THAT STRANGE MUFFLED UTTERANCE.

My wife and I are on the very last of Patrick O’Brian’s Aubrey-Maturin series (which we have been reading since July 2011), Blue at the Mizzen, and I was pleased with this linguistic description (on pp. 42-43 of my paperback):

Stephen bowed: but when they had put on formal clothes he said, ‘Interpret, is it? As I told you before I do not speak – not as who should say speak – Portuguese. Still less do I understand the language when it is spoke. No man born of woman has ever understood spoken Portuguese, without he is a native or brought up to comprehend that strange blurred muffled indistinct utterance from a very early, almost toothless, age. Anyone with a handful of Latin – even Spanish or Catalan – can read it without much difficulty but to comprehend even the drift of the colloquial, the rapidly muttered version. . .’

This has been my experience with Portuguese as well, and it was nice to see it set out so forcefully. (No knock on Portuguese, of course; it sounds very pleasant indeed. It is merely unintelligible and in fact unidentifiable. It is one of the few languages—Armenian is another—I’ve heard spoken on the NYC subway and been unable to name.)

Comments

  1. Richard Feynman’s description of learning to speak Portuguese (or as he called it, “Feynman’s Portuguese”) is quite entertaining — it’s in the chapter “O Americano, Outra Vez!” in “Surely You’re Joking, Mr. Feynman!” (see Google Books or this online excerpt).

  2. My experience of Portuguese as well. “Those are lovely sounds, my dear, but surely you don’t intend them for phonemes?”

  3. I am constantly mixing up Portuguese with the sound of Slavic languages, and have been for more than a decade. Since I speak several Slavic languages, it drives me nuts that I can’t place the language until I finally realize it’s Portuguese. :-)
    Incidentally, I just finished the Aubrey-Maturin books a few weeks ago myself. _21_, the final unfinished book, is worth checking out despite not having an ending!

  4. marie-lucie says:

    I suppose you are talking about European Portuguese, not Brazilian. Here is an anecdote I think I mentioned here several years ago:
    In the 60′s a new type of songs became very popular, and they were played almost daily on CBC radio. The sounds were those of French, but I could not understand the words at all! It was as if someone had recorded French words, cut the tape into syllable-size pieces, and put the whole thing back together at random against the original music. After a few weeks of careful listening, I heard one word I recognized: coraçao (with tilde on the a next to the end)! It was Brazilian music. I did not understand much else, but at least I could identify the language. I think I might have had more trouble with European Portuguese, even though I had taken a few classes in it some years before.

  5. Bill Walderman says:

    “Anyone with a handful of Latin – even Spanish or Catalan – can read it without much difficulty but to comprehend even the drift of the colloquial, the rapidly muttered version. . .’”
    Anyone with a handful of German can read it without much difficulty but to comprehend even the drift of the colloquial, the rapidly muttered version . . .
    Portuguese : Iberia :: Danish : Scandinavia

  6. _21_, the final unfinished book, is worth checking out despite not having an ending!
    Oh, believe you me, we’ve got a copy waiting to briefly still the anguish. Thanks for the reassurance—I had no idea what to expect from it!

  7. David Marjanović says:

    This has been my experience with Portuguese as well

    It’s also the experience Brazilians have with European Portuguese. At least that’s what one Brazilian told me.

    Anyone with a handful of German

    That’s not quite enough, but German plus English works fairly well.
    BTW, what happened to the Logeion thread? Opening it shows a long line of question marks and nothing else at all.

  8. marie-lucie says:

    I would add: Anyone [not a native speaker] with a handful of English and also knowing French … would read English much more easily than they could decode casual spoken conversations.

  9. I too had the experience, many years ago, of first hearing Portugese and thinking it had to be Slavic. I’ve wondered whether the development of its pronunciation out of Hispanian Latin might have been influenced by eastern Europeans who entered the peninsula during the Völkerwanderungen. The Alans – Iranian speakers – are known to have settled Portugal; is it possible that their number actually included Slavs swept up in the migrations?

  10. Mark Liberman posted not long ago on vowel reduction and elimination in European Portuguese. Some good audio examples:
    http://languagelog.ldc.upenn.edu/nll/?p=4366
    This is the biggest difference between European and Brazilian Portuguese. The Brazilians like to say they sing when they speak; they certainly do pronounce the vowels a lot more fully. I speak decent Brazilian. Portuguese people understand me (and very occasionally flatter me by thinking I’m Brazilian), but I have a much harder time understanding them.

  11. J.W. Brewer says:

    Have I mentioned before the experience of flipping channels in a Swiss hotel room and coming to a channel where the audio was sufficiently exotic and incomprehensible (sort of like an angry German trying to speak Italian) that I excitedly theorized it must be the Romansh channel? Way cool! But alas it was merely Portuguese. (Because of gastarbeiter there are likely substantially more Lusophones living in Switzerland than, um, Romanshophones.)

  12. I too have found European Portuguese to sound more “Slavic” than “Romance”; Brazilian Portuguese, with its heavy nasalization and open syllables, sometimes sounds dimly “Chinese” (I remember hearing some Brazilian saying “Eu acho que nao” and mistaking it for an utterance in some East Asian language). But I agree with Marie-Lucie’s point that it can sound oddly French-like too, not least because of its (strong) uvular r.
    DCBob: I guarantee you that the Slavic-like phonology of (European) Portuguese today is not due to any kind of outside influence: the major phonological transformations took place over the past five centuries, when Portuguese speakers were overwhelmingly monolingual and when within Portugal no significant groups of non-native speakers were shifting to Portuguese.

  13. Can I add my vote to those who think Portuguese sounds Slavic? Not being a linguist, I couldn’t tell you why I think that, but it does …

  14. BTW, what happened to the Logeion thread? Opening it shows a long line of question marks and nothing else at all.
    That’s distressing, but I’m not getting that effect. What browser are you using? And is anybody else having that problem? (I’ll probably be closing that thread within a day if nobody else comments in it.)

  15. marie-lucie says:

    Etienne: it can sound oddly French-like too, not least because of its (strong) uvular r.
    Not only that, but also the nasal vowels. When I heard coração (a word I knew) it sounded like a hypothetical French word “courasson” (the nasal vowel was not quite like the French one, but close enough to fool me had I not known the Portuguese word).
    I have a (not very recent) book on Brazil, in which the French author uses “Saint-Paul” as the name of the city of São Paulo, not just because that is the French equivalent of the saint’s name (since such names are not necessarily translated) but because the Portuguese name sounds almost like the French one (and so that the French readers will not think of it as “sa-o-po-lo”).

  16. Bill Walderman says:

    “German plus English works fairly well.” Re Danish.
    English is assumed, of course. And you have to know how to undo the second sound shift, since Danish was apparently heavily influenced by Plattdeutsch, not standard Hochdeutsch, during the heyday of the Hanseatic League. Many Danish words seem to be calqued on Plattdeutsch forms of standard German words.

  17. I too had the experience, many years ago, of first hearing Portugese and thinking it had to be Slavic.
    Me too, when I went to Brazil – well, being in Brazil I knew it was Portuguese, of course, but nevertheless it sounded like Russian. I’m glad to see it wasn’t just me.

  18. Trond Engen says:

    I had the question marks in the Logeion thread earlier today. It also happened in another thread a couple of days ago. I’m not sure which one — the Roman slavery one maybe, since I remember hoping for a reply. I have been reading in Mozilla Firefox on my laptop, Safari on my iPad, and whatever it’s called on my non-i Phone, and I think it happened in all of them..

  19. Trond Engen says:

    The Watkins thread, I meant. Except maybe it wasn’t. It could have been anyone.

  20. But are you still getting it? And is this something I, the Ignorant Blogger, have to worry about, or is it a problem on your end?

  21. When I saw this characterization of Portuguese, I just knew that someone here was going to drag Danish into it :) And it only took two hours and one minute!
    I’ve always characterized Danish a very blurry language, and you can ask any Scandinavian (including Danes) about the warm potato, so far be it from me to say that the comparison isn’t suitable, but I feel impelled to insist that it is far from impossible to learn to understand spoken Danish. I arrived in Denmark knowing no more than half a dozen words in Danish and having no background whatsoever in German and left ten and a half months later able to speak, read, and write the language, as well as able to carry on conversations with a wide variety of people on many different topics. It probably helped that I was living with Danish families and attending Gymnasium full time, so I was very much immersed. It also helped immensely that when I changed host families in late November, my hosts and I made the decision together that we would not use any English unless it was absolutely necessary. I always had difficulty following the words in Danish music, but I think that is a separate issue.
    AJP, Trond Engen, Sili, and anyone else who speaks a Scandinavian language are all more than welcome to disagree with me about the intelligibility (or lack thereof) of spoken Danish and I will respect whatever feelings they have about it because their feelings in the matter are as correct as mine are. It is a very blurry language and has undergone quite a lot of sound changes. (In spite of this, I still feel like Danish is spelled about as phonetically as English is, albeit in radically different ways.) I’d be interested to hear how other Scandinavian speakers perceive Danish pronunciation.

  22. I had no idea that Portuguese tended to sound Slavic. I had always imagined that it must sound a bit like French due to the abundance of nasalized vowels. Now I am going to have to go find some audio samples and see what I think it sounds like.

  23. Bill, what two sound shifts are you talking about having to undo in order to understand Danish? And are you talking about written or spoken Danish?

  24. what two sound shifts are you talking about having to undo in order to understand Danish?
    Just one–the second or High German sound shift. And I’m talking about written Danish–no one understands spoken Danish.

  25. I remember watching BBC News and hearing someone speak English with a heavy Russian accent. Then it turned out it was José Mourinho, the then manager of Chelsea.

  26. “No one understands spoken Danish.”
    I do ;) Quite well twenty years ago, in fact, but I haven’t really had the chance to use it much in the last two decades. I need to spend more time on the Internet listening to Danish broadcasts to get my skills back.
    I don’t know German to compare with not in the sense of speaking or reading German, so I’m afraid you’ll have to describe the sound change to me. Sorry about that. (I do have a bachelors in linguistics, so you don’t have to be afraid to get technical. If you get beyond my level of technical I’ll just ask questions so that I can understand.) Thanks.
    Actually, when I run into German, I generally use my knowledge of Danish to help me decipher the German. Same thing when I run into Dutch.

  27. I agree with Marie-lucie on portugese, to me it immediately sounds like french (until I hear enough to realize it isn’t). I suspect it’s the nasal vowels too. I’ve probably heard the brazilian variety more than iberian type – is anyone keeping tally of the votes by continent?
    I got along well enough with written danish (or at least well enough for a tourist’s purposes) from English and a little german. Knowing what the second sound shift was and how to undo it might have helped, I don’t know. In its place, creativity worked fine.

  28. Here are the details of the second consonant shift. The core of the shift is the transformation of voiceless stops into (a) affricates initially, when geminated, and when following a liquid or nasal; (b) geminated fricatives intervocalically; and (c) single fricatives finally. Note that /tr/ remained unshifted. In addition, voiced stops lost voicing.
    Not all these changes made it into the standard: in particular /k/ > /(k)x/ did not, and only /d/ > /t/ consistently appears among the devoicing changes. What is more, Standard German has many borrowings from Low German and more recently English, which are unshifted.

  29. SFReader says:

    Danish generally sounds like some English dialects in northern England.
    Only more blurry, so you couldn’t understand a word of what they are saying.

  30. Hamburg-accented German can be imitated perfectly* by an English person simply by assuming a Yorkshire accent. I find Danish (and Plattdeutsch) sound more like Norfolk English.
    *At least to my satisfaction.
    Isidora, the problem for us in understanding spoken Danish is that German and Norwegian & Swedish pronounce pretty much all the written letters, whereas Danish (& English and French, of course)… Maybe the solution would be for other Scandinavian countries to offer a 1-week course at school to learn the Danish rules.

