I’ll let you in on a secret that might be a bit surprising; The people of Khotan were never described as Saka or Saka descendants by their contemporaries. Their kings/dynasties did not identify as Saka like we see with the Indo-Saka dynasties. The people did not refer to themselves as Saka either, and going by the fragments of history regarding them, they did not have any memory of their nomadic ancestors and supposed linguistic forebears.
So how come there is a consensus that the medieval urban buddhists of Khotan were of Saka origin then?
It mostly comes down to linguistics, or at least it was in this field where the seed was planted.In the onset of the 20th century, British-Hungarian archaeologist sir Aurel Stein underwent four expeditions into Central Asia, including the Taklamakan Desert. In the later expeditions he managed to find many manuscripts that were in several languages such as Sogdian, Uyghur and Tocharian. Some of the documents were seemingly in an unknown Iranian language, as uncovered by indologist Rudolf Hoernle. But it was J. Kirste who dubbed the language ‘Khotanisch’. From what I can tell, it was in 1913 when the language was linked to the Indo-Saka by Heinrich Lüders, and since then it has been known as Khotanese Saka.
But are the Saka a good historical match for the spread of this language? Personally I don’t think so. I will divide my views in three categories; linguistics, history and archaeology. It will be a bit of a long read by the way, I spent some few months working on it, and shelved it multiple times due to new data coming out and me getting bored with the topic (also going to have to admit that the finished version was in the vault for over a month because I was too lazy to fix the references), as well as preparing for a fight last week (KO victory), but it is here finally! If you need a tl;dr version I’d suggest reading the article “Tocharian B etswe ‘mule’ and Eastern East Iranian *” by M. Peyrot, as a lot of our positions on this topic seem to overlap.
As I laid out earlier, The Khotanese language was dubbed Saka because it was deemed to be identical to the language of the Saka of South Asia by Heinrich Lüders, one of the earliest linguists to study this language. It is interesting to note that Hoernle had this to say about that suggestion :
In that portion of Central Asia, as it is now well known, there once prevailed, in the early centuries of the Christian era, two distinct languages, which are now quite extinct, and have to be laboriously recovered from the oblivion. Broadly speaking one was spoken in the north, the other in the south. The northern language has been named ‘Tokhari’ by Dr. F. W. K. Muller, and the southern,’Northaryan’ by Professor E. Leuman, and Saka language by professor H. Luders. None of these names however, based as they are on more or less disputeable ethnic or historical consideration, has met with general acceptance.
Ahead of his time I’d say. However that may be, it seemed that in the decades afterwards the Saka origin of this language had been accepted. Janos Harmatta and Harold W. Bailey are two examples of renowned linguists that extensively worked on Khotanese which supported the notion of it being a Saka language.
It should be pointed out that the Saka language in South Asia is poor in attestation, and most of the linguistic data comes by way of onomastics, the study of attested names. These names were often attested on coinage, or sometimes in inscriptions and edicts.
In this day and age not all linguists agree with the assessment that Khotanese and Saka were identical languages. As Michael Peyrot wonderfully summarises in a single sentence :
In sum, the evidence that the Saka language is Khotanese or an earlier form of it is weak.
A peculiar feature of Khotanese and Wakhi is that the Proto-Iranic *xw sound developed into ss rather than sp in most Old Iranic languages. Aspa, the old Iranic word for horse, was rendered as assa in Khotanese. In Scythian and Sarmatian names, the sp sound is present in names such as Ispaka.
“In Avestan, the reflex of Proto-Indo-Iranian *´cu˘ is sp, e.g. aspa- ‘horse’, as in most other languages; in Old Persian it is s, e.g. asa- ‘horse’; and in Ossetic it is fs, e.g. Dig. æfsæ ‘mare’ (<!PIIr. *Ha´cu˘a¯), probably from earlier sp. However, in Khotanese, Tumšuqese, and Wakh¯ı the reflexes of the cluster are the palatals ´s´s, ´s, and š, respectively, e.g. Khot. a´s´sa-‘horse’, Tumš. bi´sa- ‘all’ (∼ Khot. bi´s´sa- < PIIr. *u˘i´cu˘a-), and Wakh¯ı yaš ‘horse’ (Windfuhr:).” 
Yet amongst the Indo-Saka, you had a few names containing the typical old Iranic *sp, such as Spala; Spalagadama, Spalahora and Spalirisa. Harmatta argued that these were derived from Parthian *spaδa (army) , but I wonder if the argument arose because of the assumption that Khotanese and Saka were identical, making it impossible for it to have been a native Saka-Khotanese term due to the presence of *sp. This would be a bit of circular reasoning I’d say. After all, Pontic Scythian names containing Spala have been attested by way of Greek inscriptions (Σπαλω).
Here is one example that predates the Parthians, from a Greek inscription on a vase found near Olbia. For the sake of argument, let’s say that Harmatta was correct and that these had a Parthian origin.
You also have the issue that names such as Aspavarma and Vanaspara amongst the Indo-Saka kings do not have the Khotanese ss soundshift, and considering the role of the horse in Saka society I doubt they wouldn’t use their native term for horse but a borrowed Iranian one. it would be Asva in Indo-Aryan languages of course. Particularly not because many of the Indo-Saka names were compound names with an East Iranic prefix and an Indo-Aryan suffix.
I also don't see any names of kings that contain the Khotanese ss shift either, so like Peyrot I doubt the claim that Khotanese-Tumshuqese and the language of the Saka of Sistan and South Asia were identical languages.
Coin of Aspavarma
The archaic nature of Khotanese within Iranian languages is rather fascinating. Linguists have noted that the language retained Daevic terminology according to L.G Herzenberg . Here are some examples:
kamalä "head" - Avest. kamәrәδә "head of a demon"
kṣīa "teacher" - Avest. t̰kaēša- “false teaching”
ggalū "family" (from *garduv-<gard-)
gaṇḍye "building" from *garganta- - Avest. gәrәδә "house of the demon";
hīnā- "army" - Avest. haēnā- "enemy army";
paśa uda- "mouth" from *pa(ti)-zafta - Avest. zafar- "mouth of the demon, mouth";
ggośtä- "hand" from *gabasti- "hand" - Avest. gava- "demon's hand, paw", etc.
This would suggest a period of development outside of the influence of Iranians practising early forms of Zoroastrianism, or a separation from the main cluster of Central Asian Iranian languages of the later bronze and early iron age before this religion developed and spread.
The iranian language of Khotan was not alone. The neighbouring cities such as Kashgar and Yarkand all spoke very closely if not identical languages. Furthermore, you also have the Tumshuqese language, which was attested in 9th century texts from Tumxuk, in the northwestern part of the Tarim Basin. This language was closely related to Khotanese, however both languages had already diverged considerably at the time of attestation, and their divergence unlikely occurred outside of the Tarim Basin.
Given that the iron age Scythian steppe nomads were East Iranian speakers, with languages that in all likelihood diverged from the main cluster of Iranian languages in West Asia and southern Central Asia since the bronze age, it certainly wouldn’t be surprising that the language of the Saka and Khotanese, another seemingly archaic East Iranian language would have significant parallels with the language of the Saka. But to say that the languages were identical, or that the Indo-Saka language was an earlier form of Khotanese is a bit of a stretch in my eyes.
The great migration of the Saka
Saka can be a bit of a confusing term at times. Persians used it for Iranian steppe nomads of Central Asia such as the Sakra Tigraxauda and the Saka Haumavarga, but also called the Pontic Scythians Saka Paradraya. The usage of Saka as explained by Strabo seems to be in line with how academics use Saka nowadays, referring to the Scythian nomads east of the Caspian and on the central asian steppes.
The ‘Saka’ which are of relevance here would be the ones which lived around the western Tian Shan, as well as the Ferghana valley and the Alai-Pamir region. These people would match up the best with Saka Haumavarga, and the Saka para Sugd, which are likely the same people. The Amyrgians mentioned by Herodotus are in likelihood part of the same ‘cluster’ of Scythian nomads, if not the exact same people. Through increased contact with Central Asia the Han Chinese came to be aware of the Sai or Se, living in the Ili valley region.
With the onset of the Xiongnu empire, you had the beginning of the migration era of the Central Asian Steppes. After a period of wars and conflicts for 30 years, the Yuezhi were finally defeated by the Xiongnu in 176 BC in east Xinjiang or western Gansu, their king was killed and his head was turned into a cup. This caused the Yuezhi to break up, with the Han Chinese referring to the main migratory branch as Da Yuezhi and the Yuezhi groups which remained (and thus were either part of the Xiongnu or allied with “Qiang'' tribes, as Xiao Yuezhi. The Da Yuezhi migrated to the Ili river valley, which brought them to the territory of the Saka the Han were familiar with at the time.
The Saka, pressured by the migrating Yuezhi, moved southwards, towards the ancient region of Bactria. However in turn the Yuezhi were pressured by the Wusun, who then were part of the Xiongnu empire and came to conquer the Ili region. This caused the Yuezhi to move southwards into Bactria and conquer the region around 130 BC, which drove the Saka even further south, into eastern Iran and South Asia.
