An apostrophe generally indicates an omission, but in this case I would favour shan't over sha'n't for readability and consistency.
Sha'nt looks wrong, and I've never heard that their should only be a single apostrophe at the first omission.
Looking for a "general rule", an article called That Cute Li'l Ol' Apostrophe that claims:
We never use more than one apostrophe to a word.
While the general rule is to use the apostrophe in place of the last missing letter, such as in "shall not -> shan't", if we need to choose between missing letters that we'd normally pronounce and those that are silent, use the apostrophe to denote the missing sounds.
Another claims the following:
Well. at first glance, it appears to me that our way of ‘making’ contractions is to 1) lop off the last half of the first word and 2) smash it together with the “not”, contracting the “o” with an apostrophe. See for yourself…
shan’t= shall (minus the “-ll”) + not (minus the “o”) = sha n’t
Which is then moved together (sha->n’t) to spell: shan’t. Personally, I believe that this contraction (judging by the way we use the word, and say it) is an ‘evolved creature’ from the two contractions “shouldn’t” and “can’t”; as opposed to its parentage being shall and not. It just makes more sense. [Shouldn’t + can’t= shan’t]
The OED lists both shan't and sha'n't as colloquial contractions of shall not, but not sha'nt. Shan't appears in 139 quotations, sha'n't appears in 25, and sha'nt is in only five.
Project Gutenberg's out-of-copyright books are usually older and don't necessarily reflect contemporary use, but searching their August 2003 CD of 600 ebooks: there are 589 results in 103 books for for shan't, 122 results in 29 books for sha'n't, and only three results in two books for sha'nt.
Another common contraction, won't, comes from woll not (an archaic version of will not). It also has two chunks of letters omitted. Should this be wo'n't or wo'nt? Motivated Grammar writes:
Did the contractions won’t and shan’t spring into English fully formed, like Athena from Zeus’s noggin? No, interestingly. Samuel Johnson’s Dictionary (printed in 1855), has wo’n't, as do some (modern) editions of Lewis Carroll’s Alice in Wonderland and The Ohio Educational Monthly in an article from 1868. Likewise, sha’n't was commonplace in the old days, plastered across the pages of the dreadful Victorian novels that I had to read in AP English as a lesson as to what happens to those who show an interest in reading. Books like Evelina; or, The history of a young lady’s entrance into the world (why did every single book in those days have to have a subtitle?)
Now the interesting thing is that won’t and shan’t live side-by-side with wo’n't and sha’n't in these old books. Some quick results on Google Books between 1600 and 1800: 777 won’ts, 57 wo’n'ts; 216 shan’ts, 73 sha’n'ts. Between 1600 and 1700: 48 won’ts, no wo’n'ts; 1 each of shan’t and sha’n't. So it seems it was never the case that the multiple-apostrophe form was more common. For some reason or another, English writers have always preferred a single apostrophe over strict application of “put apostrophes wherever a letter’s missing”. (Michael Quinion guesses that the double-apostrophe form was a later edition, suggested by logic-minded grammarians, that died out because it was a pain to write and looked weird.)
Quinion pointed out shan't is actually older than sha'n't and summarised:
The abbreviation, as you say, strictly demands the extra apostrophe, and it was probably the influence of logically minded eighteenth-century grammarians who persuaded many people to put the extra one in to start with — but whenever did logic ultimately matter in language?
So: the rules aren't really clear, but shan't is the most common, sha'n't is somewhat old fashioned, and sha'nt is extremely rare.
Why isn't English consistent?
Because it's evolved from a big mish-mash of several other languages over some 1,500 years.
Why isn't Winnie the Pooh mandatory reading for all English speakers?
Because nothing is mandatory reading for all English speakers. If Winnie the Pooh is mandatory reading, what else should be mandatory? Where do you stop? The only mandatory reading for English speakers should be the English language.
The answers to your questions are:
- Yes, the shift was and continues to be a gradual and chaotic process.
- No, there was not a a deliberate and possibly collective decision taken by the educational institutions, the media, or the government of the time.
Yes, there were a few changes that Webster tried, but the history of English spelling is much, much, much more complicated than that. English has never had a single agreed-upon spelling, even within any one country. There is no such thing as the American spelling, nor for that matter is there the British spelling, when discussing the English language. At most, there are individual words that may show a general predilection, but that shifts over time and within writers and publishers. It isn’t a simple us-vs-them situation almost in any pairing. There is too much historical variation.
Here is a summary of my findings below sorted by year:
always inflection, inflexion
always paralyzed, paralysed
always tire, tyre
1680 surprize, surprise
1770 surprise, surprize
1810 jewelry, jewellery
1835 colored, coloured
1835 deputize, deputise
1845 armor, armour
1845 connection, connexion
1905 aluminum, aluminium
1905 woolen, woollen
1910 esophagus, oesophagus
1910 fetus, foetus
1915 leveled, levelled
1920 center, centre
1930 savior, saviour
1940 miter, mitre
1940 specter, spectre
1970 acknowledgment, acknowledgement
1975 theater, theatre
1980 scepter, sceptre
never ameba, amoeba
never glamor, glamour
There is a lot more variation than even that summary shows; please see the more detailed charts below, including some that go the other way by showing British usage. Each of these really is a separate case, and must be considered individually. There is no single, solitary, discrete, and agreed-upon thing called American spelling versus British spelling. There is a vast continuum of conflicting tendencies, some weaving back and forth time across the Atlantic, and across time. Even when one spelling “supersedes” another, it merely comes out ahead in the popularity contest. It usually does not wholly supplant the former spelling.
