Trans Eurasia travel

Your virtual guide to Eurasia! Let's travel together!

Central Asian Identities before 1917

By Paul Bergne, The Birth of Tajikistan

As the only people in Central Asia whose language is not Turkic, present-day Tajiks proudly claim to be descendants of the early pre-Islamic inhabitants of the region. Over the last 1000 years, it is claimed, successive waves of invading Turkic nomads gradually ousted them, first from the best grazing pastures and then, as the invaders adapted to the conditions of settled agriculture, from the best arable land also. This process was summed up in a long document or "working paper" prepared for the study of the region in advance of the new Soviet government's so-called National Territorial Delimitation of 1924 which divided Central Asia into new republics on ethnic lines. In this document, Tajik history was described as follows:

The Tajiks are the only people [in Central Asia] of Iranian origin, who since time immemorial have been living in the frontiers of Bukhara and the Turkestan Autonomous Soviet Socialist Republic (see below for a description of this polity). The conquering Turkic peoples enslaved them. Part of them became totally Turkicised and adopted the language of their conquerors, while the rest, although they kept their language, took refuge in the mountainous and semi-mountainous regions of Samarkand Oblast' and Bukhara, and in the valleys of the mountain rivers and the basin of the Syr Darya and the Zeravshan where they were driven by their conquerors.

According to the great Russian orientalist, Academician V.V. Bartol'd, the use of the word "Tajik" was first recorded in the literature on Central Asia by the historian Beikhaki who reported a senior Iranian so describing his nationality when speaking to Mas'ud of Ghazni in 1039. The possibly related word "Tat" was also used to define Iranian peoples, although originally it may have been used by the nomads of Central and Inner Asia for all settled peoples. Compare Mahmud al Kashgari's mysterious quote "There is no man without a hat and no Turk without a Tat". Despite attempts by Tajik nationalists and others to link the name Tajik to the Persian word for crown "taj", the usually, but not universally, accepted explanation of the origin of the word "Tajik" is that it stems from the Persian word "Tazi" i.e. "Arab". In the early Islamic period in Iran and Central Asia, this came to be applied in a general way by the locals to all the newly arrived Muslims. But, as time went on, this application was refined and, eventually limited to the Persian element which had come to the region with the Arab army. The true origin of the word "Tat" is even more uncertain.

The Tajiks' claim to be the region's oldest inhabitants has a superficial ring of truth, although it simplifies the issue somewhat and deserves closer examination.

Many of the 8th-century texts unearthed in the excavations of Mount Mug near Panjikent in the Zeravshan valley, are written in Soghdian - an Eastern Iranian language - which suggests that the dominant culture of the period was indeed Iranian. Nonetheless, the names of Turkic rulers are recorded as having defeated the previously dominant regional power, the Hephthalites, in a battle probably near Bukhara, in the 560s AD. After that defeat, the Hephthalite kingdom (as to whose ethnic identity there is still considerable uncertainty) appears to have split up into different princedoms. Some owed allegiance to the Turks in the north, others south of the Oxus to the Sasanians, while others maintained an independent existence in Kashmir and adjoining regions until the late 7th century. Soghdian coins also mention the Turks in the early 8th century. Bactrian documents discovered in Northern Afghanistan over the last ten years also confirm that, before the Arabs arrived in 739 AD, Bactria (largely in present-day Northern Afghanistan) had already been under the at least sporadic control of Turkic overlords.

It therefore seems plausible that the invading Muslim Arabs and their Persian-speaking allies found on their arrival in Central Asia in the early eighth century AD a mixed population using Eastern Iranian languages as the medium of cultural and official communication, and dominated successively by a mixed Sasanian, Hephthalite and Turkic aristocracy.

