ChapterPDF Available

The general noun-modifying clause construction in Tundra Nenets and its possible origin

January 2017

In book: Noun-Modifying Clause Constructions in Languages of Eurasia: Reshaping theoretical and geographical boundaries. (pp.147-178)
Publisher: John Benjamins
Editors: Matsumoto, Comrie, Sells

Authors:

Tundra Nenets (Uralic) exhibits unambiguous relative clauses and sentential complements of nouns, but I show that it also has a previously unstudied but structurally distinct GNMCC. The GNMCC covers a diversity of functions although its usage is restricted in various ways. The paper suggests that it has a direct parallel in the non-sentential domain in terms of its syntactic behaviour , the morphosyntactic expression of the constructional ingredients and the basic semantics: it is modelled after the non-clausal compound-like structure employed for the very general purpose of modifying one noun by reference to another noun. Such modification-by-noun constructions served as a historical source of Tundra Nenets GNMCCs, which emerged when the modifying dever-bal noun was reanalysed as heading a clausal domain.

Content uploaded by Irina Nikolaeva

Content may be subject to copyright.

is is a contribution from Noun-Modifying Clause Constructions in Languages of Eurasia.

Rethinking theoretical and geographical boundaries.

Edited by Yoshiko Matsumoto, Bernard Comrie and Peter Sells.

is electronic le may not be altered in any way.

e author(s) of this article is/are permitted to use this PDF le to generate printed copies to

be used by way of oprints, for their personal use only.

Permission is granted by the publishers to post this le on a closed server which is accessible

to members (students and sta) only of the author’s/s’ institute, it is not permitted to post

this PDF on the open internet.

For any other use of this material prior written permission should be obtained from the

publishers or through the Copyright Clearance Center (for USA: www.copyright.com).

Please contact rights@benjamins.nl or consult our website: www.benjamins.com

Tables of Contents, abstracts and guidelines are available at www.benjamins.com

John Benjamins Publishing Company

 ./tsl..nik

 

e general noun-modifying clause

construction in Tundra Nenets

and its possible origin

Irina Nikolaeva

School of Oriental and African Studies, University of London

Tundra Nenets (Uralic) exhibits unambiguous relative clauses and sentential

complements of nouns, but I show that it also has a previously unstudied but

structurally distinct GNMCC. e GNMCC covers a diversity of functions

although its usage is restricted in various ways. e paper suggests that it has

a direct parallel in the non-sentential domain in terms of its syntactic behav-

iour, the morphosyntactic expression of the constructional ingredients and the

basic semantics: it is modelled aer the non-clausal compound-like structure

employed for the very general purpose of modifying one noun by reference to

another noun. Such modication-by-noun constructions served as a historical

source of Tundra Nenets GNMCCs, which emerged when the modifying dever-

bal noun was reanalysed as heading a clausal domain.

Keywords: noun-noun compounds, modication, reanalysis, deverbal nominals

1. Introduction

is chapter examines NMCCs in Tundra Nenets, a language that belongs to

the Samoyedic branch of the Uralic language family. Apart from Tundra Nenets,

the Samoyedic branch includes Forest Nenets, Tundra Enets, Forest Enets, and

Nganasan (these are usually classied as Northern Samoyedic languages), as well

as the only living representative of the Southern Samoyedic sub-branch, Selkup,

which consists of several dialects. Tundra Nenets is spoken in the Arctic part of

European Russia and north-western Siberia between the Kanin Peninsula in the

west and the Yenisei river delta in the east. Administratively the Tundra Nenets

territory presently includes the whole Nenets District in Arkhangelsk Province,

parts of some northernmost regions in the Komi Republic, several regions of the

148 Irina Nikolaeva

Yamal-Nenets District in Tyumen Province, and most of the Ust-Yeniseisk region

of the Taimyr District in the Krasnoyarsk Region.

e ocial population gure for the Nenets people was 44,610 in the 2010

census, with 21,926 as the actual number of speakers. ese numbers include both

the Tundra and the Forest Nenets, since the ocial trend in Russia is to view them

as a single language. According to Ethnologue (Lewis 2009), there may be about

1,300 speakers of Forest Nenets, so Tundra Nenets may be currently spoken by

about 20,500 people. However, native language prociency varies from one area

to another and the estimates are not always accurate. e number of uent native

speakers is decreasing because the continuous shi to Russian proceeds faster than

the natural growth of the population. e language is highly endangered.

Unless indicated otherwise, the Tundra Nenets data cited in this chapter come

from my eldwork supported by an ELDP grant awarded to Tapani Salminen and

a grant from the Academy of Finland awarded to Larisa Leisiö, project number

125225. e transcription is based on Nikolaeva (2014). I also use other sources,

most importantly, Tereščenko (1965), abbreviated here as T. In Section2 I present

the basic grammatical characteristics of the language paying particular attention

to the structure of s and the dependent verbal forms. Sections 3 and 4 introduce

two frequent types of NMCCs that can be unambiguously characterized as rela-

tive clauses and sentential noun complements, respectively. However, in Section5

I argue that Tundra Nenets additionally has a GNMCC that covers a diversity of

functions, even though its usage is quite heavily restricted by both semantic and

possibly syntactic factors in comparison to other languages which have a similar

construction. Section6 argues that the Nenets GNMCC is modelled aer a non-

sentential construction in which a noun modies another noun, and some kind of

general semantically unspecied relation obtains between the two. In Section7 I

further speculate about the possible origin of the GNMCC in Tundra Nenets and

provide a concluding discussion.

2. e basic properties of the language

As described in more detail in Nikolaeva (2014), in Tundra Nenets the major open

word classes (nouns, verbs, adjectives and adverbs) can be distinguished by mor-

phological and syntactic criteria. Nouns have the following inectional categories:

number, case and possessive. ere are three numbers (singular, dual and plu-

ral), and seven cases (nominative, accusative, genitive, dative, locative, ablative

and prolative). More specialized local meanings are expressed by postpositional

phrases. e nominative case and singular number are formally unmarked and

therefore I do not indicate them in glosses. e possessive category realizes the

Chapter 7. e general noun-modifying clause construction in Tundra Nenets 149

person/number of the possessor by means of person/number axes attached to

the head nominal in the possessive construction.

e verbal inectional categories are agreement, tense and mood. Subject

agreement in person/number is obligatory, whereas object agreement is only in

number and ‘optional’ in the sense that only objects with particular semantic and

information structure-related features trigger agreement. e present tense is

formally unmarked (and therefore not indicated in glosses), while inectionally

expressed tenses are the past, future, future-in-the-past and habitual. All verbs

are lexically divided into two aspectual classes, perfective and imperfective, which

largely determine the meaning of the tenses. Imperfective verbs denote present

situations in the present and past situations in the past, while perfective verbs de-

note immediate past situations in the present and remote past situations in the

past. Nenets exhibits 15 inectional moods including the imperative, hortative,

optative, conjunctive, necessitative, interrogative, probabilitative, obligative, po-

tential, evidential etc., all of which are formed by suxation. e category of tense

is marginal in non indicative moods. As in many Uralic languages, negation is

periphrastic and formed by means of an inected nite negative verb and a non

nite connegative form of a lexical verb.

Syntactically, Tundra Nenets is a head nal SOV language. Verb nality is fair-

ly rigid; matrix clauses, therefore, are consonant with the head nality evident in

noun phrases, postpositional phrases and non nite dependent clauses, but the or-

der of non verbal constituents in the matrix clause is largely motivated by informa-

tion structure. Major grammatical functions, such as the subject and direct object

are rather easily identied by a cluster of grammatical properties. In particular, the

subject is a grammatical element associated with a signicant number of syntactic

behaviours such as being the main controller of coreferential relations, both with-

in the clause and clause externally. e case marking patterns of core arguments

follow accusative alignment. As in many other languages, independent pronouns

tend to be omitted unless specic discourse considerations obtain, while posses-

sive and predicate-argument agreement morphology, although referred to here as

‘agreement’ for convenience, may be construable as incorporated pronouns.

Importantly for the present topic, we can distinguish three main types of ad-

nominal dependent (with the exclusion of determiners and quantiers, which will

not be discussed here). First, the language has possessive constructions. In the

possessive construction lexical possessors stand in the genitive and do not nor-

mally trigger possessive agreement on the head: Wera-h ti (Wera- reindeer)

‘Wera’s reindeer’. However, when the lexical possessor is a local topic or somehow

pragmatically prominent for the interpretation of the whole , it can be cross-

referenced by the 3rd person possessive marker on the head, while remaining

-internal: Wera-h te-da (Wera- reindeer-3) ‘Wera’s reindeer’. Pronominal

150 Irina Nikolaeva

possessors must be expressed by the person/number agreement morphology on

the head and an independent pronoun is optional: (mən´°) te-m´i (1 reindeer-

1) ‘my reindeer’. ese constructions have typical possessive meanings such as

part-whole (xəbew°ko-h yík°c´a ‘partridge’s neck’), kinship and other relations be-

tween humans (teta-h puxac´a ‘the master’s wife’), legal ownership (xada-h m´aq

‘grandmother’s yurt’), activities and their results (n´erka-h wadir ‘growing of the

willow tree’), intrinsic properties such as size, smell and shape (xal´a-h ŋəbt° ‘smell

of sh’), and so on. However, it is well known that the range of (alienable) posses-

sive meanings can be extended indenitely, subject to the world knowledge and

belief system of a particular linguistic community (Seiler 1983; Taylor 1996) and

to contextual factors (e.g. Hawkins 1981; Sperber & Wilson 1986; Barker 1995;

Partee & Borschev 1998). Barker (2011) refers to such unspecied possessive in-

terpretations as ‘pragmatic’. As has been shown in detail in Nikolaeva (2003) and

Ackerman & Nikolaeva (2013), in Tundra Nenets possessive constructions are

characterized by high frequency and convey many pragmatic interpretations. at

is, the possessive construction has a virtually innite range of meanings, as long as

the relation between the possessor and the possessed is construed as semantically

and pragmatically plausible.

e second adnominal construction, attributive modication, typically in-

volves words of the adjective lexical class which denote some gradable proper-

ty concept and are one-place predicates whose only argument is identied with

the head noun. Canonical modication is usually treated as set intersection

(Higginbotham 1985; Larson & Segal 1995, and many other references): the de-

notation of the adjective identies a certain subset in the denotation of the head

noun by specifying which instantiation of the class is meant. Since modication

narrows the concept associated with the head noun, it serves the purpose of clas-

sifying the respective entity.

