I don’t seem to have posted about Tocharian, which is a bit surprising because it’s always been one of the Indo-European languages I’ve found most intriguing, for its unexpected location, its equally unexpected developments, and the history of its discovery. Matthew Scarborough has continued his survey of Indo-European etymological dictionaries (overview, Anatolian) with Indo-European Etymological Dictionaries for the Perplexed: Tocharian Languages, and far from being a mere list of a few books, it’s a general introduction to the field, with examples of etymologies, images of book pages, and a nice photo of an actual text. He starts by saying “I am no specialist in Tocharian and I have never studied Tocharian formally,” but he has some strong views which he makes clear, e.g.:
One of the more notable differences in the Dictionary of Tocharian B to most other works the Indo-European side of things is that Adams continues to reconstruct a fourth a-colouring laryngeal *h₄ which left no trace in Anatolian. If you’re familiar with some of Adams’s other work like The Oxford Introduction to Proto-Indo-European and the Proto-Indo-European World this oddity of reconstruction will probably not come as a huge surprise. My own hot take on fourth-laryngealism is that if you’re reconstructing a fourth laryngeal, there are a small handful of etymologies that you desperately want to be true, but the evidence is not actually bearing it out, so you reconstruct a whole new phoneme as a last ditch effort to make them work. Needless to say, it gives me the impression of a lower bar of rigour set for the acceptance of any given etymological proposal in general, and so I would tend tread a bit more cautiously with the etymological proposals made here. That said, Adams is in every respect one of the major world authorities on the Tocharian languages and he has also published extensively on Indo-European as well, so one should not be entirely dismissive of his proposals either.
So I was glad to see his post, and I hope you’ll take a moment to investigate this minor member of the I-E family.
Tocharian is indeed interesting. It will be great when/if DNA results become available from early Tocharians.
As of now, there are only old (primitive by today’s standards) results from the Tarim Basin Mummies, but there is neither evidence that these people spoke Tocharian, or that the results are actually accurate (they could be contamination).
Yes, it’s quite frustrating!
The heavenly axes of Sieg and Siegling fall on your head if you ever call the noble Tocharian languages “minor” again!
Thanks again for another shoutout on this series!
The first quoted sentence reads to me like prepositions were selected (or omitted) at random. Do others find the use of in, to and the void before ‘the Indo-European’ more felicitous than i do?
@Ryan: I think Scarborough has simply managed to use three specifically British constructions in a row.
That’s my take on it as well.
Ryan:
Vague in is exceedingly common in all forms of English nowadays, and is nothing new either. You can, for example (h/t the AHD), split something in two, be in love or in debt or in command or even in pursuit, be paid in cash, write a poem in English or in rhyme, be six feet in height, or trust in my judgment.
Different to is indeed mostly non-North-American nowadays, and though Matthew is a Canadian, he’s been living and working in the UK for some time now.
Finally, I agree that there is a preposition missing before Indo-European, most likely on. Saying most other works the Indo-European side of things is (I think) ungrammatical in every variety of English.
> most other works the Indo-European side of things
I believe NPs with “side” as head can be used to modify nouns, although I don’t think I can list the exact conditions, as in
“Best pizza this side of the river!”
When I first read the bit in question, I thought it was a case of that, or maybe a jocular extension of the construction to a context where it’s marked. By on further thought, I agree with JC it’s probably just an error.
What perplexed me more was the passage
> a fourth a-colouring laryngeal *h₄
I think he means “a fourth, a-colouring laryngeal *h₄” with a comma, right? They’re not all a-coloring, are they?
He could reconstruc’ a laryngeal just like a-colorin’ a fourth.
“… most other works the Indo-European side of things” is notably informal but nevertheless perfectly grammatical for me. Maybe it’s a US/UK thing, as Rodger C and Hat suppose.
David E: “… most other works the Indo-European side of things” is notably informal but nevertheless perfectly grammatical for me…
That is how understood it at first reading. Afterwards I thought that the lack of a preposition could be just a typing error, and was not sure of which interpretation to choose, but I can now agree in good conscience with you and with dainichi.
*sigh*
I can never get a break with my by-now very mixed spoken and written idiolect…
Correct – *h₂ and *h₄ would be a-coloring, *h₃ o-coloring, and *h₁ not coloring at all.
