To quote Wikipedia: “Seediq, also known as Sediq, Taroko, is an Atayalic language spoken in the mountains of Northern Taiwan by the Seediq and Taroko people.” That article is impressively detailed for such a minor language, and the Phonology section hints (“The stressed syllable is usually the penultimate one, and is pronounced with a high pitch. In the Truku dialect stress is on the final syllable resulting in loss of first vowel…”) at why there are such varying pronunciations here ([seˈʔediq], [səˈdiq], [səˈʔəɟiq]). But the reason I bring it up is that I discovered it via a link to Indigenous Cultural Translation: A Thick Description of Seediq Bale:
Indigenous Cultural Translation is about the process that made it possible to film the 2011 Taiwanese blockbuster Seediq Bale in Seediq, an endangered indigenous language. Seediq Bale celebrates the headhunters who rebelled against or collaborated with the Japanese colonizers at or around a hill station called Musha starting on October 27, 1930, while this book celebrates the grandchildren of headhunters, rebels, and collaborators who translated the Mandarin-language screenplay into Seediq in central Taiwan nearly eighty years later.
As a “thick description” of Seediq Bale, this book describes the translation process in detail, showing how the screenwriter included Mandarin translations of Seediq texts recorded during the Japanese era in his screenplay, and then how the Seediq translators backtranslated these texts into Seediq, changing them significantly. It argues that the translators made significant changes to these texts according to the consensus about traditional Seediq culture they have been building in modern Taiwan, and that this same consensus informs the interpretation of the Musha Incident and of Seediq culture that they articulated in their Mandarin-Seediq translation of the screenplay as a whole. The argument more generally is that in building cultural consensus, indigenous peoples like the Seediq are “translating” their traditions into alternative modernities in settler states around the world.
Wow, does that sound interesting! Had I but world enough and time…
why there are such varying pronunciations
Thank you @Hat. Indigenous Taiwaners are these days known as ‘mountain people’, but this is misleading as to how their languages evolved. (There’s more variety in indigenous languages than in the whole of the rest of Austronesian.)
Before European contact, the indigenous people lived spread out on the plains, and only went to the mountains for hunting. I guess these ‘varying pronunciations’ reflect regional accents. The Portuguese/Dutch/Chinese invaders/indentured labourers took the plains (chiefly for growing sugar cane initially), pushing the indigenes into the marginal/upland valleys. Usual story.
And then the Japanese plundered the mountains for timber, so clashing with their last strongholds.
When I lived in Taiwan I was curious about the indigenous people and their languages, but never got around to investigating — too busy trying to navigate work and life.
It seems tendentious to use “Truku” to refer to the dominant dialect of the “Taroko” language when it seems 99%+ likely that these are just alternative transliterations (mediated via various non-Roman scripts) of the same word. Some of the fatalities from the recent earthquake in Taiwan were in the Taroko Gorge national park, which I’ve visited and is lovely but not the sort of place you want to be when landslides etc. are being triggered by seismic events. (I think somewhere in that park is where I managed to snap a photo where two different road signs are in frame using rival romanizations of the same nearby toponym, as sort of a synecdoche for Taiwan’s general glorious cacophony of inconsistent romanization.)
Separately, I am somewhat confused by what exactly is meant by the vogueish college-town locution “settler state” in what hat has block-quoted, because in the context being discussed you have three separate groups: a) the indigenes with their Austronesian L1’s, by 1930 mostly driven up into the hills; b) the Han people who had arrived over the prior three centuries from Fukien/Hokkien under varying political-etc. circumstances and who had eventually taken over the lowlands; and c) the Japanese, who were at the time the imperial/colonial overlords of both a’s & b’s. Obviously “settler” is pejorative (I think you’d say “migrant” otherwise?), but who is supposed to be cast as the villain here? Or is it that the a’s are now engaged in historical revisionism about their actual historical relationship with the c’s (sometimes “rebel” but other times “collaborator”?) in order to improve their current tactical stance vis-a-vis the b’s?
Whether you agree with the politics or not, how is “settler state” confusing? It’s (maybe) the Chinese, then (maybe) the Portuguese, then (maybe) the Dutch (& maybe the Spanish), then the Chinese, then the Japanese, and then the Chinese again.
[edited: the above makes it sound like I’m skeptical that Taiwan should be considered a settler state, which, to be clear, I am not]
@Matt, you have that chronology wrong.
After the Austronesians (there for millennia) came the Portuguese, then the Dutch. There might have been Chinese/Muslim traders passed by the island before European contact, but it was the Europeans who brought the Chinese as indentured labour. (The Portuguese, as usual, took a relatively hands-off approach.)
So unlike in Mainland China, non-Sinitic languages dominated until ~C16th.
Then I kinda agree with @JWB’s drift that “settler state” is a tendentious label. OTOH C16th is no earlier than most states that come to mind with that label, and where the bloody Europeans (the ones I would cast as villains) caused huge movements of people: USA, Australia, Fiji, … The indigenes were “driven up until the hills” from C17th onwards. The Japanese didn’t drive them any further so much as made their retreats economically untenable. (And then massacred them if they expressed any opposition.)