In Tundra Nenets attributive adjectives show concord in number (obligatory,

at least in certain dialectal varieties), as well as case and possessive person/number

(highly optional and very infrequent, typically restricted to particular registers):

(1)

serako-q

white-

te-q

reindeer-

‘white reindeer ()’

serako-xədənt°

white-.2

te-xədənt°

reindeer-.2

‘from your white reindeer’

is is a true instance of attributive concord between subconstituents of the same

, which is only present if the relevant features are realized on its head (see

Nikolaeva 2005). Adjectives can be extracted out of the host , in which case they

Chapter 7. e general noun-modifying clause construction in Tundra Nenets 151

arguably form their own  that can (though does not have to) be separated from

the semantic head by other constituents; they must bear relevant case, number and

possessive inections in this instance:

(2)

wen´ako-x°q

dog-.

(serako-x°q)

white-.

si°rŋa-d°m

look-1

(serako-x°q)

white-.

‘I am looking at the dogs, the white ones.’

In (2) the adjective serako ‘white’ forms its own , either following its semantic

head ‘dogs’ and separated by a pause from it, or postposed aer the verb; in both

instances it bears the dative plural inection, reecting a relation with the seman-

tic head. e function of such constructions is to indicate some kind of aer-

thought. Moreover, the semantic head can be completely absent, in which case the

adjective itself assumes the argument status and bears the respective morphology.

(3)

serako-x°q

white-.

si°rŋa-d°m

look-1

‘I am looking at the white ones.’

Note that possessors can also undergo extractions, but do not occur in construc-

tions analogous to (3), that is, the possessor cannot represent the whole phrase.

Finally, Tundra Nenets exhibits a special construction type in which a noun

serves to modify another noun as a kind of attributive modier without turn-

ing into a full-edged adjective. is construction type deserves more discus-

sion because it is directly relevant for the topic of the paper and will be addressed

in Section6.

Turning now to dependent clauses, there are virtually no (native) complemen-

tizers in the language. Subordinate clauses make use of non nite verbal forms,

which are primarily or exclusively used in dependent contexts. ere are essen-

tially three types of non-nite verb: (i) participles, (ii) clausal nominalizations

which are termed ‘innitives’ in Salminen (1997), but will be referred to here as

‘action nominals’ aer Nikolaeva (2014), and (iii) converbs. All of them represent

so-called ‘mixed categories’ in the sense that they exhibit properties of verbs and

one other part of speech (adjectives, nouns or adverbs, respectively). Participles

are attributive forms mostly used for relativization; they will be addressed in more

detail in Section3. e main function of action nominals is complementation

and adverbial clauses, and in these functions they can take various grammatical

cases, e.g.:

(4)

sira-h

snow-

xolka-qma-xəd°

melt-.-

earth

yilʹeŋkə°

revive.3

‘Aer the snow melts, the earth revives.’ (T 142)

152 Irina Nikolaeva

yukon°h

here

to-wa-mt°

come-.-.2

ŋətʹe-lŋa

wait-.3

‘He started waiting for you to come here.’ (T 389)

In (4a) the action nominal is in the ablative and denes an adverbial temporal

clause, while in (4b) it stands in the accusative and heads a complement clause.

But, as will be shown below, action nominals have a rather general meaning and

participate in GNMCCs. Converbs are not compatible with case markers and pre-

dominantly serve in adverbial subordination, although some converbs are also

used in other functions. One example of the conditional converb is shown below.

(5)

xæ-b°qna-nt°

go-.-2

mənʹ°

1

ŋod°q

too

xan°-tə-d°m

go--1

‘If you go, I will go too.’ (T 396)

e clausal status of syntactic domains dened by non-nite verbs is evident from

the fact that they typically express tense oppositions, are syntactically opaque and

preserve the argument structure of the base verb (see Ackerman & Nikolaeva 2013

for more detail). Most grammatical functions in non-nite clauses are realized in

the same way as in main/independent clauses, except for the subject. e subject

exhibits all relevant behavioural properties but essentially takes the same form as

the possessor in possessive constructions. at is, pronominal subjects are oen

dropped but must be represented by person/number markers selected from the

possessive paradigm and oen cumulated with case marking, as is evident in ex-

ample (4b) above. I will refer to them as ‘subject agreement’, although these mark-

ers may have pronominal force. eir host diers depending on the construction:

in (4) and (5) subject agreement is hosted by the dependent non-nite verb, while

in other constructions, as will be explained below, it is located on the head nomi-

nal. Dependent lexical subjects stand in the genitive case and do not normally

trigger agreement, but in some instances a discourse-salient lexical subject may be

cross-referenced by the 3rd person possessive ax either on the dependent verb or

the head nominal. Again, this pattern is reminiscent of the marking of the lexical

possessor in possessive constructions.

3. Relative clauses

Relative clauses are headed by participles. ere are four types of participles in

Tundra Nenets: imperfective participles in -t(ʹ)a/-n(ʹ)a (with phonological vari-

ants), perfective participles in -miə, future participles in -mənta, and negative par-

ticiples in -mədaweyə. e imperfective, perfective and future participles denote

Chapter 7. e general noun-modifying clause construction in Tundra Nenets 153

events which, respectively, are simultaneous with, precede or follow the event

denoted by the main clause in time. e negative participle expresses negation

in the past.

e participial strategy is employed for the relativization of subjects and accu-

sative direct objects and involves gapping. at is, the relativized argument is not

overtly represented in the relative clause. Relativized subjects are extracted from

the relative clauses and are not cross-referenced by agreement:

(6)

ŋəcʹeki°

child..

toxolamp°-da

teach-.

nʹe

woman

‘the woman who teaches children’

When a direct object is relativized, the relative clause may contain its own subject.

As mentioned above, the morphosyntactic expression of the dependent subject is

identical to the expression of the possessor in the possessive construction: the lexi-

cal subject stands in the genitive and may trigger optional 3rd person agreement,

while the pronominal subject corresponds to pronominal person/number axes

and may be additionally expressed by a free-standing pronoun. Interestingly, in

participial relative clauses subject agreement morphology is hosted by the relativ-

ized head noun. at is, from a descriptive perspective, agreement is not located

on the word whose argument requirement it appears to satisfy (the non-nite

verb) and is external to the domain dened by the relative clause itself, resulting in

some kind of locality violation. For example, in (7a) the 1 marker on the head

noun ‘dog’ references the pronominal subject of the relative clause ‘I’, while in (7b)

the lexical subject stands in the genitive and triggers optional 3 agreement on

the head. Clause boundaries are shown with brackets.

(7)

[(mənʹ°

1

tʹenʹana

yesterday

ŋəw°la-w°dawey°]

feed-.

wenʹako-mʹi

dog-1

‘the dog which I didn’t feed yesterday’

[Petʹa h

Petya 

ŋəw°la-w°dawey°]

feed-.

wenʹako /

dog /

wenʹako da

dog 3

‘the dog which Petya didn’t feed’

is rather unusual pattern of subject marking in relative clauses is found in a

number of other languages spoken in Central and Eastern Eurasia, including some

Turkic, Tungusic and Mongolic languages (see Ackerman & Nikolaeva 2013 for an

analysis and a typological overview).

Participles also participate in the relativization of the possessor, but only if the

possessed noun functions as a relative clause subject. Unlike other types of depen-

dent lexical subject, in such constructions the subject can optionally stand in the

nominative. However, possessor relativization involves a dierent recoverability

154 Irina Nikolaeva

strategy, namely, resumption: the relativized possessor is represented by the 3rd

person possessive ax hosted by the possessed noun. In (8) the possessed noun

‘knife’ functions as the relative clause subject and may stand either in the nomi-

native or genitive, but in both instances it must bear the 3rd person resumptive

pronominal.

(8)

[xər°-da

knife-3 /

xər°-nta

knife-.3

məlʹ°-wi°]

break-.

xasawa

man

‘the man whose knife broke’

In terms of their position, participial relative clauses behave very much like regular

adjectives: they do not have to be adjacent to the head. Incontestable adjectives can

be placed between the relative clause and the head.

(9)

[(pida)

3

ta-wi°]

give-.

səwa

good

te-da

reindeer-3

‘the good reindeer he/she gave’

e linear sequence in which an adjective precedes the modifying relative phrase

is ungrammatical.

(10)

*səwa

good

[(pida)

3

ta-wi°]

give-.

te-da

reindeer-3

‘the good reindeer he/she gave’

Moreover, like adjectives, participles can stand in apposition to the head (11a) or

form semantically null-headed phrases where the participle itself assumes some

head properties and takes case and number reecting the role of the missing se-

mantic head, e.g. the plural and the accusative in (11b).

(11)

t´uku°

this

reindeer

səwa,

good

[pidər°

2

xada°-we-r°]

kill-.-2

‘is reindeer is good, the one you killed.’

[pidər°

2

wol°tampə-wimt°]

dislike-....2

t´en´ewəə-d°m

remember-1

‘I remember the ones you disliked.’

e only positional dierence between participles and adjectives is that the rela-

tive clause must precede the possessor to the head, as in (12a), while adjectives

follow the possessor (12b).

