Current studies in Slavic linguistics 9789027206138, 9789027270962, 9027206139

Table of contents :
1. List of contributors
2. Abbreviations
3. Impersonals and Beyond in Slavic (by Guiraud-Weber, Marguerite)
4. Section I: Morphosyntax
5. Binding and Morphology Revisited (by Franks, Steven L.)
6. Possessor Raising and Slavic clitics (by Zimmerling, Anton)
7. The Slavonic Languages and the Development of the Antipassive Marker (by Janic, Katarzyna)
8. Clitic SE in Romance and Slavonic revisited (by Marelj, Marijana)
9. Section II: Syntactical relations
10. The Lazy Speaker and the Fascination of Emptiness: Colloquial Russian from a Typological Perspective (by Weiss, Daniel)
11. Is the Polish Verb isc an Auxiliary to be?: A Corpus-Based Study of the Construction isc + Infinitive (by Sikora, Dorota)
12. Towards Evidentiality Markers in Albanian and Macedonian Bilingual Political Discourse (by Makartsev, Maxim)
13. A strange variant of Russian ctoby-construction: Irreality and tense-marking (by Letuchiy, Alexander)
14. Section III: Impersonal constructions
15. Impersonal Constructions in Serbian: A description within a Meaning Text linguistic model (by Milicevic, Jasmina)
16. Interpretation and voice in Polish SIE and -NO/-TO constructions (by Krzek, Malgorzata)
17. Dative-infinitive constructions in Russian: Taxonomy and semantics (by Israeli, Alina)
18. On the Nature of Dative Arguments in Russian Constructions with "Predicatives" (by Say, Sergey)
19. Russian Adversity Impersonals and Split Ergativity (by Schlund, Katrin)
20. Section IV: Lexical semantics
21. Morphological and lexical aspect in Russian deverbal nominalizations (by de Valdivia, Gloria)
22. Lexical synonymy within the semantic field POWER (by Dobrovol'skij, Dmitrij)
23. Collocations with nominal quantifiers: Semantics and combinability (by Beliakov, Vladimir)
24. Polysemy Patterns in Russian Adjectives and Adverbs: A corpus-oriented database (by Reznikova, Tatiana)
25. Language index
26. Name index
27. Subject index

Citation preview

Current Studies in Slavic Linguistics

Studies in Language Companion Series (SLCS) This series has been established as a companion series to the periodical Studies in Language. For an overview of all books published in this series, please see http://benjamins.com/catalog/slcs

Editors Werner Abraham

University of Vienna / University of Munich

Elly van Gelderen

Arizona State University

Editorial Board Bernard Comrie

Max Planck Institute, Leipzig and University of California, Santa Barbara

William Croft

University of New Mexico

Östen Dahl

University of Stockholm

Gerrit J. Dimmendaal University of Cologne

Ekkehard König

Free University of Berlin

Volume 146 Current Studies in Slavic Linguistics Edited by Irina Kor Chahine

Christian Lehmann University of Erfurt

Marianne Mithun

University of California, Santa Barbara

Heiko Narrog

Tohuku University

Johanna L. Wood

University of Aarhus

Debra Ziegeler

University of Paris III

Current Studies in Slavic Linguistics Edited by

Irina Kor Chahine Université de Provence

John Benjamins Publishing Company Amsterdam / Philadelphia

8

TM

The paper used in this publication meets the minimum requirements of the American National Standard for Information Sciences – Permanence of Paper for Printed Library Materials, ansi z39.48-1984.

Library of Congress Cataloging-in-Publication Data Current Studies in Slavic Linguistics / Edited by Irina Kor Chahine. p. cm. (Studies in Language Companion Series, issn 0165-7763 ; v. 146) Includes bibliographical references and index. 1. Slavic languages--Syntax. 2. Slavic languages--Grammar. 3. Slavic languages--History. I. Chahine, Irina Kor. PG201.C87 2013 491.8’045--dc23 2013034842 isbn 978 90 272 0613 8 (Hb ; alk. paper) isbn 978 90 272 7096 2 (Eb)

© 2013 – John Benjamins B.V. No part of this book may be reproduced in any form, by print, photoprint, microfilm, or any other means, without written permission from the publisher. John Benjamins Publishing Co. · P.O. Box 36224 · 1020 me Amsterdam · The Netherlands John Benjamins North America · P.O. Box 27519 · Philadelphia pa 19118-0519 · usa

Table of contents List of contributors Abbreviations Impersonals and beyond in Slavic Marguerite Guiraud-Weber & Irina Kor Chahine

vii ix 1

section i.  Morphosyntax Binding and morphology revisited Steven Franks

25

Possessor Raising and Slavic clitics Anton Zimmerling

43

The Slavonic languages and the development of the antipassive marker Katarzyna Janic

61

Clitic SE in Romance and Slavonic revisited Marijana Marelj & Eric Reuland

75

section ii.  Syntactical relations The lazy speaker and the fascination of emptiness: Colloquial Russian from a typological perspective Daniel Weiss

91

Is the Polish verb iść an auxiliary to be? A corpus-based study of the construction iść + infinitive Dorota Sikora

123

Towards evidentiality markers in Albanian and Macedonian bilingual political discourse Maxim Makartsev

139

A strange variant of Russian čtoby-construction: Irreality and tense-marking Alexander Letuchiy

149

 Current Studies in Slavic Linguistics

section iii.  Impersonal constructions Impersonal constructions in Serbian: A description within a Meaning Text linguistic model Jasmina Milićević

169

Interpretation and voice in Polish SIĘ and –NO/–TO constructions Małgorzata Krzek

185

Dative-infinitive constructions in Russian: Taxonomy and semantics Alina Israeli

199

On the nature of dative arguments in Russian constructions with “predicatives” Sergey Say Russian Adversity impersonals and split ergativity Katrin Schlund

225 247

section iv.  Lexical semantics Morphological and lexical aspect in Russian deverbal nominalizations Glòria de Valdivia, Joan Castellví & Mariona Taulé

267

Lexical synonymy within the semantic field POWER Dmitrij Dobrovol’skij & Ludmila Pöppel

281

Collocations with nominal quantifiers: Semantics and combinability Vladimir Beliakov

297

Polysemy patterns in Russian adjectives and adverbs: A corpus-oriented database Tatiana Reznikova, Ekaterina Rakhilina, Olga Karpova, Maria Kyuseva, Daria Ryzhova & Timofey Arkhangelskiy

313

Language index

323

Name index

325

Subject index

329

List of contributors

Timofey Arkhangelskiy National Research University Higher School of Economics Khitrovskiy pereulok, dom 2/8 korpus 5, room 308 109028 Moscow Russia [email protected]

Steven Franks Dept of Linguistics Memorial Hall 322 Indiana University 1021 E. Third Street Bloomington, IN 47405-7005 USA [email protected]

Vladimir Beliakov Université Toulouse II CLLE-CNRS (UMR 5263) 1 avenue des Romarins 31470 Fonsorbes France [email protected]

Marguerite Guiraud-Weber Aix-Marseille University Résidence Le Chambord, bâtiment I avenue René Cassin 13100 Aix-en-Provence France [email protected]

Joan Castellví Dep. Lingüística General Gran Via 585 08007 Barcelona Spain [email protected]

Alina Israeli WLC, American University 4400 Massachusetts Ave., NW, Washington DC 20016 USA [email protected]

Dmitrij Dobrovoľskij Russian Language Institute Russian Academy of Sciences 18/2 ul. Volkhonka 119019 Moscow Russia [email protected]

Katarzyna Janic Université Lyon 2 86 rue Pasteur 69007 Lyon France [email protected]

 Current Studies in Slavic Linguistics

Olga Karpova Graduiertenschule für ­Geisteswissenschaften Göttingen Historische Sternwarte Geismar Landstraße 11 37083 Göttingen Germany [email protected] Irina Kor Chahine Université de Nice – Sophia Antipolis UFR LASH – Département de Russe 98 boulevard Édouard Herriot 06204 Nice Cedex 3 France [email protected] Malgorzata Krzek Linguistics Section, SELLL Percy Building, Newcastle University Newcastle NE1 7RU United Kingdom [email protected] Maria Kyuseva National Research University Higher School of Economics Khitrovskiy pereulok, dom 2/8 korpus 5, room 308, 109028 Moscow Russia [email protected] Alexander Letuchiy National Research University Higher School of Economics Khitrovskiy pereulok, dom 2/8 109028 Moscow Russia [email protected]

Maxim Makartsev Institut slavyanovedenija RAN Leninsky pr-t, 32a 119334 Moscow Russia [email protected] Marijana Marelj Utrecht Institute of Linguistics OTS Utrecht University Trans 10 3512 JK Utrecht The Netherlands [email protected] Jasmina Milićević French Department Dalhousie University 6135 University Avenue Halifax (NS) B3H 4P9 Canada [email protected] Ludmila Pöppel Institute of Slavic Languages and Literatures Stockholm University SE-10691 Stockholm Sweden [email protected] Ekaterina Rakhilina National Research University Higher School of Economics Khitrovskiy pereulok, dom 2/8, korpus 5, room 308 109028 Moscow Russia [email protected]



List of contributors 

Eric Reuland Utrecht Institute of Linguistics OTS Utrecht University Achter de Dom 20 3512 JP Utrecht The Netherlands [email protected]

Dorota Sikora ATILF-CNRS (UMR 7118) 44, avenue de la Libération BP 30687 54063 Nancy Cedex France [email protected]

Tatiana Reznikova National Research University Higher School of Economics Khitrovskiy pereulok, dom 2/8 korpus 5, room 308 109028 Moscow Russia [email protected]

Mariona Taulé Dep. Lingüística General Gran Via 585 08007 Barcelona Spain [email protected]

Daria Ryzhova National Research University Higher School of Economics Khitrovskiy pereulok, dom 2/8 korpus 5, room 308 109028 Moscow Russia [email protected] Sergey Say Institute for Linguistic Studies Russian Academy of Sciences Tuchkov per. 9 199004 St. Petersburg Russia [email protected] Katrin Schlund Heidelberg University Slavisches Institut Schulgasse 6 69117 Heidelberg Germany [email protected]

Glòria de Valdivia Dep. Lingüística General Gran Via 585, 08007 Barcelona Spain [email protected] Daniel Weiss Slavisches Seminar der Universität Zürich Plattenstr. 43 CH-8032 Zürich Switzerland [email protected] Anton Zimmerling Institute for Modern Linguistic Research, SMSUH Smolnaya str. 37-88 125195 Moscow Russia [email protected]

Abbreviations

1 first person 2 second person 3 third person a agent-like argument abl ablative abs absolutive acc accusative adj adjective adv adverb(al) aor aorist ap antipassive aux auxiliary com comparative comp complementizer cop copula dat dative def definite dir direct erg ergative f feminine fut future gen genitive imp imperative ind indicative indef indefinite indir indirect inf infinitive

intr intransitive impers impersonal inst instrumental ipf imperfective loc locative m masculine neg negation neut neuter nom nominative obj object obl oblique part participial pass passive pf perfective poss possessive pred predicative prep preposition prs present pl plural pst past q question particle/marker ref reflexive sbj subject sg singular tr transitive v verb voc vocative

Impersonals and beyond in Slavic Marguerite Guiraud-Weber & Irina Kor Chahine

Aix-Marseille University / University of Nice – Sophia Antipolis, France

This book grew out of the selection of papers presented during the 6th Annual Meeting of the Slavic Linguistics Society (2011, Aix-en-Provence, France). The Meetings of the SLS include presentations in both theoretical and applied approaches, and in keeping with this tradition in this volume we have provided equal consideration to various issues of Slavic linguistics, from morphology to syntax and semantics. While the papers collected in this volume attempt to throw light on various aspects of Slavic linguistics, often from a typological point of view, most of them have an underlying common subject, namely expression of the impersonal. Far from being a marginal phenomenon, the means of expressing the impersonal is one of the main issues in Slavic languages. This topic is also closely related to other questions discussed in this volume, such as zero sign, passive voice, and lexis. For this reason, although they are inevitably grouped into several chapters, most ­articles often discuss issues that are partially related to each other. The issue of the impersonal can be considered in a variety of ways, i.e. it may refer to grammatical person, verbal syntax, or voice usage. An extensive ­bibliography has been developed on this question. Impersonal constructions constitute only a part of this vast domain. We know that these constructions exist to some extent in various languages. In some languages, such as Arabic, Persian, or Japanese, they are relatively rare, while in others, like Slavic languages, they turn out to be especially abundant. However, even in the Slavic languages this domain is varied. Among five Slavic languages belonging to three different groups, namely Russian, Polish, Czech, Serbo-Croatian,1 and Bulgarian, it is in Russian that the system of impersonal constructions is most developed. These constructions exist to a lesser extent in other Slavic languages. It should be noted that the development of impersonal constructions does not always appear to be a result of geographical language contact. Nonetheless, it is necessary to conduct further research to establish a detailed typology of impersonal constructions in the Slavic languages.

.  Here we do not make a distinction between Serbian and Croatian.



Marguerite Guiraud-Weber & Irina Kor Chahine

Impersonal constructions have been well known since antiquity; in particular they are well attested in Greek and Latin. In modern European languages, among non-referential markers of impersonal constructions we can point out it in ­English (it’s raining), es in German (es regnet), or even the non-referential il in French (il pleut) and ça in colloquial French (ça tonne) (Maillard 1985). Often used to express meteorological phenomena or sensations (cf. Malchukov & Siewierska 2011), these constructions are based on verbal syntax that allows either a personal reading, or an impersonal one, depending on context (je me souviens…/il me souvient…). In Slavic languages, the domain of impersonal constructions is probably the richest in comparison to other European languages. Impersonal constructions in Slavic languages are very common and include various syntactic models with a large range of semantics, including expressing existence, sensations, physical or mental states, object of communication, and modality. Moreover, these constructions cannot be considered exclusively in the context of verbal syntax, because the core predicate can be expressed not only by a verb, but also by a predicative form (noun or adverb serving as a predicate), or even by the entire construction. Below we will briefly discuss general issues associated with impersonal ­constructions in Slavic languages and will present an overview of this vast domain in order to provide a necessary context for this collection of papers. Moreover, the topic of the impersonal will allow us to consider by extension other issues covered in this volume. It’s worth mentioning that articles included in this volume on Slavic languages contribute to the discussion about impersonal and related constructions recently compiled by Andrej Malchukov and Anna Siewierska and edited in 2011 by John Benjamins, which concerns a wide range of languages and linguistic families. 1.  Impersonal sentences: A problem of definition Impersonal sentences are attested in all Slavic languages, although their variants do not appear in every language, and their frequency is not the same in different languages. Therefore, it is a heterogeneous domain, despite the fact that they originated from the same archaic Balto-Slavic model derived from the Indo-European one. Their heterogeneity and diversity have been examined in multiple studies devoted to this issue; however, Slavic impersonal sentences have not yet received a completely exhaustive description. Diachronic studies have brought to light a variety of ways the impersonal ­sentences have evolved in each language. Historical data show that very often an identical or quasi-identical structure in two different languages is not necessarily the same linguistic phenomenon, and that visible resemblances may be



Impersonals and beyond in Slavic

misleading. It is therefore necessary to acknowledge that the role an impersonal construction plays in the system of one language may not be the same that it plays in another. This represents a serious difficulty for potential comparison between languages. Another difficulty concerns terminology. The term impersonal refers to various syntactic constructions, and, according to the linguistic schools and traditions in each country, is used differently. As early as in 1883 Franz Miklosich introduced the term of “subjectlose Sätze” to underline the absence of subject as a common property of these sentences, and since then, this term has known some development, in particular in the Czech and Polish schools. However, when referring to the notion of “sentences without a subject,” we have to know what subject is, because this notion is not always clearly defined and remains implicit in many studies. An attempt to provide further clarification consists in talking of sentences without the nominative case, since nominative is typically the case of the subject (Wolińska 1978; Guiraud-Weber 1984). However, this formulation is not entirely satisfactory either. It should be indeed clarified that although the nominative is certainly the standard case for grammatical subject, it can also play other syntactic roles. Despite the inconvenience of the term impersonal, it has finally become accepted and universally recognized; in particular within studies written in ­Russian and English. It should be noted, however, that this is a purely conventional term, and many scholars have shown its inadequacy (Zolotova 1973; ­Creissels 1991; M ­ alchukov & Siewierska 2011, and others). 2.  The issue of subject The conventional term of impersonal sentence is now used, on the one hand, merely as a label applying to all sentences that do not conform to the basic pattern of subject in nominative + agreed predicate; on the other hand, it is used in a narrow sense for designating sentences that we would call purely impersonal, and which we cannot classify according to morphological properties of the predicate alone, as was proposed in Galkina-Fedoruk 1958 (and by other scholars p ­ ursuing this tradition), but according to syntactic and semantic properties of impersonal sentences as a particular syntactic structure. In this structure, in addition to the predicate, there is very often a recurring nominal element in an oblique case. The latter refers to the object or person concerned by the events designated by the predicate. And in this case, the non-agreed predicate takes the form of 3rd person of singular in present or the form of neuter then it is in past tense. It is possible to show that the strictly impersonal sentence, regardless of the variant it represents, always excludes controlled action by an animate agent. Thus, the





Marguerite Guiraud-Weber & Irina Kor Chahine

lack of n ­ ominative, which is the only case appropriate for designating voluntary action by an animate agent, is semantically justified. If the events designated by the predicate refer to a person, the person will be expressed by an oblique case and play the role of a simple observer, a passive participant, or, at best, an experiencer, but never of an active agent. It has long been observed that some oblique cases in impersonal sentences behave similarly to a nominative subject. This is especially the case for dative (type (1a)) (cf. Say, this volume), to a lesser extent for accusative (type (1b)), and even for instrumental (type (1c)) (cf. Schlund, this volume). (1) Russian a. Mne dušno. I:dat suffocate:pred ‘I’m suffocating.’ b. Menja znobit. I:acc chilly:prs.3sg ‘I feel chilly.’ c. Vetrom sorvalo kryšu. wind:inst tore-off:neut roof:acc ‘The wind tore off the roof.’

These similarities are at the origin of a new term in Russian linguistics, namely sub”ekt (“semantic or logical subject”), which is opposed to podležaščee (grammatical subject). The frequency of these terms in Russian linguistics used in all Russian textbooks and papers shows the importance of this phenomenon in the Russian language, which explains why some linguists attempt to describe syntax in semantic rather than in formal terms (Zolotova 1973). Studies devoted to the notion of subject in different world languages, in ­particular those by Edward Keenan (1976) and Bernard Comrie (1981), have shown that there is a certain number of morphological, syntactic, and semantic proprieties that allows for identifying the subject. These features are not the same in all languages, but can be similar. If, in the case of Russian, we temporarily renounce to define the subject in discrete terms (as an argument that agrees with the verb in number, person and gender), it becomes apparent that certain subject properties (such as indespensability, control over the reflexive pronouns, position in front of the predicate, topic of the sentence) not only characterize the subject in nominative, but can occur in other oblique cases, in which case the latter play the role of the argument enabling the predicate quality. This approach requires a scalar definition of the subject, such as some arguments will be considered more “subject” than others and will be placed closer to the nominative in a subjectival scale. From this point of view, it becomes possible to justify the linguistic insight that



Impersonals and beyond in Slavic

underlies the sub“ekt/podležaščee dichotomy. This approach has been adopted for Russian in Testelec 2001 and Guiraud-Weber 2003. It supports preferential treatment of certain nominal arguments in oblique cases within impersonal sentences because these arguments express the quality of the predicate. However, only Nominative (or its syntactic equivalents) can be considered as grammatical subject in Slavic. The choice of other cases is determined by ­general semantics of the sentence. Thus, on examples from Slavic languages we can see that Genitive indicates the absence or non-existence; Dative, and sometimes Accusative case, designates physical and mental state; Instrumental characterizes the inanimate destructive force, and Locative – a space. In addition, the two latter cases can be used with prepositions to designate other specific situations (Guiraud-Weber 1984). This fact shows that not only the lexis, but also the ­syntactic structure takes part in forming the meaning of the sentence. 3.  E  volution of impersonal sentences in Russian and across other Slavic languages The domain of impersonal sentences in Russian is very broad, especially in comparison to other Slavic languages. Over many centuries it has continued to become richer. Historical studies have demonstrated the following trends: 1. Expansion of the negative existential sentences (see (2a)). Studies by N. Ju. Švedova have shown that lexical means of predicates in these constructions have been continually renewed ever since the 18th century (Švedova 1964). This expansion, involving a great number of verbs in modern Russian, is described in a number of publications (Ickovič 1974; Babby 1980; Padučeva 1992; Partee & Borschev 2007; Guiraud-Weber 2003); 2. Emergence, at the end of 19th century, of constructions with the nominal argument “s + instrumental” (see (2b)), obviously related to the deletion of the desemantized subject, like delo ‘fact’, položenie ‘situation’, vopros ‘question’ (Zolotova 1966); 3. Decline of passive impersonal constructions and their substitution with impersonal constructions using the nominal constituent “o + locative” (see (2c)) (Švedova 1964); 4. Lexical and syntactic reorganization of constructions with dative argument (see (3c–d)), which since the 19th century has had the semantic function of experiencer replacing the construction “dlja + genitive” (see (3a–b)). During the 19th century, serious confusion prevailed in this domain, and variant with dlja are not any more attested in today’s Russian (Švedova 1964).





Marguerite Guiraud-Weber & Irina Kor Chahine

(2) a. Ego ne vidno. he:acc neg visible:pred ‘He is not visible.’ b. S den’gami bylo ploxo. with money:inst was:aux.neut tight:pred ‘Money was tight.’ c. Ob ètom govorilos’ v gazetax. about this:loc discussed:neut.ref in newspapers:loc ‘This was discussed in the newspapers.’ (3) a. Dlja nego kazalos’ strannym for he:gen seemed:neut strange:inst ‘It seemed strange to him.’ b. Dlja nego bylo prijatno videt’ for he:gen was:aux.neut nice:pred see:inf ‘It was nice for him to see.’ c. Emu kazalos’ strannym he:dat seemed:neut strange:inst ‘It seemed strange to him.’ d. Emu bylo prijatno videt’. he:dat was:aux.neut nice:pred see:inf ‘It was nice for him to see.’

In West Slavic languages, in particular in Polish and Czech, the domain of strictly impersonal sentences appears to be less extensive; while constructions with indefinite human subjects – “man-constructions” – are abundant and diverse. In South Slavic languages, like in Serbo-Croatian, the domain of impersonal sentences is more limited, and it is especially restricted in Bulgarian because of the absence of noun declension. What follows is a brief survey of the main types of impersonal sentences in Slavic languages. Impersonal sentences describing environment exist in all Slavic languages: (4) a. Ru. Na ulice xolodno. in street:loc cold:pred ‘It’s cold outside.’ b. Bg. Dušno. stuffy:pred ‘It’s stuffy.’ c. S-Cr. Smrklo se. grown-dark:neut ref ‘It has grown dark.’

They regularly occur without any nominal support, but necessarily refer to a space, often where the speaker is located (according to deictic rules). In such i­ mpersonal



Impersonals and beyond in Slavic

sentences, the situations are always localizable, unlike non-localizable situations depicting feelings, emotions, beliefs, or psychical states of individuals (cf. Boguslavskij 1991). Thus, in this type of impersonal sentence, the nominal element referring to location is not just an adverbial phrase which, by definition, could be deleted. It semantically represents necessary support to the predicate, without which there would be no message. S. Say (this volume) addresses in part the relation of these predicatives referring to location, which he considers to be “objective”, with predicatives that are more likely to use dative, and which he considers “subjective” (i.e. (5)). The first type of sentences is usually accounted for as “meteorological phenomena” (cf. Galkina-Fedoruk 1958; Birjulin 1994) even though it also comprises sentences like (6a–b) have nothing to do with weather events. They simply describe the space concerned by the conversation, like construction (6c) where the space is characterized from the olfactory perspective. Thus the latter construction should also be included in this category (­Guiraud-Weber 1979). (5) Emu xolodno. he:dat cold:pred ‘He is cold.’ (6) a. V komnate pusto. in room:loc empty:pred ‘The room is empty.’ a. Na lestnice tixo. on stairway:loc quiet:pred ‘The stairway is quiet.’ b. V komnate paxnet jablokami. in room:loc smell:prs.3sg apples:inst ‘It smells like apples in the room.’

Sentences of negative existence are extremely frequent in Russian (see (7) below) and involve a large number of verbs, including verbs that are not traditionally perceived as verbs of existence (vyjti ‘to go out’, projti ‘to pass though’, vypast’ ‘to fall’, rasti ‘to grow’, etc.). In Polish, genitive may occur only in the existential sentences with the verbs być and zostać (both of which also function as auxiliary verbs). The verb ‘to be’ contains two forms: być and mieć, functioning in complementary distribution (8a, b). No other verb can serve as a predicate of existence. However, unlike Russian, Polish generalizes the usage of genitive for the direct object under negation, which is used in all circumstances and is even extended to adverbial modifier in colloquial Polish (8c, d) (Guiraud-Weber 2003; Guiraud-Weber  & Zaremba 2007). In Serbo-Croatian, existential sentences use genitive when the sentence is negative but have a tendency to use this case in affirmative sentences as well (9a,  b). However this construction is only limited to one single verb of existence represented by forms of the verbs imati – biti, used in complementary





Marguerite Guiraud-Weber & Irina Kor Chahine

distribution like in Polish: the verb imati ‘to have’ is used in present tense, and biti ‘to be’ in other verbal forms (9c, d). In contrast, in Czech, sentences of negative existence have disappeared. Their decline began in the 17th century, and is now complete (Hausenblas 1958). Even negative sentences with the verb ‘to be’ use nominative in Czech (10) (Žaža 1978). (7) Russian Otca ne bylo doma father:gen neg was:neut at-home ‘The father was not at home.’ (8) Polish a. Jest / był chleb is:3sg / was:3sg.m bread:nom.m ‘There is/was some bread.’ b. Nie ma / nie było chleba neg have:prs.3sg / neg was:neut bread:gen ‘There is/was no bread.’ c. Nie chcę zupy neg want:prs.1sg soup:gen ‘I do not want soup.’ d. Nie pracowalem calej nocy neg worked:1sg all:gen night:gen ‘I did not work all night.’ (9) Serbo-Croatian a. Nema knjige. have:3sg.neg book:gen ‘There is no book.’ b. Ima knjige / knjiga. have:3sg book:gen / book:nom.f ‘There is a book.’ c. Nema ga have:prs.3sg.neg he:gen ’He is not here.’ d. Nije bilo knjige neg was:neut book:gen ‘There was no book.’ (10) Czech Otec nebyl doma father:nom.m was:3sg.m.neg at-home ‘The father was not at home.’

All Slavic languages use the dative case for experiencer with predicates describing a person’s physical or mental state (1a), (5), (11a–d):



Impersonals and beyond in Slavic

(11) a. Pl. Miło mi pleased:pred I:dat ‘I like it.’ b. Cz. Petrovi otrnulo Peter:dat felt-better:neut ‘Peter feels better.’ c. Bg. Težko ti hard:pred you:2sg.dat ‘It’s hard for you.’ d. S-Cr. Ne radi im se neg work:prs.3sg they:dat ref ‘They cannot work.’

In West Slavic languages, most impersonal constructions with modal predicate (such as Russian (12a–b)) have been replaced by personal constructions with loan-translations from German müssen, dürfen, and haben. Thus the Czech verbs muset, smĕt, and mít, like their Polish counterparts musieć and mieć, require a nominative ­subject (Wolińska 1978; Weiss 1987, 1988; Žaža 1998). (12) Russian a. Mne nado rabotat’ I:dat need:pred work:inf ‘I need to work.’ b. Emu nel’zja kurit’ he:dat should-not:pred smoke:inf ‘He should not smoke.’

Russian negative constructions, such as (13a), with a dative experiencer and a negative predicative construction are also present in Bulgarian (13b). In Czech, this construction appears both in negative and affirmative forms (13c) (Mrázek 1990; Žaža 2002), which is also the case in Serbo-Croatian (13d, e). Polish, however, does not have similar constructions. (13) a. Ru. Irine ne do smexa. Irina:dat neg prep laugh:gen ‘Irina does not feel like laughing.’ b. Bg. Ne mi e do šega. neg I:dat is:3sg prep joke:gen ‘I do not feel like joking.’ c. Cz. Petrovi bylo / nebylo do smíchu. Peter:dat was:neut / was:neut.neg prep laugh:gen ‘Peter feels/does not feel like laughing.’



 Marguerite Guiraud-Weber & Irina Kor Chahine

d. S-Cr. Irini je do smeha. Irina:dat is:3sg prep laugh:gen ‘Irina feels like laughing.’ e. S-Cr. Nije Irini do smeha. is:neg Irina:dat prep laugh:gen ‘Irina does not feel like laughing.’

Other types of impersonal sentences with a dative experiencer are “lacking” in West Slavic languages. Originally occurring in Old Slavic, Russian infinitive sentences like (14a, b), are frequent nowadays (Veyrenc 1979; Maurice 1996; ­Fortuin 2000), but as shown by A. Israeli (this volume), who offers a detailed and well illustrated study of infinitive constructions in Russian, their usage remains syntactically and stylistically limited. (14) Russian a. Emu xodit’. he:dat play:inf ‘It is his turn (to play).’ b. Vam ne ponjat’. you:2pl.dat neg understand:inf ‘You will not be able to understand.’

The dative plus infinitive construction is neither common nor productive in other Slavic languages. Only some relics of it remain in Polish (Wolińska 1978; Weiss 1993) and in Serbo-Croatian (Vojvodić 2007) while it seems to have completely disappeared in Czech (Žaža 1998). Thus, there is an obvious disparity between impersonal sentences with a dative experiencer that are widely used in Slavic ­languages on the one hand, and, on the other hand, infinitive impersonal sentences with dative attested only in some Slavic languages. The two constructions also present functional differences. This leads some scholars to suggest that these constructions contain two different types of dative “resulting from distinct case-marking strategies” (Franks 1995). In this volume, S. Say suggests some possible origins of the dative constituent with Russian predicatives (adverbial predicate). On the basis of semantic relationship between these predicatives and the argument structure of the corresponding short and long adjectives, he shows that they represent a heterogeneous class of elements and distinguishes three possible strategies. Unlike the dative experiencer discussed earlier, the accusative argument is much less common in impersonal sentences in Slavic languages. This construction appears in Russian with a limited number of verbs (15a-b). The accusative argument should not be confused with a direct object in accusative because the



Impersonals and beyond in Slavic

impersonal sentence uses an intransitive verb (15c-d). This type of accusative is thus an argument of an intransitive verb. In contrast, in the construction (17a) below, the accusative kryšu is a regular direct object of the transitive verb sorvat’. The construction with the accusative occurs with a much more limited number of predicates in other Slavic languages (16): (15) Russian a. Ego lixoradit. he:acc tremble:prs.3sg ‘He has a fever.’ b. Eë peredërnulo. she:acc shuddered:neut ‘She shuddered.’ c. *Ja ego tošnju.  I:nom he:acc make-sick:prs.1sg Intended: ‘*I make him sick.’ d. *On eë peredërnul.  he:nom she:acc shuddered:3sg.m Intended: ‘*He shuddered her.’ (16) a. Pl. Poderwało go. shuddered:neut he:acc ‘He shuddered.’ b. S-Cr. Strah vas bilo. scary:pred you:2pl.acc was:aux.neut ‘It was scary.’ c. Bg. Trese me. tremble: prs.3sg I:acc ‘I am trembling (from fever).’

Russian constructions with the instrumental case depicting damage caused by a natural phenomenon or an unknown force (17a–b) have a corresponding ­construction without instrumental in West Slavic languages (18a–b, 19a–b). (17) Russian a. Vetrom sorvalo kryšu wind:inst tore.off:neut roof:acc ‘The wind tore off the roof.’ b. Ego ranilo oskolkom he:acc wounded:neut shrapnel:inst ‘He was wounded by shrapnel.’



 Marguerite Guiraud-Weber & Irina Kor Chahine

(18) Polish a. Zasypało drogę. covered:neut road:acc ‘The road was covered.’ b. Zalało pole. flooded:neut field:acc ‘The field was flooded.’ (19) Czech a. Bratra zabilo. brother:acc killed:neut ‘My brother was killed.’ b. Souseda ranilo. neighbor:acc injured:neut ‘The neighbor was injured.’

According to most scholars of comparative Slavic studies, the “impersonal” instrumental, which is used, as we know, for a meteorological phenomenon or a force endowed with its own energy, takes a central place in East Slavic languages, in particular in Russian, but it has a rather peripheral status in other Slavic languages (Mrázek 1964; Ivić 1965). From the structural point of view, we can refer to R. Mrázek, who was aware of the specificity of the “impersonal” instrumental and its structurally indispensability, and compared it to a subject (Mrázek 1964), although this form should be placed at the far end of the subjectival scale (see above, § 2). On the other hand, from the typological point of view, it is possible to compare these models to ergative constructions that assign an oblique case to the subject of transitive verbs. Thus, K. Schlund (this volume) attempts to show the origin of Slavic instrumental constructions using formal, semantic, and pragmatic markers that make them similar to ergative constructions2 (in particular the example of Hindi is used), and discusses a possible correlation between the emergence of impersonal sentences in Slavic and the grammaticalization of the verb ‘to have’. Finally, since the neutralization of the role of agent is proper not only to ergative and impersonal but also to passive constructions we will return to this issue later (see § 5).

.  Another paper in this volume presents a comparative study between Slavic (accusative) languages and ergative languages (cf. Janic, this volume).



Impersonals and beyond in Slavic 

4.  Impersonal and indefinite subject It is worth mentioning another problem related to identifying impersonal constructions. In order to identify strictly impersonal sentences, we must eliminate all instances when the subject is missing due to the influence of context or pragmatic, lexical, and stylistic constraints, i.e. all elliptical or deleted subjects. The cases of zero subject must be eliminated as well. The ellipsis and zero subject always denote a significant absence. The elliptical subject may however be reinstated, which is not possible for zero subject: no lexeme can substitute a zero subject without changing the meaning of the sentence. Among sentences without a nominative subject, i.e. impersonal sentences “in a large sense,” are those types which refer to a generic or indeterminate subject. These sentences have different characteristics in each Slavic language, and their typology is yet to be examined. In Russian there are two well-known variants of such sentences: one with zero subject, designating a person or a group of indeterminate persons, in principle excluding the speaker (20); and one referring to a “universal person,” which can include the speaker (21). (20) Na zavode bastovali at factory:loc striked:3pl ‘  There was a strike at the factory.’ (21) Ego ne provedëš he:acc neg trick:prs.2sg ‘He can’t be fooled.’

Both constructions, unlike impersonal sentences in a narrow sense, generally include an action verb. They also have a very real, although indeterminate, agent. They contain different forms of verbal predicates: while the impersonal sentence strictly speaking has a predicate in the 3rd person singular, which can be considered a mark of non-agreement, in sentences with zero subject, the verb is in the 3rd person plural (for the first type) or in the 2nd person singular (for the second type). Constructions of the first type occur in most Slavic languages, notably in Bulgarian, Polish and Czech. As for the second type of constructions, with the exception of Russian, most other languages express the generic human subject with a word etymologically related to the lexeme ‘man’ (such as Man in German or on in French) (pl. człowiek, bg. čovek, cz. človĕk, s.-cr. čovjek): pl. Człowiek nie wie co ma robić ‘One does not know what to do’; bg. Čovek ne znae nikoga kakvo može da se sluči ‘You never know what might happen’. However, these Slavic ­constructions are personal.

 Marguerite Guiraud-Weber & Irina Kor Chahine

A third variant of zero subject, which is not always mentioned in descriptions of Russian or is confused with strictly impersonal sentences, may be added to the two types presented above, cf. (22) Russian a. Pod krovat’ju zaševelilos’. under bed:inst moved:neut ‘Something moved under the bed.’ b. Nad golovoj gudelo. over head:inst buzzed:neut ‘Something buzzed over his head.’

Its specificity has been noted by V.L.Georgieva (Georgieva 1969, 1978). Indeed, the zero subject in these sentences can either refer to a person or to an animal, as well as to an indefinite object. In other words, it does not necessarily refer to a person, and its predicate is in the 3rd person singular. One can observe a certain correspondence of such sentences with the constructions existing in other Slavic languages that also use a verb in the 3rd person singular without an explicit subject, but in this case the zero subject refers to a person (not to an inanimate object), and very often to the speaker (23–24). Such constructions always use a reflexive verb (cf. Krzek, this volume; see also reflexive constructions below, § 5). (23) Polish a. Robi się. do:prs.3sg ref ‘[I/we/they] do it.’ b. Czyta się. read:prs.3sg ref ‘[I/we/they] read.’ (24) Serbo-Croatian a. Ide se. walk:prs.3sg ref ‘[I/we/they] walk.’ b. Govorilo se. talked:neut ref ‘[I/we/they] talk.’

It is precisely these forms of indefinite subject that have not been lexicalized in Slavic languages (Russian, Polish, and Czech) that D. Weiss (this volume) ­discusses in a broader context of what he calls “syntactic emptiness”, which includes the referential zero lexemes Øpeople and Øelement, the generic ‘you’ Øyou, as well as the zero copula and omission of verbs of action. Based on diachronic data and placed in a typological perspective, his conclusions on the phenomena



Impersonals and beyond in Slavic 

of “syntactic ­emptiness” observed in colloquial Russian are in accordance with the Eurasian orientation of the latter.

5.  Impersonal and related constructions The absence of agent or its occasionally minor status makes impersonal sentences closer to other models that share some of their features. This often makes identifying these constructions difficult. J. Milićević (this volume) discusses this very issue. Through the theoretical background of the model “Meaning ⇔ Text”, she describes morphological mechanisms, such as diathesis and verbal derivation that operate within different constructions without a subject in Serbian as compared to other Slavic and European languages. She notes that impersonal constructions often appear as a result of these morphosyntactic transformations. While similarities between impersonal and passive constructions can be observed on the synchronic level, they can be seen on the diachronic level as well. In the history of Slavic languages, the development of impersonal sentences is often related to passive constructions. Thus, in Polish, the passive suffix -no/-to became an impersonal marker (Pisarkowa 1984), which can be found in sentences like (25), with a direct object making the passive reading impossible. Today, they function as active constructions (cf. Krzek this volume). (25) Podano herbatę. served:part.neut tea:acc ‘Tea was served.’

Similarly, in Russian, the relatively late emergence of impersonal constructions with instrumental, such as (17a) may have originated from the expansion of the instrumental agent in passive constructions. In Old Russian, the agent was expressed by the form “ot + genitive,” which subsequently became outdated and fell out of use (Ivić 1965). It is interesting to note that in similar modern SerboCroatian structures the “real” passive construction is used: an inanimate agent is encoded by the instrumental (26a), while an animate agent is encoded by the phrase “od + genitive” (26b). (26) Serbo-Croatian a. Put je zatrpan snegom. road:nom.m be:3sg covered:part.m snow:inst ‘The road is covered with snow.’ b. On je odlikovan od predsednika. he:nom be:3sg awarded:part.m by president:gen ‘He was awarded by the president.’

 Marguerite Guiraud-Weber & Irina Kor Chahine

Furthermore, there are similarities between passive constructions and sentences with “o + locative”, such as (27), which replaced the former impersonal passive (see §3 above) (Švedova 1964). (27) Ob ètom govorilos’ v gazetax. about this:loc discussed:neut.ref in newspapers:loc ‘This was discussed in the newspapers.’

The classification of these constructions among impersonal sentences is not always unanimous. In this construction, the postfix -sja attached to a verb can be seen as a morphological marker of impersonal subject originating from a verbal derivation, which is often confused with the process of passivization (see various types of passive constructions in Plungjan 2003). In addition to these examples of direct links between the impersonals and passive, it should be noted that the domain of impersonal constructions has been gradually expanding at the expense of the passive. Thus, in Russian, in sentences with a predicate in the 3rd person plural, the role of the agent is minimal (zero subject), which makes them closer to passive constructions where the agent plays a peripheral role. Some scholars think that this construction compensates for an absent position in the passive paradigm, which does not possess all aspect and tense forms (Wiemer 1996). Besides, it is also possible to view impersonal ­sentences as a type of construction where the voice opposition is neutralized. On the other hand, the relationship between impersonal and reflexive ­constructions is based on operations that affect the predicate-argument structure. From a certain point of view, this operation, which detransitivizes a transitive construction, can be linked to the phenomenon observed in ergative languages where this type of construction contains a “dedicative anti-passive marker.” It is in this context that K. Janic (this volume) proposes analyzing Slavic reflexive morphemes, namely ru. -sja, pl. się, and s-cr. se. Despite the fact that these reflexive morphemes appear in constructions with varied semantics (impersonal, passive, reflexive, reciprocal), M. Marelj and E. Reuland present a formal, unified analysis of these morphemes, which allows them to derive their pure-reflexive and near-reflexive readings and to point out the differences between South/West Slavic ­languages, which have developed a system of clitics, and East Slavic languages where the reflexive morpheme is realized only as the verbal affix -sja. It is also using the formal properties of reflexive items in various languages (ru. sebja, drug druga, pl. siebie, cz. sebe, bg. sebe si, en. himself, each other…) that S. Franks (this ­volume) shows what is relevant for “binding” in the morphological structure of these anaphors. Finally, A. Zimmerling compares syntactic properties of Slavic possessive clitics (i.e. ru. sebe, bg. mi) with syntactic properties of possessive constructions (e.g. ru. u X-a est’ Y) in terms of syntactic operation described as ­Possessor ­Raising, and



Impersonals and beyond in Slavic 

concludes that this concept is not really applicable to Slavic languages for which the term Possessive Shift seems to be more appropriate. 6.  The role of lexis Studies of Slavic impersonal constructions allow us to measure to which point syntactic structures are dependent on lexis, since most constraints and limitations in syntax are realized through the lexicon. For this reason, studies on lexis still remain an important part of research, and our discussion has left a special place for them. 6.1  Interaction between lexis and grammar Besides lexical meaning in impersonal constructions (see especially the article by S.Say) and lexical issues marginally touched upon in most publications, the ­interaction between lexis and grammar is a focal point of interest for certain authors. The first is a comparative study on evidentiality in Macedonian and ­Albanian. Mainly related to the lexical domain in Slavic languages (ru. jakoby, mol, pl. ponoć, jakoby, cz. prý/prej, mac. navodno…), the category of evidentiality in ­Macedonian possesses additional grammatical resources, in particular, it uses forms of the ­verbal perfect which presents a problem of interpreting homonymous forms (evidential versus perfect). Thus, M. Makartsev (this volume) attempts to analyze mechanisms for decoding the meaning of utterances using materials of simultaneous translation of political discourse in both languages. Strategies used in political discourse also take into account lexical choice, which will be discussed below (cf. article by D. Dobrovol’skij & L. Pöppel). The interaction between lexis and grammar is also observed in deverbal nouns. The question posed by G. de Valdivia, J. Castellví and M. Taulé (this volume) is whether the aspect of the verb (imperfective/perfective) at the origin of Russian deverbal nouns (such as formirovanie ‘formation’, zakaz ‘order’) has any impact on the lexical aspect of the latter (i.e. event, result, or state). Using an experimental framework and statistical data, the authors reexamine this question and show that the verbal aspect cannot be viewed as a determinant in lexical denotation of ­deverbal nouns, although it makes an important contribution to it. Furthermore, in this section we can also consider cases of lexical desemantization in favor of a particular grammatical function. Thus, it has been shown that in many languages verbs of motion and position regularly undergo this kind of semantic shift (Majsak 2005). Confirming this fact in Polish data, D. Sikora offers a detailed analysis of the Polish verb iść ‘to go’, which is currently used as a true auxiliary.

 Marguerite Guiraud-Weber & Irina Kor Chahine

To a certain extent, the interaction between lexis and grammar also concerns the usage of conjunctions: these items have their own semantics, but their usage is impossible without the grammatical constructions in which they appear. A. Letuchiy (this volume) discusses this issue in his study on tripartite constructions with the lexemes čtoby and kogda, and points out a particular verbal agreement (past tense in the temporal clause), which is not mentioned in Russian grammars. Analyzing this feature contributes to the general theory of subordination and complex clauses. 6.2  Analysis of lexical meaning The domain of lexical semantics is very broad. Papers included in this volume focus mainly on two fundamental semantic relations: synonymy and polysemy. The analysis of lexical synonyms requires above all exposing their s­ emantic properties and can be supplemented by diachronic data. In this manner, D. Dobrovol’skij and L. Pöppel study two pairs of synonyms that originated from the same semantic field, power: revolucija ‘revolution’ vs. perevorot ‘coup’, and mjatež ‘revolt’ vs. vosstanie ‘uprising’. In order to differentiate between these terms, lexicographic information alone is not sufficient, and their distinctive features have been elucidated by comparing current contexts to the contexts of 1917. Like M. Makartsev’s paper (see above), the study of this lexical class represents a fundamental basis for discourse analysis, in particular for political discourse, because it reveals linguistic strategies often deployed for manipulation. Issues of polysemy discussed in this volume concern nouns, adjectives, and adverbs. V. Beliakov’s study of lexical items in semi-phrasemes (this v­ olume) reexamines combinations of two nouns, so-called “collocations”, which are used to express a large quantity in Russian, such as kuča voprosov ‘a heap of questions’. He distinguishes two classes of quantifying nouns that have different relationships with the metaphorical transfer. It is also to the description of metaphoric as well as metonymic patterns that T. Reznikova et al. (this ­volume) devote their paper, but this time, it is applied to the lexical field of quality in Russian. This is an experimental study introducing a new methodology that involves compiling a database of qualitative lexemes (adjectives, adverbs, predicatives). This database allows the authors to bring to light recurrent correlations between different parameters of qualitative lexemes as well as to draw more general conclusions on the lexical system functioning as a whole, such as “systemic organization” of vocabulary. This electronic treatment of lexicological data reveals that current theories of semantic changes need further e­ laboration, which encourages discussion on the universality of these phenomena from a cross-linguistic perspective. Papers included in this volume explore a number of linguistic issues in Slavic languages, ranging from morphology to syntax and semantics. The volume assigns



Impersonals and beyond in Slavic 

equal importance to each of them since the authors' differing approaches contribute to our understanding of the issues from a broad variety of angles.

Acknowledgements We are grateful to the following scholars for their perceptive comments on papers in this volume at early stages: Christine Bonnot (Institute for Oriental languages, INALCO, Paris, France), Greville G. Corbett (Surrey University, United Kingdom), Paul Garde (Aix-Marseille University, France), Marguerite Guiraud-Weber (AixMarseille University, France), Steven Franks (Indiana U ­ niversity-Bloomington, USA), Frank Y. Gladney (University of Illinois at Urbana-Champaign, USA), Grigory Kreydlin (Russian University for Humanities, RGGU, Moscow, Russia), Vladimir Plungian (Institute for Linguistics, Moscow, Russia), Robert Roudet (University of Lyon III, France), Marek Świdziński (University of Warsaw, Poland), Maria Szupryczyńska (University of Toruń, Poland), Nicolas Tournadre (Aix-Marseille University, France), Hélène Włodarczyk (University Paris IV-Sorbonne, France), Daniel Weiss (Zurich U ­ niversity, Switzerland), Charles Zaremba (Aix-Marseille University, France). Our thanks must also be extended to Werner Abraham and Elly van Gelderen, editors from John Benjamins Publishing Compagny for giving us the possibility to publish our volume in Studies in Language Companion Series. In addition, we would like to express our sincere gratitude to Cornelis (Kees) H.J. Vaes for his suggestions and patience in preparation of this publication. And finally, personal thanks to Irina Mikaelian, and especially to Natalia Partenheimer for their help with translating and editing various parts of the volume.

References Babby, Leonard. 1980. Existential Sentences and Negation in Russian. Ann Arbor MI: Karoma. Birjulin, Leonid A. 1994. Semantika i sintaksis russkogo impersonala: Verba meteorogica i ix ­diatezy. München: Verlag Otto Sagner. Boguslavskij, Igor' M. 1991. Lingvističeskij processor i lokativnye obstojatel’stva. Voprosy jazykoznanija 1: 69–78. Comrie, Bernard. 1981. Language Universals and Linguistic Typology. Oxford: OUP. Creissels, Denis. 1991. Approche des constructions asubjectales, improprement désignées comme impersonnelles. In L’impersonnel: Mécanismes linguistiques et fonctionnement ­littéraire, Michel Maillard (ed.), 47–57. Grenoble: Ceditel. Fortuin, Egbert L.J. 2000. Polysemy or Monosemy: Interpretation of the Imperative and the Dative-infinitive Construction in Russian. Amsterdam: Institute for Logic, Language and Computation.

 Marguerite Guiraud-Weber & Irina Kor Chahine Franks, Steven. 1995. Parameters of Slavic Morphosyntax. Oxford: OUP. Galkina-Fedoruk, Evdokija M. 1958. Bezličnye predloženija v sovremennom russkom jazyke. Moskva: MGU. Georgieva, Valentina. 1969. K voprosu o granicax bezličnyx predloženij russkogo jazyka. Učenye zapiski Moskovskogo Gosudarstvennogo Pedagogičeskogo Instituta 341: 63–68. Georgieva, Valentina. 1980. Bezličnye predloženija. In Istoričeskaja grammatika russkogo jazyka, Viktor Ivanovič Borkovskij (ed.). Moskva: Nauka. Guiraud-Weber, Marguerite. 1979. O sintaksičeskoj prirode konstrukcij tipa: ‘V komnate paxnet jablokami’. Russian Linguistics 4: 291–301. Guiraud-Weber, Marguerite. 1984. Les propositions sans nominatif en russe moderne. Paris: Institut d’Études Slaves. Guiraud-Weber, Marguerite. 2003. Le sujet en russe. In Le sujet [Bibliothèque des Faits de Langues], 63–72. Paris: Ophrys. Guiraud-Weber, Marguerite. 2003. Ešče raz o russkom genitive otricanija: Vzgljad so storony, Russian Linguistics 27: 363–384. Guiraud-Weber, Marguerite & Zaremba, Charles. 2007. Incidence de la négation sur les deux premiers actants de la phrase en polonais, russe et tchèque. In La négation [Travaux du Cercle linguistique d’Aix-en-Provence] Christian Touratier & Charles Zaremba (eds), 49–67. Aix-en-Provence: Presses Universitaires de Provence. Hausenblas, Karel. 1958. Vývoj předmĕtového genitivu v češtině. Praha: Nakladatelství ČSAV. Ivić, Milka. 1965. On the origin of the Russian sentence type: (Ego) zavalilo snegom. Die Welt der Slaven X(3–4): 317–321. Ickovič, Viktor A. 1974. Očerki sintaksičeskoj normy. In Sintaksis i norma, Galina ­Aleksandrovna Zolotova (ed.), 43–106. Moskva: Nauka. Keenan, Edward. 1976. Towards a universal definition of subject. In Subject and Topic, Charles N. Li (ed.), 305–333. New York NY: Academic Press. Maillard, Michel. 1985. L’impersonnel français de “il” à “ça”. In Autour de l’impersonnel, Jacques Chocheyras et al. (eds), 51–118. Grenoble: Ellug. Majsak, Timur A. 2005. Tipologija grammatikalizacii konstrukcij s glagolami dviženija i glagolami pozicii. Moskva: Jazyki slavjanskix kul’tur. Malchukov, Andrej & Siewierska, Anna (eds). 2011. Impersonal Constructions: A Cross-linguistic Perspective [Studies in Language Companion Series 124]. Amsterdam: John Benjamins. Maurice, Florence. 1996. Der modale Infinitiv in der modernen russichen Standardsprache. München: Verlag Otto Sagner. Mrázek, Roman. 1964. Sintaksis russkogo tvoritel’nogo. Praha: Státní pedagogické nakladatelství. Mrázek, Roman. 1990. Sravnitel’nyj sintaksis slavjanskix literaturnyx jazykov. Brno: Univerzita J.E. Purkynĕ. Padučeva, Elena V. 1992. O semantičeskom podxode k sintaksisu i genitivnom sub”ekte glagola byt’. Russian Linguistics 16(1): 53–63. Partee, Barbara H. & Borschev, Vladimir B. 2007. Existential sentences, BE, and the genitive of negation in Russian. In Existence: Semantics and Syntax, Ileana Comorovski & Klaus von Heusinger (eds), 147–190. Dordrecht: Springer. Pisarkowa, Krystyna. 1984. Historia składni języka polskiego. Ossolineum: PAN. Plungjan, Vladimir A. 2003. Obščaja morfologija. Vvedenie v problematiku. Moskva: Editorial URSS. Švedova, Natalia Ju. (Ed.). 1964. Izmenenija v sisteme prostogo i osložnennogo predloženija v russkom literaturnom jazyke XIX veka. Očerki po istoričeskoj grammatike russkogo literaturnogo jazyka XIX veka. Moskva: Nauka.



Impersonals and beyond in Slavic 

Testelec, Jakov G. 2001. Vvedenie v obščij sintaksis. Moskva: RGGU. Veyrenc, Jacques. 1979. Les propositions infinitives en russe. Paris: Institut d’Études Slaves. Vojvodić, Dojcil. 2007. Ob infinitivnyx predloženijax s sub”ektom v datel’nom padeže v russkom jazyke i ix èkvivalenty v serbskom i pol’skom. Zbornik Matice srpske za slavistiku 71–72: 557–581. Weiss, Daniel. 1987. Polsko-niemieckie paralele w zakresie czasowników modalnych. Sprachund Kulturkontakte im Polnischen, Gerd Hentschel & Alek Pohl (eds), 131–156. München: Otto Sagner. Weiss, Daniel. 1988. Konstrukcje typu: mieć (coś) do załatwienia, być do załatwienia oraz dać (coś) do załatwienia: Stopieǹ ich sfrazeologizowania dawniej i dziś. In Z problemȯw frazeologii polskiej i słowianskiej, Mieczysław Basaj & D. Ryc (eds). Wrocław. Weiss, Daniel. 1993. Infinitif et datif en polonais contemporain: Un couple malheureux? In Actes du VI colloque international de linguistique romane et slave, Stanisław Karolak (ed.), 443–487. Cracovie: WSP. Wiemer, Björn. 1996. Analityczne passivum w języku rosyjskim i polskim. In Semantyka a konfrontacja językowa, Violetta Koseska-Toszewa & Danuta Rytel-Kuc (eds). Warszawa: SOW. Wolińska, Olga. 1978. Konstrukcje bezmianownikowe we wspȯłczesnej polszczyźnie. Katowice: Uniwersytet Sląski. Zolotova, Galina A. 1966. K razvitiju složno-padežnyx konstrukcij. In Razvitie sintaksisa sovremennogo russkogo jazyka, 147–173. Moskva. Zolotova, Galina A. 1973. Očerk funkcional’nogo sintaksisa russkogo jazyka. Moskva: Nauka. Žaža, Stanislav. 1978. K problematice negativních konstrukcí se slovesem esse v ruštině a v češtině. In Ceskoslovenské přednášky pro VIII mezinarodni sjezd slavistů v Záhřebu. Prague: Academia Praha. Žaža, Stanislav. 1998. K jednomu typu českych a ruskych vět s funkčnim predikativem. Sbornik Praci Filozofické Fakulty Brněnské Univerzity, 89–92. Žaža, Stanislav. 2002. Ešče raz k xarakteristike otricatel’nyx konstrukcij v russkom jazyke. Sborník Prací Filozofické Fakulty Brněnské Univerzity, 51–56.

section i

Morphosyntax

Binding and morphology revisited Steven Franks

Indiana University – Bloomington, USA One of the earliest and most compelling observations to emerge in response to Chomsky’s binding theory was the suspicious cross-linguistic correlation between the morphological structure of anaphors and their interpretive potential. Specifically, morphologically simplex anaphors (e.g. Russian sebja) are subject-oriented and can be bound by more distant antecedents than complex anaphors (e.g. English himself), which generally require local antecedents, the grammatical function of which is not restricted to Subject. This dependence, however, was inconsistent with the GB architecture of the grammar, in which PF and LF were mediated by S-structure and, under Minimalism, is incoherent: meaning/LF cannot in principle be read off of morphology/PF. Instead, it must be the anaphor’s LF “morphological” structure that somehow determines its binding properties. This paper therefore explores Slavic data for which PF and LF morphology diverge and argues that it is invariably the anaphor’s presumed LF morphology that is relevant for binding. Keywords:  anaphora; binding; Logical Form (LF); morphology; parameter

1.  The problem and its original solution The problem this paper addresses is empirical, but the argumentation will be largely conceptual. The empirical problem is cross-linguistic variation in the behavior of anaphora such as Russian sebja or English himself. Although there is considerable debate over the exact properties of different reflexive items in various languages, it is clear that fundamental differences exist. Consider the Russian examples in (1) and (2), based on Rappaport (1986): (1) Milicioner1 rassprašival arestovannogo2 o sebe1/*2. ‘The policeman1 asked the suspect2 about himself1/2.’ (2) a. Professor1 razrešaet assistentu2 [PRO2 proizvodit’ opyty nad soboj1/2]. ‘The professor1 allows his assistant2 [PRO2 to conduct experiments on himself*1/2].’ b. Ivan1 čital [Sašinu2 stat’ju o sebe1/2]. ‘Ivan1 read [Sasha2’s article about himself?*1/2].

 Steven Franks

Although (1) is unambiguous, the constructions in (2) are technically ambiguous.1 That is, sebe in (1) must refer to the Subject milicioner, whereas soboj in (2a) can refer either to the main clause Subject professor or to the understood PRO Subject of the infinitive (itself controlled by assistentu), and sebe in (2b) can refer either to the Subject Ivan or to the “Subject” of the noun phrase, namely Saša. English judgments are precisely the reverse: in (1), himself can be either the policeman or the suspect and, in (2), himself can only refer to the most local antecedent, which is PRO in (2a) and Saša in (2b). This is a typical example of the nature of the empirical problem, although the properties that differentiate Russian sebja from English himself  hardly exhaust the possibilities found in languages of the world. The conceptual issue hinges on the fact that these properties are not accidental, but rather exhibit consistent patterns. Consideration of diverse anaphors among the world’s languages leads for example to the conclusion that no language could have a lexical item with the morphological properties of Russian sebja but which behaved like English himself; similarly, no language could have a lexical item with the morphological properties of English himself but which behaved like Russian sebja. Indeed, one of the earliest and most compelling observations to emerge in response to the binding theory of Chomsky (1980, 1981) was the suspicious correlation between the morphological structure of anaphors and their interpretive potential. Morphologically simplex anaphors look like heads (and can be inflected for case where the language calls for it). Russian sebja is a simplex anaphor. Complex anaphors, such as English himself on the other hand, have the internal structure of a phrase. The specific correlation, as initially noted by Yang (1983), is that simplex anaphors are subject-oriented and can be bound by more

.  Much ink has been spilled debating the interpretation of Russian sebja and svoj; Chapter 9 “Vozvratnye mestoimenija” of Padučeva 1985 in particular offers a lengthy treatment of the interpretative possibilities. The purpose of the present short paper is not to offer a comprehensive theory of anaphora, but rather simply to make a point about their analysis within a particular model of grammar. It is beyond the scope of this paper either to deal with models other than GB/Minimalism or to examine data sets more complete than those required to make the basic point. Consideration of “real” corpora, as suggested by an anomymous reviewer, would only serve to muddy the waters, since such data are not only more complex than needed for the purposes of making the point at hand, but would also introduce irrelevant factors due to their unreliability, inconsistency, and heavy pragmatic conditioning. I also put aside reflexive possessive pronouns of the svoj type (although these behave the same in Slavic as sebja reflexives, and are not treated simply for reasons of space), as well as Slavic postfixes/clitics of the -­ sja/się type (which are not treated because they are not subject to the binding theory but rather directly manipulate predicate-argument structure, hence behave very differently from their true pronominal counterparts).



Binding and morphology revisited 

distant antecedents than complex anaphors can. Also, complex anaphors generally require local antecedents, and the grammatical function of the antecedent of complex anaphors is typically not restricted to Subject. Thus, Russian sebja and Chinese ziji ‘self ’ are simplex, subject-oriented, and allow long-distance antecedents, whereas English himself and Chinese ta-ziji ‘s/he-self ’ are complex, not subject-oriented, and do not allow long-distance antecedents. This is stated in (3):2 (3) a. Simplex (=head) Russian sebja, Chinese ziji ‘self ’ fi subject-oriented and allow long-distance antecedents b. Complex (=phrase) myself, himself, Chinese wo-ziji ‘I-self ’, ta-ziji ‘s/he-self ’ fi not subject-oriented and do not allow long-distance antecedents

In short, the claim was that, for any given anaphor, its set of possible antecedents depended on its internal morphological structure. This was a welcome discovery, as the correlation provided a realistic avenue for addressing the projection problem: How do children fix the interpretive properties of anaphors? The “principles and parameters” approach once popular in Government and Binding (GB) syntax is not viable here for several reasons, the most compelling being that different anaphors – even within a single language – can work differently (as one might expect, in a manner consistent with their morphological structure). Noting this, Wexler and Manzini (1987: 55) among others were led to what may be called the “Lexical Parameterization Hypothesis” in (4): (4) Lexical Parameterization Hypothesis Values of a parameter are associated not with particular languages, but with particular lexical items in a language.

This is, however, only part of the story. A major issue neglected by Wexler and ­Manzini’s statement has to do with learnability. While lexical properties must indeed be fixed in acquisition, leaving the matter as in (4) disregards the role of bootstrapping. That is, to the extent that the so-called “parametric settings” for particular lexical items are derivative of other, independently discernable ­properties – specifically, form or meaning – these “settings” can be bootstrapped from the item itself. But this of course means that there are no lexical “parameters” in the ­traditional sense, since the “settings” are in fact parasitic on aspects of form or meaning that must be learned regardless. There are other inadequacies in (4) related to matters of learnability. With respect to binding domain, as implied in Wexler and Manzini’s application of Berwick’s “Subset Principle,” (4)

.  See e.g. Cole et al. 1994 for discussion.

 Steven Franks

would require children first to adopt the grammar generating the smallest binding domain and then gradually to expand that domain as required by positive evidence. The gist of this principle is stated in (5): (5) Subset Principle Given a choice between competing grammars generating languages of which one is a proper subset of the other, the child selects the grammar that generates the smaller language.

Empirical evidence however suggests this not to be true, at least with respect to binding.3 Moreover, even if correct, to apply the Subset Principle in the acquisition of longdistance anaphors would require children to analyze complex sentences, contrary to the Degree-0 (“plus a little”) learnability hypothesis put forward in Lightfoot (1989). With respect to orientation, the Subset Principle would require children to posit the smallest set of possible antecedents (namely Subjects only) and to expand that as needed. Finally, and most relevant for present purposes, such an approach would not capture the correlation between allowing greater syntactic distance between an anaphor and its antecedent and limiting the set of possible antecedents to Subjects. Instead, then, the correct approach to variation in the binding behavior of anaphors must somehow capitalize on the morphology of the actual anaphor. Indeed, in the GB period various ingenious technical solutions were entertained in order to implement the correlation between an anaphoric item’s interpretive properties of long-distance binding and subject-orientation, on the one hand, and its presumed morphological properties, on the other. While it is not my purpose to review these technical solutions here, credible implementations involved phrasal versus head (anaphor) movement (e.g. Pica 1987; Battistella 1989; or Cole & Sung 1994) and phrasal versus head (antecedent) domain determination (e.g. Progovac 1992). Under all such approaches, however, the syntax included some mechanism that allowed an anaphor’s morphology to exert direct impact on its choice of an antecedent. In short, Logical Form (LF) properties were made to depend on Phonetic Form (PF) properties. 2.  A conceptual challenge This dependence, however, was inconsistent with the GB architecture of the grammar, in which PF and LF were mediated by S-structure. Within such a model of grammar, meaning (the so-called “LF–side”) cannot in principle be read off of morphological structure (the “PF–side”). This renders problematic the often-stated .  See (Bailyn 1992) for Russian speaking children or (Franks & Connell 1996) for English speaking speaking children.



Binding and morphology revisited 

fact that the two appear to correlate for anaphora. Why this problem with the morphologically-driven approach to anaphor variation did not bother practictioners of GB at the time remains a mystery.4 In any event, the problem of letting the morphology of a particular lexical item delimit its possible antecedents is even more acute under Minimalism.5 The reason a PF–side morphological approach to a matter of interpretation is incoherent under Minimalism is because antecedency is determined in LF, not in the syntax per se. Hence, any such account would not have access to the anaphor’s overt morphological structure, visible only on the PF–side of the grammar. Instead, it must be the anaphor’s covert morphological structure, available on the LF–side, that somehow determines its binding properties. The postulation of LF “morphological” structure may seem like an oxymoron: How can LF have its own morphology? But what is meant here is simply the existence of structure, which a little reflection shows must be the case, since both LF–side and PF– side representations are read off of whatever structures are provided by the syntax. Morphemes are pairings of morphophonological entities (“vocabulary items,” in Distributed Morphology terms) with sets of morphosyntactic features, including both closed-class formal syntactic features and open-class semantic features (“roots,” in Distributed Morphology terms). Morphemes concatenate with each other in highly structured ways. Thus, Russian sebja consists of [+anaphoric, +reflexive] and [+accusative], while English himself consists of [+pronominal, +masculine, –plural] and [+anaphoric, +reflexive].6 This could be represented crudely as in (6), where SEBJA and HIMSELF abbreviate the relevant feature sets:7 (6) a.   SEBJA b.   HIM + SELF [+anaph, +refl, +acc]

[+pron, +m, −pl]    [+anaph, +refl]

The point is that the syntax provides a highly articulated feature structure and that this structure needs to map into both PF and LF representations. These pairings, .  One might take this as an inadequacy of the model itself. Indeed, the general observation that PF and LF seem to operate in tandem has led to several promising alternatives to GB/Minimalism, the most compelling of which, in my view, is Jackendoff ’s (2002) “Parallel Architecture” model. .  This is true of standard “Y-model” Lexicalist Minimalism. In systems augmented by Distributed Morphology, “the two arms of the Y-model join again at the conceptual interface” (Siddiqi 2009: 14). .  No special significance should be imputed to the particular features, which are meant to be purely illustrative. .  Since it is not obvious whether Russian sebja should be analyzed as bimorphemic or not, i.e. whether [+anaph] and [+acc] are two feature sets or just one, I represent them together. I also put aside the issue of case in English.

 Steven Franks

however, are not always one-to-one. Consider some different kinds of examples: Russian (7a) kuplju versus English (7b) will buy or Macedonian (8a) knigana versus English (8b) that book: (7) a. Russian kuplju b. English will buy (8) a. Macedonian knigana b. English that book

In each case, the Slavic language has an inflected form of a single word whereas English has two independent words, each occupying its own head position in the syntax. Yet the feature content is more or less the same. Thus, on the PF–side, the morphology is different in Slavic and English – in Slavic the features are realized on a single head, whereas in English they require two heads (hence, phrasal structure) in order to be expressed. This is indicated in (9) and (10): (9) a. [V+T kuplju] b. [T will] [V buy] (10) a. [N+D knigana] b. [D that] [N book]

The syntax provides both V and T(ense) features in (9), and both N and D(eterminer) features in (10). These are manipulated on the PF–side, but on the LF–side the morphology is the same. That is, in their syntactic structures, both Russian and English have [TP [VP … ]] and both Macedonian and English have [DP [NP … ]]. In other languages, however, the syntax may be different. Thus, following a series of publications of which Bošković (2008) is representative, nominals in Russian (and many other languages) do not project a DP. Similarly, there is probably no reason to posit a TP for a language such as Chinese. Let us return in this light to anaphors and the problem of how to exploit the observed correlation between structure and interpretation. If we take the interpretive properties of anaphors to derive from their morphological structure, then it is going to have to be the structure at LF rather than PF that matters. That is, the difference in interpretative properties between himself and sebja must derive from the fact that the latter has so-called φ-features (i.e. person, number, gender), whereas the former does not. These features, then, are going to require phrasal structure for English himself, but not for Russian sebja. What this means for our purposes is that one should look for anaphors where their PF and LF morphological structures arguably diverge and explore what happens.8 The prediction is clear: the anaphor’s

.  One might wonder how case features fit into PF and/or LF structures and, in particular, whether they induce phrasal status. The Slavic facts clearly show that case is completely



Binding and morphology revisited 

presumed LF morphology should be relevant for binding, not its surface morphological structure. Slavic languages, it turns out, provide such data, and it is to examples of this sort that we now turn. 3.  Some Slavic data In Slavic, as elsewhere, PF morphology for the most part mirrors LF morphology. This is hardly surprising, as the transparency of one-to-one mappings makes them more easily parseable. It is for this reason that reciprocals, for example, are generally bipartite. Thus, not only does English each other consist of two distinct elements, but so also does Russian drug druga. The reason is because, in semantic terms (cf. Heim, Lasnik & May 1991), reciprocals are by their very nature bipartite: they necessarily contain a distributor, such as English each, and a reciprocator, such as English other. In accordance with their complex semantic structure, at LF reciprocals are universally phrasal, regardless of how they appear at PF. And since they are always phrasal – in the sense of (3b) – we expect them never to be subject-oriented and never to allow long-distance antecedents. We can then ask what would happen with a reciprocal which seems to be a head morphologically (i.e. on the PF–side). As stated, we expect its superficial structure to be irrelevant. Reflexives, on the other hand, vary across languages (and even within them: recall Chinese ziji ‘self ’ versus wo-ziji ‘I-self ’, ta-ziji ‘(s)he-self ’, etc.) in terms of whether or not they also embody φ-features. If they do, then they are expected to behave as if phrasal, and if they do not, then they are expected to behave as heads. This contrast was apparent for English himself versus Russian sebja in (1)–(2) above. We can then ask what would happen with a reflexive which was morphologically phrasal but completely lacked φ-features, rendering it a head for LF purposes. Since it would be a head for interpretive purposes, we would expect it to be subject-oriented and to allow long-distance antecedents. In this section we will see that Slavic languages demonstrate both these predictions to be correct. 3.1  Back to Russian In his application of the classic binding theory to Russian, Rappaport (1986) noted that Russian drug druga is unlike sebja in that it resembles English each other in irrelevant; e.g. Russian locative sebe in (1) or instrumental soboj in (2a) count as simplex (heads), not complex (phrases). One possibility, in keeping with minimalist argumentation, is that case features are stripped away in the mapping to LF, since they have no semantic interpretation (contra, e.g. Jakobson 1936/1971). On the other hand, case features must be retained in the mapping to PF, since they do have a morphological interpretation. If so, we have another argument that it is LF rather than PF morphology that matters for binding.

 Steven Franks

terms of its binding properties.9 Compare his examples in (12) and (14) with those in (1) and (2), repeated as (11) and (13): (11) Milicioner1 rassprašival arestovannogo2 o sebe1/*2. ‘The policeman1 asked the suspect2 about himself1/2.’ (12) Milicionery1 rassprašivali ix2 drug o druge1/2. ‘The policemen1 questioned them2 about each other1/2.’ (13) a. Professor1 razrešaet assistentu2 [PRO2 proizvodit’ opyty nad soboj1/2]. ‘The professor1 allows his assistant2 [PRO2 to conduct experiments on himself*1/2].’ b. Ivan1 čital [Sašinu2 stat’ju o sebe1/2]. ‘Ivan1 read [Sasha2’s article about himself?*1/2].

(14) a. My1 poprosili ix2 [PRO2 nalit’ drug drugu*1/2 čajku]. ‘We asked them2 [PRO2 to pour each other*1/2 some tea].’

b. Pisateli1 čitali [vospominanija xudožnikov2 drug o druge?*1/2]. ‘The writers1 read [the artists’2 reminiscences about each other?*1/2].’