  31. Sir JCass says:

    Portuguese sounds like a French cat trying to speak Spanish to me. All those nasals and “ão” sounds.

  32. Logeion works for me, using Firefox

  33. Trond Engen says:

    Both threads work for me now. That’s all I can say. I’ve no idea what causes the problem or what you or anyone does to fix it. If it happens again I might try peaking into the sourcecode. As if that would tell me anything.

  34. Trond Engen says:

    Isidora: I used all I had to say about Danish in the recent thread. Now I’m just enjoying the Portuguese-bashing.

  35. Trond Engen says:

    But I will say that we could need an equally weird Slavic language for the sake of balance.

  36. Trond Engen says:

    Maybe Albanian can do. There are some who see it as the phonologically reduced sister of Baltic and Slavic.

  37. SFReader says:

    @Trond Engen
    —we could need an equally weird Slavic language
    Polish wins easily.
    Just try to pronounce Szczeczin…. ;-)

  38. SFReader says:

    Typical Polish phrases:
    BYCZKI
    W trzęsawisku trzeszczą trzciny,
    trzmiel trze w Trzciance trzy trzmieliny,
    a trzy byczki spod Trzebyczki
    z trzaskiem trzepią trzy trzewiczki.
    BZYK
    Bzyczy bzyk znad Bzury zbzikowane bzdury,
    bzyczy bzdury,bzdurstwa bzdurzy
    i nad Bzura w bzach bajdurzy,
    bzyczy bzdury, bzdurnie bzyka,
    bo zbzikowal i ma bzika.
    CHRZąSZCZ
    Trzynastego w Szczebrzeszynie
    chrząszcz się zaczął tarzać w trzcinie.
    Wszczeli wrzask Szczebrzeszynianie
    - Cóż ma znaczyć to tarzanie?
    Wezwac trzeba by lekarza,
    zamiast brzmieć,ten chrząszcz sie tarza!
    Wszak Szczebrzeszyn z tego słynie,
    że w nim zawsze chrząszcz brzmi w trzcinie!
    A chrząszcz odrzekł nie zmieszany:
    Przyszedł wreszcie czas na zmiany!
    Dawniej chrząszcze w trzcinie brzmialy,
    teraz będą sie tarzały.

  39. Trond Engen says:

    Polish is weird in the opposite way. Its orthography makes it look unpronounceable, true, but phonologically it’s not that far off from its neighbors..

  40. SFReader says:

    No, no!
    Russians can’t pronounce any of that even if it was spelled in Cyrillic.

  41. Marc Leavitt says:

    At the risk of mere repetition:
    Despite speakng French and Italian, wih a working knowledge of Spanish, my Portugese cleaning ladies are almost totally incoprensible when they speak to one another. And yes, I can read and understand a lot of Portuguese.

  42. Yes, I vividly remember an encounter outside the Hall of Graduate Studies at Yale (where I both lived and took classes—it was a real luxury being able to pad down the hall from my room to a classroom in my slippers, carrying my morning coffee) where a Polish professor was gathering people interested in taking a class and I joined the little group and tried to imitate a word (don’t remember which one, but it had a y in it). He looked at me with infinite disgust and told me I sounded like a Russian. I slunk away abashed.

  43. AJP
    Um…I think you’d need more than just a week…but that would be a workable solution to the problem for Norwegian students, since you wouldn’t be expecting them to speak Danish, only understand it. (Some ofthose Daninsh sounds are murder to learn to make until you get used to it. Rød grød med fløde, anyone? No, I guess not at berries are out ofseason at the moment. Too bad.) I digress.
    As was pointed out, Danish orthography is not phonetic like Norwegian, but the places where it is not phonetic are, by and large, quite rule-governed, or I wouldn’t have been able to learn it by intuiting the rules. I did have people teaching me things, but most of my vocabulary and phonology were intuited due to the environment in which I learned it.
    So it stands to reason that if someone came up with a brief course of study where they exposed students to Danish allophones and their environments and to the all of the main things about Danish that make it the spelling non-phonetic and taught this in a systematic way being sure to include an awful lot of audio material, it should be possible to unblur Danish to a great degree for Norwegian and Sweedish speakers, and perhaps for others, in a short period of time.
    I think in learning Danish that I had two advantages that the Swedes and Norwegians do not have. One of them was that I was not starting from a a native language that I would have had every reasonable expectation ought to be mutually intelligible with Danish, and the second was that, being a native English speaker, I came to Danish with no preconception at all that a language ought to be spelled phonetically, just that it ought to have spelling and pronunciation rules that could be figured out with observation. (Yes they did start out with phonics in my American elementary school, but they did us a disservice by failing to teach us the more advanced phonics rules in English, so I had to figure those out on my own, from observation. Stupid educational theories. Some people at certain times and places didn’t even get proper phonics instruction.)

  44. I thought they must have rules and that they weren’t simply jamming, so thanks for confirming it.
    I only need to understand it so that I can watch their far superior TV. I was thinking of taking a course, but then I heard (or rather, didn’t hear) the subtleties of the vowel pronunciations and decided it would be too difficult if you weren’t living there, surrounded by native speakers.

  45. J.W. Brewer says:

    I took a reasonable number of my undergrad linguistics classes in HGS although I lived a few blocks away. Despite not living under the same roof I did, however, manage to attend at least one 9 am Semantics class there in my bedroom slippers (I had not realized I was not wearing more outdoors-appropriate footwear until I was already 2/3 of the way there and couldn’t be bothered to double back). The Ling Dep’t has subsequently moved into different digs; don’t know whether it was a promotion or an exile.

  46. Hat: Normally I am quite good at perceiving and adequately reproducing minimal pairs in any foreign language, even with minimal pairs differing solely in terms of vowel length or tone, but Polish is the one language that has defeated me (although I have never tried Danish): for the life of me, despite having the pair repeated in a clear voice a dozen times over by a very patient native speaker, I couldn’t even come close to reproducing the difference between TRZY and CZY.
    Re Danish: A Swedish fellow linguist has often told me that Danish has two phonemes, schwa and glottal stop, and indeed acoustically that is the impression it gives.
    I understand Trond Engen’s point about an “equally weird” Slavic language: Polish phonology has its quirks, but no Slavic language (or dialect, as far as I know) has undergone the types of changes which swept Danish, French or Old Irish: that is to say, a loss of intervocalic stops and large-scale loss of unstressed syllables which completely transformed the appearance of words and the phonology of the language. A Slavic scholar of my acquaintance once told me that teaching Introductory Indo-European linguistics to francophones was so much more frustrating than teaching it to anglophones or to Slavic speakers: the inherited Indo-European element in French has undergone so many sound-shifts that there are very few cognates left where the French word is still visibly related to an Indo-European root.
    And I suspect it is no accident that the languages (Old Irish, Danish, French) which underwent these disruptive changes are located in North-western Europe: my guess, for what it is worth, is that a Germanic-Insular Celtic areal feature, subsequently spreading to Proto-Old French, consisted of an exceptionally strong realization of a stressed syllable, thereby opening the door to large-scale loss of unstressed syllables. No living Slavic language is part of this language area (? I wonder whether some extinct Slavic languages such as Polabian might have belonged to it), explaining the non-existence of a “Slavic Danish”.

  47. I would say Polish sounds like Portuguese, but it would be hard to confuse Portuguese with Russian, or the South Slavic languages formerly known as Serbo-Croatian. I also think anyone with a solid grasp of Spanish or Italian can pick up the basics of spoken Brazilian Portuguese pretty quickly if they have a mind to. I took a year of Portuguese in college and had no trouble with it, nor did any of the other Spanish speakers in my class.

  48. Etienne, I think that claim is slightly overstated: Danish has a variety of initial consonants, it’s everything after that that is mush. Lars [Henrik] Mathiesen once challenged people who didn’t know Danish to figure out how his name is pronounced: I promptly replied “[ˈləː ˈməʔəʔə]“. He answered, “No, but it probably sounds that way to foreigners.”

  49. Jeffry House says:

    The suggestin that Norwegian students have a week to learn Danish pronunciation might benicely co-ordinated with the scholarly institution known as “russefeiring”. In my day, it was to Copenhagen that all russ travelled.

  50. Trond Engen says:

    I think what makes Danish so special is that it’s in the middle of the reorganizing process, on full display, like going through a pupal stage without the pupa. I don’t know enough about Portuguese to say if it’s the same thing.

  51. I also think anyone with a solid grasp of Spanish or Italian can pick up the basics of spoken Brazilian Portuguese pretty quickly if they have a mind to.
    Yes, Brazilian Portuguese is not a problem. Maturin was speaking of Portuguese Portuguese, and so perforce was I.

  52. My daughter’s going through the pupal stage of russefeiring at the moment. It’s the most absurd convention in which they willingly pay several thousand kr. for a hideous pair of red dungarees and a peaked cap, wear them for a week and then never again, stay up all night, get drunk and throw up etc. and then, when it’s all over, they take their final exams! Ridiculous. Nothing about going to Copenhagen has been mentioned, perhaps they’re keeping it as a surprise for me.

  53. I’ll be the tenth to say that spoken (European) Portuguese has sounded like Russian or some slavic language to me. Glad to hear it’s not just me!

  54. marie-lucie says:

    Etienne: And I suspect it is no accident that the languages (Old Irish, Danish, French) which underwent these disruptive changes are located in North-western Europe: my guess, for what it is worth, is that a Germanic-Insular Celtic areal feature, subsequently spreading to Proto-Old French, consisted of an exceptionally strong realization of a stressed syllable, thereby opening the door to large-scale loss of unstressed syllables.
    Of course I had read about the strong Germanic stress being responsible for the weakening and eventual loss of most post-stress consonants in French, but I did not realize that this process also applied to Irish (do you mean there was no such loss in the other Celtic languages?). I understand you to mean that the process was not occurring independently in French and Irish but was due to the same Germanic influence. Can you elaborate about the historical conditions of this Germanic influence in Irish?
    For French, I think that when the Franks, Wisigoths, etc took over a territory (Northern Gaul) which had been Celtic but by that time had largely switched to Latin, these Germanic speakers learned the local Latin spoken by the natives, but with their own accent with its very strong stress, and since they were the apex of society, the Latin speakers imitated that accent. On the other hand, if the strong stress had already existed in Gaulish, Germanic influence might have reinforced an already existing, heavily-stressed pronunciation. But since the Franks remained largely in the Northern part of their new territory, where French developed, not in the Southern part where Occitan did not undergo the same drastic changes and remained close to Latin (as did Spanish, Catalan, Italian), it is likely that strong stress was a Germanic, not a Celtic feature. Do I understand right?

  55. marie-lucie says:

    The sounds of Portuguese : A Cuban friend of mine one day said this about the pronunciation of (most likely Brazilian) Portuguese: Tiene la sensualidad del francés y la claridad del español.

  56. “no Slavic language (or dialect, as far as I know) has undergone the types of changes which swept Danish, French or Old Irish: that is to say, a loss of intervocalic stops and large-scale loss of unstressed syllables which completely transformed the appearance of words and the phonology of the language.”
    The loss of the short vowels u and i certainly produced a lot of weird consonant clusters.
    Russian mgnovenie for example.

  57. marie-lucie says:

    Bill W, loss of unstressed vowels, leaving the consonants intact, does result in weird clusters but the consonants are still recognizable. Losing consonants, especially in the middle of words, makes those words much harder to recognize in comparison with related languages which have not lost those consonants.

  58. @a lot of weird consonant clusters.
    West and some South Slavic languages have them in abundance.
    Thus, Russian morkov’ (carrot) becomes mrkva in Croatian and mrkvicka in Czech….