(12)

[Wera-h

Wera-

xada-wi°]

kill-.

(mənʹ°)

1

te-w°

reindeer-1

‘my reindeer killed by Wera’

Chapter 7. e general noun-modifying clause construction in Tundra Nenets 155

(mənʹ°)

1

serako

white

te-w°

reindeer-1

‘my white reindeer’

us, the rough linear structure of the extended  in Tundra Nenets can be repre-

sented as follows: relative clause – possessor – attributive modier – head.

With respect to their concord properties, participial relatives are just like ad-

jectives too, since they allow optional concord in number, case and person/num-

ber, although case concord is rather infrequent. ese features are illustrated be-

low in various combinations.

(13)

[xənako-x°nan°

sledge-.1

pod´erpə-da-q]

tie-.-

weno-n°

dog..-1

‘the dogs () I harness to my sledge.’

[narey°-h

spring-

s´erta-wi°-naq]

make-.-1

ŋəw°rə-naq

food-.1

‘the food () which we prepared in spring’

[yəda-we-waq

shoot-.-1

noxa-naq]

Arctic.fox-.1

pūmna

behind

‘behind the Arctic fox we shot’

[wol°tampə-we-mt°]

dislike-.-.2

xoba-mt°

skin-.2

‘the skin () that you disliked’

In sum, participles are clearly attributive structures: distributionally and in terms

of their concord properties they behave very much like regular adjectives, al-

though there is a minor positional dierence too. e participial strategy is the

primary (and perhaps the only, depending on the analysis of GNMCCs, see be-

low) dedicated relativization strategy in Tundra Nenets. It is highly constrained by

grammatical functions. In accordance with the well-known typological tendency,

it targets the highest positions on the hierarchy of grammatical functions, the sub-

ject and the direct object, as well as possessors, although relativization of possess-

ors requires a dierent recoverability strategy.

4. Noun complementation

Sentential complements are headed by dierent verbal forms which I referred to as

‘action nominals’. ese forms express sentential arguments of complement-tak-

ing verbs, as exemplied in (4b) above. ere are two types of action nominal: the

imperfective action nominal in -m(ʹ)a expresses the relative present and future,

whereas the perfective action nominal in -(o)qm(ʹ)a expresses the relative past.

156 Irina Nikolaeva

Marginal future forms will not be addressed here, as they are not accepted by all

consultants. As shown in (4b), sentential objects of complement-taking verbs usu-

ally stand in the accusative, just like regular objects. In contrast, action nominals

that serve as complements of nouns must be in the genitive. In fact, nouns do not

take accusative objects in Tundra Nenets. e dependent complement clause may

contain a subject, which receives regular possessive marking as in other types of

dependent clause. But unlike in the relative clauses described in the previous sec-

tion, subject marking is hosted by the dependent verb itself (the action nominal)

rather than the head noun. is does not preclude possessive marking on the head

noun, but in this instance it has a purely possessive rather than an agentive mean-

ing (see example (14c) below).

e same forms are used for noun complementation. Generally speaking,

lexical nouns taking sentential complements are not numerous in Tundra Nenets;

speakers always prefer to use complement-taking verbs instead. Moreover, the

relevant meanings are oen rendered by non-nite verbal forms (that is, action

nominals), possibly substantivized to various degrees. To cite just one example,

Tundra Nenets seems to lack a general word meaning ‘trace, mark’, so there is no

literal equivalent of the English phrase ‘the trace of where you were lying’. Instead

Tundra Nenets employs the action nominal of the verb ‘to lie’: (pidər°) yūs´eda-

qma-r° (2 lie-.-2) ‘where you lay; the trace from your lying’. But the

verb yūs´eda- is only used for humans and cannot describe animals. e action

nominal of the verb waqnə- ‘to lie (of animals)’ has to be used instead: wen´ako-h

waqn°-qma (dog- lie-.) ‘where the dog lay; the trace from the dog’s ly-

ing’. Obviously, these constructions do not represent noun complements. Since

noun complementation is not generally widespread, it has not received much at-

tention in the existing literature on Tundra Nenets (but see Nikolaeva 2014).

However, a certain number of complement-taking nouns do exist and they

fall into three semantic and formal classes. e rst class includes nouns that de-

note content (e.g. pad°r ‘paper, letter, document’, kniga ‘book’ (borrowed from

Russian), sʹidʹaŋk° ‘photograph’), means of communication (e.g. wada ‘word,

news, rumour’, yun° ‘news’, ləx°rʹo ‘conversation’), and the compound expression

yī(-ŋuq) ‘thought’. e peculiarity of this group is that the sentential complement

in the genitive is oen introduced by the postposition nʹamna ‘about’, sometimes

in combination with the imperfective participle of the copular verb ŋǣda ‘being’.

us, the literal translation of, for instance, (14b) in its full version is something

like ‘the book being about how the girl went to the forest.’

(14)

Wera-h

Wera-

nʹe-m

woman-

mʹi-wa-h

take-.-

(nʹanma)

about

yun°

news

‘the news that Wera got married.’

Chapter 7. e general noun-modifying clause construction in Tundra Nenets 157

nʹe

woman

ŋəcʹeki°-h

child-

pedara-n°h

forest-

xæ-wa-h

go-.-

nʹamna

about

(ŋæ-da)

be-.

kniga

book

‘the book about the girl going to the forest’

nʹelʹe-wa-nʹi

marry-.-.1

(nʹamna)

about

ləx°rʹo-doh

conversation-3

‘their conversation about me getting married’

e distribution of nʹamna and ŋæda is dicult to generalize about: in some in-

stances the consultants claimed that both of them are impossible to omit, in other

instances only nʹamna was required, and some phrases were accepted as gram-

matical without either nʹamna or ŋæda. ere seems to be a fair amount of varia-

tion here. What is clear, however, is that this pattern is at least supercially similar

to non-clausal noun complementation. e postposition nʹamna ‘about’ can in-

troduce a non-clausal complement of content-denoting nouns, and the genitive is

the usual case marking for objects of postpositions. But s or oblique s as ad-

nominal dependents are not very typical of Tundra Nenets (see Nikolaeva 2014).

e relation between the dependent oblique phrase and the head must be overtly

expressed by the participial form of a verb. e omission of such a participle is

generally impossible, although this is not necessarily reected in the English trans-

lations. us, the semantically ‘empty’ participle of the copular verb ŋæ- ‘to be’ is

required in (15):

(15)

wenʹako-h

dog-

nʹamna

about

*(ŋæ-da)

be-.

kniga

book

‘the book about a dog.’

As can be seen here, this is largely parallel to (14b), although the obligatoriness of

ŋæda in (14b) is questionable.

Second, there is a small group of abstract words such as mʹan° ‘ability, able’,

pʹir° ‘possibility’ and the very frequent word sʹer with the literal meaning ‘thing,

business’ but oen used to mean ‘possibility, way, tool, reason’ or the like. ese

words are rather similar to the ‘light’ nouns in languages like Japanese. In Tundra

Nenets they behave formally as nouns (or perhaps, adjectives) but are mostly used

in the predicative function, in which instance they seem to be fully or partially

grammaticalized to express various modalities:

(16)

tʹukoxəna

here

mənc°ra-wa-nʹi

work-.-.1

pʹir°

possibility

nʹī

.3

ŋa-q

be-

‘It’s not possible for me to work here.’

158 Irina Nikolaeva

tola-wa-nt°

read-.-.2

sʹer

thing

ŋa

be.3

‘You ought to read.’

Wera-h

Wera-

mənc°ya-m

work-

sʹerta-wa-nta

do-.-.3

mʹan°

able/ability

‘Wera is able to do the work.’

As can be seen here, the non-nite predicate stands in the genitive, as is normally

the case with sentential complements of nouns, but it is not entirely clear whether

we are dealing with monoclausal or biclausal structures in this instance.

Finally, Tundra Nenets has relational nouns that do not denote content and

take sentential complements expressed by action nominals in the genitive form.

ese are nouns expressing perceptions, senses, emotions and products of men-

tal activity, exemplied in (17). Some other semantic types of relational nouns

are shown in (18).

(17)

Wera-h

Wera-

xalʹa-m

sh-

tal°tamp°-wa-h / tal°tamp°-wa-nta

fry-.- / fry-.-.3

ŋəpt°

smell

‘the smell of Wera frying the sh.’

pidər°

you

xalʹa-m

sh-

tal°tamp°-wa-nt°

fry-.-.2

ŋəpt°

smell

‘the smell of you frying the sh’

pidər°

2

xino-qma-nt°

sing-.-2

nʹurcʹəw°

pleasure

‘the pleasure of your singing’

(18)

sʹencʹeləwa-xəna

hotel-

yilʹe-wa-h

live-.-

mʹir

price

‘the price of staying in a hotel’

nosʹi-qməna

polar.fox-.

xanʹe-wa-h

hunt-.-

tenc°-q

way-

ŋoka-q

many-

‘ere are many ways of hunting for polar foxes.’ (T 690)

Some other relational or quasi-relational (that is, construable as relational in

some but not all their meanings) nouns that belong to this group are mesiq ‘hab-

it’, nʹed°bʹa ‘payment’, wato ‘aim, goal, condition’, taro ‘use’, xərwabco ‘wish’, soxa

‘probability (of what can happen)’, xəra ‘reason’, and wæncaxa ‘sin’.

e genitive complement construction has served as a historical source of

dependent adverbial clauses introduced by postpositions. Many postpositions

in Tundra Nenets and other Uralic languages are known to go back to relation-

al nouns, oen with locational meanings. Both complement-taking nouns and

postpositions take the genitive object, and the former may develop into the latter,

Chapter 7. e general noun-modifying clause construction in Tundra Nenets 159

typically in a frozen case form. For example, in (19a) we have the postposition

‘during’ which is historically the dative of the noun meaning ‘length’. In (19a)

‘when, because’ is historically the genitive form of the multifunctional noun sʹer

‘thing’ mentioned above.

