Evidentiality in Uzbek and Kazakh

6.1.2 Differences between Uzbek and Kazakh 
As described in Chapter 1, Uzbek and Kazakh belong to different branches of Turkic and arose 
from very different sociolinguistic situations, yet their means of expressing evidential and related 
meanings are quite similar. Through much of this work, it has been possible to refer to cognate 
morphemes in each language as though they were the same (e.g. -(i)b/-(I)p, ekan/eken). The 
similarities between Uzbek and Kazakh suggest that what has been described in this work will be 
similar to what occurs in the other Turkic languages of Central Asia. There are, however, a 
number of minor differences between Uzbek and Kazakh, at least as far as the topic of this work 
is concerned. By briefly explaining those differences I hope to provide a groundwork for 
anticipating where other languages may vary. 
Four principal differences in the morphosyntactic distribution of forms separate Uzbek 
and Kazakh. The first two differences, at least in terms of the the topic of this work, are the 
ordering of ekan/eken and the question particle and the ability of ekan/eken to follow the past 
tense -di/-DI. 

---

162 
In Uzbek, it is possible for the question particle to follow the evidential marker: ekan-mi.
In Kazakh, this is not allowed: *eken-be. When the question particle follows the evidential 
marker in Uzbek, the preceding item must not be a finite form of the verb; that is, in evidential 
and rhetorical questions, where the past tense -di may precede the evidential marker, the question 
particle must intervene between the past tense marker and the evidential marker: 
(232) a. 
qil-di-mi-kan? (Uz) 
do-
PST
-
Q
-
EVID
‘Did he (apparently) do?’ 
b. 
*qil-di ekan-mi? 
do-
PST EVID
-
Q
While the past tense marker -di is limited to questions in Uzbek, in Kazakh there is a 
growing tendency, although still rare, for the cognate past tense maker -DI to appear in non-
interrogative contexts. 
(233) Täwelšilik el-imiz kel-di eken.  (Kaz) 
 
Freedom country-1
PL
come-
PST EVID
 
 
‘Freedom has come to our country!’
 2
While still rare in independent clauses, the combination of finite (i.e. conditional, past) forms of 
the verb with the eken has become quite common in Kazakh cause-and-effect constructions 
formed with the complementizer deb: 
(234) De-gen-men, dağdarïs kel-di eken dep, qol quwsïr-ïp qara-p otïr-uw-ğa
say-
PRF
-1
SG
crisis come-
PST EVID COMP
arm cross-
CVB
look-
CVB
sit-
INF
-
DAT
bol-ma-y-dï. (Kazakh) 
be-
NEG
-
PRES
-3 
2
Elemes, Kerim. 2010. “Abay ädebiyetin tanïğanmen, Ibrahimniŋ ädebiyetine boylay almay 
žürmiz.” Interfax-Qazaqstan, 6 May. Accessed 6 Oct 2010. 
http://www.interfax.kz/?lang=kaz&int_id=quotings_of_the_day&news_id=59 

---

163 
‘As I’ve said, because a crisis has apparently come, we can’t sit around waiting with our 
arms crossed.’
3
It remains to be seen whether this pattern will continue to expand beyond this sort of 
construction, and whether any semantic change will accompany this morphosyntactic change.
Constructions of this type are not found in Uzbek. 
A third difference between Uzbek and Kazakh is the presence in Uzbek of a particle -dir, 
derived from the verb tur- ‘to stand’. This morpheme is not present in Kazakh, but in Uzbek, it 
attaches to non-finite predicates and indicates a high degree of confidence. 
(235) Alloh so’z-i-dan chiroyli so’z yo’q-dir! (Uzbek) 
Allah word-3-
ABL
beautiful word 
NEG
.
EXIST
-
MOD
‘There is surely no word more beautiful than the word Allah!’
4
Most forms of the verb, and all non-verbal predicates are unmarked for 
STATUS
/
MODALITY
, and 
therefore can employ -dir to indicate this. It is especially common for -dir to be attached to the 
perfect -gan, which is unmarked not only for confirmativity, but for any form of 
STATUS
/
MODALITY
. 
(236) qiynoq-qa sol-ish uchun dunyo-ga kel-gan-dir-miz. (Uzbek) 
suffering-
DAT
undergo-
NMLZR
for world-
DAT
come-
PRF
-
MOD
-1
PL
‘Surely, we came into this world in order to suffer.’
5
In Uzbek, then, -dir can be seen as a sort of complement to ekan, which likewise attaches to non-
finite predicates. 
The fourth and final major difference between Uzbek and Kazakh is the form and 
behavior of the modern reflexes of older *er-miš. The behavior of these forms is fully discussed 
in Chapter 5. The Uzbek reflex of this form is emish, and it behaves quite similarly to ekan, 
3
Arnur, Asqar. 2009. “Ulïqpan Žoldasov: Almatïnï aralamağanïma üš žïl boldï.” Dastan 
Studiyasï, 11 Sep. Accessed 6 Oct 2010. http://www.dastan-studio.kz/?p=1904 
4
2007. “Allah” Youtube, 3 Sep. http://www.youtube.com/watch?v=kpgHtXphPz8 
5
2008. Ma’rifat, 9 Jan. Accessed 15 Feb 2011. 
old.marifat.uz/uzbl/xalqaro_hayot/oyparastlar_olkasi.mg 

---

164 
except that when employed to indicate non-firsthand information source, the only possible 
interpretation is one of reportativity, rather than the full range of non-firsthand information 
source. Like ekan, emish may be employed to express admirativity. These facts indicate that 
emish should be considered non-confirmative, so it is perhaps due to issues of markedness that 
emish is somewhat more restricted than ekan in its possible interpretations. 
The Kazakh reflex of *er-miš is the clitic -mIs. Unlike Uzbek emish, -mIs may appear 
essentially anywhere, even after finite or confirmative forms of the verb. Moreover, it is not 
employed to express admirativity. Given the morphological promiscuity of this form, its ability 
to co-occur with confirmative forms, and its inability to express admirativity, we must conclude 
that Kazakh -mIs has evolved into a marker of reportativity, and that it does not interact with the 
non-confirmativity paradigm outlined in the previous chapters.