One of the puzzles of Central Asian history is why the Tajiks of the region's cities and plains have, since the Islamic conquest, spoken a Western Iranian language (Persian i.e. the language of Fars) rather than a descendant of Soghdian, Khorezmian or Bactrian, all languages that appear to have been in common and official use locally up to that conquest. One of today's Eastern Iranian languages, Yaghnobi, spoken by a small community living until recently5 in the mountains of Tajikistan alongside the Yaghnob river (a tributary of the Zeravshan), is a clear descendant of ancient Soghdian. If the Yaghnobis are to be regarded, on these linguistic grounds, as the true descendants of the Soghdians, and the speakers of Pamiri languages in Gornyi-Badakhshan as descendants of other pre-Islamic Eastern Iranian peoples, what does it mean that most Tajiks speak the language of Fars, some 1000 miles to the southwest?

Indeed, is their language any sort of guide as to their ethnic origins?

The commonest explanation of how the Central Asian Iranians replaced their own languages - Soghdian, Khorezmian, Bactrian etc - with the language of Fars, is that the latter was brought by the large Persian element in the invading Arab army which settled throughout the region after the advent of Islam in the 8th century. Given the strong (Sasanian) tradition of imperial administration, it was natural that they took over the management of the new provinces of the Arab caliphate. Their own language became the language of officialdom and the court ("Dar" from where "Dari" - the usual name for the Persian dialect spoken in Afghanistan and, by extension of the terminology, in Central Asia) throughout the region, right up to the Soviet period.

With the passage of time, it is surmised, the bulk of the populace of the region also adopted the language. The Iran scholar Richard Frye attributes some of the dialectical difference in modern Tajik to the fact that Kulyab, the mediaeval Kuttal, in the south of the country, was ruled for a long time by the Middle-Persian-speaking Sasanians. They therefore adopted the New Persian language faster than the peoples of the north of what is now Tajikistan, who formed part of various Soghdian-speaking kingdoms.

Once Islam had taken firm root in Central Asia in the 8th century AD, it was not long before a dynasty of local governors, who became known as the Samanids, established a strong local state, owing nominal allegiance to the distant Abbasid caliphate in Baghdad but to all intents and purposes independent. The Samanids, whose rule lasted approximately from AD 819 to 999, established New Persian (Farsi) as the language of administration and culture. As already noted, it is to the Samanids that today's Tajik nationalists look back when seeking historic inspiration for the non-Turkic legitimacy of their contemporary state. The fact that the Samanid capital Bukhara is now in Uzbekistan has, since the establishment of Tajikistan, rankled with many nationalist Tajik historians. For their part, Uzbek patriots react to this Tajik grievance with a mixture of irritation and suspicion. This potential bone of contention was kept reasonably well buried during the Soviet period but, since independence, has again surfaced to cause tension between the two nations.

Over the thousand years following the arrival of Islam, the original Pre-Islamic population of Central Asia was gradually either driven from the better land, or assimilated, by a succession of invaders, starting with the Arabs and Persian-speakers, and continuing with ever more successful and numerous waves of Turkic peoples. The best known of these were led by dynasties such as the Karakhanids (who overthrew the Samanids) and Seljuqs. Others arrived as assorted tribes fighting with the Mongol armies in the 13th century and, finally, arriving in the 15th and 16th centuries, as the confederation that became known as Uzbeks. By the middle of the 19th century according to the early Russian traveller Khanykov, the Uzbeks were already the dominant group in the area. The dominance of the Uzbeks was also noted by other authorities before the revolution.

However, although from an ethnic point of view the local population became increasingly infiltrated, diluted, and dominated by Turkicspeakers, the position of Persian as the traditional language of government and continuing contacts with the heartland of present-day Iran ensured that Persian culture and language were to prove resilient in the administration and in social life. This was especially true of large centres like Samarkand and Bukhara, even where the ruling dynasty was Turkic. In the mountains of western Tajikistan, Persian continued to be spoken, while, in some remote valleys and in the Pamirs, Eastern Iranian languages also survived.

