ArticlePDF Available

The Status of the Non-Finite - OmstO Morpheme in Erzya

January 2011
Linguistica Uralica 47(1):41

DOI:10.3176/lu.2011.1.03

Authors:

University of Helsinki

This article approaches the +OmstO formative used in deverbal inflection from a concatenational perspective. It describes the morphological distinction between the elative-case non-finite in +Om+stO (sams → samsto'to arrive'; oznoms → oznomsto'to pray'; molÍems → molÍemste'to go') and its counterpart the elative-case deverbal noun in +OmA+stO (sams →samosto'to arrive'; oznoms → oznomasto'to pray'; molÍems → molÍemasto'to go'). These parallel forms are then subjected to further morphological and semantic inspection: (1) compatibility with the three declination types, indefinite (zero), possessive (cross-referential adnominal person morphology) and definite (determiner) marking, and (2) compat- ibility with functions attributed to the elative case of other common nouns. Finally, a tendency is discerned according to which both inflection types can be regarded as elativecase nouns distinguished by a [±progressIve] parameter, whereby the so-called gerund is indicates a process, and the so-called deverbal noun a result.

Content uploaded by Jack Rueter

Content may be subject to copyright.

The$status$of$the$non-finite$-OmstO%morpheme$in$Erzya$

Abstract%

This article approaches the +OmstO formative used in deverbal inflection from a

concatenational perspective. It describes the morphological distinction between the elative-

case non-finite in +Om+stO (‘to arrive’ sams => samsto; ‘to pray’ oznoms => oznomsto; ‘to

go’ moĺems => moĺemste) and its counterpart the elative-case deverbal noun in +OmA+stO

(‘to arrive’ sams => samosto; ‘to pray’ oznoms => oznomasto; ‘to go’ moĺems =>

moĺemasto). These parallel forms are then subjected to further morphological and semantic

inspection: (i) compatibility with the three declination types, indefinite (zero), possessive

(cross-referential adnominal person morphology) and definite (determiner) marking, and (ii)

compatibility with functions attributed to the elative case of other common nouns. Finally, a

tendency is discerned according to which both inflection types can be regarded as elative-

case nouns distinguished by a [±PROGRESSIVE] parameter, whereby the so-called gerund is

indicates a process, and the so-called deverbal noun a result.

Preliminaries%

Grammars of Erzya lack consistency in the description of the morphology of gerunds,

deverbal nouns and infinitives. Let us therefore introduce ourselves to a concatenational-

morphological perspective of: verb stem types; the deverbal noun in OmA, and the locative-

case infinitive in OmO (elsewhere known as the nominative or even lative infinitive).

The tradition of verb-stem analysis varies, whereas vowel loss and retention in

different cells of verbal inflection has been the target of research in most traditions of Erzya

grammar (cf. Evsev'ev 1963: 193-198). More recent presentations of the language have

discerned a subset of stems relevant to the morphological correlation observed between finite

conjugation forms and non-finite forms. In the modern Erzya literary language the illative

infinitive in ms is attested in three vowel contexts, e.g. a+ms, e+ms and o+ms (Cyrillic

а+мс, я+мс, е+мс, э+мс о+мс [no instances of ё+мс attested]). The general consensus

advocates a system of two verb types, which, in an almost over-simplified way, can be

established by the parameter [±VOWEL RETENTION] before the formative ś IND.PRETI.PRED-

3SG, (see Pall 1996: 20; Zaicz 1998: 188-189; Bartens 1999: 122; Tsypkaikina 2000: 146;

Trosterud 2006: 250-251).

When we apply the two-way split directly to the three stem vowels (a, e and o), it will

be noted that all verbs ending in -ams retain the stem-final -a before the indicative preterit I

3SG formative ś. The infinitives ending in the mid vowels (o and e), however, cannot be

associated with such a straight-forward system. Verbs ending in ems and oms must be

learned separately; some retain their mid vowel before the ś formative and others lose it; a

general rule of thumb is that deverbal verbs in the so-called frequentative formatives śe, ńe

and kšnO retain their stem vowels, (cf. Evsev'ev 1963: 198; Pall 1996: 20; Trosterud 2006:

250-251).

On method of indicating stem-vowel retention of loss in dictionaries is to insert a pipe

“|” at the appropriate break point. This is precisely what the most recent Erzya-Russian

Dictionary (henceforth ERD 1993) does. Providing no specifics, the ERD gives forms, such

as maks|oms, -ś ‘to give’ and soda|ms, -ś ‘to know; to recognize’, whereby the reader is

readily aware of the appropriate IND.PRETI.PRED-3SG forms maksś and sodaś, respectively.

