CC BY
21
Competition between
'Who' and 'Which' in Slavic Light-Headed Relative Clauses*
Maria A. Kholodilova
Institute for Linguistic Studies of the Russian Academy of Sciences; National Research University Higher School of Economics St. Petersburg, Russia
Конкуренция местоимений 'кто' и 'который' в славянских относительных предложениях без существительного в вершине
Мария Александровна Холодилова
Институт лингвистических исследований РАН; Национальный исследовательский университет "Высшая школа экономики" С.-Петербург, Россия
Abstract
The relativization systems of most Slavic languages include relative pronouns that can be conventionally labelled as 'who' and 'which' and differ in a number of logically independent parameters (etymology, animacy, grammaticality of attributive contexts, and morphological distinction for number and gender). Prior research has shown that the choice between 'who' and 'which' in Slavic languages is largely dependent on the head type. Some of the languages allow the 'who' pronouns to be used with pronominal heads, but not with nouns in the head, while in others, the pronominal heads in the plural are also ungrammatical with the pronoun 'who.'
* This study was supported by the Russian Foundation for Basic Research (project 14-06-31212). I wish to thank Alexander Piperski and the anonymous reviewers for their comments. All remaining errors are my own.
This is an open access article distributed under the Creative Commons Attribution-NoDerivatives 4.0 International
The present study aims to complement the available qualitative data on the distribution of the relativizers with quantitative data and to propose a unified account for all the observed tendencies. A corpus-based study was conducted in order to establish language-internal statistical tendencies comparable to the known grammatically restrictions. The results show much agreement between the qualitative and quantitative tendencies. Thus, the head 'those,' unlike the head 'that,' is incompatible with the relativizer 'who' in Slovak, Polish, Upper Sorbian, and Lower Sorbian languages, while the same tendency is quantitative in Czech, Slovene, Serbo-Croatian, Ukrainian, Belarusian, and the older varieties of Russian. Corpus data suggest that there is also a stronger tendency for the relative pronoun 'who' to be avoided with the head 'those' than with the head 'all.' One more relevant parameter is the semantic type of the clause, maximalizing semantics being the preferred option for 'who.' I suggest that all these and some other tendencies can be subsumed under a macro-parameter of the extent to which the head is integrated into the relative clause.
Key words
Slavic languages, relativization, relative pronouns, interrogative pronouns, light-headed relatives, free relatives
Резюме
В большинстве славянских языков система релятивизации включает относительные местоимения типов 'кто' и 'который', противопоставленные по ряду признаков (этимология, одушевленность, допустимость атрибутивных контекстов, состав парадигмы). Из существующих исследований известно, что на выбор между этими местоимениями во многом влияет тип вершины, в частности, местоимения типа 'кто' более допустимы в некоторых языках при вершинах, выраженных местоимениями ('те', 'все', 'тот' и др.), а из них — при вершинах в единственном числе. Целью настоящего исследования является дополнение известных тенденций по распределению релятивизаторов количественными данными и обобщение доступных данных по допустимости и частотности относительных местоимений в рамках единого подхода. Для получения количественных данных, сопоставимых с известными различиями по допустимости, статистически значимые количественные различия устанавливаются на материале корпусов славянских языков. Согласно проведенному исследованию, количественные и качественные тенденции по распределению местоимений 'кто' и 'который' хотя бы частично подчиняются общим закономерностям. Так, в словацком, польском и лужицких языках вершина 'те' в отличие от вершины 'тот' не допускает использования местоимения 'кто', тогда как в чешском, словенском, сербо-хорватском, украинском, белорусском и в русском языке XVIII века та же тенденция проявляется в количественных предпочтениях. Кроме того, корпусные данные позволяют установить различие между вершинами 'все' и 'те', которое не приводит к грамматическим запретам ни в одном из исследованных языков. Еще одним параметром, влияющим на распределение относительных местоимений, оказывается семантический тип относительной клаузы: местоимение 'кто' наиболее предпочтительно в максимализирующих относительных предложениях и наименее предпочтительно в нерестриктивных. Можно предположить, что эти и некоторые другие тенденции в дистрибуции местоимений 'кто' и 'который' определяются одним макропараметром — степенью вложенности вершины в относительную клаузу.
Ключевые слова
славянские языки, релятивизация, относительные местоимения, вопросительные местоимения, бессубстантивные относительные предложения, безвершинные относительные предложения
1. Introduction
Slavic languages exhibit considerable variation in the use of relativizers, which manifests itself both language-internally and cross-linguistically, see among others [Bauer 1967; Бауэр 1967; KrizkovA 1970; Got^B, Friedman 1972; Murelli 2011]. A particular problem within the variation system concerns the competition of the relative pronouns that can be labelled as 'who' and 'which.' The former label is meant to cover the relativizers which directly continue interrogative pronouns of the protolanguage with the meaning 'who,' refer to people, do not decline for number and gender, and cannot be used attributively. The latter label refers to relativizers of a different origin, applicable to both animate and inanimate objects, which decline for number and gender, and can, under some conditions, be used attributively. Relative pronouns of these classes are present in most Slavic languages [Got4B, Friedman 1972].
Language-internally, the distinction between 'who' and 'which' gives rise to a competition between syntactic synonyms, as illustrated in (1) by two Serbian translations from the same Russian original. The translation in (1a) contains the relative pronoun koji 'which,' while a similar meaning in (1b) is rendered by ko 'who.'
(1) Mikhail Bulgakov. The Master and Margarita (Parasol)
Serbian1
a. I upravo tu Ivan Nikolajevic konacno izgubi onoga and exactly here Ivan Nikolaevich finally lost him
koji mu je bio toliko potreban. which for.him was so necessary
b. I eto tu Ivan Nikolajevic konacno izgubi onoga and this here Ivan Nikolaevich finally lost him
ko mu je bio toliko potreban. who for.him was so necessary
'And it was here that Ivan Nikolaevich definitively lost him whom he needed so much.'
1 All originally Cyrillic examples in the paper are transliterated. Examples are given in professional translations extracted from parallel corpora wherever possible.
Cross-linguistically, the variation results in differences between Slavic languages with respect to the grammatically and frequency of relative pronouns in different contexts. A preliminary idea of the variation is given by (2), where Slavic translations of a single passage contain either 'who' (2a) or 'which' (2b). The original Russian extract is given under (2a) and contains the pronoun 'who.'
(2) Mikhail Bulgakov. The Master and Margarita (RNC)
'Those who saw him for the first time. . .'
The distribution of Slavic 'who' and 'which' has been surveyed in the literature, most notably by KrizkovA [1970] for all the modern standard Slavic languages, see also [Browne 1986; Kordic 1994; 1999; Mitrovic 2012; Павловиъ 2012] for Serbo-Croatian, [Zubaty 1918; Svoboda 1967; Kar-lik 1988] for Czech, [Darovec 1985: 109-111; PodhajeckA 2010; eadem 2012] for Slovak, [Polanski 1967: 78-79; Fasske, Michalk 1981: 615-617, 626627] for Upper Sorbian, [Janas 1976: 184-185] for Lower Sorbian, and [Зализняк, Падучева 1975; Spencer 1993: 38-41; Никунласси 2008] for Russian. However, most of these studies, with the notable exception of [Kordic 1994; idem 1999: 196-197], are based only on qualitative data. To the best of my knowledge, no prior study has attempted to consider this whole subsystem of variation between relativizers in view of both qualitative and quantitative data. At the same time, such data could shed some light on the consistency of relative clause patterning across Slavic languages.
The primary objective of the present study is therefore to bring together the newly acquired quantitative data and the available qualitative data. The competition between 'who' and 'which' is most consistently observed in light-headed relatives, i.e., relative constructions without a noun in the head,
a. 'who'
te, kto videl ego vpervye...
tyja, xto bacyujaho upersynju
ty, kdo ho videli poprve
those who saw him first
Russian
Belarusian
Czech
b. 'which'
ti, jaki bacyly joho vperse...
ci, ktorzy widzieli go po raz pierwszy
oni koji su ga vidjeli prviput
those which saw him first
Ukrainian
Polish
Croatian
whether overt or elided.2 I will therefore primarily confine the discussion to this class of relative constructions, even though some remarks on a more general picture will also be given.
