Sart and Tajik

The romantic idea of the nation wreaked havoc on older notions of community and identity. Armed with an understanding of the world that saw it divided into discrete groups, amenable to rigorous, "scientific"' classification if only sufficient "objective" data could be obtained, Russian officials and scholars proceeded to find the objective reality behind every label they encountered in their new domains. The enumeration and classification of the population that ensued created new understandings of old labels. The complexities of Central Asian identity were nowhere better demonstrated than in the case of the "Sarts." The career of this label in the half-century of Russian rule demonstrates the forces at work in shaping identities in Central Asia.

For reasons that remain unclear, "Sart" became the term most commonly used by the Russians to denote the sedentary population of Central Asia after the conquest. It was used in several different ways. "In common parlance and every day life," a German geographer wrote in 1914, "the Russians use 'Sart' in much the same way as British colonists would speak of 'niggers.' It is applied to all and sundry 'natives' whose dress does not single them out at once (Jews, Turkmen, Kirghiz) or who are not evidently foreigners (Europeans, Afghans, Chinese, Hindu, etc.)."^[38] Officials and scholars sought to use the term in a more precise manner, to apply it only to the "real" Sarts. The precise demarcation of the community united behind the label remained in question, but officials and scholars never doubted that the acquisition of sufficient objective information would provide the answer. The Sarts existed as an organic entity; the problem was to define them precisely. The answers could be sought in the realms of science or history, but not social practice, for how the people defined themselves was of very little importance to the concerns of "science." For physical anthropologists, craniological measurements provided a key to the Truth that was often clouded by social

[38] W. Rickmer Rickmers, The Duab of Turkestan: A Physiographic Sketch and Account of Some Travels (Cambridge, 1913), 5

conventions of naming. Similarly, although orientalists exhaustively examined the etymology of the term and its occurrences in historical texts, they did not deign to look at how the term was used in actual practice (but then that has never been the concern of orientalism).

The earliest Russian observers often saw no difference between the terms "Sart" and "Tajik." Iu. D. Iuzhakov, arriving in Central Asia with the armies of conquest, reported that the two terms were synonymous and both referred to the sedentary population of the region. His informants, most likely Tatar or Qazaq interpreters, had apparently told him that this population had descended from Jews and Iranians, an explanation he found convincing. Iuzhakov felt he knew the natives well enough to report, "In their terrible greed for money and their thievery, they exceed even the Jews. In their manners, the tone of their conversations, their cowardice, in the pettiness of their interests, and in the complete absence of political tact, they are, precisely, Jews." The fact that some Sarts spoke "their own language ... a mixture of Turkic and Persian in which the Turkic element strongly predominates," whereas others spoke a variety of Persian, was not of sufficient importance for him to override the common genetic origins as a marker of identity.^[39] The stereotypes Iuzhakov used continued to be invoked down to the end of the old regime, but the linguistic distinction, so unimportant to him, soon emerged as all-important, and "Sart" came to be applied in Russian bureaucratic practice exclusively to the Turkic-speaking parts of the sedentary population of Central Asia, while "Tajik" was reserved for those of Iranian speech, the widespread bilingualism ignored for being too cumbersome.

Such a definition of "Sart," which distinguished Sarts from Tajiks on the one hand and other Turkic-speaking Central Asians on the other, proved difficult to establish in practice. Science came to the rescue. A. Bogdanov used craniological data to argue that Sarts and Ozbeks were distinct peoples.^[40] The anthropologist N.A. Aristov suggested a narrower definition of "Sart" to rescue the term from popular misuse. Real Sarts were, for Aristov, "sedentary Turks and Turkicized natives who have already lost their tribal way of life and the tribal divisions connected with it," and the term should only be applied to them.^[41] The notion of

[39] Iu. D. Iuzhakov, "Sarty ili Tadzhiki, glavnoe osedloe naselenie Turkestanskoi oblasti," Otechestvennye zapiski 173 (1867): 398-400.

[40] A. Bogdanov, "Antropometricheskie zametki otnositel'no turkestanskikh morodtsev," Izvestiia Obshchestva liubitelei estestvoznaniia, antropologii i etnografii , 1888, no. 5, 85-87.