  59. m-l: I don’t think there’s any doubt that the strong initial stress in the Germanic languages is independent of the strong initial stress in Old Irish (note that this is the common ancestor not only of Modern Irish but also of Scottish Gaelic and of Manx). If there was a similar development in Gaulish, I know nothing about it; I had understood that Gaulish was fairly close to Vulgar Latin phonologically, and that’s part of the reason why the language shift in Gaulish-speaking lands was fairly easy. Certainly Welsh and Breton have no such effects.
    But there’s also evidence that pre-Classical Latin bore an initial stress that was strong enough to trigger vowel weakening, as in the pairs amicus/inimicus, capio/recipio, and in the word brūma ‘winter solstice, winter’ < brev(iss)ima ‘shortest’. By Classical times, however, the stress had shifted to where it remains today.

  60. “Russian morkov’ (carrot) becomes mrkva in Croatian and mrkvicka in Czech”
    These aren’t the result of phonemes dropping out, in fact, they aren’t really consonant clusters–they’re syllabic resonants, in a direct line of descent from Proto-Indo-European. (Maybe the carrot-word isn’t, but the phonemes are.) Russian realized the syllabic resonants with vowels in a variety of ways.
    Don’t we have syllabic resonants in English, too, in some environments, but due to unstressed vowels dropping out: butter, little?

  61. Marie-Lucie, John Cowan: actually, there IS evidence that the initial stress of Old Irish, (pre-Classical) Latin and Germanic are (areally!)related. According to this view, initial stress was originally an areal feature of the Westernmost Indo-European languages (Celtic, Germanic, pre-Classical Latin): incidentally, there is some evidence that Early Basque also had word-initial stress.
    However, this original initial stress was certainly not especially dynamic: it caused vowel weakening in Latin (cf. the examples given by John Cowan above) but does not appear to have caused any significant loss of unstressed syllables.
    What broke the continuity of this original “initial-stress” area was the expansion of Latin, with its new stress system. It would appear that contact with this system is what caused both Basque and Brythonic Celtic to lose their original word-initial stress systems.
    So you see, John Cowan, that the Old Irish-Germanic shared word-initial stress is, in fact, a case of an originally continuous (areal) feature being preserved in peripheral zones, i.e. in languages spoken just outside the Roman Empire to the West (Old Irish) and to the East (Early Germanic).
    Late/Vulgar Latin kept the stress system of the Classical language, but it became much more dynamic, and loss of unstressed syllables followed as a result: this is a pan-Romance feature. What is difficult to establish is what the relationship is between this growing prominence of stress in Vulgar Latin and a similar situation in Early Germanic languages: Gothic WAIT “I know” (Compare the Greek cognate (W)OIDA or Latin VIDEO) already exhibits a loss of its inherited final (thus unstressed) vowel, making it look (already!) like a modern Germanic language (compare Dutch (IK) WEET).
    So: did Vulgar Latin stress grow more dynamic, with Early Germanic acquiring this feature through contact? Or was it the reverse? Unfortunately, chronologically, it is unclear which language was the first to develop dynamic stress. Hence it is unclear who influenced whom.
    Old French, however, almost certainly owes its more-dynamic-than-what is-usual-for-Romance stress to contact with Frankish: quite apart from the fact that Old French borrowed Frankish segmental phonemes (/h/ and /w/), making further phonological influence more than possible, it is telling that the only Romance varieties which are Old French-like in terms of the dynamics of stress are varieties of Rumauntsch, which, by a remarkable coincidence (yeah, right…) were also heavily Germanic-influenced.
    Moving on to the British Isles: thus, Brythonic Celtic (Welsh, Cornish, Breton) had Latin-type stress as a result of contact with Latin, with Old Irish preserving the original word-initial stress system. Both branches of Celtic subsequently subsequently underwent a radical seris of phonological and grammatical transformations through loss of unstressed syllables (but because in each branch stress fell on a different syllable in any word of more than three syllables, the lost syllable of a word found of one branch is preserved in the cognate word in the other branch, which does ease the task of reconstruction somewhat).
    For chronological reasons I suspect that in the British isles it was contact with Vulgar Latin which led to Brythonic stress becoming more dynamic, and in turn this stronger realization of stresss spread from Brythonic to Old Irish.
    My apologies for the length of this comment, but I’ve an excuse: language contact involving Late/Vulgar Latin is a specialty of mine, if I may say so myself. I can supply relevant scholarly references to whomever is interested.

  62. Rodger C says:

    My inpression was that Old British had a uniform penultimate stress, which isn’t how Latin works.

  63. My apologies for the length of this comment
    No apologies needed! If you’ve got something to say, take as long as you need to say it, is my philosophy.

  64. Al Moonlight says:

    “The sounds of Portuguese: A Cuban friend of mine one day said this about the pronunciation of (most likely Brazilian) Portuguese: Tiene la sensualidad del francés y la claridad del español.”
    I did study with a Brazilian teacher, but I always felt that in its sound Br. Portuguese compared to Spanish as Mississippi English does to Standard American English.

  65. marie-lucie says:

    Merci, Etienne. It did not feel long at all to read your comment.
    About Celtic (Gaulish) and Latin stress: there are a number of place-names in France where the same original Gaulish name occurred in different places, and has evolved sometimes with the Gaulish stress pattern, sometimes with the Latin one, so that the originally identical names end up quite different in French, for instance with Gaulish Nemausus ending up as Nîmes in the South and Nemours in North-central France. I quote these examples from Henriette Walter’s Le français dans tous les sens (I forgot the title of the English translation) which gives a few others.

  66. David Eddyshaw says:

    @Etienne:
    There is no reason to suppose that Brythonic ever had word-initial stress, and it has never shown a stress system dependent on syllable heaviness like Latin. It is surely quite impossible that the Old Irish stress system could be due to Brythonic influence, unless all that you mean is that the languages both had a stress accent system of *some* sort. This is not exactly a rare phenomenon cross-linguistically, and it hardly seems necessary to assume borrowing as a cause.
    Khalkha Mongolian has lost all final short vowels and lost distinctive vowel quality in all remaining short vowels in non-initial syllables; I don’t think anyone attributes this to Latin influence.

  67. marie-lucie says:

    DE, of course nobody would claim Latin influence on Mongolian given the distance between the territories, but with Celtic and Germanic we are talking about neighbours of Latin, spoken during the same historical period, so an influence there (in whatever direction) is not a priori to be ruled out of the question, which is why it is worth discussing.

  68. David Eddyshaw says:

    @marie-lucie:
    True enough. I didn’t explain myself very well. The Mongolian example was meant just to illustrate the fact that the phenomena in question are so very common as not really to *need* a specific explanation.
    Given that the actual details of the stress assignment rules in Latin, Brythonic and Goidelic don’t match at all, there doesn’t seem to be a lot left apart from the mere existence of syllables with stress. I suppose you might argue (as I take it Etienne is doing) that what has been borrowed is the change from some other system, like the IE pitch-accent system, or a completely absent stress system, to some sort of “dynamic accent” system, but surely you would expect then at least *some* further resemblance among the new systems? In the absence of anything like that, and given how common stress accent systems of one kind or another are anyway, isn’t the hypothesis of borrowing simply redundant?

  69. Rodger C: Old British (=Proto-Brythonic) indeed had uniform penultimate accent, which differs from the Latin system, granted, but it is much more Latin-like than a system of word-initial stress.
    Marie-Lucie: the issue of the stress placement of Gaulish place names is a *mess* that could spawn a lot of doctoral dissertations. The problem is that language shift from Gaulish to Latin surely did not take place uniformly across all of Gaul, and thus many instances of different stress placement of the same place-name in different parts of Gaul may relate to chronological differences as to when the place-name was borrowed into the local Vulgar Latin, as opposed to actual regional differences in Gaulish involving stress placement.
    David Eddyshaw: First of all, there is every reason to suppose that Brythonic originally had word-initial stress. See Schrijver, Peter. 1995. STUDIES IN BRITISH CELTIC PHONOLOGY, pp. 16-20 for an excellent summary of the arguments presented by several Celtic scholars defending this view, and Schrijver’s own additional arguments.
    Second, a clarification: the Brythonic influence which I suspect made itself felt upon Old Irish did not involve the *placement* of stress within the word: it involved the stress becoming much more dynamic in both languages, yielding phenomenal (and, tellingly, cross-linguistically unuusal) losses of unstressed syllbles in both branches of Celtic.
    Third, “Surely quite impossible”, you wrote. Err, why? I wrote above that I see this influence as likely for “chronological reasons”: the fact that Brythonic and Old Irish underwent these radical phonological changes at about the same time is already suggestive.
    Furthermore, this is the period when Christinaity was introduced to Ireland by Brythonic speakers: tellingly, the Latin loanwords in Old Irish form several strata, the oldest of which share all the sound changes which transformed primitive Irish (circa 500 AD) into Old Irish (circa 800 AD), the youngest of which do not, and some of which exhibit some of these changes. Finally, these Latin loanwords also show phonological signs of having entered Primitive/Old Irish via Brythonic speakers (and not directly from Latin/Romance speakers).
    So: “Surely quite impossible”? I beg to differ. If Brythonic speakers enjoyed enough prestige to introduce Christianity (complete with Latin loanwords) into Ireland, and if the Latin loanwords in question reveal that it was during the several centuries during which they were being borrowed by the Irish that the major phonological transformations which transformed primitive into Old Irish took place, then I would maintain that it is not only possible, but in fact very likely, that we are dealing with Brythonic influence upon Irish.
    In this sense I see this alleged influence as an insular parallel to the Frankish impact upon the Vulgar Latin of Northern Gaul which was to become Old French.

  70. marie-lucie says:

    Etienne, thank you for clarifying the differences between the changes in the various Celtic varieties, as well as the relevant dates and histories.
    language shift from Gaulish to Latin surely did not take place uniformly across all of Gaul,
    I agree. Apparently there is evidence that the Massif Central (a very mountainous area) maintained its Gaulish speech much later than the rest of the country.
    and thus many instances of different stress placement of the same place-name in different parts of Gaul may relate to chronological differences as to when the place-name was borrowed into the local Vulgar Latin, as opposed to actual regional differences in Gaulish involving stress placement
    That makes sense, but I had not thought about regional differences within Gaulish, as opposed to a Gaulish vs. Latinized pronunciation of the same names. But the example of Nemausus seems to go against your interpretation: with Southern Nîmes, surely we sould have expected such an important city in the Provincia Romana, with its typical Roman civic monuments, to have been much more (and earlier) Latinized than the then insignificant town or village which became Nemours (however, Wikipedia relates this name to an ancient Nemora, but I don’t know how acceptable this is).

  71. When I studied Linguistics in the mid-sixties a professor mentioned that (Iberian) Portuguese nasality was reminiscent of Polish.
    About the same time, perhaps the same professor said that those who learned Spanish first had trouble with Portuguese later. I also read this many years later, but the same source said that the reverse was also true.
    Have any here learned Portuguese first and had trouble learning Spanish later?

  72. I should say I learned Spanish fairly thoroughly and found (as I think the Hatman said above) that I could read Portuguese with little trouble, but I’ve never had a need to understand spoken Portuguese.

  73. Etienne: Thanks for the clarification. I had never heard anything about Old British having initial stress, but I agree that if so, it might well have been communicated to Primitive Irish.

  74. marie-lucie says:

    I have a Brazilian “facebook friend” whom I met a few years ago, and on her page she communicates in both English and Portuguese. I know a little Portuguese, but her exchanges with her Brazilian friends are often very hard to understand because they are so colloquial, not standard written Portuguese (even in the Brazilian version).
    I think that the idea that that starting with Spanish makes it hard to learn Portuguese (or vice versa) probably refers to interference due to the very similarity of the languages, which leads learners to infer the existence of even more similarity and to guess (often wrongly) what a word or form should be in the lesser-known language. This also happens between French and Spanish or Italian, and between the latter two. I guess it also happens between Dutch and German, and even more between the Scandinavian languages.

