Evidentiality in Uzbek and Kazakh

Chuvash r 
u/ăv 
ă 
Khalaj d 
Ag 
ux 
Lena t 
ïa 
I 
Sayan d 
Ag 
Ig 
Yenisey z 
Ag 
Ig 
Altay-Siberian y 
Ag 
Ig 
Altay y 
uu 
I 
Kyrgyz y 
oo 
I 
Kipchak 
West-Central Kipchak y 
Aw 
I 
Oghuz y 
Ag 
I 
Norm 
Turkic Central 
Turkic 
Southeastern y 
Ag 
IK 
While the classification shown above manages to account for most of the Turkic 
languages, there are a few that may not be readily assigned to any branch. Ili Turki, spoken 
along the Ili River in Xinjiang, shows how difficult it can be to classify the Turkic languages. On 
the one hand, it exhibits the Kipchak labialization of *g after low vowels (*AG → Aw), on the 
other, the Southeastern fortition of *g after high vowels (*IG → IK) (Hahn 1991). It may be the 
case that Ili Turki represents a fifth branch of Central Turkic, or it may be that these features 
arose from contact between speakers of Southeastern and Kipchak languages. 
4
Data from Tekin (2005), names of branches from Schönig (1999). 

---

31 
Crimean Tatar is another example of a language that cannot be easily classified. Features 
of both Oghuz and Kipchak are found mixed within the language, which exhibits both Ag and 
Aw as reflexes of Proto-Turkic Ag. While the original Turkic language of the Crimea was 
Kipchak, migrations of speakers of Oghuz languages to the Crimea during periods of Ottoman 
rule resulted in Oghuz and Kipchak-based varieties existing side by side. Subsequent population 
upheavals (namely, the migration of Greek Orthodox speakers to the Donetsk region, the 
decimation of Jewish and Karaite speakers during the Holocaust, and the deportation of Crimean 
Tatars to Central Asia) resulted in further mixing of these varieties, leading to the difficulty in 
classifying modern Crimean Tatar (Polinsky 1992). Due to the great similarity between most 
branches of Turkic and their proximity to one another, the mixing of languages of different 
branches is not uncommon, making classification a challenging task. 
Despite occasionally difficulties in classifying certain languages, a phylogram based 
upon a neighbor-adjoining network (Saitou and Nei 1987) reveals that the basic divisions that 
have been proposed thus far still hold. The tree in Figure 2 was created using SplitsTree (Huson 
and Bryant 2006) and takes into account the modern reflexes of eighteen phonological features, 
including the reflexes of *d and final velars. 

---

32 
Figure 2: Neighbor-Adjoining Tree of the Turkic Languages 
While Kazakh is solidly a Kipchak language, Uzbek does not always meet the criteria for 
being classified as Southeastern. In certain circumstances it, too, has lost final *G after high 
vowels, as in the adjective forming suffix *lIG, which has become -li in modern literary Uzbek 
(e.g. tog’-li ‘mountainous’). According to Schönig (2007), this indicates that Uzbek occupies an 
intermediate position between the Kipchak and Southeastern languages. As the loss of *G 
occurs only in this context, however, it is more likely that this is a contact effect. In support of 

---

33 
the contact hypothesis is the existence of a ‘dialects’ of Uzbek that are clearly more Kipchak 
than Southeastern (Daniyarov 1975). The official literary dialect of Uzbek, which is clearly 
more Southeastern in nature despite occasional Kipchak influences, is the dialect that is used in 
this study. In comparing Turkic languages from two different branches, it is hoped that the 
claims made about Uzbek and Kazakh may be applicable to a broad range of other Turkic 
languages.