Cognitive Linguistics Research 4 Editors Rene Dirven Ronald W. Langacker
Mouton de Gruyter Berlin· New York
A Geograph...
14 downloads
595 Views
47MB 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
Cognitive Linguistics Research 4 Editors Rene Dirven Ronald W. Langacker
Mouton de Gruyter Berlin· New York
A Geography of Case Semantics The Czech Dative and the Russian Instrumental
Laura A. Janda
1993 Mouton de Gruyter Berlin· New York
Mouton de Gruyter (formerly Mouton, The Hague) is a Division of Walter de Gruyter & Co., Berlin
@) Printed on acid-free paper
which falls within the guidelines of the ANSI to ensure permanence and durability
Library of Congress Cataloging in Publication Data
Janda, Laura A. A geography of case semantics: the Czech dative and the Russian instrumental/Laura A. Janda. p. em. - (Cognitive linguistics research; 4) Includes index. ISBN 3-11-012672-9 1. Grammar, Comparative and general - Case. 2. Semantics. 3. Cognitive grammar. 4. Czech language - Case.' 5. Russian language Case. 6. Russian language - Grammar, Comparative Czech. 7. Czech language Grammar, Comparative Czech. I. Title. II. Series. P240.6.J36 1993 415-dc20 93-8324
Die Deutsche Bibliothek - Cataloging in Publication Data
Janda, Laura A.: A geography of case semantics: the Czech dative and the Russian instrumental/Laura A. Janda. - Berlin; New York: Mouton de Gruyter, 1993 (Cognitive linguistics research; 4) ISBN 3-11-012672-9 NE:GT
© Copyright 1993 by Walter de Gruyter & Co., D-10785 Berlin. All rights reserved, including those oftranslation into foreign languages. No part ofthis book may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopy, recording, or any information storage and retrieval system, without permission in writing from the publisher. Printing: Werner Hildebrand, Berlin Binding: Dieter Mikolai, Berlin Printed in Germany
Contents ix
Acknowledgements Preface
.
xi
Part I - Theory
o.
Introduction
3
1. 1.1 1.1.1. 1.1.2. 1.1.3. 1.1.4. 1.2. 1.2.1. 1.2.2. 1.2.3. 1.2.4. 1.3. 1.3.1. 1.3.2. 1.3.3.
What cognitive semantics is. . . . . . . . . . . . . . . . . . Support for cognitive semantics from other disciplines. Psychology . . . . . . . Neurobiology . . . . . Artificial Intelligence. Summary . Case as a semantic entity. Case lacks formal autonomy. Case bears subjective as well as objective meaning. Case meaning and lexical semantics. . . . . . . . . . . Summary . What a cognitive description of case should look like . Schematic representation of case meaning. Network structure . A test of the cognitive model. . . . . . . ..
4
2. 2.1. 2.1.1. 2.1.2. 2.1.3. 2.1.4.
Cognitive semantics compared with other descriptions of case Case semantics in previous traditions. . . . . . . . . . . . . .. Inventories of uses (Mrazek, Potebnja) . Semantic features (Hjelmslev, Jakobson, van Schooneveld). Semantic metalanguage (Wierzbicka) . Semantic/syntactic dichotomy (Kurylowicz, Isacenko) . Case as a purely syntactic phenomenon (Chomsky)... Symbolic logic (Sorensen, Mel'cuk) . "Deep" case (Fillmore) . Localist theories of case (Anderson, Freidhof).
2.1.5. 2.1.6. 2.1.7. 2.1.8.
7 7 8 9 10 10 10 11 13
15 15 15 17 24 25 25 25 26 27 27
29 30 31 32
vii
vi
2.2. 2.2.1. 2.2.2. 2.2.3. 2.3. 2.3.1. 2.3.2. 2.3.3. 2.4.
How the cognitive model has been implied by various authors Relatedness of meanings . Hierarchical relations among meanings. . . . . . . . . . . . . . . Relational invariance . . Advantages of the cognitive approach A departure from empiricism . Both diversity and unity of case meaning are accommodated Many of the best insights of various scholars are incorporated An ideal description of case: Skalicka's challenge .
33 33 36 37 37 38 38 38 39
Part II - Analysis
3. 3.0.1. 3.0.2. 3.0.3. 3.1. 3.1.1. 3.1.2. 3.2. 3.2.1. 3.2.2. 3.2.3. 3.2.4. 3.3. 3.4. 3.4.1. 3.4.2. 3.5.
Why the Czech dative and the Russian instrumental? . 43 . 43 Schemas in pictorial and prose notation Indirect object serves as the prototype of the Czech dative .. 47 Cognitive network of the Czech dative . 52 Schema 1 - indirect object. . 54 Schema 1 a - governed dative . 69 Schema 1 b - dative in impersonal constructions . 79 . 81 Schema 2 and variants - free dative Schema 2 a - dative of beneficiary - transition to indirect object . 90 . 91 Schema 2 b - intransitive expressions of free dative . 92 Schema 2 c - free dative in copular sentences Further syntagmatic variants of schema 2 . 93 Dative network bound by syntagmatic variants of the schemas 95 . 97 Reflexive uses of the dative Reflexive network . 97 Reciprocal network . 107 Semantic extension via mapping - pragmatic uses ofthe dative 110
4. 4.1. 4.1.1. 4.1.2. 4.2. 4.3.
Comparison of cognitive networks - the Russian dative. Schema 1 - indirect object. . . . . . . . . . . . . . . . . . . . Schema 1 a governed dative. . . . . . . . . . . . . . . . . . Schema 1 b - impersonal dative. . . . . . . . . . . . . . . . . Schema 2 and variants - free dative. . . . . . . . . . . . . . Dative reflexives. . . . . . . . . . . . . . . . . . . . . . . . . . .
.. .. .. .. .. ..
113 113 116 128 130 134
5. 5.0.1. 5.0.2. 5.0.3. 5.1. 5.1.1.
5.1.2. 5.1.3. 5.1.4. 5.2. 5.2.1. 5.2.2. 5.2.3. 5.3. 5.3.1. 5.3.2. 5.3.3. 5.3.4. 5.4. 5.4.1. 5.4.2. 5.5.
. Analysis of the Russian instrumental Overview of the network. . . . . . . . . . . . . . . . . . . . . . . . Instrumental network paradigm. . . . . . . . . . . . . . . . . .. Summary of the instrumental's semantic role . Conduit instrumental - schema 1 and syntagmatic variants. Schema 1 a - instrumental of instrument and agent, verbal government . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Schema 1 b - instrumental of inexplicable force . . . . . . .. Schema 1 c - instrumental of sensation . Subnetwork of schema 1 and the paradigm ofdisappearing participants . . Schema 2 - instrumental of setting Instrumental of space. . . . . . . . . . . . . . . . . . . . . . . . . . Instrumental of time . . . . . . . . . . . . . . . . . . . . . . . . . . Summary of the instrumental of setting . . . . . . . . . . . . . . Attributive instrumental . . . . . . . . . . . . . . . . . . . . . . quantitative.......... . Attributive instrumental qualitative........... ', . . . Attributive instrumental Attributive instrumental - comparative . . Summary of the attributive instrumental . . . . . . . . . Instrumental used with a preposition. . . . . . . . . . . . . . . . Comitative instrumental. . . . . . . . . . . . . . . . . . . . . . . . Proximate instrumental . . . . . . . . . . . . . . . . . . . . . . .. Instrumental network bound by paradigm of peripherality and by alliance. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
139 139 139 142 143 150 162 162 164 164 166 167 170 170 173 175 179 182 182 182 185 188
6. 6.1. 6.2. 6.3. 6.4.
The Czech instrumental . 191 Conduit instrumental. . . . . . . . . . . . . . . . . . . . . . . . .. 191 Instrumental of setting . 195 Attributive instrumental . . . . . . . . . . . . . . . . . . . . . . 196 Instrumental with a preposition . . . . . . . . . . . . . . . . . . . 197
7.
Concluding remarks and possible universals . . . . . . . . . .. 199
Notes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 203 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 218
Acknowledgements The initial research for this project was funded by Fulbright Grant number 8545620. Additional funding for the research and for preparation of the manuscript was provided by the University of Rochester. I would like to thank the many individuals who offered support, criticisms and commentary on drafts of this work: Henning Andersen, Leonard Babby, Greg Carlson, Charles Carlton, Rene Dirven, Michael Flier, Frank Gladney, Ron Harrington, George Lakoff, Ronald Langacker, Brygida Rudzka-Ostyn, Alla Shustorovich, and Charles Townsend. I would also like to thank my colleagues at Charles University and the Institute for Czech Language in Prague who assisted me in my research there: Prernysl Adamec, Alla Bemova, Oldfich Leska, Petr Pit'ha, and Petr Sgall. Much of the research into secondary source materials could not have been completed without the expert assistance of the Interlibrary Loans Staff at the University of Rochester. Likewise, I received invaluable help in preparing the final version of the manuscript from the Consulting Staff of the Faculty Computing Resource Center, and from Sean Singh in particular. Drawings in the figures were executed by Terri Beers. I would also like to express my gratitude to Mouton de Gruyter for making this series possible and for collecting further comments from their anonymous readers. Finally I must thank my family for traveling with me to Czechoslovakia and patiently enduring all the extra hours I put in to complete this book.
Preface
It is sometimes appalling to note how many basic linguistic terms that we use
frequently and with ease lack an operational definition. The ambiguity of concepts such as "subject" and "noun" has motivated the invention of further metalinguistic terminology, such as "subjecthood", "agentivity" and even "nouniness" in the search for adequate circumscription of these supposedly simple terms. What indeed, for example, is an "indirect object"? Most people's first reaction to this question would be to state that an indirect object is the argument of the verb instantiated by John in a sentence like Sally gave John a book. But what does the word "like" mean in this definition, and given such a definition could we sort clausal arguments gathered from natural language data into indirect objects and non-indirect objects? And is there a definition of indirect object that applies to all languages, or is this term to some extent language-specific? Let us suppose that we are dealing with a hypothetical case-marking language (and that this language has none of the exotic peculiarities associated with the English dative-shift). Here is a sampling of three-argument clauses in that language: Sally-NOM gave John-DAT Sally-NOM took John-DAT ('Sally took the book from John.') Sally-NOM bought John-DAT Sally-NOM brought John-DAT Sally-NOM baked John-DAT Sally-NOM introduced John-DAT Sally-NOM explained John-DAT Sally-NOM wrote John-DAT Sally-NOM told John-DAT Sally-NOM paid John-DAT Sally-NOM broke John-DAT ('Sally broke John's arm') Sally-NOM envied John-DAT
book-ACC book-ACC book-ACC book-ACC cake-ACC her friend-ACC problem-ACC letter-ACC her name-ACC bill-ACC arm-ACC his success-ACC
One soon finds the definition given above to be empirically inadequate. In which of the sentences above does John function as an indirect object? Certainly not in all of them. But just which of these sentences are enough "like" the sentence in the
xiii
xii
definition above to contain an indirect object and which are not; where do we draw the line? A closer look at these sentences makes us wonder whether we should restrict our view to three-argument clauses. Indeed there are likely to be many two-argument clauses that contain arguments that are very similar to those in the three-argument clauses. Some examples: Sally-NOM wrote Sally-NOM introduced-REFL ('Sally introduced herself to John. ') Sally-NOM ran-away ('Sally ran away from John.') Sally-NOM paid Sally-NOM told-off Sally-NOM nodded Sally-NOM hurt Sally-NOM helped Sally-NOM believed Sally-NOM ruled
John-DAT John-DAT John-DAT John-DAT John-DAT John-DAT John-DAT John-DAT John-DAT John-DAT
Although our intuition tells us that the common-sense definition captures something essential about the notion indirect object, it lacks necessary specification. The prototypical case has been correctly identified, but we need to know where we can go from there and how far. A theory of the structure of grammatical categories is required, and it is just such a theory that is in the making in this book. Although all of these constructions yield grammatical sentences in Czech, I would argue that most but not all of them instantiate the indirect object, and that the observed variety among those that do results from systematic extensions of the prototypical case instantiated with the verb meaning "give". The question of what is and is not an indirect object becomes all the more pressing when we expand our view to all case usage, for indirect objects must be successfully distinguished from other constructions, including those containing dative-governing verbs. As I will show in Part two, all members of the category of indirect object are related to the central prototype by relationships of synonymy, antonymy (allowing the verb meaning "take" to participate in indirect object constructions), and metonymy (allowing the direct object to be named in the verb and thus absent from surface structure). This gives the category a clear, operational definition, making data analysis straightforward and unproblematic.
A quick glance at dative constructions in the languages of central and eastern Europe demonstrates the need for a certain amount of elasticity in the definition. Languages in the northeastern part of this territory (Russian and Ukrainian) do not admit extensions of the indirect object category via antonymy, and thus lack dative constructions for verbs meaning "take" and further extensions based on this one, whereas languages to the south (German, Czech, Lithuanian, Romanian, SerboCroatian, and Greek) have a more fully extended category. Clearly there must be room in the theory for the growing and pruning of peripheral branches in the development of individual languages. This book will outline the specifics of a theory of grammatical categories, as applied to the use of case in two Slavic languages. In so doing it will take up issues of: -
what determines the shape of grammatical categories to what extent diachronic development is reflected in synchronic structure syntactic and semantic uses of case and their roles in case categories the role of concepts such as passivization, subject and indirect object in the structure of case categories.
Part I -
Theory
O. Introduction In 1980, Anna Wierzbicka wrote in The case for surface case that "cases have
fallen on hard times". Indeed, in the decades that have passed since Roman Jakobson's landmark essay "Beitrag zur allgemeinen Kasuslehre" (1936 [1971]), there has been no great leap forward, although new angles have been sought by Fillmore (1968), Anderson (1971) and others. Cases have largely remained as inscrutable as they were to Bloomfield and Jespersen.! In part, they are themselves the cause of this stagnation, for cases are inherently difficult to investigate, as observed by Brecht and Levine:
"there is a general consensus that the category of case is more resistant to analysis than other grammaticized semantic domains like gender, number, person, and tense. The reason for this, it would seem, is that while these latter grammatical categories can be more or less readily identified with certain pragmatic categories of the real world, it is much more difficult to find real world correlates for the putative referents of case."2
There is, however, growing evidence that the root of the problem runs much deeper, that it in fact lies at the foundation of our theoretical framework. Investigators researching parallel themes in the fields of linguistics, psychology, philosophy, neurobiology and artificial intelligence have suggested that there is a fundamental flaw in our twentieth-century empiricist world-view of symbolic logic and set theory that renders it inadequate to account for phenomena of human cognition. A new framework, known in these various disciplines as cognitive grammar, prototype theory, tensor network theory and connectionism, is being put forth. It requires essential changes in the way we conceive of semantics. The aim of the present volume is to test the possibilities of cognitive grammar using data on the dative and instrumental in Czech and Russian. If this framework is indeed better suited to the task of describing cognitive categories, it should not only shed new light on case semantics, but also incorporate the best hypotheses of traditional research in linguistics.
5
1.
What cognitive semantics is
In Neurophilosophy Patricia Churchland discusses the profound influence of one's theoretical framework on the way that data is interpreted, using specific examples from the study of the human brain and cognition. She notes that the invention of symbolic logic by Frege, Russell and Whitehead "changed how people thought about mathematics, about logic, and about language. And it gave new life to empiricism by holding out the vision that the whole of science - even the whole of one's cognitive system - might be systematized by logic in the way that the whole of mathematics was systematized" (Churchland 1986: 252). This theoretical framework has since become doctrine and pervades virtually all work done in the sciences in this century. The implications this framework has for semantics are fundamental and have until recently been accepted without question. For one thing, logical empiricism entails the perfect separation of knowledge of meaning from knowledge of facts (Churchland 1986: 267), implying that the study of semantics must be entirely divorced from the study of other linguistic phenomena. In addition to being perceived as autonomous, meaning has been presumed to be composed of indivisible building blocks, symbols to be manipulated, symbols with no shape or structure more complicated than that provided by set theory. These assumptions form a common thread that runs through much of linguistic theory of the past fifty years.3 Research on cognition carried out by psychologists in the 1970s4 indicated that the presumed model could not account for the way in which human beings store and access meaning. Based on these findings and on further work carried out on natural language, a group of linguists began to develop a new theory of semantics which has come to be known as cognitive semantics.f The fundamental concepts involved are presented in Lakoff 1982 and 1986. Briefly, Lakoff states that meaning is organized in cognitive categories which have a network structure. At the center of a given category is a prototype member of the category. Other members are placed in the category according to their relationship to the prototype, which may be very close or peripheral, thus giving the category a radial structure. As a result, rather than an unanalyzable bounded set of presumably homogeneous members, cognitive semantics provides for a hierarchically structured network of interrelated members, joined by their relationships to the prototype. Langacker (1987: 49) summed up the goals of cognitive semantics quite concisely: "the only appropriate basis for natural language semantics is a subjectivist theory of meaning that successfully accommodates conceptualization, cognitive domains and the various dimensions of conventional imagery."
Crucial to an understanding of radial categories is the fact that "the central case does not productively generate all [the] subcategories [Le., peripheral members]. Instead, the subcategories are defined by convention as variations on the central case ... and have to be learned" (Lakoff 1987: 84). Cognitive semantics therefore, unlike approaches that invoke a core plus rules, does not aim for absolute predictive power; its goal is instead to seek the principles which motivate the structures of extant categories in natural languages. It follows, therefore, that one of the important claims of cognitive semantics is that languages contain entire structured categories, rather than cores plus rules for generating such categories. This means that a given category is an integral part of a language, and not constantly and predictably generated. Without this claim it would be impossible to explain why the "same" category (i.e., one based on the same prototype ["core"] and having the same principles linking members ["rules"]) would vary from language to language, a fact that will be demonstrated in the analysis. A second essential characteristic of cognitive categories is that they are generally not built from primitives, but are defined via the prototype, which is itself a gestalt. 6 This does not mean that there are no "parts" to a prototype, but rather that the perception of the parts is secondary and must be consciously guided, whereas the perception of the whole is direct. An example of a radially-structured category is the English word mother (as explicated in Lakoff 1987: 83). At the center of the category stands the prototype, which defines mother as a woman "who gave birth to the child, suplied her half of the child's genes, nurtured the child, is married to the father, is one generation older than the child, and is the child's legal guardian". The category mother, however, also contains many non-prototypical members that stand one or more steps removed from the prototype because they lack one or more of the characteristics of the prototype listed above. Thus there are mothers who did not give birth - step-mothers, adoptive mothers, and foster mothers; mothers who did not nurture - birth mothers and surrogate mothers; and mothers who are not married to the father - unwed mothers. Modern science has produced an extreme example of mother, the "genetic mother" (who has only contributed an egg, but never bore or nurtured the child) that is a relatively peripheral member of this category. The very fact, however, that a genetic mother is referred to as a mother, demonstrates the way in which relationships to the prototype bind the network in a radial structure. The presence of category members that share no overlapping characteristics (such as birth mother vis-a-vis foster mother) rules out a feature analysis. All of these groups of women can be called "mothers" only by virtue of their relationship to the prototypical mother.
-'
6
The implications of the cognitive model for semantics are far-reaching, for this theoretical framework makes it possible to perform coherent and penetrating semantic analyses without losing sight of the integrity of a given category, and also to discuss why a category has the members it does. Under a set-theory approach one could only list subsets, and since there is no formal device for showing their interrelationships, this has produced fragmentary accounts of categories." in which one literally cannot see the forest for the trees. The only alternative was to view the undifferentiated sets as wholes, which could only be characterized in vague terms that denied the actual variety of members they contained.f By positing an internal structure for categories, cognitive semantics avoids the dilemma of choosing between unity and diversity presented by set theory. It enables the linguist to seek as much detail in his description as desired without endangering the integrity of the category. The network may become increasingly intricate, but by virtue of the fact that its structure is based on interrelationships, constant reference is made to the prototype and those members closest to it. Also, the use of conventional imagery to caption the kinesthetic image-schernas? associated with a category produces descriptions which are intuitively satisfying as well as formally elegant. Cognitive semantics is, of course, new and has to date a limited number of subscribers, yet a number of linguists who do (or did) not work within this framework have made statements indicating that they reject a symbolic logic/set theory approach and some have even advocated a network structure for meaning. 10 Sgall et al. (1986: 10) affirms that "linguistics cannot be reduced to a part of mathematics, since there is a major difference between natural language and the formal languages of logic." Potebnja ([1958]: 70, 431) agrees that symbolic logic is not adequate to describe grammatical and semantic relations and argues that it is the linguist's task to "posit formal meanings which are more concrete, to make them as distinct as possible and to show their genetic relationships", a statement which suggests that (sub )meanings are interrelated. The authors of Russkaja grammatika [Russian grammar] (Svedova 1982: 479) are even more specific in describing the structure of case meaning. They claim that cases are polysemantic, that each case has its own system of meanings and that some meanings are basic and central, whereas others are semantically peripheral. Likewise Nunberg (1979: 179) comments that "we could ... assume that all of the several uses of a form are connected by a network of referring functions". Plewes (1977: ix) set out from a Jakobsonian framework, and at a more concrete level found structure not accounted for by distinctive features: "the contextual variants of a single invariant combination, while related at an abstract level, may separate into distinct semantic groupings at intermediate levels of abstraction". Schlesinger
7
investigated the semantics of the instrumental in a number of languages and concluded that the meanings comitative, instrument and manner form a continuum. He then pondered the remaining meanings of the instrumental and asked "are cognitive structures constituted of isolated continua, or perhaps these continua are components of a more complicated system?" (Schlesinger 1979: 321). These linguists were obviously groping for a model that would provide the kind of structure that cognitive semantics offers. When the discussion turns more specifically to case and the various theories that have been applied to it, there will be more mention of statements which support the cognitive framework made by linguists who did not work with this model.
1.1. Support for cognitive semantics from other disciplines 1.1.1. Psychology As mentioned above, it was work in psychology that stimulated the development of cognitive semantics in the field of linguistics. In a series of experiments probing human categorization of natural objects, Rosch (1973b: 111) found no support for the way that "psychological and linguistic research has tended to treat categories, as though they were internally unstructured - that is, as though they were composed of undifferentiated, equivalent instances - and as though category boundaries were always "well defined"." Contrary to the tenet of empiricist logic that category membership must be all or none, Rosch's work with human categorization demonstrated that category boundaries are not necessarily definite and that categories are internally structured (Mervis and Rosch 1981: 109). Their structure is based on a gradient of relationships to a prototype, which is encoded as a mental image that represents the category as a whole. 11 It should be noted that the term mental image may refer to something considerably more abstract than a picture, as noted by Shepard (1978: 130), who has also suggested that the neural representation of an image probably reflects the structure of the image, a hypothesis which has been confmned by research in neurobiology. 12 Cognitive semantics, then, has clearly adapted Rosch's prototype theory without major alteration. The central member of the cognitive category is the prototype and is captioned by conventional imagery, represented in a profile (a representation parallel to a mental image). The structure of the semantic category is determined by the nature of relationships between the profiles of various members and the prototype. As will be seen below, it is also significant that
9
8
Rosch's work (Mervis and Rosch 1981: 104) suggests that categories are processed holistically, i. e., that when reference is made to any member, reference is also made to the whole category, for this notion is supported by research in neurobiology and artificial intelligence. 1.1.2. Neurobiology Patricia Churchland, herself a philosopher, spent several years examining recent developments in the study of the brain in order to discoverwhat implications they might have for theories of meaning and cognition and to construct a "unified science of the mind-brain", termed "neurophilosophy" (Churchland 1986: ix). She found that the results of researchin neurobiology contradictthe "familiar and virtually doctrinalcomputer metaphorfor the mind's representations and computation" (Churchland 1986: 252) on severalcounts.It can no longer be claimed that the brain stores and processes finite bits of information in vast sequences of discrete steps. For one thing, problems of access would become astronomically difficult (Churchland 1986: 395). Also, given what is known about the speed at which neurons fire and the number of steps it takes to work through an ordinary problem (such as reaching for an object) in the linear fashion of a digital computer, it is necessary to predict a processing time several orders of magnitude greaterthan the time whichthe brain actually takes to perform such tasks (Churchland 1986: 35 and Cottrell 1985: 7). Insteadof storing information in "centers", it appears that the brain accomplishes this by using networks of neurons which map the information in the brain. Whenever any part of a network is accessed, activation spreads throughout its pattern. Problem-solving is not carried out by serial computation, but rather by matrix multiplication, i. e., the pattern of activity of one neural net is mapped onto another. This is known as Tensor Network Theory, and provides a relatively straightforward solution for the conversion of vector information, such as that necessary for catching a ball (which involves coordinating the movementof the ball with that of the hand). Scientific evidence for the validity of this theory is provided by studies of brain anatomy and by neuronal experimentation and, in addition, it gives an intuitively appealing account of the operation of the brain. After all, we certainly do not compute vectors when catching a ball, we simply match our perception of the ball's movement with the muscular movements necessary to make our hand intercept its path, or, in other words, we appear to map perceived movement onto muscular movement. The discovery that "neurons are plastic ... their informationally relevant parts.grow and shrink ... [and] this appears to be essential to their functioning as informa-
tion-processing units" (Churchland 1986: 35) has further ramifications for learning. A theoretical framework based on sets and symboliclogic would predict that the brain would simply add more and more bits of information. Tensor Network Theory, however, indicates that learning involves the expansion of neural nets and this entails the automatic integration of new information in established patterns, a process not readilyaccounted for by logicalempiricism. Rosch's work makes plausible a claim of psychological reality for the structure of meaning invoked by cognitivesemantics, and neurobiological research further suggests that it may have physicalreality (based on the anatomy and function of the brain) as well. It would be premature at this point to make either such claim, but it is clear that the fundamental precepts of cognitive semantics are at least compatible with the findings of neurobiologists and psychologists, and in fact they are more compatible than those of recent theories of semantics.
1.1.3. Artificial intelligence Computer scientists have also traditionally worked within the framework of symbolic logic and, as mentioned above, it was their model for information storage and processing that came to be' accepted as the metaphor for brain function. The field of artificial intelligence, which grew out of a desire to mimic the brain's abilitieswith machines, had by the 1970sreached such an impassethat Dreyfus wrote an extensive expose of its failures. In his conclusion he asks, "Is an exhaustive analysis of humanreasoninto rule-governed operations on discrete, determinate, context-free elements possible? Is an approximation to this goal of artificial reason even probable? The answer to both of these questions appears to be, No" (Dreyfus 1979: 303). Dreyfus, however, did not lay the blame with the computers, but rather with the theoretical framework assumed by their programmers. Although many computer scientists chose to reject Dreyfus' conclusions, there were some who shared his contention that fundamental theoretical changes would have to be made before the field could progress. They began to experiment with alternative meansof storing and processing information, by using systemscharacterized by massive parallelism and high connectedness. It is this feature of the resultant movement in computer science which gives it its name: Connectionism. Experiments have been conducted in the application of connectionism to simulation of phenomena as diverse as visual perception and motor control,13 but only one piece of work will be cited here because it deals specifically with case. Cottrell (1985) employed a connectionist model in writinga program which would disambiguate both word sense and case function in
10
English. He cites a desire to imitate real brain function and to avoid the pitfalls of existing theories as his motivation for choosing the connectionist model. 14 Rather than using symbol-passing (the computational equivalent of linguistic features or of discrete bits of semantic information), Cottrell's program is designed to store and access imformation in networks. When a word is referenced, the entire network of its meanings and functions is activated and then allowed to "relax" to a consistent [with other input] interpretation" (Cottrell 1985: 207). Cottrell had his program analyze a series of English sentences and found it to be efficient at this task. In addition, it was found to be comparable to the functioning of the brain in some ways: "One test of the validity of the model presented ... is to evaluate its adequacy at accounting for neurological data .... The model is shown to be adequate for explaining some of the overall effects" (Cottrell 1985: 161).
1.1.4. Summary Both psychology and neurobiology fail to provide evidence that human categorization is organized in a manner consistent with set theory and symbolic logic. In both cases we infer that information is not stored in discrete and unrelated bits, but rather in hierarchized structures which interact as wholes. The application of these principles to computer science has produced a fruitful new movement after years of stagnation, and it seems plausible that this approach would likewise be a powerful tool in advancing our linguistic understanding of semantics.
1.2. Case as a semantic entity
There are three obstacles to the semantic description of case. The first is formal and relatively minor; the second two have to do with the nature of case semantics and will be treated in more detail.
1.2.1. Case lacks formal autonomy No single case has a unique surface representation by means of which it is signaled, a fact that prevents case from fitting neatly into the structuralist schema of "one form, one meaning". Burston (1977: 51) put it succinctly: "a certain oddness remains in calling a "sign" a linguistic element which, whenever it occurs,
11
lacks formal autonomy". She further comments, however, that "this difficulty can be avoided if cases are envisaged not as signs, but as features of grammatical nature which exist only in combination with other features". This fact does not hinder most linguists, who are nevertheless willing to posit meanings for cases, and indeed Kilby (1982: 75) asserted that "any adequate description of the Russian case system must at some point make reference to the meanings expressed by the cases".
1.2.2. Case bears subjective as well as objective meaning A second obstacle is the permeation of case meaning by a subjective element. The fact is that case actually encodes not one, but two basic functions which are themselves related in a non-trivial way. In addition to referring to the role of a substantive in a narrated event, case also encodes the speaker's interpretation of that role. This second function is quite salient in so-called "pragmatic" uses of case, such as the ethical dative, and can be decisive in determining, for example, whether possession will be encoded by the genitive or the dative or whether a predicate substantive will be marked by the nominative or instrumental. The distinction between these two functions is dominant in the work of van Schooneveld, who terms the second of them "transmissional deixis" and attributes it to an observation by Jakobson that distinctive features can be deictic and thereby refer to the speech event itself. 15 An example of the ethical dative illustrates what van Schooneveld has termed "transmissional deixis". The addressee of the utterance in (1) has no relationship to the narrated event, he has neither witnessed it, nor was he previously aware of it, yet the speaker subjectively imposes a dative relationship that holds between the hearer and the narrated event by inserting the dative pronoun vdm "you" into his narrative. (1)
"J 0, a na universite" pokracova Yes and at university-Lac continued potichu pan Keval, "tam se quietly Mr. Keval-NOM there REFL-ACC vdm dnes seprala pnrodovedecka YOU-DAT today fought natural-science-NOM fakulta s historickou." (Capek) department-NOM with historical-INST
12
13
, "Yes, and at the university", continued Mr. Keval quietly, "(hey, you know what?) the natural sciences department had a fight with the history department today." , The notion that case meaning has two functions is upheld by other linguists as well, although most refer to it as an opposition of semantics to pragmatics. These two functions interact with each other in a complex manner. Adamec (1978: 35) , gives an example of a sentence in which the deictic function does more than present the speaker's point of view - it actually forces a different interpretation of the narrated event:
(2)
a.
b.
goroda. On videl ekzoticeskie cities-Ace He-NOM saw exotic 'He saw exotic cities.' Emu videlis' ekzoticeskie Him-DAT saw-refl exotic goroda. cities-NOM 'He saw exotic cities (in his dreams).'
(2a), with an ordinary nominative subject, makes a neutral statement, whereas (2b), which has an impersonal dative construction, can only be interpreted as referring to fantasy, i.e. to imagined or dreamed cities. Furthermore, Langacker (1987: 50) points out that "there are alternate ways of construing a situation, and alternate ways of symbolizing a given construal", suggesting that the deictic function of case is itself non-simple. It appears, then, that case can carry two kinds of meaning that are complex and interact with each other. Given such a situation, one might ask how the linguist is to proceed - should the functions be untangled and dealt with separately? Nunberg (1979: 143) notes that the distinction between these two functions is anything but clear-cut: "the semantics/pragmatics distinction cannot be validated even in principle". Wierzbicka (1980: 144) likewise asserts that there is no "abyss between truth-functional and other "subjective" or "pragmatic" meanings". Van Schooneveld (1978: 216) provides a clue as to why this is so when he affmns that all case features carry transmissional deixis (i.e., that the deictic function is always present). What this means is that case always makes reference to the speaker's interpretation of the narrated event, and that in fact no narrated event is ' encoded directly, unmediated by this interpretation.18
The subjective moment of case meaning is most saliently present in examples of semantic extension via mapping. It involves the application of basic meanings (here, case relationships) to domains other than perceived objective reality. Case relationships can thus be mapped onto subjective realms, such as the speech-act domain, and they can also be subjectively applied to perceived reality. This description of subjective case meaning as subjective mapping of objective (semantic) case meaning accounts for the intimate relationship between these two kinds of case meaning and easily accommodates van Schooneveld's notion of transmissional deixis and the observations of other scholars mentioned in this section. More will be said about semantic extension via mapping and its role in the structure of case categories in section 1.3.2., and abundant examples of subjective application of case meaning will be provided in the analysis.
1.2.3. Case meaning and lexical semantics The third obstacle to a description of case semantics involves the objective function of case meaning. As mentioned above in the Introduction, there is a lack of concrete, real-world correspondence for the relationships encoded by case. It seems that the kind of meaning expressed by case is different from that expressed by other morphemes. Indeed, case belongs to the realm of grammar rather than lexicon, and even by the standards of grammatical relations, it is elusive. While one can point to a real instantiation of gender, for example, it is hard to do this for a case role like the dative. Potebnja ([1958]: 35-37) states that every word contains two discrete kinds of information: 1. lexical and 2. grammatical, but then qualifies this statement by adding that these two kinds of information are inseparably combined in the minds of speakers: "the grammatical form is an element of the meaning of a word and springs from the same basis as its essential meaning". Thus it would appear to be artificial to divide these two kinds of meaning, since they do not differ in essence. In commenting on Jakobson's description of case, Sangster (1982: 114) notes that "if syntactic phenomena are sign vehicles, they should display semantic properties just like any other set of linguistic forms". Langacker (1987: 17) is even frrmer in asserting that "grammar and lexicon form a continuum" . Although Talmy (1986: 1), like Langacker, uses a cognitive framework and does not deny that lexical and grammatical information share a common basis, he views them as "two subsystems [which] have distinct semantic functions, ones that are indispensable and complementary", and thus is able to make several valuable observations about the semantic properties specific to grammatical
15
14
elements. He points out that grammatical, as opposed to lexical, elements are characteristically relativistic, that is, they refer to relative rather than absolute values of magnitude, shape, rate, material and other properties. They use these relativistic specifications to break up an event into parts and participants and thereby "the grammatical specifications in a sentence ... provide a conceptual framework or, imagistically, a skeletal structure or scaffolding, for the conceptual material that is lexically specified" (Talmy 1986: 1). These observations further lead him to "speculate that the cognitive function of such classification [of grammatical elementsinto small, closed classes which are relativisticrather than absolute] lies in unifying contentful material within a single conceptual systemand in rendering it manipulable - i.e., amenable to transmission, storage, and processing- and that its absence wouldrendercontentan intractable agglomeration" (Talmy 1986: 30). Talmy explains not only how the grammatical pole of the informational continuum differs from the lexicalpole, but also why this difference exists: it is crucial to the efficientfunctioning of language. Yet the fact that grammatical elements function more to organize ideas than to specify them does not entail a difference between them and lexicalelementsin basic cognitive structure. Indeed, like units of lexical meaning, "grammatically specified notions can be seen to patternin categories, and the categories, in tum, in integrated systems".19 The hypothesis that grammatical information differs in some ways from lexical information yet forms a continuum with it accounts for some of the intimateways in which these two semantic subsystems interact. Without a framework that allows for overlap between them, it would be hard to explain the following phenomena. For one thing, there appears to be a high correlation between case and noun meaning. Greenberg (1974: 33) tabulated the information given in a Russian frequency dictionary and found that "the distribution of frequency over the cases is far from random in relation to the semantic characteristics of the noun". Thus a given case is associated with certain semantic groups of nouns whose meaning is compatible with that whichthe case role entails. Popova (1970: 95) notes that when the noun's lexical meaning does not correspond to the concrete meanings of the case, metaphor is often necessarily invoked, as with the use of an abstract rather than a concrete noun in the instrumental phrase kormit' obescanijami "nourish with promises". In addition, Greenberg (1974: 34) found that the "polysemy of individual nouns [is] to a great extent disambiguated by case". Conversely it is also possible for the lexical meaning of the noun to determine which submeaning of a case is expressed, as noted by Potebnja, Pit'ha and Levine.20 In both instances we see the grammatical and lexical elements complementing each other as they engage in a semantic tug-of-war. In Russian and Czech, at least, there is evidence that the lexical end of the continuum is
encroaching on the grammatical end. In both languages the use of bare case is losing ground to prepositions.U Indeed, Apresjan (1973: 5) was quite correct in asserting that "there exists a bilateral dependence between the syntactic [i.e., grammatical] and semantic [i.e., lexical] properties of linguistic expressions." Within this framework it is clear that case always bears a meaning function, and the fact that it belongs to syntax does not diminish its semantic freight. As will be shown below, this view directlycontradictsChomsky'sGovernmentand Binding Theory, and any model that treats case as a phenomenon of syntax that can be wholly divorced from semantics. Thus I agree with Wierzbicka(1980: xi) that surface case is never arbitrary. If it were, it would be necessary to assume that case markers are semantically void at least part of the time, a notion which is incompatible with the basictenets of cognitive semantics. 22 1.2.4. Summary To sum up, the following assumptions are inherent in the cognitive framework: (i) (ii) (iii)
(iv)
case is always meaning-bearing case meaning has a constant objective moment that can be subjectively applied case meaning involves the organization of rather than the specification of information case meaning is notessentially different fromlexicalmeaning in structure.
1.3. What a cognitive description of case should look like
The present section will integrate the basic premises set out above in a specific description of how case meaning is to be approached in a cognitive framework. 1.3. 1. Schematic representation of casemeaning The remarks made by Talmy (1986) about the nature of grammatical meaning have already put some restrictions on the kind of description that case should be given:
17
16
(i)
(li)
it should refer to the overall organization of events, giving just their skeletal structure, and it should be relatively abstract, without reference to any absolute measure of magnitude, shape, rate, etc.
Langacker (1987) and Smith (1985 and 1987) have both presented preliminary descriptions of case that are consistent with these principles. As mentioned above, the use of conventional imagery is central to the cognitive model. The relativistic nature of grammatical meaning should not be an obstacle to captioning it imagistically, for "the fact that an internal representation is more abstract than a picture does not entail that such a representation is nonvisual" (Shepard 1978: 130). Langacker (1987: 1-18) devotes considerable discussion to developing an appropriate means of captioning the conventional imagery of case. His approach makes two assumptions about the idealized cognitive models (IeMs) of predications: (i)
(li)
functions of case must be sought in the semantic structure of sentences and distinguishes nine types of sequences of beginning-middle-end ordering which are vaguely correlated to participants in Langacker's action chains.
O~E) setting
human beings conceive of the world as containing discrete objects at specific locations which move and interact with each other through direct physical contact ("billiard-ball model")23 in perceiving these interactions, human beings organize scenes into participants and settings and segment clusters of action into discrete events ("stage model").
By combining these models, Langacker (1987: 5) is able to produce a caption of.the "normal observation of a prototypical action", which includes an agent that acts on a patient. The event is structured in an action chain in which circles (an arbitrary, neutral shape) symbolize participants and an arrow symbolizes interaction (movement of energy).24 Case markers "profile", or focus attention on, specific parts of the action chain. In this prototypical action chain, the "head" (leftmost participant) is the nominative and the "tail" (rightmost participant) is the accusative. The basic unit referred to here is clearly the clause; the prototypical action captioned in Figure 1 corresponds to a prototypical clause with a subject and object. Thus Langacker gives a concrete interpretation of a fundamental concept of cognitive semantics formulated by Talmy (1986: 1): that "a sentence [evokes] in the listener a particular kind of experiential complex, ... a "cognitive representation"." The assumption that a study of case must be founded on sentence structure is not unique to Langacker and Talmy, or to cognitive semantics for that matter. Serensen (1957: 31) states that "the sentence is the unit from which the analysis of case must start". Ulicny (1984: 17) likewise insisted that the relational
viewer
Figure 1. Using his action-chain notation, Langacker (1987: 28) goes on to define six case roles which "are sufficiently fundamental and cognitively salient to be considered archetypal". His article does not examine the details of how cases are realized in specific natural languages, but rather gives general outlines for conducting such an analysis.
1.3.2. Network structure Although Langacker (1987: 39) deals only with universal prototypes, he makes it clear that he is simply outlining the frontiers; that in-depth studies of natural language cases can and should be undertaken according to this model. Smith 1985 brings us one step closer to actually realizing a cognitive description of the semantics of a case. His aim is to provide a unified account of the various uses of
18
19
the dative in German. He suggests the following schemas to caption "the defining property" of the nominative, accusative, and dative in German: 25
~ nom
-0 ace
dat
Figure 2 Figure 2.
dat's sphere of control Smith (1985: 393) has chosen the schema that captions the prototype of the dative because "the dative entity is viewed by the speaker as simultaneously acted upon as well as an actor in its own right", a property that he later (Smith 1987: 455) terms "bilateral involvement".26 In this schema the dative combines the properties of both the nominative (head of chain) and accusative (tail of chain). Smith then goes on to discuss various types of dative constructions in German and the schemas appropriate for them. For sentences which have nominative and accusative as well as the dative, Smith gives the schema shown in Figure 3. This schema is valid both for sentences in which the dative represents an indirect object, as well as those in which it serves other purposes, as in the sentences (dative forms and their equivalents are in roman type): (3)
(4)
Figure 3.
---~
nom
dat
Fritz offnet derDame die Ttlr, FritZ-NOM opens the lady-DAT the door-Ace 'Fritz opens the door for the lady.' Der Arzt hat mir das Leben gerettet. The doctor-NOM has me-DAT the life-NO saved 'The doctor saved my life.'
Smith (1985: 394) notes that verbs that require the dative always contain the implication that the dative entity is the agent of some further action. In considering the verb helfen "help", for example, "a moment's thought will convince one that when one is helped, one is helped to do something, to effect some further (unspecified) action". Smith uses the schema shown in Figure 4 to represent the dative governed by verbs, in which the dotted lines symbolize an unspecifed implied action.
unspecified implied action
Figure 4.
Finally, Smith discusses the dative used in impersonal constructions, as in (5).
nom
dat
Figure 5.
20 (5)
21
Es it-NOM '[ am cold.'
ist is
mir me-DAT
kall. cold
He proposes that "we can conceive of the abstract force [here, coldness - LAJ] as exerting an effect on the dative [entity] which in turn causes an internal physical reaction within the [dative entity]" (Smith 1985: 396). Smith posits the schema found in Figure 5 to caption this type of dative. The internal reaction is symbolized by a squiggly arrow in the dative landmark. Smith has put forward a series of schemas as well as a prototype for the dative. He does not outline the network that they comprise, but the notion that the submeanings should be clustered in a network structure is implicit in cognitive semantics and in other models of description as well. Vinogradov (1972: 142) stated that the meanings of the Russian dative form a whole, integrated system. In describing the uses of the Polish dative, Wierzbicka refers to its meanings as a "family" (Wierzbicka 1986: 419), and as we shall see in chapter 2, her treatment of the Russian instrumental (Wierzbicka 1980) implies a specific network structure, as does Mrazek's (1964). In a single terse sentence, Langacker (1987: 39) indicates how the schemas suggested by Smith can be arranged to form a network structure: "a case category is generally complex, comprising a network of alternate senses connected by relationships of schematicity and semantic extension". This concise description of the relationships that structure case categories is worth further scrutiny, for it contains an essential insight from which all the specific types of relationships involved can be winnowed. "Schematicity" identifies the abstract way in which the members of a category are united. "Semantic extension" refers to the system of definite positions which individual schemas occupy in the network. In conducting the extensive investigation presented in the analysis I have identified the following classification of subtypes of semantic extension: I.
Variations in the shape of schemas a) b)
paradigmatic variation - variation in the portion of the schema which represents the case-marked entity syntagmatic variation - variation in the composition of the event chain (removal of participants, replacement of the verb with a copular construction)
II.
Metaphoric extension a) b)
extension to different semantic fields - through relationships of synonymy, antonymy, and metonymy extension to different domains - through mapping of relationships onto the domain of subjective perception or the speechact domain.
