CC BY
8
DOI: https:// doi.org/ 10.15688/jvolsu2.2022.5.4
UDC 8ПЛ61.Г36 LBC 81.411.2-22
Submitted: 24.10.2021 Accepted: 16.05.2022
WHAT MAY STATISTICS TELL US ABOUT NULL SUBJECTS IN RUSSIAN?1
Abstract. The article highlights the linguistic phenomenon of abandoning the expletively presented syntactic subject (the Null Subject Phenomenon) in different languages and focuses on its exemplification by the Russian language. It is shown that despite the existing number of studies on this issue, there is lack of precisely formulated criteria that determine the choice of linguistic means to represent a syntactic subject in the languages, which allow both options (partial null subject languages). Based on the discussion of literature on the topic and on the statistical analysis of 16,718 sentence samples from the spoken language, print journalism, and fiction subcorpora of the Russian National Corpus, the article argues that the choice between null and overt subjects in Russian sentences depends on factors such as tense, person, and style, but that this dependency is rather weak, owing to the fact that the Russian language system is now in the process of change. Null-subjectness is not an exclusive syntactic parameter and should be studied along with other aspects like semantics and pragmatics. The statistical data from this study support previously discussed diachronic and acquisition data, and confirm that languages do not nicely distribute themselves into distinct groups, but inhabit a constantly changing continuum.
Key words: pro-drop languages, null subject, corpus study, syntax, continuum, syntactic subject, Russian.
Citation. Eismont P.M. What May Statistics Tell Us About Null Subjects in Russian? Vestnik Volgogradskogo gosudarstvennogo universiteta. Seriya 2. Yazykoznanie [Science Journal of Volgograd State University. Linguistics], 2022, vol. 21, no. 5, pp. 41-49. DOI: https://doi.org/10.15688/jvolsu2.2022.5.4
Аннотация. Статья посвящена проблеме возможности или невозможности опущения эксплицитно выраженного синтаксического субъекта (null subject phenomenon) в различных языках мира; данное явление рассмотрено на примере русского языка. Показано, что, несмотря на большое количество исследований по этой проблеме, точные критерии, определяющие выбор способа выражения синтаксического субъекта в языках, допускающих оба возможных варианта (partial null subject languages), не сформулированы. В результате статистического анализа 16 718 финитных глагольных клауз из трех стилистических подкорпусов Национального корпуса русского языка (публицистика, устная речь и художественная литература) выявлено, что эксплицитная реализация синтаксического субъекта или его опущение в русском языке связаны не только с традиционно выделяемыми в лингвистике факторами (время и лицо глагольной формы), но и с фактором стиля текста, однако показано, что эта связь проявляется слабо. С учетом имеющихся в литературе диахронией ческих данных и современных данных, полученных при изучении процесса развития речи, делается вывод, что в настоящее время в русском языке происходит изменение норм опущения синтаксического субъекта. Р-< Установлено также, что языки не распределяются по данному критерию на несколько закрытых групп, а § образуют постоянно изменяющийся континуум.
I Ключевые слова: языки pro-drop, нулевой субъект, корпусная лингвистика, синтаксис, континуум, ® синтаксический субъект, русский язык.
Polina M. Eismont
St. Petersburg University, Saint Petersburg, Russia
УДК 811.161.1'36 ББК 81.411.2-22
Дата поступления статьи: 24.10.2021 Дата принятия статьи: 16.05.2022
ЧТО СТАТИСТИКА МОЖЕТ РАССКАЗАТЬ О НУЛЕВОМ СУБЪЕКТЕ
В РУССКОМ ЯЗЫКЕ?1
Полина Михайловна Эйсмонт
Санкт-Петербургский государственный университет, г. Санкт-Петербург, Россия
Цитирование. Эйсмонт П. М. Что статистика может рассказать о нулевом субъекте в русском языке? // Вестник Волгоградского государственного университета. Серия 2, Языкознание. - 2022. - Т. 21, №2 5. - С. 41-49. -(На англ. яз.). - DOI: https://doi.Org/10.15688/jvolsu2.2022.5.4
Introduction
Null Subject Phenomenon (also known as "Null Subject Parameter," or "Pro-Drop Parameter") has been a consistent interest in linguistic studies since its first appearance in early 1980s. The parameter originally divided languages into two groups: those that require explicit syntactic subjects in any syntactic structures ("non-pro-drop languages," or "non-null subject languages") and those that allow subject omission ("pro-drop languages," or "null subject languages")2 Overt, or non-null, subjects in nonnull subject languages are licensed by syntactic structure and do not add any specific pragmatic nuances, while in null subject languages overt subjects are not structural and therefore the scope of supplied pragmatic shades is very large. Consequently, scholars have suggested that null subject languages be further subdivided. Huang [1984] has suggested the existence of the so-called "radical null subject languages," and Holmberg [2005] has more recently introduced the group of "partial null subject languages". Nevertheless, these groups do not have rigid boundaries, settled by some precise absolute criteria, which has led Duguine [2014] to note the "chaotic character" of null-subject phenomenon.
