Languagehat links to a couple posts found over on New York Magazine's Vulture branch. It's all about a squabble brought about by a copy editing topic on HBO's Wire.
Jump to the comments. NNALL (whose comment is an "Editor's Pick") adds some observations and opinions regarding the extent to which some copy editors are driven by their idiosyncrasies. One claim caught my especial attention (influenced certainly by the fact that languagehat also found it worth quoting).
Technically you can't perform an autopsy on a different species, so if a racehorse died mysteriously and some reporter wrote that veterinarians hoped to learn more after an autopsy, it would be changed to 'necropsy' or 'post-mortem examination.'
I'm not sure what "Technically" means in this case. That's often used for an etymological argument. You might have heard a claim like technically, 'jejune' means 'empty' and I'm sure you've heard something like technically 'decimate' means get rid of one tenth. (This last topic has been booming on ADS-L for about a week and a half.)
So what is it that technically keeps a pathologist from performing an autopsy on anything but another human? Is it like saying that homicide can only be murder of a human because the homi- part means man/human? There's nothing in the etymology or autopsy that refers to humans. Or is it the same species part? The comment says the issue is "a different species" so does this mean that if horses were smart enough (and had opposable thumbs) they would be able to perform autopsies on other horses?
There's aut(o)- meaning 'self' and there's -(o)psy related to optic. From that we get the seeing for oneself sense of the word. I don't understand from where this prescription claims its logic. Can't a veterinarian see for herself why a racehorse died? And please don't offer the obvious explanation that prescriptions are rarely based on good logic. Copy editors often have a decent argument for their style preferences. What's the argument here?