Svenja Bonmann and Simon Fries, “Linguistic Evidence Suggests that Xiōng-nú and Huns Spoke the Same Paleo-Siberian Language” (Transactions of the Philological Society, 16 June 2025, open access):
Abstract
The Xiōng-nú were a tribal confederation who dominated Inner Asia from the third century BC to the second century AD. Xiōng-nú descendants later constituted the ethnic core of the European Huns. It has been argued that the Xiōng-nú spoke an Iranian, Turkic, Mongolic or Yeniseian language, but the linguistic affiliation of the Xiōng-nú and the Huns is still debated. Here, we show that linguistic evidence from four independent domains does indeed suggest that the Xiōng-nú and the Huns spoke the same Paleo-Siberian language and that this was an early form of Arin, a member of the Yeniseian language family. This identification augments and confirms genetic and archaeological studies and inspires new interdisciplinary research on Eurasian population history.
[…]Conclusion
Our investigation has shown that (a) there are several Old Arin loanwords in Proto-Turkic and Proto-Mongolic, (b) the Jié couplet, Xiōng-nú titles and glosses betray Arin features and thus probably reflect an old form of Arin, (c) Hunnish personal names likewise seem to be Arin in origin ultimately and (d) the Yeniseian hydronyms and hydronym-derived toponyms along the westward migration route of the Huns are predominantly Arin suggesting a correlation between speakers of Arin and the Huns. In a variation of a word by the master detective Sherlock Holmes (in the short story ‘The Adventure of the Devil’s Foot’) it can therefore be established in our view that while each of these pieces of evidence are suggestive, together they are conclusive, because they independently corroborate the implications of each other.
It therefore seems an inevitable conclusion to us that Huns and Xiōng-nú both spoke the same early form of Arin that we have tentatively termed Old Arin here and that consequently the linguistic and thus most probably also the ethnic core of the Huns derived from the Xiōng-nú. These findings corroborate recent archaeological and genetic findings and show that the application of the methodology sketched out here can lead to substantial insights into the linguistic history even of regions such as Inner Asia that are at present underresearched and the history of which is much less perfectly understood than that of many other parts of the world such as Central Europe or the Mediterranean. It is to be hoped that future archaeological excavations may uncover autochthonous texts of the Xiōng-nú or the Huns (perhaps in the recently identified Xiōng-nú capital Lóng Chéng or in southeastern Europe) that allow for further testing of our Old Arin hypothesis. The synthesis of historiographical, archaeological, genetic and linguistic data and the continued application of the methodology presented here will then hopefully gradually lead to an ever deeper understanding of the linguistic history of Inner Asia and similarly underresearched parts of the world so that one day we can draw a consummate picture of the linguistic evolution of mankind.
Wikipedia has a pretty thorough article on Arin; Svenja Bonmann turned up on LH a couple of years ago in relation to the Kushan script. This is exciting stuff! Thanks go to Y, who sent me the link and added “So Turkish göl ‘lake’ is ultimately a Yeniseian loanword?”
I expected this paper to swiftly turn up here because its claim is so sensational. However, I see a number of telltale signs of insufficient peer review here. For example, the authors speak of “a borrowing from Arin into Proto-Uralic”, but there is a time gap of two millennia because their early Arin they talk about and Proto-Uralic.
They also talk about a sound shift “*š > *l” in Chuvash, and base their claim of a direction of borrowing for ‘silver’ from Arin to Turkic, but reconstructions of Proto-Turkic have commonly operated for years with a phoneme *ɫ that gave two different outcomes across Turkic. Good peer review would have demanded that the authors confront Antonov & Jacques’ paper “Turkic kümüs ‘silver’ and the lambdaism vs sigmatism debate”.
All in all, I frankly don’t feel that this paper is worth very much, except for a way to draw more attention to Yeniseic which is always good.
Well, rats. But as you say, it’s good to have more attention to Yeniseic!
Having just read the paper, I was about to look where on LH I can find the best place for the discussion … I vaguely remembered that Yeniseic and the Jie verse have been discussed here before. And lo and behold, here is the new thread and a substantive review – thank you!