The study needs a consistent definition of the notions 'who' and 'which,' which is introduced in part 2. I go on to sketch the qualitative data on the distribution (part 3) and present the quantitative data (part 4). A discussion follows in part 5 to specify the observed tendencies. In part 6, I summarize the tendencies and propose a unifying macro-parameter of variation, namely the extent to which the head is integrated into the relative clause.
2. The definition of 'who' and 'which' pronouns
As stated in the introduction, I base the distinction between 'who' and 'which' on several properties. The preliminary list of properties and sources of information follows below:
1) whether the relativizer directly continues an animate interrogative pronoun ('who') of a proto-language (Proto-Slavic *ktfo in the case of Slavic languages) [Vaillant 1958; Bauer 1967; Tpybaheb 1987: 125];
2) synchronic semantic restrictions on its use with non-human referents [KrizkovA 1970];
3) the presence of number and gender distinctions in its paradigm (individual grammatical descriptions);
4) the grammaticality of attributive uses [KrizkovA 1970].
It is important to note that the terms 'who' and 'which' are conventional labels for comparative concepts in the sense of [Haspelmath 2010] and need not reflect the properties of the English pronouns who and which.'3
Two common Slavic patterns with respect to the properties listed above are presented in Table 1. I preliminarily define 'who' and 'which' pronouns as those which adhere to these patterns.4
2 The term "light-headed relatives" for the relative clauses without a noun in the head was coined by Citko [2000; 2004]. Another common term for this group of relative constructions is "false free relatives" [De Vries 2002].
3 The English who does happen to fall into the 'who' group, whereas which does not quite correspond to the set of features of the 'which' group because it does not have distinct plural forms.
4 An anonymous reviewer suggests Slavic possessive relative pronouns ('whose') should be regarded as possessive forms of the 'who' pronouns. My data suggest that these pronouns consistently show different syntactic properties and different patterns of use, see among others [Be^hhoba 1988; Kholodilova 2013]. They are therefore not included in this survey.
Table 1. The differences between the typical Slavic 'who' and 'which'
stems from the interrogative 'who' human reference only does not have number or gender distinctions cannot be used attributively
pattern 1 (~ 'who') + + + +
pattern 2 (~ 'which') - - - -
Although the proposed definition is new, the distinction itself is commonly drawn in papers specifically addressing relative clauses [KrizkovA 1970; Зализняк, Падучева 1975] as well as general grammatical descriptions of Slavic languages and is often construed as that between adjectival and substantive relativizers. I follow [Bauer 1967: 301] and [KrizkovA 1970: 13-17] among others in classifying Slavic relative pronouns in the manner reflected in Tables 2 and 3, and I attempt only to articulate the underlying intuition more precisely.
Two additional remarks are due.
First, a straightforward classification according to etymological origin is problematic for Upper Sorbian stoz 'who,' Bulgarian kojto 'which,' and Macedonian koj 'which' and kojsto 'which.' All these pronouns retain the etymological root which directly continues *kUo only in (some of) the forms other than the nominative [Vaillant 1958: 411, 417]. This complex etymology is reflected in Tables 2 and 3 as "+/-" for Upper Sorbian stoz 'who,' which does not have any restrictions specific to the use of the non-nominative forms, and "-/+" for Bulgarian and Macedonian pronouns, which only use this stem, unlike the nominative forms, in animate contexts. In other words, if these Bulgarian and Macedonian pronouns are viewed as single items, it is natural to assume that the non-animate accusative forms are either derived from the koj stem or lacking altogether, and in either case the etymologically animate stem is only used in a subclass of oblique forms. In all cases of an intermediate status, I consider this property non-decisive for the classification of the pronouns and rely on other features.5
Second, among the relativizers with the properties of 'which' pronouns, I consider in this study only those with high frequency.
Tables 2 and 3 summarize the data on the definitional properties of Slavic pronouns. I consider here only the modern standard varieties of the languages, while some deviations from these descriptions are reported for dialects and older varieties.
5 Note that this criterion is etymological rather than synchronical. Bulgarian,
Macedonian, and Serbo-Croatian pronouns classified here as 'which' can be used in the contexts of the interrogative 'who' in the modern language, but do not directly continue the Proto-Slavic *k"bto.
Table 2. Slavic 'who' pronouns
only stems from the interrogative 'who' no number or gender distinctions cannot be used attributively
human = 'who' = 'who' + smth.
East Slavic Russian kto + + + +
Ukrainian xto + + + +
Belarusian xto + + + +
Polish kto + + + +
West Slavic Czech kdo + + + +
Slovak Upper Sorbian kto stoz + +6 + +/- + ? + +
Lower Sorbian chtoz + + ?7 +
South Slovene kdor + + + +
Slavic Serbo-Croatian (t)ko + + + +
Table 3. Slavic 'which' pronouns
only human stems from the interrogative 'who' no number or gender distinctions cannot be used attributively
Russian kotoryj - - - -
East Slavic Ukrainian Belarusian kotryj jakyj katory jaki - - - -
Polish ktory - - - -
West Slavic Czech Slovak Upper Sorbian ktery ktory kotryz - - - -
Lower Sorbian kotaryz - - - -
Slovene kateri - - - -
South Serbo-Croatian koji - - - -
Slavic Macedonian koj kojsto - -/+ -/+ - -
Bulgarian kojto - -/+ - -
6 According to [Bartels, Spiess 2012], the relative 'who' can refer to inanimates in older literary Lower Sorbian and non-standard Upper Sorbian, unlike the present-day standard languages.
7 Mucke [1891: 428] provides a paradigm of the interrogative chto, which includes instrumental plural and dual forms; however, these forms are not mentioned by Janas [1976: 179].
Macedonian koj and kojsto, as well as Bulgarian kojto, do not have a counterpart belonging to the 'who' group and are therefore not considered below in any detail.
As is evident from the data in Tables 2 and 3, the four properties strongly tend to pattern together in Slavic languages. It is therefore impossible to tear them apart, building on the Slavic data only. Although a comparative concept based on a single property would be more desirable, choosing any one of the properties listed above in this study would be largely arbitrary and could yield incorrect conclusions.
3. Qualitative data on the distribution of Slavic 'who' and 'which'
To the best of my knowledge, the most systematic overview of the distribution in question is provided by KrizkovA [1970], who describes the use of 'who' and 'which' in questions (3), correlatives (4), and postnominal relative constructions headed by the pronouns 'someone,' 'nobody,' 'each,' etc. (5), 'that' (6), 'all' or 'those' (7), and nouns (8). The paper also contains some data on constructions with 'first,' 'last,' 'that' in the feminine, and personal pronouns in the head, as well as cleft constructions; the evidence of this data, however, is more sketchy, and I leave it out in the following discussion.
(3) question, Carlo Collodi. Pinocchio (RNC)
Kto vas oskorbljaet? Russian
Kto wam ubliza? Polish
Tko vas vrijeda? Croatian
who you insults
'Who is insulting you?'
(4) correlative, Ivo Andric. The Bridge on the Drina (Parasol) Xto ne bacyc' zaraz, toj nepabacyc' nikoli. Kto nevidi teraz, ten neuvidi nikdy. Ko ne vidi sada, taj nece nikad vidjeti. who doesn't see now that will never see 'The one who does not see now will never see.'
(5)
each' in the head, Ivo Andric. The Bridge on the Drina (Parasol) . . Koznomu, xto joho xotiv sluxaty . . . . . kazdemu, kdo ho byl ochoten poslouchat. . . . . vsakomur, kdor ga je hotel poslusati . . .
each.one who him wanted to listen . . . to everyone who wanted to listen to him.'
Belarusian
Slovak
Serbian
Ukrainian
Czech Slovenian
(6) 'that' in the head, Nikolai Ostrovsky. How the Steel was Tempered (Parasol)
a. . . . zvjarnuusja en da taho, kaho zvali Moc'kam . . . obratise on onome koga su zvali Mocka
turned he to the one whom called Motka
b. . . . priwobroci so k tomu, kotrehoz bechu Motku wolali . . . turned to the one which called Motka
'He addresses the one who was called Motka.'