(19)

yilʹe-wa-nta

live-.-.3

yampən°h

during

ŋoka

many

wæwa-m

bad-

palʹeə-sʹ°

swallow-.3

‘He suered (literally: swallowed) many bad things during his life.’ (T

438)

səqnʹe-wa-nta

become.wet-.-3

sʹer°h

when

pəne-w°

coat-1

səŋkoworŋa

become.heavy.3

‘My coat became heavy because/when it became wet.’ (T 531)

In these examples postpositions are fully grammaticalized and do not have any

nominal properties, but there may be various intermediate stages.

It is important to emphasize that the noun complement construction is mod-

elled aer the possessive construction. is can be seen from the fact that the

head noun may host the 3rd person singular possessive ax. As mentioned in

Section2, the genitive lexical possessor optionally triggers possessive agreement

on the head. In a similar manner, the head noun of the genitive action nominal can

optionally be marked by the 3rd person possessive ax:

(20)

pūnʹah

back

sʹur°xəl-ma-h

return-.-

sʹer-ta /

thing /

sʹer

thing-3

‘the way of returning back’

ti-m

reindeer-

yark°-wa-h

catch-.-

səbʹe

very.much

nʹencʹelta

ease.3

‘the great ease to catch a reindeer with the lasso’ (T 852)

mər°ner-cʹ°

argue-.

ŋobk°na

together

yilʹe-wa-h

live-.-

taro-da

use-3

‘the use of living together arguing’ (T 237)

sʹencʹeləwa-xəna

hotel-

yilʹe-wa-nt°

live-.-.2

mʹir / mʹir-ta

price / price-3

‘the price of your stay in a hotel’

Example (20d) shows especially clearly that the 3rd person possessive ax on

the head does not express agreement with the dependent subject, like in relative

clauses: the dependent subject here is 2nd person and is represented by the agree-

ment morphology on the action nominal itself. e 3rd person possessive marker

on the head appears to target the dependent clause as a whole, so that the em-

bedded clause headed by the genitive action nominal formally acts as some kind

of ‘possessor’.

160 Irina Nikolaeva

5. General noun-modifying clause constructions

e previous two sections have described two distinct strategies to encode rela-

tivization and noun complementation. In this section I argue that Tundra Nenets

additionally has a version of GNMCC, although its usage may be restricted in

various ways. e construction involves the same verbal forms, action nominals,

which are employed in complementation addressed in Section4, but in GNMCCs

they stand in the nominative rather than genitive. us, the grammatical case of

the non-nite verbal form is indicative as to which of the two constructions we

are dealing with. It is also possible to use the modal converb in -(sʹ)ə or the pur-

posive converb in -məncʹə to express future or modal meaning in the GNMCCs,

but these are fairly rare and for the most part will not play a role in the subsequent

discussion. For ease of presentation the section is structured into two parts: the

relativization-like function of GNMCCs in which the head noun is coreferential

with a clausal argument or adjunct is described in Section5.1 and other functions

in Section5.2, but I hope it will become clear at the end that these two sides cannot

be easily kept apart.

5.1 Relativization-like function

In relative-like constructions with nominative action nominals, agreement with

the dependent subject is hosted by the head noun, much like in the participial

relative clauses, cf.:

(21)

(mənʹ°)

1

nū-wa

stand-.

pʹa-mʹi

tree-1

‘the tree on which I am standing’

(mənʹ°)

1

nū-qma

stand-.

pʹa-mʹi

tree-1

‘the tree on which I stood’

However, nominative action nominals are used to ‘relativize’ dierent grammati-

cal functions. (Hereaer I dispense with the scare quotes, which should however,

always be understood in references to relativization.) ey never relativize subjects

and direct objects, but are employed for lower positions on the accessibility hierar-

chy for which participles are totally excluded. For instance, they serve to relativize

oblique and indirect objects, which normally stand in various case forms. e in-

direct object is in the dative. Example (22) shows that it can only be relativized by

means of the action nominal; participles are impossible in this instance.

Chapter 7. e general noun-modifying clause construction in Tundra Nenets 161

(22)

Petʹa-h

Petya-

kartink°-m

picture-

məneqləbta-qma /

show-. /

*məneqləbta-wi°

show-.

pʹīrʹibtʹa

girl

‘the girl to whom Petya showed the picture’

A similar construction is used for the relativization of the ablative objects in the

Nenets equivalents of ‘the girl from whom you stole the money’ and ‘the man from

whom I bought the gun’, as well as to relativize locations (23), instruments (24) and

expressions of time (25). Respective examples are cited below with an indication of

the local case typical of the noun in the non-relativized version.

(23)

mʹin-ma

walk-.

sʹexare-waq

road-1

‘the road on which we are walking’ (prolative)

(mənʹ°)

1

xər°-m

knife-

temta-qma

buy-.

lapka-mʹi

shop-1

‘the shop where I bought the knife’ (locative)

(24)

ŋuda-mʹi

hand-.1

məda-qma

cut-.

xər°-mʹi

knife-1

‘the knife with which I cut my hand’ (locative)

(25)

toxodənə-°

study-.

xæ-qmʹa

go-.

yalʹa-doh

day-3

‘the day when I went to study’ (locative)

Note that these constructions are not ambiguous: to my knowledge, there are

no verbs in Tundra Nenets that can take two or more oblique objects, and these

constructions cannot be interpreted as subject or object relativization even in the

absence of the overt subject or object in the dependent clause, because relativiza-

tion of subjects and objects requires participles. So, temta-wi° lapka-mʹi (buy-.

 shop-1) can only mean ‘the shop which I bought’, as opposed to temta-

qma lapka-mʹi (buy-. shop-1) ‘the shop where I bought (it)’. However,

ambiguity may arise with respect to the relativization of adjuncts, in particular,

expressions of location, as in (30):

(26)

pida

3

xæ-qmʹa

go-.

mər°-ta

city-3

‘the city where he went / the city from where he le’

Normally the direction ‘to’ is expressed by the dative case and the direction ‘from’

by the ablative case, but since the dependent clause does not contain any ex-

plicit indication of its relation with the head, this relation can be understood in

both ways in (26).

162 Irina Nikolaeva

Turning now to the objects of postpositions, it seems there are no syntactic

constraints that would prohibit relativizing out of s, but relativization is heavily

constrained semantically. Objects of rarely used postpositions with highly spe-

cialized adverbial meanings do not normally relativize. For example, the object

of the postposition xaw°na ‘apart from’ is not relativizable, so the Tundra Nenets

equivalent of ‘the girl apart from whom I didn’t see anyone’ is unacceptable. But

the objects of frequent local postpositions can be relativized:

(27)

toxontabco-h

y-

mən°te-qma

fall-.

tol°

table

‘the table onto which the y fell’ (nʹih ‘onto’)

mʹūd°-naq

caravan-.1

mʹi-ma

move-.

soti°

hill

‘the hill over which our caravan is moving’ (nʹimnʹa ‘over’)

Again, the relation between the head noun and the dependent clause is under-

specied; thus, the most typical interpretation of nʹenecʹə°-h nū-wa pʹa (man-

stand-. tree) is ‘the tree on which the man is standing’ (relativization of the

object of the postposition nʹinʹa ‘on’), but it can also mean ‘the tree next to which’,

‘behind which’, or ‘under which’, depending on the context. All these meanings are

expressed by postpositions.

e lexical semantics of the items that are involved in the construction can

give additional clues. e phrases below are provided with the most typical default

interpretations, which depend on the identity of the subject participant: cats typi-

cally sit on the table or under the table but not at the table, while humans typically

sit at the table. Alternative interpretations are dispreferred but not totally excluded.

(28)

košk°-h

cat-

ŋamtʹo-wa

sit-.

tol°

table

‘the table on / ?at which the cat is sitting’

nʹenecʹə°-h

man-

ŋamtʹo-wa

sit-.

tol°

table

‘the table at / ?on which the man is sitting’

e lexical semantics of the modifying verb also plays a role. It is possible to rela-

tivize the object of a postposition if the verb is frequently collocated with the re-

spective postpositional phrase. For instance, both the complex expression yi-yad-

er- ‘to think about (literally: mind-walk)’, on the one hand, and the verbs xinoq- ‘to

sing’ and tolaŋo- ‘to read’, on the other hand, are combined with the oblique object

expressed by the postpositional phrase headed by the postposition nʹamna ‘about’.

But in the former case the oblique object is virtually obligatory, while singing and

reading about something or somebody is not mentioned with the same frequency

Chapter 7. e general noun-modifying clause construction in Tundra Nenets 163

in Tundra Nenets discourse; it is much more frequent just to refer to the process

of singing or reading without specifying the content of the song or the reading

material. As a result, relativization of the object of nʹamna with the verb ‘to think’

is perfectly ne, while relativizing the object of nʹamna when it is collocated with

‘to sing’ or ‘to read’ was judged as unacceptable or very marginal, see the following

contrast between (29) and (30):

(29)

tʹiki°

this

pʹīrʹibtʹa-h

girl-

nʹamna

about

yī-mʹi

mind-1

yaderŋa

walk.3

‘I am thinking about this girl.’

yī-nʹi

mind-

yader-ma

walk-.

pʹīrʹibtʹa-mʹi

girl-1

‘the girl about whom I am thinking’

(30)

tʹiki°

this

pʹīrʹibtʹa-h

girl-

nʹamna

about

xinoqŋa-d°m

sing-1

‘I am singing about this girl.’