Hence ERD 1993 provides readers and language learners with ready access to the workings

of the Erzya verb stems. The verb sato|ms, -tś ‘to suffice’, however, takes us be surprise, i.e.

in addition to the two types demonstrated above, there is a small subset of verbs ending in

ťems and toms that insert a T before the IND.PRETI.PRED-3SG ś formative, something

reminiscent of the inessive-case forms attested in the Kozlovka variety of Erzya ‘oven’

kaštomo => kaštot+so ‘in the oven’, (cf. Bubrikh 1930: 22). For a three-way split in the

Erzya verb-stem types, see table (1.), analogous information on noun-stem types is available

in Rueter Adnominal person in the morphological system of Erzya (forthcoming).

Erzya verb-stem system attests to a three-way split

Verb stem type

Gloss

+INF+ILL

+IND.PRETI.PRED-3SG

Consonant-

final

‘to give’

maks+Om+s

maksoms

maks+ś

maksś

‘to bind’

sod+Om+s

sodoms

sod+ś

sodś

‘to go’

moĺ+Om+s

moĺems

moĺ+ś

moĺś

Vowel-final

‘to scold; to

curse’

śudo+Om+s

śudoms

śudo+ś

śudoś

‘to carve’

lakśe+Om+s

lakśems

lakśe+ś

lakśeś

‘to know; to

recognize’

soda+Om+s

sodams

soda+ś

sodaś

‘to repeat’

polaksa+Om+s

polaksams

polaksa+ś

polaksaś

T-stem

‘to suffice’

satoT+Om+s

satoms

satoT+ś

satotś

‘to become

tongue-tied’

kundśeťeT+Om+s

kundśeťems

kundśeťeT+ś

kundśeťeťś

Table 1.

In observing the three verb types illustrated in table (1.), we will note that the

archiphone O has reflexes in ZERO with vowel-final and T-stem types, and mid vowels (o or

e) depending on the front-back harmony triggered by the adjacent vowels or palatalized/non-

palatalized consonants of the preceding stem, e.g. ‘to bind’ sod+Om+s => sodoms, and ‘to

go’ moĺ+Om+s => moĺems.

The deverbal noun in OmA and infinitive in OmO present us with a second variety of

vowel harmony; mid-low dissimilatory vowel harmony. Thus, while the archiphone O, as

indicated above, has three reflexes at the surface level according to assimilatory vowel

harmony, (Ø, o, e), the archiphone A has a reflex in a when the preceding surface vowel is a

mid vowel, and an o when the preceding surface vowel is a low vowel, see table (2.).

Deverbal nouns and locative-case infinitives in the verb-stem system of Erzya

Verb stem type

Gloss

+DVN

+INF+LOC

Consonant-

final

‘to give’

maks+OmA

maksoma

maks+Om+O

maksomo

‘to bind’

sod+OmA

sodoma

sod+Om+O

sodomo

‘to go’

moĺ+OmA

moĺema

moĺ+Om+O

moĺeme

Vowel-final

‘to scold; to

curse’

śudo+OmA

śudoma

śudo+Om+O

śudomo

‘to carve’

lakśe+OmA

lakśema

lakśe+Om+O

lakśeme

‘to know; to

recognize’

soda+OmA

sodamo

soda+Om+O

sodamo

‘to repeat’

polaksa+OmA

polaksamo

polaksa+Om+O

polaksamo

T-stem

‘to suffice’

satoT+OmA

satoma

satoT+Om+O

satomo

‘to become

tongue-tied’

kundśeťeT+OmA

kundśeťema

kundśeťeT+Om+O

kundśeťeme

Table 2.

It should be noted that although the modern standard attests a distinction between

deverbal nouns and locative-case infinitives in all but a-stem verbs, this has not always been

the situation. One need only consult a publication of Kirillov (1987: 40) …eź ćiďardt

kunsolomanzo… ‘he [Okhima] didn't have the patience to listen to him’ to observe deverbal-

type forms in -oma- where -omo- would be expected. Here the word form kunsolo+ma+nzo

‘to listen to him’ appears to have a ma segment before the 3SG formative OnzO, something

which occurs in some forms of Erzya spoken in the Atyashevo raion. Therefore it cannot be

taken for granted that even speakers of Erzya would be in full agreement upon the distinction

between deverbal nouns and infinitive forms.

In this treatise of Erzya, I annotate the Om+O infinitive as locative-case, a matter

which by no means is unproblematic. Other annotations call it a nominative or a lative, they

have their morphological short-comings. It is most likely the notion of ZERO marking that

has been deemed sufficient when designating this infinitive as a nominative. Bartens,

however, (FUF XXXVIII 247-; 1979: 25-26) indicates the problems of a nominative-case

interpretation and points out the semantic and pattern-based criteria for a lative or locative

interpretation. An additional morphological criterion can be cited in opposition to the

“nominative” designation, namely, the formative OzO used in marking a nominative singular

subject-function possessum of a 3SG possessor never occurs with the deverbal Om+O

formative sequence; instead it is the formative OnzO the one that co-occurs with all other

cases and the nominative plural that is attested. Hence the transitive verb forms, e.g.

saj+eme+nze ‘to take it/him/her’, the only ones that are compatible with possessor indexing,

cannot be glossed as nominative singular. A nominative singular gloss would require an

ungrammatical form *saj+eme+ze. Morphologically, we are presented with two choices: we

can interpret the OmO infinitive form as one with syntactic functions used in appositive

expressions such as what is found in associative-collective numerals (kolmo+ńe+nze

‘him/her/it and two others (literally: the three of him/her/it)’), with no case designation, or we

can appropriate a case name, such as locative, which would set the non-finite verbal case

patterns in parallel with those of some adverb/adposition case patterns, see tables (3a-b.).