Where the Russian language falls has been a matter of debate for decades. Some linguists argue that Russian should be considered a null subject language [RuZicka, 1986; Perlmutter, Moore, 2002], while others suggest it is a non-null subject language [Franks, 1995; Gordishevsky, Avrutin, 2003]. A third group of scholars claim that Russian is a typical partial null subject language [McShane, 2009; Madariaga, 2018]. Section Two discusses the merits of each position. In Section Three I present a new statistical study of null subjects in Russian based on data from the Russian National Corpus (henceforth RNC)3. The data shows that the choice between null and overt subjects in Russian depends on such factors as tense, person, and style, but that this dependency is weak, and Russian is in the process of change. The data can also help us understand some tendencies in modern Russian syntax and pragmatics that can aid in the development of
computational dialogue systems, text processing, and machine translation software.
The problem of null subjects
A number of groups or a continuum?
Null Subject Parameter was first formulated in the 1980s as a set of criteria for separating null subject languages (NSL) from non-null subject languages (non-NSL) [Rizzi, 1982]. These criteria initially included the possibility of silent referential subjects, free subject inversion, and other properties of formal syntactic structure. Further studies showed that the typology of languages in relation to this parameter is much more complicated and NSLs are not homogeneous. Lindseth [1998] proposed three absolute properties of a true NSL: 1) only null pronominal subjects are stylistically unmarked; 2) only null pronouns can function as bound variables; 3) only null third plural pronominal subjects can have arbitrary reference.
Following Lindseth, Barbosa [2011] distinguished three typological patterns of NSLs: 1) languages with rich subject agreement which allows subjects to be freely dropped (the so-called consistent NSLs [Holmberg, 2005]); 2) languages that have agreement and referential null subjects in restricted contexts (partial NSLs4); 3) languages that lack agreement but allow both null subjects and null objects (radical NSLs, or discourse NSLs)5. In 2018, Barbosa added one more group ofthe so-called "semi pro-drop languages" that only have impersonal and quasi-argumental null subjects [Barbosa, 2019]. Lastly, Cognola and Casalicchio compare expletives in non-NSLs and in Romance dialects that are classified as NSLs and conclude that they follow different rules and should not be regarded as homogeneous [Cognola, Casalicchio, 2018]. They restrict the use ofnull subjects in partial NSLs with such factors as person (not every person allows null subjects), asymmetry between main and embedded clauses (null subjects tend to embedded clauses) and interpretation of generics (only 3rd person may have arbitrary generic interpretation).
Thus, the original division NSLs and non-NSLs developed into a more nuanced separation of five different types, but this seems to have largely
complicated things. The most decisive distributing factors are the existence and the role expletive pronouns play and the dependence of null and overt subjects from text form and discourse structure. Chinese scholars differentiate discourse-oriented and sentence-oriented languages according to the place of null referential subj ect antecedents : if an antecedent can be found somewhere in the discourse, but not necessarily within the same sentence, then we are dealing with a discourse-oriented NSL (later they were renamed as a "radical NSL"); if not, we are dealing with one of the other groups [Huang, 1984]. Interestingly however, and as I consider more below, Huang's test does not work for Russian which is considered to be a partial NSL, but allows antecedents in both discourse and extra linguistic reality (cf. [Tseytlin, 1976] for a discussion of situational versus contextual ellipsis).