(7) 'those' in the head, Umberto Eco. The Name of the Rose (Parasol)
a. I te, kto ubival obezumevsix kajal'scikov. . . and those who killed crazed penitents
b. A ci, ktorzy zabijali oszalalych pokutnikow. . . A oni koji su ubijali pomahnitale pokajnike. . . and those which killed crazed penitents
'And they who killed the crazed penitents. . .'
(8) noun in the head, Dan Brown. The Da Vinci Code (RNC) . . . vrag, o kotorom ee preduprezdali
. . . nepriatel', pred ktorym ju vystrihali . . . sovraznik, pred katerim so jo posvarili enemy about which they warned her
'. . . the enemy they had warned her about'
Belarusian Croatian
Upper Sorbian
Russian
Polish Serbian
Russian
Slovak
Slovene
Table 4 summarizes KrizkovA's [1970] data. Some of the data points are corrected and highlighted in bold.8
First, pace Krfzkova, Upper Sorbian does allow 'who' to be used with indefinites [Fasske, Michalk 1981: 616], cf. a corpus example:
(9) Upper Sorbian, Nikolai Ostrovsky. How the Steel Was Tempered (Parasol)
Ale bydlenje bese prozdne, a nichto tu njebese, kohoz by so woprasal but flat was empty and nobody there wasn't whom would ask
'But the flat was empty, and there was nobody to ask.'
Second, corpus data show that Lower Sorbian allows 'which' with the head 'that':
8 Kri'zkova [1970] sometimes does not state explicitly that one of the pronouns is ungrammatical in a given context, but she gives lists of languages using different strategies which could be expected to be exhaustive.
Table 4. The use of 'who' and 'which' in Slavic languages,
according to the data in KRiz kovA [1970] with minor corrections
questions correlatives 'someone,' 'nobody,' 'each' 'that' 'all,' 'those' nouns in the head
Russian who who/which who who/which who/which which//? who
Ukrainian who who who who/which who/which which/-who
Belarusian who who who who/which who/which which/}who
Czech who who who who/which who/which which9
Slovene who who which who/which who/which which
Polish who who who who/which which which
Lower Sorbian who who who/which who/which which which
Slovak who who who/which who/which which which
Upper Sorbian10 who who/which who/which who/which which which
Serbo-Croatian11 who/which who/which who/which who/which who/which which
(10) Lower Sorbian, James 2:13 (DOTKO)12
Njefmilny fiud fmejo ten, kotaryz fmilnofcz nejo zynit Merciless judgment will.have that which mercy didn't do 'For judgment is without mercy to the one who has shown no mercy.'
Third, the data on the heads 'somone,' 'nobody,' and 'each' in Lower Sorbian are missing from [Krizkovâ 1970]. Both 'who' and 'which' pronouns are attested in such contexts in DOTKO. Cf. also the following example:
(11) Lower Sorbian, Antoine de Saint-Exupéry. The Little Prince (ASPAC)
... mimo nekogo, z kotarymz by mogat se derje wulicowas ... without someone with which would be able well talk 'without anyone that I could really talk to'
9 Slovak and Polish examples of this kind can also be found, although they are extremely rare.
10 Czech examples of 'who' pronouns with a noun in the head can be found, but they are extremely infrequent.
11 'Who' is attested more widely in older literary Lower Sorbian and non-standard Upper Sorbian [Bartels, Spiess 2012: 227].
12 Lower Sorbian writing in DOTKO texts deviates from the modern orthography; see a description of the correspondences at http://dolnoserbski.de/korpus/psawopisne_ warianty.
Fourth, Slovene examples with 'which' following the head 'all' are also attested:13
(12) Slovene (Fida Plus)
Vsekakor pa vsi, kateri so sodelovali pripopravilu, zasluzyo pohvalo! anyway but all which participated in repairs deserve praise 'Anyway, everyone who took part in the repair deserves praise!'
Most of the data in the table is consistent with the hierarchy in (13), the positions to the left being increasingly preferred for the 'who' pronouns.
(13) questions > correlatives > 'someone,' 'nobody,' 'each'; 'that' > 'all,' 'those' > heads with nouns
The only exception to this rule is represented by Russian correlatives, which unexpectedly allow the pronoun 'which.' Three comments are due here. First, standard Russian does not allow 'which' in any of the relative constructions without a noun in the head, i.e., standard Russian data clearly fits into the hierarchy in (13). Second, 'which' in non-standard Russian also occurs in the "intermediate" group, namely, with the head 'each,' in a clearly non-elliptical context, and these examples are relatively acceptable, at least when the head is feminine, as in (14).14
(14) Russian, T. Putilova ("Komsomol'skaya Pravda," 2002)
Pravda, po ego slovam, on zenilsja na kazdoj, kotoruju celoval.
truth according.to his words he married on each.F which kissed 'However, according to him, he married every girl he kissed.'
Third, even though correlatives with 'which' are slightly more acceptable and frequent than this intermediate step, it is only true for the correlatives with the relative pronoun in the plural. However, comparable conditions are impossible for the group 'someone,' 'nobody,' and 'each,' and the data discussed in 5.2 show that plural number is an independent parameter which makes 'which' pronouns more acceptable.
It is therefore possible to suggest that the hierarchy in (13) only holds if other things are equal, i.e., if the features unspecified in the group descriptions coincide.
13 Slovak and Polish examples of this kind can also be found, although they are extremely rare.
14 The examples with 'nobody' and 'someone' are, however, ungrammatical, which suggests that this group is not quite homogeneous. I will not go into further details in this paper.
The greatest variation in the table is observed in the 'that' and 'all,' as well as 'those' columns, which is why further discussion is mainly confined to this class of light-headed relatives.
4. Quantitative data on the distribution of Slavic 'who' and 'which'
The basic idea behind this study is to propose a unified account for qualitative and quantitative data on the distribution of 'who' and 'which.' The desired quantitative data point must therefore be directly comparable to the qualitative data in Table 4. It is possible to establish such comparative data by looking at information on the presence of a quantitative tendency established language-internally, which can be compared to the qualitative tendencies summarized in (13).
A potentially attractive option for a quantitative cross-linguistic study is to increase comparability of the data by using parallel corpora, such as ASPAC, InterCorp, the parallel subcorpus of the RNC, or Parasol, as advocated for Slavic languages by von Waldenfels [2014]. However, this methodology seems less suited for this study. Using parallel texts does not seem to make the data of basic interest for this study any more reliable, specifically, the data on language-internal quantitative tendencies (or, to put it in other words, the differences between columns in Table 4). It could be more useful if the intended results included not only a comparison of contexts, but also a quantitative comparison of languages (i.e., differences between rows in Table 4). However, this aim does not seem to be fully achievable. Only one text in all the Slavic varieties considered in this study can be currently found in the corpora mentioned above. This is largely due to the fact that very few texts are available in Lower Sorbian. Moreover, even if Lower Sorbian is excluded from consideration, we still end up with only four texts. The constructions in question are relatively infrequent; therefore the small amount of data could simply be non-su fficient. To give one example, the four texts mentioned above with multiple translations contain only a total of nine examples of Russian relative clauses with the head 'all' immediately preceding relative pronouns.15 Moreover, the longest of these texts is of Russian origin, which would also be a major problem for this study. For instance, it is plausible that the sentences originally in Russian would be more likely to be calqued in languages which are closer to it structurally and lexically. Finally, if only a few translated texts are taken into account, it is likely that the differences "between rows of Table 4" will reflect the individual translators' preferences rather than the intrinsic properties of the languages.
15 I only consider the first Russian translation when more than one variant is available.
To sum up, consistent use of translated texts would not provide any advantages for the core of this study and could in some respects be harmful. I therefore do not intend to collect directly comparable quantitative data on the absolute frequency of relativizers in different languages, but only the data on the tendencies of their use. This implies, among other things, that not only the corpora and texts, but also the principles of the search, such as the search query and sampling, need not be exactly the same across the languages, as long as there is no independent reason to expect that these features can reverse the correlations.
Table 5 presents the data obtained in this study.
