Evidentiality in Uzbek and Kazakh

1.2.3 Uzbek and Kazakh: Areal Importance 
 
Given the central position of Turkic within the Eurasian evidentiality belt, it is 
noteworthy that Uzbek and Kazakh, two important, widely-spoken languages spoken near the 
center of this belt, have not been paid much attention in discussions of evidentiality. This 
situation is not unique, however, as there exist only a few in-depth descriptions of evidentiality 
and related phenomena in the Turkic languages. Turkish, as the best-studied Turkic language, 
has been paid the most attention; in Johanson and Utas’ (2000) collective overview of Turkic and 
related languages, there are three papers on it, and many more exist elsewhere. In this same 
volume, works have been devoted to Salar (Dwyer 2000) and Gagauz (Menz 2000). Elsewhere, 
there are descriptions of evidentiality in the Mishär dialect of Tatar (Tatevosov 2007), Altay 
(Skribnik and Osnova 2007), and Cypriot Turkish, which is considerably different from 
Anatolian Turkish (Demir 2003). Aside from references in broader works covering the entire 
Turkic language family (such as Johanson 2000; 2003), there is almost nothing written on Uzbek 
or Kazakh. In order to understand the importance of these two languages within this belt, I 
summarize here the sociolinguistic situation in Central Asia, paying special attention to Uzbek-
Tajik contacts and to the expression of evidentiality in Tajik. 

---

34 
The sociolinguistic situation in Uzbekistan and Kazakhstan and, indeed, in all of Central 
Asia, is characterized by the different contact situations experienced by sedentary and nomadic 
populations. Sedentary Turkic populations of Central Asia usually co-existed with speakers of 
Persian languages, who had lived in the region before the arrival of Turkic peoples. The contact 
between these urbanized Turks and Persians resulted in the early adoption of Islam by these 
urbanized Turks (as well as the accompanying loan words needed to express Islamic religious 
and cultural concepts) and affected the languages of both populations. By the 19
th
Century, the 
term Sart came to be applied to sedentary Central Asian Turkic-speakers, while speakers of 
Persian came to be called Tajiks (Barthold 2010). Nomadic Turkic peoples of Central Asia, on 
the other hand, came to be called Kazakhs. During the Soviet era, Sart came to be seen as 
derogatory (due in part to now disproved Soviet hypotheses that Sart was derived from Turkic 
sarï it ‘yellow dog’) and the term Uzbek, which was a neutral term for urbanized Turks with no 
specific tribal associations (very often from the Khwarezm region), was applied instead 
(McChesney and Shalinsky 2010). 
The broadness of the term Uzbek is exemplified by the wide variety of Turkic language 
types spoken by peoples called Uzbek. While this work focuses solely upon standard Uzbek, 
which is based upon the dialects of Tashkent and Samarkand, there are other so-called dialects of 
Uzbek that are clearly of non-Southeastern origin. Within Uzbekistan, there are varieties of 
Turkic that belong to the Oghuz and Kipchak families, spoken mostly in the Xorazm region.
Despite their obvious differences from the Southeastern-based standard, these dialects are still 
referred to as Uzbek (Abdullaev 1961). The Uzbek dialects of Afghanistan are also sufficiently 
divergent that they may not belong to the Southeastern branch of Turkic, as they exhibit the 
Kipchak labialization of *AG to Aw and retain the voiced velar in *IG, as opposed to the 