I can never get a break with my by-now very mixed spoken and written idiolect…
Eh, pay it no mind — people around here pride themselves on their pickiness (and sometimes prickliness), but it’s all in fun.
Scarborough writes:
The latter change, loss of *d, seems readily explicable. There are a few instances of possible *d ~ *h₁ in mainstream PIE: most famously the putative *dḱm̥tóm > h₁ḱm̥tóm > ἑκατόν (can’t say I find that one terribly convincing, though). From a glottalic perspective, we’d be talking about a debuccalization of /tˀ/ > /ʔ/, which seems reasonably natural.
But what about Tocharian *d > *ts? Does anyone know what the glottalic theorist explanation is for that?
I did propose a little while back that the mediae series could have been affricates, later fortited to stops except for **b which instead merged with *w. I don’t think I had the Tocharian data in mind when I suggested that, but it would mean that Tocharian preserves the original value for *d.
*lightbulb moment*
That would explain a whole bunch of things, you know. (At least if we interpret it as: voiceless affricates > voiceless fricatives > voiced fricatives > voiced stops.)
First of all, the lack of *b: a system with a /pf/ but no /f/ will quickly end up as the opposite. /f/ can then turn into /h/ and disappear, or it can turn into /v/ and merge into /w/ as has happened in Mandarin for example.
Second, the lack of affricates itself! It’s typologically really unusual at least for the region. All attempts to reconstruct Proto-Nostratic or even just Proto-Indo-Uralic assume that affricates were once present and were then lost en masse in PIE, often in strange ways.
Third, the strictness of the constraint on two voiced plosives in the same root could have been a constraint against two affricates at some point, which makes at least as much phonetic sense to me.
Fourth, the Moscow School has the voiced series as the result of unconditional voicing within a chain shift: *t’ t d > *t d dʰ. That’s a bit odd. Unconditional voicing of fricatives makes more sense.
Finally (for now!), as all the glottalicists etc. keep pointing out, the voiced series is rarer than the other two. That, too, fits for affricates.
/f/ […] can turn into /v/ and merge into /w/ as has happened in Mandarin for example.
I’m pretty sure that’s not what happened. Rather, the Old Chinese bilabials /b/, /p/, /pʰ/, and /m/ became labiodental /v/, /f/, /f/, and /ʋ/ (presumably nasalized) respectively. This did not happen in the branch leading to the Min languages, which lack /f/ to this day, even in borrowings from Mandarin. Then as part of the general devoicing of initials in pre-Mandarin, these became /f/, /f/, /f/, /w/. At no point did /f/ > /v/.
So, the Moscow school has the traditional tenuis series as originally ejective? That fits the Caucasian data better, maybe? Off the top of my head, I can’t remember any of that data other than mḳerdi, so I’m not sure how reliable it is.
I’m not aware of a good explanation of Winter’s law that doesn’t involve glottalization.
and *h₁ not coloring at all.
With the exception of Luwic, where *ē > *ī but *eh₁ > *ā (reconstructed as going via a PAn. *ǣ by Melchert, though I’m not sure if this is optimal routing).
I did propose a little while back that the mediae series could have been affricates
Sounds doable for synchronic typology. For diachronic typology that leaves one glaring hole however: several IE groups have spirant reflexes for either the *T series (Germanic, Iranian, *p > *ɸ in Celtic) or for the *Dh series (Italic, later Greek, Germanic), none has them for the *D series.
There are a few languages that have /tsʼ/ but no /tʼ/ (e.g. Hadza, Sandawe, Iraqw) which could also be relevant for the *d >> /ts/ shift. I don’t know if any of these examples involve an actual shift *tʼ > /tsʼ/ however. Maybe they rather involve *tʼ on one hand, *ts *dz on the other being lost in some fashion.
The Nostratic argument for the glottalic theory is based on Afrasian *Tʼ ~ PIE *D correspondences, as charted by Bomhard. (Dolgopolsky sets up instead AA *Tʼ ~ PIE *T and AA *T ~ PIE *D, which, in a display of either nondisprovability or some kind of old complexity, is not really based on substantially poorer data.)
I forgot in my enthusiasm: *h₁ḱm̥tóm would give either **ekatón or (perhaps more likely) **hikatón in Greek, but definitely not he-.