But the best known of the nomads are those who took away Bactriana from the Greeks, I mean the Asii, Pasiani, Tochari,1 and Sacarauli, who originally came from the country on the other side of the Iaxartes River that adjoins that of the Sacae and the Sogdiani and was occupied by the Sacae.- Strabo, Geography, Book 11, chapter 8, section 2
The Saka movements were not wholesale population movements however, as volume 96 of the Book of Han mentions how the Wusun kingdom (Ili region) contains both the Saka and Yuezhi. Perhaps the Sacarauli mentioned by Strabo were Saka that joined the migratory Yuezhi-Kushan groups.
The invasion of Parthia by these Saka was felt by the cities of Merv, Hecatompylos and Ecbatana, which were all attacked by these Saka groups. The Parthian king Praates II met his fate fighting against the invading Saka. As the Saka migrated very close to the core territory of the Parthian empire, the Parthians took this threat very seriously. Eventually the Saka were defeated by Mithridates II who reconquered the Parthian regions. The Iranian province of Sistan still owes its name to these Saka migrants.
Bust of a Saka male from the Karachalchyan burial in Uzbekistan, who went by the name of Davidski.
In South Asia this migration led to a period of Saka-ruled dynasties in the northwestern part of the subcontinent. These Saka groups invaded the territory of the Indo-Greek kingdom, which was founded by Graeco-Bactrians. Unfortunately we don’t have any historical sources from the Indo-Greek kingdom and have to rely on onomastics by way of minted coins for the most part.
The first dynasty of the Indo-Saka was founded by Maues. An intriguing detail is that Maues’ rule only began in the first century BC, with the starting date estimated between 98 BC and 85 BC, and ending in either 60 BC or 57 BC. This means that the start of his dynasty occurred decades after the migrating Saka had reached South Asia, which likely was somewhere in between 130 and 120 BC. Maues may even have been born within the modern geographic boundaries of South Asia but it is impossible to confirm or deny that.
In any case, it seems that at least for a few decades that the Saka which invaded the Indo-Greek kingdom still accepted. Perhaps they initially vassalized, or were given a territory that nominally was still ruled by the Indo-Greeks but in practice was already territory ruled by the Saka. Without any historical texts describing the history of the Indo-Greek kingdom this cannot be properly investigated, making it an unfortunate knowledge gap.
The Indo-Greeks didn’t simply vanish as Maues was allied with Indo-Greek kings such as Artemidoros and his wife Machtene was of Indo-Greek origin. However, with the successive expansions of the Indo-Saka dynasties, but also territories being taken back by Indians and the Parthian empire expanding eastwards, their power dwindled and by 10 AD the political power of the Indo-Greeks was no more.
Coin of Maues
A few centuries later there were the northern and western satraps, both ruled by descendants of these invading Saka. The last of these dynasties were defeated and ousted by Chandragupta II in the 4th century AD, which may have served as inspiration for the legendary Vikramaditya, who also defeated and ousted the Saka.
When it comes to the topic of Saka migrations to the western Tarim Basin, the book of Han mentions their migrations in the vicinity of this region. But the attention is in the details! The Saka migrations as describedin volume 96 of the book of Han lead to the invasion of states situated to the northwest of Kashgar, as others travelled south through a pass in the Pamir Mountains, and ended up in the kingdom of Jibin, which is generally associated with Kashmir. Which you can check out yourself, because the Hanshu is available on Wikisource: 漢書/卷096上- 维基文库，自由的图书馆
The Han were quite meticulous in describing the location and routes of the Saka and the Saka were not mentioned in the context of the Tarim Basin, whereas their migrations to places further away from China such as Jibin were attested. The same chapter of the Hanshu which describes the Saka migration also gives descriptions of Yarkand, Kashgar and Khotan. Thus the movement of Central Asian Saka in the context of Xinjiang can only be limited to the most western parts of Xinjiang going by historical records, the Tarim basin city states such as Yarkand, Kashgar and Khotan being too far east.
What I find striking is that these same records are often referred to as if they mention a migration of Saka people to the Tarim Basin. One example:
Chinese accounts confirm that the Sakas residing in the Upper Ili Valley moved southwards to the western corner of the Tarim Basin, when they were compelled to leave by the Yuezhi, who in turn were driven from western China by the Xiongnu at the latest from 174 BC onwards (Posch 1995: 99; Yu 1995: 387, 388, 407; Benjamin 2003: 1). 
Generally there is a revolting door of references from other authors rather than the direct records themselves, but when tracking it all down the affirmation of these Chinese records are that Iranian names were recorded for the Kings of Khotan and that the kingdom had the same ruling dynasty since its inception….
Another point I would like to highlight is that if the people of the western Tarim Basin had an ancient Saka origin, it would be very peculiar that the Han never picked up on this. The Han Chinese were definitely not perfect in regards to their knowledge of “western barbarians' ', but they had a good understanding of the events, and they also had a good memory. For example, in a 10th century description of the city of Cumuda in the eastern Tian Shan, it was recalled that the people there were the descendants of Xiao Yuezhi who founded the state. This was more than a thousand years after the foundation of said city state.
The Han were perfectly aware of who the Saka were, and where they had been, yet they never mentioned any Saka origin for the medieval people of Khotan. And neither did the people of Khotan themselves for what it is worth.
As far as the historical migration of ‘Saka’ goes, there doesn’t seem any evidence that any of these Saka populations actually migrated to the Tarim Basin and founded kingdoms there.
Bailey quite early on already recognized that it was unlikely that such an event would have occurred, because there were no records of any invasion in Khotan since they entered the realm of historical texts circa 200 BC (although there actually were during the Kushan period, but that's a different story). His assumption was then that a different Saka population had migrated to the region in earlier times. Why this was seen as the most logical assumption is something I don’t understand, but I imagine it was so because linguists had decided that Khotanese was a Saka language for a while at this point.
So then, maybe different Saka populations migrated into the Tarim Basin? Perhaps at an earlier time frame? Then the historical peoples called Saka migrating into southern Central Asia were not the linguistic forebearers of the Khotanese, but another population of Saka steppe nomads, thus making Khotanese Saka a valid description for the medieval Iranian languages attested in western Xinjiang? Let’s find out.
Saka materials in the Tarim Basin
The first thing to address is the presence of Scytho-Siberian-like artefacts and material influence in the Tarim Basin, which sometimes are linked to movements of ”Saka” ancestral to the medieval Khotanese.
Satma Azar, an iron age site from the Tarim Basin could serve as a good introduction of this topic. Here is a description of the site :
On two of the vessel’s sides is an incised image of a deer with two legs, the one with at
least three tines of antlers and the other with eight tines of antlers (Plate 5). On both animals a spiral line forms the transition from the front and hind legs to the body. In the early Iron Age steppe cultures, the spiral was a decorative element used in the representation of animals and belonged to the syntax of the Scytho-Siberian Animal Style of the early 1st millennium BCE.5 This vessel with the two deer shows obvious references to the Scythian steppe art, also seen on petroglyphs from western Mongolia, southern Siberia or Kazakhstan. Similar wooden pots were found in Xinjiang at Zaghunluk near Qiemo and at Yanghai, Shan Shan county, about 100 km east of Turfan. Those vessels date from the late 9th century BCE but they also feature other engraved animals such as ibexes, wolves, camels and desert antelopes.
Also at Jumbulakum, an oval-shaped, flat plate was found with a similar deer engraved on its outer bottom. These wooden receptacles testify to contacts between the steppe world and oases of the Taklamakan Desert. A further piece of evidence for cultural contacts with the steppe cultures north of the Tian Shan was found on the ground of House H4, namely a 3 × 3 cm bronze head of a horse which may have been part of a buckle. Its shape is very similar to bronze ornaments from the steppe cultures in northern Xinjiang, western Mongolia and Tuva in southern Siberia.
One issue with attributing these type of influences to a migration of “Saka nomads” ancestral to the Khotanese is that Scytho-Siberian cultural influences in the Tarim predate the historical migration of the Saka by many centuries, and in some cases predating the first historical mention of Saka in general. Scythian-style weapons, ornaments, horse gear, clothing and artistic forms were present across the Tarim basin since the early iron age. Some of the later bronze age artefacts already look quite similar to their contemporaries in South Siberia and East Kazakhstan, which may hint at a material connection to these regions prior to the formation of Scythian cultures. An example are these later bronze age weapons from Sebeir:
Furthermore, it is argued that the usage of these Scythian-related artefacts arose due to contact and influence, and that this should not be interpreted as a common Scythian cultural origin . After all, the subsistence economies of the people in the Tarim Basin and on the steppes were quite different, but more on that later.
If you’d want to argue that on basis of these material affinities to Scytho-Siberian peoples that the iron age inhabitants of the western Tarim basin were of Saka origin, then you’d also have to extend this to the people of the eastern half of the Tarim Basin and Dzungaria, regions where'd you’d expect early Tocharians to live.
Sites like Yanghai are by russian archaeologists are sometimes included in the wing of the Scytho-Siberian cultures due to the material influence (note; this doesn’t mean they were Scythians per se), and deer iconography similar to those at Atma Sazar was present as well. I’d even argue that this region has even more of a Scythian/Siberian material influence than the people within the Tarim Basin.