  75. Trond Engen says:

    Since I’m not a foreign learner, I don’t really know, but it seems to me that guessing forms would work quite well between Scandinavian languages. Even where the standard languages differ in vocabulary or syntax, the wrong guess will often be known as a specialist term or an archaic or dialectal word.

  76. Rodger C says:

    It occurs to me that I’ve read that original initial stress is found or inferred in so many IE languages that PIE (or was it just Greek?) may have had a light initial stress in addition to its pitch accent. Can someone clarify what I vaguely remember, and/or advise as to the current status of this hypothesis? (My readings in historical linguistics are extensive but mostly took place decades ago.)

  77. Marie-Lucie: for place-names in Southern France there is yet an extra complication, in that a language other than Latin and Gaulish was present: Greek. Don’t forget that Marseilles was originally a Greek-speaking city, which forged an early alliance with the Romans, and thus it is likely via the Greek language that Romans would have learned the major place-names of Southern France.
    (So: Just how exactly *would* native Latin speakers have phonologically adapted a Gaulish place name they had just learned either from Greek-speaking native Gaulish speakers or from native Greek speakers’ rendition of a native Gaulish form?)
    The above is not pure conjecture, by the way: W. Von Wartburg, examing the Gaulish and Greek loanwords in Provencal, showed that in many instances the Greek words must have entered local Vulgar Latin via Gaulish, or vice-versa: clearly Latin was superimposed upon a region where Gaulish-Greek bilingualism was already well-established.
    John Cowan: actually, if Brythonic was once a language with word-initial stress, just like Old Irish, then is is quite possible that this system goes back to Proto-Celtic itself, and was thus inherited by both Proto-Brythonic and Proto-Goidelic. To repeat myself, my belief is that through contact with Brythonic Old Irish stress became much heavier/more dynamic. But stress remained unchanged in terms of word position in both languages.
    Again, this is very much like the impact Frankish must have had upon the Vulgar Latin of Northern Gaul.

  78. if Brythonic was once a language with word-initial stress, just like Old Irish, then is is quite possible that this system goes back to Proto-Celtic itself
    That would entail that initial stress (even weak initial stress) was already present in Primitive Irish, in which case we have to ask why no trace of reduction is present in that language.

  79. I’m guessing that Trond is right, but I’m not a native speaker. However, my best Danish friend told me about how she once picked up a book and started reading. She wasn’t having trouble understanding, but it struck her as strange that a printed book should have that many misspelt words. Eventually, she looked in the beginning of the book and discovered that it was Norwegian, not Danish. We were teenagers when she told me this, so she probably hadn’t had the time to acquire a lot of specialized knowledge yet, although she was by nature quite an intelligent person.
    I do remember reading somewhere that there are some identical words with radically different meanings between the languages. I believe the examples given were between the Swedish and Danish words ‘lejlighed’ (either ‘apartment/flat’ or ‘occasion’ in Danish) and ‘væresse’ (either ‘existence’ or ‘room – in a building’) but no one ever said what these words meant in Swedish. Or I may have misremembered.

  80. Further to m-l’s comment re interference, I recall that my Brazilian buddy, when we were in Spanish classes together in the sixties, had an exchange with the prof in which he was certain that he had learned ‘nel’ as a Spanish word. When the prof said he was ‘probably thinking in Portuguese’ he boggled, and was unable to think at all for several seconds.
    I must email him and see if he recalls, and if he had further boggles, and if indeed he still speaks Spanish.

  81. SFReader says:

    @marie-lucie at March 31, 2013 09:15 AM
    –I guess it also happens between Dutch and German, and even more between the Scandinavian languages.
    It also happens in Slavic languages.
    Most extreme case I know is Ukrainian “uroda” (beauty) which means “ugly” in Russian ;-)

  82. Confusion of homonyms and homophones with widely differing meanings occurs within English, specifically between British and American varieties. Here are two examples, one of each:
    1. In David Lodge’s Small World, a Japanese who has never visited the U.K. is translating a contemporary British novel about working-class life, full of slang not found in his his dictionaries. He writes to the author whenever he finds something he doesn’t understand. When the hero of novel says something along the lines of “I could really use a fag tonight”, the translator asks the author “Is he thinking of indulging in homosexual intercourse? If so, why does he mention it to his wife?”
    2. I’ve read that during World War II, a bunch of American soldiers in Britain training for D-Day were attending services at a local church when the vicar thanked the U.S. for being Britain’s “succour” in time of need. They heard the word as “sucker” and were very offended.

  83. Yes, the Brits and the Americans have long engaged in homophonic intercourse.

  84. John Emerson says:

    This may be the time for me to ask how it happened that Norwegian Danish and Swedish got distinctive alphabets.
    Also, is there a common Scandinavian lnguage for e.g. legal documents? Or are there people who make their livings translating from Danish to Swedish, from Swedish to Norwegian, and from Norwegian to Danish (etc.) (It sounds like an insane-making job to me).

  85. John Emerson says:

    Responding to someone above: One of the classic Portuguese cantigas has a refrain “e u e?” (in some texts) which is exactly “et ou est?”

  86. Is there a difference in pronunciation between ‘sucker’ and ‘succour’? It seems to me the problem is that the American troops were unfamiliar with the word ‘succour’ — a ‘class’ issue, not one of ‘national varieties’.

  87. @Isidora
    With regard to your friend and her Norwegian book, I think we need to keep in mind that (1) Scandinavian languages are supposed to be pretty close to each other (2) Norwegian has two written varieties, one of which is based on Danish.
    The problem you mention can occur even between dialects of the same language. For instance, even though ‘Mongolian’ is supposed to be one language, the meaning of words in Mongolia and Inner Mongolia is often quite different. One of the first I heard about was the word зарах, which means ‘to sell’ in Mongolia, ‘to throw away’ in Inner Mongolia. An Inner Mongolian friend was initially puzzled by all the perfectly good cars in Ulaanbaatar with signs saying that they were to be junked, when they were actually for sale.

  88. SFReader says:

    If you look it up, Khalkha Mongolian has a range of meanings for the word, including “to expend, spend, consume, disburse” which is quite similar to Inner Mongolian use of the word.
    turn over зарах [Нэр үг]
    to expend зарах [Үйл үг]
    to overbook зарах [Үйл үг]
    to use зарах [Үйл үг]
    to employ зарах [Үйл үг]
    to consume зарах [Үйл үг]
    to disburse зарах [Үйл үг]
    to dissipate зарах [Үйл үг]
    to drop зарах [Үйл үг]
    to fence зарах [Үйл үг]
    to go зарах [Үйл үг]
    to hand-out зарах [Үйл үг]
    to market зарах [Үйл үг]
    to misappropriate зарах [Үйл үг]
    to negotiate зарах [Үйл үг]
    to outlay зарах [Үйл үг]
    to put out зарах [Үйл үг]
    to realize зарах [Үйл үг]
    to sacrifice зарах [Үйл үг]
    to sell зарах [Үйл үг]
    to slave зарах [Үйл үг]
    to spend зарах [Үйл үг]
    to use up зарах [Үйл үг]
    to vend зарах [Үйл үг]
    to vent зарах [Үйл үг]
    to ware зарах [Үйл үг]
    to lay зарах [Үйл үг] [худ.]
    to send зарах [Үйл үг]
    to nfffer зарах [Үйл үг] [Scot.]
    lumbering зарах
    niffer зарах
    skittle away зарах
    tine зарах
    carry зарах
    exert зарах

  89. SFReader says:

    I am not sure whether the word “niffer” is English or Scots, but it’s nice to know that it translates into Mongolian as “zarakh”.
    Does anyone know what exactly does it mean?

  90. marie-lucie says:

    Bathrobe: An Inner Mongolian friend was initially puzzled by all the perfectly good cars in Ulaanbaatar with signs saying that they were to be junked, when they were actually for sale.
    This sort of thing is frequent between Canadian and European French. I was surprised to see chars usagés advertised in Canada, since usagé to me means ‘worn out by constant use’. Of course the sign referred to ‘used cars’, which in France would have been voitures d’occasion ‘secondhand cars’.

  91. quite similar to Inner Mongolian use of the word
    Yes, I realise that the original or ‘root’ usage as shown in dictionaries is quite broad, but actual usage seems to have solidified into certain narrower meanings that lead to the kind of puzzlement that I mentioned.

  92. i wouldnt say to drop, to go, to lay, to carry are all zarakh, maybe other dialects use those meanings, otherwise yes, all other meanings match it, dont know about tine niffer skittle away

  93. Rodger C says:

    @Bathrobe: As an American, I don’t think I’ve ever actually heard the word “succour” outside the phrase “comfort and succour all those who in this transitory life,” etc.

  94. @Roger C
    I’ve hardly ever heard it, either; it’s not part of my everyday vocabulary. I also doubt that it crops up very often (if at all) in the speech of most ordinary Englishmen. Perhaps it’s used more in the UK than the US, but it seems to me that well-educated Americans would have at least guessed at what the vicar was saying rather than taking offence.

  95. marie-lucie says:

    Bathrobe: a bunch of American soldiers in Britain training for D-Day …
    This description does not suggest a “well-educated” American audience.

  96. Trond Engen says:

    I may have mentioned this before. During a national debate about police violence a couple of decades ago, someone said that a dangerous subkultur had been allowed to grow in the force. A huge row broke out, since it was understood as subbkultur “trash culture”.

  97. marie-lucie says:

    Mongolian “zarakh”
    There is a difference between “translation” and “meaning”, since a very general word such as Mongolian zarakh can be translated in a variety of ways depending on the context. Compare for instance English get, which can be understood in multiple ways depending on the following word.
    It seems to me that the basic meaning of zarakh could be ‘to act on or with’, which could cover a variety of actions depending on the thing acted on or with, hence the many possible translations.

  98. Rodger C says:

    @Trond: The word “subculture” has acquired similar pejorative associations in English too, unconected with any pun.

  99. John Cowan: why should we expect reduction? This initial stress may have been similar, acoustically, to Proto-Indo-European stress, which was more a pitch then a stress accent.
    After all, we find no large-scale loss of unstressed syllables in Sanskrit or Homeric Greek, both of which inherited the original stress system of Indo-European.

  100. As I recall the story, the soldiers in the church were mostly enlisted men. I wonder if “succour” might be found in the Book of Common Prayer? If so, it would have been familiar to most of the English of that era – and all the vicars – while unfamiliar to Americans unless they were Episcopalians or read the kind of archaic literature in which the word would be found. From a quick Google it looks like it’s still used in Indian journalism: “Some succour for birds this summer”, “Insurance succour for twins with palsy”, “seeking succour for terror-tainted released”, and “Perpetual Succour High School” – all in the The Times of India.

  101. Michael: Bingo. The traditional burial rite has the minister say, when the body is committed to the ground, “In the midst of life we are in death: of whom may we seek for succour, but of thee, O Lord, who for our sins art justly displeased? ”

  102. why should we expect reduction? This initial stress may have been similar, acoustically, to Proto-Indo-European stress, which was more a pitch then a stress accent.
    Then it’s pitch, not stress. They’re two different things. With pitch you wouldn’t expect reduction; with stress you would.

  103. Gettng back to the question marks for moment: the first “Random book from my library” is pretty consistently Ð’ дебрях Уссурийского края by Владимир Клавдиевич Арсеньев vel sim, in both IE and Firefox. Someone somewhere is playing games with the encoding.

  104. SFReader says:

    @marie-lucie
    I think the basic meaning was “to use”, which then expanded to include “use up, expend, spend, sell, throw up, exploit” and so on and on…

  105. Rodger C says:

    @John Cowan: Also there’s the phrase I quoted, which is from the intercessions in the communion service.