If we were looking for a secondary villain, that would be the KuoMinTang/military dictatorship. Who suppressed not only the indigenes but also the Fukien/Hokkien/Hakka cultures and languages, imposed Putonghua, and massacred (Han) political opponents who suggested the Republic should aim to be a liberal-democratic state.
@Hat never got around to investigating
I didn’t visit Taiwan at that time; but my impression is that you’d have had a hard time finding any indigenous peoples. Now there are plenty of cultural centres and festivals. But I note those centres were all built since the ~1990’s liberalisation. The Taroko culture(s) are very visible in that gorge — which is indeed spectacular — and their cultural organisations appear to be in command of the tourism.
I was struck by that sentence as only yesterday I was trying to think of a language that had final stress, and I didn’t manage.
French names spoken by Americans usually have final stress, whereas they usually have initial stress when spoken by British people, and the two are equally wrong. Take Monod, for example which is /mə’nəʊ̯/ and /’mɔnəʊ̯/ in the two cases, whereas the French pronounce it with virtually no stress. So French doesn’t count as a language with final stress.
My experience is almost entirely with European languages, but even there my impression may be wrong, so I wonder if any expert would like to comment on this classification:
Consistent initial stress: Hungarian, Finnish, Czech
Initial stress frequent: German, English
Usually penultimate stress: Italian, Spanish, Catalan, Portuguese, Polish, Welsh*
Mainly penultimate stress, but final stress not uncommon: Greek?** Turkish?
No strong stress: French
Strong and unpredictable stress: English, Russian
*Welsh may have consistent penultimate stress, but there is a frequent poster here who will certainly know.
**I once went from Venice to Haifa in a Greek ship. There were lots of announcements that began with something like παρακαλή παρακαλώ (checking with Google Translate suggests that it was probably παρακαλώ παρακαλώ).That may have given me the idea that final stress was common.
@ac-b
You have English in twice. Is the first one wrong or do you mean different varieties (e.g., I would say esp. Indian English) have different stress patterns?
@a c-b whereas the French pronounce it with virtually no stress. So French doesn’t count as a language with final stress.
I agree French doesn’t count as a language with final stress.
OTOH when I proffered an observation that French doesn’t have syllable stress within words, I was strongly told off (IIRC by DM) that all languages (with polysyllabic words) have within-word stress. This was in context of a discussion about stress-time vs syllable-time languages. (Within-word stress is not phonemic in French. That’s a separate consideration. As you can see on that wiki page, the phenomenon is complicated and messy.)
I of course can’t remotely remember when or what was the ostensible topic of the thread — probably something totally unrelated.
BTW do you have a less obscure example than ‘Monod’? Neither Larousse online nor wiktionary nor my Oxford French-English mini include it. Monod appears to be a personal name: it may not be of ‘proper’ French origin at all (Huguenot?). In any case, Proper Nouns are unrepresentative and hazardous when it comes to pronunciation patterns.
Welsh may have consistent penultimate stress, but there is a frequent poster here who will certainly know
So it does; though there are some exceptions, mostly the result of historical vowel fusion (as in Cymraeg itself.)
There are plenty of languages with consistent final stress. Without getting too exotic, Turkish immediately comes to mind (though there are plenty of exceptions, notably in place names.) Old Welsh must have been another; it is unclear at what stage, exactly, the stress shifted back a syllable.
Many languages have word-level stress as a phonological reality without it ever being even potentially contrastive: French is one. (Kusaal is another.) The matter is complicated by the fact that “stress” is not manifested phonetically in exactly the same way across languages, and involves a whole complex of features like relative loudness, duration and pitch which may not agree.
It seems tendentious to use “Truku” to refer to the dominant dialect of the “Taroko” language when it seems 99%+ likely that these are just alternative transliterations (mediated via various non-Roman scripts) of the same word.
They are certainly alternative transliterations, but why “tendentious”? I can think of all sorts of reasons why both versions would occur here, from failure of copyediting to the use of different spellings in different contexts. Compare US place names such as Pittsburg/Pittsburgh, Allegheny/Alleghany/Allegany, and Hoosic/Hoosac/Hoosick.
“Obviously “settler” is pejorative (I think you’d say “migrant” otherwise?)”
‘Settler state’ usually refers to a state based on settler colonialism — a subset of migration, but obviously not just interchangeable with that term. I’ve been a migrant in several European countries, but it would be pretty absurd to claim that I was ever part of a process of Americans ethnically displacing the mighty Belgian hordes (however negative my presence there might have been for the country). I feel like terms like this get mocked just because they’re fairly general, but broad terms have their place, and it’s really not at all hard to pop over to Wikipedia to find out what useful demarcations a term was intended to have before dismissing it just because it’s popular among the wrong sort of people (college-town dwellers in this case, apparently).
