The Tocharian Trek.

As I said a few years ago, Tocharian is one of the Indo-European languages I’ve found most intriguing; now Ali Jones at writes about a very promising project called TheTocharianTrek:

The research is helping to pin down where the Tocharians were located in the period between 3,500 BC, when they may have left their ancestral home, and their first written history in 400 AD. In sum, the initiative is mapping the migration route from the PIE homeland all the way to China.

Through the journey, the Tocharians brought their dialect of PIE into contact with people speaking different languages. This influenced and changed the way the Tocharians spoke until finally their recorded languages evolved. Archaeological and genetic evidence suggests that the Tocharians first moved to southern Siberia.

[Professor Michaël] Peyrot and his research colleagues have sought to provide a linguistic assessment of this route. Their work reveals that, indeed, some of the quirkiest features of the language fit very well with tongues spoken in southern Siberia.

“Languages preserve precious information about their prehistory through the effects of language contact,” said Peyrot. “Observing the effects of language contact, such as borrowed words, enables us to draw conclusions about the proximity of the speakers of different languages and at which point in time the contact took place.” As an example of a borrowed word, he cited a term for sword in a language strand known as Tocharian B: “kertte” was taken from “karta” in Old Iranian.

The research team has concluded that the Tocharians arrived in the Tarim Basin in around 1,000 BC—later than was previously thought. As result, their window of influence in the Tarim Basin has narrowed and the Tocharians are being assigned a more muted role in the prehistory of the area than they have traditionally been given.

Instead, the project has found a strengthened role for Iranian languages and peoples in the area, especially Khotanese, its relative Tumshuqese and Niya Prakrit. All influenced Tocharian.

The project is also piecing together which languages left the PIE community first and when. As their work enters its final phase, the researchers agree with the theory that the Tocharians may well have left the PIE family second and certainly well after the Anatolians, a group of ancient languages once spoken in present-day Turkey.

The piece goes on to discuss weather terminology and says:

The ultimate goal is to create an atlas that maps where the words were used and when. The completed atlas is due to be available on the university’s website beginning in late 2023.

Exciting stuff — thanks, Dmitry!


  1. This study by Peyrot, from 2019, argues for contact between Proto-Tocharian and Proto-Uralic, on typological grounds. I would be curious to know what Uralists think about it.

  2. The general gist of Uralic influence seems to hold up, I haven’t seen anyone outright question it, though many details remain debated.

    Tocharian loanwords in Samoyedic remain probably underresearched; there’s been a decent number of recent proposals from TTT member Abel Warries and I’ve got a handful of unpublished finds myself around too. So far no finds of loanwords in the opposite direction though, interestingly enough! but maybe more lexicographic work will help with that.

    Something getting from Tocharian even into Ugric (or into Hungarian, Mansi, Khanty separately) could be possible, has really not been checked systematically by anyone; there seems to be nothing already into Proto-Uralic though.

  3. Trond Engen says

    @Y: Thanks!

    The uralicist hath spoken. I’ll just for the sake of discussion say that it’s interesting, but it’s based on internal reconstruction and should be backed up by other evidence. Another thought is that this looks like a Sprachbund with Yukaghir as a more peripheral member. One objection would be that it’s hard to imagine especially Yeniseian going through a Sprachbund phase without clear traces of syntactic borrowing or morphological leveling.

  4. Christopher Culver says

    I vaguely recall some scholarship which suggests that Yeniseian was spoken considerably further to the south in prehistory. In that case, there would be no need for it to be immediately adjacent to pre-Tocharian and pre-Samoyedic during their time of contact.

  5. David Marjanović says

    argues for contact between Proto-Tocharian and Proto-Uralic

    No, between Pre-Tocharian and “an early form of Samoyedic”. (Open access! I read the paper a few years ago and recommend it.)

    One objection would be that it’s hard to imagine especially Yeniseian going through a Sprachbund phase without clear traces of syntactic borrowing or morphological leveling.

    When the grammar is just too different, such things may not happen. Basque and the surrounding Romance languages have basically been spending the last 2000 years approaching their sound systems to each other’s. Grammar? Nope. OK, Basque has been emphasizing “and” (ta) over the comitative case in the last 3 or so centuries, but if that’s all…

    I vaguely recall some scholarship which suggests that Yeniseian was spoken considerably further to the south in prehistory.

