Ni: Negative concord ц in Russian

Размещено на http://www.allbest.ru/

Статья по теме:

Частица ни как вершина ц в русском языке

Д. Тискин, Санкт-Петербургский государственный университет



Тискин Даниил Борисович - магистрант кафедры общего языкознания филологического факультета, Санкт-Петербургский государственный университет

Статья посвящена особенностями синтаксиса и семантики союза/частицы ни в русском языке в свете теории координации М. Митровича. Основные наблюдения статьи состоят в том, что ни, хотя и не может функционировать в качестве аддитивной частицы, способен погружаться внутрь каждой из сочиняемых составляющих, если его позиция всякий раз одна и та же. Мы предполагаем, что отсутствие непарных употреблений ни, которые классифицировались бы как частицы, связано с необходимостью проверки признака [+npi], что по тем или иным причинам возможно только посредством юнктивной вершины J, несущей тот же признак. Допустив, что ни разлагается на не и и, мы объясняем требование тождества позиции ни при вложении тем, что не и весь предшествующий ему материал, одинаковый во всех клаузах, претерпевает постсинтаксический спуск к каждому из конъюнктов, что создает эффект вложенности ни в каждый из конъюнктов.

Ключевые слова: сочинение, отрицание, эффект крысолова, клитики, NPI, фонологическая форма, передвижение Across-the-Board.



Tiskin Daniel В. - МА student at the Department of General Linguistics of Faculty of Philology, Saint Petersburg State University

Ni: Negative concord ц in Russian

This paper discusses the syntax and semantics of the Russian coordinating conjunction/particle ni in the light of Mitrovic's theory of conjunction. The crucial observation is that ni, although unable to function as an additive particle (in contrast to the positive i), can occur embedded into each of the conjuncts provided that it occupies parallel positions in the conjuncts. I suggest that the lack of particle uses of ni is due to its need to check the [+npi] feature, which can be done via the J(unction) head with the same feature. Assuming that ni is decomposable into ne 'not' + i, the parallelism requirement is explained if the post-syntactic lowering of ne (and possibly some preceding material, hence the embedded occurrences of ni) onto each of the conjuncts is stipulated.

Key words: coordination, negation, pied-piping, clitics, NPI, Phonological Form, Across-the-Board movement.



Introduction

In a series of publications [Mitrovic, 2014; Mitrovic, Sauerland, 2014], Moreno Mitrovic has been defending a unified structure for conjunctions and disjunction across languages, whether iterated (as in the Russian i Vasja, i Petja `both Vasya and Petya') or not. Importantly, his proposal explains why in some languages the same lexical items are used both as coordinating conjunctions and as additive particles. By and large, Russian coordinators fit the predictions of this account rather neatly. However, there are unresolved issues, which the present paper will try to bring to the foreground.



Main material

The data that will be in the centre of our attention concern the syntax of the Russian NPI conjunction/particle ni. Its distribution is more complex than predicted by Mitrovic's theory alone as it differs from that of its positive counterpart i. In particular, it has no additive uses; on the other hand, it shares with i the ability to occur embedded into each of the constituents it conjoins, as in (1).

(1) Я	не хочу	ни	половины	счастья,
Ja	ne xocu	ni	poloviny	scast'ja,
I	NEG want	Ng	half		happiness,
Ни	половины	горя	не	хочу! (Ye. Yevtushenko)
Ni	poloviny	gorja	ne	xocu!
Ng	half	woe	NEG	want!

`I want neither half-happiness nor half-woe!'

I will discuss the explanatory possibilities for both empirical points and . connect my proposal for cases like (1) to similar data featuring the negative particle ne.

The paper is structured as follows. In Section 2 I review Mitrovic's theory of coordination. In Section 3 the relevant Russian data are presented. Section 4 discusses the possible ways of explaining the data. Section 5 concludes.

Mitrovic's theory of coordination

According to Mitrovic, the structure of coordination is universal cross- linguistically. Apart from the conjoined constituents themselves, a coordinate structure minimally contains two p heads (after mo, which serves this function in Japanese), one for each conjunct, and one J head (for junction). Across languages, or sometimes within a given language, there are several options as to which of the ingredients are overt and which are silent. For instance, for Russian one may have either (2a) or (2b).