In this case, the aboriginal populations of Taiwan have been nearly completely ousted by a settler population that is largely sinitic in background, and whose ‘Chinese-ness’ remains pretty central (if obviously fraught) to its identity as a state. It matters in some ways that the Qing’s relationship to this sinicization was very different from Britain’s relationship to its settler-colonies, but it probably doesn’t make that big of a difference to the aboriginals.
One complexity is what’s being counted as the “state” in “settler state.” From 1683 to 1895 Taiwan was ruled by the same Manchus that ruled mainland China, as part of the same polity or “state.” There were as of 1683 already a quite significant number of Han Chinese on the island, mostly arrived in the previous six decades or thereabouts, but I don’t think they yet controlled the whole island or necessarily outnumbered the indigenes, although a few minutes of googling didn’t give me any good numbers. Very substantial additional Han migration or “settlement” during the Manchu centuries was necessary for the aboriginal ethnicities to be swamped/ousted.
Historical developments in which members of the dominant ethnic group of a polity migrate to and eventually numerically/culturally/etc. dominate the prior population of an adjoining peripheral region (all without crossing an international boundary) bear some similarities to classic “settler colonialism,” but also exhibit some differences. Would one refer to Inner Mongolia (where Han incomers have swamped the remaining Mongol population) as a “settler state”? If not, is that because it’s not a “state” but merely a province of China (which of course many people claim is all that Taiwan is)? Because it’s not an island?
Is the Kaliningrad Oblast a settler state? Indeed, while the population turnover there was comparatively rapid and total, it can be viewed as simply a recent episode in a sequence that’s been going on for a millennium or more of back-and-forth shifts in the de facto boundary between German-dominated parts of Europe and Slav-dominated parts, as shifts in migration patterns, forces of cultural assimilation, and/or the fortunes of war push the line in one direction or another.
Taiwan is a settler state *now*, is all that the (rather general) reference in the quotations says. I suppose the PRC would deny that Taiwan is a state, but I think we can be a bit more realistic here.
For the rest, the more usual term for the phenomenon is ‘settler colonialism’, which sidesteps the question of whether an independent state is at play, or if we’re dealing with regions within larger states, probably because that’s not really the important feature for most purposes. But I think it would have been a bit pedantic for this article to have said ‘settler-colonialist states, provinces, regions, and similar entities’.
thick description
I’ve encounted this expression before, and had read that Clifford Geertz promoted it. What I just learned by netsay is that he (by acknowledgment) borrowed it from Ryle, The Concept of Mind (1949), where “thick description” is contrasted with “thin description”.
The reason I tracked this down is that the German counterpart is dichte Beschreibung. dicht means “dense” or “close together” in many contexts, but also also “wiped out” [by alcohol/drugs] in others. I considered “dense” to sound more refined than “thick” in this context. But if Ryle used “thick/thin”, it’s a palpable instance of Ryle style and I have sent deprecation packing.
My copy of CM is within sight and arm’s length, so I could check it – but it’s in a glass-fronted bookcase in front of which piles of other books are stacked.
ETA: I bet the expression “wiped out” is as dated as you can get, but that’s the way they did it back then.
“The Earth is flat” is wrong; “the Earth is a sphere” is wrong; but “both of these are equally wrong” is wronger than both…
French has prepausal stress, and varying amounts of stress at the end of “intonational phrases” in general (very often larger than words by any definition of “word”); I find that very salient. What it does not have is vowel reduction in unstressed syllables, except for e; that’s expected for a language that isn’t stress-timed.
Nobel Prize in Physiology or Medicine.
Try Monet or Rousseau.
It’s not just proper names. Lots of more recent borrowings from French receive initial stress in BrE and final stress in AmE, as in ballet or café.
I hadn’t expected everyone here to be familiar with Jacques Monod, famous though he was, and one of the few Nobel Prizewinners whose work is remembered long after the initial fuss. Maybe not as famous as Crick and Watson, but getting there. Who can remember who Katalin Karikó is, and what she did to earn her a Nobel Prize in 2023?
Anyway, I didn’t choose Monod because he was famous, but because his name was the example that made me realize for the first time that the way Americans pronounce French names is very different from how British people do. When I went to Berkeley in 1967 I was very familiar with someone I thought of as Monno, and it took a moment to realize that the person the head of the laboratory called M’no was the same chap.
Incidentally, there are lots of Monods notable enough to have Wikipedia articles. Probably better known than Jacques today in France is Théodore Monod (at least, he gets mentioned much more often on television).
Well of course. What do you expect when you see the phrase “French names”? Place names, I suppose, but what else?
Are you suggesting that the Huguenots were not “proper” French people?
I’ve certainly heard of Jacques Monod, but then I would have, wouldn’t I?
I’ve always stressed him on the final syllable, and indeed do this consistently with disyllabic French names: Chirac, Macron, Jaurès, Foucault, Céline, Houellebecq… I think, speaking English, I stress the antepenult in trisyllables, though: Mitterand, Pompidou … on the other hand, I say Sarkozy with stress on the final, but that is probably because it’s such a weird name anyway. And I stress the last syllable in Chateaubriand, but that is probably because of its obvious compositional structure.