The entire analysis is devoted to the presentation of examples of these types of variation, so I will give only brief explanations here. Relations of schema shape can be defined as topological variations in the event chain, and they follow two axes, the paradigmatic and the syntagmatic, involving variations either in the role of the case-marked entity or in the syntactic environment in which it finds itself, respectively. Smith's schemas, which appear in Figures 3 and 4, illustrate paradigmatic variation. In Figure 3 the characteristic mark of the dative is its sphere of control, whereas in Figure 4 the dative is recognized as an agent of an unspecified implied action. Syntagmatic variants are produced by alterations in the length and composition of the event chain with the characteristic mark of the profiled case held constant. An example would be a variation on Figure 3 realized by an intransitive verb, so that the accusative entity is removed from the schema and the verb interacts directly with the dative's sphere of control. The effect is equivalent to the difference between the two English sentences He read a book to me and He read to me. Metaphoric extension may project the relationships captioned by a schema onto a different semantic field, either by making use of relationships of synonymy, antonymy, and metonymy, or by mapping case relationships onto the domain of the speaker's subjective perception of the speech-act domain.I? The Czech indirect object provides numerous examples of the first kind of extension. Narrowly defined, the indirect object is the recipient of the verb ddt "give (to)" . Synonymy effects the spread of this case usage to clauses with other verbs that indicate the bestowal of possessions, such as koupit "buy (for)", poslat "send (to)". Antonymy further extends this usage to verbs having the opposite meaning, i.e., those which indicate the removal of possessions, e.g., vzit "take (from)", krdst "steal (from)". Extension via metonymy occurs with verbs which connote the giving of signals, money, benefit and harm, all of which take the dative in Czech, in which the direct object is already subsumed by the verb and therefore not overtly expressed. The ethical dative in (1) above illustrates the last type of variation, via mapping onto the speech-act domain. Here the speaker invokes a
22
sphere of control on the part of the hearer and makes the claim that the narrated event is in the hearer'ssphere. The four types of variation are presented in the order of their prominence in structuring case categories. Paradigmatic variants account for the basic structure of the network. Syntagmatic variants are local variations on a given paradigmatic variantand, because the syntagmatic variants are frequently syntactically parallel, they often help to build links between the paradigmatic variants. The two types of extension also produce localsubvariants, eitherof a paradigmatic variantor of one of its syntagmatic counterparts. Thus the relationships which structure case categories are arranged hierarchically, such that each successive type produces local variations on the resultsof morebasictypesof variation. Governed case in the cognitive model As Smith's work suggests, governed case (specifically the dative governed by verbs) is not to be dismissed as a mere syntactic phenomenon, for it is part of the semantic whole represented by the network of a given case. Indeed, it would be contrary to the spirit of cognitive semantics to leave any uses of case out of the scope of the description, since this would imply that they have no semantic role, that part of their network is defective, void. The importance of case semantics in verbalgovernment is supported by many scholars, and the verbs associated with this phenomenon are traditionally listed in semantic groups.28 In examining the difference between oblique and accusative objects, Plewes (1977: 100) affirms that "the syntactic processes remain the same; the difference is one of semantics". Gorbacevic (1971: 238) found that the significance of semantics is so great that it has caused some Russian verbs to switch their government. Porazat'sja "be struck (by)" governed only the instrumental in Tolstoy's time, but now, due to semantic analogy with dative-governing verbs such as udivljat'sja "be surprised (at)" and izumljat'sja "be amazed (by)" it appears most frequently in association with the dative. Even Popova (1974: 178), who at the outset of her study of verbal case government states that "the governed uses are semantically unmotivated and are maintained onlyby tradition", admitsin her conclusion that deviantuses of the dative and instrumental arise due to contamination from verbs having a similar meaning. She cannot deny the influence of semantics on case selection, even though it leadsher to an inherentcontradiction. If it is recognized that semantics contributes to the selection of cases governed by verbs, it stands to reason that the use of case with prepositions should also be included in the semantic network. Indeed, many scholars of various orientations
23
agree with Kurylowicz, Isacenko and Vinogradov that prepositions serve to enhance or extendcase meaning, ratherthanreduceit to an entirelysyntactic phenomenon.j? It was this observation that inspired van Schooneveld (1978: 219) to hypothesize a single "cardinal semantic pattern" for bothcases and prepositions. If case governed by prepositions is to be included in a semantic description, then, we might ask, how shall we place it in the network? Should it occupy a special part of the network, or perhaps have a separate, related network of its own? Research on trendsin prepositional usage suggeststhat these questions are inappropriate. El'zbutas (1967) found that in Russian some phrases that previously used the bare dative nowrequire prepositions and conversely, some expressions that used to contain prepositions now appearwithjust the dative. Apparently the line which we commonly assume to divide prepositional from prepositionless case is indistinct and crossable: prepositional case interactswith the semantic network, although its role is perhaps more elusive. As noted by Isacenko (1965: 100), the fact that a preposition of necessity always brings its own semantic freight to an expression, and yet does not have an independent existence in isolation from case, makes it impossible to determine the exact semantic contribution of case in prepositional phrases. An analysis of the use of prepositions will not yield anything aboutthe "general"or "basic" meaning of cases. Thus, in cognitive terms, prepositional case belongs in the semantic network of a given case, but its placeis usually pertpheral.F' Networks and linguistic universals Questions concerning the status of the network also present themselves. What universals, if any, will it reveal? Dahl (1985) providesguidelines for answering this and related questions. He recognizes two approaches to the search for language universals: that of the naive universalist, who seeks universal categories which can be attributed wholesale to any language, and that of the sophisticated universalist, who threshes out universal principles. This implies that what we should be seeking is a set of universal strategies in the organization of case categories rather than universal definitions of given cases. A hypothesis of this sort is implicit in Langacker's (1987: 26) work, for he presents case role "archetypes" which he applies to various languages, yet asserts that "the search for all-encompassing schematic characterizations would not appear promising; case semantics is better approached in terms of language-specific families of senses organized around prototypical values". Wierzbicka (1986: 386) opens her discussion of the Polish dative with an assumption that bears a striking
24
resemblance to this hypothesis: "a case has one core meaning, on the basis of which it can be identified cross-linguistically ..., and a language-specific set of other, related meanings, which have to be specified in the grammatical description of the given language". The hierarchy of variations suggested above provides a first approximation toward a set of universals of case semantics. The implications of this hierarchy and the specific universal predictions it entails will be discussed in the conclusion, after the full weight of data has been brought to bear upon them.
1.3.3. A test of the cognitive model The aim of this book is to take the precepts of cognitive semantics to their logical conclusion; to put them to the test, with semantic data from two of the most intractable problems of Slavic linguistics: the semantics of the dative and instrumental cases in Czech and Russian, respectively. In addition to presenting a prototype and related schemas for all of the various meanings of these two cases, I will show how they are arranged in networks and state the relationships among them that give the networks their structure. The discussion will focus on the Czech dative (as compared with its Russian counterpart), and the Russian instrumental (as compared with its Czech counterpart). The comparisons will serve to show how slight network variations account for the differences in case usage observed in natural languages. This is intended as a contribution to the description of the Slavic case system as well as a test of the validity of the cognitive model.
2.
Cognitive semantics compared with other descriptions of case
In this chapter the cognitive model will be compared with other models that have been used to account for case usage. In addition, it will be demonstrated that many scholars have made statements which are consistent with the cognitive model, even when this entailed contradicting their own assumptions. The fact that the cognitive model can accommodate both the major contributions of most scholars as well as such apparent contradictions suggests that it is the model of choice, and indeed, it might have been chosen by most of the authors cited herein if it had been available to them when they were conducting their research.
2.1. Case semantics in previous traditions
For the sake of the present discussion the following remarks are necessarily brief and touch ony on relevant points. They are not complete summaries of the frameworks they present.
2.1.1. Inventories of uses (Mrazek, Potebnja) Perhaps the most exhaustive treatments of case have taken the form of lists of uses. Some examples of this approach include the descriptions of the Russian instrumental proposed by Mrazek (1964) and Potebnja ([1958]), as well as the inventories commonly encountered in grammars. Aside from thoroughness and attention to detail, however, this method does not have much to recommend it. It involves little more than the compilation of data. Such list-writing is superlatively atomistic and is generally frowned upon because "no synthesis follows the classification. The method consists, in fact, of splitting up into several parts what is a complex linguistic unit" (Burston 1977: 30). Thus the integrity and internal order of the semantic category are overlooked. Commentaries on the common origins or relatedness of meanings are often sprinkled in on an ad hoc (and inconsistent) basis in order to fill out the description and to reduce its choppiness, but these constitute improvements in presentation rather than alterations in theoretical framework. As mentioned in the previous chapter, the network structure of a cognitive description facilitates captioning of the unity and infrastructure of a category without
27
26
sacrificing any of the detail sought in the list model. It also solves one of the problems inherent in the list model, that of answering the question, how much detail is enough? Depending upon how rigorously it is applied, the list model will produce any number of submeanings. This possibility for variation is problematic for an approach the express purposes of which are to discover how many submeanings are present and to describe them. Such variation is easily incorporated in the cognitive model, since it strives instead to draw a map of semantic structure, which can be drawn to whatever scale desired, thus allowing for greater or lesser detail.
discussion of feature-marking theory that semantic features like [+direction] or [+peripheral] are either present or absent (i.e., they are digital ), it is more likely that such feature-marking is analog (i.e., the features can be present in varying degrees or strengths)". There are indications, however, that Jakobson intentionally left room for such hedges, and that the apparent immutability of his system is more an artifact of the way his followers interpreted it than of how he actually conceived it.33 Necessary compromises to this approach are also predicted by the cognitive model, for when absolute invariant semantic features were posited, they created monolithic case categories with no infrastucture. Jakobson's concept of relative invariance brings his system very close to the sort of description required by cognitive semantics, and will be discussed toward the end of this chapter.
2.1.2. Semantic features (Hjelmslev, Jakobson, van Schooneveld)
Hjelmslev (1935-37) theorized that each case has a single abstract meaning, derived from a matrix of semantic features. His work inspired Jakobson's (1936 [1971]) "Beitrag zur allgemeinen Kasuslehre", which has been a subject of vigorous debate ever since. Perhaps the most ambitious author in this regard is van Schooneveld (1978), who sought a single grand semantic pattern for all morphemes that serve to structure Russian sentences. The fact that Burston (1977) succeeded in applying the very same system of features which van Schooneveld worked out for Russian to a language with a very different case system (Pali) is enough to suggest that something is amiss. It appears to be too easy to twist universal invariants around to fit the data and, conversely, this sort of approach fails to account for variation among languages. This approach has been roundly criticized, for both excessive vagueness and excessive power, and also for lack of verifiability.U Both Skalicka (1950: 135) and Isacenko (1965: 95) find this sort of description elegant, but far too abstract to be of practical value. Isacenko asks how a general meaning such as the one Jakobson assigned to the dative - "expresses the existence of the object as independent of the action" - could possibly account for meanings such as those expressed by dative subjects in expressions like mne nado "I need (lit: needed to me)" and mne xocetsja "I feel like (lit: wants-refl to me)" .32 Wierzbicka (1980: xv-xvi) sums up the major objections: "formulas like '[+peripheral, -affected]' are not self-explanatory and one could stretch them in many different ways to make them fit the facts; these formulas have limited predictive power [and are] too vague, too general to be empirically adequate." Even scholars who use Jakobson's system often criticize it and find themselves forced to hedge on some of its basic assumptions. Plewes (1977: ,263264) comments in a footnote that "although we have implicitly assumed in the
2.1.3. Semantic metalanguage (Wierzbicka)
Wierzbicka clearly recognized the drawbacks of the preceding two models: that lists were too atomistic and features too abstract. She attempted to build a fresh model to eliminate these problems34 by implementing a semantic metalanguage in her description. This would eliminate the need for features, and the expressions set in metalanguage would be self-explanatory and specific to given submeanings, rather than overly vague. Also, similarities in definitions of submeanings would point to their interrelatedness, reducing the atomistic appearance of the list of submeanings. She applied her method to the Russian instrumental, producing metalanguage defmitions of seventeen submeanings. This was a bold attempt which incorporated several of the goals of a cognitive description: it strove to reconcile the diversity of meanings with their overall unity and to demonstrate their relatedness. Hints to the structure of the category of the instrumental, however, remain diffusely scattered through the metalinguistic definitions. Relationships among submeanings are implied, but not made explicit. Still, Wierzbicka's monograph is quite significant, for it presents a very clear statement of the problems involved and provides a useful tool for the semantic description of case categories.
2.1.4. Semantic/syntactic dichotomy (Kurylowicz, Isacenko)
Another approach that has had many followers suggests that case is really two distinct phenomena which should be treated separately: semantic case and syntactic case. This distinction is clearly derived from the observation that
28
grammatical meaning differs from lexical meaning, discussed as "the third obstacle to a description of case semantics" in the preceding chapter. Kurylowicz (1960) outlined this theory,35 and even though he did not subject it to rigorous verification, it is evident that he found the syntactic and semantic components more difficult to disentangle than one might presume. Each case was stated to be either primarily syntactic (grammatical) or semantic (concrete) and to have both primary and secondary functions based on this very same distinction. In addition, the concrete cases were said to be progressing toward grammaticalization, whereas the grammatical cases were moving in the opposite direction, toward adverbialization (lexicalization). And it was claimed that concrete case occupies an intermediate position between adverbs (lexical meaning) and purely syntactic forms. Thus, even though the theoretical purpose of Kurylowicz's approach is to identify a bifurcation of semantic and syntactic functions, we observe a tendency to emphasize the relationships between these two functions as poles of a spectrum of case values rather than as a concerted effort to draw boundaries between them. 36 Kurylowicz did not, as Apresjan (1964: 33) points out, give an operational definition for distinguishing the two types of case. In spite of the lack of resolution in the supposed semantic/syntactic interface, a number of linguists have followed Kurytowicz's lead. Isacenko (1965) applied this approach to Russian, and identified many of the primary and secondary functions of the cases of that language. His work is, however, hardly an exhaustive treatment of Russian according to this model, and also contains many insights not derivable from this distinction. DeGroot (1965) makes use of the same distinction and gives an equally generalized account. Deserieva (1974) tallies up all of the semantic and syntactic uses of each case in Russian and then puts these values in a formula by means of which she arrives at figures describing the abstractness of each case. The nominative is, according to her, 100% syntactic and therefore very abstract, whereas the dative, instrumental and locative are quite concrete, as predicted by Kurylowicz. Although Deserieva is the only scholar to rigorously implement Kurylowicz's classification, it remains just that - a classification of uses into two broad categories, for she does not contribute anything new to the theoretical framework. It is interesting to note, however, that when she has identified a particular use as semantic or syntactic, even this designation is not necessarily absolute. Seven of the syntactic uses of the genitive are "partly semantic" and she admits that there are crossovers between the two categories. Perhaps the most recent revival of the syntactic/semantic dichotomy is to be found in Babby (1986), who uses it to show that Chomsky's view of case is grossly unbalanced, ignoring its semantic uses.
29
There have been a number of attempts to isolate the syntactic function of case, which inevitably disintegrate into unwieldy and unenlightening inventories. Case uses that are essentially identical are assigned to different categories because of minor details of syntactic structure and conversely uses that are significantly different are lumped together whenever they show identical structure. This is true of the accounts found in Worth 1958 and in the 1970 and 1982 Academy grammars (Svedova et al. 1970 and 1982). It must be noted that both Academy grammars resorted to semantic captions for syntactic categories, again suggesting that the two functions are invariably intertwined. Talmy's work, as cited in section 1.2. of this monograph, shows that the assumption underlying this type of approach - that grammatical meaning is different from lexical meaning - is both true and false. Those who subscribe to the semantic/syntactic dichotomy are correct in seeing that the two kinds of meaning are not identical, but incorrect in pushing this distinction too far. Their meanings can be accessed by the same methods, although they tend to focus on different aspects of the semantic structure of sentences. It is the failure to recognize this subtle yet crucial fact that has permitted the development of this approach. Kurytowicz's deliberate omission of specific boundaries amounts to a hedge on his basic theoretical outlook, and such hedging is predicted by the cognitive model. There is one other, yet related, way in which the semantic/syntactic dichotomy approach is incompatible with the cognitive model. As Oliverius (1972: 104), points out, Kurylowicz's model must be rejected because it requires one to accept the existence of empty morphemes. Kurytowicz implies that the syntactic uses have no meaning, a corollary to his theory which is also derivable from his failure to recognize the difference between grammatical and lexical meaning. Grammatical meaning is different, but not non-existent. As noted above in section 1.2.3., the basic principles of the cognitive model suggest that a single morpheme cannot be meaningful in some contexts but devoid of meaning in others.
2.1.5. Case as a purely syntactic phenomenon (Chomsky) Chomsky is mentioned here not so much because he has made any special contribution to our understanding of case - in fact, he has said remarkably little about case 37 - but rather because his theory has been so influential that it must be reckoned with. Chomsky's treatment'f of case is restricted almost entirely to discussions of syntax; indeed, a semantic analysis of case is not one of his aims at all. 39 Jakobson (1971: 494) notes that "Chomsky has made an ingenious attempt to construct a 'completely non-semantic theory of grammatical structure' " and
30 wonders whether he is merely playing devil's advocate, since the very idea strikes Jakobson as preposterous. 40 According to Starosta (1976: 1), "Chomsky's configurational definition of grammatical functions is incorrect as a representation of surface grammatical relations." Starosta tries to develop a new notation which will "allow us to represent the fact that these relations have their own characteristic semantic content, a fact that a pure configurational representation is incapable of coping with" (Starosta 1976: 18). Criticisms of Chomsky's theory for its lack of a semantic component come not only from without, but, more importantly, from those who accept his framework as well. Bachman (1980: 47) uses Chomsky's transformations to test the status of the subjective dative in Russian, but finds it "more intuitively appealing to have morphological case assigned in accordance with basic semantic relations than by the application of a superficial mechanical operation called 'case marking' ". More remarkably, Grumet's (1983) dissertation, written entirely within the Chomskyan framework, proves exhaustively that it is impossible to provide an adequate account of the English genitive without making reference to semantics ("the lexicon") and even suggests that this may be true for all cases. She presents a diagram showing the various submeanings of the English genitive and how they overlap with each other, a figure which in many ways resembles a cognitive network. Babby (1987: 136) also follows Chomsky's method, but rejects his characterization of case as semantically vacuous: "the Russian data cast serious doubt on one of the central assumptions of Government and Binding case theory, namely, that case distribution is exhaustively determined by structural relations between the case assigner and assignee". From a cognitive standpoint, Chomsky's treatment of case is an extreme and truncated version of that put forth by Kurylowicz: not only is the semantic status of syntactic case denied, but the existence of other uses of case is altogether ignored. It comes as no surprise that some of Chomsky's followers have tried to put semantics back into the formula.
2.1.6. Symbolic logic (Serensen, Mel'cuk)
Another highly formalistic approach is the application of symbolic logic to case. Serensen (1957: 39) makes very generous use of formulas, but also makes it clear that he will "not in this work give any description of the semantic contents of the linguistic elements which are being treated". After a lengthy definition of terms, Mel'cuk (1986: 42-44) presents a two-page formula to account for case, apologizing for its "clumsy and involved construction", for he "did [his] best to come up
31
with something simpler and more digestible, but failed". Like Serensen, Mel'cuk is concerned primarily with the hows of case distribution rather than with the whys of case meaning.t! Whereas the use of features is a gesture toward mathematical reductionism, the implementation of symbolic logic is an example of the empiricist approach par excellence. From the comments made on this sort of framework in the preceding chapter, it should be clear why it is not suited to semantic inquiry and, consequently, why Serensen and Mel'cuk do not include case semantics in their formulas. Cognitive semantics invokes the use of conventional imagery to caption case. If verbal explanations such as features or metalanguage are one step removed from the use of such images (profiles), words being in some sense less immediate and more abbreviated than schemas, then the assignment of symbols takes us two steps away from the cognitivist's goal.
2.1.7. "Deep" case (Fillmore)
Of articles on case, Fillmore's (1968) "Case for Case" has perhaps provoked more response than any other. Still, his work, as well as that of the "case grammarians" who followed his lead, has little bearing on case as it is understood in this volume, for Fillmore's focus was entirely different. Indeed, what he describes is properly not called "case" at all.42 He describes instead a set of universal relations. The fact that he calls them "cases" has inspired many to criticize Fillmore of having "specifically rejected the importance of surface structure morphology".43 It appears, however, that Fillmore was misinterpreted. Fillmore (1977: 67) recognized the source of this confusion and issued a clarification, stating that "an account of the uses of surface-structure cases requires more than a theory of deep cases". In other words, his framework is not appropriate for examining the semantic structure of specific surface cases, although surface cases are used to express the relations which he calls "cases". In addition to the allegation that Fillmore ignores surface case, opponents point out some more substantive flaws in his theory. Bachman (1980: 186) finds Fillmore's proposals "thought-provoking and heuristically valuable, but unworkably abstract".44 A related problem is that of correlating the two levels that Fillmore invokes, and, particularly, "assigning a given surface structure noun phrase to one of the underlying cases".45 The most frequent criticism is also the most serious - that neither Fillmore nor any of his followers has succeeded in producing a definitive list of "cases". Given that one of the aims of the model is to arrive at a universal set of "cases", failure to do so is hard to overlook.46
33
32 Since Fillmore's relations are identified with case roles,47 his descriptions of them are relevant to our discussion. Indeed, the labels he uses for them, such as Agent and Experiencer, are in many cases appropriate for use in prose explanations of cognitive captions of case meanings, as will be seen in the remaining two chapters. Even more important is the way in which Fillmore (1977: 74) approaches meaning. According to him, grammatical meaning necessarily identifies both "figure" and "ground", as well as "a perspective on [this] scene". This statement could easily be used to describe cognitive schemas. The following two sentences (Fillmore 1977: 80) define the same phenomenon which Langacker refers to as "profiling": "We recognize scenes or situations and the functions of various participants in these scenes and situations. We foreground or bring into perspective some possibly quite small portion of such a scene." Certainly, Fillmore's perspective on the kind of grammatical meaning involved in case roles is not incompatible with that of cognitive semantics.48 Although Fillmore shares some of the concepts basic to a cognitive understanding of case meaning, his framework departs from our model in important ways. Within the cognitive framework, semantics is viewed as immediately relevant to all parts of language. There is no need, therefore, to subdivide language into various levels. Semantics is a part of language in all its aspects and does not require its own, separate leve1. 49 A corollary is the notion that cases form semantic wholes, all the parts of which are interrelated. Because he relegates semantics to the "deep" level, Fillmore treats it as if it were independent of surface morphemes, thus the various meanings of a surface case are scattered about among the "deep cases", with no indication of their overall unity.50 Although Fillmore's description of "deep cases" is intriguing, his system is no more straightforward, no less complex and no less ambiguous than the surface system; which begs the question: why would any language opt for a two-tiered system if it could make do with one level?
2. 1.8. Localist theories of case (Anderson, Freidhof) In spite of the fact that in some versions localism incorporates the theoretical concepts of Fillmore and Chomsky, it is essentially a feature system-! with a special theme - everything is considered derivable from physical location or movement. As Miller (1986: 296) puts it, "localism is the thesis that semantic structures should be constructed in terms of objects being located in a place or moving from one place to another". Although the rigid application of semantic features is irreconcilable with cognitive semantics for reasons already given
above, this basic premise concerning the centrality of location and movement, however abstractly conceived, is clearly consonant with the first of Langacker's (1986a) assumptions, that our perception is organized by means of the "billiardball model", The sentence quoted from Miller above could apply equally to the schemas drawn by Langacker and Smith, as well as to a localist description.V In most versions,53 however, localism produces vast systems of features and trees, adding little to an understanding of case as a coherent semantic structure.
2.2. How the cognitive model has been implied by various authors
Although none of the authors cited thus far in the present chapter used (or, perhaps, was even aware of) the cognitive model, many have made statements that cannot be derived from their own theoretical framework, but instead support the view of semantics proposed in this volume. It appears that these scholars found their framework flawed and tried to correct for perceived deficiencies. In every case, these corrections point to a structure that the given framework did not accommodate.
2.2. 1. Relatedness of meanings As was mentioned above, those who consider their main concern to be the cataloging of all the submeanings of a case frequently mention that certain of them are related to each other, even though from the standpoint of theory there is no reason for these authors to do so.23 This is particularly true of Mrazek (1964), who suggests some relationships with other meanings for nearly every meaning of the Russian instrumental which he identifies. Although he does so in an unsystematic fashion, Mrazek makes enough such statements to indicate that the submeanings are not as haphazard or independent as his method might otherwise imply. In fact, if one gathers together all of the statements which Mrazek has made about the relationships between meanings, and arranges his meanings according to them, the network in Figure 6 is obtained. It is curious that a listwriter actually authored a network, albeit implicitly.
34
35 Subject in Passive Construction Manner ......- -...... Predicate
Object
Group or Formation
Space
Time
Integral Part
Delimitation
Cause Other Concrete Means
PseudoComitative
Figure 6. Mrazek's network
Wierzbicka (1980) also produced a list of meanings for the Russian instrumental and claimed that they were all related to each other because they shared similar statements in the semantic metalanguage of their definitions. Wierzbicka does not make the system of relationships explicit either, but, like Mrazek, she leaves enough clues for a network to be derived. When analyzed into components, it is found that every meaning is related to others by sharing one, two, three or four components. Arranging these meanings according to their relatedness produces a considerably more complex network than that implied by Mrazek, as shown in Figure 7. Triple lines indicate that four components are shared, double lines indicate three shared components, and single lines indicate two shared components (inclusion of relationships based on only one shared component creates a network too complex to be easily accommodated in a two-dimensional drawing).55 '
1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17.
Instrument Object in Action Sentences Body Part Object in Reflexive Action Sentences Emotions Inexplicable Force Agent in Passive Sentences Matter in Action Sentences Matter in Process Sentences Comparison Manner Space Time Duration Transport Personal Characteristics Predicate
Figure 7. Wierzbicka's network
36
Although Deserieva (1974 & 1977) also fails to provide networks for cases, her discussion of them as "semantic fields" is even more suggestive of this sort of structure. She concludes (1977: 60) that the basis for any description of a grammatical category lies in the structure of the semantic field it commands, and that special attention must be paid to the semantic center of a category. It is difficult to conceive of a structure that would fit her description without meeting the specifications of the networks invoked by cognitive semantics. As mentioned in chapter 1, Schlesinger did a cross-linguistic study of the relationships that hold between the instrumental and two other relations which are frequently expressed by the same case: manner and comitative. In both instances he found that there is no hard and fast boundary between the meanings; "the instrumental and comitative are really only two extreme points on what is a conceptual continuum" (Schlesinger 1979: 308). Schlesinger extrapolates from his findings to infer that a category is composed of a system of continua. This is significant because cognitive semantics makes the same claim: nodes in a network are places where instantiations tend to cluster, but this does not mean that the semantic space between them is empty; on the contrary, there usually are transitional examples.
2.2.2. Hierarchical relations among meanings It is customary among both catalogers and feature-specifiers 56 to speak of "basic" meanings for cases, but this device reaches its apex in the works of Roman Jakobson. Jakobson (1936 [1971]: 35) made several distinctions relevant to the status of case meanings. In addition to the general meaning (that represented by features), he recognized both a principal meaning (comparable to the prototype of cognitive semantics) and particular meanings (the constellation of various submeanings and their instantiations). He noted that "the particular meanings ... are not a mechanical accumulation, but form rather a regular hierarchy of particular meanings" grouped about the principal meaning. This hierarchy again implies that case has a center-periphery semantic structure, precisely that which cognitive semantics proposes. Jakobson's features are also found to be apt in describing the schemas that will be presented in the analysis for the dative and instrumental, and are particularly relevant to prototype schemas. As such, Jakobson's feature specifications can largely be retained as an umbrella system for intercase organization, since doing so would not compromise the cognitive model in any way. It is interesting to note that Jakobson held the lexical environment to be responsible for generating the particular meanings (also known as "combinatory variants" for this reason) from the general one. In the cognitive model proposed
37
in chapters 3 and 4 the difference between some submeanings will also be attributed to choice of lexical items and to extralinguistic knowledge. Pragmatics plays an elemental role in specifying the subjective content of an expression, whose outlines have been drawn by case.
2.2.3. Relational invariance On the face of it, the notion of combinatory variants based on general meanings proposed in Jakobson 1936 [1971] seems irreconcilable with the invariant semantic features introduced in Jakobson 1958 [1971]. How can variants and invariants coexist in a single model? This contradiction in terms is only apparent, for it belies a subtler concept grasped by Jakobson, but lacking logical means for expression in the basic theoretical framework that he used. The invariant features display themselves to varying degrees in actual instantiations, a phenomenon that Jakobson referred to as relational invariance, the seeming oxymoron hailed by Sangster (1982: 78) as "probably the most powerful construct in Jakobson's linguistic arsenal". It is this hedging on invariance that allows Jakobson to posit features without making them so immutable and the categories they define so monolithic that the hierarchical structure which he previously proposed would be inconceivable. The concept of relational invariance made it possible for Jakobson to invoke a unifying force for a category without denying the existence of an array of submeanings. 57 It is, however, a clumsy term which is not integrated in an obvious way into the theoretical framework in which it appears.58 Cognitive semantics can produce the same result - accommodating both unity and diversity - without compromising its theoretical premises, by reference to a central, unifying prototype, and a surrounding network of related meanings.
2.3. Advantages ofthe cognitive approach
Various reasons for recommending the cognitive model of semantics have been mentioned here and there up until now; it would be useful at this point to gather them together for purposes of reference.
39
38
2.3.1. A departure from empiricism The very fact that the study of case semantics seems to have reached a dead end (indicated in the Introduction) suggests that a new approach may be needed. In the past half-century, virtually all of semantics has been couched in the settheoretical framework of logical empiricism. Cognitive semantics seeks centerperiphery structure rather than set boundaries, and thus applies a theoretical framework that has proved useful in psychology, neurology, and artificial intelligence.
·2.3.2. Both diversity and unity of case meaning are accommodated
- In some instances meanings are distinguished by lexical environment and extralinguistic knowledge - Case meaning is never absent, even when "governed" by verbs and prepositions
(Jakobson, Potebnja, Pit'ha, Levine) (Oliverius, Plewes, Gladney, van Schooneveld)
2.4. An ideal description ofcase: Skalicka's challenge In 1950, Skalicka issued a challenge that has not yet been met. He outlined the following criteria which an ideal account of case should fulfill. It should be able to:
Since the structure of the semantic network of a case can be drawn as grossly or scrupulously as desired, questions of both overall coherence and differentiation among submeanings can be addressed with equal facility. The set-theory approach of semantic features and exhaustive inventories always requires that one of these aspects be emphasized at the expense of the other.
c)
2.3.3. Many of the best insights of various scholars are incorporated
I would like to argue that cognitive semantics is uniquely adapted to fulfilling these criteria.
Since the relevant items have been covered in previous sections, they will merely be summarized here.
- A description of case should be founded on semantics - Cases have many meanings - Case meanings are interrelated - Differentiation between case meanings is continuous rather than discrete - Case meanings are unified and arranged in a hierarchical structure rather than being haphazard - Case meanings can be explained in terms of location and movement of physical objects in a scene
(Jakobson, Starosta, Grumet, Babby) (Mrazek, Wierzbicka, Kurytowicz, Jakobson) (Mrazek, Wierzbicka) (Schlesinger) (Jakobson)
(Anderson, Miller, Fillmore)
a) b)
(i)
describe the relation of case to reality describe the meaning of case, such that the description is both complete and apt for practical use describe the relations between case and other forms.
The idealized cognitive models captioned in schemas proposed by Smith and Langacker provide the the best approximation to a link between reality and description via reference to human perception of reality.59Cognitive semantics also recognizes the hallmarks of grammatical meaning as well as their essential identity with lexical meaning, which has more obvious realworld correlates. (ii) Cognitive semantics endeavors to embrace all meanings of a case in a single network. The schemas have an immediacy that features, lists, and symbols lack, and that makes both their meanings and their relationships to each other largely self-explanatory. The simplicity and elegance of the system give it great potential for practical applications such as language teaching. (iii) Most descriptions of case either ignore 60 governed case altogether or relegate it to a completely separate category/'! The cognitive model achieves a unified description that includes the relations of case to prepositions and verbs. Relations to other parts of the sentence (such as the subject and direct object, see the analysis in part two) are also captured in the schemas, and the focus on the internal structure or the category facilitates distinction be-
40
tween its semantic contribution and that of prepositions, verbs, and other lexical units.
Part II -
Analysis
3.
Why the Czech dative and the Russian instrumental?
Ceskjjazyk tvoiil cely ceskj ndrod v celt historii, skrze jazyk si poddvaji ruce vsechna pokolenf, v jazyce se obriiii vyvoj naseho ndroda od samycli zacatkli at do dneska, je to dilo kolektivni a tradicni. " It
[All ofthe Czech people throughout history have
worked on the creation ofthe Czech language. Through languageall generationsjoin hands, in language we see the development ofour nation from the very beginning right up to the present day. It is a collective, traditionalcreation.] Frantisek Trdvnicek
If cognitive semantics is to be challenged by the facts of natural language, it would of course be best to assign it as intricate a problem as possible. The dative and instrumental may display the lowest frequency,62 but they are semantically more complex than any other cases. Wierzbicka (1980: 1) comments that "the Russian instrumental is particularly well-known for its mind-boggling qualities", and indeed, several entire monographs have been devoted to it.63 The Czech dative has received somewhat less attention in scholarly literature. 64 Its inclusion here is motivated by a subjective observation that the dative appears to be somewhat more complex in Czech than in Russian,65 and by the desire to use material from more than one language. In addition, it will be observed that the reflexive dative pronoun si produces interesting deformations of the Czech dative network.
3.0.1. Schemas in pictorial and prose notation Figure 8 presents the elements of pictorial and prose notation that will be used in the composition of schemas to caption submeanings throughout the analysis. Although each system of notation is independently adequate for description of the
44
45
cognitive categories of case, they have complementary strengths. For the sake of interpreting individual schemas, the prose notation'v has an immediacy which the pictorial schemas lack, yet the pictorial schemas bring into relief the relationships that hold among schemas in a network in a way not possible in prose. Combined, these two systems of schemas resolve the dilemma of how to represent with equal force the two seemingly antithetical natures of cognitive categories - the diversity of individual members vs. the overall unity of the category. The salience of individual meanings provided by prose notation is balanced by the salience of systematic relationships of members to the prototype visible in the pictorial notation. Yet these two notations stand in a one-to-one mapping relationship to each other. Thus they are perfectly joined and can be said to be but two expressions (verbal and graphic) of a single system. The elements of pictorial and prose notation as they appear in Figure 8 and their interpretation are explained as follows: A.
B.
C.
D.
Labeled circles (an arbitrary, but neutral shape) represent case-marked noun phrases. They will be represented by their case names in prose notation, i.e., "a nominative", "an accusative", etc. A double arrow represents the expression of a non-copular verb (including constructions containing the existential be with a predicate adjective or participle, which are functionally indistinguishable from non-copular stative verbs, cf. section 5.1.1.). So that prose notation will yield grammatical English statements, it must be sensitive to the valence of the verb. Thus if the verb has both a subject and an object, it will be rendered "acts on", as in "a nominative acts on an accusative". If the verb has only a subject it will be rendered "acts", as in "a nominative acts". If the verb has only an object, it will be rendered "an action takes place on", as in "an action takes place on an accusative". Lastly, if a verb lacks both subject and object, it will be rendered merely "an action takes place". A single arrow indicates the movement of an accusative object to a dative indirect object. This element occurs only in schemas describing indirect objects. A single arrow is translated as "to bring to" in prose schemas, as in "a nominative acts on an accusative to bring it to a dative" - the prose schema of the indirect object construction. A dotted arrow indicates that the element that it marks has potential for relative independence from the action chain. This relative independence is defined disjunctively as potential agenthood and/or inviolate integrity. The element thus marked either has the ability to react to the given situation, becoming an agent of a further action; or resists actual contact or penetra-
E.
F.
G.
H.
I.
1.
K.
tion. This is a typical characteristic of datives and has the prose translation "retains independent status". A line encircling the action chain and labeled "setting" represents the setting of a given clause, likewise called "a setting" in prose notation. Prose schemas describing an action chain contained in a setting will include the phrase "in a setting". Viewpoints external to the setting will be labeled with capital letters and signalled in prose by phrases such as "viewed from A". A diagram of a dative in an ellipse with single arrows radiating from it to the periphery of the ellipse represents what is known in prose as "a dative's sphere of control". The ellipse symbolizes a domain that in some sense belongs to the.dative (variously realized in context as the body, possessions, feelings, speech-act domain, etc. of the dative referent) and the arrows symbolize the dative's control over this domain. Any infringement on this domain or any action pertaining to elements located within it is indirectly experienced by the dative. A double arrow that represents verbal action and proceeds through a circle representing a case-marked (invariably instrumental) noun phrase indicates that the action takes place "via" the instrumental. The instrumental in this case acts as a conduit or vehicle for the action. As will be shown in the analysis below, context determines the precise interpretation of the instrumental as an instrument of the action, a manner of performance of the action, a path guiding the action, or as an agent of a passive action. An instrumental attached to another case-marked element indicates that the latter element is "seen as" or construed as an instrumental. The instrumental acts as an attribute - quantitative, qualitative or comparative - of the other element. Again, context disambiguates the role of the instrumental attribute, and this will be indicated in the analysis. A double line indicates the presence of a copular verb and will be translated as "is". A single line indicates a bond between two elements, translated as "together with" and interpretable as a comitative relationship. If two elements are located in close proximity to each other, but not related by any of the symbols described above, the term "close to" will appear in prose notation. This term may be negated as "not close to" if two elements are non-proximate.
46
47 Pictorial Schemas
A.
Prose Schemas
0;0;0;0 nom
ace
dat
"a nominative" ; "an accusative" ; "a dative" ; "an instrumental"
inst
B. "acts on" ; "acts" ; "an action takes place (on)" interpretable as a (non-copular) verb
> C.
"to bring to"
• D:
- -
-
"retains independent status" interpretable as any of the following: potential agency, resistance to integration in action chain, resistance to contact/penetration
~
\.
E.
F.
A
Oetting
@
G.
"(in) a setting" "viewed from A"
"a dative sphere of control"
3.0.2. Indirect object serves as the prototype of the Czech dative
e
"an action takes place via an instrumental"
:>
mst
H.
Q)
>
"a nominative seen as an instrumental acts"
nom inst
I.
0=0 nom
J.
"a nominative is a nominative'" interpretable as a copular verb
nom'
ginSI
"a nominative together with an instrumental" interpetable as a comitative relationship
0 0
It will be argued here that the central meaning (also known as the prototype) of the dative rests in the indirect object construction. This claim may seem facile and obvious 67 , but it conceals a cognitive structure, the intricacy of which becomes apparent as soon as one tries to defme the term "indirect object". The term represents a skewed concept that is positively identified only in very limited circumstances. An indirect object sensu stricto can be said to exist only in the presence of the verb ddt "give" and its closest synonyms; otherwise the import of the dative quickly fades into other meanings. Observe the tension between the indirect object and other meanings (possessive, beneficiary) in the following examples, all of which have the same basic nominative-dative-accusative case structure:
(6)
nom
K
In addition to the terms given above, prose notation contains several more items, including the self-explanatory words "horizontally", "vertically", and the conjunction "and". If one element in a pictorial schema has two labels, such as "nom/dat" or "setting/inst", this relationship of identity is captioned with the words "be" or "be one" in prose notation, as in "the nominative and the dative are one", "the setting is an instrumental". The basic network of a case category consists of its prototype and paradigmatic variants of the prototype. The prototype is labeled schema 1 and all paradigmatic variants are labeled with consecutive numerals (thus they are known as schema 2, schema 3, etc.). The prototype and each of the paradigmatic variants have a local subnetwork comprised of syntagmatic variants. Syntagmatic variants are labeled with the numeral of the paradigmatic variant followed by a letter (thus syntagmatic variants of schema 2 are labeled schema 2a, schema 2b, etc.). The presence of semantic extensions is mentioned in the text where appropriate, but not included in the schemas, since they represent variations not in the shape, but in the application of the schemas.
mu
dala
him-DAT
gave flower-Ace
kytku.
'Ludmila gave him a flower.'
"a nominative close to an instrumental" interpretable as proximity
(7)
Figure 8. Elements of the schemas
Ludmila
Ludmila-NOM
Ludmila
mu
dala
pusinku
Ludmila-NOM
him-DAT
gave
kiss-Ace
na on
49
48
celo. forehead-ACC 'Ludmilakissed his forehead.'
(8)
(9)
(10)
Ludmila mu koupila Ludmila-NOM him-DAT bought 'Ludmilabought him a hat.' Ludmila mu uvarila Ludmila-NOM him-DAT cooked 'Ludmilacooked kashafor him.' Ludmila mu ukazala Ludmila-NOM him-DAT showed 'Ludmilashowed him the way home.'
klobouk.
(13)
Ludmila mu vzdycky Ludmila-NOM him-DAT always 'Ludmila always cookedwell for him.'
dobre varila. well cooked
(14)
Uz jsem mu am-AUX him-DAT Already 'I have already written to him.'
napsala. written
(15)
Ludmila Ludmila-NOM
hat-Ace
kasi. kasha-ACC
mu
fekla, ze
him-DAT
told
that
chce wants
jft to-go
domu
cestu way-ACC
domu. home
(11)
UZ jsem mu poslala Already am-AUX him-DAT sent doporucenf do Prahy. recommendation-Ace to Prague-GEN 'I already sent the recommendation to him in Praguelfor him to Prague.'
(12)
Ludmila mu rekla svojejmeno. Ludmila-NOM him-DAT told her name-Ace 'Ludmilatold him her name.'
While (6) represents a canonical indirect object, (7) shows a drift toward expressing possession and (8), (9) and (10) present a trend in which the meaning of indirect object graduallygives way to a benefactive reading and the dative referent is less and less likely to be interpreted as a recipient. (11) can be read either way, depending upon whetherthe referent of mu is in Prague or not and whether he is the subjector the addressee of the letter. (12) is a transitional example which shows the ties that exist between the indirect object and an important group of dative-governing verbs that express communication. The fact that manyof these verysame verbscan appearin similarconstructions withoutthe accusative objectfurther demonstrates that an operative definition that would allow unequivocal identification of all indirect objects as opposed to other dative constructions simply cannot be formulated. Compare (13), (14) and ,(15) with (9), (11) and (12),respectively:
home 'Ludmilatold him that she wanted to go home.'
Given these data, we can conclude that the purest kind of indirect object occupies just one narrow corner of the cognitive network of the dative, which elsewherespreads out to display a plethoraof meanings. If the indirectobject has such a limitedand marginal status, how can we assertthat it is central to the dative as a whole? The indirect object may not cooccur with a large number of verbs, but the frequency of these verbs is unusually high and the indirect object fulfills other requirements for recognition as the center of gravity in the cognitive network of the dative. It should be noted that the scene invokedby the verb meaning "give" is very basic in the experience of human beings. We encounterthis situation (and therefore potential "pure"indirect objects) manytimesevery day. It is a basictenet of cognitive semantics that idealized cognitive models (IeMs) are grounded in human experience. 68 Furthermore, Lakoff (1986: 4) assertsthat a "central claim" of cognitive semantics is that "meaningful conceptual structures arise from two sources: 1. from the structured nature of bodilyand socialexperience and 2. from our innatecapacity to imaginatively projectfrom certain well-structured aspects of bodily and interactional experience to abstract conceptual structures." Clearly 1. describes the positionof the indirect object and 2. accounts for the derivation of related meaningsfrom it, as demonstrated in the examplesabove. Thus, the indirect object must be realized as conceptually central to the dative's network, even though it is very narrowly defined. In order to draw a schema for the indirect object (and, subsequently, for all other meanings of the dative), it is first necessary to arrive at a definition of the roles of the nominative and accusative, since the dative interacts with both. The
51
50
Prose Schema
Pictorial Schema
O~:O nom
A nominative acts on an accusative in a setting.
acc
Figure 9. Prose Schema
Pictorial Schema
o
schema presented in Figure 9 (a modification of Langacker 1987: 5 "normal observation of a prototypical action", which appears in Figure 1, chapter 1 of this volume) will be postulated as a definition for these two cases. A comprehensive treatment of the cognitive networks of the nominative and accusative is beyond the scope of the present work, although a few comments should be made. The schema in Figure 9 captions only the prototypes of these two cases, and these are the only parts of their cognitive structure that are relevant to our definition of the dative. The prototype of the nominative is its most interactive version. It has at least two other uses: 1. a naming function (used in lists, dictionaries, signs and in some appositive expressions) which gives a bare nominative, as in Figure 10, and 2. a predicate function, in which two nominatives are joined by a bond of equation, as in Figure 11. In addition to prepositional uses, the accusative is associated with time expressions. In some prepositional uses, the preposition is not obligatory and a direct object can be used instead as in (16b) (this example and (17) below are cited by Hajicova et at 1979: 158): (16)
a.
preskocit
potok
to-jump-over stream-ACC A nominative in a setting.
b.
nom
pfeskocit
pres
potok
to-jump-over '.
across
stream-ACC
'to jump over a stream' This suggests that such prepositional uses are actually extensions of the direct object. The accusative of time likewise shows close ties with the direct object. Compare the imperfective and perfective expressions in (17):
Figure 10. Pictorial Schema
Prose Schema
(17)
a.
plakat celou noc
b.
proplakat
celou noc
to-cry-through
all night-ACC
to-cry all night-ACC
Oi=====lO nom
A nominative is a nominative' in a setting.