(2) a.	и Маша,			и	Даша
	[jp [pp i Masa], [j,	0	[pp	i	Dasa]]]
	p M.	J°		p	D
	`both Masha and Dasha.'
b.	Маша	и			Даша
	[jp [pp 0 Masa] [j,	i	[pp	0	Dasa]]]
	p M.	J°		p	D
	`Masha and Dasha.'

Here is how this syntax is mapped onto semantics. First, each name undergoes a type-lift (known as ident from [Partee, 1987]) which maps an individual to the set of individuals identical to it. Second, each type- lifted name combines with its p head, yielding the set of sets whose subset is the type-lifted individual. Finally, the two sets of sets are intersected, which is the semantic import of the J head.

(3) a. ident(§Masd§) = {x | x = Masha}

Ы ({x | x = Masha}) = { S | {x | x = Masha} C S }

[J]({ S | {x | x = Masha} C S }, { S | {x | x = Dasha} C S }) =

= { S | {x | x = Masha} C S } A { S | {x | x = Dasha} C S } =

= { S | {x | x = Masha} C S & {x | x = Dasha} C S } =

= Masa i Dasa]

The step (3b), which gives the denotation of p Masa (realised as i Masa or as 0 Masa), is applicable without any changes to cases like (4), where i is used as an additive particle; on Mitrovic's analysis, what (4) says is that the set of those who came has the set {Masha} among its subsets.5

(4) Пришла	и	Маша.
Prisla	i	Masa.
came	p	M.
`Masha came, too.'

Mitrovic's account predicts that the variety of coordinator lexemes should be limited to two types [Mitrovic, Sauerland, 2014, s. 43]: “J-type” coordinators, which are able to conjoin propositions but cannot double or function as additive particles, and “p-type” coordinators, which are not suited for propositional coordination but can double and can have particle uses, as in (4).6

Ni and the constraints on its use

In general, ni is a negative polarity item (NPI), i.e. it is licensed by negation. This is true of both major groups of uses ni has, which I will call conjunctive, (5) and quantificational, (6)-(7). The latter type may therefore be called a negative concord quantifier (see [Werle, 2002, van der Auwera, Alsenoy, 2016] for the term).

(5) Не	люблю	ни	улыбок,	ни	слез. (G. Ivanov, RNC7)
Ne	ljublju	ni	ulybok,	ni	slez.
neg	like.lSG	Np	smiles	Np	tears

`I like neither smiles nor tears'

5 This is not the final analysis Mitrovic and Sauerland give for sentences like (4), but it suffices here to illustrate the point, i.e. the unified treatment of conjunctive and additive uses.

О 6 What Mitrovic's system does not make clear - all diachronic considerations in his papers notwithstanding - is why Slavic languages employ the same lexical item (e.g. Russian i) for both J and p. Examples marked with “RNC” are taken from the Russian National Corpus, http://ruscorpora.ru.

(6) В	доме	не	раздавалось	ни	(одного)	звука.
V	dome	ne	razdavalos'	ni	(odnogo)	zvuka.
in	house	neg	be.heard.psT	Ng	single	sound

`Not a sound was to be heard in the house.'

(7) В	доме	не	было	ни	(*одной)	души.
V	dome	ne	bylo	ni	(*odnoj)	dusi.
in	house	NEG	was	Ng	single	soul



`There was no one in the house' (odnoj OK if souls, not people, intended).

One diagnostics that may be used to distinguish between the two types is J-type coordination: quantificational uses are easily conjoinable with i `and', whereas conjunctive uses resist the insertion of i

(8) В	доме	не	было ни	одного	человека
V	dome	ne	bylo ni	odnogo	celoveka
in	house	NEG	was Ng	single	man
и	ни	одной	собаки.
i	ni	odnoj	sobaki.
J	Ng	single	dog

`There was neither a man nor a dog in the house.'

(9) В	доме	не было	ни	человека,
V	dome	ne bylo	ni	celoveka,
in	house	NEG was	Ng	man
(и)	ни	собаки.
(-)	ni	sobaki.
J	Ng	dog

`Neither the [definite] man nor the [definite] dog was in the house.'

As can be seen, the list of the uses of ni does not include additive particle uses, which would be equivalent to the English `not... either'.

Whereas we find i in positive contexts (4), ni is blocked in the corresponding negative contexts and i is used instead:



(10) a. *Не	пришла	HU	Маша.
*Ne	prisla	ni	Masa.
neg	came	NЦ	M.
b. Ne	prisla	i	Masa.
neg	came	ц	M.

`Masha didn't come either.'