*/? xinoq-ma

sing-.

pʹīrʹibtʹa-mʹi

girl-1

‘the girl about whom I am singing’

So frequency is important here. An additional factor that could make (30b) less ac-

ceptable is that the intransitive verb xinoq- ‘to sing’ can be transitivized by means

of the causative/applicative sux to derive the verb xino-pta- ‘to sing about’. e

direct object it takes can easily be relativized by means of the primary participial

strategy.

In some instances the dependent clause contains a resumptive element, e.g. an

anaphoric adverb with a locational meaning. Such elements explicitly reect the

syntactic and semantic role of the relativized element. us, example (30) above

can be disambiguated by the insertion of the anaphoric adverb tənʹad° ‘from

there’: pida tənʹad° xæqmʹa mər°ta can only mean ‘the city from where he le’.

Relativization of objects of postpositions that is otherwise dicult can be achieved

by means of the resumption strategy too: the dependent clause contains a post-

position with the 3rd person resumptive marker targeting the head noun. In this

instance the semantic relation between the dependent clause and the head noun is

overtly expressed by the postpositional word, so ambiguity does not arise. In (31)

the postposition yeqy°də-nta (instead-3) bears the resumptive pronoun.

(31)

yeqy°də-nta

instead-3

mənc°ra-wa

work-.

pʹīrʹibtʹa-mʹi

girl-1

‘the girl instead of whom I work

164 Irina Nikolaeva

Without the resumptive postposition yeqy°dənta this phrase can only be under-

stood as expressing the most typical relation between the head and the dependent

clause, namely, as ‘the girl for whom I work’, that is, as the relativized variant of (32).

(32)

tʹiki°

this

pʹīrʹibtʹa-x°na

girl-

mənc°ra-d°m

work-1

‘I work for this girl.’

ere is one additional consideration here. In Tundra Nenets pronominal objects

of postpositions are expressed by possessive morphology on the postposition, cf.

the following forms of the postposition ŋil°na ‘under’: tol°-h ŋil°na (table- un-

der) ‘under the table’, (mənʹ°) ŋil°na-nʹi (1 under-1) ‘under me’ and (pida)

ŋil°na-nta (3 under-3) ‘under him/her’. In such constructions the free-stand-

ing pronoun is optional, so possessive morphology on the postposition can have a

pronominal force. But personal pronouns, including their bound variants on post-

positions, only refer to humans (marginally, animals) and can never refer to inani-

mate entities, so for instance ŋil°na-nta cannot mean ‘under it’. e same concerns

3rd person possessive morphology on postpositions which has a resumptive func-

tion. is implies that resumption is not always possible; in fact, it is only possible

if the relativized noun to which the resumptive pronoun refers denotes a human

or an animal, cf.:

(33)

nʹamna-nta

about-3

tolaŋgo-wa

read-.

pʹīrʹibtʹa-mʹi

girl-1

‘the girl about whom I am reading’

* nʹamna-nta

about-3

tolaŋgo-wa

read-.

mar°q-mʹi

city-1

‘the city about which I am reading’

Example (33b) is ungrammatical because the resumptive pronoun on the postpo-

sition refers to the inanimate head noun ‘city’. e phrase in (33b) would be gram-

matical without the postposition but only in the meaning ‘the city in which I am

reading’. Similarly, it is impossible to employ resumption to convey the following

meanings: ‘the tree under which the girl is standing’, ‘the dog in front of which the

reindeer is walking’, ‘the hill behind which there is a river’ and so on.

Finally, recall from Section3 that relativization of possessors is accomplished

by means of participles if the possessed noun is the subject of the relative clause. In

this instance it bears a resumptive possessive ax. Action nominals are also used

for the relativization of possessors but only if the possessed noun is a non-subject

in its clause. Resumption is required in this instance and this does not depend on

animacy/humanness: unlike possessive morphology on postpositions, possessive

morphology on nouns can target inanimate entities.

Chapter 7. e general noun-modifying clause construction in Tundra Nenets 165

(34)

yad°-mta

wall-.3

ŋolʹepʹada-qma

paint-.

xar°də-mʹi

house-1

‘the house the wall of which I painted’

is kind of relativization is not constrained by the nature of the possessive rela-

tionship: both alienable and inalienable possessors can be relativized.

To conclude this subsection, we have seen thus far that nominative action

nominals are employed in the function that looks like relativization, at least super-

cially. ey relativize all clause-level elements except subjects and direct objects,

which require the participial strategy, as well as certain sub-clause level elements,

namely, the most easily recoverable objects of postpositions. e dependent clause

contains a gap corresponding to the relativized element, so there is no overt ex-

pression of the relation between the head and the dependent clause. Other sub-

clause level elements (possessors and less frequent postpositional objects) can only

be relativized if their content is recovered by a resumptive pronoun in the depen-

dent clause. It is also worth mentioning that some sub-clause level elements do not

seem to be relativizable under any conditions, even by means of resumption. ese

are objects of comparison, which are normally expressed by the ablative  within

the adjectival phrase, and oblique -internal elements, so, for example, ‘the girl

about whom the grandmother told a tale’ is not possible. e accusative object of

the grammaticalized converb of the transitive verb nʹaq- ‘to have as a companion’,

nʹacʹ°, which forms a kind of serial verb construction adding a comitative argu-

ment to the lexical verb, cannot be relativized either; therefore the exact Tundra

Nenets equivalent of ‘the man with whom I work’ cannot be constructed, and vari-

ous periphrastic escape strategies have to be used to express the relevant mean-

ings. But this is the only instance of the serial verb construction in the language.

5.2 Extended functions

In addition to relativization-like functions, nominative action nominals bear the

argument function for relational head nouns. I have shown in Section4 that sen-

tential noun complementation is usually achieved by means of the genitive action

nominal, but nominative action nominals represent an alternative way of express-

ing essentially the same meaning.

(35)

mašina-h

car-

mʹi-ma /

move-. /

mʹi-ma-nta

move-.-.3

mūh

sound

‘the sound of the car moving’

Formally, the nominative construction is identical to that described in the previ-

ous subsection for the relativization of non-core grammatical functions: the action

166 Irina Nikolaeva

nominal stands in the nominative, the head noun is not represented in the depen-

dent clause, there is no explicit expression of the relation that holds between the

head and the dependent clause, but the head noun can host dependent subject

agreement. is contrasts with the genitive construction in which agreement is on

the dependent verb. Examples (36) clearly show the dierence in the location of

agreement in the nominative and genitive action nominal constructions.

(36)

ŋuda

hand

yesʹa-mt°

metal-.2

yo-qma

lose-.

wada-r° / *wada

word-2 / word

or:

ŋuda

hand

yesʹa-mt°

metal-.2

yo-qma-nt°

lose-.-.2

wada / *wada-r°

word / word-2

‘the news/rumour that you lost your ring’

škola-m

school-

yolcʹe-qmʹa-nt°

nish-.-.2

pad°r

paper

or:

škola-m

school-

yolcʹe-qmʹa

nish-.

padəl°

paper.2

‘the document conrming that you nished school’

mənc°ra-qma

work-.

nʹed°bʹa-r°

payment-2

or:

mənc°ra-qma-nt°

work-.-.2

nʹed°bʹa

payment

‘the payment for your work’

yewako-m

orphan-

yar°da-qma

oend-.

wæncaxa-da

sin-3

or:

yewako-m

orphan-

yar°da-qma-nta

oend-.-.3

wæncaxa

sin

‘the sin of him oending an orphan’

Taken together, the data in (35) and (36) demonstrate that variability is possible

for all groups of complement-taking nouns mentioned above: content, percep-

tional and other relational nouns. It is dicult to generalize as to what motivates

the choice of one construction over the other. Speakers cannot clearly identify se-

mantic dierences, although there were also comments indicating that the genitive

construction may have a more ‘abstract’ meaning. For instance, (37a) was claimed

to be about the present need to kill the reindeer, while (37b) is rather perceived to

be a general statement, not necessarily referring to the present situation.

Chapter 7. e general noun-modifying clause construction in Tundra Nenets 167

(37)

ti-m

reindeer-

xada-wa

kill-.

yolcʹə-l°

time-2

to°

come.3

‘Your turn came to kill a reindeer.’

ti-m

reindeer-

xada-wa-nt°

kill-.-.2

yolcʹ°

time

to°

come.3

‘Your turn came to kill a reindeer.’

However, this issue needs further investigation. Note also that, unlike (37b), (37a)

is in fact ambiguous: it can be understood as the ‘relativization’ of a time adjunct

as ‘the time at which you kill a reindeer’.

e important point is that a single nominative construction can be used, rst,

for relativization-like phenomena, where the head noun plays a semantic and syn-

tactic role in what looks like a modifying clause, and second, for sentential noun

complementation, where the whole dependent clause serves as an argument to

the head but the latter has no syntactic role in it. What is more, nominative action

nominals can modify head nouns which are neither argument-taking nor easily

reconstructable inside the dependent clause. Instead some kind of loose ‘frame’

relation determined on semantic and/or pragmatic grounds obtains between the

head and the dependent. e relation can be interpreted in dierent ways, for ex-

ample, as reason or cause:

(38)

ŋarka

big

nʹū-m

child-

yo-qma

lose-.

yedʹa-da

pain-3

‘the pain from him losing a grown up child’

mənc°ra-qma

work-.

yesʹa-da

money-3

‘the money because of which he worked’

nʹen°sʹuwo-qma

quarrel-.

sʹo-waq

song-1

‘the song because of which we have quarrelled’

ese examples cannot be understood as involving relativization of the object of

the postposition yeqm°nʹa ‘because’, since such objects are not relativizable with-

out resumptive pronouns. e postposition actually has two meanings, ‘because’

and ‘for’, as in (39a), which has two interpretations. e ‘for’-type object can be

relativized by means of resumption (39b), while the causal relation is simply ex-

pressed by the nominative action nominal (39c).