I don’t think the generalisation about UK stress is valid, though admittedly I may not be typical, and practically everyone I am likely to talk to about famous French people knows at least some French, so they may not be typical either.
Come to think of it, I say “Belmóndo”, but I hereby declare that that doesn’t count. Also “Garibáldi”, after all, and he was born in France, after all (I’ve just discovered that his birthplace, certainly in France now, actually was in France at the time of his birth, even, though it didn’t stay that way for long. Another great Frenchman!)
I say Sarkozy with stress on the final, but that is probably because it’s such a weird name anyway.
That’s ironic, because in this case, your usual antepenultimate = initial stress would be etymologically correct…
I presume that French speakers ignore the Hungarian stress rules.
I may take to pronouncing the umlaut in order to épater les bourgeois.
Not only do French speakers ignore the Hungarian stress rules, but they also mangle Sárközy [ˈʃaːrkøzi] into Sarkozy [saʁkɔzi].
I’d occasionally wondered about the origin of the name, and now I know: “Sárköz (Hungarian: ‘mud alley’ or ‘mud passage’).”
If Jacques Monod’s recent ancestors had been Jewish rather than Protestant would you still have said he was not of ‘proper’ French origin?
I don’t recall anyone complaining that Lionel Jospin (also from a Protestant family) was not of ‘proper’ French origin when he was Prime Minister at the turn of the century.
What about Alexandre Dumas père, was he of ‘proper’ French origin?
What about Alexander Pushkin, was he of ‘proper’ Russian origin?
Sárköz (Hungarian: ‘mud alley’ or ‘mud passage’).
We can now reasonably surmise why there are so many Hungarians in countries where Hungarian is not understood.
I find there are several small towns in Germany that 200+ years ago had a Schlammgasse that was later renamed.
When Sarkozy was President he visited Hungary, and they showed some excerpts, Hungarian television reports. Not surprisingly, they pronounced his name in typical Hungarian way, completely different from how it was pronounced in France.
Are you suggesting that the Huguenots were not “proper” French people?
FFS you people! This is a language site.
I’m talking about the words, not the people. And I’m saying the pronunciation of Proper Names is no guide as to the sound pattern of the language. When I asked for a less obscure example, I was looking for ordinary words, _not_ more bloody Proper Names.
a c-b has failed to land any claim about the stress pattern of French. Thank you to @DM for answering the actual question.
Why assume proper names don’t follow “the sound pattern of the language”? Proper names, even if not originally French, get Frenchified by French speakers using French phonology, just like loanwords in any language. In any case, Jongseong Park already mentioned a couple of common nouns (ballet, cafe), and Wikipedia has a much longer list of French loans with different stress in BrE vs. AmE, both common and proper. Some of the most surprising to my American ears are BARRage, MASSage, MIRage. We’ve previously discussed Br GARage vs. Am garAGE.
David E: “I don’t think the generalisation about UK stress is valid, though admittedly I may not be typical”
The generalization does not have to be universally 100.0000% true to be valid as a generalization, and you are in fact not typical, according to the Longman Pronunciation Dictionary (2000), Routledge Dictionary of Pronunciation for Current English (2017), and Oxford Dictionary of English (the synchronic one, 2010), all of which give only first-syllable stress in BrE for Rousseau, Monet, and Degas. Foucault is more variable, for some reason; the dictionaries disagree, and Longman gives both options.
You can check on Youglish: it is not 100% universal, but ROU-sseau and MO-net are clearly common in BrE. Both sound strange to my American ears, but the evidence is loud and clear. Other names may be more variable. (Of course, there’s more than one Rousseau, but there’s no indication of any pronunciation difference between them.)
Why assume proper names don’t follow “the sound pattern of the language”?
Take for example the composer Debussy. The family name originated as ‘de Bussy’, originating from Bussy-le-Grand, Côte-d’Or. As such, the first syllable would be pronounced with unstressed reduced vowel, as per @DM’s explanation.
But Claude-Achille was not a country hick, was an urbanite born in Paris, and took offence at ‘deBussy’ being pronounced that way. So I (in my poor French) try to pronounce with three equally-stressed syllables.
There are bazillions of examples (in all languages I’m aware of) of people moving country and taking their names and their pronunciations with them. ‘Huguenot’ “a combination of a Dutch and a German word.” says wikip. “Huguenots controlled sizeable areas in southern and western France.” well known for not pronouncing French ‘properly’ — as in Parisianly.
So I am not claiming ‘Monod’ is non-French; but neither could I see evidence confirming that one way or the other. Then I asked for other examples.
a c-b made a claim about French-in-general, as yet unsupported. Furthermore what’s really pissed me off is his underhand suggestion I’m being or would be anti-semitic or religiously discriminatory. I’m surprised our moderator hasn’t called that out. I’m offended.
get Frenchified by French speakers using French phonology
Maybe. Sometimes. Eventually. For ‘Monod’ no evidence presented. In support for a general claim about French stress-patterns at a language-aware site, not remotely good enough, not even reaching the level of ‘anecdotal’. And that claim was not at the time about loans from French into other languages.