Rappaport had no explanation for the facts about reciprocal drug druga versus reflexive sebja, simply stating them, but we now know that these facts follow from intrinsic properties of the particular lexical items. The reciprocal – reflecting both its bipartite drug druga morphology and its bipartite semantics – necessarily has two pieces, as schematized in (15).10 (15)     EACH + OTHER [+distributor]

[+anaph, +recipr]

In this regard, all reciprocals, including Russian ones, are expected to behave in a similar fashion to English himself. They both have phrasal structure, and it is this structure that causes them to be strictly locally-bound and also that allows them to be bound by non-Subjects. Although this phrasal structure happens to correlate with PF morphology in both Russian and English, which is the canonical direct .  As an anonymous reviewer reminds me, there has of course been much discussion of binding in Russian since this early publication. Rappaport’s work was however seminal in presenting the core facts and in making the core observations. Moreover, given the conceptual goals of the present paper, consideration of additional subtleties would be otiose and ­distracting to the purpose at hand. .  Recall that the capitalized EACH and OTHER refer to the meanings of the items, not to the words themselves.



Binding and morphology revisited 

correspondence between form and meaning, my claim is that the binding properties of reciprocals actually depend on the fact that they have complex morphological structure in LF. This is universal, given what reciprocals mean.11 3.2  Polish West Slavic provides a particularly telling demonstration of the fact that what matters is semantic rather than morphological structure.12 In Polish, the reflexive behaves just as in Russian in terms of its binding properties. Consider, for ­example, the sentences in (16), from Reinders-Machowska (1991), which – just like Russian (11) and (13a) – differ from English with respect to the possibilities of a non-subject or long-distance antecedent: (16) a. Jan1 opwiada Piotrowi2 o sobie1/*2. ‘Jan1 talks to Peter2 about himself1/2. ‘

b. Maria1 kazała Piotrowi2 [PRO2 zbudować dom dla siebie1/2]. ‘Maria1 told Peter2 [PRO2 to build a house for himself2/*herself1].

However, in Polish the reflexive morpheme can also function as a reciprocal, and when it does, the judgments change accordingly.13 Reinders-Machowska (1991: 147) cites the Polish examples in (17) as a relevant minimal pair: (17) a. Chłopcy1 czytali [dziewcząt2 wspomnienia o sobie1/2]. ‘The boys1 read [the girls’2 memories about themselves*1/2].’ b. Chłopcy1 czytali [dziewcząt2 wspomnienia o sobie*1/2] ‘The boys1 read [the girls’2 memories about each other*1/2].’

.  The data reported for child English by Read and Hare (1979) constitute the only exception to this of which I am aware. Young children seem to be able interpret each other with more distant antecedents; this should presumably derive from their not yet having worked out its bipartite semantics. .  An anonymous reviewer states that Russian sebja is like its West Slavic counterparts in having reciprocal function. This is simply not true, as an attempt to translate any of the West Slavic examples with reciprocal siebie/sebe into Russian would demonstrate. .  This is not to imply that disambiguation is impossible. The examples in this article are chosen for their simplicity, but, as an anonymous reviewer observes, the emphatic sam “may serve a disambiguating purpose.” The reviewer provides the following contrast for Polish: (i) Ci koledzy podziwiają siebie samych. ‘These colleagues admire themselves.’ (ii) Ci koledzy podziwiają (tylko) siebie. ‘These colleagues admire (only) each other.’

 Steven Franks

Thus, when Polish siebie has a reflexive meaning, as in (17a), it patterns with Russian sebja. On the other hand, when it has a reciprocal meaning, as in (17b), it patterns with English each other, Russian drug druga, and English themselves. This is because its LF structure in (17a) is something like (18a), and its LF structure in (17b) is something like (18b):14 (18) a.   SOBIE [+anaph, (+refl,) +loc]

b.   OP + SOBIE

[+distributor]

[+anaph, (+recipr,) +loc]

The symbol “OP” in (18b) represents an LF Operator which functions as a logical distributor. This Operator, although present in the syntax and semantics, happens to have no morphological exponence. That is, it is silent in PF. But the absence of an overt distributor has no impact on sobie’s interpretative possibilities, which are determined in LF. There are various ways that one might implement the fact that, at PF, the two versions in (18) are not differentiated. For the sake of concreteness, I assume that the sob-/sieb- root has just the feature [+anaph]. Taking nondistinctness to be a prerequisite for vocabulary insertion (cf. e.g. Siddiqi 2009), this allows the locative form sobie to be inserted into any syntactic slot with (at least) the features [+anaph, +loc]; whether it is going to be interpreted as a reflexive or reciprocal is immaterial. That is, the anaphoric vocabulary item in Polish is simply underspecified: its interpretation as a reflexive or reciprocal depends on syntactic features which are interpretable hence retained in LF, but, from the perspective of PF, the intended reading can only be worked out in context.15 Returning to (17a)/(18a), the syntax provides something like [+anaph, +refl, +loc], which can be replaced by

.  In (18) I place “+refl” and “+recipr” in parentheses because the lexical item SOBIE does not actually make this distinction. .  In the course of considering comparable Czech data, Toman (1991: 170) states that “it is more adequate to say that a regular reflexive pronoun in Czech may have a reciprocal reading rather than singling out a reciprocal subclass.” This view, however, disregards the morphological correlation in (3) that is the point of the present paper, and has nothing to say about why the same item shows different restrictions on the range of antecedents depending on whether it has a reflexive or reciprocal function. Toman goes on to suggest that the “reciprocal reading is simply a function of the plurality of the antecedent,” because the reciprocal reading is unavailable in the absence of a plural potential antecedent. But this is simply because all anaphors need antecedents: the boy cannot bind each other just as it cannot bind herself. If there is a potential plural antecedent, as e.g. in Polish (17), then either reading is available. In short, the syntax and LF distinguish reciprocals from reflexives, the vocabulary item SIEBIE does not, hence is compatible with either. Toman’s confusion in fact highlights my basic argument



Binding and morphology revisited 

(“spelled–out as”) sobie. For (17b)/(18b) we have two nodes in the syntax, roughly [+distributor] and [+anaph, +recipr, +loc]. Here one might imagine that the Polish reciprocal becomes morphologically simplex through the PF–side operation of fusion, which combines these two feature sets into one, i.e. [+distributor, +anaph, +recipr, +loc]. Once again, the form sobie will be inserted under nondistinctness.16 But – whatever the technical implementation – for LF purposes only the local antecedent dziewcząt ‘the girls’ is accessible when sobie is interpreted as a reciprocal. Similarly, if this account is on the right track, we expect Objects and not just Subjects to be accessible antecedents, but only when the anaphor has a reciprocal interpretation. This is indeed correct, as demonstrated in (19):17 (19) a. Chłopcy1 zapytali dziewczęta2 o siebie1/*2. ‘The boys1 asked the girls2 about themselves1/2.’ b. Chłopcy1 zapytali dziewcząt2 o siebie1/2. ‘The boys1 asked the girls2 about each other1/2.’

In (19a), siebie has a reflexive interpretation and is subject-oriented, but in (19b), where siebie has a reciprocal interpretation, it is free to refer to the Object as well. Thus, reflexive siebie patterns like Russian sebja and not English themselves, as expected since both are heads, but reciprocal siebie patterns like Russian drug druga and English each other, just as if siebie were a phrase here. In short, morphological PF–LF mismatches exist and, as expected, although pronunciation necessarily reflects PF, interpretation necessarily reflects LF. 3.3  Czech In Czech as well, the reflexive behaves like its Russian counterpart. However, as in Polish, in Czech the same vocabulary item can also function as a reciprocal, and – again just as in Polish – the judgments vary accordingly. Avgustinova et al. (1999) and Toman (1991) provide considerable Czech data which make these

that, because under GB/Minimalism matters of PF do not determine interpretation, the actual form of the reflexive/reciprocal is technically irrelevant to LF. .  Alternatively, the feature [+distributor] is deleted, or it is retained but there is no ­vocabulary item to insert into that slot. However this is formalized, it is part of the process of Spell–Out and not relevant to LF. .  An anonymous reviewer points out that the word order dziewcząt wspomnienia would be preferable as stylistically unmarked. This does not however affect the judgments in (17), which is drawn directly from Reinders-Machowska (1991).

 Steven Franks

points (although only Avgustinova et al. recognize the correlation between interpretative function and range of possible antecedents). First, when sebe functions as a reflexive, it can be long-distance bound, both out of an infinitival clause as in Avgustinova et al.’s (20a) and out of a possessed NP, as in Toman’s (20b):18 (20) a. Vedoucí1 diplomové práce doporučil Petrovi2 [PRO2 citovat sebe1/2]. ‘The supervisor1 of his thesis recommended to Peter2 [PRO2 to cite himself*1/2].’ b. Jana1 zahodila [Karlovy2 básně o sobě1/2]. ‘Jana1 threw away [Karl2’s poems about himself2/*herself1/her1].

Similarly, the examples in (21) show that, as expected, a Subject antecedent is required:19 (21) a. V zrcadle Jan1 ukázal Petra2 sobě1/??2. ‘Jan1 showed Peter2 himself1/2 in the mirror.’

b. *Jan1 mluvil s Marií1 o sobě1/*2. ‘Jan1 talked with Maria2 about himself1/?herself2.’ c. Jana1 navštívila Krla2 kvůli sobě1/*2. ‘Jana visited Karl for her own benefit.’

Now consider how Czech sebe behaves when it has a reciprocal interpretation. The examples in (22), from Avgustinova et al. (1999), show that in this case an Object antecedent is perfectly acceptable: (22) a. Zamotal do sebe několik pohádek. ‘He entangled several fairy tales into each other.’ b. Položili meče (ostřím) proti sobě. ‘They laid the swords (with the blades) against each other.’

.  Toman (1991: 155) however cites (i) to argue that reflexivization in Czech, like himself in English but unlike sebja in Russian, “is constrained to clauses”: (i) *Karel1 nas nutil, oholit sebe1. ‘*Karel1 forced us to shave himself1.’ .  These examples raise various peripheral questions. For one thing, it is unclear why the Object antecedent is marked as highly degraded rather than “*”; the best way to refer to Pietra here would be to use jemu samému. This example also comes up against the issue of whether the direct object c-commands the indirect object (as needed to obtain a reading in which Petra is the antecedent), or the other way around. (21b) also raises a c-command issue, since s Marií branches, although it is worth noting that the same branching in English with Mary does not seem to prohibit binding by Mary. For this reason I also offer (21c), from Toman (1991), which avoids the c-command problem.



Binding and morphology revisited 

Finally, one can once again construct minimal pairs in which the set of possible antecedents correlates with the reflexive versus reciprocal interpretation. The former is given in (23), the latter in (24):20 (23) a. Pan Novák1 poštval sousedy2 proti sobě1/*2. ‘Mr. Novak incited the neighbors against himself/themselves.’ b. [Markrabě a vévoda]1 poručili tělesným strážcům2 [PRO2 chránit sebe1/2]. ‘[The markgrave and the earl]1 ordered the bodyguards2 [PRO2 to protect themselves*1/2].’

(24) a. Pan Novák poštval sousedy proti sobě. ‘Mr. Novak incited the neighbors against each other.’ b. [Markrabě a vévoda]1 poručili tělesným strážcům2 [PRO2 chránit sebe*1/2]. ‘[The markgrave and the earl]1 ordered the bodyguards2 [PRO2 to protect each other*1/2].’

As a reflexive, sebe in (23b) is ambiguous, but as a reciprocal, sebe in (24b) can only refer back to the local PRO subject (itself controlled by the main clause Object, tělesným strážcům). 3.4  Bulgarian On the flip side, one might expect mismatches that go in the opposite direction, that is, morphologically complex anaphors which seem not to warrant any complexity at LF. These are more difficult to pin down because, as discussed by König and Siemund (2000) inter alia, many languages exhibit an emphatic reflexive that is morphologically phrasal since it is composed of a simplex anaphor plus an adnominal intensifier, such as German sich selbst. These have logophoric and other special focus properties that introduce complications not only for the traditional binding theory but also for the morphological predictions discussed in the present paper.21 Since, however, these predictions are based on PF rather than LF morphology, there is some “wiggle room”; in particular, intensifier and/or focus features may imply an Operator in LF, so that the anaphor is indeed phrasal.

.  (23a) is from Toman (1991) and (23b) from Avgustinova et al. (1999). They add ne však již své poddané ‘but not their subjects’ to coerce the long-distance reading. .  See König and Siemund (2000) for details, as well as Kibrik and Bogdanova (1995) for discussion of Russian sam. I thank an anonymous reviewer for drawing my attention to this latter work.

 Steven Franks

Putting this problem aside, one likely place to look for a true mismatch is Bulgarian. This language exhibits a clearly phrasal anaphor, namely sebe si, which consists of the simplex anaphor sebe plus the adnominal “dative” clitic si. Like reflexives in other Slavic languages, the form sebe si lacks φ-features: it is unmarked for person, number, or gender, hence is featurally nondistinct from any antecedent. Unlike other Slavic languages, however, Bulgarian does not use sebe without the otiose clitic si.22 Since this si is completely redundant semantically in Bulgarian, it should not have an LF correlate, hence we should expect Bulgarian sebe si to pattern with Russian sebja. According to the data below, from Schürcks (2003), this prediction is borne out: (25) a. Ivan1 razkazva na doktora2 [istorii za sebe si1/*2]. ‘Ivan1 tells the doctor2 [stories about himself1/2].’ b. Ivan1 pročete [Petrovata2 statija za sebe si1/2]. ‘Ivan1 read [Peter’s2 article about himself?*1/2.]’

c. Ivan1 popita bašta2 si za [Petrovata3 statija za sebe si1/*2/3]. ‘Ivan1 asked his father2 about [Peter3’s article about himself*1/*2/3].’

In Bulgarian (25a) sebe si exhibits subject-orientation and in (25b) it exhibits long-distance binding, over the Subject Petrovata;23 (25c) combines both of these properties. The range of possible interpretations thus contrasts with the English sentences in (25), indicating that Bulgarian sebe si – PF appearances to the contrary – functions as a head rather than a phrase. That is, by all available diagnostics, Bulgarian sebe si behaves as if simplex for purposes of the binding theory, demonstrating the irrelevance of its complex morphological structure on

.  In Macedonian, the si is optional in objective positions, but obligatory in dative ones. The following are offered by Mišeska Tomić (in press): (i) Go zede so sebe(si). ‘(S)he took it with herself/himself.’ (ii) Si go dade sebe*(si). ‘(S)he gave it to herself/himself.’ Although I do not examine possessives here, it is worth noting that the Bulgarian possessive svoj only optionally occurs with clitic si, in this respect resembling both sebe and svoj reflexives in Macedonian. The ramifications of this and related phenomena are explored from a larger Balkan perspective in Franks (in press). .  Note that long-distance binding out of infinitives, although predicted, cannot be demonstrated for Bulgarian since this language lacks infinitives.



Binding and morphology revisited 

the PF–side. It has an LF structure similar to Russian (6a), as indicated in (26a), but undergoes morphological fission to produce a PF structure more like (26b):24 (26)

a.    SEBE + SI



b.   sebe + si

[+anaph, +refl, +obj] [N, +anaph, +refl, +obj]  [K, +anaph, +refl, +dat]

Before concluding, it is worth noting that Bulgarian (but, for some reason, not Macedonian), has a colloquial alternative to sebe si which is morphologically reminiscent of English himself. As Schürcks (2003, 2006) shows, however, neither c-command nor locality requirements seem to pertain to Bulgarian nego si. Thus, not only can local Objects antecede, as in (27), but so can distant and even noncommanding expressions, as in (28): (27) Ivan1 popita Penčo2 za nego si1/2. ‘Ivan1 asked Pencho2 about himself1/2.’ (28) a. Ivan1 kazva, [če doktor t2 mrazi nego si1/2]. ‘Ivan1 says [that the doctor2 hates himself*1/2].’

b. [Sin t1 na [Ivanovija2 brat3]] kritikuva nego si1/2/3. ‘[The son1 of [Ivan2’s brother3]] criticizes himself1/*2/*3.’

While it is not surprising that, in contrast to sebe si in (25), nego si in (27) can be anteceded by the Object Penčo, the data in (28) show that nego si is not comparable to English himself, since it behaves as if literally “exempt” from the binding theory. Replacing nego si in (28) with sebe si, as in (29), shows that sebe si is, on the other hand, perfectly “well behaved”: (29) a. Ivan1 kazva, [če doktor t2 mrazi sebe si*1/2]. b. [Sin t1 na [Ivanovija2 brat3]] kritikuva sebe si1/*2/*3.

What is going on with nego si is thus considerably more complex than what would be expected just by adding φ-features to the anaphor. Although space considerations prevent further discussion here, Schürcks (2003, 2006) offers an answer in terms of the notion of “point of view” put forward in Kuno (1987).25

.  In (26b), N indicates that sebe is a full nominal and K indicates that si is a case clitic. .  Franks (in press) argues that nego, as well as its counterparts in Greek, Turkish, and Albanian, freely refers as a pronoun, and that it binds the reflexive element (si in Bulgarian) within the extended nominal projection. This pronoun is however “protected” by additional functional structure, so that no local disjoint reference effects arise either, hence even local c-commanding antecedents are possible in e.g. (27) and (28).

 Steven Franks

4.  Conclusion This paper has addressed the problem of how the morphological structure of anaphors can feed the two interface components, LF and PF, in a way that leads to the right interpretative results. Assuming that it is only PF–side morphological processes which can manipulate structures, LF morphology should not encode anything more than what has been provided by the syntax. I have argued that PF morphology, on the other hand, can encode less than required by LF, as in the case of Polish and Czech “head” reciprocals, or more than LF, as in the case of B ­ ulgarian “phrasal” reflexives. However, since binding is established in LF, s­ emantic interpretation is necessarily oblivious to the facts of overt morphology. This conclusion, inevitable under any approach that relegates binding to an intepretative semantic level such as GB/Minimalism’s Logical Form or Jackendoff ’s more generic Conceptual Structure,26 has the advantage of providing a way around prima facie counterexamples to the PF morphological generalizations in (3).27 What you see may be less than what you get (West Slavic reciprocals) or it may be more (Bulgarian reflexives). Of course, one essential question remains unresolved: How does an anaphor’s LF structure come to impact on the set of possible antecedents it can consider and why, in particular, does phrasal versus head structure in LF matter? In this short paper I have put this far from trivial matter of execution aside, leaving it for future research.

References Avgustinova, Tania, Gardent, Claire & Oliva, Karel. 1999. Binding of reciprocals with particular respect to Czech, CLAUS-Report 109, Universität des Saarlandes. Bailyn, John F. 1992. LF movement of anaphors and the acquisition of embedded clauses in Russian. Language Acquisition 2: 307–335. Battistella, Edwin. 1989. Chinese reflexivization: A movement to INFL approach. Linguistics 27: 987–1012.

.  Although it should be noted that in Jackendoff ’s (2002) architecture of grammar, unlike under GB/Minimalism, Conceptual Structure can in principle directly correspond to ­morphology. .  More of these are offered in Huang 1996, who therefore rejects the morphology–­ semantics correlation. My response is of course that the correlation needs to be expressed at a more abstract level, namely through the mediation of syntactic structure.



Binding and morphology revisited 

Bošković, Željko. 2008. What will you have, DP or NP? In NELS 37: Proceedings of the 37th annual meeting of the North East Linguistic Society, Vol. 1, Emily J. Elfner & Martin Walkow (eds), 101–114. Amherst MA: GSLA. Chomsky, Noam. 1980. On binding. Linguistic Inquiry 11(1): 1–46. Chomsky, Noam. 1981. Lectures on Government and Binding. Dordrecht: Foris. Cole, Peter & Sung, Li-May. 1994. Head movement and long-distance reflexives. Linguistic Inquiry 25(3): 355–406. Franks, Steven. in press. Bulgarian NEGO SI is a Balkan anaphor. Linguistique balkanique 52(2/3), Sophia. Franks, Steven L. & Connell, Phil J. 1996. Knowledge of binding in normal and SLI children. Journal of Child Language 23: 431–464. Heim, Irene, Lasnik, Howard & May, Robert. 1991. Reciprocity and plurality. Linguistic Inquiry 22(1): 63–101. Huang, Yan. 1996. A note on the head-movement analysis of reflexives. Linguistics 14: 833–840. Jackendoff, Ray. 2002. Foundations of Language. Oxford: OUP. Jakobson, Roman. 1936[1971]. Beitrag zur Allgemeine Kasuslehre: Gesamtbedeutungen der russischen Kasus. In R. Jakobson, Selected Writings, II: Words and Language, 23–71. The Hague: Mouton. Kibrik, Aleksandr E. & Bogdanova, Ekaterina A. 1995. SAM kak operator korrekcii ožidanij adresata. Voprosy jazykoznanija 3: 28–47. König, Ekkehard. & Siemund, Peter. 2000. Intensifiers and reflexives. In Reflexives: Forms and Functions, Vol. 1 [Typological Studies in Language 40], Zygmunt Frajzyngier & Traci Walker (eds), 41–74. Amsterdam: John Benjamins. Kuno, Susumo. 1987. Functional Syntax: Anaphora, Discourse, and Empathy. Chicago IL: The University of Chicago Press. Lightfoot, David. 1989. The child’s trigger experience: Degree-0 learnability. Behavioral & Brain Sciences 12: 321–334. Mišeska Tomić, Olga. 2012. A Grammar of Macedonian. Bloomington IN: Slavica. Padučeva, Elena V. 1985. Vyskazyvanie i ego sootnesennost’ s dejstvitel’nost’ju. Moscow: Nauka. Pica, Pierre. 1987. On the nature of the reflexivization cycle. In Proceedings of NELS 17, Joyce McDonough & Bernadette Plunket (eds), 483–499. Amherst MA: University of Massachusetts. Progovac, Ljiljana. 1992. Relativized SUBJECT: Long-distance reflexives without movement. Linguistic Inquiry 23(4): 671–680. Rappaport, Gilbert. 1986. On anaphor binding in Russian. Natural Language and Linguistic Theory 4(1): 97–120. Read, Charles & Hare, Victoria Chou. 1979. Children’s interpretations of reflexive pronouns in English. In Studies in First and Second Language Acquisition, Fred R. Eckman & Ashley J. Hastings (eds), 98–116. Rowley MA: Newbury House. Reinders-Machowska, Ewa. 1991. Binding in Polish. In Long-distance Anaphora, Jan Koster & Eric J. Reuland (eds), 137–150. Cambridge: CUP. Schürcks, Lilia. 2003. Binding and Bulgarian [Groningen Dissertations in Linguistics 44]. Groningen: University of Groningen. Schürcks, Lilia. 2006. Binding and point of view in Bulgarian. Zeitschrift für Slawistik 51: 386–405.

 Steven Franks Siddiqi, Daniel. 2009. Syntax within the Word: Economy, Allomorphy, and Argument Selection in Distributed Morphology [Linguistik Aktuell/Linguistics Today 138]. Amsterdam: John Benjamins. Toman, Jindřich. 1991. Anaphors in binary trees: An analysis of Czech reflexives. In Longdistance Anaphora, Jan Koster & Eric J. Reuland (eds), 151–170. Cambridge: CUP. Wexler, Kenneth & Manzini, M. Rita. 1987. Parameters and learnability in binding theory. In Parameter Setting, Thomas Roeper & Edwin Williams (eds), 41–76. Dordrecht: Kluwer. Yang, Dong-Whee. 1983. The extended binding theory of anaphors. Language Research 19: 169–192.

Possessor Raising and Slavic clitics* Anton Zimmerling

Moscow State University for the Humanities (SMSUH), Russia The paper discusses syntactic features of Slavic possessive clitics and Slavic constructions with so-called “Possessor Raising”. I prove that only a minority of Slavic languages have true phrase-level (NP-level or DP-level) possessive clitics and argue against a generalized syntactic account of all Slavic constructions with possessive operators. The weak aspects of the PR hypothesis are that it takes the mapping of syntax and possessive semantics to be iconic and the rules/principles of extracting a NP/DP-level possessive operator out of the NP/DP to be trivial. These assumptions are poorly justified empirically. Slavic languages typically apply different case-marking for non-agreeing phrase-level possessive operators and non-agreeing clause-level possessive operators. A group of languages, including Modern Russian, lacks phrase-level possessive clitics. For this group the PR hypothesis cannot be validated. Most cases where the PR hypothesis has been proposed in previous Slavic studies do not conform to the definition of Raising as a syntactic operation, since the identity of structures with a clause-level or phrase-level possessive operator cannot be established. Such cases must be reanalyzed in terms of Possessive Shift, i.e. alternation of a true possessive construction with a NP/DP-level possessive element and its quasi-synonym, a pseudo-possessive construction with a case-marked verbal argument. Keywords:  possessor raising; possessive shift; clitics; Slavic languages

1.  Possessor Raising and possessive shift In descriptive terms “PR” refers to a quasi-synonymic transformation where a phrase-level possessive operator located in a NP/DP and expressed by a clitic/free pronoun/NP is made an argument of the clausal predicate (cf. Szabolcsi 1983). * The paper is written with financial support from the Russian Ministry of Education and Culture, project 2012–1.2.1-12-000-3004-9859 ‘Models of Case Grammar and Natural Languages’. The author is grateful to the audience of SLS-6 conference (Aix-en-Provence, University of Provence, September 1-3, 2011) and, personally, to Steven Franks, Frank ­ Gladney and Grigory Krejdlin for the valuable comments. I am also grateful to the anonymous ­reviewers for their criticism. All responsibility for shortcomings is my own.

 Anton Zimmerling

PR and other instances of raising can be analyzed as stages in the derivation of a sentence (cf. Don-Won Lee 2004; Pei-Jung Kuo & Yi-An-Lin 2008). In Slavic, specifically, in Russian studies, a description of possessive constructions in terms of PR became standard after the publications of Alexander Kibrik, (cf. Kibrik 2000, 2003), though there is no consensus whether underlying possessivity should be treated as a purely semantic relation (cf. Seliverstova 1990) or as a syntactic feature associated with some presumably prototypical possessive constructions, be it bare genitive NPs in all languages of the world, cf. Kibrik (2003: 307) or languagespecific constructions such as the Russian construction with the preposition u + genitive NP, cf. Mel’čuk & Iordanskaja 1995; Raxilina 2000: 54. The term “Possessive Shift” is introduced in this paper for a quasi-synonymic relation of sentences with a possessive operator that may be pragmatically ­equivalent in some contexts but have different syntactic structure. I argue that different placement of a possessive element may either change syntactic structure or preserve it, depending on the value morphosyntactic parameters assume in a given language. In some Slavic languages phrase-level and clause-level possessive operators are marked with different morphological cases. For instance, Russian phrase-level possessives are genitives, while Russian clause-level possessives are datives. Pairs of sentences like (a) ru. Ona ne [NP doč’ PetrovaGEN/ egoGEN doč’] ‘She is not Petrov’s daughter/his daughter’ ~ (b) Ona emuDAT ne doč’/ ­PetrovuDAT ne doč ” should be analyzed as Possessor Shift, i.e. as an alternation between different structures. In Modern Russian Possessive Shift cannot be analyzed as Raising, since both the syntax and the argument marking in (a) and (b) are different. In Old Church Slavonic, Modern Bulgarian, and Macedonian Possessive Shift is bound to the use of pronominal clitics. These are marked by the same overt case (dative), both on the phrase-level and on the clause-level. For this group of languages, a PR analysis of clausal possessive forms remains possible. In Modern Serbo-Croatian, dative possessives in clausal 2P are marginally acceptable, according to Pennington 2010, but phrase-level dative possessives in SC are ungrammatical. A similar result can be shown for Modern Russian, a language lacking pronominal clitics. Here, clause-level dative possessive pronouns are marginally acceptable, cf. (1a), while phrase-level dative possessives are ungrammatical, cf. (1b). (1) Russian a. Ja sebe ne vrag. I:nom ref.dat neg enemy:nom ‘I am not an enemy to myself.’ b. *Ja vstretil vraga sebe.  I:nom met:pf.m enemy:acc ref.dat Intended: ‘I met my own enemy.’



Possessor Raising and Slavic clitics 

The two oldest Slavic idioms – Old Church Slavonic (OCS) and Old Northern Russian (ONR) – exemplify two extremes: ONR completely lacked dative possessives, while in OCS they were common both in clausal-second position (2P)1 and on the phrase-level. The clausal 2P in Slavic languages typically hosts argument and reflexive pronominal clitics (cf. Dimitrova-Vulchanova 1999; Franks & King 2000; Zimmerling 2008). Following Zimmerling & Kosta 2013, I claim that the majority of Slavic languages only have clause-level possessives and that these pattern with argument dative clitics. Modern Bulgarian and Macedonian give the best chances to check the Raising hypothesis, since these two languages both seem to have DPlevel dative clitics (cf. Mišeska Tomić 2004; Franks et al. 2004) and mechanisms allowing for extracting dative clitics out of DP. In Franks & King 2000 and Franks 2008 pronominal and auxiliary Bulgarian clitics are analyzed as verb-­adjacent proclitics in syntax but at as phonetic 2P enclitics that do not stand clause-­initially due to a presumably non-syntactic condition. Dimitrova-Vulchanova 1999; ­Zimmerling 2006 and Zimmerling & Kosta 2013 analyse all Bulgarian clustering clitics as 2P elements. There are two competing accounts of Bulgarian possessive clitics. According to Schürcks & Wunderlich 2003, Bulgarian dative possessives raise out of DP to a position overtly resembling clausal 2P2 where they cluster with other 2P clitics, such as Bulgarian yes-no particle li, in (2a) and (2b). (2) Bulgarian a. Tja nameri= li [DP užasni-te= si greški]? she found:3sg Q   horrible-the ref.dat mistakes ‘Did she find her horrible mistakes?’ b. Tja nameri= li= si [DP užasni-te___ greški]? she found:3sg Q ref.dat   horrible-the mistakes ‘Did she find her horrible mistakes?’

.  I am adopting a traditional view that ‘clausal 2P’ is a position or a block of adjacent positions that can be defined in terms of syntax, cf. Progovac 1996 or syntax-prosody interface, cf. Zaliznjak 2008 and do not take into account an alternative hypothesis raised by Agbayani & Golston 2010 who argue that ‘2P’ is an epiphenomenon and claim that clitic hosts of 2P clitics always lie clause-external to them. Agbayani and Golston’s analysis is difficult to apply to Slavic clustering clitics. .  The exact definition of the position taken by Bulgarian dative possessives depends on the analysis of Bulgarian pronominal and auxiliary clitics – whether they are explained as verb-adjacent elements in syntax or as 2P elements. I am adopting a 2P analysis to Bulgarian clustering clitics and classify them as enclitics (X=CL) in this paper. The problem of clitic hosts is however of minor importance for the analysis of examples like (1b), since clauselevel clustering clitics and clusters like =li=si in (1b) end up in verb-adjacent positions, as an anonymous reviewer justly pointed out.

 Anton Zimmerling

According to Cinque & Krapova 2013 Bulgarian possessives do not raise in sentences denoting inalienable possession (3a–b) and are never extracted out of PPs, cf. (4). (3) Bulgarian a. Toj mi= se= izkrjaska [PPv [DP uxoto]]. he me:dat.sg ref.acc shouted:3sg   in   ear-the ‘He shouted in my ear.’ b. Toj se= izkrjaska [PPv [DP uxoto= mi]]. he ref.acc shouted:3sg    in   ear-the me:dat.sg ‘He shouted in my ear’. (4) Bg. *Az ì= mislja [PPza [DP očite___]]   I her:dat think:1sg    for   eyes-the Intended: ‘I think of her eyes.’

I adopt Cinque and Krapova’s view and treat Bulgarian examples with inalienable possession like (3a–b) as instances of Possessive Shift, i.e. as pairs of different syntactic structures with a possessive operator, along the same lines as Russian examples (1a–b). Bulgarian sentences with alienable possession, such as (4), however seem to be compatible with a PR analysis. 2.  Possessivity: Semantics and syntax In this section, I discuss the basic facts concerning the semantics and syntax of possessive constructions. A seminal analysis of Russian existential and possessive predicates has been proposed in Arutjunova & Širjaev 1983 who analyze both semantic and pragmatic/communicative features of Russian possessive constructions. An analysis in terms of PR has been proposed for Russian by Kibrik (2000, 2003) who takes for granted that the prototypical function of genitive phrases in Russian and beyond is ‘encoding of the possessive semantic relation’.3 However, the interpretation of all phrases of the type [NP N0 Ngen] as encoding the possessive relation is too broad and does not provide any basis for separating possessive relation of the type

.  The feature ascribed in Kibrik (2003: 307) to all genitive phrases is more likely to be associated with a bulk of constructions called ‘weak genitives’ in Graščenkov 2007. Standard (i.e. ‘strong’) genitive markers link together two arguments, expressed by two NPs, X and Y. The tag ‘weak genitives’ is reserved for languages which apply the same markers both for linking two nominal expressions together as well as for adjoining relative clauses, adjectives, pronouns. Graščenkov (2007: 44) claims that weak genitive markers tend to be phrasal affixes, not true case affixes. If this claim is true, the notion of ‘weak genitive construction’ is not ­applicable to most Slavic phrase-level possessive operators.



Possessor Raising and Slavic clitics 

‘X has/owns Y’ from the WHOLE: PART relation or the characterization r­elation ‘X has feature Y’ etc. which are analyzed usually differently since ­Arutjunova & Širjaev 1983, cf. also Raxilina 2000: 36–55. Therefore, some semantic and syntactic conditions must be imposed on the configuration of Possessors and Possessees. Mel’čuk and Iordanskaja 1995 and Raxilina (2000: 54–56) discuss the distribution of Russian constructions ‘Y X-a’ and ‘u X-a est’ Y’ at some length and arrive at the conclusion that both constructions encode the whole-part relation, though in a different perspective: in a bare genitive construction the most communicatively salient component is WHOLE, while the u + gen construction brings about a PART-­ oriented perspective, cf. Raxilina 2000: 54.4 Mel’čuk and Iordanskaja (1995: 152) and ­Raxilina (2000: 156) in this context claim that in the Russian u + gen construction, cf. U MašiGEN glazaNOM.PL golubyeADJ.NOM.PL, ‘Masha’s eyes are blue’ the possessor is always focal. The term ‘focalization’ is infelicitous here, since under a standard reading cf. (5a) the possessor is topical (thematic), not focalized, while true focalization in Russian normally requires a different word order and accent marking, cf. (5b). In the notation of (5a–b) I use accent tags ‘↗’ for a standard Russian topic accent (a steep rise followed by a steep fall on posttonics if any) and ‘↘’ for a standard ­Russian focus (rheme) accent; the lowercase tags ‘T’ and ‘R’ stand for Topic and Focus respectively, while the lowercase tag ‘R Proper’ refers to the part of the Focus constituent which takes the focus accent. The lowercase tag ‘Contr’ indicates that the corresponding communicative constituent is contrastive. Curly brackets stand for communicative constituents, square brackets stand for formal constituents. (5) Russian a. {T [PP U ↗Maši]} {R [NP glaza] {R Proper [PredP ↘golubye}]}. by Masha:gen    eyes:nom       blue:nom.pl ‘Masha’s eyes are blue.’ b. {T CONTR [NP ↗Glaza] [PredP golubye]}       eyes:nom    blue: nom.pl {R CONTR [PP u ↘Maši]}.        by Masha:gen

‘It is Masha (not someone else) whose eyes are blue.’

Kibrik (2003: 309) introduces another term for Mel’čuk’ and Iordanskaja’s ‘focalized possessors’ and claims that Possessor is ‘a local focus of empathy in a genitive phrase’. This is a terminological improvement. However, Kibrik’s syntax-oriented approach to genitive NPs (and to other structures analyzed in his theory as derived from genitive phrases by virtue of PR and other transformations) is not directly

.  This point defended by Mel’čuk and Iordanskaja (1995) and Raxilina (2000) goes back to Krejdlin’s analysis (1979), which is explicitly stated in Raxilina 2000: 54.

 Anton Zimmerling

compatible with Mel’čuk’ and Iordanskaja’s semantics-oriented approach. Any WHOLE-oriented structure, including Russian bare genitive NPs (i.e. prototypical possessive phrases in Kibrik’s theory) won’t be acknowledged as conveying possessive semantics in Mel’čuk and Iordanskaja’s and Raxilina’s theories. A further problem with Mel’čuk and Iordanskaja’s and Raxilina’s description of Russian data is that neither of them is framework-neutral. They strongly rely on principles of dependency syntax and construction grammar respectively. In Mel’čuk’s dependency syntax, where the notion of constituency does not play any role, the distribution of bare genitive phrases of the type ‘Y X-a’ and preposition structure ‘u X-a Y’ can be interpreted as a partial contrast of two Russian constructions each of which has its idiomatic semantics. This approach is illustrated by contexts where the segment ‘Y X-a’ can be substituted with a segment ‘Y u X-a’, cf. Russian 〈Ščeki Maši〉 vspyxnuli ‘Maša’s cheeks flushed slightly’ vs 〈Ščeki u Maši〉 vspyxnuli ‘the same’ without a clear difference in semantics, cf. Mel’čuk and Iordanskaja (1995: 147) as well as by contexts where such a substitution cannot be made and only one construction is possible, cf. Russian 〈Glaza u Maši〉 golubye ‘Maša’s eyes are blue’ but not *〈Glaza Maši〉 golubye ‘Maša’s eyes are blue’. The substitution test in contexts like 〈Ščeki Maši〉 vspyxnuli ~ 〈Ščeki u Maši〉 vspyxnuli brings about an illusion that expressions like ‘Y X-a’ and expressions like ‘Y u X-a’ are possessive operators of the same level. This is entirely misleading and will be proven wrong in any version of constituency analysis. Indeed, an adjacent position of the Possessor (X) and Possessee (Y) in sentences 〈Ščeki u Maši〉 vspyxnuli or 〈Nožka u stola〉 gnilaja ‘〈The leg of a table〉 is rotten’ results from a movement transformation, specifically, extraction of a head noun out of an NP. This transformation has a clear semantic load. In example (6a) entire VP is focal, with the focus accent on the NP ščeki (which has the role of Possessee). In the derived structure (6b) this NP is topicalized and made part of the Topic, while the verb vspyxnuli is in narrow focus. (6) Russian a. {T [PP U ↗Maši]} {R [VP vspyxnuli {R PROPER [NP ↘ščeki]}}.     by Masha:gen    flushed-slightly:pl     cheeks:nom ‘Masha’s cheeks flushed slightly.’ b. {T {T PROPER [NP ↗ščeki]i} [PP u        cheeks:nom    by Maši]} {R [VP ↘vspyxnuli] ti}.5 Masha:gen    flushed-slightly:pl

‘Masha’s cheeks flushed slightly.’

.  Remarkably, in Kibrik’s theory (2000: 308) instances with a topicalized Possessee are classified with PR (ekstrapozicija vnešnego posessora) not with Possessee extraction (ekstrapozicija obladaemogo).



Possessor Raising and Slavic clitics 

That the moved NP [NP ščeki] forms in the topicalized structure (6b) one and the same communicative constituent with the PP [PP u Maši] certainly does not prove that they form one formal constituent. It is evident that PPs of the type ‘u + gen’ are not phrase-level, but clause-level possessive operators.6 Following Seliverstova 1990 and a bulk of preceding literature going back to Benveniste 1960, I specify that as a semantic relation, the possessive relation is strictly binary and assume that imposing a condition that the Possessor is animate gives the best chances for keeping apart possessive relation ‘X has/owns Y’ from other semantic relations, including the WHOLE: PART relation ‘Y is a part of X’ and the attributive relation ‘X has a feature Y’. The Possessee is normally inanimate. In a typical communicative reading, the Possessor is a Topic/Theme, while the Possessee and the verb/auxiliaries are parts of the Focus/Rheme; cf. Arutjunova & Širjaev 1983. (7) Ru. {T [PP U ↗Maši]} {R[VP est’ kvartira [PPv ↘Lubercax]]}.    by Masha:gen   be:3sg flat:nom    in  Lubercy:loc ‘Masha has a flat in Lubercy.’

The possessive relation intersects with the WHOLE: PART relation, though not all sentences expressing the WHOLE: PART relation are possessive in the specified sense. The possessive relation is grammaticalized in a different way with alienable and inalienable possession, cf. Aikhenvald (1998: 93). In some languages this asymmetry also affects the whole: organic part relation, but most Slavic languages keep them apart. In Modern Russian the meaning ‘the leg of a table’ can be expressed both by a bare genitive NP, cf. [NP nožka [NP stola]] and by a PP with a preposition ot ‘of, from’, cf. [NP nožka [PP ot [NP stola]],7 but with an animate Possessor the insertion of ot is no longer possible: rot Maši ‘Masha’s mouth’, but not *rot ot Maši. Claims raised elsewhere that Russian and probably other Slavic languages belong to the class of world’s languages which allow inanimate possessors, cf. Herslund & Baron 2001 are based not on semantic considerations but on the observation that these languages apply similar constructions for encoding possession and for expressing other semantic relations, cf. (8a–d) and (9a–d). Examples (8a) and (8c) where the Possessor is a Topic and the remaining part of the sentence is in Focus can be regarded as basic, while variants (8b) and (8d) show topicalization of the Possessee.

.  This fact is acknowledged by Russian academic grammars as well, cf. Russkaja grammatika (1982: 149–151), where u + gen phrases are recognized as the so called ‘determinants’, i.e. immediate daughters of S. .  Selective restrictions on the use of Russian ot are mentioned in Raxilina (2000: 43).

 Anton Zimmerling

(8) Russian a. {T [PP ↗U Maši]} {R [VP Ø [PredP [NP gniloj ↘zub]}. by Masha:gen be:prs.ind    rotten:nom.m tooth:nom ‘Masha has a bad (lit.: rotten) tooth.’ b. {{T [NP ↗zub] i [PP u Maši]} {R [VP Ø [PredP ↘gniloj ti]}}. tooth:nom    by Masha:gen be:prs.ind      rotten:nom.m ‘Masha’s tooth is rotten.’ c. {T[PP ↗U Maši]} {R[VP sgnil [NP ↘zub]}.    by Masha:gen   rotted-away:m tooth:nom ‘Masha’s tooth rotted away.’ d. {T [NP ↗Zub] [PP u Maši]} {R[VP ↘sgnil]}.8    tooth:nom   by Masha:gen   rotted-away:m ‘Masha’s tooth rotted away.’

In a similar way, in a construction encoding the WHOLE: PART relation, one gets two variants with a topical argument with the role ‘WHOLE’ and the rest of the sentence in broad focus, cf. (9a) and (9c) and two variants with a topicalized argument with the role ‘PART’, cf. (9b) and (9d). (9) Russian a. {T [PP ↗U stola]} {R[VP Ø [PredP [NP gnilaja ↘nožka]}. by table:gen   be:prs.ind    rotten:nom.f leg:nom.f ‘The table has a rotten leg.’ b. {{T [NP ↗nožka] i [PP u stola]} {R [VP Ø leg:nom by table:gen be:prs.ind [PredP ↘gnilaja ti]}}.      rottenNOM.SG.F.

‘The leg of the table is rotten.’

c. {T [PP ↗U stola]} {R [VP sgnila [NP ↘nožka]}. by table:gen   rotted-away:f leg:nom ‘The table’s leg rotted away.’ d. {T [NP ↗Nožka] [PPu stola]} {R [VP ↘sgnila]}.9 leg:nom   by table:gen    rotted-away:f ‘The leg of the table rotted away.’

.  The linear order in (8d) can also be linked with a different communicative reading where the extracted NP is focalized and the rest of the sentence is deaccented: {R [NP ( Zub] [PP u Maši]} [VP sgnil]}. .  The linear order in (9d) can also be linked with a different communicative reading where the extracted NP is focalized and the rest of the sentence is deaccented: {R [NP ( nožka] [PP u stola]} [VP sgnila]}.



Possessor Raising and Slavic clitics

The parallelism of (8a–d) and (9a–d) can be interpreted differently. A straightforward solution is to admit that Russian allows inanimate possessors. This kind of analysis, is however based on the assumption that all non-locative uses of the u + gen construction are possessive.10 The validity of this assumption is far from self-evident and it is reasonable to test an alternative hypothesis that there is no direct mapping between possessive semantics and the structure of the u + gen construction. Constructions expressing the possessive relation are labeled “possessive”, though they usually express non-possessive meanings in the same language as well. (10) Russian a. U menja est’ karta. by me:gen be:3sg map:nom ‘I have a map.’

b. U menja est’ podozrenie čto P. by me:gen be:3sg suspicion:nom that P. ‘I have a suspicion that P.’

Russian (10a) and its English equivalent are true possessive sentences, while Russian (10b) and its English equivalent may be called pseudo-possessive, cf. Zimmerling 2000. The notion of “pseudo-possessivity” can also be applied to syntax if there is no real possessive relation between the predicate arguments on the surface level, but such a relation can be reconstructed by postulating a transformation, cf. Szabolcsi 1983. Along these lines, Russian (11a) meaning ‘X sewed a button on Y’s coat’ should allow for possessive structure (11b) with the intended meaning ‘Y’s button’. However, this analysis is on the wrong track, since the Russian sentence with the reconstructed NP pugovica rebenka ‘child’s button’ is ill-formed, as shown by the ungrammaticality of (11c): (11) Russian a. Ona [VP prišila rebenku pugovicu]. she:nom   sew:f child:dat button:acc ‘She sewed a button on the child〈’s clothes〉.’ b. [NP pugovica [NP rebenka].   button:nom   child:gen ‘the child’s button.’

.  It is evident that locative and non-locative uses of Russian u + gen phrases have different properties, cf. U našego doma tri magazina ‘There are three shops by our house’(locative u + gen phrase) vs U našego doma tri dveri ‘Our house has three doors’ (non-locative u + gen phrase).



 Anton Zimmerling

c. *Ona prišila [NP pugovicu [NP rebenka]].11  she:nom sew:f   button:acc   child:gen Intended: ‘She sewed a button on the child’s clothes.’

It is reasonable to reduce the notion of quasi-possessivity to those situations where two different possessive constructions seem to be competing, as in (12a–b). (12) Russian a. ?On [vP grubym priemom [VP slomal [NP nogu Aršavina    he   rough:inst.m mode:inst    broke:m   leg:acc A:gen /ego nogu]]. /his leg:acc

‘He broke Arshavin’s leg/ his leg in a rough way.’

b. On [vP grubym priemom [VP slomal [NP Aršavinu he   rough:inst.m mode:inst   broke:m   A:dat /emu] [NP nogu]]. /him:dat   leg:acc

‘the same’, lit. ‘He broke the leg to Arshavin/to him.’

Both (12a) and (12b) are well-formed, although standard Russian favors (8b). In Russian, complex NPs with a special possessive marker/possessive genitive to some extent alternate with constructions of other types, though there are exceptions where complex NPs with a bare genitive NP cannot be used; cf. (12c) above and (13b) below. (13) Russian a. [NP probka [PP ot [NP butylki]]].    plug:nom    of   bottle:gen ‘the cork from the bottle.’ b. ??[NP probka [NP butylki]].   plug:nom   bottle:gen Intended: ‘the cork from the bottle.’

3.  PR and external Dative possessor In Baker 1988 the term “Possessor Raising” is reserved for the inversion of the arguments of a ditransitive verb, where the [+ Animate] argument takes the position

.  Russian (11c) is well-formed only in the meaning ‘X made use of Y’s button and sewed it on some place not necessarily related to Y’s clothes’. Such a reading is however unnatural and requires a special context.



Possessor Raising and Slavic clitics 

of Direct Object, as in English He gave flowers [to Kate] > he gave Kate flowers, while a construction where a possessive argument has been added to an intransitive sentence, as in Russian ty u menja molodec ‘you did well’, is called External Possessor Construction.12 This delimitation seems too rigid for two reasons. First, argument inversion in transitive sentences does not depend on possessivity. Second, the opposition of transitive vs intransitive verbs is not always clear-cut. In Norwegian, one and the same verb allows transitive (14a) and non-transitive uses (14b). Instances of prepositional government (14b) can also be analyzed as combinations of phrasal verbs with a post-verbal particle. Sentences where the Possessor is Case-marked with prepositionless Accusative and the Possessor is a PP pattern with PR, as in (14c): (14) Norwegian a. Hunden slikket [DP hånden hans]. dog-the licked:3sg   hand-the his. ‘The dog licked his hand.’ b. Hunden slikket [PP på [DP hånden hans]]. dog-the licked:3sg on   hand-the his. ‘The dog licked the hand on him’, lit. ‘the dog licked on his hand.’ c. Hunden slikket ham [PP på [DP hånden]]. dog-the licked:3sg him:obl.m     on   hand-the lit. ‘The Dog licked him on the hand. ’

Slavic languages provide a parallel to these Norwegian examples, since Slavic Aktionsart prefixes corresponding to Slavic prepositions have functions similar to those of Norwegian prepositives like på ‘on’ which behave as Aktionsart particles. However, this parallel is only partial. Modern Russian allows for structures without PR resembling (14a) and (14b), but not structures with PR like (14c). (15) Russian a. Sobaka lizala / liznula [NP ego ruku]. dog:nom licked:ipf.f / licked:pf.f   his hand:acc ‘The dog was licking/licked once his hand.’ b. Sobaka polizala [NP ego ruku]. dog:nom asp-licked:pf.f   his hand:acc ‘The dog licked his hand (several times).’ c. *Sobaka liznula ego [PP po ruke].  dog:nom licked:pf.f him:acc   on hand:loc Intended: ‘The dog licked his hand (once).’ .  The term ‘External Possessor’ (ru. vnešnij posessor) is used in Kibrik (2000, 2003:  308–318) who, contrary to Baker, extends it to transitive predicates as well.

 Anton Zimmerling

In (15a–b) the indeclinable possessive pronoun ego ‘3sg.m.poss’ is NP-internal, just as the indeclinable possessive pronoun hans ‘3sg.m.poss’ in Norwegian in examples (14a–b). In the Norwegian example (14b) one deals with a non-­transitive or semi-transitive predicate slikke på handen lit. ‘to lick on one’s hand’, where the element på can be analyzed both as a preposition belonging to a PP [PP på handen] or as part of the phrasal verb slikke på ‘to lick on smth’. ­Contrariwise, the Russian sentence (15b) is with the predicative polizala is transitive and does not involve any instance of preposition government. Finally, the Norwegian example (14c) demonstrates PR, where the oblique pronoun ham may be viewed as a raised external possessor, while the ungrammaticality of (15c) indicates that the construction with PR in this case is impossible.

4.  Slavic possessive clitics Slavic pronominal Dative possessive clitics have non-trivial features and may be used both as NP/DP-level clitics and as clause-level clitics across Slavic languages; cf. Franks & King 2000; Zaliznjak (2008: 35). Unambiguous DP-level clitics are attested in Bulgarian and Macedonian (cf. Franks et al. 2004) while the nature of the NP/DP-distinction in Polish remains controversial (cf. Rutkowski 2002). There are two competing views of Bulgarian Dative possessive clitics. Schürсks & Wunderlich 2003 basing on examples like (2) above, claim that PR is generalized in this language. On the contrary, Cinque & Krapova 2013 argue that PR is only possible in some sentences expressing alienable possession. Modern Russian has phrase-level Dative possessives but does not allow NP-level Dative possessives. Recall (1), repeated here as (16): (16) Russian a. Ja sebe ne vrag. I:nom ref.dat neg enemy:nom ‘I am not an enemy for myself.’ b. *Ja vstretil vraga sebe.  I:nom met:m enemy:acc ref.dat intended: ‘I met my own enemy.’

In (16) the element sebe is a reflexive clitic. Grigory Krejdlin (p.c.) points out that in (16a) the form sebe is no longer used as a reflexive marker and should be preferably analyzed as a discourse particle rather than as a pronoun. Irrespective of the validity of this claim, sebe in (16a) is an unambiguous clause-level element, and the ill-formedness of (16b) cannot be explained by the alleged pronoun/­particle distinction. It is worth mentioning that in Slavic languages which make use of pronominal



Possessor Raising and Slavic clitics 

Dative possessive clitics there is no obvious contrast in the syntax of non-agreeing possessive personal pronouns and non-agreeing possessive reflexives. Old Russian had both clause-level possessive clitics and NP-level possessive clitics attached to nominal heads. Clause-level dative possessive clitics normally merged in clausal 2P, while NP-level possessive clitics did not have a fixed position in a clause. Applying this criterion, one can easily establish that the second entry of 2p.Sg. ti in (17a) is a NP-level element while the first entry of ti is a clause-level element, since it takes clausal 2P and is not adjacent to any element of a NP. (17) Old Russian а. Čto vozdamŭ =ti [PP protivou what render:1sg you:dat.sg   for [NP blagodĕjaniju =ti]]?13    benefaction you:dat.sg ‘What can I render for your benefaction?’ b. [NP brata =ti Romana] Bogŭ pojalŭ.14   brother:acc you:dat.sg Roman:acc God took:prf.3sg.m ‘God took from you (your) brother Roman’ or ‘God took your brother Roman (from you).’

Meanwhile, sentences like (17b), where the Dative Possessive clitic splits the clause-initial NP and is placed after a noun, remain ambiguous since we lack reliable criteria to establish whether we deal with a clause-level element related to the predicate or with a NP-level element. Turning back to the Bulgarian data that motivated the contradicting syntactic accounts of PR and Cinque and Krapova’s hypothesis that Bulgarian only allows PR with alienable possession, I restate basic observations. First, Bulgarian blocks extraction out of PPs headed by a lexical preposition, as in (4) repeated here (in a modified notation) as (18). (18) Bg. *Az =ìi mislja [PP za [DP očite __i]]   I her:dat think:1sg   for   eyes-the Intended: ‘I think of her eyes.’

.  The example is from the Ipatyevskaja chronicle [1199], list 244. The first =ti stands after the verb vozdamŭ, not after the first phonetic word, wh-word čto, since the latter could act as an optional Barrier triggering late clitic placement, cf. Zaliznjak (2008: 55). .  The example is from the Ipatyevskaja chronicle [1180], list 217. All varieties of Old Russian allowed for inserting 2P clitics into fronted NPs, and in Old Novgorod Russian such clitic placement was obligatory, cf. Zaliznjak 2008; Zimmerling 2009; Zimmerling 2012.

 Anton Zimmerling

Second, in sentences like (3a) repeated here as (19a) the Dative clitic mi is a clauselevel element taking clausal 2P and not a raised DP-level possessive clitic, since such a derivation would involve extraction out of the PP [PP v [uxoto=mi]], in violation of island constraints, cf. Cinque & Krapova 2013 consequently, (19a) and (19b) have different syntactic structure and do not exhibit PR. (19) Bulgarian a. Toj mi= se= izkrjaska [PP v [DP uxoto]].15 he me:dat ref.acc shouted:3sg     in    ear-the ‘He shouted in my ear.’ b. Toj se= izkrjaska [PP v [DP uxoto =mi]]. he ref.acc shouted:3sg     in     ear-the me:dat ‘He shouted in my ear’.

I claim that (19a) and (19b) differ both syntactically and semantically, (19b) being a true possessive sentence containing a DP [DP uxoto=mi] and (19a) being a pseudopossessive sentence where the Dative clitic mi is a verbal argument related to the main predicate izkrjaska ‘shouted’. In other words, pairs of sentences like (19a–b) are examples of Possessive Shift, despite the morphological similarity of the possessive and the quasi-possessive Dative clitics. Cinque and Krapova seem to arrive at a similar conclusion, albeit in a different way. They argue that in spite of the fact that a clause-level dative clitic and a DP-level dative clitic appear to be the in same case on the surface, they still have different underlying Cases. In their opinion, in (19a) the clause-level clitic mi expresses underlying Dative Case and the role of Addressee/Benefactor, while in (19b) the DP-level clitic mi expresses underlying Genitive Case and the role of Possessor. This explanation in terms of underlying Cases is possible but it may turn out to be redundant if no decisive proof is found for the fact that Bulgarian has PR in other types of constructions. Indeed, the PR hypothesis can be applied to Bulgarian sentences that do not involve extraction out of a island PP and express alienable possession, as in (2) restated below as (20). (20) Bulgarian a. Tja nameri= li [DP užasni-te= si greški]? she found:3sg Q   horrible-the ref.dat mistakes ‘Did she find her horrible mistakes?’ b. Tja nameri= li= sii [DP užasni-te ti greški]? she found:3sg Q ref.dat   horrible-the mistakes ‘Did she find her horrible mistakes?’ .  For reasons specified above in the Footnote 3, I gloss Bulgarian pronominal and auxiliary clitics as enclitics. These forms are usually glossed as verb-adjacent proclitics: toj mi=se= izkrjaska.



Possessor Raising and Slavic clitics 

In (20b) the dative reflexive si occupies a position in which Bulgarian clause-level clitics form a clusterclausal 2P and clusters there with another clause-level clitic, the question particle li. In Slavic languages, only clause-level clitics can cluster, cf. Zimmerling 2012; Kosta & Zimmerling 2013,16 but this important fact unfortunately does not provide an independent basis for the verification of the Raising analysis in Bulgarian. Si in (20b) could equally well be a base-generated element merged in 2P or a raised element extracted out of a DP.17 Consequently, one cannot prove whether the alternation (20a–b) exhibits PR or Possessive Shift.

Conclusions Slavic data do not refute the hypothesis of Possessor Raising in Universal Grammar, but this hypothesis has a limited applicability in Slavic languages. The weak sides of the PR hypothesis are that it takes the mapping of syntax and possessive semantics to be iconic and that the rules/principles of extracting a NP/DP-level possessive operator out of the NP/DP are trivial. The analysis has shown that these assumptions are poorly justified. Universal semantic relations may correspond to similar yet not identical syntactic patterns. Constructions expressing alienable and inalienable possession often express other predicative meanings as well. Moreover, sentences with possessive predicates can be pragmatically equivalent to sentences with pseudo-possessive constructions. Slavic languages typically apply different case-marking for non-agreeing phrase-level possessive operators and non-­ agreeing clause-level possessive operators. A group of languages including Modern Russian lacks phrase-level possessive clitics: both Russian possessive datives and Russian possessive phrases of the type u + gen are clause-level operators. For this

.  The verification of this claim depends on the analysis of Bulgarian and Macedonian DPs containing combinations of a definite article and possessive clitics, cf. užasni-te=si greški in (20a), If we impose a condition that clusters are not occasional sequences of adjacent clitics but fixed combinations of clitics of the same level, combinations like Bulgarian [[užasni-te]=si ] greški ‘one’s awful mistakes’, Bulgarian [[mlada-ta]=mu] žena ‘his young wife’ can be ruled out,  since the definite article is likely merged at an earlier stage than the dative possessive pronoun. .  Unlike Serbo-Croatian, Burgenland Croatian, Slovene, Vojvodina Rusin, Czech, and Slovak, Bulgarian lacks Clitic Climbing. Anonymous reviewer points out that the absence of Clitic Climbing in Bulgarian is due to the fact that Bulgarian lacks infinitives and suggests that Possessor Raising and Clitic Climbing instantiate one and the same mechanism. This observation sets an interesting perspective, though Clitic Climbing (i.e. extraction of clitics out of embedded non-finite clauses) finds a much better empirical support than PR.

 Anton Zimmerling

group the PR hypothesis cannot be validated, unless we advance a special syntactic theory, for instance, postulate a mismatch of underlying and surface cases and rules transforming bare genitive NPs into preposition genitive phrases or dative NPs. As attested in modern South Slavic languages, Old Church Slavonic and Old Russian, dative pronominal and reflexive clitics are a special means of marking the possessive semantics. However, true NP/DP-level dative possessive clitics are found in an absolute minority of Slavic languages, notably Bulgarian, Macedonian and Old Church Slavonic, while clause-level dative possessive elements are a more common option. Most cases where the PR hypothesis has been proposed in previous Slavic studies do not conform to the definition of Raising as a syntactic operation, since the identity of structures with a clause-level or phrase-level possessive operator cannot be established. Such cases must be reanalyzed in terms of Possessive Shift, i.e. alternation of a true possessive construction with a NP/DP-level possessive element and its quasi-synonym, a pseudo-possessive construction with a case-marked verbal argument. The PR hypothesis is still possible for a subclass of Bulgarian constructions with alienable possession and a dative clitic, but the alternative hypothesis of Possessive Shift may be applied here as well, since there is no independent verification that the extraction of a possessive clitic has taken place. The most general explanation of the fact that Slavic languages for the most part block PR as a syntactic operation is that Slavic constructions with non-agreeing NP/DP-level dative possessive clitics seem to be secondary and less common than Slavic constructions with clause-level dative clitics. It is wrong to assume that a Slavic clausal clitic must/may be a raised NP/DP-level element just because the sentence has possessive semantics. A verification procedure is needed.

References Agbayani, Brian & Golston, Chris. 2010. Second-position is first-position. Wackernagel’s law and the role of clausal conjunction. Indogermanische Forschungen 115: 1–21. Aikhenvald, Alexandra Y. 1998. Warekena. In Handbook of Amazonian Languages, Vol. 4, ­Desmond C. Derbyshire & Geoffrey K. Pullum (eds), 225–440. Berlin: Mouton de Gruyter. Arutjunova, Nadejda D. & Širjaev, Evgenij N. 1983. Russkoe predloženie. Bytijnyj tip. Moscow: Nauka. Baker, Mark C. 1988. Incorporation: A Theory of Grammatical Function Changing. Chicago IL: The Chicago University Press. Benveniste, Emile. 1960. “Être” et “avoir” dans leurs fonctions linguistiques. Bulletin de la Société de linguistique de Paris 55(1): 113–134. Cinque, Guglielmo & Krapova, Iliyana. 2013. The case for genitive case in Bulgarian. In Studies in Generative Grammar. Nominal Constructions. Slavic and Beyond, Lilia Schürcks, Anastasia Giannakidou & Urtzi Etxeberria [Studies in Generative Grammar 116]. Berlin: Mouton de Gruyter.



Possessor Raising and Slavic clitics 

Dimitrova-Vulchanova, Mila. 1999. Clitics in the Slavic languages. In Clitics in the Languages of Europe, Henk van Riemsdijk (ed.), 83–121. Berlin: Mouton de Gruyter. Don-Won Lee. 2004. Possessor-raising in existential constructions. Studies in Generative Grammar 14(2): 235–242. Franks, Steven. 2008. Clitic placement, prosody and the Bulgarian verbal complex. Journal of Slavic Linguistics 16(1): 91–137. Franks, Steven, Junghanns, Uwe & Law, Paul. 2004. Pronominal Clitics in Slavic. Journal of Slavic Linguistics 12(1–2): 3–36. Franks, Steven & King, Tracey Holloway. 2000. A Handbook of Slavic Clitics. Oxford: OUP. Graščenkov, Pavel V. 2007. Tipologija possessivnyx konstrukcij. Voprosy jazykoznanija 3: 25–54. Herslund, Michael & Baron, Irène. 2001. Introduction: Dimensions of possession. In Dimensions of Possession [Typological Studies in Language 47], Irène Baron, Michael Herslund & Finn Sørensen (eds), 1–25. Amsterdam: John Benjamins. Kibrik, Aleksandr E. 2000. Vnešnij possessor kak rezul’tat rasščeplenija valentnostej. In Slovo v tekste i slovare, Sbornik statej k 75-letiju Ju. D. Apresjana, Leonid Lejbovic Iomdin & Leonid Petrovic Krysin (eds), 434–446. Moscow: Jazyki russkoj kultury. Kibrik, Aleksandr E. 2003. Konstanty i peremennye jazyka. Sankt-Petersburg: Aletheia. Krejdlin, Grigorij E. 1979. Služebnye slova v russkom jazyke (semantičeskie i sintaksičeskie aspekty ix izučenija. Ph.D. dissertation, Moscow. Kosta, Peter & Zimmerling, Anton. 2013. Slavic clitic systems in a typological perspective. In The Structure of Nominal Expressions. Slavic and Beyond [Studies in Generative Grammar 116], Lilia Schürcks, Anastasia Giannakidou & Urtzi Etxeberria (eds). Berlin: Mouton de Gruyter. Mel’čuk, Igor & Iordanskaja, Lidia 1995. *Glaza Maši golubye vs Glaza u Maši golubye: Two Russian constructions. In Russkij jazyk v modeli “Smysl Tekst”, Igor Mel’čuk, 135–164. Moscow: Jazyki russkoj kultury. Mišeska Tomić, Olga. 2004. The South Slavic pronominal clitics. Journal of Slavic linguistics 12(1–2): 213–248. Pennington, Josh J. 2010. Kombinovanje objekta adnominalnog posesivnog dativa s dopunama glagola u jednoj klauzi u bosansko-hrvatsko-srpsko-crnogorskom. Ms, University of Ohio. Pei-Jung Kuo & Yi-An-Lin 2008. Possessor raising, resumptive pronouns, and phases. Paper presented at The 39th annual meeting of the North East Linguistic Society (NELS39), Ithaca, November 7. Progovac, Ljiljana. 1996. Clitics in Serbian/Croatian: Comp as the second position. In Approaching Second: Second Position Clitics and Related Phenomena, Aaron L. Halpern & Arnold M. Zwicky (eds), 411–428. Stanford CA: CSLI. Raxilina, Ekaterina V. 2000. Kognitivnyj analiz predmetnyx imen: Semantika i sočetaemost’. ­Moscow: Russkie slovari. Russkaja grammatika. 1982. Russkaja grammatika, Vol. 2: Sintaksis, Natal’ja Yul’evna Švedova (ed.). Moscow: Nauka. Rutkowski, P. 2002. Noun/pronoun asymmetries: Evidence in support of the DP hypothesis in Polish. Jezikoslovje 3(1–2): 159–170. Seliverstova, Olga N. 1990. Kontrastivnaja sintaksičeskaja semantika. Moscow: Nauka. Szabolcsi, Anna. 1983. The possessor that ran away from home. The Linguistic Review 3: 89–102. Schürcks, Lilia & Wunderlich, Dieter. 2003. Determiner-possessor relation in the Bulgarian DP. In From NP to DP, Vol. 2: The Expression of Possession in Noun Phrases [Linguistik Aktuell/ Linguistics Today 56], Martine Coene & Yves D’hulst (eds), 121–139. Amsterdam: John Benjamins.

 Anton Zimmerling Zaliznjak, Andrej A. 2008. Drevnerusskie enklitiki. Moscow: Jazyki slavjanskoj kul’tury. Zimmerling, Anton. 2000. Obladat’ i byt’ rjadom. In Logičeskij analiz jazyka. Jazyki prostranstv, Nina D. Arutjunova & Irina B. Levontina (eds), 179–188. Moscow: Indrik. Zimmerling, Anton. 2006. Encoding strategies in word order: The evidence of Slavic languages. The 1st meeting of the Slavic Linguistic Society, Bloomington, 8–10 September. Zimmerling, Anton. 2008. Porjadok slov v slavyanskix, germanskix i romanskix jazykax. In Ot Imen k Faktam [Slavjano-germanskie issledovanija 3], Alexander A. Gugnin & Anton V. Zimmerling (eds), 165–239. Sankt-Peterburg: Aleteia. Zimmerling, Anton. 2009. ‘Klitiki v prostranstve drevnerusskogo jazyka’. Review of Zaliznjak (2008). Russkij jazyk v nauchnom osvesjchenii 1(17): 259–277. Zimmerling, Anton. 2012. Sistemy porjadka slov s klitikami v tipologičeskom aspekte. Voprosy jazykoznanija 4: 3–38. Zimmerling, Anton & Kosta, Peter. 2013. Slavic clitics: A typology. Language Typology and ­Universals (STUF) 66(2).

The Slavonic languages and the development of the antipassive marker Katarzyna Janic

Laboratoire Dynamique du Langage, Université Lumière Lyon 2, France This article deals with the development of dedicated antipassive markers in a crosslinguistic perspective, with a special attention given to Slavonic languages. Initially, this marker was associated exclusively with ergative languages in which it was treated as a valence reducing operator. Attached to the verbal form, it ditransitives a transitive construction without affecting the semantic content of a sentence. This led many scholars to insist on a simple dichotomy according to which ergative languages possess antipassive constructions due to the presence of a dedicated antipassive marker, whereas languages of accusative alignment are considered implicitly to be deprived of this type of operation and this is because of the lack of a specialized antipassive marker. Thus, the presence of a dedicated antipassive marker was treated as one of decisive criteria in the recognition of the antipassive in accusative languages. The recent and expanded crosslinguistic investigations reveal, however, that ergative languages present a whole range of variations concerning the antipassive marker. Among those ergative languages which developed a dedicated antipassive marker, i.e. a marker specialized in the antipassive function, in some of them this marker is also used in a middle domain. This means that in these languages the antipassive marker is polyfunctional, being related diachronically to other grammatical categories, mostly reflexivity. This article shows that a similar morphological correlation also exists in accusative languages, in particular in Russian. We argue that all Slavonic languages attest a dedicated antipassive marker that evolved from reflexivity by hand of middle domain. Keywords:  antipassive; reflexive; polysemy; valence change

1.  Introduction This article that subscribes to the typological approach deals with antipassive constructions. In particular, it investigates the morphosyntactic strategy, i.e. the  antipassive marker that is employed by a language to derive antipassives.

 Katarzyna Janic

A  large subset of crosslinguistically attested antipassive constructions is triggered by an antipassive marker. Languages differ, however, in terms of the nature of the antipassive marker. While some languages make use of a dedicated antipassive morpheme, others adopt a more common strategy, i.e. they employ a polyfunctional marker. It is also possible to come across with languages that attest more than one antipassive marker. Finally, the typological investigation in the development of antipassive markers points to its possible diachronic correlation with other grammatical categories. In this article, we will investigate antipassive markers building on a sample of genetically unrelated languages. We will concentrate both on ergative and accusative languages. A special attention will be given to Slavonic family. Our aim is, first, to reanalyze the notion of a dedicated antipassive marker, introduced initially in ergative languages, second, to describe syntactic and/or semantic properties exhibited by the a­ ntipassive marker to yield a new insight into the synchronic status of the ­antipassive marker.

2.  Definition of the antipassive construction A number of subsequent scholars have wrestled with the proper definition of the antipassive. Initially, the antipassive construction was associated exclusively with languages of ergative alignment: “The antipassive is a construction typical for ergative languages and occurs along with ergative construction as a morphosyntactic alternative for the same transitive proposition” (Cooreman 1993: 50). The essence of Cooreman’s definition is that it posits the semantic coherence between the antipassive and the corresponding transitive clause. The following example taken from West Greenlandic Eskimo, an ergative language, confirms this observation: (1) a. Jaaku‑p ujarak tigu‑a‑a. Jacob‑erg stone:abs take‑tr.ind‑3sg.erg/3sg.abs ‘Jacob took stone.’ b. Jaaku ujaraj‑mik tigu‑si‑vu‑q. Jacob:abs stone‑inst take‑ap‑intr.ind‑3sg.abs ‘Jacob took stone.’

(Bittner 1987: 200)

In terms of the morphosyntactic alternation, we can observe that in (1a), the transitive verb ‘take’ agrees both with ergative subject and absolutive object. In (1b), the same verb is marked with the antipassive marker ‑si and no longer agrees with the object that is now in oblique, i.e. instrumental case. Finally, the subject argument that was initially in ergative case, assigns the absolutive one. Thus, the antipassive alternation in West Greenlandic Eskimo involves both a change in the agreement pattern on the verb and in the morphological case marking.



The Slavonic languages and the development of the antipassive marker 

Dixon (1994: 146) provides a more detailed definition of the antipassive phenomenon. He lists a range of formal features typical for antipassive derivations in ergative languages: “(a) Antipassive applies to an underlying transitive clause and forms a derived intransitive, (b) the underlying Agent becomes Subject of the antipassive, (c) the underlying Object argument goes into a peripheral function, being marked by a non-core case, adposition, etc.; this argument can be omitted, although there is always the option of including it, (d) there is some explicit formal marking of an antipassive construction.” Note that unlike Cooreman (1993: 50) who suggests that in antipassives “the verb phrase may or may not be explicitly marked as intransitive”, Dixon 1994 argues in favour of the obligatory marking on antipassives. The opinion according to which in ergative languages the presence of verbal marker is not crucial is also supported by Palmer (1994) and Polinsky (2005). These scholars observe that in ergative alignment, antipassive constructions present some other formal characteristics in favour of the antipassive analysis. Building on the above definitions, the antipassive should be understood as a syntactic operation that detransitivizes transitive constructions. The change in the verbal valence is mostly triggered by an explicit marker due to which the object argument is either suppressed or demoted to the oblique position.

