The Language of Wool.

I’m reading Leskov’s «Некрещеный поп» (The Unbaptized Priest, translated by James Y. Muckle as “The Priest Who Was Never Baptized”), which uses a great many dialectal and Ukrainian words, and one of them is the extremely interesting волна ‘lamb’s wool.’ One point that struck me is the uncertainty as to stress; Dahl, and following him Vasmer, give initial stress, vólna, and so does my 1908 11th edition of Makaroff’s Russian-French dictionary, but both my Словарь ударений [Dictionary of stresses] (1984) and my three-volume Russian-English dictionary (1997) give it with final stress (volná). The Wiktionary entry gives initial stress (вóлна, вóлны)… but the audio file has end stress! Clearly it historically has initial stress, but being a dialectal word and probably obsolete to boot, my guess is that it’s been assimilated to the far more common end-stressed волна ‘wave.’

It’s the etymology that’s worth noting, however; it’s a basic Indo-European word, given by Wiktionary as *h₂wĺ̥h₁neh₂ and widespread in the major branches: Hittite ḫulanaš, Old Armenian gełmn, Ancient Greek λῆνος, Latin lāna, and many descendants in Balto-Slavic, Celtic, Germanic, and Indo-Iranian. Slavic vьlna survives as the basic word for ‘wool’ in all branches but Russian, where for some reason it’s been replaced by шерсть (and here one feels the lack of a Russian equivalent of the OED). One of the Germanic descendants is, of course, English wool, so I looked that up in the AHD… and was astonished to find for an etymology only “[Middle English wolle, from Old English wull.]” Surely they didn’t dispute the Indo-European origin of that word? I checked the Indo-European Roots Appendix and found that there was nothing corresponding to *h₂wĺ̥h₁neh₂. I looked up words like vellus (“[Latin, wool.]”) and lanugo (“[Middle English, pith, from Latin lānūgō, down, from lāna, wool.]”); there was no indication that they were in any way related. The whole Indo-European cluster appears to have been inadvertently ignored and omitted; you’d think by the fifth edition they might have noticed!

An interesting factor in the Leskov story is that the brutal Dukach, disliked by everyone in the village and therefore unable to find godparents for his newborn son there, tells the equally disliked Kerasivna (the villagers think she’s a witch) to take him to a nearby village to have him baptized, adding that she should make sure the priest there doesn’t spoil the boy by naming him Ivan or Nikola. She says of course she won’t allow a Christian boy to be called by a “Moscow name” like Nikola, and Dukach agrees: “Никола самый москаль.” He winds up being called Savva. I hadn’t realized there was such a sharp geographical division of acceptable names (at least in the late 1820s, when this is supposed to have occurred).

Comments

  1. John Cowan says:

    It might be class rather than geographical: perhaps Nikola is a townsman’s name. Cf. Kevinismus.

  2. Neither московськое nor москаль refer to class; they are purely geographical (literally “Muscovite,” but the latter is a Ukrainian insult for Russians in general).

  3. OK, I’ve finished the story, and a passage towards the end provides more of an explanation: Kerasivna says that Dukach told her that Nikola was the first Muscovite and he won’t help the Cossacks but will do everything in the Muscovite way (Никола и есть самый первый москаль, и он нам, казакам, ни в чем не помогает, а все на московськую руку тянет), whereas Saint Savka is a Cossack himself and will be very good to them (Этот из казаков и до нас дуже добрый). I’m not sure which St. Nikola(i) is meant, but at least the idea is clear.

  4. John Cowan says:

    they are purely geographical

    Sure. But who lives in Moscow? The bourgeoisie.

  5. There was barely any bourgeoisie anywhere in Russia in the 1820s, and it certainly wouldn’t have been on the radar of village Cossacks.

  6. Stu Clayton says:

    Did Cossacks even have radar technology in 1820 ? It’s possible, I suppose, since there were airports during the Revolutionary War, a few decades earlier.

  7. They cleverly stole it from some visiting Germans.

  8. Moskal, of course, means Russian, not necessarily living in town. But occasionally it might have meant (by now, it is just an ethnic slur) a soldier (in tzar’s army), apparently the most known type of Russian outside Russia. Here’s to the hope of it remaining firmly in the past (Soviet joke: How different nationals travel abroad? German in a car, French by train, Englishman by plane, Russian … ¯\_(ツ)_/¯… in a tank?)

    Modern (standard?) Ukrainian word for “wool” is vovna and for “wave” it’s khvyla. The latter comes from the same Slavic word, but takes detour through Polish fala. By some mysterious circumhappenstance the same word (in the form khvylyna) also means “minute”.