Zaghunluq which was also mentioned is the site of the famous Cherchen man, which is sometimes connected to Saka as well. The linguistic attribution if the inhabitants of this site is still very much up for debate, as it more or less lies within the border territory of the assumed “Khotanese” distribution and the distribution of the putative ‘Tocharian C’ language, although the dates of this iron age cemetery are early enough that it doesn’t have to be connected to either. While some of the materials at Zaghunluq such as horse riding gear could be connected to iron age nomadic populations, the cemetery had been in use for several centuries and it and the materials found are unmistakably tied to the local traditions of the Tarim Basin. I doubt these have anything to do with th Saka.
Djumulak Kum or Yuansha Gacheng is an archaeological site of a fortified town or ‘city’ dating to the middle first millennium BC. It had been connected to the Saka by J.P Mallory, on the basis of pointy hats, wooden burial chambers and similarities in pottery to Saka peoples (which ones?). These connections are a bit of a stretch if you’d ask me. Especially because high pointed hats and wooden chambers also show up at Zaghunluq and the Subeishi culture, which lay to the east of the assumed distribution of the Khotanese-Tumshuqese languages and lie more in the region of the historical Jushi peoples.
I’ve seen no mentions of burials with obvious signs of the Scythian triads, affinities of these people to steppe nomads, an establishment of a nomadic elite, nomadic influence in subsistence economies or anything of the like in the relatively few articles and books that cover the town. The only connection I’ve seen made is similarities in textile weaving techniques and clothing design patterns.
The economy of the people at Djumulak Kum was quite reliant on agriculture. They had intricate irrigation networks covering several kilometres. Wheat, barley and millet were extensively grown by the people of Djumulak Kum, which had large grain storages for all their produce.
There simply isn’t anything to suggest that the people at Yumulak Kum were of Saka origin. If that single cemetery is representative of a Saka presence, which I doubt it is, it wouldn’t be the type of presence that would cause a massive linguistic shift in my eyes, because so far it would be a single cemetery that does not suggest a considerable presence or ruling elite.
East Pamir Saka
Perhaps only slightly relevant for this topic, are the ‘Saka’ sites of the Pamir mountains, in particular the Eastern Pamirs as some of these are within the borders of modern day Xinjiang. Given that these sites are sometimes referred to in the context of Saka languages in the Tarim, I figured let’s cover them as well.
The Pamir Saka material culture is very poorly studied, with most of the data coming from excavations from the 50s and 70s. The proposed dates generally seem to be within the 7th and 3rd centuries BC, although multiple opinions about the precise dates exist. The sites are mostly distributed on the southeastern part of the eastern Pamirs.
Aleksandr Bernhstam excavated these sites in the 1950s, and connected these sites to the Saka due to the presence of kurgan burial rites, scythian animal motifs and the lack of evidence for settlements.
Litvinsky also had his excavations in the region, and looked into these Saka sites. Unlike Bernstham however, he connected them to the Saka Haumavarga. Litvinsky then worked this into his theory of primordial ancient connections and longstanding interactions between the Saka and South Asia, of which the Pamir Saka would be proof of. Funnily enough in the first century of Saka and Indic contact the Indians conflated Saka and Greeks, referring to distant lands as Sakayavanam. To me this is not reflective of long standing connections.
Bernshtam and Litvinsky noted that there weren’t any settlements and assumed that they were nomads. More recent archaeological data in Central Asia have shown that many sites which initially were held to be nomadic did show signs of continuous habitation and agriculture, making it likely the amount of nomadism has been overestimated in academic works of the past. Therefore it is quite unfortunate that most of the archaeological data on these burials is quite old.
While I’ll be the first to admit that the combination of these types of artefacts, weaponry and mounds is quite Saka-like, there are some peculiar features. The metallurgy used here is less refined than what we are accustomed to in contemporary Saka sites on the steppe belt, and the kurgans are not as extravagant either. I also noticed that many of the burials had flexed/crouched positions, rather than the supine burial positions of the Saka populations of the Tian shan.
I do wonder if there isn’t a degree of continuation from the traditions of preceding bronze age populations, because kurgan burials were practised by the Andronovo groups of the Pamirs, flexed burial positions were common in the bronze age,and in Xiangbaobao there is cremation, which was a feature of Fedorovo Andronovo groups.
Both the domestic architecture and grave-type of the Andronovans survived among the Pamir Saka. In Tegirman-say, Kyzyl-Rabat and Vorukh they buried the dead within stone circles that enclosed stone cists, which were set in a pit and covered by stone slabs (Litvinsky 1972: 134-135). This rite is preserved in some districts of Afghanistan, in Ishkashim, Vahan, Darvaz and in Yagnob among Iranian peoples of the mountainous regions of Tadzhikistan (Andreev 1927: 53; Andreev and Polovtsev 1911: 17-18; Rakhimov 1956: 69). 
That’s pretty dope.
Anthropologically the Saka of the eastern Pamirs stood out from the other Saka populations in that their physical type was typical of southern central asia, and even bore resembles to some populations in South Asia, such as those of the Swat valley. They were not described as a mixture between a local southern central Asian and a northern steppic population either, the latter of which are generally described as mostly Andronovo-derived with minor “Mongoloid” features.
At Xiangbaobao, a site in the Southwestern Xinjiang near Tashkurgan linked to the East Pamir Saka, there is some more revealing information as the site contains more than just burial data.
The Xiangbaobao cemetery showed two different burial practices: cremation and inhumation. From the cremation-type burials very few objects were recovered: a copper ring, a piece of iron and several fragments of pottery and bones. These graves were assigned to the Iron Age. The inhumation-type graves contained pottery vessels, similar to the ones from the Tajikistan part of the Pamirs99, copper, bronze, iron and gold ornaments and implements. Etched beads of the Saka type were also recovered (Figure 22). The second group of Xiangbaobao graves, the richest recovered so far in the Xinjiang Pamirs region, was C14-dated to 900-400 BCE100, it should be ascribed to the Iron Age. 
There is evidence for fixed settlements and agriculture. If this is indicative of the Pamir Saka as a whole, then it would fall in line with the trend that archaeologists in the past overestimated the degree of nomadic herding which was practised in many regions of Central Asia.
Livintsky connected these ‘Pamir Saka’ to the Saka Haumavarga, but it is more logical to identify those with the more obvious Saka groups to the west of this region. The Saka Haumavarga bordered the Achaemenid empire and were turned into vassals by Cyrus the Great. This is an event that clearly did not happen in the east Pamirs, a territory that lay to the east of the Achaemenid satraps and wasn’t part of the empire. The Saka Haumavarga also lived on the Amyrgian plain, which wasn’t in the Pamirs either.
Thus we have no historic basis to refer to these people of the eastern Pamirs as Saka, only an archaeological one. I think there can be a good argument made that what we see here is a local emulation of Saka culture and fashion, perhaps specifically amongst the elites of this society.
At the time of this cultural development the significant neighbours would have been the various peoples inhabiting the eastern border regions of the Achaemenid empire, The Saka of the western Tian Shan and Pamir-Alay and South Asia. As these people clearly weren’t Zoroastrian, there may not have been a religious/cultural aversion against nomadic Iranian culture as we’ve seen in other parts of Central Asia.
These people would have had an economy where livestock is important, which meant they lived in a society where the old Indo-European trope of livestock raiders and livestock protectors comes into play. In the Scythian culture in typical Indo-European fashion the warrior elite and the presentation of such in life and death was an important aspect of their culture and they really mastered it. Let’s be real, it's mostly why we all think they were badass people to begin with. This is the exact type of cultural influence I can see being adopted by iron age pastoral peoples, without the need of significant migration or an elite replacement from Scythian groups.
I think that the archaeological Scythian affinity could be interpreted as an expression of Saka influence rather than an expression of Saka identity, and thus I would be a bit careful with assuming that these people would have considered themselves to be part of the Saka peoples - or if they were described as such by contemporary peoples. Although it is certainly possible they considered themselves to be part of the Scythian sphere, and that their contemporaries called them Saka, we simply don’t know for sure.
What happened to the East Pamir Saka culture is also not clear. The final dates of these burials would be contemporary with the conquests of Alexander and the Graeco-Bactrian kingdom. What follows after that in terms of archaeology is quite unclear.
There are later burials linked to the period of Saka and Yuezhi-Kushan migrations, but there isn’t nearly enough information to determine if the end of the Pamir Saka signified that they had migrated away, or that their material traditions had changed. I personally think the second is more likely. The region probably became dependent on the Graeco-Bactrian kingdom due to its location on the silk road, and would have been open to all sorts of cultural influences from different directions, which could have changed the material traditions of the people.
As far as I know, the first major political state to have included eastern Pamirs into their territory without a doubt was the Kushan empire, using it as a springboard for military campaigns in Xinjiang. Later on the eastern Pamirs were also an important area for the Hephthalites.