  106. SFReader says:

    Starostin’s etymology gives as basic meaning
    Proto-Mongolian: *ǯaru-
    Meaning: to send, employ
    Russian meaning: посылать, давать задание
    Written Mongolian: ǯaru-
    Khalkha: ʒar-
    Buriat: zara-
    Kalmuck: zar-
    Hence, I think this is what happened. First it meant to send, to employ, then widened to include to use, to exploit, then widened again to expend, spend, consume, throw up, to market, to sell etc.

  107. Etienne, let me summarize what you have said above:
    1) Dynamic stress is an areal feature of Insular Celtic and Germanic that later spread to Old French from the latter.
    2) Word-initial stress is an areal feature of Celtic, Germanic, pre-Classical Latin, and possibly Early Basque.
    3) Classical Latin acquired a novel penult/antepenult stress system, contact with which caused Basque and Brythonic to lose word-initial stress; the latter acquired uniform penultimate stress.
    4) Vulgar Latin then acquired dynamic stress, as did Early Germanic; it is not clear which one acquired it from the other.
    5) Old French and Rumantsh acquired their extreme dynamic stress from Germanic (specifically Frankish in the former case).
    6) Dynamic stress (with no change in the stressed syllable) spread from British Latin to Old British to Old Irish.
    Is that the correct story as you see it?

  108. marie-lucie says:

    SFR, looking at the whole list of meanings, I think that “act upon/with”, etc fits better than “use”, but I don’t want to start an argument about it.

  109. marie-lucie says:

    SFR, also, I don’t know what the difference is supposed to be between “to employ” and “to use”.

  110. SFReader says:

    Looking at Russian meaning, “to employ” means to “give a task”

  111. “First it meant to send, to employ, then widened to include to use, to exploit, then widened again to expend, spend, consume, throw up, to market, to sell etc. ”
    the root zar means a message as in zar medee, zarlakh – announcement, zarax meaning to spend, expend will be more like zartsuulax, but zarakh is okay too in that sense, to sell is another direct meaning, to employ meaning is more of using someone’s labor informally, if one feels lazy and asks someone to do something for him/her that’s the most direct usage of the word zarakh, it doesnt mean to hire someone to do a job, zarts is a serf, but to exploit or to slave are too strong words to match zarakh, there are other words for those words
    all other meanings would require some or other modification of zarakh, not the word itself i guess, though when i read the list, all the english words really match zarakh, the comparison with get is useful i guess
    about the meaning of act upon/with, the english words in the list can be summarised that way i guess, but zarakh, i am not sure, it’s the meaning of spend/expend mostly

  112. marie-lucie says:

    Thank you, read.

  113. SFReader says:

    —the root zar means a message as in zar medee, zarlakh – announcement
    Yes, this fits nicely with “to send, to employ, to give a task” being a primary meaning.
    Comparison with “message” is instructive. It’s derived from Latin verb “mittere” (to send, to throw) and its million derivates in Latin
    permittere, emittere, missivus, committere, admittere, missus, intermittere, remittere, submittere, circummittere, intromittere, supermittere, transmittere, praemittere, conmittere, demittere, immittere, inmittere, amittere, ammittere, mittens, omittere, praetermittere, manumissio, missum, dimittere, missa

  114. my pleasure, m-l

  115. John Emerson says:

    My guess is that “succour” in the US is purely Episcopal (Anglican) and that it’s not a matter of educational level.

  116. John Cowan: not exactly. I’d summarize my views like this:
    1-Germanic, prehistoric Latin, all or at least some Celtic varieties (certainly those which were the ancestral forms of the Insular Celtic languages), and possibly Early Basque, shared an areal innovation whereby (for the Indo-European languages) the Indo-European pitch accent was fixed on the initial syllable of a word. This is a matter of stress position, not stress realization.
    2-In the transition from Prehistoric to Classical Latin this word-initial stress was lost and replaced by the Classical Latin system (“novel penult/antepenult system”, in your own words)
    3-Through language contact with Latin, Basque and Brythonic lost their word-initial stress systems and acquired more Latin/Romance-looking ones (involving fixed penultimate stress in the case of Brythonic).
    4-Before the fall of the Western Roman Empire both Germanic and Vulgar Latin stress become a more dynamic one, with unstressed syllables being lost in both languages. Because the change took place at about the same time, at a period when intense language contact was taking place, this similarity is unlikely to be a coincidence, but the direction of the influence (Was it a Germanic innovation that spread to Vulgar Latin, or vice-versa?) is unclear.
    5-What is clear is that, at a later date, Old French and Rumauntsch owed their exceptionally dynamic stress to Germanic influence (specifically Frankish, in the Old French case).
    6-Back in the British isles, during the late Roman imperial period, the stress systems of both branches of Celtic were also becoming more dynamic, indeed more so than the stress systems on the continent. For chronological reasons this is likely to be an instance of Latin/Romance influence upon Brythonic, with the latter language in turn influencing Primitive Irish.
    IMPORTANTLY, changes 1-3 involve the position of the stress in a word only: changes 4-6 all involve stress being changed in its realization (from a pitch stress to a dynamic stress), all the while keeping the same position in the word.
    So the Old Irish-Germanic commonality (dynamic word-initial stress) is not due to a single contact event (which is why your point 1 is misleading): rather, word-initial *pitch* stress was originally (pre-Roman Empire) an areal feature of the ancestral forms of both languages. As for this initial pitch stress becoming dynamic in both languages, this is certainly due to later (late Roman Empire) language contact (see my own point 4).
    Hope this is clearer.

  117. marie-lucie says:

    JE: My guess is that “succour” in the US is purely Episcopal (Anglican) and that it’s not a matter of educational level.
    But not all US Protestants belong to that denomination. In any case, the word is so rare that even churchgoers who don’t listen carefully to sentences with traditional but unusual wording, not only might not really know its meaning but might fail to recognize the word in a different context, as happened in this example.

  118. J.W. Brewer says:

    “Succour” appears about a half-dozen times in the KJV, which would have been widely known by non-Episcopalian US Protestants at least back during the WW2 era, but generally appears there as a verb rather than noun (“succourer” is in Rom. xvi:2).

  119. marie-lucie says:

    succour
    I once got into a conversation with a taxi driver who told me he hated the KJV version used in his church because he could not understand it, and he asked me if there was a way of reading the Bible in normal language. Such a person might have seen and heard the word succour but have only a vague idea of what it meant, even if he could match the sound and the spelling.
    This word seems to be an old borrowing from French le secours, which is part of general vocabularly, with the meaning of ‘assistance, especially in an emergency’. This is a noun, but there is also the verb secourir, a transitive verb meaning ‘to give assistance to …’. And yet other nouns, of more recent vintage: le secourisme ‘first aid training’ and le/la secouriste ‘person with first aid training’.

  120. John Emerson:
    “Succour” is certainly not “purely Episcopalian”. I just got back from As You Like It, where I heard it once, and that reminded me to look it up in Schmidt’s Shakespeare Lexicon and Quotation Dictionary, which lists roughly a dozen instances of the noun and half a dozen of the verb. I’m sure it’s found in other literary works, though some of those uses may be more Episcopalian than literary – for instance, I wouldn’t be surprised to see it in a Barbara Pym novel.

  121. Marie-Lucie, the OED says:

    Middle English sucurs, socurs, socours, etc., < Old French (Anglo-Norman) sucurs, soc(c)ours, etc. (modern French secours) = Italian soccorso < succursus , n. of action < succurrĕre ‘to succour’. The final -s was at an early date apprehended as the plural suffix and a new singular (succour) came into existence, the plural of which is identical with the old singular.
    German succurs (from Old French) is used in the military sense, and Middle Dutch secors, socoers, in the general sense.

    Etymonline adds that succurrere in Latin meant literally “run to help,” from sub ‘up to’ + currere ‘to run’

  122. I was raised in the Anglican Church of Canada and beame quite accustomed to the word succour, but I have encountered it rarely since the age of about sixteen.
    I contacted my Brazilian buddy; he can’t recall attempting to use the word nel in Spanish class, but thinks the interference came from Italian, which he was accustomed to hearing where he grew up in Brazil, and recalled the song Nel blu di pinto di blu. He said the Portuguese word would be no.

  123. Trond Engen says:

    John E.: I forgot to answer this. Scandinavian languages have different alphabets and, more disturbing, different orthographic conventions because of the distinct orthographic traditions of Denmark and Sweden, and maybe especially because of different choices in regularizing and deornamentizing in the 18th/19th century. Early 20th century attempts to bring them closer were only partly successful, since few of the suggested means were undertaken completely in all three countries. The letter &ltå&gt was taken into the Danish and Norwegian alphabets, but not in the same place as in the Swedish, and the letters &ltæ&gt and &ltø&gt weren’t replaced by &ltä&gt and &ltö&gt (or vice versa). Swedish replaced initial &lthw-&gt with &ltv-&gt, while Danish and Norwegian Bokmål didn’t (and Nynorsk couldn’t), but it did not replace &lt-ck&gt with &lt-kk&gt.
    There’s not much common lawmaking as far as I know. When there is, the laws are written up separately in the three countries.

  124. Etienne: Much clearer, thanks. Obviously you are better equipped to summarize your views than I was!
    But I have concerns. The transition from pitch to stress accent in Latin must have occurred between (1) and (2), or we wouldn’t have the inimicus vowel reductions. So the accent must have been dynamic already in Classical times (even though it became more so in the post-Classical period). The Old Latin period was not one of heavy contact between Germanic and Latin, obviously.
    What’s more, dynamic stress is pervasive in Germanic, and I know of no direct evidence for a period before strong dynamic initial stress: there is Verner’s Law, but it is undatable. Umlaut effects and syllabic loss are already there before the beginning of our data. So I think the coincidence of timing required by your (4), whose force is crucial to your (6), is not yet demonstrated satisfactorily.
    Indeed, considering how limited the survival of the PIE pitch accent is (Lithuanian, Latvian along with initial dynamic stress, Serbo-Croat — Norwegian and Swedish pitch accent is secondary), it’s questionable how much we actually need an explanation of pitch > stress changes. Given the independent evolution of Europe and Asia on this point, independent evolution of Brythonic and Celtic doesn’t seem so unlikely.

