How the Corded Ware Culture Was Formed.

It’s been a couple of years since we got into the whole Indo-Europeans-and-Corded-Ware thing (e.g., here), so I thought I’d post Re-theorising mobility and the formation of culture and language among the Corded Ware Culture in Europe, by Kristian Kristiansen, Morten E. Allentoft, Karin M. Frei, Rune Iversen, Niels N. Johannsen, Guus Kroonen, Łukasz Pospieszny, T. Douglas Price, Simon Rasmussen, Karl-Göran Sjögren, Martin Sikora, and Eske Willerslev, from Antiquity 91. Here’s the Abstract:

Recent genetic, isotopic and linguistic research has dramatically changed our understanding of how the Corded Ware Culture in Europe was formed. Here the authors explain it in terms of local adaptations and interactions between migrant Yamnaya people from the Pontic-Caspian steppe and indigenous North European Neolithic cultures. The original herding economy of the Yamnaya migrants gradually gave way to new practices of crop cultivation, which led to the adoption of new words for those crops. The result of this hybridisation process was the formation of a new material culture, the Corded Ware Culture, and of a new dialect, Proto-Germanic. Despite a degree of hostility between expanding Corded Ware groups and indigenous Neolithic groups, stable isotope data suggest that exogamy provided a mechanism facilitating their integration. This article should be read in conjunction with that by Heyd (2017, in this issue).

And here’s an intriguing excerpt:

The new data conforms well to the reconstructed lexicon of Proto-Indo-European (Mallory & Adams 2006), which provides important clues that the subsistence strategy of early Indo-European-speaking societies was based on animal husbandry. It includes, for instance, terms related to dairy and wool production, horse breeding and wagon technology. Words for crops and land cultivation, however, have proved to be far more difficult to reconstruct. These results from historical linguistics are supported by similar evidence from archaeology (Andersen 1995; Kristiansen 2007). With the recent study by Kroonen and Iversen (in press), we can now demonstrate how social and economic interaction with existing Neolithic societies also had a corresponding linguistic imprint. This should not surprise us, as similar results are well documented from the interaction of Yamnaya societies with their northern Uralic-speaking neighbours (Parpola & Koskallio 2007).

Thanks, Trevor!

Comments

  1. Trond Engen says:

    Thanks for bringing it here. I read it several weeks ago, well before living memory, but I see that I noted:

    – The proposed system of “youthful warbands” of younger sons led by an older mentor spearheading settlement in new areas is supported by Baltic, Irish and Indo-Iranian folklore, but it seems like a good description also of Viking Age raids and migrations.

    – Nomadism isn’t only about exploiting seasonal differences in grazing conditions, it’s about adding value to goods by bringing it from one end of your range to the other. Yamnaya peoples were in a unique position to take advantage of the first waves of globalisation from the 4th millenium BCE. The paper suggests that one result of this globalisation was a first wave of the plague, weakening the less resistant settled populations in Western Eurasia.

  2. David Marjanović says:

    Open access!!!

  3. Bathrobe says:

    Most adult women (between 28 and 42 per cent) were of non-local origin and had a different diet during childhood.

    Is 28-42% “most”?

  4. Trond Engen says:

    Open access!!!

    Yes. But unfortunately only that article, not its companion article by Heyd, or any of the other interesting stuff.

  5. I am reading China Condensed, essentially a history of the past 5000 years written for young overseas Chinese (in English, I hasten to say) and published in Singapore. Among its very, very broad scale maps it includes one of the area under the Tang Dynasty (618-907 AD) which shows most of the country as Tang, the Tibetan region as Tubo, and the far northeast Pacific, north of Korea, 1000 ks or so north of “Silla” (Korea), as “Magyar”.

    I can find no reference to the Magyars originating further east than the current Altai area, several thousand kilometers further west than the map suggests. With Hat’s permission, can anyone elucidate, please ?

  6. Trond Engen says:

    Rereading the Wikipedia article on Corded Ware, I see that:

    The Corded Ware Culture also shows genetic affinity with the later Sintashta culture, where the proto-Indo-Iranian language originated.

    This is stated more clearly in the article on the Sintashta culture:

    Allentoft et al. (2015) found close autosomal genetic relationship between peoples of Corded Ware culture and Sintashta culture, which “suggests similar genetic sources of the two,” and may imply that “the Sintashta derives directly from an eastward migration of Corded Ware peoples.” Sintashta individuals and Corded Ware individuals both had a relatively higher ancestry proportion derived from the early farmers of Central Europe, and both differed markedly in such ancestry from the population of the Yamnaya Culture and most individuals of the Poltavka Culture that preceded Sintashta in the same geographic region.

    So “Russian” Corded Ware became a founding element of Shintashta and hence Indo-Iranian! If Shintashta really was Proto-Indo-Iranian, that is — models are changing fast now,

  7. Trond Engen says:

    Most adult women (between 28 and 42 per cent) were of non-local origin and had a different diet during childhood.

    I read right past that, interpreting it as “28-42% of the total sample (17-25 of 60 individuals of all genders and ages) were women who grew up elsewhere”. But it’s a strange way to say it, so I may have been wrong.

  8. SFReader says:

    Re: “Magyars” on Pacific.

    Most likely it referred to

    “The Mohe, Malgal, or Mogher were a Tungusic people in ancient Manchuria. ”
    https://en.wikipedia.org/wiki/Mohe_people

  9. Very plausible. Write the publisher to complain!

  10. SFReader, Hat : That seems plausible, though how it became Magyar… The publisher is the respected Straits Times newspaper group, so its strange.

  11. I don’t care how respected a publisher is, these days nobody’s perfect. And that’s not an error that leaps to the eye; “Magyar” looks normal, while Mohe, Malgal, and Mogher all look weird.

  12. SFReader says:

    I think this was the original map

    https://i.ytimg.com/vi/wzGuM1gwdVo/0.jpg

    I have a feeling that next generations of editors would replace Mohe with Mohican and wouldn’t even notice anything wrong

  13. Is 28-42% “most”?

    For me at least, while “the most things” can mean a plurality, “most things” must be a majority. The was a LLOG flurry about when and if it means a supermajority.

  14. David Marjanović says:

    “Most” is definitely some kind of editing error (authors half-reading their text a few times and changing it incompletely), and I agree that it is probably meant to refer to the total sample rather than just the women.

  15. Greg Pandatshang says:

    If the original Corded Ware language was Early Proto-Germanic (real early: like 2,000 years early), then why isn’t Macro-Germanic a more widespread family when it first comes into historical view? I guess one would assume that it once was widespread but was later swamped almost everywhere by subsequent Celtic and Balto-Slavic incursions. That’s fine, not implausible, it just requires an additional premise.

  16. David Marjanović says:

    Or it’s only late northern Corded Ware – which, it appears, the in-press paper by Kroonen & Iversen will be about – that’s specifically Pre-Proto-Germanic, and the rest was some kind of “Proto-West-IE” (common ancestry of Germanic and Italo-Celtic perhaps). The Bell Beaker culture has occasionally been associated with the origin of Celtic – though, AFAIK, not on more evidence than geography and age.

    Anyway, I’ve now read this paper and (through channels darkly) Heyd’s. Heyd basically provides an overview over the findings since 2013 and then warns against going too far and drawing too simple conclusions, before ending firmly on the optimistic side; some of the points he brings up are already mentioned by Kristiansen et omnes. An interesting fact Heyd brings up is the appearance of cultural commonalities (sandals or sandal soles of durable material in graves; anthropomorphic stelae) that show up all over Europe, from Spain to Yamnaya, a while before the latter’s expansion. On the genetic side, Heyd implicitly calls for more and larger samples, including of Bell Beaker people.

  17. David Marjanović says:

    Oh, considering Macro-Germanic, check this mystery out.

  18. Trond Engen says:

    I think the assumption is that the intruding language was ancestral to both Proto-Germanic and Proto-Balto-Slavic. At 2900 BC it must have been almost undifferentiated Core IE, at least in the mouth of the intruders, but Corded Ware spread so quickly over such a vast area that it can hardly have had a single language for long, if at all.