3.  Antipassive in accusative languages If the product of the antipassive operation is a derived construction in which the object argument loses its properties of a core argument, and if there is a requirement that this type of operation should be triggered by an explicit marker on a verb without affecting the semantic content of a sentence, it seems that languages with accusative system also attest this type of phenomenon: (2) Polish a. Chłopiec uchwycił klamk‑ę. boy:nom grasped:m door-handle‑acc ‘The boy grasped the door handle.’ b. Chłopiec uchwycił się klamk‑i. boy:nom grasped:m ap door-handle‑gen ‘The boy grasped the door handle.’

In (2b), the verb uchwycił is formally marked by the reflexive-based marker się due to the presence of which the object is demoted to the oblique position that in Polish is marked by the genitive case. The sentence does not get, however, a reflexive interpretation. This means that się performs a structural, but not semantic function. Both (2a) and (2b) refer to the semantically transitive event in which

 Katarzyna Janic

the subject referent performs the act of grasping on some other distinct entity. The semantic requirement that the antipassive and the construction from which it is derived should designate the same event is fulfilled. 4.  Distribution of the antipassive construction The analysis of antipassive constructions often raises the question about its ­distribution. The term antipassive was initially proposed by Silverstein (1976) with reference to ergative languages. For a long time, this type of construction was treated as a mirror image of the passive typically associated with accusative languages. Apart from Silverstein, also Dixon (1979, 1994) and Bok-Bennema (1991), among others, restricted exclusively the analysis of the antipassive to ergative languages. The recent investigation on the antipassive phenomenon, however, has extended the antipassive analysis also to accusative languages. Heath (1976), Givón (1984), Say (2005), Creissels (2006), Knjazev (2007), among others, claim that it is possible to find instances of antipassive in morphologically and syntactically accusative languages. Polinsky (2005), in her studies on the antipassive phenomenon, provides a whole list of accusative languages that attest antipassive alternations. Surprisingly, none of those belongs to the ­Slavonic family. It should also be stressed that sometimes the transitive/antipassive alternation is recognized in accusative languages, however, under different labels. For instance, Geniušienė (1987) terms the antipassive derivation ­absolutive, ­Lichtenberk (1991): depatientive, Haspelmath (2004): potential ­deobjective, Janko-Trinickaja (1962): active-objectless, Rivero and Sheppard (2003): accusative indefinites. Levin (1993) analyses the antipassive alternation under the section Characteristic Property of Agent Alternation. Regardless of the different labels, they all deal with the same type of a syntactic construction, i.e. the one in which the verbal valence is reduced by an explicit marking due to the presence of which Patient loses its properties of a core argument. 5.  The antipassive construction in accusative languages A discussion on the distribution of antipassives raises the following question: why was the antipassive not recognized in accusative languages? One of the reasons may be related to the question of visibility, i.e. the antipassive derivation tends to be more visible on the morphosyntactic level in ergative rather than in accusative languages. This has been particularly illustrated in (1), where the transitive/ antipassive alternation not only triggers a change in the case marking of the core



The Slavonic languages and the development of the antipassive marker 

arguments but also a change in the agreement pattern. However, this argument can be easily beaten off, since these are morphosyntactic consequences resulting from the general properties of ergative systems rather than distinctive features of antipassive constructions.1 Another reason why the antipassive was often refused in the syntactic description of accusative languages may be related to the question whether or not a language possesses a dedicated antipassive marker. Unlike accusative languages, those of ergative alignment are considered to develop a dedicated antipassive marker (e.g. West Greenlandic Eskimo). Although it would be difficult to argue against it, this observation needs to be specified a bit. First of all, it is legitimate to claim that in ergative languages, antipassive constructions are frequently triggered by a dedicated antipassive marker. However, it should be stressed that in some ergative languages in which antipassive constructions are triggered by the antipassive marker, this marker is not always limited to the antipassive derivation only. This means that we deal with the marker that is polyfunctional in nature, i.e. in addition to its initial reflexive function, it shows also the antipassive use (e.g. Australian languages from Pama-Nyungan family). Note that a similar correlation is observed in accusative languages, in particular in the Slavonic family. These languages employ a polyfunctional reflexive-based marker that specialized also in antipassive derivations. Yet some other ergative languages attest antipassives without explicit marking on a verb (e.g. Bezhedukh from Caucasian languages, Basque, Tongan from Oceanic languages, Inuktitut from Greenland and Canada). Even if it is frequent to come across with a dedicated antipassive marker in ergative languages, it should be stressed that such situation is far from being universal. Consequently, the argument according to which only ergative languages attest the antipassive and this is due to the presence of a dedicated antipassive marker can be easily called into question. 6.  On dedicated markers The crosslinguistic study shows that languages hardly ever possess a dedicated marker specialized in one particular function only. Among few examples, we can mention, for instance, the dedicated impersonal ending ‑no/‑to in Polish. In most .  In general, a language can be ergative in terms of its morphology, i.e. the case marking on the noun and in terms of the agreement on the verb. In a fully ergative language, a transitive construction is the one in which the subject and object arguments are marked by the ergative and absolutive case respectively, and both are indexed on the verb. Note, however, that sometimes languages may present a split-ergative pattern, i.e. they may be, for instance, ergative in its nominal morphology but accusative in the verb agreement.

 Katarzyna Janic

cases, such marker may present various degrees of specialization. For instance, impersonal constructions in French may involve the dedicated impersonal pronoun on; however, it is well-known that the same pronoun is also used in French as a 3rd singular personal pronoun. A dedicatedness of an antipassive marker has been particularly investigated by Polinsky (2005) who claims that this marker can often be related to other grammatical categories such as aspect/modality or detransitivizing operations (e.g. middle, reflexive). With respect to the middle/reflexive construction, one may treat the morphological correlation between reflexive and antipassive as a pure morphological coincidence. But taking into account that these two constructions display a strong semantic affinity and that in the diachronic approach the semantic affinity is considered to be a potential factor triggering a diachronic evolution of a marker, we argue in favour of the analysis according to which the morphological correlation between reflexivity and antipassive is not accidental but historically related and that in terms of the evolution it is the antipassive that developed from reflexivity. So what matters is not how much a dedicated marker is specialized, but whether or not a given language possesses such a marker is susceptible to perform a given function. 7.  Polyfunctional antipassive marker in ergative languages Even though the main focus of this article is on the development of the antipassive marker in Slavonic languages, since most previous work on transitivity and antipassive marking has built on or related it to ergativity, it makes sense to look first at the antipassive marking in ergative languages. Note in advance that in the syntactic description of ergative languages, once the antipassive construction is identified, there is a very strong tendency to argue that it is triggered by a dedicated antipassive marker, regardless of whether we deal with polyfunctional or monofunctional morpheme. Such situation is observed in particular in Chukchi: “In our sample, languages with a dedicated antipassive marker include (…) Chukchi (…)” (Polinsky 2005: 438). In fact, Chukchi attests two different antipassive markers: the prefix ena‑/ine‑ and the suffix ‑tku/‑tko. The choice of the marker as well as the possible semantic contrast between its different realisations is obscure. What is of interest, however, is that none of these affixes specialises exclusively in the antipassive derivation. The former one, for instance, can also be used as a 1sg object marker or to express reciprocity. But it is the latter that needs a special attention. As documented by Nedjalkov (2006), the suffix ‑tku/‑tko occupies a special position in the system of Chukchi grammar, due to its t­ ypologically distinctive p ­ olyfunctionality. Apart



The Slavonic languages and the development of the antipassive marker 

from the antipassive, it can also perform the reflexive, reciprocal, and anticausative function. It can also derive denominal verbs and be used to express iterative meaning on the verbs. Finally, the instances in which a given marker is used to express count/noncount meaning of nous are also attested. It can also function as a 1pl object marker. In this study, we are particularly interested in its morphological correlation with reciprocity. The reciprocal-antipassive polysemy of ‑tku is i­ llustrated in (3) and (4) respectively: (3) a. ətri qlawəl‑mel ukwet‑ə‑tku‑ɣʔet they:abs man‑as kiss‑ə‑rec‑aor.3pl ‘They kiss each other as men do.’ (4) a. ʔəttʔ‑e juu‑nin dog‑erg bite‑aor.3sg:3sg ‘The dog bit him.’ b. ʔəttʔ‑ən nə‑jɣu‑tku‑qin dog‑abs ipf‑bite‑ap‑3sg ‘The dog bites.’

(own glosses) (Nedjalkov 2006: 222)

Note that the suffix ‑tku, both in its reciprocal and antipassive use, detransitivizes a transitive construction. Once attached to the verb, it renders a transitive construction intransitive. Now, only the subject argument is indexed on the verb. Other formal characteristic of intransitivity involves a change in a case frame of the core arguments: the ergative subject is replaced now by the absolutive. Obviously, languages differ in how polyfunctional-like the antipassive marker is. Apart from the antipassive-reciprocal morphological correlation, languages may also make use of the antipassive marker to derive other middle operations. For instance, Australian languages from the Pama-Nyungan family use an antipassive marker that is diachronically related to reflexivity (Terrill 1997). The following examples illustrate the reflexive and antipassive uses of ‑:di marker in Yidiny, the ergative language of Pama-Nyungan family: (5) a. Yiɲdu:ŋ buɲa:‑ŋ mayi‑Ø buga‑ŋ. this‑erg woman‑erg vegetables‑abs eat‑prs ‘This woman is eating vegetables.’ b. Yiŋu‑Ø buɲa‑Ø buga‑:di‑ŋ. this‑abs woman‑abs eat‑ap‑prs ‘This woman is eating [something].’ (6) a. Wagu:da‑Ø bambi‑:di‑nu. man‑abs cover‑ap‑pst ‘The man has covered himself.’

(Foley & Van Valin 1984: 172–173)

 Katarzyna Janic

Data from Chukchi and Yidiny can be seen as part of a more general pattern whereby, the antipassive marker can be related to detransitivizing processes. This observation is confirmed by Polinsky (2005: 438) who claims that “Languages also make use of antipassive marker that are syncretic with other categories, thus functioning simultaneously as (for example) detransitivizers (…) markers”.

8.  The antipassive marker in accusative languages With regard to antipassive marking in accusative languages, a situation is less complex than in ergative languages. Apart from a few examples (e.g. Nahuatl from UtoAztecan family), most accusative languages to derive the antipassive, they use a polyfunctional marker that is either diachronically related to reflexivity (e.g. Slavonic languages and some Turkic languages such as Chuvash, Tuvan, Tatar) or to reciprocity (some Bantu, Turkic, and Oceanic languages). It is plausible that a more detailed analysis on valence changing operation in accusative languages would reveal more instances of languages with a dedicated antipassive marker. The aim of the following section is to investigate the antipassive marker in Slavonic languages. 8.1  Polyfunctional antipassive marker in Slavonic languages Slavonic languages belong to the Indo-European family. Like all members of this group, they did not develop a marker specialized only in the antipassive function. Instead, they refer to a highly polyfunctional reflexive-based marker SE. The reflexive and antipassive use of this morpheme is illustrated in (7) and (8) respectively. (7) a. On umy‑l‑sja.(Russian) he:nom wash:pf‑m.pst‑ref ‘He washed [himself].’ (Knjazev 2007: 680) (8) a. Sobaka kusaet malčik‑a. dog:nom bites boy‑acc ‘The dog bites a/the boy.’ b. Sobaka kusaet‑sja. dog:nom bites‑sja ‘The dog bites (is a biter)’.

(Say 2005: 426)

This type of antipassive derivation is peculiar to all Slavonic languages. In (8b), the verbal form kusaetsja is marked by a morpheme ‑sja that is claimed to behave like an antipassive marker of ergative languages. This means that it plays a s­ tructural but



The Slavonic languages and the development of the antipassive marker 

not semantic function. Indeed, attached to the verb, it detransitivizes a ­transitive construction without affecting its semantic content. Like the corresponding transitive clause, the antipassive designates a two-participant event in which the second argument is syntactically constrained, though semantically present. Indeed, a native speaker of Russian conceptualizes a given sentence as being semantically transitive in which a dog participant performs the act of biting on some other distinct entity. In Russian, like in many other Slavonic languages, there is an association of this construction with generic interpretation. The phrase can be paraphrased as “the dog is in a habit/has a tendency to bite [others]”. The examples of the antipassive use of ‑sja can be expanded by (9) and (10): (9) Etot byk bodaet‑sja. this:nom.m bull:nom butt:3sg.prs‑ap ‘This bull butts.’ (10) Petux kljuet‑sja. cock:nom peck:3sg.prs‑ap ‘The cock pecks.’

(Russian) (own glosses) (Nedjalkov 2006: 19) (own glosses) (Nedjalkov 2007: 297)

It may happen that the reflexive-antipassive polysemy is realized on the same verb. Depending on the context, such sentence either denotes a reflexive event in which się is used anaphorically pointing back to a previously mentioned referent, or as a valence reducing operator that blocks the overt realisation of the object argument: (11) Nie chlap się. neg splash:ipf.2sg ref/ap ‘Stop splashing yourself!’ or ‘Stop splashing [me]!’

(Polish)

In Slavonic languages the same morpheme is used to cover both reflexive and reciprocal function. Note, however, that different sources gloss this element ­differently. In Russian, for instance, ‑sja is often labeled as ‘reflexive/reciprocal marker’. Some scholars also use the term ‘reflexivity’ as shorthand to cover both the reflexive and reciprocal function. Yet, some others gloss this form as reflexive or reciprocal depending on the function it performs in a given sentence. Regardless of the varying terminology, it remains unquestionable that in ­Slavonic languages the SE morpheme is diachronically associated with the reflexive (and reciprocal) function. Taking into account that in those languages, the same morpheme that is used to mark the reflexive-antipassive polysemy is also used to express reciprocity, the question arises whether or not it is also possible to come across with reciprocal-antipassive polysemy. In fact, Knjazev (2007: 136) observes

 Katarzyna Janic

with regard to Russian language, that ‘almost all Russian reflexive ­reciprocals allow the absolutive [antipassive] use as well.’ To illustrate this situation, this author provides a following example: (12) a. Posmotr‑i, dv‑e korov‑y boda‑jut‑sja. look:pf‑imp two‑nom.f cow‑nom.pl butt:ipf‑3pl.prs‑ref ‘Look, two cows are butting each other.’ b. Bud’ ostorožen, korov‑y boda‑jut‑sja. be:imp careful:sg.m cow‑nom.pl butt:ipf‑3pl.prs‑ref ‘Be careful, cows butt.’ (Knjazev 2007: 681)

Example (12) shows that the same derivative bodajutsja can be used both in the reciprocal, (12a), and in the antipassive construction, (12b). Contrary to its reciprocal use in which ‑sja performs both semantic and structural function, ‑sja in its antipassive use is basically limited to the structural function only, i.e. it detransitivizes a transitive construction without affecting the semantic content of the sentence. Note that unlike reciprocal constructions, antipassives do not impose any constraint on the plurality of the subject referent. Building on the examples ­provided from ­Slavonic languages, the antipassive may appear both with a singular and plural subject. 8.2  Dedicated antipassive marker in Nahuatl Nahuatl, a langue belonging to a branch of the Uto-Aztecan family, is highly significant for our discussion for two reasons. Firstly, as in the case of Slavonic languages, it manifests an accusative alignment, secondly, to derive antipassive constructions, it employs a marker specialised in the antipassive function only. As pointed out by Nouguier-Voisin (2005), Nahuatl developed two markers specialized exclusively in this function. The prefix ‑tla demotes an indefinite inanimate object, (13), whereas ‑te denotes an indefinite animate object, (13b). In both instances, the object argument loses its properties of a core argument, i.e. it is removed from the clause which renders it intransitive: (13) a. Ø‑tla‑cua (he)‑ap1‑eat ‘He eats [something].’ b. Ø‑te‑notza. (he)‑ap2‑call ‘He calls [somebody].’

(Launey 1994: 48)

It should be stressed that contrary to Slavonic languages, the antipassive ­markers ‑tla and ‑te are diachronically related neither to reflexivity nor to reciprocity. Nahuatl possesses a separate marker ‑mo that initially performs the reflexive f­unction ­specialised in the middle domain. The following example illustrates its reflexive use:



The Slavonic languages and the development of the antipassive marker 

(14) Ø‑mo‑tta‑c in Quetzalcohua. (he)‑mid‑see‑tam def Quetzalcohua ‘Quetzalcohua sees himself.’

(Launey 1994: 174)

8.3  The nature of the middle Turning now to a general issue of a middle marker in Indo-European languages, according to Kemmer (1993), the middle marker is said to be spread steadily over a very broad set of uses. Regardless of their exact nature, it is unquestionable that they are all related diachronically, having one common ancestor, i.e. reflexivity. The semantic extension of a reflexive marker into middle domain results from a grammaticalization process involving both semantic and formal aspects of grammaticalization. In terms of the semantic change, a reflexive morpheme SE is argued to lose its nominal semantic content acquiring a new and different semantic function. This means that at some point of its development, it started functioning as a verbal marker of different types of events: “from a marker with relatively concrete, referential/nominal function (a reflexive pronoun), it acquired a semantic function which was verbal, i.e. event-centered rather than object-centered, i.e. middle rather than reflexive” (Kemmer 1993: 161). With respect to the formal aspect of grammaticalization, SE gave up its pronominal properties and developed into an operator on the verbal valency. Following Kemmer (1993), we can consider SE in Slavonic languages to be a middle marker in its late stage of development. This is due to the fact that among different semantic functions, e.g. autocausative, decausative so on, this marker has also specialized in syntactic domain deriving the construction such as passive, antipassive or impersonal where it is generally assumed to play only a s­tructural function. So, it extended its range of uses to those in which it has become semantically meaningless. This observation remains in agreement with Kemmer (1993: 236) who claims that “Even if the erstwhile a middle marker still has productive uses in a language in the late stages of middle development, as might be the case if it has been reinterpreted as a marker of a specific function (e.g. the Russian ‘propensative’ [antipassive] use if the middle marker ‑sja), it can in many cases no longer be considered a middle marker but a marker of some other semantic/pragmatic category”.

9.  Conclusion This article deals with dedicated antipassive markers in a crosslinguistic perspective, with a special attention given to Slavonic languages. As mentioned as the beginning of this paper, the question on dedicated antipassive markers is

 Katarzyna Janic

i­ ntriguing and complex, and the expanded typological survey calls for its clarification. Introduced initially in the syntactic description of ergative languages, this term was used with reference to markers that trigger the antipassive operation, regardless of its degree of specialization in the antipassive function. It was shown, however, that whereas some ergative languages, e.g. West Greenlandic Eskimo, indeed developed a marker specialized in the antipassive function only, some others, e.g. Chukchi, Yidiny, employ (a) polyfunctional morpheme(s). Furthermore, the presence of a dedicated antipassive marker was often treated as one of decisive criteria in the recognition of the antipassive in accusative languages. As a result, many scholars argue in favour of an analysis according to which ergative languages possess antipassive constructions due to the presence of a dedicated antipassive marker, whereas languages of accusative alignment are said to be deprived of this type of operation and this is because of the lack of a specialized antipassive marker. But the fact that the same morphological correlation between reflexivity and antipassive that was observed in Yidiny, also exists in accusative languages, in particular in Slavonic languages, suggests that the antipassive should not be ruled out in accusative languages due to the lack of a dedicated antipassive marker. Following Terrill (1997), who insists on the diachronic link between reflexivity and antipassive in Australian Pama-Nyungan family, we consider that in Slavonic languages, reflexive marker also evolved into antipassive diachronically by hand of middle domain, not mentioning those accusative languages (for instance Nahuatl) that indeed developed ­dedicated antipassive morphology.

Abbreviations abs = absolutive acc = accusative aor = aorist ap = antipassive def = definitive erg = ergative f = feminine gen = genitive ind = indicative inst = instrumental intr = intransitive imp = imperative ipf = imperfective

m = masculine mid = middle neg = negation nom = nominative pf = perfective pl = plural prs = present pst = past rec = reciprocal ref = reflexivity sg = singular tam = time, aspect, mode tr = transitive



The Slavonic languages and the development of the antipassive marker 

References Bittner, Maria. 1987. On the semantics of the Greenlandic antipassive and related constructions. International Journal of American Linguistics 53: 194–231. Bok-Bennema, Reineke. 1991. Case and Agreement in Inuit. Berlin: Mouton de Gruyter. Cooreman, Ann. 1993. A functional typology of antipassives. In Voice: Form and Function [Typological Studies in Language 27], Barbara A. Fox & Paul J. Hopper (eds), 49–88. Amsterdam: John Benjamins. Creissels, Denis. 2006. Syntaxe générale: Une introduction typologique, Tome 2: La phrase. Paris: Hermes Science–Lavoisier. Dixon, Robert M.W. 1979. Ergativity. Language 55(1): 59–138. Dixon, Robert M.W. 1994. Ergativity. Cambridge: CUP. Foley, William A. & Van Valin, Robert D. Jr. 1984. Functional Syntax and Universal Grammar. Cambridge: CUP. Geniušienė, Ema. 1987. The Typology of Reflexives. Berlin: Mouton de Gruyter. Givón, Talmy. 1984. Syntax: An Introduction, Vol. 1. Amsterdam: John Benjamins. Haspelmath, Martin. 2004. Valence change. In Morphology: A Handbook on Inflection and Word Formation, Geert Booij, Christian Lehmann & Joachim Mugdan (eds), 1130–1145. Heath, Jeffrey. 1976. Antipassivization: A functional typology. In Proceedings of the Second Annual Meeting of Berkeley Linguistics Society, Henry Thompson, Kenneth Whistler, Vicki Edge, Jeri J. Jaeger, Ronya Javkin, Miriam Petruck, Christopher Smeall & Robert D. Van Valin Jr. (eds), 202–211. Berkeleyo CA: University of California. Janko-Trinickaja, Nadija A. 1962. Vozvratnye glagoly v sovremennom russkom jazyke. Moskva: Nauka. Kemmer, Suzanne. 1993. The Middle Voice [Typological Studies in Language 23]. Amsterdam: John Benjamins. Knjazev, Jurij. 2007. Reciprocal Constructions in Russian. In Nedjalkov (ed.), 673–708. Launey, Michel. 1994. Une grammaire omniprédicative: Essai sur la morphosyntaxe du nahuatl classique. Paris: CNRS Éditions. Levin, Beth. 1993. English Verb Classes and Alternations: A Preliminary Investigation. Chicago IL: University of Chicago Press. Lichtenberk, Frantisek. 1991. Reciprocals and depatientives in To’aba’ita. In Currents in Pacific Linguistics: Papers on Austronesian Languages and Ethnolinguistics in honour of George W. Grace, Robert Blust (ed.), 171–183. Canberra: Australian National University. Nedjalkov, Vladimir. 2006. Chukchi reciprocals (with an appendix on Koryak and Itelmen). In Voice and Grammatical Relations: In honor of Masayoshi Shibatani [Typological Studies in Language 65], Tasaku Tsunoda & Taro Kageyama (eds), 217–248. Amsterdam: John Benjamins. Nedjalkov, Vladimir. 2007. Reciprocal Constructions [Typological Studies in Language 71]. Amsterdam: John Benjamins. Nouguier-Voisin, Sylvie. 2005. Antipassif et langues accusatives. In Linguistique typologique, ­Gilbert Lazard & Claire Moyse-Faurie (eds), 193–206. Villeneuve d’Ascq: Presses universitaires du Septentrion. Palmer, Frank R. 1994. Grammatical Roles and Relations. Cambridge: CUP. Polinsky, Maria. 2005. Antipassive constructions. In The World Atlas Of Language Structures, Martin Haspelmath, Matthew S. Dryer, David Gil & Bernard Comrie (eds), 438–439. Oxford: OUP.

 Katarzyna Janic Rivero, María L. & Sheppard, Milena. 2003. Indefinite reflexive clitics in Slavic: Polish and ­Slovenian. Natural Language and Linguistic Theory 21(1): 89–155. Say, Sergey. 2005. The pragmatic motivation of antipassive in Russian. In Pragmatics Today, Piotr Cap (ed.), 421–440. Frankfurt: Peter Lang. Silverstein, Michael. 1976. Hierarchy of features and ergativity. In Grammatical Categories in Australian Languages, Robert M.W. Dixon (ed.), 112–171. Canberra: Australian Institute of Aboriginal Studies. Terrill, Angela. 1997. The development of antipassive constructions in Australian languages. Australian Journal of Linguistics 17: 71–88.

Clitic SE in Romance and Slavonic revisited* Marijana Marelj & Eric Reuland

UiL OTS, Utrecht University, The Netherlands

In Slavonic and Romance languages there is a range of constructions that feature what is referred to as the “the reflexive clitic” se/si/się. Adhering to the law of parsimony, we propose a unified analysis of the clitic across different reflexive-like and passive-like constructions. We put forth generally unnoticed evidence that the reflexive clitic construction in Slavonic languages allows two readings, and derive both the pure-reflexive and the “near-reflexive” readings. Finally, we show that within the Slavonic languages there is an array of intriguing differences across the constructions under consideration that splits the group into South/West Slavonic, on the one hand, and East Slavonic, on the other hand. We show that all the relevant differences can be derived from low-level differences in the functional system. Keywords:  reflexivization; proxy-reading; pure-reflexive reading; non-reflexive uses of SECL; ambiguous X0/XP elements

1.  SE-Puzzles In Slavonic and Romance languages, there is a whole range of constructions that feature what is referred to as “the reflexive clitic” se. Some of the uses of SECL (the list is by no means exhaustive) are given below for Serbo-Croatian (SC):

*  For comments and useful discussions we are grateful to Željko Bošković, Denis Delfitto, Alexis Dimitriadis, Luca Ducceschi, Martin Everaert, Steven Franks, Dagmar Schadler, Anna Volkova, the two anonymous reviewers of SLS 2011 as well as the audiences of 2010 UiL OTS Interface Meetings, the 2011 Mayfest of the Department of Linguistics of the University of Maryland, and the 2011 6th Meeting of the Slavic Linguistics Society, in Aix-en-Provence. Of course, we take full responsibility for all remaining errors. This work was supported by grants from the Netherlands Organization for Scientific Research – NWO (project numbers 275-70021 and 360-70-330), which we hereby gratefully acknowlege.

 Marijana Marelj & Eric Reuland

(1) Serbo-Croatian a. Maks se kupaReflexive Max:nom SECL bath:3sg ‘Max bathes.’ b. Maks i Petar se ljube Max:nom and Peter:nom SECL kiss:3pl ‘Max and Peter kiss each other.’

Reciprocal

c. Kuća se gradiPassive house:nom SECL build:3sg ‘The house is being built.’ d. Ovde se puno radiImpersonal here SECL a-lot work:3sg ‘[One/They/We] work a lot here.’

Even a cursory examination of the examples in (1) reveals that these constructions have very little in common, either semantically or syntactically. They all, however, share the presence of the clitic.1 The question that arises is whether the presence of the clitic is a mere coincidence or there is something deeper at stake here. The individual members of the SECL-construction family are no less i­ ntriguing. The reflexive SECL-construction is an excellent example to illustrate this. Though typically discussed in reference with SELF-anaphors (e.g. sebe in Serbo-Croatian and se stesso in Italian), clitics in Romance and Slavonic intriguingly also allow a near-reflexive reading, the so-called proxy reading. The existence of proxyreadings was for the first time posed as a puzzle for binding theory by Jackendoff (1992). He observed that expressions such as English himself in (2) can be interpreted not only as the person Ringo, but also as Ringo’s statue (Jackendoff 1992).

(2) (Upon a visit in a wax museum:) All of a sudden Ringo started ­undressing himself.

This rarely acknowledged fact about Romance and Slavonic languages illustrated in (3) is one of the central points in our discussion. (3) Serbo-Croatian a. Marko je sebe pokrio na fotografiji. Marko:nom aux self covered:m on photograph:loc OK‘Marko covered (Image of) Marko.’

.  Since we will be generalizing over different languages that have a highly similar form of the clitic, we adopt the following terminological convention: an italicized lower case form (such as se or si) refers to a particular lexical item, the form in capitals (SECL) to the class of items.



Clitic SE in Romance and Slavonic revisited 

b. Marko se pokrio na fotografiji. Marko:nom SECL covered:m on photograph:loc OK‘Marko covered (Image of) Marko.’

Moreover, such proxy-readings are not available in all Slavonic languages. In ­Russian, for instance, the reflexive construction cannot give rise to the proxy reading. (4) Ru. Marko zakrylsja na fotografii. Marko:nom covered:m on photograph:loc *Intended: ‘Marko covered Image of Marko.’

The goal of this paper is to show that it is possible to account for the presence of the clitic in all these constructions in a uniform way. We further provide an account for the availability of a proxy reading in Romance and the majority of Slavonic languages, while also accounting for its ban in languages like Russian. The paper is organized as follows. In Section 2, we examine the role, nature, and the categorial status of the clitic. This examination allows us to propose a ­unified account of the clitic, for both its reflexive (Section 3) and passive uses (­Section 4). Section 5 sums up our main findings. 2.  Clearing the ground: SECL in Romance and Slavonic 2.1  The role of SECL We argue that what the verb forms in (1) intuitively share is that they are all in some sense reductions of the same, full-fledged transitive verb form. Following Marelj (2004), we argue that given the Lexicon Uniformity Principle of Reinhart (Reinhart 2002, et seq.), the common denominator of the constructions in (14) is that they are derived by application of some operation that affects the arity (valence) of the predicate.2 In all cases the basic entry is tampered with and changing its valence also requires case-feature checking. Under this view it becomes possible to explore whether se has a uniform role to play across these different derivations. Clearly, a proposal that accounts for an array of constructions but postulates a single SECL is preferred to the one that ­postulates a multiplicity of SECL-entities.3 .  See Babby (2010) for some interesting observations. .  As a Reviewer correctly points out, this has been one of the focal points of research on the topic of -sja/SE, going back to at least Chvany (1975) and Babby and Brecht (1975), both calling for a uniform account of -sja.

 Marijana Marelj & Eric Reuland

2.2  The nature of SECL 2.2.1  How reflexive is the reflexive clitic? Assuming that there is only one SECL, we must agree that SECL is not reflexive, since there are only two constructions in (1) that could be referred to as “anaphoric”. Bearing in mind the lex parsimoniae, it is clear that we need to divorce reflexivity from any “anaphoric” properties of the reflexive clitic. As shown in Reuland (2001, 2011) [± pronominal] and [± anaphoric] are not primitive features of natural language. Hence, SECL cannot be endowed with some [+ anaphor] feature. It is not inherently reflexive or anaphoric; such properties are epiphenomenal (see Marelj & Reuland 2011 for elaboration and references). Rather, its behavior should be determined by the way its morpho-syntactic features interact with its syntactic environment. 2.2.2  Relevance of phi-features The fact that SECL in Romance and Slavonic is underspecified in terms of phi-­ features is sometimes taken to entail that it is an anaphor. As shown by the ­examples in (1), SECL is not intrinsically an anaphor. To the extent in which SECLtype clitics show the behavior of anaphors in some environment, this must reflect the way in which this underspecification interacts with the environment. SECL will enter an Agree relation with the first available source of phi-features (finite AGR), and hence match in features with the latter. Thus, though fully underspecified for phi-features in Slavonic, the clitic is not inert, i.e. insensitive to the environment it finds itself in (see Marelj & Reuland 2011 for elaboration). In a nutshell, following Marelj (2004), we argue that its feature deficiency allows SECL to be used as a member of the “mop up squad”; a functional element that can be merged and mop up offending features that need to be taken care of in syntax. However, this does not entail that it must be always used in this fashion. In fact, as we demonstrate shortly, another use is possible as well. In what follows, for ease of exposition, we will group the SECL-constructions from (1) under the labels “reflexive” and “passive-like”. In the reflexive-like constructions, all the thematic roles are realized in syntax, whereas in the passive-like ones, one of the roles is saturated and the variable is bound either existentially or generically. Thus, the difference between passives and middles, for instance, has nothing to do with the role of SECL in these constructions (see Marelj 2004 for discussion and elaboration). 2.3  Categorial Status of SECL Chomsky (1995) allows for the existence of elements whose status is ambiguous in that they are simultaneously X0s and XPs and suggests clitics as possible ­candidates. Starting from Chomsky’s suggestion, Bošković (1997) builds up an



Clitic SE in Romance and Slavonic revisited 

account of c­ litics as non-branching elements with ambiguous XP/X0 status. In what follows, we will develop a proposal along these lines for SECL. Recall that we argue that there is one element SECL. The variation in the roles it plays is determined by its locus of merger. We propose the following structure for those uses of the SECL, when the clitic is first merged in a non-argumental position with the effect of being used as a mop-up element where the source position of the clitic/locus of its merger is the K(ase) P(hrase) (14). In such cases, SECL is first merged in the specifier position of a KP, where it absorbs an offending ACC case. We argue that if acc is not checked, the derivation will crash (i.e. we assume that the Inverse Case Filter – see Bošković 1997 and Bošković & Lasnik 2007 – holds). (5)

KP

Spec se

K′ K

VP

The other possibility is for SECL to be merged with the verb as the internal argument and subsequently to move higher up in the structure, checking Case. This is what underlies the use of SECL in derivations that give rise to proxy readings.4 Importantly, under both scenarios, as an ambiguous element, SECL can either move to an XP or an X0 position. But once head-adjoined, the clitic can no longer move to an XP position (though it can move further with the relevant head, of course).5 3.  The role of SECL in the derivation of reflexives 3.1  First merging SECL in an argument position Though typically discussed with reference to SELF-anaphors (recall (2)), proxyreadings are not limited to reflexives, they also occur with bound pronominals: (6) Grisham claims that he is even more suspenseful in Swahili.(Safir 2004)

.  This proposal that SECL merges as an argument is our point of departure from much of the current literature with the exception of Labelle (2008) (for a detailed elaboration of this point, see Marelj and Reuland (2011)). .  For an insightful discussion about how phi-features can be related to the categorial status of anaphors, see Franks (this volume).

 Marijana Marelj & Eric Reuland

A proxy-interpretation is not always available, however. If undressing himself in (7) is replaced by the undressing without an object, the proxy-interpretation is not there.6 (7) All of a sudden, Ringo started undressing.

(only Ringo)

Intriguingly enough, unlike in English, in Slavonic and Romance language, the proxy-readings are not restricted to SELF-anaphors; the morphologically simple clitic allows proxy-readings as well. Labelle (2008) and Reuland (2008, 2011) note that proxy-readings are available in French constructions with clitic se. In ­Serbo-Croatian for some speakers proxy-readings are not generally available in constructions with se, rather sebe is required. Crucially, however, there are instances (8) in which the proxy reading is fully available at least, for some speakers.7 (8) SC Iznenada se ugledao na video zapisu suddenly SECL saw:m on tape:loc ‘He suddenly saw himself on the tape.’

(himself=his visual image)

Notice that the very fact that proxy-readings are available with the clitic entails that these constructions must be semantically transitive: there must be two distinct variables, as the notion of the “open predicate” from Sells et al. (1987) (9)). (9) lx.ly [R (x, y)]

(2 semantic arguments)

Semantically, then, there is no difference between the clitic se and the other ­pronominal clitics, apart from the fact that due to its phi-featural underspecification, the clitic must necessarily enter into an AGREE relation with the nearest source of values for its features, which happens to be the subject, which means that it gets co-valued with the subject of the clause. This shows two things: (i) clitic constructions must allow for an independent object argument to be projected; (ii) it must be a relatively superficial factor that determines the availability of proxy-readings.

.  Though true, simply stating that English does not have SE anaphors is not enough here. As illustrated in (i), the Dutch counterpart of (7) contains zich, but has the same interpretative possibilities as the English example i.e. it does not allow a proxy interpretation. (i) Plotseling begon Ringo zich te ontkleden (only Ringo) Suddenly began:3sg Ringo ZICH to undress:3sg ‘All of a sudden Ringo started undressing.’ .  The availability of proxy readings in constructions with SECL seems to vary by speakers and languages. In Czech, on the other hand, proxy-readings are allowed on a much freer basis (see Reuland & Winter 2009). Italian speakers also have mixed judgments, and the same holds true for Spanish (see Marelj & Reuland 2011 for discussion and references).



Clitic SE in Romance and Slavonic revisited 

3.2  The Analysis: Clitics and proxy-readings We start by exploring the conditions a clitic system must meet in order for a proxyinterpretation to be available, as was illustrated in (3), for instance. The analysis we propose is inspired by Baauw and Delfitto (2005, henceforth B&D), and subsequent discussion with Denis Delfitto. We adopt what we see as the essence of this proposal (although, various details in our analysis d ­ iffer). We assume that B&D’s position that pronominal clitics are special and unlike tonic pronouns encode λ-abstraction, should be somewhat qualified. Whereas λ-expressions in general do not restrict the arguments they take in terms of phifeatures, the λ-expressions resulting from clitics typically do. Rather than saturate the argument position, the operator shares with λ-expressions that it needs an argument for the expression to be valued. In fact, we claim that this follows from the conception of clitics as X0/XP. The crucial difference between SECL and other pronominal clitics is that SECL is deficient in phi-features. Hence, it enters into an Agree relation with the argument in [Spec, TP], mediated by the T-system. As discussed in Reuland (2001, 2011), copying/overwriting feature bundles by Agree is the syntactic mechanism for encoding binding dependencies. Just like Internal Merge/Remerge creates different occurrences of the same element, copying/ overwriting values by Agree creates different occurrences of the same feature bundle, subsequently yielding different occurrences of the same variable in logical syntax. In line with our analysis, once the clitic is in A′-position, it is interpreted as an operator. The probe-goal relation between T and the clitic results in sharing of all the features concerned. In short, the phi-features on Marko, and se end up being occurrences of the same morpho-syntactic object (10) (see Marelj & Reuland 2011 for elaboration and references). Note further, that due to its categorial status, the dependency between the clitic and its antecedent does not form an A-chain (10). Though the subject – Marko – as well as the lower se2 – are in A-positions the intermediate se1 is an X0 in an A′-position. This means that there is no uniform 〈Marko-se〉 A-chain (with ‘uniformity’ understood as in Chomsky 1995). (10) Marko…… T se1 v ……. …V (se2) ñ ñ ñ XP/A X0 /A′ XP/A

Consequently, for the interpretation system the seX 0-seXP chain will constitute an independent operator-argument expression in line with the general interpretation strategy for clitics. Thus, the clitic reflexive maintains the syntactic arity of the predicate it is an argument of. The clitic construction remains semantically and syntactically transitive. This result immediately explains the availability of proxyreadings with reflexive clitics.

 Marijana Marelj & Eric Reuland

As a final point, let us go back to the seemingly puzzling contrast between South and West Slavic, on the one hand and Russian on the other hand with respect to the availability of proxy-readings. Recall, whereas in Serbo-Croatian reflexive se-construction can give rise to proxy-readings, their Russian counterparts can give rise to pure-reflexive reading only. We argue that morphosyntactic properties of these functional elements are relevant, to the extent that the categorial status of an element determines its behavior. Recall that to account for the proxy-readings, we capitalized on the fact that the clitic has an ambiguous X0/XP status. Note now that Russian -sja is an affix, categorially an unambiguous X0. As such it simply cannot meet the conditions necessary for the proxy reading to arise. Hence, the (dis)allowance of proxy-­ readings can be derived form a low-level difference in the functional systems of different Slavonic languages. To sum up, by restricting ourselves to the tools of narrow syntax (i.e. Merge, Agree), and adopting an independently motivated interpretation rule for clitics, we derive the properties of the reflexive clitic construction. Moreover, the analysis is entirely based on formal properties of the clitic SE, namely its feature deficiency, which makes it an active goal for probing, and its ambiguous X0/XP status. This allows a uniform treatment of SECL across the derivations it appears in, as necessitated by data in Romance and Slavonic. 3.3  First-merging SECL in a non-argument position As we saw, nothing prevents the clitic from being merged in the object position, but nothing requires this either. Being a clitic, it can also merge in the functional domain, where it can check Case, or absorb a theta-role (see Section 3.3 for discussion of the notion of absorption). In the present section we explore what happens if the clitic is first-merged in a position where it just checks/absorbs ACC (see also Franks 1995 for an analysis where the “mopping up” function is the sole function of the reflexive clitic). This must be the situation in cases where proxy-readings do not arise. (13) exemplifies Serbo-Croatian, but carries over to parallel cases in other clitic Slavonic languages as well as Romance languages. Evidence discussed extensively in Marelj & Reuland (2011) shows that Romance languages and clitic ­Slavonic languages exhibit lexical reflexivization (just like e.g. Dutch, English, or Russian). In short, unlike their semantically transitive counterparts, inherently reflexive predicates undergo a lexical operation that reduces the valence of the predicate. Reinhart (2002) and Reinhart & Siloni (2005) formalize this kind of reduction as Bundling (11). The Bundling operation reduces the internal argument of the predicate that it operates on, and bundles the internal role (theme) and the e­ xternal role (agent) into a composite agent-theme role.



Clitic SE in Romance and Slavonic revisited 

(11) Internal Reduction/Bundling (Reinhart 2002): a. Vacc (θ1, θ2) à Rs(V) (θ1,2) (where θ1,2 – also written as [θ1 – θ2] stands for the Bundling of θ1 and θ2 b. Rs (V)(θ1,2) ßà θ1,2 (λx (V (x,x))) c. shaveacc (θ1, θ2): Lucie shaved him. d. Rs(shave)( θ1,2): Max shaved.

As indicated in (11c), in their canonical transitive use, verbs like wash, or shave have an Accusative (ACC) Case feature, requiring them to check ACC Case on their object. In English, the Bundling operation fully removes this Case feature, and the lexical reflexive form appears without an object. In languages such as Dutch, the Bundling operation leaves a Case residue, as in (12), which must be checked and eliminated for the derivation to converge (Chomsky 1995). (12) Rs acc (scheren)(θ1,2): Max scheerde zich

What is needed, then, is the merger of a nominal element with properties that make it amenable to a role as an expletive. Zich is a nominal element with an impoverished phi-feature specification: no number, no gender, just 3rd person, and plausibly also without a D-feature. Thus, the role of zich in (12) is no other than that of taking care (by checking, absorbing) of the residual Case. In a sense, then, the insertion of zich facilitates a “mop up” operation. Due to the fact that zich is phi-feature deficient, it may enter a syntactic chain with the subject, due to Agree, mediated by the intervening heads in the extended projection of the verb (see Reuland 2001, 2005 and 2011 for details).8 Being chain-linked to its antecedent it forms one object with the latter that is visible at the interface with the ­interpretive system. As an inherently (lexically) reflexive verb kupa ‘washes’ is inserted in the structure; it has a composite [Agent-Theme]-role to assign, and is syntactically one-place; but it has a residual ACC feature that needs to be checked. The verb moves to K and se is merged into the structure in the Spec, KP. Irrespective of the details of Case “assignment/checking”, the clitic se takes care of the ACC feature of the verb. Because there is an external role to be discharged, the vP layer is projected, in which dete is merged as the external argument, and receives the [Agent-Theme] role. The lexical verb moves to v, followed by cliticization of se. Subsequently, (nominative) Case of dete is checked by T. Finally, dete moves, by whatever principles underlie EPP effects, to Spec-TP. The resulting structure has the required interpretive properties, as represented in (13b).

.  This entails that the mop-up element matches in features with what we may informally call its antecedent.

 Marijana Marelj & Eric Reuland

(13) a.

TP Spec dete[Ag-Th]

T′

TAGRNom

vP Spec

v′

dete[Ag-Th],Nom

v se kupa [Ag-Th]

KP

Spec

K′

se 〈checks Acc〉 se cliticizes

K kupa [Ag-Th] Acc

VP

V kupa [Ag-Th] Acc

b. dete (λx [x kupa[Ag-Th]])

Again, no special operations need to be invoked in syntax. The NP dete gets to be associated with both the theta role Theme and Agent, as required. 3.4  Discussion The non-argumental derivation with SECL and the argumental derivation with SECL together express all the relevant properties in a straightforward manner. In addition, both derivations are in principle available in all clitic languages, unless a particular further property of the environment blocks one of them. However, the analysis does raise a non-trivial issue. Given this analysis one would expect proxy-readings to be readily available in reflexive clitic constructions to all speakers and for all predicates. This expectation is not borne out. Proxy-interpretations are in general more readily accessible in cases with a complex reflexive. Moreover, across languages, proxy-readings seem less likely to be available with grooming verbs, unless there is a complex reflexive, see (14). (14) ??/*Ringo se češlja /üsebe češlja   Ringo:nom SECL comb:m /self comb:m ‘Ringo combs.’ (??/*holds for the proxy-reading)



Clitic SE in Romance and Slavonic revisited 

For speakers who do not allow proxy readings, we argue the argumental derivation is blocked by an economy preference (much like the preference for the bound variable construal of pronominals over their coreferential construal, see Reuland 2011 for elaboration and references).9 Having analysed the role of SECL in reflexives, we will now turn to the role of SECL in the derivation of passives, and see how a low-level difference in the locus of merger derives the difference between reflexives and passives. 4.  Deriving se-passives We argue that the role of SECL in passives is to absorb both case and a theta-role (15). (15) NPi

TP Tk

T′

dete Th Nom T se speck

vP

v′

vj

se absorb Ag kupa Ag

speck

KP

se absorbs Acc

Kj

K′

spec kupa Ag Acc

VP Vj

V′

kupa Ag Th Acc

NPi dete Th

The verb kupa ‘washes’ merges with dete ‘child’ and discharges the Theme role. As in the case of lexical reflexives, SECL enters the derivation in the Spec, KP and checks case. Since there is an external role to assign, the outer vP shell is projected. At this point the remaining theta role Agent has to be checked.10 SECL moves to .  Space prevents us from discussing this issue further, but the interested reader is referred to Marelj and Reuland (2011) for an in-depth account of the blocking principle that accounts for this. .  Logically, there are two possibilities: either NP dete or SECL move to Spec, vP to check it. This choice ties in with the current debate as to whether movement of an argument from a theta-position into a theta-position is licit. See Chomsky (1995, 2004) for arguments that it is

 Marijana Marelj & Eric Reuland

Spec, vP as an XP (this movement cannot target a higher position due the Minimal Link Condition). Movement of SECL to Spec, vP will ensure checking of the theta role Agent. From the specifier of vP, SECL cliticizes onto T, allowing dete to move to Spec, TP, and the passive derivation obtains. Taking the standard assumption that traces are inert for attraction, the trace of the clitic also poses no barrier for subsequent attraction of the object; so the object dete moves directly from its base position to the specifier of TP, where it checks nominative case. At this point it becomes important to spell out with more precision what we mean with “theta-role absorption”. We take thematic roles to be features on the verb. Thematic roles need to be discharged – assigned to XPs (participants in the eventuality underlying the verb). But why do thematic roles need to be discharged? They need to be discharged because they cannot be interpreted on the verb itself but only on elements that enter into a thematic relation with the verb. If the theta-role is not discharged, then it needs to be neutralized on the verb itself. This is the role of SECL. It absorbs the role and the role becomes interpretable on the verb by being associated with a special [+arb, +human] interpretation. What makes SECL a good “absorber” is the fact, due to its phi-feature deficiency, that it need not project an independent argument. In all the passive-like SECL constructions, the “implicit role” is interpreted with the [+arb, +human] flavor. The only way to interpret (16) in Polish is relating to humans, not cats. The same is true in Romance (see, e.g. Cuervo 2003 for Spanish) (16) SC Na ulici se mjauče in street:loc SECL meow:n ‘[One/people] (should) meow in the street.’

(Marelj 2004)

The ban on by-phrases in se-passives and impersonals in South and West Slavonic and Romance (17) follows under the analysis presented here. (17) SC Kuća se gradi *od strane radnika house:nom SECL build:f  by the workers:gen ‘The house is being built.’ (Marelj 2004)

Recall that SECL has an ambiguous XP/X0 status. Unlike in copular passive, where the external-theta role in passives is present syntactically, but realized on a ­non-argument (see Baker et al., 1989, among others), the situation with SECL

not allowed, and Hornstein (2001), among others, that it is. For current purposes we can stay uncommitted, because in all cases instantiating the structure of (14) movement of the object into Spec, vP will be blocked independently, since it gives rise to a violation of an independent grammatical principle (Inability to Distinguish Indistinguishables, see Reuland 2011, and Marelj & Reuland 2011 for references and elaboration). This leaves movement of SECL as the only option for the derivation to converge.



Clitic SE in Romance and Slavonic revisited 

­ assives is different. Though the theta-role is not assigned in the argument posip tion, it is absorbed in the argument position. This means that the role cannot be reassigned to an adjunct position. Again, this boils down to the categorial status of the clitic as an ambiguous XP/X0 element. As a consequence, we can explain the “unusual” behavior of East Slavonic (Russian, Belorussian, and Ukrainian), which, freely and uncontroversially, allow by-phrases in their sja-passives and impersonals. It is the categorial difference between the clitic and the suffix that accounts for the differences among the Slavonic languages. Unlike SECL, sja is an affix – an unambiguous X0.11 It exhibits none of the positional freedom of canonical syntactic clitics. Since the external role in sja-passives and impersonals is not absorbed in an argument position, we expect that it can always be reassigned to an adjunct – quite like in copular passives in English (18): (18) Ru. Dom stroitsja (plotnikami) house:nom built:m  carpenters:inst ‘The house is being built by carpenters’

(Fehrmann et al. 2010)

5.  Summary and conclusion In Slavonic and Romance languages there is a plethora of constructions that feature what is referred to as “the reflexive clitic” se/si/się. Adhering to the law of parsimony, we proposed a unified analysis of the clitic across these different constructions. Moreover, we put forth previously unnoted empirical evidence that the reflexive clitic construction in Slavonic languages allows two quite different readings, and – within our unified approach to the clitic – we derived both the pure-reflexive and the near-reflexive readings. Finally, under our account, we derived an array of intriguing differences across Slavonic from low-level variation in the functional system. Specifically, these differences can be reduced to whether or not a language has syntactic clitics, and to how valence reduction of a verb affects its Case properties.

References Baauw, Sergio & Delfitto, Denis. 2005. New views on reflexivity: Delay effects in acquisition, cross-modular Principle B and reflexive clitics in Romance. Probus 17(2): 145–184. Babby, Leonard. 2010. The Syntax of Argument Structure [Cambridge Studies in Linguistics 120]. Cambridge: CUP.

.  For an extensive discussion with respect to the morphological status of -sja, see (Babby & Brecht 1975) as well as (Junghanns 1996).

 Marijana Marelj & Eric Reuland Babby, Leonard & Brecht, Richard. 1975. The Syntax of Voice in Russian. Language 51(2): 342–367. Baker, Mark, Johnson, Kyle & Roberts, Ian. 1989. Passive argument raised. Linguistic Inquiry 20(2): 219–252. Bošković, Željko. 1997. The Syntax of Nonfinite Complementation: An Economy Approach [­Linguistic Inquiry Monograph 32]. Cambridge MA: The MIT Press. Bošković, Željko & Lasnik, Howard (eds). 2007. Minimalist Syntax: The Essential Readings. Oxford: Blackwell. Chomsky, Noam. 1995. The Minimalist Program. Cambridge MA: The MIT Press. Chomsky, Noam. 2004. Beyond explanatory adequacy. In Structures and Beyond: The Cartography of Syntactic Structure, Vol. 3, Adriana Belletti (ed.). Oxford: OUP. Chvany, Catherine. 1975. BE-Sentences in Russian. Columbus OH: Slavica. Cuervo, Maria Cristina. 2003. Datives at Large. Ph.D. dissertation, MIT. Fehrmann Dorothee, Junghanns, Uwe & Lenertová, Denisa. 2010. Two reflexive markers in Slavic. Ms, University of Leipzig & University of Göttingen. Franks, Steven. 1995. Parameters of Slavic Morphosyntax. Oxford: OUP. Hornstein, Norbert. 2001. Move! A Minimalist Theory of Construal. Oxford: Blackwell. Jackendoff, Ray. 1992. Mme Tussaud meets the binding theory. Natural Language and Linguistic Theory 10(1): 1–31. Junghanns, Uwe. 1996. SJA-Verbs in Russian: Phonology, morphology, or syntax? In ZAS Papers in Linguistics, Vol. 6, Artemis Alexiadou et al. (eds), 66–80. Berlin: ZAS. Labelle, M. 2008. The French reflexive and reciprocal se. Natural Language and Linguistic Theory 26(4): 833–876. Marelj, Marijana. 2004. Middles and Argument Structure across Languages. Utrecht: LOT. Marelj, Marijana & Reuland, Eric. 2011. Deriving reflexives – Deriving the lexicon-syntax parameter. Ms, Utrecht University. Reinhart, Tanya. 2002. The theta system: An overview. Theoretical Linguistics 28: 229–290. Reinhart, Tanya & Siloni, Tal. 2005. The lexicon-syntax parameter: Reflexivization and other arity operations. Linguistic Inquiry 36(3): 389–436. Reuland, Eric. 2001. Primitives of binding. Linguistic Inquiry 32(2): 439–492. Reuland, Eric. 2005. Agreeing to bind. In Organizing Grammar: Linguistic Studies in Honor of Henk van Riemsdijk, Hans Broekhuis, Norbert Corver, Riny Huybregts, Ursula Kleinhenz & Jan Koster (eds). Berlin: Walter de Gruyter. Reuland, Eric. 2008. Anaphoric dependencies: How are they encoded? Towards a derivationbased typology. In Reciprocals and Reflexives. Theoretical and Typological Explorations, Ekkehard König & Volker Gast (eds), 502–559. Berlin: Mouton de Gruyter. Reuland, Eric. 2011. Anaphora and Language Design. Cambridge, MA: The MIT Press. Reuland, Eric & Winter, Yoad. 2009. Binding without identity: Towards a unified semantics for bound and exempt anaphors. In Anaphora Processing and Applications [Lecture notes in Artificial Intelligence], Lalitha D. Sobha, Antonio Branco & Ruslan Mitkov (eds). Berlin: Springer. Safir, Ken. 2004. The Syntax of Anaphora. Oxford: OUP. Sells, Peter, Zaenen, Annie & Zec, Draga. 1987. Reflexivization variation: Relations between syntax, semantics and lexical structure. In Working Papers in Grammatical Theory and ­Discourse Structure, Masayo Iida, Stephen Wechsler & Draga Zec (eds), 169–238. Stanford CA: CSLI.

section ii

Syntactical relations

The lazy speaker and the fascination of emptiness Colloquial Russian from a typological perspective Daniel Weiss

University of Zurich, Switzerland The paper presents an overview of different syntactic gaps in modern colloquial Russian, and focuses on so-called ‘syntactic emptiness’, i.e. the zero copula, the referential zero lexemes Øljudi and Østixii, the generic ‘you’ Øty2 and omission of verbs of action [ØV], taking into account the pragmatical impact of these phenomena (text economy, explicitness, politeness). In addition, the study provides an historical review of colloquial empty slots in Russian (on the example of Avvakum’s ‘Autobiography’) and approaches it from a typological perspective, specifically, by comparing it to West Slavic languages (Polish, Czech) as well as to other languages, which corroborates the idea of the “Eurasian orientation” of colloquial Russian. Keywords:  syntactic emptiness; zero sign; ellipsis; colloquial Russian; typology

1.  Introduction Russian colloquial syntax has long been explored as an object of its own, cf. ­Zemskaja 1973 and Lapteva 1976, but seldom has its divergence from formal Standard Russian (Kodificirovannyj Literaturnyj Jazyk) been analysed systematically from a typological point of view. The present paper aims at characterising this divergence in terms of explicitness /implicitness and textual redundancy/ economy. To this purpose, a broad range of different empty syntactic slots will be a­ nalysed, including zero argument marking, missing predicates and absent sentential connectors. Special attention will be devoted to possible “quirks” that might not only distinguish colloquial from formal Standard Russian, but also from genetically related languages (Polish and Czech) and Standard Average European (SAE) as a whole. The paper is organised as follows: Section 1 is devoted to an overview of the various phenomena in question, Section 2 provides a first typological ­assessment,

 Daniel Weiss

in Section 3 the impact of economy on politeness management is discussed, ­Section 4 sheds some light on the historical development (this part is based on Avvakum’s autobiography), and Section 5 provides a first comparison to ­Polish and Czech. In view of the almost overwhelming diversity of the phenomena involved, the following terminological distinctions are necessary. Unlike in most syntactic work on similar topics, where the term ‘zero’ is used in a rather lose way, in what follows zero signs will be strictly kept apart from ellipses: whereas the latter require unequivocal recoverability of the missing information and belong to the textual level, the former have no overt synonyms and are part of the linguistic system (language).1 This implies that zero signs are irreplaceable and hence obligatory, whereas ellipses can be either optional or (as in the case of hidden subjects of infinitives, converbs, etc.) obligatory. Since this dichotomy is insufficient to capture the specific properties of verb omission and asyndetic clausal linking, we need to introduce a third category for which the preliminary term ‘omission of an abstract hyperlexeme’ will be coined. Moreover, we will have to account for various intermediate phenomena such as the hybrid status of the generic ‘you’, missing subjects that allow a double, triple or even quadruple interpretation or partial verb omission. As for the typology of syntactic emptiness, I propose a threefold distinction embracing referential, predicative and connective empty slots. The latter, however, will be less elaborated than the former two types. Moreover, the following subtypes will be excluded from this description due to their omnipresence or at least ‘pan-European’ nature, which makes them unattractive for comparative p ­ urposes: (i) agentless passives, (ii) missing indefinite objects such as in Sveta poet, Roditeli edjat or Ja čitaju, (iii) missing generic (in)direct objects as in Zdes’ xolodno (cf. Mne xolodno) or Ego ugrozy pugajut. More importantly, (iv) conjunction reduction will be likewise disregarded, including all subtypes such as gapping, sluicing, stripping or stranding (of the auxiliary). For the same reason, (v) so-called ­situational ellipses of the type Like some coffee?, Some coffee? or Coffee? (Heine 2011) will be left aside. To put it more generally, a whole range of ready-made micro-texts or mini-dialogues is irrelevant for our purpose: besides offers of the type just quoted, this concerns orders (Mne porciju syra!), commands (Ruki vverx!, K oružiju!), slogans (Bliže k massam!, Doloj kuxonnoe rabstvo!), road sign instructions (Ne uveren – ne obgonjaj!), conversational routines (Spokojnoj noči!, Ne za čto!), and, finally, proverbs and popular sayings (Pervyj blin komom, Baba s vozu,

.  This interpretation of the two terms is in line with a long standing tradition that can be traced back to Ch. Bally and R. Jakobson, cf. Mel’čuk 1995: 188, 199. For a more fine-grained differentiation, see for instance Apresjan 1986 and Šmelev 1998.



The lazy speaker and the fascination of emptiness 

kobyle legče). All these genres contain one or more empty slots, cf. (Esli) (ty) ne (copula) uveren, (to) (ty) ne obgonjaj!, but they do so in other languages as well: for various reasons (minimal attention time, high frequency of use, etc.), such microtexts universally tend to display maximal brevity and are therefore left beyond the scope of this paper. The functional varieties of Russian involved extend from authentic spontaneous speech to its derivatives, i.e. written genres imitating spoken usage in some way or other, notably personal letters including e-mails, press reports and ­feuilletons, interviews, literary fiction (in particular, chunks of direct speech, free indirect speech and “skaz”), comics, genres of Internet communication such as blogs, forums or chats; moreover, certain genres such as diaries that are oriented towards maximal economy also prove useful. 2.  A first overview After these preliminary remarks, we can now proceed to the overview of the different subtypes that will be examined here. Within the referential domain, these are the following: 1. Ellipses: 2. Zero lexemes:2 a. subject Øi/j/k a. indefinite personal Øljudi b. direct object Øi/j/k b. indefinite impersonal Østixii c. possessor Øposs d. head NP e. relative pronoun Øirel Intermediate case: generic ‘you’

ty2 vs. Øty2

The categories ‘subject ellipsis’ and ‘object ellipsis’ encompass both missing ­personal and relative pronouns, although the latter type is only of marginal significance due to its low frequency. For reasons of space, only coreferential ellipses will be considered; for a non-coreferential use, see example (7) below. In traditional terminology, the indefinite personal zero corresponds to the neopredelenno-ličnaja forma and the indefinite impersonal zero to the bezličnaja forma, but whereas the traditional terms refer to separate meanings of the 3rd plural and the 3rd singular neuter of any verb meeting the semantic conditions, the approach chosen here posits two separate zero lexemes which require corresponding verb agreement. As

.  The symbolical designations of these two zeroes are taken over from Mel’čuk 1995.

 Daniel Weiss

for the generic ‘you’ (traditionally called obobščenno-ličnaja forma), it is usually treated as another zero lexeme; in view of its use in oblique cases and its partial interchangeability with the overt pronoun ty this solution is, however, inappropriate. In this paper, it is considered a separate lexeme ty2 with a full case paradigm and the two allolexes ty and Øty. The predicative empty slots comprise three different subtypes: zero copula in the present anaphoric non-coordinative ellipsis of verbs omission of verbs of action

Øbyt’ ØVi ØV

As will be shown below, the vast majority of ØV represent the above-mentioned third type ‘omission of an abstract hyperlexeme’. Empty connection [ØСonn] is again a heterogeneous category which also consists of zeroes, ellipses and omissions of an abstract hyperlexeme. Since this domain will be left aside in the remainder of the present paper, a few remarks will do to characterise its complexity; for abundant examples, see Širjaev 1986. First, not every lack of a conjunction or another overt connector (adverb, particle) should be considered an empty slot. This notably holds for conditional and concessive clauses of the type Pridi ona vovremja, ja by ne ušel, which in view of their formal markers (prosody, inverted word order, free choice of mood) should be considered instances of a separate construction, cf. Fortuin 2011. Second, in many cases there is no adequate substitute which could replace the prosodic break or the corresponding punctuation: as is shown in Breuer 2002, such textual caesurae as the switch from the narrator’s to the character’s perspective or from the story to its evaluation or its summary, moreover, flashbacks into the prehistory, etc. may prevent the insertion of an overt connector. The same holds true for connections which embrace a whole sequence of sentences, e.g. a motivation of a statement like Ippolit Matveevič ne ljubil svoej tešči: here the author cannot introduce every sentence by a causal conjunction since the latter would always join the subsequent sentence with the immediately preceding one. And third, even if all syntactic conditions are met, one has to account for two different situations: either one may reconstruct the missing element unambiguously as, for example, ‘and’ or ‘or’, or else, one cannot reconstruct one single connector, but only some abstract semantic relation, e.g. cause – effect or action – goal. In the former case, we are dealing with an ellipsis, in the latter with the omission of an abstract hyperlexeme. Empty slots often occur in clusters, for instance in double or even triple combinations of different types. The following pair of sentences illustrates a referential/ predicative “package”: (1) Дома Øljudi Øbyt’ в восторге. (2) Так бы Øi и сразу ØV!



The lazy speaker and the fascination of emptiness 

The next pair provides examples for the combination ‘referential + connective’: (3) Øi Слепил голую бабуj с могучей грудью и [Øi Øj]3 на площади поставил – ØConn Øj символизирует материнство Øk, понимаете ли.  (Московский комсомолец, 1990)

(4)  Проект новых автодорoжных правилi: ØConn Øty2 Нарушил Øi раз – Øljudi ØV прокол в пaспорте, ØConn Øty2 нарушил два ØConn Øty2 – второй прокол, ØConn Øty2 нарушил три – в графу пятого пункта вписывается: Еврей.  (Советский Союз в зеркале политического анекдота: 422)

Example 3 contains an irreplaceable connective zero, located at the point where the character’s perspective is introduced. The “triple package” ‘connective + referential + predicative’ is represented by the following example: (5)  Юра всегда сторонился таких сухариковi – ØConnØi Øbyt’ слабосильны, но [Øi Øbyt’] обидчивы и [Øi Øbyt’] мстительны. (A. Рыбаков)

After this preliminary overview let us now enter into the details. The discussion of the different subtypes will focus on those peculiarities that might prove to be exclusively characteristic of colloquial Russian. 2.1  Referential ellipses The most discussed topic in the existing literature is subject ellipsis, which has mostly been examined by adherents of generative syntax when discussing whether modern Russian is a PRO-drop, a half PRO-drop or even no PRO-drop language at all.4 Unfortunately, this whole debate mainly focuses on written formal speech, a notable exception being Grenoble 2001; besides this, colloquial registers are included in the analysis of subject omission in the Russian National Corpus by Zdorenko 2010. In colloquial usage, subject ellipsis may affect the participants of the on-going discourse (deictic ellipsis), cf. (6) Øi Øbyt’Уверен, что Øj срежетесь!

.  These two ellipses have been marked for the sake of completeness, although they represent gapping within conjunction reduction, which, as may be recalled, has been excluded from our overview. .  For a comprehensive overview ot these studies, see Meyer 2009 and Meyer 2012 (unpublished Habilitation Thesis); for a quantitative account of anaphora in a Russian short story, see Kibrik 1996.

 Daniel Weiss

As this example shows, subject ellipsis can occur with predicates both with and without personal marking: the present srežetes’ has a personal ending, whereas in the main clause person remains unmarked due to the missing copula with uveren. This example is also revealing in another respect: reference of elliptical subjects may switch; here it does so from ‘I’ to ‘you’. Both properties are also characteristic for anaphoric subject ellipsis, as is shown by example (3) above and the following excerpt from Berija’s diary:5 (7) Øi Договорился с Кобойj, что замом Øi назначаю Аркадия. … Кобаj сказал: «Но чтобы товарищ Масленников не обиделся, пусть пока остается у тебя замом». Øi Спорить не стал. То нe было ни одного зама, теперь два, зато один дей­ствующий. (Лаврений Берия, Тайный дневник 1941–1945)

This time, the lack of person marking is due to the preterite forms dogovorilsja and stal.6 The example also illustrates the non-coreferential use of ellipses, cf. odnogo zama – dva [zama] – odin [zam]. It should be added that in the case of the interclausal coreferential link between naznačaju and dogovorilsja the subject ellipsis is stylistically neutral: such examples may for instance occur in scientific discourse, cf. Weiss 1993: 61. In the above-mentioned diary, one even finds examples with more remote (intersentential) antecedents, cf. (8)  Обобщили данные по [полякам Андерса]i. Øj Направил Кобе доклад. [Панфилов и Жуков]k тоже докладывают, что Øi ведут себя хамски.

Such uses are, however, less acceptable. In Weiss 2008: 368 a peculiar case of egoreference was noted: Ju. Daniel’ is undoubtedly referring to himself, but grammatically, the missing pronoun may represent all three persons since this author often utilises autodialogues, addressing himself in the 2nd person, and sometimes adopts another perspective, referring to himself in the 3rd person: (9)  Ишь, как Øi расфилософстовался, босяк, клошар несчастный!  (Ю. Даниэль)

Similar grammatical ambiguities appear in contexts where it is unclear whether we are dealing with reported or direct speech, cf. Na tebja žalujutsja s Kavkaza – v

.  This diary is by all evidence not authentic (see http://rutracker.org/forum/viewtopic. php?t=3766539). This does, however, not matter for our purpose since the syntactically empty slots occurring in it are all judged acceptable by native speakers. .  The default assumption for diaries is certainly that unmarked sg verb forms will refer to the author. This is, however, not mandatory: a reference to Koba could as well have been realised elliptically in this text.



The lazy speaker and the fascination of emptiness 

1942 g. Øi gotovil sdaču, podbiral agenturu (Øi = ty vs. on). Such examples undermine the very definition of ellipisis since the missing information is ­unambiguously recoverable only with regard to reference, whereas grammatical person remains underdetermined. The impact of free indirect speech on the interpretation of subject ellipsis is analysed in a Ph.D. thesis by O. Artyushkina (2010): as she notes on p. 116, the omission of the subject pronoun as in i Øi počti ne slyšal vystrela v Ėrcgercoga! … Ob”javila vojnu Avstrija Serbii – kak ne zametil Øi (Lenin v Cjurixe) is a characteristic marker of this type of reported speech in A. Solženicyn’s work. Translators thus face the dilemma of whether to render this as the 1st or 3rd person pronoun by adopting either the character’s or the author’s perspective. These findings are remarkable from a typological point of view: colloquial Russian turns out to be an optional and radical PRO-drop language since it does not necessarily indicate the category of person. This property separates it from West and South Slavic where the richer verbal morphology allows to always express person; at the same time, colloquial Russian is getting closer to modern Chinese, a radical PRO-drop language (Bisang 2011). Another interesting observation concerns Russian sign language: here, the (deictic) expression of ‘I’ and ‘you’ is mandatory, whereas anaphoric ellipsis with the 3rd person remains possible (Kibrik & Prozorova 2007). Despite its greater overall grammatical poverty, this language thus marks person more consistently than colloquial Russian. It goes without saying that these observations have to be corroborated by empirical, corpus-based work. This will allow to determine the ratio of combinations with and without personal marking of the predicate.7 In a Master thesis on anaphorical and deictic subject ellipsis in colloquial use and fiction (P.Sanaev), defended at the Zurich Slavic Department by Daniela Kovačević, this last factor turned out to be statistically significant; antecedent properties such as thematicity, subject status and personal reference had, however, a much greater impact on the choice between ellipsis and pronoun. Moreover, what seems to matter is the opposition of 1st and 2nd vs. 3rd person: according to my findings, the first two tend to be elided more often than the third one.8 All these figures may vary according to the genre; it is by no means obvious that authentic spontaneous speech shows the same ratio of overt and elliptical subjects as literary fiction or Internet communication. ­Incidentally,

.  Unfortunately, this distinction has not been taken into account in Zdorenko 2010. .  This is in line with the results obtained by Grenoble 2001 and Zdorenko 2010: 128 f., where the percentage of null subjects was highest in second-person contexts and lowest with the third person. As both authors explain, this was mostly due to the high frequency of the subjectless verb forms znaeš’ and ponimaeš’.

 Daniel Weiss

certain formal genres likewise require ellipsis of the first person, e.g. official forms.9 More empirical research will probably also provide evidence for additional restrictions like the following: in Iordanskaja and Mel’čuk 2011 it is argued that in parenthetical constructions of the type Položenie, *Ø sčitaju, uxudšaetsja, Položenie, *Ø sčitaet, uxudšaetsja and Položenie, kak *Ø sčitaet, uxudšaetsja subject ellipsis is blocked. The constraints of object ellipsis are still poorly described. Berger 1989 treats them together with subject ellipsis, and MacShane 1999 compares object ellipsis in Russian, Polish and Czech. The following example illustrates the interplay of both subject and object ellipsis and of overt and elliptical anaphora: (10)  Хозяйкаj Соловейчикаi, худенькая молодая женщина с тихим лицом, накрыла на стол. Что побудило ееj сойтись с Борисомi? Øi уедет и забудет ееj. Øi понравился Øj? Øj Пожалела ссыльногоi? Рядом с нейj Борисi с его замашками столичного волокиты выглядел жалко.  (A. Рыбаков)

But deictic object ellipsis is also possible, cf. ego-reference in example (18) below or in Stalin vzorvalsja, nazval Øja zmeej v očkax (Berija, Tajnyj dnevnik). The elision of relative pronouns (cf. Zemskaja 1973: 238–240; Kručinina 1968) has a markedly colloquial flavour. The following examples, quoted from Woolhiser 1990: 18, contain missing relative pronouns with subject, direct object and indirect object function, respectively: (11) Снесли сарайi Øirel во дворе был. (12) Ты видел запискуi Øirel Петров тебе оставил? (13) Девчонкаi Øirel он все пристает сидит впереди.