(39)

tʹuku°

this

nʹe-h

woman-

yeqm°nʹa

for / because

xinoqŋa-d°m

sing-1

‘I am singing because of / for this woman.’

168 Irina Nikolaeva

yeqm°nʹa-nta

for-3

xinoq-ma

sing-.

nʹe-mʹi

woman-1

‘the woman for / ? because of whom I am singing’

xinoq-ma

sing-.

nʹe-mʹi

woman-1

‘the woman because of / *for whom I am singing’

It is dicult to see how the contrast between (39b) and (39c) can be motivated on

purely syntactic grounds without, for instance, assuming that the two meanings of

yeqm°nʹa are actually associated with two dierent structures. A better alternative

would be to accept that the object of yeqm°nʹa is not relativizable without resump-

tion, just like the objects of other ‘rare’ postpositions, but the dependent clause is

interpreted as standing in a loose causal relation to the head noun. is solution

is supported by the fact that acceptability of the causal relation is not unlimited: it

depends on the semantics of the relevant parts, cf.:

(40)

nʹaŋor-oqma

become.fat-.

ŋəw°rə-r°

food-2

‘the food because of which you became fat’

? xaŋkulə-qma

get.ill-.

ŋəw°rə-r°

food-2

‘the food because of which you got ill’

* wæwawna

badly

xonʹo-wa

sleep-.

ŋəw°rə-r°

food-2

‘the food because of which you sleep badly’

Example (40a) ‘the food because of which you get fat’ is well-formed, but (40c) ‘the

food because of which you sleep badly’ was not judged as acceptable, apparently

because the semantic relation between the head and the dependent clause is atypi-

cal and therefore infrequent. Example (40b) ‘the food because of which you got

ill’ was judged as marginal since, according to my consultants, ‘the Nenets do not

normally eat food from which they get ill’.

e relation between the head and the dependent clause may be that of pur-

pose, although this is less frequent. e purposive converb here serves for future

time reference, cf.:

(41)

xarəd°-m

house-

sʹerta-wəncʹ°

do-.

yesʹa-r°

metal-2

‘the money for you to build a house’

xarəd°-m

house-

sʹerta-qma

do-.

yesʹa-r°

metal-2

‘the money you spent on building a house’

Chapter 7. e general noun-modifying clause construction in Tundra Nenets 169

Example (41b) is in fact ambiguous: the second interpretation ‘the money you re-

ceived for building a house’ involves a relationship of consequence/result. In (42)

it is even dicult to clearly name the type of semantic relation that holds between

the head and the dependent; most likely it involves some kind of manner or gen-

eral association.

(42)

xalʹa-m

sh-

pʹirʹe-qma

boil-.

ŋəw°rə-mʹi

food-1

‘the meal I made by boiling sh’

tedor-oqma

ght-.

sʹarka-doh

wine-3

‘wine such that they fought when they drank it’

pidər°

2

nʹanʹi

.1

nʹat°-wa

be.friendly-.

wada-r°

word-2

‘the word of you being friendly to me’

ŋəcʹeki°

child..

toxolamp°-wa

teach-.

mənc°ya-r°

work-2

‘your work of teaching children’

e genitive action nominal is not normally found for this group of examples,

although occasionally speakers can produce it.

What this discussion generally shows is that there exists a very general seman-

tically undetermined association between the nominative action nominal and the

head noun. In some instances the head noun is interpreted as an argument or ad-

junct to the action nominal, while in other instances this relationship is reversed:

the action nominal is an argument or some kind of semantic adjunct to the head.

Since a single construction is used for all these purposes and since the dependent

clause does not have to contain an explicit marker of the relation between the de-

pendent and the head, we are dealing with GNMCCs. It covers all types of NMCC:

argument and adjunct NMCCs, frame NMCCs and NMCCs involving argument-

taking head nouns. is analysis implies that what was referred to as ‘relativiza-

tion’ in Section5.1 is not actually relativization at all, whereas ‘true’ relativization

in Tundra Nenets only targets the highest grammatical functions, as described

in Section3.

Note that according to the Introduction to this volume, “the strict denition

of GNMCC excludes syntactic constraints on the relation between the head noun

and the modifying clause”. But in Tundra Nenets what can be rather condently

characterized as GNMCC does appear to be restricted syntactically: GNMCCs do

not serve for the ‘relativization’ of subjects and direct objects, as well as most sub-

clause level elements. If the latter restriction can perhaps be accounted for by non-

syntactic factors such as low frequency, semantic oddness, processing diculties

170 Irina Nikolaeva

or simply the lack of an appropriate context, the former is more dicult to ex-

plain in this manner. However, I suspect that the relevant distribution may have

to do with the presence of a dedicated alternative strategy. Even if GNMCCs are

in principle available to relativize subjects and objects, they would be blocked by

participial relative clauses. is means that syntactic restrictions on GNMCCs are

not only due to the structural properties of the construction itself, but also to the

availability of competing constructions.

6. GNMCCs and noun-noun compounds

e aim of this section is to provide additional details about the structure of

GNMCCs in Tundra Nenets and to show that they have parallels in the non-sen-

tential domain.

As mentioned in Section2, Tundra Nenets exhibits three types of simple ad-

nominal dependent. e third type instantiates what can be referred to as ‘modi-

cation-by-noun’ (Spencer 2011; Nikolaeva & Spencer 2012; Spencer 2013): a head

noun is modied with reference to another nominal entity. e modifying expres-

sion denotes a subset of the entities denoted by the head noun, so the semantics of

the whole phrase can be eectively characterized as intersective modication, just

like in true adjectival modication. e meaning of the dependent is part of the

complex meaning of the semantic modier as a one place predicate which picks

up a subset of the entities denoted by the head. However, unlike in standard ad-

jectival modication with simple property concepts, the semantic structure of the

modier is more complex because the modier has a purely noun-like denotation

and, arguably, belongs to the lexical class of nouns (cf. Beck 2002: 88; Nikolaeva &

Spencer 2012; Spencer 2013).

Modifying nouns in Tundra Nenets stand in the nominative and precede the

head noun within the complex : [N N]. ey can render, for instance, the

following semantic relationships: material (yes´a ləbtey°ko ‘iron box’), temporal

characteristics (yunuy° m´erc´ako ‘spring wind’), spatial characteristics (pedara

yəxa ‘forest river’), kind-species relationships (xasawa ŋəc´eki° ‘boy; literally:

man(ly) child’; yox°rey° wen´ako ‘stray dog’), and the like. But crucially, any list

would be misleading as the range of relations is potentially innite, much like

in (non-lexicalized) noun-noun compounds in English. According to Downing

(1977), Spencer (2011) and other literature, the only real function of English

noun-noun compounding is to permit a very general relationship to be expressed

between two nouns. e structure of compounds does not contain an overt indica-

tor of the semantic relation between the two relevant nouns, so they are open to

multiple interpretations. is is also true for Tundra Nenets. Constructions with

Chapter 7. e general noun-modifying clause construction in Tundra Nenets 171

modifying nominative nouns do tend to acquire xed lexicalized meanings de-

pending on the meaning of their parts, but ultimately their interpretation can only

be determined in specic contexts and dened in terms of the (contextually given)

understandings of the relevant nouns. For example, the nouns denoting containers

are expected to stand in the relationship of content and container when they head

a modifying noun, so the most typical interpretation of sax°r xidʹa (sugar-cup) is

‘cup of sugar’. However, in an imaginary context describing cups made of dierent

materials this expression can in principle be understood to mean ‘cup made of

sugar’. e expression yunuy° ləx°rʹo (spring-conversation) will typically be under-

stood as ‘conversation about spring’ where ‘spring’ is an argument of the content

noun, but it but can also mean ‘conversation held in spring’.

is relation between two nouns seems to be no less free than the semanti-

cally vague relation between the possessor and possessee in (alienable) possessive

constructions, but there is an essential dierence too. e main function of the

possessive construction is known to be the identication of the possessee through

its relation with the possessor, which serves as a pragmatic anchor (Hawkins 1981;

Fraurud 2001; Koptjevskaja Tamm 2000) or reference point (Langacker 1993 and

other works) for the conceptualization of the latter. erefore possessors are typi-

cally referential. In contrast, modifying nouns serve to classify, describe or qualify

the head noun rather than to identify it by providing a reference point. ey are

usually non-referential and express a property.

So modifying nouns are somewhat ‘in-between’ possessors and adjectival

modiers, and they do not have canonical syntactic properties of either of them.

In Tundra Nenets, while both nominal and adjectival modiers may agree with

their head, their patterns of concord dier. In particular, nominal modiers op-

tionally agree in number and, rarely, possessive person/number but, unlike adjec-

tives, they can never agree in case.

(43)

yes´a-q

metal-

xid´a-q

cup-

‘metal cups’

yes´a-r°

metal-

xid´a-r°

cup-

‘your metal cups’

* yes´a-xədənt°

metal-.2

xid´a-xədənt°

cup-.2

‘from your metal cup’

Unlike possessors, nominal dependents in the nominative are inseparable from

the head. ey must follow the possessor, cf. mənʹ° ŋuda yesʹa-mʹi (1 hand

metal-1) ‘my ring’ and the ungrammatical *ŋuda mənʹ° yesʹa-mʹi. Adjectives

172 Irina Nikolaeva

normally follow possessors, cf. Wera-h n´ud´a tubka (Wera- small axe) vs.

*n´ud´a Wera-h tubka (small Wera- axe) ‘Wera’s small axe’, but they must pre-

cede the nominative noun and cannot be inserted between it and the head.

(44)

səwa

good

yes´a

metal

xid´a /

cup /

*yes´a

metal

səwa

good

xid´a

cup

‘good metal cup’

e example in (44) demonstrates that with respect to position adjectives behave

dierently from attributive nouns: only the latter must be adjacent to the head,

thereby implicating the status of nominal compounds for these constructions.