'to cry all night' In the first case the verb is intransitive and the time expression would normally be identified as an accusative of time. The second verb is, however, transitive and the accusative complement is obligatory and arguably a direct object, analogous to that in (18):
nom'
Figure 11.
52
(18)
53
precfst
celou knihu to-read-through whole book-ACC 'to read a whole book'
3.0.3. Cognitive network of the Czech dative Schemas
Figure 12 contains the schemas for the major submeanings of the Czech dative and the cognitive network they comprise. It should be kept in mind that the network is in many ways a convenient artifice. It shows only major submeanings and a skeletal structure that holds them together. A truer diagram would look more like a map of the Milky Way, for it would show areas of high density, where instantiations are thickly clustered about the major submeanings, as well as sparser areas where there are only occasional transitional examples between them. 69 For all its accuracy, such a map would be difficult both to plot and to read. The network contains a brief explanation of the paradigmatic relationship between schemas 1 and 2. It is the result of an operation on the scope of the dative, which ranges from a mere potential as an independent participant in schema 1 to the maximally external scope of the sphere of control in 2. Both schemas have local subnetworks of syntagmatic variants, which together account for all of the dative's uses. Other types of syntactic variants also exist, but since they have little relevance for either the shape of schemas or the meaning of the dative, they will not be treated separately. Such syntactic variants are to be found in participial clauses, which for our purposes will be treated no differently than the finite verb phrases from which they are derived. (19) and the participial clause in (20) will therefore be treated identically. (19)
V Ufade mu dati potvrzenf, In office-LOC him-DAT gave confirmation-Ace 'They gave him a letter of confmnation at the office.'
(20)
Precetl potvrzent, dane mu v read confirmation-Ace given him-DAT in ufade, aZ doma. office-Lee when home 'He read the letter of confirmation given to him at the office when he got home.'
Pictorial
Usage
Prose
A nominative acts on an accusative to bring it to a dative. The dative retains independent status.
A nominative acts on an accusative in a dative's sphere of control in a setting.
Indirect object (syntagmatic variants account for governed dative & impersonal dative)
Free Dative
Network rR---externalization
~ofdative
..
.t;\
~
Figure 12. Czech dative network
55
54
Another example of a trivial sort of syntactic variant is the placement of the contents of a message in a separate clause. Thus example (15) will be treated like (12). Another factor important in shaping the network is the fact that the lexical meanings of nouns as well as extralinguistic knowledge can be used to disambiguate meanings of the dative. This makes it possible for a single schema to represent more than one submeaning. Schema 2 captions a number of submeanings, many of which are distinguished by the choice of the dative's referent. The specific relevance of both syntactic variants and the role of lexical and extra-linguistic information with regard to each submeaning will be taken up indivi-dually as the schemas are discussed in turn in the remainder of this chapter.
Schema 1 and its Syntagmatic Variants Schemas Pictorial
Prose
3.1. Schema 1 - indirect object
Schema 1 in Figure 13 builds on Figure 9 by including the dative. It captions the actual transferral of an accusative direct object to the dative. Note that all of the participants in this schema are distinct. The nominative is the head of the action chain and is responsible for the movement of the accusative. The dative both receives the accusative entity and shows a potential for carrying out some further action with it (symbolized by the dotted arrow). Note that if this diagram were cut into three pieces, they would correspond quite closely to the canonical profiles of these three cases worked out by Smith (1985), cited in chapter 1, Figure 2. This schema is quite in harmony with common prose descriptions of the dative. Jakobson's feature specification for the dative is [+directional, +peripheral]. Jakobson (1936 [1971]: 48) expands this as meaning that the dative "signifies peripheral status ... and involvement in an action", and that it "denotes the existence of its referent as independent of the action". The fact that the dative does not participate directly in the prototypical action likewise indicates its peripheral status, and its potential as an agent attests to its independent existence. The schema is also an accurate caption of the two characteristics of the dative recognized by Burston (1977: 273-274): "a) the participant in question regards the dative referent as a kind of beacon exerting an attractive force, [and] b) [the dative] entity is remote, able to escape total integration inside the narrated event".
Usage
A nominative acts on an accusative to bring it to a dative in a setting. The dative retains independent status.
Indirect object
A nominative acts on a dative in a setting. The dative retains independent status.
Governed Dative
An action takes place on a dative in a setting. The dative retains independent status.
Impersonal Dative
- - - - - - setting
Local Subnetwork
~IOSSOf~IOSSOf~ ~acc
~nom
Figure 13. Schema 1 and its syntagmatic variants
"<::
57
56
The dative plays a unique role; it is both an object of a verb and a subject of a potential verb, both controled and controling. The very nature of this dual role suggests that the dative referent must be capable of both absorbing an action and reacting to it. In addition to prompting the choice of human referents, this feature of the dative (expressed in various ways throughout the network) makes it possible for it to pass subjecthood tests in impersonal constructions (as demonstrated by Bachman [1980]) and also accounts for the readings of modality and benefit or harm (Bachmann [1980]) often associated with the dative (inanimate objects and, to some extent, animals are excluded because they lack the ability to respond and thus sense harm or benefit).70 The dative's potential role as an agent of further action is the element that distinguishes it from other cases and it is the variations in this potential role which account for the various meanings of the dative. This hallmark element places important restrictions on the classes of nouns that can appear in the dative in such constructions. They must signify referents capable of doing something with the accusative entity, which accounts for the predominance of nouns denoting persons and personal pronouns in the dative case in Russian, cited by Greenberg (1974: 25). We cannot say:
object category are produced by semantic extension via relationships of synonymy, antonymy and metonymy and do not have as strict distribution constraints as "true" indirect objects. First I will discuss the distribution constraints peculiar to "true" indirect objects, and then I will outline the shape of this category. A "true" indirect object cannot be replaced by a prepositional phrase
"True" indirect objects are the only datives with beneficiary meaning that cannot be replaced by a prepositional phrase such as pro + accusative "for". Example (23b) is unacceptable, whereas (24b) and (25b) are quite normal. (23)
b.
(24) (21)
*Dala jsem svoje knihy Gave am-AUX own books-ACC 'I gave my books to the room. '
a.
a.
pokoji. room-DAT
b.
because the room is incapable of doing anything with the books. This type of sentence normally requires a human referent for the dative. It is, however, possible to say: (22)
Dala jsem svoje knihy gave am-AUX own books-ACC 'I gave my books to the library.'
knihovne. library-DAT
(25)
a. b.
but this is because we can assume that there are people in the library who will take responsibility for cataloging and shelving the books.Z! It is no wonder that Belicova (1982: 67) labels a central semantic feature of the dative "personicnost" [personhood] . The Czech indirect object is itself a prototype-based category within the category of the dative. The central member of this category is instantiated by constructions with verbs meaning "give" in which the dative can neither be replaced by a prepositional phrase nor omitted. Peripheral members of the indirect
Ludmila mu him-DAT Ludmila-NOM 'Ludmila gave him a flower.' *Ludmila dala kytku Ludmila-NOM gave flower-Ace 'Ludmila gave a flower for him.' Ludmila mu him-DAT Ludmila-NOM 'Ludmila bought him a hat.' Ludmila pro nej Ludmila-NOM for him-ACC klobouk. hat-ACC 'Ludmila bought a hat for him.'
dala gave
kytku. flower-Ace
pro for
nej, him-ACC
koupila bought
klobouk. hat-ACC
koupila bought
Ludmila mu uvarila kasi. Ludmila-NOM him-DAT cooked kasha-ACC 'Ludmila cooked kasha for him. ' Ludmila pro nej uvarila Ludmila-NOM for him-ACC cooked kasi. kasha-ACC 'Ludmila cooked kasha for him. '
The use of the preposition pro is constrained by its positive connotations to beneficiary readings. If harm rather than benefit is involved, the use of the dative usually signals possession of the object and affectedness via that possession.
59
58
(26)
Ales
ndm
us-DAT 'Ales wrecked our car.'
Ales-NOM
naboural
auto.
ukradla
wrecked
car-ACC
stole book-ACC 'I stole a bookfor you from the library.' J jsem pro tebe I-NOM am-AUX for you-ACC
b. Since the semantics of the verb are in conflict with pro, its use is marginal at best, as in (27). (27)
*Ales
naboural
auto
pro
nas.
Ales-NOM
wrecked
car-ACC
for
us-ACC
Semantic extension of the indirect object with verbs of "taking" and "intransitive taking" (discussed below) is incompatible with pro, but can sometimes be used with od "from", with very minor alterations in meaning.
a.
Ten pan
nam
vzal
hodne
That man-NOM
us-DAT
took
much
penez. money-GEN
b.
'That man took a lot of money from us (he was dishonest).' Ten pan vzal od nas hodne That man-NOM
took
from
us-GEN
a
knihovny
much
a. 1 stole a book from your library (dative of possession). b. Hey, you know what, 1 stole a book from the library (ethical dative). With respect to such limitations, verbs like poslat "send" and prodat "sell" function like ddt; the use of a prepositional phrase with these verbs entails a radicalchangein the semantic structure of the sentence, such that the objectof the preposition is no longeridentifiable with the indirectobject.72 (31)
a.
Teta
ti
yOU-DAT 'Auntie sent you a gift.' poslala Teta
Aunt-NOM
money-GEN
'That man took a lot of moneyfrom us.'
b.
Aunt-NOM
a.
knihu.
The dative in (30a) is, however, triply ambiguous, and the reading which is replaceable with pro is the leastlikelyof the three. The other two meanings of this sentence are:
penez.
(29)
ukradla
Ludmila
z from
library-GEN stole book-ACC 'I stole a book for you from the library.'
'Ales wrecked the car for us.'
(28)
knihu.
utekla. us-DAT ran-away 'Ludmilaran awayfrom us (and we can't do anything about it).' Ludmila od nas utekla.
gave
poslala
darek,
sent
gift-Ace
darek gift-ACC
pro
tebe.
for
you-Ace
'Auntie sent a gift for you (rightnow 1 have the gift, not you).'
ndm
Ludmila-NOM
b.
Ludmila-NOM
from
us-GEN
(32)
a.
ran-away
'Ludmilaran awayfrom us.'
Kdo
b.
yOU-DAT 'Who sold that car to you?' vas Kdo za who-NOM
If the dative referentis not the possessor of the directobject,it is possible for a "taking" verb to indicate benefitto the dative referent, and in such cases pro is admissible, as in (30).
(30)
a.
Ja
jsem
ti
z
knihovny
I-NOM
am-AUX
you-DAT
from
library-GEN
wim
who-NOM
for
prodal
to auto?
sold
that car-ACC
you-GEN
prodal
to auto?
sold
that car-ACC
'Who sold that car for you?' (32b) implies that the referent of you is the seller rather than the buyer, and that who is you's agent. Dative objects of verbs of communication, such as lid "say" and psdt "write" also resist replacement by a prepositional phrase; and this is true for all
60
61
of the dative-governing verbs. The dative-governing verbs, however, have no accusa-tive directobjects and are thus already sufficiently distinctfrom verbs that can takeindirect objects. Eventhough verbsof communication can optionally take direct objects, their link to the dative-governing verbs as transitional elements is strong enough to exempt them from this rule.73 Thus dative arguments of verbs of communication are not to be identified as centralexamplesof indirect objects, even though they do not permitthe substitution of a prepositional phrase.
(35)
a.
Uz
jsem
ti
poslala
ty
Already
am-AUX
you-DAT
sent
the
dopisy. letters-ACC 'I have already sent you the letters.'
b.
UZ
jsem
poslala
ty dopisy.
Already am-A UX sent 'I have already sent the letters.'
the letters-ACC
Ludmila mu uvarila Ludmila-NOM him-DAT cooked 'Ludmilacooked kashajor him.'
kasi.
A "true"indirect object cannot be omitted (36) The absence of an accusative direct objectlikewise exemptsthe dative-governing verbs from the secondlimitation that distinguishes true indirectobjectsfrom other datives: a true indirect object cannot be omitted without rendering a sentence ungrammatical. For dative-governing verbs this is a purely configurationalconstraint. They are transitive and require some object. Again, however, there is no need to invokea rule to distinguish governed dativesfrom indirectobjects. Verbs of communication, as well as the near-synonyms of ddt, poslat and prodat, are subject to this rule, and all of these verbs are distinguished from ddt on the basis of this criterion.
a.
b.
a.
b.
Ludmila
mu
rekla, ze
Ludmila-NOM jit dornu. to-go home
him-DAT
said
that
that
wants to-go
a.
Takova holka
by
prodala
Such girl-NOM
would d'dblovi.
sell
svojimatku
(37)
a.
would sell
'A girl like that would sell her own mother.'
kytku. flower-Ace
chce
Ludmila-NOM
gave
flower-Ace
wants
'Ludmilagave a flower.'
home
It is the optionalnature of the dative in (34), (35) and (36)75 which has earned it the appellation "free dative",76 because it acts as a free modifier that entersinto a relationship with the sentence as a whole. The assertion that the free dative modifies the whole sentence rather than acting as a direct participant (indirect object) is supported by examples with more than one free dative. (38) contains both an ethical dative ti "you", which refers to the interlocutor, and a dative of possession mu "him", which refers to the owner of the house.
(38)
devil-DAT 'A girl like that would sell her own mother to the devil. ' Takovaholka by prodala svoji matku. Such girl-NOM
Ludmila mu dala Ludmila-NOM him-DAT gave 'Ludmilagave him a flower.'
kytku.74
own mother-ACC
b.
kasha-ACC
'Ludmilacookedkasha.'
dala
'Ludmilasaid that she wanted to go home.' (34)
kasi.
*Ludmila
'Ludmilatold him that she wanted to go home.' Ludmila rekla, ze chce jit domu. Ludmila-NOM said
uvarila
Ludmila-NOM cooked
b. (33)
Ludmila
kasha-ACC
own mother-ACC
To
v-~->;:->"
mu
hotel
That-NOM you-DAT him-DAT burned '(Hey, you know what?) His house was on fire.'
dum. house
Example (39), which appeared recently in the Czech periodical Mlady svet, containsthreesuch datives: an ethicaldative ti "you", a possessive dative si "her own (reflexive)", and an emotional dative mi "on me".
62
(39)
63
Pustila
jsem
dceru
Let
am-AUX
na daughter-ACe to ona ti she-NOM yoU-DAT zlomila nohu!
hory a si mountains-Ace and REFL-DAT mi tam me-DAT there broke leg-ACC 'I let my daughter go to the mountains and you know what, she broke her leg on me!'
dorucovat odvddet postupovat sveiovat vracet
"deliver" "tum over, return" ,'surrender' , "put in (someone's) custody, care" "return"
Antonymy
The role of the free dative will be taken up in detail in section 3.2.
The semantic import of the clause can be reversed, using verbs which denote "take" rather than "give", as in the examples below.
Semantic extension of the indirect object
(40)
As mentioned above, the indirect object is itself a semantic category based on prototypical three-argument constructions with the verb ddt "give". The Czech indirect object has been semantically extended via relationships of synonymy, antonymy and metonymy.
Ten pan
by
That man-NOM
would
'That man would take your last penny.' Nekdo
piedklddat
piisunovat strkat
cpdt vnucovat
piihrdvat piindset presentovat odevzddvat doddvat pieddvat
ndm
us-DAT 'Somebody stole our car.'
Synonymy
nabizet
"hand" "offer" "serve, place before" "draw toward" "push toward" "thrust upon' "force upon" "pass" "bring" , 'present" "hand in" "add" "hand over"
took
poslednf haler.
Somebody-NOM
poddvat
vzal
last penny-ACC
(41)
We have already seen near-synonyms of dat "give" with datives that function as indirect objects, such as poslat "send (to)" and koupit "buy (for)". Further examples appear in the table below:
ti you-DAT
ukradl
auto.
stole
car-ACC
Examples like these indicate the presence of a cognitive bridge between the indirect object and the possessive meaning of the dative, for when one takes something from someone else, it is necessarily assumed that the object was in the possession of the dative referent. A sampling of "taking" verbs that can be accompanied by an extended indirect object is presented in the table below. brat
odnimat krdst zabavovat rvat
trhat skubat vyfouknout
cmajmout slohnout
stipnout
"take" "take" "steal" "seize" "wrest from" "yank away" "wrench away" "snatch away" "swipe" "pinch" "pinch"
65
64
uzmout
zcizovat vyvlastnovat
"take away" "appropriate' , "expropriate' ,
(stdvat se)
(piihdzet se)
(zoot se) (jevit se)
Metonymy - intransitive giving and taking, incorporation of direct object into the verb, giving of benefit and harm
Intransitive taking ujiidet utikat
In metonymic extensions of the Czech indirect object the direct object is subsumed in the semantics of the verb and therefore is not expressed as an overt accusative. Because this type of extended indirect object construction has the same argument structure as (and some degree of semantic overlap with) the governed dative (Schema la), it produces transitions between these two schemas. Both "giving" and "taking" constructions are extended in this way via intransitivization, such that the subject (physically or perceptually) "gives" itself to or "takes" itself from the dative referent. Examples and tables below illustrate the range of such "intransitive giving and taking" verbs.I? (42)
(43)
ndm Vladimfr se us-DAT Vladimir-NOM REFL-ACC 'Vladimir introduced himself to us.'
(odddvat se) (venovat se) (obetovat se)
(zaprodavai se)
(sveiit se) patiit piisluset
stacit piipadat
chybet (vyhybat se)
(vymykat se) (zirdcetse) (vzdalovat se)
(odrozovat se) (odcizovat se)
introduced
odnistat
utekla.
Intransitive giving
(piibliiovat se) (piedstavovat se)
uprchat schdzet
predstavil.
ndm Ludmila ran-away Ludmila-NOM us-DAT 'Ludmila ran away from us.'
nadbihat
uchazet unikat
"catch up to, overtake" "approach' , "introduce self to" "devote self to" "devote, dedicate self to" "sacrifice self to" "be a traitor for, sell self to (the enemy)" "entrust self to; confide in" "belong to" "belong to; be owed to" "be enough, suffice" "fall to (someone's) lot; occur to; seem to"
"happen to" "happen to" "seem to; appear to (in a dream)" "be evident to"
"ride away (leaving the dative referent behind)" "run away (leaving the dative referent behind)" "escape" "escape; leak out; avoid; elude" "flee; escape" "be missing; lacking" "be missing; lacking" "avoid" "wrench loose; get away from" "get lost on" "become estranged from" "renounce (one's nationality), defect" "become estranged, alienated" "outgrow"
Some of these verbs are statives (stacit, zdat se), and consequently their import is "be available to", "at the disposal of" rather than "give self to"; or, in the case of chybe; schdzet, listed under intransitive taking verbs, the meaning conveyed is "not be available to" rather than "take self from' ,. A number of these verbs also have ties with expressions of dative of beneficiary (schema 2), for they describe intransitive actions which have an effect on the dative entity. With the relevant verbs in the intransitive giving and taking groups, the intransitive affective reading is produced by an evaluation relative to a beneficiary that is directly encoded in the verb, whereas the verbs used with the free dative are neutral in this respect. However, no rigid boundaries can be drawn here either, for as expected there is a range of verbs rather than distinct sets. The use of the dative with such verbs is motivated both by their inherent meaning and by their relationship to "give" (or "take"). Compare the use of the dative with intransitive taking verbs in (44) and (45) with its use as a free modification with other intransitive verbs as in (46).
67
66
(44)
Unika ndm plyn Leaks us-DAT gas-NOM 'We have a gas leak in the apartment.'
v in
byte.
apartment-LOC
(45)
Ujel ndm Drove-away us-DAT 'We missed the bus.'
(46)
Miminko ndm place v noci. Baby-NOM us-DAT cries in night-LOC 'The baby cries at night (and we are affected by this).'
autobus. bus-NOM
Whenever the verbs unikat "escape" and ujet 'ride off are used, the presence of an experiencer is implied, whether or not it is expressed. The same cannot be said of plakat "cry"; the use of the dative in (46) is motivated by circumstances other than mere verbal semantics.
blaJwpftlt (poroucet se) zehnat
gratulovat kondolovat radit* promfjet* piipominat spilat naddvat
zloiecit hrozit (modlit se) (rouhat se) (chlubit se) zalovat
(odporoucet se) Incorporation of direct objectin the verb The specification of the nature of the direct object by some verbs is sufficient to make its actual realization superfluous, or at least optional (verbs which can optionally take accusative direct objects are followed by an asterisk below). Most of these verbs involve the giving of signals, verbal or nonverbal, and are thus related to verbs of communication; their relation to expressionscontaining indirect objects has already been discussed. Most of these verbs also bear close ties to the governed dative (in its meanings of subordination and domination), or to benefit and harm.78 Giving verbal signals piisvedcovat lichotit pochlebovat piitakdvat poroucet" naiizovat" dekovat
piizvukovat pflit
"consent' , ,'flatter', "fawn upon" "pay lip service to" "command" "order, direct" "thank" "second, give support to" "wish"
odpovidat" lhdt volat79 telefonovat tykat vykat
Giving nonverbalsignals salutovat piikyvovat tleskat kynout mdvat (smdt se) signalizovat* telegrafovat* psdt
Giving punishment nafackovat napldcat
"wish well" "give regards to" "bless" "congratulate', "give condolences" "advise" "forgive' , "remind" "abuse; call names" "abuse; tell off' "curse" "threaten" "pray" "blaspheme' , "boast" "complain" "take one's leave of" "answer" "lie" "call" "call (on the telephone)" "say ty to; speak on informal terms with" "say vy to; speak on formal terms with"
"salute" "consent (lit. 'nod') to" "applaud" "wave to" "wave to" "laugh at" "signal" "telegraph' , "write to"
"slap" "spank"
68
69
nasekat
namldtit nabit
"whip" "thrash" "beat"
hrozit (hnusit se) (znechucovat se) (SJrU1t se)
, 'threaten' , "disgust" "disgust" "laugh at"
Giving money or gifts
platit obetovat piispivat
"pay" ,'sacrifice' , "contribute' ,
Verbs of benefit or harm The intransitive giving and taking verbs that invoke the presence of a beneficiary provide evidence of a link between that group and the verbs of benefit and harm, which are a logical metaphorical extension of it. The notion of benefit is clearly an abstraction of giving, for the beneficiary receives some good; harm is its converse, for good is taken away or evil given. Here the tendency for the notion of harm or benefit (which entails the existence of a beneficiary) to be incorporated in the verb is quite pronounced and their relationship to expressions using the dative of beneficiary (and thus to schema 2 and its variantsjis tangible. Benefit vyhovovat (hodit se) (libit se)
prospivat sluset svedcit lahodit chutnat
ptipijet
odzvanet Harm vadit skodit
ublizovat piekdzet nesedet
"comply; satisfy, please" "suit (someone's) purpose; please" "appeal to" "do (someone) good, benefit" "become, suit" "agree with; be good for" "be pleasant to, please" "taste good to" "drink to" "toll (a bell) for (someone who is deceased)"
"hamper, trouble" "harm" "hurt" "be in the way of, hinder" "bother"
3. 1. 1. Schema 1a - governed dative The removal of the accusative participant from the syntagm of schema 1 produces schema Ia, in which a nominative acts directly on a dative. This schema captions all uses of the dative traditionally termed government (with the exception of those which were shown to result from semantic extension of the indirect object in the preceding section).
Dative verbs It is a well-known fact that the verbs said to govern oblique cases form semantic groups rather than random aggregates. 80 The classifications proposed for such verbs tend to be rather ad hoc, for they are usually based upon the perception by individual authors of the greater or lesser salience of shared semantic features of verbs. Each author offers a different list of semantic groups, each of which is equally arbitrary and disputable. What is lacking is a principled means for approaching verbal government which will prevent it from degenerating into a disjunctive inventory. The schema of the cognitive network that represents the governed dative effectively fills this need. It provides a definition of the dative's semantic contribution to sentences using dative-governing verbs, upon which an orderly classification of such verbs can be based. In schema 1a we see the nominative and the dative neatly juxtaposed. These two participants differ only in that the nominative is an actual agent whereas the dative merely retains independent status (realizable as potential agenthood or inviolate integrity). This is all that the schema tells us and all that we can ascribe to the semantics of the dative. Yet it provides enough information to outline the groups of verbs involved. The simple juxtaposition of two entities with a similar capacity for agenthood implies a power struggle with each battling for control. Only the semantics of the individual verbs determine whether the outcome will favor the nominative or the dative or neither. Verbs that are truly dative-governing (distinct from those that participate in extension of the indirect object) do indeed fall into three categories
70
71
corresponding to the possible outcomes of the nominative/dative conflict. The system of categorization that this produces is essentially the same as that suggested by Andersen (1970), who affirms that all dative-governing verbs denote implicating relationships. He distinguishes two types of implicating relationships: symmetrical relationships, in which "the verb is necessarily relevant to both the subject and to the dative object" (Andersen 1970: 5), and transitive relationships, which include subordination. 81 These groups (and their subgroups) are artifacts of verbal rather than case semantics, and are used here for convenience. They represent semantic fields in which verbs tend to cluster, but there are some verbs that are tangential to one or more field, or can belong to one group or another, depending upon context.
Symmetrical relationships The following verbs describe situations in which the nominative subject and the dative object are approximately matched in strength. Symmetrical relationships (podobat se) (rovnat se) odpovidat (vyrovnat se) konkurovat piekazet brdnit odmlouvat odporovat oponovat (vzpirat se) (protivit se)
"be similar" "be equal" "correspond to" "compete with, keep up with" "compete with" "be in the way of, interfere" "be in the way of, prevent, resist, oppose" "contradict, talk back" "oppose, contradict" "oppose, object to, argue with" "oppose, defy, refuse" "oppose, rebel, resist"
Some verbs are given in parentheses because they require the use of the reflexive particle se, which is historically the short form of the accusative reflexive pronoun,82 and thus in some sense they can be said to have direct objects. Yet such verbs are commonly listed with dative-governing verbs,83 and with good reason, for they pattern with these verbs. The presence of the particle merely provides added motivation for using the dative.
All of these verbs assume that the dative entity is either providing a value with which the nominative can be compared or is exerting a force which can be struggled against. As soon as the balance of power is tipped in favor of either the nominative or the dative, transitive relationships are obtained.
Transitive relationships - subordination Verbs of subordination describe a situation in which the nominative entity places itself in a situation controlled by the dative entity. Of the three submeanings here attributed to the dative-governing verbs, this meaning is most clearly derivable from the schema, depicting as it does the nominative approaching the dative, the pontential agent of a further action. It comes as no surprise, then, that this meaning represents the largest and most diversified subgroup among the dative-governing verbs. This group also shows the most obvious relationships to the indirect object, for when one subordinates oneself to another, one transfers oneself (or some aspect of oneself) to the other's power. Thus it is in these verbs that we see the most direct link between schemas I and la, via the intransitive giving verbs. The presence of this concept of giving is noted in the glosses. Because this subgroup is very large, it has been further subdivided for the sake of convenience.
Subordination "Serve" otrocit (zaproddvatse) slouzit pomdhat asistovat (oddavat se) (venovat se) (obetovat se)
"be a slave to" "be a traitor for; sell (GIVE) self to" "serve" "help (GIVE help to)" "assist (GIVE assistance to)" "devote self to (GIVE self to)' , "devote, dedicate self to (GIVE self to)" "sacrifice (self) to (GIVE to)"
"Subordinate self to" podlezat podkuiovat (podddvat se)
"kowtow to" "toady to" "submit to; GIVE in to"
72
73
(podrobovat se) (podiizovat se) (vzddvat se)
"submit to; conform to" "submit to; conform to" "surrender to"
"Belong to" patiit piisluset (sveiovat se)
nalezet
"belong to" "belong to; be owed to" "entrust self to; confide in" "be one's duty to; belong to"
"Fall under the influence of' holdovat "indulge in; pay homage to" propadat "become obsessed with; addicted to; subject to" podlehat "be subordinate to; succumb to; GNE in to" povolovat "GIVE in to" hovet "GIVE in to" "Yield to/respect power of' (plizpusobovatse) ustupovat (klanet se) (dvolit se) (vyhovovat se) (zavdecovat se) vdecit dostat duvelvvat stranit vent
rozumet (divit se) (podivovat se) (obdivovat se) (ucit se) (zvykat si)
"adjust to; adapt to; conform to" "make way for; yield to; make concessions to" "bow to; make obeisance to" "court, woo" "be convenient for; comply with; accommodate" ,'oblige; ingratiate self to' , "be indebted to" "be true to (one's word)" "trust" "side with" ,'believe" "understand' , "be surprised by, wonder at" "be surprised by, wonder at" "admire" "study; learn" "become used to"
In the last set of verbs several (duverovat and those that follow) refer to the subordination of one's mind to another. In a number of these verbs the benefactive aspect of their meaning is particularly salient (e.g., pomdhat, ustupovat,
klanet se, vyhovovat), and this is indicative of the relationship between dativegoverning verbs and the dative of beneficiary expressions captioned by schema
2b. Extension of dative government via antonymy motivates the use of the dative case with verbs which denote the annulment of subordination. Those in the first two groups given below are closely related to the verbs which refer to symmetrical relationships, and the third group provides a transition to the intransitive taking verbs.
Annulment of subordination "Resist" odoldvat (vzpirat se) (protivit se) vzdorovat
"resist; stand up to" "oppose; resist" "oppose; rebel; disobey" "defy; resist"
"Fight back"
odpldcet (mstit se) (kompenzovatse)
"repay; retaliate against" "take revenge on" "repay" (positive connotations)
"Remove self from subordinating relationship" odvykat (si) "get out of the habit of; get unused to" odnistat "outgrow" (odrozovat se) "renounce (one's nationality), defect" (odcizovat se) "become estranged, alienated" (vymykat se) "wrench loose; get away from"
Transitive relationships - domination The next group of verbs represents the converse of subordination - domination which is realized when the balance of power is tipped in favor of the nominative. All of these verbs denote the imposition of power over another, and the use of the dative is significant because it is possible to rule only those entities which have the will and capacity to act on their own. Domination here implies that the nominative entity forces the dative entity to be the agent of acts which are of its (the no-
74
75
minative's) choosing. The nouns used in the dative with such verbs almost invariably refer to individual humanbeings or groupsthereof.
(47)
zavidel Korega mu envied Colleague-NOM him-DAT '(His) colleague envied his success.'
uspech. success-Ace
Domination vladnout vevodit
dominovat piedsedat
ucarovat imponovat poroucet (tesit se) (vnucovat se) naslouchat 84
"govern; rule" "rule over;dominate" "dominate', "chair, act as chairman of" "put undera spell; captivate" "impress' , "command" "enjoy, have at one's disposal (e.g. velke oblibe "great popularity")" "force oneselfon" "listen in on; tap itelefonnim hovorum "telephone conversations' ')"
Unlike the subordination group, verbs of domination are not accompanied by a set of dative-governing verbs that expressthe annulment of this relationship. The reasons for this go beyond linguistics. Human beings are endlessly struggling to escape situations in which they are subordinated to others, and there are many verbs to describe such actions. Situations in which one must make an effort to escape from dominating others are significantly rarer, and there are correspondingly fewer verbs to describe them. Czech does have a verb abdikovat "abdicate" , for example, but it patterns with verbs meaning "leave", and uses the prepositional phrase z+genitive or the bare genitive case. The fact that there are newly borrowed words in all three groups (e.g., konkurovat, asistovat, dominovat) attests to the vitality of the semanticfreight of the dative in such expressions. If the use of the dative were semantically unmotivated and maintained only by tradition85 we would expect new verbs to opt for accusative directobjects.
(48)
zavidel.
envied
This verb clearly has no place among the dative governing verbs and must constitute a specialcase. Its use of the dative does not springfrom ties to "giving" verbs, from a symmetrical or transitive relationship, or from the dative of beneficiary. The verb zdvide: is rathera resultof the lexicalimplication of the dativeof possession in the verb itself. The semantics of zdvide: "envy" entail that the object of the verb (expressed or not) be in the possession of another person. This second party must, therefore, have the sort of personal sphere which is characteristic of the dative. In examples whereenvy is not directedat a specific possession, the nominative (the "envier") interactswith the dative's personal sphere, as in schema 2g. Schema 2g is closely related to schema 2b, which captions the dative of possession as well as otherintransitive free dative uses. Government by nouns Examples of government of the dative case by a noun are infrequent and invariably derived from the verb phrasesalreadycovered. Grepl and Karlik (1986: 268) give only the three examples cited here and affirm that all are collapsed versionsof sentences with corresponding verbs. (49)
slib bratrovi promise-NOM brother-DAT 'a promise to brother'
(50)
prfkaz vojakum order-NOM soldiers-DAT 'an order to the soldiers'
One leftoververb - zdvidet envy' The preceding analysis includes all of the verbs which are said to govern the dative case in Czech, with one notable exception: zdvidet' "envy". Althoughthis verb usually has both an accusative and a dative complement (as in 47), it frequentlyappears withjust the dative case (as in 48).
Kolega mu Colleague-NOM him-DAT '(His) colleague envied him.'
(51)
proof
rodicum
wishes-NOM parents-DAT '(best) wishes to (one's)parents'
76
77
Prose
Pictorial
,'grateful' , "helpful" "used,accustomed' , "alien, strange" cizi ,'conspicuous" zjevnj jasny "clear" srozumitelny ,'understandable' , "accessible' , piistupny dostupny "accessible" "familiar' , znamj "dear" milj ,'precious" vzdcny (ne)pfijemnj "(un)pleasant' ,
vdeCnj napomocny zvyklj
Schema 2g
Usage
A nominative acts on a Dative government of dative's sphere of verb z8videt 'envy'control in a setting. transitional between Free dative & Governed dative
Figure 14. Schema 2g
vdecit pomdhat
zvykat si
odcizovatse jevit se
(subordination) (subordination) (subordination) (annul. of subor.) (intrans giving) (intrans giving) (intrans giving) (intrans giving) (intrans giving) (intrans giving) (benefit) (benefit) (benefit/harm)
The few adjectives that do not bear etymological relationships to dative-governing verbs do at least bear logical relationships to groups of dative-governing verbs.
Government by prepositions In all three examples the dative is governed by a deverbal noun; in a fuller version the dative's role would approximate that of the indirect object or of the dative governed by a verb.
Government by adjectives Cases of government by an adjective are likewise usually the result of the abbreviation of a verb phrase. As indicated in the table below, most of the adjectives that are said to take the dative case are related to verbs that were listed among the dative-governing verbs or verbs that participate in semantic extension of the indirect object in preceding sections. Dative-governing adjectives: "similar, resembling" podobnj "equal" rovny , 'repulsive" odpomy "adverse" protivny "loyal" vemj
Related verbs: podobat se rovnat se odporovat protivit se vent
(symmetrical) (symmetrical) (symmetrical) (symmetrical) (subordination) .
In many ways, k "to" ranks first among dative-governing prepositions. 86 It is the only preposition which uniquely manifests the semantic content of schema Ia; the remainder of the dative-governing prepositions bear relationships to schema la which are analogous to those of the dative-governing verbs, as will be shown below. Bethin distinguishes two meanings of k in concrete instantiations which are disambiguated by the nature of the dative referent.S? If the dative referent is human, the prepositional phrase identifies a destination. Thus in usage with motion verbs k lekati "to the doctor' 'patterns with other destinational prepositional phrases such as: do Prahy "to Prague" and na still' 'onto the table". K is exclusively reserved for destinations which are human beings. In its second meaning, k identifies not actual, but approximate destinations, and could be translated as meaning "in the direction of" or "approaching". Here k is not limited to cooccurrence with nouns denoting human beings; it can be combined with nouns referring to objects or places which are approached, provided that the subject of the pre-position does not actually make contact with the object of the preposition, as in (52).
79
78
(52)
Vladimir pristoupil k Vladimir-NOM stepped-up to 'Vladimir stepped up to the window.'
oknu. window-DAT
Prepositional phrases describing such approximate destinations occupy a unique position in Czech and do not pattern with the k, do, na phrases mentioned above. An examination of the two meanings of k and their relationship to schema la reveals that they are two manifestations of a single concept. In both meanings, the object of the preposition retains a certain inviolate integrity; it remains apart from the subject of the preposition. This independence of the dative referent is symbolized by the dotted arrow in the pictorial schema. The distinction of two meanings is a consequence of extralinguistic reality. Human beings constitute a special subset of destinations because they are neither entered nor contacted when they are reached. 88 Since other kinds of destinations are usually subject to some form of physical encroachment, those that are not require an alternate means of expression. Most metaphorical expressions, such as vztah k "attitude toward; relationship to" , preclude the notion of physical arrival and are extensions of this second meaning of k. A significant number (nine out of twenty-two) of the remaining dative prepositions are secondary formations which include k : v pomeru k
vzhledem k nehlede k s ohledem k se zietelem k smerem k ve smeru k s piihlednutim k ve vztahu k
'in proportion to' 'in view of 'regardless of' 'with regard to' 'with respect to' 'in the direction of 'in the direction of 'with regard to' 'in relation to'
The use of the dative case with all other prepositions is motivated by many of the same kinds of relationships that motivate its use with verbs. Symmetrical relationships proti naproti oproti
"against" "opposite" "in comparison with"
naodpor vstiic v twit napiic umerne(k)
vuei Subordination kviili
"in opposition to" "meeting; facing" "meeting; facing" "athwart' , "in proportion to" "towards; against; in the face of; compared with" (elements of "benefitlharm" perceptible in vuei )
"due to"
Annulment of subordination "in defiance of; despite" navzdory vzdor "in defiance of; despite" Direct object incorporated dfk(y)
"thanks to" (cf. dekovat "thank" in preceding section)
3. 1.2. Schema 1b - dative in impersonal constructions
In schema lb we see that the setting (or some unnamed portion of it) exerts a force on the dative entity, which as a consequence experiences some internal event. 89 This schema captions a variety of constructions in which the dative arguably functions as a subject.P? Typical examples express states induced by prevailing conditions, illness, or the passage of time. (53)
Je mu Is him-DAT 'He is cold.'
zima. cold
(54)
Bylo mi Was me-DAT 'I was sad.'
smutno. sad
(55)
Bratrovi
spatne, (in Grepl- Karlfk 1986) bad
bylo Brother-DAT was 'Brother was sick.'
80
81
(56)
je hodne pres Me-DAT is much across 'I am well over twenty years old.'
dvacet twenty
Mn«
let. years-GEN
(63)
Pacientu Patient-DAT
se REFL-ACC
uZ already
lepe better
dycha.
breathes 'The patient (can) already breathe better.'
If parts of the body are specified, these expressions show a close relationship
to the dative of possession. (57)
(58)
mu Is him-DAT 'His legs are cold.' Je
srdci heart-toe 'I was sad at heart.'
V
In
zima
cold
na on
mi me-DAT
nohy. legs-Ace
bylo was
smutno. sad
Alternatively, the expression can contain a passive verb (which shows neuter singularagreement and lacks an antecedent, and mayor may not be accompanied by the reflexive particle se). There is a further relationship between this type of expression and a free dative (schema 2f) that will be explained below. Example (60) indicates the links that exist between schemata 3 and 4, for chutnat "taste good" is a dative-governing verb. (59)
UZ mi svitlo! (in Grepl- Karlik 1986) Already me-DAT flashed 'Now I remember/understand!'
(60)
Chlapci chutnalo. (in Grepl- Karlfk 1986) Boy-DAT tasted-good '(It) tastedgood to the boy.'
(61)
Nemocnemu se Patient-DAT REFL-ACC 'The patient felt relief.'
(62)
Jak se ti chodf v How REFL-ACC you-DAT walks in tech novych botach? those new shoes-Loc 'How do you like the feel of those new shoes?'
In the last two examples the passive verb expresses an intransitive activity ascribable to the dativereferent. When transitive verbs undergothis process, the resulting expressions belong more to the dative of beneficiary (schema 2b) than among impersonal constructions, for they acquire subjects (via promotion of direct objectsto subjectposition), as in examples (64) and (65). (64)
Petrovi se studovala Peter-DAT REFL-ACC studied lehce. (in Grepl- Karlik 1986) easily 'Peter found studying matheasy.'
(65)
Basnicky se mi Poems-NOM REFL-ACC me-DAT (in Grepl- Karlik 1986) 'I find it hard to memorize poems.'
~zko
hard
pamatujf, memorize
The use of the dative in example (63) implies a modalreading(indicated by the insertion of "can" in the translation). In constructions with infinitiverather than passive verb forms, modality is quite salient, but such constructions are strongly markedas archaic in Czech. Note that the use of the dativein (66) is accompanied by archaic features suchas the formjest and the infinitival ending -ti. (66)
ulevilo. (in Grepl- Karlik 1986) relieved
matematika mathematics-NOM
jest is 'All people (must) die.'
Vsem lidem
umffti. (in Grepl- Karlik 1986)
All people-DAT
to-die
3.2. Schema 2 and variants- free dative
Both semantically and syntactically thisis the most complex section of the dative's cognitive network. Here all that is required is that the dative's sphere of control be affected in any way at all. This is an extrapolation from the way in which the da-
83
82
tive is affected as an indirect object. The strict definition of recipient is expanded to include beneficiaries of actions, possessors of objects and even interlocutors. A human animate being (these meanings of the dative are restricted to nouns which fit this description) possesses a personal space; anything that infringes on or exists in this space enters into a relationship with this person describable by the dative.P! This infringement is invariably evaluated as either positive (benefit) or negative (harm). Because the referent of a free dative is not necessarily the recipient of some object, a variety of predications is possible. In all cases, it is the entire predication rather than just an object which enters into a dative relationship with an entity. Thus the action of both intransitive and copular sentences can be experienced by the dative, as diagrammed in 2b and 2c, respectively. Also, since the schemas by themselves invoke only the general notion of affectedness,92 this concept is often further specified by lexicon and extralinguistic knowledge.
Schemas Pictorial
Prose
A nominative acts on an accusative in a dative's sphere of control in a setting.
Usage
Free Dative
A nominative acts on an accusative to bring it to a dative's sphere of control in a setting.
Transitional examples between Indirect Object & Schema 2
A nominative acts in a dative's sphere of control in a setting.
Intransitive expressions of Free Dative
A nominative is a nominative' in a dative's sphere of control in a setting.
Free Dative in copular constructions
Schema 2 - dative of possession, ethical dative and emotional dative A negatively-evaluated action on an object can result in a dative relationship with an entity only when the object is already in a relationship with the dative entity, as in schema 2. In such a case the dative experiences harm via its relationship to the object, and such a relationship may be realized as possession by the dative either in the physical realm or in the speech-act domain (in which case the dative finds a narrated event in its sphere of control).
Dative of possession This label is used for the sake of convenience rather than accuracy. A more apt name might be "dative of affectedness via possession". Possession acts merely as a vehicle for affecting the dative entity, thus this expression is different in value from ordinary expressions of possession by means of the genitive case or possessive.93 Possession is emphasized but not established by this use of the dative. Although it is possible to find examples for which this difference is so subtle as to be insignificant (as in [67]), the dative usually carries with it a greater context entailing its perception of the event as good or bad. Thus in (68a) which uses the possessive, the son need never know of his mother's death, whereas in (68b), the use of the dative implies that the son knows what has occurred and has reacted in some way. Given that the dative (68b) describes the most usual scenario, it is much more frequent than the neutral possessive (68a). '
Figure 15. Free dative - schema 2 and syntagmatic variants
84
85 Schemas Pictorial
Prose
A nominative seen as an instrumental acts in a dative's sphere of control in a setting.
Usage
02a
dative's sphere of cont~ol less extensive
Dative of attitude, relationship
-f9J 1088
of
ace
P
action of nom realized as identity with predicate nom'
dative's sphere of control less extensive A nominative in a dative's sphere of control in a setting.