As the author rightly points out, the simultaneous omission of prepositions is even more remarkable, cf. (Woolhiser 1990: 20): (14) Ты знаешь девчонкуi Øpraep Øirel он все получает письма?

In other Slavic languages, such examples would simply be ungrammatical. A syntactically comparable situation obtains when relative pronoun ellipsis clusters with a missing head noun, cf. (15) Øi Øirel Мусор собирает не приходила?

.  The following excerpt is taken from the current form to be filled in by foreign citizens when applying for a visa to the RF: Ob otvetstvennosti za soobščenie ložnyx svedenij v zajavlenii… Øi preduprežden. S soderžaniem statej 26, 27 zakona «O porjadke vyezda iz Rossijskoj Federacii …» Øi oznakomlen.



The lazy speaker and the fascination of emptiness 

But missing nominal heads also occur without relative clauses, if there are some modifiers left, cf. well-known examples such as those quoted in Zemskaja (1973: 238–240): (16) Ødempron/N В шубе не приходила?

This procedure is again markedly colloquial; it may even affect forms of address, cf. below, Section 3. Moreover, it combines with indeclinable comparatives, which produces strings such as in the following proverb from a prison camp: (17) Øindefpron [Ørelpron Øbyt’] Тяжелее стакана в руки не беру.

In this case, nominal categories such as number or gender also remain unmarked. The next example is taken from a comic, where we see Donald Duck being chased by four wolves, whereas Daisy and a male observer are watching this scene. D ­ onald now shouts: (18) А-а-а! Оставьте Øi! Здесь есть Øindefpron пожирнее меняi!

Since he obviously is referring to the bystanders, the missing head noun may be deciphered as ljudi or čelovek. The former variant as well as the zero version are semantically ambiguous in that they may refer to a single person or else (if the speaker is being impolite by including Daisy) both people. Again, a similar situation is hard to imagine in another Slavic language. It may be added that the same indeclinable comparative may also function as an anaphoric reference as in the following excerpt from Berija’s diary: (19) Есть наглые людиi, а Øi наглее поляков Øj не видел.

In this case grammatical number and gender are unequivocally recoverable. Literature on possessor ellipsis is even scarcer: the only pertinent titles are Šmelev (1998: 171–74), Mikaelian (2002: 112–14) and Raxilina (2010: 296–97). Consequently, we do not even know whether colloquial usage differs from standard language in this respect. Beyond any doubt, however, is the additional ambiguity produced by this type of ellipsis. This holds true for deictic as well as anaphoric uses. For instance, the following reference is ambiguous in a context where there has been no previous mention of the wife in question: (20) Расскажите это Øi жене!

i = moej/i = svoej (vašej)

This type of situational ambiguity is predicted by Šmelev 1998: 171–172. In the following case the hearer interprets the hidden possessor deictically, identifying him with the speaker, whereas the latter is anaphorically referring to some previous antecedent.

 Daniel Weiss

(21) A. От него дяди сохранилось письмо/ адрессовaнное матери // Б. Их ведь не одно было // публиковались же A. Его маме/ а не моей // (Русский язык в его функционировaнии I, 37)

2.2  Referential zeroes Let us now turn to referential zeroes. The zero subject lexeme Øljudi is stylistically neutral, hence not bound to colloquial usage, and has the following properties: indefinite speaker-exclusive or else generic reference (the latter usually includes the speaker) and ephemeral status (it does not establish sequences of coreferential links). It is not subject to any syntactic or semantic constraints except in the passive (for details, see Mel’čuk 1995: 183). It triggers not only verb agreement but functions also as a controller of converbs (adverbial participles) and reflexives. The grammatical plural agreement does not necessarily denote semantic plurality: an utterance such as Smotri, idut! may refer to a single person approaching the speaker. The potential of referents may be narrowed down by a spatial or institutional adjunct, cf. v redakcii pišut, s Kavkaza dokladyvajut, etc. The generic use is often related to implicit deontic modality, cf. Tak ne vedut sebja! Purely descriptive utterances are, however, also common, cf. (22)  И если ты случайный прохожий, забредший “на огонек”, тебя обязательно Øljudi усадят за стол, пригласят разделить радость. Принято приходить на свадьбу со своими гостями. (АиФ 28/1993: 6)

This example also shows the contrast of speaker-exclusive Øljudi and speaker-inclusive ty2, both used with generic reference. The indefinite personal zero is widely used in contexts that reveal a passive, if not helpless attitude of the potential agent: it is not the potential agent who acts, but some anonymous power acts on him. This perspective manifests itself in e.g. Tam xorošo kormjat where the ‘European view’ displays the inverse relation, cf. On mange bien là-bas, Dort isst man gut, Tam się dobrze je, etc., but also in Kak vas zovut?, Nas raz”edinili (re-initiating an interrupted a phone call) or Tam (ne) puskajut. That the latter may indeed reflect a different interpretation of the cultural script in question is confirmed by the reaction of an American visitor in the next example: (23)

О  дин из известных амeриканских ученых, немного знавший русский язык, побывав в Москве, с недоумением рассказывал, что не смог понять значение вопроса, котoрый ему Øljudi задали около Новодевичьего монастыря. Какой-то гражданин, приняв его за русского, спросил: «Øljudi Пускают?» “Если бы он меня спросил, открыто ли, я бы его понял”, – объяснял американец. (Баранов/ Казакевич, Парламентские дебаты: традиции и инновации, М. 1991: 21)



The lazy speaker and the fascination of emptiness 

This language-inherent de-agentisation of the individual was certainly reinforced by Soviet institutions which deprived their subject of its autonomy in many everyday life situations, to mention but the notorious čto dajut? or vybrosili X to be heard when queuing for some consumer good. The restrictions noted at the beginning of this section need some refinements. For instance, as is shown in example (23), the ephemeral indefinite reference established here by zadali can be made more precise in the subsequent string of text, cf. Kakoj-to graždanin…sprosil. Definite anaphora is likewise possible, but requires pronominal reinforcement: (24) Øljudi Убили Листьева… Эти силы хотят навязывать нам свою логику, хотят нас запугать. (Горбачев, Огонек 11/1995: 43).

On the other hand, the zero lexeme Øljudi may function in contexts where it indirectly refers to a single definite antecedent, for example the person calling in the following example: (25) После похорон Черненко в Кремле раздался звонок: – Алло, это Кремль? Вам генсеки не нужны? – Вы что – дурак или больной? –  Да, – Øljudi радостно ответили в трубке, – и дурак, и старый, и больной!

Such links may be best described as inferred coreferential anaphora. They work even in the interaction of two persons, as is shown by the next, rather tricky sequence of mere zero subjects with regularly alternating reference, quoted from Bulygina & Šmelev (1997: 321): (26)  Из рубки Øljudi свистнулиi в машину, и когда из машины Øljudi ответилиj, Øljudi крикнулиi туда, чтоб Øljudi не отходилиj от резерва. Да я уже два часа у резерва стою! – Øljudi ответилиj из машины.  (Ю. Казаков)

Not surprisingly, such inferred coeferential uses of the zero lexeme compete with subject ellipses: (27)  В окопах возле дороги сидели людиi. Огарков подъехал к нимi и дрожащим голосом спрoсил про свою дивизию. Ему Øi/Øljudi ответили:…(Казакевич)

And finally, even the ban on speaker reference may be ruled out in specific pragmatic contexts. Bulygina and Šmelev (1997, 350 f.) discuss a case with displaced roles: the speaker refers to himself by Øljudi and to the hearer by the 3rd sg. (28) Ейi Øljudi по-человечески говорят, а онаi не понимает.

 Daniel Weiss

Whereas this pattern shows a regular behaviour, there are also two pragmatic idioms where Øljudi refers to the speaker exclusively, viz. tebé govorjat/tebjá sprašivajut and Vam zvonjat/vas bespokojat/s vami govorjat iz…; for details, see Weiss 2008: 378 f. The indefinite impersonal zero lexeme Østixii shows a wide range of uses; the limits of its productivity, however, have not been explored so far, nor is there any in-depth description of its stylistic variation according to the parameter ±colloquial use. The richest data is probably provided by Guiraud-Weber 1984, which is also the source of most of the following examples. These are selected according to the criterion of translatability: one will have a hard time when trying to find direct equivalents in other Slavic languages. An important subtype describes physical processes caused by natural forces or violent (e.g. military) action that affect the human body, cf. (29) Правую щеку у Рыбникова Østixii безобразно раздуло. (30) После этого Østixii ему перекосило рот. (31)  Самого-то чего Østixii в такую дугу согнуло? Если хорошо-то жил – чего же согнулся? От хорошей жизни?

The anticausative reflexive form sognulsja in the last example illustrates a construction with similar semantics (unknown or unspecified cause). A military force is involved in examples such as Østixii Ego na opuške ranilo, Østixii Artilleriju razbilo, etc. Unlike in other Slavic languages except Polish, in Russian the cause may be added as an adjunct, cf. Ego ubilo molniej/udarilo tokom/razmorilo ot žary, Lodku uneslo vetrom, Derevo slomalo burej, etc. In a case such as Ego pereexalo tramvaem, quoted by A. Wierzbicka (1988: 224, 225), even Polish does not allow for a literal translation. Both A. Wierzbicka and I. Mel’čuk (1995: 194 f.) still postulate a zero subject Østixii for such sentences. However, this approach raises serious doubts: why should we posit an additional argument if the cause is overtly expressed, and what would be the meaning of such a zero? destiny? the “elements”? a demon? At any rate, unlike Mel’čuk, who considers such sentences non-synonymous with their counterparts with overt subjects (žara ego razmorila, veter unes lodku, tramvaj pereexal ego), most native speakers do not feel any difference between both variants. The whole phenomenon attests not the omnipresence of Østixii, but simply a marked preference for subjectless sentences which is a distinctive feature of Russian syntax anyway. Note that the same controversy arises when the cause is expressed in the preceding context, cf. (32) Я наскочил на мину и мне оторвало руку.10(Мирнев)

.  This example contradicts Mel’čuk’s observation that examples like Soldaty brosilis’ na Ivana, i srazu že sil’nym udarom ?ego sbilo s nog are of doubtful acceptability.



The lazy speaker and the fascination of emptiness 

Our last referential candidate is pretty specific: not only can it not be considered a real zero sign, but it is unique in having an overt alternative. As was stated in the beginning, the empty generic ‘you’ alternates with ty2, which already occurred twice in example (22), and the two versions do not show any semantic difference. But then, the same holds true for subject or object ellipsis, so why not treat Øty2 as a simple ellipsis of ty2? For the purpose of this analysis I decided to proceed otherwise. First of all, so far no study has been devoted to the distribution of the two variants; what has become obvious, though, is that besides a large intersection where both variants are possible, both of them also have their proper exclusive domains of use. For example, contrastive contexts and prepositional government require overt ty, but in the following contexts the latter would be blocked: (33) Его Øty2/*ты не разберешь! (34) Тебя Øty2/*ты никак не застанешь дома!

On the contrary, subject ellipsis as examined in Section 1.1 is never mandatory except with generally subjectless non-finite constructions such as infinitives and converbs. Thus it seems advisable to treat the empty and the overt variant of ty2 as two allolexes of the same lexeme with a full case paradigm. As for the remaining properties of this lexeme, Guiraud-Weber 1990, Šmelev 1996 and Bulygina & Šmelev 1997 provide ample information. At the distinction of Øljudi, its reference is not only ego-inclusive but even ego-centered in that it generalises the speaker’s own experience by including the hearer, much in the sense of: “it is not only my experience, you are familiar with such situations, too.” Therefore it bears a touch of empathy. It may, however, also switch to speaker exclusion as in the following example: (35)  Днем яi всегда крепко себя держуi, из меняi ни “оха”, ни вздоха Øty2 не выжмешьj, а ночью Øty2 проснешьсяi и вся Øty2 подушка мокрая от слез.  (М. Шолохов)

This example is particularly tricky since it not only shows two opposite interpretations of Øty2 but also its interplay with explicit ego-reference. When negated, the generic ‘you’ often occurs with the perfective present to express impossibility of the action involved, cf. example (33). Otherwise, it is ­frequent in generalising conditional clauses, cf. example (4) above or example (36): (36)  Kазалось бы, Øty2 заболел – Øty2 иди и лечись на здоровье. Но тут-то и начинаются “маленькие хитрости” нашей системы страховой медицины. (АиФ 5/1994: 13)

Moreover, with regard to its equivalents in other languages it shows at least two “quirks”: first, it combines freely with the homophonous form of address ty1 in the same sentence, as is shown in example (34), and second, it is also compatible with

 Daniel Weiss

situations where the relation between speaker and hearer requires the polite form of address vy, cf. (37) – Виктор Иванович, если вы, как и прежнее правительство Гайдара, не будете разъяснять народу свою политику (чего хотите, какими путями пойдете), то ваши действия будут обречены на провал. Реформу без поддержки нарoда Øty2 не сделаешь. (АиФ 5/1994: 2)

And finally, as in the case of Øljudi (see example (27)), a use of Øty2 may likewise be inter­pretable as an elided form of address (ty1): in a conversation with a foreigner, one may for example explain him how a self-service shop works in the following words: (38) Здесь самообслуживание – Øi/Øty2 сам берешь все, что тебеi/ty2 надо.

To sum up, the generic use of ‘you’ in modern Russian is reminiscent of parallel uses in many European languages, but it shows a series of characteristics that again reveal a preference for the ‘zero strategy’ even in comparison with other Slavic languages. 2.3  Verb omission As may be recalled, the predicative empty slots comprise three different subtypes: zero copula Øbyt’ in the present, anaphoric non-coordinative ellipsis of verbs [ØVi] and omission of verbs of action [ØV]. The latter turns out to be the most striking “quirk” of colloquial Russian. Before tackling this subject, the other two types will be briefly commented on. As is well known, the zero copula has no substitute since est’ means something else;11 literature on this subject is vast, therefore there is no need to dwell on this. Anaphoric ellipsis other than gapping (cf. Deti skučajuti na lekcii, a vzroslye ØVi na doklade) is attested in monological strings like (39). (39) Совсем забыл, куда сунулi проклятые деньги!… Куда это я их ØVi?

What is most striking here is the re-emergence of the overt subject ja that was elided in the preceding sentence; moreover, we also find an object pronoun ix instead of its ellipsis. We thus have come across the nominal skeleton which is necessary for verb omission, see below. Verbal anaphora may also occur in a dialogical string, cf. K. Čukovskij’s famous verse: – Vyxodite iz berlogi i vraga na roga podnimitei-ka! – My vraga by ØVi na roga by…; here, grammatical person and mood change according to the alternation of speakers.

.  It may be noted that this is another verb form not marked for person. It occurs however rarely without overt subject, except in anaphoric uses.



The lazy speaker and the fascination of emptiness 

The omission of verbs of action has been examined by numerous authors, cf. Širjaev 1973, Mel’čuk 1974, 1995; Wiemer 1996; Ščadneva 2000, McShane 2000; Saj 2002; Мažara 2010, 2011 and Weiss 1993, 2011a; moreover, a Ph.D. thesis on this subject is currently being prepared by Je. Мažara. Therefore, in what follows I will just recall the most essential properties of verb omission. According to Mel’čuk 1974, any verb of action can be omitted provided that there remain at least a bipartite structure of the type illustrated by example (39). To put it more precisely, the sentence must contain at least two arguments or an argument and some adjunct such as bystro, domoj, vniz golovoj, komu?, v institut, s pelenok, etc. There is no requirement for one argument to be the grammatical subject: as was already shown in example (2), verb omission is compatible with missing subjects. The absent verb cannot be in the scope of total negation; on the other hand, partial negation remains possible, cf. Ja ØV ne pro tebja. The somewhat vague term ‘verb of action’ has been modified in Wiemer 1996 who pointed out several classes of verbs that may likewise be omitted although they do not denote any action at all, to quote but one example: A ona ØV v slezy. A poorly studied aspect of verb omission concerns grammatical categories. Tense, aspect, and even mood often admit different readings, cf. Kak tebe Neapol’ (nravitsja/ponravilsja)? There are no illocutional restrictions: for example, verb omission may occur in offers: Vino budeš’?, wishes: Xorošo by ØV pod duš! I mne by takoe ØV! Vot by ØV na Sevan!, requests: A musor ØV potom! and even questions how to interpret a dream: Prisnilsja tank. K čemu Øv ėto? As these examples show, auxiliaries such as by or budeš’ are preserved when marking an illocution; otherwise, e.g. in the passive, they undergo deletion. Verbless requests may be spelled out in the imperative or the perfective present tense, cf. A musor potom vynesi/ vyneseš’!. Verb omission may be totally or partly phraseologised or completely unbound. Besides idiomatic descriptive phrases such as S X-a vse ØV kak s gusja voda, one finds numerous pragmatic idioms, cf. threats: Já ØV tebja/tebe!, warnings: S X-om ØV [by] ostorožnej!, criticisms: Srazu by tak ØV! Ty opjat’ za svoe ØV?, oaths: Ja sjuda bol’še ni nogoj ØV!, excuses: Ja ne naročno/nečajanno ØV!, or stereotypical questions: Kak ty na ėto ØV? Kak tebe ØV X? As can be seen, some of these idioms also violate the actional meaning constraint. Syntactic idioms (in the sense of Apresjan et al. 1978) are represented by Opjat’/večno vy ØV so svoimi X-ami (ØV = pristaete/ lezete/suetes’)!, cf. Stojnova 2011, or the somewhat non-standard Ja vozmutilas’, napomnila ej sej fakt, ona ØV takaja: «Daaa?» (Sokolova & Makarova 2011). On the other hand, non-bound uses are illustrated by the following examples: (40) Это вы фарш ØV? (покупаете/приготовляете)?(Širjaev 1973) (41) Это вы Ремарка (хвалите/читаете/купили)?(Širjaev 1973) (42)  Вы правы, так бы любой на вашем месте ØV. (отреагировал, поступил, подумал, сделал)(Мažara 2010)

 Daniel Weiss

These examples reveal a second continuum typical of verb omission: its meaning seldom lends itself to unequivocal reconstruction, most examples being located somewhere in between both ends of the scale of explicitness, i.e. unambiguous recoverability and completely vague interpretation. This notably holds true for empty slots that may be interpreted either as verbs of motion or communication (Širjaev 1973), cf. Ty kuda? (sobiraeš’sja/zvoniš’), My v institut (idem; zvonim, pišem). In the majority of cases only an abstract meaning shared by a whole group of existing verbs may be reconstructed, cf. financial transaction: Ja emu ØV rubl’ (dal, odolžil, zaplatil…) or the emanating of bad odour: Ot nego ØV vodkoj, kak iz kabaka (Čexov) (neset, vonjaet, razit, zaezžaet). Such examples raise the question of whether they should be analysed in terms of ambiguity or rather vagueness (cf. Мažara 2011). Moreover, in some cases one hesitates whether to interpret the missing verb as a verb of action or the zero copula, cf. when a mother addresses her baby as follows: Čto ž ty ØV / Øbyt’ na kolenočkax vse vremja? The cues for the appropriate interpretation may be provided by (i) the situation in which the dialogue occurs, cf. Kuda ty bez šapki?, (ii) the preceding context, cf. A pokojnica-to – prosti Gospodi, vrednaja byla baba! Kak ėto on s nej ØV tridcat’ let? (vyderžal, prožil) or (iii) the linguistic structure of the given sentence (lexical filling and case frame) as in the citation from Čexov; in many cases a combination of two of these three factors is required to grasp the intended intention. It goes without saying that all this input information has to be checked against background knowledge stored in scripts, frames and the like. At any rate, contrary to a widespread opinion the bare case frame represented by the arguments and adjuncts “left behind” is practically never sufficient. For a more in-depth discussion of the whole question, see Weiss 2011a. Given this complex scheme of semantic processing, the traditional view according to which it suffices to distinguish (a) so-called zero verbs of motion, communication, physical action, etc., on the one hand, and (b) contextually bound, spontaneous omissions, on the other hand, turns out to fail completely, cf. (Weiss 2011a: 138–139, 149–152). If there are some privileged semantic groups such as verbs of motion or speaking, they owe this status merely to their higher frequency and may thus come to mind first; this does, however, not justify the assignment of a separate theoretical status to them. Idioms aside, all verb omissions are somehow context-bound and have to be processed in the way sketched out above. Among the questions still to be solved is Mel’čuk’s claim that in principle any verb of action may be subject to omission, including even most abstract tokens as A my ego ØV po t (= prodifferenciruem). So far, I know of just one counterexample: in (43) Потом Ø1 другой костюм и быстро Ø2 на сцену.(Мažara 2010)

the first missing verb is easily decoded as a verb of dressing (nadel); the second must be a verb of motion such as pobežal. But neither in this example nor in such



The lazy speaker and the fascination of emptiness 

script-induced contexts as taking a shower or going for a swim would we interpret the missing verb as a verb of undressing, cf. *on srazu kal’sony i pod duš or *ona srazu koftočku i pryg! V vodu. If a verbless realisation is intended, one has to add some adverb, such as (portki) doloj. One final example that contains not less than five missing verbs will illustrate different effects achieved by this procedure. Besides text economy, it allows ­accelerating and dramatizing the on-going narrative: (44)  Как-то вечером, уже после лова, берет он мольбертик свой, надевает шлепки и – Ø1 на лоно. Закат рисовать. И там, в этой густой траве, его что-то колетi в основание большого пальца ноги. И шипит, как закипающий чайник. Он травку раздвигает, а там Øbyt’ хиромантия какая-то, с карандашик величиной. Он ей Ø2 пинка! Она его – еще разик Øi3. Обозлился пейзажист, сгреб этого гаденыша в кучку и нам несет: мол, мужики, червяк какой-то. Да еще щиплется. … А мы ему Ø4: Коля, ты с ума сошел, это же эфа! (Огонек 1992/8)

Even considering that this fragment contains a zero copula and one of the omissions is anaphoric, whereas another one replaces a support verb (dal pinka), such an accumulation of missing verbs attests a pretty careful stylistic elaboration.

3.  General assessment: Economy vs. implicitness Our little overview has already provided ample evidence that textual economy is particularly dear to speakers of colloquial Russian: more often than not, they encode as little referential and/or predicative information as possible. In Gricean terms, they thus comply with the third submaxim of the maxim of quantity: “Be brief ”, or, to put it more explicitly: “Produce the minimal linguistic information sufficient to achieve your communicational ends!” (Levinson 2000: 114). This formulation is flexible enough to reflect the different preferences of linguistic communities: obviously, the component “sufficient to achieve your communicational ends” is interpreted more  liberally by a speaker of Russian than, for instance, a speaker of English or French. But this amount of economy has its price to be paid by the hearer: when the speaker withholds essential information, it is the hearer who bears the heavier burden by decoding the missing elements in the message. More precisely, he faces many ambiguities that may or may not be resolved by the aid of the surrounding context: in the present study, examples (9, 20, 21, 27, 38, and 40–42) allowed such ambiguous readings. No wonder that even cooperative hearers do not always succeed to find the appropriate interpretation. Such was the case in example (21), where the speaker had to correct the hearer’s assumption

 Daniel Weiss

about the elided possessor. The next example presents a more elaborate referential misunderstanding: (45) А потом… там в коридире Øi/Øljudi делали линолеум, Б вам? А  вам! Мыi делали, я с соседом. Ну и конечно, я все-таки устала и все, но все-таки еще было ничего. Б  ты сама делала линолеум? Ты сама приколачивала линолеум? Я думала, ты такие вещи не делаешь. А господи! И обои сами Øi клеили и все. (МФ)

Here, the hearer B justifies her failure (vam?) by arguing that she did not suppose that the speaker did such work on her own. By allowing for such mismatches, our lazy speaker thus violates another, not less powerful pragmatic principle, viz. Grice’s first submaxim of quantity: “Make your contribution as informative as is required by the present purpose of information exchange”. This contradiction between maximal economy and maximal explicitness is by no means new,12 but colloquial Russian manifests a strikingly unbalanced relation between the two extremes. What’s more, it may even turn out to be pretty unique in combining both referential and predicative indeterminateness. A more advanced approach to the interaction of brevity and vagueness is offered in an illuminating paper on numerical measurements by M.Krifka. His first rule of thumb says: “a. Short, single numbers suggest low precision levels. b. Long, complex numbers suggest high precision levels.” (Krifka 2002: 433). The author then goes on by relativizing the notion of short expressions to the so-called expression space; for instance, one thousand may be opposed to two thousand, three thousand…, to nine hundred, one thousand one hundred,… or nine hundred ninety-nine, one thousand and one,… This amounts to say that “the average complexity of expressions in more fine-grained expression spaces is greater than the average complexity of expressions in less fine-grained expression spaces” (Krifka 2002: 439). Obviously, what holds for numbers is not less true for referential and predicative expressions. The equivalent of Krifka’s expression space would then be the universe of discourse shared by speaker and hearer in a given string of dialogue combined with the communicative tasks to be achieved by this dialogue. But the fact remains that natural languages differ in their search for precision within the same universe of discourse. Thus, as is shown in an experiment presented by Stoll & Bickel 2009, referential density seems to depend on the given language: Russian speakers tended to be more explicit in their identification of new protagonists when retelling Dan Slobin’s Pear story than speakers of Belhare did when facing the same task. In the latter language spoken in Nepal, referents were usually .  Bisang (2009: 35) traces it back to Von der Gabelentz.



The lazy speaker and the fascination of emptiness 

introduced by the indefinite pronoun ‘one’, whereas Russian speakers in the same situation used full nouns like xozjain, drugoj mužčina, devočka, mal’čik s korzinoj. 4.  The impact on politeness management According to the principle of social iconicity established by Leech 1977, we may posit a correlation between the complexity of an utterance addressed to another person and its degree of politeness. This does not hold for conversational routines such as greetings, thanksgivings, excuses and the like since these formulae are not part of the real politeness management. Although the principle of social iconicity is only a rule of thumb which Leech himself abandoned later,13 the idea has remained popular ever since. On the other hand, Western politeness theories have long been attacked because of their Universalist claim: systematic comparison with non-European cultures raised serious doubts about their applicability elsewhere. It therefore seems to be tempting to test this correlation on uses of various empty slots in spoken Russian: does the shortened version sound more impolite than its full counterpart? Our first case illustrates a situation typical for official discourse, e.g. in public transport or at the custom office. Public service employees may address their clients by simply referring to their outer characteristics (appearance, personal belongings, etc.), thus eliding the head noun, cf.: ØN S čemodanom/v očkax, ­projdite! [= Devuška s čemodanom,…]; ØN S rebenkom, k pasportnomu stolu! ØN U kogo Øbyt’ bilety, projdite v salon! This “concealment” of the addressee would certainly be considered most disrespectful in other languages. For Russian, it is simply mentioned as a feature of “half-official discourse with a certain touch of rudeness” in Zemskaja 1987: 224. Two remarks should be added to this. (i) The mentioned type of behaviour is asymmetrical since the passenger cannot address the public service employee in a similar way, and (ii) the effect of the FTA (= face threatening act) also depends on the attributes in question: as is rightly remarked in Buchenau 1997: 175, it matters whether body parts or clothes are involved or whether only items as an umbrella or a suitcase are mentioned. In a similar way, a metonymical transfer as in Seredina! or even Vperedi! is not likely to raise the

.  In Leech 1983, this correlation is formulated more carefully: it is now situated between the scales of indirectness and politeness. Indirectness in its turn is certainly related to the length of the utterance, as is suggested, for example, by the series of examples presented in Brown and Levinson (1987: 142–144), but it would be an overstatement to generalise this ­correlation.

 Daniel Weiss

addressee’s self-respect, although we may hear Moskva! in a museum addressing a group of visitors from Moscow who will not take offense at all. In our next two cases, the lack of subject is combined with verb omission. When using such pragmatic idioms as Tak by Ø2SG/PL srazu ØV!, the speaker has no choice between the familiar form of address ty and its polite (distant) counterpart vy. This has, however, no impact on the politeness management: all my informants agreed that both utterances can be addressed to other people without regard to on what terms (familiar or distant) they are with the speaker. What remains relevant is the illocutionary content, not the form of the utterance: we do not dare to direct criticism or warnings at addressees more powerful than ourselves. In a similar vein, a directive speech act with omission of the action verb does not sound more rude than its overt variants, cf. A musor Ø2SG/PL potom ØV[= vyneseš’/vynesete/ vynesi(te)!] Again, the lack of a person marker and hence the non-availability of the polite form is no obstacle. As for the additional choice between the imperative and the indicative in the overt formulation, one may even argue that the use of the perfective present vyneseš’/vynesete could be felt to cause more harm than the omission since it presupposes a very familiar or even a subordinative relationship between speaker and addressee. Finally, let us have a brief look at possessor ellipsis. It may be recalled that example (20) (Rasskažitej ėto Øi/j žene!) was ambiguous in that the wife’s husband may be identical with the speaker or the hearer if the preceding context or the discourse situation does not contain an explicit clue. According to my informants, it is, however, less polite to omit the pronoun of the second person than to employ it; in the case of the first person pronoun, its absence or presence has no similar effect. This is in line with the overall assumption that the addressee deserves more respect and therefore a “full treatment” in terms of person marking. To sum up, eliding reference to the addressee turns out to be potentially ­sensitive in terms of the politeness management, if there is no simultaneous verb omission involved. This provides only partial confirmation for the overall rule ‘the longer, the more polite’: surprisingly enough, the shorter structures with both ­subject and verb missing do not cause any face threatening acts. 5.  A glimpse at history: Avvakum’s Žizneopisanie The historical development of the different types of empty slots being discussed here is poorly studied. This is only partly due to the fact that colloquial language is seldom reflected in older written sources: apart from the famous birchbark d ­ ocuments, from the 17th century on we also have personal letters,



The lazy speaker and the fascination of emptiness 

conversation manuals (“razgovorniki”), narrative genres such as merchant tales, ­autobiographies, captions on “lubki”, etc. Of course, these documents seldom represent pure ­colloquial language but rather a peculiar mixture of formal and informal registers, which varies considerably from one document to another. Unfortunately, historical research often does not pay sufficient attention to these colloquial ­elements. Such an extraordinary text as Avvakum's Žizneopisanie contains a remarkable deal of colloquial morphosyntax,14 including not only articlelike postfixed forms of the demonstrative pronoun,15 historical present forms, the “illustrative” use (“primerno-nagljadnoe značenie”) of the perfective present, narrative infinitives (cf. “I ja paky svetu-Bogorodice dokučat’”) or reduplication (“Protopopica bednaja bredet-bredet, da i povalitsja – kol’zko gorazdo!”), but also all types of empty slots.16 Among the latter, the evolution of subject ellipsis is so far understood best. To begin with, the Novgorodian birchbark letters attest a peculiar compensatory mechanism in that in the 1st/2nd person perfect either the subject pronoun or the auxiliary must be missing, cf. Øi dalъ jesmьi or ja Øaux dalъ, but not *Ja dalъ jesmь (this latter variant however is frequent in chronicles); the same holds for sentences with the copula in the present, cf. Øi vinovatъ jesmьi or ja Øbyt’ vinovatъ. The 3rd person, however, is marked by a zero auxiliary anyway, which results in mere dalъ (cf. Zaliznjak 2004: 178–179). According to Meyer 2009,17 this picture begins to change with the shrinking use of the auxiliary/copula byt’, which eventually led to its complete loss in the 17th century. This process was accompanied by a spread of overt subject pronouns in the 1st/2nd person in written sources such as Avvakum’s Žizneopisanie or the Tale of Frol Skobeev. However, zero subject marking in the preterit remains frequent in Avvakum and even occurs with remote antecedents that may have been established four sentences before; besides this,

.  This text has been examined several times from a linguistic point of view, but the focus always lay on the distribution of Church Slavonic elements vs. vernacular vocabulary and morphology, see Sørensen 1957; Worth 2000 or Kuße 2009. Its colloquial layer was never analysed separately, although this promises to be a most rewarding task. .  This phenomenon has already been examined in a whole series of studies, the latest one being Mendoza 2011. .  It should be noted that the controversy about the authenticity of many parts of this document, whose autograph is not preserved (Scheidegger 1999), has no bearing on this analysis since Scheidegger argues that most of the presumable additions to the main text were made by contemporary Old Believers. .  For more details about the history of null subjects in Russian, see Meyer (unpublished).

 Daniel Weiss

reference of the subject ellipsis may switch to a new antecedent. Both cases are illustrated in the following excerpt: (46)  После тово вскоре схватав Никонi Даниила, в монастыре за Тверскими вороты, при царе остриг голову и, содрав однорятку, ругая, отвел в Чюдов в хлебню и, муча много, сослал в Астрахань. Венец тернов на главу ему там Øljudi возложили, в земляной тюрме и уморили. После Данилова стрижения Øljudi взяли другова, темниковскаго Даниила ж протопопа, и посадили в монастыре у Спаса на Новом. Таже протопопа Нероноваj Иванна – в церкве скуфью Øi снял и посадил в Симанове монастыре, опосле Øi сослал на Вологду, в Спасов Каменной монастырь, потом в Кольской острог. А напоследок, по многом страдании, Øj изнемог бедной – Øj принял три перста, да так и умер. (146)18

On the one hand, these uses of the ellipsis are reminiscent of more archaic, medieval narrations; on the other hand, however, they seem to reflect colloquial usage of that time. As for the general evolution outlined above, Meyer (2009) states that it is in line with the so-called verb impoverishment thesis, according to which the lack of person marking had to be compensated by a reinforcement of subject pronouns. As for the 3rd person, the overt pronoun appeared later, probably because the zero marking remained associated with this person for a longer period. As a result, formal Russian has evolved into a partial PRO-drop language in which pronouns have become more or less mandatory where grammatical person is not marked elsewhere, i.e. by verb forms in the same clause or the main clause. It has to be emphasised, though, that this holds only for formal speech: in spontaneous speech as well as in all written genres imitating it, the expression of person may be completely absent, as examples (6–10) have already shown. From a typological point of view, this result is most revealing, since it now turns out that spoken Russian has made a step back from a stage with full personal marking to a stage where the expression of this category has become optional. In terms of a traditional Humboldtian or Skalička typology, the evolution of the Russian preterit and the present with zero copula may be described as a partial transition from simple fusion (vinovatъ jesmь) or analytical fusion (dalъ jesmь) to isolation (ja vinovat, ja dal). But how should the new personless forms such as dal be treated? Are we to consider them as a backlash into the polysynthetic stage where expression of grammemes is optional? However this may be, the whole case clearly confirms my initial claim that from a typological point of view, colloquial Russian offers far more fascinating insights than Codified Standard Russian (KLJa).

.  The numbers in brackets refer to the pages in the edition by Robinson 1963. Cf. also feb-web.ru/feb/ avvakum/texts/rob/rob-139-.htm.



The lazy speaker and the fascination of emptiness 

Avvakum also offers an interesting case of possessor ellipsis which in modern Russian would probably be avoided, cf. (47)  В таже времена пришла ко мне с Москвы грамотка. Два брата жили у царицы вверху, а оба умерли в мор и з женами и з детми; и многия друзья и сродники померли. (148)

Does this refer to the author’s religious brethren (cf. the formula “brat’ja i srodniki naši”, which is still in use in modern prayers) or to his real brothers? The lack of closer identification would not even exclude an indefinite reading (‘two brothers’, ‘not my two brothers’). From the subsequent sentence, however, we may conclude that the author is referring to his personal sphere; besides this, in the preceding text there once was an implicit mention of two brothers by means of a comparative, cf. (48)  а брату моему меньшому бояроня Васильева и дочь духовная была. ­Так-то бог строит своя люди (144)

This is, however, too remote a reference to be remembered by the reader, and the exact number of Avvakum’s brothers remains a riddle.19 A cooperative speaker of contemporary Russian would avoid such incomplete wording and add a possessive marker, e.g. dva iz moix brat’ev. As for the zero lexeme Øljudi, it seems to be a fair guess that it has a Common Slavic genealogy. This does, however, not imply that all its uses attested in contemporary Russian may be traced back to common Slavic. In particular, the “indefinite-personal form” which comes to function as a quasi-definite zero with an inferred coreferential link to this antecedent (cf. example (25)) may well represent a Russian innovation. In the same way, the pragmatic idioms mentioned in Section 1.2 are probably of recent origin. As for the structural ambiguity in contexts allowing both an interpretation as a zero or a subject ellipsis, it seems worth noting that they also occur in Avvakum’s Žizneopisanie, as is shown by the ­following example where those who arrested Avvakum might not be the same people as those who transported him to the Andron’ev Monastery: (49)  Таж меня взяли от всенощнаго [Борис Нелединской со стрелцами]i; человек со мною шестьдесят Øi взяли: их в тюрму Øi отвели, а меня на патриархове дворе на чеп Øi посадили ночью. Егда ж розсветало в день неделный, Øi/Øljudi посадили меня на телегу, и ростянули руки, и везли от патриархова двора до Андроньева монастыря, и тут на чепи кинули в темную полатку, ушла в землю, и Øj сидел три дни, Øj ни ел, ни пил; во тме сидя… (146–7)

.  As Robinson (1963: 241) notes, the author may have had as many as four brothers.

 Daniel Weiss

Moreover, the last sentence of this citation presents another instance of elliptical subject change: the last overt antecedent of Øj is menja. In modern colloquial ­Russian, such a subject switch would rather be avoided. The zero lexeme Østixii postulated as subject of “impersonal” forms such as rvet, uneslo, etc. likewise is a reliable candidate for Common Slavic origin. It occurs frequently in the narrative parts of Avvakum’s Žizneopisanie, for example in his depiction of the sufferings he had to endure when being exposed to the disastrous forces of nature in his Siberian exile, cf. (50) Н  а том же Хилке в третьее тонул. Барку от берегу Østixii оторвало водою, – людские стоят, а мою Østixii ухватило, да и Østixii понесло! Жена и дети остались на берегу, а меня сам-друг с кормщиком Østixii помчало. Водаi быстрая, Øi/Østixii переворачивает барку вверх боками и дном… (151)

This excerpt also contains a denotation of the immediate cause in the instrumental, cf. otorvalo vodoj. Moreover, perevoračivaet is ambiguous in that its subject could also be an anaphoric ellipsis of voda. Verb omission is already attested in a XIVth century Novgorodian agreement and in Afanasij Nikitin’s ‘Journey beyond the Three Seas’ (Kopotev 2008: 71). In Avvakum it occurs several times. The following fragment contains 2 instances of ØV: (51)  указал на Никона митрополита. Царь ево и послушал, и пишет к нему послание навстречю: преосвященному митрополиту Никону и новгороцкому и великолуцкому и всеа Русии радоватися, и прочая. Егда ж приехал, с нами ØV, яко лис: челом ØV да здорóво! (146)

In the case of s nami jako lis one hesitates whether to interpret this as an instance of the zero copula or a missing action verb, such as povel sebja. Čelom is the remainder of the pragmatic idiom b’ju čelom; this truncated form occurs four times in the whole text. On the whole, we thus may state that all major types of emptiness already occur in Avvakum’s Žizneopisanie, some of them being subject to even fewer constraints than today. Remarkably, verb omission is also attested, which, as one may surmise, is a Russian innovation; for its lack in other Slavic languages, see next section. 6.  A first comparison to Polish and Czech Literature on empty reference and predication in other Slavic languages is extremely scarce: besides Seo 2001 on the frequency of null subject in five



The lazy speaker and the fascination of emptiness 

­ ifferent Slavic languages, McShane 1999 on object ellipsis in Russian, Polish d and Czech and McShane 2000 on verb omission in the same three languages, there are only some preliminary observations on subject ellipsis in Meyer 2009, verb omission in Russian vs. Czech in Mažara 2010 and in Russian and Polish in Weiss 2011a, 2011b available; additional comparative remarks may be found in Berger 1991 and Weiss 1993. To make things worse, most of these studies do not focus on colloquial usage. Thus, this section can but outline a very sketchy and unbalanced overview of the Polish and Czech equivalents of the phenomena discussed so far. Subject ellipsis is expected to be more widespread, if not regular in West Slavic due to the sufficient morphological marking of person and number in the verb forms. However, as has been shown above, in colloquial Russian the lack of person marking in the preterit and conditional (and in the present with zero copula) is no obstacle to elliptical subjects, and coreferential intrasentential subject ellipsis is even stylistically unmarked. On the other hand, spoken Polish is quite permissive of overt subject pronouns in positions where these would be avoided in formal registers (Pisarkowa 1969; Nilsson 1992), and the same seems to hold for spoken Czech (Meyer 2009). On the whole, the difference between the three spoken varieties may well turn out to be insignificant in this respect. As for direct object ellipsis, the systematic overwiev by McShane 1999 reveals an evident pattern: Russian allows for the largest range of ellipsis, Polish slightly less so, whereas Czech avoids it most often. This study does not differentiate between formal and colloquial registers, which, however, does not seem to matter significantly for object ellipsis. Moreover, it devotes much space to coordination, which, as may be recalled, was excluded from the present study. The speaker-exclusive zero Øljudi does exist in West Slavic, but it is far more restricted than in Russian.20 First of all, unlike its Russian counterpart it is markedly colloquial (it would for example not be used in scientific discourse), second, it has a potent rival, viz. the construction with the reflexive pronoun se/się used in indefinite personal sentences, which provides the most frequent equivalent of Øljudi in translations from Russian; this pattern is absolutely regular in ­Polish where unlike in Czech it embraces also transitive verbs with overt objects and the copula. Moreover, the Polish forms in -no/-to of the type wydzielono, zamknięto offer a third equivalent for the preterit and the conditional.21 Third, the expansion of the indefinite personal model with reflexive marker in West Slavic may

.  For Polish see Bogusławski 1984, for Czech Berger 1991 for more details, see Meyer (unpubl.). .  For all three variants of indefinite personal verb forms in Polish, see Puzynina 1993.

 Daniel Weiss

be responsible for the absence of the generic and speaker-including use of Øljudi, which seems to be an East Slavic innovation; see Meyer (unpubl.), ch. 5.5.3.2. And finally, in cases such as Russian Sjuda idut or Vam zvonjat the indefinite pronoun imposes itself in both West Slavic languages, cf. Ktoś tu idzie, Někdo sem jde; Do pani ktoś dzwoni, Někdo vám volá. On the whole, the Czech and Polish zero subject Øljudi cannot introduce new ­referents, cf. “In an overwhelming majority of the examples 3rd person plural refers to a group of people specified locally (40%), temporally (2%) or by the semantics of the predication and/or the wider context (51%), or by several of these factors” (Dušková 1973, 16 f.). Thus, the label ‘indefinite personal’ turns out to be inappropriate for West Slavic Øljudi. The lesser functional load of the zero subject in West Slavic also explains the lack of such particularities as the (inferred anaphoric) verbal interaction attested in example (26) or the pragmatic idioms with unique reference to the speaker such as tebé govorjat or vas bespokojat iz… As for the speaker-inclusive generic ‘you’, Russian allows for two special uses which would be unacceptable in West Slavic: the simultaneous appearance of Øty2 and the addressing ‘you’ as in example (34) as well as the combination with an addressee who requires the honorific vy as in (37). The impersonal zero Østixii may in principle be combined with the designation of the cause in the instrumental in Polish, but not in Czech. But even in Polish, many literal translations of Russian examples such as (30–31) are rejected by informants. The intermediate position of Polish on the scale of preference for syntactical emptiness is confirmed by the distribution of the zero copula in the present tense: whereas in Russian this empty verb form is simply not replaceable, its Polish equivalent occurs only in strictly defined syntactical contexts (mostly indeclinable predicatives and the pronominal subject to) and is most often optional, and in Czech this type of zero is practically obsolete. Verb omission (ØV) requires a differentiated treatment. In both Polish and Czech, this type is far more restricted and less frequent than in colloquial Russian. Thus, Mel’čuk’s rule that any verb of action may be omitted provided there remains a bipartite syntactical skeleton is utterly unthinkable in West Slavic, and the literal translation of Russian examples such as Ty ėto sama?, Kak tebe X?, On domoj; Ty skoro?, Ja nenaročno!, etc. provides ungrammatical Czech and Polish equivalents. On the other hand, Polish and (maybe to a lesser extent) also Czech admit the omission of verbs of motion, communication and physical violence quite freely in the narrative register (both in written and oral form), where it often serves to accelerate and dramatize the ongoing narration, but not in the discourse register (Weiss 2011b; 2011c). Besides this, Polish offers a rich choice of idioms with optional missing verbs of communication: for example, the construction X ØV na Y z mordą ‘X starts insulting Y’ may be varied in not less than 28 ways due to the replacement of morda by



The lazy speaker and the fascination of emptiness 

other nouns denoting the mouth, cf. z gębą, z pyskiem, z buzią, and the substitution of metaphorical verbs of motion such as wyskakiwać, wylatywać, wyjeżdżać, napadać, rzucać się and wsiadać. And finally, if utterances like Ty to sama? On nie o tym, Ja tylko na chwilę do not occur in spoken Polish, they are now well attested in ­Internet forums, blogs and chats. In this way, we observe the emergence of new verbless sentence patterns in Polish Internet communication that are strongly reminiscent of colloquial Russian. Nevertheless, the unrestricted character of verb omission is a quirk that probably distinguishes colloquial Russian not only from its Slavic cognates but also from all other European language families including the Baltic and Finno-Ugric neighbours. At the distinction of referential emptiness, even Chinese does not know a similar phenomenon (W. Bisang, p.c.).22 7.  Conclusions To sum up, our overview of different syntactic gaps in modern colloquial Russian (Section 1) has not only revealed an impressive diversity, but also a whole range of quirks not found elsewhere in Europe. The most remarkable among them is the possible omission of almost any action verb, followed by the omissibility of head nouns, which leave behind ‘stranded’ modifiers (e.g. v šube = ženščina v šube), and of relative pronouns in varying syntactic roles. Other phenomena are less constrained than elsewhere, e.g. the zero copula, the referential zero Øljudi and the generic ‘you’ Øty2. Moreover, different empty slots may co-occur within the same clause and thus combine into empty clusters. All this has a considerable impact on the balance of text economy and explicitness (Section 2). Surprisingly enough, it does not endanger politeness management: even the omission of a head noun denoting the addressee is admissible in certain social settings (Section 3). As our historical digression (Section 4) has shown, all markedly colloquial empty slots discussed in Section 1 occur already in Avvakum’s ‘autobiography’. And finally, the overall assumption formulated in Section 1 is corroborated by evidence from West Slavic languages (Section 5): modern colloquial Russian shows a marked preference for the zero strategy in many referential and predicative uses where speakers of Polish and Czech utilise overt pronouns or verbs. Thus, colloquial Russian turns out to be located closer to the extreme pole on the scales of both textual economy and implicitness, or else: the “East Asian solution” of grammatical marking. Interestingly enough, the same “Eurasian orientation” of colloquial Russian

.  On the other hand, in Salishan “all types of full sentences can make it without a verb” (Mel’čuk 2003: 207 f).

 Daniel Weiss

also manifests itself in a completely unrelated innovation, viz. the beginnings of verb serialisation; cf. Weiss 2012.

Abbreviations Øbyt’ =  zero copula ‘to be’ in the present ØСonn = empty connection Øi/j/k =  coreferential ellipsis of subject/object NP Øirel =  ellipses of relative pronom Øljudi =  zero indefinite personal lexeme ‘people’

Øposs = possessor’s ellipses Østixii = zero indefinite impersonal lexeme ‘elements’ Øty2 =  zero lexeme for generic ‘you’ ØV = omission of verbs of action ØVi =  anaphoric non-coordinative ellipsis of verbs.

References Apresjan, Jurij D., Iomdin, Leonid L. & Percov, Nikolaj V. 1978. Ob”ekty i sredstva modeli poverxnostnogo sintaksisa russkogo jazyka. Makedonski Jazik, g. XXIX: 125–171. Apresjan, Jurij D. 1986. Tipy sintaksičeskogo otsutstvija v russkom jazyke. In Problemy semantiki predloženija: Vyražennyj i nevyražennyj smysl, 111–114. Krasnojarsk. Artyushkina, Olga. 2010. Le discours indirect libre en russe. Thèse de doctorat, Université Paris IV-Sorbonne. Berger, Tilman. 1989. Die transphrastische koreferentielle Subjekts- und Objektsellipse im Russischen. In Slavistische Linguistik 1988, Wolfgang Girke (ed.), 9–34. München: Otto Sagner. Berger, Tilman. 1991. Zum denotativen Status “unbestimmt-persönlicher” Konstruktionen im Tschechischen. In Words are Physicians for an Ailing Mind, Festschrift für A. Bogusławski, Maciej Grochowski & Daniel Weiss (eds), 71–81. München: Otto Sagner. Bisang, Walter. 2009. On the evolution of complexity – sometimes less is more in East and mainland Southeast Asia. In Language Complexity as an Evolving Variable, Geoffrey Sampson, David Gil & Peter Trudgill (eds), 34–49. Oxford: OUP. Bisang, Walter. 2011. Sprachliche Komplexität und Arealität: Ostasien, Westafrika und Kreolsprachen. Paper held at the German Department, University of Zurich, December 15. Bogusławski, Andrzej. 1984. Polskie nieidentyfikacyjne wyrażenia osobowo-referencjalne. Polonica 10: 49–69. Breuer, Astrid Y. 2002. Asyndese? Zum Problem einer ‘negativen’ Kategorie. München: Otto Sagner. Brown, Penelope & Levinson, Stephen C. 1987. Politeness: Some Universals in Language Usage. Cambridge: CUP. Buchenau, Klaus. 1997. Die Distanzanrede im Russischen, Polnischen und Deutschen und ihre historischen Hintergründe. Frankfurt: Peter Lang. Bulygina, Tatiana V. & Šmelev, Aleksej D. 1997. Jazykovaja konceptualizacija mira (na materiale russkoj grammatiki). Moskva: Jazyki russkoj kul’tury.



The lazy speaker and the fascination of emptiness 

Dušková, Libuse. 1973. Man-Sätze in Czech and English. Philologica Pragensia 16: 5–37. Fortuin, Egbert. 2011. Iconicity, economy and frequency: the paratactic conditional perfective present construction in Russian. Russian Linguistics 35(1): 89–123. Grenoble, Leonore. 2001. Conceptual reference points, pronouns and conversational structure in Russian. Glossos 1(1). 〈http://www.seelrc.org/glossos/issues/1/grenoble.pdf〉. Guiraud-Weber, Marguerite. 1984. Les propositions sans nominatif en russe moderne. Paris: ­Institut d’Etudes Slaves. Guiraud-Weber, Marguerite. 1990. La structure de la personne indéterminée: Le sujet zéro en russe et le pronom on en français. Revue des Etudes slaves 62(1–2): 197–209. Heine, Lena. 2011. Non-coordination-based ellipsis from a construction grammar perspective: The case of the coffee construction. Cognitive Linguistics 22(1): 55–80. Iordanskaja, Lidia & Mel’čuk, Igor. 2011. Illocutive parenthetical verbs in Russian. In Proceedings of the 5th International Conference on Meaning-Text Theory (MTT’11), Barcelona, Igor M. Boguslavsky & Leo Wanner (eds), 120–133. 〈http://meaningtext.net/mtt2011/ proceedings/〉. Kibrik, Aleksandr A. 1996, Anaphora in Russian narrative prose: A cognitive calculative account. In Studies in Anaphora [Typological Studies in Language 33], Barbara A. Fox (ed.), 255–303. Amsterdam: John Benjamins. Kibrik, Aleksandr A. & Prozorova, Evgenija B. 2007. Referencial’nyj vybor v russkom žestovom jazyke.< http://www.dialog-21.ru/dialog2007/materials/html/34.htm>. Kopotev, Mixail V. 2008. Principy sintaksičeskoj idiomatizacii. Xel’sinki. Krifka, Manfred. 2002. Be brief and vague! And how Bidirectional Optimality theory allows for verbosity and precision. In Sounds and Systems. Studies in Structure and Change. A Festschrift for Theo Vennemann [Trends in Linguistics 141], David Restle & Dietmar Zaefferer (eds), 439–458. Berlin: Mouton de Gruyter. Kručinina, Irina N. 1968. Konstrukcii s mestoimeniem “kotoryj” v sovremennom russkom jazyke. Voprosy jazykoznanija 2: 82–88. Kuße, Holger. 2009. Sprachliche Markierungen des Religiösen bei Avvakum. In Religion und Rhetorik, Holt Meyer & Dirk Uffelmann (eds), 65–81. Stuttgart: Kohlhammer. Lapteva, Olga A. 1976. Russkij razgovornyj sintaksis. Moskva: Nauka. Leech, Geoffrey N. 1977. Language and Tact. Linguistic Agency: University of Trier, Series A, Paper No. 46. Leech, Geoffrey N. 1983. Principles of Pragmatics. London: Longman. Levinson, Stephen C. 2000. Presumptive Meanings: The Theory of Generalized Conversational Implicature. Cambridge MA: The MIT Press. Мažara, Jekaterina. 2010. Swiss cheese and the lazy speaker: The omission of verbs in Russian and Czech. Oslo Studies in Language 2(1): 231–242. Мažara, Jekaterina. 2011. Missing verbs in Colloquial Russian: Vagueness or ambiguity? In Proceedings of the 2010 POLYSLAV Conference. Munich: Otto Sagner. McShane, Marjorie J. 1999. The ellipsis of accusative direct objects in Russian, Polish and Czech. Journal of Slavic Linguistics 7(1): 45–88. McShane, Marjorie J. 2000. Verbal Ellipsis in Russian, Polish and Czech. The Slavic and East European Journal 44(2): 195–233. Mel’čuk, Igor. 1974. O sintaksičeskom nule. In Tipologija passivnyx konstrukcij. Diatezy i zalogi, Aleksander A. Xolodovič (ed.), 343–361. Leningrad: Nauka. Mel’čuk, Igor. 1995. Russkij jazyk v modeli ‘Smysl ¤ tekst’ [Wiener Slawistischer Almanach. Sonderband 39]. Wien.

 Daniel Weiss Mel’čuk, Igor. 2003. Levels of dependency in linguistic description: Concepts and problems. In Dependency and Valency. An International Handbook of Conemporary Research. Vilmos Agel, Ludwig M Eichinger, Hans-Werner Eroms, Peter Hellwig & Hans-Jurgen Heringer (eds), 187–229. Berlin: De Gruyter. Mendoza, Imke. 2011. On demonstrative pronouns and the question of the article in 17th ­century Russian. Russian Linguistics 35(2): 245–265. Meyer, Roland. 2009. Zur Geschichte des referentiellen Nullsubjekts im Russischen. Zeitschrift für Slawistik 54: 375–397. Meyer, Roland. 2012. (unpublished). The History of Null Subjects in North Slavonic. A C ­ orpusBased Diachronic investigation. Habilitation Thesis. University of Regensburg, Germany, May 2012. Mikaelian, Irina. 2002. La possession en russe moderne. Eléments pour la construction d’une catégorie sémantico-syntaxique. Paris: ANRT. Nilsson, Barbro. 1992. Personal Pronouns in Russian and Polish. A Study of their Communicative Function and Placement in the Sentence. Stockholm. Pisarkowa, Krystyna. 1969. Funkcje składniowe polskich zaimków odmiennych. Wrocław: Ossolineum. Puzynina, Jadwiga. 1993. Die sogenannten “unbestimmt-persönlichen” Formen in der polnischen Sprache. In Studies in Polish Morphology and Syntax, Gerd Hentschel & Roman Laskowski (eds), 31–61. München: Otto Sagner. Raxilina, Ekaterina V. (ed.). 2010. Lingvistika konstrukcij. Moskva: Azbukovnik. Robinson, Andrej N. 1963. Žizneopisanija Avvakuma i Epifanija. Issledovanie i teksty. Moskva: Izd. AN SSSR. Saj, Sergej. 2002. Elliptičeskie konstrukcii: Struktura i funcionirovanie (na materiale russkogo jazyka). Diplomnaja rabota (neopubl.). St. Peterburg. Ščadneva, Valentina. 2000. Diskursivno-obuslovlennye neverbalizovannye komponenty vyskazyvanija. Tartu. Scheidegger, Gabriele. 1999. Endzeit. Russland am Ende des 17. Jahrhunderts. Frankfurt: Peter Lang. Seo, Seunghyun. 2001. The Frequency of Null Subject in Russian, Polish, Czech, Bulgarian, and Serbo-Croatian: An Analysis According to Morphosyntactic Environments. Bloomington IN: Indiana University. Širjaev, Evgenij N. 1973. O nekotoryx pokazateljax nezameščennyx sintaksičeskix pozicij v vyskazyvanijax razgovornoj reči. In Russkaja razgovornaja reč’, Elena Andreevna Zemskaja, Margarita Vasil'evna Kitajgorodskaja & Evgenij N. Širjaev (eds), 288–317. Moskva: Nauka. Širjaev, Evgenij N. 1986. Bessojuznoe složnoe predloženie v russkom jazyke. Moskva: Nauka. Šmelev, Aleksej D. 1998. Tipy “nevyražennyx valentnostej“. Semiotika i informatika 36: 167–176. Šmelev, Aleksej D. 1996. Referencial'nye mexanizmy russkogo jazyka. Tampere [= Slavica Tamperensia IV]. Sokolova, Svetlana & Makarova, Anastasia. 2011. Čto-to kak-to neponjatno: tipologija kontekstov diskursivnogo upotreblenija. Paper held at the Conference Russkij jazyk: konstrukcionnye i leksiko-semantičeskie podxody, St.Petersburg, March 24–26. Sørensen, Hans C. 1957. Die stilistische Verwendung kirchenslavischer Sprachelemente in der Autobiographie Avvakums. Scando-Slavica 3: 154–175. Stojnova, Natalia. 2011. Opjat’/večno vy so svoimi konstrukcijami! Paper held at the Conference Russkij jazyk: Konstrukcionnye i leksiko-semantičeskie podxody, St. Petersburg, March 24–26.



The lazy speaker and the fascination of emptiness 

Stoll, Sabine & Bickel, Balthasar. 2009. How deep are differences in referential density? In Crosslinguistic Approaches to the Psychology of Language. Research in the Tradition of Dan Isaac Slobin, Jiansheng Guo, Elena Lieven, Nancy Budwig, Susan Ervin-Tripp, Keiko Nakamura & Seyda Ozcaliskan (eds), 543–555. New York NY: Psychology Press. Weiss, Daniel. 1993. Die Faszination der Leere. Die moderne russische Umgangssprache und ihre Liebe zur Null. Zeitschrift für slavische Philologie LIII(1): 48–82. Weiss, Daniel. 2008. L’autoréférence en russe contemporain. In Questions de linguistique slave.  Études offertes à Marguerite Guiraud-Weber, Robert Roudet & Charles Zaremba (eds), 367–382. Aix-en-Provence: Publications de l’Université de Provence. Weiss, Daniel. 2011a. Bezglagol’nye konstrukcii russkoj razgovornoj reči: ix tipologija i status v lingvističeskom opisanii. In Slovo i jazyk. Sbornik statej k vos’midesjatiletiju Ju.D. ­Apresjana. Igor M. Boguslavskij, Leonid L. Iomdin & Leonid P. Krysin (eds), 139–155. Moskva: Jazyki slavjanskix kul’tur. Weiss, Daniel. 2011b. Dekapitacja zdań w polszczyźnie współczesnej. O konstrukcjach ­bezczasownikowych i ich odpowiednikach rosyjskich. Poradnik Językowy 1, Festschrift for A. Bogusławski, 111–121. Weiss, Daniel. 2011c. Pralinko, ja do Ciebie żartem, a Ty poważnie do mnie czyli o pozornej rusyfikacji polskiej komunikacji internetowej. Prace filologiczne LX, Festschrift for M. ­Grochowski, 299–319. Weiss, Daniel. 2012. Verb serialization in North East Europe: The case of Russian and its FinnoUgric neighbours. In Grammatical Replication and Grammatical Borrowing in Language Contact. Björn Wiemer, Bernhard Wälchli & Björn Hansen (eds), 619–655. Berlin: Mouton de Gruyter. Wiemer, Björn. 1996. Klassifikacija nulevyx skazuemyx v russkom jazyke po ix leksičeskim i referencial’nym xarakteristikam. Studia z filologii polskiej i słowiańskiej 33: 245–273. Wierzbicka, Anna. 1988. The Semantics of Grammar [Studies in Language Companion Series 18]. Amsterdam: John Benjamins. Woolhiser, C. 1990. “Missing prepositions” in colloquial Russian relative clauses. In Topics in Colloquial Russian, Margaret H. Mills (ed.), 17–30. Frankfurt: Peter Lang. Worth, Dean S. 2000. Avvakum as Az i Ja. Južnoslovenski filolog LVI: 227–233. Zaliznjak, Andrej A. 2004. Drevnenovgorodskij dialekt. Moskva: Jazyki russkoj kul’tury. Zdorenko, Tatiana. 2010. Subject omission in Russian: A study of the Russian National Corpus. In Corpus-linguistic Applications. Current Studies, New Directions, Stefan T.  Gries, ­Stefanie Wulff & Mark Davies (eds), 119–133. Amsterdam: Rodopi. Zemskaja, Elena V. et al. (eds). 1973. Russkaja razgovornaja reč’. Moskva: Nauka. Zemskaja, Elena V. 1987. Russkaja razgovornaja reč’: Lingvističeskij analiz i problemy obučenija. Moskva: Russkij jazyk.

Is the Polish verb iść an auxiliary to be? A corpus-based study of the construction iść + infinitive* Dorota Sikora

ATILF-CNRS (UMR 7118), Nancy, France The aim of the present paper is to investigate the degree of grammaticalization displayed by the Polish verb iść ‘to go’ in its infinitive constructions. A significant number of such VP show features usually associated with ongoing auxiliarization (ambiguity in context, desemanticization, released selection constraints and loss of grammatical properties of the verb). However, semantic bleaching is far from being complete as long as antonyms of iść are banned from infinitive complement position. Considering available corpus data, we assume that the grammaticalization of iść in its infinitive constructions has presently attained the stage of bridging contexts (Heine 2002). Keywords:  grammaticalization; auxiliarization; Polish motion verbs; semantic change

1.  Introduction The present paper deals with the infinitive constructions (henceforth IC) of the Polish verb iść ‘to go’, focusing on those that display clear signs of ongoing auxiliairization. In its basic sense, iść denotes the motion event of walking, performed under agentive control of the walker. However, as illustrated in the example (1), in some ICs, the lexical meaning seems replaced by another: (1) X idzie siedzieć. X go:3sg be-in-prison:inf ‘X is going to be sent to prison’

(Internet)

*  I am very grateful to anonymous reviewers for the fruitful remarks to a previous version of this paper.

 Dorota Sikora

In the example (1), iść is by no means motion verb. As shown in Sikora (forthcoming), such an IC cannot be reduced to a simple goal expression.1 It turns to assume a new role, roughly speaking that of future marker of siedzieć ‘to go to prison’: by uttering (1), the speaker indicates that the person X will be in prison before long. That is to say that the spatial meaning of iść has been replaced with a functional one. The question arising here is to what extent iść behaves as a grammatical marker and, possibly, an auxiliary. 1.1  Polish future tenses Polish grammatical system offers two ways of expressing future temporal reference. The first one relies on a complex structure composed of the future form of the auxiliary być ‘to be’ followed (or preceded) by the infinitive or the past of the main verb. Such a future tense characterizes imperfective verbs as, for example, czekać ‘to wait’. In order to say that one will be waiting, both – male and female Polish speakers – are free to state: będę czekać (be:fut.1sg+ipf.inf), as well as they have possibility to mark gender using past form of the main verb: będę czekał (be:fut.1sg + ipf.pst.m) or będę czekała (be:fut.1sg + ipf.pst.f). Perfective verbs, as poczekać ‘to wait (pf)’, express future reference by using simple forms. Although from purely synchronic point of view, they could be considered present exponents, semantically they only convey future meaning. The non-actual present reading, currently expressed by such simple forms in some South Slavic languages (Bulgarian, Serbo-Croatian, Macedonian), is no more relevant for speakers of contemporary Polish, although it still may be identified in a fine-grained semantic analysis.2 1.2  Auxiliarization – a case of grammaticalization Scholars generally agree to define grammaticalization as linguistic change consisting in a transfer of a lexical unit from lexicon to grammatical system of the language (cf. Heine et al. 1991). In a broader sense, grammaticalization designates the study of the variational process mentioned above.

.  As a matter of fact, unlike the subordinate clause Idę do biblioteki, żeby się pouczyć ‘I am going to the library to study a bit’, the IC under analysis does not allow separate temporal and spatial localization for the event described in the main and in the subordinate clause. .  For an exhaustive discussion concerning Polish future forms, see Vetters & Skibińska 1998.



Is the Polish verb iść an auxiliary to be? 

In the present paper, the term grammaticalization is used in the first, restricted meaning, Consequently, I assume that auxiliarization is one of possible ways of grammaticalization relevant for verb classes: auxiliarized verbal lexeme progressively develops into auxiliary acquiring a new, partially or fully, grammatical content. Following that definition, grammaticalization (i.e. auxiliarization) is a gradual diachronic unidirectional3 process with uncertain results. As noticed in Brinton & Traugott (2005: 26): “… textual evidence s­uggests that many changes involve periods of relative indeterminacy in which it is not clear whether the older or the newer usage is in evidence (…) Indeed, the first steps may never result in change, in the sense of acceptance by a community of speakers”. Following Heine (2002), the graduality of grammaticalization process ­materializes in four stages. During the initial phase the verb fully preserves its lexical content in all uses. Some ambiguities appear in the second stage, that of bridging contexts, in which new inferences, open the way for a new significance. A switch contexts phase may then occur, when the evolving item starts an independent life in contexts incompatible with the source meaning. The final stage of c­ onventionalization is characterized by a double life of the word: on the one hand, it remains part of the lexicon as content verb, on the other, it has another existence in grammar, as it happened to have in English and to avoir ‘to have’ in French. 2.  How the motion verb iść is special “The moment a verb is given an infinitive complement, that verb starts down the road of auxiliariness” – says Bolinger (1980: 297). However, although the semantic chain MOTION > ASPECT > TIME seems largely applicable across languages,4 any single motion verb does not necessarily become a future marker. Before we focused on iść, let us get a closer look at the whole set of Polish motion verbs. Forsyth’s (1970) distinction between determinate and indetermi-

.  As shown in Prévost 2003, although unidirectionality is a definitional property of grammaticalization, it does not preclude reversibility of the process in some cases. See also Croft (2000, chapter 8). .  Heine & Kuteva 2002 quote fourteen languages across the world in which the source verb of the future auxiliary is a motion verb. See also Marchello-Nizia 2006 (especially Chapter 3), Croft 1990.

 Dorota Sikora

nate Russian motion verbs holds in Polish. The Table 1 portrays semantic relations between the two series of items forming corresponding classes.

Table 1.  Polish determinate and indeterminate motion verbs Determinate verbs

Indeterminate verbs

Translation

1.

iść

chodzić

‘to go, to walk’

2.

jechać

jeździć

‘to ride’

3.

biec

biegać

‘to run’

4.

lecieć

latać

‘to fly’

5.

frunąć

fruwać

‘to fly 〈birds〉’

6.

płynąć

pływać

‘to swim, to sail’

7.

pełznąć

pełzać

‘to crawl’

8.

leźć

łazić

‘to climb’

In Polish, the so-called indeterminate verbs (see Table 1) allow infinitive complements, but their constructions convey iterative meaning as in (2): (2) Ale mam już swoją ulubioną restaurację, do but have:1sg already own:acc favourite:acc restaurant:acc to której chodzę jeść. which:gen go:1sg eat:inf. ‘But I already do have my favorite restaurant where I usually go to eat.’ (Internet)

As Traugott (2010: 98) puts it: “there must be compatibility between the earlier (“source”) and the later (“target”) meaning”. The question to be answered is whether the semantic components of iść encourage the evolution of this verb towards grammatical uses? The issue is interesting because iść, as well as all Polish motion verbs (except for some stylistically marked and somewhat constrained items as przemieszczać się ‘to move’), are manner verbs. They differ then at that point from the English verb to go or the French aller ‘to go’, considered path verbs, without any specification of manner the motion is carried out. Nevertheless, recent studies in space semantics have shown that Talmy’s (2000) typological bipolarity is rather an idealized conceptualization. In fact, different languages appear as much or less close to one of the two edges formed by the Path lexicalization pattern on one side, and the Manner lexicalization pattern on the other (for details, see Levin, Beavers & Shiao Wei 2009).



Is the Polish verb iść an auxiliary to be? 

The semantic specificity of Polish determinate verbs has been described in terms of linearity (Laskowski 1999) or directionality (Włodarczyk 1980). Given available data, I suggest that determinate verbs, including iść, display a very ­particular lexicalization pattern conflating manner and path (Sikora 2009). Such a semantic model is crucial for aspectual properties of determinate Polish motion and transport verbs as long as they denote an event progressing along a spatially ordered Path, possibly towards a goal. For further discussion of the semantics of motion verbs, see Tenny 1995. The goal complement position may be fulfilled with an infinitive. Concomitantly, the manner lexicalized by iść relies only on physical performances of agentive subject. These semantic ­components – progressive motion and a particular manner – imply a third one, namely intentionality in a double sense. The motion denoted by iść is intentional, first, because the walker is supposed to control his/her motion, second, by reason of the goal-oriented path. The subject intention is then to reach the goal. The link between intention and future action is clear enough: by uttering something like the example (3): (3) Idę dzwonić na policję. go:1sg call:inf to police:acc ‘I am going to call the police.’

(NKJP)5

The speaker signals the intention of calling. Let us observe however that iść does not really need infinitive complement to express subject’s intention. In the example (4), iść is used as pragmatic marker of intentionality with reference to the action announced in the right – future – context. (4) Idę, napiszę smsa, będę czekał go:1sg write:pf.1sg sms:acc, be:fut.1sg wait:ipf.pst.m na odpowiedź. for answer:acc ‘Here I go, I will write a text message and I will wait for an answer.’ (Internet)

In sum, these three semantic features of iść provide a perfect ground for the pragmatic procedure that Traugott & König (1991), Traugott (2010: 101) refer to as invited inferencing: “Say no more than you must, and mean more thereby”.

.  Narodowy Korpus Języka Polskiego (Polish National Corpus).

