New Approaches to Slavic Verbs of Motion (Studies in Language Companion Series)

New Approaches to Slavic Verbs of Motion Studies in Language Companion Series (SLCS) This series has been established...

Author: Victoria Hasko | Renee Perelmutter

402 downloads 1230 Views 2MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!

Report copyright / DMCA form

DOWNLOAD PDF

New Approaches to Slavic Verbs of Motion

Studies in Language Companion Series (SLCS) This series has been established as a companion series to the periodical Studies in Language.

Editors Werner Abraham University of Vienna

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

Christian Lehmann University of Erfurt

Brian MacWhinney

Carnegie-Mellon University

Marianne Mithun

University of California, Santa Barbara

Heiko Narrog

Tohuku University

Johanna L. Wood

University of Aarhus

Volume 115 New Approaches to Slavic Verbs of Motion Edited by Victoria Hasko and Renee Perelmutter

New Approaches to Slavic Verbs of Motion Edited by

Victoria Hasko University of Georgia

Renee Perelmutter University of Kansas

John Benjamins Publishing Company Amsterdamâ•›/â•›Philadelphia

8

TM

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

Library of Congress Cataloging-in-Publication Data New approaches to Slavic verbs of motion / edited by Victoria Hasko, Renee Perelmutter. p. cm. (Studies in Language Companion Series, issn 0165-7763 ; v. 115) Includes bibliographical references and index. 1. Slavic languages--Verb. 2. Motion--Terminology. I. Hasko, Victoria. II. Perelmutter, Renee. PG169.N49â•…â•… 2010 491.8’0456--dc22 2009044948 isbn 978 90 272 0582 7 (Hb ; alk. paper) isbn 978 90 272 8863 9 (Eb)

© 2010 – 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

Contributors introduction Verbs of motion in Slavic languages: Paths for exploration Victoria Hasko and Renee Perelmutter

ix 1

Part I.â•‡ Diachrony of motion expressions chapter 1 Clause and text organization in early East Slavic with reference to motion and position expressions Sarah Turner chapter 2 Indeterminate motion verbs are denominal Johanna Nichols chapter 3 Common Slavic “indeterminate” verbs of motion were really manner-of-motion verbs Stephen M. Dickey chapter 4 PIE inheritance and word-formational innovation in Slavic motion verbs in ‑i‑ Marc L. Greenberg

15

47

67

111

Part II.â•‡ Synchronic approaches to aspect chapter 5 Perfectives from indeterminate motion verbs in Russian Laura A. Janda

125

vi

New Approaches to Slavic Verbs of Motion

chapter 6 Aspects of motion: On the semantics and pragmatics of indeterminate aspect Olga Kagan chapter 7 Verbs of motion under negation in Modern Russian Renee Perelmutter

141

163

Part III.â•‡ Typological approach to the study of Slavic verbs of motion chapter 8 Semantic composition of motion verbs in Russian and English: The case of intra-typological variability Victoria Hasko chapter 9 Motion events in Polish: Lexicalization patterns and the description of Manner Anetta Kopecka chapter 10 The importance of being a prefix: Prefixal morphology and the lexicalization of motion events in Serbo-Croatian Luna Filipović

197

225

247

chapter 11 Variation in the encoding of endpoints of motion in Russian Tatiana Nikitina

267

chapter 12 Verbs of rotation in Russian and Polish Ekaterina V. Rakhilina

291

chapter 13 Aquamotion verbs in Slavic and Germanic: A case study in lexical typology Maria Koptjevskaja-Tamm, Dagmar Divjak and Ekaterina V. Rakhilina

315

Table of contents vii

chapter 14 Metaphorical walking: Russian idti as a generalized motion verb Tore Nesset chapter 15 Russian verbs of motion: Second language acquisition and cognitive linguistics perspectives Kira Gor, Svetlana Cook, Vera Malyushenkova and Tatyana Vdovina Author index Language index Subject index

343

361

383 387 389

Contributors

Svetlana Cook 2106 Jiménez Hall University of Maryland College Park, MD 20742, USA [email protected] Stephen M. Dickey University of Kansas Department of Slavic Languages & Literatures Wescoe Hall 1445 Jayhawk Boulevard, Rm. 3051 Lawrence, KS 66045-7590, USA [email protected] Dagmar Divjak University of Sheffield Russian and Slavonic Studies Jessop West 1 Upper Hanover Street Sheffield S3 7RA, UK [email protected] Luna Filipović University of Cambridge Department of Linguistics Sidgwick Avenue Cambridge CB3 9DA, UK [email protected] Marc Greenberg University of Kansas Dept. of Slavic Languages & Literatures 1445 Jayhawk Blvd., Rm. 2133 Lawrence, KS 66045-7590, USA [email protected]

Kira Gor University of Maryland School of Languages, Literatures and Cultures 3215 Jiménez Hall College Park, MD 20742, USA [email protected] Victoria Hasko University of Georgia Department of Language and Literacy Education 125 Aderhold Hall Athens, GA 30602, USA [email protected] Laura A. Janda University of Tromsø Det humanistiske fakultet N-9037, Tromsø, Norway [email protected] Olga Kagan Hebrew University of Jerusalem Emek Ayalon 48/8 Modiin 71700, Israel [email protected] [email protected] Anetta Kopecka Max Planck Institute for Psycholinguistics Nijmegen, the Netherlands Institut des Sciences de l’Homme Laboratoire Dynamique du Langage 14 avenue Berthelot 69363 Lyon Cedex 07, France [email protected]

˘

New Approaches to Slavic Verbs of Motion

Maria Koptjevskaja-Tamm Stockholm University Department of Linguistics C342, Södra huset, Frescati S-106 91 Stockholm, Sweden [email protected] Vera Malyushenkova 197 Prospect Obuchovskoy Oborony Apt. 203 St. Petersburg 192174, Russia [email protected] Tore Nesset University of Tromsø NO-9037 Tromsø, Norway [email protected] Johanna Nichols University of California, Berkeley Slavic Languages and Literatures #2979 6303 Dwinelle Hall Berkeley, CA 94720, USA [email protected] Tatiana Nikitina 5706 S. Blackstone Ave Apt. 1 Chicago, IL 60637, USA [email protected]

Renee Perelmutter University of Kansas Department of Slavic Languages & Literatures Wescoe Hall 1445 Jayhawk Boulevard, Rm. 2127 Lawrence, KS 66045-7590, USA [email protected] Ekaterina Rakhilina Institute for the Russian Language, Moscow Volkhonka 18/2 Moscow 119019, Russian Federation [email protected] Sarah Turner University of Waterloo Department of Germanic and Slavic Studies Waterloo, ON N2L 3G1, Canada [email protected] Tatyana Vdovina 8807 Plymouth St. Apt. 6 Silver Spring, MD 20901, USA [email protected]

introduction

Verbs of motion in Slavic languages Paths for exploration Victoria Hasko and Renee Perelmutter University of Georgia / University of Kansas

Our nature consists in motion.

1.

(Pascal, Pensées)

Multiple perspectives on the expression of motion

The relation of humans to space is fundamental to our physical functioning, cognition, and linguistic expression. One of the ways that humans relate to space is through motion, i.e., a series of changes in spatial relations such as change of location or orientation (Miller & Johnson-Laird 1976). Motion is an integral part of our daily lives. We are cognitively primed to process motion events and to express our conceptualization of these events through language, which makes the linguistic encoding of motion meanings an important and necessary semantic domain in all world languages. It is, therefore, not surprising that widespread interest has been generated in the expression of motion events, which is currently being closely investigated by theoretical, cognitive, and applied linguists all over the world (e.g., see recent volumes by Aurnague, Hickmann, & Vieu 2007; Han & Cadierno forthcoming; Hickmann & Robert 2006; Levinson & Wilkins 2006; Shay & Seibert 2003). Despite the universality of motion, there is a growing consensus that there is no universal way of expressing spatial meanings. The specific surface structures and conceptual categories encoding motion events exhibit a high degree of cross-linguistic variation (Berman & Slobin 1994; Bowerman & Levinson 2001). Each human language provides a powerful and diverse pool of resources for its speakers to encode spatial relations and motion experiences – basic as well as refined. Acquisition of spatial cognitive categories by children has been shown to be influenced by language-specific patterns (Bowerman & Choi 2001; Gumperz & Levinson 1996; Slobin 1996), and studies investigating second language acquisition by adult learners have similarly documented the influence of the first

˘

Victoria Hasko and Renee Perelmutter

language and the motion-encoding categories acquired in childhood (Cadierno 2004; Hasko 2009; Inagaki 2002; Slobin 2005). A study of any linguistic domain is most effective when situated typologi‑ cally (Slobin 2003) in a contrastive network of interactive (psycho)linguistic factors. The most prominent and comprehensive typology of the lexicalization patterns and conceptual elements involved in the encoding of motion meanings across languages was put forth by Leonard Talmy (1985, 1991, 2000), who outlined systematic lexicalization patterns and conceptual categories that underlie the encoding of motion events cross-linguistically. An impressive body of studies confirms the robustness of the typological patterns proposed by Talmy (e.g., Hickmann & Robert 2006; Strömqvist & Verhoeven 2004; Verhoeven & Strömqvist 2001; Slobin 1996). At the same time, fine-grained intra-linguistic studies report that this typology fails to capture important factors in the expression of motion meanings in individual languages. Idiosyncratic variations from the typology have been noted with respect to lexis, rhetorical style, narrative construction, morphosyntax, cultural detailing, functional properties of entities and figures, reference systems, and modality (e.g., Bavin 2004; McNeill 2005; Özyürek & Kita 1999; Shay & Seibert 2003; Slobin 2003). The linguistic encoding of motion events thus emerges as a phenomenon that merits in-depth analysis from both a cross-typological and a language-specific (or at least language group-specific) perspective.

2.

Expression of motion in Slavic languages: Goals of the project

The present volume adds to the body of research on motion talk by offering a collection of interdisciplinary studies investigating how motion is expressed within the Slavic language family. As one of the largest families among Indo-European languages, Slavic languages warrant attention and should not be overlooked in the study of the cross-linguistic domain of motion due to the sheer size of the language family. Furthermore, while the linguistic expression of motion has only recently garnered extensive scholarly attention cross-linguistically, the rich lexical system of common verbs of motion has long captured the attention of Slavic linguists. For readers who are unfamiliar with the nomenclature used in the Slavic scholarly literature, it is important to understand that the term verbs of motion and the majority of the associated research efforts have largely focused on the study of a small, closed lexical category of paired motion verbs, with different West and East Slavic languages retaining between 9 and 20 pairs (Sussex & Cubberley 2006). This class of verbs of motion is notorious among Slavic linguists for their idiosyncratic

Verbs of motion in Slavic languages

behavior in their lexical, semantic, syntactical, and aspectual characteristics (Forsyth 1970; Isachenko 1960; Stilman 1951; Vinogradov 1960). In spite of previous work on the topic, the baffling uniqueness of Slavic motion verbs continues to attract scholarly interest. In fact, to date, discussion of the topic has remained rather disjointed – if not erratic – geographically, topically, diachronically, and terminologically. Thus, readers will find that publications on motion structures are dispersed throughout journals, books chapters, and dissertations, which may not only vary in terms of quality and the language of publication but may also be outdated and/or difficult to obtain. Therefore, the goal of this project is to unify a wide breadth of recent studies examining Slavic motion talk from multiple points of view. The contributors to the volume have joined in the discussion of Slavic motion talk from diachronic, theoretical, typological, comparative, cognitive and acquisitional perspectives with a particular focus on verbs of motion, the nuclei of the lexicalization patterns for encoding motion. This interdisciplinary effort is aimed both at accounting for unique semantic and syntactic properties of Slavic motion verbs that have long baffled linguists and learners of Slavic languages and at situating Slavic languages within the larger typological framework (Talmy 1985, 2000; Slobin 2005).

3.

Organization and overview of chapters

The volume is organized into three parts that address the main areas of investigation. Articles in Part I, Diachrony of motion expressions, question how the expression of motion evolved from Proto-Slavic to the present day, with a particular emphasis on the development of a special aspectual system for motion verbs. Contributions to Part II entitled Synchronic approaches to aspect question the place of motion verbs in the aspectual system of modern Slavic languages. Finally, articles in Part III, Typological approach to the study of Slavic verbs of motion, seek to situate Slavic languages within the broader context of the cross-linguistic study of motion expressions.

3.1

Diachrony of motion expressions

Investigators of Slavic languages are unusually lucky in that ample evidence of linguistic production is available since the introduction of literacy to the Slavs in 864AD. Old Church Slavic is attested in texts dating since the 10th century; the earliest Old Russian texts date from 11th century (Schenker 1996). In addition, scholars can draw on data from other Indo-European languages in order

˘

˘

Victoria Hasko and Renee Perelmutter

to reconstruct the earliest stages of Slavic. Since motion constructions in modern Slavic languages are so rich and complex, it is only natural to ask how they evolved over time. The articles in this section examine the evolution of word order patterns and the evolution of aspectual categories, in order to shed light on diachronic processes involving motion verbs. Turner investigates word order patterns in clauses that contain intransitive verbs of motion and position. Examples for this study were collected from a number of Early Slavic texts with varying textual organization principles (temporal, spatial, and personal). Turner is interested in determining to what extent the three types of word order generally discussed for modern Russian, i.e., informational, (giving additional information about a referent that is already under discussion), presentational (introducing a new referent), and expressive are applicable to the analysis of medieval data. Turner’s study shows that these modern categories are not helpful for analyzing the earlier texts, and that pragmatic considerations, as well as textual organization principles, only partially account for the wealth of word order patterns found in these texts. In Indeterminate verbs of motion are determinate, Nichols discusses the formation of Proto-Slavic indeterminate verbs. In Proto-Indo-European, verb roots were basic and underived. After the breakup of Proto-Balto-Slavic, Slavic shifted from the inherited verb-based derivation to neutral and, in some instances, even noun-based derivation. To illustrate this shift in derivation patterns, Nichols traces the development of the indeterminate category, using the indeterminate motion verb xoditi ‘to go’ as an example. She argues that the vocalism, the -i- stem suffixes, and the very indeterminate aspect of these verbs can all be traced to their denominal origin. The development of the indeterminate category of Slavic motion verbs is further discussed in two companion articles by Greenberg and Dickey. Dickey draws on early Slavic as well as modern South Slavic data for his analysis of the putative category of indeterminate verbs in Common Slavic. While determinate motion as a conceptual category is widespread in the world’s languages, the category of indeterminate motion is typologically unusual. The opposition between determinate and indeterminate motion verbs is familiar from modern East and West Slavic languages; however, it is unclear whether this opposition existed in Common Slavic. Dickey examines a variety of Old Church Slavic and Old Russian data to show that verbs such as xoditi ‘walk’, ězditi ‘ride’, běgati ‘flee’, nositi ‘carry’, and other verbs classified as indeterminate could be used in determinate contexts; thus, the determinate/indeterminate opposition is not useful for the earlier stages of Slavic. Instead, Dickey uses semantic and pragmatic analysis of the earliest examples to argue that the putative class of indeterminate verbs of motion in Common Slavic could be interpreted as Manner verbs.

Verbs of motion in Slavic languages

Greenberg uses comparative historical phonological analysis to discuss indeterminate imperfective verb formation in his paper entitled PIE inheritance and word-formational innovation in Slavic motion verbs in -i-. He examines verb formation patterns such as the uniquely Slavic *jazditi ‘to go’ and *letěti ‘to fly’, which are analyzed as including a verbal root, a manner suffix, and a verb-class suffix. These putative indeterminate imperfectives are innovative in form precisely because they add a manner meaning to the range of lexical meanings available to Proto-Slavic motion verbs. Greenberg’s analysis is thus in agreement with Dickey’s in arguing that unprefixed imperfectives in Common Slavic originated as manner-of-motion verbs.

3.2

Synchronic approaches to aspect

The place of motion verbs within the aspectual system of Russian is a hotly debated topic. Unlike the majority of Russian verbs, which form aspectual pairs, simplex motion verbs are usually said to have three distinctions: two imperfectives (the determinate and the indeterminate) in addition to the perfective (Isachenko 1975; Zaliznjak & Shmelev 2000; Forsyth 1970). The significance, frequency, function, and usage of various aspectual encodings of motion verbs, as well as their very “deviance”, have been questioned. The articles in this section utilize typological, semantic, functional and cognitive approaches to explore the aspectual properties of motion verbs. Janda debates the status of motion verbs as “exceptional” in the aspectual system of Russian. Motion verbs have been viewed as exceptional since they have two stems, the determinate and the indeterminate; in addition, the prefixation of indeterminate imperfectives as a rule yields imperfective (rather than the expected perfective) verbs such as uxodit’-imp ‘to leave’ formed from xodit’-ind.imp ‘to walk.’ Using data collected from the Russian National Corpus, Janda shows that not just imperfectives, but also a variety of perfectives can be derived from indeterminate stems, such as iznosit’-pf ‘to wear out’ from nosit’-ind.imp ‘to wear’. Examining these perfectives as “exceptions to an exception”, Janda applies her Cluster Model of Russian aspect to argue that motion verbs are not only not exceptional, but also actually rather prototypical in the system of Russian aspect: these verbs have generalized the semantic distinction of completable versus noncompletable actions at a lexical/morphological level of the indeterminate/determinate distinction. Kagan’s article aims to offer a unified semantic account of the indeterminate imperfective. Kagan treats the indeterminate imperfective as the default aspect in that it can be compatible with virtually any aspectual interpretation.

˘

˘

Victoria Hasko and Renee Perelmutter

One potential difficulty with this approach lies in the fact that the indeterminate imperfective is not routinely used to encode single motion events; it is usually associated with multiple motion events with generic, habitual, and iterative meanings. To show that the indeterminate imperfective can indeed encompass any aspectual meaning, Kagan analyzes single motion event sentences, such as odin raz plaval-ind-imp bukval’no naperegonki s akuloj ‘[I] once swam while being chased by a shark’, showing that indeterminate imperfective verbs do not necessarily imply event plurality. Kagan devotes a substantial part of the article to the analysis of “annulled action”, or ‘there-and-back-again’, meaning of the indeterminate imperfective, showing that the “there-and-back-again” part of the journey is not necessary for the indeterminate imperfective to be used. Finally, she argues that restrictions placed on the usage of the indeterminate imperfective stem from the competition between this aspect and the perfective. Perelmutter uses statistical frequencies and a corpus of examples gathered from the open web to examine the behavior of Russian motion verbs under negation. Negated motion constructions are found to be less detailed than their affirmative counterparts with respect to manner and path of motion. Specific manner verbs (such as kovyljat’ ‘to hobble’) and motion verbs that specify movement along the path rather than towards a goal (such as obplyvat’ ‘to swim around’) are less likely to appear under negation. In addition, the frequencies of various aspectual markings differ in affirmative and negative clauses. It has been commonly assumed in the literature that imperfective is the default aspect under negation; however, this is not supported by the data. Perelmutter shows that aspectual choice depends on the semantic meanings encoded by the various negated constructions; aspectual choice often depends on the relation between the person who fails to arrive and an observer of motion at the goal of the motion trajectory.

3.3

Typological approach to the study of Slavic verbs of motion

While Slavic languages are placed within the group of satellite-framed languages according to Talmy’s typology (Talmy 1985, 2000), the chapters in this section exemplify cases of both inter-typological variation that exists between non-Slavic and Slavic languages and notable intra-typological differences among Slavic languages. The impact of the typological variation on lexicon, syntax, and discourse, as well as such applied fields as translation studies and second language acquisition, is discussed at length in each of the papers with the goal of more accurately positioning the Slavic languages along the typological continuum. Hasko examines intra-typological differences that exist in the semantic composition of verbs of motion in Russian and another satellite-framed language,

Verbs of motion in Slavic languages

English. This study draws on the elicited speech of adult monolingual speakers of Russian and English as the participants filter their perceptions of visuallyportrayed motion scenes into verbalized events during a communicative task of story-telling. Corpus-supported findings allow Hasko to demonstrate that although English has traditionally been viewed as the prototypical satellite-framed language due to its rich manner lexicon, the verb-of-motion repertoire in Russian is characterized by an even wider range of Manner nuances and higher frequency of encoding Manner, along with additional semantic categories that can also be encoded by the verb internally such as Path, aspect, and unidirectionality/nonunidirectionality of motion. She argues that although broad inter-typological differences provide the basis for highlighting systematic distinctions in how speakers of satellite-framed and verb-framed languages linguistically carve out the conceptual domain of motion, finer cases of intra-typological variation are necessary for advancing and fine-tuning the typology. A similar investigation of the manner-of-motion lexicon is conducted by Kopecka with regard to Polish verbs. Drawing on a corpus of written prose, Kopecka argues that Polish does not exploit the satellite-framed pattern as productively as English. She shows that the lexical repertoire of Polish manner verbs is both smaller and less fine-grained that that of English, and that the combination of manner verbs with path satellites in directed motion constructions is more restricted in the former. Kopecka calls for finer intra- and cross-linguistic research of motion structures, arguing that the diversity of manner verbs and their morphosyntactic combinability cannot be predicted on the basis of the satellite-framed pattern alone and that consideration of other factors independent of motion-event typology may contribute significantly to the semantic granularity of motion expression in a given language. The position of Serbo-Croatian within Talmy’s typology of motion expressions is addressed by Filipović. Using electronic corpus data and extensive dictionary data, she describes the distinguishing characteristics of the motion domain in Serbo-Croatian and argues that prefixes play a crucial role in shaping the discourse of the language. Filipović details multiple spatio-temporal levels of meaning at which prefixes operate and illustrates how prefixes govern the expression of manner, path, and temporal characteristics of motion events. Specifically, she describes the role of prefixes in the morphosyntactic processes of combinatory potential and morphological blocking, which, she opines, determine such central characteristics of Serbo-Croatian motion talk as verb choice (path vs. manner), degree or presence/absence of path and manner elaboration, and frequency of manner verbs in discourse. She concludes by citing implications of her findings for the fields of translation and second language acquisition.

˘

˘

Victoria Hasko and Renee Perelmutter

Nikitina discusses the alternation in prepositional phrases that mark endpoints of motion in Russian. While goal arguments of motion verbs have to be marked with directional prepositional phrases (such as sobaka bežit v prixožujuacc ‘the dog is running into the hall’), some lexical verb classes allow the endpoint to be marked with either a locational or prepositional phrase (postav’ vazu na stol-acc / na stole-prep ‘put the vase on the table’). The verb classes that allow both markings include change of position verbs, such as sadit’sja ‘to sit down’, and change of state verbs, such as prjatat’ ‘to hide’. Nikitina argues that if an endpoint of motion is encoded by a locational phrase, the event of motion is inferred rather than expressed overtly. In addition to addressing this variation, the paper raises interesting questions regarding the typology of motion expressions: while the existing distinction of satellite-framed vs. verb-framed typology accounts for the directional encoding of motion endpoints, it does not fully account for instances in which the endpoint of motion is not overtly encoded. Several papers conduct in-depth analyses of the idiosyncratic semantic characteristics of select Slavic verbs of motion or sub-domain of motion. Rakhilina provides a detailed contrastive account of the semantic variation parameters within the domain of rotation in Russian and Polish. She uses corpus-based methodology to focus on the semantics and usage of five Russian and five Polish manner verbs of rotation, four of which are etymological cognates. Her fine-grained analysis illustrates that even close and genetically-related languages, such as Russian and Polish, might show notable variation in comparable semantic fields by privileging different parameters for lexical distinction. Even when the distinctive parameters in both languages seem similar or identical, this may not be the case because the linguistic weight accorded to individual values may vary. In the case of verbs of rotation, for example, the speed of rotation is relevant for Polish verbs, but not for their Russian counterparts. Similarly, Koptjevskaja-Tamm, Divjak, & Rakhilina undertake a fine-grained typological examination of another semantic sub-domain of motion manner, that of aquamotion. Their corpus-based comparison of verbs denoting aquamotion in the Slavic languages of Russian and Polish with their counterparts in the Germanic languages of Dutch, English and Swedish demonstrates that closely related languages show significant typological differences in their aquamotion systems. They offer an explanation of the historical processes that can create, modify, and/ or obliterate lexical variation in individual semantic domains of motion manner and beyond. Both of the aforementioned papers underscore the importance of cross-linguistic the importance of lexical typological description, on the basis that it may be more powerful in revealing inter-linguistic diversity than analysis based solely on the genetic classification of languages.

Verbs of motion in Slavic languages

Nesset adds to the semantic analyses of the verb-of-motion repertoire in Slavic languages. Specifically, he disambiguates the semantic behavior of the Russian verb idti, which typically encodes unidirectional motion on foot but can also be used in a number of metaphorical senses with an inanimate motion figure. By drawing on the notions of prototype, anthropocentrism, embodiment, and metaphor as grounded in cognitive linguistics, Nesset argues that idti can be used as a generalized motion verb in metaphors due to its unique position with respect to other verbs of motion in Russian (i.e., due to the fact that idti renders prototypical human motion). He further explicates the process of partial source-domain mapping, which takes place when idti is used metaphorically (specifically, the exclusion of the specification “motion on foot”), and argues for an affinity between generalized, metaphorical idti and “goal-oriented” contexts. Nesset’s paper demonstrates the value of the cognitive approach to understanding the linguistic categories we use to speak and reason about metaphorical motion and other conceptual domains. The connection between intra- and inter-typological variability and the field of second language acquisition is mentioned in several papers (Hasko; Filipović) and is examined at length by Gor, Cook, Malyushenkova, & Vdovina. The authors describe a series of experimental studies aimed at determining factors that could lead to different learning outcomes/levels of control over Russian verbs of motion in perception and production by late second language learners and heritage language learners of Russian. Overall, the results of the study lend support to the view that the system of Russian verbs of motion is not fully acquired by even highly proficient second language learners of Russian, both early and later starters. The problems encountered by the learners are traced back to intra-linguistic typological variability of motion expressions in Russian and English. The study provides a detailed comparison of the participants’ performance, interpreted within the image-schematic framework developed in cognitive linguistics, and outlines hypothesized sources of difficulty in the use and perception of verbs of motion.

4.

Summary

This collection presents a joint, interdisciplinary enterprise into the intriguing domain of motion meanings, informed by a multiplicity of approaches and perspectives. It offers a comprehensive and up-to-date reference for those interested in the study of Slavic motion verbs, their unique semantics, structure, and aspectual behavior. The multifaceted discussion of how motion meanings are encoded in Slavic languages additionally propels the field of typological linguistics forward, as the volume situates the discussion of the semantic categories

˘

10

Victoria Hasko and Renee Perelmutter

idiosyncratic to Slavic languages within a broader framework of typological research. This integration of the two prolific research paradigms (Slavic linguistics and typological linguistics) has been long missing and will hopefully spur a new generation of inter- and intra-typological studies informed by the insights that this collection provides.

Bibliography Aurnague, M., Hickmann, M. & Vieu, L. 2007. The Categorization of Spatial Entities in Lan‑ guage and Cognition. Amsterdam: John Benjamins. Bavin, E. 2004. Focusing on “where”: An analysis of Warlpiri frog stories. In Relating Events in Narrative: Typological and Contextual Perspectives, S. Strömqvist & L. Verhoeven (eds.), 17–35. Mahwah NJ: Lawrence Erlbaum. Berman, R. & Slobin, D. I. (eds.). 1994. Relating Events in Narratives: A Crosslinguistic Develop‑ ment Study. Hillsdale NJ: Lawrence Erlbaum. Bowerman, M. & Choi, S. 2001. Shaping meanings for language: Universal and languageÂ�specific in the acquisition of spatial semantic categories. In Language Acquisition and Con‑ ceptual Development, M. Bowerman & S. C. Levinson (eds.), 475–511. Cambridge: CUP. Bowerman, M. & Levinson, S. C. (eds.). 2001. Language Acquisition and Conceptual Develop‑ ment. Cambridge: CUP. Cadierno, T. 2004. Expressing motion events in a second language: A cognitive typological perspective. In Cognitive Linguistics, Second Language Acquisition, and Foreign Language Pedagogy, M. Achard & S. Neimeier (eds.), 13–49. Berlin: Mouton de Gruyter. Forsyth, J. 1970. A Grammar of Aspect: Usage and Meaning in the Russian Verb. Cambridge: CUP. Gumperz, J. J. & Levinson, S. C. (eds.). 1996. Rethinking Linguistics Relativity. Cambridge: CUP. Han, Z.-H. & Cadierno, T. (eds.). Forthcoming. Linguistic Relativity in Second Language Acqui‑ sition: Evidence of First Language Thinking for Speaking. Clevedon: Multilingual Matters. Hasko, V. 2009. The locus of difficulties in the acquisition of Russian verbs of motion by highly proficient learners. In Special Forum on Teaching and Learning Russian Verbs of Motion, V.Â€Hasko (ed.), Slavic and Eastern European Journal 53(3): 360–385. Hickmann, M. & Robert, S. (eds.). 2006. Space in Languages: Linguistic Systems and Cognitive Categories. Amsterdam: John Benjamins. Inagaki, S. 2002. Japanese learners’ acquisition of English manner-of-motion verbs with locational/directional PPs. Second Language Research 18: 3–27. Isachenko, A. V. 1960. Grammaticheskii stroi russkogo iazyka v sopostavlenii s slovatskim: Mor‑ fologiia. Bratislava: The Slovak Academy of Sciences. Isachenko, A. V. 1975. Die russische Sprache der Gegenwart. München: Max Hueber Verlag. Levinson, S. C. & Wilkins, D. (eds.). 2006. Grammars of Space: Explorations of Cognitive Diver‑ sity. Cambridge: CUP. McNeill, D. 2005. Gesture and Thought. Chicago IL: University of Chicago Press. Miller, G. A. & Johnson-Laird, P. N. 1976. Language and Perception. Cambridge MA: Belknap Press of Harvard University Press.

Verbs of motion in Slavic languages

Özyürek, A. & Kita, S. 1999. Expressing manner and path in English and Turkish: Differences in speech, gesture, and conceptualization. In Proceedings of the Twenty First Annual Con‑ ference of the Cognitive Science Society, M. Hahn & S. C. Stoness (eds.), 507–512. Mahwah NJ: Lawrence Erlbaum. Pascal, B. 2007. Pensées. Sioux Falls SD: NuVision. Schenker, A. 1996. Dawn of Slavic. New Haven CT: Yale University Press. Shay, E. & Seibert, U. (eds.). 2003. Motion, Direction and Location in Languages: In Honor of Zygmunt Frajzyngier. Amsterdam: John Benjamins. Slobin, D. I. 1996. From “thought and language” to “thinking to speaking”. In Rethinking Lin‑ guistic Relativity, J. Gumperz & S. Levinson (eds.), 70–96. Cambridge: CUP. Slobin, D. I. 2003. Language and thought online: Cognitive consequences of linguistic relativity. In Language in Mind: Advances in the Investigation of Language and Thought, D. Gentner & S. Goldin-Meadow (eds.), 157–191. Cambridge MA: The MIT Press. Slobin, D. I. 2005. Relating events in translation. In Perspectives on Language and Language De‑ velopment: Essays in Honor of Ruth A. Berman, D. Ravid & H. B. Shyldkrot (eds.), 115–129. Dordrecht: Kluwer. Stilman, L. 1951. Russian Verbs of Motion: Going, Carrying, Leading. New York NY: King’s Crown Press. Strömqvist, S. & Verhoeven, L. (eds.). 2004. Relating Events in Narrative: Typological and Con‑ textual Perspectives. Mahwah NJ: Lawrence Erlbaum. Sussex, R. & Cubberley, P. 2006. The Slavic Languages. Cambridge: CUP. Talmy, L. 1985. Lexicalization patterns: semantic structure in lexical forms. In Language Typol‑ ogy and Syntactic Description, T. Shopen (ed.), 57–149. Cambridge: CUP. Talmy, L. 1991. Path to realization: A typology of event conflation. Proceedings of the Berkeley Linguistic Society 17: 480–519. Talmy, L. 2000. Toward a Cognitive Semantics, Vol. I & II. Cambridge MA: The MIT Press. Timberlake, A. 2004. A Reference Grammar of Russian. Cambridge: CUP. Verhoeven, L. & Strömqvist, S. (eds.). 2001. Narrative Development in a Multilingual Context. Amsterdam: John Benjamins. Vinogradov, V. V. (ed.). 1960. Grammatika russkogo iazyka. Moscow: Izd-vo Akademii nauk SSSR. Zalizniak, A. A. & Shmelev, A. D. 2000. Vvedenie v russkuiu aspektologiiu. Moskva: Iazyki rusÂ� skoi kul’tury.

11

part i

Diachrony of motion expressions

chapter 1

Clause and text organization in early East Slavic with reference to motion and position expressions Sarah Turner

University of Waterloo

This article presents a pragmatic analysis of constituent order in clauses containing intransitive verbs of motion and position drawn from a range of early East Slavic sources. The influence of context on constituent order accounts only partially for the diversity of syntactic patterns attested, and the article suggests that functional descriptions of clause patterns (e.g., ‘presentational’) used in work on modern languages are of doubtful value in the study of pre-modern material. The analysis reveals some notable ways in which conventions of text organization observed by early writers differ from those of modern texts, and the article concludes that these conventions of text organization are themselves a significant obstacle to elucidating general principles of clause organization in early East Slavic.

1.

Aims and background

Intransitive verbs of motion and position are of interest for the study of constituent order because in many languages these verbs are found in two distinct types of clause in which the major constituents appear in different orders relative to one another. Such is true of many modern Slavic languages, including Bulgarian (Dyer 1992:â•›82–83), where the loss of case might have been expected to give rise to a fixed order of syntactic constituents. These groups of verbs have also attracted attention in non-Slavic languages. In English, for instance, they figure prominently among the verbs which are used in the locative inversion construction, and it has been argued that this construction cannot be understood with reference to the syntactic properties of the verbs as such, but is rather to be considered a pragmatically controlled discourse phenomenon (Levin & Rappaport Hovav 1995:â•›215–277).

16

Sarah Turner

This article investigates constituent order in clauses containing intransitive verbs expressing motion and position in a diverse range of pre-modern East Slavic sources and considers the extent to which the distribution of differently ordered clauses can be accounted for pragmatically. Previous studies of pragmatic aspects of early East Slavic clause organization such as Bogoljubova (1970) have examined just single sources. The clauses analysed in the corpus additionally include an overt subject and a constituent (X) which denotes the destination, origin or location of the motion, or the location of the subject. In Contemporary Standard Russian (CSR) two patterns of organization for clauses comprising a subject, a verb and an X constituent are particularly common, SVX as exemplified in (1), and XVS, as in (2): (1) a. b.

Ivan pošel v školu Ivan went to school ‘Ivan has gone to school’ Kniga ležit na stole book lies on table ‘The book is on the table’

(2) a. b.

V komnatu vošla devuška in room entered girl ‘A girl came into the room’ Na stole ležit kniga on table lies book ‘There’s a book on the table’

When the intonation centre falls on the final constituent, SVX clause patterns such as those in (1) are typically taken to have a theme-rheme structure ST/(VX)R (e.g., Adamec 1966:â•›57–59). There is less consensus about the analysis of clauses such as those in (2). Adamec (1966:â•›65) treats X as a given, thematic constituent, without, however, providing any explicit contextual justification for doing so. More recently, Grenoble (1998:â•›161–163) has differentiated clearly between X(V)S orders in which the X constituent is explicit in the context and is therefore thematic, and X(V)S orders in which the X constituent cannot be considered thematic. She describes the former as ‘informational’, just as the SVX-ordered sentences under (1) are, and the latter as a presentational word order which is commonly found at the start of stories, as in (3): (3) a.

V bol’ničnom dvore stoit nebol’šoj fligel’, okružennyj in hospital yard stands small outbuilding surrounded celym lesom repejnika, krapivy i dikoj konopli entire forest burdock nettles and wild hemp

Clause and text organization in early East Slavic

b.

‘In the courtyard of the hospital there is a small outbuilding surrounded by an entire forest of burdock, nettles and wild hemp.’ (Čexov [1977]:â•›72) V gorod D., v otdel’nom kupe pervogo klassa, pribyl na in town D in separate compartment first class arrived on gastroli izvestnyj čtec i komik g. Feniksov-Dikobrazov 2-j. tour famous reciter and comic Mr Feniksov-Dikobrazov 2nd ‘The famous reciter and comic Mr Feniksov-Dikobrazov the Second arrived on tour in town D. in a separate compartment of the first-class (Čexov [1976]:â•›175) carriage.’

Thus a single syntactic pattern can either give additional information about a referent which is already under discussion, or it can introduce into the discourse a new referent which is likely to become the theme of subsequent clauses. A third type of clause is also identified in which the falling, characteristically rhematic intonation is placed on the first syntactic constituent, rather than on the last, as in (1) and (2) above. This mode of discourse organization is found commonly in colloquial speech, and it is widely described as expressive (e.g., Grenoble 1998:â•›164). Other orders of constituents are also possible in modern Russian, but they are not found as frequently in the standard language as types (1) and (2). The SXV and XSV orders shown in (4) are typically described as colloquial or expressive variants of (1) and (2) respectively, and they differ from their neutral counterparts in the placement of the intonation centre on the middle constituent: (4) a. b.

Grigorij po gastronomam begaet Grigorij around grocery.shops runs ‘Grigorij is running around the grocery shops.’ Ko mne brat priexal to me brother arrived ‘My brother has come to stay with me.’

(5) a. b.

Pošel starik v les went old.man to forest ‘{The/An} old man went into the forest.’ (Adamec 1966:â•›66) Stojal nedaleko ot doma odinokij dub stood not-far from house lone oak.tree ‘Not far from the house there stood a lone oak tree.’ (Adamec 1966:â•›68)

(Adamec 1966:â•›71)

(Adamec 1966:â•›72)

Various interpretations of clause patterns such as those in (5) in which the verb precedes the other constituents have been put forward, for example Adamec (1966:â•›66–68), Janko (2001:â•›201–206) and Yokoyama (1986:â•›284–288, 295). Whatever their merits in relation to modern Russian, for a number of conceptual

17

18

Sarah Turner

reasons these interpretations of verb-initial clauses cannot easily be applied to pre-modern material (Turner 2005:â•›45–47, 104–105). All the constituent orders described above are found in the early medieval period also. However, in some early sources the problematic VSX and VXS orders occur much more frequently than in the modern language, and in some sources the XVS order which is common in CSR is attested rarely, if at all. In the description of the data which follows, constituents are marked as thematic or rhematic with strict reference to the contextual relationships they exhibit. From the modern perspective, this approach may be felt to be artificial: after all, it has long been recognized that the intuitively satisfactory concept of the theme cannot be strictly equated with contextual features, but is ultimately defined by the speaker’s or writer’s communicative intentions (e.g., Kovtunova 1980:â•›193–194). However, there is much to be said for offering as precise a description of the early material as possible. The study of word order in the Pskov First Chronicle mentioned earlier (Bogoljubova 1970) is compromised by its tendency to make the analyses fit the straightforward theme-rheme model of the clause which the author borrows from the modern language, regardless of the actual data; the results can be puzzling and inconsistent. When there is doubt about the applicability of the model to the variety of language under consideration, it seems advisable to prioritize descriptive accuracy over interpretative nuance and conceptual assumptions. There can anyway be difficulties in assigning a contextual label to a constituent in a completely principled manner: in a long text, decisions about the persistence of thematic status are essentially arbitrary, for instance. Nevertheless, context remains the most objective criterion on which to develop a pragmatic interpretation of the material. Particular instances of analytical difficulty are discussed as appropriate in Section 5.

2.

Sources

All the texts in the corpus are thought to have been composed prior to the end of the fourteenth century, though in the interests of trying to make the subÂ�classification of the material outlined below viable, some texts whose manuscript tradition postdates that time are included alongside texts attested from the earlier period. Where possible, editions which preserve the original orthography and punctuation have been used, but for (d)–(f) it has been necessary to use normalized versions of the texts.

Clause and text organization in early East Slavic

Žitie prepodobnaago otьca našego theodosija igumena pečerьskago (‘The life of our venerable father Theodosius, abbot of the Caves Monastery’). From the Uspenskij sbornik (Kotkov 1971:â•›71–135). b. BG: Sъkazanije i strastь i poxvala svjatuju mučeniku borisa i glěba; Sъkazanije čjudsъ · stoju strastьrpьcu xvou · romana · i dvda (‘The story and passion and encomium of the holy martyrs Boris and Gleb’; ‘The story of the miracles of Christ’s holy passion-sufferers Roman and David’). As above (Kotkov 1971:â•›42–58; 58–71). c. Nov: the Novgorod First Chronicle in the Synodal copy up to and including 118v, where the hand changes. Citations are given according to the reproduction of the manuscript in Dietze (1971). d. ID: Žitьe i xoždenьe danila rusьskyja zemli igumena (‘The life and pilgrimage of Danil, an abbot of the land of Rusь’) (Proxorov 1980:â•›24–115). e. SN: Ot strannika stefanova novgorodca (‘By the traveller Stefan of Novgorod’) (Speranskij 1934:â•›50–59). f. Const: Skazanie o svjatyx mestex, o kostjantinegrade i o svjatyx moščex spasšixsja vo ierusalimě, a sobranyx kostjantinom caremь v naricaemyi carьgrad (‘The story of the holy places, of Constantinople, and of the holy relics saved in Jerusalem and collected by Emperor Constantine in Constantinople’) (Speranskij 1934:â•›128–137). g. BB: Novgorod birch-bark letters from before the end of the fourteenth century (the first four sections of Zaliznjak 2004). These sources provide little material of relevance to the study but are included because they offer the most reliable insight into the syntactic possibilities of the early East Slavic vernacular. It would be inadvisable to base any conclusions on the statistical data presented in Section 4, but the presence or absence of various patterns of clause organization is nevertheless a matter of interest. a. Th:

Nov and ID provide most of the material for the study. The data from sources (a)–(c) are most typically clauses in which the protagonists’ movements are recounted. Sources (d) and (e) also contain a strong, spatially oriented narrative line, but those particular sections of the texts provide little material of relevance in the present connection, because they are written in the first person and the subject is expressed as a nominal constituent only rarely. The data drawn from sources (d)–(f) are instead typically found in passages which describe the topography of the places visited.

19

20 Sarah Turner

3.

Selection

The clauses for analysis were selected on a lexical basis. They contain verbs which fall into the two core semantic categories associated with presentational word orders in CSR: verbs of existence and verbs of motion. This latter group is further limited syntactically to exclude not only transitive verbs, but also reflexives, their frequent appearance in modern presentational clauses notwithstanding: the prosodic status of the reflexive particle in East Slavic in the period of interest (Zaliznjak 2004:â•›188–189) introduces an unwelcome additional variable. The corpus does not include imperatives, direct or indirect interrogative clauses, and relative clauses in which either the subject or the spatial expression is the relative pronoun or adverb. Occasional examples of clauses with compound tenses or with split subject phrases are set aside. If there is any doubt about whether a form of byti (‘to be’) is being used in an existential or in a copular function, it has been excluded from further consideration. Acknowledged quotations from the Bible and other translated texts are omitted, as are instances in which the division of the text into clauses is open to question. Several common idioms involving prepositional phrases have been set aside or treated specially in the classification. Finally, to simplify classification, clauses containing more than one non-consecutive X constituent have been omitted. Clauses of this degree of complexity are reasonably frequent in ID and Const, rare or unattested in the other sources. The final corpus contains 697 clauses.

4.

The data

In Tables 1 and 2 the data are categorized according to the basic arrangement of the syntactic units S, V and X. Sequences of more than one spatial expression are grouped together with single X constituents. No account has been taken of the presence or absence of other constituents such as adverbs or temporal adjuncts. Thus the tables should not be taken to indicate that all clauses in the SVX category, say, are syntactically identical. This simple classification serves merely as a convenient framework for organizing the detailed examination of the relationship between the syntactic and pragmatic characteristics of the clauses to follow. The extent of the similarity of the clauses in each category will be discussed below. Table 1 presents the raw data provided by the corpus, and Table 2 indicates the distribution within the individual sources of the various clause patterns under consideration. Percentages are rounded to the nearest whole number. Notably, some clause patterns are attested rarely, if at all, in some sources, yet are very frequent in others. Clauses with an initial spatial expression are

Clause and text organization in early East Slavic

Table 1.â•‡ Frequency of clause patterns Th BG Nov ID SN Const BB Total

SVX

SXV

VSX

VXS

XVS

XSV

Total

â•⁄ 31 â•⁄â•⁄ 8 â•⁄ 71 â•⁄ 36 â•⁄â•⁄ 5 â•⁄ 10 â•⁄â•⁄ 4 165

â•⁄ 6 â•⁄ 1 â•⁄ 5 27 â•⁄ 1 â•⁄ 3 â•⁄ 5 48

â•⁄â•⁄ 9 â•⁄â•⁄ 5 176 â•⁄ 55 â•⁄â•⁄ 1 â•⁄â•⁄ 4 â•⁄â•⁄ 7 257

â•⁄ 7 â•⁄ 8 21 17 â•⁄ 1 10 â•⁄ 2 66

â•⁄â•⁄ 3 â•⁄â•⁄ 1 â•⁄â•⁄ 0 â•⁄ 79 â•⁄ 21 â•⁄ 25 â•⁄â•⁄ 4 133

â•⁄ 2 â•⁄ 0 â•⁄ 2 11 â•⁄ 8 â•⁄ 5 â•⁄ 0 28

â•⁄ 58 â•⁄ 23 275 225 â•⁄ 37 â•⁄ 57 â•⁄ 22 697

Table 2.â•‡ Distribution of clause patterns within the sources (%) % of text

SVX

SXV

VSX

VXS

XVS

XSV

Th BG Nov ID SN Const BB

53 35 26 16 14 18 18

10 â•⁄ 4 â•⁄ 2 12 â•⁄ 3 â•⁄ 5 23

16 22 64 24 â•⁄ 3 â•⁄ 7 32

12 35 â•⁄ 8 â•⁄ 8 â•⁄ 3 18 â•⁄ 9

â•⁄ 5 â•⁄ 4 â•⁄ 0 35 57 44 18

â•⁄ 3 â•⁄ 0 â•⁄ 1 â•⁄ 5 22 â•⁄ 9 â•⁄ 0

extremely uncommon in the group of texts which focus on action (Th, BG and Nov) and are more typically found in the descriptive texts (ID, SN and Const): 94% of all the XVS clauses and 86% of all the XSV clauses in the corpus are found in this latter group of texts. In other respects, the statistics do not straightforwardly support a two-way subdivision of the corpus. Clauses in which the subject appears as the first of the constituents under consideration are proportionally more frequent in the first group of texts than the second, but within that first group they are the clearly dominant pattern only in Th (53%). At 64%, the proportion of VSX clauses in Nov far exceeds those in the rest of the corpus, but the proportion of VSX clauses in the other texts of the first group does not differ greatly from the proportions found in the more descriptive ID. In three texts, two loosely characterized as action texts and one as descriptive, a single pattern occurs in more than half the clauses collected: SVX in Th, VSX in Nov and XVS in SN. Elsewhere, and most notably in ID, there is greater diversity in the organization of motion and position expressions. Detailed analysis of the material in context will seek to account for these strikingly different frequencies of attestation.

21

22

Sarah Turner

5.

Analysis

In the headings of the subsections below, the subscript symbols T (theme) and R (rheme) refer to the contextual status of all non-pronominal constituents, following the methodological position outlined in Section 1.

5.1

S-first clauses

5.1.1 ST/(VX)R The clauses in this group attest a diverse range of pragmatic characteristics. The majority have a subject which is explicit in the recent context. In a smaller number of examples, the subject is a first-person pronoun which can be considered thematic by reference to the deictic parameters of person (Yokoyama 1986:â•›32). In Th and the later parts of Nov (particularly from 6724), the thematic subject is commonly followed by the particle že. Examples of že are sporadic elsewhere: in the earlier parts of Nov it appears after third-person pronouns, and in ID after first-person pronouns. Its use is almost exclusively limited to animate subjects, though to some extent that is an accident of the data, for inanimate subjects occur very rarely in sources (a)–(c). A wider survey of clauses from ID shows that že can be used after inanimate subjects as well. In the current corpus, however, inanimate subjects which are already present in the context are more typically accompanied by demonstrative modifiers than by particles: (6)

[Tъgda že velikii nikonъ i drougyi čьrnьcь stgo miny then PART great Nikon and other monk Saint Mina manastyrja {...} tako otъidosta] ... Velikyi že nikonъ otъide vъ Monastery so left great PART Nikon left to ostrovъ tьmoutorokanьskyi island Tmutarakanь [Then the great Nikon and another monk of the Monastery of S. Mina {…} went away] … The great Nikon went to the island of Tmutarakanь.’ (Th 35b. 1)

. See also Kardaševskij (1948) for observations about the complementary distribution of clauses containing že and an initial conjunction i in the Primary Chronicle.

Clause and text organization in early East Slavic

(7)

[a ottuda estь blizь rěka velika] i ta rěka poide izь and from.there is near river great and that river starts.out from ezera genisaritьskago lake Galilee ‘[and not far from there is a great river] and that river rises in the Sea of Galilee.’ (ID 94)

The X constituents in this subsection of material are usually prepositional phrases, occasionally simple inflected noun phrases, and, very occasionally, adverbs of place. In some instances, they are prototypically rhematic: neither available in nor inferable from the context, as in (8). In other instances, the X constituent may be present or predictable in the context; it may be a pronoun or may be modified by a demonstrative. However, the presence of a subject which is still more salient in the context or situation makes a rhematic reading seem more plausible than a thematic one, as in (9). If the verb phrase contains more than one X constituent, or an additional constituent of another type, the later constituent can often be construed as the new information which forms the writer’s principal point, as in (10): (8)

[grad že samaria velikъ estь velmi i obilenъ estь vsěm dobrom] city PART Samaria great is very and abundant is all good grad-ot samarija stoitъ meži dvěma gorama vysokima city.DEM Samaria stands between two mountains high ‘[the city of Samaria is extremely big and has plenty of all good things.] The city of Samaria stands between two high mountains.’ (ID 84)

(9)

togda azъ xudyi nedostoinyi v tu pjatnicju vъ 1 čas dni idox kъ then I miserable unworthy on that Friday at 1 hour day went to knjazju tomu balъdvinu prince that Baldwin ‘Then I, worthless wretch, on Friday in the first hour, went to that Prince Baldwin.’ (ID 106)

(10)

azъ že poklonivsja grobu gospodnju i ključarevi {...} izidox I PART having.bowed tomb Lord’s and sacristan exited iz groba svjatago s radostiju velikoju from tomb holy with joy great ‘Having bowed down at the Lord’s tomb and bowed to the sacristan {…}, I left the Holy Sepulchre with great joy.’ (ID 114)

. These constituents are not always referential, and thus they cannot always be described in terms of contextual affiliations.

23

24

Sarah Turner

Thus a large proportion of the clauses in this category exemplify the ST/(VX)R pattern described earlier for CSR and, in the terms set out there, may be said to have an informational function. Examples can be found in all of the texts, including the birch-bark letters, where one of the three SVX orders is of this type.

5.1.2 (SVX)R In a smaller proportion of the SVX clauses (approximately 22 out of 165), there is no contextual basis for analysing either of the nominal constituents as thematic. Occasionally, the presence of a particle may suggest that a thematic reading of the subject is nevertheless to be preferred: (11) posъlanii že priidoša otъ stopъlka na lьto nočь envoys PART arrived from Svjatopolk to Lьto night ‘Envoys came from Svjatopolk to the Lьto by night.’ (12) a doroganici ti šli vъ gorodo and people.of.Dorogani PART went to town ‘And the people of Dorogani have gone to town.’

(BG 11b. 13)

(BB B91 550)

Elsewhere, however, it is a matter of judgement whether the subject can be regarded as a so-called imposed theme (Yokoyama 1986:â•›59–61), or whether the clause is to be taken as a pragmatically undivided whole. Some such clauses have a non-specific subject (13); in others, the subject is a proper noun ((14) and (15)): (13) 7 čelověkъ ili 8 vъstavjat na pleča odinomu čelověku 7 people or 8 stand on shoulders one man ‘Seven or eight people stand on the shoulders of one man.’

(SN 141v)

(14) kjurъ mixailъ · pobeže peredi is pronьska Lord Mixail ran earlier from Pronьsk ‘Lord Mixail fled earlier from Pronьsk.’

(Nov 74r. 3)

(15) litva vъstala na korělou Lithuania rose against Karelia ‘Lithuania has risen against Karelia’.

(BB A6 590)

Again, clauses of the contextual type (SVX)R are distributed throughout the entire corpus. The greatest frequency of examples is found in Nov. That observation is to some extent the product of the analytical procedures followed: contextual relationships were assessed only within units of text bounded by the temporal expressions characteristic of the chronicles (Vъ lět · 6525 · ‘In the year 6525’; Tomь ž let · ‘In the same year’ etc.). For the early part of Nov, which consists of laconic paraphrases of non-Novgorodian chronicles, there may be a case for allowing for

Clause and text organization in early East Slavic

thematic continuity across the year entries, as the chronicler picks out information about the few figures of greatest concern to him. Example (15) appears to be a complete text in its own right (Zaliznjak 2004:â•›244), but since there is no way of knowing whether it formed part of a more extensive exchange of information, or whether it really is the bald statement it appears to be to the modern reader, there no basis for any conclusive pragmatic analysis. As Gippius (2004) has argued, it may well be misleading to apply modern assumptions about the purpose and conventions of written communication to pre-modern material. Furthermore, even if it could be shown that the clause was not prompted by any preceding exchange, there would still be scope for debate about how its constituents should be interpreted. Studies of word order in CSR which have drawn primarily on creative writing have tended to analyse clauses with pre-verbal subjects in discourse-initial environments as stylistically motivated T/R structures (Kovtunova 1980:â•›193), whilst a more recent study (Janko 2001:â•›149–153) has shown that at least in some registers of modern Russian a newsworthy subject in a newsworthy clause may be realized with a rhematic intonation, particularly if it contains several post-verbal constituents. Clearly, pre-modern written material offers no basis for furthering the discussion of how modern written texts come to be realized in speech. Taken in their own right, however, the examples in this section prompt two observations. Firstly, the same indeterminacies in the relationship between the syntactic and pragmatic organization of the clause seem to have existed then as now: the syntactic pattern cannot be taken as a reliable guide to the pragmatic properties of the clause at either period. That said, in pre-modern texts of the sort examined here it does appear to have been unusual for clauses containing an intransitive verb of motion or position to have the subject as the first of the major constituents in circumstances of contextual independence: the total number of possible (SVX)R clauses in the corpus is anyway relatively small, and those with definite subjects might actually be the result of an unintended hiatus between the writer’s perspective on the communicative situation and the reader’s perspective on the text.

5.1.3 (SV)RXT In approximately 20 instances in the corpus a subject which is not predictable from the preceding context stands as the first major constituent in the clause, and a prepositional phrase containing a noun or pronoun which is already present in the context follows the verb. Analogous patterns of clause organization in CSR are described (e.g., Adamec 1966:â•›72) as colloquial, for they are often found in everyday speech, and as expressive, though it is not clear whether this latter

25

26 Sarah Turner

interpretation alludes to a general perception that speech is of itself somehow more vivid than writing; to stylistic considerations which at particular moments may prompt a writer to incorporate direct speech into his narrative; or simply to an assumption that a statistically minor pattern must be qualitatively marked in some way. Thus the notion of expressiveness as used for the modern language is too vaguely defined to be of use for medieval material, when not even the inscrutable evidence of a native speaker’s intuition can support it. Furthermore, the current corpus offers no secure grounds for assessing the status of clauses with this discourse structure in the colloquial language of earlier periods. The texts which may be assumed to come closest to representing that variety of language – the birch-bark letters – anyway provide very few examples relevant to this study, and there are no clauses of this type among them. Although it is pertinent to consider whether any particular factor or group of factors links the examples in the (SV)RXT group and differentiates them from the XT(VS)R examples described in Section 5.3.1 below, any interpretations of the difference between the patterns must be based on observations of the medieval data themselves rather than argued from general assumptions, be they about the Russian language in particular or about wider principles of synchronic linguistics. That said, on a first inspection many of the examples in the (SV)RXT category do seem to lend themselves to an expressive interpretation. Several of them occur in close proximity to the particle se (‘behold’), when the context clearly suggests an element of surprise on the part of the speaker(s): (16) i se pьsъ čьrnъ sta predъ mnoju and PART dog black stood before me ‘And, behold, a black dog stood before me.’ (17) se bo mьnitь sja namъ prividěnije bystь vъ crkvi PART for seems REFL us apparition was in church ‘For it seems to us that there was an apparition in the church.’

(Th 44a. 14)

(Th 47a. 1)

However, in other examples in which se appears either the syntactic organization of the clause differs, as in (18), where the rhematic subject and verb appear in the opposite order from that seen in (16) and (17), or both the syntactic and the pragmatic organization differ, as in (19), where the thematic X constituent interrupts the rheme: (18) i se priidoša stranьnici vъ gradъ tъ and PART arrived travellers in town that ‘And travellers arrived in that town.’

(Th 28c. 4)

Clause and text organization in early East Slavic

(19)

i se v jedinou noščь tьmě soušči velicě priidoša na nja and PART on one night dark being great arrived against it razboinici robbers ‘And one night when it was very dark robbers moved against it.’ (Th 46c. 23)

Thus no single clause pattern is unambiguously associated with these potentially expressive environments. The particle se is of limited usefulness as a diagnostic of expressiveness, for its appearance is restricted to only the highest-style texts, primarily Th. A further point of contrast between the two sets of examples just given is that the former – with the full inversion of the theme and the rheme – both occur in direct speech, whilst the latter are found in the authorial narration. As indicated in the opening paragraph of this section, it is problematic in principle to associate speech with expressiveness, but it is worth noting in passing that of the seventeen examples of direct and indirect speech found in the texts studied here, only one of them is organized in a way which would be considered stylistically neutral in CSR. The remaining sixteen clauses have diverse combinations of syntactic and pragmatic characteristics and include both rheme-theme ordered clauses and rheme-themerheme ordered clauses. Although the use of direct and indirect speech is again limited, this time largely to Th and Nov, there would seem to be at least the beginnings of a case for associating deviations from a straightforward theme-rheme order with some sort of special effect, even if it is hard to specify precisely what that effect might be. However, difficulties of interpretation arise when (SV)RXT ordered clauses appear in contexts which do not offer any strong grounds for supposing an expressive effect to be intended other than the constituent order itself. Examples

. Studies of CSR such as Kovtunova (1980:â•›202–203) commonly distinguish between whole and partial inversions of the neutral pattern. Whatever the merits of this approach for CSR, it cannot be taken over directly into work on the pre-modern language: the effects attributed to the different types of inversion (e.g., ‘epic’) presuppose the existence of a system of literary styles, an assumption which stands in need of justification for the period preceding the formation of the standard language. . [ili mьniši jako] azъ otidoxъ ot vasъ ‘[or do you think that] I have left you?’ (Th 65b. 21). The modern equivalents to such clauses are the exception to the generalization that in neutral styles of CSR the intonation centre falls on the final syntactic constituent of the clause (Adamec 1966:â•›70).

27

28

Sarah Turner

(20) and (21) are taken from passages of authorial narration, rather than direct speech, and do not include any particles which suggest surprise or wonder: (20) i blgoouxanije isxožaše otъ crkve and sweet-smell emanated from church ‘And a sweet smell was emanating from the church.’ (21) i smrad isxodit ot morja togo and stench emanates from sea that ‘And a stench comes from that sea.’

(Th 46d. 10)

(ID 58)

Whilst the writers might well have wished to convey surprise in these circumstances, there are problems in arguing a case for an expressive interpretation with reference to the organization of the clause, unless it is claimed that the constituent order is chosen by analogy to more obviously expressive examples such as (16) and (17). For CSR, the expressiveness associated with rheme-theme clauses is accounted for by their relationship to neutral theme-rheme patterns. The neutral order is inverted, either wholly or partially, to produce a special effect. Yet in Th and Nov, in which the arguably expressive clauses with se or in/direct speech are concentrated, the potentially neutral variants of (SV)RXT clause patterns – XT(VS)RÂ€– are attested rarely, if at all. There are no examples of XVS patterns for clauses indicating motion or position at all in Nov. Th has just three XVS clauses, one of which itself has a rheme-theme rather than a theme-rheme structure. Rhematic subjects and thematic spatial complements are more often configured in VXS order rather than XVS order, as will be illustrated in Section 5.2.4. Therefore any paradigmatic relationships which might exist between different patterns of clause organization cannot be adequately described in terms of the straightforward inversion of the theme and the rheme, as is done for CSR. Moreover, since Th and Nov do not attest with any frequency an alternative configuration for clauses containing this combination of syntactic and pragmatic characteristics, there is some doubt about how meaningful it is to talk of stylistically motivated variants of a neutral order of pragmatic constituents at all.

5.1.4 SRXTVT/R SXV clause patterns are attested most frequently in ID, and among them the most common pragmatic structure is SR(XV)T. The examples are largely formulaic, however: 11 out of 27 take the form (i) S tu ležit/ležat. In most instances, the . The fact that several of the clauses either have verbs prefixed with iz- or have mnogo or a related form in the subject phrase is curious, but ultimately does not appear to be significant in its own right: there are many clauses with similar lexical features but different syntactic and pragmatic arrangements.

Clause and text organization in early East Slavic

clauses occur in a context in which it has already been established that the place in question is a renowned burial site, and thus there is a case for analysing the verb as thematic. Examples such as (22) are a further illustration of the problems associated with the concept of expressive word orders: (22)

[v toi peščerě ležit mati svjatago savy] i feodosieva mati in that cave lies mother Saint Sava and Theodosius’ mother tu ležit there lies ‘[In that cave lies the mother of S. Sava] and Theodosius’ mother lies there too.’ (ID 56)

Although to the modern ear this example certainly has an informal ring, it is not obviously an emphatic or emotionally charged statement; it merely suggests a relaxed approach to discourse organization, in which the writer is guided by an association of ideas. Whether the transition from XVS to SXV arises because an afterthought genuinely occurs to the author, because he is striving for a little elegant variation, or simply because he feels no normalizing pressure is beyond debate. Studies of the syntax of modern colloquial Russian (esp. Lapteva 1976) have shown that, the absence of a codified norm notwithstanding, that variety of the language is nevertheless highly normalized: it is characterized by the frequent use of constructions which are not found in CSR, yet there is a great deal of variation within those constructions themselves, and many different syntactic sub-types can be identified. A tendency to make heavy use of formulaic expressions whilst preserving a high degree of syntactic variation can be seen in early East Slavic, too. The same basic formula involving tu (‘there’) which was illustrated in (22) can appear in different orders as well, including, notably, SVX. Although there is a fairly strong tendency for adverbial X constituents to appear in pre-verbal positions, example (23) shows that there is no hard-and-fast syntactic rule to that effect: (23)

[v toi že peščerě 300 svjatyx otecь ležit ·] i svjatyi aleksandrъ in that PART cave 300 holy fathers lie and Saint Alexander ležit tu i grob magdalyni marьi i glava eja lies there and tomb Magdalene Mary and head her ‘[In that same cave three hundred holy fathers lie] and S. Alexander lies there and the tomb of Mary Magdalene and her head.’ (ID 28)

Nor are there any evident regularities in the ways the text develops which could help to explain the distribution of the clause patterns seen. Whilst example (22) appears at the end of a section of text, other SR(XV)T clauses appear mid-section in lists of burial places, and in this respect they resemble example (23), as well as variations on the formula with an (XV)TSR pattern illustrated in Section 5.3.1.

29

30

Sarah Turner

The other thematic X constituents in this subcategory of ID take the form of personal pronouns or noun phrases which include a demonstrative modifier. That is to say, the thematic analysis is invariably supported by their morphological or lexical properties, and does not proceed from an examination of the context alone. Generally speaking, in the analysis of real text it is unusual to have strong contextual grounds for analysing a non-referring expression such as a verb as a constituent of the theme, as was done for (22) and (23). Verbs are more typically rhematic. In the present corpus, occasional examples of SRXTVR patterns are found in ID and BB, where the X constituents exhibit the same properties as in SR(XV)T patterns: (24) i blagojuxanie čjudno ot nix isxodit and sweet-smell marvellous from them emanates ‘And a marvellous smell emanates from them.’ (25) a maksime · ivane · široki tu že byli and Maksim Ivan Široki there PART were ‘And Maksim and Ivan Široki were there.’

(ID 74)

(BB G53 178)

5.1.5 ST(XV)R Of the two clause patterns in which the subject appears as the first of the major constituents, the SXV pattern has a smaller proportion of thematic subjects than the SVX patterns discussed in Section 5.1.1 (approx. 38% and 70% respectively). A curiosity of the data from ID is that all four SXV examples there in which the verb is byti (‘to be’) exhibit this pragmatic structure, in addition to two clauses with different verbs. Similarly in Th, all the byti clauses fall into this category: (26)

to že blaženyi jego radi vъ velicě pečali i skъrbi PART PART blessed him because.of in great concern and distress boudjaše was ‘And on his account the blessed one used to be in great concern and distress.’ (Th 49c. 14)

The examples there are all variations on a formula, however, and there is anyway greater diversity amongst the clauses with other verbs than there is in ID.

5.1.6 Other types In the few examples of more complex clauses the main rhematic constituent is arguably none of the constituents of interest here, but another constituent of the clause entirely:

Clause and text organization in early East Slavic

(27) i snězi na nej ležat čres lěto and snows on it lie over summer ‘And snows lie on it over the summer.’

(ID 94)

As ever, some examples in the corpus fall into a grey area in which it is debatable how much latitude is to be allowed in the identification of thematic constituents: (28) i voda svjataja ot jazvъ gvozdinnyx ot nogu ego idet and water holy from wounds nails from feet his goes ‘And holy water comes out of the holes in his feet made by the nails.’ (SN 138r)

In a strict contextual approach to example (28), the clause would be analysed as rheme-only, because neither of the medial X constituents is available in the textual context. However, both are presumably part of the reader’s background knowledge of the Crucifixion, and the second medial constituent could also be analysed as thematic on the grounds of natural association. Comparison with example (24), in which the X constituent has unambiguously thematic properties, speaks in favour of the second reading.

5.2

V-first clauses

Table 2 shows that clauses in which the verb stands before its subject and is the first of the major constituents to appear are distributed unevenly through the corpus. Clauses ordered VSX are the heavily predominant pattern in Nov; they feature fairly prominently in BB, ID, BG and Th; and in the other sources they are a minor pattern. As with the subject-first clauses, the verb-first clauses demonstrate a variety of contextual characteristics. In contrast to the subject-first clauses, however, only in very few instances can the early position of the verb in the clause be understood as a reflection of its thematic status, and even these instances are not beyond dispute. For instance, example (29) from Nov can be analysed in two ways: (29)

[Togo ž lět · prisla velikyi knzь vsevolodъ · vъ novъgorodъ · rekja that PART year sent grand prince Vsevolod to Novgorod saying tak · vъ zemli vašei ratь xoditь a knzь vašь snъ moi stoslavъ thus in land yours war goes and prince your son my Svjatoslav malъ · a daju vy snъ svoi stareišii kostjantinъ {…}] Vъ to ž small and I.give you son my older Kostjantin In that PART lět · pride knjazь kostjantinъ · vsevolodicь · vnoukъ gjurgevъ · vъ year arrived prince Kostjantin Vsevolodič grandson Gjurgi’s in

31

32

Sarah Turner

novъgorodъ · mcja marta · vъ · 20 · na stgo gerasima Novgorod month March on 20 on Saint Gerasimus ‘[In the same year grand prince Vsevolod sent to Novgorod saying, ‘There is war in your land, and your prince, my son Svjatoslav, is young, so I’m giving you my older son Konstantin {…}]. In that same year prince Konstantin Vsevolodič, grandson of Gjurgij, arrived in Novgorod on March 20 on the feast-day of S. Gerasimus.’ (Nov 72r. 9)

Since Konstantin’s arrival in Novgorod is the expected outcome of Vsevolod’s instructions, the verb and its arguments may be taken as thematic, and the adjunct phrase which gives the date of his arrival may be taken as the rheme. Alternatively, the whole clause could be analysed as rhematic, as is done here: the section of the clause comprising the verb and its arguments serves to confirm that the expected event did indeed take place, and the adjunct phrase forms a secondary rheme. Such examples with possible thematic verbs are confined to Nov.

5.2.1 (VSX)R The most frequently attested pragmatic pattern in the corpus is also largely restricted to Nov. The annalistic structure of the early parts of the chronicle and the discontinuous narratives in the later parts result in a large number of clauses in which the verb and its arguments are all rhematic: (30) Vъ lět · 6634 · xodi vsěvolodъ · kъ ocju kyjevou in year 6634 went Vsevolod to father Kiev ‘In the year 6634 Vsevolod went to his father in Kiev.’

(Nov 11r. 16)

(31) Togo ž lět · ide antonъ arxjepspъ novgorodьskyi na tъržьkъ that PART year went Anton archbishop Novgorodian to Toržok ‘That same year archbishop Anton of Novgorod went to Toržok.’ (Nov 92v. 3)

Approximately 71% of VSX clauses in Nov and 51% of VSX clauses in the corpus as a whole are of this type. The few possible examples from the other literary texts involve clauses in which the X constituent is an abstract noun or non-referring expression: (32) i tečaaxou ljudije na pozorъ and ran people to spectacle ‘And people would run to see the spectacle.’

. Cf. the examples from modern Russian discussed in Janko (2001:â•›167–170).

(BG 19a. 25)

Clause and text organization in early East Slavic

Of the seven VSX clauses in the birch-bark letters, two examples occur discourseinitially and can securely be analysed as rheme-only clauses. One of them, (34), would seem to be a traditional formula: (33) prišьlь iskoupnikь is polocьka arrived freed.prisoner from Polotsk ‘A freed prisoner has arrived from Polotsk.’

(BB V15 636)

(34) estь grad · mežu nobomъ i zemleju is city between heaven and earth ‘There is a city between heaven and earth.’

(BB G63 10)

Thus in these two groups of sources the presentational discourse function attributed to (XVS)R clause patterns in CSR can be fulfilled by a different configuration of syntactic constituents. However, the VSX word order is not associated exclusively with a rheme-only pragmatic structure even in those sources, and in the rest of the corpus it shows stronger associations with the two structures described below.

5.2.2 VRSTXR The next most common subset of VSX clauses in Nov has a thematic subject in the first post-verbal position: (35)

[Tъgda ž novgorodьci poslaša sja po brat ego · po then PART people.of.Novgorod sent REFL for brother his for mьstislava vъ rous · i vъnide mьstislavъ vъ novъgorodъ · mscja · Mstislav to Rusь and entered Mstislav to Novgorod month nojabrja vъ · 1 · na stouju bezmezdьnikou kъzmy i damijana ·] November on 1 on Saints unmercenaries Cosmas and Damian A na zimou ide mьstislavъ sъ novgorodьci na čjudь and in winter went Mstislav with people.of.Novgorod against Čjudь na očelou against Očela ‘[Then the people of Novgorod sent to Rusь for his brother Mstislav, and Mstislav entered Novgorod on November 1 on the feast-day of SS Cosmas and Damian.] And in the winter Mstislav and the men of Novgorod marched against the Čjudь and the Očela.’ (Nov 42v. 10)

In most of the other sources which attest VSX word orders with any frequency, this type is the dominant pattern, heavily so in ID and in Const, where it only occurs in clauses with byti, less heavily so in Th. In ID, many of the subject phrases include a demonstrative modifier:

33

34

Sarah Turner

(36) i vxodjat rěky ty vъ ezero genisaritьskoe and enter rivers these in lake Galilee ‘And these rivers flow into the Sea of Galilee.’

(ID 94)

(37) i estь manastyr-etъ na ustьi and is monastery-DEM on estuary ‘And the monastery is on the estuary’.

(ID 54)

The first of these clauses exemplifies a textual progression characteristic of ID. The noun modified by the demonstrative is already salient in the immediately preceding clause, in this particular instance as its theme. In other instances, typically involving clauses with byti, the thematic subject phrase is co-referential with the rheme of the immediately preceding clause, and the VRSTXR clause gives precise information about the location of a place whose historical and religious significance has just been established. Example (37) is unusual for ID in that the antecedent of the demonstrative phrase is located several clauses earlier. In Nov, too, post-verbal thematic subjects are usually separated from co-referential phrases by more than one intervening clause.

5.2.3 (VS)RXT Examples of the third pragmatic structure associated with VSX clauses are rare in the large set of data from Nov; they make up a larger proportion of the material from ID; there is one secure example and three possible ones in BB; and they are proportionally most common in the small number of examples from BG, with additional examples to be found in the other text with a particularly strong influence from Russian Church Slavic, Th: (38) i stoitъ trostie veliko po rěkam těm and stands rushes great along rivers those ‘And tall rushes line those rivers.’ (39) [byl]a žaloba peredo vami i popъemi was complaint before you and priests ‘There was a complaint before you and the priests.’ (40) i se pride věstьnikъ kъ nemou and PART arrived messenger to him ‘And a messenger came to him.’ (41) i isxožaaše blgaja vonja otъ grobou jeju stoju and emanated sweet smell from tombs their holy ‘And a sweet smell would emanate from their holy graves.’

(ID 86)

(BB G57 276)

(BG 9a. 22)

(BG 19b. 26)

Clause and text organization in early East Slavic

(42) se bo vъ jedinъ ot dnii pride kelarь kъ semou blaženomou PART for on one of days arrived cellarer to this blessed.one ‘One day the cellarer came to this blessed one.’ (Th 50b. 17)

Again, the thematic analysis of the X constituent is usually supported by its lexical and/or grammatical properties: phrases containing demonstrative modifiers and personal pronouns make up many of the examples. As with some of the (SV)RXT clauses discussed in Section 5.1.3, here too there is no compelling reason to suppose that the placement of the thematic constituent after the rheme should have produced any special emphatic or expressive effect. In contrast to the earlier material, none of the examples in this group occurs in direct or indirect speech. The only explanation that suggests itself for the contrast in constituent order between the verb-first example (41) and the subject-first example (21) is the absence of any rigid norm of clause organization.

5.2.4 VXS Outside of Nov, there is less diversity in the pragmatic structures associated with VXS constituent orders: VRXTSR is the predominant pattern. In ID, nominal X constituents in medial position may be accompanied by a demonstrative modifier; in Th and BG, the X constituent is typically a personal pronoun; and in all the more extensive sources adverbial X constituents are also found post-verbally: (43) isxodit iz gradišča togo vasanьskago 7 rěkъ emanates from town that Bashan 7 rivers ‘From the town of Bashan flow seven rivers.’

(ID 86)

(44) tъgda že pristoupi kъ njemou ključarь then PART stepped.up to him sacristan ‘Then the sacristan approached him.’

(Th 53d. 17)

(45) idoša tamo oubozii went there poor.men ‘Poor men went there.’

(BG 21b. 9)

(46) estь tu pridel cerkovь svjatago mixaila is there chapel church Saint Michael ‘There is a chapel of the church of S. Michael there.’

(Const 409v)

Nov contains approximately equal numbers of rheme-only and VR XTSR clauses (8 and 7 respectively), as well as several examples like (37) in which the rheme may be another constituent altogether. Ad hoc explanations for the contrast between (VSX)R and (VXS)R patterns involving syntactic complexity or differences in empathy value can sometimes be suggested, but no consistent motivating factor is evident.

35

36

Sarah Turner

5.3

X-first clauses

Clauses in which the X constituent stands as the first of the constituents considered are found most frequently in the sources where topographical description features prominently, and for this reason clauses containing verbs of motion are heavily outnumbered by clauses containing verbs of position in this subset of the corpus. The XVS clause pattern is entirely absent from Nov and BG; the XSV pattern is absent from BB.

5.3.1 XVS Clauses in this group usually have the structure XT(VS)R. The X constituent is typically either a spatial adverb or a spatial prepositional phrase which includes a demonstrative modifier: (47) tu ležit glava ioanna zlatoustago there lies head John Chrysostom ‘The head of S. John Chrysostom lies there.’ (48) v tyx polatax estь čaša in those rooms is font ‘In those rooms there is a font.’ (49) na to město priteče skoro svjataja bogorodica to that place ran hastily holy mother.of.god ‘The holy Mother of God ran hastily to that place.’

(SN 141r)

(Const 418v)

(ID 38)

Again, there are a few instances of complex clauses in which another constituent is the principal part of the rheme. These aside, other patterns are very rare. Occasionally, the context makes it possible to argue for a (XV)TSR analysis, as in (50). If the contextual criterion for thematic analysis is applied strictly, there is a possible example of a rheme-only clause in SN. ID contains a small number of examples of (XV)RST patterns, as in (51), and the only XVS clause in BG is of this type: (50)

[i v toi usypalьnici ležat svjatii otci · telesy jako živy · i ležit and in that vault lie holy fathers bodies like alive and lie ixъ tu bolě 7- sot ·] tu ležit svjatyi kirijakъ ispovědnik · them there more 7 hundred there lie Saint Cyriacos confessor tělom vesь cělъ · tu ležita ksenofontova syna · ioannъ i arkadie body all whole there lie Xenophon’s sons John and Arkadios ‘[and in that vault lie holy fathers, their bodies as though they were alive. More than seven hundred of them lie there.] S. Cyriacos the confessor lies there, intact in body. The sons of Xenophon, John and Arkadios, lie there.’ (ID 74)

Clause and text organization in early East Slavic

(51) [i estь rěka ta nyně suxa] pod zemleju tečetь rěka ta and is river that now dry under ground flows river that ‘[And that river is dry now.] That river flows underground.’ (52) vъ velicě pečali bjaaše volodimirъ in â•›great anxiety was Volodimir ‘Volodimir was in great anxiety.’

(ID 66)

(BG 8d. 30)

In the last instance, the rheme-theme order might conceivably be considered expressive, but the same cannot be said for (51). That example calls to mind the type of modern Russian clause discussed in Fougeron (1989:â•›218–219), where an explanatory nuance links a clause with the rheme in initial position with the preceding text. However, that nuance is absent from the other examples found in ID, and no consistent and principled explanation for these divergences from the basic XT(VS)R pattern suggests itself.

5.3.2 XSV The pragmatic characteristics of the examples in this subset of data are generally similar to those of the XVS clauses. The X constituent is always thematic, normally either an adverb or a prepositional phrase with a demonstrative modifier. The subject is normally the primary rhematic constituent. In the absence of any compelling reason for doing otherwise, the verb is analysed as a constituent of the rheme, though in some examples taken from passages in which the location of saints’ relics is the overarching topic of discussion there may be a case for a thematic analysis of the verb ležati (‘to lie’). On several occasions, the rhematic status of the medial subject is marked by the particle i: (53) tu i sekira noeva ležit there also axe Noah’s lies ‘Noah’s axe is also there.’ (54) tu i dom ego bylъ there also house his was ‘his house was also there.’

(SN 140v)

(ID 40)

Again, in some instances ad hoc hypotheses regarding the difference between XT(VS)R and XT(SV)R types can be put forward, but none provides a consistent explanation for the variation in the order of the non-initial syntactic constituents.

. Cf. Section 5.1.4.

37

38

Sarah Turner

Table 3.â•‡ Approximate frequencies of pragmatic clause types SVX

SXV

VSX

VXS

XVS

common ST(VX)R (> 25 examples)

(VSX)R VRSTXR

VRXTSR

XT(VS)R

fairly common (SVX)R (SV)RXT

SRXTV(T/R?) (VS)RXT ST(XV)R

(VXS)R

uncommon (< 5 examples)

6.

XSV

XTSRV(T/R?) (XV)RST ?(XV)TSR ?(XVS)R

Discussion

Table 3 summarizes the pragmatic characteristics of the clauses in the corpus which were described in Section 5. Because of the uncertainties encountered in the course of analysis, the frequency of the clause patterns is indicated only approximately. Whilst VSX clauses are strongly associated with two sets of pragmatic characteristics, clauses which begin with a nominal constituent show stronger affinities with one set of pragmatic characteristics than with others, most markedly so in the case of XVS clauses. The predominance of one pragmatic pattern in the XVS category is particularly notable in the light of the exposition of CSR given in Section 1, which showed that there XVS clauses are associated with two types of pragmatic organization. A number of points emerge when the distribution of clause types among the sources is considered. They relate to three matters: (i) principles of clause organization itself; (ii) principles of text organization; and (iii) the usefulness of assigning descriptive labels to clauses of different forms.

6.1

Clause organization

6.1.1 Theme-rheme patterns Clauses which are straightforwardly divided into thematic and rhematic domains are strongly attested in all the sources in the corpus apart from BB, where their rarity is likely to be a product of the small number of relevant examples in the corpus and the fragmentary nature of the documents. In Th, BG and Nov, the dominant theme-rheme clause follows the pattern ST(VX)R, whereas in ID, SN and Const XT(VS)R is especially prominent. However, all the sources also contain clauses in which the theme-rheme progression is not observed. Alternative patterns typically involve the placement of

Clause and text organization in early East Slavic

a thematic constituent medially, after a rhematic verb. In the case of Nov and ID, the interposed thematic constituent is typically the subject; for BG and Const, it is the spatial expression; the data from Th and SN are respectively too varied and too sparse to allow for generalization. One factor encouraging this medial placement of the theme seems to be the use of the verb byti (‘to be’): all but one of the 13 verb-first clauses in Const begin with estь (‘is’), though estь may also appear in later positions in the clause; over a third of all byti clauses considered in ID begin with the verb and have a thematic medial constituent. Beyond that, however, no other local factors can be discerned which prompt this type of clause organization with any regularity or consistency. A less frequent alternative pattern involves the placement of a thematic constituent in the final position in the clause. ID has occasional examples of both S and X constituents in final position when thematic. In Th and BG the final thematic constituent is usually the spatial expression, and the other constituents can appear either in VS- or SV-order. Since BG attests no XVS patterns at all and Th has just two examples of XT(VS)R, there would seem to be grounds for proposing that in texts which demonstrate a particularly strong influence from Church Slavic there is some sort of syntactic restriction on the sorts of constituents which can occupy initial position in the clause, but clearly this hypothesis would need to be tested against a much larger and more diverse sample of data.

6.1.2 Rheme-only clauses The rough outline of the data given in Table 3 suggests that the preferred order of constituents in a rheme-only clause has changed between the medieval and modern periods. The preferred placement of the spatial X constituent seems to have shifted from final to first position. Most of the (VXS)R clauses in the premodern data are preceded by a temporal expression such as Vъ lět · 6524 (‘In the year 6524’) or tъgda že (‘then’). In modern rheme-only clauses, sequences of temporal and spatial expressions are now permissible, where previously they seem to have been avoided. However, these contrasts between the two periods do not directly support a larger conclusion about directions of linguistic change, for it is not clear to what extent the (VSX)R pattern may in fact be considered typical of early East Slavic in general. The (VSX)R clauses in the present sample of premodern material are heavily concentrated in Nov. As discussed in Section 5.1.2, the (SVX)R clauses are problematic; and the (XVS)R examples are marginal. The lack of (VSX)R clauses outside Nov does not in itself disprove the hypothesis that this clause pattern demonstrates the preferred syntactic configuration of rhemeonly clauses in early East Slavic more widely, but the lack of rheme-only clauses of any form in sources other than Nov means that there are no grounds for accepting it either. In short, the diversity of the data relevant to Section 6.1.1 and the

39

40 Sarah Turner

uniformity of the data relevant to Section 6.1.2 both make it difficult to discern any general principles which might have governed clause organization in early East Slavic writing.

6.2

Text organization

6.2.1 Division and continuity That there should be many examples of rheme-only clauses in Nov is not a surprise in itself, for the text is highly segmented. It is more surprising that there are so few to be found elsewhere. The other long text in the corpus, ID, is also broken down into smaller sections, though on a topical rather than a chronological basis. The heading often establishes a theme for the first phrase of the new section, as in (55). At other times, the referent of the heading stands late in a clause that has opened with a phrase which is very closely modelled on the last phrase of the preceding section: (55)

O VIFLEOMĚ about Bethelem Vifleem že svjatyi estь na jugь licь otъ ierusalima Bethlehem PART holy is to south direction from Jerusalem svjatago, 6 verstъ vdalee holy 6 versts distant ‘ON BETHLEHEM Holy Bethlehem is in a southerly direction from holy Jerusalem, six versts away.’ (ID 62)

(56) {...} a ot tripolja do rěki sudii 60 verstъ and from Tripolis to river Sudija 60 versts O VELICĚ ANTIOXII about great Antioch Na toi bo rěcě estь antioxija velikaa on that PART river is Antioch great ‘{…} and from Tripolis to the river Sudija [?] is 60 versts. ON ANTIOCH THE GREAT On that river stands Antioch the Great’.

(ID 82)

Thus in ID the division of the text into sections does not always break the narrative line in the way that it does in Nov, with the result that there are correspondingly very many fewer points in the text at which rheme-only clauses can appear. To the modern reader the narrative in ID may appear to be more conventionally structured than in the chronicle text, but in this respect its organization actually differs from CSR.

Clause and text organization in early East Slavic

6.2.2 Organizing parameters The high frequency of XT(VS)R patterns in some of the sources draws out a further point of difference between the literary idioms of the medieval sources and CSR. The strong inter-clausal links between many of the pre-modern texts means that their narratives develop on a single plane. In CSR, rheme-only clauses are found in environments in which scenes are set as background to the storyline. The information they convey is tangential to the main narrative, whose overarching coherence is not compromised if there is a break in chain of clause-initial themes. These two levels in text construction are not found in the pre-modern sources studied here. There does not appear to be any incidental information in them. The topographical descriptions in ID, SN and Const constitute the narrative rather than augment it, and the clauses with verbs of motion and position that they attest almost invariably have a contextually available thematic constituent which links the clause to what has gone before and which helps to create the text. The insistent use of demonstrative modifiers in ID underscores the density of the text-forming connections. The texts in the corpus can be divided into three groups according to their dominant organizing feature. ID, SN and Const are organized with reference to space, Nov with reference to time, BG and Th with reference to person. These different structuring parameters go some way towards accounting for the distribution of data seen in the corpus: the high proportions of XT(VS)R patterns in the spatial group, of ST(VX)R in the person group, and of clauses in which none of the arguments of the verbs is thematic in Nov. They do not resolve the many difficulties in the interpretation of the material discussed above, and it remains impossible to distil from the material any principles of clause organization which operate consistently across individual texts of any length, let alone groups of texts divided according to subject-matter, genre, or their orientation towards Church Slavic. Nevertheless, they do help to explain why to the modern reader Nov and ID both feel to be written in a manner that is rather far removed from the conventions of CSR. In works of literary writing, temporal and spatial expressions typically present background information; in works of historical writing or in guidebooks they may present new, rhematic information within the main narrative line; but they rarely serve as the thematic ties which link clauses into text. For these reasons, the temporal organization of the text in Nov makes for narrative discontinuity, whilst the spatial organization of ID, SN and Const make for texts which cohere almost . This density of information which to the modern reader appears to be undifferentiated for importance is found within the clause, too, especially in Nov. In example (29), for instance, it is difficult to say whether the temporal expression at the end of the clause is central or incidental to the chronicler’s point.

41

42

Sarah Turner

to the point of monotony. Th and BG build their person-oriented narratives out of information which for the modern reader is more typically thematic and found in the foreground. Furthermore, there is a wider range of possibilities for using synonymous or anaphoric substitutes for subject phrases denoting persons, or indeed for omitting them altogether, than there is for spatial or temporal expressions, with the result that these texts create an impression of narrative diversity whilst maintaining a tight textual structure.

6.3

Interpretative descriptions of constituent order

The viability of interpreting particular patterns of variation within the sources as expressive inversions of neutral word orders has been discussed at some length in Section 5.1.3 and the arguments will not be repeated here, save to say that the doubts about the concept of expressive word orders which were raised there are just one area of concern which has arisen in the course of this study in connection with the approach to word order set out in the introduction which assigns descriptive labels to particular clause patterns. At first sight, this approach appears to offer an elegant short-hand for summarizing the pragmatics of clause organization in Russian. In an analysis of constituent order which utilizes the units of the theme and the rheme, three possible pragmatic structures are each associated with a particular label: T/R: informational R/T: expressive R only: presentational

Since the model was drawn up with reference to CSR, its inability to deal with common features of the pre-modern sources such as the placement of major thematic constituents within rhemes cannot automatically be seen to compromise it. However, even without testing this model against real data, it is evident that there could be conceptual difficulties with it, for two reasons. Firstly, there is an inconsistency in the reference made to different descriptive parameters: the syntactic categories of the clause constituents are apparently not considered relevant to the identification of informational and expressive constituent orders, yet the rheme-only type of clause is associated with a syntactic feature: the postposition of the subject to the verb. Thus it is unclear which of the labels should be assigned to a clause which exhibited the pragmatic characteristics of a rheme-only clause . More problematic is the fact that pragmatic units can be split in CSR, too. There, the placement of the rheme between thematic constituents is both frequent and unmarked: see Note 4.

Clause and text organization in early East Slavic

but the syntactic properties of one of the divided types, or vice versa. The material presented in Janko (2001:â•›149–153) shows that such clauses exist in at least some varieties of modern Russian as used by educated speakers. For pre-modern Russian, where the postposition of the subject to the verb is not as strongly associated with rheme-only clause patterns, the relationship between syntactic and pragmatic aspects of clause organization will clearly be still more complicated to describe with reference to larger terms. Secondly, and more fundamentally, the different labels for the clause types merely add descriptive possibilities rather than an overarching explanatory framework to a theme-rheme approach to the analysis of constituent order. The choice of label rests upon an analysis of form. It would be circular to seek to explain a formal difference in the organization of two clause patterns with reference to these terms, when they are arrived at in the first place on the basis of an analysis of form. This being so, it is not evident that any substantial benefit is derived from using them. If the concepts of ‘informational’ and ‘presentational’ constituent orders are used in a vaguer way which depends not on the immediate pragmatic properties of the clause but, loosely speaking, on its semantic effect, it is not clear which of the labels would be more appropriate for an example such as (57): (57)

A nad dverьmi estь spasъ čjudotvornyi ikona vysoko · and above doors is saviour miracle.working icon high tot spasъ mnogo bolьnyx iscěljaet · pred tem spasom viselo that saviour many invalids cures before that saviour hung ponikadilo · retjaz želězna · k toi retjazi privjazan stьkljanokь s chandelier chain iron to that chain attached phial with maslom a pod stьkljanikom stoit stolpec kamen · a na stolpъci oil and under phial stand pillar stone and on pillar čaša {i drevo noeva kovčega} okovana želězom kovčežnym cup and wood Noah’s ark forged iron ark.ADJ ‘And high above the doors is the miracle-working icon of the Saviour. That icon cures many invalids. In front of the icon was hanging a chandelier on an iron chain. To that chain is secured a phial with oil, and under the phial there is a small stone pillar, and on the pillar is a cup {and wood of Noah’s ark} forged from iron from the ark.’ (Const 410v)

Whilst there are a strong formal and contextual grounds for analysing these clauses as theme-rheme patterns, it is not entirely clear in what sense they can be said to be informational. Certainly, each clause does give additional information about the referent of the thematic constituent, but it arguably does so only because the author wants to introduce a new referent entirely into the text, and he chooses

43

44 Sarah Turner

to build up his larger description of the interior of the church by locating the items whose existence he wishes to note with reference to known points in space. Although the main rhematic constituent may reappear as the theme of the following clause, none of the constituents typically has any prolonged presence in the text: they persist only as long as is necessary to provide a point of orientation with reference to which another item can be presented. In modern texts where the narration of action is primary, the distribution of so-called informational and presentational constituent orders correlates to some extent with the distinction between foreground action and background description, and from this correlation the descriptive labels derive a degree of intuitive satisfactoriness. In texts with a strong tendency towards description of surroundings, however, the distinction between informational and presentational clauses is not even intuitively satisfactory in the way that it can be for action-oriented texts. In short, then, greater linguistic insights are to be gained by discussing differences in the basic pragmatic structures at clause level with reference to specific features of textual cohesion, foregrounding and backgrounding, and, if appropriate, authorial style, rather than with reference to elusive pseudo-functional descriptions.

7.

Conclusion

Even for a relatively small corpus of material defined with reference to a small group of verbs, general principles influencing clause organization in pre-modern sources remain elusive. In this study it has become clear that the obstacles to generalization do not lie simply in classic areas of difficulty such as determining the relationship between Church Slavic and vernacular conventions or establishing the a/typicality of the language of the chronicles. The problem arises in part because, as shown in Section 6.2, several of the medieval sources studied here observe conventions of text organization which limit the kinds of pragmatic environments that they attest.

Clause and text organization in early East Slavic

References Primary sources Čexov, A. P. 1976. Sredstvo ot zapoja. In Polnoe sobranie sočinenij i pisem v tridcati tomax: Sočinenija. Tom četvertyj: 1885–1886, N. F. Belčikov (ed.), 175–180. Moscow: Nauka. Čexov, A. P. 1977. Palata № 6. In Polnoe sobranie sočinenij i pisem v tridcati tomax: Sočinenija. Tom vos’moj: 1892–1894, N. F. Belčikov (ed.), 72–126. Moscow: Nauka. Dietze, J. (ed.). 1971. Die erste Novgoroder Chronik nach ihrer ältesten Redaktion (Synodalhand‑ schrift). Leipzig: Edition Leipzig. Kotkov, S. I. (ed.). 1971. Uspenskij sbornik XII–XIII vv. Moscow: Nauka. Proxorov, G. M. (ed.). 1980. Xoždenie igumena Daniila. In Pamjatniki literatury drevnej Rusi: XII vek, L. A. Dmitriev & D. S. Lixačev (eds.), 24–115. Moscow: Xudožestvennaja literatura. Speranskij, M. N. 1934. Iz starinnoj novgorodskoj literatury XIV veka. Moscow: Izdatel’stvo Akademii Nauk SSSR. Zaliznjak, A. A. 2004. Drevnenovgorodskij dialekt, 2nd edn. Moscow: Škola “Jazyki russkoj kul’tury”.

Secondary sources Adamec, P. 1966. Porjadok slov v sovremennom russkom jazyke. Prague: Academia. Bogoljubova, N. D. 1970. K voprosu o porjadke slov v Pskovskoj pervoj letopisi. In Voprosy stil‑ ja i istoričeskoj grammatiki: Učenye zapiski tom 115: Trudy kafedry russkogo jazyka, N.Â€A. Donec (ed.), 69–98. Riga: Latvijskij gosudarstvennyj universitet. Dyer, D. L. 1992. Word Order in the Simple Bulgarian Sentence: A Study in Grammar, Semantics and Pragmatics. Amsterdam: Rodopi. Fougeron, I. 1989. Prosodie et organisation du message: Analyse de la phrase assertive en russe contemporain. Paris: Libraire C. Klincksieck. Gippius, A. A. 2004. K pragmatike i kommunikativnoj organizacii berestjanyx gramot. In Novgorodskie gramoty na bereste (Iz raskopok 1997–2000 godov): Tom XI, V. L. Janin et al. (eds.), 183–232. Moscow: Russkie slovari. Grenoble, L. A. 1998. Deixis and Information Packaging in Russian Discourse. Amsterdam: John Benjamins. Janko, T. E. 2001. Kommunikativnye strategii russkoj reči. Moscow: Jazyki slavjanskoj kul’tury. Kardaševskij, S. M. 1948. Porjadok slov v Povesti vremennyx let. Učenye zapiski Moskovskogo oblastnogo pedagogičeskogo instituta 12(1): 33–52. Kovtunova, I. I. 1980. Porjadok slov. In Russkaja grammatika. 2: Sintaksis, N. Ju. Švedova (ed.), 190–214. Moscow: Nauka. Lapteva, O. A. 1976. Russkij razgovornyj sintaksis. Moscow: Nauka. Levin, B. & Rappaport Hovav, M. 1995. Unaccusativity: At the Syntax-Lexical Semantics Inter‑ face. Cambridge MA: The MIT Press. Turner, S. 2005. A Pragmatic Approach to Constituent Order in Early East Slavic Sources. PhD dissertation, University of Oxford. Yokoyama, O. T. 1986. Discourse and Word Order. Amsterdam: John Benjamins.

45

46 Sarah Turner

Appendix. Abbreviations used in interlinear glosses adj dem part refl

adjective demonstrative particle reflexive

chapter 2

Indeterminate motion verbs are denominal Johanna Nichols

University of California, Berkeley

Slavic indeterminate verbs of motion are better analyzed as denominal verbs than as causatives. They owe their *-o- vocalism, their *-i- stem suffixes, and their indeterminate aspectuality to their denominal origin. Some of them may originally have been Indo-European causatives, but were reanalyzed as denominal in a larger typological shift of the Slavic verbal lexicon from verb-based to noun-based. The same type shift also facilitated the formation of denominal transitive verbs such as sušiti ‘dry’ and čistiti ‘clean’, and their secondary reflexives provided new intransitive verbs. These changes, together with the incipient development of aspect pairs, took place as Common Slavic moved into the European cultural sphere.

1.

Introduction

In Proto-Indo-European, verb roots were basic to the lexicon and the word formation. Underived verb roots were an open class, many derivational word families were based on verbs, and many nouns that one might expect to be basic were in fact deverbal, e.g., *dent- ‘tooth’, a participial formation from ‘eat’ (Pokorny 1959:â•›289). I will call this the verb-based lexical type; essentially it means that the rules of word formation expect verbs as input. Here I argue that sometime well after the breakup of Proto-Balto-Slavic and not long before historical times the Slavic lexicon underwent a shift of its derivational type from the inherited verbbased type to the neutral or even noun-based type that is more typical of modern European languages. The history of motion verbs reflects this shift, and the formation of the indeterminate Aktionsart in East Slavic and Polish is a consequence of the larger set of changes. This paper follows this history of Common Slavic verbs such as *xoditi to show how the shift was implemented. . I use Proto-Slavic (abbreviated PS) for the long period up to about the beginning of the first millennium, and Common Slavic (CS) for the period in the mid first millennium from the

48 Johanna Nichols

The received account of xoditi and similar verbs is that they are deverbal derivations using Indo-European causative morphology, which had or tended to develop iterative functions, perhaps because of an ultimate origin in denominal verbs (Kurylowicz 1964:â•›77), perhaps because the present tense of the Slavic verbs of this type was an old optative (Darden n.d.). Commonly cited in support of the iterative function is Greek phoreō ‘wear’ (i.e. habitually) vs. pherō ‘carry (something somewhere on a particular occasion)’. The causative morphology consisted of -o- grade in the verb root plus conjugational suffixes infinitive/aorist *-i:- and present *-ei- (CS *-i-) (Schenker 1995:â•›133). The same segmental morphology formed factitives (i.e. denominal transitives) from both adjectives and nouns (e.g., CS *sušiti < PS *soux-i:-tei ‘dry (something) out’, *kositi < *kos-i:-tei ‘mow, harvest hay with scythe’), but there was an accentual difference: the present tense formative *-ei- was circumflex in the causative, yielding mobile stress paradigms, while the factitive *-ei- was acute, and denominal paradigms had either fixed stress on the acute suffix or fixed stress following the stress of the noun (Vaillant 1966:â•›408, 414–415, 432–433; Stang 1957:â•›112). There are many departures from this rule, however (Stang 1957:â•›111–113; Vaillant 1966:â•›519). Machek (1968) says that *xoditi is denominal, derived from the noun *xod‘gait; movement, route’, adducing other Indo-European cognates and pointing out the structural and semantic similarity of Greek éks-odos ‘exodus’ to Slavic is-xodъ. He does not comment on the accentology. He does not derive the other similarly formed motion verbs *nositi ‘carry’, *voditi ‘lead’, *voziti ‘convey’, *goniti ‘chase’, *broditi ‘wade’, *laziti ‘crawl’ from nouns, appparently because he does not adduce IE cognate nouns. Trubachev (1974–) ascribes a denominal derivation to most of these other verbs, without, however, giving the argument or other IE cognate nouns (only for *broditi are cognate nouns cited), though he describes *laziti and *goniti as deverbal (iteratives), giving no argument for why he considers them differently derived from the others; for none of these is the accentology mentioned. Of the motion verbs Vaillant 1966 considers only *jězditi denominal, because the putative source noun *jězda has a noun-forming suffix (also found in uzda ‘bridle’ and a few others). Trubachev (1974–) denies that *jězda has a suffix but still considers this noun basic in the word family and *jězditi derived from it. Evidently neither Machek nor Trubachev considered the motion verbs as a set with parallel derivation. Vaillant does, but he regards the indeterminates as ancient iteratives derived from the determinates, i.e. as deverbal, except for *jězditi with its (in his view) clear nominal suffix. beginning of dialect differentiation through the dispersal up to historical times. CS forms are transcribed here using OCS transliteration but with unresolved liquid diphthongs (*vort- instead of vrat-, etc.).

Indeterminate motion verbs are denominal

I argue here that xoditi, whatever the origin of its morphology, is denominal in historically attested Slavic and that all motion verbs with the same morphology have been either formed or reanalyzed as denominal, not in Indo-European but more recently, in late Proto-Slavic or Common Slavic. This is probably also true of many non-motion verbs with the same morphology and of motion verbs with different morphology. Indo-European parallels are probably instances of independent parallel development, and are basically irrelevant to the Slavic derivation. Such a sweeping reanalysis is a radical change, but I argue that it is the most parsimonious explanation given the larger derivational morphological context. To make the case for the reanalysis I will show that it fits into the larger directionality of verbal derivation and valence (§2 below), that there is a nominal source for the root or stem of verbs like xoditi (§3), that the accentual differentiation of causatives from factitives has been effaced in Slavic (§4), that indeterminate verbs as a class (where they exist, e.g., Russian) bear the analysis as denominal (§5), that indeterminate semantics is a plausible outcome of denominal formation (§6), and that deadjectival verbs support the general idea that a verb’s aspectuality can depend on its derivational history (§7). This approach accounts for important points of Slavic word formation and aspectuality, but it raises some other questions about the Indo-European and early Proto-Slavic ancestral forms (§8). Let me emphasize that this is not an etymological study; the structure and origins of Proto-Slavic causative and factive morphology and the etymologies of at least some of the indeterminate verbs are well understood (see Greenberg, this volume, and Darden n.d. among others) and this paper does not question them but simply proposes that a reanalysis of aspects of Proto-Slavic word formation took place in Common Slavic.

2.

Proto-Slavic and Common Slavic verb derivations in *-ei- (> -i-) / -i:-

As mentioned above, Proto-Slavic inherited causative and factitive derivational patterns, both of which produced derived transitive verbs with -o- root vocalism and -i- conjugational suffixes. In addition, corresponding to some causatives and many factitives were stative intransitives suffixed with -ě- and without any change in the root vocalism; or (less frequent) inchoative intransitives suffixed with -nŏ-. In prehistoric CS times the inherited reflexive enclitic came to be used as a detransitivizer, and by early historical times reflexivization is well established in the formation of both (syntactic) passives and (lexical) intransitives. This account of Slavic transitivization and detransitivization is drawn from Gołąb (1968). The stages in this prehistory are summarized in Table 1.

49

50

Johanna Nichols

Table 1.â•‡ Slavic intransitive-transitive pairings POS

Earlier PS

Late PS

CS

Aktionsart

V

u:k-noN-

vyk-no˛-ti

-vyk-no˛-ti -vyk-a(j)-ti

inchoative iterative?

ouk-ei- / -i:-

uč-i-ti

uč-i-ti uč-i-ti sę

±telic ±telic

nes-

nes-ti

nes-ti nos-i-ti nes-ti sę nos-i-ti sę

±telic atelic activity? ±telic (passive) atelic (passive)

be:l-e:be:l-ei- / -i:-

běl-ě-ti běl-i-ti

běl-ě-ti běl-i-ti běl-i-ti sę*

state accomplishment ±telic (passive, “middle”)

keistkoist-ei- / -i:-

čistcěst-i-ti

čist-ъ cěst-i-ti > čist-i-ti čist-i-ti sę

(adjective) ±telic ±telic (passive, “middle”)

V Adj

sux-noNsouxsoux-ei- / -i:-

sъx-no˛-ti suxsuš-i-ti

sъx-no˛-ti sux-ъ suš-i-ti suš-i-ti sę

inchoative (adjective; state) accomplishment accomplishment

N

kos-ei- / -i:-

kos-i-ti

kos-i-ti kos-i-ti sę *

±telic ±telic (passive)

Adj

POS = part of speech of the root. * marks reflexive forms that were probably only passives. ±telic = a verb that can be either activity or accomplishment depending on context (in modern languages they are accomplishments when prefixed).

The regular use of reflexivization as a detransitivizing device led to lexical dissociation of transitives from the old intransitives: thus učiti ‘teach, study’: -vykno˛ti ‘learn’ > ‘become accustomed’ drifted apart semantically, and some intransitives have dropped out entirely. The result was pairs of plain (transitive) and reflexive (intransitive) verbs with the same root vocalism and complete parallelism between deverbal, deadjectival, and denominal formations.

3.

Unsuffixed deverbal nouns

Proto-Slavic inherited an Indo-European unsuffixed deverbal noun formation with -o- vocalism: xodъ, ‘gait, pace, motion’ and many others. In OCS these are attested with a variety of functions: agent, instrument, location, manner, general fact. In Slavic, at least by CS times, these nouns were probably most common with

Indeterminate motion verbs are denominal

prefixes and in compounds, but unprefixed independent forms are attested for at least some of them in OCS. Compounds with these deverbal nouns seem to have been formed only from transitive roots as second element, with a noun stem (the semantic object) as first element. The two clearly reconstructable Proto-Slavic compounds with verbal second elements, medvědь ‘bear’ (*med- ‘honey’ + either *věd- ‘know’ or *ěd- ‘eat’) and vojevoda ‘military commander’ (*voj- ‘troops, army’ + vod- ‘lead’) (Kiparsky 1975:â•›344) are of this type. The following forms and examples were found by searching Lindstedt (1994) and checking Sadnik & Aitzetmüller (1955) and Alekseev & Gerd (2006). Unprefixed: xodъ ‘gait, manner of walking’ (manner) (exs. (1)–(2) below) ‘stance, bearing, figure’ (manner?) (3) ‘motion, movement’ (general fact; manner?) (4) (1) i xodъ vьlovьnъ iměti and gait ox.adj have.inf ‘and have an ox’s gait’, ‘and walk like an ox’

(Supr. 3115125)

(2) upvajo˛šte końьskъimъ xodomъ relying horse.adj.instr pace.instr ‘relying on (their) horses’ (swift) pace’

(Supr. 1016105–6)

(3) lice děviče . xodъ děviče . oči děviči face girl.adj bearing girl.adj eyes girl.adj ‘a girl’s face, a girl’s figure/bearing, a girl’s eyes’

(Supr. 2002211–2)

(4) otъ kameni i otъ drěva tvoreni . xoda ne imo˛šte from stone and from wood made motion.gen not having ‘made of stone and wood, and immobile’ (Supr. 10251007–8)

Prefixed (selected examples): vъxodъ ‘entrance, entryway’ zaxodъ ‘sunset’ isxodъ ‘end, decease, death’; ‘(Biblical) exodus’ prixodъ ‘arrival’ uxodъ ‘fugitive’

(place; general fact) (5) (general fact) (6) (general fact) (7) (general fact?) (8) (general fact) (9) (agent)

(5) zaključiti imъ vьsь vъxodъ vъ grъčьsko˛ zemjo˛ forbid.inf them all entry into Greek land ‘forbid them any entry into the Greek land’

(Supr. 3149226)

51

52

Johanna Nichols

(6)

i nikomu ottoli sъměti po slъnečьněěmъ zaxodě vъně and no one thence dare.inf after sun.adj set outside domu svoego obrěsti sę house.gen refl.gen find.inf refl ‘and thenceforth no one was to be outdoors after sunset’ (Supr. 1027103–5)

(7) do isxoda dušę moeję to â•›end.gen soul.gen my.gen ‘to the end of my life’

(Supr. 1102203–4)

(8) vъ isxodъ iilevъ otъ egipta in â•›exit Israel.adj from Egypt ‘during the departure of Israel from Egypt’ (KJV: ‘when Israel went out of Egypt’) (Ps. sin. 149.14) (9) ožidajo˛št-iimъ prixoda tvoego waiting.instr.pl arrival.gen your.gen ‘to the ones waiting for you to arrive’, ‘awaiting your arrival’ (Supr. 1073125) (10) viděšę Lazara uxoda sьmrьti bivъša (sic) saw.pl Lazarus.gen fugitive.gen death having.beaten (intended: having.been) ‘they saw that Lazarus escaped death’ (lit. ‘saw Lazarus who was the/an escapee from death’) (Supr. 3028118)

Compounds: There are no canonical OCS compounds in -xodъ, but here are two with second elements from other verbs: vojevoda ‘commander’ (agent) vodonosъ / vodonosь ‘water jug’ (instrument) (11) slyšaaxo˛ otъ mnogъ ježe tvorěše vojevoda mo˛ky krьstijanomъ heard.pl from many that made commander torment Christians.dat ‘they heard from many that the commander tormented Christians’ (Supr. 1010128–9, 1010201) (12) obrěte sę vodonosь isplьnena vody found refl jug filled water.gen ‘the water jug was found full of water’

(Supr. 3141225–6)

For a full list of such nouns in OCS see Meillet (1961:â•›215–225). Tokens of them are not particularly frequent, and most have no unprefixed attestations. A number of the words are found only in Suprasliensis. Their relational semantic functions (agent, instrument, etc.) vary, but all share the property of being lexicalized (rather than syntactically formed nominalizations) and therefore referring not to particular events but to types, properties, and non-specific or general facts. Of the

Indeterminate motion verbs are denominal

examples I have found the only one that is close to referring to a specific event is isxodъ ‘Biblical Exodus’ in (8) above; but even this is used not so much as a specific event nominalization but as a lexicalized name. In their frequent use with prefixes and in compounds, and the range of their nominalization types, these nouns are reminiscent of the verbal elements of English conversion-derived words such as input, throughput, outbreak, etc. and exocentric compounds such as breakwater, scarecrow, pickpocket. The difference is that English can easily use conversion and use uninflected generic verbal forms like put, break while Slavic languages, with their more explicit and extensive inflection, must form unambiguously nominal forms that have case and gender. The unprefixed examples are particularly important to the development of indeterminate verbs. Reconstructable unprefixed nouns with cognate Slavic verbs from Meillet (1961:â•›215–225) are the following (‡ = not found in canonical OCS; ‘locust’ not in Meillet but canonical). Meaning types shown in parentheses are inferred from dictionary glosses (unlike those illustrated above, where meanings have been determined from text examples). From intransitives with agent subject:

běgъ‡ brodъ xodъ stonъ‡ pro˛gъ

‘flight, running away’ (manner) ‘ford, fording’ (manner) ‘pace’, etc. (examples above) (manner) ‘groan’ (sensory) ‘locust’ (agent)

From stative intransitives or intransitives with non-agent subject:

blěskъ‡ ‘brilliance, shine’ (sensory) cvětъ ‘blossom’ (result) (ne-)do˛gъ ‘illness, ailment’ duxъ ‘breath, spirit’ OCS gladъ, PS *gold- ‘hunger’ (cf., žlьděti ‘want’) gnojь ‘excrement, pus’ gromъ ‘thunder’ (sensory) kvasъ ‘kvas’ (beverage) (product) morъ ‘deaths, mortality, die-off ’ OCS mrakъ, PS *mork- ‘darkness’ (sensory) OCS mrazъ, PS *morz- ‘frost, freezing’ OCS rastъ, PS *orst- ‘growth, stature’ sluxъ ‘sense of hearing; rumor, news’; (du/pl) ‘ears’ (fact; instrument) OCS smradъ, PS *smord- ‘stench’ (sensory) studъ ‘shame, disgrace’

53

54

Johanna Nichols

světъ tokъ tro˛sъ znojь zvonъ

‘light’ (sensory) ‘course, run; current’ ‘earthquake’ ‘heat’ ‘sound’ (sensory)

From transitives:

gonъ ‘chase, hunting, driving’ kolъ ‘stake, pole; plug’ (instrument) kovъ ‘forging’ krovъ ‘roof, pavilion’ kruxъ ‘crumb, morcel’ kupъ ‘purchase; business’ lovъ ‘hunting, fishing’ rovъ ‘ditch, trench’ (product) strojь ‘structure, organization’ tvorъ ‘creation, something created’ (product) vozъ ‘cart, chariot’ varъ ‘heat’ vidъ ‘aspect, appearance’ OCS vratъ, PS *vort- ‘neck’ žьdъ ‘wall’ (product)

Misc. and uncertain:

blo˛dъ grobъ lo˛kъ plotъ ro˛gъ rokъ stogъ* vojь

‘sin’ ‘coffin’ ‘bow’ (instrument?) ‘fence, wall’ (product) ‘mockery’ (fact) ‘time limit, deadline; law’ ‘pile, heap’ (product) ‘soldier’ (attested only as pl. voji ‘army’)

As Darden notes (n.d.:â•›1–2), the valence of the *-i- verbs corresponding to these nouns is predictable: where the noun comes from a verb that is intransitive and agentive, the *-i- verb is intransitive; otherwise it is transitive. Table 2 summarizes the possibilities. Table 2 shows that the *-i- verbs corresponding to these -o- grade unsuffixed nominals have predictable valence: all are transitive except those corresponding to manner nouns from active intransitives. The Aktionsart of the output verbs also seems to be predictable: if the input nominal is instrument, means, or the

Indeterminate motion verbs are denominal

Table 2.â•‡ Semantics and valence of *-i- verbs corresponding to deverbal nouns and adjectives Input

Output

nominal

semantics

хodъ ‘gait’

manner

1

Intransitive(Sa) activity

xoditi ‘go; walk’

nosъ

fact; manner?