There is some evidence that, at least in the Middle Ages, relations between Tajiks and Turks were not always good. Bartol'd quotes instances of mistrust between the two nationalities in the Khorezmshah period around AD 1200. However, by the first half of the 19th century, while the Turkic peoples still thought of the Tajiks as a separate group, they seem no longer to have regarded them as in any way hostile, nor vice versa. The general consensus seems to be that, after centuries of living together in a society where ethnic distinctiveness was much less important than differences in way of life (i.e. settled as against nomadic communities) or Islamic solidarity, Tajiks and settled Turkic peoples got along perfectly well. Intermarriage was, for example, very common. Indeed, in many communities, the notion of ethnicity, as later defined under the Soviets in linguistic or ethnic terms, was either very vaguely formulated or did not exist at all.

In this Islamic Central Asian environment, where ethnicity was of little consequence, the process of assimilation between the latest arrivals, the Uzbeks, and the Iranian/Persian/Turkic/Arab cocktail of peoples whom they found, produced in due course a composite identity of mixed ethnic make-up. In terms of language, its representatives were usually bi-lingual, but eventually preferred a Turkic language strongly influenced by Persian in both vocabulary but also in vocalisation (e.g. lack of characteristic Turkic vowel harmony). This composite identity became known as "Sart".

The meaning of the term shifted over the centuries. By the Mongol/ Timurid period, "Sart" had come to mean "Tajik" and, at the end of the 14th century, the language and literature of the Sarts was described as being what wasn't Turkish i.e. Iranian. That this identification with Tajik lasted even as late as the 19th century, is supported by the account of J. Klaproth, a German traveller journeying in Kazan and Siberia in mid-century. The numerous Bukharans he spoke to told him that the Turkic peoples called them "Sarts" although they called themselves Tajiks.

By the late 19th century, possibly as a result of the increasing influence of Pan-Turkic ideas, the term was sometimes given derogatory connotations, including a popular etymological derivation from the words "sari it" i.e. "yellow dog". However, this contemptuous attitude was by no means universal. Many groups were perfectly happy to call themselves "Sarts" even though they did not always understand what this meant. This is clear from the observations of the Russian orientalist I.I. Zarubin, made while studying the so-called "Turks" who had been identified in various censuses in the rural areas of Central Asia in the late 19th and early 20th centuries. The Russian scholar S.N. Andreev had earlier identified them as a mix of various Turkic elements (Temir-Kabuk, Bakhrin, Burkut). "But," says Zarubin, "they do not really know what they are. They call themselves Turks. But their Turkmen and Kyrgyz neighbours call them "Sart" which word they also use for Tajiks. A.D. Grobenkin regards them as Uzbeks and indeed their language is very similar and they admit to Uzbek origins".

The same author noted that the inhabitants of Zaamin and Kaisma, whom the census of the Jizzakh Uezd (District) (situated in what is now Uzbekistan) identified as either Uzbeks or Tajiks, could be classified as "Sarts". He added that P.S. Skvartskii (of the 1913 Imperial Russian Land Commission) had called them Tajiks, albeit Turkicised, but added: "On the other hand they themselves deny this saying 'we are Sarts' ".

Again according to Zarubin, Andreev noted on 25 May 1915 that "Uzbek-Sarts" live in all the settlements (in Bishkent) together with a small admixture of an assimilated Tajik element, who have lost the conception of any sort of division into tribes and no longer remember their ethnic origins, but who do not consider themselves to be "Turks". As they told Andreev "we call ourselves Uzbeks but  the Tajiks call us Sarts".

In this mist of uncertainty, by the end of the 19th century the generally accepted meaning of the word "Sart" had, as Bartol'd noted, come to mean "an Uzbekized urban Tajik".