 Dorota Sikora

3.  Evidence of ongoing auxiliarization The example (1) was the first insight of the change undergone by iść. It is time now to turn toward corpus data searching for linguistic evidence of ongoing ­auxiliarization. The present section focuses on pragmatic ambiguity in context (3.1) and semantic bleaching (3.2). 3.1  Pragmatic ambiguity in context Corpus data show a number of examples with two or three possible readings. Let us recall that, according to Heine 2002, such ambiguity is typical of bridging contexts. A standard request to Polish National Corpus (Narodowy Korpus Języka Polskiego, NKJP) brings up 49% of ICs with future reading, while in 23.3% of occurrences, iść may be thought of as well as modal marker of intentionality and temporal marker of future. In some rare sentences (3%), all three readings – spatial, modal and ­temporal – remain available, as illustrated in the example (5) below: (5) Panie, co pan robi, idę zawołać policjanta. Mister:voc what Mister:nom do:3sg go:1sg call:inf policeman:acc. ‘Hey, Mister, what are you doing, I am going to call a police officer.’ (NKJP)

3.2  Semantic bleaching (desemanticization) Scholars still lack agreement as to the nature of desemanticization. Should we define it as semantic bleaching, i.e. as loss of lexical meaning of the source verb, or rather in terms of its generalization as claimed in Bybee (2003)? However, beyond current discussions, there seems to be an agreement on how desemanticization of a lexical unit may come to light. 3.2.1  Releasing selection constraints Used as content verb, i.e. as full lexeme referring to a motion event, iść selects dynamic infinitives at complement position. But more the auxiliarization process advances, wider the range of selected complements can be. As a matter of fact, corpus data clearly show that iść selects a varied set of infinitives by no means dynamic. The most frequent among them denote emotions (tęsknić ‘to long for’, rozpaczać ‘to despair’, zapomnieć ‘to forget’, etc.), intellectual processes (myśleć ‘to think’, analizować ‘to analyze’, zastanawiać się ‘to wonder’, wątpić ‘to doubt’, etc.), inactivity (odpoczywać ‘to rest’, leniuchować ‘to laze’), postures (leżeć ‘to lie’), body processes (spać ‘to sleep’, trawić ‘to digest’). The examples (6) and (7) ­provide



Is the Polish verb iść an auxiliary to be? 

instances of ICs in which weakened selection constraints do not block verbs incompatible with spatial meaning of iść. (6) Tak więc (i) idę zapomnieć o osłabieniu i złym So then (i) go:1sg forget:inf about feeling-weak:loc and bad:loc samopoczuciu i (ii) idę zrobić coś dla siebie… feeling:loc and (ii) go:1sg do:pf.inf something:acc for myself:gen ‘So, I am going to forget feeling weak and bad, and I am going to do something for myself…’ (Internet) (7) Pożarłam trzy kawały kurczaka i zagryzłam devoured:pf.f three:acc big-pieces:acc chicken:gen and bit-into:pf.f jogurtem. Idę trawić i myśleć (pozytywnie): “Jestem yogurt:inst. go:1sg digest:inf and think:inf (positive.adv) be:1sg na dukanie”. on dukan-diet:sg.loc ‘I devoured three big pieces of chicken and a yogurt. Now I am going to digest it and to think positive: “I am following the Dukan diet”. (Internet)

Let us observe that the verb trawić ‘to digest’ provides particularly strong evidence of desemanticization of iść. Such a complement precludes spatial reading, given the fact that digestion cannot be separated from human body and, consequently, conceptualized as a goal to reach. As put forward in Abraham (1990), a grammaticalized item assumes what he refers to as a twin role in the language. On the one hand, it continues being used as full lexeme, and on the other, it assumes a function of grammatical marker. That is to say, the item lives some kind of double life, independently in lexicon and in grammar. If so, then the auxiliary and the lexeme may happen to appear together in the same construction, as shown in (8) and in (9): (8) On va aller à Montparnasse, il a décidé.(Frantext) (9) I’m going to go and pick them up now.(BNC)

It would be exaggerated to claim that the auxiliarizing iść is free to select its own source lexeme at complement position. However, the corpus provides some striking uses, as those in the examples (10) and (11), revealing that the Polish verb has possibly started the way towards such a double life. Let us investigate on how far it went. (10) Idę iść. Będę wieczorem. go:1sg go:inf be:fut.1sg evening:inst ‘I am going to walk. I will be back tonight.’

(Internet)

 Dorota Sikora

(11) …idę myśleć idąc. Dosłownie. Nic mi tak nie go:1sg think:inf walk:ger literally nothing:nom I:dat so-well neg poprawia humoru jak szwendanie. improve:3sg humor:gen as wandering:nom ‘I am going to think walking. Literally. Nothing makes me feel better than wandering around.’ (Internet)

ICs with iść used as inflected auxiliary and, subsequently, as infinitive complement, remain infrequent: as for now, two instances could be isolated.6 Although in the example (11), the complement position is fulfilled with myśleć ‘to think’, the whole clause brings a strong evidence to support the claim of desemanticization undergone by iść. Obviously, using the IC idę myśleć, the speaker seems perfectly aware of the fact the inflected form idę is no more a content verb. S/he found necessary to add the gerund idąc in order to specify the circumstances of the thinking process. Basically, a grammaticalized verb is also free to introduce infinitives antonymous of the source lexeme, as illustrated in the examples (12) and (13) where stay and rester express the opposite of to go and aller (see Vetters & Lière 2009). (12) He is going to stay here for a while. (BNC) (13) Tout va rester à sa place, comme il est. (Frantext)

Comparatively, in Polish, an IC with iść followed by zostać (PF) or zostawać (IPF) ‘to stay, to remain’ is infelicitous, as shown in (14a) and (14b). (14) a. *Idziesz zostać tutaj?  go:2sg stay:pf.inf here? ‘Are you going to stay here?’ b. *Wszystko idzie zostawać na miejscu.  everything.nom go:3sg remain:ipf.inf in place:loc ‘Everything will remain at the same place’

Data discussed in the present section suggest that iść cannot be yet considered ­full-fledged auxiliary. The bleaching process, though incomplete, has attained the stage in which some selection constraints are released: stative verbs as well as, in some rare cases, iść at the complement position are allowed. It is reasonable to assume then that in some uses (examples (10) and (11)), the verb is able to play the Abraham’s twin role of both, functional marker and main verb. In short, despite clear signs of ongoing semantic bleaching, in most of its uses, iść seems at least able

.  The other one, extracted from Andrzej Stasiuk novel “Mury Hebronu” appears as stylistically marked and could hardly be analyzed as illustration of an ongoing change occurring in discourse. It should rather be looked at as a writer’s literary distortion of every day language.



Is the Polish verb iść an auxiliary to be? 

to assume twin role, without living yet the double life of an auxiliary. We are still dealing with a lexical unit with grammatical uses. 3.2.2  Loss of grammatical properties Weakening lexical constraints are not the only mark of auxiliarization. Further evidence of ongoing bleaching appears in comparison of grammatical p ­ roperties displayed by iść in its different VPs. Let us recall some examples discussed in previous sections and compare them from the angle of their ability to form p ­ refixed forms and their reactions to the change of grammatical tense. In (15), we are clearly dealing with a motion verb. In the case of (6), the spatial reading cannot be definitively excluded, but such interpretation is strongly constrained as long as we need a special context switching it on. On the contrary, in (11), iść shifts to functional meaning of future marker. (15) Idę na rynek. go:1sg to market:acc ‘I am going to the farmers market.’ (6) Tak więc (i) idę zapomnieć o osłabieniu i złym so then (i) go:1sg forget:inf about feeling-weak:abl and bad:abl samopoczuciu i (ii) idę zrobić coś dla siebie… feeling:abl and (ii) go:1sg do:pf.inf something:acc for myself:gen  ‘So, I am going to forget feeling weak and bad, and I am going to do something for myself…’ (Internet) (11) … idę myśleć idąc. Dosłownie. Nic mi tak nie go:1sg think:inf walk:ger literally nothing:nom I:dat so-well neg poprawia humoru jak szwendanie. improve:3sg humor:gen as wandering:nom ‘I am going to think walking. Litterally. Nothing makes me feel better than wandering around.’ (Internet)

The sentence (15) is slightly ambiguous as to the temporal reference with a ­possible reading in terms of current or intended displacement. However, used as main verb, iść is available for a large set of prefixes (po-, przy-, za-, prze-, pod-, do-), that can be applied to derive perfective forms (15¢). In (6), only po- may be selected (6¢), whereas in (11) even that prefix becomes problematic (11¢). (15′) po- Pójdę na rynek. go:pf.1sg to market:acc ‘I am about to go to the farmers market.’ przy- Przyjdę na rynek. come:pf.1sg to market:acc ‘I will come to the farmers market.’

 Dorota Sikora

za- Zajdę na rynek. come:pf.1sg to market:acc ‘I will make a little trip to the farmers market.’ prze- Przejdę na rynek. move:pf.1sg to market:acc ‘I will move (from here) to the farmers market.’ pod- Podejdę na rynek. go:pf.1sg to market:acc ‘I will go to the farmers market (once on my way).’ do- Dojdę na rynek. join:pf.1sg to market:acc ‘I will join the farmers market.’ (6′) Tak więc (i) pójdę zapomnieć o osłabieniu i so then (i) go:pf.1sg forget:inf about feeling-weak:abl and złym samopoczuciu i (ii) pójdę zrobić coś bad:abl feeling:abl and (ii) go:pf.1sg do:pf.inf something:acc dla siebie… for myself:gen ‘So, I will forget feeling weak and bad, and I will do something for myself…’ (11′)

?Pójdę

myśleć idąc. Dosłownie. Nic mi  go:pf.1sg think:inf walk:ger literally nothing:nom I:dat

tak nie poprawia humoru jak szwendanie. so-well neg improve:3sg humor:gen as wandering:nom ‘I am going to think walking. Literally. Nothing makes me feel better than wandering around.’

To summarize, the above results, we can conclude that more the process has advanced, less prefixes remain available for derivation. We could even go further, assuming that in the case of strongly grammaticalized use (see example (11¢)), such derivation is not possible at all, given the fact that (11¢) sounds oddly to ­Polish speakers. Another way of checking whether a lexical item tends to auxiliarize is to use the verb in another grammatical tense than present. The reaction displayed by the verb is similar in all sentences under analysis: used at past tense, iść recovers its initial spatial meaning. The examples (15″), (6″) and (11″) refer to a motion event only. (15′′) Szedłem na rynek. went:ipf.m to market:acc ‘I was walking to the farmers market.’



Is the Polish verb iść an auxiliary to be? 

(6′′) Tak więc (i) szedłem zapomnieć o osłabieniu so then (i) went:ipf.m forget:inf about feeling-weak:abl i złym samopoczuciu i (ii) szedłem and bad:abl feeling:abl and (ii) went:ipf.m zrobić coś dla siebie… do:pf.inf something:acc for myself:gen ‘So, I was walking to forget feeling week and bad, and I was walking to do something for myself ’ (11′′) ?…szedłem myśleć idąc. Dosłownie.  went:ipf.m think:inf walk:ger literally ‘? I was going to think walking. Literally.’

It is worthy to notice that the change of grammatical tense in (15″) has just transferred the event into the past, while in (6″) and in (11″), the informational content of the sentences has been attained. In other words, (6) and (11) on the one hand, (6″) and (11″) on the other provide different kind of information. To sum up, the verb iść displays different grammatical behavior when used with functional sense or with lexical meaning. 4.  Semantics of the auxiliarized iść In the present state of Polish corpora, a reliable statistic study about all aspects of iść in ICs is hard to realize. As a matter of fact, although users have access to collocations, extracting instances of infinitive complements only is not possible yet. Otherwise, only collocations with minimally five occurrences in the corpus are provided in answer to such a query. Nevertheless, random outputs obtained in such a corpus study bring to light some regularity and encourage few explanatory generalizations. First, let us notice that corpora offer several examples of grammaticalized ICs inflected at different persons, as illustrated in (16): (16) Nowi klienci, którzy teraz idą wznawiać new:nom.pl clients:nom who:nom.m now go:3sg renew:ipf.inf bądź zawierać ubezpieczenia przynoszą ze sobą or get:ipf.inf insurance:pl.acc bring:3sg with refl.inst świeży pieniądz. fresh:acc money:acc ‘New clients, who at present are going to renew their insurance or get a new one, are bringing in fresh money.’ (NKJP)

 Dorota Sikora

However, the majority of the auxiliarized occurrences are inflected at first person singular. That very fact matches with the invited inferencing discussed in the ­Section 3. As noticed by De Mulder (p. c.), the first person form is the most appropriate for expression of intentions. No one but the doer can better talk about what s/he wants to, and probably will, do. Second, it is interesting to get a quick look on the contexts in which the IC under analysis appears. Some of them, as the one in the example (17), provide striking evidence of the functional role that the auxiliarized iść assumes in today’s Polish. (17) Pozdrawiam i idę rozmrozić kawał mięsa, potem greet:1sg and go:1sg defrost:pf.inf big-piece:acc meat:gen then je przyprawię. it:acc season:pf.1sg ‘I am sending my regards, and I am going to defrost a big piece of meat, then I will season it.’ (Internet)

Obviously, the IC idę rozmrozić conveys a future meaning. However, as shown in (17′), the clause cannot be replaced with perfective future of the infinitive complement rozmrozić ‘to defrost’, unless the sentence loses coherence: the present of pozdrawiam ‘(I) (am) greet(ing)’, which is a performative act of greeting the addressee, and the future cooking activities would appear then as logically disconnected. (17′)

?Pozdrawiam i

 greet:1sg

idę rozmrożę kawał mięsa, potem and go:1sg defrost:pf.1sg big-piece:acc meat:gen then

je przyprawię. it:acc season:pf.1sg

The analysis of the example (17) points the function assumed by auxiliarized iść in ICs: it provides the speaker with a mean to tie up the present state and future actions. That is to say that the auxiliarized iść is currently relevant in its role of future marker as long as the expressed intention of doing something is the first condition – partly satisfied at the moment of speech – of the act referred to by the infinitive complement. 5.  Conclusions It is time now to sum up the above considerations and to evaluate the degree of auxiliarization displayed by iść in its ICs.



Is the Polish verb iść an auxiliary to be? 

Given available data, the auxiliarization of the verb iść in ICs appears as still ongoing process, far from being completed. The future meaning remains often contextually constrained and the selectional restrictions due to the source meaning are not fully released. The pragmatic ambiguity of the large number of examples under analysis indicates that the auxiliarization process has reached the bridging context phase (Heine 2002): the one in which such ambiguity is not resolved, given the fact that none of the three possible readings (spatial, modal and temporal) is defeasible in context. However, some examples point towards the third stage, that of switch contexts, where iść appears in environments incompatible with its original spatial meaning. That is to say, in such contexts, only the “new” grammatical meaning (i.e. that of future marker) is involved. As shown in recent works on grammaticalization, such evolution may continue as well as it may never result in the next stage. The path, or pattern, of grammaticalization and, to be precise, of auxiliarization followed by iść is, of course, by no means language-specific. We know – since Bybee, Perkins & Pagliuca 1994, Heine & Kuteva 2002 among others – that this is one of the universal conceptual channels of grammaticalization. The main aim of this paper was to provide evidence of similar conceptual device at work in the case of the IC under analysis. Unfortunately, given the state of development of Polish corpora, it is impossible to predict whether iść will continue the way of auxiliarization. Hardly 10% of texts gathered in the IPI Corpus of Polish are labeled “older prose”, but these texts come from nineteenth century novels. Consequently, it is impossible to build diachronic sub-corpora representative for different periods of Polish language history. In other words, we lack data to run statistics and identify the tendency in the auxiliarization undergone by iść. However, some insights are provided from the development of perfect present form in today’s Polish. Sentences like (18), considered inappropriate thirty years ago, are currently used and understood as conveying particular aspectual and temporal information, different from that provided by the corresponding past and present forms. (18) Dom ma zrobione zewnętrzne ocieplenie. house:nom have:3sg done:part external:acc insulation:acc ‘The house has external insulation.’ (Internet)

As stated (among others) in Paprocka-Piotrowska (2008), Polish tenses are poor in modal and aspectual information. Neither of the tenses described in handbooks of grammar can refer to a result state as the “present perfect” in (18) does. I suggest then that the auxiliarized verb iść with its strong modal, aspectual and temporal charge enables speakers to fill one of such gaps in Polish temporal system, saying no more than necessary, and meaning more thereby.

 Dorota Sikora

Abbreviations acc adv fut gen ger inf inst ipf

= accusative = adverb = future = genitive = gerund = infinitive = instrumental = imperfective

nom pst pf prs pl sg voc

= nominative = past = perfective = present = plural = singular = vocative

References Abraham, Werner. 1990. Die Grammatikalisierung von Auxiliar- und Modalverben. Beiträge zur Geschichte Sprache und Litteratur 112: 200–208. Bolinger, Dwight. 1980. Wanna and the gradience of auxiliaries. In Wege zur Universalien ­Forschung: Sprachwissenschaftliche Beiträge zum 60. Geburtstag von Hansjakob Seiler, Gunter Brettschneider & Christian Lehmann (eds), 292–299. Tübingen: Gunter Narr. Brinton, Laurel J. & Traugott, Elizabeth C. 2005. Lexicalization and Language Change. ­Cambridge: CUP. Bybee, Joan L. 2003. Mechanisms of change in grammaticization: The role of frequency. In The Handbook of Historical Linguistics, Brian D. Joseph & Richard D. Janda, 602–623. Oxford: Blackwell. Bybee, Joan L., Perkins, Revere D. & Pagliuca, William. 1994. The Evolution of Grammar. Tense, Aspect and Modality in the Languages of the World. Chicago IL: Chicago University Press. Croft, William. 1990. Typology and Universals. Cambridge: CUP. Croft, William. 2000. Explaining Linguistic Change. An Evolutionary Approach. London: Longman. Forsyth, James. 1970. A Grammar of Aspect. Usage and Meaning in the Russian Verb. Cambridge: CUP. Heine, Bernd. 2002. On the role of context in grammaticalization. In New Reflections on ­Grammaticalization [Typological Studies in Language 49], Iilse Wischer & Gabriele Diewald (eds), 83–101. Amsterdam: John Benjamins. Heine, Bernd, Claudi, Ulrike & Hünnemeyer, Friederike. 1991. Grammaticalization: A Conceptual Framework. Chicago IL: Chicago University Press. Heine, Bernd & Kuteva, Tania. 2002. World Lexicon of Grammaticalization. Cambridge: CUP. Laskowski, Roman. 1999. Czasowniki ruchu liniowego (czasowniki przemieszczania) w perspektywie kontrastywnej. Semantyka a konfrontacja językowa 2: 49–57. Levin, Beth, Beavers, John & Shiao Wei, Tham. 2009. Manner of motion roots across languages: Same or different? In Roots. Word formation from the perspective of “core lexical elements” workshop, University of Stuttgart, 10–12 June. Marchello-Nizia, Christiane. 2006. Grammaticalisation et changement linguistique. Bruxelles: De Boeck.



Is the Polish verb iść an auxiliary to be? 

Paprocka-Piotrowska, Urszula. 2008. Conter au risque de tout changer. Complexité conceptuelle et référence aux procès dans l’acquisition du français L2 et du polonais L2. Lublin: Towarzystwo Naukowe Katolickiego Lubelskiego. Prévost, Sophie. 2003. La grammaticalisation: Unidirectionnalité et statut. Le français moderne LXXI(2): 144–166. Sikora, Dorota. 2009. Les verbes de manière de mouvement en français et en polonais. Éléments pour une étude comparée des propriétés structurelles des prédicats. Thèse de doctorat sous la direction de D. Apothéloz, ATILF-CNRS (UMR 7118)-Nancy-Université. Sikora, Dorota. Forthcoming. Sur le chemin d’auxiliarisation? Verbe polonais iść dans ses constructions infinitives – étude sur corpus. Roczniki Humanistyczne, Université Catholique de Lublin. Talmy, Leonard. 2000. Toward a Cognitive Semantics, Vol. I: Concept Structuring System. Vol. II: Typology and Process in Concept Structuring. Cambridge MA: The MIT Press. Tenny, Carol. 1995. How motion verbs are special. The interaction of semantic and pragmatic information in aspectual verb meaning. Pragmatics and Cognition 3(1): 31–73. Traugott, Elizabeth C. 2010. Grammaticalization. In Historical Pragmatics. Andreas H. Jucker & Irma Taavitsainen (eds). 97–126. Berlin: Mouton de Gruyter. Traugott, Elizabeth C. & König, Ekkehard. 1991. The semantics-pragmatics of grammaticalization revisited. In Approaches to Grammaticalization, Vol. 1 [Typological Studies in Language 19], Elizabeth C. Traugott & Bernd Heine (eds), 189–218. Amsterdam: John Benjamins. Vetters Carl & Lière, Audrey. 2009. Quand une périphrase devient temps verbal: Le cas d’aller + infinitif. Faits de langues 33: 27–36. Vetters, Carl & Skibińska, Elzbieta. 1998. Le futur: Une question de temps ou de mode? Remarques générales du ‘présent-futur’ perfectif polonais. In Cahiers Chronos 2, Andrée Borillo, Carl Vetters & Marcel Vuillaume (eds), 247–266. Amsterdam: Rodopi. Włodarczyk, Hélène. 1980. Czasowniki ruchu a katergoria aspektu w języku polskim. Polonica 6: 49–56. On-line corpora: Frantext 〈www.frantext.fr〉 British National Corpus 〈www.natcorp.ox.ac.uk〉 Korpus Instytutu Podstaw Informatyki Polskiej Akademii Nauk (IPI PAN) (http://korpus.pl) NKJP: Narodowy Korpus Języka Polskiego (Polish National Corpus) (www.nkjp.pl)

Towards evidentiality markers in Albanian and Macedonian bilingual political discourse Maxim Makartsev

Institute for Slavic Studies, Russian Academy of Sciences, Moscow The article concerns the two main problems connected to evidentiality in bilingual Macedonian and Albanian political discourse: Firstly, there is a problem of matching the form in the original to its equivalent in translation, which is not trivial in case of perfect in Macedonian-Albanian and Albanian-Macedonian translation. Macedonian perfect is homonymous to some evidential forms, whereas in Albanian there is an expansion of perfect at the expense of aorist. So perfect in both languages under consideration can be better described as a frame, inside which different phenomena take place, using the suitable definition by F. Fici. The translation between the languages provides some algorithms for defining the intent of the form’s original meaning. Secondly, the problem of interaction between a grammatical and a lexical marker when used together is described. The bilingual discourse gives some insights as to how the common meaning of the phrase can be drawn from the lexical and grammatical components and then recoded in the target language. Keywords:  evidentiality; political discourse; bilingual discourse; perfect; translation

According to A. Aikhenvald, evidentiality is “a linguistic category whose ­primary meaning is [the] source of information” (Aikhenvald 2004: 3). Evidentiality takes a different place in different language systems. In this article I will focus on two languages: Macedonian, where expressing evidentiality (witnessed or ­unwitnessed) is compulsory in the past tense, and Albanian, where expressing evidentiality is optional and this grammatical meaning is usually entangled with the admirative. In Macedonian, grammatical markers of unwitnessed evidentiality (‘I didn’t witness it, so I don’t vouch for it, but they say…’) are the so-called ‘perfect-like’ l-forms, dating back to the Common Slavic ‘be’-perfect. Grammatical markers

 Maxim Makartsev

of witnessed evidentiality (‘I witnessed it, therefore I vouch for it’) are forms in ­-v-/-š-/-j-, dating back to the aorist and imperfect. When speaking about a series of events in the past on the text level, there is no way to avoid indication of whether the speaker has been a witness or not; hence, Macedonian is a language with grammaticalised evidentiality.1 In Albanian, grammatical markers of unwitnessed evidentiality are the socalled admirative forms, based on the model ‘participle of the main verb without ending + conjugated forms of the verb kam ‘to have” (practically, it is inverted apocopated perfect: cf. perfect kam shkruar ‘lit. have written’ and admirative ­shkrua-kam ‘lit. writ-have’ with the meaning ‘(it turned out that…/they say that…) I write!’ There are no grammatical markers of witnessed evidentiality and marking unwitnessed evidentiality is not compulsory, so we can say that Albanian is a language with grammaticalised evidential strategy.2 The system becomes even more difficult if we consider that apart from the grammatical markers in Albanian and Macedonian, there are lexical evidential markers, as in the majority of other European languages, even those which do not have a grammatical category of evidentiality: compare English apparently, supposedly, Russian jakoby, mol, vrode kak, Polish ponoć, jakoby, Czech prý/prej, Slovene bajè, Serbian/Bosnian/Croatian navodno, etc. This topic has received much attention recently, yet its bibliography is still not as extensive as that of grammatical(ised) evidentiality.3 The following works are dedicated especially to

.  Some interpret Balkan Slavic evidentials sub specie epistemic overtones they can gain in certain contexts, cf. Friedman 1986, who treats the opposition they express as not witnessed/unwitnessed, but confirmative/non-confirmative. I follow G. Lazard (Lazard 2001) and Zl. Guentchéva (e.g. Guentchéva 1996), who include evidentiality together with epistemic modality and mirativity into the umbrella term of mediativity, or the category which indicates a certain distance between the narrated event and the speaker. .  The Macedonian and Albanian grammatical evidential systems are described meticulously in academic grammars and other works on different aspects of the phenomenon, so I will refer readers to the following works: Lunt 1952; Koneski 1967, Foulon-Hristova 1995; Fici 2001; Mushin 2001 and Usikova 2003 for Macedonian; Lafe 1975; Sytov 1979; Buchholz & Fiedler 1987; Duchet & Përnaska 1996 and Domi 2002 for Albanian. V. Friedman in various works takes into consideration both Macedonian and Albanian evidentiality, comparing it to the situation in other Balkan (Balkan Romance and Bulgarian) and non-Balkan languages (Turkish, Georgian and Lak) (see, among other, his works Friedman 1986, 1988, 1998, 2000, 2003). .  The literature on grammatical evidentiality in Balkan languages is so vast that it both ­impossible and unnecessary to provide references to it in an article dealing with lexical evidentiality in bilingual discourse, which is a relatively new topic. The most impor­ tant ­bibliography on the issue can be found in Victor Friedman’s articles, in particular in



Towards evidentiality markers in Albanian and Macedonian bilingual political discourse 

Macedonian markers: Bužarovska 2006, Gajdova & Labroska 2008, Gajdova  & Labroska 2010; Tofoska 2008, Petroska 2012, Petroska & Тоfoska 2011, Kasapovska (foregoing). The bibliography on Albanian lexical evidential markers is the smallest one; to the best of my knowledge, there is only one article partly dedicated to them (Makartsev 2012). Here I would like to show the use of evidentiality in Albanian and Macedonian bilingual political discourse, and especially to dwell upon interactions between lexical and grammatical markers. The data for the first part of my article is taken from a bilingual Albanian-Macedonian political talk-show “Eurofokus,” which is broadcast on ALSAT-M from Skopje. The guests in the studio are Macedonian and Albanian politicians and political experts from Macedonia. Each of them chooses the language s/he wishes to speak: Macedonian or Albanian. In the studio, simultaneous translation is provided for those who need it; when the program is broadcast, subtitles are provided. That is why, in practise, we deal with one text in two linguistic dimensions. This practice is not unique to ALSAT-M and is used as well by other media in Macedonia that target minority integration. I shall focus on the full excerption of past forms from a “Eurofokus” programme.4 The main story is about citizens of two Macedonian communities, Kumanovo and Lipkovo, mainly ethnic Albanians, who started to leave Macedonia en masse for Belgium, where they demanded economic asylum. In the s­ tudio there are two journalists and four invited speakers. The majority of people in the studio speak Albanian; one of the speakers and one of the journalists speak Macedonian. In the course of the heated argument an Albanian speaker switches to Macedonian but later he turns back to Albanian. Everything said is subtitled; the speeches in Macedonian are provided with Albanian subtitles and vice versa. There is also simultaneous translation, but the Macedonian speaker is the only one who uses it, because all the others are bilingual. The majority of the discussion is in Albanian (if we count verbal forms in the programme, the proportion of Albanian to Macedonian is 5 to 1). I chose this programme because the topic mainly focuses on some events in the past (as opposed to other productions of the channel, which focus on the ­present and the future of Macedonia), and in Macedonian evidentiality is compulsorily marked only in the past. The disproportion of the language data can be easily reduced by introducing a coefficient (k = 5) to Macedonian data.

Friedman 2003. A long overview of existing literature on Balkan evidentiality with the analysis of the respecting Bulgarian, Macedonian and Albanian forms and their interpretation in the context is provided in my book (Makarcev, foregoing). .  Broadcast on 08.03.2010, total length 57’58”.

 Maxim Makartsev

If we consider the parallel text of the emission, 205 slots with aorist, imperfect or perfect in at least one of the languages are to be found. The statistics are introduced in the Table 1: Table 1.  Tenses in the original and their counterparts in translation5 From Albanian original to Macedonian subtitles

From Macedonian original to Albanian subtitles

72

29

Aor/IPF → PF

–

–

PF → PF

63

4

PF → Aor/IPF

20

4

Aor/IPF → Other

 1

–

PF → Other

 3

–

Other → Aor/IPF

 4

–

Other → PF

 1

–

Unclear

 4

–

Tenses in original → Tenses in translation5 Aor/IPF → Aor/IPF

Since only Macedonian unwitnessed forms homonymous to perfect were used, they are described in the same field of the Table under the title ‘Perfect’. ‘Other’ stands for present, future, pluperfect, participles and gaps in the translation. They are not taken into consideration, as they do not outnumber statistical observational error. ‘Unclear’ stands for the forms which cannot be heard well enough. The only Albanian admirative form is in the Present “No one took us (s’na merrkaADM/ne nè zedeAOR)” (“Eurofokus”, 08.03.2010, 34’20”), so it is put into the ‘Other’ line. It is not linked to evidentiality, because its primary meaning in this case is unpleasant surprise and embarrassment. One possible criticism for this model is that the translators are under the influence of the language they are translating from, so the result of the translation can bear certain features of the original. In this case, we would expect that the verbal forms would be translated with their structural equivalents in the second language: aorist/imperfect would be translated only by aorist/imperfect, perfect by perfect. But there is a large group of examples where an Albanian perfect is translated by Macedonian aorist/imperfect. Here we need to consider the ambivalence of perfect in Albanian and Macedonian. As has already been said, in Macedonian perfect is homonymous to some unwitnessed evidentials (to be more precise, to evidential imperfect and evidential aorist; moreover, that perfect can also be used

.  Aor – aorist; IPF – imperfect; PF – perfect.



Towards evidentiality markers in Albanian and Macedonian bilingual political discourse 

as evidential perfect, which makes us suspect evidentiality almost in every usage of Macedonian perfect).6 So in every perfect form one could suspect an unwitnessed evidential. In Albanian there is no such homonymy, but another process is currently taking place – an expansion of perfect into aorist area, which happens both in oral and in written discourse and is not reserved to a specific dialect area. Buchholz & Fiedler (1987: 130) consider this process to be a north-eastern Gheg feature, but the academic grammar of Albanian states that it happens both in spoken and in written speech with no reference to the dialects (compare also Përnaska 1982: 141, 143, 148; Asenova 2002: 252–253). So, between the areas of meanings represented by perfect in Albanian and Macedonian there is a zone of intersection, but there are nonintersecting zones as well. The left circle stands for Albanian, the right one is for Macedonian (the size of the respective zones is irrelevant):

A

B

C

A – Perfect substitutes aorist; B – Perfect for events that happened in the past and are still valid at the moment of speech; С – Unwitnessed evidentials, homonymous to perfect. Figure 1.  Semantics of perfect tense in Albanian and Macedonian

If the translation is from Albanian and the meaning is from area A, the Albanian perfect corresponds to aorist in Macedonian:7

(1) a. [Original] Azilkërkuesit nga Maqedonia, Serbia apo Mali i Zi, që kanë mbetur.PF rrugëve, jo vetëm që nuk kanë parë.PF këto shuma, por ata me më shumë fat, qe mund të kenë qjetur vende në një kamp refugjatësh, në pritje të përgjigjës, që thuan se do t’jetë negative, marrin vetëm 7 euro në javë (“Eurofokus”, 08.03.2010, 18’20”).

.  There are other unwitnessed evidentials not homonymous to the perfect, e.g. unwitnessed future (compare PF/Ev.PF/Ev.Aor. došol and Ev.Fut ḱe došol from the verb dojde ‘to come’), but they are used much less often and were not found in my data. .  In all the examples first the original is given, then translation into English, then the ­subtitles. In all cases, the English translation is made from the original, so some differences between the English translation and the subtitles are inevitable, but they are not relevant for the grammatical forms I will be concentrating on.

 Maxim Makartsev

b. The refugees from Macedonia, Serbia or Montenegro, which remained in the streets, not just that they didn’t see those money, but even those, who were more lucky and managed to find places in the refugee camps, as they are waiting for the answer (which is said to be negative), get only 7 euros per week. c. [Subtitles] No azilantite od Makedonija i od Srbija koi ostanaa.Aor na ulica, ne samo što ne gi vidoa.Aor ovie pari, tuku tie so poveké sreḱa što možea da najdat nekoje mesto vo nekoj od kampovite za begalci i se vo isčekuvanje na odgovor – za koj se veli deka ké bide negativen – dobivaat po sedum evra nedelno.

If a perfect form from B area is translated (no matter from Macedonian into ­Albanian or vice versa), there is perfect in the translation as well (compare the use of perfect in English as well):

(2) а. [Original] Për shkak se gjitha punë kjo Qeveri bân nga zyrja, asnjëherë komuna e Likovës nuk ka përjetu.PF që një kryeministër i Maqedonisë të pretë ndonjë lentë në komunën e Likovës (“Eurofokus”, 08.03.2010, 29’48”).

b. As far as all the businesses this government does through offices, never has the municipality of Lipkovo seen that a Prime Minister of Macedonia would cut a band in the municipality. c.  [Subtitles] Bidejkí ovaa Vlada svoite raboti gi pravi od kancelarija, nikogaš opština ne doživeala.PF eden minister na Makedonija da preseče nekoja lenta vo opština Lipkovo.

If a form from area C (unwitnessed evidentials, homonymous to perfect) is translated from Macedonian to Albanian, in Albanian both perfect and aorist can be used, because in Albanian it is not compulsory to express the opposition between witnessed and unwitnessed evidentiality; compare the example with Aorist:

(3) а. [Original] Mene mi e nesfatlivo deka Vladata ne moželo.PF da go ­predvidi, pošto godina dena zboruvavme (“Eurofokus”, 08.03.2010, 40’00").

b. It is incomprehensible to me that the Government couldn’t have ­foreseen it, because we have been saying this for a year. c. [Subtitles] Për mua është e pakuptueshme, që Qeveria këtë nuk arriti.Aor ta parashikojë, ngase ne ka një vit, që po flasim.

Of course the original influences the translator, so there are cases when in the translation there are structural calques, contradicting the speaker’s intentions. For example, in a story about some events which the speaker has witnessed and thus it is expected that s/he would believe in them, suddenly in the Macedonian translation a perfect, homonymous to unwitnessed evidentials is used:





Towards evidentiality markers in Albanian and Macedonian bilingual political discourse 

(4) а. [Original] Ismeti, i cili nuk pranonte të identifikohet para kamera, është një tjetër banor që ka kërkuar.PF azil në Belgjikë (“Eurofokus”, 08.03.2010, 18’40”).

b. Ismet, who didn’t want to speak for the camera, is a citizen of ­Kumanovo, who asked for asylum in Belgium. c. [Subtitles] Ismet, koj ne saka da zboruva pred kamera, e žitel na ­Kumanovo koj pobaral.PF azil vo Belgija.

The reporter met Ismet in a queue before the Belgian Bureau for Foreigners, among other refugees from former Yugoslavia who had migrated en masse several weeks before that. The reporter does not have any reason not to believe Ismet; moreover, he gives his story as one more example of the refugees. So it is not in his best interest to depict the story as untrustworthy. But the translation does him a bad turn: in this context both aorist (pobara) and present (bara) could be used in Macedonian, so use of the perfect as opposed to those unrealised opportunities is understood as doubtful reportive. In other words, an additional meaning of unbelief emerges in the translation. But examples like this are very rare, and do not change the general picture very much. What makes it difficult to describe grammatical markers in Macedonian and Albanian is that in their semantics the “clear” evidential meaning is often accompanied by an epistemic one. That is why throughout the article I have been underlining that it is essential not only whether the speaker has witnessed the events, but also whether s/he believes in the information s/he conveys. That is why in many cases (apart from the transparent: “the speaker has seen, so s/he vouches for what s/he says” and “the speaker didn’t see, so s/he doesn’t vouch for what s/he says”) one should always ask oneself the question: is it just evidentiality, is it both evidentiality and epistemic modality, or is it just epistemic modality? At the same time, lexical markers can also have several meanings. That is why on the first account we deal with very loose rules: lexical markers with multiple meanings are used together with ambivalent grammatical markers. So can we speak about rules at all? Here we can apply Peter Kehayov’s division between “analytic” and “holistic” readings of the construction consisting of a lexical and a grammatical marker (Kehayov 2008). Within the semantics of the markers, evidential and epistemic semantic components are highlighted and the common meaning of the construction is driven from the sum of the components. As an example I would like to present a part of an interview in Albanian with Macedonian subtitles.8

.  Published by the “Free Europe” channel on YouTube on 06.10.2007.

 Maxim Makartsev

(5) а. [Original] Ajo që thuhet (i) se… gjoja (ii) se simpatizantë apo anëtarë i Bashkimit Demokratik për Integrim kanë qenë.PF të paisur apo kanë ardhë.PF (iii) armatosur fare nuk qendron.

b. The thing being said (i), that… apparently (ii) sympathizers or members of the democratic Integration Union were armed or came armed (iii), that is not right at all.

c. [Subtitles] Onie što velat (i) deka navodno (ii) simpatizeri ili členovi na DUI bile dojdeni.PF.II.Evid (iii), toa voopšto ne drži.

Let me show in the Table 2 the semantic components in the Albanian original and Macedonian translation: Table 2.  Lexical and grammatical markers and their semantic components in Albanian and Macedonian (i) being said, that 〈…〉 albanian

(ii) apparently 〈…〉

macedonian albanian macedonian

thuhet se 〈…〉

velat deka

gjoja se 〈…〉

navodno

evidential

evidential

evidential & epistemic

evidential

(iii) were 〈…〉 or came 〈…〉 albanian

macedonian

kanë qenë … apo bile dojdeni 〈…〉 kanë ardhë 〈…〉 Ø

evidential & epistemic

So both evidential and epistemic components of the message are translated, but by different components: in Albanian, evidential and epistemic meanings are combined in one lexical marker (ii), and in Macedonian – in one grammatical marker (iii). Bilingual Macedonian media, besides their important role of a source of information, are a place for communication in recent years – for communication between members of different societies who speak different languages. As far as modern technologies enable the discourse to continue in two linguistic dimensions at once, it provides us with a new type of data for comparative study of grammar categories. In our case, these data help differentiate between homonymous forms (evidentials vs. perfect) in Macedonian, show the process of the expansion of perfect into the area of aorist/imperfect in Albanian and show the place of evidential and epistemic semantic components in shaping the sense of the whole utterance.



Towards evidentiality markers in Albanian and Macedonian bilingual political discourse 

References Aikhenvald, Alexandra Y. 2004. Evidentiality. Oxford: OUP. Asеnоva, Petya. 2002. Balkanskо еzikоznaniе. Vеlikо Tărnоvо: Fabеr. Buchholz, Oda & Fiedler, Wilfrid. 1987. Albanische Grammatik. Leipzig: Verlag Enzyklopädie. Bužarovska, Eleni. 2006. Pathways of semantic change: From similarity marker to sensory ­evidential. Slavia Meridionalis. Studia Linguistica Slavica et Balcanica 6: 185–208. Domi, Mahir (ed.). 2002. Gramatika e gjuhës shqipe, Vol. I: Morfologjija. Tiranë. Duchet, Jean-Louis & Përnaska, Remzi. 1996. L’admiratif albanais: Recherche d’un invariant sémantique. In L’énonciation mediatisée, Zlatka Guentchéva (ed.), 31–46. Louvain: Editions Peeters. Fici, Francesca. 2001. Macedonian perfect and its modal strategies. Македонски јазик LI-LII: 61–86. Foulon-Hristova, Jordanka. 1995. Les modes de narration en macédonien. Paris: L’Asiathèque. Friedman, Victor A. 1986. Evidentiality in the Balkans: Bulgarian, Macedonian and Albanian. In Evidentiality: The Linguistic Coding of Epistemology, Wallace Chafe & Johanna Nichols (eds), 168–187. Norwood NJ: Ablex. Friedman, Victor A. 1988. The category of evidentiality in the Balkans and the Caucasus. In American Contributions to the Tenth International Congress of Slavists, Linguistics, ­Alexander M. Schenker (ed.), 121–139. Columbus: Slavica. Friedman, Victor A. 1998. The grammatical expression of presumption and related concepts in Balkan Slavic and Balkan Romance. In American Contributions to the 12th International Congress of Slavists, Robert A. Maguire & Alan Timberlake (eds), 390–405. Bloomington IN: Slavica. Friedman, Victor A. 2000. Confirmative/nonconfirmative in Balkan Slavic, Balkan Romance, and Albanian with additional observations on Turkish, Romani, Georgian, and Lak. In Evidentials: Turkic, Iranian and Neighbouring Languages [Empirical Approaches to Language Typology 24], Lars Johanson & Bo Utas (eds), 329–366. Berlin: Mouton de Gruyter. Friedman, Victor A. 2003. Evidentiality in the Balkans with special attention to Macedonian and Albanian. In Studies in Evidentiality [Typological Studies in Language 54], Alexandra Y. Aikhenvald & Robert M.W. Dixon (eds), 189–218. Amsterdam: John Benjamins. Gajdоva, Ubavka & Labrоska, Veselinka. 2008. Katеgоrijata distanca vо makеdоnskiоt, vо bugarskiоt i vо sprskiоt jazik (gramatički i lеksički pоkazatеli). In Rеfеrati na makеdоnskitе slavisti za ХIV-оt Mеǵunarоdеn slavistički kоngrеs vо Охrid, 10–16 sеptеmvri 2008. gоdina, 79–93. Skоpје: Makеdоnska akadеmija na naukitе i umеtnоstitе. Gajdоva, Ubavka & Labrоska, Veselinka. 2010. Za еvidеncijalnоsta vо makеdоnskiоt jazik. Makеdоnski jazik LXI: 53–58. Guentchéva, Zlatka (ed.). 1996. L’énonciation médiatisée. Louvain: Editions Peeters. Kasapоvska, Božana. Forthcoming. Za sеmantikata na prеdlоgоt spоrеd vо makеdоnskiоt jazik. Kehayov, Petar. 2008. Interactions between grammatical evidentials and lexical markers of ­epistemicity and evidentiality: A case study of Bulgarian and Estonian. In Lexikalische Evidenzialitäts-Marker in Slavischen Sprachen [Wiener Slawistischer Almanach. Sonderband 72], Björn Wiemer & Vladimir Plungjan (eds), 165–201. München: Otto Sagner. Kоnеski, Blaže. 1967. Gramatika na makеdоnskiоt litеraturеn јazik. Skоpjе: Kultura. Lafe, Emil. 1975. Mbi habitoren dhe disa trajta të saj. Studime filologjike 2: 143–151.

 Maxim Makartsev Lazard, Gilbert. 2001. On the grammaticalization of evidentiality. Journal of Pragmatics 33: 359–367. Lunt, Horace. 1952. A Grammar of the Macedonian Standard Language. Skopje: Državno ­knigoizdatelstvo na NR Makedonija. Makarcеv, Maksim. Forthcoming. Ėvidеncial’nоst’ v prоstranstvе balkanskоgо tеksta (­Evidentiality in Balkan Text Space). Mоskva, Sankt-Peterburg: Nestor-Istorija. Makartsev, Maxim. 2012. Towards common Balkan lexical evidential markers. Slověne. ­International Journal of Slavic Studies 1(1): 21–51. Mushin, Ilana. 2001. Evidentiality and Epistemological Stance: Narrative Retelling [Pragmatics & Beyond New Series 87]. Amsterdam: John Benjamins. Përnaska, Remzi. 1982. De l’opposition aoriste/parfait en albanais. Cahiers balkaniques 3: 139–163. Petroska, Elena. 2012. Grammatical and lexical markers of evidentiality in Macedonian. ­Proceedings from the Seventh Macedonian - North American Conference on Macedonian Studies [Balkanistica 25:2], Victor A. Friedman & Donald L. Dyer (eds), 139-151. Pеtrоska, Elena & Tоfоska, Staša. 2011. Lеksički pоkazatеli za prеkažana еvidеnciјalnоst vо makеdоnskiоt vо pоlskiоt јazik. In Zbоrnik оd Оsmata makеdоnskо-pоlska naučna kоnfеrеnciјa. Skоpје: Filоlоški fakultеt “Blažе Kоnеski”. Sytоv, Alexej P. 1979. Katеgоrija admirativa v albanskоm jazykе i ее balkanskiе sооtvеtstvija. In Prоblеmy sintaksisa jazykоv balkanskоgо, A.V. Dеsnickaja (ed.), 90–133. Lеningrad: Nauka. Tоfоska, Staša S. 2008. Glagоlski prеdikati-nоsitеli na infоrmacijata za еvidеncijalnоst vо makеdоnskiоt i pоlskiоt jazik. In XXXV Naučna kоnfеrеnciјa – lingvistika (Охrid, 11.VIII.-28.VIII.2008), 303–310. Skоpје: Univеrzitеt “Sv. Kiril i Mеtоdij”. Usikоva, Rina P. 2003. Grammatika makеdоnskоgо litеraturnоgо jazyka. Mоskva: Muravеj.

A strange variant of Russian čtoby-construction Irreality and tense-marking* Alexander Letuchiy

National Research University, Higher School of Economics, Philological department, Moscow, Russia In my paper, I discuss the choice of the verb form in constructions with a complementizer/adverbial subordinator čtoby. I show that while the situation is rather trivial in biclausal constructions, analysis of triclausal constructions where another clause is embedded under the čtoby-clause reveals a phenomenon which is not accounted for in any Russian grammatical descriptions. The marker can impose the past tense form not only to the verb in its clause, but also to the temporal clause which is embedded deeper. The existence of such an unusual construction results from both semantic and syntactic factors: namely, from the fact that the ‘unreal’ meaning carried by čtoby spreads to the temporal clause and from the syntactic properties of čtoby and kogda, the latter not imposing any formal restrictions on the verb form in the temporal clause. Keywords:  irrealis; complementation; sentential arguments; purpose constructions; Russian; tense

1.  Introduction In Russian, as in many languages, there are numerous types of embedded clauses. These types are different both semantically and syntactically. For instance, they can be divided, as it is traditionally done (see Noonan 2007 for detailed classification), into three semantico-syntactic classes: relative clauses (as in People who are from Sweden tend to be very hard-working), adverbial clauses (with the meaning

*  The research has been carried out within the project of Russian corpus-based grammar "Rusgram" (http://rusgram.ru), and the author thanks the Program for fundamental research "Corpus Linguistics" (the project "Development of Russian Corpus-based Grammar") of the Presidium of Russian Academy of Sciences for the financial support.

 Alexander Letuchiy

of time, reason, purpose, and so on, as in I was late because I lost my bag), and ­sentential arguments (as in I want to visit my friends). Another classification, which is more relevant for the present paper, is based on the choice of the verb form in the embedded clause. The verb form in the embedded clause is sometimes restricted – these restrictions are often related to individual properties of the subordinator under analysis. Consider, for instance, embedded clauses containing the subordinators kogda and čtoby. In the former, the verb can have a wide range of finite verb forms: forms of present, future or past tense and subjunctive mood. Their choice is semantically motivated: (1) a. Kogda ja priedu, ja tebe pozvonju. when I:nom come:pf1.1sg I:nom you:dat call:pf.1sg ‘When I come, I will call you.’ b. Kogda ja priezžaju, ja emu vsegda zvonju. when I:nom come:ipf.1sg I:nom he:dat always call:ipf.1sg ‘When I come, I always call him.’ c. Kogda ja priexal, ja emu pozvonil. when I:nom came:pf.m I:nom he:dat called:pf.m ‘When I came, I called him.’ d. Kogda ja priexal by, ja by emu pozvonil. when I:nom came:pf.m part I:nom part he:dat called:pf.m. ‘When I would come, I would call him.’

The variant with subjunctive mood, as in (1d), is somewhat less frequent than the other ones. However, it is not at all ungrammatical or semantically peculiar. It can be used, for instance, in a context like: (1′) A.  Why didn’t you call your uncle? Today is his Birthday. He told me he was very angry. B.  He had to wait a bit. I would call him when I came home. I couldn’t do it at my job.

Thus, the adverbial subordinator kogda does not prohibit any verb form in the embedded clause. This is not the case with the word čtoby used both as a complementizer and an adverbial (purpose) subordinator. Čtoby was historically a combination of the complementizer čto and the particle by which marks subjunctive mood and,

.  The lexical aspect (perfective/imperfective) is marked for all verbs except byt’ ‘to be’. Most examples are elicited but examples of the construction under analysis are from the Russian National Corpus (www.ruscorpora.ru).



A strange variant of Russian čtoby-construction 

more generally, meanings related to unreality, such as wish, advise, and so on. The canonical subjunctive mood forms consist of by and the past tense verb form – or, more precisely, a form homonymous to the past tense form. There are also combinations of by with infinitive which are regarded as subjunctive forms by some researchers but are not considered to be subjunctive forms by others. Today čtoby is hardly understood by native speakers as a combination of čto + by. However, linguists are not unanimous concerning the question whether the linguistics properties of čtoby can all be compositionally derived from čto + by (see Panov 1966 and Brecht 1977 for analysis of existing points of view). At least, there are many properties which distinguish čtoby from the combination of čto and by. First of all, by in this case always follows čto while in other cases this particle has a relatively free linear position. Second, the range of meanings of čtoby (e.g. the purpose meaning) cannot be explained by the meaning of its components. However, the formal restriction on the verb form of the embedded clause is retained in the contemporary language: the verb should always be in the infinitive (2a) or past tense form (2b):2 (2) a. Prixodi čtoby za nas pobolet’. come:ipf.imp in-order-to for we:acc support:inf ‘Come and support us!’ (e.g. in a football match). b. Čtoby bylo ponjatnee, ja vstavil tablicu. in-order-to was:neut clear:com I:nom inserted:m table:acc ‘I inserted a table so as to make it clearer.’

Other forms, such as future and present, are impossible: c. *Čtoby budet ponjatnee, ja vstavil tablicu.  in-order-to be:fut.3sg clear:com I:nom inserted:m table:acc Intended: ‘I inserted a table so as to make it clearer.’

Dobrushina 2012 proposes to distinguish two types of čtoby: the first one she calls ‘purpose čtoby’ and the second one ‘irreal čtoby’. She shows that the two čtoby behave in different ways and are reducible to the combination of čto and by to ­different extent.

.  Here and below I call the form with the marker -l, as bylo `was' in (2) `past tense form' for the sake of simplicity. However, we should bear in mind that in uses discussed here it composes the form of subjunctive mood together with the particle by.

 Alexander Letuchiy

Thus, the subordinator3 čtoby restricts the use of verb forms in the embedded clauses. In what follows I will examine only constructions with čtoby and I will begin with a short sketch of its functions. The principal aim of my study is to consider constructions with čtoby containing ‘double embedding’: in other words, čtoby hosts a clause embedded under the main one, but another clause is embedded deeper under the čtoby-clause. I seek to answer the question whether there are any special kinds of behavior characterizing these ‘tripartite’ structures. 2.  Functions of čtoby The word čtoby has various functions, being used both in biclausal and monoclausal constructions (see Dobrushina 2011 for details). In biclausal constructions, its main uses are (1) in purpose adverbial clauses and (2) in complement clauses marked for irrealis. First, čtoby is the main adverbial subordinator marking purpose clauses in Russian (see, for instance, (2a–b) above). Other subordinators are daby, which is much more rare than čtoby, and the combination dlja t-ogo čtoby (for that-gen. sg.n to) which includes čtoby. These complementizers do not show significant semantic differences when compared to the marker under analysis. The second function of čtoby is to form complement clauses marked for irrealis. Some verbs whereof the clausal complement denotes an unreal situation require use of čtoby (3), not the default complementizer čto, illustrated in (4). (3) Ja xoč-u čtoby priexal moj djadja. I:nom want:1sg to came:m my:nom uncle:nom ‘I want my uncle to come.’ (4) Ja ponjal čto oni menja obmanuli. I:nom understood:m that they:nom I:acc deceived:pl ‘I understood that they deceived me.’

The verb xotet’ ‘to want’ cannot be used with čto. Russian possesses another class of lexemes which are compatible with both čto and čtoby: here belong dumat’ ‘to think’, somnevat’sja ‘to doubt’ and so on. For some of them, such as somnevat’sja the semantic difference between the two

.  Here and below, the marker čtoby is often called ‘subordinator’ in order to avoid mention of all its syntactic functions. The marker is glossed as ‘to’ or ‘in order to’, depending on the function in the given context.



A strange variant of Russian čtoby-construction 

variants is very little and not at all obvious. Others, such as dumat’, take čto as a default option, while čtoby is possible under certain conditions. For dumat’, čtoby is ­possible mainly when the verb is negated: (5) Somnevajus’ čto/čtoby oni tebja ponjali. doubt:1sg that/to they:nom you:acc understood:pl ‘I have some doubts that they understood you.’ (6) Dumaju čto/*čtoby on tebja ponjal. think:1sg that/to he:nom you:acc understood:m ‘I think that he understood you.’ (7) Ne dumaju čtoby on tebja ponjal. neg think:1sg to he:nom you:acc understood:m ‘I don’t think that he understood you.’

In (5), čto and čtoby are interchangeable, and almost no semantic difference is observed between the two variants. In contrast, the verb dumat’ with čtoby seems to presuppose a greatest degree of unreality of the situation, though it is difficult to be checked against the corpus data. There is another difference, both in (5) and in (6)–(7), namely that in variants with čto, the tense of the verb in the embedded clause can be past, present or future, depending on whether the event (e.g. ‘he understands you’) took place before, after or simultaneously with the speech act. In variants with čtoby, only the past tense form is possible in the embedded clause, which is typical for all ­constructions with čtoby, and the temporal reference can be either – for instance, in (5), the situation ‘they understand you’ could take place before, after or simultaneously with the speech act, and the verb form will remain the same. Note that the choice of the verb form is organized differently depending on whether čtoby plays a role of purpose subordinator or ‘unreal’ complementizer. In the former function, the verb in the embedded clause can take either the past tense, as in (2b) or the infinitive form, as in (2a). Infinitive is mainly used when the subjects of the main and the embedded clause are co-referent, which is the case in (2a), and more rarely in different subjects constructions. In contrast, when čtoby is used as a complementizer, as in (5) and (7), i­ nfinitive is impossible in the embedded clause, and the verb always takes the past tense form. Finally, čtoby has some particle-like uses, for instance, the use represented in (8) in which it denotes an unpolite type of command, mostly in combination with the first or third person: (8) Čtoby ja ètogo Peti u nas ne videl! to I:nom this:gen.m Petja:gen by we:gen neg saw:m ‘I don’t want to see this Petja at our home.’

 Alexander Letuchiy

In what follows, only examples illustrating the first two functions (those of ­adverbial subordinator and marker of sentential arguments) will be considered. However, I will treat these two classes together, as I haven’t discover any contexts where it is relevant for the problem under analysis if čtoby is used as an adverbial subordinator or sentential argument marker. 3.  Tense in Russian complex clauses As stated above, the verb in constructions with kogda can be in the past, present or future tense or in the subjunctive mood. Importantly, tense forms in Russian, as in many other European languages, can have absolute or relative uses. The three tense forms of the indicative mood can either denote the temporal localization of the event with respect to the speech act (absolute use), or with respect to another event mentioned in the discourse (relative use). Relative uses of verb forms logically correspond to their absolute uses. The past tense denote the precedence of the event to the speech act (absolute use) or another event (relative use). Similarly, the present tense designates that the event is simultaneous (in any precise sense) with the speech act (absolute use) or another event (relative use), while the future tense means that the event follows the speech act (absolute use) or another event (relative use). See Malchukov 2001, Xrakovskij 2009 on the typology of taxis categories. Russian complementizers and adverbial subordinators differ with respect to their compatibility with absolute or relative uses of verbs forms in the embedded clause. For instance, the complementizer čto is compatible with both types of uses: (9) a. Ja znal čto on tam rabotal. I:nom knew:ipf.m that he:nom there worked:ipf.m ‘I knew that he worked there.’ b. Ja znal čto on tam rabotaet. I:nom knew:ipf.m that he:nom there work:ipf.3sg ‘I knew that he works there.’ c. Ja znal čto on tam budet rabotat’. I:nom knew:ipf.m that he:nom there be:fut.3sg work:ipf:inf ‘I knew that he would work there.’

In (9a), the past tense can be used absolutely or relatively. If the ‘absolute reading’ is chosen, the sentence means ‘I knew that he worked there before the speech act (and simultaneously to the moment when I knew it)’. In the ‘relative reading’, the sentence means ‘I knew that he had worked there earlier (before the moment when I knew it or find it out)’. The other two verb forms in (9b-c), not surprisingly,



A strange variant of Russian čtoby-construction 

can only be interpreted relatively. The present tense means ‘I knew that he worked there’ (the meaning is synonymous to the ‘absolute’ reading of (9a)), and the future tense means ‘I knew that he would work there after the moment when I knew it or find it out.’ In contrast, the adverbial subordinator kogda does not admit relative ­interpretation of any tense in the embedded clause. In (1a-c), all tense forms are interpreted absolutely, in other words, the past, present and future tense denotes the location of the event before, simultaneously or after the speech act, respectively. Note that no construction like (10) is possible: (10) *Kogda on prišel ja emu otdam den’gi.  when he:nom came:pf.m I:nom he:dat give:pf.1sg money:acc Intended: ‘When he comes I will return him the money.’

The relative interpretation of the past tense prišel ‘(he) came’ in the embedded clause (e.g. the location before the other event, otdam ‘(I) give’ is impossible). The absolute interpretation is impossible by another reason: the subordinator kogda ‘when’ requires that either both events in (10) take place in the future or both of them take place in the past. This, however, is not the case: one of the events (‘he came’) is situated in the future, the other one (‘I will return him the money’) in the past. 4.  Tripartite constructions In most languages, including Russian, a complex clause can include more than two simple clauses (these constructions are called ‘complex clauses with multiple embedding’ by Noonan 2007 and others). For instance, a temporal clause can be embedded under conditional clause to yield something like ‘If you see John when he is playing football tell me please’. Here I will focus on constructions where a temporal clause is embedded under a purpose clause or a clausal complement containing čtoby. The main issue I address is how the tense form in the ‘third’ clause, the temporal one with kogda is chosen. Grammatical descriptions do not say anything on tripartite constructions, neither in the descriptive nor in the prescriptive respect (see, for instance, Švedova (ed.) 1970, Švedova (ed.) 1980; as well as Brecht 1977 and Timberlake 2004). Neither are these constructions addressed in detail in typological literature on complex clauses. However, it would be logical to suppose that čtoby only applies restrictions on the verb form in its clause, not on the deeper one. Also, as said above, we know that kogda does not restrict the verb form in its clause in any way. Thus, the tripartite clause must follow the general pattern, the verb form in the temporal clause being chosen semantically.

 Alexander Letuchiy

This is really the case in most examples, e.g. in (11) and (12): (11) Ja ne pomnju slučaja čtoby kogda I:nom neg remember:ipf.1sg case:gen to when on podključalsja k operacii byvali provaly. he:nom participated:ipf.m to operation:dat was:iter.pl failures:nom ‘I do not remember cases when he participated in a [military] operation and it failed.’  [Vjačeslav Morozov. Admiral FSB] (12) Xoču čtoby kogda ja umru ona vspomnila want:ipf.1sg to when I:nom die:pf.1sg she:nom recalled:pf.f kakaja u nee … byla babuška. which:nom.f at she:gen   was:f grandmother:nom ‘I want that when I die she recalled which grandmother he had.’  [Anatolij Najman. Rasskazy o Anne Axmatovoj.]

For instance, in (11), the past tense is semantically motivated: the person the speaker mentions worked in the military forces before the speech act. Thus, this example illustrates an absolute use of the past tense: the tense is with respect to the speech act, not the time of other events described. Similarly, in (12), the future tense is also used absolutely. The speaker (Anna Axmatova) will presumably die after she says these words, thus, the future tense means that the time of her death is situated on the time axis after the moment of speech. In both of these examples, the tense in the temporal clause is motivated by the position of the event with respect to the moment of speech. However, there is another variant of the same construction which is much less usual and less frequently observed in the corpus data. In this variant, the verb form in the temporal clause is chosen based on the form in the čtoby-clause (the standard variant represented in (11) and (12) is also possible for these examples): (13) Ja xoču čtoby kogda ja prixodil domoj I:nom want:ipf.1sg to when i:nom came:ipf.m home:adv dver’ otkryvala žena. door:acc opened:ipf.f wife:nom ‘I want my wife to open (me) the door when I come home.’ [Film “Tišina”]. (14) (‘I want to have a flat in my mother city, Saint-Petersburg, with the ­following purpose:’) Čtoby kogda ja tuda priezžala mogla otdoxnut’. in-order-to when I:nom there came:ipf.f could:ipf.f rest:inf ‘In order to be able to rest there when I come to Saint-Petersburg.’

In (13), the past tense of the verb prixodit’ ‘to come’ cannot be used absolutely: the unreal situation ‘When I come home, the wife opens the door’ cannot take place



A strange variant of Russian čtoby-construction 

before the speech act - it is unreal, and can only be realized after the speech act. In (14), the situation is the same: the object of wish ‘When I come to Saint-Petersburg I have a place where to live’ can only be realized after the speech act. Neither can the past tense be used relatively, marking the position of the event in the temporal clause with respect to the event in the čtoby-clause. Under this interpretation, as long as the tense is past both in (13) and (14), the event ­designated in the temporal clause had to take place before the event in the čtoby-clause. But this is obviously not the case. While in (13) the event in temporal clause really precedes the one in the čtoby – though the latter directly follows the former which makes the relative use of tense rather improbable – in (14), the verb priezžat’ ‘to come’ in fact means ‘come and live for a while’, and thus, the situation in čtoby-clause occurs simultaneously with one in the temporal clause, rather then follows it (all the time when the speaker is in Saint-Petersburg she wants to have a place where to live). What is more important, as stated above, the relative use of tense is uncommon in Russian kogda-clauses. Recall that no construction like (10), repeated here as (15), is possible: (15) *Kogda on prišel ja emu otdam den’gi.  when he:nom came:pf.m I:nom he:dat give:pf.1sg money:acc ‘When he comes I will give him the money back.’

For the sake of systematic description, we should not analyze the cases in (13) and (14) as a relative use of tense that is possible in tripartite constructions, but not in bipartite ones.4 The only way to consider things like (13) and (14) is to propose that the past tense in the temporal clause is motivated by the complementizer čtoby which is situated in another clause, a purpose adverbial clause or a clausal complement. The tripartite structures under analysis call for the ‘tense agreement’ analysis proposed by Świdziński 1990 for some Polish constructions. The main clause can contain any possible tense form, the past tense of the predicate in the second clause (dver’ otkryvala žena in (13)) is entirely determined by the subordinator čtoby. But how will the language treat the third clause: will there be any ‘tense agreement’ between the third clause (kogda ja prixodil in (13)) and the main clause or the third clause will follow the pattern of the čtoby-clause? It could be tempting to mention ‘tense agreement’ when discussing (13) and (14). This term denotes cases when there is a requirement of identity of tense forms between the two parts of complex clause. However, the case in (13) and (14) seem to be in a sense opposite to this phenomenon. .  Note that the aspectual issue is left beyond the scope of the present article. At least it is evident that, though imperfective verbs prevail in the temporal clauses, as in (13) and (14), perfective forms are also used in this construction.

 Alexander Letuchiy

In tense agreement, the identity requirement is to some extent semantically motivated. For instance, the fact that the Polish subordinator ponieważ ‘because’ allow the tense forms of the verbs in the two clauses to be different, while póki ‘while’ requires them to be identical, follows from the semantics of the two units: póki denotes two simultaneous events, whereas for ponieważ, one of the situations (the reason) can precede the other one (the consequence). In contrast, the two variants in (11) and (12) vs. (13) and (14) do not seem to reflect different interpretations. The two constructions are synonymous, and the difference is only in the syntactic behavior. In the main variant, the tense is assigned to the verb in kogda-clause autonomously, based on the temporal c­ haracteristics of the event. In (13) and (14), čtoby is responsible for case assignment in kogda-clause. It is also important that the situation in (13) and (14) is typologically unusual: the situation when a particular complementizer or adverbial subordinator ­regulates the verb form in its clause is very widespread but the situation when a subordinator influences the verb form in the other clause is much more exotic. The grammaticality of the construction under analysis does not depend on the semantic function of čtoby. For instance, in (13), a clausal complement of the verb xotet’ ‘to want’ is build with čtoby, while in (14), čtoby is used as a purpose subordinator – but the past tense in the temporal clause is possible in both cases. The third, imperative-like use of čtoby is also compatible with the construction under analysis, though this construction is not found in literary texts, maybe due to the fact that this use of čtoby tends to occur in colloquial oral speech and its imitations: (16) Čtoby kogda ja prišel moi vešči byli to when I:nom came:pf.m my:pl.nom things:nom was:pl uže vystirany! already washed:part.pl ‘I want that when I come my clothes have already been washed!’ [An anecdote].

In this case, the construction is the same as in (13) and (14), except that what used to be historically the main clause is absent. The presence of čtoby in the main clause influences the verb form in the embedded temporal clause. The parameter which really matters for (im)possibility of the past tense is word order. Note that in (13), as well as (14), the temporal clause is inserted into the čtoby-clause (more precisely, it immediately follows čtoby). If the temporal clause followed the whole čtoby-clause the construction would become impossible: (13′) *Ja xoču čtoby dver’ otkryvala žena kogda  I:nom want:ipf.1sg to door:acc opened:ipf.f wife:nom when ja prixodil domoj. I:nom came:ipf.m home:adv ‘I want my wife to open (me) the door when I come home.’



A strange variant of Russian čtoby-construction 

The relevance of the word order evidences for the assumption that it is the marker čtoby itself which can sometimes impose its restrictions on the verb in the third (temporal) clause. Note that the standard variant with the present or future tense is always possible – not only can it replace the construction under analysis in (13) and (14), but remains acceptable in cases like (13′) where the word order is changed. The word order as in (13) and (14) is rather frequent in Russian. In RNC, there are 286204 examples of čtoby, 390262 examples of kogda, whereas the ordering čtoby kogda occurs in 163 examples. This figure may seem not really significant – but note that in Bulgarian one does not find in the National Corpus any examples of this sort (a few of them occur in the Google search, while za da ‘in order to’ is used 570449 times, and kogato ‘when’ occurs in 697937 contexts. Thus, in comparison to Bulgarian, Russian has a rather high percentage of contexts with insertion. Outside the Slavic family, the situation is very diverse. In any case, there are languages similar to Bulgarian, where insertion is not at all frequent: in the Arabic corpus (ArabiCorpus.byu.edu) no examples were found where ‘indama ‘when’ followed li ‘an ‘in order to’ or ‘an ‘to, that.’ Thus, if čtoby really influences the choice of the tense form in the temporal clause, the flexilibity of the Russian word order favours emergence of the construction under analysis, because constructions with insertion occur relatively often. Note that the variant illustrated in (2a) with infinitive in the embedded clause is not found in the tripartite construction under analysis – in other words, there are no examples analogous to (13) and (14) where infinitive occurs both in the čtoby-clause and in the temporal clause, even if the subject of the main clause is co-referent to the subject of the čtoby-clause and the temporal clause. This can be related to the fact that infinitive never occurs in usual, bipartite constructions with kogda – only finite forms, as in (1), are possible. The impossibility of the infinitive in the bipartite construction, which is more statistically frequent, causes its impossibility in the rare tripartite construction under analysis, as in (13) and (14). 5.  Semantic motivation of the past tense in the temporal clause Of course, the construction under analysis can be thought of as a purely ­formal type of the double embedding construction. However, the past tense in the temporal clause has a semantic motivation, too. Recall that the marker čtoby in all its uses bears a component of unreality in its semantics. It denotes that the situation in the čtoby-clause has not occurred and the speaker is not sure that it occurs, but only expresses its subjective attitude to this possible event.

 Alexander Letuchiy

Of course, the clause introduced by čtoby denotes an unreal event. For instance, the situation ‘The wife opens the door’ in (13) is unreal – by the time of speech the speaker’s wife often is out when he comes home and does not open the door, or even the speaker is not married, and his wife does not exist. The question whether the temporal clause is included into the scope of unreal semantic operator is much more difficult. On the one hand, this sentence can very well be used if the event in the temporal clause already takes place. For example, in (13) the speaker presumably comes home every evening, though his wife does not behave in the way he requires. Thus, the temporal clause can denote an unreal or real event. On the other hand, it is plausible to suppose that the unreal meaning in (13), and (14), and similar examples applies to both embedded clause (the clausal complement and the temporal clause). In this case, the meaning of this sentence can be reformulated in the following simplified way: ‘I want that the following situation take place: [I come home, by this moment my wife is at home and opens the door]’. Moreover, even if the speaker really wants only that the second part of the situation take place (‘my wife is at home and opens the door’), it does not prevent us from proposing that the first part (‘I come home’) is in the scope of the unreal operator. Note that this sentence can be used when the speaker is not yet married and, perhaps, does not work and is mainly at home. Thus, it can well be the case that the whole state of affairs including two subevents is under the scope of irrealis. In this case, the use of the past tense in the deepest embedded temporal clause in (13) and (14) is semantically motivated: the past tense form is a part of the subjunctive mood form, which is logically possible for the temporal clause, because it may denote an unreal event. 6.  Other complementizers Of course, a question arises whether the same construction is possible with other pairs of complementizers and adverbial subordinators. The question requires an additional search of data, but for the moment, almost no analogous examples have been found. In order to check it, we had to search for similar contexts with a pair of ­complementizers where the first one requires a particular verb form, while the second one allows use of any verb form, including the three tenses of the indicative mood and the subjunctive mood forms. Examples of the same construction where the first marker is čtoby and the second one a temporal subordinator, other than kogda, occurs sometimes in Google, but not in the corpus, and they are less numerous than contexts with kogda:



A strange variant of Russian čtoby-construction 

(17) Ja xoču čtoby kak tol’ko otkryvala forum tam I:nom want:ipf.1sg that as only opened:ipf.f forum:acc there srazu byla moja ljubimaja kartinka. right-away was:f my:nom.f favourite:nom.f picture:nom ‘I want that at the moment when I open the forum there was already my favourite picture.’

The statistical difference between kogda and other (e.g. temporal) subordinators results from the following two factors. First, in general kogda is by far the most frequent temporal subordinator. Second, only kogda admits all types of temporal relations between the two events: sequence, precedence, posteriority between the main and the embedded clause and, thus, there are no requirements or even tendencies regulating the verb form in the embedded clause. All other markers bear more narrow meanings: for instance, in sentential adjuncts kak tol’ko ‘as soon as’ denotes posteriority, while poka ‘while’ designates that the two events took place simultaneously. On the other hand, in Russian there are many constructions with complementizers and particles which require the verb to be in the past tense. These include, for instance, the unreal variant of the conditional subordinator esli by ‘if ’ and the same combination used as a particle with the meaning of wish, similar to English if only, as well as combinations net by and vot by which also express the meaning of wish. All of them require the verb in the same clause to be in the past tense, but no marker from this list occurs in the construction under analysis. For instance, in the construction where the marker esli by occurs in combination with kogda, followed by a verb in the past tense, the absolute use of the past tense form is observed. In (18), the speaker means that the whole sequence of events took part in the past, and the past tense priexal ‘(he) came’ denotes that the speaker came to Sochi before the moment of speech, thus, the tense is used absolutely: (18)  Esli by mne, kogda ja v 1995 godu priexal na festival’ v Sochi, kto-to skazal, chto ja budu general’nym direktorom AMiKa, ja by dolgo smejalsja.  ‘If anyone told me when I came to the Sochi festival in 1995 that I would be the General Manager of AMiK I would laugh for a long time.’

Some of my informants allow a non-canonical reading for similar contexts where the event expressed in the temporal clause has not yet taken place. Cf., for instance, (19): (19) Kak zdorovo bylo by esli by kogda on prišel how great was:neut part if part when he:nom came:pf.m tort uže stojal na stole. cake:nom already stood:ipf.m on table:loc ‘How great it would be if the cake was already on the table when he comes!’