The first column corresponds to the surveyed languages. I do not consider modern Russian data, because kotoryj 'which' in modern standard Russian is used almost exclusively with nouns in the head, see, e.g., [Spencer 1993]. Surveying non-standard data poses additional problems related to the distinction between light-headed relatives and relatives with elided heads, which are extremely frequent in non-standard texts. However, to fill this gap, the data on Russian of the 18th century is included.16
The second column lists the pronouns in the head that were included in the search queries.17
The columns 'who' and 'which' contain the raw frequencies of the corresponding constructions. In all cases, the relative pronouns were included in the queries in the nominative. All the search results were inspected manually in order to exclude the relative clauses with elided nouns in the head. The next column conveys the ratio the 'who' constructions constitute among the relatives of both types.
The last column lists the corpora used in the substudies.
16 The legitimacy of including an earlier language variety in the sample, which was questioned by anonymous reviewers, derives from the following assumption. The aim of this study is to propose a generalization that would hold true for as many language varieties as possible. I therefore build a sample of language varieties which are, first, suitable for this study and, second, as distantly related as possible. In most cases, this sample happens to coincide with the list of modern standard Slavic languages, because they are distantly related and have large corpora. In Russian, however, the modern language does not satisfy the first condition, and therefore another variety must be considered. The choice of the century is largely random. A supplementary micro-diachronic study shows a gradual decline in the frequency of the 'which' pronoun in this construction over the last three centuries, and the end of the 18th century is a cutoff point as good as any other.
17 The preferable option was to search only for the heads in the nominative; however, this would not be sufficient for Slovene, in which both constructions under study are extremely infrequent, whereas the preferred option is the indeclinable relativizer. Moreover, the nominative forms were not considered in the Slovene sample of demonstrative pronouns because they do not differentiate between singular and plural.
Table 5. Quantitative data on the use of 'who' in different contexts
head 'who' 'which' ratio of 'who' Corpora
toj 'that' 180 8 1.0
Belarusian use 'all' 43 1 1.0
tyja 'those' 105 12 0.9 RNC,
toj 'that' 64 1 1.0 parallel subcorpus
Ukrainian usi / vsi 'all' 16 0 1.0
ti 'those' 32 11 0.7
tisti 'that' 10 0 1.0
Slovene vsi 'all' 4 4 0.5 FidaPLUS
tisti 'those' 0 9 0.0
ten 'that' 91 9 0.9
Czech vsichni 'all' 78 22 0.8 InterCorp
ti 'those' 54 46 0.5
tot 'that' 206 83 0.7
Russian of the 18th c. vse 'all' 3 7 0.3 RNC, main subcorpus
te 'those' 6 161 0.0
onaj 'that' 16 17 0.5
Serbo-Croatian18 svi 'all' 0 10 0.0
oni 'those' 0 94 0.0
ten 'that' 61 14 0.8
Slovak vsetci 'all' 0 11 0.0 InterCorp
ti 'those' 0 33 0.0
ten 'that' 53 47 0.5
Polish wszyscy 'all' 0 100 0.0
ci 'those' 0 100 0.0
ten 'that' 4 20 0.2
Lower Sorbian wsykne 'all' 0 9 0.0 DOTKO
te 'those' 0 71 0.0
ton 'that' 0 100 0.0
Upper Sorbian wsitcy 'all' 0 100 0.0 HOTKO
ci 'those' 0 100 0.0
18 Serbian and Croatian were surveyed separately; however, the results did not show any difference. The results in the table are taken from the Croatian subcorpus.
Table 6 is a reminder of the differences between 'that,' 'those,' and 'all' observed in the qualitative data. The vertical lines mark the cases where 'who' is acceptable in one of the contexts, but not the other. The relativizer 'which' is attested in all these languages with all three head types.
Table 6. Grammatically of 'who' in different contexts,
according to the data in KrIzkovA [1970] with minor corrections
'those who,' 'all who' 'that who'
Belarusian + +
Ukrainian + +
Slovene + +
Czech + +
Russian + +
Serbo-Croatian + +
Slovak - +
Polish - +
Lower Sorbian - +
Upper Sorbian - +
Table 7 summarizes the results of the quantitative study. The numbers in the cells correspond to the ratio of 'who' in these contexts. The dashed vertical lines designate statistically significant differences between two contexts with p < .05.19 The normal lines are copied from Table 6. The last column reflects statistically significant differences between the leftmost ('that') and the rightmost ('those') columns.
Table 7. Quantitative and qualitative data on the use of 'who' in different contexts Summary (ratio of 'who')
'that who' 'all who' 'those who' statistically significant difference between 'that' and 'those'
Belarusian 1.0 1.0 0.9 +
Ukrainian 1.0 1.0 0.7 +
Slovene 1.0 0.5 0.0 +
Czech 0.9 0.8 0.5 +
Russian of the 18th c. 0.7 0.3 0.0 +
Serbo-Croatian 0.5 0.0 0.0 +
Slovak 0.8 0.0 0.0 + (also qualitative)
Polish 0.5 0.0 0.0 + (also qualitative)
Lower Sorbian 0.2 0.0 0.0 + (also qualitative)
Upper Sorbian 0.0 0.0 0.0 - (qualitative)
19 Two-sided Fisher's exact test was applied if at least one of the expected values was lower than 5; x2 test was used in all other cases.
Two points are worth noting. First, the qualitative and quantitative data on the use of relative pronouns with 'that' and 'those' follow a common tendency for 'who' to be less preferred (either disallowed or less frequent) with the head 'those.' As shown in Table 7, this tendency is observed or was observed in all the standard Slavic varieties considered in this study. Second, the quantitative data suggests a further difference between 'these' and 'all,' the relativizer 'who' being more frequent in the latter case. This tendency is observed in at least four of the five Slavic languages for which the frequency of both relative pronouns with plural heads is not vanishingly small.
I assume that the two differences have a common nature, even though the second one happens not to be reflected in the grammatically constraints of any of the languages in the study. I therefore propose the second interim summary (15) with the newly found distinction taken into account.
(15) questions > correlatives > 'someone,' 'nobody,' 'each'; 'that' > 'all' > 'those' > heads with nouns
5. Discussion
The hierarchy in (15) incorporates many different properties that could be responsible for the observed grammatical differences. The discussion in parts 5.1-5.2 is intended to clarify which of them could be relevant.
5.1. Positional types of relative clauses and "lightness" of the head
At least some parts of the hierarchy in (15) can be described in terms of a diachronical cline, the positions to the left being diachronically prior.
The first step in the hierarchy (15) corresponds to a well-established grammaticalization path:
(16) interrogatives > (indefinites) > relative markers [Lehmann 1984: 384; Hende-rey 2012: 48-56]
One more part of the hierarchy which can be interpreted in diachronical terms is given in (17).
(17) correlatives > light-headed relatives > relatives with nouns in the head
This subhierarchy captures the synchronical fact that relative clauses without a noun in the head are in a sense intermediate between free relatives and relatives with full heads.
The fact that free relatives often share the relativizer with (some) relative clauses with pronouns in the head was noted on a typological basis already in [Зализняк, Падучева 1975] and [Lehmann 1984]. According to Lehmann [1984: 326], if a relative pronoun coincides with an interrogative pronoun in
noun-headed relatives, it always does so in light-headed relatives. This suggests that light-headed relatives might constitute an intermediate diachronical stage between correlatives and postnominal relative clauses.20
As noted in the literature, light-headed relatives have some properties in common with noun-headed relatives, such as the lack of matching effects, a similar set of semantic interpretations, and non-compatibility with ever [Szucsich 2003; Citko 2004]. A link between light-headed relatives and relatives with a noun in the head is also part of a proposed grammatica-lization cline of relative pronouns by Heine and Kuteva [2006], see also [Mithun 2012].
5.2. Specificity, number, and semantic type
Several researchers have independently claimed that a difference relevant for the choice of relativizers in light-headed relative clauses is that between relatives with specific and non-specific reference. This idea is proposed in these or in slightly different terms (in part due to different languages) by Kriz-kovA [1970], Lehmann [1984], and Spencer [1993]. According to all these studies, 'who' is more pervasive in relative clauses with non-specific reference.
According to KrizkovA [1970], 'which' in Czech, Slovak, Russian, Ukrainian, and Belarusian can be used with the singular head 'that' only in definite specific contexts. As demonstrated by (17) and (18), this rule is not without exceptions, at least in Slovak and Czech, but the general tendency clearly holds.