Non-finite case patterns

Gloss

Ill

(Transl

Loc

Abl

Ine

Ela

Prol

‘to sing’

moram+s

moram+ks)

moram+o

moram(o)+do

moram+so

moram+sto

moram+ga

Table 3a.

In table (3a) the literary illative infinitive is presented adjacent to its translative

counterpart, familiar from some of the Erzya dialects. Functions attributed to the literary

illative form include those, which in the case of common nouns would be readily associated

with the translative, e.g. predicate complement. The ablative case presents a morphological

conflict where the consonant-final stem type associated with non-finite inflection tends to

occur in the literary language with an intermediate vowel. This intermediate vowel, however,

should not be passed off as a direct indication of deverbal noun morphology, namely, the

verb ‘to go’ moĺems attests to three variants in published Erzya literature: moĺemado,

moĺemeďe and moĺemďe. The consonant-final stem of the ablative and the inessive forms is

quite infrequent in the written language of the Erzya majority corpus (described in Rueter,

Adnominal person in the morphological system of Erzya, forthcoming). The non-finite

elative form has functions attributed to the elative form of other common nouns, i.e. source

and temporal setting, and therefore provides no arguments for an explicit illative s + ablative

DO concatenation in a synchronic treatise of the language, although a theoretical diachronic

treatise of the language would.

In table (3b.) the translative-case form or its homonym with the formative ks is used

in noun derivation of spatial adverbs/adpositions with four-case declination patterns (lative,

locative, ablative and prolative).

Spatial adverb/adposition case patterns for the Erzya word alo ‘under; below’

Transl

Lat

Loc

Abl

Prol

Adv / Postp 1

al+ks

al+ov

al+o

al+do

al+ga

Table 3b.

Thus we can observe an affinity between the two tables (3a-b): the non-finite patterns

although preferring an illative formative in the literary language, demonstrate a dialect

translative counterpart to parallel the noun-derivation practice attested for adpositions, and

the local cases locative, ablative and prolative are attested in both the non-finite morphology

and the four-case pattern of this adverb/adposition type. Differences can be observed in the

two case patterns. The spatial adverb/adposition, it would seem, is incompatible with the

temporal-function cases of the illative and elative as well as the instructive-function case, the

inessive, whereas the spatial notion attributed to the inessive would provide no additional

information to the locative, already present. The absence of the lative formative Ov in the

non-finites, however, appears problematic. In fact, the OmO infinitive is generally used

where a target function would be expected, that is, in combinations with verbs of motion, e.g.

the construction tujems (ťikše) ĺeďeme to-go_V.INF.ILL hay_N.ABS to-mow_V.INF.LOC ‘to go

make hay’ is much more common than the deverbal-noun construction with the explicit lative

case tujems ĺeďma+v to-go_V.INF.ILL to-mow_V.N+LAT ‘to go make hay’ (Kutorkin 1976:

82).

In a description of the case pattern encountered in consonant-stem non-finite

morphology, we can assume a 7-slot paradigm, lacking the three core cases (nominative,

genitive and dative), see enumeration.

+Om+s, +Om+O, +Om+dO, +Om+sO, +Om+stO, +Om+ga, +Om+ks

+INF+ILL, +INF+LOC, +INF+ABL, +INF+INE, +INF+ELA, +INF+PROL, +INF+TRANSL

The 7-slot non-finite pattern can be contrasted with the 12-slot pattern of the deverbal

noun in OmA, which attests to core-case functions not available to the non-finite derivations.

+OmA+0 +OmA+ń, +OmA+ńeń, +OmA+dO

+DV-N+NOM. SG, + DV-N+GEN, + DV-N+DAT, + DV-N+ABL,

+OmA+sO, +OmA+stO, +OmA+va, +OmA+ška,

+INF+INE, +INF+ELA, +INF+PROL, + DV-N+COMP,

+OmA+s +OmA+v +OmA+ks +OmA+vtomo,

+ DV-N+ILL, + DV-N+LAT, + DV-N+TRANSL, +INF+ABE

Earlier (Bartens 1979: 45) it has been assumed that Mordvin possessor indices affixed

onto infinitives indicate the patient as opposed to the agent. This would mean that the notion

of a distinction [±INFINITIVE] would appear manifest in transitive but not intransitive verbs.