After the revolution, the term's lack of any clear national or ethnic label or of any glorious historic association, did not suit the new Soviet regime, which was keen to create clearer national divisions in the region. With regard to the name "Sart", the Russian scholar A.A. Semenov had noted that, "because of a misunderstanding . . . the locals sometimes did not see its origin as part of the glorious past of a great country". In his recent article on the "Archaeology of Uzbek Identity", Alisher Ilkhamov also suggests that the traditional image of the "Sarts" did not accord with the social engineering embarked on by the new communist regime. The mercantile, relatively prosperous urban background of the "Sarts" was harder to reconcile with the new progressive ideals than the poorer rural Uzbek tribal types. Contemporary observers had already noticed the change. Writing in 1925 about the population of Samarkand, I.I Zarubin attributed the small number of "Sarts" registered following the 1924 National Territorial Delimitation (see below) to the "lack of clarity as to the term's meaning". By then this designation was rapidly going out of use and had largely been replaced by the word "Uzbek", a process described in detail in Obiya Chika's essay "When Faizulla Khojaev decided to be an Uzbek".

Commenting separately on the ethnic make-up of Central Asia, A.A. Semenov noted that "the largest group are the Uzbeks, assuming that the mixed Iranian/Turkic urban and village population which in past times used to be called "Sart" and speaks a local Turkic dialect (with a significant admixture of Persian words), is counted with them".

On either side of this composite "Sart" nationality were as yet unassimilated representatives of both Turkic and Tajik groups. On the Turkic side were Kipchaks, "Turks", Kyrgyz and Kara Kyrgyz (up to the 20th century the Russian names for Kazakhs and Kyrgyz respectively), Karakalpaks, Turkmen, Uzbeks etc, many of whom still lived nomadic or semi-nomadic lives. On the Tajik side, stood socalled Tajiks and Ghalcha (see below), most of whom either lived in towns or were engaged in agriculture, often at a bare subsistence level in the mountains.

It also has to be remembered that, until the Russian conquest stabilized the Pamirs, the pattern of administration in the mountainous

area of what became known as "Eastern Bukhara" was very fluid. Although these mountains were nominally part of the Bukharan Emirate, the Amir's authority was at best intermittent. Local khans, on either side of the Oxus, were usually in charge and, as often as  not, in a state of war with one another. In the middle of the 19th century, Shahr e Sabz, for example, under its Keniges Uzbek rulers, united with Gissar to throw off the yoke of the Mangit (Uzbek) Amir of Bukhara. Although both Amirs Nasrullah and Muzaffar repeatedly tried to re-conquer the town, they were unsuccessful. The English explorer Alexander "Bukhara" Burnes describes the town as independent of either Bukhara or Kunduz. Mir Muhammad Murad Beg, Khan of Kunduz in the first half of the 19th century, extended his rule north of the Oxus by defeating Mir Yar Beg of Badakhshan, and remained in at least informal control there until the 1840s. In the 1880s the Amir of Afghanistan Abdulrahman Khan sent his army north across the Amu Darya to occupy the Pamirs. Bukhara, Afghanistan, and Kunduz were not the only claimants in the Western Pamirs. The Khan of Kokand also intermittently exercised some control there, while the Eastern Pamirs were more or less continuously under his at least nominal rule. These changes in ruler, and the economic chaos they engendered, frequently prompted part of the local population to move elsewhere.

The two attacks on Shughnan and the Wakhan (1883 and 1889) by the Amir Abdulrahman Khan, and likewise the union of Karategin and Darvaz with the Bukharan emirate, sparked considerable migration of Tajiks from the mountains into the Ferghana valley. The harsh rule of Muhammad Murad Beg of Kunduz also caused intense impoverishment amongst the Tajiks of the region many of whom preferred to move elsewhere.

If one feature was common to all these khans, big and small, it was a total lack of interest in the welfare of their subjects or, as a corollary, in the ethnic composition of their population. As we shall see below, the only attempts to gauge these matters were undertaken by visiting Russians - usually military officers - whose research was neither exhaustive nor reliable. In the late 19th century, Staff-Captain Stetkevich attempted a rough count of the mountain population of the Karatagh Darya valley of Eastern Bukhara. According to his estimate, in the Karatagh and Sary-Jui "Amlyakdarstvos" (districts) the Tajiks predominated; in Regar and Sary-Asiya, the Uzbeks. Not only were these very rough estimates. It is also hard to check them against other statistics, because different census-takers chose different combinations of districts. For example, the National Territorial Delimitation's Commission of 1924, drawing on the Bukharan census of 1913, but subtracting a good number because of the assumed losses during the revolution and civil war, concluded that Karatagh/Sary Assiya was more than 66% Uzbek. But what about Regar and Sary-Jui?