I say bárrage, mássage, but miráge (and garáge.*)
I think that my pronunciation of French names is more Frenchoid than that of common-noun loans, which on reflection would not be very surprising. (After all, neither barrages nor massages have any particularly French assocations for me, unless I’m actively thinking about etymology.)
I would guess (but am too lazy to investigate) that more-familiar French persons are more likely to get initial stress in UK English in general than less-familiar ones. Bónaparte …
* Oops, no I don’t. I say gárage. Just goes to show. Never trust your Sprachgefühl, especially in your native language. I really do say miráge, though. Or do I? Miráge. Mírage. Mírage. Miráge. Alas, all that is solid melts into air …
Was message ever pronounced messáge by Modern Englophones?
It’s not my impression that in France proper, whether in the 16th century or subsequently, there was a distinct “Huguenot” set of surnames that differed noticeably from the regular “French” set. To the extent the Huguenot population was not evenly distributed regionally (and would thus have tracked any regional skew in surname distribution), that may have been a mix of that ecclesiastical position attracting more favor in some regions and others and internal migration whereby Huguenots from other parts of the country relocated to areas where there was less persecution and more toleration, thereby further increasing the local Huguenot percentage, but bringing their home-region surnames with them.
When Huguenots fled France for other locations, that may have led to some respelling of surnames to suit the orthographic conventions of their Dutch-or-English-or-German-etc.-speaking host populations and/or to follow pronunciation shifts associated with the new environment. One of the Huguenot families fairly far up my own family tree is the Lispenards,* who were modestly prominent in 18th-century New York City and environs (there is still a Lispenard St. in lower Manhattan). The claim is that the Lispenard spelling evolved from L’Espinard, which admittedly looks more cromulently French although I’ve never checked to see if there are actual Catholics in actual France bearing a surname with that spelling.
*One relevant Lispenard ancestress was born in the late 1730’s and baptized in what was officially a colonial Church of England parish, but one so Huguenot-heavy that in those days they still maintained the register of baptisms/marriages/funerals/etc. in French.
J. W. Brewer. You will be interested in the 21 Lispenards mentioned here: https://www.findagrave.com/memorial/search?firstname=&middlename=&lastname=Lispenard&birthyear=&birthyearfilter=&deathyear=&deathyearfilter=&location=&locationId=&memorialid=&mcid=&linkedToName=&datefilter=&orderby=r&plot=&page=1#sr-132244814
@AntC:
I’m not sure what is meant here. For French speakers, Debussy and de Bussy are exact homophones as [dəbysi], and I would have no idea how to find different ways to pronounce Debussy so that one would offend urbanite sensibilities but the other wouldn’t.
Many common French surnames start with the preposition de, in many cases simply indicating the family’s geographic origin. Of course, particules were features of most noble surnames, so in time it became common practice to fuse non-noble surnames in spelling as Dechavanne, Delannoy, Deveaux, etc. but to write out noble surnames as de Beaune, de La Rochefoucauld, de Montmorency, etc. There were lots of exceptions to this of course, but the association with the latter spelling and nobility was certainly there, so in the 19th century you sometimes saw people of non-noble origin imitating this. It seems that Claude Debussy also rendered his name sometimes as de Bussy, no doubt to look more impressive. But this is all a matter of spelling, not pronunciation.
French does not have distinctive stress at the word level. The final syllable (disregarding any final e caduc) is usually stressed at the end of an ‘intonational phrase’ as David Marjanović puts it, which is not the same as word stress. For instance, certain vowels are phonetically lengthened at this position. Grande /ɡʁɑ̃d/ and guerre /ɡɛʁ/ are pronounced with long vowels as [ˈɡʁɑ̃ːd] and [ˈɡɛːʁ] if they are at the end of an utterance. However, this does not apply when they are not the final syllable of an intonational phrase. So Grande Guerre comes out as [ɡʁɑ̃d ˈɡɛːʁ] with only the second word stressed (and therefore with the long vowel) at the end of a phrase, or just as [ɡʁɑ̃d ɡɛʁ] without any stress if it is not at the end of a phrase and the utterance continues.
It is important to stress (pun unintended) that the intonational phrase is not the same as the word. Unless you speak in single words, a single intonational phrase will usually consist of several words. But it can sometimes be shorter than an individual word as well. A typical way for a French speaker to emphasize a given word is to break up its syllables into individually stressed units, as in impossible ! [ˈɛ̃ ˈpɔ ˈsibl], similar to what is implied in the English eye-dialect ‘Oh. My. God.’
I say gárage. Just goes to show.
And ‘Farage’? I say that as three syllables, including the ‘peuh’ at the end.
I would have no idea how to find different ways to pronounce Debussy so that one would offend urbanite sensibilities but the other wouldn’t.