The Erzya language, however, appears to follow a possessor-index strategy where the

primary argument of the verb is the possessor, and this strategy holds for non-finite in Om-

and deverbal nouns in OmA, alike, see (1), where the possessor indexed in (1a) is the S and

that in (1b) the P. Thus the nominative-case deverbal noun can be demonstrated to operate

according to the non-finite argument structure.

(1a)

źarïja ška+ń juta+ź, okojńiki,

some_PRON-Q.ABS time_N+GEN to-pass_V+PTC-Z, finally_ADV,

maŕav+ś bus+ońť sa+mo+zo.

to-be-heard_V+IND.PRETI.PRED-3SG bus_N+GEN.DEF.SG to-arrive_V+DVN+POSS-3SG.NOM.SG

(Motorkin 1997: 147) ‘After some time passed, finally, the arrival of the bus was heard.’

(1b)

paro+ńť ťej+ema+zo avoĺ šoždïńe,

good_N+GEN.DEF.SG to-make_V+DVN+POSS-3SG.NOM.SG not_PRT-NEG-CONTRAST easy_A.NOM.SG,

śe+ń kis piťńe-ze pek pokš.

that_PRON-DEM-DISTAL+GEN for_POP.ILL price_N+POSS-3SG.NOM.SG very_AD-A big_A.NOM.SG

(Radayev 1969: 79) ‘Doing something good isn't easy, [and] for that reason its value is very high.’

Studies of the Erzya language often speak of an (elative-case) gerund derived

regularly from verb stems in much the same way as the -ms illative infinitive is. Most

recently the OmstO formative (‘to sing’ moramsto, ‘to pray’ oznomsto, ‘to come back to life’

veĺmemste) has been recognized as one of the three gerund-forming morphemes: -Oź, -OmstO

and -do, (cf. Alyoshkina, EKM 2000: 222-227), contrast (Bartens 1979). The OmstO

formative, according to Alyoshkina, is used in the deverbal derivation of clausal adjuncts,

indicating the temporal frame of a non-matrix event/action/state entity, see (2).

(2)

čï+ńť valg+omsto čokšńe+ń zoŕa+ś

sun_N+GEN.DET. SG to-set_ V+INF.ELA evening_ N+GEN redness-in-the-sky_ N+NOM.DET.SG

veśe olakad+ś

the-whole _PRON-Q to-turn-pale_ V+IND.PRETI.PRED-3SG

(Alyokshina 2000: 223 [Grigoshin]) ‘As the sun set the redness in the evening sky became entirely pale.’

As discussed above, Erzya also has a deverbal-noun formative OmA which, when

declined in the indefinite elative formative stO, has varied reflexes in deverbal-noun

formative, whereas the actual elative form is constantly represented by a back-vowel form in

sto. Contrast the deverbal noun (DVN) forms and their non-finite counter parts in table (4).

Deverbal nouns and non-finites in the elative

Gloss

DVN+ELA

DVN.ELA

INF+ELA

INF.ELA

‘to give’

maks+OmA+stO

maksomasto

maks+Om+stO

maksomsto

‘to go’

moĺ+OmA+stO

moĺemasto

moĺ+Om+stO

moĺemste

‘to scold’

śudo+OmA+stO

śudomasto

śudo+Om+stO

śudomsto

‘to suffice’

satoT+OmA+stO

satomasto

satoT+Om+stO

satomsto

‘to touch’

toka+OmA+stO

tokamosto

toka+Om+stO

tokamsto

Table 4.

Since the phonological difference between these two forms hinges upon three notions:

(i) the presence of a mid/low vowel-harmony oriented vowel; (ii) the semantic functions held

by the elative for expressing temporal space, and (iii) deverbal nouns attest to the same

argument-marking strategies as the non-finite derivation, the question arises as to what extent

the two elative-case derivations actually differ from one another. Are there any semantic or

syntactic criteria for claiming that the so-called gerund in Om+stO is anything other than an

elative-case word form that is regularly derived from verb stems, and functions as a temporal

clausal adjunct as would other elative-case nouns with temporal-reference. The distinction

between non-finite and deverbal-noun forms might, in fact, be arbitrary.

Method%

The elative gerund morpheme OmstO will be segmented into a hypothetical non-finite

formative -Om and a subsequent elative marker stO. This segmentation will set it in contrast

with that of the deverbal-noun elative affix sequence OmA+stO, and thus render the elative

form for inspection morphological, semantic and syntactic:

A) On the basis of the majority Erzya text corpus, as defined in Rueter: Adnominal

person in the morphological system of Erzya (forthcoming), a brief list of morphological hits

will be given to illustrate the relative frequency of the most widely attested elative forms of

non-finite and deverbal-noun derivations.

B) The elative forms will be inspected for compatibility with the three Erzya

declination types, indefinite, possessive and determiner, whereas determiner declination is

not attested for adverbs and adpositions.