    I’d rather guess that the explosive spread of the superstrate culture was followed by a long period of amalgamation and language shift that panned out differently in different regions. The northeastern end may well have adopted Proto-Uralic. Or maybe satemization can be blamed on Uralic influence.

  19. Etienne says:

    1-The reference to the work of Rasmussen et al. 2015, on an early form of plague spreading from Siberia to the Baltic in the third millennium, is to my mind unsurprising: what is remarkable about the older Indo-European languages of Europe (attested or reconstructed) is how little substrate influence is actually found in them: some vocabulary chiefly relating to plants and animals. In nature this is not very different from the substrate influence of Native American languages upon transplanted Western European languages in the Americas. Since the latter owe their spread and reduced substrate influence to the catastrophic impact upon native Americans of imported Eurasian diseases, perhaps the spread of, and the weak substrate influence upon, Indo-European languages in Europe is due to the same factor…

    2-An additional factor may have been the adolescent warbands whose hunting and, more crucially, raiding activities took place at the (social and geographic) margins of various Indo-European societies: if these sociologically marginal groups were the ones who were most closely and regularly in contact with non-Indo-European societies, it is among such groups that I would expect to see words being borrowed, to see some contact language(s) involving Indo-European and non-Indo-European languages emerge (Pidgins? Mixed languages? Some xenolectal registers for speaking with outsiders? Perhaps all of the above, at different times and places…). If becoming an adult within an Indo-European community meant no longer being involved in these interactions with outsiders, perhaps it also meant shedding anything in one’s speech indicative of being an adolescent, including anything deriving from said contact language(s).

    3-Greg Pandatshang: actually, there is evidence that Baltic and Slavic were once in contact with an otherwise unattested branch of Indo-European which shared some features with Germanic, so the “additional premise” you point to has already been proposed in the scholarly literature (I can give the exact reference, should anyone request it downthread).

  20. Marja Erwin says:

    On Bastarnian as macro-Germanic:

    Is there any evidence that Bastarnian was only related to the Germanic languages, rather than one of the Germanic languages?

    It’s sometimes possible to identify west-Germanic usage in Gothic contexts. For example, one of Athanaric’s officers was named Lagarimanus. Ammianus book 31 chapter 3. In west Germanic, that can mean “Camp-person.” In Gothic, *Ligramanna would mean “Bed-person,” which is less likely.

    Would it be possible to identify near-Germanic in Gothic contexts, if Bastarnian were near-Germanic? e.g. words which defy Grimm’s law.

    Or to identify two branches of east Germanic in Gothic contexts?

  21. “On the genetic side, Heyd implicitly calls for more and larger samples, including of Bell Beaker people.”

    There are three new pre-print papers that have genomic data for hundreds of Mesolithic, Neolithic, and Bronze Age European samples. Including Bell Beakers and many groups from Southeast Europe.

    There is a lot of very interesting stuff in them. Like, the early Iberian Beaker people did not pass on any genes to the Beakers who spread out from around the Lower Rhine and then replaced >90% of the British population in only a few hundred years. The Eastern Beaker were closely related to Corded Ware, but very distinct in direct male ancestry.

    The Southeast European genomes have a number of surprises, which make many proposed routes of Indo-European more complicated.

    And an abstract for the first of the South Asian ancient genomics papers is out, but so far with almost no details.

    The Beaker Phenomenon And The Genomic Transformation Of Northwest Europe

    http://biorxiv.org/content/early/2017/05/09/135962

    The Genomic History Of Southeastern Europe

    http://biorxiv.org/content/early/2017/05/09/135616

    The Population Genomics Of Archaeological Transition In West Iberia

    http://biorxiv.org/content/early/2017/05/10/134254

  22. David Marjanović says:

    The northeastern end may well have adopted Proto-Uralic.

    Looks like it. As we’ve discussed before, the people who have the largest amount of Yamnaya ancestry today are found in Estonia, the culture that is most plausibly associated with Proto-West-Uralic is the one that started the Bronze Age in the region, and the Corded Ware culture not only encompassed Estonia but reached far north on the Finnish coast. (Heyd 2017 has maps.)

    Or maybe satemization can be blamed on Uralic influence.

    I suppose that’s possible, but it’s unnecessary: increasing the difference between the three velar plosive rows by fronting the frontmost one is the most expected development. I bet it happened several times – at least the Luwian/Lycian case must be independent from the rest.

    On Bastarnian as macro-Germanic:

    That’s just one of many possibilities which are practically untestable at the moment. I don’t know anything that’s not in the Wikipedia article I linked to, and I doubt anyone else does either. I merely wanted to raise the possibility; I’m not convinced or anything myself.

    what is remarkable about the older Indo-European languages of Europe (attested or reconstructed) is how little substrate influence is actually found in them: some vocabulary chiefly relating to plants and animals. In nature this is not very different from the substrate influence of Native American languages upon transplanted Western European languages in the Americas. Since the latter owe their spread and reduced substrate influence to the catastrophic impact upon native Americans of imported Eurasian diseases, perhaps the spread of, and the weak substrate influence upon, Indo-European languages in Europe is due to the same factor…

    Might the substrate in Mexican Spanish be a better comparison? The non-IE European “plants and animals” vocabulary – some of it, like goat, apparently shared with Berber, which may tell us that Early European Farmers got into North Africa – includes more or less all of agriculture except the simplest form of plow and surprisingly much of animal husbandry, not unlike how the substrate vocabulary in Mexican Spanish denotes a lot of food.

    After all, the Yamnaya and Funnel Beaker or whatever populations can’t have been terribly mismatched in size, or we’d have a lot more Yamnaya ancestry than we do.

    Outside the vocabulary, the substrate influence on the IE language families of Europe is hard to judge. Sure Proto-Germanic grammar is quite distinct from PIE grammar, for example (collapse of the tense/aspect system into just two tenses, etc. etc.), but we have next to no idea what the grammar of the substrate languages may have been like, so we can’t compare.

    If becoming an adult within an Indo-European community meant no longer being involved in these interactions with outsiders, perhaps it also meant shedding anything in one’s speech indicative of being an adolescent, including anything deriving from said contact language(s).

    Good idea.

    I can give the exact reference, should anyone request it downthread

    Yes, please!

  23. As for the ‘most’ usage. The cited paper shows where the errors arose:

    “The proportion of non-local individuals is at least 7/25 (28%) at Lauda-Königshofen and 8/19 (42.1%) at Bergrheinfeld.”

    “The number and proportion of females with distinctive strontium isotope ratios is notable and suggests that women were more mobile than males in CW society.”

  24. Etienne says:

    David: Mexican Spanish is indeed one of several “transplanted Western European languages in the Americas” I was thinking of: tellingly, while it has more substrate loans than American English or Canadian French do, its structure is otherwise untouched by the influence of any indigenous Mexican language: indeed, I myself, here at Casa Hat, had pointed to its being a good argument against assuming substrate influence as a matter of course (see this thread, http://languagehat.com/beckwith-on-indo-european/, and my May 2 12:41 comment in response to Jim).

    And here is the reference I had promised:

    •Holzer, Georg. Entlehnungen aus einer bisher unbekannten indogermanischen Sprache im Urslavischen und Urbaltischen. Wien: Verlag der Österreichischen Akademie der Wissenschaften, 1989. 231 s. Sitzungsberichte / Österreichische Akademie der Wissenschaften. Philosophisch-historische Klasse; Bd. 521.

  25. Trond Engen says:

    Thank you, Rick!

    Interesting stuff. This sequence from The Genomic History Of Southeastern Europe bears directly on the paper that spurred this thread:

    The Globular Amphora abutted populations with steppe-influenced material cultures for hundreds of years and yet the individuals in our study have no evidence of steppe ancestry, suggesting that this persistent culture frontier corresponded to a genetic barrier

    The migrations from the Pontic-Caspian steppe associated with the Yamnaya Cultural Complex in the 3rd millennium BCE made a profound contribution to the genetic ancestry of central Europe, contributing about 75% of the ancestry of individuals associated with the Corded Ware Complex and about 50% of the ancestry of succeeding material cultures such as the Bell Beaker Complex.