I’m just spitballing here, but maybe they were the Parasama Kambojas of Indian epic literature? Many scholars have placed this warlike group of Kambojas in the Gorno-Badakhshan region. Later on Greek historians called the peoples which inhabit the mountainous regions of Sogdiana and Asian Scythia Komedes or Komroi. These people also had archaeological ties to South Asia, so it is probable that their existence would be known to South Asians and if so Parasama Kamboja would make a good candidate.
Whoever they were, the Eastern Pamir Saka do not have much of a connection to the distribution of Khotanese-Tumshuqese languages in the western tarim basin though. There isn't anything to suggest a movement or even a significant material influence from these peoples in the Tarim Basin where we find these languages in the early middle ages. At best you have vague similarities in material traditions such as pottery. The presence of this Saka-associated material in western Xinjiang should not be connected to a hypothetical movement of Saka nomads into the Tarim Basin.
Ili valley Saka
You do have real genuine Saka sites in Xinjiang by the way, quite a good amount actually. These sites are in the northwest of Xinjiang, around the upper Ili river valley and the Tian Shan mountains.  Although part of Xinjiang, the Ili valley region is at a considerable distance from Khotan and Kashgar with some serious geographic boundaries in between, and is fully part of the steppe region, and thus a presence of Saka here should not be conflated with a presence of Saka in the Tarim Basin.
The iron age burials in the upper Ili valley constitute a material culture dubbed the Suodunbulake culture, and this material culture pretty much covers the whole first millennium BC. These archaeologists argued that the culture can be divided in three stages, each represented by the type sites of Qiongkeke, Suodonbulake and Yeshenkelieke. 
The Qiongkeke period would be the first phase of this material culture and has been dated to 1000-800 BC. Qongkeke I is a late bronze age Andronovo site with settlements, and is classified as being part of the Fedorovo subgroup of Andronovo sites. It is followed by Qiongkeke II in the early iron age, which had a different material culture. There were no settlements dated to this period, and some of the burial mounds were placed directly on the remains of the settlements of the bronze age Andronovo sites.
Perhaps this Deer Stone found in the region could be attributed to the Qiongkeke period, as Deer stones were part of the material traditions of early iron age Scytho-Siberians.
The Suodonbulake period is the prime era of this culture, and dates to 800-400 BC. There is an increased wealth disparity in the societies, the geographic distribution of the material culture magnifies, and the displayed wealth as well as the size of the burial mounds increase during this period also.
The Yehenlieke period is dated from 400 BC to the common era, although some argue for a 300 BC starting point. This is the final phase of the Saka culture in the upper Ili, and it is marked by a decline in geographic distribution, and new burial forms start appearing during this period also. This period is synchronous to the Saka-Wusun culture of eastern Kazakhstan, which both represent the population of this region as the Yuezhi and Wusun migrated into the Ili valley, leading to the Saka migration discussed earlier.
That bronze statue looks similar to the ‘ Kneeling Warrior”, dating to the 5-3rd century BC, also from Northwestern Xinjiang.
The Saka who lived here were in probability the exact Sai groups the Han became familiar with, and they were part of a much larger sphere of Saka culture around the Tian Shan. Saka Haumavarga, Saka Para-Sugd, Sai, Shakya, many names for the people of that cultural sphere.
I already outlined bits of the history of the Saka migrations, but both the book of Han and local archaeological sites show that not all the Saka migrated away, many in fact remained but became part of the political entities such as the Yuezhi-Kushans and the Wusun. Over time, the descendants of these Saka would have stopped being identified as such, and would merge into the Wusun and Kushan identities. Later on new political powers and identities such as the Xionites and Hepthalites arose, and the same thing happened again. Then with the Turkification of the steppes, the same thing occurred.And as new Turkic entities arose, the same process occurred again. Such is the way of the nomad.
Anyhow, the proposed archaeological evidence of a Saka presence in the western half of the Tarim Basin is rather weak and is mostly reliant on material influence, traded artefacts and loose interpretations of archaeological data and historiography. On many occasions the suggestions seem to serve as an explanation for an already accepted narrative of a Saka migration into the Tarim Basin, bringing the Khotanese language with it, which is where the loose interpretations come in.
The proposals that connect these Saka archaeological signs to the Khotanese Saka language are also not in line with the original hypothesis for a Saka origin of the kingdom of Khotan. Khotanese was precisely linked with the movement of Saka peoples in Central and South Asia due to the linguistic identical nature of Khotanese and the language of the Indo-Scythians. Whereas many of these influences are firmly in the earlier phase of the iron age, and can be found all over Xinjiang amongst various material cultures and sites, crossing several ethnic and linguistic boundaries and thus showing no direct connection to the distribution of the Khotanese-Tumshuqese languages.
Not to mention that the Saka were a historical peoples and there is little to suggest that any of the putative influence would’ve come from nomadic populations that were considered Saka, rather than other groups of Iranian steppe nomads.
Iron age societies of the western Tarim Basin
The sites of the final bronze age western Tarim Basin in all probability were derived from Andronovo populations, but as far as I know there are no official Andronovo sites within the Tarim Basin itself. There are Fedorovo sites on the Tian Shan in Xinjiang, so I’m imagining that these would also be the Andronovo populations to primarily move into the Tarim Basin during the later bronze age, although there also could have been movements coming from Siberia, by way of the Karasuk or related cultures.
The lifestyle of these bronze age pastoral peoples continued during the early iron age, starting from the tenth century bc. Many of the pastoralists would have lived in the foothills of the Tian Shan, Pamir and Kunlun mountain chains, which would be relatively hospitable compared to the Taklamakan desert. The Liushui cemetery, dating to the 1020-926 BC is a prime example of such. It is assumed that the site is the summer pasture of a mobile pastoral population, which likely practises vertical seasonal pastoralism, and a significant portion of their diet came from meat and dairy, with limited grain consumption. 
Some of the weaponry, particularly arrowheads, were described as similar to those of Arzhan. The pottery was linked to Karasuk. Horse riding was practised, and the horse mandiles and bits found in some graves imply it was considered prestigious in their culture.  This is a clear example of what I described earlier, that the populations in the Tarim Basin very early on had connections to the iron age Scythian cultures and their bronze age predecessors. The people at Liushui may have been derived from bronze age migrants from those regions, or only influenced by them. It is impossible to say without genetics. Perhaps these were the people responsible for the Scytho-Siberian influences we see during the late bronze age and early iron age?
However, the iron age of western Xinjiang was a period of change, and the continuation of the bronze age pastoral lifestyles was about to be impacted by this change. We first see this in the most western areas of Xinjiang in the Pamir mountains. At Xiabandi, a late bronze age site inhabited by Andronovo peoples (and also the predecessor of the Pamir Saka sites in Xinjiang I discussed earlier) the population can be divided into three groups based on their diet.
The first group to which the vast majority of the samples belonged was one highly dependent on dairy and meat, characteristic of the Andronovo culture. The second group had a higher intake of wheat. The third group was significantly more reliant on millet. However there was no visible difference in terms of burials between these groups.
Another good example of this change can be found at the Wupaer site on the western rim of the Tarim Basin, near Kashgar. During the later bronze age phase we have evidence of pastoral populations whose agricultural products were limited to barley and wheat. The period from 1200-400 BC was radically different to the preceding period, as agricultural influences from the southwest came in and the population became significantly more dependent on their crops. The population sizes increase significantly, and crops not only become more prevalent, they also become more diverse.
During the early phase of the iron age in the Tarim Basin, the climate developed to be conducive to high population oasis settlements. The increased temperature caused the glaciers in the mountain chains to retreat, which led to an increased water discharge into the basin . Along the rivers the riparian forests in the Tarim Basin grew in size. As the climate developed to be more favourable it drew people in, and large agricultural settlements started to spring up over the Tarim Basin.
Making use of elaborate irrigation systems, the agricultural peoples were able to thrive in environments their forebears would have seen as undesirable wastelands. Within a few centuries of this transformation do we see the development of fortified towns, sometimes referred to as ‘cities’ in western Xinjiang. These would be the predecessors to the many cities and city states of the Tarim Basin.
Djumulak Kum or Yuansha City, which I mentioned earlier, was one of those oasis settlements and interestingly this one was nearly within the centre of the Taklamakan desert. This passage describes the inhabitants well :
Although they exploited the resources of their nearest environment, the inhabitants were far from isolated, and their city was very different from the Daheyan of today. The archaeological material reveals that the residents had connections with areas of the neighboring piedmonts to the north and south, and, before the Silk Road, with regions farther away, including Chinese mainland, the Indo-Pakistani subcontinent, and Central Asian oases and steppes. Who, then, were the inhabitants of this city, buried outside its walls? They likely were farmer-shepherds, weavers, and craft workers who used irrigation and grew grain well before the Han and the arrival of Chinese settlers from the second to first centuries BC.
These farmer-shepherds were raising cattle, sheep, goats, horses, camels and chicken. Their irrigation networks were described as extensive, covering several kilometres. Consisting of small canals that would join together, this canal network allowed for irrigation over large territories and for societies without a state (e.g a clan or a village) to to all make use of this irrigation system. Wheat, barley and millet would be main crops grown in these systems.
