3.2.1.6 Inflection of nominalized structures

A salient feature of Udi (as well as of many other East Caucasian languages) is the possibility to turn any (non finite) relational form into a noun. Nominalization is done with the help of a specific subset of inflectional morphemes that is based on the absolutive marker -o (note that with nominalized forms, the absolutive is not zero!). The plural makes use of the ‘old plural’ morpheme -r in the absolutive (> -or), but applies the nominal plural in the oblique cases. Nominalized always have an oblique stem augment (generally -t’-, both singular and plural), to which the following case morphemes are added (the table includes stem augments and plural morphemes):

	Singular	Plural
ABS	-o	-o-r
ERG	-t’-in	-t’-g/-on
GEN	-t’-a(y)	-t’-g/-o(y)
DAT1	-t’-u	-t’-g/-o
DAT2	-t’-u-x	-t’-g/-o-x
ABL	-t’-u-x-o	-t’-g/-o-x-o
COM	-t’-u-x-o-l	-t’-g/-o-x-o-l
SUPER	-t’-u-l	-t’-g/-o-l
ALL	-t’-u-c^’	-t’-g/-o-c^’
ADESS	-t’-u-st’a	-t’-g/-o-st’a
BEN	-t’-enk’(ena) / -t’-ink’(ena)	-t’-g/-o-enk’(ena)

E.g.:

kala-t’-in tängi-n-ax aq’-i-ne

big-sa:obl-erg money-sa-dat2 take-aor-3sg:a

‘The big (one) took the money’.

me k’ic’i-t’-uxo xabar-te-zu-aq’-e

prox little-sa:obl-abl news-neg-1sg:a-take-perf

‘I have not asked this little (one)’.

In the speech of some Udi, the absolutive marker -o tends to be extended to the oblique. Thus we may observe forms like kala-o-t’-a(y) ‘of the big one’, kala-o-t’-uxo ‘from the big one’ etc. (even kala-o-t’-g/-o(y) ‘of the big ones’).

Nominalization can apply to:

a) standard adjectives: kala > kala-o (> kalo:) ‘the big one’

b) participles: biq’al ‘taking’ > biq’al-o ‘one who takes’; besbi ‘killed’ > besbio ‘one who has been killed // who killed’;

c) any genitive: g/ar-ey ‘boy:gen’ > g/ar-e-o ‘that what belongs to the boy’; ic^i ‘refl:gen’ > ic^i-o ‘that what belongs to oneself’, besbit’ay ‘killed:gen’ > besbit’a-o ‘that belongs to those who haven been killed // who have killed’; bezi ‘my’ > bezio ‘that what is mine’ etc., cp.:

bütün g/ar-e-t’-ux aq’-i-zu

all boy-gen-sa:obl-dat2 take-aor-1sg:a

‘I took all what belongs to the boy’.

An impressive example of the use of nominalization is given by the following passage from the Gospels (Lk 16:10):

va?-bak-al-o k’ic’i-ot’-u gölö-un-t’-u-al va?-ne, nut’ va?-bak-al-o k’ic’i-ot’-u

faithful-lv-part:pres-nom:abs little-se:obl-dat1 much-gen-se:obl-dat1-foc faithful-3sg:s neg faithful-lv-part:pres-nom:abs little-se:obl-dat1

nut’ va?-bak-al-le gölö-un-t’-u-al.

neg faithful-lv-part:pres-3sg:s much-gen-se:obl-dat1-foc

‘He that is faithful in that which is least is faithful also in what is much: and he that is not faithful in what is least is not faithful in what is much, too’.

3.2.2 Pronouns

Udi makes considerable use of its pronominal system. It covers all aspects of standard pronominal reference including personality, deixis, reflexivity, questioned referents, indefiniteness, relative subordinators etc.