    Globular Amphora is definitely not a Yamnaya intrusion into Funnel Beaker. Rather, they seem to have been fiercely opposed to mixing until they were suddenly run over. So what is Globular Amphora then?

  26. Not really relevant, but I bet someone here knows.. Do different declensions of Latin noun have closer relationships to other ancient languages or language groupings? Is there any study of whether roots found in each declension are more or less likely to have Indo-European roots, or whether words in one declension are more likely to share cognates in a given branch of IE?

  27. Trond Engen says:

    Very interesting on how Anatolian did not get to Anatolia:

    No evidence of Copper Age Balkans-to-Anatolia migration
    One version of the Steppe Hypothesis of Indo-European language origins suggests that Proto-Indo European languages developed in the steppe north of the Black and Caspian seas, and that the earliest known diverging branch – Anatolian – was spread into Asia Minor by movements of steppe peoples through the Balkan peninsula during the Copper Age around 4000 BCE, as part of the same incursions from the steppe that coincided with the decline of the tell settlements. If this were correct, then one way to detect evidence of it would be the appearance of large amounts of characteristic steppe ancestry first in the Balkan Peninsula and then in Anatolia. However, our genetic data do not support this scenario. While we find steppe ancestry in Balkan Copper Age and Bronze Age individuals, this ancestry is sporadic across individuals in the Copper Age, and at low levels in the Bronze Age. Moreover, while Bronze Age Anatolian individuals have CHG / Iran Neolithic related ancestry, they have neither the EHG ancestry characteristic of all steppe populations sampled to date, nor the WHG ancestry that is ubiquitous in southeastern Europe in the Neolithic (Figure 1A, Supplementary Data Table, Supplementary Information section 1). This pattern is consistent with that seen in northwestern Anatolia and later in Copper Age Anatolia, suggesting continuing migration into Anatolia from the East rather than from Europe.

    So Anatolian probably came into Anatolia from east. (Though, as they say, the genetic Majkop of Eastern Anatolia and the Caucasus is still poorly understood. The picture will soon be much clearer.)

  28. Trond Engen says:

    The last paragraph in my 3:44 comment is mine, not part of the quote. (Also, I forgot to remove a hard linebreak in my 6:29 comment, but that’s just ugly, not misleading.)

  29. Fixed!

  30. Do different declensions of Latin noun have closer relationships to other ancient languages or language groupings?

    The third declension is pretty much a giant closed set. It has active adjective-forming suffixes, and some of the new adjective become substantivised. Other than that, there are almost no new coinages in the third declension. It contains the wreckage of most of the old IE declensional classes. The fourth and fifth are marginal, the fifth containing only a handful of clearly ancient nouns.

    The first and the second are open sets where new nouns are constantly bring created.

  31. And there was another related paper recently about the genomics of Scythians, it is open access at Nature:

    ‘Ancestry and demography and descendants of Iron Age nomads of the Eurasian Steppe’

    “Contemporary descendants of western Scythian groups are found among various groups in the Caucasus and Central Asia, while similarities to eastern Scythian are found to be more widespread, but almost exclusively among Turkic language speaking (formerly) nomadic groups, particularly from the Kipchak branch of Turkic languages. The genealogical link between eastern Scythians and Turkic language speakers requires further investigation, particularly as the expansion of Turkic languages was thought to be much more recent—that is, sixth century CE onwards—and to have occurred through an elite expansion process.”

    “carriers of the Yamnaya culture migrated not only into Europe but also eastward, carrying west Eurasian genes—and potentially also Indo-European languages… these observations provide evidence that the prevalent genetic pattern does not simply follow an isolation-by-distance model but involves significant gene flow over large distances.”

  32. David Marjanović says:

    Wow, thanks for the preprints! Lots of reading for tonight.

    •Holzer, Georg. Entlehnungen aus einer bisher unbekannten indogermanischen Sprache im Urslavischen und Urbaltischen.

    Oh, “Temematian”. I didn’t know it was supposed to be close to Germanic. Anyway, there’s a recent paper that tries to deconstruct every one of Holzer’s example words; I’ll look for it.

  33. Ксёнѕ Фаўст says:

    check this mystery out [Bastarnae]

    And this one. I’m curious if anyone has proposed viable Germanic etymologies for names like Helveconae, Helisii, Nahanarvali and Manimi. All the h’s do look a bit Germanic and Helisii is thought by some to be related to Claudius Ptolemy’s Kalisia (and Slavic *kališь→ Kalisz), with Grimm’s Law (on the other hand, if the names are related it’s curious that Slavs would restore the k – either they lived in close proximity and had an exonym for it or some/all of the names are unrelated; otherwise it would end up as *xališь). Recently I’ve noticed that the twin gods Alcis worshipped by Nahanarvali in a grove remind of the Lithuanian word alkas ‘hill overgrown with trees / place of pre-christian worship / pagan god’, alkė ‘grove’.

  34. I don’t think Kalisz and Ptolemy’s Καλισία are the same place or the same name (unless it can be shown that both the first and the second vowel in Καλισία were long). Transmission via Germanic is entirely out of the question, since there is no way it could account for the Slavic *a. Deriving Helisii from *kalisjā (or the like) would require palatal umlaut to have operated increadibly early (and why not in Harii, Manimi or Hasdingi?).

  35. Marja Erwin says:

    I’m not familiar with Temematic.

    Just looking it up, Kortlandt suggests that Venetic and Vinithic are Temematic and a sister-group to Germanic. But he insists that the lower Vistula wasn’t at least partly Germanic, and that early Germanic loanwords in early Baltic must have come through early Slavic. (Does this also apply to early Germanic loanwords in Estonian or Finnish?)

  36. There are cartloads of early Germanic borrowings in Finnic, and many of them neither occur in Baltic or Slavic, nor bear any indication of having been filtered through an intermediary. For example, Slavic has *useręʒь < EGmc. *ausV-xriŋga- ‘earring’. The progressive palatalisation of *g shows that the *ę comes from *iN, not *eN. The uncompounded ‘ring’ word itself was not borrowed into Slavic or Baltic. But Finn. rengas, Est., Votic rõngas etc. reflect an older form, *xreŋgaz, complete with its (Proto-)Germanic inflectional ending.

  37. SFReader says:

    -Temematic

    Cimmerian sounds better

  38. The “Temematian” etymologies are mostly root equations, and therefore poor-quality evidence. For example, Slavic *proso ‘millet’ is compared with Germanic *βar-iz- ~ βar-z-a- ‘barley’ and Lat far/farris < *bʰar-s- ‘hulled wheat (spelt, emmer, einkorn)’. Germanic and Italic together point to an original s-stem with a fundamental *a in the root syllable and no evidence of a zero grade (which “Temematian” *ro allegedly reflects). Old Prussian prassan ‘millet’ is too close a relative to prove anything (it could even be a loan from Polish), but comparison with Tocharian B proksa ‘grain’ shows that we are probably dealing with *próḱs-o-m, pl. *proḱs-a-h₂. Perfect morphological agreement and no need of a Temematian shift.

  39. David Marjanović says:

    Haven’t had time to read any preprints yet.

    Just looking it up, Kortlandt suggests that Venetic and Vinithic are Temematic and a sister-group to Germanic.

    Ah. Another trip to Kortlandt’s homepage for me, then.

    Cimmerian sounds better

    Heh. 🙂 “Temematian” is supposed to indicate that the PIE tenues (p, t, k) show up as Slavic mediae (b, d, g), while the PIE mediae aspiratae show up as Slavic tenues.

  40. A bit late to the party, but regarding the mysterious appearance of the Magyars in the Northeast Pacific, I agree that they most likely meant to put the Malgal.

    It isn’t a complete surprise to see such an error appear in a map in a history published in Singapore mainly for an overseas Chinese audience, given that East Asian ethnonyms often appear in quite different forms between English and Chinese—e.g. Jurchens and Khitans are 女眞 Rǔzhēn and 契丹 Qìdān respectively in Mandarin. So it’s understandable that such mistakes would slip through the editors.

  41. January First-of-May says:

    For example, Slavic has *useręʒь < EGmc. *ausV-xriŋga- ‘earring’.