[41] N.A. Aristov, "Zametki ob etnicheskom sostave tiurkskikh plemen 1 svedeniia ob ikh chislennosti," Zhivaia starina 6 (1896): 429. Precisely this notion has been resurrected in the only post-Soviet discussion of the Sarts: O.M. Bronnikova, "Sarty v etnicheskoi is-torn Srednei Azii (k postanovke problemy)," in Etnosy i etnicheskie protsessy: pamiati R. F. Itsa (Moscow, 1993), 151-158.

"Turkicization" also evoked racial admixture (already foreshadowed in Iuzhakov), which proved compelling to the romantic imagination and soon became a characteristic trait of the Sarts.^[42] Orientalists, on the other hand, sought the answer in philology. Perhaps the most influential view was formulated by Bartol'd. For Bartol'd, "Sart" was an old Turkic term, of Sanskrit origin, meaning "merchant," which in the post-Mongol period came to be used as a synonym for "Tajik" in referring to bearers of the Persian Muslim culture of the towns, in opposition to the nomadic Turkic culture of the steppe. The distinction between Turk and Tajik was of little interest to the Ozbek conquerors of Central Asia, and after the sixteenth century, "Sart" distinguished the sedentary population of the conquered territory from the conquerors and their allies. Gradually, "under the influence of the conquerors," the town-dwellers began to call themselves Sart, "but the tribal differences between Turks and Tajiks were so great that the representatives of both peoples could not call themselves by the same name. Since the majority of the settled population now spoke Turkic, urban Turks began to be called 'Sarts,' in contradistinction to not just the nomads, but also the Tajiks."^[43]

This approach was in many ways typical of the orientalist enterprise. The etymology of the term "Sart" was the key to the business of understanding who the Sarts were. Similarly, if the term appeared in historical sources, then Sarts must exist as "a people," and today's Sarts must have something to do with the Sarts mentioned in those sources. Much of Bartol'd's evidence comes from a few scattered references in historical sources, all produced at court and usually referring only to court elites, which he sees as proof of his fundamental assumption that stable labels refer to stable communities, which retain their organic unity through the ages.^[44]

[42] This explanation especially found wide acceptance m more popular works; see, e.g., A. Kruber et al., eds., Aziatskaia Rossua: illiustrirovannyi geograficheskii sbornik 3rd ed. (Moscow, 1910), 189; or V. I. Masal'ski,, Turkestansku krai (St. Petersburg, 1913), 392-393.

[43] V. V. Bartol'd, "Eshche o slove 'Sart'" (1895), in Sochineniia , 9 vols. (Moscow, 1963-1977), II/2: 310-314; see also Barthold, "Sart," 175-176.

[44] This is also the approach taken by Yuri Bregel ("The Sarts in the Khanate of Khiva," Journal of Asian and African History 12 [1978]: 120-151), who on the basis of literary references to a division of notables among "Sarts" and 'Uzbeks" in Khiva concludes that Sarts "were definitely considered by the Uzbeks as a different ethnic group, a different people." Bregel's insistence that the Sarts were distinguished by a specific political position defined "by the role of their [sic] leaders in the government" makes it seem as if the Sart notables who laid claim to certain positions at court actually represented other Sarts m the affairs of government. This betrays a misunderstanding of the nature of both power and community in premodern Central Asia, for there is no reason to believe that Sart notables felt any affinity for, or were capable of mobilizing, peasant Sarts in the rest of the country.

One person not daunted by the complexity of the problem was Ostroumov. In a work titled Sarts: Ethnographic Sketches , he reviewed the "scientific" literature on the question for forty pages, but then concluded simply that "Sarts are the sedentary natives, predominantly of the Syr Darya and part of the Ferghana oblasts."^[45] (Ostroumov clearly never paused to wonder at the neat coincidence of ethnic distribution with recently created administrative boundaries.) What Ostroumov thought was important, however, because his control of the TWG and his stature as an orientalist allowed him to elaborate a Sart language, distinct from Ozbek and other Turkic dialects, as a literary language. Russian orientalism knew the Turkic speech of the sedentary population of Transoxiana as "Sart," and in addition to Ostroumov's exertions at the helm of TWG , grammars and dictionaries of the Sart language made their appearance.