3.2.2.1 Speech Act Participants

The speech act participants are not subcategorized according to sexus. Number distinction is lexical, ignoring any feature of inclusiveness/exclusiveness. The function of social deixis to encode politeness in a speech act is covered by the second person plural (rare). The base forms are:

Cognitive	Communicative	Linguistic	Pronoun
EGO	SAP:1	1sg	zu
TU	SAP:2	2sg	(h)un
EGO + others	SAP:1pl	1pl	yan
TU (multiple)	SAP:2pl	2pl	va?n

Case inflection: Conforming to the Silverstein Hierarchy, the unmarked forms of speech act participants have a strong agentive semantics. Hence, they behave alike for S (subjective) and A (agentive) (S=A, accusative behavior), encoding both with the help of the absolutive case. But note that in case the pronoun is qualified by an attribute (normally postponed, hence case marked), the standard case distinction may be restored, cp.:

zu s/um-ax aq’-sa-zu

I:abs bread-dat2 eat-pres-1sg:a

‘I eat the bread’

vs.

zu damdam s^ähär-ä tag/-al-zu

I:abs tomorrow town-dat1 go:fut-fut-1sg:s

‘I will go to town tomorrow’.

but:

zu k’ala-t’-in s/um-ax aq’-sa-zu

I:abs big-sa:obl-erg bread-dat2 eat-pres-1sg:a

‘I - being the oldest one - eat the bread’.

vs.

zu kala-o damdam s^ähär-ä tag/-al-zu

I:abs big-sa:abs tomorrow town-dat1 go:fut-fut-1sg:s

‘I - being the oldest one - will go to town tomorrow’.

Note that the absolutive of all personal pronouns is marked by a so-called ‘determinative’ -n except for the first prson singular).

The oblique cases of personal pronouns are based on the nominal system except for the genitive which shows a diverging pattern based on older strategies of (class?) prefixation, cp. the following table which also lists the reconstructional aspects:

Genitive:

	Genitive	< PL	PREFIX	STEM	GEN	Base form
1sg	bez(i)		*b@-	-zw@-	-i	zu < *zw@
2sg	vi		---	*g/w-	-i	un < *g/w@-n
1pl	bes^(i)		*b@-	-z^y-	-i	yan < *z^y@-n
2pl	e?f(i)		*e-	z^w@?-	-i	wa?n < *z^w@?-n

The following table lists all case forms:

	1sg	2sg	1pl	2pl
ABS	zu	un	yan	va?n / efa?n
ERG	zu	un	yan	va?n / efa?n
GEN	bez(i)	vi	bes^(i)	e?f(i)
DAT1	za	va	ya	va?
DAT2	zax	vax	yax	va?x
ABL	zaxo	vaxo	yaxo	va?xo
COM	zaxol	vaxol	yaxol	va?xol
SUPER	zal	val	yal	va?l
ALL	zac^’	vac^’	yac^’	va?c^’
ADESS	zast’a	vast’a	yast’a	va?st’a
BEN	zenk’(ena)	venk’(ena)	yenk’(ena)	ve?nk’(ena)

3.2.2.2 The ‘third person’: Deixis

The Udi demonstrative pronouns subcategorzie the speakers deicitic space in a threefold, monocentric way. By ‘monocentric’ is meant that the speaker does not make reference to the ‘region’ of another speech act participant to localize an exophoric referent, but judges the distance from the point of view of his/her own region. Hence, the correlation between demonstrative pronouns and speech act participants (e.g. ‘this’ = 1sg, ‘there’ = 2sg, that over there’ = 3sg) does not hold for Udi except for a rather informal description. ‘Threefold’ means that subcategorization takes place on three levels: a) rather near to EGO, b) more distant, rather/very distant from EGO. Contrary to many other East Caucasian languages, there is no vertical subcategorization (e.g. ‘that above EGO or another SAP’ etc.).