However, the ц status of ni is confirmed by the possibility of co-occurrence with i, which is permitted with non-referential (or non-nominal) conjuncts:

(11) Ты	не	бойся	ни	жары	и	ни	холода.
Ty	ne	bojsja	ni	zary	i	ni	xoloda.
you	neg	be.afraid.of	NЦ	heat	J	NЦ	cold

`Do not be afraid either of heat or of cold.' (V. Lebedev-Kumach)

(12) ...она	чувствовала себя	такой	одинокой,
.ona	cuvstvovala sebja	takoj	odinokoj,
she	felt	refl	so	lonely,
что	даже	не смела	ни	притворяться
cto	daze	ne smela	ni	pritvorjat'sja
cmpl	even	neg dared	NЦ	feign
и ни	скрывать	свою	боль. (Z. Gippius)
i ni	skryvat'	svoju	bol'.
J NЦ	hide	refl.poss	pain

`.. .She was feeling so lonely that she did not even dare to feign or to hide her pain.'

This is apparently the only configuration in Russian where both J and ц are overt, whereas across Slavic there are other examples, including (13) where the same word i serves both functions.

(13) southeastern Macedonian i Roska i I Ivan

I ц R. J ц `both Roska and also Ivan.' [Mitrovic, Sauerland, 2014, s. 44]

As a coordinating conjunction, ni may occur at the left periphery of each of the conjuncts (see e.g. (9)), but, crucially, this is not the only option:

it may also be embedded into each of the conjoined phrases, as in (1) as well as (14)-(15).

(14) Там	нет	ни		укреплений
Tam	net	ni		ukreplenij
there	no	Ng		fortifications
там	нет	ни		планировки	такой	постоянной. (RNC, oral)
tam	net	ni		planirovki	takoj	postojannoj.
there	no	Ng		layout	such	permanent

`There are no fortifications, there is no permanent layout there.'

(15) Это	страшная	штука,	когда	против	тебя
Eto	strasnaja	stuka,	kogda	protiv	tebja
this	terrible	thing	when	against	you
тупая,	свиная	со	щетиной	сила,
tupaja,	svinaja	so	scetinoj	sila,
stupid	pig-like	with	bristles	force
неуязвимая ни	для	логики,	неуязвимая
neujazvimaja ni	dlja	logiki,	neujazvimaja
impervious Ng	for	logic	impervious
ни	для эмоций.
ni	dlja emocij.
Ng	for emotions

`It's a terrible thing when there's a blind, pigheaded, pig-bodied force pitted against you, impervious to logic and emotion.'

(Strugatsky brothers, tr. A. Stone Nakhimovsky & A. Nakhimovsky)

There are two constraints on embedding. First, ni must be placed symmetrically within each conjunct (cf. the oddness of (16b)). This distinguishes ni from either, which can attach to a whole clause or to its subconstituent (“R[ight]-either”), provided that the constituent includes the contrastive focus (17). Therefore, there is little hope to adopt for ni the explanation given by den Dikken [2006] for either-float, which is that either originates attached to contrastive focus and optionally moves leftwards, so that when it does not move, “R-eithef' obtains.

(16) a. Нет	ни	от	отца	письма,
Net	ni	ot	otca	pis'ma,
no	Ng	from	father	letter,
нет	ни	от	матери	открытки
net	ni	ot	materi	otkrytki.
no	Ng	from	mother	postcard

`There is neither a letter from Dad nor a postcard from Mom.'

b. ??Нет	ни	письма	от	отца,
??Net	ni	pis'ma	ot	otca,
no	Ng	letter	from	father,
нет	письма	ни	от	матери.
net	pis'ma	ni	ot	materi.
no	letter	Ng	from	mother

Intended: `There is no letter either from Dad or from Mom.'

??Нет	письма	ни		от	отца,
??Net	pis'ma	ni		ot	otca,
No	letter	Ng		from	father,
нет	открытки	ни	от	матери.
net	otkrytki		ni	ot	materi.
no	postcard		Ng	from	mother

(17) (Either) John (either) ate rice or he ate beans.

[den Dikken, 2006, s. 690]

John either ate either [rice]CF or he ate [beans]CF

Either John either ate either [rice]CF or he ate [beans]CF

Second, not only the conjoined structures should be parallel but also the lexicalisation of the part preceding ni should be identical in all conjuncts. I Identity is observed in (16a) but violated in (16c), which results in unacceptability.