2

Transitive

activity

nositi ‘carry’

kosa ‘scythe’

instrument

2

Transitive

activity, ±telic

kositi ‘mow’

krasa means ‘decoration’

2

Transitive

activity, ±telic

krasiti ‘adorn’

suxъ ‘dry’

1

Transitive

accomplishment

sušiti ‘dry’

state

no. arguments valence

Example

Aktionsart

Input = the nominal and/or its source verb (e.g., xodъ ‘pace, gait’ and i-ti / šьd- ‘go’); output = the verb corresponding to that input nominal. Arguments = those implicit in the deverbal noun, and in the valence of its source verb. Sa = agentive subject of intransitive. ±telic = a verb which can be either telic (usually an accomplishment) or atelic (activity), as most unprefixed transitives can be in the modern languages (their prefixed perfectives are telic).

like the output verb is, or can be, telic (it means application or utilization of the nominal to achieve some goal); if the input nominal is a state (typically an adjective), the output verb is an accomplishment (see §7 below); if the input noun indicates manner the output verb is an atelic activity. Since the output verbs have predictable valence, Aktionsart, and conjugation type they can be regarded as derived, while the source verbs of the input nominals have a greater variety of conjugations, valence patterns, and Aktionsart types. This is a synchronic analysis of a situation that took shape in late Proto-Slavic or in Common Slavic, aided by the rise of reflexivization as a detransitivization device as discussed in §2 above. Thus the old *-iti verbs have become a single derivational class whose members are typically denominal synchronically and transitive by default.

4.

Accentuation

As noted in §1, causatives, or deverbal *-i:- verbs generally, had mobile accent while denominals had fixed accent. The infinitive/aorist formative *-i:- had acute accent and was stressed, while the present tense formative was circumflex *-ei(mobile stress) in causatives but acute (fixed stress) in denominals. Examples from modern Russian:

55

56

Johanna Nichols

učít’: učú, účiš’ xodít’: xožú, xódiš’ stydít’: styžú, stydíš’ šírit’: šírju, šíriš’

‘teach’ (old causative, i.e. deverbal) ‘go’ (indeterminate) ‘shame’ (transitive; denominal) ‘widen’ (transitive; denominal)

There are, however, many exceptions such as the following denominal verbs with mobile stress: prosit’: prošú, prósiš’ ‘ask’ (transitive; denominal) kosít’: košú, kósiš’ ‘mow’ (ambitransitive; denominal)

Since I am arguing that *-i- verbs both deverbal and denominal, regardless of origin, were reanalyzed as denominal in Common Slavic, I have to either account for this difference of accentuation or show that it is not in fact systematic. I surveyed all the verbs listed in Gołąb (1968) and Vaillant (1966:â•›Ch. XXII) and a random sample of about 5% of the modern Russian verbs in -it’ listed in Zaliznjak (1977:â•›700–740) to determine whether accent behaves as historically described and whether any such pattern is strong enough to be considered part of the grammatical marking of derivational classes. The results are not significant until some collapsing of categories is done, putting fixed root stress and fixed suffixal stress together and denominal and deadjectival together: Deverbal Denominal TOTAL

Fixed Mobile 26 (60%) 17 (40%) 33 (80%) â•⁄ 8 (20%) 59 (70%) 25 (30%) p < 0.05 (X2 = 4.09, d.f. = 1)

(Fixed = fixed root stress or fixed post-root stress. Denominal = derived from either noun or adjective root.)

There is a small but significant skewing: deverbal verbs do tend to have mobile stress and denominal/deadjectival ones fixed stress. This shows that there is some historical reality to the correlations. However, fixed stress predominates absolutely (though Vaillant 1966:â•›519 notes that mobile stress has been expanding), even in deverbals, which precludes (at least by structuralist method) regarding stress type as a synchronic grammatical marker. Therefore, all -it’ verbs can be regarded as synchronically a single class, with the same suffixal elements and arbitrary assignment to stress classes. The discussion of BCS forms in Vaillant (1966:â•›408ff.) and a more cursory survey I did of the BCS cognates to the verbs cited here, in Gołąb (1968), and about 1% of the verbs in -iti in Matešić (1965–1967), indicates that in BCS as well the accent paradigms for deverbal and denominal *-iti verbs are not discrete. BCS and Russian cognates (both those conforming and those not conforming to

Indeterminate motion verbs are denominal

the reconstructed norm of mobile stress for deverbal verbs and fixed stress for denominal verbs) often agree in accent class, suggesting that the mixture began in CS times; and they also rather often disagree in accent class, suggesting that the process of merging the two accent classes continued into post-CS times. Therefore it seems that the merger of formerly discrete causative (deverbal) and factitive (denominal) morphology began in CS, probably late in the history of CS. The question merits a close study of all the languages preserving free stress, however.

5.

Motion verb pairs in Common Slavic and Russian

The Common Slavic motion verbs paired for determinate vs. indeterminate, where the indeterminate has *o vocalism in the root and an *-i- suffix, are the following (Vaillant 1966:â•›412–413):

Determinate jьd- jěx-/jěd- lěz- nes- ved- velk- vez- žen-/gъn-a-

Indeterminate xod-i- jězd-i- laz-i- nos-i- vod-i- volč-i- voz-i- gon-i-/gan-i-a

‘go (by foot), walk’ ‘ride’ ‘crawl’ ‘carry’ ‘lead’ ‘drag’ ‘convey’ ‘drive, chase’

and possibly verz-/vrz- vorz-i- ‘tie’ (BCS only) bred- brod-i- ‘wade’

There are several other sets where the lexical meanings have drifted apart. Vaillant regards this formation as the oldest type of iterative and “une survivance” (Vaillant 1966:â•›475). The determinate-indeterminate pairs of Russian are listed in full below (Timberlake 2004:â•›414). Russian preserves all but one of Vaillant’s pairs with -o- grade and -i- suffix as pairs, and these represent over half of the Russian motion verbs. These are pairs in which the indeterminate shares its vocalism with a deverbal noun derived from the determinate, and for which I maintain that the indeterminate verb is derived from that noun. In addition, in the pairs with other suffixes on the indeterminate, most pairs have -o- grade in both forms and all pairs have cognate nouns with the same vocalism as the indeterminate.

57

58

Johanna Nichols

Determinate Indeterminate Stress of indeterminate -o- grade and -i- suffix: idti xodit’ exat’ ezdit’ nesti nosit’ vesti vodit’ vezti vozit’ lezt’ lazit’ bresti brodit’ gnat’ gonjat’

mobile fixed root mobile mobile mobile fixed root mobile mobile (present stem *gon-i-)

-o- grade and -(a)j- suffix: plyt’ plavat’ (-aj) -aj- conjugation and no difference in vocalism between the two stems: bežat’ begat’ (-aj-) letet’ letat’ (-aj-) polzti polzat’ (-aj-) taschit’ taskat’ katit’ katat’ (-aj-)

Thus it is possible to maintain that, synchronically, every indeterminate is based formally on the deverbal noun. The derivational types are similar in all Slavic languages preserving the determinate/indeterminate opposition, sometimes with a clearer predominance of the *-o- grade, *-i- suffix type in the indeterminates (e.g., Slovene: Herrity 2000:â•›227).

6.

Whence the indeterminate semantics?

In all of the OCS examples I surveyed (illustrative examples in §3 above), unsuffixed deverbal nouns of whatever derivational semantics (agent, fact, instrument, location, etc.) denote kinds or sorts and not specific individuals or events (though of course they can be used to refer to individuals). Therefore verbs derived from them are verbs of manner or typical activity and not telic unless the semantics of state change or affected object is independently present. Agentive intransitives from manner nominals naturally have meanings along the lines of ‘do the X activity’, just as the English verb tango means ‘do the tango’. Thus xoditi originally meant (and still does mean in South Slavic) ‘walk, walk around’. This and other indeterminates were originally manner verbs without goals (Dickey, this volume; Greenberg, this volume), and over time they become paired with the original motion verbs and begin taking goals. Analogously, voziti ‘convey’ must originally

Indeterminate motion verbs are denominal

have meant ‘transport in a carriage or cart, take for a ride in a carriage’ (rather like bus in English I’m going to bus them to San Diego) and then came to be paired with *vez-ti ‘convey, drive (someone/something) somewhere’. My approach here, deriving iterativity from denominal formation, differs from earlier work that relates the semantics of iteration to causatives but leaves denominals out of the equation (e.g., Darden n.d.:â•›1–2). Thus the denominal verbs need not have inherited any iterative meaning from PIE. They simply implied or presupposed generic time reference, and over time they have been incorporated into the developing aspect and Aktionsart systems of the various Slavic languages. In East Slavic and Polish, where iterative verbs have not developed as a class, the denominal motion verbs have stayed discrete from iteratives and make up a small closed set with the special aspect-like category of indeterminacy. They now take goals just as their determinate counterparts do. Elsewhere they have tended to merge with iteratives, and everywhere they have had at least some role in the formation of secondary imperfectives. In BCS and especially Bulgarian (Lindsey 2006, in preparation), where motion verbs have only minimally conflated manner with path (in the terms of Talmy 1985), the denominal verbs either have not developed or have been lost because motion verbs have not developed as a set. Initially all denominal verbs may well have had a non-telic, iterative, and/or manner sense as one possible interpretation, but since they functioned as transitivizers the telicity typical of causatives outweighed any other nuances. The situation was different for the motion verbs, since they had the same valence as their source verbs: iti and xoditi were both intransitive, nesti and nositi both transitive. The minimal difference in each such pair was the non-telicity and generic time reference of the denominal verb, and in this situation of minimal contrast it was salient enough to be lexicalized.

7.

Deadjectival verbs

The situation is somewhat different when the denominal verb is derived not from a noun but from an adjective. Here a different semantic factor comes into play: the adjective describes a state, the o-grade verb derived from it is telic and specifically a transitive accomplishment (‘cause to become X’), and secondary intransitives derived by reflexivization from that transitive preserve the telicity and a connection to the agentive semantics of the transitive. There are some revealing minimal pairs . Bulatova (1963) shows that they were formerly present in BCS but were lost. I thank an anonymous referee for alerting me to this work.

59

60 Johanna Nichols

where the reflexive verb is an accomplishment, implying successful reaching of a limit or threshold set by the agent, while the ancient intransitive is inchoative and probably best described as an achievement but without any clearcut threshold. An example is late CS sъxno˛ti (inchoative) and sušiti sę (telic), both translatable ‘dry out’ (Gołąb 1968). The modern Russian descendants soxnut’-vysoxnut’ and sušit’sja-vysušit’sja preserve a semantic distinction whereby soxnut’ means ‘dry out, get parched’ with the entity ending up dry as a result of natural causes and without human intention to reach some level of dryness, while sušit’sja means ‘dry, be drying’ in the sense of changing toward a threshold after which the object will be deemed dry for purposes of household management or technology. I surveyed all examples of 3sg present soxnet and sušitsja in the Russian National Corpus (RNC). There were 147 examples of soxnet as of early February 2008. These included many examples like (13) and (14) describing spontaneous and natural drying of objects and substances, and many like (15) describing parching or excessive drying (especially of skin or crops). Many of these present-tense examples have generic or iterative meaning as (13)–(15) do. (13) (14)

Ona nikogda ne zasyxaet na pečatnyx formax, no mgnovenno soxnet it never neg dries on printer forms but rapidly dries i ideal’no deržitsja na ljuboj poverxnosti and ideally holds on any surface ‘It (ink) never dries out in the printer, but dries rapidly [on printing] and adheres ideally to any surface.’ (RNC: “UF-pečat’: novyj klass pečatnyx texnologij”) pogoda Ø senokosnaja, seno v žaru soxnet bystro, no kosit’ weather (is) hay-mowing hay in heat dries fast but mow.inf tjaželo: … hard ‘It’s good haymaking weather, the hay dries quickly in the heat, but it’s hard to cut it’. (RNC: V. Rasputin. Vasilij i Vasilisa)

(15) ot detskogo myla koža na rukax soxnet po-strašnomu from children.adj soap skin on hands dries terribly ‘The skin of my hands dries out terribly if I use children’s soap.’ (RNC: “Krasota, zdorovje, otdyx: kosmetika i parfum”)

Clearly progressive examples, and examples referring to laundry drying, were very few with soxnut’. Most examples have Subject + Verb word order as in the above three. Reflexive sušitsja is less frequent (36 examples). It is most often used of laundry hung out to dry:

Indeterminate motion verbs are denominal

(16) Na verevkax sušitsja postel’noe bel’e on â•›lines dries bed.adj linen ‘On the clotheslines sheets are drying.’ (RNC: E. Čistjakova. “Litva vozvraščaetsja”)

Many examples are progressive in meaning, and most have Verb + Subject word order. Though sušit’sja ‘dry.refl’is a lexicalized intransitive verb and not a passive, and has only one argument, most tokens of it imply human agency: the laundry has been hung out by someone not simply to become drier but to reach some threshold of dryness after which it can be used or stored in the household. The implicit agency, threshold, and proneness to progressive temporality are inherited from transitive sušit’ ‘dry (something, especially laundry)’, which is an accomplishment. I do not have a good explanation of the word order, unless it is because given the reference to clotheslines and laundry the content of the verb ‘dry’ is fairly redundant and the verb and subject together contribute only one piece of information (as with most VS clauses). This comparison shows that deadjectival verbs can retain a distinctive aktionsart or Vendler class as a result of that derivation, and it lends support to the idea that verbs derived from nouns can retain aspectuality properties that reflect their derivational history. (Apart from the general effect of derivational history, the specific aspectuality of these deadjectival verbs is of course different from that of the denominal ones treated above.)

8.

Remaining open questions

The analysis proposes here raises some additional points and questions which, while not undermining the analysis, do eventually need to be answered. First, the pairing of inherited and denominal verbs as in the motion verbs is quite regular and transparent in Slavic, while other IE languages seem to offer fewer and lexically more dissociated pairs. This is important to my claim that the shift from verb-based to noun-based is relatively recent in Slavic compared to the other European branches. I assume this sort of shift occurred whenever an IE language entered the European cultural and linguistic sphere and acquired new contacts, and with Slavic this happened relatively recently. On the other hand, it could be that Proto-Slavic just maintained the pairings more faithfully (if so this was no doubt because of the developing opposition of determinate/indeterminate). It could even be that the putative regularity and transparency are just the artifactual result of Slavists attending to such pairings. It would be good to know just how distinctive Slavic is in this respect.

61

62

Johanna Nichols

Second, nothing has been said here about indeterminates and iteratives in -a(j) (Russ. plyt’-plavat’ ‘swim, float’, letet’-letat’ ‘fly’, etc.) except that they too generally bear analysis as denominal because the vowel grade of the iterative is generally that of the deverbal noun. In fact it appears that o-grade, -i-suffixed derivations and -aj-suffixed ones (with whatever root vocalism) are in complementary distribution: the -o-, -i- type corresponds to unsuffixed (in traditional terms, radical) verbs and the -aj- type to suffixed verbs. I am not sure that this statement is water-tight for Slavic and I do not know whether it obtains in other Indo-European branches at all. Third, suppose the complementary distribution is firm. Then we have a wordformation rule stating that the iterative is formed with -o-, -i- when the source verb is unsuffixed, and with -aj elsewhere. But I have argued that the derivation is denominal, not deverbal. Perhaps the complementary distribution proves me wrong; or perhaps it does not, but then the word-formation rule is an unusual one that derives a verb from a noun while also looking at the etymological source of that noun. Alternatively, perhaps o-grade deverbal nouns were formed only from unsuffixed verbs in the first place; but I am not sure that this was the case in Slavic. Fourth, though it is not critical to the topic of this paper, one cannot work on motion verbs without noticing that šьd-/xod-i- is mildly problematic. Traditionally it is derived from *sed- ‘sit’ with a sematic shift caused by the fact that one can go somewhere by sitting (on horseback or on a wagon), supported by the parallel shift in Gk. hodos ‘route, path’, cf., eks-odos ‘exodus’. Then it is one of very few Balto-Slavic verbs that escape Winter’s Law, which lengthens a vowel before a plain voiced stop, as in *sěd- ‘sit’ < PIE *sed- (Winter 1978). Kortlandt (2007) posits a reduplicated *sisd- to account for the failure, a solution that works but is not parsimonious. In addition, the initial *x- is problematic. Traditionally it is explained as due to the ruki rule applying after prefixes such as *pri-, *per-, *vъ-, and *u, but other verbs with initial *s- do not undergo the ruki rule. Perhaps prefixes were found only on motion verbs in early Proto-Slavic, but this is an additional complicating assumption. In short, problems arise with the traditional etymology of šьd-/xod-i-. Finally, the verb-framing or manner-salient linguistic type (Slobin 2004; Talmy 1985), in which motion verbs can conflate manner and still take goals, is not very common outside of northern Europe, so it cannot just be assumed that Proto-Indo-European, Proto-Balto-Slavic, and early Proto-Slavic were of this type. It could be that Common Slavic changed to this type upon entering the European linguistic sphere and coming into close contact with Germanic, so that manner salience, reflexivization, and a noun-based lexicon were a package of contact effects. Denominal derivation of verbs was one of the grammatical devices re-

Indeterminate motion verbs are denominal

cruited to help form goal-taking manner+motion verbs. More precisely, it formed pure manner-of-motion, non-goal-taking verbs (Dickey, this volume), leaving the original verb (determinates iti, lezti, etc.) to take goals.

9.

Conclusions

To summarize, indeterminate verbs arose as denominal verbs, and they owe their vocalism, their *-i- stem suffixes, and their indeterminate aspectuality to their denominal origin. The nouns they were formed from were originally deverbal, but by a combination of reanalysis and derivational productivity these deverbal nouns became the heads or bases of their word families. The development of new *-ograde, *-i-suffixed verbs from these nouns is part of a larger shift of the Slavic verbal lexicon from verb-based to noun-based. When this shift occurred some new verbs were derived from nouns and some inherited Indo-European causatives were reanalyzed as denominals. As another, separate part of this shift, denominal verbs such as sušiti ‘dry’ and čistiti ‘clean’ were created and their secondary reflexives provided new intransitive verbs or supplanted the original intransitives from which the input nouns and adjectives had originally been derived. The reanalysis of *-o-grade, *-i-suffixed verbs as denominal did not require that the *-o-grade nouns they were based on were productive, but only that they existed (or could be formed). Some of these nouns could be ancient and even Indo-European, e.g xodъ ‘pace, gait’ with its Greek cognate, but others may be more recent. The subsequent fates of the deverbal nouns varied from language to language. In Russian, for instance, -xod and -vod are productive and form compounds and prefixed words with a variety of senses (including recent formations such as lunoxod ‘moon exploration vehicle’, atomoxod ‘atomic or nuclear powered vessel’, kartoflevod ‘potato grower’, volnovod ‘(electrical) wave guide’, etc. (these words from Zaliznjak 1977). Others are less productive; -laz, for instance, forms only a handful of compounds all with the same meaning ‘one who climbs X’, e.g., verxolaz ‘high-rise construction worker’, stenolaz ‘wallcreeper’ (bird species, Tichodroma muraria). The unsuffixed deverbal noun plays the same role as conversion does in English derivation. Thus the combination of inherited deverbal nouns and a shift of derivational type which made those nouns basic to word families, taking place in the same general time frame as modern Slavic aspect began to take shape, gave rise to the distinctive class of indeterminate motion verbs.

63

64 Johanna Nichols

References Alekseev, A. A. & Gerd, A. S. (eds.). 2006. Slovar’ staroslavjanskogo jazyka. St. Petersburg: St. Petersburg University Press. Bulatova, R. V. 1963. Sud’ba sootnosimyx par glagolov dvizhenija v serbsko-xorvatskom jazyke. Uchenye zapiski Instituta slavjanovedenija AN SSSR 27: 223–258. Darden, B. J. n.d. Balto-Slavic factitive-iteratives. MS, University of Chicago. , accessed December 2008. Gołąb, Z. 1968. The grammar of Slavic causatives. In American Contributions to the Sixth Inter‑ national Congress of Slavists, H. Kucera (ed.), 71–94. The Hague: Mouton. Herrity, P. 2000. Slovene: A Comprehensive Grammar. London: Routledge. Kiparsky, V. 1975. Russische historische Grammatik, Vol. 3: Entwicklung des Wortschatzes. Heidelberg: Carl Winter. Kortlandt, F. 2007. Winter’s law again. Ms, Leiden University. , accessed December 2008. Kurylowicz, J. 1964. The Inflectional Categories of Indo-European. Heidelberg: Carl Winter. Lindsey, T. 2006. Bulgarian motion verbs: Manner and path in a Balkan context. Presented at SLS Inaugural Meeting, Bloomington IN. Lindstedt, J. 1994. Codex Suprasliensis. In Corpus Cyrillo-Methodianum Helsingiense. Helsinki: Helsinki University, Department of Baltic and Slavic Studies. Machek, V. 1968. Etymologický slovník jazyka českého. Prague: Academia. Matešić, J. 1965–1967. Ruckläufiges Wörterbuch des serbokroatischen. Wiesbaden: HarrasÂ� sowitz. Meillet, A. 1961. Études sur l’étymologie et le vocabulaire du vieux slave, Part 2. Paris: Institut d’études slaves. Pokorny, J. 1959. Indogermanisches etymologisches Worterbuch. Bern: Francke. Sadnik, L. & Aitzetmüller, R. 1955. Handwörterbuch zu den altkirchenslavischen Texten. Hague: Mouton. Schenker, A. M. 1995. The Dawn of Slavic. New Haven CT: Yale University Press. Severjanov, S. 1922/1954. Sinajskaja psaltyr’. Petrograd. (Photomechanical reprint: Graz: AkadeÂ� mische Druck- u. Verlagsanstalt, 1954). Slobin, D. I. 2004. The many ways to search for a frog: Linguistic typology and the expression of motion events. In Relating Events in Narrative, Vol. 2: Typological and Contextual Perspec‑ tives, S. Stroemqvist & L. Verhoeven (eds.), 19–257. Mahwah NJ: Lawrence Erlbaum. Stang, C. 1957. Slavonic Accentuation. Oslo: Universitetsforlaget. Talmy, L. 1985. Lexicalization patterns: Semantic structure in lexical forms. In Language Typol‑ ogy and Lexical Description, Vol. 3: Grammatical Categories and the Lexicon, T. Shopen (ed.), 36–149. Cambridge: CUP. Timberlake, A. 2004. A Reference Grammar of Russian. Cambridge: CUP. Trubachev, O. N. (ed.). 1974. Ètimologicheskij slovar’ slavjanskix jazykov Moscow: Nauka. Vaillant, A. 1966. Grammaire comparée des langues slaves. Vol. III: Le Verbe. Paris: Klincksieck. Winter, W. 1978. The distribution of short and long vowels in stems of the type Lith. ésti : vèsti, mèsti and OCS jasti : vesti : mesti in Baltic and Slavic languages. In Recent Developments in Historical Phonology, J. Fisiak (ed.), 431–446. The Hague: Mouton. Zaliznjak, A. A. 1977. Grammaticheskij slovar’ russkogo jazyka: Slovoizmenenie. Moscow: Russkij jazyk.

Indeterminate motion verbs are denominal

Sources Ps. sin. = Severjanov (1954) Supr. = Lindstedt (1994) Russian National Corpus (RNC): www.ruscorpora.ru

Appendix 1. Abbreviations used in interlinear glosses ADJ DAT GEN INF INSTR NEG REFL

adjective (derivational suffix) dative case genitive case infinitive instrumental case negation reflexive

65

chapter 3

Common Slavic “indeterminate” verbs of motion were really manner-of-motion verbs* Stephen M. Dickey University of Kansas

This paper attempts to reinterpret the class of Common Slavic indeterminate verbs of motion, e.g., xoditi ‘walk’, nositi ‘carry’, as manner-of-motion verbs. I examine the attestations of indeterminate verbs of motion in contexts of determinate motion taken primarily from Old Church Slavic and Old Russian texts to form the basis for this reinterpretation. Following a discussion of the “manner-of-motion verb hypothesis”, I argue that the development of Common Slavic manner-of-motion verbs into the functionally peculiar class of indeterminate verbs of motion in North Slavic arose as a result of the effects of prefixation in the nascent Slavic aspectual system. Finally, a hypothesis of the development of the Slavic correlation determinate/indeterminate in verbs of motion is offered.

1.

Introduction

Since Van Wijk’s influential work on Slavic aspect (1927, 1929), nearly all Slavists have assumed that Slavic pairs of simplex verbs of motion (VoM), such as Old Church Slavic (OCS) iti-xoditi ‘go on foot’, nesti-nositi ‘carry’, etc., express a distinction in determinacy, i.e., OCS iti and nesti are determinate, whereas xoditi and nositi are indeterminate (cf., Češko 1951; Dostál 1954; Němec 1958; Schuyt

* This paper has benefited from ongoing consultations with Marc L. Greenberg, and is intended as a companion piece to his article in this volume. I would like to thank Geoffrey Horrocks, Michael Shaw and John Younger for their kind help with issues involving Greek, and Svetlana Vassileva-Karagyozova for help with Bulgarian; I am also grateful for comments provided by Henning Andersen and two anonymous reviewers that have helped to improve this paper. I alone am responsible for any remaining errors and inaccuracies.

68 Stephen M. Dickey

1988). Van Wijk (1927:â•›99) characterizes the difference between the two kinds of VoM, though without using the specific terms “determinate” and “indeterminate”, in the following way: determinate VoM “express motion occurring at a specific time and in a specific direction”, whereas indeterminate VoM express “temporally and spatially unspecified motion […], including repetition or habits, abilities (such as: all birds fly; snakes slither, but horses run), or also a one-time motion event without a fixed direction, i.e., back and forth”. According to Dostál (1954:â•›36), “indeterminate actions are compound and lack a definite goal”. Van Wijk’s and Dostál’s definitions are reformulated for the purposes of this paper as follows: the opposition determinate/indeterminate distinguishes between motion occurring along a single goal-directed trajectory (determinate) versus motion not consisting of a single goal-directed trajectory (indeterminate). In addition to the expression of ability (i.e., the ability to walk, run, etc.), the category of indeterminacy has been assumed to have two primary instantiations. One is that of habitual repetition (1a), and the other is that of aimless motion, i.e., continuous motion not occurring along a single (goal-oriented) trajectory (1b): (1) a. b.

i xoždaašete roditelě ego po vьsě lěta vъ i[erusali]mъ and went parents his along all years in Jerusalem vъ prazdnikъ pasxy in holiday passover ‘and his parents went every year to Jerusalem to the feast of Passover’ (OCS; Codex Zographensis; Zog:â•›200) Xodę že pri mori galileiscěmь vidě simona · i anьdrěa walking EMPH by sea galilee.ADJ saw Simon and Andrew brata togo simona brother that Simon ‘Walking near the Sea of Galilee he saw Simon and Andrew the brother of Simon’ (OCS; Codex Zographensis; Zog:â•›48)

The idea that Common Slavic pairs, such as iti-xoditi ‘go on foot’, expressed the opposition determinate/indeterminate, in which indeterminate verbs expressed habitual repetition and aimless motion (in addition to ability; this latter meaning is, in my view, connected with aimless motion as an activity and thus irrelevant to the particular concerns of this paper), is hereinafter termed the indeterminacy hypothesis. Determinate motion as a conceptual category is relatively unproblematic and probably universal in languages of the world. However, the category of indeterminate motion is unusual, if not unique, from the perspective of linguistic typology. . Note that VoM refers to simplex verbs unless otherwise specified.

Common Slavic “indeterminate” verbs of motion

The indeterminate members (e.g., OCS xoditi ‘walk/go on foot’) of these VoM pairs are unusual as a special class, and their meaning is, as far as I am aware, unattested as a special, formally marked category outside Slavic. Though this opposition has undoubtedly become grammaticalized as a quasi-aspectual distinction in the modern East and West Slavic languages (hereinafter: North Slavic), it is less than clear that the same kind of determinate/indeterminate opposition existed in Common Slavic, contrary to typical assumptions. At least two Slavists have doubted the utility of thinking of Slavic VoM in terms of the opposition determinate/indeterminate. Vaillant (1939:â•›295–296) observes that for Common Slavic the terms determinate and indeterminate are “inexact” because the opposition is, in fact, between verbs of motion (verbes de mouvement) and stative verbs (verbes d’état). Veyrenc (1966) goes further, arguing that in modern Russian the opposition is not between determinacy and indeterminacy, but rather between motion verbs proper (verbes du deplacement) and a kind of activity verb (verbes du fonction). The validity of Veyrenc’s arguments cannot be discussed here, but the fact that such an argument can even be cogently made regarding modern Russian is another indication that stages of Slavic more obscured by the mists of time might deserve re-examination and re-assessment. Expanding on ideas outlined briefly in Greenberg and Dickey (2006), this paper bases its arguments on attestations of indeterminate verbs in contexts of determinacy in OCS and Old Russian, which reveal that the traditional category of indeterminacy does not account for the full range of data in Common Slavic, and thus that this synchronic notion is of relatively little explanatory value when applied to earlier stages of Slavic. It is argued instead that the Common Slavic indeterminate VoM were actually manner-of-motion verbs and that attestations of indeterminate VoM in contexts of determinacy are relics of that Common Slavic state of affairs. On the basis of this redefinition, a hypothesis of the development of the class of indeterminate verbs in Slavic is then presented. Manner-of-motion verbs (MoMV) are VoM that foreground the “dimensions that modulate motion, including motor pattern, rate, rhythm, posture, affect, and evaluative factors” (Slobin 2004:â•›255, fn. 5). MoMV exist in a wide array of languages and have received considerable attention in the general linguistic literature (e.g., Talmy 1985, 1991, 2000; Slobin 2004; Horrocks and Stavrou 2007). However, with the notable exception of Filipović (2007), they have received very little attention in the Slavic linguistic literature; to my knowledge, there has been

. It is important to be aware that the East Slavic languages and West Slavic languages differ with regard to the details of the morphology and usage of determinate and indeterminate VoM. For a comparison of Russian and Czech in this regard, cf., Cummins (1983).

69

70 Stephen M. Dickey

no discussion of North Slavic indeterminate VoM with regard to the category of manner-of-motion verbs to date, apart from Greenberg and Dickey (2006). The work of Talmy (1985, 1991, 2000), which has established the typology in use, divides languages into two types according to the way they express the path and manner of motion: Satellite-framed languages (S-languages) and verbframed languages (V-languages). In S-languages, the path of motion is expressed by a “satellite” particle/prefix and not in the motion verb itself, e.g., English go in, go out. In V-languages, the path is expressed directly in the verb itself, e.g., Spanish entrar ‘go in’, salir ‘go out’. As S-languages express the path of motion in a particle/prefix outside the verb, the motion verb itself is therefore essentially free to express the manner of motion as a “co-event verb”, as in English The bottle floated out. In contrast, V-languages do not express manner in the main verb expressing the core motion event (i.e., the path of the motion). V-languages can express the manner of motion in an adpositional phrase or gerundive, such as a co-event gerundive, as in Spanish La botella salió flotando ‘The bottle exited floating’, and tend not to express manner at all. It should be pointed out that recent works, such as Slobin (2004), have concluded that this typology should be considered to represent a cline, with S-languages on one end and V-languages on the other, as opposed to a single binary distinction. Although Talmy’s account of the differences between S- and V-languages in the coding of manner of motion provides a basic explanation of why MoMV do not occur with the meaning of goal-directed motion in some languages, it cannot explain the full range of data. An insightful analysis of the behavior of MoMV in various languages is offered by Horrocks and Stavrou (2007:â•›608–609), who link the use of MoMV for goal-directed motion to other elements of aspectual systems. First, they set up a parameter involving the possibility of unaccusativization of manner-of-motion verbs, i.e., whether a language can telicize MoMV with a goal prepositional phrase (PP) and focus on the attainment of the telos with a temporal in PP. The values for this parameter do not correspond to the SÂ�language/V-language division: although English, Greek and the Slavic languages are all S-languages, English allows this construction, e.g., He walked to the park in five minutes, whereas Greek and Slavic do not. Horrocks and Stavrou correlate the possibility of such MoMV use with the possibility of the resultative use of adjectives in sentences such as beat the metal flat: such usage is possible in English, but not in Greek or Slavic. Further, Horrocks and Stavrou (2007:â•›609) observe that there is an exact correlation between the possibility of the resultative use of adjectives and the absence of viewpoint aspect morphologically-encoded in verb . Note that Filipović (2007) analyzes BCS, which, like all of South Slavic, currently lacks a determinate : indeterminate opposition in VoM.

Common Slavic “indeterminate” verbs of motion

Table 1.â•‡ Horrocks’ and Stavrou’s (2007) aspectual typology: Values for English, Greek and Slavic Language English Greek Slavic

Unaccusativitization of MoMV

Resultative adj.

Viewpoint aspect

+ – –

+ – –

– + +

forms, i.e., a “grammaticalized opposition between perfective and imperfective aspect that is independent both of issues of lexical aspectual character and of telicity.” This claim is very relevant to Slavic at all stages, both early (with an aspectual system based on the aorist : imperfect opposition) and late (with the familiar derivational aspectual systems). Table 1 summarizes Horrocks’ and Stavrou’s (2007) aspectual typology with regard to English, Greek, and Slavic. Before beginning the discussion of indeterminate verbs in OCS and Old Russian, it is worthwhile to describe briefly the usage of MoMV in Bosnian/ Croatian/Serbian (BCS). The congeners of the indeterminate VoM of North Slavic are, in fact, MoMV in BCS (i.e., there is no determinate/indeterminate opposition in BCS); BCS also appears to allow its MoMV in a greater variety of contexts than Bulgarian. This makes BCS MoMV such as hodati ‘walk’ unique in a cross-Slavic perspective. It comes as no surprise that BCS uses MoMV for non-goal directed motion: (2) Hodali smo po gradu. walked AUX about city ‘We walked around the town.’

(BCS; Internet)

BCS also uses MoMV for repeated roundtrips: (3)

Mi smo nakon rata hodali u školu bosi čim bi svanulo we AUX after war walked in school barefoot as soon as would dawn proljeće. spring ‘After the war we walked to school barefoot as soon as spring broke.’ (BCS; Internet)

In this respect, the usage of MoMV in BCS is identical to the use of their indeterminate congeners in North Slavic. However, unlike North Slavic indeterminate VoM, BCS MoMV can be used as co-event gerundives:

. URLs of all Internet examples are listed in Appendix 2.

71

72

Stephen M. Dickey

(4) a. b.

Mamić je na današnju tiskovnu konferenciju koja je održana Mamić aux on today’s press conference which aux held u bolnici, došao hodajući, te je novinarima pokazao da može in hospital came walking and aux journalists showed that can micati rukama. move arms ‘Mamić came walking to today’s press conference, which was held in the hospital, and showed the journalists that he could move his arms.’ (BCS; Internet) Krenuo sam velikim autom polako prema zapadu, jezdeći širokom set out aux large car slowly toward west riding wide cestom obasjanom blistavim lampama. road illuminated shiny lamps ‘I set out in my large car slowly toward the west, driving down a wide road illuminated by shiny lamps.’ (BCS; Internet)

Perhaps strangest of all from a North Slavic perspective is the ability of BCS MoMV to express the process of goal-directed motion, which is shown in the following examples: (5) a. b.

Na prvoj stanici izašla sam van. Hodala sam do ugla at first stop went out aux outside Walked aux up to corner i onda skrenula desno. and then turned right ‘At the first stop I went out. I walked as far as the corner and turned right.’ (BCS; C:â•›110) Četiri su dana hodali do Beča. four aux days walked up to Vienna ‘They spent four days walking as far as Vienna.’ (BCS; Internet)

Examples (5a–b) resemble the unaccusativization of English MoMV in examples such as He walked to the park in five minutes. However, the resemblance is only ostensible. In He walked to the park in five minutes, the MoMV has been unaccusativized by the addition of a goal PP and thus expresses the endpoint of a telic action (i.e., the crossing of the boundary of the park). But the predicates in (5a–b) do not, in fact, express the crossing of the boundary of the landmarks (the corner and Vienna, respectively); rather, the goal PP is, as Horrocks and Stavrou (2007:â•›609) suggest, an “adjunct, specifying the arbitrary end-point” of the movement in question. Horrocks and Stavrou term this phenomenon pseudo-unaccu‑ sativization and point out that proof of the difference is that in languages such as Greek, pseudo-unaccusativized MoMV cannot co-occur with an in PP expressing the time it took for the process to reach its endpoint. Their analysis is confirmed

Common Slavic “indeterminate” verbs of motion

for BCS in the case of (5a–b) by two native speakers who unequivocally reject in PPs in these examples, as well as by the absence of such collocations on the Internet: (6) a. b.

*Za pet minuta sam hodala do ugla i onda skrenula desno. for five minutes aux walked up to corner and then turned right ‘I walked up to the corner in five minutes and turned right.’ (BCS) *Za četiri dana su hodali do Beča. for four days aux walked up to Vienna ‘They walked as far as Vienna in four days.’ (BCS)

Note that in the case of (5b), the inclusion of the bare accusative četiri dana ‘for four days’ contrasts starkly with the unacceptability of *za četiri dana ‘in four days’ in (6b). These data show clearly that BCS congeners of North Slavic indeterminate VoM behave, in fact, like MoMV in other languages such as Greek, cf., the data presented by Horrocks and Stavrou (2007), i.e., they are atelic VoM that express manner of motion. This fact is important because the usage of BCS MoMV described above appears to match the usage of the putative indeterminate VoM in OCS and Old Russian (to be discussed in depth in Section 2). As I have suggested before (Dickey 2005:â•›32, 44), the BCS aspectual system appears to be conservative compared to the aspectual systems of all other Slavic languages (i.e., North Slavic as well as Bulgarian and Macedonian). I thus consider it possible that the current behavior of BCS MoMV provides a kind of window into the situation in OCS and Old Russian.

2.

Indeterminate verbs in contexts of determinacy in Old Church Slavic and Old Russian

In Section 1 it was noted that Common Slavic indeterminate verbs have been assumed to have two basic interpretations, habitual repetition (i.e., repeated trips) and aimless motion (continuous motion not occurring along a single goal-oriented trajectory). The notion of indeterminacy as an explanatory category is tenable insofar as most attestations of indeterminate VoM in early Slavic texts express either repeated trips or continuous non-goal oriented motion. The problem is that in OCS and Old Russian (as well as in older stages of Czech, Polish, and BCS), indeterminate VoM are attested also in contexts of determinate motion. This . A Jakobsonian structuralist might object that indeterminate VoM are unmarked for motion occurring along a single trajectory and can thus occur in contexts of determinacy with

73

74

Stephen M. Dickey

descriptive fact is not new, but has been recognized by Dostál (1954:â•›484–489), who observes that indeterminate verbs could express uninterrupted, determinate (nepřerušovaný a determinovaný) motion. Similar observations were previously made by Češko (1951), in terms of her own hypothesis of the meaning of OCS indeterminate VoM (a discussion of which cannot be given due to limitations of space). Anstatt (1997) discusses the actional/aspectual differences between Old Russian and Modern Russian VoM (determinate and indeterminate) and concludes that Old Russian indeterminate VoM were aspectually vaguer than their Modern Russian counterparts. Despite the recognition of such OCS and Old Russian usage, however, no one has taken what I view to be the logical next step: reconsidering the overall semantic nature of the indeterminate VoM. This section presents and discusses particularly clear examples of such usage in OCS and Old Russian, though space limitations preclude a treatment of the full set of relevant data. These cases basically break down into two types: the occurrence of participles and gerundives profiling a manner co-event, which specifies either the manner of motion event or a concomitant transitive action (e.g., carrying something), and finite (preterite) usage. Modern Russian equivalents are given for some examples to show how modern usage differs. The intransitive VoM xoditi ‘walk’, ězditi ‘ride’, and běgati ‘flee’ are discussed first, followed by the transitive VoM nositi ‘carry’, voditi ‘lead’, and voziti ‘transport’.

2.1

Xoditi ‘walk’

Let us first consider examples of xoditi ‘walk’. Example (7) contains two instances of the participle xodę ‘walking’ in a context which precludes repetition and nongoal directed motion: (7)

vъ četvrъtŏjŏ že stražŏ nošti ide kъ nimь is[ousъ] xodę in fourth EMPH watch night went toward them Jesus walking po morju · [26] i viděvъše i oučenici xodęštь po along sea [26] and having.seen him disciples walking along morju sъmętošę sę gl[agol]jŏšte · ěko prizrakъ estъ · i otъ sea distress self saying that ghost is and from straxa vъzъpišę · fear shouted.out

impunity. I reject this view for a host of reasons that cannot be adequately presented here; some important arguments against invoking a markedness relationship in the case of indeterminate VoM are made by Kučera (1980).

Common Slavic “indeterminate” verbs of motion

‘During the fourth night watch Jesus went to them walking upon the sea. [26] And seeing him walking on the sea the disciples became distressed, saying that it was a ghost, and they shouted out with fear.’ (OCS; Zographensis; Zog:â•›21)

Češko (1951:â•›333) points out that in contrast to OCS, modern Russian must use determinate verb forms, e.g., uvideli ego iduščim po morju ‘saw him going [determinate] upon the sea’, because he was heading straight toward the boat. Note that the first instance resembles the structure used to encode manner of motion in verb-framed languages such as Spanish, i.e., with an adverbial participial phrase: La botella entrò flotando a la cueva ‘The bottle entered floating into the cave’. Example (8) contains a finite verb form, an imperfect, as an independent predicate, though the sense is a description of the manner of Peter’s motion after he gets out of the boat and as he approaches Jesus. (8)

i izlězъ is korablě petrъ · xoždaaše na vodaxъ · i pride and having.gone.out out ship Peter walked on waters and came ky is[ouso]vi. to â•›Jesus ‘And having climbed out of the boat, Peter walked on the water, and came to Jesus.’ (OCS; Zographensis; Zog:â•›21)

Again, the context is that of a single goal-oriented trajectory, but we have a socalled indeterminate verb form. If one assumes that xoditi was, in fact, a kind of MoMV meaning ‘walk’, its occurrence in (7)–(8) can be motivated very easily. There is a clear reason for

. In any discussion of OCS translations of the Bible, the influence of the original NT Greek must be taken into account. The forms of xoditi in (7)–(8) translate forms of Greek peripatō, a New Testament Greek verb that meant ‘walk’ in various contexts. Note that the prefix in peripatō would indicate a meaning of ‘walk about’, which recalls the traditional account of OCS xoditi as an indeterminate VoM; however, peripatō does not occur solely in the meaning of ‘walk about’, and many of its occurrences in the New Testament mean simply ‘walk’ in a very underspecified sense (and often in a metaphorical sense, ‘live’) without a specifically perambulatory meaning. This verb also occurs in the New Testament with at least one directional preposition, cf., en sofiai peripateite pros tous eksō ‘Walk in wisdom toward them [that are] without’ (Colossians 4:â•›5), so that Schawaller’s (1990:â•›130) characterization of peripatō as “unspecified with regard to direction” (richtungsmäßig unbestimmt) should be taken literally (she does not adequately consider all the data in her discussion). Though peripatō seems to have meant exclusively ‘walk about’ in Ancient Greek, starting with New Testament Greek it occurs in contexts such as that of the OCS example (8) above, in which a perambulatory interpretation is inappropriate: apēlthe pros autos ho Iēsous, peripatōn epi tēs thalassēs ‘Jesus went toward them walking on the water’. Such usage is also attested in Medieval Greek

75

76

Stephen M. Dickey

profiling the manner of motion in these narratives: movement along a surface of water is ordinarily impossible for humans not in some kind of floating craft, and thus the focus of the episode is on the highly unusual (if not miraculous) manner in which the characters move across the water. Moreover, the assumption that xoditi, in these examples, is a manner-of-motion verb meaning ‘walk’ avoids the problems involved with accounting for the use of a specifically indeterminate VoM in a context of determinacy. Finally, it would be patently absurd to assume that the use of xoditi in these examples asserts that Jesus and Peter were walking about, this way and that, on the water. Examples of verba percipiendi with complement clauses containing indeterminate VoM, such as in ex. (7), are numerous in old Slavic texts, including older West Slavic texts, which are considerably younger than OCS texts and the Old Russian chronicles. A particularly interesting Old Czech example from the translated Tristran Epic is given, along with the Middle High German original in (9): (9) a. A když malo nahoru kroči, malého mužíka tu ihned but when little uphill stepped small man here immediately zuoči, an po lese tu chodieše a velmi sě bojieše. spied who along wood here was.walking and very feared ‘But when he [Tinas] stepped a little uphill, he immediately spie a little man there, who was walking in the wood and very afraid.’ (Old Czech; Tristran Epic; ATE:â•›254] b. […] dô sach he den getwerg vor im in dem holze gân. there saw he the dwarf before him in the wood go ‘[…] there he saw the dwarf before himself going in the wood.’ (Middle High German; Tristran Epic; ATE:â•›255)

In this case, there is nothing in the Middle High German original that would indicate a non-goal oriented trajectory; on the contrary, gân ‘go’ was the Middle High German generalized go-verb, and in this instance the trajectory that it expresses is simply located in the wood. One could assume that the Old Czech as well, cf., the following example from John the Damascene: aneōichtheisōn de tōn petrōn, eksēlthen ho hagios peripatōn, tous ofthalmous echōn ekpepēdēkotas ‘when the rocks were opened, the holy man came out walking, his eyes having fallen out’. In this example the focus is on the manner because the saint was thought to be dead and thus, unable to walk. And in Modern Greek peripatō simply means ‘walk’, whether in a perambulatory sense or not, cf., the following example form Horrocks and Stavrou (2007:â•›612): O Orestis perpatuse ja to stathmo, otan… ‘Orestis was walking for the station, when…’. Thus, in New Testament Greek peripatō appears to have had a non-perambulatory sense of ‘walk’, i.e., it could express ‘walk’ as manner-of-motion, and accordingly the New Testament Greek data is not problematic for the analysis given here, that xoditi was translating NT Greek peripatō as a MoMV.

Common Slavic “indeterminate” verbs of motion

translator reconstrued the situation; however, it is just as simple to assume that the Old Czech translation profiles the physical activity actually witnessed by Tinas, that of the dwarf walking, with a MoMV that, as such, makes no reference to the trajectory of the motion event. This interpretation is supported by another passage, which occurs just a little later in Tristan and makes explicit reference to going around: (10)

Tak dlúho s tiem okolo jidechu, až krále na to navedechu, so long with that around went until king on that led.on žež jim to slíbi zajisté, střieci chtě Izaldy […] that him that promised firmly watch would Izalda ‘They went about with that [trying to persuade the king] so long, until they led the king to promise them firmly that he would keep guard over Izalda […]’ (Old Czech; Tristran Epic; ATE:â•›258–260)

This example contains a context typical for an indeterminate verb in the modern Slavic languages: non-goal oriented motion with a nuance of habituality. And yet, the Old Czech does not contain the “indeterminate” VoM choditi. If Old Czech choditi was specifically an indeterminate version of ‘go’, it would be very difficult to explain the choice of okolo jidechu, lit. ‘went around’, in this example. On the other hand, the choice makes complete sense if Old Czech choditi profiled the activity of walking as a manner of motion, since walking as an activity is completely irrelevant to the situation depicted in (10). Thus, assuming that Old Czech choditi was a MoMV accounts for the occurrence of chodieše in (9a), while allowing for the contrasting occurrence of okolo jidechu in (10). Old Russian participial usage similar to (7) is shown in example (11): (11)

Polovcě že, priemьše gradъ, zapališe ognemь, i ljudi Polovtsians EMPH having.taken city torched fire and people razdiliša i vedoša ja i vežě k serdoboljamъ svoimъ i split.up and led them and tents to families own and srodnikomъ svoimъ; mučimi zimoju i ocěpljajemě, u alъčbě i kin own tormented winter and torn in hunger and v žažě i v bědě, pobleděvši lici i počernivše telesy in â•›thirst and in misery turned.pale faces and blackened bodies neznaemoju stranoju, jazykomъ ispalenomъ, nazi xodjašče i bosě […] unknown country tongue ‘The Polovtsians, having taken the town, set it alight, and split up the people, and led them and their tents to their families and kin; tormented,

. NB: the Middle High German original of this passage is missing and not given by the ATE.

77

78

Stephen M. Dickey

frostbitten, hungry, thirsty and miserable, with pale faces, blackened bodies, in an unknown land, with swollen tongues, walking naked and barefoot […]’ (Old Russian; PVL; BLDR 1:â•›256)

In this example, the transitive determinate VoM vedoša ‘led’ indicates a linear trajectory towards a goal, whereas nazi xodjašče i bosě ‘walking naked and barefoot’ is an adverbial phrase describing how the poor captives were moving on the journey. Again, the Modern Russian translation (BLDR 1:â•›257) does not employ an indeterminate VoM, but opts for the manner-of-motion verb bredja ‘trudging’. The information in the clause indicates solely an interest in their condition and not in any spatio-temporal parameters of their motion. Assuming that xoditi here is a MoMV comports quite well with the nature of the information given in the passage. An interesting example of Old Russian xoditi in a possible context of determinacy is given in (12) along with its Modern Russian translation: (12) a. b.

Onъ že izyde izъ grada sъ uzdoju i xožaše skvozě he EMPH went.out out city with reins and walked through pečeněgy, glagolja: «Ne vidě li konja niktože?» Pechenegs saying not seen INTR horse no one ‘He went out from the town with the reins and walked through the Pechenegs, saying: “Has no one seen a horse?”’ (Old Russian; PVL; BLDR 1:â•›114) On že vyšel iz goroda, derža uzdečku, i prošel he EMPH went.out out city holding reins and went.through čerez stojanku pečenegov, sprašivaja ix: «Ne videl li kto-nibud’ through camp pechenegs asking them not saw INTR anyone konja?» horse ‘He went out from the town with the reins and went through the Pechenegs, saying: “Has no one seen a horse?”’ (Modern Russian; BLDR 1:â•›115)

Example (12a) is ambiguous because the trajectory expressed by the preposition skvozě ‘through’ is necessarily linear only if its object is narrow, cf., svkozě vorota ‘through the gate’ versus ězditi skvozě mixailovu volost’ ‘ride through/about the Mixailov district’; the plural pečeněgy ‘Pechenegs’ provides no clarity in this regard. In my view, a linear trajectory is likely, due to the overall context of the boy attempting to make his way through the Pecheneg camp and secure aid for the . I am grateful to an anonymous reviewer for pointing this out and providing these examples.

Common Slavic “indeterminate” verbs of motion

besieged city of Kiev. Though the interpretation indicated by the Modern Russian translation given in (12b), prošel čerez stojanku pečenegov ‘passed through the Pecheneg camp’ is not decisive in this regard, it is nevertheless worth pointing out that its comports with this construal. In addition, a close consideration of the context independently motivates xoditi as a MoMV. First, the boy is pretending that he has lost his horse and is carrying a set of reins, which brings into focus the manner of his locomotion (walking vs. riding horseback). Second, he is placing himself in considerable danger by leaving the cover of Kiev and passing so close to the Pechenegs, so it is natural to focus on the fact that he is limited to foot travel on his way through the camp (and has to make a break for it once he nears the river). Thus, I conclude that there is no motivation for a specifically indeterminate VoM in this passage, and consider it very likely that (12a) is another instance of xoditi in a context of determinacy. Another prominent use of xoditi in contexts of determinacy involves military campaigns. In the Primary Chronicle and the Novgorod Chronicle, xoditi occurs very frequently when someone embarks on a military campaign, as shown in the examples in (13). (13) a. b.

Jaroslavъ xodi na jatvjazě, synъ Svjatopolčь, i pobědi ja, Jaroslav walk on Jatvjags son Svjatopolk’s and vanquished them i, prišedъ s vojny posla Novgorodu i poja and having.come from war sent to.Novogorod and took Mьstislavlju dščerь sobě ženě, Volodimerju vnuku, měsjaca Mstislav’s daughter to.self wife Vladimir’s granddaughter month maija vъ 12, a privedena bystь ijunja vъ 29. may in 12 and brought AUX June in 29. ‘Jaroslav, son of Svjatopolk, marched on the Jatvjags and defeated them, and arriving from the war, sent [envoys] to Novgorod and took Mstislav’s daughter for his wife, Vladimir’s granddaughter, on the 12th of May, and she was brought on the 29th of June.’ (Old Russian; PVL; BLDR 1:â•›304) Xodiša novgorodci voinoju za more, v Polnuju rěku, marched Novogorodians in.war beyond sea in Polnaja river i mnogo voevaša, i vzjaša Ljuderevъ gorod sumьskogo knjazja and much warred and took Ljuder’s city suomi prince i piskuplj; i pridoša v Novgorod vsi zdorovi. and bishopric and came in Novgorod all healthy ‘The Novgorodians marched in war beyond the sea, to the River Polnaja, and warred much, and took Ljuder’s city of the Suomi prince and the bishopric; and they came to Novgorod all healthy.’ (Old Russian; NPL:â•›371)

79

80 Stephen M. Dickey

In such examples, xoditi appears to express a trajectory to a single destination. The sense of going on a military campaign is easily derived from xoditi as a manner-of-motion verb ‘walk’, as the chief mode of such campaigns was originally marching or going on foot. In many languages, it is common to refer to the movements of warriors in terms of the manner of their motion, i.e., for viewing such movement as a special kind of activity, as opposed to mere ‘going’. Examples are easy to find; compare, for instance, the difference between English The legions marched on Rome and The legions went to Rome. Johannet (1957:â•›86) suggests (as does an anonymous reviewer) that aorist forms of xoditi in the Chronicles function specifically to express a single two-way trip as opposed to a simple trajectory to a goal, as do past-tense forms of Modern Russian xodit’ ‘go [indeterminate]’, e.g., Včera my xodili v kino ‘Yesterday we went to the cinema [and came back]’. Anstatt (1997:â•›16), on the other hand, points out that a two-way trip is evident from the context only in some cases, and does not posit ‘single roundtrip’ as a specific meaning of xoditi. While it is true that in such military contexts, aorist forms of xoditi are often followed by some mention of the return (often a formulaic expression such as i pridoša … vsi zdorovi ‘they all came [back] healthy’), in my view the idea that xoditi in examples such as (13) serves primarily to express a single two-way trip is problematic for several reasons. First, both determinate VoM and prefixed VoM also occur in clear contexts of a twoway trip, as shown in (14): (14) a. b.

Ide knjazь Jaroslavъ sъ novogorodci na Emь i povoeva went prince Jaroslav with Novgorodians on Häme and conquered vsju zemlju, i polonъ privede beščisla. all land and plunder brought numberless ‘Prince Jaroslav went with the Novgorodians against the Häme and conquered the entire land, and brought limitless plunder.’ (Old Russian; NPL:â•›65) Priidoša Polovci prъvoe na Ruskuju zemlju voevatъ; Vsevolodъ came Polovtsians first on Russian land wage war Vsevolod že izide protivu imъ měsjaca fevralja vtoryi denь; i emph went.out against them month February second day and bivšisь imъ, pobědiša Vsevoloda, voevaša i otъidoša. fought them defeated Vsevolod warred and went.away ‘The Pechenegs came for the first time to wage war on the Russian land; Vsevolod went out against them on the second day of Febrary; they fought, and defeated Vsevolod, waged war and left.’ (Old Russian; NPL:â•›183)

Common Slavic “indeterminate” verbs of motion

Example (14a) is practically identical in terms of narrative structure to the context of a single two-way trip in (13). Given the fact that determinate VoM such as iti ‘go’ and prefixed VoM such as priiti ‘arrive’ occur in the Old Russian chronicles in the very same context of a subsequent departure, it is hard to imagine how Old Russian xoditi ‘walk’ could be functioning to express single two-way trips on par with Modern Russian xodit’ ‘go [indeterminate]’. Second, though Anstatt (1997:â•›15) does attest an example from a birchbark document that lacks a military context, the great majority of occurrences of goal-directed xoditi attested in Old Russian involve attacks, and a majority of those include the phrase xoditi na + accusative, i.e., ‘march on/against’. If goal-directed xoditi occurred in Old Russian merely to signal two-way foot travel, we should expect it to occur in contexts outside of warfare much more frequently than it does. This is not to say that goaldirected occurrences of xoditi were limited at all times and for all speakers to the meaning of ‘march/go on a military campaign’. I would, however, argue that the occurrences of goal-oriented xoditi in contexts of warfare can be best explained by assuming that the meaning of xoditi ‘go on a military campaign’ was derived from its basic manner-of-motion meaning, i.e., ‘walk’. I do consider it possible that the examples in (13) represent some early stage of the ‘single two-way trip’ function of indeterminate VoM that developed by the Modern Russian period, and in my view this function can be quite plausibly derived from the aspectual nature of Old Russian MoMV, such as xoditi ‘walk’, as atelic verbs that do not profile the crossing of the boundary of a destination (cf., the discussion of the BCS exx. in (5) in Section 1) applied to the motion involved in the military campaigns of Old Rus’. These military campaigns were not simple trips to destinations, but rather processes of moving groups of warriors over long distances; nor were they ordinarily missions conceived for the purpose of remaining at the destination, much less permanently occupying it; rather, they were raids or sorties carried out for the purpose of inflicting as much damage as possible, pillaging, looting and securing a peace agreement from the opposing side, followed by the return trip home. In view of these realities, I suggest that the original meaning of these verb forms is that of the manner of motion as a process proceeding in a certain direction; inasmuch as xoditi acquired associations with two-way trips based on the style of warfare in Old Rus’, this was a secondary phenomenon.

. Renee Perelmutter points out that the Modern Russian equivalent of Old Russian xoditi na in these contexts is in fact perfective determinate pojti na, lit. ‘go on’. Given the entrenchment of the roundtrip meaning of xodit’ in Modern Russian, the difference also speaks against the idea that Old Russian xoditi in cases such as (13) specifically expressed a roundtrip construal of the motion event.

81

82

Stephen M. Dickey

This hypothesis is the simplest account, in view of the other data for xoditi given in (7)–(9), (11) that are problematic for the indeterminacy hypothesis.

2.2

Ězditi ‘ride’

Though ězditi ‘ride’ occurs relatively infrequently in OCS texts, it is nevertheless attested in contexts of determinacy, as shown in (15); in this example, the context (that of Acts 8:â•›28) is what Candace’s eunuch was doing on his way back from Jerusalem (i.e., reading the prophet Isaiah): (15) bě že vьzvraštae se i ězde na kolesnici svoei was emph returning self and riding on chariot own ‘he was returning and riding on his chariot’ (OCS; Praxapostolus slepčensis; SJS IV:â•›940)

In the participial construction, ězde ‘riding’ is on equal footing with vьzvraštae se ‘returning’, though the latter clearly expresses the actual motion of the event; both may be taken to be backgrounded relative to the predicate ‘was reading’, as in the original Greek ēn te hupostrefōn kai kathēmenos epi armatos autou, kai aneginōske ton profētēn Hēsaian ‘[he] was returning and sitting on his chariot, and reading the prophet Isaiah’ (I have not had access to any of the OCS versions, but I assume that they likewise employ the imperfect for the reading event; note that here ězde translates Greek kathēmenos ‘sitting’, which is not even a motion verb). Old Russian attests ězditi more frequently, and it is not surprising that there are numerous attestations of this verb in contexts of determinacy. An interesting example is given in (16), in which the participle indicates the manner of the goaloriented motion event vъzide ‘ascended’ in the main clause. (16)

Vъzide ty ězdę na svojei krъvi, jako na kolesnici. ascended you riding on own blood as on chariot ‘You ascended riding on your blood, as on a chariot.’ (Old Russian; Služebnaja mineja za sentjabr’ Sinoldal’skoj tipografskoj bib‑ lioteki XII v.; MSDJa 3:â•›1624)

In example (17a), the context is an attack by the Russian princes on the Polovtsians in 1185 and is determinate, as the Russians leave their homes with one destination, the Polovtsian positions; in (17b) Vsevolod’s forces make a journey to Bulgaria.

Common Slavic “indeterminate” verbs of motion

(17) a. b.

Volodimerъ že Glěbovičь vnukъ Jurgevъ · ězdęše naperedu v Vladimir emph Glebovič grandson Jörg’s riding ahead in storožix s Perejaslavci. advance.guard with Perejaslavs ‘Volodimer Glebovič, grandson of Jörg, was riding ahead in the advance guard with the Perejaslavs.’ (Old Russian; LavL:â•›395) [I] pride v zemlju Bolgarskuju […] i perestoęv tu v [and] came in land Bulgarian and stood.across there two d[ь]ni · poide na tretii d[ь]nь k Velikomu gorodu · storožemъ days went on third day to great city advance.guard naperedъ ězdęščim[ъ]. ahead riding ‘And [Vsevolod] came to the Bulgarian land […] and having spent two days there, on the third day set off for the Great City, the advance guard riding ahead.’ (Old Russian; LavL:â•›389)

The meaning of ězditi ‘ride’ in these examples is particularly clear in (17b), in which prefixed forms of iti ‘go’ occur as generalized motion verbs: pride ‘came’ in fact refers to travel by boat, and only poide ‘set off ’ refers to overland travel. In this light, the occurrence of ězdęščim[ъ] ‘riding’ is hardly necessary from the perspective of narrating their movement, but rather is motivated by a focus on the manner in which they move and refers to a particular military function of horsemen. Note that both examples appear to contain a set military phrase including the words (storoži) ězditi naperedъ ‘(advance guard) ride ahead’. In all the early modern Slavic languages, ězditi ‘ride’ and its derivatives are attested with meanings of military (and other kinds of official) horsemanship, and it is possible that these meanings were already in development in Common Slavic in a manner similar to the meaning of xoditi ‘march/go on a military campaign’ discussed above. Another specific use of ězditi that recalls the military meaning of xoditi ‘march/go on a campaign’ is its occurrence in the First Novgorod Chronicle to refer to trips made by members of the Orthodox priesthood on church missions, exemplified in (18). (18)

I toi oseni ězdilъ vladyka Semeonъ vъ Pьskovъ i měsjacь and that autumn rode bishop Simeon in Pskov and month sudilъ. judged ‘And that autumn Bishop Simeon rode to Pskov and acted as judge for a month.’ (Old Russian; NPL:â•›411)

This example contains reference to a single goal-oriented trajectory, though the limitation of the following activity to a month implies a subsequent departure.

83

84

Stephen M. Dickey

Inasmuch as this context may be taken to represent a two-way trip, the commentary on the ex. in (13) applies here too, as there are repeated references in Old Russian chronicles to short trips by members of the priesthood to perform some religious duty, make an appointment, etc. Such trips were not for the purpose of staying at the destination.

2.3

Běgati ‘flee/run’

In OCS and Old Russian, unprefixed běžati/běgati tends to mean ‘flee’ as opposed to simply ‘run’. It should be pointed out that fleeing is in fact always determinate in some sense, as this notion typically involves moving as directly as possible away from some danger: there is no aimless flight equivalent to aimless walking around. Thus, it is reasonable to assume that the only opposition according to traditional accounts of VoM semantics would be determinate běžati vs. durative or iterative běgati. In any case, Dostál (1954:â•›500) observes that běgati ‘flee’ could express not only repeated but also “uninterrupted”, i.e., determinate motion events. Given the frequency with which běžati/běgati occurs in OCS and Old Russian texts, it is not surprising that there are numerous attestations of běgati in contexts of determinacy. Consider first the Old Russian example in (19): (19) I prinesoša ì k Berestьju, běgajušče s nimъ. and brought him to Berest’je fleeing with him ‘And they carried him to Berest’e, fleeing with him.’ (Old Russian; BLDR 1:â•›186)

This is another example of a participle of an indeterminate verb expressing a manner co-event by adverbially modifying a prefixed verb expressing determinate motion. The Modern Russian translation (BLDR 1:â•›187) resorts to bežavšie ‘having fled’, as it cannot employ a verbal adverb of the indeterminate begat’ ‘run’ in such a context. Another Old Russian example contains the three possible forms of ‘flee’: a prefixed proto-perfective verb, determinate běžati, and indeterminate běgati, all referring to the same event: (20)

I bě sěča zla, i odva odolěvъ k věčeru and was fighting wicked and hardly having.overcome to evening Jaroslavь. I poběgoša pečenězě razdno i ne vědaxusja, kamo Jaroslav and fled.away Pechenegs all over and not knew whither běžače, i ovii, běgajušče, tonjaxu v Sitomli, iněi že vo iněixъ fleeing and these fleeing drowned v Setomli others emph in other rekaxъ, i tako pogiboša, i prokъ ixъ proběgoša i do sego dni. rivers and thus perished and remainder their fled.through up.to this day

Common Slavic “indeterminate” verbs of motion

‘And there was a cruel battle, and Jaroslav barely emerged victorious by evening. And the Pechenegs fled off in all directions and did not know to where they were fleeing, and some, fleeing, drowned in the Setomli; others [drowned] in other rivers, and perished, and the remainder have fled to this day.’ (Old Russian; BLDR 1:â•›192)

Given the close proximity of the various forms of ‘flee’ in this example, they deserve an explanation. The first form, poběgoša ‘fled/ran off ’, is spatially ablative (po- being at the time spatially ablative and not ingressive (cf., Dickey, to appear)). The second, determinate form, kamo běžače ‘to where they were fleeing’, refers specifically to some actual motion along a definite trajectory. The third, indeterminate form, běgajušče ‘fleeing’, does not refer to motion as such but merely communicates the activity in which the Pechenegs were engaged when they drowned. Given the determinate context, Modern Russian employs the directional prefixed impf ubegaja ‘running away’. OCS attests an imperfect form of běgati in a determinate context, given in (21). (21) Psalmъ davydovъ egda běgaše otъ lica [a]vesečouma s[y]na svoego. psalm david’s when was.fleeing from face Absalom son own ‘A psalm of David when he was fleeing from Absalom his son.’ (OCS; Psalterium Sinaiticum; Leskien 1962:â•›215)

This is a remarkable example of an indeterminate verb occurring in a context of determinate motion. It translates the Greek imperfect apedidrasken ‘was running away/fleeing’ and thus represents the literal meaning of the verb and not the metaphorical ‘shun, avoid’. As in example (20), the best way to interpret this example is in terms of what David was doing, i.e., what activity he was engaged in, as opposed to where he was going. Such an interpretation calls to mind a stative construal (cf., again Vaillant’s 1939 view cited in Section 1) of the flight, akin to English ‘be on the run/lam’, and obviates the issue of the exact nature of his trajectory.

2.4

Nositi ‘carry’

OCS and Old Russian nositi ‘carry’ are attested in the same kinds of usage as the intransitive VoM discussed above, i.e., participial constructions modifying a directional motion event expressed by a determinate or prefixed verb as well as finite main clause usage. This section considers nositi ‘carry [indeterminate]’, as it is the most common transitive VoM in OCS and Old Russian. I assume that nositi profiles the agent-patient relationship of nesti ‘carry [determinate]’ with

85

86 Stephen M. Dickey

the motion component backgrounded. In this sense, nositi ‘carry’ is likewise relatively stative (cf., Vaillant’s 1939:â•›295–296 view, cited in Section 1). It is for this reason that its gerundive may co-occur with determinate and prefixed VoM as a supporting event (i.e., a “manner” co-event), as in example (22). (22)

I nosę sobě krьstь izide vъ naricaemoe kranievo město · eže and bearing to.self cross went.out in called skull.adj place which glagoletъ sę evreisky golъgofa. speaks self Hebrew Golgotha ‘And himself bearing the cross he went out to [a place] called the place of the skull, which in Hebrew is called Golgotha.’ (OCS; Evangelarium Assemani; Assem:â•›237)

In (22) there is a very clear path to a single destination, the trajectory along which is profiled by izide ‘went out’; the participle nosę ‘bearing’ simply expresses a coevent, the activity that Jesus was engaged in as he went. Note that there is variation among OCS texts here; Zographensis and Marianus both contain determinate nesy ‘carrying [determinate]’ instead of nosę ‘carrying [manner]’. I interpret this variation as indicating that at this time the MoMV were undergoing semantic changes whereby they would become indeterminate VoM in the modern sense in North Slavic. Example (23) shows nositi in Old Russian with the same function: (23)

I pridosta kъ goramъ kievьskymъ, i uvidě Olgъ, jako Oskoldъ i and came to hills Kievan and saw Oleg how Oskold and Dirъ knjažita, i poxoroni voi vъ lodьjax, a drugyja nazadi Dir rule and sheltered troops in boats and others behind ostavi, a samъ pride, nosja Igorja moloda. left but alone came carrying Igor’ young ‘And they arrived to the Kievan hills, and Oleg saw that Oskold and Dir ruled, and hid his troops in boats, and left the others behind, and came alone, carrying the young Igor’.’ (Old Russian; BLDR 1:â•›76)

Example (24) shows an imperfect form of nositi that likewise occurs to specify the manner of Svjatopolk’s flight after he has taken ill. (24)

I kь večeru odolě Jaroslavъ, a Svjatopolkъ běža. Běžaščju že and to evening overcame Jaroslav and Svjatopolk fled fleeing emph emu, i napade na nь běsь, i raslaběša kosti ego, i ne him and attacked on him demon and weakened bones his and not možaše sěděti na koni, i nošaxutь ì vь nosilkaxъ. I prinesoša could sit on horse and carried him in stretcher and carried.to

Common Slavic “indeterminate” verbs of motion

ì k Berestьju, běgajušče s nimъ. him Berest’je fleeing with him ‘And Jaroslav was victorious by evening, and Svjatopolk fled. He was fleeing, and a demon fell upon him, and his bones gave out, and he could not sit on his horse, and they carried him on a stretcher. And they carried him to Berest’je, fleeing with him.’ (Old Russian; BLDR 1:â•›186)

Here the aorist and absolutive participle of běžati express the inception and continued directional path of the flight, and at the end, the aorist prinesoša ‘carried to’ expresses the arrival at the destination. In the middle, the indeterminate imperfect nošaxutь ‘carried’ expresses nothing more than the manner in which his entourage kept him moving. Note that Modern Russian (BLDR 1:â•›187) prefers the determinate nesli ‘carried’ in this context.

2.5

Voditi ‘lead’

In OCS and Old Russian voditi ‘lead’ occurs less frequently than nositi ‘carry’, and much less frequently than xoditi ‘walk’. Nevertheless, it is attested in determinate contexts, as shown in (25). (25)

Zělo ljutě běsęštъ sę junoša priveden bystъ kъ blaženououmou very harshly possessed youth brought was to blessed Antoniju · Vъzьrěvъ že velikyi Antonii na junošŏ, glagola Antonii having.seen emph great Antonii on youth said vodęštiimъ i · něstъ se moje dělo. leading.dat.pl him is.not this my affair ‘A youth very wickedly possessed by a demon was brought to the blessed Antonii. Having looked upon the youth, he said to those leading him: This is not my affair.’ (OCS; Suprasliensis; Supr:â•›171)

This substantive use of a participle of voditi refers to those leading (bringing) the youth, i.e., the unnamed agents of the determinate predicate privedenъ bystъ ‘was brought’, and the function of the verb is simply to refer to them in terms of the agent-patient configuration profiled by vesti-voditi (i.e., leading a human on foot) and not with regard to the actual trajectory of the motion. In other words, it refers to their role, and not their actual movement. A passive example is given in (26): (26) azъ že ěko agnę bezlobno · vodimъ na zakolenie I emph like lamb unwicked being.led on slaughter ‘and I, like a meek lamb, am led to the slaughter’ (OCS; Paroemiarion Grigorovičianum; SJS 1:â•›206)

87

88

Stephen M. Dickey

Here repetition and aimless motion are precluded by the context, though this example is metaphorical in nature, and the sense is one of an enduring state as opposed to real motion. This calls to mind the occurrence of indeterminate voditi in metaphorical expressions of the direction of a path, exemplified in (27). (27) Sladъka jestъ slastь vodęštija v grěxъ. sweet is pleasure leading in sin ‘Sweet is the pleasure leading to sin.’

(OCS; Suprasliensis; Supr:â•›350)

In Modern Russian such phrases can only contain determinate vesti ‘lead’, e.g., put’, veduščij v pogibel’ ‘the path leading to ruin’, strast’ veduščaja k uspexu ‘the passion leading to success’. The stative unidirectionality of such metaphorical phrases is incompatible with Modern Russian indeterminate VoM. In OCS and Old Russian it was the stativity of the context that motivated the MoMV; actual movement by some trajector would be required for determinate vesti to occur.

2.6

Voziti ‘transport’

Data for voziti ‘transport’ are scarce both in OCS and Old Russian; in both languages, however, we find attestations in contexts of determinate motion. In OCS texts this verb occurs primarily as a reflexive, voziti sę ‘sail’, a passive example of which is given in (28). (28) i obrětъše korablь · vozimъ vъ funikiju and having.found ship being.sailed in Phoenicia ‘and they found a ship being sailed to Phoenicia’ (OCS; Apostolus Christinopolitanus; SJS 1:â•›207)

The context in (28) is clearly one of determinate motion, though the ship was probably not moving when the disciples boarded it; the passive participle vozimъ ‘being sailed’ provides a description of the vessel, as opposed to any actual movement. In Old Russian voziti commonly occurs as the reflexive voziti sja meaning ‘ford, cross a river’ (Modern Russian perepravljat’sja), which is likewise attested in contexts of determinacy, as shown in (29): (29)

I ubojavъšisja, počaša sja voziti na odinu stranu kъ svjatěi and having.taken.fright began ford on one side to holy Sofěi, rkušče: «položim glavy svoja u svjatěi Sofěi». Sophia saying we.will.lay heads own by holy Sophia ‘And frightened, they [the Tatars] began to cross to one side to Holy Sophia, saying: “We will lay down our heads at Holy Sophia”.’ (Old Russian; NPL:â•›310–311)

Common Slavic “indeterminate” verbs of motion

In this example the Tatars begin to cross a river; though there are many of them, they are presented as a single collective entity going in one direction.

2.7

Synthesis

This section has discussed occurrences of the putative “indeterminate” VoM xodi‑ ti ‘walk’, ězditi ‘ride’, běgati ‘flee’, nositi ‘carry’, voditi ‘lead’, and voziti ‘transport’ in determinate contexts. Such usage has been acknowledged by Češko (1951) and Dostál (1954), though neither makes much effort to account for it in their discussions of the nature of indeterminacy in OCS. Other VoM, such as letati ‘fly’, were presumably used in the same way in OCS and Old Russian, but they occur much less frequently than the VoM discussed here, and I have found no examples of their usage in determinate contexts (others, such as gъnati-goniti ‘chase’, have been omitted due to space considerations). As was noted the beginning of this section, there are also more examples of such usage for these verbs that have not been included; Dostál (1954:â•›495), for instance, counts 23 cases of xoditi ‘walk’ in determinate contexts in his OCS corpus. It should also be pointed out that xoditi ‘walk’ occurs in OCS imperatives with determinate meaning; this usage has been discussed by Stern (2002), who also points out the problem posed by such data for assuming indeterminacy to be the category expressed by these VoM in OCS and Old Russian. Thus, I conclude that the use of the so-called indeterminate VoM in contexts of determinacy was a real part of older stages of Slavic, presumably going back to Common Slavic. How one treats this data in an analysis of VoM in Common Slavic and early historical Slavic is another matter entirely; Section 3 discusses the possibility of treating the so-called indeterminate VoM as MoMV.

3.

Indeterminate verbs of motion as manner-of-motion verbs in Common Slavic

If the data in Section 2 are taken at face value, the indeterminacy hypothesis cannot be maintained: the use of OCS and Old Russian VoM of the xoditi type in the context of a single goal-oriented trajectory falsifies the hypothesis that these verbs were originally indeterminate in the sense of modern North Slavic. Thus, the issue becomes that of determining the precise semantic nature of these verbs. The hypothesis advanced here is that at some point the so-called indeterminate VoM as lexical items were not primarily aspectual in nature; rather, they were MoMV. It is difficult to determine whether data such as those adduced

89

90 Stephen M. Dickey

in Section 2 represent relic material or indicate that the the so-called indeterminate VoM were actually MoMV at the time of OCS and early Old Russian. OCS and especially Old Russian texts show that VoM, as well as the Slavic verbal system in general, were undergoing a period of considerable change from whatever system(s) existed in Common Slavic to a post-migration situation in North Slavic, where the determinate/indeterminate opposition existed in VoM alongside an increasingly grammaticalized system of verbal aspect (the latter of which was based on prefixation as a way of creating pf verbs). At the very least, the OCS and Old Russian data discussed in Section 2 represent relic material from an earlier state of affairs. However, the Old Czech data (and similar early West Slavic data) seem to indicate the possibility that the North Slavic “indeterminate” VoM functioned like MoMV well into the historical period. As this issue cannot be resolved here, this paper takes the conservative view that the aforementioned data are relic material from an earlier Common Slavic state of affairs. This view is henceforth termed the MoMV hypothesis. What are the consequences of saying that the Common Slavic indeterminate VoM were actually MoMV? First, it allows the study of these verbs to benefit from some valuable general linguistic literature on MoMV, thereby relieving it of the reliance on an ad hoc and overly structuralist Slavistic taxonomic category that in my view does not stand up to scrutiny. It also allows one to analyze these verbs primarily as motion verbs, taking into account the distinct properties of VoM as discussed, e.g., by Talmy (1985, 1991, 2000), Radden (1988), Levin and Rappaport Hovav (1992, 1995) and Tenny (1995), and Slobin (2004), counter to the practice of previous treatments, such as Češko (1951) and Dostál (1954) or even Stern (2002), who do not recognize VoM to be a specific class of verbs with special properties relevant to aspectual issues. As pointed out, the MoMV hypothesis can make sense of data that has not been adequately considered before. Perhaps most importantly for Slavic linguistics, the MoMV hypothesis can remove some of the danger of circularity when considering questions such as the role (or lack thereof) played by Common Slavic indeterminate VoM and the determinate/indeterminate opposition in the development of Slavic aspect. It must be stressed that the class of Slavic indeterminate verbs did not arise in a unified way. On the contrary, the etymological evidence indicates that the class of Slavic indeterminate verbs were the aggregate of distinct derivational processes spanning the Indo-European, Balto-Slavic and Slavic periods (for details, cf., Greenberg in this volume). It is curious that the transitive VoM nesti-nositi ‘carry’, vesti-voditi ‘lead’, and vezti-voziti ‘transport’ form a compact morphological class, including consistently o-grade roots in the indeterminates; they also appear to be the oldest VoM pairs, either being inherited from Indo-European (vesti-voditi and vezti-voziti) or arising relatively early in the Balto-Slavic stage (nesti-nositi).

Common Slavic “indeterminate” verbs of motion

Assuming that iti was originally a generalized go-verb before it was a determinate correlate of xoditi ‘walk’ (see Section 3.3), it appears that there was a time when the aforementioned transitive VoM existed as pairs, whereas there appear to have been no corresponding pairs of intransitive VoM. Why should this have been the case? I believe the answer is to be found in semantic differences between transitive and intransitive VoM. Intransitive MoMV profile the manner of locomotion of the agent alone, with the result that there are a relatively high number of intransitive MoMV profiling the numerous ways in which a mover can move, e.g., air- or waterborne motion or high-velocity activities such as running; importantly, any intransitive VoM can in one way or another be substituted by go (Common Slavic iti; cf., the data in Section 3.3 below). In contrast, transitive VoM profile particular agent-patient configurations, as opposed to particular manners of agent locomotion per se (with the possible exception of the distinction on foot vs. by vehicle, which is expressed in both intransitive and transitive VoM in North Slavic). The transitive VoM are in no way equivalent to each other (i.e., carrying is very different from leading, which is very different from transporting), nor can they be substituted by go without losing the crucial information about the distinct roles of agent and patient. I consider these facts to be the reason why transitive VoM would be lexically distinct and why there would be pairs of transitive VoM from the outset: the determinate correlate (e.g., vesti ‘lead’) profiled the agent-patient relationship as movement, whereas the so-called indeterminate correlate (e.g., voditi ‘lead/be at the head of ’) profiled the agent-patient relationship as such, with any movement backgrounded. The MoMV hypothesis may appear at first blush to apply more directly to intransitive VoM. However, if the particular agent-patient configurations of the transitive VoM are considered “manners” of motion, or – in Talmy’s terms – coevents that defocus the actual path of motion itself, then they can be treated together with the intransitive VoM by the MoMV hypothesis. Thus, the MoMV hypothesis allows Slavic transitive and intransitive VoM to be meaningfully differentiated while also being treated together profitably where they behave as a class. In my view, this is superior to the indeterminacy hypothesis, which, as far as I am aware, has never even considered the differences between transitive and intransitive VoM. As suggested above, the MoMV hypothesis might shed light on the development of Slavic aspect. In particular, if the aggregate nature of the historical classes of North Slavic indeterminate verbs is considered with regard to conclusions reached by Bermel (1997) and Nørgård-Sørensen (1997), i.e., that the full grammaticalization of Russian aspect occurred relatively very late (post-fifteenth century), then it seems necessary to reconsider what these verbs really were in Common Slavic, both as lexical items and with regard to their possible role in the

91

92

Stephen M. Dickey

development of Slavic verbal aspect. However, solving or even organizing all of the issues involved in the relationship between the Common Slavic xoditi-type VoM and the rise of Slavic aspect far exceeds the scope of this paper. The remainder of this paper attempts to do three things: Section 3.1 suggests answers to some questions that arise if we accept the MoMV hypothesis, namely, why the Common Slavic MoMV develop into indeterminate verbs and lose their ability to express directional motion in modern North Slavic; Section 3.2 briefly sketches a hypothesis of the development of the class of MoMV in Common Slavic; Section 3.3 briefly discusses Common Slavic iti ‘go’ as a generalized go-verb.

3.1

The relationship between MoMV and indeterminate VoM

The MoMV hypothesis allows data of the kind presented in Section 2 to be explained straightforwardly. However, as pointed out in Section 1, the expression of ongoing directed motion is not characteristic of indeterminate verbs in modern North Slavic. This section considers the MoMV hypothesis with regard to the traditionally recognized main interpretations of OCS and Old Russian xoditi-type verbs, i.e., aimless motion and the repetition of a goal-oriented trajectory. The most important characteristic of MoMV is that they are atelic activity verbs (cf., Levin and Rappaport Hovav 1995; Tenny 1995; Smith 1997). The default function of MoMV is to present some manner of motion as a kind of activity. Thus, as Levin and Rappaport Hovav (1995:â•›201) point out, MoMV “in isolation do not describe directed motion”; similarly, Tenny (1995:â•›49) observes that “[m]anner-of-motion verbs, describing pure manner, cannot encode a locational terminus in their most basic sense”. In other words, motion not directed toward a goal, i.e., aimless motion, is the basic meaning of MoMV cross-linguistically. As far as I am aware, Tenny (1995:â•›55–56) is the only scholar to motivate the link between MoMV and aimless motion (for English, which allows free unaccusativization of MoMV). She observes that when English MoMV occur with a path that is long relative to the contextual time frame, the resulting meaning is one of an incomplete single traversal of the path, which she calls the wandering-around reading, e.g., Laura hiked the Long Trail for a week. Tenny points out that in the case of a non-linear surface, the potential non-directionality of the wanderingaround reading is much more prominent, e.g., The women canoed Lake Memph‑ ramagog for several days or Mary walked the hills of Scotland for days. Tenny’s wandering-around reading clearly corresponds to the aimless motion meaning of Slavic indeterminate VoM, i.e., that of continuous motion not occurring along a single (goal-oriented) trajectory. So it is, in fact, easy to link the type of motion

Common Slavic “indeterminate” verbs of motion

prototypically expressed by MoMV with one of the two main interpretations of North Slavic indeterminate VoM. The repetition of a goal-oriented trajectory is a more complex issue. Though some Slavic indeterminate VoM have cognates in other Indo-European languages that qualify as either habitual, iterative, or iterative-intensive verbs, van Wijk (1927:â•›99) points out that Slavic xoditi-type verbs were not specifically habitual (“iterative” in van Wijk’s terms) verbs (cf., his definition of indeterminacy given in Section 1).10 Further, according to Greenberg (in this volume), only three of the Slavic indeterminate VoM can trace their etymologies to Indo-European. While this fact does not preclude the possibility of Slavic inheriting an iterative or habitual category from Indo-European, it can hardly be considered compelling evidence for such a view. And inasmuch as this was possible, it is essential to distinguish between iterative verbs and habitual verbs: regardless of the original nature of Slavic voziti ‘transport’, to link the putative habituality of this verb (i.e., ‘transport now and again’) to the iterativity of Gothic wagjan ‘shake [repeatedly]’ would be woefully misguided. As van Wijk (1927:â•›99) points out, it is very difficult based on comparative Indo-European evidence to determine with certainty the aspectual nature of xoditi-type VoM in Proto-Slavic. Here I only argue that one need not assume that Common Slavic xoditi-type VoM were necessarily special habitual verbs in order to produce the modern class of indeterminate VoM. First, according to Bybee, Perkins, and Pagliuca (1994:â•›171, 232–233) the path of development whereby progressive grammemes (which we can associate with typical uses of atelic verbs to express ongoing actions in Slavic, regardless of the lack of a specific progressive grammeme) generalize to habituals is attested (cf., their discussion of the development of the Armenian past progressive marker -um), whereas they have no evidence of habitual grammemes generalizing to include progressive meanings. Accordingly, the development whereby a specifically habitual xoditi ‘walk’ comes to express habituality, ongoing activities, and single goal-oriented trajectories seems less likely than an atelic xoditi ‘walk’ generalizing to include habitual meanings. In this respect, it is also interesting that Tenny (1995) independently motivates a semantic connection between English atelic MoMV and repetition. It should be pointed out that parts of her analysis are based on the behavior of English MoMV when combined with direct objects (e.g., walk the Appalachian Trail and paddle the canoe), which cannot occur in Slavic due to its morphosyntactic structure. 10. Iterative verbs express actions consisting of multiple instances of the same action on a single occasion, e.g., knock, clap, or in an explicit derivation, Modern Russian poxlopat’ ‘pat a few times’; habitual verbs express actions that repeat on separate occasions, e.g., frequent, or in an explicit derivation Cz chodívat ‘go as a habit, tend to go’.

93

94 Stephen M. Dickey

Furthermore, as pointed out in Section 1, English MoMV undergo unaccusativitization quite easily (i.e., they profile the crossing of the boundary of the landmark), unlike OCS and Old Russian MoMV (see below, however). Nevertheless, Tenny’s observations are in my view quite useful in an analysis of Common Slavic MoMV. Tenny’s analysis, which unfortunately cannot be recounted here in detail, compares English MoMV with two common types of telic verbs, incremental-theme verbs (e.g., eat up) and change-of-state verbs (e.g., wash clean). Following Krifka (1992), Tenny bases her analysis on the fact that incremental-theme verbs and change-ofstate verbs are assessed as having or not having reached their terminus according to the degree of affectedness of the object. Tenny terms the objects of such verbs measuring objects, which measure out in a scalar manner the degree of progress of the predicate toward completion. Tenny then contrasts the characteristics of mea‑ suring out of incremental-theme verbs and change-of-state verbs with the measur‑ ing out of MoMV, which only optionally take a distance phrase, e.g., walk to the bridge. Unlike incremental-theme verbs and change-of-state verbs, which show strong tendencies to measure out, MoMV “are always optionally measuring-out verbs and in fact, are non-measuring in their most basic sense” (Tenny 1995:â•›49), as they are atelic in their default meaning. Thus, in their basic sense MoMV do not combine with a goal PP. However, when they do, Tenny (1995:â•›58) notes an important difference between distance as the measuring scale of MoMV and the consumption, creation, and other changes of state that function as the measuring scales of ordinary telic verbs: “It is a special property of distance as a measuring-scale in the lexical semantics of these verbs, that it may be easily traversed any number of times.” She concludes that repetitive readings are more available to MoMV than to ordinary telic verbs, which tend to express contextually unique changes-of-state. Thus, if one assumes that Common Slavic xoditi-type VoM were actually MoMV similar to those of other languages, there is no difficulty in explaining the prominent meaning of habitual repetition expressed by modern North Slavic indeterminate VoM, as MoMV are already independently linked with repetition. Two points require comment here. First, this account is in fact not an attempt to demonstrate that Slavic xoditi-type VoM were ever really unable to express habitual events and developed this ability secondarily. Rather, the MoMV hypothesis argues that xoditi-type verbs were aspectually vague from the outset, and there is no reason, in fact, to believe that there was ever a time when they could not express repetition. The link outlined by Tenny (1995) is additional evidence that МоМV have a propensity for the meanings that have become so salient in the use of North Slavic indeterminate VoM. Ultimately, however, the reason why habitual repetition and aimless motion have become such salient, if “disjoint”, meanings of the Slavic indeterminate verbs is that since the time of OCS and Old Russian, North Slavic xoditi-type VoM have lost the ability to express single goal-oriented trajectories.

Common Slavic “indeterminate” verbs of motion

This brings us to the most important remaining issue: why North Slavic MoMV would lose their ability to occasionally profile a single goal-oriented trajectory in evidence in the data presented in Section 2. This question cannot be answered fully here, but the answer certainly involves two processes that were underway in Common Slavic and the early historical period. First, the Common Slavic intransitive MoMV xoditi ‘walk’ and ězditi ‘ride’ (and perhaps some others, such as bežati ‘flee’ and letěti ‘fly’) at some point went from being loosely opposed to iti ‘go’ to being each paired with a lexically identical determinate verb that had the function of foregrounding actual movement (on the basis of the already existing transitive pairs nesti-nositi ‘carry’, vesti-voditi ‘lead’, and vezti-voziti ‘transport’), whereupon the functional range of the MoMV presumably underwent some (slight) reduction in their functional scope. Note, however, that the older transitive MoMV could express motion along a single goal-oriented trajectory, so that this cannot have been the crucial development, but rather a necessary precondition. I assume that the real change occurred as prefixation became more and more important in the expression of perfectivity in the nascent aspectual system. In this regard, it is instructive to compare Horrocks’ (2004) and Horrocks and Stavrou’s (2007) account of a similar loss of the ability of ancient Greek MoMV to express goal-directed motion. As Greek has always had a system of viewpoint aspect (the aorist/imperfect opposition remained constant from ancient to modern Greek), what caused the inability of Greek MoMV to express goal-directed motion was a loss of the distinction between motion and location in the case assignment of Greek prepositions (for details, see Horrocks 2004). The relevant change in Slavic did not involve the distinctions encoded by prepositions (the case distinction between motion and location has for all intents and purposes remained constant in North Slavic), but the change from an aspectual system based on an aorist/imperfect opposition to one based on derivational morphology (including prefixation as an important marker of perfectivity with directional predicates). Horrocks and Stavrou’s (2007:â•›625–626) observation that verbs as lexical entities are aspectually fixed as either telic or atelic in all languages with morphologically expressed viewpoint aspect is well taken, but from a Slavic point of view it seems that aspectual systems in which one and the same verb occurs in both the aorist and the imperfect are relatively lexically-unencumbered: the aorist encodes temporal boundedness and the imperfect temporal unboundedness in both telic and atelic verbs. In contrast, a system such as that of Slavic based on prefixation involves much more lexical specification, especially in the case of allative motion. With the advent of perfectivizing prefixation, the Slavic system, in which the aorists of determinate VoM were already a primary means of expressing complete trajectories (albeit with source-orientation by default), specifically allative motion began to require the presence of a prefix in VoM, cf., OCS iti ‘go’ versus ‘pri-iti ‘come/arrive’ (lit.,

95

96 Stephen M. Dickey

‘to-go’) and do-iti ‘reach’ (lit., ‘up to-go’). It seems to me that the salience of such lexically transparent prefixation must be connected to the development of MoMV into indeterminate VoM in North Slavic: as a consequence of the development of such a lexically-based system, unprefixed MoMV, which already had only a weak tendency to combine with goal PPs, began to lose whatever functions of the expression of single goal-oriented trajectories that they had up to that point. Crucial in this regard has been the North Slavic allative prefix pri- ‘to [contact]’, which with very few exceptions expresses contact with the goal landmark (i.e., the final path to goal traversal) including the crossing of the threshold of the goal landmark. This tendency is the source of the semantic nature of Modern Russian prixodit’ (impf)/prijti (pf) ‘come/arrive’, which cannot refer to the approach to the landmark, but only to the actual crossing of the boundary of the landmark, cf., Vot *prixodit/idet Sergej ‘Here comes Sergej’. With this in mind, I suggest that North Slavic VoM developed as follows. Once North Slavic established pri- ‘to [contact]’ as its allative prefix in its come-verbs (e.g., Modern Russian prijti, Czech přijít, both ‘come/arrive’), its unprefixed determinate go-verbs (MRus idti, Czech jít), which always expressed goal-oriented trajectories, began to be increasingly limited to the expression of ongoing directed motion (i.e., without profiling the arrival at the goal). As determinate VoM (e.g., MRus idti, Czech jít ‘go’) shifted to the exclusive expression of ongoing directed motion, the MoMV (each of which were increasingly correlated with individual determinate VoM, e.g., xoditi-iti ‘go on foot’, cf., Section 3.2) lost this function; recall from above that MoMV do not naturally profile directed motion. When MoMV lost the function of expressing single goal-oriented trajectories, they were left with the functions of expressing aimless motion and repeated goal-oriented trajectories, thereby becoming “indeterminate” VoM.11 But this is not nearly the complete story. Like North Slavic, South Slavic languages developed such aspectual prefixation, yet their MoMV are not currently North Slavic-style indeterminate VoM.12 Why should this be? I suggest that it is connected to a major difference between North and South Slavic allative prefixation: as pointed out above, North Slavic generalized pri- ‘to [contact]’ as its allative prefix (e.g., Modern Russian prijti, Czech přijít, both ‘come/arrive’), and the North Slavic come-verbs signal the crossing of the boundary of the goal, and thus with a goal PP are unaccusative predicates (e.g., Modern Russian prijti v 11. The development of the single two-way trip function of East Slavic indeterminate VoM requires further investigation. 12. Ivanova (1962) argues that Bulgarian has lost the determinate : indeterminate opposition as opposed to never having developed it; Bulatova (1963) does the same for BCS. I disagree with their views; see below.

Common Slavic “indeterminate” verbs of motion

dom, Czech přijít v dům ‘come to the house’). South Slavic generalized do- ‘up to’ in this function (e.g., BCS doći, Bulgarian dojda, both ‘come’).13 Such VoM prefixed with do- ‘up to’ differ from VoM prefixed with pri- in that they do not assert contact with the goal, i.e., crossing its boundary, but express only the traversal of a trajectory up to the goal (cf., Filipović 2007:â•›73, who characterizes do- as bound‑ ary-reaching as opposed to boundary-crossing). The goal PP specifies whether the boundary of the goal is crossed or not; the examples in (30) express the crossing of the boundary of the goal, whereas those in (31) express merely reaching it, i.e., going as far as the goal.14 (30) a. b.

Sâžaljavam, če dojdox v kâštata ti, i izplašix decata I regret that I’ve come in house you and frightened children ti. you ‘I regret that I have come to your house and frightened your children.’ (Bulgarian; Internet) Žao mi je, što sam ti došao u kuću i preplašio tvoju sorry me is that aux you come in house and frightened your djecu. children ‘I regret that I have come to your house and frightened your children.’ (BCS; = (30a))

(31) a. b.

Ot xižata prodâlžixme pârvonačalno po kolarskija pât, from hut we continued at.first along carriage path po kojto dojdoxme do xižata … along which we.up.to.went up to hut ‘From the hut we continued at first along the carriage path, along which we reached the hut …’ (Bulgarian; Internet) Došli smo do doma, pozvonili na vrata, otvorila je up.to.went aux up to house rang on door opened aux ljubazna gospođa. polite lady ‘We reached the house, and rang the doorbell; a polite woman opened it.’ (BCS; Internet)

13. This discussion ignores the imperfective partners of these verbs, which share the prefix of the perfective except for Bulgarian imperfective idva ‘come’; whether Bulgarian idva is an older alternative to imperfective doxoždam ‘idem’ or a recent development does not affect the arguments presented here. 14. Bulgarian does not employ dojda as a reach-verb as extensively as BCS, but this is a relatively late development, as is clear from OCS data, in which doiti ‘go up to’ occurs frequently.

97

98 Stephen M. Dickey

(Note that North Slavic languages also have reach-verbs prefixed with do-, e.g., Modern Russian dojti do avtobusnoj ostanovki and Czech dojít k autobusové za‑ stávce ‘go up to/reach the bus stop’.) Thus, come-verbs in BCS and Bulgarian are vague and, as such, merely express motion up to some point. Although both uses of BCS doći and Bulgarian dojda allow in PPs to express how long it took to arrive at/reach the landmark, and as such are unaccusative (telic) verbs, the spatial meaning of do- ‘up to’ imposes an essentially arbitrary limit on the motion and, as such, resembles Horrocks and Stavrou’s (2007) notion of pseudo-unaccusativization of VoM (i.e., the facultative limitation of a trajectory with an arbitrary endpoint). This arbitrary limit expressed by do- minimally modifies the trajectory of a motion event, and likewise gives a minimum of information about the pathto-goal expressed by the verb – the trajectory simply continues ‘up to some point’. This information is simpler than that conveyed by most, if not all other Slavic prefixes (e.g., OCS iz- ‘out of ’, pri- ‘to [contact]’, sъ- ‘off of ’, vъ- ‘into’), which I think is the reason why, as Filipović (2007:â•›73–74) points out, BCS do- ‘up to’ (along with its antonym od- ‘away from’) combines with more VoM than any other goal prefix and that VoM prefixed with do- combine with a greater variety of goal prepositions than VoM with other goal prefixes. I suggest that as do- was generalized as the main allative prefix in BCS and Bulgarian, the lower specificity of this prefix created an effect different than that of pri- in North Slavic: inasmuch as the aorists of determinate VoM expressed complete trajectories in older stages of BCS and Bulgarian, a potential redundancy arose between these verbs and the do- verbs (as both the come- and reach-verbs), as the latter only arbitrarily limit a trajectory, a limitation which is inherent to any finite motion event. This can explain why in BCS and Bulgarian with the exception of the generalized go-verb (BCS ići, Bulgarian ida), all of the (unprefixed) determinate VoM were lost, leaving only the erstwhile MoMV to express all imperfectively construed motion, e.g., BCS nositi and Bulgarian nosja ‘carry’, BCS voditi and Bulgarian vodja ‘lead’, including ongoing directed motion, as shown in (32). (32) a. b.

Kâde nosiš tova dete? where you.carry that child ‘Where are you carrying this child?’ Kamo nosiš to dijete? where you.carry that child ‘Where are you carrying that child?’

(Bulgarian; Internet)

(BCS; = (32a))

According to this hypothesis, BCS and Bulgarian never really developed a determinate/indeterminate opposition, which contradicts the findings of Ivanova (1962) for Bulgarian and Bulatova (1963) for BCS. This issue cannot be discussed in detail here. I will point out only that Ivanova refers to OCS as having had

Common Slavic “indeterminate” verbs of motion

indeterminate VoM, a view that the data presented in Section 2 call into question. As for BCS, Bulatova adduces examples of BCS xoditi-type verbs in the expression of ongoing motion, e.g., the military use of xoditi discussed in SectionÂ€2.1, but fails to see the problems they pose for her view. If BCS never developed the determinate/indeterminate opposition, then its current system of MoMV described briefly in Section 1 may very well offer useful insights into the usage and nature of MoMV in Common Slavic.

3.2

A hypothesis of the development of the determinate/indeterminate correlation

It is now possible to attempt a brief sketch of the development of the pairs of verbs that produced the determinate/indeterminate opposition in North Slavic. However, in order to account for all of the core verbs, some preliminary remarks are necessary regarding the aspectual properties of the verbs that came to comprise the members of the determinate/indeterminate opposition. One way of characterizing MoMV is that they foreground a manner of motion as an activity while backgrounding the movement itself. The MoMV, both transitive and intransitive, may thus be characterized as having been atelic (cf., Tenny 1995; Stern 2002) and basically stative (cf., Vaillant’s 1939 view cited in Section 1). (Here the term stative is simply understood as meaning ‘no change of state’ as opposed to referring specifically to Vendlerian states.) The determinate VoM were more complicated. They may be characterized as foregrounding movement and backgrounding manner of motion (originally, generalized iti ‘go’ completely defocused manner, whereas the transitive VoM necessarily retained the agent-patient configuration as part of the foreground). Directional motion is potentially terminative, yet also frequently construed as ongoing and not reaching its terminus. Following Radden’s (1988) characterization of go-verbs as source-oriented by default, I consider iti ‘go’ as well as the other unprefixed determinate VoM to have been source-oriented, or weakly ablative (it is for this reason that their aorist forms occurred regularly in sequences of events that by necessity included the beginning of the motion in the viewing frame). In aspectual terms, we may say that the determinate VoM profiled the onset of the motion as well as any amount up to the entire trajectory. Including *pleuti-ploviti ‘float’ yields the following system of pairs at the earliest recognizable stage, given in Table 2. Here I suggest that the stative -ě- suffix evident in běžati ‘flee’ (< *běgěti) and letěti ‘fly’ be taken at face value, which means that these verbs were originally stative and occupied the same position as the other stative MoMV. In other words,

99

100 Stephen M. Dickey

Table 2.â•‡ Early system of pairs of onset and manner/stative VoM onset

manner/stative

nesti vesti vezti *pleuti iti

nositi voditi voziti ploviti

Table 3.â•‡ Pairs of onset and manner/stative VoM, stage 2 onset

manner/stative

nesti vesti vezti *pleuti *běgti *lekti iti

nositi voditi voziti ploviti běžati letěti

běžati meant basically ‘be fleeing/on the run’ and letěti meant ‘be in flight/airborne’. This yields “stage 2”, the situation given in Table 3. This view assumes a Common Slavic congener to a Baltic unsuffixed verb evident in Lithuanian lėkti ‘fly’ (regarding this and Common Slavic *běgti, cf., Greenberg in this volume), i.e., the derivation of the stative VoM běžati ‘flee’ and letěti ‘fly’ presupposes the prior existence of unsuffixed correlates. At this point we can anticipate the later shift of běžati ‘flee’ and letěti ‘fly’ from manner verbs to determinate verbs by noting their distinctive status as high-velocity MoMV. At some time (early) in the Slavic period, xoditi ‘walk’ is derived (for details, see Greenberg in this volume). It is likely that subsequently the innovative correlation iti-xoditi became an organizing force in the development of the new opposition onset-of-motion verb/manner-of-motion verb, which begins to give a general binary structure to VoM in Slavic. I consider it possible that it was only after the relatively late derivation of ězditi ‘ride’ as a MoMV (cf., Greenberg and Dickey 2006; Greenberg in this volume) that xoditi ‘walk’ became narrowly correlated with iti ‘go’. At some relatively late point in time (based on contemporary North Slavic variation in the pairs), ězditi ‘ride’ becomes paired with ěti ‘travel’ (future West Slavic) or with ěxati ‘ride’ (future East Slavic), yielding the following picture in Table 4. At a later point in time, -a- suffixation began to draw other verbs into the opposition, e.g., tešti/tekno˛ti-tekati ‘run, flow’. Due to the high linearity of their

Common Slavic “indeterminate” verbs of motion 101

Table 4.â•‡ Pairs of onset and manner/stative VoM, stage 3 onset

manner/stative

iti ěti ~ ěxati nesti vesti vezti *pleuti *běgti *lekti

xoditi ězditi nositi voditi voziti ploviti běžati letěti

Table 5.â•‡ Pairs of onset and manner/stative VoM, stage 4 onset

manner/stative

iti ěti ~ ěxati nesti vesti vezti *pleuti běžati/běgno˛ti* letěti tešti/tekno˛ti

xoditi ězditi nositi voditi voziti ploviti běgati letati tekati

* Běgno˛ti was a secondarily suffixed formation, cf. Gorbachov (2007:â•›51).

trajectories, the high velocity manner/stative verbs became the determinate onset verbs and were replaced in the expression of manner by new manner/stative běgati and letati. I can only assume that, due to its stative meaning, ploviti carved out its own niche as a verb meaning ‘float’ (note that this is a stative notion) and was replaced as a manner verb for ‘swim/navigate’ by plavati. This yields the picture in Table 5. Stage 4 is already well into Common Slavic, and at some point these correlations resemble proto-aspectual correlations enough to be dragged into the nascent aspectual system, but by virtue of the particular semantic qualities of VoM, they never really get fully integrated. The hypothesis of a shift of běžati ‘flee’ and letěti ‘fly’ from MoMV to determinate VoM requires some kind of empirical support. Circumstantial evidence for the shift would be relics of běžati ‘flee’ and letěti ‘fly’ in non-directional contexts or otherwise in functions that were the exclusive domain of xoditi ‘walk’, nositi ‘carry’, etc. Though letěti ‘fly’ is very infrequent in OCS and Old Russian, there are two OCS examples that appear to be just that, cf., the examples in (33):

102 Stephen M. Dickey

(33) a. b. c.

mъnogoočitii xerouvimi i šestokrilataa serafimь okrъstь stojęšta many-eyed cherubims and six-winged seraphim around standing i letěšta and flying ‘many-eyed cherubims and a six-winged seraphim standing around and flying’ (OCS; Euchologium Sinaiticum; Dostál 1954:â•›235) i vьzide na xerouvimъ i letě · letě na kroulou větrъnju and went.up on cherubim and flew flew on wing.dual wind’s ‘and he ascended a cherubim and flew, flew on the wings of the wind’ (OCS; Psalterium Sinaiticum: SJS II:â•›113) xodęi na krilu větrъnju walking on wing.dual wind’s ‘walking on the wings of the wind’ (OCS; Psalterium Sinaiticum: SJS IV:â•›776)

In example (33a) the cherubims and the seraphim are stationary, yet in the air. This situation does not involve any goal-oriented trajectories, let alone a single one. In (33b) the second and third clauses simply tell what God does, i.e., what activity He is engaged in, and the context (Ps. 18) does not indicate a goal-oriented path. Note the parallel with the MoMV xodęi ‘walking’ in (33c). The only difference between the two is that the latter has perhaps a habitual sense. I interpret (33a–b) as relics indicating that letěti ‘fly’ could at some point foreground flying as an activity as opposed to goal-directed motion. There are no clear examples of běžati ‘flee’ with non-directional meaning, which is understandable given that such usage would indeed contradict its lexical meaning. However, there are attestations of its participial forms expressing manner co-events for determinate predicates, as shown in (34). (The translations are Huntley’s.) (34) a. b.

běžęšte izlězošę i[z] syna fleeing they.went.out out tower ‘they came fleeing out of the tower’ (OCS; Suprasliensis; Huntley 1968:â•›45) oudalixъ sę běžę i vъdvorixъ sę vъ poustynjŏ I.distanced self fleeing and I.settled self in desert ‘I departed fleeing and settled in the wilderness.’ (OCS; Suprasliensis; Huntley 1968:â•›51)

These examples contain participial adverbs of běžati used to express manner coevents, just as in the data given for the recognized “indeterminate” VoM (i.e., MoMV) in Section 2. Huntley (1968:â•›51) points out that the Psalterium Sinaiticum employs the “indeterminate” correlate in the very same sentence, given in (35).