"Sart" also appeared as a category in the all-Russian census of 1897. The census made a brave attempt to reduce the empire's ethnic complexity to the simplicity of numbers. Although the census did not have a category for "nationality" as such, it did classify people according to native language, which was believed to be the primary attribute of a nation. The purpose was largely defeated in Central Asia. The census counted Ozbeks and Sarts separately but left a large part of the population classified simply as "Turkic" (tiurkskii ) (see Table 7). This could have been an attempt to distinguish between "Ozbek" and "Turk"—the census does not make this clear—but since the same classification was used in other regions of the empire for very different groups of Turkic speakers, it confounded not only local statistics but also those at the all-Russian level.^[46] (There were other instances of less than consistent usage: In some tables, Tatars were counted separately, but in others they appeared only as speakers of "Turko-Tatar languages." In any case, "Tatar" covered the Turkic languages of the Volga, Urals, Crimea, and

[45] N. P. Ostroumov, Sarty. Etnograficheskie materialy (obshchu ocherk ), 3rd ed. (Tashkent, 1908), 3.

[46] See comments by Guido Hausmann m Henning Bauer et. al., Die Nationalitaten des Russischen Reiches in der Volkzalung yon 1897 , 2 vols. (Stuttgart, 1991), 1: 244-245.

Transcaucasia.) Nevertheless, enumeration produced new understandings of community. To say that in 1897 in the three core oblasts of Turkestan there were 951,337 Sarts who were only Sarts and nothing else transformed the meaning of the term by abstracting it from the contours of local relations and oppositions.

It is difficult to judge whether the official use had any resonance among the population itself. To be sure, official favor for the term "Sart" led to its use, especially in bureaucratic contexts, and the TWG helped popularize the new understanding of the term. This usage was also accepted by Tatar writers, who found the notion of racial admixture as the explanation for the origins of the group quite compelling. The Jadids of Central Asia, however, were resolutely opposed to the label.

Criticism of the official use of the term came as early as 1893, when concern that Russian functionaries should learn local languages led to the creation of language courses for them. One of the languages scheduled to be taught was Sart. Sher Ali Lapin, whom we will encounter again in Chapter 8, an interpreter in the chancellery of the governor of Samarqand oblast and a Qazaq himself, argued in a lecture that "there is neither a Sart people, nor a Sart language." Rather, the word was a contraction of sarï' it , "yellow dog," a derogatory appellation used by Qazaq and Qïrghïz nomads for all sedentary people, regardless of ori-

gin. "We have no basis for calling the language of the Sarts 'Sart,' since the language of the Sarts includes both Tajik and the language of the sedentary Ozbeks; therefore the language should be called . . . the Ozbek language, in the dialect of the sedentary Ozbeks."^[47]

Lapin was taken to task by Bartol'd, then working on his doctorate in Tashkent, who was surprised that "Mr. Lapin decided to speak with such aplomb about things with which he apparently has not the slightest acquaintance." Rejecting the explanation of the origin of the term put forward by Lapin as mere folk etymology, Bartol'd fixed the explanation in the proper realm of high culture, citing references to the term in the literary and historical sources of the post-Mongol period.^[48] When Lapin had the temerity to reply in print, Bartol'd heaped further condescension on him. Writing in Turkestanskie vedomosti , where Lapin had responded, Bartol'd presented a long list of faulty citations and misquotations committed by Lapin, before concluding: "In printing Mr. Lapin's article, the editors proceeded from the opinion that the inclination to scientific work on the part of a native in any case represents a gratifying phenomenon that must be supported by every means. No one disputes this; but no one expects native writers to attain at once the scientific standards established by European science as a result of long experience."^[49] Having crushed the native beneath the weight of the long experience of European science, Bartol'd went on to indulge in orientalism's fetish with literary etymologies, elaborating in the process his theory of the origins of Sart. No more was heard from the native side for a decade and a half.