C) The functions of the elative case most prominently the temporal function of the

OmstO adjunct will be illustrated by means of representative examples.

D) Cross-referential adnominal-person marking on the two sets of word forms will be

inspected for argument reference, whereas possessor indexing indicating anything other than

P-argument reference has previously been cited as a criterion by which to distinguish Om

non-finites from infinitives, (cf. Bartens 1979: 45).

Analysis%

In a UNIX environment we can extract all words of pertinent form by means of the

tool ‘egrep’ and the regular expression ‘м(|а|о)ст(о|э)’. This will render three types of word

forms: Hits: (i) non-finite forms in OmstO (eŕamsto ‘while living’, ťuŕemste ‘while

fighting’); (ii) deverbal nouns in OmA+sto (eŕamosto ‘from [the] life’, ťuŕemasto ‘from [the]

fight’), and misses (śeĺmste ‘from (the) eye(s)’, kaštomsto ‘out of (the) oven’). In tables (5-6)

we will observe the sheer frequency with which non-finite elative forms surpass the deverbal-

noun elative forms, on the one hand, and the fact that both word types attest to three

declination types, i.e. unlike adverbs and adpositions, word forms with the Om+stO

formative can take definite singular declination, which indicates a notion of reference,

something associated with nouns. The hits column group in table (5) has been divided into

indefinite declination, possessive declination and definite declination to provide the reader an

idea of co-occurrence frequency for each of the declination types, e.g. ‘to go’ moĺemste;

moĺemsteń, moĺemsteť, moĺemstenze, moĺemsteńek, moĺemsteŋk, moĺemstest; moĺemsteńť.

21 most frequent non-finite elative Om+stO in Erzya majority corpus

Gloss

Word form

Hits

Indef

Def

Total

‘to go’

moĺ+emste

688

775

‘to pass’

juta+msto

608

682

‘to depart;

to bring’

tuj+emste

406

537

‘to arrive’

sa+msto

316

367

‘to come/go out’

ĺiś+emste

251

293

‘to watch’

van+omsto

254

269

‘to sleep’

udo+msto

221

239

‘to live’

eŕa+msto

205

254

‘to enter’

sova+msto

195

227

‘to talk’

korta+msto

201

222

‘to ride’

ard+omsto

195

206

‘to stay’

ašťe+mste

185

200

‘to study’

tonavtńe+mste

184

193

‘to make’

ťej+emste

126

136

‘to walk; to visit’

jaka+msto

106

115

‘to fight’

ťuŕ+emste

111

114

‘to meet’

vast+omsto

101

‘to fall’

pra+msto

‘to think’

aŕśe+mste

‘to finish’

pŕad+omsto

‘to tell (in detail)’

jovtńe+mste

Total

4533

372

104

5177

Table 5.

The minimal attestation of co-occurrence with definite-declination forms in table (5)

is only slightly smaller than that of co-occurrence with second-person possessor indexing,

such that 2SG shares the same level of attestation as the definite-declination forms. First

person possessor indexing comprises over twice the attestations found for that of second

person, and the attestation for third person possessor indexing is again approximately five

times of that attributed to first person possessor indexing. All in all, however, attestation of

deictic declination (possessive and definite combined) comprises only about one tenth of all

non-finite elative word forms. While both OmstO and OmA+stO word types prefer indefinite

elative declination forms, it will be observed that definite marking exceeds that of possessor

indexing in the OmA+stO deverbal noun forms, where indefinite and definite forms

contribute to the bulk of all attestations, see table (6).

21 most frequent deverbal nouns in elative OmA+stO from Erzya majority corpus

Gloss

Hits

Indef

Det

Total

Word form

1SG

1PL

2SG

2PL

3SG

3PL

‘to live’

eŕamosto

138

260

‘to arrive’

samosto

‘to tell (in detail)’

jovtńemasto

‘to begin’

ušodomasto

‘to talk’

kortamosto

‘to think’

aŕśemasto

‘to stand up’

śťamosto

‘to meet’

vastomasto

‘to meet (recipr)’

vastovomasto

‘to sleep’

udomasto

‘to strive’

bažamosto

‘to fall’

pramosto

‘to fight’

ťuŕemasto

‘to finish’

pŕadomasto

‘to write’

śormadomasto

‘to love’

večkemasto

‘to sit down’

ozamosto

‘to lance’

śalgomasto

‘to walk; to visit’

jakamosto

‘to pray’

oznomasto

‘to turn’

purdamosto

Total

237

184

467

Table 6.

Of the 42 verb forms shown in tables (5-6) only 11 are consistently representative of

frequent non-finites and deverbal nouns. These eleven do not include the three most frequent

verbs of the non-finite forms, which also means that the verbs jutams ‘to pass’ and tujems ‘to

leave’ with complete possessive paradigms are not represented. The verbs represented in

table (7) have deverbal nouns that can be characterized as [+COUNT], but they are by no

means consistent in their other semantic characteristics, i.e. some indicate events (samo

‘arrival’, ťuŕema ‘fight’) while others might be seen as the product of the action (kortamo

‘talk, [what was said]’, aŕśema ‘thought’) indicated by the verb.