In contrast, Wratil [2011] argues that there is a continuum of partial NSLs. While Camacho focuses on the distinction between thematic and expletive subjects, and suggests four possible combinations: 1) null thematic subjects + null expletive subjects; 2) overt thematic subjects + null expletive subjects; 3) null thematic subjects + overt expletive subjects; 4) overt thematic subjects + overt expletive subjects [Camacho, 2013]. The study of diachronic changes [Madariaga, 2018; Kinn, Rusten, Walkden, 2016; Simonenko, Crabbé, Prévost, 2018], of dialects of NSLs [Camacho, 2013; Cognola, Casalicchio, 2018], of first language acquisition [Wratil, 2011; Pinto, 2006] and of heritage languages [Nagy et al., 2011; Bidese, Tomaselli, 2018] testify that as languages pass from one group to another, pronouns and syntactic structures will often change their properties and behave differently both in history and in language usage. Taken as a whole, these findings prove that the distribution of null subjects is too diverse and we should consider NSLs as a continuum with no rigid boundaries between classes.
Is Russian a partial NSL?
The position of Russian is still vague. At first, it was described as a NSL because it lacks expletive empty pronouns. Subsequently, it was considered to be a non-NSL because its complete syntactic structure requires an explicit subject position. Today however, you often find Russian on the list of languages illustrating partial NSLs.
At the same time, many studies show that Russian often behaves as a discourse or consistent NSL, and it differs too much from typical partial NSLs such as Finnish or Brazilian Portuguese [Madariaga, 2018; Cognola, Casalicchio, 2018]. On the one hand Gordishevsky and Avrutin [2003] among others point out that the null subject is merely an optional strategy in Russian, which differs it from the majority of other Slavic languages that are canonical NSLs [Lindseth, 1998]. On the other hand, Bizarri argues that Russian exhibits properties that are closer to consistent NSLs than to the partial ones [Bizarri, 2015]. And yet others argue that the distribution of null subjects differs significantly in spoken and written Russian with null subjects more frequent in spoken language [McShane, 2009]. This relies on the idea that null subjects in Russian are not structural but are licensed by context and text form, as well as the idea that Russian is a discourse-oriented language [Bizarri, 2015] which makes it closer to radical NSLs than to the partial NSLs (cf. [Barbosa, 2019] for their similarity).
This disagreement raises the question about the nature of this process: is it syntactic or pragmatic? McShane [2009] provides four criteria for answering this question for Russian: 1) syntactic configuration (is it a coordinate or subordinate clause); 2) avoidance of redundancy (including repetitions, series of co-referential actions, elaboration, etc.); 3) avoidance of long-winded formulations; 4) stylistic nuances with only the first one being syntactic. However, the conditions and parameters that were used as the focus in most past studies have not considered the usage of null subjects because such a complex phenomenon requires large statistical and psycholinguistic research to understand its functioning in real speech data (cf.: [McShane, 2009; Bizarri, 2015]). For typical NSLs, null subjects are stylistically unmarked while overt subjects emphasize the referent or signal the change of referents, but this is not true in Russian. In Russian, both null and overt subjects may be stylistically neutral or indicate some idea depending on the context of communication. Interaction between topicality, morphology, and null subjects is significant for any NSL [Cognola, Casalicchio, 2018], and Russian is not an exception.
Material and methods
In this section I present some statistics of null and overt subjects in modern Russian. The RNC has become a fruitful source of information about
any aspect of modern Russian and its diachronic changes. The size of the RNC is more than 600 million of tokens. The RNC includes twelve subcorpora: the main subcorpus, dialectal subcorpus, poetry subcorpus, subcorpus of spoken Russian, etc. Rich metadata and detailed linguistic annotation allows one to compare various parameters in texts from different times, styles, and genres. This is a very important advantage for studying the Null Subject Phenomenon, since the use of null or overt subjects is determined not only by the syntactic characteristics of the language, but also by other pragmatic factors discussed in Section Two.