(18) Slovak, Love and Death (Intercorp, subtitles)
Ten, ktory zije s mecom, zomrie mecom. that which lives with sword dies with.sword
{Political assassination doesn't work. Violence leads to violence.} 'He who lives by the sword dies by the sword.'
(19) Czech, My Name Is Bruce (Intercorp, subtitles)
Ten, ktery se bude chtit utkat s Guan-Dim, —
that which will want to encounter with Guan-Di
utka se take se samotnou smrti.
will encounter too with itself death
'He who would confront Guan-Di,—would also confront death itself.'
20 Diachronically, however, the development of Slavic relatives did not follow this scenario unidirectionally. The group of relatives with nouns in the head is attested for Slavic 'who' in a few Slavic non-modern varieties which do not allow this construction now, at least, older varieties of Serbo-Croatian [Dezso 1982: 219; Murelli 2011: 314], Middle Polish [Senderska 2013: 333], and older Lower Sorbian [Bartels, Spiess 2012: 227]. It was also more frequent in Middle Russian [Борковский 1973].
Keeping this in mind, it is natural to suppose that this tendency could account for the observed difference between 'that' and 'these.' Plural heads could just be more likely to be specific, because the singular 'that' pronouns frequently participate in constructions with generalizing semantics. To check whether specificity could be the sole relevant factor, a pilot study was conducted based on the 18th-century Russian material. The relative clauses with the head 'those' were compared to relative clauses with the head 'that,' further divided into relative constructions with specific reference (20) and relative constructions with non-specific reference (21).
(20) Russian, Nikolai Karamzin, 1793 (RNC)
Esli on sam postal tebja — tot, kotorogo strasnoe prokljatie if he himself sent you that of.which terrifying curse gremit vsegda v moem sluxe. . . thunders always in my hearing
'If he sent you himself, he whose terrible curse always sounds in my ears. . .'
(21) Russian, Denis Fonvizin, 1788 (RNC)
Basn' ucit, cto tot, kotoryj pervee vsex prinimaet mody,
fable teaches that that which earlier than.everyone accepts vogue
i tot, kotoryj derzitsja stariny, — oba ravnye duraki. and that which keeps.to the.old both equal fools
'The fable teaches us that the one who is the first to accept the new fashion and the one who keeps to the old are equally fools.'
The study was designed to check whether specificity correlates with the choice of the relative pronoun in this Russian variety and whether this difference could account for the deviant behavior of the plural heads.
In the singular, the difference was drawn between the constructions with generalized reference, classified here as non-specific, and all other cases. The difference in specificity is difficult to observe in the plural, which is why no attempt was made to classify the occurrences of the plural into these two groups.
The results are given in Table 8. The data show statistically significant differences in both pairs. The difference between specific and non-specific 'that' (x2, p < .05) suggests that this property was relevant for Russian of the 18th century. The difference between specific 'that' and all the occurrences of 'those' (x2, p < .05) suggests that even if we presume that all the examples with 'those' are specific, the number is still a different factor.
Table 8. Frequency of 'who' and 'which' in Russian of the 18th century The impact of number and specificity (RNC)
'who' 'which' ratio of 'who'
'that' non-specific 79 21 0.79
specific 29 7121 0.29
'those' 4 96 0.04
To sum up, both number and specificity influenced the choice of relative pronouns in at least one Slavic variety. As the number distinction shows up in all the Slavic varieties considered in this study, it is likely to be an independent parameter in at least some of these varieties.
I suggest that the data on number distinctions presented so far can be summarized as follows:
(22) SG > PL
(23) no (semantically regular) sg/pl distinction > sg/pl distinction
As in all the hierarchies above, the properties at the left make the choice of 'who' more favorable. The hierarchy in (22) is a proposed summary for the discussion of the difference between 'that' and 'those,' and (23) is a suggested way to account for the difference between 'all' and 'those.'22
Bulgarian, which is not considered in the core of this study, gives some additional data in favor of (22). As mentioned in part 2, Bulgarian does not distinguish between 'who' and 'which' in the terms of this study, and the two etymological roots constitute a single paradigm. However, according to [Hh-qo^OBA 1986; H^heb 2012: 192], the forms derived from the former 'who' stem can only be used in singular masculine animate contexts, i.e., the remnants of the 'who' pronoun in Bulgarian are preferred in the singular in accordance with the hierarchy in (22).
Returning to the question of specificity, it is plausible to suggest that this distinction can be rephrased as that between maximalizing and restrictive relative clauses, as described by [Grosu, Landman 1998], maximalizing relative clauses corresponding to the non-specificity condition. As claimed in [ibid.], maximalizing relatives demand universal or definite semantics. While definiteness of examples like (20) can easily be explained by the presence of a definite pronoun in the head, the universal semantics of (21) is likely to result from a maximalizing operation. With this assumption, the generalization
21 About half of the examples in this group have reference to God. This subgroup has a stronger tendency to contain the pronoun 'which.' The tendencies presented in Table 7 hold true if these examples are excluded from the survey.
22 This generalization could also capture the colloquial Russian data discussed in footnote 14.
of the observed data might be laid out as follows: the relative pronouns of the 'who' class prefer maximalizing relative clauses over restrictive relative clauses.23
A further distinction between semantic types of relative clauses is drawn in East Slavic languages, where 'who' is possible in (a subclass of) restrictive relative clauses with nouns in the head and impossible in appositive relative clauses with nouns in the head.24 As illustrated by (24)-(29), Russian kto, Ukrainian xto, and Belarusian xto 'who' allow only the restrictive interpretation, while kotoryj, kotry, and jaki 'which' are grammatical with both restrictive (24a), (26a), (28a) and appositive (24b), (26b), (28b) interpretations.
(24) Russian, M. Sergeev. The Magic Galosha (RNC)
Vzroslye, kto v etot cas byl svoboden ot raboty, prinesli stulja. . . grown-ups who in this hour was free from work brought chairs
'The grown-ups who were free at that time brought chairs.'
a. OK'Some of the grown-ups were free.'
b. *'All the grown-ups were free, and the speaker knows it.'
(25) Russian, (constructed)
Vzroslye, kotorye v etot cas byli svobodny ot raboty, prinesli stulja. . . grown-ups which in this hour was free from work brought chairs
'The grown-ups(,) who were free at that time(,) brought chairs.'
a. OK'Some of the grown-ups were free.'
b. OK'All the grown-ups were free, and the speaker knows it.'
(26) Ukrainian, M. Frolova. The Ball in the Firmament (RNC)25
Zbihlys' susidy, xto buv nedaleko. . . came.running neighbors who was not.far 'The neighbors who were nearby came running.'
a. OK'Some of the neighbors were nearby.'
b. *'All the neighbors were nearby, and the speaker knows it.'
23 Alternatively, both (20) and (21) can be viewed as maximalizing, universal semantics being diachronically the first step by which interrogative pronouns acquire maximalizing semantics [Belyaev, Haug]. In this case, this tendency reflects the general inclination of interrogative pronouns to be used in diachronically early contexts.
24 An anonymous reviewer suggests that these relative clauses could be free relatives juxtaposed to the noun. The data that prove kto-relatives can modify nouns in the head includes the agreement of the predicate in these clauses with the head [Xo.noflH.noBA 2015] and the fact that the head can contain the word tot 'that' without the deictic meaning, which is impossible without a modifying relative clause.
25 I would like to thank Viktor A. Stegnij for the grammaticality judgments of the Ukrainian sentences and Zhanna Duko for the judgments of the Belarusian sentences.
(27) Ukrainian (constructed)
Zbihlys' susidy, kotri buly nedaleko. . .
came.running neighbors which were not.far
'The neighbors(,) who were nearby(,) came running.'
a. OK'Some of the neighbors were nearby.'
b. OK'All the neighbors were nearby, and the speaker knows it.'
(28) Belarusian, I. Chigrinov. Blood Acquittal (RNC)
Palonnyja, xto jasce stajau lja scjany, pacali aziracca na kryk captives who still stood.SG near wall started look.around on cry
i adyxodzic' na sjaredzinu dvara. and step.away on middle of.yard
'The captives who were still standing near the wall started to turn to the cry and move to the center of the yard.'
a. OK'Some of the captives were near the wall.'
b. *'All the captives were near the wall.'