Verbs with high frequency non-finite versus deverbal noun dichotomies

Gloss

Non-finite ELA

Attestation

Gloss

Deverbal ELA

Attestation

Total

‘to live’

eŕa+msto

254

‘life’

eŕa+mosto

260

514

‘to arrive’

sa+msto

367

‘arrival’

sa+mosto

418

‘to sleep’

udo+msto

239

‘sleep’

udo+masto

247

‘to talk’

korta+msto

222

‘talk’

korta+mosto

235

‘to walk; to visit’

jaka+msto

115

‘visit’

jaka+mosto

118

‘to fight’

ťuŕ+emste

114

‘fight’

ťuŕ+emasto

120

‘to meet’

vast+omsto

101

‘encounter,

meeting’

vast+omasto

112

‘to think’

aŕśe+mste

‘thought’

aŕśe+masto

‘to fall’

pra+msto

‘fall’

pra+mosto

‘to tell (in detail)’

jovtńe+mste

‘story’

jovtńe+masto

‘to finish’

pŕad+omsto

‘end’

pŕad+omasto

Total

1659

398

2057

Table 7.

Typical misses include elative-case nouns, such as, those found in table (8).

Common misses generated by ‘м(|а|о)ст(о|э)’ expression

Word form

Elative-case gloss

Word form

Elative-case gloss

śeĺm+ste

‘eye’

kaštom+sto

‘oven’

čama+sto

‘face’

vaĺma+sto

‘window’

eźem+ste

‘bench’

utom+sto

‘grainery’

ťuŕma+sto

‘jail’

rajkom+sto

‘raion committee’

ferma+sto

‘farm’

pojma+sto

‘flood plain’

Table 8.

The elative case represented by the formative -stO has two basic meanings – source

and location. While source is primarily associated with notions of space (3), material (4),

spatio-temporal starting point (5) and separation (6), capacity (7) and temporal setting(8)

appear to convey the meaning of location.

(3)

kudo+sto+ńť

house_N+ELA+DET.SG

‘out of the house’

(4)

śija+sto

silver_N+ELA

‘out of silver’

(5a)

veĺe+ste veĺe+s

village_N+ELA village_N+ILL

‘from village to village’

(5b)

śokśe+ste tundo+s

autumn_N+ELA spring_N+ILL

‘from autumn to spring’

(6)

ava+sto+nzo

mother_N+ELA+POSS-3SG

‘from its/his/her mother’

(7)

pŕavt+sto

boss_N+ELA

‘from the position of boss’, ‘in the position of boss’

(8)

eŕva čï+ste

every_Q.ABS day_N+ELA

‘every day’

The notion of temporal setting conveyed in (8) is, in fact, parallel to that observed in

(2), above. Temporality, it would seem, is determined by the referent itself, i.e. the word čï

‘day’ is a quantity of time, a duration. Since the adjunct in (2) čï+ńť valg+om+sto ‘as the

sun set’ indicates an activity simultaneous to that of the main predicate verb, one might

readily draw a parallel between these two temporal entities. Speaking of temporal entities,

naturally, brings us back to deverbal nouns, which should also bear temporal meaning. The

verb pŕadoms ‘to finish, to complete’ provides us with a near minimal pair in pŕadomsto (the

non-finite) and pŕadomasto (the deverbal noun). In two examples from the same collection

of short stories, we can observe both word forms in the same function. In (9) a notion of

durative temporal space might be entertained as a setting for a meeting (vastńems ‘[to

happen] to meet; to meet repeatedly’), indicating that the non-finite form conveys a notion of

continuous or ongoing action.

(9)

mejeĺśe+de sïń vastńe+śť kize+ń

last_PRON-DET+ABL they_PRON-PERS-3PL.NOM to-meet_V+IND.PRETI.PRED-3PL summer_N+GEN

pŕad+om+sto+ńť, źardo son

to-end_V+INF+ELA+DET.SG when_PRON-REL.ABL he_PRON-PERS-3SG.NOM

uskśe+ś gruzovoj avtomašina+so stanćija+v śuro

to-haul_V+IND.PRETI.PRED-3SG freight_A.ABS automobile_N+INE station_N+LAT grain_N.NOM.SG

(Abramov 1974: 85) ‘The last time they had met at the end of summer when he was hauling grain to the

station in a dump-truck.’

The deverbal noun in (10), on the other hand, could be construed to indicate a point in

time, which would lead to a dichotomy, duration versus reference point. This dichotomy,

however, could also be attributed to the time frames implicated by the predicate verbs; the

verb vastńems ‘to meet’ in (9) entails a durative time frame, whereas the verb sams ‘to arrive’

in (10) indicates a point in time.