I focus on the influence of three main factors that are considered decisive for choosing between null and overt subject: 1) tense, 2) person, 3) style. Tense shows the role of agreement, since person is expressed morphologically in present forms, but in past tense this information is hidden. Person corresponds to a) the difference between first and second person behavior on the one hand, and third person behavior on the other hand; b) the arbitrary interpretation of generic null subjects which was repeatedly indicated as one of the most important factors for considering Russian as a partial NSL [Holmberg, Nayudu, Sheehan, 2009]. Style shows the difference between spoken language, print journalism, and fiction which reflects the distinct topic-focus structure of these text forms and can show the role of the discourse structure in the choice of subjects6.
The process of data collection was the following. First, I limited the search within three subcorpora: 1) the spoken subcorpus, which contains domestic conversations, public lectures, transcripts of TV talkshows, etc.; 2) the print journalism subcorpus, which
contains Russian newspapers and magazines published after 2000; 3) the fiction subcorpus, which contains novels, short stories, and other fiction writing originally in Russian and published after 1980. I created a series of specific search queries focusing on verbs of all possible forms - i.e., three tenses (present, past, and future), three persons (first, second, and third), and two numbers (singular and plural). I repeated this series of search queries for each subcorpus and received a sample of more than 25,000 sentences. I then looked through the search results and eliminated all sentences with the verb "быть" ("to be") because it lacks finite forms in the present and is widely used in compound verb forms which may skew my statistics. Then I manually tagged each sentence with either null subject, overt pronominal subject, or overt nominal subject (only in third person verb forms). In complex sentences I tagged only the initial clause to avoid any asymmetry between main and embedded clauses ([Cognola, Casalicchio, 2018] for an overview).
I further tagged as null subjects only null referential subjects, and eliminated all generic ones. In Russian there are three types of null generic sentences: impersonal where the verb is in the third person form either in present or past; second person generalized human construction; indefinite personal construction (third plural both in present and past); in all these structures the use of syntactic subjects is forbidden, if the position is fulfilled we get another sentence with different meaning (compare examples 1 and 2). Referential null subjects may be fulfilled: if we use an overt subject in place of a null one, we get neither another sentence with different meaning nor an ungrammatical sentence (compare examples 3 and 4). The resulting clause
Examples:
(1) Iz truby kapalo.
from pipe [GEN.SG] drip [PST.3SG.N] Something was dripping from the pipe.
(2) Iz truby kapalo vino.
from pipe [GEN.SG] drip [PST.3SG.N] wine [ACC.SG] Wine was dripping from the pipe.
(3) Khochesh chaj? - Khochu, spasibo! want [PRS.2SG] tea [ACC.SG] want [PRS.1SG] thank you Do you want some tea? - Yes, I do, thank you!
(4) Ty khochesh chaj? - Khochu, spasibo! 2SG want [PRS.2SG] tea [ACC.SG] want [PRS.1SG] thank you Do you want some tea? - Yes, I do, thank you!
will be grammatically correct, but may need some editing because of its position in the discourse.
Thus, in Russian in the case of a generic null subject, the very fact of its absence is syntactically and semantically meaningful, while in the case of a referential null subject its absence is pragmatically neutral in many contexts. Kasevich [2004] refers to the former as "subject reduction" and to the latter as "subject ellipsis," since referential null subjects always correspond to some verb argument, while subject reduction 7 (as well as expletives in non-NSLs) do not fulfill any arguments and are only structural8.
This initial analysis revealed that the data needed to be slightly modified. The complexity of Russian tenses and its aspectual system result in a deficient verb paradigm for perfective verbs which have lost the present forms, and so Russian has only five tense-aspect forms [Bondarko, 1971]. Consequently, in this analysis the forms of perfective future were combined with the forms of imperfective present, which have the same set
Table 1. The final set of analyzed sentences
S(N) S(Pron) NULL Total
Spoken no-past 1 person - 1254 709 1963 5480 6647
2 person - 939 1067 2006
3 person 587 623 301 1511
past all 261 619 287 1167 1167
News no-past 1 person - 690 655 1345 3861 4871
2 person - 423 552 975
3 person 1213 204 124 1541
past all 423 456 131 1010 1010
Fiction no-past 1 person - 1110 743 1853 4417 5200
2 person - 666 849 1515
3 person 645 250 154 1049
past all 421 253 109 783 783
Total 3550 7487 5681 - - 16718
Table 2. Chi-squared tests of the statistical significance of tense, persons, and styles
Factor X2 df P
Tense 715.886 2 < 0,001
Person 286.195 2 < 0,001
Style 756.146 4 < 0,001
Spoken
Tense 150.838 2 < 0,001
Person 163.815 2 < 0,001
Print journalism
Tense 176.227 2 < 0,001
Person 37.418 2 < 0,001
Fiction
Tense 647.808 2 < 0,001
Person 97.356 2 < 0,001
of inflexions9, and these figures were labeled as "no-past." The plural and singular forms were also combined as there was little reason to expect this factor to be determinant. In addition, search results provided by the RNC comprise quite short context and it is sometimes impossible to distinguish the form of second person plural from the second person polite form which are homonyms. The final set of analyzed sentences is presented in Table 1.