Common Slavic “indeterminate” verbs of motion 103

(35) oudalixъ sję běgaję i vъdvorixъ sję vъ poustyni (OCS; Psalterium Sinaiticum; Huntley 1968:â•›51; = (34b))

Given that in OCS and Old Russian such participles expressing manner co-events regularly employ the “indeterminate” VoM/MoMV, it is reasonable to view běžęšte and běžę ‘fleeing’ in (21) as relics of its usage when it was a stative MoMV correlated with *běgti ‘flee [onset]’. This view is supported by another piece of circumstantial evidence: the MRus (substandard) determinate verbal adverb begja/ beža ‘running’ almost never expresses a manner co-event for other VoM, i.e., *On prišel begja/beža ‘He came running’ or *On otpravilsja begja/beža ‘He departed running’. (These verbal adverbs occur adjoined to non-VoM, e.g., oni kričali beg‑ ja/beža ‘they shouted, running’.) In other words, determinate begja/beža does not perform the functions in (35), which is an indication that běžati was a MoMV at the time when it could. To recapitulate, exx. (34)–(36) provide evidence that běžati ‘flee’ and letěti ‘fly’ were not solidly determinate in OCS, as they are attested in functions typical of MoMV. I take this to be an indication that they were originally stative verbs (a view which comports with their stative suffix) correlated with unsuffixed onset verbs (*běgti and *lekti, respectively). The reason why they shifted from the manner/stative correlates to the onset correlates lies in their lexical status as high-velocity predicates, which express very linear motion as a default. As high-velocity strongly implies directional motion, these verbs would be much more likely than other MoMV to become determinate (onset) VoM. This hypothesized special status of běžati and letěti as high-velocity MoMV also makes sense in light of Horrocks’ (2004:â•›186–188) observations on the special behavior of a few ancient Greek MoMV, such as piptō ‘fall’ and pēsō ‘jump’, which entail or strongly imply directional movement. Horrocks points out that these verbs did not require allative prepositions + accusative to express goal-oriented movement, and that they have not lost the ability to express goal-oriented motion in Modern Greek, unlike other MoMV. Though Horrocks does not mention high velocity per se, the verbs he discusses are high velocity, and it should be reiterated that high-velocity tends to imply directional movement (but not vice versa). Though the details of the unusual patterning/behavior of high velocity verbs differ in Greek and Slavic, it nevertheless seems that this category is relevant to an analysis of MoMV, and that it can account for the oddity of the stative -ě- suffix in these determinate verbs straightforwardly in the MoMV hypothesis.

104 Stephen M. Dickey

3.3

Common Slavic iti as a generalized go-verb

Another point important for the MoMV hypothesis has been mentioned previously in passing. Unlike the situation in the contemporary North Slavic languages, it seems that Common Slavic *iti did not in fact mean ‘go on foot’, but was rather a generalized VoM (i.e., ‘go’). Ferrell (1951), Kuznecova (1961), and Staniševa (1962) indicate that this was the case for OCS and Old Russian, cf., (17b) and the following examples: (36) a. b. c.

i idošę vъ pousto město korabljem edini. and they.went in empty place by.ship alone ‘and they went to a deserted place by ship alone.’ (OCS; Zographensis; Zog:â•›57) i vsědъ na konь, vborzě poexa, poemъ s soboju and having.sat on horse quickly rode having.taken with self Klimjanta… Idjasta v sobě vidjasta zarju veliku. Klimjant they.went in self they.saw light great ‘and mounting his horse, he galloped off quickly, having taken Klimjant along… The two of them went and saw a great light.’ (Old Russian; BLDR 1:â•›246) I semu sja podivuemy, kako ptica nebesnyja izъ irьja idut… and this wonder how bird heavenly out paradise go ‘And we wonder at how the heavenly birds come from paradise…’ (Old Russian; BLDR 1:â•›460)

Though Kuznecova and Staniševa only consider iti and xoditi as a pair in this regard, almost all of their examples of generalized motion include iti and not xodi‑ ti.15 Ferrell (1951:â•›297) observes that examples of xoditi as a generalized VoM are “much rarer”. Infrequent occurrences of xoditi as generalized VoM may be interpreted as a consequence of xoditi becoming lexically more closely identical to iti as its indeterminate partner, whereby it would assume some lexical meanings of the latter along the way. Returning to Common Slavic iti as a generalized VoM, it should be pointed out that before it became lexically correlated to xoditi, it would have been involved not only in a binary relationship with xoditi ‘walk’, but with other intransitive MoMV as well, e.g., ězditi ‘ride’, ěxati ‘gallop’, tešti ‘run’, letěti ‘fly’, and plavati ‘swim, roam’. It was only when iti ‘go’ and xoditi ‘walk’, as well as ěti

15. Kuznecova’s (1961) examples with xoditi as a generalized go-verb are limited to xoditi in its military meaning (cf., Section 2.1) and metaphorical meanings of ‘live [a certain way]’, both of which in fact involve its meaning of ‘walk’ as a MoMV.

Common Slavic “indeterminate” verbs of motion 105

‘travel’/ěxati ‘ride, rush’ and ězditi ‘ride’, became members of binary correlations that the basis for the future determinate/indeterminate opposition was in place.

4.

Concluding remarks

This paper has attempted to reinterpret the nature of the putative class of indeterminate verbs of motion in Common Slavic, in light of the general linguistic literature on manner-of-motion verbs. The empirical basis for this reinterpretation consists of cases in which the so-called indeterminate verbs of motion occur in contexts of determinate motion. It has been argued that viewing these putative indeterminate verbs of motion as manner-of-motion verbs allows for a more accurate description of these verbs as verbs of motion. It has been argued that the aspectual properties of manner-of-motion verbs overlap to a considerable extent with those of indeterminate verbs of motion (i.e., aimless motion and repetition), which allows us to make sense of Common Slavic xoditi-type verbs of motion as manner-of-motion verbs. It has further been suggested that the reason why the Common Slavic manner-of-motion verbs lost the ability to express single trajectories of goal-oriented motion in North Slavic lies primarily in the increased role of prefixation as a marker of perfectivity in the Slavic aspectual system and is a consequence of the effects of the establishment of pri- ‘to’ as the main allative prefix in those languages. Finally, a sketch of the development of pairs of verbs of motion that eventually expressed the North Slavic determinate/indeterminate opposition has been offered, beginning with the material known to have been inherited from Indo-European and continuing through late Common Slavic. According to this hypothesis, determinate VoM expressed the onset plus some further amount of the motion, whereas the manner verbs were aspectually stative (atelic). Common Slavic iti was originally a generalized verb of motion ‘go’; it later became correlated with xoditi ‘walk’, which is assumed to be a crucial step in the development of binary correlations of (determinate) verbs of motion expressing the onset of motion and stative manner-of-motion verbs. It is hypothesized that běžati ‘flee’ and letěti ‘fly’ originally belonged to the stative manner-of-motion verbs and eventually shifted to the determinate verbs of motion, due to their status as high-velocity verbs of motion. The manner-of-motion verb hypothesis as presented here leaves an array of issues unaddressed, such as what the manner-of-motion verb hypothesis has to say concerning the issue of the development of aspectually-paired prefixed verbs of motion (e.g., Old Church Slavic prixoditi/priti, ‘come’). Another issue is what the manner-of-motion verb hypothesis means for the assignment of Late Common Slavic to the class of satellite-framed languages, and what the differing behaviors

106 Stephen M. Dickey

of MoMV in various stages of Slavic and in the individual Slavic languages can tell us about the verb-framed language/satellite -framed language typology in general. These issues require further investigation.

References Anstatt, T. 1997. Verben der Bewegung und Zielaktanten im Altrussischen. In Slavistische Lin‑ guistik, T. Berger & J. Raecke (eds.), 9–28. Munich: Verlag Otto Sagner. Bermel, N. 1997. Context and the Lexicon in the Development of Russian Aspect. Berkeley CA: University of California Press. Bulatova, R. 1963. Sud’ba sootnosimyx par glagolov dviženija v serbsko-xorvatskom jazyke (glagoly ići-xoditi, *nesti-nositi, *vesti-voditi, *vesti-voziti). Učenye zapiski Instituta slav‑ janovedenija 37: 223–258. Bybee, J., Perkins, R. & Pagliuca, W. 1994. The Evolution of Grammar. Tense, Aspect and Modal‑ ity in the Languages of the World. Chicago IL: University of Chicago Press. Češko, E. 1951. K istorii slavjanskix glagol’nyx vidov: Osnovy glagolov dviženija v Zografskom kodekse. Učenye zapiski Instituta slavjanovedenija 3: 328–344. Cummins, G. 1983. On the aspect of motion verbs in Czech and Russian. Folia Slavica 6(1–2): 7–52. Dickey, S. 2005. S-/Z- and the grammaticalization of Slavic aspect. Slovene Linguistic Studies 5: 3–55. Dickey, S. To appear. The varying role of po- in the grammaticalization of Slavic aspectual systems: sequences of events, delimitatives, and German language contact. In Collected Stud‑ ies of the 2nd Congress of the Slavic Linguistics Society, SLS 2007, Berlin Meeting, B. Dvořák & E. Grišneva (eds.). Berlin: Zentrum für allgemeine Sprachwissenschaft. Dostál, A. 1954. Studie o vidovém systému v staroslověnštině. Prague: Státní nakladatelství. Ferrell, J. 1951. On some phases of the semantic shift of certain verbs of motion in literary Russian as between the pre-Petrine and post-Petrine periods. American Slavic and East European Review 10(4): 294–301. Filipović, L. 2007. Talking about Motion. A Crosslinguistic Investigation of Lexicalization Pat‑ terns. Amsterdam: John Benjamins. Gorbachov, Y. 2007. Indo-European Origins of the Nasal Inchoative Class in Germanic, Baltic and Slavic. PhD dissertation, Harvard University. Greenberg, M. & Dickey, S. 2006. Slavic *jazditi ‘to ride’ and its implications for the category of (in-)determinacy. In Jezikovna predanost. Akademiku prof. dr. Jožetu Toporišiču ob 80-let‑ nici, M. Jesenšek & Z. Zorko (eds.), 153–158. Maribor: Slavistično društvo Maribor. Horrocks, G. 2004. Aspect and verbs of movement in the history of Greek: Why Pericles could ‘walk into town’ but Karamanlis could not. In Indo-European Perspectives, J. Penney (ed.), 183–194. Oxford: OUP. Horrocks, G. & Stavrou, M. 2007. Grammaticalized aspect and spatio-temporal culimination. Lingua 117: 605–644. Huntley, D. 1968. Old Church Slavonic běžati-běgati. International Journal of Slavic Linguistics and Poetics 11: 45–52.

Common Slavic “indeterminate” verbs of motion 107

Ivanova, K. 1962. Sledi ot edna izčeznala opozicija v glagolnata sistema na sâvremenija bâlgarski ezik. Izvestija na Instituta za bâlgarski ezik 8: 409–417. Johannet, J. 1957. De l’aoriste imperfectif dans la Chronique laurentine. Revue des études slaves 24(3–4): 81–87. Krifka, M. 1992. Thematic relations as links between nominal reference and temporal consitution. In Lexical Matters, I. Sag & A. Szabolsci (eds.), 29–53. Stanford CA: CSLI. Kučera, H. 1980. Markedness in motion. In Morphosyntax in Slavic, C. Chvany & R. Brecht (eds.), 15–42. Columbus OH: Slavica. Kuznecova, A. 1961. O tipax smyslovyx otnošenij v gruppe glagolov dviženija v drevnerusskom jazyke. Vestnik Moskovskogo Universiteta 1: 34–48. Leskien, A. 1962. Handbuch der altbulgarischen Sprache. Heidelberg: Carl Winter Verlag. Levin, B. & Rappaport Hovav, M. 1992. The lexical semantics of verbs of motion: The perspective from unaccusativity. In Thematic Structure: Its Role in Grammar, I. Roca (ed.), 247–269. Berlin: Foris Publications. Levin, B. & Rappaport Hovav, M. 1995. Unaccusativity at the Syntax-Lexical Interface. Cambridge MA: The MIT Press. Němec, I. 1958. Genese slovanského systému vidového. Prague: Nakladatelství Československé akademie věd. Nørgård-Sørensen, J. 1997. Tense, aspect and verbal derivation in the language of the Novgorod birch bark letters. Russian Linguistics 21: 1–21. Radden, G. 1988. The concept of motion. In Understanding the Lexicon: Meaning, Sense and World Knowledge in Lexical Semantics, W. Hüllen & R. Schulze (eds.), 380–394. Tübingen: Max Niemeyer. Schawaller, D. 1990. Fortbewegungsverben im griechischen Neuen Testament und ihre altkir‑ chenslavische Übersetzung. Frankfurt: Peter Lang. Schuyt, R. 1988. The Morphology of Slavic Verbal Aspect: A Descriptive and Historical Study. Amsterdam: Rhodopi. Slobin, D. I. 2004. The many ways to search for a frog: Linguistic typology and the expression of motion events. In Relating Events in Narrative, Vol. 2: Typological and Contextual Perspec‑ tives, S. Strömqvist & L. Verhoven (eds.), 219–257. Mahwah NJ: Lawrence Erlbaum. Smith, C. 1997. The Parameter of Aspect. Dordrecht: Kluwer. Staniševa, D. 1962. Iz istorijata na slavjanskite glagoli za dviženie. Izvestija na Instituta za bâl‑ garski ezik 8: 593–600. Stern, A. 2002. The Verbs of Motion in Old Russian Texts: A Comparative Grammatical Analysis of a Nascent Verb Class. PhD Dissertation, University of California at Los Angeles. Talmy, L. 1985. Lexicalization patterns: Semantic structure in lexical forms. In Language Typol‑ ogy and Syntactic Descriptions, Vol. 3: Grammatical Categories and the Lexicon, Ian Shopen (ed.), 57–149. Cambridge: CUP. Talmy, L. 1991. Path to realization: A typology of event conflation. In Proceedings of the Seven‑ teenth Annual Meeting of the Berkeley Linguistics Society, L. Sutton, C. Johnson & R. Shields (eds.), 480–519. Berkeley CA: Berkeley Linguistics Society. Talmy, L. 2000. Toward a Cognitive Semantics. Vol. II: Typology and Process in Concept Structur‑ ing. Cambridge MA: The MIT Press. Tenny, C. 1995. How motion verbs are special: The interaction of semantic and pragmatic information in aspectual verb meanings. Pragmatics and Cognition 3(1): 31–73. Van Wijk, N. 1927. Die sogenannte Verba iterativa und die Bezeichnung der wiederholten Handlung im Altkirchenslavischen. Indogermanische Forschungen 45: 93–104.

108 Stephen M. Dickey

Van Wijk, N. 1929. Sur l’origine des aspect du verb slave. Revue des études slaves 9(3–4): 237– 252. Vaillant, A. 1939. L’aspect verbal du slav commun; sa morphologisation. Revue des études slaves 19(3–4): 289–314. Veyrenc, J. 1966[1980]. Russe idti et xodit': mouvement de déplacement et mouvement de fonction. In Études sur le verbe russe, 180–200. Paris: Institut d’études slaves.

Sources Assem = Kurz, J. (ed.). 1955. Evangeliarium Assemani. Díl II. Prague: Nakladatelství ČeskosÂ� lovenské Akademie věd. ATE = Bamborschke, U. (ed.) 1968. Das altčechische Tristan-Epos. Wiesbaden: Otto HarÂ� rasowitz. C = Gromača, T. 2004. Crnac. Zagreb: Durieux. BLDR 1 = Lixačev, D. et al. (eds.). 1997. Biblioteka literatury drevnej Rusi. Tom 1. St. PetersÂ� burg: Nauka. LavL = Editorial collective. 1926. Polnoe sobranie russkix letopisej. Tom pervyj: Lavrent’evskaja letopis’. Leningrad: Izdatel’stvo Akademii Nauk SSSR. NPL = Nasanova, N. (ed.). 1950/1969. Novgorodskaja pervaja letopis’ staršego i mladšego iz‑ vodov. The Hague: Mouton. SJS = Kurz, J. (ed.). 1966–1997. Slovník jazyka staroslovenského. Prague: Academia. Supr = Severjanov, S. (ed.). 1904[1956]. Codex Suprasliensis. Graz: Akademische Druck- und Verlagsanstalt. Zog = Jagić, V. (ed.). 1954. Codex glagoliticus, olim Zographensis nunc Petropolitanus. Graz: Akademische Druck- und Verlagsanstalt.

Appendix 1. URLs of Internet sources, listed by example number (2) http://mojaispovjest.blogger.ba/arhiva/2005/07/05 (3) http://www.katolicki.info/forum/viewtopic.php?f=11&t=174 (4a) http://www.index.hr/sport/clanak/mamic-hvala-vam-za-hrvatsko-hrabro-srce/296060. aspx (4b) http://ddadd.blog.hr/2005/05/261979/brfont-colorwhitefontbrbthe-lost-chepbluesbbrfont-colorwhitefontbr.html (5b) Marks, Ljiljana. “Povijesne osobe u hrvatskim usmenim predajama: između povijesti i mita.” http://www.hrvatskiplus.org/index.php?option=com_content&view=article&id= 467%3Amarks-jelacic&catid=36%3Acitanje-proze&Itemid=70&limitstart=1 (30a) http://www.pomak.eu/board/index.php?topic=1607.0 (31a) http://clubs.dir.bg/showthreaded.php?Board=turizym&Number=349571&page= 15&view=collapsed&sb=3&part= (31b) http://www.svjetlost.mojblog.hr/p-neka-druga-djetinjstva/85992.html (32a) Vazov, Ivan. Pod igoto. http://www.slovo.bg/showwork.php3?AuID=14&WorkID=5850 &Level=3

Common Slavic “indeterminate” verbs of motion 109

Appendix 2. Abbreviations used in interlinear glosses aux adj dat dual emph intr pl

auxiliary adjective dative dual emphatic particle interrogative plural

chapter 4

PIE inheritance and word-formational innovation in Slavic motion verbs in â•‚iâ•‚ Marc L. Greenberg University of Kansas

The unprefixed imperfective verbs of motion with present tense in -i (such as Russian vodit’, vozit’, bežat’), most of which are considered indeterminate in the modern languages, developed over a lengthy period from Proto-Indo-European to the disintegration of Proto-Slavic. The final period of their development in Slavic shows striking innovation in the formal and semantic structures, including quasi-serialization in the compounding of verbal stems in such a way that the main lexical verb is modified by a manner component, e.g., *jaâ•‚ ‘travel’ + -sd- ‘sit’ = jazd-i-ti ‘ride’. This innovative period in the development of motion verbs correlates with the period of migrations, which are seen as the end of the previous state of equilibrium in the Slavic speech community.

1.

Introduction

Languages can be classified into two groups on the basis of whether their motion verbs typically represent path or manner in the verb. The Indo-European languages, except for Romance, fall into the second category (Talmy 1985:â•›75). Indo-European languages, again with the exception of Romance, are referred to as satellite-framed (S) languages (as opposed to verb-framed [V] languages), where S-languages typically encode path in prefixes and prepositions and V-languages encode path in the verb itself (Slobin 2004:â•›219ff.). Slavic languages are not only typical representatives of S-languages, but they also demonstrate a much more frequent occurrence of manner verbs in narrative than other non-Romance Indo-European languages (Slobin 2004:â•›225). Based on analysis of adult narrations of the well-known “frog-story,” Slobin finds that some languages use manner verbs with lesser frequency than other languages. While typically V-languages have lower incidence of manner verbs and S-languages higher, such S-languages as Dutch and German use one quarter and English pne third as many manner verbs as Russian, which uses virtually exclusively manner verbs. The discrepancy

112 Marc L. Greenberg

in this analysis is attributed to different lexicalization patterns, whereby the Germanic languages all have a non-manner path verb with the meaning ‘come’, but Russian has only manner verbs with prefixes (e.g., prijti ‘come on foot’, priexat’ ‘come by conveyance’, priletet’ ‘come flying’) (Slobin 2004:â•›227). This point can be extended to other Slavic languages, which have the same general lexicalization patterns mutatis mutandis. Moreover, Slavic developed the manner of motion verbs by creating a special type of verb class that isolated manner from path, as will be discussed forthwith. From a diachronic viewpoint, Slavic represents a special case within the IndoEuropean language family in that Slavic has developed in addition to verbs of lative, directional motion (“determinate”) verbs, also a contrasting class of nonlative, non-directional verbs focusing on manner-of-motion (“indeterminate”), e.g., R idti ‘to go (on foot)’, xodit’ ‘walk, walk around’. While the determinate verb OCS iti ‘go’ has widely attested cognates across the Indo-European family (PIE *h1ey‑ > Hi īt ‘go!’, Sk éti ‘go’, Gr eımi ‘I go’, Lt eō, īre ‘I go, to go’, Old Prussian ēit ‘goes’, Tocharian B yam  ‘goes’, Gheg Albanian ve ‘goes’), the corresponding indeterminate xoditi ‘walk’ (as a verb) has a history only within Slavic (see below). For this reason a diachronic examination of the indeterminate stems is warranted. The present paper has the modest goal of surveying the word-formation processes of the Slavic unprefixed imperfective verbs of motion with present tense in ‑i‑, most of which are now considered to have “indeterminate” aspect (exemplified by R vodit’, vozit’, bežat’, nosit’, xodit’, ezdit’, lazit’, letet’; Sln voditi, voziti, ploveti, bežati, nositi, hoditi, jezditi, leteti) in order to gain insight into their role in the development of the Slavic lexicon. Given that the verbs of motion had their start in the Proto-Indo-European stage, but continued to develop from heterogeneous material and came to form a class with special aspect characteristics within the formation of Slavic, the examination of this layer of the lexicon affords also an opportunity to consider correlations between the verbal innovations and the stage of social development connected with the Slavic speech community at the relevant moments.

. The paper grew out of discussions with my departmental colleague, Stephen M. Dickey, to whom I am grateful for encouraging me to work outside of my usual research topics and for helping me see the larger framework in which my observations on the etymology of Slavic verbs of motion fit. I am also grateful to the volume editors and anonymous referees for several helpful suggestions.

PIE inheritance and word-formational innovation 113

2.

Inherited and innovative material

In the following section the relevant PIE-to-Slavic material for the formation of the Slavic motion verbs with present tense in â•‚iâ•‚, which generally correspond to modern Slavic “indeterminate” stems, is outlined. An attempt is made to distinguish the stems that indicate a direct continuation of PIE formations and those that are formed anew or display formal innovation beginning in either the BaltoSlavic or Slavic-only stages. The only exception to the delimitation to the group of verbs leading to modern “indeterminates” is the inclusion of the verbs *běžati and *letěti, which are treated together with others on formal grounds. Nevertheless, these correspond to a focus on mannerâ•‚ofâ•‚motion semantics (see Dickey, this volume, as well as Jakulis [2004:â•›168–170] with regard to Baltic). The material is culled primarily from three etymological dictionaries, ESSJ, ÈSSJa, and Rix et al. (2001). These three sources were selected because they are among the most up-to-date and detailed etymological dictionaries that relate Slavic material to their putative PIE origins; the ÈSSJ is, at the time of this writing, available through vol. 33 (*orzbotěti – *orzmajati [sę]). Other sources are noted. The emphasis is on giving a general outline of the developments rather than rehearsing the received etymologies and the details supporting them. The relevant etymological dictionaries can be consulted for this purpose. Where my own explanations go beyond the received etymologies, additional comments are added in the discussions of the relevant formation. The discussion is truncated somewhat in that it tacitly refers to the semantic developments for the relevant verbs adduced in Stephen Dickey’s paper in this volume.

2.1

Indo-European inheritance

PIE *wodh-eye‑, PS *vod‑i‑ ‘lead’ OCS voždo˛, voditi ‘lead’, R vožú, vódiš’, vodít’, BCS vòditi, Sln vọ´dim, vodíti; BgÂ€vódja, Cz vodím, vodit; Po wodzić, Li vadaũ, vadýti ‘lead’, La vadît; HiÂ€u-wattemi ‘schaffe herbei’, Av vāδaiieiti ‘leads’. PIE *wog’h‑eye‑, PS *voz-i ‘go by wagon’ OCS vožo˛, voziti ‘lead’, R vožú, vóziš’, vozít’, BCS vòziti, Sln vọ´zim, vozíti; Cz vozím, vozit; OCS vožo˛, voziti ‘travel’, Vedic vāhayati ‘makes travel’, Gr ὀχέομαι ‘I travel’, Go â•‚wagjan ‘to shake’.

The Gothic attestation’s semantics perhaps points to a common original meaning ‘go by means of a cart, i.e., take a bumpy ride’.

114 Marc L. Greenberg

PIE *plow(H)‑eye‑, PS *plov‑ě‑ ‘float’ and *plov‑i‑ ‘cause to float’ Sln plovím, plovẹ´ti ‘navigare, natare, volare, undare’, plovím, plovíti ‘schiffen’; Sk plāváyati ‘cause to float’, Av frāvayōit ‘it could sail off ’, Gr plō´ō ‘I swim, sail’, Lt perplovere ‘to flood, allow water in (a vessel)’.

2.2

Balto-Slavic innovations

(PIE *bhegwâ•‚), PS *běg‑ě‑ ‘flee’, later ‘run’ OCS běžo˛, běžati, R bežát’, Sln bežím, bežáti, R bežát’; Li pabėgė´ti ‘flee’.

The PIE causative (*bhogw-éye-) is reflected in Gr. fobéō ‘I chase off ’ and the medio-passive in Greek (fobéomai ‘I flee’) shows approximately the same meaning found in the Balto-Slavic stative formation. The e-grade stative formation (length in the root being from the operation of Winter’s Law) points to a Balto-Slavic innovation that originally meant ‘flee’, as in Lithuanian, which later became interpreted in Slavic also as a manner verb meaning ‘run’. Li bégu ‘I run’, bégau ‘I ran’, which indicates that the form originated in the PIE imperfect from the theme of the present-tense (contrast liekù ‘I remain’, likaũ ‘I remained’ with zero-grade, which is formed on the aorist stem) (Bubenik 1997:â•›82–83). The unsuffixed form is also attested in Slavic, albeit limited to Polish and East Slavic (Po biec ‘flee’, R dial. beč’ ‘idem’, OUk běčy, běhčy, běhty, Uk bíhty, [ÈSSJa vol. 2, s.v. *běgt’i]). The meaning ‘flee’ is also evident in the durative in â•‚aâ•‚, OCS běgati: otъpuštenomъ že imъ byvъšemъ • tako teko˛šte běgaaxo˛ aky ogn’emъ gonimi ‘when they were released, they fled, running as though they were chased by fire’ (from Codex Suprasliensis, cited in Cejtlin et al. 1994, s.v. běgati). The contrast between the two is still found in Slovene teči ‘run’ vs. bežati ‘run away, flee’. (See also Dickey in this volume.) (PIE *h1nok’â•‚), PS *nos‑i‑ ‘carry’ OCS nošo˛, nositi ‘carry’, BCS nòsiti, Sln nọ´sim, nosíti, Cz nosím, nosit, R nósiš’, nosít’.

ÈSSJa asserts that *nositi is “functionally” an iterative of *nesti, but in terms of word-formation a denominal in â•‚iti from *nosъ, citing the o-grade in Lith. našta, Latv. nasta ‘dress’, Li panašė´ti ‘to become similar’ (ÈSSJa vol. 15:â•›210), which also seems to have developed together with Slavic, cf., Slovene oponašati, BCS oponášati ‘to imitate’ (see Snoj 2003:â•›474). The most elegant solution would be to assume that the “indeterminate” stem was built from nesti in the Balto-Slavic stage by analogy to the existing relationships in vez- : *voz‑, *ved‑ : *vodâ•‚. See also *xodâ•‚, below.

PIE inheritance and word-formational innovation 115

2.3

Slavic innovations

(PIE *sedâ•‚), PS *xodâ•‚ ‘walk’ OCS xoždo˛, xoditi ‘walk’, Sln họ´dim, hodíti, BCS hòditi, Bg xódja, Cz chodit, Po chodzić, R xódiš’, xodit’.

The lack of Balto-Slavic cognates suggests that a direct connection to the alleged IE congeners (which Rix et al. list as “causative-iteratives” Vedic sādáyati ‘to sit, cause to sit’, Younger Av â•‚šāδaiieiti ‘sits’, Middle Welsh anho- ‘chase away’) is only apparent; he provides an alternative explanation of derivation from the nominal root xodъ ‘gait, course’, and explains the unexpected anlaut change x > s as an extension from the prefixed forms (Rix et al. 2001:â•›513–515). To the extent that the Slavic and Indo-Iranian roots can be connected, it is perhaps to a more primitive notion of the IE o-grade of *sed‑ along the lines of ‘positioning one’s legs either to sit or stride’. Similarly Gr hodós ‘path’ developed from this starting point. With regard to Slavic, we may posit a post-nominal formation from xodâ•‚ as ‘gait, march’, particularly in light of OR xoditi ‘dvigat’sja, peredvigat’sja, stupaja nogami’ (ÈSSJa vol. 8:â•›48). The formation then fits with the analogical pattern *ved‑/*vod‑ → *vod‑i‑ti; thus *eidâ•‚ ~ *xed‑/*xod‑ → *xod‑i‑ti. Although the leading form does not come straightforwardly from the present tense stem (*(e)i(d)â•‚), rather, the e-grade correlate to *xodâ•‚ is be found in the participle ‑šed‑ъši and in deverbal forms, cf., Old Novgorodian šestnikъ ‘newcomer’ (Zaliznjak 1995:â•›681). (PIE *yoh2â•‚, *sedâ•‚), PS *ja‑sd‑i‑ ‘travel while sitting, ride’ OCS jaždo˛, jažditi ‘ride’, Sln jẹzdim, jẹ´zditi ‘ride a horse’, BCS jèzditi ‘ride a horse’, Bg jázdja ‘ride a horse’, Ma jazdi ‘rides’, Po jeździć ‘ride’, Pl jezdĕt ‘ride a horse’, US jězdzić ‘ride, control with reins’, R ézdiš’, ézdit’ ‘ride’, Uk jízdyty ‘ride’.

This verb has been subject to a wide variety of etymological treatments, the details of which are sketched in Greenberg and Dickey (2006), which also contains this author’s explanation, as summarized here. The stem is formed as a compound of ja- < PIE *yoHâ•‚ + the zero-grade *â•‚sdâ•‚ ‘sit’, the simplex verb form of which is reflected in PIE congeners, e.g., Li jóti ‘to ride’, ‘travel’, La jât ‘to ride, travel’, Vd yā´ti ‘travel’; the same form is also present in West Slavic languages: Cz jet, US jěć, OPo jać all meaning ‘ride’. Slavic innovated a new verb by creating a term that meant ‘travel’, i.e., both ‘to go a long distance, longer than can be reasonably reached on foot’ and ‘to do so while sitting on a horse’, a meaning that has been preserved to this day in the South Slavic languages; the meaning has been generalized to ‘ride (by any means of conveyance)’ in the modern West and East Slavic languages.

116 Marc L. Greenberg

Although one can construct circular arguments to claim that PS *jazditi is a denominal, cf., PS *borzda˝ (Ru borozdá, Sn brázda) < PIE *bhorH ‘work with a sharp object’ + *â•‚sdâ•‚ = ‘(result of) sharp object sitting (in the earth)’ = ‘furrow’, cognate with En to bore ‘to drill a hole’ (Rix et al. 2001:â•›80, Snoj 2003:â•›s.v.) or that it was formed directly as a verb as PS *borzdı˝ti (R borozdít’) ‘to make a furrow’, the formation is reminiscent of serial-verb constructions in which foregrounded and backgrounded or accompanying verbal actions are juxtaposed in a single clause. Therefore it is reasonable to assume that the verb was formed as such and that it is not necessary to assume a preliminary stage at which a noun *ja˝zdâ•‚ was formed that in turn gave rise to a verb (as, for example, Vaillant 1966:â•›77). (PIE *loh1g’hâ•‚), PS *laz‑i‑ ‘crawl’, later ‘climb’ and ‘go’ OCS laziti ‘repere’, Sln lazim, láziti ‘crawl’, BCS la˝ziti ‘to go, walk’, Bg lázja ‘crawl on all fours’, Cz lazit ‘crawl, climb’, R lázit’ ‘crawl’, Uk lázyty ‘crawl’.

No convincing PIE formations of **loh1g’hâ•‚eyeâ•‚ are given (see ÈSSJa, vol. 14:â•›66), so it is reasonable to assume that this is an example of another Slavic manner verb formed by analogy to the type *vedâ•‚/*vod‑iâ•‚ in which the e-grade served as the leading form *lězti < PIE *leh1g’hâ•‚, cf., La lẽzêt ‘slide’, OHG læ¯ge ‘flat’. Aside from the bleaching of the meaning in BCS to ‘go’, the term is also used in a more specific meaning with regard to beekeeping, BCS dial la˝ziti med ‘to remove honeycomb from a hive’, OR laziti pčelъ, medъ ‘idem’, Br lázic’ ‘to remove the honey from a hive’, OPo połazić ‘obtain honey from a hive’. These meanings reflect the focus on forest apiculture that prevailed among the Slavs, which required tree-climbing to retrieve honey from hives, cf., OR drěvolazъ ‘beekeeper’ (Bezlaj 1948:â•›57). (PIE *lekâ•‚) PS *le‑t‑ě‑ ‘fly’ OCS lešto˛, letěti ‘fly’, Sln letím, letẹ´ti ‘fly, start’, BCS lètīm, lèt(j)eti ‘idem’, Bg letjá ‘fly, flutter’, Cz letěti ‘fly, move fast’, lećeć ‘fly, move fast’, R letét’ ‘fly’.

PS *letěti is derived as a stative verb (cf., běžati) from an unsuffixed form *lekâ•‚ ‘bend (as of a wing)’ (cf., Li lekiù, lė˜kti ‘fly, bolt, fall’, La lèkt, lecu ‘jump, fly [rare]’ and lę˜kāt ‘fly, jump’; the Baltic forms imply an earlier simplex form *lek‑ti that has ceased to exist in Slavic) and the suffix â•‚tâ•‚, which denotes a steady state resulting

. It is beyond the scope of this paper to deal with the problem of the status of such a formation in Slavic as a “true” serial verb or merely a compound. However, in this regard I defer to Crowley (2005:â•›959), who says that “verbal compounding and verb serialization [...], in fact, exhibit a variety of similarities crosslinguistically, with the only real difference being that we are dealing with either words or phrases.” See also Joseph (1990) on this problem with respect to Greek.

PIE inheritance and word-formational innovation 117

from a repeated action, cf., the parallel formation in Bg dъxtjá ‘give off a pleasant smell’, Sln dehtím, dehtẹ´ti ‘idem’, BCS dàht(j)eti ‘idem’, derived from *dъxâ•‚ ‘breath’; R pyxtét’ ‘blow (imperf.)’, krjaxtét’ ‘groan (imperf.)’. The relevant cognates to the root include ON leggr ‘leg, bone’, whence also English leg; Gk laktídzō ‘I kick’.

3.

Discussion and conclusion

The survey of the material above demonstrates that the formation of a set of largely “indeterminate” stems was a lengthy process with origins in Proto-Indo-European and continuing through the Balto-Slavic and Proto-Slavic stages. All of the material ultimately originates in PIE, but some of it was recombined at these later stages evidently to meet evolving communicative needs. The separation of a Balto-Slavic stage in the stratification above seems a bit tenuous, but the criterion for doing so was whether cognates could be found for the whole formation, including both root and suffix. So, for example, one might see the semantic development of the word for ‘fly’ (from a root meaning ‘bend, flap’ developing to ‘flutter’ and, finally, generalized to ‘fly’) as a Balto-Slavic innovation, but the formal addition of â•‚tâ•‚ in Slavic indicates further development beyond the Balto-Slavic stage. Of particular interest are the two Slavic-only formations indicating novel stem formations that include a verbal root + ‘manner’ suffix + verb-class suffix, *jazditi and *letěti. These formations look anomalous in the context of the short list of “indeterminate” verbs, but in fact they reflect innovative processes in Slavic verb formation. For example, PS *ja‑ti ‘travel’ is part of a derivational nest that includes both *ja-x‑a-ti ‘travel fast’ with an intensive suffix â•‚xâ•‚ (R éxat’, Sln jáhati) and *ja-zd-i-ti ‘travel while sitting (on a horse, chariot, etc.)’, cf., PS *ma‑x‑ati (R maxát’) ‘wave one’s hand’ vs. *ma-j‑a-ti, *ma-j-o˛ ‘to shake something’ (OCS ma‑ jati, majo˛, Sln majáti, májem ‘idem’; Li móti, móju, La mãt, mãju, both ‘wave’). Both *jazditi and *letěti can be further contextualized by a new etymology for a later development in BCS. Here, a novel explanation of BCS lèbd(j)eti ‘hover’ is presented. Rather than connect it with PS *lebetati/*lebьtati ‘sway, shake’, said to be of “onomatopoeic” origin (see ÈSSJa vol. 14:â•›78–80), BCS lèbd(j)eti can be seen as a novel formation created as a serialization of leâ•‚ ‘fly’ + â•‚bъdâ•‚ ‘be’, whereby the second element is treated as a zero-grade to related roots byâ•‚, *bo˛dâ•‚ ‘be’ or, possibly even the homophonic bъdâ•‚ (< PIE *bhewdhâ•‚ ‘awake’) (OCS bъděti ‘be awake, stand vigil’ or even a parallel formation to the double-prefixed form *sъnabъděti ‘to supply, care for’, R snabdít’, Cz snabdět, Cr snabdjeti). Here again the second morpheme denotes a manner, in this case the meaning ‘remain in place, stay still’, thus leâ•‚ ‘fly’ + â•‚bdâ•‚ ‘remain in place’ = ‘hover’. This derivational model has remained robustly productive in BCS, e.g., lepŕhati ‘leteti, mahati,

118 Marc L. Greenberg

udarati krilima’, lepŕtati ‘kretati se u vazduhu šumno mašući krilima, leteti’, ‘mahati, udarati krilima (o pticama)’, lepúhati ‘leteti proizvodeći slabiji ili jači šum krilima, lepršati (o pticama)’ (all imperfectives) (Bjeletić 2006:â•›115–117; see the entire monograph for a wider discussion and many more examples). Although Bjeletić analyzes these compounds as “expressive,” there is no reason not to view the first morpheme as lexical, the second morpheme as contributing a manner meaning, and the resulting compound as being tantamount to a serialization. The word-formational heterogeneity of the class of indeterminate verbs is underscored by the fact that iâ•‚suffixed verbs *běžati and *letěti in the Balto-Slavic and Slavic stages, respectively, for determinate verbs of motion with indeterminate stems, formed with the durative â•‚aâ•‚ suffix, *běgajo˛, *běgati and *letajo˛, *letati, respectively. Proceeding from the observation that these new formations in the “indeterminate” verbs of motion innovate precisely by adding “manner” meanings to the lexical material for motion verbs, we agree with Dickey’s view (this volume) that in Proto-Slavic unprefixed imperfectives were manner-of-motion verbs, with the relics of this usage in older attestations of the modern Slavic languages. Though it is difficult, if not impossible, to know in subtle detail what the prehistoric formations meant since no contexts are available for their usage at the moment of their introduction, it does seem certain that at the Slavic stage the manner meaning was salient, given that the new formations explicitly add material to the compounds to emphasize manner. This focus on manner of motion and, in particular, the onset of word-formational creativity in the quasi-verb-serialization compounds may actually have been motivated by extra-linguistic causation, namely, the migration of Slavic speakers from their core territory to roughly their present-day locations. At the outset it should be stated that this is a mere suggestion about the possible connection between word-formational innovation and extra-linguistic factors and should not be seen as deterministic. For example, one might suppose that postColumbian Spanish would develop a high degree of manner salience during the period of colonization and its attendant long-range migration, but this is decidedly not the case. Nevertheless, there may be reasons for this discrepancy: the Latinate world had already developed a maritime vocabulary connected with trade and warfare at least by the early Middle Ages, long before the Columbian period (see McCormick 2001:â•›404ff.). This is not the case with the Slavs, who are widely viewed to have been sedentary agriculturalists before their migrations to the west and south. Moreover, a parallel development of new technology (the “second. Bjeletić views this pattern as being of PIE provenience, citing Indo-Aryan evidence (2006:â•›377–382).

PIE inheritance and word-formational innovation 119

ary products revolution”) and new vocabulary in Indo-European can be observed with the rise of terms pertaining to the wheel, the wagon, and wool-cultivation (Anthony 2007:â•›59ff.) In the first centuries A.D. the Slavic-speakers appear in history at a moment when they have emerged as a separate linguistic entity and are, in a real sense, in motion. Prior to this, as suggested by the archaeological record (the Zarubinec Culture and the correlated Slavic hydronymy as well as the multi-ethnic Černjaxov Culture), the early Slavic speech community is connected with a sedentary agricultural society in the central and upper Dnepr basin (Doluxanov 2000:â•›180–185). The Černjaxov Culture emerges at a time of climate change and socio-political crisis in the Roman Empire which sets the stage for migrations of peripheral ethnic groups to migrate (Doluxanov: loc. cit.) The sparse (and ambiguous) descriptions of a Proto-Slavic ethnos in the earliest historical records, intriguingly enough, refer to manners of motion. For example, Tacitus (56–117 AD) remarks (in Germania 46) that the Venethi are primarily trained as foot soldiers and thus differ from the Sarmatians who operate with horse and cart (“quia et domos figunt et scuta gestant et pedum usu ac pernicitate guadent: quae omnia diversa Sarmatis sunt in plaustro equoque viventibus”) (Gindin et al. 1994:â•›39). Later, the Slavs are observed navigating the waterways in their migration across the Danube and throughout the Balkans by means of monoxyles, according to Priscus (5th c. AD) (Gindin et al. 1994:â•›84–85, 91). The diffusion of Slavic speakers from the 6th century onwards through the Balkans, reaching the Aegean Sea and Constantinople, into central Europe, to the mouth of the Elbe and beyond, implies not just motion, but adaptation to new landscapes, climates, and lifestyles that would have required innovative means of expressing interactions with the world, both their own interactions and those of newly encountered groups. In short, the migrations of the Slavs from their core area in the Dnepr basin constituted a period of punctuation of their social, accordingly, linguistic development, following a period of equilibrium. Such a period of punctuation predicates the comparatively rapid splitting of a stable language community into daughter languages (Dixon 1997:â•›73–85). This is in fact what occurred in the case of Slavic as a consequence of the migratory period.

References Anthony, D. W. 2007. The Horse the Wheel and Language: How Bronze-Age Riders from the Eur‑ asian Steppes Shaped the Modern World. Princeton NJ: Princeton University Press. Bezlaj, F. 1948. Drobci iz pradavnine. Slovenski čebelar L/1–2: 14–21; L/3–4: 57–64.

120 Marc L. Greenberg

Bjeletić, M. 2006. Iskovrnuti glagoli. Tipovi ekspresivnih preverbalnix formanata. Belgrade: Institut za srpski jezik SANU. Bubenik, V. 1997. The verbal system of Old Church Slavic. In Tense and Aspect in Indo-Eu‑ ropean Languages: Theory, Typology, Diachrony, J. Hewson & V. Bubenik (eds.), 82–102. Amsterdam: John Benjamins. Cejtlin, R. M., Večerka, R. & Blagovaja, È. 1994. Staroslavjanskij slovar’ (po rukopisjam X–XI vekov). Moscow: Russkij jazyk. Crowley, T. 2005. Serial verb construction. Encyclopedia of Linguistics 2: 959–960. New York NY: Fitzroy Dearborn. Dixon, R. M. W. 1997. The Rise and Fall of Languages. Cambridge: CUP. Doluxanov, P. M. 2000. Istoki ètnosa. Saint Petersburg: Evropejskij dom. ESSJ = Bezlaj, F. 1976–2007. Etimološki slovar slovenskega jezika, Vols. I–V. Ljubljana: Mladinska knjiga (Vols. I–III) & Založba ZRC (Vols. IV–V). ÈSSJa = Trubačev, O. N. & Žuravlev, A. F. (eds.). 1974– . Ètimologičeskij slovar’ slavjanskix ja‑ zykov, vols. 1– . Moscow: Nauka. Gindin, L. A. et al. 1994. Svod drevnejšix pis’mennyx izvestij o slavjanax, Tom I (I–IV vv.). Moscow: Vostočnaja literatura, RAN. Greenberg, M. L. & Dickey, S. M. 2006. Slavic *jazditi ‘to ride’ and its implications for the category of (in-)determinacy. In Jezikovna predanost. Akademiku prof. dr. Jožetu Toporišiču ob 80-letnici, M. Jesenšek & Z. Zorko (eds.), 153–158. Maribor: Slavistično društvo Maribor. Jakulis, E. 2004. Lietuvių kalbos tekėti, teka tipo veiksmažodžiai. Vilnius: Vilniaus universiteto leidykla. Joseph, B. D. 1990. On arguing for serial verbs (with particular reference to Modern Greek). In When Verbs Collide: Papers from the Ohio State Mini-Conference on Serial Verbs [Ohio State Working Papers in Linguistics 39], B. D. Joseph & A. Zwicky (eds.), 77–90. Columbus OH: Department of Linguistics. McCormick, M. 2001. Origins of the European Economy. Communications and Commerce AD 300–900. Cambridge: CUP. Rix, H. et al. 2001. Lexikon der indogermanischen Verben. Die Wurzeln und ihre Primärstammbil‑ dungen. Wiesbaden: Dr. Ludwig Reichert Verlag. Slobin, D. I. 2004. The many ways to search for a frog: Linguistic typology and the expression of motion events. In Relating Events in Narrative, Vol. 2: Typological and Contextual Perspec‑ tives, S. Strömqvist & L. Verhoeven (eds.), 219–257. Mahwah NJ: Lawrence Erlbaum. Snoj, M. 2003. Slovenski etimološki slovar. Ljubljana: Modrijan. Talmy, L. 1985. Lexicalization patterns: semantic structure in lexical forms. In Language Typol‑ ogy and Syntactic Description, Vol.III: Grammatical Categories and the Lexicon, T. Shopen (ed.), 57–149. Cambridge: CUP. Vaillant, A. 1966. Grammaire comparée des langues slaves, t. III. Paris: Éditions Klincksieck. Zaliznjak, A. A. 1995. Drevnenovgorodskij dialekt. Moscow: Jazyki russkoj kul’tury.

PIE inheritance and word-formational innovation 121

Appendix 1. Language abbreviations Av BCS Bg Br Cz En Go Gr Hi La Li Lt Ma O OCS OHG ON Pl Po PS R Sk Sln Uk

Avestan Bosnian-Croatian-Serbian Bulgarian Belarusian Czech English Gothic Greek Hittite Latvian Lithuanian Latin Macedonean Old Old Church Slavic Old High German Old Norse Polabian Polish Proto-Slavic Russian Sanskrit Slovene Ukrainian

part ii

Synchronic approaches to aspect

chapter 5

Perfectives from indeterminate motion verbs in Russian Laura A. Janda

University of Tromsø

Motion verbs are “exceptions” to the Russian aspect system because: (1) they have two stems (determinate and indeterminate); and (2) prefixation of indeterminate stems yields imperfective verbs (cf., uxodit’i ‘leave’). Furthermore, as an exception to (2), there are some prefixed perfective verbs derived from indeterminate stems, such as poxodit’p ‘walk for a while’ and iznosit’p ‘wear out’. This article addresses these “exceptions to the exceptions” from the perspective of the “cluster” model of Russian aspect (Janda 2007), using a database extracted from the Russian National Corpus. I argue that the motion verbs are prototypical in their aspectual behavior and that a single distinction, that of Completability, can account for the aspectual behavior of both motion and non-motion verbs.

1.

Introduction

The standard pattern of aspectual derivation in Russian involves taking an imperfective base verb and adding a prefix to get a Perfective verb (Timberlake 2004:â•›401–407). The indeterminate stems such as xodit’i ‘walk’, nosit’i ‘carry’, letat’i ‘fly’ defy this generalization when they form imperfective prefixed verbs such as prixodit’i ‘come’, otnosit’i ‘carry off ’ and uletat’i ‘fly away’. However, one cannot argue for a generalization that indeterminate stems are simply an exception to the rule because sometimes the addition of a prefix does yield a perfective verb, such as sxodit’p ‘go somewhere and come back’, iznosit’p ‘wear out’ and poletat’p ‘fly for a while’. To make matters worse, sometimes adding the same prefix to an indeterminate stem can create both a perfective and an imperfective verb, as in the case of sxodit’p ‘go somewhere and come back’ vs. sxodit’i ‘descend’ and výxodit’p ‘nurse,

. Note that perfective verbs are tagged with a superscript “p” and imperfective verbs are tagged with a superscript “i” throughout this article.

126 Laura A. Janda

rear’ vs. vyxodít’i ‘exit’. Finally, there is the problem of why some of the perfectives formed from indeterminate stems can have secondary imperfectives, as in the case of iznosit’p-iznašivat’i ‘wear out’ and vyxodit’p-vyxaživat’i ‘nurse, rear’, but others cannot, such as sxodit’p ‘go somewhere and come back’ and poletat’p ‘fly for a while’. The aim of this article is to explain all of these phenomena without using any devices beyond those already necessary to explain the Russian aspect system in general. This explanation shows that the motion verbs are a coherent subsystem of the overall aspect system, and that not only are the motion verbs not exceptions, rather they make the distinctions that motivate the entire system most salient and are thus prototypical. The explanation is illustrated with data from the Russian National Corpus (henceforth RNC; www.ruscorpora.ru). The argument presented in this article can be contrasted with existing arguments in the literature. Most handbooks of Russian grammar merely list the behavior of the motion verbs as “exceptional” without an explanation of why that is the case (Townsend 1975:â•›134; Švedova et al. 1982:â•›590–591; Wade 1992:â•›354– 356). Isačenko (1960:â•›337–344) and Zaliznjak & Šmelev (2000:â•›87–95) offer a different proposal, which suggests that perfectives from indeterminate stems such as poletat’p ‘fly for a while’ are the norm, and imperfectives such as prixodit’i ‘come’ are not formed by prefixation at all, but are rather the result of secondary Imperfectivization. This argument is flawed for a number of reasons (cf., Nesset 2008), among them: (a) it assumes that the motion verbs form a closed subsystem, removing the need to integrate them into the aspect system as a whole; (b) it gives priority to the prefixed perfectives of indeterminate stems, despite the fact that prefixed imperfectives of this type are vastly more common; (c) it assumes that there is an exceptionless rule in Russian that prefixation always yields a perfective verb, despite the existence of counterexamples such as vygljadet’i ‘look like’ (cf., Browne 1978); (d) it necessitates the creation of a whole new set of classes of secondary imperfectives, since otherwise all secondary imperfectives in Russian contain the -aj suffix; (e) it ignores the formal properties of Russian morphology, denying that prixodit’i ‘come’ is formed from pri- and xodit’. This article will confront the same set of issues, offering an alternative that is more coherent with the overall system of Russian aspect. The “cluster” model used here is compared to the traditional “pair” model in Section 2, with special attention to motion verbs. A key concept in the cluster model is Completability, which motivates the aspectual behavior of all verbs, including motion verbs. The essence of the argument is that indeterminate motion verbs express Non-completable situations, which facilitates their use to build . Note the difference in stress in these two verbs. Following the conventions of Russian, all perfective verbs prefixed in vy- receive stress on that prefix (cf., Browne 1978).

Perfectives from indeterminate motion verbs in Russian 127

certain types of prefixed perfectives. In order to provide a basis for the analysis, an empirical study was conducted, and the data are presented in Section 3. The remaining four sections are devoted to the four types of products obtained when a prefix is added to an indeterminate motion verb: an imperfective verb (SectionÂ€4), a Complex Act Perfective (Section 5), a Single Act Perfective (Section 6), and a Specialized Perfective (Section 7). The final section (8) summarizes the findings of this study.

2.

Previous relevant research and the cluster model

Traditionally it is asserted that Russian verbs exist as aspectual “pairs” (cf., Vinogradov 1938; Šaxmatov 1941; Bondarko 1983; Čertkova 1996; Zaliznjak & Šmelev 2000). The “pair” model states that Russian verbs (with the exception of the defective perfective and imperfective isolates and biaspectual/anaspectual verbs) exist as “pairs” consisting of a perfective and an imperfective partner. For example, it is claimed that Russian has a pair of verbs that express ‘write’: a perfective napisat’p and an imperfective pisat’ i. It is certainly the case that such aspectual partnerships exist, but these partnerships are usually embedded in larger clusters. For example, napisat’p ‘write’ and pisat’ i ‘write’ have aspectual relationships with a number of other verbs, among them: popisat’p ‘write for a while’, perepisat’p ‘rewrite’, perepisyvat’ i ‘rewritei ’, and poperepisyvat’p ‘rewrite for a while’. The cluster model of Russian aspect (Janda 2007) recognizes an aspectual cluster as a group of verbs that are aspectually related to a single lexical item. The cluster model extends the traditional pair model by recognizing that: (a) most verbs are aspectually related to more than just one “partner”, (b) there are four types of perfective verbs in Russian (Natural Perfective, Specialized Perfective, Complex Act Perfective, and Single Act Perfective), and (c) Completability largely determines what kinds of perfectives are related to a given imperfective. The aspectual cluster of a motion verb, letet’- letat’i ‘fly’, can be used to illustrate the cluster model, with corresponding examples from non-motion verbs krepnut’i ‘get stronger’, pisat’i ‘write’, vjazat’i ‘tie’, rabotat’i ‘work’, dut’i ‘blow’, and stonat’i ‘moan’ to show parallels. The maximal cluster structure includes all four types of perfectives. Natural Perfective: poletet’p ‘fly’, okrepnut’p ‘get stronger’, napisat’p ‘write’, svjazat’p ‘tie’. A Natural Perfective is denotationally equivalent to the corresponding imperfective verb, differing only in aspect, which has led to the claim that

128 Laura A. Janda

prefixes in such forms are semantically “empty”. Secondary imperfectives are not typically formed from Natural Perfectives. It is possible (though not usual) for a verb to have more than one Natural Perfective and there can be various morphological relationships between a Natural Perfective and the corresponding imperfective verb. Not all imperfectives have a corresponding Natural Perfective; rabotat’i ‘work’, dut’i ‘blow’, and stonat’i ‘moan’ all lack a Natural Perfective, though they do form other kinds of perfectives, as noted below. Specialized Perfective: priletet’p ‘arrive flying’, perepisat’p ‘rewrite’, razvjazat’p ‘untie’, pererabotat’p ‘revise’, razdut’p ‘inflate’. It is not uncommon for a verb to form many Specialized Perfectives; a typical motion verb will have over a dozen Specialized Perfectives with various prefixes. As described below in Section 7, metonymy facilitates semantic shifts that motivate Specialized Perfectives, and this is the mechanism behind Specialized Perfectives that are formed in clusters lacking a Natural Perfective, as well as Specialized Perfectives from indeterminate stems, such as iznosit’p ‘wear out’. Specialized Perfectives usually have corresponding imperfectives, such as priletat’i ‘arrive flying’, perepisyvat’i ‘rewrite’, razvjazyvat’i ‘untie’, pererabatyvat’i ‘revise’, razduvat’i ‘inflate’, iznašivat’i ‘wear out’. Complex Act Perfective: poletat’p ‘fly for a while’, popisat’p ‘write for a while’, porabotat’p ‘work for a while’, podut’p ‘blow for a while’, postonat’p ‘moan for a while’. A Complex Act Perfective describes a situation in which temporal limits are set on an activity that is engaged in without a result or change of state. The majority of Complex Act Perfectives are formed with the po- prefix, though pro- ‘for a given period of time’, za- ‘begin’, ot- ‘stop’ are also commonly used to form this type of perfective. A verb may form several Complex Act Perfectives, but secondary imperfectives are typically not formed. Single Act Perfective: sletat’p ‘fly someplace and come back’, dunut’p ‘blow once’. A Single Act Perfective is a semelfactive verb that selects a single cycle in a series of repeated actions that are conceptualized as identical. The indeterminate verb letat’i ‘fly’ can refer to a series of roundtrips, from which it is possible to . The “empty” prefix is a traditional term that I reject for reasons that go beyond the scope of this article, cf., Vinogradov (1938); Isačenko (1960); Tixonov (1962); Forsyth (1970). Note that the Natural Perfectives of motion verbs are all formed with po- and have an ingressive nuance. Zaliznjak & Šmelev (2000:â•›111) list idtii-pojtip ‘go, walk’ and bežat’i-pobežat’p ‘run’ alongside a number of other “pairs” sharing this property, such as čuvstvovat’i-počuvstvovat’p ‘feel’ serdit’sja’i-rasserdit’sjap ‘become angry’ and videt’i-uvidet’p ‘see’. For more on the history of poand its relationship to motion verbs, see Dickey (2007). . E.g., the three Natural Perfectives vyčistit’p, očistit’p and počistit’p of čistit’i ‘cleanse’; biaspectual verbs like blokirovat’p/i ‘block’ where the same form can be both Perfective and Imperfective; suppletive verbs such as kupit’p-pokupat’i ‘buy’; and verbs like dat’p ‘give’ where the Natural Perfective is morphologically the “base” form from which the Imperfective davat’i is derived.

Perfectives from indeterminate motion verbs in Russian 129

extract a single roundtrip; likewise dut’i ‘blow’ can refer to a series of puffs, from which it is possible to extract a single puff. Single Act Perfectives are typically formed using the s- prefix for motion verbs and the -nu suffix for other verbs. It is usually the case that only one Single Act Perfective can be formed and that there are no secondary imperfectives. Two things about the distribution of the various types of perfectives should be noted. One is that a motion verb can form all four types of perfectives. This fact can be used to argue for the prototypicality of motion verbs. The other is that whereas Natural Perfectives such as poletet’p ‘fly’ and Specialized Perfectives such as priletet’p ‘arrive flying’ are formed from the determinate stem of a motion verb, Complex Act Perfectives such as poletat’p ‘fly for a while’ and Single Act Perfectives such as sletat’p ‘fly someplace and come back’ are formed from the indeterminate stem of a motion verb. This division of labor among the stems of a motion verb corresponds to construals of Completability vs. Non-completability among non-motion verbs, as argued below. Verbs that have a Completable construal can form a Natural Perfective. Verbs that have a Non-completable construal can form a Complex Act Perfective, and might additionally have a Special Act Perfective. This distribution and the relationship between Completability and concepts such as boundedness and telicity are described in more detail below. Prototypicality is invoked in linguistics to assert that linguistic categories have a radial structure centered about a prototype (see particularly Lakoff 1987; Geeraerts 1995; Croft & Cruse 2004; and Lewandowska-Tomaszczyk 2007). Membership in a category is determined by overlapping clusters of properties, which serve as a measure of family resemblance. The most prototypical member of a category has the densest set of shared properties and relationships to other members of the category (Geeraerts 1995:â•›25; Croft & Cruse 2004:â•›78, 81; Lewandowska-Tomaszczyk 2007:â•›155). The prototypical member of a category is typically associated with embodied experiences such as motor interactions and movements and represents an “Idealized Cognitive Model” of the category (Lakoff 1987:â•›56, 68–76). These definitions support the suggestion that motion verbs are prototypical, since they have the densest set of properties (with all four types of perfectives) and reference embodied movement. Furthermore, as argued below, Completability is a key concept in the aspectual system of Russian and is motivated by the determinate vs. indeterminate distinction of motion verbs. A Completable situation is one that makes progress and will usually reach a natural conclusion if it is continued, whereas a Non-completable situation is . There do appear to be some Single Act Perfectives formed with the s- prefix involving nonmotion verbs as well, as we see in sglupit’p ‘do a single stupid thing’, but this is the topic of future research.

130 Laura A. Janda

one that can be engaged in without necessarily making any progress. Most Russian verbs are ambiguous in their expression of Completability, as we see in igrat’i ‘play’, which can be Completable when one is playing a piece of music, but can be Non-completable if one is just amusing oneself. Thus the Completable expression igrat’i sonatu ‘play a sonata’ has the Natural Perfective sygrat’p sonatu ‘play a sonata’, and the non-Completable expression igrat’i v kukly ‘play with dolls’ can have a Complex Act Perfective which merely sets boundaries for non-goal-directed activity, such as poigrat’p v kukly ‘play with dolls for a while’. A few verbs in Russian express unambiguously Completable situations, such as krepnut’i ‘become stronger’, since it is not possible to engage in becoming stronger without making progress along the scale of strength. Because such verbs are Completable, they have Natural Perfectives like okrepnut’p ‘become stronger’, but because they cannot express a Non-completable situation, they do not have any associated Complex Act Perfective verbs. Some verbs, such as stonat’i ‘moan’ and rabotat’i ‘work’, unambiguously express Non-completable situations, since no amount of moaning or working will necessarily lead to a conclusion. These Non-completable verbs lack Natural Perfective partners, but can readily form Complex Act Perfectives, such as postonat’p ‘moan for a while’ and porabotat’p ‘work for a while’. Specialized Perfectives involve a Completable construal that can be achieved via metonymy from stems that are otherwise Completable or Non-Completable (see Section 7 for more on metonymy). Some Non-completable situations can be conceived of as a collection of repeated events, any one of which can be extracted and viewed on its own. Dut’i ‘blow’ can be conceived of as a series of puffs, motivating the formation of the Single Act Perfective dunut’p ‘blow once’, but rabotat’i ‘work’ is non-homogeneous, so this verb lacks a Single Act Perfective. In the cluster model Completability emerges as the most pervasive distinction among verbs and plays a large role in determining what sorts of perfectives can be formed. Completability is no less important for motion verbs; in fact, this concept is even more salient for this group of verbs. Completability vs. non-Completability is actually the same as the determinate vs. indeterminate distinction. What is special about motion verbs is that they have taken this semantic distinction to the level of a lexical/morphological formal distinction, specializing one set of verbs (determinate stems) for Completable functions and another set (indeterminate stems) for Non-completable functions. Let us take idtii vs. xodit’i ‘walk’ as an example. With the determinate stem idtii ‘walk’, any engagement in the activity leads to progress in the given direction, making this expression fully parallel to the unambiguous Completability expressed by krepnut’i. Indeed I would argue that idtii and the other determinate motion stems serve as the metaphorical source domain for the understanding of all Completable events (cf., Lakoff 1987; Lakoff

Perfectives from indeterminate motion verbs in Russian 131

& Johnson 1980, 1999 for mechanics of metaphorical mapping). The physical goal of idtii ‘walk’ maps onto the result of a Completable event, and physical progress along the path to that goal maps onto metaphorical progress toward attaining a result. Thus igrat’i sonatu ‘play a sonata’ is a metaphorical journey from one note to the next leading to a result, understood in the same way as idtii v magazin ‘walk to the store’ is an actual journey from one step to the next. By contrast, an indeterminate stem like xodit’i ‘walk’ uses non-directed motion as the source domain for Non-completability, giving us a parallel between actions like xodit’i po parku ‘walk in the park’ and igrat’i v kukly ‘play with dolls’. Thus all non-Completable activity is metaphorically understood as non-directed motion. Among the possible interpretations of the indeterminate stems is that of repeated roundtrips, as in deti xodjati v školu ‘the children attend school (i.e., go there and back repeatedly)’. A indeterminate stem thus can express a collection of repeated cycles, each of which brings the individual back to their original location. It is possible extract a single cycle from the mass of repetitions represented by xodit’i ‘walk’ to get just one roundtrip, motivating the existence of sxodit’p ‘walk someplace and come back once’, a Single Act Perfective parallel to formations such as dunut’p ‘blow once’. Table 1 summarizes the relationships between determinate vs. indeterminate and Completable vs. Non-completable, showing the parallels between motion and non-motion verbs. The difference between motion and non-motion verbs is that non-motion verbs can be ambiguous for Completability, whereas motion verbs have two stems that unambiguously mark this distinction. There are, however, examples where this line is crossed, and an indeterminate motion verb can acquire a Completable

Table 1.â•‡ Determinacy, Completability, and the formation of perfectives Motion verbs

Non-motion verbs

Natural Perfectives

Determinate stems poletet’p ‘fly’

Completable construals napisat’p ‘write’

Specialized Perfectives

Determinate stems priletet’p ‘arrive flying’

Completable construals perepisat’p ‘rewrite’

Complex Act Perfectives

Indeterminate stems poletat’p ‘fly for a while’

Non-completable construals popisat’p ‘write for a while’, podut’p ‘blow for a while’

Single Act Perfectives

Indeterminate stems sletat’p ‘fly someplace and come back’

Non-completable construals dunut’p ‘blow once’

132 Laura A. Janda

construal and form a Specialized Perfective, but this is only possible when metonymy is present and the verb has for all practical purposes ceased to be a motion verb. This type of formation is described in Section 7. Completability, unlike telicity and boundedness, refers to both imperfective and perfective situations. In reference to the former, Completability can be used to show the relationship between imperfectives and the perfectives that can be derived from them. An imperfective verb that has a Completable construal can potentially form two kinds of telic perfectives, Natural and Specialized Perfectives, both of which are also Completable. An imperfective verb that has a non-Completable construal can potentially form two kinds of atelic perfectives, Complex Act Perfectives (which are also bounded), and Single Act Perfectives (which are punctual). If metonymy can lend a Completable construal to a verb that is otherwise Non-completable, it may be possible to form Specialized Perfectives.

3.

Data on prefixed indeterminate verbs

Table 2 presents data collected from the RNC. This study examines non-reflexive forms of the indeterminate stems in the presence of three prefixes: za-, po-, and s-. The initial round of data collection extracted all forms of indeterminate stems with these prefixes from the disambiguated subcorpus of the RNC (totaling six million words), thus culling a total of 1767 examples for fourteen stems. The data for eight of the stems, namely brodit’i ‘roam’, gonjat’i ‘chase’, katat’i ‘roll’, lazit’i ‘climb’, plavat’i ‘swim’, polzat’i ‘crawl’, taskat’i ‘drag’, vozit’i ‘convey’, are fairly meager and are not included in Table 2. The six verbs that are in Table 2 provide 1588 examples, thus presenting 90% of the original data. The column under each of the prefixes gives the total raw frequencies of prefixed forms of each verb found in the RNC subcorpus. All of these examples were analyzed individually to determine whether they represented a Complex Act Perfective, a Single Act Perfective, or something else (a Prefixed imperfective or a Specialized Perfective). Columns marked # CA and # SA list the raw frequencies of examples that can unambiguously be identified as Complex Act and Single Act Perfectives, respectively. Columns marked % CA and % SA report these data as relative frequencies rounded to the nearest integer. Thus the first row of the table can be interpreted as follows: 80 examples of zabe‑ gat’ were found, of which 30 were Complex Act Perfectives, constituting 38% of the total zabegat’ examples; 12 examples of pobegat’ were found and all 12 were Complex Act Perfectives, so this gives 100%; 86 sbegat’ examples were found, and

. Thanks are due to John Korba for collecting this data in November 2007.

Perfectives from indeterminate motion verbs in Russian 133

Table 2.â•‡ Examples of prefixed indeterminate stems extracted from RNC za-

# CA

% CA

po-

# CA

% CA

s-

# SA

% SA

begat’ ‘run’ vodit’ ‘lead’ ezdit’ ‘ride’ letat’ ‘fly’ nosit’ ‘carry’ xodit’ ‘walk’

â•⁄ 80 â•⁄ 97 â•⁄â•⁄ 4 â•⁄ 12 â•⁄ 38 323

30 â•⁄ 0 â•⁄ 0 â•⁄ 2 â•⁄ 0 â•⁄ 6

38% â•⁄ 0% â•⁄ 0% 17% â•⁄ 0% â•⁄ 2%

â•⁄ 12 â•⁄ 18 â•⁄â•⁄ 9 â•⁄â•⁄ 5 â•⁄ 43 200

12 â•⁄ 2 â•⁄ 9 â•⁄ 5 â•⁄ 4 64

100% â•⁄ 11% 100% 100% â•⁄â•⁄ 9% â•⁄ 32%

â•⁄ 86 169 103 â•⁄ 28 â•⁄ 38 323

â•⁄ 58 â•⁄â•⁄ 8 103 â•⁄ 18 â•⁄â•⁄ 0 193

â•⁄ 67% â•⁄â•⁄ 5% 100% â•⁄ 64% â•⁄â•⁄ 0% â•⁄ 60%

Totals

554

38

â•⁄ 7%

287

96

â•⁄ 33%

747

380

â•⁄ 50%

58 of them, or 67%, were Single Act Perfectives. Examples from this database are used as illustrations in the following four sections. There are two sources of possible distortion in Table 2. One results from a stress difference between perfective prefixed forms of begat’i ‘run’, which are stressed on the stem, and imperfective prefixed forms of begat’i ‘run’, which are stressed on the suffix. In other words, whereas there is a morphological difference between these forms, it is not reflected in the orthography, and both forms were collected and subsequently submitted to manual analysis. A similar problem with the opposite results obtains for ezdit’i ‘ride’, which has a distinct form for prefixed imperfectives, namely – ezžat’. Thus Table 2 reports only data for prefixed perfectives for this verb. Overall, Table 2 tells us that prefixed perfectives are a robust phenomenon among indeterminate verb stems, though their distribution varies greatly. The lowest figures are for za- which is best attested for zabegat’p ‘begin running’. Pocreates Complex Act Perfectives from all six verbs, at varying relative frequencies. The same may be true of s- and the Single Act Perfectives, but there were no attestations in this database for nosit’i ‘carry’, whereas sbegat’p ‘run someplace and back once’, sletat’p ‘fly someplace and back once’, and sxodit’p ‘walk someplace and back once’ were particularly well attested. There are, however, idiosyncratic effects on the data in almost every cell of this table (idiomatic expressions, various types of competition among homonymous verbs), which obviate any statistical analysis. These effects are discussed in more detail in the following sections, each of which is devoted to describing one type of prefixed derivation from the indeterminate motion verbs.

134 Laura A. Janda

4.

Prefixed imperfectives from indeterminate stems

The use of the indeterminate stem to form the prefixed imperfective partners of perfectives formed from the determinate stem is a trivial observation, but it is worth noting that this pattern is fully compatible with the cluster model. It certainly makes sense that the determinate stem is used to form perfectives since it is the stem that unambiguously expresses Completability, the concept that motivates both Natural and Specialized Perfectives (which are the prototypical perfectives in the Russian aspect system). Non-Completability is a prototypical concept for imperfective verbs, so it also makes sense that the indeterminate stems are used to form imperfectives. Prefixed imperfectives are present in the database in most columns headed by za- and s- in Table 2. Thus, for example, 17% of attestations of zaletat’ and 64% of attestations of sletat’ were identified as Complex Act or Single Act Perfectives respectively; the remaining attestations were of imperfective verbs meaning ‘stop by flying’ and ‘fly down’. Ezdit’i ‘ride’ provided the following exceptions: zaezdit’p ‘torture by riding’, poezdit’p ‘ride for a while’ and s”ezdit’p ‘ride someplace and come back once’ are perfective only by definition. Furthermore, while there are no attestations of imperfective uses of verbs homonymous with pobegat’p ‘run for a while’, or poletat’p ‘fly for a while’, there are homonymous verbs such as povodit’i ‘move’, ponosit’i ‘curse’, and poxodit’i ‘resemble’. However, in most cases, the meanings of the imperfectives formed with za- and s- are largely predictable and these verbs usually serve as aspectual partners to the corresponding prefixed Specialized Perfectives, with za-prefixed forms used to indicate deviation from a path (cf., Janda 1985, 1986) and s-prefixed forms used to indicate descent or departure. Examples (1) and (2) illustrate these predictable types: (1) Vpročem, ne budem za-begat’i vpered, dal’še budet ešče xužě. however not will by-run.inf ahead further will.be even worse ‘However, we aren’t going to be running ahead, further on it will be even worse.’ (Mixail Bulgakov. Teatral’nyj roman, 1936–1937) (2)

On s-xodit s tribuny i zanimaet svoe mesto v he down-walks.ipfv from tribunal and takes own place in prezidiume. presidium ‘He descends from the tribunal and takes his place in the presidium.’ (Vasilij Aksenov. Zvezdnyj bilet, 1961)

Relevant factors can be summarized as follows:

Perfectives from indeterminate motion verbs in Russian 135

– The verb zavodit’i is exclusively used as an imperfective, but appears rarely in the meaning of ‘lead off to a place’; nearly all other uses of this verb are metaphorical and involve the idiomatic use of zavodit’i as a synonym for načinat’i ‘begin’. – All imperfectives formed with po- are idiomatic. – All 16 attestations of povodit’i as a non-Complex Act verb are imperfective, but all of them involve the use of the verb in an idiomatic construction to describe movements of body parts (golovoj, plečami, brovjami, etc.). – All 39 attestations of ponosit’ as a non-Complex Act verb are also imperfective, but all of them are of a single idiomatic type, in the meaning ‘curse’. – All 136 attestations of poxodit’ as a non-Complex Act verb are likewise imperfective, and again of a single idiomatic type, in the meaning ‘resemble’. – All 38 attestations of snosit’i are imperfective, and they are nearly evenly distributed among three meanings, the latter two of which are idiomatic: ‘carry down’, ‘tear down’, and ‘endure’. – All 130 attestations of sxodit’ as a non-Single Act verb are imperfective, and the majority of these come from the idiom sxodit’i s uma ‘go crazy’. – Lack of attestation in this database does not entail that a form does not exist in Russian, though it does imply that such forms are rare.

5.

Prefixed Complex Act Perfectives from indeterminate stems

The Non-completable meaning of the indeterminate stem can motivate the formation of prefixed Complex Act Perfectives. The database presents two types of these verbs, namely verbs prefixed with za- meaning ‘begin to X’ and verbs prefixed with po- meaning ‘do X for a while’. The situation concerning the latter group, the po-prefixed delimitatives, is fairly straightforward. All six verbs form Complex Act Perfectives with po-, which account for 100% of the prefixed perfectives for those verbs. Only three of the verbs have za-prefixed Complex Act Perfectives attested in the database, namely begat’i ‘run’, letat’i ‘fly’, and xodit’i ‘walk’. The incidence of zaxodit’p ‘begin to walk’ is very low, given the overall high frequency of that verb. The use of zabegat’p and zaletat’p to mean ‘begin to run’ and ‘begin to fly’ is consistent with my earlier findings concerning the meaning of za- ingressive verbs, where I stated that such verbs refer only to the initiation of an unexpected motion (Janda 1985, 1986). Since begat’i ‘run’ and letat’i ‘fly’ are salient and often rapid motions, they are the ones that should be most likely to be surprising and thus to form Complex Act ingressives, as is borne out in the data. Examples (3) and (4) illustrate typical po- and za- prefixed Complex Act Perfectives from the database:

136 Laura A. Janda

(3)

Ona po-xodila nemnogo, po-smotrela na raznye she a.while-walked.pfv a.bit, a.while-looked at various lampočki i skazala, čto xočet spat’. lamps and said that wants sleep.inf ‘She walked for a while, looked for a while at the various lamps and said that she wanted to sleep.’ (Andrej Gerasimov. Čužaja babuška, 2001)

(4) I togda [on] vdrug vstal i za-begal po kamere. and then he suddenly got.up and begin-ran.pfv around chamber ‘And then [he] suddenly got up and began running around the chamber.’ (Jurij Dombrovskij. Xranitel’ drevnostej, 1964)

The cluster model predicts that there are no secondary imperfectives associated with these verbs.

6.

Prefixed Single Act Perfectives from indeterminate stems

Since all six verbs in the database express Non-completability and can form Complex Act Perfectives with po- or za- or both, all six could potentially form Single Act Perfectives prefixed with s-, and indeed dictionaries list this option for all six verbs. However, there are no attestations in this database of the use of snosit’p ‘carry someplace and back once’ as a Single Act Perfective and attestations of svodit’p ‘lead someplace and back once’ in this meaning are rare. For the remaining four verbs, however, the use of s-prefixed Single Act Perfectives is abundantly present in the database. It may be that transitivity plays a role here, and that the idea of taking a single roundtrip is more likely to be expressed with intransitive verbs such as sbegat’p ‘run someplace and back once’, s”ezdit’p ‘ride someplace and back once’, sletat’p ‘fly someplace and back once’, and sxodit’p ‘walk someplace and back once’ than with transitive verbs such as svodit’p ‘lead someplace and back once’ and snosit’p ‘carry someplace and back once’. The distribution of perfectives relative to this meaning is very simple: all s-prefixed perfectives formed by these verbs are Single Act Perfectives. Example (5) illustrates this use: (5)

Mnogim s-letat’ v kosmos xočetsja, a èto many roundtrip-fly.pfv.inf into space want.refl but that dorože stoit – 20 millionov dollarov. more.expensive costs 20 million dollars. ‘A lot of people would like to take a trip to outer space, but it’s pretty expensiveÂ€– 20 million dollars.’ (Izvestija, 2002.05.16)

Perfectives from indeterminate motion verbs in Russian 137

According to the cluster model, it is not usually possible to form secondary imperfectives from the Complex Act verbs, nor are any attested in the RNC.

7.

Prefixed Specialized Perfectives from indeterminate stems

In reference to verbs, metonymy makes it possible to use a verb to refer only to a subset of an activity or a contiguous activity (cf., Peirsman & Geeraerts 2006). Thus, for example, the Specialized Perfective vyigrat’p ‘win’ refers to a subset of playing, namely the playing that leads to winning a game. And since winning is a perfective expression of a Completable activity, it can form an aspectual partner, the secondary imperfective vyigryvat’i ‘win’. Because subsets of activity can be more specifically targeted, metonymy makes it possible for even verbs that normally describe Non-completable situations to acquire a Completable construal and thus form Specialized Perfectives, as in pererabotat’p ‘revise’, which describes only work on something that is done but requires another round of effort. In relation to indeterminate motion verbs, the semantic shift entailed by metonymy arguably detaches the verb from its status as a motion verb, making it into a more garden-variety imperfective base verb. Thus iznosit’p ‘wear out’ selects only the portion of nosit’i that refers to carrying clothing on the body (wearing it) and doing that in such a way that damage occurs. Similarly, looking after the health of someone involves a lot of walking to and fro, metonymically referenced in vyxodit’p ‘nurse, rear’. Because these are Specialized Perfectives, they form secondary imperfectives iznašivat’i and vyxaživat’i. The database contains only one group of examples of Specialized Perfectives formed from indeterminate stems, namely examples of zaezdit’p ‘wear out by riding around’. All four examples of this verb in the database have this meaning, as illustrated in example (6): (6)

Delo tol’ko načinaetsja, a vy už nas s Elenoj event only begins but you already us with Elena za-ezdili. excessive-rode.pfv ‘The event is just beginning, and you have already worn me and Elena out with all the riding around.’ (Valentin Rasputin. Novaja professija, 1998)

Theoretically this Specialized Perfective also has a secondary imperfective, zaezživat’i ‘wear out by riding around’, which is listed in dictionaries though its use is unattested in the RNC. A study that included other prefixes and therefore more Specialized Perfectives would be expected to find attestations of secondary imperfectives, as predicted by the cluster model.

138 Laura A. Janda

8.

Conclusion

Though the majority of prefixed forms of indeterminate stems are imperfective verbs, the formation of prefixed perfectives from indeterminate motion verbs is also a robust and well-behaved phenomenon. Consistent with the behavior of other verbs in the lexicon that unambiguously express Non-completability, the indeterminate motion verbs form three types of perfectives: Complex Act Perfectives that express engagement in an activity that is bounded in time; Single Act Perfectives that express a single cycle of a repeated action, namely a single roundtrip; and Specialized Perfectives that narrow reference to only a subset of the action described by the stem.

References Bondarko, A. V. 1983. Principy funkcional’noj grammatiki i voprosy aspektologii. Leningrad: Nauka. Browne, W. 1978. Aspectual morphology and vygljadet’. Folia Slavica 2: 46–53. Čertkova, M. Ju. 1996. Grammatičeskaja kategorija vida v sovremennom russkom jazyke. Moscow: Moscow State University. Croft, W. & Cruse, D. A. 2004. Cognitive Linguistics. Cambridge: CUP. Dickey, S. M. 2007. A prototype account of the development of delimitative po- in Russian. In Cognitive Paths into the Slavic Domain, D. Divjak & A. Kochanska (eds.), 326–371. Berlin: Mouton de Gruyter. Forsyth, J. 1970. A Grammar of Aspect. Cambridge: CUP. Geeraerts, D. 1995. Representational formats in cognitive semantics. Folia Linguistica 29: 21– 41. Isačenko, A. V. 1960. Grammatičeskij stroj russkogo jazyka v sopostavlenii s slovackim. Morfolo‑ gija. Bratislava: Slovak Academy of Sciences. Janda, L. A. 1985. The meaning of Russian verbal prefixes: semantics and grammar. In The Scope of Slavic Aspect [UCLA Slavic Studies 12], M. Flier & A. Timberlake (eds.), 26–40. Columbus OH: Slavica. Janda, L. A. 1986. A Semantic Analysis of the Russian Verbal Prefixes ZA-, PERE-, DO- and OT[Slavistische Beiträge 192]. Munich: Otto Sagner. Janda, L. A. 2007. Aspectual clusters of Russian verbs. Studies in Language 31(3): 607–648. Lakoff, G. 1987. Women, Fire and Dangerous Things. Chicago IL: University of Chicago Press. Lakoff, G. & Johnson, M. 1980. Metaphors We Live By. Chicago IL: University of Chicago Press. Lakoff, G. & Johnson, M. 1999. Philosophy in the Flesh. New York NY: Basic Books. Lewandowska-Tomaszczyk, B. 2007. Polysemy, prototypes and radial categories. In The Oxford Handbook of Cognitive Linguistics, D. Geeraerts & H. Cuyckens (eds.), 139–169. Oxford: OUP. Nesset, T. 2008. Path and manner: An image-schematic approach to Russian verbs of motion. Scando-Slavica 54: 135–158.

Perfectives from indeterminate motion verbs in Russian 139

Peirsman, Y. & Geeraerts, D. 2006. Metonymy as a prototypical category. Cognitive Linguistics 17(3): 269–316. Šaxmatov, A. A. 1941. Sintaksis russkogo jayzka. Leningrad: Učpedgiz. Švedova, N. Ju. et al. 1982. Russkaja grammatika. Moscow: Nauka. Timberlake, A. 2004. A Reference Grammar of Russian. Cambridge: CUP. Tixonov, A. N. 1962. K voprosu o čistovidovyx pristavkax v russkom jazyke. Trudy Samar‑ kandskogo gosudarstvennogo universtiteta im. A. Navoi, Novaja serija 118: Issledovanija po russkomu jazyku, 31–57. Townsend, C. E. 1975. Russian Word-Formation. Columbus OH: Slavica. Vinogradov, V. V. 1938. Sovremennyj russkij jazyk. Grammatičeskoe učenie o slove. Moscow: Učpedgiz. Wade, T. 1992. A Comprehensive Russian Grammar. Oxford: Blackwell. Zaliznjak, A. A. & Šmelev, A. D. 2000. Vvedenie v russkuju aspektologiju. Moscow: Jazyki russkoj kul’tury.

Appendix 1. Abbreviations used in interlinear glooses inf ipfv pfv refl

infinitive imperfective perfective reflexive

chapter 6

Aspects of motion On the semantics and pragmatics of indeterminate aspect Olga Kagan

The Hebrew University of Jerusalem

In this paper, I investigate the semantic and pragmatic properties of the indeterminate imperfective aspect exhibited by verbs of motion in Russian. As pointed out by Forsyth (1970), indeterminate verbs are associated with a wide range of usages, which include iterativity, habituality, motion in multiple directions, and naming the type of motion. While iterativity seems to constitute a good candidate for a property that unifies all these usages, I argue that the indeterminate aspect does not entail event repetition. I argue instead that this aspect constitutes an identity function and that the aspectual restrictions imposed on sentences which contain indeterminate verbs are pragmatic, rather than semantic, in nature. I show that these restrictions are derived from independently motivated pragmatic principles.

1.

Introduction

In this paper, I discuss the aspectual properties of verbs of manner of motion in Russian. These verbs are characterized by special aspectual properties, since in addition to the typical perfective/imperfective opposition, they exhibit a further aspectual distinction, sometimes referred to in the literature as the determinate/in‑ determinate contrast. Determinate and indeterminate verbs do not differ in terms of their lexical meaning but have different aspectual usages, which are discussed in Forsyth (1970), Zaliznyak and Shmelev (2000), Isačenko (2003), Timberlake (2004), and Romanova (2007), among others. The determinate/indeterminate contrast is illustrated in (1). (In this paper, I concentrate mainly on imperfective predicates.)

142 Olga Kagan

(1) a. b.

Maša šla po parku. Masha walked.DET.IPFV in park ‘Masha was walking in the park.’ Maša xodila po parku. Masha walked.INDET.IPFV in park ‘Masha used to walk in the park / Masha (has) walked in the park.’

(1a) reports a single event of Masha walking in the park in a single direction. In contrast, (1b), out of context, can be associated with a wide range of readings. It may report a single event of Masha walking about in the park, i.e. walking here and there in various directions. It may also receive a habitual reading, according to which Masha used to walk in the park in the past, or a (non-habitual) iterative reading, meaning that Masha walked in the park more than once. However, (1b) seems to be incompatible with the only meaning that is available for (1a), i.e. it cannot report a single event of Masha walking in a single direction. Table 1 below lists verbs of motion that exhibit the determinate/indeterminate dichotomy. Most verbs are taken from Forsyth (1970). The main goal of this paper is to investigate the semantic and pragmatic properties of indeterminate imperfective verbs. I will argue that indeterminate is the default aspect that does not impose any semantic constraints on the aspectual properties of the predicate. The aspectual restrictions imposed on sentences which Table 1 Determinate imperfective

Indeterminate imperfective

English translation

idti bežat’ jexat’ plyt’ letet’ polzti lezt’ vesti vezti nesti nestis’ taščit’ katit’ katit’sja gnat’

xodit’ begat’ jezdit’ plavat’ letat’ polzat’ lazit’ / lazat’ vodit’ vozit’ nosit’ nosit’sja taskat’ katat’ katat’sja gonjat’

walk run ride, go (by car, train, etc.) swim, sail fly crawl climb drive, lead transport carry run quickly (informal) drag roll (transitive) roll, wallow (intransitive) race

Aspects of motion 143

contain indeterminate verbs are pragmatic, rather than semantic, in nature. I will discuss in detail the mechanism by which these restrictions are derived. The paper is organized as follows. I will begin with a brief introduction of the Russian aspectual system in Section 2. Section 3 is devoted to a brief discussion of the determinate verbs, which stand in an aspectual competition with the indeterminates. In Section 4, I propose a formal semantic analysis of indeterminate aspect and demonstrate how independently-motivated pragmatic constraints restrict the usage of indeterminate verbs. The latter issue is further investigated in Section 5, which focuses on the “there and back” motion meaning component associated with certain constructions that contain indeterminate verbs.

2.

The perfective/imperfective distinction in Russian

In Russian, every verb is morphologically marked for aspect – perfective or imperfective. Developing a formal analysis of the perfective/imperfective distinction in Russian is beyond the scope of this paper. (For detailed analyses of the semantics of Slavic aspect, see Smith (1991), Padučeva (1996), Filip (1999, 2000, 2008), Borik (2002), Kagan (2007b), and references therein). Below, I briefly discuss the perfective/imperfective opposition on a descriptive level. The list of usages provided below is based on discussion in Smith (1991), Padučeva (1996), and Zaliznyak and Shmelev (2000), among others. Perfective aspect is normally used to encode a single completed or temporally delimited event which has a set terminal point (see e.g., Krifka 1992:â•›50). In the most typical case, a perfective predicate is telic (although see Borik 2002 and Filip 2003 for arguments against the telicity approach to perfectivity). (2) Ivan postroil dom. Ivan built.PFV house ‘Ivan built / has built a house.’

For instance, (2) entails that Ivan finished building the house; according to this sentence, the building event has reached its natural endpoint. According to Padučeva (1996), the semantics of perfective verbs involves a change of state. In the case of (2), this is the transition from non-existence to existence of the house. In turn, imperfective aspect is characterized by a wide range of usages, some of which are introduced below. . Part of the discussion below is based on native speakers’ judgments. My informants are monolingual native speakers of Russian between the ages of 25 and 60, who live in Russia.

144 Olga Kagan

Example (3) illustrates the progressive usage of the imperfective. According to Padučeva (1996), this is the basic usage of imperfective aspect. (3) Lena čitala knigu, kogda vošel Miša. Lena read.IPFV book when entered.PFV Misha ‘Lena was reading a book when Misha came in.’

Example (3) asserts that Lena was engaged in the process of reading at the temporal interval contributed by the embedded clause. Similarly to its English progressive counterpart, the sentence does not entail that Lena finished reading the book. Another usage of the imperfective is exhibited in habitual or iterative sentences. (4) a. b.

Lena myla posudu každyj den’. Lena washed.IPFV dishes every day ‘Lena used to wash the dishes every day.’ Lena zvonila ves’ den’. Lena phoned.IPFV whole day ‘Lena phoned all day.’

(4a) is a habitual sentence. Due to the presence of an imperfective verb, it can receive a habitual interpretation even in the absence of the adjunct každyj den' (every day). (4b) does not involve habituality or genericity; however, it does denote a set of repeated events, an interpretation that again is associated with imperfectivity. Imperfective verbs also appear in sentences that exhibit the so-called State‑ ment of Fact convention of usage, discussed extensively in Forsyth (1970), Comrie (1976), Smith (1991), Padučeva (1996) and Gronn (2003), among others. Such sentences convey the information that an event of a certain type did take place. The focus in this case is not on how many times the event occurred or on whether it took place once or repeatedly, but rather on the fact that this kind of eventuality has been instantiated. This usage is exemplified in (5), which asserts roughly that Lena has had the experience of taking the medicine in question. (5) Lena (uže) prinimala eto lekarstvo. Lena already took.IPFV this medicine ‘Lena has (already) taken this medicine.’

Importantly, a single event of the specified type is sufficient to make sentences of this type true, even if telic event predicates are involved (Padučeva 1996:â•›39–40). . See also Padučeva (1996:â•›10, 32–52) for a discussion of different subtypes of Statement of Fact.

Aspects of motion 145

For instance, (5) entails that Lena took the medicine at least once. This point is supported by the fact that, under the Statement of Fact interpretation, an imperfective verb is perfectly compatible with such adjuncts as (tol’ko) odin raz ((only) once): (6) Lena prinimala eto lekarstvo tol’ko odin raz. Lena took.IPFV this medicine only one time ‘Lena took this medicine only once.’

This reveals that imperfectivity is compatible with both telic and atelic event descriptions (for a detailed discussion of this point, see Filip 1999, 2000; Borik 2002; and Kagan 2007b; among others.)

3.

Determinate imperfective verbs

Typically, for a given verb, all the different usages of imperfective aspect involve exactly the same form. Verbs of manner of motion are special in this respect, since they have two distinct imperfective forms, determinate and indeterminate, which divide their aspectual functions in a certain way. While this paper focuses primarily on the indeterminate aspect, it is important to begin with a brief discussion of determinate verbs, which are minimally contrasted with indeterminate ones. As discussed by Forsyth (1970), sentences that contain a determinate imperfective verb denote a single event of motion in a single direction. (7) a. b.

Lena šla po ulice (okolo desjati minut). Lena walked.DET.IPFV on street near ten minutes ‘Lena was walking / walked down the street (for about ten minutes).’ Katja bežala v magazin. Katja ran.DET.IPFV to shop ‘Katja was running to a shop.’

For instance, (7a) encodes a single event of Lena walking down the street, in one direction. This sentence is not interpreted habitually. In this sense, determinate aspect is similar to the English progressive. English progressive sentences do not typically receive a habitual reading, unless such an interpretation is forced due to the presence of an overt iterativity-inducing operator (e.g., frequently; see Bennett and Partee (1972:â•›25–26) for discussion). The case is similar with determinate aspect in Russian. Properties shared by imperfective determinate and progressive aspect are not limited to the fact that both typically denote a single event. For instance, both types of aspect are associated with lack of commitment that the reported event reaches its inherent endpoint (if the predicate is inherently telic). Whenever a

146 Olga Kagan

determinate imperfective verb combines with a goal phrase, as in (7b), the resulting sentence does not entail that the goal is achieved. Thus, (7b) does not entail that Katja reached the shop. The same holds true for the English translation of the sentence. This suggests that an analysis of determinate aspect can be based on one of the existing approaches to the English progressive. This could be a partitivity approach, along the lines of Krifka (1992). Under such an approach, a sentence with a determinate verb would be analyzed as encoding an event that constitutes part of a potentially longer event of motion. Alternatively, an intensional approach of Dowty (1979) can be adapted to capture the semantics of such sentences. Such an analysis is proposed in Kagan (2007a). According to this analysis, sentences with determinate verbs report a single event of motion which reaches its natural endpoint in inertia worlds (cf., Dowty 1979), i.e. worlds in which events reach their normal completion without being interrupted. Since the actual world need not constitute an inertia world, the event need not reach its natural endpoint in reality. It is not crucial for our current purposes how exactly the semantic contribution of determinate aspect is formalized. The important point is that these sentences encode a single event of motion in a single direction, which does not necessarily reach a natural endpoint in the actual world. With this generalization in mind, let us now turn to indeterminate aspect.

4.

Indeterminate imperfective verbs

4.1

Descriptive facts

Indeterminate aspect is compatible with a wider range of aspectual interpretations. Below, I discuss several possible readings that are listed by Forsyth (1970).

4.1.1 Multiple directions Firstly, indeterminate imperfective aspect is found under the so-called multiple directions reading. This is a usage whereby a sentence with an indeterminate verb encodes what is conceived of as a single event, which is an event of motion “of one or more subjects in various unspecified directions” (Forsyth 1970:â•›321). For instance, (8) means that Lena was walking about in the park, walking here and there, rather than that she was moving in any particular direction. (8) Lena xodila po parku. Lena walked.INDET.IPFV in park ‘Lena was walking about in the park.’

Aspects of motion 147

4.1.2 Iterativity Secondly, indeterminate aspect is used in habitual and iterative sentences. For instance, (9) is a habitual sentence. Importantly, it receives a habitual reading even in the absence of the adverbial často (often); it cannot be interpreted as encoding a single event. (9) Lena (často) ezdit v derevnju. Lena often rides.INDET.IPFV to village ‘Lena (often) goes to the village.’

4.1.3 Naming the type of motion (generic sentences) In addition to multidirectional and iterative meanings, Forsyth notes that indeterminate verbs can be used to “denote the type of motion as a generalized phenomenon without reference to any specific occasion” (1970:â•›321). This usage is exemplified in (10): (10) a. b. c.

Naš rebjonok uže xodit. our child already walks.INDET.IPFV ‘Our child can walk already’ Lena vodit mašinu. Lena drives.INDET.IPFV car ‘Lena drives a car. (= Lena can drive a car)’ Pticy letajut. birds fly.INDET.IPFV ‘Birds (can) fly.’

These examples illustrate a type of generic sentence. Rather than denoting a given event or set of events taking place in the actual world at a particular temporal interval, they receive a modalized reading. For instance, (10c) encodes a certain non-accidental pattern, which is predicted to hold not only in reality but also in alternative possible worlds in which birds behave normally (Asher and Morreau 1995). In turn, (10b) reports Lena’s ability to drive a car; she may not do so in reality, because she has no car for instance, but she is predicted to do the driving in potential appropriate situations. Forsyth claims that, since indeterminate aspect is characterized by a wide range of different usages, it cannot be provided a unified account. Rather, it should be treated as a default aspect, which lacks a unified semantics and which is used whenever determinate aspect is inappropriate. While I agree with Forsyth that the indeterminate is a default aspect (a view also found in Jakobson 1984), I believe that it can be given a unified account. In particular, I propose that indeterminate aspect should be analyzed as an identity

148 Olga Kagan

function (a function that returns the same value that was used as its argument). In other words, the indeterminate is the default aspect that does not impose any aspectual restrictions and is compatible with any aspectual interpretation. One potential weakness of such an analysis is the fact that indeterminate verbs are not typically used to report single events of motion in one direction. Rather, indeterminate aspect seems to be associated with event plurality, as it is found in iterative, habitual and generic sentences. The multiple directions reading can also be conceived of as involving event plurality: the overall event of motion in multiple directions can be conceived of as consisting of a plurality of sub-events, each of which involves motion in a single direction. If indeterminate aspect constitutes an identity function, how can we account for these facts? In the next subsection, I argue that, despite the fact that indeterminate aspect is usually associated with event plurality, this does not follow from its semantics. I will show that indeterminate aspect is, in fact, semantically compatible with the single, unidirectional motion event reading. Then, in Section 4.3, I provide a formal semantic analysis of this aspect and argue that the restrictions imposed on its usage are derived from a pragmatic principle.

4.2

Indeterminate aspect and single events of motion in a single direction

4.2.1 Negation The fact that sentences with indeterminate verbs do not entail event plurality is revealed under negation. Negative sentences that contain an indeterminate verb do not only deny the fact that an event took place repeatedly, or more than once; they also assert that the event in question did not take place even once. (11) Lena (ešče) ne ezdila v derevnju. Lena yet NEG rode.INDET.IPFV to village ‘Lena didn’t go to the village / hasn’t gone to the village yet.’

Thus, (11) entails that Lena did not go to the village even once. A single event of going to the village is sufficient to make the sentence false, which, in turn, means that a single event is sufficient to make its affirmative counterpart true. In fact, the same reasoning shows that imperfective verbs in general do not normally entail iterativity in Russian. Thus, (12) below entails that Lena did not call even once. (12) is made false by a single calling event, which suggests that a single calling event makes its affirmative counterpart true.

Aspects of motion 149

(12) Lena ne zvonila. Lena NEG phoned.IPFV ‘Lena didn’t call / hasn’t called.’

Indeed, although imperfective verbs can be used in iterative and habitual constructions, a sentence with such a verb can be made true by a single event of the specified type, as revealed, for example, under the Progressive, Statement of Fact, and Annulled Result usages (see e.g., Forsyth 1970; Smith 1991; and Section 4.2.3 below). I believe that such sentences as (11) and (12) do prove that the verbs involved do not entail iterativity, even when found in affirmative clauses. It is true that the relation between a negative sentence and its affirmative counterpart is complex and not symmetric in various respects (see Horn 1989 for a detailed discussion). Still, the meaning of such sentences as (11) and (12) is compositional. If the denotation of the verbs ezdila (rode) and zvonila (phoned) contains only pluralities of events, it is unclear why their negation results in a denial of a single event, which is not even included in their denotation. In contrast, if we assume that the denotation of these verbs contains single events as well as pluralities, the semantics of (11) and (12) follows straightforwardly.

4.2.2 Interrogatives The same conclusion can be drawn if we consider interrogative clauses with indeterminate verbs. For instance, consider the polar interrogative in (13): (13) Ty (kogda-nibud’) ezdil vdol’ etogo berega? you ever rode.INDET.IPFV along this beach ‘Have you ever driven along this beach?’

A single event of the addressee driving along the beach in one direction is sufficient in order for the answer to (13) to be positive. This suggests that the proposition with which this question is associated is made true by a single unidirectional event of motion.

4.2.3 Statement of fact interpretation The compatibility between indeterminate aspect and single event interpretation is revealed in upward-entailing environments as well. For instance, it can be observed in sentences that exhibit the Statement of Fact convention. Statement of Fact sentences with verbs of motion typically involve indeterminate aspect. (14) a. On uže plaval vniz po reke. he already swam.INDET.IPFV down on river ‘He has swum / sailed down the river.’

150 Olga Kagan

b. c. d.

My uže begali naperegonki (ot vorot do we already ran.INDET.IPFV chasing.each.other from gates to fontana). fountain ‘We have already run a race from the gates to the fountain.’ Lena uže letala vo Franciju. Lena already flew.INDET.IPFV to France ‘≈ Lena went to France by plane.’ Ja uže vodila Diminu novuju mašinu. On dal mne I already drove.INDET.IPFV Dima’s new car he let me projexat’ ot ofisa do magazina. drive from office to shop ‘I have already driven Dima’s new car. He let me drive it from the office to the shop.’

As discussed in Section 2, a single event of the specified type is sufficient to make a Statement of Fact sentence true. Each of the sentences in (14) entails that the kind of event denoted by the verb and its arguments has been instantiated at least once. A single event of moving in the specified direction is sufficient to make each sentence true. Moreover, all the sentences are compatible with the adjunct odin raz (once), which makes it especially clear that the indeterminate verb is compatible with a single event interpretation. Such compatibility is further exhibited in a number of attested examples discussed below. (15) Odnaždy ja uže xodil po etoj tropinke. once I already walked.INDET.IPFV on this trail ‘I walked this trail once.’ (Hodgson 1994)

It is important to emphasize that (15) is acceptable even if the speaker walked the trail only in one direction. Thus, it can be followed felicitously by (16): (16)

Ja togda prošel po nej ot podnožija do ručja, a potom I then walked.PFV on her from foot.hill to spring and afterwards svernul na druguju dorogu. turned.PFV on different road ‘I walked it from the foot (of the hill) to the spring and then took a different road.’

. Such sentences as (14c) are strongly associated with a “there and back”, or roundtrip reading (e.g., Forsyth 1970; Timberlake 2004). Thus, (14c) suggests that Lena went to France and back. I believe, however, that the “back motion” meaning component is not entailed by sentences of this type. They are, in fact, made true by a single unidirectional event of motion. See Section 5 for discussion.

Aspects of motion 151

Further, the verb phrase (VP) in the second sentence in (17) below clearly encodes a single event, most likely of swimming in one direction – towards the shore. (17)

Ja ekstremal… Odin raz plaval bukval’no naperegonki I thrill.seeker one time swam.INDET.IPFV literally on.competition s akuloj. …ja molotil po vode rukami i nogami with shark I thrashed on water hands.INSTR and feet.INSTR izo vsex sil, čtoby dobrat’sja do berega bystree. from all forces so.that reach.INF to shore quicker ‘I am a thrill-seeker. I once swam while chased by a shark. I thrashed the water with my hands and legs with all my might in order to reach the shore more quickly.’ (Internet; Komsomol’skaja Pravda, July 2007)

The fact that motion in a single direction is sufficient to satisfy a Statement of Fact sentence is further illustrated by (18). (18)

Odin raz ezdil s Kurtzem na poezde po maršrutu one time rode.INDET.IPFV with Kurtz on train along route Moskva-Petrozavodsk… Vernee, dva raza – obratno Moskow-Petrozavodsk more.exactly two times back Petrozavodsk-Moskva tože v poezde exal. Petrozavodsk-Moscow also in train rode.DET.IPFV ‘I rode with Kurtz by train on the route Moscow-Petrozavodsk once. In fact, twice: I rode back, from Petrozavodsk to Moscow, also by train.’ (Internet; “Oxotničji Sobaki,” 2008)

This example demonstrates that a single trip in one direction (from Moscow to Petrozavodsk) is sufficient in order for the speaker to say ezdil. In this particular case, the memory that the return journey took place as well leads the speaker to state that an event of motion was instantiated twice.

4.3

Indeterminate aspect: The analysis

The above discussion reveals that indeterminate aspect does not encode event plurality. I propose that the denotation of an indeterminate verb is not restricted aspectually. Essentially, a sentence with an indeterminate verb means that a motion event of the type specified by the verb and its arguments has been instanti‑ ated. The semantics of indeterminate aspect is formalized in (19): (19) [[Ind]] = λPλe.P(e)

152 Olga Kagan

Thus, I propose to analyze indeterminate aspect as an identity function. This analysis accounts for the fact that indeterminate predicates are semantically compatible with both single event and plurality of events interpretations. It also predicts the compatibility of indeterminate aspect with the Statement of Fact interpretation, since sentences that exhibit this convention entail that an event property is instantiated. This is precisely the semantics that I propose to assign to sentences with indeterminate verbs. In addition, the proposed analysis captures the negation facts introduced in Section 4.2.1. A negative sentence with an indeterminate verb is predicted to entail that the event property in question has not been instantiated, which means that even a single event of the specified type did not take place. The prediction is borne out, as revealed above. The question that remains unanswered at this stage is why affirmative sentences with indeterminate verbs do not typically receive a “single unidirectional motion event” meaning. Why is the usage of this aspect intuitively associated with event plurality? Why is (20) below not likely to be used to encode a single motion event, unless a Statement of Fact interpretation is invoked? (20) Lena begala v storonu supermarketa. Lena ran.INDET.IPFV in direction supermarket.GEN ‘Lena ran in the direction of the supermarket.’

I propose that the restriction under discussion is not semantic in nature, but rather pragmatic. The usage of indeterminate aspect is subject to an independentlymotivated informativeness constraint, which can be derived from the Gricean Maxim of Quantity (1975). This constraint has been divided in the literature into two principles, Maximize Assertion and Maximize Presupposition. The formulation that these principles receive in Sauerland (2003b) is provided below: (21) Maximize Assertion: Use the most informative assertion that is true. Maximize Presupposition: Use the most informative presupposition that is satisfied.

. I assume an approach to the imperfective that treats it as a default aspect which should be analyzed as an identity function. (For discussion of the default aspect approach to imperfectivity, see Forsyth 1970; Klein 1995; Kagan 2007b; and references therein.) Under this approach, indeterminate aspect should be treated as an instance of the “ordinary” imperfective, whereas determinate verbs are characterized by special aspectual properties. If one assumes a different analysis of the imperfective, for instance one according to which imperfectivity encodes atelicity or unboundedness (e.g., Smith 1991), the approach to indeterminate aspect would have to be modified accordingly. In particular, this aspect should be subject to exactly the same restrictions as the ordinary imperfective.

Aspects of motion 153

If two grammatical forms are compatible with the same interpretation, we are instructed by (21) to use the one that has a more restricted meaning, i.e. the one that contributes stronger entailments or presuppositions. If the less restricted form is used, the hearer concludes that the use of the more restricted item was impossible because its presuppositions or semantic restrictions are not satisfied. This inference is often quite strong and not easy to cancel; Sauerland even assigns it the status of an Implicated Presupposition, in case two forms or assertions that differ in the strength of their presuppositions are involved. These constraints have been claimed to determine the choice between competing grammatical forms in different languages and within the framework of various phenomena; for instance, Heim (1991), who introduces Maximize Presupposition, argues that this principle determines the choice between definite and indefinite NPs. Sauerland (2003a) claims that the same principle restricts the usage of the plural morphology, and in Kagan (2007b), it is proposed that Maximize Assertion determines in certain cases the choice between perfective and imperfective verb forms in Russian. Turning back to verbs of motion, crucially, a determinate form is clearly more informative than its indeterminate counterpart. Therefore, whenever the former is appropriate, the latter is ruled out by Maximize Assertion. This is why indeterminate aspect is not generally used to report a single, unidirectional motion event: this interpretation is compatible with the more informative determinate form. The distribution of indeterminate verbs therefore reveals the important role that is devoted to pragmatic restrictions in Russian aspectual system. It has been demonstrated above that sentences with indeterminate verbs are made true by a single unidirectional event of motion. At the same time, such sentences are not typically used if such an event is reported (unless the special Statement of Fact interpretation is invoked). This discrepancy between truth conditions and usage can now be accounted for. Semantically, indeterminate aspect is perfectly compatible with single event interpretation. But pragmatically, when a single event is encoded and determinate aspect is appropriate, indeterminate aspect is ruled out by Maximize Assertion.

. It should be noted that, although highly restricted, the use of determinate verbs in Statement of Fact sentences is not absolutely impossible (Padučeva 1996:â•›44–46). But crucially for our purposes, whenever a determinate verb does appear in a Statement of Fact clause, the resulting interpretation differs from the one obtained with indeterminate verbs as illustrated in Section 4.2.3 (see e.g., Padučeva 1996:â•›44). Therefore, Maximize Assertion does not rule out the use of indeterminate aspect in such sentences as (14)–(15). They involve interpretation that is inherently incompatible with the semantics of determinate verbs.

154 Olga Kagan

5.

The “back motion” meaning component

In this section, I discuss a special meaning component that arises in sentences that contain an indeterminate verb combined with a goal phrase. Such sentences tend to be interpreted as reporting “there and back” motion (Forsyth 1970; Timberlake 2004). As noted in Timberlake (2004), indeterminate verbs can be “used to express … a roundtrip on a single occasion” (p. 412). A relevant example is provided below. (22) Odin raz Lena uže letala vo Franciju. one time Lena already flew.INDET.IPFV to France ‘≈ Lena went to France by plane once.’