All demonstrative pronouns can (in their nominalized form) be used for anaphoric purposes, covering the notion of English ‘he, she, it’. In many instances, the localization of a referent in the discourse is decisive for the use of pronouns, e.g.:

na?yne bez vic^-en arum bi-ne-‘q-i. s^e-t’-u mo-t’-xo härisä hazir-b-es-ax bu-t’u-q’-sa.

yesterday my brother-erg wheat buy-3sg:a-$-aor dist-sa:obl-dat1 prox-sa:obl-abl porridge:abs prepare-lv-inf-dat2 want-3sg:io-$-pres.

‘Yesterday, my brother brought some wheat. He wants to prepare a porridge of it.’

Yet,. it should be noted that the distal has the highest frequency with respect to anaphorics.

The base form of the demonstratives is documented by their attributive form:

Proximal me

Media ka

Distal t’e

These forms are sensitive neither for case nor for number, eg.:

me kala k’odz^

prox big house:abs

‘this big house’

ka p’a? vic^i

med two brother

‘those two brothers (over here)’

t’e g/i-mxox

dist day-pl:abs

‘those days’

When nominalized, the distal has a different stem (s^e-) which cannot be used in attributive function (*s^e adaramar ‘that man’ (recte: t’e adamar).

Nominalization is based on the inflectional pattern of other nominalized forms (cp. 3.2.1.6). Yet, especially in the dialect of Vartashen the formation of the absolutive differs from that of other nominalized forms in that it takes an additional ‘determinative’ -n- is also present in personal pronouns (except 1sg). In Nidzh, this element is dropped. The oblique case forms are those of other nominalized forms:

	PROX	MED	DIST
ABS	me-n-o	ka-n-o	s^e-n-o
ERG	me-t’-in	ka-t’-in	s^e-t’-in
GEN	me-t’-a(y)	ka-t’-a(y)	s^e-t’-a(y)
DAT1	me-t’-u	ka-t’-u	s^e-t’-u
DAT2	me-t’-u-x	ka-t’-u-x	s^e-t’-u-x
ABL	me-t’(-u)-x-o	ka-t’(-u)-x-o	s^e-t’(-u)-x-o
COM	me-t’(-u)-x-o-l	ka-t’(-u)-x-o-l	s^e-t’(-u)-x-o-l
SUPER	me-t’-u-l	ka-t’-u-l	s^e-t’-u-l
ALL	me-t’-uc^’	ka-t’-uc^’	s^e-t’-uc^’
ADESS	me-t’-u-st’a	ka-t’-u-st’a	s^e-t’-u-st’a
BEN	me-t’-enk’(ena)	ka-t’-enk’(ena)	s^e-t’-enk’(ena)

The plural forms are derived as follows (see 3.2.1.6 for the remaining case forms):

Proximal: me-n-o-r, oblique me-t’-g/-

Medial : ka-n-o-r, oblique ka-t’-g/-

Distal : s^e-n-o-r, oblique s^e-t’-g/-

The stem vowel -e- of the proximal and distal tends to be assimilated to a subsequent -o-, thus we likewise observe forms like mono, s^onor, s^ot’g/on (erg) etc. The resulting stem vowel -o- may also be extended to case forms that lack -o- in the suffix, e.g. mot’u / s^ot’u (dat1), even mot’a(y) / s^ot’a(y) (gen).

3.2.2.3 Reflexive pronoun

Reflexive strategies in Udi are pronominal based. The reflexive pronoun is ic^ which in fact is a noun with uncertain meaning (see 4.6 for the description of the functional scope of this pronoun). ic^ can be both used attributively and in nominal function. It lacks nominalization because of its former nominal status, cp.:

adamar-en ic^ xunc^i-n-ax be?-ne-g/-i

man-erg refl sister-sa-dat2 observe-3sg:a-$-aor

‘The man observed his sister’.

adamar-en ic^-ux gölö t’ap’-ne-xa

man-erg refl-dat2 much hit-3sg :a-lv:pres

‘The man hits himself very much’.

The reflexive pronoun can be inflected for case and number, cp.