Formulaic expressions
loss of noml dative's sphere of loss of control less action exte~
action of nom causes it to be seen as inst ~
~ An action takes place on a dative's sphere of control in a setting.
Transitional examples between Free Dative & Impersonal Dative
A nominative acts on a dative's sphere of control in a setting.
Dative government of verb zavi~t 'envy'- transitional between Free Dative & Governed Dative
Figure 15. Free dative - schema 2 and syntagmatic variants, continued
Figure 15. Localsubnetwork
(67)
a.
b.
(68)
a.
b.
Osus slzy lidf. (in Pit'ha 1971) Dry tears-Ace people-GEN 'Dry the people's tears.' Osus slzy lidem. Dry tears-Ace people-DAT 'Dry the people's tears.'
Jeho matka zemrela, His mother-NOM died 'His motherdied.' matka. Zem'fela mu mother-NOM Died him-DAT 'His motherdied.'
~
87
86
The contention that the dative expresses affectedness via possession rather than possession per se is upheld by the fact that there is no rule preventing this dative from co-occurring with other possessives. In some contexts, (69) was judged acceptable (although marginally so, due to the clutter of pronominal forms) by three out of five native speakers, even though it contains both the dative mu "him" and the possessive jeho "his". (69)
Kdyz se vratil, videl, ze mu returned saw that him-DAT When REFL-ACC shorel jeho dum. burned-up his house-NOM 'When he returned, he saw that his house had burned down (and he was affected).'94
,'her" is somehow alienated from her own arm; perhaps it was cut off prior to the present event. (72b) is altogether unacceptable, for it is impossible to imagine a context in which Peter could be sufficiently detached from his throat while still feeling regret. (71)
b.
(72)
Possessive and dative forms can also coexist when a single object is possessed is some way by two different persons: the possessive indicating one possessor, and the dative the other. (70) is a suggestion made by one cleaning lady to another, and here the notion of possession is interpreted rather broadly. (70)
moje podlahy Kdybys umyla pani Sulcove; my floors-Ace If-you washed Mrs. Schulz-DAT umyla i bych ti tak washed you-DAT and then would-I jejf okna. (in Pit'ha 1971) her windows-Ace 'If you would wash my floors for Mrs. Schulz, I would wash her windows for you.'
Here the dative and possessive forms have virtually equal force, demonstrated by the fact that they are interchanged from one clause to the other. Two kinds of possession are involved: 1. possession via a work relationship and 2. actual ownership. In the first clause a possessive marks (a) and a dative marks (b), yet in the second clause it is the other way around.95 The so-called dative of possession is distinguished from possessive and genitive forms by its range of environments. Datives frequently appear in sentences where the direct object (usually a part of the body) is inalienably possessed by the dative referent. In such sentences the use of a possessive or genitive is strange or unacceptable. For (68a) this results in a dry statements of facts, unusual in the situation which it describes. (71b) is strange, for it implies that the referent of jeji
a.
a.
b.
Zlomil jf ruku. arm-Ace Broke her-DAT 'He broke her arm.' Zlomil jejf ruku. Broke her arm-ACC 'He broke her arm.'
Petrovi Lftost sevrela Peter-DAT Regret-NOM tightened (in Grepl - Karlik 1986) 'Regret tightened Peter's throat.' *Lftost sevrela hrdlo Regret-NOM tightened throat-ACC 'Regret tightened the throat of Peter.'
hrdlo. throat-ACC
Petra. Peter-GEN.
The rule concerning inalienable possession is often overridden by the following rule in which the dative is contrasted with other forms of possession: if the item possessed is modified, the dative form is less felicitous'<', presumably because its use concentrates attention on the object's attributes rather than on its relationship with its possessor. Thus (73) resists replacement of the dative with a possessive, and possessive forms are preferred in (74) and (75). (73)
jf Honza se her-DAT Honza-NOM REFL-ACC ocf. eyes-GEN 'Honza looked into her eyes.'
podfval looked
(74)
podfval Honza se looked Honza-NOM REFL-ACC 'Honza looked into her blue eyes.'
do to
do to
jejfch modrych ocf, her blue eyes-GEN
89
88
(75)
Sledoval kazde jeji hnutf, (Capek) Followed every her movement-ACC 'He followed her every movement.'
As Petr Pit'ha has observed, the dative of possession is invoked by lexical content and extralinguisticknowledge rather than syntax.f? He gives a number of examples which illustrate this fact. In (76) one must know who the room belongs to in order to determine whether the dative describes a beneficiaryor a possessor. (76)
narration. This reading of the dative, like the dative of possession, requires further specification via lexicon or extralinguistic knowledge. When the necessary requirements are fulfilled (the dative referent is the interlocutorand detached from the event), the meaning of the dative is interpreted on a pragmatic plane as a mapped extension of the free dative. This use of the dative illustrates the close interaction of semantics and pragmatics (discussed in chapter 2) and the way in which they are integrated in the cognitivemodel. (79)
ti mel silnou horecku. Vcera jsem Yesterday am-AUX YOU-DAT had strong fever-Ace '(Hey, you know what?) I had a high fever yesterday.'
(80)
Ten Petr ti ukradl auto. That Peter-NOM YOU-DAT stole car-Ace '(Hey, you know what?) That Peter went and stole a car.'
mi pokoj. Uklidil Cleaned me-DAT room-ACC 'He cleaned up my room/the room for me.'
In the following examples lexical content also plays a role (77) refers to a surgical procedure performed on the referent's tongue, whereas the tongue in (78) can only be a part of a slaughtered animal. (77)
Ukar jt chtel rozrfznout Doctor-NOM her-DAT wanted to-cut-open 'The doctor wanted to cut open her tongue.'
jazyk,98 tongue-ACC
(78)
Reznik:
jazyk. tongue-ACC
jt
chtel
rozrfznout
Butcher-NOM her-DAT wanted to-cut-open 'The butcher wanted to cut open the tongue for her.'
The second example is ambiguous, since it could also be read as a dative of possession. Context determines whether or not it will be recognized as a mapped extension of the dative. Since the ethical dative is a pragmatic device of the speech-act domain used by speakers on hearers, it is restricted to use with second person pronouns. In this way it is distinguished from another mapped dative, the emotional dative, which requires first person pronouns.
Emotional dative Semanticextensionvia mapping
Ethicaldative The use of the ethical dative is largely subjective (pragmatic); it is a device employed by the speaker to capture the hearer's attention.P? The speaker maps the case relationship captioned by schema 2 onto the speech-act domain, using the dative sphere to claim the existence of a relationship between the hearer and the narrated event. The interlocutor typically has no relationship to the narrated event or its participants: in actual fact he has neither experienced the event nor does he have any of the participants in his personal sphere. Yet the speaker projects this sort of a dative relationship onto his interlocutor in order to involve him in his
The emotional dative also belongsto the realm of pragmatics, for it instantiates the subjective use of case. In effect, the speaker claims that he has a relationship to an event, even though he is not really involved in it. Again, the speaker is mapping a dative relationship onto the speech-act domain. Here the speaker implies that the narrated event belongs to his sphere of control. It comes as no surprise that the use of this dative is often associatedwith acts of aggression on the part of the speaker. The sentence in (81) might be uttered by a policeman while arresting thieves he has caught red-handed. The policeman need not have witnessed the event, nor need he be the possessor of the stolen goods. By his use of the dative he is asserting his authority in the given situation, thus the personal sphere which he invokes is that of the authorityvested in him.tOO
90
(81)
91
Co jste ndm What-ACC are-AUX us-DAT 'What have you stolen here (on us)?'
tu here
ukradli? stole
Dative of solidarity
A variation on the emotional dative can be used to show solidarity or sympathy l O1 when a dative relationship already exists for the interlocutor. The speaker uses the first person plural to place himself in the hearer's sphere and thus in the same relationship with the narrated event as that experienced by the hearer. Thus the speaker indicates that he shares the hearer's personal sphere even though he is not affected by the event in the same way that the hearer is. The sentence in (82) could be used by a mother in comforting her child. (82)
Ty zIe deti ndm rozbily Those mean children-NOM us-DAT broke vid'? see 'Those mean children broke our toys, did they?'
hracky, toys-ACC
3.2. 1. Schema 2a - dative of beneficiary - transition to indirect object The distinctions between schemas 1 and 2 are quite subtle. Even given the foregoing arguments which establish the special function of die indirect object, one might ask whether or not it would be possible to include it here rather than giving it a separate schema. Indeed, it appears that schema 1 is merely a degenerate version of schema 2, in which the personal sphere has been collapsed. The dative of the indirect object must be recognized as a potential agent of a further action, but the same cannot be said of the free dative. All that is required of the free dative is that it experience the action; it need not carry it any farther. The examples given in preceding sections of this chapter, however, demonstrate that there is no hard and fast boundary, but rather a gradient between these two datives. Occasionally one can find instantiations which can be interpreted as indirect objects or as free datives, for example (11) above 102 and (83).
(83)
Podej mu boty. (in Pit'ha 1971) Hand him-DAT shoes-ACC 'Hand him (his) shoes.'
The dative mu "him" in this example designates the recipient, but also indicates the possessor. The choice of the dative is multiply motivated in this instantiation of schema 2a Clear examples of this dative are given in (8), (9) and (10), to which I will add only two: (84)
Sel bratrovi pro cigarety. Went brother-DAT for cigarettes-Ace 'He went to get some cigarettes for his brother. '
(85)
Devce natrhalo mamince mother-DAT Girl-NOM picked 'The girl picked some flowers for her mother.'
kvetiny. flowers-ACC
The verb vysvetlovat "explain" is a metaphorical instantiation of schema 2a; it describes a situation in which the nominative agent makes ideas (the accusative direct object) available to the dative recipient. For practical reasons, this use of the dative is overwhelmingly associated with benefit. If something is done to an object, this action can be viewed as taking place on behalf of a third party only if it is evaluated positively. The object is prepared for or made available to the dative entity which previously had no relationship to the object; thus the object is introduced into the dative's personal sphere. If the action were viewed negatively, it would rarely if ever facilitate such motion.
3.2.2. Schema 2b - intransitive expressions of free dative All of the meanings of the free dative discussed above have intransitive equivalents which can be captioned by schema 2b. Since these meanings remain essentially unchanged, I will only give a few examples here.
92
93
Dativeof beneficiary (86)
Miminko ndm place v noci. Baby-NOM us-DAT cries in night-LOC 'The baby cries at night (and we are affected by this).'
Dative of possession (87)
(91)
ztratil pes. (in Grepl - Karlik 1986) Piiteli se lost dog-NOM Friend-DAT REFL-ACC '(My) friend's dog got lost.'
Ethical dative (88)
pragmatic; in the first type the speaker tries to make an impression on the hearer and in the second the speakercomplains about the subject of the sentence. The sentence below (91) was uttered by a woman talking to a toddler who was afraid of her vacuum cleaner. She wanted to impress upon the child that the vacuum cleaner was a good thing to have, and did so by projecting her affirmation of this into the child's personal sphere. Neboj Not-fear
se, ten luks je REFL-ACC that vacuum cleaner-NOM is ti tak hodny! you-DAT so good-NOM 'Don't be afraid, (hey, you know what?) the vacuum cleaner is so nice!'104
An example of the emotional dative is seen in (92), which is markedlysarcastic in tone.
Vcera
nemohl usnout. jsem ti could-not to-fall-asleep Yesterday am-AUX YOU-DAT (in Grepl Karlik 1986) '(Hey, you know what?) YesterdayI could not fall asleep.'
(92)
To je mi That-NOM is me-DAT (in Grepl - Karlik 1986) 'That is a nice mess (jor me) !'
pekny poradekl fine order-NOM
Emotional dative 3.2.4. Furthersyntagmatic variants of schema 2 (89)
(90)
nebudes Ty nam ttl here will-not You-NOM us-DAT (in Grepl - Karlik 1986) 'You are not going to goof off here (on us)!' Rostou ndm Grow us-DAT 'Our teeth are comingin.'
Ienositl to-be-lazy
There are several classes of less frequent uses of the free dative which, due to their characteristic syntactic structures, suggest further minor variations on schema 2. The meaningswhich they invoke do not deviatefrom those treated above in any significantway.
zoubky.103 teeth-NOM
3.2.3. Schema2c - free dativein copularsentences Only mappedextensions of this syntagmatic variantexist. The predicate attributed to the subject is brought into a dative relationship with either the hearer (ethical dative) or the speaker (emotional dative). The function of this dative is predictably
Schema2d Schema 2d is a close relative of 2b. In it, a dative relationship is projected over a predicate instrumental. The structure and meaning of the predicate instrumental (now losing ground in Czech) will be addressed in chapter 5. Examples such as (91), (92) and (93) are collectively referred to by Czech linguists 105 as dativ zietelovy "dative of attitude, relationship".
94
(93)
95
Karel byl mym detem druhym otcem. Karel-NOM was my children-DAT secondfather-INST 'Karel was a second father to my children.'
(97)
a. b.
Schema2e In schema 2e we see a collapsed version of schema 2. A simple bare nominative
(see Figure 9), usually an abstractnoun denotingglory or woe is put in a relationship with the dative referent. This is typical of formulaic exclamations, and is common in slogans. (94)
Slava Komunisticke strane Ceskoslovenska! Glory-NOM Communist party-DAT Czechoslovakia-GEN 'Glory to the CommunistParty of Czechoslovakia!'
Schema2f Schema 2f captionsimpersonal constructions which include a free dative. (95)
mi Naprselo Rained me-DAT 'It rained into my room.'
do to
pokoje. room-GEN
Examples such as (95) give evidence of transition between schemas 2 and 1b. Sentences captioned by schema 2f express externally realized affectedness, whereas in those captioned by lb the dative referent experiences the situation described by the verb in his own being. Compare the similarities between (96b), which belongs to schema 2f, and (97b), which belongs to schema lb. (96)
a.
b.
Prselo pokoje. do room-GEN Rained to 'It rained into the room.' do Prselo mi me-DAT Rained to 'It rained into my room.'
pokoje. room-GEN
Bylo zima, Was cold 'It was cold.' Bylo mi Was me-DAT '1 was cold.'
zima. cold
Although (96b) and (97b) have much in common and their syntactic structures are virtually identical, they do not enjoy the same status in the cognitive network of the dative. (96b) clearly has ties with the free dative which (97b) does not, and (96b) represents a marginal phenomenon. Both of these facts are reflected in the treatmentof theseexamples by the cognitive model. Schema 2g Schema 2g is transitional between schemas 2 and la. It is associated with only one verb zdvidet "envy", and it is not surprising that this verb is traditionally grouped with those that govern the dative. A discussion of this verb and its schema is includedin the sectionon the governed dative, above.
3.3. Dative network bound by syntagmatic variantsofthe schemas
The network presented at the beginningof the analysisrepresents the probable genesis of the dative category in which paradigmatic variation is responsible for the derivation of schema 2 from schema 1. Syntagmatic variants of these two schemas produce a secondaryset of relationships that further bind the schemas in their network. Schema 2a occupies a transitional zone between schemas 1 and 2, as do the verbs of benefit and harm. The two schemas are also linked by two bridges, each composed of syntagmatically parallel variants. The first bridge follows a path along la - 2g, both of which lack an accusative object in their schemas. The second bridge stretches across lb - 2f which lack both an accusative object and a nominative subject in their schemas. Schema 2 is thus linked to all of schema I's variants and shows more vigorous development of syntagmatic variants than schema 1.106 This observation emphasizes the dynamic forces at work in cognitive networks and has importantimplicationsfor language change. Schema 2 could conceivably come to dominate the network, relegating schema 1 to
96
97
peripheral status, in whichcase it might come to be expressed by a prepositional phrase. This has indeed happened with the Slavic comitative instrumental, which was purportedly a "core" meaning of the CommonSlavic instrumental, but now appears only in prepositional phrases, as will be shown in chapter 5. 3.4. Reflexive uses ofthe dative
The Czech dative reflexive particle si exhibits surprisingly high frequencylO7 and semantic range, particularly in the spoken language. Standard texts of Czech usually describe meanings related to the indirect object, and mention the possibility of reciprocal uses. Learners are thenencouraged simplyto memorize verbscommonly accompanied by si, leaving the remaining "extra" uses of si unexplained. 108 In the present framework, however, all uses of the dativereflexive are semantically motivated by their relationships to other uses of the dative. The meanings of the particle si involve deformations and extensions of schemas already presented. The resulting schemasare organized in two related networks - one containing reflexive meanings and the othercontaining meanings of reciprocity.
Intransitive Giving & Taking
Dative-governing verbs used impersonally
3.4.1. Reflexive network The starting point for this subnetwork of the dative is the type of construction in which si functions as an indirect object. 109 Classic examples of course involve the verbs ddt "give" and vzit "take", which become virtually synonymous when used with the reflexive pronoun, as is clear from schema Refll. Ddt si is somewhat restricted, for it refers almost exclusively to food (a meaning shared with vzit si ), and vzit si has among its uses the meaning "marry", which is usually classedas idiomatic.
Figure 16. The binding of the Czechdative network via syntagmatic parallelism and transitional examples
(98)
(99)
Prosfrn
te, rekni cfSnfkovi, you-ACC tell waiter-DA T dam rybu. T-will-give fish-Ace 'Please tell the waiterthat I'll havefish.'
ze
si
I-ask
that
REFL-DAT
Stale Still
Petr Peter-NOM
nerozumfm, l-not-understand
proc why
si REFL-DAT
vzal took
98
99
Evu. Eva-ACC 'I still don't understand why Peter married Eva.' As one progresses from verbs meaning "give" and "take", which are central to the building of indirect object constructions in Czech, to less and less proximate synonyms, one moves to a situation in which the accusative object is brought into the sphere of the nominative/dative rather than being directly presented to it, as in schema Refl2a. Observe this gradual transition as illustrated in sentences (100) through (103). (100) Vezmi si destnfk, REFL-DAT umbrella-ACC Take 'Take an umbrella (jor yourself). , (101) Koupfm si motorku. I-will-buy REFL-DAT motorcycle-Ace 'I'll buy myselfa motorcycle.' (102) Vypujcil si knihu. Borrowed REFL-DAT book-ACC 'He borrowed a book (jor himself). ,
si mfr. 111 (103) Prejeme We-wish REFL-DAT peace-ACC 'We want peace (jor ourselves).' (102) describes a situation in which an object becomes a temporary member of a dative's personal sphere, and (103) expresses a desire for a state (peace) to become part of a dative's environment. Other typical examples, (104) to (106), invoke nominative subjects which avail themselves of opportunities or information by bringing them into their dative sphere.
si jiz obstarala praci (104) Vera Vera-NOM REFL-DAT already found work-ACC na univerzite. at university-Lac 'Vera has already (after much trouble) found herself a job at the university.' .
Schemas Pictorial
Prose
Usage
A nominative acts on an accusative to bring it to a dative in a setting. The dative retains independent status. The nominative & the dative are one.
Reflexive Indirect Object
A nominative acts on an accusative in a setting. The setting is a dative sphere of control.
Reflexive Free Dative (Possession & Beneficiary)
A nominative acts on an accusative to bring it to a dative's sphere of control in a setting. The nominative and the dative are one.
Transitional Examples between Refl 1 & Refl 2 .
setting
A nominative acts in a setting. The setting is a dative sphere of control.
Network
Figure 17. Reflexive dative network
Intransitive expressions of Reflexive Free Dative (self indulgence)
100
101
(105) Zapsala
jsem
si
Wrote-down
am-AUX
REFL-DAT
ale
ji
ted'
jejf novou adresu, her new address-Ace
(109) a.
nemohu najft. but now it-ACC l-ean-not to-find 'I wrote down her new address tfor myself), but now 1 can't find it.' si
vstfpil
do
REFL-DAT
engrained
to
jsem
Wrote-down
am-AUX
jejf novou adresu
b. (106) Seifertovy verse jsem Seifert's verses-Ace am-AUX pameti,
Zapsala
si
do
REFL-DAT
adresare.
her new address-Ace to address-book-GEN 'I wrote down her address in my address-book.' Zapsala jsem si REFL-DAT Wrote-down am-AUX
jejf novou adresu. her new address-Ace 'I wrote down her new address (Jor myself). ,
memory-GEN 'I have engrained Seifert's verses in my memory.'
(110) a. It should be noted that this schema also captions reflexive uses of the prepositional phrase k sobe.
Dite
si
odstranilo
trfsku
z.
Child-NOM
REFL-DAT
removed
sliver-Ace
from
prstu finger-GEN
(107) Eva
k Eva-NOM to zacata jfst. started to-eat
sobb
pritahla
REFL-DAT
pulled-to
talft plate-ACC
a and
b.
'The child removeda sliver from his finger.' Dite si odstranilo trfsku. Child-NOM
REFL-DAT
removed
sliver-ACe
'The child removeda sliver (Jor himself).,
'Eva pulled the plate toward herself and started to eat.' The tendency to unite opposites (i.e., semantic extension via antonymy) that was observed for nonreflexive uses of the dative has some correlates for si; besides verbs that denote the bringingof objects into the nominative/dative's sphere, there are some that entail the removalof objectsfrom the sphere,as in (108): (108) Dfte Child-NOM
si
odstranilo
trfsku
z
REFL-DAT
removed
sliver-Ace
from
prstu. finger-GEN
'The child removeda sliver from his finger.' Examples (105), (106) and (108) all bear some resemblance to reflexive dative of possession, and thus constitute transitional instantiations between schema Refl2a and its sister schema Refl2, which captions meanings of benefit and harm as well as possessive uses. Note, however, that the presence of a possessed object, and therefore the reading of possession, is optional for most transitional uses.
Although the phrase used in (106), vstipit si do pameti "engrain in one's memory", is listed by Kralfkova (1977) as representing an indirect object construction, the synonymous phrase ulozit si do pameti "commit to memory" is classed with dative of possession since the si in this second phrase can be replaced by the possessive sve. This again demonstrates the close relationships between the two main schemasof the reflexive network. Two important features distinguish schema Refl2 from Refl2a. First, there is no movement of the accusative object into the dative sphere - it is already there. Second, the roles of dative and setting merge. The resulting field is a special private universe belonging to the subject of the sentence. This second feature is perhaps not very significant for garden-variety reflexive dative pronounsthat indicate possession, but it figures prominently in the semantics of the remaining uses of si. The meaning of this feature for the dative reflexive pronoun could best be captured in English by the phrase for one's own sake .112 Typical constructions use both objects which are inalienably possessed (111) as well as ones which are not (112). (111) Umyj si Wash REFL-DAT
ruce hands-ACC
a and
pale
si
then
REFL-DAT
102
103
je hezky utri! them-ACC nicely wipe 'Wash your hands and then wipe them off (for yourself) nicely!' rozkladajf (112) Tibet'ane ve Lhase si set-up Tibetans-NOM in Lhasa-toe REFL-DAT kramky prfmo na ulici. shops-Ace right on street-LOC 'Tibetans in Lhasa set up shop (for themselves) right on the street.' Most uses of si, however, involve some sort of evaluation of the event as being either to the benefit or detriment of the nominative/dative participant. If a verb denoting harm is used with a possessed object, the reading is that the event must have been an accident, but that the agent is at fault for being careless, since it took place entirely within his dative sphere, Le., he did it to himself. usekl si (113) Petr cut-off Peter-NOM REH...-DAT (in Grepl - Karlik 1986) 'Petr cut off his thumb with an axe.' (114) Matka si Mother-NOM REH...-DAT (in Grepl - Karlik 1986) 'Mother broke her glasses.'
rozbila broke
palec thumb-ACC
sekyrou. axe-INST
bryle. glasses
(116) Dej Give
si
mu him-DAT
pres hubu jak REFL-DAT across up-sec how chces, je mi to jedno. 113 you-want is me-DAT that-NOM one-NOM 'Give him a sock in the jaw (for your own enjoyment) if you like, I don't care.'
In the absence of an accusative direct object, as in schema Refl2b, harm via possession ceases to be an option, and the use of the dative reflexive is always associated with a meaning of "off in one's own world" or "according to one's own wishes/beliefs", or a combination of these meanings (which are clearly compatible). (117)
Zill si tam jako Lived REFL-DAT there like 'They lived it up there like gods.'
behove. gods-NOM
The use of the dative reflexive in this meaning has a variety of kinds of status. It can be conventionalized, grammaticized, or lexicalized. The following context sets up an example illustrating the conventionalized use of si. A friend of mine
Unless the verb clearly expresses harm to the nominative/dative referent, as in the previous two examples, the connotation of the dative reflexive is invariably "for one's own good/comfort/enjoyment". Here the dative reflexive clitic pronoun is no longer an essential component of the clause. Instead it is added to a construction that is already grammaticallycomplete. The speaker maps a dative relationship onto his subjective perception of the narrated event, thus the following are examples of the second type of semantic extension, i.e, extension via mapping onto the speech-act domain. (115) Petr Peter-NOM
prostudoval. (in Grepl- Karlik 1986) studied-through 'Petr has already studied that book (for his own good).'
si
ttl knihu
uz
REH...-DAT
that book-ACC
already
used to work nights while her husband was in medical school, so she would prepare dinners for him to heat up. Invariably she would return to find his meal burned beyond recognition, and when she asked what had happened, he would reply:
si! (118) Cetl jsem Read am-AUX REFL-DAT 'I was reading (for my own enjoyment, not noticing anything beyond the small universethus created)!' The fact that the verb elst "read" becomes intransitive in combination with subjective si is clearly demonstrated by the following two examples (both from Syllaba's biography of Jan Gebauer), which illustrate the different valence patterns of elst with si as an indirect object (119) and with the subjective si of selfindulgence (120). With the subjective si the object appears in a prepositional phrase; it cannot appear as an accusative direct object.
105
104
(119)
(120)
jeste 0 Stftnem musf se pracovat About Stitny-LOC must REFL-ACC to-work still malo kdo monograficky. Krome Erbena in-monograph-form Except Erben-GEN few who-NOM z neho co cetl, a from him-GEN what-ACC read and ti nasi filosofove, kterf 0 nem those our philosophers-NOM who-NOM about him-LOC si precetli filosofujf, zbezne REFL-DAT read philosophize frequently jen to, co je vytisteno a and just ma~ACC wha~NOM ~ prinred cemu namnoze nerozumejf, what-DAT largely not-understand 'A monograph on 5tftny must be prepared. Aside from Erben, hardly anyone has read any of his works, and our philosophers who philosophize about him often have read (for themselves) only what has been published, and even that they largely fail to understand. ' Pres to, ze to byla vlastne Across that-ACC that that-NOM was actually vedecka kniha a prvnf, jiz jsem scientific book-NOM and first-NOM which-Ace am-AUX cetl, zaujala mne a cftal jsem and read-ITER am-AUX read captivated me-ACC si v ill vzdy rad, takze REFL-DAT in it-LOC always glad-NOM so jsem ji umel skoro nazpamet'. am-AUX it-ACC knew almost from-memory 'Even though it was actually a scientific book and the first one which I had read, it captured my interest and I always liked to read (Jor my own enjoyment) from it, so I knew it almost by heart.'
Other verbs that have a conventionalized use of si include hrdt "play" , when it describes the kind of play that children engage in and myslet "think", when its meaning approximates "have an opinion" .114 Si is conventionally added to tici "say" when the message is uttered on the speaker's own initative, motivated by his own desires, as in (121).
(121)
Mamo, nemej starosti! Jestli budu mit Mom- VOC don't-have worries-Ace If I-will to-have hlad, tak si reknu, hunger-ACC then REFL-DAT l-will-say 'Don't worry, mom! If I'm hungry, I'll say so (according to my own good/desires). '
In some environments, this use of the dative reflexive has become grammaticized. This is true of certain verbs expressing intransitive actions which are performed for the agent's comfort or convenience, actions performed for his own enjoyment, or actions which are strictly limited to the agent's sphere of mental awareness. The table summarizes the obligatory uses of si in these meanings.U'' Comfort/Convenience sednout si lehnout si diepnout si stoupnout si 116 odpocinout si hovet si 117 oddechnout si
odskocit si vymyslet si
stezovat si Enjoyment zatancit si
popovidat si 118 hazet si kopat si 119 piihnout si
Awareness pamatovat si vsimat si
uvedomit si piipoustet si
"sit down" "lie down" "squat" "take a standing position" "rest" "rest" "be relieved, take a rest" "relieve oneself (by going to the restroom; lit: "leap away")" "make things up" "complain' ,
"do some dancing" "have a chat" "play ball (throwing)" "play ball (kicking)" "get drunk (lit: "bend"; describes motion of drinking)' ,
"remember' , "notice" "become aware" "accept (a distressing fact)"
107
106
The present explanation of the use of si with these verbs as resulting from grammaticization of a semantic extension of the dative category is particularly valuable given the fact that a similar use of the dative reflexive is attested neither in Old Church Slavic nor in other Slavic languages,120 and thus it cannot be motivated on historical grounds. Unless we recognize the connection between the semantics of these verbs and the pragmatics of the Czech dative reflexive, such examples can only be classed as exceptions resulting from the capriciousness of language. The pragmatic use of the dative reflexive to indicate self-indulgent behavior has also been fixed in the Czech lexicon via the word sobec "egotist, selfish person" , which is formed from the long-form dative reflexive (sob€) with a suffix used for nouns denoting human agents (-ec ).121 When this subjective si is used with reference to inanimate objects personification results, as in: (122)
chtely, Hodinky si sly, jak Watch-NOM REFL-DAT walked how wanted 'The watch ran however it wanted to (to suit itself). ,
(123)
V te debe pfisla jeste 0 jednoho syna, byl In that time-LOC came more about one son-Ace was maly, zustal doma, ale tady si small stayed home but even here REFL-DAT ho smrt nasla. him-ACC death-NOM found 'At that time (i.e., during the war) she lost one more son. He was small and stayed home, but even here death found him (for its own designs).'
If the speaker maps a dative relationship onto the speech-act domain, grouping the hearer with the narrated event, the implication is that the message is being given for the hearer's own good. Because the speaker has taken it upon himself to define the hearer's sphere of control, the tone is authoritative and somewhat threatening, especially well-suited to the anonymous commands that blare out over loudspeakers in the Prague metro, as in (124). (124)
Vystupte si z bezpecnostnfho pasul 122 Step-out REFL-DAT from safety zone-GEN 'Step out of the danger zone (for your own good)!'
This use of the reflexive si emphatically excludes the speaker, and the force of this self-imposed exclusion is quite clear in examples such as (125). (125)
HIed' si sam Watch REFL-DAT self-NOM 'Mind your own business!'
sebe! REFL-ACC
In these last two examples si is wielded like a lasso, and used to keep one's interlocutor in his place.
3.4.2. Reciprocal network
The presence of plural referents for the dative reflexive opens up the possibility that its meaning may be interpreted as reciprocity (i.e., to each other). Whereas the preceding network of reflexive schemas consituted a variation on the basic dative network, that containing schemas 1 and 2, the network of reciprocity largely excludes variations on schema 2 and is therefore more limited in scope. Like both the basic dative network and the reflexive dative network, the network of reciprocity uses the indirect object as its starting point, and this meaning is captioned in schema Recipl. Here an exchange of objects is involved. (126)
Dali si facky/darky. Gave REFL-DAT slaps/gifts-ACC 'They exchanged slaps/gifts.'
(127)
President a generalnf tajemnik si President-NOM and general secretary-NOM REFL-DAT podali ruce. put-forth hands-Ace 'The president and the general secretary streched forth their hands (to each other, in anticipation of a handshake).'
(128)
Vymenili si kosile. Exchanged REFL-DAT shirts-Ace 'They swapped shirts (to each other).'
109
108 Pictorial
Prose
Usage
A nominative acts on an accusative to bring it to a dative, and a nominative' acts on an accusative' to bring it to a dative' in a setting. The dative and the dative' retain independent status. The nominative and the dative' are one. The nominative' and the dative are one.
A nominative acts on a dative, and a nominative' acts on a dative' in a setting. The dative and the dative' retain independent status. The nominative and the dative' are one. The nominative' and the dative are one.
Reciprocal Indirect Object
It is also possible to get readings of possession, in which case the accusative objects are not mutually transferred, but merely acted upon. This is only possible for verbs denoting harm.
si ruce. (129) Zlomili Broke REFL-DAT arms-Ace 'They broke each other's arms.'
Reciprocal Dative Government
Reciprocal A nominative acts on an Free Dative accusative in a dative's (Possession) sphere of control, and a nominative' acts on an accusative' in a dative's sphere of control in a setting. The nominative and the dative' are one. The nominative' and the dative are one.
Examples like (129) suggest the presence of a dative sphere similar to that in schema 2; cf. schema Recip2. It should be noted, however, that this schema is quite marginal in the reciprocity network, and that its use is restricted almost exclusively to instantiations describing harm to parts of the body. As with the non-reciprocal dative, the indirect object is subject to extension. With reciprocity both the entire network and the category of the indirect object have been contracted. Here we see only examples of semantic extension via metonymy in which the direct object (a message) has been subsumed in the semantics of a verb of communication. Verbs of communication iikat si "say to each other" povidat si "converse with each other" tykat si "say ty to each other" vykat si "say vy to each other" psdt si "write to each other" dopisovat si "write letters to each other" Next come verbs which can be properly said to govern the dative, and it comes as no surprise that the same semantic groups of verbs that were found to be responsible for non-reciprocal dative government crop up here again. Given that all reciprocal datives express mutual action, all can be said to instantiate a symmetrical relationship. Some of them also appear without si under the heading "symmetrical relationships" in the discussion of the basic dative network, for example:
Network
~IOSSOf~exter":alization ~ ~acc
~ofdatlve
Figure 18. Reciprocal dative network
~
odporovat si piekazet si
"oppose, contradict each other' , "be in each other's way, interfere with each other"
The remaining verbs with which si is used in a reciprocal meaning invoke transitive relationships, specifically those of serving, subordination, and yielding:
110
111
pomoci si
rozumet si vent si
1.
"help each other" "understand each other" "believe each other"
Understandably, the group of verbs which denote domination is not represented here; for although we often find people yielding to each other in mutual relationships, it is hard to think of examples of relationships in which people mutually dominate each other. Curiously, the prepositionalphrase k sobe can have the reciprocal reading "to each other" only when the dative reflexivesobe refers to the object of the sentence, as in (130):
The speakercontrolsa sphereencompassing a) the narrated event only b) the hearerand the narrated event 2. The hearercontrolsa sphereencompassing a) the narrated event only b) the speakerand the narrated event. A representative sampling of mapped extensions of the dative is reproduced
below, with a paraphrase of the pragmatic moment of each type of usage given in double quotes. i)
vsechno pradlo, jsem alllaundry-ACC Folded am-AUX jsem ponozky k sobe. am-AUX socks-ACC REFL-DAT to 'I folded all the laundryand matchedthe socks.'
(130) Slozila
a and
dala gave
mapping onto the domain of subjective perception - self-indulgence ''The subject is engaging in an action in his own sphere." se tu dreme a We-NOM REFL-ACC here toil and on si sed! v hospode! he-NOM REFL-DAT sits in pub-Lee 'We're toiling away here and he's sitting in a pub (enjoying himselj)!'
(131) My
3.5. Semantic extension via mapping - pragmatic uses ofthe dative Perhaps the most strikingcharacteristic of the Czech dative is its capacityfor pragmatic usage, examples of which have been presented where relevant throughout this chapter. It would be appropriate at this point to collect these examples and assess their implications. All of the pragmatic uses of the Czech dative involve semantic extension via mapping of schema 2, one of its syntagmatic variants, or reflexive equivalents thereof. There are two types of mapping that take place: i) mapping onto the domain of subjective perception, and ii) mapping onto the speech-actdomain. The first type of mappinghas rather limited distribution - it occursonly with the dative reflexive si - and carries the implication that the dative referent is acting selfindulgently. The second type has four further subtypes, classed according to the way in which the speaker, hearer, and narrated event are grouped with respect to each other. Because here dative relationships are mapped onto the speech-act domain, this usage is limited to first- and second-person pronouns and the reflexive si (when used to refer to a first- or second-person participantin the speech event. i.e. the speaker or the hearer). A speaker may map the dative onto the speech-actdomainin order to make any of the following claims:
- plus conventionalized, grammaticized, and lexicalized uses of the dative reflexive. ii)
mapping onto the speech-act domain- the speaker makes the following claims:
la) The speaker controls a sphere encompassing the narrated event complaints ''The narratedevent is in my sphere, and I'm stuck with it!" (132) To
je mi That-NOM is me-DAT 'That's a fine messfor me!'
pekny poradekl fine order-NOM
1b) The speakercontrols a sphereencompassing the hearer and the narrated event - assertion of authority "You and your actions are in our sphere, so behave!"
112
(133) Jak How
tu
ndm
jezdfte?
4.
here
you-drive 'How are you driving here on us?' (uttered by a policeman chastizing a us-DAT
motorist) 2a) The hearercontrols a sphereencompassing the narrated event - ethical dativewith nonreflexives, authoritative warningwith reflexives ''The narrated event is in your sphere- take notice of it." (134) Ten caj
ti
me
zvedl.
That tea-NOM YOU-DAT me-ACC picked-up 'Hey, you know what, that tea picked me up.'
"Your actions are in your sphere, therefore your responsibility. I'm just telling you this for your own good." (135) Vystupte Step-out
si
z
bezpecnostnfho pasu,
from safety zone-GEN 'Step out of the danger zone (for your own good). , REFL-DAT
2b) The hearercontrolsa sphereencompassing the speaker and the narrated event- solidarity "I share your sphere and your experience of the narratedevent." (136) Ty zle deti Those mean children-NOM
ndm
rozbily
us-DAT
broke
hracky, toys-Ace
vid'?
Comparison of cognitive networks dative
A brief comparison of the cognitive networks of the dative in Czech and Russian is instructive for the insights it gives on the plasticity and evolution of case categories. Since both languages developed from a single protolanguage, we would expect the structure of their case categories to be essentially similar, and indeed they are. The basic structure of the dative remains the same for both languages, and significant variations are largely limited to the peripheries of these structures and to transitional zones withinthem, or to the relative frequency of usage of various parts of the network. 124 The following comparison will highlight significant differencesbetweenthe two languages.
4.1. Schema 1 - indirect object
As in Czech, true examples of the Russian indirect object are instantiated only with the verb dat" "give" and its closest synonyms, where the dative can neither be replaced by a prepositional phrase, nor omitted from the clause. Thus, only (137) illustrates the dative of schema 1; (138) - (141) show the transition to schema 2 (via schema 2a), and (142) deviates from schema 1 in the direction of schema Ia, for it is closely related to one of the transitionalgroups of verbs (denoting "giving verbal signals") that links these schemas. (137) Ljudmila
dala emu Ludmila-NOM gave him-DAT 'Ludmila gave him a flower.'
cvetok. flower-Ace
see
'Those mean children broke our toys, did they? (uttered by a mother comforting her child)' Given that pragmatics entails the grouping of speech act participants relative to the narratedevent, and that schema 2 of the dative (and all syntagmatic variants of this schema) invokes a boundary, it is surely no coincidence that speakers would find the dative personal sphere a handy device for corralling their interlocutors. 123 Such pragmatic uses are a logical extension of the dative's meaning in schema 2: they are motivated on semantic grounds and consistent with the network of the dative outlinedin this chapter.
the Russian
(138) Ljudmila
kupila
Ludmila-NOM bought 'Ludmila bought him a hat.'
(139) Ljudmila
emu
sapku.
him-DAT
hat-ACC
prigotovila emu him-DAT Ludmila-NOM prepared 'Ludmila prepared lunchfor him.'
obed. lunch-ACC
114
115
(140) Ljudmila pokazala emu Ludmila-NOM showed him-DAT 'Ludmila showed him the way home.'
dorogu way-ACC
(141) Ljudmila prislala Ludmila-NOM sent 'Ludmila sent him a letter.'
pis'mo. letter-ACC
emu him-DAT
emu (142) Ljudmila skazala Ludmila-NOM told him-DAT 'Ludmila told him her name.'
domoj. home
(145)
*A vtobus Bus-NOM
nam
us-DAT 'The bus rode off on us.'
(snit'sja) (grezit'sja) (merescit'sja)
svoe imja.
(143) Poet predstavilsja Poet-NOM introduced-REFL sobravsimsja i nacal eitat' gathered-PAST ACf PART-DAT-PL and began to-read svoistixi. own verses-Ace 'The poet introduced himself to the people who had gathered and began to recite his verses. ' vzjal took
(kazat'sja)
"introduce self to' , "fall to one's lot; pass to (by inheritance)" "come to; let self be caught" "seem to" "appear to (in a dream)" "appear to (in a dream)" "appear to, haunt"
own name-ACC
The Russian indirect object is considerably more constrained than its Czech counterpart, for it can be extended only to intransitive giving (143), but not to taking (144), nor to intransitive taking (145), as is the case with Czech.
(144) *On mne He-NOM me-DAT 'He took my last kopeck.'
Intransitive giving (predstavljat'sja) (dostat'sja) (davat'sja)
poslednjuju kopejku. last kopeck-Ace
uexal. rode-off
There are a number of groups of verbs which link schema la to schema 1. Intransitive giving verbs and verbs that incorporate a direct object both fall in this transitional zone. Verbs of benefit and harm provide examples of lexicalization of the semantics of the dative relationship captioned by schema 2 and its variants, thus linking schema la to both schema 1 and schema 2. All of these transitional verbs represent semantic extension of the indirect object via metonymy.
Incorporation of direct object "Giving verbal signals" poddakivat' tykat' vykat' govorit' otvecat' sovetovat' l'stit' grozit' zvonit' napominat' velet' prikazyvat' (molit'sja)
"say yes to" "say ty to; speak on informal terms with" "say vy to; speak on formal terms with" "speak" "answer" "advise" ,'flatter' , "threaten' , "call" "remind" "order" "order" "pray"
"Giving nonverbal signals" aplodirovat' "applaud" rukopleskat' "applaud" (radovat'sja) "rejoice at" (ulybat'sja) "smile" (usmexat'sja) "grin" (smejat'sja) "laugh" (klanjat'sja) "bow before" maxat' "wave" morgat' "blink; wink" "wink" podmigivat' kivat' "nod"
116
pisat' signalizirovat' telegrafirovat' radirovat' demonstrirovat'
117
"write" "signal" "telegraph' , "radio" "demonstrate' ,
podpevat' sootvetstvovat' (ravnjat'sja) (upodobljat'sja) podrazat' soiuvstvovat'
"Giving money or gifts" platit' zertvovai' kompensirovat'
"pay" "donate" "compensate"
In addition to the verbs of verbal signals given above, Popova (1974) has noted the deviant but popular use of blagodarit' "thank" with the dative rather than the accusative case, as in standard usage. This fact and the presence of numerous new borrowings (signalizirovat', telegrafirovat', radirovat', demonstrirovat', kompensirovat') confirm that the bond between schemas I and la is strong and cognitively actual.
4.1.1. Schema la - governed dative Government by verbs The system of relationships cued by Russian verbal government of the dative is based on the same principles as that of Czech verbal government, but is somewhat more conservative. Both symmetrical and transitive relationships are instantiated, but of the transitive relationships, only that of subordination is realized in Russian (domination is marked instead with the instrumental). The group of verbs denoting annulment of subordination is more weakly developed, and most of its members can also be considered examples of verbs denoting symmetrical relationships. Russian has a variety of groups of verbs that occupy the transitional zone between schemas 1 and Ia, and correlate well with the groups identified for Czech, with two exceptions: Russian lacks intransitive taking verbs (a fact which is consistent with the failure of the Russian indirect object to extend to verbs of taking), and verbs which denote the giving of punishment. Symmetrical relationships akkompanirovat' vtorit'
"accompany (in music)" "play second part to (in music); echo"
soboleznovat' simpatizirovat' perecit' protivorecit' prepjatstvovat' mesat'
vozrazat' protivodejstvovat'
(protivit'sja) (soprotivljat'sja) protivoborstvovat'
"sing along; echo" "correspond to" "be equal to" "become like; assimilate to" "imitate' , "be sympathetic to" "sympathize with, condole with" "sympathize with" "contradict' ' "contradict' ' "hinder, interfere" "hinder; annoy" "oppose" "oppose; counteract" "oppose; resist" "resist; oppose" ,'oppose, fight against"
As is the case with Czech, there is significant semantic overlap among groups of verbs. The three verbs which refer to musical accompaniment overlap with verbs denoting 'serve' of the subordination group. The last five verbs on the list might also be taken as denoting annulment of subordination. Soiuvstvovat' "be sympathetic to" and simpatizirovat' "sympathize with" have palpable ties to the verbs of benefit. The presence of the new borrowing simpatizirovat' "sympathize with" attests to the vitality of this portion of the dative's cognitive network.