This may be dictated by the nature of g, a focus particle that requires that the material in its clause with which it does not associate be [+given]. We will see below that an alternative (or rather complementary) explanation for the identity constraint can be given in terms of post-syntactic doubling.

Towards an analysis

Let us summarise the peculiarities of ni that require a special analysis not reducible to Mitrovic's account of coordination and conditions on NPI licensing.

Pairing requirement, (10a): ni cannot be used as an additive particle and can only serve as iterated conjunction (quantificational uses aside).

Symmetry requirement, (16c): if ni is embedded into conjuncts, it should occupy the same position within all conjuncts.

Sameness requirement, (16b): if ni is embedded into conjuncts, the parts of the conjuncts preceding ni should be identical.

Given the pairing requirement, embedded uses cannot be analysed as mere juxtaposition of clauses containing each a token of additive ni. In the remainder of this section I will consider alternative explanations for the constraints.

The pairing requirement

The pairing requirement does not hold for i, as (18) witnesses:

(18) Не	пришел	Петя,	не	пришел	{и,	*ни}	Вася.
Ne	prisel	Petja,	ne	prisel	{i,	*ni}	Vasja.
neg	came	P.	neg	came	p	Ng	V

`Petya didn't come; neither did Vasya.'

Therefore, I conclude that there is something specific for ni that creates the constraint. Assuming the structure for conjunction as in (2), I suggest that ni...ni is licensed not by negation directly but rather by a special feature (e.g. [+npi]) of the corresponding J head, itself licensed by negation. (The J head need not, however, be lexically different from the default, as the use of i in (11)-(12) makes clear.) Given this, in the absence of (overt or covert) J ni cannot be licensed. This rules out additive uses of ni. Its positive counterpart i has no such restriction and is therefore available as a discourse particle or as a conjunction, both under negation and elsewhere.

Symmetry and sameness

As regards the symmetrical placement of ni in and the sameness of the material (if any) preceding ni, ideally one would like those requirements to follow from one's account of the embedded uses of ni. In other words, first one needs to explain how it is possible at all for ni to occur not at the periphery of the constituent it operates on but rather somewhere in the middle of it.

A problem with J above ni

One simple idea would be to argue that in (1) and (14) the licensing needs of ni are satisfied by a silent J head scoping over both conjoined clauses. Although the tokens of ni are embedded into their clauses, they are still within the scope of J[+npi] and are therefore licensed. Moreover, this seems to be corroborated by the fact that too deep embeddings (crossing a clause boundary) are disallowed:

russian semantic conjunction ni

(19) *Она	не	сказала,	что	пришел	ни	Вася,
*Ona	ne	skazala,	cto	prisel	ni	Vasja,
he	neg	said	cmpl	came	Ng	V.
она	не	сказала,	что	пришел	ни	Петя.
ona	ne	skazala,	cto	prisel	ni	Petja.
he	neg	said	cmpl	came	Ng	P.

Intended: `Neither about Vasya nor about Petya did she say they will come.'

However, a problem that arises immediately is that there is no way to ensure that the silent J will have the desired [+npi] feature as J is not in the scope of either of the negations in (1) and (14).

The same problem infects a solution based on the idea that structures like (1) and (14) are the result of (parallel) topicalisation in more basic configurations, e.g. ni ukreplenij [ tam net], niplanirovki takojpostojannoj [тоис tam net] for (14). This granted, the “more basic” configuration is again one where ni... ni conjoins whole clauses; thus the null J is not in the scope of negation and thus cannot bear [+npi]. (But see Section 5 for the attested data of this sort, indicating that some account of those is needed anyway.)

Embedded ni as a post-syntactic effect

A different line of explanation starts from the observation that the Russian negative particle ne is syntactically “distributed over” conjunction in some cases, including those often (although not unanimously) characterised as “expletive negation” [Paducheva, 2014], such as (21).

132

(20) ... невозможно .nevozmozno	не ne	ужаснуться uzasnut'sja	и i	не ne	восхититься vosxitit'sja
impossible	NEG	get.horrified	J	NEG	admire

одновременно. (RNC)

odnovremenno.

Simultaneously

`...So that one cannot but get horrified and admire at the same time.'

(21) Боишься,	как бы	кто-нибудь	не	пришел
Bois'sja	kak by	kto-nibud'	ne	prisel
be.afraid.2sG	CMPL	who-NIBUD'	NEG	came
и не i ne J NEG	отнял otnjal took.away	его у ego u he. ACC PREP	тебя. tebja. you.GEN.SG

`You are afraid that someone might come and take him away from you.'