When the issue arose next, it was taken up with the Tatars, who used the term routinely. In 1911, Behram-bek Dawlatshaev, the highest-ranking interpreter in Bukharan service, broached the topic in Shura , perhaps the most respected magazine in the Tatar world. "Are we Sarts or Turks?" he asked the editors in the regular question-and-answer section of the magazine. Turkestan means "the land of the Turks," he asked, "so why is it that we are called 'Sarts'? Is it that in earlier times Turks lived in this 'land of the Turks,' but later left it, leaving their name behind? If so, then where did the people called 'Sart,' that is, us, come from and when? . . . And how is it that we inherited Turkic literature? Did the Turks

[47] Quoted in Bartol'd, "O prepodavanii tuzemnykh narechii v Samarkande" (1894), m his Socbineiia? ?, II/2: 303-304.

[48] Ibid.

[49] Bartol'd, "Vmesto otveta g-nu Lapinu" (1894), in Sochineiia , II/2: 508-309.

leave it to us? Or did we take it, and the land, from them by force?"^[50] And so forth. No Tatar writer responded, but Behbudi joined in with a lengthy article in which he argued that the origins of the word "Sart" were unknown and that it was used pejoratively only by the northern neighbors of Central Asia (Qazaqs and Tatars, from whom the Russians took it). "Upon asking Qazaqs 'Whom do you call a "Sart,'" I usually received the answer, 'Those who travel around our steppe' (meaning all traders)." The term was not used by those labeled "Sart" themselves, Behbudi further argued: "Those who have no interaction with the Russians or are unfamiliar with the press think it is the Russian word for 'Muslim.'" Finally, Behbudi noted, there were ninety-two tribes in Turkestan, but none was called Sart.^[51]

Behbudi was backed by his friend Baqa Khoja, who in a long article denied the existence of a Sart people. "The inhabitants of Turkestan, that is, Turan and Transoxiana, are, from the point of view of race and nationality [jinsiyat wa qawmiyat ], predominantly Turks and Tajiks." The opposition of Turk and Tajik had become a metaphor among "oriental poets," but neither old Arabic, nor Persian histories, geographies, or dictionaries contained the word "Sart." Quoting Russian authors in the original, he went on to show the many, often contradictory, explanations given for the word. "To call the Ozbek Turkic inhabitants of the five oblasts of Russian Turkestan and the khanates of Bukhara and Khiva 'Sart' is an injustice, the despotism of opinion, the cause of doubt and division; [in short,] a huge mistake."^[52]

The same conclusions were reached by Dawlatshaev when he himself answered the question he had raised. The matter was simple: "We Turkestanis are Turkic Ozbeks who belong to more than one hundred tribes [awmaq ] of the Mongol people [qawm ] and the Turkic race [urugh ]." The proof was simple and lay in the Turkic speech and literature of the region, as well as in its ruling dynasties. The 92 tribes "renowned from olden times to the present" had increased to over zoo, he argued, and he appended a list of 111 tribal names then current. "Now, if foisting the name 'Sart' on [the population of] Turkestan, composed of 'more than a hundred Turk-Mongol tribes,' whose history, literature, language, and

[50] Sbura , 15 August 1911, 504.

[51] Mahmud Khoja bin Behbud Khoja, "Sart söze majhuldur," Shura , 1 October 1911, 581-582.

[52] Samarqandi Baqa Khoja bin Sayyid Hadi Khoja, 'Sart söze asïlsizdïr, Shura , 15 December 1911, 754-757.

customs are Turkic, is not an injustice, what is?"^[53] The debate flared up again, but this time in Behbudi's Ayina , in early 1914, in response to an article by the young Bashkir historian Ahmed Zeki Velidi in which he spoke of "Sarts." Behram-bek Dawlatshaev again expressed his displeasure with "those writers who, not knowing that we Turkestanis belong to over a hundred Turkic tribes, call us 'Sart,' as well as those who, knowing full well that we are Turks, call us 'Sart' by way of insult."^[54] Two weeks later, Ayina published an open letter with seven signatures expressing displeasure over the use of the term "Sart" when "everyone knows that the population of Turkestan is composed of Ozbek, that is Turkic, Tajik, that is, Persian [Fars khalqi ], and Arab (Khoja) groups."^[55] Behbudi took this opportunity to republish the 1911 articles from Shura in modified form.^[56]