Related to this is the fact that, unlike adjectives, nominative nouns cannot stand

in apposition to the head, be postposed or function in null-headed s, so (45) is

strictly ungrammatical, unlike (2) and (3) cited above.

(45)

*xid´a-n°h

cup-

(yes´a-n°h)

metal-

si°rŋa-d°m

look-1

(yes´a-n°h)

metal-

‘I am looking at the cup, the metal one.’

*yes´a-n°h

metal-

si°rŋa-d°m

look-1

‘I am looking at the metal one.’

Because of the adjacency restriction, I will refer to the relevant structures as noun-

noun compounds although concord facts suggest that they behave somewhat dif-

ferently from ordinary compounds, where one might expect a single edge inec-

tion marker for number.

Crucially, this compound-like structure exhibits formal parallels to what was

analysed as the GNMCC in Section5. First, we saw that modication-by-noun ex-

hibits a distinct pattern of concord: modifying nouns optionally agree in number,

very infrequently in person/number, but never in case. As demonstrated in (46),

concord on nominative action nominals has the same distribution.

(46)

yilʹe-qmʹa /

live-. /

yilʹe-qmʹa-x°h /

live-.- /

yilʹe-qmʹa-x°yu-nʹi

live-.--1

mʹa-k°yu-nʹi

yurt--1

‘(two) yurts in which I lived’

yilʹe-qmʹa /

live-. /

yilʹe-qmʹa-mʹi /

live-.-1 /

*yilʹe-qmʹa-x°nanʹi

live-.-.1

mʹa-k°nanʹi

yurt-.1

‘in the yurt in which I lived’

Examples (46) illustrate the relativization-like function of GNMCCs, but the same

is observed in other functions too:

Chapter 7. e general noun-modifying clause construction in Tundra Nenets 173

(47)

nʹelʹe-wa /

marry-. /

nʹelʹe-wa-r° /

marry-.-2 /

*nʹelʹe-wa-mt°

marry-.-.2

yunə-mt°

news-.2

‘the news () that you are getting married’

nʹen°sʹuwo-qminaq

quarrel-...1

sʹoyonaq

song..1

‘the songs because of which we quarrelled’

Second, nominal modication is compound-like in the sense that the components

of the construction are inseparable. Likewise, the nominative action nominal must

be adjacent to the head: unlike the ‘true’ relative clauses shown in (11) and (12),

it cannot be postposed to the head as some kind of aerthought and cannot be

separated from it by adjectives.

(48)

yilʹe-qmʹa-mʹi

live-.-1

mʹa-k°nanʹi

yurt-.1

tʹecʹ°ə-sʹ° /

cold-

* mʹak°nanʹi yilʹeqmʹamʹi tʹecʹ°əsʹ°

‘It was cold in the yurt where I lived.’

ŋopoy°

only

yilʹe-qmʹa

live-.

mʹaq-mʹi/

yurt-1

* yilʹeqmʹa ŋopoy° mʹaqmʹi

‘the only yurt where I lived’

e requirement that the dependent clause be adjacent to the head ensures that in

the situation when the head takes its own possessor the GNMCC cannot be used.

Instead, the action nominal stands in the genitive and is oen accompanied by the

postposition nʹamna ‘about’, cf. (a) and (b) below:

(49)

nʹelʹe-wa

marry-.

xəra-mʹi

reason-1

‘my reason to get married’

nʹelʹe-wa-nt°

marry-.-.2

(nʹamna) /

about /

*nʹelʹe-wa-r°

marry-.-2

(mənʹ°)

1

xəra-mʹi

reason-1

‘my reason for you to get married’

(50)

xaŋkulə-wa-(mʹi)

get.ill-.-1

yedʹe-w°

pain-1

‘my pain from getting ill’

174 Irina Nikolaeva

xaŋkulə-wa-nt° /

get.ill-.-.2 /

*xaŋkulə-wa-r°

get.ill-.-2

(mənʹ°)

1

yedʹe-w°

pain-1

‘my pain because of you getting ill’

As shown in (12) above, the true relative clause in Tundra Nenets must precede

the possessor to the head and so does the genitive action nominal (49b, 50b). In

contrast, the nominative action nominal cannot stand in this position, much like

non-clausal nominal modiers. is conrms again that GNMCCs can be viewed

as sentential corollaries of noun-noun compounds.

7. Remarks about diachrony

We have seen that Tundra Nenets has three major structural-semantic types of ad-

nominal dependent: possessors, attributive adjectives, and modifying nouns. ey

all have sentential parallels in the domain of NMCCs in terms of basic semantics,

syntactic behaviour and the morphosyntactic expression of the ingredients of the

construction. Since Tundra Nenets NMCCs demonstrate remarkable similarity to

non-clausal adnominal dependents, this appears to suggest that some diachronic

mechanisms may have played a role here.

Non-nite verbs oen start as deverbal nouns or deverbal adjectives formed

by derivational processes, and only gradually acquire clausal status. As suggested

in Harris & Campbell (1995: 310–312), they have the potential for being reanal-

ysed as possessing a complex clausal structure because of their inherent dual na-

ture: since they are formed on verbal bases, deverbal nominals are open to an

interpretation as a verb and thereby provide a categorial basis for reanalysis as the

head of a clausal domain.

In Tundra Nenets, genitive action nominals employed in complementation

are formally similar to possessors as they take the genitive and do not show at-

tributive concord. Moreover, they are compatible with the 3rd person possessive

marker on the head noun, as is also observed in regular possessive constructions

with the lexical possessor. is suggests that the genitive action nominal origi-

nates as the formal ‘possessor’ of the head noun. It later acquired a clausal status,

and its own possessor was reanalysed as the subject. For instance, [pidər° xalʹa-m

tal°tamp°-wa-nt°] ŋəpt° (2 sh- fry-.-.2 smell) ‘the smell of

[you frying the sh]’ goes back to something like ‘the smell of [your frying of the

sh]’ or, more literally, ‘[your sh frying]’s smell’. e rationale behind using the

possessive construction in noun complementation is that the clause headed by

the action nominal is an argument of the head noun. In a similar manner, pos-

sessive phrases are oen understood as conveying a two place relation, in which

Chapter 7. e general noun-modifying clause construction in Tundra Nenets 175

possessors function as (subject) arguments to their heads (Kayne 1994; Szabolcsi

1994; Laczkό 1997; Partee 1997, among many others).

In contrast, Tundra Nenets participles are very much adjective-like in terms

of their semantics (essentially, set intersection) and formal properties, e.g. con-

cord. Nikolaeva (1999) and Ackerman & Nikolaeva (2013) argue in detail that

the relative clause structure with the unusual agreement pattern found in Tundra

Nenets and some other languages of Eurasia emerged as the attributive construc-

tion rather than a construction with a possessor-like deverbal noun. e historical

development involved the reanalysis of constituent structure and the acquisition

of clausal status by the dependent verbal form, the participle. It may be represent-

ed as follows: pidər° [xada-wi° te-r°] (2 kill-. reindeer-2) ‘your killed

reindeer’ > [pidər° xada-wi°] te-r° ‘the reindeer you killed’. In this initially posses-

sive structure with an attributive modier person/number marking was local to

the possessive phrase. At the next stage of historical development, the modifying

deverbal form was reanalysed as being in constituency with the possessor which,

in turn, acquired subject status. e position of agreement, however, remained

unaltered, hence the violation of locality in the resulting construction.

e third type, nominative deverbal nouns (that is, GNMCCs), patterns to-

gether with modication-by-noun constructions, in which the modifying noun

stands in the nominative and forms a compound-like structure: concord is very

restricted and the dependent clause must be strictly adjacent to the head from

the le, unlike in the previous two types. In addition to these formal similarities,

there are functional parallels. In noun-noun compounds the exact nature of the

relationship between the head and modier has to be dened pragmatically rather

than in terms of the xed semantic representations of the constituents. Similarly,

the acceptability and meaning of GNMCCs heavily depend on the semantics of

their parts, the head noun and the action nominal. In many instances it is com-

puted fairly unambiguously, especially if the whole dependent clause can be inter-

preted as an argument of a complement-taking noun. But it does not have to be

understood as an argument and there is always space for variation: GNMCCs with

the same lexical content can be interpreted in several dierent ways, as long as the

interpretation is semantically and pragmatically plausible.

is formal and semantic parallelism, together with the assumption that non--

nite dependent clauses originate as non-clausal deverbal nouns, suggests that Tundra

Nenets GNMCCs are analogically based on modication-by-noun constructions

and go back to them historically. ey emerged when the modifying deverbal noun

in the nominative was reanalysed as heading a clausal domain, and the possessor

was reanalysed as its subject, similar to what is observed in the previous type: mənʹ°

[xalʹa-m pʹirʹe-qma ŋəw°rə-mʹi] (1 sh- boil-. food-1) ‘my sh boil-

ing meal’ > [mənʹ° xalʹa-m pʹirʹe-qma] ŋəw°rə-mʹi ‘the meal I made by boiling sh’.

176 Irina Nikolaeva

A short note on variation is in order here. Speakers occasionally accept a vari-

ant of GNMCC in which subject agreement is hosted by the action nominal rather

that the head noun. is is observed in relativization-like structures (51) and also

occurs, albeit less frequently, in other functions (52).

