This orphaned posting seems to have attracted no comments from others and I have nothing to add. But that doesn't mean I didn't read it with interest.

I'm reminded of the suffix -ois in French, which has mostly become pronounced (and written) as -ais in words such français and anglais but survives in personal names derived from these: the first name is always François and not Français, and I've only seen the family name Angloys written like that. I can't immediately think of any nationalities that still end in -ois, but plenty of words relating to cities, like lillois

What has always struck me as odd about these is that -ais sounds quite different from -ois, and ят sounds quite different from -ют, even unstressed, so I wonder what drives these changes. Were they originally affectations of a small part of the population?

How is Angloys pronounced (in England)?

Athel, French words finishing with -oi or -ois were pronounced -oué in some distant past. There is an old song in which this pronunciation has been kept: La rose au boué (La rose au bois).
http://fr.lyrics-copy.com/chansons-populaires/la-rose-au-boue.htm
Roi (or roy) was famously pronounced [rwe].

How is Angloys pronounced (in England)?

I don't think it's a name you come across much in England, any more than you come across people called French in France.

We do have some Langlois here. Some François (patronymic) as well, in larger quantities.

I've met someone in Quebec whose name was English, en anglois dans le texte, which is quite weird when you think of it.

The paradigmatic nationality in -ois is danois. This is an unconditioned split, which means nobody has a clue why it affects some words and not others, like /æ/-tensing in Philadelphia.

I don't think of 'English' as a weird surname. Consider an Englishman living in Scotland, Wales or Ireland at the time his surname became fixed. Everyone calls him 'English'. Does that make sense? It's just a guess.

oi

In Old French and Early Middle French this diphthong was pronounced [oy], and the words containing it which borrowed into Middle English still have it, as in choice from OF chois (now written choix), noise from OF noise 'annoyance', and many others. So an English person named Angloys would pronounce [oy] as in boy.

Later the diphthong [oy] evolved into [we] or [wE] (here E = the letter epsilon representing roughly the pronunciation of written è), and in low-class Paris this new diphthong usually evolved into [wa], which later became generalized to Standard French. But some instances of [wE] became [E] instead.

It seems that [wE] became [wa] in cases where it was inseparable from the word, as with the examples above, but lost the [w] to become just [E] when it was perceived as a suffix: in the pair François 'Francis' (the male name) and Français 'French', both from the Latinized Germanic name Franciscus (cf Italian Francesco), the first one is perceived as an unanalyzable name and has kept the diphthong, but the second is obviously derived from the name of the country France and the suffix has lost the [w] sound (this applies also to the female counterparts). This is a general rule, although there are a few exceptions which are more or less explainable: in the case of danois 'Danish' (nationality), the word is not derived directly from the name of the country 'Danemark', only from a part of the name which does not exist independently, so that the adjective is not perceived as directly related to the country name. (But this reasoning is not valid for suédois from Suède, unless the suffix is analogical to that of danois).

For names of inhabitants of provinces or cities, regional preferences undoubtedly played a role, as well as the extent of the spread of change [wE] to [E]: those dialects where the change had not occurred, or only in verb endings, for instance, would have converted the preserved [wE] to [wa] under the influence of the new standard, and there was no competition from the new [E].

English, French, etc as last names: In French there are names like Langlois, Lefrançois, Lallemand, Lebreton, Lenormand, Lepoitevin, etc which obviously incorporate an adjective referring to national or regional origin (English, French, German, Breton, Norman, Poitou inhabitant, etc), preceded by the article le like other descriptive adjectives, as in Lebrun, Leblond, Lenoir, Leblanc, Legris, Levert, Lejeune, Lagacé, etc. In every case the name means The ... one. In English or German such adjectives are also used as last names, but without the article.

Some (informal) dialects of Canadian French still have [we] or [wɛ] in some words, don't they? I think I hear that on the Boston subway (we're the nearest largish American city and popular for school bus tours).

I thought the discussion at Anatoly's blog was pretty interesting - I was unaware, for example, that Moscow speech was apparently considered more or less the acceptable standard in pre-revolutionary Russia rather than St Petersburg speech. There also seems to be dispute about whether "Khodyut" was really still acceptable in the early 20th century in "better" social circles. But we do have actual recordings of people like Tolstoy and Bunin speaking Russian - it seems like it would be fairly easy to establish how they pronounced unstressed -ят.

MMcM: Some (informal) dialects of Canadian French still have [we] or [wɛ] in some words, don't they?

If they have this pronunciation in some words, then they should have it in all relevant words. An exception would be if they had learned some of these words in a school context where standard [wa] was prevalent, and the others in the family, etc context which retained the old pronunciation. There would be no other reason to change it.

@Athel, in casual speech, unstressed -ют and -ят could be quite indistinguishable.

Vladimir Vysotsky uses the -yut ending (and similar "ungrammatical" things) in a lot of his songs, obviously for humorous effect, along with a lot of other interesting features, like frequent okanie and yakanie, as well as "vo" where "v" is required by the accepted rules. This kind of thing is meant to indicate "negramotnost'" (pun intended), and points to the existence of such forms in dialects.

I disagree with Steve that -yut and -yat would be indistinguishable in casual speech, since yu is never reduced to i, the way ya is. I'm pretty sure I would notice khodyut instead of khodyat.

No one has ever been "like Tolstoy", Vanya. Tsk.

My knowledge of Russian is pretty poor but I do remember reading that /(j)u/ was the only vowel that did not reduce to /a/ or /i/ when unstressed.

I don't think it's an issue of the sounds being indistinguishable. The most common verb ending in the third person is -ут (ют). Applying that rule to other verbs is a bit like saying "I winned" instead of "I won." People joke around with this today, particularly with the verb хотеть (to want) -- хочут instead of хотят. But I'm sure that first form was standard in some regions (and may still be today).

Yes, I was going to say what mab did: it's not a phonetic thing, it's regularization across conjugations.

"it's not a phonetic thing, it's regularization across conjugation"

Okay, that makes sense. I do think that point could have been clearer in the post.

Could the 3rd pl. unstressed -yut ending for verbs in -i- be explained at least in part because the unstressed 3rd plural -yat ending is indistinguishable from the 3rd singular ending in -it due to vowel reduction?

Bill--

I don't reduction has anything to do with it, since these endings occur in dialects without reduction, as well as in other languages (Polish, etc.) without this kind of reduction, and therefore supposedly pre-date reduction.

Some (informal) dialects of Canadian French still have [we] or [wɛ] in some words, don't they?

If I may start a little topic drift... My meanwhile famous thesis supervisor pronounces oi the same way as the Parisians around him (and this includes poêle and moelle), but oin is to him a diphthong of o followed by his in sound (which is [ẽ] rather than [ɛ̃] or [æ̃] or suchlike). I find this especially interesting because to native Parisians moi(s) and moins is a minimal pair for nasality, that is, oin is the exact same thing as oi except that it's nasal.

I pronounce -oin as o + in ([æ̃]), for what that's worth.

My Montréal friend and I used to joke like this in the 1980's:

Ç'toé? Ouais, ç'moé.