  125. Etienne says:

    John Cowan: on your first point: it is true that vowel reductions do point to *initial* stress in Latin as having been dynamic at some point in time, but crucially, there is no comparable reduction of unstressed vowels in Classical Latin.
    But all this means is that the shift in stress position from prehistoric to Classical Latin probably took place after the realization of stress had been weakened and had become either a very weak stress or a pure pitch accent. We could also assume that the realization of the new (Classical) Latin system was unrelated to that of the older (prehistoric Latin) initial stress. Take your pick.
    On your second point: this is a little trickier: hang on:
    Actually, the evidence of Celtic and (the oldest stratum of) Latin loanwords points to their having entered Germanic at a time when unstressed syllables were still fully realized. This is true of Umlaut as well: the name of ATTILA is preserved in Old High German as ETZEL, indicating that umlaut and subsequent reduction of unstressed /i/ took place at the earliest during the late Roman Imperial period.
    A Latin word such as VINUM, in order to yield forms such as English WINE, Dutch WIJN, German WEIN, or more tellingly, Old High German WIN, must have been borrowed as a bissyllabic noun (*winan or *winam) in Early Germanic, before unstressed syllable loss took place, as otherwise VINUM could have simply been adapted as (say) Old High German WINA or WINU (cf. Old High German GEBA “gift”: such a word was phonotactically possible).
    Tellingly, apart from the loss of final short open (or closed by a nasal) vowels, there is little reduction of unstressed syllables in Gothic: a Latin loanword into Gothic such as ASILUS from ASINUS (the replacement of /n/ by /l/ was probably a morphological adaptation of the loan to a common Gothic ending) preserves every syllable of the original.
    The same argument applies to Old Irish and its latinisms, incidentally: Latin VINUM was borrowed into pre-Old Irish and is attested in Old Irish as FIN. This indicates that it too must have been borrowed as a bissyllabic noun: had the word been borrowed at a stage AFTER Old Irish had lost the primitive Irish final syllables the Latin word would have been borrowed as *FINA or the like, as phonotactically such a noun was quite possible in Old Irish (cf. BATA “stick”).
    Thus, Old Irish and Early Germanic have more in common than a dynamic initial stress system: this stress system became a dynamic one and unstressed syllables were lost, in both instances, *AFTER* a first wave of Latin loanwords had entered.
    Recall, in turn, that Latin loanwords entered Old Irish via Brythonic. So the loss of final (unstressed) syllables that we observe in a Latin loanword such as Welsh GYYN, also from VINUM, must also, in Brythonic, postdate the entrance of the first Latin loanwords.
    So: just after Christianity and the first waves of Latin loanwords had entered Goidelic, Brythonic and Early Germanic, we witness the rise of a much more dynamic realization of stress, as attested by the massive loss of unstressed syllables. Quite a coincidence, if language contact is unrelated to this change.
    Oh, and of the three groups, Brythonic is the only one where stress had earlier shifted position (from an initial to a penultimate position) and become more Latin-like. By yet another odd coincidence, Brythonic is the only one of the three languages whose speakers were politically incorporated in the Roman Empire, and thus may be assumed to have been more heavily exposed to Latin.
    So: I hope you will understand that I stand by my points 4 and 6 because the evidence, to my mind, is rock-solid.

  126. marie-lucie says:

    Hungarian ATTILA vs Germanic ETZEL
    Attila too, as a Hungarian word, had/has initial stress. Perhaps the area of initial stress was formerly even wider than we know?
    Etzel : is this the original of the name Edsel?

  127. Correction: the Welsh reflex of Latin VINUM is GWYN, not GYYN.
    Marie-Lucie: Actually, ATTILA is etymologically a Gothic name, meaning “Little father”. Its transmission from Gothic to Old High German (along with a number of other Gothic words found in High German, notably in Bavarian dialects)certainly took place long before Hungarian began spreading in Pannonia.
    While some Hungarian nationalists/scholars like to claim a continuity of some kind between the Huns and the Hungarians, the fact is that not a single word of Gothic origin (Gothic was the lingua franca of the Hunnic Empire, and indeed as a result we know next to nothing of the Hunnic language itself) can be found in Hungarian.
    This suggests that, linguistically, by the time the first speakers of Hungarian settled in Pannonia Gothic was no longer spoken there: my understanding is that Slavic and *possibly* Romance were the languages of Pannonia at the time of the initial spread of Hungarian (as usual, should you or anyone on this blog want some supporting references, the State Department will disav-err, I mean, I will gladly supply said references).
    As for the word-initial stress of Czech and Slovak, this is traditionally explained as an instance of German influence, and this may be true of Hungarian as well. Interestingly, in all the three non-German languages this word-initial stress has been much weaker than stress in German or Germanic itself, if the fate of unstressed syllables is any indication.
    Finally, as for ETZEL and EDSEL: I haven’t a clue.

  128. SFReader says:

    Some of the population might have still spoken Avar language, thought to be an Oghur Turkic language closely related to Hunnic.

  129. Etienne: I don’t think Hungarian initial stress is likely to be a contact phenomenon, since it is pretty pan-Uralic and in fact is reconstructed for Proto-Uralic. I am also surprised to hear you say it is a weak stress, since to my ear at least it is very strong, stronger than in English, though not accompanied by any reduction phenomena. This may be because of vowel harmony, however, particularly on the view that in vowel harmony part of the quality of a non-initial vowel is directly and synchronically derived from the preceding vowel (having no, or only a partial, intrinsic quality of its own) rather than being a diachronic copy or imitation of it.
    Edsel apparently is indeed a variant of Etzel, though there is another theory that it is an umlauted variant of adel ‘noble’.

  130. Edsel apparently is indeed a variant of Etzel
    Whodathunkit!

  131. marie-lucie says:

    Merci, Etienne et JC.
    JC, I would guess that Edsel was a respelling of Etzel, which looked too German to be acceptable in the US at the time. Edsel has always looked bizarre to me. I only know about Edsel Ford, but was the name ever common?
    In any case, it cannot be a variant of Adel, which would have been Edel not Edsel. But edel does exist in German, as the adjective meaning ‘noble’, while the noun Adel is ‘nobility, aristocracy’ (from the Collins dictionary).

  132. J.W. Brewer says:

    Gothic in etymology or not, the Hungarians do like their supposed Hunnishness. The only actual person named “Attila” by his parents I have ever met was a college contemporary of ethnic-Magyar background (IIRC, his family had managed to relocate to Ontario after somehow getting out of Ceaucescu-ruled Transylvania). I don’t think there’s another ethnic group in North America where that’s a boy’s name with positive rather than negative valence.

  133. Isidora says:

    English-speaking Eastern Orthodox Christians in the United States use the word ‘succor.’ I wouldn’t say that we use it constantly at all, but it is a part of our liturgical vocabulary. I am guessing that how often it is used depends on a couple of things. One of them is which service is being served (for instance, I cannot remember ever hearing it used during the Divine Liturgy or during the fixed portions of the All-Night Vigil.)
    I strongly suspect that the frequency of use of ‘succor’ and other specific words depends on the translation into English. We, quite unfortunately, have several dozen different translations currently in use, as opposed to practically every other part of the world where there is Orthodoxy, there is only one text of the all the prayers and services for each county/language group. It’s absolutely maddening to be visiting another parish and saying the Lord’s Prayer (or practically any other prayer, for that matter) and not know which of several commonly used English texts is going to be used.
    Then there is the issue that many of our liturgical books still in print and still in use were translated into English many, many decades ago. In a certain translation of the Book of Needs (Трзвник – I sincerely hope I spelled that correctly and hope that someone will please correct me if I got it wrong.) there is a prayer where the priest prays (among many other things) for “protection from the invasion of aliens” while the kids try not to giggle, and there is another place where the phrase “Thine All-Holy and adorable Spirit” is used, and even I have to refrain from rolling my eyes and just remember what the word really means. There was no fault with the translation when it was made, but it was made long enough ago that a couple of very unfortunate semantic shifts have crept into American English while the liturgical text has remained unchanged.

  134. It’s Требник (for some reason that Russian Wikipedia article is linked to the English Wikipedia one Breviary, which is misleading). Here‘s a Church Slavic Trebnik from 1882. And yes, it’s too bad you have to deal with a bunch of outdated translations!

  135. Etienne says:

    John Cowan: okay, for Hungarian its initial stress goes back to Proto-Uralic. Even so, I wonder: could there be a historical connection between Proto-Uralic initial stress and that found in Western Indo-European languages?
    Before the expansion of Slavic Proto-Germanic might well have been in close contact with some Uralic languages, and indeed in the case of Proto-Fennic a solid argument has been made that its phonology deviates from that of Proto-Uralic because of Germanic influence (the many Proto-Germanic-looking loanwords in Fennic certainly do indicate that there was language contact).

  136. J.W. Brewer says:

    Au contraire, the variousness and multiplicity of English translations of Orthodox liturgical texts (including the ubiquitous experience of hearing people standing next to you pray/chant in slightly different forms of words than you yourself are) is one of the glories of Anglophone Orthodoxy, because it has saved us from the fate of standardization around a lousy translation. Since pretty much every new standardized English-language liturgical text cranked out in the Roman Catholic or Protestant world in the generation following 1960 (the time frame in which a lot of Orthodox parishes in the US were switching into English from their heritage languages) was (imho) tin-earedly dreadful, that was a very significant risk. That said, translating competently into the archaic KJV-ish register of English is quite hard to do well w/o specialized training even if (modern) English is your native language, and . . . many pious and well-intentioned translators have lacked either the specialized training or the native-speaker intuitions. And to boot a lot of the underlying texts are written in a Greek or Slavonic so technical/florid/turgid/”thick” as to create great difficulties even for translation into quote contemporary unquote English.
    All of that said, I keep coming back to wanting to adapt the analysis of Le Corbusier (who was discussing the architecture of New York City): “A hundred times have I thought [the Englishing of Orthodox liturgical texts] is a catastrophe and 50 times: It is a beautiful catastrophe.”

  137. I was thinking that Trzvnik was a pretty extreme consonant cluster for Russian. —Joe Btfsplk, or was it Mr. Mxyzptlk?

  138. Hat, thank you so much for correcting my spelling, especially considering that I managed to put a wrong consonant into it in addition to the incorrect vowel and you still recognized the word I was trying to spell. I’m thinking I ought to leae off trying to type even single words in Cyrillic until I have brushed up on my alphabet. It’s been years since I’ve really needed to use it except for reading the inscriptions on icons, etc., and those are all in Church Slavonic script, not Russian.
    That is definitely weird for Trebnik to be linked to Breviary. It’s remininscent of when people call an Orthodox prayer rope a Rosary. The two are not the same thing. They have distinctly different physical forms, and, more importantly, they are used in an entirely different fashion.

  139. Attila shows up in the Niebelungenlied as Etzel, with both umlaut and the second sound shift. And Theodoric of Verona shows up as Dietrich von Bern.

  140. Isidora: Well, rosary is the most familiar name for a device used to count prayers as one recites them. Such structures are used by many religions. Quoth Wikipedia:

    Islamic prayer beads, called “Misbaha” or “Tasbih”, usually have either 99 or 33 beads. Buddhists and Hindus use the Japa Mala which usually has 108 beads, or 27 which are counted four times. Baha’i prayer beads consist of either 95 beads or 19 beads strung with the addition of five beads below. The Sikh Mala also has 108 beads. The secular Greek “komboloi” has an odd number of beads—usually one more than a multiple of four, e.g. (4×4)+1, (5×4)+1. Roman Catholics use the “Rosary” (Latin “rosarium”, meaning “rose garden”) with 54 with an additional five beads whereas Eastern Orthodox Christians use a knotted “Rosary” with 100 knots, although “prayer ropes” with 50 or 33 knots can also be used. Although Anglo-Catholics have used the Dominican rosary since the 19th century, in the 1980s Rev. Lynn Bauman from the Episcopal church in the United States introduced a Rosary for Anglicans with 33 beads.

    Indeed, bead originally meant ‘prayer’, and is related to bid ‘ask, pray’, and more distantly to Latin festus (sense uncertain) in manifest and infest.

  141. Well, I’m going to hazard a guess that part of the issue with Orthodox liturgical texts in the original languages being “technical/florid/turgid” is that those texts are often being used to carefully express very precise theology as poetry, often containing paradox. It is fairly common to find word/idea plays in the hymnography. These things are always going to be in the very nature of the texts themselves and it doesn’t make them easy.
    A separate issue with translating into English is that the original languages of Greek a d Slavonic have vastly different grammatical structures, conventions, and preferences than English does, to the point where the languages are pretty much incompatible if one tries to do anything so ill-advised as to translate the grammar of the original literally into English, a language in which adjectival nouns are few and far between and which does not choose to make extraordinarily efficient use of its participles the way some other languages do.
    Oh, and add to this that sometimes things have to be translated to fit with certain melodies. I say sometimes because a great deal of Orthodox music is engineered so that just about any text whatsoever can be chanted or sung to that melody, but not always.
    Personally, I have a very strong preference for the archaic register for a number of reasons, and J.W. Brewer is absolutely correct that takes specialized knowledge to do properly. Last week, I was at a service and found out afterwards that the only copy available of one of the hymns specific to that day was in a ‘you’ translation, so our poor reader was forced to conjugate on the fly as he chanted. He did fine, but that was making minor adjustments, not doing wholesale translation from one of the original languages.