 Alexander Letuchiy

Some native speakers consider two interpretations as possible for (19): (i) The person the speaker tells about has already come, and the cake was not on the table. The speaker says that it would have been great if the cake had been on the table (counterfactual conditional construction). In this case, the choice of past tense in the temporal clause is not related to the presence of esli by in the main clause. It denotes that the event (the arrival of the hero) took place before the moment of speech. (ii) The person the speaker tells about has not yet come, and the speaker says that it will be great (in the future) if the cake is on the table (real conditional construction). In this case, we observe the construction under analysis, where the use of the past tense prišel in the deepest (temporal) clause is motivated by the ­presence of esli by in the main clause. However, conditional constructions with the i­ nterpretation (ii) is never found in real texts. Similarly, the construction under analysis is not found with vot by and other expressions of wish. The question why the non-standard past occurs almost exclusively with čtoby is not at all simple. However, it supposedly results from the fact that the behavior of by in čtoby differs from its behavior in all other constructions. For instance, as mentioned below in Section 7, by in čtoby cannot be doubled by another by in the same clause, as opposed to all other instances of by. More precisely, by in čtoby seems to lose its autonomous status and does not behave as a particle – this is why it can be doubled by another particle by. It is thus not surprising that by in čtoby does not strictly obey to the restriction valid for all other instances of particle by – namely, that by in a clause cannot lead to apparition of a past tense form in another clause.

7.  By in the temporal clause The construction under analysis could seem to be just copying the verb forms under the influence of the marker čtoby. What makes the situation more complicated is the fact that the particle by sometimes occurs in the temporal clause: (20) Čtoby kogda narod videl by ego v jaščike, in-order-to when people:nom saw:ipf.m part he:acc in box:loc nemedlenno podčinjalsja immediately obeyed:ipf.m ‘So that the people obey him immediately when they see (saw) him in the box (i.e. in the TV)’.



A strange variant of Russian čtoby-construction 

In usual bipartite constructions with čtoby, by, as a rule, cannot be used in the embedded clause: (21) Ja sdelal èto čtoby menja (*by) ponjali. I:nom did:pf.m this:acc in-order-to I:acc   part understood:pf.pl ‘I did it so that they understand me.’

Note that the general rules of Russian admit two instances of by in the same ­sentence – for instance, the first of them can occur in the unreal variant of the ­conditional subordinator esli by ‘if ’, and the second one can be situated after the finite verb: (22) Esli by možno bylo by čto-to sdelat’… if part possible was:neut part something:acc do:pf.inf ‘If it only was possible to do something…’

However, čtoby rarely occurs with this type of doubling, which may be due to the fact that the particle inside čtoby is fully lexicalized. The fact that by becomes possible in tripartite constructions makes it evident that the past tense form in the temporal clause occurs not only under the influence of čtoby. If it was the case, the behavior of this temporal clause had to be the same as that of the čtoby-clause, in particular, the particle by had to be ungrammatical. However, examples like (20) show that the situation in these two clauses is not the same. While the properties of čtoby-clause is fully defined by the presence of čtoby, and by is almost impossible, since the complementizer already contains this particle, the temporal clause is embedded deeper, and čtoby is not a component of this clause. Thus, the presence of by is due to the fact that the temporal clause can fall into the scope of the unreal semantic operator, as said in Section 5. It is not the case that čtoby simply can impose its grammatical requirements to all verbs in the embedded clauses – the unreal meaning is present and can be manifested also by the particle by.

8.  Conclusions and explanations In the present article, a special Russian construction was discussed. Its peculiarity is that the presence of the complementizer/purpose subordinator čtoby influences not only the choice of the past tense form in the same clause, but also the ­occurrence of the same form in the deeper embedded temporal clause. Though Barentsen 1995 and Padučeva 1996 proposes that many of Russian verb forms can

 Alexander Letuchiy

have special uses and readings in embedded clause, there is no mention of any special properties of the tripartite constructions with double embedding. This is partly due to the fact that these constructions are relatively rare in literary texts and especially in the oral discourse. The Russian phenomenon may seem to enter naturally into the class of ­phenomena discussed by Świdziński 1990: in these examples, so-called ‘tense agreement’ is discussed. The term means that for some types of complex clauses, there is a condition that if one part has a tense characteristics, the same tense characteristics should be present in the other part. Moreover, the author shows that some tenseless forms, such as imperatives, are semantically (based on tense agreement) close to those bearing a particular tense (for instance, future for imperatives). However, there are important differences between the Russian case and the Polish constructions discussed by Świdziński. First of all, in the čtoby-construction the variant where the form of the verb in the kogda-clause is dictated by čtoby can always be replaced by the main variant where the tense in the kogda-clause is chosen semantically – in the examples pointed at by Świdziński, the tense agreement is obligatory. Second, the Russian construction very often contains the default temporal subordinator kogda which is not subject to any strict rules regulating temporal order of events. This makes a possible analysis of the variant illustrated by (13), (14) as tense agreement between the two parts of the construction problematic (there can be any temporal order between the two events, correspondingly, the hypothetic tense agreement would not correspond to any natural semantic relation between the events). The most straightforward way to explain the existence of the construction under analysis is to propose that the subordinator čtoby influences not only the clause it introduces but also the temporal clause. However, this fact demands an explanation why this is possible almost exclusively in combinations čtoby + kogda. The explanation seems to be as follows: čtoby is one of the few Russian subordinators which strictly regulate the choice of the verb form. In contrast, kogda is the least restrictive subordinator, both formally (the verb in the temporal kogdaclause can be in present, past, future or subjunctive, see (1)) and semantically (the two events in the kogda-construction can be situated in different ways on the time axis). As a result, the boundary of the temporal clause, marked with kogda, which does not impose its own restrictions, is transparent for the influence of čtoby. The fact that the temporal clause often occurs immediately after the marker čtoby facilitates the penetration of the grammatical features imposed by čtoby into the following (temporal) clause. At the same time, another possible explanation, namely that kogda behaves in examples like (13) and (14) as a co-ordinate conjunction, and the past tense (e.g.



A strange variant of Russian čtoby-construction 

prixodil and otkryvala) in (13) in a sense mark two members of the conjunction relation does not seem to be true. Examples like (20) show that čtoby-clauses and kogda-clauses do not behave symmetrically, e.g. with respect to the occurrence of the particle by. In the clause introduced by čtoby by is rarely doubled and this doubling is not fully grammatical (see (21)), while in the temporal clause the doubling becomes grammatically perfect. What is important, and this fact is also illustrated by examples like (20), is that the choice of the past tense in the temporal clause has semantic grounds. If the čtoby-clause marks an unreal situation, it is likely that the temporal clause is also inside the scope of this unreal semantic operator. The question whether the unreal meaning spreads to all language material embedded under the irrealis marker or only within one clause requires a thorough typological consideration. The data analyzed above are important for the general theory of s­ ubordination and complex clauses. In the syntactic research, the authors (e.g. Świdziński 1990; Noonan 2007) mostly restrict themselves with bipartite sentences assuming (explicitly or implicitly) that the behavior of structures with double embedding can easily be derived from structures with single embedding. This, however, is not the case. In the Russian construction under analysis, the verb form in the main clause is relevant for the choice of the form in the third clause – thus, tripartite structures with double embedding have a unique property which cannot be discussed on standard structures with single embedding. This means that the analysis of structures with double embedding can also be useful for descriptions of other languages and typological investigations. Our data also put forward the whole group of problems related to the t­ ripartite constructions with double embedding. These constructions contain two embedded clauses. Each of them can have its own formal and semantic properties, and their relation to each other, as well as to the main clause are far from being trivial. These constructions allow multiple variants of temporal reference and uses of tense forms in each of the clauses – this variance is inaccessible in usual c­ onstructions with one main and one embedded clause.

Acknowledgements I am grateful to Sergey Say and the two anonymous reviewers for his valuable comments and remarks. I would also like to thank the audience of my talks at the conferences “Slovo i jazyk” (Moscow, Russian Academy of Sciences, February 2-4, 2010) and “The 6th meeting of the Slavic Linguistic Society” (Aix-enProvence, University of Provence, September 1-3, 2011) for their questions and comments.

 Alexander Letuchiy

Abbreviations 1, 2, 3 acc adv com f fut gen imp ipf irr iter

= first, second, third person = accusative = adverb = comparative = feminine = future tense = genitive = imperative = imperfective = irrealis = iterative

m = masculine n = neutral neg = negation nom = nominative part = subjunctive particle pf = perfective pl = plural prs = present tense pst = past tense sg = singular

References Barentsen, Adrian. 1995. Shifting points of orientation in Modern Russian. Tense selection in ‘reported perception’. In Reported Speech: Form and Functions of the Verb [­Pragmatics & Beyond New Series 43], Theo A.J.M Janssen & Wim van der Wurff (eds), 15–55. ­Amsterdam: John Benjamins. Brecht, Richard. 1977. Čtoby or čto and by. Folia Slavica 1: 33–41. Dobrushina, Nina R. 2012. Subjunctive complement clauses in Russian. Russian Linguistics 36(2), 121–156. Malchukov, Andrej L. 2001. Opyt isčislenija taksisnyx otnošenij. In Issledovanija po jazykoznaniju. Festschrift for A.V. Bondarko, Sadje A. Shubik (ed.), 186-197. Saint-Petersburg: Nauka. Noonan, Michael. 2007. Complementation. In Language typology and syntactic description, Timothy Shopen (ed.), 52-150. Cambridge: CUP. Padučeva, Elena V. 1996. Semantičeskie issledovanija. Мoscow: Jazyki russkoj kul’tury. Panov, Michail V. 1966. Russkij jazyk. In Jazyki narodov SSSR, Vol. 1: Indoevropejskije jazyki, 55–122. Moscow: Nauka. Švedova, Natalia Ju. (ed.). 1970. Grammatika sovremennogo russkogo literaturnogo jazyka. ­Moscow: Nauka. Švedova, Natalia Ju. (ed.). 1980. Russkaja grammatika. Moscow: Nauka. Świdziński, Marek. 1990. Interclausal syntactic connections in contemporary Polish. Revue des Etudes Slaves 62(1–2): 427–438. Timberlake, Alan. 2004. A Reference Grammar of Russian, Cambridge: CUP. Xrakovskij, Viktor S. (ed.). 2009. Tipologija taksisnyx konstrukcij, Мoscow: Znak.

section iii

Impersonal constructions

Impersonal constructions in Serbian A description within a MeaningText linguistic model Jasmina Milićević

Dalhousie University, Halifax, Canada The paper discusses impersonal constructions in Serbian within the Meaning-Text framework. It provides a definition and a typology of impersonal constructions, based on the notions of semantically full/empty and phonetically non-zero/zero subject lexemes, and describes the interaction of impersonal constructions with verbal voices and derivation. Cet article porte sur les constructions impersonnelles serbes telles qu’envisagées au sein de la théorie linguistique Sens-Texte. Il propose une définition et une typologie des constructions impersonnelles, s’appuyant sur les notions de sujet sémantiquement plein/vide et de sujet phonétiquement non-zéro/zéro, et décrit l’interaction de ces constructions avec la voix et la dérivation verbales. Keywords:  dependency syntax; impersonal constructions; Serbian; verbal voice; verbal derivation

1.  Introduction The paper deals with impersonal constructions in Serbian from the viewpoint of the Meaning-Text linguistic theory (Mel’čuk 1974 and 1988: 43–91; Kahane 2003), more specifically its dependency approach to syntax (Mel’čuk 1988 and 2009). ­Section 1 provides a definition of the notion of impersonal construction and ­discusses its uses, especially in light of its interaction with grammatical voice and derivation. By interaction I mean the fact that some impersonal constructions participate in an essential way in the implementation of some voices and verbal derivations, which tends to blur distinctions between these phenomena. Section 2 offers a sketch of a formal description of the impersonal constructions in Serbian. Section 3 is reserved for a conclusion.

 Jasmina Milićević

1.1  What is an impersonal construction? Let me start by the following definition (adapted from Mel’čuk 1995: 17ff and 2006a: 470ff): An impersonal construction [= IC] is a surface-syntactic construction whose subject is either a semantically full indefinite impersonal pronoun, i.e. such that it is vaguely interpretable as some kind of (natural or supernatural) ‘force’, although it lacks reference, or a semantically empty, i.e. dummy, pronoun.

Both semantically full indefinite impersonal and semantically empty pronouns functioning as subjects can be non-zero (overt, i.e. having phonetic realization) or zero (covert, i.e. with no phonetic realization), depending on the language ­considered. These distinctions allow for the following typology of ICs (a zero ­subject is notated “Ø”): 1. ICs with a semantically full (indefinite impersonal) pronoun as subject 1a. non-zero

This is a logical possibility for which I do not have an attested example.

1b. zero (1) a. Rus. Ø (3, SG, NEU) EgoACC ubil+o3, SG, NEU

‘[It] him killed’ ≈ ‘A natural force killed him’.

b. Alutor Ø(3, PL) Na +ininitə +tkəni +ɣəm [They] 3plsubj cause.rash pres 1sgobj

≈ ‘Some natural forces cover me with rash’. (Kibrik et al. 2004: 397)

2. ICs with a semantically empty (dummy) pronoun as subject 2a. non-zero (2) a. Ger. Es regnet ‘It rains’. b. Fr.  Il arrive des accidents ‘It happens some accidents’ = ‘Accidents happen’.

2b. zero

(3) a. It. Ø (3, SG) Piove ‘Rains’ = ‘[It] rains’.

b. Lat. Ø (3, SG) MeACC pudet ‘[It] shames me’ = ‘I am ashamed’.

Semantically empty pronouns could be further divided into meteorological (­similar to full indefinite impersonal pronouns in that they allow, at least in some ­languages, the ‘force’ interpretation) and expletive (genuine empty pronouns); cf. (2a) and (3a), on the one hand, and (2b) and (3b), on the other. However, I do



Impersonal constructions in Serbian 

not think that this distinction is warranted in all languages; at any rate, it does not seem to be the case for languages of common European type, where all semantically empty pronouns formally behave in the same way (namely, they impose the same type of agreement). In other words, it does not seem economical to foresee two lexemes eng. IT, fr. IL, ger. ES, etc. one meteorological and one expletive. The existence of zero subjects in (1) and (3) is postulated because they trigger verbal agreement: 3sg agreement of the main verb, accompanied in some languages by gender agreement of the main verb or the participle.1 Zero subjects of which I am talking here are real lexical units, supplied with their own dictionary entries, being “special” only in the sense that they have a phonologically null signifier. They are “physically” present in the syntactic structures of corresponding sentences and should not be confounded with elements that have undergone deletion during sentence production (like dropped subjects of declarative sentences in a PRO-Drop language or dropped subjects of imperatives).2 According to our definition, semantically full indefinite personal pronoun i.e. a pronoun meaning ‘some unspecified people’, whether non-zero (cf. (4a)) or zero (cf. (4b)), does not define an IC.3 (4) a. Fr.  On distingue plusieurs types de toxines… lit. ‘One distinguishes several types of toxins…’ Eng. They raised taxes on gasoline yet again. b. Rus. Ø  (3, PL) Ego ubil+ i3, PL lit. ‘[They] him killed’ ≈ ‘Some people killed him’. Serb. Ø  (3, PL) Kaž+u3, PL da je to davno bilo lit. ‘[They/People] say that happened long ago.’

.  In languages with no agreement, such as Lezgian, where no underlying subject can be postulated, sentences corresponding to some null subjects sentences above are considered subjectless, i.e. lacking the subject altogether (Mel’čuk 1988: 228–230). .  See Weiss (this volume) for a distinction between zero lexeme in the syntactic structure and ellipsis of a lexeme from the syntactic structure. .  In some accounts, for instance in Siewierska (2007) and Creissels (2008), such pronouns are accepted as subjects of ICs; Radovanović 1990 refers to sentences like the second one in (4b) as impersonal. Generally speaking, the range of subjects admitted as defining an IC is quite broad (cf. Krzek 2010: 67): basically any non-canonical subject qualifies, a canonical subject being a subject realized by a verbal argument that is fully referential and manifests the typical morphosyntactic properties of subjects in the language considered. Different criteria are used to define an IC (Siewierska, op. cit.): (1) semantic (lack of agentivity; non-specific human reference); (2) syntactic (no subject or only a pleonastic = semantically empty subject, overt or not); and (3) morphological (the main verb lacking any personal specification or being invariably 3pl).

 Jasmina Milićević

As for empty pronouns like ES in ger. Es comen drei Studenten ‘There come three students’ and its English equivalent, THERE, they are not subjects (because the verbs do not agree with them) but rather quasi-subjects, which means that the corresponding construction is not impersonal: at best, it could be called pseudo-impersonal. Turning now to ICs in Serbian, I will tentatively assume that this language has only the IC of type 2B, i.e. the IC with semantically empty dummy subject. The norm for a Slavic language is to have the ICs of types 1B and 2B (although some Slavic languages – Sorbian, Belarus, Northern Russian dialects – do appear to also have ICs defined by dummy non-zero subjects; cf. Franks 1995: 323ff). (5) Serbian a. Ø (3, SG, NEU) Vrte +l +o mi se u glavi. Turn(Vrefl) act.part 3, sg, neu I.dat refl in head ‘It turned in my head’ = ‘I was dizzy’. b. Ø (3, SG, NEU) Svanu +l +o je. Dawn act.part 3, sg, neu is ‘It dawned.’

Although Serbian has some constructions that at first sight seem to feature (semantically full) indefinite impersonal subjects, for instance (5a), the ‘natural force’ interpretation is not entirely convincing. (I have not come upon any such interpretation in the literature, which, without proving anything, is nevertheless significant.) Also, from a formal viewpoint the “suspect” construction in (5a) does not differ in any way from the construction in (5b), which is without any doubt of type 2B: in both cases, the main verb exhibits the 3 sg agreement and the participle the 3 sg neuter agreement with the zero subject.4 Finally, positing the two ICs would lead in some cases to overly fine and, to my mind, unnatural distinctions (for instance, we would have to treat differently Hladno mi je ‘Cold to.me is’ = ‘I am cold’ and Hladno je ‘Cold is’ = ‘It is cold’; i.e. to consider the former an instance of the IC with the indefinite impersonal subject and the latter as implementing the IC with the dummy subject). This is why I prefer to stick to a single IC, at least for the time being. 1.2  Impersonal constructions & voice/derivation In much of relevant literature, impersonal has been opposed to passive and considered a “distinct grammatical category” with respect to the latter. (It is not clear .  The main verb, the clitic past auxiliary, je ‘is’ has actually been deleted from sentence (5a) because the clitic sequence *[se je] is unacceptable: Vrtelo mi *[se je] u glavi ⇒ Vrtelo mi se u glavi.



Impersonal constructions in Serbian 

whether this means ‘a different grammeme of the inflectional category of voice’ or ‘an altogether different inflectional category’.) For instance, sentence (6a) is considered to be a subjectless passive and the one in (6b) impersonal (the examples are taken from Ðurković 2004 and Blevins 2006). (6) a. Serb. Marku je preće +n +o To.Marko is threaten pass.part 3, sg, neu

od strane ubica on part of. killers



‘Marko was threatened by the killers.’

b. Pol. Tu się pije wódkę Here refl drink.3sg vodka.acc ‘One drinks vodka here.’

For us, (6a) is indeed an instance of the passive, more precisely full demotional passive. As for (6b), while we agree that it is not passive, we do not treat it as impersonal tout court but rather as a particular voice type called (subjectless) suppressive.5 At this point, a word on the MTT theory of voice (Mel’čuk 2006b) is in order. Voices involve modifications of the lexicographic, or basic, diathesis of the verb, without (substantially) changing its propositional meaning, the diathesis being a correspondence between the verb’s Semantic and Deep-Syntactic Actants (roughly, what is known as linking). Active corresponds to a zero modification of the basic diathesis, while the three logically possible types of modification (and their combinations) – permutation, suppression and referential identification – yield passive, suppressive and reflexive voice types. Passives are further differentiated along two axes: “full ~ partial” and “promotional ~ demotional”, while suppressives and reflexives are further divided into subjectless, objectless and absolute. MTT thus operates with a richer inventory of voices than most frameworks.6

.  Interestingly, Blevins (op.cit.) mentions “suppression of the reference to the subject” as a primary function of impersonals. Suppression is mentioned also in Krzek (2010) in connection with the Ukrainian and Polish -to/-no constructions, which for us are instances of the suppressive. .  This is of course a major reason for divergent data analyses, like the one just illustrated. Further example: our absolute suppressive (Ovde se dobro živi ‘Here refl well lives’ = ‘One lives well here’) is called impersonalized/deagentivized clause (Serbian grammar), reflexive impersonal (Krzek, op. cit.) and surface impersonal (Ðurković, op.cit). The differences are not merely terminological, though: some constructions that we do not treat as impersonal are regularly considered as such; in particular, the reflexive passive (Kuća se gradi ‘House.nom refl builds’ = ‘The house is being built’) and deontic/modal expressions taking a clausal complement (Treba [da ga zovem] ‘Should [that(Conj) him call.1sg]’ = ‘I should call him’).

 Jasmina Milićević

For the verbs in (6), the following diathesis modifications apply (basic diathesis—in grey shading—is a part of the Government Pattern of a lexeme L, X, Y, etc. being the variables filling L’s semantic actantial slots and I, II etc. its deep-syntactic actants; the second part of the Government Pattern specifies the surface expression of the actants): Serb. PRETITI(Vtr) ‘to threaten’ active X⇔I

Y ⇔ II

Z ⇔ III

Nnom

Ndat

Ninstr da ‘that’ PROP

ful demotional passive

⇒

X ⇔ IV

Y ⇔ II

Z ⇔ III

od (strane) ‘on part’ Ngen

Ndat

Ninstr da ‘that’ PROP

In the modified diathesis, X, the author of the threat, gets demoted to DSyntactant IV, while Y and Z, the person threatened and the content of the threat, respectively, stay in place. Note that Z is not expressed in sentence (6a); cf. the following sentence, featuring an expression of this actant: Marku Y= II = Indirect.Object je prećeno smrću 〈da će biti ubijen〉Z=III=Oblique.Object ‘To.Marko is thretened with. death 〈that [he] will be killed〉’. Pol. pić(Vtr) ‘to drink’

active subjectless suppressive X⇔I

Y ⇔ II

Nnom

Nacc

⇒

X⇔—

Y ⇔ II Nacc

In the modified diathesis, X, the Agent, cannot be expressed, while Y, the Patient, stays in place. This is an example of suppression applied to a transitive verb, which retains its direct object. Crucially, a demotion or suppression of the DSyntA I can require, in a non PRO-Drop language and/or a language with agreement, the introduction of a dummy (zero or non zero) subject. This operation does not affect the diathesis of the verb; it only concerns its surface-syntactic actants and is carried out in the transition towards the surface-syntactic level of representation (for illustrations, see Section 2). Another example of this “automatic” introduction of a dummy subject—the dummy non zero IL(pron, impers)—is provided by the French sentence in (7): (7) Fr. demotional (impersonal) passive IlSubj a été procédé à une enquêteY=II=OblObj par la policeX=III=AgentCompl. ‘It has been proceeded to an inquiry by the police’. cf.: La policeX=I=Subj a procédé à une enquêteY=II=IndirObj ‘The police proceeded to an inquiry’.



Impersonal constructions in Serbian 

However, in some instances the impersonalization can be freely chosen by the speaker—for communicative or stylistic reasons; cf. the following examples from French, where each voice has a personal and impersonal variants. (8) Fr. active a. Trois personnesX=I=Subj sont venues

‘Three persons have come’.

b. IlSubj est venu trois personnesX=I=Quasi-subj ‘It has come three persons’. (9) Fr. full promotional passive a.  Deux parapluiesY=I=Subj ont été volés par mon frèreY=II=AgentCompl cleptomane ‘Two umbrellas have been stolen by my kleptomaniac brother’. b.  IlSubj a été volé deux parapluiesY=I=Quasi-subj par mon frèreY=II=AgentCompl cleptomane It has been stolen two umbrellas by my kleptomaniac brother’. (10) Fr. reflexive (short) passive a. Des journauxY=I=Subj se vendent ici

‘Newspapers refl sell here’.

b. IlSubj se vend des journauxY=I=Quasi-subj ici ‘It refl sells newspapers here’.

The opposition “personal ~ impersonal” is thus orthogonal to voice oppositions, although in some languages some voices are always used in impersonal constructions. In other words, impersonal constructions are in principle compatible with any voice, as well as with different verbal derivations. Valence-changing deverbal derivation modifies the basic verbal diathesis, just like voices do, but, in contrast to them, it modifies as well the meaning of the verb in a substantial way. Here is an example of derivation from Serbian, coupled with an obligatory impersonalization (note the 3 sg neuter agreement of the verb); the resulting construction is called dispositional impersonal dative (the corresponding derivation rule is provided in Section 2). (11) Vesn+i se spava +l +o Vesna dat refl sleep act.part 3, sg, neu ‘Vesna felt like sleeping 〈singing, walking, …〉’

〈pevalo, šetalo, …〉

cf.: Vesn+a je spava+l +a Vesna nom is sleep act.part 3, sg, fem ‘Vesna slept 〈sang, walked, …〉’

〈pevala, šetala, …〉

In addition to these two uses, i.e. a free use to reflect a communicative or ­stylistic intent of the speaker and an ancillary use to “service” voices and derivations, ICs can be used inherently, as it were; this is the case of the constructions in (2a), (3a) and (5b), implementing meteorological expressions, as well as those in (3b) and (5a), featuring expressions of psychological/physiological states and the like, which involve lexical units lacking the Deep-SyntA I in their basic diathesis

 Jasmina Milićević

(impersonal predicates in the sense of Babby 1989). Voice-wise, one could say that these i­ nherent impersonal constructions are active or reflexive by default; alternatively, they could be deemed not to have voice at all (because no variation along voice lines is possible). 2.  Impersonal constructions in Serbian As has been established in 1.1, Serbian has only one type of IC. The subject of this IC is a semantically empty, surface lexeme Ø(3, SG, NEU);7 the construction can be used inherently, with lexical impersonals, or in the ancillary capacity, as one of the means contributing to the implementation of certain voices and verbal derivations. 2.1  The dummy zero subject IC with lexical impersonals This use of the Serbian IC was first illustrated in (5b); additional examples follow. (12) a. Grme +l +o je thunder act.part 3, sg, neu is ‘It thundered.’ b. Bi +l +o miX=II= Indir.Obj je hladno Be act.part 3, sg, neu I.dat is cold ‘I was cold’ c. Probada +l +o gaX=II=Dir.Obj je u leđimaY=III=Obl.Obj Stab act.part 3, sg, neu he.acc is in back ‘He was feeling a stabbing pain in his back’ cf.: Probada +l +a su Stab act.part 3, pl, fem are ga Y=II=Dir.Obj leđaX=I=Subj he. acc back(pl.tant, neu).pl.nom

‘His back was giving him a stabbing pain.’

As mentioned previously, voice-wise these constructions are active or reflexive by default. These are the “real” impersonals, in the sense that everyone treats them as such. They include, first of all, what could be called zero-place predicates, i.e. lexical

.  A surface lexeme is a lexical unit appearing at the surface-syntactic level of representation. MTT distinguishes Deep- and Surface representation levels in syntax. Basic structures at both levels are dependency trees; they differ with respect to linguistic elements labeling the trees: deep- (semantically full) vs. surface (both semantically full and structural) lexical units and deep- (cross-linguistically valid) vs. surface (language-specific) dependency relations.



Impersonal constructions in Serbian 

units that do not have semantic actants at all (meteorological verbs and expressions, like the one in (12a)) and, consequently, lack DSynt actants, as well as lexical units without an inherent DSynt actant I. The verb in (12c) admits both the impersonal and “personal” constructions, the former being the unmarked variant; this variation is handled via the verb’s Government Pattern modifications (cf. Mod-1 and Mod-2 below). There is a small number of verbs denoting unpleasant sensations that follow this pattern (Milićević 2009: 107–114). Here are the Government Patterns of the lexical impersonals featured in the sentences (12b) and (12c) (GRMETI ‘to thunder’ featured in sentence (12a), as a zero-place predicate has no Government pattern): HLADNO(Adv) ‘cold’

PROBADATI(Vintr) ‘to stab’ = ‘cause stabbing pain’ Mod-1



Mod-2

X ⇔ II

X ⇔ II

Y ⇔ III

Ndat

Nacc

u ‘dans’ Nloc

DSynt-level ⇔ SSynt-level GRMETIIND,PAST BITI(Aux)PRES

subjectival

auxiliary

Y⇔I Nnom

DSynt-level ⇔ SSynt-level BITI(CoP)IND,PAST II

BITI(AUX)PRES

HLADNO subjectival copular

II Ø (3, SG, NEU) GRMETIACT.PART

X ⇔ II Nacc

JA

BITI(COP)ACT.PART Ø (3, SG, NEU) attributive HLADNO indir.objectival JA

2.2  The dummy zero subject IC with voices Three voices are “serviced” by the IC in Serbian: full demotional passive, absolute suppressive and subjectless suppressive.8 .  These are minor voices, applicable to restricted classes of verbs, to which we can add short  (reflexive) passive (Ovaj lek se pije posle jela lit. ‘This medication refl drink after

 Jasmina Milićević

(13) Full demotional passive O tomeY=II je pisa+n +o About that is write pass.part 3, sg, neu u novinama (od strane stručnjakaX=III). in newspapers on part of. experts ‘It was written about that in the newspapers (on part of experts).’ ‘Experts wrote about that in the newspapers.’ (14) Subjectless suppressive O tomeY=II se pisa+l +o u novinama. About that refl write act.part 3, sg, neu in newspapers ‘Newspapers wrote about that’. ‘It was written about that in the newspapers.’ (15) Absolute suppressive a. U Siriji se ginu+l +o za slobodu. In Syria refl die  act.part 3, sg, neu for freedom ‘In Syria people were dying for freedom’. b. Al’ se nekad dobro je +l +o.9 Really refl then well eat act.part 3, sg, neu ‘Boy, didn’t people eat well in old days’.

The diatheses of the verbs involved, as well as the corresponding diathetic modifications, follow: PISATI(Vintr) ‘to write (about)’ active X⇔I

Y ⇔ II

Nnom

o ‘about’ Ndat

⇒

demotional passive X ⇔ III

Y ⇔ II

od (strane) ‘on part’ Ngen o ‘about’ Ndat

subjectless suppressive ⇒

X⇔—

Y ⇔ II o ‘about’ Ndat

eating = ‘This medication is to be taken after a meal’) and objectless reflexive (Jovan se brije lit. ‘J. refl shaves’). Major voices in Serbian include active and full promotional—participial and reflexive—passive (Penzioni sistem u Srbiji je finansiran 〈se finansira〉 od strane države lit. ‘Pension system is Serbia is financed 〈refl finances〉 on part of.state’ = ‘Pensions in Serbia are financed by the State’). .  Note, again, the deletion of the auxiliary from sentences (14) and (15), for the same reason as in (5a).



Impersonal constructions in Serbian 

(PO)GINUTI(Vintr) ‘to die violently’

JESTI(Vtr) ‘to eat’

active absolute suppressive active X⇔I Nnom

⇒

X⇔—

X⇔I

Y ⇔ II

Nnom

Nacc

absolute suppressive ⇒

X⇔—

Y⇔—

It is only normal to have recourse to an IC with these voices, since all of them make “their own” subject inexpressible – and a subject is needed for verb agreement. It is important to note that the IC is not a means of implementation of these voices, which have their own expressions means: the -n participle and the reflexive marker se, respectively. Demotional passive is used with a small set of intransitives, in particular communication and opinion verbs, either bi-actantial, as above, or three-actantial, as in (6a). The expression of the demoted Agent is strained, as it is with the participial passive and, in particular, reflexive passive (this is a well-known feature of Serbian). Subjectless suppressive is formed from at least bi-actantial intransitive verbs, while absolute suppressive can be formed either from intransitives (15a) or from intransitively used transitives (15b). Unlike Polish (cf. (6b)), Serbian does not allow suppressives from transitive verbs retaining their direct object.10 Although the Agent of the verb in the suppressive cannot be expressed, it is very much present in the background as the corresponding sentences are interpreted as having unspecified human reference. In fact, the human reference interpretation is forced upon any verb in the suppressive (Mjaukalo se ‘There was meowing’ necessarily implies ‘by people’), a phenomenon well known from other languages having this voice (cf. Siewierska 1984: 100ff ; Blevins 2006: 236; Krzek 2010: 75). To the extent that this specific interpretation could be regarded as adding meaning, something that voice modifications are not supposed to do, are we really dealing with voices in cases like (14) and (15)? In fact, there are voices that contribute meaning to some extent: potential passive in several languages, the adversative passive in Japanese, the Mayan passive, which adds the meaning ‘on purpose’, are cases in point. Alternatively, we could say that the unspecified human reference is a condition for the use of the suppressive.

.  In this respect, Croatian is like Polish; cf. the following subjectless suppressives formed from bi-transitive verbs: Stanare se isterivalo silom lit. ‘Tenants-acc refl evict-past.3.sg.neu by force’ = ‘Tenants were evicted by force’, Vozačeacc se upozorava na sklizak kolnik lit. ‘Driversacc refl warn-pres.3.sg.neu at slippery road’ = ‘Drivers are warned of slippery roads’.

 Jasmina Milićević

Here are the SSyntP and SSyntS for sentence (14): DSynt-level

SSynt-level

⇔

PISATISUBJ.SUPPRESSIVE, IND, PAST ATTR II

BITIPRES subjectival

Loc-in

auxiliary

Ø (3, SG, NEU)

PISATIACT.PART

II

TO

adverbial

oblique.objectival

NOVINEPL

aux.relfexive

O

prepositional

U

prepositional

TO

SE

NOVINEPL

2.3  The dummy zero subject IC with verbal derivations Dispositional, illustrated in (11), is an example of verbal derivation accompanied by the dummy zero subject IC. This is a very productive derivation; it can take as base any intransitive or intransitively used verb P, adding to ‘P’ the meaning ‘feel like’. At the DSynt-level, the derivateme feel.like appears as function of the verbal stem; at the SSynt-level, it is marked by se, while the Experiencer is expressed as an obligatory indirect object and a zero dummy inserted for agreement. Here are the two structures for the sentence Vesni se spavalo ‘Vesna felt like sleeping’: SPAVATI(Vintr) ‘to sleep’ active dispositional X⇔I X ⇔ II ⇒ Ndat Nnom

DSynt-level

SSynt-level

⇔

BITIPRES

feel.like(SPAVATI)IND, PAST

OBLIG

II VESNA

refl-aux subj auxiliary SE

Ø (3, SG, NEU)

SPAVATIACT.PART indir-obj VESNA



Impersonal constructions in Serbian 

The corresponding derivation rule looks like this: Sem-level

⇔ ‘P’

‘feel like’

2

1

1 ‘X’

DSynt-level

SSynt-level

⇔

L(‘P’)

feel.like(L(‘P’)) aux.refl II L(‘X’)

subj

indir-obj SE

L(‘X’)

Ø (3, SG, NEU)

To close this subsection, let us note that there is a variety of dispositional from transitive verbs, called dispositional passive dative, not accompanied by an IC; for instance: (16) Je +l +i su miY=II=IndirObj se kolačiX=I=Subj Eat act.part 3, pl, masc are I.dat refl cookie(masc).pl.nom ≈ ‘Cookies were the target of my gastronomic desire’ ‘I felt like eating cookies.’ cf.: AnaX=I=Subj je je +l +a kolačeY=II=DirObj Ana.nom is eat act.part 3, sg, fem cookie(masc). pl.acc ‘Ana ate cookies.’

3.  Conclusion The IC has been characterized as a surface-syntactic construction featuring a semantically full indefinite impersonal pronoun or a semantically empty pronoun in the role of subject. Three uses of ICs have been identified and illustrated using data from Serbian, which has only one type of IC—a dummy zero subject IC. 1. Free use, reflecting the communicative or stylistic intent of the speaker. In Serbian, this use is restricted to cases like (12c), where a lexeme allows ­Government Pattern modifications and one of these involves the impersonal construction. (In French, and perhaps other non Pro-Drop languages, the IC with a dummy non-zero as subject is used in this way; cf. French examples (8b), (9b) and (10b)). 2. Inherent use, with lexical impersonals, i.e. lexical units lacking ASyntP I in their basic diathesis; cf. examples (5) and (12). 3. Ancillary use, to provide a subject for voices and verbal derivations with which the verb has become subjectless as a result of diathesis modification; cf. examples (6), (13)–(15) and (11), respectively.

 Jasmina Milićević

The approach presented here defines independently voice modifications (and ­verbal derivation), on the one hand, and impersonal constructions, on the other, and then describes the way they interact. This is possible, to a large extent, because of the fact that we use a richer inventory of voices than is usually the case so that the distinctions between different voices do not hinge on the possible or obligatory presence of an impersonal construction. Why not say that the impersonal construction is the marker of a particular voice (or derivation), if it always accompanies it? Because it would blur the picture: the IC does participate in the implementation of some voices (on surface-syntactic level), but the voices have their own markers (se, -l, and -n participles). A parallel can be made with the cases like Ger. Nacht+Ø ‘night’ ~ Nächt+e ‘nights’, where the pluralization is accompanied by an alternation in the stem: however, the alternation a ⇒ ä does not mark the plural, the suffix does (although the alternation “reinforces” the information carried by the suffixation).

Acknowledgments I am grateful to François Louis, Igor Mel’čuk and Wayles Browne for their comments on a previous version of this paper. Thanks are also due to two anonymous reviewers, whose remarks helped me improve some important aspects of the text.

References Blevins, J. 2006. Passive and impersonal constructions. In Encyclopedia of Language and ­Linguistics, Vol. 9, 2nd ed., Keith Brown (ed.), 236–239. Amsterdam: Elsevier. Babby, Leonard H. 1989. Subjectlessness, external subcategorization and the p ­rojection ­principle, Zbornik za filologiju i lingvistiku 32: 7–40. Reprinted in Journal of Slavic ­Linguistics 10: 341–388. Creissels, Denis. 2008. Impersonal and Related Constructions: A Typological Approach, ­Lectures given at the University of Tartu, June 02–03. Franks, Steven. 1995. Null subject phenomena. In Parameters of Slavic Morpho-Syntax, Steven Franks, 287–333. Oxford: OUP. Ðurković, Milja 2004. Passive and Impersonal in English and Serbian, Newcastle Research Centre for English and Applied Linguistics Working Papers 10: 49–101. Kahane, Sylvain. 2003. The Meaning-Text theory. In Dependency and Valency. An I­ nternational Handbook of Contemporary Research, Vol.1, Vilmos Agel, Ludwig M. Eichinger, ­Hans‐­Werner Eroms, Peter Hellwig, Hans Jürgen Heringer & Henning Lobin (eds), ­546–570. Berlin: Mouton De Gruyter. Kibrik, Alexander E., Kodzasov, Sandro V. & Muravyova, Irina A. 2004. Language and Folklore of the Alutor People, Megumi Kurebito (ed.). Suita: Ōsaka Gakuin Daigaku Jōhōgakubu/­ Nakanishi Printing (Japan).



Impersonal constructions in Serbian 

Krzek, Malgorzata. 2010. Some Aspects of subjects of impersonal constructions in Polish.­ Newcastle Working Papers in Linguistics 16: 66–87. Mel’čuk, Igor. 1974. Opyt teorii lingvističeskix modelej Smysl-Tekst, Moskva: Nauka. (Reprinted in 1999, Moskva: Jazyki russkoj kul’tury). Mel’čuk, Igor. 1988. Dependency Syntax: Theory and Practice. Albany NY: State University of New York Press. Mel’čuk, Igor. 1995. Syntactic, or lexical, zero. In The Russian Language in the Meaning-Text Perspective, Igor Mel’čuk, 169–211. Wien: Wiener Slawistischer Almanach. Mel’čuk, Igor. 2006a. Zero sign in morphology. In Aspects of the Theory of Morphology, Igor Mel’čuk, 469–516. Berlin: Mouton de Gruyter. Mel’čuk, Igor. 2006b. Voice. In Aspects of the Theory of Morphology, Igor Mel’čuk, 181–262. ­Berlin: Mouton de Gruyter. Mel’čuk, Igor. 2009. Dependency in natural language. In Dependency in Linguistic Description [Studies in Language Companion Series 111] Alain Polguère & Igor Mel’čuk (eds), 1–110. Amsterdam: John Benjamins. Milićević, Jasmina. 2009. Schéma de régime: Le pont entre le lexique et la grammaire. Langages 176(4): 94–116. Special issue La représentation des structures prédicat-argument, Xavier Blanco & Pierre-André Buvet (eds). Radovanović, Milorad. 1990. O bezličnoj rečenici. In Spisi iz sintakse i semantike, Milorad Radovanović, 209–219. Novi Sad: Dobra vest. Siewierska, Anna. 1984. The Passive. A Comparative Linguistic Analysis. London: Croom Helm. Siewierska, Anna. 2007. Ways of impersonalizing: Pronominal vs. verbal strategies. In ­Languages and Cultures in Contrast and Comparison [Pragmatics & Beyond New Series 175], María de los Ángeles Gómez González, J. Lachlan Mackenzie & Elsa M. González Álvarez (eds), 27–61. Amsterdam: John Benjamins.

Interpretation and voice in Polish SIĘ and –NO/–TO constructions Małgorzata Krzek

Newcastle University, UK This paper discusses various properties of two Polish impersonal constructions with a particular focus on the function of the –NO/–TO suffix and the SIĘ particle. Following Fassi-Fehri (2009), it is assumed that the –NO/–TO suffix is best analysed as a head of VoiceP located between TP and vP. A similar analysis is not available for the SIĘ particle. It is also argued that Fassi Fehri’s (2009) generalization, according to which generic pros cannot be associated with the third person active inflection, only with the third person passive one, although true for Arabic, cannot be extended to account for the derivation and interpretation of impersonal constructions in Polish. Keywords:  Polish; impersonal constructions; voice; genericity

1.  Introduction This paper provides a general overview of morphosyntactic properties of P ­ olish impersonal SIĘ and –NO/–TO constructions. It demonstrates that whereas the –NO/–TO suffix is best analysed as the head of VoiceP (following Fassi Fehri 2009), the particle SIĘ cannot be given a similar analysis (contra Fassi Fehri 2009). What is more, Fassi Fehri’s generalization, according to which generic pros cannot be associated with the third person active inflection, only with the third person passive one, although true for Arabic, Irish, and Italian, cannot be extended to account for the phenomena found in Polish impersonal constructions. This paper is organized as follows: Section 2 presents two types of impersonal constructions found in Polish with a concise description of their morphosyntactic properties. Section 3 provides evidence that both constructions can be interpreted generically. Section 4 discusses passivization of two impersonal constructions. Section 5 reveals that the SIĘ particle and the –NO/–TO suffix cannot be analysed in a uniform way. Following Fassi-Fehri (2009), it is argued that the –NO/–TO suffix is base-generated as the head of a functional projection VoiceP. This projection is situated between TP and vP (Kratzer 1996; Collins 2005 and Sigurðsson 2011;

 Małgorzata Krzek

to mention just a few), and it is the locus of the impersonal active voice. A similar analysis is not available for the SIĘ particle, which can be analysed as either a functional head or an argument. The paper finishes with conclusions.

2.  Two types of impersonal constructions Polish has a large number of constructions that can be regarded as impersonal. This paper is, however, concerned with only two types of impersonal constructions, namely the SIĘ construction and the –NO/–TO construction. Both of these have a distinct morphological marking on the main verb. Despite the fact that morphological marking is different for the two constructions, they nevertheless share a number of morphosyntactic properties. 2.1  The SIĘ construction This structure is a cognate of one of the types of the so-called si/se construction in Italian and Spanish. It consists of a verb and a SIĘ particle. The verb is in the default 3rd person singular neuter form, as in (1): (1) a. Tańczy się tutaj często. dance:3sg.neut ref here often ‘[One] dance here often.’ b. Buduje się tutaj szkołę. build:3sg.neut ref here school:acc ‘[One] is/are building/builds/build/ a/the school here.’ c. Jadło się maliny ate:neut ref raspberries:acc ‘[One] was eating/ate raspberries.’ d. Będzie się dawać wszystkim prezenty. aux.fut ref dawać:inf everybody:dat gifts:acc ‘[One] will give everybody gifts.’

The sentences in (1) show that SIĘ impersonals can be used with intransitive (1a), monotransitive (1b-c), and ditransitive (1d) verbs with distinct forms for each tense (Kibort 2008). In (1a) and (1b) the impersonal verb refers to the present, in (1c) to the past, and in (1d) to the future. When an impersonal SIĘ refers both to the present and the past, the verb form is marked 3sg.neut. It is also clear that accusative objects are retained, as in (1b), (1c), and (1d). As noted by Kibort (2008), reflexive impersonals can also be formed from both unergative (2) and unaccusative predicates (3).



Interpretation and voice in Polish SIĘ and –NO/–TO constructions 

(2) Pracowało się ciężko. worked:neut ref hard ‘[One] worked hard.’ (3) W tym domu umiera się spokojnie. in this:loc house:loc die:3sg.neut ref peacefully ‘In this house [one] dies peacefully.’

2.2  The –NO/–TO construction The –NO/–TO construction uses otherwise an uninflected verb form with a –NO/–TO suffix as exemplified in (4) and (5): (4) Kupowano kwiaty. bought:impers flowers:acc ‘[People] were buying/bought flowers.’ (5) Tańczono do białego rana. danced:impers until white:gen morning:gen ‘[People] danced until dawn.’

The construction can be formed from both transitive (4) and intransitive (5) verbs. Just like in the case of reflexive impersonals, transitive verbs in the –NO/–TO construction also retain the accusative object. However, unlike reflexive impersonals, the –NO/–TO construction can only refer to the past tense. Another property of this construction, as noted by Kibort (2008), is that it can be formed from both unaccusative (7) and unergative predicates (6) including the iterative form of the verb ‘to be’ (8), as is illustrated by the following examples: (6) Pracowano ciężko. worked:impers hard ‘[People] worked hard.’ (7) Umierano tysiącami. died:iter.impers thousands:instr ‘[They] were dying/died in thousands.’ (8) Bywano często w tej restauracji. was:iter.impers often in this:loc restaurant:loc ‘[People/They] was/used to come to this restaurant often.’

2.3  Subjects in impersonal constructions The following subsections provide evidence that in Polish impersonals active syntactic subjects are projected. In order to substantiate the presence of subjects, a number of tests will be employed (e.g. availability of agent oriented ­adverbials,

 Małgorzata Krzek

availability of PRO in the embedded clause, availability of a reflexive or other ­anaphora, availability of the agentive phrase). 2.3.1  Availability of agent oriented adverbials The availability of agent oriented adverbials, such as celowo ‘on purpose’, dobrowolnie ‘voluntarily’, confirms the presence of an agent (or experiencer/undergoer).1 Agent oriented adverbials are allowed in both impersonal constructions, as the following examples demonstrate: (9) a. Budowano tutaj autostradę celowo. built:impers here motorway:acc on-purpose ‘[People/They] built a motorway here on purpose.’ b. Jadło się celowo dużo malin. ate:neut ref on-purpose a-lot-of raspberries:gen ‘[One/They/We] ate a lot of raspberries on purpose.’

2.3.2  Control and binding Participation in control and binding relations constitutes another argument supporting the presence of underlying subjects. Bondaruk & Charzyńska-Wójcik (2003) observe that the SIĘ and the –NO/–TO impersonals can share their subjects with embedded infinitive clauses (10), with present and past participle forms (11), and in subject-raising constructions (12). (10) a. Próbowano zrozumieć ten problem. tried:impers understand:inf this:acc problem:acc ‘[One/They] tried to understand this problem.’ b. Chciało się odkryć prawdę. wanted:neut ref discover:inf truth:acc ‘[One/They/We] wanted to discover the truth.’ (11) a. Przygotowując przedstawienie, zapomniano prepare:part.prs performance:acc forgot:impers o kostiumach. about costumes:instr

‘In preparing the performance, [one/they] forgot about costumes.’

.  As demonstrated in Section 2.1, the SIĘ and the –NO/–TO impersonals can be formed from unaccusative verbs. The subjects of these verbs, as noticed by Kibort (2004), are not agents but experiencers/undergoers, thus agent-oriented adverbials will not be felicitous in their environment. This fact, however, cannot be taken as an argument that the syntactic subject is not projected in these clauses.



Interpretation and voice in Polish SIĘ and –NO/–TO constructions 

b. Wychodząc z domu, zamykało się drzwi leave:part.prs from house:gen closed:neut ref door:acc na klucz. on key.

‘On leaving the house, [one/they/we] locked the door.’

(12) a. Zaczęto wyglądać bardzo poważnie. began:impers look:inf very serious:adv ‘[One/They] started to look very serious.’ b. Zdawano się tego nie dostrzegać. seemed:neut ref this:gen neg notice:inf ‘[They/One/We] seemed not to notice this.’  (Example (b) adapted from Kibort 2004: 272)

As regards binding, Kibort (2004) observes that the covert subject of the SIĘ and the –NO/–TO impersonals is also capable of binding reflexive and reflexive-­ possessive pronouns that need to be bound by the subject, as illustrated by ­examples in (13) and (14) respectively. (13) a. Oglądano siebie/się2 w lustrze. looked:impers self/ref in mirror:loc ‘[People/They] looked at oneself/themselves in the mirror.’  (adapted from Kibort 2004: 273) b. Kupowało się sobie prezenty. bought:neut ref self:dat gifts:acc ‘[One/They/We] bought oneself/themseves/us gifts.’ (14) a. Czytano swoje książki. read:impers.pst own:acc books:acc ‘[People/They] read one’s/their books.’ b. Słuchało się swojej muzyki. listened:neut ref own:gen music:gen ‘[One/They/We] listened to one’s/their/our music.’

2.3.3  Nominative subjects and oblique phrases Another piece of evidence supporting the claim that Polish impersonals do in fact have covert subjects comes from the observation that they do not accept the ­surface

.  In Polish the reflexive pronoun siebie ‘self:acc’ is, in very restricted contexts, interchangeable with a multifunctional enclitic form się (Nagórko 1998; Kibort 2004). Cf. about siebie and its variants in other Slavic languages see Franks, this volume.

 Małgorzata Krzek

expression of a nominative subject, as in (15). They also do not accept the expression of the agent in an oblique phrase, as demonstrated by ­sentences in (16):3 (15) a. *Maria czytano ciekawe książki.  Maria:nom read:impers.pst interesting:acc books:acc Intended: ‘Maria read interesting books.’ b. *Maria je się dużo malin.  Maria:nom eat:3sg ref a-lot-of raspberries:gen Intended: ‘Maria eats a lot of raspberries.’ (16) a. */?Kupowano samochód przez ojca.  bought:impers car:acc by father:acc Intended: ‘A car was bought by my/his/her father.’ b. */?Je się dużo malin przez Marysię.  eat:3sg.neut ref a-lot-of raspberries:gen by Marysia:acc Intended: ‘A lot of raspberries are eaten by Marysia.’

3.  Interpretation of the impersonal SIĘ and –NO/–TO constructions This section presents a very general overview of interpretation of the two impersonal constructions in question. This is to illustrate that in both constructions generic reading is available. It is, however, dependent on a number of factors such as type of a predicate, aspect and temporal/spatial frames. With unspecific time reference and imperfective aspect both constructions are interpreted generically regardless of the type of predicate. This is illustrated by sentences in (17) and (18).4 (17) a. Je się tutaj owoce.(Transitive) eat:3sg.neut ref here fruit:acc ‘[One] eats here fruit.’

.  The reason why the sentences in (16) may be acceptable is that the NPs in the prepositional phrases (PPs), although superficially similar to agent oblique phrases, do not have to be in fact interpreted as such (Bondaruk & Charzyńska-Wójcik 2003). They are not have to be instigators of the action denoted by the impersonal verbs; they can be interpreted as participant and may express either the cause of negative circumstances, or may be regarded as a kind of instrument. .  Due to the space limitation, sentences with unergative predicates have been omitted. For a more detailed analysis of the interpretation of impersonal constructions in Polish, see Krzek 2010.



Interpretation and voice in Polish SIĘ and –NO/–TO constructions 

b. Na tą chorobę umiera from this:acc disease:acc die:3sg.neut się powoli.(Unaccusative) ref slowly

‘From this disease [one] dies slowly.’

(18) a. Jedzono dużo owoców.(Transitive) was-eating:impers a-lot-of fruit:gen ‘[One] was eating a lot of fruit.’ b. Umierano powoli.(Unaccusative) died:iter.impers slowly ‘[One] was dying slowly.’

With the specific time reference and imperfective aspect, the –NO/–TO construction is still interpreted as generic (as illustrated by example in (19)), whereas the SIĘ construction clearly acquires new readings. It can be interpreted as referring to a group of people (arbitrary reading) that may or may not include the speaker (specific reading). The generic reading is not that clear, but may be, nevertheless, still available with transitive and unergative predicates. (19) W xix wieku jedzono dużo owoców.(Transitive) in xix century:loc ate:impers a-lot-of fruit:gen ‘[People] was eating/ate a lot of fruit in the 19th century.’ (20) a. Jadło się wczoraj dużo owoców.(Transitive) ate:neut ref yesterday a-lot-of fruit:gen ‘[One/They/We/I] was eating/ate a lot of fruit yesterday.’ b. Pracowało się wczoraj ciężko.(Unergative) worked:3sg.neut ref here hard ‘[One/They/We/I] was working hard yesterday.’

Examples in (17)–(20) demonstrate that in certain contexts the subjects of both constructions can be interpreted as generic. 4.  Voice Phrase and generic subjects In this section, I attempt to see whether the analysis of impersonal constructions as expounded by Fassi Fehri (2009) can be extended to account for both the interpretation of the two constructions in question and the function of the SIĘ particle and the –NO/–TO suffix, and as a consequence whether Holmberg’s generalization (2005) can be maintained. It will be argued that the SIĘ particle and the –NO/–TO suffix cannot be given a uniform analysis, as the two elements

 Małgorzata Krzek

perform different functions within a clause. The –NO/–TO suffix, but not the SIĘ particle, is best analysed as the head of the functional projection (VoiceP) located between vP and TP. Crucially, it will be demonstrated that the –NO/–TO suffix heads the active impersonal Voice Phrase, not a passive impersonal Voice Phrase, as argued by Fassi Fehri (2009) for Arabic, Italian and Irish. What follows is that third person generic pro is also available in active sentences in consistent null subject languages (CNLS) of which Polish is an example, not only in passive ones as maintained by Fassi Fehri (2009). This observation, in turn, provides counterevidence to H ­ olmberg’s generalization according to which, in consistent null subject languages, null subjects are interpreted exclusively as referential, whereas a third person pro is exclusively generic in partial null subject languages (e.g. Finnish). 4.1  Theoretical assumptions Within the generative framework, Passive Voice has been extensively studied (Jaeggli 1986; Baker at al. 1989; Collins 2005, etc.). Other types of Voice and the category of Voice have not been as widely investigated. There is, however, a growing consensus that Voice is a functional head located between v and T, either morphologically marked or unmarked (e.g. Kratzer 1996; Sigurðsson 2011; Fassi Fehri 2009). According to Sigurðsson 2011, any predicate is embedded under some Voice head, and a Voice head commonly alters the argument structure of basic (i.e. lexical, non-derived) predicates. Voice is also a cover term, much as Aspect in Cinque (1999), that is there are number of mutually exclusive Voice-type heads, including passive and active Voice, VoicePASS and VoiceACT (Sigurðsson 2011). This is, in essence, the view that I adopt here, p ­ ointing out, however, that in the case of Polish impersonal –NO/–TO construction, Voice head does not affect the argument structure of a predicate, that is, arguments are not added, incorporated/suppressed, expletivized or demoted. This fact is based on the observation that in this construction verbs behave in the same way as in sentences in active personal Voice, that is they assign accusative case to their objects, oblique phrases are not allowed, and agent oriented adverbials are possible with verbs that assign agent theta roles to their external arguments. Most importantly, however, the presence of the –NO/–TO suffix excludes passivization. The reason being that if it is assumed that there can only be one Voice head per phrase, then the fact that passive auxiliaries cannot be merged ­suggests that the position reserved for this type of auxiliaries is filled in by some other element. It is argued that this element is the –NO/–TO suffix (see ­Section 4.3). The same analysis, however, is not available for the particle SIĘ, as in this case ­passivization although restricted is nevertheless possible with the SIĘ ­construction (see Section 4.2).



Interpretation and voice in Polish SIĘ and –NO/–TO constructions 

4.2  Passivization in the SIĘ construction The SIĘ impersonals can be formed from a passivized predicates, as shown by examples in (21). (21) a. Było się lubianym / oskarżonym. aux.pst ref liked:part / accused:part ‘[One] was liked/accused.’ b. Zostało się pobitym.5 aux.pst ref beaten:part.pf ‘[One] was beaten.’ c. Zostawało się powoli wdrażanym w aux:pst.iter ref slowly introduced:part in bieżące sprawy. current:acc affairs:acc

‘[One] was slowly introduced in the current state of affairs.’

It is worth noting, however, that the process of passivization of the SIĘ construction is not entirely free and there are some constraints imposed on its derivation. It appears that although one of the characteristic features of passives is that object is promoted to subject position, and at the same time subject is demoted, the object of the SIĘ impersonal construction cannot be not promoted to the subject position, and the subject cannot be demoted. This is illustrated by the examples in (22). This seems to suggest that passivized impersonal SIĘ constructions might not be derived from its active counterpart. (22) a. Nagradzało /doceniało się ciebie. awarded:neut /appreciated:neut ref you:acc ‘[One/They/We/I] awarded/appreciated you.’ b. *Ciebie /ty było się nagradzanym / docenianym.  you:acc /you:nom aux:pst ref awarded:part / appreciated:part ‘You were awarded.’ c. Było się nagradzanym / docenianym / zmartwionym. aux:pst ref awarded:part / appreciated:part / worried:part ‘[One/They/We/I] was/were awarded/appreciated/worried.’

.  The examples in (21b) and (21c) are deemed somewhat controversial. They were judged as acceptable by 53.70% of the native speakers interviewed, and this is the reason why they were included. However, because the discussion does not focus on types of passive auxiliaries available with the SIĘ impersonal construction, they do not have a bearing on the final conclusion.

 Małgorzata Krzek

4.3  Passivization in –NO/–TO construction Contrary to the SIĘ impersonals, the –NO/–TO impersonals cannot be formed from passivized predicates. This is illustrated by examples in (23). (23) a. (*Zostało) siano pszenicę na wiosnę.   aux.pst sowed:impers wheat:acc in spring ‘Wheat will be sowed in spring.’  (adapted from Dziwirek 1994: 186–187) b. (*Było) znaleziono niemowlę w koszu (*przez lekarzy).   aux.pst found:impers baby:acc in basket:loc (*by doctors) ‘A baby was found in a basket by doctors.’ (Lavine 2005: 8)

As sentences in (23) illustrate, the passivization of the –NO/–TO construction yields ungrammatical results, regardless of the type of an auxiliary (both are excluded) and a past participle (both perfective and imperfective are excluded) used. Both passive auxiliaries and oblique phrases are available in Polish canonical passive so the fact that they are impossible in the –NO/–TO construction is not a fact about passive in Polish in general but rather specific to the –NO/–TO construction. The fact that passivization is excluded additionally confirmed when ­Polish –NO/–TO construction is compared with the cognate construction found in ­Ukrainian, examples of which are given in (24). (24) Ukrainian

(from Lavine 2005: 9)

a. Nemovlja bulo znajdeno u košyku likarjami. baby:acc aux.pst found:impers in basket:loc doctors:instr ‘A baby was found in a basket by doctors.’ b. Inozjemcja bude posadženo do v’jaznyci foreigner:acc aux.fut placed:impers to prison:gen hlavoju urjadu. head:instr government:gen

‘A foreigner will be put in prison by the head of government.’

As illustrated by examples in (24), Ukrainian –NO/–TO construction allows overt tense marking auxiliaries and by-phrases, both of which are excluded in Polish –NO/–TO construction. These data confirm that it is the –NO/–TO ­suffix in Polish that is responsible for the unavailability of passive in the –NO/–TO construction;6 this gives rise to the analysis of the –NO/–TO suffix as the head of VoiceP.

.  Kibort (2008) claims that they can be formed from the iterative form of the verb ‘to be’ as illustrated by example in (i). The following example contains an impersonal iterative form of the verb bywano ‘to be’ with an imperfective passive participle bitymi ‘beaten’:



Interpretation and voice in Polish SIĘ and –NO/–TO constructions 

5.  Generic subjects and VoiceP We are now in a position to discuss the function of the –NO/–TO suffix and the SIĘ particle and to demonstrate how the analysis suggested for the a­ forementioned elements relates the hypothesis put forward by Fassi Fehri (2009), according to which in consistent null subject languages generic subjects are only available in Passive Voice. Having assumed that different types of Voice heads are mutually exclusive, and that there can only be one Voice head in a clause, the fact that the –NO/–TO ­suffix blocks passivization suggests that it is merged in the head of VoiceP. However, the –NO/–TO suffix is not a passive morpheme and does not i­ntroduce a passive voice head as confirmed by the observations that both oblique phrases and overt e­ xpressions of nominative subject are excluded. In other words, the argument s­tructure of a predicate in the –NO/–TO c­ onstruction remains intact. Thus, the construction is still active. It is assumed here what

?*Bywano bitymi. (i)   was:impers.iter beaten:part.pl ‘[One] was beaten.’

(adapted from Kibort 2008: 265)

Kibort (2008) seems to overstate the ability of the –NO/–TO construction to passivize as all of my informants considered this sentence ungrammatical. However, the acceptability of sentences containing the impersonal iterative form of the verb ‘to be’ rises somewhat when they are coupled with perfective passive participles. This is illustrated by the sentence in (ii). (ii) ?Bywano oskarżonymi.  was:impers.iter accused:part.pl ‘[One] was accused.’ The examples containing the iterative form of the verb ‘to be’ with the –NO/–TO suffix become more acceptable when followed by an adjectival/nominal/PP complement, as illustrated by the following example. (iii) Bywano szczęśliwymi/ w tym domu rzadko. was:impers.iter happy:PL/ in this:loc house:loc rarely ‘[People/They] was happy/in this house rarely.’

The reason behind this upsurge in acceptability seems to be that what the –NO/–TO suffix attaches in the sentences above is not a passive auxiliary but a copula. As such the copula bywano is incompatible with imperfective passive participle, as in (i). Because, as discussed in Footnote 6, the constructions such as the one in (ii) are subject to dual interpretation as examples of either stative passive or a copula plus deverbal adjective, they are higher on the acceptability scale.

 Małgorzata Krzek

the –NO/–TO s­uffix does is to introduce an impersonal voice. This analysis ­coupled with the fact that generically interpreted subjects are readily available in this c­ onstruction shows that generic subjects are available not only with third person passive inflection as argued by Fassi Fehri (2009) but also when the inflection is non-agreeing. Therefore, it appears that Fassi Fehri’s generalization cannot be extended to account for the derivation and interpretation of the –NO/–TO construction in Polish. As for the SIĘ construction, it is argued that the SIĘ particle cannot reside in the head of VoiceP, as it does not block the merge of a passive auxiliary. In other words, the SIĘ construction, just like the –NO/–TO construction, is active as both oblique phrases and overt expressions of nominative subject are excluded, but contrary to the -NO/–TO suffix, the SIĘ article does not block passivization. As the generic interpretation is readily available, it is evident, contra Fassi Fehri (2009), that generic subjects can be associated with third person active inflection, not only with the passive one. As regards the possible functions of the SIĘ particle, it may be that of either a functional projection or a nominal expression (Saloni 1975). According to Saloni (1975), the particle SIĘ, is a syntactic subject to which nominative case is assigned. Saloni (1975) does not state explicitly whether he considers SIĘ an argument or, more likely, an expletive that is first merged in the Spec TP. The latter analysis of SIĘ appears to be more promising especially if one takes into consideration the fact that SIĘ is devoid of any phi-features (see Kibort 2004 for details). However, much more work needs to be done before any such conclusion is made.

6.  Conclusion This paper provides a general overview of morphosyntactic properties of two Polish impersonal constructions: the –NO/–TO construction and the SIĘ ­ ­construction. It has been demonstrated that the –NO/–TO suffix should be best analysed as the head of VoiceP. A similar analysis is not available for the SIĘ particle. It has also been shown that generic pros are found in active Voice in consistent null subject languages, not only in passive Voice, as argued by Fassi Fehri (2009). Consequently, it appears that also Holmberg’s (2005) generalization according to which, generic subjects are not found in consistent null subject languages will need to be revisited.



Interpretation and voice in Polish SIĘ and –NO/–TO constructions 

Abbreviations acc adv fut aux gen impers inf inst ipf iter

= = = = = = = = = =

accusative adverb future auxiliary genitive impersonal infinitive instrumental imperfective iterative

nom loc part pf prs pl pst ref sg

= = = = = = = = =

nominative locative participial perfective present plural past reflexive singular

References Baker, Mark, Johnson, Kyle & Roberts, Ian. 1989. Passive arguments raised. Linguistic Inquiry 20: 219–251. Bondaruk, Anna & Charzyńska-Wójcik, Magdalena. 2003. Expletive pro in impersonals passives in Irish, Polish and Old English. Linguistische Berichte 195: 325–362. Cinque, Guglielmo. 1999. Adverbs and Functional Heads. A Cross-linguistic Perspective. Oxford: OUP. Collins, Chris. 2005. A smuggling approach to the passive in English. Syntax 8(2): 81–120. Dziwirek, Katarzyna. 1994. Polish Subjects. New York NY: Garland. Fassi Fehri, Abdelkader. 2009. Arabic silent pronouns, person and voice. Brill’s Annual of Afroasiatic, Languages and Linguistics 1: 1–38. Holmberg, Anders. 2005. Is there a little pro? Evidence from Finish. Linguistic Inquiry 36(4): 533–564. Jaeggli, Osvaldo A. 1986. Arbitrary plural pronominals. Natural Language and Linguistic Theory 4(1): 43–76. Kibort, Anna. 2004. Passive and Passive-like Constructions in English and Polish, Ms, U ­ niversity of Cambridge. Kibort, Anna. 2008. Impersonals in Polish: An LFG perspective. Transactions of the Philological Society 106(2): 246–289. Kratzer, Angelina. 1996. Severing the external argument from its verb. In Phrase Structure and the Lexicon, Johan Rooryck & Laurie Zaring (eds), 109–138. Dordrecht: Kluwer. Krzek, Malgorzata. 2010. Some aspects of subjects of impersonal constructions in Polish. ­Newcastle Working Papers in Linguistics 16: 66–87. Lavine, James. E. 2005. The morphosyntax of Polish and Ukrainian -no/-to. Journal of Slavic Linguistics 13(1): 75–117. Nagórko, Alicja. 1998. Zarys Gramatyki Polskiej, Warszawa: PWN. Saloni, Zygmunt. 1975. W Sprawie się. Język Polski 55: 25–34. Sigurðsson, Halldór Á. 2011. On the new passive. Syntax 14: 148–178.

Dative-infinitive constructions in Russian Taxonomy and semantics Alina Israeli

American University, Washington, USA The article provides a taxonomy of infinitive constructions both without any particles and with the particles ne and li. It takes into account verbal aspect and delineates the semantic limits of each group. The imperfective types Dat + INFIPF and Dat + nе + INFIPF constitute predictions based upon an omen or conjecture in the first case, and a condition p in the second case. Perfective Dat + nе + INFPF conveys the Speaker’s opinion that the Subject is physically unable to perform V. All the constructions with li and nе + Dat + INFIPF express the Speaker’s opinion that the Subject, of all people, should or should not, can or cannot, perform V. Questions, their limitations and the effect of aspect are discussed separately. Keywords:  dative; infinitive; modality; quasi-synonymy; verbal aspect; predictions

0.  Introduction The need for a taxonomy of infinitive constructions is predicated upon a number of considerations: Studies that discuss infinitive constructions (Ebeling 1984; Zolotova 1990; Fortuin 2005, 2007) discuss them all together, without making a clear-cut distinction between types of infinitive constructions. They include constructions with nominative subject as well as with dative of the underlying subject (Ebeling 1984; Zolotova 1990), infinitive constructions in subordinate clauses as well as in main clauses, infinitives as grammatical subject of a sentence (Zolotova 1990), ­infinitives in imperatives as well as in exclamations, such as (1)  Странные люди эти Крессе! Сказать при мальчишке такую вещь! (Булгаков)(Ebeling 1984: 99) ‘These Kresse are strange people! How could they say such a thing in front of a boy!’

 Alina Israeli

Aspectual distinctions are typically not part of the discussion of infinitive constructions. As a result we do not have any classifications of the infinitive constructions, and we do not possess a full list of them, their descriptions or their limitations. Veyrenc (1973) discusses two types of dative-infinitive constructions, ­emphasizing the fact that the first type represents the modality of DEVOIR or obligation, while the second type represents the modality of POUVOIR or possibility or rather non-possibility: (2) [*]Ему вернуться домой.   ‘Il faut qu’il rentre chez lui, il doit rentrer chez lui’ ‘He must return home’

(Veyrenc 1973: 351)

(3) Ему не вернуться домой. ‘Il ne lui sera pas donné de rentrer chez lui, il ne pourra rentrer chez lui, il n’est pas question qu’il rentre chez lui’ (Veyrenc 1973: 352) ‘He cannot return home’

Veyrenc (1979: 20) laments that grammars continue to put forth the impressionistic idea of “infinitive of fatality”, “infinitive of impossibility”, “infinitive of necessity” as well as “infinitive of deliberation” and “optative infinitive”. He provides a syntactic classification and a semantic classification which consists of four types: a. Negative with nuclear center: (4) Ему не пройти. ‘Il n’est pas question qu’il puisse passer.’ ‘It’s impossible for him to pass.’ (5)  А Лизе за Паншиным не быть, не беспокойся; не такого мужа она стоит. (Тургенев. Дворянское гнездо) ‘And Liza will not marry Panshin, don’t worry; she is worthy of a ­better ­husband.’ (6) Что делать? Не пороть же его, каналью! (Чехов. Единственное средство) ‘What is to be done? One cannot whip the scoundrel!’

b. Negative with actantial center: (7) [?] Не ему же это понять.   ‘Ce n’est pas lui qui peut comprendre cela.’ ‘It is not he who could understand this.’



Dative-infinitive constructions in Russian 

c. Positive with nuclear center: (8) Ему ходить. ‘C’est à lui de jouer.’ ‘It is his turn (in a game).’ (9) Всё одно погибать! (Чехов. Сущая правда) ‘It’s all the same, I will die.’

d. Positive with actantial center: (10) Ему же было ходить, а ты пошёл. ‘C’était à lui de jouer, et c’est toi qui a joué.’ ‘It was his turn, yet you went.’

Examples (4)–(10) and their French translations are from Veyrenc (1979: 53). This is a rather inconsistent classification, which does not reveal differences due to aspect, placement of the negative particle ne and other factors. Also, only one construction among those on Veyrenc’s list has an obligatory particle že, the one in example (6); in all other cases the particle že is optional, and this is important. Example (10) is the past tense of example (8), so it should not represent a separate group, in which case it is not clear what would remain in group (d). In my classification, Veyrenc’s examples from group (a) fall into different groups: example (4) belongs to group 4, example (5) belongs to group 3, and ­example (6) with the obligatory particle že is not part of this study. Not only are Veyrenc’s (2) and (7) incorrect, but a number of recent studies give either incorrect infinitive sentences or incorrect translations, which points to the fact that there are a lot of misunderstandings and lack of description of scope and limitations for each type. (11) [*] Мне не уходитьipf.   *‘It’s not for me to leave.’

(Moore & Perlmutter 1999: 226)

(12) Нам не миновать гибели. *‘It’s not for us to avoid disaster.’

(Timberlake 2004: 8)

(13) Мне не сдать экзамен. *‘It is not (in the cards) for me to pass the exam.’ (14) Тебе налить? ‘Shall I pour you [some]?’ *‘Will you pour?’

