The question remains what was the role of language in the whole delimitation process? It was recorded (by ethnographers, travellers, and officials) in what languages people spoke, widely either Turkic or Iranian (Persian or Eastern Iranian). For the Soviet bureaucracy language was an important signifier in determining ethnicities and drawing boundaries to separate them, Sengupta suggests (cf. Sengupta 2002). But how is this possible in a region where people live intermingled with each other and more than often are multilingual. The Central Asian territory witnessed many feudal states and empires where indigenous population spoke one language, the ruling elite spoke another, and the intelligentsia used a third language to exchange their thoughts. Namely in the 19th century, Persian was the administrative language in courts and among the urban folk, Turkish was the language of the Uzbek elite, who were multilingual, and the intelligentsia wrote in Persian and Arabic and later also in Turkish.
Regarding the role of language, many authors agree that the lingua franca in the region before the Soviet invasion for centuries was Persian and (academically) Arabic. Sengupta’s findings on language use, namely Persian and Turkic in the Bukhara emirate (18-20th century) are worth to mention here. It is however important to make it clear that the ruling elite in Bukhara consisted of Turkic speaking Uzbeks whereas the indigenous population was largely Persian speaking which was also the case in previous empires (of the Arabs, Mongols, Timurids, and Sheybanis):
“Tajiki was the official language of the Emirate. As the language of administration, it was spoken at the court by the Emir, and his mostly Iranian officials. All foreign correspondence was in Iranian as were all official decrees to the citizens. At the same time being a Turk, as the leader of the Turkish chieftains and tribesmen, the Emir addressed his chiefs of Ils (tribes) and Ulusses (appanages) in Turkish” (Sengupta 2002: 49).
Again, language in a multilingual region is not the indicator of ethnicity and can’t be used to delineate groups along linguistic or ethnic lines. Many scholars speak of so called Turkified Iranians and Persified Turks, the evident example of the latter is the Persian speaking Hazara people in Afghanistan, who are believed to be Persified Mongols i.e. having somewhat Mongol physical features they speak Dari/Persian.
Since this section is about Uzbeks and Tajiks the reader might ask about the differences between them. Uzbeks are one of the most, to my opinion, Persified (at least linguistically, but also culturally) Turk people in the region, who preserved their language, but with an enormous influence of Persian language and settled down earlier than other nomads. As Sengupta writes about increasing influence of the Persian language on Central Asian Turkish languages: “Whereas Central Asian Karakhanid (11th century) contained 1.6 percent of Persian loanwords in its texts, early Chaghatay (14th century) contained 26 percent and classical Chaghatay (15th century) 50-60 percent” (Sengupta 2002: 91).
According to Bergne’s findings, Andreev (in ‘The Ethnography of Tajiks’ 1925) studied physical and linguistic differences of Tajiks in the mountain areas and in plains. He and several other Russian military officers noted that Tajiks in the mountains looked more ‘European’ than the ones “who had succeeded in keeping their position in the plains” (Bergne 2007: 11). Further Bergne adds that Lt. Colonel Snesyarref of the Russsian General Staff in his description of Eastern Bukhara (1906) describes Tajik plainsmen as being “mixed with Turkic stock”, the latter remained unexplained and undescribed by Bergne. When it comes to differences in the way of life, Bergne notes that Tajiks are “mainly settled and engaged in agriculture” whereas Uzbeks “were still at least semi-nomadic and engaged in stock-raising” (ibid.).
What is understood as the Uzbek nation nowadays was constructed by the Soviets, according to Sengupta. “The designation Uzbek, for instance, is being used in the Soviet sense to mean nation, whereas previously it had been used to mean a tribal classification of a dominant dynastic tribal tier, the Shybanids. Similarly, prior to 1924 there was no single Uzbek language that was prevalent in the region” (Sengupta 2002: 104). Just as we have seen with the term Tajik, the ‘so called’ Tajiks themselves had various other criteria as whom to consider Tajik, on the basis of religion, or origin (location-wise). This again shows how unclear the terms were at those years of delimitation and are indeed still. But still is language a determinant of ethnic affinity?
 One of the examples of Persianisation of Uzbeks is elopement or bride-kidnapping which was popular among nomad Kyrgyz, Kazakh and Turkmen, and also among Tajik-speaking Roma of Central Asia. Kyrgyz and Kazakh believe it to be their cultural heritage and tradition, whereas among the sedentary population elopement is regarded as the worst offence and socially unacceptable. This attitude to elopement is largely shared by settled urban Uzbeks too. This assumption of mine doesn’t have any scientific proof whatsoever, but obviously this custom could be an interesting subject for future research.
 Chaghatay is one of the Turkic languages, claimed by Uzbeks to be ‘early Uzbek’. Sengupta reveals that “historically Turki was not a single language but a combination of dialects” (ibid. 90). Some pan-Turkist leaders, such as Gaspirali, strived to create one common Turki language, which could serve as a lingua franca among Turkic Moslems of the Russian Empire, but apparently he failed (cf. Sengupta 2002: ibid.).