  9. I’m not sure which St. Nikola(i) is meant

    Nicholas Wandermaker, who else?

    I am not sure whether Mykola is less of a Cossack name then Nikolai is a Muscovite one, but Ivan is obviously is no less Ukrainian than Russian. Dukach is not the first nationalist who cannot keep his facts straight.

  10. Nicholas Wandermaker, who else?

    I thought of him, but what makes him самый москаль?

    I am not sure whether Mykola is less of a Cossack name then Nikolai is a Muscovite one, but Ivan is obviously is no less Ukrainian than Russian.

    Yes, I was puzzled by that.

  11. Weren’t shtetls just burgs filled with bourgeoisie? Livestock traders, leather dealers, makers of crafted items,, metal fabricators …

    Perhaps the ethnic/religious differences so overwhelm other distinctions that it was rarely conceived in those terms. Or is so little of the Pale Russia proper that the statement still holds?

  12. Translation:

    Santa Claus is Putin himself, doesn’t help us Ukrainians at all, but favors Russians only….

  13. About that one. It should be noted that tzar at the moment was a very unsaintlike Nicolas. But that clearly was not a consideration for Dukach. As for Mr. Pu, he doesn’t seem to believe that Ukrainians exist at all.

  14. SFReader says:

    Leskov is making Ukrainian character to issue ridiculously absurd statement for amusement of his Russian readers.

    Mocking Ukrainian nationalism as it were.

    For Leskov’s readers Ukrainian nationalism seemed a delightfully absurd thing.

    A kind of what a Victorian would have thought about modern day Jacobites – ideology so old fashioned, so 18th century that only some really backward types in the wilds of Scottish Highlands could still believe in it…

  15. PlasticPaddy says:

    Polish Wikipedia says

    Nazwa współcześnie bywa przywoływana w języku rosyjskim, ukraińskim i białoruskim jako umotywowany politycznie pejoratywny archaizm i polonizm, obraźliwe określenie Rosjan[70].

    So in Ukrainian moskal is archaism/polonism.

    I remember the term used in poczatek (“The beautiful Mrs. Seidenmann” in english). In one episode, an ethnic German rescues Mrs. Seidenmann, whom he has never met and who claims to be Mrs. Gostomka, a polish officer’s widow, from the Gestapo in Warsaw. Afterwards he says something to the effect that the moskal would have brought out another woman and then taken him for questioning when he recognised her as Mrs Gostomka.

  16. In Bulgarian “вълна” is stressed on the first syllable in the “wool” meaning, and on the second syllable in the “wave” meaning; also, the first is uncountable while second is countable, thus:

    вЪлна, вЪлната : wool
    вълнА, вълнАта, вълнИ, вълнИте : wavе (singular, singular definite, plural, plural definite.)

    There’s also the derived adjective вЪлнен — woollen (made of wool), and, most importantly, a _lot_ of adjectives, verbs and participles derived from the “wave” meaning: вълнУвам (се), развълнЕн, вълнУващ etc. : excited, exciting, upset, driven, uplifted, facsinating…

  17. Leskov is making Ukrainian character to issue ridiculously absurd statement for amusement of his Russian readers.

    Huh, I didn’t even think of that. Sadly plausible.

  18. V : Thanks for the Bulgarian information!

  19. Nikolai for Saint Nicholas of Myra of course (i.e., the original Santa Claus).

    He is a profoundly important saint in the Orthodox Church.

  20. David Marjanović says:

    It only just occurred to me that “wave” in German is Welle, which looks like it ought to be related to Wolle but really can’t be.

  21. Nikolai for Saint Nicholas of Myra of course (i.e., the original Santa Claus).

    He is a profoundly important saint in the Orthodox Church.

    Yes, that was D.O.’s suggestion, and I realize he’s a profoundly important saint in the Orthodox Church, but I repeat my earlier question: what makes him самый москаль?

  22. Although D.O.’s reminder that it was the name of the current tsar may explain it.

  23. PlasticPaddy says:

    I think it may go back further. It seems Nicholas was perceived to have helped Muscovite princes acquire and defend their territory. See e.g., here:
    https://www.rbth.com/special_projects/discovering_russia_1/2017/06/09/the-mozhaisk-st-nicholas-cathedral-from-prokudin-gorsky-to-the-present_779855

  24. Aha, very likely — an excellent find!

  25. Ксёнѕ Фаўст says:

    I’ve read somewhere that St. Nicholas was more celebrated among Russians while St. George was more popular among Ukrainians. Some even think this reflects the relative popularity of Veles and Perun in pre-Christian times (the East Slavic version of St. Nicholas took many Veles’s characteristics, and St. George took Perun’s).