	Singular	Plural
ABS	ic^	ic^-g/-on
ERG	ic^-en	ic^-g/-on
GEN	ic^-i	ic^-g/-o(y)
DAT1	ic^-u	ic^-g/-o
DAT2	ic^-u-x	ic^-g/-o-x
ABL	ic^-u-x-o	ic^-g/-o-x-o
COM	ic^-u-x-o-l	ic^-g/-o-x-o-l
SUPER	ic^-u-l	ic^-g/-o-l
ALL	ic^-u-c^’	ic^-g/-o-c^’
ADESS	ic^-u-st’a	ic^-g/-o-st’a
BEN	ic^-enk’(ena)	ic^-g/-o-enk’(ena)

Note that the reflexive pronoun lacks a case form to encode the absolutive plural (*ic^-ux).

As it is common in many other East Caucasian languages, the reflexive pronoun tends to appear in a syntactically motivated ‘redulpicated’ form especially when used in ‘close reflexive structures’ (as opposed to ‘long distance reflexivity’, cp. 4.6). The standard way of reduplication is the use of the ergative marked reflexive followed by the reflexive in a case form that conforms to the general case frame of the sentence, e.g.:

s^ägird-g/-on ic^-en ic^-bos^ fikir-q’un-b-i

disciple-pl-erg refl-erg-refl-pp(in) thought-3pl:a-make-aor

‘The disciples thought by themselves’.

s^e-t’-in ic^ partal-ax ic^-en ic^-laxo la-ne-xa

dist-sa :obl-erg refl shirt-dat2 refl-erg- refl-pp(on) put=on-3sg :a-lv:pres

‘(S)he put on his/her shirt.’

Note that with postpositions, the expected case form is often missing (the bare stem being used instead).

3.2.2.4 Interrogative pronouns

a) who? [+hum], sometimes [+anim], is expressed by s^u which is inflected on a nominal basis:

ABS s^u, ERG s^in, GEN s^i, DAT1 s^u, DAT2 s^ux etc.

There is no number distinction as for instance i n some Samur languages (Tabasaran, Agul, Rutul) or in Archi.

b) ‘what’ [-hum] has a numer of allomorphs:

e (Vartashen, used attributively only)

he (Nidzh, used attributively only)

k’a (Nominalized)

ek’a (Vartashen, nominalized)

hik’ä (Nidzh, nominalized)

The nominalized form ek’a (hik’ä) in inflection according to the pronominal paradigm (cf. 3.2.1.6): the element -k’- is replaced by the stem augment -t’- [note that the pronoun k’a ‘what’ cannot be inflected at all]:

ABS ek’a, ERG et’in, GEN et’a(y), DAT1 et’u, DAT2 et’ux usw.

The element e- serves to derive a number of other interrogative pronouns such as e-q’ara ‘how much’, e-ma ‘how much’, e-vaxt’ ‘when’, e-dz^ürä ‘how’ etc.

c) ‘where?’ [LOC]: ma.

The locative function is basically ‘essive’. An ablative meaning is represented by mac^xo ‘where from’ which perhaps stems from *ma c^e-xo ‘from which side’. An allative can be indiacted by adding the dative1 (maa // maya). The element ma serves to derive a great number of locative adverbs.

d) ‘which one?’, ‘which?’

This interrogative strategy has been grammaticalized with the help of the locative pronoun ma (see above) which is nominalized in the way of demonstrative pronouns (cp. 3.2.2.2):

SG PL

ABS ma-n-o ma-n-o-r

ERG ma-t’-in ma-t’-g/-on

GEN ma-t’a(y) ma-t’-g/-o(y)

DAT1 ma-t’-u ma-t’-g/-o

DAT2 ma-t’-u-x ma-t’-g/o-x

etc…..

Note that mano is also used in attributive function:

ma-n-o-a me g/ar-mug/-oxo haq’ullu

which-sa:abs-abs-3sg:s:q prox son-pl-abl smart

‘Which of these son is the smartest one?’

ma-n-o s/um-ax bu-va-q’-sa

which bread-dat2 want-2sg:io-$-pres

‘Which bread do you want?’