Subordination "Serve" sposobstvovat' pomogat' sode.jstvovat' slusit' (porabotit'sja)
"assist, facilitate" "help" "assist, help; contribute to" "serve" "become a slave to"
"Subordinate self to" prisjagat' (v vernosti) podlezat'
"swear (allegiance) to" "submit to"
119
118
(povinovat'sja) (podvergat'sja) (podcinjatsja)
(pokorjat'sja) sledovat' predsestvovat' soputstvovat'
"obey" "undergo, be subject to" "submit to; subordinate self to" "submit to; resign self to" "follow; comply with" , 'precede' , "accompany' ,
ment with this verb is gaining popularity due to analogy to other dative-governing verbs that denote amazement (cf. Gorbacevic 1971). As mentioned above, Russian does not develop the use of the dative with verbs meaning annulment of subordination to the same extent that Czech does. Russian has fewer verbs under this heading, and one subgroup present in Czech, labeled "remove self from subordinating relationship", is altogether missing in Russian. 125
"belong to" "devote, give self to" (also said of a woman giving herself in marriage)
Annulment of subordination
"Belong to"
prinadlezat' (otdavat'sja)
"Fall under influence of' potakat' (poddavat'sja) (sdavat'sja) "Yield to/respect power of' ustupat' popustitel'stvovat' verit' doverjat' (doverjat'sja) vnimat' (divit'sja) (udivljat'sja)
(izumljat'sja) (radovat'sja) (klanjat'sja) (poklonjat'sja) radet' (obucat'sja) (u cit 'sja)
"indulge' , "give in to" "surrender; yield to"
"yield to" "tolerate, put up with" ,'believe' , "trust" "trust in; confide in" "attend to" "be surprised at, marvel at" "be surprised at" "be amazed at" "rejoice in" "humiliate self before" , 'worship' , "oblige" "learn" "study"
Nearly half of the verbs of subordination are reflexive and could be glossed as "give self to in an X relationship/in an X way" , meaning that they overlap with intransitive giving verbs. To their ranks could be added the verb porazat'sja "be struck by" which usually appears with a conduit instrumental, but dative govern-
"Resist" prepjatstvovat' (protivit'sja) (soprotivljat'sja)
"hinder, interfere" "oppose; resist" "resist; oppose"
"Fight back" protivoborstvovat' mstit' otplacivat'
"oppose; fight against" "avenge" "repay; requite"
Transition to dative of affectedness of schema 2 In terms of overall numbers, Russian verbs of benefit and harm, which occupy the transitional zone between schemas 1 and 2, are just as significant as their Czech counterparts. Their distribution, however, deviates sharply from that observed for Czech. Whereas in Czech this group is equally divided between benefit and harm, there are over fifty percent more Russian verbs of harm than of benefit. This appears not to be an isolated phenomenon for, as noted below, Russian instantiations of schema 2 and variants have a marked tendency to express harm and threats. ,'Benefit" blagoprijatstvovat' pokrovitel'stvovat' ugozdat' (godit'sja) (nravit'sja)
"favor" "patronize' , "please" "please, be convenient' , "please"
120
(prigljanut'sja) (poljubit'sja) idti
121
"become attractive to" "become attractive to" "become, suit"
ravnosil'nyj analogicnyj s ortogonalen identicen izomorfnyj
"be rude to" "be impertinent to" "injure, harm" "betray" "avenge, take revenge on" "threaten' , "annoy" "hinder, annoy" "get on nerves of" "bore" "bore, repel" "become repulsive to" "grow hateful to"
znakomyj s
"Harm" grubit'
derzit' vredit' izmenjat'
mstit' grozit' dosazdat' mesat' nadoedat' naskucit' ostocertet' oprotivet' opostylet'
"equivalent' , "analogous" ' 'orthogonal" "identical' , "isomorphic" "acquainted"
Subordination pokornyj poslusnyj predannyj podvlastnyj podotcetnyj s blagodamyj vernyj
Intransitive giving dostupnyj d izvestnyj d ponjatnyjd
"submissive, obedient" "obedient" "devoted" "subject to, dependent on" "accountable to" ,'grateful" "loyal"
pokorjat'sja
blagodarit' 127 verit'
"accessible' , "known, familiar" "understandable' ,
Government by adjectives Benefit The dative in Russian is also governed by adjectives, nouns and adverbs, and in every case, the semantic grouping of these lexical items follows the same pattern found for verbs. On the whole, the use of the dative with words other than verbs is more widespread in Russian than in Czech, evidenced both by the relatively large number of adjectives which participate in this kind of government and by the participation of nouns and adverbs, which instantiate schema la government rarely, if at all, in Czech. Superscripts are used to identify those adjectives that can alternatively be followed by prepositional phrases: superscript s marks adjectives that can also be followed by the comitative preposition s and the instrumental case; superscript d marks adjectives that can be followed by the preposition dlja "for" and the genitive case. 126 related verbs
Symmetrical podobnyj odnorodnyj ravnyj s proporcional'nyj s
"similar' , "similar' , "equal" "proportional' ,
upodobljat'sja ravnjat'sja
roo
scastliv dorogojd milyj d poleznyj d
"glad" "happy; fortunate" ' 'dear; expensive" "dear" "useful"
Harm protivnyj
nenavistnyj d vrednyjd
"repulsive' , "hateful" "harmful' ,
vredit'
It is interesting to note that it is the adjectives expressing symmetrical relationships which allow alternative expression with the comitative preposition (with the exception of podotcetnyj "accountable to", for which government disambiguates meaning - with the dative it describes a subordinating relationship; with the preposition s "with", it describes symmetrical, mutual accountability). Alternative expression with the preposition dlja is possible only for adjectives with transi-
122
123
tional as opposed to central (to schema la) meanings. The verbs that govern the dative by incorporating a direct object have no counterparts among the adjectives. There are in addition many adjectives that are conventionally used with the dative, but are closer in meaning to schema 2 (affectedness, although some of these could also be recognized as denoting benefit or harm), all of which can also be followed with dlja "for"; plus a smaller group of adjectives that specify possession (also schema 2).
Affectedness udobnyj d prijatnyj d obidnyj d obremenitel'nyj d neozidannyj d
dalekij d nuznyj d trudnyj d sympaticnyj d blitkij d cennyj d
"convenient' , ,'pleasant" "offensive" "burdensome' , "unexpected', "distant" "needed" " difficult" "nice" "near; familiar" ,'valuable"
Possession prisuscij
rodnoj cuzdyj cuzojd privycnyj d svojstvennyj d xaraktemyj d obscij d
,'inherent' ' "own; native" "alien" "someone else's; alien" "usual, customary" "related' , "characteristic' , "common"
Government by nouns In addition to collocations derived from verbal phrases, such as vino gostjam "wine for the guests" and pis'mo otcu "a letter for father", and nouns denoting an object designated for a beneficiary, as inpamjatnik Petru Pervomu "a monument to Peter 1" , nouns which govern the dative in Russian denote symmetrical
relation-ships, relationships of subordination, benefit, harm, and the objects incorporated in verbs transitional between schemas 1 and la. Symmetrical upodoblenie akkompanirovanie soboleznovanie soiuvstvie soprotivljaemost' protivlenie
protivopostavlenie protivorecie protivoves oppozicija al'ternativa dissonans kontrast podrazanie podrazatel' parallel' Subordination pomosc' sodejstvie poklonenie
predanie obucenie sledovanie povinovenie podcinenie poslusanie vernost' lojal'nost' sposobstvovanie sluzenie slutba prisluinicestvo predannost' predostavlenie
"likening; assimilation" "accompaniment (musical)" "sympathy, condolence" "sympathy' , "capacity to resist" "opposition' , "opposition, contrasting" "contradiction' , "counterbalance' , "opposition' , "alternative" "dissonance" "contrast' ' "imitation' , "imitator' , "parallel' ,
"help" "assistance, help" "worship' , "committing, handing over" "instruction, training" "following, compliance" "obedience" ,'submission' , "obedience" "loyalty" "loyalty' , "assistance' , "serving' , ,'service" "subservience' , "devotion' , "granting, concession"
related verbs upodobljat'sja akkompanirovat' soboleznovat' soiuvstvovat' soprotivljat'sja protivit'sja
protivopostavljat' protivorecit'
podrazat', podrazat'
pomogat' sodejstvovat' poklonjat'sja predavat' obucat'sja sledovat' povinovat'sja podcinjat'sja poslusat'sja 128
verit' sposobstvovat' sluzit' sluzit' slutit'
124
ustupka popustitel'stvo poslablenie potakanie Giving verbal signals prikazanie nakaz
porucenie napominanie zapros prigovor soobscenie rasskaz pis'mo raport privetstvie otvet sovet predlozenie Giving nonverbal signals rukopleskanie poklon cest' pocet slava svidetel'stvo Giving money or gifts kompensacija subsidija kredit podarok
zertva Benefit pokrovitel'stvo otrada
125
"concession' , "tolerance' , ' 'indulgence" "indulgence' ,
"order" "order; mandate" "instruction' , "reminder' , "inquiry" "sentence, verdict" "report, communication" "story" "letter" "report" "greeting" "answer" "advice" "proposition"
"applause' , "bow" "honor" "honor, esteem" "glory" "evidence, testimony"
"compensation' , "subsidy" "credit" "gift" "donation; sacrifice"
"partonage' , "delight"
ustupat' popustitel'stvovat' potakat'
pol'za
razdol'e
"use" "freedom' ,
prikazyvat'
Harm izmena mest' vred
napominat'
Government by adverbs
pisat'
There is a limited group of Russian adverbs which govern the dative case. Most of them are either analogous to the verbs denoting symmetrical relationships, or are transitional to schema 2 (describing generalized affectedness) in which case they can also be followed with the preposition dlja "for".
otvecat' sovetovat'
rukopleskat' klanjat'sja
kompensirovat'
iertvovat'
pokrovite1'stvovat'
"betrayal' , "vengeance, revenge" "harm"
Symmetrical vratrez s proporcional'no anologicno parallel'no sootvetstvenno podobno
"contrary to" "proportionally' , "analogously' , "in parallel to" "correspondingly' , "similarly' ,
Intransitive giving napererez
"across someone's path"
Affectedness nuzno d vazno d trudno d prijatno d
"necessary' , ' 'important" "difficult" ' 'pleasant' '
obidno d neobxodimo d Possession srodni
"offensive' , "necessary' ,
"related to"
izmenjat' mstit' vredit'
126
A few of the adverbs above are also listed as prepositions below, for in Russian the boundary between these two classes of words is fluid.
Government by prepositions Russian has not just one, but two primary prepositions that govern the dative. In addition to k "to", which it shares with Czech, Russian has the primary preposition po "along; according to" .129 Although k does not differ in any significant way from its Czech counterpart, the presence of po in Russian requires further explanation. Here, the independent status of the dative entity serves to indicate a path for the nominative to follow. 130 In essence, po acts much as the verbs of subordination, for in following the path indicated by the dative referent, the nominative referent subordinates itself to the dative referent.U! The following examples illustrate both concrete and metaphorical paths marked by po and the dative case. (146) Deti sli po tropinke. Children-NOM walked along path-DAT 'The children walked along the path.' (147) Sosedka vsegda odeta po Neighbor-NOM always dressed along poslednej mode. latest fashion-DAT '(Our) neighbor is always dressed according to the latest fashion.'
127
About a third of the secondary prepositions which govern the dative in Russian are built with k or po or both: naputi k po napravleniju k po otnoseniju k primenitel'no k gljadjapo smotrjapo sudjapo
"on the way to" "in the direction of' ,'in relation to" "in conformity with" "depending on" "in accordance with" "judging from"
The remaining secondary prepositions denote either symmetrical or subordinating relationships, with one exception, navstreiu "meeting, towards", which denotes intransitive giving. Symmetrical ne v primer v kontrast vopreki v protivoves naperekor v protivopoloznost' podobno soglasno s
soobrazno s sootvetstvenno sorazmemo
"unlike" "in contrast to" "in spite of; against" "as a counterbalance to" "in defiance of' "as opposed to" "like, similar to" "in accordance with" "in conformity with" "according to, in compliance with" "in proportion to"
There is yet another meaning of po, that of "apiece, each" as in: (148) Vospitatel'nica dala detjam po jabloku. Day-care-worker-NOM gave children-DAT along apple-DAT 'The day-care worker gave the children an apple apiece.' This meaning is, however, a variant of the meaning given above, further specified by context. It appears only in constructions with indirect objects, and in every case the indirect object is plural, whereas the noun phrase governed by po is singular. 132 It is the disharmony between the singular object being given and the plural recipients that suggests a distribution is necessary. Po is appropriate, for it implies that each object will indicate a single path (to one of the recipients).
Subordination blagodarja vsled
"thanks to, due to" "following after"
Intransitive giving navstrecu
"meeting, towards"
128
129
(154) Kak
4.1.2. Schema lb - impersonal dative
zit'? US-DAT to-live 'How are we to live?' . nam
How
The use of the dative in impersonal constructions, which is frequent in Czech only in category-of-state constructions, but fairly uncommon with reflexive verbs and altogether vestigial with infinitives, is a vigorousand productive phenomenonin Russian. These constructions figure prominently in the system of modal expressions in Russian, whereas Czech has developed modal verbs in their stead. Instantiationsof the category-of-state impersonal dative usually refer to prevailing conditions (expressed adverbially or with a prepositional phrase) which affect the dative entity, as in:
(155)
by poucit'sja u starikov. Young-DAT CONDIT'L to-leam-REFL at old-men-GEN (in Svedova et al. 1982) 'It's the young men who shouldlearn from the old men.'
Molodym
(156) Lucse by Better CONDIT'L
Mne
(150)
Nam
ne
do
pomolcat'. to-be-silent
(in Svedova 1982) 'It wouldbe betterfor him to keep quiet.'
xolodno/grustno. Me-DAT cold/sad 'I am cold/sad.'
(149)
emu
him-DAT
smexa. (in Svedova 1982)
The so-called "second dative" is arguably an elliptic example of this type of impersonal dative. Miller(1986: 310)correctly suggests that a sentence such as:
Us-DAT
not to laughter-GEN 'We don't feel like laughing.'
(157) My We-NOM
Examplesin which the adverb denotes manneror necessity are transitional between the category-of-state type and the infinitival type, for they combine elements of both types of construction, as in: (151)
(152)
trudno citat' Me-DAT hard to-read _'It's hard for me to read in German.' Mne
nado dumat' US-DAT necessary to-think 'We have to think aboutthe future.'
Nam
po-nemecki. in-German
o
buduscem.
about future-LOC
Even without qualifying adverbs, infinitival impersonal dative constructions are clearly modal, as in (153) and (154). Their hypothetical implications can be furtherenhanced by the conditional particle by , as in (155) and (156). (153) Cto
nam
us-DAT 'What are we to do?'
What-ACC
delat'? to-do
poprosili
Ivana
pojti
odnomu.,
asked
Ivan-ACC
to-go
one-DAT
'We asked Ivan to go alone.' "could be seen as a loose end left by the development of paratactic structures to hypotactic ones; that is by the development of two bits of syntax - My poprosili Ivana ["We asked Ivan"] and (emu) pojti odnomu ["(he) should go alone"] merely juxtaposed into a single piece of syntax". However, despite the fact that this example has appeared in numerous scholarly works, it may be a "freak", a stray example whichdoes not representa valid type of construction at all. Native speakers of Russiando not acceptthis example, nor does Svedova 1982 mention it at all. When realized with the infinitive of the verb byt' "be", the impersonal construction is closely related to the invocations whichinstantiate 2e, as in (158): (158) Byt'
lokomotivu ispravnym! (in Bachman 1980) To-be locomotive-DAT in-good-order-tnsr 'May the locomotive be in good order!'
With subjectlessreflexive verbs, the impersonal dative expresses either involuntary action or state (159) - (160), or a modal meaning roughly equivalent to English can/cannot, as in (161).
131
130
(159) Emu
(in Levine 1986) 'She stepped on his briefcase.'
ikaetsja (in Adamec 1978)
Him-DAT hiccups-REFL 'He has the hiccups.'
(160) Mne
xocetsja
Me-DAT wants-REFL 'I feel like going home.' (161)
pojti
The idiomatic phrase znat' im cenu "know their worth", that contrasts with the possessive zoot' ix cenu "know their price" also points to the fact that inalienable possession is essential in the use of the Russian dative of possession. The use of the dative specifies inherent-qualities, as in (165), but the possessive is a neutral modifier which refers instead to a superficial or incidental possession. Thus case serves to disambiguate the noun cena "worth; price".
domoj.
to-go home
Btojptice ne letitsja. That bird-DAT not flies-REFL 'That bird cannot fly.'
(165) Edigej
Edigej-NOM
vspomnilos'
detstvo. (in Bachman 1980)
Me-DAT remembered-REFL childhood-NOM 'Memories of childhood came to me.'
The use of the free dative is somewhat more restricted in Russian than in Czech. Schema 2 has only possessive uses, and the scope of possible possession is narrower here than was observed for Czech. Russian admits the use of dative marking only for inalienable possession (163), or with objects perceived of as metaphorically inalienably possessed (164).133
(167) On
Edigej-NOM
verbljudu
prignul
bent camel-DAT i ... osadil ego. (Ajtmatov) neck-ACC and saddled him-Ace 'Edigej contrived to bend the camel's neck and ... reined him in.'
She-NOM
*Rebenok
oom
Child-NOM
got-clever-REFL
seju,
(164) Ona
them-DAT ozidal, (Ajtmatov)
placet noc'ju. us-DAT cries night-INST 'The child cries at night (and we are affected by this).'
(166)
izlovcilsja,
on
In intransitive constructions (captioned by schema 2b) dative of beneficiary (166) - (167) and dative of possession (168) - (169) are rarely instantiated. Of the following examples, only (167) and (169) are acceptable utterances, although all of these sentences could be translated as examples of free datives in Czech.
4.2. Schema 2 and variants-free dative
(163) Edigej
Znal
Knew he-NOM
got-upset
cenu, no takogo oborota ne worth-Ace but such tum-GEN not expected 'Edigej became very upset He knew their worth (i.e., he knew what they were like), but he never expected such a tum of affairs. '
Some verbs will admit nominative subjects in such constructions, thus forming examples transitional between schemas la and lb, as in (162), which, however, lack the strong modal connotations of 1b, although they do express involuntary events.
(162) Mne
im
ocen' rasstroilsja.
very
nastupila
emu
na
stepped-on
him-DAT
on
portfel'. briefcase-Ace
mil. (in Isacenko 1965)134 me-DAT kind-NOM 'He is kind to me (i.e., I find him nice)'
mne
He-NOM (168)
"Rebenok nam Child-NOM US-DAT 'Our child got hurt.'
(169) Malys
Child-NOM
usibsja.
hurt-REFL
zalez
ej
na
koleni.
climbed-up
her-DAT
on
knees-Ace
(in Levine 1984) 'The child climbed up on her lap'
132
133
Mapped extensions of schema 2 in the form of ethical datives and emotional datives do appear in intransitive constructions, but their use in Russian (as opposed to Czech) is more constrained in terms both of frequency and of connotation. In Russian these datives rarely appear outside of the context of warning, threat or complaint, as in the ethical dative of (170), and in the emotional dative of (171).
(176); some have modal readings, functioning as elliptic versions of the impersonal dative of schema 1b (with determiners such as nuino "need", xocetsja "wants" -REFL removed) and thereby bridging the zone between schemas 2 and lb, as in (178); and some express beneficiaries without any special connotation, as in (179) - (180).
trudu!
(175) Slava (170) Vmesto
In-place-of vam
vraca
prislali
doctor-GEN sent nalecitl (in Svedova 1982) will-heal
studenta:
on
student-ACC
he-NOM
Glory-NOM labor-DAT 'Glory to labor!'
you-DAT 'Instead of a doctor they sent a student: you'll see what healing he can do!'
(176) "Net" gonke vooruzenijl (in Svedova 1982) No race-DAT arms-GEN "'No" to the arms race!'
mne You-NOM me-DAT (in Svedova et al. 1982)
(177) Posevnym rabotam- udarnye tempy. (in El'zbutas 1967) Sowing work-DAT shock-work tempo-NOM 'An accelerated tempo for the sowing.'
(171) Ty
ne
opozdaj
smotri!
not
be-late
look
'Watch that you aren't late on me!'
Amerike
(178) Zacem In copular constructions (captioned by schema 2c), both dative of beneficiary (172) and ethical dative (173) are instantiated.
tebe ne I-NOM you-DAT not 'I'm not your judge.'
(172) Ja
(173) Vot
tebe
sud'ja. (in Svedova 1982)
judge
"Amerika"? America-DAT America-NOM 'Why does America (need/want) "Amerika" (name of a film)?'
Why
(179) Da
Yes
prazdnik!
you-DAT and holiday-NOM 'That's some holiday for you!'
Here
i
cto
turistu
morozy,
esli
est'
and
what-NOM
tourist-DAT gostinic?
frosts-Nou
if
is
teplo
ujut
warm and
comfort-NOM hotels-GEN
'And what is freezing weather to the tourist if he has the warmth and comfort of hotels?' (180) On
A free dative can be used in a predicate instrumental construction (schema 2d) to mark a beneficiary, as in (174): (174) On
He-NOM
byl
was
im them-DAT
sovetcikom, counselor-INST
'He was a counselor for them.' The dative of beneficiary expressed in constructions with bare nominatives (schema 2e) has a range of uses: some are formulaic exhortations, as in (175) -
korolju. godfather-NOM king-DAT 'He's the king's godfather (i.e., someone who rules the king).' kum
He-NOM
There exists a variant of schema 2e, in which the noun phrase appears in conjunction with a preposition rather than as a bare nominative, also with a modal reading, as in (181):
135
134
(181) Kuda
Prage do Moskvy! Where-to Prague-DAT to Moscow-GEN 'Prague just can't compare with Moscow!'
(183) Oxota
The transitional schema which links schemas 2 and 1b of the Czech dative, schema 2f, is missing in Russian, although, as mentioned above, examples of 2e with modal readings fill this role. Schema 2g which captions the semanticrole of the dative with the single Czech verb zavide: "envy" has a precise equivalent in Russian in the verb zavidovat' "envy".
celoveku
pisat' -
pust'
sebe
Desire-NOM
man-DAT
to-write
let
REFL-DAT
piset.
Komu
kakoe delo? (Ajtmatov)
to-write Whom-DAT what matter-NOM 'If someone has the desire to write, let him write (for his own satisfaction). Whose businessis it?'
(184) A
tomu
And
then-DAT
muz
ee
husband-NOM her 4.3. Dative Reflexives
The combination of free dative and reflexivization (as in the Czech si of selfindulgence) is rather infrequent,135 perhaps because Russian has lost the short form dative reflexive pronoun. "Pure" reflexive indirect objects are rarely observed with the Russian dative reflexive, but there are a number of verbs which are to a greater or lesser degree synonymous with dat' "give", and thus instantiate examples transitional to schema refl2, such as kupit' sebe "buy oneself" , iskat' sebe "seek for oneself', and predstavit' sebe "imagine to oneself". The reflexive dative of self-indulgence appears with both transitive and intransitive verbs (Le., both schemas Refl2 and Refl2b), but is less popular in Russian than in Czech, and seems to require more context to support its use. (182) - (184) provide representative examples of Russian use of the dative reflexive. (182) Nervy,
doli no
Nerves-NOM must
eto
obyknovenno.
byt', to-be
podguljali
nemnozko,
no
acted-up
a-bit
but
Pripadok,
nado
dumat' ,
Attack-NOM
must sebe
to-think
that-NOM
ordinary-NOM
uze
koncilsja.
already
finished-REFL Lie-down-REFL
Lozites'
REFL-DAT
spat'. (Cexov) to-sleep 'It seems that your nerves were a bit shaken, but that's not unusual.
The attack is probablyover already. Lie down and go to sleep (for your own sake).'
i and sidel
dejstvitel'no
bylo
xot' by cto,
really sebe
was
whatever-NOM
opustosivsijsja
sat
REFL-DAT
devastated-NOM-REFL
i smurnoj, s grustnym, otresennym vidom, and gloomy-NOM with sad, estranged look-INST vse ze na poxorony testja all-NOM (emphatic) to funeral-ACC father-in-law-GEN priexal, i molca kuril sebe came and silently smoked REFL-DAT vonjueie, brosovyesigarety. (Ajtmatov) putrid, low-class cigarettes 'And ideed, no matter what might happen, her husband sat (according to his own desire) devastated and gloomy, with a sad, estranged look, after all, he had come to his father-in-law's funeral, and he silently smoked (for his own pleasure) putrid, low-class cigarettes.' In general, the dative reflexivesubnetworkof Russian is quite similar to that of Czech, albeit less frequently instantiated. The reciprocal dative, however, appears to be missing in Russian. In its place, Russian uses the dative version of drug druga "one another" (i.e., drug drugu "to one another"), an all-purpose phrase for expressing reciprocity which has the advantageof being modifiablefor all case and prepositional uses. Czech has an equivalent phrase, jeden druhemu, but it does not enjoy the high frequency of its Russian counterpart.
136
137 Czech
Intransitive taking
Reflexive Clauses
Russian
Figure 19. Comparison of Czech and Russian dative
Figure 19 presents an impressionistic comparison of the networks of the Czech and Russian dative. Portions of networks which enjoy stronger expression in one language than in the other are marked in boldface, and some parts of the networks have been enlarged to show variation in realization of subtypes of given schemas. In general terms, it appears that schema 2 is stronger in Czech, whereas schema 1's syntagmatic variants are stronger in Russian. Although schema 2 vies successfully with schema 1 for primacy in the Czech network, the Russian network continues to be centered about schema 1 and is apparently more conservative than its Czech counterpart Also, we see that metaphoric extension is far more prevalent in Czech than in Russian. Metaphoric extension via antonymy is absent in Russian, and although the full range of extension via mapping is present, its use is far more limited in Russian. I venture to speculate that geography has played an important role in the divergent development of these networks. In particular, the influence of German on Czech is quite evident. It is plausible that German influence has supported the expansion of the indirect object to constructions with taking verbs and for the development (by borrowing and loan translations) of modal verbs,136 which has caused schema lb to wither considerably. Expressions of dative of affectedness (schema 2 and variants) are well-represented in German (cf. Smith 1987), perhaps stimulating the growth of this part of the network. Isolated from such influence, Russian has made fewer non-Slavic innovations in its dative network.
5.
Analysis of the Russian instrumental Cto napisano perom ne vyrubis'toporom. 'What's written with a pen can't be cut down with an axe.' Russian proverb
°
5. .1. Overview of the network At first blush, the Russian instrumental appears to be an intractable tangle of meanings, accessible only to an encyclopedic account. When instrumental constructions are captioned by image-schemas as in figure 20, however, a simple pattern emerges. The instrumental has only four roles: it can serve as a conduit for the action of the verb ("conduit instrumental" - schema 1 and variants), it can become the setting for the action ("instrumental of setting" - schema 2 and variants), it can be juxtaposed with one of the non-oblique participants ("attributive instrumental" - schema 3 and variants), or it can be merely proximate to one of the participants ("comitative and proximate instrumental" - schema 4 and variants). Of these roles, the first one (schema 1) is the prototype and is used to express the largest number of meanings, making it the center of gravity as well. Schema 1 bears direct relationships to two of the remaining schemas. Schema 2 is achieved by delocalization of the instrumental to the background; and the process of shifting the focus of the instrumental from the event as a whole to one of the participants produces schema 3. A delocalization similar to that which produces 2 from 1 likewise produces 4 from 3.
5.0.2. Instrumental network paradigm Jakobson's feature specification of [+peripheral] for the instrumental embraces all of the roles outlined above. In no case is the instrumental a full participant in the event chain; it serves merely as an adjunct to the action or to one of the participants (schemas 1 and 3), or as an accompaniment for a participant (schema 4), or as the setting (schema 2). The instrumental is indeed the peripheral case par excellence, as Jakobson suggested, for it assumes every conceivable kind of nonessential function. The fact that all of these peripheral functions can be shown to cohere in a unified network supports Jakobson's thesis. 137 It also lays to rest
140
141 Network
Schemas Pictorial
Prose
Usage participant
() Inst
A nominative acts on an accusative via an instrumental in a setting.
Conduit Instrumental
axis of focus of reference
------;:usetting
A nominative acts on an accusative in a setting. The setting is an instrumental.
delocalization of instrumental
reference of instrumental shifted from event to participant
reference of instrumental shifted from event to participant
Instrumental of setting
event
A nominative seen as an instrumental acts on an accusative in a setting.
Attributive Instrumental
A nominative together with an instrumental acts on an accusative in a setting.
Comitative (and Proximate) Instrumental
delocalization of instrumental
internal ..c-----------t~ external axis of scope of reference
------;:usetting
Figure 20. Russian instrumental network
Figure 20. Russian instrumental network
claims that the feature [+peripheral] is actually meaningless and was invented by Jakobson as a default device to deal with what is in reality an incoherent jumble.138 Indeed the entire networkcan be conceived of as a paradigm of peripherality definedby two variables: 1. reference pointof peripherality, expressible as either the event itself or a participantthereof, and 2. relative integration of peripheral element,which may be conceived of as internal or external to the reference point. This disjunctive definition produces all of the instrumental's schemas:
143
142
This paradigm of peripherality is expressed graphically in the diagram of the network in Figure 20.
the instrumental of manner, and the instrumental of instrument, all of which are represented by schema 1 and its variants. The uses associated with schema 3 (predicate instrumental, comparison and quantitative uses) are also proximate in Mrazek's network. Although Wierzbicka makes different distinctions and gives them different labels, all of the "core" meanings in the network derived from her work (i.e., those which are most tightly bound: 1, 2, 3, 4, 5, 8) are likewise subsumed by schema 1.
5.0.3. Summary of the instrumental's semantic role
5.1. Conduit instrumental- schema 1 and syntagmatic variants
The trim lines of this network seem insufficient to account for the rather baroque display of meanings associated with the instrumental. The schemas are, however, designed to caption the semantic freight of the instrumental only, and the further contributions of other aspects of instrumental constructions (syntax and lexicon) play an essential role in distinguishing meanings. Users of Russian zero in on specific meanings by accessing knowledge of syntactic structure and of the noun class that the substantive marked with the instrumental belongs to. An appreciation of the role of syntactic structure and lexical choice in disambiguation of case meaning is not unique to cognitive grammar. Jakobson (1936 [1971: 48]) states that "everything other than peripheral status is given in individual uses of the instrumental by the actual meaning of its referent and by the context, but not by the case form."139 Thus if the structure of an expression suggests schema la, the meaning instantiated is further determined by the meaning of the lexical items: if the noun denotes an instrument, we have instrumental of instrument; if the noun denotes an agent, we have a passive; if the noun denotes a body part, we have an instrumental object; if the noun denotes the logical object of the verb, we have verbal government of the instrumental, etc. The role of the lexicon and encyclopedic knowledge in disambiguating case meaning will be indicated in the detailed description of the instrumental which follows. It should be noted that the network presented in Figure 20 bears some resemblance to the networks derived from Mrazek's and Wierzbicka's work (figures 6 and 7). The instrumental of instrument occupies a central role in all of the networks. Both Mrazek and Wierzbicka identify a larger number of submeanings than in the present work (because both of them attribute semantic effects resulting from lexical choice to the instrumental), but their groupings of meanings are similar. Thus space and time (represented by schema 2 and variants in the present network) are proximate in the network derived from Mrazek's work, as are the agent in impersonal constructions, the subject in passives, the instrumental object,
Schema 1 is central to the Russian instrumental. Associated with it are three syntagmatic variants, all of which involve a modification of the syntactic structure of the construction which the instrumental finds itself in, but do not entail any essential change in the instrumental's role. All the variants are produced by the same operation, that of removal of one or both of the direct participants in the event chain.
-
Internally peripheral to event: conduit instrumental, schema 1 Externally peripheral to event: instrumental of setting, schema 2 Internally peripheral to participant: attributive instrumental, schema 3 Externally peripheral to participant: proximate instrumental, schema 4.
Schema 1 - instrument, manner As mentioned above, the instrumental in this schema and in all of its local variants, serves merely as a conduit for the action of the verb, and does not have the status of a nominative, accusative or dative participant. The instrumental remains, therefore, relatively indifferent to the structure of the construction and retains the same relationship to the action even when participants are removed (in contrast to the dative, which, as we saw above, assumes different relationships with the removal of participants). Schema 1 captions the canonical action (of a nominative entity on an accusative entity) ,140 with the addition of the instrumental, which channels the event, but does not participate in it. This account of the instrumental marks a departure from that suggested by Langacker (1986), who distinguishes the instrumental from the accusative largely by their relative positions in the event chain (i.e., the accusative is the furthest "downstream"), although he does refer to dinstinctions in their case roles (Langacker 1987) labels the accusative an "energy sink", and the instrumental a "noncentral participant"). Passivization is viewed by Langacker simply as a reassignment of roles according to position, such that the downstream participant becomes the subject. The roles of the accusative and instrumental are, however, more
144
145 Schemas
Schemas Pictorial
Prose
Usage
A nominative acts on an accusative via an instrumental in a setting.
Instrument
A nominative acts via an instrumental in a setting.
Instrument, Passive Agent, Verbal Government
A nominative is a nominative' via an instrumental in a setting.
Pictorial
Prose
An action takes place via an instrumental in a setting.
~~~OT
Instrument
loss of acc
~
action of nom realized
~1 ' as identity Inexplicable force
-------.;s~etting
Figure 21. Conduit instrumental- schema 1 and syntagmatic variants
Impersonal Instrumental
Local Subnetwork
- - - - - - s....etting
An action on an accusative takes place via an instrumental in a setting.
Usage
~with predicate nom'
1 a
loss of acc
~
loss of nom
1
c
Figure 21. Conduit instrumental, continued distinct than Langacker's description suggests (the accusative entity is created or affected by the action whereas the instrumental entity is not). Also, Russian has a number of constructions (captioned by schema la) which contain only a moninative and an instrumental, therefore necessitating the use of some overt mechanism (here, the recognition that the instrumental is not a participant on a par with the accusative, symbolized by the double arrow that passes through the instrumental but not the accusative) to prevent the downstream participant from being identified as an accusative. And, as I shall argue below, a mere reordering of roles according to position is not sufficient to account for the complex phenomena of voice in Russian.
146
147
Instrumental of instrument
Abstract objects can serve as instruments in appropriate contexts as well, as in (190).
This use of the instrumental is largely self-explanatory: it is instantiated wherever the instrumental entity denotes an implement used in realizing the given action, as in (185) - (187). (185) Ivan rezal xleb Ivan-NOM cut bread-Ace 'Ivan cut bread with a knife.'
noiom. knife-INST
(186) Kolxozniki ubirali roz' Collective-farmers-NOM harvested rye-ACC kombajnom. combine-INST 'The collective farmers harvested rye with a combine.' skatert'ju. (187) Mama pokryla stol tablecloth-INST Mother-NOM covered table-ACC 'Mother covered the table with a tablecloth.'
Some authorsl'[! recognize such examples as illustrations of three separate submeanings - instrument, means, and ancillary material- but this finer distinction is drawn entirely by encyclopedic knowledge (which identifies referents as instruments, other kinds of objects, or matter) and cannot be attributed to the instrumental case. Although Staniseva (1958b: 78) supports this distinction, she admits that it can be hard to draw in some cases, and that one and the same word may invoke different submeanings in different contexts. She suggests that the instrumental case of platok "handkerchief" instantiates instrument in (188), but ancillary material in (189). proter ocki platkom. (188) On glasses-Ace handkerchief-INST He-NOM wiped 'He wiped his glasses with a handkerchief.' (189) On zakryl okno platkom. He-NOM covered window-Ace handkerchief-INST 'He covered the window with a handkerchief.'
(190) Monax dolzen svjazat' Monk-NOM must tie-up molcaniem. silence-INST 'A monk must seal his lips with silence.'
usta svoi lips own-ACC
This use of the instrumental case often entails the recognition of a part-whole relationship, in which the part serves as an instrument for the whole. In this example the subject, a pig, uses a part of his body, his snout, as an instrument. (191) Vyryla svin'ja jamu rylom Dug-out pig-NOM hole-ACC snout-INST legla v nee. lay-down in it-ACC 'The pig dug out a hole with its snout and lay down in it.'
i and
Extension via metonymy is quite widespread. Part-whole relationships are exploited in a wide variety of intransitive constructions captioned by schema la, as will be shown below. Examples which take advantage of this relationship also provide clear evidence of a transition between the instrumental of instrument and the instrumental object, as the comparison below indicates. The first group of phrases contains verbs followed by accusative direct objects. Here the instrumental is optional and when used identifies instruments that are parts of the subjects' bodies. The instrumental is an obligatory argument of the verbs in the second group, where it marks the logical object. The fact that verbs built from the same root (Um "squeeze" and xlop "slap") produce expressions in both groups further supports the argument that these two submeanings of the instrumental are related. Instrumental of instrument ubit' (kogo-to) to-kill (someone-Ace) 'kill (someone) with one's fist'
-------------------_•._
kulakom fist-INST
.
148
149
pal'cami nngers-nesr 'crush (something) with one'sfingers'
sial'
(etc-to)
to-crush
(something-Ace)
prixlopnut' (etc-to)
rukoj
to-slap
hand-INST
(something)-ACC
razrezanie nozom signalizacija flagami
"cutting with a knife" "signaling withflags"
stuk kulakom zagruzka rabotojl43
"knock with a fist' "burden of work"
'slap (something) with one's hand' Instrumental of manner Instrumental object brosat'(sja)
kamnjami to-throw-REFL stones-INST 'throw stones (at something)' 142
pozat'
(vo
etc-to)
(in
something-A CC)
plecami
(194) vzjat'
to-squeeze shoulders-ixsr 'shrug one's shoulders' xlopat'
city-ACC 'to take a city with spears.'
to-slam door-INST 'slam a door' In the extreme case, the part is identified with the whole, and the nominative entity is in effect wielding itself as an instrument. This explanation motivates the presence of the instrumental reflexive pronoun soboj in constructions like those in (192) and (193) and in sam soboj "by oneself" (self-NOM REFL-INST; lit: "oneself by oneself '). predstavljaet soboj iskljucenie. That-NOM represents REFL-INST exception-NOM 'That (in itself) constitutes an exception.'
(192) Eto
North-NOM
gorod
to-take
dver'ju
(193) Nort
Scholars frequently identify a further submeaning l44 which is found in constructions formally identical with those of the instrumental of instrument, but is distinguished from the latter in that it answers the question "how?" as in (194):
javljaet
soboj primer
is
REFL-INST
This distinction is further motivated by the fact that instrumental noun phrases with this meaning have a marked tendency to become adverbialized (witness words such as ukradkoj "stealthily" and kakim obrazom "in what way" , which are even listed as separate entries in dictionaries). It is clear, however, that the criterion of answering the question "how?" cannot serve as an operational definition of instrumental of manner vis avis other submeanings, for it is applicable to most examples of instrumental of instrument as well. What we have here is a group of constructions in which an abstract instrumental of instrument 145 acts not only as a conduit for the action, but also as a characterization of it, thus modifying the action. Most examples invoke a part-whole distinction between a certain aspect of the action and the action as a whole. In (195) stealth is an inherent part of the act of copying someone else's answers, but it is distinguished from the whole action and used as a label for it. (195) Ucenik
example-NOM
Pupil-NOM
patriotizma
tovarisca.
patriotism-GEN 'North is (in himself) an example of patriotism.'
comrade-GEN
Most verbal constructions which contain an instrumental of instrument retain this case usage when deverbalized, as these phrases attest:
kop'jami. spears-msr
ukradkoj
stealthily
spisal diktant u copied dictation-Ace at
'The pupil stealthily copied the dictation from his neighbor.' A similar case could be made for instantiations of other adverbialized "instrumentals of manner", such as:
150
begom
sagom polzkom verxom razom tajkom
151
"at a run" "at a walk" "at a crawl" "riding" "at once" "secretly',
b.
5.1.1. Schema 1a - instrumental of instrument and agent, verbal government The removal of the accusative entity from the event chain releases the predicate from satisfying the rather strict requirement of transitivity. The predicate is thus freed to take on a variety of forms: it may appear as a simple finite verb, as a reflexive verb, or as the verb "be" with a predicate adjective or participle. Consequently the predicate also expresses a variety of voices - active, passive and middle voice l46 - and its arguments can lay claim to a number of roles: agent, patient, object and instrument As will be shown below, both of these parameters represent a continuum rather than a series of distinct choices, and in each case distinctions can be made only by reference to the form of the verb and by analyzing information on noun classes or extra-linguistic knowledge. Xodova (1958a: 128 - 30) notes that there is no definitive line to be drawn between reflexive constructions which are passive (or middle voice) and those which are merely intransitive. Keenan (1985: 254) similarly observed that Russian -sja constructions can be identified as active or passive only on semantic grounds. Examples of this gradient will be seen below. Instrumental in passive constructions marks both logical subjects (agents) and instruments, a fact which further complicates the issue. In most cases, objects are identified as instruments and human beings as agents as in (196): Pis'mo napisano Letter-NOM written-PASTPASSPART 'The letter was written with a pen.'
Cexovym. CexOV-INST
The instrumental in (196a) clearly cannot be construed as an agent, but the instrumental in (197) can be so construed, even though it does not refer to a human being:
Not only is the definition of this submeaning vague and insufficient, but it makes reference exclusively to factors other than the meaning of the instrumental itself (i.e., we determine whether the noun phrase answers the question "how?" on the basis of context and extra-linguistic knowledge). It cannot therefore be recognized as an essential submeaning of the instrumental. In the present work it will be subsumed under the instrumental of instrument.
(196) a.
Pis'mo napisano Letter-NOM written-PASTPASSPART 'The letter was written by Cexov,'
perom. pen-INST
byli oskorbleny (197) V se Everyone-NOM was of!ended-PASTPASSPART etimi slovami. those words-INST 'Everyone was offended by those words.' Example (196) demonstrates that reference to noun class is necessary to distinguish between instruments and agents and (197) further shows that extralinguistic knowledge (understanding of the metonymic relationship between the speaker and the offensive words he has uttered) is essential in some instances.
Instrumental of instrument and instrumental of agent For the sake of organization, I will follow one parameter, that of morphological realization of the predicate, through the matrix of instrumental expressions outlined above, noting examples of both instrument and agent and areas of significant overlap or ambiguity.
Active constructions The simplest of these constructions are common instrumentals of the type covered in schema 1 with the accusative removed, as in: (198) Ja sejcas budu pisat' tol'ko etim karandasom. I-NOM now will to-write only this pencn-nssr 'Now I will write only with this pencil.' Instrumentals can also mark instruments used in actions that are obligatorily intransitive, such as motion:
153
152
(199) My priexali We-NOM arrived 'We arrived by train.'
poezdom. train-INST
the instrument of growing effected by the yard. Examples such as (205) and (206) bear a strong resemblance to passive constructions, both with reflexive verbs and with participles. 147
Extension via metonymy realized in the overlay of part-whole image-schemas is a recurrent theme, and may invoke parts of the nominative entity, as in (200) and (201), or reduplicate the whole of the action thus intensifying the expression (common in colloquial speech), as in (202): (200) On sidel spinoj k He-NOM sat back-INST to 'He sat with his back to the audience.'
zrlteljam. viewers-DAT
(201) On svetilsja He-NOM shone-REFL 'He shone with joy.'
radost'ju. joy-INST
(202) Propadi Get-lost 'Get lost!'
propadom! (Ajtmatov) getting-lost-INST
ty
YOU-NOM
(205) Kuxnja kisit muxami. Kitchen-NOM swarms flieS-INST 'The kitchen is swarming withflies.' (206) Dvor zaros travoj. Yard-NOM grew-over graSS-INST 'The yard became overgrown with grass.'