    I haven’t seen that outside of attempts to tie to, specifically, Burushaski and often also to the archaeological Karasuk culture.

  6. Trond Engen says

    David M.: When the grammar is just too different, such things may not happen. Basque and the surrounding Romance languages have basically been spending the last 2000 years approaching their sound systems to each other’s. Grammar? Nope.

    Point. It also occured to me that if Yeniseian instead is a common substrate, carrying the sound system over would be expected, bringing morphology along well nigh impossible.

  7. Trond Engen says

    More on Afanasievo and Tocharians, starting with the comments:

    Indo-European and the Yamnaya
    They Perished Like Avars (a comment I meant to leave in another thread)
    Son of Yamnaya (obviously)

  8. Trond Engen says

    A study from last year that I hadn’t seen:

    Kumar et al: Bronze and Iron Age population movements underlie Xinjiang population history, Science (2022)

    The Xinjiang region in northwest China is a historically important geographical passage between East and West Eurasia. By sequencing 201 ancient genomes from 39 archaeological sites, we clarify the complex demographic history of this region. Bronze Age Xinjiang populations are characterized by four major ancestries related to Early Bronze Age cultures from the central and eastern Steppe, Central Asian, and Tarim Basin regions. Admixtures between Middle and Late Bronze Age Steppe cultures continued during the Late Bronze and Iron Ages, along with an inflow of East and Central Asian ancestry. Historical era populations show similar admixed and diverse ancestries as those of present-day Xinjiang populations. These results document the influence that East and West Eurasian populations have had over time in the different regions of Xinjiang.


    Beginning with ancestral sources traced to indigenous ANE-derived and BA western and eastern Steppe populations, Xinjiang population structure can be characterized by waves of incoming gene flow and admixture from surrounding populations adding to the extant ancestry. The BA Xinjiang region contained four major ancestries, which included Tarim_EMBA1 (Xinj_BA1_TMBA1) (16), Afanasievo (Xinj_BA5_oAfan), Northeast Asian (Xinj_BA7_oEA), and BMAC (Xinj_BA6_aBMAC), with Tarim_EMBA1 possibly being indigenous to the region given its presence among diverse BA individuals (Fig. 3B). The Mongolian Chemurchek inhabitants near the Altai region were further linked to Xinjiang through the Chemurchek culture of BA northern Xinjiang, demonstrating the BA movement of people across the Altai region. Thus, we not only find support for both the Steppe and Bactrian oasis hypotheses (5, 13–15), but the identification of additional ancestries suggests further complexity of the EBA populations in Xinjiang. Additional sampling of pre-BA and EBA populations will be necessary to further characterize the succession of ancestries established in Xinjiang during this time period.

    Later, in the LBA, ancestry present in the Central Asian BMAC populations becomes more pronounced, which was likely to have entered Xinjiang though the IAMC route along with the Steppe_MLBA populations, such as the Andronovo, Sintashta, and Dali (Botai related ancestry) populations (Fig. 3B). The entrance of Steppe MLBA (~3900 B.P.) into Xinjiang correlates with the arrival of the eastern Fedorovo subculture of the Andronovo (~3750 to 3500 B.P.) from the Tianshan Mountains (39). The IA is marked by an increase in movement and admixture of Steppe, Central Asian, and East Asian people into the Xinjiang region. The IA oversaw a continuation of Steppe_MLBA ancestry with greater genetic affinity to Central Asian populations containing BMAC ancestry. We also observed ancestries derived from South Asian Hunter Gatherer (Onge) in the LBA and IA, which suggests the movement of populations either from Central Asia already carrying this ancestry or from the Indus periphery region through the Pamir mountain regions into southern Xinjiang (11). Concurrently, an IA influx of East Asian ancestry from the eastern Steppe of present-day Mongolia is also observed, which may be tied to the westward expansion of the Pazyrk Xiongnu into Xinjiang. These admixed ancestries related to Steppe, East Asian, and Central Asian people established in the IA have been maintained since that time and are still prevalent in both the HE and present-day Xinjiang, linking the past with present-day populations. Whereas aspects of this reconstructed population history find support in the archaeological record, several insights can be gained by comparing the newly generated genomic data with previous archaeological and historical evidence.