3.2.2.5 Relative pronoun

The relative pronoun corresponds to the interrogative pronoun mano (see above) to which the general subordinator -te is added (note that the clitic -te can appear later in the clause), cp.:

t’e adamar-ux ma-n-o-r-te t’eg/i ar-i-q’un gölö busa-q’un

dist man-pl:abs rel-sa:abs-abs-pl-sub today arrive-aor-3pl:s much hungry-3pl:s

‘Those men who have arrived today are very hungry’.

[alternative: t’e t’eg/i ari adamarux gölö busaq’un]

The inflection of the relative pronoun conforms to that of the interrogative pronoun mano.

3.2.3 Numerals

In Udi, there are two way systems of numerals: a) the autochthonous system which is basically vigesimal, and b) the Azeri system which tends to be more and more adopted by Udi speakers. In this section, only the Udi forms are presented.

The basic ordinals (1-10) are as follows:

1 sa

2 p’a?

3 xib - xe.b - xe.?b

4 bip’

5 qo

6 u?q

7 vu?g/ - vug/

8 mu?g/ - mug/

9 vui

10 vic’

Numbers from 11 to 19 are formed by adding -(e)c’c’e to the first decade which is derived from vic’ ‘ten’ (an opaque element -e is added). The numerals of the first decade may experience slight (in parts idiosyncratic) modifications (especially in allegro speech):

11 sac’c’e

12 p’a?c’c’e

13 xibe’c’ce

14 bip’ec’c’e

15 qoc’c’e

16 u?qec’c’e

17 vu?g/ec’c’e

18 mu ?g/ec’c’e

19 vuic’c’e - vuiec’c’e

The decades are:

20 q’a - q’o

30 saq’ovic’ (lit. ‘one twenty ten’)

40 p’a?q’o (lit. ‘two twenty’)

50 p’a?q’ovic’ (lit. ‘two twenty ten’)

60 xibq’o (lit. ‘three twenty’)

70 xibq’ovic’ (lit. ‘three twenty ten’)

80 bip’q’o (lit. four twenty’)

90 biq’q’ovic’ (lit. ‘four twenty ten’)

The hundreds are based on bac/ ‘hundert’:

100 (sa)bac/

200 p’a?bac/

300 xibbac/

400 bip’bac/ - bip’p’ac/

500 uqbac/

600 qo?bac/

700 vu?g/bac/

800 mu ?g/bac/

900 vuibac/

1000 hazar

5376 : uqhazar xibbac/ xibq’o u?qec’c’e

5000 300 3x20 16

The ordinals are derived from cardinals with the help of the suffix -(u)mdz^i (loan from Azeri), e.g.

sa(u)mdz^i first

p’a?umdz^i second

xibumdz^i third

bip’umdz^i fourth

qo(u)mdz^i fifth

u?qumdz^i sixth

vu?g/umdz^i seventh

mu?g/umdz^i eighth

vuiumdz^i nineth

vic’umdz^i tenth

Ordinals can be nominalized with the help of the nominalizing suffix -o (cp. 3.2.1.6), eg.: saumdz^io ‘the first one’; compare the two following sentences:

s^e-t’-in p’a?c’c’e-t’-uxo xabar-re-aq’-i

dist-sa:obl-erg twelve-sa:obl-abl news-3sg:a-take-aor

‘(S)he asked the twelve (ones)’.

s^e-t’-in p’a?c’c’e-umdz^i-t’-uxo xabar-re-aq’-i

dist-sa:obl-erg twelve-ord-sa:obl-abl news-3sg:a-take-aor

‘(S)he asked the twelfth one’.

A now rather obsolete, but in fact autochthonous way of encoding ordinals is the use of the genitive with the cardinals, e.g. p’a?u?n ‘second’, xib-un ‘third’ etc.