PS: Antonov & Jacques 2011 is available here
https://shs.hal.science/halshs-00655014/document
Funny that in their “variation” of the Sherlock Holmes quote, they take out the detective’s careful qualification:
Why do they think they will find texts written in Hunnic? I would have thought the Huns would not have written down their oral poetry or other intrinsically Hunnic stuff, using instead (a) Greek/Roman/Chinese/whatever scribal slaves to write down proclamations in the regionally used written language for the corresponding conquered peoples and (b) horse couriers to convey oral instructions and take oral messages from army commanders. Maybe they should look instead for these proclamations and see if there are any Hunnic features, esp. names that bolster their theory,
According to Priscus, Attila’s Huns widely used Latin for trade and international communications, while otherwise relying on Gothic and Scythian languages for communications with the allied and subjugated peoples?
Generally from the descriptions of the Huns’ life in the travelogue of Priscus, one wouldn’t even guess that they were anything but regular Scythians of the land.
There are countless personal names of the Huns and I’m a little surprised why just 3 of them made it to the new paper’s authors’ analysis.
BTW another recent study by Bonmann et al., “Towards a New Reconstruction of the Proto-Yeniseian Sound System. Part I: Word-Initial Consonants” (here, paywalled), if I recall correctly, invalidated many of the etymologies proposed to link Yeniseian and Dené.
Christopher Culver wrote:
>the authors speak of “a borrowing from Arin into Proto-Uralic”, but there is a time gap of two millennia because their early Arin they talk about and Proto-Uralic.
Perhaps I’m misunderstanding something, but a search of the linked paper for the letter sequence uralic turns up only one instance, which doesn’t seem to address any such borrowing:
>Yeniseian languages are usually considered remnants or survivors of the original linguistic diversity of Siberia, historically spoken in retreat areas as the result of several waves of superimposition or displacement by expanding Uralic/Samoyedic, Turkic and Tungusic languages.
The only protos mentioned in the paper are proto-Turki, proto-Mongolic and proto-Yeniseian.
The early Arin/old Arin of the authors seems intended to describe a language spoken in the first centuries CE. What are the hypothesized ages of proto-Turkic and proto-Mongolic? Wiki offers that proto-Turkic might have been spoken 2,500 years ago, and proto-Mongolic is “verly close to the Middle Mongolic language” spoken at the time of Genghis Khan. So neither seem to be 2,000 years before the hypothesized old Arin.
I’m also not finding this reference to Chuvash in the paper:
>They also talk about a sound shift “*š > *l” in Chuvash
Mr. Culver, are you referring to another paper by these authors that makes you skeptical of their linguistic knowledge, or what am I missing?
Huh. Maybe CC assumed this post was about a different paper he’d read recently.
I missed the deadline for erasing that last comment. I realize now that there is a sideways reference that doesn’t use the word Chuvash.
Also, now that I follow Dmitry’s link and learn the title of Antonov and Jacques’s paper*, it does seem troubling that the current authors don’t deal with it when their argument depends in significant part on Turkish kümüš and lambdaism. Hmm.
* – Their paper is called “Turkic kümüš ’silver’ and the lambdaism vs sigmatism debate”.
The P-U is from a footnote in a prepublication version (fn. 9, p. 12): “We would like to thank a reviewer for pointing out to us the similarity of Arin kus ‘birch’ with Proto-Uralic *kuće ‘birch’. This might either be a coincidental similarity or reflect a loan from Old Arin into Proto-Uralic.”
That was removed from the final version. It sounds like a disposable, incidental afterthought.
The Chuvash is referred to as Bulghar.
The new paper also makes no reference to the Zeng / Viazov paper we discussed here (mostly about Seima-Turbino and Uralic) that highlights “cis-Baikal late Neolithic/Bronze Age” genetic markers as “a tracer-dye for prehistoric mobility associated with the spread of Yeniseian languages.”
Zeng, et al., don’t mention finding this ancestry in any Hunnic or Xiong-nu genomes that were available to them. It’s not clear what is available, so maybe there could be a relationship that Zeng et al. didn’t know about. But Bonmann and Fries boldly state that their work “confirms genetic and archaeological studies” without any reference to what seems to be the signal work on Yeniseian prehistoric genetics.
lambdaism
Their [Antonov & Jacques 2011] paper is called “Turkic kümüš ’silver’ and the lambdaism vs sigmatism debate”
This?