Sorry for the confusion. The appeal to Mandarin is only intended for the last step, the merger of /v/ into /w/ as in 万 wàn “10,000”, which is used for /v/ in extended bopomofo.
Kartvelian, yes.
Oh, that’s easy: it’s pretty much the same thing as prefortis clipping in English. Lengthening vowels before modally voiced plosives comes naturally to me*, and not doing it before breathy-voiced ones makes sense because the aspiration needs the extra breath for itself.
* Unlike the voiced plosives themselves, admittedly.
Couldn’t that simply mean that *ē > *ī was completed before *eh₁ > *ē happened, and the new *ē turned into *ā as the third step?
I took a quick look at the evidence for unconditioned /*d/ (> /*dz/) > /ts/ in Tocharian. On the surface, it doesn’t look very compelling. The classic example is “ten”, déḱm̥ Toch. śäk~śak. This, of course, doesn’t show /ts/ but a sibilant fricative. This fits the model but requires an additional step of palatalization. Tocharian has extensive palatalization affecting dentals and velars, but the rules don’t seem to be entirely reliable. Ringe cites as further evidence of unconditioned /*d/ > /ts/ two words that had originally *dʰ, but are posited to have first become *d via Grassmann’s law, and thence /ts/. But in both of those cases, /ts/ precedes /e/, so the same palatalization affecting śäk~śak would be expected. Ringe must rely on analogy to explain that anomaly.
I looked up IE roots in *d in Wiktionary and found two with reflexes in Tocharian (*déḱm̥ is mysteriously absent): the “dexter” root, Toch. täk; and the “domus”~”timber” root, Toch. tam~täm. In both of those cases, /*d/ > /ts/ as well as palatalization both fail.
If we conclude that, in our current state of knowledge, Tocharian palatalization is not entirely predictable, then instances of ts or ś from *d can be explained as unusual palatalized forms.
Interesting. It’s always presented as textbook knowledge…
It will be great when/if DNA results become available from early Tocharians.
Not a year later you got you wish granted. Not quite Tocharians but a population which predated them by 7 to 9 centuries in Shirenzigou in the Far Eastern fringes of Tian Shan, and showed cultural similarities with both Yanbulak culture of the Tarim Basin and with the Pazyryk mound motifs in Altai to the North.
Shirenzigou show DNA continuity with both today’s Uyghurs and with the Afanasievo culture of the broader Altai, and clearly aren’t descended from Andronovo / BMAC / proto-Iranian populations. Rather, their Steppe DNA is traced back directly to Yamnaya people, before the back-migrations from Europe which occurred in the European Steppes after the Afanasievo peoples left and went East.
Today’s Uyghurs have some Iranian ancestry in their DNA as well, but it wasn’t yet the case in II .. IV centuries BCE in Shirenzigou.
The various layers of Asian ancestry were already abundant in Shirenzigou, ranging from 30% to over half. Most of it came from the nomads of South Siberia / NE China, but some came directly from the Chinese.
https://www.cell.com/current-biology/fulltext/S0960-9822(19)30771-7
“Besides the cultural exchange between the eastern Tianshan Mountains and the Altai region, a special funeral ritual appeared in the Shirenzigou site, whereby the upper torsos were disturbed. This was a common custom in the Neolithic Gansu corridor, in Northwestern China [18].”
A stake through the heart?
A genetics blog suggests that the Shirenzigou people had way too much East Asian, especially North-Eastern, DNA to be plausible ancestors of the Tocharians … but that they are quite similar to the Huns whose DNA from the following centuries from the nearby Tian Shan regions has been published earlier.
http://eurogenes.blogspot.com/2019/07/they-mixed-up-tocharians-with-huns.html
So we may be back to square one with respect to the genetic roots of the Tocharian speakers.
Even so, their West Eurasian component is pure Pit Grave without any European Farmer admixture, so they’re still evidence that the same holds for the actual Tocharians.
And what does “Huns” even mean…
I’ve now read the post and all comments. While the author’s speculation about Botai ancestry is interesting and should be tested at some point, I can’t distinguish the paper’s result from “the ‘Tianshan Huns’ had Pre-/Para-Tocharian ancestry”.