  142. J.W. Brewer says:

    I should perhaps clarify that I don’t mean florid/turgid (or even “technical”) as pejoratives — the difficulty arises I think because of different aesthetics within different cultural / literary traditions — it is to this day I think not uncommon for certain sorts of Balkan/Levantine rhetoric to sound (if translated more or less word for word) so unnecessarily flowery and overwrought for the context that a stereotypical plainspoken Anglo-Saxon will incorrectly infer insincere flattery or evasive duplicity when the speaker is simply employing a rhetorical style suitable for the context according to the lights of his own speech community. But translating non-literally in order to avoid creating that misimpression raises a host of problems and risks of its own, and in any event it is reasonably common cross-linguistically for liturgical language to be in a distinct register which (rightly or wrongly) comes to take on numinous associations. Indeed, one can say the very foundation / basement level of Orthodox liturgical texts is comprised of the LXX Greek version of the Psalms, which is a highly problematic translation from the Hebrew into often odd and awkward-sounding Greek, at least if you go by the standards of what most modern Anglophone theorists of translation (including those engaged in Scriptural translation) think makes for good translation practice. But there it is, and the Church has used that version for two millenia as perhaps a sign that modern theorists of translation do not have a monopoly on wisdom.

  143. Colin Batchelor says:

    Re Attila as a boy’s name, the Icelandic variant Atli is not uncommon. The wikipedia page even has a histogram: http://is.wikipedia.org/wiki/Atli_(mannsnafn)

  144. Trond Engen says:

    Atle is common in Norway too, although clearly limited to my cohort. Much like my name.

  145. marie-lucie says:

    A younger friend of mine married a Norwegian man called Atle.
    Trond seems to be very common in Norway. I had seen it many times in writing but only learned very recently how to pronounce it.

  146. There was a very famous mathematician, Atle Selberg. I never realized his name was the equivalent of Attila.

  147. I had seen it many times in writing but only learned very recently how to pronounce it.
    Is it /tron/ or /trun/?

  148. marie-lucie says:

    LH, /trun/ is what I learned (from a Norwegian friend). Trond, can you confirm?

  149. We Hattics were dancing the selfsame Portuguese jig back in 2007 at “Lusophone Nations Unite”. In the comments there’s a discussion of Frøken Smilla and how she’s Miss Smilla in BrE but just Smilla in AmE, at least in the title. Which reminds me of how Strindberg’s Fröken Julie gets her title as the daughter of a count, and therefore the play should be called “Lady Julie” in English.

  150. David Marjanović says:

    Not all these changes made it into the standard: in particular /k/ > /(k)x/ did not,

    The change of long /kː/ to [g̊χ] and of initial /k/ (where length wasn’t phonemic) to [g̊χ] or [χ] is today limited to the Alemannic and South Bavarian dialects. The Central Bavarian dialects participated, but then undid it (except they didn’t get the aspiration back).

    and only /d/ > /t/ consistently appears among the devoicing changes.

    In spelling, the others consistently don’t appear after Old High German… but I think there was something more complicated going on here:
    1) The voiced fricatives (other than [z], which was distributed between [s] and [r]) turned into voiceless lenis plosives: [β~v d~ð ɣ] became [b̥ d̥ g̊].
    2) Being voiceless, these sounds were often written p t c/k by monks who were used to Latin and local Romance languages. This shows up in Lombardic, Old Alemannic and Old Bavarian.
    3) A few hundred years later, /θ~ð/ turned into [d] which must have quickly devoiced by analogy with all other plosives; the already existing /d/ got out of the way by turning into an unaspirated fortis [t] as found in the surrounding Romance and Slavic languages. Since then, the old /d/ has always been spelled t, and the new one has pretty consistently been spelled d.
    The oldest mention of potatoes in Upper Austria, IIRC, dates from 1644 and is spelled ertöpfl. Comparison with the modern Erdäpfel shows three interesting features: 1) /d/ spelled t; 2) the second e was already silent and the /l/ syllabic; 3) L-Umlaut – syllabic /l/ rounding the vowel in the preceding syllable (Wikipedia says this is much older); 4) the /l/ that triggers this change is still there – it has turned into a vowel since ([ɪ] or thereabouts).

    What is more, Standard German has many borrowings from Low German and more recently English, which are unshifted.

    There are borrowings from Low German, but I wouldn’t say “many”; and the English ones, with very few exceptions, aren’t nativized enough in pronunciation or spelling to be mistaken for native.

    Hamburg-accented German can be imitated perfectly* by an English person simply by assuming a Yorkshire accent.

    I’ve heard this in the other direction: Yorkshire English is easier to understand for people who know Low German than Standard English is.

    Russians can’t pronounce any of that even if it was spelled in Cyrillic.

    That’s true, but they have rather little trouble understanding it when they hear it.
    Polish, Czech and Slovak are all mutually intelligible most of the time, and I do mean the standards, not just intermediates like Silesian. Polish and Croatian is a combination that requires more of an effort, but the Croat understood the Pole better than vice versa the one time I witnessed such an attempt at bypassing English. (The conversation ended in: “Nie rozumiemy się!” – “Razumjemo se!” :-D meaning “we don’t understand each other” – “we do”.)

    I couldn’t even come close to reproducing the difference between TRZY and CZY.

    Does it help if I explain it in theoretical terms? :-)
    Most affricates on this planet have a single place of articulation. For instance, the English ch doesn’t really consist of an alveolar [t] and a postalveolar [ʃ] – the whole thing is postalveolar.
    The Polish cz is just such a thing.
    The Polish trz, on the other hand, is not. It really does begin with an alveolar [t] that is released into a postalveolar [ʃ]. If you say it too slowly, it comes out as [tsʃ].
    The only other such case I know is the Tyrolean (c)k [g̊χ], a velar plosive released into a uvular fricative. Unlike the Polish situation, it does not contrast with a similar but better-behaved affricate.
    Postalveolar stops don’t seem to occur outside of Australia, and there are no IPA symbols for them. People usually resort to the symbols for palatal stops.

    “Russian morkov’ (carrot) becomes mrkva in Croatian and mrkvicka in Czech”
    These aren’t the result of phonemes dropping out, in fact, they aren’t really consonant clusters–they’re syllabic resonants, in a direct line of descent from Proto-Indo-European. (Maybe the carrot-word isn’t, but the phonemes are.) Russian realized the syllabic resonants with vowels in a variety of ways.

    Nope. The syllabic consonants of Czech, Slovak, Slovene and Serbocroatian aren’t inherited from PIE; they’re usually not in the same places.

    Don’t we have syllabic resonants in English, too, in some environments, but due to unstressed vowels dropping out: butter, little?

    Of course. One big difference is that they were sometimes stressed in PIE, like in *wĺkʷos “wolf”.

    there’s also evidence that pre-Classical Latin bore an initial stress that was strong enough to trigger vowel weakening

    Indeed; for this reason, it is textbook wisdom that Proto-Italic and Proto-Celtic both had initial stress and indeed may well have inherited it from a common ancestor – which, I guess, may have borrowed it from some kind of Vasconic sub- or adstrate. Proto-Germanic presumably borrowed it from Celtic (like so many other things!) after Verner’s law had run its course.

    It occurs to me that I’ve read that original initial stress is found or inferred in so many IE languages that PIE (or was it just Greek?) may have had a light initial stress in addition to its pitch accent. Can someone clarify what I vaguely remember, and/or advise as to the current status of this hypothesis?

    Never heard of it, but it strikes me as unnecessary: once a more complex stress system is lost, consistent initial stress is one of the obvious possibilities, and globally a common one.

    Sanskrit or Homeric Greek, both of which inherited the original stress system of Indo-European

    Their stress systems only line up in about 3/4 of the words. This is part of the reason why, as I explained about a month ago, some people think PIE had phonemic tones, and sequences of tones evolved into placements of stress according to different rules in Homeric Greek and Vedic Sanskrit as well as into the pitch accent* system of Proto-Balto-Slavic. This view has made it into Wikipedia.
    (Classical Sanskrit has predictable stress.)
    * meaning what phonologists, as opposed to IEists, mean by “pitch accent”: a system where the stressed syllable of a word can have one of several tones, while the pitches of the unstressed syllables are predictable. Found in Swedish, Norwegian and some Danish dialects, in Ancient Greek (long vowels only), in Latvian, Lithuanian, Serbocroatian and Slovene, as well as in Shanghainese and (marginally) Japanese and maybe still some kinds of Korean.

    Attila too, as a Hungarian word

    Didn’t exist before 19th-century nationalism.

    Some of the population might have still spoken Avar language,

    Unlikely. As soon as their empire was gone, the Avars “disappeared like the Avars” (згинули какъ обры), as the medieval Russian chronicles say of steppe peoples that disappeared without a trace.

    thought to be an Oghur Turkic language closely related to Hunnic.

    There were West Turkic names in the Avar and the Hunnic empires, but how much that really amounted to is not clear.

    Attila shows up in the Ni[...]belungenlied as Etzel, with both umlaut and the second sound shift.

    Yep.

    And Theodoric of Verona shows up as Dietrich von Bern.

    Yep, with second sound shift at all four opportunities: [θ] to [d̥], [ð] to [t], [kʰ] to [xː], [v] to [b̥].

    But there it is, and the Church has used that version for two millenia as perhaps a sign that modern theorists of translation do not have a monopoly on wisdom.

    No, that sounds like the sheer weight of tradition to me. In German, the Lord’s Prayer starts with Vater unser, the only case in the entire language where anything however vaguely adjective-like comes behind the noun it modifies, and the Apostolic Creed is a long line of sentence fragments that don’t form sentences – many people probably think that’s a deliberate, if unusually far-reaching, adaptation to mindless/meditative chanting in choir, but it’s simply the word order of the original.

  151. borrowings from Low German
    A German can hardly go to sea without Low Germanizing himself (Selbstplattung?) But I admit that the Standard German story is nothing like as big as the Danish story, where you can’t talk about anything without Low German borrowings.
    and more recently English
    Consider Keks, Konzern, Streik, Tipp, Scheck, Stopp, Rekord, Standard, all 19th century and all assimilated phonetically and orthographically (except the last). A hundred years ago, Germans consistently wrote Klub, Klan, Kode, Hurrikan rather than the angliform spellings they use now, and Handicap seemingly took over from Handikap only in the mid-1990s.
    Vater unser
    In Old English too it was Fæder ure, but the tradition was then broken, and in Middle English it was retranslated as Our fadir and remains so. The Apostle’s Creed is a little odd in English too, but nothing like the German: “the third day he rose again from the dead” is syntactically unnatural — I’d expect on at the beginning in ModE. And speaking of the Creed, what gives with /ð/ in father, and likewise in mother, hither, gather, together? Grimm’s law all over again, or a countermove to /ð/ > /d/ in the rest of Germanic? And why so sporadic? Brother might account for the first two, but hardly the rest.

  152. meaning what phonologists, as opposed to IEists, mean by “pitch accent”
    David, the definition you give is the only one I know. What is this special Indo-Europeanist sense?

  153. marie-lucie says:

    JC: the third day he rose again
    This is the French order: le troisième jour, no need for a preposition.
    father, together, etc
    I think this is just intervocalic devoicing of the fricatives, as in bath/bathe, breath/breathe or leaf/leaves, life/lives.

  154. Marie-Lucie: No, the trouble is that those words should have /d/ rather than /ð/: ME fader, moder, hider, gader, togeder (but not brother, which has been fricative since OE days), and similarly Modern Dutch vader, moeder. Apparently it’s a sporadic sound-change that happened around the beginning of Early Modern English as far as the written record goes, though it may of course have been earlier in speech.

