Questions framed as “Does not X ...?” “Is not Y ...?” “Should not Z ...?” and the like are certainly grammatical, and to me they seem every bit as unambiguous as the corresponding forms “Does X not ...?” “Is Y not ...?” and “Should Z not ...?” Nevertheless, they are not the forms normally used in contemporary English speech or writing; and when they do appear, it tends to be in the context of someone straining to sound serious and impressive—which, unfortunately, often translates into sounding stilted. I think of this form as “Letters to the Editor English” because that section of the newspaper occasionally publishes content from writers who, being uncertain of their voice, attempt to compensate with artificially formal diction.
The phrasing seems to arise out of back formations of the contractions that a person would use in everyday speech (“doesn’t,” “Isn’t,” “shouldn’t”) into their full-dress forms (“does not,” “is not,” “should not”)—but without any adjustment in word order to reflect how people generally use those uncontracted forms.
The result is text that reads like this excerpt from the Portland association notes in Bulletin of the National Association of Credit Men (June 1916):
In the Portland weekly letter these very searching questions are asked, bearing upon the acceptance of a compromise where fraud is recognized:
Does not this compromise place a premium upon crime?
Is not this kind of compromise exactly like fining a man twenty-five dollars for stealing one hundred dollars?
Is not this similar to taking morphine to cure the colic, disregarding the cause of the trouble, and taking chances of acquiring the habit?
Does not every settlement like this encourage others to perpetrate the same fraud?
Does not this kind of settlement almost insure for the participants further losses of the same nature in the future?
Is not this sort of settlement an injury and an insult to the worthy, steady, conservative, reliable customers who are paying one hundred cents on the dollar for all their bills?
The notion that the “Does not X ...?” was the standard form in earlier centuries is not strongly corroborated by a Google Books search. Search results yield countervailing examples at least as far back as Nicholas Rowe, "Some Account of the Life, &c. of Mr. William Shakespear" (1709):
Orestes imbrues his hands in the blood of his own mother ; and that barbarous action is performed, though not immediately upon the stage, yet so near, that the audience hear Clytemnestra crying out to Ægysthus for help, and to her son for mercy : while Electra her daughter, and a Princess (both of them characters that ought to have appear'd with more decency ) stands upon the stage and encourages her brother in the Parricide. What horror does this not raise! Clytemnestra was a wicked woman, and had deserv’d to die ; nay in the truth of the story, she was kill’d by her own son ; but to represent an action of this kind on the stage, is certainly an offence against those rules of manners proper to the persons, that ought to be observed there.
In the eighteenth century the “Does not this ...?” form is more frequent in Google Books results than the “Does this not ...?” form, but the total number of matches for both is very small. Neither form occurs in any great numbers in Google Books results until the first half of the nineteenth century, during which period (by my count) a Google search finds 13 unique matches for “Does not this ...?” and 30 unique matches for “Does this not ...?” indicating that the modern preference for “Does this not ...?” had already taken hold by 1850.
I didn’t run comparable searches for “Is not X ...?” versus “Is X not ...?” or for “Should not X ...?” versus “Should X not ...?” or for any other such pair of alternatives. Still, I feel fairly confident that the results for the pair I did check are representative of those for the larger population of such pairs.