The “way too much East Asian” part is apparently argued from nothing more than a painting with pretty much destroyed faces showing beardless men.
From the post:
“Here’s a famous wall painting of Tocharian princes from the cave of the sixteen sword-bearers in the Tarim Basin, dated to 432–538 AD. They don’t look like guys with a lot of Ulchi-related admixture to me, but I might be wrong.”
They’re beardless, and their faces are pretty much destroyed, so I’m not sure what the author is even talking about. His speculation about Botai-related ancestry is interesting and should be tested, but so far it’s still just speculation…
The Xiaohe mummies, presumably considerably older than that, were already described as Asian-European admixed (based on uniparental markers, so the composition and timing of admixture is an open question), and their facial features are compatible with the idea of mixed ancestry, right? Eastern Scythians had substantial Asian / NE admixture as well although they were a part of the cultural continuum which spread all the way through Pannonia.
If the bones of historical-period Tocharians are processed for DNA, we’ll doubtless find some Iranian-like ancestry. And we’ll face unresolved questions about the link between it and the language…
Hello chaps,
Unfortunately, those Shirenzigou nomads aren’t relevant to the Tocharian question. There are a number of reasons for that, including…
– they appear to be a group directly ancestral to the Tian Shan Huns
– their R1b subclade doesn’t seem to be of the Indo-European, Afanasievo and Yamnaya type, but rather a Hunnic-specific lineage that was probably native to East Central Asia
– the genome-wide ancestry of the Tian Shan Huns is easily modeled with formal methods as derived from Andronovo groups, early Central Asian farmers and East Asians, and this is also likely to be true for the Shirenzigou nomads
So I’d say that Ning et al. should have put out a paper on the origins of the Hunnic confederacy, and in particular the Tian Shan Huns, rather than enter the Tocharian debate without any ancient samples from sites in the Tarim Basin where Tocharian was attested.
They also should have used Eneolithic Central Asian farmers in their mixture models, because not doing so for Central and South Asian test samples often makes it very difficult to distinguish between steppe admixture sources that are poor and rich in early European farmer ancestry, like Afanasievo and Andronovo samples, respectively.
In other words, the chance that the Shirenzigou nomads were significantly derived from the Afanasievo people and thus closely related, let alone ancestral to Tocharians, is slim to none.
Hey, I had looked forward to reading that paper when I got home to my laptop, and now it’s shot down.
Davidski: So I’d say that Ning et al. should have put out a paper on the origins of the Hunnic confederacy, and in particular the Tian Shan Huns, rather than enter the Tocharian debate without any ancient samples from sites in the Tarim Basin where Tocharian was attested.
That sounds like a peculiar narrowness of scope. Any reason for that? With an almost entirely Chinese team I hope it’s not European bias in favour of Indo-European genetics. But still, could it have played a role, e.g. when formulating a topic of study in the proposition for collaboration with the Max Planck Institute in Jena?
They also should have used Eneolithic Central Asian farmers in their mixture models, because not doing so for Central and South Asian test samples often makes it very difficult to distinguish between steppe admixture sources that are poor and rich in early European farmer ancestry, like Afanasievo and Andronovo samples, respectively.
Would it be difficult to rerun the analysis with Central Asian Farmers in the model?
This reminds me of another question I have had brewing for a while. As the amount of avilable data on ancient populations increase, should studies from a few years ago be rerun and updated? Or are the implications of new data for the first studied populations always clear for those in the field?
Trond Engen,
I guess the authors were going for the strongest angle they could find and it seems they overshot their mark. A paper about the origins of Huns would’ve been fine, but I guess Tocharians are sexier than Huns, or something?
Eventually the genotype data for these samples will become available and it’ll be possible to check whether the Shirenzigou nomads got their steppe ancestry directly from Afanasievo or indirectly from Andronovo. But Andronovo looks a much better bet.
Also, I can tell you that at least one team of scientists will be looking at this issue very closely and you’re likely to see the results in a new paper soon.
Great, let us know when/if it appears!
I take all your other points, and of course I’m looking forward to the next paper. But this point strikes me as wholly irrelevant: they could, if we overlook your other points, simply be the descendants of Pre-Tocharians that stayed behind when everyone else went off to settle in the Tarim Basin, and eventually ended up contributing to the formation of the Huns. (Or whatever a “Tianshan Hun” really is.)