    There’s a Ukrainian anecdote about a peasant who misunderstood what they said in a village that ‘God has died’ (it was the Good Friday). When he went back home his family began wondering about who shall be the next God. His mother suggested the Mother of God should be God now. But he protested that it was not a woman’s business to be God. His younger brother said it should be St. Nicholas as the eldest. Everybody shuddered at the suggestion because St. Nicholas is a moskal’ and he would do good to moskali only. Finally his older brother suggested that St. George should be God now and everybody was pleased by that, as he’s ‘one of their own’.

  26. Fascinating!

  27. According to Boris Uspensky, cult of St. Nicholas replaced cult of Veles, Slavic god of earth, waters, and the underworld.

    He was venerated in northern Russian lands while the southern Ukrainian lands tended to favor St. George, whose cult replaced older cult of Perun, Slavic god of thunder.

    It was believed by Ukrainians that St. George is Ukrainian and St. Nicholas is Russian (Muscovite).

    Uspensky says this Ukrainian-Russian opposition of St. George and St. Nicholas reflects conflict of Perun and Veles in Slavic pagan religion which, in turn, is a legacy of an older Indo-European myth of a hero slaying dragon or serpent.

  28. This is very surprising. St. George is the patron saint of Moscow. Moscow founder (sorta) was Yuri. Main military decoration was the Cross of St.George. I am raking my brain trying to find anything special about St.George veneration in Ukraine and find blanks (this one shouldn’t count for much). Maybe Russian elevation of St.George was an elite project and was buoyed up as Russia’s military might increased. St.Nicolas is certainly venerated in Russia the most after Christ and Mary (I would even say that Mary comes before Jesus), that part of Ксёнѕ Фаўст’s anecdote is very plausible, but George and Ukraine?

  29. SFReader says:

    Moscow founder (sorta) was Yuri

    The founder of Moscow prince Yuri Dolgoruki died in Kiev in 1157 as a Grand Prince of Kiev.
    His father, prince Vladimir Monomakh died in Kiev in 1125 as a Grand Prince of Kiev.
    His grandfather, prince Vsevolod Yaroslavich died in Kiev in 1093 as a Grand Prince of Kiev.
    His great-grandfather, prince Yaroslav the Wise died in Kiev in 1054 as a Grand Prince of Kiev.
    His great-great-grandfather, prince Vladimir the Great died in Kiev in 1015 as a Grand Prince of Kiev.
    His great-great-great-grandfather prince Svyatoslav died in battle on the Dnieper river in 972, needless to say, he was Grand Prince of Kiev too.
    His great-great-great-great-grandfather prince Igor the Old was slain by rebel tribes in 945 and he was the first historical Grand Prince of Kiev.

    No wonder Moscow got patron saint from Kiev too..

  30. BTW, in this genealogical line Yuri is the first one commonly known by his Christian name. Vladimir Svyatoslavich was christened as Vasily, Yaroslav Vladimirovich was christened as George, Vsevolod Yaroslavich was Andrei, Vladimir Vsevolodovich was Vasili. I have never heard about Yuri’s pagan name and cannot quickly find whether he had one.

  31. Trond Engen says:

    This was a time of onomastic flux. The brother of Great Prince Yuri Vladimirovich, Great Prince Mstislav Vladimirovich, appears prominently in the Norse sagas as Haraldr. It seems that not only did he get the Russian name Mstislav and the Christian name Fjodor but a Germanic one, this one after his maternal grandfather, king Harold Godwinson of England. Great Prince Mstislav was married to Kristina, a daughter of king Inge of Sweden. Their two eldest daughters are known by Norse names, Ingiborg and Malmfrid, and were married to king Knut of Denmark and king Sigurd of Norway respectively. The daughter who married the Byzantine emperor Alexios Komnenos is known as Eupraxia.

    Anyway, Norse language and culture must have been even more important at the Kievan court at his time than before or later, since other princes were known by adapted forms of their Russian names — Jarisleiv, Visevald and Valdemar. I’ll rush to add that Scandinavian royals for generations had married into Slavic dynasties not only in Kievan Rus but along the Baltic coast, so there may have been bilinguality on both sides.

  32. I don’t know whether this was an issue at the time, but the name Yaroslav has been sometimes reinterpreted as a Germanic name meaning “lord of Slavs.” I have a Ukrainian colleague by that name, for which that was the only meaning he learned for his name growing up.

  33. David Marjanović says:

    With jarl? That’s not a Germanic word order, though.

  34. not a Germanic word order

    Which one?