The sentence seems to assert that Lena went to Paris and came back. Given that, as argued above, indeterminate aspect does not entail event plurality, it is unclear how the “back motion” meaning component emerges. It cannot be contributed by the goal phrase and, thus, does not seem to constitute part of the compositional semantics of the VP. Crucially, note that even if the indeterminate verb had entailed event plurality, the roundtrip reading would still not be derived compositionally. The attachment of a goal phrase would result in a VP denoting a plurality of motion events directed towards the goal, i.e. a plurality of flights to France. Example (22) would therefore be predicted to be unacceptable due to the presence of odin raz (once). Padučeva (1996) discusses the bidirectional subtype of Statement of Fact sentences. This discussion includes prefixed verbs of motion (e.g., prixodit’ (arrive)) and predicates other than verbs of motion (e.g., brat’ (take)). Padučeva points out that the action denoted by these verbs is naturally associated with motion (or other action) in the opposite direction. Crucially, the latter event does not fall under the denotation of the original predicate, but rather constitutes an implicature. According to Padučeva, the bidirectionality implicature is an instance of the annulled result meaning component that consistently accompanies certain subtypes of Statement of Fact sentences. (The relation between the bidirectionality meaning component and annulled result state is discussed in Section 5.2.2.) Below, the nature of the “back motion” meaning component of such sentences as (22) is further investigated. First, additional evidence is provided that this component is not entailed by VPs that consist of an indeterminate verb and a goal. Second, I discuss the nature of the “there and back” reading, ways in which it may arise, and its relation to the original, literal meaning of the VPs.

Aspects of motion 155

5.1

Not an entailment

VPs of the type introduced above are indeed quite strongly associated with the “there and back” reading on the intuitive level. However, the “back motion” meaning component is not entailed by such VPs under their literal, purely compositional meaning. First, the “back motion” meaning component does not (or at least is much less likely to) arise in sentences that do not contain a goal phrase, such as (14d) above. If this meaning component is entailed by the verb in (22), this contrast is definitely unexpected. Second, (22) is true even if Lena went to France by plane but went back by train. Namely, it is true if the “back motion” is not of the manner specified by the verb. Moreover, if we add an adjunct to the VP, this adjunct, again, can be interpreted as modifying the “there” event but not the “back motion”. Thus, (23) below can be followed felicitously by (24). This makes it clear that the adjunct PP need not modify the back journey: (23) Odin raz ja uže ezdila v Eilat na starom avtobuse. one time I already rode.INDET.IPFV to Eilat on old bus ‘I went to Eilat on an old bus once.’ (24)

Eto bylo užasno. K sčastju, obratno Dima otvjoz menja na it was horrible to luck back Dima drove.PFV me on svoej mašine! self car ‘This was terrible. Fortunately, Dima took me back on his car.’

What we see is that the “back” event may be of a different manner than the one specified by the verb; it need not conform to restrictions imposed by the adjuncts present in the sentence. In addition, its goal is certainly not the goal specified in the sentence. (For instance, Eilat is not the goal of the back journey associated with (24), but rather the source.) All of this suggests that motion “back” cannot be semantically encoded by the VP. Also, negation facts prove that sentences like (22) entail only the “there” event. If we negate such sentences, as is exemplified in (25), the resulting sentence entails that the “there” journey did not take place. Thus, (25) entails that the speaker did not go to Paris by plane. However, had the corresponding affirmative sentence encoded two events (motion there and back), we would expect the negative sentence also to be true if the speaker did go to Paris but did not come back. Namely, the non-realization of one of these events would be sufficient for the sentence to be negated. This does hold for (26), with the VP explicitly encoding motion in two directions. This sentence can be felicitously followed by (27). But (25) cannot

156 Olga Kagan

be followed by (27), nor even by a sentence meaning “I only came to Paris”. This results from the fact that if the speaker came to Paris by plane, that is sufficient for (25) to be false, whatever the state of affairs with the back journey is. (25) Ja ne letala v Pariž. I NEG flew.INDET.IPFV to Paris ‘I didn’t go to Paris by plane.’ (26) Ja ne letala v Pariž i obratno. I NEG flew.INDET.IPFV to Paris and back ‘I didn’t go to Paris and back by plane.’ (27) Ja priletela tuda, a obratno ujexala poezdom. I flew.DET.PFV there and back rode.DET.PFV train.INSTR ‘I came there by plane and went back by train’.

Furthermore, the fact that sentences under discussion do not entail the “back motion” is supported by native speakers’ judgments regarding the following scenario. Lena is now in Paris; she has come to Paris by plane, by Austrian Airlines. A lottery is announced, with the following terms of participation: (28)

Každyj, kto letal v Pariž “Avstrijskimi avialinijami”, each.one who flew.INDET.IPFV into Paris Austrian.INSTR airlines.INSTR možet učastvovat’ v loteree. can participate in lottery ‘Anyone who flew to Paris by Austrian Airlines may participate in the lottery.’

My informants were asked whether, under this scenario, Lena is eligible for participation. All the informants gave a positive answer, despite the fact that Lena is assumed to be still in Paris and not to have flown away.

5.2

How is the “back motion” component derived?

In this section, I discuss a mechanism whereby the “back motion” meaning component can be derived. In fact, I will propose that there are two, probably interrelated, ways in which it comes to be associated with indeterminate verbs. Firstly, I believe that sentences with an indeterminate verb and a goal phrase contribute the “back motion” component via an implicature (along the line of Padučeva 1996). In order to discuss this pragmatic mechanism in more detail, some background in the pragmatics of the perfective/imperfective competition in Russian is needed. Below, I briefly review the ways in which this competition is affected by the Maximize Assertion principle as proposed in Kagan (2007b). Then, in Section

Aspects of motion 157

5.2.2, I demonstrate how the pragmatic constraint in question accounts for the emergence of the “back motion” implicature in sentences with verbs of motion. Secondly, in Section 5.2.3, it is shown that in some cases, VPs that contain an indeterminate verb and a goal phrase are used to relate to a whole trip, which includes the state of the subject occupying the goal location. It will be suggested that these cases involve a (conventionalized) shift in the meaning of the VP, whereby the latter is used to encode eventualities that fall beyond the scope of its original, literal meaning.

5.2.1 Perfective/imperfective opposition and Maximize Assertion In Kagan (2007b) I argue that the choice between perfective and imperfective verbs in Russian is partly governed by the Maximize Assertion principle, which has been introduced above. I assume that perfective predicates are used to denote single delimited events, whereas the imperfective constitutes a default aspect (e.g., Jakobson 1984; Forsyth 1970). Imperfective verbs are compatible with telic and atelic single event, as well as plurality of events, readings. Crucially, this view makes the relation between perfective and imperfective verbs partly similar to the relation between determinate and indeterminate imperfectives. In both cases, the first member of the opposition is the marked one, given that it is more restricted in meaning. The second member of the opposition is the default form. As a result, the perfective/imperfective contrast is sensitive to Maximize Assertion in the same way as the determinate/indeterminate opposition. If a single delimited event is to be encoded, both the perfective and the imperfective forms are semantically compatible with the intended meaning. In this case, Maximize Assertion forces the speaker to use the more informative perfective. In turn, if the speaker chooses an imperfective verb, the hearer concludes that the more restricted perfective form was inappropriate. This reasoning gives rise to various meaning components. In Kagan (2007b) I discuss how the semantics of perfectivity and imperfectivity, in combination with Grice’s Cooperative Principle, accounts for the choice of aspect in various contexts and environments. 5.2.2 Motion back as annulled result state Let us now turn again to sentences exhibiting the “there and back” component and consider these sentences in the context of the perfective/imperfective competition. (29) (Odin raz) Lena letala vo Franciju. one time Lena flew.INDET.IPFV to France ‘Lena went to France by plane (once).’

158 Olga Kagan

Suppose that Lena has gone to France and is still there at the speech time. Suppose further that the speaker is aware of this state of affairs. Under these circumstances, she is expected to use a perfective verb of motion while reporting the event. Sentences with perfective verbs entail that a single delimited event took place (and reached its natural endpoint) and contribute a strong implicature that the result state brought about by this event holds. This is precisely the kind of meaning that we want in the given context. And once a perfective verb is appropriate, we are required to use it (rather than an imperfective one) by Maximize Assertion. Sentences with two different perfective verbs could be used to convey the relevant meaning. More precisely, these verbs would contain the same root but different prefixes. The choice depends on the anchor from whose point of view the situation is described. If the anchor (possibly but not necessarily the speaker) is located in France (the goal), the appropriate perfective verb would be priletet’ (arrive by flying.DET.PFV) and the resulting sentence (30): (30) Lena priletela vo Franciju. Lena arrived.by.flying.DET.PFV in France ‘Lena has come to France by plane.’

If the anchor is associated with the source location, the verb would be uletet’ (fly away.DET.PFV) and the sentence (31): (31) Lena uletela vo Franciju. Lena flew.away.DET.PFV to France ‘Lena has gone to France by plane.’

In the case of (30), the reported event is Lena’s arrival in France, and the corresponding result state is Lena being in France. In (31), the reported event is Lena leaving the contextually supplied source location with the intention of going to France. The implicated result state is one of Lena not being in the source location. Suppose, however, that instead of uttering (30) or (31), the speaker uses an indeterminate imperfective verb of motion and utters (29). The hearer concludes that the restricted and informative perfective verb was not appropriate. This means that some meaning component contributed by perfective aspect does not hold. There are two main candidates for such a component. One possibility is that Lena went in the direction of France but did not reach the goal, i.e. the encoded event did not reach its natural endpoint. (Note that such a situation makes inappropriate (30) but not (31); the latter is made true as soon as the flying event begins.) In this case, it is the determinate imperfective that we expect to be used. It is the most appropriate form, as it is also informative and encodes a single unidirectional event of motion that does not necessarily reach its natural endpoint.

Aspects of motion 159

The other possible reason for the inappropriateness of perfective aspect is that Lena did arrive in France, but the result state brought about by the encoded event no longer holds at the speech time. In the case of (30), this means that she is no longer in France; in the case of (31), that she is again in the source location. In numerous contexts, these two states would be predicted to coincide. In this case, indeterminate imperfective aspect will be used. In other words, the hearer is predicted to conclude, on the basis of the fact that the informative perfective form was not chosen, that perfectivity was not appropriate because the result state does not hold. This, in turn, leads to the conclusion that Lena is no longer in Paris and is likely to have come back to her original location. This is how the “there and back” interpretation arises. As pointed out above, inferences that are based on the speaker’s choice to use a less restricted form are very strong and not easily cancellable. It is important to note that the above discussion apparently illustrates a more general mechanism, whereby numerous Statement of Fact sentences are accompanied by the annulled result state meaning component (Padučeva 1996). We are now in a position to explain why the “there and back” reading does not emerge in sentences that do not contain a goal phrase. Such sentences involve atelic event predicates, activities which are not associated with an inherent endpoint and are therefore not expected to be realized as perfective VPs. Therefore, the usage of an imperfective verb does not give rise to implicatures of the kind discussed above.

5.2.3 Using indeterminate verbs to encode a whole trip In the previous section, it was suggested that VPs containing an indeterminate verb and a goal phrase are associated with the “there and back” reading for pragmatic reasons. An additional, probably related, fact should be noted. In some cases, a shift in meaning (possibly coercion) takes place, whereby such phrases come to be used to encode a trip as a whole – which includes motion in two directions as well as, most crucially, the intermediate state of the subject occupying the goal location. Thus, one may utter (32a) with the intention to communicate that Lena has been at the post office, while (32b) may be used to inform the addressee that Katja has been in Paris in the summer. (32) a. b.

Lena xodila na počtu. Lena walked.INDET.IPFV to post.office ‘Lena went to the post office.’ Etim letom Katja ezdila v Pariž this summer Katja rode.INDET.IPFV to Paris ‘Katja went to Paris this summer.’

160 Olga Kagan

This usage is characterized by two special properties. First, the manner of motion meaning component becomes much less salient. The verb ezdit’ (ride/drive.INDET. IPFV), for example, can be used even if the goal was reached by plane or by ship. Thus, it comes close to losing its manner meaning component, and becomes similar in this respect to the English verb go, which does not encode manner. Further, the likelihood of obtaining the “whole trip” reading decreases in the presence of adverbials that modify specifically the event of motion. Apparently, this happens because under this reading, an event of motion (on which the original semantics of the verbs in question is based) becomes less salient; rather, the focus is on the fact that the trip as a whole took place and that the subject has been in the location denoted by the goal phrase. Second, as stated above, the event reported by such sentences includes the state of the subject occupying the goal location. Thus, (32b) can be uttered as an answer to the question How did Katja spend the summer?. Clearly, it is Katja’s stay in Paris, or her trip as a whole, that is of importance for the purposes of answering the question, rather than the several hours that she spent in the plane or train. Thus, (32b) reports an event that includes the state of Katja being in Paris. But obviously, the state of being in a certain location (without necessarily undergoing any motion during that period) is not included in the original, literal denotation of the verbs of motion. A related fact is that under the discussed usage, the eventuality encoded by the VP is not conceived of as a set of two events (motion “there” and “back”) but rather as a single (complex) event, i.e. the whole trip. The above discussion suggests that examples of this type exhibit a special usage of indeterminate verbs, one that goes beyond their original semantics and involves some kind of shift in the meaning of the V or VP. A VP that originally denotes an event of motion conventionally comes to be used to encode a trip, of which this event of motion constitutes the first stage. It is possible that we deal with coercion. A more detailed investigation of the issue is needed, which would address the question of how exactly the “whole trip” reading is derived and how it is related to the original denotation of indeterminate verbs. I leave such an investigation for future research.

6.

Conclusion

To sum up, in this paper I have proposed a unified analysis of indeterminate imperfective verbs of motion in Russian. According to this approach, indeterminate aspect constitutes an identity function. The usage of these verbs is restricted by pragmatic principles, in particular, by Maximize Assertion, which dictates that

Aspects of motion 161

out of two (or more) competing grammatical forms, the one with stronger entailments has to be used. Pragmatic constraints play an important role in Russian aspect. Due to Maximize Assertion, indeterminate verbs of motion come to be associated with event plurality (or at least with a plurality of directions). The same pragmatic principle has been proposed to constitute one source of the “there and back” reading that sometimes arises in sentences in which indeterminate verbs combine with a goal phrase. Finally, as argued in Kagan (2007b) and discussed briefly in Section 5.3.1, the restrictions imposed on the usage of imperfective aspect are also pragmatic in nature and are due to the competition between this (default) aspect and the more informative perfective. These facts together create a more general picture of strong interrelation between semantic and pragmatic principles in the realm of Russian aspect. One therefore has to be careful while determining whether a given meaning component associated with a particular aspectual form constitutes an inherent part of its semantics, or rather emerges as a result of the application of a pragmatic rule. Distinguishing between the two options may not be a trivial matter.

References Asher, N. & Morreau, M. 1995. What some generic sentences mean. In The Generic Book, G.Â€N. Carlson & F. J. Pelletier (eds.), 300–338. Chicago IL: University of Chicago Press. Bennett, M. & Partee, B. 1972. Toward the Logic of Tense and Aspect in English. Bloomington: IN: Indiana University Linguistics club. Borik, O. 2002. Aspect and Reference Time. PhD dissertation, University of Utrecht. Comrie, B. 1976. Aspect. Cambridge: CUP. Dowty, D. R. 1979. Word and Meaning in Montague Grammar. Dordrecht: Reidel. Filip, H. 1999. Aspect, Eventuality Types and Nominal Reference. New York NY: Garland. Filip, H. 2000. The quantization puzzle. In Events as Grammatical Objects, C. Tenny & J. Pustejovsky (eds.), 3–60. Stanford CA: CSLI. Filip, H. 2003. Prefixes and the delimitation of events. Journal of Slavic Linguistics 11(1): 55– 101. Filip, H. 2008. Events and maximalization. In Theoretical and Crosslinguistic Approaches to the Semantics of Aspect, S. Rothstein (ed.), 217–256. Amsterdam: John Benjamins. Forsyth, J. 1970. A Grammar of Aspect. Cambridge: CUP. Grice, H. P. 1975. Logic and conversation. In Syntax and Semantics, Vol. 3: Speech Acts, P. Cole & J. L. Morgan (eds.), 41–58. New York NY: Academic Press. Gronn, A. 2003. The Semantics and Pragmatics of Russian Factual Imperfective. PhD dissertation, University of Oslo. Heim, I. 1991. Artikel und Definitheit (Articles and definiteness). In Semantik: Ein internation‑ als Handbuch der zeitgen¨ossischen Forschung, A. von Stechow & D. Wunderlich (eds.), 487–535. Berlin: de Gruyter.

162 Olga Kagan

Hodgson, W. H. 1994. Golos v noči. In Šedevry anglijskogo gotičeskogo rasskaza, I. N. Vasiljeva, (ed.), 511–530. Slovo. Horn, L. R. 1989. A Natural History of Negation. Chicago IL: University of Chicago Press. Isačenko, A. V. 2003. Grammatičeskij stroj russkogo jazyka v sopostavlenii s slovackim. Moscow: Jazyki slavjanskoj kultury. Jakobson, R. 1984. Russian and Slavic Grammar. Berlin: Mouton. Kagan, O. 2007a. On the semantics of verbs of motion in Russian. In Proceedings of Israel As‑ sociation for Theoretical Linguistics 23 (IATL 23): 1–15. Kagan, O. 2007b. On the semantics of aspect and number. In Annual Workshop on Formal Ap‑ proaches to Slavic Linguistics: The Stony Brook Meeting, A. Antonenko et al. (eds.), 185–198. Ann Arbor MI: Michigan Slavic Publications. Klein, W. 1995. A time-relational analysis of Russian aspect. Language 71: 669–695. Krifka, M. 1992. Thematic relations as links between nominal reference and temporal constitution. In Lexical Matters, I. Sag & A. Szabolcsi (eds.), 29–53. Stanford CA: CSLI. Padučeva, E. V. 1996. Semantičeskie issledovania. Moscow: Jazyki Russkoj Kul’tury. Romanova, E. 2007. Constructing Perfectivity in Russian. PhD dissertation, University of Tromsø. Sauerland, U. 2003a. A new semantics for number. In Proceedings of SALT 13, R. Young & Y.Â€Zhou (eds.), 258–75. Ithaca NY: CLC, Cornell University. Sauerland, U. 2003b. Implicated presuppositions. Paper presented at Polarity, Scalar Phenom‑ ena, Implicatures. University of Milan Bicocca. Smith, C. 1991. The Parameter of Aspect. Dordrecht: Kluwer. Timberlake, A. 2004. A Reference Grammar of Russian. Cambridge: CUP. Zaliznyak, A. A. & Shmelev, A. D. 2000. Vvedenie v russkuju aspektologiju. Moscow: Jazyki Russkoj Kul’tury.

Appendix 1. URLs of Internet examples, listed by example number (17) http://vesti70.ru/stats/full/?id=10202. (18) http://www.hunter.ru/read.php?f=4&i=13400&t=13393&v=f

Appendix 2. Abbreviations used in interlinear glosses det gen indet instr ipfv pfv

determinate genitive indeterminate instrumental imperfective perfective

chapter 7

Verbs of motion under negation in Modern Russian Renee Perelmutter University of Kansas

This article examines the behavior of Russian motion verbs under negation. Negated motion constructions differ from affirmative motion constructions in two respects: (1) frequency of high manner verbs, as well as verbs that specify path through prefixation; (2) aspect marking. Using detailed statistical frequencies gathered from the web as well as analysis of specific constructions, I show that negated motion events are significantly less detailed in terms of the specification of manner and certain types of path. In addition, the common assumption that imperfective is predominant under negation is not supported by the data. Aspectual choice for motion verbs depends on spatial relations between the moving figure and an observer of motion at goal or origin of the motion trajectory.

1.

Introduction

In this article, I show that lexicalization patterns of motion events in Modern Russian differ significantly depending on whether the motion events are negated or not. Modern Russian is classified as a satellite-framed and high-manner language by Talmy (1985, 2000) and Slobin (2004); verbs specifying path (in the prefix) and manner of motion (in the verb) are expected to appear frequently. Under negation, however, the frequency of high-manner and path-specifying verbs of motion (VoM) is significantly lower than in affirmative contexts. Examples (1) and (2) demonstrate relative frequencies for a high-manner verb kovyljat’ ‘to waddle’: (1)

Odnaždy teplym sirenevym utrom kovyljal po doroge utenok once warm purple morning waddle.ipfv upon road duckling Krjachik Krjachik

164 Renee Perelmutter

‘Once upon a time, on a warm purple morning, Krjachik the duckling waddled on the road.’ (M. Pljatskovskij. “Romashki v janvare”) (kovyljal ‘waddled’: 56,500 hits in google.com) On ne kovyljal po savanne, sognuvshis’ kak ego predok – he neg waddle.ipfv upon savannah bent like his ancestor avstralopitek. Australopithecus ‘He did not waddle upon the savannah, bent like his ancestor the Australopithecus.’ (L. N. Erdakov. “Čelovek v biosfere”) (ne kovyljal ‘did not waddle’: 50 hits in google.com)

(2)

Using frequency counts and data from a large untagged corpus (google.com), I ask and propose answers to the following questions: (a) what kind of motion verbs tend to appear under negation, and why? and (b) how does verbal aspect interact with lexicalization patterns under negation?

1.1

Corpus and statistical data

The data for this article was gathered using the open web accessed through Google.com, rather than a corpus of selected and edited texts, such as the Russian National Corpus (http://ruscorpora.ru). According to Meyer (2004), any artificially created and controlled corpus can provide only a “snapshot” of the speakers and writers of any particular language, even if the corpus is large and contains full texts rather than text excerpts. The Russian National Corpus (RNC) has a number of important disadvantages, which include: 1. Size: the sample of texts in RNC is significantly smaller than those available on the web, thus it does not represent the variety of usages that can be retrieved through the open web. 2. Uniformity: the selection of registers (mostly literary texts) available through the RNC limits the kind of results that will be obtained; 3. Control: since this corpus is carefully edited and balanced, new and nonstandard usages will not be reflected in RNC. The open web as a corpus has the following advantages over RNC or any other closed corpus:

. On web as corpus, see Thelwall (2005); Kilgarriff and Grefenstette (2003).

Verbs of motion under negation in Modern Russian 165

1. Size: mining the web for data is advantageous since this is the largest dataset available for natural language processing (Keller & Lapata 2003). 2. Heterogeneity of registers: the open web retrieves a wide variety of registers, including literary texts, newspaper articles, analytical/scientific writing, as well as colloquial writing genres (on-line forums and weblog articles). 3. Lack of control: since the open web is not edited, new and evolving usages will be retrieved. The open web searches also have a number of disadvantages, which have to be taken into account when analyzing the examples and compiling statistics, including the following: 1. Statistical noise: web counts contain more noise than counts obtained from a well-edited, carefully balanced corpus (Keller & Lapata 2003:â•›460) due to such factors as page repetitions and citations. The web contains replicated documents, such as mirror sites (one or more copies of a collection of pages) and multiple URLs referencing the same file. The computational linguists who view the web as content-centered suggest that for statistical purposes page duplicates should be excluded (Thelwall 2005:â•›521). 2. Statistical results are unstable: unlike data in a closed corpus, the amount and nature of texts retrieved by Google or other search engines varies greatly over time, both due to the dynamic nature of web content and to the fact that changes can be made to the index and the database of the search engine, and depending on which Google server is accessed (Keller & Lapata 2005:â•›5). Statistical processing involving the open web presents unique challenges that did not confront scholars working with traditional methodologies. Taking into account the instability of the results retrieved and the possibility of noise, to what extent can we ever trust the numbers retrieved by statistical examination of the open web? Keller and Lapata (2003) give a partial answer when they show that the results generated by Altavista and Google are highly correlated with frequencies obtained from two standard closed and edited corpora for English. Given the limitations offered by the different kinds of corpora, the importance of statistical study using the open web lies not in exact numbers themselves, but in relative frequencies, i.e. proportions of the results for different morphological realizations relative to each other. Simple statistics such as word counts, or simple frequency estimations, are useful for linguistic research by providing suggestive answers, even if the statistics are not entirely accurate (Thelwall 2005:â•›518; Meyer et al. 2003). Statistics are suggestive, for example, when an unexpected scarcity of

166 Renee Perelmutter

a certain construction is observed under negation. Some of those correlate with a certain verbal aspect. Modern Russian verbs usually distinguish two aspects, perfective and imperfective. Perfective aspect involves a temporal boundary, most often an endpoint/goal that is reached, thus prochital ‘finished reading.pf’. On the other hand, the imperfective aspect does not involve such a boundary, thus chital ‘read for a while, was engaged in reading.ipfv’. Verbs tend to be classified into aspectual pairs, perfective and imperfective; prefixed verbs of motion such as prijti-prix‑ odit’ ‘to arrive.pf.ipfv’ follow this pattern (on the category of aspect in Russian, see overviews by Forsyth 1970; Zaliznjak and Šmelev 2000; Timberlake 2004). However, simplex (unprefixed) motion verbs are deviant in that they are usually grouped not in pairs, but in groups of three: for example, the basic meaning ‘to go’ can be expressed by the perfective pojti ‘to go.pf’, determinate imperfective idti ‘to go in a single direction.det.impf’, and indeterminate imperfective xodit’ ‘to go in multiple directions. indet.impf’. One example of statistical deviance correlating with aspect is determinate imperfective under negation. For example, the determinate imperfective plyl ‘swam (towards a goal)’ accounts for 38.8% of all basic swimming verb usages in affirmative clauses; however, under negation ne plyl ‘did not swim (towards a goal)’ accounts for only 2.8% percent. I use relative frequencies rather than exact numbers to discuss specific patterns of usage of VoM in affirmative and negative contexts for different aspects of both simplex and prefixed VoM. The following verbs have been examined for statistical frequency (listed in order indet.ipfv-det.ipfv-pfv): – – – – –

xodit’-idti-pojti ‘to go’ ezdit’-exat’-poexat’ ‘to go by vehicle’ begat’-bežat’-pobežat’ ‘to run’ plavat’-plyt’-poplyt’ ‘to swim’ and prefixed verbs based on that stem letat’-letet’-poletet’ ‘to fly’ and prefixed verbs based on that stem

. The term construction is used in this article following Construction Grammar (Fillmore and Kay 1993; Goldberg 1995), in which a construction is defined a basic linguistic unit of form and meaning, where the form may be as small as a morpheme or as large as a phrase, and the meaning may include syntactic, semantic, and pragmatic elements. . Although competing theories have been advanced, asserting that VoM are not deviant in the verbal system of Russian (see Janda, this volume). . High manner verbs were chosen from Dan Slobin’s (2006) unpublished list of manner verbs in four languages Verbs of Manner of Human Motion. These verbs were chosen since they do not appear in idiomatic expressions which could skew the statistical count.

Verbs of motion under negation in Modern Russian 167

– polzat’-polzti-popolzti ‘to crawl’ and prefixed verbs based on that stem – non-basic manner verbs: melkat’ ‘to flicker’, plestis’ ‘walk dragging one’s feet’, kovyljat’ ‘waddle’ The verbs examined range from basic motion verbs such as xodit’-idti-pojti ‘to go’ to prefixed VoM obpolzti-obpolzat’ ‘crawl around [something],’ and high-manner motion verbs such as plestis’ ‘drag one’s feet’. Verbs were sampled: for basic VoM, the determinate and the indeterminate imperfective and the perfective, for prefixed VoM, the perfective and the imperfective. Data usage in this paper follows the Principles of Reuse and Enrichment of Linguistic Data as outlined in Lewis et al. (2006). While not all internet sources are attributed to specific authors, when available, author and title of work are cited immediately following an example. When neither the author nor the title are available (in case of some forum posts and short newspaper articles), the website name is cited. Full URLs of all examples gathered online are documented in Appendix 1.

2.

Negative spaces are less detailed

Negation is said to be more complex than affirmation, since every negative expression involves the supposition of its affirmative counterpart (Clark 1974; Gazdar 1979; Horn 1989; Lambrecht 2000). However, this presupposed affirmative counterpart differs from a ‘real’ affirmative construction in that it involves a lesser degree of detail for both manner and path of motion. Table 1 shows statistics for affirmative and negated motion verbs from basic (‘go’) to higher manner specifying verbs (‘waddle’). As Table 1 shows, the amounts and relative frequencies of negated VoM decrease the more manner-specific а verb is. Negation is less frequent with highmanner verbs of motion; basic manner verbs are more likely to appear under negation than manner-specific verbs. The picture is less clear for path-specifying verbs. Table 2 shows the statistical distribution for verbs specifying path through prefixation, with four stems xod'go’, let- ‘fly’, plav- ‘swim’, polz- ‘crawl’. Seven path prefixes were tested: u-, ot-, pri-, do-, pod-, pere-, ob-, and data for both imperfective and perfective aspects was collected. Note that all seven prefixal combinations are possible for manner stems examined, but not all are compatible with higher manner verbs, e.g., prikovyljal . Please note that while the data for aspectual distribution appears in Tables 1 and 2 in this section, discussion of aspect is postponed to Section 5.

168 Renee Perelmutter

Table 1.â•‡ Degrees of manner specificity Negative

Affirmative

Total

Basic motion verb ‘to go’ xodil â•⁄â•›â•›661,000 shel â•⁄â•›â•›383,000 poshel 1,100,000 all aspects 2,144,000

11.74% â•⁄ 4.94% â•⁄ 8.53% â•⁄ 8.16%

â•⁄ 4,970,000 â•⁄ 7,370,000 11,800,000 24,140,000

88.26% 95.06% 91.47% 91.84%

â•⁄ 5,631,000 â•⁄ 7,753,000 12,900,000 26,284,000

Basic manner of motion ‘to swim’ plaval â•⁄â•⁄â•›â•›32,900 plyl â•⁄â•⁄â•⁄â•⁄â•›â•›â•›982 poplyl â•⁄â•⁄â•⁄â•⁄â•›â•›â•›747 all aspects â•⁄â•⁄â•›â•›34,629

â•⁄ 6.49% â•⁄ 0.15% â•⁄ 0.13% â•⁄ 1.95%

â•⁄â•⁄â•›â•›474,000 â•⁄â•⁄â•›â•›674,000 â•⁄â•⁄â•›â•›589,000 â•⁄ 1,737,000

93.51% 99.85% 99.87% 98.05%

â•⁄â•⁄â•›â•›506,900 â•⁄â•⁄â•›â•›674,982 â•⁄â•⁄â•›â•›589,747 â•⁄ 1,771,629

High manner ‘to waddle’ kovyljal â•⁄â•⁄â•⁄â•⁄â•⁄â•⁄ 49 pokovyljal â•⁄â•⁄â•⁄â•⁄â•⁄â•⁄ 66

â•⁄ 0.09% â•⁄ 0.38%

â•⁄â•⁄â•⁄â•›56,500 â•⁄â•⁄â•⁄â•›17,500

99.91% 99.62%

â•⁄â•⁄â•⁄â•›56,549 â•⁄â•⁄â•⁄â•›17,566

High manner ‘to walk slowly’ plelsya â•⁄â•⁄â•⁄â•⁄â•⁄ 289 poplelsya â•⁄â•⁄â•⁄â•⁄â•⁄â•⁄ 96

â•⁄ 0.34% â•⁄ 0.06%

â•⁄â•⁄â•⁄â•›84,700 â•⁄â•⁄â•›â•›156,000

99.66% 99.94%

â•⁄â•⁄â•⁄â•›84,989 â•⁄â•⁄â•›â•›156,096

Table 2.â•‡ Prefixed VoM Affirmative ‘to go’ – imperfective shel ‘went.det.ipfv’ xodil ‘went.indet.ipfv’ uxodil ‘went away’ otxodil ‘went away’ doxodil ‘reached’ prixodil ‘came’ podxodil ‘came close’ matched’ perexodil ‘crossed’ obxodil ‘walked around’

Negative

Total

5,580,000

94.75%

309,000

5.25%

5,889,000

4,390,000

88.01%

598,000

11.99%

4,988,000

1,510,000

85.89%

248,000

14.11%

1,758,000

401,000

77.12%

119,000

22.88%

520,000

306,000

81.10%

71,300

18.90%

377,300

1,850,000

84.90%

329,000

15.10%

2,179,000

1,010,000

81.72%

226,000

18.28%

1,236,000

522,000

89.22%

63,100

10.78%

585,100

218,000

95.69%

9,820

4.31%

227,820

Verbs of motion under negation in Modern Russian 169

Table 2.â•‡ (continued) Affirmative ‘to go’ – perfective poshel ‘went’ ushel ‘went away’ otoshel ‘went away’ doshel ‘reached’ prishel ‘came’ podoshel ‘came close’ matched’ pereshel ‘crossed’ oboshel ‘walked around’ ‘to fly’ – imperfective letel ‘flew.det.ipfv’ letal ‘flew.indet.ipfv’ uletal ‘flew away’ otletal ‘flew away’ priletal ‘reached by flying’ podletal ‘approached flying’ doletal ‘flew close, arrived’ pereletal ‘crossed by flying’ obletal ‘flew around’ ‘to fly’ – perfective poletel ‘flew’ uletel ‘flew away’

Negative

Total

10,300,000

90.67%

1,060,000

9.33%

11,360,000

8,750,000

94.12%

547,000

5.88%

9,297,000

1,280,000

93.31%

91,700

6.69%

1,371,700

1,660,000

77.32%

487,000

22.68%

2,147,000

12,900,000

92.61%

1,030,000

7.39%

13,930,000

4,430,000

91.51%

411,000

8.49%

4,841,000

3,310,000

96.78%

110,000

3.22%

3,420,000

926,000

89.38%

110,000

10.62%

1,036,000

825,000

97.62%

20,100

2.38%

845,100

810,000

86.45%

127,000

13.55%

937,000

131,000

90.66%

13,500

9.34%

144,500

37,300

96.51%

1,350

3.49%

38,650

125,000

92.89%

9,570

7.11%

134,570

26,900

96.97%

841

3.03%

27,741

54,800

84.05%

10,400

15.95%

65,200

18,500

98.56%

271

1.44%

18,771

24,300

98.57%

353

1.43%

24,653

945,000

94.65%

53,400

5.35%

998,400

830,000

94.35%

49,700

5.65%

879,700

170 Renee Perelmutter

Table 2.â•‡ (continued) Affirmative otletel ‘flew away’ priletel ‘reached by flying’ podletel ‘approached flying’ doletel ‘flew close, arrived’ pereletel ‘crossed by flying’ obletel ‘flew around’ ‘to swim’ – imperfective plaval ‘swam.indet.ipfv’ plyl ‘swam.det.ipfv’ uplyval ‘swam away’ otplyval ‘swam away’ priplyval ‘reached by swimming’ podplyval ‘approached swimming’ doplyval ‘swam close, arrived’ pereplyval ‘crossed by swimming’ obplyval ‘swam around’ ‘to swim’ – perfective poplyl ‘started swimming’ uplyl ‘swam away’ otplyl ‘swam away’ priplyl ‘reached by swimming’ podplyl ‘approached swimming’

Negative

Total

234,000

98.60%

3,320

1.40%

237,320

1,100,000

96.67%

37,900

3.33%

1,137,900

147,000

99.45%

811

0.55%

147,811

248,000

80.36%

60,600

19.64%

308,600

121,000

98.19%

2,230

1.81%

123,230

84,700

98.27%

1,490

1.73%

86,190

482,000

93.45%

33,800

6.55%

515,800

438,000

98.01%

8,900

1.99%

446,900

27,400

97.34%

749

2.66%

28,149

12,200

94.21%

750

5.79%

12,950

7,480

94.92%

400

5.08%

7,880

17,700

97.19%

512

2.81%

18,212

4,880

94.68%

274

5.32%

5,154

24,900

97.51%

636

2.49%

25,536

423

99.76%

1

0.24%

424

399,000

98.23%

7,200

1.77%

406,200

163,000

94.85%

8,850

5.15%

171,850

94,800

98.95%

1,010

1.05%

95,810

143,000

97.27%

4,020

2.73%

147,020

74,800

99.03%

736

0.97%

75,536

Verbs of motion under negation in Modern Russian 171

Table 2.â•‡ (continued) Affirmative doplyl ‘swam close, arrived’ pereplyl ‘crossed by swimming’ obplyl ‘swam around’ ‘to crawl’ – imperfective polzal ‘crawled.indet.ipfv’ polz ‘crawled.det.ipfv’ upolzal ‘crawled away’ otpolzal ‘crawled away’ pripolzal ‘reached by crawling’ podpolzal ‘approached crawling’ dopolzal ‘crawled close, arrived’ perepolzal ‘crossed by crawling’ obpolzal ‘crawled around’ ‘to crawl’ – perfective popolz ‘started crawling.pfv’ upolz ‘crawled away’ otpolz ‘crawled away’ pripolz ‘reached by crawling’ podpolz ‘approached crawling’ dopolz ‘crawled close, arrived’ perepolz ‘crossed by crawling’ obpolz ‘crawled around’

Negative

Total

50,600

84.98%

8,940

15.02%

59,540

69,100

98.27%

1220

1.73%

70,320

633

99.84%

1

0.16%

634

201,000

92.58%

16,100

7.42%

217,100

285,000

98.66%

3,880

1.34%

288,880

13,800

93.32%

988

6.68%

14,788

11,800

91.83%

1,050

8.17%

12,850

7,100

96.45%

261

3.55%

7,361

16,000

97.91%

342

2.09%

16,342

4,430

86.51%

691

13.49%

5,121

10,300

98.35%

173

1.65%

10,473

1240

99.92%

1

0.08%

1,241

346,000

99.30%

2,450

0.70%

348,450

163,000

97.85%

3,580

2.15%

166,580

61,500

99.25%

464

0.75%

61,964

76,100

97.55%

1910

2.45%

78,010

156,000

99.64%

556

0.36%

156,556

118,000

92.32%

9,810

7.68%

127,810

39,000

98.24%

698

1.76%

39,698

704

99.86%

1

0.14%

705

172 Renee Perelmutter

‘came waddling’ is possible, but *ukovyljal ‘went away waddling’ is not. Verbs are bolded if the percentage of hits under negation is comparatively high; note that these verbs are usually prefixed with do- ‘until’ which involves goal-reaching. As can be seen from Table 2, in general nonprefixed VoM are more frequent than prefixed ones under negation. Among the VoM specifying path through prefixation, negated verbs which mark the goal or the origin of motion (with prefixes such as u-, ot-, pri-, do-) are more frequent than verbs with a specification of path which does not involve goal or origin (i.e., prefixes such as pere-, ob-). As can be seen in Table 1, high-manner stems rarely combine with prefixes of path; prefixes that do appear with high-manner stems tend to specify goal of motion rather than origin or goal-less path. In conclusion, statistical data shows that (a) negated motion tends to be less specific regarding manner than affirmative motion; (b) when path is specified, negated motion tends to specify path towards a goal or less frequently, origin.

3.

Prefixation and motion trajectory in negated VoM

Prefixed VoM can specify goal, origin, or movement along a trajectory that involves neither goal nor origin. In this section I will show that the specification of goal is most frequent under negation, and that the presence of an observing entity closely correlates with negated motion events.

3.1

Trajectory involving goal of motion

Motion events specifying goal (involving verbs with prefixes do-, pri, pod-) are frequent in affirmative clauses as well as in negative clauses. However, affirmation differs from negation in the type of goal and the level of detail in which the goal is described. In affirmative clauses, the goal of motion can be either a physical locus or a person in a locus. In negated clauses, the goal is predominantly a person in locus. This goal is not reached (hence the negation), but it is set up with an expectation to be reached; there is a person in the locus who is waiting for the arrival of the moving figure. The person in the locus has the semantic role of observer (Padučeva 1992, 1997, 2004). In this section, I will show that the presence of observer is crucial for understanding negated motion verb constructions, but not as important for affirmative clauses.

Verbs of motion under negation in Modern Russian 173

3.1.1 Locus in affirmative clauses When a physical locus is involved, prefixed motion often occurs with detailed description of this locus. Thus, in (3), the speaker describes the goal of motion (his dacha) in great detail: (3)

Byla rannjaja vesna, kogda ja priexal na dachu, i na dorožkax was early spring when I arrived.pfv to dacha and on paths eše ležal prošlogodnij temnyj list. […] ja odin brodil sredi still lay.ipfv last year’s dark leaf I alone wandered among pustyx dač, otražavšix steklami aprelskoe solnce empty dachas reflected.gen.pl glass.instr.pl. April.adj sun ‘It was early spring when I arrived to the dacha, and the dark leaves of last year still lay on the paths. I wandered alone between the empty dachas, which reflected the April sun in the glass of their windows.’ (L. Andreev. “Na stancii”)

Similarly in (4), the speaker plans to spend the night in Umhausen, which is the goal of motion. Unlike in (3), where the locus itself becomes the focus of the narrative, Umhausen of (4) is important because an event takes place there – the camping ground is full. (4)

Ja dovolno bystro doexal do gorodka Umhausen, I â•›rather fast reached.pfv to little.town Umhausen no v kempinge, v kotorom ja planiroval zanočevat’, but in camping-site in which I planned.ipfv spend.night.inf.pfv mne skazali, chto on zakryt. me.dat said.3pl that it closed.adj ‘I reached Umhausen rather quickly, but at the camping site where I planned to spend the night I was told that it was closed, and I had to move on.’ (E. A. Luchin. “S Vizborom v Alpax”)

Though there are people at the locus, they do not have an important part to play in the narrative (thus the impersonal reference). Goal-oriented motion towards a person at a certain locus is possible with affirmative clauses, thus in (5), the commander arrives at the battalion in order to speak with the soldiers (goal of motion). The speaker is among these soldiers:

174 Renee Perelmutter

(5)

Čerez neskolko dnej komandir divizii priletel k nam v after few days commander division.gen flew.pfv to us into polk. On proizvel razbor našix dejstvij. batallion he accomplished.pfv analysis our actions ‘After a number of days the division’s commander arrived (by flying) to our battalion. He analyzed our actions.’ (A. G. Litvin. “Vykhod iz mertvogo prostranstva”)

Thus for affirmative motion events, movement towards the goal may involve: (a) focalized physical locus as goal, described in detail; (b) physical locus as goal, which is important because of events occurring in that locus; (c) a person as goal.

3.1.2 Locus in negated clauses Since the moving figure does not arrive at the physical destination, it is highly unlikely that the locus will be focalized. As a rule, loci of negated clauses specifying goal are generic and do not involve detailing. The lush description of the dacha in (3) would be impossible under negation. Frequent loci of negated motion events include well-established locations such as domoj ‘(to) home’, sjuda ‘hither’, na rabotu ‘to work’, et cetera. In (6), the locus is simply stated as the city of Riga. (6)

Tolja ne priexal v uslovlennyj den’ v Rigu. Tolja neg come.pfv in appointed day in Riga Ja zabespokoilas’, pozvonila dočeri i zjatju. I became.worried.pfv called.pfv daughter and son.in.law ‘Tolya did not arrive on the appointed day in Riga. I became worried, called my daughter and my son-in-law.’ (O. Bulkina. “Kak oni umerli”)

For negated motion events, physical locus as goal of negated motion usually involves a person expecting the arrival of the moving figure, as in (6), where the speaker expects Tolya. The speaker has the semantic role of an observer, or a perceiving entity (a person, often the speaker of the utterance) who expects the appearance of the moving figure in the locus which he/she occupies. The notion of observer as a semantic category was introduced and developed by Padučeva (1992, 1997, 2004), and has been used by Borschev and Partee (2002) and Perelmutter (2005) among others to discuss the distribution of genitive versus nominative case under negation. The statistically prevalent genitive of negation (mamy ne bylo na rabote ‘mother-gen wasn’t at work’) involves, according to Padučeva, a perceiving entity/agent who is expecting the appearance of the absentee in the locus. The notion of ‘observation’ is similar to ‘expectation,’ a regular feature of negated clauses according to Tottie (1991), Horn (1989), and others. The presence of an observing entity is more prominent for negated motion, since

Verbs of motion under negation in Modern Russian 175

such an observer occupies the locus which the moving figure fails to reach; the observer is often the goal of motion. Thus in a joke (7), a hunter is expecting the arrival of a a sloth at his locale: (7)

Sižu ja v zasade, – načal oxotnik, – polzet ko mne lenivec… sit.1sg I in blind began hunter crawls to me.dat sloth Den’ polzet, dva polzet, tri polzet… Tak i ne dopolz do zasady! day crawls two crawls three crawls so and neg crawl.pfv to blind ‘I am sitting in a blind, the hunter began, and a sloth is crawling towards me… He crawls for a day, he crawls for two, he crawls for three… He never did reach (by crawling) the blind!’ (“Anekdoty pro oxotnikov”)

In (8), a moving object (tomato) thrown by other girls routinely does not reach the observer, Maria, who therefore has no reason to participate in the ensuing fights: (8)

Marija ne učastvovala v potasovkax, možet byt’ potomu, Maria neg participated.ipfv in skirmishes may be.inf because čto pomidor nikogda ne doletal do ee uglovoj kojki. that tomato never neg flew.pfv until her corner bed ‘Maria did not participate in the skirmishes, maybe because the tomato never reached (by flying) her corner bed.’ (N. Čertkova. “Mjagkij mir”)

The moving object’s failure to materialize in the locus impacts other events in the narrative, usually influencing the observer’s later actions. Thus in (8), Maria does not participate in fights because she hasn’t been hit by a flying tomato. In (9), a failure of a rock star to appear on a concert enrages fans, who subsequently have to be pacified by the police: (9)

Jurij Antonov ne prišel… na sobstvennyj concert Jurij Antonov neg come.pfv to self concert vo MXAT imeni Čexova, into Moscow-Artistic-Academic-Theatre name.gen Chekhov.gen čem vyzval vozmuščenie poklonnikov by which provoked.pfv outrage fans.gen ‘Yuri Antonov did not show up to his own concert in Chekhov’s MXAT (Moscow Art Academic Theatre), by which he provoked the outrage of his fans.’ (Novosti NEWSru.com)

The failure of Antonov to appear provokes an emotional reaction from the observers. I show in Perelmutter (2005) that such an emotional response to unexpected absence is frequent also with the genitive of negation, as in (10), where regret connected with the absence is expressed:

176 Renee Perelmutter

(10)

Žaleju, čto v oktjabre 1993 goda menja ne bylo v Moskve sorry.1sg that in October 1993 year me.gen neg was in Moscow i sredi zaščitnikov belogo doma. and among defenders.gen white.acc house.acc ‘I am sorry that in October 1993 I wasn’t in Moscow, among the defenders of the White House.’ (BBC News forum)

Negated be-clauses with the absentee in genitive and negated motion events are comparable, since both involve absence – failure to reach the locus – which is observed by somebody in the locus. Absence, expressed through genitive of negation or through negated motion towards a goal, can impact other events and provoke emotions in the observer/narrator. While both affirmative and negated clauses allow a person to be the goal of motion, the person as goal in affirmative clauses is not an observer: he/she is not, as a rule, majorly impacted by the arrival of the moving figure, and there is usually no emotional reaction. In narratives, arrival of a person at goal (be it a physical locus or a person) is often just one event in a sequence of other events. On the other hand, failure of the moving figure to arrive at goal is often focalized in narratives due to the fact that (1) other events in the narrative are impacted by the absence, and (2) observer’s feelings about the absence are made known. In this respect, negated motion events with observer at goal are semantically similar to negated be-clauses with genitive of negation.

3.2

Trajectory that specifies origin

As can be seen from Table 2, prefixes of origin (u-, ot-) are rarer under negation than the prefixes of goal; in addition, their distribution under negation depends on aspect (see Section 5 for discussion). Perfective that marks origin usually involves a successful departure from the observer’s locus. Negated or failed departure can provoke emotional reactions from the observer(s), as in example (11), where the tourist’s failure to board the plane provokes feelings of surprise from the travel agency workers: (11)

ja poexal v J.K. Tour Travels. Rabotniki etoj firmy I went.by.driving.pfv in J.K. Tour Travels employees this.gen agency.gen byli očen’ udivleny, počemu ja ne u-letel, polučil were very surprised why I neg away-flew.pfv received.pfv ot nix kuču izvinenij. from them heap excuses.gen

Verbs of motion under negation in Modern Russian 177

‘I went to J. K.Tour Travels. The agency’s employees were very surprised that I hadn’t flown away, and I received many apologies from them.’ (turizm.ru forum, “Ulet v Goa”)

Most negated VoM specifying origin of departure occur in temporal clauses where the act of motion does indeed take place, as in (12), where a grandmother is experiencing intense emotions while sending her grandchildren away: (12)

Babuška dolgo stojala na kraju šljuza, poka paroxod grandmother long stood on edge dock.gen until steamer ne ot-plyl tak daleko ot berega, čto passažirov, neg from-sailed.pfv so far from shore that passengers.gen.pl stojavšix na palube, stalo ne vidno standing on deck began neg seen ‘Grandmother waited for a long time on the edge of the dock, until the steamer sailed so far away from the shore that one could no longer see the passengers standing on the deck.’ (O. Uvarkina. “Suxar’”)

Even though the construction here employs negation, the motion away from the observer does indeed take place; this aspectual idiom does not universally appear with negation (cf., English ‘until it sailed away’).

3.3

Motion trajectory specifying neither goal nor origin

In Russian, some prefixes of path specify neither goal nor origin, but rather give additional detail about the trajectory. Some prefixes include ob-‘around’ and pere‘across’. Verbs with these prefixes are rare under negation. The infrequent negative clauses with these verbs usually involve a pre-planned trajectory that for some reason does not take place. In (13), the only example of ob-‘around’ with plyt’ ‘to swim’, the speaker dreams about swimming around the Adalary rocks, but this does not come to pass. Note the affirmative ja objazan ix obplyt’ ‘I have to swim around them’ in the following clause: (13)

Ja ne ob-plyl Adalary (eto byla moja mečta s togo I neg around-swam.pfv Adalary this was my dream from that momenta kak ja ix vpervye uvidel ja rešil čto prosto moment that I them first saw.pfv I decided that simply objazan ix ob-plyt’). obligated them.acc around-swim.inf.pfv ‘I did not swim around the Adalary rocks (it was my dream from the moment I saw them, I decided that I really have to swim around them).’ (Artek.org forums)

178 Renee Perelmutter

Similarly in (14), we find denial of trajectory with the prefix pere- ‘over’ that was previously suggested in an affirmative clause: (14)

Pere-bros’ ego v ogorod, – posovetovala tetka. over-throw.pfv him into vegetable.patch advised auntie Ja podošel k izgorodi i švyrnul ego okamenevšimi rukami. I approached to fence and threw him stony.instr.pl hands.instr.pl Prokljatie! On, konečno, ne pere-letel čerez zabor, a uselsja curse he certainly neg over-fly.pfv over fence but sat.refl.pfv na nego, rasplastav tjaželye krylja. on it having.spread heavy wings ‘Throw [the rooster] over to the vegetable patch, – the woman advised. I approached the fence and threw him with stony hands. Damnation! He didn’t, of course, fly over the fence but perched upon it, spreading [his] heavy wings.’ (F. Iskander. “Petux”)

Since appearance of negated motion events closely correlates with the appearance of observer at a boundary of motion (either arrival or departure), it is not surprising that prefixed verbs that specify movement along the path and do not involve an observer don’t often appear under negation. To summarize, path-specifying VoM appear under negation most often when they encode movement towards a goal, which is either the observer or a location where the observer expects the moving figure to appear. VoM specifying origin rather than goal are less prominent, and VoM specifying trajectory which does not involve either goal or origin are rare.

4.

Aspect and VoM under negation

Isačenko (1962) among other scholars points out that the imperfective is preferred under negation. Merrill (1985:â•›130) remarks that “even a relatively cursory examination shows that the imperfective is relatively more frequent than it is in positive declarative contexts.” Forsyth writes that both perfective and imperfective can occur under negation, but “there is a certain tendency to switch to the imperfective in negative statements in past tense” (Forsyth 1970:â•›103). According to Forsyth, perfective under negation is a non-performance of a potential single action, whereby an expected or desired result is not produced: my . My own statistical examination of verbs under negation shows no pronounced preference for either aspect with ‘base’ verbs; there are idiosyncratic preferences depending on lexicalization patterns, as will be shown below.

Verbs of motion under negation in Modern Russian 179

ne pročitali etu knigu ‘we did not read this book’, where the reading was planned but for some reason was not accomplished. The negated imperfective, on the other hand, indicates an action that has never taken place at any time: my ne čitali etu knigu ‘we did not read this book’, where the action was not planned. Imperfective is used with negation that is “general, comprehensive and vehement” (Forsyth 1970:â•›117). A number of scholars propose that aspectual choice involves the same or similar contextual considerations for affirmative and negative statements (e.g., Šigurov 1993; Timberlake 2004). Timberlake (2004:â•›418) points out that aspectual choice “revolves around the way in which the speaker conceptualizes the possible occasions for an event.” The perfective places a bounded event, whether it had occurred (positive clause) or not (negated clause) into a sequence of events. The imperfective allows for further change. According to Forsyth, VoM follow the general pattern for negation (‘kinetic’ presentation of the perfective vs. general denial of imperfective), but the possibility of using the indeterminate imperfective “provides a third stage further removed from the reality of performance” (Forsyth 1970:â•›339). The negated per‑ fective denotes non-performance of the action at a specific juncture: a departure which a subject does not make at a specific time: (15)

On ešče ne poexal v Leningrad. Ego poezdka namečena tol’ko na he yet neg go.pf to Leningrad his trip planned only on buduščij mesjac. next month ‘He hasn’t gone to Leningrad yet. His trip is planned only for next month.’ (Forsyth 1970:â•›340)

The determinate imperfective implies an absence of motion at a given moment, or absence of the tendency to perform it: (16) Zdes’ ljudi ne šli, a stojali. here people neg walk.det.ipfv but stood.ipfv ‘Here people did not walk, but stood.’

(Forsyth 1970:â•›340)

The indeterminate imperfective denotes negation of any of the meanings associated with these verbs, such as general motion in unspecified direction or a habitual motion. It implies a total denial of a hypothetical journey: (17) V voskresenje ja nikuda ne ezdil. on Sunday I nowhere neg went.indet.ipfv ‘On Sunday I did not go anywhere.’

(Forsyth 1970:â•›341)

180 Renee Perelmutter

How do these aspectual differences play out in the statistical distribution? This depends on whether the verb of motion is prefixed or simplex.

4.1

Prefixed VoM

For prefixed verbs of motion, the perfective is significantly more frequent than imperfective for all manner verbs examined (fly, swim, crawl), and especially for those specifying the goal of motion. Table 3 showcases the breakdown for three prefixed verbs: ‘to fly’, ‘to swim’, and ‘to crawl’ (for additional data see Table 2). The perfective is more frequent with manner verbs specifying goal or origin. The perfective reports an expected arrival or departure of the moving figure in the observer’s locus; predictably, the failed arrival/departure triggers further events or emotional reactions. Thus in (18), the absence triggers an event (consumption Table 3.â•‡ Prefixed VoM under negation Verb.ipfv-pf

Imperfective

Perfective

‘to fly’ uletal-uletel ‘flew away’ otletal-otletel ‘flew away from’ priletal-priletel ‘flew to’ podletal-podletel ‘flew near’ doletal-doletel ‘reached by flying’ pereletal-pereletel ‘flew across or over’ obletal-obletel ‘flew around’

13,500 â•⁄ 1,350 â•⁄ 9,570 â•⁄â•⁄â•›â•›841 10,400 â•⁄â•⁄â•›â•›271 â•⁄â•⁄â•›â•›353

21.36% 28.91% 20.16% 50.91% 14.65% 10.84% 19.15%

49,700 â•⁄ 3,320 37,900 â•⁄â•⁄â•›â•›811 60,600 â•⁄ 2,230 â•⁄ 1,490

78.64% 71.09% 79.84% 49.09% 85.35% 89.16% 80.85%

‘to swim’ uplyval-uplyl ‘swam away’ otplyval-otplyl ‘swam away from’ priplyval-priplyl ‘swam to’ podplyval-podplyl ‘swam near’ doplyval-doplyl ‘reached by swimming’ pereplyval-pereplyl ‘swam over’ obplyval-obplyl ‘swam around’

â•⁄â•⁄â•›â•›749 â•⁄â•⁄â•›â•›750 â•⁄â•⁄â•›â•›400 â•⁄â•⁄â•›â•›512 â•⁄â•⁄â•›â•›274 â•⁄â•⁄â•›â•›636 â•⁄â•⁄â•⁄â•⁄â•›â•›1

â•⁄ 7.80% 42.61% â•⁄ 9.05% 41.03% â•⁄ 2.97% 34.27% 50.00%

â•⁄ 8,850 â•⁄ 1,010 â•⁄ 4,020 â•⁄â•⁄â•›â•›736 â•⁄ 8,940 â•⁄ 1,220 â•⁄â•⁄â•⁄â•⁄â•›1

92.20% 57.39% 90.95% 58.97% 97.03% 65.73% 50.00%

‘to crawl’ upolzal-upolz ‘crawled away’ otpolzal-otpolz ‘crawled away from’ pripolzal-pripolz ‘crawled to’ podpolzal-podpolz ‘crawled near’ dopolzal-dopolz ‘reached by swimming’ perepolzal-perepolz ‘crawled across-over’ obpolzal-obpolz ‘crawled around’

â•⁄â•⁄â•›â•›988 â•⁄ 1,050 â•⁄â•⁄â•›â•›261 â•⁄â•⁄â•›â•›342 â•⁄â•⁄â•›â•›691 â•⁄â•⁄â•›â•›173 â•⁄â•⁄â•⁄â•⁄â•›â•›1

21.63% 69.35% 12.02% 38.08% â•⁄ 6.58% 19.86% 50.00%

â•⁄ 3,580 â•⁄â•⁄â•›â•›464 â•⁄ 1,910 â•⁄â•⁄â•›â•›556 â•⁄ 9,810 â•⁄â•⁄â•›â•›698 â•⁄â•⁄â•⁄â•⁄â•›â•›1

78.37% 30.65% 87.98% 61.92% 93.42% 80.14% 50.00%

Verbs of motion under negation in Modern Russian 181

of alcohol). The observer (Ljoša) is expecting the arrival of the figure (Saša) and tries to save some vodka for him. However, the vodka is drunk by others because of Saša’s absence and despite Lyoša’s protests: (18)

Ljoša pereodevaetsja, dostaet butylku i vypivaet bol’she Ljoša changes.clothes.refl takes.out bottle and drinks.pfv more poloviny. Narod ego javno ne odobrjaet, no ostatok half people.sg him clearly neg approve.sg but rest do-pivajet, nesmotrja na to, čto Ljoša gromko kričit pro completely-drinks.pfv despite on that that Ljoša loudly yells about Sašu, kotoryj ešče ne pri-plyl, i kotoromu vodka Saša that yet neg arrived-sailed.pfv and which.dat vodka nuzhnee. more.needed ‘Ljoša … changes his clothes, takes out a bottle and drinks more than half. People are not approving, but finish the bottle despite Ljoša’s loud protests about Saša, who had not arrived here (by sailing/swimming) yet and who needs the vodka more.’ (A. Savvateev. “Polomet’ – “El’ba” – tri otcheta”)

Example (19) showcases an example of negated departure from a locus originally shared by the figure and two observers. In this example, the two observers discuss the figure (a secret agent nicknamed “Professor”). One observer believes that Professor flew away on the plane, as scheduled, while the other explains how this secret agent tricked the surveillance into believing he flew away, while in reality the planned departure did not occur. Due to the presence of two observers who hold opposing opinions about the departure, we find both the expected affirmative event (“Professor departed”) and the expected negated event (“Professor didn’t fly away”): (19)

Professor ne u-letel. […] A vam Artist s čistoj professor neg away-flew.pfv but you.dat Performer with clean sovest’ju doložil, čto ob’’ekt nabljudenija otbyl v Moskvu. conscience reported that object surveillance.gen departed for Moscow. Vy xotite sprosit’, otkuda ja eto znaju? Skažu. Ja sam za nim You want ask.inf from.where I this know Tell.1sg.pf I self after him sledil. Ja ne veril, čto Professor u-letit. followed I neg believed.ipfv that Professor away-fly.pfv ‘The Professor did not fly away. And Artist told you in all honesty that the object of surveilance left for Moscow. You want to know how I know this? I will tell you. I followed him myself. I did not believe that he would fly away.’ (V. Levashov. “Ubit’ Demokrata”)

182 Renee Perelmutter

So far in examples (18) and (19), we saw that perfective is used with failed arrival/departure when an observer is present at the locus to remark on the failure of the figure to arrive or depart. Perfective can also be used when the figure does arrive/depart, but fails to arrive or depart by predictable means. This is exemplified by (20), a narrative interval in which a man is supposed to be killed by a charging rhinoceros. At the last moment, the man escapes by magical means. The figure’s movement away from the locus which is shared with another being (in this case, the enraged rhinoceros) does happen, however, the escape does not occur through the expected modes of motion such as crawling or running. (20)

Zver’ nastig ego i […] vonzil svoi roga. No […] roga beast caught.up him and stuck self horns but horns nosoroga votknulis’ ne v čeloveka, a v glinu, v dno. rhinoceros.gen.sg struck neg in man but in clay in bottom I vovse ne potomu, chto on promahnulsja, a potomu, chto and at.all neg because that he missed.refl but because that čeloveka na etom meste vdrug ne okazalos’ Net, on man.gen on this place suddenly neg located.refl no he ne ot-prygnul v storonu, ne ot-polz i neg away-jumped.pfv to side, neg away-crawl.pfv and ne ot-bezhal. On prosto-naprosto ischez. neg away-run.pfv He simply-simply disappeared ‘The beast reached him and impaled the man. But the rhinoceros’s horns stuck not the man but the mud at the bottom. And not because the rhinoceros missed, but because the man suddenly wasn’t there. No, he did not jump away, did not crawl away and did not run away. He simply disappeared.’ (G. Razumov. “Kosmičeskij majak”)

To summarize, negated motion verbs appear in perfective when an expectation of motion is thwarted – either when the motion itself does not take place, or when the motion happens in the manner unexpected by the observer. The imperfective of prefixed motion verbs is different from the perfective in that the motion might indeed have happened at some point in time, and that the lack of motion at this specific temporal juncture may become an attribute of the whole situation, rather than an event. As a rule, an observer is not involved in this situation. Imperfective negated motion often appears with specifications of time, such as davno ‘for a while’, pjat’ let ‘five years’, etc. In (21), a woman complains that her husband has not come home for a while. This is canonical use of the imperfective, i.e. a situation of absence which continues at the moment of speech; and a canonical use of the negated motion event as described above: the failure of the moving

Verbs of motion under negation in Modern Russian 183

figure to arrive at the woman’s locus is combined with an emotional reaction. However, the ongoing absence of the husband in the wife’s locus is an attribute of her daily life rather than a sudden event of absence when motion is expected: (21)

Inoj raz požaluetsja žitel’nica poselka sosedke, other time complain.refl.pfv dweller village.gen neighbor.dat čto muž davno ne pri-ezžal domoj a deti that husband long-time neg arrived-drove.ipfv home and children rastut, i trudno ej odnoj upravljat’sja s nimi. grow and hard her.dat alone.dat manage.refl with them ‘At times a village woman would complain to her neighbor that her husband didn’t come home in a long time, and the children are growing and she is having hard time taking care of them.’ (Russia-today.ru news)

In example (22), a negated motion event with davno ‘for a while’ has a slightly different meaning. Motion does take place, but after a significant temporal interval: (22)

V takie “zavedenija” davno ne pri-ezžal into such establishments long.time neg arrived-drove.ipfv i, kak pokazala praktika, lučše by voobšče tam ne and as showed.pfv practice better would at.all there neg pokazyvalsja showed.up.refl.ipfv ‘I haven’t arrived at such establishments for a long time, and as practice showed, I would have been better off not going there at all.’ (“Prodavcam ne nužny den’gi”)

In (23) and (24), the figure does indeed arrive at the locus, and the negated motion refers to the time lapsed since the last visit. The fact that the negated motion does in fact take place accounts for the extremely detailed description of the locus – something which does not happen when the negated motion does not take place (as discussed in Section 4). (23) Davno ne pri-letal v Šeremet’evo-2, no tam ničego long.time neg arrived-flew.ipfv into Šeremet’evo-2 but there nothing ne menjaetsja smotriš’ na vse, kak v sovetskom kino neg changes.refl look.2sg.ipfv on everything like in Soviet movie […] bol’šie priemnye, patriarxal’no-otečeskie, bezlično-zabotlivye, big reception.rooms patriarchal-fatherly faceless-caring surovo-spravedlivye vorota Sovka. I daže vrode kislymi strict-just gates Sovok.gen and even seems sour.instr ščami kak-to potjanulo cabbage.soup.instr somehow wafted.pfv

184 Renee Perelmutter

‘I haven’t flown into Sheremet’yevo-2 airport for a while, but nothing changes there: you are looking at everything like in a Soviet movie – big receptionrooms, patriarchal, faceless but caring, strict but just gates of the Soviet Union. And even the smell of sour cabbage soup wafted from somewhere.’ (turizt.livejournal.com)

The negated motion here describes the properties of the absentee (he was absent from Russia/Moscow for a long time). Such an imperfective negated motion event is functionally similar to the nominative of absence construction ne byl ‘wasn’t’Â€+ nom, such as mama ne byla na rabote ‘mother.nom wasn’t at work’, as opposed to the genitive of absence mamy ne bylo na rabote ‘mother.gen wasn’t at work’ (Perelmutter 2005). In the nominative of absence construction, the focus is on the properties of the absent individual, and specifications of a temporal interval, such as davno ‘for a long time’ are frequent. Example (24) shows an example of nominative with the ne byl ‘wasn’t’ + nom construction. Here, a young dentist returns to Moscow after a 1.5 year absence, and describes his impressions of the locus: (24)

Rejsom Los-Andželes-Moskva pri-leteli neskol’ko čelovek […] flight.instr Los-Angeles-Moscow arrived-flew.pfv few people Vrač Artur – v tom čisle. On ne byl v Moskve poltora goda: doctor Artur in that number. he.nom neg was in Moscow 1.5 years “Očen’ zdes’ veselo. Prosto neverojatno. Vse krasivye, very here joyful simply unbelievable all beautiful.pl ulybajutsja, u vseh zuby xorošie”. smile.pl.refl to everybody teeth good ‘A couple of people flew in on the Los-Angeles-Moscow flight – and immediately (went) to the Vogue Café. Arthur, the doctor, was among them. He wasn’t in Moscow for a year and a half: “It’s very happy here. It’s incredible. Everybody is beautiful, smiling, and they all have good teeth”.’ (E. Egereva. “Glamorama”)

The two constructions in (23) and (24) are similar. Both the negated imperfective motion event and the negated be-event include a person who didn’t visit a locus for a long time, finally arrives at this locus, and gives a detailed description of the locus. The focus on absentee and his/her experiences and properties often coincides with the first person viewpoint, where the moving figure and the observing entity are one and the same, as in (23) and (24). However, first person is not a requirement. In the third person account of (25), the absentee’s arrival after a long temporal interval emphasizes the importance of his arrival for the observers. Negated imperfective is used here to indicate that the figure arrived after a long absence:

Verbs of motion under negation in Modern Russian 185

(25)

Neskol’ko let ne pri-ezžal v Kursk [… ] professor iz few years neg arrived-drove.ipfv to Kursk professor from Kolomny Aleksandr Auer. Teper’ on poznakomil kolleg so Kolomna Alexandr Auer now he acquainted colleagues.acc with svoimi naučnymi poiskami poslednego self.instr.pl scientific.instr.pl searches.instr.pl recent.gen.sg vremeni. time.gen.sg ‘Alexandr Auer, a professor from Kolomna and a regular participant of the Fet readings, did not come to Kursk for a number of years. Now he shared his most recent research with his colleagues.’ (T. Antipenko. “Fetovskie čtenija”)

Both the imperfective and the perfective aspects can appear in one sentence, as in (26), where the observer expects the appearance of a young boy, his friend: after the thwarted expectation of the perfective ne prišel ‘did not come’, the observer continues to wait. The appearance of the imperfective celuju nedelju ne prixodil ‘did not come all week’ signals that the boy did come after all. Note the emotional reaction of the observer: (26)

Rovno nedelju nazad, malyš ne pri-šel na alleju, exactly week ago boy neg arrived-came.pfv to boulevard ne pri-šel i na sledujuščij den’, i vsju nedelju neg arrived-came.pfv and on next day and all week ne pri-xodil. Nikogda ne dumal, čto budu tak. neg arrived-came.ipfv never neg thought.ipfv that will.1sg so pereživat’ i ždat’ ego worry and wait him ‘Exactly a week ago, the boy did not come to the boulevard, did not come the next day, did not come all week. I never thought that I would worry so much and wait for him.’ (“Monetki”)

To summarize, the perfective of prefixed VoM indicating goal and origin appears in situations of thwarted expectation where an observer is usually present; the imperfective appears in situations of sustained non-arrival in the locus, where the non-arrival becomes a property of the situation rather than a singular event; or in situations where a long non-arrival implies eventual arrival, and where the previous non-arrival is a property of an absentee. Thus, the negated motion events described in the imperfective often involve motion events that actually do take place. The constructions involving negated prefixed VoM of goal are semantically similar to the genitive and nominative absence constructions with byt’ ‘to be’. Negated perfective is similar to genitive of negation in that it involves thwarted expectations, observers, and an impact of absence on the observer, such as an

186 Renee Perelmutter

emotional reaction; the negated imperfective is similar to nominative of negation in that it focuses on the absentee’s properties and often includes a temporal specification.

4.2

Simplex VoM

Table 4 tests five VoM for aspectual distribution: xodit’-idti-pojti ‘to go’, ezdit’exat’-poexat’ ‘to go by vehicle’, letat’-letet’-poletet’ ‘to fly’, plavat’-plyt’-poplyt’ ‘to swim’, polzat’-polzti-popolzti ‘to crawl.’ For the five VoM examined, in affirmative clauses indeterminate imperfective, determinate imperfective and determinate perfective are more or less equally distributed, with a slight preference towards the perfective for three out of five verbs (‘to go’, ‘to fly’, and ‘to crawl’) or towards the indeterminate imperfective for two verbs (‘to go by vehicle’, ‘to swim’). Under negation, indeterminate imperfective is significantly more prominent for manner verbs ‘to swim’, ‘to fly’, ‘to crawl’. Determinate imperfective is significantly less prominent (under 20%) for all five VoM examined. Indeterminate imperfectives of simplex VoM do not involve an observer, and usually signal “a total denial of a hypothetical journey” (Forsyth 1970:â•›341)Â€– however, it is not a journey that is denied (which would imply an intent or a destination), but the fact that a motion of this kind takes place in a specific time period: Table 4.â•‡ Basic VoM under negation – aspect Affirmative

Negative

xodil ‘went.indet.ipfv’ šel ‘went.det.ipfv’ pošel ‘went.pfv’

â•⁄ 4,390,000 â•⁄ 5,580,000 10,300,000

21.66% 27.53% 50.81%

â•⁄â•›598,000 â•⁄â•›309,000 1,060,000

30.40% 15.71% 53.89%

ezdil ‘went by vehicle.indet.ipfv’ exal ‘went by vehicle.det.ipfv’ poexal ‘went by vehicle.pfv’

â•⁄ 3,240,000 â•⁄ 2,600,000 â•⁄ 3,360,000

35.22% 28.26% 36.52%

â•⁄â•›320,000 â•⁄â•⁄â•›59,900 â•⁄â•›346,000

44.08% â•⁄ 8.25% 47.66%

letal ‘flew.indet.ipfv’ letel ‘flew.det.ipfv’ poletel ‘flew.pfv’

â•⁄â•⁄â•›â•›810,000 â•⁄â•⁄â•›â•›825,000 â•⁄â•⁄â•›â•›945,000

31.40% 31.98% 36.63%

â•⁄â•›127,000 â•⁄â•⁄â•›20,100 â•⁄â•⁄â•›53,400

63.34% 10.02% 26.63%

plaval ‘swam.indet.ipfv’ plyl ‘swam.det.ipfv’ poplyl ‘swam.pfv’

â•⁄â•⁄â•›â•›482,000 â•⁄â•⁄â•›â•›438,000 â•⁄â•⁄â•›â•›399,000

36.54% 33.21% 30.25%

â•⁄â•⁄â•›33,800 â•⁄â•⁄â•⁄â•›8,900 â•⁄â•⁄â•⁄â•›7,200

67.74% 17.84% 14.43%

polzal ‘crawled.indet.ipfv’ polz ‘crawled.det.ipfv’ popolz ‘crawled.pfv’

â•⁄â•⁄â•›â•›201,000 â•⁄â•⁄â•›â•›285,000 â•⁄â•⁄â•›â•›346,000

24.16% 34.25% 41.59%

â•⁄â•⁄â•›16,100 â•⁄â•⁄â•⁄â•›3,880 â•⁄â•⁄â•⁄â•›2,450

71.78% 17.30% 10.92%

Verbs of motion under negation in Modern Russian 187

(27) V etot den’ nikto ne plaval. Otdyxali, tancevali. in that day nobody neg swam.indet.ipfv rested.3pl, danced.3pl ‘That day nobody was swimming. [People] rested, danced.’ (Forsyth 1970:â•›341)

A total denial of motion often involves motion towards a specific destination, as in (28), where the rock star Boris Grebenschikov never goes to vote: (28) Boris Grebenščikov nikogda ne xodil na vybory Boris Grebenščikov never neg went.indet.ipfv to elections ‘Boris Grebenschikov never voted.’ (V. Zvereva. “Ne chastyj greben’”)

Or movement on a specific type of vehicle, or even a single named vehicle, as in (29), (29) Belgorodskij gubernator ne plaval na «Titanike» Belgorod.adj governor neg swam.indet.ipfv on Titanic ‘The governor of Belgorod never sailed on the “Titanic”’. (M. Bukov. “MK proigral”)

This type of denial of motion is extremely rare with prefixed VoM, and has different semantics when it appears. For example, a search for ne xodil na vybory retrieved 10,200 hits; a search for ne prixodil na vybory retrieved 14 hits, which involved an expectation of observers in the locus and their adverse reaction to absence, as in (30), where a deputy is so upset that people do not vote that she mentions mutilation as a means of encouraging voters: (30)

deputat Dinara Moldoševa otmetila, čto ran’še otrubali ruki deputy Dinara Moldoševa noted that previously cut.3pl hands i primenjali različnye sankcii k tem, kto ne pri-xodil and applied.3pl different sanctions to those who neg arrived-came.ipfv na vybory to elections ‘Deputy Dinara Moldoševa remarked that in the old days [they] cut hands and applied various sanctions to people who did not come to the elections’. (CIS-news.info)

When total denial of motion is needed, only the imperfective of a simplex VoM can be used, which explains the greater frequency of indeterminate imperfectives of the simplex VoM versus the determinate imperfectives, since a type of motion expressed by the determinate imperfective can also be expressed by an imperfective of prefixed VoM. Thus, the formula of absence of motion for a period of time, followed by motion towards a locus, with observer present, appears

188 Renee Perelmutter

in (31). In this example, a kitten does not go to its new owners (the observers), despite expectation; it is implied that the motion does eventually take place: (31)

Pervye dni on daže ne šel k nam na ruki, first days he even neg went.det.ipfv to us to hands sil’no kričal, vidimo emu bylo bol’no, sejčas on stal strongly screamed, apparently he.dat was painful now he became gorazdo spokojnee. considerably calmer ‘During the first days he didn’t even come to our hands, cried hard – apparently he was in pain, now he has become much calmer.’ (L. Kuranova. “Kotiku 4 goda”)

What is the difference between the indeterminate imperfective of a simplex VoM and the imperfective of the prefixed VoM? From the data, it appears that the difference lies in the intent of the moving figure. For the prefixed VoM, there is no intent for the figure to move despite the expectation of an observer in the locus. For the simplex VoM, the intent to move is there, together with the reasoning why the movement does not happen. When a prefixed VoM is used, the focus is on the interaction between the observer and the absentee. In (32), the young man wants to apologize to the observer for his absence, but does not know what to say, since he did not want to come: (32)

Prosti, ja ne pri-xodil k tebe vsë èto vremja, ja … forgive.2sg I neg arrived-came.ipfv to you all this time I Zapnuvšis’, on zamolčal. On ponjal, čto […] na samom faltering he fell.silent.pfv he understood that on real.prep dele on prosto ne hotel obščat’sja. thing.prep he simply neg wanted.ipfv communicate. inf ‘I am sorry I did not come to you all this time, I … Faltering, he fell silent. He realized that … he just did not want to talk [to her].’ (Ya. Zorin. “Istočnik”)

Similarly in (33) – if a child did not yet arrive to ask for a dog, the desire to come is not yet there. Note the observer’s expectation of future arrival: (33)

Esli vaš malyš ešče ne pri-xodil k vam s pros’boj if your little.one still neg arrived-came.ipfv to you with request zavesti sobaku to ne rasslabljajtes’. Vse ešče vperedi! acquire.pet.inf dog.acc then neg relax.refl. everything yet before ‘If your little one has not yet come.ipfv to you asking to get a dog, don’t relaxÂ€– everything is still in front of you!’ (A. Bisembaeva. “Vybiraem zverja”)

Verbs of motion under negation in Modern Russian 189

On the other hand, if the determinate imperfective of šel ‘went’ is used, the focus is on the moving figure rather than on the interaction between the figure and the observer; the desire to move is present, but is not acted upon for various reasons. For example in (34), the speaker recalls that when as a young boy he considered telling his mother about his toothache, he often did not go to her: (34)

Ja ne šel k nej vot po kakoj pričine: ne somnevajas’ I neg went.det.ipfv to her here on which reason: neg doubting v tom, čto ona dast mne aspirin, ja v to že vremja in that.prep that she give.pfv me.dat aspirin, I in this part time znal, čto ona […] na sledujuščee utro povedet menja k knew that she on next morning lead.pfv me.acc to zubnomu vraču. dental doctor ‘I did not go to her (my mother) for the following reason: not doubting that she would give me aspirin, I knew that she would not stop at that, and the next morning would take me to the dentist.’ (C. Lewis. “Vo čto eto obxoditsja”)

Forsyth’s example zdes’ ljudi ne šli, a stojali ‘people did not walk.det.ipfv here, but stood’ which he takes to illustrate an absence of motion at a given moment, or absence of the tendency to perform it, I believe illustrates a different construction, namely ‘figure moved not in manner A, but in manner B’. This construction is not frequent, but when it does occur, determinate imperfective is indeed used for the denied manner of motion, as in (35): (35)

On šel na rabotu, daže ne šel, a bežal, he went.det.ipfv to work even neg went.det.ipfv but ran.det.ipfv Eži voobšče redko xodjat. hedgehogs at.all rarely walk.indet.ipfv ‘He walked to work, or rather [he] didn’t work but ran. Hedgehogs rarely walk.’ (“Ežik, kotoryj čto-to ponjal”)

To summarize, the determinate imperfective of simplex VoM is not frequent under negation since it is usually appears in a single construction, that of an absence of expected/planned motion for a period of time; a similar meaning is more frequently expressed for VoM with prefixes of goal/origin which emphasize an observer towards whom the figure moves, and the observer’s expectations. The indeterminate imperfective expresses a total denial of motion which routinely cannot be expressed with prefixed verbs. The perfective of basic simplex verbs (pojti ‘to go.pfv’, poexat’ ‘to go.pfv by vehicle’) is statistically more prominent than the imperfective, which is due to a large number of fixed collocations/idioms that exist with these verbs, such as ne pošel na pol’zu ‘did not help/did not go well for…’ as in (36):

190 Renee Perelmutter

(36)

Kakim budet mir bez detskogo smexa i plača? What.instr will.be world without children.adj laughter and tears Interesno, togda posmotrite etot fil’m. Skažu, čto čelovečestvu Interesting then see.2pl.pfv this movie say.1sg that mankind.dat otsutstvie detej ne pošlo na pol’zu. lack children.gen neg went.pfv to help ‘What would the world be like without children’s laughter and tears? If this interests you, then go see the movie. I will say that the absence of children did (“Ditja Čelovečeskoe”) not benefit mankind.’

The number of perfectives falls sharply for the simplex VoM which do not participate in idioms (such as popolzti ‘to crawl’), and the indeterminate imperfective becomes the most frequent.

5.

Conclusions

Two kinds of motion events are possible under negation: the most frequent is the motion event that never occurred. The less frequent is a motion event that occurred but appears in a negated construction for a reason other than denial that the motion ever took place (e.g., motion after a significant temporal interval). Motion events that did not take place tend to be less detailed in terms of the specification of manner and path of motion. High-manner verbs such as kovyljat’ ‘to wobble’ are extremely rare; prefixed VoM tend to describe a movement towards goal, or from an origin, rather than a detailed specification of the ground covered by the moving figure. In addition, detailed descriptions of a locus do not appear with negated motion events which did not happen. This is to show that although a negative statement must always reference a corresponding affirmative statement, this ‘paired’ affirmative space shows less detail than a regular affirmative space depicting an actual motion event. Motion events that did not happen usually involve not only the moving figure but an observer who is either at goal or at origin of motion, and who is expecting the motion or is otherwise involved in it. The observer imagines the motion event and its non-occurrence can impact him/her. Both imperfective and perfective aspects are possible with this construction, though perfective is preferred. The observer at locus and his/her expectation occurs with most perfective VoMs describing events which did not take place. In some negated VoM constructions the motion actually does take place – such is the usage of the imperfective to indicate motion that takes place after a significant interval. Motion in this case becomes an attribute of the moving figure

Verbs of motion under negation in Modern Russian 191

or of the whole situation. Motion that did take place can have more detailed specification of path and manner than motion which did not take place.

References Borschev, V. & Partee, B. H. 2002. The Russian genitive of negation in existential sentences: The role of theme-rheme structure reconsidered. In Travaux du Cercle Linguistique de Prague 4, E. Hajicova & P. Sgall (eds.), 185–250. Amsterdam: John Benjamins. Clark, H. H. 1974. Semantics and comprehension. In Current Trends in Linguistics 12, T. Sebeok (ed.), 1291–1498. The Hague: Mouton. Fillmore, C. J. & Kay, P. 1993. Construction Grammar Coursebook. Ms, University of California-Berkeley. Forsyth, J. 1970. A Grammar of Aspect: Usage and Meaning in the Russian Verb. Cambridge: CUP. Gazdar, G. 1979. Pragmatics: Implicature, Presupposition, and Logical Form. New York NY: Academic Press. Goldberg, A. E. 1995. Constructions: A Construction Grammar Approach to Argument Struc‑ ture. Chicago IL: University of Chicago Press. Horn, L. R. 1989. A Natural History of Negation. Chicago IL: University of Chicago Press. Isačenko, A. V. 1962. Die russische Sprache der Gegenwart. Formenlehre. Halle: Niemeyer. Keller, F. & Lapata, M. 2003. Using the Web to obtain frequencies of unseen bigrams. Computa‑ tional Linguistics 29(3): 459–484. Keller, F. & Lapata, M. 2005. Web-based models for natural language processing. ACM Transac‑ tions on Speech and Language Processing 2(1): 1–31. Kilgariff, A. & Grefenstette, G. 2003. Introduction to the special issue on the Web as corpus. Computational Linguistics 29(3): 333–347. Lambrecht, K. 2000. Information Structure and Sentence Form. Cambridge: CUP. Lewis, W. D., Farrar, S. & Langendoen, T. 2006. Linguistics in the internet age: Tools and fair use. In Proceedings of the EMELD06 Workshop on Digital Language Documentation: Tools and Standards: The State of the Art. Lansing MI. Merrill, P. 1985. Aspect as evaluation: The case of negation. In The Scope of Slavic Aspect, M.Â€Flier & A. Timberlake (eds.), 129–152. Columbus OH: Slavica. Meyer, C. F. 2004. Can you really study language variation in linguistic corpora? American Speech 79 (4): 339–355. Meyer, C. F., Grabowski, R., Han, T., Mantzouranis, K., & Moses, S. 2003. The World Wide Web as linguistic corpus. In Corpus Analysis: Language Structure and Language Use, C. F. Meyer & P. Leistyna (eds.), 241–254. Amsterdam: Rodopi. Padučeva, E. V. 1992. O semantičeskom podxode k sintaksisu i genitivnom subjekte glagola BYT’. Russian Linguistics 16(1): 53–63. Padučeva, E. V. 1997. Roditelnyj subjekta v otricatel’nom predloženii: sintaksis ili semantika? Voprosy Iazykoznaniia 2: 101–116. Padučeva, E. V. 2004. Dinamičeskie modeli v semantike leksiki. Moskva: Jazyki slavjanskoj kul’tury. Perelmutter, R. 2005. Case choice in Russian genitive/nominative absence constructions. Rus‑ sian Linguistics 29(3): 319–346.

192 Renee Perelmutter

Slobin, D. I. 2004. The many ways to search for a frog: Linguistic typology and the expression of motion events. In Relating Events in Narrative, Vol. 2: Typological and Contextual Perspec‑ tives, S. Strömqvist & L. Verhoeven (eds.), 219–257. Mahwah NJ: Lawrence Erlbaum. Slobin, D. I. 2006. Verbs of Manner of Human Motion. Ms, University of California-Berkeley. Šigurov, V. V. 1993. Tipologija upotreblenija atributivnyx form russkogo glagola v uslovijax otri‑ canija dejstvija. Saransk: Izdatelstvo moskovskogo universiteta. Talmy, L. 1985. Lexicalization patterns: Semantic structure in lexical forms. In Language Typol‑ ogy and Syntactic Description: Grammatical Categories and the Lexicon, T. Shopen (ed.), 57–149. Cambridge: CUP. Talmy, L. 2000. Toward a Cognitive Semantics. Cambridge MA: The MIT Press. Thelwall, M. 2005. Creating and using Web corpora. International Journal of Corpus Linguistics 10(4): 517–541. Timberlake, A. 2004. A Reference Grammar of Russian. Cambridge: CUP. Tottie, G. 1991. Negation in English Speech and Writing: A Study in Variation. San Diego CA: Academic Press. Zaliznjak, A. A. & Šmelev, A. D. 2000. Vvedenie v russkuiu aspektologiiu. Moskva: Jazyki russkoj kul’tury.

Appendix 1. Sources of Internet examples, listed by example number 1. http://bookz.ru/authors/mihail-plackovskii/d45968cf82ef/1-d45968cf82ef.html 2. http://ecoclub.nsu.ru/isar/books/erdakov/4.htm 3. http://leonidandreev.ru/rasskazy/na_stantsii.htm 4. http://turizm.lib.ru/l/luchin_ewgenij_anatolxewich/cvizboromvalpach2.shtml 5. http://www.victory.mil.ru/lib/books/memo/litvin/13.html 6. http://kulichki.com/akter/publ/papanov_a1.htm 7. http://anekdoti.ru/jokes.php?joke_category=99 8. http://topos.ru/article/6162 9. http://www.newsru.com/cinema/07jun2007/antonov.html 10. http://news.bbc.co.uk/hi/russian/talking_point/newsid_1451000/1451670.stm 11. http://www.turizm.ru/india/stories/p-2056.html 12. http://www.wplanet.ru/index.php?show=text&id=5258 13. http://phorum.artek.org/showthread.php3?threadid=522 14. http://lib.krnet.ru/alt/FISKANDER/isk_rassk5.txt 18. http://lib.ru/TURIZM/HITCHHIKE/elba98.txt 19. http://www.levashov-book.ru/levashov/3/1/546/ 20. http://world.lib.ru/r/razumow_g/pravdaivimisel.shtml 21. http://www.russia-today.ru/2004/no_16/16_local_admin_2.htm 22. http://mk.ru/44800.html 23. http://turizt.livejournal.com/34000.html 24. http://www.afisha.ru/article/1004/ 25. http://www.old.kurskcity.ru/events/fetch.html 26. http://www.cirota.ru/forum/view.php?subj=79459 28. http://www.mkset.ru/news/music/7108/ 29. http://www.compromat.ru/page_17656.htm

Verbs of motion under negation in Modern Russian 193

30. http://www.cis-news.info/read/80565/ 31. http://forum.vetby.ru/index.php?showtopic=6276&mode=threaded&pid=11031 32. http://www.likefeo.narod.ru/avtor9.html 33. http://www.detki.kz/kids/sorvanec/zver/ 34. http://wap.zavet.ru/book/02apol/002/031.htm 35. http://lifeizlife.aha.ru/rabb/skazka1.html 36. http://www.blogus.ru/pop/blogs/blogdetail.aspx?id=5059&pid=402055

Appendix 2. Abbreviations used in interlinear glosses acc adj dat det gen indet inf instr ipfv neg nom part pfv pl prep refl sg

accusative adjective dative determinate genitive indeterminate infinitive instrumental imperfective negation nominative particle perfective plural prepositional reflexive singular

part iii

Typological approach to the study of Slavic verbs of motion

chapter 8

Semantic composition of motion verbs in Russian and English The case of intra-typological variability Victoria Hasko

University of Georgia

The study examines and provides evidence for intra-typological variability in the semantic domain of motion in Russian and another satellite-framed language, English. By drawing on parallel corpora comprised of oral narratives in Russian and English, the study contrasts the semantic composition of motion verbs in these two languages. The results reveal that although English has traditionally been viewed as the prototypical satellite-framed language due to its rich Manner-of-motion lexicon, the verb-of-motion repertoire in Russian is superior in its semantic and structural capabilities for encoding motion-related nuances. The verbs attested in the English corpus were monomorphemic, whereas the verbs retrieved from the Russian data consistently and variedly co-encoded such categories as Manner, Path, unidirectionality/non-unidirectionality of motion in space, and aspect through diverse semantic-to-surface associations.

1.

Introduction

During the last several decades the field of linguistics has witnessed the emergence of a rich body of typological studies investigating the domain of motion. Theoretical, applied, and cognitive linguists are studying and comparing expression of motion meanings in diverse language families, ranging from Samoyedic (Shilova 2002) to Nilo-Saharan (Mietzner & Treis 2007). In spite of the fact that all humans share the same basic components of perception, languages differ not only in the surface structures used to encode motion, but also in the amount of explicit attention they pay to various aspects of motion events and spatial relationships. There is a growing consensus that while a universal formula for describing motion events lexically, syntactically, or semantically does not exist (e.g., see Berman & Slobin 1994; Filipović 2007; Hickmann & Robert

198 Victoria Hasko

2006; Slobin 2006), similarities and differences characteristic of motion talk can be captured by a systematic cross-linguistic typology (Talmy 1985, 1991, 2000). The comparative study reported in this chapter contributes to the typological research by outlining fine-grained intra-typological differences that exist in the encoding of motion between Russian and English. This study draws on the elicited speech of adult monolingual speakers of Russian and English and focuses on how the participants filter their perceptions of visually-portrayed motion scenes into verbalized events during a communicative task of story-telling.

2.

Cross-linguistic typology of motion events

The surface structures and systematic patterns used by speakers of various languages to express motion meanings have been analyzed most extensively through the typological framework developed by Talmy (1985, 1991, 2000). Talmy describes several semantic schemata and their systematic relationship with various surface elements involved in the linguistic representation of motion events. Different types of semantic-to-surface associations are considered in Talmy’s research through the lens of lexicalization patterns which, as he noted, can be systematized into similar groups within a cross-linguistic typology. In Talmy’s terms, a Motion event is understood as a situation containing either motion or the continuation of a stationary location (ibid.). A basic motion event is described as consisting of four internal components: Motion, Figure, Ground, and Path. Motion itself can express two motive states: the presence of motion or locatedness. Other internal components include an object, i.e., a Figure which is moving or located with respect to another object, i.e., Ground. The course that a Figure follows or the site that it occupies is referred to as the Path. This framework specifies that a motion event can be associated with external co-events, which most often include Manner and Cause. Manner refers to the subsidiary action or state in which motion might occur, e.g., in English one can move by running, marching, skipping, tiptoeing, waltzing, hightailing, storming, etc. out of the room. Cause is a Co-event which has a causal interrelationship with a Motion event and can be observed in such English verbs as carry, bring, and throw. The verbal complex represents the focal point of the typology of lexicalization patterns proposed by Talmy (ibid.). Specifically, his prevailing interest is in the expression of the conceptual core schema of Path and Manner. Based on the linguistic representation of these two components and their subsequent salience, Talmy (ibid.) divides the world’s languages into two general typological groups: satellite-framed and verb-framed languages. In verb-framed languages (V-languages), Path is salient because it is readily expressed in an obligatory

Semantic composition of motion verbs in Russian and English 199

backgrounding constituent, i.e., the main verb root. Manner, on the other hand, is encoded through an optional foregrounding constituent (such as an adverb or gerundive phrase), and, as a result, is less salient and commonly omitted, e.g., (1) Le chat est sorti de la pièce (en courant). the cat exited.PATH the room (by running.MANNER) ‘The cat ran out of the room.’

Examples of V-languages include Romance and Semitic languages, Japanese, Tamil, as well as most Bantu and Mayan languages. In satellite-framed languages (S-languages), on the other hand, Manner is highly salient, because it is readily included under backgrounding in the main verb root. Path is also expressed in a backgrounding constituent in S-languages, because it is encoded by an element or satellite associated with the main verb (particles, prefixes, agglutinative morphemes, etc. depending on the language), e.g., (2) Kot kradetsia v komnatu. cat sneaks.MANNER in.PATH room ‘The cat is sneaking into the room.’

Sample S-languages include Slavic, Germanic, and Finno-Ugric families. Examples (1) and (2) show that languages in both groups, irrespective of their typological membership, consistently encode Path but differ in their attention to Manner of motion. Berman & Slobin (1994) offer a cognitive-linguistic explanation for this fact. In their view, S-languages allow for a more economical expression of Manner, since they accumulate Manner nuances on the main verb in the predicate; in addition, their syntax permits Path satellites to be adjacent to or even added onto the same verb. In contrast, V-languages are less economical as they encode Path and Manner in separate expressions, i.e., Manner is expressed in a subordinate construction (as en courant in (1)) additive to the main Path verb which increases processing load; hence, these subordinate constructions are frequently omitted in speech to ease processing demands. Cross-linguistic research reports robust evidence in support of the prototypical dichotomy in the habitual expression of Path and Manner observable in the linguistic performance of S- vs. V-language speakers (e.g., see volumes by Aurnague, Hickmann, & Vieu 2007; Hickmann & Robert 2006; Strömqvist & Verhoeven 2004; Verhoeven & Strömqvist 2001; also Slobin 2003, 2005, 2006). Slobin has convincingly demonstrated that S-languages use motion verbs of Manner frequently and habitually, and that these languages are also characterized by wider lexicons for encoding motion events, whereas Manner lexicons in V-languages are smaller and the Manner verbs that are present occur in actual

200 Victoria Hasko

production with reduced frequency. Salience of Manner in S-languages has been reported to result in numerous differences in verbal accounts of motion events in S- vs. V-languages, including ease of lexical access to Manner-of-motion lexicon, increased imagery and understanding of Manner verbs and more metaphorical extensions of Manner verbs in the motion talk of S-languages. Although Talmy’s typology is continuously bolstered by new studies reporting support for the general inter-typological characteristics it outlines, a number of researchers have commented on the limitations of the typology with regard to its ability to capture with precision which motion structures are used and which motion nuances can be encoded in each individual language, arguing that the typology proposes tendencies rather than an exact characterization of lexicalization patterns of motion talk within each individual language (Berman & Slobin 1994). For example, availability of motion verbs for encoding Manner in select contexts has been documented for some V-languages (e.g., Aske 1989; Inagaki 2002; Kopecka this volume; Naigles et al. 1998; Slobin & Hoiting 1994). Other studies have pointed out idiosyncratic variations in terms of lexis, rhetorical style, narrative construction, morphosyntax, cultural detailing, functional properties of entities and figures, reference systems, and modality that exist both inter- and intratypologically (e.g., Bavin 2004; McNeill 2005; Özyürek & Kita 1999; Shay & Seibert 2003; Slobin 2003; Wienold 1995). These data suggest that while inter-typological differences result in dramatic distinctions in how speakers of S- and V-languages linguistically carve out the conceptual domain of motion, finer cases of intra-typological variation should not be overlooked. Thus, thorough description of finer intra-typological differences appears to be necessary for advancing and fine-tuning the field of typological linguistics and for creating a more comprehensive and precise catalogue of diverse typological differences as manifested in the linguistic encoding of motion in individual languages. Moreover, better understanding of intra-typological differences can also advance research and practice in such applied areas as second language acquisition and language teaching methodology (through the study of positive and negative interlingual transfer), investigations of linguistic relativity, and bilingual cognition. Thus, indirect evidence for peculiar intra-linguistic variability between Russian and English comes from the field of L2 pedagogy. Previous studies investigating the effects of intra-typological similarity in the expression of motion reported that positive transfer occurs when learners’ first and second languages belong to the same typological group of motion talk, even if these languages are otherwise as distant as English and Japanese (cf., Yu 1996). Based on these data, we should expect a facilitative effect in second language acquisition of motion structures by English-speaking learners of Russian because both languages belong to the same S-group. Yet, in reality, Russian verbs of motion have traditionally been singled out as one of the most difficult topics for English-

Semantic composition of motion verbs in Russian and English 201

speaking learners of Russian. In fact, Russian verbs of motion are believed to be a topic so challenging, that it “damages [L2 learners’] self-assuredness of expression and hampers further acquisition of the language” (Pahomov 1977:â•›1). For the sources of the acquisitional difficulties experienced by English-speaking learners of Russian to be fully understood and pedagogically addressed, it is necessary to first identify the differences that exist intra-typologically between these two S-languages (cf., Hasko forthcoming, 2009). Hence, my goal is to analyze the form-tomeaning mappings in the verbal complex of Russian and English through a study of attested motion talk. Before I proceed with the discussion of the study, I offer a synthesis of the literature which provides a foundation for my analysis.

3.

Russian and English verbs of motion

3.1

Manner

Within Talmy’s typology, Russian and English are both classified as S-languages, which means that the two languages share significant similarities in how they encode the conceptual core schemas pertaining to the motion domain. Thus, both are believed to have rich lexicons for expressing Manner in the verb root. Numerous studies have attested to the range of the Manner verbs repertoire in English (e.g., Berman & Slobin 1994; McNeill 2005; Slobin & Hoiting 1994; Slobin 2003, 2004, 2005, 2006.). These typological studies are written in English and, for the most part, on English, because English is used as a prototypical example of a Manner-rich language and as a benchmark for inter-typological comparisons. It is a well-documented fact that the salience of Manner in English is highly pronounced irrespective of genres and discourse contexts. A sign located in the San Diego Zoo and quoted by Slobin (2006:â•›60) serves as a fitting illustration of the variety of Manner verbs in English; the sign reads: DO NOT TREAD, MOSEY, HOP, TRAMPLE, STEP, PLOT, TIPTOE, TROT, TRAIPSE, MEANDER, CREEP, PRANCE, AMBLE, JOG, TRUDGE, MARCH, STOMP, TODDLE, JUMP, STUMBLE, TROD, SPRING, OR WALK ON THE PLANTS.

While there are few empirical studies contrasting the Russian and English Manner-of-motion lexicons, Slobin (2006) reports that Russian speakers show at least an equal propensity for encoding Manner-of-motion nuances in comparison to English speakers. The review of Slavic lexicological literature suggests that Russian Manner verbs can encode a very wide range of meanings. Levontina & Shmelev (2005:â•›76) focus on the variety of lexical means to describe locomotion on foot in

202 Victoria Hasko

Russian and comment on the “astonishing” number of commonly used motion verbs to describe this type of movement, including idti ‘go’, brodit’ ‘wander’, bresti ‘toil/drag oneself along’, plestis’ ‘trudge/toil along’, tashchit’sia ‘drag oneself along’, proitis’ ‘walk past somebody, often showing off ’, prokhazhivat’sia ‘stroll up and down; stroll’, pokhazhivat’ ‘frequently visit some place’, shliat’sia ‘gad/loaf about’, shatat’sia ‘gad/loaf about’, sloniat’sia ‘loiter about’, shagnut’/shagat’ ‘step’, marshiro‑ vat’ ‘march’, vyshagivat’ ‘pace’, stupit’/stupat’ ‘step carefully’, shestvovat’ ‘proceed, march proudly’, kovyliat’ ‘hobble; toddle’, semenit’ ‘mince along’, topat’ ‘stamp’, raskhazhivat’ ‘walk/strut about’, razgulivat’sia ‘walk/stroll insatiably; debauch’, guliat’ ‘walk/stroll’, progulivat’sia ‘have a leisurely stroll/walk’, proshvyrnut’sia ‘quickly run on an errand’, krast’sia ‘slink/sneak’. This list is not exhaustive and does not include the vocabulary associated with such semantic sub-domains as those expressing motion up and down, in a circle, along Paths with barriers and obstacles, and along or without roads, aquamotion, and quick motion on foot, along with an extended domain encoding motion by vehicle. Corpus-based evidence from studies investigating such closely-related language to Russian as Serbo-Croatian and Polish seems to suggest, nevertheless, that Slavic languages have smaller and less finely-grained lexical repertoires of Manner verbs in comparison to English and other Germanic languages (Koptjevskaja-Tamm, Divjak, & Rakhilina this volume; Kopecka this volume; Filipović 2007). While both Russian and English are clearly languages with rich lexicons for encoding nuances of motion Manner, contrastive usage-based studies are needed to identify the exact position they occupy along the typological continuum in comparison to each other and to other S-languages (including the Slavic language family) with regard to the salience of Manner in their lexicons and the frequency with which their speakers choose to encode Manner in actual production.

3.2

Path

As far as the encoding of Path is concerned, both languages are again classified in the typological literature as similar in this respect, because they rely on satellites to mark motion Path, with the difference being that in English unbound verbal particles and/or adverbs (as in run to and go home) serve as satellites, whereas in Russian bound prefixes appended to motion verbs play this role (as in vbegat’ ‘inrun’, vyezzhat ‘out-drive’, uletat ‘away-fly’). In comparison to the starkly different patterns attested inter-typologically, this intra-typological morphosyntactic dissimilarity between Russian and English could be considered minor. Nevertheless, close analysis of the form-to-meaning mappings in the encoding of Path in these two languages shows noteworthy differences.

Semantic composition of motion verbs in Russian and English 203

In English, as already mentioned, satellites are mostly free-standing (except for a few low-frequency motion verbs with prefixes like over- and under-), and no restrictions have been reported with respect to the catalogue of Path satellites that various translocational Manner-of-motion verbs can be associated with. Slobin (2003) has additionally shown that because English satellites are unbound, this morphosyntactic freedom allows for relative ease of building complex Path constructions and granularity of Path segmentation; as a result, sub-trajectories can be stacked, as in (3): (3) And it goes down but it rolls him out down the rainspout out into the sidewalk into a bowling alley. (McNeill 2005:â•›196)

As for Russian, the role of motion prefixes in encoding spatial meanings is underscored in the typological literature (Talmy 2000) and widely acknowledged in the work of Slavic linguists (e.g., Cienki 1989; Janda 1988; Israeli 2002; Nikitina this volume). According to Wade (2000), up to sixteen prefixes can attach to certain Russian verb roots with each prefix imparting a different spatial meaning. Launer (1987) analyzed 14 Russian verbs of motion that are particularly apt to combine with various prefixes and counted no fewer than 160 new verbs derived through spatial prefixation. In addition to prefixes, Slavic linguists have traditionally considered prepositions as parts of speech with distinct lexical and grammatical satellite-like meanings integral to the encoding of locative, static, and directed motion events (Gvozdev 1973; Vinogradov 1960). The reasons for treating both prefixes and prepositions as equally important Path-encoding elements are many. Phonologically, Path prefixes and unbound prepositions have compelling similarities and, in fact, many commonly associated prefixes and prepositions are completely homophonous (cf., Matushansky 2002). The conceptual unity of the prefix-preposition combinations in the expression of Path is revealed in constructions in which a combination of both surface components is obligatory in Russian, even when an unbound satellite alone (as in (4a–b)) or no satellite at all (as in (4c)) is used in English to convey Path, e.g.,: (4) a. On v-bezhal v dom. he in-ran into house ‘He ran inside the house.’

. This description refers to one of the episodes from the popular Tweety Bird cartoon in which Sylvester the cat tries to climb up the inside of the drainpipe next to Tweety’s window. When Tweety drops a bowling ball down the drainpipe, Sylvester swallows it and rolls down the drainpipe.

204 Victoria Hasko

b. c.

On obo-shel vokrug doma. he around-went around house ‘He walked around the house.’ On u-shel s raboty. he away-went from work ‘He left work.’

Murav’eva (2006) argues that for L2 learners of Russian, the ability to convey and/or interpret a motion event depends on their level of understanding of how the semantics of a prefix, verb, and preposition interact with each other in the context of an utterance; a similarly holistic view on the morphosyntax of motion Path is evident in the recent methodological materials developed for children studying Russian as a native language in elementary schools (L’vov, Goretskii, & Sosnovskaia 2008). The close bond between prefixes, motion verbs, and associated prepositional phrases manifests itself in unique form-to-meaning constraints in the expression of Path in Russian. Several restrictions have to do with the number of segments a Path description can include in Russian. Thus, morphosyntactically, only one prefix-satellite is allowed per verb in Russian (in contrast to English where satellites are easily stackable, as in (3)). Outside of prefixes, Path can be further elaborated through prepositional phrases, but because of the close semantic prefix-preposition connection (Ferm 1990), multiple prepositional phrases with varying meanings cannot be associated with a single prefix. This means that the potential for building complex motion Paths with stacked-satellite phrases is somewhat limited in Russian. Slobin (2003, 2005, 2006) has indeed attested that English speakers tend to produce Path descriptions of greater lengths than speakers of Russian. Corpus-informed work of Filipović (2007) and Kopecka (this volume) on Serbo-Croatian and Polish suggests that speakers of these languages do not have the same freedom in the simultaneous expression of Manner and Path that speak‑ ers of English enjoy. Both Filipović and Kopecka (ibid.) hypothesize that the simpler the semantics of the motion verb (i.e., the less pronounced the nuance of Manner is), the more likely these verbs are to receive a Path prefix. Such a restriction does not exist in English where, they argue, all Manner verbs can potentially be followed by any of the unbound satellite phrases. This constraint seems to hold for Russian as well, e.g., verbs idti ‘walk on foot’ and prygat’ ‘jump’ combine with most Path prefixes, while guliat’ ‘walk leisurely; stroll’ and kovyliat’ ‘hobble’ combine with very few. The question remains as to how pronounced the aforementioned constraints are in Russian. Their pervasiveness and the scope of their influence on the narrative style of Russian need to be ascertained and assessed through a corpus-based investigation of production data.

Semantic composition of motion verbs in Russian and English 205

3.3

Aspect

Most typological studies capitalize on the encoding of Manner vs. Path because these two categories are central to the inter-typological distinctions outlined by Talmy. But the semantic structure of verbs of motion is much more complicated than that. In fact, Talmy (2000) describes and illustrates 35 different semantic categories that occur cross-linguistically in the verb complex, i.e., the root, affixes, and associated elements. In reference to English and Russian, such central categories as aspect, tense, voice, mood, and person (as well as gender in the case of the Russian verb in the past tense) regularly occur in the verb complex. Because Russian is a synthetic language, the verb is characterized by an impressively complex conjugational paradigm, notably more elaborate than that of English, an analytic language. There is common agreement among Slavic linguists that the Russian verb is the most intricate category of the language in terms of the range of meanings (lexical, grammatical and procedural (expressed by Aktionsarten)) that it can impart, and that the verb has the richest and most complex sub-system of grammatical forms in the language. While the space here does not permit the discussion of all verbal categories, I deem it necessary to highlight the differences in the encoding of grammatical aspect in the two languages, because aspect plays a distinct role in the formation of the lexicalization patterns used to encode motion in Russian. Grammatical aspect, which characterizes how the event unfolds or exists in reference to a time-line (Bussmann 1996; Zalizniak & Shmelev 2000), is a category that dominates the Russian verbal system in that each Russian verb (with the exception of the verb byt’ ‘be’ (Chertkova 1996)) is marked for aspect in all of its forms, including infinitives, imperatives, and participles. Russian aspect is an obligatory binary category which includes sovershennyi vid ‘perfective aspect’ and nesovershennyi vid ‘imperfective aspect’. Most Russian verbs form aspectual pairs whose members can be formally distinguished from each other by a number of characteristics, such as suffixation, derivation from different roots, stress, prefixation, or a combination thereof, e.g., compare roniat’ – uronit’ [ipfv – pfv] ‘drop’, prygat’ – prygnut’ [ipfv – pfv] ‘jump’, rassypat’ – rassypat’ [ipfv – pfv] ‘spill’. With regard to motion talk, the obligatory encoding of aspect means that categories of Manner, Path, and aspect can all co-occur within Russian motion verbs, e.g., (5) On pri-ekhal v Moskvu. he to.PATH-drove.MANNER.PFV in Moscow ‘He arrived in Moscow.’

206 Victoria Hasko

In (5), the Path of motion is rendered by the combination of the prefix pri- and preposition v, and Manner is expressed in the verb root, where the imperfective aspect is also marked. But, as previously mentioned, aspectual form-to-meaning connections are not restricted to the verb root, and derivation of perfective verbs from imperfectives is most frequently associated with prefixation in Russian (Dickey 2006). This is directly relevant to the discussion of motion talk, because prefixes also commonly serve as slots for Path satellites. That is why many Russian verbs of motion, particularly perfective procedurals, are subject to morphological blocking: when the prefixal slot is occupied by an aspectual prefix, Path is rendered solely by an unbound satellite(s), e.g., consider (6): (6) Iz aeroporta my po-ekhali domoi. from airport we “po”.INC-drove.PFV home ‘From the airport we drove home.’

