Juglandine Linguistics.

Brian Wallheimer reports for Phys.org on what I must consider a dubious hypothesis:

Purdue University research shows that ancient languages match up with the genetic codes found in Persian walnut (Juglans regia) forests, suggesting that the stands of trees seen today may be remnants of the first planned afforestation known in the world.

In a paper published in the journal PLoS One, Keith Woeste, a research geneticist for the U.S. Department of Agriculture’s Forest Service and a Purdue adjunct assistant professor of forestry, found that the evolution of language and spread of walnut forests overlapped over wide swaths of Asia over thousands of years. He believes as traders traversed the Silk Roads, connecting Eastern Europe and Africa with far-East Asia, they purposely planted walnut forests as a long-term agricultural investment.

The paper is “Ancient Humans Influenced the Current Spatial Genetic Structure of Common Walnut Populations in Asia,” by Paola Pollegioni, Keith E. Woeste, Francesca Chiocchini, Stefano Del Lungo, Irene Olimpieri, Virginia Tortolano, Jo Clark, Gabriel E. Hemery, Sergio Mapelli, and Maria Emilia Malvolti (not, n.b., by Woeste alone). One thing that gives me pause is that none of the authors has any connection to linguistics. But I do love walnuts, and I couldn’t resist using the word “juglandine” (which I created on the basis of Latin juglans, jugland– ‘walnut (tree)’ before discovering that it actually exists — but only, so far as I can tell, as a noun meaning “An alkaloid found in walnut leaves,” and thus I am staking my claim to it as an adjective), so here it is for your delectation. Shell and enjoy. (Thanks again go to Trevor for the link.)

Comments

  1. The first sentence makes the claim seem even more insane than it is. I was half expecting to read about someone trying to translate DNA sequences into strings of phonemes.

  2. @Keith Ivey

    Same here.

  3. Ecaxtly. Had to reread twice, thinking at first that plant DNA encoding human languages must be a hoax

  4. Lucy Kemnitzer says

    But there’s nothing very radical about the contention that people planted walnut trees along the Silk Road. We have lots of ancient tree planting and forest grooming all over the world. The walnut DNA being distributed in swathes instead of concentrically seems like sound evidence to me. The thing that bothers me is the abstract doesn’t talk about the supposed time frames.

    I have no way of judging the linguistic evidence. It’s not necessary, in my mind, to support the idea that people planted walnuts along the SIlk Road a long time ago. If unrelated languages along the Silk Road have similar worlds for walnut, that could either have happened in ancient times like they seem to be suggesting, or it could have happened in more recent times. How would you evaluate the evidence for that?

  5. What they seem to observe is that walnuts in China vs. Central Asia vs. Middle East and the Caucasus form distinct genetic clusters – which the geography would explain nicely – but straight geographic distances “as the crow flies” don’t explain all the variation (a pattern surely familiar to us from the earlier language-diffusion non-lingusts’ papers) and there is residual correlation with the migrations of peoples (but when the terrain has obstacles to moving of peoples – such as deserts, seas, or high mountain ranges – then these are also obstacles for dissemination of plants, so there should be some correlation… ). They cite earlier papers claiming that maize and sorghum genetic variation also correlated with ethnic substrates.

    Interestingly, walnuts can be grafted but apparently even in China, which is considered the place where grafting originated, walnuts tend to be grown from seeds! Sounds like the humans didn’t see much need for walnut selection?

  6. Lucy Kemnitzer says

    I wish I could edit the earlier comment-another article on the Phys.org site quotes Indiana experts worrying about “warmer,drier weather” doing the walnut trees in. But walnut trees live wild in the valleys of California, where we have definitely warmer and drier weather than they have in Indiana. I imagine what they really mean is that the walnut industry might have a hard time sustaining itself-I’m sure the trees are more productive with summer water.

  7. So, did some nucular equivalent of Johnny Appleseed plant walnut trees all along the Silk Roads? Given a peripatetic disposition and a few decades to work with, a single juglandomaniac might well have sufficed.

    Juglans, juglandis looks like it may be related to glans, glandis, ‘acorn’ and then ‘glans penis’ and then ‘gland’ – the last two seem to refer more to the shape of the acorn cap than the acorn nut itself. (As I recall, the first organ called a gland was the adrenal gland, which is shaped rather like an acorn cap perched atop the acorn-kidney.)