    Andlát Egils Skalla-Grímssonar
    https://bookdown.org/eythorbj/Egils_saga/kafli-andlat-egils-skalla-grimssonar-.html

  35. PlasticPaddy says:

    Jarl does for compound names as first element but I think that jaroslav is a stretch. Compare kvetoslav, which is clearly slavic (unless you seriously think the first element is something like hvit).

  36. It only just occurred to me that “wave” in German is Welle.

    I always kind of assumed Polish fala was borrowed from Germanic. Not the case?

  37. Yup, it’s from Welle.

  38. “St. George, whose cult replaced older cult of Perun, Slavic god of thunder. ”

    Another saint who replaced Perun is the Old Testament prophet St Elijah. Sv. Ilija is the patron saint of my town in Croatia.

  39. I’m surprised nobody mentioned “волнуха” – woolly milkcap.

  40. Ксёнѕ Фаўст says:

    Yup, it’s from Welle.

    But the v → f sound change is unusual. Is there a German dialect which devoices here?

    Another saint who replaced Perun is the Old Testament prophet St Elijah. Sv. Ilija is the patron saint of my town in Croatia.

    Yes. In fact the Slavic St. George is a probably a mix of Perun, Jarilo-Jarovit and original St George (who is probably a Christian projection of the IE and Semitic storm god myth anyway). In Poland it seems to me that Perun was partly merged with St Peter (and partly remained in a demonized form as Pieron, Piorun and in the common noun piorun ‘thunderbolt’) and Jarovit with St John and St George (both are lunar heroes killing snakes/dragons in Polish sayings and songs).

    This is very surprising. St. George is the patron saint of Moscow. Moscow founder (sorta) was Yuri. Main military decoration was the Cross of St.George. I am raking my brain trying to find anything special about St.George veneration in Ukraine and find blanks (this one shouldn’t count for much). Maybe Russian elevation of St.George was an elite project and was buoyed up as Russia’s military might increased.

    I’m by no means sure but this sounds very likely. St Nicholas as the primary patron of the peasant masses, merchants and so on and St George’s cult secondarily spreading in the military, which contained plenty of Cossacks from the south. Note how it was the Cossacks who conquered Siberia for Russia.

    Uspensky says this Ukrainian-Russian opposition of St. George and St. Nicholas reflects conflict of Perun and Veles in Slavic pagan religion which, in turn, is a legacy of an older Indo-European myth of a hero slaying dragon or serpent.

    Ah, so it was him. I didn’t remember in detail.

  41. There were quite a few Ioann’s in the Left Bank Cossack branch of my ancestors in the 1700s. Not a single Nikolay / Mykola, nor Savva. The names were of course largely drawn from the Orthodox Calendar, but there were enough calendric Saints to avoid picking a name the parents wouldn’t like.

    Various saint Savva’s are assigned to over 30 days in the Orthodox calendar. Some of these days would also have a Saint John, some also a Saint Nicholas, and perhaps only one day would honor all 3 names. By looking it in the calendar, you should be able to figure out which Savva and which Nicholas were actually “in the game”. But I’m sure Leskov just wanted to illustrate the preposterous views of the uneducated Ukrainians, nothing more.

    As to the St George the patron of Moscow, he only officially appeared on the coat of arms in the 1500s. Graced the coins a bit earlier (where the Dragon-Slayer design has been inherited from the Novgorod coins of the 1400s). The first mentions of the Cossacks as of Christian-aligned irregular forces of the borderlands are also from the 1400s, but the mass escapes of peasants of Muscovy to the Cossackdom didn’t seem to have started until about 1600. So there doesn’t seem to be any reason to believe that St. George’s becoming the patron saint of Moscow, vs. St. George becoming popular among the Cossacks, had any influence on each other. It must be just parallel developments, of adopting the “armed forces saint” by entities too fond of their military might.

  42. John Cowan says:

    There’s a Ukrainian Catholic cathedral of St. George in Lviv, which is effectively the capital of Western Ukraine. Until 2005 it was the mother church of Ukrainian Catholics, though that role has now been moved to a newly built cathedral in Kyiv, the Cathedral of the Resurrection of Christ. There is a Cathedral of St. Nicholas in Kyiv too, though currently secularized.

    My local Ukrainian Catholic church is St. George’s, on the corner of East 7th St. and Taras Shevchenko Place (a short north-south street between 2nd Ave. and 3rd Ave. that connects East 6th with East 7th).

  43. David Marjanović says:

    But the v → f sound change is unusual. Is there a German dialect which devoices here?

    Not that I know of.

    Maybe it’s a hypercorrectivism? After all, the weird e → a change that seems downright regular for German borrowings in Polish makes sense as undoing real and imagined umlaut…

  44. There is a Cathedral of St. Nicholas in Kyiv too, though currently secularized

    I used to walk by nightly, returning from the dance hall to my hotel room. It is a beautiful music hall now, but it used to be the cathedral of the Polish community (not an Eastern Rite Catholic church).