(Fleisher 2006: 4)

(Maurice 1996: 91 in Fortuin 2000: 264)

Altogether there are four formative elements of these constructions: (1)  “dative subject” (Dat), (2) infinitive (INF), (3) the negative particle ne, and (4) the p ­ articles

 Alina Israeli

li, že and by. This study will focus on two of the particles: ne and li. The various combinations with these two particles theoretically create ten possible constructions (see Table 1): Table 1.  Dative-infinitive constructions in positive and negative sentenses Dative-infinitive constructions

Realisation

1.

Dat + INFIPF

ему петь

2.

Dat + INFPF

(что) ему спеть

3.

Dat + nе + INFIPF

ему не петь

4.

Dat + nе + INFPF

ему не спеть

5.

nе + Dat + INFIPF

не ему петь

6.

nе + Dat + INFPF

*не ему спеть

7.

Dat + li+ INFIPF

ему ли петь

8.

Dat + li + INFPF

?ему ли спеть

9.

Dat + li+ nе + INFIPF

ему ли не петь

10.

Dat + li + nе + INFPF

ему ли не спеть

The specific reasons for the parentheses, the asterisk and the question mark above will be discussed in the corresponding sections below. In the next section I will examine each type separately. In the semantic descriptions “I” means the Speaker and “S” means the dative Subject. 1.  Dative-infinitive constructions 1.1  Model Dat + INFIPF The first construction Dat + INFIPF has a number of subtypes. The main distinction between them is the sentencial stress: INFIPF +Dat vs. Dat + INFIPF. However, the first model must be divided in its turn in several subtypes, as follow: 1.1.1. INFIPF +Dat

1.1.1.1. Dat+INFIPF 1.1.1.2. INFIPF +Dat+NP 1.1.2. Dat + INFIPF 1.1.3. Dat + V(NṔ +) INFIPF

1.1.1  Model INFIPF +Dat The construction INFIPF +Dat means ‘I predict there shall be S’. The verb used in this group is exclusively byt’ ‘to be’.



Dative-infinitive constructions in Russian 

This subtype represents the speaker’s prediction, sometimes based on an omen or some observed sign. The prediction pertains to the future manifestation of some negative or positive event, more often bad than good: (1.1)  Да, мы – мистики; крестьянин тоже мистик, когда у него – «свербит в пояснице» и он утверждает: быть грозе. [Андрей Белый. На рубеже двух столетий (1929)] ‘Yes, we are mystics; a peasant is also a mystic, when he has pain in the small of the back and claims: there will be a thunderstorm.’ (1.2) Катерина. Тиша, на кого ты меня оставляешь! Быть беде без тебя! Быть беде! [А. Н. Островский. Гроза (1860)] ‘Katerina. Tisha (Tixon), in whose charge are you leaving me! There shall be trouble without you!’ (1.3)  Если в город вошел еще один военный – быть войне! – решил доктор Струве, углядевший в окно проезжающего мимо полковника. – Надо готовить полевой госпиталь! [Дмитрий Липскеров. Сорок лет Чанчжоэ (1996)] ‘ “If one more military man enters the town, there shall be war!” decided Doctor Struve, after noticing in the window a passing colonel. “It’s time to prepare the field hospital!” ’ (1.4)  Даже скорей была она язычница, брали в ней верх суеверия: что на Ивана Постного в огород зайти нельзя – на будущий год урожая не будет; что если метель крутит – значит, кто-то где-то удавился, а дверью ногу прищемишь – быть гостю. [Александр Солженицын. Матренин двор (1960)] ‘If anything, she was a pagan, and, above all, superstitious: if you went into the vegetable garden on the day of the beheading of St. John, that meant there would be a bad harvest next year; if a storm was whirling the ­snowflakes round and round, it meant that someone had hanged himself; if you caught your foot in the door, it meant there would be a visitor.’

A guest or visitor (1.4), in the same sequence of predictions as a bad harvest and ­suicide by hanging, does not constitute a positive event. Positive predictions seem to be possible only when made by the person who will deliver the positive outcome: (1.5)  Осматривая Заячий остров, Петр I вырезал тесаком два куска дерна, положил их крестом и произнес: «Здесь быть городу!» (Элла Кричевская. Литературная жизнь старого Петербурга) ‘Examining Rabbit Island, Peter the Great cut out with his Greek sword two pieces of turf, put them criss-cross and announced: “Here there shall be a city!”

 Alina Israeli

1.1.1.1  Model Dat+INFIPF .  Dat + INFIPF means that ‘after some doubts or ­deliberations the decision has been made that there will be S’. The verbs in this subgroup are also usually byt’ ‘to be’ or occasionally another existential ­substitute. We can see signs of discussion and deliberation in the following examples. We can also see that the existence of the Subject in question was previously in doubt or prohibited. This is the case of the new magazine, the possibility of a heart transplant, or the Second Vatican Council. Unlike events in 1.1.1, these events are viewed very positively by the Speaker. (1.6)  Я абсолютно согласен с вашей точкой зрения и почти уверен, что инициатива будет поддержана. Думаем, что мы можем уже поздравить друг друга. Журналу – быть! Он встал и пожал руки писателям… [Василий Аксенов. Таинственная страсть (2007)] ‘I totally agree with your point of view and am almost sure that the initiative will be supported. We think that we can already congratulate each other. The magazine shall exist! He got up and shook the writers’ hands…’ (1.7)  Наконец в 1986 году после долгих переговоров, совещаний, увещеваний в верхах было решено “ операции быть ”, и я пересадил пациенту донорское сердце. [Анна Маева. Дар на всю жизнь (1999) // «Здоровье», 1999.03.15] ‘Finally in 1986 after long negotiations, meetings, and admonishings it was decided at the highest level that “the operation will take place”, and I ­transplanted the donor’s heart to a patient.’ (1.8)  25 декабря 1961 года Папа огласил апостольскую конституцию – «­Humanae salutis». Это значило, что Собору – быть. 2 февраля 1962-го объявлено о дате его созыва. [Александр Архангельский. 1962. Послание к Тимофею (2006)] ‘On December 25, 1961, the Pope proclaimed the apostolic constitution “Humanae salutis”. This meant that the Vatican Council would be formed. On February 2, 1962, the date of its convocations was announced.’

The very well known example from Mayakovsky (1.9) seems to be an exception insofar as the use of a verb other than byt’. Considering that it is a poem that requires rhyme and meter, this example should not be considered as standard for the infinitive constructions: (1.9)  Я знаю – город будет,/я знаю – саду цвесть,/когда такие люди/в стране в советской есть! (В. Маяковский. Рассказ Хренова о Кузнецкстрое и о людях Кузнецка. (1929) ‘I know there shall be a city,/I know a garden shall bloom,/when in the Soviet country/there are such people!’

No other example with a verb other than byt’ with the meaning of prediction has been attested.



Dative-infinitive constructions in Russian 

1.1.1.2  Model INFIPF +Dat+NP.  INFIPF +Dat+NP means ‘I predict that S will be NP’ or ‘I predict that S will V (NP)’. Аll of the above examples represent the existential use of byt’. However, byt’ can also be a copula of a predicate that describes future state or condition of the Subject. These are Zolotova’s (1990: 345) examples (1.10‒1.11) as well as examples from RNC (1.12–1.13): (1.10) Быть тебе без бороды! (Пушкин) ‘[I predict that] you will lose your beard!’ (1.11) [Я вот всегда говорила:] нашей графине быть за генералом (Герцен) ‘[I always said:] our countess will marry a general.’ (1.12) Я ему еще на первом курсе говорил: «Быть тебе, Угрюмов, адмиралом!» [Вячеслав Морозов. Адмирал ФСБ (2004) // «Наш современник», 2004.03.15] ‘I said to him back when he was a first year student, “Ugryumov, you will be an admiral!”’

What is interesting here, among other things, is that the predictions are typically addressed in the singular; the plural form of address means that there is more than one addressee: (1.13) – Видим ваши хитрости! Только нам это нипочем. Хоть всех самолучших бойцов с завода поставьте, а быть вам битыми! [П. П. Бажов. Широкое плечо (1948)] ‘We see your clever moves! Only it’s nothing to us. Even if you put in all of your very best workers from the plant, you will still be beaten!’

Not only the verb byt’ could manifest itself in predictions, even though it may be more common. We do find other verbs in this construction: (1.14) Жить тебе, Светка, в болоте, станешь ты кикиморой, выйдешь замуж за лешего… [Юлия Буцик. «Не убий прохожего ботинком!» (2004) // «Уральский автомобиль» (Миасс), 2004.01.15] ‘You will live, Sveta, in a swamp, you will become a hobgoblin and you will marry a wood-spirit…’ (1.15) Вариться тебе в котле! (Руслан Тарба. История о непокорном бычке…) ‘You will boil in a cauldron!’ (1.16) Маяться тебе еще весь день. Не надейся так легко отделаться. (http:// www.forum-volgograd.ru/archive/topic/10406–2.html) ‘You will still suffer for the rest of the day. Don’t expect to get out of it so easily.’

Not all the predictions are of a negative variety, although that is more common. We do find positive predictions as well:

 Alina Israeli

(1.17) Долго жить тебе, Михеевна! – сказал ей Озеров. [Михаил Бубеннов. Белая береза/части 1–2 (1942–1952)] ‘You will live a long life, Mixeevna! Ozerov told her.’

1.1.2  Model Dát + INFIPF Dat + INFIPF means ‘it is the Subject that has to do V, not me, and no one else, so no one should meddle’. In all of the following examples (1.18–1.20) the pronouns are stressed: (1.18) Не такую жену хотела она своему Саше [сыну]. Ну да ему жить… ́ Ирина поплакала и устремила все свои чаянья на Снежану [дочь]. (Виктория Токарева. Своя правда) ‘She wanted a different wife for her [son] Sasha. Oh well, it is for him to live… [i.e. it’s his life] Irina cried for a while and diverted all of her hopes onto [her daughter] Snezhana.’ (1.19) Но самому тебе не справиться. Тебе должен помочь мудрый и опытный человек. Это твое дело, кого пригласить на помощь. Тебé решать. [Вальтер Запашный. Риск. Борьба. Любовь (1998–2004)] ‘You won’t manage alone. You need help from a wise and experienced person. It is your business whom you should ask for help. It is for you to decide.’ (1.20) – Тебе учиться, не мне, я свое отучилась. (Грани разума. http://litprom. ru/thread10983.html) ‘It’s for you to study, not me, I am done studying.’

In example (1.18), the mother thinks, “it’s her son’s life, and even though she does not like her daughter-in-law it is not her problem”. Similarly in (1.19), “it’s for you to decide or it’s your job”. In (1.20) the speaker (the mother) says specifically: “you have to study, not me”. 1.1.3  Model Dat + (NP ́ +) INFIPF Dat + (NṔ +) INFIPF means an impending action, difficult or unpleasant one. This construction does not have an emphatic stress unless there is an NP p ­ receding the verb. There is usually an adverbial designating either the time of the event or ­timespan between the moment of speech and the time of the event. (1.21) Эх, жалко, что мне завтра рано вставать! [Ю. О. Домбровский. Факультет ненужных вещей, часть 1 (1978)] ‘Oh, too bad, I have to get up early tomorrow!’ (1.22) Я хочу, чтобы эта книга была у тебя. Скоро тебе сдавать экзамен. [Андрей Геласимов. Ты можешь (2001)] ‘I want you to have this book. You have to take the examination soon.’



Dative-infinitive constructions in Russian 

(1.23) Пришло, видимо, время ему выговориться, и скоро ему умирать. [Владимир Березин. Свидетель // «Знамя», 1998] ‘Apparently the time has come for him to unload his soul, and soon he will have to die.’

The absence of a time adverbial is possible when both interlocutors are aware of the proximity of the event in time, as in (1.24) where the interlocutor has to get a good night sleep before the road, and in (1.25) where there is reason for asking about the weather is the proximity of the business trip: (1.24) [Лиза, Алена Фалалеева, жен, 22, 1985] Ты бы тоже выспался! Тебе машúну вести. Ехать долго/а дорога/ты говорил/плохая. [Алексей Балабанов. Груз 200, к/ф (2007)] ‘ “You also should get enough sleep! You have to drive. It’s a long way, and the road, you said, is bad.” ’ (1.25) А меня интересует, как там в Питере – снега и грязи ещё много? А то мне в командирóвку ехать. (http://lilitochka.ru/viewtopic. php?id=1889&p=38) ‘And I am interested, how is it in St. Petersburg, is there still a lot of snow and dirt? For I have to go on a business trip.’

The absence of the time adverbial could also signal the immediacy of the action: (1.26) – Нам выходить, – вполголоса сказал он и едва заметно подтолкнул ее к выходу из вагона. (А. Маринина. Шестерки умирают первыми) (Fortuin 2005: 39) ‘ “We have to get off,” he said sottovoce and barely noticeably pushed her towards the exit from the train.’

In (1.24) the stress is on mašinu, in (1.25) on komandirovku, and in (1.27) on èkzamen: (1.27) – У меня возникла идея, – говорил Парфенов. Ну хорошо, идея. А мне завтра экзáмен сдавать. [Елена Ханга. Про все (2000)] ‘ “I have an idea,” Parfenov was saying. Okay, an idea, but I have to take an exam tomorrow.’

1.2  Model Dat + INFPF The type Dat + INFPF does not seem to occur without a question word or without a question intonation. It will be discussed with questions below (see § 2.6). 1.3  Model Dat + ne + INFIPF The construction– Dat + nе + INFIPF or nе + INFIPF + Dat–has the element of a fatalistic, usually negative prediction. Its meaning is: ‘action V will never again

 Alina Israeli

(or no longer) take place’. The fact that the meaning is never again or no longer is important for understanding this construction. The negative prediction in this construction is connected to some outside event or condition p which precludes V from ever taking place again. (3.1) – А уезжаю я из деревни, – насмешливо сказал он. – В городе Томске теперь буду жить, у сына Володьки. Так что не пить нам больше водку девятого мая… [Виль Липатов. Деревенский детектив/ Кто уезжает, а кто остается… (1967–1968)] ‘ “And I am leaving the village,” he said mockingly. “From now on I am going to live in the city of Tomsk with my son Volodya. So we won’t be able to drink vodka on May 9th ever again…” ’

Because the speaker is moving (p), he will no longer be able to drink vodka with his interlocutor. In (3.2), Nonna’s lover, who was renting an apartment for her, just died. So because of his death (p), she will never again live in this apartment: (3.2)  – Квартира эта, значит, не ваша? – уточнил Стас, рассматривая ключи.  – Нет, не моя. – Нонна поправила глаза, подрисовала губы, тяжело вздохнула и только сейчас подумала о том, что больше ей здесь не жить. (Н. Леонов, А. Макеев. Эхо дефолта) ‘ “So this apartment is not yours?” Stas wanted to get it straight, examining the keys. “No, not mine.” Nonna fixed her eye make-up, put on lipstick and sighed ­heavily, and only now thought about the fact that she will never again live here.’

Similarly in (3.3): if she stays with him (p), he will no longer work with Gurov as per Mariana’s condition: (3.3)  – Но ты сама посуди – какая нас с тобой ждет жизнь? Если я останусь с тобой, у Гурова мне уже не работать – это условие Марьяны.  – Ну и что? Обойдемся без твоего Гурова! – пробормотала она, сама не понимая, что такое говорит. [Татьяна Тронина. Русалка для интимных встреч (2004)] ‘ “ But just think, what kind of life awaits us? If I stay with you, I will never again be able to work with Gurov, such is Mariana’s condition.” “So what, we’ll do fine without your Gurov!” she muttered, herself not fully understanding what she was saying.’

The condition p could in fact be the actions of the speaker or some other ­participant associated with the speaker. In other words, the statement that the Subject will no longer be able to perform V is due to the Speaker’s actions. Thus



Dative-infinitive constructions in Russian 

the Speaker’s actions against the Subject are euphemistically veiled by a statement that the Subject will no longer perform V: (3.4)  Эти кадровые перемены пройдут уже под патронажем Центризбиркома. «Мы будем избавляться от тех, кто не умеет работать по новым технологиям или скомпрометировал себя политической ангажированностью», – заметил Вешняков. …«Видимо, через год многим уже не сидеть в этом зале», – заметил один из руководителей регионального избиркома. [Александр Садчиков. Многим тут не сидеть…. К выборам Центризбирком начинает чистку своих рядов (2002) // «Известия», 2002.05.31] ‘These personnel changes will take place already under the auspices of the Central Election Committee. “We will get rid of those who do not know how to work using new technologies or who have compromised ­themselves by political involvement,” noted Veshnyakov [president of CEC]. … “­Apparently in a year many people will no longer be sitting in this hall,” noted one of the leaders of the regional election committee.’

This construction can be used as a threat; in (3.5) the condition p is lying: (3.5) – Ей не жить, если будет врать! (Распутин. Дочь Ивана, мать Ивана) ‘She will not live if she lies!’

P could also be the unspoken current state of affairs. In the case of (3.6), it is the antisemitism in the Soviet Union: unless it changes, Jews will not be able to ­perform actions V: (3.6) Не ходить вам в камергерах, евреи! … Не сидеть вам ни в Синоде, ни в Сенате. (А. Галич. Предостережение) ‘Jews, you will not be chamberlains! You will not sit either in the Synod or in the Senate.’

So what is wrong with (11)? (11) [*]Мне не уходитьipf.   *It’s not for me to leave. (Moore & Perlmutter 1999: 226) ‘It’s not in the cards, not in my future to leave.’

Not just the translation, which is of course wrong, even if the sentence had been correct. In all of the examples in this section, whether drinking with a buddy, living in an apartment, working with Gurov, being a chamberlain, or sitting in the Senate, the action is one that the Subject wants to perform, but because of a condition or situation p will not be able to do in the future. In order for (11) to be correct, uxodit’ has to fit the category of a desired action that is impossible due to some event p.

 Alina Israeli

Let us first examine the verb podnimat’ ‘to lift’ as in (3.7): (3.7) a. *Мне не поднимать книгу с пола.  ‘I will no longer pick the book up from the floor.’ b. ?Мне не поднимать книги с пола.  ‘I will no longer pick the books up from the floor.’

Example (3.7a) refers to one possible act of picking up a book from the floor, which no longer will be performed despite the Subject’s desire, while (3.7b) refers to the activity of picking up books from the floor, which will no longer take place. As (3.7a) shows, this type of infinitive construction does not tolerate single acts, but only activities, as in (3.7b), whose questionable quality stems from the non-desirability of performing such acts. Insofar as (11) refers to a single act, it is impossible. If a context could be found where uxodit’ ‘to leave’ means an activity, then it would become possible: (11′) Мне уже не уходить в море на долгие месяцы. ‘I will no longer be able to go away to the sea for long months.’

In this case the speaker would want to continue going out to the sea as he used to, but something happened (p) so that this is no longer possible. This construction up to this point has meant wistful impossibility. There is another but very specific usage which could be considered antonymous to subtype 1.1.1, that is with ne byvat’ or ne byt’. It means the prediction of an impossible event, that is of an event that can never take place or a state that can never occur. (3.8)  Враки! Свадьбе не бывать! Что ж из того, что он камер-юнкер. [Н. В. Гоголь. Записки сумасшедшего (1835)] ‘Lies! The marriage will never take place! So what if he is a chamber-junker.’ (3.9) – Только треба и то помнить: хоть тресни синица, а не быть ей журавлем! [В. Я. Шишков. Емельян Пугачев. Книга вторая. Ч. 1–2 (1939–1945)] ‘ “One only needs to remember this: no matter what a chickadee would do, she could not be a crane!” ’

However, the construction may mean that the speaker or the group associated with him, will not allow something to happen, the English equivalent being ‘over my dead body’. (3.10) Революции дедушка не принял. А бабушка говорила: «Неужто ты – Симонов! – будешь мужичонков обучать? ! Не бывать этому!» И всей семьей они уехали под Смоленск в свое имение, село Барадуличи. [Людмила Гурченко. Аплодисменты (1994–2003)] ‘Grandfather did not accept the revolution. And grandmother used to say, “­Really, you, Simonov! Will you teach peasants?! That will never happen!” And the whole family went to their estate near Smolensk, the village of Baradulichi.’



Dative-infinitive constructions in Russian 

Here also belongs Veyrenc’s example (5), particularly since the phrase is uttered by Marfa Timofeevna, Liza’s great-aunt, who may have some influence in the matter: (5)  A Лизе за Паншиным не быть, не беспокойся; не такого мужа она стоит. (Tургенев. Дворянское гнездо) ‘And Liza will not marry Panshin, don’t worry; she is worthy of a ­better ­husband.’

This group includes two anomalies, two expressions that follow the pattern grammatically, but whose meaning can only be understood if one gets to the extreme point of the verb: –– (кому) не привыкать = N will never have to get used to X because N is used to X –– (кому) не занимать (опыта etc.) = N will never have to borrow experience because N is very experienced (3.11) Африканскому кочевнику не привыкать к жаре, а каково ему придется на морозе? [Юрий Нечетов. Духи благосклонны (2004) // «За рулем», 2004.02.15] ‘An African nomad is used to the heat, but how will he feel when there is frost?’ (3.12) Только на предприятии не привыкать к трудностям, справлялись здесь и не с такими проблемами, а значит, есть опыт и знания. [«Светотехника» и «Электротехника» – 12 лет в рынке (2004) // «Пермский строитель», 2004.05.11] ‘Except that at the enterprise they are used to difficulties, they have managed even worse problems, which means they have experience and knowledge.’ (3.13) Однако Гоярыну, в свою очередь, было не занимать опыта. [Дмитрий Емец. Таня Гроттер и колодец Посейдона (2004)] ‘However, in turn, Goyaryn had plenty of experience.’ (3.14) самомнения твоему свекру не занимать. [Новые родственники – как уживаемся? (форум) (2008)] ‘your father-in-law has a very high opinion of himself.’ (3.15) Выдумки ребятам не занимать. [Татьяна Морозова. Мы говорим: «Нет – капитализму!» (2004) // «Правда», 2004.10.29] ‘Children are extremely inventive.’

1.4  Model Dat + ne + INFPF Unlike the previous types, this one does not deal with predictions of any kind. It  only states the perception of facts. This construction describes the Subject’s  perceived ability or, to be more precise, physical or mental inability to perform V:

 Alina Israeli

(4.1)  И, когда я встретил свою тогда ещё будущую третью жену Наталью, и узнал, что она живёт там же, где и остальные, но в соседнем доме, то пришёл к выводу, что от судьбы мне не уйти… [Светлана Ткачева. День влюбленных… (2003) // «100% здоровья», 2003.01.15] ‘And when I met my then future third wife Natalya and found out that she lives at the same place as the others but in the next building, then I ­concluded that I could not escape my fate…’ (4.2)  Она знала: абсурд ей не сыграть. [Галина Щербакова. Актриса и милиционер (1999)] ‘She knew: she won’t be able to play the absurd.’ (4.3)  Я поискал швейцара, его не оказалось. Я пошел назад, но не смог разыскать квартиру, где жил Пастернак. Это был дом с переходами, коридорами и полуэтажами. Я понял, что мне не выбраться до утра, и покорно сел на ступеньку. Лестница была чугунная, под ногами копошилась ночь. Вдруг дверь открылась. Я увидал Пастернака. [И. Г. Эренбург. Люди, годы, жизнь. Книга 2 (1960–1965)] ‘I looked for a doorman, there wasn’t any. I went back, but could not find the apartment where Pasternak lived. This was a building with passages, hallways and split-level floors. I realized that I would not be able to get out till morning, and I submissively sat down on the step. The staircase was made of cast iron, night was swarming under my feet. Suddenly a door opened. I saw Pasternak.’

The absence of Subject means that the situation is generalized, that is no one can ­perform V, although often, as in (4.4), it is based on the Speaker’s inability to ­perform V: (4.4) Нина Константиновна закудахтала: – Милая, дорогуша, вас прямо не узнать! Вы стали совсем, ну почти совсем как до… до вашего происшествия. [И. Грекова. Перелом (1987)] ‘Nina Konstantinovna started to squawk: “Darling, sweetheart, one ­cannot even recognize you! You became just, well, almost just like before… before the ­incident.’ (4.5) – Просто запираем дверь и уходим, – спокойно рассказывала Майечка, – замок хороший, фирменный, такой сразу отмычкой не открыть. [Дарья Донцова. Микстура от косоглазия (2003)] ‘ “We simply lock the door and leave,” Maya was saying calmly. “The lock is good, brand-name, you cannot open it at once with a picklock.” ’

The mistranslated example (12) from Timberlake and others with ne minovat’ which means ‘will not be able to avoid/escape’ also belongs to this type: (12) Нам не миновать гибели. *It’s not for us to avoid disaster. (Timberlake 2004: 8) ‘We will not be able to avoid death, i.e. we are doomed.’



Dative-infinitive constructions in Russian 

(4.6)  «Смертного часа нам не миновать,» – писал он брату Александру в 1888 году, то есть двадцати восьми лет от роду, … [Руслан Киреев. Чехов. Посещение Бога // «Нева», 2004] ‘ “We cannot avoid death,” he wrote to his brother Alexander in 1888, that is at the age of twenty-eight,…’ (4.7)  Фетинья из хозяйского разговору поняла, что Юрку кнута не миновать. [П. П. Бажов. Шелковая горка (1947)] ‘Fetinya from the conversation of the masters understood that Yurko will not avoid punishment by the whip.’

Ne minovat’ is commonly used with a generalized Subject: (4.8)  Отсрочка кадровой революции не повлияет на расстановку сил в Думе – большой чистки не миновать. [Александр Садчиков. Вопрос решен, но отложен. В Думе обсуждают кандидатуры нового спикера (2001) // «Известия», 2001.06.28] ‘The delay of the personnel revolution will not have any effect on the distribution of power in the Duma, a big purge is unavoidable.’

Fleisher’s example (13), which means perceived inability and belongs to this group (i.e. type 4), was misclassified as a mixture of type 3 to mean predestination (“not in the cards”) and type 5 below (“it is not for me to V”). (13) Мне не сдать экзамен.(Fleisher 2006: 4) *It is not (in the cards) for me to pass the exam. ‘I won’t be able to pass this exam.’

1.5  Model ne +Dat + INFIPF This construction means that ‘it is not up to the Subject to perform V because there is someone else who could do it better or knows better etc.’ Most often the one who could do it better is either the addressee or a higher power. It is often stated explicitly as to who is better suited than the Subject. The translation pattern attributed by Timberlake, Moore & Perlmutter, and partially by Fleisher ‘it’s not for me/us to V’ actually would be correct for this type and this type only. The most prototypical example is ne (mne/nam/emu/…) sudit’ – ‘it’s not up to (me/us/him/etc.) to judge’. Most typically it includes verbs of communication that transmit information: govorit’ ‘to say’, ob”jasnjat’ ‘to explane’ , učit’ ‘to teach’, etc. (5.1)  А собственно полемику с Солженицыным я пытался свести к минимуму и быть при этом предельно корректным. Не мне судить о том, в какой мере всё это удалось… [Валерий Прайс. Семён Резник: «Делай что должно, и пусть будет что будет…» (2003) // «Вестник США», 2003.10.29] ‘And the polemics with Solzhenitsyn proper I tried to reduce to a minimum and to be at the same time absolutely polite. It is not for me to judge to what extent I succeeded…’

 Alina Israeli

(5.2)  Не мне говорить об этом. Вам лучше известно о преступных деяниях, совершенных бывшими руководителями МВД, настоящими государственными преступниками. [Обращения родственников репрессированных командиров Красной Армии к руководителям страны (1937–1960)] ‘It’s not for me to speak about this. You know better about the criminal deeds committed by the former leadership of MVD, real state criminals.’ (5.3)  Не мне писать о Зайцеве-художнике. Есть специалисты-искусствоведы, есть коллеги, есть женщины, в конце концов… [Григорий Горин. Иронические мемуары (1990–1998)] ‘It’s not for me to write about Zaitsev the artist. There are specialist art ­critics, there are colleagues, there are women after all…’ (5.4)  – Это древний обычай, – сказал дедушка, – и не нам отменять обычаи. Без традиций общество деградирует. (К. Булычев. Глубокоуважаемый микроб) ‘ “This is an ancient custom,” said grandfather. “And it’s not up to us to ­abolish customs. Without customs the society goes downhill.” ’

An example of the higher power is provided by (5.5): (5.5)  А сколько жить отмерено, это не нам знать. [И. Грекова. Перелом (1987)] ‘And how long life will be, that’s not for us to know.’

1.6  Model *ne + Dat + INFPF This construction simply does not exist. All of the following four constructions have the particle li. 1.7  Model Dat + li + INFIPF The meaning of this type is close to that of type 5: ne emu pet’, ne mne sudit’. However, unlike type 5 where there is an explicit or implicit appropriate Subject who could and should perform the action V, in this case there are no such implications. It simply denies the given Subject the right to do so: the Subject may not and should not perform V. In (7.1) the Subject has no right to talk about freedom since he never sacrificed or risked himself for freedom, and in (7.2) the Subject in question is not suited for the medical profession: (7.1) – Оставьте, – махнул рукой Свинецкий. – Вам ли говорить о свободе? Жертвовали ли вы собой? [Дмитрий Быков. Орфография (2002)] ‘ “Oh please,” Savitsky made a disparaging gesture. “How could you speak of freedom? Did you risk yourself?” ’



Dative-infinitive constructions in Russian 

(7.2)  Вот уж недоразумение: неловкая, неприветливая, ей ли работать в медицине, с детьми! [Лев Дурнов. Жизнь врача. Записки обыкновенного человека (2001)] ‘It is a total mismatch: she is awkward, gloomy, how could she work in medicine, with children!’

1.8  Model Dat + li +INFPF This group is limited to rare examples with mental verbs, such as ponjat’ ‘to understand’ ‘zabyt’ ‘to forget’ and a few others. That is why spet’ ‘to sing’ cannot be used, hence it is marked with a question mark. This type is quite rare and pertains to high style. Many examples in RNC are from poetry, which underscores the elevated style of this construction. This construction questions the Subject’s ability or appropriateness to perform the action V, and consequently asserts the opposite. In other words, how could the Subject of all people perform V? (8.1)

А юность была – как молитва воскресная… Мне ли забыть ее? (Анна Ахматова) ‘Youth was like a Sunday prayer… How could I forget it?’

(8.2)  При виде ее спокойно-мечтательного лица Казарин ощутил страшное одиночество: ей ли понять его смятение? [Дмитрий Быков. Орфография (2002)] ‘Seeing her calm-dreamy face Kazarin felt awful solitude: how could she, of all people, understand his confused soul?’ (8.3)  Вам ли забыть, как строго римская церковь осуждает грех самоубийства? [Н. Н. Шпанов. Ученик чародея (1935–1950)] ‘How could you, of all people, forget how severely the Roman Catholic Church denounces the sin of suicide?’ (8.4)  И мне ли оставить тебя, – я русский, сын русского, я из самых недр твоих. [А. М. Ремизов. Взвихренная Русь (1917–1924)] ‘How could I of all people, a Russian, a son of a Russian, abandon you, I am from your bowels.’

1.9  Model Dat + li + ne + INFIPF This construction means that the Subject of all people should do V: if anyone should do V, it is the Subject who should do it. Out of 95 examples on RNC, 77 are with the verb znat’ ‘to know’ (mne/nam li ne znat’, etc.) and 18 are with other verbs. Most of the examples with verbs other than znat’ are from 19th century prose.

 Alina Israeli

In example (9.1), the speaker talks to God (capital Тy), so if anyone knows what awaits Apenushkin, it is God. In (9.2) the author, who spent nearly twenty years in Kolyma, states that she of all people should know what а Kolyma blizzard is like. (9.1)  «Я прошу Тебя: дай мне силу и дай мне желание спасти человека, имя которого Ты знаешь. Потому что хватит оперировать всеобщими категориями добра и зла, во имя их творилось всё зло». «Тебе ли не знать, что Апенушкина ждёт гибель. Дар прозрения позволяет мне видеть то, что произойдёт с Твоего соизволения.» [Анатолий Азольский. Лопушок // «Новый Мир», № 8, 1998] ‘ “I beg Thee: give me the strength and the desire to save the man whose name Thou knowest. Because it is time to stop operating in general ­categories of good and evil in whose name all the evil was created.” “Thou of all beings should know that death awaits Apenushkin. The gift of ­clairvoyance allows me to see what will happen with Thine permission.” ’ (9.2)  Колымские бураны – не шутка. Уж я-то должна это знать. Я, конечно, знала. Мне ли не знать… Сколько историй о гибели целых этапов и отдельных людей наслышалась я за свой срок! [Е. С. Гинзбург. Крутой маршрут (1990)] ‘Kolyma blizzards are not a joke. I should know. And of course I knew. I of all people knew… I’ve heard so many stories of deaths of entire groups and individuals during my prison term.’ (9.3)  В ходу были паспорта старого образца с фото три на четыре (мне ли не помнить этого)… [Николай Климонтович. Фотографирование и прочие Игры (1988–1990)] ‘Old fashioned passports were then in use, with photos of three by four [centimeters] (I of all people should remember that)…’

1.10  Model Dat + li + ne + INFPF There are altogether 12 examples in the Russian National Corpus of this type, hardly a commonly used construction. It has the same meaning as type 9 (§1.9 above), except the verb is in perfective. It means: who other than the Subject would be able to accomplish V?. (10.1) Девушка с красным дипломом университета, Лилька могла мгновенно о нем забыть, если надо было что-то объяснить родному народу. Ее уважали технички, электрики, слесаря, даже продавщицы, потому что Лилька всегда держала в кармане пультик на спуск некоей себя другой, не знающей французского и латыни и не почитающей Аполлинера. Так вот… Ей ли не смочь? [Галина Щербакова. Восхождение на холм царя Соломона с коляской и велосипедом (2000)] ‘A young woman who graduated from the university with top honors, Lilya could instantly forget it [and use foul language] if she needed to



Dative-infinitive constructions in Russian 

explain something to her own people. She was respected by cleaning ladies, ­electricians, workers and even saleswomen, because Lilya always kept in her pocket a switch for a different self, the one who does not know French or Latin and does not revere Apollinaire. So… How could she of all people not be able to?’ (10.2) – Виталий. Нам нужно развестись. И это было бы–все. Она же сказала–стоя у мойки, чистя сковородку или еще что-то, сказала, не обернувшись (Талий и сам в это время не смотрел на нее, но ему ли не суметь определить по голосу, даже не видя, смотрит она на него или нет!), сказала в какой-то своей паузе. [Алексей Слаповский. Талий // «Дружба народов», 1999] ‘Vitaly. We have to get divorced. And that was all. She said it standing by the sink, scrubbing the frying pan or something else, said without turning around (Taly himself was not looking at her at that time, but he of all people should be able to determine by her voice, even without looking, whether she is looking at him or not!), she said it in some kind of pause of hers.’

2.  Dative-infinitive constructions in interrogative sentenses All constructions with Dative in interrogative infinitive sentences are presented in Table 2.

Table 2.  Dative-infinitive constructions in interrogative sentenses Dative-infinitive constructions in question

Realisation

1.

(Dat) + INFIPF?

петь?

2.

Dat + INFIPF?

мне петь (, что ли)?

3.

Datkomu + INFIPF?

кому петь?

4.

wh-+ Dat + INFIPF?

как ему петь?

5.

(Dat) + INFPF?

(мне) спеть?

6.

wh- + (Dat) + INFPF?

что (мне) спеть?

2.1  Question (Dat) + INFIPF [петь?] Questions with imperfective verbs, transitive or intransitive, mean ‘Should I/we begin V?’. The action itself is expected. There is a stress on the verb. This construction is often found where food is being served: (Q1.1) [Виолетта] Горячее подавать? [Эльдар Рязанов, Эмиль Брагинский. Вокзал для двоих, к/ф (1982)] ‘Should I serve the main course?’

 Alina Israeli

(Q1.2) [Полищук] Ну что/начинать? [Юрий Мамин, Владимир Вардунас. Праздник Нептуна, к/ф (1987)] ‘So, should we begin?’ (Q1.3) [К., жен, 16] Тушенку открывать/ да? [Разговор в компании, приехавшей отдохнуть на остров на Волге // Из коллекции Саратовского университета, 2002] ‘Should I open the can of meat now?’

Question words in this case cannot represent the Subject but only the indirect Object: (Q1.4) [В., жен] Кому наливать? [Ш., муж, 42] Никите/ ему ищ этого мало будет// [Разговоры в компании, отдыхающей на Волге (день 2-й) // Из коллекции Саратовского университета, 2002] ‘ “To whom should I pour (some)?” “To Nikita, this won’t even be enough for him.” ’

This subtype allows the use of the 3rd person Subject. In this case the Speaker is a proxy for the Subject: (Q1.5) Ему начинать? ‘Should he begin?’ (Q1.6) Ей одеваться? ‘Should she get dressed (start getting dressed)?’

2.2  Question Dat + INFIPF [мне петь (, что ли)?] This type of question signals the perceived obligation or expections on the part of the Speaker to perform action V. There is a stress on the verb. (Q2.1) При мне скучала, подтирая лужи от зонтиков со словами “ходют [sic!] тут, капают, а мне убирать что ли”. (http://oio-oio.blogoda. ru/b/503781496/242529.html?bpg=0) ‘She was bored in my presence, wiping puddles from the umbrellas ­saying “they are walking around, dripping, and I am supposed to clean up or what?” ’ ( Q2.2) [Татьяна З., жен, 23] На бензин-то ради бога/если б на икру я бы не дала. [Татьяна, жен, 30] А я чё/он подошёл и взял. Мне спрашивать что ли? [Разговор об аквапарке, одежде и котах // практика, 2008] ‘ “For gas of course, but for caviar I would not have given.” “What, he came and took. Am I supposed to ask or what?” ’

The fact that there is a specific time does not make the statement less general (cf. type 1.1.3 above):



Dative-infinitive constructions in Russian 

(Q2.3) – Может, оно и к лучшему? – Так мне завтра приходить? – тоже шепотом спросил Митя. [Андрей Житков. Кафедра (2000)] ‘ “Maybe it’s for the better?” “So should I come tomorrow?” Mitya asked in a whisper as well.’

2.3  Question Datkomu + INFIPF [кому петь?] Questions of the type komu + V fall into two groups: those where the Subject can be designated by the question word and those where the Subject may never be designated by the question word. The question Datkomu + INFIPF describes the first type. This type of question typically occurs in a group setting where different ­participants take turns performing V. The question means ‘whose turn it is to do V?’. Very rarely do we find an open-ended question which means ‘no one will perform V’. (Q3.1) Один за Волка играет, другой – за поросенка. Кому ходить первому – решайте сами. [Три поросенка и серый волк // «Трамвай», 1991] ‘One plays the Wolf, the other the piglet. Who goes first is up to you.’ (Q3.2) Каждый из игроков в марьяж кладет по фишке в каждую коробочку и затем разрешается вопрос, кому сдавать карты. (http://party.fome.ru/ ras-6-28.html) ‘Each of the players in “marriage [card game]” places a token in each box and then the question must be decided as to who deals the cards.’ (Q3.3) Сестренки-снежинки всегда соглашались. Да и на самом-то деле, не все ли равно, кому водить первым? (Н. Павлова. Миллион миллионов и одна снежинка) ‘Sisters-snowflakes always agreed. And indeed, does it really matter who is going to be “it” first?’

There is a possible variation of this question when there is a choice as to which one of the possible Subjects will perform the action, as in the title of the article discussing two soccer teams (Q3.4): which one will win and consequently sing its national anthem: (Q3.4) Барыс – Динамо, предисловие. Кому петь гимн? (http://www.pressball.by/ news.php?t=0709&id=67955) ‘Barys – Dynamo, foreword. Who will sing the anthem?’ (Q3.5) В конце концов всё оборачивалось банальнейшей ситуацией – кому пропадать? Либо ему, либо Краснянскому. [Василь Быков. Бедные люди (1998)] ‘After all everything turned into the most banal situation – who should die? Either he, or Krasnyansky.’

 Alina Israeli

This type of question is often found in subordinate clauses; somebody has to chose, decide, define etc. as to who from a clearly defined group of people should perform V. (Q3.6) Искусство же тренера состоит в том, чтобы только по ему одному известным признакам определить, кому выходить на лед в тот или иной ответственный момент. (http://www.hock.ru/0203/p/0/292.htm) ‘The art of the coach consists of figuring out, based on the features known only to him, who should step on the ice at any given decisive moment.’

An open-ended question, implying that no one will perform V is represented by (Q3.7): (Q3.7) – Моя обязанность за порядком смотреть. Случится чего, – говоря, старушка, мелко перемещаясь, загородила коридор, – случится, и кому отвечать? [Роман Сенчин. Елтышевы (2008) // «Дружба Народов», 2009] ‘ “My responsibility is to watch over the order. If anything happens,” while speaking the old woman, making small steps, barricaded the hallway, “if anything happens, who will answer?” ’

Interestingly enough, there is no perfective counterpart to this question. 2.4  Question wh-+ Dat + INFIPF [как ему петь?] This subtype includes rhetorical questions or subordinate clauses introduced by a wh- word: (Q4.1) Откуда им, бедным, знать, что, «беспричинные богатство и знатность порождают смуту»? [Иван-дурак и миллиардер Шахновский (2003) // «Завтра», 2003.08.22] ‘How could these wretched folk know that “unfounded wealth and nobility cause riots”?’ (Q4.2) Позже посмотрел «Волосы» и понял, что мне делать в жизни. [Валерий Кичин. Москва у стен Нотр-Дам (2002) // «Известия», 2002.05.21] ‘Later I saw “Hair” and understood what I am supposed to do in life.’ (Q4.3) Каково ей [Слуцкой] сражаться в Америке против Америки? [Игорь Порошин. Против Америки с любовью. Ирина Слуцкая идет второй после короткой программы (2002) // «Известия», 2002.02.20] ‘How does it feel to her [Slutskaya] to fight in America against America?’

2.5  Question (Dat) + INFPF [(мне) спеть?] Unlike Q-1 (see §2.1.) where the action is expected, in this case it is a question of whether or not the action itself should take place. The stress is on the verb.



Dative-infinitive constructions in Russian 

(Q5.1) – Мне вам вечером позвонить? (А. Терехов. Каменный мост) ‘Should I call you in the evening?’ ( Q5.2) [Ж1, жен, 60] тебе помыть персик? [И43, муж, 60] угу // [Разговор за ужином с женой // Из материалов корпуса «Один речевой день», подготовленного группой А.С.Асиновского, 2009] ‘ “Shall I wash you a peach?” “Yeah.” ’ (Q5.3) Водки налить? [Денис Нейманд и др. Жесть, к/ф (2006)] ‘Should I pour (you) vodka?’ (Q5.4) [Сергей, Александр Феклистов, муж, 45, 1955] Сонечка/те налить? [Владимир Меньшов, Марина Мареева. Зависть богов, к/ф (2000)] ‘Sonya, shall I pour you some?’

And consequently example (14) earlier can only mean ‘Shall I /should I’. (14) Тебе налить? (Maurice 1996: 91) Shall I pour you [some]? *Will you pour [some]?

(both Fortuin 2000: 264)

This subgroup also allows a proxy question for a third party: (Q5.5) Eму прийти? ‘Should he come?’ (Q5.6) Ей позвонить? ‘Should she call?’

The only pragmatic context that could allow other persons without the Speaker being a proxy is the case of a suggestions: (Q5.7) Может, тебе слинять отсюда, а? Тут могут быть, ну, ты понимаешь. [Василий Аксенов. Новый сладостный стиль (2005)] ‘Maybe you should make yourself scarce, huh? There could be here… well, you understand.’

2.6  Question wh-+ (Dat) + INFPF [что (мне) спеть?] The type Dat + INFPF does not seem to occur without a question word. C ­ onsequently type 1.2 is discussed in this section. (Q6.1) А Самоделкин стоял на городской площади, не зная, куда ему пойти. [Юрий Дружков (Постников). Приключения Карандаша и Самоделкина (1964)] ‘And Samodelkin stood on the city square, not knowing where he should go.’

 Alina Israeli

This type has a high degree of generalization which triggers the absence of Dat. The “generalized” use often refers to the Speaker or the protagonist whose point of view is represented in the narrative. Let us examine one example from RNC: (Q6.2) К кому поехать, кому сообщить, с кем посоветоваться о деньгах? [Ю. Н. Тынянов. Малолетный Витушишников (1933)] ‘To whom should she go, who should she inform, who should she ask for advice about the money?’

The larger, immediately preceding context tells us that it is Varen’ka Nelidova whose point of view is represented. While this is author’s text, it is Varen’ka ­Nelidova, who has received an envelope with two hundred thousand rubles, who is asking the question in (Q6.2). We are dealing with the case of discours indirect libre and therefore the translation should state not ‘to whom should one go’ but rather ‘to whom should she go’. We find other similar examples (Q6.3–Q6.5); in (Q6.3) we have to add “he” in the translation: (Q6.3) Мастерским виражом он оторвался от невидимой погони и задумался. У кого спросить, кто такой Георгий Валентинович? [Анатолий Азольский. Лопушок // «Новый Мир», № 8, 1998] ‘With a masterful turn he pulled away from the invisible chase and started thinking. Who should he ask as to who Georgy Valentinovich is?’

This is also typical of advice columns, real or cited, so in translation we have to add “one” or “you”: (Q6.4) Что подарить своим сотрудникам, клиентам и партнерам? [Электронное объявление (2004)] ‘What should one/you give as a gift to one’s/your colleagues, clients and partners?’ (Q6.5) Так и возникает у городского жителя старый советский вопрос: куда пойти учиться? [Сергей Баричев. Учиться, учиться и еще раз? (2004) // «Бизнес-журнал», 2004.08.17] ‘That’s when a city dweller asks the old Soviet question: Where should one/ you go to study?’

This explains why Veyrenc’s example (2) is incorrect or rather impossible ­without a question word: (2) [*]Emu vernut’sjapf domoj. It can be improved either by making ith into a question of type Q-5 above (see §2.5.): (2’) Emu vernut’sjapf domoj? ‘Should he return home?’ or by adding a question word: (Q6.6) И главное – как ему вернуться домой?! (http://fictionbook.ru/author/ anna_ustinova/uchenik_maga/) ‘And the main thing: how could he return home?’



Dative-infinitive constructions in Russian 

3.  Conclusion In conclusion, the article has analyzed the semantics and pragmatics of various combinations of a dative-subject, a verb and the two particles nе and li. It has demonstrated that aspect is a factor; perfective and imperfective verbs produce different semantic outcomes and have different combinatory abilities. While there are mathematically ten combinatory possibilities of the four components, only nine actually exist. Addionally, one of those (Dat + INFPF) can occur only in combination with a question word. Semantically, the first type (Dat + INFIPF) has four subtypes, three of which either have the power of prediction or involve an announcement of an impending action. The second type is the one that occurs only in questions. The third type (Dat + ne + INFIPF) describes the impossibility of a previously possible action due to some event or condition or threat in case some condition is not eliminated. The fourth type (Dat + ne + INFPF) means physical impossibility to perform the action V by the Subject. Types five and seven (the sixth type does not exist) are quasi-synonymous, stating that the Subject is not the one who should do V. Types eight, nine and ten are rhetorical questions about the Subject’s likelihood of being able to do V.

References Ebeling, Carl L. 1984. On the meaning of the Russian infinitive. In Signs of Friendship, to ­Honour André G.F. van Holk, Slavist, Linguist, Semiotician, Joost J. Van Baak (ed.), 97–130. ­Amsterdam: Rodopi. Fleisher, Nicholas. 2006. Russian Dative Subjects, Case, and Control. Ms, University of California. Fortuin, Egbert L.J. 2000. Polysemy or Monosemy: Interpretation of the Imperative and the Dative-infinitive Construction in Russian. Amsterdam: Institute for Logic, Language and Computation. Fortuin, Egbert. 2005. From necessity to possibility: The modal spectrum of the dative-­infinitive construction in Russian. In Modality in Slavonic Languages. New Perspectives, Björn ­Hansen & Petr Karlík (eds), 39–60. München: Otto Sagner. Fortuin, Egbert L.J. 2007. Modality and aspect: Interaction of constructional meaning and aspectual meaning in the dative-infinitive construction in Russian. Russian Linguistics 31(3): 201–230. Maurice, Florence. 1996. Der modale Infinitiv in der modernen russischen Standardsprache ­(Slavistische Beiträge 340), München: Otto Sagner. Moore, John & Perlmutter, David M. 1999. Case, agreement, and temporal particles in Russian infinitival clauses. Journal of Slavic Linguistics 7(2): 219–246. Timberlake, Alan. 2004. A Reference Grammar of Russian. Cambridge: CUP.

 Alina Israeli Veyrenc, Jacques. 1973. Structure de la proposition infinitive impersonnelle (en russe contemporain). Revue des Etudes Slaves 49: 351–358. Veyrenc, Jacques. 1979. Les propositions infinitives en russe. Paris: Institut d’Etudes Slaves. Zolotova, Galina. A. 1990. O russkix infinitivnyx predloženijax. In Carina Amicorum: Carin Davidsson Septuagenariae 23.3.1990 oblata, Martina Björklund, Helena Lundberg & Janina Orlov (eds), 341–350. Turku: Åbo Akad.

On the nature of dative arguments in Russian constructions with “predicatives”* Sergey Say

Institute for linguistic studies, Russian Academy of Sciences, St.Petersburg Based on diachronic corpus data, the ability of Russian predicatives (категория состояния, items like холодно ‘it is cold’, грустно ‘it is sad’, приятно ‘it is pleasant’) to co-occur with dative Experiencers is compared to the ability of corresponding adjectives to co-occur with Experiencers. The hierarchy (predicatives > short forms of adjectives > long forms of adjectives) is put forward, which reflects decreasing ability to co-occur with dative Experiencers. The predicatives are divided into several classes based on how their syntactic properties are related to properties of corresponding adjectives. It is shown that none of the usual types of analysis (structural-syntactic, derivational, constructional) adequately captures the properties of all the groups of predicatives, but rather that each of these approaches is best suitable for a particular subtype of predicatives. It is hypothesized that a possible way out could be to assume that in the experiential predicative construction the dative position is directly linked with the Experiencer, so that this structure is “blind” to the role structure of the head lexeme. Keywords:  Russian; dative; argument structure; predicatives; adjective; corpus linguistics; derivation; constructional approaches; diachrony

1.  Outlining the problem Russian “predicatives” (“предикативы”, also known as “категория состояния”, “предикативные наречия” etc. in Russian linguistics) are a class of forms that can

* The study was supported by a grant from Russian Foundation for Humanities, No. 11-04-00179a “Verb argument structure variation and verb classification in languages of various structural types” and by the RAS fundamental research programme “Corpus linguistics” (project “Development of a corpus-based reference system on Russian syntax” ). I am indebted to Alexander Letuchiy, Maria Ovsjannikova and the two anonymous referees for useful comments that helped me a lot when editing this paper. The usual disclaimers apply.

 Sergey Say

only function in the position of a (non-verbal) predicate and are not able to take canonical subjects in the Nominative case:1 (1)  Мне было приятно хранить эту жалкую тайну. (Сергей Довлатов. Чемодан (1986)) ‘I was glad to keep this poor secret.’ (2) Право, мне жаль, что я его обманул. (А. И. Куприн. Куст сирени (1894)) ‘I really regret that I have deceived him.’

Like all non-verbal predicates, predicatives take overt copula быть ‘to be’ in all tenses (1), except for the present indicative, in which they are used without overt copula, as in (2).2 Constructions with predicatives are usually viewed as impersonal (subjectless) (Vinogradov et al. 1953: 632; Guiraud-Weber 1984). However, infinitives and subordinate clauses, when they are present in constructions with predicatives, do show some syntactic properties of subjects (Letučij 2012). In any case, the copula in these constructions is always in the default agreement form, e.g. 3rd person singular in the future tense (будет) or neuter singular in the past tense (было). Whether this non-agreeing pattern is due to the lack of subject or to the lack of case property in clause-like subjects is irrelevant for the present discussion. Although there are predicatives that are not homonymous to any other expression, like жаль ‘it is a pity/regretful’ in (2), the vast majority of predicatives have the form of expressions of other types (for an overview see V ­ inogradov et al. 1953: 633–634; Cimmerling 2003), most frequently of neuter “short” (i.e. predicative, as opposed to “long”, i.e. attributive) forms of adjectives. Traditional ­grammars usually differentiate between neuter short adjectives and predicatives based on whether there is an overt nominative subject; hence курение вредно ‘smoking (noun) is harmful’ is viewed as a clause with nominative subject, zero copula and neuter short form adjective that agrees with the subject in gender, whereas курить вредно (same meaning, literally ‘to smoke is harmful’) is often viewed as an impersonal clause with predicative.

.  This is the definition that is used throughout this paper. In the literature the term предикативы is sometimes understood differently, either in a wider sense, e.g. including agreeing short forms of adjectives (Timberlake  2004: 281), or in a narrower sense, e.g. as limited to non-agreeing modal predicates only (Švedova 1980: 705). .  In fact, non-finite verbal predicates (independent infinitives and participles) behave similarly to non-verbal predicates in this respect, that is, take an overt copula (cf. мне было не выйти ‘I could not escape’) in all tenses except for the present (cf. мне не выйти ‘I cannot escape’). On these constructions see Israeli, this volume.



On the nature of dative arguments in Russian constructions with “predicatives” 

Many researchers analyze predicatives as a separate part of speech (starting with Ščerba 1928/2004 and then Vinogradov 1947), whereas others are against this view (Babby 1974). The former analysis naturally suggests itself for those few predicatives that can’t be used otherwise, like жаль ‘it is a pity’; it would be problematic to treat these words as belonging to any major part of speech. However, adjective-like predicatives can be treated either (1) as just neuter short forms of adjectives used in particular types of syntactic environment and endowed with particular types of meaning or (2) as separate lexemes that formally coincide with neuter short forms of adjectives. Evidently, the choice between the two analyses is largely dependent upon the criteria for part-of-speech differentiation and identification.3 This controversial topic is not directly addressed below. What is beyond doubt, though, is that “adjectival” predicatives are closely related to ­corresponding adjectives and share many of their semantic and syntactic ­properties, cf., e.g. the discussion in Bonč-Osmolovskaja (2003: 144–145). A possible line of research, then, is to compare individual grammatical properties of predicatives with those of corresponding adjectives. It is exactly this line that is pursued here. The particular property that will be explored is the ability of the predicative to co-occur with the dative argument denoting a sentient participant whose role can be broadly characterized as “Experiencer” (including perceiver, bearer of a modal state, sometimes cogniser, etc.). The main empirical question of the study is the following: how the predicative’s (in)ability to take a dative E ­ xperiencer can be related to (a) whether the Experiencer is present in the semantic structure of the corresponding adjective and, if yes, to (b) what is the syntactic position of the Experiencer in constructions with the corresponding adjective. Evidence on this empirical question is further relevant for a more abstract theoretical issue: what is the nature of the relationships (both on the semantic and syntactic levels) between argument structures of predicatives and corresponding adjectives? One can outline beforehand the following calculus of possible answers to that latter question.

.  Many predicatives coincide in form not only with adjectives, but also with adverbs. Some researchers even view predicatives as a subclass of adverbs, cf. “предикативные наречия” in Švedova (1980: 705). However, for the topics discussed in this paper the relationship between predicatives and adjectives is more relevant than the relationships between these two types of entities and adverbs.

 Sergey Say

A. The predicative and the corresponding adjective are manifestations of the same lexical entry and have the same lexical argument structure (thematically motivated or given by lexical rule), including the dative slot. Under this assumption, the differences between predicatives and adjectives can only be accounted for by very general principles of syntax (similar to, e.g. differences between finite and non-finite verb forms), while the dative is subcategorized for and the Experiencer is semantically selected by the unitary lexeme as such. B. There is a derivational relation between the predicative and the adjective. In particular, predicatives can be related to adjectives via a certain kind of rule that influences argument structure (e.g. the dative slot can be introduced) and changes the semantics of the lexeme correspondingly. C. The dative position in the argument structure stems from the syntactic properties of the construction as such. This position is associated with a particular component in the meaning of the construction. It is not subcategorized for by the predicative lexeme, although this lexeme has to somehow match the meaning of the construction. In previous literature on the topic it was often tacitly assumed that predicatives form a homogeneous class and much more effort was applied to propose various across-the-board analyses than to unearth relevant differences between predicatives in terms of their argument structure (but see a notable exception in (Kustova 2002)). Importantly, in the present study it is not a priori assumed that predicatives must empirically show a uniform pattern of behaviour and, consequently, it is not assumed that they have to call for a uniform analysis in terms of the possibilities outlined above. On the contrary, I will try to figure out the individual patterns of semantic and syntactic correspondence between argument structures in constructions with predicatives and adjectives, establish groups of predicatives based on these patterns and eventually point out those theoretical analyses that will perform better for capturing the observed empirical findings in each case. The ultimate theoretical goal of this study is to compare several types of analyses (which seem to be mutually exclusive) and to show that, somewhat surprisingly, their adequacy varies depending on the choice of the data. Thus, essentially this study is data-driven: theoretical interpretations are chosen based on what is suggested by particular sets of data. The alternatives in A, B and C above were presented in a rather informal ­general manner. For any theorist, formal implementations of these analytic patterns would depend on which particular theoretical model is used. However, I assume that essentially the choice between the alternatives listed in A, B and C is relevant for various syntactic models, from traditional, to generative, to constructional etc. For example, the two extreme possibilities, “lexical” (A) and “syntactic” (C),



On the nature of dative arguments in Russian constructions with “predicatives” 

were actually proposed for various predicatives as early as in Peškovkij (1956 and earlier editions). In particular, Peškovskij argued that in мне было холодно ехать ‘I was cold when traveling’ “дательный падеж не связан ни со словом было, ни со словом холодно в отдельности, а только с обоими словами вместе, т.е. с самой безличной конструкцией, соответствуя дательному других безличных предложений” (Peškovskij 1956/2001: 356), which corresponds to analysis C here; by contrast, for мне было свойственно краснеть ‘I used to blush’ it was argued that, despite surface similarity, “дательный падеж зависит исключительно от прилагательного свойственно, управляющего вообще, во всяком положении дательным падежом” (ibid.), i.e. falls under the analysis in A. The remainder of the paper is organised in the following way. Section 2 contains a brief discussion of methodological issues. In Section 3 I propose a four-way classification of predicatives based on their and corresponding adjectives’ ability to co-occur with dative arguments. Sections 4 to 6 elaborate on this classification, providing information on the semantic interpretation of the predicative’s dative argument and on its relation to the argument structure of the corresponding adjective. Such an approach makes it possible to propose the second level of classification of predicatives, so that each of the three sections is devoted to a particular type of predicatives. Section 7 contains conclusions, including the comparison of analytic models that are mentioned with respect to individual classes of predicatives discussed in Sections 4–6. 2.  Data and methodology The data have been gathered from the Russian National Corpus (RNC; www.­ ruscorpora.ru); unless stated otherwise, all numbered examples given in separate paragraphs are taken from this source. Attention is paid not only to whether a particular type of structure is possible and attested in the RNC, but also to the frequency of various constructions, including diachronic changes in frequency.4 The procedure of gathering quantitative data consisted of two stages.

.  The preference for natural examples and corpus frequencies over contrived examples and introspection-based grammaticality judgements is largely due to ideological reasons (­decades-long debate on the nature of evidence in linguistics only shows that this issue remains largely a matter of personal conviction) and would not be discussed here in any detail. It can only be mentioned that for the constructions at issue such an approach can make a difference, since there are many types of uses that are somewhere in the grey zone between grammaticality and ungrammaticality. All this said, I don’t intend to say that grammaticality judgments will not be used in the study, though.

 Sergey Say

i. Automatic search for 53 different adjective-like predicatives and corresponding short and long adjectives in various syntactic environments. The list of predicates is a slightly shortened variant of the list that was used in ­Bonč-Osmolovskaja (2003: 147–148). The automatic search method was a pilot study that yielded rough estimates for overall frequencies of the very elements under study and their combinations with a dative nominal in either of the two adjacent positions, preliminary diachronic trends etc. However, these estimates were rather inaccurate, as there was a lot of corpus junk in search results. ii. Based on this pilot study, a more precise (and time-consuming) method was used for a small group consisting of 12 predicatives that co-occur with dative arguments (“X” in the translations corresponds to that argument, when overt): безразлично ‘X is indifferent’, грустно ‘X is sad’, известно ‘it is known to X’, любопытно ‘X is curious’, обидно ‘it hurts X/it is a pity to X’, приятно ‘it is pleasant for X’, скучно ‘X is bored/has a tedious time’, странно ‘it seems strange to X’, темно ‘it is (too) dark for X’, тесно ‘X is cramped’, тяжело ‘it is heavy/hard for X’, холодно ‘X is cold’. For each of these predicatives 100 random uses from modern (1979 onward) texts were manually analyzed. For every token it was established: (a) whether there is an overt argument that corresponds to an Experiencer-like participant; if yes, (b) whether this ­participant is coded with a dative NP or in some other fashion. The same procedure was then repeated for random 100-token samples of examples with corresponding short (only masculine and feminine forms were searched for the sake of unambiguousness) and long adjectives. Overall, 300 uses were analyzed for every stem under study. In order to implement the diachronic dimension, another 300-example sample was gathered for each stem in the same fashion from texts created between 1700 and 1850. Thus, in principle there were 600 examples per adjectival stem, although in some cases the actual groups of examples under analysis were smaller, because there were not enough tokens in the RNC. 3.  Co-occurrence with the dative: An overview The observed distribution of co-occurrence with the dative arguments on the part of predicatives, corresponding short adjectives, and long adjectives suggests the following two implicational generalizations:

(3) If the long (attributive) form of an adjective can co-occur with a dative argument, its short (predicative) form can co-occur with a dative argument as well. The reverse is not true.





On the nature of dative arguments in Russian constructions with “predicatives” 

(4) If the short (predicative) form can co-occur with a dative argument, then the corresponding predicative can co-occur with a dative argument as well. The reverse is not true.

In other words, the following hierarchy of forms represents their co-occurrence with the dative:

(5) Predicatives > Short (predicative) adjectives > Long (attributive) adjectives

Items that are higher on this hierarchy can show better or equal ability to co-occur with datives than items that are lower on the hierarchy.5 The hierarchy in (5) yields a four-way classification of predicatives, as schematically represented in Table 1. Table 1.  Co-occurrence of predicatives and adjectives with the dative: possible distributions Type 1

Type 2

Type 3

Type 4

predicative

–

+

+

+

short adjective

–

–

+

+

long adjective

–

–

–

+

Type 1 encompasses predicatives that cannot co-occur with the dative argument. Both predicatives in the following example belong to this group (cf. *мне тихо, *мне солнечно): (6)  На дворе октябрь, а солнечно и тихо, как летом. (А. П. Чехов. Вишневый сад (1904)) ‘It’s October outside, but it’s as sunny and as quiet as if it were summer.’

As follows from the implicative hierarchy in (5), adjectives corresponding to these predicatives, even in their short form, cannot co-occur with the datives as well: (7)  Но голос его был по-прежнему тих (*мне). (Ю. О. Домбровский. Хранитель древностей, часть 2 (1964)) ‘But his voice was still quiet.’

.  Although this hierarchy is proposed based on yes-or-no parameters (grammaticality and occurrence in the RNC), it is also supported by quantitative data: for the majority of stems items higher on the hierarchy show comparable or higher ratios of uses with overt datives than items lower on the hierarchy. For example, the predicative тесно ‘X is cramped’ is used with a dative argument in some 50% of occurrences, whereas for the corresponding short form adjective the ratio is around 10% and for the long form adjective it is below 0,5%. See also the data for приятно ‘pleasant’ in Table 2.

 Sergey Say

Predicatives of this type are not discussed below, as the study is devoted to the nature of the dative argument. It may be noted in passing, however, that such predicatives mostly denote properties of the setting (пустынно ‘it is deserted’), including atmospheric states (солнечно ‘it is sunny’) that are viewed as objective, independent of any sentient observer. Type 2 consists of predicatives that can co-occur with dative NPs, but correspond to adjectives that cannot. It will be shown below that this discrepancy can manifest itself in two different semantic-syntactic patterns, so that Type 2 will be further divided into two subtypes. These subtypes are conventionally labeled as subtype холодно ‘it is cold/X is cold’ (see Section 4) and subtype грустно ‘it is sad/X is sad’ (see Section 5). Predicatives of type 3 can take dative arguments (8) and are related to ­adjectives that can also take dative arguments when used in the short form (9): (8)  Я говорю, нам тесно, нас 6 человек в комнате. («Отечественные записки», 2003) ‘I say, we are cramped, we are six persons living in one room.’ (9)  Наш уголок нам никогда не тесен. (Людмила Гурченко. Аплодисменты (1994–2003)) ‘Our corner never seems too small for us.’

In the 100-token sample of contemporary uses of short forms тесен or тесна there were 11 examples that contained an overt dative Experiencer as in (9). The long form of the adjective (тесный ‘cramped’, ‘small’, ‘tight’) was not registered in co-occurrence with such NPs in either of its two 100-token samples. Similar discrepancies were observed for тяжело ‘it is hard/heavy’, скучно ‘it is boring’, странно ‘it is strange’ and corresponding adjectives. This fact allowed me to (conventionally, see immediately below) classify these items as belonging to Type 3. Predicatives of Type 4 can also take dative arguments (like those in Types 2 and 3) and are related to adjectives that can also take dative arguments in both short and long form, cf. приятно ‘it is pleasant’, известно ‘it is well-known’. However, even in this group the frequency of co-occurrence with the dative is lower for long adjectives than for the other two types of forms. It might be conjectured then that there is no strict boundary between types 3 and 4: this distinction was conventionally based on counts from small samples of examples, but it can be somewhat blurred if larger counts are taken into account. Indeed, although тесно ‘X is cramped’ was classified as belonging to type 3 based on the small working sample, one can infrequently attest examples like (10): (10)  Тот … неуклюже развернувшись в тесной ему канцелярии … двинулся потихоньку к дверям. (Олег Павлов. Казенная сказка (1993)) ‘He awkwardly turned around in the office, which was too small for him, and slowly directed his steps towards the door.’



On the nature of dative arguments in Russian constructions with “predicatives” 

The distinction between types 3 and 4 will not be relevant for further discussion, because in both groups the predicatives are such that they can co-occur with dative Experiencers and are related to adjective that can do so as well. Ultimately, the distinction between types 3 and 4 is related to the syntax and semantics of adjectives (short vs. long form) and is thus beyond the scope of this paper. Thus, Sections 4 and 5 are devoted to two subtypes of predicatives classified as belonging to type 2, i.e. to predicatives that can co-occur with dative Experiencers, whereas corresponding adjectives cannot; predicatives of types 3 and 4 are analyzed together in Section 6. 4.  Subtype холодно ‘it is cold/X is cold’ Among those dative-taking predicatives that are related to adjectives not cooccurring with datives this subtype is distinguished based on the following criterion: when in their long form, the adjectives in question have a meaning that does not presuppose any Experiencer. The contrast between the predicative and the adjective is illustrated by the following example: (11) Почему ей не холодно в холодной воде? (Google) ‘Why is not she cold when swimming in cold water?’

The predicative denotes a “personal-feeling” temperature that can be caused by either external or internal conditions. The presence of a sentient Experiencer is indispensable for this meaning; moreover, it is exactly the state of the Experiencer that is expressed by these constructions.6 If overt, the Experiencer is expressed with a dative NP. By contrast, the long form adjective with the same stem describes the temperature of some objects or substance objectively, placing it lower than average on some contextually conditioned scale: холодная вода is water that is colder than usual for particular circumstances. The coldness of water does not depend on the actual presence of any Experiencer; in other words, the semantic representation of the adjective as such lacks references to any actual Experiencer. This semantic property of the long form adjective is echoed by its syntactic property, viz. by the lack of the dative argument in the syntactic structure (cf. *холодная мне вода, intended meaning ‘the water which is cold for me’). In other words, the predicatives in this subtype have a subjective meaning that is crucially different from the meaning of the long adjective. This can be easily seen in (11) where the water is described as ‘cold’, but the Experiencer is not ‘cold’ when swimming in the water.

.  For the semantic types of temperature expressions see Koptjevskaja-Tamm (2007).

 Sergey Say

In fact, the correspondence between the two readings (personal feeling vs. objective temperature) and two types of structures (predicatives vs. adjectives) is not as straightforward as outlined above. For example, those uses of predicatives in which there is no overt dative Experiencer are open to two types of interpretation and sometimes are ambiguous. Consider the following two examples: (12)  В прошлом году об эту пору уже снег шёл, если припомните, а теперь тихо, солнечно. Только что вот холодно… Градуса три мороза. (А. П. Чехов. Вишневый сад (1904)) ‘Last year at this time the snow was already falling, if you remember, and now it’s nice and sunny. Only it’s rather cold… There’s three degrees of frost’. (13) Как холодно, у меня руки закоченели. (А. П. Чехов. Вишневый сад (1904)) ‘It is freezing (= I am freezing), my hands are numb with cold.’

In (12) the reading of холодно is objective, which is implied by a series of meteorological statements in previous discourse and mentioning of objective measurement (three degrees Celsius) in the next clause. This use of the predicative does not presuppose an Experiencer. The situation is slightly more complicated in (13). There is no overt dative argument either, but the context unequivocally shows that the reading is “personal feeling”, the Experiencer being the speaker (cf. “my hands are numb with cold”). A natural way of analyzing construction in (13) is to view it as containing a predicative холодно that has a dative slot which is semantically associated with the Experiencer (cf. кому холодно? ‘who is cold?’, lit. ‘to whom is cold?’), but remains overtly unoccupied. This non-overtness of the Experiencer is due to more general principles of discourse organization in Russian, like omission of speech-act participants; for possible reasons of omitting dative arguments in constructions with predicatives, see Guiraud-Weber (1984: 175–177). Of course, there can also be ambiguous or even intermediate uses, in which it would be difficult to figure out whether there is an Experiencer within the speaker’s perspective. The following generalizations can be thus proposed for this pair. The adjectival stem as such denotes an objective property of external reality which is related to human senses but does not presuppose any actual instance of sensation; the dative slot does not belong to the argument structure of the adjective (the long form adjective never has dative dependents). The corresponding predicative can be used in two constructions. The one that lacks the dative slot is similar to the adjectival construction, the difference being that the objective property is predicated of the setting in general (12), rather than of a particular object or substance. The one that does have a dative slot (11, 13) differs crucially: it describes actual sensation on the part of a sentient being, namely, the kind of sensation that is ­similar to the sensation which humans normally experience when placed in an environment with the corresponding objective property. When overtly expressed, the referent associated



On the nature of dative arguments in Russian constructions with “predicatives” 

with the dative slot is interpreted as the bearer of actual sensation. Like many other referents in Russian syntax, this referent can also remain implicit due to ellipsis.7 Similar relationship in the semantic and syntactic properties of predicatives vis-à-vis corresponding adjectives is observed for other units belonging to this subtype. Most adjectives in this type denote physical properties of objects or the environment in general, especially in terms of temperature or illumination, cf. темный ‘dark’, прохладный ‘cool’, светлый ‘light’, душный ‘stifling, stuffy’, теплый ‘warm’. Simplifying somewhat, the predicatives in this group are ambiguous: their interpretation is related to whether the dative argument slot is or is not present. It is crucial that this type of semantic correspondence is productive to the extent that the expected kind of subjective sensation is conceivable as a natural reaction to the physical property in question. The following example is illustrative in this respect: (14) Я кликаю на фото – все равно мне мелко. (Google) ‘I click the picture (on the Internet, – S.S.), but still I can’t see it properly, because it is too small’.