Reflexiveconstructions In constructions with reflexive -sja verbs, we observe the same tension between
However, there are many examples in which the identification of the instrumental as an instrument is not so secure. Compare the following two examples, which are both structurallyand semantically almost identical. Although the basses in the first of these examples are most likely to be recognized as instruments of the orchestra, they can also be construed as agents; the voices, however, are more properly agents of the verb and the nominative entity, the forest, is merely a location. (203) Orkestr gremit basami. Orchestra-NOM booms basses-nesr 'The orchestra is booming with basses.' (204) Les sumel golosami. Forest-NOM made-noise voiCeS-INST 'The forest was noisy with voices.' With many stative verbs the instrumental can only be recognized as an agent. Swarming is not something that the kitchen does by means of flies, nor is grass
active and passive and between instrumentalof instrument and of agent. The subjects of middle voice verbs such as ljubovat'sja "admire, be delighted with" and zanimat'sja "study, be occupied with" most resemble agents, and their instrumental arguments can be recognized as instruments. There are also many constructions in which the instrumental is non-obligatory, but retains a clear identification as an instrument, as in (207), which again demonstrates the use of a partwhole schema in distinguishing betweeninstrumental and nominative: (207) Ona usmexnulas' tol'ko uglami gub. She-NOM grinned-REFL only comers-INST lips-GEN 'She grinned with just the corners of her lips.' As seen above in the section on schema 1, the part-whole schema can be applied to the action as well, and this instrumental of instrument is commonly known as instrumental of manner (here the dance represents all or part of the verbal action of tossing about): (208) Teper' on i novye dva Now and he-NOM and new two-NOM meeutsja otcajannoj cecetkoj. (Majakovskij). toss-REFL desperate tap-dance-INST 'Now both he and the new pair toss about in a desperate tap-dance.'
154
155
However, as the voice of the verb shifts through middle to passive, the possibility for the instrumental to assume the role of agent increases. As a result there are many examples of minimal pairs (especially for clear passives) in which the class of the lexical item can decide the interpretation of the instrumental (209), as well as examples which are ambiguous (210) - (211): (209) a.
Sud'by
Rossii
Fates-NOM Russia-GEN salonnoj boltovnej.
b.
resalis'
ne
decided-REFL not
salon chatter-INST 'The fate of Russia has not been decided by salon chatter.' Sud'by Rossii resalis' Fates-NOM Russia-GEN velikimi vozdjami
decided-REFL
greatleaders-INST 'The fate of Russia has been decided by great leaders.' (210) Dver' otkryvaetsja voditelem. 148 Door-NOM opens-REFL driver-INST 'The door is opened by the driver. ' (211) Dusa smutilas' Soul-NOM troubled-REFL vozbudilas' gorjacej krov'ju stirred-up-REFL hot blood-INST 'His soul was troubled, his flesh was stirred
v
nem,
in
him-LOC
plot'. (Ajtmatov) flesh-NOM up by hot blood.'
At the other end of the spectrum there are reflexive constructions which are clearly passives and admit instrumental agents only: (212) Takie knigi citajutsja sirokimi massami. Such books-NOM read-REFL broad masses-INST 'Such books are read by the broad masses.' (213) Ljubye skandaly tnt obyeno smakujutsja Any scandals-NOM here usually gobble-up-REFL pressoj. preSS-INST 'Here all sorts of scandals are usually gobbled up by the press.'
Constructions with predicate adjectives and participles Copular constructions completed by adjectives or participles state a predicate of the subject (rather than the mere existential "be") in a manner often semantically indistinguishable from that stated by finite verbs. Thus the choice of constructions is somewhat arbitrary (Russian, for example, sometimes uses finite verbs where English uses predicate adjectives, as in belet' "be/show white" and gordit'sja "be proud", and the reverse is also true, as in poxoz "resemble"). The adjectival and participial predicate constructions express a verbal function of the subject, which, like other verbal constructions, displays a range of voice phenomena, and, toward the passive end of this spectrum, the instrumental may mark agents as well as in-struments. Constructions with participles which serve the purposes of subordina-tion differ from those which appear in predicates only in that their clause has been reduced (because the subject already appears in the main clause) and will not be treated separately. Constructions with predicate adjectives and instrumentals frequently make use of the part-whole image schema (and in the case of 216, recognize the whole as a part) in order to set the instrument apart from the nominative entity.149 (214) Ona byla bledna She-NOM was pale-NOM 'She was pale in the/ace.'
licom.
face-INST
(215) Petr simpaticen svoej iskrennost'ju. Peter-NOM likeable-NOM own sincerity-INST 'Peter is likeable/or his sincerity.' (216) lena Edil'baja soboj ladnaja, Wife-NOM Edil'baj-GEN and REFL-INST harmonious-NOM i domu xozjajka... (Ajtmatov) and house-DAT lady-NOM 'Edil'baj's wife is both good-looking, and the lady of the house ...'
Since predicate adjectives describe attributes, entities which can clearly be recognized as instruments are usually parts of the subject. When instantiations deviate from this pattern it becomes possible 150 to recognize instrumental entities as effecting actions (as agents), as in (217):
156
157
(217) Tjanulas' zizn' leak u vsex, kto Drew-REFL life-NOM how at everyone-GEN who-NOM zivet, bogataja dlinnymi gorestjami i lives rich-NOM long sorrows-INST and bednaja korotkimi radostjami. (Tolstoj; in Gadolina 1958a) poor-NOM short joys-INST 'Life dragged on - as it is with everyone who lives - rich in lingering sorrows and poor in briefjoys.' Here sorrows and joys can be considered either an integral part of life or external agentswhich influence its quality. A range of interpretations of voice is also possible, for life can either be construed to express richness and poorness (in which case it is an agent) or can be said to have been made rich and poor by sorrows and joys (in whichcase it is a patient). With comparative adjectives the instrumental isolates the part which provokes the comparison (Le., the part by which one object is perceived as superior or inferiorto another object): (218) Losad' soseda lucse moej i Horse-NOM neighbor-GEN better mine-GEN and i krasotoj. siloj strength-INST and beauty-INST 'The neighbor's horse is better than mine both in strength and in appearance.'151
The conjunctions cern and tern (lit: "by what means, by that means") which are used to construct many comparative expressions are, of course, historical instrumentals. The so-called instrumental of measure is a special case of the exploitation of a part-whole relationship in instrumental of instrument constructions. Here a dimension of the nominative entity is singled out as an abstract instrument by means of which the wholeis characterized. This use of the instrumental is particularly widespread in technical literature. It frequently appears in reducedclauses analogous to participialuses. The units of measurement can appear in a bare accusative or in a prepositional phrase. (219) xolodil'nik refrigerator-NOM
emkost'ju dvenadcat' volume-INST twelve-NOM
tysjac tonn thousands tons-GEN 'a refrigerator with a twelve-ton capacity' (220) obuv' cenoj ot vos'mi footwear-NOM price-INST from eight-GEN dvenadcati rublej twelve rubles-GEN 'footwearfrom eight to twelve rubles in price' (221) progulki prodolzitel'nost'ju walks-NOM duration-INST 'walks an hour in length'
v
cas
in
hour-Ace
do to
This part-whole distinction is extended to the whole in adverbialized instrumentals such as celikom "altogether" and polnost'ju "entirely". As examples (196) - (197) demonstrate above, with passiveparticiples both instrumental of instrument and of agent can be instantiated. Often even lexical information is not sufficient to disambiguate them, for their interpretation is dependent on context or on the construal preferred by the speaker. Thus (222) - (224) have a number of possible interpretations requiringthe assignment of various sets of roles to the participants (222) Oni ob"edineny obscim gorem. common grief-INST They-NOM united-PASTPASSPART 'They are united by shared grief.' a. They (AGENT) have united themselves by shared grief (INSTRUMENT). Something or someone (AGENT) has united them (PATIENT) by b. sharedgrief (INSTRUMENT). They (PATIENT) are united by shared grief (AGENT). c. (223) Na skam'e sidit zenscina, okutannaja On bench-toe sits woman-NOM wrapped-PASTPASSPART cemoj sal'ju. (Staniseva 1958b) black shawl-INST 'On the benchsits a woman wrapped in a black shawl.' a. The woman (AGENT) has wrapped herself with the shawl (INSTRUMENT).
158
159
b. c.
Something or someone (AGENT) has wrapped the woman (PATIENT) with the shawl (INSlRUMENT). The woman (PATIENT) has been wrapped by the shawl (AGENT). (This interpretation is probably accessible only in the context of literary fiction.)
(224) Stroitel'stvo
oplaceno
bankom.
paid-PASTPASSPART bank-INST 'The construction was paid for by the bank.' a. The construction (PATIENT) was paid for by the bank (AGENT). b. Someone (AGENT) paid for the construction (PATIENT) by using the bank (INSlRUMENT). Construction-NOM
Examples of this sort show both the versatility and the limitations of the instrumental: its narrow definition as a conduit allows it to be assigned various roles. The ambiguity arising from the use of instrumentals is commonly exploited by writers and politicians.
count for the many examples presented in this text for which the interpretation of voice is ambiguous andlor context-sensitive. The accommodation of ambiguity in the roles of agent vs. instrument is crucial to a description both of these examples and of the instrumental of inexplicable force (cf. section 5.1.2.), where ambiguity produces a unique effect. - Passive constructions bear a strong structural resemblance to other constructions and should therefore be integrated into a comprehensive description of them. This larger group of constructions further participates in a significant paradigm of instrumental constructions, in which one after another of the participants of the event chain is removed (schema 1 and its local variants). If passives were isolated from this paradigm (i.e., described as transformations of active expressions rather than as full-fledged instantiations of the instrumental), both the unity of the paradigm and the role of passives within it would be lost The foregoing argument has established that the integrity of the instrumental case in the instrumental of both instrument and agent is best seen as a single category: the conduit instrumental. These two manifestations of that instrumental can be disambiguated only by context, thus there is no support for a distinction within the semantics of the instrumental case.
Role of passive constructions The reader might at this point object to the treatment of passives contained herein. In this model, passives do not occupy a separate space of their own, nor are they considered simple transformations of active expressions, but are rather integrated into the overall account of the instrumental. The rationale for this is as follows: - Passive constructions in fact represent a diffuse rather than distinct phenomenon. As argued above, the route from active to passive is a smooth transition rather than a sudden leap. - A passive construction entails a statement about the nominative entity, and is therefore not merely a rearrangement of an active expression. It is a statement of a qualitatively different kind, although of course it bears a special relationship to the corresponding active construction. - If passive constructions were simply derived from active ones via transformations, then we would expect a straightforward mapping of roles from the active construction. The observed ambiguity of roles in passive constructions is problematic both to a transformational model and to Langacker's treatment of passives as active expressions with reassigned roles. Since such models require an all-or-nothing interpretation of passivization, they are inadequate to ac-
Instrumental object There are two more groups of constructions which also belong to schema la as instantiations of the conduit instrumental: the instrumental object and the instrumental governed by a verb. The relationship between the instrumental object and the instrumental of instrument has already been mentioned above. It is commonly agreed 152 that the instrumental object is a variant of the instrumental of instrument. When the object is marked as an instrumental, it is not the goal of the action, but serves only to manifest it. 153 This distinction is clear in examples which contrast the use of accusative and instrumental objects with the same verb: (225) Dvigatel'
krutit kolesa. turns wheels-ACC 'The engine turns the wheels.'
Engine-NOM
(226) Krutit' To-tum
rulem -
delo
ne
steering-wheel-INST
affair-NOM
not
161
160
sloznoe. complicated-NOM 'Turning a steering wheel isn't a hard thing to do.' The wheels in (225) are the goal of the action described therein, but in (226) the steering wheel is an instrument used by the driver for turning. In most cases, however, such near-minimal pairs cannot be constructed. The instrument is frequently a part of the body and is often redundant as well, as in (228):
Verbs governing the instrumental "Govern" vedat' verxovodit' zavedovat' zloupotrebljat' * komandovat'
dirizirovat' (227)
Passazir
maxal platkom/rukoj. waved handkerchieflhand-INST 'The passenger waved a handkerchief/his hand. '
Passenger-NOM
(228)
V otvet inzener tol'ko kivnul In answer-Ace engineer-NOM only nodded golovoj. head-INST 'In response the engineer only nodded his head.'
The distinction between accusative and instrumental objects can be restated as a claim about transitivity: accusative constructions are transitive, whereas instrumental constructions are intransitive. This claim is supported by evidence that inherently intransitive verbs can take instrumental objects, as in the chess term: (229) idti ferzem to-walk queen-INST 'to move (one's) queen'
Verbal government The second group of constructions, those in which the instrumental is governed by the verb, is in turn closely related to instrumental object constructions. 154 It is distinguished from the latter by two features: 1. it involves a closed class of verbs which share the meaning "dominate", and 2. it is always an obligatory argument. The following is a list155 of the members of this class..
pravit' predvodite I 'stvovat' rukovodit' upravljat'
"Possess' , vladet' vorocat'
obladat' ovladevat' * raspolagat'
"manage" "lord it over" "superintend' , "misuse" "command" "conduct (a musical group)" "govern" "lead" "lead, direct" "govern, administer"
,'possess' , "have control of' ,'possess' , "take possession of; master" "have at one's disposal"
"Evaluate" prenebregat' * "despise" dorozu' "value" *paired with corresponding perfective In addition to the shared semantic content of "dominate", these verbs display interesting phenomena. As a group they most resemble statives,156 which explains why only an exceptional few (starred above) have perfective counterparts. The meaning of these verbs can be generalized by the paraphrase "be the leader/ owner/evaluator", and the complement is then the people or objects 157 which make the role of leader/owner/evaluator realizable. Without soldiers a general is not a commander, and without possessions a tycoon is not an owner. Thus the governed instrumental can be viewed as yet another instance of the conduit instrumental, for it too marks the entity by means of which the verb is expressed by the subject.
162
163
5.1.2. Schema lb - instrumental of inexplicable force
vejat' "waft") and noun phrases in expressions of sensations such as smell (233)
and warmth (234). Schema 1b is arrived at by removing the nominative participantfrom the syntagm in schema 1. The resulting construction contains a verb (with neuter singular marking, but nothing to agree with), an accusative patient, and an instrumental. Here the ambiguity inherent in the instrumental between instrument and agent produces a heightened effect. Although the instrumental entity is readily recognizable as the agent, the fact that the subject position remains empty carries the implication that there is some unnamed entity (fate, nature, God) which has the power to wield the agent as an instrument. Such constructions are most often found in descriptions of disasters; typical examples appear in (230) - (232): ubilo molniej. killed lightning-INST 'He was killed by lightning.'
(233) So
tjanulo
stepi
edva
From steppe-GEN wafted barely ulovimoj polynnoj prel'ju. (Ajtmatov)
perceptiblewormwood rot-INST 'From the steppe wafted the barely perceptible musty smell of wormwood.'
(234) Ot
reki potjanulo proxladoj. From river-GEN wafted chill-INST 'A chill began to come off the lake.'
(230) Ego
Him-ACC
Since it lacks both a nominative and an accusative, this construction is even more under-determined than those captioned by schemas la and lb and is therefore multiply ambiguous. Example (233) can have any of the following interpretations:
(231) Mal'cika
elektrickoj. zadavilo ran-over commuter-train-INST 'The boy was run over by a commuter train.'
Boy-A CC
(232) Ego Him-ACC
obdalo
zapaxom
poured-over
scent-INST
i and
vlagoj
moisture-INST
a. b. c.
Something (AGENT) wafted using wormwood rot (INSlRUMENT). Wafting took place by means of wormwoodrot (AGENT). Something (AGENT) like wormwoodrot (COMPARISON) wafted.
cvetov. (in Xodova 1958a) flowers-GEN
'He was seized by the scent and the moisture of the flowers.' The parallels between this construction and the passive construction (both entail focus on the patient and allow for ambiguity between instrumental of instrument and of agent) 158 support their inclusion in the same paradigm (that of the "disappearing participants" in the local variantsof schema 1).
The first two possibilities are the same as those encountered with the instrumental of inexplicable force; the third is new and gives evidence of a transition between schema 1 and schema 3 (which captions instrumental of comparison). The verb paxnut' "smell" can form personal constructions which, because they are less ambiguous, clarify this transition. (235) Dvor
paxnet
senom.
Yard-NOM
smells 'The yard smells of hay. '
hay-INST
5.1.3. Schema 1c - instrumental of sensation (236) Babuska
vsegda
paxnula
pirogami.
always
smelled
pies-INST
In schema lc both the nominative and the accusative participants have been re-
Grandmother-NOM
moved, leaving behind only a subjectless verb (which has neuter singular agreement) and an instrumental noun phrase. This impersonal construction employs a limited number of verbs tpaxnut' "smell", vonjat' "stink", tjanut' "waft",
'Grandmother always smelled ofpies.' In the first example, the most likely reading is that there is hay in the yard which is giving off the smell. The hay is therefore the instrument of the yard
__________________IIIIIIIIIL__
11111
165
164
(which plays the role of agent) in producing the smell (it could be that there is something else in the yard which smells like hay, in which case we would have a comparison, but this is much less likely). The likelihood of reading the second example as an instrumental of instrument is much less. This sentence will usually be interpeted as one of comparison, meaning that grandmother smelled like pies (although the case could be made that she is actually using pies as an instrument to produce the smell). These two examples are representative of the continuum that exists between instrument and comparison and of the factors which influence interpretation along this continuum. Our knowledge of the fact that hay can be found in yards enables us to recognize the hay as an instrument of smell in the first example, and likewise we know that grandmothers usually do not carry pies with them at all times, and we can thus recognize the instrumental in the second example as evoking a comparison. It is the presence of the nominative entity in these examples which makes it possible for us to determine whether it is actually using the instrumental or being compared with it. This ambiguity is more difficult to resolve when no nominative entity is named.
5.1.4. Subnetwork of schema 1 and the paradigm of disappearing participants The subnetwork of schema 1 and its local variants unifies a number of submeanings of the instrumental (commonly identified as instrument, agent, manner, delimitation, inexplicable force, sensation) in a single paradigm of constructions. This is a paradigm of constructions related to the prototype by relative completeness of the syntagm of arguments (nominative, instrumental, accusative; where nominative and accusative are variable and instrumental is constant). All logical possibilities are represented in the paradigm, and transitions are observed both between members of the paradigm and between this paradigm and others in the instrumental network.
agent), variations in syntactic structure will not be considered separately here; all variants will be symbolized by the canonical event (which is composed of a nominative, a verb and an accusative). The nature (i.e., spatial or temporal) and shape of the setting are determined by the semantics of the noun phrase marked as instrumental. The setting may be very general (without a focused shape), or it may take the shape of a path or an opening. In addition, the setting may be singular or reduplicated (if the noun phrase is plural), and reduplicated settings may be discrete or contiguous. Details of all of these variants will be outlined below. The use of the same schema to accommodate expressions of both space and time is well-motivated. In her survey of the function of the brain, Churchland (1986: 200) writes that "it is surely implausible to suppose that the brain deals with space and time separately; from what we do know of its operations, there is spatiotemporal integration". Talmy (1986: 9) comments on the linguistic realization of this phenomenon; he finds "an extensive homology between the representation of space and that of time". Russian itself provides examples independent of the instrumental case in which space and time are treated equivalently.D? Uses of the Russian instrumental of space and time are tradi-tionally held. to be very closely related. 160 Space is generally percieved to be the more basic concept, and the use of schema 2 with noun phrases dentoing time constitutes semantic extension via synonymy of the spatial instrumental of setting, motivated by analogy of space and time.
Pictorial
Prose
A nominative acts on an accusative in a setting. The setting is an instrumental.
5.2. Schema 2 - instrumental ofsetting ------....settinglinst
In schema 2 the instrumental found in schema 1 has been generalized to the point
at which it serves no longer as a conduit for the action, but rather as a spatial or temporal backdrop. It has in effect merged with the setting. Because the instrumental in this schema relates to the event chain as a whole, and because the composition of the event chain does not serve to disambiguate the meaning of the instrumental (as was sometimes the case with the instrumental of instrument and
Figure 22. Schema 2
Usage
Setting (space, time)
166
5.2.1. Instrumental of space In its most typical uses the instrumental of space marks the physical setting for an action described by a verb of motion. prostornoj, svetloj komnatoj. (237) Oni prosli spacious, bright room-INST They-NOM walked-through (in Bulygina 1958a) 'They walked througha spacious, bright room.' storonu i v (238) On rezko svernul side-ACC to and He-NOM sharply turned posel suxodolom naiskos'. (Ajtmatov) went riverbed-INST diagonally 'He turned sharplyto the side and walked diagonally across the riverbed.' Adnominal uses of this instrumental appear with nouns derived from motion verbs, such as ezda polem "a ride through thefield" (in Leska Ms). If the noun phrasespecifiesa path, the setting takes the appropriate shape: kraem (239) Pegij pes, beguscij edge-INST Skewbald dog-NOM running-PRESACTPART morja (Ajtmatov) sea-GEN 'A skewbald dog, runningalong the edge of the sea' (240) Sestra bredet trexverstnoj Presneju. (Majakovskij) Sister-NOM wanders three-verst long Presnja Street-INST 'Sister wandersdown three-verst long Presnja Street.' Since a path can be metaphorically conceivedof as a method (cf. English way) for accomplishing a goal, we see some overlap between this instrumental and the conduit instrumental (especially in its meanings of instrumentand manner). Thus the path becomesa meansof reaching a goal:
prosveseenie. (241) Vot kakimiputjami rasprostranjaetsja education-NOM Here what ways-INST spreads-REFL (Lermontov; in Bulygina 1958a) 'Here are the ways in which education is propagated.'
167
Some path-shaped instrumentals of space have becomefixed expressions, such as the adverbs krugom "around; entirely (lit: in a round path)", sledom "right behind; immediately (lit: in a path following)" (note the transition to time), and putem "properly (lit: in a path)", the last of which is also used as a preposition meaning "by means of'. In a usage whichis alreadyrare in Russianthe noun phrase denotes an opening through which something passes. This providesa clear transition between schema 2 and the instrumental of instrument of schema 1a, for here the instrumentalentity acts as a conduitfor the action: (242) Krov' sla nosom i Blood-NOM went nose-INST and (in Bulygina 1958a) 'Blood went through the nose and throat. '
gorlom. throat-INST
If space is depicted by a plural count noun, the setting again adapts itself accordingly,this time by means of reduplication:
(243) V lesu mestami esce lezal In forest-LOC plaCeS-INST still lay (in Wierzbicka 1980) 'In the forest there was still snow lying in places.'
sneg. snow-NOM
5.2.2. Instrumental of time Since time is usually conceived of as being linear, the instrumental of time frequently instantiatespath-shaped schemas.There is some overlap between time and space, as shown in the following two examples. The first describes a journey, which is a path through both time and space, and the second is an idiomatic expression in which a certain period of time (that of bearing young) is conceived of as a space through which the animalwalks. (244) Dorogoj on Road-INST he-NOM 'While traveling he read.' (245) Ona She-NOM
cital, (in Malaxovskaja 1958)
read
verbljudica molodaja, udojnaja, camel young, good-milker-NOM
169
168
'With these thoughts Edigej walked away in that deep night, having followed the spacerocketright up untilit completely disappeared.'
idet. (Ajtmatov)
vtorym okotom
second young-bearing-season-INST walks 'She is a young camel, a good milker, in her second season of bearing (248)
young.'
Fierce winter-INST v Prage.
Like the instrumental of space, that of time shows evidence of transitional examples linking it to the conduitinstrumental of instrumentand manner. The construction in (246) can be construed as instantiating eitherschema la or schema2: (246) Celovek Man-NOM
bez
pamjati.
proslogo ...
without
memory-GEN past-GEN
sposoben capable-NOM
zit'
tol'ko segodnjasnim dnem. (Ajtmatov) to-live only today's day-INST 'A man with no memory of the past ... is capable of living only in/for/by means ofthe present day.'
The most familiar uses of the instrumental of time are examplescontaining the following words, frequently referred to as adverbs: utrom dnem vecerom noc'ju zimoj
vesnoj letom osen'ju
Prague-toe
'We happened to be in Prague during the fierce winter of 1987.' Second, as Kucera(1966) pointsout, this use of the instrumental of time is not restricted to a closed class of nouns denoting parts of the day and year, as has been assumed. Examples (249)and (250)illustrate the productivity of the instrumentalof time withothernounsdenoting fixed periods: (249)
zastavili Velikim postom menja forced Great Lent-INST me-ACC (Gor'kij; in Kucera 1966) 'During Great Lent they mademe fast.'
govet'. to-fast
«
mother-in-law-NOM died. 'During that March of snowstorms and shivering cold mother-
in-law died.'
The classification of these words as adverbs is not unproblematic. First, all of thesewords admitadjectival modifiers: uxodil
okazalis' happened-REFL
That March- snow-stormy, freezing-cold-INST tesca pomerla. (Vasil'ev: in Russkaja grammatika 1982)
Like the instrumental of space, that of time can appear in adnominal position:
takimi mysljami
in
1987 my 1987 we-NOM
(250) Tem martom - metel'nym, oznobistym
"in the morning" "in the afternoon" "in the evening" "in the night" "in the winter" "in the spring" "in the summer" "in the fall"
progulka vecerom "a walk in the evening'" (in Leska MS).
(247) S
Ljutoj zimoj
Edigej
With such thoughts-INST walked-away Edigej-NOM kosmiceskuju raketu toj glubokoj noc'ju, provodiv that deep night-INST having-followed space rocket-Ace
do
samogo polnogo ee isceznovenija. (Ajtmatov)
to
most complete its disappearance-GEN
Perhaps the strongest argument for considering these forms instrumentals rather than dispensing with them by classing them as adverbs is paradigmatic. The words listed above have an obvious relationship to the paradigm of the instrumental of setting, and their artificial removal from that paradigm would both deny the relationship that theybear to othermembers of the paradigm and leave a significantgap in it. Because the nomenclature for temporalphenomena usually specifies whether periodsof time are serialor contiguous (whereas the spatiallexiconis more general), the instrumental of time showsmorelexicallydifferentiated subtypesthan the instrumental of space. With pluralnounsdenoting noncontiguous time periodsthe setting is reduplicated:
171
170
(251)
Yecerami ja caste brodil odin Evenings-INST I-NOM often wandered alone-NOM po gorodu. (Tendrjakov; in Kucera 1966) along city-DAT 'In the evenings I often wandered about the city alone.'
Note that the adverbs dvazdy, trizdy "twice, three times" are historical instrumentals which developed from this part of the network. They joined the class of adverbs when changes in nominal declension rendered them unrecognizable as instrumentals. If the periods of time described in the noun phrase are necessarily contiguous, duration is emphatically expressed: (252)
Eto It-NOM
celovecestvu
tradicionnoe privetstvie, traditional greeting-NOM vekami.
znakomoe known-NOM
mankind-DAT ccnumes-rssr 'It is a traditional greeting, which has been known to mankind for centuries.'
it is juxtaposed with one of the event chain participants. It thus becomes an attribute of that participant, while maintaining perceptible ties to the conduit instrumental, for the attributive instrumental describes the way in which the participant it modifies is perceived. Since the instrumental is dependent on a single participant, its interpretation is largely indifferent to the structure of the phrase it is in, and again the canonical event chain represents all syntactic variants. The actual range of this instrumental has distinct limits. It can be juxtaposed with: the nominative of an intransitive clause (3a), the nominative subject of a copular verb (3a'), a bare nominative (3b), the accusative of a transitive clause (3c), or the dative of an impersonal clause (3d). It appears that the attributive instrumental is only.very rarely juxtaposed with the nominative of a transitive clause. This is probably because the participant which has the juxtaposed instrumental must be the focus 16 1 of the clause. All of the participants listed above are indeed focal elements in their respective clauses. In a transitive clause focus is on the object which has undergone some process rather than on the subject which has not. Thus we can say (253a) and (254a), but not (253b) and (254b), since that would require the clause to have two foci: (252)
a.
Anja poet solov'em. Anja-NOM sings nightingale-INST 'Anja sings like a nightingale.'
b.
*Anja spela etu pesenku Anja-NOM sang that song-ACC 'Anja sang that song like a nightingale.'
a.
Vanja smotrit volkom. wolf-INST Vanja-NOM looks 'Vanja looks like a wolf.' *Vanja est makarony Vanja-NOM eats macaroni-Ace 'Vanja eats macaroni like a wolf.'
5.2.3. Summary of the instrumental of setting To sum up:
1.
2.
The spatial instrumental of setting is extended via metonymy to the temporal domain. Spatial and temporal instrumentals of setting are distinguished by the meaning of the case-marked noun phrases and by the greater context in which they occur. Variants are conditioned by the grammatical number of the instrumental noun phrase, and by recognition of these entities as discrete or contiguous.
5.3. Attributive instrumental The schematic extension which produces schema 3 from schema 1 is the opposite of that which produces schema 2. Rather than being generalized to the point that it merges with the setting, the instrumental in schema 3 is localized to the point that
(253)
b.
solov'em. nightingale-INST
volkom. wolf-INST
The fact that the subject of an intransitive clause patterns with the object of a transitive clause rather than with the subject of a transitive clause gives evidence of absolutive-ergative distinctions in what is predominantly a nominative-accusative language. 162
172
173 Schemas Pictorial
Prose
Usage
3. A nominative seen as an instrumental acts on an accusative in a setting.
Attributive Instrumental
Example (255)is the exception that proves the rule. Note that even though the attributive instrumental is here juxtaposed with the subject of a transitive verb, focus has been placed on the subject by the markedOVS word order in this utterance and by the use of a transitive phrase profit' slezy "shed tears" that is synony-mous with the morefrequent intransitive verb plakat' 'cry": pomnis',
skol'ko
slez
rebenkom
You-NOM
remember
how-many
tears-GEN
child-INST
prolil
ja? (in Potebnja [1958])
poured-out
I-NOM
(255) Ty
3a.
3a'.
A nominative seen as an instrumental acts in a setting.
A nominative seen as an instrumental is a nominative' in a setting.
3b. A nominative seen as an instrumental in a setting.
3c.
A nominative acts on an accusative seen as an instrumental in a setting.
'Do youremember howmanytearsI poured out as a child?' The submeaning instantiated by an attributive instrumental is determined by its relationshipto the participant with which it is juxtaposed, and this relationshipis in tum a function of the semantics of the two noun phrases. Thus such examples provide no evidence for variation on the part of the instrumental case itself. Since both noun phrases havethe samereferent, they may differin their characterization of that referent both quantitatively and qualitatively. The instrumental may therefore use different units of measurement in referring to the participantit is juxtaposed with (a phenomenon also known as tvoritel'nyj sovokupnosti [instrumental of group or formation]), or it may present a characteristic which is not evident in the noun phrase with which it is juxtaposed (also known as predicate instrumental). Yet a thirdpossibility exists: that the instrumental may describe an entirely different referent, thusinvoking a comparison between that referentand the one marked by the participant with which it is juxtaposed. All three of these possibilities are well-represented in Russian, and there is significantoverlap among these categories. 5.3.1. Attributive instrumental - quantitative
3d.
An action takes place on a dative seen as an instrumental in a setting. The dative retains independent status.
The quantitative attributive instrumental can be juxtaposed only with subjects of intransitive verbs and with objects of transitive verbs. The noun phrases which represent these participants denote pluralcount nouns, and the instrumental noun phrase denotesa group whichgathers the individuals into a single mass (256) and (257), or a series of identicalgroupings (258) - (260). (256) Oni
Figure 23. Schema 3 and syntactic variants
They-NOM
xoteli
byt'
vse
vmeste
pod
wanted
to-be
all-NOM
together
under
174
175
dozdern, s det'mi, vsej sem'ej. (Ajtmatov) rain-INST with children-INST wholefamily-INST 'They all wanted to be together in the rain, with the children, as a whole family.'
(257) Mysli Thoughts-NOM
roem
pronosilis'
swarm-INST
rushed-REFL in
v
ee
utra
dom
From moming-GEN house-NOM
Larinoj
gostjami
Mrs. Larin-GEN
guests-INST
celymisem'jami sosedi all full-NOM whole families-INST neighbors-NOM s"exalis'. (Puskin; in Staniseva 1958c) gathered-REFL
'Mrs. Larin's house has been full of guests since morning - neighbors gathered there in whole families. ,
tysjacami
rozdalis' born-REFL
moi uceniki.
mousands-msr my disciples-NOM (Majakovskij) '...so that my disciples may be born by the thousands.'
so-that
(260) Fasisty
unietozali
Fascists-NOM destroyed
ljudej people-Ace
vedrami. nosjat vodu buckets-INST Firemen-NOM carry water-Ace 'Firemencarry water by the bucket.'
5.3.2. Attributive instrumental-qualitative
ves' polon -
(259) ...etob
(262) Pozamye
her
golove. head-toe 'Thoughtsrushedaboutin her head like/in a swarm.' (258) S
Further, there are some mass nouns whichcan be quantified with a juxtaposed instrumental, as in (262):
Here the instrumental describes a (usually temporary) characteristic of the participant with whichit is juxtaposed. In predicate position this instrumental contrasts with constructions using a predicate nominative. The predicate nominative, however, is not juxtaposed with the subjectbut rather equatedwith it. The nominative describessomething which is equivalentto the subject, whereas the instrumental describesonly an attribute. 163 This distinction is borne out by comparingsimilar sentenceswhich vary in case usage. (263a) is an equation of two uniquelyspecified noun phrases and uses the nominative. (263b) labels something other than an inherentcharacteristic, and it does so with the instrumental: (263) a.
b.
millionami. miilions-INST
'The fascists killed people by the millions.' Example (257) demonstrates the overlapthat exists between attributive instrumentals of quantification and comparison. Some noun phrasesdo not in and of themselves denote groups but take on this meaning in the context of this usage, as in (261), where odna kniga "one book" becomes a unit for grouping stories:
Adam
byl
pervyj celovek
na
zemle.
Adam-NOM
was
first man-NOM
on
earth-LOC
'Adam was the first man on earth.' Adam byl pervym zemledel'cem. first farmer-INST Adam-NOM was 'Adamwas the first farmer.'
Although this use is most frequently associated with verbs meaning "be" or "become" (264), it is used with other verbs as well (265) - (268). (264) Amerikanskij eitatel'
stal vymirajuscej osob'ju. became dyingindividual-INST 'The American readerhas become a dying breed.'
American reader-NOM
(265) Segodnja (261) Ego rasskazy izdali odnoj knigoj. His stories-Ace published one book-INST 'His stories were published as a book.'
Today
Sajmerden
sidit
Sajmerden-NOM sits 'Today Sajmerden is sitting as dispatcher.'
dispetcerom. (Ajtmatov) dispatcher-INST
176
177
(266) V
sil'nyj moroz sljuna spittle-NOM In stong frost-Ace zemlju. (in Potebnja [1958])
l'dom
padaet na
pervogo russkogo universiteta.
ice-INST
falls
first Russian university-GEN 'Lomonosov is the founder of the first Russian university.'
on
earth-Ace 'In extreme cold spittle falls to the ground as ice.'
(267) I
vse ze
imenno
takie stampy
And
still precisely these stock-phrases-NOM sluzat postojannoj zivoj bazoj living basiS-INST serve constant-INST and
Further support for the argument that this use of the instrumental is characteristically attributive comes from the fact that it does not require a full noun phrase and may be employed with merely an adjective (implying that the instrumental is a modifier, whose head is the noun phrase with which it is juxtaposed): (272) K
semanticeskix novoobrazovanij. 'And still it is precisely these stock-phrases which serve as a constant and living basis for semantic neologisms.' on
mectal
cern
o
Boy-INST
he-NOM dreamed 'As a boy he dreamed of airplanes.'
samoletax. about airplanes-toe
Like the quantitative attributive, this instrumental can also be interpreted as invoking comparison: druz'jami. (269) My rasstalis'
osobnjakom. stands loner-INST 'The Slavic instrumental case stands alone.'
zacem
mudrit',
ne
why
to-be-clever
not
ravno,
gde
even-NOM
where to-be
byt'
dele takom matter such-LOC
tern
lucse. (Ajtmatov)
by-that-much better
'What's the point in dragging about and complicating things, Who cares where you are buried, in such matters, it's the sooner the better.' (273) Volcata
rodilis'
Wolf-cubs-NOM
slepymi.
born-REFL 'The wolf-cubs were born blind.'
parted-REFL tnends-ncsr 'We parted as friends/like friends. '
stoit
bystrej,
by-how-much faster
We-NOM
(270) Slavjanskij tvoritel'nyj padez
tjanut',
what-DAT to-drag vse Ii all-NOM INTERROGPART zarytym, v buried-PASTPASSPART-INST in
semantic neologisms-GEN
(268) Mal'cikom
cemu
To
blind-INST
A sampling of the syntactic variants of the qualitativeattributive instrumental
Slavic instrumental case-NOM
The use of the instrumental with the verb javljat'sja "be; serve (as)" is wellmotivated in the present model. With this verb, the nominative noun phrase always indexes a specific referent, whereas the instrumental invokes a general category which the nominative's referent belongs to. The instrumental serves as a label applied to the nominative:
Although all of the examplesof the qualitative attributive instrumentalgiven above appear in intransitive clauses (schema 3a), this subtype is syntactically the most versatile. The following examples illustrate other kinds of constructions in which this instrumentalappears: Juxtaposed with nominativesubject of a copular verb (schema 3a') Deviceju
ona
pevun'ja,
javljaetsja
osnovatelem
Girl-INST
she-NOM
sonstress-NOM
is-REFL
founder-INST
zamuz,
zamolknet.Jvt
(274)
(271) Lomonosov Lomonosov-NOM
a but
vyjdet will-get
married will-hush-up 'As a girl she's a songstress, but (when) she gets married she'll hush up.'
178
179
zverem i pognal ego. (Ajtmatov) beast-INST and drove-away him-ACC 'Karanar (name of a camel) rushed at him like a wild beast and drove him away.'
Juxtaposedwith a bare nominative (schema3b) (275) Sluzba oficerom v Sovetskoj armiiService-NOM officer-INST in Soviet army-LOC prestiznoe zanjatie. prestigious occupation-NOM 'Service as an officer in the Sovietarmy is a prestigious occupation.'
(280) Eto vino -
voda water-NOM This wine-NOM 'This wine is very watery.'
vodoj. water-INST
Juxtaposed with accusative object (schema3c) (276) On pomnit menja He-NOM remembers me-ACC 'He remembers me as a hussar.'
5.3.3. Attributive instrumental- comparative
gusarom. hussar-INST
The comparative attributive instrumental is instantiated when the instrumental noun phrase and the noun phrase with which it is juxtaposed necessarily refer to two distinct entities. This perception of incompatibility between the two noun phrases is dependent not only on lexicon, but also on context, as the following two examples prove. In both of them a person is referred to as an animal, but in the first examplecontext (Hindubeliefin reincarnation) allows for the.animal and the person to be identified as the same individual, whereas in the second this is impossible and the instrumental can only be interpreted as a comparison:
(277) Oni ne privolokli pokojnika They-NOM not dragged deceased-Ace nepogrebennym domoj. unburied-PASTPASSPART-INST home 'They did not drag the deceasedhome unburied.' Juxtaposed with a dative in an impersonal clause (schema3d) (278) Bee junkerom slucilos' Still cadet-INST happened-REFL 'It so happenedthat I was still a cadet...'
mne me-DAT
Indii (281) V seitaetsja, cto kaZdyj celovek, In India-toe considers-REFL that every person-NOM kogda-to, esce do svoegorozdenija, pobyval do at-some-time still to own birth-GEN stayed to etogo pticej, iIi zverem kakim-nibud', iIi that-GEN bird-INST or beast some-sort-INST or nasekomym. (Ajtmatov) insect-INST 'In India they believe that every person was, at some time before his birth, a bird, or some kind ofbeast, or an insect.'
byt'... to-be
Tautological instrumental The tautological instrumental is a varietyof the qualitative attributive instrumental. It can be formed either by augmenting the instrumental noun phrase in an intransi-
tive clause with an identical nominative noun phrase as in (279), which instantiates schema 3a, or by adding an identicalinstrumental noun phrase to a predicate nominative in a copular sentence, as in (280), which instantiates schema 3a'.I65 In either case, the effect is one of intensification. (279) Karanar kinulsja na Karanar-NOM rushed-REFL at
nego him-ACC
(282) I esli kak-nibud' na mig And if somehow for moment-Ace Udastsja mne zabyt'sjaSucceeds-REFL me-DAT t010rget-REFL Pamjat'ju k nedavnej starine Memory-INST to not-distant old-times-DAT
zver' beast-NOM
1
180
181
Lecu ja vol'noj, vol'noj pticej. Fly I-NOM free, free bird-INST (Lermontov; in Xodova 1958b) 'And if I somehow manage to sink into reverie for a moment, through my memory I fly like a free, free bird to old times not so long gone.'
Schemas Pictorial
Prose
A nominative together with an instrumental acts on an accusative in a setting.
Usage
Comitative with preposition .§.
The use of the instrumental for comparison is highly circumscribed. It can be juxtaposed only with nominatives, and must be clearly recognizable as referring to an entity other than that specified by the nominative. (283) and (284) present typical examples which fulfill these requirements: A
(283) Dekabr'skij rassvet vstaet nad Moskvoj December dawn-NOM rises above Moscow-INST gorjackoj tifoznoj. (Majakovskij) fever typhoid-INST 'The December dawn rises above Moscow like typhoid fever:' (284) nos kartoskoj (in Leska Ms) nose-NOM potato-INST 'a nose like a potato'
A nominative close to an instrumental acts on an accusative in a setting.
----""":s:;";e'tting
B
A dawn is clearly not the same thing as typhoid fever, and, despite similarities in size and shape, a nose cannot be a potato. When, however, the two noun phrases could plausibly be referring to the same entity, instrumental of comparison ceases to be an option. Thus because a person could indeed be an orator, even though we perhaps know that the individual involved is not one, we cannot say:
A nominative close to an instrumental & an instrumental' acts on an accusative in a setting.
Location with prepositions za - viewed from A, horizontal Rod - viewed from A, vertical pered - viewed from B, horizontal nad - viewed from B, vertical
Location with preposition mezdu
setting
(285) *On govorit He-NOM speaks 'He talks like an orator.'
oratorom. orator-INST
The choice of verb in this example, as in example (236), disallows a qualitative reading of the instrumental, for it describes not a state of existence or characteristic action, but rather an activity that can be performed in various ways. Thus the comparative reading is favored, but the lack of resolution between the referents of the nominative and the instrumental renders this utterance ungrammatical.
Schema 4 subnetwork
0) ~
Loss of bond
bet~~en ~ 4 a redUPlication-B participant & of instrumental 4a' Instrumental
Figure 24. Schema 4 and variants
183
182
5.3.4. Summary of the attributive instrumental Like the conduit instrumental, the attributive instrumental unifies several uses of the instrumental which are usually recognized as distinct submeanings. Here the predicate instrumental, which is frequently considered rather problematic and exceptional is shown to function as a regular member of a paradigm in which instrumentals are juxtaposed with event chain participants as attributes.
5.4. Instrumental used with a prepositon All prepositional uses of the instrumental in Russian involve the removal of the instrumental entity from the event chain. In some cases (uses with the preposition s "with") it retains a bond with one of the event chain participants, but in the remainder it is merely proximate to the participant. The instrumental case itself specifies only the relative position of the instrumental to the event chain participant; all of the remaining information (on whether there is a bond, and, if not, whether orientation is vertical or horizontal and where the relationship is viewed from) is supplied by the accompanying prepositions. l 66
5.4. 1. Comitative instrumental Comitative uses of the Russian instrumental are instantiated with the preposition s "with" and are captioned by schema 4. There is strong evidence of a transitional zone linking the conduit instrumental (specifically of instrument and manner) with the comitative instrumental.lv" As the following examples show, several factors contribute to the construal of a peripheral object as either an instrument or an accessory. These examples are grouped according to obligatoriness of prepositional or bare case usage:
Bare instrumental only: (286) vspaxivat' polja volami oxen-INST to-plow fields-ACC 'to plow the fields with oxen'
(287) Krestjane obrabatyvali zemlju Peasants-NOM worked land-Ace 'The peasants worked the land with hoes.'
motygami. hOeS-INST
(288) Xuligan razbil okno Hooligan-NOM broke window-Ace 'The hooligan broke the window with a rock.'
kamnem. rock-INST
(289) Etoj xitrost'ju on ubil dvux zajcev. That clever-trick-INST he-NOM killed two hares-ACC 'By that clever trick he killed two hares.'