    A whole slew of the Eastern rite Ukrainian Catholic churches in the Americas are consecrated in the name of St. Nicholas though, including their principal cathedral in Chicago. I am positive that it had nothing to do whatsoever with the Cossacks’ prejudices, as the Cossacks wouldn’t touch anything Catholic with a six foot pole. St Nicholas is popular among the emigrants because he’s also the protector of all those traveling away from home / across the seas.

    For exactly the same reason St Nicholas is the favorite consecration of a church among the seafaring Pomors of Russia’s North, where a folk saying goes, от Холмогор до Колы 33 Николы ( ~~ You’ll find 33 St Nicholas churches along the Pomor Coast).

  45. Ксёнѕ Фаўст says:

    Maybe it’s a hypercorrectivism? After all, the weird e → a change that seems downright regular for German borrowings in Polish makes sense as undoing real and imagined umlaut…

    Maybe. Or it mixed with chwila (itself an old loan), also fila in Old Polish (the group chw dialectally simplified to /f/). Chwila could once mean ‘weather’, in Belarusian хвіля can mean ‘storm’ and in Ukrainian хвиля is ‘wave’. I’ve had the impression that in many of the e → a words there’s a colouring effect of the German /r/ (warsztat, rachować) even if it doesn’t cover all instances. Huta, sztuka could be an argument in favour of the umlaut-undoing hypothesis but I’m wondering whether this isn’t just something dialectal in nature.

    The imagined-feature idea made me think about many Germanic place names in -(e)n failing to decline in Polish. It looks as if people thought they’re all plurals. So you say we Frankfurcie but w Essen, w Liverpoolu but w Southampton (it’s less consistent with English names and there may be vacillation).

  46. David Marjanović says:

    Huta, sztuka could be an argument in favour of the umlaut-undoing hypothesis but I’m wondering whether this isn’t just something dialectal in nature.

    Shouldn’t happen in these words. The number of umlaut-blocking factors increases from north to south, and indeed my dialect lacks umlaut in such words as drücken, hüpfen, Brücke* – but not in Hütte or Stück, and how words from that far south would end up in Polish isn’t clear to me either.

    * That one may not count: it may not be an exact cognate, but a *n- stem instead of a *ja- stem, the *-j- being the umlaut trigger. Indeed, from the 18th century the masculine *n- stem accusative einen Brucken is attested. Similar things may hold for some other cases.

  47. PlasticPaddy says:

    Welle and Quelle(n) are seemingly unrelated. Is chwila a borrowing (or cognate) of quelle(n)?

  48. David Marjanović says:

    Can’t be a cognate (Slavic /x/ comes from PIE */s/ under certain conditions, and from PIE */kχ/, not from PIE */g/), and Polish doesn’t have a problem with word-initial [kv], so there’s no motivation to borrow it as [xv].

  49. Trond Engen says:

    What about an older borrowing from Gmc. *xwīlō-, the source of ON hvílr “rest”, Eng. while “period of time”, Ger. Weile?

  50. PlasticPaddy says:

    @dm
    What Trond said.
    Why Welle > chwila and then (later) Welle > Fala? Hypercorrection later ok but why not Welle > wila/vila earlier?

  51. Trond Engen says:

    That was supposed to be Gmc. *xʷīlō-

    And I see that I didn’t read the thread properly, because an old loan is exactly how it was intoduced to the discussion.

    @PlasticPaddy. I understand how a word with /w-/ might be contaminated with one with /hw-/. I have more trouble with the rest of the word. Duden online says:

    mittelhochdeutsch welle = Reisigbündel; zylindrischer Körper; Wasserwoge, althochdeutsch wella = Wasserwoge, zu mittelhochdeutsch wellen, althochdeutsch wellan = wälzen, zu wallen

    Wiktionary says Welle is from Gmc. *wallijō-, which would mean that it’s derived from the o-grade of the verb.

  52. David Marjanović says:

    Why Welle > chwila

    Probably chwila is straight from *xʷīlō, and then the meaning shifted from “time” to “weather” – compare temps in French –, then on to “storm”. How “wave” in Ukrainian comes into this is less evident, but perhaps by confusion with Polish fala?

  53. Ксёнѕ Фаўст says:

    Indeed, clearly chwila was borrowed when the h- was still around. Welle never had an h- from what I see. The meaning ‘wave’ could be from ‘storm, bad weather’ which cause waves. Apparently fala could mean ‘storm’, ‘torrential rain’ in the past. Kashubian has wała ‘wave; rain cloud’. Looks like the two meanings ‘waves’ and ‘bad weather’ were synced in the worldview of old.