It’s fairly clearly that being an ‘artiste’, our man affected all sorts of sensibilities. (He at times wished to be known as Achille-Claude rather than v.v.) So I’m not making a claim there’s a single ‘proper’ way to pronounce the name. I am claiming only that when it comes to Proper Names, you can’t take them as representative of the language at large. Especially of something so subtle as stress patterns. How would you hazard ‘Saint-Saëns’?
the intonational phrase is not the same as the word.
well known for not pronouncing French ‘properly’
Regularly when holidaying in S/W France, I heard the ‘intonational phrase’ for prices pronounced ‘ving-te’, whether or not there was a ‘-et-un’ following, with (if anything) more stress on the ‘-te’.
And I’m pretty sure that if you were doing the individually stressed units business, impossible ! would have four syllables, again with stress on the fourth.
[I am perhaps over-sensitive when it comes to French. When I got to actual France, I discovered after 5 years of school French, that I’d been taught utter bollox. Nobody in France talks like Racine or Molière.]
I am claiming only that when it comes to Proper Names, you can’t take them as representative of the language at large. Especially of something so subtle as stress patterns.
Actually, you can. As has been said, there is no difference at all between the patterns of common nouns (or other words) and proper names in French. Your impression is your own invention.
The final syllable (disregarding any final e caduc) is usually stressed at the end of an ‘intonational phrase’
You could analyse that as full-word-final stress which is lost phrase-internally; there are parallels elsewhere. Mooré, for example, deletes underlying word-final short vowels phrase-internally, but keeps them before pause. Biblical Hebrew, too, essentially preserved the underlying stress patterns of full words before pause, but subjected them to stress dissimilation or outright loss elsewhere.
Kusaal, which basically has stress on all word roots, drops stress on word-final syllables which don’t precede pause unless they are superheavy CV:C; though it also adds word-final stress before several enclitics which have no actual segmental form of their own at all. Just to be awkward. (This is not completely unique to Kusaal: Tongan has this too, though only with one inaudible enclitic, compared to Kusaal’s three, which have distinct tonal effects from one another, showing that there really are three different ghostly enclitics.)
Such an analysis of French stress obviously makes sense historically, but I admit that it’s unnecessarily complicated synchronically.
As has been said, there is no difference at all
Sorry, @Hat: when were you last in France buying vegetables at a market stall in Salignac-Eyvigues? (‘Eyvigues’ stressed (if anything) on the third syllable.) For me it was over 40 years ago, and I remember it particularly vividly because the French was so totally not what I’d been taught in school. Perhaps pronunciation has changed since, but Proper Names are more likely to preserve a historical pronunciation.
I note Napoleon complaining how impossible it was to manage a country where so few people spoke the Language. (His own French had a thick Corsican accent.) I note De Gaulle making a similar complaint about the non-uniformity of cheeses.
I have a strong impression what I’m getting told here is coming from the Académie.
@AntC: The pronunciation of Saint-Saëns is a different matter because it’s an issue of whether a phonological segment /s/ is present at the end or not. Your claim about Debussy seems to be about phonological stress, which is absent in mainstream versions of French at least.
French speakers of southwestern France, a traditionally Occitan-speaking region, may well have applied some elements of phonological stress to French, though it’s not something I’ve noticed. I’m not saying it doesn’t exist; rather, as the variety I’m most familiar with is Parisian French, I’m probably conditioned to tune out any indication of phonological stress in French.
In French, the t in vingt is pronounced when it is followed by another number as in vingt-quatre /vɛ̃tkatʁə/. The lexicalized exceptions to this are quatre-vingt-dix /katʁəvɛ̃dis/ and quatre-vingt-onze /katʁəvɛ̃ɔ̃zə/. The t also is always pronounced in liaison if it is followed by a noun that the number modifies that starts with a vowel, as in vingt ans /vɛ̃tɑ̃/.
Parisian French regularly deletes final /ə/ in normal speech when it is possible (exceptions are cases like dis-le !). So utterance-finally, the above examples come out as [vɛ̃t.ˈkatʁ], [ka.tʁə.vɛ̃.ˈdis], and [ka.tʁə.vɛ̃.ˈɔ̃ːz] respectively. Exceptions are in singing – where final /ə/ can even be pronounced after a vowel, as in the line ‘Allons enfants de la Patrie’ of La Marseillaise with its [pa.tʁi.ə] – and in emphasis.
So you’re right, impossible ! can be four syllables, and I’ve also heard merde ! as two syllables from Parisian French speakers plenty of times. But I’m fairly sure that the final /ə/ is frequently if not usually dropped in this kind of emphasis as well – in my head, both versions sound correct.
The retention of final /ə/ in normal speech is however a feature stereotypically associated with the South of France. This can be pronounced even when there is no etymological /ə/ as in ciel, parc, or vingt. I don’t claim to be familiar with these types of accents, but my guess is that the speakers you heard have vingt stored with the unetymological final /ə/ and are pronouncing the t everywhere, with the possible exceptions of quatre-vingt-dix and quatre-vingt-onze.