If Stetkevich and other Russians had a reasonably clear idea in their minds as to the difference - at least in the mountains - between Tajiks and others, the Tajiks themselves were not always so sure. Nor were they a homogeneous community. As S.N. Andreev discovered when he visited the region as late as the 1920s, the local Tajiks at that time still had rather unclear ideas of their own identity. When asked to describe themselves, the (Eastern-Iranian-language-speaking) inhabitants of the reaches of the Upper Pyandzh (now in the Mountainous Badakhshan district of Eastern Tajikistan) described themselves as "Tajiks" and the Persian speakers as "Farsigu" (Persian speakers). On the other hand, the Yaghnobis of the upper Zeravshan described themselves as Yaghnobis and their Persian-speaking neighbours, like the Falgartsy, as "Tajiks". For their part, the Falgartsy described themselves as Tajiks and the Yaghnobis as "Ghalcha", a word widely used in the region (including incidentally by the Chinese imperial administration in Altishahr, probably to indicate an Ismaili mountain Tajik from Badakhshan) for the Tajiks of the mountains. In short, Persian-speakers who used the word "Tajik" to describe themselves were generally the mountaindwellers of areas like the Zarafshan mountains, Karategin, Darvaz and the hilly parts of Khojand.  The Persian-speakers of the cities in the plains, such as Samarkand and Bukhara, were more inclined to identify themselves in regional terms, e.g. as Samarkandis or Bukhara'is.

In his study, "The Ethnography of the Tajiks", Andreev emphasised the difference, both in language and physical appearance, between the Tajiks of the mountains (including Ghalcha) and those who had succeeded in keeping their position in the plains. In the second group he included both the inhabitants of the cities of Samarkand and Bukhara, which remain to this day largely Tajik-speaking, and of Khojand, which was not part of the Tajik ASSR under the 1924 dispensation, but was to be transferred to the Tajik SSR in 1929.

The physical appearance of the former struck him as remarkably "European", while the plainsmen were obviously of much more mixed racial origins. Lt Colonel Snesyareff of the Russian General Staff, in his Description of Eastern Bukhara dated 1906, also noticed the physical difference between the Tajiks of Karategin and Darvaz (i.e. the mountains) and those of the plains. The former looked "European" while the latter were mixed with Turkic stock. He also noticed a difference in the way of life as between Tajiks and Uzbeks, with the former being mainly settled and engaged in agriculture, while the latter were still at least semi-nomadic and engaged in stock-raising.

Like most outside observers of foreign peoples at the time, Snesyareff was not shy about subscribing to national stereotypes. In his summary,  the Tajiks were "sympathetic, kind and patient" but "deceitful and miserly". As for the Uzbeks, they were "horribly lazy, lovers of freedom, proud, hospitable, kind, thriftless, simple and straight". He concluded that, while the Uzbeks were more political than the Tajiks, they had not yet developed an overall national identity. As we shall see, Soviet rule was to change this, often to the continuing disadvantage of the Tajiks.