Results and discussion
All three factors proved to be statistically significant for the choice between null and overt subjects (Table 2).
These results show that Russian null subjects are linked with morphology (i.e., agreement), reference (i.e., person), and discourse features (i.e., topic distribution in different text forms). Moreover, the data - which shows a significant number of null subjects in the initial clauses -refutes Holmberg's statement [2016] that in partial
NSLs null subjects can only be in embedded clauses (or we should admit that Russian is not a partial NSL).
To ascertain the statistical significance of these factors I compared the tense and person in total and within each style (Table 3, the most interesting and unexpected results are italics).
These comparisons verified the role of tense and person for the choice between null and overt subjects, but showed that although correlated, their association is weak in most cases and should be considered only as predisposition and not as some strict rule. At the same time, the data falsified the conventional statement that spoken Russian favors null subjects in contrast to written Russian. Rather, the difference was marked most obviously by the text form (e.g., newspaper versus fiction), which confirms the previous findings of the role
of text form (cf. : [Simonenko, Crabbé, Prévost, 2018], for French). Furthermore, the comparison of first and third persons challenged the previously suggested significance of the person distinction for Russian [Holmberg, Nayudu, Sheehan, 2009], and it supports the idea that referential and generic third person subjects should be considered separately in future studies since they follow different syntactic and pragmatic rules (cf. : [Kinn, Rusten, Walkden, 2016; Rosenkvist, 2018], for Germanic languages).
The only factor that showed a tighter correlation with the choice between null and overt subjects is tense, and it is significant for each style (Table 3). This finding proves that rich agreement -despite being disputed in many works recently [D'Alessandro, 2015] - is meaningful for Russian in each analyzed subcorpus. Russian speakers are
Compared parameters Yates continuity correction Chi-square test Cramer's V
X p-value V association
Total, tense, S(N) / S(Pron) 25G.447 < G,GG1 G.151 weak
Total, tense, S(Pron) / Null 113S.994 < G,GG1 G.329 middle
Total, tense, S(N) / Null 642Ш5 < G,GG1 G.263 middle
Total, 1st vs 2nd persons, S(Pron) / Null 19G.171 < G.141 weak
Total, 2nd vs 3rd persons, S(Pron) / Null 191.495 < G,GG1 G.177 weak
Total, 1st vs 3rd persons, S(Pron) / Null 17.797 < G.G51 no association
Total, spoken / Print journalism, S(N) / S(Pron) 6SS.673 < G,GG1 G.299 middle
Total, spoken / fiction, S(N) / S(Pron) 144.916 < G,GG1 G.13S weak
Total, Print journalism / fiction, S(N) / S(Pron) 1S2.171 < G.165 weak
Total, spoken / Print journalism, S(N) / S(Null) 46G.72G < G,GG1 G.27G middle
Total, spoken / fiction, S(N) / S(Null) 72.594 < G.1G9 weak
Total, Print journalism / fiction, S(N) / S(Null) 161^3 < G,GG1 G.164 weak
Total, spoken / Print journalism, S(Pron) / S(Null) 16.492 < G,GG1 G.G43 no association
Total, spoken / fiction, S(Pron) / S(Null) 16.4S7 < G.G41 no association
Total, Print journalism / fiction, S(Pron) / S(Null) G.G63 0.802 G.GG3 no association
Spoken, tense, S(N) / S(Pron) 67.G3G < G,GG1 G.126 weak
Spoken, tense, S(Pron) / Null 36.2S1 < no association
Spoken, tense, S(N) / Null 151.S99 < G,GG1 G.21S middle
Spoken, 1st vs 2nd persons, S(Pron) / Null 116.269 < G.172 weak
Spoken, 2nd vs 3rd persons, S(Pron) / Null Ш7.Ш < G,GG1 G.192 weak
Spoken, 1st vs 3rd persons, S(Pron) / Null 3.312 0.069 G.G35 no association
Print journalism, tense, S(N) / S(Pron) 0.959 G.GG2 no association