Anyway, from my own intuition, the whole thing about the final syllable of an intonational phrase being stressed in French (which I’ve read in plenty of descriptions of French phonology) is more of a general observation and the reality is more nuanced. For example, I have a hard time hearing vingt ans with only the final syllable stressed; it sounds more natural as [ˈvɛ̃.ˈtɑ̃] or even [ˈvɛ̃.tɑ̃]. Same for ballet [ˈba.(ˈ)lɛ] or café [ˈka.(ˈ)fe]. It might be something to do with how ‘light’ the final syllable is in these cases.
French speakers of southwestern France, a traditionally Occitan-speaking region, may well have applied some elements of phonological stress to French, though …
Thank @Jongseong; I’ve spent far more time in West/South/South-East France than in Paris. (Indeed what I remember most strongly in Paris was the preponderance of Arab speakers.)
… it’s not something I’ve noticed
Hmm? I rather thought that those ‘extra’ final syllables are what Parisians particularly notice when they head off on vacances. Here’s a well-known comic song conforming to the caricature. Note ‘place’ has two syllables (unlike Parisian French), stressed on the second. (The refrain also contrasts ‘d’placé’ two syllables, Parisian pronunciation, stressed on the first.) (When I say ‘stressed’ I mean in the if-anything sense. Certainly not to the extent of English within-word stress.)
‘Allons enfants de la Patrie’ … with its [pa.tʁi.ə]
La Marseillaise written 1792. Perhaps all French of that vintage retained those extra vowels in pronunciation. And perhaps that usage continues in Occitan-aligned France — at least as of last Century?
I’ve also heard merde ! as two syllables
There’s an avant-garde play which opens with the author pacing the stage pronouncing ‘merdre !’ vehemently (two syllables, stressed on the second). It got past the censor because that was claimed to be a nonsense word.
… more of a general observation and the reality is more nuanced.
I’ve heard a great variety of pronunciations throughout France. (Possibly wider variations than regional accents in Britain — France is much bigger, after all.) A very elderly woman (as at 1980’s) in a much more elderly farmhouse near Salignac, whose French was beyond comprehension even for somebody with a Masters in French Language (that wasn’t me, I hasten to add). And who insisted on plying us with illegal hooch, despite us trying to head off for the Channel Ferry. Illegal? “Ce n’est pas contre la loi, c’est à côté de la loi.”
Oh, re The pronunciation of Saint-Saëns, as well as the final /s/ or not, I’d draw attention to what the diaeresis is telling us about historical pronunciation. As in: Proper Names you can’t take them as representative of the language at large.
@AntC: I did say that the “retention of final /ə/ in normal speech is however a feature stereotypically associated with the South of France.” It’s what Parisians associate with speakers from that region, along with the [ŋ] in the nasalized vowels. What I meant is that I didn’t notice phonological stress from regional French speech. Are there any speakers who contrast minimal pairs that are otherwise pronounced identically by the placement of stress?
Singing pronunciation is well known to be conservative, and the retention of final /ə/ in the traditional Occitan areas or in Francophone Africa suggests that final /ə/ was pronounced until no more than a couple of centuries ago in the history of French. I wouldn’t be surprised if patrie could be pronounced with a final /ə/ in normal speech when La Marseillaise was written in the late 18th century.
The trema in Saint-Saëns is interesting, and it is not hard to imagine that it could have been pronounced historically with two different vowel sounds given that Saëns comes from Sidonius. But the French Wikipedia article on the trema suggests that in this case, it is not a diaeresis but a relic of another historical usage of the trema which marked a vowel when it was not pronounced. They also give the example of de Staël /stal/, which I’ve also wondered about before.
There are also names of foreign origin that get the trema in French, as in Citroën [sitʁɔɛn] from Dutch Citroen [siˈtrun], where it is treated as a diaeresis in pronunciation despite the etymology. You can even see the spelling Laëtitia for the name more usually spelled Laetitia and pronounced [letisja]. I’m guessing it is possible to use the spelling pronunciation [laetisja], but it doesn’t sound very natural to me – I would just say [letisja] even if it’s spelled with a trema.
I don’t know about message, but I learned here not too long ago that balcony used to be stressed on the second syllable in BrE. (Da balkon is stressed on the last syllable to this day, but that may be from German. I didn’t check).
Are there any speakers who contrast minimal pairs that are otherwise pronounced identically by the placement of stress?
Not that I observed. (IOW I don’t think this is a phonemic feature.)
Going by the ‘music’ of the phrasing, it seemed like a prosodic feature: you have to tack on a vowel if the phrase would otherwise end in a consonant, to carry the falling tone “stressed at the end of an ‘intonational phrase’”, as we’re putting it.
Synchronicity!? Over at LLog, myl has made two posts in two days about the acoustics of focus-like/stress-like pronunciations in French.
It’s complicated. And fersure isn’t the fairy-story I was told in school.
[saʀ̥͡xkoˈzi], or just Sarko [saʀ̥͡xˈko]. Definitely obligatorily with [o]. /r/ isn’t specified for voicing anymore, it often gets devoiced even between vowels these days, but it’s not a straightforward [χ].
