You are correct that that is what I would do. But these are rules of type 1, we can derive them from common usage for the formal written register of an academic research paper; the general community of people reading and writing such things would actually be following these rules already in that context, as opposed to the various sorts of prescriptivist rules from category 2 which I’ve referenced, which people only follow now and then when it pops into their head to think to do so or in response to remarks from an editor who’s been trained (for no good reason) to have these rules pop into their head. When you violate the rules of type 1, the readers you are addressing are bound to notice the jarring deviation from the expected style. When you violate the rules of type 2, not so much (because these aren’t actually normally naturally followed; if they were, we could observe them as common usage and they’d be of type 1).
And to reference the OP a little, I really don’t think you’d find “alright” to be rare compared to “all right” in academic literature, certainly not enough to make it jarring to read or condemnable to write.
Above I’ve just re-requoted what you requoted in your previous post, as you are right to imply that it speaks directly to the issues I was raising. Sorry I missed it before!
In the above, you say “Don’t say ain’t” is a rule of type one, as long as we understand that it ranges only over some communities, in some contexts. Okay, I get it now. I can talk this way.
In my previous posts I was ascribing views to you which you do not subscribe to. My apologies.
I think there is a little bit of a problem in the notion that empirical research can distinguish between proper use and a gaffe, unless “proper” is just defined in terms of statistical majority or something like that. Is that how it’s done? Otherwise, the best I can imagine is that empirical research could conclude “Don’t say ain’t in context Y” is a rule on the basis of the fact that speakers tend to admonish each other not to say “ain’t” in context Y–which, you’ll notice, isn’t a fact about the usage of “ain’t” but rather a fact about what speakers say about the usage of “ain’t.” That can’t be the way it’s supposed to work, since allowing for rules based on what speakers say the rules are supposed to be is exactly what anti-language-mavenism is aimed not to allow.
Also, it seems to me that there can be cases in which a usage is inappropriate, because people say its inappropriate, even though in fact, the rule about its being inappropriate is often broken without anyone noticing. An analogy would be something everyone says is wrong, but that most everyone does at least occasionally anyway.
But I’m having a hard time coming up with examples (which is usually a bad sign ) so I’ll just have to go think about it some more.