  155. marie-lucie says:

    JC, could it be a matter of dialect difference?

  156. Bill Walderman says:

    Thanks for correcting me on syllabic consonants in Slavic.

  157. Bill Walderman says:

    Is the American English phoneme in words such as “perfect”, “first”, “heard”, “further” analyzed as a syllabic consonant?

  158. Bill Walderman says:

    John Wells answered my question in his blog (21 Dec 2011):
    “In terms of phonology, I would say that syllabic consonants are not phonemes, i.e. not part of our underlying sound system. Rather, they are derived by rule from an underlying string of ə plus a non-syllabic sonorant consonant. I call the rule Syllabic Consonant Formation, and it takes the general form
    “ə [+son] → [+syll] / …
    “Two segments are reduced to one, with the sonorant consonant retaining its various attributes (place, nasality/laterality, etc) as it acquires syllabicity.
    “The conditioning environment of the rule (shown here just as “…”) is pretty complex. It varies according to different accents and different speaking styles, and also depending on which consonant is concerned. . . .
    “Although the AmE NURSE vowel could in principle be analysed as a strong (= stressable) syllabic r̩, this would not fit the above rule, which requires a weak ə as part of the input. So I treat the NURSE vowel in both BrE and AmE as a primitive, ɜː ~ ɝː. The second vowel of AmE father, however, does fit, and I analyse it accordingly: ˈfɑːðər → ˈfɑːðɚ.”

  159. As soon as their empire was gone, the Avars “disappeared like the Avars” (згинули какъ обры), as the medieval Russian chronicles say of steppe peoples that disappeared without a trace.
    Interesting—according to Vasmer, Old Russian обьринъ ‘Avar’ is cognate with a bunch of words meaning ‘giant’: Slovenian obər, Czech obr, Slovak obor, Old Polish obrzym (modern Polish olbrzym), and Sorbian hobr.
    John Wells answered my question in his blog
    Well, if you believe in things like [+son] → [+syll], that is.

  160. David Marjanović says:

    A German can hardly go to sea without Low Germanizing himself (Selbstplattung?)

    :-D True! All German sea vocabulary is a mixture of Low German, English (Reling…?), Dutch and who knows what else. That had to be expected from geography.
    (Selbstplättung sounds better, though that would immediately be understood as “flattening oneself”.)
    As far as English borrowings are concerned, Scheck and especially Standard could just as easily be from French… but I do admit it’s suspicious that Scheck has gone completely native orthographically, and that Standard is stressed on the first syllable.

    A hundred years ago, Germans consistently wrote [...] Klan, Kode, Hurrikan rather than the angliform spellings they use now

    Yes; what’s going on here is that German spellings suggest German pronunciation, which isn’t used. Only Klub has gone native, with /ʊ/, and therefore often keeps its k.
    Pronunciation doesn’t matter in the case of Clan, but it’s such a technical term that there’s little incentive to nativize its spelling.

    the third day he rose again

    Interestingly not kept in German: am dritten Tage auferstanden von den Toten / aufgefahren in den Himmel / er sitzt zur Rechten Gottes / des allmächtigen Vaters / von dort wird er kommen / zu richten / die Lebenden und die Toten.

    what gives with /ð/ in father, and likewise in mother, hither, gather, together?

    Dialect mixture. Some Early Modern English dialect turned /d/ between vowels into /ð/. Some other dialect did the opposite. Shakespeare wrote murtherere, right?

    What is this special Indo-Europeanist sense?

    Simply that the stressed syllable has (much) higher pitch than the others but is not (appreciably) louder than them. Not really something remarkable.
    It gets really funny when IEists write Wikipedia articles that link to the article on “pitch accent”, because that article only describes the other sense!

    John Wells answered my question in his blog (21 Dec 2011):

    I thought your question was whether syllabic consonants existed in English, not whether they were separate phonemes.
    And while Wells’s answer works (unsurprisingly enough) for English, it doesn’t for PIE, where there was no ə; instead, zero-grade syllables lacked a vowel, and the next most vowel-like sound ended up as the syllable nucleus.

    Well, if you believe in things like [+son] → [+syll], that is.

    As far as I can tell, “ə [+son] → [+syll]” just means “ə plus a sonorant* becomes a syllabic sonorant**”. It’s not a different theory, it’s just a fancy abbreviated notation.
    * which has the feature “[+son]”
    ** which has the feature “[+syll]”

  161. I know, but the very sight of such notation gets my hackles up. Having to take a theoretical-ling course to get my degree felt like a Soviet student having to take a course in diamat.

  162. Shakespeare wrote murtherere, right?
    He did, and murther, murtherer are the native English words, though now lost. Despite appearances, Modern English murder and friends are borrowings from Anglo-French and mediaeval Latin, which picked them up from some Continental Germanic language where /ð/ > /d/ already. No doubt the existence of murther reinforced murder, as did the Anglo-Latin legal term murdrum for a fine payable by the hundred (a collection of villages) when a stranger was found dead (later, only if it could not be proved that he was English rather than French).

  163. SFReader says:

    I always thought that Slavic obres are somehow related to Western European ogres.
    Of course, the obres are derived from Avars and ogres from Hungarians and the latter were successors of the former as we just proven…

  164. marie-lucie says:

    DM: “ə [+son] → [+syll]” just means “ə plus a sonorant* becomes a syllabic sonorant…”. It’s not a different theory, it’s just a fancy abbreviated notation.
    I agree. So many of these abbreviated formulas do not clarify anything, they are just shorthand for what can be said in a longer sentence, like writing a + b = c instead of saying it aloud.
    LH: I know, but the very sight of such notation gets my hackles up
    Me too! I gave up these notations years ago when I tried to use them to express the French rule changing i to the sound y (as in English you) before a vowel (as in French Marie + suffix -on Marion, pronounced Maryon). This is a very simple change when explained in words, but in feature notation just about every feature of /i/ had to be changed to make it /y/, so each side of the formula was a column of feature words, making the formula much harder to decode than an explanatory sentence.

  165. marie-lucie says:

    (There should be a rightward arrow (>) before Marion).

  166. Trond Engen says:

    Hat: Is it tron or trun?
    That’s a good question, and the answer is “yes”. When I lived near Oslo until I was 13 I was called trun, but I had family elsewhere who used tron. Then I moved to Bergen and became tron with uvular r. Where I live now, back east but farther from Oslo, I’m tron with apical r, but at home I’m trun. And it’s all OK. Actually, the only pronunciation I find a little jarring is trun with a narrow u, as if it were long. That’s from people with natural tron who try to accomodate me.

  167. What an interesting answer—I’m glad I asked!

  168. David Marjanović says:

    Impressive. What do you mean by “narrow u”?

    I agree. So many of these abbreviated formulas do not clarify anything, they are just shorthand for what can be said in a longer sentence, like writing a + b = c instead of saying it aloud.

    But it looks so much more sciency and mathematical and logical! Argh.

  169. David Marjanović says:

    What an interesting answer

    That’s what I mean by “impressive” when it’s still “early” in the “morning”… :-]

  170. Trond Engen says:

    Sensible followup question. I was busy, and I still am, but briefly:
    This depends on the phonemization of Norwegian, but I understand that the most common way to describe it is:
    ubort”away”, hvor “where”, vond “painful”
    u: bord “table”, mor “mother”, von “hope”
    o kort “short”, hånd “hand”, vott “mitten”, vått “wet (n.)”
    o: kår “choire”, hån “chair”, våt “wet (m./f.)”
    The long versions of the vowels are generally denser, narrower than the short, making it equally possible to follow historical phonology and treat it as
    ubort”away”, hvor “where”, vond “painful”
    o: bord “table”, mor “mother”, von “hope”
    o kort “short”, hånd “hand”, vott “mitten”, vått “wet (n.)”
    å(:) kår “choire”, hån “chair”, våt “wet (m./f.)”
    (Norwegian orthography isn’t particularly helpful.)
    My name can be pronounced with u or o according to either phonemization. What sounds wrong is a short version of u: (in the former) or o: (in the latter). It’s not that it bothers me in any way, but it sounds phonologically wrong. It would be equally jarring in vond. `Well, maybe not, since I wouldn’t hear it in isolation.

  171. Trond Engen says:

    I’ve finally worked my way down the thread.
    I agree about [+syll] etc. Until now I thought it was a dark secret, revealing my truly shallow understanding of phonology.
    Is it just that German maritime vocabulary is Low German, or do Germans have a register of (inherited or acquired) regionalisms for holiday purposes? To the former I’ll say that also Norwegian maritime vocabulary is a mash of Dutch, English and Low German. To the latter I can say that this is how I first read John’s comment. No doubt because of something I’d just been thinking about. My daughter asked me this weekend why I say spise about eating at home and ete in a tent.

  172. Is it just that German maritime vocabulary is Low German?
    Yes, for (as David says) obvious geographical reasons. People all around the North and Baltic Seas passed their professional and personal vocabulary to each other like a venereal disease. The result is that if we only had the modern forms of the Germanic languages, we’d see well enough that they were related, but we’d never be able to work out just how. (Ringe and the Fuluffyans actually did test runs with the same software they used to work out the IE tree, and got a complete hairball.)
    Hat, David, Marie-Lucie: There actually was a time before algebraic notation, y’know, and it was painful. “The eighth part of the first quantity, multiplied by twice the second quantity, plus five, equals zero, and nine times the first quantity minus five equals ten.” Painful.
    I agree that rigid feature notation can be a nuisance, but /i/ > /j/ is surely just [-syll]?

  173. marie-lucie says:

    JC, I am not arguing about the desirability and usefulness (or not) of abstract algebraic notation, but of phonological “feature matrices” (and those features as the basic units of phonology, rather than phonemes).
    /i/ > /j/ is surely just [-syll]?
    You would think so, but at the time (decades ago so I don’t remember the exact formula) I was frustrated because almost all the features were different in the two columns for /i/ and /j/.

  174. marie-lucie says:

    Trond: I say spise about eating at home and ete in a tent
    Probably because the conditions are different: at home you probably sit at a table and eat from a plate with knife and fork, while in a tent you probably eat mostly with your hands in whatever position feels least uncomfortable. Perhaps there is a difference as with dine and eat in English: you don’t usually dine while camping in a tent.
    What about animals eating?

  175. Trond Engen says:

    Animals eter. That’s what I told my daughter when she asked. But that’s not it. It’s me switching to a more folksy register in that situation. I know I also say skau for forest rather than skog in such situations.

  176. Is German Speise unrelated to English spice? One of the countless questions I’ve never bothered to inquire about?

  177. marie-lucie says:

    Trond, I did not mean that you were identifying yourself as an animal, but ete must be older and more fundamental, and spise a more fancy or “genteel” word introduced into the language at a more recent time (like English dine).

  178. Empty: Nope. Spice < Old French espice < Latin species ‘goods, spices’ < ‘kind, sort’; but Speise < Old High German spisa < Latin expensa ‘measured’.

  179. Trond Engen says:

    Well, I do eat like an animal, but how should you know? You’re about right, but spise is the everyday word, not fancy at all.
    Is it clear that spise is from expensa rather than espice?

  180. Well, I don’t have a German-specific etymological dictionary, but Wiktionary thinks so.

  181. marie-lucie says:

    Trond, even if spise is an everyday word in present-day Norwegian, it does not mean that it has always been so. In North America a diner is a rather low class restaurant, but the verb to dine is a borrowing from French, and like most such borrowings it was originally considered high class.

  182. Indeed, diner was originally short for dining car (on a train), where the food was high-class indeed.

  183. Dinner in the diner,
    Nothing could be finer,
    Than to have your ham and eggs in Carolina.

  184. marie-lucie says:

    JC, diner from dining car
    How the mighty have fallen!

Speak Your Mind

*