Or lambdacism?
The paper’s title/header page at HAL Open Science is c-less, but the very next page’s title and throughout the body is c-full. ‘Sibilantlessism’?
Gammacism Stigmatism?
The printers’ devil has escaped from that other thread?
> The printers’ devil has escaped from that other thread?
LOL. I had merely cut and pasted without noticing the inconsistency.
As mentioned, it is in footnote 9 in the PDF. Strange that Wiley would have different content in the plain HTML text and in the linked PDF.
Ah! The footnotes are present in the “Full Text“ version, but because of the way they’re presented as pop-ups, they can’t be searched.
So the authors did include this nonsensical reference to a loan from early Arin into proto-Uralic. Hoo boy.
*ɬ, rather; a lateral fricative that merged into *l in West/Uncommon Turkic but became [ ʃ ], Arabic-style, in Rest/Common Turkic.
(For generations there was a major conflict between “lambdacists” and “sigmatists” in Turkology. Likewise “zetacists” and “rhotacists”… all quasi-ideologically charged because the classical Altaicists were sigmatists and zetacists.)
Probably those are the only 3 that make any sense.
Mind you, that’s up from 0 previously.
Here.
Antonov and Jacques recognize the ‘silver’ word as a Wanderwort (all the way to Austro-Asiatic), but don’t come up with a source. Bonmann et al.’s argument is good, on its face: the word, with the meaning ‘tin’, is reflected with an initial k- in Arin, and with a zero in Ket, Yugh, and Assan. This is a regular correspondence, from which Bonmann et al. had reconstructed PY *w or so (going to Arin k via *gw). It’s hard to reconcile this with anything but an Arin-like source.
It’s too bad that they don’t address Antonov and Jacques, but that doesn’t invalidate the whole paper.
N.B. The paper has abstracts in English, German, Latin, and Greek. Other recent papers in Trans. Phil. Soc. have an English-only abstract, and one peper by an Italian author has abstracts in English and Italian. I am mystified.
Y, the authors write on page 11: “A form *kumiš ‘silver’ (> Arin kumɨš) was borrowed into Proto-Turkic”. But, as ought to be well known to anyone working with the diachrony of the Turkic family, the original final consonant of the Turkic word likely wasn’t *š but *ɫ. So, in this light the Arin word with final -š looks more like a borrowing from Common Turkic into Yeniseic than a Yeniseic borrowing into Proto-Turkic. That’s pretty devastating in my book, and moreover, it’s just one of many omissions that led me to finding this paper unconvincing. That’s not to say that the thesis that the Huns spoke Yeniseic etc. isn’t true, only the argument will need to come from some future paper written with more rigor, and peer-reviewed better.
Well, maybe Pre-Arin [ ʃ ] was borrowed as Proto-Turkic [ɬ] because that was the most similar sound available (cross-linguistic parallels for this would be nice, though); or it was borrowed into Proto-East Turkic (aka Common Turkic), and then etymologically nativized into West Turkic and Mongolic with [l] (and spread from there).
However, wouldn’t Turkic have borrowed [u…ɨ] as such instead of fronting and further-harmonizing it to [y…y]? That seems like a strong argument for borrowing in the other direction (from East Turkic to Arin). And [u…i] would likely have ended up as [u…ɨ] in Turkic as well, wouldn’t it?
@CC: Indeed, but then how would you explain the initial k- in Arin but not in the other Yeniseian languages? From the little I can tell, both your argument and this one are significant, and they should be reconciled somehow.
Also, the non-Arin languages have an l instead of a sibilant, e.g. Ket umɔl’a. The authors don’t touch on that as an explanation for the varying loans into Turkic, nor do they offer a good Yeniseian-internal explanation for the variation (see fn. 8, p. 11). I don’t know what to make of all that.
Victor Mair also just posted about this. It’s a very Mair post, but not in the way you might think.
For anyone who doesn’t feel like wading through Mair’s prolix prose, here’s the nub of it:
I’m not sure he grasps that language need have nothing to do with physical culture.