    First, although the diffusion of culture is not always accompanied by population movements (32), we observed an overall concordance between the two in Xinjiang populations. For example, the major genetic influences present at the earliest settlements of north and west Xinjiang populations can be related to the coexistence of people with different cultural backgrounds—e.g., Afanasievo, Chemurchek, and Okunevo (Fig. 3A and supplementary text). Also, the shift in population ancestries can be associated with proposed population movements. Specifically, the Afanasievo-related ancestry in BA individuals is consistent with the concurrent appearance of the Yamnaya culture in the Altai-Sai region, and the Steppe_MLBA ancestry in LBA individuals can be attributed to the expansion of the Steppe_MLBA culture into Xinjiang (see the supplementary text, which provides archaeological backgrounds). In the IA, genetic affinities with nomadic groups such as the Saka reveal the widespread presence of these groups across the entire Steppe region. Overall, we detect Tarim_EMBA ancestry surviving into the IA and HE, which implies the cohabitation of populations descended from Steppe EMBA, BMAC, and local Tarim_EMBA people in Xinjiang (Fig. 3B). These findings also give genomic evidence for the broad demographic processes underlying the spread of numerous languages in Xinjiang that have survived into historical times, such as the introduction of Tocharian languages by populations associated with the Afanasievo culture. An increasing mobility and movement of the Sakas in the IA and the establishments of the Saka states lasting into the HE aided in the expansion of Indo-Iranian languages, such as Khotanese, in Xinjiang as well.

    Despite the widespread population movements documented here, the degree of genetic continuity that has been maintained in Xinjiang over the past 5000 years is noteworthy. Although genetic continuity has been observed in isolated environments or regions with relatively high cultural homogeneity, such as Northeast Asia (40), dynamic interactions between populations with diverse ancestries and cultures are more likely to result in major population shifts and turnover, such as in the Oceania archipelago and Europe (28, 41). However, this has not been the case for Xinjiang populations, where at least two different instances of genetic continuity are observed. The first is the genetic continuity (Steppe ancestry) from BA individuals to LBA and IA individuals, which represents a case in which a core Steppe ancestry has been maintained despite the addition of an extensive influx of diverse ancestries. The second case is the stability of Xinjiang population diversity from the HE to the present day, despite the turmoil of successive external ruling powers over the past 2000 years. A major reason may be that this mixed ancestry was prevalent not just in Xinjiang but throughout Central Asia, so dynamic population movements would not result in major genetic shifts. These findings indicate that genetic and archaeological evidence can provide distinct yet complementary insights into population history. This, in turn, further emphasizes the importance of a multidisciplinary approach to uncover the complex histories of regions like Xinjiang, where persistent interactions between various populations and cultures occurred.

    The paper isn’t easy to follow, or trawl for details on single elements, but it seems that the Afanasievo, the Okuneva (who were their neighbours up north in the Minusinsk Basin) and the Chemurcheck (who took up nomadic herding on the Western Mongolian Plateau after learning animal husbandry from them) may have formed an economic and political community, with the Afanasievo as an initially dominant minority element. This multi-culture could well have become linguistically Afanasievan (Indo-European) with an Okunevan (e.g. Yeniseian) substrate. They spread out in a wide region around the Altai searching for pastures and presumably copper ore. We find them in the Early Bronze Age in the fertile Ili Valley in Western Sinkiang. With the advent of the Iron Age about 1000 BCE they seem to have been replaced or subdued by the Saka, and I don’t where they went after that, but they leave written documents along the northern rim of the Tarim basin, just across the Tian-Shan range, from around 400 CE.

  9. Trond Engen says

    I learn that recent archaeology in the Ili valley divide the metal age into an Andronovo era (c. 1900-1000 BCE) and a Saka era (after c. 1000 BCE). That might weaken my assertion about continuity from the early Bronze Age, but I’ll note that there’s some debate (as always) over whether the Andronovo phase is echt Andronovo or a local development influenced by it. It may be more scathing that there’s no evidence of metal work older than c. 1900 BCE. But absence of evidence etc.

    Zhi, Festa: Archaeological Research in the Ili Region: A Review, Asian Perspective (2020)
    Wang et al: Copper metallurgy in prehistoric upper Ili Valley Xinjiang China, Archaeological and Anthropological Sciences (2019)

  10. David Marjanović says

    The paper isn’t easy to follow, or trawl for details on single elements

    On the upside, it’s in open access, unlike the last few Science papers in this thread!