    That’s where Russian серьга, серёжка comes from, right?

    (Apparently Wiktionary says it’s from Turkic; though the supposed Turkic forms sound sufficiently similar that for all I know they were borrowed from Slavic originally.)

  42. Vasmer says “Предполагают преобразование из др.-русск. усерязь (ХI в.)…; однако трудно понять появление вторичного -г- в великорусск.; ср. блр. по златои серазѣ, еще в лютеранском катехизисе 1562 г. …. Правильнее считать русск. слово заимств. из др.-чув. *śürüɣ “кольцо”, чув. śǝrǝ, śørǝ “кольцо”, тат. jözök, тур., чагат. jüzük – то же.” [The suggested derivation from ORuss. усерязь is difficult because of Great Russian secondary -г-; it’s preferable to consider it a borrowing from Old Chuvash *śürüɣ ‘ring.’]

  43. Vasmer’s objection isn’t quite valid. The progressive palatalisation was blocked if the velar was followed by *u or *ū at the time (later *ъ, *y). The blocking endings included probably the gen.pl., the acc.pl. of masculines and feminines, the instr.pl. of masculines and neuters, and the gen.sg. of feminines. After the regular operation of the sound change, most — but not all — case forms had a palatalised consonant. The difference was soon levelled out by analogy, but “soft” and “hard” variants sometimes survived side by side, hence e.g. *ęʒa (Polish jędza) ~ *ęga (Russ. баба-яга) ‘witch, hag’ (analogical after the gen.sg./acc.pl. *ęgy). The Russian idiom ни зги не видно ‘you can’t even see the path’ (= ‘it’s pitch dark’) retains the original, unpalatalised form of the gen.sg. of *stьʒa ‘path’ (*stьgy > z(d)gi). PSl. *useręʒь must have been accompanied by forms like the acc.pl. *useręgy or gen.pl. *useręgъ, so an analogical reversal of the progressive palatalisation was possible as long as those forms lingered on.

    It’s Great Russian, not Belorussian, by the way.

  44. Oops, thanks. Fixed now.

  45. Ксёнѕ Фаўст says:

    As for Temematic, apart from the possibility of alternative etymologies, the sound changes seem unlikely (unless in a very long time frame combined with substrate effects) to me, with the voicing swap (t → d, dh → t).

    As for Kalisz, yeah, I’m aware the source form would have to be kālīsj- to yield *kališь regularly (for it’s not Kalesz or Kolesz). Although -išь was a popular hypocoristic suffix so it’s possible that, for instance, *kalьšь would be altered to *kališь via folk etymology, I think. According to wikipedia there was a German study in this decade that claimed to confirm the identity of Kalisia and Kalisz but I have no access to it.

    “Der Ort gilt als sicher lokalisiert. Ein interdisziplinäres Forscherteam um Andreas Kleineberg, das die Angaben von Ptolemaios neu untersuchte, bestätigt anhand der entzerrten antiken Koordinaten die bisherige Lokalisierung von Kalisia beim heutigen Kalisch (Kalisz) in der Woiwodschaft Großpolen in Polen. Die polis an einem Übergang der Prosna[3] war offenbar eine Station an der Bernsteinstraße.”

  46. See the 99.9% exhaustive treatment of the subject by Willem Vermeer (it would be 100% exhaustive if he discussed the ‘[ear]ring’ word). I found this article only a few minutes ago, haven’t seen it before, and am reading it now with enormous interest. This example caught my eye:

    Noun *orbotędźь as reflected in Old Polish robociądz ‘boy’, Old Czech robotěz ‘slave’. Old Russian has rabotjagъ ‘slave’ with unmodified -g-.

    So the Slavic prototype of the word robot could be combined with the Germanic suffix *-iŋg- already at the time of the progressive palatalisation. Now they have met again:

    http://www.springer.com/la/book/9783319400013

    (to be sure, the first time it was an occupational term derived from *orbota ‘work, toil’, and now it has returned as a nomen actionis).

  47. Trond Engen says:

    Rick: ‘Ancestry and demography and descendants of Iron Age nomads of the Eurasian Steppe’

    A continuity test was performed between the two Iron Age groups (‘West’ and ‘East’) and a large set of contemporary Eurasian populations (n = 86, Supplementary Table 19). For western Scythian-era samples, contemporary populations with high statistical support for a genealogical link are located mainly in close geographical proximity, whereas contemporary groups with high statistical support for descent from eastern Scythians are distributed over a wider geographical range. Contemporary populations linked to western Iron Age steppe people can be found among diverse ethnic groups in the Caucasus, Russia and Central Asia (spread across many Iranian and other Indo-European speaking groups), whereas populations with genetic similarities to eastern Scythian groups are found almost exclusively among Turkic language speakers (Supplementary Figs 10 and 11).

    Afanasevo lasted to 2500 BCE and was replaced by the culture of the neighbouring Siberian forest people, who took up animal husbandry. I’ve wondered if those were the (Pre-)Proto-Turks. Some 1000 years later Andronovo came eastwards and into old Afanasevo lands with the Scythians, Now we learn that Eastern Scythians had mixed local-Yamnaya ancestry, and that the only modern descendants of Eastern Scythians are Turkic speakers. So maybe they included Turkic speakers from the outset, and that the Turkic element came to dominate when non-Scythian Turkic became influential on the Eastern steppe.

  48. David Marjanović says:

    the only modern descendants of Eastern Scythians are Turkic speakers

    Any Ossetes in the dataset?

    (I haven’t had time to read the preprints yet. Soon…)

  49. David Marjanović says:

    I’m aware the source form would have to be kālīsj- to yield *kališь regularly […]. Although -išь was a popular hypocoristic suffix […]

    Sure, that takes care of having to assume . But if the name is supposed to be Germanic, it has to be Northwest Germanic, because hardly existed in East Germanic – and isn’t Kalisz rather too far east for that?

    I’m curious how this interdisciplinary team “de-distorted” the “coordinates”.

  50. because *ā hardly existed in East Germanic

    Do we really know that? Gothic is written in Greek script (modulo handwriting issues), and Greek script has never had a method, however limited or fallible, for distinguishing ā from ă, unless the former happens to have a circumflex accent. So the contrast might have been robust in Gothic and just not representable in writing.

  51. Gothic had ā of native origin at least from denasalised *[ãː] < *[aŋx]. I discuss some other possible sources of Gothic ā in an article which is still in press. Nonetheless, “Calisia” does not offer a segmental environment in which a long ā could have developed.

  52. (I mean, denasalised *[ãː] < *[aŋ] when followed by *[x])

  53. SFReader says:

    Looked up what Polish etymological sources say about Kalisz.

    It turns out they believe that the word comes from Slavic ‘kał’ (feces, excrements, filth).

    For the sake of young Kaliszane, I hope better Cimmerian etymology would be found than Big Shit! (Kaliszcze)

  54. David,

    Yes. Ossetians are included.

    They are most similar genetically to Turkic speaking Balkars and Kumyks, followed by the Northeast Caucasian speaking Lezgians and Chechens, and the Northwest Causasian speaking Adyghs, then Iranians, then Iraqi and Iranian Jews, Abkhazians, Georgians, and the Turkish.

    These populations are all very closely related, and all have almost identical Scythian ancestry. That’s some major language diversity for such closely related peoples.

  55. The simplest solutions are the best as long as no facts contradict them. There is a rare but attested Old Polish dithematic name Kalimir in which the first element is connected with the noun kał ‘excrement’ (now a medical term, formely meaning ‘mud, dirt’) via the kalić ‘defile, soil, disgrace’ (only its iterative kalać is used in present-day Polish. The normal hypocoristic form of Kalimir would have been Kalich, and the possessive derivative of that, Kalisz, would perfectly fit into the Polish place-naming pattern. This is, so to speak, the default etymology, involving no special pleading. It leaves unexplained the similarity of Kalisz (which under this hypothesis can’t be pre-mediaeval) to Ptolemy’s Καλισία, but then the similarity (which looks superficial and may be accidental) is the only reason why the identification is insisted upon by some. No other Polish city, no matter how important, has a demonstrably pre-Slavic name (except when it’s derived from the name of a river or a mountain — these are often old — and well, there may be some West Baltic names in the northeast). Remember that the time gap between Ptolemy and the earliest attestation of Kalisz as the name of a town in Poland (1107) is about a thousand years.