Or that DNA is physical and not mathemagical.
“The Xiōngnú were not hunter-gatherers and fishermen of the Yenisei Valley. I am amazed and dismayed that the linguists who propose that the origins of Xiōngnú language are to be found in Ket, Yeniseian, or other Paleo-Siberian language are oblivious to this basic reality of existence and ecology.”
Similarly, the Welsh are not horse nomads of the Pontic-Caspian steppe. It is extraordinary that anyone could seriously maintain that Welsh is an Indo-European language in view of this basic reality of existence and ecology.
Also
https://en.wikipedia.org/wiki/Classification_of_Pygmy_languages
But in the face of such radical misunderstanding, mere evidence, however copious, is powerless.
There’s a difference. I suspect Mair’s gut reaction is against the idea that the Xiongnu would ever again “revert” to hunting and gathering.
(To be clear, though people still use the term “cultural reversion” in this day and age, Mair has not explicitly done so that I know of.)
I note Mair’s characteristically crass and ill-informed response to DM’s entirely reasonable comment on LLog. I don’t think there is any point in trying to engage with such people.
However, he is evidently mellowing. When I commented on one of his posts* that his source (the BBC) might not be 100% correct on the linguistic issue, Mair simply deleted my comment. (It was subsequently rescued by Mark Liberman, who made a post of his own about it, and then restored in situ by Mair.)
* I’m never going to do that again.
Mair makes my eyes hurt on the inside, so I only go on LL if Hat tells us there is something worth looking at.
I still read Liberman’s Language Log posts, and Mair’s if they look like they might be interesting (which is a decided minority). I just keep in mind that Mair is thin-skinned and prone to delve into areas that he knows little about.
If I may venture into analyzing* the behavior of two people I don’t know, I think it is interesting to compare how Pullum and Mair responded to losing their spouses, through the lens of Language Log. Pullum gradually pulled back from the project, seeming to find that posting and (even more) dealing with comments was not worth his time, which could be spent on other things. Life’s too short, and all. Mair, in contrast, threw himself headlong into posting on Language Log, perhaps feeling that it could help promote his late wife’s long-running project of supporting a rationalized transliteration of Sinitic languages.
* I heard someone use psychoanalyze recently to mean “analyze someone’s behavior from a psychological viewpoint,” and I had to object to misuse of the neologism. My grandfather may have been disrespected as a theoretical Freudian psychoanalyst, but he was a talented clinician and medical doctor, who believed in the value of classical psychoanalysis to treat patients dealing with past trauma. (His specialty was treating adult patients who had lost a parent or close friend in childhood.)
Idly scrolling through recent LLog posts, I see that Prof Mair is sufficiently terrifying in person that a brave TSA functionary stole his walking stick in order to prevent him hijacking an aeroplane with it. (This appears to have been long before the current autogolpe.)
I shall accordingly only make nice comments about him henceforward.
In case anyone’s wondering, I haven’t seen it yet (and won’t this evening). I’ll reply if I haven’t been banned.
if I haven’t been banned.
Your comment is still there. My experience is more like DE’s: the comment gets deleted with no response or explanation. (But I haven’t been banned — at least not yet.)
I think that another great weakness of Bonmann & Fries’ paper is a lack of a dating, however speculative or tenative, for Proto-Yeniseic. The argument in the paper also seems to assume that the Yeniseic family had broken up into its constituent branches before the migration northwards, and then multiple branches made the trek north. More should have been said about this.
I’ve had the experience that comments of mine were gone the next time I visited the page and then reappeared a few days later with no comment. I think that’s just the spam filter.
That’s consensus, AFAIK. One strange thing about the paper is how much background knowledge it assumes.
Are there estimates of the time of the breakup of Yeniseic other than by glottochronology?
I was thinking of hydronyms. Yeniseic hydronyms, including some (in contiguous geographic areas) that show sound shifts which are lacking in the attested languages, cover a pretty large area in south-central Siberia. Irtysh is one. I can’t look up the sources this or next week.
I haven’t been banned, and the response is much less bad than I feared, so I replied yesterday. No reaction yet, though someone else brought up genetics.