  54. Russian волна – wave – is still connected to the word for disturbance. Волновать(ся) when applied to sea means “storm” and when applied to a person means “anxiety” (it’s a verb, but the picture is clear)

  55. Trond Engen says:

    I got the idea that ‘wool’ could be from the meaning “twist, turn” of the root of Welle.

    Looking up ull in Bjorvand & Lindeman, I find that they reconstruct Gmc *wullō- f. < *wulnō- < PIE *H₂₃wĺ̩Hno-. A stressed zero-grade and a suffix *no- could mean a nominalization of a verbal adjective. There are also forms like Latin vellus n. and Arm. gelmn that seem to go back to a full grade root H₂₃wel(H)-. The first laryngeal is necessary for the Anatolian forms but is difficult to square with Greek lḗnos without initial vowel. The second laryngeal is necessary for i.a. Lat. lāna. In Bjorvand 1994 it’s suggested that there may instead have been laryngeal metathesis in Anatolian (*H₂₃wĺ̩no- < *wĺ̩H₂₃no-).

    As for velle v., the Germanic strong verbs *wellan-/*wallan- and the weak causative *wallijan- all have *-ll-. B&L suggest analogical leveling after *-ll- > *-ln- in the present forms, but it could also be from the verbal noun, since the forms meaning “wave” have an n-suffix all over IE. The basic Gmc root *wel- can be united with I-Ir and B-Sl forms to PIE *wel-H- ~*wl̩H-.

    We see that with laryngeal metathesis in the “wool” word, the roots could be homonymous in PIE. Without metathesis, they could also be homonymous, since I don’t think there’s anything that prevents us from suggesting an initial laryngeal in the “Welle” word.

    And if the verbal roots are homonymous, they could also be the same, meaning “twist, turn”. The o-grade or full-grade adjective derivation would have meant something like “prone to twisting, turning” and gave the word for waves and turbulent water in most languages. The zero-grade adjective derivation would have meant “twisted, turned”. The nominalization could also mean “wave” but came to mean “wool” (<- “curly”) in most languages.

    Bonus: I think current opinion is that Germanic lost the laryngeals very late. We might ponder the possibility that the /x/ of Slavic xvila “wave” is a reflex of the Pre-Germanic laryngeal.

  56. David Marjanović says:

    *H₂₃wĺ̩Hno-

    Crazily, that would become *H₂₃wĺ̩no- already in PIE, giving us an alternation between *H₂₃wĺ̩Hneh₂- (preserved in Indic, Slavic, Italic and Germanic, and likely others) and *H₂₃wĺ̩no- (preserved in Anatolian and Greek).

    The missing initial laryngeal in Greek can’t be explained that way. Perhaps it’s a Crotonian loan…

    I think current opinion is that Germanic lost the laryngeals very late.

    Isn’t that just based on Cowgill’s law?

  57. Trond Engen says:

    David M: Crazily, that would become *H₂₃wĺ̩no- already in PIE

    Interesting. The Saussure effect didn’t make it into B&L’s discussion of the “wool” word. I guess it’s showing that it’s 20 years since the manuscript for the first edition was finished, and most entries must be several years older. OTOH, these are very old observations.

    The missing initial laryngeal in Greek can’t be explained that way. Perhaps it’s a Crotonian loan…

    Or perhaps it wasn’t there at all. I don’t think there’s any evidence for it outside Anatolian.

    I note with satisification that Andrew Bird quotes van Beek (2011) in deriving Gk. oúlos from PIE *wel- “twist, turn”, but with less satification the remark “Not *h₂wolh₁-“.

    Isn’t that just based on Cowgill’s law?

    Maybe. I didn’t mean into Grimmly times but at least long enough into the Pre-Grimm phase to play a role in some of the sound changes, but it may well be that the there are more satisfactory explications out there. I think Piotr had something to say about this, but I can’t find it now.

  58. David Marjanović says:

    I don’t think there’s any evidence for it outside Anatolian.

    Only Anatolian and Greek would preserve any evidence for it in the first place. Balto-Slavic would have the “acute” tone as evidence that there was at least one laryngeal somewhere in the syllable, but that wouldn’t tell us where (interestingly – I’m not sure what phonetic sense that could make).

  59. Trond Engen says:

    Only Anatolian and Greek would preserve any evidence for it in the first place.

    So I thought, but there’s a passage in B&L that reads to me as if the laryngeal is reconstructed without the Anatolian evidence, and I wouldn’t be too dogmatic about it.