What (20) says is that one is bound to get horrified as well as to admire, so the negation must be interpreted above the conjunction. What is going on in (21) is less clear, but here too, coming in and taking away are seen as parts of the same complex situation. To the best of my knowledge, such cases of negation doubling have not been previously investigated. A similar phenomenon (22) is discussed in Meyer and Sauerland [2016], where such cases are declared instances of Across-the-Board (ATB) movement [de Vries, 2017] of may (and consequently of Jane) without the deletion of the second lower copy (23); only the higher copies are interpreted.

Jane may sing or Jane may dance.

A possible interpretation: `Jane may (sing or dance)' (? > V)

Jane2 may1 [[ Jane2 may1 sing ] or [ Jane2 may1 dance ]].

The obvious difference between (22) and (20)-(21) is that may is syntactically allowed and interpretable (albeit differently) both with and without movement, whereas ne in (20)-(21) must be doubled for the sentences to be grammatical. Nevertheless, an analysis along these lines may tentatively be proposed. Alternatively, instead of the usual leftward ATB movement one may stipulate rightward ATB movement of the negation from a position above the conjunction onto each of the conjuncts. Such movement would be problematic at the level of syntax, but given that ne is a clitic and needs a host, post-syntactic movement [Embick, Noyer, 2001] is a more plausible option. In search of a host, ne stumbles upon a conjunction, whose all parts undergo various operations in parallel - the property known as the Coordinate Structure Constraint. Therefore, a copy of ne lowers down onto each of the conjuncts.

Coming back to embedded ni, as in (1), (14) and (15), I suggest that ni is decomposable into negation and i (following the spirit, if not the letter, of Mitrovic's approach to coordinators). The negation, perhaps expletive (as ni and i are often interchangeable), originates higher than the conjunction

but moves downwards in search of a host and merges with each of p's. The material preceding the negation may optionally also be moved, which gives the impression of the embedded ni.

tam net ne [[i ukreplenij], [i planirovki...]] ^

tam net ne [[(ne + i) ukreplenij], [(ne + i) planirovki.]] ^

tam net ne [[ni ukreplenij], [ni planirovki.]] ^ (optionally)

tam net ne [[tam net ni ukreplenij], [tam net ni planirovki.]] = (14)

As long as the same material gets copied, the parallelism and sameness requirements are satisfied.

5. Conclusion

In the preceding sections I have pointed out several idiosyncratic features of the Russian p-class lexical item ni. Like other p's, it can serve as an iterated conjunction (licensed by negation, although more on that below), but it lacks additive uses. In spite of that, as a conjunction it need not occupy a peripheral position in a conjunct but may be embedded into it, provided that the embedding is symmetrical in all conjuncts and that the material before ni is always the same. To account for those features, I suggested that the licensing of ni is indirect and proceeds via the corresponding J head, in the absence of which ni is not licensed (thus disallowing additive uses). Furthermore, ni is decomposable, and the negation which it has as a part moves post-syntactically, in some cases together with some preceding material, which creates the observed syntactic and lexical parallelism in the cases of embedded ni.

This said, there remain unresolved issues (see also footnote 5). As noted above, p-type coordinators across languages are predicted to be limited to DP conjunctions and unable to conjoin at the propositional level. However, this does not seem to hold either for i or for ni in Russian: (25) is just as acceptable with iterated i as it would be without the first token of it; in (26) iterated ni conjoins two VPs; and in (27) it conjoins two clauses.

(25) И	солнце	всходило,	и радуга	цвела. (L. Derbenyov)
I	solnce	vsxodilo,	i raduga	cvela.
p	sun	rose	p rainbow	bloomed

`The sun used to rise, and the rainbow used to bloom.'

(26) Сидит, на	свет	не	глядит,	ни	ухи
Sidit,	na	svet	ne	gljadit,	ni	[VP uxi
sits	on	world	NEG	looks,	Ng	fish-soup
не	ест,	ни	вина	не	пьет.	(A. Ostrovsky)
ne	est],	ni [VP	vina	ne	p'et].
neg	eats,	Ng	wine	NEG	drinks

`[He] is sitting still, not looking onto the world, neither eating his fish-soup nor drinking wine.'