Print journalism, tense, S(Pron) / Null 15G.39G < G.216 middle
Print journalism, tense, S(N) / Null 14S.932 < G,GG1 G.22G middle
Print journalism, 1st vs 2nd persons, S(Pron) / Null 13^ < G.G7S no association
Print journalism, 2nd vs 3rd persons, S(Pron) / Null 34.G37 < G,GG1 G.163 weak
Print journalism, 1st vs 3rd persons, S(Pron) / Null 12.143 < G,GG1 G.GS7 weak
Fiction, tense, S(N) / S(Pron) 362.117 < G.33G middle
Fiction, tense, S(Pron) / Null 34.296 < G,GG1 G.G92 no association
Fiction, tense, S(N) / Null 512.S56 < G.42G quite strong
Fiction, 1st vs 2nd persons, S(Pron) / Null S4.355 < G,GG1 G.159 weak
Fiction, 2nd vs 3rd persons, S(Pron) / Null 4G.342 < G,GG1 G.146 weak
Fiction, 1st vs 3rd persons, S(Pron) / Null G.463 0.497 G.G15 no association
Table 3. Paired comparisons of the choice between null and overt subjects in all tenses, persons, and styles
likely to experience some difficulties to co-reference agents of past actions, but they may rely on inflectional gender information although this factor seems to be skipped in the majority of studies. The strongest association between tense and the choice of nominal overt subject versus null subject was found in fiction. This finding needs further research as fiction follows specific rules of reference, permits sentences much longer than both spoken and journalistic language, favors the serial sequences of actions, etc. All these properties may influence the choice of subject [McShane, 2009].
Conclusion
The statistical analysis of Russian data shows that Russian deviates from expectations that fit NSLs, non-NSLs, or partial NSLs. It allows null subjects in all persons, all analyzed styles, and both main and embedded clauses, as well as possessing a rich agreement system (this matches consistent NSLs). Furthermore, Russian is co-referential to antecedents but they may be extrasentential, and it allows null objects (this matches discourse NSLs). And lastly, although rich in its subject agreement system, Russian is still deficient here and it includes arbitrary generic pronouns including the third person (this matches partial NSLs), while at the same time null subjects are not restricted to any finite verb forms or styles and third person referential null subjects may be found in clauses of any type (in contrast to partial NSLs).
All in all, Russian presents a very complicated example of the Null Subject Phenomenon which consequently needs to be studied by means of corpus and psycholinguistic methods. Together with the data from the history of Russian and from Russian acquisition, these findings show that Russian is likely to be currently undergoing a process of change, but the direction, reasons, and modes of this change are still unclear.
NOTES
1 The work is supported by the research grant number 20-012-00290 'Oral and written narrative as a secondary text: peculiarities of production by various categories of Russian speakers' from the Russian Foundation for Basic Research.
Работа выполнена при поддержке Российского фонда фундаментальных исследований (про-
ект N° 20-012-00290 «Устный и письменный нарра-тив как вторичный текст: особенности порождения разными категориями носителей русского языка»).
2 As I do not discuss the problem of pro in this paper, I from now on use only the term "null subject".
3 RNC is available at www.ruscorpora.ru.
4 They are also called "non-full pro-drop languages" [Cardinaletti, 2012].
5 Holmberg argues that radical NSLs is a false term because it suggests some radical rules of argument eliding [Holmberg, 2016].
6 A detailed analysis of topic prominence and topic chains must stay outside the scope of the present study, although recent studies have shown it to be a significant factor for the choice between null and overt subjects [Frascarelli, Casentini 2019].