  56. SFReader says:
    It turns out they believe that the word comes from Slavic ‘kał’ (feces, excrements, filth).
    and
    Piotr Gąsiorowski says:
    kał ‘excrement’ (now a medical term, formely meaning ‘mud, dirt’)

    In Croatian “kal” still means mud, though it is not the ordinary word for “mud”. The ordinary word is “blato”, and there are “muddy” places in Croatia – notably Blato on the island of Korčula. The etymology of “kal” as given by the Croatian Dictionary portal is:

    pre-Slavic & Old Slavic kalъ (Russian. kal, Polish kał) ← ie. *keh2lo-: “dark” (lat. caligo: “fog”, skr. kālas: “blackish”)

  57. David Marjanović says:

    Do we really know that?

    The in brought, thought is of PGmc. origin as Piotr said; it’s written a in Gothic, and given the preservation of vowel length where we can see it it was most probably still present here, too. In addition, Ringe’s 2006 book says that *-aja- (limited to a few verb forms) was contracted to *-ā- on the way to PGmc., so apparently this was preserved in Gothic as well. That makes two sources for a very rare phoneme. Seeing as Gothic didn’t undergo the > shift of Northwest Germanic, any further occurrences must have been derived within Gothic by processes yet unknown – I’m looking forward to Piotr’s paper.

  58. A recent paper, Milk and the Indo-Europeans by Sagart and Garnier, claims to show linguistic evidence to support the idea that the Corded Ware and Yamnaya folks introduced into Europe genes for lactose tolerance, and hence milk-drinking. The idea doesn’t seems far-fetched, but I can’t evaluate their arguments.

  59. Y,

    There are some major problems with that theory about milk. The biggest one being that they didn’t actually find high levels of lactase persistence in these ancient groups, but rather imputed the alleles based on modern statistics, which is not appropriate in this case.

    When imputing the −13.910:T frequency based on surrounding markers, Allentoft et al. reported a high LP allele frequency (approximately 20%) in ancient steppe populations. However, the major haplotype currently carrying the −13.910:T allele is also found carrying the −13.910:C (LNP) allele both in high frequency in modern Eurasian populations and in one early Neolithic individual. Therefore, the −13.910 genotype cannot be reliably imputed from surrounding sequences, whether from modern or ancient data.

  60. There are problems with the linguistic argument as well. It’s only sketched in the presentation, but my impression is that they for example take the absence of some agricultural terms from Indo-Iranian or their presence there with non-agricultural meanings as evidence of their “regional” character, rather than the consequence of the fact that a group of IEs became specialised in nomadic pastoralism and spent a long time roaming the steppes with their livestock. After all, Proto-Oceanic lost the AN words for ‘rice’ and ‘millet’, and yet the authors don’t regard them as regional.

  61. David Marjanović says:

    Lactose tolerance wasn’t widespread in early Bell Beaker times either, according to a comment here (which says that the preprint, which I still haven’t read, shows this).

  62. Greg Pandatshang says:

    Has anyone looked at the genetic affinities of the non-national Uralic-speaking peoples? (i.e. everyone but Finns, Estonians, and Hungarians—and, really, I tend to disregard the Hungarians, as the invading Magyars are the textbook example of high linguistic impact along with minimal genetic impact)

    I’m trying to wrap my head around the secnario where northeastern Corded Ware adopts/is adopted by … Proto-Uralic? Or Proto-[subset of Uralic]? Wikipedia suggests a wide range of possible time depths for Uralic.

    If it’s simply that an IE-speaking Corded Ware élite class conquered the Pre-Proto-Uralics and eventually adopted the commoners’ language (the Norman/Bulgarian model) we’d expect to see the tale told in both genetics and language contact effects. Is it possible that the entire Uralic family originated as a creole combining IE with an otherwise unknown substrate? Sort of the opposite of Kortlandt’s view that IE originated as a Uralic dialect with radical NW Caucasian contact effects.

  63. SFReader says:

    In Finnish language, Orja means ‘slave’.

    Looks like an Uralic elite class conquered Pre-Proto-Aryan tribes and not vice versa.

  64. Lars (the original one) says:

    That is suggestive, but one word in one modern language can just be a coincidence. Is this orja reconstructed to anything like the time depth we’re talking about?

  65. Etienne says:

    Greg: on the Italian dialect thread (next to last comment) I gave a reference to an Uralicist who argues that there was indeed shift from Indo-European to Uralic in northeastern Europe.

    As for creolization in Uralic and Indo-European: creoles are not mixed languages, and thus to speak of a creole as consisting of equal parts of a superstrate and of a substrate is inaccurate. Creole typological distinctiveness has as its source the pidginization of the source language (i.e. what in creolistics is called, inaccurately from the vantage point of historical linguistics, the superstrate), and not the influence of the substrate language(s).

    Now, pidginization entails (near-) total loss of bound morphemes, and thus the trouble with any claim that Indo-European or Uralic might have been creolized (as opposed to influencing one another deeply, nota bene!) is the reality that both proto-languages had plenty of bound morphemes.

  66. It’s really annoying that there are two definitions of sub/superstrate, the temporal one (from geology), whereby the substrate was spoken at an earlier time and the superstrate at a later one, and the power/prestige one, whereby the substrate is spoken by the lower classes and the superstrate by the upper. Often enough they coincide, but sometimes they doesn’t.

  67. Marja Erwin says:

    “Is this orja reconstructed to anything like the time depth we’re talking about?”

    Anthony says it’s a common Finno-Ugric root, generally meaning “southerner,” and in Finnish and a few other languages, “slave.” He attributes it to contact between Indo-Iranian Sintashta and Finno-Ugric Volosovo. (2007, p. 385)

  68. Greg Pandatshang says:

    Wiktionary purports cognates in Erzya and Udmurt.

  69. Lars (the original one) says:

    common Finno-Ugric root, generally meaning “southerner,” and in Finnish and a few other languages, “slave.”

    But possibly common Uralic if the Erzya cognate is valid? Or might that be a borrowing?

    In any case it would seem likely that the original meaning was ‘southerner’ and that the shift to ‘slave’ happened in individual branches at a time where the specific southerners were no longer the original Pre-Proto-IE speakers. In fact the consensus seems to be that ‘Aryan’ only became an autonym in Indo-Iranian, whatever you think of its PIE ancestry,

    It is of course possible that a Pre-Proto-Uralic-speaking elite once conquered the Pre-Proto-IE speakers and that the latter were already calling themselves ‘Aryan’s at the time, in which case the Finnish meaning of the word would be original, but any number of other scenarios are more likely so ‘looks like’ is a bit strong.

  70. So the traditional connection between Aryan on the one hand, and Old Irish aire ‘free, noble’ and Gaulish names like Ariogaisus ‘noble spear’ on the other, is now abandoned?

  71. SFReader says:

    Russian archaeology recognizes so called Seima-Turbino phenomenon – traces left by groups of bronze metal workers and warriors of same origin spread along astonishing distances from Finland to Mongolia (including remote Siberian Arctic!) in middle of 2nd millenium BC.

    In what capacity they were expanding is not very obvious, elite/mercenaries/highly paid foreign experts are some of the suggested explanations.

    Anyway, it is clear that they were relatively small groups of foreigners by no means representing the general population of the regions they are found. It is thought that mapping spread of Seima-Turbino phenomenon matches perfectly with Uralic expansion while some point to Aryan Indo-Europeans instead.

    Of course, it could a combination – say, spread of an Uralic military elite accompanied by their Aryan (Orja) metal workers (disparagingly referred to as ‘slaves’).

  72. Lars (the original one) says:

    So the traditional connection […] is now abandoned? — not from what I can see quickly, but its use as an ethnic autonym is no longer dated to PIE, only to II. Wikipedia talks about a ‘root’ *haerós (unclear what that would be in a modern reconstruction) meaning ‘member of ingroup’ — with the real root *haer- meaning ‘put together’. Adolescent warrior bands turning into elites/nobility again?