(10)

juta+ź ťeĺe+ń pŕad+oma+sto+ńť veĺe+ńťeń

to-pass_V+PTC-Z winter_N+GEN to-end_V+DV-N+ELA+DET.SG village_N+DAT.DET.SG

sa+ś geolog+oń apokške gruppa –

to-arrive_V+IND.PRETI.PRED-3SG geologist_N+GEN small_A.ABS group_N.NOM.SG --

kolmo ćora+t dï vejke ava.

three_NUM-CARD.ABS man_N+PL.NOM and_CONJ one_NUM-CARD.ABS woman_N.NOM.SG

(Abramov 1974: 26) ‘At the end of last winter, a group of geologists came to the village – three men and

one woman.’

Another verb kortams ‘to talk’ also provides us with examples of non-finite versus

deverbal noun variation in kortamsto and kortamosto respectively. Here the non-finite or

short form is consistently associated with the speech act, whereas the longer or deverbal noun

indicates the product, see (11-12).

(11)

korta+m+sto+ńť pšťi+ste van+ś,

to-talk_V+INF+ELA+DET.SG sharp_A+ELA to-look_V+IND.PRETI.PRED-3SG.

tonavtńe-ś, meźe-ńť kona

to-study_V+IND.PRETI.PRED-3SG, what_PRO-N-INTER+GEN.DET.SG which_PRO-DET-INTER.ABS

keď-se dï koda kund-ï

hand_N+INE and_CONJ how_PRO-ADV-INTER to-hold_V+IND.PRES.PRED-3SG

sonze večkeviks+eś jarsa+m+sto+nzo.

he_PRO-PERS-3SG.GEN.POSS-3SG beloved_N+NOM.DET.SG to-eat_V+INF+ELA+POSS-3SG

(Kutorkin 1987: 253) ‘While talking he [Mikoĺ] watched attentively, studied, what, how and with which

hand his beloved held onto things when she ate.’

(12)

no kiŕillov+ńeń korta+mo+sto+ńť vejke+jak val

but_CONJ Kirillov_PRP+DAT to-talk_V+N+ELA+DET.SG one_NUM-CARD.ABS+CLT word_N.NOM.SG

a čaŕkoďev+i.

not_PRT-NEG to-understand_V+IND.PRES-PRED-3SG

(Shcheglov 1968: 29) ‘But Kirillov cannot comprehend a single word from what is said.’

Although activity versus product might be attested in some instances, there appears to

be free variation in some publications between short and long forms. It seems that the short

forms cannot be used for expression of product entities. If, however, the short form indicates

a duration and not a point in time, then it can be attested as with verbs, such as, sïrgoźems ‘to

wake up’, see (13). Hence it is conceivable that the short Om+stO form can be associated

with source functions, as well.

(13)

kaštom+ost udal+o ćiŕk meŕev+ś

oven_N+POSS-3PL behind_POP+LOC chirp_IDEOPH to-suddenly-say_V+IND.PRETI.PRED-3SG

ťeĺe+ń udo+m+sto+nzo sïrgoź+ića ćiŕkun.

winter_N+GEN to-sleep_V+INF+ELA+POSS-3SG to-awaken_V+PTC-CA cricket_N.NOM.SG

(Kutorkin 1976: 236) ‘Behind the stove a cricket awakening from its winter sleep gave a chirp.’

Cross-referential adnominal-person marking, possessor indexing, is, as illustrated in

tables (5-6), compatible with both non-finite and deverbal-noun forms. Possessor indexing,

in Erzya, is primary-argument oriented, i.e. with the non-finite elative forms of intransitive

verbs the personal affix indicates the S argument possessor, but with transitive verbs there are

further stipulations to be dealt with, see (14-16).

(14)

Baška-baška ťev+ť+ńe+ste, vejke-vejke

separate_ADV-separate_ADV thing_N+PL+DET.PL+ELA, one_NUM-CARD.ABS-one_NUM-CARD.ABS

meĺga moĺe+ma+sto+st maŕav+i

after_POP to-go_V+N+ELA+POSS-3PL to-be-heard_V+IND.PRES.PRED-3SG

son+ś+enze ška+ńť ikeĺe+v

it_PRO-PERS-3SG+REFL+POSS-3SG.OBL time_N+GEN.DET.SG forward_ADV+LAT

šašto+ma+ś, koso, meźe dï

to-move_V+N+NOM.SG.DET, where_PRO-SPAT-INTER+INE, what_PRO-N-INTER.NOM.SG and_CONJ

źardo moĺ+i, ťe+ń laŋg+s

when_PRO-ADV-TEMP.ABL to-go_V+IND.PRES.PRED-3SG, this_PRO-DET+GEN on_POP+ILL

jav+i baška meĺ avtor+oś…

to-share_V+IND.PRES-PRED-3SG separate_ADV thought_N.NOM.SG author_N.NOM.SG.DET…

(Gobunov 1993: 14) ‘From separate things, from their succession you can feel the progression of time

itself, where, what and when it happens, this is something the author addresses explicitly.’