(29) Belarusian, I. Chigrinov. Blood Acquittal (RNC)
Palonnyja, jakija jasce stajali lja scjany, pacali aziracca na kryk captives which still stood.PL near wall started look.around on cry
i adyxodzic' na sjaredzinu dvara. and step.away on middle of.yard
'The captives(,) who were still standing near the wall(,) started to turn to the cry and move to the center of the yard.'
a. OK'Some of the captives were near the wall.'
b. OK'All the captives were near the wall.'
It is therefore possible to provisionally summarize the differences between semantic types of clauses which favor the use of 'who' as follows:
(30) maximalizing > restrictive > appositive
6. Summary
The tendencies discussed in part 5 are brought together in the following schema (see Figure 1).
Most of these generalizations are based on the hierarchy (15) which is justified by the data of numerous Slavic languages. The second part of the hierarchy (34) is less reliable, because it is backed only by the data of East Slavic languages.
Figure 1
'who' <-► 'which'
head:
(31) no external head > external head without a pronoun > external head with a noun
(32) no semantically regular sg/pl distinction > sg/pl distinction
(33) sg > PL clause:
(34) maximalizing > restrictive > appositive
The choice of relative pronoun is usually regarded as the sole argument in favor of a separate analysis of light-headed vs. noun-headed relative constructions, see, e.g., [Szucsich 2003; Citko 2004]. It follows naturally from the hierarchies related to the properties of the head that light-headed relatives do not constitute a homogeneous typologically relevant grammatical class. In other words, if this distinction is preserved, it is typologically more useful to postulate a gradual opposition between "lighter" and "heavier" heads rather than a binary opposition between light vs. noun-headed relatives.
Building on the data in (31)-(34), I would like to propose that all these differences can be derived from a single macro-parameter, namely, the extent to which the head is integrated into the relative clause. The more integrated heads are more likely to take 'who' in the relative clause.
This macro-parameter would naturally explain (34), because, according to Grosu and Landman [1998: 126], the semantic impact of the material external to the relative clause on the overall semantics of the construction declines along the following hierarchy:
(35) Simplex XPs—Appositives—Restrictives—Maximalizers—Simplex CPs
As a head gets more integrated into the relative clause it is natural to assume that it is less likely to have an independent impact on the overall semantics; see also the discussion in [^ah^ep 2012: 411-417].
Integration is also more natural for a head that is simply smaller segmentally and structurally (31) and has less "semantic material" (32). The difference in (33) could be regarded as that between the more and the less frequent options, which are, accordingly, less or more informative. The more informative plural form can have more impact on the semantics of the construction.
The generalization is additionally supported by four other phenomena in Russian, all of which involve some kind of interaction between the material inside the relative clause and the material outside the relative clause. These phenomena are the following:
— Agreement of the relative clause predicate with the head in Russian relative constructions [XonoflHnoBA 2015].
— Inverse attraction in non-standard Russian [Kholodilova 2015]. Under inverse attraction, the head of the relative clause is case-marked as if it belonged to the dependent clause, as in (36); see, e.g., Bianchi [2000] for an overview of attraction phenomena in relative clauses.
(36) Non-standard Russian (Yandex search engine)
■Vsem, komu eto nado, sami razberutsja.
to.everybody to.whom this is.necessary themselves will.figure.out
'Everybody who needs it (lit. to whom it is necessary) will figure it out himself.'
— Case attraction in non-standard Russian [Kholodilova 2015], i.e., constructions in which the relative pronoun receives case from the main clause, as in (37).
(37) Non-standard Russian (Yandex search engine)
Bol'suju sobaku zvali Dunkan, kotoruju pomen'se — Tom. big.ACC dog.ACC they.called Dunkan which.ACC smaller Tom
'The big dog was called Dunkan, the smaller one was called Tom.'
— Transparent relatives in non-standard Russian [XonoflH.oBA 2016]. Transparent relative constructions can be roughly defined as those which allow the nominal predicate of the relative clause to participate in the main clause, for instance, by triggering agreement as in (38), and see [van Riemsdijk 2001] for more details.
(38) Non-standard Russian (RNC)
V kazdoj strane byli i est' to, cto nazyvaetsja "social'nye problemy". in every country were and is/are that what is.called social problems
'In every country, there are what is called social problems'
Table 9 summarizes the grammatical tendencies these constructions have in common. The list might be non-exhaustive. Fuller data on these tendencies can be found in the works cited above.
Table 9. Russian relative constructions with a grammatical interaction
between the main and the subordinate clause: Common tendencies
relative pronoun choice predicate agreement inverse attraction case attraction transparency
"lightness" of the head no head > pronoun (logically impossible) (logically impossible) + +
pronoun > noun + + + + +
semantic types maximalizing > restrictive + + +
restrictive > appositive + + +
Finally, some phenomena suggest that the distinction between more and less integrated heads is typologically relevant.
First, the idea that the opposition between internal and external relative heads is gradual was advanced by Lander [.Hah^ep 2012: 403-417] on the basis of Adyghe and some typological data.
Second, inverse attraction (see the definition above) in Standard Moksha and Standard Udmurt is only possible if the head of the clause does not contain a noun [Kholodilova, Privizentseva 2015], and, according to Aghaei [2003], as cited in Cinque [2015], inverse attraction in Persian is impossible in appositive relative clauses. These facts suggest that inverse attraction tends cross-linguistically to show tendencies similar to those listed in Table 9.
A small piece of evidence in favor of the overall idea is also supplied by Finnish punctuation as related to pausation. If some of the heads are more integrated into the relative clause than others, it is natural to expect that this difference will be reflected in their intonation patterns, which can in turn be reflected by the punctuation. According to [Itkonen, Maamies 2012], a comma can be absent before a relative clause in Finnish if and only if it is headed by a demonstrative pronoun. A corpus study presented in Table 10 shows that a further difference can be found between the singular and the plural of the relative demonstrative pronoun, and commas are left out more often if the head is in the singular. This means that Finnish punctuation obeys at least two of the tendencies described above, the tendency for the demonstrative pronouns to be more integrated than nouns (31) and the tendency for singular to be more liable to integration than plural (33).
Table 10. Commas in Finnish relative clauses with pronominal heads and the pronoun joka 'which' (InterCorp)
no comma comma ratio of missing commas
se 'that' 284 591 0.3
ne 'those' 154 1208 0.1
To summarize, I suggest, following [^AHflEP 2012: 403-417], that the level of integration of the head into the relative clause is a typologically relevant gradual macro-parameter. I also argue that this macro-parameter includes the opposition between nouns and demonstrative pronouns in the head, as well as that between plural and singular demonstrative pronouns. Finally, I suggest that this macro-parameter manifests itself, among other things, in the choice of relative pronouns in Slavic languages.
Bibliography
Sources
ASPAC, Amsterdam Slavic Parallel Aligned Corpus compiled by Adrian Barentsen26
DOTKO, Dolnoserbski tekstowy korpus (www.dolnoserbski.de/korpus/, accessed via "Kontext" interface at kontext.korpus.cz)
HOTKO, Hornoluzicky textovy korpus (korpus.cz/corpora/run.cgi/first_form, accessed via "Kontext" interface at kontext.korpus.cz)
InterCorp (https://ucnk.ff.cuni.cz/intercorp/, accessed via "Kontext" interface at kontext.korpus.cz)
FidaPLUS, Korpus slovenskega jezika
Parasol (von Waldenfels, Ruprecht and Meyer, Roland (2006-): ParaSol, a Corpus of Slavic and Other Languages. Available at parasolcorpus.org. Bern, Regensburg)
RNC, Russian National Corpus (ruscorpora.ru), the main and parallel subcorpora
Literature Aghaei 2003
Aghaei B., "Case attraction: Evidence for raising analysis for relative clauses in Farsi. Paper presented at SALA 23" (unpublished manuscript). Bartels, Spiess 2012
Bartels H., Spiess G., "Restrictive relative clauses in the Sorbian languages," Sprachtypologie und Universalienforschung, 65/3, 2012, 221-245. Bauer 1967
Bauer J., "K vyvoji vztaznych vet v slovanskych jazycich," Slavica Slovaca, 2, 1967, 297-320. Belyaev, Haug
Belyaev O., Haug D., "Correlatives in Indo-European and Beyond" (unpublished manuscript). Bianchi2000
Bianchi V., "Some Issues in the Syntax of Relative Determiners," in: The Syntax of Relative Clauses, Amsterdam, Philadelphia, 2000, 53-81. Browne 1986
Browne W., Relative Clauses in Serbo-Croatian in Comparison with English, Zagreb, 1986. Cinque 2015
Cinque G., "Three Phenomena Discriminating between "Raising" and "Matching" Relative Clauses," Semantic-Syntax Interface, 2, 1, 2015, 1-27. Citko 2000
Citko B., "Parallel Merge and the Syntax of Free Relatives" (PhD Thesis, New York, 2000).