    (There’s a lookalike in Latin (ad)haerere which might (Watkins) be from *ghais- — not the same root as written, and probably not at all, but I wouldn’t trust Wikipedia to know the difference).

  73. David Marjanović says:

    The trick is that the Corded Ware/Battle Axe Culture is Neolithic/Chalcolithic; those battle axes were made of polished stone. The Seima-Turbino phenomenon introduced bronze, for which there’s a Uralic wanderwort (traditionally reconstructed as PU *wäśkä, but the vowels don’t line up between at least some of the Uralic branches). That sounds like a reason for a language shift.

    It hasn’t been that long since we discussed a whole special issue full of papers about this and related issues. Where was that post again…

  74. David Marjanović says:

    Here’s the special issue in any case. Open access.

  75. So the traditional connection between Aryan on the one hand, and Old Irish aire ‘free, noble’ and Gaulish names like Ariogaisus ‘noble spear’ on the other, is now abandoned?

    In IIr. *aryá-, the first /a/ can reflect any short non-high vowel, and we don’t even know for sure that the /r/ comes from PIE *r rather than *l. If, as usually proposed, it’s related to RV arí- ‘guest, stranger, alien’ (hence the originally possessive adjective aryá- ‘hospitable’, substantivised as ‘host, master of the house’ –> ‘lord, leader’, then vr̥ddhied into ā́r(i)ya- ‘noble, nobly born, having the qualities of an aryá-), we can at least exclude *o (no Brugmannian lengthening in arí-) but not the other possibilities. Note that the meaning ‘noble’ and the connotations of close ingroup solidarity, as in aryamán- ‘comrade, friend, best man’, are not original but derived within Indo-Iranian, which makes the equation with Celtic Ario- less plausible that it might seem at first blush. On the Celtic side, Ario- may come from *pr̥h₃-jo- ‘foremost, leading’, cf. the common prefix *(φ)ari- ‘in front of, fore-‘ (OIr. air-, Gaulish are-, etc.). Old Irish aire ‘freeman, nobleman’ is a k-stem (*arisak-), not really parallel to anything in Indo-Iranian.

  76. Trond Engen says:

    On the Celtic side, Ario- may come from *pr̥h₃-jo- ‘foremost, leading’, cf. the common prefix *(φ)ari- ‘in front of, fore-‘ (OIr. air-, Gaulish are-, etc.).

    Could the common Germanic naming element harja- be a borrowing (or mis-translation) of early Celtic *(φ)arjo-?

  77. Of course haria- has its own inner Germanic etymology (*xarja-z ‘troop, host, army’), but there’s little doubt that early Germanic speakers may have identified it folk-etymologically with the (H)aria- proterotheme in the names of their Celtic neighbours. The Roman authors were not very consistent as they recorded Germanic tribal and personal names. We find both Chario- (especially in the west) and Hario- (especially in the east), and since at that time the Romans dropped their own (h)aitches, an initial H- was sometimes omitted or added regardless of etymology. We still don’t know whether Ariovistus had a Celtic or a Germanic name.

    Similarly, Slavic -mirъ ‘peace’ was more or less identified with ON -mærr ‘illustrious’ or conflated with it as -měrъ, since they both functioned as second members of dithematic names, sometimes with cognate first members, as in Valdamarr vs. Volodiměrъ/Vladimir.

  78. Greg Pandatshang says:

    Incidentally, another blogger I read regularly, normally non-linguistic, once asked his commenters the following trivia question: which two modern countries have names that are derived from Arya? I of course correctly answered Iran and South Ossetia. Sadly, the blogger in question himself remained mired in the delusion that the answer should be Iran and Ireland.

  79. David Marjanović says:

    there’s little doubt that early Germanic speakers may have identified it folk-etymologically with the (H)aria- proterotheme in the names of their Celtic neighbours.

    Hm.

    der vor dem Heer herzog, ward Herzog genannt
    “the one who trekked along in front of the host was called duke”

    That’s from a slightly disgusting poem from the 19th or 20th century and an ideological statement, not a serious etymological hypothesis. But now I wonder if the West Germanic “duke” words reflect just this conflation of “army” and “first”.

  80. David Marjanović says:

    Ha! Ireland is the copper island next to the tin island, all in Phoenician.

  81. OHG herizogo = OE heretoga < *xarja-tuɣan- ‘army-leader’. The second member of the compound, *tuɣ-an-, is actually a cognate of dux (ziehen = dūcere). Heri, here were perfectly normal words in both languages at the time, and these compounds were fully transparent semantically (which was not always true of Germanic and Celtic proper names).

  82. South Ossetia…

    If you mean Alania, that’s North Ossetia.

  83. South Ossetia calls itself Respublikæ Khussar Iryston, which I suspect is what he means.

  84. Ah, but the self-designation ir is no longer regarded as derived from āryana-. The Alani, on the other hand, were etymologically ‘the Aryans’ (with the characteristic “Sarmatian” lambdacism of palatalised *r, shared also by Ossetian), so Alania counts.

    Edited to add:

    P.S. Johnny Cheung’s study of Ossetian historical phonology supports the derivation of Ir, Iron and Iryston from *wīra- ‘man, hero’.

  85. Ah, once again I am behind the times! Thanks for updating me.

  86. Trond Engen says:

    Piotr: OHG herizogo = OE heretoga < *xarja-tuɣan- ‘army-leader’.

    Yes, I didn’t mean to suggest that the word itself came from Celtic, only its popularity as a naming element. A name like *harjagastiz makes more sense if understood as “noble guest” at the time of coinage.

    I even started wondering if German Ehre f. “honour, fame, respect” and relatives could be borrowed from Celtic, since the IE etymology strikes me as thin. ON eir f. “protection, mercy, help” says no, though. But I do wonder if the WGmc. meaning could be coloured by Celtic.

  87. The impact of Celtic on Germanic is clear. For example, all those names of tribal chieftains ending in *rīk- are copies of a Celtic pattern, and *rīk- itself is a loan. Slavic *mīrъ was partly identified with Germanic *mēri-, but at an earlier date *mēri- was for similar reasons partly identified with Celtic *māro-. There are even dithematic cognates (calques?) like *touto-rīg- (cf. Tudor) and *þeuða-rīk- (Theodoric), and it’s hard to believe that the coining of such strictly cognate tribal names as the Brigantes and the Burgundians was completely independent.

  88. ‘Noble host’, more probably. The meanings ‘host’, ‘guest’, ‘stranger’, and ‘enemy’ take forever to sort out, and in some IE languages it never quite happens.

  89. It reminds me of the Indo-Aryan vocative or *ari-, *arai > are!, an exclamation used to get attention (often reduplicated as arere!). It means literally “hey, stranger!” (etymologically, at any rate).

    John: it might also be interpreted as a bahuvrihi: “having noble guests”.

  90. Marja Erwin says:

    Related Celtic loanwords reached Old Norse. Rikr, ambatt, etc.

  91. The northeastern end may well have adopted Proto-Uralic.

    As was just argued by Kallio at our conference Contextualizing historical lexicology earlier this week in Helsinki (slides generally forthcoming at some point): probably Proto-Finnic specifically, and with the substrate at the time being clearly Balto-Slavic and not (para-)Germanic. The Volga-Kama area languages (Mordvin, Mari, Udmurt) show no indication of ever having been anywhere else or having ever taken over any substrates, at least not individually.

    There are cartloads of early Germanic borrowings in Finnic, and many of them neither occur in Baltic or Slavic, nor bear any indication of having been filtered through an intermediary.

    I still wonder about vocabulary with a “Northwestern European” distribution covering Germanic, Finnic and Balto-Slavic. Two of these are related, but that doesn’t mean such vocabulary can’t have originated in one of these groups specifically and been loaned to both of the others.