Prepositional instrumental only:
resal etu zadacu so (290) Inzener Engineer-NOM solved that problem-Ace with specialistom po komp'juteram: specialist-INST along computers-DAT 'The engineer solved that problem with a computer specialist.' (291) Slepec s sobakoj peresel Blind-man-NOM with dog-INST walked-across ulicu. street-Ace 'The blind man crossed the street with a dog.' (292) Oxotnik posel na olenja s Hunter-NOM went at deer-Ace with 'The hunter went after the deer with a gun.'
ruz'em. gun-INST
Examples (286) - (290) demonstrate the role which the nature of the instrumental referent has in determining whether it is a participant's instrument or accomplice in a transitive action. Inanimate objects (hoes, rocks), animals (oxen), and parts of the subject (his cleverness) can be fully subjugated and wielded by the subject as instruments. Human beings acting according to their own volition (specialists) cannot be conceived of as instruments at all: (291) - (292) show that animacy and volition are not the only factors which contribute to a comitative readng. In clauses with motion verbs requiring no conveyance, for example, it is in-
185
184
conceivable that any instrument could be needed; here animals and even inanimate objects can qualify as adjuncts of the subject. The use of nouns denoting abstractqualities with the preposition s to describe states attendant to the action furtherprovides examples bridging comitative instrumentaland conduit instrumental of manner: (293) S
udovol'stviem
vam
pleasure-INST you-DAT 'I will help you with pleasure.'
With
pomogu.
(297) Cto
I-will-help
What-NOM
We-NOM
kupili
divan
s
vysokoj spinkoj.
bought
couch-ACC
with
high back-INST
podnjal
Edigej-NOM
raised
nee
s
golovu,
gljanul
na
head-ACC glanced gor'koj usmeskoj. (Ajtmatov)
at
with bittergrin-INST 'Edigejraisedhis headand looked at her with a bitter grin.'
her-ACC
Although the instrumental entity may also be a participant in a mutualaction, its participation is only implied by the semantics of the verb. The utterances in (296a) certainlyimply (296b), but this is purely a consequence of extra-linguistic knowledge (the fact that we know that we can neitherfight nor converse with someone unless he is also fighting or conversing with us). The instrumental clause in each case gives a distinct expression rather than just a transformation of the corresponding clause; in (296a) one participant (the nominative entity) is viewed as controlling the action, in which the instrumental entity is merelya secondary participant. (296b) viewsneitherentityas dominant. (296) a.
b.
On He-NOM
borolsjalbesedoval
s
bratom.
fought/conversed
with
brother-INST
'He fought/conversed with his brother.' Brat'ja borolis'/besedovali. Brothers-NOM
fought-REFUconversed
'The brothers fought/conversed.'
toboj?
with
YOu-INST
(298) S
proizosel him-INST happened 'A funny thinghappened to him.' nim
With
smesnoj slucaj. funny case-NOM
5.4.2. Proximate instrumental
'We boughta couch with a high back.' (295) Edigej
s
'What's with you?'
The overlayof a part-whole schemamakesit possiblefor parts of a participant to be presented as accessories which accompany it in a comitative relationship: (294) My
There is a type of construction in Russian in which a (usuallyhuman) referent is placedin a comitative relationship with an abstract objector event. Here the fact that the only human being in the sentence is not a full participant but rather the secondary associate of something else carries with it the implication that the personnamedis a victim of the situation which he is thus associated with:
Four prepositions which appear with the instrumental, za "behind", pered "in front of", pod "under", and nad "above", all encode the same relationship of the instrumental entity to a participant, that of simple proximity, as in schema 4a. The prepositions themselves further specify this proximityin terms of two parameters. The orientation of the entities maybe horizontal (za and pered ) or vertical (pod and nad); and the relationship may be viewedfrom a point opposite the instrumental entity (pered and nad) or from a point opposite the (usually nominative) participant (za and pod). Since horizontal and vertical orientation are treated synonymously, as are opposite viewing arrangements, all of these uses illustrate semanticextension via synonymy and antonymy. This model has the obviousadvantageof revealing the perfectparadigm constituted by these prepositions. These prepositions are mainlyused in expressions of physical locationor in straight-forward metaphorical extensions of such expressions. Only a few examples deserve furthercomment. 'ZLl- "behind"
This prepositionhas a number of idiomatic uses originally based on the concrete meaning. 'ZLl obedom "during lunch" designates a period of time, but it is specified by the position one takes at that time;i.e., seatedbehindthe table facing one's meal. The Russian way of saying "one after another" also employs za:
187
186
(299) On kurit sigaretu za He-NOM smokes cigarette-Ace behind sigaretoj. cigarette-INST 'He smokes one cigarette after another.'
za (304) Brat sxodil v magazin after Brother-NOM went to store-Ace 'Brother went to the store to get cigarettes.'
papirosami. cigarettes-INST
Pered - "in front of' This is use of za evokes the image of a series of objects lined up so that all subsequent objects stand behind the first from the perspective of the viewer. The modem-day adverb zamuzem "married (of women)" is of course based on a prepositional phrase inspired by the traditional position of the married woman who stands behind her husband. Since the object of za is located between the viewer and another object, it frequently refers to barriers, as in (300), which may also be metaphorical, as in (301):
In addition to expressions of physical location, this preposition is occasionally used in comparisons. Here an item is compared to another by being placed before
it: (305) Ivan pered nim Ivan-NOM before him-INST 'Ivan is nothing compared to him.'
nieto. nothing-NOM
(300) Za derev'jami ne vidat' lesa. Behind trees-INST not to-see forest-GEN 'You can't see the forest for the trees.'
Pod - "under"
(301) No, za zuejskimi xlopotami, ne uspel. But behind worldly cares-INST not managed (in Bulygina 1958b) 'But, due to worldly cares, he didn't manage [to do it].'
Since cities were usually built on hills, this preposition was used to describe the physical location of suburban areas which lay below them, and continues to be used no matter what the geography of the city, as in pod Moskvoj "near Moscow". Since titles and names typically appear over doors or at the top of documents they are also associated with this preposition: pod nazvaniem "with the name/title" .
Alternatively, za can set up a situation in which an object is conceived of as following behind the object of za, as in (302), andthis usage is also subject to metaphorical extension, as in (303). Furthermore, the object sought may be stationary, and thus may describe only the endpoint of the path, in which case it is conceived of as a goal, as in (304):
Nad - "above"
(302) Sobaki begut za zverem. Dogs-NOM run after beast-INST 'The dogs are running after the beast.'
ee dvizenijami. za (303) On sledil her movements-INST after He-NOM folowed (in Bulygina 1958b) 'He followed her movements (i.e., he observed her).'
This preposition can be used metaphorically to referto the subject of research, as in rabota nad dissertaciej "work on [one IS] dissertation" .
Reduplicated proximate instrumental mezdu "between" There exists yet one more preposition which instantiates the proximate instrumental, but requires that the instrumental entity be multiple (Le., either plural (306) or marked on more than one object (307)). Because this schema (4a') contains two (or more) points of reference, further specifications (position of viewer, orientation) are unneccesary, This preposition is meidu "between; among".
188
189
(306) Mezdu soldatami proizoslo Among soldiers-INST happened (Puskin; in Svedova 1982)168 'The soldiers fell into confusion.'
smjatenie. confusion-NOM
nadeemsja na ulucsenie otnosenij hope-REFL at improvement-Ace relations-GEN mezdu SSA i Sovetskim Sojuzom. between USA-INST and Soviet Union-INST 'We hope that there will be better relations between the USA and the Soviet Union.'
(307) My We-NOM
In the cognitive model, the prepositonal uses of case are integrated into the overall account,169 and the relative semantic contributions of case and prepositions are carefully distinguished.
Paradigmatic Relationships
r Ieb r -J rdelocalization
shift from event to participant
shift from event to participant
delocallzatlon
5.5. Instrumental networkbound by paradigmofperipherality and by alliance In addition to the paradigmatic bonds between schemas that hold the instrumental network together, the tracking of transitional examples linking schemas reveals another set of alliancesbinding the same network. As opposed to the Czech dative that was shown to have two schemas vying for dominance, the secondary system of alliances of the Russian instrumental further confirms schema 1 in the role of prototype.170 Because all of the schemas except schema 1 deemphasize the relationship between the instrumental and the event chain (in schema 2 the instrumental is delocalized to the point of becoming the setting, and in schemas 3 and 4 the instrumental uses a participant rather than the event as its point of reference), the role of syntagmatic parallelism in providing secondary binding for the network is far less significant for the instrumental than for the dative.
Alliance via Transitional Examples
~
paxnut' 'smell' condUit/comparison (exx. 235-236)
coLe determined by context (exx. 286-292)
inst of space as opening conduit/space (ex. 242)
--0 2
Figure 25. Two realizations of the Russian instrumental network
6.
The Czech instrumental
Aside from certain spatial uses of this case, Czech's instrumental is overall less intricate than that of Russian.
6.1. Conduit instrumental
The local subnetwork of the conduit instrumental in Czech makes less use of the instrumental in verbal government (schema la) and the construction which expresses inexplicable force (captioned by schema lb of that network), but is otherwise comparable in its semantic application. Like Russian, Czech uses the central member of the conduit subnetwork to express the use of instruments in effecting transitive actions. dykou. (308) Vrah zabil obet' Assassin-NOM killed victim-Ace dagger-INST 'The assassin killed his victim with a dagger.'
(309) Uhasil prsty cigaretu. Extinguished nngers-rssr cigarette-ACC 'He extinguished the cigarette with his fingers.' Intransitive constructions (captioned by schema la) display the same range of voice phenomena as those found in Russian. In active constructions the instrumental most frequently marks an instrument in the realization of the action. As with Russian, the so-called "instrumental object" (as in (310) is subsumed under this heading rather than being recognized as a separate phenomenon. (310) Umel dobre malovat Knew-how well paint 'He could paint well with a brush.'
stetcem. (in Grepl- Karl1k: 1986) brush-INST
(311) Skrz Jugoslavii jsme Through Yugoslavia-Ace are-AUX 'We rode through Yugoslavia by train.'
jezdili vlakem. rode train-INST
192
193
(312) "Mluvte!" volala netrpelive a dupla nohou. Speak called impatiently and stamped foot-INST , "Speak!" she called out impatiently and stamped herfoot.' In some examples, such as those below, the noun phrase marked by the instrumental is arguably the agent rather than the instrument of the action. Such examples evidence a transition between active and passive constructions: (313)
Leto sviti Zivjmi barvami. Summer-NOM shines lively colors-INST 'Summer shines with lively colors.'
(314) Obyvatelstvo trpf malarii. matana-msr Populace-NOM suffers 'The populace suffers from malaria.' In passive constructions the instrumental usually identifies the agent, although in some examples the interpretation of the instrumental as instrument or agent is dependent on the meaning of the noun phrase so marked. In the first example below, the knife is clearly the instrument. The ink in the second example likewise serves as instrument, but the scholar must be recognized as the agent. The last two examples are more typical of passive constructions - in both, the instrumental noun phrase unambiguously represents the agent. (315)
Timhle nozem se This knife-INST REFL-SCC 'This knife cuts well.'
(316)
a.
b.
dobre krajf, well cuts
Dopis byl napsan Letter-NOM was written-PASTPASSPART cervenym inkoustem. red ink-INST 'The letter was written with red ink.' Dopis byl napsan Letter-NOM was written-PASTPASSPART velkym vedcem. great scholar-INST 'The letter was written by a great scholar.'
(317) Fakulta se naplnila College-NOM REFL-ACC filled-up (in Grepl - Karlfk 1986) 'The college became filled with students.' (318)
studenty. students-INST
Diichodce byl prepaden Pensioner-NOM was attacked-PASTPASSPART chuligdny, hooligans-INST 'The pensioner was attacked by hooligans.'
Because Czech includes verbs of domination under dative government, its use of the instrumental with verbs is almost entirely restricted to instrumental objects (body parts and other conduits). There are only four verbs which might qualify as instrumental-governing: 171 vlddnout disponovat hospodatit opovrhovat
"rule; know how to use" "make use of, have at one's disposition, take advantage of" "manage; spend" "scorn"
It seems, though, that the instrumental entities associated with the first three verbs are employed as instruments in actions which are implied but not further specified by these verbs. The first verb, vlddnout "rule", distinguishes between instruments and other objects by governing different cases, thus providing good evidence that the instrumental is restricted to marking instruments. The dative is used with human referents (which cannot be wielded as instruments), as in vladnout lidu "rule the people" , but the instrumental is used with inanimates, such as vlddnout domem "rule the house". When vladnout means "know how to use", its objects are necessarily instruments and the use of the instrumental is required, as in vlddnout stetcem "know how to use a paint-brush", and nevlddnout svym casem "not know how to use one's time". The remaining two verbs are used exclusively with inanimates which are conceived of as instruments, and the last of these, hospodatit is used more frequently with the comitative preposition s than with the bare instrumental. Schema la' (a minor variation on schema Ia, in which the intransitive verb is replaced by a copular existential and a nominative predicate) is also present in Czech:
194
195
(319) Svetoznamy ftlozofErazim Kohak World-famous philosopher Erazim Kohak-uou
je
rodem
is
birth-INST
Cech.
(322)
skublo. Peter-INST shuddered 'Peter shuddered.'
Petrem
Czech-NOM
'The world-famous philosopherErazim Kohak is a Czech by birth.'
6.2. Instrumental ofsetting The instrumental of inexplicable force is very rare in Czech. It is cited in a footnote in Grepl- Karlik 1986: 116 as "possible", but native speakers do not accept or use such examples (one of which appears below). Failure to develop this particular branch of the instrumental network seems to be motivated by the same forces which have caused some impersonal uses of the dative to wither away in that language: Czech does not tolerate sentences which lack a nominative subject as readily as Russian does. The few examples of inexplicable force instrumentals in Czech that can be found in grammars refer to sensations of temperature, and, if recognizedat all, they must be classed as variants of the instrumental of sensation (schema lc), rather than as examples of instrumental of inexplicable force of the kind found in Russian, for their semantic scope more nearly approximates the former than the latter.
Transitional examples linking the conduit instrumental to the instrumental of setting in which the instrumental entity is an opening are rare in Russian, but common in Czech, and can even be used with non-motionverbs: (323) Kocka
dirou, pes oknem. crawls hole-INST dog-NOM window-INST 'The cat crawls through a hole, the dog through the window.' leze
Cat-NOM
(324) Med'a
se
Teddy-bear-NOM
REFL-ACC
diva oknem na looks window-lNST at
mesfcek, moon-ACC
(320) ?Rozffaslo
ho
zimou. (in Grepl- Karlik 1986)
him-ACC cold-INST 'He was shaken by the cold.'
'The teddy bear is looking through the window at the moon.'
Shook
The Czech instrumental of sensationwhich appears in subjectlessconstructions is virtually identical to its Russian counterpart, and commonly refers to temperature and smells:
The use of the instrumental to describe a path is likewise more vigorous in Czech than in Russian. Walking down a street is usually described with an instrumental in Czech, but with a po phrase in Russian: Celetnou ulici. (Czech) am-AUX CeletnaStreet-INST 'We walked down Celetnd Street. '
(325) Sli
jsme
Walked
(321) Trochu A-bit
zavanelo
cpelo smelled bahnem.
trochu gasoline-INST a-bit
benzinem,
bog-INST wafted 'It smelled a bit ofgasoline and there wafted a bit of the scent ofa bog.' In addition, Czech has some impersonal verbs which refer to sudden body movements and use the same argument structure as the instrumental of sensation. Among them are skubnout "shudder" and trhnout "jerk":
(326) My We-NOM
sli
po
walked
along Nevskij prospekt-DAT
Nevskomu prospektu. (Russian)
'We walked down Nevskij prospekt.' Czech can also employ referents which are not themselves paths to indicate the shape of the trajectory:
196
197
(327) Projdeme
pes
tema kabdtama.
these coats-INST 'Let's walk through (going) along these coats (uttered in a department store).'
friend-INST always 'A dog is an animal only from the point of view of a zoologist, for me, a dog is always a friend.'
Let's-walk-through
The instrumentalof time, however, has all but died out in Czech, leaving only fossilized expressions which are markedly archaic, such as dnem i noci "day and night" (which is more frequently expressed using prepositional phrases: ve dne i v noci ). Discrete plural uses of the instrumental of time are encountered in Czech (e.g. chvflema "from time to time"), but plural and contiguous uses are not.
plitelem. 173
vzdy
dog-NOM
The first sentence leaves the question of whether the town is a location or a special context undecided, and both cases can be used. When the locative phrase is moved to the left, as in (328b), the town becomes an environment which gives the attribute special relevance (i.e., all towns are expected to have a full set of craftsmen). Examples (329a) - (329b) likewise show that the use of the instrumental is influenced by the contextin which the attribute is viewed.
6.3. Attributive instrumental 6.4. Instrumental with a preposition
Attributive uses of the instrumental in Czech are largely limited to qualitative descriptions,172 and even here the use of these expressionsis flagging. In modern Czech attributes are usually marked with the nominative unless there is a context to support the idea that the object is being viewed in a special way, in a particular environmentor from a point of view which makes the attribute stand out from the participant rather than being merely equated with it. The following examples illustrate this point:
The Czech prepositional instrumental differs from its Russian counterpart only in that the Czech equivalent of pered, pled' 'before" is somewhat more versatile. Like za, it can be used with motion verbs, but has of course the opposite meaning; rather than describing motion following behind,it describes motion going before, as in: (330) Uprchli
(328) a.
Jeho otec His father-NOM
byl
krejcflkrejcim
v
was
tailor-NOMltailor-INST
in
Bydzove.
(329) a.
b.
do
Anglie
pred
Escaped in-time to England-GEN before Hitlerovym rasistickym zdkonikem. (in Bulygina 1958b)
Hitler's rascist legislation-INST 'They escaped in time to England (fleeing) before Hitler's racist legis-
Bydiov-LOC
b.
vcas
'His father was a tailor/the tailor in Bydzov.' (NOM or INST) Jeho otec byl v Bydzove krejcim. His father-NOM was in Bydiov-LOC tailor-INST 'His father was the tailor in Bydzov.' (INST only) Pes
je
zvffe.
Dog-NOM
is
animal-NOM
'A dog is an animal.' (NOM only) Pes je zviletem Dog-NOM is animal-INST hlediska zoologa, point-of-view-Gun
jen
z
only
from
pro
mne
zoologist-GEN for
me-ACC
je is
lation.' Pled can also be applied to the domain of time, where it refers to points preceding the present on the time line, as in pled tiemi roky "three years ago", and tyden pled tim, jak odjel "a week before (the time when) he left". As the figure shows, the basic structureof the instrumentalcase is identical for the two languages. Variations show up mainly in the peripheral branching of schemas and in the way in which they are applied to non-prototypical (i.e., nonspatial or non-physical) domains. It is apparent that although parts of the network structure have waxed and waned through the individual histories of Czech and Russian, the overall shape of the categoryhas been retained, and much of the evolution has proceded in a parallel fashion. Despite variations, case categories are remarkablystable and predictable in form.
198
7.
Czech comparison (archaic) qualitative 3
)-------i 4
space
Russian comparison qualitative quantitative
3 )-------i 4
)-------{ 2 space time
I
contiguous time
Figure 26. Comparison of Czech and Russian instrumental
Concluding remarks and possible universals
The foregoing analysis of Czech and Russian should suffice to convince the reader that surface case categories do indeed have internal structure and that the cognitive model is adequate to the task of describing this structure. Furthermore, an examination of the internal structure of cases: - Makes it possible to differentiate between case semantics and semantics contributed by other sources (e.g. lexicon, context, encyclopedic knowledge), since it obviates what is systematically related to the given case and what is not. Reveals significant patterns that might otherwise go unnoticed. Examples include: the paradigm of the indirect object in Czech which contains giving, taking, and intransitive giving and taking verbs Czech dative verbal government with verbs denoting symmetrical relationships, and transitive relationships of both domination and subordination, as well as annulment of subordination (a pattern that is repeated in the semantics of dative-governing adjectives and prepositions and in reciprocal uses of the dative reflexive si) the array of mapped extensions of the Czech dative, which correlates with their significance as pragmatic rather than obligatoy grammatical uses of case the Russian conduit instrumental, which unites the instrumentals of instrument, passive agent, verbal government, inexplicable force and impersonal constructions in a single syntagm of disappearing participants the prepositional uses of both the Czech and the Russian instrumental, which form a compact and gapless system. The correlation of markedness with relative distance from the prototype provides a new expression for markedness theory. Markedness is indeed compatible with the hierarchical branched structure of cognitive categories. And, since the marked or unmarked value of features can vary from culture to culture and change with time,174 the implementation of markedness theory would not compromise the observed plasticity of the network structure. The central schema of a given category is the unmarked member; paradigmatic variants thereof are necessarily marked. This explains in part why syntagmatic variants of the central schema are semantically more significant than those of other schemas. According to marked-
201
200
ness theory, further (meaningful) variation is possible only for the unmarked member (schema 1). Both the dative and the instrumental bear this out. The cognitive model incorporates notions put forth by scholars in previous traditions (such as the centrality of spatial movement proposed by localists; structuralist markedness theory and claims that submeanings are interrelated and that the dative and instrumental have the feature [+peripheralD, and actually further empowers these notions by placing them in a framework which makes these descriptions apt and accessible. The cognitive model suggests universal strategies in the evolution and structure of case categories: The semantics of any given case category in any given natural language represent a single, coherent cognitive category. b) The cognitive category of a case has internal structure, determined by the following relationships of variation: (I) Variations in the shape of schemas (i) paradigmatic (ii) syntagmatic Metaphoric extension (II) (i) via synonymy, antonymy, or metonymy (ii) via subjective mapping of perceptions or mapping onto the speech-act domain c) The relationships that structure the network are arranged hierarchically, such that each successive type produces local variations on the results of more basic types of variation. This hierarchy of types of variation determines both synchronic structure and diachronic development. This does not mean that there is a stage-by-stage lock-step development that proceeds via successive applications of these types of variation. Rather, '-the claim is only that these are the strategies that are employed, and that when a paradigmatic variant exists it can be further subject to any of these types of variation and extension (whereas a syntagmatic variant can be subject only to further syntagmatic variation or metaphoric extension). In every case the actual realization of these types of variation is simply an option , that mayor may not be exercised. Also, the synchronic structure of a given category need not mirror its diachronic development; indeed the dynamic forces at work in semantic categories can mask the record of historical development. d) The hierarchy of variants reveals the relative status of category members along the scale ranging from grammatical (syntactic) to purely semantic.
e)
a)
f)
g)
Core paradigmatic and syntagmatic variants are more likely to be obligatory and essentially grammatical. Uses that are less central to the category (and in particular those produced by means of metaphoric extension) are more likely to be largely semantic and non-obligatory. This universal is a realization of the intricate relationship between syntactic and semantic uses of case recognized by Kurylowicz, recast in terms of the cognitive framework. The overall structure of the network that specifies which variations are indeed realized is language-specific and is a part of the language to which it belongs. It is not something that is constantly generated from some "coreplus-rules". The differences in case category structure that are peculiar to Czech and Russian provide ample proof of this point. Although the overall shape of a case category is quite stable, the internal structure is dynamic and subject to constant adjustment. A case category does not simply reach a state of development and freeze there. Case uses may be extended or discarded as the need arises. This type of change, via branching and pruning of the internal structure is responsible for most of case category evolution. The development of variants actually enhances cohesion rather than causing the case category to weaken into an overextended and overly complex conglomeration of uses, because such variants engender new alliances between members of the network and serve to reinforce its structure. While case categories playa primary and clearly definable role in the semantics of sentence structure, concepts such as passivization, Subject, Object, Indirect Object emerge as secondary notions based on typical case usage, but influenced also by lexical meaning and context.
Notes Chapterone 1.
2. 3.
4. 5.
6. 7. 8.
9.
10. 11. 12. 13. 14. 15.
Cf. the oft-quoted statements that "case-categories ... resemble various situations of the world, but never with any consistency", Bloomfield (1966: 272) and that cases are "one of the most irrational parts of language in general" Jespersen ([1965]: 186). For a discussion, see Burston (1977: 29-30). Brecht - Levine (1986: 18). See also Skalicka (1950: 134-135), who likewise points out the obstacles to semantic description which are unique to case. Note, for example, the basic division made by Chomsky between syntax and the lexicon (which contains semantics), and the set theory approach imposed by feature systems. Although logical empiricism currently dominates linguistic theory, its ascendancy is recent. As Geeraerts (1987: 694) correctly points out, the "prestructuralist, historicalphilological tradition of semantic research" was not founded on these principles and is not incompatible with the cognitive paradigm. Citations and a summary of this work will be provided below in section 1.1.1. This group includes Fauconnier, Lakoff, Langacker, and Talmy. Accounts of specific linguistic phenomena which use this framework have been undertaken by Brugman, Deane, Lindner, Rudzka-Ostyn, Smith and Sweetser. See the bibliography for citations of relevant works. For a summary of how cognitive semantics developed, see Janda (1986: 44-60). The most detailed account to date of what cognitive semantics entails and of its relationships to other disciplines is Lakoff (1987). Lakoff (1987: 279-280) argues that most, if not all, cognitive categories cannot be defined using primitives. See, for example, DeGroot's (1965) article, in which he examines the thirty traditionally recognized uses of the Latin genitive and Mrazek's (1964) book, which gives an encyclopedic account of a similar number of functions for the Russian instrumental. Cf. the generalizations produced by the feature specifications of Jakobson and van Schoonefeld, which have been widely criticized (for more discussion, see section 2. below). For a fuller account of the irreconcilability of unity and diversity within a settheoretical framework, see Janda (1986: 33-43). "Kinesthetic image schemas" make use of basic experiential structures such as in-out, updown, and part-whole (cf. Lakoff 1987: 271-275). "Conventional imagery" refers to the depiction of image-schemas by means of figures which conventionally represent entities, actions, etc., as described below. A detailed discussion of comments made by scholars outside the school of cognitive semantics supporting this framework will be undertaken in section 2.2. Cf. Rosch 1973a and Merivs - Rosch 1981: 92. Cf. Churchland 1986 and see section 1.1.2. below. For a bibliography, see Rochester Connectionist Papers: 1979-1985. "It is precisely on this count that existing theories have broken down... The approach taken here is to use a computational paradigm that is similar to the human brain in form and functional capabilities." Cottrell (1985: 4). Van Schooneveld 1978: 11. See also van Schooneveld (1986).
204 16.
17. 18.
19. 20.
21.
22.
23.
24.
25. 26. 27. 28. 29.
205 Cf. Sgall, Petr, Eva Hajicova & Jannila Panevova (1986: 34), who postulate that "linguistic meaning comprises pragmatic as well as semantic phenomena." DeGroot (1965: 192) makes a similar distinction when he recognizes that "the semantic units of the wordcontent of a language - words and word-groups - fall into three categories. They are either attitudinal, or referential, or both. According to this principle we can also classify the semantic units ... of a case." Note that Adamec refers to these two functions as denotative (semantic) and presentative (grammatical, or pragmatic/deictic). Indeed, as Flier has pointed out (1988 personal communication) this characteristic is not unique to case: "This notion of an objective structure and a necessary subjective filtering of it has a direct counterpart in the use of procedurals, in which a quantified interpretation produces clear formal and semantic limitations on the surface." Ibid.: 9. Cf. also Cottrell (1985: 57-58) who found the same semantic phenomena to be associated with cases as with polysemous words. Potebnja ([1958]: 486) suggests that the lexical meaning of the word is the only factor which determines whether or not comparison is being invoked when the instrumental is used in predicate position. Both Pit'ha (1971) and Levine (1984 and 1986) conclude that the interpretation of a dative form as indicating possession is dependent on the selection of the lexical item which is in the dative. In conversational Russian, Zemskaja (1983: 96) notes that declension patterns are becoming weaker at the same time that prepositions are being used more and more frequently, and Schupbach's (1986) research reveals a similar trend in technical written Russian. In a personal communication (1987), Adamec confirmed that colloquial Czech also shows an increased use of prepositions vis-a-vis bare case. For an argument against the existence of empty morphemes (specifically prefixes, although largely applicable to morphemes as a whole class), see Janda (1986: 19-26). Cf. also Oliverius (1972: 98): "the notion of empty morphemes is a contradictio in adiectio totally unacceptable in any type of linguistics considering language as a sign-system", and Langacker (1987: 26): "the assumption that a governed morpheme is ipso facto semantically empty is erroneous". This notion is of course neither new nor unique to cognitive semantics. Prud'homme (1981: 11) notes that localist case theory is based on the premise that "humans perceive the world around them with respect to the physical location of its various entities, and the movement of these from one place to another". The choice of elements in this action-chain notation is somewhat arbitrary, but not random. For example, the identification of entities as circles, the use of arrows to symbolize actions, and the left-to-right orientation of the diagrams. The use of these symbols conforms to their conventional use in our society and is therefore wellmotivated. This, however, does not constitute a claim of psychological reality of the schemas. Smith further elaborates on these schemas in his 1987 dissertation, but these details go beyond the scope of the present discussion. But as will be seen in the analysis of the Czech dative, Smith overstates the case somewhat. The dative is always a potential and never an actual agent. Fauconnier has produced a detailed account of the mechanisms responsible for mapping in natural language. Sweetser gives numerous examples of the mappings of English modal verbs and conjunctions onto the speech-act domain. Cf. for example Svedova 1970: 492. Kurylowicz (1960: 134-135) writes that prepositions cannot be said to govern case; a preposition is rather a sous-morpheme [submorpheme] of case. Isaeenko (1965: 98) in a
30.
similar vein stated that when cases are used with prepositions, they are not dependent on the preposition, but are the morphological complementation of the prepositional phrase. Cf. also Belicova (1982: 92), who believes that "the semantic basis for the meaning of the prepositional dative must of course be sought in the meaning of the dative"; Bethin (1983b: 466), who entertains "the possibility that perhaps the occurrence of the preposition is to a certain degree dependent on the appearance of the dative case and that the preposition might be somehow predictable"; Gladney (1986: 132), who doubts that prepositions can really be said to govern case at all and is convinced instead that prepositions add "specificity to the case relationship being expressed but [do] not fundamentally alter it"; Svedova 1982, 1: 476), whose authors assert that "znacenija bespredloznyx padezeji padezejs predlogami sovpadajut" [the meanings of prepositional and non-prepositional case coincide]; Smith (1987: 458), wbo notes parallels "between the occurrence of DAT in the prepositional and clausal realms"; and Rudzka-Ostyn (personal communication: 1988), who writes "given the fact that the same preposition can combine with two or more different cases and vice versa, the notion of government could be replaced with that of semantic compatibility." Obviously the locative case in Slavic would constitute a special problem, since there are no examples of bare usage in the modem languages. Its solution goes beyond the scope of the present volume, although two possibilities can be suggested. Either a single prepositional usage would be found to be central or, barring this, an abstract idealized central concept would need to be posited, which would serve as an "empty center" around which the network would be arranged. Certainly this core concept was once a viable participant in the network and was orphaned when the bare locative went out of use. Majakovskij resurrected this latent concept by using the locative without a preposition in his poetry only a few decades ago. Indeed, the locative must have some meaning, for if it were semanticallyvoid it would have been abandoned.
Chapter two 31. 32. 33.
34. 35.
Cf. Brecht & Levine (1986: 21): "few investigators have sbared Jakobson's optimism that all case use, whether syntactic or semantic, can be unified under a highly restricted correlative set of invariant semantic features". Cf. also Plewes (1977: 64): "the very abstractness of ... invariant definitions weakens or at least attenuates - their descriptive power to the point that nothing really significant is being claimed". Chvany (1986: 125-126) claims that Jakobson's system was not as rigid as most seem to believe. According to her, "subsequent generations have, by and large, lost sight of the question marks built into his brilliantly reductionist model. The reason for this is easy to see. The maximal expansion of the system to eight cases, and its modeling in the metaphoric cube, suddenly made it non-negotiable.... The parts fit together so neatly that to touch any of them would destroy the whole. So, for twenty-five years [since Jakobson's 1958 article], controversies have raged while the cube "in its crystalline elegance" (as van Schooneveld puts it) remained inviolate." A third goal was to restore the status of surface cases as semantic entities of themselves, a status which she felt Fillmore had denied. Cf. the section on Fillmore's "deep" cases below. This approach is in many ways parallel to that of Tesniere (1959), wbo distinguisbed between actants , which were part of syntactic structure and circonstants, which were part
206
36.
207 of the meaning structure of a sentence. This distinction reappears in Sgall, Petr, Eva Hajicova & Jarmila Panevova (1986), with the terms "inner participants" and "free modifications". An important difference, however, is that the object of these authors is not to discuss the status of specific surface cases. As a consequence, the designation of inner participants and free modifications is conducted independent of the case structures of natural languages. Thus the various uses of, for example, the Russian instrumental are scattered about among these two categories with no indication of their interrelatedness. Mel'cuk (1986: 57) accepts both Kurylowicz's distinction and his terminology. He recognizes syntactic case as one "which conveys no meaning of its own, [but] encodes, on the morphological level, a particular syntactic relation", yet states that this" "meaningless" case contributes to the meaning of the sentence, but only indirectly - through its syntactic structure". Thus a semantic component must be recognized even in syntactic
47. 48. 49.
50. 51.
case. 37. 38.
39. 40.
41. 42.
43. 44. 45. 46.
Cf. Kilby (1977: 1): "the solutions offered by transformational grammarians to the problem of case ... are ... found to be surprisingly unrevealing." In Chomsky 1965: 221-222, case receives no more than a comment in a footnote. References to case are more frequent in his recent work (Chomsky 1982, 1986, 1988), but remain restricted to comments on the assignment of case. Both case and its governing elements are considered "semantically empty". For a relevant summary of Chomsky's theory of case, see Brecht - Levine (1986: 23-25). Cf. Bachman (1980: 37): "it apparently was not Chomsky's purpose to develop fresh definitions of the traditional syntactic categories, but rather to develop a practical notational convention." Cf. Jakobson's (1958 [1971]: 182) assertion of the central importance of semantics and his oblique allusion to Chomsky's theory as "naibolee mexanisticeskoj, javstvenno otmirajuscej raznovidnost'ju amerikanskogo distribucionizma" [the most mechanistic and clearly moribund variety of American distributionalism]. After all, it can be said without hyperbole that Jakobson devoted much of his lifetime to studying the correlation between form and meaning. It is hardly surprising that he would find fault with a model which ignored this correlation. Mel'cuk does recognize the existence of semantic case, but his discussion of it is quite independent of his symbolic logic framework and is based on Kurylowicz's distinction. Cf. Kottum (1979: 2) who suggests that "case" denote only nominal inflectional forms and that some other term be used for Fillmore's "cases", and Mel'cuk (1986: 78): "I don't think the term "deep case" itself is entirely felicitous; to me, it is rather a misnomer, and if linguists are to use it, they should use it cautiously." Given that what Fillmore does is to try to discover all the possible relations between entities in a sentence, it is fair to say that his "cases" belong at least as much to the realm of philosophy as to that of linguistics. Plewes (1977: 2). Cf. Wierzbicka (1980: xvii): "the whole idea of "case grammar" rested, to a considerable extent, on the assumption that surface cases cannot be characterized in semantic terms". Cf. also Burston (1977: 39): "we cannot accept a conception of grammar where nonobservationally accessible categories are considered as relevant". Plewes (1977: 48). See, for example, Burston (1977: 40); Belicova (1982: 132); and Dahl (1985: 4-5). It may be that the origin of the quest for quantitative delimitation of the set of cases properly rests with Fillmore's followers and detractors rather than with Fillmore himself. This issue is implied but not overtly stated in Fillmore (1968); Fillmore's non-comniittal posture on this issue is discussed in Fillmore (1977: 70-72).
52. 53. 54.
55.
56. 57. 58. 59.
60. 61.
Or, more specifically, they represent some of the submeanings of surface cases. For more on Fillmore's understanding of semantics and, in particular, prototypes, see Fillmore 1982. See Janda 1986: 55-59 for a discussion of cognitive semantics and the supposed "levels" of language. Note also that by giving his "cases" their own "deep" level, Fillmore violates one of Jakobson's most basic principles, that of the immediate correspondence between form and meaning: "the semantic minimums of a language can be stated only with reference to their formal counterparts and vice versa" (quoted in Sangster [1982: 48]). This is why it appears that Fillmore neglects surface case and why the above-mentioned objections have been raised. Localism uses features, rather than constructs from Chomsky or Fillmore (which serve other purposes in this theory), to describe case semantics; thus it is the feature system that is relevant to the present discussion. Dahl (1985: 19) also mentions the similarity in basic intent of localism and cognitive semantics and suggests that "the prototype approach [i.e., cognitive semantics] seems to be a possible way of making sense of localism without accepting its oddities." For example, Anderson (1971), Kilby (1977) and Freidhof (1978). See, for example, Potebnja ([1958]: 431-504), who makes several statements about relationships between the meanings which he lists for the Russian instrumental. Staniseva (1958c, a contributor to the Bernstejn volume, an inventory of meanings of the Russian instrumental in which each author wrote on one or more meanings) noted that the tvoritel'nyj sovokupnosti [instrumental of group or measure] is related to the instrumentals of manner and comparison as well as to the comitative instrumental. It must be admitted that the drawing of such figures is in some sense an abuse of Mrazek's and Wierzbicka's work, since they never intended this. It remains a fact, however, that their indications of the interrelatedness of meanings is complete enough to make the deduction of entire networks possible. Neither of the networks presented here is the same as the network that will be given for the instrumental in chapter 4, although they share some essential similarities with it. Others use this concept as well. Cf. Kurytowicz's (1960) designation of primary and secondary functions for cases. Belicova (1982: 30) states that each case must be a class of relations with its own hierarchy. Thus Jakobson avoided the paradox of integrating unity and diversity inherent in any model which uses invariants. See Janda (1986: 38-42). The awkwardness of this concept probably accounts for the fact that many structuralists did not incorporate it in their own work and thus treated features as absolute, rather than relative, entities. Cf. Langacker's billiard-ball and stage models. Human perception of reality is recognized by cognitive grammarians as playing a central role in categorization. It is our perception of reality which de facto is reality as far as cognitive semantics is concerned, for we do not, and in fact cannot, have access to any other, more privileged, more objective perspective of reality. Cf. Lakoff (1987: 238 fO. For example, Mrazek (1964), Wierzbicka (1980 and 1986). E.g., Kurylowicz (1960).
208
209
Chapterthree 62.
63. 64. 65.
66.
67.
68. 69. 70.
71.
72.
73.
According to Nikonov (1961: 16), who uses written sources, the dative and the instrumental each account for about 7% of overall case usage in Russian. In transcripts of actual conversation Zemskaja (1983: 139) found that the dative and instrumental accounted for only 4.1% and 5% of case usage, respectively. I do not have figures for Czech, but a survey of the most recent reverse dictionary (Tesitelova et al. 1986), which includes frequencies for forms of individual nouns, indicates that the dative and instrumental have frequencies similar to those found in Russian. For Russian, see Bernstejn (1958), Mrazek (1964), Prud'homme (1981) and Wierzbicka (1980). For Czech, see Ulicny (1984). Although the Czech dative figures prominently in a number of the works cited herein, I can name only one article dedicated exclusively to it: Pit'ha (1971). The so-called "free dative" is indeed more frequent and variegated in Czech than in Russian, as will be shown below. It is interesting to note that the only case other than the Russian instrumental to which Wierzbicka (1986) devotes a whole article is also a West Slavic dative (in this case, Polish). The prose notation used here is inspired by Wierzbicka's metalanguage. The elements listed in Figure 8 and used as building blocks in the analysis are abbreviations of phrases which could further be fully articulated into statements containing only the universal semantic primitives identified by Wierzbicka (1980b: 10). This notion is not universally held. Wierzbicka (1986: 386) gives the "core meaning" of the Polish dative as "the case used to designate the recipient in sentences of GNING". Vinogradov (1972: 140) agrees and reminds us that the very word "dative" is derived from the (Latin) verb meaning "give". Kurylowicz, however, lists this among the secondary functions of the dative, asserting that its meaning pour [for] in prepositional phrases is primary. Lakoff (1982: 48-49): "the categories used in ICMsare not objective, but rather based on experiential aspects of human psychology. The ICMs are understood via such a basis in human experience." I am indebted to aldrich Leska of the Institute for Czech Language in Prague for this starmap metaphor. Wierzbicka (1986: 396-397) shows that animals make awkward dative referents in some types of sentences. She cites that the Polish sentence Pomalowalam bude Rexowi "I painted Rex's kennelfor him" "sounds odd because it implies that Rex (a dog) wished for the kennel to be painted". Greenberg's (1974: 25) work supports this claim, for he found that nouns referring to animals are much less frequent than the norm for the dative case. Note that it is also possible to use the prepositional phrase do knihovny "(in)to the library" here in place of the dative, an operation which is not permissible when the dative refers to a person. This use, however, invokes a different meaning of the verb ddt which would have to be translated as "put". Note that the same is true for ddt, for which the immediate recipient (indirect object) must also be identified. mi dala darek pro tebe. Teta Aunt-NOM me-OAT gave gift-ACC for you-ACC 'Auntie gave me a gift for you.' Here me is the indirect object, and the referent of you has not yet received the gift. Lists of dative-governing verbs appear in section 3.3. Verbs of communication will be discussed with other semantic extensions of the indirect object.
74.
75. 76. 77. 78.
79.
80.
81.
(37b) could be rendered grammatical if provided with a context in which a specific referent for the elided indirect object had already been identified. Thus it could appear in the following discourse: Davidovi k Co jste daly what-Ace you-AUX gave David-OAT to narozenindm? birthday-OAT What did you give David for his birthday?' Ja jsem mu dala knihu, Hana I-NOM am-AUX him-OAT gave book-ACC Hana-NOM dala kravatu a Ludmila dala kytku. gave tie-sec and Ludmila-NOM gave flower-Ace 'I gave him a book, Hana gave a tie, and Ludmila gave a flower.' The examples which are here contrasted with ddt do not, however, require this kind of contextual support. In the remaining cases, an indirect object has not been specified and then elided, rather it has not been specified at all. Although the dative is optional in (33) as well, verbs of communication bridge the gap between the indirect object and datives governed by verbs like volat "call", as has been noted above, and therefore represent a function quite different from that of the free dative. This is my translation of the Czech term dativ volny , used by Smilauer (1966) and Grepl - Karlik (1986). Reflexive verbs (those followed by se) appear in parentheses because they can be said to have a direct object (at least in a historical sense). Clearly, however, they are a systematic part of this phenomenon. Note especially the relation of ptisvedcovat, lichotit, pochlebovat, ptitakdvat, salutovat, ptikyvovat to subordination; the relation of poroucet, natizovat to domination; the relation of dekovat, piizvukovat, prot, blahoptdt, poroucet se, zehnat, gratulovat, kondolovat, radit, tleskat, promijet, ptipominat and the verbs denoting giving of money and gifts to verbs of benefit; and the relation of sptlat, naddvat, zloiecit and the verbs denoting giving of punishment to verbs of harm. This verb can take either an accusative direct object or a dative complement to describe the person called, depending upon the speaker's intended meaning. If this verb is used to mean "call someone to come, or to get his attention", the person called appears in the accusative case. If, however, the calling involves some sort of message, the person called is in the dative, since there is already an inherent direct object. Virtually all scholars who have described verbal government in Czech and Russian have devised classifications of groups of verbs according to semantic features. The following works pertaining to verbal government of the dative can be cited: Grepl - Karlik (1986), Belicova (1982), Plewes (1977), Svedova (1970 and 1982), Andersen 1 (970,) Smilauer (1966). Of these, only Andersen (1970) presents a principled approach rather than an unconnected list of headings for groups of verbs, cf. below. Plewes (1977) sets out to find significant groupings while relating their meanings to lakobson's features, but does not accomplish this. The symmetrical relationship is observed when the strengths of the nominative and dative referents are equally matched. Subordination is present when the nominative is weaker than the dative, thus subjecting itself to the dative's control. This relationship is called transitive because of its capacity for establishing lines of hierarchy. Thus if A is subordinate to B and B is subordinate to C, A is necessarily subordinate to C. Andersen presents only these two relationships because they are sufficient to accommodate the data provided by Russian dative-governing verbs, to which his article is limited. There is a
210
82.
83. 84. 85.
86.
87.
88.
89. 90.