  11. David Marjanović says

    Remarkable amounts of Anatolian Farmer ancestry – up to 43%.

  12. Trond Engen says

    Yes, in some. I don’t have the paper in front of me, but I think those are believed to be Sintashta (or maybe Andronovo) Indo-Iranians. I gave up trying to unite all data for each individual.

  13. Trond Engen says

    Me: We find them in the Early Bronze Age in the fertile Ili Valley in Western Sinkiang. With the advent of the Iron Age about 1000 BCE they seem to have been replaced or subdued by the Saka, and I don’t where they went after that,

    No, that’s not it. Except from this single skeleton which they say is unadmixed Afanasievo, the other finds that are archaeologically Afanasievo or Chemurcheck are concentrated up north. These cluster together in the PCA plot. I’m confused because the “pure” specimen is rather like Sintashta. But in the admixture bars they all look the same. I guess there are things going on that isn’t shown in the plots.

    Looking at the PCA plots I’m leaning towards the old suggestion that the Tocharians are the Wusun. They show up at the eastern end of the Silk Road, together with the Yueshi, a couple of centuries BCE, when the Han start pitting the steppe peoples up against eachother. Maybe this is the old alliance of Afanasievo and Chemurcheck?

    Anyhow, the Wusun and the Yueshi start fighting eachother, and then moving westwards in turns under pressure from the Xiongnu. The Yueshi first took the Ili Valley from the Saka. Then the Wusun took it from the Yueshi. The Yueshi then went on to Bactria and founded empires, while the Wusun stayed and eventually were allowed to form a sort of buffer state between the Xiongnu and Han China, sinking gradually into obscurity. This is exactly the period when the Tocharian languages are attested.

  14. This Wikipedia page has a “Chemurchek culture and contemporary cultures and polities” map, useful for those of us who have a hard time keeping in mind which culture was where. (Can’t link to the map directly — all the labels disappear.)

  15. Trond Engen says

    Or maybe the Tocharians are both the Yueshi and the Wusun, if the fallout and breakup were messy, if either A or B is a Yueshi stay-behind group, or if either represents the local delegation of the Kushan empire.

  16. Trond Engen says

    That history of the Wusun and Yuezhi* peoples is just my summary of (mostly) Wikipedia. Ideally, this is where John Emerson would join in with all the Yuezhi details from Chinese, Mongolian, and Tibetan sources.

    * Did I write ‘Yueshi’ up there? So I did. I devoice on your lousy Pinyin!

  17. David Marjanović says

    Aren’t the Yuezhi more likely the Actual Tocharians, the ones who wrote an Iranian language in the Unknown Kharoṣṭhī script?

    unlike the last few Science papers in this thread!

    …in… the… Son of Yamnaya thread.

  18. Trond Engen says

    The Kushan Script Deciphered this July. I was just pondering how to think about that.

  19. David Marjanović says

    Oh. Remarkable how I half-remembered that and mixed it up with the knife reported on LLog.

  20. (Can’t link to the map directly — all the labels disappear.)

    Yeah, wp seems to have a real problem with maps. The thumbnails on the page are too small to read. Magnifying in-situ is temperamental[**]. (I got it to work eventually.) Clicking gives you a full-screen map with no labels — useless.

    [**] And randomly magnifies other wp pages you might have open.

  21. Lars Mathiesen (he/him/his) says

    Chrome, at least, and maybe Firefox, have the feature that the zoom level is shared between pages with the same host name. I’m not sure if “wp” here is Wikipedia or WordPress, but it would certainly be a thing for separate blogs hosted under (Wheras blogs with their own boughten domain name would not share zoom levels, even if actually hosted on wordpress.\code/>com).
    (*) so called

  22. Lars Mathiesen (he/him/his) says

    Ha, my trick worked (writing wordpress.<code/>com to avoid autolinking), but I misspelled it the second time and forgot to check inside the edit window….

  23. David Marjanović says

    I vaguely recall some scholarship which suggests that Yeniseian was spoken considerably further to the south in prehistory.

    I haven’t seen that outside of attempts to tie to, specifically, Burushaski and often also to the archaeological Karasuk culture.

    Oops, sorry, that depends on how considerable “considerable” is. The evidence from hydronyms is clear that Yeniseian drifted north in historical times; but the names don’t extend as far south as where the Karasuk culture once was, IIRC.

