I will use the example of the analysis of the Bukhara censuses in 1926 provided by Bergne and Sengupta to illustrate how unclear the notions of ethnicity and identity were, and furthermore, how fluid and flexibly changed they were. I will further show how language use was confused and how non homogenous communities were forced to be claimed as such. Referring to Sengupta’s ‘fluid identities’, it must be stressed that multiple identities were the predominant characteristic of the Central Asian region (cf. Sengupta 2002).
Bukhara and Samarkand are two cities disputed between Uzbekistan and Tajikistan. Both cities are within Uzbek borders, and both of them are primarily Tajik-speaking. There are a number of other settlements within Uzbek territories, such as villages and towns, where Tajik inhabitants predominate. It is not difficult to find, especially in the mountainous area of Nurata, villages with only Tajik speaking inhabitants. The same is true to the oblasts (districts) of Syrdarya, Sukhandarya, Kashkadarya, Bukhara, Samarkand, Dzhizak, and the Fergana valley. But because these two cities, Samarkand and Bukhara, are among the most ancient and famous cities in the region (indeed, they are the second and third largest cities of Uzbekistan) and played an important role as centres of science, culture and religion at the Silk Road there is a bigger sense of loss on the part of Tajikistanis, and ‘defensiveness’ on the part of Uzbekistanis. These two are currently, together with Khiva, the biggest tourist destinations in the whole of Central Asia.
Both Sengupta and Bergne note that there were no accurate and reliable figures on the population of Bukhara and moreover of its ethnic composition before the late 1920s (cf. Sengupta 2002, Bergne 2007). According to Sengupta, the only source of information regarding the overall population of Bukhara was in form of travel accounts, where only approximate numbers ranging from 2 to 3.5 millions were mentioned. A survey after the October revolution revealed a number of 1,531,015 (cf. Sengupta 2002: 62).
The discrepancy between both, censuses as well as travel and ethnographers’ accounts is evident. Huge differences were observed by Sengupta between the outcomes of censuses in 1920 and in 1926. Such as in the 1920 census of the population of Turkestan the following ethnic groups were identified: 1) Kyrgyz 2) Sarte-i-Tajik 3) Turkmen 4) Russian. Obviously, no separate Uzbek ethnic group existed (ibid. 76). Whereas the censuses of 1926 in Bukhara gave different numbers, namely, out of 41,839 people, 27,823 called themselves Uzbeks and only 8646 called themselves Tajiks (cf. Sengupta 2002: 68). It is interesting that a non-existent ethnic group in the 1920s became the majority in the course of only six years. Important to note that every time numbers in Bukhara are mentioned both Sengupta and Bergne speak of a so called ‘misidentification’. Sengupta writes: “[F]rom different parts of the region came reports that the people were innocently misidentifying themselves and it would be difficult to assign ethnic distinctions on a clear basis” (ibid. 76). She further suggests “cases of deliberate misidentification for political reasons, particularly in the Tajik-Uzbek case” took place (ibid.). She also concludes that often “estimates were determined by political exigencies” (ibid.).
Both Sengupta and Bergne refer primarily to the field research conducted by a Russian ethnographer Sukharyova in Bukhara in the 1940s and 1950s. According to Sengupta, she also recorded the language in which the interviews were conducted. This made some most exciting findings possible: “While the majority of the population spoke Tajik, they identified themselves as Uzbeks” (ibid.). The works of O. A. Sukharyova, in the words of Sengupta illustrate that “delineating the population of the city as Uzbek or Tajik would be problematic given the fact that in most cases there was no congruence between language use and ethnic identity” (ibid. 66).
Sengupta argues that “incorrect identification of the ‘ethnic’ criteria” resulted in a number of occasions when thousands of Tajiks “declared themselves as Uzbeks in their identity papers and were therefore recorded as such in all demographic records” (ibid. 58). Bergne, too, points on this strange event. Moreover, many ethnographers seem to agree and repeat sentences like ‘a number of Tajiks declared/think of themselves as Uzbeks’. It is interesting that ethnographers already call them Tajiks, even if respondents identify themselves as Uzbeks. This sentence as such is strange. It shows that researchers are either biased or seem to know beforehand that the respondent is Tajik but calls himself Uzbek. Ethnographers obviously assumed respondents to be Tajik because they were Tajik-speaking. Sukharyova refers to ethnographic findings of other Russian scholars (such as P. Savalev and Khanikov) who agreed that the majority of the city’s population was Tajik but she admits that these “accounts define ethnic affinity on the basis of language” (ibid. 67).