211 third possibility, i.e., the converse of subordination - domination - the inclusion of which would not threaten the basic tenets of Andersen's classification, since it is also a transitive implicating relationship. Although Russian ignores this possibility (it uses the instrumental for this function, as discussed in chapters 4 and 5), Czech utilizes it. Although for some verbs (those which can appear with or without se) this particle functions like a direct object, its salience in this role has faded significantly here. Almost all dative-governing reflexive verbs either do not have corresponding nonreflexive forms or require the reflexive particle to express the meanings which are in force when they govern the dative. Thus se has become part of the anatomy of the verb itself (marking intransitivity) rather than remaining part of its valence pattern. Some reflexive verbs which governed the dative up until recently have even taken this trend one step farther and are actually becoming ordinary transitive verbs. Witness the verb ucitse "learn, study" (which requires se in this meaning), which in contemporary Czech has accusative complements. Even educated native speakers (among the younger generation) consider the use of the dative with this verb archaic. Cf. Smilauer (1966,) Plewes (1977), and Grepl- Karlik (1986). The last verb, naslouchat, may seem a bit out of place here until one reflects on the considerable power that is wielded in situations where this sort of practice is common. As noted above, Popova (1974: 178) claims that this is so for dative-governing verbs in Russian, but then herself goes on to show that deviant uses of the dative with verbs that do not normally govern the dative are the result of contamination from dative-governing verbs having similar meanings. Both her conclusions and mine prove that her basic premise must be incorrect. The frequency of k , which is 3202 if one counts bare uses, or 3394 if one includes uses in secondary prepositions (frequency figures are from Tesitelova, Marie, Jan Petr & Jan Kralik (1986) and are based on a corpus of 540,000 word tokens), is an order of magnitude greater than the next contender, proti "against", with a frequency of 397. The frequency of all other dative-governing prepositions is below fifty. K is also the only dativegoverning preposition listed as being a primary preposition by Tesitelova, Marie, Jan Petr & Jan Kralik and it is the only preposition that participates in the formation of secondary prepositions. Cf. Gladney (1986: 139), who reviews all the dative-governing preposition formations of Russian and concludes that the only true prepositions are k and po. Since po governs the locative case in Czech, k remains as the dative preposition par excellence for that language. Bethin's (1983b) work is on Russian, but the use of k in Czech is not significantly different from its use in Russian. Note that the meaning of other instantiations is arrived at by means of metaphorical extension from concrete meanings. The use of either do or na with a human destination would be so concrete as to be absurd. Do would imply that the person was sliced open, and na would indicate that the subject stepped upon or draped himself around the object of the preposition. Note, however, that this construal is language-specific. It is probably the lack of externalized potential agency that has prompted Belicova (1982; 42) to state that dative subjects are characteristically "inactive". Bachman (1980) presents a very thorough and convincing case for the subjecthood of such datives in Russian, showing that they pass most of the tests which nominative subjects do. According to Bachman, dative subjects are animate, refer to the discourse topic, and mark the experiencer or logical subject of a situation. He identifies three subtypes: 1. category-of-state sentences (Mne xolodno "I am cold' '), 2. infmitive sentences (Nam esce nazadexat' "We still (have to) go back"), and 3. sentences with the reflexive particle -sja (Ptice ne letitsja "The bird (can)not fly"). His classification is convenient for Czech as
91. 92.
93.
94.
95. 96.
97. 98. 99. 100. 101. 102.
103. 104. 105. 106.
well, and I have adopted it with a few modifications. Miller (1986: 307) discusses the first two types of dative constructions (ignoring the third one) and concludes that they are semantically identical. He notes that in both instances a situation has come to an individual. Wierzbicka (1986: 404) likewise suggests that the dative can encode the idea of a "personal sphere". The notion of affectedness represents the dative's total semantic contribution to such expressions. The schemas are therefore very accurate in that they caption only the information relevant to case meaning. Pit'ha (1971: 310) correctly asserts that the dative is not the bearer of the possessive meaning in such constructions. Kottum (1979: 125) observes that the dative of possession in Polish is not entirely synonymous with the use of possessives. In a similar vein, Levine (1984: 495) notes that "sentences with the dative are more informative than the corresponding sentences with the genitive" in Russian. Compare this combined use of possessive and dative forms with English expressions such as his house burned down on him. Note that both (68) and (69) actually belong to schema 2b, but since they are examples of the dative of possession which have significant frequencies, they are included in the discussion here. The sense of the interchangeability of the two kinds of expression is eliminated in English, where these dative constructions do not exist. Since two possessives cannot modify a single noun, only the benefactive reading can be retained in the translation. Levine (1984: 498) found this to be true for Russian as well: "The presence of the modifier characterizes the body part more explicitly and individuates it with 'respect to the possessor, and more generally, to the event." Pit'ha undated: 131: "The notion of dativus possessivus used in grammars for generations was introduced from the domain of extralinguistic reality on the basis of an observation of the content of certain utterances. This notion has never been syntactically defmed." These examples are adapted from sentences cited in Pit'ha 1971. Grepl and Karlik (1986: 104) compare this usage with that of the vocative, and note that in both cases an addressee is required. Note the use of the first-person plural (plural majesticus ) by the singular speaker. This further enhances the authoritative tone, for the policeman thereby implies that he is the representative of a larger body (i.e., the state). I am indebted to Petr Pit'ha for pointing out the existence of this variation on the emotional dative. Note, however, that poslat "send" fails one of the tests for true indirect objects, i.e., it is possible to omit the dative object in a sentence with this verb without rendering it . ungrammatical. This example was given to me by Petr Pit'ha in a personal communication. Like (82), it would be uttered by a mother comforting her child. This example does not mean that' 'the vacuum cleaner is nice to you" . In order to express this idea a prepositional phrase na tebe would be used. E.g. Havranek and Jedlicka, Smilauer, Grepl and Karlik. The meanings of schemas 1, la, and Ib are largely dependent upon the structure of the action chain. The dative in schemas Ia and Ib derives meaning from comparisons of it with other participants, and these comparisons are invoked by the syntactic structures of these constructions. Schema Ia compares the dative with an accusative, the canonical object. Schema Ib compares the dative with a nominative, the canonical (logical) subject. The meaning of schema 2 is defined by the presence of a dative sphere, and thus more independent of the structure of the action chain.
212 107. 108.
109.
110.
111.
112.
113. 114. 115.
116.
117.
118.
119.
213 This statement is not based on actual statistics, but rather on the impressions of learners of Czech. Tesitelova, Marie, Jan Petr & Jan Kralik (1986) do not give a frequency count for si . Cf. Sova (1962: 139, 439); Heim (1982: 56); and Naughton (1987: 20). A happy exception is Townsend (1981: 166), who mentions the use of si as a "stylistic feature" of spoken Czech, and tells us that it "may ... lend an expressive or modal meaning". Kralfkova (1977) has done a partial inventory of verbs which commonly have listings with si in dictionaries. She found thirty-one verbs for which si 's role is analogous to that of the indirect object. In every case, si can be replaced by sobi in emphatic sentences. Not all of her examples, however, illustrate schema Refll: some are captioned by Refl la, some by Refl2, and yet others occupy a transitional position between these two schemas. Kralfkova's sole criterion for sorting verbs concerned the possible replacement of si by sobe or the reflexive possessive svitJ . She did not take semantics into account, which explains the discrepancies between her classification and mine. It is surprising that ddt si does not appear anywhere in Kralfkova's lists. Given the high frequency of this collocation, this could hardly be an oversight. This omission was probably motivated by the fact that ddt si is considered idiomatic in Czech, and idioms are often avoided in linguistic studies. In the current framework, however, the semantics of idioms is considered an important and integral part of a cognitive category. Examples of this type (along with vylomil si zub "he knocked out his (own) tooth", which in this classification would belong among the possessive uses of schema Refl2), are characterized by Havranek and Jedlicka (1960) as "volnydativ s obecnym vyznamem prospechu" [free dative with the general meaning of benefit]. The notion that reflexivization might involve something more than merely identifiying the subject and object with the same participant is neither new nor unique to the present work. Cf. Maldonado (1987: 9), who concludes that "a different strategy from the classical subject/object identification is necessary to explain the formation of reflexives in Spanish." I am indebted to Petr Pit'ha (personal communication, 1987) for this example. There are verbs that use si to indicate the evaluation of something relative to one's own opinions but require a direct object: cenitsi "value" and vdzit si "respect'" To my knowledge, this table contains all obligatory uses of si with intransitive verbs; uses that are not examples of indirect object or dative of possession. Stezovat si "complain" and the verbs of awareness do take objects, but they appear in the genitive case or in prepositional phrases rather than as accusatives. Stoupnout si is not on a par with sednout si and lehnout si, for it is not the neutral verb for' 'stand up" (which is vstdt ); its use is limited to situations in which a standing posture is most convenient. Many of the verbs in the table have been culled from Tesitelova, Marie, Jan Petr & Jan Kralik (1986); information given therein on obligatoriness of si was checked against the intuitions of native speakers. Si is now required when this verb is used in its original meaning of "rest", although hovet also appears without si in the meaning "give in to'" however, this meaning is listed in etymological dictionaries as an innovation. Both zatancit si and popovidat si represent productive types of verbs. the prefix zacombines with si and simplex verbs that denote physical exercise, and the prefix pocombines with si and verbs which denote talking. This participation of si in wordformation gives further evidence of its grammaticization. The meaning of si in the two verbs hdzet si and kopat si is related to its meaning with the verb hrdt si "play". Both of these verbs exist without si, but have different meanings; hdzet means simply' 'throw" and kopat means nothing more than "kick".
120.
121. 122. 123.
This generalization requires the following qualification. The dative reflexive appears in Slovak in both lahnut' si "lie down" and sadnut' si "sit down", and in Bulgarian it appears with the verb legna si "lie down". In the remaining Slavic languages it is not associated with these verbs. The use of si with verbs of comfort is sporadic and isolated in Slovak and Bulgarian, whereas in Czech it is systematically applied. The obligatory use of si with the verbs in the table is an innovation in Czech. Gebauer (1958) gives attestations of odpocinout si from the last quarter of the nineteenth century; at that time, lehnout still appeared without si. Czech grammars from the tum of the century list lehnout and sednout with and without si . A less complete description of the status of pragmatic si in Czech was published as part of Janda (1988b: 196-197). Linguists at Charles University offered me this example and expressed their dismay at the unseemly proliferation of si in metro announcements. The projection of boundaries from the real-world domain onto the speech-act domain is consistent with Fauconnier's (1985) theory of mental spaces. Note that Sweetser (1984) has discovered similar projection of real-world relations onto the speech-act domain in the semantics of English modals and conjunctions.
Chapterfour 124. 125.
126. 127. 128. 129.
130. 131.
132. 133.
134. 135.
This finding does not, of course, preclude the possibility that there might exist unrelated languages with vastly different case category structures. This gap is probably attributable to the lack of extension of the Russian indirect object to taking verbs, since the verbs of removal from subordinating relationship naturally overlap with intransitive taking verbs. Without the taking verbs and intransitive taking verbs, there is not enough support in the network for this group of verbs. Information on which adjectives admit prepositions is taken from Svedova (1982,) VoL 2. In standard Russian blagodarit' "thank" governs the accusative, but, as noted above, the dative often accompanies this verb in colloquial usage. Poslusat'sja ' 'obey" governs the genitive rather than the dative. Po exists in Czech, but it differs from the Russian preposition in both meaning and government. In Czech po means "after" and governs the locative case. Po + loc meaning "after" does exist in Russian, but its use is very limited. Cf. Flier's (1975) description of po as a relatively unmarked preposition/prefix that displays a wide range of connotational variants. Note that in Russian the verb sledovat'''follow, comply" is dative-governing. Czech, which lacks dative-governing po also lacks dative government for its corresponding verb sledovat, which can only take accusative objects. If the noun phrase governed by po is plural the accusative, rather than the dative, is used. Cf. Levine's (1986) detailed study of the relationship between inalienable possession and the dative of possession. He notes that (164) entails that "the briefcase is like an appendage of the possessor". If, however, the possessed object is modified, alternate means of expressing possession (genitive case, possessives) must be used, as is the case with Czech, cf. Levine (1984: 498) and section 3.2.2. of this volume. Note that for the purposes of cognitive description, constructions with predicate adjectives are equivalent to constructions with intransitive verbs, as described in section 5.1.1. This usage does, however, enjoy some popularity in colloquial speech.
215
214 136. Czech has borrowed muset "must" from German (cf. mussen ), and has broadened its uses of sme: "be allowed to" to approximate German durfen. The modal use of the verb mit "have" in expressions like: Mel by to vedet "He ought to know that (lit: (He) had (conditional) that to know)" may have developed under the influence of German modals as well, for all German modals use haben "have" as their auxiliary.
146. 147.
Chapter five 137. Note that this network is also compatible with characterizations of the instrumental put forth by other scholars. Cf. Burston (1977: 181): "when the speaker uses an instrumental ending, he presents the entity referred to by the substantive stem as playing a role which is not essential in the narrated situation"; and van Schooneveld (1978: 208), who terms the instrumental "the case of accompanying circumstance". 138. Cf. Plewes (1977: 255-256): "...I submit that there is not [a shared semantic trait] and that the feature [+peripheral] is a residual feature that appears by default when some semantic feature is needed but none of the others fits". Plewes is discussing instrumental verbal government in particular, but his challenge of Jakobson's feature specification stands. Cf. also Kilby's (1986: 331-332) question, which is fueled by the same motives: "Is this use of the term "peripheral" merely a metaphorical use which gives the impression of consensus without in fact representing anything substantive in the way of agreement, or is it a genuine unifying factor?" A skeptical view of the validity of this feature likewise inspired the condemnation of Jakobson's system as "too abstract" by Skalicka (1950), Isacenko (1965) and others. 139. Note that Greenberg (1974) has stressed the role of lexical meaning in case meaning, and Kilby (1986: 330) has pointed out that this phenomenon is particularly important for the instrumental: "The crucial point about the instrumental, I suggested, was that the classes of nouns, and the classes of governing elements which occurred with the instrumental were higly specific, and therefore it is possible to fit a large number of different uses into the same case... For example, consider the instrumental of instrument, location, time; personal characteristics, etc., which could not be confused with each other because of the virtual complementarity of the noun phrases which occupy these positions." 140. As conceived of by Langacker (1987); cf. Figure 1. 141. Cf. Staniseva, (1958b); Mrazek (1964); and Wierzbicka (1980a). 142. Although the example brosat' kamnjami is frequently cited in the literature, cf. Jakobson (1936 [1971]), Mrazek (1964), Wierzbicka (1980a), its use is considered old-fashioned by contemporary speakers, who prefer the reflexive form brosat'sja kamnjami. 143. These examples cited in Leska, Substantivni determinace . In fact, all meanings of the Russian instrumental are instantiated adnominally, cf. Gadolina (1958b). Because adnominal usage is common to the entire network and parallels usage in full clauses, a systematic explanation of this phenomenon will not be presented. 144. Among those who distinguish an instrumental of manner are Staniseva (1958b), Mrazek (1964), Wierzbicka (1980a). 145. Staniseva (1958b) likewise refers to the instrumental of manner as an abstraction of the instrumental of instrument; Schlesinger (1979) affirms that the instrumental of manner and instrument form a continuum; and Mrazek (1964) places these two submeanings in close proximity.
148. 149.
Wierzbicka (1980a: 22) adds a "semantic antipassive" (for instrumental objects) to this list of voices associated with the instrumental. For a recent survey of literature on voice in Russian, see Toops (1987). Compare (205) and (206) with the following true passives, the first with a reflexive verb and the second with a participle: Stol pokryvaetsja skatert'ju. covers-REFL tablecloth-INST Table-NOM 'The table is covered with a tablecloth.' Zemlja pokryt snegom. Ground-NOM covered-PASTPASSPART snow-INST 'The ground is covered with snow.' Xodova (1958a: 128) insists that the driver in this example can only be construed as an instrument, an extreme and controversial interpretation. Note that there is a small group of related constructions which employs (mostly deadjectival) nouns in predicate position and requires a slight modification of the schema (to schema 1a On velikan rostom. He-NOM giant-NOM stature-INST 'He is a giant in stature.' Bogatyr' ty budes' s vidu i Bogatyr'-NOM YOU-NOM will-be from appearance-oms and kazak duso]. (Lermontov; in Gadolina 1958a) Cossack-NOM soul-INST 'You will be a bogatyr' in appearance, and a Cossack in soul.' This becomes possible, but not necessary. In the following example the instrument is not a part of the nominative entity (albeit this sort of example is strictly literary): Oni bogaty grabitel'stvom. They-NOM rich-NOM thievery-INST 'They are rich through thievery.' This use of the instrumental has been extended to many other kinds of constructions, in which we see the basic constructionreduced and embedded in another clause, as in: Xlopocut nabirat' ucitelej polki t
) .
150.
151.
They-are-busy
to-gather
Cislom
pobolee,
teachers-am cenoju
regiments-NOM
podesevle.
greater price-INST cheaper (Griboedov; in Gadolina 1958a) 'They busy themselves trying to gather regiments of teachers. Greater in number and cheaper in price.' Alternatively, the comparative may appear as an adverb, as in: Marija Ivanovna zivet etatom vyse, Maria Ivanovna-NOM lives floor-INST higher 'Maria Ivanovna lives one floor up.' 152. See Gibson (1984: 26), who states that the entity marked as an instrumental object "is clearly the means used to bring about the motion specified by the verb", Mrazek (1964; cf. also Figure 6 this volume) and Wierzbicka (1980a); cf. also Figure 7, this volume, who identify the instrumental object as bearing the closest possible relationship to the instrumental of instrument. Belicova (1982: 70) likewise states that the instrumental object is a means for the realization of the action. 153. This phenomenon has received considerable attention in the literature already. For more details, the reader is referred to Potebnja [1958], Jakobson 1936 [1971], Wierzbicka Number-INST
217
216
154.
155. 156. 157. 158.
159.
160.
161. 162. 163. 164.
165.
166.
167.
(1980a), Gibson (1984). Both Konecna (1969: 347) and Plewes (1977: 237) describe the semantics of this type of instrumental using the word manipulate. Plewes (1977: 237) considers instrumental object constructions as a subtype among governed instrumentals, and Wierzbicka (1980a: 161) suggests a semantic formula similar to that used for instrumental of instrument. Compiled from lists cited in Plewes (1977) and Fowler (1987). Cf. Fowler (1987: 13) who also notes that these verbs resist passivization. Wierzbicka (1980a: 161) has argued that governed instrumental arguments invariably represent people, but this is only true of the first subcategory. It might be added that the constructioncaptioned by schema Ib also has a decidedly passive "feel" to it, but it is difftcult to determine whether this is a product of the features listed above or a subjective opinion influenced by the fact that this construction is translated by English passives. Note the structural identity of the accusative of time and the accusative of measure. Prefixes which describe motion through space are often utilized to describe similar paths through time. The pervasiveness of this phenomenon is apparent in the proverb zizn' profit' ne pole perejti [Life is no picnic (lit: "Living through life is not (like) walking across a field)"], in which a stretch of time (life) is compared to and expanse of space (a field). Potebnja ([1958]) and Malaxovskaja (1958) consider the instrumental of time an extension of the instrumental of space. Mrazek (1964) and Wierzbicka (1980a) speak of the instrumental of time as being closely related to the instrumental of space. Jakobson (1936 [1971]) places the instrumentals of time and space side-by-side, and Isacenko (1965: 100) implies that both uses of the instrumental should be subsumed under the category "adverbial usage of case". The term "focus" as used here correlates most frequently with "topic" or "theme", but in transitive sentences it is identified with the "comment" or "rheme". It can be defined as the participant most affected by the action of the verb. Further evidence of such distinctions in Russian is presented in Janda (1986: 71-77) and Janda (1988a: 335-339). Cf. Rothstein (1986: 313) who states that "in copular sentences, nominative implies equation, whereas instrumental implies ascription". Note the overlap between the attributive instrumental and instrumental of setting here. This instrumental is also defming a certain period of time and could be translated: "while she's a girl..." . As proof of the fact that these are two kinds of tautological instrumental and are formed differently, note that in the first example the nominative phrase zver' "beast" is omissible, whereas in the second example only the instrumental phrase vodoj "water" can be omitted. This suggests that the first example is created by adding the nominative noun phrase, whereas the second is formed with the addition of the instrumental. Cf. Gladney (1986: 130) who traces the semantic value of prepositions from pre-Slavic times through Russian: "in construction with a noun this [indoeuropean] preposition added specificity to the case relationship being expressed but did not fundamentally alter it. This original state can still be discerned in modern Russian." Cf. also Staniseva (l958a: 41) who notes that historically the use of a preposition was optional in expressing the comitative instrumental. Cf. Schlesinger (1979: 308) who states that "the instrumental [of instrument] and comitative are really only two extreme points on what is a conceptual continuum", thus suggesting that instrument and manner form a similar continuum (Schlesinger [1979:
168. 169.
170.
316]), and entertains the possibility that these continua may form a system (Schlesinger [1979: 321]). The use of metdu with a plural which refers to more than two objects is now archaic. The preposition sredi+ genitive' 'among" now serves this purpose. Note that prepositional uses are usually left out of discussions of the Russian instrumental. They do not figure into the works of Mrazek (1964), Wierzbicka (1980a), and Prud'homme (1981) at all. Historically, however, 1 and 4 were originally collapsed into a single central prototype, cf. Potebnja [1958], Mrazek (1964), Staniseva (1958a). Neither the comitative instrumental nor the instrumental of setting (at least in spatial uses) required a preposition. As formal distinction between the conduit and comitative instrumentals became established, the network unfolded and the former (schema 1) came to dominate it.
Chapters six 171.
172.
Note that the first three verbs would fall under the second grouping used to classify Russian verbs, that entitled "possess", and the fourth verb belongs to the third group, that entitled "evaluate". Quantitative uses of the attributive instrumental are missing from the Czech paradigm. Comparative uses of the instrumental are bookish and artificial; ct. Smilauer (1972: 282), who gives this example (which is stylistically marked as belonging exclusively to the written language by the words nyn; "now" andjit "'lready''': Kone leteli nynf ne jiz g{pem, a le Horses-NOM flew now not already arrow-INST but
bleskem.
173.
lightning-INST 'Now the horses flew, not likean arrow, but like lightning.' Grepl and Karlik (1986: 295) consider words such as bleskem to be fully adverbialized in the modern standard language (implying in this case that the meaning is closer to "very quickly" than to "like lightning"). These examples are taken from Ulicny (1984), who notes that the use of the instrumental is more subjective than the use of the nominative.
Chapter seven 174.
Cf. Jakobson and Pomorska (1988) for discussions of this phenomenon.
219
References Adamec, Premysl
1978
Obrazovanie predlozenij iz propozicij v sovremennom russkomjazyke (Acta Universitatis Carolinae, PhilologicaMonographia LXIX-78). Prague: Universita Karlova. Andersen, Henning 1970 "The dative of subordination in Baltic and Slavic", in: Thomas F. Magner & William R. Schmalstieg (eds.), Baltic Linguistics. University Park: Pennsylvania State University Press, pp. 1-9. Anderson, John M. 1971 Thegrammar0/ case. Cambridge: Cambridge University Press. Apresjan, Jurij. D. "0 sil'nom i slabom upravlenii: Opyt kolicestvennogo analiza", Voprosy 1964 jazykoznanija 13: 32-49. 1973 "A description of semantics by means of syntax", Linguistics 96: 5-32. Babby, Leonard H. 1986 "The locus of case assignment and the direction of percolation: Case theory and Russian", in: Richard D. Brecht & James S. Levine (eds.), Case in Slavic. Columbus: Slavica, pp. 170-219 .1987 "Case, prequantifiers, and discontinuous agreement in Russian", Natural Language and Linguistic Theory 5: 91-138. Bach, Emmon & Robert T. Harms (eds.) 1968 Universals in linguistic theory. New York: Holt, Reinhart & Winston. Bachman, Ronald David 1980 The subject potential of the dative case in modem Russian. [Unpublished Ph.D. Dissertation, Ohio State University.] Belicova, Helena 1982 Semantickd struktura vet» a kategorie padu. Prague: Academia. Bernstein, Samuil B. (ed.) 1958 Tvoritel'nyj padez v slavjanskix jazykax. Moscow: Izdatel'stvo Akademii Nauk SSSR.
Bethin, Christina Y. 1983a "The meaning of po + dative case in spatial sentences", Russian Language Journal 37: 27-34. 1983b "On the spatial dative in Russian", Slavic and East European Journal 27: 465-477. Bloomfield, Leonard 1966 Language. New York: Holt, Rinehart and Winston. Brecht, Richard D. & James S. Levine, 1986 "Case and meaning", in: Richard D. Brecht & James Levine (eds.), Case in Slavic. Columbus: Slavica, pp. 17-35. Brecht, Richard D. & James S. Levine (eds.) 1986 Case in Slavic. Columbus: Slavica. Brugman, Claudia 1981 Story of OVER. [Unpublished M.A Thesis, University of California at Berkeley.] Bulygina, AM. 1958a "Tvoritel'nyj mesta bespredloznyj", in: Samuil B. Bernstejn (ed.), Tvoritel'nyjpadet v slavjanskixjazykax. Moscow: Izdatel'stvo Akademii Nauk SSSR, pp. 246-265.
"Tvo~tel'n~j ~esta s predlogami", in: Samuil B. Bernstejn (ed.), Tvoritel'nyjpadez v slavjanskix jazykax. Moscow: Izdatel'stvo Akademii Nauk SSSR, pp. 266-287. Burston, Monique A 1977 A semantic analysis of the Pali case system. [Unpublished Ph.D. Dissertation, Cornell University.] Chomsky, Noam 1965 Aspects ofthe theory 0/ syntax. Cambridge: MIT Press. 1982 S~me concepts and consequences 0/ the theory 0/ government and binding. Cambridge: MIT Press. 1986 Knowledge oflanguage: Its nature, originand use. New York: Prager. 1988 Language andproblemsofknowledge. Cambridge and London: MIT Press. Churchland, Patricia 1986 Neurophilosophy. Cambridge: MIT Press. Chvany, Catherine 1986 "Jakobson's fourth and fifth dimensions", in: Richard D. Brecht & James S. Levine (eds.), Case in Slavic. Cloumbus: Slavica, pp. 107-129. Cole, Peter & Jerrold M. Sadock, (eds.) 1977 Grammatical relations. (Syntax and Semantics 8). New York: Academic Press. Cottrell, Garrison Weeks 1985 A Connectionist approach to word sense disambiguation. [Unpublished Ph.D. .. Dissertation, University of Rochester.] Dahl, Osten 1985 Casegrammar and prototypes. Duisburg: L.AU.D. Deane, Paul in press "Polysemy and cognition", Lingua. Deserieva, Tamara I. 1974 Stuktura semanticeskix polej cecenskix i russkixpadetej. Moscow: Nauka. 1977 "Nekotorye problemy grammaticeskoj semantiki v svjazi s osobennostjami formalizacii v estestvennyx jazykax", Voprosy jazykoznanija 4: 57-60. Dreyfus, Hubert L. 1979 What computers can't do. New York: Harper & Row. El'zbutas, Ju. 1967 "Upotreblenie datel'nogo padeza v sovremennom russkom jazyke", Russkij jazyk v Skole 28: 85-90. Fauconnier, Gilles 1985 Mental spaces: Aspects of'meaning construction in natural languages. Cambridge: MIT Press. Feldman, Jerome A, Dana H. Ballard, Christopher M. Brown & Gary S. Dell (eds.) 1985 Ro.chester connectionistpapers: 1979-1985. Rochester: Departments of Computer SCience and Psychology at the University of Rochester. Fillmore, Charles J. 1968 "The case for case", in: Emmon Bach & Robert T. Harms (eds.), Universals in Linguistic Theory. New York: Reinhart & Winston, pp. 1-88. 1977 "The case for case reopened", in: Peter Cole & Jerrold M. Sadock (eds.), GrammaticalRelations. Syntax & Semantics 8. New York: Academic Press, pp. 59-81. 1982 "Frame semantics", in: Linguistic Society of Korea (eds.), Linguistics in the morning calm. Seoul: Hanshin, pp: 111-138. Fowler, George 1987 Oblique passivization in Russian. [Unpublished MS]
1958b
221
220 Freidhof, Gerd
1978
Kasusgrammatik und lokalerAusdruck im Russischen. (Specimina Philologiae
Slavicae 15; Beitrage zur Kasusgrammatik der slawischen Sprachen 1) Munich: Sagner. Gadolina, M.A. 1958a "Tvoritel'nyj ogranieenija", in: Samuil B. Bernstejn (ed.), Tvoritel'nyj padet v slavjanskixjazykax. Moscow: Izdatel'stvo Akademii Nauk SSSR, pp. 201-221. 1958b "Tvoritel'nyj priimennoj", in: Samuil B. Bernstejn (ed.), Tvoritel'nyj padet v slavjanskixjazykax. Moscow: Izdatel'stvo Akademii Nauk SSSR, pp. 289-312. Gebauer, Jan 1958 Historickd mluvnice jazyka eeskeho. Vol. III: Tvaroslovi, Part II: Casovanf. Prague: Nakladatelstvi CSAV. Geeraerts, Dirk 1987 "Cognitive grammar and the history oflexical semantics", in: Brygida Rudzka-Ostyn (ed.), Topics in Cognitive Linguistics. Amsterdam: Benjamins, pp. 647-678. Gibson, Margaret I. 1984 "Some changes in the use of instrumental case", Russian Language Journal 38: 2535. Gladney, Frank Y. 1986 "Prepositions and case government in Russian", in: Richard D. Brecht & James S. Levine (eds.), 130-151.
Gorbacevic,Kirill Sergeevic
1971 Izmenenie normrusskogo literaturnogo jazyka. Leningrad: Prosvescenie. Greenberg, Joseph H. 1974 "The relation of frequency to semantic feature in a case language (Russian)", Working Papers on Language Universals 16: 21-45. Grepl, Miroslav & Petr Karlik 1986 Skladba spisovne ceStiny. Prague: Statni pedagogicke nakladatelstvi. Groot, A. Willem De 1965 "Classification of cases and uses of cases", in: Morris Halle (compiler), For Roman Jakobson. The Hague: Mouton, pp. 187-194. Grumet, Joanne Sher 1983 The Lexical Genitive Case Hypothesis. [Unpublished Ph.D. Dissertation, New York University.] Hajieova, Eva et al. 1979 "Transducing components of functional generative description", Eksplicitnoe opisanie jazyka i avtomaticeskaja obrabotka V. Halle, Morris (compiler) 1965 ForRomanJakobson. The Hague: Mouton. Havranek, Bohuslav & Alois Jedlicka 1960 Ceskd mluvnice. Prague: Statnf pedagogicke nakladatelstvi. Heim, Michael 1982 Contemporary Czech. Columbus: Slavica. Hjelmslev, Louis 1935 [1937] "La categoric des cas", Acta Jutlandica, Vol. 7, No.1; Vol. 9, No.2. Isacenko, Aleksandr. V. 1965
Grammaticeskij stroj russkogo jazyka v sopostavlenii s slovackim, Cast' 2: Morjologija. Bratislava: Izdatel'stvo Slovackoj akademii nauk.
Jakobson, Roman O. 1936 "Beitrag zur allgemeinen Kasuslehre: Gesamtbedeutungen der russischen Kasus", Travaux du Cercle linguistique de Prague VI, pp. 240-288. 1958 "Morfologiceskie nabljudenija nad slavjanskim skloneniem", in: American Contributionsto the Fourth International Congress of Slavicists. The Hague: 's Gravenhage, pp. 127-156. [1971] [reprinted in: Selected Writings, Vol. 2 (second edition). The Hague: Mouton, pp. 23-71]. [1971] [reprinted in: Selected Writings, Vol. 2 (second edition). The Hague: Mouton, pp. 154-183]. 1971 "Boas' view of grammatical meaning", in: Selected Writings, Vol. 2 (second edition). The Hague: Mouton, pp. 489-496. Jakobson, Roman & Krystyna Pomorska 1988 Dialogues. Cambridge: MIT Press. Janda, Laura A.
1986
A semantic analysis of the Russian verbal prefixes ZA-, (Slavistische Beitrage 192). Munich: Sagner.
PERE-, DO-,
and OT-.
"The mapping of elements of cognitive space onto grammatical relations: An example from Russian verbal prefixation", in: Brygida Rudzka-Ostyn (ed.), 323-339. 1988b "Pragmatic vs semantic uses of case", Chicago Linguistic Society 24: 189-202. Jespersen, Otto 1924 Thephilosophy ofgrammar. London and New York: Allen, Unwin, Holt & Co. [1965] [reprinted: New York: W.W. Norton]. Keenan, Edward L. 1985 "Passive in the world's languages", in: Timothy Shopen (ed.), Language typology and syntactic description, Vol. 1. Cambridge: Cambridge University Press, pp. 243281. Kilby, David A. 1977 Deepand superficial casesin Russian. (Specimina Philologiae Slavicae 14; Beitrage zur Kasusgrammatik der slawischen Sprachen 2). Munich: Sagner. "A functional approach to case in Russian", Papers in Slavonic Linguistics 1: 731982 90. 1986 "The instrumental in Russian: On establishing a consensus", in: Richard D. Brecht & James D. Levine (eds.), Case in Slavic. Columbus: Slavica, pp. 323-337. Koneena, Dana. 1969 "0 vyznamech paduv soucasne ceStin~. zejmena v mfstnfch urcenfch", Slovo a slovesnost 30: 347-356. Kottum, Steinar Egil 1979 A semantic distinctive feature analysis of the Polish case system with a comparison to Russian. [Unpublished Ph.D. Dissertation, Indiana University.] Kralfkova, K vetoslava 1977 Problematika pasiva pii explicitnim popisu ceStiny. [Unpublished MS] Kucera, Vladimir 1966 "K adverbializacii russkogo tvoritel'nogo vremeni", Ceskoslovenskd Rusistika 11: 97-102. Kurytowicz, Jerzy 1960 Esquisses linguistiques. Wroclaw-Krakow: Ossolineum. Lakoff, George 1977 "Linguistic gestalts", Chicago Linguistic Society 13: 236-286. 1988a
223
222 "Categories", in: Linguistic Society of Korea (eds.), Linguistics in the morning calm. Seoul: Hanshin, pp. 139-193. 1986 Cognitive semantics, Berkeley Cognitive Science Report No. 36. 1987 Women,fire, anddangerous things: What categories reveal aboutthe mind. Chicago: University of Chicago Press. Lakoff, George & Mark Johnson 1980 Metaphors we live by. Chicago: University of Chicago Press. Langacker, Ronald W. 1986 Foundations of cognitive grammar. Vol. I: Theoretical prerequisites. Stanford: Stanford University Press. 1982
1987
Transitivity, case, and grammatical relations: A cognitive grammar prospectus. Duisburg: L.A.U.D.
1988 Foundations of cognitive grammar. Vol. II. [Unpublished MS.] Leska, Oldrich n.d. Substantivni determinace. [Unpublished MS.] Levine, James S. 1984 "On the dative of possession in contemporary Russian", Slavic and East European Journal 28: 493-501. 1986 "Remarks on the pragmatics of the "inalienable dative' in Russian", in: Richard D. Brecht & James S. Levine (eds.), Case in Slavic. Columbus: Slavica, pp. 437-451. Lindner, Susan 1981 A lexico-semantic analysis of verb-particle constructions with UP and OUT. [Unpublished Ph.D. Dissertation, University of California at San Diego.] Linguistic Society of Korea (eds.) 1982 Linguistics in the morning calm. Seoul: Hanshin. Magner, Thomas F. & William R. Schmalstieg (eds.) 1970 Baltic linguistics. University Park: Pennsylvania State University Press. Malaxovskaja, L.S. 1958 "Tvoritel'nyj vremeni", in: Samuil B. Bernstein (ed.), Tvoritel'nyj padet v slavjanskixjazykax. Moscow: Izdatel'stvo Akademii Nauk SSSR, pp. 222-245. Maldonado, Ricardo 1987 Accidental reflexives. [Unpublished MS] Mel'cuk, Igor A. 1986 "Toward a definition of case", in: Richard D. Brecht & James D. Levine (eds.), Case in Slavic. Columbus: Slavica, pp. 36-85. Mervis, Carolyn & Eleanor Rosch 1981 "Categorization of natural objects", AnnualReview of Psychology 32: 89-115. Miller, James E. 1986 "A third look at the second dative", in: Richard D. Brecht & James D. Levine (eds.), Case in Slavic. Columbus: Slavica, pp. 296-311. Moore, Timothy E. (ed.) 1973 Cognitive development and the acquisition of language. New York: Academic Press. Mrazek, Roman 1964 Sintaksis russkogo tvoritel'nogo. (Opera Universitatis Purkynianae Brunensis. Facultas Philosophica 94). Prague: Statnf pedagogicke nakladatelstvL Naughton, James 1987 Colloquial Czech. London and New York: Routledge and Kegan. Nikonov, Vladimir A. 1961 "Borba padezej", International Journal ofSlavic Linguistics and Poetics 4: 13-3'3.
Nunberg, Geoffrey 1979 "The non-uniqueness of semantic solutions: Polysemy", Linguistics and Philosophy 3: 143-184. Oliverius, Zdenek 1972 "A contribution to the semantic analysis of Russian affixal morphemes. (On the material of Russian cases and prefixes)", in: Dean S. Worth (ed.), The Slavic word: proceedings of the International Slavistic Colloquium at UCLA. The Hague: Mouton, pp.96-119. Paprotte, Wolf & Rene Dirven (eds.) 1983 The ubiquity of metaphor. Amsterdam: Benjamins. Pit'ha, Petr n.d. "Remarks on the description of possessivity", Papers in Computational Linguistics, 129-137. 1971 "Existuje dativ posesfvnf?" Slovo a slovesnost 32: 301-311. Plewes, Stanley Frank 1977 Invariance and case function in Czech and Russian. [Unpublished Ph.D. Dissertation, Princeton University.] Popova, Zinaida D. 1970 "K teoriii padeznogo znacenija", Voprosy jazykomanija 4: 92-101. 1974 "Prostorecnoe upotreblenie padeznyx form i literaturnaja norma", in: Zolotova, Galina A. (ed.), Sintaksis i norma. Moscow: Nauka, pp. 176-186. Potebnja, Aleksandr A. 1888 Iz zapisokpo russkoj grammatike 1-2. Voronez: Tip. N. D. Goldstein. [1958] [reprinted: Moscow: Gosudarstvennoe ucebno-pedagogiceskoe izdatel'stvo]. Prud'homme, David Henri 1981 The Russian prepositionless instrumental and case theory. [Unpublished Ph.D. Dissertation, University of Alberta.] Rosch, Eleanor 1973a "Natural categories", Cognitive Psychology 4: 328-350. 1973b "On the internal structure of perceptual and semantic categories", in: Timothy E. Moore (ed.), Cognitive developments and the acquisition of language. New York: Academic Press, pp. 111-144. Rothstein, Robert A. 1986 "Equation vs. ascription: The nominative/instrumental opposition in West Slavic", in: Richard D. Brecht & James D. Levine (eds.), Case in Slavic. Columbus: Slavica, pp. 312-322. Rudzka-Ostyn, Brygida 1983a Cognitive grammar and the structure of Dutch UIT and Polish WY. [Unpublished MS]
1983b
"Metaphoric processes in word formation: The case of prefixed verbs", in: Wolf Paprotte & Rene Dirven (eds.), The ubiquity of metaphor. Amsterdam: Benjamins, pp.209-241. Rudzka-Ostyn, Brygida (ed.) 1987 Topics in cognitive linguistics. Amsterdam: Benjamins. Sangster, Rodney B. 1982 Roman Jakobson and beyond: Language as a system ofsigns. Berlin: Mouton. Schlesinger,Izchak 1979 "Cognitive structures and deep structures: The case of the instrumental", Journal of Linguistics 15: 307-324.
224
225
Schooneveld, Cornelis H. van
1978
Semantic transmutations: Prolegomena to a calculus of meaning. Vol. 1: The cardinal semantic structure of prepositions, cases, and paratactic conjunctions in contemporary standard Russian. Bloomington: Physsardt.
"Jakobson's case system and syntax", in: Richard D. Brecht & James D. Levine (eds.), Case in Slavic. Columbus: Slavica, pp. 373-385. Schupbach, Richard D. 1986 "Grammatical innovations in technical Russian: Decline of the 'instrumental of means' ", Slavic andEastEuropean Journal 30: 487-508. Sgall, Petr, Eva Hajicova & Jarmila Panevova 1986 The meaning of the sentence in its semantic and pragmatic aspects. Prague: Academia. Shepard, Roger N. 1978 "The mental image", American Psychologist 33, 125-137. Shopen, Timothy (ed.) 1985 Language typology and syntactic description. Vol. 1. Cambridge: Cambridge University Press Skalicka, Vladimir 1950 "Poznamky k theorii padu", Slovo a slovesnost 12: 134-152. Smilauer, Vladimir 1966 Novoceskd skladba. Prague: Stamfpedagogicke nakladatelstvi. 1972 Nauka 0 eeskemjazyku. Prague: Statnt pedagogicke nakladatelstvi. Smith, Michael B. 1985 "Event chains, grammatical relations, and the semantics of case in German", Chicago Linguistic Society 21: 388-407. 1987 The semantics of dative and accusative in German: An Investigation in cognitive grammar. [Unpublished Ph.D. Dissertation, University of California at San Diego.] Serensen, Hans Christian 1957 Studies on casein Russian. Copenhagen: Kommission hos Rosenkilde og Bagger. Soya, Milos 1962 A practical Czech course. Prague: Stamf pedagogicke nakladatelstvi. Staniseva, D.S. 1958a "Tvoritel'nyj sociativnyj", in: Samuil B. Bernstejn (ed.), Tvoritel'nyj padet v slavjanskixjazykax. Moscow: Izdatel'stvo Akademii Nauk SSSR, pp. 41-75. 1958b "Tvoritel'nyj instrumental'nyj", in: Samuil B. Bernstein (ed.), Tvoritel'nyj padet v slavjanskixjazykax. Moscow: Izdatel'stvo Akademii Nauk SSSR, pp. 76-126. 1958c "Tvoritel'nyj sovokupnosti", in: Samuil B. Bernstejn (ed.), Tvoritel'nyj padet v slavjanskixjazykax. Moscow: Izdatel'stvo Akademii Nauk SSSR, pp. 193-200. Starosta, Stanley 1976 "A place for case", Language Learning 26: 1-36. Svedova, Natalja Ju.(eds.) 1970 Grammatika sovremennogo russkogo literaturnogo jazyka. Moscow: Akademija Nauk SSSR. 1982 Russkaja grammatika. Moscow: Nauka. Sweetser, Eve Eliot 1984 Semantic structure and semantic change: A cognitive linguistic study of modality, perception, speech acts, and logical relations. [Unpublished Ph.D. Dissertation, University of California at Berkeley.] 1986
Talmy, Leonard
1986
The relation of grammar to cognition. Berkeley: Berkeley Cognitive Science Report
No. 45. Tesitelova, Marie, Jan Petr & Jan Krilik
1986
Retrogrddni siovnik spisovne eestiny. Prague: Academia.
Tesniere, Lucien
1959 Elements de syntaxe structurale. Paris: Klincksieck. Toops, Gary H. 1987 "Russian contextual causatives", Slavic andEastEuropean Journal 31: 595-611. Townsend, Charles E. . 1981 Czech through Russian. Columbus: Slavica. Ulicny, Oldfich 1984 "Instrumental v strukture ceske vety", Informacai bulletin Usttedni knihovny Pedagogicke fakulty Univerzity Karlovy, Supplementum XVII. Vinogradov, Victor V. 1972 Russkij Jazyk: (Second edition). Moscow: Vyssaja skola. Wierzbicka, Anna 1980a The casefor surface case. (Linguistica Extranea Studia 9) Ann Arbor: Karoma. 1980b Lingua mentalis: Thesemantics of natural language. Sydney: Academic Press. 1986 "The meaning of a case: A study of the Polish dative", in: Richard D. Brecht & James D. Levine (eds.), Case in Slavic. Columbus: Slavica, pp. 386-426. Worth, Dean S. 1958 "Transform analysis of Russian instrumental constructions", Word 14: 247-290. Xodova, Kapitolina I. 1958a "Tvoritel'nyj padez v stradatel'nyx konstrukcijax i bezlienyx predlozenijax", in: Samuil B. Bernstejn (ed.), Tvoritel'nyj padet v slavjanskix jazykax. Moscow: Izdatel'stvo Akademii Nauk SSSR, pp. 127-158. 1958b "Tvoritel'nyj prevrascenija i sravnenija", in: Samuil B. Bernstejn (ed.), Tvoritel'nyj padet v slavjanskix jazykax. Moscow: Izdatel'stvo Akademii Nauk SSSR, pp. 181192. Zemskaja, Elena A. 1983 Russkaja razgovornaja rec'. Moscow: Russkij jazyk. Zolotova, Galina A. (ed.) 1974 Sintaksis i norma. Moscow: Nauka.