26 I am grateful to Adrian A. Barentsen for access to the corpus data.
-2004
Citko B., "On Headed, Headless, and Light-headed Relatives," Natural Language and Linguistic Theory, 22/1, 2004, 95-126.
Darovec 1985
Darovec M., "Vztazne vety v slovencine," Studia Academica Slovaca, 14, 1985, 107-131. Dezso 1982
Dezso L., Typological Studies in Old Serbo-Croatian Syntax, Budapest, 1982. de Vries 2002
de Vries M., "The Syntax of Relativization" (PhD Thesis, University of Amsterdam, 2002).
Fasske, Michalk 1981
Fasske H., Michalk S., Grammatik der obersorbischen Schriftsprache der Gegenwart. Morphologie, Bautzen, 1981.
Gol^b, Friedman 1972
Gol^b Z., Friedman V. A., "The Relative Clause in Slavic," in: P. Peranteau, J. Levi, G. Phares, eds., The Chicago Which Hunt. Papersfrom the Relative Clause Festival, Chicago, 1972, 30-46.
Grosu, Landman 1998
Grosu A., Landman F., "Strange Relatives of the Third Kind," Natural Language Semantics, 6, 1998,125-170.
Haspelmath 2010
Haspelmath M., "Comparative Concepts and Descriptive Categories in Cross-linguistic Studies," Language, 86/3, 2010, 663-687.
Heine, Kuteva 2006
Heine B., Kuteva T., The Changing Languages of Europe, Oxford, 2006.
Henderey 2012
Henderey R., Relatives in Time and Space: A Case Study in the Methods of Diachronic Typology, Amsterdam, Philadelphia, 2012.
Itkonen, Maamies 2012
Itkonen T., Maamies S., Uusi kieliopas, 5. painos, Helsinki, 2012.
Janas 1976
Janas P., Niedersorbische Grammatik für den Gebrauch der Sorbischen Erweiterten Oberschule, Bautzen, 1976.
Karli'k 1988
Karlik P., "Relativni vety z hlediska reference," Sbornik praci Filozoficke fakulty brnenske univerzity. A: Rada jazykovedna, 37/A36, 59-66.
Kholodilova 2013
Kholodilova M., "Animacy Distinction in (Mostly) Slavic Possessive Relative Pronouns. Talk at CoQuaT 2013, Leipzig, August 14, 2013" (unpublished manuscript).
-2015
Kholodilova M., "Case Attraction in Non-standard Russian. Talk at 'Insufficient Strength to Defend Its Case': Case Attraction and Related Phenomena, Wroclaw, September 18-19, 2015" (unpublished manuscript).
Kholodilova, Privizentseva 2015
Kholodilova M., Privizentseva M., "Inverse Attraction in Finno-Ugric Languages. Talk at 'Insufficient Strength to Defend Its Case': Case Attraction and Related Phenomena, Wroclaw, September 18-19, 2015" (unpublished manuscript).
KoRDic 1994
KoRDic S., "Relativna recenica s formalnim antecedentom," Radovi Zavoda za slavensku filologiju, 28-29, 1994, 81-93.
--1999
KoRDic S., Der Relativsatz im Serbokroatischen, München, 1999. Kr îzkovâ 1970
Kr îzkovâ H., "Relativni vety v soucasnych slovanskych jazycich," Slavia, 39, 1970, 10-40. Lehmann 1984
Lehmann Ch., Der Relativsatz: Typologie seiner Strukturen; Theorie seiner Funktionen; Kompendium seiner Grammatik, Tübingen, 1984.
Mithun 2012
Mithun M., "Questionable Relatives," in: B. Comrie, Z. Estrada-Fernândez, eds., Relative Clauses in Languages of the Americas: A Typological Overview, Amsterdam, Philadelphia, 2012.
Mitrovic 2012
Mitrovic I., "Relative Clauses in Serbian. Talk at 'Syntax Lab,' September 11, 2012" (unpublished manuscript).
Mucke 1891
Mucke K. E., Historische und vergleichende Laut- und Formenlehre der niedersorbischen (niederlausitzisch-wendischen) Sprache mit besonderer Berücksichtigung der Grenzdialekte und des Obersorbischen, Leipzig, 1891.
Murelli 2011
Murelli A., Relative Constructions in European Non-standard Varieties, Berlin, Boston, 2011. Podhajeckâ 2010
Podhajeckâ K., "Komunikacny potencial operatora kto v sucasnej pisanej slovencine," in: Zbornikplnych prispevkov z XIX. kolokvia mladychjazykovedcov (Trnava — Modra-Harmonia, 18.-20.11.2009), Trnava, 2010, 276-282.
--2012
Podhajeckâ K., "Operatory co, kto a ktory ako prostriedky rozvijania vypovede," in: M. Chovanec, ed., Studentska vedecka konferencia. Zbornik prispevkov, 7, Presov, 2012, 757-764.
Polanski 1967
Polanski K., Skladnia zdania zlozonego w jçzyku gornoluzyckim, Wroclaw, 1967. Senderska 2013
Senderska J., "Zdania wzglçdne w dobie sredniopolskiej," Polonica, 33, 2013, 327-342. Spencer 1993
Spencer D., "Aspects of the Syntax of Relative Clauses in Colloquial and Standard Russian" (PhD Thesis, University of Toronto, 1993).
Svoboda 1967
Svoboda K., "K nekterym otazkam vet vztaznych," Slavica Slovaca, 2/4, 1967, 334-339. Szucsich 2003
Szucsich L., "The Structure of Relative Clauses in Slavic," in: P. Kosta et al., eds., Investigations into Formal Slavic Linguistics, 2, Frankfurt a. M., 2003, 697-713.
Vaillant 1958
Vaillant A., Grammaire comparée des langues slaves, 2: Morphologie, Lyon, 1958. van Riemsdijk 2001
van Riemsdijk H., "A Far from Simple Matter: Syntactic Reflexes of Syntax-pragmatics Misalignments," in: I. Kenesei, R. M. Harnish, eds., Semantics, Pragmatics, and Discourse. Perspectives and Connections. A Festschrift for Ferenc Kiefer, 2001, 21-41.
von Waldenfels 2014
von Waldenfels R., "Explorations into Variation across Slavic: Taking a Bottom-up Approach," in: B. Szmrecsanyi, B. Wälchli, eds., Aggregating Dialectology and Typology: Linguistic variation in text and speech, within and across languages, Berlin, New York, 2014, 290-323. Zubaty1918
Zubaty J. "Jenz, ktery, kdo, co," Nase fee, 2, 1918, 37-44. Бауэр 1967
Бауэр Я., "К развитию относительных придаточных предложений в славянских языках", Вопросы языкознания, 5, 1967, 47-59.
Беличова 1988
Беличова Е., "Придаточные предложения относительные и посессивность в современных славянских языках", in: Ю. Н. Караулов, ред., Язык: система и функционирование, Москва, 1988, 22-30.
Борковский 1973
Борковский В. И., Сравнительно-исторический синтаксис восточнославянских языков. Сложноподчиненные предложения, Москва, 1973.
Зализняк, Падучева 1975
Зализняк А. А., Падучева Е. В., "К типологии относительного предложения", Семиотика и информатика, 6, 1975, 51-101.
Илиев 2012
Илиев И. Г., Теория на относительността. Българските относителни местоимения — произход и развой, Пловдив, 2012.
Ландер 2012
Ландер Ю. А., "Релятивизация в полисинтетическом языке: адыгейские относительные
конструкции в типологической перспективе" (Дисс____канд. филол. наук, Москва,
Институт востоковедения РАН, 2012).