Centuries later at Karadong, a site not too far from Djumulak Kum, do we see the development and continuation of this irrigation centred agricultural lifestyle. The settlements at Karadong were surrounded by gardens, orchards with fruit trees and blackberry bushes, vineyards and irrigated terraces.
Remains of a circular water tank at Karadong.
However it's interesting to note that by this time the degradation of the area was already noticeable as the regions surrounding Karadong had already gone through significant desertification again. Eventually the settlement would be abandoned, and swallowed up by the desert.
You can notice the desertification if you look at the Tamarisk and poplar mounds that grow into the region:
For comparison, see the snapshot of three Tamarisk brushes below. The lower alluvial plain is further south and within a more deserted area.
I got sidetracked for a bit there talking about foliage mounds, let’s refocus on the topic at hand.
> Yuansha City’s establishment was among those oasis settlements [12, 16] sprung up in the Tarim Basin shortly after 2.8 ka BP, when the Mediterranean populations came into the Tarim Basin and started the Iron Age civilization . It was coeval with neoglacial retreat of ~2.8 ka BP’s in the mountains surrounding the Tarim Basin [34, 35]. During the onset of warm climates, the river’s (meltwater driven) discharge was probably increased [36, 37], and it might have facilitated oasis development and provided suitable conditions for the immigrated civilization.
The article mentioned a migration of “mediterranean peoples” which in this context likely means a movement of populations with significant ancestry from bronze age southern central asian agricultural populations, credited with starting the iron age civilization of the Tarim Basin, which I assume refers to the foundation of the oasis towns.
From what I can tell, the populations of the western Tarim Basin during antiquity and the middle ages were of the same anthropological type, and thus they would have had a degree of genetic similarity or would’ve been derived from the same people.
If a migration of agricultural people into the Tarim Basin occurred I think the best candidate would be the Chust culture of early iron age Central Asia, centred around the Ferghana valley.
The Chust culture was a late bronze/early iron age material culture centred in the Ferghana valley. The economy of the Chust culture was one of settled, agricultural populations. Their exact origins are a bit unknown but it is quite obvious to tell that the Chust culture has a material and genetic relation to the contemporary agriculturalists of southern central Asia, such as the Yaz culture. This would immediately make the Chust culture a viable candidate for an Iranian speaking community. Apparently they migrated into the Ferghana valley and had contacts with their more pastoral northern neighbours, also of Indo-Iranian origin.
The physical remains of the Chust culture were also described as being of the Mediterranean type, differing from the preceding inhabitants of the Ferghana valley which resembled the Andronovo more in terms of physical characteristics. The Pamir “Saka” were described as having similar physical characteristics, as were the inhabitants of southern Central Asia of the iron age and also the inhabitants of Khotan.
Skull of a woman from the Chust culture
The only thing I noticed though is that the pottery of the Tarim Basin wasn’t painted, whereas the pottery of Central Asian agriculturalists, including Chust, generally was painted. Luckily this exact topic was addressed in the book “The Cultures of Ancient Xinjiang, Western China: Crossroads of the Silk Roads”, which also suggests that the Chust, Burgulyuk and Aketala could be seen as being part of a single cultural phenomenon :
> Moreover, besides the complete identity of the stone tools in the painted pottery culture of the Fergana valley and the Aketala cultural group, at least four sites of the Chust culture (all in the east of the Fergana valley) yielded, besides the painted pottery, grey-black ceramics, which could be compared to the Aketala group (Zadneprovskiy 1962: 28; Zadneprovskiy 1997: 51). Would it then not be more appropriate to consider the Chust culture of Fergana, together with the Burgulyuk culture of the Tashkent region, simply as the local western version of one cultural phenomenon that extended into the Tarim Basin up to the Late Bronze and Early Iron Age? It seems that the separation of regions with highly similar archaeological complexes is due rather to the history of research, caused by the order in which phenomena were studied by different research teams located within different contemporary administrative borders.
By the way, Peyrot was ahead of me here. I didn’t realise it at first but the Aketala site and the similar sites in Xinjiang Peyrot referred to are argued to be part of the Chust culture as you can see above. It seems like a great candidate for the dispersal of a single or very closely related Iranian languages, especially as they shared the same niche pottery tradition across their spread implying a close relation to one another. Furthermore the distribution of these grey-black ceramics perfectly correlates with the known historic distribution of Khotanese-Tumshuqese languages.
One thing interesting to note is that Satma Azar (9-6th century BC) had phallic objects similar to those of the early bronze age Xiaohe culture, and that buried individuals had similar face paint markings. As this clearly involves a degree of cultural continuity over the centuries, to me it suggests that newcomers and earlier inhabitants in the Tarim Basin had contact and influenced one another, and that the ethnogenesis of the iron age inhabitants of the western tarim basin was the result of various populations intermixing over a period of centuries.
Although quite a bit to the west of the Tarim Basin, I think that for the topic of potential linguistic dispersals an interesting archaeological site to look at would be the Jirzankal cemetery in southwestern Xinjiang. Given the clear presence of exposure burials and fire altars it has been suggested we are looking at a site with a zoroastrian religious presence. Jirzankal is also the oldest site with evidence of cannabis with enhanced THC potency , which I think is pretty nifty.
However there are burials with different rites and structures such as primary burials covered with stone kurgans which were not commonly used by Zoroastrians which might reflect a multireligious society at Jirzankal or perhaps a population that was in the process of adopting Zoroastrian rites. The people of Jirzankal were also described as a mobile, pastoral population with a dietary focus on animal products .
The population at Jirzankal also had a considerable amount of migrants, with one third of the samples having isotope values inconsistent with local origins. The people at Jirzankal were described as having both western and eastern eurasian features , although it is unclear if this refers to people of mixed ancestry or the presence of both western and eastern peoples at Jirzankal.
Given the aforementioned '”Daevic'' elements of the Khotanese language, one might wonder how relevant the possible Zoroastrian site of Jirzankal would be to the Khotanese language. I think it has little to no relevance to Khotanese, but perhaps it could explain the presence of the Sarikoli language in Xinjiang?
Jirzankal is very close to modern day Tashkurgan, a town in western Xinjiang mainly inhabited by Sarikoli people. During the middle ages it was the capital of a medieval kingdom called Sarikol, to which the Sarikoli owe their name. In the Han dynasty it was a town of the Puli kingdom. That being said it is also possible that the ancestors of the Sarikoli only came to these regions later, as the Jirzankal cemetery dates to the middle iron age.
Despite having converted to Ismailism centuries ago, it seems that the Sarikoli people of Tashkurgan have kept some of their old Zoroastrian traditions alive, which you can read more about here: Chasing the Fire-Worshippers of Tashkurgan
To summarise this all, my current best estimate would be something like this:
During the later bronze age and the earliest phase of the iron age, speakers of unknown Indo-Iranian languages lived in the western Tarim Basin. Could be Indo-Aryan, could be Iranian, could be both, could be neither, who knows?
From around 1200-900 BC there is an entry of agricultural populations into western Xinjiang. These people seem to be related to the people of the Chust culture, a material culture from Central Asia which had been connected to Iranian speaking peoples.
During the early part of the iron age, these agricultural people spread through the Tarim Basin, with a distribution that overlaps perfectly with the later attested Iranian languages at Khotan and Tumxuk, as well as places such as Yarkand and Kashgar where the same language was spoken. The agricultural settlements of the iron age agricultural population grow into towns, and later turn into city-states. I’d imagine there was a period of several centuries of coexistence between these oasis farmers and the preceding agro pastoral populations of western Tarim Basin across the iron age.
After 500 BC do we see movements of Zoroastrian religious practices towards western Xinjiang. East Iranian languages such as Sarikoli could have been spoken by the people spreading the religion. Tashkurgan became an important centre for zoroastrianism by the time of antiquity. I’d suggest that this is a cutoff date for when Khotanese could have entered the Tarim Basin, because the historical distribution of Khotanese as well its relative Wakhi clearly suggest that this linguistic grouping lay to the east of the other East Iranian languages, and was separated relatively early from them. Details in the Khotanese language suggest that this split happened before the “Daeva-Asura flip” characteristic of Zoroastrianism.
Kingdom of Khotan
The founding myth of Khotan states that it was founded by colonists from Taxila during Ashoka’s reign in the 3rd century BC. There are four versions and they all differ on the details. For example in one version the city was founded by Kunala, the son of Ashoka but in another it was founded by noblemen from Taxila who were exiled after having blinded Kunala. At the same time, a Chinese colony was formed in the east, and these two groups eventually merged, laying the foundation for Khotan being divided into a western and eastern city.
Given the lack of historical descriptions from this era, it is a bit difficult to state to which degree this foundation myth reflects the truth. Some see this myth as an explanation for the Indian and Chinese cultural influences in Khotan, as they do not mention the Iranian peoples of Khotan. The usage of Kharosthi script in Khotan has been linked to the origin story, but others have argued for it to have come from influence from the Kushan period for, as Khotan and the Kushans were affiliated, and the Khotanese joined the Kushans on campaigns in South Asia.