David Marjanović,
The Tian Shan Huns aren’t related to Afanasievo, but to Andronovo, and this is very important, because the Shirenzigou nomads can only be related to the Huns or to Afanasievo, not to both. The two things are mutually exclusive.
Of course, once the link between the Shirenzigou nomads and Afanasievo is debunked, then there’s no way, for now, to link the Tocharians to Afanasievo with ancient DNA.
And yes, the Shirenzigou nomads, Tian Shan Huns, Tian Shan Saka and other closely related groups with steppe ancestry in the region might be the descendants of Pre-Tocharians. But there’s no evidence for this either, and it’s not what the Ning et al. paper was arguing anyway, for one because it excludes Afanasievo from the picture.
I see, thanks.
The Tian Shan Huns aren’t related to Afanasievo, but to Andronovo
Isn’t it the biggest point the Ning paper tries to make, that in their analysis the new DNA samples completely lack European farmer ancestry, and therefore may be related to the early migrants East of Afanasievo but not to the later-origins Andronovo?
I understand that their full data isn’t available, but the Tian Shan Huns data is, and thectwo groups partly overlap on a principal component plot. But does it make you sure that the minor farmer derived component must be present in both groups?
Even Yamnaya had some Early European Farmer ancestry, mostly ultimately from Anatolia. I think the point is that without Central Asian Farmers in the model of the Shirenzigou nomads’ ancestry, it’s impossible to tell how much of their Anatolian ancestry that came from Europe, and without that you can’t tell Andronovo apart from Afanasievo. (Well, I think it should also be possible to look at the Non-Anatolian European element.)
here are only old (primitive by today’s standards) results from the Tarim Basin Mummies, but there is neither evidence that these people spoke Tocharian, or that the results are actually accurate (they could be contamination).
This thread seems to be a proper place to cite the new paper on Xiaohe horizon Tarim basin mummies DNA.
https://www.nature.com/articles/s41586-021-04052-7
These people turned out to be completely unrelated to the Indo-European peoples to the North or West of them, and mostly composed of the Ancient North Eurasian ancestry stream with some East Asian ancestry added to it. In contrast, earlier Bronze Age people from the Northern foothills of Tian Shan were indeed, as suspect, strongly genetically linked to Afanasievo culture (also with some added Ancient North Eurasian ancestral component).
In other words, Middle Bronze Age Tarim Basin appears to be culturally, but, surprisingly, not genetically influenced by the Indo-European steppe herders which preceded it by a few centuries in the same general region but on the other side of the mountains. No need to remind you that we don’t know what happened in the huge time gap between the Tarim mummies and the Tocharian era.
And, yes, not all of the early DNA results got confirmed. In particular, the much-discussed R-haplogroups of the Tarim Y-chromosomes turned out to be not of the Yamnaya sub-branch, but more ancient in origin, ultimately related to the same Siberian hunters-gatherers who passed a different sub-branch of it to the Yamnaya as well…
Fascinating! It’s amazing to me that we’re suddenly discovering so much about these ancient peoples.
Fan Zhang et al (2021): The genomic origins of the Bronze Age Tarim Basin mummies (linked by Dmitry above).
The discussion happened here.
It’s amazing to me that we’re suddenly discovering so much about these ancient peoples.
Not just ancient ones, and not just peoples.
I occasionally look at breitbart on my cellphone, to see what brain chow is being dished out to overexcited parts of the USA. This week there were large ads there for an “Illumina sequencer” (they don’t appear in my notebook Chrome), unlike the usual ones for stiffening agents, knives and camouflage drag.
The Illumina website sez:
# Our innovative next-generation sequencing (NGS) platforms deliver exceptional data quality and accuracy, at a massive scale. View benchtop and production-scale sequencer comparison tables, and find tools designed to help you choose the right platform for your needs. … (human, plant, animal genomes, exomes …) #
Why does anyone think right-wingers would like to buy a DNA/RNA sequencer ?? What would they get up to with one ? Prove that Trump is descended from the messiah ?
I shudder at the mischief that has already been done with home etymology kits.
Why does anyone think right-wingers would like to buy a DNA/RNA sequencer ??