All seconded. (Except that the reduced vowel is definitely rounded and more or less front, not central; it’s closer to /ø/ than to any other French vowel, and native speakers often equate the two.)
Allegedly, Danton sometimes spelled himself D’Anton before he had enough of this and participated in the Revolution.
…yes, if you’re speaking very slowly; and even then mostly just when /r/ follows.
Also, will somebody please update the dictionary transcriptions of the nasal vowels? I’ve never heard an actual [ɑ̃]. In Québec it’s a central [ä̃]; in Europe it’s conservatively also central, but rounded, so let’s try [ɒ̈̃], while Kids Today often move it past [ɔ̃] and threaten to merge it with the on sound, which is [õ]. About the latter, even m-l once said that the dictionary transcription as [ɔ̃] is obsolete, and she preserves a whole list of remarkable archaisms like a full-blown phonemic vowel length contrast. I’ve only heard [õ] on both sides of the Atlantic, except from Germans; beau and bon are a minimal pair for nasality. The in sound has drifted in opposite directions just like an has: in Québec it’s [ẽ], in Europe it’s mostly [æ̃], but Kids Today sometimes shift it to [ɐ̃] (in free variation as far as I’ve noticed). That leaves un, which remains [œ̃] for the rather few people who haven’t merged it into in.
Listen again, then: the stress is consistently on the third. The fourth, which is very long but not stressed, is precisely such an e caduc that is allowed to resurface in singing whenever it’s needed to fill the meter.
To have a third syllable, it would need to be spelled Eyviguès. It’s not. The fr.wikipedia article doesn’t mention the pronunciation at all; the en.wikipedia one says: “French pronunciation: [saliɲak ɛviɡ]” – as expected.
(It adds: “Occitan: Salanhac e Aivigas“.)
Pronouncing -re as a syllable is largely limited to higher registers even in this environment, though. You’re more likely to hear [k̟atvæ̃(n)dis], [k̟atvæ̃õz] with the whole thing dropped. (And a vowel at the end of onze is limited to the circumstances where any other e caduc surfaces – except in strong southern accents, where it’s more likely to come out as [ˈɔŋzɐ] or suchlike.)
Oh no – that’s [pa.tʀ̥͡xiː.i.ɵ] with the i doubled!
It’s particularly popular as a crescendo: Eh, merde !!! as if spelled mère DEUX !!!.
That’s interesting, because they all feel unequivocally end-stressed to me (if phrase-final; if not, as in c’était vingt ans plus tard, they’re unstressed entirely).
I listened to the first minute. It’s entirely in unremarkable Parisian*, except for the resurfaced e of place (because singing), which is, unremarkably, unstressed. I hear no difference between this and, say, blackbird in English (as opposed to the end-stressed black bird), so I’m confused about what you hear as stress.
* I mean, these days you can’t even distinguish Paris from Brussels or Lausanne from this sample; but there’s no southern accent in it whatsoever.
You can’t take the spelling of proper names as representative of the language at large. Their pronunciation fits into the sound system of the language at large just as much as Featherstonehaugh, Cholmondeley, Menzies and their ilk do in English.
Yup, colleague of mine.
Isn’t Staël from Dutch, though, where ae was long used for /aː/?
No, it’s beyond that – it’s from an earlier stage of the language that had a lot of penultimate stress (i.e. the last vowel of the word-or-so was stressed unless it was e).
To have a third syllable, it would need to be spelled Eyviguès.
French wikipedia was presumably written by Parisians/northerners. Unless you’ve been to the place and heard the locals pronounce it, I don’t think you can make any claim.
Happisburgh to you.
which is all about actual spoken French and how profoundly different its syntax is from written or spoken-formal French.
Maybe you heard the actual Occitan; I’m sure Aivigas has three syllables (…and is stressed on the second).
spoken by Americans usually have final stress, whereas they usually have initial stress when spoken by British people,
Can anybody explain what’s going on with the cadence/stress patterns of this dude.
The delivery seems altogether too smooth (AI generated? [**] heavily sound-engineered), but chiefly I notice the lack of cadence. Sentences go on for many words with little phrasing or pauses and such that I want to say please take a breath but then they suddenly end with a down-sweep.
At 1:30 the “corrrect” and “larrrge” seem weird. And “starrr’ling conclusion” a little later.
The volcano eruption topics started in Hawaii; I’m guessing that’s where the accent is from. Polynesian languages have a very different stress pattern compared to English. Would that be an influence?
[**] Seems unlikely: copes (too!) smoothly with complex technical terms and foreign names for islands/mountains.
@AntC. “Can anybody explain what’s going on with the cadence/stress patterns of this dude.”
The “speaker” is not a human being.
Yeah, that’s AI, and it’s just guessing at the names. Doesn’t sound even a little human.
Thank you. I don’t know whether to be more impressed at how much AI is getting right these days (or if not right, at least close enough to be understandable); or more annoyed the author can’t use their own voice (since a lot of it is opinionated or speculative), and/or break their stream-of-consciousness into more digestible chunklets.