3.2 The referential center: Nouns and Pronouns

The referential center of an Udi clause is represented by nouns or pronouns. Actants in S-function or A-function can be deleted or can be inferred from the context resp. context, cp.

ba-ne-k-e sa pasc^'ag/ me-t'-ai ba-ne-k-e-i xib g/ar. sa vaxt'-a fikir-re-b-i .....
be-3sg:s-$-perf one king prox-sa:obl-gen be-3sg:s-$-perf-past three son.one time-dat1 thought-3sg:s-lv:trans-aor...
'There was a king; he had three sons. Once he thought..'

The agreement clitic -ne- in fikirrebi is coreferential with pasc^'ag/, yet sa vaxt'a fikirrebi starts a new clause. This leaves us with the impression that Udi (in parts) follows the typology of a 'pro-drop' language.

Open referential entities are marked for case and number; nominalized adjectives, participles, numerals etc. make use of a specific set of inflectional morphemes (stem augemts etc.) they distinguishes them from nouns.



3.2.1 Nouns

The inflectional type of nouns is marked by a relatively high degree of allomorphy both in case and number marking. In most cases, the allomorphy is not predictable but must be regarded as being a part of the paradigmatic set that underlies the lexical entry. The order of suffixes is STEM-NUMBER-CASE which reveals the expected affinity of number marking to derivational features.




3.2.1.1  Number

The singular of Udi nouns is generally unmarked (but see below 3.2.2 fornominalized adjectives). the numeral sa can be added to denote a referentiallyunspecified entity (sa pasc^'ag/  'a (certain) king').

The standard distributive plural is nowadays marked by the suffix -ux which undergoes a significant shift in the oblique case forms (cp. 3.2.1.2). This morpheme can call for another preceding plural element that belongsto an older stratum of number marking in Udi. These 'old' plural morphemesare -ur-, -m-, and -q'- (in combination also -ur-m- ). There distribution is lexically determined though we can assume that their former function was connected with soome kind of lexical subcategorization [perhaps [human/animate] vs. [nonhuman/inanimate]). The younger morpheme -ux most probably is a loan morpheme that stem from the Old Armenianplural marker -kc (already a spirans in the source ('Azeri')variety of Old Armenian?). The alternative vocalization -ox stemsfrom the oblique cases (see 3.2.1.2). -ox may also appear in reduplication( -xox).

Historically speaking, the -ux morpheme seems to have had a more collective function. This can be inferred from those nouns that either have lexicalized the -ux element (such as c^ubux 'woman' (< coll.), imux 'ear' (< paral), elmux 'spirit' (< coll.) etc.) and from rare pairs such as pop 'a single hair', pop-ur 'single hairs', pop-ur-ux 'hair' (coll).

The -ur-plural without further plural element is characteristic for nominalized adjectives etc. (kala-o-r < *kala-o-(u)r) but may also appear with certain nouns. In at least three nouns, it has become lexicalized:

adamar 'man' < *adam-ar < *adam-ur (coll. reading),
g/ar 'boy' < Middle Armenian tg/a 'boy' (< Aramaic t.alya: 'boy') > Udi *g/a (initial CC-cluster is reduced) + -(u)r > g/a-r). Plural marking is in analogy to:
xinär 'girl' < *xin@-(u)r.




3.2.1.2 Case

Morphologically speaking, the Udi case system consists of eleven cases. The case morphemes are added to the lexical stem (if given, + plural morpheme) according to three inflectional types:

a) Absolutive inflection (AI): The zero-marked absolutive serve as the basis for the oblique cases;
b) Oblique inflection (OI): the oblique cases are characterized by an additional morpheme ('stem augment', SA), normally -n-,
c) Ergative inflection (EI): the ergative case serves as the basis for the other oblique cases.

EI is rather uncommon in Udi, whereas the distribution of AI and OI is somewhat 50:50. With nouns, AI or OI is lexically determined.