In regards to the foundation myth, the coin of Gurgamoys from the first century AD features Prakit on one side and Chinese on the other:
In the early first century AD Khotan was ruled by Yarkand and it had a population of roughly 20.000 people. Khotan revolted, eventually conquering Yarkand themselves and ended up controlling an important portion of the Silk Road. A bit over a century later and the population of Khotan had quadrupled, now being over 80.000.
The Han exerted their powers into the Western regions and subdued Khotan during the Han-Xiongnu war period in the first century AD. Other city states in the area temporarily came under the control of the Xiongnu or the Wusun as well. Eventually the western regions rebelled and temporarily became independent of Han rule. Khotan then seemingly became a vassal of the Kushan empire, and joined in campaigns to South Asia.
This is the period that the Silk road properly gets going, and Khotan goes through a heck of a time. Around this time exotic goods such as the Sampul Tapestry now flow into the Tarim Basin:
Another legend relating to Khotan is that of how silk was introduced in the area, detailed by Tang dynasty monk and Scholar Xuangzuang. Aurel Stein described the myth in his works on ancient Khotan. According to Stein, this story was quite popular in Khotan, finding itself represented on a painted panel he discovered . The story goes as follows:
“To the south east of the royal city, five or six li-’. so the Hsi-yil-chi tells us, there was a convent known as the Lii-shc Saiigluirama, founded by ilie queen of a former ruler, to whom tradition ascribed the introduction of sericulture into Khotan. In old times the country knew nothing of either mulberry trees or silkworms. Hearing that China possessed them, the king of Khotan sent an envoy to procure them ; but at that time the ruler of China was determined not to let others share their possession, and he had strictly prohibited seeds of the mulberry tree or silkworms’ eggs being carried outside his frontiers. The king of Khotan then with due submission prayed for the hand of a Chinese princess. When this request had been acceded to, he dispatched an envoy to escort the princess from China, taking care to let the future queen know through him that, in order to assure to herself fine silk robes when in Khotan, she had better bring some mulberry seeds and silkworms with her.
The princess thus advised secretly procured mulberry seeds and silkworms’ eggs, and by concealing them in the lining of her headdress, which the chief of the frontier guards did not dare to examine, managed to remove them safely to Khotan. On her first arrival and before her solemn entry into the royal palace, she stopped at the site where subsequently the Lu-shê convent was built, and there she left the silkworms and the mulberry seeds. From the latter grew up the first mulberry trees, with the leaves of which the silkworms were fed when their time had come. Then the queen issued an edict engraved on stone, prohibiting the working up of the cocoons until the moths of the silkworms had escaped. Then she founded this Sanghārāma on the spot where the first silkworms were bred; and there are about here many old mulberry tree trunks which they say are the remains of the trees first planted. From old time till now this kingdom has possessed silkworms which nobody is allowed to kill, with a view to take away the silk stealthily. Those who do so are not allowed to rear the worms for a succession of years.”
During the period of the Tang dynasty Khotan was subdued and became a vassal state of the Tang dynasty, but then the TIbetans invaded in the 8th century AD.This happened again a century later, when Khotan became part of the Tibetan empire. In 851 Khotan managed to become independent, and then remain the only state in the Tarim Basin to not be conquered by various Turkic-ruled entities, peaking in size around 1000 AD.
It was around this time that the Buddhist kingdom of Khotan was invaded by the Kara-Khanid Khanate. After a long war which had some back and forth moments, Khotan lost and was conquered by the year 1006 by Yusuf Kadir, khan of Kashgar. This process lead to the spread of islam in the region, as well as the spread of eastern Turkic (Karluk) languages which are spoken by the Uyghurs today. The language of the historical Uyghurs, Old Uyghur, is actually a Siberian Turkic language which is a different sub-branch of Common Turkic.
In later parts of the 11th century AD it was noted that the people of this region still spoke Iranian languages by Mahmud al-Kashgari :
The most elegant of the dialects belongs to those who know only one language, who do not mix
with Persians, and who do not customarily settle in other lands. Those who have two languages and who mix with the populace of the cities have a certain slurring (rikka) in their utterances - for example, Soy-daq, Kancak and Aryu. The second category are such as Khotan, Tiibiit and some of Tangut - this class are settlers in the lands of the Turks. I shall now outline the language of each of their groups.
The language of the people of Jabarqa is unknown because of their distance and the interposition of the Great Sea between them and Mafln. The people of Masin and of Sin have a language of their own, although the sedentary population know Turkic well and their correspondence with us is in the Turkic script. Also the language of Yajuj and Majuj is unknown because of the Barrier and the interposition of the mountains and the sea that is near Masin. Tiibiit have a language of their own. Khotan also have both a script and a language of their own. Both of these do not know Turkic well.
However in the decades to come Khotanese was spoken less and less, and by the end of the 11th century AD the region was completely turkified. Unfortunately during the conquest of the Tarim basin states and the ensuing islamification, many of the beautiful frescos and statues were defaced and destroyed. The same also occurred further east with the Buddhist Uyghurs kingdoms such as Qocho.
Al-Kashgari also recorded this poem relating to this exact topic:
We came down on them like a flood,
We went out among their cities,
We tore down the idol-temples,
We shat on the Buddha's head!
As you can see, the story of Khotan and the neighbouring city states can hardly be described as a migration of Saka nomads, which then spread all over the western tarim basin to found several kingdoms that all spoke the same language. The populations in the western Tarim Basin started out as agro pastoral peoples, but during the iron age there was a heavy agricultural shift and their societies gradually developed into proto-urban communities, which were the founding blocks of the city-states of the Tarim during the antiquity, city-states such as Kashgar and Khotan.
Scytho-Siberian style artefacts and material influence were present in the Tarim Basin since the onset of the iron age, and even in the bronze age this region was connected to the Siberian cultures. Actual archaeological sites that show a clear affinity to the archaeological imprints of the Saka around the Tian shan or the pamirs, suggestive of a presence, do not seem to be within the Tarim Basin, which would be consistent with the Han dynasty era historical records which despite popular belief do not mention Saka invading or even migrating to the western Tarim Basin. Neither is there evidence of such an invasion in the centuries prior.
Thus to sum all of this up, the notion that Khotanese was a language of Saka nomads which formed a kingdom in the Tarim Basin is one that needs to be strongly reconsidered by academics in my opinion.
Evidence for such a migration is not to be found in historical sources and cannot be confirmed with currently known archaeological data. From a linguistic point of view, the claim that Khotanese is identical to the language of the Indo-Saka is at the very least debatable, and is probably not the case going by the fragments of Indo-Saka attestations we have. The areas where these languages were found or were assumed to have spoken are all intensely connected to iron age agricultural populations rather than steppe nomads.
From where I stand, the “Buddhist Saka kingdoms” of the Tarim Basin are nothing but a modern historical invention. Which is a quite strong statement to make, I know, but given the points I laid out above relating to the historical, archaeological and linguistics elements of the debate there is quite some weight behind that statement.
As I had nearly finished this entry, a new bomb dropped in terms of genetics from Xinjiang. Interestingly many of the sites and cultures mentioned earlier were present in this article.
I covered quite a bit of the article in a previous blog entry of mine:
I will give them a brief overview here as well. I’ll try and keep the breakdowns as simple as can be, particularly because many of the samples have mixed origins and can be modelled all sorts of ways when using proximate sources.
Wutulan - Soudunbulake
Wutulan is a site of the Soudunbulake culture, the material culture of the Saka in the Ili valley I briefly discussed earlier. These are the eastern wing of the people who were called “Sai” in the Hanshu, before their migration deeper into Central Asia.
As you can see from the bronze and iron age models, the ethnogenesis of these samples can be roughly described as 50% Siberian steppe nomads and 50% Central Asian agriculturalists. The Chust culture might also be important for the Tian Shan nomads as a substrate, since they migrated into the Ferghana valley from the south.
The vast majority of samples from this article come from the same region and period as these samples and look quite closely related to them if not downright identical. It is likely that they were Saka too, but for some reason their burials weren’t classified as being part of the Suodonbulake culture.
Furthermore, they are pretty much identical to the samples from Kyrgyzstan labelled as “Tian Shan Saka” in previously released articles. This shows that the archaeological affinity this culture has to the Saka material cultures on the steppes around the Tian Shan mountains in Central Asia is also reflected genetically because they basically are the same peoples.
Unfortunately only two of the samples from Liushui, the early iron age site with mounted pastoralists were available, and some more data would have been helpful in this case.
C1235 has to be one of the all-time strangest samples I’ve come across. It looks like all the genetic substrates of bronze age Indo-Iranians thrown together but without the Indo-Iranian ancestry to match. And crazy enough this sample is from the iron age!
Khovsol_BA are samples from the Deer-Stone Khirigsuur culture of northern Mongolia, Aigyrzhal_BA and Dali_EBA are from Kyrgyzstan and Kazakhstan respectively and would be realistic proxies for ancestries from populations around the inner-asian mountain corridor.
The Y-DNA haplogroup of this sample was Q-L330, which was a Y-DNA haplogroup typical of these south siberian and Northwest Mongolian pastoral populations, and is later found carried by Scytho-Siberian populations across the iron age steppes.