---

35 
Southeastern fortition of *IG to *IK (Boeschoten 1983). In all of these cases, the peoples 
speaking these varieties of Turkic are sedentary, so the term Uzbek has been applied to these 
peoples and their languages, regardless of linguistic affiliation. 
As explained in the previous section, Uzbek and Kazakh belong to separate branches of 
the Turkic language family. The differences between these two languages were increased due to 
intense Uzbek-Tajik contact. The original language of the Uzbeks was presumably similar to 
other Turkic languages, exhibiting an eight-vowel system with vowel harmony, suffixing 
morphology, and SOV word order. All of these features have been affected by contact with 
Tajik. With regard to phonology, Uzbek has lost its vowel harmony and has neutralized the 
distinction between the high front vowel /i/ and the high back vowel /ɨ/. In many contexts, the 
low back vowel /a/ came to be rounded to /ɔ/ (orthographic ), e.g. ol- ‘to take’ (Kazakh al-), 
oyoq ‘foot’ (Kazakh ayaq). Uzbek morphology has been supplemented with a number of Persian 
prefixes, such as be- ‘without’ (thus, be-ma’ni ‘without meaning, meaningless, absurd’). Many 
of these prefixes have nativized and are applied to native Turkic vocabulary: be-tinchlik ‘without 
peace, disorderly’. While SOV word order has been generally maintained in Uzbek (Tajik and 
standard Persian are, after all, SOV themselves), the head-final nature of older Uzbek has been 
altered by the addition of the Persian complementizer ki. In addition, the older Turkic strategy of 
nominalizing a complement clause and inserting it into the object slot is still available and 
preserves SOV word order in (7), but the new, borrowed strategy using ki results in SVO word 
order in (8) (Soper 1987). 
(7) 
Ali [Hokim yiqil-gan-i-ni] bil-a-di. 
Ali Hokim fall-
NMLZR
-3-
ACC
know-
PRES
-3 
‘Ali knows that Hokim has fallen.’ 

---

36 
(8) 
Ali bil-a-di-ki [Hokim yiqil-di]. 
Ali know-
PRES
-3-
COMP
Hokim fall-
PST
‘Ali knows that Hokim has fallen.’ 
The effects of Persian contact are not limited to Uzbek; other Turkic languages in contact 
with Persian have undergone similar changes, including Khalaj, Qashqa’i, and Azerbaijani 
(Soper 1987). Kazakh, however, has retained most Turkic typological characteristics, as contacts 
between the nomadic Kazakhs and sedentary Tajiks have been less intense and prolonged. 
The Persian influence upon sedentary populations is well attested historically. Texts in 
Chagatay, the medieval literary language of Turkic Central Asia and ancestor of Uzbek, are 
written in Perso-Arabic script, borrow heavily from Perso-Arabic vocabulary, and calque from 
Persian constructions. While all Central Asian Turkic peoples claim Chagatay literature as part 
of their heritage, the language actually used in this literature is that of sedentary Turks, that is, 
the ancestors of today’s Uzbeks. 
With regard to the expression of evidentiality and related meanings, it is Turkic that has 
influenced Persian, and not the other way around. Proto-Indo-European appears not to have 
grammaticalized evidentiality, and the closest relatives of the Iranian languages, the Indic 
languages, by and large also lack grammaticalized evidentiality. In most of the major Iranian 
languages, there exists something like grammaticalized evidentiality, mediativity, or indirectivity 
(see for Persian: Jahani 2000; Utas 2000; for Kurmanji: Bulut 2000). Of the most interest for the 
purposes of this work is, of course, Tajik. Fortunately, there exists a considerable body of work 
on evidentiality in Tajik, as this will allow for the future comparison of Tajik with Uzbek. 
According to Friedman (1979), evidentiality in Tajik functions similarly to other 
languages within the Eurasian evidentiality belt. There is a contrast between the simple past, 
which is confirmative, and the perfect, which is not marked for confirmativity. Complex or 

---

37 
doubled perfect forms bear similar non-confirmative meanings to those in Turkish and Balkan 
Turkic, namely, reportedness, admirativity, and dubitativity. Lazard (2000) illustrates the 
remarkable similarities between the normative description of Bulgarian and the Tajik system, 
which is shown in Table 9. Note that he employs a neutral/mediative distinction, rather than a 
confirmative/non-confirmative distinction. 
Table 9: Confirmativity in Bulgarian and Tajik