Fighting off the 'were'-wolves

It's a good day, with the sun shining again and, over at Lingua Franca, a post by Geoff Pullum addressing Philip Corbett’s unnecessary anxiety about the was vs. were choice in the New York Times's columns.

As Pullum notes, many of us waffle on the choice of was or were in the subjunctive, or irrealis, construction. Merriam-Webster's Dictionary of English Usage quotes Swift, Byron, Thackeray, Frost, and other worthies opting for was where a purist would use were. (Byron: "I wish H. was not so fat.") 

Like these august writers, and like many educated users of English, I go back and forth in actual usage. I learned early on that it was OK to be blasé about was vs. were, thanks to Bergen and Cornelia Evans’s Dictionary of Contemporary American Usage (1957). Back in the mid-20th century, a lot of language authorities thought the subjunctive were was dying, and the Evanses were among them: "Was has been used as a past subjunctive in literary English for more than three hundred years and is the preferred form today." 

It's unlikely the Evanses had statistical proof for that "preferred form" judgment, and I doubt that it's even true. But given the number of subjunctive wases they saw in educated writing, they were confident in saying there were only two constructions in which were was clearly preferred: "If I were you," more or less a fixed phrase, and the literary "were I in a desert." 

We haven't abandoned subjunctive were as quickly as the midcentury mavens predicted. But if it disappeared overnight, as Pullum points out, there would be no loss of clarity; for other English verbs, the irrealis form is already the same as the simple past, or preterite: "If he had the money, he'd buy a generator." Is there any example in print of a failure to use the irrealis were causing genuine ambiguity? I don't think I've ever seen even a made-up example.*

By contrast, the recent blending of may and might – saying “if he had a raft, he may have survived” when you know that he didn’t survive -- can cause actual misunderstanding. (Way back in 2002 and 2003, Language Hat was bemoaning the development: "One of the changes going on in English that distresses me the most ... is the obsolescence of the contrary-to-fact past 'might have.'") Yet the takeover by may continues, and few notice the ambiguity. 

Why, then, should the meaningless was/were distinction persist, where no ambiguity ensues? I guess it's just one of those quirks of peevology.

*The myth of the "misplaced" only, on the other hand, has inspired dozens of made-up examples of alleged ambiguity, of the "I ate only the almonds/I ate the only almonds" genre. Yet I''m still waiting for a reader to show me an edited, printed example of an ambiguously placed only.