The other sample from luishui had Q-Y6826, which is a different sub branch of Q1b, and could even have come from a steppe ancestor depending on the subclade, as Afanaisevo samples also carried Y-chromosome haplogroups downstream of Q-Y6826, and have a shared clade with one Corded Ware sample from Bohemia.
The results (Table 3) allow us to conclude that the whole phase of activity started between 1108 and 893 B.C. (95% probability range) or most likely between 1017 and 926 B.C. (68%). The directly dated tombs span the interval between 149 and 414 y (95%) or most likely between 173 and 298 y (68%). The whole phase seems to have finished between 760 and 493 B.C. (95%) or most likely between 750 and 630 B.C. (68%).
It could be that around this time a Siberian-Central Asian admixed population entered the tarim basin and settled on the southern rim. I have no idea how this would have occurred, but it must’ve happened some way. Considering that the material culture of the Deer Stone Khirigsuur is pretty much an Andronovo/Karasuk derivative, this may have implications for the introduction of siberian-style metallurgy, deer iconography or other traits associated with iron age steppe nomads into the Tarim Basin.
Jirzankal also was rather surprising. I would’ve imagined that the predominant people there would’ve been similar to modern day Pamiri peoples, with a sizable amount of pilgrims from both the west and the east. That didn’t exactly turn out to be the case though. C1192 was close, but the rest of the samples from Jirzankal sport a rather unique genetic profile.
Given Jirzankal’s proximity to the “Pamir Saka”, I wondered if I erred big time with that whole big section on why I think Pamir Saka weren’t actually Saka. But then I got to thinking. Physical anthropology has fallen a bit out of favour out here in the west, but I’m of the opinion that bones do not lie. And the physical anthropology of the East Pamir Saka painted a very different picture than what we are seeing here, which were noted to be very similar to the contemporary populations in southern Central asia and even northern parts of South Asia. Given the way that Scythian physical remains tended to be described I think the ~22% (going up to 28%) Siberian ancestry would have been noticed at least if it was a common component of the Pamir Saka peoples. It was noticed for the samples at Jirzankal though which were described as having mixed western and eastern eurasian features, which is consistent with their genomic profile.
All these samples had a quite noticeable amount of Siberian ancestry, which is characteristic of Scytho-Siberian populations. Yet if you look at their amount of steppe ancestry, it is quite low, which is rather uncharacteristic of Scytho-Siberian populations. If the Siberian ancestry at Jirzankal came from Scytho-Siberian peoples, it would probably be the case that the donor population would at the very least have a steppe-to-siberian ancestry ratio comparable to Tasmola, Pazyryk and early Cimmerians (1-to-1 basically).
Given the low amount of steppe ancestry, in such a scenario these populations would pretty much be the sole source of steppe ancestry for several of these samples, with the rest of the ancestry coming from un-admixed IAMC and BMAC populations. Under normal circumstances I’d call that a highly unlikely scenario for the iron age. But normal circumstances don’t seem to apply to prehistoric Xinjiang. Staying a bit grounded however, we don't have evidence of any unadmixed BMAC or Aigyrzhal_BA type population in the Tarim Basin or Central Asia by the iron age and we know that the Scytho-Siberian expansions only commenced around the end of the 11th century BC.
I think the samples seen at Liushui could be of relevance though. If you consider the geography, the Siberian ancestors of the Liushui samples very likely followed a migration route that went past the eastern Pamirs and if not, it wouldn’t have taken long for contact to have been established. Especially given that the people at Liushui were mounted.
If such a population is considered to be the main source of Siberian ancestry, then the steppe-to-central-asian ancestry ratio would also be consistent with late bronze age/early iron age Indo-Iranian populations from southern Central Asia as the main source for steppe ancestry, with the Siberian ancestry coming from people similar to sample C1235 from Liushui.
What is interesting is that many samples here were under Q1b. The Liushui sample had Q-L330,whereas one of the jirzankal samples had Q-YP4000 which is another Scytho-SIberian lineage that in likelihood is connected to populations such as those of the Deer Stone Khirigsuur complex. The presence of multiple Siberian Q lineages could suggest that the Liushui man wasn’t a loner, and that he was part of a larger population that reached the southern rim of the Tarim Basin. C1192 also had Q1b, but he had a completely different subclade which has also been found in several (related) Afanasievo samples and in one early Corded Ware sample from Czechia. This clade is undoubtedly linked towards its steppe ancestry.
In which context these people came here is hard to say, but it was noted that several of the Jirzankal samples had isotope ratios not consistent with them being locals to the region. Three genetic clusters were present at Jirzankal, and we can’t say which one were locals and which ones were migrants.
What is interesting to note is how C1192 is the only sample to have no excess Tarim_EMBA ancestry, has more steppe ancestry and has more central asian farmer ancestry than the rest of the bunch. This makes me wonder if he was a migrant from further west or partially derived from migrants. Or it may be that this individual was a local and that the other samples were recent migrants from the east, although given that they are the majority this seems to be a bit unlikely.
These Jirzankal samples are quite a puzzling bunch. From a genetic standpoint I think you can describe their ethnogenesis in two manners. Either as a Scytho-Siberian population migrating through the Pamirs, probably from the west, who then significantly intermix with a genetic relict population of the bronze age that essentially is just like Aigyrzhal_BA or even BMAC-like. The alternative is that a population with significant South Siberian ancestry but with a negligible amount of steppe ancestry reached the southwestern Xinjiang and became intermixed with the other migrants to the area. Given that we have evidence of the latter but not really the former, I find this the more likely scenario but I would need more data to properly confirm or deny this.
The Siberian ancestry could explain the usage of stone burial mounds at Jirzankal as well as the pastoral lifestyle and diet practised by the people of Jirzankal. The pottery used, as well as the fire altars and musical instruments could be attributed to the Central Asian side of the people at Jirzankal.
Zaghunluq, the burial site which contains the world famous Cherchen Man or Ur-David, is not only interesting from an archaeological perspective, but also from a genetic perspective. These peoples had a high frequency of R-PH155 carriers present at the site. This was the paternal haplogroup of the early Tarim Mummies, descendants of a small population of foragers which migrated into the area, still going strong more than a thousand years later and after so many different waves of migration had entered the region.
The Tarim_EMBA ancestry comes in quite heavy in this model, but some of this is more likely going to be representative of ancestry from ANE-rich populations on the IAMC rather than being actual ancestry from the Tarim Basin natives as the ancestors of the Zaghunluq peoples definitely would have mixed with those upon entering the Tarim Basin. That said, there definitely was a contribution of the Tarim_EMBA peoples to their genome. Here is a model with TKM_Gonur1_BA_o included:
Further points are that their steppe ancestry was not very high, and that their east eurasian ancestry seems best mediated by upper_yellow_river_ln and Ulaanzuukh_LBA, which you can see as geneflows from the tibetan plateau and the gobi desert respectively. One of the samples had Y-dna O1a. There might also be a small amount of khovsgol_ba present but they model equally well without it, there is practically no difference between the average distances with Khovsgol_BA included or excluded, thus it may me that the percentage of khovsgol_BA ancestry seen overfitting.
The Zaghunluq mummies such as the Cherchen Man have been connected to various peoples. The site is located on what pretty much would be the border region between “Tocharian C” and Khotanese-Tumshuqese speakers and has been attributed to both. A connection to Saka has been made too based on the riding equipment and clothing.
However from their genomic ancestry we can tell that the Zaghunluq peoples in terms of steppe ancestry were mainly Indo-Iranian derived, and had significant ancestry from Central Asia and the Tarim basin, but not so much from South Siberia and with a genetic profile that is considerably different from those of iron age Scytho-Siberians. The ancestry of the Zaghunluq samples reaffirmed what I said earlier about this archaeological site likely having nothing to do with Saka nomads, but everything with the iron age native inhabitants of the Tarim Basin.
I think you could even make the argument that these may not have been Indo-European speaking peoples, but I find that a little unlikely. Certainly possible though. If they spoke Indo-European languages, it would certainly be some Indo-Iranian dialect.
Hetian - Sampul
There also were a few samples from Hetian and Sampul, areas where Khotanese would have been spoken at the time. Were these samples Khotanese speakers and if so, were they genetically representative? That unfortunately can’t be said for certain, especially not as this period saw both the rise of Silk road trade, but also a tumultuous period of history with migrations and conquests such as those by the Xiongnu, Wusun or the Kushans. Therefore it may be that some of these samples were not native, or perhaps had parents/ancestors which came from other regions.
That being said, this is about as good insight into the area as we are going to get, so let’s have a look.
Using two more references gets me something like this, but at this point I am bordering on overfitting. Not much of a difference anyways.
The two main components seen are Steppe_MLBA and BMAC related ancestries, in comparable frequencies to one another. Then you have East Asian ancestry represented by Upper_Yellow_River_LN which are samples from the Longshan culture . When using a single reference this was the one with the best results, but it might be that there is variation in terms of east eurasian source e.g Tibeto-Burman or Sinitic speaking peoples. This is followed by Tarim_EMBA which could represent actual ancestry from the Tarim Basin, or it might function as a proxy for the many ANE-rich central Asian populations. Finally you have south Siberian related ancestry represented by Khovsgol_BA. The typical siberian ancestry of Scytho-Siberians unfortunately cannot be used as a tracer dye for Scythian related ancestry in this area due to the presence of such ancestry at Jirzankal and Liushui, to the west and east of Hetian respectively. The presence of Siberian ancestry in the historical periods could be ascribed to these populations on the southern rim of the Tarim Basin.