The encoding of aspect in English is dramatically different from the aspectual meanings that the Slavic verb imparts (Binnick 1991; Comrie 1976), but here I would like to highlight the lack of morphological constraints in the encoding of aspect and the relative infrequency of the grammatical marking of aspect, in contrast to Russian. In English, a verb in its infinitive form is not marked for aspect, and auxiliaries and suffixes are needed to impart aspectual meanings (perfect and continuous) to the verb in its other forms. Structurally, the two aspects in English are distinguished as follows: the perfect aspect is marked by the auxiliary verb have + ed-participle, and the progressive aspect is marked by the auxiliary verb be + ing-particle; thus, there are no structural limitations with respect to the coexpression of aspect and Path, as can be the case in Russian. At the same time, the frequency with which aspectual distinctions are grammatically encoded in English is dramatically lower. Whereas in Russian, as already mentioned, each verb must be marked for grammatical aspect, in English it is quite common to have non-perfect, non-progressive verbs, i.e., no grammatical aspectual marking on the verb (Celce-Murcia & Larsen-Freeman 1999). Zero aspect is “overwhelmingly the most common in all (conversation, fiction, newspaper language, and academic prose) registers”, according to Biber et al.’s (1999:â•›461) analysis of a 40-million-word corpus. They report that about 90% of all verbs in English are unmarked for aspect. In sum, the brief discussion above highlights aspect as a category that warrants our attention as we compare the expression of motion events in Russian and English. Aspect adds one more category to the semantic composition of Russian verbs of motion: Manner, Path, and grammatical aspect can all co-occur in the Russian verb internally; however, evidence for the morphological restrictions for simultaneous encoding of Path and aspect in the prefixal slot calls into

Semantic composition of motion verbs in Russian and English 207

question the actual frequency of motion verbs co-expressing all three categories internally.

3.4

Directionality

Curiously, typologically-based investigations of motion talk have largely ignored a semantic category that is central to the expression of motion in Russian (and in most Slavic languages). As detailed as the catalogue of the semantic categories described in the seminal cross-linguistic typology by Talmy (2000) is, even his work does not describe the idiosyncratic encoding of unidirectionality and nonunidirectionality of motion in space (hereinafter (non)-unidirectionality) that a select closed class of motion verbs formally expresses in Russian. The expression of (non)-unidirectionality is rendered by fourteen paired unprefixed verbs of motion in Russian. This closed class of verbs that formally mark (non)-unidirectionality can be argued to be one of the most thoroughly researched and scrupulously described verb classes in the field of Slavic linguistics (e.g., see Foote 1967; Isachenko 1960; Mahota 1996; Stilman 1951). Due to their semantic idiosyncrasies, this group of paired verbs of motion has become known as glagoly dvizheniia ‘THE verbs of motion’ in Slavic linguistic and pedagogical literature. The rigid boundaries for inclusion into the (non)-unidirectional group of motion verbs are based on a seemingly isolated “pairedness” phenomenon. The primary semantic feature of this verbal class is that when unprefixed, these verbs come in imperfective pairs that essentially state the same type of motion Manner, but still provide different information about the motion event (Foote 1967; Isachenko 1960). While a variety of terms have been proposed to capture the different facets of the opposition (including its sub-aspectual nature, e.g., see papers in Part 2 of this volume), I largely agree with Ward (1965) who argued that the primary meaning common to one group of these verbs is the idea of motion in one direction, while the verbs serving as their paired counterparts lack this meaning of unidirectionality; hence, my preference for the terms “unidirectional” and “non-unidirectional”. The surface structures that render (non)-unidirectionality include verb roots, suffixes, and/or changes in stress pattern. It is without dispute that (non)-unidirectionality of motion can potentially be encoded in all languages implicitly (e.g., contextually) and explicitly through

. The group of paired verbs of motion is preserved in West and East Slavic; among Southern Slavic languages, only Slovenian has preserved the trace of the distinction (Sussex & Cubberley 2006).

208 Victoria Hasko

various types of circumlocution. However, (non)-unidirectionality is obligatorily expressed in Russian when certain common types of motion events are described (e.g., motion on foot, swimming, running, driving). Thus, when a running Figure is to be described in Russian, the speaker must attend to the directionality of the figure’s Path and verbally frame the event as either unidirectional or non-unidirectional by choosing between bezhat’ [uni] and begat’ [non] ‘run’, e.g., consider (7a–b): (7) a. b.

Sportsmeny begut po dorozhkam. sportsmeny run.UNI along tracks ‘Sportsmen are running on the tracks (in one direction).’ Deti begaiut po dvoru. children run.NON in yard ‘Children are running in the yard (in no particular direction; back and forth).’

(Non)-unidirectional verbs are unprefixed and refer to types of motion events that are durative and involve non-boundary-crossing or reaching, i.e., their semantics does not permit spatial or aspectual prefixation. Arguably, the addition of a Path-encoding or an aspectual meaning subdues the (non)-unidirectionality nuance, i.e. (non)-unidirectionality is salient only in unprefixed motion verbs. To sum up this discussion, a close comparison of motion talk in Russian and English reveals noteworthy differences between the two languages. In English, there are no restrictions for encoding Manner, Path, and aspect in the verbal phrase, but all of these categories do not commonly co-occur in the verb itself. In Russian, the categories of Manner, Path, aspect, and the idiosyncratic category of (non)-unidirectionality can be packed internally within motion verbs, but with certain morphological and lexical restrictions at play. Therefore, although prior investigations of inter-typological descriptions of motion would suggest that the two languages are similar in how they encode motion (in contrast to the group of V-languages), it is also obvious that intra-typological differences that exist in the semantic structure of Russian and English motion verbs have to be ascertained and attested in actual speech production for us to gauge the extent of these differences, not only in terms of what is possible but also in terms of what is habitual for the speakers of both languages. Consequently, the present study undertakes an empirical investigation with the goal of ascertaining and describing the intratypological similarities and differences in the use of motion verbs by speakers of Russian and English during a spontaneous oral narrative task.

Semantic composition of motion verbs in Russian and English 209

4.

The study

4.1

Participants

Fifty nine participants took part in the study: 30 monolingual speakers of Russian pursuing undergraduate studies in Eastern Russia (mean age = 19 y. o.) and 29 monolingual speakers of English enrolled in a large urban university in the U.S. Northeast (mean age = 22 y. o.). Prior to the interview sessions, participants were pre-screened to confirm their native speaker status in accordance with their target language group and to verify that their proficiency in a foreign language was minimal (in order to eliminate possible cross-linguistic interference).

4.2

Procedure and stimulus

The intent of this investigation is to analyze and compare the use of motion verbs by speakers of English in Russian in actual production. Therefore, to collect samples of authentic language use, I chose a spontaneous oral story-telling as a task of high ecological validity that we habitually engage in every day. Narrative elicitation has gained particular popularity in recent second language acquisition research as well as in the social sciences (Creswell 2003). Pavlenko (2002) points out several benefits of narrative elicitation such as the contextualized nature of elicited speech samples and a relative measure of control (i.e., all participants are asked to describe the same stimulus); the resulting narratives therefore allow for a meaningful contrastive analysis. A wordless picture book Frog, where are you? (Mayer 1969) served as a visual stimulus for eliciting motion-rich descriptions. Prior to the onset of story-telling, participants were given 5 minutes to look through the book. The whole session took under 15 minutes, on average, and was audio recorded. Upon completion of the story-telling task, all audio recordings were transcribed and stored electronically as textual corpora.

4.3

Data analysis

To analyze the use of motion structures, the study adopted contrastive corpus analysis, a methodological approach in which electronic corpora comparable in size, domain, genre, and other characteristics are formed from language samples collected from several speaker groups, with all participants being similar in age, socio-economic and educational background, etc. (Pavlenko & Driagina 2007). In this study, two oral speech corpora were created from the narratives produced by speakers of Russian and English and contrasted with the purpose

210 Victoria Hasko

of ascertaining similarities and differences in the semantic composition of motion verbs and the frequency with which the verbs partake in encoding such central categories of motion as Manner, Path, aspect, and, in the case of Russian, (non)-unidirectionality. To ensure quick and easy retrieval of motion elements, all motion verbs, corresponding Path satellites (bound and unbound), and aspectual prefixes marking temporal boundaries of motion events were tagged as such in both corpora. The concordancing software WordSmith Tools was used to retrieve motion elements for analysis. Throughout the analysis, I distinguish between types (units of meaning or words) of motion elements in order to measure variety and tokens (lexical items) which are used to measure frequency of motion expressions in both corpora.

5.

Findings

Table 1 summarizes the comparison of the corpora in terms of size, diversity of motion verb lexicon, and frequency of motion verb usage. The word count, adjusted for articles in the English data, shows that the Russian speakers produced slightly longer stories (mean = 399 words/narrative) than the English-speaking participants (mean = 367 words/narrative), although the difference is not statistically significant (Mann-Whitney U-test, p > 0.05). Yet, comparison of motion elements usage reveals important differences between the two corpora with respect to both the variety and frequency of motion verb usage. As far as the attested variety of motion verbs is concerned, the Russian-language narratives are characterized by a notably richer verb-of-motion lexicon employed by the narrators: 141 different motion verb types were identified in the Russian corpus vs. 70 in the English data. The Russian corpus also has more instances of motion verb tokens in comparison to the English data (1,073 vs. 684, respectively); the difference in the number of motion verbs retrieved from the two datasets is highly significant (Mann-Whitney U-test, p < 0.001). Overall, the frequency of motion verbs in the . Counting verb types in Russian is more complicated than in English for several reasons. First, individual motion lemmas typically come in imperfective–perfective verb pairs (e.g., prygat’ [ipfv] – prygnut’ [pfv] ‘jump’. Because both imperfective and perfective verbs encode the same motion concept, I counted them as one lemma. Secondly, it is important to bear in mind that in Russian several motion verbs can be formed from the same motion root when a Path prefix is added onto the motion stem (e.g., bezhat’ ‘run’, vy-bezhat’ ‘run out’, v-bezhat’ ‘run in’), but the newly derived Path verbs all refer to the same Manner concept while encoding different Path nuances. For the purpose of assessing the lexical variety in the domain of motion in Russian, I counted simple motion verbs and the prefixed verbs derived from them as one motion lemma.

Semantic composition of motion verbs in Russian and English 211

Table 1.â•‡ Corpus size, variety and frequency of motion verbs

Russian corpus n = 30 English corpus n = 29

Mean N of words

N of motion verb types

N of motion verb tokens

N of motion verbs per 100 words

399

141

1,073

9

367

â•⁄ 70

â•⁄â•›684

5

corpora is approximately 9 motion verbs per hundred words in the Russian corpus vs. about 5 in the English corpus. Assessment of Manner salience in the two corpora serves as the first point of comparison. Based on the primary lexical meaning imparted by the verb root, all verbs were divided into the categories of Manner and Non-Manner, including Path, Basic Motion, and Location. The overwhelming majority of verb types in the Russian and English narratives fall into the category of Manner (see FigureÂ€1), whereas non-Manner verb types constitute a small proportion of the attested motion verb lexicons (10% and 16% in Russian and English, respectively). Comparatively, however, the Manner lexicon attested in the Russian corpus (N = 128) is strikingly more varied than the catalogue of Manner verb types found in the English narratives (N = 59). The magnitude of this lexical disparity is remarkable, especially considering that English is typically viewed as the benchmark language as far as the lexical diversity of Manner verbs is concerned. As for the rate of the actual occurrence of Manner descriptions in the data, Figure 2 demonstrates that the encoding of Manner tokens is omnipresent in both

140

100%

120

90%

100

70%

80%

80

60%

60

40%

50% 30%

40

20%

20 0

10% Manner verbs Russian

Other English

Figure 1.â•‡ Variety of motion verb types

0%

Manner Russian

Non-Manner English

Figure 2.â•‡ Frequency of motion verb types

212 Victoria Hasko

Table 2.â•‡ Distribution of Path-encoding structures across morphosyntactic categories

Russian corpus (n=30) English corpus (n=29)

Verbs, types/tokens

Bound elements; types/tokens

Unbound elements; types/tokens

Total, types/tokens

8/34

17/580

41/705

66/1,319

7/39

–

36/606

43/645

corpora and, alternately, that non-Manner verbs are dramatically less frequent in both the Russian and English narratives. Yet, in comparison to the English participants, the Russian narrators expressed Manner more regularly. In fact, nonManner verbs account for 11% of all motion verb tokens in the Russian data vs. 29% in the English narratives. The English-speaking participants were therefore almost three times more likely than the Russian narrators to opt for a deictic or basic motion verb instead of a Manner verb. Let us now consider the encoding of Path in the narratives. As discussed earlier, the typology of motion events suggests that Path in satellite-framed languages is typically expressed by non-verb elements. The corpora under analysis reveal a less clear-cut picture (see Table 2 which summarizes how Path-encoding structures are distributed across morphosyntactic categories in Russian and English). First, instances of motion verbs encoding Path in a root were attested in both Russian and English corpora, although they only account for a small number of motion verb types/tokens (8/34 and 7/39, respectively). Secondly, Russian Manner verbs serve as the primary means for encoding Path in the prefixal position: 17 different types / 580 tokens of bound Path satellites were retrieved from the Russian corpus where they combined with various types of Manner verbs to derive new motion lemmas co-expressing Manner and Path. Finally, both language groups made use of non-verb Path elements: 41 typesÂ€/ 705 tokens of unbound Path elements were attested in the Russian corpus and 36 different types / 606 free-standing satellites in the English narratives. Contrary to the previous reports that Path satellites are habitually stacked in English, I found that the overwhelming majority of Path descriptions (78%) consist of single-segmented clauses, while 21% are double-segmented. In contrast, double-segmented clauses are prevalent in the Russian narratives and account for 66% of all Path descriptions, whereas only 30% of Paths are single-segmented. In both corpora, Path descriptions with 3 or 4 segments are rare (1% in the English corpus and 4% in the Russian corpus). Figures 3 and 4 offer a graphic comparison of the Path lengths in the data.

Semantic composition of motion verbs in Russian and English 213

2 seg 21%

3 or 4 seg 1%

3 seg 4% 1 seg 30%

2 seg 66%

1 seg 78%

Figure 3.â•‡ Path length in English

Figure 4.â•‡ Path length in Russian

The category of (non)-unidirectionality, unattested in the English corpus, was formally expressed in the narratives produced by the Russian narrators. The frequency of (non)-unidirectional lemmas proportionately to other motion verbs in the Russian corpus is summarized in Table 3. The total of 59 tokens was retrieved (49 unidirectional and 10 non-unidirectional), which constitutes 5.5% of the attested Russian motion verb lexicon. Besides encoding (non)-unidirectionality, these verbs played an additional word-building role as they served as the basis for derivation of new prefixed motion verbs with (non)-unidirectional stems (e.g., plyt’ [uni] – plavat’ [non] => vyplyt’ [Path, pfv] – vyplyvat’ [Path, ipfv]). Thus, 39.5% of the total number of motion verbs in the Russian corpus were derived from (non)-unidirectional lexemes, i.e., the (non)-unidirectional verbs have been shown to have a high combinatory potential as far as their co-occurrence with Path satellites is concerned. With regard to aspect, we have established that 100% of Russian verbs encode aspect internally. As for the distribution of this binary category in the corpus, the majority of motion verbs occurred in perfective (76% of all tokens) and 24% were Table 3.â•‡ Frequency and representation of (non)-unidirectional verbs in the Russian corpus Unprefixed (non)-uni verbs Tokens Proportion

Prefixed verbs derived from (non)-uni verbs

59 (83% uni; 17% non) 424 5.5% 39.5%

Other motion All motion verbs verbs 590 55%

1,073 100%

214 Victoria Hasko

Table 4.â•‡ Aspectual marking in the corpora Russian corpus (n = 30) English corpus (n = 29)

Perfective

Imperfective Perfect

Progressive

Zero

76%, of which 16% are procedurals N/A

24%

N/A

N/A

N/A

N/A

1%

26%

73%

used in imperfective aspect. Besides the encoding of perfective/imperfective, a class of verbs with idiosyncratic lexical qualities pertaining to aspectuality was attested in the Russian data; this class of Aktionsarten accounts for 16% of all perfective verbs in the corpus. On the other hand, the English corpus show that “zero” aspect (i.e., lack of progressive or perfect marking) is predominant in the descriptions of motion events in English. Specifically, 73% of motion verbs in the English corpus are not marked for aspect. Progressive aspect, the second most common type, accounts for 26% of all motion descriptions in the English narratives, and 1% of motion verbs are marked for perfect. Finally, I turn to the semantic composition of motion verbs. Motion verbs retrieved from the English corpus are monomorphemic and express one semantic category at a time in a verb root, either Manner or Path. Motion verbs attested in the Russian corpus are characterized by a more complex morphosemantic makeup. As the results show, the most lexically diverse and well-represented group in the Russian corpus is that of Manner verbs. We have seen that internal co-expression of Path, (non)-unidirectionality, and aspect/aspectuality is possible in the Russian motion verbs. The question arises how habitually Russian speakers co-express all or a combination of these categories. As Figure 5 demonstrates, the most Manner, (N)U Aspect 5%

Manner Aspect 20%

Manner Path Aspect 75%

Figure 5.â•‡ Semantic composition of Russian Manner verbs

Semantic composition of motion verbs in Russian and English 215

common composition of the Manner verbs represented in the Russian narratives encompasses all three categories of Manner, Path, and aspect accounting for 75% of all Manner descriptions. Manner verbs co-occur with the single category of aspect in 20% of all Manner descriptions. Internal co-expression of Manner, (non)unidirectionality, and aspect was attested in 5% of Manner expressions.

6.

Discussion

In both languages the verb root was attested as the nuclei for expressing motion Manner: Manner verbs surfaced as the most lexically diverse and most frequently represented motion verb group. An impressive variety of Manner verb types was found in each of the corpora, including such Manner-related nuances as “difficulty”, “velocity”, “stealth”, “awkwardness”, “aimlessness”, “attitude”, and “motor pattern”; however, the Russian narratives display almost a doubly more varied Manner lexicon and a notably higher frequency of encoding Manner in comparison to the English corpus. The English-language participants were shown to be almost three times more likely than the Russian narrators to opt for a deictic or basic motion verb instead of a Manner verb than the Russian narrators (as 29% vs. 11% of motion verb tokens in the English and Russian corpora, respectively, are non-Manner). For example, in one Frog Story scene the frog is portrayed sneaking out of the jar while his owner is sleeping. None of the Russian speakers used a basic motion verb but rather specified Manner through a motor pattern or an instantaneous motion verb (e.g., liagushka vyprygnula / ubezhalaÂ€/ uskakala / vykarabkalas’ / vylezla / vykralas’ / ischezla ‘the frog jumped out /Â€ran away / jumped away / clambered out / climbed out / sneaked out / disappeared’). The use of a verb signifying Basic Motion on foot vyiti ‘go/get out’ would be ungrammatical in Russian with a non-human Figure. In the English data, the frog’s disappearance is described through Manner verbs (including sneak, creep, climb, jump, and escape), but in contrast to Russian, basic motion verbs such as get and go were also attested, e.g., (8) a. The little boy discovers that the frog is gone. b. The frog is trying to get out of his jar.

Thus, Russian emerged as a more Manner salient language than English. This conclusion is in contrast with the studies investigating such Slavic languages as Serbo-Croatian (Filipović 2007) and Polish (Kopecka this volume) which report greater diversity of motion Manner repertoires in English. The dissimilar behavior of Russian and these two Slavic languages confirms that in typological research

216 Victoria Hasko

of motion talk membership in a language family does not guarantee identical behavior with respect to the expression of motion meanings or the same position along the typological continuum of Manner salience. Based on the data at hand, Russian is more Manner salient than English (a Germanic language) and, in turn, English is more Manner-salient than Serbo-Croatian and Polish (Slavic languages like Russian). At the same time, I would like to caution that the overall size of Manner lexicons should not be the only point of comparison in cross-linguistic studies, as there may be differences in how richly individual sub-domains are encoded in each language. For example, the corpora analyzed here demonstrate that the sub-domains expressing such Manner nuances as “velocity”, “stealth”, and “posture” seem to be particularly saturated in Russian in comparison to English. Several papers in this volume provide evidence for cross-linguistic intra-typological variation specific to such sub-domains as aquamotion, rotation, motion on foot, and saccadic motion, both among Slavic and between Slavic and non-Slavic S-languages (see other papers in Part 3 of this volume). In addition to the superior Manner salience, the results reveal a more extended repertoire of Path elements and significantly more frequent encoding of Path in the Russian corpus; the count of Path tokens is twice as high in the Russian data as compared to the English narratives. Moreover, Russian offers more morphosyntactic choices for form-to-meaning mappings with regard to the expression of Path. The examples in (9) illustrate encoding of Path internally in verb roots (9a) and bound satellites (9b), as well as externally through unbound Path elements (9c). (9) a. b. c.

Mal’chik s-pustilsia za sobakoi. boy down-lowered.himself after dog ‘The boy descended after the dog.’ Sobachke prishlos’ u-bezhat’. dog had to away-run ‘The dog had to run away.’ On predlagal svoemu drugu vernut’sia obratno. he suggested his friend to.return back ‘He suggested to his friend that they should come back.’

The Russian narrators drew on prefixes-satellites and unbound elements with equal frequency, but only free-standing satellites were attested in English. Verbs encoding Path in verb roots (which is a characteristic of V-languages) account for a small number of motion lemmas in both corpora and, therefore, could be considered deviations from the typological tendency or typological leakage (cf., O’Connor) rather than a rule.

Semantic composition of motion verbs in Russian and English 217

Morphosyntactically, only one prefixal slot exists in Russian, and it was previously hypothesized (Slobin 2005) that this characteristic of the Russian morphosyntax, along with the strong conceptual bond between a prefix and an associate unbound Path element, might restrict the ability of its speakers to express Path as freely as in English, where satellites are easily stackable. The findings of the present study do not provide support for these conjectures – either in reference to Russian or English. Contrary to the previous findings that Path satellites are habitually stacked in English, I found that the overwhelming majority (78%) of Path descriptions associated with each individual motion verb overwhelmingly consist of single-segmented clauses, as in (10): (10) The deer carries him away and drops him into the pond.

In contrast, the majority of Path clauses in the Russian data (66%) are longer and consist of two segments, as in (11): (11) Sobaka dazhe za-lezla v banku. dog even in-climbed into jar ‘The dog even stuck its head in the jar.’

How can we reconcile the Path-related morphosyntactic restrictions with the fact the Russian narratives have longer Paths than the English corpus? It would seem that we need to re-evaluate whether the morphosyntactic phenomena at hand are in fact restrictions or not. I would argue that what characterizes the encoding of Path in Russian is the presence of morphosyntactic rules which regulate the expression of Path through the combination of prefixes and conceptually related unbound Path elements. Prefixes are not optional in Russian when descriptions of boundary-crossing and reaching motion events are offered (cf., Hasko 2009), and their semantic bond with free-standing satellites is likely to compel Russian speakers to use both. For example, in (11) above, the prefix za- is not grammatically omissible given that a boundary-crossing event is described. The verb za‑ lezt’ ‘in-climb’ was in fact attested in the Russian corpus with an associated Path phrase 100% of the time because the semantics of the prefix alone is not specific enough and needs to be accompanied by another satellite specifying the exact direction and the nature of boundary crossing. Consider another example with the same verb in (12); it illustrates how a different Path preposition gives a new interpretation to the entire motion event: (12) Mal’chik v eto vremia za-lez na derevo. boy in this time on-climbed on tree ‘In the meanwhile the boy climbed the tree.’

218 Victoria Hasko

Thus, I would argue that the complexity of the Russian morphosyntax does not restrict the freedom of encoding Path nuances but rather regularizes and enforces the systematic encoding of fine-grained nuances pertaining to Path. In English, because satellites are independent semantically and structurally, their co-occurrence is not required and, as the corpus analysis revealed, not as systematic or habitual as in the Russian motion talk. In addition to the lexical diversity and habitual occurrence of the categories of Manner and Path in the Russian narratives, they also contain instances of the formal expression of (non)-unidirectionality, with the group of unidirectional verbs represented more consistently. While proportionately to the whole corpus these verbs constitute only a small group, contextual analysis reveals that the unidirectional verbs serve as the unique means for framing the narrative plot of the Frog Story as an explicitly unidirectional adventure – the boy and the dog are on a quest to find the lost pet frog. Most of the motion events in the storybook are in fact pictorially presented as trips from one point to the other with the starting point and the endpoint of these trajectories explicitly provided. Unidirectional verbs in the Russian corpus verbally portray events as goal-oriented trips in one direction: first, a trip from the house to the forest, then from the forest to the swamp, etc., as in (13a–b). (13) a. b.

Shli oni dolgo / do-breli do lesa. walked.UNI they for a long time / until-wandered up.to forest ‘They walked for a long time until they wandered into the forest.’ [Olen’] bezhit bezhit i vdrug v ovrag. deer runs.UNI runs.UNI and suddenly into ravine ‘The deer keeps running and suddenly (falls) into a ravine.’

Remarkably, (non)-unidirectional verbs are all Manner verbs except for the pair idti-khodit’ which encodes basic motion on foot. Therefore, the roots of these verbs compactly and simultaneously co-express nuanced details pertaining to motion Manner and the directionality of motion in space. While the conceptual differences between Russian perfective/imperfective and English perfect/progressive aspects are beyond the scope of the paper, the collected data reveal a dramatic difference in the frequency with which the category of aspect was unit. One-hundred percent of all Russian motion verbs were marked for aspect internally, but only 27% of motion descriptions were marked as +perfect or +progressive. In addition to the general perfective/imperfective distinction, procedural nuances pertaining to the temporal characteristics and/or the intensity of motion events were encoded through perfectivizing prefixation; such nuances include inception of motion (as in (14a), comitative motion (as the accompaniment to another action, as in (14b)), and delimitative (i.e., temporally-

Semantic composition of motion verbs in Russian and English 219

bounded motion, as in (14c)). Sixteen percent of all perfective verbs encode such procedural meanings. (14) a. b. c.

[Мal’chik] schastlivyi po-polz domoi cherez boloto. boy happy “po”.INC-crawled home through marsh ‘The boy started crawling home through the marsh happily.’ [Mal’chik’] shiol ochen’ medlenno po-tiraia svoi ssadiny. boy walked very slowly “po”.COM-rubbing his bruises ‘The boy was walking very slowly rubbing his bruises.’ Sova po-kruzhilas’, po-kruzhilas’ nad nim i u-letela. owl “po”.DEL-circled “po”.DEL-circled over him and away-flew ‘The owl hovered over him for a while and left.’

To summarize, the corpora collected for the study attested that Russian motion verbs can internally pack and co-encode the categories Manner, Path, (non)unidirectionality of motion in space, as well as aspect and aspectuality; the semantic composition attested most frequently includes Manner, Path and aspect. Even though the morphosyntax in Russian is fairly stringent in comparison to English in that only one semantic category or element can be encoded in the prefixal position (i.e., Path vs. aspect, and, in the case of Path, only one satellite per verb), the frequency of encoding of these categories is indisputably higher in the Russian corpus compared to the English. The categories are distributed across such surface elements as the verb root (Manner, (non)-unidirectionality, and Path in select verbs), prefixes (Path), and suffixes ((non)-unidirectionality)); all of these elements can be involved in aspectual marking. The complexity of the semantic make-up and the morphosyntax of the Russian motion lexicon are in stark contrast to the verbal lexicon in English, where motion verbs are monomorphemic. Slobin (2006) hypothesized that one of the psycholinguistic factors that facilitate regular and frequent encoding of a semantic domain in a language is its expression through a single morpheme, because presumably it is easier to access a simple monomorphemic lexical item. The findings of this study contradict this conclusion, as the semantic composition of Russian verbs of motion is more complex than that of English verbs, and yet the encoding of the motion domain is more lexically diversified and habitual in Russian. These insights represent only the initial step towards understanding the nature of the differences in the expression of motion in the two languages but not their psycholinguistic causes. My preliminary hypothesis is that the differences can be attributed to the highly regularized morphosyntax that underlies the encoding of verbal categories in Russian. It appears that the rigidity of the morphosyntactic rules helps regulate the lexico-semantic encoding of motion events, as speakers are forced to obligatorily assign the morphosyntactic slots to each of the semantic categories and, therefore,

220 Victoria Hasko

notice and differentiate between various spatial and temporal nuances of motion events as the form-to-meaning assignment takes place. Further studies employing psycholinguistic experimentation and in-depth cognitive linguistic analysis are needed to explain both the divergences in encoding motion meanings and the causes for this dissimilarity between Russian and English and similar cases of intra-typological variability.

7.

Conclusion

The present study intended to conduct a contrastive investigation of motion verb usage during a spontaneous oral narrative task with the goal of verifying and attesting the extent of the hypothesized intra-typological variation between two S-languages (i.e., Russian and English) in the domain of motion. The findings unambiguously confirm that in spite of their shared membership in the group of S-languages, Russian and English present a noteworthy case of intra-typological variability with regard to the semantic composition and occurrence of motion verbs. As expected, motion events were encoded in both corpora frequently and variedly, but the internal structure of Russian motion verbs was shown to be more complex semantically and structurally contributing to the greater specificity of the attested motion descriptions. The analysis was built on the insights drawn from the theoretical work by Slavic linguists and general typological research. This interdisciplinary approach has shown that comprehensive studies of individual languages are of paramount importance as they can bring to light idiosyncratic categories that are not typically considered in broad typological research; alternately, studies limited exclusively to the specialized literature on an individual language (or even a language family) might lack the points of comparison that cross-linguistic typological work offers. Thus, the sole focus on (non)-unidirectional verbs characteristic of the work of Slavic scholars only analyzes the tip of the iceberg when it comes to understanding how motion is encoded in Russian. I hope that the insights gained from the study will facilitate and instigate further research aimed at understanding how speakers of different languages make sense of and communicate motion relations, as well as other conceptual categories central to human experiences.

Semantic composition of motion verbs in Russian and English 221

References Aurnague, M., Hickmann, M. & Vieu, L. 2007. The Categorization of Spatial Entities in Lan‑ guage and Cognition. Amsterdam: John Benjamins. Aske, J. 1989. Path predicates in English and Spanish: A closer look. Proceedings of Berkeley Linguistics Society 15: 1–14. Bavin, E. 2004. Focusing on “where”: An analysis of Warlpiri frog stories. In Relating Events in Narrative: Typological and Contextual Perspectives, S. Strömqvist & L. Verhoeven (eds.), 17–35. Mahwah NJ: Lawrence Erlbaum. Berman, R. & Slobin, D. I. (eds.). 1994. Relating Events in Narratives: A Crosslinguistic Develop‑ ment Study. Hillsdale NJ: Lawrence Erlbaum. Biber, D., Johansson, S., Leech, G., Conrad, S. & Finegan, E. 1999. Longman Grammar of Spoken and Written English. Harlow: Longman. Binnick, R. I. 1991. Time and the Verb: A Guide to Tense and Aspect. Oxford: OUP. Bussmann, H. 1996. Routledge Dictionary of Language and Linguistics. London: Routledge. Celce-Murcia, M. & Larsen-Freeman, D. 1999. The Grammar Book: An ESL/EFL Teacher’s Course. Boston MA: Heinle & Heinle. Chertkova, M. 1996. Grammaticheskaia kategoriia vida v sovremennom russkom iazyke. Moscow: Moscow State University. Cienki, A. J. 1989. Spatial Cognition and the Semantics of Prepositions in English, Polish, and Russian. München: Verlag Otto Sagner. Comrie, B. 1976. Aspect: An Introduction to the Study of Verbal Aspect and Related Problems. Cambridge: CUP. Creswell, J. W. 2003. Research Design: Qualitative, Quantitative and Mixed Methods Approaches. London: Sage. Dickey, S. M. 2006. Aspectual pairs, goal orientation, and po-delimitatives in Russian. Glossos 7: 1–32. Ferm, L. 1990. Vyrazhenie napravleniia pri pristavochnykh glagolakh peremeshcheniia v sovre‑ mennom russkom iazyke: K voprosu prefiksalno-predlozhnogo determinizma. Uppsala: Academiae Upsaliensis. Filipović, L. 2007. Talking About Motion: A Crosslinguistic Investigation of Lexicalization Pat‑ terns. Amsterdam: John Benjamins. Foote, I. P. 1967. Verbs of Motion [Studies in Modern Russian Language 1]. Cambridge: CUP. Gvozdev, A. N. 1973. Sevremennyi russkii iazyk. Moskva: Akademiia Nauk. Hasko, V. Forthcoming. The role of thinking for speaking in adult L2 speech: The case of (non)unidirectionality encoding by American learners of Russian. In Linguistic Relativity in Sec‑ ond Language Acquisition: Evidence of First Language Thinking for Speaking, Z.-H. Han & T. Cadierno (eds.). Clevedon: Multilingual Matters. Hasko, V. 2009. The locus of difficulties in the acquisition of Russian verbs of motion by highly proficient learners. In Special Forum on Teaching and Learning Russian Verbs of Motion, V.Â€Hasko (ed.), Slavic and Eastern European Journal 53(3): 360–385. Hickmann, M. & Robert, S. (eds.). 2006. Space in Languages: Linguistic Systems and Cognitive Categories. Amsterdam: John Benjamins. Inagaki, S. 2002. Japanese learners’ acquisition of English manner-of-motion verbs with locational/directional PPs. Second Language Research 18: 3–27.

222 Victoria Hasko

Isachenko, A. V. 1960. Grammaticheskii stroi russkogo iazyka v sopostavlenii s slovatskim: Mor‑ fologiia. Bratislava: The Slovak Academy of Sciences. Israeli, A. 2002. Russian verbs of motion: Focus, deixis, and viewpoint. Cahiers Chronos 10: 97–118. Janda, L. 1988. The mapping of elements of cognitive space onto grammatical relations: An example from Russian verbal prefixation. In Topics in Cognitive Linguistics, B. Rudzka-Ostyn (ed.), 327–343. Amsterdam: John Benjamins. Launer, M. 1987. The semantic structure of verbs of motion in Russian. Russian Language Jour‑ nal 41: 77–105. Levontina, I. B. & Shmelev, A. D. 2005. Na svoikh dvoikh: leksika peremeshcheniia v russkom iazyke. In Kliuchevye idei russkoi iazykovoi kartiny mira, I. B. Levontina & A. D. Shmelev (eds.), 76–95. Moscow: Iazyki slavianskoi russkoi kul’tury. L’vov, M. P., Goretskii, V. G. & Sosnovskaia, O. V. 2000. Metodika prepodavaniia russkogo iazyka v nachal’nykh klassakh. Moscow: Akademiia. Mahota, W. 1996. Russian Motion Verbs for Intermediate Students. New Haven CT: Yale University Press. Matushansky, O. 2002. On formal identity of Russian prefixes and prepositions. In Phonological Answers (and Their Corresponding Questions), A. Csirmaz, L. Zhiqiang, A. Nevins, O.Â€Vaysman & M. Wagner, 217–253. Cambridge MA: MIT Working Papers in Linguistics. Mayer, M. 1969. Frog, Where Are You? New York NY: Dial Press. McNeill, D. 2005. Gesture and Thought. Chicago IL: University of Chicago Press. Mietzner, A. & Treis, Y. (eds.). 2007. Encoding Motion: Case Studies from Africa. Special Issue of Annual Publication in African Linguistics 5. Murav’eva, L. S. 2006. Glagoly dvizheniia v russkom iazyke. Moscow: Russkii iazyk. Naigles, L. R., Eisenberg, A. R., Kako, E. T., Highter, M. & McGraw, N. 1998. Speaking of motion: Verb use in English and Spanish. Language and Cognition Processes 13: 521–549. O’Connor, L. 2009. All typologies leak: Predicates of change in Lowland Chontal of Oaxaca. In New Challenges in Typology: Transcending the Borders and Refining the Distinctions, P. Epps & A. Arkhipov (eds.), 343–364. Berlin: Mouton de Gruyter. Özyürek, A. & Kita, S. 1999. Expressing manner and path in English and Turkish: Differences in speech, gesture, and conceptualization. In Proceedings of the Twenty First Annual Con‑ ference of the Cognitive Science Society, M. Hahn & S. C. Stoness (eds.), 507–512. Mahwah NJ: Lawrence Erlbaum. Pahomov, G. 1977. Bull, boar, and orbit/trajectory: On presenting the verbs of motion. Russian Language Journal 31: 1–5. Pavlenko, A. 2002. Emotions and the body in Russian and English. Pragmatics and Cognition 10: 201–36. Pavlenko, A. & Driagina, V. 2007. Russian emotion vocabulary in American learners’ narratives. Modern Language Journal 91: 213–234. Shay, E. & Seibert, U. (eds.). 2003. Motion, Direction and Location in Languages: In Honor of Zygmunt Frajzyngier. Amsterdam: John Benjamins. Shilova, V. V. 2002. Semanticheskaia klassifikatsiia glagolov dvizheniia nenetskogo iazyka. Lin‑ guistica Uralica 38(4): 255–270.

Semantic composition of motion verbs in Russian and English 223

Slobin, D. I. 2003. Language and thought online: Cognitive consequences of linguistic relativity. In Language in Mind: Advances in the Investigation of Language and Thought, D. Gentner & S. Goldin-Meadow (eds.), 157–191. Cambridge MA: The MIT Press. Slobin, D. I. 2004. The many ways to search for a frog. In Relating Events in Narrative: Typologi‑ cal and Contextual Perspectives, S. Stromqvist & L. Verhoeven (eds.), 219–257. Mahwah NJ: Lawrence Erlbaum. Slobin, D. I. 2005. Relating events in translation. In Perspectives on Language and Language De‑ velopment: Essays in Honor of Ruth A. Berman, D. Ravid & H. B. Shyldkrot (eds.), 115–129. Dordrecht: Kluwer. Slobin, D. I. 2006. What makes manner of motion salient? In Space in Languages: Linguistic Systems and Cognitive Categories, M. Hickmann & S. Robert (eds.), 59–81. Amsterdam: John Benjamins. Slobin, D. I. & Hoiting, N. 1994. Reference to movement in spoken and signed languages: Typological considerations. Proceedings of the Twentieth Annual Meeting of the Berkeley Lin‑ guistics Society: 487–505. Stilman, L. 1951. Russian Verbs of Motion: Going, Carrying, Leading. New York NY: King’s Crown Press. Strömqvist, S. & Verhoeven, L. (eds.). 2004. Relating Events in Narrative: Typological and Con‑ textual Perspectives. Mahwah NJ: Lawrence Erlbaum. Sussex, R. & Cubberley, P. 2006. The Slavic Languages. Cambridge: CUP. Talmy, L. 1985. Lexicalization patterns: semantic structure in lexical forms. In Language Typol‑ ogy and Syntactic Description, T. Shopen (ed.), 57–149. Cambridge: CUP. Talmy, L. 1991. Path to realization: A typology of event conflation. Proceedings of the Berkeley Linguistic Society 17: 480–519. Talmy, L. 2000. Toward a Cognitive Semantics, Vol. I & II. Cambridge MA: The MIT Press. Timberlake, A. 2004. A Reference Grammar of Russian. Cambridge: CUP. Verhoeven, L. & Strömqvist, S. (eds.). 2001. Narrative Development in a Multilingual Context. Amsterdam: John Benjamins. Vinogradov, V. V. (ed.). 1960. Grammatika russkogo iazyka. Moscow: Izd-vo Akademii nauk SSSR. Wade, T. 2000. A Comprehensive Russian Grammar. Malden MA: Blackwell. Ward, D. 1965. The Russian Language Today: System and Anomaly. London: Hutchinson University Library. Wienold, G. 1995. Lexical and conceptual structures in expressions for movement and space: With reference to Japanese, Korean, Thai, and Indonesian as compared to English and German. In Lexical Knowledge in the Organization of Language, U. Egli, P. E. Pause, C.Â€Schwarze, A. Von Stechow & G. Wienold. (eds.), 301–340. Amsterdam: John Benjamins. Yu, L. 1996. The role of L1 in the acquisition of motion verbs in English by Chinese and Japanese learners. Canadian Modern Language Review 53: 191–218. Zalizniak, A. A. & Shmelev, A. D. 2000. Vvedenie v russkuiu aspektologiiu. Moscow: Iazyki russkoi kul’tury.

224 Victoria Hasko

Appendix. Abbreviations com del inc ipfv non pfv uni

comitative delimitative inchoative imperfective non-unidirectional perfective unidirectional

chapter 9

Motion events in Polish Lexicalization patterns and the description of Manner* Anetta Kopecka

Max Planck Institute for Psycholinguistics, Nijmegen, the Netherlands

This chapter investigates the expression of motion events in Polish (a satel‑ lite-framed language) and examines the semantic granularity of Manner in the descriptions of motion in written prose. It shows that despite its typological similarity to English in encoding Manner in the main verb and Path in a satellite Polish does not exploit its satellite-framed pattern as productively as English. The types of fine-grained Manner components that Polish lexicalizes in the verb (e.g., velocity, motor pattern, attitude, etc.) are less diverse than those found in English (cf., Slobin 2004a, 2004b, 2006) and the lexical repertoire of Manner verbs is smaller. Furthermore, although many Manner verbs can combine freely with Path satellites, and occur in directed motion constructions, for some of them such combinations are much more restricted. Therefore, to elaborate the expression of Manner, Polish relies on modifying expressions and tends to distribute Manner information across the sentence between the main verb and a subordinate clause or some adverbial expression.

1.

Introduction

As is well known, languages differ considerably in the types of structural devices they use to express motion and in the ways they assign semantic information to lexical and grammatical material. Following Talmy’s seminal typology of verbframed and satellite-framed languages (Talmy 1985, 1991), recent research has * I would like to thank Bhuvana Narasimhan for inspiring discussions on issues discussed in this paper. I am also grateful to Victoria Hasko, Diana Lewis and two anonymous reviewers for helpful comments and suggestions. This work was supported by the European Commission Marie-Curie Fellowship. I am solely responsible for views and results presented in this paper.

226 Anetta Kopecka

contributed significantly to our understanding of cross-linguistic variation in this domain and of the effects of language-specific patterns on the representation of motion in discourse. It has been shown, for instance, that lexicalization patternsÂ€– that is whether Path is expressed in the main verb or in a verb satellite (a particle or a prefix) – have crucial consequences for narrative style and the ways the semantic dimension of Manner is elaborated in motion descriptions (e.g., Berman & Slobin 1994; Slobin 2004b, 2006; Strömqvist & Verhoeven 2004). Satel‑ lite-framed languages, such as English in (1), in which Path is expressed in a verb particle and the verb slot is free for the encoding of Manner, tend towards greater specification of Manner. By contrast, verb-framed languages, such as French in (2), in which the main verb expresses Path, tend either to omit Manner or to express it periphrastically, in an adverbial expression, to compensate for its absence in the main verb. (1) English (a satellite-framed language) a. (…) a few minutes later Charles plunged through the door. [Tar 423] b. I went down the back staircase – low and claustrophobic, scarcely three feet wide – through the kitchen and butler’s pantry, and out onto the back porch. [Tar 625] (2) French (a verb-framed language) a. Colin descendit du métro, puis remonta les escaliers. ‘Colin got off the metro, then went up the stairs.’ b. Ils sortirent de la chambre et passèrent dans l’entrée. ‘They went out of the room and passed into the entryway.’

[Via 21] [Via 143]

While much research has focused on the impact of typological patterns on discourse, the typological characteristics of languages have also been shown to have consequences for the specificity of the motion verb lexicon, in particular for its size and diversity. Slobin (e.g., 2004a, 2004b, 2006) has noticed that the availability of an open verb slot for the encoding of Manner has led satellite-framed languages, such as English, to develop a larger lexicon of manner-of-motion verbs than verb-framed languages and to lexicalize a greater variety of Manner distinctions in the verb. He has observed that English has several hundred verbs lexicalizing fine-grained semantic components of Manner such as velocity (e.g., run, sprint), motor pattern (e.g., hop, jump), force dynamics (e.g., step, tread) or attitude (e.g., amble, saunter). In contrast, verb-framed languages have a much smaller inventory of Manner verbs, rarely exceeding one hundred. Furthermore, it has been argued that the availability of specific grammatical devices for the encoding of Path allows satellite-framed languages to use a great variety of Manner verbs in expressions of directed motion (e.g., walk, run, saunter, dance), in contrast to verb-framed languages, such as French, in which the use of Manner verbs

Motion events in Polish 227

(e.g., marcher ‘walk’, courir ‘run’) is much more restricted and often limited to the expression of activities (Aske 1989; Beavers 2008; Beavers et al. in press; Horrocks & Stavrou 2007; Levin & Rappaport 1998). Yet despite numerous studies on the expression of motion across languages, the relationship between language-specific patterns and the semantic diversity of Manner verbs, as well as the degree of compositionality of Manner verbs and Path satellites, have not been the subject of much study, and we still lack a systematic cross-linguistic account of such correlations (see however Filipović 2007; Jovanovic & Martinovic-Zic 2004; Matsumoto 2003). In this paper, I investigate the first of these issues and examine the relationship between typology and the semantic variety of Manner verbs in Polish. The combinability of Path satellites and Manner verbs at the constructional level, which requires a distinct (corpusbased) investigation, will be addressed in this paper only indirectly. Polish, like English, is satellite-framed and encodes Manner in the main verb and Path in a satellite (a verb prefix), as in (3). Hence, beyond the fact that they differ in the morphological type of the satellite, English and Polish have a similar typological pattern in which the verb slot is open to encode Manner. The question therefore arises whether Polish is also prolific in its lexicalization of fine-grained components of Manner and whether it exploits the main verb in a productive way, as was predicted for satellite-framed languages by previous studies. To address this question, I will examine the semantic granularity of Manner expressions in the descriptions of motion and analyze both the verbs and the adverbial expressions in order to explore how Manner information is organized in the sentence. (3) a. b.

Prze-lazł przez pociąg między wagonami. through-creep.3SG.M.PST through train.ACC between carriages.INS ‘He crept along the train between carriages.’ [Kon 14] Wicio ze-skoczył z roweru przed Wicio.NOM off-jump.3SG.M.PST off bike.GEN in front of żelazną furtą. iron.ACC gate.ACC ‘Wicio jumped off the bike in front of an iron gate.’ [Kon 22]

In the following, I first present the method and the data selected for this study (§2), and then explore the types of lexical means used to describe motion, with a specific focus on the expression of Manner (§3). I then examine the lexicalization of Manner in the verbs, and analyze the distribution of the fine-grained semantic components of Manner in motion descriptions between verbs and the modifying expressions that accompany them (§4).

228 Anetta Kopecka

2.

Text-based study: Data and annotation

The expression of motion in Polish has not been the subject of much study, in contrast to the large body of research on some other Slavic languages (see e.g., Filipović 2007; Jovanovic & Martinovic-Zic 2004; Stosic 2002 for the in-depth analyses of Serbo-Croatian). Moreover, research exploring the domain of motion in Polish has mainly focused on dictionary-based surveys of verbs (see e.g., Bojar 1979; Grochowski 1973), and relatively little typological and corpus-based work in this semantic domain has been done. This study aims to investigate the expression of motion from a typological perspective and on the basis of motion descriptions in written prose. Two Polish novels were chosen on the basis of the abundant descriptions of motion they contain: Kronika wypadków miłosnych ‘Chronicle of Amorous Accidents’ (Konwicki 1974) and Los powtórzony ‘The destiny repeated’ (Wiśniewski 2004). The analysis was carried out on the descriptions of spontaneous motion of animate and inanimate entities (e.g., biegać ‘run’, pływać ‘swim’) as opposed to caused motion brought about by an external agent (e.g., ciągnąć ‘pull’, pchać ‘push’). In order to examine the semantic diversity of Manner verbs, all the descriptions of motion were extracted from the texts, regardless of the type of construction a given verb occurs in. Hence, the data set selected for the analysis includes non-directed motion constructions, as in (4), where the prepositional phrase (when used) indicates the location of motion, and directed-motion constructions, as in (5) and (6), where the prepositional phrase expresses the Ground toward (or away from) which motion is directed; in this latter case, the Path of motion, typically expressed by a verb prefix, may be unbounded as in (5), or bounded as in (6), where it is indicated by the prefix. (4) Non-directed motion a. Monika biegała po plaży jak oszalała. Monika.NOM run.3SG.F.PST over beach.LOC like mad ‘Monika was running all over the beach like mad.’ [Wis 110] b. Wczoraj włóczyłem się trochę koło torów, w lesie, yesterday roam.1SG.M.PST.REFL little near rails.GEN in forest.LOC nad rzeką. above river.INS ‘Yesterday, I roamed a little near the rails, in the forest, by the river.’ [Kon 82] (5) Directed-motion: unbounded Path a. Jadę do Warszawy, do samej Warszawy. ride.1SG.PRS to Warsaw.GEN to herself Warsaw.GEN ‘I’m going to Warsaw, to Warsaw itself.’ [Kon 7]

Motion events in Polish 229

(6) Directed-motion: bounded Path a. Cecylia pode-szła do gramofonu. Cecylia.NOM under-walk.3SG.F.PST to gramophone.GEN ‘Cecylia walked up to the gramophone.’ b. Alina w-biegła do cukierni. Alina.NOM in-run.3SG.F.PST to pastry-shop.GEN ‘Alina ran into the pastry-shop.’

[Kon 63]

[Kon 85]

All told, the selected corpus consists of 1,429 motion verb clauses (including the main verb and the modifying expression). As we can see in Table 1, the directed motion constructions are the most frequent in the data and include 859 bounded Path expressions, and 422 unbounded Path expressions. The descriptions of nondirected motion comprise 148 occurrences. All the lexical and grammatical means (i.e., verbs, modifiers, satellites, prepositions, case markers) used in the descriptions of motion were annotated according to the Berkeley manual “Coding of Motion Events in Texts” designed by Dan Slobin and his colleagues (February 2005 version). Table 2 illustrates the syntactic and semantic parameters according to which the data were coded. Each construction was coded for the type of the event it expresses, directed or non-directed motion. Verb satellites and prepositional phrases were coded for their type and/or meaning. The mode, tense and aspect of the verb were annotated as well as was the syntactic type of the verb (i.e., whether intransitive or transitive), and its semantics (i.e., whether it lexicalizes Manner or Path or both elements, or whether it is neutral). The modifying expressions were coded for Table 1.â•‡ Types of events in motion descriptions Event type

Directed motion

Non-directed motion Total

bounded Path unbounded Path Token Token frequency

859 60.1%

422 29.5%

148 10.4%

1429 100%

Table 2.â•‡ Data annotation Event type

Satellite PP

non-directed motion prefix directed motion adverb (bounded vs. unbounded Path)

Verb

locative mode directional tense aspect

Verb type

Modifier

Verb/ Modifier

Vmanner Vpath Vpath+manner Vneutral

adverbial nominal comparative participial adjectival descriptive

attitude, effort, noise, posture, rhythm, suddenness, velocity, etc.

230 Anetta Kopecka

their lexical type (i.e., whether it is an adverb, a nominal expression, a prepositional phrase, and so forth). Finally, verbs and modifiers were annotated for the fine-grained semantic component of Manner they lexicalize (e.g., attitude, effort, velocity, etc.). Before going into the details of the data, it is important to point out that, in contrast to the semantic components of Path that have been described in detail (see e.g., Talmy 2000), there has been relatively little work on the semantic components of Manner, and we still lack a unified model for the description of Manner verbs. What is even more problematic is that the semantic dimension of Manner has not been clearly defined yet. Scholars differ on the types of components they consider as part of the Manner dimension and describe Manner verbs at a greater or lesser level of granularity (cf., Ikegami 1969; Miller & JohnsonLaird 1976; Narasimhan 2003; Slobin 2004a, 2004b, 2006; Snell-Hornby 1983). For example, some scholars (e.g., Levin 1993; Slobin 2004b, 2006) consider the conveyance (e.g., vehicle) on which the motion of the agent is dependent to be a semantic component of Manner, whereas others (e.g., Frawley 1992; Ikegami 1969) consider it to be a distinct semantic dimension. In the study presented here, Manner is defined as the feature characterizing how the Figure moves as opposed to the feature characterizing where the Figure moves to or from. This interpretation of Manner encompasses not only different ways of moving the body, but also other semantic dimensions related to the movement. More specifically, it includes various physical (e.g., effort, posture, velocity) and psychological (e.g., attitude or emotional state) features, as well as some more external aspects such as the conveyance and the noise associated with the movement.

3.

Motion descriptions in written texts

Intra-linguistic variation in the expression of motion is widespread, and languages often offer different lexical and morphosyntactic strategies to describe motion, some of which might be more dominant than others (e.g., Filipović 2007; Kopecka 2006; O’Connor 2007; Slobin 2004b). Such is the case with Polish. Since the present study aims to investigate the expression of Manner, I will mainly focus on the lexical means here, namely the verbs and the modifying expressions which are the main locus of such information in this language.

. The lack of a unified descriptive approach is possibly due to the diversity of Manner verbs, which are (as far as we know thus far) much more numerous in the languages of the world than Path verbs.

3.1

Motion events in Polish 231

Motion verbs: Types and tokens

Four semantic categories of motion verbs were found in the data: verbs encoding Manner, verbs encoding Path, verbs conflating both Path and Manner and a neutral verb encoding motion itself with no further specification. There are also a few more complex verbal expressions encoding either Manner or Path of motion. Table 4 below shows the number of types and tokens of different verbs within each semantic category. Predictably, Manner verbs predominate in the descriptions of motion in Polish. This category of verbs is the most diverse and the most extensively used. In particular, 66 different types of Manner verbs were found in the data, representing 1098 tokens, that is 76.8% of all the verbs. All the verbs included in this category are monomorphemic and encode different dimensions of Manner (e.g., biec ‘run’, płynąć ‘swim’, lecieć ‘fly’). Verbs encoding Path form a smaller category. There are 13 different types of Path verbs comprising 198 tokens, that is 13.9% of the data. It is important to note that this category contains two kinds of verbs: (a) monomorphemic verbs such as cofać się ‘move back’, mijać ‘pass by’ or wracać ‘go back, return’, and (b) morphologically complex verbs such as zbliżać się ‘approach’ and oddalać się ‘move away’. The morphologically complex verbs have been formed by simultaneous prefixation and suffixation of non-verbal elements encoding distance and are interpreted in Polish as single lexical units, i.e., despite the formal link that can be established between the simple form and the derived form, the meaning of the prefix is merged with the meaning of the lexical base. The category of verbs encoding both Path and Manner consists of 16 types of verbs and 93 tokens (6.5% of the data). This category includes verbs such as wspinać się ‘climb up’, uciekać ‘run away’ or przewrócić się ‘fall over’ which, diachronically, are composed of a prefix (wz- ‘up’, u- ‘away’, prze- ‘through’, respectively) and a verb stem, but for which the relation between form and meaning is not semantically transparent. Thus, these verbs are perceived in contemporary Polish as wholes rather than as morphologically complex verbs.

Table 3.â•‡ Verb types and verb tokens in motion descriptions Verb type

Vmanner Vpath

Vpath+manner Vneutral

Vmanner expression Vpath expression Total

Type Token Token frequency

â•⁄â•⁄ 66 1098 76.8%

16 93 6.5%

2 8 0.6%

â•⁄ 13 198 13.9%

â•⁄ 1 31 2.2%

1 1 0.1%

â•⁄â•⁄ 99 1429 100%

232 Anetta Kopecka

The data also display one neutral verb, ruszyć ‘move’, occurring in 31 tokens (2.2% of the data). In addition, there are two verbal expressions, uczynić krok ‘make a step’ and postąpić kilka kroków ‘(lit.) step a few steps’, describing Manner, and one verbal expression, zatoczyć łuk ‘describe a circle’, describing Path. However, the use of such complex expressions is not frequent in the data. It is important to note that, like some other Slavic languages, Polish has a few lexical pairs among verbs of motion; in each pair, the so-called ‘determinate verb’ (cf., Jakobson 1971) typically expresses motion in one direction, while the ‘indeterminate verb’ typically expresses motion in random directions or habitual and/or repeated motion (e.g., iść/chodzić ‘walk.DET/INDET’, biec/biegać ‘run.DET/ INDET’, płynąć/pływać ‘swim.DET/INDET’ or wrócić/wracać ‘go back, return.DET/ INDET’). Although the two verbs forming a lexical pair lexicalize the same component of motion – that is, either Manner or Path – they differ in terms of fine aspectual features and have therefore been counted separately here. Overall, the data comprise 99 different types of verbs and/or verbal expressions, representing 1429 tokens in total. Note that this is a fairly good sample, considering the fact that a search on a monolingual dictionary (USJP 2004) returned approximately 140 types of monomorphemic verbs of motion. All the verbs found in the data are listed in Appendix 1. Table 4 provides a survey of the distribution of different types of verbs in the expression of directed and non-directed motion. The greatest lexical diversity was found in the expression of bounded Path events, which are also the most frequent in the data; these included 41 types of Manner verbs, 10 types of Path verbs and 13 types of verbs lexicalizing both Path and Manner. The expression of non-directed motion, which is the least frequent, presents less diversity: there are 30 types of Manner verbs, 2 types of verbs lexicalizing Path and Manner and no verbs lexicalizing Path alone. This distribution indicates that some verbs have more semantic and/or aspectual affinity for some types of events than for others. However, as already mentioned, this issue shall be investigated in a distinct (corpus-based) study and will not be pursued further here. Table 4.â•‡ Number of verb types in the descriptions of different types of events Event type

Vmanner Vpath Vpath+manner Vneutral Vmanner expression Vpath expression Total

directed motion â•… bounded Path 41 â•… unbounded Path 31 non-directed 30 motion

10 â•⁄ 7 –

13 â•⁄ 6 â•⁄ 2

– 1 1

2 – –

– 1 –

66 46 33

3.2

Motion events in Polish 233

Modifying expressions: Types and tokens

Let us now look at the kinds of modifying expressions used in the descriptions of motion. Note, first, that 202 out of 1,429 verb clauses are modified, usually by one but occasionally by two modifiers. This represents about 14% of the descriptions, as shown in Table 5. As we can see in Table 6, various lexical strategies have been used to modify the main verb. The most frequent are (a) adverbs (e.g., powoli ‘slowly’, spiesznie ‘quickly’), which represent 101 tokens (and 49 types), and (b) nominal and/or prepositional expressions, which represent 86 tokens (and 51 types). This latter kind of expression consists of a noun or a nominal phrase marked by a case marker (e.g., biegiem ‘run.INS’) or associated with a preposition (e.g., na palcach ‘on tiptoe.LOC’). Among less frequent modifiers are the descriptive expressions (e.g., rozstawiając szeroko nogi ‘standing astride’) as well as participial (e.g., kulejąc ‘hobbling’), comparative (e.g., jak szalona ‘like a fool’) and adjectival (e.g., wy‑ prostowani ‘straight.M.PL’) expressions. Overall, 212 modifying expressions were found in the descriptions of motion, all of which provide some information about the way the process expressed in the main verb is realized. All the modifiers found in the data are listed in Appendix 2. Table 5.â•‡ Verb clauses in motion descriptions Clause type

Motion verb

Token Token frequency

Motion verb + modifier(s)

1,227 85.9%

202 14.1%

Total 1,429 100%

Table 6.â•‡ Manner modifiers in motion descriptions Expression Adverbial NP/PP Descriptive Participial Comparative Adjectival Total Type Token Token frequency

4.

â•⁄ 49 101 47.6%

51 86 40.6%

10 10 4.7%

8 9 4.2%

3 3 1.4%

3 3 1.4%

124 212 100%

Lexicalization and distribution of Manner

In Talmy’s terms, lexicalization “is involved where a particular meaning component is found to be in a regular association with a particular morpheme” (Talmy 1985:â•›59). However, the relation between the semantic components of motion and the morphological (lexical or grammatical) elements is not one-to-one. Several

234 Anetta Kopecka

semantic components can be lexicalized in one morphological element, just as one semantic component can be distributed among several morphological elements (Talmy 1985, 2000; Sinha & Kuteva 1995). In the following, I first examine the semantic components of Manner Polish lexicalizes in verbs, in order to assess the diversity of motion verbs found in the data. I then consider the distribution of Manner components between the main verb and the modifying expressions to investigate the role played by modifiers in motion descriptions.

4.1

The lexicalization of Manner in the verb

As we have seen earlier (§3.1), Manner verbs are the most frequent category in motion descriptions in Polish, and the variety of types found in the written texts studied represents well the motion-verb lexicon of this language. The question to be investigated now is how diverse and expressive Polish Manner verbs are. Table 7 below shows all the Manner verbs found in the data (including verbs that conflate both Manner and Path), and classifies them according to the sorts of finegrained semantic elements they lexicalize. Among all the verbs, the most numerous are verbs depicting “ways of walking”. There are about 30 different verbs (some far more frequent than others) lexicalizing a semantic component such as effort (e.g., toczyć się ‘(lit.) roll’ meaning ‘walking heavily’), attitude (e.g., stąpać ‘step’ meaning ‘walking steadily with attention’), unsteadiness (e.g., chwiać się ‘stagger, wobble’), and aimlessness (e.g., łazić ‘dawdle, stroll’). The second group in terms of relative diversity is formed by verbs lexicalizing “velocity”. There are about 15 different verbs depicting different degrees of swiftness, including quick motion (e.g., biec ‘run’), fast motion away from a danger (e.g., uciekać ‘run away’), motion with great speed and energy (e.g., gnać ‘speed’), and so forth. There is also a relatively numerous group of verbs lexicalizing “medium” and relating to the environment through which or in which the Figure moves, such as air (e.g., lecieć ‘fly’), water (e.g., płynąć ‘swim, navigate’) or conveyance (e.g., jechać ‘ride’). These verbs can be used in reference to vehicles upon which motion may depend (e.g., lecieć ‘fly’ for planes, płynąć ‘swim’ for boats, jechać ‘ride’ for cars) or in reference to human locomotion (płynąć ‘swim’ and lecieć ‘fly’ for motion with great speed).

. The translations of verbs provided in Table 7 are based on a bilingual Polish-English dictionary (Stanisławski 1999). Inevitably, many of the translations are not fully equivalent and only some shades of meaning overlap between Polish verbs and their English counterparts.

Motion events in Polish 235

We shall also emphasize the group of verbs lexicalizing “contact” between the Figure and the Ground. This group comprises several verbs that imply some friction (or lack of it) between the two entities (e.g., czołgać się ‘crawl on one’s belly’, ślizgać się ‘slide, slither’). There are, further, a few smaller groups of verbs lexicalizing such components as “motor pattern” (e.g., skoczyć ‘jump’), “figure” (e.g., sypać się ‘scatter (for a collection of entities)’), “posture” (e.g., siąść ‘sit’), and so forth. Nonetheless, these semantic components seem to be less often lexicalized in verbs, and hence the semantic groups that these verbs form are smaller. The set of verbs that lexicalize both Manner and Path express mainly motion directed down, up or through a reference entity. Among them, there are verbs depicting (a) involuntary motion directed down and having different degrees of impact (e.g., paść ‘fall’, runąć ‘fall, crash’), (b) voluntary motion up that involves some use by the agent of hands and feet (e.g., wspinać się ‘climb’), and (c) voluntary motion through an obstacle involving some force and effort (e.g., przedrzeć się ‘break through’). Although Manner verbs are relatively numerous in Polish, this text-based survey shows nonetheless that they are not as diverse as Manner verbs found in English (cf., Slobin 2004a, 2004b, 2006). While English lexicalizes a great variety of fine-grained components of Manner in the verb, and has a large lexicon of several hundred manner-of-motion verbs (including about fifty verbs depicting different ways of walking), in Polish, the process of lexicalization of fine nuances of Manner in the verb appears to be less productive, and the class of Manner verbs to be less diverse. We may note, for example, that, there are only two verbs lexicalizing “motor pattern” (not only in this data, but also in Polish verb lexicon in general), the determinate verb skoczyć and the indeterminate verb skakać meaning ‘jump’. In English, this semantic class is more diverse and includes verbs such as bound, hop, hurdle, jump, leap, skip, spring and vault. Similar differences can be noticed with respect to ‘travel’ verbs: while Polish has one ‘travel’ verb, podróżować, English has several verbs including journey, tour, travel, trek, and voyage. Furthermore, English has conventionalized the use of some nouns as verbs (e.g., boat, cycle, sail, skate); such a process of conversion represents an important source of lexical innovation in English, augmenting even further the diversity of the motion verb lexicon (cf., Clark & Clark 1979). In Polish, this process is not productive,

. Verbs siąść ‘sit.DET’ / siadać ‘sit.INDET’ are the only posture verbs in Polish that can be used in the expression of spontaneous motion from one place to another. They are typically used to express motion into or out of a place (e.g., wsiąść do pociągu ‘(lit.) to sit into the train’ meaning ‘to get onto the train’, wysiąść z pociągu ‘(lit.) to sit out of the train’ meaning ‘to get off of the train’).

236 Anetta Kopecka

Table 7.â•‡ Manner components in Polish verbs of motion Manner of motion Verbs Walking (ways of) iść ‘walk’ (413), chodzić ‘walk’ (113) effort toczyć się ‘roll’ (8), dotrzeć ‘reach (with some difficulty)’ (6), leźć ‘creep’ (6), gramolić się ‘clamber up, scramble’ (3), dyrdać ‘scurry’ (2), pchać się ‘push one’s way’ (2), brnąć ‘wade’ (1), ciągnąć się ‘crawl, drag oneself ’ (1), człapać ‘shuffle’ (1), grząść ‘sink, wade’ (1), taczać się ‘roll’ (1), telepać się ‘jolt, rattle’ (1), wlec się ‘straggle’ (1) attitude uczynić krok ‘take a step’ (6), stąpać ‘step’ (4), kroczyć ‘tread, pace’ (3), postąpić krok ‘step a step’ (2), (za)kradać się ‘sneak, creep’ (2), przechadzać się ‘stroll’ (2), przejść się ‘take a walk’ (2), spacerować ‘saunter, stroll’ (1), maszerować ‘march’ (1), zdążać ‘keep pace’ (1) unsteadiness chwiać się ‘stagger, wobble’ (2), (za)toczyć się ‘reel, stagger’ (1), kuśtykać ‘hobble, limp’ (1) aimlessness łazić ‘dawdle, stroll’ (2), włóczyć się ‘roam’ (1) Velocity biec ‘run’ (73), uciekać ‘flee’ (18), biegać ‘run’ (16), rzucić się ‘rush’ (14), pędzić ‘run, speed’ (8), uciec ‘flee’ (8), gnać ‘speed’ (5), mknąć ‘rush, dash along’ (3), rwać się ‘tear, rush’ (3), gonić ‘speed’ (2), puścić się ‘set off, dart’ (2), spieszyć się ‘hurry’ (2), ścigać (się) ‘pursue, chase’ (2), wyprzedzić ‘outpace’ (1), szarpnąć się ‘jerk, tear’ (1) Medium air (or velocity) lecieć ‘fly, run quickly’ (51), latać ‘fly, run quickly’ (12) water płynąć ‘swim, navigate’ (4), pływać ‘swim, navigate’ (4), nurzyć się ‘plunge, immerse’ (3), nurzać się ‘plunge, immerse’ (1) conveyance jechać ‘ride’ (173), jeździć ‘ride’ (46) Contact sunąć (się) ‘glide, slide’ (12), czołgać się ‘crawl’ (3), pełzać ‘creep’ (2), suwać się ‘glide, slide’ (2), pełznąć ‘creep’ (1), ślizgać się ‘slide, slither’ (1) Motor pattern skoczyć ‘hop, jump’ (33), skakać ‘hop, jump’ (1) Figure sypać się ‘scatter’ (2), lać ‘flow’ (1), rozproszyć się ‘disperse, scatter’ (1) Posture siadać ‘sit’ (14), siąść ‘sit’ (6) Dancing tańczyć ‘dance ‘ (8), walcować ‘waltz’ (1) Circuition krążyć ‘circle, hover’ (4) Travel podróżować ‘travel’ (1) Other (-)nieść się ‘carry oneself ’ (2), wyłonić się ‘appear, emerge (unexpectingly)’ (1), drgnąć ‘budge’ (1) Manner + Path falling paść ‘fall’ (25), padać ‘fall’ (5), runąć ‘fall, crash’ (5), zwalić się ‘fall, tumble down’ (1), przewrócić się ‘tumble down’ (3) climbing wspinać się ‘climb up’ (14), wspiąć się ‘climb up’ (3), wdrapywać się ‘climb, clamber’ (1) getting through (prze)drzeć się ‘break through’ (1), (prze)dzierać się ‘break through’ (1)

Motion events in Polish 237

and there are very few verbs (e.g., żeglować ‘sail’, saneczkować ‘sledge’) derived from nouns and encoding vehicles.

4.2

The distribution of Manner between the verb and the modifier

Let us now examine the use of modifying expressions with verbs of motion. The use of Manner modifiers in descriptions of motion is a very common strategy employed by both verb-framed and satellite-framed languages. However, Manner modifiers appear to have a different function in the two types of language. Through a close examination of English (a satellite-framed language) and Turkish (a verb-framed language), Özçalişkan and Slobin (2003) and Slobin (2004b) observed that in Turkish, the sole function of the modifier is to “compensate” for the absence of Manner in the main verb, whereas in English, such expressions “qualify” Manner verbs and thereby enhance the semantic granularity of Manner. Since different types of verbs have been found in the Polish data (Manner verbs, Path verbs and so forth) the question arises then as to the function of modifying expressions in descriptions of motion in Polish. Namely, are the modifying expressions more likely to accompany Path verbs (or the neutral verb of motion) to “compensate” for the absence of Manner in these verbs, or are they more likely to enhance Manner expressivity by accompanying Manner verbs?

4.2.1 The distribution of modifiers and their semantics Let us look first at the distribution of modifiers. The results in Table 8 show that the modifying expressions in the data are associated with different types of verbs: Manner verbs, Path verbs, verbs lexicalizing Manner and Path, and the neutral verb of motion. Among all the types, this last is the most frequently modified: 45.2% of the instances (14 out of 31 of tokens) have an additional Manner specification. Manner verbs represent the second type in terms of the frequency of Manner modification received: 14.3% (158 out of 950 tokens) of Manner verbs are combined with a modifier. Interestingly, Path verbs are among the least often modified: 9.6% (19 out of 179 tokens) of this type have a modifying expression.

Table 8.â•‡ Distribution of Manner modifiers with verbs of motion Clause type

Vmanner

Vpath

Vpath+manner

Vneutral

Vmanner expression

Verb Verb+modifier Total

â•⁄ 950 (85.7%) â•⁄ 158 (14.3%) 1108 (100%)

179 (90.4%) â•⁄ 19 (9.6%) 198 (100%)

73 (88%) 10 (12%) 83 (100%)

17 (54.8%) 14 (45.2%) 31 (100%)

7 (87.5%) 1 (12.5%) 8 (100%)

238 Anetta Kopecka

Table 9.â•‡ Semantic components conveyed by the modifying expressions Manner component

Number of Examples expressions (tokens)

attitude velocity vehicle posture slowness suddenness effort unsteadiness walking softness noise (or lack of) rhythm leaping other

35 31 30 27 26 20 10 â•⁄ 8 â•⁄ 7 â•⁄ 4 â•⁄ 3 â•⁄ 1 â•⁄ 1 â•⁄ 9

leniwie ‘lazily’, ostrożnie ‘carefully’ szybko ‘quickly’, biegiem ‘at a run’ motocyklem ‘by motorcycle’, dorożką ‘by hackney’ na palcach ‘on tiptoe’, na czworakach ‘on all fours’ wolno ‘slowly’, bez pośpiechu ‘in no hurry’ raptem ‘suddenly’, nagle ‘suddenly’ mozolnie ‘laboriously’, ciężko dysząc ‘breathing heavily’ kulejąc ‘hobbling’, potykając się ‘stumbling’ pieszo ‘on foot’, piechotą ‘by foot’ miękko ‘softly’, lekko ‘lightly’ cicho ‘silently’, w straszliwym łoskocie ‘with a terrible rumble’ w rytmie tanga ‘in a tango rythm’ zeskakując ‘jumping off ’ po omacku ‘in the dark, intuitively’, na oślep ‘without seeing, instinctively’

The modifying expressions convey various details about the way the motion expressed in the main verb takes place. As we can see in Table 9, most of them specify the attitude of the protagonist (35), the velocity of motion (31), the vehicle upon which motion of the protagonist depends (30), posture (27), slowness (26) or suddenness of motion (20). Other semantic components such as, for instance, noise or rhythm accompanying motion are less frequently expressed in the modifying expressions.

4.2.2 The function of modifiers in motion descriptions The distribution of modifying expressions with different types of verbs suggests that, depending on the type of the verb they co-occur with, they might have different functions in the expression of motion. In particular, in the context of the neutral verb or Path verbs, which are semantically underspecified with respect to Manner, the modifiers provide some information about the way the protagonist’s motion expressed in the main verb takes place and thus compensate for the absence of Manner in the verb. This is shown in the examples in (7), where the modifiers accompany the verbs ruszyć ‘move’ and cofnąć się ‘move back.DET’ and convey the velocity of motion.

Motion events in Polish 239

(7) a. b.

Ktoś ruszył szybko w jego stronę (...). someone.NOM move.3SG.M.PST quickly in his.ACC direction.ACC ‘Someone moved quickly in his direction (...).’ [Kon 192] Cofnął się w pośpiechu do kuchni. move.back.3SG.M.PST in hurry to kitchen.GEN ‘He went back to the kitchen in a hurry.’ [Wis 31]

Nevertheless, in the context of such semantically underspecified verbs, the modifying expressions may also be used to foreground Manner information. Consider the examples in (8) below. In (8a), the verb wrócić ‘go back.DET’ is accompanied by the adverb pieszo meaning ‘on foot’, and in (8b), the verb wracać ‘go back.INDET’ by a noun biegiem ‘at a run’. As we have seen earlier, Polish has in its lexical repertoire verbs conflating these components, namely iść ‘walk.DET’ / chodzić ‘walk. INDET’ and biec ‘run.DET’ / biegać ‘run.INDET’ that could be used to depict similar events. However, while these latter verbs typically background Manner information, the use of an alternative modifying strategy foreground Manner component, thereby putting more emphasis on how the motion expressed in the main verb is realized (see Talmy 2000). (8) a. b.

Pieszo wrócił do muzeum on foot return.DET.3SG.M.PST to museum.GEN ‘(lit.) He went back to the museum on foot.’ [Wis 72] Wicio biegiem wracał pod Ostrą Bramę. Wicio.NOM run.INSTR return.INDET.3SG.M.PST under Ostra Brama.ACC ‘Wicio was going back to Ostra Brama at a run.’ [Kon 91]

In the context of Manner verbs, the contribution of the modifying expressions appears to be more subtle and may depend on the semantic specificity of the verb they accompany. First, as is the case in English (cf., Özçalışkan and Slobin 2003), in Polish the main function of the modifying expressions in the context of Manner verbs is to elaborate the Manner expressed in the main verb, as in the examples in (9) in which the adverbial expressions niepewnie ‘hesitantly’ and chyłkiem ‘furtively’ provide further information about the way the protagonist’s motion takes place. (9) a.

Wicio pod-szedł niepewnie, kłaniając się Wicio.NOM up-walk.DET.3SG.M.PST hesitantly bowing z daleka. from distance.GEN ‘Wicio approached hesitantly, bowing from a distance.’ [Kon 102]

. Ostra Brama is the name of a city gate.

240 Anetta Kopecka

b. Wtedy chyłkiem w-biegł do pokoju Lowka. then furtively in-run.3SG.M.PST to room.GEN Lowka.NOM ‘Then Lowka ran furtively into the room.’ [Kon 156]

In other cases, the modifying expression seems to elaborate and to compensate, at the same time, for the absence of a fine-grained semantic component in the Manner verb itself. Consider the examples in (10). In both examples, the main verb expresses Manner of motion, namely walking and jumping, whereas the modifier specifies how this motion is carried out: in (10a), it indicates that the process of walking is realized on tiptoe, and in (10b), it indicates that the process of jumping is realized on one leg. Note that, while Polish distributes these semantic components between the main verb and the modifier, English offers verbs – tiptoe and hop – which conflate such semantic elements. (10) a. b.

Potem w-szedł na palcach do domu (...) then in-walk.3SG.M.PST on tiptoe.LOC to house.GEN ‘Then he tiptoed into the house (…)’ [Kon 218] (lit.) ‘Then he walked on tiptoe into the house (...)’ Skoczyliśmy na jednej nodze do szpitalnej kaplicy. jump.1PL.M.PST on one.LOC leg.LOC to hospital.ADJ.GEN chapel.GEN ‘We hopped to the hospital chapel.’ [Kon 54] (lit.) ‘We jumped on one leg to the hospital chapel.’

Finally, the distribution of Manner between the main verb and a modifying expression may also be driven by combinatorial constraints. Although the process of combination of Manner verbs with Path expressions is very productive in Polish, these combinations are not free for all the Manner verbs (cf., Narasimhan et al. 2006). Thus, some verbs, such as, for example, zataczać się ‘stagger’ or kuleć ‘hobble’ do not combine with Path prefixes and cannot occur in the slot of the main verb in directed motion constructions. When used in the expression of directed motion, they are typically realized in their nonfinite participial form, as in the examples in (11), in which they modify the verbs iść ‘walk.DET’ and biec ‘run. DET’ respectively. (11) a. (…) zataczając się pode-szła do płyty kuchennej. staggering REFL under-walk.3SG.F.PST to stove.GEN kitchen.GEN ‘She staggered to the kitchen stove.’ [Kon 91] (lit.) ‘She walked staggering up to the kitchen stove.’ b. Pies pod-biegł kulejąc (...). dog.NOM under-run.3SG.M.PST hobbling ‘The dog hobbled up [to me]) (…).’ [Kon 50] (lit.) ‘The dog ran hobbling up [to me] (…)’

Motion events in Polish 241

Note that, while in Polish these Manner verbs occur in nonfinite form (subordinate to the main verb), English verbs such as stagger and hobble do combine with Path expressions and (like most of the verbs lexicalizing Manner) can occur in the slot of the main verb in directed motion constructions. In Polish, the possibility of combining a particular Manner verb with a particular Path expression seems to depend on the semantic properties of the verb. Some verbs that have a very salient Manner meaning appear to be more constrained in their combinatorial possibilities than verbs in which the Manner meaning is less specific. For example, verbs encoding aimlessness (e.g., błąkać się ‘rove’, włóczyć się ‘roam’), leisure (e.g., przechadzać się ‘saunter’, spacerować ‘stroll’) or impairment (e.g., kuleć ‘hobble’, kuśtykać ‘limp’) and that express processes which are extended in time and space cannot occur in the slot of the main verb in directed motion constructions and cannot be conceptualized as change-of-state. This shows that satellite-framed languages can vary not only in the sheer size of the Manner verb lexicon and in the range of Manner ingredients lexicalized in verbs, but also in the extent to which their Manner verbs can combine with Path expressions.

5.

Conclusion

This paper has addressed the relationship between lexicalization patterns and semantic diversity of Manner verbs by investigating descriptions of motion in Polish. The study shows that despite its typological similarity to English in encoding Path in a satellite and Manner in the verb, Polish does not exploit the slot of the main verb as productively as English. In Polish, the size of the Manner verb lexicon, although still substantial, appears to be smaller, and the sorts of fine-grained semantic components of Manner lexicalized in the verbs are less diverse. In particular, the text-based survey has revealed 66 types of Manner as well as 13 verbs lexicalizing Path, 16 conflating both Path and Manner, and one neutral verb of motion. Manner verbs lexicalize a variety of semantic components, such as “ways of walking” – including effort, attitude, unsteadiness, aimlessness – as well as “velocity”, “medium”, “contact” and so forth. Nonetheless, while some semantic classes (e.g., “ways of walking”, “velocity”) are highly diverse and semantically nuanced, others (e.g., “motor pattern”, “posture”) are much less varied and convey less fine distinctions. Therefore, to elaborate the expression of Manner, Polish relies heavily on modifying expressions, which convey fine-grained shades of meaning and enhance the granularity of Manner expression. This study suggests that two satellite-framed languages can vary importantly in the ways they make use of the satellite-framed pattern. Thus, it appears that Polish and English differ not only in the diversity and semantic specificity of Manner

242 Anetta Kopecka

verbs, but also in the combinability of Manner verbs with Path expressions. While English can combine a wide variety of Manner verbs in directed motion constructions, in Polish there seem to be restrictions on the use of Manner verbs in such constructions. This suggests that the diversity of Manner verbs and their morphosyntactic combinability cannot be predicted on the basis of the satellite-framed pattern alone. Other factors – semantic, aspectual and constructionalÂ€– which are independent of motion-event typology need to be taken into account, as they may contribute significantly to the semantic granularity of motion expression in a given language. Further intra- and cross-linguistic research based on larger samples of data and on systematic analyses of the relationship between Manner verbs and Path expressions is therefore needed in order to identify these factors and examine their interplay in the expression of motion events.

References Aske, J. 1989. Path predicates in English and Spanish: A closer look. Proceedings of the Fifteenth Annual Meeting of the Berkeley Linguistics Society 1–14. Beavers, J., Levin, B. & Tham, S. W. In press. A Morphosyntactic Basis for Variation in the Encoding of Motion Events. Journal of Linguistics 46(3). Beavers, J. 2008. Scalar complexity and the structure of events. In Event Structures in Linguistic Form and Interpretation, J. Dölling, T. Heyde-Zybatow & M. Shäfer (eds.), 245–268. Berlin: Mouton de Gruyter. Berman, R. & Slobin, D. I. (eds.). 1994. Relating Events in Narratives: A Crosslinguistic Develop‑ ment Study. Hillsdale NJ: Lawrence Erlbaum. Bojar, B. 1979. Opis semantyczny czasowników ruchu oraz pojęć związanych z ruchem. Warszawa: Wydawnictwa Uniwersytetu Warszawskiego. Clark, E. V. & Clark, H. H. 1979. When nouns surface as verbs. Language 55: 767–811. Filipović, L. 2007. Taking About Motion: A Crosslinguistic Investigation of Lexicalization Pat‑ terns. Amsterdam: John Benjamins. Frawley, W. 1992. Linguistic Semantics. Hillsdale NJ: Lawrence Erlbaum. Grochowski, M. 1973. Klasyfikacja semantyczna jednomiejscowych czasowników ruchu współczesnego języka polskiego. Prace filologoczne XXIV: 99–116. Horrocks, G. & Stavrou, M. 2007. Grammaticalized aspect and spatio-temporal culmination. Lingua 117: 605–644. Ikegami, Y. 1969. The Semological Structure of the English Verbs of Motion. Linguistic Automation Project. New Haven CT: Yale University. Jakobson, R. 1971. Selected Writings, vol. 2. The Hague: Mouton de Gruyter. Jovanovic, J. & Martinovic-Zic, A. 2004. Why manner matters: Contrasting English and Serbo-Croatian typology in motion description. In Discourse Across Languages and Cultures, C.Â€L. Moder & A. Martinivic-Zic (eds.), 211–226. Amsterdam: John Benjamins. Kopecka, A. 2006. The semantic structure of motion verbs in French: Typological perspectives. In Space in Languages: Linguistic Systems and Cognitive Categories, M. Hickmann & S.Â€Robert (eds.), 83–101. Amsterdam: John Benjamins.

Motion events in Polish 243

Levin, B. 1993. English verb Classes and Alternations. Chicago IL: The University of Chicago Press. Levin, B. & Rappaport Hovav, M. 1998. Morphology and lexical semantics. In The Handbook of Morphology, A. Spencer & A. Zwicky (eds.), 248–271. Oxford: Blackwell Publisher. Matsumoto, Y. 2003. Typologies of lexicalization patterns and event integration: Clarifications and reformulations. In Empirical and Theoretical Investigations into Language: A Festschrift for Masaru Kajita, S. Chiba et al. (eds.), 403–418. Tokyo: Kaitakusha. Miller G. A. & Johnson-Laird, P. N. 1976. Language and Perception. Cambridge MA: Harvard University Press. Narasimhan, B. 2003. Motion events and the lexicon: a case study of Hindi. Lingua 113: 123– 160. Narasimhan, B., Kopecka, A. & Özyürek, A. 2006. Crosslinguistic variation in motion event en‑ coding: do constructions play a role? Paper presented at the 4th International Conference of Construction Grammar, 1–3 September 2006, Tokyo, Japan. O’Connor, L. 2007. Motion, Transfer, and Transformation: The grammar of change in Lowland Chontal. Amsterdam: John Benjamins. Özçalişkan, S. & Slobin, D. I. 2003. Codability effects on the expression of manner of motion in Turkish and English. In Studies in Turkish Linguistics, S. Özsoy, D. Akar, M. NakipogluDemiralp, E. Erguvanli-Taylan & A. Aksu-Koç (eds.), 259–270. Istanbul: Bogaziçi University Press. Sinha, C. & Kuteva, T. 1995. Distributed Spatial Semantics. Nordic Journal of Linguistics 18: 167–199. Slobin, D. 2004a. How people move. Discourse effects of linguistic typology. In Discourse Across Languages and Cultures, C. L. Moder & A. Martinovic-Zic (eds.), 195–210. Amsterdam: John Benjamins. Slobin, D. 2004b. The many ways to search for a frog: Linguistic typology and the expression of motion events. In Relating Events in Narrative: Typological and Contextual Perspectives, S.Â€Strömqvist & L. Verhoeven (eds.), 219–257. Hillsdale NJ: Lawrence Erlbaum. Slobin, D. 2006. What makes manner of motion salient? Explorations in linguistic typology, discourse, and cognition. In Space in Languages: Linguistic Systems and Cognitive Catego‑ ries, M. Hickman & S. Robert (eds.), 59–81. Amsterdam: John Benjamins. Snell-Hornby, M. 1983. Verb-Descriptivity in German and English: A Contrastive Study in Se‑ mantic Fields. Heidelberg: Winter. Stosic, D. 2002. “Par” et “à travers” dans l’expression des relations spatiales: Comparaison entre le français et le serbo-croate. PhD dissertation, Université de Toulouse Le-Mirail. Strömqvist, S. & Verhoeven, L. (eds.). 2004. Relating Events in Narrative: Typological and Con‑ textual Perspectives. Mahwah NJ: Lawrence Erlbaum. Talmy, L. 1985. Lexicalization patterns: Semantic structure in lexical forms. In Language Typol‑ ogy and Lexical Description, Vol. 3: Grammatical Categories and the Lexicon, T. Shopen (ed.), 36–149. Cambridge: CUP. Talmy, L. 1991. Path to realization: A typology of Event Conflation. Proceedings of the Seven‑ teenth Annual Meeting of the Berkeley Linguistics Society, 480–519. Talmy, L. 2000. Toward a Cognitive Semantics, Vol. II. Cambridge MA: The MIT Press.

244 Anetta Kopecka

Source of the data Novels

Polish English French

[Kon] Konwicki, Tadeusz. 1974. Kronika wypadków miłosnych. Warszawa: Czytelnik. [Wis] Wiśniewski, Janusz. 2004. Los powtórzony. Warszawa: Prószyński i S-ka. [Tar] Tartt, Dona. 2006. The secret history. London: Penguin Red Classics. [Via] Vian, Boris. 1963. L’écume des jours. Paris: Éditions Pauvert.

Dictionnaires

Kurcz, I., Lewicki, A., Sambor, J., Szafran, K. & J. Woronczak. 1990. Słownik frekwencyjny pol‑ szczyzny współczesnej, t. 1&2. Kraków: Polska Akademia Nauk. Stanisławski, J. 1999. Wielki słownik angielsko-polski, 2 volumes. Warszawa: Philip Wilson. Stanisławski, J. 1999. Wielki słownik polsko-angielski, 2 volumes. Warszawa: Philip Wilson. [USJP] Uniwersalny słownik języka polskiego. PWN, Warszawa 2004.

Coding manual

Slobin, D. 2005. Coding Motion Events in Texts. Berkeley Coding Manual.

Appendix 1. The inventory of verbs found in the data NB. Verbs that cannot occur without a prefix are marked by (-). Manner verbs (66 types, 1098 tokens) iść ‘walk’ (413), jechać ‘ride’ (173), chodzić ‘walk’ (113), biec ‘run’ (73), lecieć ‘fly’ (51), jeździć ‘ride’ (46), skoczyć ‘jump’ (33), biegać ‘run’ (16), rzucić się ‘rush’ (14), siadać ‘sit’ (14), latać ‘fly’ (12), sunąć (się) ‘glide, slide’ (12), pędzić ‘run, speed’ (8), tańczyć ‘dance’ (8), toczyć się ‘roll’ (8), leźć ‘creep’ (6), siąść ‘sit’ (6), gnać ‘speed’ (5), krążyć ‘circle, hover’ (4), płynąć ‘swim’ (4), pływać ‘swim’ (4), stąpać ‘step’ (4), czołgać się ‘crawl’ (3), gramolić się ‘clamber, scramble’ (3), kroczyć ‘tread, pace’ (3), mknąć ‘rush, dash’ (3), nurzyć się ‘plunge, immese’ (3), rwać się ‘tear, rush’ (3), (-)kradać się ‘sneak, creep’ (2), (-)nieść się ‘carry oneself ’ (2), przechadzać się ‘stroll’ (2), chwiać się ‘stagger, wobble’ (2), dyrdać ‘scurry’ (2), gonić ‘speed’ (2), łazić ‘creep’ (2), pchać się ‘push one’s way’ (2), pełzać ‘creep’ (2), przejść się ‘take a walk’ (2), puścić się ‘set off, dart’ (2), ścigać (się) ‘pursue, chase’ (2), spieszyć się ‘hurry’ (2), suwać się ‘glide, slide’ (2), sypać się ‘pour (for a collection of entities)’ (2), zatoczyć się ‘reel, stagger’ (1), brnąć ‘wade’ (1), ciągnąć się ‘crawl, drag oneself ’ (1), człapać ‘shuffle’ (1), drgnąć ‘budge’ (1), grząść ‘sink, wade’ (1), kuśtykać ‘hobble, limp’ (1), lać ‘flow’ (1), maszerować ‘march’ (1), nurzać się ‘plunge, immese’ (1), pełznąć ‘creep’ (1), podróżować ‘travel’ (1), skakać ‘sit’ (1), ślizgać się ‘slide, slither’ (1), spacerować ‘stroll’ (1), szarpnąć się ‘jerk, tear’ (1), taczać się ‘roll’ (1), telepać się ‘jolt, rattle’ (1), walcować ‘waltz’ (1), wlec się ‘straggle’ (1), włóczyć się ‘roam’ (1), zdążać ‘keep pace’ (1) Path verbs (13 types, 198 tokens) wrócić ‘go back, return’ (102), wracać ‘go back, return’ (53), cofnąć się ‘move backward’ (13), minąć ‘pass’ (6), skręcić ‘turn’ (6), zbliżyć się ‘approach’ (5), mijać ‘pass’ (4), zbliżać się ‘approach’ (3), oddalać się ‘move away’ (2), cofać się ‘move backward’ (1), oddalić się ‘move away’ (1), przybyć ‘arrive’ (1), zboczyć ‘deviate’ (1)

Motion events in Polish 245

Path+Manner verbs (16 types, 93 tokens) paść ‘fall’ (25), uciekać ‘flee, run away’ (18), wspinać się ‘climb up’ (14), uciec ‘flee, run away’ (8), dotrzeć ‘reach (with some difficulty)’ (6), padać ‘fall’ (5), runąć ‘fall, crash’ (5), wspiąć się ‘climb up’ (3), przewrócić się ‘tumble down’ (3), przedzierać się ‘break through’ (1), wyprzedzić ‘outpace’ (1), zwalić się ‘crash, fall’ (1), rozproszyć się ‘disperse, scatter’ (1), wdrapywać się ‘climb, clamber’ (1), wyłonić się ‘appear, emerge (unexpectedly)’ (1), przedrzeć się ‘break through’ (1) Neutral verb (1 type, 31 tokens) ruszyć ‘to move’ (31) Manner verbal expressions (2 types, 8 tokens) uczynić krok ‘take a step’ (6), postąpić krok ‘step a step’ (2) Path verbal expressions (1 type, 1 token) zatoczać łuk ‘describe a circle’ (1)

Appendix 2. The inventory of Manner modifiers found in the data Adverbs (49 types, 101 tokens) powoli ‘slowly’ (11), wolno ‘slowly’ (9), szybko ‘quickly’ (7), pośpiesznie ‘hurriedly’ (6), raptem ‘suddenly’ (6), pieszo ‘on foot’ (4), ciężko ‘heavily’ (3), mozolnie ‘laboriously’ (3), nagle ‘suddenly’ (3), chyłkiem ‘furtively’ (2), cicho ‘silently’ (2), energicznie ‘energetically’ (2), lekko ‘lightly’ (2), leniwie ‘lazily’ (2), ostrożnie ‘carefully’ (2), spiesznie ‘hastily’ (2), tyłem ‘by walking backwards’ (2), majestatycznie ‘majestically’ (2), lękliwie ‘timidly’ (1), bardzo wolno ‘very slowly’ (1), bezceremonialnie ‘in a familiar fashion’ (1), gniewnie ‘in anger’ (1), gwałtownie ‘impetuously, violently’ (1), jak najwolniej ‘as slowly as possible’ (1), łagodnie ‘gently’ (1), lojalnie ‘loyally’ (1), miękko ‘softly’ (1), mocno ‘strongly’ (1), nerwowo ‘nervously’ (1), niedbale ‘carelessly’ (1), niem‑ rawo ‘sluggishly’ (1), nieoczekiwanie ‘unexpectedly’ (1), niepewnie ‘unsteadily, hesitantly’ (1), niespiesznie ‘without haste’ (1), niezauważenie ‘unnoticeably’ (1), niezdarnie ‘clumsily’ (1), nisko ‘lowly’ (1), obłędnie ‘madly’ (1), ociężale ‘heavily, languidly’ (1), ostro ‘sharply’ (1), ostrożnie ‘carefully’ (1), raptownie ‘suddenly’ (1), rozpaczliwie ‘in distress’ (1), rychło ‘quickly’ (1), sennie ‘sleepily’ (1), spokojnie ‘quietly’ (1), stępa ‘at a canter’ (1), wysoko ‘high’ (1), wytrwale ‘persistently’ (1) Nominal and/or prepositional expressions (51 types, 86 tokens) biegiem ‘at a run, running’ (6), na palcach ‘on tiptoe’ (6), pociągiem ‘by train’ (6), samochodem ‘by car’ (4), w popłochu ‘in panic’ (4), na oślep ‘without seeing, instinctively’ (3), samolotem ‘by plane’ (3), autobusem ‘by bus’ (2), motocyklem ‘by motorcycle’ (2), na czworakach ‘on all fours’ (2), na motocyklu ‘by motorcycle’ (2), na rowerze ‘by bike’ (2), po omacku ‘in the dark, intuitively’ (2), taksówką ‘by taxi’ (2), w pogoń ‘in pursuit’ (2), windą ‘by lift’ (2), wolnym krokiem ‘with a slow step’ (2), bez pośpiechu ‘in no hurry’ (1), całą siłą ‘with all force’ (1), dorożką ‘by hackney’ (1), do ucieczki ‘in flight’ (1), galopem ‘at a gallop, galloping’ (1), kaczym chodem ‘by duck walk’ (1), kilka kroków ‘a few steps’ (1), metrem ‘by metro’ (1), na boki ‘from side to side’ (1), na czterech nogach ‘on all fours’ (1), na jednej nodze ‘on one leg’ (1), na przednie łapy ‘on fore legs (for animals)’ (1), na rowerach ‘on bikes’ (1), na złamanie karku ‘in a headlong rush’ (1), parę kroków ‘a few steps’ (1), piechotą ‘on foot’ (1), podmiejskich pociągów ‘on suburban trains’

246 Anetta Kopecka

(1), policyjną nyską ‘by police car’ (1), prawie biegiem ‘almost running’ (1), spacerem ‘at a stroll, strolling’ (1), strużką ‘in a trickle’ (1), szybkim krokiem ‘with a quick step’ (1), w milczeniu ‘in silence’ (1), w pośpiechu ‘in a hurry’ (1), w rytmie tanga ‘in a tango rhythm’ (1), w straszliwym łoskocie ‘with a terrible rumble’ (1), wężowym ruchem ‘with a snake movement’ (1), z całych sił ‘with all force’ (1), z furią ‘in a fury’ (1), z nadludzkim wysiłkiem ‘with an superhuman effort’ (1), zakosami ‘in zigzags’ (1), ze skrzypieniem ‘with a creak’ (1), ze wstrętem ‘with repugnance’ (1), z wysiloną swobodą ‘with forced ease’ (1) Descriptive expressions (10 types, 10 tokens) chwiejąc się na nogach ‘wobbling on one’s legs’ (1), ciężko dysząc ‘breathing heavily’ (1), czepiając się ściany ‘hanging on the wall ’ (1), głową do przodu z góry ‘with the head forward’ (1), jakbym miał coś ważnego ‘as if he had something important’ (1), nie dotykając ziemi ‘not touching the ground’ (1), rozstawiając nogi ‘astride’ (1), w nie najlepszym szyku ‘not in the best order’ (1), wysoko podnosząc łapy ‘lifting high its paws’ (1), z odchyloną do tyłu głową ‘with the head drawn back’ (1) Participial expressions (8 types, 9 tokens) kulejąc ‘hobbling’ (2), klucząc ‘dodging’ (1), kryjąc się ‘hiding’ (1), potykając się ‘stumbling’ (1), przedzierając się ‘breaking through’ (1), skradając się ‘slinking’ (1), zataczając się ‘staggering’ (1), zeskakując ‘jumping down’ (1) Comparative expressions (3 types, 3 tokens) jak oszalały ‘like mad’ (1), jak szalona ‘like a fool’ (1), jak cielę ‘like a calf ’ (1) Adjectival expressions (3 types, 3 tokens) szybkich ‘quick.Adj’ (1), wyprostowani ‘straight.Adj’ (1), żywiej ‘more vividly.Adj’ (1)

Appendix 3. Abbreviations acc adj det indet ins f gen loc m nom prs pst pl sg

accusative adjective determinate indeterminate instrumental feminine genitive locative masculine nominative present past plural singular

chapter 10

The importance of being a prefix Prefixal morphology and the lexicalization of motion events in Serbo-Croatian* Luna Filipović

University of Cambridge

The focus of this paper is the expression of motion events in Serbo-Croatian in the context of Talmy’s (1985) typology. Attested electronic corpus data and extensive dictionary data have been used for the purpose of analysis. Prefixes and two morphosyntactic processes that they are responsible for play the crucial role here. These processes are termed combinatory potential and morphological blocking. They guide verb choice (Path vs. Manner verb), the presence of detail regarding Path of motion (whole vs. partially expressed Path), the presence or absence of Manner elaboration, and the frequency of manner of motion verbs in general. Consequences for cross-linguistic contrasting and the study of acquisition and translation are also discussed.

1.

Introduction: Where the journey begins

The starting point in a study of lexicalization is to choose the level at which to ask a research question, which could vary from language-specific, language-family specific, or possibly universal. I chose to start from language-specific issues on this occasion, indicating how the methodology and results could prove rewarding when applied in broader cross-linguistic contexts. A typological inquiry into * I refer to Serbo-Croatian as one language because it is indeed linguistically one language even though nowadays it seems to be politically correct to refer to Serbian and Croatian separately. I have used sources from both languages and the analysis presented here holds for both variants. 1. I express my thanks to the editors, Victoria Hasko and Renee Perelmutter, for their hard work on editorial matters as well as comments regarding this paper in particular. All potential errors that may have still remained are exclusively mine. 2. “Lexicalization” refers to rendering experience into language (cf., Talmy 1985).

248 Luna Filipović

whether Serbo-Croatian neatly fits in the typological profile ascribed to it (cf., Talmy 1985; Slobin 1996, 1997) brought interesting questions to the forefront, namely what role of frequency and use of prefixed manner of motion verbs plays in the lexicalization of motion events, and whether similar patterns and restrictions could be found in other Slavic languages (cf., Filipović 2007a). Talmy (1985) divides all the languages into two groups: verb-framed and satellite-framed. Lexicalization of motion events was one of the central features of Talmy’s semantic typology of languages, because of the universality and ubiquity of motion in language and experience. A motion event consists of a number of components, namely Figure, Manner, Path, and Ground. In the expression Mary walked into the room, Mary refers to the Figure, walked to the Manner of motion, into to the Path, and the room to the Ground component. The central component is the Path of motion because motion is essentially defined as a change of location (cf., Talmy 1985, 1991, 2000). Slavic and Germanic languages are classified as satellite-framed because they typically express this key component of Path outside of the verb, in elements called satellites (e.g., adverbs, prefixes, etc.) and the manner of motion in the verb (as in Serbo-Croatian utrčati ‘into-run’ and English run in/out/across, etc.). Verb-framed, the other typological group, includes Romance languages among others, and lexicalizes Manner in an adjunct and Path in the verb (as in Spanish entrar corriendo ‘enter running’). The consequence of these typological preferences is that speakers of verb-framed languages tend to exclude information about Manner in motion expressions or provide less detailed accounts than speakers of satellite-framed languages in all contexts of language use (cf., Slobin 2006; Filipović 2007b). A number of restrictions and constraints to Talmy’s (1985) typology has been proposed (e.g., Aske 1989; Slobin 1996, 1997, 2000; Filipović 2001, 2002, 2006, 2007a; Zlatev & Yangklang 2002), but the crucial postulates are still viable, whereby more room is provided for certain languagespecific features.

2.

Outline of the present study

Serbo-Croatian is classified as a satellite-framed language on a par with English. However, an in-depth study based on extensive corpus and dictionary data (Filipović 2007a) has revealed that it does not fall neatly within this typological group. A number of constraints condition the differences in the lexicalization . Talmy (1985, 2000) differentiates between satellites and prepositions, which I do not because the distinction does not bear relevance to the argument and it is not always tenable either theoretically or practically (cf. Filipović 2007a).

The importance of being a prefix 249

patterns between Serbo-Croatian (and perhaps some other Slavic languages) and the rest of the satellite-framed languages. Not all manner of motion verbs can be freely used in all motion expression, as is the case in English, due to lexical and morphosyntactic restrictions. The central role of prefixes in this context becomes evident when we look at the occurrence of manner of motion verbs in attested examples. Differently-prefixed manner of motion verbs in Serbo-Croatian show different frequencies in dictionary and corpus data, whereby verbs prefixed with certain prefixes are significantly more frequent than others which are nearly (or completely) absent. In addition, prefixed manner verbs are not licensed in all motion expressions. I argue that the restrictions described in this paper affect the lexicalization of motion events in Serbo-Croatian and, as a result, its position within Talmy’s typology. The database for the research presented here consists of both dictionary and corpus data. Two dictionaries were used: one by Matica Srpska and one by the Serbian Academy of Sciences and Arts. The corpus data comes from the Croatian National Corpus on-line. First, a list of all prefixed manner verbs was created based on their presence in the dictionary data. Then I checked for their occurrences in the corpus data using forms for all three genders and two numbers in past perfective tense (perfekat). I noticed in the dictionary data that the number of verbs prefixed by OD-/DO- (typically indicating movement to/from the speaker/scene, thus providing deictic information) is much higher than those prefixed by any other prefix in Serbo-Croatian. I wanted to investigate whether verbs prefixed by other prefixes may still occur in the corpus even though their dictionary presence was unattested. It might have been the case that they simply had not been listed, but that their combination with other prefixes was still possible and perhaps used by speakers. Thus, I first collected all the verbs prefixed by OD-/DO- from the dictionaries, and then checked in both the dictionaries and corpus for confirmation that those verbs could be prefixed by other prefixes. I put prefixes indicating different directions instead of OD-/DO- onto manner of motion verbs. The prefixes considered are: IZ- ‘out of ’, U- ‘into’, PRE- ‘across/over’, PRO- ‘through’, NA- ‘onto’, and POD- ‘under’ because these cover the most common and frequent spatial relations. I then searched the corpus for examples of all prefixed manner verbs, attested and constructed. The first thing I noticed in the corpus was that many verbs that exist in the dictionaries prefixed with OD-/DO- do not appear with other prefixes. Therefore, I was able to establish that if a verb is prefixed by only one kind of prefix, it is . Serbo-Croatian prefixes appear in capital letters. They are morphophonemes that stand for all the phonetically changeable forms, which depend on the initial sound of the verb they precede (e.g., OD- is realized as either od- or ot-).

250 Luna Filipović

prefixed by OD-/DO-. The prefixed verbs that I constructed with prefixes other than OD-/DO- and which did not exist in the dictionaries were also absent from the corpus data. The frequencies of prefixes that form the basis of the present claims are relevant because they confirm the preference for the deictic OD-/DOverbs in Serbo-Croatian motion expressions, even when verbs prefixed otherwise are possible. The presence of the deictic OD-/DO-verbs is significantly higher than those of verbs prefixed otherwise. The frequencies of all items are absolute and not random because I included all the occurrences found in the corpus. Motion expressions in English are not conditioned by verb+preposition combinability as is the case with Serbo-Croatian prefix+verb occurrences. The use of prepositions with manner verbs and their accumulation is unconstrained in English, and any manner verb can combine with any of the prepositions, which is not the case with prefixed manner verbs in Serbo-Croatian. It is not the use of prepositions per se that is restricted in Serbo-Croatian, but rather the prefixed verbs combinability with prepositions. Therefore, general preposition frequencies are not of interest here and are not further discussed. I analyzed expressions of physical motion only and excluded all metaphorical uses of motion verbs and senses of prepositions that did not relate to the direction of motion. The reason why I carried out a corpus search on selected forms (prefixed perfective verbs in Serbo-Croatian) is a principled one. These forms are typically used in descriptions of experienced events in all kinds of narrative. Perfekat (past perfective) is the most frequent past tense. It also contrasts neatly with English past simple tense forms, which have been the subject of much research in motion lexicalization, thus providing fruitful ground for cross-linguistic comparisons. Perfekat in Serbo-Croatian and past simple tense in English are by no means absolute equivalents, but many of their uses coincide enough for the purposes of this study. These forms are also the key data which offer the relevant insights into the intratypological differences that are the focus of my study and the main support for my arguments. Iterative and habitual meanings are not included, because they are seen as a separate topic for research. . This freedom may be universally restricted depending on vector constraints; cf., Bohnemeyer (2003) for the proposed “argument uniqueness constraint” (AUC), which concerns the semantic uniqueness of Path argument roles that can be mapped onto Ground-denoting expressions within single clauses. It requires each expression of a semantic argument expressed in a single clause to be assigned a unique argument role (Bohnemeyer 2003:â•›110). AUC also affects the coding of Path roles in such a way that in simple clauses only one Path role (e.g., source, goal, route, etc.) can be assigned to exactly one Path-expressing element. This is an apparently universal constraint on the coding of motion events at the syntax/semantics interface (ibid.). This constraint blocks examples like: *Sally walked across the hall to the entrance out of the library cross-linguistically (ibid.).

The importance of being a prefix 251

3.