    If the two are related, the adjective from juglans should surely be ‘juglandular’. If this form is new, and catches on, perhaps I’ll get a brief credit in the next revision of the OED.

  8. @Lucy, your wild californian walnut trees are Juglans californica. Related, but not the same species.

  9. Juglans, juglandis looks like it may be related to glans, glandis, ‘acorn’
    Accoding to de Vaans etymological dictionary, it is related:
    Lat. iūglāns is probably a calque on Greek Διὸς βάλανος ‘chestnut’, with the gen.sg. *djowes(>Iovis), or with iū- taken from Jupiter.

  10. Lat. iūglāns is probably a calque on Greek Διὸς βάλανος ‘chestnut’, with the gen.sg. *djowes(>Iovis), or with iū- taken from Jupiter.
    Though the elements composing each are cognate, and there’s a further wonderful cognate in dialectal Armenian տկողին tkołin ‘hazelnut’.

  11. This is not the first of these kinds of analyses – there was a paper earlier tis year, also in PLoS One, about the distribution of the boab tree in Northern Australia, and the genetics correlating with linguistic groups in the area.

    Here it is: http://journals.plos.org/plosone/article?id=10.1371/journal.pone.0119758

  12. @ Alex Fink: Very intersting; Martirosyan (s.v. kaɫin) agrees with that analysis. So this may be at least a common formation by some Southern European IE languages.

  13. Could someone who understands the etymology of tkołin ‘hazelnut’ explain it for us non-Armenophones?

  14. This is what Martirosyan has:

    I assume that the form reflects PArm. *tukaɫin > *tukuɫin (vocalic assimilation, on which see 2.1.26.4) and can be derived from QIE *diuos *gʷlh2-eno- ‘divine acorn’, cf. Gr. *διϝός βάλανος ‘chestnut’ and Lat. iūglāns ‘walnut’ (on which see Walde/Hofmann 1, 1938: 727; Schrijver 1991: 273). On *tu/tw-, see HAB s.v. tiw ‘day’; see also s.v. ciacan ‘rainbow’. As is pointed out by Laufer 1919: 369, 3691, the pattern of Gr. Διός βάλανος “acorn of Zeus” is comparable to that of Pers. šāh-bal(l)ūt ‘the edible chestnut’ < “acorn of the Shah, royal acorn”; cf. Pahl. šāh-balūt ‘id.’, Arm. šahpalut ‘id.’, an Iranian loan, Łarabaɫ šmbálut‘ ‘chestnut’ [Hübschmann 1897: 272; HAB 3: 486a].

    On kaɫin “acorn”:

    Since Ayvazovsk‘i, Pictet, et al., connected with Gr. βάλανος f. ‘acorn’, Lat. glāns, glandis f. ‘acorn, beach-nut; missile discharged from a sling’, Russ. žëlud’, SCr. žȅlūd ‘acorn’, Lith. gìlė, dial. gylė̃ ‘acorn’, Latv. zĩle ‘acorn’, etc. [HAB 2:495-496]

  15. One thing I don’t understand is how we get the i in kaɫin from a pre-form *gʷlh2-eno-, as the laryngeal should have coloured *e > a, while Arm. i normally goes back to PIE *i, i:, e:.

  16. Thanks, that’s very helpful! (I reformatted the quotes for greater readability.)

  17. This all seems rather, ahem, nutty.

  18. Ow!

  19. Quick, someone send the ambulance to Chez Hat! Our host is suffering from a surfeit of punitive action.

  20. “If Persian walnut trees could talk…”

    I don’t know how anyone could resist an article that begins like that.

  21. George Gibbard says

    It’s disappointing the article doesn’t seem to include primary linguistic data, i.e. the various local names for ‘walnut’. It looks like they instead compared the genetic family tree for the walnut populations with Ethnologue’s and Ruhlen’s linguistic trees. This is remarkable since one would expect walnuts (and their names) to be borrowable, so I guess the finding is that they aren’t. So we are apparently to assume, for example, that as Turkic populations spread they brought their own walnuts, refusing to buy walnuts from trees planted by members of other linguistic stocks. Maybe I should have realized that this is what they were doing from your description, but I didn’t.