To prove to the purchaser (and his* friends, associates and local Ortsgruppenleiter) that he is entirely free of all forms of untermenschlich genetic contamination, I would imagine.
Any enterprising “genetics” company could surely provide a service guaranteed to satisfy such a clientele, at minimal expense to the provider. Hell, now I’ve thought of it, it seems like a colossal business opportunity just waiting to be taken …
* Of course. Though he might buy one for his SO. She better pass …
Stu: Why does anyone think right-wingers would like to buy a DNA/RNA sequencer ?? What would they get up to with one ? Prove that Trump is descended from the messiah ?
I don’t follow the market for personal tests, but I think Illumina must be a new provider and may have a limited marketing budget. I would guess they made a survey for interest in buying a DNA test from an unmerited company in correlation with media habits, and conclued that Breitbart gave most clout for the money. That doesn’t answer the question, though. It could be demographic — maybe a combination of age, purchasing power and education.
Are you sure it’s aimed at right-wingers and not at you personally? Are you now reading, or have you ever read, genetics papers on your phone?
I’ve never seen an extreme-right reaction to the genetic discoveries of the last 10 years, actually – that the White Race is a mixture of four components, that blue eyes are native to Europe but light skin and light hair are mostly not, or that what could have been called “the purest Aryans” are the Estonians…
(…whose language has a vowel even more nasty and Asiatic than Russian…)
David M.: Are you sure it’s aimed at right-wingers and not at you personally?
Good point.
I’ve never seen an extreme-right reaction to the genetic discoveries of the last 10 years, actually – that the White Race is a mixture of four components, that blue eyes are native to Europe but light skin and light hair are mostly not, or that what could have been called “the purest Aryans” are the Estonians…
Perhaps not. But I have the clear impression that genetic nationalism these days have recalibrated and defend a vulnerable, delicately balanced hybrid.
Are you now reading, or have you ever read, genetics papers on your phone?
Certainly not ! I take twenty winks even when such subjects come up here, the only possible source of interest inference. I regularly delete all data on my phone, and look at breitbart at most once a month.
I’ve never bought a stiffening agent either. Maybe the advertisers are scraping the bottom of the barrel to find something I might buy.
a vulnerable, delicately balanced hybrid
Fat, ignorant and angry. Not much vulnerability there. Relax, dollinks !
It’s certainly possible that Google Ads doesn’t find enough of a market for its sequencer ads, so it justs foists them on everyone whose desires aren’t known in useful detail…
Youtube has now given up on convincing me to spend money on getting to know fascinating Slavic ladies and wants me to date single moms in my local area. Maybe I should google something about mattresses to steer it back to shallower waters.
Or I should get around to deleting my Google account and watch Youtube anonymously with a VPN and an adblocker that works. It’s not hard, but then neither is doing the dishes. Or not watch Youtube, you say?
large ads there for an “Illumina sequencer”
Maybe looking for buyers in law enforcement agencies? It isn’t your typical garage toy, they are expensive, the consumables are expensive too, and it’s hard to make productive use without added bioinformatics expenses.
… Why does anyone think right-wingers would like to buy a DNA/RNA sequencer ?? What would they get up to with one ?
My upbringing taught me that rightwing culture has deep roots and can be very sophisticated; Orwell has related things to say about doublethink. Genomics these days is a playground and playgrounds can attract bullies and trolls. `Who knows what Evil lurks in the mind of man?’ says Lamont Cranston.
\On a lighter note see for ex
https://en.wikipedia.org/wiki/Orphan_Black
https://en.wikipedia.org/wiki/They_Saved_Hitler%27s_Brain
https://en.wikipedia.org/wiki/That_Hideous_Strength
Dmitry: Maybe looking for buyers in law enforcement agencies?
Yes, there’s demographics for you.
Oh, that makes some sense.
I never get any ads on YouTube. The trick seems to be to switch “acceptable ads” off; that means “ads that Google, owner of YouTube and oh BTW Chrome, wants you to find acceptable”. Fuck this waste of bandwidth and electricity, I’m not going to buy anything anyway!
They Saved Hitler’s Brain! That takes me back. “Mach schnell! Mach schnell!!”
I would not have expected to see They Saved Hitler’s Brain and That Hideous Strength together.