For the Ngramaniacs
In all cases, the putative “American” spelling is in blue, and the putative “British” spelling is in red. Notice how the answers are all over the board, and some rather surprising, too.
center vs centre
When did center supersede centre in American English? Answer: Around 1910. Ngram
theater vs theatre
When did theater supersede theatre in American English. Answer: During the late 1970s. Ngram
armor vs armour
When did armor supersede armour in American English? Answer: Around 1850. Ngram
glamor vs glamour
When did glamor supersede glamour in American English? Answer: Never, because it has always been a minority usage. Ngram
woolen vs woollen
When did woolen supersede woollen in American English? Answer: around 1905. Ngram
tire vs tyre
When did tire supersede tyre in American English? Answer: Never, because it was always dominant. Ngram
Well, and when did tyre supersede tire in British English? Answer: in the early 1940s. But it only lasted a couple of decades before falling back into minority usage again. Ngram
miter vs mitre
When did miter supersede mitre in American English? Answer: Around 1920. Then again around 1940. Then again around 1970. Ngram
acknowledgment vs acknowledgement
When did acknowledgment supersede acknowledgement in American English? Never. It has always been that way. Ngram
What about the other way around? When did acknowledgement supersede acknowledgment in British English? Answer: Around 1970. Ngram
scepter vs sceptre
When did scepter supersede sceptre in American English? Answer: Around 1980. Ngram
savior vs saviour
When did savior supersede saviour in American English? Answer: Around 1930. Ngram
deputize vs deputise
When did deputize supersede deputise in American English? Answer: Around 1810. Ngram
But what about the other way around? When did deputise supersede deputize in British English? It’s arguable, but the general answer is that it never did. Ngram
ameba vs amoeba
When did ameba supersede amoeba in American English? Answer: Never. Ngram
jewelry vs jewellery
When did jewelry supersede jewellery in American English? Answer: Around 1810. Ngram
inflection vs inflexion
When did inflection supersede inflexion in American English? Answer: It probably never did, because it inflection was probably always the dominant spelling of that word in American publications. Ngram
The data from before 1800 is not of sufficiently high quality to use for drawing conclusions from.
But what about the other way around? When did inflexion supersede inflection in British English? Answer: Mostly never.
Or, depending on the quality of the data, perhaps around 1805. But if so, it only did so for 7 years, up through around 1812. There was a minor and short-lived blip in the early 1920s when inflexion seems to have edged out inflection again in British English for a couple years tops, and then again between 1948–1962. Since then, the inflexion spelling has been on wane in British publication such that by 2000 the inflection spelling dominated the inflexion spelling by a factor of more than 5:1. Ngram
But what about the other way around? When did connexion supersede connection in British English? Answer: Around 1820. But it only did so for around 30 years, up through around 1850. Ngram
colored vs coloured
Since this one seems to be everybody’s favorite peeve, when did colored supersede coloured in American English? Answer: Around 1840. Ngram
leveled vs levelled
When did leveled supersede levelled in American English? Answer: Around 1915. Ngram
fetus vs foetus
When did fetus supersede foetus in American English? Answer: around 1910. Ngram
Hm, but what about the other way around? When did foetus supersede fetus in British English? It didn’t: it gave up the lead around 1970. Ngram
paralyzed vs paralysed
When did paralyzed supersede paralysed in American English? Answer: Always. ngram
And the other way around: when did paralysed supersede paralyzed in British English? Answer, around 1830, but it has been losing a lot of ground recently. Ngram
specter vs spectre
When did specter supersede spectre in American English? Answer: around 1940. Ngram
esophagus vs oesophagus
When did esophagus replace oesophagus in American English? Answer: Around 1910. Ngram
But what about British usage? The answer is that oesophagus seems to have lost to esophagus around 1980, but then may have returned. It’s hard to tell. Ngram
aluminum vs aluminium
When did aluminum supersede aluminium in American English? Answer: around 1900. Ngram
And what about the other way around? When did aluminium supersede aluminum in British English? Answer: around 1850, but it has lost a great deal of ground of late. Ngram
surprize vs surprise
It turns out that surprize was once the dominant form, not surprise, but this didn’t last even a hundred years. Ngram
Then again, here is the British Ngram for the same pair. Notice that even the Brits used the z for a good while, possibly even for longer than in North America:
Summary
I hope you can now see why I said what I said: that there can be no clear answer here. Everything is different. You have to look at each word-pair separately, and you should make sure you aren’t wrong about Britain, either.
Best Answer
The OED shows many forms of be that have occurred over the centuries, including numerous negative forms. Ben’t is among them, having been in use from the seventeenth century onwards, and it may still be heard in some parts of England today. It has probably always been a non-standard form, in so far as it is possible to make such distinctions before a fully standardised form of English emerged.
There are similar forms which may represent similar pronunciations, so it is may not be clear which one the maid actually used.
The supporting citations show it being used with I, with the meaning I am not, but that doesn’t necessarily mean it is not found with other persons.