(27) Ни он	ничего не	сказал, ни	он а	не	смогла
Ni on	nicego ne	skazal, ni	ona	ne	smogla
Ng he nothing NEG said Ng произнести слов прощания. proiznesti slov proscanija. utter words farewell.GEN	she	NEG	could

`Neither he said anything nor she could utter the words of farewell.' (from an online dictionary)

What is more, (26)-(27) pose a licensing problem:11 assuming that ni is not itself negative but rather an NPI and that there is no unpaired ni, we expect to find a negative element licensing both tokens of ni at once; but there is no such element in (26) or (27). Moreover, ni. ni conjoins the clauses that are already negative, i.e. the conjunction scopes above the negations; i. i would be equally acceptable here.

(28) Я	не	заставлял	его	ехать	ни	в	Москву,
Ja	ne	zastavljal	ego	exat'	ni	v	Moskvu,
I NEG ни в ni v Ng to	forced Петербург. Peterburg. St. Petersburg	him	go	Ng	to	Moscow,

`I didn't force him to go either to Moscow or to St. Petersburg' = `For both cities x: I didn't force him to go to x.' [Bylinina, 2003]

Given that ni scopes above negation in (26)-(27), I suggest that examples like (28) belong to the same class; non-finiteness of the complement clause facilitates the LF movement of [ni. ni... ] (cf. Bylinina's (2003) observation that пі-conjunctions in object position are easier to extract than those in subject position). As for now, there seems to be no explanation for how ni is licensed in such configurations.



References

1. Abels, 2002 - Abels K. Expletive (?) negation. Proceedings ofFASL 10. Toman J. (ed.). Bloomington, ID, 2002. Pp. 1-20.

2. van der Auwera, Alsenoy, 2016 - van der Auwera J., Alsenoy V. On the typology of negative concord. Studies in Language. 2016. Vol. 40. No. 3. Pp. 473-512.

3. Bylinina, 2003 - Былинина Е.Г. О синтаксисе отрицания в русском языке: отрицательные местоимения и конструкция «ни X, ни Y» // Материалы Международной конференции «Диалог - 2003». М., 2003. [Bylinina E.G. On the syntax of negation in Russian: Negative pronouns and the construction ni X, ni Y. Proceedings of the international conference “Dialogue 2003”. Moscow, 2003. URL: http:// www.dialog-21.ru/media/2617/bylinina.pdf.]

4. den Dikken, 2006 - den Dikken M. Either-float and the syntax of co-or-dination. Natural Language & Linguistic Theory. 2006. Vol. 24. No. 3. Pp. 689-749.

5. Embick, Noyer, 2001 - Embick D., Noyer R. Movement operations after syntax. Linguistic Inquiry. 2001. Vol. 32. No. 4. Pp. 555-595.

6. Meyer, Sauerland, 2016 - Meyer M.-C., Sauerland U. Covert Across-the-Board Movement Revisited. Presented at NELS 47. 2016.

7. Mitrovic, 2014 - Mitrovic M. Morphosyntactic atoms of propositional logic: a philological programme. PhD diss. University of Cambridge, 2014.

8. Mitrovic, Sauerland, 2014 - Mitrovic M., Sauerland U. Decomposing coordination. Proceedings of NELS 44. Vol. 2. Iyer J., Kusmer L. (eds.). 2014. Pp. 39-52.

9. Paducheva, 2014 - Падучева Е.В. Эксплетивное отрицание и семантика союза пока // Язык, константы, переменные: памяти Александра Евгеньевича Кибрика. СПб., 2014. С. 339-350. [Paducheva E.V. Jazyk, konstanty, peremennye: Pam- jati Aleksandra Evgen'evica Kibrika. Daniel M.A. et al. (eds.). St. Petersburg, 2014. Pp. 339-350.]

10. Partee, 1987 - Partee B.H. Noun phrase interpretation and type-shifting principles. Studies in discourse representation theory and the theory of generalized quantifiers. Groenendijk J. et al. (eds.). De Gruyter, 1987. Pp. 115-143.

11. de Vries, 2017 - de Vries M. Across-the-Board Phenomena. Blackwell Companion to Syntax. 2nd ed. Everaert M., van Riemsdijk H. (eds.). Wiley, 2017.

12. Werle, 2002 - Werle A. A typology of negative indefinites. Papers from the 38th Meeting of the Chicago Linguistic Society. Vol. 2. 2002. Pp. 127-143.

Размещено на Allbest.ru

...