Beyond such "ethnographic" debates, "Ozbek" also appeared occasionally in Jadid literature as synonymous with the nation. In a poem published in 1916, for instance, Hamza used Turkistan eli ("the people of Turkestan") and Ozbek eli ("the Ozbek people") interchangeably in exhorting the nation to "not sleep in this age of progress."^[57]

Not everyone in Central Asia shared the Jadids' position. Not surprisingly, TWG took the lead in criticizing the Jadids. A student from Osh wrote in 1913 to criticize those who wanted to protest the use of the term on two counts. First, "Sart" was not a pejorative term but rather carried connotations of "royal descent" and "philosopher." Second, the author asked if a change in terminology would make the people of Turkestan stronger or more developed?^[58] Other writers argued that this search for roots was a form of nationalism and divisive of the Muslim community and that labels were not important, since "the name of the renowned and developed Nemets [German] nation comes from the word nimoi , which means 'mute.' But they respect this name and do not worry about changing it. They have not lagged behind because of this name, but are the most developed; the cause for their renown is not their name, but their good morals."^[59] An author, writing under the pseudonym "Sart,

[53] Behram-bek Dawlatshaev, "Turkistanlilar," Shura , 1 January 1913, 12-15.

[54] Behram-bek Dawlatshah, "Sart masalasi," Ayina , 19 February 1914, 300.

[55] "Tashkanddan gila=opka," Ayina , 1 March 1014, 354.

[56] "Sart sozi majhuldur!" Ayina , 22 March 1914, 314-315; 29 March 1914, 338-340; 5 Apr, 1914, 362-365; 12 April 1914, 386-388; 19 April. 1914, 478-480.

[57] Hamza, "Dardiga darmon istamas," in Tola asarlar toplami , ed. N. Karimov et al., 5 vols. (Tashkent. 1988-1989), II: 29-30.

[58] Mirza Qadirjan Qabiljanbayev, "Haqqamyat," TWG , 10 January 1913.

[59] Mulla Alim, "Po povodu pis'ma o siove 'Sart'," TWG , 20 January 1913.

son of Sart [Sart oghli sart ]," contended that "we Sarts do not hate the name, since our faith does not consider names and lineages important."^[60]

Nevertheless, the Jadids' opposition to the term "Sart" was significant for various reasons. In disowning it, the Jadids were as far from the pre-Russian usage of "Sart" as a social marker as were Russian scholars and functionaries, but whereas the latter searched for the nation hiding behind the label, the Jadids rejected it because, they argued, there was no nation there. Nations were objective identities, but their objectivity was defined by race; hence the concern with biological origins, which made Tajiks into Iranians and Khojas into Arabs. The fact that Behbudi, with his cosmopolitan tastes, moderate politics, and Muslim education, was so prominent in demanding the use of the "proper" names for the people of Central Asia indicated how far Turkism had crept in around the edges into local discourses of identity. "In our age, the 'national' [milliyat ] question has taken precedence over the question of religion among Europeans," wrote a writer who unfortunately remained anonymous, "so there is no harm if we too occasionally discuss the 'Sart' question, which is considered a national question, and thus remember our nation."^[61] Indeed, the Tatar writer Abdurrauf Muzaffar made the point, quite popular in Ottoman circles at the time, that "religion exists only on the basis of the nation and national life A religion without a nation is destroyed."^[62]

It is equally important that criticism of the use of "Sart" was directed against Turkist authors. The discourse of Turkism was polyphonic, and the debate described above was an attempt by Central Asian writers to define their own version of Turkism. More significant, especially with hindsight, is the fact that the distinction between "Ozbek" and "Turk" disappears entirely. The ninety-two tribes mentioned by Behbudi were the ninety-two tribes of the Ozbek confederation in the aftermath of the Shaybani conquest of Transoxiana and did not encompass the entire Turkic-speaking population of the region. No wonder, then, that they turn into "more than one hundred tribes" in the hands of Dawlatshaev. Ozbekness became, for the Jadids, a defining feature of the Turkic-speaking population of Central Asia. We are reminded of Wickmer's ob-

[60] Sart oghli Sart, "Otvet zhurnalu 'Aina'," TWG , 27 April, x and 4 May 1914. The pseudonym is significant, since this form was popular with Turkist authors m the Ottoman empire.