    Balto-Slavic would have the “acute” tone as evidence that there was at least one laryngeal somewhere in the syllable, but that wouldn’t tell us where (interestingly – I’m not sure what phonetic sense that could make).

    Yes. Something something Winter’s law something. My impression is that everybody agrees that there’s something there but nobody agrees what.

  60. David Marjanović says:

    Word-initial laryngeals do show up in Sanskrit inside compound words, as unexpected vowel lengthening. I don’t know if there is one in this case.

    Reduplication has the same effect in potentially all IE branches: even without a theory of PIE root structure, we can tell the eat root was *h₁ed- and not just *ed- from the long vowel in ate < PGmc. *ēt- < PIE *h₁e-h₁d-. But I really don’t expect “wool” to form any reduplicated words.

    Something something Winter’s law something.

    That’s not Winter’s, which says that the “plain voiced” plosives lengthened immediately preceding vowels and is phonetically unremarkable both with and without a Leiden-type glottalist interpretation.

  61. Trond Engen says:

    Oh, I remembered it being tied to Winter’s law, but maybe that’s glottalism. I have read some things by Kortlandt on Balto-Slavic tone. But I’ve also read Miguel Carrasquer Vidal and Jay Jasanoff and several others, so at least I think my confusion is fairly broadly based and up to date.

  62. David Marjanović says:

    That’s a great way to broaden one’s confusion!

    I should have mentioned that Winter’s law entails not just lengthening, but also acuting; for Kortlandt it’s just a special case of laryngeals causing acute tone.

  63. Trond Engen says:

    Speaking of Carrasquer Vidal, here’s a short reply to Andrew Byrd’s categorization of final voicing as “crazy” in the paper above. Short: They aren’t voiced, they’re “unreleased, implosive, or pre-glottalized”, and that adds up with their origin as pronominal endings.

  64. David Marjanović says:

    Interesting, but highly speculative as usual, involving a lot of internal reconstruction. And, yes, explicitly glottalist.

    No mention of the voicing found within the separate history of Italic: Latin ab, sub = Greek apo, hypo = IIRC Sanskrit apa, upa indicate that Italic underwent a round of apocope and voiced the resulting word-final /p/. At that point it really should have been actual phonetic voicing. (…But maybe not, given this glottalist explanation of Lachmann’s law, which is the next level of crazy otherwise.)

    Word-final voicing of fricatives is not particularly crazy: Old Norse has done it across the board, English does it with plural -s, 3sg -s, of (thereby split off from off) and for some people apparently still if – that probably started as general voicing in the least stressed positions. Maybe PIE did the same and extended it to the few cases of word-final plosives.

    Finally, if the “plain voiced” plosives weren’t voiced, glottalism does not automatically follow. Central Bavarian has word-final (and -internal*) lenition of postvocalic /t/ and /tː/ to /d/**, with no voice, glottalizaton or aspiration anywhere in earshot. Take such a sound system, drape it over pretty much any substrate, and you might end up with final voicing.

    * The exception is that /tː/ is preserved after formerly long vowels and diphthongs. That’s right, overlong syllables are selectively preserved instead of, as is usual, eliminated with extreme prejudice – how’s that for a crazy rule!
    ** There may be one or two word-internal examples of /kː/ to /g/.

  65. Another mechanism to end up with final voicing in principle: levelling of sandhi, which we can see traces of in Latin in prefixation, but which likely would not have been indicated in writing otherwise.

    even without a theory of PIE root structure, we can tell the eat root was *h₁ed- and not just *ed- from the long vowel in ate < PGmc. *ēt- < PIE *h₁e-h₁d-

    I dunno, it would be easy to propose that that’s from *e-ed- or *e-əd- as derived sometime before zero grades were really zeros.

  66. David Marjanović says:

    sandhi

    Good old Vedic Sanskrit had utterance-final devoicing, but intervocalic voicing across word boundaries. *-s disappeared in the process, with compensatory lengthening plus the usual oddities in vowel quality, because [z] was banned at some point.

    But then, it remains unclear if PIE had vowel-initial words in the first place.

    I dunno, it would be easy to propose that that’s from *e-ed- or *e-əd- as derived sometime before zero grades were really zeros.

    That would seem to be difficult to reconcile with this hypothesis about the development of zero-grade between different meanings of PIE.

  67. John Cowan says:

    Latin ab, sub = Greek apo, hypo = IIRC Sanskrit apa, upa indicate that Italic underwent a round of apocope and voiced the resulting word-final /p/.

    That may be rather intervocalic voicing, since ab is a clitic. The story is quite complicated even descriptively; thus L & S (cit. om.):

    ap, af, ab (av), au-, ā, ă; aps, abs, as-.