    There are already a small number of known cases where something Germanic goes thru Finnic and then ends up in East Baltic (one long since clear example being ‘ship’: G *flawją̄ → F *lawja > *laiva (> Fi laiva etc.) → Latv. laiva, Lith. laivas); and at least one likely case where something Baltic goes through Finnic and ends up in Germanic (B *žaras ‘branch’ → F *šara > *hara ‘branch’, *harapa ‘rake’ (> Fi harava etc.) → Old Norse harfr > Eng. harrow). Sufficiently early loans directly between B and G would likely be harder to identify, though. Doesn’t Baltic have a bunch of words that “fail Satemization”?

    Is this orja reconstructed to anything like the time depth we’re talking about?

    There is zero doubt that at least the Finnic + Mordvin + Permic words for ‘slave’ are related, which would traditionally put it as a “Finno-Permic” root. (OTOH my contribution to the conference argues that “Finno-Permic” is not a subgroup as much as something like a contact/diffusion zone featuring distinctive IE loanwords.) A competing loan etymology by Koivulehto derives this as an older Satem-type loan from PIE *werǵ- ‘work’ though, which seems semantically more straightforward. So my trust in Uralic speakers superstrating “Aryans” specifically isn’t very high.

    The Samic word for ‘south’, ‘southwest’, ‘west’ (the cardinal direction terms amusingly enough rotate depending on what direction rivers run in in an area) is also an exact phonetic match, but it takes a bit of storytelling to connect the semantics, and there seems to be a complete lack of evidence for the supposed intermediate stages like ‘general name for southerners’.

    Of course haria- has its own inner Germanic etymology (*xarja-z ‘troop, host, army’)

    Connecting the last side of the triangle: Kümmel now argues that initial laryngeals were retained quite late in Indo-Iranian. I wonder if this Germanic term could not be a loanword from II *h₂arya-.

  92. I wonder if this Germanic term could not be a loanword from II *h₂arya-.

    But there are exact cognates in Baltic and Celtic pointing to *kor-jo- and also meaning ‘troop, army’. Even Greek has several personal names with Koiro- and the noun κοίρανος ‘warlord, ruler, commander’ (+ derivatives).

  93. Greg Pandatshang says:

    Thanks, Piotr. I never knew that Iron and Iran are false friends. Or that Alan and Iran are, it seems, trusty comrades. (Like the song says “Pinot Noir / Find out who your true friends are”)

  94. David Marjanović says:

    Related Celtic loanwords reached Old Norse. Rikr, ambatt, etc.

    They’re in Gothic, too, and therefore were most likely already present in Proto-Germanic – fitting the fact that they’ve undergone Grimm’s law (kr).

    the cardinal direction terms amusingly enough rotate depending on what direction rivers run in in an area

    That’s like the ancient Egyptians calling the Euphrates “the river that runs the wrong way around”! 🙂

    Doesn’t Baltic have a bunch of words that “fail Satemization”?

    Yes, but they could be Celtic loans at least as easily as Germanic ones, AFAIK.

    I wonder if this Germanic term could not be a loanword from II *h₂arya-.

    Intriguing idea, but apart from *kor-jo- there’s the sheer geographic distance. The temporal distance is a bit extreme, too: we’d need to postulate that westernmost “Scythian” kept *h₂ for 1500 years longer than the eastern side of II until Germanic had undergone Grimm’s law…

  95. Trond Engen says:

    j: The Samic word for ‘south’, ‘southwest’, ‘west’ (the cardinal direction terms amusingly enough rotate depending on what direction rivers run in in an area)

    Not only South Saami. Eastern Norwegian rivers traditionally define the local north-south axis. Nord i dala means “further up in the valley”. (I remember I first encountered this in an old yearbook for Den norske Turistforening “The Norwegian Accociation of Hikers”, where it was told as an amusing fact that the hostess of a certain mountain lodge in Jotunheimen was of the old school who still had the cardinal directions 180 degrees shifted, since the valley fell to the north.) I’ve never looked into North Swedish usage, but it wouldn’t surprise me if it’s the same there.

  96. Riverine coordinates are a standard alternative to egocentric and absolute ones in various languages. I believe m-l has said that the Tsimshianic-speaking cultures she studies use riverine coordinates, and in Hawaiian-flavored English they have seen some use as well.

    On Larry Niven’s Ringworld, a huge ring spinning around its sun, as big as the Earth’s orbit and a million miles wide, with gazillions of hominins of many species living on it, the standard directions established by the first explorers from Earth are spinward (the way the Ring turns), antispinward, port (left when facing spinward), and starboard (right when facing spinward).

  97. Ксёнѕ Фаўст says:

    Nahanarvali: the bastards keep bugging me and I’ve been wondering if they could be Balto-Slavic:

    naha- would be the equivalent of the Slavic preposition & prefix *na ‘on’ with ‹h› representing the acute (glottalization); Lith. nuo ‘from’, Prussian no, na ‘on’ are the same thing
    naru- Narew river (I think it’s related to Slavic *noriti ‘submerge’, *po-norvъ ‘burrowing worm/slug’, Lith. nerti ‘dive, submerge’)
    -(a)l- adjectival suffix (cf. Slavic *-lъ found e.g. in the resultative/perfect participle and in a few old adjectives like *milъ, Germanic *-laz)

    So the name would mean: [those living] on the Narew.

    Additionally their deities’ name makes better sense in Baltic than in Germanic (unless it’s a Germanic without Grimm’s law).

    What do the more knowledgeable think?

  98. Eastern Norwegian rivers traditionally define the local north-south axis. Nord i dala means “further up in the valley”.
    Which is a bit different from the Sami usage: direction terms there keep the same meaning on both sides of the watershed division, so glosses like “left of downstream” won’t work. Perhaps it would be more accurate to say that they’re defined with respect to the reindeer herding areas, which are themselves (usually) defined with respect to major rivers.

    The Atlantic coast up to at least Magerøya was “always” held by the Norse though (already by the time the Sami languages arrived), so the Atlantic coast and its north-west-ish flowing rivers might never have been super relevant. Even if we supposed that there once were Atlantic-side Sami-speaking groups who e.g. used nuorti to mean ‘right of downstream’ = ‘west’ ~ standard Northern Sami ‘east’, that probably would have gotten wiped out after the rise of reindeer nomadism in the Middle Ages.

    we’d need to postulate that westernmost “Scythian” kept *h₂ for 1500 years longer than the eastern side of II until Germanic had undergone Grimm’s law
    No need to wait that long: early IE *h₂ gets loaned as *k into Uralic, and Anatolian /ḫ/ as /kʰ ~ k/ into Greek. The same could in principle work for Germanic (likewise even Baltic).

    Celtic and Greek cognates however sound like a stronger counterargument. Direct loaning is indeed not really an option, and even mediated loaning doesn’t work if these show *o.

  99. Marja Erwin says:

    “On Larry Niven’s Ringworld, a huge ring spinning around its sun, as big as the Earth’s orbit and a million miles wide, with gazillions of hominins of many species living on it, the standard directions established by the first explorers from Earth are spinward (the way the Ring turns), antispinward, port (left when facing spinward), and starboard (right when facing spinward).”

    Marc Miller’s Traveller uses coreward, rimward, spinward, and trailing over a significant portion of the galaxy. Although it doesn’t address distances port/north or starboard/south of the galactic plane.

  100. Trond Engen says:

    What do the more knowledgeable think?

    That’s not me, but I like it. I gave up trying to make it work with Germanic. *nēhwa- “near”.

  101. Trond Engen says:

    The Atlantic coast up to at least Magerøya was “always” held by the Norse though (already by the time the Sami languages arrived), so the Atlantic coast and its north-west-ish flowing rivers might never have been super relevant.

    I should say that the riverine system as far as I know was purely Southeast Norwegian, used in the watersheds emptying into the Skagerrak, where the rivers mosty run parallel, just as in Northern Sweden. In western and northern Norway, north was always along the coast. (Also, I have no idea if these local systems were all-encompassing. The seafaring coastal population obviously maintained the knowledge of the celestial directions and saw the litoral directions as a practical local adaptation. I suspect this to have been the case even for many inland dwellers with the riverine directions.)