    Here is what I’ve gleaned about names for walnuts. ‘Walnut’ in Germanic is ‘Roman nut’ (or, in Britain, OE wealh-hnutu might be translated ‘Welsh nut’). In Russian, ‘walnut’ is often орех ‘nut (par excellence)’, but can be гре́цкий оре́х ‘Greek nut’. Both of these usages suggest walnuts were introduced to northern Europe in fairly recent times. The classical terms seem not to have stuck in Romance and Greek: in Romance, the Latin nux seems to mean the walnut by default, and Modern Greek καρύδι from the diminutive of Ancient Greek κάρυον ‘nut’ is the wiktionary translation of both ‘walnut’ and ‘nut’ (likewise Albanian karrë). The Arabic collective noun ɟawz- (singulative ɟawzat-) is said by wiktionary to derive from Middle Persian gōz, and Aramaic gawzā and Hebrew ʔĕgōz are also considered to be of Persian origin.

    (In vernacular Arabic varieties, the reflex of ɟawz- ‘walnuts’ can also mean ‘husband’ (Classical Arabic zawɟ-, thought to be from Greek ζέυγος ‘couple’); likewise the reflex of ɟawzat- ‘walnut’ can mean ‘wife’.)

    For what it’s worth the common walnut, alias English walnut, is also called the Persian walnut.

    Wiktionary deserts us, unfortunately, for ‘walnut’ in Central Asian languages (also Caucasian languages). I note that Urdu has [ɐχɾoˑʈ]. This word is a bit odd in that [χ] normally occurs in Persian or Arabic words, while [ʈ] normally occurs in native Indo-Aryan words. Does anyone know the history of this word? Could it be Pashto?

  22. George Gibbard says

    I neglected to mention that Turkish of Turkey has ceviz, from Arabic.

  23. Well, Platts says axroṭ is for Skt. akṣoṭa, for what it’s worth.

  24. Wiktionary deserts us, unfortunately, for ‘walnut’ in Central Asian languages (also Caucasian languages).

    Uzbek for walnut is yong’oq/ёнғоқ, Kirgiz, жаңгак, Kazakh, жаңғақ, and Turkmen, хоз (хо:з). (Kirgiz ж is like an English j, and Kazakh ж is like a French j).
    Chechen for nut is бӀар and for walnut бочабӀар.
    Finally, the Japanese word is kurumi:

    くるみ【胡桃】 ローマ(kurumi)
    【植】 〔クルミ科クルミ属 (Juglans) のうち, 食用になるものの総称〕 a walnut (tree); 〔その実〕 a walnut.
    ▲クルミの殻 a walnut shell.
    ▲クルミを割る crack a walnut
    ・クルミを拾いに行く go gathering walnuts.
    __むきグルミ a skinned walnut.
    __クルミ科 the walnut family; Juglandaceae. △juglandaceous adj.
    クルミ油 walnut [nut] oil.
    クルミ割り 《a pair of》 nutcrackers.

  25. And the Pashto word is غوز (ғуз).

  26. The modern Persian word is گردو gerdu; they say هر گردى گردو نيست har gerdi gerdu nist, ‘not every round thing is a walnut’ (i.e., all that glitters is not gold), and I would guess that it is in fact originally ’round thing’ (and thus from the root گرد gard– ‘to turn around’). I don’t know when it replaced the native goz, from which the Turkmen хоз is presumably borrowed (or whether the Pashto ғуз is borrowed or cognate).

    Mallory and Adams, in The Oxford Introduction to Proto-Indo-European and the Proto-Indo-European World (p. 161), say:

    From the West Central region comes […] *h2er– ‘nut’ (e.g.Lith ruošutỹs ‘nut’, Rus orékh ‘nut’, Alb arrë ‘walnut, nut tree’, dialectal Grk árua ‘nut’) perhaps Proto-Indo-European if Hit harau– ‘poplar’ is cognate but the Hittite meaning is certainly distant […]

    For the Baltic words, Vasmer says “Ср. лит. ríešutas ‘орех’, ríešas — то же, riešutỹs, вост.-лит. ruošutỹs (с уменьш. -ut- аналогично ноготь), лтш. riẽksts — то же, др.-прусск. bucca-reisis ‘буковый орешек.”

  27. Here is what I’ve gleaned about names for walnuts. ‘Walnut’ in Germanic is ‘Roman nut’ (or, in Britain, OE wealh-hnutu might be translated ‘Welsh nut’).