  24. See now “Tocharian Bilingualism, Language Shift, and Language Death in the Old Turkic Context,” by Hakan Aydemir (Sino-Platonic Papers 337 [Dec. 2023], open access):

    […] we do not know at all what happened to the Tocharians, when, how, and why they disappeared, or when the Tocharian languages died out. This study tries to solve these fundamental questions from the perspective of Turkic historical linguistics. In connection with this, the Turkic background of Tocharian-Turkic interethnic and linguistic contact is first examined in order to understand how the Tocharian-Turkic language contact came about in the first place and how long it lasted.

  25. Thanks @Hat for linking that. I have questions … (about relying on soundalikes for random seemingly everyday words, rather than specialist vocabulary like horse equipment or thole pins). But also I have ‘form’ over there which makes me reluctant to even ask.

  26. uTexas announcement (a year+ ago) of Dr. Aydemir’s research plan.

  27. David Eddyshaw says

    Publication in the Sino-Platonic Papers does not inspire confidence: it is the home of some really quite extraordinary nonsense, and edited by a man who has a very large and very evident axe to grind, and who lacks any comprehension of the basics of historical linguistics.

    Still, the paper deserves to be looked at on its own merits. (An evident proto-Oti-Volta cognate did catch my eye, though …)

  28. … the paper deserves to be looked at on its own merits.

    Sadly, it seems the paper that deserves looking at is not this one, but another “ This period, together with other linguistic data, will be covered in another article. ” [p. 9 — all too ‘Deep Thought’] So those wanting posited cognates can inspect only the fewer than a dozen single-syllable roots pp 7 ~ 9.[**] (Which anyways date to very early in posited contact; you’d expect borrowings/influences from the different episodes of contact over the several centuries the paper claims.)

    Those few pages were what raised my questions above. I assumed at that stage there’d be much more in an addendum or somewhere.

    What I’d expect (but tell me I’m being dumb) is comparanda showing this form doesn’t fit the sound pattern of language X, so must be a borrowing from language Y, which it does fit. And/or this form is only in Turkic languages that were in close contact with Toch; other Turkic languages show a quite unrelated form. I don’t even get a strong feel for the direction of borrowing.

    There’s a bunch of possible ethnonyms/toponyms in Turkic languages of the Tarim basin possibly showing Toch influence.

    [**] Oh, those who hold their breaths until cont. p42 get

    I assume that the OTu. op in Argu dialect is a reflex of the TochA ops- ‘ox.’ Namely, according to Pinault, ops- goes back regularly to the form *ops(o) ( TochB okso ‘ox’, i.e., -ps- < *-ks-).

    Really? The reflex of PIE *uks-en is TochB ‘okso’? That seems all _too_ neat.

  29. I had the same immediate reactions as David E. and AntC, and this morning I just stopped reading after a few paragraphs, underwhelmed by the borrowings in Old Turkic. I’ll take it up again after Christmas, when days are long and slow and I have more patience.

  30. David Marjanović says

    I’ve downloaded the paper, but haven’t had time to even begin reading it.

    Does it mention Iranian/Iranic? There must be a decent helping of words from that family that were borrowed into both Turkic and Tocharian.

  31. haven’t had time to even begin reading it

    You might do best to start with the Summary pp 74 ff, especially the posited timeline. (Apologies if I’m ‘teaching my grandmother …’.)

    Does it mention Iranian/Iranic?

    It should also be noted that while a significant part of the Tocharians became Turkicized, another part became Iranianized, and another part became Sinicized. However, their Iranianization and Sinicization processes will be discussed in a separate study. [p. 5]

    The ‘meat’ being “in a separate study” seems to be a recurring pattern. The guy’s a Turkicist not an Iranianist, so reasonable, I suppose.

    That said, there are plenty of mentions in the footnotes of findings/speculations[**] from Iranianists.

    What worries me methodologically is the risk of finding a borrowing into Turkic and presuming it’s from Toch (and taking that as evidence Toch was still spoken), without first eliminating the possibility it’s from Iranian.

    [**] Indeed many of the mentions of Iranian are suggestions of mis-attributions: what had been thought as an Iranian borrowing into Turkic (or v.v. or into Toch) is more likely Toch into Turkic — note 121 p. 47, for example. Indeed on a closer reading, I tend to think the footnotes carry more ‘meat’ than the text.