The basic shape of the case morphemes is illustrated below:


AI
 OI
 EI
Absolutive
 -ZERO
 g/ar
 mex
 xe
Ergative
 -en
 g/ar-en
 mex-n-en
 xe-n
Genitive
 -Vy, -un
 g/ar-i
 mex-n-ay
 xe-n-ey
Dative1
 -V
 g/ar-a
 mex-n-u
 xe-n-e
Dative2
 -V-x
 g/ar-a-x
 mex-n-u-x
 xe-n-e-x
Ablative
 -V-xo
 g/ar-a-xo
 mex-n-u-xo
 xe-n-e-xo
Comitative
 -V-xol
 g/ar-a-xol
 ---
 ---
Superessive
 -V-l
 g/ar-a-l
 mex-n-u-l
 xe-n-e-l
Allative
 -V-c^'
 g/ar-a-c^'
 mex-n-u-c^'
 xe-n-e-c^'
Adesssive
 -V-st'a
 g/ar-a-st'a
 mex-n-u-st'a
 xe-n-est'e
Benefactive
 -en-k'ena, -enk'
 g/ar-enk'ena
 ---
 ---

g/ar 'son'; mex 'scythe'; xe 'water' (note that OI is more common for xe).

A number of words show a genitive-ergative syncretism (morpheme -in ). cp. gädä 'boy' (Persian loan), ERG/GEN gäd-in (also gädinen vs. gädinay). Among them there are the follwoing terms for body parts that show loss of final consonant:

bul 'head' -> ERG/GEN b-in, DAT1 b-e etc.
kul 'hand' -> ERG/GEN k-in, DAT1 k-e etc.
pul 'eye' -> ERG/GEN p-in, DAT1 p-e etc.

In analogy: tur 'leg' (mostly probably a loan from Georgian) -> ERG/GEN tur-in, DAT1 tur-e
Loss of final consonant is also characteristic for k'odz^ 'house' (GEN k'odz^-in, but DAT1 k'u-a etc.)


General make-up of the inflectional paradigm:

The functional cases ERG and GEN are opposed to the rest of the case forms that basically serve to encode locative functions. This opposition is mirrored by the fact that all locative are derived from the dative1. The only excpetion is the so-called benefactive (-en-k'ena) which in fact is an ergattive to which the postposition -k'ena 'like, as' is added. Hence a phrase like

g/ar-en baba-enk'ena kag/e.z-ax cam-p-i-ne
son-erg father-ben letter-dat2 write-lv:trans-aor-3sg:a
'The son wrote the letter for the father'

really reads: 'the son wrote the letter as the father (would do it)'.

Allomorphs and functional (metaphorical) extensions:
CASE
 ALLOMORPHS
 Functional (metaphorical) extension
Ergative
 -en, -n, -in
 Instrumental, modal
Genitive
 -a(y), -e(y), -e(y); -un, -in
 S=A-marking with masdars
Dative1
 -a, -e, -i, -u
 Locative (allative, essive), Time, A>IO-marking
Dative2
-ax, -ex, -ix, -ux
 Locative (allative, essive) > O[def]-marking, A>IO-marking
Ablative
 -axo, -exo, -ixo, -uxo
 Partitive, backgrounded agent (rare), S=A in potential mood
Comitative
 -axol, exol, -ixol, uxol
 ---
Superessive
 -al, -el, -il, -ul
 Possessive in 'have'-clauses
Allative
 -ac^', -ec^', -ic^', -uc^'
 ---
Adessive
 -ast'a, -est'a, -ist'a, -ust'a
 Possessive in 'have'-clauses
Benefactive
 -enk'ena, -enken, -enk'
 IO-marking (rare)

In the plural, the allomorphy is suspended due to the following phonetic process:

*-ux-V- > -ux-o- > -ug/-o > -g/-o-

Hence, we have the following paradigm:
CASE
 Morpheme
 Example: g/ar-mux 'sons'
Ergative
 -on
 g/armug/on
Genitive
 -o(y)
 g/armog/o(y)
Dative1
 -o
 g/armug/o
Dative2
-ox
 g/armug/os
Ablative
 -oxo
 g/armug/oxo
Comitative
 -oxol
 g/armug/oxol
Superessive
 -ol
 g/armug/ol
Allative
 -oc^'
 g/armug/oc^'
Adessive
 -ost'a
 g/armug/ost'a
Benefactive
 -oenk'ena, -oenken, -oenk'
 g/armug/oenk'(ena)


NOTE: In Nidzh, ablative and comitative merge in one form (-xun).






j