    Which is of course misleading. Germanic *walxaz (transferred from the ethnonym of the Celtic Volcae before the operation of Grimm’s Law) originally referred to all Celts, continental or insular, and then, by extension, to all non-Germanic westerners and southerners (Celtic or Romance-speakers). Wealh mostly referred to the Brittonic Celts in Anglo-Saxon England, but the compound corresponding to OE wealhhnutu is the same all over the place: OIc. valhnot, German Walnuss, etc. — all of them probably translating Late Latin *nux gallica ‘Gaulish nut’, cf. Old French nois gauge.

    Middle High German Walch and the associated adjective welsch meant French/Italian, since there were no continental Celts left. In modern usage, welsch has come to be regarded as pejorative and politically incorrect, but it survives in compounds like Walnuss and Welschkohl ‘Savoy cabbage’. Speakers of Proto-Slavic borrowed walch from their West Germanic neighbours and applied them to whichever Romance nation they were familiar with. Thus the descendants of early Slavic *volxъ came to mean ‘Italian’ in Polish and Czech, for example, but ‘Vlach/Romanian’ among the Eastern and Southern Slavs (volox and vlah). Therefore, the fact that Polish orzech włoski and Czech vlašský ořech (calqued from German) SEEM to mean ‘Italian nut’ is the result of a series of historical misunderstandings.

    https://en.wikipedia.org/wiki/Walhaz

  28. originally referred to all Celts, continental or insular and then, by extension, to all non-Germanic westerners and southerners

    Is there actual evidence for this particular line of semantic development? My understanding is that it meant ‘Romans’, whether Romance-speaking or Celtic-speaking, almost as soon as it was generalized.

    Two nice Hungarian instances of the root are given in the article you linked: Olaszország ‘Italy’, Oláhok (obs.) ‘Wallachian’.

  29. Hebrew ʔĕgōz

    Which in modern Hebrew is just “nut”; “walnut” is egoz melekh “king nut” (presumably a calque on the Persian šāh-bal(l)ūt mentioned by Martirosyan in Hans’s quote above).

    So Zeus’s acorn was a chestnut for the Greeks, a walnut for the Romans, and a hazelnut for the Armenians? Interesting. Is there any particular reason why the god was associated with any of these nuts?

  30. Is there actual evidence for this particular line of semantic development? My understanding is that it meant ‘Romans’, whether Romance-speaking or Celtic-speaking, almost as soon as it was generalized.

    The name does seem to derive from Celtic *wolko- borrowed some time in the mid-first millennium BC. At that time the (pre-Proto-)Germani and the Romans were not yet aware of each other’s existence (there are no Latin loans showing the effects of Grimm’s Law), but contacts between the Celts and the ancestral Germani were regular and intensive (there are early loans to demonstrate it). After the Romanisation of Gaul, however, the British Isles became the last stronghold of the Celts. In the 5th/6th centuries the invading Ingvaeonic tribes in Britain, unlike most other Germani at the time, could meet real living Celts at close quarters, and they reserved the word Wealas for them, never applying it to the “Romans” (including the Romanised Franks in Gaul). They also distinguished the Old Brittonic language (wylisċ ~ welisċ) from Latin (læden; note the Brittonic phonetics). Only in some compounds does wealh convey a meaning like ‘Gaulish, Mediterranean’ or more generally ‘foreign’ (wealhhnutu, wealmoru ‘parsnip’).

  31. the Ingvaeonic tribes in Britain, unlike most other Germani at the time, could meet real living Celts at close quarters, and they reserved the word Wealas for them

    Quite so. But after four centuries, the Britons surely were pretty well Romanized in all but language, and even to some extent in language. So that is not really evidence that it meant ‘Celts’ at that time.

  32. I don’t think that is a reasonable assumption. They were at the outer fringe of the empire and lost their romanitas very quickly once the troops left; it seems to me more likely that they were only superficially Romanized, and that mainly in urban contexts.

  33. David Marjanović says

    In the 5th/6th centuries the invading Ingvaeonic tribes in Britain, unlike most other Germani at the time, could meet real living Celts at close quarters, and they reserved the word Wealas for them, never applying it to the “Romans” (including the Romanised Franks in Gaul). They also distinguished the Old Brittonic language (wylisċ ~ welisċ) from Latin (læden; note the Brittonic phonetics).

