") When no such request is made and the publisher instead offers to pay the author to write something, it is implied that they want something original. That implicit request is part of the contractual agreement between the author and the publisher. When the author publishes something unoriginal and then takes money for the service of publishing something that is original, then the author has effectively stolen the money.
I knew it was a sort of cop out, a disfavored option. But since it was my own work appearing under my own name in a signed opinion/satirical slot, I thought a re-run of some of my better, less-seen work was preferable to offering nothing. I didn't yet realize that 'self-plagiarism' was even a thing. I got the message from the disappointed tsk-tsk look/tone from my editor.
I think it wound up running with a disclaimer about having appeared before elsewhere. Does precise self-attribution cure self-plagiarism?