(51)

to-qma-mʹi

arrive-.-1

yalʹa-rʹi-x°na

day--

‘on the same day as I arrived’

xər°-m

knife-

mʹiso-qma-mʹi

give-.-1

xasawa

man

‘the man to whom I gave the knife’

(52)

? pidər°

2

škola-n°h

school-

xæ-qmʹa-r°

go-.-2

sʹidʹaŋk°

photograph

‘the photograph of how you went to school’

ti-m

reindeer-

wadaba-wa-r° /

raise-.-2 /

wadaba-wa-nt°

raise-.-.2

ləx°rʹo

conversation

‘the conversation about you raising the reindeer’

škola-m

school-

yolcʹe-qmʹa-r°

nish-.-2

pad°r

paper

‘the document conrming that you nished school’

I believe this variation reects the intermediate transitional status of the construc-

tion. In relative clauses agreement is consistently on the head noun (Section3),

while in complement and adverbial clauses agreement is consistently on the verbal

form (Section4). It is likely that in GNMCCs agreement originally patterned with

relative clauses because this is also the pattern of possessive agreement observed in

the compound-like structure which, I have hypothesized, served as their analogi-

cal source, cf. mənʹ° ŋuda yesʹa-mʹi (1 hand metal-1) ‘my ring’ cited above.

Agreement on the verbal form is secondary and results from the interfering inu-

ence of complement clauses, whose functions partly intersect with the functions of

GNMCCs. In other words, variation in the surface shape of the GNMCC in terms

of the position of agreement has been facilitated by other constructions.

On a more general note, we can hypothesize that the availability of GNMCCs

in a language correlates with the presence of a comparable non-clausal structure.

If a language has an undetermined adnominal structure that expresses a variety of

functions, the latter may give rise to the GNMCC. is condition is neither neces-

sary nor sucient, but can be interpreted as a diachronic tendency. Although the

grammar permits or even encourages the relevant diachronic path, it does not

deterministically mandate its appearance among many alternatives (cf. Ackerman

& Nikolaeva 2013). erefore, one can motivate GNMCCs but cannot predict that

they will necessarily occur in specic languages.

Chapter 7. e general noun-modifying clause construction in Tundra Nenets 177

Abbreviations

 ablative  locative

 accusative  modal

 action nominal  negation

 conditional  past

 converb  participle

 dative  perfective

 dual  plural

 future  prolative

 genitive  purposive

 inchoative  singular

 imperfective

References

Ackerman, Farrell & Nikolaeva, Irina. 2013. Descriptive Typology and Linguistic eory: A Study

in the Morphosyntax of Relative Clauses. Stanford CA: CSLI.

Barker, Chris. 1995. Possessive Descriptions. Stanford CA: CSLI.

Barker, Chris. 2011. Possessives and relational nouns. In Semantics: An International Handbook

of Natural Language Meaning, Vol. 2, Claudia Maienborn, Klaus von Heusinger & Paul

Portner (eds), 1109–1130 Berlin: De Gruyter.

Beck, David. 2002. e Typology of Parts of Speech Systems. e Markedness of Adjectives.

London: Routledge.

Downing, Pamela. 1977. On the creation and use of English nominal compounds. Language 55:

810–842.

doi: 10.2307/412913

Fraurud, Kari. 2001. Possessives with extensive use: A source of denite articles? In Dimensions

of Possession [Typological Studies in Language 47], Irène Baron, Michel Herslund & Finn

Sørensen (eds), 243–267. Amsterdam: John Benjamins.

doi: 10.1075/tsl.47.14fra

Harris, Alice & Campbell, Lyle. 1995. Historical Syntax in Cross-linguistic Perspective. Cambridge:

CUP.

doi: 10.1017/CBO9780511620553

Hawkins, Roger. 1981. Towards an account of the possessive constructions: NPs and the N of

NP’. Journal of Linguistics 17: 247–269.

doi: 10.1017/S002222670000699X

Higginbotham, James. 1985. On semantics. Linguistic Inquiry 16(4): 547–593.

Kayne, Richard S. 1994. e Antisymmetry of Syntax. Cambridge MA: e MIT Press.

Koptjevskaja-Tamm, Maria. 2000. Romani genitives in cross-linguistic perspective. In

Grammatical Relations in Romani: e Noun Phrase [Current Issues in Linguistic eory

211], Viktor Elšík & Yaron Matras (eds), 123–149. Amsterdam: John Benjamins.

doi: 10.1075/cilt.211.08kop

Laczkó, Tibor. 1997. Action nominalization and the possessor function. Acta Linguistica

Hungarica 44 (3–4): 413–475.

Langacker, Ronald. 1993. Reference-point constructions. Cognitive Linguistics 4: 1–38.

doi: 10.1515/cogl.1993.4.1.1

178 Irina Nikolaeva

Larson, Richard & Segal, Gabriel. 1995. Knowledge of Meaning. Cambridge MA: e MIT Press.

Lewis, M. P. (ed.). 2009. Ethnologue: Languages of the World, 16th edn. Dallas TX: SIL

International. <http://www.ethnologue.com/>

Nikolaeva, Irina. 1999. Ostyak. Munich: Lincom.

Nikolaeva, Irina. 2003. Possessive axes as markers of information structuring: Evidence from

Uralic. In International Symposium on Deictic Systems and Quantication in Languages

Spoken in Europe and North and Central Asia. Collection of papers, Pirkko Suihkonen

& Bernard Comrie (eds), 130–145. Izhevsk: Udmurt State University, and Leipzig: Max

Planck Institute for Evolutionary Anthropology.

Nikolaeva, Irina. 2005. Modier-Head person concord. In Morphology and Linguistic Typology.

Proceedings of the Fourth Mediterranean Morphology Meeting (MMM4) Catania, 21–23

September 2003, G. Booij, E. Guevara, A. Ralli, S. Sgroi & Sergio Scalise (eds), 221–234.

Bologna: Universití degli Studi di Bologna.

Nikolaeva, Irina. 2014. A Grammar of Tundra Nenets. Berlin: Mouton de Gruyter.

doi: 10.1515/9783110320640

Nikolaeva, Irina & Spencer, Andrew. 2012. Possession and modication: A perspective from

canonical typology. In Canonical Typology, Dunstan Brown & Greville G. Corbett (eds),

207–238. Oxford: OUP.

Partee, Barbara H. 1997. Genitives: A case study. Appendix to eo M.V. Janssen,

‘Compositionality’. In Handbook of Logic and Linguistics, Johan van Benthem & Alice ter

Meulen (eds), 464–470. Amsterdam: Elsevier.

Partee, Barbara H. & Borschev, Vladimir. 1998. Integrating lexical and formal seman-

tics: Genitives, relational nouns, and type-shiing. In Proceedings of the Second Tbilisi

Symposium on Language, Logic, and Computation, Robin Cooper & omas Gamkrelidze

(eds), 229–241. Tbilisi: Center on Language, Logic, Speech, Tbilisi State University.

Salminen, Tapani. 1997. Tundra Nenets Inection. Helsinki: Suomalais-Ugrilainen Seura.

Seiler, Hansjakob. 1983. Possession as an Operational Dimension of Language [Language

Universals Series, 2]. Tübingen: Narr.

Spencer, Andrew. 2011. What’s in a compound. Review article of Lieber and Štekauer (eds), e

Oxford Handbook of Compounding. Journal of Linguistics 47: 481–507.

Spencer, Andrew. 2013. Lexical relatedness. Oxford: OUP.

doi: 10.1093/acprof:oso/9780199679928.001.0001

Sperber, Dan & Wilson, Deidre. 1986. Relevance: Communication and Cognition. Oxford:

Blackwell.

Szabolcsi, Anna. 1994. e noun phrase. In Hungarian syntax [Syntax and Semantics 27], Ferenc

Kiefer & Katalin É. Kiss (eds), 179–274. New York NY: Academic Press.

Taylor, John R. 1996. Possessives in English. An Exploration in Cognitive Grammar. Oxford: OUP.

Tereščenko, Natalʹja M. 1965. Nenecko-russkij slovarʹ (Nenets-Russian dictionary). Moscow:

Sovetskaja ènciklopedija.

The Wiru Noun-Modifying Clause Construction

Article

Full-text available

Jan 2021
Ocean Ling

The Typology of Parts of Speech Systems: The Markedness of Adjectives

Book

Full-text available

Jan 2013

David Beck

A Grammar of Tundra Nenets

Book

Full-text available

May 2014

Irina Nikolaeva

Possession and modification—a perspective from Canonical Typology

Chapter

Full-text available

Nov 2012

The chapter addresses a set of semantic, syntactic, and categorial criteria for canonical attributive modification and canonical inalienable possession. Canonical attributive modification is expressed by a dedicated word class (adjective) denoting a property concept, while canonical possession is inalienable possession of a relation noun (kin term, (body) part). Irina Nikolaeva and Andrew Spencer argue that modification-by-noun and alienable possession constructions can be analysed as non-canonical variants of canonical modification and possession. In many languages the same morphosyntax is used to express various combinations of the four constructions and the authors explore some of the typological variation in terms of selective violations of canonical properties. © editorial matter and organization Dunstan Brown, Marina Chumakina, and Greville G. Corbett 2013. © the chapters their several authors 2013. All rights reserved.

Descriptive typology and linguistic theory: A study in the morphosyntax of relative clauses

Book

Full-text available

Jan 2014

Descriptive grammarians and typologists often encounter unusual constructions or unfamiliar variants of otherwise familiar construction types. Many of these phenomena are puzzling from the perspective of linguistic theories: they neither predict nor, arguably, provide the tools to insightfully describe them. This book analyzes an unusual type of relative clause found in numerous related and unrelated languages of Eurasia. While providing a detailed case study of Tundra Nenets, it broadens this inquiry into a detailed typological exploration of this relative clause type. The authors argue that an understanding of this construction requires exploring the (type of) grammar system in which it occurs in order to identify the (set of) independent constructions that motivate its existence. The resulting insights into grammar organization illustrate the usefulness of a construction-theoretic syntax and morphology informed by a developmental systems perspective for the understanding of complex grammatical phenomena.

On semantics

Article