    But how could they have remembered that Wealas referred to Celts while they were holed up in Jutland for hundreds of years, and how could they have recognized the Celts in Britain as Wealas?

    It seems more likely that Wealas would have been applied to any people encountered in Britain, and then narrowed down to “Briton” that was eventually distinguished from “Frenchman”. If læden has Brittonic phonetics, then most likely it was borrowed in place to refer to the liturgical language which the invaders hadn’t been aware of back home, not the “Welsh” language the “Welsh” people actually spoke.

    Somewhat similarly, welsch in German used to refer variably to French and/or Italian (still Welschschweiz = Suisse romande), but never to Spanish or Romanian for example.

    …That’s interesting about “Gaulish nut”, BTW. Walnuss is regularly explained as welsche Nuss, but the implication is always that it’s Italian, not French.

  34. But how could they have remembered that Wealas referred to Celts while they were holed up in Jutland for hundreds of years, and how could they have recognized the Celts in Britain as Wealas?

    The North Sea Germani were pretty mobile and very much aware of what was happening around them. Some of them apparently served as mercenaries in Roman armies before Gaulish died out completely (well, Celtic armies too if Vortigern really invited the Jutes); they certainly accompanied the Franks in their migrations and pirate raids (just as some Franks accompanied them during the conquest of Britain). The southeastern coast of Roman Britannia wasn’t called Litus Saxonicus for nothing. One of the oldest Anglo-Frisian runic objects found in England, the Undley Common bracteate, features a Roman emperor’s head and the Lupa Capitolina suckling Romulus and Remus. Even if it represents some kind of cargo cult, the people who owned such objets de luxe surely knew who the Romans were; they were familiar with their trade goods as well as ideas, symbols and institutions. I am not saying thet they had a linguistic definition of “Celtic” (or “Germanic”, for that matter — the East Germani were regarded as more Scythian than Germanic). For example, the term Wealas was not applied to Gaelic-speakers. I’d rather suggest that they were aware that the Britittonic Celts were kinda sorta Gaulish.

  35. David: “Gaulish nut”, BTW. Walnuss is regularly explained as welsche Nuss, but the implication is always that it’s Italian, not French.

    Northern Italy was Gallia cisalpina for a long time.

  36. it seems to me more likely that they were only superficially Romanized, and that mainly in urban contexts.

    The Welsh language shows such profound Latin influence, including on the verb system, as to suggest nearly universal bilingualism verging on imminent replacement.

    The southeastern coast of Roman Britannia wasn’t called Litus Saxonicus for nothing.

    Um, Saxonicum.

  37. The Welsh language shows such profound Latin influence, including on the verb system, as to suggest nearly universal bilingualism verging on imminent replacement.

    Huh. OK, I stand corrected.

  38. There is an entry for walnut in an etymological dictionary of Ossetic, which can be downloaded at iratta.com (The word is ængūz, V.1, p.157.)

  39. Can you summarize the etymology they give for those of us too lazy to download the whole dictionary?

  40. Daghestani languages
    Avar: цІулакьо
    Aknadi: цIулакьа
    Bagvalal: цIулакьа
    Bezhta: ГЬЁЛ1Е
    Hinux: вáхъи
    Godoberi: цIулакъа
    Dargi: хив
    Tindi: цIулакьа
    Tsez: ваIхъар

  41. Um, Saxonicum

    [blush]

  42. The Welsh language shows such profound Latin influence, including on the verb system, as to suggest nearly universal bilingualism verging on imminent replacement.

    That’s right. Nevertheless, when the British Celts were left to their own resources, their vernacular languages survived, whereas Insular Latin didn’t (by contrast to the situation in Gaul and many other former Roman provinces). The Latinisation of Old Brittonic could be compared to that of Proto-Albanian, which also survived the Great Migrations and the Slavic invasion of the Balkans (but so did Balkan Latin).

  43. I entirely agree with Rodger C., the influence of Latin upon Welsh (and Breton, and Cornish) is of such a nature that we must assume mass bilingualism, which in turn implies the existence of a large number of Latin L1 speakers in Roman Britain, who may indeed have made up a majority of the population of Roman Britain at some point.