  102. Re: Nahanarvali. Beware of lookalikes!

    A. Is there any reason to locate the Nahanarvali on the Narew?

    B. Slavic tribal names derived from rivers (with or without a prepositional prefix) almost invariably end in -jane (pl.). As for the Balts, the usual ethnonymic suffix is -iai. No such names end in -al-, and the suffix *-lo- forms deverbal adjectives (*milъ and Lith. míelas ‘nice, dear’ are not exceptions: *meiH- ‘ripen, flourish’ was a verb root). As for prefixes, one could expect po- (as in the name of the Polabians), possibly nad- or za-, but hardly na-.

    C. The early Slavic name of the Narew was *Nary (gen. *Narъve) — an ū-stem (like *kry, *krъve ‘blood’). In Tacitus’s time, this would have been *Nārū (gen. *Nāruwe). I don’t think one could expect to find the development *Nāruwjāne (?) > *Nāruwljāne so early — /l/-epenthesis is a late, dialectal development, so such a source of the /l/ in Nahanarvali can also be ruled out.

  103. David Marjanović says:

    No need to wait that long:

    Good point.

    In the meantime, I found (and already can’t find the source anymore) that an exact cognate of κοίρανος, Herjann, was an epithet of Óðinn.

  104. Marja Erwin says:

    Is Tacitus the only source for the name Naharvali?

  105. Greg Pandatshang says:

    Have we considered the possibility of a isolated band of peripatetic narwhals?

  106. SFReader says:

    Nahanarvali could be a name formed from a town.

    -l’ – is quite common Slavic ending for towns – Yaroslavl, Pereyaslavl, Przemyśl.

    The hypothetical town on Narev would have been called something like Nanarevl’ (Slavicized form).

  107. Marja Erwin says:

    If Nahanaruali is the Roman transcription, and if it is Germanic, then perhaps “People at [nigh] the little narrows”? How often does “narrows” refer to passes and other land features?

    Is there any reason, besides the name, to associate a branch of the Lugii with the Narew?

  108. -l’ – is quite common Slavic ending for towns – Yaroslavl, Pereyaslavl, Przemyśl.

    Nope. The ending is *-jь < *-jos. In Yaroslavl’ and Pereyaslavl’ the -l- is “intrusive”: in many dialects (East and partly South Slavic) an epenthetic lateral was inserted between a labial (*m, *p, *b, *v) and /j/, or rather */j/ came to be pronouced [ʎ] after labials. This is why Russian and Ukrainian have земля́ while Polish has ziemia, Czech země, and Upper and Lower Sorbian zemja. In Przemyśl, the -l- is not a suffix but part of the personal name (Przemysł) to which the possessive suffix *-jь was added.

    Even assuming for the sake of the argument that /l/-epenthesis is oder than most people think, the lateral should directly follow the labial.

  109. Trond Engen says:

    Greg P.: Have we considered the possibility of a isolated band of peripatetic narwhals?

    I was going to suggest an isolated band of peripatetic Nahuatl, but I agree that your suggestion has more going for it.

  110. January First-of-May says:

    Nope. The ending is *-jь < *-jos. In Yaroslavl’ and Pereyaslavl’ the -l- is “intrusive”: in many dialects (East and partly South Slavic) an epenthetic lateral was inserted between a labial (*m, *p, *b, *v) and /j/, or rather */j/ came to be pronouced [ʎ] after labials. This is why Russian and Ukrainian have земля́ while Polish has ziemia, Czech země, and Upper and Lower Sorbian zemja. In Przemyśl, the -l- is not a suffix but part of the personal name (Przemysł) to which the possessive suffix *-jь was added.

    As visible in the city name Vladimir (where the epenthesis never happened, and the ending ultimately disappeared) and, possibly, Ryazan’ and Voronezh (though in both cases there are other competing etymologies not involving personal names).

  111. Celtic-type coins (staters) minted in Poland in areas associated with the Lugii show two goddesses, identified by Andrałojć & Andrałojć (2014, “The unknown face of the Lugian Federation – Celtic coinage in the Polish lands”) with a shield-carrying Athena (= Alcis; Alkis/Alkidemos are known epithets of Athene at Pella, Macedon) and Nehalennia (seated upon a boat, like her Zeelandic prototype). The authors suspect that the “twin gods” whose priests wore “female attire” were in fact female themselves. The Lugian Federation seems to have been ethnically mixed (at least at the time it was established), possibly with Celtic refugees from Northern Gaul and the Danubian/Carpathian region having been integrated into the local East Germanic groups at the élite level. The A&A article (which can be found in English at Academia.edu) makes a lot of sense to me. I can only add that the name of the Lugii may simply mean ‘united, oath-bound’ cf. Goth. liuga ‘marriage’, OIr. luige ‘oath’. A connection can be suspected between the Nahanarvali and Nehalennia, but unfortunately the latter name, though well attested in Roman-period inscriptions, also remains obscure (Celtic *nāwā ‘boat’?).

  112. Ксёнѕ Фаўст says:

    A. Is there any reason to locate the Nahanarvali on the Narew?

    I don’t think there’s a historical reason to precisely locate them anywhere in, roughly speaking, present day’s Poland (except that not near the sea coast as it was occupied by other, non-Lugian peoples). The Lugii are described as a “great people”, “immense host” (so not county-sized tribelets I assume). Well, the Roman interest in peoples they didn’t directly border and weren’t at war with wasn’t particularly high.

    As for ‘prepositional’ prefixes in Slavic tribe names, there was even czres – if Czrezpienianie happened not to be attested, who would’ve thought? 🙂 Apart from Slavic, there’s no nad in Balto-Slavic afaik, so if they weren’t Slavic, they can hardly be expected to use the Slavic innovative preposition. It appears Lithuanian has prie in this meaning but na isn’t a bad candidate (and closely tied to nad etymologically). The prefixes would be constricted by whatever prepositions were present to them to describe locations relative to rivers.

    I wasn’t operating by the assumption it would be a good pre-Slavic/pre-attested-Baltic name, rather a wayward branch — I think it highly probable Balto-Slavic was more varied than just Western Baltic, Eastern Baltic and Slavic before events like the Hunnic invasion and the Slavic expansion wiped out other branches.

    As for the use of the reflexes of *-los I was expecting the criticism — indeed it most often creates deverbal adjectives in IE languages — but then there’s also Latin nubilus ‘cloudy’, Greek χθαμαλός ‘low’, -alas diminutives in Lithuanian like draugalas. Admittedly I don’t know of any geographical derivates containing it. From this point of view, it would be more convincing if an adequate IE verb could be found to derive the word from. Or a common noun.

    EDIT: Andrałojć? Very strange-sounding surname, I thought it was some Serbian-Sorbian crossbreed before a quick google search told me it’s of Lithuanian origin.

  113. Lithuanian names with the (originally patronymic) suffix -aitis are Polonised with -ojć. Andralaitis is, I think, a variant of Andriulaitis, Andriukaitis (all from Andrius). In Polish the surname is a unisex one (which is why both authors of the article use the same form), but of course in Lithuanian the wife of Mr. X-aitis would be Mrs. X-aitienė.

  114. David Marjanović says:

    peripatetic narwhals

    They’re certainly causing a commotion…

  115. Even assuming for the sake of the argument that /l/-epenthesis is oder than most people think

    I was going to change “oder” to “older” but then it occurred to me it might be a riverine pun.

  116. I was going to change “oder” to “older” but then it occurred to me it might be a riverine pun.

    We have a winner!

    The Nahrwhal/Nahuatl stuff being a distant third.

  117. I was going to change “oder” to “older” but then it occurred to me it might be a riverine pun.

    That’s a Neisse pun.

  118. Trond Engen says:

    I was going to change “oder” to “older” but then it occurred to me it might be a riverine pun.

    I try to avoid those Inn jokes.

  119. David Marjanović says:

    Oh, on Proto-Germanic , I forgot a lengthy conference handout by Eugen Hill from January 2014 which argues in considerable detail that where and there – just two words, though common ones – had this rare phoneme as the result of a regular lengthening of monosyllabic words in *-r, which happened in a total of three words (the third being here, which incidentally solves the mystery of “*ē₂“). I’ve summarized it here, with a link to the original.

Speak Your Mind

*