Никунласси 2008
Никунласси А., Приместоименно-относительные конструкции в современном русском языке (= Slavica Helsingiensia, 33), Helsinki, 2008.
Ницолова 1986
Ницолова Р., Българските местоимения, София, 1986.
Павловиъ 2012
Павловиъ С., "Систем релативизатора у српском ]езику", Зборникматице српске за славистику, 82, 2012, 71-86.
Трубачев 1987
Трубачев О. Н., ред., Этимологический словарь славянских языков: праславянский лексический фонд, 13, Москва, 1987. Холодилова 2015
Холодилова М. А., "Согласование с вершиной относительных конструкций и обособленных именных оборотов в русском языке", Русский язык в научном освещении, 2, 2015, 74-97.
-2016
Холодилова М. А., "(То,) что называется: между относительной конструкцией и устойчивым вводным оборотом", Acta linguistica Petropolitana. Труды Института лингвистических исследований РАН, 12/3, 2016, 763-780.
References
Bartels H., Spiess G., "Restrictive relative clauses in the Sorbian languages," Sprachtypologie und Universalienforschung, 65/3, 2012, 221-245.
Bauer J., "On the Development of Relative Subordinate Clauses in the Slavonic Languages," Vopro-sy jazykoznanija, 5, 1967, 47-59.
Bauer J., "K vyvoji vztaznych vet v slovanskych jazycich," Slavica Slovaca, 2, 1967, 297-320.
Belichova E., "Pridatochnye predlozheniia otno-sitel'nye i posessivnost' v sovremennykh slavianskikh iazykakh," in: Yu. N. Karaulov, ed., lazyk: sistema i funktsionirovanie, Moscow, 1988, 22-30.
Bianchi V., "Some Issues in the Syntax of Relative Determiners," in: The Syntax of Relative Clauses, Amsterdam, Philadelphia, 2000, 53-81.
Borkovsky V. I., Sravnitel'no-istoricheskii sintaksis vostochnoslavianskikh iazykov. Slozhnopodchinennye predlozheniia, Moscow, 1973.
Browne W., Relative Clauses in Serbo-Croatian in Comparison with English, Zagreb, 1986.
Cinque G., "Three Phenomena Discriminating between «Raising» and «Matching» Relative Clauses," Semantic-Syntax Interface, 2, 1, 2015, 1-27.
Citko B., "On Headed, Headless, and Lightheaded Relatives," Natural Language and Linguistic Theory, 22/1, 2004, 95-126.
Darovec M., "Vzt'azne vety v slovencine," Studia Academica Slovaca, 14, 1985, 107-131.
Dezso L., Typological Studies in Old Serbo-Croatian Syntax, Budapest, 1982.
Faßke H., Michalk S., Grammatik der obersorbischen Schriftsprache der Gegenwart. Morphologie, Bautzen, 1981.
Gol^b Z., Friedman V. A., "The Relative Clause in Slavic," in: P. Peranteau, J. Levi, G. Phares, eds., The Chicago Which Hunt. Papers from the Relative Clause Festival, Chicago, 1972, 30-46.
Grosu A., Landman F., "Strange Relatives of the Third Kind," Natural Language Semantics, 6, 1998, 125-170.
Haspelmath M., "Comparative Concepts and Descriptive Categories in Cross-linguistic Studies," Language, 86/3, 2010, 663-687.
Heine B., Kuteva T., The Changing Languages of Europe, Oxford, 2006.
Henderey R., Relatives in Time and Space: A Case Study in the Methods of Diachronic Typology, Amsterdam, Philadelphia, 2012.
Iliev I. G., Teoriia na otnositel'nostta. Bülgarskite otno-sitelni mestoimeniia — proizkhod i razvoi, Plovdiv, 2012.
Itkonen T., Maamies S., Uusi kieliopas, 5. painos, Helsinki, 2012.
Janas P., Niedersorbische Grammatikfür den Gebrauch der Sorbischen Erweiterten Oberschule, Bautzen, 1976.
Karlik P., "Relativni vety z hlediska reference," Sbornik praci Filozoficke fakulty brnenske univerzity. A: Radajazykovednä, 37/A36, 59-66.
Kholodilova M. A., "Agreement with the Head of Russian Relative Constructions and Detached Attributive Phrases," Russkijyazyk v nauchnom osve-shchenii (= Russian Language and Linguistic Theory), 2, 2015, 74-97.
Kholodilova M. A., "(To,) Cto Nazyvaetsja '(That) What Is Called': Halfway Between a Relative Construction and a Set Parenthetical Expression," Acta linguistica Petropolitana, 12/3, 2016, 763-780.
Kordic S., "Relativna recenica s formalnim antecedentom," Radovi Zavoda za slavensku filologju, 28-29, 1994, 81-93.
Kordic S., Der Relativsatz im Serbokroatischen, München, 1999.
Krízková H., "Relativní vety v soucasnych slovanskych jazycích," Slávia, 39, 1970, 10-40.
Lehmann Ch., Der Relativsatz: Typologie seiner Strukturen; Theorie seiner Funktionen; Kompendium seiner Grammatik, Tübingen, 1984.
Mithun M., "Questionable Relatives," in: B. Comrie, Z. Estrada-Fernández, eds., Relative Clauses in Languages of the Americas: A Typological Overview, Amsterdam, Philadelphia, 2012.
Murelli A., Relative Constructions in European Non-standard Varieties, Berlin, Boston, 2011.
Nikunlassi A., Relative Constructions with Pronominal Heads in Contemporary Russian (= Slavica Helsingiensia, 33), Helsinki, 2008.
Nitsolova R., Bülgarskite mestoimeniia, Sofia, 1986.
Pavlovic S., "Sistem relativizatora u srpskom je-ziku," Zbornik matice srpske za slavistiku, 82, 2012, 71-86.
Podhajecká K., "Komunikacny potenciál operá-tora kto v súcasnej písanej slovencine," in: Zborník plnych príspevkov z XIX. kolokvia mladych jazykovedcov (Trnava — Modra-Harmónia, 18.-20.11.2009), Trnava, 2010, 276-282.
Podhajecká K., "Operátory co, kto a ktory ako prostriedky rozvíjania vypovede," in: M. Chovanec, ed., Studentská vedecká konferencia. Zborník príspevkov, 7, Presov, 2012, 757-764.
Polanski K., Skladnia zdania zlozonego w jçzyku górnoluzyckim, Wroclaw, 1967.
Senderska J., "Zdania wzglçdne w dobie sredniopolskiej," Polonica, 33, 2013, 327-342.
Svoboda K., "K nekterym otázkám vet vztaznych," Slavica Slovaca, 2/4, 1967, 334-339.
Szucsich L., "The Structure of Relative Clauses in Slavic," in: P. Kosta et al., eds., Investigations into Formal Slavic Linguistics, 2, Frankfurt a. M., 2003, 697-713.
Vaillant A., Grammaire comparée des langues slaves, 2: Morphologie, Lyon, 1958.
van Riemsd^k H., "A Far from Simple Matter: Syntactic Reflexes of Syntax-pragmatics Misalignments,"
in: I. Kenesei, R. M. Harnish, eds., Semantics, Pragmatics, and Discourse. Perspectives and Connections. A Festschrift for Ferenc Kiefer, 2001, 21-41.
von Waldenfels R., "Explorations into Variation across Slavic: Taking a Bottom-up Approach," in: B. Szmrecsanyi, B. Walchli, eds., Aggregating Dialectology and Typology: Linguistic variation in text
Acknowledgements
Russian Foundation for Basic Research. Project No.
and speech, within and across languages, Berlin, New York, 2014, 290-323.
Zalizniak A. A., Paducheva E. V., "K tipologii otnositel'nogo predlozheniia," Semiotics and Informatics, 6, 1975, 51-101.
Zubaty J. "Jenz, ktery, kdo, co," Nase fee, 2, 1918, 37-44.
14-06-31212.
Мария Александровна Холодилова, магистр филологии
Институт лингвистических исследований РАН, младший научный сотрудник
Лаборатории типологического изучения языков
199053 С.-Петербург, Тучков пер., 9
Россия/Russia
hol_m@mail.ru
Received February 11, 2016