    The belief that Roman Britain only had a thin veneer of Latin and of Roman culture was to a great degree due to a famous Celtic scholar’s claim that Latin loanwords in Welsh/Breton/Cornish point to Latin in Roman Britain as having been a school/learned language only. This claim has been (quite convincingly, to my mind) refuted by a Latin scholar, but the Celtic scholar’s original claim had unfortunately remained unchallenged for so long that his views on the place of Latin in Roman Britain became (and remain) the conventional wisdom among all too many specialists, especially non-linguists.

  44. ængūz | ængozæ
    ~ The closest links are with Armenian əngoyz, ənkoyz and Georgian nigozi; then cf. Farsi gauz, gōz, gūz, Kurdish əgviz, guviz, gujz, Shugni γōz, Pashto γuz, Ishkashmi oγuzo, Turkish koz, Ancient Hebrew egōz. The urform is goz, koz. The initial n- (Ossetic æn-, Armenian ən-, Georgian ni-) is an augment that arose in the Caucasus. Armenian has a form without the augment: guz.

  45. Thanks!

  46. For example, the term Wealas was not applied to Gaelic-speakers.

    Indeed, precisely because they were not and never had been Roman, unlike the British. All this about mass Latinization in Britain simply confirms me in my views: *Walhaz meant ‘Roman’ (in culture) everywhere, not one thing here and another there.

  47. An inconvenient truth: the Anglo-Saxon invasion saved the Welsh language.

  48. According to Leslau’s Ge’ez dictionary, the Ge’ez gawz is a borrowing from Aramaic gawzā, which also entered Tigrinya. From Ethiopic Semitic it further passed into Cushitic. In modern Amharic the word is used also for ‘coconut’. Leslau says the word (which appears once in the Song of Songs and in numerous places in the Mishna) is from Persian, but he tantalizingly refers to an article by Rabin which claims a Hittite origin. I’ll take a look next time I’m at my local big library, unless someone beats me to it (the article, in Hebrew, is in Sefer Segal [1964], p.151).

  49. J. W. Brewer says

    Was the cultural Roman-ness of the Wallachians/Vlachs blatantly obvious to whoever (don’t know if it was the medieval German-speaking minority in Transylvania or someone else) first gave them that name? I’m sure modern Romanian nationalists would like to think so, but . . .

  50. The Welsh language shows such profound Latin influence, including on the verb system, as to suggest nearly universal bilingualism verging on imminent replacement.

    I know there are hundreds of Latin loanwords in Welsh, including basic vocabulary items; what are the specific Latin influences on the grammar? I’ve found this article tracing the Welsh pluperfect to Latin influence; is there more?

  51. Was the cultural Roman-ness of the Wallachians/Vlachs blatantly obvious to whoever (don’t know if it was the medieval German-speaking minority in Transylvania or someone else) first gave them that name?

    The Slavs applied it to “foreigners” inhabiting the Balkans, and Byzantine scholars, in turn, adopted the term (Βλάχοι) to refer to Romance-speakers in the region. It’s clear that the Slavs must have borrowed *walxa- from West Germanic before the breakup of Common Slavic unity, since the *-al- sequence (Slavic *-ol-) underwent the normal “liquid metathesis & pleophony” changes characteristic of various Slavic dialectal areas: it became -olo- in East Slavic, -lo- in Northwest Slavic, and la in Czech, Slovak and South Slavic. These changes took place more or less 750-800 AD, which is the terminus ante quem for the borrowing.

  52. @TR: The pluperfect is what I was thinking of. I learned this tidbit from the Cantos of Mvtabilitie blog a number of years ago. (Okay, 13 Dec. 2008.)

  53. J. W. Brewer says

    So in other words the Greek-speaking Romans (Rhomaioi) used “Blachoi” as a name for a bunch of illiterate hillbillies who they would have thought marginal at best to their own Roman culture? (I’m I think plausibly assuming little audience for the maximalist claims of much later Romanian nationalists in 8th century Constantinople.) I guess that’s semantic drift in action . . .

  54. @J. W. Brewer

    In Byzantine times it took no linguistic genius to recognise Balkan Latin as folk Latin of sorts. The stereotyped early story told by Theophylactus in the 7th century and repeated by Theophanes Confessor in the 9th has one local mule driver yelling to another “in the language of their country” to make him turn back: Torna, torna, (fratre!)