Prefixation in a spatio-temporal network

3.1

Combinatory potential

Prefixation in Serbo-Croatian has been studied in the past as a part of morphology, precisely the domain of word formation. However, the frequency and use of prefixes and prefixed verbs is largely unaccounted for. Unlike English, SerboCroatian does not allow unrestricted use of manner verbs in motion expressions, due to the processes I termed morphological blocking and combinatory potential (Filipović 2002, 2006, 2007a). Combinatory potential is a scalar relation that reflects the combinability of prefixes with verbs and prepositions that follow them. It is attested in both dictionary and corpus data, based on the number of prefixed verbs registered in the dictionary (i.e., types) and the frequency of their occurrence in the corpus (i.e., tokens). Additional complexity of Serbo-Croatian is documented by the fact that manner verbs can be both prefixed and unprefixed in motion expressions. Unprefixed manner verbs are used in expressions referring to events in which no physical boundary is being crossed by the moving Figure, as in ‘John ran in the park’. They are unrestricted in Serbo-Croatian; however, the same also holds true for verb-framed languages (e.g., Romance languages), thus this is not an exclusive feature of satellite-framed languages. Namely, it has been observed that spatial features of events, such as whether there was a bound‑ ary on the Figure’s path or not, affect the choice of expressions (cf., Aske 1989; Slobin 1996; 1997; 2000). The relevant typological differences are most noticeable in expressions that refer to one particular situation type, known as boundarycrossing. Boundary-crossing occurs when there is an explicit change of spatial configuration, as in He ran out of the house into the garden. Manner verbs are blocked in verb-framed languages in such situations, and a verb expressing Path is used instead. The manner component gets expressed in a gerund (e.g., Spanish: Salió de la casa corriendo. (‘He exited the house running’)). In expressions of nonboundary-crossing, manner of motion verbs, usually those more common and frequent, can be used in both verb-framed and satellite-framed languages (e.g., Spanish: Corrieron hacia la casa. (‘They ran towards the house’)). Nevertheless, on the whole, the information about Manner is often absent and less elaborate if present in verb-framed languages, as Slobin has shown on numerous occasions (cf., Slobin, 1996, 1997, 2000, 2003, 2006). For instance, it is habitually omitted in translation from English into Spanish in more than 50% of cases (cf., Slobin 1996, 1997). Therefore, the situation where the typological predictions hold most conspicuously is that of boundary-crossing (as in John ran into the park), where Romance languages demonstrate their preference for directional verbs and

252 Luna Filipović

omission of information about Manner in both experimental and naturalistic data, unlike English, for example, where information on Manner is omnipresent and detailed (cf., Slobin 1996, 1997, 2000, 2006). When it comes to expressing boundary-crossing in Serbo-Croatian, I noticed that manner can be expressed in the verb, and that verbs must be perfective in order to render this situation type. All perfective manner verbs are prefixed by prefixes which are also directional, and all such verbs are used to express the crossing of a boundary. There is an important and well-defined cline in prefix use, attested equally both in the verbs prefixed in the dictionary data (types) and in the number of examples found in the corpus (tokens). The type and tokens frequencies coincide because more frequent types have more numerous tokens. I grouped the verbs according to the 5 basic spatial configurations they express, and these are placed on the following cline from most frequent to least frequent: OD-/DO-, IZ-/U-, PRO-, PRE-, NA-/ POD-. The frequencies are given in Table 1, in descending order. PRO- is the only prefix that can be used to refer to movement involving two different spatial configurations. The interpretation of the prefix PRO- (i.e., whether ‘through’ or ‘past’) is resolved by the prepositions that normally follow PRO-verbs, which differ depending on this contrast in meaning. Practically speaking, a PRO-verb can be followed by the preposition pored ‘past’ or kroz ‘through’, and the meaning of the verb varies accordingly. This explains why PRO-verb types and tokens in Table 1 seem almost as numerous as IZ-/U-verbs, which are higher on the proposed cline, because, in effect, they comprise two different spatial configurations and thus are placed lower. IZ-/U-verbs are used in expressing movement within the same kind of spatial configuration, albeit in opposing directions (out vs. into a three-dimensional space). For convenience, PRO1- and PRO2- data are given together since this does not affect the point made. The great majority of manner verbs in the dictionary and corpus data are prefixed with OD-/DO-prefixes. There are significantly fewer verbs in both the Table 1.â•‡ Prefix cline: individual frequencies Prefixed manner verbs

Types

Tokens

DOODIZUPRO1- and PRO2PRENAPODTotal

118 118 â•⁄ 43 â•⁄ 43 â•⁄ 41 â•⁄ 20 â•⁄â•⁄ 5 â•⁄â•⁄ 3 391

â•⁄ 423 (31.9%) â•⁄ 396 (29.9%) â•⁄ 169 (12.7%) â•⁄ 117 (8.8%) â•⁄ 135 (10.2%) â•⁄â•⁄ 74 (5.6%) â•⁄â•⁄ 10 (0.7%) â•⁄â•⁄â•⁄ 2 (0.2%) 1326 (100%)

The importance of being a prefix 253

dictionary and corpus data towards the end of the cline. The explanation for this is based on the fact that the prefixes on the top of the cline can be attached to more verbs and employed in more expressions, because of their combinability with different prepositions that follow the verbs, the number of which decreases down the cline. This is the feature I termed combinatory potential, and it diminishes progressively down the cline in Table 1. The OD-/DO-verbs have the highest combinatory potential because (a) they can combine with all prepositions, and (b) they encompass the whole Path of motion, as illustrated in the examples (1) and (2). Namely, one OD-/DO-verb followed by different prepositions is enough to express the whole Path, whereas verbs prefixed otherwise express only limited portions of the Path, meaning that more than one verb would be necessary for the expression of Path. The phrase ‘to-the-reference-point’ in the glosses refers to the speaker’s position or scene-setting where the attention of the listener or reader is focused: (1)

U baraku preko se do-teturao into shed opposite REFL ‘to-the-reference-point’-stagger.PST.PFV.3SG.M otud iz logora. from out of camp ‘He staggered into the shed opposite from out there out of the camp.’

(2)

Do-trčao je uz stepenice u ‘to-the-reference-point’-run.PST.PFV.3SG.M be.COP up stairs into spavaću sobu. bedroom ‘He ran up the stairs into the bedroom.’

It is understood that the speaker was positioned upstairs in example (2) and that the movement was towards him/her. Similarly, in (1), the speaker was located in the shed, and the movement was towards him/her. If an OD-verb had been used (as in otrčao) in example (2) instead, it would have been clear that the speaker was positioned downstairs and that the movement was away from the speaker. This is due to the fact that the verbs prefixed with OD-/DO- also express deixis. . The term “deixis” actually refers here to spatial deixis, meaning the lexicalizing information of the position of the speaker/scene, where “scene” refers to the location where narrative attention is focused (cf., Lyons 1991:â•›170; Saeed 2001:â•›173). Similarly, Miller and JohnsonLaird (1976:â•›396) define place deixis as “the linguistic system for talking about space, relative to a speaker’s egocentric origin and coordinate axes”. There are examples of non-deictic uses of OD-/DO-verbs in the corpus, such as Oteturao se od stola do vrata ‘He staggered from the table to the door’ where no egocentric perspective is expressed (i.e., the position of the speaker or listener/reader). However, they amount to only 7% of OD-/DO- meanings. The majority of OD-/DO-verbs are used deictically.

254 Luna Filipović

This seems to be a strong feature of Serbo-Croatian, since an OD-/DO-verb is used more frequently for all directions, even though other prefixes specifying the relevant direction are possible. Compare (3) and (4): (3)

To je rekao i ot-eturao it be.COP said and ‘from-the-reference-point’-stagger.PST.PFV.3SG.M se u gostinjsku sobu. REFL into guest room ‘He said that and staggered away into the guest room.’

(4)

To je rekao i u-teturao se u it be.COP said and into-stagger.PST.PFV.3SG.M REFL into gostinjsku sobu. guest room ‘He said that and staggered away into the guest room.’

Examples like (3) found in the corpus data show that, although a deictically neutral option exists in the language (i.e., the verb in (4)), the deictically marked verb is more frequently used. Similarly, the verb isteturati se ‘stagger out’ also exists, but it is not used in the example (1) because it can only refer to one immediate boundary-crossing. By using one deictic manner verb, all of the relevant pieces of information regarding the Path and Manner of the Figure’s motion can be expressed. In fact, whenever an OD-/DO-verb is used, it is also possible to use two or more verbs prefixed otherwise. This is not done, simply because using one verb is more economical in communication. For instance, if we were to use manner verbs prefixed by non-deictic prefixes, we would have to provide three manner verbs as in (5b), instead of one deictic verb as in (5c) in the translation of the English example in (5a): (5) a. b. c.

John ran out of the yard, across the field and into the forest. Džon je is-trčao iz dvorišta, pre-trčao preko polja i John be.COP out-ran out yard across-ran across field and u-trčao u šumu. into-ran into woods Džon je o-trčao iz dvorišta preko polja John be.COP ‘from-the-reference-point’-ran out yard across field u šumu. into woods

. Cf., Verkuyl (1999) points out the importance and strong preference for these prefixes in Slavic languages, which perform so-called “aspectual slicing”, putting emphasis on the initial/ final point of activities.

The importance of being a prefix 255

Verbs prefixed with OD/DO- can combine with all prepositions. Moreover, a single OD-/DO-verb can accumulate numerous prepositions, which is not the case with verbs prefixed otherwise. Since deictically prefixed verbs are used more frequently than verbs prefixed otherwise, the consequence is that an additional piece of information (i.e., the location of the speaker/scene) is frequently found in translation from a language that does not have deixis as a strong feature of the system (e.g., English) into one that does (e.g., Serbo-Croatian). Very often, no deictic reference is given in the original (English), but exists in the translation into Serbo-Croatian. This normally happens when an adequately prefixed verb does not exist in Serbo-Croatian, as illustrated in the following example in which the prefix indicating upward movement is not licensed: (6)

He rushed up the stairs and into the bedroom. Do-jurio je uz stepenice u ‘to-the-speaker/scene’-rush.PST.PFV.3SG.M be.COP up stairs into spavaću sobu. bedroom

Verbs prefixed with prefixes other than OD-/DO- are normally followed by the prepositions that express the same direction as those prefixes. They refer to a single boundary-crossing and are followed by a single Ground expression. The accumulation of Ground expressions is also one of the characteristics of satellite-framed languages. Serbo-Croatian (and Russian, for example) seems to have restrictions when it comes to this typological feature. For instance, in Russian, accumulation of Ground expressions onto a single verb is also not licensed (cf., Filipović 2007a:â•›21–22), and the Path of motion is dissected by differently prefixed manner verbs for each directional portion (as in the example (5b)). Accumulation of Ground expression is possible in Serbo-Croatian only with OD-/DO-verbs; with verbs prefixed otherwise, it is not licensed, as illustrated in the following examples: (7) *Pre-trčao je ulicu i u poštu. across-run.PST.PFV.3SG.M be.COP street and into post office ‘He ran across the street and into the post office.’ (8) *Is-trčao je iz kuće, preko ulice u prodavnicu. out-run.PST.PFV.3SG.M be.COP out house across street into shop ‘He ran out of the house, across the street into the shop.’

The verbs in (7) and (8) can express only one boundary-crossing event and, therefore, cannot be followed by two or three different Ground elements that indicate two or three different boundaries. This issue does not arise in English

256 Luna Filipović

and represents a significant intratypological distinction. The number of prepositions that can follow the prefixed manner verbs diminishes along the cline in TableÂ€1. IZ-/U-verbs can be followed by iz ‘out’, u ‘into’, na ‘onto’, and kroz ‘through’. PRO-verbs can be followed by kroz ‘through’ and pored ‘past’, reflecting the PRO1 and PRO2 meanings (cf., Table 1). PRE-verbs can be followed by preko ‘across’, iz (‘out’) and u (‘into’) combined together (cf. (9)), and, more rarely, na ‘onto’. NAand POD-verbs can only be followed by na ‘onto’ and pod ‘under’, respectively. These non-deictic verbs are also normally followed by a single Ground element, unlike OD-/DO-verbs. They can be followed by more than one Ground element, but only on determinate occasions. This use is limited to the reference to the Grounds that are in contact (physically or conceptually understood as joined or close together in a spatial continuum), and when only one boundary-crossing occurs, for example, ‘running out of the house into the garden’ (cf., Filipović 2007a:â•›126 for detailed discussion): (9) Pre-trčao je iz svoje u komšijinu baštu. across-ran.PST.PFV.3SG.M be.COP out of his into neighbor’s garden ‘He ran out of his garden into the neighbor’s.’ (10) U-leteo je u sobu kroz prozor. into-fly.PST.PFV.3SG.M be.COP into room through window ‘He flew into the room through the window.’

When it comes to other Slavic languages, Shull (2003) argues that, in Russian, the prefix combines almost exclusively with a single preposition (e.g., OT-with the preposition ot) or with a few principle prepositions, whereas Czech prefixes combine more flexibly with a number of prepositions in more even distributions (ibid.:â•›81). Ferm (1990) confirms that this claim is too restrictive for Russian, and it would be interesting to see whether prefixed verb+preposition combinations reveal similar behavior in the context of motion expressions as those in SerboCroatian. Shull (2003) does not discuss in detail prefixed verb combinability with prepositions or accumulation of prepositions in motion expressions. It is possible that one could trace the reason for this to the impossibility of using a single prefixed manner verb to express the whole of the Path of motion in Russian and to accumulate Ground elements to a single prefixed manner verb, which would be a very similar situation to the one presented in this section for Serbo-Croatian. I concluded that the more we can say about the Path with fewer elements, the more frequently these elements will occur. The whole Path can be expressed by OD-/DO-verbs; IZ-/U-verbs can express many parts of Path, but not all of them because they cannot combine with all prepositions. At the end of the scale, NA-/POD-verbs combine solely with the phonetically equivalent prepositions

The importance of being a prefix 257

and, therefore, can express only one section of Path. Thus, speakers prefer the use of one verb instead of many in the lexicalization of Path. I showed that these features of the prefixes are not random. For instance, they do not depend on the length of verb root with which they combine, as has been suggested. For example, NA-/POD-prefixes occur with complex (e.g., three syllabic) verb roots, just like OD-/DO-verbs. Semantic complexity may bear more relevance on this occasion. Semantically simpler manner verbs (such as trčati ‘run’ or leteti ‘fly’) occur with all prefixes; whereas semantically more complex verbs have both fewer types and tokens and do not occur with all the prefixes. Nevertheless, this is true for any semantic field in language. Semantically (and morphologically) simpler manner verbs are always more frequent than the complex ones. This fact does not affect my results and claims, because there is no semantic reason why semantically more complex verbs, such as dobazati ‘get to the reference point walking aimlessly’, would occur with OD-/DO- and not with PRE- for example. Furthermore, there is no purely morphological reason (in the sense of word formation) why certain verbs occur with some prefixes and not with others, except that language users develop preferences based on providing more information with fewer means (e.g., one OD-/DO-verb instead of three prefixed otherwise as illustrated in (5c) and (5b) respectively). The preferential use of OD-/DO-verbs and the diminishing frequencies down the cline in Table 1 are dependent on their potential to combine with different prepositions and on the variety of spatial relations they can refer to. The foundations of my argument are usage-based, and they strongly support the relevance of the combinatory potential. It is valuable to note that the analysis of deictic OD-/DO- and other directional prefixes in Serbo-Croatian is aimed at drawing attention to the importance of deixis in the lexicalization process. The deictic meaning of certain prefixes in Serbo-Croatian has not been of much interest and is hardly ever mentioned in some major descriptions of the system of prefixes. It has also not been of major concern in lexicalization research, apart from a few notable studies (e.g., Choi and Bowerman 1992). Choi and Bowerman (1992) consider deixis one of the key elements in the lexicalization process, on a par with Manner and Path. Further restrictions in prefix use in Serbo-Croatian are particularly noticeable when motion along the vertical scale is to be expressed – a situation in stark contrast to its counterpart in English, in which there are no restrictions with respect to the combinations of manner verbs and directional prepositions expressing

. I hereby thank an anonymous reviewer for this suggestion, which prompted me to clarify my arguments. . No mention in two seminal studies on Serbo-Croatian (Belić 1964; Stevanović 1989).

258 Luna Filipović

upward or downward motion. The scarcity of manner verbs prefixed by ‘up/down’ prefixes in Serbo-Croatian (UZ- and S- respectively) again means that OD-/DOverbs will be used; thus, deictic information will be present almost as a default if a manner verb is used.10 Interestingly, a verb expressing just Path (e.g., the equivalents of ‘go up’ or ‘go down’) is used even more often than an OD-/DO-verb in these situations. Deictic information was always present in Serbo-Croatian translations from English, even though it was absent in the original text. This has been documented in a contrastive translation study (Filipović 1999). For example: (11)

The balloon floated up the chimney. Balon je od-lebdeo balloon be.COP ‘from-the-reference-point’-float.PST.PFV.3SG.M uz dimnjak. up chimney

(12)

Children skipped down the hill. Deca su ot-skakutala children be.COP ‘from-the-reference-point’-skip.PST.-PFV.3PL.N niz brdo. down hill

By the same token, the information about Manner is often omitted in translation from English into Serbo-Croatian (cf., Filipović 1999, 2001, 2008). This is due to the fact that an adequately prefixed manner verb does not exist. Expressing Manner out of the verb would involve using a complex paraphrase or adverbs and burden the narrative; thus it is excluded, as in the following examples: (13) … that the flagstaff has been blown down. … da je jarbol srušen. that be.COP flagstaff brought down ‘… that the flagstaff was brought down.’ (missing: by blowing) (14)

The animals rushed to the top of it. Životinje se popeše na njegov vrh. animals REFL ascend.PST.PFV.3PL.N onto its top ‘Animals climbed to its top.’ (missing: in a rush)

10. These morphological peculiarities, regarding the prefixes indicating movement on the vertical scale, are discussed in more detail in Filipović (2002, 2007a). Particularly interesting are the hybrids of verbs and the prefix S-, which seems to indicate downward motion in verbs like strčati ‘run down’, but also seems to form a hybrid in verbs like skombrljati se ‘fall down heavily’, with the verb stem *kombrljati se being non-existent in modern Serbo-Croatian.

The importance of being a prefix 259

There are many questions related to “ups-and-downs” cross-linguistically that are still waiting to be answered. For example, is the movement along the vertical axis to be treated as boundary-crossing or non-boundary-crossing? Spanish speakers seem to use manner verbs in this situation in contrast to their habitual pattern (e.g., describing a Figure sliding into a swimming pool), and thus perhaps are treating such motion as non-boundary-crossing. It could be spatial continuity or contact with the surface within a motion event itself that could affect its interpretation (cf., Naigles et al. 1998). A possible explanation may lie in the fact that this situation is not viewed as boundary-crossing at all, and that the slide and the swimming pool form a spatially united space (ibid.). Further research, within individual languages and across language families, will hopefully bring important insights in this particular area. Combinatory potential is evidently a phenomenon that connects the levels of morphology, syntax and lexical choices in the expression of events in at least one cognitive domain, that of motion. Future inquiries should focus on exploring whether the same/similar principles guide the lexicalization processes of other cognitive domains and other morphosyntactically similar languages. In this way, starting from broadly defined typological contrasts based on the lexicalization of event components, such as Manner and Path, enabled us to detect subtler differences, some of which are language-specific and some of which may be shared by a number of structurally-different languages. As a result, both similarities and differences become clearer, and studying the consequences of these lexicalization features within other fields, such as language acquisition or translation studies, becomes more straightforward.

3.2

Morphological blocking

In this study of lexicalization of motion in Serbo-Croatian, another morphological process determines lexical choices and syntactic structure of the expressions of motion events. In accordance with Talmy’s typology, Serbo-Croatian was expected to have Manner expressed in the verb and Path expressed elsewhere (in this case a prefix and a preposition). I argue (Filipović 2002, 2006, 2007a) that the relevant temporal features of events, previously disregarded, have to be taken in consideration, along with the spatial ones, in the analysis of lexicalization patterns. The reason lies in the fact that not all temporally different situation types can be expressed with manner verbs in Serbo-Croatian. Temporal features of events have not been considered on a par with spatial features, such as boundary-crossing. I established that there is a difference in lexicalization choices within a language, based on what temporal phase of an event

260 Luna Filipović

Table 2.â•‡ Spatio-temporal network for identification of motion situation types Boundary-crossing

Boundary-reaching

Non-boundary-crossing

Change-occurred Moment-of-change No-change

+ (a) + (b) –

+ (c) – (+) –

– – + (d)

→

Spatial → Temporal

is to be expressed. The crucial temporal features are change-occurred, momentof-change, and no-change; their interaction with spatial features is illustrated in Table 2. This feature network helps us distinguish among different types of motion events, and in effect provides tools for a comprehensive analysis of means used to express different types of events in individual languages, without resorting to highly language-specific categories such as aspect. The examples of the relevant situation types are given in (15a–d): (15)

a. b. c. d.

He rolled out of the shelter. He was rolling out of the shelter. He rolled to the shelter. He rolled/was rolling in the shelter.

Boundary-reaching/moment-of-change situation type is marked as (–)+ because it does not refer to a motion event strictly speaking, even though it is possible. Rather, we can say that this situation type is expressed with a different verb category, that of arriving or leaving (in Levin’s 1993 classification), as in I was arriving at London Kings Cross station when my phone rang, where the noun phrase that follows the preposition refers to goal, rather than a boundary that is reached. If we try to construct an example where the object of the preposition actually refers to a boundary (e.g., I was arriving at your door when my phone rang), we can see that the construction sounds rather unusual. If we try to use a manner of motion verb in this construction to refer to this situation type, we can see that the situation type referred to is actually that of non-boundary-crossing/no-change type (still directional, not locational): I was running to your door when my phone rang. Manner of motion verbs in Serbo-Croatian can be prefixed or unprefixed. Prefixed manner verbs are also perfective at the same time, while unprefixed manner of motion verbs are normally imperfective. Not all situation types allow the use of manner verbs in Serbo-Croatian. The constraints within the typology proposed in Filipović (2007a:â•›73–75) specify conditions for the lexicalization of motion events in Serbo-Croatian in the following ways: unprefixed manner verbs are used in the non-boundary-crossing/no-change phase; prefixed manner verbs in the boundary-crossing/change-occurred phase with combinatory potential

The importance of being a prefix 261

constraints; and directional verbs in the boundary-crossing/moment-of-change phase. The examples in (16a–d) illustrate the relevant situation types, which are equivalent to the English examples (15a–d) respectively: (16) a. b. c. d.

U-kotrljao se u sklonište. into-roll.PST.PFV.3SG.M REFL into shelter ‘He rolled into the shelter.’ U-lazio je u sklonište kotrljajući se. enter.PST.IPFV.3SG.M be.COP into shelter rolling REFL ‘He was entering the shelter rolling.’ Do-kotrljao se do skloništa. ‘to-the-reference-point’-roll.PST.PFV.3SG.M REFL to shelter ‘He rolled to the shelter.’ Kotrljao se u skloništu. roll.PST.IPFV.3SG.M REFL in shelter ‘He rolled/was rolling in the shelter.’

The examples in (16) illustrate how morphological blocking restricts the use of manner verbs. The situation in (16a) and (16c) require a perfective verb, whereas in (16b) and (16d) an imperfective verb must be used. Manner verbs can be used in all situations, except (16b), the moment-of-change phase. The examples show a stark contrast between Serbo-Croatian and English, its typological “partner”. Morphological blocking occurs because of the general impossibility to further imperfectivize a perfective manner verb, which would have been necessary in order to express the required Manner and Path. If an unprefixed imperfective manner verb is used instead, the situation type is no longer boundary-crossing/ moment-of-change, but rather changes to the non-boundary-crossing/no-change situation type, as in (16d). Secondary imperfectivization is not carried out in Serbo-Croatian. A few prefixed imperfective manner of motion verbs do exist, but they are used exclusively for conveying iterative and habitual meaning. In the whole corpus used for this study, there was no example of a prefixed imperfective manner verb in an expression of moment-of-change phase. That is why when an imperfective verb is needed in an expression of motion of the type in (16b), manner verbs are blocked and a directional imperfective verb is used instead. In this context, Shull (2003) mentions that, although further imperfectivization exists as a tendency in Czech, for example, in some domains (an example she herself deemed unusual), in motion verbs, prefixed imperfectives have a different stem from the unprefixed imperfectives. She quoted only three of them in her analysis (chodit’ ‘walk’, létat’ fly’ and běhat’ ‘run’). We are not told how productive this process is, and whether all manner of motion verbs are affected. Moreover, this makes us wonder how frequently we actually find the prefixed

262 Luna Filipović

imperfectives in Czech in moment-of-change expressions (especially in past tense reports), and whether they are more common in habitual/iterative use, thus showing restrictions similar to the use in Serbo-Croatian. The same study also claims that Russian, unlike Czech, “tolerates” further imperfectivization in manner of motion verbs (the stem remaining the same) (ibid.:â•›231). It could be the case that their use is limited to habitual or iterative meaning and not used in expressions of the moment-of-change phase, which transpires in the accounts by Gasparov (1990) and Grickat (1966/67). The extent to which this occurs and the productivity/frequency of such forms await more in-depth investigation. Further study is therefore necessary in order to establish how similar or different Slavic languages are when it comes to prefixation and secondary imperfectivization, with respect to lexicalization patterns in motion expressions. We could expect a substantial amount of variation because, as Dickey (2000) confirms, Slavic languages differ significantly in their aspectual systems. Serbo-Croatian occupies a special transitional position between the Eastern group (Czech, Slovakian, Slovenian) and the Western group (Russian, Ukranian, Bulgarian); the criterion is the use of perfective verbs in the Western group and imperfective verbs in the Eastern group for denotation of habitual events. Another trend noticed in Serbo-Croatian is the use of perfective directional verbs, such as ući ‘enter’, izaći ‘exit’, preći ‘cross’, etc., whose frequency is much higher in motion expressions than that of manner verbs on the whole. The total number of perfective directional verbs in motion expressions found in the corpus is 2,483. They are all used in expressions of boundary-crossing/change-occurred where prefixed manner verbs can also be used, although with restrictions delimited by the combinatory potential (cf., Filipović 2007a:â•›132). Directional verbs are morphologically simpler than prefixed manner verbs, and they tend to be used more often, even when an adequate prefixed manner verb exists. The frequency of imperfective directional verbs was not considered at present because, due to morphological blocking as explained earlier, they are used only in expressions of one particular situation type, namely that of boundary-crossing/moment-of-change. On the other hand, unprefixed, imperfective manner verbs were not analyzed either in terms of frequency, because they are exclusively used in non-boundary-crossing/no-change situations. Directional verbs in Serbo-Croatian show an interesting morphological feature; they seem to be formed by prefixation, as in: U- ‘into’ + ići ‘to go’ (imperfective) = ući ‘enter’ (perfective). Then again, there is ulaziti, which is an imperfective form that looks prefixed and means ‘enter’. However, it is not a prefixed form because the stem is not licensed (U- + *laziti). Normally, perfective manner verbs are prefixed and imperfective are unprefixed, sharing the same root. This is clearly not the case with directional verbs, whose imperfective forms are the only option for rendering moment-of-change situation

The importance of being a prefix 263

types because they are not constrained morphologically and semantically the way manner verbs are. Data from the dictionaries and the corpus based on the modern usage of Serbo-Croatian show that even when a prefixed manner verb is possible, it is not used as often as may be assumed. The most frequently used option for expressing motion events seems to be the use of directional verbs, which is very similar to the verb-framed pattern (cf. Filipović and Vidaković in press). Still, manner verbs show higher presence in Serbo-Croatian than other verb-framed languages. This brings us to the conclusion that the limited use of manner verbs and the preferential use of directional verbs put Serbo-Croatian in a position between a typical satellite-framed language (English) and a typical verb-framed language (Spanish) within the typology.

4.

Conclusion

In this paper, I discussed the multilayered role of prefixes and their morphosyntactic characteristics relevant to the lexicalization of motion events in Serbo-Croatian. They play a decisive role in conveying information regarding both spatial and temporal features of events in motion expressions, as well as determining lexical choices (e.g., verbs expressing Manner vs. verbs expressing Path). On the whole, the combinatory potential appears to be a highly regularized systemic feature, guided by semantic (e.g., prefix-determined preposition use) and morphosyntactic requirements (i.e., providing as much information as possible with fewer items). Together with morphological blocking, combinatory potential underlies the restrictions that had to be introduced into the typology, and simultaneously enhances the predictive power of the typology. The possibility to attest the validity of arguments presented here is vast. For example, Vidaković (2006) and Filipović and Vidaković (in press) provide evidence that English learners of Serbo-Croatian have difficulties, precisely because of the intratypological contrasts reflected in the combinatory potential and morphological blocking. English learners of Serbo-Croatian shy away from prefixed manner verbs and use unprefixed ones when they are not licensed. The errors of English learners of Serbo-Croatian occur due to the difficulty of acquiring prefixed manner verbs and result in the tendency to use unprefixed manner verbs (morphologically simpler) on all occasions, which is not licensed in situation types such as in example (17a) taken from Vidaković (2006). Following the pattern of their mother tongue, they also tend not to opt for a directional verb, whereas the directional verb is selected by native speakers who acted as controls in the experiment (ibid.). For example, a directional verb is required in (17a), and the correct answer

264 Luna Filipović

should have the form of (17b). An adequately prefixed manner verb does not exist, and the unlicensed unprefixed manner verb is used by learners in (17a): (17) a. They danced onto the balcony. Learner error: *Plesali su na balkon. dance.PST.IPFV.3PL.M be.COP onto balcony b. Izašli su na balkon plešući. exit.PST.PFV.3PL.M be.COP onto balcony dancing ‘They went onto the balcony dancing.’

This analysis helps us achieve better understanding of the difficulties learners might encounter. Implementing these findings would make learners aware of the crucial differences between their mother tongue and a second language. With an adequately adjusted teaching methodology, the learning process would be facilitated. Finally, the understanding of combinatory potential and morphological blocking makes it possible to pin down more precisely the problematic points in learning and in practicing translation (Filipović 2008). The starting point in the explanation of how events are lexicalized should include an introduction to the two universal spatial and temporal notions, such as boundary and change, respectively. This helps us tease apart language-specific features (such as the role of prefixation, which only some languages have) from universal ones (such as expressing presence/absence of change over time, which is done in all languages). This paper is an illustration of how language contrasting can be carried out within a typological framework proposed in cognitive linguistics, encompassing many levels of meaning at which prefixes seem to operate. Further research within Slavic and non-Slavic languages along the lines advocated here will undoubtedly uncover more intricacies that will help us understand and appreciate the importance of being a prefix.

References Aske, J. 1989. Path predicates in English and Spanish: A closer look. Proceedings of the Fifteenth Annual Meeting of the Berkeley Linguistics Society, 1–14. Belić, A. 1964. Savremeni srpskohrvatski knjizevni jezik II. Beograd: Naučna knjiga. Bohnemeyer, J. 2003. The unique vector constraint: The impact of direction changes on the linguistic segmentation of motion events. In Representing Direction in Language and Space, E. Van der Zee & J. Slack (eds.), 86–110. Oxford: OUP. Choi, S. & Bowerman, M. 1992. Learning to express motion events in English and Korean: The influence of language-specific lexicalization patterns. In Lexical and Conceptual Semantics, B. Levin & S. Pinker (eds.), 83–121. Cambridge MA: Blackwell. Dickey, S. M. 2000. Parameters of Slavic Aspect: A Cognitive Approach. Stanford CA: CSLI.

The importance of being a prefix 265

Ferm, L. 1990. Vyrazhenie napravleniia pri pristavochnykh glagolakh peremeshcheniia v sovre‑ mennom russkom iazyke: K voprosu prefiksalno-predlozhnogo determinizma. Uppsala: Academiae Upsaliensis. Filipović, L. 1999. Language-Specific Expression of Motion and Its Use in Narrative Texts. MPhil dissertation, University of Cambridge. Filipović, L. 2001. Language-specific use of verbs in motion expressions. In El Verbo entre el Léxico y la Gramática, A. Veiga et al. (eds.), 53–62. Lugo: Tris-Tram. Filipović, L. 2002. Verbs in Motion Expressions: Structural Perspectives. PhD dissertation, University of Cambridge. Filipović, L. 2006. Weaving the web of meaning. Languages in Contrast 6(1): 151–175. Filipović, L. 2007a. Talking about Motion: A Crosslinguistic Investigation of Lexicalization Pat‑ terns. Amsterdam: John Benjamins. Filipović, L. 2007b. Language as a witness: Insights from cognitive linguistics. Speech, Language and the Law 14(2): 245–267. Filipović, L. 2008. Typology in action: Applying insights from typological contrasts. Interna‑ tional Journal of Applied Linguistics 18(1): 42–61. Filipović, L. & Vidaković, I. In press. Typology in the L2 classroom: Second language acquisition from a typological perspective. In Inside the Learner’s Mind: Cognitive Processing in Second Language Acquisition, M. Pütz & L. Sicola (eds). Amsterdam: John Benjamins. Gasparov, B. 1990. On the “metaphysics” of Russian aspect. In Verbal Aspect in Discourse, N. B. Thelin (ed.), 191–212. Amsterdam: John Benjamins. Grickat, I. 1966/67. Prefiksacija kao sredstvo gramatičke (čiste) perfektizacije. Južnoslovenski Filolog XXVII(1–2): 185–223. Levin, B. 1993. English Verb Classes and Alternations. Chicago IL: University of Chicago Press. Lyons, J. 1991. Language and Linguistics. Cambridge: CUP. Miller, G. A. & Johnson-Laird, P. N. 1976. Language and Perception. Cambridge MA: Harvard University Press. Naigles, L. R., Eisenberg, A. R., Kako, E. T., Highter, M. & McGraw, N. 1998. Speaking of motion: Verb use in English and Spanish. Language and Cognitive Processes (13): 521–549. Saeed, J. I. 2001. Semantics. Oxford: Blackwell. Shull, S. 2003. The Experience of Space. München: Verlag Otto Sagner. Slobin, D. I. 1996. Two ways to travel: Verbs of motion in English and Spanish. In Grammatical Constructions: Their Form and Meaning, M. Shibatani & S. A. Thompson (eds.), 195–219. Oxford: Clarendon Press. Slobin, D. I. 1997. Mind, code, and text. In Essays on Language Function and Language Type, J.Â€Bybee, J. Haiman & S. A. Thompson (eds.), 437–467. Amsterdam: John Benjamins. Slobin, D. I. 2000. Verbalized events: A dynamic approach to linguistic relativity and determinism. In Evidence for Linguistic Relativity, S. Niemeier & R. Dirven (eds.), 107–138. Amsterdam: John Benjamins. Slobin, D. I. 2006. What makes manner of motion salient? Explorations in linguistic typology, discourse and cognition. In Space in Languages: Linguistic Systems and Cognitive Catego‑ ries, M. Hickmann & S. Robert (eds.), 59–81. Amsterdam: John Benjamins. Stevanović, M. 1989. Savremeni srpskohrvatski jezik, Vol. 1. Beograd: Naučna knjiga. Talmy, L. 1985. Lexicalization patterns: semantic structure in lexical forms. In Language Typol‑ ogy and Syntactic Description, Vol. 3, T. Shopen (ed.), 57–149. Cambridge: CUP. Talmy, L. 1991. Path to realization: A typology of event conflations. Proceedings of the Berkeley Linguistic Society 17: 480–519.

266 Luna Filipović

Talmy, L. 2000. Toward a Cognitive Semantics, Vol. 2. Cambridge MA: The MIT Press. Verkuyl, H. J. 1999. Aspectual Issues: Studies on Time and Quantity. Stanford CA: CSLI. Vidaković, I. 2006. Second Language Acquisition of Dynamic Spatial Relations. PhD dissertation. RCEAL, University of Cambridge. Zlatev, J. & Yangklang, P. 2002. Not either-or, but both path+motion and manner+motion. Paper presented at the International Conference on the Adpositions of Movement, University of Leuven, Belgium.

Dictionaries Rečnik srpskohrvatskog književnog i narodnog jezika (SANU). Vols. 1–15. Beogorod: Institut za srpskohrvatski jezik, 1959–1989. Rečnik srpskohrvatskoga književnog jezika. Vols. 1–6. Novi Sad: Matica srpska, 1967–1976.

Corpora (consulted between October 2000 and April 2002) British National Corpus on-line (for English): www.natcorp.ox.ac.uk. Croatian National Corpus on-line (for Serbo-Croatian): www.hnk.ffzg.hr.

Appendix. Abbreviations 3SG 3PL COP IPFV M N PFV PST REFL

3rd person singular 3rd person plural copula imperfective masculine neuter perfective past reflexive pronoun

chapter 11

Variation in the encoding of endpoints of motion in Russian Tatiana Nikitina

Stanford University

In Russian, as in a typical satellite-framed language, endpoints of motion are usually introduced by specialized directional PPs (such as combinations of a preposition with the accusative case). With a small set of verbs, however, the endpoint of motion can instead be introduced by locational PPs. This paper explores restrictions on the use of this less-studied strategy for encoding endpoints of motion. It is argued that locational PPs with an endpoint interpretation are licensed by change of state verbs, rather than motion verbs, and alternate with directional PPs that behave as result phrases. It is also shown that the choice of a construction is further influenced by a number of contextual factors, including event construal and the preposition used.

1.

Alternative ways of encoding endpoints of motion

Cross-linguistic variation in the encoding of motion events is one of the most intensively studied topics in lexical typology (Talmy 1975, 1985, 1991; Slobin 1996, 2006; Bohnemeyer et al. 2007; Beavers et al. forthc., inter alia). An especially wellresearched parameter of the variation is the distinction between so-called verbframed and satellite-framed languages (Talmy 1985, 1991, 2000). The two language types rely on different lexical means in distinguishing goals of motion from static locations. Satellite-framed languages tend to encode goals of motion with specialized adpositions, case marking, particles, or adverbs. For example, in (1) from English the endpoint of motion is unambiguously encoded by a preposition. (1) John walked / ran / came into the room.

Verb-framed languages typically lack specialized lexical means for encoding the meaning of goal and instead rely on the class of verbs of directional motion, which unambiguously determine the role of their locative argument. In such a language, the same PP can be interpreted as a goal or as a static location depending on the

268 Tatiana Nikitina

verb it combines with. In example (2), from French, the goal of motion is encoded in the same way as a static location, and only the verb distinguishes between the two readings. The PP introduces a static location with verbs that do not entail a change of location, such as with positional and manner of motion verbs (2a), but a goal with verbs of directional motion (2b). (2) a. b.

Je suis à l’école ‘I am at school.’ (static location) Je suis allé à l’école ‘I went to school.’ (motion toward a goal)

Within this typology, Russian appears to be a well-behaved satellite-framed language (see, e.g., Slobin 2003, 2004). The distinction between goals and static locations is marked by a combination of preposition and case, and a number of prepositions, including v ‘in’ and na ‘on’, can take a complement in either locative (as in (3a)) or accusative (as in (3b)) case depending on whether they mark a static location or an endpoint of motion. (3) a. b.

Sobaka spit v prixožej dog sleeps.ipfv in hall.loc ‘The dog is sleeping in the hall.’ Sobaka bežit v prixožuju dog runs.ipfv in hall.acc ‘The dog is running to the hall.’

. In traditional grammar the locative is often subsumed under the prepositional case, since for most nouns the two forms are identical; a small set of nouns, however, have a locative form that is clearly distinct from the prepositional form, cf. the locative (v) sadu and the prepositional (o) sade (see Comrie 1986:â•›93–94). Timberlake (1993:â•›836; 2004:â•›330–333) refers to the prepositional case as “locative”, and the locative case as “second locative”. In this paper I do not distinguish between the two case forms (prepositional vs. locative, or locative vs. second locative) and subsume them under the label “locative”. . I ignore the distinction between determinate (unidirectional) and indeterminate (non-unidirectional) motion verbs, since it is orthogonal to the expression of the endpoint of motion. Although determinate motion verbs have a strong tendency to co-occur with goal arguments, they can also combine, in an appropriate context, with locational modifiers; similarly, indeterminate motion verbs, while commonly modified by locational PPs, can combine with directional PPs that specify the endpoint of motion (typically, such uses are interpreted habitually; this reading reconciles, on the one hand, the lack of a specific direction suggested by the choice of an indeterminate motion verb and, on the other, the presence of an endpoint of motion suggested by the directional PP). The following two pairs of examples from the Russian National Corpus (http://www.ruscorpora.ru/) illustrate that both types of verb are compatible with either type of PP.

Variation in the encoding of endpoints of motion in Russian 269

Table 1.â•‡ Some directional and locational P + case combinations3 preposition

locational PP

directional PP

v ‘in’ na ‘on’ pod ‘under’ nad ‘above’ za ‘behind’ pered ‘in front of ’ rjadom s ‘near’ u ‘by’ k ‘toward’

v + Loc na + Loc pod + Instr nad + Instr za + Instr pered + Instr rjadom s + Instr u + Gen *

v + Acc na + Acc pod + Acc * za + Acc * * * k + Dat

Selected directional and locational P + case combinations are listed in Table 1; some Russian prepositions are complex and consist of more than one word (cf. rjadom s ‘near’ below). With some verbs, however, the distinction between goals of motion and static locations is blurred. In (4), the endpoint of motion can be expressed in two alternative ways: in the P + accusative configuration, i.e. with the specialized marking normally used to encode goals of motion, or in the P + locative combination, as if

(i) Determinate motion (letet’ ‘fly’) a. Zvuk truby letit v nebo sound of.trumpet flies.ipfv in sky.acc ‘The sound of the trumpet flies into the sky.’ (V. Aksenov, Zvezdnyj bilet) b. Letit v sinem nebe, lëgkij, kak vol’naja ptica flies.ipfv in blue.loc sky.loc light as free bird ‘[He] flies in the blue sky, as light as a free bird.’ (F. Sologub, V tolpe) (ii) Indeterminate motion (letat’ ‘fly’): a. Gluxari malo edjat xlebnyx zëren i redko wood-grouses little eat.ipfv bread grain and rarely letajut v xlebnye polja fly.ipfv in bread.acc fields.acc ‘Wood-grouses eat small amounts of bread grains and rarely fly to bread fields.’ (S. Aksakov, Zapiski ružejnogo oxotnika) b. Lastočki tabunkami letajut v poljax swallows little.herds.instr fly.ipfv in fields.loc ‘Swallows fly in the fields in small herds.’ (M. Prishvin, Dnevniki) . The table reflects current usage only; for example, the preposition pered could combine with the accusative case on the directional reading until at least the second half of the 19th century, and possibly later (Stanisheva 1966:â•›118–119).

270 Tatiana Nikitina

it were a static location (in Russian, the former variant is more frequent and tends to be preferred out of context). (4) Postav’ vazu na stol / na stole put.pfv vase on table.acc / on table.loc ‘Put the vase on the table.’

Similar variation is attested in other Slavic languages: prepositional phrases that normally express static locations can encode the endpoint of motion in Polish (Cienki 1989:â•›141–147), Czech (Belichova-Krzhizhkova 1974; Ungermanová 2005), and Ukrainian (Nedashkivska 2001). This paper addresses the problem of the directional/locational alternation in Russian. I survey the use of the two alternative ways of encoding endpoints of motion (with directional vs. with locational PPs) and suggest that the choice between them is determined by a set of pragmatic factors, as well as by a lexical constraint defining the set of verbs that allow for both variants. In this first exploratory study of the factors involved, I rely mostly on grammaticality judgments of minimal pairs of made-up examples; examples from the Russian National Corpus are provided in footnotes where relevant to illustrate the use of the less common pattern in context. . Some examples with the P + locative combination, including those attested in the corpus, may even be rejected by some speakers when provided in isolation. Judgments tend to change when the sentence is embedded in an appropriate context. The following examples illustrate the use of the P + locative combination in the Russian National Corpus (http://www.ruscorpora. ru/); due to space limitations, I cannot provide a comprehensive account of all such uses; exploration of the corpus data remains a topic for a separate study. (iii) Stol postavil sredi grota, na stole postavil vazy… table put.pfv in.the.middle.of grotto.gen on table.loc put.pfv vases ‘[He] put the table in the middle of the grotto, put vases on the table…’ (A. Beliaev, Čelovek-Amfibija) (iv) … Vasilisa nalivaet emu stakan čaju i stavit na seredine stola V. pours.ipfv him glass of.tea and puts.ipfv on middle.loc of.table ‘… Vasilisa pours him a glass of tea and puts it in the middle of the table.’ (V. Rasputin, Vasilij and Vasilisa) (v) A njanjuška stavit da stavit rjadkom ix na stole… and nanny puts.ipfv and puts.ipfv side.by.side them on table.loc ‘And nanny keeps putting [lit., ‘puts and puts’] them [toys] on the table side by side’. (V. Odoevsky, Igoša) . I have worked closely with three native speakers of Russian as my primary language consultants; all grammaticality judgments reported in this study were also checked with two more speakers.

Variation in the encoding of endpoints of motion in Russian 271

2.

Lexical constraints on the directional vs. locational alternation

2.1

Defining the class of verbs

The variation in the encoding of goals is constrained by two kinds of factors. First of all, the locational and the directional constructions differ with respect to the type of motion event they describe. As I will show in Section 3, which construction is preferred with a particular verb sometimes depends on subtle properties of the setting, such as the intended duration of the resulting state or the spatial configuration described. In addition to this, not every verb can combine with a locational PP expressing the endpoint of motion. In Russian, only a small set of verbs allow for the alternation illustrated in (4). In this section I discuss differences in the behavior of verbs with respect to the marking of endpoints of motion. As already discussed in Section 1, Russian is largely a satellite-framed language. With most verbs, the endpoint of motion must be encoded by a directional PP, i.e. by a specialized combination of preposition and case. No alternative description of goal is acceptable in examples like (5a), where the verb encodes a change of location as part of its meaning. Similarly, in (5b) there is no way of expressing the endpoint of motion with prepositions that introduce unambiguously locational (and not directional) phrases (see again Table 1 for a non-exhaustive list of such prepositions). (5) a. b.

Mal’čik v-bežal v školu / *v škole boy in-ran.pfv in school.acc / in school.loc ‘The boy ran to school.’ *Mal’čik v-bežal pered školoj / u školy boy in-ran.pfv in.front.of school.instr / near school.gen ‘The boy ran to in front of the school / to near the school.’

The verb vbežat’ ‘run in’ is derived by a prefix from the manner of motion verb bežat’ ‘run’ and, unlike the latter, entails directed motion (the entailment is in this case contributed by the prefix); thus, the non-prefixed verb bežat’ ‘run’, but not vbežat’ ‘run in’, can describe running without changing location, such as running on the spot. All verbs entailing directed motion (which I will refer to as inherently directional verbs) require the endpoint of motion to be encoded by a directional PP. In Russian, as in English, not only motion entailing verbs can be used to describe directed motion. In (6), directional PPs combine, with certain restrictions, with verbs that do not by themselves entail a change of location, including verbs that describe certain activities, or manners of motion (6a), and certain causing events (6b).

272 Tatiana Nikitina

(6) a. b.

Activities: Soldaty šagajut / begut / *pljašut v gorod soldiers march.ipfv / run.ipfv / dance.ipfv in town.acc ‘The soldiers are marching / running / *dancing into the town.’ Causing events: Korova tolknula / pixnula / *ljagnula sobaku v ozero cow pushed.pfv / shoved.pfv / kicked.pfv dog in lake.acc ‘The cow pushed / shoved / *kicked the dog into the lake.’

These examples demonstrate that in addition to inherently directional verbs, some other (but by no means all) verbs can be used to describe directed motion events and combine with directional PPs. Such verbs do not entail a change of location (e.g., pushing does not always result in motion) but can be used to describe it due to a particular kind of meaning extension (Levin and Rapoport 1988). Using the term introduced by Talmy (2000:â•›270–271), I will refer to such verbs as verbs with a lexicalized implicature of change of location. The lexicalized implicature is part of the lexical content of the verb and differs from an entailment in being defeasible. Lexicalization of the change of location implicature is subject to certain constraints on the relation between the motion event and the additional event encoded by the verb in its basic meaning; in particular, the two events must be causally related (Carter 1988:â•›178–179; Levin and Rappaport Hovav 1991; Croft 1991:â•›160–161; also Goldberg 1995:â•›61–63). For this reason, extending the meaning of a verb to describe a change of location results in characteristic constraints on interpretation. With activity verbs, directed motion is understood to be temporally coextensive with the main event (in 6a, motion is performed by means of marching), while with punctual verbs, motion is directly caused by the event and follows it (in 6b, the change of location is caused by pushing). Most importantly for our study, verbs that implicate a change of location behave exactly like verbs that entail it with respect to the encoding of goals: both require the endpoint of motion to be expressed by a directional PP. This restriction is independent of the verb’s aspectual characteristics (it holds for both perfective and imperfective verbs

. The number of Russian verbs that lexicalize a change of location as an implicature seems to be small compared to English. For example, sound emission verbs are typically disallowed with directional PPs unless derived by a directional prefix, as in (vi) (and in this case, the derived form unambiguously entails a change of location). (vi) Gruzovik pro-gromyxal / *gromyxal vo dvor truck PRO-rumbled.pfv / rumbled.ipfv in yard.acc ‘The truck rumbled into the yard.’

Variation in the encoding of endpoints of motion in Russian 273

implicating a change of location), as well as of the distinction between determinate (unidirectional) and indeterminate (non-unidirectional) motion verbs. I now turn to verbs that allow the endpoint of motion to be expressed in two alternative ways, by a directional or by a locational PP. This flexibility characterizes, in the first place, intransitive verbs of change of position and their transitive counterparts, i.e. verbs of putting (7). The example in (4) involved one of these verbs. (7)

Change of position verbs: Intransitive Transitive sadit’sja ‘sit down’ sažat’ ‘seat (caus.)’ ložit’sja ‘lie down’ klast’ ‘lay down, put, place’ v-stavat’ ‘stand up’ stavit’ ‘put, place, set, make stand’ (visnut’ ‘hang’) vešat’ ‘hang (caus.)’

In addition to change of position verbs, verbs that allow locational PPs to express the endpoint of motion include a number of change of state verbs that describe activities not directly associated with motion but often accompanied by it. I list some verbs of this kind in (8); most verbs from this group also have an intransitive counterpart derived by adding the suffix -sja. Examples of the variation between locative and directional PPs are provided in (9). (8)

Some additional change of state verbs prjatat’ ‘hide’ zatočat’ ‘incarcerate, immure’ zapirat’ ‘lock up’ zapečatyvat’ ‘seal up’ zaryvat’ ‘bury’ ustanavlivat’ ‘install’ prikleivat’ ‘stick’ prikrepljat’ ‘attach’

. While only imperfective verbs are listed in (7), their perfective counterparts are assumed not to differ in any relevant respect. More generally, no correlation has been noted between the verb’s aspectual class and the restrictions it imposes on the expression of the endpoint of motion. . The verb visnut’ ‘hang’ is occasionally used with directional PPs in the present-day language; among expressions that were accepted by the speakers consulted is visnut’ na šeju ‘hang on [smb.’s] neck’. Apart from those few expressions, however, the verb combines with locational PPs only.

274 Tatiana Nikitina

(9) a. b. c.

Pirat sprjatal zoloto v sunduk / v sunduke pirate hid.pfv gold in chest.acc / in chest.loc ‘The pirate hid the gold in a chest.’ Koldun zatočil carevnu v bašnju / v bašne sorcerer incarcerated.pfv princess in tower.acc / in tower.loc ‘The sorcerer incarcerated the princess in a tower.’ Mexanik ustanovil v mašinu / v mašine radio technician installed.pfv in car.acc / in car.loc radio ‘The technician installed a radio in the car.’

Finally, both locational and directional PPs occur with some verbs that do not describe any motion in space but rather refer to motion in a metaphorical sense, such as the verb zapisyvat’ ‘write down’ in (10), where the act of writing is represented as a metaphorical transfer of information. I leave instances of metaphorical motion to future research and do not try to account for them in the present paper. (10) On zapisal telefon v tetrad’ / v tetradi he wrote.down.pfv phone in notebook.acc / in notebook.loc ‘He wrote down the phone number in the notebook.’

Verbs that allow for variation in the encoding of the endpoint of motion differ semantically both from verbs that entail a change of location and from verbs with a change of location implicature. First of all, the verbs in (7), unlike verbs of directed motion, do not entail a change of location but rather a change in the theme’s spatial configuration. With the verb sažat’ ‘seat’, the theme ends up in a sitting position; the verbs stavit’ and klast’ differ with respect to the lexicalized resulting state (roughly speaking, in a vertical vs. horizontal position). (11) Postav’ / položi butylku na pol make.stand.pfv / put.pfv bottle on floor.acc ‘Put the bottle on the floor (in standing / horizontal position).’

Although the context may make it clear that the change of position was preceded by a change of location, the change of location is by no means entailed. For example, (12) can describe a change of position without a change of location. (12) Ja uvidel, čto butylka upala, i snova eë postavil I saw.pfv that bottle fell.pfv and again it.acc make.stand.pfv ‘I saw that the bottle fell and put it back into standing position.’

Similarly, the verbs in (8) entail a specific change of state but not directed motion, cf. the following example, where no change of location takes place.

Variation in the encoding of endpoints of motion in Russian 275

(13) Oni sprjatali korobku, nakryv eë gazetoj they hid.pfv box cover.prt.pfv it.acc newspaper.instr ‘They hid the box by covering it with a newspaper.’

Just like change of position verbs, the verbs in (8) can but need not describe a change of location in addition to the encoded change of state. Interestingly, some of the change of state verbs in question have prefixed counterparts that do entail a change of location, cf. sadit’sja ‘sit down’ vs. peresaživat’sja ‘change place, take another seat’. Although these two verbs are similar in their meaning (in both cases, the resulting state is that of sitting), the latter, but not the former, describes a complex change of location event. Due to this difference (one verb does not entail a change of location while the other one does), the two verbs differ in the way they mark the endpoint of motion. The change of state verb combines with locational as well as directional PPs; the prefixed change of location verb combines with directional PPs only. (14) a. Gosti seli na divan / na divane guests sat.down.pfv on sofa.acc / on sofa.loc ‘The guests sat down on a sofa.’10 . As noted by an anonymous reviewer, only a locational PP is possible when the verb describes a change of state that is not accompanied by any change of location. Such locational modifiers indicate where the entire change of state took place, not where the theme ended up as a result of the change: (vii) Postav’ čašku v rakovine / *v rakovinu make.stand.pfv cup in sink.loc / in sink.acc ‘Put the cup in into the standing position the sink’ (the cup is located in the sink from the very beginning, and the event only involves a change from a horizontally to a vertically oriented position) 10. As already noted above, the P + locative combination is relatively infrequent and dispreferred in the absence of context. The following examples illustrate its use with the same verb as in (14) (sest’ ‘sit down’) to encode the endpoint of motion in the Russian National Corpus; in all of them, a directional PP can be substituted for the locational one without an obvious difference in meaning. ( viii) Nu, tak sjademte zdes’, na lavočke well then sit.down.pfv.imper.pl here on bench.loc ‘Well, then, let’s sit down here on a bench, then.’ (ix)

(A. Bely, Peterburg)

On sdelal rezkoe dviženie i srazu sel na he made.pfv sharp movement and right.away sat.down.pfv on svoëm meste … own.loc place.loc ‘He made a sharp movement and right away sat down at his place.’ (V. Korolenko, Moroz)

276 Tatiana Nikitina

b. Gosti pere-seli na divan / *na divane guests pere-sat.down.pfv on sofa.acc / on sofa.loc ‘The guests changed place and sat down on sofa.’

The difference in (14) confirms once again that verbs with variable expression of the endpoint of motion differ in their lexical properties from verbs that entail a change of location. They also differ from verbs with a change of location implicature, even though the use of change of state verbs to describe a motion event is superficially similar to the meaning extension that was illustrated in (6). In both cases, the change of location is not entailed by the verb per se but is described by the verb in addition to its basic meaning. On closer examination, however, the two extension patterns turn out to be quite different. First of all, with verbs that entail or implicate change of location, the motion event is either temporally co-extensive with the other event described by the verb (if the verb describes an activity) or directly follows that event (if the verb describes a punctual event). In both cases, the two events stand in a direct causal relationship. With change of state verbs, no direct causal relationship is required: burying something or hiding it is not a means of moving it; rather, it is an action that may be accompanied by motion at some of its stages. Moreover, with change of state verbs the directed motion event is not temporally coextensive with the change of state and is often understood to precede it. Things do not move by means of being sealed up or attached, sitting down is understood to be preceded by motion and not to cause it, and no change of location takes place after something is hidden or buried. The difference in interpretation of verbs that entail or implicate motion and verbs of change of state suggests that although they are used in similar constructions (in combination with a directional PP), this similarity is in fact superficial. I return to this issue in the following section. To sum up, I have introduced three lexical classes of verbs that can express the endpoint of motion by a directional PP: verbs that entail a change of location, verbs that implicate it, and change of state verbs that sometimes allow motion to be inferred. Change of state verbs differ from verbs that implicate a change of location in the way the motion event is interpreted with respect to the other event lexicalized by the verb. They do not express directed motion as part of their extended meaning, but rather allow a change of location to be inferred as taking

(x) On sel na stupen’ke i namylil svoi dlinnye volosy i šeju … he sat.down.pfv on step.loc and soaped.pfv own long hair and neck ‘He sat down on a step and soaped his long hair and neck…’ (A. Chekhov, Kryžovnik)

Variation in the encoding of endpoints of motion in Russian 277

Table 2.â•‡ Lexical classes and the expression of endpoints of motion Verb class

Change of location

Endpoint of motion

verbs of directional motion

entailed

directional PP

verbs that lexicalize a directional motion implicature

part of extended meaning: directional PP accompanies another event or follows it

change of state verbs

inferred as preceding the change of state

directional / locational PP

place prior to the change of state. The correspondence between the lexical classes and the expression of endpoints of motion is summarized in Table 2.

2.2

Change of location and change of state: Goal arguments vs. results

In the previous section I showed that motion verbs differ from change of state verbs in the way they encode the endpoint of motion. With verbs that lexicalize directed motion, either as an entailment or as an implicature, the endpoint must be expressed by a directional PP. With change of state verbs, directional PPs alternate, under certain conditions, with locational ones. In addition, the two types of verb differ in the way they integrate motion with the other event they express: with change of state verbs the directed motion is neither coextensive with the change of state nor caused by it, but precedes it. Such differences suggest that the (non-alternating) directional complement of motion verbs may differ in its syntactic status from the (alternating) directional complement of change of state verbs. In the present section I argue that the two kinds of directional PP should be treated as components of two different constructions. Verbs that entail or implicate a change of location select for a goal argument and require it to be expressed in a directional PP. This is the only way they can express the endpoint of motion. Change of state verbs, on the other hand, can be modified by locational PPs, which specify where the change of state took place and in some cases receive a directional interpretation through inference (the change of state is inferred to be preceded by a change of location). This explains one part of the problem, namely, why the endpoint of motion can be expressed as a static location with the change of state verbs (prior motion is inferred) but not with change of location verbs (the goal argument is subcategorized by the verb and must be expressed in a conventionalized way). The other part of the problem has to do with availability of directional marking with change of state verbs. How is it to be explained if the verbs do not select for a goal argument?

278 Tatiana Nikitina

Crucially, in Russian, the use of directional PPs is not restricted to goal arguments. Directional PPs occur with a variety of change of state verbs that are not associated with motion. In such cases, directional PPs describe the resulting state of the object that undergoes a change of state. In (15), for instance, the verbs do not describe any motion in space, and the meaning of the directional PPs is clearly different from that of a spatial goal. (15) a. b. c. d. e.

rvat’ / rvat’-sja na melkie kusočki tear.ipfv / tear.ipfv-refl on small.acc pieces.acc ‘to tear (trans./intrans.) into small pieces’ raskolot’ / raskolot’-sja na časti split.pfv / split.pfv-refl on parts.acc ‘to split (trans./intrans.) into parts’ razrezat’ nožnicami na dlinnye poloski cut.ipfv scissors.instr on long.acc strips.acc ‘to cut into long strips with scissors’ krasit’ v sinij cvet paint.ipfv in blue.acc color.acc ‘to paint blue’ svoračivat’ bumagu v rulon roll.ipfv paper.acc in roll.acc ‘to roll paper into a roll’

In all the above examples the directional PPs function not as goal arguments of motion verbs (they cannot be interpreted literally in spatial terms) but rather as a kind of result PP. This suggests that the same kind of result PP could also combine with other change of state verbs, including verbs of change of position (7) and verbs of the hide class (8). In (15), where the verbs describe a change in the object’s physical properties (an object is destroyed, transformed with respect to its color, shape, etc.), the result PP specifies the resulting state of that change, i.e. the object’s acquired physical property (its consistency, color, shape, etc.). With change of state verbs that allow motion to be inferred, however, the object undergoes a change in position or some other kind of change that does not affect it physically (the object becomes hidden, sealed up, buried, etc.). With such verbs, the result PP is not associated with a physical property (physical properties remain intact) but describes a location where the object ends up by the time the change is completed.11

11. Besides the endpoint of motion, the result PP can specify the resulting position, namely, the part of the object that ends up providing support to it, cf. (xi) vs. (xii):

Variation in the encoding of endpoints of motion in Russian 279

On this account, the alternation between locational and directional PPs is due to the fact that the endpoint of motion coincides in the case of the verbs under investigation with the location where the lexicalized change of state takes place. Because of this semantic affinity (the directed motion is understood to end where the change of state takes place), the use of locational vs. directional PPs can be treated as an alternation. With verbs that describe changes in physical properties, the result PP does not have a spatial interpretation and therefore differs significantly in its meaning from a locational PP. Due to this difference, speakers of Russian do not treat the use of directional and locational PPs as semantically equivalent with verbs of change in a physical property. There is some additional evidence for distinguishing between directional PPs that are used with change of location verbs (goal arguments) and those that occur with non-physical change of state verbs (results). One striking difference between the result PPs with a spatial reading and goal arguments has to do with their compatibility with descriptions of sources of motion. Typically, any change of location that has an endpoint also has a starting point; hence, we might expect a verb that takes a goal argument to be equally compatible with a description of the starting point of motion. In (16), regular motion verbs (with and without prefixes) can combine with either goal or source phrases. They can also occur with both at the same time. (16) a. b. c. d.

Sobaka vy-bežala iz doma / vo dvor dog out-ran.pfv out.of house.gen / in yard.acc ‘The dog ran out of the house / into the yard.’ Sobaka v-bežala so dvora v dom dog in-ran.pfv from yard.gen in house.acc ‘The dog ran in from the yard into the house.’ Mal’čik bežit iz magazina / v školu boy runs.ipfv out.of store.gen / in school.acc ‘The boy is running from the store / to school.’ Gruzovik vezët brëvna iz lesa v derevnju truck carries.ipfv logs from forest.acc in village.acc ‘The truck is carrying logs from the forest to the village.’

(xi) Koška ležit na boku / na spine cat lies.ipfv on side.loc / on back.loc ‘The cat is lying on its side / back.’ (xii) Koška legla na bok / na spinu cat lay.down.pfv on side.acc / on back.acc ‘The cat lay down on its side / back’

280 Tatiana Nikitina

In contrast to verbs that entail or implicate a change of location, verbs of change of position do not freely combine with descriptions of sources of motion. In (17), no source PP can be used either in combination with a goal phrase or on its own. (17) a. b.

Položi ključi (*so stula) na stol put.pfv keys from chair.gen on table.acc ‘Put the keys (*from the chair) on the table.’ Sjad’ (*s kresla) na divan sit.down.pfv from armchair.gen on sofa.acc ‘Sit down (*from the armchair) on the sofa.’

The difference in the treatment of sources follows straightforwardly from our analysis of the directional PPs in (17) as result PPs. Goal and source arguments are typically licensed by verbs that entail or implicate a change of location, hence their similar distribution. Result PPs, however, are licensed by change of state verbs, which often cannot combine with a source phrase. In (18), for example, a transformation verb, which describes a change in physical properties of the object, is compatible with a result PP but does not allow the initial state to be expressed as a source PP: (18) Razrež’ bumagu na poloski (*iz lista) cut.pfv paper on strips.acc out.of sheet.gen ‘Cut the paper into strips (*from a sheet).’

The correlation between compatibility with a source argument and restrictions on the expression of endpoints of motion receives further support from the use of prefixed verbs in (19). (19) a. b.

Pere-loži ključi so stula na stol pere-put.pfv keys from chair.gen on table.acc ‘Take the keys from the chair and put them on the table.’ Pere-sjad’ s kresla na divan pere-sit.down.pfv from armchair.gen on sofa.acc ‘Move from the armchair to the sofa.’

As we saw in the previous section, verbs with the prefix pere- differ from their counterparts in one crucial property: they entail a change of location (the verb pere-saživat’sja means ‘change seat, go sit in a different place’). As predicted by the contrast in meaning, verbs entailing a change of location, but not simple change of position verbs, are compatible with descriptions of the source of motion (since they select for both goal and source arguments). This difference once

Variation in the encoding of endpoints of motion in Russian 281

again suggests that the construction in (17) differs from that in (19): the directional PP is a result phrase in the former but a goal argument in the latter.12 To sum up, I have shown that the use of directional phrases is not restricted to expressing goal arguments of motion verbs but extends to certain kinds of result. Unlike goal arguments, which are in general compatible with descriptions of sources of motion, result PPs do not have to be interpreted in spatial terms and are in general incompatible with descriptions of sources. In other words, verbs that lexicalize motion license goal and source arguments, while change of state verbs are modified by result PPs. This difference corresponds to the different interpretation of motion with the two types of verb. With goal arguments, the change of location event is coextensive with another event or directly caused by it (it is part of the verb’s basic or extended meaning). With result PPs, the change of location need not be coextensive with the event of change of state and never follows it: the specific position in space is acquired along with the change of state, and not after it.

3.

Directional vs. locational: The choice of goal-marking strategy

3.1

Event construal

Up to now I was concerned with the status of directional PPs when they are used with verbs from different lexical classes. In this section I turn to the difference between locational and directional PPs when they appear, with a similar meaning, with verbs of change of state. Are the two expressions equivalent when the change of state is accompanied by a change of location, and if not, what determines their distribution? The directional vs. locational alternation with change of state verbs is due to the coexistence of two different strategies of expressing the endpoint of motion. The use of a directional PP (more precisely, a result PP interpreted in spatial terms) is an instance of overt lexical coding (a satellite-framed construction). With a locational PP, the fact of motion is not expressed overtly but can be inferred from context. In this section I address factors that influence the choice between the two strategies with verbs that in principle allow for both alternatives. In particular, I distinguish between factors related to event construal, on the one hand, and structural factors related to general availability of a competing option, on the other. 12. The pattern found with change of state verbs may create an appearance of a source/goal asymmetry: verbs of putting occur with directional phrases but not with descriptions of source of motion. I suggest that this appearance is due to the fact that instead of being expressed as goal arguments, endpoints of motion are introduced by result PPs.

282 Tatiana Nikitina

Slavic languages are not unique in providing alternative ways of expressing the endpoint of motion with certain verbs. The alternation between a satelliteframed construction, on the one hand, and a construction that does not encode the directional meaning but relies on contextual inference, on the other, is attested in a variety of languages. To give just two examples, a similar alternation is found in English and in Ancient Greek, where the goal of motion is typically encoded with a specialized preposition or a P + case combination. The use of the satelliteframed strategy is illustrated in (20a) and (21a); in the (b) examples, however, the endpoint of motion is described by a PP that normally encodes static locations.13 (20) English a. John put the box into his pocket. b. John put the box in his pocket. (21) Ancient Greek (Smyth 1920:â•›368; Luraghi 2003:â•›66) a. Specialized P + case combination: eis hála lúmata bállon (Hom. Il.1.314) into sea.acc water.used.in.washing.acc.pl throw.ipfv.3pl ‘They threw the dirty water into the sea.’ b. P + case combination generally associated with static locations: kaì tà mèn en purì bálle (Hom. Od.14.429) and def.acc.neu.pl prt in fire.dat throw.ipfv.3sg ‘And these things he threw in the fire.’

The set of verbs with which both alternative expressions occur, as well as the relative frequency of the use of the two strategies, varies greatly from one language to another. Along with restrictions imposed by individual verbs, the choice of strategy is often constrained by contextual factors pertaining to what can be characterized as conventionalized event construal. Verbs of motion typically describe complex events that involve both a process of motion and a change of location (in some cases, also a causing event). Such complex events can sometimes be described either from the perspective of motion or from the perspective of the resulting state (the endpoint of motion), depending, among other factors, on certain properties of the event. For example, in a corpus study of the in/into variation in English (cf. (20)), I argued that in is less likely to mark a goal when the event described involves a prominent path of motion or is of considerable 13. I do not claim, of course, that in all such languages the use of a directional PP in alternation with a locational one is equivalent to the use of a result PP in Russian. This is clearly not so in English, where both into and in can introduce goal arguments (Nikitina 2008). Hence, the superficially similar alternation may reflect different underlying phenomena in the three languages.

Variation in the encoding of endpoints of motion in Russian 283

duration (Nikitina 2008), i.e. when the event is more likely to be construed from the motion perspective. Various factors of similar kind can be expected to affect the choice of strategy in other languages, including Russian. The difference in the construal of motion events is analogous in some respects to differences associated with various forms of grammatical aspect. Some kinds of event can be described in different aspects, e.g., in the progressive or in the perfect, depending on context-specific construal, while other kinds of events are more likely to be described by a particular aspectual form (e.g., punctual events tend to be described in the perfect rather than in the progressive). Among the factors that favor the use of locational as opposed to directional PPs is the intended duration of the resulting state. The directional PP is more likely to be used in examples where the resulting location is temporary, while the locational marking is associated with relatively permanent results. In (22), the “static” description is more compatible with the event of placing a vase of flowers on the table (which is more likely to be construed, out of context, as having a relatively long-lasting result) than with the event of setting up a table for dinner (which tends to be construed as having a relatively short-term result unless more specific context is provided). The judgments, however, appear to be rather subtle, suggesting a need for a thorough experimental or corpus study.14 (22) a. b.

Postav’ vazu / cvety na stol / na stole put.pfv vase / flowers on table.acc / on table.loc ‘Put the vase / flowers on the table.’ Postav’ čašku na stol / ??na stole put.pfv cup on table.acc / on table.loc ‘Put the cup on the table.’

Another factor relevant for the choice of expression is information structure. In a study conducted by Blazhev (1988) 10 native speakers of Russian were asked to choose between accusative and locative marking in the following pair of sentences: (23) a. V kuxne na stenu / na stene ja povesil časy in kitchen.loc on wall.acc / on wall.loc I hang.pfv clock ‘In the kitchen on the wall I hang a clock.’

14. That the locational marking sounds “more natural” in (22a) than in (22b) was confirmed by all five native speakers consulted. This does not mean, of course, that sentences describing events with a relatively short-term result never occur in the corpus; rather, they are predicted to be less frequent compared to events with longer-lasting results or require special types of context (see again Note 4 for some corpus examples of locational marking with the same verb).

284 Tatiana Nikitina

Table 3.â•‡ Speaker judgments on (23), based on Blazhev (1988:â•›64–65) sentence (23a) sentence (23b)

locative only

accusative only

loc or acc

9 5

0 0

1 5

b. Ja povesil časy v kuxne na stenu / na stene I hang.pfv clock in kitchen.loc on wall.acc / on wall.loc ‘I hang the clock in the kitchen on the wall.’

The word order difference between the two sentences corresponds to a difference in information structure: in (23a), but not in (23b), the endpoint of motion (‘on the wall’) is topicalized, along with the modifying location (‘in the kitchen’). The results of the study are presented in Table 3. They show a stronger preference for the locative marking when the PP is included in the topic (23a), suggesting that backgrounding the resulting location (by topicalizing it) favors the use of locational PPs. This result is consistent with the event construal hypothesis: (23b) is more easily construed as providing information about the new location of the clock, i.e. about the change of location component of the complex event.15 Factors related to conventionalized event construal are extremely difficult to characterize in precise terms, and a thorough investigation is needed to establish the relationship between the properties of motion events and the way of encoding the endpoint of motion.

3.2

Type of preposition as an additional factor in the variation

In the previous section I argued that the alternation between locational and directional PPs is partly constrained by factors that influence the way the complex motion event is linguistically construed. Such factors do not, however, account for all of the variation; moreover, they appear to play a significant role in a limited number of contexts. Overall, the use of locational PPs in contexts involving motion is rather restricted in Russian (it is extremely rare compared to Polish, where locational PPs are actually preferred with verbs of putting, cf., Toporov 1961:â•›295).

15. Blazhev discusses a number of other factors that can be subsumed under the notion of event construal, including aspectual characteristics of the event (96–97), the relative salience of displacement as opposed to change of position (90–102), the distance between the theme and its intended resulting location (longer distances favor the directional marking, 104–106), and many others (see also Israeli 2004).

Variation in the encoding of endpoints of motion in Russian 285

For example, locational PPs with the preposition v ‘in’ rarely occur in examples involving a change of location; they are invariably ruled out in (24). (24) a. ?? Položi ključi v korzine / v karmane put.pfv keys in basket.loc / in pocket.loc ‘Put the keys in the basket / in the table.’ b. ?? Postav’ čašku v rakovine / v xolodil’nike put.pfv cup in sink.loc / in refrigerator.loc ‘Put the cup in the sink / refrigerator.’

One might assume, based on such examples, that locational PPs are dispreferred with change of state verbs and only appear marginally in a small set of contexts. The account developed in the previous sections, however, relies heavily on the assumption that both directional and locational PPs are available with change of state verbs, and it is only due to the relative prominence of the change of location component in the complex event that the locational PP is avoided. On that account, the event construal associated with (24) strongly favors expressing the endpoint of motion by a result PP, to the exclusion of the locational variant. But are we justified in assuming that the locational variant is underlyingly available in examples like (24), in spite of the fact that it is unacceptable? The answer to this question becomes clear when evidence from a wider range of prepositional phrases is considered. As I briefly discussed in Section 1, locative prepositions differ in their inherent directional value. Some prepositions, including v ‘in’ and na ‘on’, can introduce either directional or locational phrases, depending on the case of their complement (accusative in directional PPs, locative in locational PPs). Other prepositions do not allow the case of their complement to vary and introduce only one kind of PP (see again Table 1). The preposition k ‘toward’, for example, takes a complement in the dative case and introduces directional PPs. A number of prepositions, including rjadom s ‘near’, pered ‘in front of ’, and nad ‘above’, introduce locational PPs only. Such prepositions cannot be used with verbs of motion to introduce the endpoint of motion, cf. (25).16 (25)

Mal’čik pribežal k škole / *pered školoj / boy pri-ran.pfv toward school.dat / in.front.of school.instr / *rjadom so školoj near school.instr ‘The boy ran to school / *to in front of the school / *to near the school.’

16. The prepositions in question are incompatible with the meaning of endpoint of motion independently of the verb’s prefix or aspectual class: they cannot introduce goals with any verb that lexicalizes motion.

286 Tatiana Nikitina

If my analysis is correct and certain change of state verbs can indeed express the endpoint of motion with either directional or locational PPs, we would expect them to combine freely with the unambiguously locational PPs, since no directional PP is available in such cases to express the same meaning. We would then expect inherently locational prepositions to be acceptable in examples like (24) independently of any considerations relating to the event construal. If, on the other hand, change of state verbs impose some additional restrictions on the use of locational phrases, PPs with unambiguously locational prepositions should be incompatible with them just like the locational PPs in (24). The crucial data is provided in (26). It supports the former hypothesis and rules out the latter. (26) a. b.

Položi ključi pered korzinoj / rjadom s korzinoj put.pfv keys in.front.of basket.instr / near basket.instr ‘Put the keys in front of / near the basket.’ Postav’ čašku pered rakovinoj / nad rakovinoj put.pfv cup in.front.of sink.instr / above sink.instr ‘Put the cup in front of / above the sink.’

The contrast with (24) shows that acceptability of a locational phrase with a change of state verb is dependent on the actual preposition used, and in particular, on the availability of a competing (directional) variant for describing the same resulting position. The apparently categorical pattern in (24) is explained not by additional restrictions on locative marking but rather by a preference for the directional variant. When the same configuration can be described, using the same preposition, either in a directional or in a locational phrase, the locational PP may be dispreferred. When the directional equivalent is not available (as defined by the lexical properties of the preposition), the locational variant is invariably acceptable. This once again supports the initial hypothesis that both variants are available with change of state verbs, and the choice between them is determined in context based on a number of factors, including event construal (only relevant when both variants are available) and availability (and relative markedness) of the competing variant. A full account of the directional vs. locational alternation should address differences between individual prepositions as well as the general pattern.

4.

Conclusion

The data discussed in this paper has important implications for the study of the typology of motion expressions. It suggests that the satellite-framed vs. verbframed distinction does not provide a full account of cross- or intra-linguistic

Variation in the encoding of endpoints of motion in Russian 287

Table 4.â•‡ Interpretation of directional and locational PPs Verb class

Type of PP

Interpretation

motion verbs

directional

goal argument

change of state verbs

directional

result; change of location brought about along with a change of state static location; change of location can be inferred but is left unexpressed

locational

variation in encoding endpoints of motion, since it only applies to instances of overt encoding of the directional meaning. It is not sufficient, however, to explain examples where the endpoint of motion is not encoded either by a specialized satellite or in the verb. The fact that a change of location can be inferred with locational PPs with change of state verbs suggests that an endpoint of motion does not have to be encoded overtly and a language may rely on contextual inference rather than on lexical encoding of the directional meaning.17 The study demonstrates once again that the set of expressions a language may use to describe directed motion is typically not limited to one construction. Alternative ways of expressing the same meaning can be used with different or partly overlapping classes of verbs. Accordingly, in describing the use of competing strategies in Russian I distinguished between (i) lexical constraints imposed by verbs of different classes, and (ii) factors involved in the choice of a variant when more than one option is allowed. This distinction turns out to be essential for characterizing the use of locational PPs to describe endpoints of motion. The lexical constraints determine the set of verbs that can combine with either directional or locational PPs without a difference in meaning. I suggested that the relevant distinction is that between goal arguments of motion verbs and endpoints of motion associated with verbs of non-physical change of state. Goal arguments must be expressed by directional PPs. Endpoints of motion associated with change of state verbs can be expressed by directional result phrases; alternatively, a change of state verb can be modified by a locational PP, leaving the event of motion to be inferred. This difference is summarized in Table 4. The distinction between goal arguments and results correlates, rather surprisingly, with a whole set of factors, namely, (i) with acceptability of the locational marking for endpoints of motion, (ii) with acceptability of descriptions of source,

17. For examples of inference of the directional meaning based on broader context, see Jones (1983) on French, Aske (1989) on Spanish, Nikitina (2008) on English; additional difficulties with applying the satellite-framed vs. verb-framed distinction to different languages are discussed in Zlatev and Yangklang (2004); see also references therein.

288 Tatiana Nikitina

and (iii) with the way of interpreting motion as coinciding, following, or preceding the additional event described by the verb. Finally, in cases where both locational and directional PPs can express the endpoint of motion with a given verb and a given preposition, the choice is determined by event construal, or the relative conceptual prominence of different subparts of the complex motion event. Where the resulting location can only be described by an inherently locational preposition, the locational PP is used independently of other factors.

References Aske, J. 1989. Path predicates in English and Spanish: A closer look. Proceedings of the Fifteenth Annual Meeting of the Berkeley Linguistic Society, 1–14. Berkeley CA: Berkeley Linguistics Society. Beavers, J., Levin, B. & Tham, S. W. Forthcoming. The typology of motion events revisited. Journal of Linguistics 46(3). Belichova-Krzhizhkova, E. 1974. Prostranstvennaja determinacija v russkom i v češskom jazykax. In Konfrontační studium ruské a české gramatiky a slovní zásoby: Sopostavitel’noe izučenie grammatiki i leksiki russkogo i češskogo jazykov, Vol. 1, M. Zatovkaniuk, T. I. Konstantinova & A. G. Shirokova (eds.), 93–118. Prague: Universita Karlova. Blazhev, B. I. 1988. Upotreblenie konstrukcii napravlenija i mesta v sovremennom russkom ja‑ zyke, 2nd edn. Sofia: Narodna Prosveta. Bohnemeyer, J., Enfield, N. J., Essegbey, J., Ibarretxe-Antuñano, I., Kita, S., Lüpke, F. & Ameka, F. K. 2007. Principles of event segmentation in language: The case of motion events. Lan‑ guage 83(3): 495–532. Carter, R. 1988. Compositionality and polysemy. In On Linking: Papers by Richard Carter, B.Â€Levin & C. Tenny (eds.), 167–204. [Lexicon Project Working Papers 25]. Cambridge MA: Center for Cognitive Science, MIT. Cienki, A. J. 1989. Spatial Cognition and the Semantics of Prepositions in English, Polish, and Russian. Munich: Verlag Otto Sagner. Comrie, B. 1986. On delimiting cases. In Case in Slavic, R. D. Brecht & J. S. Levine (eds.), 86–106. Columbus OH: Slavica. Croft, W. 1991. Syntactic Categories and Grammatical Relations: The Cognitive Organization of Information. Chicago IL: The University of Chicago Press. Goldberg, A. E. 1995. Constructions: A Construction Grammar Approach to Argument Struc‑ ture. Chicago IL: The University of Chicago Press. Israeli, A. 2004. Case choice in placement verbs in Russian. Glossos 5: 1–54. Jones, M. A. 1983. Speculations on the expression of movement in French. In A Festschrift for Peter Wexler, J. Durand (ed.), 165–194. Occasional Papers No 27 of the Department of Language and Linguistics, University of Essex. Levin, B. & Rapoport, T. R. 1988. Lexical subordination. CLS 24: 275–289.

Variation in the encoding of endpoints of motion in Russian 289

Levin, B. & Rappaport Hovav, M. 1991. Wiping the slate clean: A lexical semantic exploration. Cognition 41: 123–151. Luraghi, S. 2003. On the Meaning of Prepositions and Cases: The Expression of Semantic Roles in Ancient Greek. Amsterdam: John Benjamins. Nedashkivska, A. 2001. Whither or where: Case choice and verbs of placement in contemporary Ukrainian. Journal of Slavic Linguistics 9(2): 213–251. Nikitina, T. 2008. Pragmatic factors and variation in the expression of spatial goals: The case of into vs. in. In Syntax and Semantics of Spatial P, A. Asbury, J. Dotlačil, B. Gehrke & R.Â€Nouwen (eds.), 175–195. Amsterdam: John Benjamins. Slobin, D. I. 1996. Two ways to travel: Verbs of motion in English and Spanish. In Grammatical Constructions: Their Form and Meaning, M. Shibatani & S. A. Thompson (eds.), 195–220. Oxford: Clarendon Press. Slobin, D. I. 2003. Language and thought online: Cognitive consequences of linguistic relativity. In Language in Mind: Advances in the Study of Language and Thought, D. Gentner & S.Â€Goldin-Meadow (eds.), 157–192. Cambridge MA: The MIT Press. Slobin, D. I. 2004. The many ways to search for a frog: Linguistic typology and the expression of motion events. In Relating Events in Narrative, Vol. 2: Typological and Contextual Perspec‑ tives, S. Strömqvist & L. Verhoeven (eds.), 219–257. Mahwah NJ: Lawrence Erlbaum. Slobin, D. I. 2006. What makes manner of motion salient? Explorations in linguistic typology, discourse, and cognition. In Space in Languages: Linguistic Systems and Cognitive Catego‑ ries, M. Hickmann & S. Robert (eds.), 59–81. Amsterdam: John Benjamins. Smyth, H. W. 1920. Greek Grammar. Cambridge MA: Harvard University Press. Stanisheva, D. S. 1966. Vinitel’nyj padež v vostočnoslavjanskix jazykax. Sofia: Bolgarskaja Akademija Nauk. Talmy, L. 1975. Semantics and syntax of motion. In Syntax and Semantics, Vol. 4, J. P. Kimball (ed.), 181–238. New York NY: Academic Press. Talmy, L. 1985. Lexicalization patterns: Semantic structure in lexical forms. In Language Typol‑ ogy and Syntactic Description, vol. 3, T. Shopen (ed.), 57–149. Cambridge: CUP. Talmy, L. 1991. Path to realization: A typology of event integration. Buffalo Papers in Linguistics 1: 91–01, 147–187. Talmy, L. 2000. Toward a Cognitive Semantics. Cambridge MA: The MIT Press. Timberlake, A. 1993. Russian. In The Slavonic Languages, B. Comrie & G. G. Corbett (eds.), 827–886. London: Routledge. Timberlake, A. 2004. A Reference Grammar of Russian. Cambridge: CUP. Toporov, V. N. 1961. Lokativ v slavjanskix jazykax. Moskva: Akademija Nauk SSSR. Ungermanová, M. 2005. Locative complements of verbs of movement in Czech: Some typical structures and their interpretation. In Adpositions of Movement, H. Cuyckens, W. de Mulder & T. Mortelmans (eds.), 87–113. Amsterdam: John Benjamins. Zlatev, J. & Yangklang, P. 2004. A third way to travel: The place of Thai in motion-event typology. In Relating Events in Narrative, Vol. 2: Typological and Contextual Perspectives, S.Â€Strömqvist & L. Verhoeven (eds.), 159–190. Mahwah NJ: Lawrence Erlbaum.

290 Tatiana Nikitina

Appendix. Abbreviations acc dat def gen instr ipfv loc neu prt pfv pl refl

accusative dative definite genitive instrumental imperfective locative neutral participle perfective plural reflexive

chapter 12

Verbs of rotation in Russian and Polish Ekaterina V. Rakhilina

Institute of Russian Language, Moscow

The aim of this article is to provide an account of possible parameters of semantic variation within the domain of rotation in Russian and Polish. The paper presents a corpus-based analysis (supplemented by questionnaires) of five Russian and five Polish verbs, both cognates and not. The results of this study show a considerable difference between verbal frames of rotation lexicalized in both languages despite of their close relations. In particular, the speed of rotation is relevant for Polish verbs, but not for their Russian counterparts. On the contrary, the difference between controlled and uncontrolled rotation is relevant for Russian, but not for Polish. It is argued that the comparison of cognates from genetically related languages is not only possible, but can be extremely useful in the typological perspective.

1.

Introduction

The paper describes the semantics of verbs of rotation (like turn, spin, revolve, etc.) in two closely related Slavic languages, Russian and Polish. The core of this semantic group consists of four pairs of etymological cognates, which historically derive from the same proto-Slavic root and still preserve a significant phonetic resemblance: cf., Russian / Polish krutit’sja / kręcić się, vertet’sja / wiercić się, vraščat’sja / obracać się, kružit’ / krążyć. Both Russian and Polish dictionaries treat . The analysis is based on corpus data (taken mainly, though not exclusively, from RNC, PWN and IPI PAN corpora; for the detailed information see the list in the end of the paper), as well as on specifically targeted questionnaires that had to supplement corpus examples when needed. I sincerely appreciate the assistance of Irina Prokof ’eva in collecting Polish data; I am also grateful to Magdalena Danilewiczowa, Olga Katrečko and Valentina Kul’pina for their comments on the Russian version of this paper (see Rakhilina & Prokof ’eva 2004), and to all the native speakers, both Russian and Polish, who filled in the questionnaires and tested the corpus examples. I would like to thank Vladimir Plungian, Pavel Novičkov, and Alex Loony for their comments and suggestions, which have been particularly helpful in the preparation of this paper, as well as two anonymous reviewers of the fist version of this paper.

292 Ekaterina V. Rakhilina

these verbs as close synonyms: at least, their semantic representations look nearly identical. Thus, Evgen’eva (1981) uses vertet’sja and vraščat’sja for the explanation of krutit’sja; vraščat’sja, kružit’sja while explaning vertet’sja and ‘to move along the circle’, ‘make circles’ for all the three (krutit’sja, vraščat’sja, kružit’sja). The same similarity and vagueness is present in Polish dictionaries as well (PSJP and SJPSz): wirować is used for the semantic explanation of kręcić się; obracać się is explained with the help of both kręcić się and wirować: and so on (a more detailed analysis of Polish and Russian dictionaries was given in Rakhilina & Prokof ’eva 2004). Unexpectedly, the members of these pairs display considerable and significant semantic differences. For example, it is impossible to substitute one verb for another when describing quite simple and frequent situations of circular motion, e.g., compare the following equivalent examples from Russian (1a) and Polish (1b): (1) a. b.

Zemlja vraščajetsja (*krutitsja / ??vertitsja / * kružitsja) vokrug Solnca. Earth rotates around Sun Ziemia krąży (*obraca się / *kręci się / *wierci się / *wiruje) Earth rotates wokół Słońca. around Sun ‘The Earth circles the Sun.’

This derives from the well-known fact that semantic changes in natural languages occur much faster than phonetic ones: while one and the same phonetic form can be preserved for centuries, semantic changes can occur several times within a single generation’s lifespan. Therefore the comparison of cognates is a legitimate part of a lexical-typological study, cf., Majsak & Rakhilina (2007) and Koptjevskaja-Tamm (2008). In our case, only two related languages are taken into account. In this respect, I follow the approach adopted in Fillmore & Atkins (2000) seminal study of the concept of crawling in English and French. Based on “linguistic evidence” in Wierzbicka’s sense (1985), as well as on principles of the Moscow Semantic School (see Apresjan 2000, among others), this paper is focused on what can be called . Some interesting reflections on the semantics of Polish rotation verbs can be found in Bojar (1979). However, this book considers the whole range of Polish motion verbs and not specifically the domain of rotation. . Here, as usually, the asterisk is used to indicate ungrammatical sentences; the double question mark stands for sentences, the grammaticality of which is very marginal. . All the verbs of rotation are translated in glosses as ‘rotate’. Subtle semantic differences between various Russian and Polish verbs are discussed directly in the bulk of the paper.

Verbs of rotation in Russian and Polish 293

“contextual behavior” of verbal lexemes. This type of data allows us to isolate a set of semantic parameters relevant to each pair of cognates, such as inner/outer axis of rotation, incomplete/complete revolution, rotation above the Landmark, rotation combined with forward motion, etc. The article is divided into 6 sections. Sections 2–4 describe contrastive semantics of the Russian and Polish cognates: krutit’sja & vertet’sja vs. kręcić się & wiercić się (Section 2), vraščat’sja vs. obracać się (Section 3), and kružit’(sja) vs. krążyć (Section 4). Section 5 includes a discussion of the Polish verb wirować, which does not have a Russian cognate, but occupies an important place within the domain of Polish verbs of rotation. Section 6 summarizes the results of the analysis presented.

2.

Russian krutit’sja & vertet’sja vs. Polish kręcić się & wiercić się

In this section I will consider two Russian verbs, which are very close to each other in meaning – krutit’sja and vertet’sja (a native speaker of Russian can easily remember the first lines of a well-known song in which both verbs seem to express one and the same meaning: krutitsja, vertitsja šar goluboj ‘a blue ball is turning round’). The Russian verbs will be compared with their Polish correlates kręcić się and wiercić się.

2.1

Russian

Both Russian verbs refer to the inner axis of rotation, so they can describe rotation of the wheel. Interestingly, in some cases concerning wheel rotation, these verbs are not interchangeable because in the context of a purposeful (or goal-Â�directed) movement vertet’sja is normally excluded, cf.: (2) S lëgkim žužžaniem krutilos’ / *vertelos’ vereteno. with light whirr rotated spindle ‘The spindle was turning with a slight whirr.’ (Kuprin. “Olesja”)

. The terms “Trajector” (a moving entity, or Figure) and “Landmark” (a stable entity, setting the coordinates of the movement, or Ground) are used in the description of verbs of motion following Langacker (1987:â•›231ff.) and other studies written within the framework of Cognitive Grammar. . Most of the Russian literary examples are taken form Russian National Corpus. They are specified for the author and marked “RNC”. The rest of the Russian literary examples from

294 Ekaterina V. Rakhilina

(3) Celyj den’ krutitsja / *vertitsja polivalka. whole day rotates sprinkler ‘The sprinkler is rotating the whole day.’

(Granin. RNC)

Therefore, it is krutit’sja, and not vertet’sja, which is used in negative constructions with the specific meaning ‘does not work, is broken’ (used for rotating parts of artifacts): (4) Švejnaja mašinka slomalas’: u neë ručka ne krutitsja. sewing machine broke at it handle not rotates ‘The sewing machine is out of order/has broken: its handle does not work.’

Incompatibility of vertet’sja with the idea of purpose or goal can be also illustrated by the context of animate subjects: (5) On nabljudal, kak pod kupolom cirka krutilis’ / *vertelis’ gimnasty. he observed how under top of.circus rotated gymnasts ‘He watched the gymnasts doing acrobatics / spinning high up under the big top of the circus.’

On the contrary, the context of uncontrolled motion prefers vertet’sja: (6) Skripja pružinami, vertelsja (??krutilsja) na kojke Gvozdev. creaking with.springs rotated on bed Gvozdev ‘Gvozdev was tossing and turning in his bed, with the bedsprings creaking.’ (Evgen’eva 1981)

Accordingly, in cases when the verbs seem to be interchangeable, there is always a slight shift to uncontrolled motion if vertet’sja is used, and a shift to purposeful interpretation for krutit’sja. Thus, the wheels of a moving car are normally described by means of krutit’sja (purposeful motion), and the wheels of an upturned car (uncontrolled motion) are more likely to be described by means of vertet’sja. This is reflected in the behavior of different Trajectors that typically move following the inner axis of rotation: some of them are drawn to vertet’sja and some to krutit’sja (Table 1) depending on the characteristics of the motion they display. The table shows that the verb krutit’sja is semantically more general than vertet’sja, as it is able to cover all the cases available for the latter. This is obviously motivated by the fact that an uncontrolled situation can easily be reinterpreted as purposeful, thus allowing the use of krutit’sja in the contexts where vertet’sja is appropriate, while the converse is false. well-known classical works are specified for the author and the title. Polish literary examples are all taken from Polish corpora.

Verbs of rotation in Russian and Polish 295

Table 1.â•‡ Russian krutit’sja and vertet’sja Trajector

krutit’sja

vertet’sja

koleso ‘wheel’ karusel’ ‘carousel’ fljuger ‘weather vane’ motor ‘motor engine’ mel’ničnoe koleso ‘millwheel’ rul’ ‘steering wheel’ koleso obozrenija ‘observation wheel’

+ + + + + + +

+ ? + ? + ? ?

Both verbs demonstrate a semantic extension to the domain of ‘chaotic motion’ (i.e., aimless motion in different directions), which seems to be very typical for verbs of rotation in general. Vertet’sja describes chaotic motion of a person turning round spontaneously in different directions, for example, while sitting on a chair or standing in front of a mirror. In this case, the use of krutit’sja (presupposing the goal-directed interpretation of the same situation) is impossible. Another type of chaotic motion described by vertet’sja bleaches the idea of rotation, cf., vertitsja u menja pod nogami ‘be under one’s feet’ (lit. ‘turns under my feet’). The internal axis of rotation is changed into the external Landmark specifying the space where the chaotic motion takes place. Similar contexts are also characteristic of krutit’sja, which adds a component of purposefulness demanding the Landmark to be animate, cf.: (7) On vsegda krutitsja / *vertitsja vozle staršix. he always rotates near elders ‘He always hangs around with an older crowd.’

or: (8) […] ètot kapitaniška, kotoryj krutitsja / *vertitsja okolo neë. this little.captain, which rotates near her ‘[…] that captain fellow who is always following her around.’

(NBARS)

Note that purposefulness is characteristic for other metaphorical uses of krutit’sja as well, like krutit’sja / *vertet’sja s utra do večera ‘to run around from dawn till dusk’ (lit. ‘to turn from morning to night’, cf., also den’gi krutjatsja / *vertjatsja (lit. ‘the money rotates’) meaning money being active), where the human controller and manager of the whole situation is presupposed.

. A similar (though not identical) effect is observed for kružit’ and its cognate (see below).

296 Ekaterina V. Rakhilina

In case the chaotic motion is applied to purely inanimate subjects, such as subconscious ideas or unpronounced words, the difference between krutit’sja and vertet’sja is neutralized, cf.: (9) Množestvo myslej vertelos’ / krutilos’ u menja v golove. a.lot of.thoughts rotated with me in head ‘A lot of thoughts were going round in my head.’ (Bulgakov. RNC)

2.2

Polish

The Polish pair kręcić się ~ wiercić się differs from its Russian cognates in many respects. There is practically no semantic overlapping of the verbs: wiercić się is used only with animate subjects, and the whole domain of inanimate Trajectors with the inner axis of rotation is covered by kręcić się invariably. Neither a wheel nor a weather vane can be described by means of wiercić się. Thus, the first line of a well-known song of Bulat Okujava, “Poka Zemlja eščë vertitsja” (‘While the Earth is still going round’), cannot be translated into Polish with the help of the cognate wiercić się: the actual translation is Dopóki nam Ziemia kręci się. As for animate subjects, the dictionaries consider kręcić się (as well as wiercić się) possible with them; however, according to our data, wiercić się is strongly preferred in contexts like wierci się na krześle ‘to twist on the chair’: (10)

Niech pan się tak nie wierci – powiedział tata – powrzuca let mister himself so not rotate – responded father – will.throw pan wszystkich do wody! mister everybody to water ‘Please, sit still and don’t twist like this, said the father, or you will throw us into the water!’ (PWN)

Polish wiercić się, like Russian vertet’sja and its quasi-synonym krutit’sja, requires an internal axis of rotation, but the semantics of the other Polish cognate kręcić się is somewhat broader. It can be also applied to long, flexible Trajectors wrapped around a certain external axis. This axis is not arbitrary; it reproduces the original prototypical form and orientation of the Trajector itself, e.g., curly hair or a winding road. In Russian, this additional semantic domain is covered by a special verb, vit’sja, e.g., compare the following example from Polish (11a) and its Russian translation (11b): (11) a. Droga kręciła się miejscami samą krawędzią pól. road rotated in.places by.very edge of.fields

Verbs of rotation in Russian and Polish 297

b. V nekotoryx mestax doroga vilas’ vozle samogo kraja polja. in some places road twined near very edge of.field ‘In some places the road was winding along the very edge of the field.’ (Grochola)

The difference is that Russian vit’sja expresses both active motion (as ‘waving flags’) and the resulting state of this motion (as ‘curly hair’), while Polish kręcić się cannot describe the actual rotation, but only the state resulting from it. No matter how different the meaning of kręcić się is from its Russian cognate, it displays the same semantic shift from rotation to chaotic motion. In these cases (similarly to krutit’sa), it occurs mainly with animate Landmarks, cf.: (12)

Około księcia kręciło się niemało dworaków służących i near prince rotated a.lot of.servants serving and schlebiających dla własnej korzyści, ale orli umysł Jeremiego pleasing for own profit but eagle’s thought of.Jeremiah wiedział dobrze, co o kim trzymać. knew well what about whom to.keep ‘There were always a lot of servants around the prince, who were serving and wheedling him for their own benefit, but Jeremiah’s keen eye discerned who was who perfectly.’ (Sienkiewicz)

However, in this case the range of uses of the Polish verb is also broader as compared to Russian. For example, kręcić się can describe not only purposeful motion in the vicinity of a person (or an inanimate object), but also chaotic motion inside some limited space, which normally corresponds to Russian begat’ po, krutit’sja being exluded in these contexts. Cf., the following example from Polish (13a) and its Russian translation (13b): (13) a. b.

Dom jest czysty i zadbany, po pokoju kręci się dwójka dzieci. house is clean and groomed on oom rotates pair of.children Dom čistyj i uxožennyj, po komnate begajut (*krutjatsja) dvoe house clean and groomed on room run (*rotate) pair detej. of.children ‘The house is clean and well-managed; two children are running around in the room.’ (Grochola)

. This semantic development is a typical example of what can be called “metonymic shift” in meaning; for a recent discussion of these issues, cf., Traugott & Dasher (2002), Kustova (2004), and Padučeva (2004).

298 Ekaterina V. Rakhilina

3.

Russian vraščat’sja vs. Polish obracać się

Etymologically and morphologically, the closest correlate to Polish obracać się is Russian oboračivat’sja, but in contemporary Russian its only meaning is ‘to turn one’s head when called’. This meaning (among others) is retained by obracać się, but it belongs to the periphery of the semantic domain of rotation; therefore the comparison of the true pair of cognates oboračivat’sja ~ obracać się is of little interest. Oboračivat’sja has an Old Church Slavonic counterpart obraščat’sja in Russian, which is also the true cognate of Polish obracać się; however, semantically it does not pertain to the rotation domain anymore, meaning only ‘address, turn to’. Thus, when looking for a Russian correlate of Polish obracać się, we have to choose another verb, not a cognate in the strict sense, but still having a close morphological structure and comparable semantics – vraščat’sja. Vraščat’sja denotes rather slow, continuous, and monotonous motion of objects, characteristic of mechanisms and their parts (turbines, parts of machines, etc.). Lasting steady rotation presupposes a sequence of turns. Thus, combinations like ??kran ne vraščaetsja (lit. ‘a tap does not rotate’) look strange: the right expression should be either ne povoračivaetsja ‘does not turn’ (because povoračivat’sja denotes an incomplete turn) or ne krutitsja (because krutit’sja is not sensitive to the number of turns). Interestingly, vraščat’sja – or rather its active participle vraščajuščijsja – is used as a general technical term for artifacts designed for rotation, cf.: vraščajuščeesja kreslo ‘rotating chair’, vraščajuščajasja dver’ ‘revolving door’, etc. The participle krutjaščijsa (from krutit’sja) or vertjaščijsja (from vertet’sja) is not used in this case, since it would denote actual and not potential rotation. Another characteristic of vraščat’sja is that it does not take animate subjects and (prototypically) body parts. In general, the mechanical nature of the motion denoted by this verb is not suitable for living beings, and it is impossible to say *mal’čik vraščalsja na stule ‘a boy was turning to and fro on his chair’. On the contrary, planetary motion is a very typical situation for vraščat’sja: being endless, orderly, and monotonous (as if mechanical) rotation with the inner or outer axis, it fully satisfies all the semantic restrictions. The corresponding deverbal noun, vraščenie, is also used as a general term for this type of motion, because vraščat’sja is applicable both to circumrotation and revolution and thus occupies the most neutral position in the semantic field. In Polish, it is kręcić się which seems to serve as a general term occupying the most neutral position in the system. Thus, kręcić się describes planetary motion and is, to a certain extent, indifferent to the place of the axis. The verb obracać się plays another role. Unlike Russian vraščat’sja, it prefers the inner axis and, like the other Russian verb, povoračivat’sja, can also describe a single rotation, both

Verbs of rotation in Russian and Polish 299

complete and incomplete. Cf.: obrócić się na pięcie / na drugi bok ‘turn on one’s heels / turn over on one’s side’; cf., also the following contexts with inanimate subjects: (14) Nie mogę otworzyć zamka: klucze nie obracają się. not I.can open lock keys not rotate ‘I cannot unlock the door: the key does not turn.’

In the case of continuous rotation, the motion should not be very quick, because individual turns have to be observable. Hence, wheels (including millwheels) are suitable for obracać się, but not carousels or weather vanes, which are faster and prefer the more general kręcić się, cf.: (15)

Patrzył z ogromnym skupieniem, jak pod białą pustynią he.looked with immense attention how under white desert sufitu obraca się (??kręci się) powoli śmiga wentylatora. of.ceiling rotates slowly propeller of.fan ‘He observed attentively the propeller of a fan rotating slowly under the white desert of the ceiling.’ (Żukrowski)

This restriction is motivated by the semantic structure of obracać się, as the corresponding motion consists of discrete turns (which can be many, few in number, occur only once, or even constitute an incomplete turn, cf., Bojar 1979:â•›38). Despite the fact that the primary meanings of Russian vraščat’sja and Polish obracać się are not very similar, both verbs display the same semantic shift yielding the reading ‘to mix in society’, as in the following examples (see also Evgen’eva 1981; SJPSz 1995): (16)

Olga […] ènergična, zamužem za amerikanskim pisatelem, vraščaetsja Olga energetic married to American writer rotates v izdatel’skom mire. in publishing world ‘Olga […] is active, married to an American writer, familiar with the publishing world.’ (Solženicyn. “Ugodilo zërnyško promež dvux žernovov”)

(17)

Schmidt-Holtz nie ma doświadczenia w dziedzinie muzyki, ale Schmidt-Holtz not has competence in field of.music but obracał się wśród artystów. rotated among artists ‘Schmidt-Holtz had no musical education, but he lived among artists.’ (PWN)

300 Ekaterina V. Rakhilina

This meaning existed in both languages in the nineteenth century, but since that time it has become less neutral in Russian and now applies only to high society, often ironically, cf.: (18) […] kogda uže sovsem stanovilos’ nevmogotu, ona govorila sebe, when already entirely became unbearable she told herself čto, verno, budni aktrisy Giacintovoj ne namnogo that probably weekdays of.actress of.Giacintova not much raznoobraznee eë budnej, […] – darom čto ona stoličnaja more.various than.her weekdays though that she metropolitan primadonna, vraščaetsja i voobšče. prima.donna rotates and in.general ‘When in despair, she told herself, that the everyday life of a famous actress Giacintova did not differ a lot from her own, though Giacintova was a prima donna and lived in a big city, frequented the society and all that’.

Polish оbracać się has no similar constraints and can describe any type of society, cf.: (19) 14-letni syn [...] obracał się w złym towarzystwie. 14-year son rotated in bad company ‘His fourteen-year-old sonÂ€has started hanging around with a bad crowd.’ (20)

4.

[…] chłopca, który już od dawna obraca się w kręgach boy which already since long rotates in circles przestępczych. criminal ‘[…] a boy who has been hanging out with the wrong guys for a long time.’ (Siemaszko)

Russian kružit’(sja) vs. Polish krążyć

In this domain, the difficulty arises from the asymmetry of Russian and Polish data: Russian has two verbs, kružit’ and its reflexive counterpart kružit’sja, both similar to Polish krążyć. Russian kružit’ (which can also be used transitively as a causative with the meaning ‘to turn smth. / smb.’) has a small sub-class of nontransitive uses with the meaning ‘make circles, hover’ as in (21). In these contexts, it is commonly regarded as a synonym to kružit’sja, cf., (22): (21) Na blednom nebe jastreb kružit. on pale sky hawk rotates ‘A hawk is making circles in the pale sky.’

(Blok. “Vozmezdie”)

Verbs of rotation in Russian and Polish 301

(22) Nad gorodom vysoko kružilis’ nemeckie samolëty. over town high rotated German airplanes ‘German planes were circling high over the town.’

Polish krążyć can easily be used to translate Russian kružit’ / kružit’sja in these contexts, cf., the following examples from the PWN: (23) Jastrząb krążył nad wodą, szukając swej ofiary. hawk rotated over water, searching its victim ‘The hawk hovered over the water waiting for its victim.’ (24)

Nasz samolot kwadrans krążył w powietrzu, bo nie było our airplane 15.minutes rotated in air because not was miejsca na lądowanie. place for landing ‘Our plane circled in the air for 15 minutes, because there was no place for landing.’

Since Polish does not have a reflexive counterpart for Russian kružit’sja, Polish krążyć could be considered a cognate for both Russian verbs, kružit’ and kružit’sja, though semantically it is identical to neither of them. The Russian intransitive kružit’ requires that the Trajector be situated above its Landmark, as in (21). This restriction appears to be rather rigid, so that the most common candidates for this role are birds, flying insects, and planes. One can see that these Trajectors are mostly living beings (except planes, but planes are normally conceptualized as birds, see Rakhilina 2000:â•›300; Rakhilina 2004) which are located far from the corresponding Landmark; normally, they are moving in search of something like food, a victim, a nest, or a place for landing. This situation is highly prototypical for kružit’, and it is reproduced in its metaphorical uses, meaning ‘to mill around’, ‘to move along a winding path, looking for something’. In this case, kružit’sja is no longer a verb of rotation, because it describes a motion without real “circles”, although the general configuration of this motion is still reminiscent of original rotation. In such cases, it is legitimate to speak about the metaphorical situation, where the main participants are basically the same: (25)

Po doroge ix zaxvatila metel’, oni dolgo kružili i priexali on road them captured snowstorm they long rotated and arrived k mestu ne v polden’, kak xoteli, a tol’ko k večeru, kogda to place not in midday as wanted but only to evening when

. In case the Trajector is not expressed in the sentence, one can assume that this role is taken up by the observer.

302 Ekaterina V. Rakhilina

uže bylo temno. already was dark ‘On their way, the snowstorm began; they spent a lot of time looking for the right place (lit. ‘making circles’) and they reached their destination not at midday as it had been planned, but at night, when it was already dark’. (Čexov. RNC)

Kružit’sja denotes quite a different type of motion: Trajector and Landmark are situated at the same level, and the motion is translational, so that the situation can be described as “progressive rotation”. This type of motion is especially characteristic of people dancing together, snowflakes, or falling leaves, cf.: (26)

I mne kazalos’, čto vse, tak že, kak i ja, ne moljatsja, and to.me seemed that everybody as well as and I not pray a prosto smotrjat na ètot dymok, kak on podnimaetsja strujkami, but simply look at this smoke how it rises in.wisps kružitsja i nesëtsja vverx, k sinemu, zamerzšemu oknu. rotates and rushes up to blue frozen window ‘And it seemed to me that all the other people did not pray either, but looked at that small stream of smoke, observing it rise in wisps and move in circles up to the blue frozen window.’ (Kaverin. RNC)

The fact that the verb kružit’sja often occurs with the other verbs of motion and looks very natural in these contexts underlines the translational component in its semantics: (27) On čuvstvoval, čto samolët kružitsja, nesëtsja k zemle. he felt that airplane rotates rushes to earth ‘He felt that the plane was spinning and hurtling to the ground.’

In contrast to kružit’, the motion described by kružit’sja has no Landmark. This characteristic of kružit’sja is related to another interesting feature of this verb, also opposed to kružit’: unlike kružit’, kružit’sja usually denotes spontaneous uncontrolled motion without any goal, set by the Landmark. Thus, kružit’sja (but not kružit’) is applied mostly to inanimate Trajectors like leaves, snowflakes, etc., as stated above. These subjects normally do not occur in the context of the purposeoriented kružit’. Additional evidence for the spontaneity of motion denoted by kružit’sja is that this verb can be applied to the special physiological state of a person, who is not able to perceive reality adequately, so that all the objects in the environment appear to be moving chaotically, making circles:

Verbs of rotation in Russian and Polish 303

(28)

Fyrkajut koni, begut v rovnoj stepi, i kažetsja Nataše, chto spit horses run in plane steppe and seems to.Natasha that kružitsja step’ i begut lošadi kak-to nazad. rotates steppe and run horses somehow backward ‘The horses are spitting, they are running through the smooth steppe, and it seems to Natasha that the steppe is turning round and the horses are running back somehow.’ (Garin-Mixajlovskij. RNC)

This kind of usages gives rise to a metonymical phrase golova kružit’sja ‘(my) head swims’. Notice that kružit’ is never possible in these “physiological” contexts. At the same time, there are examples in which these verbs are interchangeable, as illustrated at the beginning of the section. However, if we look at them more carefully, we see that the corresponding situations are not at all identical. Actually, jastreb kružitsja ‘the hawk makes circles’ is less purpose-oriented than jastreb kružit; in particular, the trajectory of the movement denoted by the reflexive verb (kružit’sja) is less determined than in the case of kružit’ and allows non-rotational motion. As for kružit’, it presupposes some goal (for example, a search for a certain victim). To illustrate our explanation, let us consider one more example of kružit’sja, concerning not birds, but flying insects: (29) I muxa na podokonnike kružitsja nad nedoedennym pirožkom. and fly on windowsill rotates over leftover patty ‘And a fly on the windowsill is circling round the leftover patty.’ (Kanovič. “Park zabytyx evreev”)

The example in (29) presupposes that a fly is just aimlessly flying above the patty, more or less around it, not as if a predatory insect coveting the patty as its victim. It is the substitution of reflexive kružit’sja with kružit’ which may yield the latter interpretation. It should be recalled that in Polish both situations are described by the same predicate krążyć, cf., examples (23)–(24) at the beginning of the section as well as a Polish example in (30a) and its translation into Russian in (30b): (30) a. b.