The conclusions of the Soviet sociologist O.A.Sukhareva, who was researching in Bukhara city as late as the 1950s, pointed to two propositions: first that, as in many large and internationally famous cities, which for centuries have acted as a pole of attraction for the inhabitants of the surrounding countryside and beyond, the urban population was very mixed. Over the centuries, the main obstacle to intermarriage had been religious rather than national or racial. Tajiks and Uzbeks, and most other Turkic peoples, were Sunni Muslims for whom intermarriage presented little or no problem. On the other hand, neither group would normally have given their daughters to (Persian-speaking) Shi'i or Jewish families, of which there were also large numbers in the city. Sukhareva's second conclusion was that many Tajiks, even as late as the 1950s, still had a weak feeling of national identity. Numerous Tajik-speakers still identified themselves as Uzbeks, often citing long family traditions and specifying from which Turkic tribe their forebears had originated. She further found  that, even within families, some members would call themselves Uzbek and others, often the female members, Tajik. She wondered whether this division might be due to a masculine inclination to identify with the prestigious military caste in the former Bukharan emirate, which was mainly drawn from families of nomadic Uzbek background. She was probably aware of Tajik claims that the Uzbeks had, during Soviet-sponsored censuses, forced Tajik families to register as Uzbeks. To encourage Tajiks to identify themselves, she therefore stressed that hers was an academic survey which had nothing to do with the government. There must be some doubt as to whether these Bukharans would have believed that a Russian conducting such an enquiry could have been totally independent of the official and Communist Party world. Nonetheless, her conclusions about the Tajiks' lack of clear national identity to some extent chimed with the earlier findings of Dr M. Gabrielyan of the Institute for Tropical Diseases in Bukhara.

Working in the city from 1928 to 1930 in the districts of Khauz e Nau and Akhtachi, Gabrielyan found that, in the former, nearly three quarters of the population described themselves as Uzbeks and, in the latter, over 90% as Tajiks; this despite the fact that they all spoke Tajik for preference. Moreover, she could When Sukhareva asked those Tajik speakers who identified themselves as Uzbeks why their mother tongue was Tajik the reply was often that their forebears had come to Bukhara, a Tajik-speaking city, and had, over the generations, adopted the language of their environment. Asked further whom they considered to be Tajiks, such respondents would often either reply "the inhabitants of Tajikistan" or the "Fars" - Persian speaking Shi'i immigrants from Iran or Merv. Those "Fars" whom she interviewed did indeed often identify themselves as Tajik. Other Tajik-speakers who identified themelves as Tajiks (rather than "Fars"), often told her that they were descended from immigrants from the countryside, especially the mountains. Her conclusion, therefore, seems to have been that Muslim Tajik-speakers often thought they were Uzbeks, but, when asked to describe Tajiks, thought they were either immigrant mountaineers, or Shi'i "Fars".

As for the Tajik-speakers still in the mountains, although they seem to have been generally more clearly aware of their Tajik identity, in isolated cases confusion reigned here also. Sukhareva quotes the research of B.Kh Karmysheva who, working in the hilly Urgut district above Samarkand, found several instances of Tajik-speaking villages which associated themselves with Uzbek tribes - and were accepted as such by their Uzbek neighbours. If these findings from the second half of the 20th century show a good deal of vagueness as to national identity still surviving in Bukhara city amongst the Tajik-speaking population, these feelings will have been even vaguer and weaker before the advent of the Soviet regime. As one old man put it to Sukhareva "before 1926 no one ever asked us whether we were Tajiks or Uzbeks". The findings also seem to show that there is at least an element of truth in the claims of Uzbek nationalists that some Tajik-speakers are in fact of Turkic origin but have for various reasons adopted the Persian language. The pre-Soviet "Sart" identity had after all been bi-lingual Uzbek and Tajik. This is not to say that the efforts of Soviet censuses of Central Asia to divide the population into ethnic groups were always free of political manipulation (in this case by the Uzbeks). On the contrary, the evidence for manipulation, indeed falsification, is strong. But, for many of the inhabitants of Bukhara city, language was not and continued not to be a clear indicator of national identity. Irrespective of people's language or their genetic origins, they reached conclusions as to their own identities for a range of different reasons. Sukhareva and Gabrilyan's research suggests that there were Tajik-speakers who genuinely believed themselves to be Uzbeks. And who was there to tell them they weren't? However, as we may already have seen in Khojibaeva's account above, it was not necessarily the rank-and-file population who were to be the main players in determining Tajik identity but rather the intelligentsia - and here influences were at work which were to complicate still further not only the identity of individuals but the path to nationhood.