    It surely took a lot of drifting for a ethnonym first picked up by strangers before 300 BC, somewhere in Central Europe, to be applied to the Gauls, the Welsh, the Walloons, the Romans, the French, the Italians, and finally the Romanians, by (in turn) the Proto-Germanic speakers, the Anglo-Saxons, the Vikings, the Germans, the Dutch, the Slavs, the Hungarians, and the Byzantine Greeks. Meanwhile the original Volcae quietly left the scene in the aftermath of the Gallic Wars and were never heard of again.

  55. I forgot to list the Franks, responsible for renaming GalliaGaule (*Walxa).

  56. One thing I don’t understand is how we get the i in kaɫin from a pre-form *gʷlh2-eno-, as the laryngeal should have coloured *e > a, while Arm. i normally goes back to PIE *i, i:, e:.

    These reconstructions are rather impressionistic and should be made more precise. In Latin, there is some evidence of early word-initial syncope between an obstruent and a liquid. The protoform of Lat. gland- may therefore be an exact cognate of the Slavic ‘acorn’ word, *gʷelh₂-end(ʰ)(i)- (the suffix is probably complex and requires more work; I’m not sure what to make of it). The Greek word is a reflex of *gʷĺ̥h₂-no-, a related word with a different apophonic grade and a simpler suffix. If the primary form was a nasal stem, perhaps the vocalism of the Armenian word reflects a hysterokinetic nominative, *gʷl̥h₂-ḗn (no laryngeal colour because of the vowel lengthening).

  57. I wanted to post some references to the discussion of Italic syncope by Brent Vine and myself; the server didn’t like them and blocked my message. Let me try again:

    http://tinyurl.com/qdvoafz
    http://tinyurl.com/oa3yayo (page 60)

  58. David Marjanović says

    It’s clear that the Slavs must have borrowed *walxa- from West Germanic

    Why West and not East a bit earlier?

  59. The word is not directly attested in East Germanic. Admittedly, this may well be an accidental attestation gap, given the imperfect documentation of Gothic (we don’t have Gothic *kunings either, but the Slavic word for ‘ruler’ is widely regarded as an East Germanic loan anyway).

  60. David Marjanović says

    the discussion of Italic syncope by Brent Vine

    Ooh, nice. I particularly enjoyed finding out that cry/crier is from shouting “O Quirites”, and crone is cognate with carrion. 😀

  61. I posted about the former eight years ago.

  62. @Piotr: *gʷl̥h₂-ḗn would indeed account for the Armenian form.
    NB that Martirosyan also notes dialectal forms with a suffix -nt- / -nd-:

    For Svedia, Ačaṙyan (HAB, ibid.) cites gaɫɛɔn (read /gaɫön/). Later, he (2003: 573) also records gyäɫɛɔnd, noting (p. 378) that the change ClArm. i > Svedia ɛɔ is irregular in this position. This form with epithetic -d is corroborated by other authors: gäɫɛnd or gäɫund (see Andreasyan 1967: 36 and 367a, respectively), gäɫɛnd [Hananyan 1995: 53, 187b]; K‘esab käɫɛnt vs. käɫɛn [Č‘olak‘ean 1986: 206b]. The final d is exceptional since the other examples of the epenthetic or epithetic d (Ačaṙyan 2003: 431; see also Hananyan 1995: 53) apply to specific conditions: -nr > -ndr and -s(-) > -sd(-).

    It is tempting to identify the final dental stop of Arm. Svedia gyäɫɛɔnd (on which see above) with that seen in Russ. žëlud’, SCr. žȅlūd, etc., and Lat. glāns, glandis. Alternatively, one might assume a contamination with Svedia hɛɔnd from Arm. (h)und ‘edible seed, grain’ (q.v.), although this seems less probable.

    Looking at the /t/ in forms like käɫɛnt and if (as it seems) Svedia has voiced the unvoiced stops, Armenian would be a further witness for the -nd suffix and confirm the stop as *-d-, not *-dh-.

  63. An inconvenient truth: the Anglo-Saxon invasion saved the Welsh language.

    With Roman rule gone and Brittonic still widely spoken, was it really certain that it would be replaced by Latin even in the less Romanized parts of the island?

  64. Good point.

Speak Your Mind

*