Nad pełnym wiadrem krążyły muchy. over full bucket rotated flies ‘The flies were circling over the full bucket.’ Nad polnym vedrom kružilis’ (??kružili) muxi. over full bucket rotated flies

(Wojdowski)

Roads, paths, trailways, etc. comprise another group of Trajectors that seem to allow the substitution of one verb for another, cf., (31):

304 Ekaterina V. Rakhilina

(31)

Tropa kružitsja / kružit meždu stvolami derev’jev, poroj otčajanno path rotates between trunks of.trees sometimes desperately ustremljaetsja prjamo vverx. rushes directly up ‘The path is winding around the tree trunks, sometimes it is desperately rushing directly upwards.’ (Iskander. “Sandro iz Čegema”)

But even in this case the semantic structure of the phrase changes with the change of the verb: kružitsja means that the path goes forward, although twists a good deal. As for kružit, it appears here in its secondary meaning ‘to move meanderingly, looking for a certain place (Landmark).10 In this respect, Polish krążyć is closer to kružit’ than to kružit’sja, as it also presupposes a certain Landmark for the spinning axis. The Landmark can be situated below the trajectory, similar to Russian kružit’ (see the examples at the beginning of the section), and also at one and the same level with the trajectory, as in (32)–(35): (32) Lódź krąży wokół wyspy. boat rotates round island ‘A boat is making circles round the island.’ (33) Korowód krąży wokół choinki. round.dance rotates round fir.tree ‘People are reeling around the fir tree.’ (34) W cyrku koń chodzi po kole. Krąży po arenie. in â•›circus horse goes on circle Rotates on arena ‘In the circus, the horse goes round. It makes circles in the arena.’ (35)

Byk zaś chcąc się uwolnić krążył wokół drzewa, owijając linę bull however wanting to.get.free, rotated round tree looping rope wokół pnia, aż całkiem skrócił pole manewru. round stub so wholly shortened field of.maneuver ‘Wishing to get free, a bull was walking in circles round the stub looping his rope around the tree, and decreasing maneuver space.’ (UG)

The translation of krążyć as kružit’ is clearly inapplicable to such situations. Furthemore, Russian has no verb of motion that could express the situation of Trajector moving round its Landmark at the same level. The closest equivalent would be

10. Apresjan (1995:â•›251) makes the stronger assertion that the trajectory remains relevant for kružit’ being used in the special contexts like these, because they reproduce not only the Landmark (as a goal of the motion) but also the trace of the motion, formed by circles (kružit’v ‘make circles’ ← krugn ‘circle’, the same for petljat’ ‘to twist’, lit. ‘to make a loop’ ← petlja ‘a loop’).

Verbs of rotation in Russian and Polish 305

phrasal: xodit’ / dvigat’sja krugami ‘to go / move in circles’. Note, that Polish krążyć describes also the circular motion of blood, as in the following example: (36) Krew zaczyna krążyć szybciej w żyłach. blood begins to.rotate quicker in veins ‘Blood in the veins begins to circulate quicker.’

So Polish krążyć does not distinguish between spinning around a Landmark or above a Landmark and can cover the domain of Russian kružit’sja at least partly. In fact, one can find examples, where krążyć applies to snowflakes or leaves moving round in the air, cf., (37): (37) W â•›powietrzu krążą żółtoczerwone liście. in air rotate yellow.red leaves ‘Flame-colored leaves are spinning in the air.’

And also: (38) Krążyć będą wokół niego w wspaniałym tańcu. rotate they.will round him in wonderful dance ‘They will reel around him in a wonderful dance.’

(Saługa)

It is interesting that Polish krążyć includes the metaphorical meanings of Russian kružit’, e.g., compare the following examples from Polish (39a) and Russian (39b): (39) a. b.

Krążyli po całej dzielnicy, zanim doszli do domu. rotated on whole district before arrived to house Oni kružili po vsemu rajonu, poka ne dobralis’ do domu. they rotated on whole district before not arrived to house ‘They had walked in circles round the whole neighborhood before they got home.’

But the class of metaphorical contexts of this kind in Polish is much bigger than in Russian, which confirms that the meaning of the metaphorical krążyć is broader than that of Russian kružit’. Cf., the following examples that cannot be translated with the help of kružit’: (40) a. Pol. b. Rus.

Po mieście krążyły patrole wojska. on town rotated patrols of.army Po gorodu xodili (*kružili) armejskie patruli. on town went (*rotated) military patrols ‘Patrols moved around the town.’

306 Ekaterina V. Rakhilina

The examples (41) and (42) are also highly characteristic of Polish krążyć, describing an abstract motion – in this particular case, news and gossip: (41) a. Pol. b. Rus.

Jakaś plotka musi jednak krążyć o nas po mieście. some gossip must however rotate about us on town Vsë-taki kakie-to spletni o nas, dolžno byt’, xodjat however some gossips about us must be go (*kružat) po gorodu. (*rotate) on town ‘It seems, however, that some whispers concerning us wander about the town.’

(42) a. Pol. b. Rus.

O panu Małyszu niespodziewanie zaczęły krążyć about mister Malysh suddenly began to.rotate dziwne opowieści. strange stories O pane Malyše vdrug stali xodit’ (*kružit’) about mister Malysh syddenly began to.go (*to.rotate) strannye rasskazy. strange stories ‘Suddenly, strange stories began to circulate about pan Malysh.’

It should be noted that in both classes of metaphorical contexts (describing people who lost their way or abstract motion of gossip or ideas) the component of rotation is eliminated and the emphasis is laid on the random nature of the motion. Some Polish examples of this kind have no Landmark at all, with only location being marked. For these cases the proper translation into Russian is not kružit’, but xodit’ or brodit’ (as in (41)–(42)), cf., also: (43) a. Pol. b. Rus.

Dym krążył po całej izbie [...]. smoke rotated on whole room Dym šël (*kružil) po vsej komnate. smoke went (*rotated) on whole room ‘There was smoke all around the room.’

(PWN)

The same effect, as it was shown in Section 1, is relevant for kręcić się. This Polish verb can also “get rid” of the Landmark (and the axis of rotation) and shift its meaning to the domain of chaotic motion. The difference is that, while krążyć is translated into Russian as more or less “slow” motion (xodit’ / brodit’ = ‘go around’), kręcić się needs begat’ ‘run about’, which denotes a much quicker type of motion. And it is not accidental: spinning round the external axis (obligatory for krążyć) has a bigger radius and makes the motion smoother and slower. This

Verbs of rotation in Russian and Polish 307

is why the chaotic motion described by krążyć seems to be not as quick as that of kręcić się and may take up more space.

5.

Polish wirować

Wirować stands apart in the group of Polish verbs of rotation, as it has no Russian cognates. Having a very precise and specific core meaning, it differs significantly from the other members of the Polish group as well. Its meaning is well-captured by the dictionaries, cf., SJPD: ‘to rotate very quickly round the inner axis or (when talking about substances and collections of objects) to whirl so that the result resembles a tornado or whirlwind’. In Russian this special kind of quick motion is not lexicalized in the domain of rotation, so, as already pointed out, wirować has no equivalents and is usually translated by the most general krutit’sja. As shown in Section 1, the latter presupposes the inner axis, but has no special emphasis on the high speed of rotation: it can denote not only slow rotation (cf., plastinka medlenno krutilas’ ‘the record was turning slowly’), but also examples like (44a): (44) a. Pol. b. Rus.

A gdy silnik zaskoczy i krąg wiruje jak szalony. and when engine will.work and circle rotates like mad A kogda motor zarabotaet i krug zakrutitsja kak bešenyj and when engine will.work and circle will.rotate like mad ‘And when the engine is started and the potter’s wheel begins to whirl madly.’

In its metaphorical meanings wirować also correlates with krutit’sja, cf., (45a) similar to (45b): (45) a. Pol. b. Rus.

Tyle pytań wiruje mi w głowie. So.many questions rotate to.me in head Stol’ko voprosov krutitsja u menja v golove. So.many questions rotate at me in head ‘A lot of questions are spinning in my head.’

It can be seen that Polish, unlike Russian, “adds speed” to the abstract motion of thoughts and ideas making circles while coming to mind or leaving one’s head. In this respect it is interesting that the same effect of “speeding up” occurs in another group of abstract motion contexts, concerning mental representations where Russian uses kružit’sja. According to the analysis presented in Section 3, kružit’sja has an external axis and therefore is even “slower” than krutit’sja. Nevertheless, the frame of “swimming” head is expressed by Polish wirować and at the same time by Russian kružit’sja.

308 Ekaterina V. Rakhilina

(46) a. Pol. b. Rus.

6.

Mam okropny ból głowy: trzęsie mną, a czasami I.have awful pain of.head it.shivers me and sometimes wszystko wiruje. everything rotates U menja užasno bolit golova, menja znobit, a at me awfully hurts head I feel.chilly, and inogda pered glazami vsë kružitsja. sometimes before eyes everything rotates ‘I have an awful headache, I feel chilly and my head swims from time to time.’

Conclusion: The systems of rotation verbs and their parameters

Let us summarize our analysis. The lexical systems of two closely related languages (Polish and Russian) structure the semantic field of rotation with one and the same (except for wirować) set of cognates, very similar phonetically, but unexpectedly different semantically. For example, Russian vertet’sja applies both to animate and inanimate subjects, whereas Polish wiercić się describes only animate Trajectors. Important differences in combinability exist also for other pairs. Thus, the scope of Polish kręcić się can include, unlike Russian krutit’sja, long and flexible objects, such as curved hair or a meandering road. On the other hand, Russian kružit’ presupposes that the Trajector is situated above the Landmark (typical examples are eagles or hawks flying above prey), while Polish krążyć is not sensitive to this restriction and tolerates the Trajector and the Landmark located at the same level (a boat going around an island, etc.). Thus, the system of rotation verbs in each language can be thought of as a set of specific and largely independent frames,11 integrating semantic parameters into holistic images. Accordingly, the Russian list of frames can be represented as follows: – krutit’sja – controlled circular motion (the number of turns is irrelevant) of the object round the inner axis; – vertet’sja – uncontrolled (and hence sometimes chaotic) circular motion (with the variation in the direction) round the inner axis; – vraščat’sja – smooth and steady circular motion with numerous turns round an internal or external axis; 11. The term “frame” is used here much in the Fillmorian sense (cf., Fillmore & Atkins 1992). It can be argued that frames (and not, for example, semantic features) are the main conceptual units upon which cross-linguistic comparison in the domain of lexical semantics must be based; for more detail, cf., also Rakhilina (2003).

Verbs of rotation in Russian and Polish 309

– kružit’ – circular motion over the Landmark, the Landmark being far under the Trajector and conceptualized as the Goal of the motion; – kružit’sja – translational motion combined with circular motion round an ideal (external) axis, designed by the entire trajectory. The Polish list of frames looks completely different in many respects. Cf.: – kręcić się – either circular motion of point-like objects round an internal or external axis or the result of imaginary circular motion of long, flexible objects (like roads, ropes, wires, etc.) round themselves; – wiercić się – circular motion with the inner axis of an animate object (primarily, a person) in different directions; – obracać się – circular motion of an inanimate object round the inner axis, when turns are distinct; – krążyć – circular motion round an external Landmark being at a certain distance from the Trajector; – wirować – speedy circular motion round an inner axis, resembling a whirlwind or tornado. Both lists appear nearly incomparable, at least at first glance. Nevertheless, in each case we deal with a certain system with its own basic priorities. Generally, the differences between the two languages may occur due to various factors. The main ones are the following: 1. Sometimes, the languages simply privilege different parameters for lexical distinction. Thus, the speed of rotation is relevant for Polish verbs, but not for their Russian counterparts. On the contrary, the difference between controlled and uncontrolled rotation is relevant for Russian, but not for Polish. 2. Even when the distinctive parameters in both languages seem similar or identical, this may not be the case because of their different internal structure: the linguistic weight accorded to individual values may vary. For example, the opposition of internal and external axis of rotation is relevant both for Russian and Polish, but within each pair of cognates the value of this parameter is not necessarily the same. While Russian krutit’sja presupposes an internal axis, Polish kręcić się allows both an internal and (though in special cases) external one. These and other similar cases may lead us to the conclusion that Russian privileges the axis of rotation much more than Polish. Accordingly, most individual deviations within the lexical systems appear to be linguistically motivated. 3. Driven by different priorities, the cognates more often than not represent a combination of different semantic features within one lexical unit, thus

310 Ekaterina V. Rakhilina

capturing parts of the semantic domains of other lexemes or even semantic fields. So, in the pair exemplified by Rus. kružit’ ~ Pol. krążyć, the Russian lexeme requires an obligatory vertical distance between Trajector and Landmark, while Polish krążyć treats this situation as a subtype of a much more general one: namely, the motion round some external Landmark at a certain distance from it. Russian, on the contrary, has no special lexeme to express circular motion of a Trajector round an external Landmark located at the same level. The only way to express this configuration is to do it indirectly, “borrowing” one of the more general verbs of non-directional motion (e.g., xodit’ ‘to walk; to move in different directions’) combined with a satellite expression (e.g., the manner adverb po krugu ‘circle-wise, in a circle’). 4. The metaphorical meanings of Polish and Russian cognates seem to display more similarity in general. This, of course, is quite natural from the point of view of the Image-scheme theoretical model of metaphors (cf., Lakoff 1987, 1990). However, closer inspection shows that one can see considerable differences when it comes to details. This is understandable too, given that metaphorical extensions of cognates may inherit (at least partially) the differences included in their original meanings. For example, both Polish kręcić się and Russian krutit’sja are used metaphorically, manifesting a shift from rotation domain to the domain of chaotic motion. In this case, they both lose the component ‘axis of rotation’ in the strict sense. However, Russian krutit’sja still maintains its Landmark (as a trace of the former axis) and denotes movement / location near a salient person. Polish kręcić się, on the other hand, may lose the former axis completely, inscribing the chaotic motion in a certain space not related to the domain of Landmark. Therefore, it can be argued that the idea of axis is somewhat stronger for the Russian system of rotation verbs than for the Polish one. The result of the aforementioned is the following – systematic semantic differences between genetically related languages are manifold, independent, and in a sense, inevitable. This is a far-reaching idea for linguistic typology. Usually, it is admitted that typologically-relevant samples must not include genetically related languages for the sake of “unbiased” representation (cf., for example, Bybee 1985 & Bybee et al. 1994; see also Maslova 2000 for a deeper discussion). But data like those discussed in this paper suggest that the division between genetically related and unrelated languages may not be very relevant for lexical typology (cf., also the detailed argumentation in Rakhilina & Plungian 2007 and Divjak et al.,

Verbs of rotation in Russian and Polish 311

this volume).12 If this is the case, our preliminary results from the comparison of Polish and Russian may be safely generalized into further attempts at lexical typological studies. In my opinion, the best way to use these generalizations is to integrate them in a working questionnaire to be applied to broader data. The important point is that the most relevant parameters revealed in our analysis, which specify the domain of non-translational motion,13 do not fit very well into the only widely recognized lexical typology of motion as proposed by Leonard Talmy (see the last version in Talmy 2000; cf., also Slobin 2003). Let us summarize these parameters once more here: – – – – – – – – –

inner / outer axis of roÂ�taÂ�tion incomplete / complete revolution rotation over the Landmark / at the same level with the Landmark rotation with / without moving forward rotation of a living being / an artifact rotation of a point-like / long and flexible object controlled / uncontrolled (spontaneous) rotation high / low speed of rotation natural / “mechanical” (= smooth and steady) rotation

I hope that this list may become a good starting point for a comprehensive lexical typology of rotation (cf., one of the first attempts in this field made in Krugljakova, forthcoming).

References Apresjan, J. D. 1995. Selected Works, Vol. II. Moscow: Jazyki russkoj kul’tury. Apresjan, J. D. 2000. Systematic Lexicography. Oxford: OUP. Bojar, B. 1979. Opis semantyczny czasowników ruchu oraz pojęć związanych z ruchem. Warsaw: University of Warsaw. Bybee, J. L. 1985. Morphology: A Study of the Relation between Meaning and Form. Amsterdam: John Benjamins. Bybee, J. L., Perkins, R. & Pagliuca, W. 1994. The Evolution of Grammar: Tense, Aspect and Mo‑ dality in the Languages of the World. Chicago IL: University of Chicago.

12. Interesting argumentation for a wider use of the data from genetically related languages in other domains of linguistic typology can be found in Kibrik (1998). 13. In fact, this domain includes not only rotation but also many other intricate types of motion. One of them is swinging, which is dealt with in more detail in Rakhilina (2001) and Rakhilina & Prokof’eva (2005).

312 Ekaterina V. Rakhilina

Fillmore, C. J. & Atkins, B. T. S. 1992. Toward a frame-based lexicon: the semantics of RISK and its neighbors. In Frames, Fields, and Contrasts: New Essays in Semantic and Lexical Organi‑ zation, A. Lehrer & E. F. Kittay (eds.), 75–102. Hillsdale NJ: Lawrence Erlbaum. Fillmore, C. J. & Atkins, B. T. S. 2000. Describing polysemy: the case of “crawl”. In Polysemy: Linguistic and Computational Approaches, Y. Ravin & C. Leacock (eds.), 91–110. Oxford: OUP. Kibrik, А. Е. 1998. Does intragenetic typology make sense? In SpraÂ�che im Raum und Zeit: in Memoriam Johannes Bechert , Vol. I, W. Boeder, C. Schroeder, K. Wagner & W. Wildgen (eds.), 61–68. Tübingen: Narr. Koptjevskaja-Tamm, M. 2008. Approaching lexical typology. In From Polysemy to Semantic Change: Towards a Typology of Lexical Semantic AssoÂ�ciÂ�aÂ�tions, M. Vanhove (ed.), 3–54. Amsterdam: John Benjamins. Krugljakova, V. Forthcoming. Sistema glagolov vraščenija v russkom i ispanskom. Cuadernos de Rusística Española. Kustova, G. I. 2004. Tipy proizvodnyx značenij i mexanizmy jazykovogo rasširenija. Moscow: Jazyki russkoj kul’tury. Lakoff, G. 1987. Women, Fire and Dangerous Things: What Categories Reveal about the Mind. Chicago IL: University of Chicago. Lakoff, G. 1990. The invariance hypothesis: Is abstract reason based on image-schemas? Cogni‑ tive Linguistics 1: 39–74. Langacker, R. W. 1987. Foundations of Cognitive Grammar: Theoretical prerequisites. Stanford CA: Stanford University Press. Majsak, T. A. & Rakhilina, E. V. (eds.). 2007. Glagoly dviženija v vode: Leksičeskaja tipologija. Moscow: Indrik. Maslova, E. 2000. A dynamic approach to the verification of distributional universals. Linguistic Typology 4(3): 307–333. Padučeva, E. V. 2004. Dinamičeskie modeli v semantike leksiki. Moscow: Jazyki russkoj kul’tury. Rakhilina, E. V. 2004. There and back: the case of Russian ‘go’. GLOSSOS 5: 1–33. Rakhilina, E. V. 2003. Polzti: put’ k xaosu. In Kosmos i xaos, N. D. Arutjunova (ed.), 415–430. Moscow: Indrik. Rakhilina, E. V. 2001. O prirode beskonečnogo dviženija: “kačat’sja”. Prace filologiczne, XLVI: 493–502. Rakhilina, E. V. 2000. Kognitivnyj analiz predmetnyx imën: semantika i sočetaeÂ�most’. Moscow: Russkie slovari. Rakhilina, E. V. & Plungian, V. A. 2007. O leksiko-semantičeskoj tipoloÂ�gii. In Glagoly dviženija v vode: leksičeskaja tipologija, T. A. Majsak & E. V. Rakhilina (eds.), 9–26. Rakhilina, E. V. & Prokof ’eva, I. A. 2004. Rodstvennye jazyki kak ob”ekt leksičeskoj tipologii: russkie i pol’skie glagoly vraščenija. Voprosy jazykoznanija 1: 60–78. Rakhilina, E. V. & Prokof ’eva, I. A. 2005. Russkie i pol’skie glagoly kolebaÂ�tel’nogo dviženija: semantika i tipologija. In Jazyk. Ličnost’. Tekst., V. N. Toporov (ed.), 304–314. Moscow: Jazyki slaÂ�vjanskoj kul’tury. Slobin, D. I. 2003. Language and thought online: Cognitive consequences of linguistic relativity. In Language in Mind: Advances in the Investigation of Language and Thought, D. Gentner & S. Goldin-Meadow (eds.), 157–191. Cambridge MA: The MIT Press.

Verbs of rotation in Russian and Polish 313

Talmy, L. 2000. Towards a Cognitive Semantics, Vol. II. Cambridge MA: The MIT Press. Traugott, E. C. & Dasher, R. B. 2002. Regularity in Semantic Change. Cambridge: CUP. Wierzbicka, A. 1985. Lexicography and Conceptual Analysis. Ann Arbor MI: KaÂ�roma.

Dictionaries Apresjan, Y. D. (ed.). 1993. NBARS: Novyj bol’šoj anglo-russkij slovar. Moscow: Russkij jazyk. Bańko, M. 2000. ISJP: Inny słownik języka polskiego PWN. Warszawa: PWN. Doroszewski, W. 1958–1969. (ed.). SJPD: Słownik języka polskiego. (10 tt.). Warszawa: PWN. Evgen’eva, A. P. (ed.). 1981. Slovar’ russkogo jazyka v 4 tomax. Moscow: Russkij jazyk. Sobol, E. (ed.). 1996. PSJP: Podręczny słownik języka polskiego. Warszawa: PWN. Szymczak, M. 1995. SJPSz: Słownik języka polskiego PWN. WarÂ�szaÂ�wa: PWN.

Corpora RNC: Russian National Corpus (http://www.ruscorpora.ru) PWN: Corpus of Polish compiled by Polish Scientific Publishers PWN (http://korpus.pwn.pl/ index_en.php) IPI PAN: The IPI PAN Corpus of Polish (http://korpus.pl/index.php)

chapter 13

Aquamotion verbs in Slavic and Germanic A case study in lexical typology Maria Koptjevskaja-Tamm, Dagmar Divjak and Ekaterina V. Rakhilina

Stockholm University, Sweden / University of Sheffield, UK / Institute for the Russian Language, Moscow, Russian Federation

The paper contrasts the verbs plyt’/plavat’ in Russian and płynąć/pływać in Polish with their correspondences in Dutch, English and Swedish against a broader typological background. The three Germanic languages use several verbs for what is covered by a pair of derivationally related verbs in each of the two Slavic languages. The Germanic languages lexicalize the activity/passivity of motion, but vary considerably as to how they carve up the conceptual space. Russian and Polish, on the other hand, use plavat’/plyt’ independently of the activity/passivity of motion and focus on the uni- or non-unidirectionality of the motion. Nonetheless, it appears that the different lexicalizations in the Swedish-EnglishDutch systems of aquamotion verbs are reflected in constructional differences in the Russian-Polish systems.

1.

Introduction

The present paper views the verbs plyt’/plavat’ in Russian and płynąć/pływać in Polish against a broader typological background, stemming from a systematic lexical-typological study of more than 50 genetically, areally and structurally diverse languages (Maisak & Rakhilina 2007b). In order to highlight the main . The international project on aquamotion (2002–2006) supported by the Russian Foundation for the Humanities (RFBR#05-06-80400a) involved a systematic lexical-typological research based on a methodology (questionnaires, checklists and corpora) elaborated, tested and improved by a group of language experts, who collected and analyzed the data in close collaboration with native speakers. The directors of the project were Timur A. Maisak (Institute of Linguistics, Russian Academy of Sciences) and Ekaterina V. Rakhilina (Institute for the Russian Language, Russian Academy of Sciences), who also both carried out research on several languages. Maria Koptjevskaja-Tamm and Dagmar Divjak contributed to the project with the

316 Maria Koptjevskaja-Tamm, Dagmar Divjak and Ekaterina V. Rakhilina

typological properties of these Russian and Polish verbs we will contrast them with the corresponding verbs in three Germanic languages. For this purpose we will look briefly at Swedish simma, segla, driva, flyta etc., at swim, sail, drift and float in English, and at zwemmen, varen, drijven in Dutch (see Divjak & Lemmens 2007; Koptjevskaja-Tamm 2007; Golubkova & Rakhilina 2007 for detailed case-studies on these languages, and Divjak 2006 for a comparison of Dutch and Russian). All these verbs are primarily used for horizontal translational (and sometimes stationary) motion of a non-liquid Figure in or on the surface of a liquid Ground (prototypically in water); this domain of Aquamotion is complementary to Terra-motion (motion on land) and to Aero-Â�motion (motion in the air). The verbs listed are opposed to mersion verbs (such as dive and sink), to self-contained motion verbs (bob) and to verbs expressing motion of the liquid itself (stream, flow). As is clear from the above enumeration, all three Germanic languages use several different verbs for what is covered by a pair of derivationally related verbs in each of the two Slavic languages. Such cross-linguistic variation in how languages categorize or carve up the conceptual aquamotion domain lay at the core of the Aquamotion project (http://aquamotion.narod.ru) and cross-linguistic variation in lexicalization patterns will likewise constitute the main focus of the present paper. This difference is particularly interesting given that all five languages under scrutiny belong to the same type of satellite-framed languages in Talmy’s (2000) classification. This is in fact one of the reasons for why we have chosen to compare Slavic and Germanic instead of comparing Slavic with more typologically diverse languages: we aim to draw attention to the considerable diversity that co-exists with cross-linguistic similarities stemming from genetic and typological relatedness among languages.

descriptions of Swedish and Dutch aquamotion systems. We are also grateful for the INTAS support (05-1000008-7917). . The data on the Slavic and Germanic languages discussed in the present paper come from various sources including several corpora (Russian National Corpus, http://www.ruscorpora. ru/; Korpus Języka Polskiego Wydawnictwa Naukowego PWN, http://korpus.pwn.pl/; the corpora at the Instituut voor Nederlandse Lexicologie in Leiden, www.inl.nl; the Parole corpus at the Swedish Language Bank, http://spraakbanken.gu.se). In working on the present paper we have benefited from the SALT Dutch-Swedish parallel corpus compiled by Gudrun Rawoens (Ghent University) who has kindly provided us with numerous examples. The examples for which the source is not explicitly given are constructed by the authors but have been checked with several native speakers. In some cases we have opted for this strategy to illustrate certain basic facts and simplify the presentation, although we are very much in favor of using authentic examples in our research.

Aquamotion verbs in Slavic and Germanic 317

In this paper, we will show that Dutch, English and Swedish lexicalize the degree of activity/passivity of aquamotion (i.e., primarily Manner of motion) by means of different verbs, i.e., they encode the source of propulsion on the verb, leaving the interpretation of directionality to optional satellites or to contextual inference. However, the three Germanic languages vary considerably as to how exactly they carve up the conceptual space and what precisely they encode on the verb. Russian and Polish, on the other hand, use plavat’/plyt’ independently of the activity/passivity of motion involved, and focus on the uni-/non-directionality of the motion that has grammaticalized. Nonetheless, it turns out that the different lexicalizations in the Swedish-English-Dutch systems of aquamotion verbs are reflected in constructional differences in the Russian-Polish systems. We will start from a general description of the aquamotion system in Swedish, English and Dutch (Section 2); this will introduce the general typological background relevant for the domain (Section 3) that will be used in the analysis of Russian and Polish (Section 4). In Sections 5 and 6 we will look more specifically into the differences among the closely related languages. Section 6 will sum up the main findings illustrated throughout the paper.

2.

A Germanic triad: Swedish, English and Dutch

Swedish has an elaborated lexical system for talking about the horizontal motion of a non-liquid Figure in or on the surface of a liquid Ground (typically water) that includes six designated (or proper) aquamotion verbs (simma, segla, driva, flyta, ro, paddla) and a number of general motion verbs. Simma denotes active self-propelled aquamotion with an animate Figure accomplished by the Figure’s (controlled) movements of body and/or limbs. The typical Figures here are human beings, but many animals likewise swim (mammals, fish, water birds and snakes). Simma cannot be used whenever aquamotion involves a vessel and/or whenever the Figure’s own contribution to motion is minimal or non-existent (passive aquamotion); a limited contribution might be caused either by the fact that the Figure’s bodily movements are not (fully) responsible for motion (for an animate Figure), or because the Figure is inanimate. Let’s consider these different cases in turn. For aquamotion involving vessels the situation is quite complex. The most prominent verb here is segla (the cognate of sail), which is obligatorily used for motion of sailing boats and/or of people aboard them. Consider ex. (1a) in which the word båt ‘boat, ship’ is used as the subject to seglade ‘sailed’. Although båt can denote any (or most) kinds of vessel, in this example it unambiguously refers to a sailing boat because of the predicate – neither rowing boats, nor canoes or

318 Maria Koptjevskaja-Tamm, Dagmar Divjak and Ekaterina V. Rakhilina

motor-driven boats and ships “sail” in Swedish. Segla can be said both about moving sailing boats and about the people aboard, i.e., segla combines with different kinds of Figure/subject (an inanimate Figure referring to a vessel vs. a human Figure). In addition, for human Figures there are different constructions available depending on whether the human being is actively navigating (with segla used as a transitive verb and the vessel encoded as its direct object, ex. (1b)) or is a passenger (with segla used intransitively and the vessel encoded as its comitative adjunct, ex. (1c)). (1) a. b. c.

Båt-en segla-de från G. till M. på e-n dag. boat-def.c.sg sail-pret from G. to M. on one-c day ‘The sailing boat sailed from G. to M. in one day.’ Jan segla-de vår båt från G. till M. på e-n dag. Jan sail-pret our.c.sg boat from G. to M. on one-c day ‘Jan sailed our sailing boat from G. to M. in one day.’ Jan segla-de med vår båt från G. till M. på en dag. Jan sail-pret with.our.c.sg boat from G. to M. on one-c day ‘Jan sailed with / on our sailing boat from G. to M. in one day.’

Different from sailing boats, neither rowing boats nor canoes can be construed in Swedish as actively moving Figures. Their motion can only be described indirectly, via reference to the corresponding activities (rowing – ro and paddling – paddla) of the human Agents aboard (2a). Strangely enough, motor-driven boats and ships do not have any designated aquamotion verbs at all; reference to their motion and to the motion of people aboard has to recruit generalized verbs of motion (åka/fara ‘to go by a vehicle’, gå ‘to go on foot’, komma ‘to come’) or even verbs that do not express motion at all (e.g., ta ‘to take’, etc.) (2b)). (2) a. b.

Vi/*Båt-en ro-dde / paddla-de över vik-en. we/*boat-def.c.sg row-pret / paddle-pret over bay-def.c.sg ‘We / *The boat rowed / paddled over the bay.’ Titta, vilk-en stor ångare som komm-er hit! look which-c.sg big.c.sg steamship rel come-pres here ‘Look what a big steamship is coming!’

Finally, passive aquamotion is covered by two verbs – driva and flyta. In both cases the Figure’s motion (or the absence of any significant motion) is to a large extent determined by the dynamics of its liquid Ground (prototypically water), sometimes in combination with another external force (typically the wind). Driva . Glossing conventions follow the Leipzig Glossing Rules available at http://www.eva.mpg. de/lingua/resources/glossing-rules.php.

Aquamotion verbs in Slavic and Germanic 319

refers to passive motion that is completely out of control and often (but not necessarily) takes on an undesirable course and/or is chaotic in an unpredictable and disturbing way. The latter connotation is primarily present when the Figure is a vessel or a human being, for which controlled aquamotion is the default (or at least a desideratum). Flyta lacks negative connotations and is more neutral as to whether the motion is uncontrolled or controlled. What matters is that the Figure is situated in water or another liquid (typically on its surface) and moves together with it without sinking. Note that if motion is determined by the dynamics of the water, it is not necessarily translational, i.e., the Figure does not necessarily change its location. In the absence of a current, when the water itself is not moving, objects on its surface will remain more or less located in one and the same place – such situations of location on the surface of water are also covered by flyta, cf. ex. (3). For inanimate Figures, except for vessels, passive aquamotion (including location on the surface of water) is the only possible kind of aquamotion; for vessels and animate Figures it competes with others, more active kinds of aquamotion, cf. ex. (4)–(6). (3)

Location in/on the surface of water/liquid Död-a fisk-ar flyt-er vid strand-en. dead-pl fish-pl float-pres by shore-def.c.sg ‘Dead fish is floating by the shore.’

(4)

Aquamotion of a human Figure Jan simma-r (a.) / seglar (b.) / flyt-er (c.) / driv-er (d.) Jan swim-pres (a.) / sail-pres (b.) / float-pres (c.) / drift-pres (d.) mot klipp-or-na. towards rock-pl-def.pl a. ‘Jan is swimming towards the rocks.’ b. ‘Jan is sailing (on) a sailing boat towards the rocks.’ c. ‘Jan is floating towards the rocks.’ d. ‘Jan is drifting towards the rocks.’

(5)

Aquamotion of sailing boats as a Figure Båten *simma-r (a.) / seglar (b.) / flyt-er (c.) boat-def.c.sg swim-pres (a.) / sail-pres (b.) / float-pres (c.) / driv-er (d.) mot klipporna. / drift-pres (d.) towards rock-pl-def.pl a. *‘The sailing boat is swimming towards the rocks.’ b. ‘The sailing boat is sailing towards the rocks.’ c. ‘The sailing boat is floating towards the rocks.’ d. ‘The sailing boat is drifting towards the rocks.’

segla

Sailing boats

(no specific aquamotion verbs)

Motor-Â�driven vessels

sail

zeilen

zwemmen

Dutch

roeien (indirectly) varen

row (indirectly)

Rowing boats

Canoes

flyta

drift drijven

float

Neutral motion / Location

Passive motion Motion out of control

driva

Neutral motion / Location

Passive motion Motion out of control

paddelen (indirectly)

paddle (indirectly)

Canoes

paddla (indirectly)

Motion of vessels and people aboard Motor-driven vessels

swim

Sailing boats

English

Active motion of an animate Figure

ro (indirectly)

Rowing boats

Motion of vessels and people aboard

Table 2.â•‡ The English and Dutch systems of aquamotion verbs

simma

Active motion of an animate Figure

Table 1.â•‡ The Swedish system of aquamotion verbs

320 Maria Koptjevskaja-Tamm, Dagmar Divjak and Ekaterina V. Rakhilina

Aquamotion verbs in Slavic and Germanic 321

(6)

Aquamotion of other inanimate Figures Is-flak-en driv-er / flyt-er mot klipp-or-na. ice-floe-pl.def drift-pres / float-pres towards rock-pl-def.pl ‘The ice-floes are floating / drifting towards the rocks.’

The Swedish facts considered above are summarized in Table 1. The aquamotion verb systems in the two related Germanic languages English and Dutch show distinctions akin to those in Swedish, but with somewhat reduced degrees of elaboration. Thus, the distinctions among swim, drift and float in English (Golubkova & Rakhilina 2007) are fairly similar to those among their cognates simma, driva and flyta in Swedish. English, however, has generalized the use of sail to cover most kinds of vessels (both sailing and motor-driven boats and ships), whereas the motion of rowing boats and canoes is treated indirectly, in the same way as in Swedish (row, paddle). The Dutch system (Divjak & Lemmens 2007) is reduced compared to both Swedish and English in two respects. First, it has only one verb for passive aquamotion (including location in water), drijven (cognate to drift and driva). And, second, it has one generalized aquamotion verb, varen, which covers all motion of vessels and people on vessels, even though there are also other optional verbs with more specific applications (zeilen, roeien, paddelen). The English and Dutch facts are summarized in Table 2. The differences among Swedish, English and Dutch notwithstanding, all three languages clearly differentiate among several unrelated verbs, each responsible for its own kind of aquamotion. The main distinctions are between active self-propelled motion of an animate Figure, active motion of vessels and people aboard (with further specifications of vessels) and passive motion of both animate and inanimate Figures (with location on the surface of water as the endpoint of aquamotion). Now, in what way are these Germanic facts relevant for a volume on motion verbs in Slavic? In order to see how, we first need to consider a typology of aquamotion verbs.

3.

From Germanic to typology: Sub-domains of aquamotion and aquamotion language systems

The Germanic facts considered in the previous section are highly relevant in a more general typological perspective as they find parallels in many languages. In fact, as shown in Maisak & Rakhilina (2007a), languages tend to carve up the aquamotion domain according to the distinctions illustrated by the three

322 Maria Koptjevskaja-Tamm, Dagmar Divjak and Ekaterina V. Rakhilina

Germanic languages, e.g., by using different lexemes and/or constructions for reference to different “kinds” of aquamotion. The main parameter underlying semantic categorization within the aquamotion domain across languages is the degree of activity/passivity of motion, which can be viewed as a scale with three main zones, or sub-domains, each centred on a particular prototype. 1. One sub-domain corresponds to active, self-propelled motion of animate Figures accomplished by the Figure’s controlled movements of body and/or limbs in water (swimming). 2. Its opposite is passive uncontrolled and non-agentive aquamotion, whereby the Figure simply follows the movement of the water in which or on the surface of which it happens to be. Since passive Figures lacking control and agentivity are prototypically inanimate, passive aquamotion commonly involves inanimate Figures. And, since motion in these cases depends on the dynamics of the water, passive aquamotion can occasionally be stationary, whereby the Figure remains more or less located in one and the same place. Passive aquamotion is therefore particularly sensitive to the presence/absence of direction, with the concomitant distinction between the maximally passive location in or on the surface of water (floating) and passive, but directed motion in water (drifting). 3. Finally, there is motion of vessels and/or people aboard (sailing), which combines properties typical of both active and passive aquamotion and can therefore be viewed as in-between the two. A sailing situation can be construed as motion of different kinds of Figure – the navigator, the vessel and passengers on board. Sailing involves, of course, agentivity of an animate (and prototypically human) entity who navigates the vessel, but whose own motion is completely dependent on the latter’s motion. The vessel itself, although moving, lacks agentivity and control over the motion, whereas “passengers” on board are typically passive, when it comes to their contribution to the motion. Figure 1 bellow summarizes possible participation of different kinds of Figures in different kinds of aquamotion and the relative degree of activity of motion typical for each of the aquamotion subdomains. As Maisak & Rakhilina (2007a) demonstrated, although languages on the whole tend to distinguish among these sub-domains by means of lexical oppositions and/or constraints on interpretation in the corresponding contexts, there is a great deal of cross-linguistic variation. One major difference is the sheer number . The four main English aquamotion verbs will be used as meta-linguistic labels for the aquamotion sub-domains, even though the distinctions among the verbs and those among the subdomains do not completely coincide.

Kind of Figure

Aquamotion verbs in Slavic and Germanic 323

Animates

Vessels

Inanimates

Motion of vessels and/or people aboard (sailing)

Passive motion (directed motion, drifting, and location in water, floating)

primarily humans Kind of aquamotion

Active self-propelled aquamotion (swimming)

Activity of motion more active

more passive

Figure 1.â•‡ The activity-of-motion scale underlying the distinctions among the main aquamotion sub-domains and possible kinds of Figure participating in each of them

of aquamotion verbs in a language, or the degree of elaboration within its lexical aquamotion system. There are, roughly speaking, three groups of systems. – Swedish, English and possibly Dutch belong to languages with rich aquamo‑ tion systems, i.e., they have more than three verbs. Indonesian, with its fourteen aquamotion verbs (Lander & Kramarova 2007), provides an example of a particularly richly elaborated aquamotion system. – Languages with middle systems have two main aquamotion verbs normally distinguishing between active and passive motion, and use a general motion verb for motion on vessels. They are, on the whole, fairly infrequent – Tamil, Persian and Maninka are the only languages with middle aquamotion systems par excellence among the 50 languages in the Aquamotion project. – Finally, quite a few languages have poor systems that seem to neutralize the distinctions among the aquamotion sub-domains or at least make these distinctions peripheral. Turkish, for instance, has only one aqua-motion verb, yüzmek, which alone covers swimming, sailing and drifting/floating. Poor lexical aquamotion systems are also found in the Slavic languages Russian and Polish, to which we turn in the next section. An interesting question is how the distinctions among the aquamotion subdomains relate to the components in a Motion event and to the different lexicalization patterns as suggested by Talmy (2000). Thus, for instance, in the face of such contrasts as Peter was swimming/floating/sailing, where one and the same human Figure is participating in different kinds of aquamotion, it might be reasonable . Note that even languages with comparable systems may show significant variation in details, e.g., with respect to the precise ways in which the different aquamotion verbs divide the semantic space. For the sake of space we will not elaborate on this point in the paper.

324 Maria Koptjevskaja-Tamm, Dagmar Divjak and Ekaterina V. Rakhilina

to suggest that the three English verbs swim, float and sail have lexicalized the Manner of aquamotion. On the other hand, all the three conflate Manner with Ground; in addition, swim conflates both Ground and Manner with Figure (since only animate Figure can swim).

4.

From typology back to Russian and Polish

For reference to aquamotion, both Russian and Polish have only two derivationally related imperfective verbs, plyt’/plavat’ in Russian and płynąć/pływać in Polish, where the basic distinction between the two verbs has to do with unidirectionality of motion. The first verb in each pair is unidirectional (7a), whereas the other one is non-(uni)directional (7b–c). (Although the presentation and examples in this section will be based on Russian, it applies also to Polish, mutatis mutandis.) (7) Russian a. Petja ply-l k skal-am. Petja “plyt’”-pret.m.sg towards rock-dat.pl ‘Petja was swimming to the rocks.’ b. Petja plava-l k skal-am (každyj den’). Petja “plavat’”-pret.m.sg towards rock-dat.pl (every day) ‘Petja swam to the rocks (every day).’ c. Petja plava-l u skal. Petja “plavat’”-pret.m.sg by rock.gen.pl ‘Petja was swimming by the rocks.’

Aquamotion verbs in Russian and Polish, like motion verbs in general, have thus lexicalized a semantic distinction that appears to be absent from the corresponding lexical field in Swedish, English, Dutch, as well as in many other languages. At the same time, the parameter of activity/passivity of motion, underlying the distinctions among the aquamotion sub-domains within the Swedish, English and Dutch systems and, further, in many other languages of the world, does not appear to be relevant for Russian and Polish. For instance, the Russian examples in (8)–(10) show ambiguities that are absent from their Swedish, English and Dutch translations. (8)

Human Figure: Active self-propelled motion vs. motion aboard Ply-vi k skalam! “plyt’”-imp towards rock-dat.pl ‘Swim/sail/row/paddle to the rocks!’

Aquamotion verbs in Slavic and Germanic 325

(9)

Vessel: Motion of vessels vs. passive motion Lodk-a plyl-a vniz po rek-e. boat-nom.sg “plyt’”-pret.f.sg down on river-dat.sg ‘The boat was sailing/drifting/floating down the river’ or ‘Someone was rowing the boat down the river.’

(10) Animate non-human Figure: Active self-propelled motion vs. passive motion (i.e., location in/on the surface of water) V bassejn-e plava-l krokodil. in â•›swimming.pool-loc.sg “plavat’”-pret.m.sg crocodile.nom.sg ‘A crocodile was swimming in the swimming pool / There was a crocodile in the swimming pool.’

Let us examine to what extent the Swedish-Dutch-English and the Russian-Polish systems are different. As we hope to show, the systems in the two groups of languages are sensitive to similar cross-linguistically relevant semantic distinctions within the aquamotion domain, even though their manifestations can be striking. First of all, the difference between plyt’ and plavat’ in reference to passive aquamotion easily translates into the distinction between drifting and floating. In the prototypical cases drifting constitutes a unidirectional motion that involves a current or a stream, flowing in a certain direction. Therefore it does not come as a surprise that drifting is expressed by the unidirectional verb plyt’, while the non-directional plavat’ refers to non-directional (or even static) floating. Thus, the normal interpretation of ex. (11a) requires that both the boat and the black logs are moving in a certain direction, while the black logs in ex. (11b) are either static or are moving chaotically, without any specific direction. (11) a. b.

Rjadom s naš-ej lodk-oj ply-l-i alongside with our-instr.f.sg boat-instr.sg “plyt’”-pret-pl čern-ye tjažel-ye brevn-a. black-nom.pl heavy-nom.pl log-nom.pl ‘Heavy black logs floated downstream alongside our boat.’ (= were moving) Rjadom s naš-ej lodk-oj plava-l-i alongside with our-instr.f.sg boat-instr.sg “plavat’”-pret-pl čern-ye tjažel-ye brevn-a. black-nom.pl heavy-nom.pl log-nom.pl ‘Heavy black logs floated around/here and there near our boat.’ (= were located)

326 Maria Koptjevskaja-Tamm, Dagmar Divjak and Ekaterina V. Rakhilina

Second, plyt’ and plavat’ show some asymmetries in their propensity to combine with expressions referring to the ground, source or goal of motion that can be related to the semantic distinctions among the three main aquamotional subdomains. To unveil these asymmetries we have compared the relative frequencies of plyt’ in the Russian National Corpus (RNC) used in different constructions and referring to different kinds of aquamotion (cf. Tables 3 and 4). The relevant constructions (or contexts) are as follows: – plyt’ not followed by ground, source or goal – plyt’ followed by ground (po ‘on’ + NP in the Dative case) – plyt’ followed by goal (k ‘towards’ + NP in the Dative case, v ‘into’ + NP in the Accusative case, or do ‘to’ + NP in the Genitive case) – plyt’ followed by source (iz ‘from’ + NP in the Genitive case, or ot ‘from’ + NP in the Genitive case) – plyt’ followed by source and goal (iz ‘from’ + NP in the Genitive case and v ‘into’ + NP in the Accusative case, or ot ‘from’ + NP in the Genitive case and do ‘to’ + NP in the Genitive case) The occurrences of plyt’ have been categorized into the by now familiar cateÂ� gories: – motion of vessels and people aboard, – aquamotion of other inanimate (including dead) Figures (i.e., passive motion), and – active aquamotion, where we have distinguished between two kinds of animate Figures – human Figures and animate non-human Figures (both not aboard), for reasons which will become evident below. We have restricted ourselves to a one-sentence context, which has sometimes not been sufficient for understanding the details of the motion (i.e., who/what is moving and in what way). However, the absolute majority of the examples could be categorized unambiguously. Note also that we have excluded the numerous metaphorical and extended uses of plyt’. We have applied different strategies for selecting and counting the occurrences of plyt’ without reference to the ground, source and goal, on the one hand, and in all the other contexts. In the first case, we have randomly selected 1000 examples among the 6362 occurrences of plyt’ in the Russian National Corpus (RNC). For the other contexts we have analyzed all the occurrences in the corpus. That means that the comparison of plyt’ across the different contexts should be based on relative rather than absolute frequencies.

Aquamotion verbs in Slavic and Germanic 327

Plyt’ not followed by ground, source or goal. As is clear from Table 3, in this context (which is the by far most frequent one) plyt’ tends to refer more or less equally often to motion of vessels/people aboard and to active aquamotion, while its passive-aquamotion readings have a significantly lower frequency. However, explicit reference to the ground, source or goal of aquation may influence the distribution of the preferred vs. dispreferred readings of plyt’. These constructions are on the whole relatively infrequent, with ground being mentioned much more often than both source and goal, and source being mentioned in very few cases (for the figures cf. Table 4). Plyt’ accompanied by ground (po ‘on’ + NP in the Dative case). In this context, motion of vessels and people aboard constitutes the most frequent reading of plyt’ (cf. ex. (12a)). In addition, these contexts also favour passive-motion readings of plyt’, as in ex. (12b), even though reference to passive motion is in general fairly infrequent. On the other hand, this context seems to disfavour active selfpropelled motion of human Figures, i.e., prototypical active aquamotion. In other words, when referring to active self-propelled motion of animate entities, plyt’ prefers not to combine with an overt indication of the ground (with (12c) being a rare example of the opposite). Perhaps somewhat unexpectedly, active motion of other animates – fish, birds, snakes, dogs, horses, fairy figures etc. – freely allows specification of the ground (ex. (12d)). (12) Plyt’ accompanied by ground (po ‘on’ + NP in the Dative case). a. Po sin-emu morj-u ply-l znakom-yj on blue-n.dat.sg sea-dat.sg “plyt’”-pret.m.sg familiar-m.nom.sg bel-yj korabl’ s dvumja vysok-imi white-m.nom.sg ship.nom.sg with two-instr tall-instr.pl naklonn-ymi truba-mi. inclined-instr.pl funnel-instr.pl ‘A familiar white ship with two tall inclined funnels was sailing on the blue sea.’ (G. Alekseev. Zelenye berega (1983–1984), quoted after RNC) b. Ply-l-i po vod-e stog-i sen-a, “plyt’”-pret-pl on water-dat.sg stack-nom.pl hay-gen.sg brevn-a, plot-y, oblomk-i izb i, log-nom.pl raft-nom.pl fragment-nom.pl hut.gen.pl and (…) dostignuv plotiny, stalkivalis’ drug s drugom, nyrjali, opjat’ vyplyvali i sbivalis’ v kuču v odnom meste. ‘Haystacks, logs, rafts and fragments of huts were floating/drifting downstream on the water and, having reached the dam, dove into the water, emerged again and bunched together in one and the same place.’ (M. E. Saltykov-Ščedrin. Istorija odnogo goroda (1869–1870), quoted after RNC)

13.6% (62)

23.3% (106)

33.4% (152)

Motion of vessels and people aboard 6.6% (30)

23.1% (105)

Passive aquamotion of Unclear contexts an inanimate (incl. dead) Figure (not a vessel) 455

Total

0% 12.5% (1)

Plyt’ + iz GEN Plyt’ + ot GEN

Expressions of source 0% 0%

0% 0%

16.2% (16) 0% 0%

11.2% (41)

* Including one occurrence in the context plyt’ + po DAT + v ACC (i.e., plyt’ + ground + goal)

0% 0%

47.5% (47) 0% 0%

Plyt’ + k DAT Plyt’ + v ACC Plyt’ + do GEN

Expressions of goal

Expressions of Plyt’ + iz GEN v ACC source and goal Plyt’ + ot GEN do ACC

1.6% (6)

Plyt’+ po DAT

Expressions of ground

Aquamotion of a human Figure

100% (7) 100% (2)

87.5% (14) 87.5% (7)

25.3% (25) 94.7% (36) 92.3% (12)

47.5% (174)

0% 0%

12.5% (2) 0%

7.1% (7) 5.2% (2)* 0%

13.7% (50)

0% 0%

0% 0%

4% (4) 0% 7.7% (1)

26% (95)

Motion of vessels Passive aquamo- Unclear and people aboard tion of an inanim./ context Aquamotion of an dead Figure anim. non-human (not a vessel) Figure

Active aquamotion

7 2

16 8

99 38 13

366

Total

Table 4.â•‡ Plyt’ referring to the different sub-domains within aquamotion when followed by expressions of ground, source or goal in the Russian National Corpus

36.9% (168)

Aquamotion of an animate non-Â�human Figure

Aquamotion of a human Figure

Active aquamotion

Table 3.â•‡ Plyt’ referring to the different sub-domains within aquamotion when not followed by expressions of ground, source or goal in the Russian National Corpus (among the 1000 randomly chosen occurrences of plyt’)

328 Maria Koptjevskaja-Tamm, Dagmar Divjak and Ekaterina V. Rakhilina

Aquamotion verbs in Slavic and Germanic 329

c. d.

Mne snilos’, čto ja ply-vu I:dat dream:pret.n.sg that I:nom “plyt’”-pres.1sg po bezbrežn-omu i spokojn-omu morj-u, on boundless-n.dat.sg and calm-n.dat.sg sea-dat.sg (…) tolkaja vperedi sebja jarko-krasočnyj mjač. ‘I dreamed that I was swimming/floating on a boundless and calm sea, pushing a bright-coloured ball in front of me.’ (R. Naxapetov. Vljublennyj (1998), quoted after RNC) Čern-yj zmej ply-l po vod-e, (…) black-m.nom.sg. serpent-nom.sg “plyt’”-pret.m.sg on water-dat.sg (…) a sled za soboj ostavljal krasnyj, počti krovavyj, i Miše stalo strašno. ‘The black serpent was swimming [forward] in the water, leaving a red, almost bloody trace after itself, and Misha got scared.’ (A. Jašin. Sladkij ostrov (1960), quoted after RNC)

Plyt’ accompanied by goal (k ‘towards’ + NP in the Dative case, v ‘into’ + NP in the Accusative case, or do ‘to’ + NP in the Genitive case). Explicit reference to the goal of aquamotion is, on the whole, much less frequent than reference to the ground and can be expressed in different ways. In our corpus data, the most frequent goal expressions consist of “k ‘towards’ + NP in the Dative case”. These are compatible with the different readings of plyt’ (cf. (13c) for passive motion), but are particularly frequent with reference to active self-propelled motion of animate entities (ex. (13a–b). The other two goal expressions – “v ‘into’ + NP in the Accusative case”, or “do ‘to, as far as’ + NP in the Genitive case” – are almost exclusively restricted to motion of vessels and people aboard (ex. (13e)). Interestingly, one of the two examples with plyt’ followed by a “v+ACC”-phrase and referring to passive motion (ex. (13d)) contains also a reference to the ground (a “po+DAT”-phrase) which, as shown above, constitutes a favourable context for such readings. (13) Plyt’ accompanied by goal a. Marija ply-l-a k bereg-u, (…) M.:nom.sg “plyt’”-pret-f.sg to shore-dat.sg bojazlivo pogljadyvaja vverx. ‘Maria was swimming towards the shore, casting timid glances upwards.’ (A. Dmitriev. Doroga obratno//“Znamja”, 2001, quoted after RNC) b. S živ-oj pojmann-oj ryb-oj with living-f.instr.sg captured-f.instr.sg fish-f.instr.sg vo rt-u vydra ply-l-a k gust-omu in mouth-loc.sg otter.nom.sg “plyt’”-pret-f.sg to thick-m.dat.sg

330 Maria Koptjevskaja-Tamm, Dagmar Divjak and Ekaterina V. Rakhilina

c. d. e.

kustarnik-u, kotorym byl pokryt bereg. shrubs-dat.sg (…) ‘Holding a living captured fish in its mouth, the otter was swimming towards the thick shrubs covering the shore.’ (I. Sokolov-Mikitin. Vydry (1923–1928), quoted after RNC) Želt-ye list’ja, plyv-ušč-ie yellow-nom.pl leaf.nom.pl “plyt’”-part.pres.-nom.pl k vodostok-u to gutter-dat.sg ‘Yellow leaves floating downstream towards the gutter’. (E. Šklovskij. Sutra pjatogo patriarxa (1990–1996), quoted after RNC) Tak poka on tam naladčikom byl, ply-l-i brevn-a (…) “plyt’”-pret-f.sg log-nom.pl po Isterv-e v Vytekl-u, on Isterva-dat.sg into Vytekla-acc.sg (…) ni razu ni za čto sučkom ne zadevši. ‘So while he worked as an adjuster there, logs floated downstream the Isterva into the Vytekla without any bough ever touching anything.’ (E. Lukin. Katali my vaše solnce (1997), quoted after RNC) My seli na kater, plyv-em do Novorossijsk-a (…) “plyt’”-pres.1pl as.far.as Novorossijsk-gen.sg (…) i popadaem v sil’nejšij štorm. ‘We took a launch, are sailing as far as Novorossijsk and get caught in a very heavy storm.’ (È. Gerštejn. Perečen’ obid (1997), quoted after RNC)

Plyt’ accompanied by source (iz ‘from’ + NP in the Genitive case, or ot ‘from’ + NP in the Genitive case), and plyt’ accompanied by source and goal (iz ‘from’ + NP in the Genitive case and v ‘into’ + NP in the Accusative case, or ot ‘from’ + NP in the Genitive case and do ‘to’ + NP in the Genitive case). Although these contexts are, on the whole, fairly infrequent, in our data they are almost exclusively restricted to motion of vessels and people aboard. This is particularly true when both source and goal are mentioned, as in ex. (14a). Example (14b) illustrates one of the rare cases in which plyt’ designates passive aquamotion when followed by an explicit reference to the source of motion. (14) Plyt’ accompanied by source or by source and goal a. Jaxta s kapitan-om Tolst-ym i šest’j-u yacht.nom.sg with captain-instr.sg T.-instr.sg and six-instr preobraženc-ami s potuše-nn-ymi preobrazhenec-instr.pl with put.out-pass.part-instr.pl ognj-ami ply-l-a iz Peterburg-a light-instr.pl “plyt’”-pret-f.sg from Petersburg-gen.sg

Aquamotion verbs in Slavic and Germanic 331

b.

v Kronštadt. in Kronstadt.acc.sg ‘A sailing boat with captain Tolstoy and six soldiers of the Preobrazhenski Regiment was sailing from Petersburg towards Kronstadt with extinguished lights.’ (È. Radzinskij. Knjažna Tarakanova (1999), quoted after RNC) Reka sovsem očistilas’ oto l’da, i tol’ko izredka na nej pokazyvalis’ belymi pjatnami zapozdavšie l’diny; verojatno, oni ply-l-i (…), they.nom “plyt’”-pret-pl iz kakogo-nibud’ bojk-ogo gorn-ogo from some.m.gen.sg vivid-m.gen.sg mountain.adj-m.gen.sg pritok-a. tributary-gen.sg ‘The river had become completely clear of ice, only some late ice-floes appeared as rare white spots; they were probably floating downward from some vivid mountain tributary.’ (D. N. Mamin-Sibirjak. Na reke Čusovoj (1912), quoted after RNC)

The constructional differences related to the distinctions in the interpretation of plyt’ can also be found, mutatis mutandi, in the behaviour of plavat’, even though the situation there is slightly more complicated. Summarizing this section, we can conclude that the cross-linguistically recurrent distinctions among the three basic aquamotion sub-domains are in fact present in Russian (and Polish) as well, yet in other ways than in languages with lexically rich aquamotion systems in general and in Swedish, English or Dutch in particular. The three Germanic languages have several individual and derivationally unrelated lexemes, each specializing in a particular sub-domain. The main parameter underlying these distinctions is the degree of activity/passivity of motion, which, to a certain degree, can be interpreted as the differences in the manner of motion of a non-liquid Figure on a liquid ground (dependent, in turn, on the nature of the Figure). Russian and Polish, on the other hand, have two derivationally related lexemes that in the first place lexicalize unidirectionality/non-directionality of motion of a non-liquid figure on a liquid ground, rather than the degree of its activity/passivity (or Manner). The Swedish-English-Dutch and the Russian-Polish systems are, however, not as distinct as they appear. First, directionality (i.e., presence/absence of aquamotion) is relevant also for Swedish and English, even though this distinction is found in only one sub-domain of aquamotion (drifting/floating within passive aquamotion) and is expressed by two unrelated lexemes. Conversely, the semantic distinctions among the main aquamotion sub-domains underlie some interesting syntactic asymmetries in the behaviour of plyt’ and plavat’, i.e., their combinability with expressions

332 Maria Koptjevskaja-Tamm, Dagmar Divjak and Ekaterina V. Rakhilina

referring to the ground, source or goal. In other words, different lexicalizations in the Swedish-English-Dutch systems of aquamotion verbs are reflected in constructional differences in the Russian-Polish systems.

5.

Genetic vs. typological similarities: Aquamotion of a non-liquid vs. liquid Figure

Since the main impetus for cross-linguistic and, in particular, typological research is finding out what aspects of a particular phenomenon are more specific or more universal, one of its tasks consists in distinguishing between cross-linguistic similarities that stem from genetic relatedness, on the one hand, and commonly attested phenomena that do not stem from genetic relatedness. These commonalities, in turn, call for an explanation. The different ways of lexicalizing the degree of passivity/activity of aquamotion of a non-liquid Figure, demonstrated by Swedish, English, Dutch, as opposed to Russian and Polish, seem to have a strong genetic component; among other things, they involve cognates in Germanic and the inherited derivational opposition in Slavic. We will now turn to a phenomenon where similarities among languages crosscut these two families, namely, aquamotion verbs that can apply to both non-liquid and liquid Figures. These are found in two of the languages considered in the present paper – in Polish and Swedish; the description of Polish and most of the relevant examples are quoted from Prokof ’eva (2007). In Polish, the verb płynąć (i.e., the unidirectional member of the pair płynąć/ pływać considered in Section 4) can, in addition, describe the motion of a river, i.e., of a liquid Figure, as in ex. (15) (15)

Ob jest rzeką nizinną, płynie doliną o Ob is river lowland płynąć.pres.3sg valley.instr about szerokości do 60 km i uchodzi do Morza Karskiego. width.instr up.to 60 km and go:pres.3sg to Sea.gen Kare.gen ‘Ob is a (lowland) river that flows through a valley that is up to 60 km wide and goes into the Kara Sea.’ (Zat. Obska, http://transsib.com.pl/rzeki.html, visited on November 11, 2008)

Płynąć competes with several other verbs for motion of liquids, each of which has its own specific meaning and range of uses: sąszyć się ‘ooze out, trickle’ (about a little stream of liquid), ciec ‘run, flow’ (about bodily liquids and thick liquids such as honey or lava that move on a surface without normally following a particular path), and lać się ‘flow, pour, stream’ (about a considerable amount of

Aquamotion verbs in Slavic and Germanic 333

liquid moving in a stream without any contact with a surface, e.g., water pouring from a tap or from a ceiling). The core uses of płynąć among these verbs have to do with motion of (water in) rivers. Crucially, rivers flow in riverbeds, their movement (or rather, the movement of the water in them) is, thus, severely restricted in its path and is ideally unidirectional, which agrees very well with the unidirectional meanings of płynąć when applied to motion of non-liquid Figures. On the contrary, the water in a sea or an ocean cannot be described by płynąć. The two other main contexts compatible with płynąć are liquids running in pipes (e.g., oil or water, ex. (16a)) and flow in blood vessels (ex. (16b)), which both show obvious similarities with water flowing in rivers. (16) a. b.

Woda ogrzana w kotle płynie water.nom.sg heated.up.f.sg.nom in tank.loc “płynąć”.pres.3sg rurami wgórę do kaloryfer-ów, pipes.instr.pl up to radiator-gen.pl (...) a woda zimna innymi rurami spływa w dół do kotła. ‘Water heated up in the tank flows up trough pipes to the radiators, while cold water flows down through other pipes to the tank.’ (“Horyzonty Techniki”. 20.1959, quoted after KJPWN) W rozgrzan-ych i elastyczn-ych żyłach in warmed.up-loc.pl and elastic-loc.pl veins.loc.pl płynie bez przeszkód krew. “płynąć”.pres.3sg without obstacle:gen.pl blood:nom.sg. ‘In warmed up and elastic veins blood can flow without obstacle.’ (J. Wojda. Za oceanu, quoted after KJPWN)

Swedish too uses one and the same verb, flyta, both with non-liquid and liquid Figures. As shown in Section 2, for a non-liquid Figure flyta describes its passive aquamotion including location in/on the surface of water. In addition, it can apply to (the water in) rivers, as in ex. (17): (17) Swedish a. Ungern-s huvudstad Budapest bruka-r kalla-s för Hungary-gen capital Budapest use-pres call.inf-refl for “Donau-s pärla” eftersom flod-en flyt-er mitt Donau-gen pearl because river-def.c.sg “flyta”-pres amidst genom stad-en och dela-r de-n i Buda och Pest. across city-def.c.sg and divide-pres it-c in Buda and Pest

334 Maria Koptjevskaja-Tamm, Dagmar Divjak and Ekaterina V. Rakhilina

b.

‘The capital of Hungary, Budapest, is often called “the pearl of the Donau” because the river runs across the middle of the city and divides it into Buda and Pest.’ (www.lfv.se/upload/Flygplatser/Umea/Program.doc, visited on November 11 2008) I älv-en flöt ren-t vatten. in river-def.c.sg “flyta”.pret clear-n.sg water ‘Clear water was running in the river.’ (Parole)

Swedish has several verbs for motion of liquid, primarily water (ströma ‘to stream, to flow, to pour’, forsa ‘to rush, to come in torrents’, droppa ‘to drop’, etc.), but the main “rival” and quasi-synonym of flyta is rinna. Rinna normally denotes relatively rapid unidirectional motion (often downwards) of a relatively small amount of water that is conceived of as a relatively narrow stream and is therefore often applied to brooks, small currents, and the like. Flyta, on the contrary, presupposes a broad flow, a large amount of water flowing slowly, smoothly and is therefore often used for depiction of big and broad rivers. However, rinna is preferred in geographic contexts for generic descriptions of the location of rivers, of where they have their sources, where they flow out etc. This can be due to the differences in the perspectivization typical of rinna and flyta: on a map rivers look like thin lines running in one particular direction. In many contexts with water as Figure, flyta lacks (almost) any elements of motion, emphasizing the big quantity, abundance of water, often undesirable, as in (18): (18) Det flyt-er omkring e-n massa vatten på köksgolv-et. subj “flyta”-pres around a-c lot water on kitchen.floor-def.n.sg ‘There is a lot of water (flowing) on the kitchen floor.’ (Parole)

Summing up, we see that both Polish and Swedish converge in having a verb that can describe aquamotion of both non-liquid and liquid Figures, with motion of (water in) rivers constituting the core of the latter uses (flowing). It is also clear that the liquid-Figure uses share significant semantic similarities with the non-liquid Figure uses of the same verb. In Polish, the distinctive property of płynąć is the pronounced unidirectionality of the motion it refers to, which explains why rivers, pipes and blood vessels provide the best conditions for such motion. In Swedish, the distinctive property of flyta in non-liquid-Figure uses is the passive, slow character of motion or even the absence of any motion, all depending on the dynamics of the water involved (which, in turn, presupposes a considerable amount of water). This slowness or even absence of motion and the presence of a big quantity of water is also characteristic of flyta in its liquidFigure uses. It would in fact be rather unexpected if liquid-Figure uses would be found in the swimming verb to the exclusion of the floating verb in a system

Aquamotion verbs in Slavic and Germanic 335

distinguishing the two. An interesting contrast between Polish and Swedish is provided by descriptions of blood, sweat and tears. Since streams of blood, sweat and tears can evoke an image of a channel, they can occasionally be depicted as Figure to płynąć; however, the relative small quantities of liquid involved in such situations are not compatible with the basic requirements on the use of flyta. Now, in contrast to Polish, Russian distinguishes between aquamotion verbs combining with liquid vs. non-liquid Figure: plyt’/plavat’ apply to non-liquid Figures, while motion of liquid Figures is described by such verbs as lit’sja and teč. The same is true for such Slavic languages as Czech, Bulgarian and a few others. On the other hand, examples like (19) attested in Old Russian (at least in the 16th–18th centuries), Ukrainian, Belorussian and in some Russian dialects testify to the existence of a situation akin to the one found in Polish: (19) a. b.

Old Russian (1688) polacъ by-lъ sozda-nъ rostrum.nom.sg be-pret.m.sg make-part.pret.pass.nom.m.sg nadъ mor-emъ, takъ čto vod-a morsk-aja over sea-instr.sg so that water-nom.sg sea.adj-nom.f.sg pod nego ply-l-a. under it.acc “plyt’”-pret-f.sg ‘A rostrum was built over the sea and the sea water was flowing under it.’ (Rimskie dejanija, 183) Ukrainian Plyv-l-a rik-a, zabut-a “plysti”-pret-f.sg river-nom.sg forgotten-f.sg.nom cily-m svit-om. whole-instr.m.sg world-instr.m.sg ‘The river was flowing, forgotten by the whole world.’ (L. Malkovič, “Tanec’ samotnosti” (2005). Literaturno-mistec’kij al’manax’, 2–3: 69, www.pu.if.ua/data/ukr/lib/e-book/alkos-2-3.pdf, last visited on November 11 2008)

Similarly, within Germanic, languages such as German, Dutch and English maintain the distinction between floating and flowing, in contrast to Swedish; Danish behaves like Swedish, while Norwegian and Icelandic have certain restricted contexts where the cognates of flyta can apply to liquid Figures. Some

. There is a handful of examples in the Russian National Corpus where plyt’ describes the motion of rivers and of liquids, e.g., Kuda plyveš, Volga? ‘Where are you flowing, Volga?’ (Valentina Oseeva. Dinka (1959), quoted after RNC). We have no explanation for these examples which strike us as very marked.

336 Maria Koptjevskaja-Tamm, Dagmar Divjak and Ekaterina V. Rakhilina

other Indo-European languages outside of the Germanic and Slavic families show similar phenomena, e.g., fluitare in Latin (Gruntova 2007) and plaukti in Lithuanian (Arkad’ev 2007) which both can describe aquamotion of both non-liquid and liquid Figures. We cannot judge to what degree these syncretic phenomena are inherited from their proto-languages or are results of later independent developments in particular languages or in particular subgroups of languages. The typologically interesting fact is that there are abundant cross-linguistic parallels to the Swedish and Polish verbs conflating aquamotion of liquid and non-liquid Figures – e.g., in Persian (šenāvar budan), Hindi (bah-), Finnish (solua/soljua), Japanese (nagareru), Hakass (ağarğa), etc. (Maisak & Rakhilina 2007a:â•›51–52). The direction of the semantic extension is not clear, i.e., whether it starts with the motion of liquid Figures and later spreads to non-liquid ones, or the other way round. Both scenarios involve simple metonymy. Thus, the (non-)existence of an aquamotion verb that can apply to both liquid and non-liquid Figures is an additional parameter of cross-linguistic variation, where even closely related languages differ. The next section will continue on the issue of closely related languages in lexical typology.

6.

Lexical typology, closely related languages and diachronic change

Closely related languages often offer instructive examples of significant typological differences that cast light on possible ways in which language systems arise and develop, e.g., expand or shrink. While this is an acknowledged and theoretically significant fact for grammatical and phonetic typology, this is probably even more evident in lexical-typological comparison, since historical changes in the lexicon take on the whole less time than grammatical changes and sound changes. Here we will briefly mention a couple of relevant phenomena in the Slavic and Germanic aquamotion systems. All the Slavic languages have poor aquamotion systems, but this poverty shows somewhat different degrees across Slavic. On the one hand, the Old Russian system was richer than the systems of modern Russian and of most other Slavic languages in having a special verb bresti for active self-propelled motion of animate Figures (swimming) while limiting plouti/plavati to motion of vessels/people on vessels and to passive motion/location (Rakhilina 2007). Bulgarian, on the other hand, has an extremely poor aquamotion system, both within Slavic and within the whole sample of languages examined for the aquamotion project (Ganenkov 2007). More specifically, while some speakers do maintain the directionality-related distinction between the two verbs pluvam/plavam, others have generalized pluvam to (almost) all aquamotion. The only gaps in its uses are

Aquamotion verbs in Slavic and Germanic 337

passive motion of inanimate Figures, including location in or on the surface of water that are covered by nosja se. The different Germanic languages, while operating with similar sets of cognates, distribute their uses in markedly different ways. Thus, the aquamotion system of Modern German has been subject to a historical change that has brought it closer to the poor Slavic systems (cf. Šemanaeva 2007). Although German has two special verbs for passive aquamotion, driften and treiben, these are used optionally. They have to compete with the verb schwimmen that has been generalized to mean both active self-propelled motion (swimming) and passive motion (drifting/floating). Its only “gap” is the motion of vessels and of people aboard, for which German uses general motion verbs as well as segeln ‘sail’, rudern ‘row’, etc. Active motion of vessels and people aboard is, on the whole, an interesting domain within Germanic, where the languages differ quite a lot, as was made clear in Section 2 (cf. Tables 1 and 2). Swedish and German split this zone into four sub-zones in accordance with the kind of vessel involved: the motion of rowing boats and canoes, and sailing boats. This splitting is easy to understand given that motion in each of these cases is accomplished in its own particular way and is driven by its own particular force. Motor-driven boats stand out here: since they are relatively late additions to the traditional vessels, the need to have a label for their motion arose quite recently. Now, although motor-driven boats move in water, they also share certain similarities to other motor-driven vehicles, such as trains, cars etc., which might be the reason for why German and Swedish use general motion verbs for this kind of aquamotion. English, on the other hand, has extended sail to these vessels, thereby reducing the number of sub-zones within motion of vessels and people aboard to three. This step is quite reasonable since motion in motor-driven boats has partly taken over the role of sailing as the fastest and most efficient type of aquamotion. Dutch has undertaken a more radical step in having one verb covering aquamotion of all vessels and people aboard – varen, which however co-exists with optional specialized verbs for motion involving particular kinds of vessels. Interestingly, its German cognate, fahren, is the general verb for going/riding/travelling-not-on-foot. It seems that Dutch first used the general motion verb varen both for “not-self-propelled” aqua- and terra- motion, but later limited it to the former, while terra-motion is covered by rijden. The Dutch system finds therefore certain parallels in the languages with middle aquamotion systems (e.g., Persian, Tamil, Maninka (Maisak & Rakhilina 2007a)), which have two designated aquamotion verbs – one for active self-propelled motion and one for passive motion/location – and use general motion verbs for motion of vessels and of people aboard. The present Dutch system is, however, still richer than these systems in

338 Maria Koptjevskaja-Tamm, Dagmar Divjak and Ekaterina V. Rakhilina

restricting varen to aquamotion and in having optional specialized verbs for motion involving particular kinds of vessels.

7.

Conclusions and implications

The main points of the paper can be summarized as follows: 1. Languages vary considerably as to how they carve up, or categorize the aquamotion domain by means of words and lexicalized expressions; they differ primarily with respect to the number of different aquamotion verbs they have, ranging from one in the “poor system” of Turkish to eight in the “rich system” of Indonesian, and secondarily with respect to the division of labour or the parameters underlying the choice among them. The main parameters underlying semantic categorization within the aquamotion domain across languages are the degree of activity/passivity of motion, which can be viewed as a scale with three main sub-domains Â�– here labelled swimming, sailing, drifting/floating – each centred around its own prototype and representing a particular Manner of motion often limited to a particular kind of Figure, and the presence/absence of direction for distinguishing between drifting and floating. An additional typological parameter is the existence of a verb that can apply to aquamotion of both liquid and non-liquid Figure. The crosslinguistic variation according to these two parameters seems to be independent of each other. 2. The aquamotion systems in Slavic are poor, typically limited to two derivationally related verbs like plyt’/plavat’ in Russian that cover the whole domain. Although the main distinction between the two verbs is apparently of a completely different nature than the distinctions among the aquamotion sub-domains, the latter are still lexicalized as the different senses within the two polysemous lexemes. This is manifested in the asymmetries in the uses of plavat’ and plyt’ and in the different constructions (in)compatible with each of the senses. Thus, the Slavic opposition, in all its uniqueness, is an interesting variation on the generally attested cross-linguistic theme. 3. Closely related languages show significant typological differences in their aquamotion systems and provide invaluable insights for fine-grained semantic analysis and for our understanding of historical processes that create, modify and obliterate such systems. In general, data from closely related languages often deserve more attention in typological studies than they normally receive, and this is particularly true for lexical typology.

Aquamotion verbs in Slavic and Germanic 339

A large portion of studies on motion verbs carried out within the latest two decades starts from the binary distinction between verb-framed and satellite-framed languages, building on Talmy’s (2000) three-fold classification of languages with respect to the encoding of spatial events. This is also true for a number of articles in the present volume. Although we have no doubts about the importance of this work in the domain of motion, we would like to emphasize that there are many more aspects to motion that deserve to be studied and many more parameters to motion verbs along which languages demonstrate systematic cross-linguistic variation. As we hope to have shown in the present paper, aquamotion is an excellent example of a domain where languages show systematic cross-linguistic variation that has previously remained unnoticed. It is worth mentioning that Talmy’s types do not necessarily predict other important aspects of cross-linguistic variation within motion verbs. Thus, the division of the European languages according to the presence/absence of deictic verbs, studied by Ricca (1993), crosscuts Talmy’s types. The same goes for aquamotion, where the two groups of satellite-framed languages, Germanic and Slavic, turn up to be very different in their lexicalization patterns. Cf. the following quote from a paper by Slobin (2003): S[atellite-framed]-languages allow for an economical expression of manner of motion in the main verb of a clause. Apparently as a consequence, these languages make habitual use of manner verbs when encoding motion events, and have developed large lexicons with many fine-grained distinctions of manner, in comparison with smaller and less differentiated manner lexicons in V[erb-framed]languages. One can say that the semantic space of manner of motion is “highly saturated” in S-languages, in comparison with V-languages. For example, French bondir doesn’t distinguish between the manners of motion encoded in English by jump, leap, bound, spring, skip, gambol; Spanish escabullirse can be translated as creep, glide, slide, slip, slither. A detailed study of 115 English manner-of-motion verbs found only 79 French counterparts, many of them of low frequency in comparison with English manner verbs (Jovanović & Kentfield 1998). By contrast, a similar study of Russian and English showed these two S-languages to be comparably saturated on this dimension (Dukhovny & Kaushanskaya 1998). (Slobin 2003:â•›161)

Slobin uses these cross-linguistic differences in the number of manner-of-motion verbs as evidence for a straightforward connection between a language type according to Talmy and the elaboration of its verbal lexicon with respect to manner of motion; this connection provides, in turn, support for Slobin’s “thinking-forspeaking” research. Although the connection pointed out in Slobin’s quote may very well hold for the lexicon of motion verbs on the whole, its different subparts can obviously show their own peculiarities. Thus, counter to the expectations, the

340 Maria Koptjevskaja-Tamm, Dagmar Divjak and Ekaterina V. Rakhilina

satellite-framed Slavic languages turn up to have significantly fewer manner-ofmotion aquamotion verbs than the verb-framed Romance languages Portuguese, Spanish and French. The latter, on the other hand, side with most of the Germanic languages, which are satellite-framed, but show a much higher degree of elaboration within their aquamotion verbal systems than the Slavic ones. Facts like these question the validity of Slobin’s straightforward connection or at least call for its modifications. We hope that the future will see more detailed lexical-typological studies of particular domains, or of particular lexical groups within verbs of motion as a complement to the more general and often imprecise holistic classifications of languages on the basis of their overall motion verb systems.

References Arkad’ev, P. 2007. Glagoly peremeščenija v vode v litovskom jazyke. In Glagoly dviženija v vode: leksičeskaja tipologija, T. Maisak & E. Rakhilina (eds.), 315–333. Divjak, D. 2006. Bewegen in water. Een zaak van lexicon of grammatica? In De taal van Pe‑ ter. Russisch-Nederlandse contacten en contrasten, E. Waegemans (ed.), 55–67. LeuvenÂ�Voorburg: Acco. Divjak, D. & Lemmens, M. 2007. Lexical conflation patterns in Dutch aquamotion verbs. In Glagoly dviženija v vode: leksičeskaja tipologija, T. Maisak & E. Rakhilina (eds.), 152–174. Dukhovny, E. & Kaushanskaya, M. 1998. Russian Verbs of Motion. Ms, Department of Psychology, University of California-Berkeley. Ganenkov, D. 2007. Glagoly peremeščenija v vode: južnoslavjanskie jazyki. In Glagoly dviženija v vode: leksičeskaja tipologija, T. Maisak & E. Rakhilina (eds.), 286–305. Golubkova, E. & Rakhilina, E. 2007. Glagoly plavanija v sovremennom anglijskom jazyke. In Glagoly dviženija v vode: leksičeskaja tipologija, T. Maisak & E. Rakhilina (eds.), 106–127. Gruntova, E. 2007. Latinskaja sistema glagolov plavanija i ee razvitie v romanskix jazykax (fran‑ cuzskom, ital’janskom, ispanskom). In Glagoly dviženija v vode: Leksičeskaja tipologija, T.Â€Maisak & E. Rakhilina (eds.), 231–266. Jovanović, J. & Kentfield, M. 1998. Manifold manner: An exploratory analysis of French and English verbs of motion. Ms, Department of Psychology, University of CaliforniaÂ�Berkeley. Koptjevskaja-Tamm, M. 2007. Švedskie glagoly dviženija v vode. In Glagoly dviženija v vode: leksičeskaja tipologija, T. Maisak & E. Rakhilina (eds.), 128–151. Lander, Yu. & Kramarova, S. 2007. Indonezijskie glagoly plavanija i ix sistema. In Glagoly dviženija v vode: Leksičeskaja tipologija, T. Maisak & E. Rakhilina (eds.), 664–693. Maisak, T. & Rakhilina, E. 2007a. Glagoly dviženija i naxoždenija v vode: leksičeskie sistemy i semantičeskie parametry. In Glagoly dviženija v vode: leksičeskaja tipologija, T. Maisak & E.Â€Rakhilina (eds.), 27–75. Maisak, T. & Rakhilina, E. (eds). 2007b. Glagoly dviženija v vode: Leksičeskaja tipologija. Moskva: Indrik.

Aquamotion verbs in Slavic and Germanic 341

Prokof ’eva, I. 2007. Peremeščenie v vode i peremeščenie vody: Glagoly płynąć/pływać ‘plyt’/pla‑ vat’’ v pol’skom jazyke. In Glagoly dviženija v vode: Leksičeskaja tipologija, T. Maisak & E.Â€Rakhilina (eds.), 305–314. Rakhilina, E. 2007. Glagoly plavanija v russkom jazyke. In Glagoly dviženija v vode: Leksičeskaja tipologija, T. Maisak & E. Rakhilina (eds.), 267–285. Ricca, D. 1993. I verbi deittici di movimento in Europa: Una ricerca interlinguistica. Firenze: La Nuova Italia Editrice. Slobin, D. I. 2003. Language and thought online: Cognitive consequences of linguistic relativity. In Language in Mind: Advances in the Study of Language and Thought, D. Gentner & S.Â€Goldin-Meadow, S. (eds.), 157–192. Cambridge MA: The MIT Press. Šemanaeva, O. 2007. Vyraženie peremeščenija v vode v nemeckom jazyke. In Glagoly dviženija v vode: leksičeskaja tipologija, T. Maisak & E. Rakhilina (eds.), 175–197. Talmy, L. 2000. Toward a Cognitive Semantics, Vol. I & II. Cambridge MA: The MIT Press.

Data sources for the examples quoted in the paper KJPWN = Korpus Języka Polskiego Wydawnictwa Naukowego PWN, http://korpus.pwn.pl/ Parole = the Parole corpus at the Swedish Language Bank, http://spraakbanken.gu.se/ Rimskie dejanija (Gesta Romanorum). 1878. Sankt Petersburg, v. 1–2. RNC = Russian National Corpus, http://www.ruscorpora.ru/

chapter 14

Metaphorical walking Russian idti as a generalized motion verb* Tore Nesset

University of Tromsø

This article explores the metaphorical use of Russian idti ‘walk’ as a generalized motion verb in collocations like poezd idet ‘(lit.) train walks’. The study seeks to explain why idti is used in such collocations, and why this usage is restricted to examples where a goal is prominent in the context (Rakhilina 2004). It is suggested that idti is used as a generalized motion verb in metaphors because it represents prototypical motion. In metaphor, it is natural to take prototypical motion as the starting point. Unidirectionality is dominant in metaphorical usage, and this explains why idti is restricted to goal-oriented contexts. The analysis involves the notions of prototype, anthropocentrism, embodiment and metaphor and lends support to a cognitive approach to linguistic categories.

Although dictionaries translate Russian idti as ‘walk’, idti is widely used metaphorically as a generalized motion verb in contexts where walking is not involved: (1) Tramvaj šel po teperešnej Kropotkina. tram went along present Kropotkin ‘The tram went along what is now Kropotkin Street.’ [Oleša: Kniga proščanija (1930–1959)]

* This paper was presented at the Slavic Cognitive Linguistics Conference in Chicago, October 2007. I would like to thank the participants for questions and discussion. Thanks also to Laura Janda, two anonymous reviewers and the editors of this volume for detailed comments on earlier versions of the paper. . The majority of examples cited in this paper are from the Russian National Corpus (www. ruscorpora.ru). For corpus examples, the source is given in square brackets (either the author and a name of a work of fiction, or the name of a periodical). Examples with no source in square brackets are from Rakhilina (2000, 2004). Exact references are given in the text.

344 Tore Nesset

(2) Ja vzjal bilet na teploxod “Alušta”, šedšij iz Odessy v Soči. I took ticket for steamer Alušta going from Odessa to Soči ‘I bought a ticket for the steamer “Alušta”, which was going from Odessa to Sochi.’ [Najman: Slavnyj konec besslavnyx pokolenij (1994)]

This paper addresses two questions concerning generalized idti. Why is idti used as the source domain in metaphorical examples of generalized motion like (1) and (2)? Why does generalized idti show affinity to “goal-oriented” contexts, as observed by Rakhilina (2000:â•›303ff., 2004:â•›7)? In this study, I will argue that idti is used metaphorically because it represents a prototypical kind of motion. It will furthermore be argued that metaphorical idti is used to describe “goal-oriented” motion because the unidirectionality component of the meaning of idti becomes more salient in metaphorical usage. The interest of my analysis goes beyond Russian. First, as observed by Majsak and Rakhilina (2007:â•›45), the generalized use of verbs like idti is quite common cross-linguistically. Second, the proposed analysis has theoretical implications in that it lends support to key concepts in cognitive linguistics, viz. prototypes, anthropoÂ�centrism, embodiment and metaphor. After a short introduction in Section 1, I will discuss prototypical motion in Section 2, and metaphorical motion in Section 3. The contribution of the paper is summarized in Section 4.

1.

Introduction: Three conceptual layers

Russian verbs of motion involve three conceptual layers, “manner”, “directionality” and “path”, which correspond to the verbal root, suffixation, prefixation. The verb root specifies manner (in the sense of Talmy 1985). Manner is a cover term that involves information about mover, gait, speed, effort, and body part (Dodge and Lakoff 2005:â•›68). The root /polz/ in polzti and polzat’ characterizes the manner of motion as crawling, while the root /l’ot/ in letet’ and letat’ indicates flying. These examples furthermore illustrate that Russian verbs of motion occur in pairs. The roots /polz/ and /l’ot/ combine with different suffixes, and the combination of root plus suffix indicates whether we are dealing with unidirectional, goal-oriented motion or not. The verbs polzti and letet’ are what we may call “unidirectional” motion verbs, whereas polzat’ and letat’ are “non-directional”. We will return to . Different authors use different terms to describe the pairs of motion verbs. In English, the most widely used terms seem to be determinate/indeterminate (Foote 1967; Forsyth 1970; Jakobson 1966; Timberlake 2004; Ward 1965) and unidirectional/multidirectional (Mahota 1996; Wade 1992). I prefer unidirectional to determinate, because the former gives a better indication of the meaning of the verbs in question. I will not use multidirectional, however,

Metaphorical walking 345

Inner layer: • Root • Indicates manner of motion

vletet’

Intermediate layer: • Root + suffix • Indicates directionality Outer layer: • Prefixed verb • Indicates path

Figure 1.â•‡ The three conceptual layers of Russian verbs of motion

the use of unidirectional and non-directional verbs in Section 3. For now it is sufficient to note that directionality (unidirectional and non-directional) represents the second, intermediate conceptual layer of Russian verbs of motion. The third and outermost layer is introduced by prefixes, which add a path (in the sense of Talmy 1985) to the meaning of the verb. For instance, in vletet’ ‘fly into’ the prefix v- indicates motion into something, while vy- in vyletet’ ‘fly out of ’ specifies the opposite path: (3) (4)

Ja pil čaj, a na čerdak vletel artillerijskij snarjad. I â•›drank tea and on attic in.flew artillery shell ‘I was drinking tea, when an artillery shell flew into the attic.’ [Novaja gazeta 2003.01.15] Iz sosednej komnaty vyletela bol’šaja temnaja ptica i tichon’ko from neighboring room out.flew big dark bird and lightly zadela krylom lysinu bufetčika. grased with.wing bald.spot of.bartender ‘A big dark bird flew out of the neighboring room and lightly grazed the bartender’s bald spot with its wing.’ [Bulgakov: Master i Margarita (1929–1940)]

The three conceptual layers are illustrated in Figure 1. In this study I will focus on the two inner layers, i.e., non-prefixed verbs, since they are sufficient to shed light on the generalized use of idti, which is the

because this term covers only one of the meanings of the relevant verbs. A more precise term would be “non-unidirectional”, which corresponds to the Russian term glagoly neod‑ nonapravlennogo dviženija employed by the Academy Grammar (Švedova 1980). However, in the following I shall use the somewhat simpler term non-directional.

346 Tore Nesset

Table 1.â•‡ Russian verbs of motion Unidirectional verb

Non-directional verb

Gloss

bežat’ vesti vezti gnat’ exat’ idti katit’ lezt’ letet’ nesti plyt’ polzti taščit’

begat’ vodit’ vozit’ gonjat’ exat’ xodit’ katat’ lazit’ (lazat’) letat’ nosit’ plavat’ polzat’ taščit’

‘run’ ‘lead’ ‘convey, transport’ ‘drive, chase’ ‘travel, ride’ ‘go, walk’ ‘roll’ ‘climb’ ‘fly’ ‘carry’ ‘swim, float’ ‘crawl’ ‘drag’

topic of this paper. I will not discuss motion verbs with the so-called postfix -sja, since they do not contribute to the issue under scrutiny in this study. In the following, we will be concerned with the thirteen pairs of verbs listed in Table 1. Some sources include more verbs, such as bresti/brodit’ ‘wander’, lomit’/lo‑ mat’ ‘break’, mčat’/mykat’ ‘rush’, and valit’/valjat’ ‘drag’. However, in presentday Russian, the members of these four pairs do not display the same semantic difference as the verbs in Table 1 and therefore do not constitute pairs of the same type as those listed in the table (Garde 1980:â•›386; Isačenko 1982:â•›427; Wade 1992:â•›340; Ward 1965:â•›250).

2.

Prototypical motion

An important contribution of cognitive linguistics has been to integrate categorization by prototypes into linguistic theory. Prototypes came to the attention of linguists from the work of Eleanor Rosch and her associates in psychology (e.g., Rosch 1978), and from the early 1980s cognitive linguists started exploring the implications of prototypes for linguistic categories (cf. e.g., Geeraerts 1989; Langacker 1987; Lakoff 1987; Nathan 1986; Taylor 1989). In a nutshell, . It should be noted that prefixed motion verbs are attested in generalized use as shown by examples like Korabl’ ušel dal’še na sever, k ust’ju Krivoj. ‘The ship went further north, to the mouth of the river Krivaja.’ [Golovanov: Ostrov, ili opravdanie bessmyslennyx putešestvij (2002)]. However, discussion of examples like this are beyond the scope of the present study.

Metaphorical walking 347

categorization by prototype involves comparing something to a central, representative subcategory. For example, this allows us to establish that robins and swallows are fairly prototypical birds, while penguins and ostriches are peripheral members of the category. Categorization by prototype differs from the traditional, Aristotelian conception where categories are defined in terms of necessary and sufficient criteria. In the Aristotelian line of reasoning, robins, swallows, penguins and ostriches are all birds if they meet the necessary and sufficient criteria for birdhood (whatever they are). However, Aristotelian categories do not enable us to say anything about central and peripheral subcategories, because all members meet the same set of criteria. Aristotelian categories lack internal structure. Cognitive linguists generally acknowledge the need to include both prototypes and Aristotelian categories in linguistics (Taylor 1989:â•›59). In the following, however, I will focus on categorization by prototype, suggesting that walking represents a prototypical kind of motion and that the Russian walking verbs idti/xodit’ are prototypical. In order to corroborate this view, we need to compare idti/xodit’ to the remaining verbs of motion in Table 1. Let us start by comparing idti/xodit’ and polzti/polzat’ ‘crawl’. The former pair involves an erect posture, while the latter denotes movement in a prone position with the body resting on the ground, or on the hands and knees. Clearly, walking represents the typical case. Once a child has started walking, s/he will prefer this kind of motion unless special conditions make crawling appropriate. In a similar vein, idti/xodit’ are prototypical compared to лезть/лазить ‘climb’. Unless the presence of an impediment forces us to climb, we prefer walking. The verbs bežat’/begat’ ‘run’ involve extra effort and speed. Again, this may be desirable in certain circumstances, but in the normal case humans walk when they move about. Movement on the ground is more natural for humans than motion in air or water. Learning to fly is very difficult (if not impossible) for humans, and although we can learn to swim, walking represents a much more convenient way of moving from the perspective of humans. Therefore, idti/xodit’ are more prototypical than letet’/letat’ ‘fly’ and plyt’/plavat’ ‘swim’. In addition to ‘swim’, plyt’/plavat’ means ‘sail’. This takes us to the motion verbs that involve vehicles, such as ezdit’/exat’ ‘ride’, which are used about riding on horseback, but also driving cars and other vehicles. The use of vehicles represents techniques developed by humans to accommodate special needs. However . As shown by Rakhilina (2007), plyt’/plavat’ also covers the meaning ‘float’, cf. examples like [...] po strogoj vode uže plyli opavšie list’ja. ‘Fallen leaves were already floating on the rough water.’ [Poljanskaja: Proxoždeni teni (1996)]. A thorough analysis of the meaning of plyt’/plavat’ is beyond the scope of the present study.

348 Tore Nesset

important as these technological revolutions are for human culture, walking remains the most natural, prototypical manner of motion. So far we have only considered intransitive motion verbs. The transitive verbs in Table 1 are vesti/vodit’ ‘lead’, vezti/vozit’ ‘transport’, gnat’/gonjat’ ‘chase’, katit’/katat’ ‘roll’, nesti/nosit’ ‘carry’ and taščit’/taskat’ ‘drag’. Among the transitive verbs, nesti/nosit’ ‘carry’ are closest to idti/xodit’. Carrying is essentially walking with the additional specification that the walker has something in his or her arms or on his or her back. Idti/xodit’ represents the simpler concept, and furthermore a more natural way of moving. Carrying is something you do if you have to move something to a different location. Otherwise, you just walk. The remaining transitive motion verbs have more complex meanings, but they all combine some kind of intransitive motion with the concept of moving an object. The transitive motion verbs therefore stand out as less prototypical. We have now compared the Russian motion verbs along six dimensions, and established that idti/xodit’ ‘walk’ are prototypical on all of them. In this way, idti/ xodit’ represent prototypical motion. Figure 2 illustrates this. Idti/xodit’ are placed in the center of the category and connected to the remaining verbs by means of dashed arrows that represent comparison along the six dimensions described above. In the terminology of Langacker (1987:â•›369ff.), the non-prototypical kinds of motion represent extensions from the central subcategory of walking. Each extension arrow is marked with a word in italics which represents the relevant dimension. In order to avoid unnecessary complications, I include only one verb pair for each dimension. Bear in mind that the figure is not intended as a complete analysis of the structure of the category. For one thing, the figure does not capture any relationships among the peripheral subcategories, although such connections polzti/ polzat’ nesti/ nosit’

posture

object idti/ xodit’

vehicle exat’/ ezdit’ “medium”

lezt’/ lazit’ impediment speed

plyt’/ plavat’

Figure 2.â•‡ Idti/xodit’ representing prototypical motion

bežat’/ begat’

Metaphorical walking 349

may well exist. The ordering of the peripheral subcategories in the figure does not reflect their degree of semantic relatedness. Second, the peripheral subcategories are not mutually exclusive. For instance, vezti/vozit’ ‘transport’ are transitive and at the same time involve a vehicle. In order to make the figure as simple as possible, I have placed each verb pair in one subcategory only. However, while the figure does not capture all the details about verbs’ meanings and the relations between them, the figure suffices to show that walking is prototypical – which is what matters for the purposes of the present study. The analysis sketched above, simple as it is, has implications for two important theoretical issues. First of all, it demonstrates the need to recognize categories organized around prototypes in linguistics. The notion of prototype enables us to capture generalizations about the relations between motion verbs; without prototypes these generalizations would not be accounted for. Second, the analysis presented above offers a strong argument in favor of an anthropocentric and embodied perspective on language. Xodit’/idti are only prototypical from the perspective of humans. If we adopted an objectivist position which ignores the fundamental role of human experience, the perspective of a fish or a bird would be just as valid as a starting point. Presumably, for a fish swimming is the prototypical kind of motion, while flying would be the prototypical kind of motion for most birds. In other words, if we do not adopt an anthropocentric approach to linguistic categorization (as suggested by Wierzbicka 1988:â•›2, 526 et passim), we are not able to make sense of the Russian motion verbs. Anthropocentrism is closely related to the notion of embodiment, which in cognitive linguistics is used to indicate that our conceptualization is grounded in experience with our bodies and how we use them in interacting with our surroundings (Lakoff & Johnson 1999:â•›36, see also Johnson 1987:â•›19ff. and Feldman 2006:â•›95ff.). This perspective is evident in the analysis presented above. Idti/xodit’ occupy a central position in the category of motion verbs exactly because they represent the prototypical way to interact with our bodies.

3.

Metaphorical motion

In the beginning of the paper, I asked why idti is used as the source domain in metaphorical examples of generalized motion like (1) and (2). We are now in a position to give a principled answer to this question. Idti is used as the source domain for metaphors because for humans it represents the prototypical manner of motion. This insight, in turn, enables us to address the second question mentioned in the beginning of the article, namely why generalized (metaphorical) idti shows affinity to “goal-oriented” contexts. Rakhilina (2000:â•›304, 2004:â•›9) considers examples like this:

350 Tore Nesset

(5) Korabl’ idet / ??plyvet v port. ship goes ??sails to harbor ‘The ship is going/??sailing to the harbor.’

She observes that idti is preferable because there is a specific itinerary provided by the adverbial в порт ‘to the harbor’. Notice that Rakhilina’s claim is not that plyt’ is excluded in sentences like (5); what she suggests is that the use of idti involves a stronger emphasis on the goal than plyt’. Rakhilina does not specify whether (5) is a constructed example, but similar examples with idti are attested in the Russian National Corpus: (6) Poslednjaja šxuna idet v gavan’. last schooner goes to harbor ‘The last schooner is going to the harbor.’ [Žemajtis: Bol’šaja laguna (1977)]

If no itinerary is specified, plyt’ is preferable to idti, as illustrated by the following example from a popular song (Rakhilina 2004:â•›10), where the focus is on the manner of motion rather than its goal: (7) Plyla / *šla, kačalas’ lodočka. swam / *went rocking boat ‘There was a boat sailing/*going and rocking.’

The generalization of the “goal-orientation” of idti is not restricted to motion in water. The generalization implies that generalized idti is particularly felicitous for streetcars and trains, which follow fixed itineraries and have no freedom of movement since they move on rails. Examples like the following are indeed very common: (8) Ja udivilsja, tramvaj šel v tom že napravlenii. I was.surprised streetcar went in same direction ‘I was surprised – the streetcar went in the same direction.’ [Svetov: Moe otkrytie muzeja, Znamja (2001)] (9) Poezd idet v Moskvu v dvenadcat’ časov. train goes to Moscow in twelve o’clock ‘The train goes to Moscow at twelve o‘clock.’

[Leskov: Nekuda (1864)]

Exat’ is less common in sentences like (8) and (9). While a search in the Russian National Corpus for poezd + idti gave 208 examples, there were only thirteen examples with poezd + exat’ in the corpus. Interestingly, eight of the examples with

. Corpus search performed on July 18, 2008.

Metaphorical walking 351

exat’ describe the speed of the train. The focus on the speed backgrounds the goal of the motion, which may make the use of exat’ more felicitous: (10) Poezd exal očen’ medlenno, ostanavlivalsja, gas svet. train drove very slowly stopping went.out light ‘The train went very slowly, stopped, and the light went out.’ [Najman: Ljubovnyj interes (1998–1999)]

For cars, which unlike trains normally do not follow fixed itineraries, both exat’ and idti are common. However, on the basis of the following examples Rakhilina (2000:â•›304, 2004:â•›9) argues that there is a subtle difference in meaning: (11) Ne volnujtes’, za nami v aèroport uže idet / ?edet mašina. NEG worry after us to airport already goes / ?drives car ‘Don’t worry, a car is already coming to pick us up at the airport.’ (12) Smotri, za nami edet / ??idet kakaja-to mašina. look after us drives / ??goes some car ‘Look a car is following us.’

Although exat’ is possible in (7) idti is preferred because the car in question follows a fixed itinerary. However, Rakhilina judges exat’ felicitous in (8), where the car most likely is moving behind us accidentally. According to Rakhilina (2004:â•›11ff.), the situation for movement in air is somewhat less clear. However, heavenly bodies, which indeed have fixed “itineraries”, combine with idti: (13)

Ne zabyvaj, solnce idet s vostoka, u vas po utram NEG forget sun goes from east with you in mornings budet žarko. will.be hot ‘Don’t forget that the sun goes from the east; it is going to be hot at your place in the mornings.’ [Ščerbakova: Mitina ljubov’ (1996)]

On the face of it, the following corpus example questions Rakhilina’s judgment that letet’ is incompatible with heavenly bodies: (14)

No on znal: ničego teper’ ne moglo, ne dolžno slučit’sja, vse but he knew nothing now NEG could NEG should happen everything bylo sčastlivoe, legkoe, solnce letelo. was happy light sun flew

. Heavenly bodies also combine with plyt’ and dvigat’sja. For a detailed account of the metaphorical use of plyt’, the reader is referred to Rakhilina (2007:â•›274ff.). The use of dvigat’sja is discussed in Rakhilina (2004:â•›15ff.).

352 Tore Nesset

‘But he knew that nothing would, nothing could happen; everything was happy, light, the sun was flying.’ [Zamjatin: Ëla (1928)]

However, upon closer inspection (14) lends further support to Rakhilina’s argument; the sun is associated here with happiness and is described as if it were able to loosen itself from its orbit – and fly. As pointed out by Rakhilina (2004:â•›11), in the Russian folk taxonomy birds are associated with freedom (cf. the saying svoboden kak ptica ‘free like a bird’). Since freedom is antagonistic to goal-bias, we should expect birds to combine with letet’, and not with idti. According to Rakhilina (2004:â•›11), this prediction is borne out by the facts. The verb idti cannot replace letet’ in the following example (Rakhilina 2004:â•›6): (15) Lastočka s vesnoju v seni k nam letit / *idet. swallow with spring to porch to us flies / *goes ‘A swallow arrives at our doorstep with the spring.’ [Volos: Nedvižimost’ (2000)]

The Russian National Corpus provides 100 examples of ptica + letet’, but only one apparent counterexample where ptica collocates with idti: (16) Pticy šli po nebu rovno, spokojno, krasivo. birds went on sky evenly quietly gracefully ‘The birds moved across the sky evenly, quietly and gracefully.’ [Jašin: Malen’kie rasskazy (1954–1962)]

However, this example is not at variance with Rakhilina’s generalization about the goal-orientation of generalized idti. The use of the adverb rovno ‘evenly’ suggests that in (16) the birds are described as if they were following a fixed itinerary across the sky. In the same way as birds, insects and dragons combine with letet’, since they do not have fixed itineraries: (17)

Čeloveku nado otdoxnut’ v puti, a strekozy i babočki human needs to.rest in trip but dragon-flies and butterflies letjat, ne otdyxaja, i desjat’, i dvadcat’, i sto verst. fly NEG resting and ten and twenty and hundred verst ‘A human must rest on his/her way, but dragon-flies and butterflies fly without rest for ten, twenty, and even a hundred versts.’ [Bragin: V strane dremučix trav (1962)]

. I did not count examples like the following, where the birds are actually walking on the ground: Sytye, belye, važnye pticy šli po plotine. ‘Plump, white, important birds were walking along the dam.’ [A. Tolstoj: Xoždenie po mukam (1941)].

Metaphorical walking 353

(18) Drakon letel, lenivo vzmaxivaja kryl’jami. dragon flew lazily flapping wings ‘The dragon was flying, lazily flapping his wings.’ [Lazarčuk & Uspenskij: Posmotri v glaza čudovišč (1996)]

The examples discussed in this section cannot do justice to Rakhilina’s (2000, 2004) detailed analysis of a substantial body of data, but suffice to show that the contexts where generalized (metaphorical) idti can replace plyt’, exat’ or letet’ have something in common: they all involve a strong focus on the goal of motion. The question is why generalized idti is goal-oriented. In order to see that, we need to consider the meaning difference between idti and xodit’, and more generally the difference between the unidirectional and non-directional verbs in Table 1. In Nesset (2000), I give a detailed analysis of the relationship, but for the purposes of the present study, the following brief overview will suffice. As suggested by the term, unidirectional verbs like idti are used about motion in one direction towards a goal. Consider the following examples: (19) Poètomu on bežit v kino, čtoby v temnote otdyšat’sja. therefore he runs to movies in.order.to in darkness catch.breath ‘Therefore he runs to the movie theater in order to catch his breath in the dark.’ [Izmajlov: Naši ljudi (1984)] (20) Kostja i Njura idut k babuške. Kostja and Njura go to grandmother ‘Kostja and Njura are walking to their granmother’s place.’ [Dubov: Ogni na reke (1966)] (21) Lisa polzla k nim s podvetrennoj storony. fox crawled to them from leeward side ‘The fox was crawling towards them from the leeward side.’ [Mamin-Sibirjak: Malinovye gory (1899)]

In examples (19)–(21) the goal is expressed by prepositional phrases. However, this is not necessary, and unidirectional verbs often combine with prepositional phrases representing the starting point: (22) My byli vynuždeny idti iz gostinicy peškom. we were forced to.go from hotel on.foot ‘We were forced to leave the hotel on foot.’ [Arxipova: Muzyka žizni (1996)]

The non-directional motion verbs are used whenever we are not dealing with motion in one direction (cf. Isačenko 1982:â•›421). It may be fruitful to distinguish between three types of situation, all of which are described in major grammars

354 Tore Nesset

such as Isačenko (1982:â•›422) and Timberlake (2004:â•›412). First, we have “motion round about”, i.e., motion in various directions with no particular goal: (23) On xodit, točno letaet; ego budto kto-to nosit po komnate. he walks like flies him as.if somebody carries around room ‘He is walking, almost flying; it is as if someone is carrying him around the room.’ [Gončarov: Oblomov (1859)] (24)

Izvestno, čto on mnogo ezdil po Evrope i pobyval well-known that he much drove around Europe and has.been daže v Amerike. even in Amerika ‘It is well known that he traveled a lot in Europe and had even been to America.’ [Rossijskaja muzykal’naja gazeta 2003.03.12]

The second type of situation involves “movement back and forth”, i.e., movement to a goal and back again: (25) Borisjuk ezdil v Avstraliju? Borisjuk drove to Australia ‘Has Borisjuk been to Australia?’ (26)

[Argumenty i Fakty 2001.04.04]

Včera on xodil v kino i dolžen byl pokazat’ mestnym, kak yesterday he went to movies and had.to AUX show locals how odevajutsja normal’nye ljudi. dress normal people ‘Yesterday he went to the movie theater and was supposed to show the locals how normal people dress.’ [Bolmat: Sami po sebe (1999)]

. The three situations represent broad categories with several subcategories, and there are uses that lie beyond the three subcategories. Consider the following example, which is more about location than movement (Majsak and Rakhilina 2007:â•›53ff. and Rakhilina 2007:â•›271): U dverej ee, na taburetke, stojal taz, i v nem plavali vyčesannye volosy. ‘By her door on a stool there was a wash-basin, and combed out hairs were floating in it.’ [Oleša: Zavist’ (1927)] However, detailed discussion of examples like this is beyond the scope of the present study. . The “back and forth” type explains the widespread use of non-directional motion verbs for repeated events. In order to go to the movie theater again in (26), for example, one has to go back home before repeating the event. Notice, however, that unidirectional motion verbs are used about repeated events, as shown by examples like Inogda Mixas’ka ostavljal ix vdvoem i bežal daleko vpered, a potom ostanavlivalsja i smotrel, kak oni idut k nemu. ‘Sometimes Mixas’ka would leave the two of them alone and run far ahead, but afterwards he would stop and watch them walk towards him.’ [Lixanov: Čistye kamuški (1967)] Here, we are dealing with repeated motion in one direction, and accordingly the unidirectional verbs bežat’ ‘run’ and idti ‘walk’ are used.

Metaphorical walking 355

(27)

Ešče buduči na gastroljax v Rige vesnoj 1959 goda, ja letal v while being on tour in Riga in.spring 1959 year I flew to Moskvu, gde repetiroval v “Sovremennike” pervuju svoju rol’ v Moscow where rehearsed in Sovremennik first my role in spektakle “Vzlomščiki tišiny”. play Thieves of.silence ‘While I was on tour in Riga in the spring of 1959, I made a roundtrip flight to Moscow where I rehearsed my first role in the play “The thieves of silence” [Kozakov: Akterskaja kniga (1978–1995)] at the “Sovremennik” theater.’

The third type of context involves what Timberlake (2004:â•›412) calls the “essentialist idea of a certain type of activity”. For simplicity, I will use the terms “ability situation”. In the following examples, the focus is on the ability to move, while directionality is irrelevant. (28) On zabyl, čto ne umeet plavat’, i, estestvenno, utonul. he forgot that NEG can swim and naturally drowned ‘He forgot he wasn’t able to swim, and, naturally, drowned.’ [Amurskij meridian (Khabarovsk), 2004.12.22] (29)

A moi deti, Tiša i Toša, vsjudu vmeste, odin ešče polzaet, and my kids Tiša and Toša everywhere together one still crawls a drugoj uže xodit. and other already walks ‘And my kids, Tiša and Toša, are always together, one is still crawling, while the other is already walking.’ [Petruševskaja: Nevinnye glaza (1998–2000)]

(30) Vidor naučilsja letat’ po-nastojaščemu. Vidor learned to.fly for.real ‘Vidor learned to fly for real.’