(15)

iŕina pavlovna čaŕkoď+ś:

Irina_PRP.NOM.SG Pavlovna_PRP.NOM.SG to-understand_V+IND.PRETI.PRED-3SG:

ťeťa+zo dï ava+zo, nať, lamo+kśť

father_N+POSS-3SG.NOM.SG and_CONJ mother_N+POSS-3SG.NOM.SG, apparently_ADV, many_Q+ITER

meŕńe+śť vejke+ńeń-vejke+ńeń

to-say_V+IND.PRETI.PRED-3PL one_NUM-CARD+DAT-one_NUM-CARD+DAT

maks+om+sto+nzo.

to-give_V+INF+ELA+POSS-3SG

(Kutorkin 1987: 105) ‘Irina Pavlovna understood: Apparently, his [Dina] father and mother had said

[that] several times when they gave/passed him [Dina] to each other’

While (14-15) illustrate the indication of primary-argument relations of S and P, (16)

would appear to indicate an A relation for the possessor-index, in as far as the verb ećems ‘to

stuff’ is understood as a transitive verb with the patient/theme indicated by the genitive-form

attribute peŋge ‘firewood’. In (17) we can observe yet another instance of an A relation for

possessor-indexing.

(16)

toń beďńak+t+ne+ń meźejak a

you_PRO-PERS-2SG.GEN pauper_N+PL+DET.PL+GEN anything_PRO-N-IND.NOM.SG not_PRT-NEG

mu+sïńźe, ― karma+ś muzgoŕď+em+e

to-find_V+IND.PRES.PRED3SG>3PL, -- to-begin_V+IND.PRETI.PRED-3SG to-mutter_V+INF+LOC

kaštom+s peŋge+ń eće+m+ste+nze matŕa.

oven_N+ILL firewood_N+GEN to-stuff-full_V+INF+ELA+POSS-3SG Matrya_PRP.NOM.SG

(Kutorkin 1976: 343) ‘There is nothing that will get to those paupers of yours [is there?] – Matrya began

to mutter as she stoked the oven with firewood.’

(17)

― ťe paro.

that_PRO-DET.NOM.SG good_A.NOM.SG.

― meźe+ś? ― kevkśť+iźe, śeĺme+nze

what_PRO-N-INTER+NOM.DET.SG ? to-ask_V+IND.PRETI.PRED-3SG>3SG, eye_N+POSS-3SG.OBL

końa+m+sto+nzo, Merkulovna.

to-close:eyes_V+INF+ELA+POSS-3SG, Merkulovna_PRP.NOM.SG

(Kutorkin 1976: 22) ‘That [is] good. – What is? Merkulovna asked him as she closed her eyes.’

No instances of deverbal nouns (OmA+sto) were detected in the Erzya majority

corpus that were both (i) derived from transitive stems, and (ii) had possessor-indexing

indicating a P-argument relation.

Results

Non-finite and deverbal forms in the elative are attested with 3 declination types

(indefinite, possessive and definite), an indication that both forms refer to entities, and are not

mere relation-words such as adverbs and adpositions might be conceived to be, see tables (5-

6) for specific data.

The Elative functions of temporal location and source can be discerned for the short

form Om+stO, which indicates that these relation words are durative in nature, and might be

distinguished from point-in-time referent deverbal nouns OmA+stO, see (9-11, 13).

Possessor indexing on non-finite targets is associated with the primary-argument

relations S with verbs[-TRANS] (13-14), P with verbs[+TRANS] (15) when no competing

indication of possessor is present, e.g. a genitive attribute may render an A reading of

possessor indices when a genitive-form attribute/P is present, (16-17).

Abbreviations:

A = Agent, ABL = Ablative, ABS = Absolutive, DAT = Dative, DEM = Demonstrative,

DEF = Definite, DVN = Deverbal noun, ELA = Elative, GEN = Genitive, ILL =

Illative, IND = Indicative, INDEF = Indefinite, INF = Infinitive, INTER =

Interrogative, INE = Inessive, LAT = Lative, LOC = Locative, NA = Not attested,

NOM = Nominative, P = Patient, PL = Plural, POSS = Possessor Index, PRED =

Predication Marker, PRES = Present, PRETI = Preterit I, PRETII = Preterit II,

PRO-A = Pro-adjective, PRO-ADV = Pro-adverb, PRO-DET = Pro-determiner, PROL =

Prolative, PRO-N = Pro-noun, PRO-Q = Pro-quantifier, Q = Quantifier, S =

Subject, SG = Singular, TRANS = Transitive, TRANSL = Translative

Sources

Abramov, Kuz'ma 1974 = Абрамов, Кузьма 1974:

Нурька морот.