3.3. The relational center: verbs


The verb morphology of Udi is rather simple. Basically, we have to deal with a number of strategies to encode tense, aspect, and mood (TAM) - categories that are all indicated by suffixes to which in parts are added (floating) modal clitics. Furthermore, the finite verb is characterized by floating agreement clitics that usually are coreferent with the S=A center. Non finite verb forms are either identical with tense/aspect stems or derived from the stem with the help of (in parts fossilized) case suffixes (infinitive/masdar, participles, converbs).


The lexical basis of Udi verbs (‘verb stem’) may vary in complexity:


a) simple verbs: usually CV, VC, or CVC(V)-stems, e.g. p’o?- ‘to satisfy’, aq’- ‘take’, biq’- ‘take’, c^ur- ‘stand’, buq’- ‘want/love’, bes- ‘to ask for’ etc. 


b) lexical basis + light verb (lv): light verbs are (quoted in the masdar):


            esun    ‘to go’                         >          lv:intrans

            baksun ‘to become’                >          lv:intrans

            *desun                                   >          lv:trans

            *tesun                                    >          lv:trans

            *k’esun                                  >          lv:trans

            *kesun                                   >          lv:trans

            *q’esun                                  >          lv:trans

            *p’esun                                  >          lv:trans / lv:intrans

            pesun ‘to say’                         >          lv:trans / lv:intrans

            besun ‘to make, do’                >          lv:trans


Those light verbs that are marked by an asterisk are not used as simple verbs and do not have a specific semantic meaning. Some of them may perhaps be reanalyzed from allomorphs of e.g. pesun in a given assimilatory context. Examples are:


usk’esun                    ‘to measure’

zoq’albesun                ‘to count’

axs/umpesun              ‘to laugh’

campesun                   ‘to write’

va?baksun                 ‘to believe, be faithful’

c^urdesun                  ‘to want’

aizesun                       ‘to rise’

buibaksun                  ‘to be full’

besbesun                    ‘to kill’

t’ap’pesun                  ‘to hit’


Many of these verb forms are based on now fossilized incorporation strategies. Normally, the incorporated element does not play a role in the valence pattern of the verb (most probably, an adequate rendering og these INC elements would be one in terms of ‘adverbs’).


c) Incorporated element + heavy (‘full’) verb. Such verbs often respect the incorporated element in their valence pattern. Examples:


xabar-aq’sun              ‘to ask’            (lit. ‘to take news [from]’ + ABL)

as^-besun                  ‘to work’         (lit. ‘to make work’)

fikir-besun                  ‘to think’          (lit. ‘to make thought’)


Typical verbs that serve to derive such complex structures are: besun ‘to make, to do’, aq’sun ‘to take’, biq’sun ‘to take’, dug/sun ‘to hit’, saksun ‘to throw’, zapsun ‘to drag’.


d) An intermediate position between b) and c) is represented by many so-called ‘oriental’ verbs, that is verbs that are based on stems from Oriental languages (Osmanic/Azeri, Persian, Arabic). The most prominent structure is represented by Azeri -mis^-participles, e.g. sinamis^besun ‘to search’, güdmis^besun ‘to supervize’, busmis^besun ‘to watch, spy on’, bürmis^besun ‘to order’, ögmis^besun ‘tro praise’, bag/is^lamis^besun ‘to forgive’ etc. Oriental verbs are also those mentioned under c).




3.3.1 Masdar and infinite


Udi has two forms that cover the functions of a masdar (verbal noun) and of an infinitive. The so-called ‘first masdar’ is derived from verbal stems with the help of the suffix -es wwhich -historically speaking - represents the old (Proto-Lezgian) dative (*-s). This structure has a strong telic function and should be treated in terms of an infinitive. The ‘second masdar’ is derived from the first one with the help of a genitive (-es-un). In fact, the first masdar has a number of other case-like correpondences which show that this masdar oncve has been treated as a noun, cp.:


Stem:               bak-                                        ‘to become’                

DAT                bak-es                                    First masdar

            GEN                bak-s-un                     Second masdar           

            ERG                bak-s-in                      Modal converb

            DAT2              bak-s-ax                     Telic converb

            ABL                bak-s-axo                   Converb (anterior)


Nopte that -es- becomes -s- if preceded by a simple consonant but is retained if a CC-cluster precedes it.


The genitive of the first masdar (probably supplied with a partitive reading) has become a separate verbal noun which again can be inflected, e.g.:


ABS                baksun

ERG                baksunin

GEN                baksunun

DAT1              baksuna

DAT2              baksunax

ABL                baksunaxo

...                     .....




3.3.2 Tense, Aspect, Mood (TAM)


The TAM-system of Udi is in parts based on older case morphemes, but also on true temporal/modal suffixes. We can distinguish three basic groups:


a) Present (imperfective)

b) Past (perfective)

c) Future-modal


The present tense is marked by the suffix -(e)s-a. Historically speaking we have to deal with a ‘first masdar’ (cf. above) to which the old copula *’a ‘to be’ has been added. The construction that correponds to an Englksih continuous form (aq’-s-a ‘is taking’).


The past tense is represented by two suffixes: -i (aorist) and -e (perfect). These suffixes are added to the stem, e.g. aq’-i, aq’-e ‘having taken // being taken’. The difference between the two tense forms in not yet fully understood. Today, -e is preferred by some speakers, but the i-past (aorist) is nearly as frequent as the e-past. Moreover, the suffix -i is the general morpheme to encode a past event/state in predicative structures, but also to mark ‘relative’ past tenses, see below. In any case it can be claimed that the e-past  has a more perfective meaning that the i-past. But contrary to the i-past (aorist), the e-past cannot be used as a free participle.


The future-modal domain is represented by three paradigms: a) the standard present-future (suffix -al), the optative (or adhortative) that often plays the role of a general modal form (suffix -a), and the optative-future (suffix -o). The present-future and the optative have in common that they imperatively call for the agreement clitic to follow these suffixes.


Additionally, all TAM mentioned so far can be mark for a reference to the relative past with the help of the suffix -i which, however, can be seperated from the basic TAM-marker by AGR-morphemes (e.g. bak-i-ne-i ‘(s)he had been’). In older texts, the present-future morpheme can appear reduplicated (-al-al) as well as in a sequence with the optative (-al-a) denoting a present optative.


Summary of Udi TAM-forms


Basic Tense

Tense form


Relative Past






Present-Future (fut1)


























The TAM-based stem opposition because more transparent with those verbs that show stem suppletion. All these verbs function as light verbs, hence their frequency is rather high. The following table lists the suppletive forms:





















Note that ex-a ‘say:pres’ lacks the masdar morpheme. The final -a is dropped when the verb is used as a heavy verb (e.g. ex-ne ‘(s)he says, ex-q’un ‘they say’).


In case esun is used as a light verb (lv:intrans), its past tense stem usually is -c-, obviously a loan from Old Armenian: -c- corresponds to the Old Armenian -cc-Aorist (cp. Old Armenian ka-cc-i ‘I stood’). Yet, some verbs like aizesun ‘to rise’ have retained the -r-stem (cp. ai-ne-z-er-i ‘(s)he rose’).


The light verb desun shows metathesis in the masdar and in the present tense forms: desun > *dsun > st’un (present: -desa > -dsa > -st’a). The original stem is retained in the other tense forms, e.g. ta-st’a (< *ta-d(e)sa) ‘giving’, but tad-i ‘having given’).