7 Kasevich [2004] suggested this term in its phonological meaning.
8 Cf. [Haider, 2019] for a discussion of semantically void arguments and expletives subjects.
9 The forms of imperfective future were eliminated at the previous stage.
REFERENCES
Barbosa P.M., 2011. Pro-Drop and Theories of Pro in the Minimalist Program. Part 2. Prornun Deletion Analyses of Null Subjects and Partial, Discourse and Semi Pro-Drop. Language and Linguistics Compass, vol. 5, iss. 8, pp. 571-587. Barbosa P.M., 2019. Pro as a Minimal nP: Toward a Unified Approach to Pro-Drop. Linguistic Inquiry, vol. 50, iss. 3, pp. 487-526. Bidese E., Tomaselli A., 2018. Developing Pro-Drop: The Case of Cimbrian. Cognola F., Casalicchio J., eds. Null Subjects in Generative Grammar: A Synchronic and Diachronic Perspective. Oxford, Oxford University Press, Oxford Scholarship Online. DOI: 10.1093/oso/ 9780198815853.003.0003 Bizarri C., 2015. Russian as a Partial Pro-Drop Language: Data and Analysis from a New Study. Annali di Ca 'Foscari. Serie occidentale, no. 49, pp. 335-362. Bondarko A.V., 1971. Vid i vremya russlogo glagola [Tense and Aspect of Russian Verb]. Moscow, Prosveshchenie Publ. 239 p. Camacho J.A., 2013. Null Subjects. Cambridge,
Cambridge University Press. 253 p. Cardinaletti A., 2012. Afterword: On Clitic Omission and the Acquisition of Subject Clitic Pronouns. Larranaga P., Guijarro-Fuentes P., eds. Pronouns and Clitics in Early Language. Berlin, De Gruyter Mouton, pp. 283-303. Cognola F., Casalicchio J., 2018. On the Null-Subject Phenomenon. Cognola F., Casalicchio J., eds.
Null Subjects in Generative Grammar: A Synchronic and Diachronic Perspective. Oxford, Oxford University Press, Oxford Scholarship Online. DOI: 10.1093/oso/ 9780198815853.003.0001
D'Alessandro R., 2015. Null Subject. Fabregas A., Mateu J., Putnam M., eds. Contemporary Linguistic Parameters. London, Bloomsbury Press, pp. 201-226.
Duguine M., 2014. Argument Ellipsis: A Unitary Approach to Pro-Drop. The Linguistic Review, vol. 31, iss. 3-4, pp. 515-549.
Franks S., 1995. Parameters of SlavicMorphosyntax. New York, Oxford, Oxford University Press. 432 p.
Frascarelli M., Casentini M., 2019. The Interpretation of Null Subjects in a Radical Pro-Drop Language: Topic Chains and Discourse-Semantic Requirements in Chinese. Studies in Chinese Linguistics, vol. 40, iss. 1, pp. 1-45.
Gordishevsky G., Avrutin S., 2003. Subject and Object Omissions in Child Russian. Proceedings of IATL 19 (Ben-Gurion University of the Negev, 16-17 June 2003). URL: http://linguistics.huji. ac.il/IATL/19/GordishevskyAvrutin.pdf
Haider H., 2019. On Absent, Expletive, and Non-Referential Subjects. Herbeck P., Bernhard Pöll B., Wolfsgruber A.C., eds. Semantic and Syntactic Aspects of Impersonality. Hamburg, Helmut Buske Verlag, pp. 11-46.
Holmberg A., 2005. Is There a Little Pro? Evidence from Finnish. Linguistic Inquiry, vol. 36, pp. 533-564.
Holmberg A., 2016. Null Subjects in Finnish and the Typology of Pro-Drop. Tamm A., Vainikka A., eds. Uralic Syntax. URL: https://andersholmberg 92428242.files.wordpress.com/2020/04/ holmberg-to-appear.-null-subjects-in-finnish-and-the-typology-of-pro-drop.pdf
Holmberg A., Nayudu A., Sheehan M., 2009. Three Partial Nullsubject Languages: A Comparison of Brazilian Portuguese, Finnish and Marathi. Studia Linguistica, vol. 63, pp. 59-97.
Huang C.-T.J., 1984. On the Distribution and Reference of Empty Pronouns. Linguistic Inquiry, vol. 15, pp. 531-574.
