@Mitch Strange.

So anything that isn't American is "wrong" when written in America? And also when it is American but older?

@Cerberus What?

@Rob @Mitch thanks again. Just sent it off now. Fuck, can you tell from that how much I care about proofreading.

Anyway it's just two words in a million-word private message to just one person nobody even knows. Just some American trombone player twice Rob's age. But it actually means the world to me. I guess things like that happen. So thank you.

@Cerberus That's just silly.

@RegDwigнt Of all that message you sent to him, of all that attempt at a connection, the one thing that they'll catch on and judge you on... 'nigh any'.

@Mitch That's probably not what you intended, but it is what it may sound like.

@Cerberus possibly. But 'anything' seems like quite a stretch to infer.

If there's a question "Is X correct?" then you specify the context.

At any rate, I think it's best to carefully word generalisations based only upon personal observation, when trying to prove a negative. Even better would be a quick check in Google Books, possibly in order to qualify the generalisation.

@Mitch But was there really a question?

"Is 'X' correct?" - In place(s) Y during time(s) Z by subculture(s) Z

You commented voluntarily on a word quoted (not said) by someone else.

@Cerberus You have to be very careful with context (the specific search too) in NGrams

comparing likes with likes

and base rates

I recommend Books, not Ngrams, to check whether, as one hypothesises, no serious person would use x.

@Cerberus NGrams gives the timeline, and then links to Book entries

it's the same site

If one finds an acceptable counter-example in the first couple of results pages, one could qualify one's own generalisation.

and search situation.

the graphs are based on the corpus that a Books search returns from

Books allows you to check the context.

Ngram has no context.

Hence my recommendation to use Books.

They're the same thing

We all know that.

so by 'use books' you mean 'look at the hits too'

I mean, use the subdomain books.google.com and check the results pages.

How you arrive there is up you.

You can go there by clicking on a link from Ngrams, or just search Books directly.

@Cerberus I prefer to start with the relative frequency graph (because that's the usual direct answer to most questions, then check the instances to make sure what is being found is what you're thinking of, as a check. Just seeing a page of instances, just the existence of a handful of hits is pretty paltry. Often those hits are from the same source, and could easily be OCR errors, or misdated reprints of much older things.