In this sentence the speaker describes not only the objective size of a particular object, but also the uncomfortable sensation that s/he gets when trying in vain to scrutinize closely a small-size picture. On the interpretative level, it would be uneconomical and counterintuitive to posit transparent and productive pairs of lexemes such as e.g. холодно1 and холодно2, so that the former is only used without the dative and describes a physical state, whereas the latter has a dative slot and denotes an actual sensation that is triggered by internal or external factors. It is more attractive to assume that the

.  Interestingly, холодный ‘cold’ and other long adjectives from this group sometimes are used with arguments that denote sentient beings, as in the following example: (i)  Если невская вода слишком холодна для сих интересных чужестранцев…  (Ф. М. Достоевский. Крокодил (1865)) ‘If the water in the Neva river is too cold for these curious foreigners…’ However, the object in the prepositional phrase headed by для ‘for’ cannot be interpreted as a usual ‘Experiencer’; rather, it corresponds to what can be called ‘evaluating subject’; cf. the notion of ориентир оценки introduced in Serdobol’skaja and Toldova (2005). Indeed, in (i) the “foreigners” are described as mentally evaluating the temperature of the water as too low (for swimming) rather than actually experiencing the sensation of being cold when swimming. By contrast, in (11) actually being in the water is presupposed.

 Sergey Say

“dative + predicative” construction is endowed with a meaning of its own, which involves an actual experiential state. The dative slot of the construction is directly associated with the Experiencer. This construction can host adjectival stems to the extent that they are semantically compatible with the meaning of the construction, even though the adjectives as such do not describe an experiential event and do not presuppose any Experiencer.8 The question of the nature of the aforementioned compatibility needs further research. At this point it can be prematurely noticed that those adjectival stems which describe such objective properties of objects that can naturally trigger particular types of experiential states (e.g. ‘cold’ or ‘dark’) are more compatible with the constructional meaning than those adjectival stems that do not have this property (e.g. presumably there is no special experiential state that is triggered by sunny weather; hence, whereas predicative солнечно ‘it is sunny’ exists, it is poorly compatible with the constructional meaning, cf. ?мне солнечно, although see footnote 8). In short, predicatives of this group are best analyzed within a constructional approach. 5.  Subtype грустно ‘it is sad/X is sad’ This is the other kind of predicatives that differ from corresponding adjectives in that they do co-occur with dative Experiencers (15), whereas the adjectives do not (16): (15) Ему было грустно одному идти через лес.  (Сергей Козлов. Правда, мы будем всегда? (1969–1981)) ‘He felt sad walking alone in the woods’. (16) Значит, жизнь далеко еще не так грустна (*мне).  (Андрей Ростовский. По законам волчьей стаи (2000)) ‘This means that life is not that sad (*to me)’.

However, the reason for this discrepancy is different from that described for the previous subtype of predicatives: the meaning of the very stem in грустный ‘sad’

.  In fact, this model is so productive that sometimes predicatives that denote physical properties of the environment and that were classified as not co-occurring with datives (Type 1 above) are used in this structure for poetic or metaphorical reasons: (ii)  Дождь в Риме промочил насквозь все мои бумажные пакеты с покупками, но все равно по настроению мне было солнечно. (Google) ‘  The rain in Rome soaked all my paper bags, but nevertheless in terms of mood I felt sunny’.



On the nature of dative arguments in Russian constructions with “predicatives” 

requires a sentient participant. Indeed, various types of mood (sad, cheerful, calm, etc.) are properties of humans and humans only. When used attributively, ­adjectives of this group can take sentient participants as their heads (грустная девушка ‘a sad girl’), when used predicatively, they can predicate of those sentient participants (девушка была грустна ‘The girl was sad’). In other words, Experiencer is an indispensable part of the semantic structure of these adjectives. Thus, the dative argument of the predicative corresponds to the internal argument of the adjective. The group of predicatives that are related to adjectives in this fashion is relatively small. Along with грустно, it chiefly encompasses a handful of other predicatives denoting emotional states: весело ‘cheerful’, спокойно ‘calm’, радостно ‘joyful’, тоскливо ‘sad, melancholic, dreary’, одиноко ‘lonely’ (however, голодно ‘hungry’ also belongs to this group). In this group, too, there is a correlation between the type of construction and the distinction between subjective and objective components in the semantics: the dative-oriented construction (construction with the predicative) correlates with a more subjective reading. This point can be clearly illustrated by the pair одинокий and одиноко ‘lonely’. The adjective одинокий can denote feeling lonely, or just living alone (e.g. одинокая женщина can refer to an unmarried woman, not necessarily implying that she feels lonely), or both. The predicative, however, always describes a particular emotional state, the feeling of loneliness: (17) Гоголю неприютно в лицее, одиноко среди сверстников, он часто переживает приступы тяжелой скуки и тоски. (Александр Воронский. Гоголь (1934)) ‘Gogol feels desolate at the lyceum, he feels lonely among his peers, he often endures attacks of devastating boredom and anguish.’

This class of predicatives is lexically limited: many adjectives that also denote internal emotional states of human beings (e.g. довольный ‘content’, угрюмый ‘­sullen’, несчастный ‘happy’) do not have corresponding predicatives of this kind: (18) *Ему было довольно/угрюмо/несчастно.  Intended: ‘He was content/in a sullen mood/happy’.

The next property of the predicatives at issue is that they convey a meaning that is somewhat different from that of corresponding adjectives, which can be seen if one compares pairs of quasi-synonymous construction like он весел – ему весело, roughly ‘he is cheerful’ – ‘he feels cheerful’. This problem is extensively studied in the literature; cf. discussion in Guiraud-Weber (1984: 173–174). It has been argued, for instance, that constructions with predicatives of this kind always denote an internal state of the Experiencer regardless of whether it is noticeable to others, whereas adjectives denote external manifestations of these states

 Sergey Say

(­Zolotova 1987: 684; Kustova 2002: 18). Predicative-based constructions of this class also belong to a larger family of Russian constructions with dative arguments that are quasi-synonymous to constructions with canonical subjects (for potential semantic invariant in the meaning of dative-oriented impersonal constructions see Guiraud-Weber 1984: 173 ff.). Among such constructions is, e.g. the so-called “modal-deagentive” reflexive construction: мне не спится ‘I can’t get asleep, I am not in the proper state for sleeping’. We are now in a position to make a generalization with respect to the nature of these predicatives: (i) they have the same sets of semantic arguments as ­corresponding adjectives; (ii) there is a difference in the syntactic status of the Experiencer participant: the internal argument (the head noun or subject in ­constructions with long resp. short form adjectives) in structures with adjectives vs. dative argument in constructions with predicatives; (iii) this syntactic difference reflects a noticeable semantic change; (iv) the pattern is severely restricted lexically. The combination of properties in (ii–iv) suggests that the relation of adjectives to predicatives is best analyzed as a valency-changing operation, namely, a derivation that demotes the subject into an oblique (dative) position.9

.  In principle, one could also consider the possibility of comparing constructions like мальчику грустно ‘the boy feels sad’ with such uses of the adjective грустный where this property is metonymically predicated of some objects, circumstances, places (грустные обстоятельства ‘sad circumstances’) etc. rather than persons (грустный мальчик ‘a sad boy’). Such an approach could be tempting in some respects, first of all, because it would allow us to abandon the idea of demotion of Experiencer to the dative position, which is unique to this subtype of predicatives within the approach advocated in the main text. However, such an alternative approach has to be refuted for a number of reasons that can be briefly summarized as follows. (1) This type of metonymy is more typical of adjectives used attributively (cf. грустный день ‘a sad day’), than of predicative short adjectives (?день был грустен ‘the day was sad’). (2) Short forms of adjectives never co-occur with dative Experiencers (*она мне грустна). (3) The two types of structure always have different meanings, even when the subject in the construction with the short form of adjective is a semi-dummy pronoun, like это ‘this’ or всë  ‘everything’. Consider the following example that contains a short form adjective грустно that agrees in gender and number with the subject (одно ‘one thing’): (iii) Одно было грустно (*ей) – нет Яшина. («Наш современник», 2004.04.15) ‘There was one thing that made her sad – Jashin was not there’ As in many other metonymic uses, this adjective means that a particular state of affairs (absence of Jashin in this case) can cause someone’s sadness, not that there was an Experiencer who was actually in a sad mood; this causal metonymic shift is a typical model of adjectival polysemy (Kustova 2002) but is not characteristic of predicatives.



On the nature of dative arguments in Russian constructions with “predicatives” 

6.  Type приятно ‘it is pleasant’ The last large group contains predicatives that are related to adjectives which semantically select Experiencers and which put them into the dative slot, at least in some of their uses (for the sake of simplicity, we will now disregard the difference between adjectives that take and do not take a dative argument when used attributively, see Section 3). In other words, unlike predicatives discussed in the two previous sections, these predicatives (20) are parallel to corresponding adjectives (19) in terms of co-occurrence with and interpretation of the dative argument: (19)  Они стали перечислять приятные им названия городских улиц, ресторанов, кино. (Василий Гроссман. Жизнь и судьба, часть 2 (1960))  ‘They started to recall the names of streets, restaurants and cinemas, that sounded pleasant for them.’ (20)  Признаюсь, мне было приятно слышать такие комплименты. (И. А. Архипова. Музыка жизни (1996)) ‘I have to confess that it was pleasant for me to hear such compliments’.

The semantics of both the adjective приятный ‘pleasant’ and the predicative приятно ‘it is pleasant’ contains obligatory reference to a sentient participant. This has an important consequence for the interpretation of such predicatives: even if the dative slot is left unoccupied, Experiencer is obligatorily interpreted as either contextually recoverable given NP or as a generalized NP, as in (21): (21) Приятно после долгой ходьбы и глубокого сна лежать неподвижно на сене. (И. С. Тургенев. Хорь и Калиныч (1847)) ‘After a long walk or a deep sleep it is pleasant to lie motionlessly on hay.’

In (21) the presence of an Experiencer in the semantic structure is further shown by its obligatory co-reference with the non-overt subject of the infinitive (лежать) and of the nominalization (ходьбы). Thus, these predicatives cannot yield an “objective” reading: unlike the personal feeling of temperature comfort (see S­ ection 4), the feeling of pleasure as such does not have any measurable ­objective counterpart. This property (semantic obligatoriness of a sentient participant) is r­ eminiscent of the previous group discussed, namely, of predicatives like грустно (see ­Section  5). The crucial difference is that adjectives like грустный ‘sad’ can be used with Experiencers in their head noun position. By contrast, adjectives like приятный ‘pleasant’ cannot be used in such a way; indeed, приятный человек can only mean ‘a pleasant person’, that is, a person who can, roughly speaking, cause positive reaction on the part of Experiencers and not the person who experiences something pleasant him- or herself. In other words, the long form adjectives

 Sergey Say

related to ­predicatives from this group are modifiers to head nouns that are interpreted as Stimuli, rather than as Experiencers. These predicatives of the приятно-type are semantically heterogeneous; they include predicatives describing emotional (приятно ‘it is pleasant’, обидно ‘it hurts X’, безразлично ‘X is indifferent’), mental (интересно ‘it is interesting’, понятно ‘it is clear/understandable’), deontic modal (нужно ‘it is necessary’), and other (полезно ‘it is profitable/useful for X’, скучно ‘it is boring/X is bored’) states, as well as two perception predicatives (видно ‘it is visible = X can see’ and слышно ‘it is audible = X can hear’), which are exceptional in that they are basically the only predicatives that can take direct objects, cf. мне слышно музыку ‘I can hear music’. It is crucial that constructions with predicatives of this group are semantically very close to constructions with short forms of adjectives. Their main deviation is that the Stimulus is syntactically realized as an infinitive, subordinate clause, adverb, etc., that is, as a constituent that has no case feature and is unable to trigger agreement, see Bonč-Osmolovskaja (2003) for the typology of Stimuli in constructions with predicatives. Thus, structures like мне было приятно здесь/гулять в парке/когда она былa рядом ‘I liked it here/walking in the park/when she was beside me’ are quite similar to e.g. мне было приятно твое общество ‘I liked your companionship’. The former are usually treated as (impersonal) constructions with predicatives, whereas the latter as constructions with agreeing short adjectives. In fact, however, they are related to each other in exactly the same way as, e.g. constructions like мне нравилось здесь/гулять в парке/когда она рядом are related to мне нравилось твое общество (roughly same meanings); needless to say, nobody assumes two lexical entries for agreeing and impersonal нравилось. Now we see that the simplest way to analyze these predicatives is to view them as syntactically determined uses of the “default” (neuter) forms of corresponding short (predicative) adjectives. The dative argument is subcategorized for by the adjectival stem as such, i.e. lexically determined. For this group, the differences between “predicatives” vs. (other) short forms of adjectives could be accounted for by very general rules of syntax that are independently motivated (e.g. patterns of default agreement). In short, for this group (and for this group only!) the analysis could easily follow such proposals as in Garde (1982) and Babby (1974). Although diverging in many respects, these scholars make an important common point: both view “predicatives” as a type of use of adjectives that can be deduced from the interaction of semantic and syntactic rules. However, even in this group of adjectival stems predicatives seem to favour the use of dative arguments if compared to the corresponding adjectives. This bias is largely due to the fact that adjectives of this group often co-occur with sentient arguments coded by prepositional phrases with для ‘for’:



On the nature of dative arguments in Russian constructions with “predicatives” 

(22)  И эта неясность для инвесторов не очень приятна. («Владивосток», 2003.09.29) ‘This uncertainty is not particularly pleasant for investors.’

Whether dative NPs and prepositional phrases with для ‘for’ express the same semantic role is a matter of debate, cf. discussion in, e.g. Serdobol’skaja & Toldova (2005). On the one hand, they are mutually exclusive, that is, an adjective can cooccur with a dative NP or with a prepositional phrase with для ‘for’, but never with both. This seems to be an argument for viewing them as morphologically distinct manifestations of the same semantic role. In all probability the two (and – for some adjectival stems – maybe even more) morphological means for coding this argument used to be non-differentiated (or weakly differentiated) semantically in the 18th century. However, it is argued by Guiraud-Weber that this subsystem ­underwent reorganization during the 19th century so that the dative came to be closely associated with the Experiencer “au sein de la construction impersonelle à proprement parler, c’est-à-dire celle qui exprime les états, les sensations et les sentiments d’un être vivant” (Guiraud-Weber 2003: 71), whereas other coding devices, including prepositional phrases with для ‘for’ in the first place, became related to evaluative meanings. Guiraud-Weber describes this tendency for the development of impersonal uses only, but it appears that the RNC provides cross-construction quantitative data that further support her idea. The relevant data are in Table 2. Table 2.  Coding the sentient argument for various uses of приятно ‘pleasant’ 1979

Dative

Other

Zero

Dative

Other

Zero

Long form

 1

 4

95

 0

7

93

Short form

32

10

58

37

4

59

Predicative

58

 8

34

36

0

64

It can be seen that in the earlier subcorpus both dative and other peripheral coding options were attested (though not equally distributed) for sentient arguments in all three types of constructions. This can be illustrated by (23) where the predicative co-occurs with the prepositional phrase headed by для: (23) … для меня было очень приятно видеть на сыне моем тот же самый крест и в том же самом крае, где я получил оный 48 лет тому назад. (М. С. Воронцов. Кавказские письма М. С. Воронцова к А. П. Ермолову (1845–1855)) Lit. ‘For me it was very pleasant to see my son wearing the same cross and wearing it in the same lieu where I was awarded such a cross 48 years before.’

 Sergey Say

In modern texts there is a more pronounced bias in the distribution of patterns for coding the sentient participant. Indeed, predicatives no longer co-occur with forms other than the dative; the example in (23) would sound somewhat awkward in modern Russian and the dative (мне) would be definitely preferable. On the contrary, when used attributively, приятный ‘pleasant’ can take prepositional phrases with для (24), whereas the dative coding of the sentient participant is disfavoured (see Table 2), although it is not altogether impossible (25). (24)  Двадцатиминутный приятный для обеих сторон разговор закончился осмотром его трубок. (Сергей Штерн. Ниже уровня моря // «Звезда», 2003) ‘The 20-minute conversation, enjoyable for both parties, ended with an examination of his pipes.’ (25)  С другой стороны – так же я раздражаю приятных мне людей меньшего дохода. (Наши дети: Дошколята и младшие школьники (форум) (2005)) ‘On the other hand, I irritate in the same way those people who are pleasant for me, but earn less than me.’

This bias nicely fits Guiraud-Weber’s semantic interpretation: the semantic core of the constructions with predicatives is the meaning of actual sensation, which is in its turn associated with the dative. Similar distributions and diachronic trends are observed for other adjectival stems from this group.10 These findings make it possible to enrich the interpretation proposed above. Indeed, as stated above, adjectival stems from this group require Experiencer in their role structure and this argument can be coded with a dative NP. However, the diachronic evidence discussed above suggests that even in this group predicatives are not just a syntactic subtype of short forms. Constructions with predicatives favour the choice of the dative option for coding the sentient participant and this fact nicely fits the semantic peculiarities of predicatives. In other words, in structures like (20) the Experiencer is both subcategorized for by the adjectival stem and belongs to the constructional pattern. An essentially similar analysis of some uses of predicatives, although in quite different terms, was suggested in Švedova (1978/2005: 283–284). On a more general level, such double motivation of arguments (lexeme-based and construction-based) is found in some versions of Construction grammar, cf.: “… the argument must satisfy the specifications of both the argument role of the construction and the participant role of the verb. That is, the argument role of the construction may be “fused” with a participant role of the verb” (Goldberg 2005: 24). .  This is to say, for other stems of this group co-occurrence with для-phrases is more typical of (long form) adjectives than of predicatives (with the exception of видный ‘visible’ and слышный ‘audible’, as these two adjectival stems generally co-occur with для-phrases very rarely).



On the nature of dative arguments in Russian constructions with “predicatives” 

7.  Discussion and conclusion 1. In the present paper, a two-level classification of Russian predicatives is ­proposed. On the first level of classification they are grouped based on whether they and their corresponding adjectives in short and long forms are able to co-occur with dative NPs denoting sentient participants. A hierarchy (predicatives > short forms of adjectives > long forms of adjectives) is put forward, which reflects decreasing ability to co-occur with dative Experiencers. On the second level of classification predicatives are grouped based on the semantic relationship between their dative arguments and the argument structure of corresponding adjectives. The following types of these relationships have been distinguished: (i) the predicative’s dative argument is not related to any of the adjective’s arguments (subtype холодно, see Section 4); (ii) the predicative’s dative argument is related to the adjective’s internal argument (subtype грустно, see Section 5); (iii) the predicative’s dative argument is related to the adjective’s dative argument (type приятно, see Section 6). 2. When analyzed in isolation, various types of predicatives call for different types of analyses. In particular, the properties of predicatives from the subtype холодно ‘it is cold/someone is cold’ (Section 4) are best captured by a constructional analysis, in which it is the construction as such that accounts for the presence of the dative argument and this argument is linked to a component in the constructional meaning. The adjectival stem has to be loosely compatible with the meaning of the construction but does not semantically select any Experiencer. The combination of the constructional and lexical meanings yields an overall interpretation; e.g. for [Х-DAT холодно]: Х experiences the feeling that is normally triggered by the environment of low temperature. Predicatives of the subtype грустно ‘it is sad/X is sad’ (Section 5) are best analyzed as separate lexemes that are derived through a valency-changing operation from corresponding adjectives (the subject is demoted to the dative slot). Finally, predicatives of the type приятно ‘it is pleasant’ (Section 6) can be generally analyzed as inflectional forms of corresponding adjectival ­lexemes. Their use is deducible from very general syntactic rules of the Russian grammar.11

.  One also has to keep in mind that there are predicatives that do not correspond to any adjective (cf. example 2) and these have to be posited as independent items in the lexicon.

 Sergey Say

3. This comparison of various possible approaches and demonstration of their relative suitability for various subsets of data was not aimed at arguing for any single one of them. Nor was it intended to claim that the subtypes of predicatives represent syntactically unrelated phenomena and to propose a very complex syntactic model that would encompass all the competing mechanisms. On the contrary, it was aimed at showing the limitations of possible individual explanations that are so usual in the domain of studying argument structures of related constructions: no one of them alone seems to be adequate for the whole range of data, only for a subset. This is a pessimistic methodological outcome of this study. Needless to say, once this pessimistic conclusion is arrived at, one starts seeking a possible way out. At this point such a solution can be suggested in a very tentative form only. Its essence is the following. The experiential predicative construction encompasses a dative slot that is directly related to the Experiencer in terms of argument-linking. Informally speaking, this constructional pattern is “blind” to the argument structure of the corresponding adjective. If the adjective itself is subcategorized for an oblique (dative or otherwise) Experiencer, the predicative construction preserves it (and “chooses” dative coding when other options are available for the adjective). If the adjective semantically selects Experiencers but normally puts them into internal argument position, then the predicative construction demotes the Experiencer into the dative slot. If the adjective itself does not select the Experiencer, then this argument is coerced by the construction. In still more informal terms, what is important for the construction is the product rather than the source (cf. the notion of product-oriented schemas in morphology (Bybee 2001: 126−129)). 4. On a methodological level, this case study raises the issue of how our theoretical models are dependent upon our data. In terms of efficiency and intuitive adequacy, the competing theoretical models (e.g. constructional, semanticcompositional, head-driven) might thus turn out to be complementary rather that mutually exclusive. In other words, it might be more sensible to seek the ways of combining these models, rather than arguing for the choice of the best among them.

References Babby, Leonard H. 1974. Towards a formal theory of “parts of speech”. In Slavic Transformational Syntax, Richard D. Brecht & Catherine V. Chvany (eds), 151–181. Ann Arbor MI: University of Michigan. Bonč-Osmolovskaja, Anastasija A. 2003. Konstrukcii s dativnym sub”ektom v russkom jazyke [Dissertacija kandidata filologičeskix nauk]. Moskva: MGU.



On the nature of dative arguments in Russian constructions with “predicatives” 

Bybee, Joan L. 2001. Phonology and Language Use [Cambridge studies in linguistics 94]. ­Cambridge: CUP. Cimmerling, Anton V. 2003. Predicativy i kačestvennye narečija: Klassy slov i napravlenija derivacii. In Rusistika na poroge XXI veka: Problemy i perspektivy, 54–59. Moskva: IRJa RAN. Garde, Paul. 1982. Russkoe “kakovo”. Revue des Etudes Slaves 54 (1–2): 79–86. Goldberg, Adele. 2005. Argument realization: The role of constructions, lexical semantics and discourse factors. In Cognitive Grounding and Theoretical Extensions [Constructional Approaches to Language 3], Jan-Ola Östman & Miriam Fried (eds), 17–43. Amsterdam: John Benjamins. Guiraud-Weber, Marguerite. 1984. Les propositions sans nominatif en russe moderne. Paris: ­Institut d’Etudes Slaves. Guiraud-Weber, Marguerite. 2003. Les phrases impersonnelles en russe au XXe siècle. La revue russe 23: 65–77. Koptjevskaja-Tamm, Мaria. 2007. Guidelines for collecting linguistic expressions for temperature concepts. Version 1. Ms. Kustova, Galina I. 2002. O tipax proizvodnyx značenij slov s èksperiencial’noj semantikoj. Voprosy jazykoznanija 2: 16–34. Letučij, Aleksandr B. 2012. O nekotoryx svojstvax sentencial’nyx aktantov v russkom jazyke. Voprosy jazykoznanija 5: 57–87. Peškovskij, Aleksandr M. 1956/2001. Russkij sintaksis v naučnom osveščenii. Moskva: Jazyki ­slavjanskoj kul’tury. Ščerba, Lev V. 1928/2004. O častjax reči v russkom jazyke. In Ščerba, L.V. Jazykovaja sistema i rečevaja dejatel’nost’, 77–100. Moskva: Editorial URSS. Serdobol’skaja, N.V. & Toldova, S.Ju. 2005. Ocenočnye predikaty: tip ocenki i sintaksis ­konstrukcii. In Komp’juternaja lingvistika i intellektual’nye texnologii. Materialy meždunarodnoj konferencii «Dialog’2005», Zvenigorod, 1–7 ijunja 2005, Irina M. Kobozeva, Aleksandr S. Narin’jani, Vladimir P. Selegej (eds), 436–443. Moskva. Švedova, Natalia Ju. 1978/2005. Dixotomija “prislovnye – neprislovnye padeži” v ejo otnošenii k kategorijam semantičeskoj struktury predloženija. In Švedova, N.Ju. Russkij jazyk. ­Izbrannye raboty, 270–284. Moskva: Jazyki slavjanskoj kul’tury. Švedova, Natalia Ju. (ed.). 1980. Russkaja grammatika. T. I. Moskva: Nauka. Timberlake, Alan. 2004. A Reference Grammar of Russian. Cambridge: CUP. Vinogradov, Viktor V. 1947. Russkij jazyk (grammatičeskoe učenie o slove). Moskva–Leningrad: Učpedgiz. Vinogradov, Viktor V., Istrina, E.S. & Barxudarov, Stepan G. (eds). 1953. Grammatika russkogo jazyka, Tom I: Fonetika i morfologija. Moskva: AN SSSR. Zolotova, Galina A. 1987. O distinktivnyx značenijax sinonimičeskix konstrukcij. Revue des Etudes Slaves 59(3): 681–686.

Russian Adversity Impersonals and split ergativity* Katrin Schlund

Slavisches Institut, Heidelberg University, Germany Most of the numerous studies on Russian Adversity Impersonals (e.g. lodku uneslo vetrom) focus on the formal make-up of these constructions and/or on their semantics. This article aims at describing Russian Adversity Impersonals from a typological point of view by linking them with ergativity. It shows that the conditions under which Adversity Impersonals occur correspond in many ways to the factors determining case assignment in languages with split ergativity. To account for the occurrence of Adversity Impersonals in Slavic, other characteristics of the Slavic languages disposing of this type of impersonal construction are discussed and compared to characteristics of languages with ergative splits. Keywords:  Adversity Impersonals; Russian; ergativity; split ergativity; relational typology; East Slavic; have- and be-languages

1.  Introduction While a lot has been done to account for the syntactic structure and semantic nature of Adversity Impersonals (henceforth AI) in countless contributions, far less has been said about AI from a typological point of view. The aim of this study is to investigate AI in terms of relational typology, on the basis of the distinction between nominative-accusative and ergative-absolutive languages.

* A number of people have contributed to this paper. I would like to thank Jadranka Gvozdanović (Heidelberg), Marguerite Guiraud-Weber (Aix-en-Provence) and Hélène Włodarczyk (Paris) for fruitful discussions and helpful advice. I also owe great thanks to Katarzyna Janic (Lyon), Igor Dreer (Negev), Oleg Chinkarouk (Paris), Robert Grošelj ­ (­Ljubljana), David Jardine (Mannheim), Lioubov Arapova, Katharina Kunz, and Aleš Půda (all Heidelberg) for their native speaker judgments, numerous examples, and patient explanations.

 Katrin Schlund

The following examples are given to illustrate what is referred to by the term AI: (1) a. Лодку унесло ветром. boat:acc carried-away:neut wind:inst ‘The boat was carried away by the wind’. (Mustajoki & Kopotev 2005: 2) b. Его ранило пулей. he:acc wounded:neut bullet:inst ‘He was wounded by a bullet.’

(Sulejmanova 1999: 63)

c. Осколком стекла девушке порезало ногу. shard:inst glass:gen girl:dat cut:neut leg:acc ‘A girl got her leg cut by a shard of glass.’ (Mustajoki & Kopotev 2005: 4)

As the above examples show, AI obligatorily consist of an active neuter past participle functioning as past tense and an affected entity in the accusative. In the prototypical AI, the cause of the event is also given and is assigned instrumental case. Sometimes, a dative experiencer can be added (1c). The term “Adversity Impersonal” was first introduced by Babby (1994). Other labels include “impersonal transitives with instrumental” (Green 1980: 10), “elemental construction” (stixijnaja konstrukcija, e.g. Mustajoki & Kopotev 2005), “uncontrolled event impersonals” (Davies 2005), “transitive impersonals” (Lavine 2010), and “unaccusative accusatives” (Lavine & Freidin 2002). Some of these alternative labels highlight the prototypical semantics of AI: an uncontrolled, inanimate, often elemental force negatively affects an animate or inanimate patient. Although some scholars have provided non-adverse examples of AI (e.g. Green 1980: 128; Davies 2005: 90), I will stick to the term “Adversity Impersonals” here as the negative uses of AI clearly prevail and as Babby’s label seems most widely used to refer to the phenomenon in question. Although AI are most typical of Russian, formal equivalents of the Russian type can be found in other East Slavic languages and Polish: (2) a. Ukr. У Луганську сильним вітром віднесло in Lugansk strong:inst wind:inst drifted-away:neut

коляску з однорічною дитиною. stroller:acc with one-year-old:inst child:inst

‘In Lugansk, a stroller with a one-year-old child was drifted away by strong wind.’1

.  Source: http://www.ukr.net/news/u_lugansku_silnim_v_trom_v_dneslo_koljasku_z_odnor_ chnoju_ditinoju-12917788-1.html, accessed 15th September, 2012.



Russian Adversity Impersonals and split ergativity 

b. Bel. Яго абдало ветрам. he:acc carried-away:neut wind:inst ‘He was carried away by the wind.’  Citation from Vasil Bykaǔ’s ‘Alpine Ballad’ (Al’pijskaja balada)2 c. Pl. Drogę zasypało śniegiem. road:acc covered:neut snow:inst ‘The road was covered with snow.’

(Włodarczyk 1996: 182)

These examples show that AI are not an exclusively Russian phenomenon. Although this paper is dedicated to exploring AI in Russian, we shall keep their incidence in other languages in mind and will return to this fact in the final part of the paper. Although AI have been of great interest in Russian linguistics for decades, it is hard to find reliable information about their historical development. Hofmann (1934) addresses two conflicting views on the origin of Russian AI. While some scholars assume that Russian AI are a continuation of an old Indo-European construction, others claim that they are actually a quite recent Russian innovation. Hofmann concludes that neither view is correct. He claims that Russian AI can indeed be attested in the earliest monuments of East Slavic literacy but that they were very rare at the time and cannot be traced back to Common Slavic or even Balto-Slavic (Hofmann 1934: 218f.). Hofmann gives some instances of AI from the First Chronicle of Pskov and the Second Chronicle of Novgorod, pointing out that in both sources the vernacular is used at least occasionally. Given that AI cannot be attested in clerical or secular documents without any reference to the vernacular of the time, Hofmann deduces that AI were originally a part of vernacular speech which eventually entered the standard language (Hofmann 1934: 210, 212). However, in contrast to his findings, a more recent study assumes that there is a Proto-Slavic origin of AI (Khodova 1958: 151). Mrázek (1956: 26) claims that AI have become more frequently used since the fifteenth century. Guiraud-Weber (1984: 285) observes that AI began to flourish in the sixteenth and seventeenth centuries, which is exactly the time the instrumental agent in the passive voice had begun to supplant its ot + genitive competitor. However, the use of the instrumental was not a preliminary for the emergence of AI. Hofmann (1934: 215) gives examples in which ot + genitive is used instead of the instrumental in AI as well. Galkina-Fedoruk states that AI (1958: 149) and impersonal constructions in general (1958: 151) are still gaining ground against their personal counterparts but she does not give any empirical data supporting her assumption. Green (1980: 25) also states that AI are “an extremely productive

.  I owe thanks to Igor Dreer for pointing out this example to me.

 Katrin Schlund

construction whose use seems to be growing […]”. Although these statements still need empirical corroboration, AI are a completely “healthy” construction in the sense that they are accepted by native speakers and regularly used in the media and in fiction (Mustajoki 2006: 54; Mustajoki & Kopotev 2005). 2.  Selection mechanisms of AI and their parallels with ergativity Relational typology starts from the assumption that there are “three core ­semantico-syntactic relations” (Dixon 1979: 61) that a language needs to express: a. The subject of a transitive clause (A) b. The subject of an intransitive clause (S) c. The object of a transitive clause (O) Languages following the ergative-absolutive pattern treat O and S formally alike but mark A differently. Nominative-accusative languages, on the other hand, unify A and S and distinguish them from O. For our purposes, it is also important to note that the ergative case, denoting A, is morphologically marked, and that the absolutive case, denoting both O and S, is morphologically unmarked (cf. Comrie 1978: 329; Dixon 1979: 60f.; Ura 2006: 113). Russian, like the other Slavic languages, belongs in principle to the nominative-accusative type. However, the distinction between ergative-absolutive and nominative-accusative languages is not absolute: languages can occupy different points on a continuum with ergativity at one end and nominativity at the other. It thus makes sense to analyze an allegedly nominative-accusative language in terms of “hidden ergativity”. In languages with split ergativity, ergative-absolutive case marking is dependent on certain morphological, semantic and/or pragmatic factors. This means that ergativity in such languages is restricted to a certain set of contexts while the nominative-accusative pattern prevails in others. Linking AI with ergativity is actually not new. Pontoppidan-Sjövall (1968) and Daly (1977, referred to in Green 1980: 145f.)3 come up with an interpretation of AI as ergative. Kacnel’son (1974: 73–92) also establishes a link between oblique

.  As Green unfortunately does not give the entire reference of Daly (1977), I have not been able to assess Daly’s views myself and will not discuss them here in detail. If Green’s (1980: 145–147) brief discussion of Daly is correct, Daly equates oblique subjects with ergative ­subjects, pointing out that many impersonal verbs in Russian denote emotional or psychological states and are semantically marked for lack of control; their semantic subjects are consequently denied unmarked nominative (agentive) case.



Russian Adversity Impersonals and split ergativity 

subject marking and ergativity but claims that the two phenomena are yet too ­different to be theoretically unified. Pontoppidan-Sjövall (1968: 60) directly parallels AI with ergative constructions in Georgian. Green (1980: 136–147) totally dismisses the idea of associating AI with ergativity, claiming that the data from ergative languages are incomplete and that the reasoning was “ad-hoc in the worst sense of the word” (ibid.: 144), “too sketchy” and “too vague to be of any real value […]” (ibid.: 146). When reading through Pontoppidan-Sjövall’s line of argumentation, one is indeed tempted to agree with Green. The problem with Pontoppidan-Sjövall’s account is her interpretation of the Georgian ergative construction as the conveyance of an uncontrolled event which “happens to” the patient (Pontoppidan-Sjövall 1968: 41; Green 1980: 140). This interpretation is erroneous. Georgian expresses the same event with a nominativeaccusative construction in the present and uses the ergative pattern only in the aorist. There is no ground to presume that the event is less volitional in the aorist than it is in the present tense. Pontoppidan-Sjövall’s misunderstanding of the Georgian ergative construction gives, however, no reason to rule out any connection between AI and ergativity. On the contrary, tense-aspect split ergativity in languages such as Georgian actually corroborates such a connection, as will be shown in Section 2.2. Davies (2005: 100–110) gives four features of AI which are essential to the nature of AI: de-emphasis of instigator, tense/aspect constraints, adverbial restrictions, and word order. Furthermore, AI are restricted to inanimate causes and to impersonal morphology (cf. Green 1980: 24). The following sections will discuss these six parameters and establish the extent to which they parallel traits of ­ergative languages. 2.1  De-emphasis of instigator Ergative languages de-emphasize the instigator of a transitive action (A) by assigning it an oblique case. S and O, on the other hand, receive an unmarked case and are usually in control of agreement. The criterion of “instigator defocusing” (Słoń 2007) holds not only for ergative constructions but also for impersonal constructions and passivization. In AI, A is demoted to the oblique instrumental case just like in the passive voice. O, however, is not promoted to the unmarked nominative as would be the case in the passive voice but remains in the accusative, the prototypical transitive object case in nominative-accusative languages. The treatment of O and A in AI is thus neither clearly nominative-accusative nor absolutiveergative. The marking of O corresponds to the active voice while A is treated like the instrumental agent in the passive voice. AI are thus neither active nor passive, a fact that has caused Mustajoki (2006: 53) to refer to AI as a “[…] contamination of the active and passive voices […]”.

 Katrin Schlund

The reason why A is not assigned nominative case in AI lies in its semantics. The instrumental phrase of AI does not denote a prototypical subject as established by Comrie (1981: 101), Keenan (1976), or Guiraud-Weber (2002: 286). Most importantly, a prototypical subject is not only the subject of a transitive clause but also in control of the transitive action expressed by the verb. The prototypical subject is thus human or at least animate. In AI, however, only inanimate nouns are admissible in the instrumental phrase. The instrumental thus denotes a nonprototypical agent – a cause.4 I will now compare AI to different types of ergative splits. Languages with split ergativity assign ergative case to the transitive subject only under certain circumstances. As will be shown in the following sections, the occurrence of AI is similarly restricted. 2.2  Tense and aspect constraints Many partly ergative languages display tense and/or aspect splits. Whenever a language has such a split, it is always the perfective aspect and/or past tense which yields ergative marking, while the nominative-accusative pattern is used in the non-past (cf. Trask 1979: 385). Empirically based, quantitative analyses such as Guiraud-Weber (1984) and Mustajoki and Kopotev (2005) shed useful light on the aspect-tense behaviour of AI. Both studies show that the perfective aspect clearly dominates over the imperfective aspect. In Guiraud-Weber’s data base, the perfective aspect is used in almost 80% of the examples (Guiraud-Weber 1984: 280). Mustajoki and Kopotev (2005) have found more than two thousand cases of AI, of which the vast majority use the perfective aspect. However, they have also identified more than 60 examples using the imperfective aspect: (3) a. Progressive: Когда меня уносило течением, они как-то when I:acc took-away:ipf.neut current:inst they somehow злорадно притихли. maliciously remained-silent:pl ‘When I was being taken away by the current, they remained ­maliciously silent.’ (Mustajoki & Kopotev 2005: 13) .  With reference to AI, the syntactic labels “transitive subject” (A) and “transitive object” (O) are considered mutually interchangeable with the semantic labels “cause” and “patient”. The syntactic terms will be used when the syntactic role of the constituents is being evaluated; the semantic labels will be preferred in other contexts. Of course, when discussing information structure (see Section 2.5), pragmatic labels will be introduced as well.



Russian Adversity Impersonals and split ergativity 

b. Iterative: За неделю его восстаналивали дважды, и дважды его in a-week it:acc restored:pl twice and twice it:acc сносило новым потоком грязи swept-away:ipf.neut new:inst stream:inst dirt:gen и камней. and stones:gen ‘Within a week they restored it [maybe a bridge or a house, KS] twice, and twice it was swept away by another stream of dirt and stones.’ (Mustajoki & Kopotev 2005: 13).

The two examples demonstrate that imperfective aspect is obligatory in strongly imperfective contexts such as progressivity and iterativity. Note, however, that past tense is used nevertheless. As mentioned above, the ergative pattern is also linked with perfectivity in some ergative languages. Montaut and Pilot-Raichorr (1994: 92) claim that Hindi restricts ergativity to the perfective aspect due to its resultative semantics. The authors point out that resultativity by definition focuses on the endpoint of an action. Consequently, O (the patient) will be in the focus of attention in a perfective action. The initiator of a perfective transitive action, however, is a mere source of the event and consequently receives peripheral syntactic marking. This explanation is indeed of some value for AI as well. AI express transitive, preferably resultative events, too, in which the instigator of the event, according to Montaut & Pilot-Raichorr, is naturally de-emphasized. In fact, the obvious affinity between perfectivity and ergativity could nicely be explained in these terms if perfectivity always resulted in ergativity. Of course, clearly nominativeaccusative languages and ergative languages without a tense-aspect split prove the opposite. Trask (1979: 398) gives a different answer to why tense-aspect splits occur in some languages, but not in others. He points out that in Indo-European ergative languages, which all have the tense-aspect split, the periphrastic perfect results from a passive participle in a formerly possessive construction, yielding a construction like the following example from Latin: (4) Mihi est factum. I:dat is done:pst.part.pass ‘I have done.’

(Benveniste 1960: 128)

However, this construction was abandoned in Latin and the Romance languages because of the emergence of the verb habere ‘to have’, which resulted in a new,

 Katrin Schlund

transitive possessive construction for the formation of an analytical perfect (Isačenko 1974: 73): (5) Habeo opus factum. have:1sg work:acc done:pst.part.pass Lit.: ‘I have the work done.’ > ‘I have done the work.’

From this it follows that ergativity is attached to perfectivity only in languages which do not dispose of a transitive possessive construction. The rise of a transitive possessive construction, again, is preconditioned by the existence of a verb have. Indeed, Trask (1979: 398) submits that in all languages whose ergative construction originates from a possessive, the verb have is absent. With these considerations in mind, it is remarkable that Russian is a be-­ language, too (cf. Isačenko 1974). Of course, no reanalysis of the possessive u menja + nominative case as perfect tense has occurred in standard Russian.5 Yet, cognate resultative structures are part of the colloquial standard language:6 (6) У меня работа написана. at me:gen work:nom.f write:pst.part.pass.f ‘I have my work written.’

(Maslov 1988: 80)

Although there is convincing evidence that ergativity is linked with perfectivity in languages lacking a transitive possessive construction with have, it is not clear why AI should be connected with the lack of the verb have, or at least with the lack of grammaticalization of have.7 Perhaps a missing link can be found if we consider that both the absence of have and the existence of impersonal constructions tend to correlate with rich inflectional morphology (especially noun morphology) and free word order. A transitive possessive construction yields reliable SVO word order, thus making inflectional morphology potentially dispensable. Impersonal and possessive constructions with be are, on the other hand, more convenient in inflectional languages. Free word order has indeed been connected with the emergence of impersonal constructions: ‘In languages with relatively free word order, the high frequency of a syntactic subject which becomes part of the rheme creates a favorable climate for the preservation and development of impersonal and 3rd-person-only structures.’ (Alisova 1969: 36, translation by Green 1980: 192)

.  It does, however, occur in some North Russian dialects (see also Section 3.). .  The typological relevance of Colloquial Russian in terms of “syntactic emptiness” is ­revealed extensively by Weiss (this volume). It is therefore noticeable that Colloquial Russian seems to be a pioneer with regard to other typologically relevant features as well. .  This question will be taken up again in Section 3.



Russian Adversity Impersonals and split ergativity 

It seems that the parallel between aspect-tense split ergativity and AI is not incidental, but only indirect. Perfectivity is more likely to focus on the result (i.e. the patient, or O) of a transitive action than imperfectivity, so perfectivity is connected with both aspect-tense split and AI. However, the aspect-tense split originates from a possessive construction while perfectivity in AI seems to be a natural byproduct of resultativity. 2.3  Animacy restrictions Animate nouns are not admitted in the instrumental phrases of AI: (7) *Женщину ранило медведем.  woman:acc wounded:neut. bear:inst Intended: ‘The woman was wounded by a bear.’

The restriction to inanimate agents (i.e. causes) is reminiscent of the animacy split found in many ergative languages. Certain ergative languages are split in accordance with Silverstein’s (1976) “animacy hierarchy”. The higher an agent ranks in this hierarchy, the greater the likelihood of the nominative-accusative pattern being used. For instance, some languages stick to the nominative-accusative pattern with first and second person pronouns but use ergative-absolutive marking for third person pronouns and full noun phrases. Other languages show ergative-absolutive behaviour only with inanimate transitive subjects, while the nominative-accusative scheme is used with animate transitive subjects. The parallel between the animacy split and AI is obvious. However, several scholars have pointed out that not any inanimate noun is acceptable in the instrumental phrase of AI. I will thus take a closer look at the semantic restrictions on the instrumental phrases of AI. Sal’nikov (1977) and Lavine (2010) similarly explain the semantics of the instrumental phrases of AI. Both argue that the noun in the instrumental must dispose of some “potential energy” (potencijal’naja ėnergija, Sal’nikov 1977: 282) or initiative force (Lavine 2010: 111). Sal’nikov (1977: 283) illustrates his point with the following examples: (8) a. *Его убило домом.  he:acc killed:neut house:inst *‘He got killed by a house.’ b. Его убило кирпичом. he:acc killed:neut brick:inst ‘He got killed by a brick.’

Davies (2005: 79) proposes adding the feature “–controlled” to the semantics of AI instrumental phrases. She shows that this feature is highly context dependent and

 Katrin Schlund

may thus apply to many nouns which are stereotypically conceived of as “+controlled” such as weapons or means of transportation (2005: 80–83). Davies (2005: 101) concludes that the semantics of AI is actually the result of the interplay between all the constituents of the construction, thus yielding a new and more complex unit than the sum of its parts would suggest. Although Davies is right to state that the special semantics of AI are created by the construction as a whole, it still holds that some nouns are more likely than others to occur in the instrumental phrases of AI. A closer look at the animacy properties of the patients of AI actually helps to find a more satisfactory explanation for the semantics of the causes as well. At first glance, one may be tempted to dismiss animacy as a useful explanation of AI patient semantics, since patients can be both animate and inanimate: (9) a. Ребенка раздавило грузовиком. child:acc ran-over:neut truck:inst ‘A child got run over by a truck.’ b. Подвал залило водой. basement:acc flooded:neut water:inst ‘The basement got flooded with water.’

(Green 1980: 173)

(Green 1980: 46)

Green (1980) assumes that one reason why AI emerged is that some kind of tension between agent and patient arises with inanimate (and thus non-prototypical) agents in canonical (nominative) transitive subject position. This tension is even greater with animate patients. In the above examples, the patients rank higher in animacy than the agents if we assume that solid, countable objects are “more like” animates than uncountable nouns. In (9a), we have an animate patient with a countable, yet inanimate noun in the instrumental phrase. In (9b), the patient is countable while the cause is not. It seems that cause and patient of AI are governed by an animacy hierarchy which functions similar to the animacy hierarchy in certain ergative splits. The “split animate” nature of AI is of course most obvious in the fact that only inanimate nouns are admitted in the instrumental phrase. However, the above reasoning also shows that the cause has to rank lower or at least equal in animacy to the patient. AI are thus a means of mediating conflicting hierarchies and serve to morphologically express finer degrees of agentivity than a canonically nominative-accusative language usually allows for. 2.4  Adverbial restrictions Adverbs expressing deliberation or control block AI: (10) *Меня преднамеренно ранило стрелой.  I:acc intentionally wounded:neut arrow:inst Intended: ‘I was intentionally wounded by an arrow.’

(Davies 2005: 108)



Russian Adversity Impersonals and split ergativity 

Notions such as deliberation and volition are closely linked with agentivity. As established in the previous section, AI are used to describe subtle degrees of noncanonical agentivity of inanimate forces. In Hindi, similar adverbial restrictions can be observed: (11) Hindi a. Raam-ø/-ne jor-se cillaayaa. raam:abs/-erg loudly shout.pf ‘Ram shouted loudly.’ (absolutive or ergative) b. Raam*-ø/-ne jaanbuujkar cillaayaa. raam:-abs/-erg deliberately shout:pf ‘Ram shouted deliberately.’ (ergative only)

(Ura 2006: 129)

With an adverb indicating a volitional action, the ergative is obligatory in Hindi even with an intransitive verb. Both Russian AI and Hindi ergative constructions are sensitive to different degrees of volition expressed by adverbial modifiers. 2.5  Word order and information structure Šeljakin (2002: 186) and Davies (2005: 107f.) claim that patient – verb – cause is the neutral word order in AI. The relative frequencies of word order patterns as attested by Mustajoki and Kopotev (2005: 10f.) are given below: n=2303 patient – verb – cause

36.6%

cause – verb – patient

26.5%

cause – patient – verb

 9.9%

patient – cause – verb

 6.0%

verb – patient – cause/verb – cause – patient

 3.0%

Although Davies’ and Šeljakin’s assumption is corroborated by the above ranking, the difference between the first and second most frequently occurring patterns is only 10%. Mustajoki and Kopotev (2005: 2f, 10f) deduce from their analysis that AI are less restricted in word order than corresponding active or passive sentences. However, they do not give any explanation for this behaviour of AI. A reason for the difference between the first and the second pattern can be found if we analyze the theme-rheme-structure of AI: (12) Дорогу занесло снегом. street:acc blocked-up:neut snow:inst ‘The street got blocked up with snow.’

(Green 1980: 184)

 Katrin Schlund

In this most frequently attested pattern, the patient forms the theme while verb and cause form the rheme. When the word order is reversed, the pragmatic roles are reversed, too, meaning that now cause and verb are the theme and the patient is the rheme: (13) Снегом занесло дорогу. snow:inst blocked-up:neut street:acc ‘The street got blocked up with snow.’

(Green 1980: 183)

As AI lack a nominative phrase, the prototypical equations “nominative phrase = subject = theme” and “accusative phrase = direct object = rheme” do not apply. Pragmatic and syntactic roles can thus be combined more freely in AI than in canonical transitive sentences. That is the reason why word order is more flexible in AI. However, the instrumental phrase (the cause) of AI is less likely to be part of the theme than the accusative phrase (the patient) as the cause is usually semantically and discursively peripheral to the patient. This explains the slight difference in frequency between the word order patterns “patient – verb – cause“ and “cause – verb – patient” established by Mustajoki and Kopotev (2005: 10f.). The verb final AI occur mostly in connection with personal pronouns, which are of course very likely to be part of the theme: (14) Повезло дураку, каким-то чудом его из was-lucky:neut fool:dat some:inst wonder:inst he:acc out-of кабины просто выкинуло. cabin:gen simply threw-out:neut ‘The fool was lucky, by some wonder he simply got thrown out of the booth.’(Mustajoki & Kopotev 2005: 11)

In the above example the verb functions as rheme, while both patient and cause form the theme. This pattern occurs quite rarely. The verb initial AI are of even less importance since they carry an artistic and often poetical note (Mustajoki & Kopotev 2005: 11). The question remains whether there are any parallels between the information structure of AI and ergative constructions. Y. Kachru (1987) conducted a study on topicality and subjecthood of ergative noun phrases in Hindi. The study empirically corroborates the assumption that ergative noun phrases are less likely to occur in topic/theme position than nominative noun phrases. In particular, ergative noun phrases are hardly ever used to initiate topics/themes but rather serve to maintain topicality (Kachru 1987: 235). If we again compare the instrumental phrases of AI with ergative noun phrases in Hindi, we find that both are less apt



Russian Adversity Impersonals and split ergativity 

to occupy the initial topic/theme position, a position which is canonically held by a prototypical subject. 2.6  Impersonal morphology Impersonal morphology has been linked with syntactic unaccusativity (e.g. Harves 2009: 429). As impersonal morphology is compulsory with AI as well, this criterion allows for including AI into the unaccusative phenomena of Russian syntax. In this context, it is interesting to note that in transitive perfective sentences in Hindi (i.e. in a context requiring the ergative pattern), either the verb agrees with the unmarked O in the absolutive, or, when O is marked with accusative, neutral (impersonal) morphology is employed (Kachru 1987: 227). Impersonal morphology has also been linked with emerging or vanishing ergativity. Comrie (1978: 330) states that impersonal morphology may occur as a “compromise” (ibid.) in languages which are in a transitional state from ergativity to accusativity or vice versa. Impersonal morphology obviously arises when no argument has the power impose agreement on the verb.

3.  Conclusion and further reasoning Although some of the observations made here are still tentative, the above reasoning showed that there are indeed formal, semantic and pragmatic parallels between AI and ergative splits. The typological view on AI allows us to interpret Russian as more ergative than Western European languages such as English, German, or French. It seems that a language can vary more easily between nominative-accusative and ergative-absolutive features when it is rich in inflectional morphology and free in word order. This flexibility is reflected in a comparatively big inventory of impersonal constructions. Even a superficial check of the relevant parameters in Russian and the aforementioned Western European languages corroborates this. The situation within the Slavic world is, however, more complex. As mentioned in the introduction, AI are productive only in Russian, and, to a lesser extent, in Ukrainian and Belarusian (cf. Khodova 1958: 155), but almost absent in West and South Slavic (with the exception of Polish, see below) (cf. Mrázek 1956: 25). Croatian, Serbian, Slovenian and Czech are rich in inflectional morphology and relatively free in word order, too, so those criteria do not hold for an intra-Slavic explanation of the occurrence of AI (cf. Guiraud-Weber 1978: 137, for a similar line of thought).

 Katrin Schlund

Ivić (1965: 318) points out that AI are productive only in languages which accept an instrumental agent in the passive voice.8 Although Ivić is right to claim a correlation between the instrumental case of AI and the passive instrumental agent, the question remains why East Slavic allows for those uses of the instrumental as opposed to the rest of the Slavic world. Therefore, one has to look for other systemic features which distinguish East Slavic from the other Slavic languages and which could account for the existence and productivity of passive instrumental agents and of AI as well. When comparing East Slavic with the rest of the Slavic family, probably the most striking difference is that East Slavic languages, and especially Russian, form a predicative possessive construction with a prepositional genitive and the verb be, and only rarely a transitive construction with the verb have. As discussed in ­Section 2.2, the possessive construction with be often forms the core of an ergative construction in the past tense. The rise of an ergative construction out of a possessive implies a reinterpretation of the possessor as agent and of stative as resultative (Trask 1979: 398). This is exactly what has happened in North ­Russian dialects and is justifiably labeled “ergative” by Orr (1989, 1991) and others. Although the possessive-ergative construction has not penetrated into standard Russian, it can be assumed that the existence of have is generally inconvenient for the emergence of impersonal constructions. As pointed out by E. Benveniste (1960), A. Isačenko (1974), and R. Orr (1989), be-languages are poor in modal auxiliaries and tend toward impersonal constructions with dative experiencers to convey modal meanings, for instance Russian možno, nado, nel‘zja (Isačenko 1974: 74). If there is indeed a correlation between the grammaticalization of have and the occurrence of AI and impersonal constructions in general, one would expect that AI were most productive in Russian, followed by Belarusian, ­Ukrainian, and Polish, for the latter languages “seem to be in a state of transition from B- to ­H-languages” (Isačenko 1974: 44). As regards East Slavic, the hypothesis of a correlation between the grammaticalization of have and the emergence of AI seems tenable although it of course needs further investigation. As regards Polish, we are faced with a problem because Polish both has various kinds of grammaticalizations of mieć unknown in East Slavic and a great variety of impersonal constructions, including AI (see example (2c)). However, it seems that Polish AI are restricted to colloquial speech and tend to be replaced by regular passives in official language (K. Janic, p. c.). Indeed, it has been established by Doros (1975: 133) that AI are less ­productive in Polish than in Russian and that this is probably due to the fact that

.  This generalization does not seem to hold for Czech (for Czech examples using instrumental passive agents see, for instance, Fried 2006).



Russian Adversity Impersonals and split ergativity 

Polish does not accept d ­ efocused instigators occurring openly in an oblique case. Note that not all impersonal constructions allow for an instigator to be openly marked; many leave the instigator unexpressed (this holds, for instance, for the majority of Polish impersonals such as się- and no-/to-constructions, cf. Krzek, this volume). Accordingly, M. Ivić (1965: 318) states that Polish AI are rather infrequent and points out that their existence is sometimes attributed to Russian influence altogether.9 It thus holds that all Slavic languages which do not dispose of productive AI of the Russian type are at the same time classified as have-languages by A. Isačenko (1974). Given that only East Slavic accepts instrumental instigators (be they agents or causes) in passive voice and AI, it is interesting to note that the ergative marker in Hindi originates from a Sanskrit instrumental ending (Orr 1991: 166). The instrumental agent in the Russian passive voice has indeed been paralleled with ergative case (for instance by Comrie 1978: 33, and Gvozdanović 2004: 344f.). The hypothesis that the existence and productivity of AI is systematically linked with the status of have and be in Slavic requires further investigation. First, the contemporary status and historical development of AI must be established empirically. Second, the history of have needs to be investigated and correlated with the findings on AI.10 Finally, other impersonal constructions with and without openly expressed instigators should be taken into consideration. The long-term objective of such an investigation would be to explain why certain impersonal constructions occur in some Slavic languages but not in others. It has been assumed in this section that the grammaticalization of have is of some importance in this respect. However, nominalization patterns and valency changing mechanisms such as passivization and antipassive phenomena should be taken into consideration, too – i.e. factors which are of general importance within relational typology.

References Alisova, T.B. 1969. Semantiko-kommunikativnyj substrat bezličnyx predloženij. In Invariantnye sintaksičeskie značenija i struktura predloženij, Nadežda D. Arutjunova (ed.), 27–36. Moskva: Nauka. Babby, Leonard H. 1994. A theta-theoretic analysis of adversity impersonal sentences in ­Russian. In Formal Approaches to Slavic Linguistics (FAS-2), The MIT meeting 1993. [Michigan Slavic

.  For a brief discussion of the translatability of Russian AI into Polish see Weiss (this volume). .  A first attempt in this direction is made in Schlund (2013).

 Katrin Schlund Materials 36], S. Avratin, Steven Franks & Ljiljana Progovac (eds), 25–67. Ann Arbor MI: Michigan Slavic Publications. Benveniste, Émile. 1960. “Être” et “avoir” dans leurs fonctions linguistiques. Bulletin de la Société de Linguistique de Paris 55(1): 113–134. Comrie, Bernard. 1978. Ergativity. In Syntactic Typology. Studies in the Phenomenology of ­Language. Winifred Lehmann (ed.), 329–394. Sussex: Harvester Press. Comrie, Bernard. 1981. Language Universals and Linguistic Typology. Syntax and Morphology. Oxford: Basil Blackwell. Davies, Laura E. 2005. A Construction-grammatical Analysis of Impersonalization in Russian. Ph.D., Princeton University. Dixon, Robert M.W. 1979. Ergativity. Language 55(1): 59–138. Doros, A. 1975. Werbalne konstrukcje bezosobowe w języku rosyjskim i polskim na tle innych języków słowiańskich [Polska akademia nauk – odział w Krakowie, Prace komisji słowianoznawstwa 32]. Warszawa: Zakład narodowy imienia ossolińskich wydawnictwo polskiej akademii nauk. Fried, Mirjam. 2006. Agent back-grounding as a functional domain. Reflexivization and ­passivization in Czech and Russian. In Demoting the Agent: Passive, Middle and Other Voice Phenomena [Linguistik Aktuell/Linguistics Today 96], Benjamin Lyngfelt & Torgrim ­Solstad (eds), 83–109. Amsterdam: John Benjamins. Galkina-Fedoruk, E.M. 1958. Bezličnye predloženija v sovremennom russkom jazyke. Moskva: Izdat. Mosk. Univ. Green, Mark Christopher. 1980. On the Syntax and Semantics of Impersonal Sentences in ­Russian: A Study of the Sentence Type “vetrom uneslo lodku”. Ann Arbor MI: University Microfilms International. Guiraud-Weber, Marguerite. 1978. Klassifikacija prostogo predloženija i ee posledstvija dlja tipologii jazykov. Revue des Etudes Slaves 51(1–2): 129–137. Guiraud-Weber, Marguerite. 1984. Les propositions sans nominatif en russe moderne. Paris: ­Institut d’Études Slaves. Guiraud-Weber, Marguerite. 2002. Sub”ektnye čerty i problema podležaščego v russkom jazyke. Revue des Études Slaves 74(2–3): 279–289. Gvozdanović, Jadranka. 2004. Morphosyntactic transparency in Bantawa. In Himalayan ­Languages: Past and Present, Anju Saxena (ed.), 341–347. Berlin: de Gruyter. Harves, Stephanie A. 2009. Unaccusativity. In Die slavischen Sprachen. Ein internationales Handbuch zu ihrer Struktur, ihrer Geschichte und ihrer Erforschung, Bd. 1: The Slavic ­Languages. An International Handbook of their Structure, their History and their Investigation, Vol. 1 [Handbücher zur Sprach-und Kommunikationswissenschaft 23(1)], Sebastian Kempgen, Peter Kosta, Tilman Berger & Karl Gutschmidt (eds), 415–430. Berlin: Mouton de Gruyter. Hofmann, Eeich. 1934. Impersonale mit Instrumental im Russischen. Zeitschrift für V ­ ergleichende Sprachforschung auf dem Gebiete der indogermanischen Sprachen 61: 209–221. Isačenko, Alexandre V. 1974. On ‘have’ and ‘be’ languages. A typological sketch. In Slavic Forum. Essays in Linguistics and Literature, Michael S. Flier (ed.), 43–77. The Hague: Mouton. Ivić, Milka. 1965. On the origin of the Russian sentence type “(ego) zavalilo snegom”. Die Welt der Slaven X(3–4): 317–321. Kachru, Yumuna. 1987. Ergativity, subjecthood and topicality in Hindi-Urdu. Lingua 71: 223–238. Kacnel’son, Solomon D. 1974. Sprachtypologie und Sprachdenken. München: Hueber. Keenan, Edward. 1976. Towards a unified definition of subject. In Subject and Topic, Charles N. Li (ed.), 303–333. New York NY: Academic Press.



Russian Adversity Impersonals and split ergativity 

Khodova, K.I. 1958. Tvoritel’nyj padež v stradatel’nyx konstrukcijax i bezličnyx predloženijax. In Tvoritel’nyj padež v slavjanskix jazykax. S.B. Bernštejn (ed.), 127–158. Mosvka: Izdatel’stvo Akademii Nauk SSSR. Lavine, James E. 2010. Case and events in transitive impersonals. Journal of Slavic Linguistics 18(1): 101–130. Lavine, James E. & Freidin, Robert. 2002. The subject of defective T(ense) in Slavic. Journal of Slavic Linguistics 10(1–2): 253–289. Maslov, Jurij S. 1988. Resultative, perfect, and aspect. In Typology of Resultative Constructions [Typological Studies in Language 12], Vladimir N. Nedjalkov (ed.), 63–85. Amsterdam: John Benjamins. Montaut, Annie & Pilot-Raichorr, C. 1994. Sémantique actancielle et cas morphologiques en hindi et en badaga. Langages 28 (113): 90–112. Mrázek, Roman. 1956. K jednočlenným větám slovesným. In Kapitoly ze srovnávací mluvnice ruské a české, I: Studie syntaktické, Karel Horálek (ed.), 7–81. Praha: Nakladatelstvi Českosl. Akad. Věd. Mustajoki, Arto. 2006. The Integrum Database as a powerful tool in research on contemporary Russian. In Integrum: Točnye metody i gumanitarnye nauki, Galina Nikiporec-Takigava (ed.), 50–75. Moskva: Letnij sad. Mustajoki, Arto & Kopotev, Mixail. 2005. Lodku uneslo vetrom: Uslovija i konteksty upotreblenija russkoj “stixijnoj” konstruckii. Russian Linguistics 29(1): 1–38. Orr, Robert. 1989. A Russo-Goidelic syntactic parallel: U nego svoja izba postavlena/tá sé déanta agam. General Linguistics 29(1): 1–21. Orr, Robert. 1991. More on embryonic ergativity. General Linguistics 31(3–4): 163–175. Pontoppidan-Sjövall, Karin. 1968. Categories of Form and Content in Language. A Study of the Personal and Impersonal Constructions in Russian. Stockholm: Almqvist & Wiksell. Sal’nikov, N. 1977. Bezličnye predloženija tipa: “Kryšu sorvalo vetrom”. Russian Linguistics 3: 271–292. Schlund, Katrin. 2013. Zur Grammatikalisierung von imet’ und der Entstehung von russischen unpersönlichen Konstruktionen des Typs vetrom sorvalo kryšu. In Sprachgeschichte und Sprachwandel im Slavischen. Festschrift für J. Gvozdanović, Alexander Bierich (ed.). ­Frankfurt: Peter Lang. Słoń, Anna. 2007. The ‘impersonal’ impersonal construction in Polish. A cognitive grammar analysis. In Cognitive Paths into the Slavic Domain. Dagmar Divjak & Agata Kochańska (eds), 257–287. Berlin: Mouton de Gruyter. Sulejmanova, Ol’ga A. 1999. Problemy russkogo sintaksisa. Semantika bezličnyx predloženij. Moskva: Dialog-MGU. Silverstein, Michael. 1976. Hierarchy of features and ergativity. In Grammatical Categories in Australian Languages, Robert M.W. Dixon (ed.), 112–171. Canberra: Australian Institute of Aboriginal Studies. Šeljakin, Mixail A. 2002. Jazyk i čelovek (k probleme motivirovannosti jazykovoj sistemy). Trudy po russkoj i slavjanskoj filologii. Lingvistika. Novaja serija VII. Tartu: Ülikooli Kirjastus. Trask, Robert L. 1979. On the origins of ergativity. In Ergativity. Towards a Theory of Grammatical Relations, Frans Plank (ed.), 385–404. London/New York NY: Academic Press. Ura, Hiroyuki. 2006. A parametric syntax of aspectually conditioned split-ergativity. In Ergativity. Emerging Issues, Alana Johns, Diane Massam & Juvenal Ndayiragije (eds), 111–141. Dordrecht: Springer. Włodarczyk, Hélène. 1996. Les traits sémantiques du sujet “anonyme” en polonais, russe et français. Semantyka a konfrontacja językova 1: 179–198.

section iv

Lexical semantics

Morphological and lexical aspect in Russian deverbal nominalizations* Glòria de Valdivia, Joan Castellví & Mariona Taulé Universitat de Barcelona, Spain

This paper examines whether it is possible to establish a relationship between the lexical denotation of the deverbal noun with the morphological and lexical aspect of its base verb. We have classified and analyzed deverbal Russian nouns in relation with their corresponding verbs to understand more deeply what is inherited from the verb by the nominalization. Keywords:  deverbal nominalization; aspect; Russian

1.  Introduction Deverbal nominalizations in Russian inherit the presence of the aspectual ­morphological marks of the base verb from which the nominalization derives. The main focus of this paper is to analyze to what extent the morphological and the lexical aspect of the base verb determines the lexical denotation of its ­corresponding nominalization, in our case its denotation as an event, as a result or as a state. The initial hypothesis of our research is that it is not possible to say that morphological and lexical aspect of the base verb determines the lexical denotation of the deverbal noun, although it seems that both have significant influence on the lexical aspect of the nominal. In order to examine these hypotheses we proceed as follows. (a) First, we analyze the different types of nominalizations on the basis of the traditional verbal classification into aspectual paired, biaspectual and uniaspectual verbs. The result of this analysis is a deverbal nominalization classification, which enables us to determine from which class of verb the nominalization is derived and whether the nominalization has inherited morphological aspectual marks – either imperfective or perfective – from the corresponding verb. (b) ­Second, we analyze each type of nominalization in examples from real data

*  We thank the anonymous reviewers of the Slavic Linguistic Society 2011 for their useful ­observations and notes.

 Glòria de Valdivia, Joan Castellví & Mariona Taulé

in order to ­establish their denotation or lexical aspect (i.e. event, result or state). The main two goals in doing this analysis are the following: first, to determine the influence of the morphological aspect of the base verb on the lexical denotation of the nominalization, and, second, to determine the influence of the lexical aspect of the base verb on the lexical denotation of the deverbal noun. The intent of this research is to understand more deeply the relationship established between the nominalization and its base verb. 2.  Morphological and Lexical aspect in deverbal nouns Nominalizations derived from a verb may preserve morphological aspectual marks from the corresponding base verb. In the case of verbs, those aspectual marks have a grammatical function. In fact, depending on the expression of the ­morphological aspect, verbs are traditionally classified into aspectual paired (pisat’/ napisat’ ‘to write’), biaspectual (ženit’sja ‘to marry’) and uniaspectual (­sglazit’ ‘to put the evil eye’). A logical question is then to ask whether or not nominalizations have m ­ orphological aspect because of the presence or the absence of such ­morphological marks as well. In fact, in other Slavic languages such as Polish, verbal nouns and deverbal nouns are distinguished by the presence of ­morphological aspect (Komur 2005). Most authors (Vinogradov 1972; S­choorlemmer 1995; Pazelskaya & Tatevosov 2003) consider that Russian nominalizations do not have such an aspect, thus aspectual marks inherited by the nominals do not have a grammatical function. Some authors claim that, even if there is not m ­ orphological aspect in the n ­ ominalizations, some inherited verbal affixes have an influence on the lexical denotation of the nominalization. ­Vinogradov (1972) postulates that the s­econdary imperfectivizing suffix -(y/i)va- gives a repetitive nuance to the meaning of the nominalization. Schoorlemmer (1995) and Zimmermann (2002) suggest that this -(y/i)va- suffix gives some complexity to the nominal giving some unambiguous event character to it. Pazelskaya & Tatevosov (2003) do not see such an influence and claim that the lexical aspect of the nominalization depends on the event structure of the corresponding verb and its internal argument. Th ­ erefore, the inheritance of the aspectual marks by the nominal poses the question about the influence of such marks on the lexical meaning of the nominalization. ­Following Vendler (1957), lexical aspect traditionally classifies verbs into states, activities, achievements and accomplishments. Nominalizations derived from verbs also have lexical aspect, which may be studied through the property of ­telicity (Pazelskaya & Tatevosov 2003) or through the denotative differences such as event, result and state. By an event we mean the action expressed by the ­corresponding base verb, a result names the concrete or abstract entity



Morphological and lexical aspect in Russian deverbal nominalizations 

related to the action and a state refers to the n ­ on-dynamic situation named by the c­ orresponding verb. Our proposal is parallel to what is proposed in literature. Event nominalizations correspond to the complex event/process/action nouns, result nouns correspond to result/result-object and state nouns to state/stateobject proposed in Apresjan 1971, Grimshaw 1990, Pustejovsky 1995, A ­ lexiadou 2001, Eberle et al. 2009, Peris & Taulé 2009, Balvet et al. 2010, among others. 3.  Classification of deverbal nominalizations This section presents a deverbal nominalization classification based both on the different verbal types and on the morphological aspect of the c­ orresponding base verb. The main point in proposing this classification is to see whether or not there is any systematic relationship between morphological aspect of the base verb and the lexical aspect of the deverbal noun. The analysis is based on a sample of 296 different types of nominalizations derived from 294 different verbs extracted from ‘The Essex Database of Russian Verbs and their Nominalizations’ (Spencer &Z ­ aretskaya, 2010). The Essex Database is composed of 7,000 verbs and their ­corresponding 5,000 nominalizations. This database codifies both morphosyntactic and semantic information, from which we took into account: the type of base verb (paired, b ­ iaspectual and uniaspectual) from which the nominalization derives, its ­corresponding aspectual pair (if there is one), and the nominalizations derived. The selected nominalizations are extracted according to a list that includes the most productive Russian nominalizing suffixes, i.e. the group of ­suffixes ending in -ij(e),1 -k(a), -stv(o), -Ø, -Ø(a) (Pazelskaya (2009) based on Švedova (1982: 157–166)). In order to provide a balanced sample of ­nominalizing suffixes, we extracted about 300 nominalizations, from which 296 instances were finally selected. These correspond to 8% of the instances in the database containing each of these nominalizing suffixes. The selection of 296 nominalizations includes every productive nominalizing suffix. In the case of the least p ­ roductive suffixes (such as, -iš, -išč(e),-otn(ja), -ot(a)) they are represented in almost its totality since there are few deverbal nouns with these suffixes. The classification of Russian deverbal ­nominalizations takes into account the type of verb – aspectual paired, biaspectual and uniaspectual – and the ­morphological aspect – perfective or ­imperfective – of the base verb. This ­ classification groups deverbal ­ nominalizations into ­symmetrical, neutralized, biaspectual and uniaspectual.

.  -ij(e) suffixes include: -nij(e), -enij(e), -anij(e), -tij(e), -an’(e), -en’(e), -n’(e), -itij(e), -t’e, -vij(e).

 Glòria de Valdivia, Joan Castellví & Mariona Taulé

1. Symmetric Nominalizations are derived from a paired verb, and are those in which different deverbal nouns are derived from each member of the aspectual paired verb. Consequently, the aspectual morphological opposition of the corresponding base verb is preserved (1). (1) izživat’IPF ‘to eliminate’ > izživanie izžitie kroj krojka kroenie Ø

The verbs kroit’/skroit’ generate their corresponding nominalizations from only one member of the aspectual pair. In the case of (2) there is no aspectual mark inherited from the base verb kroit’. These nominalizations kroj, krojka, kroenie have been derived by means of different nominalizing suffixes, that is, a zero ­morpheme, -k(a) and -enij(e). 3. Biaspectual Nominalizations are generated from a unique biaspectual ­verbal form. Consequently, there is not aspectual opposition of morphological marks neither in the original verb nor in the nominalization.

(3)  demoralizovat’IPF & PF ‘to demoralize’

> demoralizacija