  32. a) Long after Toch lost prestige as a Buddhist liturgical language, it continued as a vernacular, gradually losing ‘grammaticalisation’.

    b) Although TochA might be the older IE variant, it continued in parallel/in different areas vs TochB and indeed outlasted TochB. the Argu were undoubtedly the descendants of Yuezhi, and their language was TochA. [p. 79]

    c) That’s why there was still in C10th enough knowledge of TochA [p. 23] to translate key Buddhist texts (Maitrisimit and Daśakarmapathāvadānamālā) into Uyghur — albeit with some misunderstandings of the grammar.[**]

    [**] like modern English misunderstandings of Shakespeare or KJV?

    But please explain: if you’re translating Buddhist texts why not start from the Sanskrit?

  33. 1. Nobody knows Sanskrit.
    2. The Tocharian text is canonical or sacred in its own right.

    Compare the Vulgate in the Cactholic tradition.
    Compare the KJV in some Protestant domains.

  34. I have read the paper. It’s not so much a research paper as a synthesis hypothesis or an outline of his ongoing work, so the Sino-Platonic Papers may be a good place for it.

    Main points:
    – The Yuezhi are the Toch A (Yarki). They ruled a state where the population consisted of both Toch A and Toch B groups.
    – The historical Tocharians are the Toch B (Tukhar). They ruled the western part of the basin
    – The mobile elites of both states were chased out by the Huns and fled first to the Ili Valley and then to Bactria in the second century CE.
    – The societies rebound, united by Buddhist religion and a monastic economy,
    – The languages took the first blow after the Tibetan conquest of 840 CE. The Tibetans were Buddhists too, so the religion and the monastic economy survived and suffered through the Uighur-Tibetan wars that followed.
    – The next blow came with the Uighur conquest of 866 CE. The Tocharian elite may have become Manichaean and Uighur-speaking, but Buddhism and Tocharian spread from below, until the Uighur king converted.
    – The Uighur state was bilingual. Uighur was used for official and external communication and Toch A and B in everyday situations and religious rituals.
    – The Western basin was conquered by the Karakhanids in 1040, and islamization followed. The Toch B reorganized as a tribe and kept a separate identity within the tribal system for many generations. So did a tribe of Argu, likely Toch A from the Ili Valley. Both were bilingual.
    – The late texts in bad Tocharian reflect a period when the language was no longer spoken by the writing class but still used in Buddhist rituals.
    – Tocharian finally died out in the Karakhanid realm around 1200 CE, in the Uighur realm before 1300 CE. Paradoxically, Toch B seems to have survived longer than Toch A in the Uighur east, while Toch A survived longer than Toch B in the Karakhanid west.

    The case relies entirely on reinterpretation of written sources, and he suggests that genetics, archaeology and thorough lexical work in the rural regions of Xinjiang and eastern Kyrgyzstan and Kazakhstan might clarify the matters.

  35. Thanks very much for that summary! Interesting and plausible, at any rate.

  36. Trond Engen says

    Synchronicity strikes again. Here’s an article putting genetic sequencing for ethnic origin in grimmer light:

    Amy Hawkins, The Guardian: Academic paper based on Uyghur genetic data retracted over ethical concerns.

    The article mentions lack of informed consent by the participants, but the ethical problems seem much deeper than that.

    This isn’t exactly about ancient DNA, but it’s about using genetic anthropology in criminal investigations — and potentially about using genetic screening to isolate ethnic groups — and that makes me very uneasy. Maybe Dmitry can throw some light on the issue.

  37. @Trond The case relies entirely on reinterpretation of written sources, and he suggests that genetics, archaeology and thorough lexical work in the rural regions of Xinjiang and eastern Kyrgyzstan and Kazakhstan might clarify the matters.

    Yes, thank you for your ‘due diligence’.

    It was not clear whether the “thorough lexical work” is already further advanced than the too few examples in S-PP. Are we about to see another paper in a more ‘technical’ forum?

    I appreciate many lines of research would like genetic data wrt Uyghur ethnology — especially before PRC wipe them out completely — but I’m appalled at Elsevier and OUP (alleged). “ help the police identify suspects in cases” my foot! As far as Chinese State police are concerned, if you’re Uyghur, you’re guilty.

Speak Your Mind