[Sem’ja 2000.01.19]

Since the non-directional verbs are compatible with different kinds of situations, and since, as shown in (28)–(30), these verbs are used when directionality is irrelevant, we can conclude that non-directional verbs are semantically unmarked (cf. Nesset 2000 for further discussion). Unidirectional verbs like idti, on the other hand, must be specified for directionality. Summarizing, the meaning of idti (used non-metaphorically) involves two components: (31) The meaning of idti: a. Motion on foot at normal speed. b. Motion in one direction towards a goal.

Clearly, (31) does not offer an exhaustive description of the meaning of idti. In particular, the brief statement in (a) does not do justice to the analysis of proto-

356 Tore Nesset

typical motion in Section 2. However, (31) is sufficiently precise for a discussion of the metaphorical use of idti as a generalized verb of motion. Metaphor has been one of the cornerstones of cognitive linguistics ever since Lakoff and Johnson (1980) argued that metaphor plays a pivotal role in linguistic semantics. Metaphor can be defined as “a cross-domain mapping in a conceptual system” (Lakoff 1993:â•›203).10 When idti is used metaphorically, we can identify a mapping from a domain of walking on foot to a domain of generalized motion. Mapping means that some of the structure from the source domain (walking on foot) carries over to the target domain (generalized motion). Consider the following example: (32) Vot idet poezd – normal’no, po rel’sam. here goes train normally along rails ‘Here comes the train – as usual, along the rails.’ [Šukšin: Priezžij (1970–1974)]

In this example, the idea of motion along a path towards a goal is mapped onto the target domain. This makes it possible to make inferences. Since we know that when people are walking in one direction towards a goal, they normally reach that goal, we may infer that the train in (32) will arrive at its destination (where we are waiting for it). However, as is normally the case in metaphor, not all structure from the source domain carries over to the target domain of a metaphor. Example (32) illustrates this. Since trains do not have legs, it is clear that the part concerning motion on foot is not included in the target domain. Simplifying somewhat, we can say that the directionality component in (31b) carries over to the target domain, whereas the specification that idti involves motion on foot in (31a) does not. In the beginning of this section, I argued that idti is used metaphorically as a generalized motion verb because it represents a prototypical kind of motion. However, why is the metaphorical use restricted to “goal-oriented” contexts? The analysis sketched in this section enables us to give a principled answer to this question. As we have seen, the component involving motion on foot does not carry over to the target domain of the metaphor. As a consequence of this, the second component of the meaning, unidirectionality, becomes more salient in the target domain.

10. The cognitive linguistics approach to metaphor adopted in this article is inclusive, insofar as all examples of idti not involving actual movement on foot are classified as metaphorical. However, since cognitive linguistics endorse gradient categories, it is possible to accommodate the intuition that expressions like krov’ idet ‘the blood is flowing’ are “more metaphorical” than poezd idet ‘the train is coming’. The cognitive linguistics approach to metaphor is controversial, but a comparison with other conceptions is beyond the scope of the present study. For a recent critical discussion, the reader may consult Padučeva (2004:â•›172ff.).

Metaphorical walking 357

In a sense, in metaphors unidirectionality does not have to compete for attention with the foot-based motion component. The prediction that follows from this is that unidirectionality is dominant in metaphorical uses. This prediction is borne out by the facts discussed in the beginning of this section. By way of illustration, consider again example (9), which for convenience is repeated here: (33) Poezd idet v Moskvu v dvenadcat’ časov. train goes to Moscow in twelve o’clock ‘The train goes to Moscow at twelve o‘clock.’

[Leskov: Nekuda (1864)]

Since trains have fixed itineraries and are bound to move on rails, we expect frequent use of the “goal-oriented” generalized idti in sentences like this. As we have seen, this prediction is indeed borne out by the facts. On the other hand, we do not expect idti to collocate with ptica, since birds are associated with freedom and do not have fixed itineraries. Once again this prediction is borne out by the facts from the Russian National Corpus reviewed earlier in this section.

4.

Conclusion

The analysis of Russian motion verbs proposed in this paper suggests three conclusions. First of all, I have argued that idti is used as a generalized motion verb in metaphors because it represents prototypical motion. In metaphorical speaking and reasoning about motion, it is natural to take prototypical motion as the point of departure. Second, the analysis enables us to explain the affinity between generalized, metaphorical idti and “goal-oriented” contexts. Since unidirectionality is dominant in the target domain, we predict that generalized idti occurs in “goal-oriented” contexts, which is exactly what we observe. The third and final conclusion takes us beyond Russian and Slavic linguistics. As we have seen, the proposed analysis crucially involves the notions of prototype, anthropocentrism, embodiment and metaphor. The analysis therefore testifies to the value of these key concepts of cognitive linguistics and lends support to a cognitive approach to linguistic categories.

References Dodge, E. & Lakoff, G. 2005. Image schemas: From linguistic analysis to neural grounding. In From Perception to Meaning: Image Schemas in Cognitive Linguistics, B. Hampe (ed.), 57–91. Berlin: Mouton de Gruyter. Feldman, J. A. 2006. From Molecule to Metaphor. A Neural Theory of Language. Cambridge MA: The MIT Press.

358 Tore Nesset

Foote, I. P. 1967. Verbs of Motion. Cambridge: CUP. Forsyth, J. 1970. A Grammar of Aspect. Usage and Meaning of the Russian Verb. Cambridge: CUP. Garde, P. 1980. Grammaire russe. Paris: Institut d’études slaves. Geeraerts, D. 1989. Prospects and problems of prototype theory. Linguistics 27: 587–612. (Reprinted in Geeraerts, D. 2006. Words and Other Wonders. Berlin: Mouton de Gruyter, 3–26). Isačenko, A. V. 1982. Die russische Sprache der Gegenwart. Formenlehre. Munich: Max Hueber Verlag. Jakobson, R. 1966. Relationship between Russian stem suffixes and verbal aspects. In Selected Writings of Roman Jakobson 2: Word and Language, 198–202. The Hague: Mouton. Johnson, M. 1987. The Body in the Mind. The Bodily Basis of Meaning, Imagination, and Reason. Chicago IL: The University of Chicago Press. Lakoff, G. 1987. Women, Fire, and Dangerous Things. What Categories Reveal about the Mind. Chicago IL: University of Chicago Press. Lakoff, G. 1993. The contemporary theory of metaphor. In Metaphor and Thought, A. Ortony (ed), 202–251. Cambridge: CUP. Lakoff, G. & Johnson, M. 1980. Metaphors We Live By. Chicago IL: University of Chicago Press. Lakoff, G. & Johnson, M. 1999. Philosophy in the Flesh. New York NY: Basic Books. Langacker, R.W. 1987. Foundations of Cognitive Grammar, Vol. 1: Theoretical Prerequisites. Stanford CA: Stanford University Press. Mahota, W. J. 1996. Russian Motion Verbs for Intermediate Students. New Haven CT: Yale University Press. Majsak, T. A. & Rakhilina, E. V. 2007. Glagoly dviženija i naxoždenija v vode: leksičeskie sistemy i semantičeskie parametry. In Glagoly dviženija v vode: Leksičeskaja tipologija, T. A. Majsak & E. V. Rakhilina (eds.), 27–75. Moscow: Indrik. Nathan, G. S. 1986. Phonemes as mental categories. In Proceedings from the Twelfth Annual Meeting of the Berkeley Linguistics Society, V. Nikiforidou, M. Van Clay, M. Niepokuj & D.Â€Feder (eds.), 212–223. Berkeley CA: University of California at Berkeley. Nesset, T. 2000. Iconicity and prototypes: A new perspective on Russian verbs of motion. Scan‑ do-Slavica 46: 105–119. Padučeva, E. V. 2004. Dinamičeskie modeli v semantike leksiki. Moscow: Jazyki slavjanskoj kul’tury. Rakhilina, E. V. 2000. Kognitivnyj analiz predmetnyx imen: semantika i sočetaemost’. Moscow: Russkie slovari. Rakhilina, E. V. 2004. There and back: the case of Russian ‘go’. GLOSSOS 5: 1–33. . Rakhilina, E. V. 2007. Glagoly plavanija v russkom jazyke. In Glagoly dviženija v vode: LeksiÂ� českaja tipologija, T. A. Majsak & E. V. Rakhilina (eds), 267–285. Moscow: Indrik. Rosch, E. 1978. Principles of categorization. In Cognition and Categorization, E. Rosch & B. B. Lloyd (eds.), 27–48. Hillsdale NJ: Lawrence Erlbaum. Švedova, N. Ju. (ed.). 1980. Russkaja grammatika 1. Moscow: Nauka. Talmy, L. 1985. Lexicalization Patterns. In Language Typology and Syntactic Description 3, T.Â€Shopen (ed.), 57–149. Cambridge: CUP. Taylor, J. R. 1989. Linguistic Categorization. Oxford: OUP. Timberlake, A. 2004. A Reference Grammar of Russian. Cambridge: CUP.

Metaphorical walking 359

Wade, T. 1992. A Comprehensive Russian Grammar. Oxford: Basil Blackwell. Ward, D. 1965. The Russian Language Today. System and Anomaly. London: Hutchinson University Library. Wierzbicka, A. 1988. The Semantics of Grammar. Amsterdam: John Benjamins.

Appendix. Abbreviations * ? ?? aux neg

ungrammatical questionable phrasing extremely questionable phrasing auxiliary auxiliary

chapter 15

Russian verbs of motion Second language acquisition and cognitive linguistics perspectives* Kira Gor, Svetlana Cook, Vera Malyushenkova and Tatyana Vdovina University of Maryland

The results of three experiments comparing the processing of verbs of motion by late second language learners, American college students of Russian, and early starters, heritage speakers of Russian, are interpreted within the imageschematic framework developed in cognitive linguistics: the cross-linguistic typological approach introduced by Leonard Talmy (1985, 2000), the extension of this approach to Russian developed by Tore Nesset (2008), and the “thinking for speaking” hypothesis by Dan Slobin (1996). The results of the study support the claim that the system of verbs of motion is not fully acquired even in highly proficient second language learners. They typically lag behind not only native speakers, but also heritage speakers at the same proficiency levels.

1.

Introduction

Russian verbs of motion (VoM) is a notoriously thorny topic for late American learners of Russian as a second language (L2). While the more basic usages of * Authors are deeply indebted to Michael Long for guidance they received at every stage of the study, Scott Jackson for programming experiments with DMDX and advice in psycholinguistic methods, and Jennifer Koran for statistical analyses. . The term ‘late’ learner refers to post-pubescent learners who acquire the L2 after age 18. Early starters acquire L2 in early childhood. This distinction, widely used in SLA research, reflects the notion lying at the core of the Critical Period Hypothesis, namely, that adult learners, unlike young children, do not achieve native proficiency in the L2. If children are raised in a bilingual household or in a community that speaks a language different from the one used at home, it may be difficult to establish a difference between the L1 and L2. Heritage speakers represent a special case of early incomplete acquisition, since their L1, the language spoken at

362 Kira Gor et al.

VoM are typically acquired by intermediate learners, the more complex ones remain a challenge even for highly proficient learners. Do they present the same level of difficulty for early and late starters? A comparison of late L2 learners’ performance with that of heritage speakers of Russian at the same proficiency levels provides the answer to this question and takes the discussion one step further by identifying the possible factors leading to different learning outcomes/levels of control over VoM in perception and production. The results of three experiments comparing the processing of VoM by late L2 learners, American college students of Russian, and by early starters, heritage speakers of Russian, will be interpreted within the image-schematic framework developed in cognitive linguistics, including three separate approaches with their sets of claims that will be applied to the findings on Russian VoM: the cross-linguistic typological approach introduced by Leonard Talmy (1985, 2000), the extension of this approach to Russian developed by Tore Nesset (2008), and the ‘thinking for speaking’ hypothesis by Dan Slobin (1996). The claims of these approaches are evaluated in light of the results of an empirical study testing the perception and production of verbs of motion by highly proficient American learners of Russian. These experiments included an online Grammaticality Judgment Task with five different types of verbs of motion, a Sentence Completion experiment involving idiomatic expressions with the missing sentence-final verb of motion, and a Restricted Control experiment in which subjects heard idiomatic expressions and only the first syllable of the verb of motion in sentence-final position and were then required to produce the missing second part of the verb. All three tasks were part of a pilot data-collection battery developed to study the control of linguistic domains and discrete features in highly proficient American learners of Russian. This study addresses possible sources of difficulty in the use of verbs of motion and compares the performance of late second language (L2) learners and heritage learners of Russian at the same proficiency levels, with native speakers of Russian serving as a control group.

home, does not develop to native levels (Polinsky 2007). While there are some limitations on the generalizability of findings pertaining to heritage speakers to early L2 learners, there is an advantage to using comparisons of late L2 learners and heritage speakers. Since heritage speakers do not reach native language proficiency in their L1, they can be matched in proficiency level with late L2 learners, making it possible to investigate qualitative differences in their control of various linguistic aspects.

Russian verbs of motion 363

2.

VoM in L2 Russian

2.1

Typological perspective

This section will look at Russian verbs of motion as a linguistic domain to be acquired by L2 learners and make predictions about the problems that even highly proficient American learners may encounter while mastering it. Russian verbs of motion can be described from two different perspectives. First, they possess a unique distinction between unidirectional and multidirectional verbs of motion that pertains to a closed class of 14 pairs of basic verb stems, all imperfective, found in other Slavic languages. This is a widely accepted way of representing Russian VoM. And second, from a more recent cognitive linguistics perspective, Russian is typologically a satellite-framed language, in which manner is expressed by the verb root itself, while path is expressed by other elements associated with the verb root (Talmy 1985, 2000). The latest proposal integrates these two perspectives into one account using an image-schematic approach to Russian verbs of motion (Nesset 2008). The review below will compare Russian and English VoM while taking into account these different frameworks. According to the VoM-based language typology developed in cognitive linguistics (Berman & Slobin 1994; Slobin 1996, 2004; Talmy 1985, 2000), both Russian and English belong to the group of satellite-framed languages, in which the manner of motion (e.g., walking, running, or crawling) is encoded in the verbal root, while path (e.g., motion in and out of a space) is encoded by an element associated with the root. In English, this is a postfix, while in Russian this is a prefix. Conversely, in verb-framed languages, such as French, path is expressed by the verb root, and manner by an element associated with the verb. Unsurprisingly, verb-framed languages tend to avoid the heavy phrases required to encode manner, and thus manner remains unexpressed, as demonstrated in a cross-Â�linguistic study based on the Frog Story narratives by children and adults (Berman & Slobin 1994; Slobin 2004). The emergence of the owl from the hole in the tree elicited no manner verbs in verb-framed languages, which preferred to use the verb exit to describe the event. In satellite-framed languages, the same study showed, first, that the number of verbs encoding manner dramatically increased, and second, that there was a big disparity even among satellite-framed languages, with English speakers encoding manner in 32% of responses and Russian speakers in 100%. In other words, Russian mostly uses manner verbs with path prefixes, while English often uses the path verb come plus the postfix out to refer to the emergence of the owl. Since both English and Russian are satellite-framed languages, which differ in

364 Kira Gor et al.

the way they encode path, from a typological perspective one can expect certain processing difficulties associated with the expression of path and the appropriate use of prefixation in L2 learners of Russian that are L1 English speakers. Research on L2 acquisition and processing of VoM in cases where learners’ L1 and L2 are typologically different has produced certain evidence in support of Slobin’s ‘thinking for speaking’ hypothesis (Slobin 1996) and at the same time has demonstrated that the influence of the L1 on L2 acquisition of VoM (L1 transfer) is a complex multifaceted phenomenon (Cadierno 2004). According to Slobin, native speakers are trained by their linguistic experience to structure and verbalize motion events in a certain way, prioritizing and making explicit some aspects and details of the event and not others. Late L2 learners will then experience difficulties when reconfiguring the structure and segmentation of motion events and encoding them in L2-appropriate ways. The study by Cadierno compared the use of VoM in Frog Story written narratives by Spanish native speakers to L2 learners of Spanish whose L1 was Danish. Spanish is a verb-framed language, which expresses path in the verb and manner of motion by means of adverbial or gerundival phrases. Danish is a satellite-framed language, which expresses manner in the verb and path with the help of satellite verb particles. Two findings of the study are relevant for the present research. First, contrary to predictions based on previous findings (Berman & Slobin 1994; Slobin 1996), both Danish and Spanish speakers used the same number of verb types in their native language, and thus manner of motion was not more prominent in Danish, a satellite-framed language, than Spanish, a verb-framed language. At the same time, native speakers of Danish used fewer verb types in their L2 Spanish, which could be expected given their limited proficiency in the L2 (Intermediate Mid and Advanced on the OPI writing scale). Second, the use of satellites to encode path confirmed the expected pattern based on L1 transfer: Path was expressed more often in Danish than in Spanish as L1, and, accordingly, Intermediate L2 learners redundantly used directional adverbs in Spanish. Additionally, two more parameters, event conflation (i.e., expression of at least two locations referring to the source and goal of motion in the same clause) and elaboration of setting-description, while showing definitive language-specific patterns in Danish and Spanish as L1s, did not provide evidence of transfer from L1 Danish to L2 Spanish. In a follow-up study (Cadierno & Ruiz 2006), a group of Italian learners of Spanish was added to the groups of L1 Spanish speakers and L1 Danish learners of Spanish. Italian and Spanish, both Romance languages, are verb-framed, and the comparison of their treatment of VoM with Danish learners of Spanish was aimed at isolating L1 transfer of lexicalization patterns. The comparisons of the three groups showed that there were more similarities than differences between the two L2 groups, both of which differed considerably from the L1 Spanish group. Thus, the results

Russian verbs of motion 365

of the study lent limited support to the ‘thinking for speaking’ hypothesis for L2 learners rated as Advanced on the ACTFL writing proficiency scale.

2.2

Directionality in Russian verbs of motion

The generalized typological account outlined above fails to address the specifics of VoM in Russian, where the term is used in a narrow sense in reference to 14 pairs of verb stems, all imperfective, with one stem in the pair referring to unidirectional movement that typically has the source and the goal, and where the action occurs only once, and a multidirectional stem, which may express different types of movement, other than unidirectional (Muravyova 1980). Multidirectional verbs include roundtrips, or movement to a certain goal and then back to the source, random motion in different directions, and the ability to perform movement with a certain manner. Once a unidirectional stem acquires a path prefix (e.g., denoting movement in or out of a space), it turns into a perfective verb. Multidirectional verbs when adding a path prefix remain imperfective. The analysis of the closed class of paired verbs of motion evoking the notion of prototype, both at the level of form and meaning, has demonstrated that formal and functional prototypicality are relevant for this domain (Nesset 2000). According to Nesset, “In Russian, a non-directional verb of motion evinces more prototypical form and meaning than its unidirectional partner” (2000:â•›117). Multidirectional verbs are termed non-directional to emphasize the fact that it is the absence of directionality, their general core meaning that makes them prototypical. Unidirectional verbs are unique in that they express directionality and are therefore less prototypical. Formal properties of VoM match their meanings, with non-directional verbs using the regular default -aj- and the high-frequency productive -i- conjugational patterns, and unidirectional verbs using low-frequency unproductive -a-, -e-, and -ø- (zero-suffixed) conjugational patterns. Thus, nondirectional verbs use more common, or more prototypical, conjugational patterns than their unidirectional counterparts. Several considerations add to the complexity of the overall picture. First, based on the Russian National Corpus data, unidirectional verbs are five times more frequent than multidirectional in . Conjugational patterns are defined by the suffix, based on the one-stem verb system developed by Roman Jakobson (1948). . These observations can also be handled within markedness theory, with non-directional verbs being the unmarked member of the pair. Given that unidirectional verbs are less prototypical or more marked, according to the Markedness Differential Hypothesis (Eckman 1985), they are expected to present more difficulties for late L2 learners than non-directional (multidirectional) verbs.

366 Kira Gor et al.

the speech of L1 Russian speakers (Hasko forthcoming), despite the fact that they express fewer meanings. Second, unidirectional verbs can be argued to have a more prototypical meaning, as they uniformly refer to a motion in one direction with one or two endpoints implied. In contrast, multidirectional verbs have multiple meanings, which are likely to vary in the degree of prototypicality. Since path encoding is considered to be part of the image schema for VoM, the fact that unidirectional verbs encode path should contribute to their greater prototypicality. As a result, the existing tension between regularity, prototypicality (in Nesset’s terms), and frequency of use in paired VoM does not favor one unique set of predictions concerning the level of difficulty for their acquisition by late L2 learners and heritage speakers. In a recent account, Nesset incorporates the unidirectional/multidirectional distinction, which is formally expressed by the use of two different suffixes with the same verb root depending on directionality, into the cognitive linguistics framework and shows that it is reflected at the level of image schemas (2008). Thus, unprefixed unidirectional verbs encode path, while unprefixed multidirectional verbs do not. When path prefixes are used, “the semantic overlap between stem and prefix leads to the neutralization of the directionality contrast in prefixed motion verbs” (2008:â•›14). Additionally, Russian VoM systematically differentiate motion on foot (idtixodit’ [goUNI-goMULTI]) and by means of transportation (exat’-ezdit’ [goUNI-goMULTI]), and therefore, manner is an important characteristic of motion events, more so than in similar English motion expressions. However, idti [goUNI] also happens to be the main Russian generalized VoM, and is substituted for manner of motion verbs, e.g., letet’ [flyUNI] or plyt’ [swim/sailUNI] under specific conditions (Rakhilina 2004). This explains the fact that idti [goUNI] is by far the most frequent verb in the Russian National Corpus (Hasko forthcoming). Another property of Russian is that it uses prefixed verbs even when English uses simple bare stems to express motion with a clear source or goal, such as ‘come’ prixodit’/prijti [comeIPFV/comePFV] and ‘leave’ uxodit’/ujti [leaveIPFV/leavePFV]. Consequently, Russian speakers may be more prone to using prefixed VoM, when either the source or the goal is important than American learners. And finally, more generally, since English tends to encode path in postfixes and Russian in prefixes, one can predict that prefixation would be more used by Russian native speakers than by L2 learners of Russian. . Nesset (2008) includes 13 pairs of motion verbs in his analyses and uses the terms unidirectional and non-directional (for multidirectional) verbs. The only pair that uses suppletion is idti-xodit’ [goUNI-goMULTI], with different morphophonological processes taking place in other verb stems depending on the suffix used.

Russian verbs of motion 367

Given that English does not encode the uni/multidirectional distinction and has a different temporal and aspectual system, several predictions can be made with regard to the difficulties the Russian VoM will present for American learners. Learners will need to master the distinction between unidirectional and multidirectional stems, the meanings associated with them, the appropriate contexts for their use, and the correct aspectual choices for prefixed VoM. One of the most idiosyncratic elements in the overall system of Russian VoM is the use of multidirectional verbs with the meaning of roundtrips and their contrast with unidirectional verbs referring to motion in one direction only, which mirrors the more general use of the imperfective aspect for cancelled actions, as in (1) and (2). This contrast is not grammaticalized in English and, as a result, is notoriously problematic for English-speaking learners of Russian. (1) Ja ezdil na rabotu. I droveMULTI to work ‘I drove to work and back’. (2) Ja exal na rabotu. I droveUNI to work ‘I was driving to work.’

3.

VoM in heritage speakers of Russian

One of the goals of the present study is to establish whether late American learners of Russian process verbs of motion in the same way as heritage speakers of Russian at the same proficiency levels. Heritage speakers are characterized by incomplete acquisition (Polinsky 2007, 2008), which makes the comparisons both possible and informative. There are a few observations in the literature regarding heritage speakers’ use of verbs of motion. First, American heritage speakers of Russian overgeneralize the use of unidirectional verbs idti [goUNI] and pojti [goPFV] to the contexts that would require the use of ‘go’ in English, especially to those calling for roundtrips and therefore, the verbs xodit’ [goMULTI], sxodit’ [goPFV], etc. (Andrews 2000; Polinsky 2007; Zemskaja & Glovinskaja 2001). In other words, roundtrips appear to present considerable difficulties for heritage Russian speakers. Second, this preference for unidirectional verbs of motion applies only to the most generic stems idti and exat’ and does not involve all the paired VoM. Third, heritage speakers often choose one verb, unidirectional or multidirectional, and use it in all contexts. Fourth, it appears that the choice of VoM in American Russian interacts with aspectual usage, which for these heritage speakers is best accommodated by the Lexical Aspect Hypothesis developed in second language

368 Kira Gor et al.

acquisition research (see Bardovi-Harlig 2000) and applied to heritage learners (Pereltsvaig 2002). And finally, fifth, the observations found in the literature prompted the Prototype Hypothesis for heritage use of VoM: the choice of one ‘preferred’ verb stem to be used in contexts that call for both unidirectional and multidirectional verbs is dictated by considerations of greater prototypicality in heritage language use (Gor et al. 2009). In this understanding, prototypicality refers to the verbs that occur more consistently and frequently in heritage input and output with certain meanings, regardless of the form (inflectional pattern).

4.

The study

The three tests involving VoM reported below, GJT, Restricted Control, and Sentence Completion, were part of a comprehensive four-hour test battery developed to investigate the perception and production of a set of linguistic features and structures in highly proficient American learners of Russian as a part of the funded project Linguistic Correlates of Proficiency. Seventy paid volunteers participated in the experiment: 36 adult American learners of Russian, 24 heritage speakers of Russian living in the U.S., and 10 adult native Russian controls. The age of L2 learners was 21–56 years (mean = 32.1), heritage learners – 18–51 years (mean = 22.5), and native speakers – 20–54 years (mean = 36.7). Of all participants, 31 were male and 39 were female. Nonnative participants were pre-tested using the ILR oral proficiency interview format, and the two nonnative groups

. According to this hypothesis, “Aspectual marking in American Russian encodes the presence vs. absence of an inherent end-point associated with the verbal root: verbs that imply an inherent end-point are marked with the so-called PERFECTIVE morphology, whereas verbs that do not imply an inherent end-point are marked with the so-called IMPERFECTIVE morphology” (Pereltsvaig 2002:â•›9). . Heritage language use is expected to differ from standard language use, as the former typically involves colloquial language with little to no exposure to literary written language. . This study was funded by the Center for Advanced Study of Language at the University of Maryland. . In all three tests, stimulus material was recorded with two speakers, a male and a female, and two versions of the battery alternated the voices used in each test to avoid any possible speaker-related bias. Testing was computer-based, with each subject tested individually on Dell, Latitude/D820 computers with headsets and game pads in a quiet room. Sound recordings used in the study were digitized and processed using Praat software, and all the programming of the experiments was done with DMDX. Statistical analyses of the obtained results were performed with SAS.

Russian verbs of motion 369

were matched in proficiency levels. Both groups ranged from 1 to 4 on the ILR scale with most participants falling between ILR levels 2 to 3. The results obtained from the ten native controls were used to eliminate the items on which more than two native Russian subjects out of ten disagreed from further analyses.

4.1

Goals of the study and predictions tested

The reported study targeted both perception and production of unprefixed and prefixed verbs of motion by highly proficient late American learners and heritage speakers of Russian in a set of highly controlled experiments designed to elicit comparable responses in limited and fixed contexts. The first goal was to determine which verb types, unidirectional or multidirectional, presented more difficulties in perception and production, and whether different meanings expressed by multidirectional verbs, such as random or general motion as opposed to roundtrips, would generate different levels of difficulty. The second goal was to analyze problems arising when encoding path and manner of motion. And finally, the third goal was to identify possible differences in the treatment of any aspect of VoM between L2 learners and heritage speakers. Both Russian and English are satellite-framed languages, but they differ in several important ways (see above). Therefore, one cannot expect striking effects arising from L1 transfer, as in the case of L1 Danish and L2 Spanish, yet some tendencies can be hypothesized: e.g., the use of path prefixes, which is more common in Russian, may be more problematic for native speakers of English than Russian native speakers, along with the Slavic unidirectional/multidirectional distinction absent in English. Accordingly, the study tested the following predictions, some of them relevant for both perception and production and others for only one of the two modalities: 1. There will be significant differences in the treatment of motion verbs by late L2 learners and heritage speakers of Russian. 2. Roundtrips will present more difficulties for L2 learners, while unidirectional verbs will present more difficulties for heritage speakers at the same proficiency levels. 3. Prefixed verbs of motion will be more problematic than unprefixed ones and to a greater extent for L2 learners than heritage speakers because they encode path and involve aspectual choices. . The ILR (Interagency Language Roundtable) oral proficiency scale runs up to level 5, which corresponds to ‘highly articulate well-educated native speaker.’

370 Kira Gor et al.

4. Manner of motion will find limited expression in both heritage and L2 production: manner of motion verbs will be replaced by more generic verbs, such as xodit’, ezdit’ ‘go’. 5. Multidirectional verbs, which are more prototypical and use productive conjugational patterns, will be chosen over unidirectional verbs in production.

4.2

Grammaticality Judgment Task

The online Grammaticality Judgment Task (GJT) contained five blocks of ten sentences each (a total of 50 sentences), with five correct and five incorrect sentences per block: unidirectional unprefixed, unidirectional prefixed, multidirectional unprefixed expressing general or random motion, multidirectional prefixed with prefixes used as Aktionsart, and multidirectional prefixed roundtrips (see Table 1 for examples of each type of sentence with incorrect options provided in parentheses).10 All the sentences were presented randomly as part of a larger GJT, including several more linguistic features to avoid the serial effect. Sentences were presented binaurally through the headphones, and subjects had to press one of the two buttons on the game pad to indicate whether the sentence was correct or incorrect. Combined accuracy scores for all the five types of VoM are provided in FigureÂ€1, which contains mean accuracy scores for the Heritage and L2 groups broken down by ILR oral proficiency levels, with the score of 0.450 corresponding to 45 percent accuracy, etc. The scores line up perfectly with the OPI levels, which indicates that perceptual knowledge of VoM, as assessed by GJT, is acquired incrementally in a linear fashion. Group mean scores show a perfect Spearman correlation of 1.00 with ILR proficiency levels and plus sublevels. The next step in data analysis addressed two issues: whether the two groups, Heritage and L2, differed in their performance on VoM in GJT in general, and which types of VoM were the most difficult for each of the groups. Figure 2 represents accuracy scores for the two groups separately and for each of the five VoM types. It demonstrates that (a) overall, the Heritage group slightly outperformed the L2 group, (b) all three prefixed conditions were more difficult than unprefixed ones for the L2 group, but only one of them was more difficult for the Heritage group (i.e., with Aktionsart prefixes on multidirectional stems) and (c) prefixed roundtrips were the most difficult for the L2 group, but not for the Heritage group. The Heritage and L2 groups were further subdivided into high-proficiency (ILR 2+ and above) and low-proficiency (ILR 2 and below) subgroups, and these 10. All the examples provided in the article are very similar, but not identical to the actual sentences in the tests. Those changes were made to protect the tests, which are still in use.

Russian verbs of motion 371

Table 1.â•‡ Grammaticality Judgment Task: Testing material Type of motion verb

Number of sentences

Examples†

Unidirectional 5 Correct Unprefixed 5 Incorrect (One-way motion)

Smotri! Anton opjat’ kuda-to bežit (*begaet). look Anton again someplace runsUNI (*runsMULTI) ‘Look! Anton is running (*running around) someplace.’

Unidirectional Prefixed

5 Correct 5 Incorrect

Matrosu prikazali bystro za-lezt’ sailorDAT ordered.3.PL quickly on-climbUNI (*za-lezat’) na mačtu. (*on-climbMULTI) on mast ‘The sailor was ordered to climb the mast right away.’

Multidirectional Unprefixed (General/random motion)

5 Correct 5 Incorrect

Moj drug letal (*letel) v Pariž na prošloj my friend flewMULTI (*flewUNI) to Paris on last nedele. week ‘My friend flew (*was on his way) to Paris last week.’

Multidirectional Prefixed

5 Correct 5 Incorrect

My is-xodili (*izo-šli) we through-walkedMULTI (*through-walkedUNI) ves’ les, no gribov ne našli. all forest but mushrooms NEG found ‘We walked around all the forest, but did not find any mushrooms.’

Multidirectional Prefixed (Perfective Roundtrip)

5 Correct 5 Incorrect

Maša, nado s-begat’ (*s-bežat’) Maša necessary roundrip-runMULTI (*away-runUNI) v magazin za xlebom. to â•›store for bread ‘Masha, (it is necessary to) go buy some bread.’

†

Incorrect options are provided in parentheses and marked with an asterisk. The actual test items are similar, but not identical to the examples offered here.

two subject groups and two proficiency-based subgroups were used in two-way ANOVA analyses for all the tests in the study. A comparison of overall accuracy scores in the Heritage and L2 groups in a repeated measures ANOVA using an alpha level of 0.05 found a statistically significant difference between the two groups of subjects, F(1, 56) = 13.74, pÂ€<Â€0.01, and a significant difference between the high- and low-proficiency levels, F(1,Â€56)Â€= 48.29, p < 0.01. Statistically significant within-subjects effects for the type of VoM, F(4, 224) = 5.25, p < 0.01, and interaction between the group and type of VoM, F(4, 224) = 5.58, p < 0.01, were also found. Scheffe’s test of paired comparisons revealed significant differences between the two groups

372 Kira Gor et al.

0.9

0.8

Mean Accuracy Score

0.7

0.6

0.5

0.4

0.3

0.2

0.1

0 1

1+

2

2+

3

3+

4

ILR Level

Figure 1.â•‡ Grammaticality Judgment Task: Mean accuracy scores for L2 learners and heritage speakers

Figure 2.â•‡ Grammaticality Judgment Task: Accuracy scores for each type of verb of motion

Russian verbs of motion 373

80 70 60 50

Heritage High Heritage Low L2 High L2 Low

40 30 20 10 0

Uni

Multi Types of VoM

Figure 3.â•‡ Grammaticality Judgment Task: Responses to incorrect unprefixed verbs of motion

for Unidirectional Unprefixed, Unidirectional Prefixed, and (Multidirectional) Roundtrip Prefixed VoM types. Finally, a separate analysis was conducted to test the hypothesis that lowerproficiency heritage speakers would experience more difficulties with unprefixed unidirectional VoM than L2 learners, i.e., that they would be more willing to accept unidirectional verbs in contexts calling for roundtrips. Each of the two groups was divided into high-proficiency (2+ and above) and low-proficiency (2Â€and below) subgroups, and only accuracy scores for unprefixed unidirectional and multidirectional verbs were analyzed. In each of the two VoM types, responses to correct sentences (requiring a “yes”) and incorrect sentences (requiring a “no”) were examined separately, and it was established that, while heritage accuracy scores were consistently higher than those of L2 learners, there was one single case (i.e., expected “no” responses to multidirectional verbs) where this consistent tendency was reversed. In this case, unidirectional verbs were incorrectly substituted for multidirectional ones in stimuli sentences. Figure 3 demonstrates that low-proficiency heritage speakers were more willing than L2 learners to accept the incorrect use of unidirectional verbs instead of multidirectional ones. To summarize the results obtained in GJT, the group means showed a perfect rank-order correlation with the ILR OPI levels and sublevels, when heritage and L2 learners’ data were combined together. At the same time, significant differences were established between overall Heritage and L2 accuracy scores. When the accuracy scores of two groups were examined more closely for each VoM type, it was found that the groups exhibit similar performance on some types, and significant differences on others. First, L2 learners consistently demonstrate lower accuracy on prefixed than on unprefixed VoM, while heritage speakers do not show such a tendency. If we assume that heritage speakers exposed to Russian as an L1 in early childhood developed more native-like cognitive mechanisms

374 Kira Gor et al.

for dealing with path expression, while late American learners had to master the Russian way of systematically encoding path and in prefixes rather than postfixes, the results support the cognitive typological model of VoM developed by Talmy (1985, 2000) and indicate that heritage speakers may possess an early acquired native-like cognitive structure. Section 5 will address the issue of whether the data lend support to the ‘thinking for speaking’ hypothesis proposed by Slobin (1996). Second, according to the study, even relatively proficient L2 learners experience difficulties with perfective roundtrips, with accuracy rates lagging behind heritage speakers. Again, roundtrips do not find simple straightforward expression in English and are notoriously problematic for American learners. And finally, lower-proficiency heritage speakers showed a tendency to accept unidirectional unprefixed verbs where the context required the use of a multidirectional one. This tendency supports observations in the literature referring to low-proficiency heritage use (Andrews 2000; Pereltsvaig 2002).

4.3

Restricted Control Task

Restricted Control is a production task with very limited verbal output required from the test-taker. Subjects hear the beginning of an idiom containing a VoM in sentence-final position and the first syllable of the VoM itself, and are asked to provide the missing part of the verb. Advantages of this test are the relative ease of data collection and scoring, and a highly controlled context. The main concern is the fact that the test measured idiom knowledge, as well as knowledge of VoM, and that idiom knowledge could act as a confounding variable. This concern is addressed in the next section on Sentence Completion. Thirty sentences containing idiomatic expressions with verbs of motion were constructed for this task. Overall, the testing material included 20 prefixed verbs, both perfective and imperfective, and 10 unprefixed imperfective verbs of motion. Sentences with the first syllable of the target VoM were presented aurally through the headsets, and subjects were instructed to say the remainder of the verb into the microphone. All the responses were transcribed and used for further analyses. An example of the idioms included in the testing material is provided in (3). (3) Soldat spit – služba idet. soldier sleeps service goesUNI ‘No matter what a soldier may be doing, he is fulfilling his term of service.’

A one-way ANOVA on the three groups of subjects, Heritage, L2, and Native, with a Tukey post hoc comparison, using an alpha level of 0.05 demonstrated

Russian verbs of motion 375

significant differences, F(2, 66) = 60.05, p < 0.01. A two-way ANOVA on two groups, Heritage and L2, and two proficiency levels, high and low, demonstrated significant differences in accuracy scores for group, F(1, 55) = 65.18, p < 0.01, and proficiency level, F(1, 55) = 22.14, p < 0.01. Figure 4 represents accuracy scores of the two groups of subjects, Heritage and L2, broken down by two proficiency levels, high and low, based on their ILR OPI scores. It shows that the Heritage group outperformed the L2 group at both proficiency levels. In other words, L2 learners at the same proficiency levels as heritage speakers exhibited lower accuracy scores on VoM in this task. Finally, the analysis of errors in the choice of VoM by heritage and L2 learners revealed a similar tendency in choosing incorrect direction in both groups. They preferred to substitute multidirectional stems for unidirectional ones, and not unidirectional for multidirectional ones (see FigureÂ€5). Given that the majority of VoM in this task were prefixed (24 prefixed and 6 unprefixed), the preference for multidirectional stems may be due to their prototypicality (Nesset 2000). 90 80 Percent Accuracy

70 60 50

High Low

40 30 20 10 0

Heritage

L2

Figure 4.â•‡ Restricted Control: Accuracy scores of high and low proficiency heritage speakers and L2 learners

Response Rate (Percent)

120 100 80 Uni instead of Multi Multi instead of Uni

60 40 20 0

Heritage

Figure 5.â•‡ Restricted Control: Errors in direction

L2

376 Kira Gor et al.

To summarize the results of the Restricted Control Task, accuracy scores on VoM in Heritage, L2, and Native groups, as well as high- and low-proficiency Heritage and L2 subgroups, differed significantly, with L2 learners typically lagging behind heritage speakers. When heritage and L2 learners made errors in direction, substitutions predominantly followed one direction: multidirectional verb stems were substituted for unidirectional ones.

4.4

Sentence Completion Task

Sentence Completion differed from Restricted Control in that subjects were required to provide the entire VoM in sentence-final position. A total of 45 sentences containing idiomatic expressions with verbs of motion were constructed for the testing material (see (4) for an example of the test item). (4) Bityj nebitogo vezet. injured noninjured transporting/drivingUNI ‘There are some who really know how to take advantage of others’.

Overall, the testing material included 12 fixed and 33 loose idioms. Out of the total of 45 idioms, there were 25 prefixed VoM with 19 unidirectional (perfective) and six multidirectional (imperfective) stems, which designated general motion. In addition, 20 unprefixed verbs of motion were included, with 13 multidirectional (imperfective) verbs expressing general or random motion and roundtrips, and seven unidirectional (imperfective) verbs. All the stimulus sentences were presented aurally through the headsets, and subjects were asked to produce the last missing verb into the microphone. All responses were subsequently transcribed and analyzed. A one-way ANOVA on three groups of subjects, Heritage, L2, and Native, using an alpha level of 0.05 demonstrated significant differences, F(2, 65) = 57.46, pÂ€< 0.01. Post hoc Tukey comparisons indicated that all three groups differed from each other. A two-way ANOVA on two groups, Heritage and L2, and two proficiency levels, high and low, demonstrated significant differences in accuracy scores for group, F(1, 57) = 21.94, p < 0.01, and proficiency level, F(1, 57) = 25.53, p < 0.01. As can be seen in Figure 6, heritage speakers outperformed L2 learners in both high-proficiency and low-proficiency groups, and high-proficiency subjects were significantly ahead of low-proficiency subjects. Error analysis on Sentence Completion revealed one trend in the pattern of substitutions: generic verb roots expressing the most general kind of motion, such as idti [goUNI] and nosit’ [carry MULTI] were substituted for manner of motion roots such as lezt’ [climbUNI]

Russian verbs of motion 377

45 40 Percent Accuracy

35 30 25

High Low

20 15 10 5 0

Heritage

L2

Figure 6.â•‡ Sentence Completion: Accuracy scores of high-proficiency and low-proficiency heritage speakers and L2 learners

70

Percent Substitutions

60 50 40

Generic Manner of Motion

30 20 10 0

Heritage

L2

Figure 7.â•‡ Sentence Completion: Substitutions of manner of motion verbs

and taščit’ [haulUNI] more often than manner of motion verbs were substituted for generic in both Heritage and L2 groups (see Figure 7). Moreover, when these rare substitutions of manner of motion verbs for more generic verbs took place, Heritage speakers in 50% of responses used the same verb lezt’ [climbUNI], and L2 learners in 57% of responses used two verbs, bežat’ [runUNI] and letat’ [fly MULTI]. Given that both Russian and English are satellite-framed languages, the use of manner of motion verbs was not a cognitive issue for L2 learners; however, one can expect non-native speakers to operate with a reduced vocabulary of motion verbs, which is a general tendency in L2 vocabulary use. And indeed, both the Heritage and L2 groups reverted to generic VoM when more specific manner of motion verbs were required, leaving manner underspecified. The use of idioms as the testing material raises the issue of whether the test actually measures knowledge of VoM or merely memorization of individual idioms. The following step was taken to justify the use of idiomatic expressions to elicit VoM in two production tasks, Restricted Control and Sentence Completion.

378 Kira Gor et al.

We compared the performance of our three groups of subjects, Heritage, L2, and Native, on Restricted Control and Sentence Completion Tasks with the comparable task, Collocation Completion, which exclusively measured idiom knowledge and did not contain any VoM. Collocation Completion closely mirrored Sentence Completion in that subjects were required to provide the missing last word of an idiomatic expression. A one-way ANOVA on accuracy rates for the Heritage, L2, and Native groups using an alpha level of 0.05 demonstrated significant differences in performance for the three groups, F(2, 64) = 11.23, p < 0.01, and for the two proficiency levels F(1,Â€64)Â€= 25.88, p < 0.01. However, post hoc Tukey comparisons showed no significant difference in performance for the Heritage and L2 groups, while similar analyses demonstrated significant differences for the Heritage and L2 groups both on Restricted Control and Sentence Completion. This indicates that the differences were not due to idiom knowledge per se, but rather reflected the processing of VoM. To summarize, Sentence Completion, a highly controlled production task, produced higher accuracy scores in the Heritage than L2 groups and in both, the high and low proficiency ranges. The analysis of errors established that both groups tended to substitute generic VoM for manner of motion verbs; thus, manner expression was limited. Remarkably, heritage speakers used no more manner of motion verbs in their substitutions than L2 learners.

5.

Discussion and conclusions

It is a well-known fact that Russian VoM are difficult to master for American learners, and accordingly, a great deal of emphasis is placed on teaching VoM in an intermediate-level classroom (corresponding to ILR 1) with the assumption that while complex, they developmentally belong to the Intermediate-Advanced proficiency range. And indeed, the unidirectional/multidirectional distinction, as well as basic functions of prefixed VoM are typically acquired at the Advanced (ILR 2) level. Does this mean that highly proficient learners of Russian process VoM in a native-like way? This was the question that motivated the present study. The study was designed to (1) target highly proficient late L2 learners and heritage speakers of Russian, and (2) gauge the perception and production of VoM in a set of highly controlled experiments. This type of data collection is an alternative to the widely used oral elicitation technique making use of guided narratives, such as the Frog Story, and as such, provides additional insights into high-proficiency processing of VoM. First, it demonstrated that even highly proficient American learners of Russian do not process VoM in a native-like way.

Russian verbs of motion 379

Second, it compared perception and limited production of VoM by three groups of subjects, with heritage and native speakers, in addition to L2 learners, and identified similarities and differences in their performance. The following general results were obtained: 1. Both the Heritage and L2 groups experienced difficulties with basic uses of VoM until very high proficiency levels. The degree of difficulty varied depending on the type of VoM. 2. Neither the Heritage nor the L2 groups ever attained the accuracy rates of Native controls on either of the two tasks involving idiomatic use of VoM. 3. Heritage speakers consistently outperformed L2 learners on all the tasks, and most types of VoM within each task, which is a particularly interesting finding given that the Heritage and L2 groups were matched in proficiency and age. The study confirmed most of the predictions put forth above, with each of them addressed below. 1. As mentioned above, there were statistically significant differences between the Heritage and L2 groups, with heritage speakers producing higher accuracy scores on most, but not all VoM types. 2. On the Grammaticality Judgment Task, prefixed roundtrips were most problematic for L2 learners, while unidirectional unprefixed VoM were most problematic for heritage speakers. 3. On the Grammaticality Judgment Task, L2 learners produced lower accuracy scores on prefixed types of VoM than on unprefixed types. Heritage speakers experienced difficulties only with multidirectional verbs using Aktionsart prefixes. 4. On the Sentence Completion Task, both the Heritage and L2 groups preferred to substitute generic VoM for manner of motion verbs, and consequently, manner of motion was not adequately expressed. 5. On the Restricted Control Task, the analysis of directionality in substitutions established that both the Heritage and L2 groups preferred to use multidirectional stems instead of unidirectional stems. Overall, the results of the study support the claim that the system of VoM is not fully acquired in even highly proficient L2 learners. They typically lag behind not only native speakers, but also heritage speakers at the same proficiency levels. The problems encountered by L2 learners in this study can be traced back to the linguistic and cognitive challenges VoM present on several levels: expression of directionality specific for Slavic languages, and expression of path and manner

380 Kira Gor et al.

of motion from a typological perspective. Comparisons of heritage speakers at the same proficiency levels as L2 learners helped to isolate an L2-specific set of difficulties. Expected problems with directionality, and perfective roundtrips in particular, tend to persist even in highly proficient L2 learners. And indeed, L2 learners, but not heritage speakers found prefixed VoM more difficult to process. Apparently, while both Russian and English are satellite-framed languages, in Russian, path expression and the use of prefixes are more obligatory, especially since English mostly uses postfixes to encode path. Heritage speakers seem to develop more native-like linguistic and cognitive strategies in path expression. At the same time, manner found weaker expression both in the L2 and Heritage groups, most likely due to limited availability of manner of motion verbs or problems accessing them in non-native mental lexicon. One feature recorded in heritage speaker performance was not found in L2 learners: eagerness to accept unprefixed unidirectional VoM in place of multidirectional ones. In this case, the effect does not reflect parallelism with the general-purpose use of English ‘go,’ since no idti [goUNI] items were included in the testing material. The likely explanation is the influence of higher type and token frequency characteristic of unidirectional verbs. If this is indeed the case, heritage speakers show an inputdriven tendency compatible with early implicit exposure to VoM. And finally, both the L2 and Heritage groups preferred to substitute multidirectional stems for unidirectional, which supports the role of prototypicality in processing Russian VoM (Nesset 2000). Therefore, the study confirms that cross-linguistic differences identified in VoM typology exert a certain influence on processing of VoM even in highly proficient learners. Since the study used perception and limited production data and did not analyze samples of free speech, its conclusions cannot be used in direct support of the ‘thinking for speaking’ hypothesis (Slobin 1996), but they are definitely not in conflict with it.

References Andrews, D. R. 2000. Heritage learners in the Russian classroom: Where linguistics can help. ADFL Bulletin 3(3): 39–44. Bardovi-Harlig, K. 2000. Tense and Aspect in Second Language Acquisition: Form, Meaning, and Use. Malden MA: Blackwell. Berman, R. A. & Slobin, D. I. (ed.). 1994. Relating Events in Narrative: A Crosslinguistic Devel‑ opmental Study. Hillsdale NJ: Lawrence Erlbaum. Cadierno, T. 2004. Expressing motion events in a second language: A cognitive typological perspective. In Cognitive Linguistics, Second Language Acquisition, and Foreign Language Teaching, M. Achard & S. Niemeier (eds.), 13–49. Berlin: Mouton de Gruyter. Cadierno, T. & Ruiz, L. 2006. Motion events in Spanish L2 acquisition. Annual Review of Cogni‑ tive Linguistics 4: 183–216.

Russian verbs of motion 381

Eckman, F. R. 1985. Some theoretical and pedagogical implications of the markedness differential hypothesis. Studies in Second Language Acquisition 7(3): 289–307. Gor, K., Cook, S. V., Malyushenkova, V. & Vdovina, T. 2009. Verbs of motion in highly proficient learners and heritage speakers of Russian. In Special Forum on Teaching and Learn‑ ing Russian Verbs of Motion, V. Hasko (ed.), Slavic and Eastern European Journal 53(3), 386–408. Hasko, V. Forthcoming. The role of thinking for speaking in adult L2 speech: The case of (non)unidirectionality encoding by American learners of Russian. In Linguistic Relativity in Sec‑ ond Language Acquisition: Evidence of First Language Thinking for Speaking, Z.-H. Han & T. Cadierno (eds.). Clevedon: Multilingual Matters. Jakobson, R. 1948. Russian conjugation. Word 4(3): 155–167. Muravyova, L. S. 1980. Verbs of Motion in Russian. Moscow: Russian Language Publishers. Nesset, T. 2000. Iconicity and prototypes: A new perspective on Russian verbs of motion. Scando-Â�Slavica 46(1): 105–119. Nesset, T. 2008. Path and Manner: An image-schematic approach to Russian verbs of motion. Scando-Slavica 54(1): 135–158. Pereltsvaig, A. 2002. Aspect lost, aspect regained: restructuring of aspectual marking in American Russian. Paper presented at the NSF Workshop on Syntax, Semantics and Acquisition of Aspect, University of Iowa. Polinsky, M. 2007. Reaching the end point and stopping midway: Different scenarios in the acquisition of Russian. Russian Linguistics 31(2): 157–199. Polinsky, M. 2008. Without aspect. In Case and Grammatical Relations: Studies in Honor of Ber‑ nard Comrie, G. Corbett, & M. Noonan (eds.), 263–282. Amsterdam: John Benjamins. Rakhilina, E. V. 2004. There and back: The case of Russian ‘go’. GLOSSOS 5: 1–33. . Slobin, D. I. 1996. From ‘thought and language’ to ‘thinking for speaking’. In Rethinking Linguis‑ tic Relativity, J. Gumperz & S. C. Levinson (eds.), 70–96. Cambridge: CUP. Slobin, D. I. 2004. The many ways to search for a frog: Linguistic typology and the expression of motion events. In Relating Events in Narrative: Typological and Contextual Perspectives, S.Â€Strömqvist & L. Verhoeven (eds.), 219–257. Mahwah NJ: Lawrence Erlbaum. Talmy, L. 1985. Lexicalization patterns: Semantic structure in lexical forms. In Language Typol‑ ogy and Syntactic Description, Vol. 3: Grammatical Categories and the Lexicon, T. Shopen (ed.), 36–140. Cambridge: CUP. Talmy, L. 2000. Toward a Cognitive Semantics, Vol. II: Typology and Process in Concept Structur‑ ing. Cambridge MA: The MIT Press. Zemskaja, E. A. & Glovinskaja, M. Ya. (ed.). 2001. Jazyk russkogo zarubežja: Obščie processy i rečevye portrety. Moscow: Jazyki slavjanskoj kul’tury.

Appendix. Abbreviations IPFV MULTI PFV UNI

imperfective multidirectional perfective unidirectional

Author index

A Adamec, P.â•… 16–17, 25, 27 Aitzetmüller, R.â•… 51 Alekseev, A. A.â•… 51 Andrews, D. R.â•… 367, 374 Anstatt, T.â•… 74, 80–81 Anthony, D. W.â•… 119 Apresjan, J. D.â•… 292, 304 Arkad’ev, P.â•… 336 Asher, N.â•… 147 Aske, J.â•… 200, 227, 248, 251, 287 Atkins, B. T. S.â•… 292, 308 Aurnague, M.â•… 1, 199 B Bardovi-Harlig, K.â•… 368 Bavin, E.â•… 2, 200 Beavers, J.â•… 227, 267 Belichova-Krzhizhkova, E.â•… 270 Bennett, M.â•… 145 Berman, R.â•… 1, 197, 199–201, 226, 363–364 Bermel, N.â•… 91 Bezlaj, F.â•… 116 Biber, D.â•… 206 Binnick, R. I.â•… 206 Bjeletić, M.â•… 118 Blazhev, B. I.â•… 283–284 Bogoljubova, N. D.â•… 16, 18 Bohnemeyer, J.â•… 250, 267 Bojar, B.â•… 228, 292, 299 Bondarko, A. V.â•… 127 Borik, O.â•… 143, 145 Borschev, V.â•… 174 Bowerman, M.â•… 1, 257 Browne, W.â•… 126 Bubenik, V.â•… 114 Bulatova, R. V.â•… 59, 96, 98–99 Bussmann, H.â•… 205 Bybee, J. L.â•… 93, 310

C Cadierno, T.â•… 1–2, 364 Carter, R.â•… 272 Cejtlin, R. M.â•… 114 Češko, E.â•… 67, 74–75, 89–90 Chertkova, M.â•… 127, 205 Choi, S.â•… 1, 257 Cienki, A. J.â•… 203, 270 Clark, E. V.â•… 235 Clark, H. H.â•… 167, 235 Comrie, B.â•… 144, 206, 268 Cook, S.â•… 9 Creswell, J. W.â•… 209 Croft, W.â•… 129, 272 Crowley, T.â•… 116 Cruse, D. A.â•… 129 Cubberley, P.â•… 2, 207 Cummins, G.â•… 69 D Darden, B. J.â•… 48–49, 54, 59 Dasher, R. B.â•… 297 Dickey, S. M.â•… 4–5, 58, 63, 69–70, 73, 85, 100, 112–115, 118, 128, 206, 262 Divjak, D.â•… 8, 202, 310, 315–316, 321 Dixon, R. M. W.â•… 119 Dodge, E.â•… 344 Doluxanov, P. M.â•… 119 Dostál, A.â•… 67, 68, 74, 84, 89, 90, 102 Dowty, D. R.â•… 146 Driagina, V.â•… 209, 225 Dukhovny, E.â•… 339 Dyer, D. L.â•… 15 E Eckman, F. R.â•… 365 Evgen’eva, A. P.â•… 292, 294, 299

F Feldman, J. A.â•… 349 Ferm, L.â•… 204 Ferrell, J.â•… 104 Filip, H.â•… 143, 145 Filipović, L.â•… 7, 9, 69–70, 97–98, 197, 202, 204, 215, 227–228, 230, 248, 251, 255–256, 258– 260, 262–264 Fillmore, C. J.â•… 166, 292, 308 Foote, I. P.â•… 207, 344 Forsyth, J.â•… 3, 5, 128, 141–142, 144–147, 149–150, 152, 154, 157, 166, 178–179, 186–187, 189, 344 Fougeron, I.â•… 37 Frawley, W.â•… 230 G Ganenkov, D.â•… 336 Garde, P.â•… 346 Gasparov, B.â•… 262 Gazdar, G.â•… 167 Geeraerts, D.â•… 129, 137, 346 Gerd, A. S.â•… 51 Gindin, L. A.â•… 119 Gippius, A. A.â•… 25 Glovinskaja, M. Ya.â•… 367 Goldberg, A. E.â•… 166, 272 Golubkova, E.â•… 316, 321 Gorbachov, Y.â•… 101 Goretskii, V. G.â•… 204 Gor, K.â•… 9, 368 Greenberg, M. L.â•… 4–5, 49, 58, 67, 69–70, 90, 93, 100, 115 Grefenstette, G.â•… 164 Grenoble, L. A.â•… 16–17 Grice, H. P.â•… 157 Grickat, I.â•… 262 Grochowski, M.â•… 228 Gronn, A.â•… 144

384 New Approaches to Slavic Verbs of Motion

Gruntova, E.â•… 336 Gumperz, J.â•… 1 Gvozdev, A. N.â•… 203 H Han, Z.-H.â•… 1 Hasko, V.â•… 2, 6–7, 9, 201, 217, 225, 247, 366 Heim, I.â•… 153 Herrity, P.â•… 58 Hickmann, M.â•… 1–2, 197, 199 Hodgson, W. H.â•… 150 Horn, L. R.â•… 149, 167, 174 Horrocks, G.â•… 67, 69–73, 76, 95, 98, 103, 227 Huntley, D.â•… 102–103 I Ikegami, Y.â•… 230 Inagaki, S.â•… 2, 200 Isachenko, A. V.â•… 3, 5, 126, 128, 141, 178, 207, 346, 353–354 Israeli, A.â•… 203, 284 Ivanova, K.â•… 96, 98 J Jakobson, R.â•… 73, 147, 157, 232, 344, 365 Jakulis, E.â•… 113 Janda, L. A.â•… 5, 125, 127, 134–135, 166, 203, 343 Janko, T. E.â•… 17, 25, 32, 43 Johannet, J.â•… 80 Johnson-Laird, P. N.â•… 1, 230, 253 Johnson, M.â•… 1, 10, 131, 265, 349, 356 Jones, M. A.â•… 287 Joseph, B. D.â•… 116 Jovanovic, J.â•… 227–228, 339 K Kagan, O.â•… 5–6, 143, 145–146, 152–153, 156–157, 161 Kardaševskij, S. M.â•… 22 Kaushanskaya, M.â•… 339 Kay, P.â•… 166 Keller, F.â•… 165 Kentfield, M.â•… 339 Kibrik, A. E.â•… 311 Kiparsky, V.â•… 51

Kita, S.â•… 2, 200 Klein, W.â•… 152 Kopecka, A.â•… 7, 200, 202, 204, 215, 230 Koptjevskaja-Tamm, M.â•… 8, 202, 292, 315–316 Kortlandt, F.â•… 62 Kovtunova, I. I.â•… 18, 25, 27 Kramarova, S.â•… 323 Krifka, M.â•… 94, 143, 146 Krugljakova, V.â•… 311 Kučera, H. â•… 74 Kurylowicz, J.â•… 48 Kustova, G. I.â•… 297 Kuteva, T.â•… 234 Kuznecova, A.â•… 104 L Lakoff, G.â•… 129–130, 310, 344, 346, 349, 356 Lambrecht, K.â•… 167 Lander, Yu.â•… 323 Langacker, R. W.â•… 293, 346, 348 Lapata, M.â•… 165 Lapteva, O. A.â•… 29 Launer, M.â•… 203 Lemmens, M.â•… 316, 321 Leskien, A.â•… 85 Levin, B.â•… 15, 90, 92, 227, 230, 260, 272 Levinson, S.â•… 1 Levontina, I. B.â•… 201 Lewandowska-Tomaszczyk,Â€B. â•… 129 Lewis, W. D.â•… 167, 189, 225 Lindsey, T.â•… 59 Lindstedt, J.â•… 51 Luraghi, S.â•… 282 L’vov, M. P.â•… 204 Lyons, J.â•… 253 M Machek, V.â•… 48 Mahota, W.â•… 207, 344 Majsak, T. A.â•… 292, 344, 354 Malyushenkova, V.â•… 9 Martinovic-Zic, A.â•… 227–228 Maslova, E.â•… 310 Matešić, J.â•… 56 Matushansky, O.â•… 203 Mayer, M.â•… 209

McCormick, M.â•… 118 McNeill, D.â•… 2, 200–201, 203 Meillet, A.â•… 52–53 Merrill, P.â•… 178 Meyer, C. F.â•… 164–165 Mietzner, A.â•… 197 Miller, G. A.â•… 1, 230, 253 Morreau, M.â•… 147 Murav’eva, L. S.â•… 204, 365 N Naigles, L. R.â•… 200, 259 Narasimhan, B.â•… 225, 230, 240 Nathan, G. S.â•… 346 Nedashkivska, A.â•… 270 Nesset, T.â•… 9, 126, 353, 355, 361–363, 365–366, 375, 380 Nikitina, T.â•… 8, 203, 282–283, 287 Němec, I.â•… 67 Nørgård–Sørensen, J.â•… 91 O O’Connor, L.â•… 216, 230 Özyürek, A.â•… 2, 200 P Padučeva, E. V.â•… 143–144, 153–154, 156, 159, 172, 174, 297, 356 Pagliuca, W.â•… 93 Pahomov, G.â•… 201 Partee, B. H.â•… 145, 174 Pascal, B.â•… 1 Pavlenko, A.â•… 209 Peirsman, Y.â•… 137 Perelmutter, R.â•… 6, 81, 174–175, 184 Pereltsvaig, A.â•… 368, 374 Perkins, R.â•… 93 Pokorny, J.â•… 47 Polinsky, M.â•… 362, 367 Prokof ’eva, I.â•… 291, 292, 311, 332 R Radden, G.â•… 90, 99 Rakhilina, E. V.â•… 8, 202, 291– 292, 301, 308, 310–311, 315–316, 321–322, 336–337, 343–344, 347, 349–354, 366

Rapoport, T. R.â•… 272 Rappaport Hovav, M.â•… 15, 90, 92, 227, 272 Ricca, D.â•… 339 Rix, H.â•… 113, 115–116 Robert, S.â•… 1–2, 197, 199 Romanova, E.â•… 141 Rosch, E.â•… 346 Ruiz, L.â•… 364 S Sadnik, L.â•… 51 Saeed, J. I.â•… 253 Šaxmatov, A. A.â•… 127 Sauerland, U.â•… 152–153 Schawaller, D.â•… 75 Schenker, A. M.â•… 3, 48, 64 Schuyt, R.â•… 67 Seibert, U.â•… 1–2, 200 Šemanaeva, O.â•… 337 Shay, E.â•… 1–2, 200 Shilova, V. V.â•… 197 Shmelev, A. D.â•… 5, 126–128, 141, 143, 166, 201, 205 Shull, S.â•… 256, 261 Šigurov, V. V.â•… 179 Sinha, C.â•… 234 Slobin, D.â•… 1–3, 62, 69–70, 90, 111–112, 163, 166, 197–201, 203–204, 217, 219, 225–226, 229–230, 235, 237, 239, 248, 251–252, 267–268, 311, 339– 340, 361–364, 374, 380 Smith, C.â•… 92, 143–144, 149, 152 Smyth, H. W.â•… 282 Snoj, M.â•… 114, 116

Author index 385

Sobol, E.â•… 313 Sosnovskaia, O. V.â•… 204 Stang, C.â•… 48 Staniševa, D. 104, 269 Stavrou, M.â•… 69–73, 76, 95, 98, 227 Stern, A.â•… 89–90, 99 Stevanović, M.â•… 257 Stilman, L.â•… 3, 207 Stosic, D.â•… 228 Strömqvist, S.â•… 2, 199, 226 Sussex, R.â•… 2, 207 Švedova, N. Ju.â•… 126, 345 T Talmy, L.â•… 2–3, 6–7, 59, 62, 69–70, 90–91, 111, 163, 198, 200–201, 203, 205, 207, 225, 230, 233–234, 239, 247–249, 259, 267, 272, 311, 316, 323, 339, 344–345, 361–363, 374 Taylor, J. R.â•… 346–347 Tenny, C.â•… 90, 92–94, 99 Thelwall, M.â•… 164–165 Timberlake, A.â•… 57, 125, 141, 150, 154, 166, 179, 268, 344, 354–355 Tixonov, A. N.â•… 128 Toporov, V. N.â•… 284 Tottie, G.â•… 174 Townsend, C. E.â•… 126 Traugott, E. C.â•… 297 Treis, Y.â•… 197 Trubachev, O. N.â•… 48 Turner, S.â•… 4, 18

U Ungermanová, M.â•… 270 V Vaillant, A.â•… 48, 56–57, 69, 85–86, 99, 116 Van Wijk, N.â•… 67–68, 93 Vdovina, T.â•… 9 Vendler, Z.â•… 61, 99 Verhoeven, L.â•… 2, 199, 226 Verkuyl, H. J.â•… 254 Veyrenc, J.â•… 69 Vidaković, I.â•… 263 Vieu, L.â•… 1, 199 Vinogradov, V. V.â•… 3, 127–128, 203 W Wade, T.â•… 126, 203, 344, 346 Ward, D.â•… 207, 344, 346 Wienold, G.â•… 200 Wierzbicka, A.â•… 292, 349 Wilkins, D.â•… 1 Winter, W.â•… 62, 114 Y Yangklang, P.â•… 248, 287 Yokoyama, O. T.â•… 17, 22, 24 Yu, L.â•… 200 Z Zalizniak, A. A.â•… 141, 143, 205 Zemskaja, E. A.â•… 367 Zlatev, J.â•… 248, 287

Language index

A Albanianâ•… 112 Avestan (=AV)â•… 113–115 B Balticâ•… 100, 113, 116 Balto-Slavicâ•… 62, 90, 113–115, 117–118 Proto-Balto-Slavicâ•… 4, 47, 62 Bantuâ•… 199 BCS (Bosnian/Croatian/ Serbian, Bosnian- Croatian-Serbian)â•… 56–57, 59, 70–73, 81, 96–99, 113–117 Serbo-Croatianâ•… 7, 202, 204, 215–216, 228, 242, 247–252, 254–263, 266 Belorussian, Belarusian (=Br)â•… 116, 335 Bulgarian (=Bg)â•… 15, 59, 67, 71, 73, 83, 96–98, 113, 115–117, 262, 335–336 C Czech (=Cz)â•… 69, 73, 93, 96–98, 113–117, 256, 261–262, 270, 335 Old Czechâ•… 76–77, 90 Common Slavicâ•… 4–5, 47, 49, 55–57, 62, 67–69, 73, 83, 89–95, 101, 104–105 D Danishâ•… 335, 364, 369 Dutchâ•… 8, 111, 315–317, 320–321, 323–325, 331–332, 335, 337

E East Slavicâ•… 15–16, 19–20, 29, 39–40, 47, 59, 69, 114–115, 207 Englishâ•… 7–9, 15, 53, 58–59, 63, 70–72, 80, 85, 92–94, 111, 117, 144–146, 160, 165, 177, 197–220, 225–227, 235–242, 248–258, 261–263, 271, 282, 288–289, 292, 315–317, 320– 321, 323–325, 331–332, 335–340, 363–369, 374, 377, 380 F Finnishâ•… 336 Finno-Ugricâ•… 199 Frenchâ•… 226, 244, 268, 287, 292, 339–340, 363 G Germanicâ•… 8, 62, 112, 199, 202, 216, 248, 315–317, 321–322, 331–332, 335–337, 339–340 Germanâ•… 111, 223, 335, 337 Middle High German (=MHG)â•… 76–77 Old High German (=OHG)â•… 116 Gothic (=Go)â•… 93, 113 Greek (=Gr), Classical Greek, Ancient Greek, New Testament Greekâ•… 48, 63, 67, 70–73, 75–76, 82, 85, 95, 103, 112–116, 282 H Hakassâ•… 336 Hindiâ•… 336 Hittite (=Hi)â•… 112–113

I Icelandicâ•… 335 Indo-Europeanâ•… 2–4, 48–50, 63–64, 90, 93, 105, 112–113, 119–120, 336 Proto-Indo-European (=PIE)â•… 4–5, 47, 59, 62, 111–117 J Japaneseâ•… 199–200, 336 L Latin (=Lt)â•… 112, 114, 336 Latvian (=La)â•… 70, 75, 113, 115–117, 341 Lithuanian (=Li)â•… 100, 113–117, 336 M Macedonian (=Ma)â•… 73, 115, 121 Maninkaâ•… 323, 337 Mayanâ•… 199 Middle High Germanâ•… see German N Nilo-Saharanâ•… 197 North Slavicâ•… 67, 69–73, 86, 89–99, 104–115 Norwegianâ•… 335 O Old Church Slavic (=OCS), Church Slavicâ•… 3–4, 34, 41, 44, 48, 50–54, 58, 67–69, 71–76, 82, 84–90, 92, 94–95, 97–98, 101–105, 112–117 see also Proto-Slavic

388 New Approaches to Slavic Verbs of Motion

Old Czechâ•… see Czech Old High Germanâ•… see German Old Norse (=ON)â•… 117 Old Prussianâ•… 112 Old Russianâ•… 3–4, 67–9, 73–74, 77–94, 101–104, 335 P Persianâ•… 323, 336–337 Polabian (=Po)â•… 113–115 Polish (=Pl)â•… 7–8, 47, 59, 73, 114–115, 202, 204, 215–216, 225–242, 244, 270, 284, 291–311, 315–317, 323–325, 331–336 Proto-Slavic (=PS)â•… 3–5, 48–51, 53–55, 61–62, 93, 111, 113–119, 121 see also Old Church Slavic R Romanceâ•… 111, 199, 248, 251, 340, 364

Russian (=R), Contemporary Standard Russian (=CSR); Modern Russian (Mrus)â•… 4–9, 16–17, 25–26, 29, 32, 34, 37, 42–43, 49, 55–57, 60, 63–65, 69, 71, 74–75, 78–88, 91, 93, 96, 98, 111–112, 125–137, 141–161, 163–190, 197–220, 255–256, 262, 267–287, 291–311, 315–317, 323–332, 335–339, 343–357, 361–380 see also Old Russian

264, 270, 282, 291, 315–316, 321, 323, 332, 335–340, 357, 363, 369, 379 see also Common Slavic, East Slavic, North Slavic, Old Church Slavic, ProtoSlavic, South Slavic, West Slavic Slovene (=Sl)â•… 58, 64, 106, 114 South Slavicâ•… 4, 58, 70, 96–97, 115 Swedishâ•… 8, 315–318, 320–321, 323–325, 331–337, 341

S Samoyedicâ•… 197 Sanskrit (=Sk)â•… 112, 114 Vedic Sanskritâ•… 113, 115 Semiticâ•… 199 Serbo-Croatianâ•… see BCS Slavicâ•… 1–10, 15, 47–50, 53, 58–63, 67–73, 77, 83, 89–106, 111–119, 143, 201– 207, 215–216, 220, 228, 232, 248–249, 254–256, 262–

T Tamilâ•… 199, 323, 337 Tocharian Bâ•… 112 U Ukrainian (=Uk)â•… 114–116, 270, 335 W West Slavicâ•… 69, 76, 90, 100, 115

Subject index

A accentuationâ•… 55–56 acquisitionâ•… 200–201, 209, 364, 366–368 adverb, adverbialâ•… 29, 35–37, 74, 78, 84, 102–103, 147, 160, 225–233, 239, 245, 258, 350–352 affirmationâ•… 167, 172 affirmativeâ•… 148–149, 152, 155, 166–179, 181, 186, 190 agentâ•… 50–53, 58, 60, 85, 87, 91, 99, 174, 181, 228, 230, 235, 318 agentive, agentivityâ•… 54–55, 58–59, 322 Aktionsartâ•… 50, 54–55, 59, 205, 214, 370, 379 annulled resultâ•… 149, 154, 157, 159 anthropocentrismâ•… 9, 357, 343–344, 349, 357 aquamotionâ•… 315–340 active aquamotionâ•… 326–328 passive aquamotionâ•… 317– 319, 321–322, 325, 328, 330–331, 333, 337 aspect, aspectualâ•… 59, 67–74, 81, 89–101, 105, 112, 125–129, 134, 137, 141–161, 164, 166–168, 177–190, 198, 205–208, 210, 213–215, 218–219, 232, 262, 272–273, 283–285, 367–369 see also aspectuality, imperfective, perfective aspectualityâ•… 49, 61, 63, 214, 219 atelic, atelicityâ•… 50, 55, 73, 81, 92–95, 99, 105, 132, 145, 152, 157, 159 axisâ•… 259, 293–296, 298, 304, 306–311

B background, backgroundingâ•… 41, 44, 82, 86, 91, 99, 116, 156, 199, 239, 284 bidirectionalityâ•… 154 boundaryâ•… 72, 81, 94, 96–97, 130, 166, 178, 208, 210, 217, 251–264 boundary crossing and boundary reachingâ•… 97, 208, 217, 219, 251–256, 259–262 non-boundary-crossing and non-boundary-reachingâ•… 208, 259–262 boundednessâ•… 95, 129, 132, 152

completabilityâ•… 125–127, 129– 132, 134, 136, 138 conceptual, conceptualizationâ•… 17–18, 42, 68, 198, 200–203, 217–218, 288, 308, 315–317, 344–345, 349, 356 constraintâ•… 142–143, 152–153, 157, 161, 204, 206, 240, 248, 250, 260, 270–272, 287, 300, 322 cooperative principleâ•… 157 corpusâ•… 18–26, 30–33, 38–44, 60, 89, 125–126, 164–165, 191–192, 209–219, 227–229, 247–254, 261–263, 270, 282–283, 291, 316, 326–329, 350–352, 365–366

C caseâ•… 15, 53, 95, 104, 174, 229, 233, 267–272, 282, 285, 326–330 causativeâ•… 47–49, 55–59, 63, 114–115, 300 change-occurredâ•… 260, 262 see also no-change change of locationâ•… 248, 268, 271–277, 279–282, 284–285, 287 change of positionâ•… 273–275, 278, 280, 284 change of stateâ•… 94, 99, 128, 143, 241, 267, 273–281, 285–287 cluster modelâ•… 126–127, 130, 134–137 cognitiveâ•… 129, 197–200, 242–243, 259, 293, 343–347, 349, 356–357, 361–363, 366, 373–374, 377–380 combinatory potentialâ•… 213, 247, 251, 253, 257, 259–264

D deadjectivalâ•… 49–50, 56, 59, 61 deixis, deicticâ•… 22, 212, 215, 249–250, 253–258, 339 denominalâ•… 47–50, 55–57, 59, 61–63, 114, 116 derivationâ•… 47–49, 61–63, 93, 100, 115, 125, 133, 205–206, 213 determinacyâ•… 67, 69, 73, 76, 78–79, 82, 84, 88–89, 131 determinateâ•… 57–63, 67–69, 71–75, 80–91, 95–96, 98–103, 129–131, 141–143, 157–158, 166–167, 179, 186–189, 232, 268–269 deverbalâ•… 47–48, 50–51, 55–58, 62–63, 115, 298 diachrony, diachronicâ•… 112, 120, 336 directionalityâ•… 49, 207–208, 218, 317, 331, 336, 344–346, 355–356, 365, 379–380 distribution of Mannerâ•… 233– 234, 237, 240

390 New Approaches to Slavic Verbs of Motion

duration, durativeâ•… 84, 114, 118, 208, 271, 283

heritage speakersâ•… 361–362, 366–380

E embodimentâ•… 343–344, 349, 357 encoding of Mannerâ•… 205, 210, 226 see also Manner expression and Manner verb encoding of Pathâ•… 202, 206, 212, 216–217, 226 see also Path expression, Path prefix, and Path verb endpointâ•… 72, 98, 143–146, 158–159, 166, 218, 267–277, 279, 281–288, 321, 366 event construalâ•… 267, 281–286, 288 expressive constituentâ•… 17, 42, 118 expressivenessâ•… 26–28

I identity functionâ•… 141, 147–148, 152, 160 imperfectiveâ•… 59, 71, 111–112, 118, 125–129, 132–138, 141–146, 148–153, 156–160, 165–171, 178–190, 205–207, 260–262, 365–368 see also aspect, perfective imperfectivityâ•… 144–145, 152, 157 imperfectivizationâ•… 126, 261–262 indeterminateâ•… 47–49, 53, 56–58, 61–63, 67–81, 84–99, 102–105, 111–114, 117–118, 125–138, 141–161, 166–167, 179, 186–190, 232, 268–269 informational constituentâ•… 16, 24, 42–44 interrogative clauseâ•… 20, 149 inter-typologicalâ•… see variability intransitiveâ•… 15–16, 25, 49–50, 54–55, 58–61, 74, 85, 91, 95, 99, 104, 136, 273, 348 intra-typologicalâ•… see variability iterative, iterativityâ•… 48, 57, 59–62, 84, 93, 114–115, 141– 149, 250, 261–262 itineraryâ•… 350–352

F Figureâ•… 172–184, 188–190, 198, 208, 215, 230, 234–236, 248, 259, 293, 326–328, 331 animate and inanimate Figureâ•… 317–322, 324, 336–337 liquid and non-liquid Figureâ•… 316–317, 331–336, 338 foreground, foregroundingâ•… 42, 44, 95, 99, 199 G Goalâ•… 58–59, 68–83, 89, 92–98, 102–103, 112, 130–131, 146, 154–161, 172–178, 185, 189–190, 218, 260, 267–272, 277–282, 287, 293–295, 302–303, 309, 326–332, 343–344, 349–357, 364–366 Grammaticality Judgment Taskâ•… 362, 370–373, 379 Groundâ•… 198, 228, 235, 246, 248, 250, 255–256, 293, 316–318, 324, 326–332 H habitualityâ•… 77, 93, 141, 144

L Landmarkâ•… 72, 94, 96, 98, 293, 295, 301–302, 304–311 L2â•… see second language lexicalization patternâ•… 112, 163–164, 178, 198, 200, 205, 226, 240, 259, 262, 316, 323, 339, 364 Locationâ•… 16, 50, 58, 92, 95, 131, 157–160, 178, 198, 211, 228, 248, 255, 260, 267–288, 306, 310, 319–323, 325, 333–337, 348 locational phraseâ•… 267, 269–275, 279–288 locativeâ•… 15, 203, 229, 246, 267–269, 273, 283–286, 289–290

M Manner expressionâ•… 215, 227, 231–232, 237, 241, 378 see also encoding of Manner and Manner verb Manner verbâ•… 58, 100–101, 105, 111–112, 114, 116, 163, 166–167, 180, 186, 199–204, 211–215, 225–242, 240–241, 247–264, 339, 363 maximize assertionâ•… 152–153, 156–158, 160–161 maximize presuppositionâ•… 152– 153 metaphor, metaphoricalâ•… 75, 85, 88, 104, 130–131, 135, 200, 250, 274, 295, 301, 305–307, 310, 326, 343–344, 349, 353, 356–357 see also source domain metonymyâ•… 128, 130, 132, 137, 336 modifierâ•… 22, 30, 33, 35–37, 41, 229–230, 233–234, 237–238, 240 modifying expressionâ•… 225, 227, 229, 230, 234, 237–241 moment-of-changeâ•… 260–262 morphological blockingâ•… 206, 247, 251, 259, 261–264 morphologyâ•… 48–49, 57, 69, 95, 126, 153, 247, 251, 259, 368 morphosyntaxâ•… 107, 200, 204, 217–219 motion eventâ•… 68, 70, 74, 77, 81–85, 98, 151–154, 163, 172–174, 176, 178, 182–185, 190, 197–200, 203–210, 214, 217–220, 226–229, 242, 247–250, 259–260, 263, 267, 271–272, 276, 283– 288, 323, 339, 364, 366 motion, directedâ•… 92, 96, 98, 203, 225, 226, 229, 240–242, 271–272, 274, 276–277, 279, 287, 322–323 motion, endpoint ofâ•… see endpoint motion goalâ•… see Goal motion, goal-directedâ•… 70–72, 95, 102

motion, goal-orientedâ•… 103, 105, 173, 344 motion, habitualâ•… 68, 73, 93–94, 102, 142, 144–145, 147–149, 179, 232, 250, 261–262 motion locationâ•… see Location motion Mannerâ•… see encoding of Manner, Manner expression, Manner verb motion, negatedâ•… 163, 172, 176, 183–185 motion, non-directedâ•… 131, 228, 232 motion originâ•… see Origin motion, passiveâ•… 319–321, 323, 325–329, 336–337 motion, prototypicalâ•… 343–344, 346, 348, 357 motion, self-propelledâ•… 321–325, 327, 329, 336, 337 motion sourceâ•… see Source motion, starting point ofâ•… see starting point motion, there and backâ•… 131, 143, 150, 154–157, 159, 161 see also roundtrip multidirectionalâ•… 147, 363, 365–371, 373–376, 378–381 see also non-directional, non-unidirectional, indeterminate multiple directionsâ•… 141, 146, 148, 166 N narrative, narrationâ•… 26, 32, 40–42, 76, 111, 173, 175–176, 197, 200, 209–218, 250, 258, 363–364, 378 negationâ•… 148–149, 152, 155, 163–164, 166–167, 172, 174– 180, 185–186, 189–190 no-changeâ•… 260–262 see also change-occurred non-deicticâ•… 253–254, 256 non-directional, nondirectionalityâ•… 92, 101–102, 112, 310, 317, 325, 331, 344–346, 353–355, 365–366

Subject index 391

non-unidirectional, nonunidirectionalityâ•… 197, 207–208, 213, 268, 273, 315, 345 O observerâ•… 163, 172, 174–178, 180–182, 185–190 origin (historical)â•… 47–49, 56, 63, 113, 117 Originâ•… 16, 163, 172, 176–178, 180–181, 185, 189–190 P Paroleâ•… 334, 341 particleâ•… 20, 22, 24, 26–27, 37, 70, 193, 199, 202, 206, 226, 267, 364 Path expressionâ•… 229, 231–232, 240–242, 374, 380 see also encoding of Path, Path verb, Path prefix Path prefixâ•… 167, 172, 203–204, 210, 240, 363, 265–366, 369 Path verbâ•… 112, 199, 210, 230– 232, 237–238, 363 perfectiveâ•… 55, 71, 81, 84, 125– 138, 141, 143, 153, 156–159, 166–167, 176, 178–180, 185–186, 189–190, 205– 206, 213–214, 249–250, 252, 260–262, 365, 371, 374, 380–381 see also aspect, aspectuality, imperfective perfectivityâ•… 95, 105, 143, 157, 159 positional verbâ•… 268 prefixâ•… see Path prefix, prefixation prefixationâ•… 67, 90, 95–96, 105, 125–129, 163, 167, 172, 203, 205–206, 208, 217–218, 226–229, 231, 251, 262–264, 344, 364, 366 prepositionâ•… 75, 78, 95, 98, 103, 111, 203–204, 206, 217, 229, 233, 248, 250–263, 267–271, 284–286

prepositional phraseâ•… 20, 23, 25, 36–37, 70, 204, 228–230, 353, 270, 285, 353 presentational constituentâ•… 42– 44 presentational word orderâ•… 16 prototypeâ•… 129, 322, 338, 343–344, 346–347, 349, 357, 365, 368 prototypicalityâ•… 129, 365–366, 368, 375, 380 R reflexiveâ•… 20, 47, 49–50, 60, 63, 88, 300–301, 303 resultative adjectivesâ•… 70–71 rhemeâ•… 16, 18, 22, 26–28, 31–43 rotationâ•… 216, 291–299, 301–302, 306–311 roundtripâ•… 71, 80–81, 131, 136, 138, 128–129, 150, 154, 355, 365, 367, 369–371, 373–374, 376, 379–380 see also there and back S salienceâ•… 62, 96, 118, 198, 200– 202, 211, 216, 284 salientâ•… 23, 34, 59, 94, 118, 126, 130, 135, 160, 198–199, 208, 215–216, 241, 310, 344, 356 satelliteâ•… 70, 199, 202–204, 206, 210, 212–213, 216–219, 225–227, 229, 241, 248, 287, 310, 317, 364 satellite-framedâ•… 70, 105, 111, 163, 197–199, 212, 225–227, 237, 241–242, 248–249, 251, 255, 263, 267–268, 271, 281, 286– 287, 316, 339–340, 363–364, 369, 377, 380 second language, L2â•… 200–202, 204, 209, 264, 361–380 semanticsâ•… 49, 55, 58–59, 84, 94, 113, 143, 146–149, 151–154, 157, 160–161, 187, 204, 208, 217, 229, 291–293, 308, 356 Sentence Completion Taskâ•… 376, 379

392 New Approaches to Slavic Verbs of Motion

Sourceâ•… 155, 158–161, 250, 279– 281, 287, 317, 326–328, 330, 332, 334, 364–366 source domainâ•… 130–131, 344, 349, 356 see also metaphor spaceâ•… 1, 44, 197, 207, 218–219, 252–253, 274, 278, 281, 307, 310, 315, 317, 339, 363, 365 spatial meaningâ•… 1, 98, 203 starting pointâ•… 218, 279, 353 statement of factâ•… 144–145, 149–154, 159 story-tellingâ•… 7, 198, 209 synchronicâ•… 26, 55–56, 69 syntaxâ•… 29, 199, 250, 259 T telic, telicityâ•… 50, 55, 58–60, 71–72, 94–95, 98, 129, 132, 143–145, 157 themeâ•… 16–18, 22, 24, 27–28, 30, 34, 37–44, 114, 274–275, 284, 338 thinking for speakingâ•… 339, 361–362, 364–365, 374, 380

topicalizingâ•… 284 Trajectorâ•… 293–296, 301–304, 308–310 trajectoryâ•… 68, 73, 75–78, 80, 83, 85–89, 92–102, 105, 163, 172, 176–178, 218, 303–304, 309 transitiveâ•… 20, 47, 49–51, 54–56, 59, 61, 74, 78, 85, 90–91, 95, 99, 136, 273, 318, 348–349 typologyâ•… 68, 70–71, 198, 200–291, 207, 212, 225, 227, 242, 247–249, 259–260, 263, 267–268, 286, 310–311, 321, 336, 338, 363, 380 U unaccusativizationâ•… 70, 72, 92, 98 unboundednessâ•… 95, 152 unidirectional, unidirectionalityâ•… 88, 148–150, 152–153, 158, 198, 207–208, 210, 212, 215, 218– 219, 268, 273, 315, 324–325, 331–334, 343–346, 353–357, 363, 366–371, 373–376, 378–380

universalâ•… 1, 68, 197, 247, 250, 264, 332 V valenceâ•… 49, 54–55, 59 variabilityâ•… 197–198, 200, 202, 208, 220 varietyâ•… 18, 26, 29, 31, 50, 55, 63, 71, 98, 115, 164–165, 201, 210–211, 215, 226–227, 234– 235, 241–242, 257, 278, 282 verbâ•… see atelic, deictic, imperfective, indeterminate, intransitive, iterative, Manner verb, multidirectional, nondeictic, non-directional, non-unidirectional, Path verb, perfective, positional, reflexive, telic, terminate, transitive, unidirectional verb-framedâ•… 75, 106, 111, 198, 226, 237, 248, 251, 263, 267, 287, 339–340, 363–364 vesselâ•… 88, 114, 317–330, 333–334, 336–338

Studies in Language Companion Series A complete list of titles in this series can be found on the publishers’ website, www.benjamins.com 118 Mühleisen, Susanne: Heterogeneity in Word-Formation Patterns. A corpus-based analysis of suffixation with -ee and its productivity in English. 2010. xiii, 245 pp. 117 Spevak, Olga: Constituent Order in Classical Latin Prose. 2010. xv, 318 pp. 116 Nordström, Jackie: Modality and Subordinators. 2010. xvii, 341 pp. 115 Hasko, Victoria and Renee Perelmutter (eds.): New Approaches to Slavic Verbs of Motion. 2010. x, 392 pp. 114 Roby, David Brian: Aspect and the Categorization of States. The case of ser and estar in Spanish. 2009. xiii, 191 pp. 113 Comrie, Bernard, Ray Fabri, Elizabeth Hume, Manwel Mifsud, Thomas Stolz and Martine Vanhove (eds.): Introducing Maltese Linguistics. Selected papers from the 1st International Conference on Maltese Linguistics, Bremen, 18–20 October, 2007. 2009. xi, 422 pp. 112 Dufter, Andreas and Daniel Jacob (eds.): Focus and Background in Romance Languages. 2009. vii, 362 pp. 111 Polguère, Alain and Igor A. Mel’čuk (eds.): Dependency in Linguistic Description. 2009. xxii, 281 pp. 110 Dimmendaal, Gerrit J. (ed.): Coding Participant Marking. Construction types in twelve African languages. 2009. xvi, 389 pp. 109 Narrog, Heiko: Modality in Japanese. The layered structure of the clause and hierarchies of functional categories. 2009. xxii, 277 pp. 108 Barðdal, Jóhanna and Shobhana L. Chelliah (eds.): The Role of Semantic, Pragmatic, and Discourse Factors in the Development of Case. 2009. xx, 432 pp. 107 Butler, Christopher S. and Javier Martín Arista (eds.): Deconstructing Constructions. 2009. xx, 306 pp. 106 Vanhove, Martine (ed.): From Polysemy to Semantic Change. Towards a typology of lexical semantic associations. 2008. xiii, 404 pp. 105 Van Valin, Jr., Robert D. (ed.): Investigations of the Syntax–Semantics–Pragmatics Interface. 2008. xxiv, 484 pp. 104 Mushin, Ilana and Brett Baker (eds.): Discourse and Grammar in Australian Languages. 2008. x, 239 pp. 103 Josephson, Folke and Ingmar Söhrman (eds.): Interdependence of Diachronic and Synchronic Analyses. 2008. viii, 350 pp. 102 Goddard, Cliff (ed.): Cross-Linguistic Semantics. 2008. xvi, 356 pp. 101 Stolz, Thomas, Sonja Kettler, Cornelia Stroh and Aina Urdze: Split Possession. An areallinguistic study of the alienability correlation and related phenomena in the languages of Europe. 2008. x, 546 pp. 100 Ameka, Felix K. and M.E. Kropp Dakubu (eds.): Aspect and Modality in Kwa Languages. 2008. ix, 335 pp. 99 Høeg Müller, Henrik and Alex Klinge (eds.): Essays on Nominal Determination. From morphology to discourse management. 2008. xviii, 369 pp. 98 Fabricius-Hansen, Cathrine and Wiebke Ramm (eds.): 'Subordination' versus 'Coordination' in Sentence and Text. A cross-linguistic perspective. 2008. vi, 359 pp. 97 Dollinger, Stefan: New-Dialect Formation in Canada. Evidence from the English modal auxiliaries. 2008. xxii, 355 pp. 96 Romeo, Nicoletta: Aspect in Burmese. Meaning and function. 2008. xv, 289 pp. 95 O’Connor, Loretta: Motion, Transfer and Transformation. The grammar of change in Lowland Chontal. 2007. xiv, 251 pp. 94 Miestamo, Matti, Kaius Sinnemäki and Fred Karlsson (eds.): Language Complexity. Typology, contact, change. 2008. xiv, 356 pp. 93 Schalley, Andrea C. and Drew Khlentzos (eds.): Mental States. Volume 2: Language and cognitive structure. 2007. x, 362 pp. 92 Schalley, Andrea C. and Drew Khlentzos (eds.): Mental States. Volume 1: Evolution, function, nature. 2007. xii, 304 pp.

91 Filipović, Luna: Talking about Motion. A crosslinguistic investigation of lexicalization patterns. 2007. x, 182 pp. 90 Muysken, Pieter (ed.): From Linguistic Areas to Areal Linguistics. 2008. vii, 293 pp. 89 Stark, Elisabeth, Elisabeth Leiss and Werner Abraham (eds.): Nominal Determination. Typology, context constraints, and historical emergence. 2007. viii, 370 pp. 88 Ramat, Paolo and Elisa Roma (eds.): Europe and the Mediterranean as Linguistic Areas. Convergencies from a historical and typological perspective. 2007. xxvi, 364 pp. 87 Verhoeven, Elisabeth: Experiential Constructions in Yucatec Maya. A typologically based analysis of a functional domain in a Mayan language. 2007. xiv, 380 pp. 86 Schwarz-Friesel, Monika, Manfred Consten and Mareile Knees (eds.): Anaphors in Text. Cognitive, formal and applied approaches to anaphoric reference. 2007. xvi, 282 pp. 85 Butler, Christopher S., Raquel Hidalgo Downing and Julia Lavid (eds.): Functional Perspectives on Grammar and Discourse. In honour of Angela Downing. 2007. xxx, 481 pp. 84 Wanner, Leo (ed.): Selected Lexical and Grammatical Issues in the Meaning–Text Theory. In honour of Igor Mel'čuk. 2007. xviii, 380 pp. 83 Hannay, Mike and Gerard J. Steen (eds.): Structural-Functional Studies in English Grammar. In honour of Lachlan Mackenzie. 2007. vi, 393 pp. 82 Ziegeler, Debra: Interfaces with English Aspect. Diachronic and empirical studies. 2006. xvi, 325 pp. 81 Peeters, Bert (ed.): Semantic Primes and Universal Grammar. Empirical evidence from the Romance languages. 2006. xvi, 374 pp. 80 Birner, Betty J. and Gregory Ward (eds.): Drawing the Boundaries of Meaning. Neo-Gricean studies in pragmatics and semantics in honor of Laurence R. Horn. 2006. xii, 350 pp. 79 Laffut, An: Three-Participant Constructions in English. A functional-cognitive approach to caused relations. 2006. ix, 268 pp. 78 Yamamoto, Mutsumi: Agency and Impersonality. Their Linguistic and Cultural Manifestations. 2006. x, 152 pp. 77 Kulikov, Leonid, Andrej Malchukov and Peter de Swart (eds.): Case, Valency and Transitivity. 2006. xx, 503 pp. 76 Nevalainen, Terttu, Juhani Klemola and Mikko Laitinen (eds.): Types of Variation. Diachronic, dialectal and typological interfaces. 2006. viii, 378 pp. 75 Hole, Daniel, André Meinunger and Werner Abraham (eds.): Datives and Other Cases. Between argument structure and event structure. 2006. viii, 385 pp. 74 Pietrandrea, Paola: Epistemic Modality. Functional properties and the Italian system. 2005. xii, 232 pp. 73 Xiao, Richard and Tony McEnery: Aspect in Mandarin Chinese. A corpus-based study. 2004. x, 305 pp. 72 Frajzyngier, Zygmunt, Adam Hodges and David S. Rood (eds.): Linguistic Diversity and Language Theories. 2005. xii, 432 pp. 71 Dahl, ÃŒsten: The Growth and Maintenance of Linguistic Complexity. 2004. x, 336 pp. 70 Lefebvre, Claire: Issues in the Study of Pidgin and Creole Languages. 2004. xvi, 358 pp. 69 Tanaka, Lidia: Gender, Language and Culture. A study of Japanese television interview discourse. 2004. xvii, 233 pp. 68 Moder, Carol Lynn and Aida Martinovic-Zic (eds.): Discourse Across Languages and Cultures. 2004. vi, 366 pp. 67 Luraghi, Silvia: On the Meaning of Prepositions and Cases. The expression of semantic roles in Ancient Greek. 2003. xii, 366 pp. 66 Nariyama, Shigeko: Ellipsis and Reference Tracking in Japanese. 2003. xvi, 400 pp. 65 Matsumoto, Kazuko: Intonation Units in Japanese Conversation. Syntactic, informational and functional structures. 2003. xviii, 215 pp. 64 Butler, Christopher S.: Structure and Function – A Guide to Three Major Structural-Functional Theories. Part 2: From clause to discourse and beyond. 2003. xiv, 579 pp. 63 Butler, Christopher S.: Structure and Function – A Guide to Three Major Structural-Functional Theories. Part 1: Approaches to the simplex clause. 2003. xx, 573 pp. 62 Field, Fredric: Linguistic Borrowing in Bilingual Contexts. With a foreword by Bernard Comrie. 2002. xviii, 255 pp.

61 Goddard, Cliff and Anna Wierzbicka (eds.): Meaning and Universal Grammar. Theory and empirical findings. Volume 2. 2002. xvi, 337 pp. 60 Goddard, Cliff and Anna Wierzbicka (eds.): Meaning and Universal Grammar. Theory and empirical findings. Volume 1. 2002. xvi, 337 pp. 59 Shi, Yuzhi: The Establishment of Modern Chinese Grammar. The formation of the resultative construction and its effects. 2002. xiv, 262 pp. 58 Maylor, B. Roger: Lexical Template Morphology. Change of state and the verbal prefixes in German. 2002. x, 273 pp. 57 Mel’čuk, Igor A.: Communicative Organization in Natural Language. The semantic-communicative structure of sentences. 2001. xii, 393 pp. 56 Faarlund, Jan Terje (ed.): Grammatical Relations in Change. 2001. viii, 326 pp. 55 Dahl, ÃŒsten and Maria Koptjevskaja-Tamm (eds.): Circum-Baltic Languages. Volume 2: Grammar and Typology. 2001. xx, 423 pp. 54 Dahl, ÃŒsten and Maria Koptjevskaja-Tamm (eds.): Circum-Baltic Languages. Volume 1: Past and Present. 2001. xx, 382 pp. 53 Fischer, Olga, Anette Rosenbach and Dieter Stein (eds.): Pathways of Change. Grammaticalization in English. 2000. x, 391 pp. 52 Torres Cacoullos, Rena: Grammaticization, Synchronic Variation, and Language Contact. A study of Spanish progressive -ndo constructions. 2000. xvi, 255 pp. 51 Ziegeler, Debra: Hypothetical Modality. Grammaticalisation in an L2 dialect. 2000. xx, 290 pp. 50 Abraham, Werner and Leonid Kulikov (eds.): Tense-Aspect, Transitivity and Causativity. Essays in honour of Vladimir Nedjalkov. 1999. xxxiv, 359 pp. 49 Bhat, D.N.S.: The Prominence of Tense, Aspect and Mood. 1999. xii, 198 pp. 48 Manney, Linda Joyce: Middle Voice in Modern Greek. Meaning and function of an inflectional category. 2000. xiii, 262 pp. 47 Brinton, Laurel J. and Minoji Akimoto (eds.): Collocational and Idiomatic Aspects of Composite Predicates in the History of English. 1999. xiv, 283 pp. 46 Yamamoto, Mutsumi: Animacy and Reference. A cognitive approach to corpus linguistics. 1999. xviii, 278 pp. 45 Collins, Peter and David Lee (eds.): The Clause in English. In honour of Rodney Huddleston. 1999. xv, 342 pp. 44 Hannay, Mike and A. Machtelt Bolkestein (eds.): Functional Grammar and Verbal Interaction. 1998. xii, 304 pp. 43 Olbertz, Hella, Kees Hengeveld and Jesús Sánchez García (eds.): The Structure of the Lexicon in Functional Grammar. 1998. xii, 312 pp. 42 Darnell, Michael, Edith A. Moravcsik, Michael Noonan, Frederick J. Newmeyer and Kathleen M. Wheatley (eds.): Functionalism and Formalism in Linguistics. Volume II: Case studies. 1999. vi, 407 pp. 41 Darnell, Michael, Edith A. Moravcsik, Michael Noonan, Frederick J. Newmeyer and Kathleen M. Wheatley (eds.): Functionalism and Formalism in Linguistics. Volume I: General papers. 1999. vi, 486 pp. 40 Birner, Betty J. and Gregory Ward: Information Status and Noncanonical Word Order in English. 1998. xiv, 314 pp. 39 Wanner, Leo (ed.): Recent Trends in Meaning–Text Theory. 1997. xx, 202 pp. 38 Hacking, Jane F.: Coding the Hypothetical. A comparative typology of Russian and Macedonian conditionals. 1998. vi, 156 pp. 37 Harvey, Mark and Nicholas Reid (eds.): Nominal Classification in Aboriginal Australia. 1997. x, 296 pp. 36 Kamio, Akio (ed.): Directions in Functional Linguistics. 1997. xiii, 259 pp. 35 Matsumoto, Yoshiko: Noun-Modifying Constructions in Japanese. A frame semantic approach. 1997. viii, 204 pp. 34 Hatav, Galia: The Semantics of Aspect and Modality. Evidence from English and Biblical Hebrew. 1997. x, 224 pp. 33 Velázquez-Castillo, Maura: The Grammar of Possession. Inalienability, incorporation and possessor ascension in Guaraní. 1996. xvi, 274 pp. 32 Frajzyngier, Zygmunt: Grammaticalization of the Complex Sentence. A case study in Chadic. 1996. xviii, 501 pp.

31 Wanner, Leo (ed.): Lexical Functions in Lexicography and Natural Language Processing. 1996. xx, 355 pp. 30 Huffman, Alan: The Categories of Grammar. French lui and le. 1997. xiv, 379 pp. 29 Engberg-Pedersen, Elisabeth, Michael Fortescue, Peter Harder, Lars Heltoft and Lisbeth Falster Jakobsen (eds.): Content, Expression and Structure. Studies in Danish functional grammar. 1996. xvi, 510 pp. 28 Herman, József (ed.): Linguistic Studies on Latin. Selected papers from the 6th International Colloquium on Latin Linguistics (Budapest, 23–27 March 1991). 1994. ix, 421 pp. 27 Abraham, Werner, T. Givón and Sandra A. Thompson (eds.): Discourse, Grammar and Typology. Papers in honor of John W.M. Verhaar. 1995. xx, 352 pp. 26 Lima, Susan D., Roberta Corrigan and Gregory K. Iverson: The Reality of Linguistic Rules. 1994. xxiii, 480 pp. 25 Goddard, Cliff and Anna Wierzbicka (eds.): Semantic and Lexical Universals. Theory and empirical findings. 1994. viii, 510 pp. 24 Bhat, D.N.S.: The Adjectival Category. Criteria for differentiation and identification. 1994. xii, 295 pp. 23 Comrie, Bernard and Maria Polinsky (eds.): Causatives and Transitivity. 1993. x, 399 pp. 22 McGregor, William B.: A Functional Grammar of Gooniyandi. 1990. xx, 618 pp. 21 Coleman, Robert (ed.): New Studies in Latin Linguistics. Proceedings of the 4th International Colloquium on Latin Linguistics, Cambridge, April 1987. 1990. x, 480 pp. 20 Verhaar, John W.M. S.J. (ed.): Melanesian Pidgin and Tok Pisin. Proceedings of the First International Conference on Pidgins and Creoles in Melanesia. 1990. xiv, 409 pp. 19 Blust, Robert A.: Austronesian Root Theory. An essay on the limits of morphology. 1988. xi, 190 pp. 18 Wierzbicka, Anna: The Semantics of Grammar. 1988. vii, 581 pp. 17 Calboli, Gualtiero (ed.): Subordination and Other Topics in Latin. Proceedings of the Third Colloquium on Latin Linguistics, Bologna, 1–5 April 1985. 1989. xxix, 691 pp. 16 Conte, Maria-Elisabeth, János Sánder Petöfi and Emel Sözer (eds.): Text and Discourse Connectedness. Proceedings of the Conference on Connexity and Coherence, Urbino, July 16–21, 1984. 1989. xxiv, 584 pp. 15 Justice, David: The Semantics of Form in Arabic. In the mirror of European languages. 1987. iv, 417 pp. 14 Benson, Morton, Evelyn Benson and Robert F. Ilson: Lexicographic Description of English. 1986. xiii, 275 pp. 13 Reesink, Ger: Structures and their Functions in Usan. 1987. xviii, 369 pp. 12 Pinkster, Harm (ed.): Latin Linguistics and Linguistic Theory. Proceedings of the 1st International Colloquium on Latin Linguistics, Amsterdam, April 1981. 1983. xviii, 307 pp. 11 Panhuis, Dirk G.J.: The Communicative Perspective in the Sentence. A study of Latin word order. 1982. viii, 172 pp. 10 Dressler, Wolfgang U., Willi Mayerthaler, Oswald Panagl and Wolfgang Ullrich Wurzel: Leitmotifs in Natural Morphology. 1987. ix, 168 pp. 9 Lang, Ewald and John Pheby: The Semantics of Coordination. (English transl. by John Pheby from the German orig. ed. 'Semantik der koordinativen Verknüpfung', Berlin, 1977). 1984. 300 pp. 8 Barth, E.M. and J.L. Martens (eds.): Argumentation: Approaches to Theory Formation. Containing the Contributions to the Groningen Conference on the Theory of Argumentation, October 1978. 1982. xviii, 333 pp. 7 Parret, Herman, Marina Sbisà and Jef Verschueren (eds.): Possibilities and Limitations of Pragmatics. Proceedings of the Conference on Pragmatics, Urbino, July 8–14, 1979. 1981. x, 854 pp. 6 Vago, Robert M. (ed.): Issues in Vowel Harmony. Proceedings of the CUNY Linguistics Conference on Vowel Harmony, May 14, 1977. 1980. xx, 340 pp. 5 Haiman, John: Hua: A Papuan Language of the Eastern Highlands of New Guinea. 1980. iv, 550 pp. 4 Lloyd, Albert L.: Anatomy of the Verb. The Gothic Verb as a Model for a Unified Theory of Aspect, Actional Types, and Verbal Velocity. (Part I: Theory; Part II: Application). 1979. x, 351 pp. 3 Malkiel, Yakov: From Particular to General Linguistics. Selected Essays 1965–1978. With an introduction by the author, an index rerum and an index nominum. 1983. xxii, 659 pp. 2 Anwar, Mohamed Sami: BE and Equational Sentences in Egyptian Colloquial Arabic. 1979. vi, 128 pp. 1 Abraham, Werner (ed.): Valence, Semantic Case, and Grammatical Relations. Workshop studies prepared for the 12th International Congress of Linguists, Vienna, August 29th to September 3rd, 1977. 1978. xiv, 729 pp.

New Approaches to Slavic Verbs of Motion (Studies in Language Companion Series)

i> in Slavic (Studies in Language Companion Series)

Studies on German-Language Islands (Studies in Language Companion Series)

Anaphors in Text: Cognitive, formal and applied approaches to anaphoric reference (Studies in Language Companion Series)

Linguistic Diversity And Language Theories (Studies in Language Companion Series)

Linguistic Diversity And Language Theories (Studies in Language Companion Series)

Language Complexity: Typology, contact, change (Studies in Language Companion Series)

Talking About Motion: A crosslinguistic investigation of lexicalization patterns (Studies in Language Companion Series)

A Functional Grammar of Gooniyandi (Studies in Language Companion Series)

Grammatical Relations in Change (Studies in Language Companion Series)

Cross-Linguistic Semantics (Studies in Language Companion Series)

New Approaches to Enclosures

Motion, Transfer and Transformation: the Grammar of Change in Lowland Chontal (Studies in Language Companion Series, Volume 95)

The New! Beader's Companion (Companion series, The)

The New! Beader's Companion (Companion series, The)

Pathways of Change: Grammaticalization in English (Studies in Language Companion)

Approaches to Second Language Acquisition

Gender, Language and Culture: A Study of Japanese Television Interview Discourse (Studies in Language Companion Series)

Handbook of Generative Approaches to Language Acquisition

Language Policy: Hidden Agendas and New Approaches

Interdependence of Diachronic and Synchronic Analyses (Studies in Language Companion)

A Companion to the Philosophy of Language

A Companion to the Philosophy of Language

A Companion to the Philosophy of Language

The Ecology of Language Evolution (Cambridge Approaches to Language Contact)

The Ecology of Language Evolution (Cambridge Approaches to Language Contact)

Grammaticalization of the Complex Sentence: A Case Study in Chadic (Studies in Language Companion Series)

Mental States: Volume 2: Language and cognitive structure (Studies in Language Companion Series)

NMR Studies of Translational Motion

From Linguistic Areas to Areal Linguistics (Studies in Language Companion Series, Volume 90)

From Particular to General Linguistics: Selected Essays, 1965-1978 (Studies in Language Companion Series, V. 3)

New Approaches to Slavic Verbs of Motion (Studies in Language Companion Series)

i> in Slavic (Studies in Language Companion Series)

Studies on German-Language Islands (Studies in Language Companion Series)

Anaphors in Text: Cognitive, formal and applied approaches to anaphoric reference (Studies in Language Companion Series)

Linguistic Diversity And Language Theories (Studies in Language Companion Series)

Linguistic Diversity And Language Theories (Studies in Language Companion Series)

Language Complexity: Typology, contact, change (Studies in Language Companion Series)

Talking About Motion: A crosslinguistic investigation of lexicalization patterns (Studies in Language Companion Series)

A Functional Grammar of Gooniyandi (Studies in Language Companion Series)

Grammatical Relations in Change (Studies in Language Companion Series)

Cross-Linguistic Semantics (Studies in Language Companion Series)

New Approaches to Enclosures

Motion, Transfer and Transformation: the Grammar of Change in Lowland Chontal (Studies in Language Companion Series, Volume 95)

The New! Beader's Companion (Companion series, The)

The New! Beader's Companion (Companion series, The)

Pathways of Change: Grammaticalization in English (Studies in Language Companion)

Approaches to Second Language Acquisition

Gender, Language and Culture: A Study of Japanese Television Interview Discourse (Studies in Language Companion Series)

Handbook of Generative Approaches to Language Acquisition

Language Policy: Hidden Agendas and New Approaches

Interdependence of Diachronic and Synchronic Analyses (Studies in Language Companion)

A Companion to the Philosophy of Language

A Companion to the Philosophy of Language

A Companion to the Philosophy of Language

The Ecology of Language Evolution (Cambridge Approaches to Language Contact)

The Ecology of Language Evolution (Cambridge Approaches to Language Contact)

Grammaticalization of the Complex Sentence: A Case Study in Chadic (Studies in Language Companion Series)

Mental States: Volume 2: Language and cognitive structure (Studies in Language Companion Series)

NMR Studies of Translational Motion

From Linguistic Areas to Areal Linguistics (Studies in Language Companion Series, Volume 90)

From Particular to General Linguistics: Selected Essays, 1965-1978 (Studies in Language Companion Series, V. 3)

Recommend Documents