    The existence of the oldest form, ap, is proved by the oldest and best MSS. analogous to the prep. apud, the Sanscr. api, and Gr. ἐπί, and by the weakened form af, which, by the rule of historical grammar and the nature of the Latin letter f, can be derived only from ap, not from ab. The form af, weakened from ap, also very soon became obsolete. There are but five examples of it in inscriptions at the end of the sixth and in the course of the seventh century B. C. In the time of Cicero this form was regarded as archaic, and only here and there used in account-books.

    The second form of this preposition, changed from ap, was ab, which has become the principal form and the one most generally used through all periods—and indeed the only one used before all vowels and h; here and there also before some consonants, particularly l, n, r, s; rarely before c, j, d, t; and almost never before the labials p, b, f, v, m, such examples as ab Massiliensibus being of the most rare occurrence.

    By changing the b of ab through v into u, the form au- originated, which was in use only in the two compounds aufero and aufugio for abfero, ab-fugio; aufuisse for afuisse is altogether unusual.

    Finally, by dropping the b of ab, and lengthening the a, ab was changed into ā, which form, together with ab, predominated through all periods of the Latin language, and took its place before all consonants in the later years of Cicero, and after him almost exclusively. By dropping the b without lengthening the a, ab occurs in the form ă- in the two compounds ă-bîo and ă-pĕrio, q. v.

    On the other hand, instead of reducing ap to ā and ă, a strengthened collateral form, aps, was made by adding to ap the letter s (also used in particles, as in ex, mox, vix). From the first, aps was used only before the letters c, q, t, and was very soon changed into abs (as ap into ab): abs chorago, abs quivis, abs terra, and in compounds: aps-cessero, abs-condo, abs-que, abs-tineo, etc.

    The use of abs was confined almost exclusively to the combination abs te during the whole ante-classic period, and with Cicero till about the year 700 A. U. C. (=B. C. 54). After that time Cicero evidently hesitates between abs te and a te, but during the last five or six years of his life a te became predominant in all his writings, even in his letters; consequently abs te appears but rarely in later authors.

    Finally abs, in consequence of the following p, lost its b, and became -ss- in the three compounds aspello, as-porto, as-pernor. The late Lat. verb abbrevio may stand for adbrevio, the d of ad being assimilated to the following b.

    (Whew! You can send me the coveted HTML award now, please.)

  68. David Marjanović says:

    Interesting. But I wonder if the very rare and very early form af actually represents [v], not [f].

    Lewis & Short on sub don’t mention any variants, other than saying that *subs must have existed because of the occurrence of sus- as a prefix.

    Also, I could have copied the evidence for final voicing from the “crazy” paper instead of trying to remember it:

    3.2.1 The evidence.
    Anatolian
    (Melchert 1994:85) :
    pa-i-ta-aš /páyd–as/ ‘went he’ [because otherwise we’d expect -it- instead of -i-]
    Hitt. -at(+V) = Luv. -ata = Lyc. -ede = Lyd. -ad ‘it’ < PIE *od
    Old Latin
    (Meillet-Vendryes 1968:146-7):
    quod, istud, ad, apud, feced < *-Vd
    ab, ob, sub < *apo, *opi, *supo

  69. John Cowan says:

    But I wonder if the very rare and very early form af actually represents [v], not [f].

    Unlikely. As far as I know, there was no [v] or other voiced fricatives in any Italic language until the 1C, long after af was permanently retired. There had been in Proto-Italic voiced fricatives (< PIE voiced aspirated stops), but they changed into /h/ < /x/ < /gʰ/ and /f/ in all other cases. Even ζ in early Greek loanwords became /sː/, as in massa ‘barley cake’ < μᾶζα.

    As for ob and sub, they were certainly clitics as well. But perhaps the true order was intervocalic devoicing followed by loss of the final vowel.

  70. January First-of-May says:

    as in massa ‘barley cake’ < μᾶζα

    I immediately had to google whether there is any relation to matzo (the answer is apparently “inconclusive, but maybe”).

  71. David Marjanović says:

    As far as I know, there was no [v] or other voiced fricatives in any Italic language until the 1C

    There’s pretty good evidence that all intervocalic fricatives were voiced except in Latin (and of course intervocalic /s/ was voiced in Latin too, before it merged into /r/ as it also did in Umbrian).

    For instance, there’s a town whose Latin name changed from Mefania to Mevania right around the 1C.

    ζ in early Greek loanwords

    wasn’t [z] yet, but [dz]; given the absence of [dz] and [ts] in Latin, /sː/ was the best possible representation of that.

Speak Your Mind

*