It might be the case that some samples may have a bit of ancestry from the iron age steppes, such as C388 and particularly C3622. At this point in history we have a period where several nomadic or formerly nomadic entities such as the Xiongnu, Wusun or Yuezhi-Kushans had either conquered, vassalized or temporarily settled in kingdoms within the Tarim Basin. People with significant amounts of this Siberian ancestry would have been present amongst all of these three, with the Wusun likely containing the highest number of people of Central Asian “Saka” origin in their realm amongst them. This means there is by no means an exclusive connection to Scytho-Siberian ancestry and the “Saka of Khotan” for lack of a better term.
However I can also model these samples as a combination of central asian populations and populations from the tarim basin without the input of Scytho-Siberian steppe nomads with similar distances, thus I wouldn’t be certain that there is actual “central Asian Saka” rather than it being the result of overfitting.
It is uncertain which populations should be seen as the source, but what is certain is that these individuals were not of significant Scytho-Siberian origins and not consistently so either. These samples did not seem to have significant South Asian ancestry either, and they seemingly have more Tibeto-Burman related ancestries rather than Han Chinese ancestry. Thus if these samples are anything to go by, despite the origin tales of Indian and Chinese colonists the predominant ancestors of the people of Khotan would be late bronze/iron age central asian agriculturalist populations, followed by earlier inhabitants of Xinjiang of varied origins.
The spread of the Khotanese language
The authors of the article briefly discussed the topic of the Khotanese language and its spread into the Tarim Basin:
> Further, although the spread of languages is not always congruent with population histories (32), the presence of Saka ancestry in Xinj_IA populations supports an IA introduction of the Indo-Iranian Khotanese language, which was spoken by the Saka and later attested to in this region (19)
I find this a little shortsighted, because the XINJ_IA samples they are referring to are those found in the Ili valley, rather than the western portion of the Tarim Basin itself.
Furthermore, the authors also noticed this:
> The IA also shows an increase in the frequency of BMAC ancestry in the f4-statistics comparisons (figs. S19 and S20) (21). Seven IA populations were found to contain BMAC ancestry (30 to 47%), and we observed four IA populations that could also be modeled using Indus periphery ancestry sources SPGT and two with Gonur_2BA (18 to 37%) (Fig. 3A and tables S9 and S10). The increase in the appearance of BMAC ancestry suggests a substantial movement of people from either BMAC- or Indus periphery–derived populations into the Xinjiang region during the IA (Fig. 3A), most likely through the IAMC route over the Pamir and Tianshan Mountains
And what is odd is that when it comes to Tocharian, the authors either attribute it to the Afanasievo culture, or this BMAC influx (the “Oasis hypothesis). If they were realistic about it you would understand that an iron age "BMAC influx” is a roundabout way of implying iron age Indo-Iranian admixture from southern Central Asia, as by this time the Indo-Iranian tribes had dispersed all throughout Central Asia. I don’t see why this wasn’t even brought up in connection to the Khotanese language. Oh well..
Thus to sum it up, the results from ancient genetics in Xinjiang were close to the scenarios I described before the article came out, although there certainly were some surprises. During the late bronze age and early iron age Andronovo-derived pastoralist populations migrated to the Tarim Basin and intermixed with local populations there, as well as proximate neighbouring populations around the Tarim Basin. This can be shown at Zaghunluq, where a significant amount of ancestry came from IAMC-related populations as well as the bronze age natives of the Tarim Basin, which also provided the majority of the paternal lineages found at Zaghunluq.
The presence of a population derived from late bronze age pastoralists of Siberia and northern Mongolia within the tarim Basin at Liushui is a surprising discovery, and it could have implications for the attribution of Scythian-like material influences to particular ethnic groups. The presence of such ancestry later seen at Jirzankal, in an admixed form, could suggest that this ancestry became part of the local genetic substructure in the iron age.
An iron age influx of populations with a higher amount of BMAC-related ancestry was noticed by the authors, and such ancestry is considerably present amongst historic era samples in Khotanese-Tumshuqese speaking territories, which can be seen as a mix of early iron age tarim basin peoples and early iron age central asian agriculturalist samples. This could affirm the proposition made by Peyrot that the Chust/Aketala culture of iron age Xinjiang could be the linguistic ancestors of the Khotanese people.
Hoernle, R. (1913). Manuscript remains of Buddhist literature found in East Turkestan.
Peyrot, M. (2018). Tocharian B etswe ‘mule’ and Eastern East Iranian
Artur Ambartsumyan - "In memory of Leonard Georgievich Herzenberg" // Cultural representation at the Embassy of the Islamic Republic of Iran in the Russian Federation. February 2013. p. 13.
Wagner, M., Bo, W., Tarasov, P., Westh-Hansen, S. M., Völling, E., & Heller, J. (2009). The ornamental trousers from Sampula (Xinjiang, China): their origins and biography. Antiquity, 83(322), 1065–1075. doi:10.1017/s0003598x0009935x
Baumer C. (2011). The Ayala Mazar-Xiaohe culture: new archaeological discoveries in the Taklamakan desert, China. Asian affairs, 42(1), 49–69. https://doi.org/10.1080/03068374.2011.539323
Polosmak N.V. The Pazyryk Style. Archaeology, Ethnology & Anthropology of Eurasia. 2021;49(4):43-56. https://doi.org/10.17746/1563-0110.2021.49.4.043-056
Kuzmina, E.E. (2007). The origin of the Indo-Iranians.
Festa, Marcella. (2017). Prehistoric Cultures in Xinjiang: Retrospect and Prospect. ATHENS JOURNAL OF HISTORY. 3. 137. 10.30958/ajhis.3-2-4.
Meng, Q. (2021) - A study of the late prehistoric burial records of Xinjiang, northwest China (c. 3000−200BCE)
Li, S. Y. (2014). Research On Prehistoric Archaeological Cultures In The Ili River Valley Of Xinjiang
Wagner, Mayke & Wu, Xinhua & Tarasov, Pavel & Aisha, Ailijiang & Ramsey, Christopher & Schultz, Michael & Schmidt-Schultz, Tyede & Gresky, Julia. (2011). Radiocarbon-dated archaeological record of early first millennium B.C. mounted pastoralists in the Kunlun Mountains, China. Proceedings of the National Academy of Sciences of the United States of America. 108. 15733-8. 10.1073/pnas.1105273108.
XinYu ZHANG, Dong WEI, Yong WU, Ying NIE, YaoWu HU, Carbon and nitrogen stable isotope ratio analysis of Bronze Age humans from the Xiabandi cemetery, Xinjiang, China: Implications for cultural interactions between the East and West, In Journal of Chinese Science Bulletin, Volume 61, Issue 32, 2016, Pages 3509-3519, ISSN 0023-074X, https://doi.org/10.1360/N972016-00514.
Zhang, F., Wang, T., Yimit, H., Shi, Q., Ruan, Q., Sun, Z., & Li, F. (2011). Hydrological changes and settlement migrations in the Keriya River delta in central Tarim Basin ca. 2.7–1.6 ka BP: Inferred from 14C and OSL chronology. Science China Earth Sciences, 54(12), 1971–1980.
Debaine-Francfort, C., Debaine, F., & Idriss, A. (2009). The Taklimakan Oases: An Environmental Evolution Shown Through Geoarchaeology. Water and Sustainability in Arid Regions, 181–202. https://doi.org/10.1007/978-90-481-2776-4_12
Betts, A. V. G., Vicziany, M., Jia, P., & Di Castro, A. A. (Eds.). (2019). The Cultures of Ancient Xinjiang, Western China: Crossroads of the Silk Roads. Archaeopress. https://doi.org/10.2307/j.ctvwh8c04
Ren, Meng & Tang, Zihua & Wu, Xinhua & Spengler, Robert & Jiang, Hongen & Yang, Yimin & Boivin, Nicole. (2019). The origins of cannabis smoking: Chemical residue evidence from the first millennium BCE in the Pamirs. Science Advances. 5. eaaw1391. 10.1126/sciadv.aaw1391.
Guo, Yi & Lou, Jia & Shiyu, Xie & Wu, Xinhua & Fuller, Benjamin. (2019). Isotopic reconstruction of human diet in the Ji'erzankale site, Xinjiang Uyghur Autonomous Region, China. International Journal of Osteoarchaeology. 30. 10.1002/oa.2831.
M. Wang, X. Zhang, X. Wu. Preliminary study on human bones in Jierzankale Cemetery, Tashikuergan, Xinjiang / 新疆塔什库尔干吉尔赞喀勒墓地人骨初步研究
Stein, M. Aurel (1907). Ancient Khotan Volume I
Dankoff R. (1982) - Mahmud al-Kashghari - Compendium of The Turkic Dialects: Part I