Kasevich V.B., 2004. Nuli, Ellipsis, etc. [Nulls, Ellipsis, etc.]. Verbitskaya L.A., ed. Teoreticheskie problemy lingvistiki: k 140-letiyu kafedry obshchego yazykoznaniya Sankt-Peter-burgskogo universiteta [Theoretical Problems of Linguistics: To the 140th Anniversary of the Department of General Linguistics at St Petersburg University]. Saint Petersburg, St. Petersburg University Press, pp. 196-208.
Kinn K., Rusten K.A., Walkden G., 2016. Null Subjects in Early Icelandic. Journal of Germanic Linguistics, vol. 28, iss. 1, pp. 31-78.
Lindseth M., 1998. Null-Subject Properties of Slavic Languages: With Special Reference to Russian, Czech, andSorbian. München, Otto Sagner. 215 p.
Madariaga N., 2018. Diachronic Change and the Nature of Pronominal Null Subjects: The Case of Russian. Cognola F., Casalicchio J., eds. Null Subjects in Generative Grammar: A Synchronic and Diachronic Perspective. Oxford, Oxford University Press, Oxford Scholarship Online. DOI: 10.1093/oso/9780198815853.003.0007
McShane M., 2009. Subject Ellipsis in Russian and Polish. Studia Linguistica, vol. 63, pp. 98-132.
Nagy N. G., Aghdasi N., Denis D., Motut A., 2011. Null Subjects in Heritage Languages: Contact Effects in a Cross-Linguistic Context. University of Pennsylvania Working Papers in Linguistics, vol. 17, iss. 2, art. 16.
Perlmutter D.M., Moore J., 2002. Language-Internal Explanation: The Distribution of Russian Impersonals. Language, vol. 78, pp. 619-650.
Pinto M., 2006. Subject Pronouns in Bilinguals: Interference or Maturation? Torrens V., Escobar L., eds. The Acquisition of Syntax in Romance Languages, Amsterdam, John Benjamins Publishing Company, pp. 331-352.
Rizzi L., 1982. Issues in Italian Syntax. Dordrecht, Foris. 188 p.
Rosenkvist H., 2018. Null Subjects and Distinct Agreement in Modern Germanic. Cognola F., Casalicchio J., eds. Null Subjects in Generative Grammar: A Synchronic and Diachronic Perspective. Oxford, Oxford University Press, Oxford Scholarship Online. DOI: 10.1093/oso/ 9780198815853.003.0012
Rûzicka R., 1986. Funkcionirovanie i klassifikatsia pustykh kategoriy v russkom literaturnom yazyke [Functions and Classification of Empty Categories in Russian Standard Language]. Zeitschrift für Slavistik, vol. 31, pp. 388-392.
Simonenko A., Crabbé B., Prévost S., 2018. Text Form and Grammatical Changes in Medieval French: A Treebank-Based Diachronic Study. DIACHRONICA, vol. 35, iss. 3, pp. 393-428.
Tseytlin S.N., 1976. Stroenie predlozheniya i rechevaya situatsia (k probleme elliptichnosti predlozheniya) [The Structure of Sentence and the Situation of Communication (To the Problem of Elliptic Sentences)]. Funktsionalny analiz grammaticheskikh kategoriy i edinits [Functional Analysis of Grammatical Categories and Elements]. Leningrad, Herzen Institute Publishing House, pp. 37-46
Wratil M., 2011. Uncovered Pro - On the Development and Identification of Null Subjects. Wratil M., Gallmann P., eds. Null Pronouns. Berlin, Walter de Gruyter, pp. 99-140.
Information About the Author
Polina M. Eismont, Candidate of Sciences (Philology), Associate Professor, Ludmila Verbitskaya Department of General Linguistics, St. Petersburg University, Universitetskaya Emb., 7-9-11, 199034 Saint Petersburg, Russia, [email protected], https://orcid.org/0000-0002-0595-1847
Информация об авторе
Полина Михайловна Эйсмонт, кандидат филологических наук, доцент кафедры общего языкознания им. Л.А. Вербицкой, Санкт-Петербургский государственный университет, Университетская наб., 7-9-11, 199034 г. Санкт-Петербург, Россия, [email protected], https://orcid.org/0000-0002-0595-1847