[61] "Sart sozi, ma'lum bolmadi," Ayina , 19 July 1914, 923.

[62] A. Muzaffar, "Din millat, millat milliyat ila qaimdir," ST , 26 November 1914; 2 December 1914; 10 December 1914.

servation that "Sart" was used to denote all "unmarked" Turcophone groups of Central Asia. In Jadid parlance, that meaning of "Sart" was being translated into "Ozbek." This also applied to the name of the language, which was often equated with Ozbek in this period. In the early 1920s, all the "marked" groups of Central Asia (Türkmen, Qazaq, Qïr-ghïz, and, eventually, Tajik) were carved away from Turkestan, and the remaining Turcophone population became the modern Uzbek nation. The roots of these momentous changes are to be found in local discourses before 1917. The modern nations of Central Asia were not simply the work of an imperial Soviet regime bent on dividing its subject populations, the better to conquer them; rather, their origins lie in new ways of imagining the world and Central Asia's place within it. Similarly, the abolition of the term "Sart" in the early Soviet period was not "evidence of [the] ignorance of those who governed Turkestan at that time," as Yuri Bregel snidely claims, but rather the outcome of very real politics surrounding Central Asian identity in a revolutionary age.^[63]

The insistence on the "proper" identification of peoples led to the disaggregation of the sedentary population along newly drawn ethnic lines. The most difficult disentanglement was that of the Tajiks. The longstanding dichotomy of Turk and Tajik was invested with new meaning. Now the difference was seen to reside in the realm of nature and was described in a new language, such as in this description of the Tajiks of Bukhara that appeared in Sbura : "Although the Tajiks are Iranian and their language Persian, their religion is Sunni. Their name emerged from their animosity toward the Shi'is .... Their faces are straight, their women renowned for their beauty. They are assiduous and masters of commerce, but [also] deceptive and have low morals."^[64] The low levels of morality and general effeteness of the Tajiks, taken whole cloth from contemporary European anthropology, appear quite frequently in Jadid writing. Criticisms of the maktab often carried an anti-Iranian subtext. A Jadid schoolteacher complained that although old-style maktab teachers "are Turks, [they] do not know their mother tongue and do not teach it, but rather look upon it... with hatred. Instead, they waste the poor children's time with Persian fairy tales and puzzles, whose harmfulness in the present day is quite obvious."^[65]

[63] Bregel, "The Sarts of the Khanate of Khiva," 121n.

[64] "Bukhara mamleketi," Shura , 1 February 1910, 101.

[65] Dada Mirza Qari, "Muallim wa shagirdlar," Ayina , 14 December 1913, 183-4.

This re-visioning of identity remained an exercise in exclusion , since we have little evidence of any parallel assertion of Iranian or Tajik (or Aryan) identity among the Iranophone population of Central Asia. The few instances in which romantic notions of Iranianness appeared in print were all connected with the Turkist enterprise. The Transcaucasian editor of Bukhara-yi sharif published a language tree that placed Persian in its "proper" Aryan context. In another issue we read: "There are [between] 10 million and 12 million Muslims in Turkestan, Transoxiana, Bukhara, and Khiva. Approximately 7 million of them are Ozbeks, Turkmens, and Qïrghïz Qazaq, who belong to Turanian nations [umam-i turaniya ], and are Turks; their national language is Ozbek or Chaghatay Turkic. The remaining [sic] z million are Tajiks, who belong to the Aryan nations [umam-i ariyaniya ], and are of Iranian origins; their literary language is also Persian."^[66] Beyond this, however, there was little discussion of Tajik or Iranic identity in Central Asia until well into the 1920s.